The Functionalization and Characterization of Adherent

THE FUNCTIONALIZATION AND CHARACTERIZATION OF ADHERENT

CARBON NANOTUBES WITH SILVER NANOPARTICLES FOR BIOLOGICAL

APPLICATIONS

A thesis submitted in partial fulfillment of the requirements for the degree of Master of Science in Engineering

ADAM ANTHONY MALESZEWSKI B.S., Wright State University, 2007

2011 Wright State University

WRIGHT STATE UNIVERSITY

SCHOOL OF GRADUATE STUDIES

May 9, 2011

I HEREBY RECOMMEND THAT THE THESIS PREPARED UNDER MY SUPERVISION BY Adam Anthony Maleszewski ENTITLED The Functionalization and Characterization of Adherent Carbon Nanotubes with Silver Nanoparticles for Biological Applications BE ACCEPTED IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF Master of Science in Engineering.

Sharmila M. Mukhopadhyay, Ph.D.

Thesis Director

George Huang, Ph.D., PE, Chair, Mechanical and Materials Engineering College of Engineering and Computer Science Committee on Final Examination

Sharmila M. Mukhopadhyay, Ph.D.

Saber Hussain, Ph.D.

Allen Jackson, Ph.D.

Andrew Hsu, Ph.D. Dean, School of Graduate Studies

Abstract

Maleszewski, Adam Anthony. M.S.Egr., Department of Mechanical and Materials Engineering, Wright State University, 2011. The Functionalization and Characterization of Adherent Carbon Nanotubes with Silver Nanoparticles for Biological Applications.

The purpose of this project is to form silver nanoparticles (Ag-NP) attached to a hierarchical substrate for possible use in biological applications. The effectiveness of these Ag-NP-containing devices, including biofilters and biosensors, may be dramatically enhanced by the use of hierarchical structures such as carbon nanotubes (CNT), as they offer a high surface area surface suitable for cell-device interactions, while Ag-NP would be a suitable component in many such devices due to its plasmonic surface properties

(e.g. in sensor and directed energy applications) and its anti-microbial properties

(desirable for fluid filtration due to its low weight. Meaningful control over the Ag-NP sizes and degree of adherence has been achieved. The interaction between silver and human epidermal cells as would be present in Ag-NP-based wound dressings has also been investigated in vitro and is discussed.

iii

Table of Contents

CHAPTER 1. INTRODUCTION ...... 1

CHAPTER 2. RESEARCH OBJECTIVES ...... 5

CHAPTER 3. LITERATURE REVIEW ...... 6

3.1 METAL NANOPARTICLE SYNTHESIS METHODS ...... 6

3.1.1 Electroless Deposition ...... 7

3.1.2 Photodeposition ...... 10

3.1.3 Gas-Catalyzed Solid-State Reduction ...... 11

3.1.4 Laser Ablation ...... 11

3.2 SUBSTRATE TOXICITY ...... 13

3.2.1 Carbon Nanotubes ...... 13

3.2.2 Silver Nanoparticles ...... 14

CHAPTER 4. SYNTHESIS PROCEDURE ...... 19

CHAPTER 5. SAMPLE COMPOSITION AND QUANTITY...... 24

5.1 X-RAY PHOTOELECTRON SPECTROSCOPY ...... 24

5.1.1 X-Ray Photoelectron Spectroscopy Procedure ...... 24

5.1.2a X-Ray Photoelectron Spectroscopy Discussion ...... 25

5.1.2b X-Ray Photoelectron Spectroscopy Individual Scans...... 27

5.2 ENERGY DISPERSIVE X-RAY SPECTROSCOPY...... 42

5.3 SCANNING ELECTRON MICROSCOPY ...... 44

5.3.1 Scanning Electron Microscope Measurement Procedure ...... 47

5.3.2 Scanning Electron Microscopy Results ...... 53

CHAPTER 6. SONICATION-BASED ANALYSIS ...... 57

6.1 SONICATION-BASED ANALYSIS PROCEDURE ...... 57

6.2 SONICATION-BASED ANALYSIS RESULTS/DISCUSSION ...... 58

CHAPTER 7. BIOCOMPATIBILITY ASSESSMENT ...... 66

7.1 FLUORESCENCE-BASED BIOCOMPATIBILITY ANALYSIS PROCEDURE ...... 66

7.2 BIOCOMPATIBILITY RESULTS/DISCUSSION ...... 70

CHAPTER 8. SUMMARY AND CONCLUSIONS ...... 76

CHAPTER 9. FUTURE RECOMMENDATIONS ...... 78

9.1 BIOSENSOR ...... 78

9.2 BIOFILTER ...... 79

9.3 WOUND DRESSING ...... 80

APPENDIX 81

A. WETTING ANGLE ANALYSIS ...... 81

B. SAMPLE OF BACKGROUND CHOICES OF XPS SPECTRA ...... 85

C. 80° REDUCED SAMPLE 10,000X MAGNIFIED MICROGRAPH ...... 88

D. CHEMICALS/EQUIPMENT ...... 88

REFERENCES 92

List of Figures

Figure 4.1: A schematic displaying the substrate to finished hybrid structure...... 20

Figure 4.2: A schematic of the apparatus mid-reduction. For simplicity, the thermocouple and stir bar are not pictured. The stir bar would appear in the sheath which supports the sample shelf, while the thermocouple would appear adjacent to the top face of the sample...... 23

Figure 5.1: A AgNO3 control sample general scan XPS spectrograph showing a small C

1s peak due to a layer of silver nitrate on the sample surface (A 60 °C reduction temperature sample shown)...... 28

Figure 5.2: A reduced sample general scan XPS spectrograph showing a large C 1s peak due metallic nanoparticles beading on the surface revealing CNTs (sample 1 shown). ... 29

Figure 5.3: A composite spectrograph depicting XPS spectra collected in different locations of the AgNO3 control sample in the C 1s regime...... 30

Figure 5.4: A composite spectrograph collected from three reduced samples depicting

XPS spectra in the C 1s regime...... 31

Figure 5.5a: A spectrograph of a reduced sample fit with three curves lying at 284.5,

285.8, and 290.3 eV...... 32

Figure 5.5b: A composite spectrograph comparing the C 1s peaks of three reduced samples with bare HOPG, indicating the formation of a carbon bonding structure during the reduction process. The HOPG peaks lie on the bottom at the position indicated...... 33

Figure 5.6: A composite spectrograph depicting three XPS spectra collected in different locations of the AgNO3 control sample in the Ag 3d regime...... 34

Figure 5.7: A composite spectrograph collected from three reduced samples depicting

XPS spectra in the Ag 3d regime...... 35

Figure 5.8: A composite spectrograph depicting three XPS spectra collected in different locations of the AgNO3 control sample in the N 1s regime...... 36

Figure 5.9: A composite spectrograph collected from three reduced samples depicting

XPS spectra in the N 1s regime...... 37

Figure 5.10: A composite spectrograph depicting three XPS spectra collected in different locations of the AgNO3 control sample in the O 1s regime...... 38

Figure 5.11: A composite spectrograph collected from three reduced samples depicting

XPS spectra in the O 1s regime...... 39

Figure 5.12: A composite spectrograph depicting three XPS spectra collected in different locations of the AgNO3 control sample in the S 2p regime...... 40

Figure 5.13: A composite spectrograph collected from three reduced samples depicting

XPS spectra in the S 2p regime...... 41

Figure 5.14: An EDX spectrum of a 60 °C sample after synthesis (top) and the accompanying quantification (bottom)...... 43

Figure 5.15: A 60 °C reduced sample micrograph at 100kx magnification...... 44

Figure 5.16: An 80 °C reduced sample micrograph at 100kx magnification...... 46

Figure 5.17: An 80 °C reduced sample micrograph at 10kx magnification showing patches of coalesced particles...... 47

Figure 5.18: An SEM micrograph of unfunctionalized MWCNTs with measured iron particles at 250kx magnification...... 50

vii

Figure 5.19: As-collected micrograph of a 60 °C reduced sample showing silver nanoparticles at 250kx magnification...... 51

Figure 5.20: An analyzed micrograph of a 60 °C reduced sample showing measured silver nanoparticles at 250kx...... 51

Figure 5.21: An as-collected micrograph of an 80 °C reduced sample showing silver nanoparticles at 250kx...... 52

Figure 5.22: An analyzed micrograph of an 80 °C reduced sample showing measured silver nanoparticles at 250kx...... 52

Figure 5.23: Histogram comparing reduction temperatures and resulting mean particle sizes...... 54

Figure 6.1: Schematic of the top and side views, respectively, of the sonication-based analysis experimental setup...... 58

Figure 6.2: A SEM micrograph of a CNT sample after five minutes of sonication...... 59

Figure 6.3: A SEM micrograph of a 60 °C reduced sample after ten seconds of sonication...... 60

Figure 6.4: A SEM micrograph of a 60 °C reduced sample after ten seconds of sonication...... 61

Figure 6.5: A SEM micrograph depicting a sonicated 60 °C reduced sample showing an intact silica layer (left) and the underlying graphite (right)...... 62

Figure 6.6: SEM micrographs of a 60 °C reduced sample after a ten-second sonication.

...... 63

viii

Figure 6.7: A schematic representing forces exerted by sonication on CNTs and Ag-NPs during a single pulse. Given the attachment area of a CNT as A and an attachment area of a particle as a, F/A >> f/a...... 65

Figure 7.1: A micrograph of fluorescent HaCaT cells on a glass substrate after two days’ growth...... 70

Figure 7.2: A micrograph of fluorescent HaCaT cells on HOPG + CNTs after two days’ growth...... 71

Figure 7.3: A micrograph of fluorescent HaCaT cells on HOPG + CNTs + Ag NPs after two days’ growth...... 73

Figure A.1: A schematic of one wetting angle measurement...... 82

Figure A.2: A sample micrograph depicting wetting angle measurements...... 83

Figure A.3: A schematic depicting one challenge in wetting angle measurement...... 84

Figure C.1: A micrograph of an 80 °C reduced sample displaying a large patch of agglomerated silver particles...... 88

Figure D.1: Kratos X-Ray Photoelectron Spectrometer ...... 91

List of Tables

Table 5-1: A comparison of selected atomic percentages (≥ 0.5 at %) based on averaged regional quantifications of XPS spectra…………………………………………………….………………….26

Acknowledgements

First, I would like to thank my advisor, Dr. Sharmila Mukhopadhyay. Her expertise and continual guidance have shaped this thesis in more ways than can be counted. Her technical counsel in surface science has been invaluable. Likewise, the members of the toxicology research group directed by Dr. Saber Hussain at Wright Patterson Air Force

Base were hugely helpful when I regularly came to them to clear up confusions in my understanding of biology.

I would also like to thank Dr. Raghavan Srinivasan and Dr. Allen Jackson, who have supplied me with countless hours of technical discussions and explanations – they will be missed in my future career. I would also like to thank Mr. James Morely -- the first man to believe in (and encourage) my intellectual abilities who did not also attend our family’s Thanksgiving dinner.

I would also like to thank my parents, Vicky and Ed, who have supported me through the process of getting this thesis, as well as my spiritual inspirations, Ayn Rand and Adam

Carolla.

Chapter 1. Introduction

Carbon is a non-metallic element which is tetravalent in its electrically neutral state. It is remarkable in that it exhibits a wide variety of thermodynamically stable or metastable

(at room conditions) allotropes including diamond, graphite, nanotube, amorphous carbon, and buckminsterfullerene (also known as buckyball). Each of these provides remarkably different properties from one another despite sharing the same constituent element. As a result, carbon has been used in applications as diverse as fuel, jewelry, high-performance cutting blades, and writing utensils. Since the 1990s, a renaissance of carbon research has been underway to exploit the advantages of its more newly discovered allotropes. Of particular interest have been fullerenes – the family to which buckyballs and nanotubes belong. Using these nanostructures, carbon has become a highly investigated element, particularly in the fields of biomedical and electrical engineering. Because it exhibits high electrical conductivity, many stable allotropes, low chemical reactivity at room temperature, and biocompatibility, carbon is a prime candidate for many such applications.

At room temperature and standard atmospheric pressure, the most stable form of carbon is graphite. Graphite is comprised of stacked layers of single-atom-thick sheets of covalently bound carbon atoms. These layers are individually known as “graphene”. 1

Each graphene plane is bound to the next primarily by the relatively weak van der Waals forces, yielding a high degree of anisotropy to the material (i.e., the electrical and thermal conductivity is much better along the planes than through them). Each carbon atom in graphite is primarily bound to only three in-plane neighbors, which gives rise to graphite’s unique electrical properties. Since each carbon atom has four valence electrons but forms only three bonds, one valence electrons is delocalized; this imbues graphite with high electrical conductivity, a trait not shared by purely covalently bound allotropes such as diamond.

Another carbon allotrope, the single-walled carbon nanotube (SWCNT), has a structure which may be envisioned as a graphene sheet rolled into a tube (although this is not how they are formed; rather, current methods form them beginning at one end of the CNT and ending at the other). A similar structure is the multi-walled carbon nanotube (MWCNT), which is comprised of a concentric encapsulation of one or more SWCNTs by a larger

SWCNT, resulting in an ever-broadening outer diameter with each additional nanotube encapsulation. Several layers of nanotube may be encapsulated in this fashion, yielding an outer diameter of 20 nm or more – an order of magnitude greater than an individual

SWNT. These nanotubes, like the layers of graphene comprising graphite, are prone to agglomeration due to significant secondary bonding between them. They have been investigated as high-surface-area supports for catalytic applications since they can increase the surface area of a bulk substrate by several orders of magnitude [1].

The other main component studied herein is the metal silver. Silver is a transition metal with an atomic number of 47. Silver has many remarkable properties, including the highest electrical and thermal conductivity of any pure metal under room conditions. It 2 has been under increased investigation in the biological community as a supplemental bactericide to antibiotics since an increasing number of bacteria have become resistant to the effects of more traditional antibiotics.

Additionally, silver exhibits what is known as surface plasmon resonance (SPR). SPR refers to the propagation of some electromagnetic waves (known as surface plasmon- polaritons) coupled to oscillations of free electrons along the surface of a metal [2, 3].

The phenomena was first explained by Otto, Kretschmann, and Raether in 1968 [4, 5], and has led to hundreds of studies into ways to exploit this remarkable property. SPR occurs in, but is not limited to, noble metals. The mechanism for SPR is the excitation of surface plasmon electromagnetic waves (SPW) along the surface of a surface plasmon resonant material (that is, parallel to the interface between the surface of the dielectric and the metal) by an incident electromagnetic wave. These SPWs occur in surface plasmon resonant materials instead of reflection, which is the typical path for EM waves which impinge on a metal surface.

The incident angle at which the maximum amount of incident light is converted into

SPWs instead of reflected light is known as the SPR angle. This SPR angle is extremely sensitive to changes in surface bonding (such as the introduction of a biological agent onto the surface of an surface plasmon resonant material), as it is largely determined by the index of refraction within the immediate vicinity of the metal-dielectric interface. It is this concept that forms the scientific basis for optical biosensors.

Although carbon and silver have compelling biological properties, the biocompatibility of their nanoscale forms is still being ascertained. This process is desperately needed

3 because the amount manufactured and the variety of applications of nanoscale forms of carbon and silver are increasing. Because biological toxicity depends heavily on the individual factors of a substrate, I have included baseline toxicity analysis of a common exposure route, epidermal contact, of the hybrid structure synthesized.

Chapter 2. Research Objectives

The primary objective for this research was to synthesize novel high surface area, silver- nanoparticle-laden composites for use in biological applications. These particles were to be metallic, strongly adhered to the substrate, and abundant. The reasons for these requirements are the following: If the silver nanoparticles are in an oxidized form, they will not exhibit surface plasmon resonance. If they are weakly adhered, the useful life of such a composite in service will be limited. Because we are attempting to create a surface-area-dependent device, abundance is obviously critical because it directly relates to the efficacy of the device. Each of these qualities was to be demonstrated. Also, toxicity was to be evaluated to provide baseline toxicology information for this novel composite and to provide data useful for evaluating possible future applications which may involve contact with human cells.

A secondary goal was to determine if the size of the nanoparticles could be easily modified by a change in the temperature at which the nanoparticles were generated.

Chapter 3. Literature Review

The literature review has been restricted to seminal or comprehensive studies related to nanoparticle synthesis or pertinent characterizations thereof. Section 3.1 provides an explanation of the most common methods of generating nanoparticles similar in size and chemistry to those synthesized for this research, while Section 3.2 addresses the known toxicity of the components of the composite that was synthesized in the present research.

In addition to the synthesis techniques used to generate Ag-NPs, ripening techniques have been explored to enlarge the particles at the expense of the total number of particles

[6]. This process is critical for particle modification to create high performance, surface- area-dependent devices.

Section 3.2 also contains summaries of the biological interactions of various cell lines in vitro, animal systems in vivo, and bacteria in the presence of silver nanoparticles. Carbon nanotubes have also been widely examined for toxicity, and those results will be briefly summarized in Section 3.2, as well.

3.1 Metal Nanoparticle Synthesis Methods

There are several current methods of nano-metal deposition onto uneven substrates such as carbon nanotubes. These approaches include electrostatic deposition (also known as

6 citrate reduction, Turkevich, or electroless deposition methods) [7 - 12], gas-catalyzed solid-state reduction [13], laser ablation [14, 15], and radiation reduction [16, 27].

Because the research reported in this thesis is a modification of electrostatic deposition, it is given the most thorough treatment. The other methods are discussed primarily in terms of their applicability to produce large numbers of metallic silver nanoparticles strongly bound to carbon nanostructures.

For the purposes of this discussion, the following definitions apply:

Capping agent – a molecule which, in whole or, more typically, in part, bonds with a nanoparticle in order to prevent atoms from bonding to its surface. This is used to minimize the final nanoparticle size and to control its surface chemistry.

Reducing agent – for the purposes of this study, a molecule which is primarily ionically bound, the cationic portion of which spontaneously binds with the non-metallic portion of the salt to be reduced.

3.1.1 Electroless Deposition

Electroless deposition is characterized by metal salt reduction in a solution containing a reducing agent, a capping agent (which may be the same chemical as the reducing agent), and sometimes a surfactant, depending on the substrate’s wettability. An early comprehensive study (and the first to utilize electron microscopy in analyzing the process) was performed in 1951 by chemist John Turkevich, who adapted a citrate capping process from prior research [18, 19]. In this original 1951 study, Turkevich used

7 this adaptation to synthesize gold in an aqueous solution at 100 °C via chloroauric acid reduction in sodium citrate. He quantified nanoparticle sizes by electron microscopic measurement, reporting a mean diameter of approximately 20 nm.

In the many follow-up studies which have been performed since, this electroless method has been refined as a method for both synthesizing metal nanoparticles and adhering those nanoparticles to substrates which have included carbon nanotubes (CNT).

However, in the studies which used this method to attach silver nanoparticles to CNTs, the attachment has been always prefaced by a pre-attachment step to functionalize the

CNT surface via an acid etch [7 - 12]. This introduction of moieties, thought to provide an “anchor” onto the otherwise inert CNT surface, necessarily causes damage in the form of fissures in the CNT [8]. This impairs the electrical conductivity of the CNT and hinders its usefulness in biosensing applications. Additionally, if the CNTs are adhered to a substrate before reduction, the site of its attachment to the substrate may be aggressively attacked. The acid etch may also selectively target the ends of the CNTs, opening the end and wicking the metal catalyst inside via capillary action [8, 20], thereby inhibiting its ability to act as a catalyst in surface-area dependent applications.

Following the reduction, the silver nanoparticles spontaneously adhere to functional CNT side groups located on the outside of the CNT wall [20]. These groups include carboxylic (-COOH), carbonyl (-C=O), and hydroxyl (-OH) groups [21]. The simplest reaction mechanism proposed is that of ion exchange, replacing the hydrogen H° with a metal M° (Equation 1) [22].

(Equation 1)

Many silver reductions using this method have been performed using a citrate, such as sodium citrate (NaCit), reportedly both as a capping agent [7, 9] and a reducing agent

[10, 23]. It has been reported that NaCit reduces the silver ions and caps them [24, 25].

Other reducing agents such as sodium borohydride and dimethyl sulfoxide (DMSO) have been used, as well [26, 7], but in the case where DMSO and NaCit are both used (as in the adaptation presented in this thesis), the mechanism of reduction is not well understood.

Because the salt is typically dissolved in water before it is reduced via this method, one role of DMSO is to act as a solvent since DMSO is capable of dissolving both polar and non-polar components. DMSO is especially useful in reductions in the presence of

CNTs, due to their hydrophobic nature.

The electroless deposition method itself has several advantages. It has the advantage of being capable of depositing metal nanoparticles on a three dimensional surface despite the fact that part of that surface may be shrouded from a single angle (as opposed to

“line-of-sight” methods). It is also highly scalable because it requires only the metal salt, the reducing and capping agents, and inexpensive instrumentation (e.g. a low-temperature heater, stirring mechanism, etc.). It also involves relatively benign equipment and chemicals, as opposed to high-energy lasers, explosive gasses, or high temperatures. It has achieved particles sizes from under 1 nm to approximately 50 nm [8, 27].

The main disadvantage of this process is that the introduction of complex molecules to the sample yields some lingering post-reduction impurities. As a result, further processing may be required depending on the intended application.

3.1.2 Photodeposition

In the photodeposition method, CNTs and a salt are mixed in an aqueous solution and irradiated with high-energy photons in order to reduce positively-charged metal ions [28,

16, 17]. The photons are known to create moieties on the surface of the CNTs, which are theorized to serve as nucleation sites for the reduced metal as in electroless deposition by acid etching [29]. The irradiated solution contains some form of a reducing agent (i.e. the reduction is thermodynamically favorable), but irradiation is still necessary due to the reaction’s inherently slow kinetics or prohibitively large activation energy [29]. The electrolysis of water is a proposed mechanism by which the silver itself is reduced, yielding independent H° [28]. Energy has been supplied via microwaves [29], gamma rays [7, 28], and ultraviolet radiation [16] to this end. Often, capping agents such as citrate are used as well [7], and the particle diameters achieved are in the 5-30 nm range

[16, 28]

Disadvantages of this process lie in the high-energy photons required. Unfortunately, as the energy of the radiation increases, potential hazards to the health of the operator increase as well. As such, training procedures and inspections of such electromagnetic wave generators, coupled with high energy costs for operating such equipment, are current hindrances to commercialization of this method.

3.1.3 Gas-Catalyzed Solid-State Reduction

Gas-catalyzed solid-state reduction is characterized by metal reduction via a chemical reaction with a gaseous environment such as at an elevated temperature [13, 30, 31].

This process has yielded nanoparticles of a relatively small size and distribution (5 ± 2 nm), but it has significant drawbacks.

One such drawback is the need for reactive gases such as H2 or C-O in the presence of elevated temperatures (approximately 200 °C) in order to reduce the metal. In addition to the safety concerns associated with such gases, an airtight chamber is necessary to avoid oxidative effects – for example, the conversion of C-O to the less reactive CO2. The need for airtightness also limits the ability of the operator to precisely measure the effective temperature using usual methods, and thus a period of fine tuning of the heating system controls may be required to achieve the proper inner-chamber temperature [30].

Advantages of this method include that the reduction does not require capping agents, thereby eliminating troublesome contamination issues. Also, it may be used to achieve a narrow distribution of impressively small sized particles (~ 5 nm diameter).

3.1.4 Laser Ablation

Laser ablation is a process by which a solid metal is placed into a vacuum or inert environment and then evaporated or sublimated (ablated) using ultra-short bursts of light from a laser. A pure metal donor is placed in front of the desired substrate (often on the order of centimeters away), onto which the excited metal condenses [14, 15, 32]. A raw

11 vacuum may be adequate for deposition, but a gaseous atmosphere during the ablation may be used instead to restrict the expansion of the resulting plume [14]. Gases used are inert – most commonly helium or argon, which result in differing deposition rates due to the atomic number effects on the ablated material [33]. Other factors influencing particle size include laser intensity, target-substrate distance, laser wavelength, sample movement during deposition, radiation frequency, and the method and speed of changing the radiation focal point [14, 15, 32].

The disadvantages of this process include the high cost of the instrumentation required to produce and control highly directed femtosecond or picosecond-long radiation bursts.

Also, extensive optimization time is required because a detailed understanding of the interrelationships between the many controllable variables is difficult to achieve. Lastly, it is better suited for flat substrates than for 3-dimensional ones because the process requires a line of a sight from the donor to the desired substrate.

The advantages of this method include a high degree of deposition parameter control and the lack of impurities normally introduced by reducing or capping agents. It has also been used with a variety of substrates and deposited materials, including noble metals

(Au, Ag), ferromagnetic metals (Ni), semiconductors (Si), and superconductors (Mg ).

The parameters with respect to deposited particle size also remain remarkably consistent regardless of the material [32]. It also allows for exceptional control of deposited particle size from 1-200 nm [14].

A similar process has also been explored using mini-arc plasma in lieu of photons [6].

This technique relied on applying an electrical charge to the surface of the CNTs in order to attract the oppositely charged ions.

3.2 Substrate Toxicity

The toxicity of any novel material or composite should be evaluated before it is put into public use or production. In this case, there is extensive toxicity information on the constituent materials due to the tremendous interest in these materials over the past 20 years. These studies will be used to establish the level of knowledge and any significant knowledge gaps.

3.2.1 Carbon Nanotubes

The toxicity of carbon nanotubes has been a matter of great debate since they were brought to international awareness by Iijima in 1991 [34]. Because the CNTs are adhered to a substrate, one of the most serious pathways, inhalation, is not plausible in this case.

Therefore, our discussion will focus on another common exposure route; skin contact.

The in vitro toxicity results vary based on the type of CNT (MWCNT or SWCNT), the amount of catalyst remaining from synthesis, and the cell type in question. A correlation between increasing iron catalyst amount and decreasing cell viability has been observed

[35], but even nearly catalyst-free MWCNTs in their unattached state have been shown to

13 decrease viability significantly ( < 80% viable) in keratinocytes at 400 µg/mL in vitro

[36].

During evaluations of the toxic mechanisms, both MWCNTs and SWCNTs have been observed inside cell vacuoles when toxicity has also been observed [36, 37]. Therefore, it has been asserted that CNTs may have a reduced biological impact if they are attached to a substrate and therefore cannot be taken up by the cell. In support of this proposition, mouse fibroblasts have been shown to grow and proliferate on a MWCNT substrate in which the MWCNTs are constrained [38].

It has been reported that single-walled CNTs are more harmful than are MWCNTs [39,

40], but this observation is not universal [41], likely because other factors (catalyst used, agglomeration, etc.) are dominant below a certain size threshold. Colloidal carbon nanotubes are harmful to keratinocytes in moderate doses (≥ 200 µg/mL) [35, 36], at least somewhat independently of catalyst content.

In conclusion, adherent CNTs have been observed to be more biocompatible than those in solution because uptake is prevented. However, contamination of the nanotubes may also limit their biocompatibility since common catalysts used to form CNTs are not biocompatible (e.g. iron nanoparticles).

3.2.2 Silver Nanoparticles

The toxicity of Ag-NPs has been studied in considerable depth over the past decade, driven by a surge in anti-bacterial- and biosensor-related research. Various consumer

14 products seeking to exploit the bactericidal properties of silver have emerged, such as paint [42], catheters [43], water filters [25], and wound dressings [44]. Several exposure mechanisms have been investigated, including contact with skin and inhalation [44 – 46].

However, several issues are involved in obtaining objective toxicity results. Among them are unaddressed sources of error within tests [47] and difficulty in accurately relating dosages between colloidal solutions and adherent particles. In addition to these factors, the number of variables is considerable, making it difficult to relate one test to the next.

For example, particle size distributions, testing procedure used, cell line tested, and the capping agent used are all known to influence nanoparticle toxicity and are all variable from one study to the next, making comparison difficult.

Variations in reporting dosage amounts are especially problematic. Some studies report the dosage in terms of atomic percentage and in terms of weight percentage via XPS curve-fitting, respectively [27, 48], while others report silver weight added with respect to the total sample weight [49], by molar concentration in a colloid [50], and in parts per million with the mean particle radius included [51]. A standardization of reporting dosages, particularly for adherent (non-colloidal) nanoparticle toxicology information, is desperately needed.

As previously noted, the studied cell type also governs toxicity to some extent. While it has been reported that silver has a greater toxicological effect against gram-negative bacteria [52, 53], this result is not universal [49], and it is well-established that silver nanoparticles are effective against gram-positive strains as well [54]. Staphylococcus epidermidis, Methicillin-resistant Staphylococcus epidermidis (MRSE), and Methicillin-

15 resistant Staphylococcus aureus (MRSA) are common examples of gram-positive bacteria, while Escherichia coli (E. coli), Salmonella, and Legionella are examples of gram-negative bacteria. Less toxicity has been observed in eukaryotic cells than in bacteria, as well [54].

Many studies have reported a size-dependent effect of silver toxicity in the nanoscale

[55]. However, greater toxicity is observed in smaller particles at the same weight concentration because, in part, the same weight concentration of a smaller particle yields a much greater surface area. A recent study asserted that in both in vitro (A-549 epithelial cells) and in vivo (male Wistar rats’ lungs) and for a wide variety of nanomaterials, a more intuitive measure of nanomaterial toxicity should be adopted [56].

Surface area per volume rather than weight per volume was proposed as this more intuitive system. Such a change would offer a significant benefit, since weight per volume is currently used by many regulatory agencies such as the National Institute for

Occupational Safety and Health [57]. As such, these standards may be either too lax or too rigid, depending on the mean particle size. Both alternatives are injurious to the efforts of industry and research organizations that use such particles.

In selecting a more appropriate standard, the following factors should be maximized: ease of measurement, the accuracy and objectivity of a measurement, and the intuitive relationship between the measured number and resulting toxicity across a variety of cell types, if possible. This last factor refers to providing a knowledgeable researcher a clear understanding of the danger of an environment in simple terms. By analogy, let us consider a researcher who wishes to study a simple elastic collision in which two objects impact one another after travelling in opposite directions. If one wishes to study the force 16 imparted by a such collision, it would greatly aid the researchers in the field to discuss their experiments in terms of velocity and mass instead of convoluted variables such as: the angle the objects traveled from the observer’s perspective prior to impact, the apparent distance which initially separated them, the distance from the observer to the objects, the volume of the objects, and their densities. Note that these pieces of information can lead one to an understanding of collisions, but speaking in terms of the factors which are irreducibly simple – in this example, the velocities and masses of the objects – facilitates understanding.

Using any of these standards would be better than using all of them (in some cases, comparing values has been made impossible by this variety).

With it being said that surface area per volume is a better indicator of toxicity than mass per volume, some particle sizes may be too large to damage cells through a mechanism requiring cell uptake [55]. As such, surface area is not expected to have a linear relationship with toxicity through all regimes.

The sources of actual error involved in this testing (as opposed to the factors discussed which merely make it difficult to relate results from one study to another) are also considerable to be overcome in this arena. As one such example, a staple of toxicological in vitro research, the MTT assay, has recently been criticized for indicating false cytotoxicity when carbon nanotubes are used, likely due to reactions between these constituent materials and some tetrazolium salts such as MTT [47]. This and other sources of error may be critical, given that such methods are still in use with hybrid structures [45, 58].

In conclusion, in modern silver nanoparticle research, there is a lack of standardization which inhibits the comparison of studies. Many studies are incomparable due to varying methodologies, capping agents, exposure times, cell types, toxicity measurements, particle sizes, dosages, etc. As such, we can conclude relatively little from the abundance of studies available. The conclusions which are consistent include the following:

Smaller particle sizes yield more toxicity if the particle concentration by weight remains constant. Bacteria, especially gram-negative bacteria, are harmed more readily by silver contact than are eukaryotic cells. Silver has been observed to accumulate in the organs of mammals; significantly in the liver, kidneys, and spleen [59, 60], but a significant amount of short term (< 1 week) clearance has also been observed [61]. Lastly, in vitro studies have found a greater degree of toxicity than in vivo studies have, so care should be taken to avoid needless alarm based solely on the former. Quite to the contrary, repeated examinations of consumer products containing silver such as wound dressings, silver catheter coatings, and colloidal solution ingestion have shown little toxicity [62, 63, 61].

Chapter 4. Synthesis Procedure

The synthesis of the Ag-NP + CNT + HOPG hybrid structure was broken into three main parts: 1) The MWCNTs are grown on a substrate comprised of HOPG and silica (Figure

4.1). 2) Silver nitrate is dried on this substrate. 3) The silver nitrate is reduced by an adapted electroless method in order to achieve a hierarchical substrate consisting of

HOPG, silica, MWCNTs, and Ag-NPs. In this section, the details of the latter two processes are described because they have not been previously outlined elsewhere.

The utensils used in the following step were all cleaned in the following manner between each step in the experiment: The utensils were sprayed with absolute ethanol and wiped dry with cotton paper towels. All utensils were visually inspected for fibers or other contaminants and the process was repeated until none were visible.

Step 1) Multi-walled carbon nanotubes (MWCNTs) were grown on highly-oriented pyrolitic graphite (HOPG), as outlined by Mukhopadhyay, et. al. [1] [Figure 4.1].

The MWCNT-HOPG hybrid structure was then heated to approximately 110 °C using a hot plate. To protect the light-sensitive silver nitrate, the experiment was carried out in a reduced-light setting by turning off the overhead lights, leaving only indirect light from a window with blinds closed. 40 mg of was well stirred in 1 mL distilled water to achieve a molarity of 0.24 M via magnetic stirring for approximately 10 seconds on setting “8”. At this point, no salt was visible on the bottom of the beaker, indicating

19 dissolution. The MWCNTs were then wetted by spraying absolute ethanol on the surface until it was entirely wetted, and, immediately, the solution was applied by pipetting one drop onto the surface.

Figure 4.1: A schematic displaying the substrate to finished hybrid structure.

Such drops had been weighed previously on a micro-balance, and were determined to weigh approximately 30 mg/drop. After the first drop was added, the next was added after a 45-second interval to allow partial drying to occur so as to avoid causing the liquid to spill over the side of the sample. When the last drop was added, the sample was immediately covered by an aluminum foil cover to further eliminate light. The

MWCNTs were dried in this condition at the same temperature for one half hour, at 20 which time the sample was placed into a clean plastic sample holder, which was wrapped in aluminum foil and placed in a dark drawer until it was used.

The final step in the synthesis, sample reduction, took place within two days of the previous step in each experiment. The utensils were first washed in ethanol and dried by cotton towel. The alumina shelf, alumina sheath, Teflon® stir bar, and glass beakers were then submersed in the solvent dimethyl sulfoxide (DMSO) in order to remove any remaining ethanol. This DMSO was then discarded and these instruments were used while still moist.

15 mL of DMSO were then added to a 25 mL beaker containing the alumina sheath, alumina shelf and stir bar [Figure 4.2]. The purpose of the alumina sheath was to support the alumina shelf, which in turn supported the sample. This arrangement was implemented to avoid collisions between the sample and the stir bar during mixing while avoiding adhesive contamination inherent in taping the samples inside the beaker. This beaker was then placed on a hotplate stirrer. A cleaned thermocouple probe was then inserted into the solution until it rested at approximately the same height at which the top face of the sample would set during reduction. The hotplate stirrer’s stir control was set to “5” at this time to initiate stirring. The hotplate was then switched on, and the appropriate input temperature was dialed in once the solution’s temperature stabilized.

This input temperature was not constant between iterations because of room temperature variations. This process consisted of simply manually adjusting the setting higher or lower depending on the readout of the thermocouple after sufficient time was allowed for temperature stabilization.

When the solution stabilized within ± 0.5 °C of the desired reduction temperature, 4.4 ±

0.2 mg of sodium citrate dihydrate (NaCit) was measured on a calibrated microbalance and added to the solution for a concentration of 1 mM.

This reduction temperature in the DMSO, based on the thermocouple readings, was set at either 60 °C or 80 °C in the experiments. These two temperatures were used on different samples to study the effects of reduction temperature.

At this time, the room lights were turned off, leaving a dimmed, reduced light environment. Approximately 30 seconds elapsed to allow for NaCit dissolution in

DMSO, at which time a prepared sample (i.e. a -supporting sample as previously described) was inserted onto the shelf using stainless steel tweezers. The beaker was immediately covered with aluminum foil, leaving a slot for the thermocouple to remain in position throughout the experiment. One hour elapsed, during which time the temperature varied by less than 1 °C. Based on previous iterations of this experiment in which the beaker was not covered with foil, the observed solution color changed from clear to pink within five minutes; then to burgundy; and then finally to an opaque, nearly black liquid over the course of the hour. This darkening is likely due to the agglomeration of particles which remained in solution.

At this time, the sample was removed with tweezers and shaken by hand in a beaker of room-temperature distilled water. The sample was allowed to sit in the beaker for approximately 5 minutes before another vigorous agitation and was then removed with tweezers. The reason for the agitation was to remove any loosely bound particles from

22 the sample surface. The sample was then sprayed with absolute ethanol to speed drying and placed in a clear plastic sample holder until analysis could take place.

In early trial runs, the amount of silver adhered to the substrate was minimal. A large increase in the amount of silver deposited was observed after the procedure was modified so that instead of adding the silver nitrate to the DMSO + NaCit solution containing the sample, the silver nitrate was dried on the substrate prior to reduction according to the procedure in Chapter 4.

Chapter 5. Sample Composition and Quantity

5.1 X-Ray Photoelectron Spectroscopy

An understanding of the chemistry of the particles adhered to the CNTs is essential to the suitability of the hybrid structure for virtually all potential applications. To this end, the hybrid structures synthesized were evaluated using a variety of methods: X-Ray

Photoelectron Spectroscopy (XPS), Energy Dispersive Spectroscopy (EDS), and

Scanning Electron Microscopy (SEM).

5.1.1 X-Ray Photoelectron Spectroscopy Procedure

A HOPG + CNT hybrid structure was coated in silver nitrate and evaluated via XPS as a reference for peak locations, peak shapes, and relative peak intensities. This unreduced sample was prepared in the following manner: 22 mg of were dissolved and well-mixed in 0.5 mL of distilled water and were added to a CNT + HOPG sample drip- wise, using the same method outlined in Chapter 4. The sample was then placed on a 110

°C hotplate and covered from light for 30 minutes. The sample temperature was held at

50 °C overnight to ensure drying, and the sample was evaluated via XPS the following day. Spectra from three locations were collected on this control sample. It

24 should be noted that a much larger amount of silver is to be expected to be present in this sample because it had roughly an order of magnitude more silver nitrate on its surface than the reduced samples did initially (and reduction reduced the amount even further).

X-Ray photoelectron spectroscopic analysis was also carried out on two 60 °C samples and two 80 °C samples. A general scan was performed on each sample from 1100 eV to

0 eV, and any visible peaks or expected peaks were analyzed using a fine scan. Each peak location was then offset such that the C 1s peak energy was set to 284.5 eV. These peaks were quantified using Linear or Shirley fits, as indicated in Appendix B. As a rule, a Shirley background fit was chosen when the noise surrounding a particular peak was low and the background level varied from one side of a peak to the other. A linear fit was used if either of these conditions were not met.

These three samples’ spectra will be presented overlaid with one another for the sake of brevity.

5.1.2a X-Ray Photoelectron Spectroscopy Discussion

Each spectrum has been charge-corrected to C 1s at 284.5 eV. Each carbon main peak was approximately 284.3 eV before this correction, so all peaks were shifted 0.2 eV higher from their as-collected positions. The regions were quantified according to the boundaries and background fits detailed in Appendix B. These quantification data were compiled in Table 5-1 for all of the reduced and control samples (the data were averaged for like samples).

Based on this data, the nitrogen-to-oxygen ratio in the control sample is 1:3, which is expected of a nitrate sample (AgNO3). However, the silver-to-nitrogen ratio is 2:1, which is inconsistent with the expected silver nitrate value of 1:1. Since XPS is an extremely surface-sensitive technique (< 10 nm penetration depth), this discrepancy likely indicates partial reduction on the surface of the AgNO3 control sample, perhaps resulting from the mild heat and indirect light in the laboratory.

At % Control Sample At % Reduced Sample Ag 21.9 3.7 N 10.0 0.8 O 31.4 8.6 C 36.1 86.4 S 0.6 0.5 Table 5.1: A comparison of selected atomic percentages (≥ 0.5 at %) based on averaged regional quantifications of XPS spectra.

The reduction method increased the silver-to-nitrogen ratio from 2:1 to nearly 5:1 on the sample surfaces. From the C 1s satellite peaks and the O 1s full width at half maximum

(FWHM) and atomic percentage of the reduced samples, we conclude that several bonding states involving oxygen are present. The peak locations are consistent with these bonding states which include, in addition to the nitrate-oxygen bond in nitrate ( ), an

O-O bond found in C-O-O, and a C-O bond, both of which are suitable moieties responsible for facilitating metal nanoparticle attachment to the otherwise nonreactive carbon nanotube surface.

5.1.2b X-Ray Photoelectron Spectroscopy Individual Scans

Notable features can be observed in the general scans of the control and reduced samples.

The small nitrogen peak observed in the reduced samples is larger in the control sample due to the nitrogen found in . The Ag 3d 5/2 peak is the most intense peak in the spectra of the control samples, but the C 1s peak is most intense in the reduced sample. The reason for this difference is that the control sample had 22 mg on the surface, while the reduced sample had only 2.4 mg before reduction and significantly less after. Because XPS is a superficial technique, this layer of silver nitrate shrouds most of the underlying carbon signal in the control samples, while the metallic particles on the reduced samples “bead”, exposing the carbon nanotube beneath and resulting in a higher C 1s peak.

Figure 5.1: A AgNO3 control sample general scan XPS spectrograph showing a small C 1s peak due to a layer of silver nitrate on the sample surface (A 60 °C reduction temperature sample shown).

This C 1s peak of the control sample was located at 284.3 eV before charge correction to 284.5 eV, a well established value for highly-oriented pyrolitic graphite

(HOPG) [64]. A silver nitrate coating is shown in Figure 5.1, while a reduced sample is shown in Figure 5.2. The main peak in this reduced sample corresponds to the C-C bonds of CNTs, which should appear between 284.4 and 285 eV [64 - 68]. The carbon measured in the silver nitrate coated sample can be attributed to substrate carbon and/or to contaminants from the atmosphere adsorbed onto the silver nitrate surface.

Figure 5.2: A reduced sample general scan XPS spectrograph showing a large C 1s peak due metallic nanoparticles beading on the surface revealing CNTs (sample 1 shown).

Figure 5.3: A composite spectrograph depicting XPS spectra collected in different locations of the AgNO3 control sample in the C 1s regime.

While the control sample displays only one prominent C 1s peak (Figure 5.3), secondary peaks can be observed in the region 1-7 eV higher than the C 1s peaks indicating C-C bonds in the reduced samples (Figure 5.4). These secondary peaks are consistent with common carbon bonding states. Three curves were used to fit these satellite peaks (Figure 5.5a). In addition, an XPS spectrograph of HOPG was also compared with those of the reduced samples and was found to be lacking the satellite peak near 286 eV (Figure 5.5b).

The satellite peaks, appearing at 285.8 eV and 290.3 eV, may correspond to C-O (~285.4 eV) bonds and C-O-O (~288.9 eV), respectively [68, 64]. There is a fourth possible peak near 288 eV which may correspond to C=O [64, 69]; however, it is extremely faint and would require more detailed experimental analysis to accurately evaluate.

Figure 5.4: A composite spectrograph collected from three reduced samples depicting XPS spectra in the C 1s regime.

Figure 5.5a: A spectrograph of a reduced sample fit with three curves lying at 284.5, 285.8, and 290.3 eV.

5/2

3/2

Figure 5.6: A composite spectrograph depicting three XPS spectra collected in different locations of the AgNO3 control sample in the Ag 3d regime.

The Ag 3d 5/2 peak on the control sample lies at 368.0 eV (Figure 5.), which agrees with the literature value for silver nitrate [67]. However, the silver peak locations of other potential silver-containing molecules, such as metallic silver and silver oxide, are also near this position. The AgNO3 peak lies between 368.2 eV and 368.8 eV [70, 71], while AgO is at 368 eV [67], and Ag2O is at 367.9 eV [72]. Metallic silver, silver nitrate, and both oxide forms of silver all fall within 0.3 eV of each other.

Data for the Ag 3d 3/2 peak of AgO and AgNO3 are not available, and the Ag2O 3/2 peak has been reported as 373.9 eV [72] (a separation of 6 ± 0.1 eV; the same as metallic silver

[73]), so this satellite peak is not helpful in identifying the bonding type.

For these reasons, the silver peaks can be used only to identify the amount of silver and not the bonding type.

5/2

3/2

Figure 5.7: A composite spectrograph collected from three reduced samples depicting XPS spectra in the Ag 3d regime.

The average Ag 3d 5/2 peak location for the three samples post-reduction was 368.2 eV

(Figure 5.7), a small shift relative to the control sample. Some authors argue that a lower-energy shift in 3d 5/2 peak location occurs during silver reduction to metallic silver

[71, 74], while a slight higher-energy shift over the course of the reduction of roughly

~0.2 eV has been observed in this work. However, this is a small shift (within the estimated error of 0.3 eV < x < 0.5 eV), so taken alone, it does not rule out silver reduction.

Silver nitrate has been reported within 1 eV of its currently-observed position and metallic silver has been reported higher than my observed location (368.3 and 368.4 eV)

[75, 76].

Figure 5.8: A composite spectrograph depicting three XPS spectra collected in different locations of the AgNO3 control sample in the N 1s regime.

The observed nitrogen peak positions of the AgNO3 control sample are centered at 406.1 eV (Figure 5.). However, the literature once again reports higher binding energies than are here observed – publications place this peak between 406.6 and 407.2 eV [70, 71].

Figure 5.9: A composite spectrograph collected from three reduced samples depicting XPS spectra in the N 1s regime.

The reduced samples (Figure 5.) exhibit a virtually identical N 1s location to the nitrate control (avg. 406.2 eV), but the peak is much weaker with respect to the strength of the silver peak. This indicates that most of the silver nitrate present in the AgNO3 control sample’s spectra has been reduced. 37

Figure 5.10: A composite spectrograph depicting three XPS spectra collected in different locations of the AgNO3 control sample in the O 1s regime.

The main O 1s peak of the AgNO3 control sample is located at 531.7 eV (Figure 5.10), while the O1s nitrate peak has been reported from 532.3 to 532.9 eV [70, 71]. The lower binding energy observed may indicate partially reduced silver nitrate on the surface, an interpretation which is also supported by the atomic ratios found in Table 5-1.

Figure 5.11: A composite spectrograph collected from three reduced samples depicting XPS spectra in the O 1s regime.

The mean observed O 1s peak in the reduced samples is located at 532.2 eV (Figure

5.11). The FWHM has increased from 1.5 to 2.0 eV from the AgNO3 control sample, indicating that more than one bonding state of oxygen is present after reduction.

It should be noted that the highest FWHM in Figure 5.11, pictured in green, is 2.3 eV, while the others are roughly 1.9. This wider peak is of the 80 °C reduction temperature sample.

The O 1s peaks of the oxides of silver, Ag2O and AgO, are reported to be between 528.4 and 530.9 eV [77 - 79]. This indicates that a silver oxide layer, if it exists, is a small 39 portion of the oxygen used – especially since 530.9 eV is one of the highest values reported, with more common values lying near 529 eV. This makes the presence of significant surface oxidation on the reduced samples unlikely because the observed peak lies primarily higher than this value.

However, the oxygen found in silver nitrate is located between 532.3 and 532.9 eV [70,

71], so the presence of unreduced AgNO3 is not ruled out on this basis.

Figure 5.12: A composite spectrograph depicting three XPS spectra collected in different locations of the AgNO3 control sample in the S 2p regime.

A contaminant, sulfur, was found in both types of samples. In the AgNO3 control sample, S 2p peaks were observed at 167.7 and 168.8 eV (Figure 5.12). However, the signal is too noisy to have a high level of confidence in the location of the 168.8 eV peak.

Figure 5.13: A composite spectrograph collected from three reduced samples depicting XPS spectra in the S 2p regime.

The reduced S 2p primary peak found on the reduced samples is located at approximately

162.9 eV (Figure 5.13). The origin of the sulfur and the change in its 2p binding energy observed here are not clearly understood.

5.2 Energy Dispersive X-Ray Spectroscopy

Energy Dispersive X-Ray Spectroscopy (EDX) was used for collecting qualitative as opposed to quantitative information. It was not calibrated specifically for this substrate or this particular elemental composition. EDX spectra were taken at 2000x magnification, 15 kV accelerating voltage, and a working distance of 8 mm.

These spectra show that a significant portion of silver has been attached to the surface, but they provide little clarification of the degree of reduction for reasons explained below

(Figure 5.14).

Figure 5.14: An EDX spectrum of a 60 °C sample after synthesis (top) and the accompanying quantification (bottom).

As is expected, the amount of silver observed has decreased with respect to carbon from the XPS spectra of the reduced samples, as EDX analysis penetrates much more deeply into the sample than does the alternative method. Unfortunately, the N Kα peak appears at approximately 0.4 keV, which lies in the low energy, higher-noise region of the spectrum. This value also lies between two large peaks of this sample; O K and C K.

Since the nitrogen amount indicated by XPS is small and in combination with these

43 factors, the nitrogen peak is obscured in this spectrum. This spectrum confirms that we have deposited silver.

5.3 Scanning Electron Microscopy

In SEM micrographs, the state of silver nanoparticles for both the 60 °C and 80 °C samples (the preparation of which is discussed in Chapter 4) can be observed. A 2 kV accelerating voltage was used with a 1.5 mm working distance.

Figure 5.15: A 60 °C reduced sample micrograph at 100kx magnification.

The 60 °C reduced samples contained Ag-NPs with a mean diameter of less than 5 nm and showed a high degree of dispersion (Figure 5.15). Conversely, the 80 °C reduced samples showed coalescence of silver particles, yielding larger particles (Figure 5.16) with agglomerated patches of coalesced particles as broad as several micrometers across

(Figure 5.17, Appendix C).

Closer inspection of the 80 °C reduced samples indicates that the coalesced particles are mostly distinct, with constituent particles mostly in the 20-40 nm range (Figure 5.16). It is not surprising that these particles have grown as a result of the application of heat, as metal nanoparticles are well known to ripen even at low temperatures [6].

Figure 5.16: An 80 °C reduced sample micrograph at 100kx magnification.

Figure 5.17: An 80 °C reduced sample micrograph at 10kx magnification showing patches of coalesced particles.

5.3.1 Scanning Electron Microscope Measurement Procedure

In order to gain a quantitative grasp of the effect of reduction temperature on particle sizes, measurements of the sizes and quantity of particles have been carried out using the following methodology: The number of particles visible in several (from three to five)

SEM micrographs of areas on a sample have been manually counted and measured using the Scandium© software application [Figure 5.19 - Figure 5.22].

Scanning electron microscopy was carried out at 2 kV in either backscatter or secondary electron mode at a magnification of 250kx. These images are captured at a working distance of 1.5 mm. The particles were clearly visible in both, and a side-by-side comparison followed by a particle count revealed no significant difference in the number or radii of the particles (data not included). The two were compared with one another in this analysis.

In order to perform this analysis, locations were chosen in an unbiased manner, i.e., the sample itself was imaged in low magnification mode (approximately 50x), followed by the random selection of an area to collect images for analysis. The specimen would be moved horizontally in low magnification mode to ensure that the analyzed images were sufficiently far apart to accurately represent the whole sample.

In the following analyses, two types of samples were evaluated: those reduced at 60 °C and those reduced at 80 °C. Three 60 °C samples and two 80 °C samples were evaluated in this manner. To depict these samples, a total of 18 micrographs from the 60 °C samples and 14 micrographs from the 80 °C samples were evaluated.

These measured values were then exported into a Microsoft Excel© file, where the data were analyzed. In Figure 5.23, a histogram depicting particle diameters, the error values were computed as follows: First, the diameter values of all the micrographs counted for a particular reduction temperature were rounded to the nearest whole number (in nm).

Then, the number of particles which fell into each of these 1 nm bins (e.g. 1.5 to 2.5 » 2 nm bin) were tallied on a per image basis (e.g. 15 of the 2 nm particles for image one were counted, 19 for image two, etc.) to determine how much variation was present

48 between locations. The last bin includes all particle diameters greater than 14.5 nm , and is labeled “>15 nm”. The mean value for each bin is also calculated; for example, 20 particles were counted at 2 nm in image #1, 10 in image #2, so the mean value is reported as 15 particles. The standard deviation of all the totals from each micrograph per bin is calculated using the standard (n – 1) formula:

σ =

This standard deviation is then used to determine the standard error of the mean formula, where n is the number of values and σ is the standard deviation:

In order to normalize the number of particles so that they could be compared between the micrographs of the heavily particle-laden 60 °C reduced samples and those of the relatively sparsely populated 80 °C reduced samples, this mean value is divided by the highest mean value reported (e.g. if most particles were in the 5 nm bin, which had an average of 30 particles per image, the 2 nm bin from the previous example will now read

15/30, or 0.5 in its normalized value). This process is carried out for each bin, such that the tallies of the most populous bin is 1 and all others are a fraction of it. The normalized standard error of the mean, as is reported in Figure 5.23, is divided also by the highest mean value for any bin, so that the ratio between mean counts and error remains constant before and after normalization.

In order to distinguish these particles from the iron particles deposited on the CNTs in the

CNT synthesis process, an identical analysis was performed of a sample that was generated in the same CVD batch without the addition of silver [Figure 5.18]. Six such micrographs were evaluated.

Figure 5.18: An SEM micrograph of unfunctionalized MWCNTs with measured iron particles at 250kx magnification.

Figure 5.19: As-collected micrograph of a 60 °C reduced sample showing silver nanoparticles at 250kx magnification.

Figure 5.20: An analyzed micrograph of a 60 °C reduced sample showing measured silver nanoparticles at 250kx.

Figure 5.21: An as-collected micrograph of an 80 °C reduced sample showing silver nanoparticles at 250kx.

Figure 5.22: An analyzed micrograph of an 80 °C reduced sample showing measured silver nanoparticles at 250kx.

5.3.2 Scanning Electron Microscopy Results

The bare sample which was investigated provided a baseline particle number. This sample was adorned with only the catalyst iron particles from the CNT synthesis. This evaluation revealed an average of approximately 20 particles per image at 250kx magnification (ppi) over the six locations investigated. The visible area at one location at this magnification is approximately 480 x 383.6 nm, or just over 180,000 nm2, which may be used in order to ascertain the number of visible particles per area.

In contrast, the ppi values of the 60 °C or 80 °C reduced samples were 330 and 190 ppi, respectively. As such, we can conclude that the vast majority of the particles observed in the micrographs of the reduced samples are silver particles and not particles of the iron catalyst.

Figure 5.23 shows the full histogram for these functionalized samples. 3.53 nm is the average diameter of the 60 °C reduced samples’ particles, while 7.45 nm is the average diameter of the 80 °C reduced samples’. The higher temperature has increased the diameter of the particles, as was expected because it has been noted that increased temperature has been observed to increase the number of nuclei formed as a function of time [6, 80], thereby extending the time spent in the ripening phase and, presumably, the rate of particle growth as well.

Figure 5.23: Histogram comparing reduction temperatures and resulting mean particle sizes.

It should be noted that a calculation of the amount of silver present based only on the ppi and diametric distributions is a considerable underestimate for two main reasons.

One reason is that nanoparticles smaller than 1 nm cannot be measured via traditional

SEM. There are many particles which are observed in both the 80 °C and 60 °C reduced samples which are near this size, suggesting that a significant number of these very small particles may go uncounted.

The second cause of error is due to shadowing effects which may cause some particles to go uncounted. There are three ways this occurs: First, the particles may be located on the side opposite the one visible in the SEM micrographs of a CNT, rendering them partially or wholly invisible because of shielding by the CNTs (see schematic in

Appendix A). Second, particles may go uncounted due to shielding by the other, larger silver particles above them (Figure 5.21). Third, the CNTs bearing Ag-NPs may be shielded by CNTs above them. The last shadowing effect may be especially significant because the substrate itself is completely invisible under the CNT “forest,” indicating that a large number of CNTs and therefore the Ag-NPs which adorn them are hidden from view.

An alternative method for determining particle abundance is as follows. The number of particles present in the average area may be found by

Where is the number of attachment sites per area, i.e., the number of places the

CNTs are attached to the substrate per area, is the average CNT length, and ppl is the average number of particles per length. In order to estimate the number of particles per length, further SEM analysis was performed. Two to three top-layer CNTs per micrograph were selected and the particles on their surfaces were counted. This number was then divided by the length of the CNTs to determine ppl.

By manually counting particles in six micrographs of three 60 °C samples, an average ppl of 65 ± 5 particles/ was ascertained. Both the number of attachment sites per area and the average CNT length have been conservatively estimated as 170 CNT/ and 20 55 for CNTs grown on this substrate and in this manner [1]. Therefore, the ppa of 60 °C reduced samples may be conservatively estimated as 220,000 ± 17,000 particles/ .

Estimating the number of particles per substrate area in this way relies on one important assumption: The concentration of particles on CNTs is independent of the depth of the

CNTs, i.e., the CNTs located nearer to the substrate have a similar concentration of Ag-

NPs as do the CNTs further from it.

Chapter 6. Sonication-Based Analysis

6.1 Sonication-Based Analysis Procedure

A 60 °C reduced sample was placed in a conical, graduated test tube containing 10 mL of distilled water. The test tube was held upright in a 5.5 gallon ultrasonic cleaner as is shown in Figure 6.1. The ultrasonic cleaner was filled to roughly 1 ½ inches below the top of the tank, which was below the top cusp of the test tube. The sample was sonicated for a set duration, measured by a manually-operated stopwatch. The sample was then removed, sprayed with ethanol, and allowed to dry.

Figure 6.1: Schematic of the top and side views, respectively, of the sonication- based analysis experimental setup.

6.2 Sonication-Based Analysis Results/Discussion

The aim in this sonication-based analysis was to test the durability of the hybrid structure by causing sample failure via mechanical oscillation, i.e., to irreversibly change the hybrid structure in a manner which would significantly compromise its utility in applications such as biosensation or biofiltration. To that end, a 60 °C reduced sample

58 was sonicated for five minutes, as outlined in the introduction to this chapter. A micrograph of the resulting substrate is presented in Figure 6.2.

Figure 6.2: A SEM micrograph of a CNT sample after five minutes of sonication.

The flat, white portions visible in Figure 6.2 are patches of silica that have remained intact on the graphite surface. The “valleys” observed are pits in the graphite substrate.

This heterogeneous response to sonication is likely due to the varying graphitic planar orientations that are known to lead to anisotropy in HOPG. This sonication was observed

59 to remove all CNTs, which provides little information about how well the Ag-NPs adhered to the CNTs.

An identical 60 °C reduced sample was tested using a much shorter sonication time, 10 seconds. This was found to provide a snapshot of the early stages of sample failure under this type of rapidly oscillating mechanical force. The results are shown in Figure 6.3.

Figure 6.3: A SEM micrograph of a 60 °C reduced sample after ten seconds of sonication.

Figure 6.4: A SEM micrograph of a 60 °C reduced sample after ten seconds of sonication.

Figure 6.5: A SEM micrograph depicting a sonicated 60 °C reduced sample showing an intact silica layer (left) and the underlying graphite (right).

Five micrographs were taken, in a manner similar to that of those taken to compare 60 °C and 80 °C reduced samples in Section 5.3.1, of the areas that remained covered with

CNTs. The only procedural difference between the previous study and the sonication examination is that, in the latter, low magnification mode was used to avoid areas which were rendered bare by sonication.

Figure 6.6: SEM micrographs of a 60 °C reduced sample after a ten-second sonication.

Based on these five micrographs of sonicated samples, 332 ± 41 ppi were counted, while an average of 328 ± 35 ppi were observed on unsonicated samples. This result demonstrates that an appreciable amount of silver was not removed via mechanical oscillation, though an appreciable number of CNTs were removed.

In summary, silver remained attached to the CNTs at least until CNTs were removed from the surface under an oscillating mechanical force. However, this does not imply that the Ag-CNT atomic bonds are stronger than those binding the CNTs to the substrate.

For geometric reasons, the force exerted on a particular CNT may be much greater than that exerted on a silver particle that has a much smaller surface area (Figure 6.7). This study demonstrates that a robust hybrid structure has been synthesized which has particles that will not easily separate from their substrate.

Figure 6.7: A schematic representing forces exerted by sonication on CNTs and Ag- NPs during a single pulse. Given the attachment area of a CNT as A and an attachment area of a particle as a, F/A >> f/a.

Chapter 7. Biocompatibility Assessment

7.1 Fluorescence-Based Biocompatibility Analysis Procedure

Hybrid structure samples were evaluated in vitro for compatibility with keratinocytes, which are the most common cell type found in the epidermis. The HaCaT cell line, an immortal cell line of keratinocytes, was used in our biocompatibility studies. This cell line was chosen due to its retention of key in vivo properties despite long-term differentiation. For example, after long-term passaging, differentiation-specific keratins are expressed, unchanging DNA fingerprint patterns are observed, and the cells have been found to be nontumorgenic [81].

These HaCaT cells were cultured in the following manner: First, the cells were thawed from their frozen state (cryogenic vials containing roughly 106 cells, and 1 mL of solution containing 72% growth medium, 20% fetal bovine serum (FBS), and 8% DMSO, immersed in liquid nitrogen) by immediately immersing the vials in a 37° C water bath.

When the vial had completely thawed, the exterior of the vial was sterilized by spraying with ethanol. Then, the contents were poured into a sterile conical tube. 30 mL of complete growth medium (CGM) were added into the tube slowly, so as to reduce osmotic shock. The cells were then pipetted to evenly disperse and seed them into a flask, in which they were then placed into a 37° C environment with 5% CO2 and 100%

66 humidity. The medium was replaced the following day to provide additional nourishment.

The cells were then passaged several (more than three) times in the following manner:

Roughly four days after passaging, when the cells were 80% to 100% confluent, the medium was aspirated. 10 mL of phosphate-buffered saline solution (PBS) were added to remove any remaining trypsin-inhibiting solution in addition to any non-adhered cells.

The PBS was aspirated, and then 3 mL of 0.25% trypsin were added, followed by incubation in the typical HaCaT growth environment for ten minutes. When all cells had released from the flask, 10 mL CGM was added and the resulting solution was pipetted to disperse the cells. The cells were then resuspended in fresh growth medium and in a new flask at a 1:10 ratio. The process was repeated until the cells are needed.

In each iteration of this experiment, three samples were tested for HaCaT toxicity: a portion of a bare, sterile glass slide (control); MWCNTs on a graphitic substrate, as outlined previously; and silver nanoparticles (Ag-NPs) adhered to the previous sample, as outlined in Chapter 4.3. A 24-well plate and the samples were irradiated for 30 minutes with a 75 Watt, 254 nm wavelength Phillips UV lamp positioned approximately two feet away to sterilize the samples. The samples were placed in separate wells and were rinsed with CGM to remove any loose particulates which had collected on the surface during functionalization. 2 mL of fresh CGM were then added to the samples in the well plate, completely immersing them, and 125,000 cells per mL of CGM were added to each well.

The cells were allowed to grow for 2 days at 37 °C in a 5% , 100% humidity environment until the control slide was roughly 100% confluent with HaCaT cells.

The samples were labeled with fluorescent markers to make cell morphologies visible and to determine their viability, according to the following procedure: The CGM was aspirated from each well. Each sample was then washed twice with room temperature

PBS, which was then aspirated. Each well was fixed with a solution comprised of 4% paraformaldehyde and 96% PBS for ten minutes. The PBS washing step was repeated.

The cells on each sample were then permeabilized for five minutes with a solution comprised of 0.1% Triton X-100 in a PBS solvent by pipetting the solution into each well, followed by aspiration. The wells were again rinsed twice with PBS, followed by adding a solution comprised of 10 mg BSA powder dissolved in 1 mL PBS to each well, where it remained for 30 minutes until aspiration. For the remainder of the procedure, a light controlled (red-lamp only) environment was used to avoid photobleaching of the fluorescent markers. Alexa Fluor 555 phalloidin was added to the BSA solution prepared in the previous step in the ratio of 1:40 by volume – once the cells are permeabilized, phalloidin was able to bypass the cell membrane. This solution was well-mixed by vortexing, then 1 mL was added to each well for 20 minutes, followed by aspiration.

Each well was again aspirated and the samples were then twice rinsed with PBS. The samples were air dried for one to two minutes, followed by the application of several drops of Prolong Gold reagent with Dapi counterstain to each sample.

The role of these components is as follows: The Prolong Gold reagent counteracts the effects of photobleaching, and the Dapi counterstain emits a blue fluorescent light when bound to a nucleus. Each sample was then placed upside down an unused well. The cells were then allowed to stain and cure in darkness (covered by aluminum foil) overnight and viewed the following day.

The cells were then imaged using an Olympus IX71 Inverted Microscope. The lights were dimmed, and each sample was viewed using the TRITC and DAPI filters. These filters are designed to narrow the range of wavelengths which impinge on the sample and which reach the objective lens; they are known as excitation filters and barrier filters, respectively. The former is necessary to avoid unnecessary photobleaching of the sample, while the latter ensures that only fluorescent light is reaching the objective, since other light sources (such as reflected excitation light) may obfuscate cell information.

These images were digitally collected, which was immediately followed by switching the light filter to collect an image of the same location to observe the other fluorescent stain.

Each sample was observed in at least three locations on the samples’ surface, and typical images are shown here of the control and CNT samples (Figures 7.1, 7.2). However, the areas imaged of the silver samples were selectively chosen so that images were possible, as will be explained in the results section. The images of the nuclear and cytoplasmic stains were collected separately and overlaid in the composite images shown. The colors of each were digitally modified to increase the contrast between them.

7.2 Biocompatibility Results/Discussion

Figure 7.1: A micrograph of fluorescent HaCaT cells on a glass substrate after two days’ growth.

The negative control in this experiment was glass, which is known to be biocompatible.

The glass was sprayed with isopropyl alcohol, allowed to dry, and sterilized under a UV

70 lamp for roughly ten minutes at a distance of two feet. The fluorescence microscopy shows healthy, confluent cells after two days’ growth, as expected (Figure 7.1).

Figure 7.2: A micrograph of fluorescent HaCaT cells on HOPG + CNTs after two days’ growth.

Adherent MWCNTs on graphite have been evaluated using this method also. After the same time period as the control (two days), islands of viable cells resulted with some cells showing morphological signs of unhealth. Specifically, some nuclei were ill- defined, suggesting they were in the process of cell death. However, much of the area was covered with apparently healthy cells (Figure 7.2). Previous research suggests that we may have expected full, but delayed confluence of those cells, but this hypothesis was not tested herein.

Figure 7.3: A micrograph of fluorescent HaCaT cells on HOPG + CNTs + Ag NPs after two days’ growth.

The cells introduced to the final hybrid structure of silver nanoparticle-decorated carbon nanotubes showed the most toxicity. There were very few cells left which had not degraded beyond exhibiting fluorescence – in fact, while the control and CNT samples were imaged “randomly,” this image had to be sought after, since the vast majority of the

73 sample appeared completely dark. In both trials of this experiment, there were fewer than ten fluorescent cells per 0.25 sample – approximately 100% toxicity. The few images which were obtained revealed cells in the process of death based on their morphologies and weak fluorescence (this image was much dimmer than the others before image processing was performed) (Figure 7.3).

There are likely several contributing factors to these apparent toxicity findings. One likely contributor was the small size of the silver particles that adorned the CNTs, coupled with the large number of particles present. These Ag-NPs were primarily less than 8 nm in diameter, which likely contributed to the toxicity because there is a significant negative correlation between particle size and toxicity on a per-mass basis, as discussed in Chapter 3. It is also important to note that silver nitrate is much more toxic than metallic silver, so its presence, even if slight, may have been an important contributing factor to the toxicity of the samples. Both of these factors likely worked in concert to produce the observed toxicity, which was conclusive and apparent in both trials performed.

The mechanism of cell death cannot be determined via this method and there is significant debate in the literature as to the mechanisms of nanoparticle-induced apoptosis or necrosis.

One potential flaw in our study is that it is possible that the fluorescent stains are neutralized by silver or a contaminant which is present in the final hybrid structure but not in the previous ones. We cannot rule this out because the exact composition of the final sample is uncertain. This possible effect could mask viable cells, although that

74 seems unlikely due to the past reported toxicity of silver nanoparticles in smaller amounts, coupled with the fact that very few, but some, cells did exhibit fluorescence.

Chapter 8. Summary and Conclusions

Successful attachment of silver nanoparticles on surface attached nanotubes has been accomplished. It is accepted in the literature that surface activation of CNTs is required – every other electroless deposition in the literature, to our knowledge, reportedly relies on acid-activated nanotubes (i.e. containing a large number of side groups on the nanotube surface, such as –COOH), via an acid pre-etch) [7 – 9, 11, 28]. However, the current process does not utilize an acid pre-etch, thereby avoiding damage to the CNT walls, whose integrity are important to various material properties such as thermal and electrical conductivity as well as mechanical strength.

Additionally, to my knowledge, this is the first successful adherence Ag-NPs to the surface of CNTs that are strongly adhered to a larger substrate via electroless deposition.

Such structures can be used for future devices where the benefits of high surface area

CNT is combined with the robustness of larger solid structures.

The Ag-NPs that have been adhered can be somewhat altered in size as desired, from particles smaller than 6 nm in diameter to larger particles of 20-60 nm. They do not wash away with vigorous rinsing in ethanol or water -- or even with sonication until hybrid structure failure via other mechanisms. The very dense CNT-adhered Ag-NPs produced are not biocompatible in the 60 °C reduced form with human skin cells in vitro. Future

76 investigations are needed to determine if and how the toxicity level depends upon concentration, distribution, and surface properties of silver particles.

Chapter 9. Future Recommendations

Possible biological applications of a high surface area, surface plasmon resonant, antimicrobial nanocomposites include three types of devices: Biosensors, biofilters, and wound dressings. Each of these branches will be considered below.

9.1 Biosensor

An electrochemical biosensor is defined by Thévenot, et al. as, “A self-contained integrated device, which is capable of providing specific quantitative or semi-quantitative analytical information using a biological recognition element (biochemical receptor) which is retained in direct spatial contact with an electrochemical transduction element

[82].” Biosensing is a growing area of silver nanoparticle (and other surface plasmon resonant materials) application.

The extremely high surface area of my particles could provide an attractive substrate for biosensing, since SPR is a surface-dependent phenomenon. This coupling of a hierarchical substrate and a ligand specific to the analyte would be a prime candidate for a highly sensitive electrochemical biosensor.

9.2 Biofilter

Biofiltration is the passing of a fluid through some sort of barrier, blocking some or all biological agents from passing through or neutralizing those agents, while allowing the bulk of the fluid through. Constituting an omnipresent threat throughout human history, waterborne diseases are responsible for millions of deaths annually according to the

World Health Organization [83]. In addition to other steps such as improved hygiene practices and nutrition intake, portable, reusable, storable filtration devices are needed in order to combat this threat.

Biofiltration via passing contaminated liquid over a silver-NP functionalized substrate has been shown to be a viable method of sterilizing water [25], and is likely limited by the surface area of the silver per volume. Given that my hybrid structure boasts a large amount of Ag-NPs, and that carbon foams are readily functionalized by the CNTs I have used in this study, this seems to be a promising application.

A researcher could generate CNTs on commercially available, interconnected, highly porous carbon foams by reported methods [1]. This would then be followed by either force or gravity-driven throughput of contaminated water, depending on pore diameters, volume processing speed requirements, and surface activity, among other factors. Then,

Kirby-Bauer diffusion tests for measuring bactericide efficacy could be used to determine its efficacy.

9.3 Wound Dressing

A barrier dressing is a material which is placed over a wound or surgical graft site with the goal (in the case of the latter) being, “to promote autograft survival by providing an antimicrobial barrier and a local environment that supports adherence and eventual healing [84].” There are several commercial products which currently appeal to this role that use silver deposited on a substrate. One of the more widely tested is known as

ActicoatTM, a hybrid structure used to treat burn victims [44, 85]. However, in a broader sense than only barrier dressings, silver compounds have been used to prevent or limit infection for centuries [86, 87].

Because of the size effects noted by other researchers, this material may still be suitable for such an application – that is, perhaps larger particle sizes such as those seen in the 80

°C samples presented herein or even larger would be more biocompatible, and yet still bactericidal.

Additionally, the samples were shown herein to contain some impurities – nitrates particularly – which could harm the cells. Additional purification may further mediate the deleterious effects of the hybrid structure on human cells (by dissolution of additional nitrate, for instance), while still effectively inducing apoptosis in bacteria. Lastly, it is important to remember that silver has a history of proving to be more toxic in simpler systems – bacteria more than eukaryotic cells and in vitro more than in vivo.

Appendix

A. Wetting Angle Analysis

An analysis of the wetting angle of the particles was performed in order to gain some insight into the specifics of the nanoparticle bonding (Figure A.1). The procedure was very similar to previous SEM imaging: A location on the sample was chosen indiscriminately and four 250 kx micrographs were captured. These were then examined using an Angle tool in Scandium as is shown in (Figure A.2). Both visible edges of the particle were evaluated and averaged to determine the wetting angle.

A total of 9 particles were measured for wetting angle. The reason for this small size is that particles had to both be very near the edge of the CNT (as seen by the incident SEM electron beam) so that it could be viewed in profile and the particle had to be large enough to get a reliable measurement.

One criticism of this method is that it is difficult to ensure that a particle which appears to be lying on the edge is not in fact lying slightly behind it (Figure A.3). While this is certainly a possible source of error, it does not compromise the results in a significant way for two reasons: 1) No NPs of this size (< 20 nm) have been observed with a wetting angle over 90°, despite examining dozens of images. 2) Particles are readily found where CNTs border one another, suggesting that these interfaces are

81 thermodynamically favorable and that our apparent wetting angle is close to the actual wetting angle; that is, less than 90°.

Figure A.1: A schematic of one wetting angle measurement.

Figure A.2: A sample micrograph depicting wetting angle measurements.

Figure A.3: A schematic depicting one challenge in wetting angle measurement.

The wetting angles measured are less than 90°, suggesting a wettable surface. After a review of the literature, these less-than-90° wetting angles have been observed in other silver-CNT work as well [7, 28]. There are two commonalities between these cases: that the CNTs are functionalized with moieties as explained in the literature review and that each procedure utilizes surfactants of one kind or another. What is actually being observed in this case may not be that silver wets CNTs, but rather that moieties, surfactants, or both create a more energetically favorable surface for silver wetting. As far as the surfactant is a contributor, it may simply facilitate the initial nucleation or adherence of the silver to the moieties by removing tiny air pockets, after which time it may be irrelevant. The purpose of these functional groups is to create a low-energy bonding form for silver instead of silver remaining as spherical particles in solution. It is this lower-energy interface that also creates wettability. In the absence of such functional groups, high surface tension materials do not wet the surface of CNTs independently of the application of high pressures or chemically reactive environments [88].

B. Sample of Background Choices of XPS Spectra

In this section, example boundary choices are presented which were used for the quantification in Table 5.1 because the choices of these boundaries have a significant impact on the quantification results.

In each case, either Shirley or Linear fits were used. The former was used when the background noise level changed substantially before and after a peak and the scan had a high signal-to-noise ratio, while the latter was used if both conditions were not met.

C. 80° Reduced Sample 10,000x Magnified Micrograph

Figure C.1: A micrograph of an 80 °C reduced sample displaying a large patch of agglomerated silver particles.

D. Chemicals/Equipment

Chemicals:

Alexa-fluor 555 Phalloidin (Life Technologies, Inc.)

Bovine Serum Albumin (BSA) (Thermo Fisher Scientific Inc.)

Complete Growth Medium (CGM) (RPMI-1640 from ATCC supplemented with 1% Pen

Strep from Fisher Scientific Inc. and 10% BCS from Invitrogen, Inc.)

Dimethyl Sulfoxide, (DMSO) (CH3)2SO (Sigma-Aldrich Co.)

Fetal Bovine Serum (FBS)

HaCaT cells

Highly Oriented Pyrolitic Graphite (HOPG) (Koppers, Inc.)

Hyclone Bovine Calf Syrum (BCS) (Thermo Fisher Scientific Inc.)

Paraformaldehyde (Electron Microscopy Sciences)

Penicillin-Streptomycin Solution (Pen-Strep) (Thermo Fisher Scientific Inc.)

Phosphate-Buffered Saline Solution (PBS) (Life Technologies, Inc.)

ProLong Gold antifade reagent with DAPI (ProLong Gold) (Life Technologies, Inc.)

RPMI-1640 with L-Glutamine (RPMI) (Media Tech, Inc.)

Silver Nitrate, AgNO3 (Sigma-Aldrich Co.)

Trisodium Citrate Dihidrate (NaCit), C6H5Na3O7∙2H2O (Sigma-Aldrich Co.)

Triton X-100 Detergent Solution (Triton) (Thermo Fisher Scientific Inc.)

Trypsin with EDTA (Trypsin) (Life Technologies, Inc.)

Equipment:

 SIP Brand Ultrasonic Cleaner 5.5 gal.

Scientific Products Division of Baxter Diagnostics Inc.

Input: 115 V, 200 W, 2 A, 50/60 cycles

Cat. C6450-55, Model ME 5.5 S, Serial # 104B6083

 JSM-7401F FE SEM

 Kratos AXIS ULTRA (Fig 4.1)

Aluminum monochromatic anode K-alpha

 Thermolyne Nuova 2 Plate Stirrer

Model SP18425

 Polygon Spinbar with Ring

12 x 4.5 mm, Teflon

Figure D.1: Kratos X-Ray Photoelectron Spectrometer

References

1. Mukhopadhyay, Sharmila M; Karumuri, Anil;Barney, Ian T.. “Hierarchical nanostructures by nanotube grafting on porous cellular surfaces.” Journal of Physics: D Applied Physics, v. 42 issue 19, 2009, p. 195503-195503.

2. Liz-Marzán LM. “Tailoring surface plasmons through the morphology and assembly of metal nanoparticles.” Langmuir: The ACS Journal Of Surfaces And Colloids, v. 22 issue 1, 2006, p. 32-41.

3. Shalaev, V.M. Nanophotonics with Surface Plasmons. 2007 Elsevier.

4. Otto, A. “Exitation of nonradiative surface plasma waves in silver by the method of frustrated total reflection.” Z Phys 216, 1968, p. 398-410.

5. Kretschmann, E.;Raether H.. “Radiative decay of non radiative surface plasmons excited by light.” Z. Naturforsch, Teil A, 23, 1968, p. 2135-2136.

6. ZHU, LIYING;LU, GANHUA;MAO, SHUN;CHEN, JUNHONG;DIKIN, DMITRIY A.;CHEN, XINQI;RUOFF, RODNEY S.. "Ripening of silver nanoparticles on carbon nanotubes." NANO, v. 2 no. 3, 2007, p. 149-156.

7. Yan, Dongsheng;Wang, Feng;Zhao, Yongbin;Liu, Jingjun;Wang, Jianjun;Zhang, Lianghu;Park, Ki Chul;Endo, Morinobu. “Production of a high dispersion of silver nanoparticles on surface-functionalized multi-walled carbon nanotubes using an electrostatic technique.” Materials Letters, v. 63 issue 2, 2009, p. 171-173.

8. Byeon, Jeong Hoon;Hwang, Jungho. “Morphology of metallic nanoparticles as a function of deposition time in electroless deposition of metal on multi- walled carbon nanotubes.” Surface and Coatings Technology, v. 203 issue 3-4, 2008, p. 357-363.

9. Wang, Feng;Arai, Susumu;Park, Ki Chul;Takeuchi, Kenji;Kim, Yong Jung;Endo, Morinobu. “Synthesis of carbon nanotube-supported nickel–phosphorus

nanoparticles by an electroless process.” Carbon, v. 44 issue 7, 2006, p. 1307- 1310.

10. Rodríguez-Gattorno, G.;Díaz, D.;Rendón, L.;Hernández-Segura, G. O.. “Metallic nanoparticles from spontaneous reduction of silver(I) in DMSO. Interaction between nitric oxide and silver nanoparticles.” J. Phys. Chem. B, v 106 issue 10, 2002, p. 2482–2487.

11. Fam, D.W.H.;Tok, A.I.Y.;Palaniappan, Al.;Nopphawan, P.;Lohani, Anup;Mhaisalkar, S.G.. “Selective sensing of hydrogen sulphide using silver nanoparticle decorated carbon nanotubes.” Sensors and Actuators, B: Chemical, v. 138 issue 1, 2009, p. 189-192.

12. Wang, T.;Kaempgen, M.;Nopphawan, P.;Wee, G.;Mhaisalkar, S.;Srinivasan, M.. “Silver nanoparticle-decorated carbon nanotubes as bifunctional gas- diffusion electrodes for zinc–air batteries.” Journal of Power Sources, v. 195 issue 13, 2010, p. 4350-4355.

13. Lago, R. M.;Tsang, S. C.;Lu, K. L.;Chen, K.;Green, M. L. H.. "Filling carbon nanotubes with small palladium metal crystallites: the effect of surface acid groups." J. Chem. Soc. issue 13, 1995 p. 1355-1356.

14. Alonso, J.C.;Diamant, R.;Castillo, P.;Acosta–García, M.C.;Batina, N.;Haro- Poniatowski, E.. “Thin films of silver nanoparticles deposited in vacuum by pulsed laser ablation using a YAG:Nd laser.” Applied Surface Science, v. 255 issue 9, 2009, p. 4933-4937.

15. Giorgetti, E;Giusti, A.;Arias, E.;Moggio, I.;Ledezma, A.;Romero, J.;Saba, M.;Quochi, F.;Marceddu, M.;Gocalinska, A.;Mura, A.;Bongiovanni, G.. “In situ production of polymer-capped silver nanoparticles for optical biosensing.” Macromolecular Symposia, v. 283-284 issue 1, 2009, p. 167-173.

16. Li, J.;Xu, J.;Dai, W.;Fan, K.. “Dependence of Ag deposition methods on the photocatalytic activity and surface state of TiO2 with twistlike helix

structure.” Journal of Physical Chemistry C, v. 113 issue 19, 2009, p. 8343- 8349.

17. Henglein, A.;Giersig, M.. “Formation of colloidal silver nanoparticles: capping action of citrate.” J. Phys. Chem. B, v. 103 issue 44, 1999, p. 9533-9539.

18. Turkevich, J.;Stevenson, P. C.;Hillier, J.. "A study of the nucleation and growth processes in the synthesis of colloidal gold", Discuss. Faraday. Soc. v. 11 1951, p. 55-75.

19. Hauser and Lynn, Experiments in Colloid Chemistry (McGraw Hill, 1g40), p. 18.

20. Serp, P.;Corrias, M.;Kalck, P.. “Carbon nanotubes and nanofibers in catalysis.” Applied Catalysis A General, v. 253 issue 2, 2003, p. 337-358.

21. Kyotani, T;Nakazaki, S.;Xu, W.;Tomita, A.. “Chemical modification of the inner walls of carbon nanotubes by HNO3 oxidation.” Carbon, v. 39 issue 5, 2001, p. 782-785.

22. Ebbesen, T.;Hiura, H.;Bisher, M.;Treacy, M.;Shreeve-Keyer, J.;Haushalter, R.. “Decoration of carbon nanotubes.” Advance Materials, v. 8 issue 2, 1996, p. 155-157.

23. Lee, P.;Meisel, D.. “Adsorption and surface-enhanced raman of dyes on silver and gold sols.” J. Phys. Chem. v. 86 issue 17, 1982, p. 3391-3395.

24. Kamat, P.;Flumiani, M.;Hartland, G.. “Picosecond dynamics of silver nanoclusters. Photoinjection of electrons and fragmentation.” J. Phys. Chem. B, v. 102 issue 17, 1998, p. 3123-3128.

25. Jain, P.;Pradeep, T.. “Potential of silver nanoparticle-coated polyurethane foam as an antibacterial water filter.” Biotechnology and Bioengineering, v. 90 issue 1, 2005, p. 59-63.

26. Shirtcliffe, N. “Reproducible preparation of silver sols with small particle size using borohydride reduction: for use as nuclei for preparation of larger particles.” J. Colloid Interface Sci. v. 211 issue 1, 1999, p. 122–129.

27. Krasimir Vasilev;Vasu R Sah;Renee V Goreham;Chi Ndi;Robert D Short;Hans J Griesser. “Antibacterial surfaces by adsorptive binding of polyvinyl- sulphonate- stabilized silver nanoparticles.” Nanotechnology, v. 21 issue 21, 2010, p. 215102.

28. Cveticanin, Jelena;Krkljes, Aleksandra;Kacarevic-Popovic, Zorica;Mitric, Miodrag;Rakocevic, Zlatko;Trpkov, Djordje;Neskovic, Olivera. “Functionalization of carbon nanotubes with silver clusters.” Applied Surface Science, v. 256 issue 23, 2010, p. 7048-7055.

29. Raghuveer, M.;Agarwal, S.;Bishop, N.;Ramanath, G.. “Microwave-assisted single-step functionalization and in situ derivization of carbon nanotubes with gold nanoparticles.” Chem. Mater. v. 18 issue 6, 2006, p. 1390-1393.

30. Cavicchioli, Maurício;Varanda, Laudemir C.;Massabni, Antonio C.;Melnikov, Petr. “Silver nanoparticles synthesized by thermal reduction of a silver(I)– aspartame complex in inert atmosphere.” Materials Letters, v. 59 issue 28, 2005, p. 3585-3589.

31. Varanda, L. C.;Jafelicci, M.;Tartaj, P.;O'Grady, K.;González-Carreño, T.;Morales, M. P.;Muñoz, T.;Serna, C. J.. “Structural and magnetic transformation of monodispersed iron oxide particles in a reducing atmosphere”, Journal of Applied Physics, v. 92 issue 4, p. 2079, Academic Search Complete, EBSCOhost, viewed 21 February 2011.

32. Amoruso, S.;Ausanio, G.;Bruzzese, R.;Vitiello, M.;Wang, X “Femtosecond laser pulse irradiation of solid targets as a general route to nanoparticle formation in a vacuum.” Physical Review, v. 71, 2005, p. 33406-1-33406-4.

33. Geohegan, David B.;Puretzky, Alex A.;Duscher, Gerd;Pennycook, Stephen J.. “Time-resolved imaging of gas phase nanoparticles synthesis by laser ablation.” Applied Physics Letters, v. 72 issue 23, 1998, p. 2987-2989.

34. Iijima, S. “Helical Microtubules of Carbon.” Nature, 354, 1991, p. 56-58.

35. Shvedova, Anna A.;Castranova, Vincent;Kisin, Elena R.;Schwegler-Berry, Diane;Murray, Ashley R.;Gandelsman, Vadim Z.;Maynard, Andrew;Baron, Paul. “Exposure to Carbon Nanotube Material: Assessment of Nanotube Cytotoxicity using Human Keratinocyte Cells.” Journal of Toxicology & Environmental Health: Part A, v. 66 issue 20, 2003, p. 1909-1926.

36. Monteiro-Riviere, Nancy A.;Nemanich, Robert J.;Inman, Alfred O.;Wang, Yunyu Y.;Riviere, Jim E.. “Multi-walled carbon nanotube interactions with human epidermal keratinocytes.” Toxicology Letters, v. 155 issue 3, 2005, p. 377-384.

37. Zhang, Leshuai W.;Liling Zeng;Barron, Andrew R.;Monteiro-Riviere, Nancy A.. “Biological Interactions of Functionalized Single-Wall Carbon Nanotubes in Human Epidermal Keratinocytes.” International Journal of Toxicology (Taylor & Francis), v. 26 issue 2, 2007, p. 103-113.

38. Correa-Duarte;M. A.;Wagner;N.;Rojas- Chapana;J.;Morsczeck;C.;Thie;M.;Giersig. “Fabrication and Biocompatibility of Carbon Nanotube-Based 3D Networks as Scaffolds for Cell Seeding and Growth.” Nano Letters, v. 4 issue 11, 2004, p. 2233-2236.

39. Schrand, Amanda M.;Dai, Liming;Schlager, John J.;Hussain, Saber M.;Osawa, Eiji. “Differential biocompatibility of carbon nanotubes and nanodiamonds.” Diamond & Related Materials, v. 16 issue 12, 2007, p. 2118-2123.

40. Seoktae Kang;Moshe Herzberg;Debora F. Rodrigues;Menachem Elimelech. “Antibacterial effects of carbon nanotubes: size does matter!.” Langmuir, v. 24 issue 13, 2008, p. 6409-6413.

41. Grabinski, Christin;Hussain, Saber;Lafdi, Khalid;Braydich-Stolle, Laura;Schlager, John. “Effect of particle dimension on biocompatibility of carbon nanomaterials.” Carbon, v. 45 issue 14, 2007, p. 2828-2835.

42. Kumar, Ashavani;Vemula, Praveen Kumar;Ajayan, Pulickel M.;John, George. “Silver-nanoparticle-embedded antimicrobial paints based on vegetable oil.” Nature Materials, v. 7 issue 3, 2008, p. 236-241.

43. Kassler, Jackie;Barnett, Josh. “A Rehabilitation Hospital's Experience with Ionic Silver Foley Catheters.” Urologic Nursing, v. 28 issue 2, 2008, p. 97-100.

44. Huang, Yuesheng;Li, Xiaolu;Liao, Zhenjiang;Zhang, Guoan;Liu, Qun;Tang, Jin;Peng, Yizhi;Liu, Xuesheng;Luo, Qizhi. “A randomized comparative trial between Acticoat and SD-Ag in the treatment of residual burn wounds, including safety analysis.” Burns (03054179), v. 33 issue 2, 2007, p. 161-166.

45. Lu, Wentong;Senapati, Dulal;Wang, Shuguang;Tovmachenko, Oleg;Singh, Anant Kumar;Yu, Hongtao;Ray, Paresh Chandra. “Effect of surface coating on the toxicity of silver nanomaterials on human skin keratinocytes.” Chemical Physics Letters, v. 487 issue 1-3, 2010, p. 92-96.

46. Jun Ho Ji;Jae Hee Jung;Sang Soo Kim;Jin-Uk Yoon;Jung Duck Park;Byung Sun Choi;Yong Hyun Chung;Il Hoon Kwon;Jayoung Jeong;Beom Seok Han;Jae Hyeg Shin;Jae Hyuck Sung;Kyung Seuk Song;Il Je Yu. “Twenty-eight-day inhalation toxicity study of silver nanoparticles in Sprague-Dawley rats.” Inhalation Toxicology, v. 19 issue 10, 2007, p. 857-71.

47. J. M. Wörle-Knirsch;K. Pulskamp;H. F. Krug. “Oops They Did It Again! Carbon Nanotubes Hoax Scientists in Viability Assays.” Nano Letters, v. 6 issue 6, 2006, p. 1261-1268.

48. Chun, Ja Young;Kang, Hyun Ki;Jeong, Lim;Kang, Yun Ok;Oh, Ju-Eun;Yeo, In- Sung;Jung, Sung Youn;Park, Won Ho;Min, Byung-Moo. “Epidermal cellular response to poly(vinyl alcohol) nanofibers containing silver nanoparticles.” Colloids and Surfaces B Biointerfaces, v. 78 issue 2, 2010, p. 334-342.

49. Rujitanaroj, Pim-on;Pimpha, Nuttaporn;Supaphol, Pitt. “Wound-dressing materials with antibacterial activity from electrospun gelatin fiber mats containing silver nanoparticles.” Polymer, v. 49 issue 21, 2008, p. 4723-4732.

50. Lok, C. M.;Ho C. M.;Chen, R.;He, Q. Y.;Yu, W. Y.;Sun, H.;Tam, P. K.;Chiu, J. F.;Che, C. M.. “Proteomic analysis of the mode of antibacterial action of silver nanoparticles.” J Proteome Res v. 5 2006, p. 916–924.

51. Hsu, Shan-hui;Tseng, Hsiang-Jung;Lin, Yu-Chun. “The biocompatibility and antibacterial properties of waterborne polyurethane-silver nanocomposites.” Biomaterials, v. 31 issue 26, 2010, p. 6796-6808.

52. G. Martínez-Castañón;F. Martínez-Gutierrez;J. Martínez-Mendoza;Facundo Ruiz. “Synthesis and antibacterial activity of silver nanoparticles with different sizes.” Journal of Nanoparticle Research, v. 10 issue 8, 2008, p. 1343- 1348.

53. Shrivastava, S.;Bera, T.;Roy, A.;Singh, G.;Ramachandrarao, P.;Dash, D.. “Characterization of enhanced antibacterial effects of novel silver nanoparticles.” Nanotechnology v. 18, 2007, p. 225103.

54. Alt, Volker;Bechert, Thorsten;Steinrücke, Peter;Wagener, Michael;Seidel, Peter;Dingeldein, Elvira;Domann, Eugen;Schnettler, Reinhard. “An in vitro assessment of the antibacterial properties and cytotoxicity of nanoparticulate silver bone cement.” Biomaterials, v. 25 issue 18, 2004, p. 4383-4391.

55. Wei Liu;Yuan Wu;Chang Wang;Li, Hong C.;Thanh Wang;Liao, Chun Y.;Lin Cui;Zhou, Qun F.;Bing Yan;Jiang, Gui B.. “Impact of silver nanoparticles on human cells: effect of particle size.” Nanotoxicology, v. 4 issue 3, 2010, p. 319- 30.

56. Duffin, Rodger;Lang Tran;Brown, David;Stone, Vicki;Donaldson, Ken. “Proinflammogenic Effects of Low-Toxicity and Metal Nanoparticles In Vivo and In Vitro: Highlighting the Role of Particle Surface Area and Surface Reactivity.” Inhalation Toxicology, v. 19 issue 10, 2007, p. 849-856.

57. CDC - NIOSH Pocket Guide to Chemical Hazards -Silver (metal Dust and Soluble Compounds, as Ag)." Centers for Disease Control and Prevention. 18

Nov. 2010. Web. 23 Feb. 2011. .

58. Garza-Ocañas, L.;Ramirez-Cabrera, M.;Zanatta-Calderon, M.T.;Lujan-Rangel, R.;Ferrer, D.A.;Yacaman, M.J.. “In vitro toxicity assessment of silver nanoparticles in Chang liver cells and J-774 macrophages.” Toxicology Letters, v. 196, 2010, p. S284.

59. Tiwari, Dhermendra K;Jin, Takashi;Behari, J.. “Dose-dependent in-vivo toxicity assessment of silver nanoparticle in Wistar rats.” Toxicology Mechanisms & Methods, v. 21 issue 1, 2011, p. 13-24.

60. Lankveld, D.P.K.;Oomen, A.G.;Krystek, P.;Neigh, A.;Troost – de Jong, A.;Noorlander, C.W.;Van Eijkeren, J.C.H.;Geertsma, R.E.;De Jong, W.H.. “The kinetics of the tissue distribution of silver nanoparticles of different sizes.” Biomaterials, v. 31 issue 32, 2010, p. 8350-8361.

61. David Roe;Balu Karandikar;Nathan Bonn-Savage;Bruce Gibbins;Jean-Baptiste Roullet. “Antimicrobial surface functionalization of plastic catheters by silver nanoparticles.” Journal of Antimicrobial Chemotherapy (JAC), v. 61 issue 4, 2008, p. 869-869.

62. Vlachou, Evangelia;Chipp, Elizabeth;Shale, Elizabeth;Wilson, Yvonne T.;Papini, Remo;Moiemen, Naiem S.. “The safety of nanocrystalline silver dressings on burns: A study of systemic silver absorption.” Burns (03054179), v. 33 issue 8, 2007, p. 979-985.

63. Moiemen, Naiem S.;Shale, Elizabeth;Drysdale, Kate J.;Smith, Gary;Wilson, Yvonne T.;Papini, Remo. “Acticoat dressings and major burns: Systemic silver absorption.” Burns (03054179), v. 37 issue 1, 2011, p. 27-35.

64. “BE lookup table for signals from elements and common chemical species.” (1999). Handbook of Monochromatic XPS Spectra, vol. 1, The Elements and Native Oxides, John Wiley, New York.

65. Tselesh, A.S.. “Anodic behaviour of tin in citrate solutions: The IR and XPS study on the composition of the passive layer.” Thin Solid Films, v. 516 issue 18, 2008, p. 6253-6260.

66. V.I. Nefedov, “X-ray photoelectron spectroscopy of chemical compounds.” Handbook, Chemistry, 1984.

67. Wagner C. D. “Handbook of X-Ray Photoelectron Spectroscopy”. Perkin- Elmer Corporation, Physical Electronics Division, Eden Prairie, Minn. 55344 (1979)

68. Pirlot, C.;Willems, I.;Fonseca, A.;Nagy, J. B.;Delhalle, J.; “Preparation and characterization of carbon nanotube/polyacrylonitrile composites.” Advanced Engineering Materials, v. 4 issue 3, 2002, p. 1-11.

69. Hozumi, Atsushi;Inagaki, Masahiko;Shirahata, Naoto. “Spatially defined silver mirror reaction on a micropatterned aldehyde-terminated self-assembled monolayer.” Applied Surface Science, v. 252 issue 18, 2006, p. 6111-6114.

70. Kumar Kaushik V. J. Electron Spectrosc. Relat. Phemon. v. 56, issue 273, 1991.

71. Ana Maria Ferraria;Sami Boufi;Nicolas Battaglini;Ana Maria Botelho do Rego;Manuel ReiVilar. “Hybrid systems of silver nanoparticles generated on cellulose surfaces”, Langmuir 2010, v. 26 issue 3.

72. Tjeng L.H. Phys. Rev. B 41, 1990, p. 3190

73. Chimentao, Ricardo José;Medina, Francesc;Sueiras, Jesús Eduardo;Fierro, José Luís García;Cesteros, Yolanda;Salagre, Pilar. “Effects of morphology and cesium promotion over silver nanoparticles catalysts in the styrene epoxidation.” Journal of Materials Science, v. 42 issue 10, 2007, p. 3307-3314.

74. Godbole, P.D.;Mitra, A.;Pasricha, R.;Mandale, A.B.;Patil, K.R.. “Deposition and characterization of silver nano-films by a novel solid liquid interface reaction technique (SLIRT).” Materials Letters, v. 59 issue 14/15, 2005, p. 1958-1961.

75. Romand R. J. Solid State Chem. v. 25 issue 59, 1978.

100

76. Schoen G. J. Electron Spectrosc. Relat. Phenom. v. 1, issue 377, 1972.

77. Janardhanan, Revathi;Karuppaiah, Murugan;Hebalkar, Neha;Rao, Tata Narsinga. “Synthesis and surface chemistry of nano silver particles.” Polyhedron, v. 28 issue 12, 2009, p. 2522-2530.

78. Yatsimirskii K.B. Phys. Lett. 52, 481 (1977).

79. Sharma J., Garrett W.L., Owens F.J., Vogel V.L. J. Phys. Chem. 86, 1657 (1982).

80. Turkevich, J.;Stevenson, P.;Hillier, J.. “The formation of colloidal gold.” J. Phys. Chem., 1953, v. 57 issue 7, p. 670–673.

81. Boukamp, P.;Petrussevska, R.;Breitkreutz, D.;Hornung, J.;Markham, A.;Fusenig, N.. “Normal Keratinization in a Spontaneously Immortalized Aneuploid Human Keratinocyte Cell Line.” The Journal of Cell Biology, Volume 106, March 1988, p. 761-771.

82. Thévenot, Daniel;Toth, Klara;Durst, Richard;Wilson, George. “Electrochemical biosensors: recommended definitions classification.” Analytical Letters, v. 34 issue 5, 2001, p. 635-659.

83. "WHO | Burden of Disease and Cost-effectiveness Estimates." Web. 06 Apr. 2011. .

84. Silver, G. M.;Robertson, S. W.;Halerz, M. M.;Conrad, P.;Supple, K. G.;Gamelli, R. L.. “A silver-coated antimicrobial barrier dressing used postoperatively on meshed autografts: a dressing comparison study.” Journal of Burn Care & Research, v. 28 issue 5, 2007, p. 715-719.

85. Poon, Vincent K.M.;Burd, Andrew. “In vitro cytotoxity of silver: implication for clinical wound care.” Burns (03054179), v. 30 issue 2, 2004, p. 140-147.

86. Klasen, HJ. “Historical review of the use of silver in the treatment of burns. I. Early uses.” Burns (03054179), v. 26 issue 2, 2000, p. 117-130.

101

87. Klasen, H. J.. “A historical review of the use of silver in the treatment of burns. II. Renewed interest for silver.” Burns (03054179), v. 26 issue 2, 2000, p. 131- 138.

88. Dujardin, E.;Ebbesen, T.W.;Hiura, H.;Tanigaki, K.. “Capillarity and wetting of carbon nanotubes.” Science, v. 265 issue 5180, 1994, p. 1850-1852.

89. J. Kimling;M. Maier;B. Okenve;V. Kotaidis;H. Ballot;A. Plech. “Turkevich method for gold nanoparticle synthesis revisited.” J. Phys. Chem. B, v. 110 issue 32, 2006, p. 15700–15707.

90. Harris, J. W. and Stocker, H. "Spherical Segment (Spherical Cap)." §4.8.4 in Handbook of Mathematics and Computational Science. New York: Springer- Verlag, p. 107, 1998.

91. Kern, W. F. and Bland, J. R. "Spherical Segment." §36, Solid Mensuration with Proofs, 2nd ed. 1948, p. 97-102.

92. Weisstein, Eric W. "Spherical Cap." From MathWorld--A Wolfram Web Resource. http://mathworld.wolfram.com/SphericalCap.html

102