Interface Driven and Bio-Mimetic Design of 3D Hybrid Materials

Home , Exfoliation joint

2018

ABSTRACT

Interface Driven and Bio-mimetic Design of 3D Hybrid Materials

Peter SamoraOwuor

 What role does the interfacial region plays in macro-scale materials

properties?

 Is the same effect of interfacial region at macro-scale applies to the nano-

scale materials?

 Is there a means to modify the interface region to assemble 3D hybrid

structures?

 What are the resulting applications of such design in materials?

To address the aforementioned questions, novel synthetic and biomimetic strategies were employed. The first detour of the thesis delves into the chemical process where functionalization and freeze-drying methods are used to fabricate porous carbon nanotubes (CNT) with self-stiffening behavior. The chemical approach is

then applied to zero-dimensional SiO2 nanoparticles to fabricate three-dimensional nanostructures with improved fire-retardant capability. Next, the thesis explores the physical methods in assembling 3D structures where graphene oxide foam is chemically and physically reinforced with polymer molecule to fabricate an oil absorption and electrical resistant foam. The thesis further develops a new method to functionalize hexagonal boron nitride (h-BN) to enable their networking forming property resulting in high porous foam for CO2 absorption. The above solutions all relates to what is referred to as ‘hard interface’ therefore there was a need to explore ‘mobile interface’ like those found in nature. In this regard, a new area of study was developed; solid-liquid composite in macro-scale materials. Here, the thesis presents two new approaches; high damping composite by addition of liquid metal in a polymer matrix and optical and stiffness switching of a phase change composite. Finally, the thesis attempts to combine the two interfaces in hybrid materials. The most important contribution of this thesis is the new techniques which can be used to design advanced composites. Furthermore, a new subset of solid-liquid composites which have never been looked at in terms of mechanical properties is brought-forth. Finally, the peer-reviewed papers published should form a basis for future scientists with plans to pursue this field.

Acknowledgments

Since joining Rice University in the fall of 2014, I have received support and encouragement from many different people and here I would like to acknowledge those people. First, I would like to express gratitude to my major advisors: Professor

Pulickel M. Ajayan and Professor Jun Lou for giving me the opportunity to work on this research project, their support and guidance and fruitful discussions. I wish to express my sincere thanks to Professor C. Fred Higgs III for accepting to be in my committee and his constructive discussion on my work.

I would also like to thank Dr. Chandra Sekhar Tiwary for his advice and great help he offered to me; training on instruments, advice and lots of valuable literature among others.

I would like to thank my colleagues at MSNE department for their help and friendship and for creating conducive atmosphere that broadened my understanding of other cultures. I would also like to thank MSNE staffs for their tireless help.

Special thanks go to Mr. & Mrs. Curry for their ever ending support and lastly to my parents and whole family for their encouragement.

Contents

Acknowledgments ...... iv

Contents ...... v

List of Figures ...... ix

Nomenclature ...... xviii

I. Overview ...... 19

II. Introduction ...... 21 2.1. Role of Interfaces in Materials Design ...... 21 2.2. Materials Design Approaches...... 24 2.2.1. Liquid exfoliation ...... 24 2.2.2. Mechanical exfoliation ...... 26 2.2.3. Chemical vapor deposition ...... 27 2.2.4. Sol-gel method ...... 28 2.3. Synthesis of Graphene Oxide ...... 30 III. Role of Interface Engineering in Assembling 3D Structures from Zero and One- dimensional Materials ...... 32 3.1. Background ...... 33 3.1.1. Introduction ...... 33 3.2. Self-Stiffening Behavior of Reinforced Carbon Nanotubes Spheres ...... 33 3.2.1. Background ...... 34 3.2.2. Preparation of CNT Sphere and Composite...... 35 3.2.3. Physical Characterization ...... 36 3.2.4. Simulation Details ...... 36 3.2.5. Results and Discussion ...... 37 3.3. Scalable Bottom-up Assembly of Ultra-Low Density Multifunctional Three-

Dimensional Nanostructures from SiO2 ...... 45 3.3.1. Background ...... 45

3.3.2. Synthesis of SiO2 3D Nanostructures ...... 46 3.3.3. Physical characterizations...... 47

3.3.4. Simulation Details ...... 48

3.3.5. Mechanical response of 3D SiO2 Structures ...... 49 3.4. Conclusion ...... 64

IV. Surface Modification of 2D Materials into 3D Hybrid Structures ...... 65 4.1. Background ...... 66 4.2. High Toughness in Ultra low Density Graphene Oxide Foam ...... 66 4.2.1. Background ...... 66 4.2.2. Preparation of interconnected Graphene Oxide Foam ...... 68 4.2.3. Physical Characterization ...... 68 4.2.4. Simulation Details ...... 70 4.2.5. Result and Discussion ...... 71 4.3. Oxidation of hexagonal boron nitride (h-BN) nano-sheets to synthesize reactive sites for assembling of 3D structures ...... 82 4.3.1. Background ...... 82 4.3.2. Oxidation of h-BN ...... 84 4.3.3. Preparation of oxidized h-BN/polyvinyl alcohol foam ...... 84 4.3.4. Physical Characterization ...... 85 4.3.5. Simulation Details ...... 86 4.3.6. Properties of Oxidized h-BN/PVA Foam ...... 87 4.4. Conclusion ...... 103

V. Bio-mimicking of ‘Mobile’ Interfaces from Nature: Liquid Metal Composite ...... 104 5.1. Background ...... 105 5.1.1. Introduction ...... 105 5.1.2. Background to solid-liquid composite ...... 105 5.2. Preparation of Gallium/PDMS Composite ...... 108 5.3. Physical Characterization ...... 109 5.3.1. Mechanical Test ...... 109 5.3.2. Optical Observation of Gallium Spheres ...... 109 5.4. Analytical Development ...... 110 5.4.1. Analytical Formulation ...... 110 5.4.2. FE-SPH Modeling Scheme ...... 114 vii

5.5. Mechanical Behavior of Liquid Metal Composite ...... 118 5.6. Analytical model based prediction ...... 126 5.7. Finite Element (FE) Modeling with Smooth Particle Hydrodynamics (SPH) Coupling for Liquid Inclusion ...... 129 5.8. Conclusion ...... 134 VI. Bio-mimicking of Multi-functionality in Nature: High Stiffness Polymer Composite with Tunable Transparency ...... 135 6.1. Background ...... 136 6.1.1. Introduction ...... 136 6.1.2. Background on Multi-functional Materials ...... 136 6.2. Preparation of Phase Change Composite...... 138 6.3. Physical Characterization ...... 139 6.4. Simulation Details ...... 140 6.5. High Stiffness and Tunable Transparency ...... 141 6.6. Conclusion ...... 157

VII. Interface Engineering Applied to Hybrid Building Blocks ...... 158 7.1. Introduction ...... 158 7.1.1. Background to Hybrid Building Blocks ...... 158 7.2. Preparation of Hybrid 3D Materials ...... 160 7.2.1. Synthesis of polyalbumen (cross-linked egg white) ...... 160 7.2.2. Synthesis of polyalbumene (Graphene Oxide/Albumen) ...... 160 7.2.3. Synthesis of Albumin/Hydroxyapatite Composite ...... 161 7.3. Physical Characterization ...... 161 7.4. Computation Details ...... 162 7.5. High Strength and Tough Cross-linked Egg White Albumin ...... 167 7.6. Hybrid Graphene Oxide/Albumin Layered Material ...... 175 7.7. Polymerized Egg White Hydroxyapatite Nacre-like Material ...... 181 7.8. Conclusion ...... 189

VIII. Summary ...... 191

Reference ...... 194 viii

List of Figures

Figure 2.1 – (a) Interface in Nanocomposite showing a clear boundary between a matrix and nano-fillers. A transition region can exist when the nano-fillers are functionalized...... 22

Figure 2.2 – Biological materials with inherent ‘mobile’ interfaces enabling them to be multifunctional ...... 23

Figure 2.3 – Liquid exfoliation of layered materials (a) Ion intercalation. (b) Ion exchange or de-intercalation. (c) Solvents sonication.Reprinted from ref(23) with permission from America Chemical Society(check) ...... 26

Figure 2.4 – Illustration of a typical CVD process. (1) Mass transportation of reactants from the gas/liquid source to the substrate surface; (2) Adsorption of reactants on surfaces; (3) Chemical reaction on the substrate to form the film; (4) Desorption of by-products from the substrate; and (5) pumping away of the by-products and unreacted reactants...... 28

Figure 2.5 – Schematic representation of SiO2 particles synthesis by Sol-gel method ...... 30

Figure 3.1 – Morphology of CNT Spheres. (a) CNT sphere/PDMS showing structural integrity in presence of polymer. (b) Optical image of CNT sphere embed in PDMS matrix. (c) Molecular representation of CNT sphere showing function groups (d) Scanning electron microscope (SEM) images show the polymer get into the spheres and is well distributed. (e) CNT spheres after freeze drying. (f) TEM images showing multi wall CNT in close proximity due to strong van der Waal forces between individual CNT and functionalization which induces carboxylic functional groups...... 38

Figure 3.2 – Mechanical behavior of CNT spheres/PDMS (a) Self-stiffening behavior of nanocomposites due to dynamic loading. Pure PDMS does not show any self-stiffening characteristic. (b) Increasing stiffness with relaxation, here samples are loaded in dynamic load and relaxed before another cycle of loading. There is a high stiffness increase after first cycles and reduces as the number of dynamic loading increases. (c) Static stress/strain test showing high modulus of CNT spheres/PDMS. (d) High modulus of CNT spheres/PDMS nanocomposite at low and high temperature. (e) Elastic behavior of CNT spheres under load as observed from optical images...... 42

Figure 3.3 – Simulation of compressive strain of CNT spheres/PDMS nanocomposite. (a) Mises stress (units: Pa) and displacement (units: m) distributions of CNT sphere-reinforced PDMS for a 0.20 strain. (b) Mises stress (units: Pa) and displacement (units: m) distributions of steel sphere- reinforced PDMS for a 0.20 stain. (c) Mises stress (units: Pa) distribution for the nanocomposite inclusions, steel sphere (top) and CNT sphere (bottom) for the strains of 0.05, 0.10, 0.15, and 0.20. (d) Maximum value of Mises stress in the inclusion for a range of strains from 0.05 to 0.20...... 45

Figure 3.4 – Fabrication methodology for functionalized (fSiO2) material. (a)Hydrolyzation of silane couplings in water to form reactive silanol groups, addition of SiO2 nanoparticles followed by vacuum filtration and later heating at 50°C for 3 hours. (b) Proposed mechanism of reaction between silanol groups with SiO2 nanoparticles, reactive silanol groups attach themselves to the oxygen molecules on the SiO2. Depending on silane coupling, bi-, tri- etc. connection can be achieved. This connection is responsible for covalent bonds between nanoparticles resulting in a random three-dimensional structure. .. 52

Figure 3.5 – Mechanical behavior of functionalized SiO2. (a) Discernable stiffening of functionalized SiO2 in a cyclic compressive test, functionalized

SiO2 has a stiffness more than four-fold compared to non-functionalized

SiO2sample. After loading, non-functionalized sample completely collapsed into small pieces while functionalized and GO reinforced samples show no observable damage (insets). The addition of graphene oxide, greatly enhances the stiffness to more than an order of magnitude in comparison to non- functionalized SiO2. (b) Load displacement curves showing better elastic deformation behavior of fSiO2/GO and fSiO2. Non-functionalized sample fails at very low loads (0.01N). (c) Frequency controlled test to understand the stability of functional groups on the SiO2 at high loading cycles. As expected at high frequency, non-functionalized SiO2 lose its stability unlike functionalized and GO reinforced samples which exhibit constant stiffness without noticeable loss in stiffness. (d) Nano-indentation test on two types of silane coupling agents and non-functionalized SiO2 nanoparticles. APTS (i) functionalized SiO2 takes extremely large force compared to N1-(3-Trimethoxysilylpropyl) diethylenediamine (ii) and non-functionalized SiO2 nanoparticles (iii and inset). (e, f) SEM images in-situ during nano-indentation test. Non- functionalized SiO2 structure breaks at very low loads (e), once the crack develops, it propagates in all directions. On the other hand, functionalized

SiO2 structure does not show any cracks development (f)...... 58 xi

Figure 3.6 – Computer models for hSiO2, fSiO2 and fSiO2/GO 3D amorphous

SiO2 structures. Generation of fSiO2/GO, a 3D structure combining APTS functionalized amorphous SiO2nano-particles and both hydroxyl and epoxy functionalized graphene nano-ribbons (GO-h2e2) from a 3D structure made of

APTS functionalized amorphous SiO2 (fSiO2). The 3D structure fSiO2 is generated connecting adjacent APTS functional groups belonging to different fSiO2 nanoparticles which are generated attaching APTS functional groups to the surface of a pristine amorphous SiO2 nanoparticle (pSiO2). Hydrogen passivated to obtain the hydrogen passivated amorphous SiO2 nanoparticle

(hSiO2 or simply SiO2) used to obtain the equivalent 3D structure, the pristine amorphous SiO2 nanoparticle (pSiO2) is generated carving a sphere out the an amorphous SiO2 obtained through annealing a chunk of crystalline SiO2 structure (α-quartz)...... 61

Figure 4.1 –Structural stability of pGO composite. (a) Proposed structural pGO model. (b) Freeze-dried graphene oxide foam and its scanning electron microscope images showing the typical interconnected networks of individual (c, d). (e) pGO exhibits structural stability without collapsing due to cross- links among GO layers. Corresponding SEM images reveal that nano-sheets are wetted by PDMS (f, g). (h) Three and two dimensional atomic force microscope (AFM) mapping showing homogeneously distribution of PDMS within the pGO (i) FTIR spectra GO and pGO ...... 73

Figure 4.2 – pGO mechanical properties. (a) High modulus, tensile strength and improved strain to failure of pGO foam. Interconnected GO sheets result in better adhesion hence high tensile strength and toughness. Inset - optical image of collapsed sample indicate elastic failure. (b) Compressive test shows pGO to have very high compressive strength compared to pure GO foam (inset). (c) Load vs stiffness results reveal high stiffness with increasing load for the pGO unlike pristine GO foam. (d) Cyclic loading showing stiffening behavior of pGO as the number of cycles increases. (e) Load relaxation tests showing increases of stiffness with subsequent loading, here, the sample is loaded in a dynamic manner, test stopped to let the sample relax and then loaded again after a certain number of loading cycles. (f) Temperature controlled test to study the storage modulus. High modulus of pGO at low and high temperatures. Physical and chemical cross-links induced between the interphase make the material highly resistance to dynamic load...... 75

Figure 4.3 – Morphology of fractured pGO. (a) Better wetting of the GO sheets by the PDMS. (b, c) High magnification images PDMS on the thin GO nano- xii sheets, no noticeable failure of the GO sheets shown by strains lines in response to load. (d) Nano-indentation probe of pGO by indenting on a bundle of engineered surfaces of GO area exhibiting high load bearing compared to pure GO foam. (e) A shear like test by pulling one sheet from each. Interphase engineered sheets are harder to separate than those no surface treatment. (f) Comparison of compressive strength of pGO with other lightweight materials...... 78

Figure 4.4 – Schematic of pGOstructural models for three different considered arrangements. Each snapshot in the Initial stage represents the thermalized structure (at 300 K) obtained from reactive molecular dynamics simulations (MD). The further stage snapshots were taken from a stress-strain MD simulations before structural failure occur. Notice the ductile-like deformation of the pGO unlike what happens with GO, which breaks into pieces exhibiting brittle behavior. (d) Relative displacement as a function of the applied force on the outer graphene oxide sheets. This figure shows that the presence of the polymer makes pGO more resilient to structural deformations in comparison to the pristine GO...... 81

Figure 4.5 – Characterization and morphology of h-BN/PVA foams. (a) Schematic representation depicting the h-BN nanosheets connected by polyvinyl alcohol (PVA) molecules. Typical chains contain 30 monomers (~ 6.2 nm chain length). The PVA molecules act like a glue to link the nanosheets through van der Waals interactions between the hydroxyl groups in the PVA and Boron/Nitrogen atoms in the nanosheets. (b) Structural stability of the h- BN/PVA foam. (c) Mechanical stability of the lightweight foam carrying a vial without any visible degradation. (d) Stable structure of the foam when subjected to liquids. (e) Pristine foam disintegrates in presence of water while h-BN/PVA maintains its stable interconnected morphology (f-g) Low magnification and high magnification image of the foam...... 88

Figure 5.0 – Radial distribution function (RDF) of (a) PVA polymer into the adsorbed h-BN layer, (see Fig. 1a); (b) CO2 in the gas phase into the adsorbed PVA/h-BN composite, (see Fig. 4c). Results from reactive molecular dynamics simulations at 300K. Notice the high affinity between CO2 molecules and the PVA polymer (see text for discussions)...... 102

Figure 5.1 – Mechanical Behavior of PDMS-Ga. (a) Natural composites which utilizes liquids inclusions. (b) Load variation as a function of compressive displacement for PDMS-Ga (blue curve) and pristine PDMS (black curve). PDMS shows classical polymer behavior while PDMS-Ga shows metal like xiii behavior with high modulus. (c, d) Contact angle tests showing high hydrophobicity between gallium and PDMS. The high hydrophobicity is maintained for a long time showing there is no chemical interaction between the two materials. (e) Dynamic compressive tests illustrating the stiffening phenomena of PDMS due to the addition of liquid metal gallium. Stiffness is increased almost four times compared to pristine PDMS matrix. (f) High stiffness of PDMS-Ga is also observed when gallium spheres are arranged in a PDMS matrix, three layers of gallium spheres exhibit high stiffness compared to two layers...... 120

Figure 5.2 – Response of the PDMS-Ga composite to Load. (a) Load vs displace curve depicting high modulus of the PDMS-Ga. (b - d) Gallium spheres changes shape to an ellipse under load and spring back to spherical shape when unloaded. (e) Storage modulus increases with temperature in PDMS-Ga (f, g) Digital images after impact test. High damping capability of PDMS-Ga where fractures grow in lateral direction. (h) Schematic of an impact test. High damping during impact of PDMS-Ga where fractures grow in lateral direction. (i) Optical images after impact loading showing fracture field near individual Ga. The impact energy is absorbed by liquid Ga and does not allow the load transfer to the base of PDMS...... 125

Figure 5.3 – Airy's stress function based analytical modeling. (a) Schematic diagram of the PDMS matrix and Gallium inclusion subjected to uni-axial compression. (b) Variation of specific energy density of the inclusion with respect to applied stress. (c) Variation of specific energy density of the PDMS matrix with respect to applied stress...... 129

Figure 5.4 – Comparison of stress-strain behaviour of obtained from experimental data, FE-SPH model and Ariy's stress function based analytical model. Peak near the point A1 is due to the impact response. Strain after 0.1 is considered to be large strain behaviour. (b) Comparison of impact energy transmission for inclusion material as PDMS, Gallium, Air and Gallium with 4 spheres having same volume as single sphere. (c - f) Von-Misses stress distribution for inclusion material as PDMS, Gallium, Air and Gallium with 4 spheres having same volume as single sphere at time t=0.75ms, velocity of the impactor is 3.428 m/s...... 131

Figure 5.5 – Kinetic energy study of the Composite. (a - b) Development of kinetic energy and strain energy in PDMS matrix, Gallium Inclusion and in PDMS-Ga Composite system. Kinetic energy and the strain energy are normalized with the impact energy provided. (c - d) Von-Misses stress xiv distribution in PDMS matrix and the resultant velocity of the Gallium particles in the core at time tA= 0.75 ms and tB= 1.25 ms respectively. (e) Comparison specific total energy density obtained from of Eshelby theory, present Airy's stress function based analytical model and FE-SPH simulation model for varying core size. Variation of increasing viscosity in the analytical model is also shown. Based on inclusion interaction behaviour with the matrix three region can be defined, region A where surface tension of the inclusion play vital role. Region B where viscosity of the inclusion plays important role in energy transfer. Region C where Fluid flow of the inclusion derives the energy transfer mechanism...... 133

Figure 6.1 –Optical switching of the pc-composite under heat contact heating (a) and non-contact heating by applied magnetic field (M) (b). When the material is subjected to the mentioned conditions, heat melt the wax particles within the polymer matrix resulting in similar refractive index of the two materials hence the optical switching to transparent observed...... 142

Figure 6.2 –Optical transmittance control of pc-composite. (a) Gradual increase in optical transmittance as temperature rises with a ‘complete’ transparency of the polymer blend at a peak transmittance of 80% at 45°C (bar graph). (b) Optical switching behavior of a pipe made of up pc-composite showing its ability to change into transparency in presence of hot water and return to its initial condition when immersed in cold water bath. (c) Optical switching behavior of the pc-composite with 0.75 mass% loading of magnetic nanoparticles with application and removal of AC magnetic field, showing the transition from opaque to translucent on application of AMF of 4kA/m and frequency 375kHz. (d) Gradual increase in optical transmittance with time by a noncontact heating through the application of AFM. Up to ~ 50% transparency was observed, the sample returned to its initial (opaque) condition when the AC field was removed. The ‘0‘ and ‘1‘ represent the application and removal of AFM...... 146

Figure 6.3 –TGA (a) Weight loss with temperature showing high stability of the pc-composite but with reduced onset decomposition temperature. (b) Differential thermal analysis (DTA) showing two peaks for pristine wax associated with melting and sublimation which are absent in the blend composite. The melting peak is at 60°C while sublimation peak is at 360°C. 148

Increase of both stiffness and optical transmittance of the composite as the temperature increases...... 150

Figure 7.1 – Morphology of poly-albumen (a) Egg white is extracted from an avian egg and a cross-linker added to polymerize it resulting in a flexible and tough material. (b) Large scale (cm) polymerized structure (c) Optical and SEM images of the crystalized pristine egg white characterized by smooth surfaces with visible cracks on the surface. (d) Optical images of polybumen showing homogeneous ridges spread all over the sample forming sheets like cramps. Layered structured of polybumen as seen on the SEM. The individual cross-linked sheets appear to be closely connected to each other...... 168

Figure 7.2 – Artistic visualization of ovalbumin/amine binding (b) Highest- affinity binding pockets (colored sites) on the gray ovalbumin surface complete with their binding energies indicated on top of the colored bars representing them, the units are kcal/mol. Site # 5 is highlighted to be the highest binding site and the structure of Glutamic Acid is shown as the most abundant amino acid. (c) Glutamic acid/glutamic acid peptide bonding. (d) Glutamic acid radical/glutamic acid radical bonding through condensation. (e) Glutamic acid radical/amine bonding through condensation. Configurations I and II are made up of two identical five glutamic acid chains connected through their radicals and through cross-linker. The right chain moves downward 0.05 Å/ps while the respective highlighted interfacial bonds, β, are calculated as a function of the distance, δ, and travelled by the right chain until rupture. (f) Through initial/final bond differences, Y, and equilibrium/rupture δ differences, X, show the brittle nature of configuration I, the ductility and stiffening effect of configuration II...... 171

Figure 7.3 – Mechanical behavior (a) High compressive stiffness of egg white that breaks within few loading cycles unlike polybumen which sustain many loading cycles without showing any deformation. (b) High toughness and strength of poly-albumen. (c) Quantified toughness in tension mode for egg white, poly-albumen (5%) and poly-albumen (10%). (d) Stiffness increases with increase in frequency. (e) High stiffness with aging. (f) SEM surfaces of pristine egg white showing cracks propagating in all directions. (g) Bridge like network connecting the individual cross-linked sheets preventing brittle failure in poly-albumen. (h) AFM modulus mapping of pristine egg white. (i) Inhomogeneous high modulus distribution regions connected by low modulus areas which provide the elasticity observed in poly-albumen...... 174 xvi

Figure 7.4 – Polygrabumin hybrid material (a) GO is added to egg-white and amine-based cross-linker; diethylenetriamine (DETA) added. (b) Surface appearance where the material exhibit wrinkled surface typical of ductile materials, side view shows a layered-like appearance throughout the thickness of the material (c)...... 176

Figure 7.5 – Mechanical effect of polygrabumin through molecular dynamics simulations. Configurations I and II are made up of two identical five-glutamic acid chains connecting their R-groups through DETA (I), and through DETA and graphene oxide (II). The right chain moves downward 0.05Å/ps while the respective highlighted interfacial bonds, β, are calculated as a function of the distance, δ, and travelled by the right chain until rupture. a) Final/initial bond differences (Y), equilibrium/rupture δ differences (X) showing the stiffening impact of configurations I and II...... 178

Figure 7.6 – Mechanical response of the polygrabumin material. (a) Stress- strain curve showing high strength and ductility (b) Quantified ultimate strength. (c) Increase in stiffness with time, high stiffness is observed as the number of days increases. (d, e) Atomic force microscope (AFM) local modulus mapping of the sample, showing the distribution of GO, polygrabumin...... 180

Figure 7.7 – Synthesis of hydroxyapatite. (a) Bones are dried, soft tissue and fats removed by hydrogen peroxide (H2O2), selective etching is then done by immersing the bones in 30-40% concentration of citric acid (C6H8O7). (b) SEM of hydroxyapatite nanocrystals from bones showing ‘brick’ like morphology with high surface area. (c,d) TEM images of the crystals with rectangular like shapes and their polycrystalline nature from diffraction pattern (e)...... 182

Figure 7.8 – Fabrication of polymerized egg white and hydroxyapatite (HA) (a) An artistic visualization of the process adopted to synthesize HA/egg white, DETA/egg white is added to HA cured at 50˚C. (b) Hydroxyapatite/egg white composite with its layered structure (c). (d) Interconnected HA crystals by a matrix of polymerized egg white ...... 184

Figure 7.9 – The Mechanical characterization (a) Compression test showing high load carrying capability of HA/egg white with a characteristic step-wise failure. (b) Local in-situ loading exhibiting a high modulus of the sample. (c) AFM modulus mapping of HA with no polymer with a modulus of 0.6 to 8.9 GPa. (d) Extremely high modulus (30-180GPa) of reinforced HA. (e) In-situ xvii

SEM showing the failure of HA with no polymer unlike reinforced HA (f) which does not exhibit major failure...... 186

Figure 7.10 – Illustrative structure of hydroxyapatite (light-blue oval) interfacing with water molecules, DETA and normal/zwitterionic glutamic acid. Glutamic acid is taken here as an example of amino acid present in eggwhite. Coulomb (dash-dot blue lines) and hydrogen interactions (dash blue lines) at the interface play key roles in the composite’s toughness...... 189

Nomenclature

PDMS Polydimethylsiloxane

GO Graphene Oxide

HA Hydroxyapatite

CNT Carbon Nanotubes

GA Gallium

TGA Thermogravimetric Analysis

DMA Dynamical Mechanical Analysis

TEOS Tetraethylorthosilicate

Chapter 1

I. Overview

The discovery of Graphene, carbon nanotubes and subsequent other nanomaterials led to an explosion in research geared towards utilizing their intriguing mechanical, physical and chemical properties. While the physical properties of nanomaterials have been extensively explored, the assembly in a bottom-up approach to design hybrid 3D nanostructures by taking advantage of their interfacial properties still needs a deeper inquiry. This thesis scope is to explore different methodologies which can be employed to design 3D nanostructures by taking advantage of the interfacial interplay between building blocks through chemical, physical or combination of the two. Firstly, a brief introduction of the implications of interfacial region in materials is discussed (Chapter II), then detail study of the assembly of zero-dimensional silicon dioxide (SiO2) into 3D nanostructures and self-stiffening behavior of carbon nanotube spheres (Chapter III) 20

follows. Secondly, high toughness in lightweight graphene oxide foam and lightweight hexagonal boron nitride foam for CO2 absorption (IV) are discussed. The above discussion is mainly focused on what is normally referred to as 'hard' interfacial interaction. Therefore, 'mobile' interfacial region as normally observed in natural designs is mimicked for the preceding exploration where high stiffness in liquid metal/polymer composite (V) and controlled optical switching and high stiffness of a multifunctional composite is brought forth (VI). Having toured the landscape of conventional designs and borrowed some natural designs principles, the discussion of hybrid designs by a combination of natural and synthetic building blocks is attempted (VII). Finally, a synopsis of the whole thesis is presented and future directions suggested (VIII).

Chapter 2

II. Introduction

2.1. Role of Interfaces in Materials Design

Interfaces are pervasive in material systems and are accountable for mechanical, optical, electrical, magnetic and chemical properties. Interfaces play a critical role not only in synthetic but also biological materials. The importance of interfaces is prevalent from nano- to macro-scale in defining material's properties1.

Generally, the field of interface encompasses solid-solid phase boundaries, grain boundaries, free surfaces and solid-liquid interfaces. From a material design perspective, solid-solid phase and solid-liquid interfaces are two important parameters when it comes to assembling of individual building blocks to fabricate materials with improved properties. A distinct demonstration of the interfaces importance can be found in composite materials which are a combination of two or more materials to take advantage of the synergy of their desirable properties. For 22

instance, in fiber reinforcement composite the matrix provides a means for the transfer of the load while the fiber carries the load2–4. The result is a 'tough and strong' material. The properties are closely linked to the intrinsic properties of individual phases as well as the nature of the interface between the phases. In nanocomposites for example, the large surfaces areas as the result of the nanometer size of the phases are important in nanocomposites design1. In polymer nanocomposites the interface normally divides the organic and inorganic phases into two clear regions (see Figure 2.1). The inorganic nanoparticles reinforce the matrix forming weak bonds such as van der Waals or hydrogen across the hetero- interfaces of the two materials. However upon functionalization or modification of the inorganic, a transition region across the interface can exist (enlarged view of nanoparticle Figure 2.1). This transition region can result in strong bonding like covalent leading to stronger nanocomposites5–11. Majority of synthetic materials possess what is normally referred to as 'solid' interfaces.

Figure 2.1 – (a) Interface in Nanocomposite showing a clear boundary between a matrix and nano-fillers. A transition region can exist when the 23

nano-fillers are functionalized. Reprinted from ref (MRS) with permission from Materials Research Society (MRS).

Biological materials on the other hand use a complete different approach to design their interfaces. In majority of biological materials such nacre, skin membrane, inter-vertebral discs etc. (Figure 2.2) utilizes an extremely soft and mobile interface between hard components. This soft interface aids in transfer and quick distribution of load. It has been theorized the 'mobile' interfaces in biological materials allows them to be tough and strong which has remained a challenge in synthetic materials. In the following sections, the thesis will discuss the conventional top-down design methodology and compare it with the bottom-up method. Methods to synthesize nano-based based building blocks which make bottom-up assembly possible will also be addressed. Finally, a summary of nature- derived building blocks will be presented.

Figure 2.2 – Biological materials with inherent ‘mobile’ interfaces enabling them to be multifunctional

2.2. Materials Design Approaches

A broad classification of nanostructures synthesis divides methods into two; those that build from top-down and those that build from bottom-up. The top-down approach is characterized by miniaturizing or breaking down bulk materials specifically macro-crystalline structures while keeping the original integrity. Both of these methods have their advantages and disadvantages. A major problem with top- down method is the imperfection of surface structure and general damage to the significant crystallographic structure. The imperfections are problematic when high quality device are needed however this method leads to bulk production of nanomaterials cutting cost tremendously.

On the other hand, bottom-up approach builds nanostructures from their atomic scale constitutes such as molecules or atoms. The major advantage of bottom-up approach is the ability to tailor the structures and imperfections can greatly be avoided. While the synthetic advanced materials are adopting this method in droves recently, nature has perfected this method where all the growth takes advantage of it. Due to its versatility and high yield5,12–19, top-down method is the most widely used to synthesize nanoparticles. Major top-down methods to fabricate nanomaterials will be discussed first followed by bottom-up methods.

2.2.1. Liquid exfoliation

Liquid exfoliation20–25 is perhaps the most preferred method to exfoliate two- dimensional materials (Figure 2.3). It uses liquids as tool to separate layered materials into monolayers. It relies on the liquids to weaken the interaction force 25

between layers thereby separating them. The method always results in high yield of monolayer due to high contact between the liquids and the layered materials. The only downside to this method is the breakdown down of the monolayers due to the strong liquids used. The breakdown reduces the lateral size of the exfoliated monolayers leading to limited lateral size. For fundamental studies, this may be not a major problem but for large application such in nanocomposites, bigger lateral size is important to increase aspect ratio of the monolayer. There are mainly two approaches to exfoliate the layered materials with aid of liquids. First approach utilizes surface energy matching between the liquids/solvents with layered material where the solvents tend to the exfoliated flakes. Solvents with required surface energy are selected to prevent the exfoliated monolayers from aggregating together.

Commonly used organic solvents include N, N-dimethylformamide (DMF), isopropanol (IPA), N-Methyl-2-pyrrolidone (NMP), acetone etc. and their mixtures23.

As alluded to earlier, liquid exfoliation does not lead to single monolayers but mostly hybrids of few layers. To obtain single monolayers a second method alkali metal intercalation and dissolution is used. Alkali metals such as sodium, lithium etc. is driven into the van der Waal space of a specific layered material through electric field or diffusion. The method is normally referred to as intercalation by alkali metal.

The aggressive reaction of the alkali metal with water leads to lithiated material de- intercalated by water producing hydrogen gas between layers. It is this hydrogen bubbles that expand the spacing between layers facilitating the delamination between layers. The method has been widely used to exfoliate MoS2 and high yield 26

graphene sheets by the use of FeCl3 which is de-intercalated later by hydrogen peroxide.

2.2.2. Mechanical exfoliation

Mechanical exfoliation was among the first method used to obtain monolayer from layered materials, the most famous being graphene19 and others like MoS221, h-BN20 etc. The mechanical exfoliation uses adhesives like scotch tape placed on a crystal to cleavage it sequentially until monolayer are obtained. A huge advantage of mechanical exfoliation is the ability to obtain very high quality monolayers. The method however, has its own drawbacks notably, the size of the sample is normally 27

small (submicron) despite the micrometer size of starting flake. This disadvantage limits this method to produce samples for fundamental studies only not for large scale application. Another disadvantage is that the force required to delaminate a single layer depends on the interaction between layers. For materials with strong interaction between the van der Waals gaps, it is difficult to obtain the monolayer.

2.2.3. Chemical vapor deposition

In its simplest form, chemical vapor deposition (CVD) consists of five steps

(Figure 2.4); (1) mass transport of reactants; (2) adsorption of reactants on substrate surface; (3) chemical reaction on the substrate surface; (4) desorption of un-reacted products and by-products; and (5) pumping away of un-reacted products and by-products. Adsorption and desorption are the two processes which occur fast meaning deposition rate is controlled by the competition between mass transportation (in both steps one and five) and surface reaction rate occurring at step three. Mass transportation is a function of gas flow and its pressure. On the other hand, surface reaction rate is an exponential proportional to temperature therefore surface chemical reaction rate tends to be dominant at high temperatures.

Mass transportation control the reaction on the substrate and increasing it by more supply of reactants like liquid or gas precursor will lead to deposition rate of the thin films. Deposition tends to starts at different nucleation sites growing and coalescing together. Low-temperature results in slow surface reaction no matter the change in mass transportation rate leading to constant deposition rate. Slow reaction rate is desirable due to reactions happening at large surface simultaneously 28

which result in conformal and uniform deposition. However, temperature selection should be ideal because at low temperatures, the reaction may not occur due to thermodynamic reasons.

2.2.4. Sol-gel method

The sol-gel technique has been around for a long and is a well-established method for production of colloidal nanoparticles from mostly liquid phase26–28. It has recently been modified as an efficient method to produce advance nanomaterials and coatings. It is well adapted to the production of oxide 29

nanoparticles and composite nano-powders. The sol-gel is generally selected as result of its desirable advantages such as flexible versatility, low temperature as well as flexible rheology for easy of shaping into complex shapes. In addition, it is well suited to for the fabrication of hybrid organic-inorganic nanomaterials. Starting precursors are oxides of alkoxides due to liability of M-OR bond, in situ tailoring processing and commercial availability. Sol-gel method is a preferred method for synthesizing SiO2 nanoparticles. Briefly, The SiO2 synthesis by sol-gel involves two processes; hydrolysis and condensation of metal alkoxides (TEOS, (Si(OR)4) like tetraethylorthosilicate (TEOS, Si(OC2H5)4), an inorganic salts (Na2SiO3) in presence of mineral acids (HCl) or a base like NH3 acting as catalysts. A simple flow chart for the production of SiO2 nanoparticles from silicon alkoxides is shown below. The general reaction can be written as;

Figure 2.5 – Schematic representation of SiO2 particles synthesis by Sol-gel method

2.3. Synthesis of Graphene Oxide

In its simplest form, graphene oxide is an oxidized graphite to intersperse carbon layers by oxygen molecules then reduced with an aim of separating the layers to individual of few graphene oxide. Graphene oxide is an insulator as a result of disruption of sp2 bond network. However, once reduced its conductivity is greatly improved. Presence of functional groups makes it an interesting material to assemble into three dimensional nano-structures. Graphene oxide is one the nanomaterial widely used in this thesis therefore its befitting to discuss how it is synthesized. Hummers' method is the well-known method to synthesize graphene oxide29. Many modifications have been done to the Hummers method21,25,30,31 though the fundamental principle remains the same. The Hummers method consists of steps which must be followed to obtain high quality graphene oxide and they include; 1) Addition of graphite flakes (2g) and NaNO3 (2g) in 50 mL of H2SO4 (98%) in a 1000 mL flask and temperature maintained at (0-5°C) achieved in ice bathe while stirring. 2) Addition of potassium permanganate (6g) at this temperature very slowly. The rate of addition should be carefully controlled to maintain the reaction temperature no more than 15°C. 3) Ice bath is then removed and mixture stirred at elevated temperature of 35°C until turns into brownish and stirring continues for two days. 4) The mixture is then diluted with addition of water 100ml. Reaction temperature is then ramped to 98°C the mixture turns into brown color. 5) Further 31

dilution by addition of water (200ml) and stirring continued. 6) 10ml H2O2 is then used to terminate the reaction that changes the color to yellow. 7) 10% HCl is used to purify the solution followed by de-ionized water severally and 8) finally filtration is done and drying under vacuum at ambient conditions.

In summary, the above discussed methods for synthesizing nanomaterials are not exhaustive as there hundreds of methods. The thesis chose to focus on the methods which were extensively used for the purpose of the work presented herein.

Chapter 3

III. Role of Interface Engineering in Assembling 3D Structures from Zero and One-dimensional Materials

3.1. Background

3.1.1. Introduction

Building 3D structures in bottom-up assembly using zero, one-dimensional building blocks is still a major challenge. This chapter discusses a simple chemical functionalization method applied to carbon nanotubes (CNT) and silicon dioxide

(SiO2) nanoparticles to synthesize 3D structures with improved mechanical properties. The first part of the chapter deals with CNT while second part dwells on

SiO2 surface modification by silanization chemistry. This chapter is based on two papers: Owuor et al, Advanced Engineering Materials 19(5), 2017 and Owuor et al,

ACS Nano, 11(1), 806-813, 2016

3.2. Self-Stiffening Behavior of Reinforced Carbon Nanotubes

Spheres

Strong van der Waals forces between individual carbon nanotubes pose a major hurdle for effective use of nanotubes as reinforcement in nanocomposite due to agglomeration. However, the van der Waals forces in combination with functionalization of carbon nanotubes can be utilized to design nanocomposites mimicking stiffening behavior normally observed in biological materials such as fibrin gels, health bones, actin filaments in cytoskeletons etc. 34

3.2.1. Background

Nanotechnology is allowing researchers and engineers to design very sophisticated materials using bottom-up designs which were impossible to achieve using conventional methods. Carbon nanotubes (CNTs) have received a lot of attention from researchers due to their exceptional mechanical properties,3233 and extraordinary thermal and electrical conductivity34,35,36. The interphase between the

CNTs and the matrix is crucial to achieve excellent mechanical properties in a composite37. Though high damping can be achieved by incompatibility between the

CNTs interphase and the polymer38, high strength nanocomposites require compatibility between the two phases for efficient load transfer across the CNTs and matrix39. Addition of moieties on the CNTs40–42 which are chemically similar to a desired matrix has been widely researched43–45. Even though excellent results have been achieved, high loading of CNTs still pose a huge setback due to agglomeration

43,46. To overcome this problem, CNTs forests grown by chemical vapor deposition

(CVD)44 and CNTs films47 can achieve higher percentage loading in a polymer matrix. However, CNTs will buckle or kink when loaded along their central axis39.

It has previously been shown that self-stiffening48 effect of CNTs forests infused with compliant polydimethysiloxane (PDMS) polymer can be achieved.

Stiffness increases as the number of cyclic loading increases. To further improve on the content of CNTs and take advantage of their recovery from compressive cycles, three dimensional foams have shown remarkable mechanical properties49. These 3- dimensional (3D) carbon structures can be an effective way to achieve high content 35

CNT in polymer to take advantage of desirable mechanical properties of carbon nanotubes and in essence design high stiffness and damping nanocomposites. To create such CNTs spheres, functionalizing individual nanotubes by attaching carboxylic functional groups (-COOH) on the surfaces. Continuous dynamic loading on the PDMS encapsulated CNT spheres lead to subsequent higher stiffness, modulus and strength of the composite.

3.2.2. Preparation of CNT Sphere and Composite

Carbon nanotubes powder was used as a starting material where it was functionalized by attaching carboxylic functional groups (-COOH) in presence of nitric acid for 3-4 days in an oil bath at 70-80˚C. After functionalization, the multiwall CNTs solution (50mg) was sonicated in DI water (5 mL) for 1 hour.

Glutaraldehyde solution (35 μL), resorcinol (6 mg) and borax (0.5 mg) added and the solution sonicated further for 4 hours. Controlled volume of the solution was then dropped directly in a liquid nitrogen bath to form CNT spheres with diameters in the range of (φ ≈ 20µm to 3-4 mm). The spheres were further transferred in a freeze-drying instrument (Millrock Technology) and kept at -50˚C and 40 mtorr. The composite of carbon nanotube spheres and an elastomeric polymer (PDMS) were fabricated by mixing the required amount of monomer to hardener (10:1) by weight. First an initial layer of polymer is poured into a predetermined sized mold and semi-cured at 60˚C for 5 minutes then the spheres are arranged on top, the final layer of the polymer poured, and the composite cured at 70˚C for one hour. 36

3.2.3. Physical Characterization

Dynamical mechanical analysis tests were conducted on a DMA Q800 (TA,

USA). All the tests were conducted in compression mode at ambient conditions.

Three types of tests were conducted; strain sweep, load unload and temperature ramp tests. The temperature ramp tests were conducted from 30 to 150˚C at ramp rate of 3˚C /min. Samples of pristine PDMS and carbon nanotube spheres/PDMS were glued on a steel bar, selected weights, varying height, and an indentor used to impact samples. Each weight and height was recorded, and using the well- established potential energy formula (mgh). Energy absorption was calculated for both pristine and CNT spheres samples. Scanning electron microscope (SEM) images were taken on a FEI Quanta 400 ESEM at scanning electricity of 15 to 20 kV.

Optical images were taken on a goniometer.

3.2.4. Simulation Details

The computational model has dimensions of 2x2x1 mm. By virtue of symmetry and using appropriate boundary conditions on the face perpendicular to the 1 mm side, the model is able to represent a 2x2x2 mm cubic composite sample.

The finite element method simulations and analysis were carried out using Abaqus software50. A displacement was enforced on the top face of the rectangular model

(figure 3a) to simulate the compressive strain of the material at strain values of

0.05, 0.10, 0.15, and 0.20. The load-unload simulation was done in two steps: first by enforcing a compressive displacement of the top face of the rectangular model, and second by enforcing a displacement of the top face of the rectangular model back to 37

its initial position in the vertical direction. The materials properties were assigned as follows: PDMS was given a density of 0.97 kg/m3, elastic modulus of 2.66 MPa, and a Poisson’s ratio of 0.551; CNT sphere was assigned a density of 3 kg/m3, modulus of elasticity of 10 kPa, and negligible Poisson’s ratio; finally, the metallic inclusion was assigned the materials properties of stainless steel 302, with a density of 7861 kg/m3, elastic modulus of 193 GPa, and Poisson’s ration of 0.2552.

3.2.5. Results and Discussion

Figure 3.1(a) shows CNT spheres embedded within an elastomer, polydimethylsiloxane (PDMS). High structural integrity of the CNT sphere is maintained after being used as reinforcement in the polymer. Further optical images in Figure 3.1(b) and (e) show almost perfect spherical shape of the CNT spheres.

Presence of polymer does not interfere with the CNT spherical shape. This shows functionalization (Figure 3.1(c)) and strong van der Waals forces are beneficial to maintain the structure of the CNT spheres as shown by transmission electron microscope image (Figure 3.1(g)). Scanning electron microscope (SEM) images show the polymer is infused into the spheres and is well distributed (Figure 3.1(d)). 38

Figure 3.1(a) illustrates the stiffening behavior of CNT spheres/PDMS composite under dynamic loading in a compressive mode. There is an apparent stiffness improvement of the CNT spheres/PDMS composite compared to neat

PDMS. Furthermore, there is a clear self-stiffness increase as loading cycles increase for the CNT sphere reinforced composite unlike neat PDMS which does not show any self-stiffening with number of loading cycles (Figure 3.1(a), part II). It has been suggested that thicker crystallized shell encompassing CNTs can greatly improve the strength in semi-crystalline polymers such as PDMS53. Furthermore, modeling has also shown that polymers chain can adsorbs on the CNTs, therefore formation of well-oriented polymer layers at the interface can greatly contribute to improved van der Waals interaction within the CNTs resulting in efficiency load transfer6.

Experimental work has shown that addition of chain alignment along the polymer

CNTs interface can contribute to self-stiffening of CNTs forests grown by CVD method48. Organization of polymer chain at the interface region is crucial to achieve stiffening behavior. Enhanced van der Waals coupled with chemical functionalization aggregate many individual CNTs to form a porous sphere which increases the density of alignment of polymer chain around CNTs resulting in huge interaction thereby high stiffness observed. Close proximity of the CNTs can further increase the size of crystallized shell around CNTs during dynamic loading.

Interestingly, van der Waals forces between nanotubes is detrimental for their effective dispersion in conventional nanocomposites45,54 to achieve high mechanical 40

properties. However, the same van der Waals force in conjunction with functional groups can be highly beneficial for self-stiffening nanocomposites. The self- stiffening property of the CNT spheres/PDMS composite is further demonstrated by performing load and relaxation tests by subjecting the composite to a compressive load, remove the load, and load it again after 2 hours (Figure 3.1(b)). Stiffness jumps to 26% after the first loading and continues to do so for subsequent loading cycles though with lower magnitude compared to previous loading (inset Figure

3.1(b)). The self-stiffening effect was also exhibited by a thermoplastic polymer polycarbonate confirming that the effect is mostly due to the strengthening effect of

CNT spheres in the polymer matrix. In dynamic loads, densification of the CNTs takes place progressively leading to high density of the CNT spheres just like densification of bones in dynamic loads. Figure 3.1(c) shows static compression behavior of the two samples, from which it is evident that the CNT spheres/PDMS composite can withstand more load compared to the pristine PDMS, especially at higher loads. The CNT spheres/PDMS composite shows high modulus and negligible hysteresis compared to neat PDMS, further showing high strength of the CNT reinforced composite. It is believed that at higher loads, the PDMS is able to transfer the load to the highly elastic CNT spheres despite the weak interfacial bonding between the CNTs and PDMS55. The CNT spheres absorb this load without any permanent deformation due to their ability to withstand significant compressive force without collapsing56. Change in storage modulus of the CNT spheres/PDMS with temperature shows samples with CNT spheres with high storage modulus in 41

comparison with pristine PDMS sample. CNT spheres appear to hinder the chain movement in the polymer, hence the high modulus observed57. In fact, the modulus and stress increase with temperature (Figure 3.1(d)). Morphological changes of the

CNT spheres under compression load observed under an optical microscope (Fig

2e) depict the spheres taking the load and once the load is removed, no permanent deformation due to high elasticity of the CNTs sphere. High elastic behavior of CNT spheres makes it possible for them to deform while taking the load and springing back to their original shape with no noticeable permanent damage. Strain lines take an elliptical shape (white lines) but once the load is removed, they return to their straight shape further reinforcing the elasticity of the spheres within the PDMS matrix. High van der Waals forces in conjunction with functional groups aid in maintaining the spheres in their form and result in efficient load sharing between individual CNTs. 42

Using a CNT sphere as reinforcement inside the PDMS matrix for mechanical applications is advantageous over conventional reinforcing materials such as steel, 43

as demonstrated by finite element method (FEM) simulations of the composite. By virtue of symmetry, only half of the unit cell corresponding to the composite sample was modeled, as shown in Figure 3.3(a) and (b). The model shown in the figure represents spherical filler within a PDMS matrix, for two cases, each with different filler (CNT sphere and steel) and with material properties that greatly differ from each other in terms of elastic modulus, density, and Poisson’s ratio. The purpose of this simulated experiment is to establish a strict difference in the behavior of the composites with the two different fillers. Other assumptions and materials properties are given in simulation details.

The CNT spherical filler within the PDMS matrix is able to deform under a compressive pressure, as shown in the stress (all units of stress in contour maps, unless otherwise noted: Pa) and vertical displacement (all units of displacement in contour maps, unless otherwise noted: m) contour maps of Figure 3.3(a). Such deformation allows the PDMS directly above and below the CNT inclusion to be in a lower stress state than the composite sample with a steel inclusion (Figure 3.3(b)- top). The steel inclusion is not deformed under the applied compressive strain of

0.20 on the composite (Figure 3.3(b)-bottom), while the CNT inclusion suffers deformation at the top and bottom regions of the sphere under the applied compressive strain of 0.20 (Figure 3.3(a)-bottom), which allows the deformation to be transferred more efficiently form the PDMS matrix to the CNT sphere inclusion in comparison to the case with the steel inclusion. The high porosity of the CNT sphere was modeled by assuming a negligible Poisson’s ratio such that the compressive 44

strain in the axial direction does not cause a significant expansion of the material in the transverse direction, as observed in Figure 3.3(c). The high stiffness (high elastic modulus) of the steel spherical filler is the cause of negligible strain in the spherical inclusion and a stress field of about two to three orders of magnitude higher than the CNT spherical filler under a compressive axial strain of 0.20. In fact, over a range of applied strains from 0.05 to 0.20, the maximum von Mises stress is at least two orders of magnitude higher for the steel inclusion than the CNT filler, as seen in Figure 3.3(c) and (d). This shows the increased ability of the CNT sphere/PDMS composite to undertake applied compressive strains. This ability of deformation of the CNT sphere within the polymer matrix allows the material to withstand larger compression than the case of using a hard, metallic inclusion.

3.3. Scalable Bottom-up Assembly of Ultra-Low Density

Multifunctional Three-Dimensional Nanostructures from SiO2

3.3.1. Background

Interconnecting nano-size materials (CNTs58,59, graphene60,61 hBN,62 metallic nanoparticles63 and ceramics64 is a fertile ground to build 3D architecture for wide applications in automotive, aerospace, biomedicine, electronics packaging, energy storage, conversion etc. In order to connect CNTs and other carbon-based materials, electrical, mechanical, and chemical methods can be used. Among all, the chemical methods of connecting nanomaterials are simple and easily scalable. The chemical functionalization of CNTs can produce 3D porous interconnected structure and improve mechanical properties58. Extending this method to interconnecting inorganic and hybrid nanomaterials can be an interesting and promising research area. Inorganic nanoparticles are well known for their chemical stability65, high- temperature resistance65, and lightweight66,67. In this class of materials, silicon dioxide nanoparticles are preferred due to their low toxicity68–70, stability65 and 46

ability to be functionalized by a wide range of chemical functional groups65. The chemical nature of SiO2 permits easy attachments of organic moieties which has made SiO2 nanoparticles widely applicable as reinforcing fillers in polymers71, strengthening fillers in concrete, additives in rubber/plastics72, and non-toxic platforms for drug delivery73,74. Furthermore, the hydrolysable nature of the bonds provides an effective stress relaxation mechanism between the interfaces of organic and inorganic materials, which leads to improved durability and adhesion74.

However, directly fabricating lightweight nanostructured materials using inorganic nanoparticles as building blocks has proven elusive.

This section demonstrates that functionalized SiO2 nanoparticles can form well-defined 3D nanostructured materials. By employing a simple vacuum filtration method, 3D of SiO2 structures are assembled from its constituent 0D nanoparticles.

Mechanical properties are correlated with the degree of functionalization and show a close relationship between stiffness and functionalization. Functionalized SiO2 exhibit high stiffness and structural stability compared to non-functionalized SiO2.

Furthermore, reinforcing the SiO2 with small amount of graphene oxide (GO) as a filler, a multifold increase in strength and stiffness is observed. Abundant functional groups on GO nanosheets provide more anchoring points for the functionalized SiO2.

3.3.2. Synthesis of SiO2 3D Nanostructures

Silicon dioxide nanoparticles (US Research Nanomaterials Inc, USA) were functionalized by addition of silane coupling agents. First, 100ml of de-ionized water

(DI) was poured in a 1000ml flask then 1%, 5% and 10% of silane coupling agents 47

added (3-aminopropyl) triethoxysilane, N1-(3-Trimethoxysilylpropyl) diethylenediamine, Sigma Aldrich) based on the amount of DI water. The mixer of water and silane was stirred vigorously and left for eight hours for complete hydrolysis. Plus, 50g of SiO2 nanoparticles (average diameter 40nm) were later added and mixing under magnetic stirrer (Thermo Scientific, USA) continued for three hours. For graphene oxide reinforced samples, 0.1, 0.5 and 1% were added at this stage too. Vacuum filtration was later performed using a 55mm diameter filtration paper with pore size of 42µm. The mixer was left to filter for 24 hours. To ensure complete removal of solvent, the samples were later subjected to a temperature of 50°C in an oven (Across International, China) for 3 hours. Required sample dimensions were later cut using a razor for various characterizations.

Graphene oxide (GO) was synthesized via an improved Hummer’s method25. Briefly, a 9:1 acid mixture of concentrated H2SO4: H3PO4 (360mL: 40mL) was added to a mixture of natural graphite flake (3g) and KMnO4 (18g). The reactants were then heated to 50°C and stirred for 12h. The mixture was cooled to room temperature and slowly poured into 500 mL of iced de-ionized water. Then 30% H2O2 (10 mL) was added and stirred for 2 h at room temperature. The resulting mixture was washed with DI-water, 30% HCl, and ethanol and filtered. The residual filtrate containing the GO was collected and dried at 60°C for 24h.

3.3.3. Physical characterizations

Dynamical mechanical analysis tests were performed on a DMA Q800 (TA

Instrument, USA). All the tests were conducted in a compression mode at ambient 48

conditions. For the entire test, 1Hz cycle was selected unless otherwise stated. This cycle is useful to detect sensitive changes in the samples at low frequency.

Controlled force and strain tests were also done. For the strain rate controlled test, each sample was strained at 0.25% and held for a specified time. Load controlled test was done by ramping the force from 0.01N to 2N at 0.01N/min. Nano- indentation test on the SiO2 samples was tested by quasi-static uniaxial compression loading inside a scanning electron microscope (SEM) using a SEM PicoIndenter, xrPI85 (Hysitron, Inc., USA). Thermogravimetric analysis (TGA) was done on a Q600

(TA Instruments, USA) equipped with a DTA to study thermal stability of the three dimensional SiO2 material. The material was subjected to a temperature ramp test

(30 – 700°C at 10°C/min).

Optical images were taken using MM Optical machine. Scanning electron microscope (SEM) was done on a FEI Quanta 400 ESEM, at scanning electricity of 15 to 20 kV.

3.3.4. Simulation Details

The SiO2 computational structures are generated with: 1) the Reax Force Field

(ReaxFF) parameterizations for silicon carbide75, 2) a timestep of 0.25 femtoseconds, and the 3) Nosé-Hoover thermostat at 300 K. A 36 Å × 64 Å × 42 Å chunk of α-quartz in a 200 Å × 200 Å × 200 Å simulation box with periodic boundary conditions to isolate abutting units is equilibrated (NVT) for 50 picoseconds up to

3000 K to disrupt its short and long range orders. The resulting amorphous SiO2

(aSiO2) is then quenched down to 300 K for 10 picoseconds and further equilibrated 49

(NVT) at 300 K for 50 picoseconds. SiO2 and fSiO2 3D structures are created as a 2 ×

2 × 2 super-cells of the respective nanoparticles. In the fSiO2 3D structure, fSiO2/fSiO2 nanoparticle interfaces were manually made by connecting roughly 20 % of nitrogen endings of adjacent APTS functional groups.

The equilibration (NVT) for 50 picoseconds of the 3D models of SiO2, fSiO2, and fSiO2/GO resulted in a 77 Å × 81 Å × 77 Å, a 89 Å × 91 Å × 84 Å, and a 86 Å × 89 Å ×

99 Å simulation boxes with periodic boundary conditions. The computation of the stress-strain relationships on the aforementioned 3D models proceeded by: 1) further equilibrating the respective structures in an isothermal-isobaric ensemble

(NPT) while compressing them by 25 % for 5 picoseconds and 2) stretching them back by ~30 % for 5 picoseconds.

3.3.5. Mechanical response of 3D SiO2 Structures

Vacuum assisted filtration was used to fabricate the three-dimensional silicon dioxide structure (SiO2). Vacuum filtration has been used widely to prepare carbon nanotubes films76, graphene oxide films77,78, and other 2D materials79. As depicted by the schematic representation of the method in Figure 3.4(a), different types of silane at required concentration were first added to water and left for 8 hours before the addition of a specified amount of SiO2 nanoparticles. The extended time of silane in water is crucial for complete hydrolysis80. Addition of SiO2 was accompanied by magnetic stirring for three hours to ensure homogeneous distribution of the silane coupling agents on the SiO2 nanoparticles. Later filtration was done and the samples were left for a day to ensure all the solvent has been 50

removed. Figure 3.4(b) shows the proposed mechanism of silane modification on the SiO2 nanoparticles. Typically, SiO2 nanoparticles contains hydroxyl groups on their surfaces81–83. Silanol derivatives from hydrolysis of silane in deionized water

(DI) are reacted with SiO2 where hydroxyl groups in silanol are covalently bonded to the hydroxyl on the surface of SiO2 through a condensation reaction by modifying the SiO2 with long alkyl chains. The functionalized SiO2 (fSiO2) then forms a network of -Si-O-Si- bonds resulting in a random three-dimensional structure when excess silanol and DI water are filtered out and dried in an oven at 50°C for 3 hours. X-ray photo spectroscopy (XPS) characterization was used to quantify the chemical groups on samples. As expected, non-functionalized exhibits only Si and O chemical elements. Functionalized SiO2 and graphene oxide reinforced samples on the other hand show addition of C and N elements possibly from the silane functional groups.

Further analysis reveals reduction of O element from non-functionalized SiO2 mainly due to their reaction with silanols to form networks. Non-functionalized SiO2 can still form a three-dimensional structure but weak bonding between SiO2 particles does not allow them to form stable structures. High magnification optical image shows isolated crumps of the SiO2 nanoparticles. The weak forces between

SiO2 nanoparticles are not enough to maintain the structural stability of the materials in large scale. Unlike silane deficient SiO2 samples, silane-treated SiO2 tend to congregate together in large interconnected chunks of SiO2 forming large 3D structures. In addition, the digital image shows the structure composed of a three- dimensional structure with minimal observable manufacturing defects. Silane 51

coupling of SiO2 introduces chemical bonding responsible for maintaining high attraction forces among SiO2 nanoparticles. The addition of graphene oxide to the functionalized SiO2 further improves the structural stability of the materials as shown by optical and digital images of the sample. To further understand the morphology of the SiO2 samples, high-resolution scanning electron microscope images (SEM) were taken. The SEM images reveal major cracks on SiO2 samples

(Figure 3.4(c)). The absence of chemical interactions among SiO2 nanoparticles could be the reason for these large scale cracks on the sample. On the other hand,

SEM images of fSiO2 samples show large area SiO2 crumpled together without any observable cracks (Figure 3.4(d)) most likely due to the presence of silane functional groups. Microstructural analysis of fSiO2/GO samples shows large areas anchoring of SiO2 nanoparticles on GO sheets (Figure S3). GO has been shown to possess a lot of functional groups on the basal plane84. It is theorized that that these functional groups such as carboxyl, hydroxyl etc. provide more anchoring sites for

SiO2 nanoparticles to adhere on the surface of GO sheets. Further high resolution transmission electron microscope (TEM) images reveals better connection of functionalized SiO2 and GO reinforced samples compared to SiO2 samples. As observed in the optical images (Figure S2), SiO2 shows scattered small areas of SiO2 nanoparticles possibly due to the lack of strong interactions between individual SiO2 nanoparticles (Figure 3.4(e)). Contrary to SiO2 samples, fSiO2 samples exhibit large groups of nanoparticles in close proximity (Figure 3.4(f)). 52

groups with SiO2 nanoparticles, reactive silanol groups attach themselves to

the oxygen molecules on the SiO2. Depending on silane coupling, bi-, tri- etc. connection can be achieved. This connection is responsible for covalent bonds between nanoparticles resulting in a random three-dimensional structure. 53

The effect of functionalization on the mechanical properties of different 3D

SiO2 structures was investigated using a dynamical mechanical analysis (DMA) instrument. Samples were loaded to a compression strain of 0.25% and held in an isothermal mode for a number of specified cycles. Silane-functionalized SiO2 (fSiO2) shows a significantly higher stiffness compared to that of non-functionalized silicon dioxide samples (SiO2)(Figure 3.5(a)). There is an apparent self-stiffening behavior of functionalized SiO2 in all concentrations of silane. The highest stiffening behavior is exhibited by 1vol% silane concentration. The general trend is that a higher concentration of silane leads to a higher stiffness. A higher silane concentration leads to more silanol formation in water which in turn leads to a higher number of networks within the structure, hence the direct improvement in stiffness with increase in concentration. However, 10vol% silane concentration shows low stiffening in comparison to 5vol% silane concentration due to functionalization saturation. Saturation of silane within the nanoparticles could result in a thin shell like cover around the nanoparticles. Owing to this thin layer, when the structure is loaded, there is a tendency of the nanoparticles to slide over each other unlike the unsaturated silane functionalized-nanoparticles which tend to lock over each other when loaded hence the decrease stiffness observed.

Self-stiffening has been observed in a model random packed granular chains85. The self-stiffening in such systems was attributed to the length of connected granular chains. The long chains are essential to form high density entanglements which make them tighten and lock in presence of strain to prevent 54

any failure due to shear. The presence of functional groups may allow the

SiO2nanoparticles to form long chains which entangle and self-stiffen. For non- functionalized SiO2, this connection between SiO2 does not exist, hence easily broken apart in presence of a force. An observation of the samples after the test showed a good structural integrity of fSiO2. Unlike SiO2 samples which broke into small pieces after the test, fSiO2 and fSiO2/GO samples maintained their structures with no noticeable failure (inset of Figure 3.5(a)). The addition of graphene oxide to the functionalized SiO2 (fSiO2/GO) makes it 15 times stiffer than the non- functionalized SiO2. The extraordinary stiffness exhibited by the addition of graphene oxide could be explained by an increased interactions due to functional groups on the surfaces of GO such as epoxide, carboxyl, hydroxyl etc.84,86. These moieties may create covalent bonding as well as non-covalent interactions between

GO and functionalized SiO2 surfaces. It can be theorized that there is a tendency to form covalent bonds between the GO and SiO2. In fact, as shown by scanning electron microscope earlier, there is a high density of SiO2 nanoparticles attached to

GO nanosheets. Load displacement curves are shown in Figure 3.5(b). Such curves should normally exhibit two regions for materials built from particulate building blocks: deformation stage and densification stage. For non-functionalized SiO2, the deformation stage is completely absent due to failure of the sample at very low loads of 0.01N (pre-load force). The small force is enough to break the SiO2 samples.

For this sample, densification commences at a displacement of 120μm. The deformation stage for functionalized samples is characterized by a non-linear curve 55

with “pop-in” features showing possible local shear failure of the sample, and they take as much as 1N before reaching densification stage. The addition of graphene oxide in the SiO2three-dimensional structure results in a much stiffer sample taking as much load as the functionalized SiO2 without apparent transition between deformation and densification stages.

To compare effects of different functional groups, additional similar load- unload experiments of these SiO2 samples functionalized by different chemical groups are done. The aforementioned comparison reveals that silane functional groups are better compared to epoxide functional groups. The epoxy functional group has a comparable stiffness to non-functionalized SiO2 but four times less than silane-treated SiO2. Further comparisons between different types of silane coupling agents show that 3-aminopropyl triethoxysilane (APTS) exhibits a higher stiffness than all the silanes tested such as N1-(3-Trimethoxysilylpropyl) diethylenetri-amine and (3 Glycidyloxypropyl) trimethoxysilane. The high stiffness induced by APTS could be attributed to its shorter alkyl spacer. This shortness of alkyl spacer has been shown to be as a result of stronger electron interaction between the functionality and the silicon atom87. In all the silanes used to functionalize the SiO2 nanoparticles, APTS had the shortest alkyl spacer.

What’s more, frequency controlled test also shows high stiffness with an increase of frequency for fSiO2 and fSiO2/GO samples. At higher frequencies, SiO2 starts to lose its stiffness due to the lack of connection between individual

SiO2nanoparticles (Figure 3.5(c)). We further monitored the concentration of GO 56

concentration in the SiO2-Silane structure and observed that stiffness increases with

GO concentration. However, concentrations of GO above 1% do not show any further improvement in the stiffness. This observation could be due to agglomeration of GO nanosheets at higher concentration where the majority of moieties could be shielded from the functionalized SiO2 nanoparticles thereby limiting covalent linkages.

To better understand the local deformation and failure mechanism in such materials, an in-situ nano-indentation test inside SEM is used88,89. Representative load-displacements curves are shown in Figure 3.5(d). The test was done on SiO2 and two different types of silane coupling agents, (3-aminopropyl) triethoxysilane

(APTS) and N1-(3-Trimethoxysilylpropyl) diethylenediamine. All samples were loaded with same amount of force under load control for consistency. As shown in

Figure 3.5(d), functionalized SiO2 shows a very high elastic modulus which is much stiffer than that of the SiO2. APTS (indicated by i) exhibits high modulus than the N1-

(3-Trimethoxysilylpropyl) diethylenediamine (indicated by ii). The unloading portion shows a typical hard elastoplastic behavior for the two functionalized SiO2 samples. Unloading curve of SiO2 shows a non-linear zigzag behavior reminiscent of softening due to the fractured sample (indicated by iii). Concurrent SEM images during indentation show that SiO2 deformation originates from the indentor tip, a stress concentrated area. When loaded, weak forces between SiO2 nanoparticles make them easily breakable and crack starts to propagate in all directions as shown in Figure 3.5(e). This type of failure is repeated no matter where the indentation 57

location is chosen on the material. As opposed to non-functionalized samples, silane-treated 3D SiO2 does not show any noticeable failure for the same load

(Figure 3.5(f)). The presence of covalent bonding on SiO2 nanoparticles does not allow for the loading force to easily fracture the material.

The atomic scale of functional groups and 2D graphene oxide sheets is hypothesized to play a crucial role in interconnecting and building SiO2 3D porous architecture, which cannot be observed directly in experiments. Therefore, a detailed MD simulation was performed. 58

Figure 3.5 – Mechanical behavior of functionalized SiO2. (a) Discernable stiffening of functionalized SiO2 in a cyclic compressive test, functionalized

SiO2 has a stiffness more than four-fold compared to non-functionalized 59

functionalized SiO2. (b) Load displacement curves showing better elastic

deformation behavior of fSiO2/GO and fSiO2. Non-functionalized sample fails at very low loads (0.01N). (c) Frequency controlled test to understand the stability of functional groups on the SiO2 at high loading cycles. As expected at high frequency, non-functionalized SiO2 lose its stability unlike functionalized and GO reinforced samples which exhibit constant stiffness without noticeable loss in stiffness. (d) Nano-indentation test on two types of silane coupling agents and non-functionalized SiO2 nanoparticles. APTS (i) functionalized SiO2 takes extremely large force compared to N1-(3-Trimethoxysilylpropyl)

diethylenediamine (ii) and non-functionalized SiO2 nanoparticles (iii and inset). (e, f) SEM images in-situ during nano-indentation test. Non-

functionalized SiO2 structure breaks at very low loads (e), once the crack develops, it propagates in all directions. On the other hand, functionalized

SiO2 structure does not show any cracks development (f).

The generation of the computer model for pristine amorphous SiO2 nanoparticles (pSiO2) used in the hydrogen passivated (hSiO2 or SiO2), APTS functionalized (fSiO2), and APTS functionalized with graphene oxide three- dimensional (3D) models as well as the calculation of mechanical properties of the aforementioned structures were done employing a combination of the scientific visualization and analysis software for atomistic simulation, OVITO90, and classical molecular dynamics (MD) with its numerical implementation in the large-scale atomic/molecular massively parallel simulator91(LAMMPS). The details are described in simulation details section. Using OVITO90, a 35 Å diameter spherical nanoparticle of pristine amorphous SiO2 (pSiO2) is carved out of aSiO2. Additionally,

1) hydrogen atoms are attached to roughly 5 % of the surface of pSiO2 nanoparticle 60

to create the hydrogen-passivated nanoparticle (hSiO2 or SiO2 for simplicity) and

APTS functional groups are attached to roughly 4 % of the surface of pSiO2 nanoparticle to create the APTS-functionalized SiO2 nanoparticle (fSiO2). The

APTS/pSiO2 interface in light green highlight of Figure 3.6 encompasses a silicon atom and three highly reactive oxygen atoms to be attached to the surface of pSiO2.

The 3D structure made up of fSiO2 nanoparticles and both epoxide and hydroxyl groups attached to 35 Å × 25 Å graphene nano-ribbons, was created by connecting

APTS nitrogen endings to carbon atoms belonging to the previously described graphene oxide nano-ribbons. Such a graphene oxide structure, i.e., GO-h2e2, has been chosen for its stability92 and its likelihood to match the experimental randomness of graphene functionalization. As the box compression/tension or loading/unloading occurred in the z-direction, the respective stresses and strains at each step were computed. The above description of the creation of 3D models of

SiO2, fSiO2, and fSiO2/GO is illustrated in Figure 3.6. 61

Figure 3.6 – Computer models for hSiO2, fSiO2 and fSiO2/GO 3D amorphous

SiO2 structures. Generation of fSiO2/GO, a 3D structure combining APTS functionalized amorphous SiO2nano-particles and both hydroxyl and epoxy 62

functionalized graphene nano-ribbons (GO-h2e2) from a 3D structure made of

the surface of a pristine amorphous SiO2 nanoparticle (pSiO2). Hydrogen

passivated to obtain the hydrogen passivated amorphous SiO2 nanoparticle

(hSiO2 or simply SiO2) used to obtain the equivalent 3D structure, the pristine amorphous SiO2 nanoparticle (pSiO2) is generated carving a sphere out the an

amorphous SiO2 obtained through annealing a chunk of crystalline SiO2 structure (α-quartz).

The uniaxial stress-strain relationship depicted in Figure 3.6is in qualitative in agreement with the experimental mechanical testing in Figure 3.6. In fact, under the same 25 % uniaxial load, the addition of GO to the fSiO2 3D structure provides

~200 % more maximum stress than the fSiO2, and ~600 % more maximum stress than SiO2. Such a superior mechanical performance on the part of fSiO2/GO and fSiO2 overSiO2 is attributed to the interfacial bonding between nanoparticles and GO nanosheet-ribbons, and among nanoparticles. Assuming the percentage of hydrogen passivation at the interface to be enough to reduce the amount of O-O, O-Si, or Si-Si bonds at the nanoparticle interface, the presence of hydrogen atoms drastically reduces the amount of covalent bonding necessary for a good mechanical performance. The mechanical performance of fSiO2 has therefore expectedly been confirmed greater in our simulations in view of the N-N/O/Si bonds connecting two adjacent APTS functional groups, an APTS nitrogen ending to oxygen/silicon atoms on the nanoparticle as well as the lack of hydrogen atoms that increases the amount of O-O, O-Si, and Si-Si bonds at the interface. In addition to the aforementioned 63

bonds in the fSiO2 3D structure, the insertion of GO resulting into the fSiO2/GO 3D structure enables APTS nitrogen endings, oxygen and silicon atoms at the nanoparticle interface to covalently bind with carbon and oxygen atoms on the nano-ribbon, hence its superior mechanical performance, see SiO2, fSiO2, and fSiO2/GO nanoparticle interfaces in the respective red, blue and black circles above the plot in Figure 3.6. It is worth noting that the unloading phase allows for more recovery in the case of the fSiO2/GO 3D structure by virtue of the greater number of bonds as well as stronger bonds than those of fSiO2, and SiO2.

Figure 3.7 -(a) SiO2, (b) fSiO2, and (c) fSiO2/GO nanoparticle interfaces in the respective 3D models. (d) Uniaxial (z-direction) stress/strain curve for hSiO2

(red curve), fSiO2 (blue curve), and fSiO2/GO (black curve) 3D amorphous SiO2 structures.

3.4. Conclusion

In summary, this chapter has shown that simple surface chemistry is critical in taking advantage of nano-building blocks for assembling 3D structures in a bottom-up process. For instance, the strong van der Waals forces in combination with functionalization of CNTs can be utilized to design nanocomposites that mimic stiffening behavior normally observed in biological materials. Furthermore, silanization chemistry is also used to create siloxane networks in SiO2 nanoparticles resulting in 3D structures which show high stiffness. Functionalized SiO2 exhibits efficient flame retardant when exposed to a flame. This lightweight material could be useful in various applications such as automotive and aerospace. 65

Chapter 4

IV. Surface Modification of 2D Materials into 3D Hybrid Structures

4.1. Background

The discovery of Graphene93, and subsequent other two-dimensional materials led to an explosion in research geared towards utilizing their intriguing mechanical, electronic, physical and chemical properties. While the physical properties of nanomaterials have been extensively explored, the assembly in a bottom-up approach to design hybrid 3D nanostructures by taking advantage of their interfacial properties still needs a deeper inquiry. In this chapter, a discussion of synthesizing 3D structures from graphene oxide and hexagonal boron nitride is presented. The discussion is based in the following papers; Owuor et al., ACS Nano

11(9), 8944-8952, (2017) and Owuor et al., Advanced Materials Interfaces 4(10),

(2017).

4.2. High Toughness in Ultra low Density Graphene Oxide Foam

4.2.1. Background

Graphene oxide foams are example of bottom-up assembly structures with excellent electronic94,95 and mechanical properties96–102. These low-density porous foam materials have shown high capability in energy absorption under high compressive deformations. Other areas have also benefited from these three dimensional foams, such as biomedical103, fuel cells104–106, sensor107–109, supercapacitors110111, lithium-ion batteries112–114, high thermal interface conduction115, flexible electronics116 and high density and fast electron transfer95. 67

The recent freeze drying32,62 method has proved to be an elegant and cost effective way to produce the foams. Despite the attractive physical properties of graphene oxide foams, their stability needs improvement. For instance, they are easily degraded when placed in flowing solution, either by the collapsing of their layers or just structural disintegration into small pieces. Bending and buckling of the cell walls are main causes of failure when subjected to compression loads117. Low bending modulus of the individual nano-sheets in graphene oxide have curtailed their use as effective load carrying materials. To exacerbate this problem, individual nano-sheets are normally held together by week van der Waal forces118,119 resulting in very poor shear strength among the nano-sheets. Recently, several reports have shown that connecting these nano-sheets either by welding60 or CNTs can result into foam with very high stiffness61. Addition of nanoparticles and 2D sheets such as zirconia64, iron oxide63 and hBN62 improves mechanical properties. Deliberate design of the interface between these 2D layers can prove instrumental in fabricating foams with high mechanical stability without degrading their inherent properties, such as high strength and low density.

A unique method to engineer the interface between graphene oxide nanosheets by controlled addition of a thin layer of polydimethylsiloxane (PDMS) to form interconnected graphene oxide foam (pGO) (Figure 1a) is presented. PDMS is selected due to its high affinity to GO nano-sheets forming physical and chemical cross-links120. Detailed mechanical characterization under tensile and compressive loading are performed. 68

4.2.2. Preparation of interconnected Graphene Oxide Foam

GO was synthesized by improved Hummer’s method as previously mentioned in chapter two. To prepare the foam, 100 mg of GO powder was sonicated in 10mL of water, 20µL glutaraldehyde and 2mg of resorcinol. The solution was then poured in plastic cuvettes and frozen under liquid nitrogen. GO foam is obtained after lyophilization of the frozen mixture on a freeze-drier. To synthesize polymer interconnected graphene oxide foam (pGO), a two parts polydimethylsiloxane (PDMS) was mixed and a syringe used to pour one drop and a time into the GO foam in the cuvettes and left to cure for two days. The PDMS was controlled in the foam to have 0.5, 1, 1.5 wt.% etc.

4.2.3. Physical Characterization

The oxidative thermal stability of GO, pGO was studied by Thermogravimetric

Analysis (TGA) under inert environment (Argon gas, at flow rate of 45mL/min) using TA Instrument (Q600). For repeatability and consistence of data, around 10-

15 mg of each sample was weighed prior to testing. Temperature ramp test was used for all samples where temperature was ramped from 30°C to 700°C at a heating rate of 10°C/min. Weight loss and DTA thermograms were extracted from raw data. For mechanical characterization, all tests were conducted in a compression mode at a frequency of 1Hz with amplitude of 20µm. Multi frequency- strain, controlled force and multi-strain tests were done. Temperature ramp tests were performed from room temperature to 200°C at a heating rate of 5°C/min. 69

Scanning electron microscope (SEM) images were taken on a FEI Quanta 400

ESEM at scanning electricity of 15 to 20 kV. Optical images were taken on a goniometer. Local Modulus Mapping was conducted by atomic force microscope

(AFM) PeakForce technique (Multimode 8, Bruker Santa Barbara, CA) with contact mode. Relative stiffness quantification method121 was employed, in which the deflection sensitivity was calibrated by a ‘Ramp’ on the flat sample surface, and the spring constant of each specific cantilever was calibrated by thermal tuning. It should be noted that the deflection sensitivity and spring constant need to be carefully performed before each characterization. Herein, contact-mode cantilever

(FESP, Bruker, SantaBarbara, CA) with a spring constant about 5N/m was adopted.

The photo-thermal induced resonance measurements were finished using NanoIR2

(Anasys Instruments, Santa Barbara, CA). Infrared Analysis (IR) spectra were collected by directly exposing the composite surface to a tunable optical parameter oscillator laser beam.122 Contact mode AFM probe (PR-EX-nIR2) with stiffness between 0.07 and 0.4N/m was used for the thermal characterization. The experimental data were collected in contact mode using Analysis Studio software

(Version 3.9, Anasys Instruments, Santa Barbara, CA). The laser irradiation was conducted in air with a 10.6 µm laser (CO2 laser) cutter system (Universal Laser

System, Model XLS10MWH) with fixed frequency of 3 kHz and fixed speed of laser scribing of 15.24 cm s-1 with image density of 500 pulses per inches, over a 0.25 cm2 area. The laser duty cycles was varied from 1 to 100% (75 W). 70

4.2.4. Simulation Details

The fully atomistic molecular dynamics simulations were carried out under a NVT ensemble (at 300 K) controlled by the Nosé−Hoover thermostat, as implemented in the open source code large-scale atomic/molecular massively parallel simulator (LAMMPS)91. The reactive force field ReaxFF123,124 was used in all simulations. This force field can reliably handle chemical reactions, as formation and/or breaking bonds and it uses a general relationship between bond distance/bond energy and bond order. Its parameterization is based on density functional theory (DFT) calculations and/or experimental data (when available).

The typical errors in heat of formation values are around 2.8 kcal mol−1, which is comparable to experimental ones. 71

Structural model systems were composed of graphene oxide square sheets of

41 Å in length and PDMS polymer chains made up of 30 repeating units

(monomers). To simulate the stress-strain process we first thermalized the entire system for 600 ps at 300 K, followed by an application of a time dependent external force on each carbon atom in the GO sheets in opposite directions in order to pull the sheets apart (see Figure 4). The applied force maintained until the sheets breaks apart. In this study we considered three different arrangements of graphene oxide

(GO) sheets covered with PDMS polymer as described in the section Simulation

Studies.

4.2.5. Result and Discussion

Solution processing method was used to produce the covalently interconnected graphene oxide layers62, which were freeze-dried to produce a stable three dimension structure (Figure 4.1 (b)). Abundant –OH groups present in

GO layers were covalently linked to each other to form a defined porous three dimensional network of interlinked sheets97, as shown by scanning electron images in Figure 4.1(c)and (d). Graphene oxide layers in the foam do not exhibit any visible collapse on the macro-scale (Figure 4.1 (e)). Individual layers of graphene oxide in pGO are well connected to each other by the PDMS (smooth surfaces on the GO foam) as is clearly depicted in Figure 4.1(f). Such connections are important to prevent the layers folding on each other due to their own weight (Figure 4.1(g)).

Atomic force microscopy (AFM) was further used to understand the distribution of the PDMS within the graphene oxide foam. Two-dimensional and its corresponding 72

3D mapping of the pGO reveal a substantial homogeneous-like distribution of PDMS

(light blue and blue colors) in the GO foam. Local modulus values ranges from 39 to

155MPa, encompassing both modulus for PDMS and GO. The mapping closely resembles SEM images for the pGO (Figure 4.1(h)). The detailed characterization tests were conducted to study the chemical interface of PDMS with GO sheets.

Fourier transform infrared spectroscopy (FT-IR) results exhibit graphene oxide peaks in the pGO foam. These peaks occurs at 1040cm-1, 1420cm-1, 1630cm-1 corresponding to C-O, C-O and C=C, respectively (Figure 4.1(i)). Thermal tests showed high thermal and structural stability of the pGO induced by PDMS, in comparison to GO foam. The GO foam exhibits three transitions losses of weight, the first at 100°C, mainly due to adsorbed water in the foam, 14% at 150°C and finally a substantial loss of more than 35% at 230°C, which is an indicative of functional groups removal from GO foam61 and reduction to reduced GO (rGO). On the contrary, pGO shows gradual losses of weight above 480°C, where only 12% of the weight is lost, which implies better surface and interconnects engineered of GO sheets with high thermal stability. 73

(AFM) mapping showing homogeneously distribution of PDMS within the pGO (i) FTIR spectra GO and pGO

A comparative study on the mechanical properties of GO and pGO. Tensile engineering stress-strain curves are shown in (Figure 4.2(a)). The pristine GO foam exhibits almost negligible strength in tension due to the weak inter-phase tensile strength between the nano-sheets, whereas pGO shows multifold increase in tensile strength and strain values. A higher PDMS loading leads to an increase in tensile yield strength (0.17MPa) and strain to failure (14%). Furthermore, the tensile stress-strain curves show peaks and valleys which are characteristics stepwise failure of the pGO sheets. Compressive load-unload tests on the other hand, show high compressive strength of pGO (0.49MPa) in comparison to the GO foam

(0.02MPa) (Figure 4.2(b) and inset), which represents an improvement of 2350%.

The high strength was present in five repeated cycles. pGO has very high strain as high as 40% without any observable deformation. Unlike pGO, GO foam collapsed after certain load under compression. Changes in stiffness with load shows high stiffness of pGO in low and high loads (Figure 4.2(c)). Cycling tests by loading the materials with the same loadings but at hundreds of cycles similar to a fatigue tests also reveal high stiffness of the pGO foam. For instance, pGO-1 (1% PDMS) led to an improvement of stiffness by 65 orders of magnitude. Similarly, this value increases to an extraordinary ̴ 2800% for pGO-2(1.5% PDMS) (Figure 4.2(d)). In addition, load-relaxations tests were also performed and changes in stiffness monitored.

Samples were loaded and let to relax for 2 hours before another dynamic 75

compression (inset schematic on Figure 4.2(e)). For each follow-up test, the stiffness continues to increase (Figure 4.2(e)). Storage modulus of pGO showed high values in high and low temperature with pGO-2 having the highest modulus of 0.29MPa while

GO foam has a meager value of 0.005MPa (Figure 4.2(f)).

Figure 4.2 – pGO mechanical properties. (a) High modulus, tensile strength and improved strain to failure of pGO foam. Interconnected GO sheets result in 76

better adhesion hence high tensile strength and toughness. Inset - optical image of collapsed sample indicate elastic failure. (b) Compressive test shows pGO to have very high compressive strength compared to pure GO foam (inset). (c) Load vs stiffness results reveal high stiffness with increasing load for the pGO unlike pristine GO foam. (d) Cyclic loading showing stiffening behavior of pGO as the number of cycles increases. (e) Load relaxation tests showing increases of stiffness with subsequent loading, here, the sample is loaded in a dynamic manner, test stopped to let the sample relax and then loaded again after a certain number of loading cycles. (f) Temperature controlled test to study the storage modulus. High modulus of pGO at low and high temperatures. Physical and chemical cross-links induced between the interphase make the material highly resistance to dynamic load.

To elucidate the failure mechanism of the pGO, scanning electron microscopy

(SEM) images of the fractured surfaces are analyzed. As shown in Figure 4.3(a),

PDMS adheres well to the GO sheets and no noticeable damage on the GO sheets100 were observed as shown by clear strain lines (Figure 4.3(b)). Further strain lines or pulling on the GO foam sheets in response to load is seen where the PDMS prevent the sheets from pulling from each other and increases their bending modulus

(Figure 4.3(c)). As the load increases, the sheets start to pull from each other in a progressive manner, but the rate is highly reduced due to the presence of the PDMS.

PDMS addition proves to be an effective way to increase the strength of the GO foam.

To further shed light on the failure mechanism of the graphene oxide interconnects with PDMS, quantitatively study on the local failure behavior by in-situ compression tests are done. First GO sheet are pulled out of plane (inset Figure 4.3(d)). High resistance to deformation is observed unlike pristine GO foam, which shows very low resistance to failure (Figure 4.3(d) and insets). The current compression shows 77

a good out of plane interconnection of GO sheet. Next, compression at in plane interface was performed by compression of the interface region of two layers (insets

Figure 4.3(e)). As shown in Figure 4.3(e), presence of PDMS increases the interconnections between two layers of graphene resulting in high strength compared to un-connected GO layers. Comparison of compressive strength with other lightweight materials shows that pGO does have high values compared to majority of materials such as carbon nanotube aerogels and other foams (Figure

4.3(f)) 78

Figure 4.3 – Morphology of fractured pGO. (a) Better wetting of the GO sheets by the PDMS. (b, c) High magnification images PDMS on the thin GO nanosheets, no noticeable failure of the GO sheets shown by strains lines in response to load. (d) Nano-indentation probe of pGO by indenting on a bundle of engineered surfaces of GO area exhibiting high load bearing compared to pure GO foam. (e) A shear like test by pulling one sheet from each. Interphase 79

engineered sheets are harder to separate than those no surface treatment. (f) Comparison of compressive strength of pGO with other lightweight materials.

For a GO foam, failure occurs mainly due to weak cell wall/struts, which constitutes the 3D interconnected network. Van der Waals interactions among the sheets are not strong enough to make the structure resistance to loads. In contrast, pGO foam makes the struts or cell walls strong enough to elongate in response to applied load.

In addition, the presence of PDMS at interconnects of GO foam greatly improves its strain transfer property117. The zig-zag behavior of stress-strain curve could be explained by in-homogeneity on size of the struts/cell walls, for example thickness, where some of the struts/cell walls will fail earlier than others in a random manner.

High stiffness of pGO can still be linked to better interconnect in the nano-sheets as well as densification, in addition to better fibril alignments125 of the graphene oxide sheets. Tan delta values are good indicators of the damping properties of a material.

Engineered inter-phases of GO foam lead to an extremely better energy damping.

In order to gain insights into the influence of the PDMS in the mechanical properties of the pGO foam, a proposed simplified structural model is used (Figure

4.4(a)). To investigate the stress-strain behavior of the interconnected foam, three different arrangements of graphene oxide (GO) sheets interconnected with PDMS are considered. The orientations are: (a) two neighboring three-layers GO anchoring two single layer GO sheets connected with PDMS; (b) two neighboring three-layers

GO with their relative edges facing each order and with PDMS in between; and (c) rotated by 90°. All cases are presented in Figure 4.4. This figure shows for each 80

considered orientation two snapshots: the thermalized structure (@300K) in the initial stage of the stress-strain molecular dynamics simulation and also in a further stage, but before the complete failure. The long chain PDMS clearly shows its capability of holding two GO sheets until a remarkable strain value. Figure 4.4 clearly shows that the presence of the PDMS makes the structures more flexible, presenting ductile-like deformations, unlike what happens with graphene oxide foams, which breaks into pieces exhibiting a well-known brittle behavior. The chemical interaction of PDMS with high surface area GO gives resistance to straining or separation of layer very similar to as reported in our previous works32,62. In

Figure 4.4(d) the relative displacement of the outer GO sheets as a function of the applied force is shown. This arrangement was considered in order to provide a reliable comparison (in terms of the energies involved) between the pGO and the GO foam. From the Figure the necessary force to pull the GO sheets apart is larger in the case where the foams are connected with PDMS, which is consistent with pGO being more resilient to deformations than pure GO. This can be explained by the high affinity between PDMS and GO nano-sheets, which is confirmed in our local deformation using in-situ compression experimentally tests discussed in the previous section. 81

simulations before structural failure occur. Notice the ductile-like deformation of the pGO unlike what happens with GO, which breaks into pieces exhibiting brittle behavior. (d) Relative displacement as a function of the applied force on the outer graphene oxide sheets. This figure shows that the presence of the polymer makes pGO more resilient to structural deformations in comparison to the pristine GO.

4.3. Oxidation of hexagonal boron nitride (h-BN) nano-sheets to

synthesize reactive sites for assembling of 3D structures

Weak van der Waals forces between inert hexagonal boron nitride (h-BN) nanosheets make it easy for them to slide over each other resulting in an unstable structure in macroscopic dimensions. Creating interconnections between these inert nanosheets can remarkably enhance their mechanical properties. However, controlled design of such interconnections remains a fundamental problem for many applications of h-BN foams. Here, a simple methodology to oxide the h-BN nanosheet is presented. Oxidation creates 'defects' which are used as linking position for polymer molecules.

4.3.1. Background

With its naturally occurring layered structure126, hexagonal boron nitride (h-

BN) has emerged as an important 2D material, possessing excellent properties such as high temperature stability, high thermal conductivity as well as mechanical strength127. Unlike its carbon graphene ‘cousin’, h-BN planar networks are composed of boron and nitrogen atoms arranged in a hexagonal structure, resulting 83

in strong covalent bonds in an in-plane direction, but with weak van der Waals forces between layers127.

The fabrication of three-dimensional porous, nano-structures from interconnected, 2D materials such as graphene100,128, clays129, metal dichalcogenides130,131, etc. has proven to be a promising technique to exploit their properties for several applications including catalysis114,132,133, energy storage134, biological and environmental applications58, mechanical damping49 and gas sequestration135. Graphene and its derivatives like graphene oxide (GO) have received much attention mainly due to their easy integration into 3D structures. The same cannot be said of few layers h-BN, which has proven to be difficult to synthesize as 3D constructs unless used as reinforcements20 in nanocomposites.

Although chemical vapor deposition (CVD) methods37,136, spark plasma welding techniques137,138, and surface chemistry methods58,135 have achieved considerable success in designing graphene based nano-engineered 3D structures, there still exist enormous challenges in replicating the same for h-BN structures, in particular in large quantities for large scale applications. Recently, the freeze-drying method61,62 has proved to be an efficient method to produce 3D nanostructures from their 1D or

2D constituents49,58,59. This method can be improved further with surface chemistry, enabling the design of 3D structures with tailored surface properties. This combination is an inexpensive method to mass-produce selective macroscopic 3D nanostructures, but has not been effective in producing scaffolds such as h-BN foams. Three (3D) foams are essential for various applications due to their high 84

porosity, hence their utilization as air filters139 or gas absorption materials. With a controlled porosity and chemical functionalization in these 3D foams, they can serve as effective absorbents for a large class of different materials.

4.3.2. Oxidation of h-BN

For the preparation of oxidized BN (f-BN), 1 g of h-BN was dispersed in a 9:1 acid mixture (200 mL) of H2SO4:H3PO4 (99% H2SO4) and stirred at 50oC for 2h. 1.5 g of KMnO4 was then added to the mixture and stirred at 50oC for 6 h. The mixture was then poured over ice followed by the addition of H2O2 and stirred for 1h. The resulting mixtures were washed with 200mL DI-water, 200 mL of 30 wt% HCl aqueous solution and 200 mL of ethanol. The collected h-BN was dried at 50oC for

24h.

4.3.3. Preparation of oxidized h-BN/polyvinyl alcohol foam

0.1 g of f-BN was added to 10 mL of DI-water and sonicated for 1h to aid in exfoliation. After sonication, the mixture was stirred at room temperature for 10h.

After that, resorcinol (11mM), glutaraldehyde solution (22mM), and sodium tetraborate (0.06mM) were added. The resulting materials were stirred for 2h.

Then, 0.1g of 1wt. % Polyvinyl alcohol (PVA) solution (solvent: DI-water) was added to the BN solution and the mixture was sonicated at room temperature for 2 h. The resulting materials were then frozen under liquid nitrogen and subsequently freeze dried for 24h. 85

4.3.4. Physical Characterization

Mechanical characterization was investigated on a dynamic mechanical analyzer (DMA, Q800 TA instruments). All tests were done in a compression mode clamps at a frequency of 1 Hz except frequency sweep tests where the amplitude was kept at 20µm. Temperature ramp tests were performed from 30°C to 200°C at a heating rate of 5°C/min. The loading condition for stiffness test was done by applying amplitude of 20µm with a static force of 0.01N while maintaining a force track of 125% on the dynamic mechanical analyzer (DMA, Q800 TA). Stiffness, storage modulus and glass transition temperature were then obtained from the data. Scanning electron microscope (SEM) images were taken on a FEI Quanta 400

ESEM, at a scanning electricity of 15 to 20kV. The laser irradiation was done in air with a 10.6 µm laser (CO2 laser) cutter system (Universal Laser System, Model

XLS10MWH) at a fixed frequency of 3 kHz. The laser speed was a fixed laser scribing of 15.24cm s-1 with an image density of 500 pulses per inches on an area of 0.25cm2.

The laser duty cycles were varied from 1 to 100% (75 W). X-ray diffraction was done on Rigaku D/Max Ultima II with Cu K alpha radiation (1.542 nm) while Raman measurements were done on a Renishaw Raman Spectrometer with a 532nm laser.

The thermal stability of h-BN/PVA foam was studied under Thermogravimetric

Analysis (TGA) in an inert environment (Argon gas, at flow rate of 100mL/min) using a TA Instrument (Q600). Around 10-15 mg of each sample was weighed prior to testing. The samples were subjected to a temperature ramp of 10°C/min from

30°C to 700°C. 86

4.3.5. Simulation Details

The fully atomistic molecular dynamics simulations were performed under a NVT ensemble (at 300K) controlled by the Nosé−Hoover thermostat, as implemented in the open source code large-scale atomic/molecular massively parallel simulator (LAMMPS)91. The fully atomistic reactive force field ReaxFF124,123 was used in all simulations. This force field can accurately estimate chemical reactions as bond formation and/or breaking due to its parameterization through density functional theory (DFT) calculations and/or experimental data. ReaxFF uses a general relationship between bond distance/bond energy and bond order. The error in heat of formation values is around 2.8 kcal mol−1, comparable to experimental data. The time step was chosen to be 0.2fs (femtoseconds) to ensure the numerical stability of the simulations. Our structural model system was composed of hexagonal Boron Nitride square sheets of 41Å in length and PVA polymer chains made up of 30 repeating units (monomers). To simulate the stress- strain process we first thermalized the entire system for 100ps at 300K, followed by an application of a time dependent externally applied force on each Boron/Nitrogen atoms in the h-BN sheets along opposite directions in order to pull the sheets apart

(see Figure 3d). The force was applied until the sheets broke apart. In this study, we considered three different arrangements of hexagonal Boron Nitride sheets covered with the PVA polymer as described in the results and discussion section.

For the CO2 adsorption simulation, we kept the same parameters and ensemble used in the stress-strain simulation. In the RDF analysis, we first 87

equilibrated the system (depicted in Figure 4c) for 100ps and further take the averages for another 50ps

4.3.6. Properties of Oxidized h-BN/PVA Foam

In Figure 4.5(a), the proposed schematic structure of interconnected h-BN nanosheets with PVA is presented. PVA molecules act as cross-linker to link the h-

BN nanosheets. The lightweight, (0.029 g/cm3) interconnected highly porous foam is shown in Figure 4.5(b), with high mechanical stability. The foam exhibit a brown color appearance unlike the typical boron white color largely due to the

PVA addition through the in-situ freeze-drying method. The structural stability of foam is evidenced by its ability to carry a vial (2.7 g) with no visible structural damage, contrary to what is observed for pristine h-BN (Figure 4.5(c)). A major drawback of freeze-dried formed foams is their disintegration in the presence of liquids, restricting their usefulness in various applications, which can be overcome using PVA addition. When the two foams are immersed in water, PVA connected foam and pristine h-BN foam into water and in less than a minute, the pristine h-BN foam broke into small pieces, unlike the interconnected h-BN/PVA foam (Figure 4.5(d-e)). The absence of interconnected networks in pristine h-BN foam could contribute to the observed structural collapse. Furthermore, the thermal stability of the h-BN is not significantly affected by the addition of PVA, as shown by the TGA thermograms. The low magnification SEM images shown in

Figure 4.5(f) reveal that the porous architecture of h-BN sheets is interconnected. 88

The h-BN/PVA foam at high magnification shows h-BN sheets are connected through thin polymer layers.

Figure 4.5 – Characterization and morphology of h-BN/PVA foams. (a) Schematic representation depicting the h-BN nanosheets connected by 89

polyvinyl alcohol (PVA) molecules. Typical chains contain 30 monomers (~ 6.2 nm chain length). The PVA molecules act like a glue to link the nanosheets through van der Waals interactions between the hydroxyl groups in the PVA and Boron/Nitrogen atoms in the nanosheets. (b) Structural stability of the h- BN/PVA foam. (c) Mechanical stability of the lightweight foam carrying a vial without any visible degradation. (d) Stable structure of the foam when subjected to liquids. (e) Pristine foam disintegrates in presence of water while h-BN/PVA maintains its stable interconnected morphology (f-g) Low magnification and high magnification image of the foam.

Raman spectroscopic analysis is a useful tool to obtain information on the lattice vibration modes of the foams. In Figure 4.5(a), we present Raman spectra of h-BN and h-BN/ PVA foams. For exfoliated h-BN, a characteristic Raman peak at

1366-1373cm-1 is normally associated with the E2g phonon mode140,141. For the pristine h-BN foam, the Raman peak is at 1366.7cm-1 with a shift of 1.5cm-1 from the Raman peak of bulk h-BN (1365.2cm-1). This shift could be an indication of the reduction of the number of layers142 compared to the bulk h-BN. This observation reveals an efficient exfoliation method of the pristine h-BN prior to being interconnected with PVA. On the other hand, the h-BN/PVA spectrum does not show the same h-BN characteristics. The normally low-intensity h-BN mode disappears due to addition of PVA molecules. PVA modes become dominant in the spectrum with a mode centered at 875 cm-1. X-ray diffraction data showed that the majority of crystals are normally oriented along the direction (002)130. This characteristic can be found at 26.7° for both h-BN and PVA interconnected foams, which is suggestive that addition of PVA does not significantly interfere with the h-BN crystallinity (Figure 4.5(b)). Through the oxidation reaction, hydroxyl groups 90

(-OH) were formed on the surface or edge plan of BN, and the result was confirmed by FT-IR. Comparative FT-IR spectra for raw h-BN and oxidized h-BN (f-BN) are shown in Figure 4.5(c). The broad band at 3310cm-1 can be assigned to hydroxyl groups (B-OH) at the edge side of h-BN143. Furthermore, the characteristic peak of B-

N-O appeared at 1100 cm-1144. This result clearly conformed to the presence of hydroxyl groups on the surface or edge side of h-BN via oxidation reaction. The introduced hydroxyl groups on the surface/edge side of h-BN can form the hydrogen bond with PVA145. Furthermore, PVA can acts as an effective cross-linker connecting the joints between individual h-BNs in the h-BN foam, thereby preventing slip behavior and stress concentration between the h-BN sheets and enhancing the mechanical properties of finally produced h-BN foam145,146. Detailed

XPS fitting was done to understand the functionalization of h-BN. The functionalization of BN sheets is evident from presence of a peak at 194eV resulting from the chemical shift of B1s due to B2O3. In addition to this, we also observe chemical shifts due to C-N bond in N1s and C1s, again indicating functionalization

(Figure 4.5(d-f). 91

Figure 4.6 – (a) Raman spectra exhibiting the h-BN peak in pristine foam and a weaker peak from the h-BN/PVA foam due to the addition of PVA molecules. (b) X-ray diffraction pattern of pristine and PVA connected foam showing that 92

the presence of PVA does not significantly interfere with the crystallinity of the h-BN nanosheets. (c) FTIR analysis of h-BN and h-BN/PVA (d, e, f) XPS fitting showing functionalization of h-BN sheets.

The interconnected h-BN/PVA foam is expected to exhibit enhanced mechanical properties. We subjected the foam to compressive loading cycles and as can be seen in Figure 4.5(a). At low loads, the foam appears to maintain its structural integrity unlike pristine h-BN which breaks immediately without taking any load. However, as the load increases, the foam undergoes irreversible damage.

Analyzing individual load-unload cycles, a non-linear failure mechanism in the foams147 like those observed in metallic microlattices148 and elastomers;149,150 a

Hookean region, a plateau and finally densification accompanied with an increase in load is observed. Each failure region is characterized by different failure mechanisms. For instance, linear elastic failure dominates in the Hookean region while the plateau failure is due to post bucking of the foam branches and interconnected nodes147. In the densification region, as has been widely reported, the majority of internal micro-walls break in large numbers without separating from each other rather pilling on top each other, which results in dense loading regions.

The load-unload curves also exhibit hysteresis loops, which indicate better energy absorption. In a detailed study of h-BN foam grown by CVD, Yin et al.147 attributed these hysteresis loops to kinking and buckling of microstructures, friction between branches and cracks developed during initial compression. There is a steep increase in stiffness for the initial 2000 cycles, the increases of stiffness with number of 93

cycles does not seem to end, only reducing the rate of increasing after 2000 cycles.

The interconnected foams therefore exhibit some form of self-stiffening in a dynamic load. The self-stiffening observed here could be due to well aligned PVA molecules which may result in a better load transfer between individual h-BN layers. In addition under dynamic load, the PVA molecules do not have enough time to relax. The fact that there is a large number of layers forming an interconnected network, results in a well-distributed load within the network. However, compaction of the porous structures due to dynamic loads can also contribute to such stiffness especially at high loads. Temperature ramp tests show a linear relationship between storage modulus and temperature (Figure 4.5(c)). The tests were carried out at different single point frequencies, in which the storage moduli increase with an increase in temperature. The lowest modulus is observed for 1Hz frequency and highest modulus at 22Hz, which is the double value at 1Hz.

Frequency sweep tests also exhibit a linear relationship between modulus and frequency (Figure 4.5(d)). Higher modulus at high frequency values could be attributed to the short time range the foam is given to relax. At such high frequencies, PVA molecules, which interconnect the h-BN nanosheets do not get enough time to relax, hence the high storage modulus.

Scanning electron microscopy (SEM) was used to study the morphology of the h-BN/PVA to elucidate their inherent, enhanced mechanical properties compared to the pristine h-BN. As shown in Figure 4.5(e), the foam is highly porous forming a network-like structure. Individual h-BN nanosheets can be seen 94

connected to each other due to the presence of PVA molecules, as clearly depicted by high magnification SEM (Figure 4.5(f)). These types of linkages between h-BN layers is important to design such macroscopic 3D structures composed entirely of their 2D building blocks. The presence of functional groups on the edges of h-BN nanosheets provides anchoring points for the polymer molecule to join them together, thus forming very strong and stable structures. In fact, these interconnections prevent crack propagation along the whole structure by deflecting cracks once they develop within the internal structure. To further shed light on the above-mentioned observations, full atomistic molecular dynamics simulations was done and discussed below. 95

Figure 4.7 – Mechanical response of interconnected h-BN/PVA foams. (a) Typical load-unload curves showing high structural integrity of the foams where they can sustain multiple load-unload cycles without noticeable failure. The foams can sustain loads as high as 0.7N. (b) Self-stiffening behavior of the 96

foam, there is an increase in stiffness over the number of loading cycles. (c) Storage modulus vs temperature at different frequencies, with highest modulus of 15kPa recorded at 22Hz. (d) Storage modulus vs frequency at different temperatures. The foam is able to retain high modulus at high temperature mainly due to thermal stability of h-BN. (e) Low magnification SEM of PVA/h-BN foam showing a network structure. (g) High magnification SEM of PVA/h-BN foam.

To elucidate the role of the PVA polymer in the enhanced mechanical and structural properties of the h-BN/PVA composite simplified structural model as depicted in Figure 4 is used. To mimic the stress-strain response of the composite, three different arrangements of hexagonal boron nitride (h-BN) sheets embedded by PVA polymer molecules are proposed. These arrangements were chosen in order to consider explicitly the interaction between the PVA polymer molecules with either the edges and the basal plane of the BN sheets. The orientations are: Case I -

(Figure 4.5(a)) two neighboring three-layer h-BN anchoring two single layer h-BN sheets embedded with PVA; Case II - (Figure 4.5(b)) two neighboring three-layer h-

BN with their relative edges facing each order and with PVA in between; and Case III

- (Figure 4.5(c)) rotated by 90°. The Figure 4.5(a-c) present for each considered orientation two snapshots: (initial stage) the thermalized structures, at 300K during

100ps; (further stage) the thermalized structures under a small strain before its complete failure. Figure 4.5(a-c) shows that the polymer makes the structures more flexible, presenting a ductile-like deformation, unlike what happens with pristine h-

BN foams, which easily breaks apart for small loading, exhibiting a well-known brittle behavior. 97

To show the enhancement in the mechanical response of the h-BN/PVA composite compared to pristine h-BN foams, MD simulations is done to compare the force necessary to pull these two structures apart (using Case I as example of arrangement), as depicted in Figure 4.5(d). This plot shows the relative displacement of the outer h-BN sheets (center of mass) as a function of the applied time dependent force on this h-BN sheet. This arrangement attempts to mimic the experimental conditions and it is expected to give a reliable comparison (at least in terms of the energies involved) between the h-BN/PVA composite and the pristine h-BN foams. As evidenced in Figure 4.5, clearly the necessary force to pull the h-BN sheets apart is larger to the cases where the foams are filled with the PVA polymer, which results in a more resilient composite in relation to pristine h-BN foams. From the MD simulations inference can be made that strong van der Waals interactions is attributed to the high affinity between the hydroxyl groups present in PVA and h-BN nanosheets, which is consistent with the experimental observations discussed above. 98

Figure 4.8 – (Upper) Snapshots of Stress-Strain molecular dynamics simulations for the h-BN/PVA composite comparing three different arrangements ((a) Case I, (b), Case II and (c) Case III) in two different stages. 99

Initial stage shows thermalized structures at 300 K after 100 ps. The further stage snapshots show the structures under strain. All simulations were performed under a NVT ensemble. Notice the ductile-like deformation of the h-BN/PVA composite. This behavior is unlikely to happen with pristine h-BN foams. (Lower) Relative h-BN nanosheets displacement as a function of the applied force on the upper h-BN sheets. This figure shows that the presence of the polymer (case 2) makes the composite more resilient to deformation (smaller relative displacement) when compared to the pristine h-BN structure.

Lightweight h-BN can be explored for various applications. Here, one application; CO2 absorption is demonstrated. For an effective CO2 absorption material, porosity and chemical stability are important parameters to be considered.

The porosity distribution and surface area of the macroscopic h-BN/PVA foam were measured through nitrogen adsorption/desorption isotherms. As seen in Figure

4.5(a), the curve reveals a type II adsorption/desorption isotherms for h-BN/PVA foams151. Interconnected networks formed by the PVA molecules could be responsible for such high surface areas by connecting the individual layers to an ordered porous structure58. The surface area of the foam has increased from a value of 30.7m²/g to 124.4m²/g after 1wt% PVA is added. The foam revealed a better pore size distribution as calculated from the Barret-Joyner-Halend (BJH) method. As mentioned above, porous and lightweight materials possessing high surface area are crucial for effective CO2 absorption152,153,154. The foam can absorb and store CO2

(Figure 4.5(b)). The foam captured more than 320 mg/g of CO2 at 54 bars. This high- improved CO2 capture and storage is attributed to the enlarged and effective surface 100

area, ordered porosity and presence of nitrogen atoms within the internal structure of the foam Figure 4.5(c).

Figure 4.9 – Surface area and CO2 absorption of the foam (a) Nitrogen absorption isotherm of PVA interconnected foam. (b) CO2 uptake of the foam

absorbing more that 340% of CO2 at 65Bar and comparison with similar recently reported porous materials. (c) Snapshots (time evolution) from 101

reactive MD simulations of the CO2 adsorption processes (300K during 100ps) towards the steady state.

To further elucidate the mechanism of CO2 adsorption in the h-BN/PVA foam, the Radial Distribution Function (RDF) of the PVA polymer in the adsorbed h-BN layer and CO2 in the gas phase, the structural model used in this analysis is depicted in Figure 4.5(c). Reactive molecular dynamics simulations at room temperature (see

Simulation details section). Succinctly, RDF describes how the local time-averaged

(50 ps) density varies as a function of the distance of a reference atom. To make the discussion clearer the RDF is separated into two plots (see Figure 4.5(a)), one describing the PVA polymer/h-BN sheet interactions and the other one for the CO2 adsorbed in the PVA/h-BN composite. Figure 4.5(a) shows the RDF of the equilibrated system of PVA polymer on top of h-BN layer. This plot shows that both hydroxyl (black curve) and CH3 (blue curve) groups have good affinity to h-BN layer, but the hydroxyl groups have stronger interactions with h-BN, in comparison to CH3 groups, as expected. Figure 4.5(b) shows the RDF case of CO2 gas atmosphere adsorbed into the h-BN/PVA foam. This plot shows separately the CO2 interactions with the h-BN layer (purple and green curves) and the PVA polymer (blue and red curves). CO2 shows good affinity in both cases to polymer and h-BN sheets.

However, the CO2 interactions with the hydroxyl and CH3 groups presented in the

PVA polymer are much stronger when compared to h-BN sheets and CO2-hydroxyl groups have the highest affinity of all. 102

Figure 6.0 – Radial distribution function (RDF) of (a) PVA polymer into the

adsorbed h-BN layer, (see Fig. 1a); (b) CO2 in the gas phase into the adsorbed PVA/h-BN composite, (see Fig. 4c). Results from reactive molecular 103

dynamics simulations at 300K. Notice the high affinity between CO2 molecules and the PVA polymer (see text for discussions).

4.4. Conclusion

In conclusion, this chapter has shown that judicious modification of two dimensional materials can be utilized to assemble 3D materials with improved mechanical properties. For instance, the interface engineering between graphene oxide nano-sheets is important in designing lightweight structures that maintain their structural integrity without compromising on some of key physical properties.

Same strategy is used to design h-BN/polyvinyl alcohol foam with high CO2 absorption capacity.

104

Chapter 5

V. Bio-mimicking of ‘Mobile’ Interfaces from Nature: Liquid Metal Composite

105

5.1. Background

5.1.1. Introduction

The previous chapters have dealt with what is called ‘hard’ interfaces.

Natural materials on the other hand have been able to achieve contrasting properties by using what is called ‘mobile’ interfaces. Therefore, this chapter and next will discuss the bio-mimicry of natural mobile interfaces. Firstly, in this chapter exploration of the mechanical behavior of liquid metal reinforced polymers and shed light on its response to dynamic loads. This chapter is based on the paper

Owuor et al., Advanced Materials Interfaces, 4(16), 1700240.

5.1.2. Background to solid-liquid composite

Solid-Solid interface mechanism understanding of composite inclusions, when extended to solid-liquid interface design of composite using Eshelby theory, indicates possibility of decreasing effective stiffness with increasing liquid inclusion in a solid matrix. In contrary, experimental evidence suggests high stiffness and enhanced dynamic energy absorption in a soft polymer (polydimethylsiloxane

(PDMS)) with high bulk modulus liquid inclusions (gallium). The basic deformation mechanism is governed by the hydrostatic stress causing shape change of the liquid inclusion in the large deformation regime and strain hardening of soft polymer matrix. In addition, dynamic visco-elasticity and fluid motion also play significant role. 106

Stiffness and impact strength are among the most important design parameters which can be optimized to a great extent in composite having solid-solid microstructure, nanostructures and their interface155–160. To improve these two properties, higher modulus (stiffness) reinforcements such as ceramic161,162 particles are most commonly added to compliant soft polymer matrices163; improvements of stiffness16,164 up-to ~50% can be achieved by this approach.

Further, drastic improvement can be achieved with the help of addition of nano- sized8,164,165 reinforcement such as silica166, nano-clay161, aluminum oxide167, carbon nanotubes9,10,163,168–170 (single and multi-wall), graphene oxide171, graphene nanoplatelets172, nanocellulose173etc. However, engineering of the interface of reinforcement and matrices in case of such composites to achieve high stiffness and toughness remains a challenge10. A compromise has to be made between toughness and stiffness where one is improved at the expense of the other. On the contrary, nature is replete with examples where both the properties are in optimum e.g. nacre174–177. Natural composites (such as deep sea animal’s skin, as shown on

Figure 5.1(a) are also designed for high stiffness and strength but using a different approach, where immiscible liquid (oil in such case) is embedded in a compliant solid178. Inter-vertebral discs separating spinal cord vertebrae are also designed with an inner liquid nucleus (nucleus pulposus) surrounded by a tough outer coating (as shown in Figure 5.1(a), and type one collagen (annulus fibrous)179,180.

They act as efficient shock absorbers to prevent fractures of the vertebrae due to compression loading and provide mobility to the spine allowing it to bend and twist 107

with minimal damage. These adaptations allow materials to take a very high hydrostatic stress (deep sea animals) high compressive and torsional force181,182

(inter-vertebral discs) by utilizing liquid components.

Although Eshelby’s theory183 predicts a loss of stiffness in a composite due to an increased liquid content, which actually neglects the large deformation of the confined fluid. Our lab has shown recently showed that liquid filled polymeric spheres can improve stiffness of compliant polymer matrix184. In extension, here the use of liquid gallium metal inclusion in a compliant elastomeric polydimethylsiloxane (PDMS) leads to high stiffness and high energy absorption in the composite. Flexible conducting polymers sheet have been made by addition of liquid metal185,186 however detailed mechanical understanding has not been attempted for large volume liquid metal in polymer. The liquid gallium inclusion is assumed to influence primarily through its bulk modulus besides causing interfacial tension due to the metal/metal-oxide and polymer cohesion. Compressive tests show a high effective stiffness of the composite when reinforced with liquid gallium inclusion where the gallium is in liquid state at room temperature (at 30°C). This type of composite exhibits high energy absorption where a spherical inclusion of liquid gallium diffuses in the cracks which are initiated from the interface of the liquid inclusion. To understand this phenomenon, analytical model is developed by incorporating visco-elasticity in the inclusion model. Interfacial surface tension effect which is a quasi-static phenomenon, was identified as one of the factors that enhances the stiffness of the composite187. Additionally, it is observed that surface 108

tension is a small energy contributor in the deformation process. In the case of transient dynamic and high impact energy absorption, experimental results show significantly higher effective stiffness and energy absorption, which is most likely due to mechanisms other than bulk modulus and interfacial tension. Dynamic motion of the fluid and the viscosity could be important factors here also. With the help of the analytical model, fluid-structure interaction model and finite element method for solid and a smooth-particle hydrodynamic method for the liquid inclusion (together called FE-SPH Model) are used. Numerical simulation results successfully explain the large deformation and viscosity dependent phenomena of stiffening and high energy absorption as further discussed in the result section.

5.2. Preparation of Gallium/PDMS Composite

Commercially available PDMS Sygard 184 (Dow Corning) was prepared by mixing 10ml of the monomer to 2ml of the curing agent in a vial. Polydimethylsilane

(PDMS) elastomer was selected due to its optical transparency and excellent elasticity. The mixture was mixed manually for approximately three minutes and later transferred to bath sonicator (Branson 2800) for 15 minutes and later poured into petri dish which acted as our mold. The polymer was semi-cured at 60°C for 5 minutes to increase its viscosity and prevent gallium spheres settling on the bottom of the dish. Gallium (99.99% trace metals basis) from Sigma Aldrich was heated in hot water bath at 80°C for 3 minutes in a syringe. Air was used to blow the gallium into a cold iced water bath where it formed small spheres. These gallium spheres were transferred onto the semi-cured surface of PDMS and final layer of PDMS 109

poured on top. The composite was left to cure under vacuum for 24 hours and post- cured at 60°C for 2hrs. The sample was cut careful to include gallium spheres.

Required dimension samples were cut for compression tests. The neat samples were prepared following the same methodology but with no gallium inclusions.

5.3. Physical Characterization

5.3.1. Mechanical Test

Specimens were cut and tested under load- unload and hysteresis tests were performed. The loading rate for hysteresis test was 1N/min. The test was done in three blocks, first samples were ramped from 0N to 10N, unloaded at same loading rate to 0N, ramped again to 15N, unloaded to 0N and these cycles repeated three times. Dynamic compressive testing was conducted on a TA Instruments Q800 DMA at 5% peak strain. The frequency was maintained at 5 Hz to allow for cyclic compressive loading eliminating the risk of resonance in the sample. All tests were conducted isothermally at ambient conditions. In addition, thermal loading and unloading tests were conducted in the temperature range of -40°C to 100°C. For simulation details, please see supporting information. Scanning electron microscope

(SEM) images were taken on a FEI Quanta 400 ESEM at scanning electricity of 15 to

20 kV.

5.3.2. Optical Observation of Gallium Spheres

A goniometer was used to observe the deformation of gallium sphere in the matrix and also measure the contact angle. Compressive loads were applied on 110

various different samples without and with gallium inclusion in PDMS matrix and deformation was studied.

5.4. Analytical Development

5.4.1. Analytical Formulation

Analytical solution for 2D inclusion problems is developed using Airy stress functions. We assume an infinite domain of matrix with a circular inclusion having dissimilar material and subject to uni-axial compressive load. The relationship between stress components and displacements are expressed as,

∂u ∂v ∂u ( σ = λ + + 2μ x ∂x ∂y ∂x (1)

∂u ∂v ∂v ( σ = λ + + 2μ y ∂x ∂y ∂y (2)

∂u ∂v ( τ = μ + xy ∂y ∂x (3)

Where λ and μ are Lame's constants. The liquid inclusion is modeled by incorporating a visco-elastic model. Shear component of the liquid plays a vital role in defining the fluidic property. The visco-elastic model is introduced in the shear components of stresses and strains by replacing the shear modulus by complex form.

   1 i 0 (

(4) 111

whereη is the visco-elastic co-efficient.

The momentum conservation equations are expressed as

  xy  xy  y ( x   0   0 x y x y , (5)

To deal with the inclusion geometry, the stresses are expressed in (r,θ) coordinate as

2 2 (  r   x cos    y sin   2 xy sin cos (6)

2 2 (   x sin    y cos   2 xy sin cos (7)

2 2 (  r   x sin cos   y sin cos  2 xy cos  sin   (8)

Furthermore, these stresses can be expressed with help of Ariy's stress function ϕ(r,θ)

1  1  2  2 (      r r r r 2  2  r 2   1   , ,  r     (9) r  r   Reported studies (Ref 188) provides a basic formulation in which a matrix with a circular solid inclusion subjected to uni-axial loading is solved using the polarco-ordinate system. We further modify the stresses for the inclusion and the matrix by including the fluidic property of the inclusion. 112

  I 1 K M   I 1 K M cos2 (  I  0    0   rr 2K I  M  4 I  2 M 2K M  I  2 M (10)

1  1 cos2  (  I   1 K M  I     2 0  2 I  1 K I  M K M  I   M    (11)

  K I K M   I sin2 (  I   0 0  r 2 I K M   M (12)

  M K M  M    M  a2  I 1 K M    3a4  I cos21  0  I I  0  I  M  0      0 cos2     rr      2 2r 2 2 I   M  K I  M  2 2r 4 K M  I   M  ( M 3 I    2a  0 cos21  (13)   I    r 2 K M  I   M 

(

(14)

  M  3a4  I cos21  ( 0  I  2 I M M  cos2  a   1 K     0     0 r 4 M I M 2 M I M (15) 2 2r K     2r K    

Where the superscript I indicates inclusion and superscript M indicates matrix. The displacements for inclusion and the matrix system in polar co-ordinates are solved using ϕ which gives, 113

r 1 K I 1 K M ( u I  0    rr 16 I  81 K I  M (16)

M ( I r 1 K sin 2 u  0  4K M  I   M  (17)

2 M 2 I M 4 M a  0 r 0 1 K  a  0 1 K  r 0 cos2 a  0 cos2 urr  M  M  M I M I M  M  3 M I M  4r 8 4r 2    K   4 4r K     ( a4  I cos2 a2 cos2 1 K M  a2  I cos2 1 K M  0  0  0 (18) 4r3 M K M  I   M  4rK M  I   M  4r M K M  I   M 

 a4   I   M  a2r 2 1 K M   I   M  r 2 K M  I   M sin2 ( uM  0         4r3 M K M  I   M  (19)

Where 'a' is the radius of the circular inclusion and σ0 is the applied stress.

The material properties defined by bulk modulus K and shear modulus µ. The strain components in polar co-ordinate system are expressed as

urr 1 u urr u 1 u u (  rr  ,    ,  r    r r  r r r  r (20)

The strain energies of the liquid inclusion in and for the surrounding matrix system are determined as

I 1 2 a I I I I I I  ( U     rr rr       drd 2 0 0   r r  (21)

2 b M 1 M M M M M M ( U   rr  rr     r  r  drd 2 0 a (22)

Where a is the inclusion diameter and b is the sample size. In the above formulation, addition component of energies can be accounted in terms of surface 114

tension coefficient for the inclusion-matrix interface. However, for simplicity, we omit this term, since neglecting such term due to surface tension will give a lower estimate of stiffness and can be a good measure to evaluate the other phenomena involving stiffness mismatch, finite deformation, and viscoelasticity particularly for larger size of inclusion. Viscosity of the gallium is taken form Assael et al189 and shear modulus is taken from Lyapin et al190.

5.4.2. FE-SPH Modeling Scheme

PDMS matrix is modeled using polymer material model. To define the constitutive model of the PDMS matrix, experimental stress-strain curve are used.

Compressive and tensile responses both are required. Compression data obtained from the experiments and tensile stress-strain data from Tae et al.191 are used. The density of PDMS matrix is taken as 970 kg/m3and poison's ratio as 0.44. Matrix is modeled using tetrahedral finite elements, which give constant stress per element and nodal pressure is averaged to alleviate volumetric locking. This kind of finite element formulation is best suited for a material having higher poison's ratio192.

Gurneisen model is used to define the liquid state of Gallium inclusion. The density of gallium used in the simulation is taken as 6102 Kg/m3 and dynamic viscosity is taken as 1.5e-3. The speed of sound in gallium is 2860m/s193 and bulk modulus is determined from this. Smoothed particle hydrodynamics (SPH) elements are used to model fluid region. This is a mesh-free collocation194–196. The interaction between

SPH fluid elements (particles) and the tetrahedral solid elements are controlled by defining adaptive solid-SPH interaction. This method reduces the contact between 115

SPH and solid elements, thus allowing the fluid particles to interact with all solid elements as the solid fracture initiates

Conservation of mass, momentum and energy are written as,

 (  .v  0 t (23)

v (  .v  v  p  0 t (24)

u ( SPH  .u  pv  0 t (25)

Where ρ is the density, t is time and v is the velocity. Fluid Pressure is denoted by p and  represents the tensor product; u is the specific internal energy.

Smooth particle hydrodynamics is a discrete element method, which has the advantage of modeling fluids and solids, which can handle extreme deformations.

Absence of finite element like grid is one of the major differences between the SPH and FEM. The governing equations are resolved on the particles. For brief explanation of the calculation methods please see ref192. A particle approximation function is defined in as

h f (x)   f (y)W(x  y, h)dy ( (26)

Where W is the kernel function, defined as

1 ( W (x,h)   (x) h(x)d (27) 116

where d is the number of space dimension and h(x) is the smoothing length, which varies in time. W(x,h) needs to have maximum at the center of particle. We use a cubic B-spline function for β

 3 3 1 d 2  d 3 for d 1  2 4  ( 1 3 (d)  C (2  d) for 1 d  2 2 (28) 0 for 2  d  

Where C is a constant of normalization that depends on the number of space dimension.

The SPH method is based on a quadrature formula for moving particles

((xi (t)) i 1..N, where xi(t) is the location of the particle i, which moves along the velocity field v. Now the particle approximation can be defined as shown in Eq (36)

N h (  f (xi )   w j f (xi )W (xi , x j h) j1 (29)

m j where w j  is the weight of the particle. The weight of the particle (SPH ) j varies proportionally to the divergence of the flow. The particle approximation for the gradient of the function is

N h m j (  f (xi )   f (x j )Aij  f (xi )Aij  j1  (SPH ) j (30)

1  xi  x  where A   j  ij d 1   h  h  117

and the particle approximation of the partial derivative  is x a

N  f  m j ( h    a (xi )   f (x j A (X i , X j ))  x  j1  j (31)

Where A is the operator defined as

  1 (x , x ) xi  x j A(x , x )  i j   , A is the component α of the A i j hd1(x , x ) x  x  h(x , x )  i j i j  i j  vector.

Finally the SPH estimates of the governing equations are written as

  ( (SPH ) j W ji drj W ji dri W ji drh   mi  .  .  .  t i  rj dt ri dt rh dt  (32)

dv  p p  ( j   m  j  i  W  i  2 2  j ji dt i  (SPH ) j (SPH )i  (33)

dE  dv u dp u d  (  i (SPH )i i (SPHi i    mi vi    dt i  dt pi dt i dt  (34)

drj dr dr Here  v , i  v and h  0 dt j dt i dt

SPH formulation is used along with a equation of state in LS-DYNA. The

Gruneisen equation of state with cubic shock velocity-particle velocity defines pressure for compressed material as 118

2    0  a 2  0C c 1 1 c  c    2  2  ( P      E  2 3  0 c I c c (35) 1 S1 1c  S2  S3 2   c 1 c 1 

1 c  1 The compression μc is defined in terms of relative volume V as V , and Ec is the internal energy per initial volume , C is the intercept of the curve, S1, S2,

S3 are the coefficients of the curve. γ0 is the Gruneisen gamma and a is the first order volume correction to γ0. Constants C, S1, S2, S3, γ0 and 'a' are user defined input parameters.

5.5. Mechanical Behavior of Liquid Metal Composite

Figure 5.1(b) shows the compressive load-displacement behavior of a liquid gallium (Ga) spherical (volume fraction of 43.70 ± 2.28 %, 3.11 ± 0.15 mm diameter) inclusion into an elastomer PDMS matrix as compared to that of pristine PDMS.

Initial compressive elastic stiffness of PDMS-Ga is significantly higher compared to that of PDMS. PDMS show strain hardening which is typical to such polymers. On the other hand, the PDMS-Ga composite shows a metal-like behavior with a high compressive modulus with yielding and softening, but most of which are recoverable by unloading. PDMS-Ga modulus is almost ten times higher compared to PDMS. Dynamic mechanical analysis (Figure 5.1(e) and (f)) clearly shows higher stiffness of PDMS-Ga compared to PDMS. It does not show plasticity. We observe high stiffness of the samples with gallium inclusions in both low and high loads. An 119

improvement in stiffness was up to four times higher for gallium inclusion samples after only a second loading cycle (inset Figure 5.1(e)). Similar experiment with water sphere shows a decrease in stiffness hence showing high stiffness can only be achieved with high bulk modulus liquids. Controlling arrangement of liquid gallium inclusions in hierarchical manner further shows that three layers have highest stiffness (Figure 5.1(f)). The geometric confinement related phenomena of increased stiffness and similar increase in the dynamic stiffness due to viscosity are not captured in the standard dilute mixture model of effective elastic composite.

Hence the experimental results cannot be elucidated with the classical rule of mixture197. To develop our understanding of the mechanism involved in high stiffness and energy absorption due to the addition of liquid gallium inclusion in polymer, we first analyze the interaction between the Ga and polymer using contact angle tests. Figure 5.1(c) and d show representative images of the contact angles for liquid gallium on the polymer (PDMS). Liquid gallium drop shows incomplete wetting on the polymer (Figure 5.1(c) and (d)) with a contact angle of approximately 160o. Liquid Ga rapidly forms thin (<~5nm) oxide layer which protects further oxidation of the metal in static condition of loading. The thickness of the oxide layer is in the order of nm and is much smaller than Ga sphere having diameter in the range of mm. We do not consider this oxide layer in mechanical analysis because the critical yield surface stress is 0.5−0.6 N/m197, above this stress, the oxide skin ruptures. We believe the oxide layer had insignificant effect on 120

behavior of gallium spheres in loading as the stress used was in orders of magnitudes above this critical value.

PDMS shows classical polymer behavior while PDMS-Ga shows metal like behavior with high modulus. (c, d) Contact angle tests showing high hydrophobicity between gallium and PDMS. The high hydrophobicity is maintained for a long time showing there is no chemical interaction between the two materials. (e) Dynamic compressive tests illustrating the stiffening phenomena of PDMS due to the addition of liquid metal gallium. Stiffness is increased almost four times compared to pristine PDMS matrix. (f) High stiffness of PDMS-Ga is also observed when gallium spheres are arranged in a PDMS matrix, three layers of gallium spheres exhibit high stiffness compared to two layers.

In order to understand the strain distribution in PDMS-Ga composite as compared to PDMS, scratch lines are drawn on both samples and then samples are subjected to compressive load and deformation pattern are imaged. A comparison of pristine PDMS and PDMS/Ga clearly shows that the geometric shape combined with the liquid-polymer stiffness at the interface seems to have important role which needs a more detailed mechanistic understanding. We observed this effect independent of loading direction. Furthermore, the gallium spheres seem to pull the

PDMS elastomer near the interface and bend inwards instead of outward as is typical with conventional materials when under compressive loads. In the process of liquid gallium spheres deforming, they take the excess stress from the PDMS elastomer matrix forming an ellipsoid shape which returns to their original spherical shape once the load is removed. The deformation was radially compressive along the loading path and tensile on lateral sides against the compliant surrounding of PDMS matrix. The radial expansion is necessary for energy absorption180. Additionally, a higher energy absorption and dissipation 122

capability of the infused gallium samples is demonstrated by viscoelastic tests. High energy absorption and dissipation are the characteristic of highly elastomeric rubber198 which has an ability to undergo large deformation under load and return to their original shape. In essence, the gallium inclusion in the PDMS elastomer transform the elastomer into a system that can dissipate even higher energy by internal viscosity of the liquid and fluid flow under dynamic condition. Besides that, the directional distribution of the strain paths involving geometric non-linear effect becomes useful here. Eshelby's theory predicts lower effective stiffness of composite compared to that of solid matrix when liquid inclusion with bulk modulus is considered. According to this theory, the effective stiffness reduces with increasing volume fraction of liquid inclusion. In Eshelby's theory the effective strain of the composite is related to Eigen strain of the inclusion through Eshelby's tensor (P), where P is the function of geometry of the inclusion. Hence the effective modulus depends on geometry of the inclusion and its volume fraction. Here it is expected that the higher the volume fraction of the liquid inclusions, lesser is the effective stiffness199. However, Style et. al 200have shown recently that the small radii inclusions do not follow Eshelby’s theory. They showed that small isolated droplets are better in resisting deformation in a soft matrix but not a stiff matrix. This is because of emergence of surface tension as the dominant effect, as surface area to volume ratio increases with decreasing radius of sphere. However, effects of larger size of inclusions (mostly observed in nature) and dynamic loading are not yet understood. In our present study, we observe high stiffness of composite with liquid 123

inclusion in soft matrix even for large size of sphere and under high load (large deformation and dynamic load).

To understand the role of strain gradient surrounding the liquid inclusion on the effective properties, we patterned the same volume of Ga into four spheres each of 1.5mm diameter separated by a distance of ~1.5mm from each other. The compressive force vs displacement curve is shown in Figure 5.2(a). It clearly shows even higher modulus of four-sphere composite as compared to a single sphere composite. The in-situ imaging during loading as shown in Figure 5.2(b-d) reveal changing shape from sphere to an ellipse which is completely recovered upon unloading. We observe strain localization near Ga and the strain lines connecting the two spheres. Such type of strain path can contribute in higher strain energy density201. We have shown with contact angle results that there is incomplete wetting between liquid gallium and PDMS, which is the case of larger diameter sphere with relative flat surface (as compared to surface tension within smaller confinement). But it is clear that there is a better load transfer from the elastomer to the liquid metal spheres when they undergo large deformation, which is without plastic yielding. We believe that the shape of the gallium inclusion in the matrix plays a critical role in enhancing the stiffness by smoothing the ripples and stress concentration in the unit cell of composite. In other words, it shows transfer of load from top to bottom layers during an impact from the top, which is caused by conversion of the compressive dynamic strain into other components of strain at the liquid interface and within the inclusion. In order to quantify such behaviour, we 124

measured storage modulus as a function of temperature (Figure 5.2(e)). Effective storage modulus increases in the PDMS-Ga while it reduces in pristine PDMS.

In order to demonstrate the impact resistance, we further performed impact dynamic test with the samples (Figure 5.2(h)). The composite shows a high damping during impact, the Ga inclusion fractures and spread like a sheet inside the polymer

(Figure 5.2(f) and (g)). The optical image shows the fracture pattern near individual

Ga spheres when several drops are embedded in a PDMS matrix. The liquid Ga inclusions absorb the impact energy and do not allow the load to transfer to the base leading to lateral propagation of cracks in the material (Figure 5.2(i)).

125

individual Ga. The impact energy is absorbed by liquid Ga and does not allow the load transfer to the base of PDMS.

5.6. Analytical model based prediction

To obtain a closed form solution to the deformation behavior of the liquid inclusions in a polymer matrix, we consider an isotropic elastic plate problem with a circular inclusion with a dissimilar viscoelastic property (Figure 5.3(a)). Closed form solution for a centre hole problem was obtained for various boundary conditions using Airy's stress function approach188. In this approach the stresses and displacements are derived considering displacement continuity at the interface of inclusion and the matrix. Far-field uniform uniaxial stress is applied with vanishing shear stress at the outer boundary of the matrix. We have introduced complex shear modulus for the inclusion and the matrix material by including the viscoelastic loss parameter, which are to capture the effect of fluidic inclusion as well as the visco-elastic behavior of the PDMS matrix system. The strain energy developed in the liquid inclusion and the PDMS matrixes are calculated (see detailed formulation in Supplemental information). Figure 5.3(b) shows the strain energy density (normalized with density) developed in the region of inclusion filled with different media. Figure 5.3(c) shows the energy density of the matrix region. The liquid inclusion of Gallium in PDMS absorbs more energy as compared to pure PDMS matrix and air filled PDMS matrix and in all these cases the matrix region absolves almost same energy. From the above observations, it is clear that under quasi-static 127

stress, most of the energy developed due to applied stress are taken into the liquid inclusion, which makes the PDMS-Ga system a more energy absorbing system than air filled one and pure PDMS

128

129

5.7. Finite Element (FE) Modeling with Smooth Particle

Hydrodynamics (SPH) Coupling for Liquid Inclusion

In order to understand the large deformation and dynamic behavior of the composite, which becomes difficult to model analytically, we used FE-SPH modeling scheme for gas/liquid inclusion in solid matrix. Details are given in supporting information. We compare the experimentally obtained compressive stress-strain behavior of the composite under quasi-static loading with simulation results based on analytical model (linear-viscoelastic-viscoplastic model) and the FE-SPH model

(nonlinear viscoelastic model). Fluid motion can be considered viscoplastic because there is no elastic equilibrium. Figure 5.4(a) shows good agreement of experimental results with simulation results where simulation results tend to match with experimental results as the strain becomes large. Higher magnitudes of strain will cause large deformation as well as fluid motion, which are studied later. Figure

5.4(a) also clearly shows a higher stiffness compared to pure PDMS and even PDMS

 u  filled with air. When engineering strain  Engg   is measured on the sample  x 

1 composite, the large strain (Green strain) is given by     2 . Form L arg e Engg 2 Engg 130

    1 this, the error in strain measurement can be estimated as  Larg e Engg    . We   Engg   Engg  2 identify 5% or higher error as indication of large strain regime in the following analysis. Since the Ga inclusion is liquid of high density and viscosity, two additional important aspects can be of further advantage, one is the fluid dynamic viscosity effect in developing interfacial shear stress based energy absorption and the other is the fluid motion (sloshing and circulation). To demonstrate these advantages, we apply impact dynamic load on PDMS-Ga composite. An indenter plate having 5 kg mass is considered to apply impact force from a 60 cm height as in the experimental condition. At the bottom of the sample, we keep a solid plate as an armor plate

(made of Al) which is modeled as elastic plate. We define the impact energy transmission as the ratio of total energy of bottom armor plate and the input kinetic energy. A significant drop in the impact energy transmission through PDMS-Ga composite is observed as seen in Figure 5.4(a). In Figure 5.4(a) and (b), the time at

A1 marks the onset of contact of the impactor or plate on the sample. t=0.75ms indicates onset of large strain as explained earlier. Figure 5.4(c) and (f) shows the

Von-Misses stress path for the different cases. Shear bands are more prominent on the sides of the Ga inclusion as compared to air inclusion indicating that shear energy in polymer also become strongly connected to the deformation mechanism of the inclusion.

131

spheres having same volume as single sphere. (c - f) Von-Misses stress distribution for inclusion material as PDMS, Gallium, Air and Gallium with 4 spheres having same volume as single sphere at time t=0.75ms, velocity of the impactor is 3.428 m/s.

Portioning of the kinetic energy and the strain energy in the samples are shown over elapsed time in Figure 5.5(a) for PDMS matrix, Ga inclusion and PDMS-

Ga composite medium. Figure 5.5(b) shows zoomed graph of the kinetic and strain energies. At t=0.75 ms the kinetic energy in the inclusion builds higher than the matrix medium, indicating more energy absorption into the inclusion. At t=1.25 ms the kinetic energy in the inclusion greater than that of the matrix medium, at this stage the inclusion will start exhibiting fluid circulation as shown on Figure 5.5(c) and (d). Figure 5.5 shows the von-Misses stress distribution in PDMS-Ga system and the velocity vectors at various regions inside the fluidic inclusion.

Figure 5.5(e) shows the variation of total energy density (normalized with mass density) as function of inclusion size. Eshelby’s theory predicts decreasing total energy density as the inclusion size increases, whereas the analytical model based on airy’s stress function shows increasing total energy. The total energies will also increase with increasing viscosity of the liquid. The FE-SPH simulation predicts nearly linear increase in the total energy of the system as the inclusion size is increased. Inclusion size as length scale can be divided into three zones based on the mechanism of energy transfer between the inclusion and the matrix. First zone is surface tension dominated region indicated by A in the figure where inclusion size is very small200. The second zone is viscosity driven where the energy transfer/ 133

absorptions are driven by the viscosity of the liquid and strain rate due to dynamic loading. The third zone is driven by fluid flow inside the matrix inclusion involving large strain in the matrix and rotation of the fluid in the inclusion.

Figure 5.5 – Kinetic energy study of the Composite. (a - b) Development of kinetic energy and strain energy in PDMS matrix, Gallium Inclusion and in 134

PDMS-Ga Composite system. Kinetic energy and the strain energy are normalized with the impact energy provided. (c - d) Von-Misses stress distribution in PDMS matrix and the resultant velocity of the Gallium particles

in the core at time tA= 0.75 ms and tB= 1.25 ms respectively. (e) Comparison specific total energy density obtained from of Eshelby theory, present Airy's stress function based analytical model and FE-SPH simulation model for varying core size. Variation of increasing viscosity in the analytical model is also shown. Based on inclusion interaction behaviour with the matrix three region can be defined, region A where surface tension of the inclusion play vital role. Region B where viscosity of the inclusion plays important role in energy transfer. Region C where Fluid flow of the inclusion derives the energy transfer mechanism.

5.8. Conclusion

In this chapter, it has been shown that it is possible to design novel composite materials exhibiting higher stiffness by the addition of liquid inclusions.

Based on classical composite theories, this is an unprecedented phenomenon because they predict lower stiffness with liquid inclusions. It is believed that liquid metals and other liquid materials with high bulk modulus and high viscosity can be used to design high energy absorption and high stiffness materials. Like the inter- vertebral discs, it possible to design composite materials with liquid inclusions but still providing the required mechanical stiffness and strength. These results can form a basis for the development of new multifunctional materials utilizing liquids inclusions spanning over very small to very large length scales. Further, these types of composites can find wide applications, for example as high energy absorption materials, shock absorbers and in designing bio-mimetic structures such as artificial inter-vertebral discs. 135

Chapter 6

VI. Bio-mimicking of Multi- functionality in Nature: High Stiffness Polymer Composite with Tunable Transparency

136

6.1. Background

6.1.1. Introduction

Like the previous chapter, this chapter discusses the bio-mimicking of multi- functionality normally associated in natural materials. In the same breath, a mobile interface between a polymer and reinforcement is employed. In this case, semi- crystalline polymer (paraffin wax) is used as reinforcement components. This discussion is based on a paper; Owuor et al., Materials Today, 5(21), pp. 475482,

2018.

6.1.2. Background on Multi-functional Materials

Multifunctional materials possess unique properties where more than one function is integrated within the material157,202. These functionalities can include magnetic, electrical, optical and power generative working in synergy in an efficient manner. These kinds of materials present tremendous advantages for design of high performance structural materials by leveraging on weight, power consumption, strength, etc. Nature is replete with multifunctional materials203,204 performing more than one specific function12,205,206. Biological materials possess healing, sensing, load bearing and other functionalities which are inherently built within an organism’s structures207–210. In most cases, these natural designs can be tuned for specific functions and environments. For example, animal muscles stiffen209 in response to applied load, and some plants open and close their leaves in response to touch. These processes are controlled by various stimuli such as touch, heat and 137

light. The marine animals known as cephalopods’ skin for instance not only protect, regulate temperature and used it as a sensing tool but also act as an excellent camouflage211 mechanism by rapidly changing their color depending on their surrounding environment. The color changes are light induced and brought about by cytoelastic sacs of pigment called chromatophore211, which can be tuned to display any color. Scientists and engineers frequently attempt to mimic nature’s complex structural designs212 to solve complex human problems12,213 though this has proven a daunting task over the years.

Recently various multifunctional materials have been designed and proposed11,214. The majority of these systems integrate a traditional material with an additional functionality. As an example, multi-functionality of gold can be realized at the nanoscale by designing nanoscale particles that can be used for optical imaging215, drug delivery and cellular uptake216. Functionalized carbon- based nanomaterials217–221 are another important multifunctional materials that offer combinations of high mechanical properties, optical tunability, and superior electrical and thermal properties222. Furthermore, hierarchical multiphase19,219,221,223–226 materials have been designed by utilizing metallic- intermetallic laminate227 which can be tailored to perform more than one function.

Engineered materials with tunable properties rely heavily on a switchable chemistry and morphology228 to create a change in their functionalities such as strength, toughness, stiffness and transmittance229. Most of these multi- 138

functionalities have been demonstrated by nanomaterials, but it remains a grand challenge to achieve the same abilities for macro-scale materials.

This work demonstrates a phase change (pc) composite at the macro-scale by combining a phase change material wax and polymer matrix

(polydimethylsiloxane). High stiffness and fast-changing optical transmittance properties of the pc-composite is shown. The shape controlled paraffin wax embedded in a stiff elastomer matrix, achieves very high stiffness and unlimited optical switching.

6.2. Preparation of Phase Change Composite

Polydimethylsiloxane (PDMS) (Sylgard 184, Dow Corning), iron oxide nanoparticles (purity: 99.5+%, particle size: 15-20nm, US research nanomaterials) and paraffin wax (Sigma Aldrich) were used as received without any further purification. The pc-composite was prepared by mixing PDMS with a 10 wt% wax.

First, PDMS resin (10 ml) was mixed with wax (10 wt. %) and heated at 70°C to melt the wax and allow homogeneous mixing. PDMS hardener was added and mixed thoroughly (10:1; Monomer: Hardener). Bubbles due to mixing were removed using a vacuum oven, and then poured into petri dishes to cure at 100°C for two hours.

For noncontact heating, the loading of the magnetic nanoparticles (MNP) in the pc-composite was varied in the range of 0.5 to 1 wt% to obtain maximum transparency in the temperature range from 45 to 50°C. 0.5 wt% led to insufficient heating while samples with 1 wt% of MNP were opaque even at high temperature.

Therefore, 0.75 wt% MNP were selected and loaded in blended Pc-composite 139

composite. The MNP were added to the PDMS + 5 wt% wax liquid mixture at ~

50°CC and stirred with a nonmagnetic spatula. 10 wt% of wax mixing in PDMS with equal amount of MNP results in a temperature rise of only up to 40°C, therefore, 5 wt% wax in PDMS was selected. For curing, the samples were then placed in vacuum box for 5h.

6.3. Physical Characterization

Optical transmittance test was performed on a hot plate where we increased the temperature and observed the changes of the composite from translucent to transparent. Noncontact heating of the samples was carried out in a water-cooled 5- loop copper induction coil, energized by an AC generator (Inductelec, UK), at an operating frequency of 375kHz and AC magnetic field (AMF) of 4kAm-1. A Luxtron

MD600 fiber optic thermometry unit connected to a computer was used to measure the sample temperature at 1s intervals throughout the experiment.

FTIR testing was done on a Nicolet FTIR Infrared Microscope. Optical images and contact angle tests were performed on a Goniometer. Thermogravimetric analysis (TA Q600) equipped with a DTA was used to study thermal stability of the composite and study the changes in heat flow by subjecting the composite in temperature ramp test (30 – 700°C at 10°C/min). Scanning electron microscope

(SEM) was performed on a FEI Quanta 400 ESEM, at scanning electricity of 15 to

20kV. Dynamical mechanical analysis tests were conducted on a DMA Q800 (TA,

USA). All the tests were conducted in compression mode at room temperature 140

conditions. The temperature ramp tests were done from 30 - 150˚C at ramp rate of

3˚C/min. The impact test was done by a home-made impact machine.

Spectroscopy was performed on a Cary 5000 UV-Vis-NIR spectrophotometer from Agilent Technologies. A specialized sample holder was used to hold the polymer sample in the beam path during the measurement. For the heated composite experiment, the polymer was placed for 10 minutes in a beaker of water maintained at approximately 75°C. The polymer was dried and then placed on the sample holder prior to performing the spectroscopy.

Transmission-mode brightfield imaging was accomplished on a Nikon Eclipse

Ti-U microscope equipped with a 10X objective, a standard halogen lamp, and a Pro-

EM HS camera from Princeton Instruments with a 1024x1024 pixel array. Similar to the spectroscopy experiment, the sample was first heated for 10 minutes in a hot water bath maintained at 85°C prior to the collection of images.

6.4. Simulation Details

The simulated model consisted of a polymer sphere being dropped from a height several times its diameter, accelerating down in the vertical direction due to the force of gravity, impacting a stainless steel plate at the bottom, and bouncing back up to a certain height. The only force acting on the sphere was gravity and the simulated drop test was performed without any friction forces acting on any objects, which is an ideal situation used to study the effect of the mechanical properties of the material isolated from other interactions with the environment. 141

By virtue of symmetry only half of the sphere is modeled with the appropriate boundary conditions on the symmetry face. The finite element method simulations and analysis were carried out using Abaqus software32 [32].The materials properties were assigned as follows: PDMS was given a density of

0.97kg/m3, a Poisson’s ratio of 0.4532, and an elastic modulus of 1.76kPa; Pc- composite was given a density of 0.97kg/m3, a Poisson’s ratio of 0.45 32, and an elastic modulus of 22.37kPa; finally, the metallic plate was assigned the materials properties of stainless steel 302, with a density of 7861kg/m3, elastic modulus of

193GPa, and Poisson’s ration of 0.2533.

6.5. High Stiffness and Tunable Transparency

Figure 6.1 shows an artistic visualization of pc-composite optical switching capability in two different conditions. First condition shows the heat induced changes or contact heating (Figure 6.1(a)) and applied magnetic heating changes or non-contact heating (Figure 6.1(b)). In both conditions, the material turns transparent in a reversible process. Tunable transparent is demonstrated by contact and non-contact methods. Contact method is discussed first for the Pc-composite subjected to changes in temperature. 142

The Pc-composite exhibited fast switching between translucency and transparency. The blended composite of pc-composite has a translucent appearance at room temperature, but as the temperature increases, it becomes transparent

(Figure 6.2(a)). A transparency of 80% with a peak at 45°C is observed (inset of

Figure 6.2(a)). Optical images provide evidence for the theory that the wax is homogeneously dispersed within the composite. As depicted by the optical images, it is observed that the change in transparency initiates from the bottom of the 143

sample and progresses upward, which the result of directional heating is provided by a hot plate. In the solid state both constituent materials have different refractive index (PDMS = 1.4 -1.42, wax = 1.442-1.448). In the liquid form, the refractive index of wax closely matches with that of PDMS, which result in transparency230,231. To examine structural integrity, the material is designed into a pipe shape and exposed to hot water (Figure 6.2(b)). The optical property of the pipe changes from translucent to ‘transparent’ while retaining its shape even after hundreds of thermal cycles in hot water.

In order to determine the optical transmission over a range of wavelength as a function of temperature, a series of tests were performed with an UV-Vis-NIR spectrophotometer. The room temperature profile yielded a percent transmittance below 10% across the entire spectrum. Upon being heated, the pc-composite profile closely resembled that of the PDMS-only. The topographic profiles of these two spectra are nearly identical, and the Pc-composite profile has a percent transmittance nearer to 80% for much of the region between 300 and 1200nm.

Following the tests on the spectrophotometer, the pc-composite composite was heated in a water bath at 85°C and then placed immediately onto an optical transmitted-light microscope operated in brightfield mode. The purpose of this experiment was to image changes in morphology as it cooled and to simultaneously monitor the change in transmittance. Interestingly, the initial transmittance measurement made 60 seconds after removal from the water bath was 33%, but from the 2 minute mark until 6 minutes, it remained steady near 60%. This initial 144

dip in transmittance was observed during repeated experiments. After the 6-minute mark, the transmittance gradually decreased over the course of several minutes.

The corresponding brightfield images show a lack of small features in the polymer through the first seven minutes, but at the 8-minute mark, small structures become prevalent throughout the imaging area. These structures continue to evolve through the rest of the experiment. When the paraffin wax is mixed with the polymer, the two materials form an immiscible polymer composite. Free volume within the polymer matrix may be responsible for providing regions where the macro/nanoscale wax particles can adhere to. The transparency at high temperature suggests homogenous mixing. The non-contact heating method to tune optical transmission is next discusses next.

To achieve this, magnetic nanoparticles (MNP) are dispersed in the pc- composite. The pc-composite/MNP has an opaque appearance at room temperature.

Under the applied magnetic field (AMF), the temperature of MNP rises by hysteresis and relaxation losses, which leads to a transparent composite. After ~ 8 minutes, the temperature reaches a steady state and the wax completely transforms from solid to liquid. As in the case of direct heating, the refractive index of wax closely matches that of polymer, thus, the composite becomes semi-transparent with a transparency of more than 50% (Figure 6.2(c) and (d)). After removal of the AMF, the pc- composite/MNP returns to its initial opaque state. Multiple cycles were applied and removal of AMF and found that the process is completely cyclic, i.e., opaque without

AMF (0) and semitransparent with AMP (1). The lower transparency in non-contact 145

heating relative to contact heating is due to the presence of the opaque MNP in the composite. The advantages of AMF heating include homogenous and local heating due to the dispersion of MNP in the composite. The optical switching under AMF is remotely accessible and can be used in applications where direct heating is challenging.

146

Optical switching behavior of a pipe made of up pc-composite showing its ability to change into transparency in presence of hot water and return to its initial condition when immersed in cold water bath. (c) Optical switching behavior of the pc-composite with 0.75 mass% loading of magnetic nanoparticles with application and removal of AC magnetic field, showing the transition from opaque to translucent on application of AMF of 4kA/m and frequency 375kHz. (d) Gradual increase in optical transmittance with time by a noncontact heating through the application of AFM. Up to ~ 50% transparency was observed, the sample returned to its initial (opaque) condition when the AC field was removed. The ‘0‘ and ‘1‘ represent the application and removal of AFM.

To qualitatively determine the effect of wax in polymer, we conducted thermal stability tests using a thermogravimetric analysis (TGA) instrument where samples were subjected to temperature ramp tests. We monitored the weight changes for all samples as a function of temperature. Though wax does not lower the onset degradation temperature of pristine polymer, the pc-composite has a lower decomposition temperature than pristine polymer (Figure 6.3 (a)).

Differential thermal analysis (DTA) has two peaks in wax that are associated with melting and sublimation (Figure 6.3 (b)). However, these peaks are completely suppressed in the pc-composite. One reason for this could be the encapsulation of wax particles within the polymer matrix which shield them from high temperature.

Presence of wax in the matrix has negligible effect in its ability to operate at high temperature. 148

Dynamic loading at ambient conditions shows stiffening behavior of the composite as the number of loading cycles increases (Error! Reference source not found. (a)). In comparison with pristine polymer, we observe an improvement in stiffness of more than one order of magnitude. Further improvement in stiffness is exhibited by the pc-composite at higher temperature of 35°C. The high stiffness of the composite could be due to the realignment of crystals within the wax. At 35°C, wax starts to melt where realignment of crystals is aided further by allowing them more space to rearrange them. High stiffness of pc-composite is observed in both high and low loads (Error! Reference source not found. (b)). Compression test shows high modulus of pc-composite (Error! Reference source not found. (c)). The pc-composite exhibits a linear stress-strain curve unlike the pristine polymer. The 149

latter shows non-liner behavior with low modulus and high modulus as loading increases. Furthermore, we monitored the two functions of the material to truly demonstrate how they are related with temperature. As depicted on Error!

Reference source not found. (d), the two functions, stiffness and optical transmittance increases (inset, Error! Reference source not found. (d)) with an increase in temperature. Contrary to pc-composite, pristine polymer exhibit a loss in stiffness while optical transmittance remained unchanged as the temperature increases. Loss of stiffness with increase in temperature is an expected phenomenon in polymer materials due to plasticization. The concurrent improvement in stiffness and optical transmittance is unique and advantageous in the quest for the design of multifunction materials. 150

Figure 6.4 –Mechanical behavior of pc-composite. (a) The pc-composite showing the spherical shape of the wax within the polymer matrix. (b) Load vs stiffness (c) Compression test showing high modulus of the composite. (d) Increase of both stiffness and optical transmittance of the composite as the temperature increases.

Eshelby theory 183 predicts lower stiffness due to addition of liquids but recent reports 184,232 have shown higher stiffness in polymer matrix by addition of liquids. To elucidate on this behavior of high stiffness as result of addition of liquid 151

wax in the PDMS, in-situ optical imaging under compression load at various temperatures is performed Figure 6.5(a)). The wax sphere starts to melt directionally as the temperature is raised and when cooled, it transition back to its original shape. This experiment revealed that heat can easily be transferred through the polymer matrix. Moreover, the wax remained intact within the polymer as a transparent liquid. The wax remains as isolated homogeneously dispersed polymer particles. These isolated particles absorb latent heat, melt, and change phase from solid to liquid. Deformation of the wax sphere inside the polymer reveals high strain concentration 233 development on the surface of solid wax as it turned into an elliptical shape Figure 6.5 (b)(I)). On the contrary, there is minimal strain concentration on the surface of liquid wax sphere Figure 6.5 (b)(II)). The high strain developed in the PDMS is easily transferred to the liquid sphere hence it can take more load compared to solid wax sphere. Load transfer is aided by the compliant nature of wax sphere which can change shape without any noticeable yielding. To further clarify on the deformation behavior of wax sphere as a solid or liquid inside the polymer matrix, difference in angles due to deformation was analyzed. Assuming a normal force (insert Figure 6.5 (c)), changes in angles θ1 and θ2 can be monitored with respect to force F1. By simple geometry, it can be shown that F, F1 and θ are related by 퐹 = 퐹1 cos(휃). From this equation, force distribution on the sphere can be calculated. Setting F as starting point, it can be observed that the angle on the liquid portion is always greater than the solid one. We measured the initial angles

(θ1=54, θ1= 72) to get the force behavior between the two F1 forces (liquid and 152

solid). Fixing a constant force F, we observe F1 on the liquid sphere is always higher than that on the solid with change in angles Figure 6.5 (c)). This further ascertains our assumption of better load distribution of wax liquid sphere within the polymer matrix. In addition, under dynamic loads, polymers chains will always try to align themselves in the direction of load. For a pristine polymer this will take place before the polymer start to harden Figure 6.5 (d)). On the other hand, presence of wax spheres in the polymer matrix will tend to aid the process on polymer chains arraignment around it thereby resulting in high modulus Figure 6.5 (e)). 153

Figure 6.5 –(a) Non-uniform changes in optical transmittance of a wax sphere inside polymer matrix. This arises from the non-linear heat transfer from polymer matrix to the wax sphere within it (b) Deformation of solid wax inside polymer where it turn into an elliptical shape in response to load with high stress concentration on the surface. (I) Minimal stress concentration on the surface of liquid wax sphere. (II) The high stress developed in the polymer is easily transferred to the liquid sphere thereby taking more load. (c) Relationship between force and change in angles in the wax sphere within the polymer matrix. Liquid sphere exhibit high angle change which result in high force F1 as a function of applied force (F). 154

To demonstrate the effectiveness of pc-composite in energy damping, we fabricated spherical shape samples and subjected them to drop test (Figure 6.2(a)).

We varied the drop height and observed minimal energy loss when the sphere is heated above the wax’s melting point (Figure 6.2(b)). To further demonstrate the ruggedness of the pc-composite, a thin film was used as a coating layer on a glass plate (25x75x1mm) and subjected to an impact test by dropping a steel spherical ball (1.04g, ∅ = 6.32mm). A pristine glass slide failed at a drop height of only 20cm, while the pc-composite coated glass slide did not show any observable failure for a drop from as high as 50cm (Figure 6.2(c)). The pc-composite thin film is therefore able to absorb much of the impact energy and protect the glass substrate underneath.

The mechanical response of the two materials was further studied by a simulated drop test using finite element method (FEM). Two spheres with material properties of polymer and pc-composite were dropped from various heights (inset of

Figure 6.1 (d)). These heights were normalized by the sphere’s diameter

(Height/Diameter), and chosen to be 5, 10, 15, and 20 times the sphere’s diameter.

From every starting height, the Pc-composite composite was able to bounce higher than the pristine polymer. Figure 6.2 (d) shows the difference in final height (ΔH) between the pc-composite and pristine polymer spheres, normalized by the sphere’s diameter (ΔH/D). The normalized difference in height (ΔH/D) becomes more prominent as the initial height increases, up to as much as 25% of the initial height.

Due to a higher value of elastic modulus, the pc-composite sphere does not suffer as 155

much deformation during impact as the pristine polymer sphere (Figure 6.2(e) and

(f); units: Pa). Greater deformation of the pristine polymer sphere causes an energy loss sufficient to prevent it attaining the same height as the pc-composite sphere.

This is demonstrated by the strain in the sphere at the maximum point of deformation during impact shown in Figure 6.2(f). It is shown that the pristine polymer sphere suffers greater deformation at every point (points 1-6 mapped inset of Figure 6.2(f)). Additionally, the maximum point of deformation on the pc- composite composite sphere is closer to the edge of the sphere than the pristine polymer sphere, so that less part of the material suffers deformation during impact.

The potential energy of the spheres is converted to kinetic energy as the objects are accelerated due to the force of gravity; however, the displacement of the molecular chains within the pristine polymer sphere causes a higher loss of energy than less- deformed Pc-composite composite sphere. 156

Figure 6.6 –Impact behavior of the pc-composite. (a) Schematic representation of the drop test of composite blend sphere when cold and heated above melting temperature of wax. The two spheres are dropped from different heights and bouncing height recorded (b) Low energy loss of the pc-composite above the melting temperature of wax. At room temperature the pc-composite show comparable energy loss with pristine polymer. (c) High energy absorption of the pc-composite film when used as a coating on a glass slide. (d) 157

Finite simulation of change in final height normalized by the sphere’s diameter between the pc-composite and pristine polymer spheres as a function of initial height normalized by the diameter (inset). (e) Deformation of the spheres upon impact on the metallic plate in the simulated gravity drop test (units: Pa). (f) Logarithmic strain of the pc-composite and pristine polymer spheres upon impact as measured at six different positions.

6.6. Conclusion

In summary, we have shown that designing multifunctional materials at the macro-scale can be achieved using simple polymer blends. We demonstrate a high stiffness and optical transmittance in a phase change (pc) composite by combining polymer and paraffin wax. We present a truly autonomous technique to control the optical transmittance of the pc-composite. This simple method eliminates the need for costly, sometimes toxic, and typically complex chromic materials with sophisticated circuitry to achieve the same effect. The simplicity and generality of the method makes it a viable option for accelerated application of these materials in windows as coatings.

158

Chapter 7

VII. Interface Engineering Applied to Hybrid Building Blocks

7.1. Introduction

The previous chapters have discussed interface engineering applied to synthetic materials in a ‘hard interface’ and later discussed ‘mobile interface’ in bio- mimic designs. This concluding chapter takes advantage of mobile interface in biological materials and hard interface in synthetic materials to design hybrid materials with improved and unique properties. The discussion is based on the papers; Materials Today Chemistry 9(2018) and advanced Mtrs Interfaces 2018.

7.1.1. Background to Hybrid Building Blocks

Bio-derived materials, which are designed using biological building blocks234,235 can solve some of the problems associated with synthetic materials. The 159

efficiency of materials design achieved by nature is still unmatched by their synthetic counterparts208,223,236. However, great strides have been achieved in mimicking natural designs to create multifunctional materials12,206,223,237,238. What makes natural designs so remarkable is their ability to utilize simple molecules, through the combination of hierarchy and geometry, to make very strong and tough materials146,174,224–226,239–241. As recently demonstrated by Kotov lab242, the parallel micro/nanoscale ceramics columns in a soft polymer of the tooth enamel is responsible for the extreme vibration resistance observed. The same design principles are exhibited by the well-studied nacre structure176,223,225,226,236,243. Efforts to mimic such elegant natural designs still rely heavily on the use of synthetic building blocks which themselves are still limited. For instance, the inorganic synthetic reinforcing ‘bricks’ in nacre needs surface modification to induce high affinity with the ductile polymer matrix11,208,225,240. On the other hand, natural building blocks are in abundance but difficult to use effectively. If judicious processing methods can be developed to take advantage of them, advanced environment-friendly materials with improved multi-functionality can be designed.

Protein building blocks are found in abundant202,218,244,245 are environmental friendly and can be explored for structural and bio-medical applications. Unlike other natural building blocks which require further processing or modification before use, proteins can be used in one step process. The polypeptide bonds83,246 which constitute the make-up of proteins are created by the reaction of amino acids where carboxylic and amine end groups react in a condensation reaction. 160

In this chapter, following similar approaches used to interconnect other nanomaterial like CNTs, Graphene Oxide, silicon dioxide etc., we interconnect or crosslink egg white (EW) solution with amine-based molecules to build 3D hybrid material with improved mechanical properties. Next a discussion of addition of graphene oxide in the egg white and finally effect of hydroxyapatite on the mechanical behaviour of cross-linked egg white is presented.

7.2. Preparation of Hybrid 3D Materials

7.2.1. Synthesis of polyalbumen (cross-linked egg white)

To prepare the material, amine-based crosslinkers, phenylethylamine, diethylenetriamine (DETA), urea etc. was added to egg white liquid extracted from chicken eggs. Specifically, (5%, 10%, 20%, 40% etc.) of crosslinker was added in egg white ((3ml) to initiate polymerization, mixed thoroughly, degassed in a vacuum oven and cured at 50°C for an hour.

7.2.2. Synthesis of polyalbumene (Graphene Oxide/Albumen)

To prepare the hybrid graphene oxide (GO) and cross-linked egg white material, GO is exfoliated and a selected amount of egg white and amine-based cross-linker; diethylenetriamine (DETA) (US Composites) added, the mixer is first put in magnetic stirrer for 24 hours later transferred to a bath sonicator for 20 minutes then added and the material cured at 50°C for an hour. 161

7.2.3. Synthesis of Albumin/Hydroxyapatite Composite

Hydroxyapatite (HA) nano-particles are extracted from chicken bone by a simple method. The bones are first washed to remove any impurities. Next, the bones are dried at 90C for 24hours. The bones are later immersed in a prepared 30 wt.% H2O2 to remove soft tissues and fats. Citric acid (20- 4 wt.0%) is then used to remove HA from bone leaving the collagen matrix intact. The powder is oven dried to remove any remnant solvent. The composite is made by first adding the polymerization initiator to the egg white liquid which are primary amine compounds such urea, phenylethylamine, diethylenetriamine (DETA) (Sigma

Aldrich, USA). The egg white with amine are added to the HA crystals and cured at

55˚C for an hour

7.3. Physical Characterization

Mechanical tests were done on a dynamical mechanical analysis DMA Q800

(TA Instrument, USA). Tests were conducted in a tensile and compression mode at ambient conditions. Controlled force and strain tests were done for compression test while tensile test was done by ramping the force at 0.5N/min to the limit of the

DMA (18N). The nanoscale stiffness mapping of both materials was conducted by

PeakForce technique on a Multimode 8 AFM (Bruker, USA) with contact mode.

Relative quantification method was employed to find the stiffness. The deflection sensitivity was calibrated by a ‘Ramp’ on a silicon sample surface, and the spring constant of each specific cantilever was calibrated by thermal tuning, with the 162

information of the standard stiffness of the substrate. It should be noted that the deflection sensitivity and spring constant need to be carefully performed before each characterization. The contact-mode cantilever (FESP) with a spring constant about 5N/m is used for both materials, (Bruker, USA).

7.4. Computation Details

Computational studies are carried out to understand the egg white and amine chemical affinity and the mechanical impact thereof. Pertaining to the exploration of the egg white and amine chemical affinity, Molecular Docking is a computational method that uses scoring functions to determine the preferred binding orientation of one molecule with respect to another to form a stable complex247. Such docking simulations are performed to understand the role of cross-linker in a complex egg white structure. The docking calculations elucidates on the underlying mechanism involved when the egg white cross-link with cross- linker. Egg white mainly consists of water (88%) and proteins248 (11%).

Ovalbumin constitutes over half the egg white protein by weight. The structure of

Ovalbumin was obtained from the Protein Data Bank with PDB code 1ova249, rendered using Chimera250. Figure 1a shows a bound configuration of cross-linker molecules with Ovalbumin. To determine the best possible binding sites for cross- linker, AutoLigand251 is used to locate the high affinity binding sites on the surface of the Ovalbumin protein where the cross-linker binder has the highest binding probability. The basic premise of AutoLigand is that the different elements of a ligand have different binding affinities and AutoLigand mimics this behavior by 163

generating adjoining fill volumes that generates maximum negative free energy change of binding per unit volume251. These high affinity binding pockets are then used as search spaces for docking calculations performed with AutoDockVina252.

Docking calculations were performed to identify the best binding configurations of cross-linker (flexible ligand) at each high-affinity binding pockets using

AutoDockVina, an open source program for molecular docking. The calculation of the binding affinity of the ligand in the high affinity binding sites are expected, specifically in the case of the highest affinity binding site, to produce a statistic of the amino-acid mostly responsible for the cross-linker /ovalbumin affinity, which was found to be Glutamic Acid.

Based on the previous insight, the first-principles simulations to test the mechanical quality of the interface between two glutamic acid radicals, a glutamic acid radical and a cross-linker monomer, are performed using classical molecular dynamics (MD) as numerically implemented in the large-scale atomic/molecular massively parallel simulator7(LAMMPS) using the Reax Force Field (ReaxFF) parameterizations for silicon carbide75, a timestep of 0.25 femtoseconds, and the

Nosé-Hoover thermostat at room temperature. In a manner similar to the fracture force simulations performed by Tsafack et al.253, investigating the mechanical effect of cross-linker’s insertions into ovalbumin proteins can be broken down to understanding the mixture’s interfacial bond strength and stability. The setups in I-

II are constructed under assumptions made in Figure 5.2(c-e) whereby two abutting glutamic acids can either vertically connect their amine and carboxyl groups 164

through peptide bonding (Figure 5.2(b)), or horizontally connect the carboxyl ends of their radicals through condensation (Figure 5.2(c)). Both processes involve the creation of a water molecule with H coming from the amine end of the top amino acid or the H from the hydroxyl end of one radical, reacting with OH coming from the carboxyl end of the bottom amino acid or the carboxyl end of the other radical, as the case may be. In addition to be a very common way biological cells link monomers together, the aforementioned process legitimates a comparable linking mechanism between cross-linker amine ends and glutamic acid radicals’ carboxyl ends (Figure 5.2(c)).

In configurations I and II from, two identical chains of five interconnected glutamic acids are linked through their radicals (configuration I) and through one amine monomer (configuration II). Each chain is held in place by two handles located at the carboxyl end of the top glutamic acid and at the amine end of the bottom glutamic acid leaving both the rest of the chain and the linking intermediary materials free to interact, equilibrate and move. A first equilibration process on the whole structure (NVT), except for the handles, takes place for 25 picoseconds. The subsequent equilibration process (NVT) eventuates with a series of consecutive 1 picosecond cycles in which the handles in the right chain move down by 0.05 Å while the rest of the structure reacts accordingly. The interfacial bond lengths, indicated in the respective configurations in Figure 5.2(c-e) as β, are monitored as the moving chain travels an increasing distance δ. The interfacial bond length is expected to remain reliably stable and to gradually increase before complete 165

rupture, see supplemental videos for visual rendering. The difference between the equilibrium bond length and the bond length at rupture, Y, qualitatively describes the stiffness of the interface whereas t = X/v, the time it takes to get to the end of the bond stretching process from the beginning thereof (X being the difference between

δ at the rupture and δ at the end of the equilibrium bond length, v being the speed of the moving chain) qualitatively describes the ductility of the interface. Interfacial parameters, Y and t, can hence give a qualitative insight into the mechanical behavior of the mixtures.

Testing the mechanical quality of the interface between glutamic acid and

DETA as well as DETA-functionalized glutamic acid and graphene oxide would provide a reliable insight into the effect of polyalbumene. This is performed using classical molecular dynamics (MD) in its numerical implementation within the large-scale atomic/molecular massively parallel simulator91 (LAMMPS) in conjunction with the Reax Force Field (ReaxFF) parameterizations for silicon carbide75, a time-step of 0.25 femto-seconds, and the Nosé-Hoover thermostat at room temperature. The assumptions behind the setups in Figure 2I-II are identical to those behind the setups in previous studies.

They involve the formation of peptide bonds regulating the interaction between a carboxyl and an amine ends belonging to either two adjacent glutamic acids or the R-group carboxyl end of glutamic acid and the amine end of DETA. It is further assumed that a reactive DETA amine end (-NH) connects to a carbon atom on the graphene oxide nanoribbon. The 24.875Å × 16.075Å graphene oxide 166

nanoribbon in Figure 2II contains isolated hydroxyl groups (C-OH) and bridging oxygens (C-O-C) randomly arranged on both sides of the surface. The reactive DETA amine end is connected to a carbon atom adjacent to a C-O-C oxygen bridge.

On the basis of the above assumptions, configurations I and II (Figure 2I-II) are created with two identical chains of five interconnected glutamic acids and their

R-groups linked through a single DETA monomer (configuration I) or one DETA monomer on each side connecting to a graphene oxide nanoribbon as described above. Two handles located at the carboxyl end of the top glutamic acid and at the amine end of the bottom glutamic acid hold each chain in place while leaving both the rest of the chain and the fillers free to interact, equilibrate and move. The outer carbon atoms on the graphene oxide nanoribbon are non-reactive in order to keep it steady while the rest of nanoribbon is free to interact with DETA.

The whole structures (without the handles) are first equilibrated (NVT) for

25 picoseconds. They are equilibrated again (NVT), this time with a series of consecutive1-picosecond cycles with the handles in the right chain moving down by

0.05Å as the rest of the structures react accordingly. As the right chain travels an increasing distance δ, the interfacial bond lengths for both configurations, β (C-N), are recorded. The interfacial bonds (C-N) are therefore expected to remain reliably stable and to gradually increase until they break, see supplemental videos for visual rendering.

As stated in previous studies studies,254 the difference between the average bond length before the bond stretching process and the bond length at the breaking 167

point (Y) qualitatively describes the stiffness of the interface whereas t = X/v, the duration of the bond stretching process qualitatively describes the ductility of the interface. X is the difference between δat the end of the stretching process and δ at the beginning of the stretching process, v is the speed of the moving chain.

Interfacial parameters, Y and t, can therefore provide a qualitative comparison of the mechanical behaviors of the mixtures.

7.5. High Strength and Tough Cross-linked Egg White Albumin

Albumin from egg white was extracted from chicken eggs and the

‘polymerization’ or cross-linking initiated by primary amine-based molecules like diethylenetriamine (DETA), phenylethylamine, urea etc. The discussion is mainly focused on the DETA cross-linker (Figure 5.1(a)). A desired amount of amine-based molecule was added to egg white albumin and cross-linked at 50°C for an hour resulting in a three-dimensional structure. A relatively low concentration of amine

(≤5%) tends to result in thin films while a higher concentration (˃5%) leads to the formation of foam-like structures. The crosslinking occurs mainly due to amine functional groups. A large scale (cm) size model of a ‘cup’ is fabricated as shown on

Figure 5.1(b). In understanding the morphological characteristics of material, samples were imaged under optical and high resolution scanning electron microscope (SEM). Pristine egg white shows very smooth surfaces with cracks due to the mechanical strains generated during the drying process (Figure 5.1(c)). On the other hand, cross-linked egg white exhibits a rough surface in contrast. The SEM images reveal rough surface and a sheet-like morphology (Figure 5.1(d)). Higher 168

amine concentration allows the sheet-like structures to transform to highly cross- link foam-like morphology possessing good thermal stability as seen from thermogravimetric analysis.

Indeed, through the docking of amine groups on the ovalbumin surface (see

Figure 5.2(a) for an artistic visualization), it is observed that among the five highest affinity binding sites (colored pockets laying on the gray ovalbumin surface in

Figure 5.2(b)), the affinity, measured in terms of binding energy in kcal/mol and visible on top of the colored bars indicating each site, is maximum in Site # 5

(binding energy 3.6 kcal/mol). In fact, even though both Alanine and Glutamic Acid are the most abundant amino acids in the five highest affinity binding sites, in the highest one (Site # 5), the amino-acid most abundant (~3 %) within 1.5 Å of amine monomers is Glutamic Acid. This result is to be expected on account of the glutamic acid R-group’s end (-COOH) being more reactive and more polar than Alanine’s saturated hydrocarbon R-group (-CH3).

Based on the outcome of the docking simulations and the likeliest Glutamic

Acid/amine interconnections through condensation (see computational details and

Figure 5.2(c-d)), the computer experiment with two five-Glutamic Acid chains connected through R-groups and cross-linker, with the right chain moving downward until the rupture of interconnecting bonds, can shed light on the mechanical effect of cross-linker on the egg white. Indeed, Figure 5.2(f) shows a sizeable bond stretching difference (YIII-YI≃ 0.25 Å) between configuration II and configuration I to the advantage of configuration II, 2) a very different stretching times between configurations I and II, namely, tI≃ 100 ps and tII≃ 252 ps. It can then be inferred that, in as much as configuration II provides more stiffness than configuration I and gets to rupture 152 ps later, the interfacial information of 170

configuration II is the main source of stiffness and ductility. Configuration I, because of its much smaller elongation and smaller time-to-rupture with respect to the same parameters in configurations II, does not significantly increase the stiffness nor the ductility of the composite. It can hence be concluded that ovalbumin by itself is likely to be brittle and not sufficiently stiff because it will predominantly be based on the interfacial behavior of configuration I. The addition of amine cross-linker to ovalbumin will predominantly be based on the interfacial behaviors of both configuration I and configuration II, and will therefore have a significant increase in ductility and stiffness. The very chemically favorable environment of ovalbumin thus appears to bring out the remarkable stiffening and toughening capabilities of cross-linker, thus confirming on one hand egg white’s weak crack resistance ascribed to lack of interconnections (Figure 5.2(a, f)), and on the other hand, high crack resistance ascribed to the presence of bridge-like connections in the cross- linked egg white (Figure 5.2(b, g)). 171

The mechanical properties of the material were assessed and compared with those of pristine dried egg white. The egg white is typical of brittle materials exhibiting high stiffness under cyclic load. Despite its initial high stiffness, egg white breaks up in less than 200 cycles (Figure 5.3(a)). Cross-linked albumin has low stiffness in initial loading cycles though the stiffness exhibits considerable increase and even surpasses that of pristine egg white at high loading cycles (˃3000, inset

Figure 5.3(a)). In addition, cross-linked albumin shows tendency of self-stiffening behavior which continues as the number of cycles increases. The strength was measured in tensile mode using dynamic mechanical analysis in the tension mode.

As shown in Figure 5.3(b), tensile strength is increased multifold by 5%, 10% addition of amine (inset Figure 5.3(b)). In conventional materials, achieving simultaneous increase in strength and toughness has remained elusive but here it is shown that amine groups can induce both strength and toughness concurrently

(Figure 5.3(c)). Frequency dependent tests show that the material behaves like a viscoelastic material where stiffness increases with frequency255. There is an order of magnitude increase in stiffness from samples tested at 1Hz and 10Hz (Figure

5.3(d)).

Cross-linked albumin shows a linear increase in stiffness with time (Figure

5.3(e)). The stiffness almost doubled in 24 hours and increased to more than 40-fold after aging for three days (inset Figure 5.3(e)). For the vast majority of synthetic materials, this is unexpected behavior but biological materials do indeed exhibit such traits of increasing strength with time, for example bones in mammals256. The 173

stiffness-aging behavior may be explained by the nature of the cross-linking that takes place in egg white. It is suggested that the majority of reactive groups on large- sized protein molecules is not initially exposed to amine groups all at once but progressively. Furthermore under ambient conditions cross-linked albumin loses its hydrogen bonds slowly. From a macroscopic level, SEM and modulus mapping using atomic force microscopy (AFM) can be crucial in understanding the materials’ failure mechanism. These two techniques are used here to show the brittle nature of pristine albumin (Figure 5.3(f)) as opposed to cross-linked albumin that shows bridge-like connections that prevent crack propagation (Figure 5.3(g)). To understand the crack resistance and high ductility in the materials, high resolution scanning electron microscope images were taken on the fractured samples. Egg white films developed cracks immediately after drying.

These cracks are typical of brittle materials as shown on Figure 5.3(f). Once the cracks initiate in egg white albumin, they propagate in all directions. Unlike egg white, cross-linked films in seems to prevent the catastrophic crack growth as exhibited by the ductile failure of the cross-linked films. The cross-linked multilayer films are important to prevent crack propagation and absorb much of the fracture energy generated in the material. To further explain cross-linked albumin’s high toughness, three-dimensional mapping of the modulus on AFM shows homogeneous distribution of high moduli ranging from 1 to 13GPa for egg white (Figure 5.3(h)).

Cross-linked albumin on the other hand, is characterized by a non-homogeneous distribution of the moduli which ranges from 0 to 9.8GPa (Figure 5.3(i). In between 174

these high moduli, smooth surfaces separating the moduli are observed. The distribution in modulus in the cross-linked sample connected by smooth or low modulus areas is responsible for the high toughness inherent in the cross-linked albumin where they act like glue to link or interconnect high-modulus regions.

Inhomogeneous high modulus distribution regions connected by low modulus areas which provide the elasticity observed in poly-albumen.

7.6. Hybrid Graphene Oxide/Albumin Layered Material

The graphene oxide/albumin composite (polyalbumene) was made by combining egg white, amine-based cross-linker diethylenetriamine (DETA) and graphene oxide (GO) as shown in the artistic representation inFigure 5.4(a).

Morphological analysis of the surface and the cross-section of the film exhibited a rough surface with wrinkle-like features as a result of layered cross-linking of the egg white liquid. The cross-section of the film shows clear layered structure similar nacre (Figure 5.4(b)). Such a layered structure can only be achieved by efficiency exfoliation of GO. The layered morphology is observed throughout the sample

(Figure 5.4(c)). We have recently shown that cross-linking of egg-white is a layer- by-layer process254. This mechanism is important by allowing the already exfoliated

GO to easily cross-link to the egg white. 176

materials, side view shows a layered-like appearance throughout the thickness of the material (c).

The computer model with two five-glutamic acid chains as well as R-groups of the middle glutamic acids connected through DETA (configuration I in Figure 5.5) and DETA/graphene oxide (configuration II in Figure 5.5). As the right chain moves downward until the rupture of interconnecting bonds, we can indeed indicate valuable mechanical differences between egg white/DETA and egg white/DETA/graphene oxide composites.

In fact, the plot in Figure 5.5displays a small bond stretching difference (YII-

YI≃ 0.07Å) between configuration II and configuration I to the advantage of configuration II. It also displays very different stretching times between configurations I and II, namely, tI≃ 252ps and tII≃ 120ps. It can then be inferred that, in as much as configuration II provides slightly more stiffness than configuration I and gets to the breaking point 122ps earlier, configuration II contributes to the stiffness of the composite slightly more than configuration I, but significantly less to the ductility of the composite than configuration I does. Thus it can be concluded that the addition of graphene oxide to DETA-functionalized egg white will predominantly be based on the interfacial behaviors of both configuration

I and configuration II, and will have a significant increase in stiffness as well as a moderate increase in ductility. The substantially chemically favorable environment of DETA-functionalized egg white thus appears to bring out the well-known stiffening nature of graphene oxide. 178

Stress-strain curves are typical of highly elastic materials like metals (Figure

7.6(a)). An increase in stress is accompanied by an increase in strain in the initial part of the curve. This increase is much higher for polyalbumene. The material shows a yielding stress region followed by the ultimate strength and finally, failure 179

occurring at ̴ 20% strain. The measured Young’s modulus from the linear part of the curve is 41MPa. A similar observation can be made on the ultimate tensile strength

(Figure 7.6(b)) with a value of 2.47MPa for polyalbumene. The stiffness measurement in a dynamic mechanical analysis (DMA) shows a stiffness of 2.5kN/m

(inset Figure 7.6(c)). The stiffness appears to increase with aging. The samples are left in ambient conditions and tested after seven, thirty days, etc. As shown in Figure

7.6(c), after seven days of aging, the stiffness jumped to 300kN/m and after thirty days, the stiffness is 940kN/m. Modulus mapping in the atomic force microscope

(AFM) was conducted to elucidate on the high mechanical properties of polyalbumene (Figure 7.6(e) and (f)). The material shows a distribution of high and low modulus regions which are important for the elasticity observed. The low modulus regions correspond to egg white while the high modulus regions are GO layers and the transitions regions in between those of polyalbumene. 180

7.7. Polymerized Egg White Hydroxyapatite Nacre-like Material

A simple one-step chemical process protocol is adopted to synthesize hydroxyapatatite (HA) from bones. The bones are first dried in an oven, dipped in hydrogen peroxide to remove any remaining soft tissue257 and placed in solution of citric acid. The citric acid opens the collagenous matrix to release the ‘brick’ like HA as shown in Figure 7.6(a). Unlike other methods of extracting HA bone, this method retains the collagenous matrix morphology with no degradation. Scanning electron microscopy (SEM) of the crystals shows large thin ‘bricks’ with enhanced surface area (Figure 7.6(b)). Transmission electron microscope (TEM) further prove the high surface area of the HA crystals (Figure 7.6(c) and (d)). The bricks are polycrystalline as shown by diffraction pattern from the TEM (Figure 7.6(e)). The x- ray diffraction of the bricks exhibits a typical pattern associated with polycrystalline materials. The peaks are observed at 32°, 40°, 47° and 51° with the highest peak at

32°. These peaks are similar to what is normally observed for HA. This method of producing HA in a single procedure in large quantity is important for wide applications of these “nano-bricks.” 182

Following the typical procedure to design nacre-like structure, we use liquid egg white ‘polymerized’ by primary amine compounds (diethylenetriamine (DETA) in this case) as a biopolymer to provide flexibility to the inorganic HA. Figure 7.8(a) 183

shows an artistic representation for the adopted methodology to make the material.

The diethylenetriamine is added in liquid egg white which is later added to the inorganic bricks and cross-linked/cured.

The polymerization of liquid egg white results in a very strong and tough material as shown in Figure 7.8(b). The material is inherently layered, similarly to nacre (Figure 7.8(c)). High magnification SEM images (Figure 7.8(d)) depict a material where the polymerized egg white has formed the junctions between the inorganic HA. Unlike other nacre-like synthetic materials that have to rely on the surface modification of the inorganic bricks, this work shows that bio-derived building blocks do not require such sophisticated methods to achieve structural stability. On the other hand, the inorganic bricks assembled in the absence of polymerized egg white does not possess the mechanical stability shown by the

HA/egg white as clearly exhibited by cracks development on the structure (Figure

7.8(e)). The polymerized egg white acts like an effective glue to the bricks resulting in the observed high mechanical stability.

184

Mechanical characterization is done on an Instron instrument to understand the addition of polymerized egg white in the inorganic bricks of HA. There is an increase of load with strain for HA/egg white material (Figure 7.9(a)). The curve is characterized by the zigzag form an indication of stepwise failure of individual HA crystals within the material. There is a huge increase in modulus as a result of polymerized egg white addition. Local deformation is important to understand the failure mechanism of nanocomposite. The nacre-like material deformation test was conducted in an in-situ nano-indentation test equipped with an SEM88,89. A load- displacement curve is shown in Figure 7.9(b). For consistency, the test is done under load control at the same force. For the nano-mechanical test, high modulus of

HA/ egg white is observed. The load-unload curve exhibits a completely recovery unlike the HA which has failed by this same load due to lack of ‘glue’ between the bricks. Modulus mapping on the atomic force microscope (AFM) shows a ranging modulus of 0.6 – 8.9 GPa (Figure 7.9(c)) for the pristine HA while the HA/egg white exhibits a modulus of 0.03-0.18TPa (Figure 7.9(d)). The addition of cross-linked egg white greatly improves on the modulus of the composite. Egg white by itself is an extremely brittle material with high stiffness, but once cross-linked by primary amine compounds, elasticity is induced. It is therefore believed that the high stiffness of egg white combine with the HA crystals to result in strong and tough material. In-situ indentation under SEM further shows a large hole left behind after loading on the pristine HA (Figure 7.9(e)) unlike HA/egg white which does not show any large scale observable failure (Figure 7.9(f)). With such desirable mechanical 186

properties and the easy of making, HA/egg white material can find application in biomedical field where biodegradable scaffolds are in high demand. To overcome some of the complexity designs encountered in such scaffolds, 3D printing can be an important tool to solve those issues.

AFM modulus mapping of HA with no polymer with a modulus of 0.6 to 8.9 GPa. (d) Extremely high modulus (30-180GPa) of reinforced HA. (e) In-situ SEM showing the failure of HA with no polymer unlike reinforced HA (f) which does not exhibit major failure.

The noticeable increase in toughness induced by the addition of Egg white

(EW) and diethylenetriamine (DETA) to calcium hydroxyapatite (HA) can be attributed to the nature of the chemical interactions between HA surface ions on one hand and EW amino acid terminals and water molecules on the other hand. Calcium hydroxyapatite, Ca10(PO4)6(OH)2, is a highly porous hexagonal crystal258 whose unit cell can be mapped into the illustrative oval in Figure 7.10. It is a network of six phosphate groups (PO43-) held together by coulomb interactions between the double-bonded oxygen in the phosphate group and one calcium cation (Ca++) located at the center of the structure (see dash-dot black lines in Figure 7.10).

Additional Ca++ions electrostatically connect two O-ions belonging to each phosphate group (see dotted black lines in Figure 7.10). Other Ca++ions connect the remaining O- ions belonging to two adjacent phosphate groups (see dash black lines in Figure 7.10). Hydroxyl ions OH- located in opposite sides of the circle add to the ionic nature of HA’s structure. This high porosity along with strong and localized electrostatic interactions hold PO43-, Ca++, and OH- ions together thereby giving HA not only its unique combination of lightweight and stiffness but also its brittleness.

Almora-Barrios at al.259 showed that HA surfaces show different patterns of PO43-,

Ca++, O-and OH- ions that are chemically favorable to interfacing with highly polar and/or ionic molecules. The water molecules (88%) and the amino acids248 (11%) 188

comprising egg white as well as DETA do indeed match the chemical environment at

HA surfaces in two fundamental respects:

(1) Hydrogen bonds form between water molecules in EW and OH-/O- ions at the

surface of HA. They also form between amine terminals (NH2/NH3+) in

normal/zwitterionic glutamic acid and OH-/O- ions at the surface of HA (see

dash blue lines in Figure 7.10). Additionally, hydrogen bonds form between

DETA’s amine ends (NH2) and OH-/O- ions at the surface of HA.

(2) Ionic bonds form between O/O- ends of normal/zwitterionic glutamic acid in

EW and Ca++ ions on the surface of HA (see dash-dot blue lines in Figure

7.10).

While ionic bonds at the HA/egg white interface contribute to the stiffness, hydrogen bonds, because of their reformable nature, tend to prevent crack propagation and increase the composite plastic regime, thus significantly increasing the composite’s toughness. The combination of hydrogen and ionic bonds at the interface between EW and HA is believed to be responsible for the remarkable increase in ductility without loss in stiffness.

189

7.8. Conclusion

In conclusion, this chapter has shown that primary amine-based molecules forms strong interfacial bonds with the egg white albumen resulting in a strong, tough and porous material as demonstrated and explained by experiments and molecular dynamic simulations, respectively. In addition to an adaptable self- stiffening, the new material shows an increase in stiffness with aging. Also, the new material is non-toxic. This work explored the design of bio-derived materials 190

utilizing readily available cross-linkers to polymerize biological proteins. As shown by experiments and theoretical simulations, this amine-based cross-linker tends to induce the reaction of amino groups on the cross-linker and carboxylic functional groups on the protein amino acids. Other amine-based materials like phenylethylamine and urea induce the same behavior. The non-toxic nature of the material can be utilized to design 3D scaffolds which can be investigated for their viability as biomedical application such as bone repair. Further addition of graphene oxide and hydroxyapatite into the egg white greatly improves the mechanical behavior. Though nature remains incomparable in its capability to design hybrid materials where contradicting properties like toughness and strength are optimized, this work opens the possibility that new designs routes can be used to design advanced materials with unprecedented properties.

191

Chapter 8

VIII. Summary

In summary, self-assembled and bio-mimic 3D hybrid materials have been carefully studied. Our results suggest that:

1. Surface chemistry method is developed to assemble nano-building blocks

into 3D structures in a bottom-up process. It has been shown that van der

Waals forces in combination with functionalization of CNTs can be

utilized to design nanocomposites that mimic stiffening behavior

normally observed in biological materials. Furthermore, silanization

chemistry is also used to create siloxane networks in SiO2 nanoparticles

resulting in 3D structures which show high stiffness.

2. Judicious modification of two dimensional materials can be utilized to

assemble 3D materials with improved mechanical properties. Interface

engineering between graphene oxide nano-sheets is important in 192

designing lightweight structures that maintain their structural integrity

without compromising on some of key physical properties. Same strategy

is used to design h-BN/polyvinyl alcohol foam with high CO2 absorption

capacity.

3. Solid/liquid composite has been developed exhibiting higher stiffness by

the addition of liquid inclusions. Based on classical composite theories,

this is an unprecedented phenomenon because they predict lower

stiffness with liquid inclusions. It is believed that liquid metals and other

liquid materials with high bulk modulus and high viscosity can be used to

design high energy absorption and high stiffness materials. These results

can form a basis for the development of new multifunctional materials

utilizing liquids inclusions spanning over very small to very large length

scales.

4. A multifunctional material at the macro-scale has been designed using

simple polymer blends which demonstrate a high stiffness and tunable

optical transmittance. A truly autonomous technique to control the

optical transmittance of the pc-composite is presented. This simple

method eliminates the need for costly, sometimes toxic, and typically

complex chromic materials with sophisticated circuitry to achieve the

same effect.

5. It has been shown that primary amine-based molecules forms strong

interfacial bonds with the egg white albumen resulting in a strong, tough 193

and porous material. In addition to an adaptable self-stiffening, the new material shows an increase in stiffness with aging. The non-toxic nature of the material can be utilized to design 3D scaffolds which can be investigated for their viability as biomedical application such as bone repair. Further addition of graphene oxide and hydroxyapatite into the egg white greatly improves the mechanical behavior.

194

Reference

(1) Nan, C. W.; Jia, Q. Obtaining Ultimate Functionalities in Nanocomposites:

Design, Control, and Fabrication. MRS Bull.2015, 40, 719–723.

(2) Mohammad, M.; Winey, K. I. Polymer Nanocomposites Containing Carbon

Nanotubes. Macromolecules2006, 39, 5194–5205.

(3) Zaman, I.; Kuan, H. C.; Meng, Q.; Michelmore, A.; Kawashima, N.; Pitt, T.; Zhang,

L.; Gouda, S.; Luong, L.; Ma, J. A Facile Approach to Chemically Modified

Graphene and Its Polymer Nanocomposites. Adv. Funct. Mater.2012, 22,

2735–2743.

(4) Mezzenga, R.; Ruokolainen, J. Nanocomposites: Nanoparticles in the Right

Place. Nat. Mater.2009, 8, 926–928.

(5) Hu, C.; Xue, J.; Dong, L.; Jiang, Y.; Wang, X.; Qu, L.; Dai, L. Scalable Preparation

of Multifunctional Fire-Retardant Ultralight Graphene Foams. ACS Nano2016,

10, 1325–1332.

(6) Wei, C.; Srivastava, D.; Cho, K. Structural Ordering in Nanotube Polymer

Composites. Nano Lett.2004, 4, 1949–1952.

(7) Chan, M.; Lau, K.; Wong, T.; Ho, M.; Hui, D. Mechanism of Reinforcement in a

Nanoclay/polymer Composite. Compos. Part B Eng.2011, 42, 1708–1712.

(8) Balazs, A. C.; Emrick, T.; Russell, T. P. Nanoparticle Polymer Composites : Meet 195

Two Small Worlds. Science (80-. ).2013, 314, 1107–1110.

(9) Coleman, J. N.; Khan, U.; Gun’ko, Y. K. Mechanical Reinforcement of Polymers

Using Carbon Nanotubes. Adv. Mater.2006, 18, 689–706.

(10) Veedu, V. P.; Cao, A.; Li, X.; Ma, K.; Soldano, C.; Kar, S.; Ajayan, P. M.; Ghasemi-

Nejhad, M. N. Multifunctional Composites Using Reinforced Laminae with

Carbon-Nanotube Forests. Nat. Mater.2006, 5, 457–462.

(11) Owuor, P. S.; Park, O.-K.; Woellner, C. F.; Jalilov, A. S.; Susarla, S.; Joyner, J.;

Ozden, S.; Duy, L.; Villegas Salvatierra, R.; Vajtai, R.; et al. Lightweight

Hexagonal Boron Nitride Foam for CO2 Absorption. ACS Nano2017.

(12) Ortiz, C.; Boyce, M. Bioinspired Structural Materials. Science (80-. ).2008, 319,

1053–1054.

(13) Kim, I. T.; Magasinski, A.; Jacob, K.; Yushin, G.; Tannenbaum, R. Synthesis and

Electrochemical Performance of Reduced Graphene Oxide/maghemite

Composite Anode for Lithium Ion Batteries. Carbon N. Y.2013, 52, 56–64.

(14) Rümmeli, M. H.; Rocha, C. G.; Ortmann, F.; Ibrahim, I.; Sevincli, H.; Börrnert, F.;

Kunstmann, J.; Bachmatiuk, A.; Pötschke, M.; Shiraishi, M.; et al. Graphene:

Piecing It Together. Adv. Mater.2011, 23, 4471–4490.

(15) Owuor, P. S.; Tiwary, C. S.; Koizumi, R.; Soto, M.; Hart, A. C.; Barrera, E. V.;

Vajtai, R.; Lou, J.; Ajayan, P. M. Self-Stiffening Behavior of Reinforced Carbon

Nanotubes Spheres. Adv. Eng. Mater.2017, 1–11. 196

(16) Bao, Z.; Katz, H. E.; Dodabalapur, A.; Kagan, C. R.; Andry, P.; Pomp, A.; Hart, C.

M.; Leeuw, D. M. De; Frisbie, C. D.; Marks, T. J.; et al. Ultrastrong and Stiff

Layered Polymer Nanocomposites. Science (80-. ).2007, 318, 80–84.

(17) Cheng, F.; Liang, J.; Tao, Z.; Chen, J. Functional Materials for Rechargeable

Batteries. Adv. Mater.2011, 23, 1695–1715.

(18) Ma, Y.; Chen, Y. Three-Dimensional Graphene Networks: Synthesis, Properties

and Applications. Natl. Sci. Rev.2015, 2, 40–53.

(19) Liu, K.; Jiang, L. Bio-Inspired Design of Multiscale Structures for Function

Integration. Nano Today2011, 6, 155–175.

(20) Yu, Y.; Chen, H.; Liu, Y.; Craig, V. S. J.; Wang, C.; Li, L. H.; Chen, Y.

Superhydrophobic and Superoleophilic Porous Boron Nitride

Nanosheet/Polyvinylidene Fluoride Composite Material for Oil-Polluted

Water Cleanup. Adv. Mater. Interfaces2015, 2, 1–10.

(21) Li, D.; Müller, M. B.; Gilje, S.; Kaner, R. B.; Wallace, G. G. Processable Aqueous

Dispersions of Graphene Nanosheets. Nat. Nanotechnol.2008, 3, 101–105.

(22) Cooper, A. J.; Wilson, N. R.; Kinloch, I. a.; Dryfe, R. a W. Single Stage

Electrochemical Exfoliation Method for the Production of Few-Layer

Graphene via Intercalation of Tetraalkylammonium Cations. Carbon N.

Y.2014, 66, 340–350.

(23) Gupta, A.; Sakthivel, T.; Seal, S. Recent Development in 2D Materials Beyond 197

Graphene. Prog. Mater. Sci.2015.

(24) Park, S.; Ruoff, R. S. Chemical Methods for the Production of Graphenes. Nat.

Nanotechnol.2009, 4, 217–224.

(25) Marcano, D. C.; Kosynkin, D. V.; Berlin, J. M.; Sinitskii, A.; Sun, Z.; Slesarev, A.;

Alemany, L. B.; Lu, W.; Tour, J. M. Improved Synthesis of Graphene Oxide. ACS

Nano2010, 4, 4806–4814.

(26) Li, W.; Wang, F.; Feng, S.; Wang, J.; Sun, Z.; Li, B.; Li, Y.; Yang, J.; Elzatahry, A. a;

Xia, Y.; et al. Sol − Gel Design Strategy for Ultradispersed TiO 2 Nanoparticles

on Graphene for High-Performance Lithium Ion Batteries. J. Am. Chem.

Soc2013, 18300–18303.

(27) Choi, D.; Kumta, P. N. Surfactant Based Sol-Gel Approach to Nanostructured

LiFePO4 for High Rate Li-Ion Batteries. J. Power Sources2007, 163, 1064–

1069.

(28) Li, W.; Wang, F.; Feng, S.; Wang, J.; Sun, Z.; Li, B.; Li, Y.; Yang, J.; Elzatahry, A. a.;

Xia, Y.; et al. Sol-Gel Design Strategy for Ultradispersed tio2 Nanoparticles on

Graphene for High-Performance Lithium Ion Batteries. J. Am. Chem. Soc.2013,

135, 18300–18303.

(29) William S. Hummers, J.; Offeman, R. E. Preparation of Graphitic Oxide. J. Am.

Chem. Soc1958, 80, 1339.

(30) Zhong, Y. L.; Tian, Z.; Simon, G. P.; Li, D. Scalable Production of Graphene via 198

Wet Chemistry: Progress and Challenges. Mater. Today2015, 18, 73–78.

(31) Dreyer, D. R.; Park, S.; Bielawski, C. W.; Ruoff, R. S. The Chemistry of Graphene

Oxide. Chem. Soc. Rev.2010, 39, 228–240.

(32) Sudeep, P. M.; Narayanan, T. N.; Ganesan, A.; Shaijumon, M. M.; Yang, H.;

Ozden, S.; Patra, P. K.; Pasquali, M.; Vajtai, R.; Ganguli, S.; et al. Covalently

Interconnected Three-Dimensional Graphene Oxide Solids. ACS Nano2013, 7,

7034–7040.

(33) Vaisman, L.; Marom, G.; Wagner, H. D. Dispersions of Surface-Modified Carbon

Nanotubes in Water-Soluble and Water-Insoluble Polymers. Adv. Funct.

Mater.2006, 16, 357–363.

(34) Behabtu, N.; Young, C. C.; Tsentalovich, D. E.; Kleinerman, O.; Wang, X.; Ma, A.

W. K.; Bengio, E. A.; ter Waarbeek, R. F.; de Jong, J. J.; Hoogerwerf, R. E.; et al.

Strong, Light, Multifunctional Fibers of Carbon Nanotubes with Ultrahigh

Conductivity. Science (80-. ).2013, 339, 182–186.

(35) Huang, Y. Y.; Terentjev, E. M. Tailoring the Electrical Properties of Carbon

Nanotube-Polymer Composites. Adv. Funct. Mater.2010, 20, 4062–4068.

(36) Ray H. Baughman, Anvar A. Zakhidor, de H. W. A. Carbon NanotubesOThe

Route Toward Applications. Science (80-. ).2002, 297, 787.

(37) Hashim, D. P.; Narayanan, N. T.; Romo-Herrera, J. M.; Cullen, D. A.; Hahm, M. G.;

Lezzi, P.; Suttle, J. R.; Kelkhoff, D.; Muñoz-Sandoval, E.; Ganguli, S.; et 199

al.Covalently Bonded Three-Dimensional Carbon Nanotube Solids via Boron

Induced Nanojunctions. Sci. Rep.2012, 2, 1–8.

(38) Meaud, J.; Sain, T.; Yeom, B.; Park, S. J.; Shoultz, A. B.; Hulbert, G.; Ma, Z.; Kotov,

N. A.; Hart, A. J.; Arruda, E. M.; et al. Simultaneously High Sti Ff Ness and

Damping in Nanoengineered Microtruss Composites. ACS Nano2014, 8, 3468–

3475.

(39) Carey, B. J.; Patra, P. K.; Hahm, M. G.; Ajayan, P. M. Foam-like Behavior in

Compliant, Continuously Reinforced Nanocomposites. Adv. Funct. Mater.2013,

23, 3002–3007.

(40) Dai, L.; Chang, D. W.; Baek, J.-B.; Lu, W. Carbon Nanomaterials for Advanced

Energy Conversion and Storage. Small2012, 8, 1130–1166.

(41) Premkumar, T.; Mezzenga, R.; Geckeler, K. E. Carbon Nanotubes in the Liquid

Phase: Addressing the Issue of Dispersion. Small2012, 8, 1299–1313.

(42) Venturelli, E.; Fabbro, C.; Chaloin, O.; Menard-Moyon, C.; Smulski, C. R.; Da Ros,

T.; Kostarelos, K.; Prato, M.; Bianco, A. Antibody Covalent Immobilization on

Carbon Nanotubes and Assessment of Antigen Binding. Small2011, 7, 2179–

2187.

(43) Gao, L.; Chou, T.-W.; Thostenson, E. T.; Godara, A.; Zhang, Z.; Mezzo, L. Highly

Conductive Polymer Composites Based on Controlled Agglomeration of

Carbon Nanotubes. Carbon N. Y.2010, 48, 2649–2651. 200

(44) Zeng, Y.; Ci, L.; Carey, B.; Vajtai, R.; Ajayan, P. Design and Reinforcement:

Vertically Aligned Carbon Nanotube-Based Sandwich Composites. ACS

Nano2010, 4, 6798–6804.

(45) Zhang, X.; Fan, X.; Yan, C.; Li, H.; Zhu, Y.; Li, X.; Yu, L. Interfacial Microstructure

and Properties of Carbon Fiber Composites Modified with Graphene Oxide.

ACS Appl. Mater. Interfaces2012, 4, 1543–1552.

(46) Shi, D.-L.; Feng, X.-Q.; Huang, Y. Y.; Hwang, K.-C.; Gao, H. The Effect of

Nanotube Waviness and Agglomeration on the Elastic Property of Carbon

Nanotube-Reinforced Composites. J. Eng. Mater. Technol.2004, 126, 250.

(47) Cao, A.; Dickrell, P. L.; Sawyer, W. G.; Ghasemi-Nejhad, M. N.; Ajayan, P. M.

Super-Compressible Foamlike Carbon Nanotube Films. Science (80-. ).2005,

310, 1307–1310.

(48) Carey, B. J.; Patra, P. K.; Ci, L.; Silva, G. G.; Ajayan, P. M. Observation of Dynamic

Strain Hardening in Polymer Nanocomposites. ACS Nano2011, 5, 2715–2722.

(49) Ozden, S.; Tiwary, C. S.; Hart, A. H. C.; Chipara, A. C.; Romero-Aburto, R.;

Rodrigues, M.-T. F.; Taha-Tijerina, J.; Vajtai, R.; Ajayan, P. M. Density Variant

Carbon Nanotube Interconnected Solids. Adv. Mater.2015, 11, 1842–1850.

(50) Abaqus, 2016.

(51) Wang, Z. Polydimethylsiloxane Mechanical Properties Measured by

Macroscopic Compression and Nanoindentation Techniques, University of 201

South Florida, 2011.

(52) AISI Type 302 Stainless Steel, cold rolled to 703 MPa tensile strength

http://asm.matweb.com/search/SpecificMaterial.asp?bassnum=MQ302AD

(accessed Dec 4, 2016).

(53) Coleman, J. N.; Cadek, M.; Ryan, K. P.; Fonseca, A.; Nagy, J. B.; Blau, W. J.;

Ferreira, M. S. Reinforcement of Polymers with Carbon Nanotubes. The Role of

an Ordered Polymer Interfacial Region. Experiment and Modeling. Polymer

(Guildf).2006, 47, 8556–8561.

(54) Liao, W. H.; Tien, H. W.; Hsiao, S. T.; Li, S. M.; Wang, Y. S.; Huang, Y. L.; Yang, S.

Y.; Ma, C. C. M.; Wu, Y. F. Effects of Multiwalled Carbon Nanotubes

Functionalization on the Morphology and Mechanical and Thermal Properties

of Carbon Fiber/vinyl Ester Composites. ACS Appl. Mater. Interfaces2013, 5,

3975–3982.

(55) Agrawal, R.; Nieto, A.; Chen, H.; Mora, M.; Agarwal, A. Nanoscale Damping

Characteristics of Boron Nitride Nanotubes and Carbon Nanotubes Reinforced

Polymer Composites. ACS Appl. Mater. Interfaces2013, 5, 12052–12057.

(56) Zhang, X.; Luster, B.; Church, A.; Muratore, C.; Voevodin, A. A.; Kohli, P.;

Aouadi, S.; Talapatra, S. Carbon Nanotube-MoS2 Composites as Solid

Lubricants. ACS Appl. Mater. Interfaces2009, 1, 735–739.

(57) Zhou, K.; Liu, J.; Shi, Y.; Jiang, S.; Wang, D.; Hu, Y.; Gui, Z. MoS2 Nanolayers 202

Grown on Carbon Nanotubes: An Advanced Reinforcement for Epoxy

Composites. ACS Appl. Mater. Interfaces2015, 7, 6070–6081.

(58) Ozden, S.; Narayanan, T. N.; Tiwary, C. S.; Dong, P.; Hart, A. H. C.; Vajtai, R.;

Ajayan, P. M. 3D Macroporous Solids from Chemically Cross-Linked Carbon

Nanotubes. Small2015, 11, 688–693.

(59) Koizumi, R.; Hart, A. H. C.; Brunetto, G.; Bhowmick, S.; Owuor, P. S.; Hamel, J. T.;

Gentles, A. X.; Ozden, S.; Lou, J.; Vajtai, R.; et al. Mechano-Chemical

Stabilization of Three-Dimensional Carbon Nanotube Aggregates. Carbon N.

Y.2016, 110, 27–33.

(60) Chakravarty, D.; Tiwary, C. S.; Woellner, C. F.; Radhakrishnan, S.; Vinod, S.;

Ozden, S.; da Silva Autreto, P. A.; Bhowmick, S.; Asif, S.; Mani, S. A.; et al. 3D

Porous Graphene by Low-Temperature Plasma Welding for Bone Implants.

Adv. Mater.2016, 1–9.

(61) Vinod, S.; Tiwary, C. S.; Machado, L. D.; Ozden, S.; Vajtai, R.; Galvao, D. S.;

Ajayan, P. M. Synthesis of Ultralow Density 3D graphene–CNT Foams Using a

Two-Step Method. Nanoscale2016, 8, 15857–15863.

(62) Vinod, S.; Tiwary, C. S.; da Silva Autreto, P. A.; Taha-Tijerina, J.; Ozden, S.;

Chipara, A. C.; Vajtai, R.; Galvao, D. S.; Narayanan, T. N.; Ajayan, P. M. Low-

Density Three-Dimensional Foam Using Self-Reinforced Hybrid Two-

Dimensional Atomic Layers. Nat. Commun.2014, 5, 4541. 203

(63) Sharifi, T.; Gracia-Espino, E.; Reza Barzegar, H.; Jia, X.; Nitze, F.; Hu, G.;

Nordblad, P.; Tai, C. W.; Wagberg, T. Formation of Nitrogen-Doped Graphene

Nanoscrolls by Adsorption of Magnetic Gamma-Fe2O3 Nanoparticles. Nat

Commun2013, 4, 2319.

(64) Chakravarty, D.; Tiwary, C. S.; Machado, L. D.; Brunetto, G.; Vinod, S.; Yadav, R.

M.; Galvao, D. S.; Joshi, S. V.; Sundararajan, G.; Ajayan, P. M. Zirconia-

Nanoparticle-Reinforced Morphology-Engineered Graphene-Based Foams.

Adv. Mater.2015, 27, 1–10.

(65) Kango, S.; Kalia, S.; Celli, A.; Njuguna, J.; Habibi, Y.; Kumar, R. Surface

Modification of Inorganic Nanoparticles for Development of Organic-

Inorganic Nanocomposites - A Review. Prog. Polym. Sci.2013, 38, 1232–1261.

(66) Walther, A.; Bjurhager, I.; Malho, J. M.; Ruokolainen, J.; Berglund, L.; Ikkala, O.

Supramolecular Control of Stiffness and Strength in Lightweight High-

Performance Nacre-Mimetic Paper with Fire-Shielding Properties. Angew.

Chemie - Int. Ed.2010, 49, 6448–6453.

(67) Wu, H. C.; Sun, P. New Building Materials from Fly Ash-Based Lightweight

Inorganic Polymer. Constr. Build. Mater.2007, 21, 211–217.

(68) Yu, T.; Malugin, A.; Ghandehari, H. Impact of Silica Nanoparticle Design on

Cellular Toxicity and Hemolytic Activity. ACS Nano2011, 5, 5717–5728.

(69) Li, L.; Tang, F.; Liu, H.; Liu, T.; Hao, N.; Chen, D.; Teng, X.; He, J. In Vivo Delivery 204

of Silica Nanorattle Encapsulated Docetaxel for Liver Cancer Therapy with

Low Toxicity and High Efficacy. ACS Nano2010, 4, 6874–6882.

(70) Ambrosone, A.; Scotto di Vettimo, M. R.; Malvindi, M. A.; Roopin, M.; Levy, O.;

Marchesano, V.; Pompa, P. P.; Tortiglione, C.; Tino, A. Impact of Amorphous

SiO2 Nanoparticles on a Living Organism: Morphological, Behavioral, and

Molecular Biology Implications. Front. Bioeng. Biotechnol.2014, 2, 37.

(71) Chrissafis, K.; Bikiaris, D. Can Nanoparticles Really Enhance Thermal Stability

of Polymers? Part I: An Overview on Thermal Decomposition of Addition

Polymers. Thermochim. Acta2011, 523, 1–24.

(72) Sengupta, R.; Chakraborty, S.; Bandyopadhyay, S.; Dasgupta, S.;

Mukhopadhyay, R.; Auddy, K.; Deuri, a S. A Short Review on Rubber / Clay

Nanocomposites With Emphasis on Mechanical Properties. Polym. Eng.

Sci.2008, 47, 21–25.

(73) Liong, M.; Lu, J.; Kovochich, M.; Xia, T.; Ruehm, S. G.; Nel, A. E.; Tamanoi, F.;

Zink, J. I. Multifunctional Inorganic Nanoparticles for Imaging, Targeting, and

Drug Delivery. ACS Nano2008, 2, 889–896.

(74) Witucki, G. . A Silane Primer: Chemistry andApplications of AIkoxy Silanes. J.

Coatings Technol.1993, 65, 57–60.

(75) Newsome, D. A.; Sengupta, D.; Foroutan, H.; Russo, M. F.; van Duin, A. C. T.

Oxidation of Silicon Carbide by O 2 and H 2 O: A ReaxFF Reactive Molecular 205

Dynamics Study, Part I. J. Phys. Chem. C2012, 116, 16111–16121.

(76) Wu, Z., Chen, Z., Du, X., Logan, J.M., Sippel, J., Nikolou, M., Kamaras, K.,

Reynolds, J.R., Tanner, D.B., Hebard, A.F. and Rinzler, A. . Transparent,

Conductive Carbon Nanotube Film. Science (80-. ).2010, 305, 1273–1276.

(77) Vinod, S.; Tiwary, C. S.; Machado, L. D.; Ozden, S.; Cho, J.; Shaw, P.; Vajtai, R.;

Galvão, D. S.; Ajayan, P. M. Strain Rate Dependent Shear Plasticity in Graphite

Oxide. Nano Lett.2016, 16, 1127–1131.

(78) Eda, G.; Fanchini, G.; Chhowalla, M. Large-Area Ultrathin Films of Reduced

Graphene Oxide as a Transparent and Flexible Electronic Material. Nat.

Nanotechnol.2008, 3, 270–274.

(79) Hong, C. Q.; Han, J. C.; Zhang, X. H.; Du, J. C. Novel Nanoporous Silica Aerogel

Impregnated Highly Porous Ceramics with Low Thermal Conductivity and

Enhanced Mechanical Properties. Scr. Mater.2013, 68, 599–602.

(80) Hongmei, J.; Zhen, Z.; Zhiming, L.; Xinling, W. Effects of Temperature and

Solvent on the Hydrolysis of Alkoxysilane under Alkaline Conditions. Ind. Eng.

Chem. Res.2006, 45, 8617–8622.

(81) Werne, T. Von; Patten, T. E. Preparation of Structurally Well-Defined Polymer

- Nanoparticle Hybrids with Controlled / Living Radical Polymerizations. J.

Am. Chem. Soc.1999, 7409–7410.

(82) Kim, Y. H.; Lee, D. K.; Cha, H. G.; Kim, C. W.; Kang, Y. C.; Kang, Y. S. Preparation 206

and Characterization of the Antibacterial Cu Nanoparticle Formed on the

Surface of SiO2 Nanoparticles. J. Phys. Chem. B2006, 110, 24923–24928.

(83) Marucco, A.; Turci, F.; O’Neill, L.; Byrne, H. J.; Fubini, B.; Fenoglio, I. Hydroxyl

Density Affects the Interaction of Fibrinogen with Silica Nanoparticles at

Physiological Concentration. J. Colloid Interface Sci.2014, 419, 86–94.

(84) Daniel R. Dreyer, Sungjin Park, C. W. B.; Ruoff, R. S. The Chemistry of Graphene

Oxide. Chem. Soc. Rev.2010, 228–240.

(85) Brown, E.; Nasto, A.; Athanassiadis, A. G.; Jaeger, H. M. Strain Stiffening in

Random Packings of Entangled Granular Chains. Phys. Rev. Lett.2012, 108, 9–

12.

(86) Stankovich, S.; Dikin, D. A.; Piner, R. D.; Kohlhaas, K. A.; Kleinhammes, A.; Jia,

Y.; Wu, Y.; Nguyen, S. T.; Ruoff, R. S. Synthesis of Graphene-Based Nanosheets

via Chemical Reduction of Exfoliated Graphite Oxide. Carbon N. Y.2007, 45,

1558–1565.

(87) Xie, Y.; Hill, C. A. S.; Xiao, Z.; Militz, H.; Mai, C. Silane Coupling Agents Used for

Natural Fiber/polymer Composites: A Review. Compos. Part A Appl. Sci.

Manuf.2010, 41, 806–819.

(88) Fischer-Cripps, A. C. Critical Review of Analysis and Interpretation of

Nanoindentation Test Data. Surf. Coatings Technol.2006, 200, 4153–4165.

(89) Schuh, C. A. Nanoindentation Studies of Materials. Mater. Today2006, 9, 32– 207

40.

(90) Stukowski, A. Visualization and Analysis of Atomistic Simulation Data with

OVITO–the Open Visualization Tool. Model. Simul. Mater. Sci. Eng.2009, 18,

15012.

(91) Plimpton, S. Fast Parallel Algorithms for Short-Range Molecular-Dynamics. J.

Comput. Phys.1995, 117, 1–19.

(92) Wang, L.; Sun, Y. Y.; Lee, K.; West, D.; Chen, Z. F.; Zhao, J. J.; Zhang, S. B. Stability

of Graphene Oxide Phases from First-Principles Calculations. Phys. Rev. B -

Condens. Matter Mater. Phys.2010, 82, 2–5.

(93) K.S. Novoselov1, A.K. Geim1, S.V. Morozov2, D. Jiang1, Y. Zhang1, S.V.

Dubonos2, I.V.Grigorieva1, a. a. F. Electric Field Effect in Atomically Thin

Carbon Films. Science (80-. ).2004, 306, 666–669.

(94) Simon, P.; Gogotsi, Y. Materials for Electrochemical Capacitors. Nat.

Mater.2008, 7, 845–854.

(95) Lin, T.; Chen, I.-W.; Liu, F.; Yang, C.; Bi, H.; Xu, F.; Huang, F. Nitrogen-Doped

Mesoporous Carbon of Extraordinary Capacitance for Electrochemical Energy

Storage. Science (80-. ).2015, 350, 1508–1513.

(96) Kim, K. H.; Oh, Y.; Islam, M. F. Graphene Coating Makes Carbon Nanotube

Aerogels Superelastic and Resistant to Fatigue. Nat Nanotechnol2012, 7, 562–

566. 208

(97) Niu, Z.; Chen, J.; Hng, H. H.; Ma, J.; Chen, X. A Leavening Strategy to Prepare

Reduced Graphene Oxide Foams. Adv. Mater.2012, 24, 4144–4150.

(98) Zhao, Y.; Liu, J.; Hu, Y.; Cheng, H.; Hu, C.; Jiang, C.; Jiang, L.; Cao, A.; Qu, L. Highly

Compression-Tolerant Supercapacitor Based on Polypyrrole-Mediated

Graphene Foam Electrodes. Adv. Mater.2013, 25, 591–595.

(99) De Volder, M. F. L.; Tawfick, S. H.; Baughman, R. H.; Hart, A. J. Carbon

Nanotubes: Present and Future Commercial Applications. Science (80-. ).2013,

339, 535–539.

(100) Chen, Z.; Ren, W.; Gao, L.; Liu, B.; Pei, S.; Cheng, H.-M. Three-Dimensional

Flexible and Conductive Interconnected Graphene Networks Grown by

Chemical Vapour Deposition. Nat. Mater.2011, 10, 424–428.

(101) Li, C.; Qiu, L.; Zhang, B.; Li, D.; Liu, C. Y. Robust Vacuum-/Air-Dried Graphene

Aerogels and Fast Recoverable Shape-Memory Hybrid Foams. Adv.

Mater.2016, 28, 1510–1516.

(102) Wen, L.; Li, F.; Cheng, H.-M. Carbon Nanotubes and Graphene for Flexible

Electrochemical Energy Storage: From Materials to Devices. Adv. Mater.2016,

28, 4306–4337.

(103) Jakus, A. E.; Secor, E. B.; Rutz, A. L.; Jordan, S. W.; Hersam, M. C. Three-

Dimensional Printing of High-Content Graphene Sca Ff Olds for Electronic and

Biomedical Applications. ACS Nano2015, 9, 4636–4648. 209

(104) Wang, H.; Wang, G.; Ling, Y.; Qian, F.; Song, Y.; Lu, X.; Chen, S.; Tong, Y.; Li, Y.

High Power Density Microbial Fuel Cell with Flexible 3D Graphene-Nickel

Foam as Anode. Nanoscale2013, 5, 10283–10290.

(105) Gu, L.; Qian, L.; Lei, Y.; Wang, Y.; Li, J.; Yuan, H.; Xiao, D. Microwave-Assisted

Synthesis of Nanosphere-like NiCo2O 4 Consisting of Porous Nanosheets and

Its Application in Electro-Catalytic Oxidation of Methanol. J. Power

Sources2014, 261, 317–323.

(106) Gong, K.; Du, F.; Xia, Z.; Durstock, M.; Dai, L. Nitrogen-Doped Carbon Nanotube

Arrays with High Electrocatalytic Activity for Oxygen Reduction. Science (80-.

).2009, 323, 760–764.

(107) Pang, C.; Lee, G.-Y.; Kim, T.; Kim, S. M.; Kim, H. N.; Ahn, S.-H.; Suh, K.-Y. A

Flexible and Highly Sensitive Strain-Gauge Sensor Using Reversible

Interlocking of Nanofibres. Nat. Mater.2012, 11, 795–801.

(108) Rim, Y. S.; Bae, S. H.; Chen, H.; De Marco, N.; Yang, Y. Recent Progress in

Materials and Devices toward Printable and Flexible Sensors. Adv.

Mater.2016, 28, 4415–4440.

(109) Trung, T. Q.; Lee, N. E. Flexible and Stretchable Physical Sensor Integrated

Platforms for Wearable Human-Activity Monitoringand Personal Healthcare.

Adv. Mater.2016, 28, 4338–4372.

(110) Liu, K.; Chen, Y.; Policastro, G. M.; Becker, M. L.; Zhu, Y. Three-Dimensional 210

Bicontinuous Graphene Monolith from Polymer Templates. ACS Nano2015, 9,

6041–6049.

(111) Zhu, Y.; Murali, S.; Stoller, M. D.; Ganesh, K. J.; Cai, W.; Ferreira, P. J.; Pirkle, A.;

Wallace, R. M.; Cychosz, K. A.; Thommes, M.; et al. Carbon-Based

Supercapacitors Produced by Activation of Graphene. Science (80-. ).2011,

332, 1537–1541.

(112) Ji, L.; Lin, Z.; Alcoutlabi, M.; Zhang, X. Recent Developments in Nanostructured

Anode Materials for Rechargeable Lithium-Ion Batteries. Energy Environ.

Sci.2011, 4, 2682.

(113) Hu, G.; Xu, C.; Sun, Z.; Wang, S.; Cheng, H. M.; Li, F.; Ren, W. 3D Graphene-

Foam-Reduced-Graphene-Oxide Hybrid Nested Hierarchical Networks for

High-Performance Li-S Batteries. Adv. Mater.2015, 1–7.

(114) Wang, Z. L.; Xu, D.; Xu, J. J.; Zhang, L. L.; Zhang, X. B. Graphene Oxide Gel-

Derived, Free-Standing, Hierarchically Porous Carbon for High-Capacity and

High-Rate Rechargeable Li-O 2 Batteries. Adv. Funct. Mater.2012, 22, 3699–

3705.

(115) Kholmanov, I.; Kim, J.; Ou, E.; Ruoff, R. S.; Shi, L. Continuous Carbon Nanotube-

Ultrathin Graphite Hybrid Foams for Increased Thermal Conductivity and

Suppressed Subcooling in Composite Phase Change Materials. ACS Nano2015,

9, 11699–11707. 211

(116) Wu, C.; Huang, X.; Wu, X.; Qian, R.; Jiang, P. Mechanically Flexible and

Multifunctional Polymer-Based Graphene Foams for Elastic Conductors and

Oil-Water Separators. Adv. Mater.2013, 25, 5658–5662.

(117) Nieto, A.; Boesl, B.; Agarwal, A. Multi-Scale Intrinsic Deformation Mechanisms

of 3D Graphene Foam. Carbon N. Y.2015, 85, 299–308.

(118) Georgakilas, V.; Tiwari, J. N.; Kemp, K. C.; Perman, J. A.; Bourlinos, A. B.; Kim, K.

S.; Zboril, R. Noncovalent Functionalization of Graphene and Graphene Oxide

for Energy Materials, Biosensing, Catalytic, and Biomedical Applications.

Chem. Rev.2016, 116, acs.chemrev.5b00620.

(119) Zhu, Y.; Murali, S.; Cai, W.; Li, X.; Suk, J. W.; Potts, J. R.; Ruoff, R. S. Graphene and

Graphene Oxide: Synthesis, Properties, and Applications. Adv. Mater.2010, 22,

3906–3924.

(120) Cao, L.; Wang, Y.; Dong, P.; Vinod, S.; Tijerina, J. T.; Ajayan, P. M.; Xu, Z.; Lou, J.

Interphase Induced Dynamic Self-Stiffening in Graphene-Based

Polydimethylsiloxane Nanocomposites. Small2016, 3723–3731.

(121) Stan, G.; King, S. W.; Cook, R. F. Nanoscale Mapping of Contact Stiffness and

Damping by Contact Resonance Atomic Force Microscopy.

Nanotechnology2012, 23, 215703.

(122) Qichi Hu, Kjoller Kevin, Myers Alan, Singh J. Kanwal, K. W. S. Nanoscale

Chemical Structure Variations in Nano-Patterned and Nano-Porous Low-K 212

Dielectrics: A Comparative Photothermal Induced Resonance and Infrared

Spectroscopy Investigation. Vib. Spectrosc.2016, 86, 223–232.

(123) Chenoweth, K.; Cheung, S.; Van Duin, A. C. T.; Goddard, W. A.; Kober, E. M.

Simulations on the Thermal Decomposition of a Poly(dimethylsiloxane)

Polymer Using the ReaxFF Reactive Force Field. J. Am. Chem. Soc.2005, 127,

7192–7202.

(124) Van Duin, A. C. T.; Dasgupta, S.; Lorant, F.; Goddard, W. A. ReaxFF: A Reactive

Force Field for Hydrocarbons. J. Phys. Chem. A2001, 105, 9396–9409.

(125) Dikin, D. A.; Stankovich, S.; Zimney, E. J.; Piner, R. D.; Dommett, G. H. B.;

Evmenenko, G.; Nguyen, S. T.; Ruoff, R. S. Preparation and Characterization of

Graphene Oxide Paper. Nature2007, 448, 457–460.

(126) Lin, Y.; Williams, T. V.; Connell, J. W. Soluble, Exfoliated Hexagonal Boron

Nitride Nanosheets. J. Phys. Chem. Lett.2010, 1, 277–283.

(127) Lee, K. H.; Shin, H.; Lee, J.; Lee, I.; Kim, G.; Choi, J.; Kim, S. Large-Scale Synthesis

of High-Quality Hexagonal Boron Nitride Nanosheets for Large-Area

Graphene Electronics Large-Scale Synthesis of High-Quality Hexagonal Boron

Nitride Nanosheets for Large-Area Graphene Electronics. Nano Lett.2012, 12,

714–718.

(128) Chen, Z.; Xu, C.; Ma, C.; Ren, W.; Cheng, H. M. Lightweight and Flexible

Graphene Foam Composites for High-Performance Electromagnetic 213

Interference Shielding. Adv. Mater.2013, 25, 1296–1300.

(129) Walther, A.; Bjurhager, I.; Malho, J. M.; Pere, J.; Ruokolainen, J.; Berglund, L. A.;

Ikkala, O. Large-Area, Lightweight and Thick Biomimetic Composites with

Superior Material Properties via Fast, Economic, and Green Pathways. Nano

Lett.2010, 10, 2742–2748.

(130) Zhou, J.; Huang, Y.; Cao, X.; Ouyang, B.; Sun, W.; Tan, C.; Zhang, Y.; Ma, Q.; Liang,

S.; Yan, Q.; et al. Two-Dimensional NiCo2O4 Nanosheet-Coated Three-

Dimensional Graphene Networks for High-Rate, Long-Cycle-Life

Supercapacitors. Nanoscale2015, 7, 7035–7039.

(131) Zhou, H.; Yu, F.; Huang, Y.; Sun, J.; Zhu, Z.; Nielsen, R. J.; He, R.; Bao, J.; Goddard

III, W. A.; Chen, S.; et al. Efficient Hydrogen Evolution by Ternary Molybdenum

Sulfoselenide Particles on Self-Standing Porous Nickel Diselenide Foam. Nat.

Commun.2016, 7, 12765.

(132) Cai, S.; Zhang, D.; Shi, L.; Xu, J.; Zhang, L.; Huang, L.; Li, H.; Zhang, J. Porous Ni-

Mn Oxide Nanosheets in Situ Formed on Nickel Foam as 3D Hierarchical

Monolith de-NO(x) Catalysts. Nanoscale2014, 6, 7346–7353.

(133) Liu, Z.; Nie, H.; Yang, Z.; Zhang, J.; Jin, Z.; Lu, Y.; Xiao, Z.; Huang, S. Sulfur–

nitrogen Co-Doped Three-Dimensional Carbon Foams with Hierarchical Pore

Structures as Efficient Metal-Free Electrocatalysts for Oxygen Reduction

Reactions. Nanoscale2013, 5, 3283. 214

(134) Rhodes, C. P.; Mcevoy, T. M.; Bourg, E.; Lubers, A. M. Multifunctional 3D

Nanoarchitectures for Energy Storage and Conversion. Chem. Soc. Rev.2009,

38, 226–252.

(135) Leonard, A. D.; Hudson, J. L.; Fan, H.; Booker, R.; Simpson, L. J.; O’Neill, K. J.;

Parilla, P. a; Heben, M. J.; Pasquali, M.; Kittrell, C.; et al. Nanoengineered

Carbon Scaffolds for Hydrogen Storage. J. Am. Chem. Soc.2009, 131, 723–728.

(136) Camilli, L.; Pisani, C.; Gautron, E.; Scarselli, M.; Castrucci, P.; D’Orazio, F.;

Passacantando, M.; Moscone, D.; De Crescenzi, M. A Three-Dimensional

Carbon Nanotube Network for Water Treatment. Nanotechnology2014, 25,

65701.

(137) Terrones, M.; Banhart, F.; Grobert, N.; Charlier, J.-C.; Terrones, H.; Ajayan, P. M.

Molecular Junctions by Joining Single-Walled Carbon Nanotubes. Phys. Rev.

Lett.2002, 89, 75505.

(138) Ozden, S.; Brunetto, G.; Karthiselva, N. S.; Galvão, D. S.; Roy, A.; Bakshi, S. R.;

Tiwary, C. S.; Ajayan, P. M. Controlled 3D Carbon Nanotube Structures by

Plasma Welding. Adv. Mater.2016.

(139) Tang, B.; Zhang, L.; Li, R.; Wu, J.; Hedhili, M. N.; Wang, P. Are Vacuum-Filtrated

Reduced Graphene Oxide Membranes Symmetric? Nanoscale2016, 8, 1108–

1116.

(140) Lin, Y.; Williams, T. V.; Xu, T. B.; Cao, W.; Elsayed-Ali, H. E.; Connell, J. W. 215

Aqueous Dispersions of Few-Layered and Monolayered Hexagonal Boron

Nitride Nanosheets from Sonication-Assisted Hydrolysis: Critical Role of

Water. J. Phys. Chem. C2011, 115, 2679–2685.

(141) Gorbachev, R. V.; Riaz, I.; Nair, R. R.; Jalil, R.; Britnell, L.; Belle, B. D.; Hill, E. W.;

Novoselov, K. S.; Watanabe, K.; Taniguchi, T.; et al. Hunting for Monolayer

Boron Nitride: Optical and Raman Signatures. Small2011, 7, 465–468.

(142) Bhimanapati, G. R.; Kozuch, D.; Robinson, J. Large-Scale Synthesis and

Functionalization of Hexagonal Boron Nitride Nanosheets. Nanoscale2014, 6,

11671–11675.

(143) Hou, J.; Li, G.; Yang, N.; Qin, L.; Grami, M. E.; Zhang, Q.; Wang, N.; Qu, X.

Preparation and Characterization of Surface Modified Boron Nitride Epoxy

Composites with Enhanced Thermal Conductivity. RSC Adv.2014, 4, 44282–

44290.

(144) Sudeep, P. M.; Vinod, S.; Ozden, S.; Sruthi, R.; Kukovecz, A.; Konya, Z.; Vajtai, R.;

Anantharaman, M. R.; Ajayan, P. M.; Narayanan, T. N. Functionalized Boron

Nitride Porous Solids. RSC Adv.2015, 5, 93964–93968.

(145) Zeng, X.; Ye, L.; Yu, S.; Li, H.; Sun, R.; Xu, J.; Wong, C.-P. Artificial Nacre-like

Papers Based on Noncovalent Functionalized Boron Nitride Nanosheets with

Excellent Mechanical and Thermally Conductive Properties. Nanoscale2015,

7, 6774–6781. 216

(146) Qi, X.; Yang, L.; Zhu, J.; Hou, Y.; Yang, M. Stiffer but More Healable Exponential

Layered Assemblies with Boron Nitride Nanoplatelets. ACS Nano2016, 10,

9434–9445.

(147) Yin, J.; Li, X.; Zhou, J.; Guo, W. Ultralight Three-Dimensional Boron Nitride

Foam with Ultralow Permittivity and Superelasticity. Nano Lett.2013, 13,

3232–3236.

(148) Torrents, A.; Schaedler, T. A.; Jacobsen, A. J.; Carter, W. B.; Valdevit, L.

Characterization of Nickel-Based Microlattice Materials with Structural

Hierarchy from the Nanometer to the Millimeter Scale. Acta Mater.2012, 60,

3511–3523.

(149) Gibson, L. J. Modelling the Mechanical Behavior of Cellular Materials. Mater.

Sci. Eng. A1989, 110, 1–36.

(150) Gibson, L. J.; Ashby, M. F.; Zhang, J.; Triantafillou, T. C. Failure Surfaces for

Cellular Materials under Multiaxial Loads-I.Modelling. Int. J. Mech. Sci.1989,

31, 635–663.

(151) Esteves, I. A. A. C.; Cruz, F. J. A. L.; Mu, E. A.; Agnihotri, S. Determination of the

Surface Area and Porosity of Carbon Nanotube Bundles from a Langmuirian

Analysis of Sub- and Supercritical Adsorption Data. Carbon N. Y.2010, 47,

948–956.

(152) Bae, Y.; Snurr, R. Q. Minireviews Development and Evaluation of Porous 217

Materials for Carbon Dioxide Separation and Capture. Angew. Chemie - Int.

Ed.2011, 50, 11586–11596.

(153) Casco, M. E.; Martı, M.; Rodrı, F. Effect of the Porous Structure in Carbon

Materials for CO 2 Capture at Atmospheric and High-Pressure. Carbon N.

Y.2013, 67, 230–235.

(154) Alessandro, D. M. D.; Smit, B.; Long, J. R. Carbon Dioxide Capture Carbon

Dioxide Capture : Prospects for New Materials Angewandte. Angew. Chemie -

Int. Ed.2010, 49, 6058–6082.

(155) Stankovich, S.; Dikin, D. a.; Dommett, G. H. B.; Kohlhaas, K. M.; Zimney, E. J.;

Stach, E. a.; Piner, R. D.; Nguyen, S. T.; Ruoff, R. S. Graphene-Based Composite

Materials. Nature2006, 442, 282–286.

(156) Podsiadlo, P.; Kaushik, A. K.; Arruda, E. M.; Waas, A. M.; Sup, B.; Xu, J.;

Nandivada, H.; Pumplin, B. G.; Lahann, J.; Kotov, N. A. Ultrastrong and Stiff

Layered Polymer Nanocomposites. Science (80-. ).2007, 318, 80–83.

(157) Bonderer L.J, Studart A.R, G. L. . Bioinspired Design and Assembly. Science (80-

. ).2008, 319, 1069–1073.

(158) Geim, A. K. Graphene : Status and Prospects. Science (80-. ).2009, 324, 1530–

1534.

(159) Ajayan, P. M.; Schadler, L. S.; Giannaris, C.; Rubio, A. Single-Walled Carbon

Nanotube-Polymer Composites: Strength and Weakness. Adv. Mater.2000, 12, 218

750–753.

(160) Schadler, L. S.; Giannaris, S. C.; Ajayan, P. M. Load Transfer in Carbon

Nanotube Epoxy Composites. Appl. Phys. Lett.1998, 73, 3842.

(161) Giannelis, E. Polymer Layered Silicate Nanocomposites. Adv. Mater.1996, 8,

29–35.

(162) Balazs, A. C.; Emrick, T.; Russell, T. P. Nanoparticle Polymer Composites:

Where Two Small Worlds Meet. Science (80-. ).2006, 314, 1107–1110.

(163) Srivastava, S.; Schaefer, J. L.; Yang, Z.; Tu, Z.; Archer, L. A. 25th Anniversary

Article: Polymer-Particle Composites: Phase Stability and Applications in

Electrochemical Energy Storage. Adv. Mater.2014, 26, 201–234.

(164) Liff, S. M.; Kumar, N.; McKinley, G. H. High-Performance Elastomeric

Nanocomposites via Solvent-Exchange Processing. Nat. Mater.2007, 6, 76–83.

(165) Kharchenko, S. B.; Douglas, J. F.; Obrzut, J.; Grulke, E. A.; Migler, K. B. Flow-

Induced Properties of Nanotube-Filled Polymer Materials. Nat. Mater.2004, 3,

564–568.

(166) Raja, S. N.; Olson, A. C. K.; Limaye, A.; Thorkelsson, K.; Luong, A.; Lin, L.;

Ritchie, R. O.; Xu, T.; Alivisatos, A. P. Influence of Three-Dimensional

Nanoparticle Branching on the Young’s Modulus of Nanocomposites: Effect of

Interface Orientation. Proc. Natl. Acad. Sci. U. S. A.2015, 112, 6533–6538.

(167) Xie, H.; Wang, J.; Xi, T.; Liu, Y.; Ai, F.; Wu, Q. Thermal Conductivity 219

Enhancement of Suspensions Containing Nanosized Alumina Particles. J. Appl.

Phys.2002, 91, 4568–4572.

(168) Koerner, H.; Price, G.; Pearce, N. A.; Alexander, M.; Vaia, R. A. Remotely

Actuated Polymer Nanocomposites—stress-Recovery of Carbon-Nanotube-

Filled Thermoplastic Elastomers. Nat. Mater.2004, 3, 115–120.

(169) Zhan, G.-D.; Kuntz, J. D.; Wan, J.; Mukherjee, A. K. Single-Wall Carbon

Nanotubes as Attractive Toughening Agents in Alumina-Based

Nanocomposites. Nat. Mater.2003, 2, 38–42.

(170) Byrne, M. T.; Gun’ko, Y. K. Recent Advances in Research on Carbon Nanotube-

Polymer Composites. Adv. Mater.2010, 22, 1672–1688.

(171) Xu, Y.; Hong, W.; Bai, H.; Li, C.; Shi, G. Strong and Ductile Poly(vinyl

Alcohol)/graphene Oxide Composite Films with a Layered Structure. Carbon

N. Y.2009, 47, 3538–3543.

(172) Chatterjee, S.; Nafezarefi, F.; Tai, N. H.; Schlagenhauf, L.; Nüesch, F. a.; Chu, B. T.

T. Size and Synergy Effects of Nanofiller Hybrids Including Graphene

Nanoplatelets and Carbon Nanotubes in Mechanical Properties of Epoxy

Composites. Carbon N. Y.2012, 50, 5380–5386.

(173) Boufi, S.; Kaddami, H.; Dufresne, A. Mechanical Performance and

Transparency of Nanocellulose Reinforced Polymer Nanocomposites.

Macromol. Mater. Eng.2014, 299, 560–568. 220

(174) Zhu, H.; Zhu, S.; Jia, Z.; Parvinian, S.; Li, Y.; Vaaland, O.; Hu, L.; Li, T. Anomalous

Scaling Law of Strength and Toughness of Cellulose Nanopaper. Proc. Natl.

Acad. Sci.2015, 112, 1–6.

(175) Sarikaya, M. Biomimetics: Materials Fabrication through Biology. Proc. Natl.

Acad. Sci.1999, 96, 14183–14185.

(176) Nassif, N.; Pinna, N.; Gehrke, N.; Antonietti, M.; Jäger, C.; Cölfen, H. Amorphous

Layer around Aragonite Platelets in Nacre. Proc. Natl. Acad. Sci. U. S. A.2005,

102, 12653–12655.

(177) Gao, H.; Ji, B.; Jäger, I. L.; Arzt, E.; Fratzl, P. Materials Become Insensitive to

Flaws at Nanoscale: Lessons from Nature. Proc. Natl. Acad. Sci. U. S. A.2003,

100, 5597–5600.

(178) Somero, G. N. Adaptations to High Hydrostatic Pressure. Annu. Rev.

Physiol.1992, 54, 557–577.

(179) Lundon, K.; Bolton, K. Structure and Function of the Lumbar Intervertebral

Disk in Health, Aging, and Pathologic Conditions. J. Orthop. Sports Phys.

Ther.2001, 31, 291-303-306.

(180) Humzah, M. D.; Soames, R. W. Human Intervertebral Disc: Structure and

Function. Anat. Rec.1988, 220, 337–356.

(181) Bogduk, N. Clinical Anatomy of the Lumbar Spine and Sacrum; Elsevier Health

Sciences, 2005. 221

(182) Neumann, D. A. Kinesiology of the Musculoskeletal System: Foundations for

Rehabilitation; Elsevier Health Sciences, 2013.

(183) Eshelby, J. D. The Determination of the Elastic Field of an Ellipsoidal Inclusion,

and Related Problems. Proc. R. Soc. A Math. Phys. Eng. Sci.1957, 241, 376–396.

(184) Alin C. Chipara, Peter Samora Owuor, Sanjit Bhowmick, Gustavo Brunetto,

Syed Asif, Mircea Chipara, Robert Vajtai, Douglas S. Galvao, Chandra Sekhar

Tiwary, P. M. A. Structural Reinforcement Through through Liquid

Encapsulation. Adv. Mater. Interfaces2016.

(185) Jeong, S. H.; Chen, S.; Huo, J.; Gamstedt, E. K.; Liu, J.; Zhang, S.-L.; Zhang, Z.-B.;

Hjort, K.; Wu, Z. Mechanically Stretchable and Electrically Insulating Thermal

Elastomer Composite by Liquid Alloy Droplet Embedment. Sci. Rep.2015, 5.

(186) Bartlett, M. D.; Fassler, A.; Kazem, N.; Markvicka, E. J.; Mandal, P.; Majidi, C.

Stretchable, High-K Dielectric Elastomers through Liquid-Metal Inclusions.

Adv. Mater.2016, 28, 3726–3731.

(187) Schonhorn, H. Surface Tension-Viscosity Relationship for Liquids. J. Chem.

Eng. Data1967, 12, 524–525.

(188) Kavati, D. R. Airy Stress Function for Two Dimensional Inclusion Problems,

2007.

(189) Assael, M. J.; Armyra, I. J.; Brillo, J.; Stankus, S. V.; Wu, J.; Wakeham, W. A.

Reference Data for the Density and Viscosity of Liquid Cadmium, Cobalt, 222

Gallium, Indium, Mercury, Silicon, Thallium, and Zinc. J. Phys. Chem. Ref.

Data2012, 41.

(190) Lyapin, A. G.; Gromnitskaya, E. L.; Yagafarov, O. F.; Stalgorova, O. V.; Brazhkin,

V. V. Elastic Properties of Crystalline and Liquid Gallium at High Pressures. J.

Exp. Theor. Phys.2008, 107, 818–827.

(191) Kim, T. K.; Kim, J. K.; Jeong, O. C. Measurement of Nonlinear Mechanical

Properties of PDMS Elastomer. Microelectron. Eng.2011, 88, 1982–1985.

(192) Hallquist, J. O. LS-DYNA Theory Manual; Califonia, 2006.

(193) Brito, D.; Nataf, H. C.; Cardin, P.; Aubert, J.; Masson, J. P. Ultrasonic Doppler

Velocimetry in Liquid Gallium. Exp. Fluids2001, 31, 653–663.

(194) Lucy, L. B. A Numerical Approach to the Testing of the Fission Hypothesis.

Astron. J.1977, 82, 1013–1024.

(195) Gingold, Robert A., and J. J. M. Smoothed Particle Hydrodynamics: Theory and

Application to Non-Spherical Stars. Mon. Not. R. Astron. Soc.1977, 181.3, 375–

389.

(196) Allahdadi, F. A. High Strain Lagrangian Hydrodynamics: A Three Dimensional

SPH Code for Dynamic Material Response (No. PL-TR-92-1054); New Mexico,

1983.

(197) Dickey, M. D. Emerging Applications of Liquid Metals Featuring Surface

Oxides. ACS Appl. Mater. Interfaces2014, 6, 18369–18379. 223

(198) Bokobza, L. Multiwall Carbon Nanotube Elastomeric Composites: A Review.

Polymer (Guildf).2007, 48, 4907–4920.

(199) Fassler, A.; Majidi, C. Liquid-Phase Metal Inclusions for a Conductive Polymer

Composite. Adv. Mater.2015, 1928–1932.

(200) Style, R. W.; Boltyanskiy, R.; Allen, B.; Jensen, K. E.; Foote, H. P.; Wettlaufer, J.

S.; Dufresne, E. R. Stiffening Solids with Liquid Inclusions. Nat. Phys.2014, 11,

1–7.

(201) Boyce, M. C.; Arruda, E. M. Constitutive Models of Rubber Elasticity: A Review.

Rubber Chem. Technol.2000, 73, 504–523.

(202) Lee, H.; Dellatore, S. M.; Miller, W. M.; Messtersmith, P. B. Mussel-Inspired

Surface Chemistry for Multifunctional Coatings. Science2008, 318, 426–430.

(203) Fabritius, H. O.; Sachs, C.; Triguero, P. R.; Raabe, D. Influence of Structural

Principles on the Mechanics of a Biological Fiber-Based Composite Material

with Hierarchical Organization: The Exoskeleton of the Lobster Homarus

Americanus. Adv. Mater.2009, 21, 391–400.

(204) Nikolov, S.; Petrov, M.; Lymperakis, L.; Friák, M.; Sachs, C.; Fabritius, H. O.;

Raabe, D.; Neugebauer, J. Revealing the Design Principles of High-Performance

Biological Composites Using Ab Initio and Multiscale Simulations: The

Example of Lobster Cuticle. Adv. Mater.2010, 22, 519–526.

(205) Omenetto, F. G.; Kaplan, D. L. New Opportunities for an Ancient Material. 224

Science (80-. ).2010, 329, 528–531.

(206) Tang, Z.; Kotov, N. A.; Magonov, S.; Ozturk, B. Nanostructured Artificial Nacre.

Nat. Mater.2003, 2, 413–418.

(207) Taylor, D. Some of Nature’s Little Tricks. Mater. Today2010, 13, 6–7.

(208) Dunlop, J. W. C.; Weinkamer, R.; Fratzl, P. Artful Interfaces within Biological

Materials. Mater. Today2011, 14, 70–78.

(209) Patterson, J.; Martino, M. M.; Hubbell, J. A. Biomimetic Materials in Tissue

Engineering. Mater. Today2010, 13, 14–22.

(210) Al-Jawad, M. Creative Approaches in Biomimetic Materials Research. Mater.

Today2014, 17, 312–313.

(211) Hanlon, R. Cephalopod Dynamic Camouflage. Curr. Biol.2007, 17, 400–404.

(212) Ge, L.; Sethi, S.; Ci, L.; Ajayan, P. M.; Dhinojwala, A. Carbon Nanotube-Based

Synthetic Gecko Tapes. Proc. Natl. Acad. Sci.2007, 104, 10792–10795.

(213) Zan, G.; Wu, Q. Biomimetic and Bioinspired Synthesis of

Nanomaterials/Nanostructures. Adv. Mater.2016, n/a-n/a.

(214) Owuor, P. S.; Woellner, C. F.; Li, T.; Vinod, S.; Ozden, S.; Kosolwattana, S.;

Bhowmick, S.; Duy, L. X.; Salvatierra, R. V.; Wei, B.; et al. High Toughness in

Ultralow Density Graphene Oxide Foam. Adv. Mater. Interfaces2017, 1700030.

(215) Stender; Marchuk; Liu; Sander; Meyer; Smith; Neupane; Wang; Li; Cheng; et al. 225

Single Cell Optical Imaging and Spectroscopy. Chem. Rev.2013, 113, 2469–

2527.

(216) Cobley, C. M.; Chen, J.; Cho, E. C.; Wang, L. V; Xia, Y. Gold Nanostructures: A

Class of Multifunctional Materials for Biomedical Applications. Chem. Soc. Rev.

Chem. Soc. Rev2011, 40, 44–56.

(217) Wei Cui, Mingzhu Li, Jiyang Liu, Ben Wang, Chuck Zhang, Lei Jiang, and Q. C. A

Strong Integrated Strength and Toughness Artificial Nacre Based on

Dopamine Cross-Linked Graphene Oxide. ACS Nano2014, 9511–9517.

(218) Studart, A. R. Towards High-Performance Bioinspired Composites. Adv.

Mater.2012, 24, 5024–5044.

(219) Li, Y. Q.; Yu, T.; Yang, T. Y.; Zheng, L. X.; Liao, K. Bio-Inspired Nacre-like

Composite Films Based on Graphene with Superior Mechanical, Electrical, and

Biocompatible Properties. Adv. Mater.2012, 24, 3426–3431.

(220) Ternary, P.; Wan, S.; Li, Y.; Peng, J.; Hu, H.; Cheng, Q.; Jiang, L. Synergistic

Toughening of Graphene Oxide Molybdenum Disulfide-Thermoplastic

Polyurethane Ternary Artificial Nacre. ACS Appl. Mater. Interfaces2015, 708–

714.

(221) Zhang, C.; Mcadams, D. A.; Grunlan, J. C. Nano/Micro-Manufacturing of

Bioinspired Materials: A Review of Methods to Mimic Natural Structures. Adv.

Mater.2016, 1–30. 226

(222) Cha, Chaenyung, Su Ryon Shin, Nasim Annabi, Mehmet R. Dokmeci, A. K.

Carbon-Based Nanomaterials: Multifunctional Materials for Biomedical

Engineering. ACS Nano2013, 7, 2891–2897.

(223) Wegst, U. G. K.; Bai, H.; Saiz, E.; Tomsia, A. P.; Ritchie, R. O.; Ortiz, C.; Boyce, M.;

Wegst, U. G. K.; Bai, H.; Saiz, E.; et al. Bioinspired Structural Materials. Nat.

Mater.2014, 14, 23–36.

(224) Bouville, F.; Maire, E.; Meille, S.; Van de Moortèle, B.; Stevenson, A. J.; Deville, S.

Strong, Tough and Stiff Bioinspired Ceramics from Brittle Constituents. Nat.

Mater.2014, 13, 508–514.

(225) Yao, H. Bin; Ge, J.; Mao, L. B.; Yan, Y. X.; Yu, S. H. 25th Anniversary Article:

Artificial Carbonate Nanocrystals and Layered Structural Nanocomposites

Inspired by Nacre: Synthesis, Fabrication and Applications. Adv. Mater.2014,

26, 163–188.

(226) Bai, H.; Walsh, F.; Gludovatz, B.; Delattre, B.; Huang, C.; Chen, Y.; Tomsia, A. P.;

Ritchie, R. O. Bioinspired Hydroxyapatite/Poly(methyl Methacrylate)

Composite with a Nacre-Mimetic Architecture by a Bidirectional Freezing

Method. Adv. Mater.2015, 2, 50–56.

(227) Vecchio, K. S. Synthetic Multifunctional Metallic-Intermetallic Laminate

Composites. Jom2005, 57, 25–31.

(228) Deng, X.; Mammen, L.; Vollmer, D. Candle Soot as a Template for a 227

Transparent Robust Superamphiphobic Coating. Science (80-. ).2012, 335, 67–

70.

(229) Yao, X.; Hu, Y.; Grinthal, A.; Wong, T.-S.; Mahadevan, L.; Aizenberg, J. Adaptive

Fluid-Infused Porous Films with Tunable Transparency and Wettability. Nat.

Mater.2013, 12, 529–534.

(230) Seeboth, A.; Lötzsch, D.; Ruhmann, R.; Muehling, O. Thermochromic Polymers-

-Function by Design. Chem. Rev.2014, 114, 3037–3068.

(231) Seeboth, A.; Kriwanek, J.; Vetter, R. Novel Chromogenic Polymer Gel Networks

for Hybrid Transparency and Color Control with Temperature. Adv.

Mater.2000, 12, 1424–1426.

(232) Owuor, P. S.; Hiremath, S.; Chipara, A. C.; Vajtai, R.; Lou, J.; Mahapatra, D. R.;

Tiwary, C. S.; Ajayan, P. M. Nature Inspired Strategy to Enhance Mechanical

Properties via Liquid Reinforcement. Adv. Mater. Interfaces2017, 4.

(233) Tiwary, C. S.; Kishore, S.; Sarkar, S.; Mahapatra, D. R.; Ajayan, P. M.;

Chattopadhyay, K. Morphogenesis and Mechanostabilization of Complex

Natural and 3D Printed Shapes. Sci. Adv.2015, 1, 18–21.

(234) Trost, B. M. The Atom Economy-A Search for Synthetic Efficiency Published

by : American Association for the Advancement of Science The Atom

Economy-A Search for Synthetic Efficiency. Science (80-. ).2016, 254, 1471–

1477. 228

(235) Whaley, S. R.; English, D. S.; Hu, E. L.; Barbara, P. F.; Belcher, A. M. Selection of

Peptides with Semiconductor Binding Specificity for Directed Nanocrystal

Assembly. Nature2000, 405, 665–668.

(236) Sellinger, A.; Weiss, P. M.; Nguyen, A.; Lu, Y.; Assink, R. a; Gong, W.; Brinker, C.

J. Continuous Self-Assembly of Organic-Inorganic Nanocomposite Coatings

That Mimic Nacre. Nature1998, 394, 256–260.

(237) Fantner, G. E.; Hassenkam, T.; Kindt, J. H.; Weaver, J. C.; Birkedal, H.; Pechenik,

L.; Cutroni, J. a; Cidade, G. a G.; Stucky, G. D.; Morse, D. E.; et al. Sacrificial

Bonds and Hidden Length Dissipate Energy as Mineralized Fibrils Separate

during Bone Fracture. Nat. Mater.2005, 4, 612–616.

(238) Wei, X.; Naraghi, M.; Espinosa, H. D. Optimal Length Scales Emerging from

Shear Load Transfer in Natural Materials: Application to Carbon-Based

Nanocomposite Design. ACS Nano2012, 6, 2333–2344.

(239) Ritchie, R. O. The Conflicts between Strength and Toughness. Nat. Mater.2011,

10, 817–822.

(240) Studart, A. R. Bioinspired Ceramics: Turning Brittleness into Toughness. Nat.

Mater.2014, 13, 433–435.

(241) Tanay, A.; Bitincka, L.; Shamir, R.; Shea, E. K. O.; Tschumi, A. I.; Kishony, R.;

Telmer, C. A.; Adler, S. A.; Subramaniam, V.; Lopez, A. J.; et al. Tough, Bio-

Inspired Hybrid Materials. Science (80-. ).2008, 322, 1516–1520. 229

(242) Yeom, B.; Sain, T.; Lacevic, N.; Bukharina, D.; Cha, S.-H.; Waas, A. M.; Arruda, E.

M.; Kotov, N. A. Abiotic Tooth Enamel. Nature2017, 543, 95–98.

(243) K Chen, B Shi, Y Yue, J Qi, L. G. Binary Synergy Strengthening and Toughening

of Bio-Inspired Nacre-like Graphene Oxide / Sodium Alginate Composite

Paper. ACS Nano2015, 9, 8165–8175.

(244) Huang, S.; Phua, S. L.; Liu, W.; Ding, G.; Lu, X. Nacre-like Composite Films Based

on Mussel-Inspired “glue” and Nanoclay. RSC Adv.2014, 4, 1425.

(245) Gumbart, J.; Aksimentiev, A.; Tajkhorshid, E.; Wang, Y.; Schulten, K. Molecular

Dynamics Simulations of Proteins in Lipid Bilayers. J. Comput. Chem.2007, 15,

80–90.

(246) Janmey, P. A.; Winer, J. P.; Weisel, J. W. Fibrin Gels and Their Clinical and

Bioengineering Applications. J. R. Soc. Interface2009, 6, 1–10.

(247) Lengauer, T.; Rarey, M. Computational Methods for Biomolecular Docking.

Curr. Opin. Struct. Biol.1996, 6, 402–406.

(248) Abeyrathne, E. D. N. S.; Lee, H. Y.; Ahn, D. U. Egg White Proteins and Their

Potential Use in Food Processing or as Nutraceutical and Pharmaceutical

Agents – a Review. Poult. Sci.2013, 92, 3292–3299.

(249) Stein, P. E.; Leslie, A. G. W.; Finch, J. T.; Carrell, R. W. Crystal Structure of

Uncleaved Ovalbumin at 1.95Å Resolution. J. Mol. Biol.1991, 221, 941–959.

(250) Pettersen, E. F.; Goddard, T. D.; Huang, C. C.; Couch, G. S.; Greenblatt, D. M.; 230

Meng, E. C.; Ferrin, T. E. UCSF Chimera - A Visualization System for

Exploratory Research and Analysis. J. Comput. Chem.2004, 25, 1605–1612.

(251) Harris, R; Olson, A. J. G. D. S. Automated Prediction of Ligand-Binding Sites in

Proteins. Proteins2008, 70, 1506–1517.

(252) Oleg, T. O. A. J. AutoDock Vina: Improving the Speed and Accuracy of Docking

with a New Scoring Function, Efficient Optimization, and Multithreading. J.

Comput. Chem.2010, 31, 455–461.

(253) Tsafack, T.; Alred, J. M.; Wise, K. E.; Jensen, B.; Siochi, E.; Yakobson, B. I.

Exploring the Interface between Single-Walled Carbon Nanotubes and Epoxy

Resin. Carbon N. Y.2016, 105, 600–606.

(254) Owuor, P. S.; Tsafack, T.; Agrawal, H.; Hwang, H. Y.; Zelisko, M.; Li, T.;

Radhakrishnan, S.; Park, J. H.; Yang, Y.; Stender, A. S.; et al. Poly-Albumen: Bio-

Derived Structural Polymer from Polymerized Egg White. Mater. Today

Chem.2018, 9, 73–79.

(255) Findley, W. N. A Modified Superposition Principle Applied to Creep of

Nonlinear Viscoelastic Material Under Abrupt Changes in State of Combined

Stress. J. Rheol. (N. Y. N. Y).1967, 11, 361.

(256) Firth, E. C. The Response of Bone, Articular Cartilage and Tendon to Exercise

in the Horse. J. Anat.2006, 208, 513–526.

(257) Murugan, R.; Ramakrishna, S. Modification of Demineralized Bone Matrix by a 231

Chemical Route. J. Mater. Chem.2004, 14, 2041–2045.

(258) Sudarsanan, K.; Young, R. A. Significant Precision in Crystal Structural Details.

Holly Springs Hydroxyapatite. Acta Crystallogr. Sect. B Struct. Crystallogr.

Cryst. Chem.1969, 25, 1534–1543.

(259) Almora-Barrios, N.; De Leeuw, N. H. A Density Functional Theory Study of the

Interaction of Collagen Peptides with Hydroxyapatite Surfaces.

Langmuir2010, 26, 14535–14542.