Aqueous Syntheses of Transition Metal Oxide Nanoparticles For

Aqueous syntheses of transition metal oxide nanoparticles

for biomedical applications

A dissertation submitted to Kent State University in partial fulfillment of the requirements for the degree of Doctor of Philosophy

By Matthew Worden December, 2015

Dissertation written by

Matthew Worden

B.Sc., University of Waterloo, 2008

Ph.D., Kent State University, 2015

Approved by

______, Chair, Dissertation Committee Dr. Torsten Hegmann

______, Dissertation Committee Member Dr. Songping Huang

______, Dissertation Committee Member Dr. Mietek Jaroniec

______, Dissertation Committee Member Dr. Min-Ho Kim

______, Department of Grad Studies Rep. Dr. Edgar Kooijman

Accepted by

______, Chair, Department of Chemistry Dr. Michael Tubergen

______, Dean, College of Arts and Sciences Dr. James Blank

Table of Contents

List of figures ...... vi

List of tables ...... xv

Dedication ...... xvii

Acknowledgments ...... xviii

Overview of thesis goals and outcomes ...... xx

Chapter 1 ...... 1

1.1 Research purpose and goals ...... 2

1.2 Introduction to synthetic methods for and applications of transition metal oxide nanoparticles ...... 3

1.2.2 Manganese oxide nanoparticles ...... 4

1.2.3 Copper oxide NPs ...... 7

1.2.4 Cobalt oxide NPs ...... 9

1.2.5 Mixed metal oxide NPs ...... 10

1.3 Experimental ...... 12

1.3.1 Instrumental methods ...... 12

1.3.2 Synthesis of bare Mn-ONPs ...... 12

1.3.3 Synthesis of Mn-ONPs-APTES ...... 13

iii

1.3.4 Synthesis of Mn-ONPs-OA ...... 14

1.3.5 Synthesis of Cu-ONPs ...... 14

1.3.6 Synthesis of Co-ONPs ...... 15

1.3.7 Synthesis of MFe2O4 NPs ...... 15

1.4 Results and Discussion ...... 16

1.4.2 Copper/Copper oxide NPs (Cu-ONPs) ...... 25

1.4.3 Co-ONPs ...... 29

1.4.4 Mixed metal oxide NPs ...... 31

1.4.5 Reduction/hydrolysis and particle formation mechanism ...... 36

1.5 Conclusions ...... 38

Chapter 2 ...... 44

2.1 Research purpose and goals ...... 45

2.2 Introduction to synthetic methods for, physical properties of, and applications of, iron oxide nanoparticles ...... 47

2.2.1 General mechanisms behind nanoparticle formation ...... 47

2.2.2 Wet chemical synthesis of quasi-spherical IONPs ...... 51

2.2.3 Synthesis of non-spherical IONPs ...... 54

2.2.4 Physical properties of IONPs and the effect of particle shape therein .... 56

2.2.5 The principles of MRI and the use of IONPs as contrast agents ...... 59

2.2.6 IONPs in targeted drug delivery and the effect of particle shape on cell

uptake ...... 63

2.1.7 IONPs in hyperthermia treatment ...... 66

2.1.8 Lyotropic liquid crystals and Triton X surfactants ...... 71

2.3 Experimental ...... 74

2.3.1 Instrumental methods ...... 74

2.3.2 Synthesis of IONBsX45 and IONBsX100 ...... 76

2.2.3 Synthesis of S-IONBsX45 and S-IONBsX100 ...... 77

2.4 Results and Discussion ...... 77

2.4.1 Evaluation of particle morphology and crystallinity ...... 77

2.4.2 Magnetic measurements ...... 87

2.4.3 Surface functionalization and characterization ...... 88

2.4.4 MRI relaxivity measurements ...... 92

2.4.5 Hyperthermia measurements ...... 96

2.4.6 Cell uptake studies ...... 99

2.5 Conclusions ...... 101

Chapter 3 ...... 111

3.1 Project Goals ...... 112

3.2 Functionalization of IONPs with silanes for bioapplications ...... 113

3.2.1 Introduction to surface functionalization of IONPs ...... 113

3.2.2 Instrumentation...... 117

3.2.3 Experimental methods ...... 118

3.3 Results and Discussion ...... 120

3.4 Conclusions ...... 142

Chapter 4 ...... 145

4.1 Project Goals ...... 146

4.2 Additional investigations into IONP shape control and the mechanisms therein ...... 147

4.2.1 Effects of temperature and lyotropic phase on IONP morphology for

synthesis in Triton X surfactants ...... 147

4.2.2 Effects of alternative surfactants on IONP morphology ...... 152

4.2.3 Evaluation of visible crystal facets under TEM as a potential means to

understand particle growth mechanisms ...... 158

4.2.4 Reduction-hydrolysis method for IONP shape control ...... 164

Chapter 5 ...... 169

5.1 Research purpose and goals ...... 170

5.2 Iron oxide nanoparticle / carbon dot composites ...... 170

5.2.1 Introduction to fluorescent carbon dots ...... 170

5.2.2 Materials and instrumentation ...... 174

5.2.3 Experimental methods ...... 175

5.3 Results and discussion ...... 176

5.4 Conclusions ...... 184

Appendix ...... 188

List of Figures

Figure 1: Powder XRD pattern for bare Mn-ONPs...... 17

Figure 2: A) and B) show TEM images of Mn-ONPs. C) shows SAED pattern of

Mn-ONPs with peaks indexed to hausmannite (Mn3O4)...... 18

Figure 3: Histogram of Mn-ONP sizes as evaluated from TEM images...... 18

Figure 4: FT-IR spectra of A) bare, B) oleic acid coated, and C) APTES coated

Mn-ONPs...... 19

Figure 5: Images of (from left to right) APTES-MnONPs in water with hexane layered on top, APTES-MnONPs in water, OA-MnONPs in hexane, and OA-

MnONPs in hexane with water layered below...... 20

Figure 6: Magnetic hysteresis loops of bare Mn-ONPs conducted at 300 K (top) and 10 K (bottom)...... 22

Figure 7: Powder XRD spectra of Mn-ONPs made with: A) 1:5, B) 1:10, and C)

1:20 molar ratios of Mn(acac)3 and NaBH4...... 24

vii

Figure 8: FT-IR spectra of Mn-ONPs made with a 1:1 molar ratio (blue, top) and

1:10 molar ratio (red, bottom)of Mn(acac)3 and NaBH4...... 25

Figure 9: MTT cytotoxicity assay of Caco2, bEnd.3, and HepG2 cells (in cell culture media) after exposure to various concentrations of Mn-ONPs-APTES after 24 hours. Viability values are expressed as a percentage of live cells, observed relative to controls which were not exposed to particles...... 26

Figure 10: Powder XRD of Cu-ONPs. Peaks highlighted in red are indexed to

Cu2O, and those in blue are indexed to elemental Cu...... 27

Figure 11: A) and B) show TEM images of Cu-ONPs. C) shows SAED micrograph of Cu-ONPs...... 28

Figure 12: Histogram of Cu-ONP sizes based on analysis of TEM images...... 28

Figure 13: FT-IR spectrum of Cu-ONPs...... 29

Figure 14: XRD patterns of Co-ONPs before (top) and after (bottom) calcination, showing the change from amorphous to crystalline Co3O4...... 31

Figure 15: TEM images of Co-ONPs. A) before calcination, B) and C) after calcination. D) shows SAED of the particles, indexed to Co3O4...... 32

Figure 16: Powder XRD of mixed metal oxide NPs. A) Shows NPs made with a

19:1 molar ratio of Fe to Mn, B) shows NPs with a 7:1 molar ratio of Fe to Mn, C) shows a 4:1 ratio of Fe to Mn. Each A – C can be indexed to the cubic crystal phase with space group Fd-3m, indicative of Mn doping into iron oxide. D) shows

NPs made with a 2:1 molar ratio of Fe to Mn. The broad peaks suggest a poorly crystallized material or a mixture of many phases...... 34

viii

Figure 17: TEM images of mixed metal oxide NPs MMO1 (A and B), MMO2 (C and D), and MMO3 (E and F)...... 35

Figure 18: TEM images of mixed metal oxide NPs MMO4...... 36

Figure 19: Schematic diagram showing classical and non-classical crystallization.

A shows the classical mechanism to create single crystalline particles. B shows an iso-oriented crystal upon fusing. C shows a mesocrystal composed of small particles coated in an organic surface stabilizer. This image has been modified from Niederberger and Cölfen.4 ...... 50

Figure 20: Schematic of the principles of MRI. A) shows the precession of a proton's magnetic moment (m) about that of an external field (B0) along the z- axis. B) shows the net vector moment, mz, before the RF pulse. C) Shows an RF pulse moving the net vector moment, mxy, into the xy-plane. D) shows the decrease of mxy as individual proton moments diphase (T2 relaxation), the incease of mz as the net vector realigns with (T1 relaxation) and the release of RF energy...... 63

Figure 21: Schematic representation of two types of magnetic relaxation mechansims that result in heat generation in an IONP. The top shows Neel relaxation, the movement of the internal magnetic moment against the orientation of the crystal structure as the external magnetic field moves. The bottom shows

Brownian relaxation, the physical movement of the entire particle to align the internal moment with the external field. This figure is adapted from Kumar and

Mohammad...... 69

Figure 22: (Top) General structure of Triton X molecule. For X100, n = 10; for

X45, n = 4 - 5. (Bottom left) Schematic representation of a lamellar phase.

(Bottom right) Schematic representation of a hexagonal phase...... 73

Figure 23: Powder XRD patterns for IONBsX100 (top) and IONBsX45 (bottom).

Peaks have been indexed to match that of bulk magnetite...... 80

Figure 24: TEM images of IONBsX45...... 81

Figure 25: TEM images of IONBsX100...... 82

TEM images of this for IONBsX45 and IONBsX100, respectively. Elongation along one edge leads to a parallelepiped as in image D. HR-TEM images of this shape for IONBsX45 are shown in figures E and F. The same shape as D seen along an edge leads to a purely rectangular shape, as in figure G. HR-TEM images H and I show examples of this shape seen in IONBsX45...... 83

Figure 27: Visualization of the discussion on the possibility of octahedral vs. rhombohedral shapes. Figures 5A and 5B show HR-TEM images of particles from IONBsX45. Similar images can be seen in other syntheses which conclude these are octahedral particles (as in image C). Figure 5D shows how two rhombohedral particles could fuse via mesoscale assembly to form one larger polyhedral particle. Figure 5E shows a TEM image of these two types of particle shapes side by side from IONBsX45...... 85

Figure 28: TEM images of IONBsX45 at varying tilt angles. These demonstrate how a rhombohedral shape can be made to look cubic or rectangular, depending on the orientation of the particle relative to the electron beam...... 86

Figure 29: POM images under crossed polarizers showing birefringent textures of lyotropic phases. A shows an image of a 50% mixture of Triton X45 in water at

35 ˚C; B shows the same along with a 2:1 mixture of FeCl3 and FeCl2, which represents the condition of the reaction mixture before hydrolysis with NaOH occurs. Fingerprint textures typical of a lamellar phase can be seen in both. C shows an image of a 50% mixture of Triton X100 in water at 30 ˚C; D shows the same along with a 2:1 mixture of FeCl3 and FeCl2. Focal conic textures typical of a hexagonal phase can be seen in both (limited transmission due to presence of iron salts), although the inclusion of the iron precursors does lower the transition temperature slightly...... 87

Figure 30: Magnetic hysteresis curves for IONBsX100 (triangles) and IONBsX45

(squares) at 300 K. Inset shows a magnified image of the coercivity for each particle set...... 89

Figure 31: (Left): Picture of particle dispersions of S-IONBsX45 and S-

IONBsX100. (Right): Schematic representation of the reaction scheme resulting in coated brick-like particles...... 90

xi at ~1600 cm-1 are indicative of COO- stretching; the broad peaks centered around ~1000 cm-1 are due to Si-O-R stretching; and the large asymmetric peaks at ~590 cm-1 correspond to stretching modes associated with Fe2+-O and Fe3+-O.

...... 91

Figure 33: TGA plots for S-IONBsX45 (top) and S-IONBsX100 (bottom) showing weight loss profiles of ~45% in both cases...... 92

Figure 34: Plots used to determine relaxivity values measured at 1.5 T for: A S-

IONBsX100, B S-IONBsX45, C S-IONPs. y-axes show the inverse of the relaxation time; x-axes show Fe concentration in mM...... 95

Figure 35: Plots used to determine relaxivity values measured at 7 T for: A S-

IONBsX100, B S-IONBsX45, C S-IONPs. y-axes show the inverse of the relaxation time; x-axes show Fe concentration in mM...... 96

Figure 36: Increase in temperature vs. time for different particle dispersions during exposure to AC magnetic field. Inset: calculated SLP values for each set of particles. AC field was set at an amplitude of 20 kA/m, and a frequency of 2

MHz...... 100

Figure 37: Schematic representation of the silanization of an IONP...... 116

Figure 38: Representations of IONPs coated (top) with AmS, and (bottom) after surface modification with BAA (indicated by the star shape)...... 122

Figure 39: Characterization of AmS-IONPs. (Top) FT-IF spectrum; (bottom) TGA weight loss plot...... 125

xii

Figure 40: Characterization of COOH-AmS-IONPs. (Top) FT-IR spectrum;

(bottom) TGA weight loss plot...... 126

Figure 41: Representative TEM images of AmS-IONPs (A and B), and EDTS-

IONPs (C and D)...... 129

Figure 42: Representation of an IONP coated with EDTS...... 130

Figure 43: Characterization of EDTS-IONPs. (Top) FT-IR spectrum; (bottom)

TGA weight loss plot...... 131

Figure 44: Example of a plot of hydrodynamic radius vs. normalized intensity for

EDTS-IONPs in DI water (black) and PBS (red). Inset shows EDTS-IONPs dispersed in PBS (left) and water (right)...... 134

Figure 45: Schematic procedure for a generalized EDC/NHS coupling to form an amide bond between a carboxylic acid and a primary amine. Step 1 involves the formation of an unstable acylurea ester between a carboxylate and EDC. Step 2 involves the formation of a semi-stable amino ester between NHS and the carboxylate. Step 3 involves the final coupling between the carboxylate and the primary amine via the formation of an amide bond...... 137

Figure 46: Representation of an IONP coated with EDTS and coupled with PEI via an amide bond...... 138

Figure 47: Characterization of PEI-IONPs. (Top) FT-IR spectrum; (bottom) TGA weight loss plot...... 142

xiii

Figure 48: TEM images of IONPs made in 50 wt% X45. Reaction temperatures are: (A and B) 25 °C, (C and D) 45 °C, (E and F) 55 °C. Each system remains in a lamellar phase...... 149

Figure 49: TEM images of IONPs made in 50 wt% Triton X100. Reaction temperatures are: (A and B) 25 °C, (C and D) 45 °C, (E and F) 55 °C...... 151

Figure 50: Chemical structures of Brij C10 (top, where n = 10), and CTAB

(bottom)...... 153

Figure 51: A and B show TEM images (A and B) of particles made in 50 wt% Brij

C10 in water with Fe precursors at 40 °C. Bottom set show POM images of 50 wt% Brij C10 in water, showing the onset of a hexagonal phase beginning above

67 °C...... 154

Figure 52: TEM results of particles synthesized in 0.5 M CTAB at 35 °C, an isotropic micelle phase. The images show no distinct particle shapes...... 155

Figure 53: Results for particles synthesized in 1M CTAB in the hexagonal phase.

A and B show TEM images, revealing brick-like particles. POM images show a type of hexagonal columnar texture beginning at 42.5 °C and persisting at the reaction temperature of 35 °C...... 156

Figure 54: Schematic representation of Miller index notation for selected planes

(top, round brackets) and directors (bottom, square brackets)...... 159

xiv in top right taken from Song and Zhang with permission. TEM image in bottom right taken is IONBX45 taken from Worden et al...... 161

Figure 56: TEM images of a single IONB. Angles in the top right are with respect to the initial stage position. d spacing values and related Miller indices are located on the bottom right of each image...... 164

Figure 57: TEM images of nanostructures made via the borohydride reduction mechanism in the presence of Triton X45. Reaction parameters are listed in

Figure 58: Schematic representation of the process for creating IONP/Cdot composites. Top image shows an IONP coated with AmS (left structure) and

EDTS (right structure). Bottom left image shows a schematic of the particle after calcination, along with a picture of the calcined particles under ambient light

(bottom middle), and illuminated under 365 nm UV light (bottom right)...... 177

Figure 59: Fluorescence spectra for AmS-IONPs before (left) and after (right) calcination...... 179

Figure 60: Fluorescence spectra for AmS-EDT-IONPs before (left) and after

(right) calcination...... 179

Figure 61: Fluorescence spectra for EDT-IONPs before (left) and after (right) calcination...... 180

Figure 62: Representative TEM images of silanized IONPs (left) and IONP/Cdots

(right)...... 181

Figure 63: FT-IR spectra of AmS-IONPs before (bottom) and after (top) calcination...... 181

Figure 64: FT-IR spectra of AmS-EDTS-IONPs before (bottom) and after (top) calcination...... 182

Figure 65: FT-IR spectra of EDTS-IONPs before (bottom) and after (top) calcination...... 182

List of Tables

Table 1: Concentration of Fe and Mn in different mixed metal oxide NP samples, as measured by ICP...... 36

Table 2: Physiochemical properties of silanized particles in water...... 92

Table 3: Relaxivity values at different field strengths for EDTS coated particles. 93

Table 4: Comparison between commercially available and literature reported iron oxide based contrast agents...... 96

Table 5: Physio-chemical properties of silanized IONPs dispersed in water. ... 126

Table 6: Effective particle diameter and zeta potential of EDTS-IONPs dispersed in water and PBS...... 132

Table 7: Physiochemical properties of PEI-IONPs dispersed in DI water and 1M

PBS solution...... 138

Table 8: Summary of results for IONB syntheses in X45 at different temperatures and lyotropic phases...... 150

Table 9: Summary of results for IONB syntheses in X100 at different temperatures and lyotropic phases...... 152

xvi

Table 10: Selected list of reaction parameters investigated for possible IONP shape control with the reduction-hydrolysis method modified with Triton X45. . 167

xvii

I dedicate this dissertation to

David Lamble for being the first to show me the wonderful world of chemistry

xviii

Acknowledgments

First and foremost, I must extend my thanks and gratitude to my PhD advisor, Dr. Torsten Hegmann. He welcomed me into his lab in way that immediately made me feel like a collaborator rather than simply a student. He always provided guidance and insight when I needed it, but also allowed me to pursue my own research interests when I had the desire.

I want to thank all the members of Dr. Hegmann’s research group with whom I’ve had the pleasure to work over the years. In particular I must thank Dr.

Vinith Yathindranath and Anshul Sharma. Vinith helped me immediately to get established in the lab, and indeed it was his work that set the groundwork of my own. Without his insight I would have struggled much more than I already did.

Anshul was the only other member of the Hegmann lab to come to Kent, and we experienced all the stages, the trials and tribulations, of graduate research simultaneously. Without both her scientific insight and her friendship my experience in Kent would have not been nearly as successful.

I owe many debts of gratitude to too many people to name at Kent State and the University of Manitoba. To highlight just a few, I wish to thank Dr. Min

Gao from the LCI for providing invaluable training me on the TEM, an instrument without which my research would not have been possible. And I wish to thank

xix

Zhizhi Sun and Dr. Don Miller at the U of M whose work demonstrated a number of exciting applications for the variety of materials I have made.

Finally, and most of all, I want to thank my wife, Angela. Suffice to say that none of this would have had much meaning without her love and support.

Overview of thesis goals and outcomes

The research presented in this thesis covers work I conducted under the supervision of Prof. Torsten Hegmann. I joined Prof. Hegmann’s group in

September of 2010 at the University of Manitoba in Wininipeg, MB, Canada, and later joined his lab at the Liquid Crystal Institute at Kent State University in

January of 2012. Broadly speaking, my research has focused on the synthesis and characterization of metal oxide nanoparticles. This work has been highly interdisciplinary and collaborative in nature, allowing me to work alongside not just materials chemists, but physicists and biologists as well. In particular, much of my work has been informed by and has contributed to research on nanoparticle-cell interactions conducted in Dr. Don Miller’s lab at the University of

Manitoba. Specifically how my work has contributed to these pharmacological studies is detailed below.

This thesis is broken down into five chapters. Each chapter, given a brief overview below, contains separate detailed introductions, conclusions, and literature reviews relevant to the individual research topics contained within.

Chapter 1 concerns investigations into a new, general synthetic method for transition metal oxide nanoparticles. Previous work by a Hegmann lab member, Vinith Yathindranath, resulted in a simple, one pot, low temperature, aqueous method for the synthesis of iron oxide nanoparticles through a reduction/hydrolysis mechanism. This method could also be easily modified to

xxi create functionalized, water dispersible particles. With this research to build from,

I worked with Vinith on expanding this method to see what other types of particles could be made with the same general reaction scheme. My research focused specifically on manganese, cobalt, copper, and mixed metal oxide nanoparticles. The results show that this reduction/hydrolysis method can in fact be expanded upon with minor modifications to make nanoparticles of Mn3O4,

Co3O4, Cu/Cu2O, and Fe3O4 doped with various concentrations of manganese.

Additionally, using the manganese oxide NPs as examples, I showed that these particles can be functionalized with both hydrophilic and hydrophobic ligands so as to make stable dispersions in various solvents.

Chapter 2 concerns attempts to change and control the shape of iron oxide nanoparticles (IONPs) using modified aqueous syntheses. This research was informed by the idea that particle shape can have dramatic effects on both the physical properties of the particles themselves as well as how these materials interact with biological systems. At the same time, there are very few examples in the literature of pharmacological studies on non-spherical IONPs, most likely because most aqueous (and thus biocompatible) synthetic methods for IONPs naturally result in quasi-spherical particles. In order to address this, I investigated the use of surfactants that can form lyotropic liquid crystal (LLC) phases in aqueous media as a means to change the shape of IONPs synthesized within such a mixture. These modifications allowed for the creation of polyhedral particles best described as “brick-like” in shape, termed iron oxide nanobricks

xxii

(IONBs). While providing a means to change the shape, the synthesis also allows for the functionalization of the IONBs directly in situ with hydrophilic silanes, allowing for stable dispersions of these particles in aqueous media.

These IONBs are crystalline and show magnetic properties similar to quasi- spherical IONPs. The importance of particle shape, however, was shown with investigations as MRI contrast agents and in hyperthermia. The IONBs are significantly more effective and efficient in these applications as compared with comparable, quasi-spherical IONPs. Furthermore, this chapter briefly discusses studies conducted by Zhizhi Sun at the University of Manitoba that show that the silanized IONBs exhibit unique cell uptake properties which are not present in comparable quasi-spherical IONPs.

Chapter 3 concerns research done on the surface chemical properties of

IONPs through the use and modification of hydrophilic silane molecules. Organo- silanes have been shown to be excellent compounds that allow for the synthesis of stable, robust, biocompatible dispersions of particles in aqueous media.

Furthermore, the terminal functional groups on these compounds often allow for additional chemical modifications, allowing for particles bonded to and associated with polymers, proteins, and a variety of biologically relevant materials. Previous work done in the Hegmann lab showed that modification with aminosilane (AmS) gave stable, positively charged dispersions of IONPs in aqueous media. My work involved changing the surface functionalities on the IONPs, through chemical modification of AmS or through the use of entirely new silanes, to create

xxiii negatively charged IONPs when dispersed in water. The cell toxicity and uptake properties of both positively and negatively charged IONPs were investigated by

Zhizhi at the University of Manitoba in Dr. Don Miller’s lab in order to better elucidate the role that surface potential can have on biologically relevant properties of IONPs. Additionally, this chapter contains my work on the functionalization of IONPs with a positively charged polymer, poly(ethyleneimine). These particles were again used in investigations in the

Miller lab at the University of Manitoba for the purpose of siRNA delivery to cells.

Chapter 4 presents research intended as a supplement to that done in

Chapter 2. It shows my attempts to better understand the role that the LLC surfactant has one particle shape through changes in reaction conditions, as well as through the use of alternative LLC surfactants not discussed in Chapter 2. The results of this research suggest that the particular LLC phase does not in fact act as a “template” for the particles. This chapter also looks at my attempts to use the crystallographic information provided by TEM imaging of particles to gain a greater understanding of the 3-dimensional shape of IONBs. Lastly, this chapter contains a summary of early investigations on IONP shape control I did using an alternative reaction scheme, and provides a discussion as to why these investigations were abandoned in favor of those discussed in Chapter 2.

Chapter 5 discusses investigations into creating composite fluorescent carbon dot/IONP nanomaterials. Carbon dots (Cdots) can be made through the

xxiv calcination of organic compounds, and recent reports have shown that organosilanes can also be used as precursors to the formation of Cdots.

Separately, IONPs functionalized with organic fluorophores are often used in cell imaging and labelling experiments. With these ideas in mind, I wanted to look at the possibility of synthesizing fluorescent IONPs via the calcination of silanized particles. I used a variety of silanized IONPs (which themselves exhibit no fluorescence properties), as described in Chapter 3, and heated them at high temperatures in conditions similar to typical Cdot syntheses. This resulted in water dispersible particles that exhibit wavelength-independent blue fluorescence. The possible mechanism behind the fluorescence properties is also briefly discussed. This work is intended as a proof-of-concept; optimization of these composite particles for use in cell imaging would be a possible avenue of investigation for future researchers.

Since joining Prof. Hegmann’s research group, I have contributed to 8 published papers, as well as one more currently in press. Additionally, two more have been submitted for peer review, and two are currently in preparation and will be ready for submission for peer review in the next few months. A list of these papers, published and otherwise, can be found below. A brief discussion of my specific contributions is included, as well as where in this thesis the data relevant to each paper may be found.

xxv

1) M. Worden, T. Hegmann. Fluorescent magnetic nanoparticles via

calcination of silanized iron oxide NPs, 2015, in preparation.

The results of this work are reported at length in Chapter 5. The entirety of the work – synthesis, characterization, and the written discussion – was done by me.

2) S. Morrison, M. Worden, V. Yathindranath, Z. Sun, T. Hegmann, D. Miller.

siRNA delivery with PEI coated magnetic nanoparticles, 2015, in

preparation.

I synthesized and characterized the PEI functionalized IONPs used in the siRNA delivery studies discussed in this manuscript. The specifics of my work that contributed to this manuscript are reported at length in Chapter 3, and the general conclusions of the above manuscript can also be found there.

3) M. Worden, M. Bruckman, M-H. Kim, N. Steinmetz, J. M. Kikkawa, C.

LaSpina, T. Hegmann. Aqueous synthesis of polyhedral “brick-like” iron

oxide nanoparticles for hyperthermia and T2 MRI contrast enhancement,

J. Mater. Chem. B, 2015, in press.

The results of this work are discussed at length in Chapter 2. The entirety of the work was done by me, with the exception of the following: MRI contrast measurements were conducted along with Dr. Bruckman, hyperthermia measurements were conducted along with Dr. Kim, XRD data was collected by

Michal Marszewski, and Dr. Kikkawa collected the magnetization data himself.

xxvi

4) Z. Sun, M. Worden, Y. Wroczynskyj, J. A. Thliveris, J. van Lierop, T.

Hegmann, D. W. Miller. Differential internalization of brick shaped iron

oxide nanoparticles by endothelial cells, 2015, submitted.

I synthesized the particles used for the cell uptake studies reported in this manuscript. The physical properties of the particles themselves were gathered by me as reported in the paper noted above (2). The general results and conclusions of this manuscript are discussed at the end of Chapter 2, and a copy of the manuscript as it currently stands can be found in Appendix A.

5) Y. Wroczynskyj, Z. Sun, D. W. Miller, M. Worden, T. Hegmann, J. van

Lierop. A simple alternative to assay based techniques for determining

iron oxide nanoparticle concentrations using magnetic measurements,

2015, submitted.

I synthesized the particles used in the studies reported in this manuscript.

6) B. M. Yao, Y. S. Gui, M. Worden, T. Hegmann, M. Xing, X. S. Chen, W.

Lu, Y. Wroczynskyj, J. van Lierop, C.-M. Hu. Quantifying complex

permittivity and permeability of magnetic nanoparticles, Appl. Phys. Lett.

2015, 106, 142406.

I synthesized the particles used in the studies reported in this manuscript.

7) A. Sharma, T. Mori, H.-C. Lee, M. Worden, E. Bidwell, T. Hegmann.

Detecting, visualizing, and measuring gold nanoparticle chirality using

xxvii

helical pitch measurements in nematic liquid crystal phases, ACS Nano,

2014, 8, 11966-11976.

I contributed to the characterization of the gold nanoparticles (specifically

TEM imaging) used in the studies reported in this paper.

8) Z. Sun, M. Worden, Y. Wroczynskyj, V. Yathindranath, J. van Lierop, T.

Hegmann, D. W. Miller. Magnetic field enhanced convective diffusion

(MECD) of iron oxide nanoparticles in an osmotic disrupted cell culture

model of the blood-brain barrier, Int. J. Nanomed. 2014, 9, 3013-3026.

I synthesized and characterized all of the IONPs (examples of which are reported at length in Chapter 3) used in these studies. The general results and conclusions of this paper are discussed at the end of Chapter 3.

9) V. Yathindranath, Z. Sun, M. Worden, L. J. Donald, J. A. Thliveris, D. W.

Miller, T. Hegmann. One-pot synthesis of iron oxide nanoparticles with

functional silane shell: A versatile general precursor for conjugations and

biomedical applications, Langmuir, 2013, 29, 10850-10858.

The bulk of the work discussed in this paper was done by Dr. Yathindranath. I contributed small samples of a variety of different particles and helped with the reproducibility of the synthetic methods.

10) V. Yathindranath, M. Worden, Z. Sun, D. W. Miller, T. Hegmann. A

general synthesis of metal (Mn, Fe, Co, Ni, Cu, Zn) oxide and silica

xxviii

nanoparticles based on a low temperature reduction/hydrolysis pathway,

RSC Advances, 2013, 3, 23722-23729.

Dr. Yathindranath and I split the work discussed in this paper approximately equally. The specifics of what I contributed are given at length in Chapter 1 of this thesis, with all of the synthesis and characterization data found therein done by me. Cell toxicity studies on the manganese oxide particles were done by

Zhizhi Sun. The paper itself was written primarily by Dr. Yathindranath.

11) Z. Sun, V. Yathindranath, M. Worden, J. A. Thliveris, S. Chu, F. E.

Parkinson, T. Hegmann, D. W. Miller. Characterization of cellular uptake

and toxicity of aminosilane-coated iron oxide nanoparticles with different

charge in CNS relevant cell culture models, Int. J. Nanomed. 2013, 8, 961-

970.

I synthesized and characterized the negatively charged COOH-AmS-IONPs used for the cell uptake studies discussed in this paper. The specifics of my work that contributed to this paper are discussed at length, and the general conclusions of this paper are also discussed briefly, in Chapter 3.

12) V. Yathindranath, M. Worden, V. Ganesh, M. Inokuchi, T. Hegmann.

Highly crystalline iron/iron oxide nanosheets via liquid crystal templating,

RSC Adv., 2013, 3, 9210-9213.

xxix

I contributed to this paper through synthesis of materials without the templating surfactant (for comparison purposes) as well as some of the characterization of of the materials via TEM imaging. This paper established the groundwork for the research discussed in Chapters 2 and 4.

13) A. Sharma, M. Worden, T. Hegmann. Nanoparticle-promoted thermal

stabilization of room temperature cholesteric blue phase mixtures,

Ferroelectrics, 2012, 431, 154-163.

I synthesized and characterized the manganese oxide nanoparticles used in part of the work discussed in this paper. These particles were based on the synthesis described in Chapter 1 of this thesis.

xxx

Chapter 1

Aqueous synthesis of manganese, copper, cobalt, and mixed metal oxide

nanoparticles

1.1 Research purpose and goals

The main goal of the research detailed in this chapter is to investigate the applicability of the reduction/hydrolysis method in the synthesis of other transition metal oxide NPs. The reduction/hydrolysis method was developed by Vinith

Yathindranath in the Hegmann lab at the University of Manitoba for the purpose of synthesizing iron oxide nanoparticles (IONPs).1 This method shows a number of benefits over other existing syntheses of IONPs: the reaction medium is aqueous-based, it can be conducted at room temperature, it requires only a single metal precursor (thus eliminating the need to precisely balance the ratio between metal precursors of different oxidation states), it is a “one-pot” reaction with a simple set-up, and it gives purse phase Fe3O4 particles with a relatively narrow size distribution under 10 nm. Additionally, this method allows for easy functionalization of the particles directly in-situ, and can be used to make both hydrophilic and hydrophobic particles. At the same time, there are relatively few general synthetic methods available that allow for the production of a wide variety of metal oxide NPs under the same general conditions, and effectively none that can be done at low temperatures in aqueous solvents.

With all this in mind, the reduction/hydrolysis method appeared to be a promising method to create a simple, low cost, general approach to the synthesis of transition metal oxide nanoparticles. The results of this research are detailed below.

1.2 Introduction to synthetic methods for and applications of transition metal oxide nanoparticles

1.2.1 General methods for transition metal oxide NPs

While variations on the basic strategy of synthesizing metal oxide NPs via thermal decomposition of metal precursors in a high boiling point solvent have been used for nearly every transition metal, only a relatively few papers have explicitly developed a single strategy to create multiple metal oxide NPs. Jana et al. developed a method to make ZnO, NiO, Cr3O4, Fe3O4, MnO and Cr2O3 NPs based on the decomposition of metal fatty acid precursors.2 The precursors were heated to 300 °C or more in a mixture of octadecene and the corresponding fatty acid (oleic, stearic, or myristic acid), resulting in spherical, monodisperse NPs, the precise size of which could be tuned by controlling the molar ratios of the reactants and solvent. Similarly, Wang et al. synthesized Mn3O4, Co3O4, CoO,

ZnO, NiO, and CeO2 NPs of various shapes and sizes via thermolysis of metal nitrate precursors in octadecylamine.3 A small number of methods using alternative reaction mechanisms have also been used. Kumar et al. showed a method for creating Fe3O4, Co3O4, CuO, and ZnO NPs by sonochemical hydrolysis and oxidation of metal acetate precursors in a mixture of water and

DMF.4 These mixtures were agitated with a high intensity ultrasonicator for several hours under argon resulting in highly crystalline particles, although all types of particles had high polydispersity indices. A recent technique, described as “solvent-deficient”, allows for the formation of the highest number of different

3 types of metal oxide NPs in the literature, including Mn2O3, Fe2O3, Fe3O4, CoO,

5 Co3O4, NiO, PdO, CuO, Ag2O, ZnO, among others. The method involves physically grinding a metal salt with ammonium bicarbonate and then calcinating the resulting powder in air between 220 – 550 °C, depending on the metal involved, for several hours. The researchers describe the overall mechanism as being similar to that of aqueous co-precipitation routes, in which metal salts are hydrolyzed in basic aqueous media to form metal oxide NPs, except in this case no solvent is needed. This solvent-deficient technique limits the potential applications of the particles, however, since there are no easy means to functionalize the resulting particles, and as such stable dispersions of the particles in any type of solvent are not possible.

1.2.2 Manganese oxide nanoparticles

Manganese oxide nanoparticles (Mn-ONPs), comprised typically MnO,

Mn2O3 or Mn3O4, though other phases exist, have been investigated primarily for biomedical and catalytic applications. These materials exhibit unique, size- dependent magnetic properties that allow for their potential use as T1 MRI contrast agents (as distinct from iron oxide NPs, which are typically used as T2 contrast agents) and as drug delivery vehicles.6,7,8,9 Catalytic applications result from the high surface-to-volume ratio of nanosized materials and the ability to control particle structure and morphology. Within this realm, Mn-ONPs have been often used to facilitate carbon monoxide oxidation.10,11

As with most transition metal oxide NPs, the most common methods for synthesizing nanosized manganese oxides involve variations on the high temperature decomposition of a manganese precursor in a high boiling point solvent. One of the most highly cited examples involves heating Mn(acac)2 in oleylamine, with the OA acting as both solvent and capping agent for the particles.12 The resulting particles – approximately 10 nm in size and composed of MnO or Mn3O4 depending on the particular reaction conditions – are roughly spherical, and the size and polydispersity can be tuned to a small extent through the addition of small amounts of water to the precursor mixture. Yin and O’Brien demonstrated a very similar synthesis of irregularly shaped Mn3O4 particles on the order of 20 nm in size.13 This method involves the rapid heating of manganese acetate in a mixture of trioctylamine and oleic acid, with the latter acting as capping agent. In a variation on the same theme, Ghosh et al. produced roughly spherical sub-10 nm MnO NPs by thermally decomposing a

Mn precursor in a mixture of toluene and TOPO (trioctylphosphine oxide), with the former acting as solvent and the latter as particle capping agent.14 Changes to this general scheme have allowed researchers some control over Mn-ONP shape as well. Hyeon et al. were able to obtain nano-spheres, “-platelets”, and – wires of Mn3O4 through the heating of manganese acetate in a mixture of water and xylene at 90 ºC, along with stabilizers such as oleylamine.15 Zheng et al. produced 1-dimensional elongated wires of Mn3O4. In their procedure MnCO3

5 was first synthesized as a precursor, which was then mixed with NaCl and ether and heated to 850 ºC for two hours in order to obtain the nanowires.16

There have been a small number of publications detailing syntheses done in aqueous conditions. Li et al. reacted aqueous mixtures of Mn(CH3COO)2·4H2O and NaOH along with polyethylene glycol as a functionalizing ligand to create

Mn3O4 nanocrystals; this method, however, results in relatively large particles

17 (~90 nm). Yang et al. synthesized monodisperse Mn3O4 NPs by heating manganese acetate and KOH in ethanol to 60 ºC for 24 hours; the particles, however, were not discrete, but instead assembled into large sheets microns in size.18

The modification of the reduction/hydrolysis method allows for the creation of irregularly shaped, pure phase Mn3O4 nanoparticles of approximately 20 nm in size without the need for high temperatures or harsh solvents. Functionalization of the particles, often a requirement for certain applications, is also quite easy using this method. The functionalizing agent is simply added directly to the aqueous dispersion after the particles have been formed, without the need for additional ligand exchange steps. (3-Aminopropyl) triethoxysilane (APTES) and oleic acid (OA) were used as example functionalizing agents; these allow for the dispersion of MnONPs in polar and non-polar solvents, respectively.

1.2.3 Copper oxide NPs

Copper oxide nanoparticles (Cu-ONPs), composed of cupric oxide, CuO,

19,20 or cuprous oxide, Cu2O, have been investigated for their use in catalysis and chemical sensing.21 Furthermore, as a p-type semiconductor with a bandgap of

2.0 eV, cuprous oxide offers potential for use in photonics. For example, Cu2O nanowires have been used in conjunction with ZnO and TiO2 to demonstrate the feasibility of solar energy conversion using inexpensive materials.22

As with other metal oxide NPs, Cu2O NPs have been synthesized in a variety of ways. One of the more popular methods involves a typical solvothermal reduction. In a synthesis devised by O’Brien et al. copper(I) acetate was dissolved in a mixture of oleic acid and trioctylamine and heated first to 180 ºC for one hour, then 270 ºC for an additional hour.23 The particles, relatively monodisperse at less than 10 nm, were then precipitated using ethanol, and could be redispersed in hexane. A similar method (in which oleic acid was replaced with tetradecylphosphonic acid) used by Hung et al. allowed for the

24 creation of core-shell Cu-Cu2O as well as hollow Cu2O NPs. Particles of pure

Cu were synthesized first via this method, while subsequent oxidation in different organic solvents allowed for the formation of cuprous oxide particles.

There are a few papers that describe aqueous syntheses as well. Murphy and Gou created Cu2O “nanocubes” through an aqueous mixture of Cu(SO4),

CTAB and sodium ascorbate heated with aqueous NaOH at 55 ºC.25 This allowed for the formation of the copper hydroxide; after cooling to room

7 temperature, cuprous oxide slowly formed and precipitated from the solution.

Relatively large (upwards of 500 nm) and polydisperse NPs are formed by this method. A certain amount of size and shape control was demonstrated by Bao et

26 al. using a very similar method. Aqueous solutions of CuCl2 and NaOH were mixed at 55 ºC, along with varying amounts of poly(vinylpyrrolidone). An ascorbic acid solution was then added after a certain amount of time, allowing for the slow precipitation of the particles. Cubic and octahedral shaped Cu2O particles were formed with this method; the particles remain several hundred nm in size.

Additionally, there exist a few procedures roughly similar in various ways to the reduction/hydrolysis method. The first, devised by Wang et al., also yields core-

27 shell Cu-Cu2O particles. In this method, CuCl2 is dissolved in an isopropanol/water mixture along with polyvinylalcohol as a particle stabilizer.

Addition of KBH4 to this mixture caused the formation of the core-shell particles.

This report, however, lacks any TEM images or other methods to determine particle size and shape, so it is difficult to compare any similarities or differences between this procedure and the current method described in this thesis. The second, devised by Dodoo-Arhin et al., uses NaBH4 to reduce a copper

28 precursor in an aqueous media to create pure Cu2O. The major difference is that surfactants are added to create reverse-micelles, inside which the particles form, a requirement not present in our procedure.

1.2.4 Cobalt oxide NPs

Cobalt oxide NPs (primarily Co3O4 and CoO) and other nanostructures have shown promising applications in catalysis, electrochemical batteries, and as components in magnetic memory devices.29 These particles are not nearly as ubiquitous as other metal oxide NPs, however. This is perhaps due to the fact that their general properties are similar to that of iron oxide NPs (eg. both are superparamagnetic when below ~50 nm in diameter, both tend to form cubic, spinel M3O4 structures), without showing substantial benefits over the more commonly investigated materials. Cobalt oxide NPs also tend to exhibit a much higher degree of toxicity as compared with IONPs, thus eliminating a huge area of research concerned with bioapplications.30

The synthesis of cobalt oxide NPs follows many of the typical methods outlined above, with a relatively few reports discussing synthetic methods designed solely for cobalt oxide nanomaterials. Modifications of Massart’s co- precipitation method (for iron oxide NPs) are common. For example, Feng et al. demonstrated a size controlled synthesis of cubic cobalt oxide NPs through a reaction involving aqueous cobalt nitrate and sodium hydroxide mixed under reflux.31 The size of the particles was tuned simply by “aging” the reaction solution for various lengths of time. Xu et al. reported a similar method hydrothermal method involving cobalt salts in basic solution under reflux conditions.32 Their procedure was modified with the inclusion of a commercial surfactant, Tween-85, which they claim caused the particles to self-assemble into

9 various larger structures, depending on the concentration of the surfactant.

Oftentimes post-synthesis calcination is required to induce crystallinity in cobalt oxide nanomaterials. Park et al. reports a synthesis involving cobalt nitrate mixed with sodium dodecylsulfate in a methanol/water solution and then reacted in an autoclave at 180 °C.33 The particles isolated afterwards were then heated to 500

°C for 3 hours, which, the authors report, helped to “improve the crystallinity and phase purity” of the resulting NPs. Similarly, Davies et al. report the need of a calcination step to induce crystallinity for their nanosized cobalt oxide particles.34

1.2.5 Mixed metal oxide NPs

Doping metal oxide NPs with other transition metal cations can often be done through simple modifications to the general techniques described above.

Each of the above methods implicitly allows for the production of mixed metal oxide NPs, although they do not provide specific data on this kind of synthesis.

Theoretically any material discussed herein can be doped with other metals, but most commonly this is done with magnetite (Fe3O4) as the “parent” material.

Variations on high temperature decomposition mechanisms are again the most popular methods. One of the most highly cited methods to create IONPs, devised by Sun et al., involves the decomposition of Fe(acac)3 in a high boiling point alcohol along with oleic acid and oleylamine.35 In a simple modification, the same researchers show that introducing Mn(acac)2 or Co(acac)2 in controlled molar ratios allows for the synthesis of monodisperse, spherical particles of CoFe2O4 or

36 MnFe2O4. In a method similar to the general scheme by Jana et al. discussed

10 above, Adireddy et al. synthesized MFe2O4 NPs, where M could be Mn, Co, Ni, or Zn.37 Although fewer in number, there are examples of mixed metal oxide NP syntheses that do not involve solvothermal mechanisms. Liu et al. made MFe2O4

NPs, where M = Mn, Co, Ni, Cu, Zn, Mg, or Cd, through a reverse micelle reaction of metal nitrates in a water/toluene/ sodium dodecylbenzenesulfonate mixture.38 The downside to this method is that the resulting particles are amorphous and require a high temperature (~350 °C) calcination step in order to become crystalline. Modifications to the co-precipitation route for IONP synthesis provide perhaps the simplest means of creating mixed metal oxide NPs. Pereira et al. created MFe2O4 NPs, where M = Co or Mn, through the hydrolysis and precipitation of metal chloride salts using high molecular weight amines and sodium hydroxide in aqueous conditions.39 The amines were found to be important not only as bases, but also as capping agents which controlled the NP growth and size. This allowed for a somewhat narrower degree of polydispersity in the particles than is typically seen using aqueous syntheses. This method has to be conducted under reflux rather than at room temperature, however. This is most likely needed in order to ensure that each metal cation underwent hydrolysis at the same rate, as otherwise the product would have consisted of particles of various phases.

1.3 Experimental

1.3.1 Instrumental methods

Powder X-ray diffraction (XRD) measurements were done with a

PANalytical X’Pert Pro Bragg-Brentano powder diffractometer equipped with an

X’Celerator detector, a diffracted beam Ni-filter, and a Cu Kα X-ray source.

Transmission electron microscopy (TEM) imaging was done on two separate instruments: a Hitachi H 7000 transmission electron microscope and an

FEI Tecnai TF20 TEM instrument at an accelerating voltage of 200 kV. Particle samples were dispersed in methanol and dropcast onto 400 mesh carbon coated copper grids. FT-IR spectra were obtained using powdered samples dispersed in

KBr pellets and analyzed with a Bruker TENSOR 27.

Data for particle size histograms were collected through analysis of TEM images using Image J® software40.

1.3.2 Synthesis of bare Mn-ONPs

The synthesis of Mn3O4 nanoparticles was performed following the

1 previously reported procedure for Fe3O4 nanoparticles, with slight modifications.

In a typical synthesis, Mn(acac)3 (704 mg, 2 mmol) was dissolved in a mixture of

50 mL ethanol and 25 mL water. The resulting black solution was purged with nitrogen for one hour. The purged solution was then heated to reflux. To this, 25 mL of a nitrogen-purged aqueous solution containing NaBH4 (75.7 mg, 2 mmol) was added with 1000 rpm mechanical stirring under a nitrogen atmosphere and

12 continued refluxing. Upon addition of the NaBH4 the dark solution changed to a lighter brown colour. Over the period of an hour it slowly turned to a dark, opaque brown, indicating the formation of the nanoparticles. After one hour the reaction was left to cool to room temperature. The particles were then isolated via centrifugation at 4000 rpm for 5 minutes, repeatedly washed with water and ethanol to remove any unreacted precursors, and then air-dried. The final product is a flakey black powder.

Reactions were also done in which the molar ratios between the

Mn(acac)3 and NaBH4 reactants were varied in order to see the effect on the products. The previous report on the synthesis of IONPs gives an ideal molar ratio of 1:10 Mn(acac)3:NaBH4. Lower ratios did not result in fully crystalline particles. This was not the case with Mn-ONPs, as ratios as low as 1:1 still gave highly crystalline particles. The results shown below will focus primarily on the

1:1 reactant ratio. Powder XRD and FT-IR data will subsequently be provided to show that higher ratios were investigated and that no measurable change to particle crystallinity was observed.

1.3.3 Synthesis of Mn-ONPs-APTES

The reaction conditions were set-up and begun as above for bare particles. After one hour the reaction was left to cool to room temperature after which point APTES (0.5 mL, 2 mmol) was added directly to the reaction mixture under continued mechanical stirring. The resulting dispersion was left to mix for

20 hours. The particles were then isolated via centrifugation at 4000 rpm for 5 minutes, repeatedly washed with water and ethanol to remove any unreacted precursor, and then either air-dried (yielding a black powder)or redispersed in water (yielding a stable dark-brown dispersion).

1.3.4 Synthesis of Mn-ONPs-OA

The reaction conditions were set-up and begun as above for bare particles. After one hour the reaction was left to cool to room temperature after which point oleic acid (2 mL, 6 mmol) was added directly to the reaction mixture under continued mechanical stirring. The resulting dispersion was left to mix for 3 hours. The particles were then isolated via centrifugation at 4000 rpm for 5 minutes, repeatedly washed with water and ethanol to remove any unreacted precursor, and then either air-dried (yielding a black powder) or redispersed in non-polar solvent (yielding a stable brown dispersion).

1.3.5 Synthesis of Cu-ONPs

In a typical synthesis, Cu(acac)2 (130 mg, 0.5 mmol) was dissolved in 60 mL of ethanol. This mixture was heated slightly to induce solvation, and then purged with nitrogen for one hour. To this, 25 mL of a nitrogen purged aqueous solution containing NaBH4 (190 mg, 5 mmol) was added at room temperature with 1000 rpm mechanical stirring under a nitrogen atmosphere. Upon addition of the NaBH4 the initially blue copper solution went through a quick series of colour changes, from yellow to brown and finally to black, indicating the formation of the

14 nanoparticles. The black particles tended to agglomerate and precipitate out of solution, collecting on the sides of the flask. Total reaction time was 10 minutes.

The particles were collected with a centrifuge, washed 3 times with water to remove any precursors and then air dried. The final product is a fine black powder.

1.3.6 Synthesis of Co-ONPs

In a typical synthesis, Co(acac)3 (710 mg, 2 mmol) was dissolved in a mixture containing 40 mL of water and 50 mL of ethanol. The resulting solution was purged with nitrogen for 1 hour. This solution was then heated to reflux. To this 10 mL of degassed water containing NaBH4 (750 mg, 20 mmol) was added with 1000 rpm mechanical stirring under a nitrogen atmosphere and continued refluxing. After one hour, the reaction was cooled and the black precipitate was washed several times with water and ethanol, isolated via centrifuge, and dried under air. This dried black powder was then calcinated at 500 °C under air for three hours, resulting in the final Co-ONPs.

1.3.7 Synthesis of MFe2O4 NPs

In a typical synthesis, Fe(acac)3 and Mn(acac)3 were dissolved in a mixture containing 40 mL of water and 50 mL of ethanol in various molar ratios totalling 2 mmol. The resulting solution was purged with nitrogen for 1 hour. This solution was then heated to reflux. To this 10 mL of degassed water containing

NaBH4 (750 mg, 20 mmol) was added with 1000 rpm mechanical stirring under a

15 nitrogen atmosphere and continued refluxing. After one hour, the reaction was cooled and the black precipitate was washed several times with water and ethanol, isolated via centrifuge, and dried under air.

1.4 Results and Discussion

1.4.1 Manganese Oxide NPs (Mn-ONPs)

Figure 1: Powder XRD pattern for bare Mn-ONPs.

Figure 1 shows the powder X-ray diffraction (XRD) pattern for bare Mn-ONPs.

The diffraction pattern can be indexed to hausmannite Mn3O4 (ICDD ref. number

00-018-0803). There is no indication of impurities due to other phases. The

16 pattern shows broadening of the peaks, a characteristic feature of nano-sized materials.

Figure 2: A) and B) show TEM images of Mn-ONPs. C) shows SAED pattern of Mn-ONPs with peaks indexed to hausmannite (Mn3O4).

Figure 3: Histogram of Mn-ONP sizes as evaluated from TEM images.

Figure 2 shows transmission electron microscopy (TEM) images of bare Mn-

ONPs. Since the particles are not functionalized in any way they tend to agglomerate; discrete particles can still be observed, however. Figure 2B shows a high resolution image, revealing a lattice plane spacing (d) of ~ 2.7 Å which is indicative of (103) planes. Figure 2C shows a selected area electron diffraction

(SAED) pattern which can be indexed to a tetragonal structure indicative of

Mn3O4. Figure 3 shows a histogram detailing the frequency of particle sizes measured from TEM images, done using Image J® software.

2.1

1.9 A) bare MnONPs

1.7

1.5 B) MnONP-OA 1.3 1.1

0.9 Intensity (arb. units)

C) MnONP-APTES 0.7

0.5

0.3 3400 2400 1400 400 -1 wavenumber (cm )

Figure 4: FT-IR spectra of A) bare, B) oleic acid coated, and C) APTES coated Mn-ONPs.

Figure 4 shows the FT-IR spectra for the bare and functionalized Mn-ONPs. For the bare particles, the significant peaks are found at 638 cm-1, 532 cm-1 and 412 cm-1; these correspond to Mn-O stretching in tetrahedral sites, octahedral sites, and the vibration of Mn3+ respectively.41 This data indicates that the particles are pure Mn3O4, which is in agreement with that found with XRD. The Mn-ONPs-

APTES show a broad peak at 1000-1100 cm-1, indicative of Si-O-Si stretching.

Oleic acid coated samples show strong C-H stretching at 2929 cm-1 and 2858 cm-1, and a carboxylate stretch at 1558 cm-1. The proper functionalization of the particles is further demonstrated through the ability to make stable dispersions of particles in various solvents, as seen in Figure 5.

Figure 5: Images of (from left to right) APTES-MnONPs in water with hexane layered on top, APTES-MnONPs in water, OA-MnONPs in hexane, and OA- MnONPs in hexane with water layered below.

The polar amine terminal groups on the Mn-ONPs-APTES allow the particles to be dispersed in water. The non-polar terminal groups on the Mn-ONPs-OA sample allow the particles to be dispersed in non-polar media, such as hexane, and stay in such media even in the presence of water. In both cases the particle dispersions remain stable for days with minimal settling.

Magnetic properties of bare Mn-ONPs were characterized with a superconducting quantum interference device (SQUID). Figure 6 shows hysteresis curves of induced magnetization (M) versus an applied field (H) at 300

K and 5 K. The Mn-ONPs show paramagnetic behavior at 300 K, while at 5 K show ferromagnetic properties. The maximum magnetization (Mmax) of Mn-ONPs at 5 K is approximately 38 emu/g in a field up to 60 000 Oe; the saturation magnetization (Ms) was not reached under these conditions. These properties

42 are typical of nanosized particles of Mn3O4.

Figure 6: Magnetic hysteresis loops of bare Mn-ONPs conducted at 300 K (top) and 10 K (bottom).

The results above were conducted on particles synthesized using a 1:1 molar ratio between Mn(acac)3 and NaBH4. Initial reactions were conducted with higher amounts of NaBH4, as had been done in earlier work on iron oxide NPs.

Further optimization of the procedure demonstrated this, however, was not necessary. Figure 7 shows powder XRD patterns of Mn-ONPs made with 1:5,

1:10, and 1:20 molar ratios of Mn(acac)3 and NaBH4. Each pattern can be indexed to pure phase Mn3O4. Figure 8 shows FT-IR spectra of Mn-ONPs made with 1:1 and 1:10 molar ratios of the Mn(acac)3 and NaBH4. This is intended to show the presence of the same peaks at approximately 638 cm-1, 532 cm-1 and

412 cm-1, demonstrating again that the same crystal phase is present regardless of the amount of NaBH4 used.

Figure 7: Powder XRD spectra of Mn-ONPs made with: A) 1:5, B) 1:10, and C) 1:20 molar ratios of Mn(acac)3 and NaBH4.

Figure 8: FT-IR spectra of Mn-ONPs made with a 1:1 molar ratio (blue, top) and 1:10 molar ratio (red, bottom)of Mn(acac)3 and NaBH4.

As noted above, Mn-ONPs in general continue to be investigated for their use in bioapplications (eg. as MRI contrast agents, drug delivery vehicles, etc.).

Such applications require these particles to be dispersed in aqueous media

(accomplished in this case through functionalization with hydrophilic APTES), and also require the particles to exhibit limited cellular toxicity. Cell viability studies on Mn-ONPs-APTES were conducted by Zhizhi Sun at the University of

Manitoba, in Winnipeg, MB, Canada. Brain endothelial (bEnd.3), liver epithelial

(HepG2), intestine epithelial (Caco2) human cell lines were exposed to these particles at various concentrations for the period of 24 hours. The particles

24 showed no significant toxicity to any type of cell at any of the concentrations used. The results are summarized in Figure 9.

1.4.2 Copper/Copper oxide NPs (Cu-ONPs)

Figure 10 shows the powder x-ray diffraction (XRD) pattern for bare Cu-

CuONPs. The spectrum can be indexed to a combination of pure Cu (ICDD ref. #

00-003-1005), which is a cubic crystal structure, and cuprite Cu2O (ICDD ref. #

00-005-0667), which is also a cubic crystal structure. The broad peaks are characteristic of nano-sized material.

Figure 10: Powder XRD of Cu-ONPs. Peaks highlighted in red are indexed to Cu2O, and those in blue are indexed to elemental Cu.

Figure 11 shows TEM images for these particles. As with our other particles there is a tendency to agglomerate since no coating is present, but discrete particles can still be seen. Figure 11B shows a high resolution image, revealing a lattice plane spacing (d) of ~ 2.4 Å which is indicative of (111) planes for Cu2O. Figure 11C shows an SAED pattern which can be indexed to the cubic structures of both Cu2O and elemental Cu. Figure 12 shows a histogram detailing the frequency of particle sizes measured from TEM images, done using

Image J® software.

Figure 11: A) and B) show TEM images of Cu-ONPs. C) shows SAED micrograph of Cu-ONPs.

Figure 12: Histogram of Cu-ONP sizes based on analysis of TEM images.

For additional confirmation about the nature of the particles we used FT-

IR, shown in Figure 13. The spectrum shows a peak centred at 628 cm-1, indicative of a Cu+1—O stretch.43 This is in agreement with the XRD data

+2 showing the presence of Cu2O. There is no indication of Cu —O stretching, once again in agreement with the XRD data demonstrating the absence of CuO in the sample.

Figure 13: FT-IR spectrum of Cu-ONPs.

1.4.3 Co-ONPs

Unlike all the other particles synthesized through the reduction/hydrolysis method mentioned thus far, the Co-ONPs isolated directly from solution after the reaction with NaBH4 were not crystalline. In order to induce crystallinity, the particles were calcinated at in air at 500 °C for 3 hours. The final product after this step was pure phase Co3O4 (ICDD ref. 00-001-1152). Figure 14 shows powder XRD analysis of Co-ONPs before and after calcination. Although the synthesis of crystalline particles directly from solution would be preferable, this calcination requirement is common for cobalt oxide NPs made in a variety of different synthetic schemes, as noted above in the introduction.

Figure 15 contains TEM images of Co-ONPs before and after calcination.

Figure 15A shows that the initial product consists of thin, creased, amorphous sheets. There is no indication of visible crystallinity. After calcination, the product appears as aggregated crystalline domains assembled into the sheet-like structure, as seen in Figure 15B. These domains are quasi-spherical and between 5 – 10 nm in size, distributed throughout the sheets.

Figure 14: XRD patterns of Co-ONPs before (top) and after (bottom) calcination, showing the change from amorphous to crystalline Co3O4.

Figure 15: TEM images of Co-ONPs. A) before calcination, B) and C) after calcination. D) shows SAED of the particles, indexed to Co3O4.

1.4.4 Mixed metal oxide NPs

The first attempts at mixed metal oxide NPs were done by adding Mn3+ cations to the synthesis of Fe3O4 NPs, while keeping a consistent 10:1 molar ratio of NaBH4 to total metal precursors the same, as in the original IONP synthesis. This was done by adding Mn(acac)3 alongside Fe(acac)3 in various

31 molar ratios. Four separate syntheses were conducted this way, with the molar ratios of the metal precursors, Fe(acac)3:Mn(acac)3, set as 19:1, 7:1, 4:1, and

2:1, herein termed MMO1, MMO2, MMO3, and MMO4, respectively. Figure 16 shows powder XRD characterization of the materials. The particles made with

19:1, 7:1, and 4:1 ratios of Fe:Mn all show highly crystalline, single phase patterns that can be indexed to the cubic structure indicative of MFe2O4, along with broadened peaks typical of nano-sized material. In contrast, the highest ratio, 2:1, no longer shows a single crystal phase. Instead, two very broad peaks are present which cannot accurately be indexed to any particular phase. This suggests the product is not fully crystalline, a mixture of various phases, or a combination of both. Figure 17 and Figure 18 show TEM images of the particles from these syntheses. In Figure 17, the particles synthesized using 19:1, 7:1 and

4:1 ratios are shown. In each case, aggregates of quasi-spherical particles between 5 and 10 nm in size can be seen, with no difference in size or morphology of the particles evident between the three samples. Crystal facets can clearly be seen in high-resolution images, demonstrating good agreement with the XRD results showing that the material is crystalline in each case. Figure

18 shows the product from the synthesis with a 2:1 ratio. In this case, no individual particles are present, instead replaced by large agglomerations with no particular morphology. No crystal facets are evident in these images, indicating that the material is largely amorphous, as suggested with XRD analysis.

Figure 16: Powder XRD of mixed metal oxide NPs. A) Shows NPs made with a 19:1 molar ratio of Fe to Mn, B) shows NPs with a 7:1 molar ratio of Fe to Mn, C) shows a 4:1 ratio of Fe to Mn. Each A – C can be indexed to the cubic crystal inverse spinel phase indicative of Mn doping into iron oxide. D) shows NPs made with a 2:1 molar ratio of Fe to Mn. The broad peaks suggest a poorly crystallized material or a mixture of many phases.

Figure 17: TEM images of mixed metal oxide NPs MMO1 (A and B), MMO2 (C and D), and MMO3 (E and F).

Figure 18: TEM images of mixed metal oxide NPs MMO4.

Neither of the above analyses provides direct evidence that the products obtained in fact contain a mixture of Mn and Fe. The XRD results of the first three syntheses indicate that only a single phase – which can be indexed to that of magnetite, Fe3O4, or maghemite, Fe2O3 – is present in the samples. If the Mn precursor formed Mn3O4 particles separately, peaks associated with this material would also be present. As that is not the case, we can rule out the presence of separate IONPs and Mn-ONPs. This does not provide proof that Mn cations have been doped into the IONPs, however. To prove this, elemental analysis using

ICP was conducted on MMO1, MMO2, and MMO4 to measure the extent to which Mn is indeed present in the final products. MMO3 was lost before ICP could be conducted and is not included in this analysis. Table 1 summarizes the results. The concentration of Fe and Mn in each sample, as measured by ICP, is shown (in mg/mL), as well as the molar ratios between each metal.

Table 1: Concentration of Fe and Mn in different mixed metal oxide NP samples, as measured by ICP.

Sample Fe Fe (mol) Mn Mn (mol) Fe:Mn ID (mg/mL) (mg/mL) molar ratio MMO1 15.06 2.697 0.8127 1.479 18.2 : 1 x10-4 x10-5 MMO2 20.96 3.752 3.024 5.504 6.81 : 1 x10-4 x10-5 MMO4 6.509 1.166 x10-4 3.339 6.079 1.92 : 1 x10-5

The ICP analysis demonstrates that Mn is in fact present in each sample measured. This confirms, along with the XRD results demonstrating the single crystalline phase in these samples, that doping of Mn cations into iron oxide particles using the reduction/hydrolysis method is possible, at least at relatively low concentrations of Mn.

1.4.5 Reduction/hydrolysis and particle formation mechanism

The mechanism behind the formation of Fe3O4 particles via reduction and hydrolysis of Fe(acac)3 with NaBH4 was investigated in the original paper by

Yathindranath et. al. There, the following reaction pathway was proposed:

+ - - 1) B(OH)3 + 7H + 8e ↔ BH4 + 3H2O E° = -0.48 V

2) Fe3+(aq) + e- ↔ Fe2+(aq) E° = 0.77 V

3+ - 3) Fe + 3OH → Fe(OH)3

4) Fe(OH)3 → FeO(OH) + H2O

2+ - 5) Fe + 2OH → Fe(OH)2

6) 2FeO(OH) + Fe(OH)2 → Fe3O4 + 2H2O

In short, the electrochemical potential of the borohydride anion in solution, seen in equation 1), is sufficient to allow for the reduction of Fe3+ to Fe2+, seen in equation 2). The dissolution of NaBH4 also causes a quick increase in the pH of the reaction medium. This in turn allows for the formation for various hydroxide species, as in equations 3 – 5. The hydrolysis and subsequent co-precipitation of these species causes the formation of magnetite nanoparticles, similar to the mechanism behind the formation of IONPs in Massart’s method.

Similar to the properties of iron, the reduction of Mn3+ to Mn2+ is favorable

(E° = +1.51 V) under the conditions provided by aqueous borohydride. Further reduction of Mn2+ to elemental Mn is unfavorable (E° = -1.81 V). As such, the formation of Mn3O4 particles follows the same pathway as that of Fe3O4 seen above, with Mn(OH)3 and Mn(OH)2 hydrolyzing in a 2:1 ratio and precipitating out

3+ 2+ as Mn3O4(s). For, Co-ONPs, the reduction of Co to Co is favorable (E° =

+1.80 V), but unlike with Mn, further reduction of Co2+ to the elemental state (E°

= -0.29 V) is also possible under the provided conditions. The condensation and hydrolysis of hydroxide species is likely more kinetically favorable than the relatively slow reduction to elemental Co, which would explain why the final product is pure Co3O4.The electrochemical properties of Cu are more complicated under the conditions this synthesis provides. The reductions of Cu2+ to Cu+ as well as from Cu2+ to elemental Cu (E° = +0.16 V, and E° = +0.34 V,

37 respectively) are both highly favorable with aqueous borohydride. This explains the existence of both cuprous oxide and elemental copper in the final product.

1.5 Conclusions

This work summarizes attempts to expand upon the reduction/hydrolysis method for the synthesis of IONPs to that of other metal oxide nanoparticles.

Through simple modifications to the existing procedure, Mn3O4, Co3O4, and

Cu/Cu2O nanoparticles, as well as IONPs doped with Mn cations, were successfully made. Additional work by another member of the Hegmann lab,

Vinith Yathindranath, showed that this method is also applicable to the synthesis of ZnO and NiO NPs, as well as silica NPs. In addition to being capable of producing all of the above types of materials as nanoparticles, the synthesis easily allows for functionalization of the particles through simply in situ introduction of surface binding molecules, as demonstrated through the functionalization of Mn-ONPs with oleic acid and aminosilane. As such, we have developed a simple, versatile method for the creation and subsequent application of a wide variety of materials, without the need for harsh organic solvents, often at temperatures under 100 °C. The bulk of this work was published in 2013.44

1.6 References

1 Yathindranath, V.; Rebbouh, L.; Moore, D. F.; Miller, D. W.; van Lierop, J.; Hegmann, T. A Versatile Method for the Reductive, One-Pot Synthesis of Bare, Hydrophilic and Hydrophobic Magnetite Nanoparticles. Adv. Funct. Mater. 2011, 21, 1457-1464

2 Jana, N. R.; Chen, Y.; Peng, X. Size- and Shape-Controlled Magnetic (Cr, Mn, Fe, Co, Ni) Oxide Nanocrystals via a Simple and General Approach. Chem. Mater. 2004, 16, 3931-3935

3 Wang, D.; Xie, T.; Peng, Q.; Zhang, S.; Chen, J.; Li, Y. Direct Thermal Decomposition of Metal Nitrates in Octadecylamine to Metal Oxide Nanocrystals. Chem. Eur. J. 2008, 14, 2507-2513

4 Kumar, R. V.; Diamant, Y.; Gedanken, A. Sonochemical Synthesis and Characterization of Nanometer-Size Transition Metal Oxides from Metal Acetates. Chem. Mater. 2000, 12, 2301-2305

5 Smith, S. J.; Huang, B.; Liu, S.; Liu, Q.; Olsen, R. E.; Boerio-Goates, J.; Woodfield, B. F. Synthesis of metal oxide nanoparticles via a robust "solvent- deficient" method. Nanoscale 2015, 7, 144-156

6 Bae, K. H.; Lee, K.; Kim, C.; Park, T. G. Surface functionalized hollow manganese oxide nanoparticles for cancer targeted siRNA delivery and magnetic resonance imaging. Biomater. 2011, 32, 176-184

7 Gilad, A. A.; Walczak, P.; McMahon, M. T.; Na, H. B.; Lee, J. H.; An, K.; Hyeon, T.; van Zijl, Peter C. M.; Bulte, J. W. M. MR tracking of transplanted cells with positive contrast using manganese oxide nanoparticles. Magn Reson Med. 2008, 60, 1-7

8 Huang, J.; Xie, J.; Chen, K.; Bu, L. H.; Lee, S.; Cheng, Z.; Li, X. G.; Chen, X. Y. HSA coated MnO nanoparticles with prominent MRI contrast for tumor imaging. Chem. Commun. 2010, 46, 6684-6686

9 Shin, J.; Anisur, R.; Ko, M.; Im, G.; Lee, J.; Lee, I. Hollow Manganese Oxide Nanoparticles as Multifunctional Agents for Magnetic Resonance Imaging and Drug Delivery. Angew Chem Int Ed, 2009, 48, 321-324

10 Han, Y.; Chen, F.; Zhong, Z.; Ramesh, K.; Widjaja, E.; Chen, L. Synthesis and characterization of Mn3O4 and Mn2O3 nanocrystals on SBA-15: Novel combustion catalysts at low reaction temperatures. Catal. Commun. 2006, 7, 739-744

11 Ching, S.; Kriz, D. A.; Luthy, K. M.; Njagi, E. C.; Suib, S. L. Self-assembly of manganese oxide nanoparticles and hollow spheres. Catalytic activity in carbon monoxide oxidation. Chem. Commun. 2011, 47, 8286-8288

12 Seo, W. S.; Jo, H. H.; Lee, K.; Kim, B.; Oh, S. J.; Park, J. T. Size-Dependent Magnetic Properties of Colloidal Mn3O4 and MnO Nanoparticles. Angew Chem Int Ed , 2004, 43, 1115-1117

13 Yin, M.; O'Brien, S. Synthesis of monodisperse nanocrystals of manganese oxides. J. Am. Chem. Soc., 2003, 125, 10180-10181

14 Ghosh, M.; Biswas, K.; Sundaresan, A.; Rao, C. N. R. MnO and NiO nanoparticles: synthesis and magnetic properties. J. Mater. Chem. 2006, 16, 106-111

15 Yu, T.; Moon, J.; Park, J.; Park, Y. I.; Bin Na, H.; Kim, B. H.; Song, I. C.; Moon, W. K.; Hyeon, T. Various-Shaped Uniform Mn3O4 Nanocrystals Synthesized at Low Temperature in Air Atmosphere. Chem. Mater., 2009, 21, 2272-2279

16 Wang, W. Z.; Xu, C. K.; Wang, G. H.; Liu, Y. K.; Zheng, C. L. Preparation of smooth single-crystal Mn3O4 nanowires. Adv. Mater., 2002, 14, 837-840

17 Hao, X.; Zhao, J.; Li, Y.; Zhao, Y.; Ma, D.; Li, L. Mild aqueous synthesis of octahedral Mn3O4 nanocrystals with varied oxidation states. Colloids Surf., A, 2011, 374, 42-47

18 Wang, N.; Guo, L.; He, L.; Cao, X.; Chen, C.; Wang, R.; Yang, S. Facile Synthesis of Monodisperse Mn3O4 Tetragonal Nanoparticles and Their Large- Scale Assembly into Highly Regular Walls by a Simple Solution Route. Small, 2007, 3, 606-610

19 O'Brien, S.; White, B.; Yin, M.; Hall, A.; Le, D.; Stolbov, S.; Rahman, T.; Turro, N. Complete CO oxidation over Cu2O nanoparticles supported on silica gel. Nano Lett., 2006, 6, 2095-2098

20 Prucek, R.; Kvitek, L.; Panacek, A.; Vancurova, L.; Soukupova, J.; Jancik, D.; Zboril, R. Polyacrylate-assisted synthesis of stable copper nanoparticles and copper(I) oxide nanocubes with high catalytic efficiency. J. Mater. Chem., 2009, 19, 8463-8469

21 Zhang, J. T.; Liu, J. F.; Peng, Q.; Wang, X.; Li, Y. D. Nearly monodisperse Cu2O and CuO nanospheres: Preparation and applications for sensitive gas sensors. Chem. Mater., 2006, 18, 867-871

22 Yang, P. D.; Yuhas, B. D. Nanowire-Based All-Oxide Solar Cells. J. Am. Chem. Soc., 2009, 131, 3756-3761

23 Yin, M.; Wu, C. K.; Lou, Y. B.; Burda, C.; Koberstein, J. T.; Zhu, Y. M.; O'Brien, S. Copper oxide nanocrystals. J. Am. Chem. Soc., 2005, 127, 9506- 9511

24 Hung, L. I.; Tsung, C. K.; Huang, W. Y.; Yang, P. D. Room-Temperature Formation of Hollow Cu2O Nanoparticles. Adv. Mater., 2010, 22, 1910-+

25 Gou, L.; Murphy, C. J. Solution-Phase Synthesis of Cu2O Nanocubes. Nano Letters, 2002, 3, 231-234

26 Huang, W. X.; Bao, H. Z.; Zhang, W. H.; Shang, D. L.; Hua, Q.; Ma, Y. S.; Jiang, Z. Q.; Yang, J. L. Shape-Dependent Reducibility of Cuprous Oxide Nanocrystals. J. Phys. Chem. C, 2010, 114, 6676-6680

27 Wang, C. Y.; Zhou, Y.; Chen, Z. Y.; Cheng, B.; Liu, H. J.; Mo, X. Preparation of Shell-Core Cu2O-Cu Nanocomposite Particles and Cu Nanoparticles in a New Microemulsion System. J. Colloid Interface Sci., 1999, 220, 468-470

28 Dodoo-Arhin, D.; Leoni, M.; Scardi, P.; Garnier, E.; Mittiga, A. Synthesis, characterisation and stability of Cu2O nanoparticles produced via reverse micelles microemulsion. Mater. Chem. Phys., 2010, 122, 602-608

29 A) Davies, T. E.; Garcia, T.; Solsona, B.; Taylor, S. H. Nanocrystalline cobalt oxide: a catalyst for selective alkane oxidation under ambient conditions. Chem. Commun. 2006, 3417-3419 B) Poizot, P.; Laruelle, S.; Grugeon, S.; Dupont, L.; Tarascon, J. Nano-sized transition-metal oxides as negative-electrode materials for lithium-ion batteries. Nature 2000, 407, 496-499 C) Li, Y.; Tan, B.; Wu, Y. Freestanding Mesoporous Quasi-Single-Crystalline Co3O4 Nanowire Arrays. J. Am. Chem. Soc. 2006, 128, 14258-14259

30 Papis, E.; Rossi, F.; Raspanti, M.; Dalle-Donne, I.; Colombo, G.; Milzani, A.; Bernardini, G.; Gornati, R. Engineered cobalt oxide nanoparticles readily enter cells. Toxicol. Lett. 2009, 189, 253-259

31 Feng, J.; Zeng, H. C. Size-Controlled Growth of Co3O4 Nanocubes. Chem. Mater. 2003, 15, 2829-2835

32 Xu, R.; Zeng, H. C. Self-Generation of Tiered Surfactant Superstructures for One-Pot Synthesis of Co3O4 Nanocubes and Their Close- and Non-Close- Packed Organizations. Langmuir 2004, 20, 9780-9790

33 Park, J.; Shen, X.; Wang, G. Solvothermal synthesis and gas-sensing performance of Co3O4 hollow nanospheres. Sens. Actuators, B., 2009, 136, 494-498

34 Davies, T. E.; Garcia, T.; Solsona, B.; Taylor, S. H. Nanocrystalline cobalt oxide: a catalyst for selective alkane oxidation under ambient conditions. Chem. Commun. 2006, 3417-3419

35 Sun, S.; Zeng, H. Size-Controlled Synthesis of Magnetite Nanoparticles. 2002, J. Am. Chem. Soc., 124, 8204-8205

36 Sun, S.; Zeng, H.; Robinson, D. B.; Raoux, S.; Rice, P. M.; Wang, S. X.; Li, G. Monodisperse MFe2O4 (M = Fe, Co, Mn) Nanoparticles. J. Am. Chem. Soc., 2003, 126, 273-279

37 Adireddy, S.; Lin, C. K.; Palshin, V.; Dong, Y. M.; Cole, R.; Caruntu, G. Size- Controlled Synthesis of Quasi-Monodisperse Transition-Metal Ferrite Nanocrystals in Fatty Alcohol Solutions. J. Phys. Chem. C, 2009, 113, 20800- 20811

38 Liu, C.; Zou, B.; Rondinone, A. J.; Zhang, Z. J. Reverse Micelle Synthesis and Characterization of Superparamagnetic MnFe2O4 Spinel Ferrite Nanocrystallites. J. Phys. Chem. B, 2000, 104, 1141-1145

39 Pereira, C.; Pereira, A. M.; Fernandes, C.; Rocha, M.; Mendes, R.; Fern├índez-Garc├¡a, M. P.; Guedes, A.; Tavares, P. B.; Gren├¿che, J.; Ara├║jo, J. P.; Freire, C. Superparamagnetic MFe2O4 (M = Fe, Co, Mn) Nanoparticles: Tuning the Particle Size and Magnetic Properties through a Novel One-Step Coprecipitation Route. Chem. Mater. 2012, 24, 1496-1504

40 Rasband, W.S., ImageJ, U. S. National Institutes of Health, Bethesda, Maryland, USA, http://imagej.nih.gov/ij/, 1997-2014.

41 Ishii, M.; Yamanaka, T.; Nakahira, M. Infrared-Absorption Spectra and Cation Distributions in (Mn, (Mn,Fe)3o4. Solid State Commun.,1972, 11, 209

42 Seo, W. S.; Jo, H. H.; Lee, K.; Kim, B.; Oh, S. J.; Park, J. T. Size-Dependent Magnetic Properties of Colloidal Mn3O4 and MnO Nanoparticles. Angew. Chem. Int. Ed., 2004, 43, 1115-1117

43 Dodoo-Arhin, D.; Leoni, M.; Scardi, P.; Garnier, E.; Mittiga, A. Synthesis, characterisation and stability of Cu2O nanoparticles produced via reverse micelles microemulsion. Mater. Chem. Phys., 2010, 122, 602-608

44 Yathindranath, V.; Worden, M.; Sun, Z.; Miller, D. W.; Hegmann, T. A general synthesis of metal (Mn, Fe, Co, Ni, Cu, Zn) oxide and silica nanoparticles based on a low temperature reduction/hydrolysis pathway. RSC Adv. 2013, 3, 23722- 23729

Chapter 2

Aqueous synthesis for shape control and bioapplications of

iron oxide nanoparticles

2.1 Research purpose and goals

As discussed blow, in almost all cases non-spherical IONPs have been made in high temperature, non-aqueous conditions and often with the use of toxic solvents which does not directly result in particles that are dispersible in aqueous media. Post-synthesis ligand exchange procedures must be conducted to allow for the particles to be dispersed in biologically relevant media. In order to avoid such extra steps, not to mention the additional costs associated with high temperature methods and organic solvents, an aqueous method that allows for the control and alteration of the IONP morphology and permits simple functionalization with hydrophilic ligands is highly desirable.

In a recent publication we reported some first steps towards this goal.1 An aqueous reduction/hydrolysis of iron(III) chloride with sodium borohydride was conducted in the presence of Triton X100, a surfactant that when mixed with water behaves as a lyotropic liquid crystal (LLC). This synthesis resulted in the formation of 2D sheets composed of a mixture of crystalline iron and iron oxide.

These sheets, while only a few nanometers thick, were much larger in other dimensions (hundreds of nanometers) and the total size was difficult to control.

To address these issues, we report a modification of that technique which offers more precise control of particles fully “nano” in all dimensions, composed entirely of a single phase of iron oxide. A mixture of iron chloride precursors was hydrolyzed in water with sodium hydroxide in the presence of a surfactant (Triton

X45 or Triton X100). The resulting iron oxide particles are a mixture of

45 anisometric polyhedral particles with “brick-like” shapes of varying aspect ratios.

The precise ratio between particle types depends on the reaction conditions. The use of Triton X45 – which can form a lamellar phase LLC in water – allows for the formation of predominantly rectangular and rhombohedral “nanobricks” labelled

IONBsX45. The synthesis with Triton X100 - which, in a 50% mixture with water, forms a hexagonal phase LLC – allows for the formation of smaller cubic and rhombohedral particles, labelled IONBsX100. These particles can be easily functionalized with siloxane molecules, allowing for dispersal in aqueous media and thus potential use in biomedical applications.

Two such applications, as noted above, are in the areas of magnetic hyperthermia and MRI contrast enhancement. The former technique involves exposing magnetic particles (either ferro- or superparamagnetic, depending on the material) to an external AC magnetic field, the energy of which is converted to heat in the particles through Brownian and Néel relaxation mechanisms. A number of publications have looked at IONPs in hyperthermia treatment of cancer and tumor cells, as well as infectious agents such as bacteria, both in vivo and in vitro, with promising results. Particle size and shape can play important roles in the overall efficiency of this process. With this in mind we compared both types of NPs, functionalized with a hydrophilic siloxane, with spherical IONPs coated with the same siloxane. The results demonstrate that both kinds of anisometric NPs are significantly more efficient than comparable spherical IONPs for hyperthermia applications.

The superparamagnetic behavior of IONPs allow them to be used as T2

(negative contrast) agents in MRI diagnostics. Indeed a number of commercial

IONP-based contrast agents have been clinically available for many years. All of these are spherical particles, however. As with the properties associated with hyperthermia, particle shape can affect the efficiency of IONPs as MRI contrast agents. We analyzed the same particle systems mentioned above, at varying magnetic field strengths, for their efficacy as T2 contrast agents. Our anisometric

NPs provided both higher r2 relaxivity values as well as higher r2/r1 ratios than comparable spherical NPs, and in fact performed similarly to the most effective commercially available particles currently on market.

2.2 Introduction to synthetic methods for, physical properties of, and applications of, iron oxide nanoparticles

2.2.1 General mechanisms behind nanoparticle formation

Broadly speaking, the formation of nanoparticles composed of any material from a solution is usually described as occurring via some combination of two types of mechanisms.2 The first is the “burst nucleation” mechanism, initially described and elucidated by LaMer and Dinegar in a study on the formation and properties of colloids.3 This involves, in the first step, a rapid increase in monomer (i.e. precursor) concentration in solution as the material is injected or dissolved. At some critical concentration point, “burst nucleation”

47 occurs in which most of the free monomers condense to form seeds, or nuclei, for further particle growth. “Ostwald ripening”, a process first used to describe the growth of precipitates in inhomogeneous solid phases, is another mechanism often cited to describe nanoparticle formation. Ostwald ripening posits that larger particles grow by scavenging material from the smaller seed particles. This growth occurs because larger particles have lower surface energy than smaller ones, and so are more energetically favorable. The smaller particles in effect slowly dissolve back into solution as the larger particles coalesce. The growth will occur until a minimum size is achieved that allows a particle to remain stable in solution, (which is dependent on the nature of the specific material that is forming particles as well as the properties of the reaction medium, including the existence of capping ligands, surfactants, temperature, etc.)

Figure 19: Schematic diagram showing classical and non-classical crystallization. A shows the classical mechanism to create single crystalline particles. B shows an iso-oriented crystal upon fusing. C shows a mesocrystal composed of small particles coated in an organic surface stabilizer. This image has been modified 4 from Niederberger and Cölfen.

The above mechanisms offer general descriptions on the formation of roughly spherical, homogeneous particles. Niederberger and Cölfen term these mechanisms as “classical” particle formation pathways, and describe alternative

“non-classical” pathways which help account for more complex particle morphologies and size distributions.4 Whereas the classical pathways are (when

49 discussing the formation of inorganic crystals) ion-mediated, the non-classical pathways are particle mediated – that is, they involve the coalescence of particles themselves into larger structures. One such pathway they term “oriented attachment”, wherein two or more nanoparticles fuse along the same crystal planes to form a larger “iso-oriented” crystal. The second pathway they discuss involves particles coated with some sort of organic stabilizing ligand. In this scenario, these particles can combine to form a larger mesocrystal via a process they term “mesoscale assembly”. These larger structures can either remain as combined inorganic-organic mesocrystals, or, depending on the nature of the organic material and reaction conditions, further fuse into single iso-oriented crystals. A visual representation of these crystallization processes is shown in

Figure 19.

Additional growth mechanisms for anisometric particles have been discussed in the literature. Z. Peng and X. Peng investigated the mechanism behind the formation of CdSe nanorods.5 They found that, at certain monomer concentrations, the growth and shape evolution of the particles was controlled by what they term “1-dimension to 2-dimension intraparticle ripening”. This growth mechanism only occurs when the bulk energy of the monomers in solution is lower than that of one of the surfaces of the crystal (but not all surfaces, otherwise no growth will occur at all). Under these conditions, the total surface energy of the particle is almost equal to that of the bulk solution, so no new monomers from the solution coalesce onto the particle. Instead, the monomers

50 on the single, high energy surface move to the lower energy surface until total surface equilibrium is reached. This is in contrast to classical Ostwald ripening which occurs by the diffusion of monomers between particles (interparticle diffusion). There are a few examples of this type of mechanism being explicitly cited as the formation pathway for various anisometric particles, although, because it only occurs under specific particle/monomer concentration conditions, it cannot account for all types of non-spherical particles.

2.2.2 Wet chemical synthesis of quasi-spherical IONPs

Iron oxide nanoparticles (IONPs, composed of either magnetite, Fe3O4, or maghemite, -Fe2O3) have been used for several decades in many disparate fields, including environmental remediation6, energy storage7, and catalysis8. In addition, an increasingly prominent area of investigation is in biomedical applications, including drug delivery, MRI contrast enhancement, and magnetic hyperthermia.9 The effects of the surface chemistry of functionalized IONPs on such biological and medical applications are fairly well established, as demonstrated in various reviews on the topic cited above (and discussed in more detail in Chapter 3). A frequently overlooked aspect of these reviews, however, is the effect that particle morphology may have on these applications. This is unsurprising, as the vast majority of publications deal with very similar core shapes and sizes, almost certainly because most synthetic methods for creating

IONPs yield particles of similar morphology (i.e. quasi-spheres in the range of a few to tens of nanometers).

One of the oldest and still most common methods for creating iron oxide nanoparticles is the so-called coprecipitation method developed by Massart. In this simple aqueous method, a mixture of Fe2+ and Fe3+ salts are hydrolyzed

10 under basic conditions in water, yielding nanoparticles of Fe3O4. The size of the particles can be roughly controlled through changes to the pH or by altering the iron salt precursors. However, while the particles are in the range of tens of nanometers size distribution remains an issue. Also, the particles must be stabilized by additional ligands in order to become dispersible in a solvent.

Without modifications this method does not offer any direct means to control particle morphology. Another aqueous method was developed recently in the

Hegmann lab, and discussed previous in Chapter 1 of this dissertation. In this method iron(III) acetylacetonate is reacted with sodium borohydride in a mixture of ethanol and water to produce pure Fe3O4 particles via a reduction/hydrolysis mechanism.11 The main benefit of this method over Massart’s method is that there is no need to precisely balance the molar ratio of two iron salts in order to obtain the desired product. This method still requires additional ligands added post-synthesis in order to stabilize the particles.

Hydrothermal methods also exist for IONP synthesis. Wang et al. developed a general method for the creation of monodisperse nanocrystals, including magnetite particles.12 The authors describe the procedure as involving the reduction of metal ions by ethanol at the interfaces between a solid metal linoleate, a liquid ethanol-linoleic acid phase and a water-ethanol solution phase

52 under hydrothermal conditions. Highly monodisperse Fe3O4 particles under 10 nm can be synthesized in this way.

In general non-aqueous methods offer a greater degree of control over particle size and homogeneity. One of the earliest and most widely cited examples, reported by Sun et al., involves iron(III) acetylacetonate reacted at 265

°C in a mixture of phenyl ether, an alcohol, oleic acid and oleylamine, with the latter two compounds acting as in-situ stabilizing ligands.13 This method produces highly monodisperse, pure magnetite particles below 20 nm in size.

Another popular method involves a reaction between an iron salt and sodium oleate to synthesis iron oleate, which can then be used as a precursor to oleate- coated magnetite nanoparticles via thermal decomposition in a high boiling solvent such as octadecene.14 This basic strategy of thermal decomposition of an iron precursor in a high boiling point solvent has been repeated and modified a number of times.15 While these various strategies allow for the formation of highly monodisperse particles in a wide range of sizes, the stabilizing agents in these reactions only allow for the particles to be dispersed in non-polar solvents. Post- synthesis ligand exchange procedures must be conducted to allow for the particles to be dispersed in biologically relevant media.

In a review on the general topic of magnetic nanoparticles, Lu et al. offer a summary of the benefits and drawbacks of the various synthetic methods which can also apply specifically to iron oxide particles.16 For the purposes of this

53 research, the main message from this comparison is that aqueous syntheses are the simplest and most energy efficient methods for creating IONPs, but in general they do not offer a means to control particle shape. Some modifications must be done to accomplish shape control in an aqueous synthetic method.

2.2.3 Synthesis of non-spherical IONPs

Although spherical particles are by far the most common, there are several methods available which produce non-spherical IONPs of a variety of shapes and sizes. Kim et al. report a synthesis of cubic magnetite particles

17 ranging in size from 20 to 160 nm. In this procedure Fe(acac)3 was added to a mixture of benzyl ether and oleic acid. The mixture was then heated to 290 °C and held there for anywhere from 30 minutes to 2 hours, with longer reaction times yielding larger particles. The particles were isolated by centrifugation and could be redispersed in non-polar solvent. In a similar procedure, Yang et al. synthesized magnetite nanocubes by heating Fe(acac)3 in a mixture of oleic acid, oleylamine, 1,2-hexanediol and benzyl ether.18 This method allowed them to obtain monodisperse nanocubes as small as 6.5 nm. Guardia et al. demonstrated somewhat tighter control on the size and distribution of nanocubes at a lower temperature by replacing oleic acid with decanoic acid.19 They theorized that this was the case because iron decanoate is less stable than iron oleate, and thus could be decomposed at a lower temperature.

Rod-like iron oxide particles have also been made. Lee et al. developed a thermal decomposition reaction of Fe(CO)5 in a mixture of DMF and various

20 imidazolium-based ionic liquids to create -Fe2O3 nanoparticles. Depending on the reaction conditions and specific ionic liquid used the particles could be short

“bars” about 50 nm long, or longer “wires” several hundred nanometers long.

Palchudhury et al. performed a relatively low temperature (150 °C) decomposition reaction of iron oleate to create what they termed “iron oxide nanowhiskers” approximately 50 nm long.21 These nanowhiskers are poorly crystalized, however. Moving away from thermal decomposition reactions, Woo et al. reported a synthesis of maghemite nanorods using a microemulsion technique involving FeCl3 mixed with water, oleic acid, and propylene oxide in an excess of benzyl ether.22 The particles formed by this process were amorphous and had to be treated at high temperature to induce formation of a Fe2O3 crystal phase. Perhaps the only method currently available that allows for some degree of control and tuning over iron oxide nanorod size was published recently by Bao et. al. They started by synthesizing iron oleate which, under normal high temperature decomposition conditions, would simply yield spherical particles.

They found, however, that repeated washing of the iron oleate complex in methanol before the thermal decomposition process allowed for the formation of

23 anisotropic, rod-like particles of -Fe2O3. They theorized that the methanol wash removed weakly-bound unidentate oleate ligands, leaving behind only bridging bidentate ligands which force a rod-like geometry on the iron oleate precursor.

Variations in reaction temperature and time allowed for some control over the length and width of the nanorods. A recent paper by Zhou et al. provides the most diverse set of IONP shapes through slight modifications to a synthesis similar to Bao nanorod procedure above. Through simple variations in the heating rate and ratio of precursors in a thermal decomposition of iron oleate in oleic acid and octadecene, they synthesized iron oxide plates, octahedrons, tetrahedrons, and cubes.24

In every case mentioned above, non-spherical IONPs have been made under high temperature, non-aqueous conditions. This means that for any hope of applying particles made with any of these methods some sort of ligand exchange process must be done, which necessarily adds another layer of complexity to the synthetic procedure. From this standpoint, not to mention the additional costs associated with high temperature methods and organic solvents, an aqueous method that allows for the control and alteration of IONP morphology is highly desirable.

2.2.4 Physical properties of IONPs and the effect of particle shape therein

Iron oxide as a general term can refer to a number of different crystal structures. This discussion will focus on magnetite, Fe3O4, and maghemite,

Fe2O3. Both materials in the bulk exist in the cubic spinel crystal structure at room temperature.25 Magnetite has an inverse spinel structure, with oxygen

3+ atoms in a ccp arrangement within which Td sites are occupied by Fe , and Oh

56 sites are randomly occupied by Fe2+ and Fe3+.26 Maghemite has a normal spinel structure, with the precise ratio of vacancies to Fe3+ ions dependent on particle size. Both materials in the bulk are ferrimagnetic at room temperature; magnetite has a Curie temperature of ~850 K. while maghemite has a Curie temperature between 820 and 986 K. Below approximately 20 nm (with the precise size dependent upon a number of factors including particle shape and influences of surface effects like site vacancies and binding ligands) both materials can exhibit superparamagnetic behavior, with blocking temperatures (the temperature at which the material goes from ferromagnetic to superparamagnetic) well above room temperature.27 Superparamagnetism describes a state in which a particle of a ferro- or ferrimagnetic material is smaller than the material’s magnetic domains. In such a state, an assembly of particles with single magnetic domains behaves like a single paramagnetic atom – that is, no remnant magnetization remains in the material after the removal of an external magnetic field.28

In addition to size, the shape of a magnetic particle can also have a significant effect on the magnetic properties.29 In bulk materials, the shape of the magnetic domains affects the saturation magnetization of the material depending on how the shape aligns with both the external field and internal axes of the crystalline structure. This is an effect of the magnetic anisotropy of a material.

The crystal structure of a material and the shape and overlap of electron orbitals are necessarily related. This relationship causes electron spins, and thus magnetic moments, to exist in a lower energy state when they are aligned along

57 particular crystal axes. The lowest energy axes are termed the “easy axes”; the material is easier to magnetize in these directions than other ones.30 In nanosized materials, particle shape becomes an important factor since a particle itself may define an entire magnetic domain. The surface area of a material can also affect magnetization. The surface of a magnetic domain has a lower spin- spin exchange coupling energy since magnetic dipoles near the surface are not coupled with as many neighbouring dipoles as those in the interior of the domain.29 This effect is negligible in bulk materials, but in nanosized materials

(with inherently large surface area-to-volume ratios) this disordered surface layer becomes important.

A few studies have been done that specifically compare the magnetic properties of IONPs of various shapes with the intention of “tuning” the properties of the particles. Noh et al. compared the saturation magnetization values of iron oxide nanocubes and nanospheres of approximately the same size and

31 composition. They found that the nanocubes have higher Ms values. They concluded that this was due to the cube having a lower surface anisotropy which was the result of the surface of the cube being terminated by a single, low energy facet. The curved surfaces of the nanospheres, in contrast, were naturally terminated by many facets which in turn increased the overall surface anisotropy.

The magnetic properties of IONPs are the main considerations in terms of their application in biomedicine. These applications include their use as MRI

58 contrast agents, in hyperthermia treatments, and as targeted drug delivery vehicles, each of which will be discussed below.

2.2.5 The principles of MRI and the use of IONPs as contrast agents

The physical principles of magnetic resonance imaging (MRI) involve the interaction between the magnetic dipole moments of individual protons and an

32 external magnetic field, B0. In the presence of a strong B0, a small but measurable number of protons will align their magnetic moments parallel to that of B0. When aligned with a magnetic field, a proton’s magnetic moment, m, has a precession about and proportional to B0 which is defined by the Larmor precession frequency equation: ω0 = γB0, where γ is the gyromagnetic ratio, a constant. When an electromagnetic (or radiofrequency, RF) pulse proportional to the Larmor frequency is applied to the protons aligned in the field, some of this energy is absorbed and causes the net magnetic moment of aligned protons to move out of plane (defined as being in the z-plane) with B0. Once the RF pulse is turned off, the excited spins move back into plane with B0 and the absorbed RF energy is released at a lower resonant frequency. This released energy is the signal measured by the MRI instrument. Two separate time-dependent measurements can be made from this information. As the excited spins relax back into alignment with B0, the measured RF energy is the result of the magnetization in the z-plane. The rate at which this occurs is defined by the

-t/T equation: mz = m(1 – e 1), where T1 is the measured value unique to the immediate environment in which the protons are found. Some of this energy is

59 lost to the surrounding tissue, or lattice. Thus T1 is termed the “spin-lattice” or longitudinal relaxation time. Another signal arises from the relaxation of the magnetic moments in phase with each other along the xy-plane. When the initial

RF pulse tilts some of the proton magnetic moments out of alignment with B0, the moments interact with each other and cause a net vector aligned toward the xy- plane. When the RF pulse is turned off, the spins quickly lose this in phase

−t/T alignment at a rate defined by the equation: mxy = msin(ω0t + φ)e 2, where the value T2 is termed the transverse or “spin-spin” relaxation time. A visual schematic of the mechanism behind MRI and the different types of relaxation is shown in Figure 20.

Magnetic contrast agents can be used to shorten both T1 and T2 relaxation

33 times. T1 agents act through the interaction of the magnetic moment of the electron cloud of the material with the magnetic moment of protons on water molecules coordinating with the contrast material. This direct interaction between

34 the protons and the contrast agent is termed the “innersphere” mechanism. T1 contrast agents are often termed “bright field” agents because the areas in which they are located appear brighter relative to surrounding tissue. T2 agents act through the effect that the magnetic field gradient caused by the material has on protons diffusing through the field. As protons move through the field, spin-spin interactions between them are decreased, which leads to a decrease in overall dephasing time (and thus in T2 time). This indirect interaction between the protons and the contrast agent is termed the “outersphere” mechanism. T2

60 contrast agents are often termed “dark field” agents because they cause the areas in which they are located to appear darker relative to surrounding tissue.35

The efficiency of a particular contrast agent is typically reported as the relaxation rate, ri, which is the inverse of the relaxation time, Ti. and where the higher the value the greater the contrast and thus efficiency of the contrast agent.36

Figure 20: Schematic of the principles of MRI. A) shows the precession of a proton's magnetic moment (m) about that of an external field (B0) along the z- axis. B) shows the net vector moment, mz, before the RF pulse. C) Shows an RF pulse moving the net vector moment, mxy, into the xy-plane. D) shows the decrease of mxy as individual proton moments dephase (T2 relaxation), the increase of mz as the net vector realigns with the external field (T1 relaxation) and the release of RF energy.

IONPs have typically been used as T2 contrast agents. Several iron oxide based contrast agents were at various points clinically available, such as

“Ferridex”, which was sold in Europe to help MRI contrast of the liver though has now since been discontinued.37 Many studies have looked at increasing the magnetic saturation properties of IONPs, through size control or through doping with magnetic ions, so as to increase the contrast efficiency of IONPs.38 More recent papers have looked at the effect of particle morphology on contrast efficiency. Iron oxide nanocubes39, and “octapods”40 have both been shown to have r2 values many times higher than comparable spherical particles. A more systematic study was conducted by Zhou et al. in which they compared the T2 relaxivities of IONPs with a variety of shapes, including spheres, cubes, octahedrons, tetrahedrons, and multibranched cores.41 They found substantial differences in contrast efficiencies related to the interplay saturation magnetization and the effective radii of the particles (which affects the diffusion of protons through the particle’s induced magnetic field).

2.2.6 IONPs in targeted drug delivery and the effect of particle shape on cell uptake

IONPs have been investigated for many years as vehicles for targeted drug delivery, with a number of reviews available on the topic.42 The main justification for the use of magnetic particles in this type of therapy is the ability to control to some extent where the particles travel to and concentrate in a body through the use of an external magnetic field. However, in order to be most

63 effective at targeting specific areas for drug treatment, particles must be able to penetrate into and through different types of cells. A great deal of research has been done on the effect that surface ligands attached to the particles have on cell uptake properties. Most if not all such research has been done using spherical particles. A growing amount of research suggests that morphology can have a profound effect on how particles of any composition behave in and interact with different biological systems. Circulation and biodistribution can differ dramatically between spherical and oblate objects due to the different hydrodynamic behaviours involved. When distributed in blood, spherical objects tend to accumulate in the center of the flow, while non-spherical objects collect along the sides. As such, non-spherical objects can exit the main vasculature more quickly and easily.43

A number of computational studies have been done on the topic of the effects of particle shape and cellular interaction. Nangia et al. modeled particles of various shapes ranging from spheres and cubes to cone and rice-like geometries in order to investigate translocation rates through cell membranes.44

In these simulations the particles were composed of gold. The interactions between particles and the modeled cell membranes were due to shells of charge

(both positive and negative, depending on the model) on the surface of the particles, rather than any intrinsic properties of gold. As such the actual particle composition did not affect the results. The researchers found that, regardless of particle shape, negative surface charges on the particles inhibited translocation,

64 while translocation rates increased with increasing positive surface charge. When these surface properties were accounted for, the effect of shape on translocation was clear and dramatic. For the maximum assigned charge density, they found that spherical particles had a free energy barrier for translocation of approximately 60 kJ/mol, giving the second lowest barrier and thus the second highest translocation rate. While these results were superior to those calculated for most of the other shapes the rice-like particles were found to interact and pass through the cell membrane instantaneously – that is, the free energy barrier was effectively zero. The researchers explain that this is due to the anisotropic shape of the rice-like particles. This property allows these particles to maximize the attractive forces between the surface charges by changing their orientation, thereby increasing the relative surface area that interacts with the cell membrane, something that the isotropic spherical particles are unable to do.

Moving away from computational models, Zhang et al. conducted experiments in which they compared cell uptake of polystyrene nanospheres with two-dimensional “nanodisks”.45 They synthesized both kinds of particles with diameters on the long axis of roughly 20 nm, but the disks were constrained to 2

– 3 nm along the perpendicular axis. The particles were then compared in cell permeation studies on HeLa cells. The researchers found that the nanodisks preferentially associated with the cell membrane, rarely passing into the cell endoplasm. This was in stark contrast to the nanospheres which entered into the cells without accumulating along the membrane. They quantified the differences

65 in membrane permeability with a parallel artificial membrane permeability assay

(PAMPA). They concluded from this that the nanodisks were retained on the membrane at an 8-fold higher ratio as compared with the nanospheres. The researchers suggest that these differences are due to the fact that the 2- dimensional property of the nanodisks allows them to enter into and between the phosolipid bilayers that compose the cell membrane. The nanospheres, which are too large to be maintained between the bilayers, disrupt the membrane to a much greater degree which leads to endocytosis. Gratton et al. conducted a similar study, investigating the influence of particle shape and size on cellular uptake for hydrogel particles.46 They concluded that, when surface charge was taken into account, particles with the largest aspect ratios were taken up by HeLa cells four times faster than spherical particles (i.e. particles with no difference in aspect ratio).

This non-exhaustive overview demonstrates how control of a particle’s shape could help researchers in fields related to biomedical applications, for example by allowing for more target specificity, increasing circulation times or increasing cellular uptake speeds.

2.1.7 IONPs in hyperthermia treatment

When exposed to an alternating magnetic field, IONPs release heat. This is the result of losses in the energy of the induced magnetic moment of the particle as it reverses direction.47 In ferromagnetic materials, the heat generation

66 is primarily the result of hysteresis losses. In superparamagnetic materials, the heat generation is the result of two types of relaxation: Néel and Brownian. The former involves “internal friction” as the magnetic moment of the single-domain particle rapidly switches direction; the latter involves external friction as the entire particle moves and rotates in a medium. A visualization of these two mechanisms can be found in Figure 21. This heating mechanism has been exploited for use clinically in hyperthermia48, which, in a medical setting, involves localized heating

(typically of a tumor) to the point where the surrounding tissue dies.

67 of the crystal structure as the external magnetic field moves. The bottom shows Brownian relaxation, the physical movement of the entire particle to align the internal moment with the external field. This figure is adapted from Kumar and Mohammad.

The efficiency of a hyperthermia agent is given as a number called SAR

(specific absorption rate) or SLP (specific loss power), defined as follows:

SLP = (CmS/m)*(∆T/∆t) (1) where C is the specific heat of the solution in which the particles are dispersed, ms is the mass of the solution, m is the mass of the magnetic material, and ΔT/Δt is the slope of the heating curve.49 This value is in effect a measure of the amount of power lost to the surroundings normalized to the mass of the magnetic material. In order to measure this, particles are typically dispersed in aqueous media and exposed to an AC magnetic field, and any subsequent change in temperature of the media is monitored over time.

Since the mechanism for heating IONPs is dependent on the amplitude and frequency of the AC field used and the SLP value as calculated does not take these factors into account, meaningful comparisons between measurements operating under different experimental conditions becomes difficult if not outright impossible.50 A number of investigations have been done seeking ways to maximize hyperthermia performance in IONPs in general, although perhaps because of the difficulties in normalizing procedures across different experiments, much of the literature on this topic is contradictory. For example

68 particle size, and size distribution therein, can have profound effects on heating efficiency. Fortin et al. found that maghemite particles of approximately 16 nm in diameter to have the highest SLP values, with a steep decline in efficiency as the particles are made smaller or larger.51 They theorize that this is due to maximization of Neel relaxation mechanisms, and the drastic changes in efficiency with other particle sizes also demonstrates the importance of narrow size polydispersity. In contrast to this study, Rosensweig found that maghemite particles exhibit the highest heating rate at a diameter of 22 nm.52 Additionally,

Rosensweig noted substantial differences between optimal particle size for magnetite and maghemite particles, concluding that the former material exhibits the highest heating rate in particles almost half the size (14 nm diameter). Other researchers have investigated the relationship between magnetic saturation (Ms), coercivity (Hc), and SLP values. Zhao et al. looked at the hyperthermia

53 performances of IONPs with varying Ms and Hc properties. They found a direct relationship between Ms and SLP (larger Ms resulted in larger SLP), and an inverse relationship between Hc and SLP. They concluded that the inverse relationship between Hc and SLP was the result of the loss of Néel relaxation effects as the particles move from superparamagnetic to ferromagnetic in nature.

Ma et al., however, did not find such an explicit relationship between Hc and SLP, instead finding a more complicated interplay between particle size, the resulting

54 magnetic properties, and SLP values. In effect they found that a large Hc value

(indicative of ferromagnetic behavior) does not necessarily result in low SLP values if the particle size is also minimized.

As noted above, particle morphology can significantly affect the magnetic properties of a material. This should, along with the obvious effects shape can have on Brownian relaxation, make shape control another means to potentially maximize SLP values. There are, however, relatively few papers on this topic.

Martinez-Boubeta et al. compared the heating efficiency of cubic and spherical

IONPs.55 They found that, for particles of similar size and with similar magnetic

(eg. Ms and Hc) properties, cubic particles showed significantly higher SLP values. Interestingly, they also found that higher Ms values in particles of the same shape but larger size did not result in higher SLP values. Hugounenq et al. looked at assemblies of 11 nm sized IONPs into larger clusters of “nanoflowers” and the effect therein on hyperthermia efficiency.56 They found that such clusters enhance the SLP values of the material by an order or magnitude or more depending on the specific conditions. They suggest that this increase is due to magnetic exchange coupling between the aggregated particles, thereby lowering the overall magnetic and surface anisotropy. In a more general review on the topic of maximizing hyperthermia performance, Dennis and Ivkov concluded that the most important physical property that influences heating efficiency is magnetic anisotropy.57 The complicated interplay between crystal structure, particle shape, polydispersity, and external magnetic field strengths make it

70 difficult to predict beforehand what an “optimal” magnetic anisotropy would be though.

2.1.8 Lyotropic liquid crystals and Triton X surfactants

Figure 22: (Top) General structure of Triton X molecule. For X100, n = 10; for X45, n = 4 - 5. (Bottom left) Schematic representation of a lamellar phase. (Bottom right) Schematic representation of a hexagonal phase.

The term “liquid crystal” (LC) can be applied to any molecule that can exhibit properties intermediate between an isotropic liquid and a crystalline solid.

Liquids contain molecules that can freely move, and the whole system is isotropic

– that is, the molecules exhibit no long range orientation or order. Molecules in a crystal, on the other hand, are locked into a particular orientation, and the whole

71 system exhibits long range order and repeating patterns. In a liquid crystal, molecules may move more freely than in a solid crystal, but they still exhibit anisotropic ordering commonly in one or two, but sometimes all three dimensions. There are two main types of LCs: thermotropic LCs in which the phase change from isotropic liquid to anisotropic LC is caused by a change in temperature, and lyotropic liquid crystals (LLCs) in which the phase change is the result of dissolution in and interaction with a solvent.58 LLC molecules are generally surfactant-like in structure, with a polar, soluble head group and a long, non-polar tail. Spontaneous ordering of the molecules in the solvent occurs due to a combination of hydrophilic and hydrophobic interactions. When mixed with a polar solvent (eg. water) the hydrophilic head group will interact with the solvent molecules while the hydrophobic tails will segregate and aggregate away from the solvent. Although this segregation and inevitable formation of order is entropically unfavorable, it is counterbalanced by a favorable total free energy of the system. The research described herein focuses solely on the use of LLC phases.

At low concentrations of surfactant, isotropic micelles are formed. As the concentration increases the anisotropic LLC phases begin to form. Within LLCs there are six major classes of structures, or mesophases, which are formed: lamellar, hexagonal, cubic, nematic, intermediate, and gel phases. The lamellar phase is defined by the formation of bilayers, in which the behavior of the surfactant molecules is roughly analogous to lipids in cell membranes.

Hexagonal phases are defined by the formation of cylindrical tube-like structures that pack together into a hexagonal lattice. There are two types of hexagonal phases: H1, the “normal” phase, in which the alkyl chains are packed into the interior of the cylinders, and H2, the reversed phase, in which water and the polar head groups are packed inside the cylinders. Cubic phases are defined by typical cubic lattice structures (primitive, body centered, face centered). These can be formed either when quasi-spherical micelles pack together in a manner equivalent to how atoms are arranged in a ccp lattice (I type), or when the micelles aggregate further into a continuous, repeating 3D network (V type). The nematic LLC phase is defined similarly to nematic phases in thermotropic LC systems; rodlike molecules or aggregates of molecules exhibit anisotropic ordering along one direction. Intermediate phases are a broad collection of structures that have, as the name implies, properties intermediate between other

LLC phases. Lastly, the long-range structure of gel phases is very similar to that of the lamellar phase. The main distinguishing property between the two phases is, as one would suspect, the high viscosity in gel phases that prevents flow.59

There are two main types of surfactants used for the purpose of forming

LLC phases in this research. Triton X45 is a commercially available surfactant that exhibits a lamellar phase at concentrations in water between 20 – 60% up to a temperature of 333K.60 Triton X100 exhibits a hexagonal phase at concentrations in water between approximately 40 – 60%, up to a temperature of approximately 305 K.61 Figure 22 shows the structure of both compounds (n =

10 for X100, 4-5 for X45), as well as schematic diagrams of how idealized LLC molecules self-assemble in lamellar and hexagonal phases.

2.3 Experimental

2.3.1 Instrumental methods

TEM imaging was done with a FEI Tecnai TF20 TEM instrument at an accelerating voltage of 200 kV. Particle samples were dispersed in methanol and dropcast onto 400 mesh carbon coated copper grids.

Powder X-ray diffraction patterns (XRD) were measured on an X'Pert PRO diffractometer manufactured by PANalytical, Inc. (Westborough, MA, USA). The experimental setup used Bragg-Brentano geometry in θ-θ configuration, copper as a radiation source (Cu Kα radiation), and a diffracted beam curved crystal monochromator to eliminate Cu Kβ. All patterns were collected in a range of 2θ values from 10.00° to 80.00° with a step size 0.05°.

The ionic relaxivity of the iron oxide particles was tested using a pre- clinical 7.0 T (300 MHz) MRI (Bruker BioSpec 70/30USR), and a Bruker Minispec mq60 relaxometer (60 MHz). A standard inversion recovery sequence protocol was used to determine the longitudinal T1 values on each of the instruments. The transverse relaxivity (r2) of the particles was calculated as the slope of 1/T2 against iron concentration. T2 relaxation times were determined using a standard

Carr-Purcell-Meiboom-Gill spin echo sequence.

Dispersions of S-IONBsX45 and S-IONBsX100 were made by sonicating dried powder in deionized water (at concentrations of 9.88 and 9.95 mg/mL, respectively). A dispersion of S-IONPs, synthesized similarly to a previous report, was also made in this manner at a concentration of (10.01 mg/mL). In a typical hyperthermia experiment, 250 μL of particle dispersion was added to a single well from a 96-well plate. The samples were exposed to a field with an amplitude of 20 kA/m and a frequency of 2.1 MHz for 3 minutes while the temperature of the sample media was monitored using a fiber optic temperature probe (Neoptix).

The hydrodynamic radius and ζ-potential of the S-IONBs were determined using a Brookhaven Zetaplus ζ-potential-DLS measurement system. The instrument specifications include a 35 mW class 1 laser at 660 nm with a scattering angle of 90°. All dispersions were measured at a concentration of ~1 mg/mL. Results listed are an average of 3 consecutive measurements.

POM images were taken with an Olympus BX-53 equipped with a Linkam

LTS420E heating/cooling stage.

Surface functionalization was analyzed through FT-IR using KBr pellet techniques. Approximately 1 mg of dried particles were mixed with approximately

150 mg of KBr, which was then pressed into a pellet. The pellet was stored in a vacuum oven at 50 °C for several hours before analysis to remove any adsorbed water. Spectra were recorded using a Magna Nicolet-500 series FT-IR spectrometer.

The amount of surface ligands on the particles was estimated via a TA instruments TGA Q500. The heating rate was set at 10 °C/min. Powdered samples were typically dried in a vacuum oven at 50 °C for 2 hours before analysis in order to eliminate any surface water.

The magnetic properties were characterized with an RF Superconducting

Quantum Interference Device (SQUID) magnetometer (Quantum Design MPMS-

XL) with reciprocating sample transport. The field was applied between −30 to

+30 kOe at 300 K.

2.3.2 Synthesis of IONBsX45 and IONBsX100

In a typical experiment FeCl3 (2 mmol) and FeCl2∙4H2O (1 mmol) were dissolved in 20 mL of degassed water and added to a 3 neck round bottom flask under nitrogen. This reaction vessel was heated to 50 °C, at which point 25 mL of degassed Triton X surfactant (either X45 or X100, depending on the experiment) was added and mechanically stirred at 100 rpm to ensure that a homogeneous mixture was formed. The vessel was then cooled to 35 °C in the case of X45, and 30 °C in the case of X100. NaOH (30 mmol) was then dissolved in 5 mL degassed water and added to the above mixture under nitrogen and 100 rpm mechanical stirring. The thick, yellow mixture quickly turned black as the NaOH was mixed in. The reaction was left to mix for 1 hour. The black product was washed with warm water and centrifuged at 10,000 rpm several times to isolate it

76 from the surfactant, then dried under nitrogen and stored as a powder under ambient conditions.

2.2.3 Synthesis of S-IONBsX45 and S-IONBsX100

Silanized IONBs (S-IONBs) were synthesized following a modification of the above procedure. One hour after the addition of the NaOH solution, 15 mL of

EDTS (45% in water) was added via syringe directly into the reaction vessel. The reaction was left to mix for 12 hours. The product was isolated via multiple washings with a water/ethanol mixture and centrifugation at 10,000 rpm, and then dried under nitrogen. The black powder could then be stored under ambient conditions or re-dispersed in aqueous solution for further use.

2.4 Results and Discussion

2.4.1 Evaluation of particle morphology and crystallinity

Figure 23 shows the powder X-ray diffraction (XRD) patterns of bare

IONBsX100 and IONBsX45. The indexed patterns match closely to that of bulk magnetite (ICDD reference code 01–089-0691) as well as maghemite.

Figure 23: Powder XRD patterns for IONBsX100 (top) and IONBsX45 (bottom). Peaks have been indexed to match that of bulk magnetite.

The most important question concerned whether the addition of the surfactants could in fact allow for some measure of control over particle shape.

TEM imaging was thus used on the bare IONBsX45 and IONBsX100 in order to answer this question. Figure 24 shows representative high-resolution (HR)-TEM images of IONBsX45. The particles are composed primarily of rectangular and rhombohedral shapes, with a distribution of sizes of approximately 15 +/- 10 nm, and with varying aspect ratios between edge lengths therein. An average d- spacing of approximately 4.9 Å for the visible lattice fringes was determined via

78 analysis with ImageJ® software and is consistent with the d-spacing (4.842 Å) associated with the (111) lattice plane of magnetite.

Figure 24: TEM images of IONBsX45.

Figure 25 shows representative HR-TEM images of IONBsX100. These particles differ noticeably from those made with X45. They include primarily rhombohedral shapes with edges of similar lengths of approximately 10 +/- 5 nm.

The longer rectangular particles seen with X45, in which perpendicular sides

79 have varying aspect ratios, are not seen with X100. An average d-spacing of approximately 4.8 Å was determined, again close to that of the d-spacing associated with the (111) lattice plane of magnetite.

Figure 25: TEM images of IONBsX100.

Further evaluation of particle morphology is shown in Figure 26, which compares TEM images of various particles with 3D representations (shown adjacent to the relevant TEM images) of shapes to which they most likely

80 conform. The figure starts from a rhombohedral “brick”. Figure A shows the face of this rhombohedron, with B (from X45) and C (from X100) showing

Figure 26: Comparison between a representative polyhedral brick-like shape with how it may appear with changes to size and viewing position. Image A shows this rhombohedral shape as seen perpendicular to its face, while B and C show HR- TEM images of this for IONBsX45 and IONBsX100, respectively. Elongation along one edge leads to a parallelepiped as in image D. HR-TEM images of this shape for IONBsX45 are shown in figures E and F. The same shape as D seen along an edge leads to a purely rectangular shape, as in figure G. HR-TEM images H and I show examples of this shape seen in IONBsX45.

81 representative particles with this shape under TEM; D shows a larger shape in which one of the pair of edges is elongated to form a parallelepiped (3D parallelogram), with E and F showing representative TEM images of this shape from IONBsX45; G shows the same shape as D as viewed along the long edge, resulting in a rectangular shape, with H and I showing representative TEM images of this shape from IONBsX45.

We acknowledge the difficulty in precisely distinguishing between octahedral and rhombohedral shapes based on the 2D information provided by

TEM. It is possible that solely one or the other shape is formed exclusively, or even a mixture. Figure 27A and Figure 27B show HR-TEM images of

IONBsX45 and IONBsX100, respectively, along with a 3D representation of an octahedron for comparison.

Other researchers, having synthesized IONPs showing similar shape profiles under TEM, have concluded that their particles are octahedral in shape.62

However, the existence of the larger rectangular and parallelepipedal shaped particles in our synthesis suggests that the rhombohedral shape is more likely.

Mesoscale assembly – in which small particles fuse together through alignment of their crystal facets to produce larger particles – is one mechanism through which different particle shapes can be obtained. Octahedral particles could not fuse together to form these kinds of shapes. On the other hand, if two rhombohedral particles fuse together in this manner they would result in a

82 parallelepipedal shape. Figure 27E shows a schematic representation of this mechanism, and Figure 27D shows a TEM image of these two types of particles side by side. Examples of images taken with TEM at various tilt angles in order to visualize the effect of sight angle on the apparent shape of a particle are shown in Figure 28.

The Triton X surfactants form LLC phases in water. Figure 29 has POM images of each surfactant as observed under experimental conditions, which demonstrates that such phases exist before particle formation. The formation of these phases may have allowed the surfactants to behave as templates in nanoparticle synthesis. The formation of discrete, layered structures could

Figure 29: POM images under crossed polarizers showing birefringent textures of lyotropic phases. A shows an image of a 50% mixture of Triton X45 in water at 35 °C; B shows the same along with a 2:1 mixture of FeCl3 and FeCl2, which represents the condition of the reaction mixture before hydrolysis with NaOH occurs. Fingerprint textures typical of a lamellar phase can be seen in both. C shows an image of a 50% mixture of Triton X100 in water at 30 °C; D shows the same along with a 2:1 mixture of FeCl3 and FeCl2. Focal conic textures typical of a hexagonal phase can be seen in both (limited transmission due to presence of iron salts), although the inclusion of the iron precursors does lower the transition temperature slightly. contribute to the shape control seen here by constraining the direction of growth during particle formation. Other than templating, most shape control in nanomaterials results from control over material growth rate and precursor

85 concentration (i.e. selective access to precursor atoms or monomers through local concentration gradients). In general this is accomplished through some combination of variation in heating rate and the overall temperature of the reaction medium, as well as the use of selective capping agents, which preferentially adsorb onto certain crystal facets. Our modified co-precipitation method requires a much lower reaction temperature than methods using adjusted heating rate and temperature as means of shape control. However, the surfactant molecules may be acting as capping agents. Changes in the nature of capping agent functional groups (i.e. the presence or absence of certain functional groups as well as changes in the ratio between different capping agents) can lead to changes in IONP morphology. Triton X45 and X100 have similar structures, differing only in the length of the ethylene oxide chain.

Variations in the chain length of capping agents have been shown to influence shape in other metal oxide particles; the longer PEO chain length of the X100 molecule could account for the restriction in diversity of the resulting particle shapes. The longer X100 chain could also prevent smaller particles from coming close enough together to fuse and form larger particles via mesoscale assembly, which would account for the absence of larger parallelepipedal and rectangular particles in the synthesis of IONBsX100. Given the speed of the hydrolysis reaction and the dynamic nature of the reaction medium, all of the above mechanisms may contribute to the final particle shapes.

2.4.2 Magnetic measurements

Saturation magnetization (Ms) values were found to be 58 and 61 emu/g at

300 K for IONBsX45 and IONBsX100, respectively. These values are typical of crystalline IONPs on the order of tens of nanometers in size. Both types of particles were found to have low coercivity (Hc) values of 18 and 20 Oe for

IONBsX45 and IONBsX100, respectively. These results are shown in Figure 30.

Figure 30: Magnetic hysteresis curves for IONBsX100 (triangles) and IONBsX45 (squares) at 300 K. Inset shows a magnified image of the coercivity for each particle set.

2.4.3 Surface functionalization and characterization

One of the many benefits of an aqueous synthesis is the ease with which particles can be coated with hydrophilic functional groups. In the present case, the IONBs were coated with N-(trimethoxysilylpropyl)ethylenediaminetriacetate trisodium salt (EDTS) through injection of the silane solution directly into the reaction media after the synthesis. Particles coated with the above siloxane are labelled S-IONBsX. EDTS can impart a high negative surface potential on the particles, which allows them to be easily stabilized in aqueous media.

Figure 31: (Left): Picture of particle dispersions of S-IONBsX45 and S- IONBsX100. (Right): Schematic representation of the reaction scheme resulting in coated brick-like particles.

Additionally, previous investigations on the cell viability and uptake properties of

EDTS-coated spherical IONPs have demonstrated the usefulness of this molecule as a functionalizing agent for IONPs in bioapplications (see Chapter 3).

The dried particles were analyzed by FT-IR and TGA to confirm the presence and binding of the EDTS surface coating (shown in Figure 32 and Figure 33,

88 respectively). The dried particles could be re-suspended in water with mild sonication, and remained stable for weeks without any sign of precipitation.

Figure 32: FT-IR spectra for S-IONBsX45 (top) and S-IONBsX100 (bottom). The broad peaks centered at ~3400 cm-1 correspond to O-H stretching; the sharp peaks at ~2915 and ~2850 cm-1 correspond to C-H stretching; the broad peaks at ~1600 cm-1 are indicative of COO- stretching; the broad peaks centered around ~1000 cm-1 are due to Si-O-R stretching; and the large asymmetric peaks at ~590 cm-1 correspond to stretching modes associated with Fe2+-O and Fe3+-O.

Figure 33: TGA plots for S-IONBsX45 (top) and S-IONBsX100 (bottom) showing weight loss profiles of ~45% in both cases.

Characterization of the particle suspensions was done with dynamic light scattering (DLS) and ζ-potential measurements. Figure 31 shows a set of images of S-IONBsX suspended in water along with a schematic representation of the reaction and the surface coating of a model particle. A summary of their properties in solution can be found in Table 2. For comparison, quasi-spherical

IONPs were synthesized and coated with the same siloxane.

Table 2: Physiochemical properties of silanized particles in water.

Material DLS (nm) Zeta (mV) pH

S-IONBsX45 50.9 +/- 1.4 -44.2 +/- 2.4 9.8

S-IONBsX100 64.0 +/- 0.4 -38.1 +/- 1.0 9.4

S-IONPs 30.3 +/- 1.0 -36.8 +/- 2.8 10.2

These particles are labeled S-IONPs. Details about these particles can be found in the Chapter 3 discussion on the surface chemistry of IONPs. It should be noted that while the particles were added to distilled water rather than a buffer, the pH of the media was basic due to the nature of the surface coating itself (which contains the conjugate base of a carboxylic acid). Both particle size and surface potential when in solution are highly dependent on pH, as well as ionic strength, which may explain the slight differences between the X45 and

X100 particles. These particle dispersions were then used for both MRI relaxivity and hyperthermia measurements described below.

2.4.4 MRI relaxivity measurements

Table 3: Relaxivity values at different field strengths for EDTS coated particles.

@ 1.5 T @ 7 T r r r r 1 2 r /r 1 2 r /r (mM−1 s−1) (mM−1 s−1) 2 1 (mM−1 s−1) (mM−1 s−1) 2 1 S-IONPs 8.8 29.6 3.4 2.5 43.9 17.7 S-IONBsX45 12.2 285 23.4 1.4 423 298 4.3 599 139 S-IONBsX100 11.8 247 21.0

The transverse (r2) and longitudinal (r1) relaxivities of S-IONBsX45 and S-

IONBsX100 are shown in Table 3. Figure 34 and Figure 35 show relaxation time vs. Fe concentration plots used to calculate r2 and r1. The properties of quasi-spherical S-IONPs were also measured for comparison. Properties at both

1.5 T and 7 T were measured. The r1 values for the three particle types are low and roughly similar. Since iron oxide materials are not typically considered viable

T1 contrast agents, this is unsurprising. The r2 values of both types of anisometric particle systems are an order of magnitude higher than the spherical IONPs, however. This is true at both 1.5 and 7 T field strengths. The ratio of r2/r1 – which is a measure of the efficiency of the contrast agent, where a low ratio is favorable for T1 agents, and a high ratio is favorable for T2 agents – is also shown. These results demonstrate that the polyhedral NBs are much more efficient T2 contrast agents as compared with the spherical NPs coated with the same surface ligands.

Figure 34: Plots used to determine relaxivity values measured at 1.5 T for: A S- IONBsX100, B S-IONBsX45, C S-IONPs. y-axes show the inverse of the relaxation time; x-axes show Fe concentration in mM.

Figure 35: Plots used to determine relaxivity values measured at 7 T for: A S- IONBsX100, B S-IONBsX45, C S-IONPs. y-axes show the inverse of the relaxation time; x-axes show Fe concentration in mM.

Table 4: Comparison between commercially available and literature reported iron oxide based contrast agents.

r2 r2/r1 Field references (mM−1 s−1) strength Feridex 41, 98.3 8.7, 4.1 1.5 T 63, 64 Resovist 61, 151 7.1, 5.9 1.5 T 64, 65 MPIO 169 - 1.5 T 65 FION 324 - 1.5 T 65 VNPs 140.28 144.6 3 T 66 Octapod IONPs 679 - 7 T 67

Table 4 shows the relaxivity properties of selected iron oxide based contrast agents reported in the literature, both commercial and otherwise.

Feridex and Combidex are two types of commercial IONP-based contrast agents, often used for baseline comparison. Their specific properties vary depending upon the literature cited (two examples are given), but in all cases both S-IONB systems show an order of magnitude higher r2 and r2/r1 values. MPIOs

(micrometer-sized iron oxide particles), aggregates of nanometer sized IONPs, have been shown to increase the r2 value as compared with dispersions of non- aggregated particles. Even still, our polyhedral particles have nearly double the r2 value seen with MPIOs. Three other types of particles that do show higher r2 values have been included as well. FIONs (ferrimagnetic iron oxide nanocubes)

−1 −1 have a reported r2 value of 324 mM s measured at 1.5 T; VNPs (virus-based

−1 −1 nanoparticles) have a reported value of 140.28 mM s , with a favorable r2/r1 value of 144.6, measured at 3 T; and octapod IONPs have a reported value of

679 mM−1 s−1 measured at 7 T. Each system contains highly anisometric

95 particles, demonstrating the important effect that particle morphology has on their efficiency in MRI applications. In all above cases, particles were synthesized using variations on thermal decomposition methods. We contend that since the relaxivity values reported are not significantly different (324 vs. 285 mM−1 s−1 for

FIONs vs. S-IONBsX45; 679 vs. 599 mM−1 s−1 for octapod IONPs vs. S-

IONBsX100), the benefits associated with our synthesis allow our particles to be viable alternatives.

2.4.5 Hyperthermia measurements

The hyperthermia performance of the S-IONBs was evaluated by exposing particle dispersions to an AC magnetic field on a custom instrument

(described in a previous report). The SLP (specific loss power, also often referred to as SAR, specific absorption rate) values of the samples were calculated in order to measure the efficiency of the particles at converting the magnetic field energy to heat with respect to the amount of iron in each sample. This was calculated using equation 1 above, where C is the specific heat of the solution

(taken to be the same as water, 4.186 J/g°C), ms is the mass of the solution, m is the mass of the magnetic material (in this case the mass of Fe, established by

ICP analysis), and ΔT/Δt is the slope of the heating curve. SLP values are highly dependent on the strength of the magnetic field, the nature of the media, how one chooses to evaluate the ΔT/Δt curve, and even the placement of the temperature probe. As such, comparisons between materials used by different researchers with different experimental setups and protocols are problematic. In

96 order to obtain an internal comparison the IONBs were compared with quasi- spherical IONPs to act as a kind of internal standard, as with the MRI relaxivity measurements. Figure 36 shows the temperature vs. time curves for the three types of particle dispersions measured, along with the calculated SLP values.

The graphs clearly show that the quasi-spherical particles (S-IONPs) perform the worst, giving rise to a mere 2.6 °C temperature change over 3 minutes, yielding an SLP value of 32.7 W/g. The IONBs show an order of magnitude more efficient heating response, with a 28.3 °C temperature change and 166 W/g SLP value for

S-IONBsX100, and a 29.2 °C temperature change and a 415 W/g SLP value for

S-IONBsX45.

Previous reports have suggested that single crystalline particles on the order of 18 nm in size are the most efficient for hyperthermia applications. This may account for the low SLP value for the S-IONPs, which are <10 nm according to the literature. However, the same report suggests that low poly-dispersity in particle size yields greater efficiency. This should preclude higher SLP values for

S-IONBsX45 and S-IONBsX100 given their high polydispersity due to the mixture of shapes produced. In the end, the complicated interplay between particle shape, surface anisotropy and overall size distribution may equally contribute to

SLP values, which makes direct conclusions as to what accounts for the numbers seen here difficult. Further work in the isolation of particular shapes from the mixtures produced will allow for a better understanding as to which properties contribute the most to high SLP values.

2.4.6 Cell uptake studies

As discussed above, changes in nanoparticle morphology can and do have profound effects on the properties of the particles in biological systems.

With this in mind, collaborators in the Miller lab at the University of Manitoba compared the cell uptake properties of S-IONBsX45 with quasi-spherical IONPs with the same surface functionality (S-IONPs, as above). The general results of these studies will be discussed below. A manuscript containing the results is currently in preparation.68

Aqueous dispersions of S-IONBsX45 and S-IONPs at various concentrations were used in uptake studies on three types of cells: MDCK

(canine kidney epithelial cells), HLEC (primary human lung endothelial cells), and brain endothelial cells (bEnd.3). The exposure to the particles on the cells was done in the presence and absence of a static external magnetic field. After 4 hours of exposure, the cells were washed several times with PBS to remove any

99 unbound particles, and then lysed with a solution of sodium hydroxide. The amount of internalized particles was calculated via a ferrozine assay, which determines the concentration of Fe in solution.

The results of these studies were twofold. First, it was found that S-

IONBsX45 showed preferential uptake in the endothelial cell lines as compared with the kidney cells. The uptake was notably enhanced with the presence of an external magnetic field. Second, within brain endothelial cells, S-IONBsX45 were internalized at concentrations up to 30x of those found with spherical S-IONPs.

This demonstrated in general that changes in IONP shape have dramatic effects on both the extent to which the particles are internalized by cells, and that the effects of shape changes are not equal across different types of cells. In order to better understand the selectivity of S-IONBsX45 towards endothelial cells, the mechanism of internalization was investigated. This was done by treating bEnd.3 cell lines with various inhibitors which selectively block different uptake pathways.

It was found that uptake of the particles was significantly decreased when inhibitors which block caveolae mediated endocytosis pathways were used, an effect unseen with other inhibitors (for clathrin mediated endocytosis, micropinocytosis, and endosome maturation). These results were further bolstered with experiments showing that endothelial bEnd.3 cells have an increase in expression of caveolin-1, a protein which facilitates caveolae mediated endocytosis, as compared with epithellial MDCK cell lines.

100

In sum, these experiments demonstrate that shape control of IONPs offers a means through which targeted delivery to specific cells can be achieved without the need for complicated surface modifications to the particles or the need for receptor-ligand interactions between the cells and the particles.

2.5 Conclusions

This study presents the synthesis of polyhedral particles of iron oxide via a modification of the aqueous co-precipitation method with Triton X surfactants. A variety of shapes – variations on cubic and rectangular “brick-like” shapes, deemed IONBs – are formed, with the precise mixture dependent on the surfactant used. The resulting particles are highly crystalline, and their surface properties can easily be modified with the in situ addition of a hydrophilic siloxane. Silanized IONBs remain stable when dispersed in water, allowing for applications in medicine. Their efficacy in two such applications, hyperthermia and MRI contrast, were investigated. Both types of particle systems, S-

IONBsX45 and S-IONBsX100, have SLP values an order of magnitude higher than spherical IONPs with the same siloxane coating. Additionally, both

“nanobrick” systems show highly favorable MRI T2 contrast properties, with r2 values comparable to the highest reported in the literature. We have shown an effective alternative strategy for the control of IONP morphology that is simple, cost-effective, water-based, and also environmentally friendly. In addition, these particles have been investigated for their applications in cell uptake and for their

101 potential as drug delivery vehicles. Recent studies by Sun et al. on S-IONBsX45 have shown that these particles are taken up in endothelial cells at a rate far greater than spherical particles with the same surface coating and reasonably similar hydrodynamic radius and ζ-potential. This demonstrates that IONP shape modification in general, and the specific shapes found in the particles discussed here, offer a potential means of targeted delivery to specific cells without the need for receptor-ligand interactions.

2.6 References

1 V. Yathindranath, V. Ganesh, M. Worden, M. Inokuchi and T. Hegmann, RSC Adv., 2013(24), 9210-9213.

2 Thanh NTK, Maclean N, Mahiddine S. Mechanisms of nucleation and growth of nanoparticles in solution. Chem. Rev. 2014;114(15):7610-7630

3 LaMer VK, Dinegar RH. Theory, production and mechanism of formation of monodispersed hydrosols. J. Am. Chem. Soc., 1950;72(11):4847-4854

4 Niederberger M, Colfen H. Oriented attachment and mesocrystals: Non- classical crystallization mechanisms based on nanoparticle assembly. Phys. Chem. Chem. Phys., 2006;8(28):3271-3287

5 Peng ZA, Peng X. Mechanisms of the shape evolution of CdSe nanocrystals. J. Am. Chem. Soc., 2001;123(7):1389-1395

6 Parham H, Zargar B, Shiralipour R. Fast and efficient removal of mercury from water samples using magnetic iron oxide nanoparticles modified with 2- mercaptobenzothiazole. J. Hazard. Mater., 2012;205–206(0):94-100

7 Qu Q, Yang S, Feng X. 2D sandwich-like sheets of iron oxide grown on graphene as high energy anode material for supercapacitors. Adv. Mater., 2011; 23(46):5574-5580

102

8 Firouzabadi H, Iranpoor N, Gholinejad M, Hoseini J. Magnetite (Fe3O4) nanoparticles-catalyzed reactions in ethylene glycol under ligand-free conditions. Adv. Synth. Catal., 2011; 353(1):125-132

9 A) J. Guo, W. Yang and C. Wang, Adv Mater., 2013, 25(37), 5196-5214. B) E. Amstad, M. Textor and E. Reimhult, Nanoscale, 2011(7), 2819. C) R. Qiao, C. Yang and M. Gao, J. Mater. Chem., 2009, 19(35), 6274-6293

10 R. Massart, IEEE Trans. Magn., 1981, 17(2), 1247-1248.

11 V. Yathindranath, L. Rebbouh, D. F. Moore, D. W. Miller, J. van Lierop and T. Hegmann, Adv. Funct. Mater., 2011, 21(8), 1457-1464.

12 Wang X, Zhuang J, Peng Q, Li Y. A general strategy for nanocrystal synthesis. Nature, 2005;437(7055):121-124

13 Sun S, Zeng H. Size-controlled synthesis of magnetite nanoparticles. J. Am. Chem. Soc., 2002;124:8204-8205

14 Park J, An K, Hwang Y, et al. Ultra-large-scale syntheses of monodisperse nanocrystals. Nat Mater. 2004;3(12):891-895

15 A) Pinna N, Grancharov S, Beato P, Bonville P, Antonietti M, Niederberger M. Magnetite nanocrystals:Nonaqueous synthesis, characterization, and solubility. Chem. Mater., 2005;17(11):3044-3049 B) Lu A, Salabas E, Schüth F. Magnetic nanoparticles: Synthesis, protection, functionalization, and application. Angew. Chem. Int. Ed., 2007;46(8):1222-1244, C) Liu H, Ko SP, Wu J, et al. One-pot polyol synthesis of monosize PVP-coated sub-5 nm Fe3O4 nanoparticles for biomedical applications. J. Magn. Magn. Mater., 2007;310(2, Part 3):e815-e817

16 Lu A, Salabas E, Schüth F. Magnetic nanoparticles: Synthesis, protection, functionalization, and application. Angew. Chem. Int. Ed., 2007;46(8):1222-1244

17 Kim D, Lee N, Park M, Kim BH, An K, Hyeon T. Synthesis of uniform ferrimagnetic magnetite nanocubes. J. Am. Chem. Soc., 2009;131(2):454-455

18 Yang H, Ogawa T, Hasegawa D, Takahashi M. Synthesis and magnetic properties of monodisperse magnetite nanocubes. J. Appl. Phys., 2008;103(7).

19 Guardia P, Perez-Juste J, Labarta A, Batlle X, Liz-Marzan LM. Heating rate influence on the synthesis of iron oxide nanoparticles: The case of decanoic acid. Chem. Commun., 2010, (33):6108

103

20 Lee C, Jeong H, Lim ST, Sohn M, Kim DW. Synthesis of iron oxide nanoparticles with control over shape using imidazolium-based ionic liquids. ACS Appl. Mater. Inter., 2010;2(3):756-759

21 Palchoudhury S, An W, Xu Y, et al. Synthesis and growth mechanism of iron oxide nanowhiskers. Nano Lett., 2011;11(3):1141-1146

22 Woo K, Lee HJ, Ahn J-, Park YS. Sol?gel mediated synthesis of Fe2O3 nanorods. Adv. Mater., 2003;15(20):1761-1764

23 Bao L, Low W, Jiang J, Ying JY. Colloidal synthesis of magnetic nanorods with tunable aspect ratios. J. Mater. Chem., 2012;22(15):7117-7120

24 Zhou, Z.; Zhu, X.; Wu, D.; Chen, Q.; Huang, D.; Sun, C.; Xin, J.; Ni, K.; Gao, J. Anisotropic Shaped Iron Oxide Nanostructures: Controlled Synthesis and Proton Relaxation Shortening Effects. Chem. Mater. 2015, 27, 3505-3515

25 Teja AS, Koh P. Synthesis, properties, and applications of magnetic iron oxide nanoparticles. Prog. Cryst. Growth Charact. Mater., 2009;55(1):22-45

26 Klotz S, Steinle-Neumann G, Str├ñssle T, Philippe J, Hansen T, Wenzel MJ. Magnetism and the verwey transition in under pressure. Phys. Rev. B., 2008;77(1):012411

27 Teja AS, Koh P. Synthesis, properties, and applications of magnetic iron oxide nanoparticles. Prog. Cryst. Growth Charact. Mater., 2009;55(1):22-45

28 Bean CP, Livingston JD. Superparamagnetism. J. Appl. Phys., 1959;30:120S- 129S

29 Jun Y, Seo J, Cheon J. Nanoscaling laws of magnetic nanoparticles and their applicabilities in biomedical sciences. Acc. Chem. Res., 2008;41(2):179-189

30 Getzlaff, Mathias. "Magnetic Anisotropy Effects." Fundamentals of Magnetism. Springer Berlin Heidelberg, 2008. 89-115.

31 Noh, S.-h.; Na, W.; Jang, J.-t.; Lee, J.-H.; Lee, E. J.; Moon, S. H.; Lim, Y.; Shin, J.-S.; Cheon, J. Nanoscale magnetism control via surface and exchange anisotropy for optimized ferrimagnetic hysteresis. Nano Lett., 2012;12(7):3716- 3721

104

32 Q A Pankhurst and J Connolly and S K Jones and,J.Dobson. Applications of magnetic nanoparticles in biomedicine. J. Phys. D., 2003;36(13):R167

33 Zhou Z, Zhao Z, Zhang H, et al. Interplay between longitudinal and transverse contrasts in Fe3O4 nanoplates with (111) exposed surfaces. ACS Nano., 2014;8(8):7976-7985

34 Laurent S, Forge D, Port M, et al. Magnetic iron oxide nanoparticles: Synthesis, stabilization, vectorization, physicochemical characterizations, and biological applications. Chem. Rev., 2008;108(6):2064-2110

35 Chavhan GB, Babyn PS, Thomas B, Shroff MM, Haacke EM. Principles, techniques, and applications of T2*-based MR imaging and its special applications. Radiographics., 2009;29(5):1433-1449

36 Laurent S., Forge D., Port M., Roch A., Robic C., Vander Elst. L, Muller R.N. Magnetic iron oxide nanoparticles: Synthesis, stabilization, vectorization, physicochemical characterizations, and biological applications. Chem. Rev., 2008;108(6):2064-2110

37 Neuberger T, Schupf B, Hofmann H, Hofmann M, von Rechenberg B. Superparamagnetic nanoparticles for biomedical applications: Possibilities and limitations of a new drug delivery system. J. Magn. Magn. Mater., 2005; 293(1):483-496

38 Lee, J.; Huh, Y.; Jun, Y.; Seo, J.; Jang, J.; Song, H.; Kim, S.; Cho, E.; Yoon, H.; Suh, J.; Cheon, J. Artificially engineered magnetic nanoparticles for ultra- sensitive molecular imaging. Nat. Med., 2007, 13, 95-99

39 Lee, N.; Choi, Y.; Lee, Y.; Park, M.; Moon, W. K.; Choi, S. H.; Hyeon, T. Water-Dispersible Ferrimagnetic Iron Oxide Nanocubes with Extremely High r2 Relaxivity for Highly Sensitive in Vivo MRI of Tumors. Nano Lett. 2012, 12, 3127- 3131

40 Zhao, Z.; Zhou, Z.; Bao, J.; Wang, Z.; Hu, J.; Chi, X.; Ni, K.; Wang, R.; Chen, X.; Chen, Z.; Gao, J. Octapod iron oxide nanoparticles as high-performance T2 contrast agents for magnetic resonance imaging. Nat. Commun. 2013, 4

41 Zhou, Z.; Zhu, X.; Wu, D.; Chen, Q.; Huang, D.; Sun, C.; Xin, J.; Ni, K.; Gao, J. Anisotropic Shaped Iron Oxide Nanostructures: Controlled Synthesis and Proton Relaxation Shortening Effects. Chem. Mater. 2015, 27, 3505-3515

105

42 A) Dobson J. Magnetic nanoparticles for drug delivery. Drug Dev Res. 2006;67(1):55-60 B) Sun C, Lee JSH, Zhang M. Magnetic nanoparticles in MR imaging and drug delivery. Adv Drug Deliv Rev. 2008;60(11):1252-1265 C) Gupta AK, Gupta M. Synthesis and surface engineering of iron oxide nanoparticles for biomedical applications. Biomater. 2005;26(18):3995-4021

43 Daum N, Tscheka C, Neumeyer A, Schneider M. Novel approaches for drug delivery systems in nanomedicine: Effects of particle design and shape. Wiley Interdisciplinary Reviews: Nanomedicine and Nanobiotechnology. 2012;4(1):52- 65

44 Nangia S, Sureshkumar R. Effects of nanoparticle charge and shape anisotropy on translocation through cell membranes. Langmuir. 2012;28(51):17666–17671

45 Zhang Y, Tekobo S, Tu Y, et al. Permission to enter cell by shape: Nanodisk vs nanosphere. ACS Appl. Mater. Interfaces. 2012;4(8):4099-4105

46 Gratton SEA, Ropp PA, Pohlhaus PD, et al. The effect of particle design on cellular internalization pathways. Proc. Natl. Acad. Sci. 2008;105(33):11613- 11618

47 Kumar CSSR, Mohammad F. Magnetic nanomaterials for hyperthermia-based therapy and controlled drug delivery. Adv. Drug Deliv. Rev. 2011; 63(9):789-808

48 Jordan A, Scholz R, Wust P, Fuñhling H, Felix R. Magnetic fluid hyperthermia (MFH): Cancer treatment with AC magnetic field induced excitation of biocompatible superparamagnetic nanoparticles. J Magn Magn Mater. 1999; 201(1):413-419

49 Shu-Yi Wang, Shujuan Huang, Borca-Tasciuc D. Potential sources of errors in measuring and evaluating the specific loss power of magnetic nanoparticles in an alternating magnetic field. Magnetics, IEEE Transactions on. 2013;49(1):255-262

50 Shu-Yi Wang, Shujuan Huang, Borca-Tasciuc D. Potential sources of errors in measuring and evaluating the specific loss power of magnetic nanoparticles in an alternating magnetic field. Magnetics, IEEE Transactions on. 2013;49(1):255-262

51 Fortin J, Wilhelm C, Servais J, Manager C, Bacri J, Gazeau F. Size-sorted anionic iron oxide nanomagnets as colloidal mediators for magnetic hyperthermia. J. Am. Chem. Soc. 2007;129(9):2628-2635

106

52 Rosensweig RE. Heating magnetic fluid with alternating magnetic field. J Magn. Magn. Mater. 2002;252(0):370-374

53 Zhao D, Zeng X, Xia Q, Tang J. Preparation and coercivity and saturation magnetization dependence of inductive heating property of Fe3O4 nanoparticles in an alternating current magnetic field for localized hyperthermia. J. Alloys Compounds. 2009;469(1):215-218

54 Ma M, Wu Y, Zhou J, Sun Y, Zhang Y, Gu N. Size dependence of specific power absorption of Fe3O4 particles in AC magnetic field. J. Magn. Magn. Mater. 2004; 268(1):33-39

55 Martinez-Boubeta, C.; Simeonidis, K.; Makridis, A.; Angelakeris, M.; Iglesias, O.; Guardia, P.; Cabot, A.; Yedra, L.; Estrade, S.; Peiro, F.; Saghi, Z.; Midgley, P. A.; Conde-Leboran, I.; Serantes, D.; Baldomir, D. Learning from nature to improve the heat generation of iron-oxide nanoparticles for magnetic hyperthermia applications. Sci. Rep. 2013, 3, 1652, 8 pp.

56 Hugounenq P, Levy M, Alloyeau D, et al. Iron oxide monocrystalline nanoflowers for highly efficient magnetic hyperthermia. J. Phys. Chem. C. 2012; 116(29):15702-15712

57 Dennis CL, Ivkov R. Physics of heat generation using magnetic nanoparticles for hyperthermia. Int. J. Hyperthermia. 2013;29(8):715-729

58 Goodby JW, Gray GW, Demus D. “Guide to the nomenclature and classification of liquid crystals.” In: Handbook of liquid crystals set. Wiley-VCH Verlag GmbH; 1998;17-23.

59 Fong C, Le T, Drummond CJ. Chem. Soc. Rev., 2012,41, 1297-1322

60 Fritscher, C.; Husing, N.; Bernstorff, S.; Brandhuber, D.; Koch, T.; Seidler, S.; Lichtenegger, H. C. \it In situ SAXS study on cationic and non-ionic surfactant liquid crystals using synchrotron radiation. J. Synchrotron Rad. 2005, 12, 717- 720

61 Ahir SV, Petrov PG, Terentjev EM. Rheology at the phase transition boundary 2. hexagonal phase of triton X100 surfactant solution. Langmuir. 2002;18(24):9140-9148

62 A. Mitra, J. Mohapatra, S. S. Meena, C. V. Tomy and M. Aslam, J. Phys. Chem. C, 2014, 118(33), 19356-19362.

107

63 M. Rohrer, H. Bauer, J. Mintorovitch, M. Requardt and H. Weinmann, Invest. Radiol., 2005, 40(11), 715-24.

64 Y. Wang, Quant. Imaging Med. and Surg., 2011, 1(1), 35-40.

65 N. Lee, H. Kim, S. H. Choi, M. Park, D. Kim, H. Kim, Y. Choi, S. Lin, B. H. Kim, H. S. Jung, H. Kim, K. S. Park, W. K. Moon and T. Hyeon, Proc. Natl. Acad. Sci. U. S. A., 2011, 108(7), 2662-2667.

66 X. Huang, B. D. Stein, H. Cheng, A. Malyutin, I. B. Tsvetkova, D. V. Baxter, N. B. Remmes, J. Verchot, C. Kao, L. M. Bronstein and B. Dragnea, ACS Nano, 2011, 5 (5), 4037–4045.

67 Z. Zhao, Z. Zhou, J. Bao, Z. Wang, J. Hu, X. Chi, K. Ni, R. Wang, X. Chen, Z. Chen and J. Gao, Nat. Comm., 2013, 4, 2266.

68 Z. Sun, M. Worden, Y. Wroczynskyj, J. Thliveris, J. van Lierop, T. Hegmann, D. W. Miller, Differential Internalization of Brick Shaped Iron Oxide Nanoparticles by Endothelial Cells, 2015, in preparation.

108

Chapter 3

Surface functionalization of iron oxide nanoparticles for stabilization in

aqueous media and bioapplications

111

3.1 Project Goals

The broad purpose of this work concerned the synthesis of biocompatible

IONPs for use in cell uptake and drug delivery investigations. As noted above, work had already been done establishing AmS as a versatile functionalizing agent for IONPs. The next area of investigation we decided to look at involved the effects that surface charge may have on bio-applications. AmS-IONPs, due to the terminal amine group, tend to have a positive surface potential in aqueous media. The work involved attempting to synthesize biocompatible IONPs with a negative surface potential, either through chemical modification of AmS-IONPs, or through the use of an entirely different silane coating. The differences between positively and negatively charged IONPs in solution with respect to cytotoxicity and cell uptake could then be investigated, with the hopes of better understanding what kind of functionalizing agent might be best for different applications. Additionally, covalent attachment of charged polymers was also tested (eg. polyethyleneimine, or PEI) for the purpose of creating a delivery vehicle for charged biomolecules like siRNA. The results of this synthetic work on surface coatings on IONPs are presented below, with references to how these materials were used in pharmacological studies conducted by other researchers provided in appendices.

112

3.2 Functionalization of IONPs with silanes for bioapplications

3.2.1 Introduction to surface functionalization of IONPs

As noted in Chapter 2, there are a huge variety of methods for creating

IONPs. For most of the applications of these materials, the particles need to be functionalized. That is, their surfaces need to be modified in some way to include chemical moieties which can protect the particles from further oxidation and aggregation, and which can allow for properties such as biocompatibility and stability in various media. Since pharmacological applications are the main end- use goals of the particles synthesized during this project, this introduction will focus on functionalizing IONPs with hydrophilic chemical groups.

Functionalization with synthetic and/or natural polymers is one of the most common methods for creating hydrophilic, biocompatible IONPs. A wide variety of amphiphilic polymers, such as poly(acrylic acid), poly(vinyl alcohol), poly(ethylene imine), poly(ethylene oxide) dextran, chitosan, etc. have been used to create biocompatible IONPs.1 Poly(ethylene glycol (PEG) is perhaps the most widely used synthetic polymer for IONP functionalization due to the very low toxicity of PEG as well as the high blood circulation times seen in PEGylated nanoparticles.2 Oftentimes the interaction between an IONP and the polymer is simply throughnon-specific adsorption, so long term stability of these composites can be an issue.

113

Coordinative binding of small-molecule conjugate bases, such as carboxylates and phoshonates, is another common method for functionalizing

IONPs.1 Citrate-coated IONPs, for example, have been used in commercially available MRI contrast agents.3 Carboxylate (COO-)/IONP bonding is highly labile, however, which again can cause long-term stability issues. The bidentate binding between phosphonates and the IONP surface is stronger than binding with carboxylates, and has led to IONPs functionalized with multifunctional phosphonic acids being investigated in cancer therapy and cell imaging.1

The formation of a silica shell around nanoparticles of all kinds is a very popular way to create highly stable and biocompatible nanomaterials.4 In terms of IONPs, modifications on the Stöber method – which involves the hydrolysis of tetraethyl orthosilane to form mesoporous silica – have been used to create core- shell composites that can allow for the controlled release of therapeutic compounds.5 Further modification of this method through the use of multifunctional organosilanes allows for the synthesis of stable, hydrophilic

IONPs with terminal moieties that can be used for additional conjugation to biologically relevant molecules. Figure 37 shows a general schematic of this

“silanization” process whereby a trialkoxysilane coats the surface of an IONP via a condensation reaction with surface hydroxyl groups on the particle.

The most ubiquitous of these silanes is 3-aminopropyltrimethoxysilane

(often termed just aminosilane, APTES, or AmS), which has been used to

114 stabilize and functionalize nanoparticles of nearly every composition, from gold6 to silica7 to, of course, iron oxide.8 More generally, silanes are available with a variety of terminal functionalities (thiol, carboxylic acid, epoxy, aldehyde, etc.), and a number of publications have described and compared these different organosilanes as capping agents on nanoparticles. Jana et al. reported that a variety of different organosilanes can be used to make stable aqueous dispersions of metal, metal oxide, and quantum dot NPs.9 The particles could be nearly monodisperse in solution, stable in different buffers, and have positive or negative surface charges depending on the terminal functionalities of the silane and the properties of the dispersion media. De Palma et al. demonstrated that organosilanes can be used in ligand exchange reactions to make hydrophobic

IONPs water dispersible.10 They found that the silane coatings were robust, protecting the IONP core from degradation in acidic or basic conditions, and could create highly concentrated dispersions of IONPs in aqueous media. Li et al. compared IONPs with different silane coatings for their ability to conjugate with and stabilize a model biomolecule, trypsin.11 They concluded that immobilization of trypsinn on silanized IONPs offered a highly durable and reusable (via easy isolation of the particles with an external magnet) means to deliver biomolecules to a substrate.

115

Figure 37: Schematic representation of the silanization of an IONP.

Previous work published in the Hegmann lab looked at combining the reduction/hydrolysis method for making metal oxide NPs (discussed at length in

Chapter 1) with the silanization process in order to create a simple, one-pot procedure for hydrophilic, biocompatible IONPs. Vinith Yathindranath, a former member of the Hegmann research group, developed the initial investigations into

IONP functionalization in collaboration with members of the Miller research group in the pharmacology department at the University of Manitoba.12 He chose AmS as the molecule to functionalize the IONPs. AmS-IONPs were found to be biocompatible, showing little cytotoxicity on HepG2 cells, and the terminal amine moiety allowed for covalent attachment with a variety of additional molecules (eg.

PEG, oleic acid, and bovine serum albumin, a common model protein) while remaining easily dispersible in aqueous media. In addition, AmS-IONPs could be used to bind molecules of small interfering RNA (siRNA) through the electrostatic

116 interactions between the positive surface charges on the AmS and the negative surface charges on the siRNA. This work demonstrated the ease of synthesis and versatility in application for silanized IONPs made via the reduction/hydrolysis method, and set the ground work for the following investigations into these materials.

3.2.2 Instrumentation

Surface functionalization was analyzed through FT-IR using KBr pellet techniques. Approximately 1 mg of dried particles were mixed with approximately

117

TEM imaging was done with a FEI Tecnai TF20 TEM instrument at an accelerating voltage of 200 kV. Particle samples were dispersed in methanol and dropcast onto 400 mesh carbon coated copper grids.

3.2.3 Experimental methods

Synthesis of aminosilane coated iron oxide nanoparticles (AmS-IONPs):

The synthesis of the particle core and the addition of the functionalizing aminosilane coating followed prior publications.12 In a typical synthesis,

Fe(acac)3 (8 mmol) was dissolved in 200 mL of ethanol and degassed for 1 hour.

NaBH4 (80 mmol) was dissolved in 200 mL of degassed water and then quickly added to the above under nitrogen and 1000 rpm mechanical stirring. Over the course of one hour, the initial red solution slowly turned opaque brown and finally black, indicating the formation of IONPs. After 1 hour, AmS (10 mL, ~40 mmol) was added directly to the reaction mixture, which was then left to mix under the same conditions as above overnight. The next day, the black mixture was vacuum filtered to remove any large aggregates, concentrated via rotavap, and then dialyzed against water (MWCO 6-8000) to remove any unreacted precursors. The particles were stored and analyzed in solution, and small portions were dried via rotavap and high vacuum to obtain dry powdered sample for TGA, IR, and TEM characterization.

Synthesis of COOH-AmS-IONPs: In a typical synthesis, 5 mL of water containing BAA (2mmol) was added dropwise to a 10 mL aqueous dispersion of

118

AmS-IONPs (~10 mg/mL). This was left to mix under ambient conditions for 3 hours. The resulting particles were isolated via centrifugation, washed several times with water, and air dried. The purified particles could then be redispersed in water.

Synthesis of carboxysilane coated iron oxide nanoparticles (EDTS-

IONPs): In a typical synthesis, Fe(acac)3 (4 mmol) was dissolved in 100 mL of ethanol and degassed for 1 hour. NaBH4 (40 mmol) was dissolved in 100 mL of degassed water and then quickly added to the above under nitrogen and 1000 rpm mechanical stirring. Over the course of one hour, the initial red solution slowly turned opaque brown and finally black, indicating the formation of IONPs.

After 1 hour, EDTS (10 mL, ~20 mmol) was added directly to the reaction mixture, which was then left to mix under the same conditions as above overnight. The next day, the black mixture was vacuum filtered to remove any large aggregates, concentrated using a rotatory evaporator, and then dialyzed against water (MWCO 6-8000) to remove any unreacted precursors. The particles were stored and analyzed in solution, and small portions were dried with a rotatory evaporator and high vacuum to obtain dry powdered sample for TGA,

IR, and TEM characterization.

Synthesis of PEI-IONPs: In a typical synthesis, a 25 mL aqueous dispersion containing EDT-IONPs at a concentration of 7.8 mg/mL was mixed with a 10 mL solution of MES buffer containing 0.011 mol EDC and 0.011 mol

119

NHS. The resulting mixture was stirred for 15 minutes. Separately, a solution containing 10 mL MES buffer and 10 mL of PEI (50% in water, MW = 2000) was made and mixed for 15 minutes. The two solutions were then combined and stirred mechanically overnight. The reaction mixture was purified via dialysis

(MWCO = 6000 to 8000) against water. The resulting PEI-IONPs were isolated from solution with a rotatory evaporator and stored under ambient conditions.

3.3 Results and Discussion

The synthesis of the IONP core followed the reduction-hydrolysis method.13 This particular synthesis allows for particles which are pure magnetite

(as opposed to mixtures of magnetite and maghemite typically seen in other aqueous syntheses), between 5 and 10 nm in size. It also allows for a simple means of functionalizing the particles directly in situ, leading to hydrophilic, dispersible particles without the need for additional ligand exchange steps. The reaction occurs in basic conditions, which results in surface hydroxyl groups attached to the particles when suspended in the reaction medium. Additionally, the alkoxy group on silane molecules introduced to this reaction medium will hydrolyze to form hydroxyl groups. Between these two sets of hydroxyl groups, a series of condensation reactions occur, resulting in an IONP core coated with a thin siloxane shell to which the various functional groups are attached.

120

Figure 38: Representations of IONPs coated (top) with AmS, and (bottom) after surface modification with BAA (indicated by the star shape).

The first silane molecule used to functionalize the IONPs was AmS. As noted above, AmS is a popular molecule for functionalizing various nanomaterials and for general surface coatings and had already been used previously

121 in similar studies in the Hegmann lab. Figure 38 shows a schematic image of an

AmS coated IONP. The nature of the surface coating was characterized by FT-IR and TGA, with results of a typical synthesis shown in Figure 39. The FT-IR shows a broad peak centered at 3400 cm-1, indicative of O-H vibrational stretching which could come from a combination of adsorbed moisture and unreacted surface hydroxyl groups. The vibrational N-H stretching peak is likely saturated by the O-H stretching peak. Symmetric and asymmetric C-H vibrational stretching peaks are seen at 2920 cm-1 and 2850 cm-1, respectively. A broad peak centered around 1600 cm-1 is indicative of N-H bending, and a smaller broad peak centered around 1450 cm-1 coming from C-H bending. The collection of broad peaks between 1000 – 1250 cm-1 come from vibrational stretching associated with the siloxane groups (Si-O and Si-C). Vibrational stretching associated with Fe-O is seen at 580 cm-1. The TGA plot shows that the amount of organic surface coating was about 10 wt%. The initial weight loss before 100

°C was assumed to be associated with adsorbed moisture. Representative TEM images of AmS-IONPs are shown in Figure 41A and Figure 41B. These images show the particles to be quasi-spherical in shape and uniformly between 5 and

10 nm in size.

The first attempts to modify the surface of IONPs in order to create a negative surface potential when dispersed in aqueous media involved reacting

AmS-IONPs with bromoacetic acid (BAA). This reaction has been shown as a means to modify amine-terminated particles with a terminal carboxylic acid

122 group.14 This carboxylic acid group is expected to impart a neutral to negative surface potential on the particles when in solution (depending on the specific pH conditions of the dispersion media). The reaction involves the elimination of the bromine atom on the α-carbon of BAA by the amine terminus on the AmS-IONPs; a schematic representation of the resulting material, COOH-AmS-IONPs, is seen in Figure 38.

123

Figure 39: Characterization of AmS-IONPs. (Top) FT-IF spectrum; (bottom) TGA weight loss plot.

124

Figure 40: Characterization of COOH-AmS-IONPs. (Top) FT-IR spectrum; (bottom) TGA weight loss plot.

The nature of the surface coating for COOH-AmS-IONPs was characterized by FT-IR and TGA, with results of a typical synthesis shown in Figure 40. The FT-IR spectrum shows an intense, broad peak centered

125 at 1410 cm-1 with a second smaller shoulder peak at 1590 cm-1. These peaks can be attributed to a combination of bending and vibration modes associated with carboxylate moieties, although it is unclear as to why the peak at 1410 cm-1 is so intense. A weak, sharp peak at 1750 cm-1 suggests that a small amount of the carboxylate groups are protonated. The TGA plot shows that the amount of organic surface coating was about 18 wt%; the increase in weight relative to the initial AmS-IONPs is attributed to the addition of the acetic acid group.

DLS and zeta-potential measurements first conducted on the aqueous dispersions containing approximately 1 mg/mL of dry particles in

DI water in order to measure the effective particle diameter and surface potential of the particles with the different surface coatings. The results can be seen in Table 5 .

Table 5: Physio-chemical properties of silanized IONPs dispersed in water. Material Zeta potential (mV) Effective diameter (nm)

AmS-IONPs + 23.9 +/- 1.2 27.0 +/- 3.2

COOH-AmS-IONPs - 17.0 +/- 3.9 26.8 +/- 8.9

As hoped, AmS-IONPs and COOH-AmS-IONPs show approximately the same size when dispersed in water (~27.0 nm and ~26.8 nm, respectively), but exhibit oppositely charged surface potentials (~+23.9 mV, and ~−17.0 mV, respectively). This showed that these two materials could

126 be used to investigate and compare the effect that the change in surface functionality, and specifically surface potential, may have on cell viability and uptake. This research was conducted by Zhizhi Sun from the Miller lab at the University of Manitoba and published in 2013.15 A general summary of the results is given below.

The cytotoxicity and cell uptake properties of AmS-IONPs and

COOH-AmS-IONPs were investigated with mouse brain endothelial cells

(bEnd.3), mouse astrocytes, and mouse neurons. An MTT assay was conducted to measure cytotoxicity at various particle concentrations, and the accumulation of particles inside the cells was measured by analysing the amount of Fe using a ferrozine assay. Neither material showed measurable cytotoxicity below a concentration of 100 μg/mL. Above this concentration differential toxicity was seen; AmS-IONPs showed increased toxicity towards neurons, while COOH-AmS-IONPs showed increased toxicity towards astrocytes. AmS-IONPs exhibited a significantly higher degree of cell accumulation across all three types of cells investigated, and this was attributed largely to the effect of the positively charged surface potential. Additionally, all studies were conducted with and without the presence of an external magnetic field. This showed very minimal effects on cytotoxicity properties, but resulted in a significant increase in cell accumulation across both sets of particles in all three types of cells.

127

Figure 41: Representative TEM images of AmS-IONPs (A and B), and EDTS- IONPs (C and D).

128

Figure 42: Representation of an IONP coated with EDTS.

Although the above results demonstrate that AmS-IONPs could be modified with terminal carboxylates via a reaction with BAA, we hoped to eliminate the need for this extra reaction step by using a silane that already had this moiety. This would allow us to maintain the single, one-pot synthesis of functionalized IONPs, as is the case with AmS-IONPs. To do so, we looked at N-(trimethoxysilylpropyl)ethylenediaminetriacetate

(abbreviated EDTS), a silane typically stored in aqueous media as a trisodium salt. A report on the surface modification of IONPs with a wide range of various silane molecules included this compound, demonstrating its capability as a hydrophilic functionalizing agent.10 There are few, if any, reports on the use of IONPs coated with EDTS in bioapplications, but this can be seen as a positive since it could open a variety of new areas of investigation.

129

Figure 43: Characterization of EDTS-IONPs. (Top) FT-IR spectrum; (bottom) TGA weight loss plot.

130

Figure 42 shows a schematic representation of an EDTS-IONP. The

EDTS molecule contains three carboxylate groups, which are shown deprotonated (though that depends on the nature of the medium in which the particles may be dispersed). The nature of the surface coating for

EDTS-IONPs was characterized by FT-IR and TGA, with the results of a typical synthesis shown in Figure 43. The FT-IR spectrum shows a typical broad peak centered at 3400 cm-1 indicative of vibrational O-H stretching, due either to adsorbed moisture or protonated COOH groups. Sharp, narrow peaks indicative of symmetric and asymmetric C-H vibrations are seen at 2920 cm-1 and 2850 cm-1, respectively. The collection of broad peaks centered between 1550 and 1650 cm-1 are indicative of asymmetric vibrations in a deprotonated carboxylate group. The broad band between

1000 – 1250 cm-1 can be attributed to vibrational stretching associated with the siloxane groups (Si-O and Si-C). Finally, the peak around 580 cm-1 is associated with vibrational Fe-O stretching. A typical TGA plot shows that the amount of organic surface coating was about 43 wt%. Representative

TEM images of AmS-IONPs are shown in Figure 41C and Figure 41D.

These images show the particles to be quasi-spherical in shape and uniformly between 5 and 10 nm in size.

One benefit of this particular functionalizing agent that was apparent immediately was that the dried powder could be easily redispersed in high concentrations in aqueous media (AmS-IONPs are difficult to redisperse

131 once they have been dried), and would remain stable indefinitely. As such,

EDTS-IONPs could be dispersed in water and other biologically relevant media, such as PBS (phosphate buffered saline), a buffer often used as a part of a growth media for cell cultures. Dispersions of EDTS-IONPs in water and 1 M PBS solution, at a concentration of 2 mg/mL, are shown in the inset of Figure 44. As can be seen, the dispersions are transparent and show no settling or precipitation of the particles. The physio-chemical properties of these dispersions were investigated by DLS and zeta potential measurements, with the results shown in Table 6. Typical plots showing the distribution of measured hydrodynamic radii are shown in

Figure 44, demonstrating a low degree of polydispersity in the particle sizes when dispersed in either medium.

Table 6: Effective particle diameter and zeta potential of EDTS-IONPs dispersed in water and PBS. Media pH Effective diameter (nm) Zeta potential (mV)

DI water 10.0 30.3 +/- 0.2 - 36.8 +/- 2.8

1 M PBS 7.4 18.0 +/- 0.4 ------

132

Figure 44: Example of a plot of hydrodynamic radius vs. normalized intensity for EDTS-IONPs in DI water (black) and PBS (red). Inset shows EDTS-IONPs dispersed in PBS (left) and water (right).

In water, EDTS-IONPs had an effective diameter of ~30 nm, typical of silanized particles as seen with AmS-IONPs and COOH-AmS-IONPs previously. They show a highly negative surface potential of approximately

-37 mV, which accounts for the high dispersibility and stability of these particles. The pH of the mixture (which can have a significant effect on both

133 measurements as it will affect the degree of protonation on the carboxylate terminal groups) was found to be 10.0. Although the water used to disperse the particles was neutral, the pH of the media increased because the

EDTS molecule used for synthesis existed as the salt of the carboxylic acid moieties. EDTS-IONPs had an effective diameter of only ~18 nm when dispersed in PBS. It’s possible that the increased ionic strength of the PBS solution relative to pure water caused the particles to repel each other to an even greater extent that would occur otherwise. The zeta potential of the PBS dispersion was not measurable as the particles in this medium tended to aggregate on the electrode.

Further work on various bioapplications of EDTS-IONPs was done, once again by Zhizhi Sun at the University of Manitoba and published in

2013.16 As done previously, a general summary of the results is given below, with a more detailed copy of the publication reproduced in Appendix

The broad goal of the pharmacological studies was to improve the permeability of IONPs across an in vitro model of the blood-brain barrier.

AmS-IONPs and EDTS-IONPs were used on confluent monolayers of mouse bEnd.3 cells (the model BBB), under normal and osmotically disrupted conditions, with and without the presence of an external magnetic field. Since the previous study on AmS- and COOH-AmS-IONPs demonstrated the impact that opposite surface charges can have on cell

134 uptake, it was thought in this study that the use of differently charged

IONPs would allow for permeability through different pathways in order to see which might be the most efficient. It was found that an intact monolayer prevented either type of IONP from permeation (which was not unexpected). When the monolayer was disrupted using D-mannitol – a compound often used to briefly open the tight junctions between BBB cells

– it was found that EDTS-IONPs could permeate the barrier to a significant degree. This permeation rate was nearly doubled through the use of an external magnetic field, which was thought to pull the IONPs through the barrier through a process termed magnetic field convective diffusion. In contrast, the AmS-IONPs showed relatively little increase in permeation rate through the disrupted monolayer. These results demonstrated once again the importance that surface charge can have on bioapplications of

IONPs and, more specifically, that EDTS coated IONPs show a great deal of promise in their use as vehicles for crossing the BBB.

The carboxylate terminal moiety allows for further modification of

EDTS-IONPs via EDC/NHS coupling. These reagents are commonly used in biochemistry as a means to conjugate large molecules such as proteins together via the formation of an amide bond between an amine and a carboxyl group (ester or acid).17 Figure 45 shows a schematic procedure demonstrating the reaction steps involved in the formation of an amide bond between a carboxylate and a primary amine via EDC/NHS coupling.

135

This method was used to couple PEI to EDTS-IONPs for the purpose of creating a high density of positive charges around the surface of the particles.

The resulting particles are termed simply PEI-IONPs. Figure 46 shows a representation of a PEI-IONP. For the purpose of clarity, the schematic shows a single amide bond between the PEI polymer and one carboxylate group on an

136 individual EDTS molecule. It’s possible that multiple amide bonds could be formed between the two molecules.

Figure 46: Representation of an IONP coated with EDTS and coupled with PEI via an amide bond.

As previously, the nature of the surface coating for PEI-IONPs was characterized by FT-IR and TGA, with the results of a typical synthesis shown in Figure 47. The FT-IR spectrum shows a small but apparent broad peak at approximately 3100 cm-1, which can be attributed to secondary amine N-H vibrational stretching. There is also a strong increase (as compared with the spectrum for EDTS-IONPs, in Figure 43)

137 in the peak centered at approximately 1210 cm-1 which is likely attributable to a C-N stretch typical of aliphatic amines. The amide carbonyl peak, formed through EDC/NHS coupling, is not evident because it is likely buried in the peaks already associated with the carboxylate C-O stretching from the EDTS molecule. The TGA plot shows a total weight loss of ~80%, which means 37% of the total weight of the dried particles can be attributed to PEI (after subtracting for the ~43 wt% of EDTS as seen in Figure 43).

Dried samples of PEI-IONPs could be redispersed in aqueous media with mild sonication. The physiochemical properties of dispersions in DI water and 1 M PBS solution were evaluated by DLS and zeta-potential measurements, summarized in Table 7.

Table 7: Physiochemical properties of PEI-IONPs dispersed in DI water and 1M PBS solution. Media pH Effective diameter (nm) Zeta potential (mV)

water 10.1 63.3 +/-1.2 +4.9 +/- 1.6

PBS 7.4 48.4 +/- 0.6 +27.5 +/- 10.6

When dispersed in DI water, PEI-IONPs had an effective diameter of

~63 nm. This was over twice the size of the precursor EDTS-IONPs, which was further indication of the presence of PEI polymers on the particles. The surface potential was relatively low, at only +4.9 mV. This is due to the basic conditions (the dispersion had a pH of 10.1) which would leave many amine groups deprotonated and thus uncharged. The dispersion properties

138 changed substantially when in PBS. The surface potential increased to

+27.5 mV, likely due to the decrease in pH conditions allowing for more amine groups to become protonated, and the effective particle diameter became smaller, at ~48 nm, likely due to the increased surface charge forcing the particles to separate more in solution.

Investigations into the ability of PEI-IONPs as synthesized above to act as vehicles for the delivery of siRNA in model biological systems were done once again in the Miller lab at the University of Manitoba. A broad overview of the results of these investigations is presented below. A full manuscript is currently in preparation.18

RNA interference – the process through which short strands of RNA

(usually called small interfering RNA, or siRNA) are experimentally introduced into a cell for the purpose of modulating the expression of certain genes19 – has become increasingly popular in investigations in possible gene therapy treatments recent years. Naked strands of siRNA usually undergo rapid degradation in biological systems, and so delivery vehicles such as polymer liposomes have been used to increase the concentration and overall targeting efficiency of siRNA.20 Functionalized

IONPs, given their well-established biocompatibility and tunable chemical and physical properties, have also been shown to be efficient siRNA carriers.21 Our previous work with AmS-IONPs showed a proof of concept demonstration that silanized IONPs could bind siRNA. The expectation

139 was that further modification with PEI would increase the overall loading efficiency of siRNA onto the IONPs and thus allow more siRNA to reach targeted cells. To investigate this, dispersions of PEI-IONPs were mixed with siRNA in various weight ratios. The expectation was that the electrostatic interactions between the positively charged PEI-IONPs and the negatively charged siRNA molecules would cause the siRNA to “stick” to the PEI-IONPs. This was measured via gel retardation assays, which showed no free siRNA molecules in solution when mixed with PEI-IONPs at weight ratios of no greater than 10:1 (siRNA:PEI-IONPs). The specific siRNA molecule used is effects the expression of Pgp (P-glycoprotein).

The ability of these dispersions of siRNA:PEI-IONP complexes to affect the expression of Pgp in a model cell line was examined. The complexes showed a decrease in Pgp expression significantly higher than that shown by free siRNA, demonstrating that PEI-IONPs could be used as more efficient delivery vehicles for siRNA therapy.

140

Figure 47: Characterization of PEI-IONPs. (Top) FT-IR spectrum; (bottom) TGA weight loss plot.

141

3.4 Conclusions

IONPs were synthesized via the reduction/hydrolysis method and functionalized with different silane molecules. This gave IONPs coated with amine and carboxylate terminal moieties which allowed for stable dispersions in aqueous media with both positive and negative surface charges. Additionally, further chemical conjugation could be done, such as covalent attachment of small molecules and a positively charged polymer, PEI. These different particles were used in pharmacological studies investigating the effects that surface charge can have on IONP uptake into cells and on translocation through a model blood brain barrier. PEI-IONPs were also shown to have a high binding affinity for siRNA, demonstrating their potential as delivery vehicles for RNA-based therapies.

3.5 References

1 Turcheniuk, K.; Tarasevych, A. V.; Kukhar, V. P.; Boukherroub, R.; Szunerits, S. Recent advances in surface chemistry strategies for the fabrication of functional iron oxide based magnetic nanoparticles. Nanoscale 2013, 5, 10729-10752

2 Lee, J. H.; Lee, H. B.; Andrade, J. D. Blood compatibility of polyethylene oxide surfaces. Prog. Polym. Sci. 1995, 20, 1043-1079

3 Taupitz, M.; Schnorr, J.; Pilgrimm, H.; Hamm, B.; Wagner, S. CMR 2005: 2.03: VSOP-C184 as contrast agent for MRI of atherosclerotic plaques: experimental results in rabbits. Con. Med. Mol. Imag. 2006, 1, 55-55

4 Jiang, S.; Win, K. Y.; Liu, S.; Teng, C. P.; Zheng, Y.; Han, M. Surface- functionalized nanoparticles for biosensing and imaging-guided therapeutics. Nanoscale 2013, 5, 3127-3148

142

5 Liu, J.; Qiao, S. Z.; Hu, Q. H.; (Max) Lu, G. Q. Magnetic Nanocomposites with Mesoporous Structures: Synthesis and Applications. Small 2011, 7, 425- 443

6 Yang, J.; Ichii, T.; Murase, K.; Sugimura, H. Site-Selective Assembly and Reorganization of Gold Nanoparticles along Aminosilane-Covered Nanolines Prepared on Indium-Tin Oxide. Langmuir 2012, 28, 7579-7584

7 Yokoi, T.; Kubota, Y.; Tatsumi, T. Amino-functionalized mesoporous silica as base catalyst and adsorbent. Appl. Catal., A. 2012, 421, 422, 14-37

8 A) Ma, M.; Zhang, Y.; Yu, W.; Shen, H.; Zhang, H.; Gu, N. Preparation and characterization of magnetite nanoparticles coated by amino silane. Colloids Surf. Physicochem. Eng. Aspects 2003, 212, 219-226. B) Mikhaylova, M.; Kim, D. K.; Berry, C. C.; Zagorodni, A.; Toprak, M.; Curtis, A. S. G.; Muhammed, M. BSA Immobilization on Amine-Functionalized Superparamagnetic Iron Oxide Nanoparticles. Chem. Mater. 2004, 16, 2344- 2354. C) Kim, D. K.; Mikhaylova, M.; Zhang, Y.; Muhammed, M. Protective Coating of Superparamagnetic Iron Oxide Nanoparticles. Chem. Mater. 2003, 15, 1617-1627

9 Jana, N. R.; Earhart, C.; Ying, J. Y. Synthesis of Water-Soluble and Functionalized Nanoparticles by Silica Coating. Chem. Mater. 2007, 19, 5074-5082

10 De Palma, R.; Peeters, S.; Van Bael, M. J.; Van, d. R.; Bonroy, K.; Laureyn, W.; Mullens, J.; Borghs, G.; Maes, G. Silane Ligand Exchange to Make Hydrophobic Superparamagnetic Nanoparticles Water-Dispersible. Chem. Mater. 2007, 19, 1821-1831

11 Li, D.; Teoh, W. Y.; Gooding, J. J.; Selomulya, C.; Amal, R. Functionalization Strategies for Protease Immobilization on Magnetic Nanoparticles. Adv. Funct. Mater. 2010, 20, 1767-1777

12 Yathindranath, V.; Sun, Z.; Worden, M.; Donald, L. J.; Thliveris, J. A.; Miller, D. W.; Hegmann, T. One-Pot Synthesis of Iron Oxide Nanoparticles with Functional Silane Shells: A Versatile General Precursor for Conjugations and Biomedical Applications. Langmuir 2013, 29, 10850-10858

13 Yathindranath, V.; Rebbouh, L.; Moore, D. F.; Miller, D. W.; van Lierop, J.; Hegmann, T. A Versatile Method for the Reductive, One-Pot Synthesis of Bare, Hydrophilic and Hydrophobic Magnetite Nanoparticles. Adv. Funct. Mater. 2011, 21, 1457-1464

14 Masotti, A.; Pitta, A.; Ortaggi, G.; Corti, M.; Innocenti, C.; Lascialfari, A.; Marinone, M.; Marzola, P.; Daducci, A.; Sbarbati, A.; Micotti, E.; Orsini, F.; Poletti, G.; Sangregorio, C. Synthesis and characterization of

143

polyethylenimine-based iron oxide composites as novel contrast agents for MRI. Magn. Reson. Mater. Phys., Biol. Med. 2009, 22, 77-87

15 Sun, Z.; Yathindranath, V.; Worden, M.; Thliveris, J. A.; Chu, S.; Parkinson, E.; Hegmann, T.; Miller, D. W. Characterization of cellular uptake and toxicity of aminosilane-coated iron oxide nanoparticles with different charges in central nervous system-relevant cell culture models. Int J Nanomed. 2013, 8, 961-970

16 Sun, Z.; Worden, M.; Wroczynskyj, Y.; van, L. J.; Yathindranath, V.; Hegmann, T.; Miller, D. W. Magnetic field enhanced convective diffusion of iron oxide nanoparticles in an osmotically disrupted cell culture model of the blood-brain barrier. Int. J. Nanomed. 2014, 9, 3013-26

17 Sehgal, D.; Vijay, I. K. A Method for the High Efficiency of Water-Soluble Carbodiimide-Mediated Amidation. Anal. Biochem. 1994, 218, 87-91

18 S. Morrison, M. Worden, V. Yathindranath, Z. Sun, T. Hegmann, D. Miller. siRNA delivery with PEI coated magnetic nanoparticles, 2015, in preparation.

19 Fire, A.; Xu, S.; Montgomery, M. K.; Kostas, S. A.; Driver, S. E.; Mello, C. C. Potent and specific genetic interference by double-stranded RNA in Caenorhabditis elegans. Nature 1998, 391, 806-811

20 Endo-Takahashi, Y.; Negishi, Y.; Kato, Y.; Suzuki, R.; Maruyama, K.; Aramaki, Y. Efficient siRNA delivery using novel siRNA-loaded Bubble liposomes and ultrasound. Int. J. Pharm. 2012, 422, 504-509

21 Mahmoudi, M.; Sant, S.; Wang, B.; Laurent, S.; Sen, T. Superparamagnetic iron oxide nanoparticles (SPIONs): Development, surface modification and applications in chemotherapy. Adv. Drug Deliv. Rev. 2011, 63, 24-46

144

Chapter 4

Further investigations into shape evolution of iron oxide “nano-bricks”

145

4.1 Project Goals

The main purpose of the work described in this chapter is to expand upon the work discussed in chapter 2 regarding the synthesis of brick-like iron oxide

NPs. The precise mechanism behind the evolution of the “IONBs” was difficult to elucidate with the data discussed in that chapter. As such, follow up work was conducted in which the effects of temperature, LLC phase, and the nature of the surfactant molecules were investigated with respect to particle shape. Reactions using the Triton X compounds described previously were conducted within particular LLC phases and also isotropic phases in order to better understand the possible “templating” effects of the LLC phase. Furthermore, alternative surfactants were also investigated to better understand the possible effects caused by the surfactant molecules other than those possibly caused by lyotropic phases.

Separate from the effects that reaction parameters may have on particle shape, analysis was done on certain high resolution TEM images of IONBs in an attempt to elucidate more information on the 3-dimensional structure of the particles.

Additionally, the last part of this chapter discusses earlier work on particle shape control conducted with the reduction/hydrolysis method, rather than modifications to the co-precipitation method as discussed in chapter 2. The

146 results of this are shown and a discussion as to why the synthesis was modified is given.

4.2 Additional investigations into IONP shape control and the mechanisms therein

4.2.1 Effects of temperature and lyotropic phase on IONP morphology for synthesis in Triton X surfactants

In addition to the reaction conditions detailed above, a number of experiments were also conducted at various temperatures. Such investigations were intended to allow for a better understanding of the effect of higher or lower temperatures within the same lyotropic phase on the resulting particle morphology, and to also investigate the effect that different phases and indeed isotropic conditions may have on particle morphology while still in the presence of the surfactant molecules.

Each experiment was set-up similarly to those described above in Section

2.3 (i.e. precursor concentrations and total reaction volumes remained the same), with the only differences being the temperature at which the reaction was carried out. Figure 48 shows TEM images of IONPs made in the presence of 50 wt% X45, with reaction temperatures at 25 °C, 45 °C, and 55 °C. Each of these temperatures still allows for the formation of a lamellar phase. Figure 49 shows

TEM images of IONPs made in the presence of 50 wt% X100, with reaction

147 temperatures at 25 °C, 45 °C, and 55 °C. (See section 2.2 for discussion of phase behavior properties of Triton X45 and X100).

148

Figure 48: TEM images of IONPs made in 50 wt% X45. Reaction temperatures are: (A and B) 25 °C, (C and D) 45 °C, (E and F) 55 °C. Each system remains in a lamellar phase.

149

Table 8: Summary of results for IONB syntheses in X45 at different temperatures and lyotropic phases. Temperature ( °C) Phase Particle shape

25 Lamellar Brick-like

35 Lamellar Brick-like

45 Lamellar Brick-like

55 Lamellar Smaller, shapeless

As seen in the TEM images in Figure , reactions at 25 °C and 45 °C in a medium containing Triton X45 both still result in “nanobricks” – mixtures of polyhedral, brick-like particles in the size range of tens of nanometers. These results are very similar to those seen in the published reaction conditions of 35 °C, detailed above in section 2.3. When the reaction temperature is set to 55 °C, however, much of the brick-like shape is lost and the resulting particles appear shapeless and undefined. A broad summary of the results is listed in Table 8 for comparison purposes. The results for similar experiments done in X100 complicate the picture. TEM images seen in Figure 49 show that there are effectively no “particles” at 25 °C, despite the fact that the surfactant exists in the hexagonal phase at this temperature. In contrast, brick-like particles appear at 45 and 55 °C, temperatures in which the surfactant exists in an isotropic state. Also, importantly, these brick-like particles are very similar to those seen in X45 at 35

°C, rather than those seen in X100, as seen in section 2.3 above. Table 9 offers a general summary of these results with X100 at different temperatures.

150

Figure 49: TEM images of IONPs made in 50 wt% Triton X100. Reaction temperatures are: (A and B) 25 °C, (C and D) 45 °C, (E and F) 55 °C.

151

Table 9: Summary of results for IONB syntheses in X100 at different temperatures and lyotropic phases.

Temperature (°C) Phase Particle shape

25 Hexagonal Shapeless

30 Hexagonal Small cubic and brick-like

45 Isotropic Brick-like similar to X45

55 Isotropic Brick-like similar to X45

4.2.2 Effects of alternative surfactants on IONP morphology

In order to better understand the role that the specific LLC phases may or may not have on the particle morphologies seen above in Section 2.3, experiments were conducted using Brij C10 and CTAB (cetyltrimethylammonium bromide) instead of the Triton X surfactants. Brij C10 is a non-ionic surfactant, similar to the Triton X family of surfactants, and can form both hexagonal and lamellar phases depending on the concentration ratio in water and the medium temperature.1 CTAB is an ionic surfactant that can form a hexagonal phase, as well as exist as isotropic micelles at lower concentrations.2 Figure 50 shows chemical structures for both of these molecules. The intent of these experiments was to help determine if the specific LLC phases used in a reaction directly affect or control the particle morphology. If this were the case, one would expect that the particle morphology of IONPs synthesized in, for example, a lamellar phase would not change regardless of the specific surfactant used to obtain that phase.

152

If, however, different shapes were seen then more complicated factors (i.e. medium viscosity, the preferential adsorption of the surfactant molecules on particular crystal facets during particle growth, etc.) would have to be considered for how the surfactant affects particle growth and morphology.

Figure 50: Chemical structures of Brij C10 (top, where n = 10), and CTAB (bottom).

A typical experiment was set-up similarly to those described above in

Section 2.2.2, with either Triton X45 or X100 replaced with Brij C10 or CTAB at a concentration necessary to obtain either a hexagonal or lamellar phase. The phase of the lyotropic mixture before the addition of sodium hydroxide was evaluated via POM imaging. The resulting particle morphology was evaluated by

TEM.

153

Figure 51: A and B show TEM images (A and B) of particles made in 50 wt% Brij C10 in water with Fe precursors at 40 °C. Bottom set show POM images of 50 wt% Brij C10 in water, showing the onset of a hexagonal phase beginning above 67 °C.

Figure 51 shows results of a reaction in a 50 wt% mixture of Brij C10 and water at 40 °C. Note that this temperature was chosen (rather than 35 or 30 °C as used with experiments done in the Triton X surfactants) because at lower temperatures the reaction medium becomes too viscous to allow for

154 homogeneous mixing. The POM images show the formation of a texture typical of a hexagonal phase, beginning at approximately 68 °C. This phase was maintained until room temperature. The TEM images reveal a mixture of polyhedral particles, roughly similar to those seen in reactions done in in the lamellar phase of Triton X45.

Figure 52: TEM results of particles synthesized in 0.5 M CTAB at 35 °C, an isotropic micelle phase. The images show no distinct particle shapes.

155

Figure 53: Results for particles synthesized in 1M CTAB in the hexagonal phase. A and B show TEM images, revealing brick-like particles. POM images show a type of hexagonal columnar texture beginning at 42.5 °C and persisting at the reaction temperature of 35 °C.

Figure 52 shows results for a reaction with a 0.5 M concentration of CTAB at 35

°C. The POM images did not reveal any textures, indicating that the medium was isotropic, and were thus not included. The TEM images from this reaction show ill-defined clusters with no clear examples of particles of any particular shape.

156

Figure 53 shows results for a reaction with a 1.0 M concentration of CTAB at 35

°C. The POM images of the reaction medium before addition of sodium hydroxide show a texture indicative of a columnar phase at this temperature. The

TEM images show brick-like IONPs, very similar to those seen when synthesized in the lamellar phase of Triton X45 as discussed in Section 2.3.

The results of these experiments with Brij C10 and CTAB were helpful in differentiating between the effects of a particular lyotropic phase (i.e. the anisotropic 3D structure formed by the LLC) and the effects of different surfactants (i.e. the specific molecule used to form the LLC phases). Examples of polyhedral and “brick-like” IONPs can be seen in the TEM images of particles made in the presence of both individual surfactants in conditions that allow for the formation of a hexagonal phase. This, along with previously discussed results with Triton X surfactants, suggests that the structure of the LLC phase itself is more important that the chemical structure of each individual LLC molecule since very different chemical moieties still allow for the formation of IONPs with similar morphologies.

Unfortunately, these results do not help make those discussed in Section

2.5.1, in which brick-like particles could be synthesized even in an isotropic phase, any clearer. It would appear that brick-like iron oxide particles can be made irrespective of the chemical properties of a surfactant, irrespective of the

157 specific type of LLC phase formed by the reaction medium, and in fact without the existence of an LLC phase at all.

The most that the results shown here can conclude is that the LLC phases formed by surfactants are not, in fact, templating the particles. Ultimately it would appear that a complicated interplay between the reaction conditions (eg. temperature), the growth rate of particular facets in the iron oxide crystal structure, and the existence of any kind of surfactant-like molecule in the reaction medium cause the formation of brick-like particles as opposed to spherical or amorphous particles.

4.2.3 Evaluation of visible crystal facets under TEM as a potential means to understand particle growth mechanisms

A number of papers have used high resolution TEM imaging to evaluate possible growth mechanisms in non-spherical IONPs. The idea involves measuring the spacing between visible crystal facets as seen in a TEM image of a single particle and, in conjunction with XRD data, relating this to atomic lattice planes. The position of the peaks in an XRD pattern can be related to inter- atomic lattice spacing through the Bragg equation:

nλ = 2dsin(θ) (2) where λ is the wavelength of the incident electron beam, d is the spacing between the atomic lattice, and θ is the angle between the crystal surface and the incident electron beam, with n indicating any whole number integer. This d

158 value can then be used to determine the Miller indices h, k and l. Miller indices are a form of notation that allow for the labelling of particular lattice planes and directors. They are defined as the reciprocal of the fractional intercepts along the x, y, and z axes. Figure 54 shows schematic representations of the Miller index notation system for planes and directors in a cubic lattice system.

Figure 54: Schematic representation of Miller index notation for selected planes (top, round brackets) and directors (bottom, square brackets).

159

The Miller indices for a particular crystal structure are related to the diffraction pattern determined by XRD through the following equation:

λ = (2asinθ)/(h2 + k2 + l2)1/2 (3) where a is the lattice constant, particular to a particular crystal structure. By combining equation 2 and equation 3, one can determine the Miller indices for a particular d value measured under TEM. This information can then be used to understand directions of particle growth as well as surface termination planes.

160

Figure 55: TEM images of IONPs showing measured lattice spacing with associated Miller indices, along with schematic representations of these planes with respect to incoming electron beam (indicated by a black arrow). TEM image in top right taken from Song and Zhang with permission.3 TEM image in bottom right taken is IONBX45 taken from Worden et al.

Song and Zhang used high resolution TEM to determine the termination planes of cubic shaped IONPs.3 They found that the cubic particles consistently showed a (220) lattice plane as observed under TEM, which led them to conclude that the particles were terminated preferentially along the {100} family

161 of planes (See Figure 55). Ho et al. came to similar conclusions on their own cubic IONPs, and further used this information to elucidate probable growth mechanisms for cubic shaped IONPs.4 They found that, in their high temperature decomposition method, spherical products resulted in conditions with high monomer concentrations. At lower monomer concentrations, however, cubic particles could be made. They suggested that this was the result of preferential growth along certain directions which would minimize the surface energy of the particles. At high monomer concentrations, the growth rate along different facets is equal, while at lower concentrations small differences in growth rates along different directors result in the formation of non-spherical particles. Simulation work by Davies et al. demonstrated that the {100} planes and {111} planes have the lowest overall and highest overall surface energies, respectively, in spinel oxide crystals.5 This led Ho et al. to conclude that the cubic particles were in fact the result of preferential growth along certain directors so as to minimize overall surface energy.

These types of investigations suggested that the same techniques may allow for a better understanding of the growth mechanisms behind the IONB particles discussed above. Initially, as evaluated under TEM, that IONBs were showed a d spacing corresponding to the (111) plane, which would correspond to the same termination plane, (111). This was surprising, since the {111} family of planes has the highest overall energy for spinel oxides, as noted above. This would suggest that the IONB growth mechanism could not be the result of the

162 system minimizing the surface energies of the particles, in contrast to the mechanism suggested for cubic IONPs. A more systematic evaluation of TEM images for a variety of IONBs revealed that the lattice termination plane was not, in fact, consistently (111). To better understand why this might be the case and what it meant in terms of IONB surface properties, tilting experiments under TEM imaging were conducted. This involves rotating the stage holding the TEM grid along an axis perpendicular to the electron beam. Figure 56 shows the results of this for a single IONB particle. As can be seen in the images, small changes in the angle of a particle relative to the incident electron beam result in completely different lattice planes becoming visible. This occurs despite the fact that the visible morphology of the particle remains relatively consistent. Thus the inconsistency of the visible crystal facets could be a result of the orientation of the particles with respect to the beam, rather than a result of the fact that the particles themselves have a variety of different termination planes. However, since there is no way to determine if the particles on a grid are completely “flat”, it is impossible to make general conclusions as to the nature of termination planes for these particles and, as such, the precise contribution that the reduction in surface energy may have on the IONB growth mechanism.

163

4.2.4 Reduction-hydrolysis method for IONP shape control

As noted above, initial attempts at controlling IONP morphology focused on modifications of the reduction-hydrolysis method previously developed in the

Hegmann lab. This involved introducing Triton X100 into the system as a hexagonal phase LLC. Ethanol was removed so as to not interfere or alter the formation of the lyotropic phase. FeCl3 was used instead of Fe(acac)3 simply due

164 to the latter’s poor solubility in a pure water phase. The product of this reaction is what we termed “nanosheets” of a mixture of iron and iron oxide – irregularly shaped particles several tens of nanometers in size, but constrained to only

~4nm wide. The results demonstrated the ability of this system to alter the morphology of the resulting particles and were published a few years ago. From here, it was a matter of investigating how different lyotropic systems would affect the shape of the final product.

The first set of experiments focused on duplicating the published procedure above with Triton X45, which can provide a lamellar phase in contrast to the hexagonal phase of the X100, to investigate what effect if any a different

LLC phase may have on IONP shape. In a typical synthesis, FeCl3 (2 mmol) was dissolved in 25 mL of water in a 2 neck round-bottom flask. This solution was degassed with nitrogen for one hour. After degassing, 25 mL of Triton X45

(degassed previously using the freeze-pump-thaw method for viscous liquids) was added to the above solution, and the resulting mixture was heated to 50 °C under 100 rpm mechanical stirring to ensure a homogeneous mixture was formed, at which point the reaction vessel was cooled down to 40 °C. Then,

NaBH4 (20 mmol) was dissolved in 5 mL degassed water and added to the

FeCl3/X45 mixture while under nitrogen flow and 100 rpm mechanical stirring.

The dispersion immediately turned black, indicating the formation of iron oxide, while the production of hydrogen gas caused the reaction mixture to foam. Any escaping foam was captured in a separate vessel. The reaction was left to mix

165 for one hour. The product was isolated by centrifugation and washed repeatedly with warm water to remove the surfactant. The final, dried product was a black, magnetic powder.

A large number of follow-up experiments were conducted in the hopes of obtaining a clearer picture as to how a particular LLC phase might affect and/or control particle morphology. Different variables were investigated: LLC concentration was changed (while still working within the relevant phase); and the concentration and type of iron precursor (eg. FeCl3 or Fe(acac)3) was altered.

Table 10 summarizes some of the experiments conducted, with reference to the relevant representative TEM images of the products in Figure 57. Broadly speaking within the X45 system there is little in the way of clear, reproducible results that allow for consistent control of particle morphology. A variety of different structures were formed across the various experiments – flat, sheet-like shapes; large crumpled structures that can be likened to parchment paper; small, flat disks; and simple particles of different sizes – but the specific contribution of the LLC phase on the final particle shape is not at all obvious. Just as importantly, no individual experiment resulted in a homogeneous product, revealing a lack of consistent control on the final particle shape.

166

Table 10: Selected list of reaction parameters investigated for possible IONP shape control with the reduction-hydrolysis method modified with Triton X45.

Wt % Fe precursor Visual description of TEM Reference Triton X45 images Figure 57 in water 50 FeCl3, 1 mmol Mixture of sheets and particles A of no particular shape

50 FeCl3, 2 mmol Same as above B

50 FeCl3, 4 mmol Crumpled “sheets” and flat C shapeless structures

20 FeCl3, 1 mmol Large, shapeless particles D

50 Fe(acac)3, 1 mmol Mixture of sheets and particles E of no particular shape

50 Fe(acac)3, 2 mmol Same as above F

The difficulty in obtaining any consistent, reproducible control over IONP morphology using this method could be related to the production of hydrogen gas during the borohydride reaction. This causes the Triton X45 surfactant to foam and bubble quite vigorously, which in turn must cause localized concentration gradients throughout the reaction medium which are impossible to control. This foaming effect was the main impetus for attempting IONP control using the co- precipitation method, discussed in Section 2.3 above. The co-precipitation reaction does not produce hydrogen gas and thus does not cause the surfactant to foam or bubble. This lack of foaming undoubtedly contributes to the far more consistent particle morphology seen using Triton X surfactants in the co- precipitation method.

167

Figure 57: TEM images of nanostructures made via the borohydride reduction mechanism in the presence of Triton X45. Reaction parameters are listed in Table 6.

168

4.3 References

1 Zhao; Zhou; Li Effects of Deposition Potential and Anneal Temperature on the Hexagonal Nanoporous Nickel Hydroxide Films. Chem. Mater. 2007, 19, 3882- 3891

2 Shikata, T.; Hirata, H.; Kotaka, T. Micelle formation of detergent molecules in aqueous media: viscoelastic properties of aqueous cetyltrimethylammonium bromide solutions. Langmuir 1987, 3, 1081-1086

3 Song, Q.; Zhang, Z. J. Shape Control and Associated Magnetic Properties of Spinel Cobalt Ferrite Nanocrystals. J. Am. Chem. Soc. 2004, 126, 6164-6168

4 Ho, C.; Tsai, C.; Chung, C.; Tsai, C.; Chen, F.; Lin, H.; Lai, C. Shape- Controlled Growth and Shape-Dependent Cation Site Occupancy of Monodisperse Fe3O4 Nanoparticles. Chem. Mater. 2011, 23, 1753-1760

5 Davies, M. J.; Parker, S. C.; Watson, G. W. Atomistic simulation of the surface structure of spinel. J. Mater. Chem. 1994, 4, 813-816

169

Chapter 5

Iron oxide nanoparticle / carbon dot composites

169

5.1 Research purpose and goals

The main goal of the work discussed in this chapter is to see if silanized

IONPs (see chapter 3) can be modified so as to create a multi-functional fluorescent, magnetic particle. There are a number of methods for creating fluorescent carbon dots (Cdots, discussed below) which suggest that, with slight modifications, similar procedures could be used so as to combine Cdots with

IONPs. This would allow for a very simple method for creating fluorescent IONPs without the need for additional surface functionalization with organic fluorophores. IONPs with different types of silane surface functionalities are investigated to see what, if any, effect this has on the fluorescent properties of the final particles. This work is intended as a proof-of-concept, with the hope that future work can be conducted on cell labelling and imaging using these particles.

5.2 Iron oxide nanoparticle / carbon dot composites

5.2.1 Introduction to fluorescent carbon dots

Carbon dots (also referred to as carbogenic nanoparticles, and herein referred to as Cdots) are a relatively new type of fluorescent nanomaterial. Like the more well-known and ubiquitous quantum dots (QDs), Cdots show excitation wavelength-dependent photoluminescence in the visible spectrum. In contrast to

QDs, which typically contain heavy metals, Cdots are comprised solely of organic molecules. This fact has led to investigations into Cdots for bioapplications, such as cell imaging and labelling, from which quantum dots are often restricted.1

170

Early syntheses of these materials involved non-chemical modification of carbon precursors, such as laser ablation of graphite or through electrochemical decomposition of carbon nanotubes.2 Subsequent investigations have allowed for the synthesis of Cdots through solely chemical means. Bourlinos et al. synthesized Cdots through a simple pyrolysis of various ammonium salts.3 A typical synthesis involved creating a salt of citric acid and a long chain amine, followed by high temperature calcination of the salt at several hundred degrees for a few hours. The Cdots were highly monodisperse, and could be made hydrophilic or hydrophobic, depending on the exact composition of the precursors. Wang et al. reported similar methods for Cdot synthesis in which they decomposed citric acid with molten lithium nitrate under argon4 as well as a high temperature thermolysis of citric acid in a mixture of octadecene and hexadecylamine.5 This same group also recently reported a synthesis of Cdots in which the amine group typically used as described above was attached to an organosilane compound.6 They proposed that the Cdots from this synthesis are composed of a mixed amide core, resulting from the acylation of the amine group, with the trimethoxysilane group remaining as a functionalizing group on the surface of the particles. This particular work inspired the idea that it may be possible to create composite IONP/Cdot materials through calcination of silanized IONPs, described herein.

The precise mechanism behind the photoluminescent properties of Cdots is still under debate. One possible mechanism is the formation of graphene-like

171 sheets within the particle which would thus mimic the PL process as seen in reduced graphene and graphene oxide (GO). In this process, the fluorescence arises not from quantum confinement effects, as in QDs, but rather from interactions between localized clusters, or “islands”, of sp2 hybridized carbons.7

When these clusters are in close proximity – but not touching, so as to avoid quenching - such as in stacks of graphene sheets, the band gap between π and

π* orbitals is lowered enough so that electrons can be excited by UV radiation.

Additionally, recent research on graphene oxide has revealed that C-O and C=O functional groups play an important role in tuning the band gap energy and the overall emission wavelengths seen in GO.8 Other researchers have noted that passivation by a variety of chemical functionalities can help increase the PL intensity, as well as contribute to emission wavelength tunability, in Cdots.9 This is often attributed to a second possible fluorescence mechanism involving surface defects on the particles. Defects on the surface of Cdots create carbon clusters with different hybridizations, resulting in PL emission properties similar to that seen in graphene and GO without the explicit formation of these chemical structures.10 A recent investigation by Gan et al. attempted to elucidate the PL mechanism in blue-light emitting Cdots.11 They found that the surface defect mechanism dominates for Cdots which emit predominately blue light, while Cdots which show more tunability in emission wavelengths have fewer defect sites.

Thus the PL mechanism is very dependent on the specific nature of the Cdots in

172 terms in size, internal composition, and surface functionalities, and can even depend on the synthetic method used.12

There are a number of examples in the literature of IONPs being made fluorescent for the purpose of cell imaging and labelling. This can be accomplished through the covalent attachment of an organic fluorophore, such as FITC13, rhodamine14, or fluorescent polymers15, to functionalized IONPs.

These types of fluorescent particles require a number of reaction and purification steps separate from the synthesis of the IONP core itself, however, and so a simplified process for creating fluorescent IONPs would be beneficial.

Composites of Cdots with metal and metal oxide nanoparticles have been made previously. Typically this involves functionalizing or doping the Cdot core, rather than vice-versa. Examples of this include Cdot with ZnO and ZnS,16 as well as with Au, Cu, and Pd.17 Li et al. synthesized titania and silica NPs with

Cdots physisorbed onto their surface for photocatalytic applications.18 The same group also reported a similar strategy for creating magnetic iron oxide nanoparticles (IONPs) coated with Cdots.19 In each of these cases the core particles are several hundred nanometers in size, and the metal oxide particles and Cdots must be made separately before being combined in a composite material. Wang et al. created smaller clusters of IONPs of approximately 50 nm in size decorated with Cdots through a one-pot solvothermal method.20 The

173 synthesis involved a reaction between ferrocene and concentrated hydrogen peroxide in an autoclave, and required 48 hours at 200 °C.

Herein we report a simple method for creating magnetic IONPs with blue- light emitting fluorescence properties. The IONPs are created through simple aqueous means and are functionalized in situ with hydrophilic silane compounds.

Calcination of these particles at 200 °C yields water dispersible particles that visibly fluoresce under UV light.

5.2.2 Materials and instrumentation

Surface functionalization was analyzed through FT-IR using KBr pellet techniques. Approximately 1 mg of dried particles were mixed with approximately

The amount of surface ligands on the particles was estimated via a TA instruments TGA Q500. The heating rate was set at 10 °C/min. Powdered

174 samples were typically dried in a vacuum oven at 50 °C for 2 hours before analysis in order to eliminate any surface water.

TEM imaging was done with a FEI Tecnai TF20 TEM instrument at an accelerating voltage of 200 kV. Particle samples were dispersed in methanol and dropcast onto 400 mesh carbon coated copper grids.

Fluorescence measurements were conducted on a Spectramax Gemini

EM Microplate Reader with a xenon flash lamp at 100-240 VAC, 50/60 Hz. A 1 cm transparent 4-wall cuvette was used to hold the samples.

5.2.3 Experimental methods

Silanized IONPs: A typical experiment was conducted as follows. IONPs were first synthesized following the reduction-hydrolysis method by

Yathindranath et al., discussed at length in Chapter 1. These bare particles were then functionalized with some combination of two different kinds of silane compounds: (3-aminopropyl)triethoxysilane (termed AmS) and N-

(trimethoxysilylpropyl)ethylenediaminetriacetate (termed EDTS), following the method published by Yathindranath et al. This was done by injecting the silane solution directly into the reaction medium containing the IONPs 1 hour after particle formation. Three sets of particles were made this way. The first used just

AmS (AmS-IONPs). The second used both AmS and EDTS in a 1:1 molar ratio

(AmS-EDTS-IONPs). The third used just EDTS (EDTS-IONPs). In each case the concentrations of silane to iron precursor was kept at a 10:1 molar ratio during

175 the reaction. The particles in each case were purified by dialysis against water, and then isolated via rotavap. The dried powders were then washed several times with water/ethanol mixtures and collected with an external magnet before being dried in a vacuum oven. Portions of each of the purified particles were kept as a powder to be used in the subsequent calcination step.

Calcination: Samples of the dried powders obtained above were placed in an oven under air at 200 °C for 45 minutes. The resulting powders were redispersed in water using an ultrasonication horn (at a concentration of approximately 1 mg/mL) for fluorescence measurements. These products after calcination were labelled AmS-IONP/Cdots, AmS-EDTS-IONP/Cdots, and EDT-

IONP/Cdots.

5.3 Results and discussion

A schematic representation of the overall reaction can be found in Figure

58. The accompanying pictures (of AmS-EDTS-IONP/Cdots) show that the calcined particles can be redispersed in water (when sonicated) and that the solution is visibly photoluminescent when illuminated under 365 nm UV light.

Dispersions of the same particles before the calcination process do not show visible photoluminescence, indicating that the calcination process is integral in the modification of the particle surface.

176

Figure 58: Schematic representation of the process for creating IONP/Cdot composites. Top image shows an IONP coated with AmS (left structure) and EDTS (right structure). Bottom left image shows a schematic of the particle after calcination, along with a picture of the calcined particles under ambient light (bottom middle), and illuminated under 365 nm UV light (bottom right).

The photoluminescent properties of each particle set dispersed in water, before and after calcination, were measured via fluorescence spectroscopy. The particle dispersions were excited with light at wavelengths between 250 and 400 nm, in 25 nm increments. Figure 59 shows the results for AmS-IONPs; Figure 60 shows the results for AmS-

EDTS-IONPs; and Figure 61 shows the results for EDTS-IONPs. In each case, the initial particle dispersions before calcination show little to no significant photoluminescence. Post-calcination, each particle set displays

177 significantly enhanced photoluminescence, with a marked difference in which excitation wavelength corresponding to the maximum emission peak. The AmS-IONP/Cdots show broad, asymmetrical emission peaks between 400 and 450 nm, with a max emission corresponding to a 325 nm excitation wavelength. AmS-EDTS-IONP/Cdots show narrower, more symmetrical peaks, with three roughly equal max emission peaks centered around 400 nm and corresponding to 325, 350, and 375 nm excitation wavelengths. EDTS-IONP/Cdots show the least intense emission peaks of the three particle sets, with broad, asymmetrical emission peaks centered around 450 nm, with a max emission roughly equal between 325, 350 and

375 nm.

Figure 59: Fluorescence spectra for AmS-IONPs before (left) and after (right) calcination.

178

Figure 60: Fluorescence spectra for AmS-EDT-IONPs before (left) and after (right) calcination.

179

Figure 61: Fluorescence spectra for EDT-IONPs before (left) and after (right) calcination.

TEM imaging was done on each set of particles before and after calcination. There is no visible difference in particle morphology for any set of particles. This is not necessarily surprising, however, since the iron oxide core, approximately <10 nm in diameter, provides the majority of electron density and thus image contrast under TEM. The thin surface silane layer does not provide enough contrast to be seen, before or after calcination. These images do demonstrate that the IONPs have not been destroyed by the calcination treatment, and that the particles remain individually separated. Figure 62 shows representative images of the particles before and after calcination.

180

Figure 62: Representative TEM images of silanized IONPs (left) and IONP/Cdots (right).

Figure 63: FT-IR spectra of AmS-IONPs before (bottom) and after (top) calcination.

181

Figure 64: FT-IR spectra of AmS-EDTS-IONPs before (bottom) and after (top) calcination.

Figure 65: FT-IR spectra of EDTS-IONPs before (bottom) and after (top) calcination.

182

FT-IR spectroscopy was used to characterize the surface chemistry of the particles before and after calcination. Surprisingly, the calcination process causes little notable differences in surface functionality of the particles. In Figure 63, AmS-IONPs and AmS-IONPs/Cdots are compared.

The only observable differences between the spectra involve the relative intensity of the broad peaks between 3200–3500 cm-1, but this is likely due to changes in the amount of adsorbed moisture which causes the O-H vibrational stretch. Figure 64 shows spectra for AmS-EDTS-IONPs and

AmS-EDTS-IONPs/Cdots. Once again the same functional groups are present in each spectrum. The only significant difference involves a decrease after calcination in relative intensity of the broad peak centered at approximately 1610 cm-1. This peak can be associated with both asymmetric vibrational stretching of carboxylate groups and scissoring of protonated amine groups.21 Figure 65 shows spectra for EDTS-IONPs and

EDTS-IONP/Cdots. As before, there is little evidence of significant changes in surface functionality between the particle sets. In their synthesis of oganosilane Cdots, Wang et. al note, through FT-IR, the formation of amide groups as the main significant difference in chemical moieties between the precursors and the Cdot products. There is no clear evidence for that in the IONP/Cdot composites, although it is possible that such a peak – centered at 1654 cm-1 – could be masked by peaks associated with

183 asymmetric vibrational stretching of carboxylate groups and scissoring of protonated amine groups.

5.4 Conclusions

While the results herein make it difficult to determine with complete certainty the precise mechanism behind the PL properties of the

IONP/Cdot composites, some broad conclusions about the nature of these materials can be made. From TEM, there is no evidence of the formation of separate Cdots on the surface of the IONPs, or of any signs that the particles have become clustered due to reactions between the surfaces of the particles. This suggests that any reactions or decomposition processes arising from the calcination process occur on the surface of individual

IONPs. Additionally, the PL data show that the emission wavelengths are largely not tunable with respect to changes in the excitation wavelength.

Rather, the composite particles emit predominately blue light. Differences in surface composition affect the PL emission spectra more than does the excitation wavelength (AmS-IONP/Cdots show max emission centered around 400 nm at an excitation wavelength of 325 nm, compared with

AmS-EDTS-IONP/Cdots which show max emission centered at 450 nm at an excitation wavelength between 350 and 375 nm). With these points in mind, quantum confinement effects arising from changes in Cdot size can be eliminated. Interactions between graphene sheets would also be

184 expected to cause some degree of emission wavelength tunability seen in other materials containing such material, and so this mechanism can also be eliminated. As such, we suggest that the formation of sp2 defect carbon clusters in thin layers on the surface of the IONPs is the primary contributors to the PL properties of these IONP/Cdot composites. Further work needs to be done on the precise effect that particular functional groups have on this process, as well as the effect that additional passivation with functional molecules post-calcination may have on PL intensity as seen in other publications.

5.5 References

1 A) Yang, S.; Cao, L.; Luo, P. G.; Lu, F.; Wang, X.; Wang, H.; Meziani, M. J.; Liu, Y.; Qi, G.; Sun, Y. Carbon Dots for Optical Imaging in Vivo. J. Am. Chem. Soc. 2009, 131, 11308-11309 B) Yang, S.; Wang, X.; Wang, H.; Lu, F.; Luo, P. G.; Cao, L.; Meziani, M. J.; Liu, J.; Liu, Y.; Chen, M.; Huang, Y.; Sun, Y. Carbon Dots as Nontoxic and High-Performance Fluorescence Imaging Agents. J. Phys. Chem. C 2009, 113, 18110-18114

2 A) Sun, Y.; Zhou, B.; Lin, Y.; Wang, W.; Fernando, K. A. S.; Pathak, P.; Meziani, M. J.; Harruff, B. A.; Wang, X.; Wang, H.; Luo, P. G.; Yang, H.; Kose, M. E.; Chen, B.; Veca, L. M.; Xie, S. Quantum-Sized Carbon Dots for Bright and Colorful Photoluminescence. J. Am. Chem. Soc. 2006, 128, 7756- 7757 B) Zhou, J.; Booker, C.; Li, R.; Zhou, X.; Sham, T.; Sun, X.; Ding, Z. An Electrochemical Avenue to Blue Luminescent Nanocrystals from Multiwalled Carbon Nanotubes (MWCNTs). J. Am. Chem. Soc. 2007, 129, 744-745

3 Bourlinos, A. B.; Stassinopoulos, A.; Anglos, D.; Zboril, R.; Karakassides, M.; Giannelis, E. P. Surface Functionalized Carbogenic Quantum Dots. Small 2008, 4, 455-458

185

4 Wang, F.; Kreiter, M.; He, B.; Pang, S.; Liu, C. Synthesis of direct white- light emitting carbogenic quantum dots. Chem. Commun. 2010, 46, 3309- 3311

5 Wang, F.; Pang, S.; Wang, L.; Li, Q.; Kreiter, M.; Liu, C. One-Step Synthesis of Highly Luminescent Carbon Dots in Noncoordinating Solvents. Chem. Mater. 2010, 22, 4528-4530

6 Wang, F.; Xie, Z.; Zhang, H.; Liu, C.; Zhang, Y. Highly Luminescent Organosilane-Functionalized Carbon Dots. Advanced Functional Materials 2011, 21, 1027-1031

7 Eda, G.; Lin, Y.; Mattevi, C.; Yamaguchi, H.; Chen, H.; Chen, I.; Chen, C.; Chhowalla, M. Blue Photoluminescence from Chemically Derived Graphene Oxide. Adv Mater 2010, 22, 505-509

8 Shang, J.; Ma, L.; Li, J.; Ai, W.; Yu, T.; Gurzadyan, G. G. The Origin of Fluorescence from Graphene Oxide. Sci. Rep. 2012, 2, 792

9 A) Chandra, S.; Pathan, S. H.; Mitra, S.; Modha, B. H.; Goswami, A.; Pramanik, P. Tuning of photoluminescence on different surface functionalized carbon quantum dots. RSC Adv. 2012, 2, 3602-3606 B) Yang, Y.; Cui, J.; Zheng, M.; Hu, C.; Tan, S.; Xiao, Y.; Yang, Q.; Liu, Y. One-step synthesis of amino-functionalized fluorescent carbon nanoparticles by hydrothermal carbonization of chitosan. Chem. Commun. 2012, 48, 380- 382

10 Lim, S. Y.; Shen, W.; Gao, Z. Carbon quantum dots and their applications. Chem. Soc. Rev. 2015, 44, 362-381

11 Gan, Z.; Wu, X.; Hao, Y. The mechanism of blue photoluminescence from carbon nanodots. CrystEngComm 2014, 16, 4981-4986

12 Baker, S.; Baker, G. Luminescent Carbon Nanodots: Emergent Nanolights. Angewandte Chemie International Edition 2010, 49, 6726-6744

13 Chekina, N.; Horak, D.; Jendelova, P.; Trchova, M.; Benes, M. J.; Hruby, M.; Herynek, V.; Turnovcova, K.; Sykova, E. Fluorescent magnetic nanoparticles for biomedical applications. J. Mater. Chem. 2011, 21, 7630- 7639

14 Cuiling Ren and Jinhua Li and Xingguo Chen and Zhide Hu and,Desheng Xue Preparation and properties of a new multifunctional material composed of superparamagnetic core and rhodamine B doped silica shell. Nanotechnology 2007, 18, 345604

15 Howes, P.; Green, M.; Bowers, A.; Parker, D.; Varma, G.; Kallumadil, M.; Hughes, M.; Warley, A.; Brain, A.; Botnar, R. Magnetic Conjugated Polymer

186

Nanoparticles as Bimodal Imaging Agents. J. Am. Chem. Soc. 2010, 132, 9833-9842

16 Sun, Y.; Wang, X.; Lu, F.; Cao, L.; Meziani, M. J.; Luo, P. G.; Gu, L.; Veca, L. M. Doped Carbon Nanoparticles as a New Platform for Highly Photoluminescent Dots. J. Phys. Chem. C 2008, 112, 18295-18298

17 Tian, L.; Ghosh, D.; Chen, W.; Pradhan, S.; Chang, X.; Chen, S. Nanosized Carbon Particles From Natural Gas Soot. Chem. Mater. 2009, 21, 2803-2809

18 Li, H.; He, X.; Kang, Z.; Huang, H.; Liu, Y.; Liu, J.; Lian, S.; Tsang, C.; Yang, X.; Lee, S. Water-Soluble Fluorescent Carbon Quantum Dots and Photocatalyst Design. Angewandte Chemie International Edition 2010, 49, 4430-4434

19 Zhang, H.; Ming, H.; Lian, S.; Huang, H.; Li, H.; Zhang, L.; Liu, Y.; Kang, Z.; Lee, S. Fe2O3/carbon quantum dots complex photocatalysts and their enhanced photocatalytic activity under visible light. Dalton Trans. 2011, 40, 10822-10825

20 Wang, H.; Wei, Z.; Matsui, H.; Zhou, S. Fe3O4/carbon quantum dots hybrid nanoflowers for highly active and recyclable visible-light driven photocatalyst. J. Mater. Chem. A 2014, 2, 15740-15745

21 De Palma, R.; Peeters, S.; Van Bael, M. J.; Van, d. R.; Bonroy, K.; Laureyn, W.; Mullens, J.; Borghs, G.; Maes, G. Silane Ligand Exchange to Make Hydrophobic Superparamagnetic Nanoparticles Water-Dispersible. Chem. Mater. 2007, 19, 1821-1831

187

Appendix

Supplementary paper on pharmacological properties of S-IONBsX45

The following paper is intended as a supplement to the work outlined in

Chapter 2. The particles synthesized and described in that chapter were used in cell toxicity and uptake investigations by Zhizhi Sun, a graduate student in Dr.

Don Miller’s lab at the University of Manitoba in Winnipeg, Canada. This paper describes a number of important and unique properties of the materials described in Chapter 2, but has not yet been accepted for publication. As such, it has been included here for the edification of the members of the dissertation committee.

188

Differential Internalization of Brick Shaped Iron Oxide Nanoparticles by Endothelial Cells

Zhizhi Sun1, Matthew Worden2, Yaroslav Wroczynskyj3, James A. Thliveris4, Johan van

Lierop3, Torsten Hegmann1,2,5, Donald W. Miller1.

1Department of Pharmacology and Therapeutics, University of Manitoba, Winnipeg,

Manitoba, Canada

2Department of Chemistry and Biochemistry, Kent State University, Kent, OH, U.S.A.

3Department of Physics and Astronomy, University of Manitoba, Winnipeg, Manitoba,

Canada

4Department of Human Anatomy and Cell Sciences, University of Manitoba, Winnipeg,

Manitoba, Canada

5Chemical Physics Interdisciplinary Program, Liquid Crystal Institute, Kent State

University, Kent, OH, U.S.A.

Correspondence: Donald W. Miller

Department of Pharmacology and Therapeutics, 710 William Avenue, University of

Manitoba, Winnipeg, Manitoba, Canada. R3E 0T6

Tel +1 204 789 3278

Fax +1 204 789 3932

Email [email protected]

189

Abstract

Nanoparticles targeting endothelial cells for potential therapeutic delivery to treat diseases such as cancer, oxidative stress, and inflammation were traditionally achieved by grafting ligand on the surface for binding to cell membrane receptors expressed on the endothelial cells. The present studies examined the effect of nanoparticle shape in regulating preferential uptake in endothelial cells compared to epithelial cells. Spherical and brick shaped iron oxide nanoparticles (IONPs) were synthesized with identical negatively charged surface coating. The accumulation of the various IONPs was examined in cultured endothelial and epithelial cells in presence and absence of external magnetic field. The nanobricks showed a significantly greater uptake profile in endothelial cells compared to nanospheres. Magnetic field exposure drastically enhanced the uptake of nanobricks but not nanopheres. Transmission electron microscopy revealed differential internalization of nanobrick in endothelial cells compared to epithelial cells. Increased expression of caveolin-1 was found in endothelial cells compared to epithelial cell lysates and nanobricks were able to interfere with caveolae-mediated endocytosis process. Furthermore, the cellular uptake of nanobricks was significantly reduced in the presence of methyl beta cyclodextran and genistein, inhibitors of caveolae-mediated endocytosis. These data collectively support the differential internalization of nanobricks in endothelial cells via caveolae mediated endocytosis. Our approach offers a general means to achieve delivery to vascular endothelium without a typical ligand receptor interaction.

190

Keywords: shape, iron oxide nanoparticles, drug delivery, nanobrick, endothelial cells, differential uptake

191

Highlights

 Nanobricks show improved cellular uptake profiles compared to nanospheres

despite of negative surface charge.

 The increased magnetic properties of the nanobricks resulted in enhanced

uptake in the presence of an external magnetic field.

 Caveolae dependent endocytosis mediate preferential uptake of nanobricks in

endothelial cells.

192

Introduction

There is a growing interest in developing iron oxide nanoparticles (IONPs) as platforms for drug delivery applications.[1-3] In this regard, IONPs provide several advantages: 1)

The ability to target to areas of interest using externally applied magnetic field, thereby increasing local therapeutic concentration and decreasing potential toxicity related to systemic circulation. 2) Monitoring capabilities for IONPs using MRI. 3) Favorable biocompatibility profile. 4) Flexibility of surface modification to build multifunctional complexes for advanced drug delivery applications.

Drug delivery applications for intracellular targets require internalization of the delivery system. The interaction between IONPs and cell membrane is largely determined by their physiochemical properties such as surface coating and shape.[4, 5] Our group has previously examined the effect of surface charge on cellular uptake of IONPs.[6] The positively charged IONPs have a significantly higher uptake profile compared to negatively charged ones, likely due to electrostatic interaction between positively charged IONPs and negatively charged plasma membrane of the cell. However, the charge related effects on internalization are non-specific as they were present in a variety of different cell types.[6] As a result, negatively charged nanoparticles have emerged as better candidates for drug delivery due to longer circulation times and reduced clearance.

Various pathological conditions such as cancer, cardiovascular disease, inflammation, and oxidative stress would benefit from the preferential delivery of nanoparticles to the vascular endothelium.[7-9] To achieve the cell specific delivery, targeting ligands are

193

often grafted onto the NPs to increase the delivery efficiency. For instance, intracellular adhesion molecule (ICAM), vascular cell adhesion molecule (VCAM), and platelet- endothelial cell adhesion molecule (PECAM-1) have been used to target endothelial cells.[10-12] However, this approach is often associated with variability in outcome due to variable receptor expression level between patients or heterogeneity in tumor endothelial cells within a patient.[13] Therefore, a generalized approach that preferentially target endothelial cells without ligand receptor interaction would be favorable.

To date, there are few reports concerning non-spherical nanoparticles. More specifically, only a handful of studies provide side-by-side comparison of spherical and non-spherical nanoparticle interactions with biological milieu. Shape may also play a role with regards to nanoparticle cell interaction. For example, recent work with theoretical modeling revealed the role of nanoparticle shape and membrane rigidity on cellular uptake.[14]

Nonetheless, our understanding of the role of IONPs shape in cellular internalization is rather limited. Recent advances in fabrication techniques have enabled generation of brick shaped IONPs. We hypothesize that alteration of IONP shape can influence both the cellular uptake in endothelial cells and the ability to augment cell uptake with application of an external magnetic field. In this work, we studied the uptake profile of nanospheres and nanobricks in various cell types. We investigated the mechanism of preferential uptake of the nanobrick IONPs in endothelial cells. By understanding the relationship between NP shape and cell surface domains, our work provides insight into the development of IONPs for specifically targeting endothelial cells.

194

195

Materials and Methods

Materials

All chemical reagents were purchased from Sigma Aldrich (St. Louis, MO) and cell culture reagents from Invitrogen Canada Inc. (Burlington, ON) unless otherwise specified.

Nanoparticle synthesis and characterization

Sphere shaped iron oxide nanoparticles were prepared under mild conditions at room temperature as previously described.[15] They were prepared by adding N-

(trimethoxysilylpropyl)ethylenediaminetriacetate trisodium salt (EDT, 3 mmol, from a solution concentration of 45% in water) (Gelest, Morrisville, PA) directly to a reaction vessel containing IONPs . The mixture was allowed to react overnight with stirring and the final product was purified by dialysis (MWCO 30000) against deionized (DI) water over 48 hours and was freeze dried and resuspended in sterile PBS to prior to experiments. Brick shaped IONPs was synthesized as described. (See attached paper for information only)

Nanoparticle crystallographic properties were probed with powder x-ray diffraction experiments using a Bruker D8 Davinci diffractometer (Billerica, MA). Both nanoparticle formulations were identified as iron-oxide through pattern-matching with a database.

Reitveld refinement of the collected diffraction patterns allowed for an assessment of the average crystallite size.

The IONP size distribution in DI water was determined initially through photon correlation spectroscopy (PCS) at a fixed scattering angle (90°) using a Horiba Nano-Partica SZ-

196

100 series instrument (Horiba Instruments Inc., Irvine, CA). The same instrument allowed for the assessment of particle surface charge (zeta potential) by the measurement of IONP electrophoretic mobilities using phase analysis light scattering.

Further characterization of the nanoparticle hydrodynamic size was done by photon correlation spectroscopy at multiple scattering angles (60°,90° and 120°) using a

Photocor FC Complex (Moscow, Russia). The scattered light intensity at each angle was fit to a distribution of Brownian relaxation times, related to the particle hydrodynamic size through the viscosity and refractive index of the suspending medium (DI water). The magnetization of dry nanoparticle powder samples were recorded at room temperature as a function of applied magnetic field (0 – 4 T) using a Quantum Design MPMS XL

SQUID magnetometer (San Diego, CA).

Cell culture

A mouse brain derived microvessel endothelial cell line, bEnd.3 (American type tissue culture collection, Manassas, VA), was used as a cell culture model of the blood-brain barrier (BBB). The bEnd.3 cells (passage number 15-30) were cultured in DMEM

(Hyclone, Logan, UT) supplemented with 10% heat-inactivated FBS (Hyclone, Logan,

UT), 50 U/mL penicillin and streptomycin (MP Biomedicals, Solon, OH) at 37°C and 5%

CO2. Cells were expanded in T-75 tissue culture flasks, and seeded at 2x104 cells per cm2 on 6 or 12 well plates for uptake and cytotoxicity studies, respectively. Culture medium was changed every 2 days. All experiments were performed on confluent monolayers (typically 4-5 days post seeding).

Cellular Uptake of IONP compositions

197

Confluent monolayers of bEnd.3 cells grown on 6-well culture plates (Costar, Lowell, MA were treated with culture media containing either nanosphere or nanobrick compositions

(2.5μg/mL – 100μg/mL of Fe). After treatment with IONPs, cells were placed in a

he IONP solutions were removed and the cell monolayers were washed 3X with ice cold phosphate buffered saline (PBS) to remove unbound nanoparticles. Cells were lysed by the addition of 500

μl of 0.2 N NaOH and IONPs content determined based on the ferrozine assay described below. Cellular accumulation was examined in both the presence and absence of a static magnetic field created by placing the cells over a platform containing cylindrical rare earth magnets (19mm diameter, 3mm height) (Lee Valley, Winnipeg,

MB). Cells remained in the magnetic field for the duration of the experiment.

For mechanistic studies of IONP uptake, cells were pretreated with chlorpromazine

-beta- cytochalas

1 h at 37 °C in the presence of the various endocytotic inhibitors. Cell association of nanobrick was determined as described below.

Additional markers of caveolae mediated endocytosis, alexa fluor 488-labeled cholera toxin subunit B (CTB) and tetramethylrhodamine conjugated bovine serum albumin

(BSA) were examined for cellular uptake. For these studies, cells were exposed to CTB

llowing 15-min pretreatment with various concentrations of the iron-oxide nanobricks. Cells were washed and lysed and fluorescence determined using a Synergy HT reader.

Analytical assay for measuring IONPs

198

Quantitative determination of IONPs content in cell and media samples was performed using the Ferrozine assay. As the Ferrozine assay is an absorbance based assay for determining soluble iron concentrations, IONPs in the cell lysate and media samples were first solubilized by adding 500 µL of concentrated HCl (~12M) to 500 µL of cell lysate or media samples. This mixture was incubated for 1 h at room temperature with gentle shaking and then neutralized with 500µL of 12M NaOH. Once the samples were neutralized, 120 µL of hydroxylamine hydrochloride (2.8 M) in 4M HCl was added and the samples incubated for 60 min at room temperature with gentle shaking. Following this incubation, 50uL of 10M ammonium acetate solution (pH 9.5) and 300 uL of 10mM ferrozine in 0.1M ammonium acetate solution was added to each sample. Absorbance was measured at 562 nm using a Synergy HT plate reader (BioTek, Winooski, VT).

Quantitative assessment of IONP concentration was based on a standard curve prepared using 1000 ppm iron atomic absorption standard (Fisher Scientific, Ottawa,

ON). Samples from the cell lysates were normalized for protein content using BCA protein assay kit (Pierce, Rockford, IL).

Electron Microscopy

The cellular localization of IONPs compositions was examined using transmission electron microscopy. For these studies, cells were incubated with IONPs at 50μg/mL concentration in media for 2 hours. After incubation, cells were washed 3X with PBS and collected using 0.25% trypsin EDTA (Hyclone, Logan, UT). After centrifugation, the cell pellets where fixed in 3% glutaraldehyde in 0.1M phosphate buffer (pH 7.3), followed by post-fixation in 1% osmium tetroxide in 0.1M phosphate buffer (pH 7.3). Cells were then dehydrated and embedded in Epon 812 using standard techniques. (13764136) Thin sections were stained with uranyl acetate and lead citrate, viewed and photographed in

199

a Philips CM 10 electron microscope. In order to eliminate observer bias, sections were examined without foreknowledge of their source.

Statistical analysis

All data were expressed as mean ± SEM. All values were obtained from at least three independent experiments. Statistical significance was evaluated using one-way ANOVA followed by post-hoc comparison of the means using the Fisher's least significant difference test.

200

Results

Physico-chemical characterization of IONPs

Physico-chemical parameters of the nanobrick and nanosphere compositions are provided in Figure 1. Both nanobricks and nanospheres have silanized coating and free carboxylic acid functional groups on the surface resulting in zeta potentials of approximately -40 mV. The TEM images confirmed the different shapes of IONPs.

Based on the TEM images, the dimensions of IONP core for the nanobricks are approximately 20 nm wide and 5 nm thick and the nanosphere is around 10 nm diameter. The hydrodynamic diameters in water as determined by photon correlation spectroscopy at 90° were 50 ± 2 nm and 30 ± 1 nm for nanobrick and nanosphere, respectively. A more complete evaluation of the size of the nanobrick sample in suspension using photon correlation spectroscopy at multiple scattering angles revealed a larger hydrodynamic diameter of 80 ± 40 nm (Figure S2). The saturation magnetization, determined by fitting the high field magnetization to a straight line after background subtraction (diamagnetic signal from the sample holder), was 50 ± 5 A m2 kg-1 and 10 ± 2 A m2 kg-1 for the nanobricks and nanospheres, respectively (Figure S3).

The saturation magnetization is the largest magnetization that a material can exhibit in an applied magnetic field. Samples with larger saturation magnetizations have greater magnetic response and thus are more favorable for targeted delivery using externally applied magnetic fields. A more detailed description of the nanoparticle characterization is provided in the Supplementary Information.

Preferential internalization of nanobrick in endothelial cells

201

Quantitative uptake analysis was performed in the bEnd.3 mouse brain endothelial cell line. (Figure 2a) There was a significantly greater uptake of nanobrick compared to

presence of external magnetic field, cell association of nanobrick was substantially increased compared to nanosphere. At the highest concentration examined

-fold and 10-fold increase in uptake of nanobricks compared to nanospheres with and without a magnetic field, respectively. This surprising finding suggests that despite the negative surface charge, brick shaped IONPs are taken up by brain endothelial cells to a greater extent than spherical counterparts. Furthermore, the shape of IONPs affected their magnetization value and ability to be influenced by application of external magnetic field.

Uptake studies with the nanobrick and nanospheres was expanded to include primary human lung and brain endothelial cells as well as Madin-Darby canine kidney (MDCK) epithelial cell line with two fold purpose: 1) To investigate whether there was any selectivity in endothelial cells; and 2) To examine whether enhanced uptake of the nanobricks was specific to brain endothelial cells compared to other endothelial beds. In the absence of magnetic field, a similar uptake profile for nanobricks were observed in

was greater cell association in epithelial cells compared to endothelial cells. (Figure 2b).

p to a greater extent in lung endothelial cells compared to brain endothelial cells. In presence of a magnetic field, the uptake of nanobricks in both brain and lung endothelial cells were enhanced, whereas no significant increase in uptake in MDCK cells were observed. To further confirm this

202

observation, transmission electron microscopy (TEM) from the various cell preparations were examined. The TEM images confirmed that nanospheres were loosely bound on the cell surface and not internalized by bEnd.3 cells (Figure 3a) or human hepatocellular liver carcinoma cell line HepG2 (Figure 3b). In contrast, large amounts of nanobricks were found inside the bEnd.3 cells (Figure 3c) but few were found inside HepG2 (Figure

3d) or MDCK cells (Figure 3e) confirming that nanobricks were selectively internalized in endothelial cells.

Internalization of nanobrick in bEnd.3 cells via caveolae mediated endocytosis

To understand the selectivity of nanobricks to endothelial cells, we examined the potential mechanism of internalization via endocytosis. Confluent bEnd.3 cell monolayers were pretreated with inhibitors for clathrin mediated endocytosis

in), macropinocytosis

(cytochalasin D), and endosome maturation (monensin) for 30 min, and uptake of nanobricks was determined (Figure 4a). There was a significant inhibition of nanobrick

findings were confirmed in TEM studies showing diminished IONP association in bEnd3 in the presence of genistein compared to controls receiving the nanobricks alone (Figure 4b,c).

None of the other treatment groups examined significantly impacted on nanobrick accumulation in bEnd3 cells (Figure 4a). These data collectively suggested that nanobrick internalization in bEnd.3 cells was mediated via a caveolae dependent endocytosis pathway.

203

To confirm whether elevated caveolae mediated endocytosis in endothelial cells contributes to the selective internalization of nanobricks observed, additional studies were performed with known markers of caveolae-mediated endocytosis. The uptake of

CTB and BSA is 6-fold and 12-fold greater in bEnd.3 cells than MDCK cells, respectively. (Figure 5a) The increase in uptake of CTB and BSA in the bEnd3 was correlated with an increase in the expression of caveolin-1 compared to epithelial MDCK cell line. Expression of caveolin-1 in another endothelial cell line hCMEC/D3 was also elevated. (Data not shown) Additional evidence of potential interaction of nanobricks in caveolae-mediated endocytosis is the ability of the nanobricks inhibit the uptake fluorescently-labeled BSA in a concentration dependent manner. (Figure 5b)

204

Discussion

There are a few publications reporting the synthesis and fabrication of different shaped

IONPs,[16, 17] however, most of these methods are thermal decomposition methods so the resulting particles are not directly dispersible in water. Shape-dependent differences in the biological properties of IONPs are limited to macrophages.[18] The results of the present study are the first to demonstrate that brick shaped IONPs may be preferentially taken up by endothelial cells. The selective uptake of the nanobricks by endothelial cells appears to be due to caveolae-mediated endocytosis, which is more prevalent in endothelial cells compared to epithelial cells examined. As nanospheres with identical surface coating and similar iron oxide core compositions were not internalized by either endothelial or epithelial cells, the present studies strongly support the shape of the IONP as having the predominant role in the observed phenomena.

It has been shown that rod shaped polystyrene NPs have enhanced antibody binding specificity compared to spherical and disk shaped NPs.[19] Using in silico and in vivo approaches, Kohlar and colleagues demonstrated rod shaped polystyrene NPs with antibody against intracellular adhesion molecule (ICAM) or transferrin receptor exhibit higher internalization in brain and lung endothelial cells than spherical counterparts under flow conditions.[20] These studies showed a shape induced accumulation of antibody conjugated NPs in endothelial cells in vivo. The amount of rod shaped NPs distributed in lung tissue was significantly higher than spherical shaped NPs.

From a biophysical standpoint, modification of shape could enhance the affinity of a negatively charged surface coating to lipid raft domain on the cell surface. An increased

205

contact area with the cell surface provides potentially more sites for interaction and has been previously identified as an important contributor to enhance nanoparticle targeting effects.[21] It is hypothesized that by changing the IONPs from sphere shaped to brick shaped, the negatively charged surface coating may interact at multiple, discrete sites on the cell membrane that contributes to selective binding of nanobricks to endothelial cells. This may provide fundamental advantages especially when second generation nanobrick compositions are created that have additional ligand targeting capabilities.

Caveolae are formed by a group of caveolin protein binding to cholesterol in lipid raft region of the cell membrane.[22] Although surface chemistry and functional groups can influence IONP cell interaction, it has been reported that negatively charged IONPs can interact with cationic lipid domains in the lipid raft.[23] Caveolae are enriched in endothelial cells and present in muscle, fibroblast, and adipocytes.[24] Following the pinch off of caveolae from the lipid raft, the fate of caveolae is dependent on the cell type in which endocytosis occurs. In non-endothelial cells, caveolae are subjected to endolysosomal system. In endothelial cells, caveolae may bypass the lysosome and transport cargo through vesicular processes across the endothelial cell layer.[25, 26] For this reason, the nanobrick IONPs may potentially be exploited for drug and gene delivery applications to tissues underlying endothelial cells such as the brain. These studies are currently ongoing.

The surface coating indeed plays a role in targeting endothelial cells. Several line of evidence suggested that positively charged sphere shaped NPs show increased cellular uptake in various cell types including endothelial cells.[6, 27-29] It was reported that the

206

negatively charged NPs with non-spherical in shape exhibit increased uptake efficiency in endothelial cells via clathrin mediated pathway.[30] However, no preferential internalization was observed in endothelial cells compared to other cell types. In the present study, the use of an inhibitor of clathrin mediated endocytosis had no effect on nanbrick IONP uptake in endothelial cells. The lack of clathrin-mediated internalization in our study may be due to the smaller diameter IONPs used compared to those previously examined (in the range of 100 nm).

Of the various internalization processes, caveolae mediated endocytosis is predominantly found in endothelial cells.[31] Therefore, targeting to endothelial cells may be achieved by interacting with caveolae localized in lipid rafts within the plasma membrane. The current study certainly points to a caveolae-mediated mechanism for the endothelial selective uptake of the nanobrick IONP. The evidence in support of this is the increased expression of caveolin in endothelial cells compared to the epithelial cells examined and the ability of inhibitors of caveolae-mediated uptake to significantly reduce nanobrick IONP accumulation in endothelial cells. In addition, the nanobrick IONPs could be used in a concentration-dependent manner to prevent the cellular uptake of two macromolecules, CTB and BSA, that are known to enter into endothelial cells through caveolae-mediated endocytosis. Previous studies grafting anionic polyelectrolytes of varied hydrophobicity to nanospheres reported endothelial cell targeting of NPs via a caveolae-mediated endocytic process.[32] These findings together suggest that non- spherical nanoparticles with negative surface charges are likely to have the greatest affinity for caveolae-based uptake.

207

Magnetic targeting presents a promising potential for drug delivery application. In presence of an external magnetic field, there is a significant increase in uptake of nanobricks compared to nanospheres. This is also reflected on the saturation magnetization. The shape and size of IONPs affects anisotropy energy that in turn governs the relaxivity. A small change in shape may have drastic effect on longitudinal and transverse relaxivities.[33] Nanobricks show potent and constant transverse relaxivity (r2) for medium and high-field MRI where gadolinium based contrast agent peaks at 20 MHz and decreases sharply at high magnetic fields. (See attached paper for information only) The change in shape and increased size results in an increased saturation magnetization value for the nanobricks compared to the nanosphere compositions. This increased magnetization observed for the nanobrick IONPs leads to increased responsiveness in a magnetic field gradient and will orient more completely with the magnetic field. As the nanobricks have a preferred direction of magnetization along their largest dimension (see Supplementary Information), an externally applied magnetic field will act to more preferentially align the smaller dimensions of the nanobrick along the cell surface, decreasing the area of interaction and thus limiting the effects of steric repulsion between the cell surface and nanobrick coating.

The preferential uptake of nanobrick IONPs within vascular endothelial cells combined with the enhanced targeting through application of external magnetic fields has several potential therapeutic applications. The ability to target to the endothelial cells within tumor microvasculature is a prime application for this technology platform. It is generally accepted that angiogenesis is crucial for tumor growth, evasion and metastasis.[34] The creation of new blood vessels to supply oxygen and nutrients to tumor cells is a

208

necessary requirement for solid organ tumor growth. Thus, anti-angiogenesis therapy has emerged as a viable treatment strategy to control tumor growth. Recent studies demonstrated the potential of PEG-PLGA nanoparticles for tumor neo-vasculature and tumor cells dual-targeting drug delivery.[35] The ability to focus an external magnetic field within the tumor stroma will not only increase the local concentration of IONPs but also facilitate improved internalization of nanobrick IONPs in endothelial cells. An anticipated result of such focused targeting of the IONPs would be enhanced delivery and potential destruction of the tumor neovasculature. While current anti-angiogenic therapies have been limited in the clinic due to the development of resistance, [36] the targeting of nanobrick IONPs to endothelial cells using shape and magnetic fields would make resistance to these delivery vehicles less probable.

209

Conclusion

Nanoparticle shape plays an important role in the cellular internalization process.

Targeting nanoparticles to endothelial cells can be achieved modifying shape from sphere to brick. Nanobricks exhibited improved cellar uptake profile compared to nanospheres counterparts despite of negative surface charge. The increased magnetic properties of the nanobricks resulted in enhanced uptake in the presence of an external magnetic field. The preferential uptake of nanobricks in endothelial cells was mediated via caveolae dependent endocytosis. Our results demonstrate that shape modification offers a general approach to achieve targeted delivery without the need for receptor- ligand interaction.

210

Figures

Fig.1

Fig 1 Physicochemical properties and TEM image of nanosphere and nanobricks.

Hydrodynamic size and charge were measured in triplicate samples using a Nano- partica SZ-100 series instrument from Horiba. Values represent the mean ± SEM (n=3).

211

Fig. 2

Fig 2 Cellular accumulation of nanobricks and nanospheres in bEnd.3 cells (A). Uptake of nanobricks in MDCK, primary human lung and brain endothelial cells (B).

Experiments were performed in the presence and absence of external magnetic field.

Values are expressed as the mean + SEM for three cell monolayers per treatment group. *** p<0.001

212

Fig. 3

Fig 3 Electron microscopy images of nanospheres (a, b) and nanobricks (c, d, e) in bEnd.3 (a, c), HepG2 (b, d), and MDCK cells (e). The arrows point to nanoparticles.

213

Fig. 4

Fig 4 Effect of various endocytosis inhibitors on cellular uptake of nanobricks in bEnd.3 cells. The internalization of nanobricks were significantly decreased by treatment with

mean + SEM for three cell monolayers per treatment group; * p<0.05 compared to control. This is confirmed by TEM that shows substantially greater internalization of nanobricks under control conditions (b) compared to cells treated with genistein (c).

214

Fig. 5

Fig 5 The caveolae dependent uptake mechanism and expression of caveolin-1 in bEnd.3 and MDCK cells (a). The ability of nanobricks to inhibit the uptake of fluorescently labeled BSA suggests a competitive binding of the nanobricks to the caveolae (b). Post hoc analysis indicate that each treatment group are significantly different (at least p < 0.01) from other groups. Values are expressed as the mean + SEM for three cell monolayers per treatment group. *** p<0.001

215

Acknowledgments

This study was funded by research grants from the Collaborative Health Research

Program sponsored by the Canadian Institutes of Health Research and Natural Science and Engineering Research Council of Canada (DWM, JvL). This work was also financially supported by the Ohio Third Frontier Ohio Research Scholar Program

“Research Cluster on Surfaces in Advanced Materials” (TH). Graduate student fellowship support provided by the Natural Science and Engineering Research Council of Canada (ZS) and the University of Manitoba (YW).

216

Reference

[1] A.K. Silva, N. Luciani, F. Gazeau, K. Aubertin, S. Bonneau, C. Chauvierre, D. Letourneur, C. Wilhelm, Nanomed., 11 (2015) 645-655. [2] M.E. Mertens, S. Koch, P. Schuster, J. Wehner, Z. Wu, F. Gremse, V. Schulz, L. Rongen, F. Wolf, J. Frese, V.N. Gesche, M. van Zandvoort, P. Mela, S. Jockenhoevel, F. Kiessling, T. Lammers, Biomaterials, 39 (2015) 155-163. [3] J.P. Chen, P.C. Yang, Y.H. Ma, Y.J. Lu, J. Nanosci. Nanotechnol., 11 (2011) 11089- 11094. [4] S. Laurent, A.A. Saei, S. Behzadi, A. Panahifar, M. Mahmoudi, Expert Opin. Drug Deliv., 11 (2014) 1449-1470. [5] K. Kaaki, K. Herve-Aubert, M. Chiper, A. Shkilnyy, M. Souce, R. Benoit, A. Paillard, P. Dubois, M.L. Saboungi, I. Chourpa, Langmuir, 28 (2012) 1496-1505. [6] Z. Sun, V. Yathindranath, M. Worden, J.A. Thliveris, S. Chu, F.E. Parkinson, T. Hegmann, D.W. Miller, Int. J. Nanomedicine, 8 (2013) 961-970. [7] J. Mai, Y. Huang, C. Mu, G. Zhang, R. Xu, X. Guo, X. Xia, D.E. Volk, G.L. Lokesh, V. Thiviyanathan, D.G. Gorenstein, X. Liu, M. Ferrari, H. Shen, J. Control. Release, 187 (2014) 22-29. [8] E.D. Hood, M. Chorny, C.F. Greineder, S.A. I, R.J. Levy, V.R. Muzykantov, Biomaterials, 35 (2014) 3708-3715. [9] T. Novokhatska, S. Tishkin, V. Dosenko, A. Boldyriev, I. Ivanova, I. Strielkov, A. Soloviev, Eur. J. Pharmacol., 718 (2013) 401-407. [10] J. Hsu, T. Bhowmick, S.R. Burks, J.P. Kao, S. Muro, J. Biomed. Nanotechnol., 10 (2014) 345-354. [11] M. Dan, D.B. Cochran, R.A. Yokel, T.D. Dziubla, PLoS ONE, 8 (2013) e81051. [12] H. Yang, F. Zhao, Y. Li, M. Xu, L. Li, C. Wu, H. Miyoshi, Y. Liu, Int. J. Nanomedicine, 8 (2013) 1897-1906. [13] K. Hida, N. Ohga, K. Akiyama, N. Maishi, Y. Hida, Cancer Sci., 104 (2013) 1391- 1395. [14] S. Dasgupta, T. Auth, G. Gompper, Nano Lett., 14 (2014) 687-693. [15] Z. Sun, M. Worden, Y. Wroczynskyj, V. Yathindranath, J. van Lierop, T. Hegmann, D.W. Miller, Int. J. Nanomedicine, 9 (2014) 3013-3026. [16] B. Karagoz, J. Yeow, L. Esser, S.M. Prakash, R.P. Kuchel, T.P. Davis, C. Boyer, Langmuir, 30 (2014) 10493-10502. [17] C. de Montferrand, L. Hu, I. Milosevic, V. Russier, D. Bonnin, L. Motte, A. Brioude, Y. Lalatonne, Acta Biomater., 9 (2013) 6150-6157. [18] W. Xia, H.M. Song, Q. Wei, A. Wei, Nanoscale, 4 (2012) 7143-7148. [19] S. Barua, J.W. Yoo, P. Kolhar, A. Wakankar, Y.R. Gokarn, S. Mitragotri, Proc. Natl. Acad. Sci. U. S. A., 110 (2013) 3270-3275. [20] P. Kolhar, A.C. Anselmo, V. Gupta, K. Pant, B. Prabhakarpandian, E. Ruoslahti, S. Mitragotri, Proc. Natl. Acad. Sci. U. S. A., 110 (2013) 10753-10758. [21] E.M. Munoz, J. Correa, R. Riguera, E. Fernandez-Megia, J. Am. Chem. Soc., 135 (2013) 5966-5969. [22] M. Murata, J. Peranen, R. Schreiner, F. Wieland, T.V. Kurzchalia, K. Simons, Proc. Natl. Acad. Sci. U. S. A., 92 (1995) 10339-10343. [23] I.M. Adjei, B. Sharma, V. Labhasetwar, Adv. Exp. Med. Biol., 811 (2014) 73-91.

217

[24] E.J. Smart, G.A. Graf, M.A. McNiven, W.C. Sessa, J.A. Engelman, P.E. Scherer, T. Okamoto, M.P. Lisanti, Mol. Cell. Biol., 19 (1999) 7289-7304. [25] A. El-Sayed, H. Harashima, Mol. Ther., 21 (2013) 1118-1130. [26] M. Simionescu, D. Popov, A. Sima, Cell Tissue Res., 335 (2009) 27-40. [27] N. Ma, C. Ma, C. Li, T. Wang, Y. Tang, H. Wang, X. Moul, Z. Chen, N. Hel, J. Nanosci. Nanotechnol., 13 (2013) 6485-6498. [28] E. Rampazzo, R. Voltan, L. Petrizza, N. Zaccheroni, L. Prodi, F. Casciano, G. Zauli, P. Secchiero, Nanoscale, 5 (2013) 7897-7905. [29] A. Verma, F. Stellacci, Small, 6 (2010) 12-21. [30] R. Agarwal, V. Singh, P. Jurney, L. Shi, S.V. Sreenivasan, K. Roy, Proc. Natl. Acad. Sci. U. S. A., 110 (2013) 17247-17252. [31] R.G. Parton, K. Simons, Nat. Rev. Mol. Cell Biol., 8 (2007) 185-194. [32] J. Voigt, J. Christensen, V.P. Shastri, Proc. Natl. Acad. Sci. U. S. A., 111 (2014) 2942-2947. [33] E.D. Smolensky, H.Y. Park, Y. Zhou, G.A. Rolla, M. Marjanska, M. Botta, V.C. Pierre, Journal of materials chemistry. B, Materials for biology and medicine, 1 (2013) 2818-2828. [34] S.M. Weis, D.A. Cheresh, Nat. Med., 17 (2011) 1359-1370. [35] G. Gu, Q. Hu, X. Feng, X. Gao, J. Menglin, T. Kang, D. Jiang, Q. Song, H. Chen, J. Chen, Biomaterials, 35 (2014) 8215-8226. [36] J.M. Ebos, R.S. Kerbel, Nat. Rev. Clin. Oncol., 8 (2011) 210-221.

218