Fluorescent and Magnetic Nanocomposites for Multimodal Imaging

THESIS

Presented in Partial Fulfillment of the Requirements for the Degree Master of Science in the Graduate School of The Ohio State University

Dhananjay Thakur

Graduate Program in Biophysics

The Ohio State University

2010

Master's Examination Committee:

Dr. Jessica O. Winter, Advisor

Dr. Ronald Xu

Dhananjay Thakur

2010

Abstract

Nanoparticles possessing photoluminescent and magnetic properties have been sought after for their applicability in fields ranging from single molecule biophysics to clinical imaging. Here, three independent strategies have been described to synthesize multifunctional nanocomposites. 1) pH sensitive, fluorescent–magnetic, nanocomposites were created through a simple aqueous procedure. Separately synthesized superparamagnetic iron oxide nanoparticles and mercaptopropionic acid (MPA)-coated

CdS quantum dots were crosslinked using 3-mercaptopropyl trimethoxysilane (MPS) as a bifunctional linker to yield CdS–iron oxide conjugates. Conjugates formed clusters of

0.1–1.0 μm diameter, with the smallest observed particle diameter ∼50 nm. Particle solubility and photoluminescent (PL) intensity were sensitive to solution pH, with the highest PL intensity and stability obtained at pH values <3.0 and MPS:Cd:Fe ratios of

1:10:1. 2) Polylactic-glycolic acid (PLGA) and carbon nanoparticle conjugates were created for live cell imaging as a potentially non-toxic alternative to heavy metal based quantum dots. Carbon nanoparticles from carbon soot were synthesized in a simple process through refluxing in concentrated nitric acid for > 12 hours. Photoluminescent properties showed a dependence on excitation wavelength. The PLGA encapsulation scheme allows for addition of other modalities such as magnetic nanoparticles or therapeutic agents. Particles displayed long-term stability in live cell cultures without photobleaching or blinking as seen in organic dyes or quantum dots. 3) Blue-fluorescent ii silica nanoparticles were obtained that displayed a higher quantum yield than previously described carbon nanoparticles. For both carbon and silica nanoparticles rapid endocytosis was seen in cultured human gioblastoma multiforme cells. Given these unique properties each particle type might find applications in biosensing, optical tracking and separation.

iii

Acknowledgments

None of this work presented here would have been possible without the influence and contributions of many remarkable individuals I met along the way over the last few years.

I am deeply grateful to Dr. Winter for her guidance, mentoring and, above all, her superhuman patience. I learnt much from her – not only in the importance of maintaining experimental rigor in my studies and ultimately putting together exquisitely crafted presentations and publications, but also in appreciating the very rewarding roles we can play in education, out-reach and mentorship for those younger than us; all, while managing a healthy work-life balance (or at least making an honest attempt to do so). My lab members – even those who were only passing through, made remarkable impressions and were all wonderful to work with. I would especially like to thank and acknowledge the contributions of Sean Hawkins, Thierno Baldet, Ellis Robinson and Lee Siers who not only did a lot of grunt work but also enthusiastically contributed to the intellectual parts of our projects. The cheerful, helpful and very talented members from Dr. Atom Sarkar‘s and Dr. Sooryamukar‘s groups were a delight to work with. Their contributions within my work have been indispensible. Dr. Hendrik Colijn helped me in the most critical parts of nanoparticle characterization – electron microscopy. Without his assistance, advice and the occasional very lightly issued scolding over really bad sample preps- the images that we desperately needed would have been impossible to take. I would also like to

iv thank Ning Han, Shreyas Rao, Dr. Gang Ruan, Kunal Parikh and Darian Richardson for their advice, help, contributions and long discussions. Without my closest friends in

Columbus, Yukti Aggarwal, Deepti Vikram, David Harris, Franco Merea, Renji Thomas,

Kevin Slaten and Ali Hassanali, life would have been indefinably lesser. Thank you all for sticking through, through the good, bad and turbulent times. I am grateful to Dr. Mike

Zhu and Dr. Ralf Bundschuh, in their roles as advisers and directors of the Biophysics program, for always being helpful, encouraging and supportive beyond expectations.

I would also like to thank the Managing Editor, Copyright and Permissions,

Nanotechnology for granting me permission to use material from our publication: pH sensitive CdS-iron oxide fluorescent-magnetic nanocomposites. Dhananjay Thakur,

Shuang Deng, Thierno Baldet, and Jessica Winter, Nanotechnology 20, 485601 (2009).

Vita

2002...... B.Sc. Physics, Fergusson College,

University of Pune

2005...... M.Sc. Physics, University of Pune

2005-2007 ...... Assistant Manager, Tata Motors Ltd.

2007 to present ...... Graduate Research Associate,

Biophysics Graduate Program,

The Ohio State University

Publications

1. Thakur D., Deng S., Baldet T.,Winter J. O., Nanotechnology. 20(48): p. 485601. 2009

2. Ruan G., Vieira G., Henighan T.,Chen A., Thakur D., Sooryakumar R., Winter J. O.

Nano Letters May 7 2007 (Web pub.).

Fields of Study

Major Field: Biophysics

Table of Contents

Abstract ...... ii

Acknowledgments...... iv

Vita ...... vi

List of Figures ...... x

Chapter 1: Introduction and Research Summary ...... 1

Chapter 2: Review of Fluorescent Magnetic Nanocomposites ...... 7

2. 1 Design and Synthesis of Quantum dot and Magnetic nanoparticle composites: ..... 7

2.1.1 Synthesis of QD-MNP heteromers: ...... 12

2.1.2 Synthesis of core-shell QD-MNP composites: ...... 12

2.1.3 Synthesis of composites through co-encapsulation: ...... 13

2.1.4 Tissue Targeting and Viability ...... 14

2.1.5 Toxicity ...... 17

2.2. Applications of Fluorescent Magnetic Composites: ...... 18

2.2.1 Cell separation and Single Molecule Manipulation: ...... 18

2.2.2 Imaging and therapy: ...... 19 vii

2.2.3 Potential applications ...... 21

Chapter 3: pH sensitive CdS–iron oxide fluorescent–magnetic nanocomposites ..... 26

3.1. Introduction ...... 27

3.2. Materials and methods ...... 30

3.2.1. Chemicals and instrumentation ...... 30

3.2.2. Synthesis of iron oxide nanoparticles ...... 30

3.2.3. Synthesis of CdS nanoparticles ...... 31

3.2.4. Synthesis of nanoparticle composites ...... 31

3.2.5. Nanocomposite characterization ...... 31

3.3 Results ...... 32

3.3.1. Nanocomposite morphology ...... 32

3.3.2. Nanocomposite absorbance ...... 37

3.3.3. Nanocomposite photoluminescence ...... 39

3.3.4. Nanocomposite magnetic properties ...... 41

3.4. Discussion ...... 42

3.5 Supporting information ...... 49

Chapter 4: Fluorescent carbon particles for live cell imaging ...... 51

4.1 Introduction ...... 51

4.2 Methods: ...... 53

viii

4.2.1 Carbon nanoparticle synthesis: ...... 53

4.2.2 PLGA nanosphere synthesis: ...... 54

4.2.3 Characterization: ...... 54

4.3 Results: ...... 55

4.4 Silica/carbon nanoparticles: ...... 62

4.5 Conclusion:...... 63

Chapter 5: Conclusion ...... 65

List of references ...... 67

List of Figures

Figure 1.1 Cadmium sulfide quantum dots ...... 2

Figure 1.2. TEM: Iron Oxide nanoparticles: Fe2O3-Fe3O4 ...... 3

Figure 1.3 pH sensitive CdS-Iron oxide composites ...... 5

Figure 2.1 Current paradigms in fluorescent magnetic composite design ...... 8

Table 2.1 Recent developments in fluorescent-magnetic composite design ...... 9

Figure 2.2 Magnetic manipulation of fluorescent magnetic micelles ...... 19

Figure 2.3 Live cell imaging with fluorescent magnetic micelles ...... 21

Figure 3.1 (a), (b) TEM images of Iron Oxide and CdS nanoparticles ...... 33

Figure 3.1 (c) STEM and TEM of composites ...... 34

Figure 3.2. STEM and EDS of CdS–iron oxide clusters ...... 35

Figure 3.3 Absorbance Spectra ...... 36

Figure 3.4. Fluorescence from MPA-capped CdS QDs and CdS–iron oxide composites 37

Figure 3.5. PL spectra for CdS–iron oxide composites ...... 38

Figure 3.6 Measurement of magnetic properties ...... 41

Figure 3.7 Fluorescence images of fluorescent magnetic composites ...... 44

Figure 3.8. XPS for bare iron oxide particles indicating that the crystals are Fe2O3-Fe3O4

...... 49

Figure 3.9. Size distribution histogram from STEM images of CdS-Iron oxide composites

...... 49

Figure 3.10 PL spectra for CdS QDs with varying pH ...... 50

Figure 4.1 Representation of an N-V center ...... 52

Figure 4.2 Fluorescence from CNP solution...... 55

Figure 4. 3 (a), (b) CNP emission and excitation spectra ...... 58

Figure 4.3 (c) CNP absorbance ...... 59

Figure 4.4 Electron and optical microscopy for size and structure determination ...... 60

Figure 4.5 Live cell imaging with PLGA-CNPs and bare CNPs ...... 61

Figure 4.6 Fluorescence and live cell imaging with silica nanoparticles ...... 63

Chapter 1: Introduction and Research Summary

Nanometer sized crystals of a wide assortment of compositions and designs are new additions in the biomedical imaging toolbox. They offer possibilities of novel therapeutic procedures and imaging modalities at length and time scales that have not been accessible with current imaging techniques. For example, magnetic resonance imaging, positron emission tomography or computed tomography, which have been sufficiently developed to become mainstream clinical techniques for tumors, clots, aneurysms and other abnormalities, do not yet provide the spatial and temporal resolutions required for single- molecule, organelle, or single-cell tracking. These requirements can be critical in cases such as rapidly metastasizing tumors. The use of bright, stable and highly localizable markers, including nanoparticles, is an attractive solution for this requirement.

Semiconductor nanoparticles, also referred to as quantum dots (QDs, figure 1.1)) were developed initially for applications in opto-electronics(1), but their potential to transform other areas such as biomedical imaging was quickly realized. Soon, methods were developed to solubilize these crystals in water by capping them with appropriate ligands(2), which triggered a revolution in nanoparticle based biomedical research.

Fluorescent nanoparticles can replace dyes for a number of reasons. Quantum dots (QDs) provide better photostability, wide absorption profiles, longer fluorescence lifetimes (10 – 1

100ns compared to 1- 10ns for organic dyes) and a high degree of customizability due to their ease of synthesis.

Figure 1.1 Cadmium sulfide quantum dots

Figure 1.1. Aqueous based cadmium sulfide (CdS) quantum dots. (a) Fluorescence emission maximum as a function of particle diameter (smallest particles are green-blue, largest particles are red) ( excitation = 365 nm). Emission wavelength increases proportional to particle size, which in this case was regulated by modifying the concentration of surface ligand mercaptopropionic acid added during synthesis. (b)

Transmission electron microscopy (TEM) image of ‗yellow‘ CdS quantum dots (a, center) showing particle diameter ~5nm.

In a different class of nanoparticles are small engineered clusters of magnetic materials that, owing to size confinement effects, display superparamagnetism: paramagnetic behavior below the Curie temperature. Their size permits temperatures below the Curie temperature to provide enough thermal energy to modify the magnetization of the entire crystallite. These properties have led to their use in many applications, some of which 2 are: as micromechanical force transducers for gene therapy(3), drug delivery(4), cell separations(5), manipulation of mechanosensitive ion channels(6), image-contrast enhancing agents in magnetic resonance imaging (MRI)(7, 8), multimodal imaging (9,

10) and for use in hyperthermia for thermal ablation of tumors(10). Shown below (Figure

2) are iron oxide nanoparticles prepared in aqueous medium through co-precipitation of iron salts in a base.

Figure 1.2. TEM: Iron Oxide nanoparticles: Fe2O3-Fe3O4

A combination of magnetic resonance and fluorescence imaging in cancer through the use of multimodal particles would greatly benefit diagnosticians and surgeons in indentifying tumors, and more significantly, distinguishing tumor boundaries from healthy tissue(11). Yet, creating stable particles with multiple functionalities has been challenging. Chapter 2 in this thesis is a literature review of the latest developments in the field, covering multimodal nanocomposite synthesis strategies, their applications and

3 an additional section discussing the still nascent field of fluorescent carbon nanoparticles

– both graphitic and diamond.

One of the major research undertakings in our group has been to create multi-functional particles with a specific application focus on tumor identification. During this effort we discovered a hitherto unknown functionality arising from functionalization of quantum dot surfaces with mercapto-propyl trimethoxysilane and mercaptopropionic acid(12).

Chapter 3 will present this data (reproduction of published work). Briefly, while developing a synthesis strategy to create fluorescent–magnetic composites for simultaneous in-vivo fluorescence and MRI tracking, we observed pH sensitive luminescence in the reaction product. These observations led to a method to synthesize pH sensitive fluorescent-magnetic nanocomposites. Separately synthesized superparamagnetic iron oxide nanoparticles and mercaptopropionic acid (MPA)-coated

CdS quantum dots were crosslinked using 3-mercaptopropyl trimethoxysilane (MPS) as a bifunctional linker to yield CdS-iron oxide conjugates (figure 1.3). The composites formed clusters 0.5 to 1.5 µm in diameter, with the smallest observed particle diameter

~50nm. Particle solubility and photoluminescent intensity were sensitive to solution pH with the highest PL intensity and stability obtained at pH values < 3.0 and MPS:Cd:Fe ratios of 1:10:1. pH sensitivity is believed to result from changes in nanoparticle solubility within the silica matrix. Given these unique properties, this material might find application in endosomal tracking, cell separations and pH sensitive detection.

Figure 1.3 pH sensitive CdS-Iron oxide composites

Figure 1.3. pH sensitive CdS-Iron oxide composites. (a) Solution containing composites; the fluorescent composites show attraction to a permanent magnet (to the right of the sample) (b) TEM shows silica encapsulated constituent CdS and iron oxide particles.

Scale bar is 10nm. (c) Schematic for an ideal silica encapsulated cluster of CdS QDs and iron oxide NPs.

However, the inherent toxicity presented by materials containing cadmium restricts the use of cadmium-based quantum dot composites to in vitro applications only. Hence, to achieve a less toxic product we examined carbon-based fluorescent nanostructures.

Following a previously described protocol by Mao et al., we synthesized carbon nanoparticles from candle soot. Hydrophobic carbon soot was refluxed in concentrated nitric acid for >14 hrs to obtain green-fluorescent carbon particles. We subsequently adapted this synthesis for biomedical imaging applications by scaling up the procedure, identifying optimum excitation wavelengths not investigated previously, incorporating these particles in PLGA nanospheres, and demonstrating in vitro cell uptake. The maximum emission peak was obtained while under excitation of 465nm wavelength.

Initial experiments with non-biofunctionalized PLGA spheres have shown non-specific endocytosis into human glioblastoma cells. The structure of these particles has yet to be defined. Transmission electron micrographs show structures below 5nm. The luminescence of these particles most likely originates from nitrogen-vacancy centers, or

N-V centers, created during the acid reflux. These N-V centers are typically defects introduced in carbon lattices by the substitution of one carbon atom with a nitrogen and an adjacent carbon with an empty site, giving rise to photoluminescence in the crystal.

These particles could also be used for multimodal imaging by incorporating additional constituent particles in the PLGA microspheres. Chapter 4 will detail this study.

In conclusion, in chapter 5, these findings will be summarized along with a brief discussion of a third particle design featuring blue-fluorescent silica particles. Possible future directions in which these particles can be further developed and adapted to specific biomedical requirements will also be discussed.

Chapter 2: Review of Fluorescent Magnetic Nanocomposites

Nanocomposites containing fluorescent and magnetic particles have served as a basic and popular test-bed for multiplexed nanoprobe designs. The two components have previously and independently been well characterized(13, 14). A major consideration behind combining these modalities is their ability to provide good spatiotemporal resolutions(15). This section will review the recent developments in synthesis and applications of methods incorporating semiconductor quantum dots (QDs) with iron oxide or gadolinium based particles, as well as those including other fluorophores such as organic dyes, graphitic carbon particles and nano-diamonds. The field of nanocomposite applications is still in a nascent state. Hence, studies employing only quantum dots or magnetic particles are also discussed in cases where the targeting and delivery techniques can be extended to nanocomposites.

2. 1 Design and Synthesis of Quantum dot and Magnetic nanoparticle composites:

A variety of methods have been devised to create composites of fluorescent semiconductor QDs and magnetic nanoparticles (MNPs). Three of the major design paradigms are: 1) Blending the two materials to create a single heteromeric particle with both optical and magnetic properties, 2) Encapsulation of independently synthesized fluorescent and magnetic particles in a polymer or silica matrix, 3) Encapsulation of

7 individual particles in a polymer or silica shell. These and others have been summarized in Table 2.1.

Of these, micellar encapsulation(16, 17) and encapsulation in mesoporous silica or silica shells(18-20) have been some of the more widely explored strategies. Representations of these paradigms are shown in Figure 2.1. Syntheses of the individual components of composites – QDs and magnetic nanoparticles have been reviewed extensively (10, 13,

21-30) and will not be discussed here.

Figure 2.1 Current paradigms in fluorescent magnetic composite design

Figure 2.1 (a) Co-encapsulation of independently synthesized QDs and iron oxide NPs in amphiphilic polymers (b) Surface modified ‗particle-blends‘, shown in core-shell configuration here but which may attain dumbbell shaped morphology as well (c) QDs and SPIONs embedded in silica particles (d) Biofunctionalized composites.

Table 2.1 Recent developments in fluorescent-magnetic composite design

Composition Morphology Size Quantum Saturation Ref. yieldc magnetization (emu/g)d

Iron oxide,CdTe Embedded core-shella ~40nm, -- 5.8 (31), Embedded core-shell 30nm, -- 2.26 (32), Co-embedded particlesb -- ~40%, 30.5 (33), Co-embedded particles ~50nm, -- 13.1 (34)

FeP, CdTe Dumbbell shaped ~6nm ~25% 20 (35)

CoPt-core, CdSe- Core-shell -- 3-5% 4.4 (36) shell

FePt, CdS Dumbbell shaped ~7nm ~3.2% -- (37)

Fe2O3, CdSe Dumbbell shaped -- 8-10% -- (38)

Fe O , CdS Co-embedded particles ~50nm ~20% 9.5 (12) Quantum 3 4 dot based Fe3O4-core, CdSe- Embedded core-shell 60-80nm ~12% 3.2 (39) ZnS shell

Fe3O4-core, CdSe- Core-shell 8nm 10-15% -- (40) ZnS shell

Co/CdSe Core-shell 2-3nm 2-3% -- (41)

CdSe/ZnS, Core-shell 25 nm 15% 0.76 (42) perfluorocarbon

CdTeSe-ZnS, Iron Surface immobilized ~100nm -- -- (43) oxide QDs on iron oxide core

CdTeSe-ZnS, Iron ------(44) oxide on carbon nanotubes continued… 9

Table 2.1 Continued CdSe/ZnS, Iron oxide Micelle encapsulated 35nm -- -- (45)

CdSe/ZnS,γ-Fe2O3 Micelle encapsulated 25nm 15% 0.76 (46)

CdSe/ZnS, iron oxide Micelle encapsulated 60-70nm -- -- (47)

Ferritin, CdSe/ZnS Cross linked particles ~25 nm -- 70 (48)

Gadolinium –doped Co-embedded particles 50-200nm -- -- (49) gold-speckled silica

Mn doped Si --- ~5nm 8.1% -- (50)

Bacterial magnetic Cross linked particles ------(51) particles, QDs

CdSe-ZnS, Au, Ag, γ- Co-embedded particles ∼20-35 nm -- 1.77 (52) Fe2O3

FITC, Fe3O4 Core-shell 13.8±5.3nm -- 53.47 (53), Embedded core-shell 73 nm -- 0.6-0.9 (54)

Rhodamine, Iron oxide Core-shell ------(55), Core-shell 36.8-85.8nm -- 57-60 (56), Liposome encapsulated ------(57) rhodamine Cross linked particles 16±4nm -- 43 (58) isothiocyanate (RITC),iron oxide

FITC/RITC, iron oxide Surface immobilized iron 50-100nm -- -- (59) oxide on silica particles oligothiophene Core-shell 30–400 nm -- -- (60) fluorescents, Fe2O3

Eu-DOTA (DOTA Functionalized zeolite 30nm -- -- (61) =tetra- crystals azacyclododecane- tetraacetic acid),Gd- DOTA- zeolite L crystals naphthyl methacrylate, Surface immobilized iron ~100nm -- -- (62) iron oxide oxide on silica particles

NaGdF4:Er3+,Yb3+/ -- 20-41nm -- -- (63) NaGdF4 continued…

Table 2.1 Continued fluorescent pyrene, iron Core shell ~250nm -- 0.44 (64), oxide Core shell ~200nm -- 1.26 (65)

Ferrocene, carbon -- -- ~1 ~10 (66) nanopaticles

Iron-oxide,sulfur- Core shell ~7nm 12% -- (67) oxidized diarylethene

Fe3O4@LaF3:Ce,Tb -- 30nm -- -- (68)

Fe2O3-Congo red -- ~15nm -- -- (69) nanoparticles

Tb-doped- Fe2O3 -- 13 nm 2.4 30 (70) nanocrystals

Hydroxyapatite, Core-shell 30nm -- -- (71) Eu3+,Gd3+

a where either QD/iron oxide are core particles surrounded by a shell on which iron oxide/QD particles are embedded. b where both types of particles are embedded in the same shell in a non-sequential manner. c Quantum yield of a fluorophore is a measurement of its fluorescence efficiency. It is typically defined as:

[Eqn 1]

Where, is the quantum yield of the sample, is the known quantum yield of a standard sample – typically a fluorescent dye that shares the same excitation wavelength as the sample. and are the gradients from plots of integrated fluorescence intensity vs. absorbance, and and are the refractive indices of the solvent. d The saturation magnetization of a magnetic nanoparticle indicates the maximum possible magnetization of a material in a given magnetic field. Hence, it as an indicator of the particle‘s efficacy as an MRI contrast agent.

2.1.1 Synthesis of QD-MNP heteromers:

QD-MNP heteromers typically appear in TEM analysis, as dumbbell shaped particles (in the case of heterodimers – each lobe of the ‗dumbbell‘ comprising of the individual constituent particles), or as one of the constituent particles, either QD or MNP forming a core, with the multiple particles other type fusing on the surface of the core(35, 37, 38).

It has been observed that in high temperature decomposition steps during the synthesis of nanocomposites, heterodimers of the constituent QD and MNP may be formed(35).

Temperature, ratios of reactants and choice of solvents used are also important factors that control the heterodimerization and surface stabilization. It is believed that interstitial lattice mismatch leads to a heteromeric structure at high temperatures in nanoparticles in close proximity or even as an evolution of a core shell particle where lattice mismatch exists between the constituent particles. Properties of the constituents – fluorescence or magnetism can be retained, though they may be differ from those of the original QDs and

MNPs as a result of impurities introduced in the particles at the heteromer interface.

2.1.2 Synthesis of core-shell QD-MNP composites:

Core shell composites may be comprised of an MNP core around which a QD layer is assembled or vice versa. The core is chosen as a seed particle that behaves as a nucleation site around which the second layer of material is deposited and allowed to grow and crystallize(72). In the case of magnetic cores and fluorescent shells it has been seen in some cases that, though magnetic properties remain unaffected(36, 40), the fluorescence properties are significantly affected through a reduction in quantum yield(36). In such

12 cases a recovery in quantum yield can often be seen simply after passivating the fluorescent semiconductor layer with a higher band gap material such as ZnS(40). Other groups have reported a loss in saturation magnetization of the magnetic core upon coating with semiconductor material without loss in coercivity(41). Precise conditions under which the core remains may remain unaffected have yet to be elucidated.

2.1.3 Synthesis of composites through co-encapsulation:

One approach for co-encapsulation of individually synthesized NPs is to employ silica or crosslinked polymer shells in which these particles are embedded. Another is to create liposomes or micelles that incorporate these particles. Of these, amphiphilic block copolymers show better results than lipid-PEG(Poly-ethylene glycol) in forming micelles that can incorporate multiple NPs(45). This is because the longer hydrophobic chains in amphiphilic block copolymers such as in poly(styrene-b-ethylene oxide), allow for more core material to be added without modifying the micellar structure as opposed to lipid-

PEG which is destabilized by the core material owing to its shorter hydrophobic chains.

A critical parameter that is taken into consideration for micelle based composite designs is p, or the packaging parameter: defined as the ratio of hydrated area of the hydrophobic chain to the molecular volume(73, 74). This parameter can be used to predict the structure formed in water: spherical if p < 1/3, cylindrical if 1/3 < p < 1/2 and a vesicle bilayer if 1/2 < p < 1. On the other hand, f, a parameter inversely related to p, is used to more accurately determine block copolymer assembly in the water phase(75). Spherical micelles are formed if f > 50%, wormlike micelles if 40% < f < 50% and unilamellar

13 polymer vesicles if 25% < f < 40%. The physical parameters that control the type of structure produced are the polymer block lengths, the concentrations of ions and surfactants, the choice of solvent and water content and temperature(73, 76-78).

In a ‗hybrid‘ synthesis using silica stabilized detergent micelles, Bakalova et al. reported an amino-functionalized silica shell encapsulating single quantum dot micelles(79), providing extra stability to the constituent particle and hence reduced toxicity. It is possible to extend this technique to introduce additional particles such as MNPs into silica stabilized micelles. Polymeric or silica matrices are also popularly used to embed

NPs along with targeting ligands or drug molecules. These composites resemble ‗plum puddings‘ or ‗raisin buns‘ under TEM(12, 18). It is suspected that thermal or pH related decomposition of the polymeric or silica substrate has as bigger a role to play in the formation and stability of these particles(12).

2.1.4 Tissue Targeting and Viability

A critical step in nanoparticle-based studies is an efficient delivery of the particles to the site of interest. This requires appropriate biofunctionalization of the particle surface and delivery techniques that are best suited for the tissue type and the specific target site which could be a membrane receptor, a filamentous protein or a cell organelle. For example in neural-imaging applications, the blood brain barrier (BBB) prevents the passage of large molecules from blood vessels into the cerebrospinal fluid. Therefore nano-composites have to be able to bypass or pass through the tight junctions between the

14 endothelial cells that line the blood vessels and make up the BBB. Most molecules pass through the barrier by passive diffusion through the cell membranes depending on their molecular weight and lipophilicity. One such molecule is the TAT protein which has been extensively studied since it demonstrates an ability to enter cells in a size- independent and receptor-independent fashion. It has a protein transduction domain that is composed of an 11-amino acid sequence containing six arginine and two lysine residues, which confer upon it a high cationic charge that is believed to be responsible for macropinocytosis (80). In one study, TAT conjugated CdS:Mn/ZnS QDs ~3 nm in size were successfully translocated across the blood brain barrier(BBB) and were seen to even penetrate neuronal cell membranes and bind to the cell nucleus(81). Larger particles

~150nm coated with TAT peptide were also successfully transported across the blood brain barrier.(82) TAT peptide (YGRKKRRQRRR), a domain from the transcriptional activator TAT protein from HIV-1, has thus proved to be an effective cell transfecting agent. Once through the BBB or in in vitro studies targeting the axon hillock or the synaptic cleft, functionalization of the probe surface with the appropriate peptide or antibody is required to bind the receptor surface. Biotin ligase can be used to biotinylate a surface protein with a lysine side chain. The biotinylated protein can then be conjugated with a streptavidin bound NP(83). Using functionalizing molecules of minimal sizes reduces cross linking and aggregation. Functionalizing QDs with peptides and antibodies for binding with SK-N-SH neuroblastoma cell membranes (84), it was observed that antibody binding led to aggregation of QDs due to cross-linking, as opposed to the well

15 dispersed QDs on the cell surface observed with peptide conjugation. Without conjugating molecules, non-specific binding and endocytosis was observed(85).

While targeting strategies are still under development, other groups have used NPs for applications that do not require site-specific targeting. For example, in tumor cell imaging, enhanced permeation and retention (or the EPR effect) in the tumor region due to increased vasculature and a change in cell fate that makes tumor cells more prone to rapidly endocytose extracellular particles. Identification and delineation of the tumor boundary are thus made easier since tumor cells appear ‗brighter‘ than surrounding healthy tissue. Another application in which specific targeting is not required is the study of embryogenesis via NPs introduced into a zygote. NP distribution through the developing embryo is tracked to map organogenesis. Slotkin et al. used electroporation in utero and ultrasound in vivo to target QDs to neural stem and progenitor cells (NSPCs)

(86). Koster‘s and co-workers injected QDs into zebrafish blastomeres and used them as lineage tracers through embryogenesis. Similarly, nanoparticles have been used for imaging in circulation. Streptavidin-conjugated QDs were injected into the heart to promote distribution of the QDs through the vasculature and imaged with laser scanning confocal microscopy. Nanoparticles can even be used to track drug delivery. In an in vivo demonstration, Gao et al.introduced wheat germ agglutinin functionalized PEG-PLA coated QDs into a mouse brain via nasal application(87). This was done with the intent of employing the properties of the nasal mucosa that enable drug absorption into the brain, while avoiding spread of the nanoparticles to other organs. The injected NPs were

16 observed to extensively distribute through all the major brain regions in a span of 4hrs after injection.

2.1.5 Toxicity

Some limited studies suggest possible toxicity concerns for iron oxide NPs. Increasing doses of 0.15 to 15 mm Fe2O3 nanoparticles resulted in diminishing viability and capacity of PC12 cells to extend neurites in response growth cues(88). However, there have been a significant amount of studies demonstrating the toxicological properties of magnetic nanoparticles containing gadolinium or iron oxide, and these are generally believed to be biocompatible. In contrast, a major source of toxicity in nanocomposites are the heavy metals, such as cadmium, used in the preparation of QDs. Uncoated CdTe particles introduced in mice brains were seen to induce rapid neuroinflammatory responses, whereas PEG coated CdTe showed no bioreactivity(89), fortifying the concept that PEGylation of the QD surface offers a degree of biocompatibility. Similarly, unmodified CdSe QDs have been known to induce neuron cell death depending on the dose(90). In this study it was noted that the damage was most likely due to disruption of calcium homeostasis. QD treatment also disrupted the functioning of voltage gated sodium ion channels.

2.2. Applications of Fluorescent Magnetic Composites:

2.2.1 Cell separation and Single Molecule Manipulation:

The most successful applications of composites have been in the area of cell separations.

Di Corato et al. reported a nearly 100% efficiency using fluorescent oligothiophene- polymer-coated iron oxide nanoparticles(60). In a novel approach using magnetic particles originating from bacteria conjugated with quantum dots, Maeda et al. reported a

92% recovery rate (51). With the view of applying nanocomposites for single molecule studies and cell manipulation, Ruan et al. have demonstrated control of fluorescent- magnetic micelles on micro-magnetic domains (figure 2.2). A nano-conveyor-belt system was created in which <100nm sized fluorescent-magnetic micelles were moved between magnetic domains created by micron sized magnetic wires and discs using digitally controllable magnetic field gradients applied through external electromagnets(45). This technology has a potential beyond nanoassembly and cell sorting into manipulating sub- cellular organelles within live cells and studying their force-movement relationships. A similar study was performed where magnetic nanoparticle manipulation was demonstrated within live cells on a ‗magnetic-dot‘ substrate and a fixed-strength permanent magnet(91).

Figure 2.2 Magnetic manipulation of fluorescent magnetic micelles

Figure 2.2 Time course from top left panel to bottom right: A fluorescent-magnetic micellar aggregate can be seen to ‗hop‘ along micro-magnetic domains. External magnetic fields were applied in the left-right direction.

2.2.2 Imaging and therapy:

Dextran sulfate coated Si-Mn composites showed near infrared excited two photon emissions and T1 weighted MRI contrast in macrophages. Since macrophages are known to accumulate near arterial plaques these particles have been promoted as a diagnostic tool for artherosclerosis(50). In another dye based composite, folic acid conjugation to the particle surface enabled endocytosis to the target cells through folate receptors(56).

Several groups have exploited the encapsulation paradigm to add drug molecules to the composite. Doxorubicin (DOX), an anti-cancer drug loaded with magnetite nanocrystals in mesoporous silica particles, was delivered to tumor cells in mice(59). The particles 19 retained their fluorescence, enhanced MR contrast in the targeted region, and led to apoptosis of the targeted tumor cells. In a similar study, micelles containing magnetic particles, QDs and DOX were shown to efficiently localize with tumor cells in vitro, in vivo and ex vivo(47). In a thermal ablation study, Xu et al. (92) deployed iron oxide – quantum dot composites to human pancreatic cancer cells. With exposure to a radio frequency (RF) treatment 99.2% of the targeted cancer cells died via apoptosis traceable through the quantum dots complexed with the iron oxide particles. Sun et al. reported

Fe3O4/CdTe magnetic/fluorescent nanocomposites conjugated with anti-CEACAM8 antibody successfully employed for immuno-labeling and fluorescent imaging of HeLa cells(32). Efficient delivery often relies on cell-selective endocytosis, which could be receptor mediated – in which case the composites have to be surface-functionalized with an appropriate ligand, peptide or antibody or rely on enhanced phagocytotic behavior displayed by invasive cancer cells(93). For example, rapid endocytosis and retention was seen in glioblastoma cells with non-biofunctionalized micellar nanocomposites (figure

2.3). Along these lines, in a recent study, it was noted that nanocomposite surface charge was a critical factor in endocytosis, and that NPs were internalized through clathrin- mediated endocytosis and macropinocytosis(94).

Figure 2.3 Live cell imaging with fluorescent magnetic micelles

Figure 2.3 Micelles containing green CdSeZnS QDs and iron oxide nanoparticles in human cultured glioblastoma cells. Large particle aggregates can be seen in a immediately after addition of particles to the cell culture. Endocytosis and retention is seen over several hours in b and c.

2.2.3 Potential applications

Applications where QDs and MNPs have been successfully implemented are also open for potential applications using fluorescent-magnetic composites. To illustrate the extent of applicability, described briefly below are applications of QDs and MNPs with a particular focus on neural system applications. Though every physiological system would require tailor-made nanocomposites, the difficulty of delivering material through the blood brain barrier, the sensitivity of neural tissue to foreign material through immunological responses, and difficulty in clearance make this system an ideal candidate to understand the applicability of nanocomposites.

2.2.3.1 Quantum dots as optical probes:

Recently, QDs have been extensively applied to neural systems. For example, several groups have investigated nanoparticles for studying neurotransmitter receptors. The binding affinity of a toxin to nicotinic receptors in neuromuscular synapses ex vivo was studied using streptavidin coated QDs(95). Also, extracellular domains of AMPA receptors were labeled with antibody tagged quantum dots(96, 97). The QDs could be accommodated in the synaptic cleft and no significant change in target molecule behavior was observed arising from possible steric hinderance stemming from the size of the QD.

Similarly, a method for video tracking of lateral diffusion of membrane molecules such as lipids and proteins has been described (98). GABA receptors were tagged with QDs.

Muscimol, which is a GABA receptor agonist, was bound to PEG-coated CdSe/ZnS core- shell QDS and used to image GABA receptors in Xenopus oocytes by Gussin et al.(99).

Streptavidin tagged QDs were also used to track GABA receptors in growth cones of mouse axons to observe reorganization of the receptors in growth cone dynamics(100).

Alternatively, potassium channel proteins have been labeled and tracked with QDs (101,

102). The diffusion coefficient for QD tagged K+ receptors was lower than that for fluorescent dye bound ones, suggesting that the size of the QD decreased receptor diffusion. This is a potential cause for concern as the bound fluorescent agent appears to alter the behavior of the target molecule.

QDs have also been used to track the dynamic motion of released neural biomolecules.

Cui et al. used quantum dot-labeled NGF to track the movement of NGF in real time in a

22 culture of rat dorsal root ganglion neurons(103) . QD-NGF conjugates were also used by

Echarte et al. (104) to track retrograde transport in PC-12 cells. Vu et al. demonstrated that beta-NGF tagged with QDs could activate TrkA receptors that initiate neuronal differentiation in PC-12 cells(105). Rosenthal et al. showed serotonin-labeled CdSe nanocrystals could inhibit serotonin transport in transfected cells(106) The large size of

QDs, which is usually a limitation, was exploited by Zhang et al. to distinguish between two fusion modes of synaptic vesicles(107). The 15nm size of the QDs they used was small enough to fit in to a synaptic vesicle but too large to escape through transient ―Kiss and Run‖ fusion pores that are 1- 5nm in size. The vesicles that were loaded with the

QDs showed no significant change in morphology or function, though in principle, the volume occupied by a QD would be 1/4th of the vesicle volume.

QDs have even been used for opto-electrical sensing. Kotov‘s group demonstrated that neurons could be photoelectricaly stimulated using QDs (108). Nadeau et al. used fluorescence energy transfer between voltage sensitive dyes and QDs to create sensors for recording glutamate induced action potentials(109). Nadeau‘s group has gone on to demonstrate that conjugation of CdSe/ZnS QDs to dopamine induced quenching of luminescence, faster photobleaching and change in the rate of blinking (110). This study paves the way for creating hypersensitive biochemical sensors. QD-dopamine conjugates have been shown to be sensitive to oxidation with higher luminosity seen near sites of increasing oxidation(111). These conjugates could thus be used as free radical sensors and can be coupled with antioxidants to reduce phototoxicity.

2.2.3.2 Superparamagnetic nanoparticles as probes and MRI contrast agents:

Since 1993, Ingber et al, using magnetic twisting cytometry(112), magnetic needles(113) and magnetic chips(114) have demonstrated magnetic actuation of receptors and performed cell adhesion studies. In a landmark, recent study demonstrating the ability of magnetic particles to manipulate mechanosensitive channels, Hughes et al. showed magnetic activation of TREK-1 channels in COS-7 cells(115). Based on similar principles, magnetofection kits are commercially available for magnetic field driven gene transfections. Hence, much interest has been shown in combining fluorescence with these particles for optical monitoring in these applications.

On the other hand, MRI applications for magnetic particles have been well proven and are employed regularly in clinical settings. In exploratory studies, iron oxide nanoparticles have been used to study migration of transplanted human neural stem cell in rodent brains (116). In another study with fetal human neural precursors, paramagnetic signals were detected with MRI up to 3 months after surgery and were seen not to cause any changes in cell viability or proliferation (117). For comprehensive reviews discussing

MRI applications, targeted hyperthermia, magnetofection, drug delivery and mechanical manipulation in the reader is referred to (14),(5),(10) and (4).

2.4 Conclusion

Considerable interest has been shown in the development of multimodal imaging agents, with a wide assortment of strategies and materials being employed (Table 2.1). As with any carrier or compound used for in vivo assays, specificity and toxicity remain the biggest hurdles to be cleared before these particles can be put to commercial use.

Additional multimodality can be offered within the fluorescent-magnetic composites in the form of multiple functions of magnetic particles as MR contrast agents and hyperthermia inducers, along with fluorophores that can report changes in local environmental variables through sensitivity to ionic concentrations, temperature, pH, etc.

Hence the outlook for future developments in fluorescent magnetic nanocomposites retains numerous possibilities in their function as both sensors and actuators.

Chapter 3: pH sensitive CdS–iron oxide fluorescent–magnetic nanocomposites

Abstract

There has been great interest in the use of nanoparticles for imaging, particularly in multimodal applications (e.g., combination of MRI and fluorescence). Yet creating particles with multiple functionalities has been challenging. Here, we report the synthesis of pH sensitive, fluorescent–magnetic, nanocomposites created through a simple aqueous procedure. Separately synthesized superparamagnetic iron oxide nanoparticles and mercaptopropionic acid (MPA)-coated CdS quantum dots were crosslinked using 3- mercaptopropyl trimethoxysilane (MPS) as a bifunctional linker to yield CdS–iron oxide conjugates. Conjugates formed clusters of 0.1–1.0 μm diameter, with the smallest observed particle diameter ∼50 nm. Particle solubility and photoluminescent (PL) intensity were sensitive to solution pH, with the highest PL intensity and stability obtained at pH values <3.0 and MPS:Cd:Fe ratios of 1:10:1. pH sensitivity is believed to result from changes in nanoparticle solubility within the silica-based matrix. Given these unique properties, this material might find application in separation, pH sensitive detection (e.g., endosomal tracking) and biosensing.

*Reproduced with permission from:

Nanotechnology 20 (2009) 485601 (9pp) doi:10.1088/0957-4484/20/48/485601 doi:10.1088/0957-4484/20/48/485601

3.1. Introduction

Nanoparticles have had tremendous influence in the field of biomedical imaging.

Particularly, fluorescent nanoparticles prepared from type II–VI (e.g. CdS, CdSe, ZnO,

ZnSe), type III–V (InGaAs/GaAs), and type IV–VI (PbSe, PbS) semiconductor materials are popular because of their unique optical properties. Their high photostability, narrow emission bandwidths, and broad excitation spectra make these semiconductor quantum dots (QDs) an attractive alternative to traditional fluorescent organic dyes(118, 119). QDs have been used for in vivo imaging, high throughput screening, fluorescent cell separation, and single molecule tracking(2, 119-121). Similarly, superparamagnetic iron oxide nanoparticles (e.g. Fe3O4, Fe2O3, FePt) have been examined as contrast agents for clinical magnetic resonance imaging (MRI) since the early 1980s(7). Iron oxide particles and iron composites reduce T2 relaxation time in tissues, functioning as positive contrast agents(10). Recently, there has been a significant amount of interest in combining nanoparticle functionalities, particularly in the biomedical imaging field(9, 10), to yield multifunctional composites. Nanoparticles with both fluorescent and magnetic properties could find application in fields ranging from in vivo imaging to single molecule tracking(72, 122-125). For example, iron oxide–QD composites could be used to produce noninvasive MR images, with the fluorescent modality used for intrasurgical imaging or post-surgery pathology studies. At the single cell level, there is a great deal of excitement about the possibility of combined magnetic force application and fluorescent tracking, which could be used for improved cell separation and studies of mechanotransduction

(115). Numerous strategies have been employed to create composites combining these

27 two functionalities. Primary examples of these strategies include micellar encapsulation

(17) and encapsulation in silica spheres (19, 20, 126). The micelle encapsulation strategy is a microemulsion process that introduces a protective shell around the particles while simultaneously creating a hydrophilic nanocomposite. It can be used to encapsulate a variety of structures: covalently linked particles, core–shell structures (where the QD- core is surrounded by a layer of magnetic material (127) or vice versa (128)) and for assembling independently synthesized quantum dots and iron oxide nanoparticles.

Particles may be coated with ligands such as trioctylphosphine (TOPO) or oleylamine in a hydrophobic core to enable encapsulation by a layer of amphiphilic polymers. The polymer layer protects the core and renders the micelles water-soluble. The silica encapsulation method involves either layer-by-layer assembly (129) or co-encapsulation through microemulsion(126). The basic premise of both micelle and silica encapsulation is to synthesize particles using an organic route and then to co-encapsulate particles in another substrate. Water solubility is introduced through the encapsulating substrate or modification of its surface. Here, we demonstrate a new synthetic route, using an entirely aqueous synthesis, to produce fluorescent–magnetic composites that exhibit tunable fluorescence and respond to external magnetic fields. Because the synthesis is performed entirely in the aqueous phase, toxic organic solvents typically employed in both QD and iron oxide nanoparticle synthesis are avoided (130, 131). Also, in view of employing these composites in biomedical applications, water-based synthesis was selected as an important initial step in elucidating the dynamic behaviour of the bare particle surface in water and subsequent changes in its properties based on the selection of surface ligands.

Cadmium sulfide (CdS) QDs were synthesized by terminating the particle surface with mercaptopropionic acid (MPA). Iron oxide nanoparticles were synthesized through arrested co-precipitation of iron salts. Nanocomposites were created using 3- mercaptopropyl trimethoxysilane (MPS), a bifunctional molecule, to crosslink independently synthesized QDs and iron oxide nanoparticles. The thiol (–SH) terminus of

MPS probably attaches to surface Cd2+ ions in CdS nanoparticles (132), whereas the siloxane (Si(OCH3)3) terminus binds to the iron oxide nanoparticle surface (133). The silane termini also stabilize nanocomposites by preventing corrosion of the nanoparticle surface commonly arising from interactions with free ions in solution and direct photooxidation (118, 134). Surface protection is a critical feature as corrosion can decrease fluorescence quantum yield. Interestingly, MPS fluorescent–magnetic conjugates were found to exhibit pH sensitive stability and fluorescence. Stable and highly fluorescent composites were formed only in a very specific pH regime. The presence of both surface ligands—MPA and MPS—was critical in the development of these structures and their optical properties. This work emphasizes the importance of surface functionalization in nanoparticle design. The pH sensitive behaviour coupled with the magnetic properties of the particles presented here could be used for applications in biosensing (pH detection), imaging, biomechanics, and cell separation.

3.2. Materials and methods

3.2.1. Chemicals and instrumentation

Cadmium chloride (CdCl2), sodium sulfide (Na2S), iron (III) chloride hexahydrate

(FeCl3·6H2O), and iron (II) chloride tetrahydrate (FeCl2·4H2O) were obtained from

Sigma Aldrich (124). Mercaptopropionic acid (MPA) and 3-mercaptopropyl trimethoxysilane (MPS), were obtained from Fluka (124). Double de-ionized distilled water (18M, Millipore) was used for all syntheses and was sparged with Ar gas for 30 min before preparing each solution. All syntheses were carried out under argon. An

Accumet AB15 pH meter with a flushable junction electrode was used for pH measurements.

3.2.2. Synthesis of iron oxide nanoparticles

Iron oxide nanoparticles were synthesized as described previously [25] through co- precipitation of iron compounds by reduction in base. FeCl2·4H2O and FeCl3·6H2O were mixed in 15 ml DI water in the molar ratio 1:2. The solution was heated to 70 ◦C under argon, at which time 0.6 ml of 50% NaOH solution were injected into the flask. The solution was then allowed to cool to room temperature. The resultant iron oxide nanoparticles were rinsed through repeated centrifugation in DI water that was sparged with Ar for 1 h before use. Aggregation was prevented by sonication for 2 h, followed by re-suspension in pH 7.0 DI water.

3.2.3. Synthesis of CdS nanoparticles

CdS QDs were synthesized following the procedure of Winter et al (85) using MPA:Cd:S ratios fixed at 22:1:0.4. Briefly, 400 μl of MPA was added to 40 ml of 5 mM CdCl2. The solution pH was raised to 6.5 by dropwise addition of 1 M NaOH. Then, 40 ml of 2 mM

Na2S solution was added. The final solution was stabilized at pH = 7.0 by adjusting pH through dropwise addition of 1 M HCl or 1 M NaOH as required. The solution was permitted to react under Ar with continuous stirring for 30 min.

3.2.4. Synthesis of nanoparticle composites

Composites were synthesized through crosslinking of constituent particles using a bifunctional linker molecule. Iron oxide nanoparticles were added to the pH 7.0 final CdS reaction mixture in the ratio of Cd:Fe = 10:1. Next, MPS was added at a ratio of MPS:Fe

= 1:1. This solution was stirred continuously for 6–8 h. Then, pH was reduced to pH 2.0 by dropwise addition of 1 M HCl. Aliquots of 3 ml were taken at various times throughout the synthesis for analysis.

3.2.5. Nanocomposite characterization

Constituent particles and their composites were characterized using electron microscopy

(EM), UV–visible absorbance spectroscopy, fluorospectrophotometry, x-ray photoelectron spectroscopy (XPS) and superconducting quantum interference device

(SQUID) magnetic measurements. Scanning transmission electron microscopy (STEM), transmission electron microscopy (TEM) and energy dispersive x-ray spectroscopy

(EDS) measurements were collected with a Tecnai F20 instrument using holey carbon laced copper 400 mesh EM grids for the composites and CdS nanoparticles and carbon film coated nickel grids (200 mesh) for iron oxide nanoparticles. Absorption spectra were recorded using a Genesys spectrophotometer (Thermo Electron Corp.).

Photoluminescence measurements were obtained at room temperature using standard quartz cuvettes on a Photon Technology International PTI-810 fluorometer (λex = 365 nm, lamp power = 72 W, detector voltage = 1110 V). Quantum yield measurements, obtained using quinine sulfate as a standard (λex = 365 nm, QY = 0.54), showed that the quantum yield for CdS QDs at pH 7.0 was ∼0.2 or 20%.

3.3 Results

3.3.1. Nanocomposite morphology

Nanocomposites were produced using individually synthesized iron oxide and CdS QD nanoparticles produced through aqueous arrested precipitation routes. Iron oxide nanoparticles were ∼10 nm in diameter (figure 3.1(a)) whereas CdS QDs exhibited diameters of 4–5 nm (figure 3.1(b)). Composites, however, exhibited much larger dimensions. The size distribution, collected from STEM, shows that 68% of the composites are less than 200 nm in diameter, 78% are less than 300 nm and 90% are less than 500 nm. Both large (figures 3.2(a) and (b)) and small (figures 3.2(c) and (d)) particles contain iron oxide and CdS nanoparticles as evidenced by EDS scans.

Figure 3.1 (a), (b) TEM images of Iron Oxide and CdS nanoparticles

Figure 3.1. (a) TEM of iron oxide particles showing a particle diameter ∼10 nm. XPS for bare iron oxide particles indicating that the crystals are Fe2O3–Fe3O4. (b) MPA coated

CdS nanoparticles. (c) Please see below. STEM images (first row, first two panels from left) showed a wide particle size distribution 100 nm–1 μm. TEM images (starting from first row, third panel from the left) show particle cores embedded in a silica-based matrix.

Each following snapshot in a sequence, shown from left to right and top to bottom, was taken after an interval of 30 s. It can be seen that the composite ‗melts‘ under the focused electron beam—MPS and MPA polymerize through increased crosslinking to form a denser mesh and the particles coalesce to form a smaller composite.

Figure 3.1 (c) STEM and TEM of composites

Figure 3.2. STEM and EDS of CdS–iron oxide clusters

Figure 3.2. STEM and EDS of CdS–iron oxide clusters where particle size is ((a), (b)) ∼1

μm, ((c), (d)) ∼100 nm. Red squares in (a) and (c) show areas scanned in each sample.

Cd:Fe ratios in both samples were similar, and independent of size.

Figure 3.3 Absorbance Spectra

Figure 3.3. Absorbance spectra for (a) MPA-capped CdS QDs, (b) MPS-conjugated;

MPA-capped CdS QDs and (c) CdS–iron oxide composites. Numbers near curves indicate pH of the solution. In ((b), (c)) absorbance curve of MPA-capped QDs without

MPS (CdS@7) is included for comparison. (d) Wavelengths at peak absorbance versus pH.

Figure 3.4. Fluorescence from MPA-capped CdS QDs and CdS–iron oxide composites

Figure 3.4. (a) From left to right: MPA-capped CdS QDs and CdS–iron oxide composites under UV excitation (365 nm). Numbers indicate pH of the solution. (b) MPS- conjugated, MPA-capped CdS QDs.

3.3.2. Nanocomposite absorbance

It has been well established that the constituent fluorescent nanoparticles, CdS quantum dots, exhibit pH sensitive growth in aqueous solution (85). This was also observed in our work, even when pH changes were made as much as 30 min after particle synthesis

(figure 3.3(a)). However, there are significant differences in pH response for iron oxide–

CdS composites (MPS: MPA = 1:220) (figure 3.3(c)). Changes in peak intensity with pH are similar to those for CdS quantum dots, but shifts in the peak absorbance wavelength were far less pronounced for composites (figure 3.3(d)) and composite peaks also exhibited significant broadening. The resistance of the peak absorbance wavelength to pH induced shifts may indicate greater stability of the capping ligand coating, providing composites with better protected surfaces. These results were similar to those obtained for MPS-conjugated CdS nanoparticles in the absence of iron oxide (MPS: MPA = 1:220)

(figure 3.3(b)), suggesting that the observed behaviour is a direct result of MPS conjugation and not iron oxide nanoparticles.

Figure 3.5. PL spectra for CdS–iron oxide composites

Figure 3.5. PL spectra for CdS–iron oxide composites, (a) as pH declines and (b) as pH is restored to 7.0. PL spectra for MPS-CdS QDs, (c) as pH declines and (d) as pH is restored. Numbers near curves indicate pH. Inset shows change in PL intensity with standard error averaged over 3 samples. PL for MPA-capped QDs (CdS@7) is included in ((a), (b)) for comparison. (e) PL intensity variation as pH is recycled between 2.0 and

7.0.

3.3.3. Nanocomposite photoluminescence

Iron oxide–CdS composites also exhibited pH dependent photoluminescence. For intermediate pH values of 5–7, virtually no fluorescence was seen (figures 3.4(a) and

5(a)). However, as pH was decreased from 5 to 1.5, fluorescence intensity increased in a linear fashion with a maximum at the lowest pH investigated (pH 1.5) (figure 3.5(a) inset). As anticipated for quantum dots prepared through an aqueous route, the peak bandwidth is broad, indicating trapped state emission(135). The location of the PL peak was difficult to discern at low pH, but appeared to be ∼475 nm for pH values 4–7, shifting to ∼525 nm at pH < 3. This behavior was mostly reversible (figure 3.5(b)).

However, as pH was increased from a minimum value of 1.5, the PL peak remained at

∼525 nm and weak fluorescence was observed at pH 5–7. Thus, upon restoration, increasing pH resulted in a near linear decline of PL intensity throughout the pH range investigated, with a maximum PL intensity at pH 1.5. It should be noted that these responses were not dependent on the duration of stirring and heating (up to 16 h) during

39 synthesis and that repeated centrifugation and re-dissolution of the particles to eliminate possible interactions with dissolved species did not alter the results, indicating the observed response likely results directly from the particle composition. For comparison,

PL spectra of MPA-capped, MPS conjugated CdS quantum dots were also evaluated.

When MPA alone was used as the capping ligand, CdS QDs exhibited high PL intensity at pH 7, which rapidly declined with reduction to pH 6 (figure 3.4(a) and supporting information available at stacks.iop.org/Nano/20/485601). As pH was further decreased from 6 to 4, PL intensity increased, reaching a maximum at pH 4. As pH declined from 4 to 1.5, PL intensity decreased with virtually no fluorescence observed below pH 2. This behaviour was reversible but PL intensity was significantly diminished (increasing peak maximum ∼20% of declining peak maximum) and very little fluorescence was observed at pH 7.0. MPS-conjugated, MPA-capped CdS quantum dots (MPS: MPA = 1:220), however, displayed profiles similar to those observed in CdS–iron oxide composites

(figures 3.4(b), 3.5(c) and (d)). This suggests that MPS, and not iron oxide, is primarily responsible for the observed pH responsive PL behaviour.

Figure 3.6 Measurement of magnetic properties

Figure 3.6. SQUID measurements ((a), (c)) iron oxide nanoparticles before conjugation and ((b), (d)) CdS–iron oxide composites.

3.3.4. Nanocomposite magnetic properties

Constituent iron oxide nanoparticles (figures 3.6(a) and (c)) were nearly superparamagnetic at room temperature, with a remanence of 1.26 emu g−1 and a saturation magnetization of ∼55 emu g−1. Both parameters are consistent with reported 41 values(133, 136). CdS–iron oxide composites (figures 3.6(b) and (d)) had near 0 remanence and a saturation magnetization of ∼9.5 emu g−1, which was significantly reduced from that of unconjugated particles. Magnetization (M) also declined with increasing field strength (H) indicating diamagnetic behaviour stemming from the silica- based matrix surrounding the nanocomposites. This is consistent with previous studies of silica conjugated nanoparticles (137). Response of composites to a magnetic field was also pH dependent. At neutral pH, iron oxide nanoparticles could be separated from composite solutions observed immediately after synthesis when a magnetic field was applied (figure 3.7(a)). These particles precipitated from solution when the magnet was removed. Even after several hours of stirring and heating, composites were not formed. In older samples (2–7 days) at neutral pH, particle flocculation was observed and high PL and magnetic sensitivity were lost (figure 3.7(b)). Flocculated particles precipitated even in the presence of a magnet. In contrast, composites formed at low pH, and both fluorescence and magnetic saturation were stable in low pH environments (figure 3.7(c)).

3.4. Discussion

Here, we report a method for creating aqueous, magnetic-fluorescent particles with pH sensitive properties, including pH sensitive formation, reduced sensitivity to absorbance shifts in response to pH change, and pH sensitive PL. The composites consist of CdS quantum dots conjugated to iron oxide nanoparticles through a bifunctional crosslinker

(MPS) containing both thiol (–SH) and siloxane (Si(OCH3)3) termini. Despite the fact that only a small amount of MPS is used (1:220 MPS:MPA molar ratio), the morphology

42 of the composites (figure 3.1(c)) indicates individual quantum dots and iron oxide nanoparticles are embedded in a silica-based matrix. EDS results confirm this observation

(figure 3.2). Similar structures were observed by Kotov‘s group for MPS–CdTe and

MPS–CdSe composites(18) in what was dubbed a ‗raisin-bun‘-type composite; however, pH sensitive behaviour was not reported. Excess MPS (MPS:Cd ratio 1:1) was essential in the formation of these complexes. MPS has also been used to create thick, porous silica shells that controlled oxidation reactions limiting introduction of defects to the QD surface (138). The observed absorbance behaviour suggests that MPS–MPA–CdS and iron oxide–MPS–MPA–CdS particles are more stable to pH changes than their MPA–

CdS counterparts. Specifically, MPA–CdS exhibit absorbance peaks that are correlated to pH. Given that QD absorbance is a function of particle size(120), absorbance peak wavelength shifts are likely caused by surface erosion of smaller particles (blue-shift, pH values 7–4). The loss of surface ligands from the CdS quantum dots allows for Ostwald ripening through addition of freely floating CdS to the evacuated surface sites, thereby increasing the size of the particle (red-shift, pH values 3–1.5). These processes are most likely initiated by ligand loss, indicating an unstable particle surface. Similar behaviour has been seen in thiol-capped CdTe crystals (139). This trend is not observed, however, for MPS–MPA–CdS or iron oxide–MPS–MPA–CdS composites, suggesting that MPS in combination with MPA ligands enhance particle stability against pH changes. Iron oxide–

MPS–MPA–CdS composites also exhibited pH sensitive PL with a near linear increase in

PL from pH 7 to pH 1.5 (figures 3.4, 3.5 and 3.7). This is in contrast to MPA–CdS nanoparticles, which also exhibited pH sensitive PL, but with a maximum at intermediate

43 pH (e.g., pH 4–5) and pH declines on either side of this maximum. It is interesting to note that at initial neutral pH values for fresh iron oxide-MPS–MPA–CdS nanocomposites, very little fluorescence was observed and the PL peak was centred near 475 nm, whereas as pH declined the PL peak shifted to ∼525 nm, and remained in this position upon subsequent pH increase (figure 3.5).

Figure 3.7 Fluorescence images of fluorescent magnetic composites

Figure 3.7. (a) Freshly prepared, pH 7 composites observed within minutes of MPS and iron oxide addition. Under UV excitation and in presence of a neodymium magnet, separation between the iron oxide and fluorescent particles is seen (magnetic field strength is 0.5 T on the right and falls off to ∼0.25 T over the width of the glass vial), (b)

CdS–iron oxide composites at pH 7 under UV excitation (365 nm) after particles were

44 allowed to mature for 2 days. High PL and magnetic sensitivity is absent. (c) CdS–iron oxide composites at pH = 2.0 at t = 0, 10 min, and 3 h.

Additionally, as pH increased, PL, which was previously virtually absent, was observed at neutral pH. This suggests that the particles undergo a ‗maturation process‘, which may be related to bond forming between the MPS and MPA ligands. MPA-only particles exhibited a similar response, but in different pH ranges. At an initial pH of 7 high PL intensity was exhibited, but this declined dramatically when pH was lowered to 6 and remained much lower when pH returned to 7. We believe that pH sensitive PL behaviour and the PL ‗maturation‘ response may result from alterations in surface passivation as a result of MPS-MPA interactions. Increased passivation reduces surface defects, which can serve as centres for non-radiative decay of the excited state, enhancing fluorescence intensity. Increased surface passivation also enhances particle solubility, and we have observed that a decline in PL intensity is invariably coupled to a loss of solubility. This implies a mechanism whereby surface ligands detach in a pH sensitive manner, exposing the nanoparticle surface and permitting non-radiative recombination and particle aggregation, which decreases fluorescence intensity. Magnetic properties were also altered in CdS–iron oxide composites. Unconjugated particles exhibit typical ferromagnetic to near superparamagnetic behaviour (133). However, iron oxide–MPS–

MPA–CdS conjugates exhibit diamagnetic behaviour in high magnetic fields. This can most likely be attributed to the presence of the silica-based coating (137). SQUID results

(figure 3.6) indicate that the silica-based matrix creates an opposing magnetic field,

45 which effectively screens the field applied to the iron oxide particles, producing a decrease in magnetism following initial saturation. Nonetheless, particles clearly respond to a magnetic field (figure 3.7(c)) supplied by handheld magnets. Thus, the unique, pH sensitive absorbance, PL, solubility, and magnetism observed in these composites all appear to result from interaction between the MPA and MPS capping ligands, which we believe form a crosslinked protective shell around the constituent particles. This hypothesis is supported by several factors. First, the magnetic data clearly indicate the presence of a diamagnetic screening material, and given the particle composition, this is most likely a silica-based matrix. Second, the composite peak absorbance wavelength is more stable against pH change, an indicator of increased surface passivation. Stability is enhanced even at low pH, where protonation of both thiolate ligands, dissociation from the CdS surface, and particle precipitation would normally be expected (140). Indeed,

MPA–CdS particles do not exhibit stability at low pH, suggesting interaction between the two ligands, possibly via crosslinking, results in this enhanced stability. Third, we have observed that particles are not formed when MPS alone is used as a capping ligand, or if

MPS is added before MPA (data not shown). Additionally, although CdS nanoparticles can be created using only MPA as a ligand, their absorbance, PL, and solubility are distinct from that of MPS-MPA–CdS and iron oxide–MPS–MPA–CdS nanoparticles.

Thus, both ligands are necessary to produce stable composites with the reported behaviour, and the order in which the ligands are added is critical. Given that MPS is added in a 1:220 ratio with MPA, this is surprising, and points to a possible interaction between the two ligands. One possibility is that enhanced stability is derived from

46 crosslinking of the siloxane termini at highly acidic (low) pH to form a silica-based matrix that protects MPA molecules from protonation and prevents their detachment from the QD surface. Iron oxide nanoparticles are also embedded within the matrix, and possibly conjugated to the siloxane termini. Such a matrix, while containing only a small amount of MPS, would stabilize MPA ligands on the nanoparticle surface, reducing their propensity to undergo Ostwald ripening and experience absorbance wavelength shifts and also enhancing PL intensity. The observed ‗maturation response‘ in PL can then be explained by the initial formation of these crosslinks, which stabilize particles against further change. The ability of composite PL to respond to pH changes indicates that these matrix bonds are electrostatic in nature, rather than covalent, and can be reversibly formed and broken, leading to the observed reversible pH sensitive response. The ability of silica-based coatings to modify nanoparticle surface properties, including solubility and photoluminescence, has been previously reported for other systems (132). For example, similar pH dependent effects on PL have been observed for MPS-capped CdS

QDs synthesized in tetrahydrofuran and methanol(134), and pH dependent stability changes have been observed for core shell CdSe/ZnS particles (141) and Fe3O4/CdTe particles encapsulated in silica (142). In the latter system, the observed increases in solubility and stability at low pH were correlated with high zeta potentials, the charge on the particle surface, leading to mutual repulsion and hence better particle dispersion.

These results confirm that quantum dot photoluminescence is directly correlated with solubility and surface passivation. Preventing degradation of the protective ligand matrix will be key to stabilizing future composites synthesized through an aqueous route. Here,

47 we show that using optimized combinations of surface modifying ligands can enhance nanocomposite stability, and can be used to imbue particles with interesting properties

(e.g., pH sensitivity). Given their fluorescent and magnetic properties, these particles could be applied in a number of applications (e.g., magnetic cell separation, multimodal imaging, and magnetic manipulation). It should be noted that the large size distribution of the composites may result in light scattering that reduces the observed fluorescence intensity, and limits biological application. This can; however, be addressed by size selective isolation of the smallest nanocomposites. Standard filtration techniques such as column and membrane filtration can be used to isolate the small composites necessary for biological applications. Another approach would be to use focused electron beams to polymerize MPS and MPA to form denser composites. This process can be seen in the sequence of TEM images above (figure 3.1(c)). This suggests that other focused energy sources such as ion beams or lasers could also be employed to induce polymerization in the composites and reduce particle size. This will be a focus of future work. In addition, the dependence of fluorescent properties on pH provides an extra modality that could be exploited for sensing the local environment, for example as a pH sensor, with subsequent magnetic separation. pH sensitive behavior could be particular useful for local sensing in a biological environment. For example, particles could be used to track the motion of endosomes, with fluorescence increasing as pH decreases and vesicles convert to lysosomes.

3.5 Supporting information

Figure 3.8. XPS for bare iron oxide particles indicating that the crystals are Fe2O3-Fe3O4

Figure 3.9. Size distribution histogram from STEM images of CdS-Iron oxide composites

Figure 3.10 PL spectra for CdS QDs with varying pH

Figure 3.10 (a) PL spectra for CdS QDs while reducing pH from 7 to 1.5, (b) PL spectra for CdS QDs while restoring pH to 7.0 from 1.5

Chapter 4: Fluorescent carbon particles for live cell imaging

4.1 Introduction

An emerging class of nanostructures includes small molecular weight, fluorescing, carbon nanostructures. Typically, structures such as fullerenes, carbon-onions, graphene, graphitic and diamond crystals, non-crystalline aggregates, fibers and nanotubes are obtained as products of both natural and artificial processes. For instance, during slow and rapid petroleum combustion, detonations, chemical vapor deposition and laser ablation of carbonaceous material(143, 144) these particles are produced in copious amounts. They have traditionally been regarded as wasteful byproducts; however, with appropriate processing and surface modifications to imbue them with bio-applicable modalities, they are now being considered as stable and potentially less toxic alternatives to heavy metal containing nanoparticles. Fluorescence in these structures arises when a carbon atom is substituted by a nitrogen atom with an adjacent vacant site – creating a nitrogen vacancy or N-V center (figure 4.1). There are several types of such configurations or ‗defects‘ that give rise to fluorescence properties. Similar crystallographic defects also give diamonds their distinctive colors. These centers can be very stable and, unlike organic dyes, do not undergo photobleaching under sustained excitation or stochastic blinking to which quantum dots are prone(145). Combined with well defined surface passivation and coating techniques using silica, polyethylene glycol,

51 polylactic-glycolic acid or other polymeric compounds, these particles can be easily adapted for biomedical applications.

vacancy

nitrogen

Figure 4.1 Representation of an N-V center

Figure 4.1. Energy minimized structure after a nitrogen atom and a vacancy are introduced to interstitial sites in a simplified array of carbon atoms (denoted by grey spheres). Red spheres indicate electrons (created in NanoEngineer-1, Nanorex Inc).

Several configurations of such defects are known to exist with combinations of predominantly nitrogen, oxygen and vacancy centers.

Several of these structures have been adapted for biological imaging, for example: nanodiamonds(146-150), graphitic crystals(151, 152), graphene-based particles(153, 154) and surface-modified, fluorescent carbon nanotubes(155-158). Adapting a process for producing fluorescent particles from candle soot described by Liu et al(159), several 52 groups have attempted modification of the particles and their surface chemistry to modulate the fluorescence (160-162). Here, synthesis of biocompatible polylactic- glycolic acid (PLGA) encapsulated carbon nanoparticles (CNPs) is described. The choice of PLGA as an encapsulating agent originates from its well characterized and extensive history of applications in drug delivery. It has been used as a nano-carrier to encapsulate

DNA, proteins, peptides, small molecular weight compounds and nanoparticles(163,

164). The use of PLGA here would enable later addition of other compounds (e.g., MRI contrast agents, receptor targeting ligands, or therapeutic drugs).

4.2 Methods:

4.2.1 Carbon nanoparticle synthesis:

To obtain carbon black, paraffin candles were burnt under a mechanically rotated aluminum drum. Deposited soot was periodically collected by scraping the drum surface gently with a clean razor blade. 1g of soot was added to 200mL of 5M nitric acid solution in a three-necked flask and refluxed at 90 deg C under constant argon pressure for at least

24 hours. Longer reflux times gave higher yields. After reflux, the solution was allowed to cool to room temperature before centrifugation at 12000RPM for 90 minutes. The supernatant contained the fluorescent particles; non-oxidized carbon particles precipitated forming the pellet. The acidic solution was neutralized with base (NaOH), or through dialysis.

4.2.2 PLGA nanosphere synthesis:

Polyvinyl alcohol (PVA, Sigma Aldrich) was added to distilled water (5-10% wt) to make 100 mL PVA solution. 0.2 g of PLGA was added to dichloromethane in a separate container to make a total volume of 1.5mL, to which 200µL of CNP solution and

0.075mL of SPAN 85 (Sorbitane trioleate, Sigma Aldrich) were added, followed by probe-sonication with a Branson Digital Sonifier S-450D for 10 minutes (duty cycle

50%). Contents of the vial were then added to 10mL of PVA solution followed by probe- sonication for 10 minutes (duty cycle 30%). This solution was added to the remaining

PVA solution and stirred at room temperature for 1 hour, before centrifugation at

4000RPM for 10 minutes. Supernatant was discarded and the collected particles were frozen at -20 deg C after which they were dried in a vacuum oven overnight at room temperature.

4.2.3 Characterization:

Particles were characterized using electron microscopy (EM), UV–visible absorbance spectroscopy, and fluorospectrophotometry. Transmission electron microscopy (TEM) images were collected with a Tecnai F20 instrument using carbon and silica based grids.

Scanning electron microscopy (SEM) images were taken on a Quanta200. Absorption spectra were recorded using a Genesys spectrophotometer (Thermo Electron Corp.).

Photoluminescence measurements were obtained at room temperature using standard quartz cuvettes on a Photon Technology International PTI-810 fluorometer (lamp power

= 72 W, detector voltage = 1110 V). CNPs encapsulated in PLGA were introduced to

54 human glioblastoma multiforme cells (OSU-GB-1,2). Images of PLGA nanospheres were taken using an Olympus BX-41 optical fluorescence microscope. Cell culture images were taken using a Zeiss 510 Meta Confocal Microscope.

4.3 Results:

Figure 4.2 Fluorescence from CNP solution

Figure 4.2. Vials containing carbon nanoparticle solution (left) and cadmium sulfide quantum dot (aqueous) solution (right) for comparison. λex = 365nm.

The originally hydrophobic carbon particles from soot are converted to hydrophilic particles through refluxing in nitric acid. Throughout the synthesis process, the insoluble soot is gradually observed to become soluble, converting the reaction solution to a dense, 55 black solution. Upon centrifugation a yellow-brown supernatant can be observed. These are the small, fluorescent particles separated from the bulk of black solute, which consists of large particles that settle to the bottom of the centrifuge tube. Though almost all of the fluorescent particles are present in the supernatant collected in the tube after the refluxed solution is centrifuged, a small quantity of fluorescent particles can be recovered from the non-fluorescing, black precipitate with repeated dissolution in DI water and centrifugation. The particles used for analysis here were from the supernatant. It was seen that reflux durations beyond 12 hours did not have any effect on yield or the photoluminescence properties of the particles. 10-14 hours of reflux time was deemed a sufficient duration-range for harvesting the product. Similar results were noted by Ray et al.(160).

Following synthesis, fluorescence was first confirmed with a handheld UV lamp (λex =

365 nm) (figure 4.2). Particle fluorescence was stable over several months in aqueous solution. Subsequent photoluminescence scans showed that emission intensity depends on excitation wavelengths with maximum emission at an excitation wavelength of

~465nm (figure 4.3a, b). Characterizations of particles obtained by Mao et al. (162) and others (160) confirm these observations. Quantum yield (QY) for bare carbon particles calculated using fluorescein as a standard (QY 95%) was 0.7 %, similar to previously obtained values (159). Absorbance spectra are broad and continuous (figure 4.3c).

Organic dye absorbance spectra are typically narrow and quantum dot spectra have discrete peaks in a generally continuous spectrum, suggesting that these particles may

56 resemble quantum dots in size and structure, but may possess multiple fluorescent units or color centers aggregated in a particulate fashion. Such multiple species of color centers with different configurations of nitrogen, oxygen and vacancy defects may explain the multiple colors seen in the CNP solution after gel electrophoresis by Liu et al(159).

Additionally, the photoluminescence peak dependence may also be explained by the existence of discrete particles - each possessing different color centers existing in the bulk solution. The presence of different color centers however, has to correlate with the particle size dependent emission seen after gel separation. Further analysis needs to be performed to determine how size quantization and color center relate to each other and contribute to the mechanism of excitation dependent fluorescence.

Electron microscopy scans of bare CNPs (without PLGA) showed in-distinct and ill- defined structures (figure 4.4a). Predominantly, TEM grids are supported by carbon films, compounding existing difficulties in obtaining high resolution TEMs arising from the small (expected 1-4nm) particle size. Additionally, carbon particle aggregation occurring after deposition of the solution on the TEM grid may also explain the large

(~5nm) sized dark spots seen in figure 4.4a. Similar aggregation behavior has been observed and discussed by Ray et al.(160), who hypothesized that the sample preparation process in which the particle solution is allowed to evaporate may lead to accumulation of the particles. SEM images (figure 4.4b) showed PLGA nanospheres with a range of size distribution. Prior to SEM, the same sample was imaged in an optical fluorescence microscope under UV illumination, showing submicron fluorescent aggregates (figure

4.4c); thus demonstrating successful encapsulation of CNPs in PLGA and visibility through a conventional microscope.

Figure 4. 3 (a), (b) CNP emission and excitation spectra

Figure 4.3 (c) CNP absorbance

Figure 4.3 (a) Photoluminescence scans with varying λex (see legend). (b) Excitation scan

(λem = 525 nm) shows a coarse but definite maximum in the region between λex 450-470 nm. (c) Absorption spectra were wide and without the typical peaks that characterize discrete band gaps such as those seen in semiconductor quantum dots.

Human patient derived glioblastoma multiforme cells showed rapid uptake of particles in vitro (figure 4.5). Localization near the cell nuclei seen in all four panels is possibly due to uptake and accumulation of particles at golgi or the endoplasmic reticulum. This phenomenon is commonly observed with nanoparticle uptake where endocytosed particles are transported to the microtubule organizing center. Though cell viability assays to quantify possible toxicity were not conducted during the duration of this study, cell death in GB cells was not seen unless high particle concentrations (volume ratios 59 greater than 1:1 between particle solution: serum) were used, in which case cell death was seen along with large intra- and extra-cellular particle-aggregates. The contribution of particle toxicity, versus cell death stemming from dilution of serum (from addition of the DI water based CNP solution), remains to be elucidated.

Figure 4.4 Electron and optical microscopy for size and structure determination

Figure 4.4. (a) TEM image of bare CNPs on a carbon substrate. (b) SEM of CNP encapsulating PLGA nanospheres. (c) Fluorescence image of PLGA-CNP nanosphere- aggregates under oil immersion at 100X.

Figure 4.5 Live cell imaging with PLGA-CNPs and bare CNPs

Figure 4.5. Representative fluorescence confocal-bright-field composite images taken with excitation wavelengths between 440 and 480nm, between 1-16 hours after introduction of CNP solution to cell culture. Rapid endocytosis was seen in samples in 61 which PLGA-encapsulated CNPs were used (a), (b) and (c), as well as samples with bare

CNPs (d).

4.4 Silica/carbon nanoparticles:

As an alternative to PLGA, silica encapsulation was also investigated. While attempting silica encapsulation of CNPs and iron oxide nanoparticles, bright blue particles were discovered in solutions containing aminopropyl-trimethoxysilane and carbon nanoparticles. The nature and structure of these particles has yet to be ascertained, but photoluminescence scans produced quantum yield measurements equal to 5%, seven times greater than QY of CNPs. While further analyses needs to be done, it is suspected that these particles may be silica nanostructures nucleated around a CNP core. A similar scheme has been described elsewhere using colloidal SiO2 with surfactants to create

CNP-silica complexes(165). Introduction of these particles into GB cell cultures showed rapid particle uptake with distribution across the entire cell with higher concentrations near the cell nucleus (figure 4.6)

Figure 4.6 Fluorescence and live cell imaging with silica nanoparticles

Figure 4.6 (a) Vials containing green CNPs (left), blue silica/carbon particles (right). (b)

Cell uptake was faster than that with green CNPs (c) particles were seen distributed throughout the cell body with concentrations most likely at the peri-nuclear region.

4.5 Conclusion:

An inexpensive method to produce biocompatible fluorescence markers less than 5nm in size has been demonstrated. The ease of synthesis, compared to high energy explosions or ion beam bombardment, gives this method wide potential to be applicable in laboratory as well as commercial settings. Separation of the nanoparticle species and a comprehensive functional and structural characterization would enable optimization of the preparation method. It is worth noting that both CNPs and silica/carbon NPs were readily taken up by cells without functionalization and potentially, at optimum concentrations, the particles could be observed in cells through a conventional microscope as well as with a naked eye.

Acknowledgements: I would like to thank Sean Hawkins for help with CNP synthesis and purification, Ning Han for help with the PLGA nanosphere synthesis and SEM, Dr.

Atom Sarkar (Neurosurgery, OSUMC) and his team for use of the confocal microscope and the GB cell line.

Chapter 5: Conclusion

Two independent methods to synthesize nanocomposites were developed using widely varying sources, ranging from conventional quantum dot syntheses to combustion byproducts from carbonaceous material. This demonstrates the wide range of possibilities for creating novel biomedical imaging tools through wet-chemical techniques. Within each of the syntheses investigated: water based composite synthesis, micelle assembly and polymer encapsulation of carbon nanoparticles, more avenues remain to be explored for optimization, as well as for new designs branching from existing ones. The pH sensitivity in the synthesis process for CdS-Iron Oxide composites revealed a novel experimental parameter that could lead to other designs with unexplored silane and siloxane compounds. The stable fluorescence from an inexpensive carbon source has provided a potentially non-toxic fluorescing agent that can replace heavy metal based quantum dots. It can be seen that strong collaboration between the traditionally varied fields of polymer science, solid state chemistry, and biochemistry is required in the design of these novel materiel. For example, both theoretical and experimental advances in polymer science are leading to optimal and stable designs of nanocomposites through the identification of the polymer head-groups and chain lengths required for stable micelles, as well as their sensitivity to parameters such as temperature, pH and ionic concentrations.

Every biomedical application involving nanoparticles requires tailor-made particle designs, with site-specific targeting being one of the most significant challenges in the design process. As a result, every available targeting technique, including antibodies, peptides, transfecting agents and physical techniques like electroporation and magnetofection need to taken into careful consideration during conceptualization of the particle design. Though rigorous short-term and long-term toxicological analyses need to be performed for all constituent particles: heavy metals, iron oxide and carbon NPs; passivation techniques using a variety of materials offer promise in offsetting potential toxic effects. Applications ranging from cancer diagnosis and therapeutics to single molecule tracking and force spectroscopy can greatly benefit from these new designs.

Magnetic nanoparticles have already seen commercial application as MRI contrast agents. The addition of other modalities to these particles has the potential of extending their applicability in fundamental research as well as clinical practice.

List of references

1. Alivisatos A. P., Science. 271(5251): p. 933. 1996 2. Bruchez M., et al., Science. 281(5385): p. 2013. 1998 3. McBain S. C., et al., Nanotechnology. 19(40). 2008 4. Sun C., Lee J. S. H.Zhang M. Q., Advanced Drug Delivery Reviews. 60(11): p. 1252. 2008 5. Zborowski M.Chalmers J. J., Magnetic Cell Separation. 2007, New York: Elsevier. 6. Hughes S., El Haj A. J.Dobson J., Medical Engineering & Physics. 27(9): p. 754. 2005 7. McCarthy J. R.Weissleder R., Advanced Drug Delivery Reviews. 60(11): p. 1241. 2008 8. Zimmer C., et al., Radiology. 196(2): p. 521. 1995 9. Pankhurst Q. A., Connolly J., Jones S. K.Dobson J., Journal of Physics D-Applied Physics. 36(13): p. R167. 2003 10. Laurent S., et al., Chemical Reviews. 108(6): p. 2064. 2008 11. Biju V., et al., Biotechnology Advances. 28(2): p. 199. 12. Thakur D., Deng S., Baldet T.Winter J. O., Nanotechnology. 20(48): p. 485601. 2009 13. Smith A. M.Nie S., Accounts of Chemical Research. 43(2): p. 190. 2009 14. Pankhurst Q. A.et al., Journal of Physics D: Applied Physics. 42(22): p. 224001. 2009 15. Minchin R. F.Martin D. J., Endocrinology. 151(2): p. 474. 16. Park J. H., et al., Angewandte Chemie-International Edition. 47(38): p. 7284. 2008 17. Park J. H., et al., Angew Chem Int Ed Engl. 47(38): p. 7284. 2008 18. Rogach A. L., et al., Chemistry of Materials. 12(9): p. 2676. 2000 19. Salgueirino-Maceira V., et al., Advanced Functional Materials. 16(4): p. 509. 2006 20. Selvan S. T., Patra P. K., Ang C. Y.Ying J. Y., Angewandte Chemie-International Edition. 46(14): p. 2448. 2007 21. Reiss P., Protiere M.Li L., Small. 5(2): p. 154. 2009 22. Veiseh O., Gunn J. W.Zhang M. Q., Advanced Drug Delivery Reviews. 62(3): p. 284. 23. Laurent S., Bridot J. L., Elst L. V.Muller R. N., Future Medicinal Chemistry. 2(3): p. 427.

24. Trindade T., O'Brien P.Pickett N. L., Chemistry of Materials. 13(11): p. 3843. 2001 25. Smith A. M., Gao X. H.Nie S. M., Photochemistry and Photobiology. 80(3): p. 377. 2004 26. Xie J., et al., Current Medicinal Chemistry. 16(10): p. 1278. 2009 27. Liu W., Zhong W.Du Y. W., Journal of Nanoscience and Nanotechnology. 8(6): p. 2781. 2008 28. Osaka T., et al., Analytical and Bioanalytical Chemistry. 384(3): p. 593. 2006 29. Giersig M.Hilgendorff M., European Journal of Inorganic Chemistry,(18): p. 3571. 2005 30. Huber D. L., Small. 1(5): p. 482. 2005 31. Ren C., et al., Nanoscale Research Letters. 4(7): p. 640. 2009 32. Sun P., et al., Langmuir. 26(2): p. 1278. 2009 33. Yan H., Zhang J. C., Yu B. W.Shen Y., Acta Materialia. 58(2): p. 726. 34. Wang G., et al., Journal of Fluorescence. 19(6): p. 939 2009 35. Deng S., Ruan G., Han N.Winter J. O., Nanotechnology. 21(14): p. 145605. 36. Tian Z.-Q., et al., Chemical Communications. 2009(27): p. 4025. 2009 37. Gu H., Zheng R., Zhang X.Xu B., Journal of the American Chemical Society. 126(18): p. 5664. 2004 38. Selvan S. T., Patra P. K., Ang C. Y.Ying J. Y., Angewandte Chemie International Edition. 46(14): p. 2448 2007 39. Li L., et al., Chemical Physics Letters. 461(1-3): p. 114. 2008 40. Du G. H., et al., Nanotechnology. 17(12): p. 2850. 2006 41. Kim H., et al., Journal of the American Chemical Society. 127(2): p. 544. 2004 42. Lim Y. T., et al., Journal of the American Chemical Society. 131(47): p. 17145. 2009 43. Shi D., et al., Advanced Materials. 21(21): p. 2170 2009 44. Shi D., et al., Applied Physics Letters. 95(22): p. 223702. 2009 45. Ruan G., et al., Nano Letters. 46. Roullier V., et al., Chemistry of Materials. 20(21): p. 6657. 2008 47. Park J.-H., et al., Angewandte Chemie International Edition. 47(38): p. 7284 2008 48. Fern, et al., Advanced Functional Materials. 18(24): p. 3931 2008 49. Sharma P., et al., Chemistry of Materials. 20(19): p. 6087. 2008 50. Tu C., et al., Journal of the American Chemical Society. 132(6): p. 2016. 51. Maeda Y., Yoshino T.Matsunaga T., Journal of Materials Chemistry. 19(35): p. 6361. 2009 52. Saha A., Basiruddin S. K., Pradhan N.Jana N. R., Langmuir. 26(6): p. 4351. 2009 53. Ge Y., et al., Nanoscale Research Letters. 4(4): p. 287. 2009 54. Li G., et al., Macromolecules. 42(21): p. 8561. 2009 55. Gallagher J. J., et al., Journal of Materials Chemistry. 19(24): p. 4081. 2009 56. Ke J.-H., et al., Biomaterials. 31(7): p. 1707. 57. Beaune G. g., MeÌ•nager C.Cabuil V. r., The Journal of Physical Chemistry B. 112(25): p. 7424. 2008

58. Perlstein B., Lublin-Tennenbaum T., Marom I.Margel S., Journal of Biomedical Materials Research Part B: Applied Biomaterials. 92B(2): p. 353 59. Lee J. E., et al., Journal of the American Chemical Society. 132(2): p. 552. 2009 60. Di Corato R., et al., Macromolecular Bioscience. 9(10): p. 952 2009 61. Tsotsalas M., et al., Chemistry of Materials. 20(18): p. 5888. 2008 62. Govindaiah P., et al., Journal of Colloid And Interface Science. 343(2): p. 484. 63. Park Y. I., et al., Advanced Materials. 21(44): p. 4467 2009 64. Nagao D., et al., Colloid and Polymer Science. 286(8-9): p. 959 2008 65. Nagao D., et al., Langmuir. 24(17): p. 9804. 2008 66. Chang I. P., Hwang K. C.Chiang C.-S., Journal of the American Chemical Society. 130(46): p. 15476. 2008 67. Min Yeo K., Ji Gao C., Ahn K.-H.Su Lee I., Chemical Communications. 2008(38): p. 4622. 2008 68. He H., Xie M. Y., Ding Y.Yu X. F., Applied Surface Science. 255(8): p. 4623. 2009 69. Skaat H.Margel S., Biochemical and Biophysical Research Communications. 386(4): p. 645. 2009 70. Zhang, et al., Journal of Materials Chemistry. 19(22): p. 3696. 2009 71. Ashokan A., Menon D., Nair S.Koyakutty M., Biomaterials. 31(9): p. 2606. 72. Salgueirino-Maceira V.Correa-Duarte M. A., Advanced Materials. 19(23): p. 4131. 2007 73. Vriezema D. M., et al., Chemical Reviews. 105(4): p. 1445. 2005 74. Tong R., et al., Mrs Bulletin. 34(6): p. 422. 2009 75. Jain S.Bates F., Science. 300(5618): p. 460. 2003 76. Lim Soo P.Eisenberg A., Journal of Polymer Science Part B: Polymer Physics. 42(6): p. 923 2004 77. Hayward R. C.Pochan D. J., Macromolecules. 43(8): p. 3577. 78. He P. T., et al., Soft Matter. 6(7): p. 1539. 79. Bakalova R., et al., Analytical Chemistry. 78(16): p. 5925. 2006 80. Kaplan I. M., Wadia J. S.Dowdy S. F., Journal of Controlled Release. 102(1): p. 247. 2005 81. Santra, Chemical Communications,(25): p. 3144. 2005 82. Liu L., et al., Biomaterials. 29(10): p. 1509. 2008 83. Howarth M., Takao K., Hayashi Y.Ting A. Y., Proceedings of the National Academy of Sciences of the United States of America. 102(21): p. 7583. 2005 84. Winter J. O., Liu T. Y., Korgel B. A.Schmidt C. E., Advanced Materials. 13(22): p. 1673. 2001 85. Gomez N., et al., Talanta. 67(3): p. 462. 2005 86. Slotkin J R C. L. H. N. D. R. S. E. C. T. H. B. S. B. G. I. G. J. G. C. T. F. H., Developmental Dynamics. 236(12): p. 3393. 2007 87. Gao X., et al., Bioconjugate Chemistry. 19(11): p. 2189. 2008 88. Pisanic T. R., 2nd, et al., Biomaterials. 28(16): p. 2572. 2007 89. Maysinger D., Behrendt M., Lalancette-Hebert M.Kriz J., Nano Letters. 7(8): p. 2513. 2007 69

90. Tang M. L., et al., Biomaterials. 29(33): p. 4383. 2008 91. Tseng P., Di Carlo D.Judy J. W., Nano Letters. 9(8): p. 3053. 2009 92. Xu Y., et al., The Journal of Physical Chemistry C. 114(11): p. 5020. 93. Pellegrino T., et al., Differentiation. 71(9-10): p. 542. 2003 94. Fabio C., et al., Small. 5(22): p. 2555. 2009 95. Orndorff R. L., et al., Nano Letters. 8(3): p. 780. 2008 96. Bats C., Groc L.Choquet D., Neuron. 53(5): p. 719. 2007 97. De Koninck P., Labrecque S., Heyes C. D.Wiseman P. W., HFSP Journal. 1(1): p. 5. 2007 98. Bannai H., et al., Nat. Protocols. 1(6): p. 2628. 2007 99. Gussin H. A., et al., Journal of the American Chemical Society. 128(49): p. 15701. 2006 100. Bouzigues C. d., Morel M., Triller A.Dahan M., Proceedings of the National Academy of Sciences. 104(27): p. 11251. 2007 101. Nechyporuk-Zloy V., et al., Am J Physiol Cell Physiol. 291(2): p. C266. 2006 102. Nechyporuk-Zloy V., et al., Am J Physiol Cell Physiol. 294(4): p. C1096. 2008 103. Cui B., et al., Proceedings of the National Academy of Sciences. 104(34): p. 13666. 2007 104. Echarte M. M., et al., FEBS Letters. 581(16): p. 2905. 2007 105. Vu T. Q., et al., Nano Letters. 5(4): p. 603. 2005 106. Rosenthal S. J., et al., Journal of the American Chemical Society. 124(17): p. 4586. 2002 107. Zhang Q., Cao Y.-Q.Tsien R. W., Proceedings of the National Academy of Sciences. 104(45): p. 17843. 2007 108. Pappas T. C., et al., Nano Lett. 7(2): p. 513. 2007 109. Nadeau J. L., Clarke S. J., Hollmann C. A.Bahcheli D. M. Quantum dot-FRET systems for imaging of neuronal action potentials. in Engineering in Medicine and Biology Society, 2006. EMBS '06. 28th Annual International Conference of the IEEE. 2006. 110. Khatchadourian, Journal of Biomedicine and Biotechnology. 2007: p. 1. 2007 111. Clarke S. J., et al., Nat Mater. 5(5): p. 409. 2006 112. Wang N.Ingber D. E., Biochemistry and Cell Biology-Biochimie Et Biologie Cellulaire. 73(7-8): p. 327. 1995 113. Mannix R. J., et al., Nat Nano. 3(1): p. 36. 2008 114. Polte T. R., et al., Biomaterials. 28(17): p. 2783. 2007 115. Hughes S., McBain S., Dobson J.El Haj A. J., Journal of the Royal Society Interface. 5(25): p. 855. 2008 116. Guzman R., et al., Proceedings of the National Academy of Sciences of the United States of America. 104(24): p. 10211. 2007 117. Almut Focke S. S. A. F. J. S. J. M. C. Z. J. S., Magnetic Resonance in Medicine. 60(6): p. 1321. 2008 118. Gao X., Chan W. C.Nie S., J Biomed Opt. 7(4): p. 532. 2002 119. Chan W. C., et al., Curr Opin Biotechnol. 13(1): p. 40. 2002 120. Alivisatos A. P., Gu W.Larabell C., Annu Rev Biomed Eng. 7: p. 55. 2005 70

121. Biju V., et al., Analytical and Bioanalytical Chemistry. 391(7): p. 2469. 2008 122. Sanvicens N.Marco M. P., Trends in Biotechnology. 26(8): p. 425. 2008 123. Cozzoli P. D., Pellegrino T.Manna L., Chem Soc Rev. 35(11): p. 1195. 2006 124. Quarta A., et al., IEEE Trans Nanobioscience. 6(4): p. 298. 2007 125. Corr S. A., Rakovich Y. P.Gun'ko Y. K., Nanoscale Research Letters. 3(3): p. 87. 2008 126. Yi D. K., et al., J. Am. Chem. Soc. 127(14): p. 4990. 2005 127. Wang, Journal of the American Chemical Society. 129(13): p. 3848. 2007 128. Gao J. H., et al., Journal of the American Chemical Society. 129(39): p. 11928. 2007 129. Zhang Y., et al., Journal of Biomedical Materials Research Part A. 85A(3): p. 840. 2008 130. Murray C. B., Kagan C. R.Bawendi M. G., Annual Review of Materials Science. 30(1): p. 545. 2000 131. Sun S., et al., Journal of the American Chemical Society. 126(1): p. 273. 2003 132. Yu W. W., Chang E., Drezek R.Colvin V. L., Biochem Biophys Res Commun. 348(3): p. 781. 2006 133. Gupta A. K.Gupta M., Biomaterials. 26(18): p. 3995. 2005 134. Wuister S. F.Meijerink A., Journal of Luminescence. 102: p. 338. 2003 135. Winter J. O., et al., Colloids and Surfaces a-Physicochemical and Engineering Aspects. 254(1-3): p. 147. 2005 136. Kim D. K., et al., Journal of Magnetism and Magnetic Materials. 225(1-2): p. 30. 2001 137. Aslam M., Fu L., Li S.Dravid V. P., Journal of Colloid and Interface Science. 290(2): p. 444. 2005 138. Li H., Shih W. Y.Shih W. H., Nanotechnology. 18(49). 2007 139. Zhang Y., et al., Journal of Luminescence. 128(12): p. 1948. 2008 140. Aldana J., Lavelle N., Wang Y.Peng X., Journal of the American Chemical Society. 127(8): p. 2496. 2005 141. Gerion D., et al., The Journal of Physical Chemistry B. 105(37): p. 8861. 2001 142. Rong H., et al., Nanotechnology,(31): p. 315601. 2007 143. Rao C. N. R., Seshadri R., Govindaraj A.Sen R., Materials Science & Engineering R-Reports. 15(6): p. 209. 1995 144. Shenderova O. A., Zhirnov V. V.Brenner D. W., Critical Reviews in Solid State and Materials Sciences. 27(3-4): p. 227. 2002 145. Chang Y. R., et al., Nature Nanotechnology. 3(5): p. 284. 2008 146. Xing Y.Dai L., NANOMEDICINE. 4(2): p. 207. 2009 147. Schrand A. M., Hens S. A. C.Shenderova O. A., Critical Reviews in Solid State and Materials Sciences. 34(1-2): p. 18. 2009 148. Mohan N., et al., Advanced Materials. 22(7): p. 843 149. Hui Y. Y., et al., Optics Express. 18(6): p. 5896. 150. Fu C.-C., et al., Proceedings of the National Academy of Sciences. 104(3): p. 727. 2007 151. Liming Z., et al., Small. 6(4): p. 537. 71

152. Wang B. L., et al., Nanotechnology. 21(2). 153. Eda G., et al., Advanced Materials. 22(4): p. 505. 154. He S. J., et al., Advanced Functional Materials. 20(3): p. 453. 155. Lu F., et al., Advanced Materials. 21(2): p. 139 2009 156. Cherukuri P., Bachilo S. M., Litovsky S. H.Weisman R. B., Journal of the American Chemical Society. 126(48): p. 15638. 2004 157. Wu H.-C., et al., Journal of Materials Chemistry. 20(6): p. 1036. 158. Liu Z., Tabakman S., Welsher K.Dai H., Nano Research. 2(2): p. 85 2009 159. Liu H. P., Ye T.Mao C. D., Angewandte Chemie-International Edition. 46(34): p. 6473. 2007 160. Ray S. C., Saha A., Jana N. R.Sarkar R., The Journal of Physical Chemistry C. 113(43): p. 18546. 2009 161. Zhang W., et al., Chemistry Letters. 38(1): p. 28. 2009 162. Mao X. J., et al., Spectrochimica Acta Part A: Molecular and Biomolecular Spectroscopy. 75(2): p. 553. 163. Anderson J. M.Shive M. S., Advanced Drug Delivery Reviews. 28(1): p. 5. 1997 164. Panyam J.Labhasetwar V., Advanced Drug Delivery Reviews. 55(3): p. 329. 2003 165. Liu R., et al., Angewandte Chemie International Edition. 48(25): p. 4598 2009