Aggregation and UptakeKinetics of Gold Nanoparticles

in Biological Cells, UsingPlasmon Coupling and ImageCorrelation Spectroscopy

A thesis submitted for the degree of

Doctor of Philosophy

by

A S M Mohsin

Centre for Micro - Photonics

Faculty of Science, Engineering and Technology

Swinburne University of Technology

Melbourne, Australia

Supervisor: A/P James Chon

2015

For my parents, family and friends.

This would not have been possible without you.

2

“And your Lord has decreed that you not worship except Him, and to

parents, good treatment. Whether one or both of them reach old age [while]

with you, say not to them [so much as], ‘uff,’ [i.e., an expression of irritation

or disapproval] and do not repel them but speak to them a noble word. And

lower to them the wing of humility out of mercy and say, ‘My Lord, have

mercy upon them as they brought me up [when I was] small.’” [Qur’an:

Chapter 17, Verses 23 - 24]

3

Declaration

I, A S M Mohsin, declare that this thesis entitled :

“Aggregation and Uptake Kinetics of Gold Nanoparticles in Bio-

logical Cells, Using Plasmon Coupling and Image Correlation

Spectroscopy”

is my own work and has not been submitted previously, in whole or in part,

in respect of any other academic award.

A S M Mohsin

Centre for Micro - Photonics

Faculty of Science Engineering and Technology

Swinburne University of Technology

Australia

Dated this day, December 04, 2015

i

ii

Abstract

Just imagine the world is out of focus, with no option of putting on glasses.

That is precisely the frustrating situation for scientists wanting to observe

molecules and living cells under optical microscopes. Anything smaller than

200 nm (objects ~200 times smaller than the width of a single human hair)

looks blurry. Abbe’s (1873) theoretical limit of resolution is about half the

wavelength of the light used, which translates to about 200 nm for visible

light. Thus, the shape of bacteria (1 µm) and mitochondria (200 nm) can be

seen, but not the internal structures. But can resolution be improved?

Recently, gold nanoparticles (AuNPs) have gained enormous interest, partly

due to their distinctive optical properties. The anisotropic shape of AuNPs

offers longitudinal and transverse surface plasmon resonance (SPR) in infrared,

and strong light absorption and plasmon coupling properties make them great

candidates for use in sensing, labelling and imaging. Although considerable

work has been done on the functionalisation of plasmonic nanoparticles

(PNPs), PNPs uptake and cytotoxicity and the qualitative and quantitative

uptake of PNPs, molecular aggregation has still not been demonstrated.

Inductively coupled plasma atomic emission (ICP - AEP) and inductively

coupled plasma mass spectroscopy (ICP - MS), together with transmission

electron microscopy (TEM), can quantify AuNPs uptake, but are not suitable

for live cell imaging due to their destructive nature. Several microscopy

iii

techniques have been used to investigate molecular activity at sub - microscopic

resolution without destroying cells, but each method has limitations. For

example, image correlation microscopy (ICM) can characterise larger protein

assemblies, but is limited to sub - microscopic levels and is highly sensitive to

background interference. Similarly, nanoparticle plasmon coupling can provide

a microscopic but not macroscopic view of cellular interactions.

The aim of this PhD was to develop a technique to quantify the uptake and

aggregation of AuNPs using image correlation spectroscopy (ICS). We pro-

posed a non - destructive microscopic optical method using image correlation

spectroscopy (ICS) together with plasmon coupling. The combination of these

techniques provides an indication of what is happening within cells, both at

the microscopic and macroscopic level. After successful demonstrations of H

- ICS under experimental conditions (i.e. dielectric samples), we quantified

AuNPs uptake and aggregation of AuNPs inside human cervical carcinoma

(HeLa) cells. Hence, my original contribution to knowledge is developing a

technique that can detect PNPs uptake kinetics and aggregation dynamics in

a live cellular environment. This technique could be used in different biological

applications including cancer therapy, drug delivery, disease diagnosis and also

for probing membrane protein stoichiometry and dynamics. Although visual

resolution itself may not be improved, the ability to ‘see’ inside cells may be

enhanced by the use of the methods discovered here.

iv

Acknowledgements

I would like to express my gratitude to my supervisor, Assoc. Prof. James

Chon, for his valuable expertise, experience and guidance and giving me the

opportunity to tackle this challenging project, and the support to finish it.

I would like to thank my co - supervisor, Prof. Saulius Juodkazis, for his

guidance and support. I would also like to thank to Assoc. Prof. Andrew

Clayton for his suggestions, guidance and fruitful discussions.

I would like to thank Dr. Adam Taylor for his support with synthesising

AuNPs, how to deal with femtosecond laser and constructive criticism. Thanks

also to Tim Chow for his support with synthesis, lab induction and high order

image correlation spectroscopy (H - ICS) simulation code. Thanks Bio 21 and

Swinburne Nanolab for use of the transmission electron microscopy (TEM) and

scanning electron microscopy (SEM) facilities. I would also like to thank Ms.

Pierrette Michaux for electron beam microscopy (EBL) fabricating a custom -

made grid and helping me perform the tedious SEM for correlating the optical

images. Thanks also to Dr. Chiara Paviolo, Katharine Adcroft, De Ming

Zhu and Matthew Quinn for training me in various methods and equipment.

I would like to thank Barbara Gillespie, Swinburne Research and Swinburne

Information Technology service (ITS). Thanks Dr. Priyamvada Venugopalan

for encouragement, and for guiding me to write my thesis on Lyx. Thanks

also to Arif, Salman, Zubaidah, Ali, Ivylo, Amit and Yala for discussions and

sharing equipment, and to all the students and members of centre for micro

v

photonics (CMP) for providing such a wonderful research environment.

Finally, I would like to thank my parents for believing in me and supporting

me through this journey. I would like to thank my elder brother Dr. A. K. M

Momin and A. N. M Mamun, who have inspired and supported me. I would

like to thank my elder sister Parveen Akther, her husband Abdul Malek Sarkar,

my sister in law Shahana Rahman and my other brothers and sisters. I would

also like to thank my nieces especially Fatema and Tanzima and nephews

Riad and Rifat for their support, love and best wishes. Last but not the least,

I would like to thank Momel, Kazi, Dr. Imran, Sayem, Maruf, Dr. Ayaz,

Dr. Zia, Dr. Tanveer, Dr. Razib, Dr. Imrul, Wasim, Ifat, Nidi Apu, Dr.

Seemin, Shirin Apu, Dana Apu, Dr. Nazia and Dr. Urmi and the “Swinburne

Bangladeshi Community” for providing a friendly and homely environment in

Melbourne. My work would not have been possible without all of the support

that I received from those around me. This thesis is dedicated to all of you.

A S M Mohsin

Melbourne, Australia

4th December, 2015

vi

Contents

Declaration i

Abstract iii

Acknowledgements v

Contents vii

List of Tables xi

List of Figures xii

1 Introduction 1

1.1 Introduction of plasmonic nanoparticles in biological cell applic-

ations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1

1.2 Application of plasmonic nanoparticles in biology . . . . . . . . 3

1.2.1 Light - scattering imaging . . . . . . . . . . . . . . . . . 3

1.2.2 Refractive index sensing . . . . . . . . . . . . . . . . . . 4

1.2.3 Assembly based sensing . . . . . . . . . . . . . . . . . . 4

1.2.4 Inter - particle coupling effects . . . . . . . . . . . . . . . 5

1.2.5 Gold plasmonic nanoparticle - cell interaction . . . . . . 6

1.3 Gold nanoparticle preparation . . . . . . . . . . . . . . . . . . . 13

1.3.1 Wet chemical synthesis . . . . . . . . . . . . . . . . . . . 14

1.3.2 Gold nanoparticle surface modification . . . . . . . . . . 17

1.4 Aim and methodology of this thesis . . . . . . . . . . . . . . . 18

vii

1.4.1 Specific aim . . . . . . . . . . . . . . . . . . . . . . . . . 18

1.4.2 Methodologies . . . . . . . . . . . . . . . . . . . . . . . . 19

1.4.3 This thesis . . . . . . . . . . . . . . . . . . . . . . . . . . 21

2 Theory and simulations of surface plasmon resonance and

plasmon coupling 25

2.1 Abstract . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25

2.2 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26

2.3 Theory of surface plasmon resonance . . . . . . . . . . . . . . . 26

2.4 Theory of surface plasmon resonance of metallic nanoparticles . 27

2.4.1 Extinction of light by a nanosphere . . . . . . . . . . . . 28

2.4.2 Extinction of light by a nanorod . . . . . . . . . . . . . 33

2.5 Basics mathematical and physical formalism behind finite dif-

ference time domain (FDTD) technique . . . . . . . . . . . . . . 36

2.5.1 Simulation setup . . . . . . . . . . . . . . . . . . . . . . 39

2.6 Finite difference time domain (FDTD) simulations of standalone

particles . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41

2.6.1 Finite difference time domain (FDTD) simulations of

gold nanospheres and nanorods . . . . . . . . . . . . . . 41

2.6.2 Quantum yield variation due to tip curvature . . . . . . 44

2.7 Finite difference time domain (FDTD) simulations of coupled

nanoparticles . . . . . . . . . . . . . . . . . . . . . . . . . . . . 46

2.7.1 Dipolar excitation coupling model and plasmon ruler

equation . . . . . . . . . . . . . . . . . . . . . . . . . . . 46

2.7.2 Numerical simulation of gold nanosphere dimer . . . . . 53

2.7.3 Numerical simulation of gold nanorod dimer . . . . . . . 55

2.7.4 Numerical simulation of gold nanosphere trimer . . . . . 57

2.7.5 Numerical simulation of gold heterodimer nanorod . . . . 57

2.8 Quantum yield of gold nanoparticles . . . . . . . . . . . . . . . 59

viii

2.9 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 60

3 Theory and simulations of image correlation spectroscopy 61

3.1 Abstract . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 61

3.2 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 61

3.3 Image correlation spectroscopy (ICS) . . . . . . . . . . . . . . . 62

3.4 High order image correlation spectroscopy (H - ICS) . . . . . . . 64

3.4.1 Interpreting high order image correlation spectroscopy

(H - ICS) for plasmon coupled particles . . . . . . . . . . 67

3.5 Factors affecting precision of image correlation spectroscopy (ICS) 68

3.6 High order image correlation spectroscopy (H - ICS) simulations 69

3.7 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 74

4 Image correlation spectroscopy of plasmon coupled gold

nanoparticles into dielectric medium 75

4.1 Abstract . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 75

4.2 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . 75

4.3 Nanoparticle plasmon coupling and simulations . . . . . . . . . 77

4.4 High order image correlation spectroscopy (H - ICS) of plasmon

coupled nanoparticles . . . . . . . . . . . . . . . . . . . . . . . 79

4.5 Experimental . . . . . . . . . . . . . . . . . . . . . . . . . . . . 80

4.5.1 Sample preparation . . . . . . . . . . . . . . . . . . . . . 80

4.5.2 Grid fabrication . . . . . . . . . . . . . . . . . . . . . . . 81

4.6 Results and discussion . . . . . . . . . . . . . . . . . . . . . . . 81

4.6.1 High order image correlation spectroscopy (H - ICS) for

plasmon coupled dielectric samples . . . . . . . . . . . . 82

4.6.2 Validating high order image correlation spectroscopy (H

- ICS) results using single particle spectroscopy . . . . . 89

4.7 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 91

ix

5 Gold nanoparticle uptake and aggregation dynamics in HeLa

cells using image correlation spectroscopy 93

5.1 Abstract . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 93

5.2 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . 94

5.3 High order image correlation spectroscopy (H - ICS) of plasmon

coupled nanoparticles . . . . . . . . . . . . . . . . . . . . . . . . 95

5.4 Surface modified AuNPs - HeLa cell sample preparation . . . . . 97

5.5 Experimental results . . . . . . . . . . . . . . . . . . . . . . . . 101

5.5.1 High order image correlation spectroscopy (H - ICS) for

gold nanoparticle incubated HeLa cell images . . . . . . 101

5.5.2 Gold nanoparticle uptake due to surface modification . . 105

5.5.3 Gold nanoparticle oligomerisation due to surface modi-

fication . . . . . . . . . . . . . . . . . . . . . . . . . . . 109

5.5.4 Effect of size on gold nanoparticle uptake . . . . . . . . 110

5.5.5 Effect of size on gold nanoparticle oligomerisation . . . . 111

5.6 Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 112

5.7 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 117

6 Conclusions and future work 119

6.1 Thesis conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . 119

6.2 Future research . . . . . . . . . . . . . . . . . . . . . . . . . . . 121

6.2.1 More than two emitter . . . . . . . . . . . . . . . . . . . 121

6.2.2 Validation by other techniques . . . . . . . . . . . . . . . 124

6.3 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 126

Bibliography 127

x

List of Tables

1.1 Effect of size and shape of gold nanoparticles (AuNPs) on

endoctytosis.[76] . . . . . . . . . . . . . . . . . . . . . . . . . . 10

2.1 Comparison of longitudinal surface plasmon resonance (LSPR)

and scattering cross - sections (Scat.CS) of different particle

morphologies for aspect ratio (AR) 2, length 40 nm PNPs,

calculated via FDTD simulations. . . . . . . . . . . . . . . . . 43

5.1 UV - vis spectrum peak (nm), size distribution (nm) and pH

for bare, PEG and maleimide coated 50 nm, 80 nm and 100

nm diameter gold nanosphere (AuNS). Here, 1 represents the

measured value 2 and 3 represents company supplied values. . . 97

5.2 Zeta potential, mobility and pH for bare, PEG and maleimide

coated 100 nm diameter AuNSs. . . . . . . . . . . . . . . . . . . 98

xi

List of Figures

1.1 (A) Transmission electron microscope (TEM) images of wet

chemically synthesised gold nanospheres (AuNSs) (diameter

80 ± 6.5 nm) after drop - casting onto a TEM grid, (B)

corresponding histogram showing size distribution. . . . . . . . . 14

1.2 (A) Transmission electron microscope (TEM) images of wet

chemically synthesised gold nanorods (AuNRs) (aspect ratio

3.8) after drop - casting onto a TEM grid, (B) corresponding

histogram showing size distribution. . . . . . . . . . . . . . . . . 15

1.3 Chemical structure of citrate, polyethylene glycol (PEG) and

maleimide. (adopted from wikipedia) . . . . . . . . . . . . . . . 16

2.1 (A) Schematics for plasmon oscillation for a sphere, (B)surface

plasmon resonance (SPR) spectrum of 40 nm radius gold

nanospheres calculated using Mie theory [129], refractive index

1.33. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27

xii

2.2 Surface plasmon resonance (SPR) spectrum of nanorods with

semi major axis 43 nm and semi minor axis 10 nm calculated

using Mie - Gans theory, [129] refractive index 1.33. While

gold nanospheres show one SPR band in the visible region, gold

nanorods show two bands: a strong longitudinal band in the

near infrared region corresponding to electron oscillation along

the long axis and a weak transverse band, similar to that of

gold nanospheres in the visible region corresponding to electron

oscillations along the short axis. . . . . . . . . . . . . . . . . . . 28

2.3 (A) Experimental measurement from Johnson and Christy for

the dielectric function of gold, showing real and imaginary parts

of the dielectric constant. [142] (B) Variation of extinction cross

- section spectra predicted by Mie theory [129] for 10 nm radius

gold nanospheres, immersed in a media with various refractive

indices (C) Variation in the extinction cross - section spectra

predicted by Mie theory [129] for gold nanospheres of various

radii immersed in a media with a refractive index n = 1.33. . . . 30

2.4 Standard Yee - cell. The electric field components are located

on the edges while the magnetic field components are located

on the face centres. Figure taken from Wikipedia. [156] . . . . . 38

2.5 Layout editor of finite difference time domain (FDTD) sim-

ulation for plasmonic nanoparticles (PNPs) structure. The

yellow rectangular box is the total field, the white rectangular

box is the total - field scattered field source and outer yellow

rectangular box is the scattered field. The pink arrow shows

the direction of propagation, k vector. The blue dot represents

the direction of the electric field . . . . . . . . . . . . . . . . . . 39

xiii

2.6 Extinction, absorption and scattering cross - sections calculated

via Mie theory [129] compared with finite difference time domain

(FDTD) simulations for, (A) 10 nm radius gold nanospheres and

(B) 20 nm radius gold nanospheres. . . . . . . . . . . . . . . . . 41

2.7 (A) Scattering cross - sections of 5 - 100 nm radius gold

nanospheres calculated via Mie theory compared with finite

difference time domain (FDTD) simulations, (B) longitudinal

surface plasmon resonance (LSPR) of 5 - 100 nm radius gold

nanospheres calculated via Mie theory compared with finite

difference time domain (FDTD) simulations. . . . . . . . . . . . 42

2.8 Schematics of different particle morphologies under considera-

tion including, (A) a prolate spheroid, (B) a spherically capped

cylinder, (C) an ellipsoidally capped cylinder and (D) a cylinder. 44

2.9 Scattering cross - sections of different morphology nanoparticles

for aspect ratio (AR) 2, length 40 nm refractive index 1.33 and

mesh size 1 nm using FDTD simulations. . . . . . . . . . . . . . 45

2.10 Finite difference time domain (FDTD) simulations of gold, (A)

dumbbell, (B) nanorods and (C) bipyramids. . . . . . . . . . . . 45

2.11 Scattering cross - sections (per unit volume) of spheres (SPs),

dumbbells (DBs), nanorods (NRs) and bipyramids (BPs) using

FDTD calculations. . . . . . . . . . . . . . . . . . . . . . . . . . 47

2.12 Schematic of the energy level splitting resulting from the dipolar

interaction of AuNR dimer, showing symmetric (ψ+) and anti

- symmetric coupling (ψ−) of excitons for (A) H aggregate geo-

metry and (B) J aggregate geometry. (C) Exciton theory picture

of the nature of the coupled longitudinal plasmon excitation in

AuNRs dimers: electromagnetic analogy to molecular orbital

theory. [160] . . . . . . . . . . . . . . . . . . . . . . . . . . . . 50

xiv

2.13 Plasmon coupling of nanoparticle at (A) weak and (B) strong

coupling regime, (C) the energy level splitting resulting from

the dipolar coupling of dimers, showing symmetric and anti -

symmetric coupling for AuNS dimer. [171] . . . . . . . . . . . . 51

2.14 Finite difference time domain (FDTD) simulation results of

20 nm radius gold nanosphere dimer plasmon coupling. (A)

scattering spectrum and (B) plasmon resonance peak shift as a

function of inter - particle separations. . . . . . . . . . . . . . . 52

2.15 Finite difference time domain (FDTD) simulation results of

40 nm radius gold nanosphere dimer plasmon coupling. (A)

Scattering spectrum and (B) Plasmon resonance peak shift as

a function of inter - particle separations. . . . . . . . . . . . . . 53

2.16 Finite difference time domain (FDTD) simulation results of gold

nanorod dimer plasmon coupling. (A) scattering spectrum of

rod for length 75 nm, width 20 nm, aspect ratio (AR) 3.8 and

(B) plasmon resonance peak shift as a function of inter - particle

separations. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 55

2.17 Finite difference time domain (FDTD) simulation results of gold

nanorod dimer plasmon coupling. (A) scattering spectrum of

rod for length 43 nm, width 10 nm, aspect ratio (AR) 4.3 and

(B) plasmon resonance peak shift as a function of inter - particle

separations. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 56

2.18 Finite difference time domain (FDTD) simulation results of gold

nano - sphere trimer plasmon coupling. (A) scattering spectrum

of a sphere - trimer with a diameter of 40 nm for different inter

- particle distances for weakly coupling regimes compared with

monomer and dimer and (B) plasmon resonance peak shift as a

function of inter - particle separations. . . . . . . . . . . . . . . 57

xv

2.19 Finite difference time domain (FDTD) simulation results of

gold nanorod hetero dimer plasmon coupling. (A) Scattering

of hetero dimer (spherically capped) rod having AR 4.3, length

43 nm, width 10 nm and AR 3.8, length 75 nm, width 20 nm

and (B) Au nanorod dimers peak wavelengths are shown for

different inter - particle distances. . . . . . . . . . . . . . . . . 58

2.20 Scattering quantum yield (QY) ratio of clusters and monomers

of gold nanospheres (AuNS ) (80 nm diameter) . . . . . . . . . . 60

3.1 Autocorrelation of an image. . . . . . . . . . . . . . . . . . . . 63

3.2 High order image correlation spectroscopy (H - ICS) correlation

functions. Autocorrelation of squared and cubed images and

corresponding high order correlation function. . . . . . . . . . . 65

3.3 Finite difference time domain simulations and quantum yield of

80 nm diameter gold nanosphere (AuNSs), (A) scattering cross

- sections of 80 nm diameter AuNS at different separation and

(B) quantum yield with respect to separation/diameter. . . . . 66

3.4 Simulated confocal laser scattering microscopy images (CLSM)

containing monomer and dimer mixture. . . . . . . . . . . . . . 68

3.5 Typical confocal laser scattering point spread function profile

from single particles (blue) and dimer (black). Red colour

spectrum indicate analytical point spread function using Vec-

torial Debye theory for objective 1.4 NA, at 715 nm wavelength

for circular polarisation. Debye theory [183] can be used

to calculate the diffraction pattern of an objective of high

numerical aperture. . . . . . . . . . . . . . . . . . . . . . . . . 70

3.6 High - order image correlation spectroscopy (H - ICS) simulation

results. The plots show the (A)N1, (B)N2 and (C) Alpha of the

simulated sample without background noise. Each data point

is averaged by analysis of 500 images. . . . . . . . . . . . . . . 72

xvi

3.7 High order image correlation spectroscopy (H - ICS) simulation

results. The plots show the (A) N1, (B)N2 and (C) Alpha of

the simulated sample with signal to noise ratio (SNR) = 3 0,

and the e - radius of the diameters is 1.2 times of that of the

monomers. Each data point is averaged by analysis of 500 images. 73

3.8 (A) Dark - field scattering images for AuNS incubated human

cervical carcinoma (HeLa) cell samples, (B) selected noise loca-

tion for high - order image correlation spectroscopy simulations,

(C) recorded noise images in high - order image correlation

spectroscopy (H - ICS) simulations and (D) recorded AuNS

attached HeLa cell images with noise correction in high order

image correlation spectroscopy (H - ICS) simulations. . . . . . . 73

4.1 (A) Transmission electron microscope (TEM) images of gold

nanoparticles (AuNSs) dropcasted onto a TEM grid, and (B)

dimer separation histogram, showing 75% of dimers are within

10% of separation of diameter. . . . . . . . . . . . . . . . . . . . 78

4.2 Gold nanoparticle (AuNP) characterisation: UV- vis spectra of

bare gold nanosphere (AuNS) of diameter 80 nm compared with

Mie theory and FDTD simulations. The UV- vis spectrum is

the ensemble spectra and red shifted compared with Mie theory

and FDTD calculated for single particle spectra. . . . . . . . . . 80

4.3 Grid fabrication: (A) scanning electron microscope (SEM)

images of magnified grid location, (B) SEM images of fabricated

full grid. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 82

4.4 Schematic diagram of laser scattering confocal set up. . . . . . . 83

xvii

4.5 Correlation of optical and SEM/TEM images of 80 nm gold

nano - sphere. (A) confocal scattering images for low concen-

tration samples and (B) corresponding correlated SEM images;

(C) confocal scattering images for high concentration samples,

and (D) corresponding un - correlated (same location) TEM

images. scale bar = 4µm. . . . . . . . . . . . . . . . . . . . . . 84

4.6 Low concentration sample :- The number of gold nanoparticles

in aggregated samples was calculated using H - ICS and the

real number. (A) calculated average monomer number <N1>

per beam area, (B) average dimer number per beam area (N2),

(C) quantum yield α2 and (D) percentage of oligomers for the

selected images (A - F represent six different locations). The

error bar represents standard error. Each data point is an

average from analysis of 20 images. . . . . . . . . . . . . . . . . 85

4.7 High concentration sample :- The number of gold nanoparticles

in aggregated samples was calculated using H - ICS and the

real number. (A) average monomer number per beam area, (B)

average dimer number per beam area N2, (C) quantum yield α2

and (D) percentage of oligomers for the selected images (A - G

represent seven different locations) and E) comparison of SEM

and H - ICS dimer numbers. Thye error bar represents standard

error. Each data point is an average from analysis of 20 images. 86

4.8 Error distributions for N1, N2 and α2 among 100 cases of

images considering the contribution of monomers and high -

order clusters (e.g. trimers and tetramers). The distribution of

error varied from 0% to - 30 % due to the presence of high -

order clusters. . . . . . . . . . . . . . . . . . . . . . . . . . . . 88

xviii

4.9 Error distributions for N1, N2 and α2 among 100 cases of images

considering monomer and dimer contribution. Discarding the

contribution of high order cluster (e.g. trimers, tetramers), only

considering the contribution of monomers and dimers, the error

can be reduced to 10 % , and the accuracy of the H - ICS analysis

can be improved. . . . . . . . . . . . . . . . . . . . . . . . . . . 88

4.10 (A) Intensity variation due to polarisation sensitivity (00-1800).

Experimental values for dimer one and two extracted from

polarisation dependent images for 00-1800, fit perfectly with

cosine, showing cos2 dependency and (B) dimer spectrum

extracted from wavelength dependent images from 700 - 900

nm wavelength matches with FDTD simulated AuNS dimer

spectrum of 2 nm separation. . . . . . . . . . . . . . . . . . . . 89

4.11 Dimer number calculated using polarization spectroscopy and H

- ICS technique and compared with the dimer number extracted

from SEM images. . . . . . . . . . . . . . . . . . . . . . . . . . 90

5.1 Gold nanoparticle characterisation: UV - vis spectra of bare,

PEG and maleimide coated AuNSs of diameter, (A) 50 nm (B)

80 nm and (C) 100 nm. . . . . . . . . . . . . . . . . . . . . . . . 96

5.2 Schematic diagram of dark field microscopy set up. . . . . . . . 99

5.3 Dark - field scattering images of 50 nm diameter bare gold nano

sphere for different incubation times. . . . . . . . . . . . . . . . 100

5.4 Dark - field scattering images of 80 nm diameter maleimide

coated gold nano - spheres for different incubation times. . . . 102

5.5 Dark - field scattering images of 100 nm diameters PEG coated

gold nano - spheres for different incubation times. . . . . . . . 103

5.6 Dark - field scattering images of 80 nm diameter bare, PEG and

maleimide coated gold nano - spheres for two days incubation. . 105

xix

5.7 Cellular uptake and aggregation kinetics of 50 nm diameter

gold nanoparticles (AuNPs) as a function of incubation time

for different surface modified gold nanospheres (AuNSs). (A) H

- ICS extracted monomer number per beam area, (B) H - ICS

extracted oligomer (mostly dimer) number per beam area, (C)

H - ICS extracted quantum yield, (D) cellular uptake of gold

nanoparticles (AuNPs) and ( E) percentage of oligomers. . . . . 106

5.8 Cellular uptake and aggregation kinetics of 80 nm diameter

gold nanoparticles (AuNPs) as a function of incubation time

for different surface modified gold nanospheres (AuNSs). (A) H

- ICS extracted monomer number per beam area, (B) H - ICS

extracted oligomer (mostly dimer) number per beam area, (C)

H - ICS extracted quantum yield, (D) cellular uptake of gold

nanoparticles (AuNPs) and ( E) percentage of oligomers. . . . . 107

5.9 Cellular uptake and aggregation kinetics of 100 nm diameter

gold nanoparticle (AuNPs) as a function of incubation time for

different surface modified gold nanospheres (AuNSs). (A) H -

ICS extracted monomer number per beam area, (B) H - ICS

extracted oligomer (mostly dimer) number per beam area, (C)

H - ICS extracted quantum yield, (D) cellular uptake of AuNPs

and ( E) percentage of oligomers. . . . . . . . . . . . . . . . . . 108

5.10 Effect of size on (different surface modified) gold nanoparticle

uptake for different incubation times. . . . . . . . . . . . . . . . 111

5.11 Effect of size on (different surface modified) gold nanoparticle

oligomerisation for different incubation times. . . . . . . . . . . 112

xx

6.1 (A) Wet chemically synthesised gold nanorods (AuNRs) drop-

casted onto a transmission electron microscope (TEM) grid.

Histograms displaying the distribution of (B) aspect ratio (red),

and (C) length (green) and width (red). Measurements were

taken from transmission electron microscope (TEM) images,

using the fit ellipse option in ImageJ, with hand fit ellipses,

to avoid threshold errors. . . . . . . . . . . . . . . . . . . . . . . 122

6.2 Confocal laser scattering, gold nanoparticle internalised cell

images. Cell membrane was stained by dye molecules, clearly

visible (red color borders) in the confocal images and inter-

nalised gold nanosphere was also clearly visible (green color

particles) while excited by laser source. Scale bar represents

15 µm. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 125

xxi

Chapter 1

Introduction

1.1 Introduction of plasmonic nanoparticles in

biological cell applications

Plasmonic nanoparticles (PNPs) exhibit excellent optical properties such as

remarkable absorption, scattering, tunability in the visible to near infrared

regions due to size, shape, orientation and tip geometry. Most importantly they

do not blink or bleach like quantum dot or dyes, offering unlimited lifetime.

These properties have enabled the use of PNPs in biological application

including biological imaging, [1–6] biolabelling and sensing [7], photothermal

cancer therapy, [8, 9] drug and gene delivery [7, 10] and probing membrane

protein. [11, 12]

However, PNP - cell interaction is poorly understood at single particle level.

Previous works were mostly focused on functionalisation of PNPs, PNPs

uptake and cytotoxicity. [13–20] Research on interactions of PNPs due to

functionalisation or different surface modification is also lacking. Qualitative

and quantitative study of PNP uptake, molecular aggregation and cellular

1

Chapter 1 2

movement with respect to size, shape, incubation time [14], surface effect

[15] and nanoparticle aggregation effects [16] have not been demonstrated

methodically.

To quantify the PNP uptake, several indirect measurements such as inductively

coupled plasma atomic emission (ICP - AEP) or inductively coupled plasma

mass spectroscopy (ICP - MS) have been performed along with destructive

transmission electron microscopy (TEM) analysis. The available method could

not quantify in - situ gold nanoparticle (AuNP) uptake and interaction due

to their destructive nature not being suited to live cell imaging. The non -

destructive techniques, such as image correlation spectroscopy combined with

dark - field scattering microscopy are ideal for characterisation purposes but

have not been introduced for plasmonic nanoparticles.

Recently, nanoparticle plasmon coupling has been introduced to investigate the

interparticle distance between two micromolecules utlising the plasmon ruler,

on the basis of spectral shift due to coupling of two gold nanospheres (AuNS).

[21–25] This tool has been used for probing membrane proteins on cell surface

receptors, [26] to follow receptor trafficking [3] and to detect aggregation of

PNPs inside cells through receptor trafficking. As a methodology, plasmon

coupling provides a microscopic view of the interaction in these applications,

but there is no macroscopic view of the interaction available at the cellular size

regime (~ 10 µm image scale). So it becomes a necessity to develop a tool that

can quantify PNP uptake and aggregation and provide the microscopic and

macroscopic view of interactions at the cellular size regime. To address this

challenging task, we combined the plasmon coupling technique with image

correlation spectroscopy. Plasmon coupling provide information on local

interactions of PNPs between two interacting molecules at the microscopic

level while the image correlation spectroscopy (ICS) tells us what is happening

at the cellular size regime providing a microscopic picture. In this work, we

Chapter 1 3

studied the feasibility of applying the combined methods (ICS and NP plasmon

coupling) to quantify PNPs uptake and aggregation dynamics for different

functionalised PNPs as a function of size and incubation time into human

cervical carcinoma (HeLa) cells.

1.2 Application of plasmonic nanoparticles in

biology

1.2.1 Light - scattering imaging

Plasmonic AuNPs scatter light strongly at the surface plasmon resonance

(SPR) frequency. [27] A 40 nm gold nanoparticle (AuNP) offers a scattering

cross - section five orders of magnitude larger than typical dye molecules. [28]

Most importantly, they are photostable and do not blink or bleach like dye

molecules. [27] Their sizes are comparable to biological systems, which is why

they are very promising for optical imaging, allowing biological labelling with

high spatial resolution. [29, 30] Furthermore, AuNPs can be conjugated with

specific targeting molecules such as proteins, antibodies, antigens and DNA

for providing molecular specific resolutions. Strong surface plasmon resonance

(SPR) scattering allows us to use much simpler and less expensive techniques

such as dark - field scattering, whereas most of the imaging techniques require

lasers and sophisticated experimental setups. [5, 31] On the other hand, in

a dark - field setup, the optical microscope is equipped with a dark - field

condenser and a microscope lense. The AuNPs are excited by white light

sources and broad spectrums are collected using a spectrogram and imaged

using a charge coupled device (CCD). Cancer cells can be identified by AuNPs

conjugated with anti - epidermal growth factor receptor (EGFR) antibodies to

target the cancer cells. [5] Depending on the disease, biomarkers of targeted

Chapter 1 4

molecules, the targeting ligand can be chosen and conjugated to a variety of

proteins, antibodies and small molecules. [32]

1.2.2 Refractive index sensing

Surface plasmons of AuNPs are heavily dependent on the refractive index of

the medium. This property becomes helpful for optical sensing of chemical

and bimolecular analytes. [33–37] The adsorption or bindings of molecules to

the AuNP surface induces refractive index change followed by a shift of surface

plasmon resonance. By following the shift of surface plasmon resonance either

red shift for increasing the refractive index or blue shift for decreasing refractive

index it is possible to sense the changes in the environment of PNPs.

In order to sense the presence of particular chemical or biological species,

AuNPs should be conjugated with recognition molecules which specifically

bind to the target molecules at the AuNP surface. The refractive index change

depends on number of molecules bound per particle, the molar mass of the

molecules, and their proximity to PNP surface and refractive index/dielectric

properties of molecules relative to the medium in which AuNPs are dispersed.

Mulvaney and co - workers reported a 40 nm shift for 0.1 refractive index

change in liquid media for gold nanorods (AuNRs) of aspect ratio 3, which is

six times higher than gold nanosphere (AuNS) sensitivity. [38, 39]

1.2.3 Assembly based sensing

The dependence of the nanoparticle plasmon resonance on the presence of

other PNPs in close proximity has been utilised for the sensing of biomolecules.

Mirkin and co - workers [21] first showed this by using a biomolecular event

where AuNPs were conjugated with DNA strands and by observing the spectral

Chapter 1 5

shift of the solution from 520 nm to 560 nm red shift, a colour change of red to

purple resulted from the assembly. This assembly - based sensing strategy has

been introduced for detection of antibody - antigen interactions and of specific

disease biomarkers (e.g. prostate cancer). [29]

1.2.4 Inter - particle coupling effects

When AuNPs come into close proximity, there is a dramatic change in surface

plasmon resonance due to the coupling of plasmon oscillations of interacting

particles. [40–47] When the light polarisation direction is parallel to the

inter - particle axis, the SPR is red shifted, due to strongly attractive inter -

particle interactions. Conversely, when the light is polarised orthogonal to the

inter - particle axis, the plasmon spectrum is blue shifted with respect to the

single particle case, due to repulsive inter - particle interactions. [48, 49] This

phenomenon has been used for probing receptor trafficking, where cell receptors

outside the cell interact with membrane proteins and being internalised can

be observed by attaching AuNPs. Indeed, that whole process can be observed

by attaching AuNPs, their aggregation state detected by a change of colour

or change of plasmon spectrum due to plasmon coupling. [3] A similar result

was reported by Reinhard [24] in 2011, in which they looked specifically at the

membrane proteins on the surface of the cells. The interaction of receptors on

the cell surface was probed by attaching PNPs and observing colour change

due to plasmon coupling. Considering that the field has become extremely

broad, specific attention is given to AuNS and cell interaction. Discussion

about the factors affecting AuNP uptake and interaction, and techniques to

investigate this uptake and interaction provides background to assist,critical

analysis of the results. A brief section about PNPs synthesis, specific research

aims and methodology to achieve the proposed aims follows afterwards.

Chapter 1 6

1.2.5 Gold plasmonic nanoparticle - cell interaction

Understanding of PNP - cell interaction at the single particle level is

poor. Previous studies mostly focused on AuNP uptake, cytotoxicity and

aggregation. Qualitative and quantitative study of PNP uptake, molecular

aggregation with respect to size, shape and incubation time has not been

demonstrated systematically.

A previous study [14] suggests that, AuNP uptake and interaction depends on

AuNP size and shape however PNP surface modification such as PEG (poly-

ethylene glycol), PAA (poly acrylic acid), PAH (poly allylamine hydrochloride)

and maleimide etc. or functionalisation (e.g.- protein, DNA) significantly

affects their interactions. [50]

In pioneering work on AuNP uptake due to size by Chan and co - workers

[14] for 14 nm , 50 nm and 74 nm diameter AuNSs, the most efficient uptake

was reported for 50 nm diameter AuNS. In addition, Jiang and co - workers

[51] reported minimal uptake for 25 - 50 nm (among 2, 10, 25, 40, 50, 70, 80

and 90 nm) transferrin coated AuNSs, attached to HeLa cells, quantified by

transmission electron microscope (TEM) and laser capture microdissection

(LCM). Wang and co - workers [52] reported maximal uptake for 45 nm

(among 45, 70 and 100 nm) AuNSs attached to HeLa cells using TEM and

dark - field microscopy. Chithrani and co - workers [53] suggested one possible

reason for minimal uptake for optimal size reporting that, 50 nm spherical

particles require minimal time to wrap around spheres, which increases AuNP

uptake compared with other diameters, in agreement with previously reported

thermodynamic calculations. [54]

Gold nanoparticle uptake due to shape was also reported by Chan and co

- workers [14] in 2006, for 14 nm , 50 nm and 74 nm diameter AuNSs and

40×14 nm and 74×14 nm AuNRs and most efficient uptake was reported

Chapter 1 7

for 50 nm diameter AuNSs. Hence, AuNR uptake was 3 and 6 fold less

for 40×14 nm and 74×14 nm AuNRs respectively. In both of these cases

AuNP uptake was quantitatively measured by ICP - AES. Another possible

reason could be, a difference in curvature of differently - shaped nanoparticles.

Rod - shaped nanoparticles can have larger contact area with cell membrane

receptors than the spherical nanoparticles when the longitudinal axis of

the rods interacts with the receptors. This could reduce the number of

available receptor sites for binding, hence reduce the uptake. The elongated

particles (AuNRs) requires greater membrane wrapping time compared with

spherical AuNSs. Another reason for AuNP uptake could be the amount of

cetyltrimethylammonium bromide (CTAB) surfactant molecules adsorbed onto

the rod - shaped nanoparticle surface during synthesis. Due to presence of

CTAB on the surface of AuNPs, the serum protein may not be able to bind

efficiently onto the gold nanoparticle surface. This would affect the uptake of

the nanoparticles. Size and shape effects of AuNPs on endocytosis are shown

in Table.1.1.

Surface charge influences PNP uptake and cytotoxicity. Since most cells (either

cancerous or normal) seem to have negative surface charge, they offer greater

permeability for cationic particles. Generally, we cannot say cationic charge

remains as such in vivo. As serum or other proteins/NH2 acids adsorb to the

PNP surface, the surface charge is altereds. On the other hand, positively

charged PNPs (CTAB coated NRs) have greater cytotoxicity than negatively

charged NPs (citrate coated NSs), but become nontoxic after being coated

with polymer biomolecules [55–57] such as PAA, PAH, maleimide or PEG. [15,

56–58] Also most negatively charged or neutral PNPs undergo non - specific

adsorption of the particles on the cell membrane.

Recently, uptake of mesoporous spherical silica particles was demonstrated by

Slowing and co - workers [59] and they observed highest uptake for positively

Chapter 1 8

charged particles. Similar trends are evident with nanorods. Alkilany and co -

workers [13] performed a systematic study of AuNRs (18× 40 nm) for varying

concentration with and without serum protein (medium). For both cases, they

observed that, negatively charged poly (4 - styrenesulfonic acid) (PSS) AuNRs

exhibited the lowest uptake at all concentrations, while nanorods coated in

PSS followed by a layer of positively charged poly - diallyldimethylammonium

chloride (PDADMAC) exhibited the highest uptake. A possible explanation

for the higher uptake could be electrostatic interactions between positively

charged nanorods and negatively charged cell membranes of HeLa cells. [59]

Arnida and co - workers [60, 61] compared the uptake of bare and PEG coated

NSs (30, 50, 90 nm diameter) with PEG coated NRs (3×10nm, 45×10 nm)

for PC-3 and RAW 2647 cells. Through TEM and ICP - MS analysis they

reported most efficient uptake for 50 nm non - PEGlated NSs. PEGlated NR

uptake was worse than for nanospheres.

Several studies suggest there is a general trend for cellular uptake relating to

surface charge, but absolute quantification of uptake mostly depends on the

chemistry of the molecule (functional group) adsorbed onto the AuNP surfaces.

Alkilany and co - workers [13] reported molecules containing quaternary

amine (e.g. CTAB and PDADMAC) exhibit high uptake, whereas molecules

containing primary amine (e.g. PAH) exhibit lower uptake. On the other

hand, a negatively charged sulfate group (PSS), shows very low uptake. These

results indicate that functional group can influence uptake, although further

evaluation is needed on a whole array of functional groups (e.g., alcohols,

carboxylic acids) before a definitive answer is obtained.

Generally, serum proteins (bovine serum) from biological media (DMEM)

adsorb to the AuNR surface slightly increasing the size of AuNPs and

altering the surface charge of AuNPs of negative BSA (bovine serum albumin)

regardless of initial AuNR surface charge. A previous study suggested that,

Chapter 1 9

[14] AuNP uptake is receptor mediated due to the presence of serum protein in

the medium. Citrate - stabilised AuNPs contains a variety of serum proteins,

hence transferrin contains two corresponding receptors [62]. Therefore citrate

- stabilised AuNPs show three times greater uptake than transferrin - coated

nanoparticles. We can conclude that, initial surface charge of PNPs is not a

simple predictor of PNP uptake and cytotoxicity; however, PNP uptake and

cytotoxicity are governed by type of adsorbed protein and their orientation on

the PNPs surface via receptor mediated endocytosis. [63, 64]

Uptake of PNPs due to different functionalisation has been also reported in

the literature. Villanueva and co - workers [65] studied the uptake of iron

oxide nanoparticles functionalised with differently charged carbohydrates in

human cervical carcinoma (HeLa) cells and observed no uptake for neutral

NPs; however, they also observed uptake of negatively charged PNPs through

nonspecific binding. In another study, the negatively charged NPs had

most efficient uptake than positively charged NPs for cerium oxide NPs, in

adenocarcinoma lung cells [66]. Ryman - Rasmussen and co - workers [50]

reported the internalisation of different surface coated quantum dot (QD)

ellipsoid NPs into skin cells in the following order: QD- COOH > QD-NH2 >

QD - PEG.

In another study, cell membrane barrier breaches were dependent on both the

type of ligand and the arrangement of the ligand. More efficient breaching

has been observed for ribbon - like arrangements than random arrangements

on the surface. [77] Plasmonic nanoparticle interactions with oligonucleotide

(peptide or protein coated) have been studied for negatively charged PNPs

on endothelial cells and the most efficient PNP uptake has been observed

despite the negatively charged PNPs for larger density of oligonucleotides on

the PNP surfaces. In addition to oligonucleotides, synthesised PNPs could be

conjugated with positively charged biological motifs for efficient internalisation

Chapter 1 10

AuN

Ptype

andsiz

eObservedeff

ects

NSs

of14

,30,50

,74an

d10

0nm

NRsof

40×

14an

d74×

14nm

[14]

Max

imal

uptake

occu

rsfor50

nmNSs

NSs

of14

,30,50

,74an

d10

0nm

NRsof

40×

14an

d74×

14nm

[53]

Uptak

ede

pend

son

sizean

dshap

eNSs

of2,

10,2

5,40

,50,

70,8

0an

d90

nm[51]

Max

imal

uptake

occu

rsfor25

-50nm

NSs

NSs

of13

and45

nm[67]

Max

imal

uptake

occu

rsfor45

nmNSs

NSs

of45

,70an

d11

0nm

[52]

Max

imal

uptake

for45

nmNSs

NSs

of10

,20,

30an

d45

nm[68]

Max

imal

uptake

occu

rsfor30

nmNSs

NSs

of5,10

,20,

30,4

0an

d50

nm[69]

Max

imal

uptake

occu

rsfor20

nmNSs

NSs

of4,

12an

d17

nm[70]

Endo

cytosis

increase

with

increasin

gdiam

eter

NSs

of10

,25an

d50

nm[71]

Endo

cytosis

increase

with

increasin

gdiam

eter

NSs

of7,

21an

d31

nm[72]

Endo

cytosis

increase

with

increasin

gdiam

eter

NSs

of25

and50

nm[73]

Endo

cytosis

increase

with

increasin

gdiam

eter

NSs

of2,

6,an

d15

nm[74]

Max

imal

uptake

occu

rsfor2nm

NSs

NSs

of50

and51

00nm

[75]

Max

imal

uptake

occu

rsfor50

nmNSs

Table 1.1 Effect of size and shape of gold nanoparticles (AuNPs) on endoctytosis.[76]

Chapter 1 11

similar to positively charged functionalised PNPs. [78–85]

Nevertheless, it has been reported that most other nanoscale macromolecules

and molecular assemblies are internalised through the process called endo-

cytosis. [63] Endocytosis is a process by which PNPs are engulfed by a cell,

which forms vesicles of invaginated portions of plasma membrane. Endocytosis

can be classified as phagocytosis, pinocytosis or clathrin - dependent and

independent receptor mediated endocytosis. Among these, receptor mediated

endocytosis is considered to be the most effective mechanism for PNPs uptake.

Banjeri and Hayes [86] suggested an endocytic pathway for cellular AuNP

uptake through the lipid bilayer, although Xia [87] and Taylor [88] have shown

that nonendosomal AuNP uptake is possible in principle. Endocytic fate and

intracellular uptake mechanisms of PNPs have been shown before. [89] After

being taken up, PNP - protein complexes are transported to lysosomes by

vesicles where CTAB is released and proteins are digested. After that, the PNP

(e.g. nanorod) aggregates are delivered to the mitochondria which are damaged

by the release of CTAB, inducing cell apoptosis and death. Cytotoxicity effect

could be minimised by coating the PNPs with an inert polymer, which would

enable lysosomal enzyme digestion.

Gao and co - workers [54] proposed a theoretical model for understanding the

receptor mediated endocytosis for spherical and cylindrical PNPs based on

energetic analysis, which was originally developed by Freund and Lin. [90]

Their study suggested that, PNPs can enter the cell via wrapping even in

the absence of clathrin or caveolin coats and the shortest wrapping time is

observed for optimal particle size. They also deduced a threshold particle

radius of about 15 and 30 nm respectively for a cylindrical and spherical

particle below which endocytosis would never happen, which is in agreement

with experimental observations by Aoyama and co - workers. [91] Decuzzi and

Ferrari [92] modified the original formula proposed by Gao and co - workers

Chapter 1 12

[54] including the contribution of non - specific interactions.

In another study, it has been shown theoretically that, depending on particle

size and membrane - particle interaction (either attractive or repulsive),

curves towards larger particles or curves away from adsorption layers of

small particles. [93] A similar study showed that, attractive interaction and

aggregation of small PNPs on the cell membrane will decrease the minimal

size of particles whereas nonspecific repulsive interactions will increase the

minimal size for effective uptake. [94] Yuan and co - workers [95] presented

a thermodynamic model providing a phase diagram in which they elucidate

uptake efficiency with particle size and density of ligands expressed on the

particle surface. From their model, they predicted the most efficient uptake

for high ligand density for PNPs of radius 25 - 30 nm. Recently, Li [96]

developed a thermodynamic theoretical model to explain the size and shape

effect of cigarlike and spherical NPs on endocytosis which suggests a minimal

NP radius exists that would overcome the thermodynamic energy barrier for

endocytosis. More recently, Dobay and co - workers [97] proposed stochastic

pi calculus, a widely - used process algebra, to simulate PNP uptake and

intracellular distributions.

Gold nanoparticle uptake quantification methods are still in their infancy,

with destructive electron microscopy together with inductively coupled plasma

atomic emission (ICP - AEP) or inductively coupled plasma mass spectroscopy

(ICP - MS). These two methods estimate the number of particles in a sample

measuring the mass of particles per unit volume without the aid of any

labels, by relying only on the properties of the particles themselves. [98] The

drawback of ICP - AEP and ICP - MS is that they require sophisticated sample

preparation procedures that are strongly dependent on instrument calibration

and show as large scatter of data within one set of samples. [99] TEM analysis

could provide high resolution images (nm scale) visualising the inner structures

Chapter 1 13

but could not be used for live cell imaging as it destroys cells. [100]

Plasmon coupling between pairs of nanospheres, nanorods, nanodiscs, and

nanoshells has been used to detect DNA - DNA, [101–103] DNA - protein, [104]

and protein - protein binary interactions. [105] Numerous researchers [21–25]

have utilised a plasmon ruler on the basis of spectral shift due to the coupling

of two AuNPs. Reinhard’s [26] group attached EGFR antibody conjugated

PNPs to EGFR protein expressing cells and probe the membrane protein onto

the cell surface utilising plasmon coupling properties of aggregations at the cell

surfaces. In a similar study, Sokolov’s group [3], attached PNPs to receptors

and specifically looked at cell signalling and follow the receptor trafficking

inside the cell using plasmon coupling.

As a methodology, plasmon coupling provides a microscopic view of the

interactions in these applications, but no macroscopic view of the interaction

is available at the cellular size regime (~ 10 µm image scale). To address

this issue, we proposed a non - destructive microscopic optical method, image

correlation spectroscopy, together with plasmon coupling. This combined

technique can provide both microscopic and macroscopic images, elucidating

local interactions and providing an idea of what is happening at the cellular

level.

1.3 Gold nanoparticle preparation

In order to attach AuNPs to HeLa cells, AuNP preparation technique must be

discussed. There are two techniques available; wet chemical synthesis and nano

- fabrication. Wet chemically prepared AuNPs are suitable for attachment

to different cells. For simplicity of preparation, functionalisation and optical

properties, we will mostly focus on AuNSs and their attachment to HeLa cells.

Chapter 1 14

(A) (B)

Figure 1.1 (A) Transmission electron microscope (TEM) images of wet chemicallysynthesised gold nanospheres (AuNSs) (diameter 80 ± 6.5 nm) after drop - casting ontoa TEM grid, (B) corresponding histogram showing size distribution.

1.3.1 Wet chemical synthesis

Turkevich pioneered monodispersed AuNSs in 1950, and it is still regarded as

the simplest and most commonly used technique. [106, 107] It involves mixing

a dilute solution of chlorauic acid with a dilute sodium citrate solution, where

citrate acts as both a reducing agent and a capping agent. Turkevich’s method

is appropriate for synthesising 10 - 20 nm diameter nanoparticles. In contrast,

Frens’ method is appropriate for synthesizing larger nanoparticles (16 - 150 nm

diameters). [108, 109] To produce larger particles, less sodium citrate is added,

so there is a reduction of citrate ions available for stabilising the particle, which

causes the small particles to aggregate into bigger ones. Several approaches

have been introduced that refine Frens’ technique. [110, 111]

Spherical nanoparticles have thermodynamically favourable geometry and they

require only simple chemical reactions. However, synthesising non - spherical

particles involves complex techniques. Martin and co - workers first reported

NR template based growth, by electrochemically depositing a rod shaped

template in the pore of an alumina membrane. [112] The width of nanorods

Chapter 1 15

(A) (B)

Figure 1.2 (A) Transmission electron microscope (TEM) images of wet chemicallysynthesised gold nanorods (AuNRs) (aspect ratio 3.8) after drop - casting onto a TEMgrid, (B) corresponding histogram showing size distribution.

can be controlled by the pore diameter of the aluminum membrane template

(e.g. 5 - 200 nm) and length can be controlled by varying the amount of

gold deposited thus the aspect ratio can be tuned. Rods of materials such as,

copper, silver and gold can be prepared offering only monolayers of rods. A

transmission electron microscope (TEM) image of synthesised AuNSs of 80 nm

diameter using this technique is shown in Figure.1.1.

Electrochemical growth of nanorods was first explored by Wang and co -

workers. [113, 114] The synthesis was conducted on a two electrode electro-

chemical cell, where a gold metal plate anode and platinum plate cathode were

immersed in an electrolytic solution consisting of cetyltrimethylammonium

bromide (CTAB) surfactant and tetradodecylammonium bromide (TOAB)

co - surfactant. Gold ions in the form of AuBr4+ produced from the gold

metal anode, then form complexes with the CTAB micelle and migrate to

the cathode, where gold ions are reduced to gold atoms. Then, to control

the aspect ratio of the nanorods a silver plate is gradually inserted into the

solution by the concentration and release rate of silver ions is produced from

Chapter 1 16

C6H5O73−

Citrate Polyethylene glycol

C2nH4n+2On+1

Maleimide H N o o

C4H3NO2

Figure 1.3 Chemical structure of citrate, polyethylene glycol (PEG) and maleimide.(adopted from wikipedia)

the redox reaction between the gold ion and the silver plate. To facilitate the

mixing of CTAB and TOAB and to assist the formation of rod - like CTAB

micelles, acetone and cyclohexane are added before electrolysis respectively.

CTAB forms a bilayer structure on the longitudinal surface of gold nanorods.

[115]

The method was further improved by Nikoobakht and co - workers, [116]

and Jana and co - workers, [114, 117, 118] who found an aqueous growth

solution containing dilute concentration of chloauric acid, ascorbic acid, silver

nitrate and a surfactant cetlytributly ammonium bromide (CTAB) formed

rod shaped micelles. Ascorbic acid is a weak reducing agent that can only

reduce Au (III) to Au (I) in the presence of high concentration of CTAB

at low pH (2.5) producing a single crystal seed structure. [114, 117, 118]

CTAB preferentially binds to {110} facets of Au crystals producing anisotropic

nanorods until reduction of all Au (I) in the solution resulting in non - Au

spherical nanorods. Addition of AgNO3 facilitates Au deposition in the correct

lattice sites, which leads to a wider distribution of geometries. [114, 119] By

varying the chloauric acid and silver nitrate concentration, the longitudinal

surface plasmon resonance (LSPR) wavelength of nanorods can be tuned

between 600 and 850 nm. Further, introducing a binary surfactant mixture

composed of CTAB and benzyldimethyl - hexadecyl - ammonium chloride

(BDAC) nanorods with an aspect ratio upto ~ 10 (LSPR :- 1300 nm) can

be produced. A TEM image of AuNRs synthesised using this technique with

Chapter 1 17

average aspect ratio 3.8 is shown in Figure.1.2

However, other shaped PNPs such as nanostars, [120, 121] bipyramids, [121,

122] dumbbells, [123] nanoplates [122] and nanotubes [124] can be produced

by changing the reaction conditions. To produce dumbbell, penta-twinned

[125] and bipyramid shape particles are produced respectively adding acetone

to their growth solution, citrate stabilised penta - twinned seeds instead of

single crystalline CTAB, or using cetyltributal ammonium bromide (CTBAB)

instead of CTAB.

1.3.2 Gold nanoparticle surface modification

The presence of surfactants (e.g. citrate for AuNSs and CTAB for AuNRs),

enhances the stability of AuNPs under aggressive conditions such as centri-

fugation and high ionic strength. Unfortunately, the strong binding of the

molecules to the gold surface makes surface hydrophobation difficult. [126]

Additionally, they induce cytotoxicity both in vitro and in vivo. To reduce

the cytotoxicity effect of AuNPs due to surfactance, AuNP surfaces can be

modified with additional coating (e.g. PEG, PAA, PAH, maleimide). The

coating material should be stable and chemically and physically inert. They

should be easily deposited on AuNP surfaces and coating thickness should

be controllable. Also, they should be nontoxic, bio - compatible and easily

modifiable with functional groups for further conjugation purposes. [127]

For example, PEG coating of AuNSs provides enhanced colloidal stability,

biocompatibility (due to the highly hydrophilic nature of PEG) and offers

reduced non - specific adsorption of molecules to particle surfaces. The

chemical structures of citrate, polyethylene glycol (PEG) and maleimide are

shown in Figure.1.3

After functionalisation AuNPs can be functionalised with biomolecules (e.g.

Chapter 1 18

DNA, proteins, antigens) by electrostatic adsorption, surface coating with

charged polymers, biofunctional ligand attachment or ligand exchange. [128]

These surface modifications can change the dielectric properties of the external

surface resulting in a shift of plasmon peak. These also influence the AuNP

uptake kinetics and aggregation dynamics. In this study, the focus is limited

to surface modified PNPs (e.g. bare, PEG and maleimide coated AuNSs) and

their consequences on aggregation and uptake, while incubated in HeLa cells.

1.4 Aim and methodology of this thesis

In this section, the specific aims of this thesis, potential research problems and

proposed techniques to achieve the intended goals will be discussed.

1.4.1 Specific aim

The aim of this PhD was study cell uptake and aggregation of gold nano-

particles (AuNPs) using image correlation spectroscopy (ICS). To accomplish

the proposed aim we divided the task into two major sections:- (1) in a

dielectric environment and (2) in a cellular environment.

The specific aim in the dielectric medium section was validating the use

of ICS on plasmon coupled AuNPs, cross - checking the number of AuNP,

with scanning electron microscope (SEM) images. In order to apply ICS for

plasmon coupled AuNPs, the quantum yield (QY) of AuNPs was extracted

using a numerical simulation called the finite - difference time - domain method

(FDTD), which can provide the scattering cross - section for different sizes and

shapes of AuNPs at particular separations for a specific excitation wavelength.

Integrating the whole spectrum over the visible to near - infrared regions, we

extracted the scattering of QY of AuNPs. This number was then used in high

Chapter 1 19

order image correlation spectroscopy (H - ICS) simulations to extract the total

particle number and to investigate aggregation details of coupled AuNPs. The

H - ICS simulated number was then compared with the number extracted

from correlated SEM images for the same location to verify the accuracy of

our method. This number was also compared with polarisation and wavelength

spectroscopy results.

After successful demonstrations of H - ICS in experimental condition, (i.e.

dielectric samples), we extended our tool to a cellular environment to explore

PNP cell interaction at the 10 - 100 µm cellular size regime. We investigated

the following for AuNPs incubated in HeLa cell samples:

• PNP uptake kinetics as a function of size, surface charge and incubation

time.

• PNP aggregation dynamics as a function of size, surface charge and

incubation time.

• The effect of PNP size on uptake and oligomerisation.

• The effect of PNP surface charge (e.g. PEG, maleimide coating) on PNP

uptake and oligomerisation.

To explore PNP aggregation and uptake kinetics, we proposed image correl-

ation spectroscopy (ICS) in conjunction with plasmon coupling. These two

proposed techniques will be discussed in the following sections.

1.4.2 Methodologies

Plasmon coupling of nanoparticles: The optical properties of AuNPs have

fascinated many scientists since ancient times. More recently, the topic has

continued to interest researchers, beginning with the developments of classical

Chapter 1 20

electromagnetic theory. Gustav Mie [129] deduced Maxwell’s equation to

explain the strong absorption of AuNS while illuminated under plane wave,

offering a rigorous scientific foundation for understanding these phenomena.

Recently, it was discovered that collective electronic oscillations, known as

surface plasmons, give rise to strong optical properties which has created

a considerable amount of interest in the scientific community. A more

dramatic optical property is the change in colour due to size, shape and

aggregation. Generally when gold (Au) or silver (Ag) NPs begin to aggregate,

they show shifts in the plasmon resonance compared with isolated particles.

To understand this phenomenon, sufficient knowledge of the electromagnetic

properties of interacting metallic nanoparticles at close proximity is required.

One of the simplest descriptions of this interaction is that of two nearby dipoles.

To deduce the interaction of two interacting molecules the magnitude of dipole

moments and inter - particle distance need to be defined, which is given by

V∝ P1P2r3 where P 1 and P 2 are the magnitude of dipole moments and r is the

inter - particle distance. For the case of AuNSs, the lower - energy resonance

corresponds to two longitudinally aligned dipoles, giving rise to red shifts in

the optical spectrum, whereas for the higher energy resonance, the coupled

dipoles cancel each other out resulting in a zero dipole moment. While for

the case of elongated and other shape or disordered nanoparticles, the dipole

- dipole interaction and red shift of spectrum are also visible but offer more

complexity. This property has been introduced in detection of nucleotides,

proteins and cells which leads to a wide range of assays and tests for medical

applications. [3, 21, 22, 26]

Exploring the plasmon coupling of different size, shape and orientation AuNPs

using FDTD simulation we can quantify the QY, which will be used in high

order image correlation spectroscopy (H - ICS) for studying the feasibility of

H - ICS for weakly coupled AuNPs.

Chapter 1 21

Image correlation spectroscopy (ICS) of plasmon coupled nano-

particles: Image correlation spectroscopy (ICS) is a characterisation method

for emitting species in random configuration which can provide the average

number of particles within a focal volume and aggregations details within an

image, by correlating the image with itself or its squared or cubed images. It is

a well known technique in cell biology and widely used to measure the transport

properties and cell membrane protein distribution in living cells. [22–24] It

has been used to investigate the organisation of supramolecular complexes

but has not been used for macromolecular complexes ofcoupled plasmonic

nanoparticles. [11, 22–24] There are a few issues that need to be resolved

for application of ICS for plasmon coupled samples embedded in a dielectric

medium. The quantum yield of PNPs drastically changes due to its size, shape,

orientation and distribution. Image correlation spectroscopy intrinsically does

not account for the distribution of quantum yield and orientation anisotropy

of the material. Therefore, we need to know the emission quantum yield

variation of plasmon coupled AuNPs. We explored the plasmon coupling effect

of AuNS dimers, trimers using numerical simulation, FDTD. Integrating the

whole spectrum over a range of 400 -100 nm, we extracted the quantum yield

for the plasmon coupled particles. This value will be used as an input to

investigate the feasibility of H - ICS for plasmon coupled AuNPs dispersed

in a polyvinyl alcohol (PVA) matrix. We will extract the total number of

particles and dimers within an image for weakly coupled AuNPs embedded in

a PVA matrix or incubated in HeLa cells.

1.4.3 This thesis

The presentation of this work has been divided into chapters - each of which

focuses on a specific portion of our aims of investigating oligomerisation in a

dielectric medium and cellular environment. The current chapter has focuseds

Chapter 1 22

on reviewing the factors affecting PNP uptake and aggregation, existing

techniques, identifying potential research problems, and finally suggesting

methodologies to address any issue and achieve intended goals.

In Chapter 2, we review the theoretical concepts of surface plasmon resonance

(SPR) and plasmon coupling. We perform the numerical simulation FDTD to

investigate the plasmon coupling effect of single AuNPs and dimers of several

sizes and shapes. These simulation results enable calculation of quantum

yield (QY) values, which can be used as an input for H - ICS to explore

the oligomerisation of plasmon coupled AuNPs in plasmonic random medium

(PRM) and cell samples.

In Chapter 3, we explore the H - ICS theory and simulation results. We

discuss important issues that need to be resolved to apply H - ICS for plasmon

coupled AuNPs. We present the simulation results of H - ICS which provide the

boundary conditions regarding concentration of emitters, quantum yield and

noise level, thus inferring when we can and cannot apply these techniques. We

finish this chapter explaining; how to interpret the quantum yield of plasmon

coupled AuNPs for this technique.

In Chapter 4, the feasibility of H - ICS for plasmon coupled nanoparticles

will be examined. We perform confocal laser scattering microscopy of the

PVA embedded plasmonic nanoparticle samples and perform H - ICS for the

acquired confocal laser scattering microscopy (CLSM) images. We extract

the total number of plasmonic nanoparticles and clusters for the acquired

images performing H - ICS simulations, and comparing these numbers with

the reference number extracted from correlated SEM images. These results

reveal, how accurately our proposed H - ICS tool performs under experimental

conditions. This technique will be then extended into a cellular environment

to explore AuNPs uptake and oligomerisation .

Chapter 1 23

In Chapter 5, we perform dark - field scattering microscopy and acquire the

PNPs internalised cell images. H - ICS is performed on acquired images to

extract the total number of particles and dimers. The simulated H - ICS results

are used to explore PNPs uptake and oligomerisation kinetics as a function of

size, surface charge and incubation time for different surface modified PNP

incubated HeLa cell images.

In Chapter 6, a conclusion to the study will be presented, including a summary

of the major achievements along with suggestions for future extensions of this

work.

Chapter 1 24

Chapter 2

Theory and simulations of

surface plasmon resonance and

plasmon coupling

2.1 Abstract

In this chapter, we present the analytical theory for metallic nanoparticles,

specifically for spherical and ellipsoidal AuNPs. Existing analytical models

- Mie theory and Mie Gans theory can only provide the analytical solution

for spherical and ellipsoidal AuNPs, but not other shapes. Due to a lack of

analytical models it becomes necessary to use available numerical simulation

techniques to gain further insight into their near field properties, particularly

quantum yield of PNPs of different sizes and shapes. To address this issue,

FDTD simulation has been introduced for standalone and coupled AuNPs,

describing basic mathematical and physics formalism behind the FDTD

algorithm and simulation setup. In addition, excitation coupling theory will be

presented to describe dipole - dipole interactions between two molecules. The

25

Chapter 2 26

FDTD simulation results will be helpful for extracting the quantum yield (QY)

of coupled AuNPs, to determine the aggregation state of AuNPs dispersed in

a PVA matrix or cellular environment.

2.2 Introduction

In this chapter, we discussed the underlying theory of surface plasmon reson-

ance, plasmon coupling and the plasmon ruler equation. FDTD simulation

results of standalone particles (e.g nanospheres, nanorods, bipyramids and

dumbbells) and coupled particles (e.g. AuNS dimers, AuNS trimers) are

presented. This simulation result will be utlised in ICS simulations discussed

in chapter 3, to investigate the oligomerisation of AuNPs in dielectric and

cellular environments.

2.3 Theory of surface plasmon resonance

Plasmonic nanoparticles are mediated by surface plasmons by localised surface

plasmon resonance (SPR), which is a charge density oscillations confined

to metallic nanoparticles embedded in dielectric media. [130–135] When

metal nanoparticles are much smaller than the wavelength of incident light,

the electromagnetic field can induce a resonance of free electrons across

the particle, known as surface plasmon resonance (SPR) (Figure.2.1.A).

The resonance wavelength strongly depends on the size, shape, and surface

composition of the particle, and the dielectric properties of the surrounding

medium.

Surface plasmons can have multiple energy bands depending on their geometry.

For a metallic nanosphere there is single energy band (Figure.2.1.B) but for

Chapter 2 27

(A) (B)

Figure 2.1 (A) Schematics for plasmon oscillation for a sphere, (B)surface plasmonresonance (SPR) spectrum of 40 nm radius gold nanospheres calculated using Mie theory[129], refractive index 1.33.

nanorods there are multiple energy bands (Figure.2.2) due to band splitting

into low and high energy bands. The high energy or transverse absorption

band (short axis) corresponds to electron oscillation perpendicular to the major

axis, while the low energy or longitudinal absorption band (long axis) results

from the oscillation of the electrons along the major axis (Figure.2.2). By

varying the aspect ratio, the longitudinal absorption can be made tunable in

the near infrared region (NIR), making the gold nanorods suitable for in vivo

applications. In the next section, we will review the underlying theory and

concepts of SPR.

2.4 Theory of surface plasmon resonance of

metallic nanoparticles

The spectrum of metallic nanoparticles depends on the size, shape and

environment of the surrounding particles. [130–135] Mie theory can give the

exact analytical solution of a sphere with an arbitrary shape, given that the

dielectric constants of the environments are known. [129, 136, 137] Gans theory

can give the analytic expression to calculate the spectra approximately for

Chapter 2 28

Figure 2.2 Surface plasmon resonance (SPR) spectrum of nanorods with semi majoraxis 43 nm and semi minor axis 10 nm calculated using Mie - Gans theory, [129]refractive index 1.33. While gold nanospheres show one SPR band in the visibleregion, gold nanorods show two bands: a strong longitudinal band in the near infraredregion corresponding to electron oscillation along the long axis and a weak transverseband, similar to that of gold nanospheres in the visible region corresponding to electronoscillations along the short axis.

rod shaped particles much smaller than the wavelength of light. [138–140]

Numerical approaches such as discrete dipole approximation (DDA), [141]

finite - difference time - domain (FDTD), or finite element calculations can

be used for other shapes. In this section, we will introduce the underlying

theories that govern SPR in spherical or rod shaped nanoparticles.

2.4.1 Extinction of light by a nanosphere

In 1908 Gustav Mie proposed a theory for an arbitrary shaped spherical particle

to solve Maxwell’s equation with the correct boundary conditions. This

solution can be used to determine the absorption, scattering and extinction

cross - section of arbitrary shaped nanoparticles. The cross - section depends

on the radius of the nanosphere R, vacuum number of the incident light k

Chapter 2 29

and dielectric function of the nanosphere where εp and its surrounding εm, are

expressed as

σscat = 2πk2

∞∑n=1

(2n+ 1)(| an |2 + | bn |2) (2.1)

σext = 2πk2 Re(an + bn) (2.2)

The coefficient an and bn can be calculated from Eq2.3 and Eq2.4

an = mψn(mx)ψn(x)− ψn(mx)ψ′n(x)

mψ′n(mx)ξn(x)− ψn(mx)ξ′

n(x) (2.3)

bn = ψ′n(mx)ψn(x)−mψn(mx)ψ′

n(x)ψ′n(mx)ξn(x)−mψn(mx)ξ′

n(x) (2.4)

Here, x = (√εm ) kR is a size parameter where R is the radius of the sphere and

m =√

εpεm

, whereεp and εm are the dielectric functions of sphere and medium

respectively. [129, 137] Also, ψ and ξ are the Ricatti - Bessel cylindrical

function of order n and ψ and ξ prime indicate differentiation with respect

to argument x. Ricatti - Bessel cylindrical function, ξn can be written as

ξn(r) = r [Jn(r)− iyn(r)]

where, spherical - Bassel function Jn(r) and yn(r) are expressed as

Jn(r) =√

π2rJn+ 1

2(r) , yn(r) =

√π2rYn+ 1

2(r)

where, ψn(r) = rJn(r) = r ×√

π2rJn+ 1

2(r) ,

and ξn(r) = r [Jn(r)− iyn(r)] = r [√

π2rJn+ 1

2(r) -i

√π2rYn+ 1

2(r)]

Also, n is the summation index of partial waves, n = 1 corresponds to the

dipole, n = 2 quadrupole and n = 3 hexapole oscillation. Only the dipole and

Chapter 2 30

(A) (B)

(C)

Figure 2.3 (A) Experimental measurement from Johnson and Christy for the dielectricfunction of gold, showing real and imaginary parts of the dielectric constant. [142] (B)Variation of extinction cross - section spectra predicted by Mie theory [129] for 10nm radius gold nanospheres, immersed in a media with various refractive indices (C)Variation in the extinction cross - section spectra predicted by Mie theory [129] for goldnanospheres of various radii immersed in a media with a refractive index n = 1.33.

quadrupole terms are included in these calculations, as the higher order terms

are negligible for this size regime.

The summation of absorbed and scattered energy is known as extinction

energy. This phenomenon can be expressed as

σabs = σext − σscat (2.5)

Therefore, to calculate the absorption and scattering cross - section the

dielectric function of the particle is required. Drude and Lorentz’s free electron

models may be used to calculate the dielectric functions, where electrons

and ions were treated as simple harmonic oscillations, considering only the

contribution of free electrons. However, for the noble metals, we have to

consider the effect of interband transitions and free electrons. The interband

Chapter 2 31

transitions are 3d->4sp, 4d->5sp, and 5d->6sp and free electrons are 4s, 5s

and 6s for copper, silver and gold respectively. For alkali metals, such as

sodium and potassium which are unstable in air, the energy of the band gap

is large, so the interband transitions do not affect the dielectric functions in

the visible range. Therefore, the dielectric constant of metal is given by the

sum of contribution from the free electron (εf ), given by Drude’s free electron

model of metal and a contribution from the interband transitions (εib). This

yields the following expressions for the dielectric constant.

ε(ω) = εf (ω) + εib(ω) (2.6)

εf (ω) = 1−ω2p

(ω2 + iγb)(2.7)

εf (ω) = 1−ω2p

(ω2 + γ2b )

+ iω2pγb

ω(ω2 + γ2b )

(2.8)

ε(ω) = 1−ω2p

ω2 + iγb+ εib(ω) (2.9)

where ωp is the plasma frequency and γb is the bulk damping constant related

to the mean free path of the electron by γb = vF/ l, where vF is the Fermi

velocity. The Fermi velocity and damping constant for bulk Au are vF =

1.4×106 ms−1 and γb = (15fs)−1. The plasma frequency is given by ωp =√ne2

ε0me(ωp for Au 13.8×1015), where n is the electron density, ε0 is the vacuum

permittivity and me is the effective mass of the electron.

Plasmon dephasing mechanisms

For small particle radius (less than 5 nm, that is dimensions less than the bulk

mean free path), the electrons collide with the surface of the particle and the

dielectric constant deviates from bulk metal broadening the line width. For

large particles, radiation damping causes broadening. [130, 143] Considering

Chapter 2 32

the electron surface scattering for small particles the damping constant is given

by

(leff ) = γb + (AvF )/leff (2.10)

where, leff is the effective path length of electrons (the average distance they

travel before scattering off a surface) and A is a constant that depends on the

electron - surface interactions. [144, 145] The effective path length depends

on the size and shape of the particles and can be calculated consistently for

arbitrary shaped particles by leff = 4VS

, where V is the volume and S is the

surface area of the particle. [144, 146] For spherically capped cylinders yields

leff = w(1− w3l), where w is the width and l is the total length ( l

wis the aspect

ratio). For spheres leff = 43R where R is the radius of the sphere and A = 0.33.

In the case of large particles with a radius above 50 nm, if the radiation

damping effect is considered, the line width is given by

Γ = γb + AvF/leff = 2hκv (2.11)

where V is the volume and k is a constant that characterises the efficiency of

radiation damping.

For noble metal nanoparticles, radiative decay occurs via transformation of the

particle plasmon into photons and nonradiative decay occurs via excitation of

the electron hole pair either within the conduction band (intraband excitation)

or between the d band and the conduction band (interband excitation). In

particular, it is advantageous to minimise the nonradiative decay (e.g. to

avoid sample heating or quenching of fluorescence from adsorbed molecules).

In other words, it is useful to maximise the quantum efficiency for resonance

light scattering. Sonnicsen and co - workers [31] found drastic reduction of

Chapter 2 33

the plasmon dephasing rate in nanorods (dephasing time 6 fs) compared

with small nanospheres (dephasing time 1.4 fs) due to suppression of the

decay into interband excitation. In contrast, the dephasing rate of nanorods

decreases dramatically with increasing red shifts, i.e.; for higher aspect ratios

such as 4.1 AuNRs, the dephasing time is lowest and reported as 18 fs. The

reduced nonradiative decay in nanorods is explained by the fact that interband

excitation requires a threshold energy of about 1.8 eV in Au, thus the process

of plasmon decay into such interband excitations is precluded for plasmons

with energy below 1.8 eV (above 688 nm). However, the increase of radiation

damping apparently outweighs this effect by far, resulting in much larger line

widths in spheres. In the near infrared region NRs have much narrower line

widths for a given resonance frequency. This is because of reduced radiation

damping and is simply a volume effect. For a sphere, the red shift arises from

retardation effects, which are only significant for large sizes. In contrast, for

nanorods the red shift arises from the shape and not the overall size of particles.

Radiation damping and electron surface scattering are negligible for NR with

widths 15 - 20 nm. However, increasing the width beyond 20 nm leads to

significant broadening. Experimental measurements of dielectric function for

Au on metal films under high vacuum by Johnson and Christy are most reliable

and widely accepted. [142] Their values are plotted in Figure.2.3.A and will

be used in the simulations performed in this thesis considering the interband

transitions and surface scattering for smaller particles and radiation damping

effect for larger particles.

2.4.2 Extinction of light by a nanorod

An advance of chemical synthesis, electron and ion beam lithography tech-

niques has allowed us to prepare anisotropic geometries from rods [116] to

bipyramids, [125, 147] dumbbells, [148] and stars [120, 121]. In 1912, Gans

Chapter 2 34

developed an extension of Mies work to calculate the absorption and scattering

cross - section of anisotropic particles, where particles much smaller than

the wavelength of incident light are approximated as ideal dipoles. The field

outside the particles is superposition of ideal dipole and incident field. If an

elongated particle is treated as ellipsoid then its polarisability in an incident

field is parallel to one of its principal axes which can be expressed as

αp = 4πabc (ε1 − εm)(3εm + 3Lp(ε1 − εm) (2.12)

where Lp is a depolarisation factor

Lp =∫∞

0ds

(s+c)2f(s) = [√

(s+ a2)(s+ b2)(s+ c2)]

For prolate spheroids, if light is polarised along the long axis of the spheroids

Lp now changes to

L1 = 1−e2

e2 (−1+12 ln

1+e1−e), L2 = L3=1−L1

2

Where e is the eccentricity of the particle given by,

e2 = 1 − ( ba)2 and Lp= (L1,L2,L3) denotes the polarisation of the incoming

field, aligned to one of the axes of the particle.

For oblate spheroids the Lp depolarisation factor is expressed as

L1 = L2 = g(e)2e2 [π2 − tan

−1g(e)]g2

2

g(e) = 1−e2

e2 , e2 = 1− c2

a2 ,L3 = 1− 2L1

The optical cross - section can be expressed as

σext = kIm(α) (2.13)

Chapter 2 35

σscat = k4

6π | α |2 (2.14)

Gans theory suggests that, anisotropic particles will split into two resonant

modes, the longitudinal mode at a higher wavelength along the long axis

and a transverse mode at lower wavelength along the short axis due to

independent polarisability. The aspect ratio of the particle determines the

position of the longitudinal peak, and has no impact on the transverse

peak which is close to the resonance wavelength of a sphere. In the

Mie - Gans theory elongated nanoparticles are approximated as spheroids;

however, TEM studies revealed that wet chemically synthesized nanorods have

hemispherically capped geometry. To account for this issue, Prescott and co -

workers [140] approximated nanorods as prolate spheroids, spherically capped

cylinders, an ellipsoidally capped cylinders and cylinders to simulate the

extinction spectra using the direct dipole approximation (DDA) method and

modifying the geometrical factor (L). Considering the above approximation,

the AuNR absorption, scattering and extinction spectra are presented in

Figure.2.2. These calculations are for a nanorod 43 nm long and 10 nm wide,

immersed in water of refractive index 1.33 (Figure.2.2). The resultant SPR

wavelength of AuNRs is far from the interband transition, giving narrower

and symmetric peak shape than for nanospheres. The absorption of nanorods

is proportional to volume and scattering, and follows V 2 relationship. Hence,

due to its dipole nature, the nanorod shows cosine square dependency on the

angle between nanorod long axis and incident direction, but the nanosphere

does not show any polarisation dependency.

Chapter 2 36

2.5 Basics mathematical and physical formal-

ism behind finite difference time domain

(FDTD) technique

Recently, different size and shape PNPs including prisms, [149] shells, [150]

cubes, [151] bipyramids, [146] and nanorods [131–135] have been synthesised

and studied optically. This exhibits promising optical properties tuning their

plasmon mode across whole visible to near infrared ranges which have been

used in different biological applications.

To study the plasmon coupling effect and employ that into ICS to investigate

the oligomerisation of PNPs into dielectric medium and cellular environments,

we are required to extrapolate the quantum yield of PNPs. Existing analytical

models - Mie theory and Mie Gans theory can provide the analytical solution

of PNP’s and infinite long cylinders as special cases [152], but for many shapes

analytical solutions are not available.

The lack of an analytical model has set a barrier to understand their plasmon

spectra, as well as to gain further insight into their near field properties

particularly, the quantum yield of PNPs. Several state of the art methods

have been developed to solve Maxwell’s equation numerically, including the

T - matrix, [153, 154] discrete dipole approximation (DDA), [155] and finite

difference time domain (FDTD). [156, 157] The T - matrix method emphasises

more far field properties (i.e. the scattered field) and it is a better developed

systems for revolution symmetry. The DDA and FDTD, however, can give

both near and far field properties due to their finite element nature. We used

finite difference time domain (FDTD) to explore the quantum yield of different

size and shape NS, NR, bipyramid and dumbbell structures and their coupling

effects.

Chapter 2 37

Finite difference time domain method is a state of the art method for solving

Maxwell’s equations in complex geometries. It can provide a direct solution

performing Fourier transformations in time and space offering a unique insight

into electromagnetic and photonics problems (e.g. complex Poynting vector

and transmission or reflection of light). The basic mathematical and physics

formalism behind the FDTD algorithm is that, it solves Maxwells curl equation

2.15 in non - magnetic materials.∂D

∂t= ∇× H (2.15)

¯D(ω) = ε0εr(ω)E(ω)

∂H∂t

= − 1µ0∇× E

where, H,E and D are the magnetic, electric and displacement fields respect-

ively. While εr(ω) is the complex relative dielectric constant.

In three dimensions, Maxwell’s equation has six electromagnetic components:

Ex, Ey, Ez and Hx, Hy and Hz. Considering the structure infinite in z

dimensions and fields independent of z we can write,

εr(ω, x, y, z) = εr(ω, x, y, z) and

∂E∂z

= ∂H∂z

=0

In this situation, Maxwell’s equation is split into two independent equations

composed of three vector quantities each of which can be solved in x - y

plane only. These are termed the TE (transverse electric) and TM (transverse

magnetic) equations where TE has Ex, Ey, and Ez components and TM has

Hx, Hy and Hz components.

For example in the case of TM, Maxwell’s equations can be reduced to:

∂Dz

∂t=∂Hy

∂x− ∂Hx

∂y, where Dz(ω) = ε0εr(ω)Ez(ω)

∂Hx

∂t= − 1

µ0∂Ez

∂y

Chapter 2 38

Figure 2.4 Standard Yee - cell. The electric field components are located on the edgeswhile the magnetic field components are located on the face centres. Figure taken fromWikipedia. [156]

∂Hy

∂t= − 1

µ0∂Ez

∂x

Therefore, the FDTD method solves these equations on a discrete spatial and

temporal grid. Each field component is solved at a slightly different location

within the grid cell, which is known as Yee cell [158] (as shown in Figure.2.4).

In our study, finite difference time domain (FDTD) solution has been used to

simulate the nanostructures of wavelength and subwavelength scale geometries

in the presence of incident illumination. To explore the optical property

of PNPs such as nanospheres(NSs), nanorods (NRs), dumbbells (DBs) and

bipyramids (BPs), single particle response and plasmon coupling property, we

employed commercial FDTD software (Lumerical Solutions 7.5). Simulation

methodology can vary whether one is interested in exploring the light scattering

from standalone particle, particles on substrates, or from a surface with

nanoscale structures. However, in this experimental case, our major focus was

Chapter 2 39

Figure 2.5 Layout editor of finite difference time domain (FDTD) simulation forplasmonic nanoparticles (PNPs) structure. The yellow rectangular box is the total field,the white rectangular box is the total - field scattered field source and outer yellowrectangular box is the scattered field. The pink arrow shows the direction of propagation,k vector. The blue dot represents the direction of the electric field

AuNS dimers and its consequences when incubated with HeLa cells. However,

in this section, we explore the plasmon coupling effect, of different size and

shape AuNPs for further investigation.

2.5.1 Simulation setup

The layout editor shows the simulation of objects, total field, scattered field and

polarisation direction. The AuNSs are simulated in the middle, a position that

can be moved easily with a mouse. Surrounding that, the yellow rectangular

box is the total field, followed by the total - field scattered field source (white

rectangular box) and scattered field (outer yellow rectangular box). The pink

arrow shows the direction of propagation, k vector. The blue dot represents

the direction of the electric field (Figure.2.5).

To simulate the AuNPs a total - field scattered field source (TFSF) was

used which surrounds the AuNPs. Two analysis groups one in the total field

Chapter 2 40

region and one in the scattered field region were used. These analysis groups

provide absorption and scattering cross - sections and angular distribution of

scattered radiation. To calculate the electric field an enhancement frequency

profile monitor can be included. Johnson and Christy’s dielectric function for

AuNPs was used considering the size effect (surface scattering and radiation

damping). [142] The mesh override region intentionally was kept large enough

to accurately resolve the location of AuNS interfaces, especially for a curved

surface and for TFSF sources which works best in uniform meshed regions. In

addition, sources require a certain amount of space (~2 mesh cells) to inject

the fields, because the fields are not physically meaningful within this region.

Hence, the monitor should not be placed in this restricted region. The rate at

which energy is removed from the incident plane wave hence, the net power

flow into the particle is considered as an absorption cross - section which can

be calculated by the analysis group located inside the TFSF source using

optical theorem. On the contrary, the net power scattered from the particle

hence, the scattering cross - section can be calculated by another analysis

group located outside the TFSF source. This group measures the net power

scattered from the particle. Afterwards, the absorption or scattering cross -

section can be calculated considering the geometrical area π×r2 and the size

parameter 2×π× rλ. For higher accuracy, simulation mesh refinement was set to

“conformal variant 1” to achieve sub - cell resolution followed by convergence

testing. Mesh override mesh size was set to 0.25 - 0.5 nm considering the

size of the PNPs structure. Simulation span was set to 2 µ m to avoid the

interaction of evanescent tails of the resonant surface plasmon modes with

perfectly matched layer (PML boundary) conditions. To reduce the light

reflections by PML layers, more PML layers were considered. Setting the

X min boundary condition to symmetric and Z min boundary condition to

anti - symmetric simulation memory and time were improved by a factor of

4 (Figure.2.6), shows the better agreement between FDTD and theoretical

Chapter 2 41

(A) (B)

Figure 2.6 Extinction, absorption and scattering cross - sections calculated via Mietheory [129] compared with finite difference time domain (FDTD) simulations for, (A)10 nm radius gold nanospheres and (B) 20 nm radius gold nanospheres.

results (Mie theory) considering higher accuracy simulation for 10 and 20 nm

radius AuNSs for a refractive index of water of 1.33.

2.6 Finite difference time domain (FDTD)

simulations of standalone particles

In this section, we will discuss a standalone particle’s response (e.g. AuNS

or AuNR) and coupling effect of dimer (e.g. AuNS and AuNR dimer) and

tetramer (e.g. AuNS tetramer) structure. We will finish the section by

presenting the quantum yield variation due to tip curvature.

2.6.1 Finite difference time domain (FDTD) simula-

tions of gold nanospheres and nanorods

A.Nanospheres

We performed finite difference time domain (FDTD) simulation (Lumerical

Solutions 7.5, Canada) to investigate the absorption and scattering cross -

Chapter 2 42

(A) (B)

Figure 2.7 (A) Scattering cross - sections of 5 - 100 nm radius gold nanospherescalculated via Mie theory compared with finite difference time domain (FDTD)simulations, (B) longitudinal surface plasmon resonance (LSPR) of 5 - 100 nm radiusgold nanospheres calculated via Mie theory compared with finite difference time domain(FDTD) simulations.

section of AuNSs with 5 - 100 nm radii. To simulate the AuNSs, a total - field

scattered field source (TFSF) was used and grid resolution 0.5 was chosen

nm for better accuracy of results (Figure.2.6 and 2.7). The FDTD simulation

reproduced the analytical results very well, both on line shapes and cross -

section at the plasmon resonance. The FDTD plasmon resonance of 10 nm

radius AuNSs (2.38 eV) agreeds very well with quasistatic approximation (Mie

theory) (2.39 eV). However, plasmon redshifts and broadens as the AuNP

grows larger due to the phase retardation effect and increasing contribution

from higher order modes. The peaks are located at 2.35 eV for 25 nm radius

AuNSs and 2.29 eV for 50 nm radius AuNSs. In the quasistatic approximation

absorption cross - section is proportional to the volume of particles, Qabs =Cabsπr2 . The relation holds roughly for smaller particles such as 10 and 25 nm

particles. However, retardation effect becomes so significant for the 50 nm

radius AuNPs that, its absorption efficiency at resonance is lower than 25 nm

radius AuNSs.

The absorption, scattering cross - section (Figure.2.7A) and In Figure2.7B

LSPR of 5 - 100 nm radius AuNSs were compared. The result shows that,

Chapter 2 43

Nanoparticles LSPR (nm) Scat. CSProlate spheroid 731 1.11×10−13

Spherically capped cylinder 769 1.26×10−13

Cylinder 878 1.66×10−13

Table 2.1 Comparison of longitudinal surface plasmon resonance (LSPR) andscattering cross - sections (Scat.CS) of different particle morphologies for aspect ratio(AR) 2, length 40 nm PNPs, calculated via FDTD simulations.

for the largest particles (above 25 nm radius) the absorption and scattering

intensities deviate from simple scaling laws of d3and d6 respectively. The

scattering peak increases with increase of NP size and dramatically increases

above 60 nm radius.

B.Prolate spheroid

FDTD calculations were performed using FDTD software (Lumerical Solutions

7.5, Canada). The Dielectric function of gold was formulated from Johnson and

Christy and corrected for size effect (surface scattering and radiation damping)

[142]. A TFSF source with its wavelength ranging from 100 - 400 nm was used

and a grid resolution of 0.5 nm was chosen for better accuracy of results. The

source direction was set along the axis of NPs. The surrounding medium was

taken as water with a refractive index of 1.33. An FDTD solution of gold

prolate spheroid (ellipsoid) with minor axis 20 nm and aspect ratio (AR) 2.1

shows LSPR at 1.696 eV (731 nm) and scattering cross - section 1.11×10−13

for refractive index 1.33 and 0.5 nm mesh size (Figure.2.9).

C.Rod shaped simulations (spherically capped)

For FDTD simulations the nanorod is modelled as a finite cylinder with both

ends capped by hemispheres (Figure.2.10B). The aspect ratio of a nanorod is

defined as the ratios between its total length and diameter. The simulations

are performed for two aspect ratios 3.8, and 4.3 for 75 nm and 40 nm major

Chapter 2 44

Figure 2.8 Schematics of different particle morphologies under consideration including,(A) a prolate spheroid, (B) a spherically capped cylinder, (C) an ellipsoidally cappedcylinder and (D) a cylinder.

axis length, and given resonance energies are 1.50 eV (827 nm) and 1.46 eV

(851 nm) respectively (Figures.2.16 and 2.17). The polarisation was set along

the major axis, mesh size was taken as 0.5 nm and refractive index as 1.33.

2.6.2 Quantum yield variation due to tip curvature

Elongated nanoparticles, especially nanorods and bipyramids are promising

for optical studies as their spectra are easily tunable by varying the aspect

ratio (Figure.2.10). Compared with nanospheres, nanorod structures of the

same volume give a larger curvature at the tips. Therefore, to investigate

the significant field enhancement due to tip curvature, we performed FDTD

simulation for NS, DB and BP. Gold nanospheres of 80 nm diameter gave

a longitudinal plasmon resonance at at 2.21 eV (560 nm). An example of

simulation for a AuBP with R (radius at equator) = 15 nm, h (total length)

= 162 nm and r (radius at the poles) = 10 nm is shown in Figure.2.10C.

Chapter 2 45

Figure 2.9 Scattering cross - sections of different morphology nanoparticles for aspectratio (AR) 2, length 40 nm refractive index 1.33 and mesh size 1 nm using FDTDsimulations.

(A)

(B)

(C)

Figure 2.10 Finite difference time domain (FDTD) simulations of gold, (A) dumbbell,(B) nanorods and (C) bipyramids.

Chapter 2 46

Dog - bone - like AuNRs (dumbbells) were also modelled as a cylinder at

the middle waist and two larger spheres at both ends, considering a total

length of 84 nm and sphere radius of 15 nm (Figure.2.10A). The FDTD

calculations were performed with 0.5 nm grid resolution and gave a longitudinal

plasmon resonance at 1.39 eV (890 nm) and 1.54 eV (805 nm) for BP and DB

respectively.

Therefore, we calculated the scattering spectrum of the above mentioned

diameter and normalised per unit volume, which shows the following sequence

NS<DB<NR<BP (Figure.2.11). A significant field enhancement is observed

as AuNRs and bipyramids show around one to three orders of magnitude higher

in cross - section than nanospheres (Figure.2.11).

2.7 Finite difference time domain (FDTD)

simulations of coupled nanoparticles

In this section we will discuss the dipolar excitation coupling model and

plasmon ruler equation. We conclude the section by presenting FDTD

simulations results of AuNS dimers, AuNR dimers, AuNS trimers and

heterodimer AuNRs.

2.7.1 Dipolar excitation coupling model and plasmon

ruler equation

When two metallic nanoparticles approach each other their plasmonic near

field couple strongly gives rise to a distance dependent wavelength shift of the

plasmon mode. Quantitative study on near field coupling between pairs of

elliptical metal particles and spheroidal nanoparticles as a function of inter

Chapter 2 47

Figure 2.11 Scattering cross - sections (per unit volume) of spheres (SPs), dumbbells(DBs), nanorods (NRs) and bipyramids (BPs) using FDTD calculations.

- particle separation was independently reported by Su [44] and Rechberger

[48] respectively. The plasmon coupling effect can be understood using the

plasmon hybridisation model, where coupled mode is treated as bonding and

antibonding of individual particle plasmon mode. [159] The plasmon coupling

of nanorod was reported by Prashant and co - workers in 2006 [160] and

Funston and co - workers in 2009. [161] The previous study [160, 161] reported

that, for the side by side geometry when polarisation is parallel to the inter -

particle axis, with a decrease of inter - particle distance longitudinal plasmon

mode blue shifted. For end to end geometry when polarisation is parallel to

the inter - particle axis, with decrease of inter - particle distance longitudinal

plasmon mode red shifted. Plasmon coupling in nanoparticle dimers obeys the

exponential red shift, according to the universal plasmon ruler equation at long

nanoparticle distances, but the model is not valid at very short nanoparticle

distances. [162] Most of the fundamental studies of plasmon coupling have

focused on the interactions between two identical metal NPs in a homodimer

where dipole - dipole plasmon mode is active. Heterodimers of metal PNPs

Chapter 2 48

exhibit much more complex plasmon coupling behaviour than homodimers

because of the symmetry breaking of heterodimers [163] which leads to the

formation of new plasmon modes (bright and dark mode) giving rise to

Fano resonance in asymmetric plasmonic nanostructures. Recently, plasmon

coupling for rod - sphere pairs, [164] trimers [165, 166] and nanoparticle cluster

arrays [167] (n<7) has also been reported.

Previous studies [25, 44, 45] have shown that plasmon coupling between two

identical metal NPs can be described using an empirical universal scaling law

using Eq.2.16

∆λλ

= Aexp[−dD

τ] (2.16)

where A is the maximum fractional plasma resonance shift, τ is the decay

constant, d is the inter - particle separation, D is the diameter of sphere/rod

length. The decay constant τ has been found to be within 0.2−0.3, irrespective

of the metal type, nanoparticle shape and the surrounding medium. Funston

and co - workers [161] showed that the fitting parameter is not good enough

for predicting the coupled plasmon wavelengths of various AuNR dimers in

Eq.2.16. Fore more universal equation, Jain and co - workers [25] reported

that, coupled plasmon energy is determined by competition between the inter

- particle coulombic restoring forces on the displaced electron cloud. The

d/D term in Eq.2.16 was replaced by Jiang fang [163] group as ( V gapV nanorod

) 13 ,

where Vgap and Vnanorod denotes the volume of the gap region and nanorod

respectively. The exponential expression can be expressed∆λλ

= Aexp[−( dD

) 1c

τ] (2.17)

where c = 3, 5, 7 are for dipole, quadrupole and octupole respectively. Using

the results of previous studies, we can define two regimes of plasmon coupling,

strong and weak coupling. When particle separation is greater than 0.1 D,

Chapter 2 49

where plasmon ruler governs, we call it weak coupling and when particle

separation is less than 0.1 D, where the plasmon mode hybridisation occurs

(band splitting), we call it strong coupling (Figure.2.13). In the weak

coupling regime, plasmon coupling obeys an exponential relationship with the

separation, known as the universal plasmon ruler equation. In the strong

coupling regime, hybridization of the plasmon energies occurs similar to

molecular bonding and anti - bonding orbitals.

The dipole – dipole interactions between two interacting molecules are

generally described in the framework of excitation coupling theory. [168, 169].

As predicted by excitation theory, excited – state levels of monomer split

into two levels e.g. lower energy level and a higher energy level relative

to the monomer excited state, upon dimerisation. In that process it forms

two possible arrangements of the transition dipoles of the dimer, e.g. in

- phase or symmetric and out of phase or antisymmetric (see Figure.2.12).

The interaction energy 2U between the molecules could be approximated by

coulombic interaction between the transition dipole moments of the monomers,

the angle and the distance between the transition dipoles 1 and 2. [170]

U = 14πε0

| µ |2

n2R3 ξ (2.18)

where ξ = cosθ12 − 3cosθ1Rcosθ2R is the orientation factor, nmis the refractive

index of the medium, | µ |2 is the squared modulus of the transition dipole

moment and R is the distance between dipole centres.

For parallel or H - type dimers (Figure.2.12.A), the interaction between two

transition dipoles of dimer is repulsive, because of the cancellation of the two

dipole moments, transition to the lower energy excited state is forbidden,

therefore it contains a single band at higher energy with respect to the

monomer (because θ12 =00,θ1R = θ1R =900 and so ξH = 1). For head to

Chapter 2 50

Monomer MonomerDimer

Monomer Dimer Monomer

𝛙𝛙+ = 𝟏𝟏

√𝟐𝟐 [𝛗𝛗𝟏𝟏 + 𝛗𝛗𝟐𝟐]

𝛙𝛙− = 𝟏𝟏

√𝟐𝟐 [𝛗𝛗𝟏𝟏 − 𝛗𝛗𝟐𝟐]

𝛙𝛙− = 𝟏𝟏

√𝟐𝟐[𝛗𝛗𝟏𝟏 − 𝛗𝛗𝟐𝟐]

𝛙𝛙+ = 𝟏𝟏

√𝟐𝟐 [𝛗𝛗𝟏𝟏 + 𝛗𝛗𝟐𝟐]

2U

2U

Excited State

Ground State

Excited State

Ground State

(A)

(B)

(C)

π*

σ*

σ

π

+ + --

H aggregate

J aggregate

++

---

-

--- ++

+

+

+-

+

Figure 2.12 Schematic of the energy level splitting resulting from the dipolarinteraction of AuNR dimer, showing symmetric (ψ+) and anti - symmetric coupling(ψ−) of excitons for (A) H aggregate geometry and (B) J aggregate geometry. (C)Exciton theory picture of the nature of the coupled longitudinal plasmon excitation inAuNRs dimers: electromagnetic analogy to molecular orbital theory. [160]

Chapter 2 51

+ -

- +

+ -

+ -

+ -

+ -

+ -

+ -

+ -

+ -

(A) (B)

(C)

Sep d > 0.1 D Sep d < 0.1 D

Figure 2.13 Plasmon coupling of nanoparticle at (A) weak and (B) strong couplingregime, (C) the energy level splitting resulting from the dipolar coupling of dimers,showing symmetric and anti - symmetric coupling for AuNS dimer. [171]

tail or J – type dimers(Figure.2.12.B), the interaction between two transition

dipoles of the dimer is attractive, and because of the two dipole moments,

transition to a higher energy state is forbidden, and therefore the spectrum

shows a single band at lower energy with respect to the monomer (because

θ12 = θ1R =00 and so ξJ = −2).

However, the excitation coupling theory could be used to elucidate the optical

spectra of AuNR dimers. For side - by - side AuNR arrangements (J aggregate)

when polarisation direction is along the inter - particle axis, transverse

polarisation, leads to a red – shift and end - by - end AuNR arrangements

(H aggregate), leads to a too small blue – shift. That’s because, transverse

plasmon dipoles are far apart even when the rods touch each other. As a

consequence, the optical properties of the AuNR dimers and their dependence

on dimer orientation appear to be qualitatively steady with the excitation -

coupling model.

Chapter 2 52

\

4x10-7 5x10-7 6x10-7 7x10-7

0

1x10-15

2x10-15

3x10-15

4x10-15

Scat

terin

g cr

oss-

sect

ion

(m2 )

Wavelength (m)

1 4 8 16 24 40 80 Monomer

0.0 0.5 1.0 1.5 2.0 2.50.00

0.05

0.10

0.15

0.20

0.25

a) FDTD b) Modified PR c) Universal PR

b) A = 1.09107, t = 0.19005c) A = 0.30115, t = 0.09033

Sep/Dia(40nm)

(A) (B)

Figure 2.14 Finite difference time domain (FDTD) simulation results of 20 nm radiusgold nanosphere dimer plasmon coupling. (A) scattering spectrum and (B) plasmonresonance peak shift as a function of inter - particle separations.

Similar reasoning could be used to describe the polarisation dependence of the

optical resonance shift in AuNS dimers. [48] In AuNSs, polarisation along the

inter - particle axis (p - pol) produces red - shift with respect to the single

AuNS resonance and polarisation along perpendicular (s - pol) axis produces

a blue - shift (Figure.2.14 and 2.15).

Recently, nanoparticle plasmon coupling has been introduced to investigate

inter - particle distance between two micromolecules utlising the plasmon ruler

concept, on the basis of spectral shift due to coupling of two AuNSs. [21–

25] This tool has been used for probing membrane proteins on cell surface

receptors, [26] following receptor trafficking [3] and detecting aggregation of

PNPs inside the cell through receptor trafficking.

Therefore, plasmon coupling provides a microscopic view of the interaction

in these applications, but there is no macroscopic view of the interaction at

the cellular size regime. In our study, we proposed to use plasmon coupling

techniques together with image correlation spectroscopy, to quantify the AuNP

uptake and oligomerisation when AuNPs were incubated in HeLa cells. By

exploring the plasmon coupling of different size, shape, and orientation AuNPs

using FDTD simulation we can quantify the QY, which will be used in H - ICS

Chapter 2 53

4.0x10-7 6.0x10-7 8.0x10-7 1.0x10-60

2x10-14

4x10-14

6x10-14

8x10-14

1x10-13

1x10-13

Wavelength (m)

Scat

terin

g cr

oss-

sect

ion

(m2 )

0.5 1 4 8 16 24 40 80 200 Monomer

0.0 0.5 1.0 1.5 2.0 2.50.0

0.1

0.2

0.3

0.4

0.5

a) FDTD b) Modified PR c) Universal PR

Sep/Dia(80nm)

b) A = 1.13252, t = 0.22513c) A = 0.42349, t = 0.09639

(A) (B)

Figure 2.15 Finite difference time domain (FDTD) simulation results of 40 nm radiusgold nanosphere dimer plasmon coupling. (A) Scattering spectrum and (B) Plasmonresonance peak shift as a function of inter - particle separations.

for studying the feasibility of the H - ICS tool, for weakly coupled AuNPs.

To acquire an understanding of quantum yield (QY) and introduce it for

image correlation spectroscopy (ICS), we need to know the plasmon coupling

effect of PNPs. In this study, we explored the plasmon coupling of AuNSs

and AuNRs dimer and investigated the QY for coupled AuNSs and AuNRs.

Hence, in experimental observation, for simplicity, we dealt only with the gold

sphere. We investigated the plasmon coupling of two symmetric dimer using

FDTD simulation. Integrating the FDTD simulated scattering intensity of

coupled gold nanosphere, we extracted the values of QY (cluster by monomer

scattering strength ratio) and introduced that number for H - ICS simulation

for determining oligomerisation.

2.7.2 Numerical simulation of gold nanosphere dimer

To explore plasmon coupling commercial FDTD software (Lumerical Solu-

tions7.5, Canada) was used. The dielectric function of gold was formulated

from Johnson and Christy and corrected for size effect (surface scattering and

radiation damping). [142] A total - field scattered field source (TFSF) with its

Chapter 2 54

wavelength ranging from 100 - 400 nm was used and a grid resolution of 0.5

nm was choosen for better accuracy of results. The source direction was set

along the axis of AuNPs. The surrounding medium was taken as water with

refractive index 1.33. We investigated the plasmon coupling of 40 nm and 80

nm diameter gold plasmonic nanoparticle dimer using FDTD simulation and

found red - shift in surface plasmon resonance (SPR) peak with decreasing

inter - particle distance (weak coupling regime), in line with previous results.

[25, 44, 45, 161] The red shift is due to coupling of plasmonic near field when

two spheres approache each other and form bonding and antibonding plasmon

modes. Additionally, the amount of red - shift is modelled with a modified

universal plasmon ruler equation, A = 1.09107, τ = 0.19005 with A = 1.13252,

τ = 0.22513 for 40 nm and 80 nm diameter AuNSs respectively (Figure.2.14

and 2.15). This value is comparable to published results. [161, 163]

We found that there are multiple peaks at strong coupling regimes for 1 nm

and 0.5 nm separations for 40 nm and 80 nm diameter NSs in Figure.2.14

and Figure.2.15 due to hybridisation of plasmon energy. At weak coupling

regimes (as an example 80 nm and 200 nm separation respectively), the

peak is almost the same position as the monomer peak (around 541 nm and

558 nm respectively for 40 nm and 80 nm diameter AuNS) and intensity

becomes double than monomer. Intensity of monomers at 541 nm and 558

nm peak wavelengths for 40 and 80 nm diameter AuNS are 2.724×10−14 m2

and 1.656×10−14 m2 respectively. Intensity of 1 nm dimer for 40 and 80 nm

diameter AuNSs at 666 nm and 785 nm peak wavelengths are 8.434×10−14

m2 and 4.66×10−15 m2 respectively. At monomer peak wavelength 558 nm

the intensity of the dimer (8.41×10−14 m2) becomes five times greater than

monomer (1.66×10−14 m2) intensity. As an example for 80 nm diameter

AuNSs, at 700 nm and 715 nm wavelength intensity of the dimer becomes

3.5 (7.70×10−14 m2) and 2 (5.16×10−14 m2) times greater than monomer

(1.66×10−14 m2) intensity.

Chapter 2 55

6x10-7 7x10-7 8x10-7 9x10-7 1x10-6 1x10-6 1x10-60

1x10-14

2x10-14

3x10-14

4x10-14

5x10-14

Sc

atte

ring

cros

s-se

ctio

n (m

2 )

Wavelength (m)

1 nm 2 nm 4 nm 10 nm 48 nm 87.5 nm 128 nm Monomer

0.0 0.5 1.0 1.50.00

0.05

0.10

0.15

Seperation/NR Length

(b) A = 0.47171, t = 0.23292(c) A = 0.18369, t = 0.08976

FDTD (b) Universal PR (c) Modified PR

(A) (B)

Figure 2.16 Finite difference time domain (FDTD) simulation results of gold nanoroddimer plasmon coupling. (A) scattering spectrum of rod for length 75 nm, width 20nm, aspect ratio (AR) 3.8 and (B) plasmon resonance peak shift as a function of inter- particle separations.

2.7.3 Numerical simulation of gold nanorod dimer

To explore the plasmon coupling of AuNR we used the same simulation setup

as discussed in section 2.5.1. A total - field scattered field source (TFSF)

of 400 - 1000 nm with its wavelength was set along the long axis of AuNRs.

Therefore, we investigated the plasmon coupling of spherically capped nanorod

dimer for nanorod length 75 nm width 20 nm and aspect ratio 3.8 and nanorod

length 43 nm width 10 nm and aspect ratio 4.3 for side - side geometry with

polarisation parallel to the long axis.

The calculated scattering spectra corresponding to longitudinal plasmon

excitation of a pair of AuNRs, appraching each other along their long axis

(i.e., end - to - end assembly) for aspect ratio (AR) 3.8 and 4.3 are shown

in Figures.2.16 and 2.17 respectively. The calculated longitudinal bands of

isolated single AuNR are shown for comparison as a black curve in Figures.2.16

and 2.17. The optical response maximum does not seem to shift its position

from that of the isolated AuNR case for larger inter- AuNR distance (D

= 117 nm and D = 128 nm of Figures.2.16 and 2.17 ). As the distance

Chapter 2 56

8x10-7 9x10-7 1x10-6 1x10-6

0

5x10-16

1x10-15

2x10-15

2x10-15

Scat

terin

g cr

oss-

sect

ion

(m2 )

Wavelength (m)

1.2 nm 2.6 nm 6.3 nm 28.5 nm 56.5 nm 117 nm Monomer

0.0 0.5 1.0 1.5 2.0 2.50.00

0.05

0.10

0.15

(b) A = 0.18083,t = 0.11068(c) A = 0.50052, t= 0.24458

FDTD (b) Universal PR (c) Modified PR

Seperation/NR Length

(A) (B)

Figure 2.17 Finite difference time domain (FDTD) simulation results of gold nanoroddimer plasmon coupling. (A) scattering spectrum of rod for length 43 nm, width 10nm, aspect ratio (AR) 4.3 and (B) plasmon resonance peak shift as a function of inter- particle separations.

decreases, the longitudinal plasmon band progressively red - shits due to

coupling of the longitudinal plasmons. The amount of red - shift is modelled

with modified universal plasmon ruler equation, with A = 0.18396, τ =

0.08976 and A = 0.50052, τ = 0.24458 for aspect ratio (AR) 3.8 and 4.3

respectively (Figures.2.16 and 2.17). This value is comparable to published

results. [161, 163]

To elucidate the nature of plasmon excitation in the coupled system, an

excitation model could be employed. [160] For the side - by – side dimer

arrangement longitudinal plasmon bonding in nature is analogous to the

formation of σv bond from two Pz orbitals. This produces maximum electric

field in the junction between the interacting AuNR. On the other hand

for the side - by - side dimer the coupled longitudinal plasmon has anti -

bonding in nature analogous to the formation of a π* bond from Px/y orbitals,

concentrating electric field on either side of the inter - particle junction. For

the case of end - to – end configuration, a new band emerges at higher energies

as the inter - AuNR distance becomes very small or the number of AuNRs

interacting in an assembly increases. [160]

Chapter 2 57

4x10-7 5x10-7 6x10-7 7x10-7 8x10-7 9x10-7

0

5x10-16

1x10-15

2x10-15

Monomer Dimer Sep 140 nm Trimer Sep 110 nm Trimer Sep 140 nm

Scat

terin

g Cr

oss-

sect

ion

(m2 )

Wavelength (m)

0 1 2 30.00

0.08

0.16

0.24

0.32

Sep/Dia

b) A=0.21836,t=0.192c) A=0.3123,t=0.46134

(a) FDTD (b) Plasmon Ruler (c) Modified Plasmon Ruler

(A) (B)

Figure 2.18 Finite difference time domain (FDTD) simulation results of gold nano -sphere trimer plasmon coupling. (A) scattering spectrum of a sphere - trimer with adiameter of 40 nm for different inter - particle distances for weakly coupling regimescompared with monomer and dimer and (B) plasmon resonance peak shift as a functionof inter - particle separations.

2.7.4 Numerical simulation of gold nanosphere trimer

We investigated the plasmon coupling of a nanosphere trimer with diameter 40

nm, as a function of inter - particle separations (Figure.2.18). We compared

the monomer and dimer with the same diameter with the trimer and found

that, peak position remains the same for weakly coupled trimers, dimers

and monomers but intensity becomes double and triple for dimer and trimer

respectively due to formation of bonding and antibonding plasmon modes. The

intensity and peak are 1.89×10−16 m2 and 543 nm, 6.01×10−16 m2 and 543 nm,

1.01×10−15 m2 and 543 nm for monomers, dimers and trimers respectively. We

used the same simulation setup as discussed in section 2.5.1. The simulation

results are shown in Figure.2.18.

2.7.5 Numerical simulation of gold heterodimer nanorod

The calculated longitudinal plasmon spectrum using FDTD simulation for

a dimer with two spherically capped nanorods with dissimilar aspect ratios

(spherically capped nanorod 1: length 75 nm width 20 nm aspect ratio 3.8

Chapter 2 58

6x10-7 8x10-7 1x10-6 1x10-6 1x10-60

3x10-15

6x10-15

9x10-15

1x10-14

2x10-14

2x10-14

Scat

terin

g cr

oss-

sect

ion

(m2 )

Wavelength (m)

1 nm 3 nm 13 nm 25 nm 55 nm 135 nm AuNR AR 3.8 AuNR AR 4.3

0 40 80 120 160880

920

960

1000

1040

1080

1120

(a) Peak Wavelength(b) Fitting Curve

Peak

wav

elen

gth

(nm

)

Seperation (nm)

(b) y = y0 + A exp (-x/t)yo = 895.4748, A = 194.345, t = 11.1824

(A) (B)

Figure 2.19 Finite difference time domain (FDTD) simulation results of gold nanorodhetero dimer plasmon coupling. (A) Scattering of hetero dimer (spherically capped) rodhaving AR 4.3, length 43 nm, width 10 nm and AR 3.8, length 75 nm, width 20 nmand (B) Au nanorod dimers peak wavelengths are shown for different inter - particledistances.

and nanorod 2 : length 43 nm width 10 nm aspect ratio 4.3) for side -by -

side geometry, polarisation parallel to the long axis is shown in Figure.2.19.

As seen in Figure.2.19, the longitudinal plasmon maximum and intensity is

around 695 nm and scattering cross - sections is 2.43×10−15 m2 nm for an

aspect ratio 3.8 and around 852 nm and 4.78×10−16 m2 for an isolated AuNR

of aspect ratio 4.3.

For a side - by – side assembly of same length rods, the anti – symmetric

coupling mode with lower energy would have a zero dipole moment, therefore

absorption would not be observed optically. However, for the case of

dissimilar rod length, red shifted absorption would be observed, due to

the lower net dipole moment, even though it possesses a lower spectral

intensity. Correspondingly, for an end - to – end assembly, the higher energy

component becomes allowed analogous to the anti – symmetric coupling due

to symmetry breaking, even though it has a relatively lower spectral intensity

(Figure.2.19.A).

Therefore, heterodimers of AuNRs exhibit much more complex plasmon

coupling behaviour than homodimers because of the symmetry breaking of

Chapter 2 59

heterodimers that leads to the formation of bright and dark plasmon modes

giving rise to Fano resonance. A single peak was observed, when asymmetric

nanorod dimers were weakly coupled, and multiple peaks were observed when

asymmetric nanorod dimers were strongly coupled due to band splitting

(Figure.2.19A). At weakly coupled regimes, an asymmetric nanorod dimer

peak is almost at the same position as monomer peak. We also found that,

with a decrease in separation between two asymmetric AuNR dimers plasmon

resonance red shifted and intensity increaseds exponentially (Figure.2.19B).

The quantum yield for asymmetric dimer could be also explored and the

quantum yield number could be used in H - ICS simulations for investigating

aggregation details due to asymmetric AuNS or AuNR dimer.

2.8 Quantum yield of gold nanoparticles

We investigated the plasmon coupling effect of different sized nanospheres and

nanorods using FDTD simulation as discussed in the previous section. From

the FDTD simulation it was observed that, for 80 nm diameter Au monomers,

dimers (1 nm separation) and trimer (1 nm separation) peak and intensity were

558 nm and 1.65571×10−14 m2, (1 nm separation) 7.85 nm and 8.43×10−14

m2 and 8.77 nm and 1.42×10−13 m2 respectively. For aspect ratio 3.8 nm, and

length 75 nm AuNR peak and intensity for monomers and dimers was 827 nm

and 2.02×10−14 m2 and 965 nm and 5.08×10−14 m2 respectively. Integrating

the total spectrum, we extracted the dimer to monomer ratio for 40 nm and

80 nm diameter AuNSs and 3.8 and 4.3 aspect ratio nanorods. The ratio of

dimer to monomer varies from 2 ~ 4, and ratio of the trimer to monomer varies

from 4 ~ 9 for different separations. Figure.2.20 shows the scattering quantum

yield (QY) ratio of clusters to monomers of AuNSs (80 nm diameter).

Chapter 2 60

Figure 2.20 Scattering quantum yield (QY) ratio of clusters and monomers of goldnanospheres (AuNS ) (80 nm diameter)

2.9 Conclusion

The use of the FDTD technique enabled us to extract the QY (ratio of clusters

to monomers), which will be required to perform H - ICS for plasmon coupled

PNPs. In the next chapter, we will discuss how we interpreted QY for plasmon

coupled AuNPs to perform the H - ICS simulations.

Chapter 3

Theory and simulations of

image correlation spectroscopy

3.1 Abstract

In this chapter we present the theory and simulation results of high order image

correlation spectroscopy (H - ICS) and the concept of interpreting the H - ICS

technique for coupled AuNPs, including how to determine the experimental

boundary conditions (e.g. dielectric medium or cellular environments) for

investigating aggregation dynamics.

3.2 Introduction

Image correlation spectroscopy (ICS) is a well - known technique in cell

biology that has been used for investigating the organisation of supramolecular

complexes (at a sub - micron scale), but has not been used for macromolecular

complexes (at nanometre scales) for coupled PNPs. Plasmon coupling is

61

Chapter 3 62

infact one of the methods used to look at 10-100 nm size regime utilising

PNPs as a ruler to measure the distance between two interacting molecules.

However, variation in scattering quantum yield (QY) becomes problematic

when introducing ICS in this size regime. These issues need to be resolved

before using ICS for AuNPs.

In this chapter, we will discuss the theory of image correlation spectroscopy

(ICS), high order image correlation spectroscopy (H -ICS) and boundary

conditions to perform H - ICS simulations. We will finish the chapter by

discussing the concept of interpreting H - ICS for coupled AuNPs utlising QY

of coupled AuNP dimer.

3.3 Image correlation spectroscopy (ICS)

Image correlation spectroscopy is a characterisation method for emitting

species in random configurations. [25] In recent decades, it has been

widely used to measure the transport properties and cell membrane protein

distribution of living cells. [22–24] The main use of ICS is to obtain information

of emitting species from confocal laser scanning microscopy (CLSM) or dark -

field scattering images by correlating the image with itself (Figure.3.1).

If the random intensity variable i, is a function of two independent variables x

and y, it is possible to define a corresponding two - dimensional autocorrelation

function

g(ζ, η) =< δi(x, y)δi(x+ ζ, y + η) >=1

NM

∑Nk=1

∑Mk=1 i(k, l)i(k + ζ, l + η)

1NM

∑Nk=1

∑Mk=1 i(k, l)

− 1,

(3.1)

Now the variance of the random function is equal to the value of the correlation

function in the limit where both ζ and η vanish. Thus we conclude that the

Chapter 3 63

⊗ =

Figure 3.1 Autocorrelation of an image.

density of the fluorescent particles can be measured by a magnitude of g (0,

0) that is

g(0, 0) = limζ→0limη→0g(ζ, η) = 1< N >

(3.2)

where, g(0, 0) is the autocorrelation function and < N > is the average number

of particles in a focal volume. The peak of the autocorrelation function g(0, 0)

indicates the average expected number of particles in a focal volume and

aggregation details of emitter species in an entire image.

In order to use ICS, the following conditions need to be understood. Image

correlation spectroscopy intrinsically does not account for the interaction

of emitters such as plasmon coupling. It also does not account for heavy

distribution in quantum yield, such as is observed in scattering images of

AuNPs. In addition, ICS does not account for orientation of anisotropic

material. Generally, when the quantum yield (QY) of the emitter is fixed,

and there is no interaction between the emitters, ICS is very accurate.

But since the QY of coupled plasmonic particles greatly depends on their

degree of aggregations, the simulated results of ICS on coupled samples

cannot be validated. Nevertheless, for polydispersed samples, higher order

autocorrelation analysis needed to be addressed as it cannot resolve the number

Chapter 3 64

density together with quantum yield, with only one correlation function. The

approach has been applied for studies of molecular aggregation, [172, 173]

analysis of ion channel kinetics, [174] image recognition, [175] and no -

equilibrium thermodynamics. [176, 177] High - order ICS has been introduced

for quantitative measurement of the number densities of different cluster sizes

present in multicomponent samples, [178, 179] but has not been introduced

for plasmonic nanoparticles. In the next section we will discuss about the

conceptual background of H - ICS and how to interpret it for plasmon coupled

particles.

3.4 High order image correlation spectroscopy

(H - ICS)

The methodology of H - ICS is analogous to ICS, except it is the autocorrela-

tion of squared or cubed intensity of the scattering images. [180] The beauty

of this technique is that, the peak value of the high order autocorrelation

function is highly dependent on the aggregates and therefore is able to provide

information about aggregates, such as dimer, concentration [181] (Figure.3.2).

Through already available analytical expressions of peak values, H - ICS

allows us to extract the concentration of emitters of different species and their

emitting quantum yield ratio simultaneously.

When a focused laser beam is scanning across the sample, the intensity

fluctuation at each pixel can be expressed as

δi(x, y) = i(x, y)− < i > (3.3)

Where i (x, y) is the intensity of the emitter measured at the pixel located at

Chapter 3 65

Figure 3.2 High order image correlation spectroscopy (H - ICS) correlation functions.Autocorrelation of squared and cubed images and corresponding high order correlationfunction.

(x, y), and < i > is the average intensity of the entire image. The spatial high

order autocorrelation function, Gm,n(0, 0), is defined as

Gm,n(ζ, η),= < δi(x, y)m >< δi(x+ ζ, y + η)n > − < δi >m< δi >n

< i >m+n (3.4)

where (m,n)εN are positive integer, (m ≤ n).

Considering three different species of aggregate present in the system, peak

values of the high order mode can be expressed following the equation

suggested by Palmer and co - workers [178]

G1,1(0, 0) =B2,

G1,2(0, 0) =4B33 ,

G2,2(0, 0) = 2B4 + 2B22 ,

G1,3(0, 0) = 2B4 + 3B22 ,

Chapter 3 66

4.0x102 6.0x102 8.0x102 1.0x1030

3x10-14

6x10-14

9x10-14

1x10-13

Strongly coupled dimer

Weakly coupled dimer

σ sca

t (m

2 )

Wavelength (m)

Two monomers

0.1 11

2

3

4

5

6

Dim

er σ

scat

/ M

onom

er σ

scat

Sep / Diameter

(A) (B)

d

d

Figure 3.3 Finite difference time domain simulations and quantum yield of 80 nmdiameter gold nanosphere (AuNSs), (A) scattering cross - sections of 80 nm diameterAuNS at different separation and (B) quantum yield with respect to separation/diameter.

G2,3(0, 0) = 16B55 + 12B2B3,

G3,3(0, 0) = 16B63 + 30B2B4 + 15B3

2+15B23 ,

where,

Bk =∑Ri=1 α

ki < Ni >

[∑Ri=1 α

ki < Ni >]k

(3.5)

where, α2 is the emitting quantum yield ratio of the aggregate to the monomer

< N1 > is the concentration of the monomer and

<N2 > is the concentration of the dimer.

Theoretically, we can extract information of samples containing emitting

particles of infinite species by putting the values of m and n with 1, 2,

3....∞. Therefore, if more species are included in one image, simultaneous

equations of higher order will have to be solved to extract information about

the samples. For simplicity, in the following we will only deal with samples

with two populations. In this case, only the first three higher order normalised

moments need to be considered. The real solutions of the above equations

represent the population densities of the two species, N1 andN2, as well as their

Chapter 3 67

quantum yield ratio α2. Likewise, by solving the six higher order normalised

moments of the above equations we can solve for three emitting species.

In order to eliminate the contribution from the noise, the noise corrected spatial

high order autocorrelation was derived as follows

Gm.n(0, 0) |NS= Gm,n(0, 0) |Image< iimage >m+n −Gm,n(0, 0) |Noise< iNoise >

m+n

(< iImage>− < iNoise >)m+n

(3.6)

Here, main idea of the noise correction equation is to subtract the noise signal

from the original measured signals Gm,n (0,0) of the acquired images. [182]

3.4.1 Interpreting high order image correlation spectro-

scopy (H - ICS) for plasmon coupled particles

Quantum yield of AuNPs drastically varies for plasmon coupling. By

simulating PNP plasmon coupling we can extract the quantum yield ratio of

the aggregate to the monomer to interpret H - ICS for plasmon - coupled NPs

(Figure.3.3). Whenever two particles are brought close together the plasmon

resonance red shifts. If they are brought much closer together multiple peaks

will form due to higher order mode and band splitting.

From the FDTD simulation (Figure.3.3A), the total scattering strength of two

interacting AuNSs can be calculated integrating the entire spectrum regime

(400 - 1000 nm wavelength). By observing the total scattering strength with

respect to separation distance, it can be concluded that, whenever two particles

are further apart, there is no plasmon coupling, and total scattering strength

becomes doubled as they act as identical monomers (Figure.3.3B). From there

we can safely assume for 80 nm diameter AuNS dimers at 10% separation

that the quantum yield ratio of the aggregate to the monomer is ~ 4, which

Chapter 3 68

Monomer

Dimer

Figure 3.4 Simulated confocal laser scattering microscopy images (CLSM) containingmonomer and dimer mixture.

is the square of the dimer. Quantum yield is the same for 40 nm and 100

nm diameter AuNSs. Conversely, QY for AuNS trimers can be computed as

9, which is the cube of the trimer. Hence, using a QY value of 4 for dimers

simultaneous Eq.3.4 can be solved, for the first three higher orders normalised

moment, to extract the values of monomer and dimer concentration and QY.

3.5 Factors affecting precision of image correl-

ation spectroscopy (ICS)

Several factors are responsible for the accuracy of ICS. Quantum yield of

PNPs might change according to size, shape, orientation, material and coupling

effect, which subsequently affects the precision of ICS. For randomly oriented

rod, such as EBL fabricated randomly distributed PNPs, precision is at

its maximum, but this is reduced for oriented rods, such as electron beam

Chapter 3 69

microscopy (EBL) fabricated rods. Spatial distribution is also responsible

for the accuracy of ICS. Precision is highest for fixed quantum yield random

position samples, and decreases for fixed quantum yield fixed position samples.

Point spread functions of PNPs also vary from monomers to clusters which

may affect the precision of ICS. Several kinds of noise (e.g. background,

mechanical, electrical or shot noise) also impact on the accuracy of ICS. For our

study, we dealt with wet chemically synthesised AuNPs which are completely

random and determined the boundary condition simulating a similar condition

as the experimental condition. Later, we report on H - ICS simulations that

determined the concentration of monomers, dimers and QY for different noise

levels to determine the accuracy of the tool (Figures.3.6 and 3.7).

3.6 High order image correlation spectroscopy

(H - ICS) simulations

In order to validate the accuracy of H - ICS on two populations of plasmonic

random media analysis, we have simulated confocal laser scanning images of

samples containing monomers and dimers (α = 4) of gold nanospheres of 80

nm diameter (Figure.3.4). In the simulation, (Figure.3.4) the two species

had different intensities: one higher, lower. The higher intensity emitter was

designated as dimer with concentration per beam area N2. The lower intensity

emitters were considered to be monomer with concentration per beam area N1.

We spincoated a dilute solution of AuNS onto a co - ordinate marked EBL

fabricated grid (details of grid fabrication are illustrated in Section 4.4.2) with

the concentration adjusted to produce isolated particles and dimers. The

scattering spectrum produces a convolution of the focal spot and AuNSs in

the sample, which acts as a point source. Exploiting the convolution between

Chapter 3 70

1100 nm

1320 nm

FWHM = 465 nm

Figure 3.5 Typical confocal laser scattering point spread function profile from singleparticles (blue) and dimer (black). Red colour spectrum indicate analytical point spreadfunction using Vectorial Debye theory for objective 1.4 NA, at 715 nm wavelength forcircular polarisation. Debye theory [183] can be used to calculate the diffraction patternof an objective of high numerical aperture.

Gaussian beam and a point scatterer, Gaussian spot profile of the focusing

objective can be deduced. Figure.3.5, shows the cross - section of one of the

spot of scattered intensity, collected at the photomultiplier tube (PMT), when

the laser is scanned across the sample, producing a focal spot profile. The

full width half maxima (FWHM) and airy disk diameter of monomers were

465 nm and 1100 nm respectively, which is close to the theoretically expected

results for 1.4 NA focusing objective for circular polarisation. Similarly, airy

disk diameter of dimer is found at 1320 nm (1.2 times that of monomers).

Therefore, the e - radius of dimer was considered to be 1.2 times that of

monomer for H - ICS simulations presented in this thesis.

An FDTD simulation was conducted to explore the effect of plasmon coupling

(Figure.3.3). Integrating the whole scattering spectrum over the visible to near

Chapter 3 71

- infrared wavelength region, we estimated the intensity of dimers to be four

times that of monomers. In the electrostatic limit, the scattering intensity

increases with volume squared. Therefore, alpha can be estimated to be 4 for

dimers of gold nanospheres of small (80 nm) diameter. Applying a similar

hypothesis, intensity of trimer was estimated to be 9 times greater than of

monomers and so on for more than three emitter species or higher orders (as

illustrated in section 2.6).

We performed H - ICS of the simulated confocal laser scanning images and

extracted the input and output N1, N2 and α of 500 simulated CLSM images

(Figure.3.6). The result shows that, input and output monomer, dimer and

quantum yield follows a similar trend for certain monomer concentrations

(between 0.1 and 1 of input N1 concentration), but deviates for other

concentrations. The results show that the background noise correction is vital

in producing accurate results. In order to eliminate the contribution of noise,

we derived the noise corrected spatial high order autocorrelation function as

discussed in Eq.3.6 where, we subtracted the noise signal from the original

measured signals Gm,n (0,0). Therefore, we conducted H - ICS simulations to

extract the outputs N1, N2 and α.

Figure.3.8A shows the dark - field scattering images for AuNS incubated HeLa

cells sample. The green square in the acquired images shows the location

of noise Figure.3.8B, and inset Figure.3.8C shows recorded noise images in

the H - ICS simulations, which were subtracted from the measured signal for

better accuracy. Cell images with noise correction were recorded in H - ICS

simulations (Figure.3.8D).

In addition, the results of H - ICS simulations are shown with and without

noise correction for images with signal to noise ratio = 30 (Figures. 3.6 and

Chapter 3 72

(A) (B)

(C)

Figure 3.6 High - order image correlation spectroscopy (H - ICS) simulation results.The plots show the (A) N1, (B) N2 and (C) Alpha of the simulated sample withoutbackground noise. Each data point is averaged by analysis of 500 images.

3.7). H - ICS analysis shows good agreement with both input N1 and α when

input N1 is larger than input N2 , i.e.; when the density of monomers is higher

than that of dimers. Output N2 is also accurate when it is close to input N1.

Furthermore, output N1, N2 and α are more accurate for input concentration

N1>0.01. This result provides a better understanding of which concentration

we can and cannot use in H - ICS simulations. For the best performance of H

- ICS, it is important to operate under the aforementioned limits.

High - order ICS can provide average monomer concentration, average dimer

concentration and quantum yield. To calculate the total number of particles,

total number of oligomers (especially dimer number) and percentage of

oligomers, the following formula was used:

• Total number of AuNPs per beam area = average number of monomers

per beam area, N1 × 1 + average number of oligomers (e.g. dimer) per

Chapter 3 73

(A) (B)

(C)

Figure 3.7 High order image correlation spectroscopy (H - ICS) simulation results.The plots show the (A) N1, (B)N2 and (C) Alpha of the simulated sample with signalto noise ratio (SNR) = 3 0, and the e - radius of the diameters is 1.2 times of that ofthe monomers. Each data point is averaged by analysis of 500 images.

(A) (A)

(B)

(C)

(D)

Figure 3.8 (A) Dark - field scattering images for AuNS incubated human cervicalcarcinoma (HeLa) cell samples, (B) selected noise location for high - order imagecorrelation spectroscopy simulations, (C) recorded noise images in high - order imagecorrelation spectroscopy (H - ICS) simulations and (D) recorded AuNS attached HeLacell images with noise correction in high order image correlation spectroscopy (H - ICS)simulations.

Chapter 3 74

beam area, N2 ×√α.

• Total number of AuNPs in that total imaging area = (total number of

AuNPs per beam area × total imaging area (pixel base)) / (π (e - radius

)2 (pixel base))

• % Oligomers = (N2) ×√α) / (total number of AuNPs per beam area)

× 100%

• % Error = (|H - ICS dimer number| - |SEM dimer number|) / |SEM

dimer number| × 100%

After, acquiring the total number of monomers and dimers for the simulated

images, these can be compared with the real number of particles extracted from

correlated SEM or TEM images to determine the accuracy of the simulations.

3.7 Conclusion

In this chapter, we discussed the underlying theory of H - ICS and concept

of interpreting H - ICS for coupled AuNPs utlising QY of AuNP dimer as

discussed in Chapter 2. From the simulation results, we can conclude that,

high - order ICS can be used for plasmon - coupled AuNP dimers. However,

the accuracy of these tools under experimental conditions, is investigated in

Chapter 4, wherein AuNPs are embedded in a PVA matrix.

Chapter 4

Image correlation spectroscopy

of plasmon coupled gold

nanoparticles into dielectric

medium

4.1 Abstract

In this chapter we present a feasibility study of Image correlation spectroscopy

of weakly coupled gold nanoparticles to understand aggregation dynamics of

plasmonic nanoparticles embedded in a dielectric medium.

4.2 Introduction

Recently, plasmonic NPs have been utlised to investigate inter - particle

distances between two macromolecules. Plasmon coupling between pairs of

75

Chapter 4 76

nanospheres (NSs), nanorods (NRs), nanodiscs and nanoshells has been used

to detect the DNA - DNA, [101–103] DNA - protein, [104] and protein - protein

binary interactions. [105] Numerous research groups [21–25] have utilised the

plasmon ruler on the basis of spectral shift due to the coupling of two AuNPs.

The technique of measuring and monitoring the dynamic distance between

biological macromolecules on a nanoscale regime has proven useful. More

recently, plasmonic NPs have been considered a promising tool for probing

membrane proteins on cell surfaces. Reinhard and co - workers [26] attached

anti - EGFR antibody conjugated PNPs to EGFR protein expressing cells and

probed the membrane protein onto cell surfaces utilising the plasmon coupling

properties of aggregations at cell surfaces. A similar study was carried out by

Sokolov and co - workers [3], in which they attached PNPs to receptors and

identified cell signalling by looking at receptor trafficking inside the cell using

plasmon coupling. They were able to detect aggregations of PNPs inside the

cell through receptor trafficking.

Over the past few decades different microscopy based techniques such as

fluorescence resonance energy transfer (FRET), [184–187] image correlation

microscopy (ICM), [188, 189] fluorescence correlation spectroscopy (FCS),

[89, 172–175, 177–179, 190–199] and image correlation spectroscopy (ICS)

[198, 199] have been used to investigate molecular activities at sub-microscopic

resolution without destroying cells. FRET is limited to detecting two closely

separated (<5nm) molecules of different types. [184–187] ICM has been used

to characterise larger protein assemblies, but is limited to sub - microscopic

level and is highly sensitive to background interference. Furthermore, FRET

and ICM are critically limited by photo bleaching. Fluorescence correlation

microscopy (FCM) has been used for describing molecular events, molecular

activities and measuring transport properties of cell macromolecules on a fast

time scale (from microseconds to seconds), but it becomes more problematic

for measurements of slower protein transport. In order to overcome these

Chapter 4 77

difficulties, ICS was developed, to measure transport properties and cell

membrane protein distribution in living cells, but this still can not provide

aggregation details at the nanometre scale. [198, 199] For poly - dispersed

samples, higher order autocorrelation analysis has been applied to studies of

molecular aggregation, [172, 173] analysis of ion channel kinetics, [174] image

recognition, [175], and no - equilibrium thermodynamics. [176, 177] Higher -

order ICS has been used for quantitative measurements of number densities

of different cluster sizes present in multicomponent samples, [178, 179] but

has not been introduced for plasmonic NPs. In these circumstances, plasmon

coupling together with image correlation spectroscopy could be a promising

tool for sensitive detection of cell biology in the 10- 100 µm regime. ICS is a

well - established technique that can provide information regarding molecular

aggregation within focal volume, but is yet to be applied to plasmonic NPs

for characterisation of aggregation at the nanometre scale regime. Here, we

proposed the use of image correlation spectroscopy together with nanoparticle

plasmon coupling to investigate PNP aggregation dynamics, when PNPs are

embedded in a PVA matrix (e.g. dielectric medium).

4.3 Nanoparticle plasmon coupling and simu-

lations

We explored the plasmon coupling of 80 nm diameter AuNP dimers using

commercial FDTD software (Lumerical Solutions 7.5, Canada) (Figure.3.3A).

The red - shift in surface plasmon resonance (SPR) was found to peak with

decreasing inter - particle distance (weak coupling regime), and the amount

of red - shift is modelled with a modified universal plasmon ruler equation,

(Eq.2.17 with A = 1.13252, τ= 0.22513). This value is comparable to published

Chapter 4 78

Sep

(A) (B)

Figure 4.1 (A) Transmission electron microscope (TEM) images of gold nanoparticles(AuNSs) dropcasted onto a TEM grid, and (B) dimer separation histogram, showing75% of dimers are within 10% of separation of diameter.

results [25, 161] as discussed in section 2.5.1.

From the FDTD simulation (Figure.3.3A), we found that, there are multiple

peaks at strong coupling regimes due to hybridisation of plasmon energy,

and at weak coupling regimes (for example, 200 nm separation) the peak

is almost in the same position as the monomer peak (around 558 nm) and

intensity is doubled (Figure.3.3A). Therefore, total scattering strength of two

interacting AuNSs can be calculated integrating the whole spectrum. By

observing the total scattering strength with respect to separation distance it

can be concluded that, whenever two particles are at further distances, there

is no plasmon coupling, and total scattering strength is doubled as they act as

identical monomers (Figure.3.3B).

We determined the inter - particle separations for around 150 dimer particles

from TEM images, where 80 nm diameter AuNSs were randomly oriented onto

a TEM grid and the number of particles of specific separations was plotted as

a function of separation (Figure.4.1). From the histogram of PNP separations,

extracted from TEM images we found that, 75% of dimers are separated

Chapter 4 79

by less than 10% of their diameter. At separation of 10% of the diameter,

scattering strength reaches upto 4, from the initial value (Figure.3.3B), which

is actually squared (four times) for dimers. This is in line with electrostatic

approximation, where σscat∼| α2 |2∼ V 2. For 80 nm or smaller particles, it is

therefore valid to use V 2 as the QY for dimer. Similarly QY for trimers will

be cubed (nine times). These results are helpful for interpreting H - ICS for

coupled PNP’s (especially in the dimeric case).

4.4 High order image correlation spectroscopy

(H - ICS) of plasmon coupled nanoparticles

Before performing H - ICS simulation of the experimentally acquired images,

we simulated confocal laser scanning images of samples that contained

monomers and dimers (α = 4) of gold nanospheres of 80 nm diameter. From

section 4.2, the QY of AuNS dimer was estimated to be four times that of

monomers. To produce more accurate results, we subtracted the noise signal

from the measured signals as discussed in Eq. 3.6. We extracted the average

monomer number per beam area (N1), the average dimer/cluster number per

beam area (N2), and quantum yield (α2) for the selected scattered images.

Hence, to determine the total number of particles of the scattered images,

total imaging area and e - radius of the point spread function of scattered

particles were required.

Thus, we calculated the total area of the acquired images, taking the average of

80 images, which was 20 ± 2 µm2 (Figure.4.5A and C). Exploiting the values

of N1, N2, total imaging area and e - radius of the particles in the formula

(as illustrated in section 3.5), we calculated the total number of monomers

and percentage of oligomers in the total imaging area. Thus, we inspected the

Chapter 4 80

Figure 4.2 Gold nanoparticle (AuNP) characterisation: UV- vis spectra of bare goldnanosphere (AuNS) of diameter 80 nm compared with Mie theory and FDTD simulations.The UV- vis spectrum is the ensemble spectra and red shifted compared with Mie theoryand FDTD calculated for single particle spectra.

aggregation details of AuNPs interaction in the acquired images. Outcomes of

these analyses are presented in the experimental section.

4.5 Experimental

In this section we discusse sample preparation, grid fabrication and experi-

mental results.

4.5.1 Sample preparation

The 80 nm diameter Au nanospheres used in this study were purchased from

NanoSeedz Ltd (Hong Kong). To prepare different concentration samples,

nanosphere solution was dispersed 1:1 into an aqueous solution of 2% PVA

(molecular weight 36 kDa). This solution was then spin - coated onto a co

- ordinate marked fabricated (5 nm Ti layer) grid attached to a glass slide,

Chapter 4 81

with the spin parameters adjusted to produce an approximately 150 nm thick

nanorod/PVA film. The use of a PVA layer ensured that, the nanosphere

remained attached to the grid during multiple SEM sessions and due to the

co - ordinate marked grid, we could correlate the optical and SEM images

accurately. UV-vis spectra, of bare AuNS of diameter 80 nm compared with

Mie theory and FDTD simulations are shown in Figure.4.2.

4.5.2 Grid fabrication

To correlate optical images, we used an electron beam lithography (EBL)

fabricated grid (Figure.4.3). The EBL fabricated grid, contained three different

sized blocks (57 µm , 97µm , 137 µm). Each block contained 16 sub - blocks

of 13 µm , 23 µm and 33 µm respectively for imaging flexibility. Glass surface

were cleaned with acetone/ethanol/methanol/H2O and a 5 nm Ti adhesion

layer was added. Later, a 100 nm poly methyl methacrylate (PMMA) layer

was introduced, followed by e - beam exposure, Au sputtering and lift - off.

Each individual block contained five rows and five columns in total 25 sub

- blocks that could be traced by corresponding row and column values. For

example, in the magnified (red marked) grid in Figure.4.3A (marked as A to

- D in a vertical direction and 1 to - 4 in a horizontal direction), the first row

first column value is A1, and the last row last column value is D4.

4.6 Results and discussion

In this section, we will present the experimental results of investigation into

oligomerisation using H - ICS simulations, which will be followed by discussion

and conclusion, validating the H - ICS results using single particle spectroscopy.

Chapter 4 82

Figure 4.3 Grid fabrication: (A) scanning electron microscope (SEM) images ofmagnified grid location, (B) SEM images of fabricated full grid.

4.6.1 High order image correlation spectroscopy (H -

ICS) for plasmon coupled dielectric samples

We prepared samples of different concentrations (e.g. optical density 0 - 10, NP

concentration 2.75× 109 to - 5.5× 1010 NPmL

). From the TEM images of different

concentrations we found, 2 - 4 ± 2, NP/µm2(from the average of 100 images)

for optical density 0 - 0.5, which we defined as the low concentration samples

and 20 - 30 ± 8, NP/µm2(value taken from average of 150 images) for optical

density 1 - 10 which we considered the high concentration sample. Prepared

samples were irradiated with Ti - sapphire femtosecond tunable (700 - 1100

nm) laser (Tsunami, Spectra - Physics) and then focused using a 1.4 numerical

aperture, an oil immersion objective lens (Olympus) and a 50 µm pinhole

(Figure.4.4). The sample was mounted onto a Piezo stage (PI) for scanning

and controlled using the program Labview. An oil immersion medium was used

to match the refractive index. The immersion oil ensured that there was no

background signal coming from the interface, and only the PNPs scatters and

scattering signals were collected using a photomultiplier tube (PMT, Oriel)

using a circular polariser. Figure.4.4 shows the schematic diagram of confocal

Chapter 4 83

Figure 4.4 Schematic diagram of laser scattering confocal set up.

laser scattering microscopy (CLSM) setup showing the major equipment.

We captured CLSM images of low (2 - 4 ± 2, NP/µm2) and high concentration

(20 - 30) ± 8, NP/µm2) samples for two mixed populations of species (e.g.

monomers and clusters). To investigate the ability of the method, the first

three higher order moments, upto G2(0, 0) of Eq.3.4 were calculated, to resolve

the monomer - dimer distributions of the captured images. We extracted the

average monomer number (N1), average dimer number (N2) and quantum

yield (α2) for six and seven different sets of 100 images each, for low and

high concentration samples respectively. To produce more accurate results,

we subtracted the noise signal from the measured signals as discussed in

Eq. 3.6. The mean values of the parameters measured from each set of

images were plotted as a function of concentration of population N1, N2 ,

and α2 (Figs.4.6A - 4.6C and Figs.4.7A - 4.7C). Examples of optical and SEM

correlated (low concentration) and uncorrelated (high concentration) samples

are shown in Figure.4.5. Acquired optical images (CLSM) of low concentration

samples matched perfectly with correlated SEM images (Figure.4.5A and B).

Chapter 4 84

(A) (B)

(C) (D)

Figure 4.5 Correlation of optical and SEM/TEM images of 80 nm gold nano - sphere.(A) confocal scattering images for low concentration samples and (B) correspondingcorrelated SEM images; (C) confocal scattering images for high concentration samples,and (D) corresponding un - correlated (same location) TEM images. scale bar = 4µm.

For high concentration samples, particles scattering overlapped due to the inter

- particle coupling effect. This coupling effect makes correlating optical and

SEM images difficult; however by tracing the EBL fabricated co - ordinate

grid, we confirmed that the optical images and SEM images were acquired

from the same location (Figs.4.5C and D).

From the noise corrected H - ICS simulated results (as discussed in section

3.5), we found that, H - ICS works convincingly, when monomer, dimer

and quantum yield concentration per beam area is above 0.1, 0.01 and 0.5

respectively, otherwise it deviates from the real values. In experimental

observations, of low and high concentration samples we found, monomer

particle density varied from 0.1 to - 0.40 particles per BA, dimer particle

Chapter 4 85

(A) (B)

(C) (D)

Figure 4.6 Low concentration sample :- The number of gold nanoparticles inaggregated samples was calculated using H - ICS and the real number. (A) calculatedaverage monomer number <N1> per beam area, (B) average dimer number per beamarea (N2), (C) quantum yield α2 and (D) percentage of oligomers for the selectedimages (A - F represent six different locations). The error bar represents standard error.Each data point is an average from analysis of 20 images.

density varied from 0.01 to - 0.05 particles per BA and quantum yield density

varied from 2∼4 (Figs.4.6A - 4.6C and Figs.4.7A - 4.7C), which are all within

the accuracy limit of H - ICS (as discussed in section 3.5). To retrieve accurate

values using the moment method, the density of monomeric particles must be

equal to or greater than the density of the dimer population otherwise the

results obtained can deviate from expected values. Additionally, numerous

kinds of background noise such as interference from surface, and detector dark

count affects the accuracy and precision of the results. The existence of noise

reduces the mean relative intensity fluctuations resulting in overestimation of

number densities and underestimation of the relative intensity yield. From the

simulated results (illustrated in Figure.3.7), an S/B ratio of 30 or higher would

be required to precisely resolve monomer - dimer distribution which perfectly

matches our experimental observations for varying concentrations.

Chapter 4 86

(A) (B)

(C) (D)

Figure 4.7 High concentration sample :- The number of gold nanoparticles inaggregated samples was calculated using H - ICS and the real number. (A) averagemonomer number per beam area, (B) average dimer number per beam area N2, (C)quantum yield α2 and (D) percentage of oligomers for the selected images (A - Grepresent seven different locations) and E) comparison of SEM and H - ICS dimernumbers. Thye error bar represents standard error. Each data point is an average fromanalysis of 20 images.

Chapter 4 87

Subtracting the noise signal from the original signal, we determined the total

number of particles and dimers in the selected regions, exploiting average

monomer number per beam area (N1), average dimer number per beam area

(N2), and e - radius of the particles. As illustrated in section 3.5 we considered,

the dimer e - radius 1.2 times that of the monomer. Additionally, with the help

of EBL fabricated co - ordinate grids, the total monomer and dimer numbers

were also determined from the correlated samples (low and high concentration

samples).

For both of the cases, (low and high concentration), we considered that,

our sample contained two emitters :- monomers and high order clusters (e.g.

dimers, trimers, tetramers). So the total number of particles was determined by

taking the sum of monomer and cluster numbers. H - ICS calculated monomer

and dimer numbers were compared with correlated real numbers of monomers

and dimers extracted from SEM. In our H - ICS simulation, we only considered

two emitter systems (monomers and dimers), and thus could not resolve the

contribution of high - order clusters (e.g. trimers, tetramers etc) as they were

considered to be dimers. An error (calculated using the formula discussed

in section 3.5) was produced when real monomer and dimer numbers were

compared with SEM images. The comparison between H - ICS and SEM

numbers showed a Gaussian distribution, concentrated at zero indicating that

the distributions frequency matches many time do they distribute (Figure.4.8).

The standard deviation of the distribution gives the error to be around 30 %

for N1, N2 and α2, which is quite acceptable for a large number of data sets.

Several factors are responsible for the error distribution and accuracy of H

- ICS. Point spread function distribution, due to the presence of trimers or

high order oligomers in the low and high concentration samples, significantly

affected the H - ICS results. Experimentally, e - radius of dimers is found

Chapter 4 88

(A) (B)

(C)

Figure 4.8 Error distributions for N1, N2 and α2 among 100 cases of imagesconsidering the contribution of monomers and high - order clusters (e.g. trimers andtetramers). The distribution of error varied from 0% to - 30 % due to the presence ofhigh - order clusters.

(A) (B)

(C)

Figure 4.9 Error distributions for N1, N2 and α2 among 100 cases of imagesconsidering monomer and dimer contribution. Discarding the contribution of high ordercluster (e.g. trimers, tetramers), only considering the contribution of monomers anddimers, the error can be reduced to 10 % , and the accuracy of the H - ICS analysis canbe improved.

.

Chapter 4 89

(A) (B)

Figure 4.10 (A) Intensity variation due to polarisation sensitivity (00-1800).Experimental values for dimer one and two extracted from polarisation dependent imagesfor 00-1800, fit perfectly with cosine, showing cos2 dependency and (B) dimer spectrumextracted from wavelength dependent images from 700 - 900 nm wavelength matcheswith FDTD simulated AuNS dimer spectrum of 2 nm separation.

to be about 1.1~2 times greater than monomers (0.46 µm) (see section 3.5).

If, focal size increases, the number of particles in the given region decreases,

therefore number of particles per focus volume increases. Discarding the high

order clusters, only considering the contribution of dimers and monomers and

carefully adjusting the e - radius of dimers (0.55 µm) the accuracy of H - ICS

was improved. In this case, standard deviation of the distribution gave an

the error around 10 % for N1, N2 and α2, significantly improving H - ICS

performance (Figure.4.9). Figure.4.8 shows the error distribution among 100

cases of images. The distribution of error varies from 0% to - 30 % due to the

presence of high order clusters. This would reduce the error by upto 10% (see

Figure.4.9).

4.6.2 Validating high order image correlation spectro-

scopy (H - ICS) results using single particle

spectroscopy

We acquired polarisation dependent images, by using a linear polariser and

varying the incident angle from 00 to - 1800 at 150 intervals. From the

Chapter 4 90

Figure 4.11 Dimer number calculated using polarization spectroscopy and H - ICStechnique and compared with the dimer number extracted from SEM images.

polarisation dependent images, we found that, some of the particles show

cosine squared dependency (Figure.4.10A), due to their dipolar characteristic.

However, with a change of polarisation the remaining particles (monomers) do

not show dependency (e.g. linear dependency).

We also performed the wavelength dependent imaging for 700 - 900 nm

wavelengths using a circular polariser at 10 nm intervals. Particle intensity

decreases with an increase in wavelength in the 700 - 900 nm spectral range

(Figure.4.10B). Experimentally acquired spectra were compared with FDTD

simulated dimers for several separations, which fit nicely with 2 nm separations,

confirming that they are dimers, in line with the polarisation dependency

results for the correlated images (Figure.4.5). Therefore, we extracted the

monomer and dimer number for the correlated images and compared these

with the H - ICS simulated results.

Figure.4.11 shows the comparison of dimer numbers at six different locations

(A - F) using H - ICS, SEM and polarisation spectroscopy techniques. As an

example, for location C the extracted dimer number is 5 which is similar to the

Chapter 4 91

polarisation and SEM extracted dimer numbers. The situation is almost the

same for the other locations, except for a few deviations due to the presence

of high - order clusters (e.g. - trimers, tetramers).

4.7 Conclusion

In this chapter, through extensive simulations and experiments, for the first

time, aggregation of randomly distributed AuNP inside a polymer matrix has

been demonstrated using ICS for plasmon - coupled nanoparticles. Monomer

and dimer numbers were calculated by performing ICS together with plasmon

coupling, and these number corresponded closely with the real number

of monomers and dimers determined by SEM/TEM images for low and

high concentration samples. This technique could be extended to cellular

environments to investigate PNP uptake and oligomerisation inside cells. This

finding paves the way for cell biologists to utlise plasmonic PNPs for labelling

and investigating PNP - cell interactions, aggregation and uptake kinetics to

understand the signaling pathways of intact cells in the unexplored regime (10

- 100 nm).

Chapter 4 92

Chapter 5

Gold nanoparticle uptake and

aggregation dynamics in HeLa

cells using image correlation

spectroscopy

5.1 Abstract

In this chapter we present a feasibility study of image correlation spectroscopy

of three different surface modified (e.g. bare, PEG and maleimide coated)

gold nanoparticles (AuNPs) of 50 nm, 80 nm and 100 nm diameter. We

bridge an existing probing technique, image correlation spectroscopy with

plasmon coupling for understanding aggregation dynamics and nanoparticle

uptake kinetics in HeLa cells. Optimum AuNP size for maximum uptake and

highest oligomerisation will be identified as a function of surface charge and

incubation time. The findings will pave the way for cell biologist to utlise this

tool to understand PNP uptake and aggregation dynamics of in - vitro cellular

93

Chapter 5 94

studies.

5.2 Introduction

Plasmonic nanoparticles exhibit excellent optical properties due to their size,

shape and tip geometry. Thus they have been extensively used in biological

applications including biological imaging, [1–6] bio labelling and sensing, [7]

photothermal cancer therapy, [8, 9] drug and gene delivery, [7, 10] and probing

membrane proteins. [11, 12] Understanding of PNP - cell interactions at the

single particle level is poor. However, previous work to understand PNP - cell

interactions has focused on cellular uptake and cytotoxicity, [13–20] in relation

to PNP size, shape, [14] surface effects [15] and aggregation effects. [16, 200]

Factors that have not been demonstrated methodically and which therefore

require further study, include quantitative and qualitative cellular uptake of

PNP, molecular aggregation and cellular movement with respect to PNPs size,

shape, concentration and incubation time.

Indeed, PNP uptake quantification methods are still in their infancy, including

destructive electron microscopy, inductively coupled plasma atomic emission

(ICP - AES) or inductively coupled plasma mass spectroscopy (ICP - MS).

TEM analysis could provide high resolution images (nm scale length) of

inner structures but cannot be used for live - cell imaging as it destroys

cells. Non - destructive microscopic optical methods such as image correlation

spectroscopy combined with dark - field scattering microscopy are ideal for

characterisation but until now have not been used for plasmonic nanoparticles.

Image correlation spectroscopy (ICS) was introduced as an imaging extension

of fluorescent correlation spectroscopy (FCS) to measure molecular transport

and aggregation in fixed cells. [190, 199, 201]

So far, no systematic study on cell uptake of AuNPs and aggregation has

Chapter 5 95

been conducted. A good characterisation method is needed to accurately

quantify AuNP aggregation behaviour. Image correlation spectroscopy (ICS)

is a powerful image analysis tool, but has not been extended to plasmonic

nanoparticle applications yet. [199] To address this challenging task, we bridge

an existing probing technique, image correlation spectroscopy with plasmon

coupling to quantify the PNP uptake and aggregation dynamics of in - vitro

cellular studies.

In chapter 4, we studied the feasibility of the H - ICS technique to investigate

AuNPs aggregation in a dielectric medium, where spherical AuNPs were

embedded in polyvinyl alcohol (PVA) at varying concentration to induce

aggregation. In the present study, we extended our technique into a cellular

environment, where different surface modified plasmonic nanoparticles were

incubated with human cervical carcinoma (HeLa) cells, to study PNP uptake

kinetics and aggregation dynamics.

For our study, we selected the bare, PEG and maleimide coated AuNSs of

50, 80 and 100 nm diameter. We acquired dark - field scattering images

and performed H - ICS utilising nanoparticle plasmon coupling. From the

simulated results, we explored the AuNP uptake and aggregation kinetics as a

function of size, surface charge and incubation time.

5.3 High order image correlation spectroscopy

(H - ICS) of plasmon coupled nanoparticles

To calculate AuNP uptake and aggregation, we performed H - ICS on the

acquired images and solved simultaneous equation 3.4, extracting N1 average

monomer number per beam area, N2 average dimer/cluster number per beam

area and α2 quantum yield for the selected dark - field scattered images. To

Chapter 5 96

(A) (B)

(C)

Figure 5.1 Gold nanoparticle characterisation: UV - vis spectra of bare, PEG andmaleimide coated AuNSs of diameter, (A) 50 nm (B) 80 nm and (C) 100 nm.

determine the total number of particles of the scattered images, we needed to

know the total cell area and e - radius of the scattered particles.

Thus, we calculated the total area of cells taking the average of 100 cells

which is 30 ± 8 µm2. Inserting the values of N1, N2, α2, total imaging area

and e - radius of the particles, we calculated the total number of monomer

and percentage of oligomers in the total imaging area, following the formula

discussed in section 3.5. Therefore, we extracted AuNP uptake and percentage

of oligomerisation for bare, PEG and maleimide coated AuNS samples of 50,

80 and 100 nm incubated with HeLa cells as a function of incubation time.

Outcomes of these analyses are presented in the experimental section.

Chapter 5 97

Samples UV - vis(nm) (Exp)1 Size(nm)2 pH3

Bare NS - 50 532 50±4 6~8PEG - NS - 50 527 50±4 7

Maleimide - NS - 50 538 - -Bare NS - 80 553 80 ± 6 6~8PEG - NS - 80 553 80 ± 6 7

Maleimide - NS - 80 552 - -Bare NS - 100 567 100±8 6~8PEG - NS - 100 563 100±8 7

Maleimide - NS - 100 566 - -

Table 5.1 UV - vis spectrum peak (nm), size distribution (nm) and pH for bare, PEGand maleimide coated 50 nm, 80 nm and 100 nm diameter gold nanosphere (AuNS).Here, 1 represents the measured value 2 and 3 represents company supplied values.

5.4 Surface modified AuNPs - HeLa cell sample

preparation

In this section, we will discuss AuNP sample preparation, HeLa cell culture

and incubation of AuNPs with HeLa cells. We will conclude this section by

presenting the experimental set up and characterisation.

AuNP sample preparation

The 50, 80 and 100 nm diameter bare and PEG AuNSs used in this study were

purchased from NanoSeedz Ltd (Hong Kong) and 50, 80 and 100 nm diameter

maleimide AuNSs were purchased from Nanopartz (USA). The solution was

diluted to prepare optical density one , (1.1 ×1010 AuNPmL

) and spectrum was

recorded using UV - vis spectrometer (Figure.5.1). Gold nanosphere solution

was then centrifuged for 15 min at 6000 rpm and the supernatant was discarded

adding a similar amount of distilled water to AuNP pellets. Then AuNPs

were sonicated for 10 min to minimise the aggregation followed by 30 min

of UV light exposure (inside a bio - safety cabinet) for sterilisation purposes.

Figure.5.1 shows the UV - vis spectrum of 50, 80 and 100 nm diameter AuNS.

Chapter 5 98

Samples Zeta (mv) Mobility pH CoatingBare NS - 100 -17.82± 2.85 -0.81± 0.62 8.2 CTABPEG - NS - 100 -22.03± 2.70 -1.15± 0.14 8.5 PEG

Maleimide - NS - 100 -26.48± 1.70 1.41 ± 0.07 8.3 Maleimide

Table 5.2 Zeta potential, mobility and pH for bare, PEG and maleimide coated 100nm diameter AuNSs.

Additionally, Table.5.1 shows the UV - vis spectrum, size distribution and pH

of 50, 80 and 100 nm diameter AuNSs.

HeLa cell culture

Dulbecco’s Modified Eagle Medium (DMEM) was purchased from Sigma

Aldrich, (AU). Foetal bovine serum (FBS), glutamine, penicillin/streptomycin

and amphotericin B were obtained from Life Technologies, (AU) and used as

received. HeLa cells were maintained in DMEM supplemented with 10% (v/v)

FBS, 1% (v/v) glutamine and 0.5% (v/v) amphotericin B in a humidified

atmosphere (95% (v/v) air, 5% (v/v) carbon dioxide) at 37 °C. Cells (1.5 x

104 cellscm2 ) were trypsinised and seeded under serum - free conditions in a µ -

slide chamber (Dksh, AU) containing a coverslip. Samples were incubated for

one day at 37°C / 5% CO2 to allow cell attachment.

Incubation of AuNPs with HeLa cells

Therefore, 100 µL of different functionalised AuNSs was added to the cultured

HeLa cells and PNP cell complexes incubated for fixed incubation times (e.g.

30 min, 1 hr, 6hrs, 12 hr, 18 hrs, 24 hrs, 48 hrs and 72 hrs). Finally, cells were

fixed with 3.7% (v/v) formalin for 10 min, followed by three rounds of washing

with phosphate buffered saline (PBS) for 5 min.

Characterisation

Absorption spectra of different functionalised (bare, PEG and maleimide

coated) AuNSs of 50, 80 and 100 nm diameters were taken on a UV - vis

Chapter 5 99

CCD

AuNP-Cell Sample l

Mirror

Halogen Lamp

Objective

Dark Field Condenser

Ti/S -100

Microscope

Figure 5.2 Schematic diagram of dark field microscopy set up.

spectrometer (Figure.5.1). To determine the surface charge of our surface

modified AuNP samples, we calculated the zeta potential using the Huckel

model on a 90 Plus particle analyzer, Brookhaven Instruments Corporation

(USA). To do so, firstly we conditioned the electrode using 0.15 M KCl . We

proceeded to the next step, whenever the conductance value exceeded 3 x 104

µS. For dilution purposes, bare, PEG and maleimide samples were mixed into

1 : 100 distilled water. Finally, we ran the program to calculate zeta potential

and mobility for three sets of measurements of individual samples for around

15 trials and took average values. The results are presented in Table.5.2.

Cell cultures on coverslip were mounted in microscopy mounting media

(Aquatex, Merck Millipore, AU). Three different diameters (50, 80 and 100

nm) of surface modified (bare, PEG and maleimide) AuNS samples of several

incubation times (30 min, 1 hr, 6 hrs, 12 hrs, 18 hrs, 24 hrs, 48 hrs and 7 2hrs)

Chapter 5 100

6hr 12hr

18hr 24hr

48hr 72hr

Figure 5.3 Dark - field scattering images of 50 nm diameter bare gold nano sphere fordifferent incubation times.

were imaged with a dark - field inverted microscope (Eclipse Ti - S, Nikon,

AU), using a 1.2 - 1.3NA (Nikon, AU) dark - field condenser and a 0.6 - 1.3NA,

100x oil immersion objective lens. White light was used for excitation, and

scattering images were acquired with a Nikon DS - Fi1c - U3 5Mb coloured

cooled digital camera.

Figure. 5.2 shows the dark - field scattering setup showing the major

components used :- a Nikon Ti - s/L100 microscope, a 100W white light as an

excitation source, focusing lenses, a 0.8 - 95NA dark - field condenser, a 0.75

NA 40x objective lense, and a Nikon colour CCD DS - Fi1c to capture the

scattering images.

Chapter 5 101

5.5 Experimental results

In this section, we discuss the experimental results of investigating the AuNP

uptake and oligomerisation for three different surface modified AuNP attached

HeLa cell samples. H - ICS simulation results will be presented to investigate

the AuNP uptake and oligomerisation as a function of AuNP size, surface

chemistry and incubation time.

5.5.1 High order image correlation spectroscopy (H -

ICS) for gold nanoparticle incubated HeLa cell

images

We captured the dark - field scattering (DFS) images of different surface

modified AuNPs (e.g. bare, maleimide and PEG) attached to HeLa cells

(Figs.5.3 - 5.6). To investigate the efficacy of the method, H - ICS simulations

were performed to resolve the distributions of the captured images (in total

1500 images). We performed the H - ICS exclusively for three surface modified

AuNPs (e.g. bare, maleimide and PEG) of 50, 80 and 100 nm diameter

attached to HeLa cells. We captured 20 dark - field scattering (DFS) images for

each incubation time for three different diameter (e.g. 50 nm, 80 nm and 100

nm) and three surface modified AuNPs samples. Dark - field scattering images

comprise two emitters (monomers and dimers only due to low optical density of

AuNPs) and H - ICS simulations were performed for the captured DFS images

for the first three higher order moments, upto G2,2 (0, 0) of equation 3.4, to

extract the values of average monomer number <N1>, average dimer number

<N2> and quantum yield α2. The mean value of the parameter was measured

from each set of images, and plotted as a function of incubation time to extract

the concentration of emitters <N1>, <N2> and their quantum yield ratio α2.

Chapter 5 102

6hr 12hr

18hr 24hr

48hr 72hr

Figure 5.4 Dark - field scattering images of 80 nm diameter maleimide coated goldnano - spheres for different incubation times.

For the case of surface modified AuNPs of 50 nm, 80 nm and 100 nm diameters,

attached to HeLa cell samples the monomer particle density varied from 0 to -

1.5, 0 to - 3.0 and 0 to - 1.5 particles per beam area (BA), dimer particle density

varied from 0 to - 0.5, 0 to - 1.2 and 0 to - 0.5 particles per BA for 50 nm, 80

nm and 100 nm diameter AuNSs respectively (Figuress.5.7 - 5.9). Quantum

yield ratio varied from 2 to ~ 4 confirming that, most of the NP attached to

HeLa cells are in dimeric form and there might be very few chances of getting

higher order clusters (e.g. trimers, tetramers or higher) and the variation is

due to the separation distance of dimers.

Chapter 5 103

6hr 12hr

18hr 24hr

48hr 72hr

Figure 5.5 Dark - field scattering images of 100 nm diameters PEG coated gold nano- spheres for different incubation times.

From the simulated results in section 3.5, it was observed, ICS performed well

when the density of monomeric particles was equal to or greater than the

density of the dimer population. As discussed in chapter 3, several kinds of

background noise such as interference from surface, detector dark counts affect

the precision of ICS. We confirmed with the simulation, an S/B ratio of 30 or

higher would be required to precisely resolve monomer - dimer distribution.

In our experimental observations, monomer density was greater than dimer

density (Figs.5.7 - 5.9). In order to eliminate the contribution of noise,

we derived the noise corrected spatial high order autocorrelation function

Chapter 5 104

as discussed in equation 3.6 where, we subtracted the noise signal from the

original measured signals. Therefore, exploiting average monomer number per

beam area <N1>, cluster number <N2> per beam area and multiplying the

N1 and N2 values with total area, total numbers of clusters in the selected

region of interest were determined.

As we used lower concentrations of AuNS (one optical density, 1.1 ×1010

AuNPmL

), we presumed our DFS images contained two emitters (monomer

and dimer), which was mimicked in our experimental observation as our

quantum yield varied between 2 - 4 (Figures.5.7C, 5.8C and 5.9C). From

our previous experimental observation illustrated in section 3.5 we have seen

dimer point spread function is larger than monomer and e - radius is 1.2

times greater than monomers (section 3.5). With the increase of point spread

functions, the numbers of particles in the given region decreases, as does

the number of particles per focus volume which significantly affects the H

- ICS results. Though, we anticipated error for our H - ICS results due to

different background noise level, point spread functions, which are minimized

considering greater number of images. In addition, the error bar was calculated

taking the average over 20 images per incubation times and placed in all

calculations (Figures.5.7 - 5.9). H - ICS simulated results of 50, 80 and

100 nm diameter surface modified (e.g. bare, PEG and maleimide) AuNSs

oligomerisation and uptake by HeLa cells are presented in (Figures.5.7D, 5.7E,

5.8D, 5.8E and 5.9D, 5.9E). From the H - ICS simulated results, an increasing

trend was observed for plasmonic nanoparticle internalisation and aggregation

with increasing incubation time.

Chapter 5 105

Maleimide

PEG Bare

Figure 5.6 Dark - field scattering images of 80 nm diameter bare, PEG and maleimidecoated gold nano - spheres for two days incubation.

5.5.2 Gold nanoparticle uptake due to surface modific-

ation

When surface modified (e.g. bare, PEG and maleimide) AuNPs attached to

HeLa cells they responded differently in accordance with surface chemistry. H

- ICS simulation was performed on different surface modified AuNPs incubated

HeLa cell images, (Figs.5.7E, 5.8E and 5.9E). For 50 nm diameter AuNSs we

can speculate that, bare, PEG and maleimide AuNPs follows an increasing

trend and become saturated after 48 hrs of incubation except for bare at 24

hrs (Figure.5.7E). Bare AuNP uptake is faster than PEG or maleimide AuNPs.

At 80 nm diameter we can presume that, PEG and maleimide AuNSs show a

similar increasing trend except maleimide AuNP uptake, is slowest (upto 24 hr)

compared with PEG AuNPs and becomes saturated after 48 hrs incubation

Chapter 5 106

(A) (B)

(C) (D)

(E)

Figure 5.7 Cellular uptake and aggregation kinetics of 50 nm diameter goldnanoparticles (AuNPs) as a function of incubation time for different surface modifiedgold nanospheres (AuNSs). (A) H - ICS extracted monomer number per beam area, (B)H - ICS extracted oligomer (mostly dimer) number per beam area, (C) H - ICS extractedquantum yield, (D) cellular uptake of gold nanoparticles (AuNPs) and ( E) percentageof oligomers.

Chapter 5 107

(A) (B)

(C) (D)

(E)

Figure 5.8 Cellular uptake and aggregation kinetics of 80 nm diameter goldnanoparticles (AuNPs) as a function of incubation time for different surface modifiedgold nanospheres (AuNSs). (A) H - ICS extracted monomer number per beam area, (B)H - ICS extracted oligomer (mostly dimer) number per beam area, (C) H - ICS extractedquantum yield, (D) cellular uptake of gold nanoparticles (AuNPs) and ( E) percentageof oligomers.

Chapter 5 108

(A) (B)

(C) (D)

(E)

Figure 5.9 Cellular uptake and aggregation kinetics of 100 nm diameter goldnanoparticle (AuNPs) as a function of incubation time for different surface modifiedgold nanospheres (AuNSs). (A) H - ICS extracted monomer number per beam area, (B)H - ICS extracted oligomer (mostly dimer) number per beam area, (C) H - ICS extractedquantum yield, (D) cellular uptake of AuNPs and ( E) percentage of oligomers.

Chapter 5 109

time (Figure.5.8E). PNP uptake of bare AuNSs dramatically increases with

shorter incubation time compared with PEG and maleimide AuNSs showing

highest uptake at 18 hrs incubation time, then reducing significantly and

becoming saturated after 48 hrs incubation. For 100 nm diameter AuNSs

we can conclude that, PNP uptake for maleimide and PEG coated AuNSs

shows an increasing trend upto 24 hrs incubation, and becomes saturated after

48 hrs incubation (Figure.5.9E). Interestingly, PNP uptakes for bare AuNSs

was higher at shorter incubation times compared with PEG and maleimide

AuNSs but became saturated after 48 hrs incubation. In general, PNP uptake

for maleimide coated AuNSs was higher for 50 nm diameter AuNPs, lower

for 80 nm diameter AuNPs and in between for 100 nm diameter AuNPs

(Figures.5.7 - 5.9). In contrast, PNP uptake was higher for 50 nm diameter

AuNPs, lower for 100 nm diameter AuNPs and in between for 80 nm diameter

AuNPs (Figures.5.7 - 5.9) for both PEG - coated and bare AuNSs. HeLa cells

responded impeccably to PEG and maleimide coated PNP. A good number

of cells were visible after 24 hrs incubation but few cells were visible for bare

PNPs after 24 hrs incubation due to the toxicity effect.

5.5.3 Gold nanoparticle oligomerisation due to surface

modification

For 50 nm diameter AuNSs we can conclude that, bare AuNSs oligomerise

faster at shorter incubation times than PEG and maleimide AuNPs (Fig-

ure.5.7D), which shows similar oligomerisation kinetics. For 80 nm diameter

AuNSs, bare AuNSs oligomerise faster (reaching maximum value of 55 % at

shorter incubation time of 24 hrs) than PEG and maleimide NPs (Figure.5.8D).

PEG AuNS oligomerisation was slower and maleimide was slowest at shorter

incubation times. For 100 nm diameter AuNSs there was similar increasing

trend for PEG, bare and maleimide AuNS, oligomerisation reached maximum

Chapter 5 110

for PEG coated AuNPs (50%) at 24 hrs and becames saturated after 24hrs

(Figure.5.9). Interestingly, oligomerisation was faster (reaching maximum at

24hrs) for 100 nm diameter PEG - coated AuNSs, but slower for maleimide

and bare AuNSs. On the basis of our results, we can speculate that,

different surface modification mediates plasmonic nanoparticle uptake and

oligomerisation differently as a function of different incubation time.

5.5.4 Effect of size on gold nanoparticle uptake

Gold nanoparticle uptake varies for different size nanoparticles and different

incubation times (Figure.5.10). Bare nanoparticle uptake rate is faster for 50

nm diameter gold nanoparticles and reaches its maximum at 18 hrs incubation

time; uptake rate is slower for 80 nm diameter and slowest for 100 nm diameter

AuNSs (Figure.5.10A). PEG nanoparticle uptake rate is faster for 50 nm

diameter gold nanoparticles slower for 80 nm diameter nanoparticles, and

in between for 100 nm diameter nanoparticles (Figure.5.10B). Surprisingly,

maleimide nanoparticle uptake rate is faster for 100 nm diameter gold

nanoparticles slower for 80 nm diameter, and in between for 100 nm diameter

(Figure.5.10C). For bare, PEG and melaimide coated AuNSs, uptake reaches

its maximum (e.g. around 400 AuNPs) after longer incubation time (e.g. 3

days). One of the possible reasons may be that, there are no unbounded

protein sites available on the surface of gold nanoparticles for further binding.

This result suggests that, different diameter AuNPs respond differently with

different surface modified particles, due to the different membrane wrapping

time required by HeLa cells. These results are in line with published results

[13, 14], which suggest, surface charge influences AuNPs uptake. Different

functional group are also responsible for distinct uptake kinetics suggesting

that, there might be an optimal size for which specific surface modified gold

nanoparticle uptake will be highest.

Chapter 5 111

(A) (B)

(C)

Figure 5.10 Effect of size on (different surface modified) gold nanoparticle uptake fordifferent incubation times.

5.5.5 Effect of size on gold nanoparticle oligomerisation

Gold nanoparticle oligomerisation varies for different size nanoparticles and

different incubation times (Figure.5.11). Bare - coated 50 nm diameter

nanoparticle show fastest oligomerisation, 80 nm diameters nanoparticles

are slower and 100 nm diameter nanoparticles are slowest (Figure.5.11A).

Interestingly, for PEG coated gold nanoparticle oligomerisation is fastest for

80 nm diamter. Therefore, 100 nm diameters PEG coated nanoparticles

show highest oligomerisation (e.g. 45%) at the shorter incubation time 24

hrs, compared with other diameters (Figure.5.11B). In maleimide coated

nanoparticles of 50 nm, 80 nm and 100 nm diameter uptake is similar at shorter

incubation times. Though similar to PEG coated AuNSs, maleimide coated

100 nm diameter AuNSs shows highest oligomerisation (e.g. 40%) after shorter

Chapter 5 112

(A) (B)

(C)

Figure 5.11 Effect of size on (different surface modified) gold nanoparticleoligomerisation for different incubation times.

incubation times (e.g. 24 hrs). For all cases, AuNP uptake becomes saturated

for longer incubation times. This result suggests that, different diameter

gold nanoparticles respond differently with different surface modified AuNSs

and there might be an optimal size for surface modified gold nanoparticles

oligomerisation.

5.6 Discussion

Plasmonic nanoparticle - cell interaction and uptake depends on PNP size

and shape. However PNP surface modification (e.g. PEG, PAA, PAH and

maleimide) or functionalisation with different biomolecules (e.g. protein,

DNA) notably affects their interaction. [50]

In 2006, PNP uptake was carried out by Chan and co - workers [14] for 14,

Chapter 5 113

50 and 74 nm diameter of AuNSs and 40×14 and 74×14 nm AuNRs. Among

these 50 nm diameter AuNSs showed most efficient uptake using ICP - AES.

Jiang and co - workers [51] reported minimal uptake for 25 - 50 nm (among 2,

10, 25, 40, 50, 70, 80 and 90 nm) transferrin coated AuNSs, attached to HeLa

cells, quantified by TEM and laser capture microdissection (LCM). Wang and

co - workers [52] reported maximal uptake for 45 nm AuNSs (among 45, 70 and

100 nm) attached to HeLa cells using TEM and dark - field microscopy. In our

study, we observed maximal uptake for 50 nm AuNS (among 50, 80 and 100

nm) for three surface modified AuNS. A possible reason for minimal uptake

due to optimal size (e.g. 50 nm diameter) could be that, 50 nm diameter

spherical particle require minimal time to wrap around spheres compared with

other diameter particles, which therefore increases AuNP uptake, which is in

agreement with previously reported experimental observation [14, 51–53] and

thermodynamic calculations. [54]

Previously, Chan and co - workers studied the AuNPs cellular uptake kinetics

of gold nanoparticles as a function of incubation time (upto 10 hr) for three

different size gold nanoparticles (diameters of 14 nm, 50 nm, and 74 nm) and

found AuNP uptake increases with an increase of incubation time. However,

AuNPs show an increasing trend upto a certain incubation time and become

saturated at longer incubation times.

Recently Sheng and co - workers [202] reported the endocytosis of AuNPs (e.g.

45 nm, 70 nm and 110 nm diameter), in various cells (the human cancer cell

lines, CL1 - 0 and HeLa). They identified 3D distribution of AuNPs in living

cells using a dark - field optical sectioning microscope. Their statistical results

showed maximum uptake for 45 nm AuNPs for both cell types. They found

the total number of AuNPs (in and on HeLa cells) were 2167 , 564 and 108 for

45 nm, 70 nm and 110 nm respectively when AuNPs were functionalised with

single - stranded DNA (ssDNA) and attached to HeLa cells for 2 hours.

Chapter 5 114

In our study using the combined technique, considering all incubation times

(30 min to 3 days), maximal AuNPs uptake by 50 nm diameter AuNS was 560

± 35, 512 ± 17 and 482 ± 16 for bare, PEG and maleimide coated AuNSs

respectively. Maximal AuNP uptake by 80 nm, diameter AuNS was 450 ±

32, 345 ± 28, 370 and ± 40 for bare, PEG and maleimide coated AuNSs

respectively. Similarly, maximal AuNPs uptake by 100 nm, diameter AuNS

was 398 ± 31, 375 ± 12 and 405 ± 10 for bare, PEG and maleimide coated

AuNSs respectively.

During these experimental analyses it was observed that, initially AuNP

uptake by cells appeared to be dependent on the geometry and surface charge of

nanoparticle, however, at later incubation time uptake became saturated and

equilibrated to a level irrespective of geometry and surface chemistry. These

slight variations in uptake rate and oligomerisation may be due to difference

in the internalisation mechanisms, whereas saturation of AuNS uptake and

oligomerisation at longer incubation times may be due to inadequate protein

site availability on AuNPs surfaces for binding, [14]

PNP uptake could be either endocytosis (or other energy - dependent

mechanisms) or membrane association. A previous study suggested that a,

greater degree of uptake is energy dependent or due to endocytic mechanisms

and some amount of nanoparticle uptake could be independent of energy

activitye (membrane association), suggesting PNPs could interact with the

membrane of the cell and facilitate uptake. [203]

Endocytosis is not always the only mechanism of internalisation, as membrane

association of nanoparticles can induce physical interactions that allow particle

internalisation. However, all geometries show very different uptake and

aggregation profiles as a function of incubation time, in line with previous

results. [13, 14, 53] These observations suggest that the mechanisms by which

the particles enter the cells vary and are dependent on their relative size and

Chapter 5 115

surface chemistry.

Size dependent AuNP uptake can be explained by thermodynamic model

of the many PNP - cell system [54, 90, 92, 95, 96] for receptor - mediated

endocytosis. [204, 205] For AuNPs uptake there are two competing energies -

one is the binding energy between ligands and receptors and the other is the

thermodynamic driving force for wrapping. Binding energy depends on degree

of ligand - receptor interaction and the diffusion kinetics for the recruitment of

receptors to the binding site. On the other hand, thermodynamic driving force

refers to the amount of free energy required to drive PNPs into the cell. These

two factors determine how fast and how many PNPs are taken up by the cell.

PNPs smaller than 40 nm diameter produces an inadequate amount of free

energy to completely wrap the PNP surfaces of the membrane due to docking.

PNPs larger than 80 nm diameter require more time to wrap the larger surfaces

area hence AuNPs uptake is reduced. The reduction of free receptors also limits

the ligand - receptor binding energy forming large membrane curvature, which

affects AuNP uptake. Considering membrane bending rigidity and ligand -

receptor binding energy, an optimal AuNPs diameter has been identified at

which the cellular uptake of PNPs is maximised. [91, 206, 207] The optimal

diameter for AuNPs uptake falls in the range of 40 - 60 nm for reasonable

values of membrane bending rigidity and ligand - receptor binding energy. In

our study we found maximal uptake for 50 nm diameter AuNSs and reduced

uptake for larger diameter AuNSs (80 nm and 100nm), which is in line with

previous findings. [14, 202]

Surface charge influence the PNP uptake and cytotoxicity. Since most cells

(either cancerous or normal) seem to have negative surface charge, they offer

greater permeability for cationic particles. On the other hand, positively

charged PNPs (e.g.CTAB coated NRs) has greater cytotoxicity than negatively

charged PNPs (e.g.Citrate coated NSs). [13] Also most of the negatively

Chapter 5 116

charged or neutral PNPs undergo non - specific adsorption of the particles on

the cell membrane. AuNP uptake due to surface chemistry has been reported

by Arnida and co - workers. [60, 61] They compared the uptake of bare and

PEG coated NSs (30, 50 and 90 nm diameter) with PEG NR (3×10nm, 45×10

nm) for PC-3 and RAW 2647 cells. They reported most efficient uptake for

50 nm non - PEGlated NSs, which was quantified by ICP - MS and TEM

analysis. PEGlated NR uptakes was worse than NSs.

In another study, Alkilany and co - workers [13] reported that, uptake

mostly depends on functional group adsorbed onto the AuNPs surface rather

than surface charge. They reported higher uptake for molecules containing

quaternary amines (e.g. CTAB and PDADMAC), lower uptake for negatively

charged sulfate groups (PSS) and intermediate uptake for molecules containing

primary amines (e.g. PAH).

In our study, we observed the AuNP uptake due to three different functional

groups such as, bare AuNSs coated with CTAB (contains quaternary amines),

PEG (contains alcohol groups) and maleimide (contains secondary amines).

Due to variation between these functional group, functional group kinetics.

We observed the maximal uptake for bare AuNS (CTAB coated), lowest

uptake for maleimide coated AuNS and intermediate uptake for PEG coated

AuNSs, which are in line with the previous study. [13, 59] One of the possible

reason could be that, CTAB coated (positively charged) AuNSs penetrate the

negatively charged cell membrane more effectively than PEG or maleimde

coated AuNPs at lower concentration. As CTAB is detrimental to cells, with

increasing incubation time, number of viable cells decreases and effectively

decreases the AuNPs uptake for bare AuNS at longer incubation times. Due to

the presence of secondary amines, maleimide coated AuNSs showed less uptake,

and consequently lower aggregation than bare and PEG AuNSs, whereas PEG

coated AuNSs showed intermediate uptake.

Chapter 5 117

In another study, Cathrine and co - workers observed that sedimentation

increases with the increase of AuNPs aggregation in cell media for 90 nm

AuNPs and 300 nm × 20 nm AuNR. [200] Oligomerisation is expected to

increase with increase of AuNP uptake.

In our study, considering all incubation times (30 min to 3 day) for all

three surface modified AuNSs (e.g. 50 nm, 80 nm and 100 nm diameter

AuNS samples) we observed highest uptake for 50 nm diameter nanoparticles

and distinct kinetics at different incubation times. In addition, AuNP

oligomerisation was different for different diameter AuNSs as a function of

surface chemistry due to variation of functional group. Interestingly, for bare

coated AuNSs, highest oligomerisation (48%) was observed for 80 nm diameter

AuNSs, lowest for 100 nm and intermediate for 50 nm diameter. Conversely,

we observed highest oligomerisation for 80 nm diameter AuNSs, lowest for 50

nm and intermediate for 100 nm diameter PEG and maleimide coated AuNSs,

hence the percentage of oligomerisation was different. These results suggest

that, due to different surface chemistry there might be an optimal size such as

80 nm diameter AuNSs, for which highest oligomerisation could be observed.

In future experiments, we would like to determine, the AuNPs uptake and

oligomerisation and effect of surface charge of different shape AuNPs (e.g.

AuNRs, dumbbells or bipyramids).

5.7 Conclusion

Understanding PNP - cell interaction is a challenging task. Previously,

various techniques have been used to determine plasmonic nanoparticles

uptake and aggregation of PNPs embedded in cell samples but no systematic

study has been performed. Here, with extensive H - ICS simulations

and experimentation, for the first time, we have demonstrated a technique

Chapter 5 118

to determine the plasmonic nanoparticles uptake kinetics and aggregation

dynamics at a previously unexplored region (10 - 100 µm) which has not been

reported yet for a cellular environment. The findings of these studies will

be helpful for probing membrane protein interactions, drug delivery, disease

diagnosis, and also in cancer therapy. These results will also provide new

insight to understanding the aggregation dynamics of PNPs inside cells and

studying the consequence of oligomerisation, cell activation and signaling

pathways, for living cells.

Chapter 6

Conclusions and future work

6.1 Thesis conclusions

This thesis has detailed my research into the feasibility of H - ICS for coupled

AuNPs, and how this tool can be exploited for quantifying AuNPs uptake

and oligomerisation kinetics. Towards this achievement several major research

areas have been explored :

1. Observations of plasmon coupling of different diameter (50, 80 and 80 nm)

AuNS dimers using numerical simulations of FDTD. These simulation results

deduce QY for AuNS dimers.

2. Simulating images of monomer - dimer mixture (the same as in experi-

mental AuNS scattering images) and utilising the QY extracted from FDTD

simulations to solve simultaneous equations, we extrapolated the boundary

conditions of precise application of H - ICS for coupled AuNPs.

3. Experimental demonstration of H - ICS for coupled AuNPs embedded in a

dielectric medium. The total monomer and dimer numbers extrapolated from

H - ICS simulations perfectly matched the total monomer and dimer numbers

119

Chapter 6 120

extracted from correlated SEM images.

4. Demonstration of polarisation and wavelength spectroscopy combined with

plasmon coupling for investigating single particles and dimers. The total

monomer and dimer numbers extrapolated from single particle spectroscopy

analyses were in line with previous H - ICS results and the number extracted

from correlated SEM images.

5. Experimental demonstration of a technique combining H - ICS, with

plasmon coupling, for cells incubated with AuNSs. AuNP uptake and

oligomerisation were successfully demonstrated - as a function of AuNS size,

surface charge and incubation time.

The SPR effect and plasmon coupling properties present in plasmonic nano-

particles enable numerous biological applications. However, there is still much

to understand about PNP - cell interactions. The majority of the research

thus far has focused on, PNP cytotoxicity and uptake. PNP uptake has been

investigated through ICP - MS and ICP - AEP along with TEM or, SEM.

Recently plasmon coupling has been used to explore molecular interactions;

although techniques used, each have limitations, most importantly they are

not suitable for live cell imaging. Thus, it was necessary to explore a

technique, that could be suitable for live cell imaging. To address this issue,

a combined technique utilising nanoparticle plasmon coupling together with

image correlation spectroscopy, has been demonstrated here for the first time.

This technique can quantify not only AuNP uptake but also AuNP aggregation

inside biological cells. The beauty of this technique is that, it is non -

destructive and thus can be used for live cell imaging purposes.

Before introducing this technique in a cellular environment, we successfully

demonstrated the combined technique in a dielectric medium. AuNPs were

embedded in PVA matrix, and a confocal laser scattering microscopy setup

Chapter 6 121

was used to acquire the scattering images. To determine the feasibility and

precision of H - ICS simulation in experimental conditions, we performed

numerical simulations of AuNS dimers to extrapolate QY. This value was then

used in H - ICS simulations to extract the monomer and dimer numbers. The

total particle and dimer numbers extracted from the H - ICS simulation were

compared with a reference number extracted from correlated SEM images,

which matched perfectly.

After successful use of H - ICS for coupled particles in a dielectric medium, the

tool was next traialled in a cellular environment. AuNSs were incubated into

HeLa cells, to investigate AuNS uptake and oligomerisation. We performed H -

ICS on dark - field scattered images, carefully interpreting the QY for coupled

AuNPs, utilising FDTD simulation results. Finally, we explored NP uptake

kinetics and aggregation dynamics as a function of size, surface chemistry and

incubation time.

6.2 Future research

The research conducted for this thesis could be further extended for anisotropic

materials and aggregation of more than two emitters on different types of

cells, considering refractive index change due to cellular environment, and

how these properties may affect AuNP uptake and oligomerisation. The

AuNP uptake and oligomerisation dynamics studied here were demonstrated

using a combination of extensive H -I CS, plasmon coupling simulations and

experimentation.

6.2.1 More than two emitter

In these H - ICS simulations, high order autocorrelation function was solved

only for two emitters, assuming that our images consisted of a monomer -

Chapter 6 122

(A)

(B) (C)

Figure 6.1 (A) Wet chemically synthesised gold nanorods (AuNRs) dropcasted ontoa transmission electron microscope (TEM) grid. Histograms displaying the distributionof (B) aspect ratio (red), and (C) length (green) and width (red). Measurements weretaken from transmission electron microscope (TEM) images, using the fit ellipse optionin ImageJ, with hand fit ellipses, to avoid threshold errors.

dimer mixture and solving the simultaneous equation up to the first three

higher order moments (see Eq.3.4). Theoretically it is possible to extract the

information of samples containing multiple emitters, which is an extremley

time consuming process, as it requires solving simultaneous equations of higher

orders. To investigate whether any higher order oligomers formed or not, we

need to extend H - ICS simulations for trimers or higher order oligomers.

For a three emitter system, we need to solve upto six higher order moments

as discussed in Eq.3.4. Therefore if more species are included in one image,

simultaneous equations of higher orders need to be solved. In addition, we need

to perform FDTD simulations for trimers, tetramers or higher order oligomers

to determine quantum yield. Utlising QY values H - ICS simulations could be

extended for coupled PNPs for multiple emitters.

Chapter 6 123

In our experimental study, we used lower optical density and less concentrated

AuNS samples and surface modified (non - specific functionalisation) PNPs for

the cellular environment. PNP uptake and oligomerisation in different types

of cells could be explored using high optical density or high concentration

AuNS samples and specific functionalised (e.g. protein or DNA molecules)

different shape and size PNPs (e.g. AuNSs, AuNRs, dumbbells, bipyramids),

and solving the simultaneous equations acquiring QY from FDTD simulations

for more than two emitters (e.g. monomer - dimer - trimer, monomer - dimer

- tetramer).

Here, wet chemically synthesised AuNRs were dropcasted onto a TEM grid

and the corresponding histogram showed size anisotropy (Figure.6.1). Due to

size anisotropy plasmonic nanoparticles offer variation in QY, hence by using

QY values into H - ICS simulations PNP uptake and oligomerisation could be

explored.

For simplicity, we attached AuNSs to HeLa cells and explored the QY

for coupled AuNS dimers using the FDTD technique. However, QY of

PNPs drastically changes due to the size, shape, orientation and coupling

of plasmonic particles. As discussed in chapter 2, tip radius of AuNPs also

influences QY. Thus, the QY of different shapes such as NRs, dumbells and

bipyramids could be examined by H - ICS. The H - ICS model described

in this work, could also be extended to multiple emitters, and QY variation

of anisotropic materials could enable exploration of numerous properties

and interactions inside biological cells. Additionally, AuNPs functionalised

with proteins, DNA, antigens, antibodies or any other micromolecules and

conjugated with different cell types could add an extra dimension to future

research in this area.

Chapter 6 124

6.2.2 Validation by other techniques

Plasmon resonance is sensitive to the environment surrounding the AuNP. If

the AuNP surroundings change, refractive index (RI) can changes significantly.

Previously, the RI of cellular components were investigated in ensembles of NPs

with significantly lower spatial resolution (µm scale). The RI values have been

reported for cell membrane (1.46 - 1.6), cytoplasm (1.35 - 1.39) and protein

(1.36 - 1.55). [208–213]

In our study, we performed the confocal laser scattering microscopy and

confirmed the internalisation of AuNPs. Staining the cell membrane by

dye molecules we captured the confocal laser scattering images (Figure.6.2).

Cell membrane were clearly visible (red color border) in the confocal images

along with internalised AuNSs (green color), when excited by la aser source.

Therefore, we used the refractive index, (n) close to water, which was the

assumed media. Analysing the results we found that, after a change of

refractive index Δn = 0.2 ~ 0.6, SPR peak changed minimally therefore would

not affect the measurement critically.

As a preliminary study, we examined the feasibility of our H - ICS technique

confirming the internalisation of AuNPs in HeLa cells, however we did not

look specifically at, whether aggregation and uptake was happening in the

nucleus or the cytoplasm. The position of the AuNPs could be helpful for

analysing AuNP uptake and aggregation, which could be achieved by confocal

laser scattering microscopy, dark field scattering microscopy or, transmission

electron microscopy. AuNP uptake and aggregation kinetics at particular

locations (e.g. cell membrane, nucleus, cytoplasm or proteins) could be

explored by, integrating live cell imaging with confocal laser scattering or dark

- field scattering, attaching the dye molecules to HeLa cells and following the

trajectory of AuNPs over a long period of time. The position could be useful

Chapter 6 125

Figure 6.2 Confocal laser scattering, gold nanoparticle internalised cell images. Cellmembrane was stained by dye molecules, clearly visible (red color borders) in the confocalimages and internalised gold nanosphere was also clearly visible (green color particles)while excited by laser source. Scale bar represents 15 µm.

to investigate cell signalling and other cellular activities.

We performed the extensive H – ICS simulations together with plasmon

coupling to investigate AuNP uptake and aggregation kinetics. We did not

perform any other experimental methods or techniques to confirm the ICS

results for cells. However, in chapters 3 and 4, match between simulation and

experimental data to confirm the validity of H – ICS. .

Nevertheless, additional confirmation of ICS results could be achieved by

Chapter 6 126

performing live cell scattering microscopy, transmission electron microscopy,

inductively coupled plasma atomic spectroscopy (ICP - AES) or inductively

coupled plasma mass spectroscopy or combining any two of these techniques.

6.3 Conclusions

To conclude, the beauty of the tool outlined in this thesis is, that it does not

require any pre or post image analysis software, sophisticated experimental

setup or expensive equipment such as lasers or detector. All it requires is, a

microscope equipped with a white light source, an objective lens and a CCD

to record the images. In addition, the image acquisition time is faster (1

~ 5 msec to acquire a single image) than other methods. Using this tool,

insights into AuNP uptake and aggregation could be found in many membrane

protein interactions responsible for cell signalling and other cellular activities.

Thus, the tool could eventually be useful for cancer cell therapy, drug delivery,

disease diagnosis and numerous other biological applications. Although visual

resolution itself may not be improved, the ability to ‘see’ inside cells may be

enhanced by the use of the methods discovered here.

Bibliography

[1] Z. Peter, W. M. C. James, and G. Min, “Five dimensional optical

recording mediated by surface plasmons in gold nanorods,” Nature 459,

pp. 410–413, 2005.

[2] K. Sokolov, M. Follen, J. Aaron, I. Pavlova, A. Malpica, R. Lotan, and

K. R. Richards, “Real time vital optical imaging of precancer using

anti-epidermal growth factor receptor antibodies conjugated to gold

nanoparticles,” Cancer Research 63(9), pp. 1999–2004, 2003.

[3] J. Aaron, K. Travis, N. Harrison, and K. Sokolov, “Dynamic imaging

of molecular assemblies in live cells based on nanoparticle plasmon

resonance coupling,” Nano letters 9(10), pp. 3612–3618, 2009.

[4] J. Aaron, N. Nitin, K. Travis, S. Kumar, T. Collier, S. Y. Park,

M. Jose Yacaman, L. Coghlan, M. Follen, R. Richards Kortum, and

K. Sokolov, “Plasmon resonance coupling of metal nanoparticles for

molecular imaging of carcinogenesis in vivo,” Journal of Biomedical

Optics 12(3), pp. 034007–034011, 2007.

[5] E. Sayed, H. Ivan, X. Huang, and M. A. El Sayed, “Surface plasmon

resonance scattering and absorption of anti egfr antibody conjugated

gold nanoparticles in cancer diagnostics applications in oral cancer,”

Nano Letters 5(5), pp. 829–834, 2005.

127

Chapter 6 128

[6] R. Mercatelli, G. Romano, F. Ratto, P. Matteini, S. Centi, F. Cialda,

M. Monici, R. Pini, and F. Fusi, “Quantitative measurement of scattering

and extinction spectra of nanoparticles by darkfield microscopy,” Applied

Physics Letters 99(13), pp. 1–3, 2011.

[7] C. J. Murphy, A. M. Gole, S. E. Hunyadi, J. W. Stone, P. N. Sisco,

A. Alkilany, E. Kinard Brian, and P. Hankins, “Chemical sensing and

imaging with metallic nanorods,” Chemical Communication , pp. 544–

557, 2008.

[8] X. Huang, I. El Sayed, W. Qian, and M. A. El Sayed, “Cancer cell

imaging and photothermal therapy in the near-infrared region by using

gold nanorods,” Journal of the American Chemical Society 128(6),

pp. 2115–2120, 2006.

[9] A. Wijaya, S. B. Schaffer, I. G. Pallares, and K. Hamad Schifferli,

“Selective release of multiple dna oligonucleotides from gold nanorods,”

ACS Nano 3(1), pp. 80–86, 2009.

[10] C. C. Chen, Y. P. Lin, C. W. Wang, g. H. C. Tzen, C. H. Wu, Y. C. Chen,

C. P. Chen, L. C. Chen, and Y. C. Wu, “Dna gold nanorod conjugates for

remote control of localized gene expression by near infrared irradiation,”

Journal of the American Chemical Society 128(11), pp. 3709–3715, 2006.

[11] H. A. Clayton Andrew, F. Walker, S. Orchard, C. Henderson, D. Fuchs,

J. Rothacker, E. C. Nice, and B. A. W, “Ligand-induced dimer-tetramer

transition during the activation of the cell surface epidermal growth

factor receptor-a multidimensional microscopy analysis,” Journal of

Biological Chemistry 280(34), pp. 30392–30399, 2005.

[12] C. Sonnichsen, M. L. ReinhardBjorn, and A. P. Jan Alivisatos, “A

molecular ruler based on plasmon coupling of single gold and silver

nanoparticles,” Nat Biotechl 23(6), pp. 741–745, 2005.

Chapter 6 129

[13] A. M. Alkilany, k. K. Nagaria Prati, l. C. R. Hexe, T. J. Shaw,

C. J. Murphy, and M. D. Wyatt, “Cellular uptake and cytotoxicity

of gold nanorods: Molecular origin of cytotoxicity and surface effects,”

Small 5(6), pp. 701–708, 2009.

[14] B. D. Chithrani, A. A. Ghazani, and n. W. C. W. Cha, “Determining the

size and shape dependence of gold nanoparticle uptake into mammalian

cells,” Nano Letters 6(4), pp. 662–668, 2006.

[15] C. Grabinski, N. Schaeublin, A. Wijaya, H. D Couto, S. H. Baxamusa,

K. Hamad Schifferli, and S. M. Hussain, “Effect of gold nanorod surface

chemistry on cellular response,” ACS Nano 5(4), pp. 2870–2879, 2011.

[16] A. Albanese and C. W. Chan Warren, “Effect of gold nanoparticle

aggregation on cell uptake and toxicity,” ACS Nano 5(7), pp. 5478–5489,

2011.

[17] T. Hauck, A. Ghazani, and W. W. Chan, “Assessing the effect of surface

chemistry on gold nanorod uptake, toxicity, and gene expression in

mammalian cells,” Small 4(1), pp. 153–159, 2008.

[18] C. J. Murphy, A. M. Gole, J. W. Stone, P. N. Sisco, A. M. Alkilany, E. C.

Goldsmith, and S. C. Baxter, “Gold nanoparticles in biology: Beyond

toxicity to cellular imaging,” Accounts of Chemical Research 41(12),

pp. 1721–1730, 2008.

[19] L. Raphael, S. Umbreen, C. Yann, and S. Violaine, “Gold nanoparticles

delivery in mammalian live cells: a critical review,” Nano Reviews 1(0),

2010.

[20] A. Wei, Q. Wei, and A. P. Leonov, “Gold nanorods as theranostic

agents,” pp. 659–681, 2011.

Chapter 6 130

[21] R. Elghanian, J. J. Storhoff, R. C. Mucic, R. Letsinger, and C. A. Mirkin,

“Selective colorimetric detection of polynucleotides based on the distance

dependent optical properties of gold nanoparticles,” Science 277(5329),

pp. 1078–1081, 1997.

[22] A. P. Alivisatos, K. Johnsson, P. P, W. Xiaogang, E. L. Troy, J. B.

Colin, P. S. , Marcel, and P. G, “Organization of nanocrystal using

dna,” Nature 382, pp. 609–611, 1996.

[23] C. Sonnichsen, B. L. Reinhard, and A. A. P. Jan, “A molecular ruler

based on plasmon coupling of single gold and silver nanoparticles,”

Nature biotechnology 23(6), pp. 741–745, 2005.

[24] B. M. Reinhard, M. Siu, l. H. Agarwa, A. P. Alivisatos, and L. Jan,

“Calibration of dynamic molecular rulers based on plasmon coupling

between gold nanoparticles,” Nano Letters 5(11), pp. 2246–2252, 2005.

[25] P. K. Jain, W. Huang, and M. A. El Sayed, “On the universal scaling

behavior of the distance decay of plasmon coupling in metal nanoparticle

pairs a plasmon ruler equation,” Nano Letters 7(7), pp. 2080–2088, 2007.

[26] H. Wang, g. G. Ron, B. Yan, L. Yang, and B. M. Reinhard, “Optical

sizing of immunolabel clusters through multispectral plasmon coupling

microscopy,” Nano Letters 11(2), pp. 498–504, 2011.

[27] J. Yguerabide and E. E. Yguerabide Analytical Biochemistry 262(0),

p. 137.

[28] P. K. Jain, K. S. Lee, I. H. El Sayed, and M. A. El Sayed Journal of

Physical Chemistry B 110(0), p. 7238, 2006.

[29] N. L. Rosi and C. A. Mirkin Chemical Review 105(0), p. 1547, 2005.

[30] A. P. Alivisatos Nature Biotechnology 22(0), p. 47, 2004.

Chapter 6 131

[31] C. Sonnichsen, T. Franzl, T. Wilk, G. von Plessen, J. Feldmann,

O. Wilson, and P. Mulvaney, “Drastic reduction of plasmon damping

in gold nanorods,” Physical Review Letters 88(0), p. 077402, 2002.

[32] E. Katz and I. A. Willner Chem Int Ed 43(0), p. 6042, 2004.

[33] J. Haes A and D. R. P. Van Journal of American Chemical

Society 124(0), p. 10596, 2002.

[34] A. J. Haes, l. W. P. Hal, L. Chang, W. L. Klein, and D. Van Nano

Letters 4(0), p. 1029, 2004.

[35] I. Yu C Journal of Analytical Chemistry 79(0), p. 572, 2007.

[36] Y. Sun and Y. Xia Analytical Chemistry 74(0), p. 5297, 2002.

[37] A. D. McFarland and D. R. P. Van Nano Letters 3(0), p. 1057, 2003.

[38] J. J. Perez, S. I. Pastoriza, L. M. Liz Marzan, and M. P Coord Chemical

Review 249(0), p. 1870, 2005.

[39] K. S. Lee and M. A. El Sayed Journal of Physical Chemistry B 110(0),

p. 19220, 2006.

[40] E. Hao and G. C. Schatz Journal of Chemical Physics 120(0), p. 357,

2004.

[41] J. J. Storhoff, R. C. Lazarides, A A Mucic, C. A. Mirkin, R. L. Letsinge,

and G. C. Schatz Journal of American Chemical Society 122(0), p. 4640,

2000.

[42] J. J. Storhoff, R. Elghanian, R. C. Mucic, C. A. Mirkin, and R. L.

Letsinger Journal of American Chemical Society 120(0), p. 1959, 1998.

[43] J. Aizpurua, G. W. Bryant, L. J. Richter, d. A. F. J. Garcia, and

T. Kelley, B Kand Mallouk Physical Review B 71(0), p. 235420/1, 2005.

Chapter 6 132

[44] K. H. Su, i. Wei, Q H, X. Zhang, J. J. Mock, D. R. Smith, and Schultz

Nanoletters 3(8), pp. 1087–1090, 2003.

[45] L. Gunnarsson, T. Rindzevicius, J. Prikulis, B. Kasemo, M. Kall, Z. S,

and G. C. Schatz Journal of Physical Chemistry B 109(0), p. 1079, 2005.

[46] L. A. Sweatlock, S. A. Maier, H. A. Atwater, J. J. Penninkhof, and

Polman Applied Physics Review B 71(0), p. 235408/1, 2005.

[47] J. J. Xiao, J. P. Huang, and K. W. Yu Applied Physics Review B 71(0),

p. 045404/1, 2005.

[48] W. Rechberger, A. Hohenau, A. Leitner, J. R. Krenn, B. Lamprecht,

and F. R. Aussenegg Optics communications 220(0), pp. 137–141, 2003.

[49] J. P. Kottmann and O. J. F. Martin Optics Letter 26(0), p. 1096, 2001.

[50] R. Ryma, R. Jessica P, E. M. R. Jim, and A. Nancy, “Surface

coatings determine cytotoxicity and irritation potential of quantum dot

nanoparticles in epidermal keratinocytes,” J Invest Dermatol 127(1),

pp. 143–153., 2006.

[51] W. Jiang, B. Y. S. Kim, J. T. Rutka, and W. C. W. Chan Nature

Nanotechnology 3, p. 145, 2008.

[52] S. H. Wang, C. W. Lee, A. Chiou, and P. K. Wei Journal of

Nanobiotechnology 8, p. 33, 2010.

[53] B. D. Chithrani and W. C. W. Chan Nano Letters 7, p. 15427, 2007.

[54] H. Gao, W. Shii, and L. Freund, “Mechanics of receptor mediated

endocytosis.,” Proceedings of the National Academy of Sciences of the

United States of America 102(27), pp. 9469–9474, 2005.

Chapter 6 133

[55] Q. Yang, L. Ying, W. Liming, X. Ligeng, B. Ru, J. Yinglu, W. Xiaochun,

Z. Yuliang, L. Yufeng, and C. Chunying, “Surface chemistry and aspect

ratio mediated cellular uptake of au nanorods,” Biomaterials 31(30),

pp. 7606 – 7619, 2010.

[56] W. Shuguang, L. Wentong, T. Oleg, S. R. Uma, Y. Hongtao, and C. R.

Paresh, “Challenge in understanding size and shape dependent toxicity

of gold nanomaterials in human skin keratinocytes,” Chemical Physics

Letters 463(1-3), pp. 145–149, 2008.

[57] N. Takuro, Y. Masato, O. Yuri, A. Yasuyuki, T. Hironobu, K. Takahito,

K. Yoshiki, and N. Yasuro, “Peg-modified gold nanorods with a stealth

character for in vivo applications,” Journal of Controlled Release 114(3),

pp. 343 – 347, 2006.

[58] C. B. Sanda and A. Simion, “Detoxification of gold nanorods by

conjugation with thiolated poly(ethylene glycol) and their assessment as

sers-active carriers of raman tags,” Nanotechnology 21(23), p. 235601,

2010.

[59] E. E. Connor, J. Mwamuka, A. Gole, C. J. Murphy, and M. D. Wyatt

Small2005 1, pp. 325–327, 2005.

[60] Arnida, A. Malugin, and H. Ghandehari, “Cellular uptake and toxicity of

gold nanoparticles in prostate cancer cells: a comparative study of rods

and spheres,” Journal of Applied Toxicology 30(3), pp. 212–217, 2010.

[61] M. M. Arnida, A. Janat, R. A, P. C M, and G. H, “Geometry and surface

characteristics of gold nanoparticles influence their biodistribution and

uptake by macrophages,” European Journal of Pharmaceutics and

Biopharmaceutics 77(3), pp. 417 – 423, 2011. Biological Barriers A

Need for Novel Tools in Nanotoxicology and Nanomedicine The 8th

Chapter 6 134

International Conference and Workshop - Biological Barriers - in vitro

Tools in Nanotoxicology and Nanomedicine.

[62] X. M. Qing, H. Li, H. Sun, and K. Ho Pharmacology Review 54, pp. 561–

587, 2002.

[63] S. Conner, S. D, and L. Sandra, “Regulated portals of entry into the

cell.,” 10.1038/nature01451 7(7), pp. 37–44, 2003.

[64] K. J. Kim and A. B. Malik, “Protein transport across the lung epithelial

barrier,” American Journal of Physiology Lung Cellular and Molecular

Physiology .

[65] V. Angeles, a. C. Magdalen, G. R. Alejandro, C. Macarena, V. V.

Sabino, J. S. Carlos, d. P. M. Maria, and M. Rodolfo, “The influence

of surface functionalization on the enhanced internalization of magnetic

nanoparticles in cancer cells,” Nanotechnology 20(11), p. 115103, 2009.

[66] P. Swanand, S. Amanda, H. Eric, S. William, and S. Sudipta, “Protein

adsorption and cellular uptake of cerium oxide nanoparticles as a

function of zeta potential,” Biomaterials 28(31), pp. 4600 – 4607, 2007.

[67] T. Mironava, M. Hadjiargyrou, M. Simon, V. Jurukovski, and M. H.

Rafailovich Nanotoxicology 4, p. 120, 2010.

[68] L. S. Rieznichenko, S. I. Shpyleva, T. G. Gruzina, S. N. Dybkova,

Z. R. Ulberg, and V. F. Chekhun Republic National Academy Science

Ukraine 2, p. 170, 2010.

[69] J. D. Trono, K. Mizuno, N. Yusa, T. Matsukawa, K. Yokoyama, and

M. Uesaka Journal Radiation Research 52, p. 103, 2011.

[70] Y. Shan, S. Ma, L. Nie, X. Shang, X. Hao, Z. Tang, and H. Wang

Chemical Communication 47, p. 8091, 2011.

Chapter 6 135

[71] X. Ma, Y. Wu, S. Jin, Y. Tian, X. Zhang, Y. Zhao, L. Yu, and X. J.

Liang ACS Nano 5, p. 8629, 2011.

[72] M. A. Sobhan, V. K. A. Sreenivasan, M. J. Withford, and E. M. Goldys

Colloids and Surfaces B Biointerfaces 92, p. 190, 2012.

[73] A. C. Sabuncu, J. Grubbs, S. Qian, T. M. Abdel Fattah, M. W. Stacey,

and A. Beskok Colloids and Surfaces B Biointerfaces 95, p. 96, 2012.

[74] K. Huang, H. Ma, J. Liu, S. Huo, A. Kumar, T. Wei, X. Zhang, S. Jin,

Y. Gan, P. C. Wang, S. He, X. Zhang, and X. J. Liang ACS Nano 6,

p. 4483, 2012.

[75] S. Huo, H. Ma, K. Huang, J. Liu, T. Wei, S. Jin, J. Zhang, S. He, and

X. J. Liang Cancer Research 73, p. 319, 2013.

[76] L. A. Dykman and N. G. Khlebtsov, “Uptake of engineered gold

nanoparticles into mammalian cells,” Chemical Reviews 114(2),

pp. 1258–1288, 2014.

[77] A. Verma, O. Uzun, Y. Hu, Y. Hu, Y. Hu, H. s. Han, n. N. Watso,

S. Chen, D. J Irvine, and F. Stellacci, “Surface structure regulated cell

membrane penetration by monolayer protected nanoparticles,” Nature

materials 422(6), pp. 588–595, 2008.

[78] Z. Matjaz and L. Ulo, “Cell-penetrating peptides: mechanism and

kinetics of cargo delivery,” Advanced Drug Delivery Reviews 57(4),

pp. 529 – 545, 2005. Protein- and Peptide-Mediated Transduction:

Mechanisms and Implications for Drug Delivery.

[79] B. Chen, Q. Liu, Y. Zhang, L. Xu, and X. Fang, “Transmembrane

delivery of the cell-penetrating peptide conjugated semiconductor

quantum dots,” Langmuir 24(20), pp. 11866–11871, 2008.

Chapter 6 136

[80] P. C. Patel, D. A. Giljohann, D. S. Seferos, and C. A. Mirkin,

“Peptide antisense nanoparticles,” Proceedings of the National Academy

of Sciences 105(45), pp. 17222–17226, 2008.

[81] J. J. Green, E. Chiu, E. S. Leshchiner, i. J. Sh, R. Langer, and

D. G. Anderson, “Electrostatic ligand coatings of nanoparticles enable

ligand-specific gene delivery to human primary cells,” Nano Letters 7(4),

pp. 874–879, 2007.

[82] V. Biju, D. Muraleedharan, K. i. Nakayama, Y. Shinohara, T. Itoh,

Y. Baba, and M. Ishikawa, “Quantum dot-insect neuropeptide

conjugates for fluorescence imaging, transfection, and nucleus targeting

of living cells,” Langmuir 23(20), pp. 10254–10261, 2007.

[83] H. Mok, J. W. Park, and T. G. Park, “Enhanced intracellular delivery

of quantum dot and adenovirus nanoparticles triggered by acidic ph via

surface charge reversal,” Bioconjugate Chemistry 19(4), pp. 797–801,

2008.

[84] L. Sun, D. Liu, and Z. Wang, “Functional gold nanoparticle peptide

complexes as cell targeting agents,” Langmuir 24(18), pp. 10293–10297,

2008.

[85] M. Thomas and A. Klibanov, “Enhancing polyethylenimine delivery of

plasmid dna into mammalian cells,” Proceedings of the National Academy

of Sciences of the United States of America 99(23), pp. 14640–14645,

2002.

[86] S. Banerji, H. K, and A. Mark, “Examination of nonendocytotic

bulk transport of nanoparticles across phospholipid membranes,”

Langmuir 23(6), pp. 3305–3313, 2007.

Chapter 6 137

[87] T. Xia, L. Rome, and A. Nel, “Nanobiology: Particles slip cell security,”

Nature Publishing Group 7(7), pp. 519–520.

[88] U. Taylor, S. Klein, S. Petersen, W. Kues, S. Barcikowski, and D. Rath,

“Nonendosomal cellular uptake of ligand free, positively charged gold

nanoparticles,” Cytometry Part A 77A(5), pp. 439–446, 2010.

[89] S. Konstantin, B. Jacques, and E. Perez, “Applications of fluorescence

correlation spectroscopy: Polydispersity measurements,” Journal of

Colloid and Interface Science 213(2), pp. 479–487, 1999.

[90] F. L B and L. Yuan, “The role of binder mobility in spontaneous adhesive

contact and implications for cell adhesion,” Journal of the Mechanics and

Physics of Solids 52(11), pp. 2455 – 2472, 2004.

[91] F. Osaki, T. Kanamori, S. Sando, T. Sera, and Y. Aoyama, “A quantum

dot conjugated sugar ball and its cellular uptake on the size effects

of endocytosis in the subviral region,” Jounal of American Chemical

Society 126, pp. 6520–6251, 2004.

[92] P. Decuzzi and M. Ferrari, “The role of specific and non specific

interactions in receptor mediated endocytosis of nanoparticles,”

Biomaterials 28(18), pp. 2915 – 2922, 2007.

[93] R. Lipowsky and H. G. Dobereiner, “Vesicles in contact with

nanoparticles and colloids,” Europhysical Letter 43(2), pp. 219–225,

1998.

[94] I. Canton and G. Battaglia, “Endocytosis at the nanoscale,” Chemecial

Society Review 41, pp. 2718–2739, 2012.

[95] H. Yuan, J. Li, G. Bao, and S. Zhang, “Variable nanoparticle

cell adhesion strength regulates cellular uptake,” Physical Review

Letters 105, p. 138101, Sep 2010.

Chapter 6 138

[96] X. Li, “Size and shape effects on receptor mediated endocytosis of

nanoparticles,” Journal of Applied Physics 111(2), p. 024702, 2012.

[97] M. P. D. Dobay, P. Alberola A, a. E. R. Mendoz, and J. O. Radler,

“Modeling nanoparticle uptake and intracellular distribution using

stochastic process algebras,” Journal of Nanoparticle Research 14(4),

2012.

[98] S. H. Tan and G. Horlick Applied Spectroscopy 40, p. 445, 1986.

[99] S. D. Tanner and V. I. Baranov Journal of American Society Mass

Spectrometry 10, p. 1083, 1999.

[100] T. M. Mayhew, C. Muhlfeld, D. Vanhecke, and M. Ochs Analytica

Analyte 191, p. 153, 2009.

[101] C. L. Pai, C. W. Wei, C. K. Liao, C. D. Chen, K. C. W. Pao, W. C R C,

and S. D. B. Y N, “Photoacoustic Imaging of Multiple Targets Using

Gold Nanorods, journal = Proceding SPIE , year = 2006, pages = 6086„”

[102] G. Decher, “Fuzzy nanoassemblies toward layered polymeric multicom-

posites,” Science 277(5330), pp. 1232–1237, 1997.

[103] K. Caruso, D. N. Niikura, Y. Furlong, and Okahata Langmuir 13,

p. 3427, 1997.

[104] D. I. Gittins and F. Caruso, “Tailoring the polyelectrolyte coating of

metal nanoparticles,” The Journal of Physical Chemistry B 105(29),

pp. 6846–6852, 2001.

[105] H. Ai, M. Fang, S. A. Jones, and Y. M. Lvov, “Electrostatic

layer by layer nanoassembly on biological microtemplates platelets,”

Biomacromolecules 3(3), pp. 560–564, 2002.

Chapter 6 139

[106] J. Turkevich, P. C. Stevenson, and J. Hillie, “A study of the nucleation

and growth processes in the synthesis of colloidal gold,” Discussions of

Faraday Society 11(0), pp. 55–75, 1951.

[107] J. Kimling, M. M, B. Okenve, B. Kotaidis, H. Ballot, and A. Plech,

“Turkevich method for gold nanoparticle synthesis revisited,” Journal of

Physical Chemistry B 110(0), pp. 15700–15707, 2006.

[108] G. Frens, “Particle size and sol stability in metal colloids,” Colloid and

Polymer Science 250(0), pp. 736–741, 1972.

[109] G. Frerns, “Controlled nucleation for the regulation of the particle size

in monodisperse gold suspensions,” Nature (London), Phys. Sci. 241(0),

pp. 20–22, 1973.

[110] B. J. Spence, “Colloidal solutions, and the refractive indices of gold,

platinum and silver,” Physical Review (Series I), 28(4), pp. 233–263,

1909.

[111] F. J. Burger and K. Sallner, “The action of ultrasonic waves in

suspensions,” Transactions of the Faraday Society 32(7), pp. 1598–1603,

1936.

[112] A. F. J. Colby, L. H. Gabor, S. Jon A, and M. Charles R, “Template-

synthesized nanoscopic gold particles optical spectra and the effects of

particle size and shape,” The Journal of Physical Chemistry 98(11),

pp. 2963–2971, 1994.

[113] S. D. Stookey and R. J. Araujo, “Selective polarization of light due

to absorption by small elongated silver particles in glass,” Applied

Optics 7(0), pp. 777–779, 1968.

[114] N. R. Jana, L. Gearheart, and C. J. Murphy, “Wet chemical synthesis

Chapter 6 140

of high aspect ratio cylindrical gold nanorods,” Journal of Physical

Chemistry B 105(0), pp. 4065–4067, 2001.

[115] B. Nikoobakht and M. A. El-Sayed, “Selective polarization of light due to

absorption by small elongated silver particles in glass,” Langmuir 17(0),

p. 6368, 2001.

[116] B. Nikoobakht and M. A. El Sayed, “Preparation and growth mechanism

of gold nanorods (nrs) using seed-mediated growth method,” Chemistry

of Materials 15(10), pp. 1957–1962, 2003.

[117] N. R. Jana, L. Gearheart, and C. J. Murphy, “Evidence for seedmediated

nucleation in the chemical reduction of gold salts to gold nanoparticles,”

Chemistry of Materials 13(7), pp. 2313–2322, 2001.

[118] N. R. Jana, “Gram-scale synthesis of soluble, near-monodisperse gold

nanorods and other anisotropic nanoparticles,” Small 1(8), pp. 875–882,

2005.

[119] D. B. Brantley, O. O. Sherine, and J. M. Catherine, “An improved

synthesis of high aspect ratio gold nanorods,” Advanced Materials 15(5),

pp. 414–416, 2003.

[120] F. Hao, C. L. Nehl, J. H. Hafner, and P. Nordlander, “Plasmon

resonances of a gold nanostar,” Nano Letters 7(3), pp. 729–732, 2007.

[121] O. Krichevski and G. Markovich, “Growth of colloidal gold nanostars

and nanowires induced by palladium doping,” Langmuir 23(3), pp. 1496–

1499, 2007.

[122] Z. Li, Z. Liu, J. Zhang, B. Han, J. Du, Y. Gao, and T. Jiang, “Synthesis

of single-crystal gold nanosheets of large size in ionic liquids,” Journal

of Physical Chemistry B 109(30), pp. 14445–14448, 2005.

Chapter 6 141

[123] C. J. Huang, P. H. Chiu, Y. H. Wang, T. H. Meen, and C. F. Yang,

“Synthesis and characterization of gold nanodogbones by the seeded

mediated growth method,” Nanotechnology 18(39), p. 0, 2007.

[124] Y. Sun, B. Mayers, and Y. Xia, “Metal nanostructures with hollow

interiors,” Advanced Materials 15(7), pp. 641–646, 2003.

[125] X. Kou, S. Zhan, C. K. Tsung, M. H. Yeung, Q. Shi, G. D. Stucky,

L. Sun, J. Wang, and C. Yan, “Growth of gold nanorods and bipyramids

using cteab surfactant,” The Journal of Physical Chemistry B 110(33),

pp. 16377–16383, 2006.

[126] S. I Pastoriza, J. J Perez, and M. L M Liz, “Silica coating

and hydrophobation of ctab stabilized gold nanorods,” Chemical

Mater 18(10), pp. 2465–2467, 2006.

[127] Q. Zhan, J. Qian, X. Li, and S. He, “A study of mesoporous silica

encapsulated gold nanorods as enhanced light scattering probes for

cancer cell imaging,” Nanotechnology 21(5), p. 055704, 2005.

[128] X. Huang, S. Neretina, and M. A. El Sayed, “Gold nanorods from

synthesis and properties to biological and biomedical applications,”

Advanced Mater 21, pp. 4880–4910, 2009.

[129] G. Mie, “Beitrage zur optik truber medien, speziell kolloidaler

metallosungen,” Annalen der Physik 330(3), pp. 377–445, 1908.

[130] U. V. Kreibig, “Optical sizing of immunolabel clusters through

multispectral plasmon coupling microscopy,” Optical Properties of Metal

Clusters; Springer-Verlag: Berlin , 1995.

[131] M. Paul, “Surface plasmon spectroscopy of nanosized metal particles,”

Langmuir 12(3), pp. 788–800, 1996.

Chapter 6 142

[132] K. L. Kelly, E. Coronado, L. L. Zhao, and G. C. Schatz, “The optical

properties of metal nanoparticles the influence of size shape and dielectric

environment,” The Journal of Physical Chemistry B 107(3), pp. 668–

677, 2003.

[133] X. Lu, M. Rycenga, S. E. Skrabalak, B. Wiley, and Y. Xia, “Chemical

synthesis of novel plasmonic nanoparticles,” Annual Review of Physical

Chemistry 60(1), pp. 167–192, 2009.

[134] M. Grzelczak, J. J. Perez, P. Mulvaney, and L. M. Liz Marzan, “Shape

control in gold nanoparticle synthesis,” Chemical Social Review 37,

pp. 1783–1791, 2008.

[135] S. Link, M. B. Mohamed, and M. A. El Sayed, “Simulation of the optical

absorption spectra of gold nanorods as a function of their aspect ratio and

the effect of the medium dielectric constant,” The Journal of Physical

Chemistry B 103(16), pp. 3073–3077, 1999.

[136] H. M, “Light scattering by small particles. von h. c. vande hulst. new

york: Dover publications, inc. 1981.,” Acta Polymerica 35(4), pp. 338–

338, 1984.

[137] C. F. Bohren and D. R. Huffman, “Absorption and scattering of light

by small particles,” Research supported by the University of Arizona and

Institute of Occupational and Environmental Health. New York, Wiley-

Interscience , pp. 1–541, 1983.

[138] R. Gans, “Uber die form ultramikroskopischer silberteilchen,” Annalen

der Physik 352(10), pp. 270–284, 1915.

[139] Y. Binghai, Yang, and Y. Wang, “Comment on simulation of the optical

absorption spectra of gold nanorods as a function of their aspect ratio and

Chapter 6 143

the effect of the medium dielectric constant,” The Journal of Physical

Chemistry B 107(34), pp. 9159–9159, 2003.

[140] S. W. Prescott and P. Mulvaney, “Gold nanorod extinction spectra,”

Journal of Applied Physics 99(12), p. 123504, 2006.

[141] B. T. Draine and P. J. Flatau, “Discrete dipole approximation for

periodic targets theory and tests,” Journal of Optical Society America

A 25, pp. 2693–2703, 2008.

[142] J. Peter B and C. R W, “Optical constants of the noblemetals,” Physical

Review B 6(12), p. 4370, 1972.

[143] M. N. C. Hu and A. Funston, “Dark-field microscopy studies of

single metal nanoparticles: understanding the factors that influence

the linewidth of the localized surface plasmon resonance.,” Journal of

materials chemistry 18(17), pp. 1949–1960, 2008.

[144] E. A. Coronad and G. C. Schatz, “Surface plasmon broadening for

arbitrary shape nanoparticles: A geometrical probability approach,” The

Journal of Chemical Physics 119(7), pp. 3926–3934, 2003.

[145] U. Kreibig and L. Genzel, “Optical absorption of small metallic

particles,” Surface Science 156, pp. 678 – 700, 1985.

[146] Liu, S. Guyot, and Philippe, “Mechanism of silver(i)-assisted growth

of gold nanorods and bipyramids,” The Journal of Physical Chemistry

B 109(47), pp. 22192–22200, 2005.

[147] Liu, S. Guyot, and Philippe, “Synthesis and optical characterization

of au/ag core/shell nanorods,” The Journal of Physical Chemistry

B 108(19).

Chapter 6 144

[148] C. J. Huang, P. H. Chiu, Y. H. Wang, T. H. Mee, and C. F. Yang,

“Growth of gold nanorods and bipyramids using cteab surfactant,” The

Journal of Physical Chemistry B 18(39), 2007.

[149] R. Jin, Y. Cao, C. A. Mirkin, K. L. Kelly, G. C. Schatz, and J. G.

Zheng, “Photoinduced conversion of silver nanospheres to nanoprisms,”

Science 294(5548), pp. 1901–1903, 2001.

[150] D. Averitt, D. Sarkar, and N. J. Halas, “Plasmon resonance shifts of

au-coated au 2 s nanoshells: Insight into multicomponent nanoparticle

growth,” Physical Review Letters 78(0), p. 4217, 1997.

[151] Y. G. Sun and Y. N. Xia, “Shape controlled synthesis of gold and silver

nanoparticles,” Science 298(0), p. 2176, 2002.

[152] F. Bohren and D. R. Huffman, “Absorption and scattering of light by

small particles,” Wiley, New York , 1983.

[153] M. I. Mishchenko, “Light scattering by randomly oriented axially

symmetric particles,” Journal of Optical Society America A 8(0),

pp. 871–882, 1991.

[154] M. I. Mishchenko, L. D. Travis, and D. W. Mackowski, “T-matrix

computations of light scattering by nonspherical particles: A review.,” J

Quant Spectrosc Radiat Transf 55(0), pp. 535–575, 1996.

[155] C. R. Purce, E M land Pennypacker, “Scattering and absorption of light

by nonspherical dielectric grains,” Astrophys Journal 186(0), pp. 705–

714, 1973.

[156] K. S. Ye, “Numerical solution of initial boundary value problems

involving maxwell s equations in isotropic media,” IEEE Trans.

Antennas Propag. 14(0), pp. 302–307, 1966.

Chapter 6 145

[157] A. Taflove and S. C. Hagness, “Computational electrodynamics the finite

difference time domain method,” Artech House, Boston , 2000.

[158] K. Yee, “Ieee transactions on antennas and propagation,” pp. 302–307,

1966.

[159] P. Nordlander, Coubre, E. Prodan, K. Li, and M. I. Stockman

Nanoletters 4(5), pp. ,899–903, 2004.

[160] K. J. Prashant, e. E. Susi, and A. E. s. Mostafa Journal of Physical

Chemistry B 110(37), pp. 18243–18253, 2006.

[161] A. M. Funston, C. Novo, T. J. Davis, and P. Mulvaney Nano Letter 9(0),

p. 1651, 2009.

[162] P. H. Cheng, G. Y. Xiao, B. K. Ling, and Y. Z. Yong Journal of Physical

Chemistry C 114(49), pp. 21123–21131, 2010.

[163] C. W. Kat, S. Lei, C. Huanjun, L. yao, w. Jiangfang, and Q. L. Hai ACS

NANO 5(7), pp. 5976–5986, 2011.

[164] S. Lei, F. Caihong, C. Huanjun, C. M. Yat, W. Jianfang, and Q. L. Hai

Nano Letter 12(3), pp. 1424–1430, 2012.

[165] A. Funston, D. TJ, C. Novo, and P. Mulvaney Phil Trans Royal Society

A 369(0), pp. 3472–3482, 2011.

[166] L. Chuntonov and G. Haran Nano Letter 11(6), pp. 2440–2445, 2011.

[167] Y. Bo and B. M. R. Svetlana, V Boriskinaand Journal of Physical

Chemistry C 115(11), pp. 4578–4583, 2011.

[168] M. Kasha, H. R. Rawls, and M. A. El-Bayoumi, “Pure applies chemistry,”

11, p. 371, 1965.

[169] M. Kasha, “Radiation research,” 20, p. 55, 1963.

Chapter 6 146

[170] B. Z. Packard, D. D. Toptygin, A. Komoriya, and L. Brand, “Physical

chemistry b,” 102, p. 752, 1998.

[171] M. Viktor, R. F. Jessica, P. S. Isabel, M. F. Alison, N. Carolina, M. Paul,

M. M. Luis, and A. F J G, “Modelling the optical response of gold

nanoparticles,” Chemical Social Review 37, pp. 1792–1805, 2008.

[172] P. W. Wiseman, J. A. E. M. H. Squier, and K. R. Wilson, “Two

photon image correlation spectroscopy and image cross correlation

spectroscopy,” Journal of Microscopy 200(1), 2000.

[173] P. W. Wiseman, “Image correlation spectroscopy development and

application to studies of pdgf receptor distributions,” PhD thesis, The

University of Western Ontario, London, ON , 1995.

[174] P. W. Wiseman, F. Capani, J. A. Squier, and M. E. Martone, “Counting

dendritic spines in brain tissue slices by image correlation spectroscopy

analysis,” Journal of Microscopy 205(2), 2002.

[175] N. O. Peterson, “Scanning fluorescence correlcation spectroscopy.

i. theory and simulation of aggregation measuremets.,” Biophysical

Journal 49, pp. 809–815, 1986.

[176] P. Wiseman and N. Petersen, “Image correlation spectroscopy. 2.

optimization for ultrasensitive detection of preexisting platelet-derived

growth factor-beta receptor oligomers on intact cells.,” Biophysical

Journal 76(2), pp. 963–977., 1999.

[177] T. Meyer and H. Schindler, “Particle counting by fluorescence correlation

spectroscopy. simultaneous measurement of aggregation and diffusion of

molecules in solutions and in membranes.,” Biophysical Journal 54(6),

pp. 983–993., 1988.

Chapter 6 147

[178] A. Palmer and N. Thompson, “Molecular aggregation characterized

by high order autocorrelation in fluorescence correlation spectroscopy,”

Biophysical Journal 52(2), pp. 257–270, 1987.

[179] H. Qian, “On the statistics of fluorescence correlation hong qian

spectroscopy,” Biophysical Chemistry 38, pp. 49–57, 1990.

[180] M. Sergeev, “High order autocorrelation analysis in image correlation

spectroscopy,” PhD Thesis, McGill University, Canada A, Montreal,

Quebec , 2004.

[181] M. Sergeev, S. Costantino, and P. Wiseman, “Measurement of monomer

oligomer distributions via fluorescence moment image analysis,”

Biophysical Journal 91(10), pp. 3884–3896, 2006.

[182] N. Peterson, C. Brown, A. Kaminski, J. Rocheleau, M. Srivastava, and

P. Wiseman, “Analysis of membrane protein cluster densities and sizes

in situ by image correlation spectroscopy,” Faraday Discuss 111(0),

pp. 289–305, 1998.

[183] G. Min, “Advanced optical imaging theory,” Springer Series in Optical

Sciences , p. 143, 1999.

[184] A. J. E. Elizabeth and M. J. Thomas, “Imaging molecular interactions

in living cells by fret microscopy,” Current Opinion in Chemical

Biology 10(5), pp. 409 – 416, 2006.

[185] S. S. Vogel, C. Thaler, and S. V. Koushik, “Fanciful fret,” Science

Signaling 2006(331), pp. re2–re2, 2006.

[186] M. Janos and E. Michael, “Energy transfer methods for detecting

molecular clusters on cell surfaces,” 278, pp. 444– 462, 1997.

[187] T. C. Tisone and J. Drobek, “Diffusion in thin film ti au, ti pd, and ti

pt couples.,” J Vac Sci Technol .

Chapter 6 148

[188] R. B. Ian, W. W. Paul, and W. H. John, “Investigating membrane

protein dynamics in living cellsthis paper is one of a selection of papers

published in this special issue, entitled csbmcb - membrane proteins in

health and disease.,” Biochemistry and Cell Biology 84(6), pp. 825–831,

2006.

[189] A. H. A. Clayton, F. Walker, S. G. Orchard, C. Henderson, D. Fuchs,

J. Rothacker, E. C. Nice, and A. W. Burgess, “Ligand-induced dimer-

tetramer transition during the activation of the cell surface epidermal

growth factor receptor-a multidimensional microscopy analysis,” Journal

of Biological Chemistry 280(34), pp. 30392–30399, 2005.

[190] G. P. Arthur and L. T. Nancy, “Fluorescence correlation spectroscopy

for detecting submicroscopic clusters of fluorescent molecules in

membranes,” Chemistry and Physics of Lipids 50(3-4), pp. 253 –270,

1989.

[191] H. Qian and E. L. Elson, “Distribution of molecular aggregation by

analysis of fluctuation moments.,” Proceedings of the National Academy

of Sciences 87(14), pp. 5479–5483, 1990.

[192] P. Sengupta, K. Garai, J. Balaji, N. Periasamy, and S. Maiti, “Measuring

size distribution in highly heterogeneous systems with fluorescence

correlation spectroscopy.,” Biophysical Journal 84(3), pp. 1977–1984,

2003.

[193] D. Magde, E. Elson, and W. W. Webb, “Thermodynamic fluctuations

in a reacting system—measurement by fluorescence correlation spectro-

scopy,” Physical Review Letters 29, pp. 705–708, Sep 1972.

[194] S. Hess, S. Huang, A. Heikal, and W. Webb, “Biological and chemical

applications of fluorescence correlation spectroscopy: a review.,”

Biochemistryl 41(3), pp. 697–705, 2002.

Chapter 6 149

[195] M. A. Medina and P. Schwille, “Fluorescence correlation spectroscopy for

the detection and study of single molecules in biology,” BioEssays 24(8),

pp. 758–764, 2002.

[196] F. K. Dennis, E. Elliot L, and P. S. Michael, “Weak dependence of

mobility of membrane protein aggregates on aggregate size supports a

viscous model of retardation of diffusion,” Biophysical Journal 76(1),

pp. 314 –322, 1999.

[197] N. Thompson, “Fluorescence correlation spectroscopy. in j. r. lakowicz,

editor, topics in fluorescence spectroscopy,” Techniques. Plenum

Press,New York 1, 1991.

[198] R. Kask, Pand Gunter and P. Axhausen, “Statistical accuracy in

fluorescence fluctuation experiments.,” European Biophysical Journal 25,

pp. 163–169, 1997.

[199] N. Petersen, P. Hoddelius, P. Wiseman, O. Seger, and K. Magnusson,

“Quantitation of membrane receptor distributions by image correlation

spectroscopy concept and application,” Biophysical Journal 65(3),

pp. 1135–1146, 1993.

[200] J. A. Yang, H. T. Phan, S. Vaidya, and C. J. Murphy, “Nanovacuums:

Nanoparticle uptake and differential cellular migration on a carpet of

nanoparticles,” Nano Letters 13(5), pp. 2295–2302, 2013.

[201] P. Wiseman and N. Petersen, “Image correlation spectroscopy

optimization for ultrasensitive detection of preexisting platelet derived

growth factor beta receptor oligomers on intact cells,” Biophysical

Journal 76(2), pp. 963–977, 1999.

[202] H. W. Sheng, L. Chia Wei, C. Arthur, and W. Pei Kuen, “Size dependent

endocytosis of gold nanoparticles studied by three dimensional mapping

Chapter 6 150

of plasmonic scattering images,” Journal of Nanobiotechnology 8(33),

2010.

[203] H. Heather, D. Nicole, T. J. Arwyn, H. Hanno, G. Hamidreza, and

L. Claus Michael, “Nanoparticle geometry and surface orientation

influence mode of cellular uptakeshape and orientation matter for the

cellular uptake of nonspherical particles,” Nano Letters 7(3), pp. 1961–

1973, 2013.

[204] P. Yang, X. Sun, J. Chiu, H. Sun, and He, “Transferrin mediated gold

nanoparticle cellular uptake,” Bioconjugate Chemistry 16, pp. 494–496,

2005.

[205] F. O. Harush, N. Debotton, S. Benita, and Y. Altschuler, “Targeting of

nanoparticles to the clathrin mediated endocytic pathway,” biochemical

Biophysical Research Communications 353, pp. 26–32, 2007.

[206] Y. Aoyama, T. Kanamori, T. Nakai, T. Sasaki, S. Horiuchi, S. Sando,

and T. Niidome, “Artificial viruses and their application to gene delivery

size controlled gene coating with glycocluster nanoparticles,” Jounal of

American Chemical Society 125, pp. 3455–3457, 2003.

[207] P. Lawrence, Fernando, K. Kandel, J. Yu, J. McNeill, P. Ackroyd,

and A. Christensen, “Artificial viruses and their application to gene

delivery size controlled gene coating with glycocluster nanoparticles,”

Biomacromolecules 11, pp. 2675–2682, 2010.

[208] C. Tregidgo, J. A. Levitt, and S. K, “Effect of refractive index on the

fluorescence lifetime of green fluorescent protein,” Journal of Biomedical

Optics 13, pp. 031218–8, 2008.

[209] J. M. Beuthan, J. H. Helfmann, and G. M Mller, “The spatial variation

Chapter 6 151

of the refractive index in biological cells,” Physical Medical Biology 41,

p. 369, 1996.

[210] R. Meyer, “Light scattering from biological cells dependence of

backscatter radiation on membrane thickness and refractive index,”

Applied Optics 18, pp. 585–588, 1979.

[211] S. Johnsen and E. Widde, “The physical basis of transparency in

biological tissue ultrastructure and the minimization of light scattering,”

Journal of Theor otical Biology 199, pp. 181–198, 1999.

[212] A. C. Curry and W. A. M, “Molecular imaging of epidermal growth

factor receptor in live cells with refractive index sensitivity using dark

field microspectroscopy and immunotargeted nanoparticles,” Journal of

Biomedical Optics 13, p. 014022, 2008.

[213] M. J. Crow, K. Seekell, J. H. Ostrander, and A. Wax, “Monitoring of

receptor dimerization using plasmonic coupling of gold nanoparticles,”

ACS Nano 5(11), pp. 8532–8540, 2011.

Chapter 6 152

Author publications

Journals

• Timothy TY Chow, Abu S. M. Mohsin and James W.M. Chon, “Im-

age Correlation Spectroscopy of Plasmon Coupled Gold Nanoparticles

for Probing Oligomerisation into Dielectric Medium”, (manuscript in

preparation expected to submit ACS Nano).

• Abu S. M. Mohsin, Timothy TY Chow, Chiara Paviolo and James

W.M. Chon, “Nanoparticle Uptake Kinetics and Aggregation Dynamics

of Different Functionalized Nanoparticles into HeLa Cells Using - High

Order Image Correlation Spectroscopy”, (manuscript in preparation

expected to submit ACS Nano).

• Adam B. Taylor, Pierrette Michaux, Abu S. M. Mohsin, and James

W. M. Chon, “Electron - beam lithography of plasmonic nanorod arrays

for multilayered optical storage”, Optics Express, March, 2014.

• Reda Kubiliute, Ksenia A Maximova, Alireza Lajevardipour, Jiawey

Yong, Jennifer S Hartley, Abu S. M. Mohsin, Pierre Blandin, James

W. M. Chon, Marc Sentis, Paul R Stoddart, Andrei Kabashin, Ricardas

Rotomskis, Andrew H.A. Clayton and Saulius Juodkazis, “Ultra -

pure, water - dispersed Au nanoparticles produced by femtosecond laser

ablation and fragmentation”, International Journal of Nanomedicine, 8,

153

Chapter 6 154

2601-2611(2013); doi; 10.2147/IJN.S44163.

Conferences

• Abu S. M. Mohsin*, Timothy T. Y. Chow, Chiara Paviolo, Andrew

H.A. Clayton and James W. M. Chon, “Cell uptake and aggregation

dynamics study of gold nanoparticles using image correlation spectro-

scopy”, Australian Institute of Physics Congress, Australian National

University, Canberra, Australia, 7th -11th December 2014.

• Arif Siddique, Timothy TY Chow, Abu S. M. Mohsin* and James

W.M. Chon, “The effect of Ti adhesion layers on the plasmonic properties

of gold nanorod arrays”, IONS - KOALA, 2014, Adelaide, Australia.

• Abu S. M. Mohsin, Timothy TY Chow, Chiara Paviolo and James

W.M. Chon, “Plasmonic gold nanoparticle aggregation characterisation

using high - order image correlation spectroscopy for cellular uptake

study”, 5th International Nano Medicine Conference, 30 June – 2 July

2014, Coogee Beach, Sydney, Australia.

• Adam B.Taylor , Abu S. M. Mohsin, Pierrette Michaux and James

W.M. Chon, “Electron - beam lithography of plasmonic nanorod arrays

for multilayered optical storage”, ANN Early Career Workshop, July

2014.

• Abu S. M. Mohsin*, Andrew H.A. Clayton and James W. M.

Chon, “Plasmon Coupling of Gold Nanoparticles and Investigating

Dimerization through Single Particle Spectroscopy of Weakly Coupled

Nanoparticle into Random Medium”, The international OSA Network of

Students (IONS), NA# 7, Charlotte, North Carolina, USA, 2nd - 4th

October 2013.

Chapter 6 155

• Timothy T. Y. Chow, Abu S. M. Mohsin *, Andrew H.A. Clayton

and James W. M. Chon, “Image Correlation Spectroscopy of Weakly

Coupled Plasmonic Gold Nanoparticles”, The International Conferences

on Surface Plasmon Photonics SPP6, Ottawa, Canada, 26th - 31st May,

2013.

• Abu S. M. Mohsin*, Timothy T. Y. Chow, Andrew H.A. Clayton

and James W. M. Chon, “Plasmon Coupling of Gold Nanoparticles for

Probing Membrane Proteins - Image Correlation Spectroscopy Study”,

20th Australian Institute of Physics Congress, University of New South

Wales, Sydney, Australia, 9th - 13th December 2012.