Raman Characterization of Colloidal Nanoparticles
using Hollow-Core Photonic Crystal Fibers
by
Jacky Siu Wai Mak
A thesis submitted in conformity with the requirements
for the degree of Master of Applied Science
Graduate Department of Electrical and Computer Engineering
University of Toronto
© Copyright by Jacky Siu Wai Mak 2011
ii
Raman Characterization of Colloidal Nanoparticles using
Hollow-Core Photonic Crystal Fiber
Jacky Siu Wai Mak
Master of Applied Science
Graduate Department of Electrical and Computer Engineering
University of Toronto
2011
Abstract
This Masters thesis investigates the ligand–particle binding interactions in the thiol–capped
CdTe nanoparticles and dye adsorbed gold nanoparticles. In the CdTe nanoparticles, Raman
modes corresponding to the CdTe core, thiol ligand and their interfacial layers were observed
and correlated to the different nanoparticle properties. To the best of our knowledge, this is the
first time that such strong Raman modes of the thiol-capped nanoparticles in aqueous solution
have been reported. In the gold nanoparticle systems, gold–citrate binding interactions were
observed as well as adsorption of the Raman dyes and binding with the polyethyleneglycol
polymer coating and phospholipid coating. These observations coincided with findings from
conventional optical techniques. In addition, gold nanoparticles were found to carbonize at high
pump power and prolonged exposure time. In summary, the two nanoparticle characterizations
demonstrated the high sensitivity and nondestructive nature of the photonic crystal fiber for
Raman spectroscopy.
iii
Acknowledgments
First, I would like to thank my supervisor, Professor Amr S. Helmy, for the opportunity to learn
about Raman spectroscopy during my summer research three years ago and the opportunity to
work on this very exciting project for my Master’s degree. He has continuously supported my
work and given me much freedom and insight to explore the many difficult problems in my
experiments throughout the degree. More importantly, he had provided me with tremendous
support during the difficult times I had with personal problems at home. I truly appreciate all of
his support and efforts throughout these two years.
In addition, I would like to thank numerous members in our group who have also helped me with
my research work. I want to thank Dr. Abdiaziz Farah for synthesizing the CdTe nanoparticles
and aiding me with the many chemistry questions I had in nano-materials. I want to thank Peter
Chan for confirming the Raman results on the CdTe nanoparticles. I want to thank Basil
Eleftheriades for the help with the photonic crystal fiber splicing experiments in measuring the
core diameter with SEM. Furthermore, I would like to thank Dr. Abdiaziz Farah, Dr. Fatima
Eftekhari, Steve Rutledge, Wen Ma and Dr. Rashid Abu-Ghazalah for the many discussions
about my experiments and the many useful comments on my manuscripts. Without the help of
my colleagues, my work would not have been accomplished in such a timely manner.
Secondly, I would like to thank Professor Gang Zheng, Professor Gilbert Walker and their group
members for providing me with the many nanoparticle samples for our study. In particular, I
would like to thank Christina MacLauglin and Glenda Sun in Professor Walker’s group for
providing the pegylated nanoparticles. In addition, I would like to thank Natalie Tam and Sophie
iv
Wang in Professor Zheng’s group for providing the phospholipid nanoparticles. Furthermore, I
thank Vivo Nano Inc. for providing the polymer encapsulated gold and silver nanoparticles.
Finally, I would like to express my love and affection to my friends, family and my girlfriend,
Sandy Chau, for their continuous love and support throughout my study. The many long hour
chats were very encouraging and have guided me through my tough times. I would like to thank
Sandy in particular for her efforts in helping me relax and enjoy life out of my busy schedule as
well as her assistance in editing my writings. Without everyone’s support, the completion of this
work would have never been possible.
v
Table of Contents
1. Introduction
1.1 Motivation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
1.2 Thesis Objectives . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
1.3 Organization of Thesis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
2. Raman Spectroscopy
2.1 Light-Matter Interaction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
2.2 Classical Treatment of Raman Scattering . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
1
1
5
5
8
9
10
2.3 Quantum Mechanical Treatment of Raman Scattering . . . . . . . . . . . . . . . . . . . . . . . .
2.4 Raman System Setup . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
2.5 Raman Scattering Efficiency . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
3. Photonic Crystal Fibers
3.1 Classifications of Photonic Crystal Fibers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
3.2 Index Guiding Mechanism . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
3.3 Photonic Bandgap Guiding Mechanism . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
12
15
19
21
23
25
30
vi
3.4 Hollow-Core Photonic Crystal Fiber for Efficient Raman Scattering . . . . . . . . . . . . .
4. Raman Characterization of Colloidal CdTe "anoparticles
4.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
4.2 Experimental Details . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
4.3 CdTe Core and Te Defects . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
4.4 Core-Thiol Interface . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
4.5 Carboxylate-Metal Complexes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
4.6 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
5. Raman Characterization of Metallic "anoparticles
5.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
5.2 Experimental Details . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
5.3 Molecular Interactions in Pegylated Gold Nanoparticles . . . . . . . . . . . . . . . . . . . . . .
5.3.1 Citrate Stabilized Gold Nanoparticles . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
5.3.2 Pegylated Gold Nanoparticles with MGITC Adsorbed . . . . . . . . . . . . . . . . . . .
5.3.3 Pegylated Gold Nanoparticles . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
5.4 Molecular Interactions in Phospholipid Gold Nanoparticles . . . . . . . . . . . . . . . . . . . .
5.4.1 Citrate Stabilized Gold Nanoparticles . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
5.4.2 CV Adsorbed Gold Nanoparticles . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
5.4.3 Phospholipid Gold Nanoparticles Adsorbed with CV . . . . . . . . . . . . . . . . . . . .
5.5 Pump Power and Duration Dependence of Pegylated Gold Nanoparticles . . . . . . . . .
5.6 Pump Power and Duration Dependence of Phospholipid Gold Nanoparticles . . . . . .
5.7 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
34
41
41
44
49
52
54
58
59
59
66
70
70
75
77
81
81
82
84
86
95
99
vii
6. Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
6.1 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
6.2 Future Work . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Appendix A Snell’s Law in Wave Optics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Appendix B Synthesis Method and Materials for Thiol-Capped CdTe Quantum Dots
Appendix C "anoparticle Signal Enhancement using HC-PCF . . . . . . . . . . . . . . . . . . .
Appendix D Raman Mode Assignments of Thiol-Capped CdTe "anoparticles . . . . . .
Appendix E Raman Mode Assignments of Gold "anoparticle Systems . . . . . . . . . . . .
Appendix F Fiber Splicing Program Optimization for Large Biological Molecules . . .
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
101
101
104
106
108
110
117
119
124
131
viii
List of Tables
Table 2.1: Laser specifications. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15
Table 2.2: Output powers of a 17 mW laser passing through the different optical density
filters. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16
Table 2.3: Specifications of the Olympus M PLAN objectives. . . . . . . . . . . . . . . . . . . . . . . 16
Table 2.4: Scanning range and resolution of diffraction gratings. A 1000µm confocal hole
size is used as a reference for resolution comparison. . . . . . . . . . . . . . . . . . . . . . 17
Table 2.5: Approximate order of magnitude for cross-sections (per molecule), σ, of the
different optical processes. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20
Table 4.1: Intensity ratios of the Te A1 mode to CdTe LO mode for CdTe QDs capped
with different thiol agents. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 51
Table 4.2: The wavenumber separations, ∆, between the symmetric and asymmetric
stretching modes of the carboxylate, and their corresponding carboxylate-metal. 57
Table 4.3: Intensity ratios of the symmetric carboxylate stretch to CdS 2SO mode. . . . . . . 58
Table 5.1: Nanoparticle samples studied for pegylated nanoparticle system and
phospholipid nanoparticle system. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 69
ix
Table 5.2: Wavenumber position of the symmetric and asymmetric stretching carboxylate
modes of citrate, their intensity ratio to νas(COO), respective wavenumber
differences and proposed complexes for a fresh batch of citrate stabilized gold
nanoparticles. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 73
Table 5.3: Wavenumber position of the symmetric and asymmetric stretching carboxylate
modes of citrate, their intensity ratio to νas(COO), respective wavenumber
differences and proposed complexes for a six-month old batch of citrate
stabilized nanoparticles. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 75
Table 5.4: Wavenumber position of the symmetric and asymmetric stretching carboxylate
modes, their respective wavenumber differences and proposed complexes in
pegylated gold nanoparticles. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 80
Table 5.5: Wavenumber position of the symmetric and asymmetric stretching carboxylate
modes, their respective wavenumber differences and proposed complexes for a
fresh batch of citrate stabilized gold nanoparticles for synthesizing the
phospholipid nanoparticle system. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 82
Table C.1: Enhancement factor of MGITC modes using HC-PCF. . . . . . . . . . . . . . . . . . . . 113
Table D.1: Proposed assignment of thiol-capped CdTe nanoparticles. Spectra are shown in
Figure 4.6 for regime between 100 and 200 cm-1
, Figure 4.8 for regime between
200 and 310 cm-1
and Figure 4.9 for regime between 310 and 1750 cm-1
. . . . . . 118
Table E.1: Proposed assignment of citrate stabilized Gold nanoparticles for pegylated
nanoparticle system. Spectrum is shown in Figure 5.5. . . . . . . . . . . . . . . . . . . . . 120
Table E.2: Proposed assignment of pegylated gold nanoparticles. Spectrum is shown in
Figure 5.9. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 121
Table E.3: Proposed assignment of MGITC adsorbed pegylated gold nanoparticles.
Spectrum is shown in Figure 5.7. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 122
x
Table E.4: Proposed Assignment of citrate stabilized Gold nanoparticles for phospholipid
nanoparticle system. Spectrum is shown in Figure 5.11. . . . . . . . . . . . . . . . . . . . 122
Table E.5: Proposed Assignment of CV adsorbed gold nanoparticles. Spectrum is shown in
Figure 5.12b. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .123
xi
List of Figures
Figure 2.1: Possible outcomes of light-matter interaction. . . . . . . . . . . . . . . . . . . . . . . .
Figure 2.2: Three possible energy transitions of molecule during the light scattering
process. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Figure 2.3: Schematic of the LabRam HR 800 Raman system. . . . . . . . . . . . . . . . . . . .
Figure 3.1: Wave interferences in the photonic bandgap of a one dimensional periodic
material. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Figure 3.2: Cross-section of a HC-PCF taken by scanning electron microscope (SEM).
Figure 3.3: Classification of PCFs based on their main light guiding mechanism. . . . . .
Figure 3.4: A ray of light incident at the interface between two dielectric materials and
refracted according to the Snell’s law. . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Figure 3.5: Dispersion diagram of PCF showing regions of light guidance by TIR and
photonic bandgap effect. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Figure 3.6: Dispersion diagram of a water-core PCF.Figure 3.6: Dispersion diagram of
a water-core PCF. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
9
12
16
22
23
25
26
28
30
xii
Figure 3.7: Dispersion diagram of a photonic crystal as a function of (a) out-of-plane
wave vector, kz, and (b) in-plane wave vector at kza/2π = 1.7 in the
irreducible Brillouin zone (inset) for the triangular lattice of air holes with
period a and radius 0.47a in ε = 2.1. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Figure 3.8: Schematic illustration of solid angles falling within the photonic bandgap
of a photonic crystal with a triangular lattice operating at a fixed frequency.
Figure 3.9: Raman scattering and signal collection in (a) conventional Raman
spectroscopy and (b) Raman spectroscopy using HC-PCF. . . . . . . . . . . . . .
Figure 3.10: Raman intensity of water mode centered at around 3400 cm-1
for varying
lengths of (a) HC-800 PCF and (b) TCT. . . . . . . . . . . . . . . . . . . . . . . . . . .
Figure 3.11: Comparison of silica Raman signals between water and thiol-capping
agent when they are filled into the central core of a HC-PCF. . . . . . . . . .
Figure 4.1: (a) Core-shell structure of CdTe QDs. (b) Molecular structure of capping
ligands. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Figure 4.2: Experimental setup of HC-PCF in HR800 micro-Raman system. . . . . . . . .
Figure 4.3: Raman spectrum obtained when the laser light is focused onto the cladding,
holey region and the center of the central core of the HC-PCF when it is
filled with water. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Figure 4.4: Optical images of HC-PCF cross-section at the (a) fused end and (b) non-
fused end after the entire central air core is filled with water. . . . . . . . . . . .
Figure 4.5: Raman intensity comparison of the OH modes of water between
conventional Raman spectroscopy using a cover slide, selective filling of
PCF and non-selective filling of PCF. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Figure 4.6: Raman spectra of CdTe QDs with different thiol capping agents in the
lower Raman shift regime between 100 and 200 cm-1
. . . . . . . . . . . . . . . . . .
33
33
35
38
39
45
46
46
48
48
50
xiii
Figure 4.7: (a) UV-vis spectra of the CdTe QDs (inset: fluorescent images of the QD
solutions). (b) PL profile of the QDs. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Figure 4.8: Raman spectra of CdTe QDs with different thiol capping agents in the mid-
Raman shift regime between 200 and 310 cm-1
. . . . . . . . . . . . . . . . . . . . . . .
Figure 4.9: Raman spectra of CdTe QDs with different thiol capping agents in the high
Raman shift regime between 310 and 1750 cm-1
. . . . . . . . . . . . . . . . . . . . . .
Figure 5.1: Schematic of plasmon resonance for a metal sphere. . . . . . . . . . . . . . . . . . .
Figure 5.2: Molecular structure of (a) pegylated gold nanoparticle system (b)
phospholipid gold nanoparticle system. . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Figure 5.3: Scheme of the nanoparticle synthesis process for (a) pegylated and (b)
phospholipid gold nanoparticles. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Figure 5.4: Raman spectra of water obtained using cover slide, HC-800 and HC-1060.
Enhancement of the OH stretching mode at 3421 cm-1
is about 101 times
and 112 times relative to the conventional technique respectively. . . . . . . .
Figure 5.5: Raman spectrum of a fresh batch of gold nanoparticles stabilized by citrate.
Inset: enlarged spectrum showing carboxylate stretching modes of citrate. .
Figure 5.6: Raman spectrum of six-month-old gold nanoparticles stabilized by citrate.
Inset: enlarge spectrum showing carboxylate stretching modes of citrate. . .
Figure 5.7: Raman spectrum of pegylated gold nanoparticles adsorbed with MGITC
obtained with 0.17 mW. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Figure 5.8: Raman spectrum of pegylated gold nanoparticles with MGITC adsorbed
and pure MGITC solution before background subtraction (a) and after
background subtraction (b). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
52
53
55
60
67
68
70
71
74
76
76
xiv
Figure 5.9: Raman spectrum of pegylated gold nanoparticles. Inset left: Raman mode
of Au–Cl at 259 cm-1
and Au–S at 305 cm-1
. Inset right: Stretching modes
of carboxylates. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Figure 5.10: UV–vis (a), TEM (b) and DLS (c) results obtained from a pegylated
nanoparticle system synthesized using the same method and materials. . .
Figure 5.11: Raman spectrum of 60 nm gold nanoparticles stabilized by citrate. Inset:
enlarged spectrum showing carboxylate stretching modes of citrate. . . . .
Figure 5.12: Raman spectrum of CV adsorbed gold nanoparticles and CV solution
without background subtraction (a) and with background subtraction (b).
Figure 5.13: Raman spectrum of CV adsorbed gold nanoparticles before and after
coating of phospholipid layer. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Figure 5.14: UV–vis spectra of the phospholipid nanoparticles throughout the synthesis
process. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Figure 5.15: Raman spectra of MGITC absorbed pegylated gold nanoparticles with
increasing power from 0.17 mW (a) to 1.7 mW (b), 4.3 mW (c), 8.5 mW
(d) and 17 mW (e). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Figure 5.16: Optical image of nanoparticle clusters around the inner wall of the PCF
core after about 1 minute of selective filling. . . . . . . . . . . . . . . . . . . . . . . .
Figure 5.17: Absorption of MGITC. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Figure 5.18: Raman spectra of citrate stabilized gold nanoparticles with increasing
power from 0.17 mW (a) to 1.7 mW (b), 4.3 mW (c), 8.5 mW (d) and 17
mW (e). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Figure 5.19: Raman spectra of polymer encapsulated gold and silver nanoparticles
obtained from Vivo Nano Inc. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
78
79
82
83
85
86
88
90
90
92
94
xv
Figure 5.20: Raman spectra of CV adsorbed gold nanoparticles with increasing power
from 0.17 mW (a) to 1.7 mW (b), 4.3 mW (c), and 17 mW (d). . . . . . . . .
Figure 5.21: Raman spectra of citrate stabilized gold nanoparticles, used for
phospholipid nanoparticle synthesis, with increasing power from 0.17
mW (a) to 1.7 mW (b), 4.3 mW (c), and 8.5 mW (d). . . . . . . . . . . . . . . . .
Figure A.1: Directions of the electric field, magnetic field and wave vector when a TE
field is incident at the interface between two dielectric materials. . . . . . . .
Figure C.1: Raman spectrum of citrate stabilized gold nanoparticles obtained using
cover slide and HC PCF with 633 nm laser. . . . . . . . . . . . . . . . . . . . . . . . .
Figure C.2: Raman spectrum of pegylated gold nanoparticles obtained using cover
slide and HC-PCF with 633 nm laser. . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Figure C.3: Raman spectra of MGITC adsorbed gold nanoparticles obtained using
cover slide and HC-PCF with 633 nm laser. . . . . . . . . . . . . . . . . . . . . . . .
Figure C.4: Raman spectra of citrate gold nanoparticles obtained using cover slide and
HC-PCF with 633 nm laser. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Figure C.5: Raman spectrum of CV adsorbed gold nanoparticles obtained using cuvette
with 785 nm laser, cover slide and HC-PCF with 633 nm laser. . . . . . . . . .
Figure C.6: Raman spectrum of phospholipid gold nanoparticles obtained using cuvette
with 785 nm, cover slide and HC-PCF with 633 nm laser. Phospholipid
was formed using (a) MHPC and DMPC and (b) SHPC and DMPC. . . . . .
Figure F.1: SEM image of the HC-1060 before (a) and after (b) fiber splicing to
collapse the cladding holes. The central core of the PCF is reduced to about
7 µm after the splicing process. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Figure F.2: Comparison of water modes with HC-800, HC-1060, HC19-1550 and
cover slide. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
96
98
107
111
112
113
114
115
116
125
125
xvi
Figure F.3: SEM image of PCF facet after splicing with program modified from that
for HC-800. The central core of the PCF after splicing is about 17 – 18 µm
and all of the cladding holes are closed except the three corner holes. . . . .
Figure F.4: SEM image of PCF fused with fiber–electrode distance at (a) 30, (b) 50, (c)
70, (d) 80, and (e) 90 µm. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Figure F.5: SEM image of three PCFs fused HC15-1990 PCFs at 50 µm from
electrode. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Figure F.6: Optical image of HC19-1550 PCF selectively filled with lung cancer cells.
The bright light at the central core shows the cancer cell solution. . . . . . . .
126
128
129
130
xvii
List of Appendices
Appendix A Snell’s Law in Wave Optics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Appendix B Synthesis Method and Materials for Thiol-Capped CdTe Quantum Dots
Appendix C "anoparticle Signal Enhancement using HC-PCF . . . . . . . . . . . . . . . . . . .
C.1 Citrate Stabilized Gold Nanoparticles for Pegylated System . . . . . . . . . . . . . . . . . .
C.2 Pegylated Gold Nanoparticles . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
C.3 Pegylated Gold Nanoparticles with MGITC Adsorbed . . . . . . . . . . . . . . . . . . . . . . .
C.4 Citrate Stabilized Gold Nanoparticles for Phospholipid System . . . . . . . . . . . . . . .
C.5 Gold Nanoparticles with CV Adsorbed . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
C.6 Phospholipid Gold Nanoparticles with CV Adsorbed . . . . . . . . . . . . . . . . . . . . . . .
Appendix D Raman Mode Assignments of Thiol-Capped CdTe "anoparticles . . . . . .
Appendix E Raman Mode Assignments of Gold "anoparticle Systems . . . . . . . . . . . . .
E.1 Citrate Stabilized Gold Nanoparticles for Pegylated System . . . . . . . . . . . . . . . . . .
E.2 Pegylated Gold Nanoparticles . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
E.3 Pegylated Gold Nanoparticles with MGITC Adsorbed . . . . . . . . . . . . . . . . . .. . . . . .
106
108
110
110
111
112
113
114
115
117
119
119
120
121
xviii
E.4 Citrate Stabilized Gold Nanoparticles for Phospholipid System . . . . . . . . . . . . . . . . . .
E.5 Gold Nanoparticles with CV Adsorbed . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
E.6 Phospholipid Gold Nanoparticles with CV Adsorbed . . . . . . . . . . . . . . . . . . . . . . . . . .
Appendix F Fiber Splicing Program Optimization for Large Biological Molecules . . .
122
122
123
124
xix
List of Acronyms
CCD — Charge-Coupled Device
CV — Crystal Violet
CW — Continuous Wave
CZE — Capillary Zone Electrophoresis
DLS — Dynamic Light Scattering
DMPC — 1,2-Dimyristoyl-sn-Glycero-3-Phosphocholine (C36H72NO8P)
EL — Electroluminescence
EM — Electromagnetic
FDTD — Finite-Difference Time-Domain
FT-IR — Fourier-Transform Infrared Spectroscopy
FWHM — Full-Width at Half Maximum
HC–PCF — Hollow-Core Photonic Crystal Fiber
HRTEM — High Resolution Transmission Electron Microscope
KKR — Korringa-Kohn-Rostoker
MGITC — Malachite Green Isothiocyanate
MHPC — 1-Myristoyl-2-Hydroxy-sn-Glycero-3-Phosphocholine (C22H46NO7P)
MOF — Micro-Structured Optical Fiber
MPA — 3-Mercaptopropionic Acid
MSF — Micro-Structured Fiber
NA — Numerical Aperture
xx
PCF — Photonic Crystal Fiber
PEG — Polyethyleneglycol
PL — Photoluminescence/Photoluminescent
QD — Quantum Dot
SC–PCF — Solid-Core Photonic Crystal Fiber
SEM — Scanning Electron Microscope
SERS — Surface-Enhanced Raman Spectroscopy
SHPC — 1-Stearoyl-2-Hydroxy-sn-Glycero-3-Phosphocholine (C26H54NO7P)
SO — Surface Optical
SPP — Surface Plasmon Polariton
TCT — Telfon Capillary Tube
TE — Transverse Electric
TEM — Transmission Electron Microscope
TG — 1-Thioglycerol
TGA — Thioglycolic Acid
TIR — Total Internal Reflection
UV-Vis — Ultraviolet-visible Spectroscopy
XPS — X-Ray Photoelectron Spectroscopy
XRD — X-Ray Diffraction
xxi
List of Publications
Journal Article
(1) J. S. W. Mak, A. A. Farah, F. Chen, A. S. Helmy, “Photonic Crystal Fiber for Efficient
Raman Scattering of CdTe Quantum Dots in Aqueous Solution”, ACS Nano, vol. 5, pp
3823-3830 (2011).
International Conferences
(1) J. S. W. Mak, A. Farah, Feifan Chen, A. S. Helmy, “Efficient Raman Sensor for
Nanoparticles using Hollow Core Photonic Crystal Fiber,” OSA Sensors, Oral
Presentation, Toronto, June 12-15, 2011.
(2) J. S. W. Mak, A. Farah, Feifan Chen, Amr S. Helmy, “Probing Quantum Dot Cores,
Their Interfaces and Thiol Cappings Non-Destructively in Dilute Solution using Raman
Scattering in Hollow Core Photonic Crystal Fiber,” OSA/IEEE CLEO-QELS, Oral
Presentation, Baltimore, May 2-5, 2011.
(3) J. S. W. Mak, Abdiaziz A. Farah, Feifan Chen, A. S. Helmy, “Photonic Crystal Fiber for
Efficient Raman Scattering of Thiol-capped Quantum Dots in Aqueous Solution,” MRS
Spring Meeting, Oral Presentation, San Francisco, April 25-29, 2011.
xxii
Local Conferences
(1) J. S. W. Mak, F. Eftekhari, Abdiaziz A. Farah, S. Rutledge, A. S. Helmy, “Raman
Spectroscopy of Colloidal CdTe Quantum Dots using Teflon Capillary Tube,” Third
BiopSys All Network Meeting (not disclosed to public), Poster presentation, Toronto,
January 8-9, 2010.
(2) J. S. W. Mak, Abdiaziz A. Farah, F. Eftekhari, A. S. Helmy, “Raman Spectroscopy of
Thiol-Capped Colloidal CdTe Quantum Dots Using Capillary Tubes,” Fourth BiopSys
All Network Meeting (not disclosed to public), Poster presentation, Toronto, May 26-29,
2010.
(3) J. S. W. Mak, Abdiaziz A. Farah, F. Chen, A. S. Helmy, “Probing Quantum Dot Cores,
Their Interfaces and Thiol Cappings Non-Destructively in Dilute Solution using Raman
Scattering in Hollow Core Photonic Crystal Fiber,” Fifth BiopSys All Network Meeting
(not disclosed to public), Poster presentation, Toronto, January 6-7, 2010.
(4) J. S. W. Mak, N. Tam, S. Wang, G. Zheng, A. S. Helmy, “Low Power Raman Detection
of Lung Cancer Cells using Hollow-Core Photonic Crystal Fiber with Functionalized
Gold Nanoparticles,” Sixth BiopSys All Network Meeting (not disclosed to public),
Poster presentation, Toronto, June 16-17, 2010.
1
Chapter 1
Introduction
(1.1) Motivation
Nanotechnology has advanced significantly in the past several decades enabling innovative
solutions for many challenging problems in a wide field of applications. Colloidal nanoparticles
whose dimension in the nanometer range are one of the most innovative material in
nanotechnology over the past two decades as they exhibit unique material, optical, electrical and
thermal characteristics that are tuneable with the size, shape and composition of the particles.1
The surface of particles can also be functionalized with different dyes, polymers or lipids to
provide the necessary functions and properties required; for instance, water solubility, reduced
toxicity, signalling, biocompatibility and bioconjugation. Due to the large design variability,
many efforts have been put into improving the different synthesis methods or on designing the
particular nanoparticle system for an application. However, there have been little efforts in
improving the tools to characterize these novel systems in the nanometer scale.
2
One of the most important technological challenges in quantum dot (QD) advancement is the
development of a cost-effective, reliable and sensitive optical monitoring system to control the
physical, chemical and size-dependent properties of QDs before, during and after their
fabrication on a nanometer scale. Currently, many analytical techniques have been employed but
they rely on changes in the nanoparticles’ material compositions and the overall nanoparticle
properties. For example, photoluminescence (PL), electroluminescence (EL), and ultraviolet–
visible spectroscopy (UV-vis) provide changes in the luminescence and absorption of the
nanoparticles. Imaging techniques such as transmission electron microscope (TEM) or high
resolution transmission electron microscope (HRTEM) provides changes in the physical
geometry and composition. TEM, HRTEM, UV-vis as well as dynamic light scattering (DLS)
and capillary zone electrophoresis (CZE) provide quantitative information in the nanoparticles
size and size distributions. Other techniques such as X-ray diffraction (XRD), and X-ray
photoelectron spectroscopy (XPS) provide information only in the crystallinity and compositions
of the nanoparticles. None of these techniques is capable of describing the binding interactions of
molecules at the nanometer scale. Therefore, the impact of the interactions on the overall
molecular structure, molecular complex, and different QD properties remain unclear. In situ
monitoring of the nanoparticle synthesis also remains challenging. Consequently, this limits our
capability to improve the properties and functionalities further from what has been achieved
today. Complex QD designs for increasing performance and functionalities in different
applications remain very challenging as a result.
To observe detail binding interactions at the molecular level, conventional techniques involve
using Fourier-transform infrared spectroscopy (FT-IR) and Raman spectroscopy whose signals
contain “fingerprints” of the molecular structures. However, strong and broad vibrational modes
3
(signals) of water in FT-IR interfere with the weak vibrational modes of nanoparticles which
severely limit its capability to obtain information on the nanoparticles and the ligands. On the
other hand, water signal is very weak in the complementary technique, Raman spectroscopy.
However, Raman signals are also extremely weak in aqueous or dilute solutions. In situ
monitoring of the nanoparticles using conventional Raman spectroscopy therefore remained very
challenging because of the solution-based environments in which nanoparticles are produced.
Surface enhanced Raman spectroscopy (SERS) can be used to enhance the nanoparticle signals;
however, metallic nanoparticles have to be mixed with the QD solution, which could alter the
structure and properties of the QDs from its as-synthesized state.
In the past decade, many fiber based platforms have been developed to enhance the sensitivity of
conventional Raman spectroscopy for particle or molecule analysis through efficient induction
and generation of Raman scattering signals. For example, Teflon Capillary Tube (TCT) was
demonstrated to enhance the detected Raman signal of aqueous based solutions, such as dilute
toluene,2 sodium carbonate,
2 benzene
2 and creatinine solutions,
3 through confining the pump
laser in the same fiber core with the solution by total internal reflection (TIR). Hollow-core silica
optical fiber was demonstrated to enhance the Raman mode of dilute carbon tetrachloride
solutions also using the same confinement mechanism.4 In addition to fiber platforms that utilize
TIR for light confinement, photonic crystal fibers (PCFs) have been utilized to enhance the
Raman sensitivity in which light is being confined by the photonic bandgap effect. In particular,
solid-core PCF (SC-PCF) was utilized in which the sample solution is confined in the cladding
holes while the pump laser is confined weakly in the solid core. Raman signals are generated by
the evanescent wave that is overlapped with the solution. 4-Mercaptobenzoic acid,5 acetonitrile
6
and Rhodamine 6G solutions7 were characterized using this technique. However, there are
4
several drawbacks found including weaker interactions with evanescence wave, smaller cladding
holes limiting particle sizes and strong signal interferences from the silica fiber. Another type of
PCF is named hollow-core PCF (HC-PCF) in which the core of the PCF is hollow allowing
solutions to be filled inside and interact with the pump laser directly. Rhodamine B was tested
using this technique.8 The solution was filled into both the core and the cladding holes of the
PCF which shifted the bandgap of the fiber and guided the pump laser weakly by the photonic
bandgap effect. It is important to mention that the characterization of nanoparticles using these
techniques were not achieved due to the low Raman cross-section of nanoparticles and the
limited sensitivity of these fiber based platforms for Raman spectroscopy.
To enable molecular level monitoring of the nanoparticles, a novel technology utilizing HC-PCF
for sensitive Raman spectroscopy has recently been developed in our research group.9, 10
The
novelty of our technique is that we can selectively fill the sample solution into only the central
core of the PCF. This enables both the pump laser and the Raman scattering signal to be guided
by both TIR and photonic bandgap effects which allow efficient induction and collection of
Raman scattering signal from every single drop of solution available. This technique is capable
of enhancing the detected Raman scattering signal for about 2 – 3 orders of magnitude without
the need of adding or mixing extra chemicals or materials with the analyte for enhancements,
which is at least an order of magnitude greater than the other fiber based Raman platforms
described. Moreover, this technique retains both the nondestructive property of Raman
spectroscopy and the high flexibility and compactness of optical fibers which make it most
suitable for remote sensing and in-situ detections of nanoparticles in their native states. This
technique has been utilized successfully for characterizing ZnO and CdTe nanoparticles post-
synthesized with rapid thermal annealing.9, 10
5
(1.2) Thesis Objectives
The goal of this thesis is two-fold. First, it is to show that the use of HC-PCF with Raman
spectroscopy can enhance the Raman scattering signal obtained from a wider variety of colloidal
nanoparticles when compared to conventional Raman Spectroscopy. Second, with enhancements
of the detected Raman signal from the novel PCF platform, we want to improve our
understanding of semiconductor and metallic nanoparticles and their binding interactions with
different ligands at the molecular level.
In particular, this thesis aims to confirm the ligand-particle binding speculated in the thiol-
capped CdTe quantum dots and correlate this information to the different properties reported in
the literature. In addition, this thesis aims to study the ligand-particle binding in different gold
nanoparticle systems throughout the entire synthesis process, which has never been achieved
before with Raman spectroscopy due to the limited sensitivity of conventional Raman technique.
Furthermore, this thesis aims to achieve a real time Raman study of the gold nanoparticles at
different measurement conditions to show laser induced effects that limit many nanoparticle
studies with Raman spectroscopy.
(1.3) Organization of Thesis
The structure of this thesis is mainly consisted of two parts. The first part involves the
fundamentals and theoretical aspects of Raman spectroscopy (Chapter 2) and PCF (Chapter 3).
In chapter 2, classical and quantum mechanical treatments of Raman spectroscopy are reviewed
to offer insights into the important parameters governing the Raman effects. In particular, the
efficiency of the Raman scattering signal is discussed and compared with other optical effects.
These analyses demonstrate an important point that the detectable Raman scattering signal is an
6
extremely weak signal compare to other optical events. Thus, an effort to enhance the detected
Raman scattering signal is required to ultimately take advantages of its “fingerprint” features for
molecular analysis and identification. To realize the many efforts that companies have made to
achieve sensitive Raman spectroscopy up until today, the design and operation of one of the most
sensitive Raman system built by Horriba, namely the LabRAM HR800, is also discussed.
Chapter 3 covers the fundamentals and guiding mechanism in PCF. Both the index and bandgap
guiding mechanisms are discussed in detail to provide a theatrical understanding of the PCF
operation. In addition, section 3.4 presents our novel technique to enhance the detected Raman
signal utilizing the hollow core type of PCF. The improved light collection efficiency and
increased light-matter interactions from conventional Raman spectroscopy are analyzed
theatrically and experimentally to demonstrate their different contributions to the detected
Raman signal enhancement. The detected Raman signal enhancements with the PCF are also
compared to the liquid core fiber, namely TCT, to demonstrate the unique advantage of PCF’s
light guiding mechanisms. Furthermore, the advantages of fiber-based detection and metal-free
detection are briefly discussed.
The second part of this thesis discusses the experimental details and results on characterizing
thiol-capped CdTe nanoparticles (Chapter 4) and metallic nanoparticle systems (Chapter 5).
Chapter 4 presents the experimental setups and results on the Raman analysis of thiol-capped
CdTe nanoparticles. One very important feature of our PCF platform described in this chapter is
the technique to selectively fill the central core of the PCF with the liquid sample of interest.
This technique enables the pump laser light and the Raman scattering signals to be guided by
both the index and photonic bandgap confinement; which subsequently lead to an enhancement
7
of the detected Raman signal. A comparison of the Raman signal between conventional
techniques, non-selectively filled PCF and selectively filled PCF shown in Figure 4.7
demonstrates the large sensitivity improvement with our selective filling technique on PCF. This
improved sensitivity enables detailed molecular bindings between the ligand and the nanoparticle
core to be observed as well as vibrational modes of the CdTe core and its contained defects.
These modes are further correlated to the QD’s PL quantum efficiency, stability, water solubility
and bio-conjugation capability. To the best of our knowledge, this is the first time that such
strong Raman modes of the thiol-capped CdTe QDs in aqueous solution have been reported.
Chapter 5 presents the characterization results of two metallic nanoparticle systems for cancer
detection. Raman spectra of pegylated and phospholipid gold nanoparticles obtained at different
stages of the synthesis process are presented and correlated to the structural changes of the
nanoparticles. These results confirmed the stability of the stabilizer, binding of the Raman dye
and coating of the polymer and phospholipid on the nanoparticle surface. Further, these results
demonstrate that efficient Raman scattering of the PCF enables metallic nanoparticle systems to
be characterized through the entire synthesis process; thus, allowing nanoparticles to be
monitored in situ for quality control. As these nanoparticles are designed for cancer detections
with Raman imaging, this chapter also investigates, in real time, the stability of the two
nanoparticle systems under different pump power exposure and exposure durations. Structural
changes in the nanoparticle system are observed with the enhanced Raman signal using HC-PCF.
These changes show the power and exposure time limit in which the pump laser can be focused
onto the nanoparticle systems for non-destructive measurements.
Finally, chapter 6 summarizes the work discussed in this thesis and provides future directions.
8
Chapter 2
Raman Spectroscopy
Raman spectroscopy is a rapid and non-destructive means of probing molecular vibrations
optically through inelastic scattering. This technique provides the unique vibrational modes or
“fingerprints” of the molecules that enable similar molecular structures to be unambiguously
distinguished and identified. This high specificity of the vibrational information to chemical
bonds and structures has led Raman to become a powerful tool in chemistry and medicine since
its first discovery in 1928 by C. V. Raman and K. S. Krishnan.11
Since Raman is an optical
measuring technique, it would involve light interacting with matter. Therefore, we will start this
chapter by discussing the physics behind light-matter interaction; this will lead to the discussion
of the Raman effect in two scales – macroscopically through classical physics, and
microscopically through quantum mechanics. We will then describe the experimental aspect of
Raman spectroscopy and examine the setup and operation of the Raman spectrometer used in
this experiment. Finally, the efficiency of Raman scattering will be discussed to provide an
9
overview of the Raman signal strength compared to other possible optical processes in
spectroscopy and between various molecules of interest in nanotechnology.
(2.1) Light-Matter Interaction
In physics, light is an electromagnetic (EM) radiation that lies in the range of the infrared, visible
light or ultraviolet in the EM spectrum. When light interacts with matters, it can be reflected,
absorbed, scattered, or passed straight through the material without any interaction (Figure 2.1).
If light is scattered by the material, it can be scattered either elastically or inelastically. An
elastically scattered EM radiation is generally called Rayleigh scattering. In Rayleigh scattering,
the EM radiation is scattered by particles that are much smaller than the EM radiation
wavelength upon interaction. The scattered light will have a wavelength that is identical to the
incident light; thus, energy is conserved during the scattering process. If the scattered light has a
wavelength that is different than the incident light, then there is an energy exchange between the
incident light and the material during the interaction. Energy of the incident light could be gained
or lost through the vibrational motion of the molecules in the material causing inelastic scattering
of light. This process is called Raman scattering. In the next two sections, we will look at this
Raman scattering process macroscopically using classical theory of physics and microscopically
using quantum mechanics.
Figure 2.1: Possible outcomes of light-matter interaction.
10
(2.2) Classical Treatment of Raman Scattering
On the macroscopic level, Raman scattering can be viewed as light emitted from oscillating
dipoles in a material induced from EM radiation. Theoretically, this phenomenon can be
described using classical physics as follows;11, 12
consider a monochromatic light (or an EM
wave) propagating in the z-direction with an oscillating electric field in the x-direction. The
strength of the electric field (Ex) at any time (t) can be expressed as:
= cos2 (2.1)
where is the maximum amplitude of the electric field and is the frequency of the
monochromatic light. When this light interacts with a material, its electric field will polarize the
cloud of electrons around each of the molecules in the material and displace them from their
equilibrium position. These electron displacements will oscillate with the electric field and
behave as an oscillating electric dipole. The oscillating electric dipole then emits light (the
scattered light) due to its oscillating electric current. Since the scattered light is dependent on the
oscillation of the electric dipole, changes of the dipole polarity will indicate how frequent the
dipole is oscillating; hence, the frequency of the radiating light. The polarity of a dipole is
determined by its dipole moment (µ):
= = cos2 (2.2)
where α is the polarizability tensor that describes the tendency of electron clouds distortion in the
presence of an external electric field. If the molecule is vibrating with a natural frequency (,
the nuclear displacement q can be written as:
= cos2 (2.3)
where is the vibrational amplitude. For small amplitude of vibration, α can be expressed as a
linear function of q using Taylor series:
11
= +
+ ⋯ (2.4)
Combining (2.2) with the first two terms of (2.4), the dipole moment can be expressed as:
= cos2 = cos2 +
cos2 = cos2 +
cos2 cos2
= cos2
+
cos!2 + " + cos!2 − "$ (2.5)
According to classical theory, the first term in (2.5) represents an oscillating dipole that radiates
light at the frequency which is the same frequency as the incident light; therefore this term
corresponds to Rayleigh scattering. In contrast, the second and third term represents an
oscillating dipole that radiates light at the frequency + and − respectively. These
radiating lights have frequency that is different than the frequency of the incident light; thus, they
are inelastically scattered or Raman scattered. The radiating light with frequency higher than that
of the incident light (i.e. + ) is referred to as anti-Stokes scattering. The radiated light
with frequency lower than that of the incident light (i.e. − ) is referred to as Stokes
scattering. It is important to note that both the Raman scattering terms can vanish if is zero
in certain vibrations. These vibrations are not Raman-active and cannot yield Raman scattering.
Thus, the rate of change of the polarizability must not be zero for vibrations to be Raman-active.
This sets the selection rule for determining which modes of vibration in a molecule yields the
Raman effect. It is not obvious from the equation, but is also implicitly smaller than ;
therefore, Raman scattering is much weaker than Rayleigh scattering, generally by three orders
of magnitude as we will discuss in section 2.5.
12
(2.3) Quantum Mechanical Treatment of Raman Scattering
On the microscopic level, Raman scattering can be described using a semi-phenomenological
model of quantum mechanics called the energy transfer model.11, 13
In this model, light can be
considered as a collection of a basic unit or “particle” called photon. The energy (E) of a photon
is proportional to the frequency of the monochromatic light (v) as follows:
= ℎ (2.6)
where h is the Planck’s constant. Molecules of the material that is interacting with light are
considered systems with many different energy states. The lowest possible state is named the
ground state, whereas all the other higher energy states are considered as the excited states.
When a collection of photons is incident onto a material and being scattered, molecules of the
material are virtually excited to a higher energy state from its initial energy state by the amount
of energy equal to the incident photon. Instantaneously, these excited molecules would return to
a lower state and emit a photon (the scattered light) with an energy that is lower, higher or equal
to the energy of the incident photon. Figure 2.2 shows the schematic of the three possible energy
transitions in the scattering phenomenon. Assuming the molecule interacting with light is
initially sat in its ground state. The molecule will be excited to a virtual state, an energy state that
lies between the electronic states of the molecule which does not actually exist, and descend
instantaneously to the lower energy states emitting a photon with energy that is equal to the
difference between the two energy states. If this lower energy state is the same as the initial state
before the transition occurred (i.e. the ground state in this case), then the emitted photon will
have the same energy as the initial photon; thus, the frequency of the scattered light is equal to
that of the incident light, namely the Rayleigh scattering. If the molecule returns to a energy state
higher than the initial state (i.e. an excited state rather than the ground state in this case), then the
13
emitted photon will have energy less than that of the initial photon; hence, scattering light with
frequency lower than that of the incident light which correspond to the Stokes lines of Raman
scattering. In order for the molecule to return to its initial equilibrium state or ground state, the
molecule will emit a phonon, which is a basic unit of vibrational energy, with energy equal to the
difference between the excited state and the final equilibrium state. The energy of the phonon is
also proportional to the vibrational frequency of the molecule by Planck’s constant (h) like
photon:
'() = ℎ'() (2.7)
If the molecule is initially sat in an excited state, the molecule might return to a state lower than
that of the initial state after it has been excited. In this case, the molecule will emit a photon with
energy greater than that of the incident photon; hence, the scattered light will have a frequency
greater than that of the incident light which corresponds to the anti-Stokes lines of Raman
scattering. The greater energy in the scattered photon is gained from the phonon that is initially
present in the molecule and retaining the molecule at an excited state primarily.
Figure 2.2: Three possible energy transitions of molecule during the light scattering process.
14
In the viewpoint of the energy transfer model, two characteristics of Raman scattering can be
deduced: first, the frequency of the Raman scattered light is relative to the electronic states of the
molecule. Hence, any perturbation field such as strain field, thermal excitation or chemical
potentials resulting from the surrounding environment can affect the Raman peak position.
Second, the number of photons scattered (also corresponding to the intensity of the scattered
light) in the Stokes and anti-Stokes lines are related to the initial state that the molecule is at.
According to Boltzmann’s distribution, most molecules are likely to be resided in the ground
vibrational state at room temperature prior to light interaction. Higher energy states will be
populated with an exponentially decreasing number of electrons as the energy of the state
increases. This implies that majority of the Raman scattered photons will be Stokes scattering
and so most Raman scattered photon will have a lower energy (and lower frequency) than the
incident photon.
It is important to emphasize that the molecule is only virtually excited and the excited state by
the incident photon is also a virtual state in scattering. The word virtual is used because such
excitation of the molecule never existed and the incident photon is never actually absorbed since
energy cannot conserve upon absorption of the incident photon when there is no actual excited
electronic state present in the molecule with energy equal to that of the incident photon. One may
also consider this process as an excitation with a zero lifetime in the virtual state; thus the
molecule will descend to a lower state instantaneously. However, if the incident photon energy
approaches electronic transition energy, a more intensity scattered light would be observed.
15
(2.4) Raman System Setup
The Raman system used in this thesis is a high resolution LabRam HR 800 system manufactured
by Horiba Jobin Yvon. This is a confocal Raman system that collects Raman spectra in the back-
scattering geometry. The main components of the system are as follows:
Laser An internal HeNe laser is equipped in the LabRam Raman system for sample excitation.
The specifications of the laser are summarized in the following table:
Table 2.1: Laser specifications.
Parameters Value
Wavelength 632.817 nm
Max. Power 17 mW
Polarization ratio 500:1
Density Filter Wheel A wheel of five different optical density (OD) filters is installed to
reduce the laser power at different levels. The OD of the filters ranges from 0.3 to 4. An option
of no filter is also available in the system. The power of the laser passing through an OD filter
(Pout) can be determined as follow:
*+, = *(-1001 (2.8)
where Pin is the laser power entering the OD filter. The output powers to the different OD filters
with the equipped 17 mW HeNe laser are listed in the table below:
16
Table 2.2: Output powers of a 17 mW laser passing through the different optical density filters.
Filter Output Power
(mW)
D0.3 8.5
D0.6 4.25
D1 1.7
D2 0.17
D3 0.017
D4 0.0017
Microscope Objectives Three Olympus M PLAN objectives are equipped in the Raman system
to focus the excitation light onto the sample. The magnifications, numerical apertures (NA), and
working distances of the three objectives are listed in the table below:
Table 2.3: Specifications of the Olympus M PLAN objectives.
Magnification NA Working Distance
(mm)
10x 0.25 10.6
50x 0.7 0.38
100x 0.9 0.21
"otch Filter A holographic notch filter is used to reject the excitation line < 120 cm-1
in the
Stokes edge.
Confocal Hole A confocal hole adjustable between 0 and 1000 µm is available through the
LabRam software to maximize the Raman scattering signal and adjust the field-of-depth in a
Raman measurement.
Spectrograph The HR800 is equipped with an asymmetric Czerny Turner spectrograph of two
different diffraction gratings (1800 g/mm and 600 g/mm) that separates an incoming light into a
frequency spectrum. The 1800 g/mm grating provides a higher wavenumber resolution in a
17
Raman measurement than the 600 g/mm grating, but also a relatively shorter scanning range per
scan. The scanning range and resolution of the two gratings are listed in the table below:
Table 2.4: Scanning range and resolution of diffraction gratings. A 1000µm confocal hole size
is used as a reference for resolution comparison.
Grating
(g/mm)
Possible Scanning
Range (nm)
Scanning Range
per Scan (cm-1
)
Resolution
(cm-1
)
600 0-2850 1241.2 ±2
1800 0-950 321.6 ±1
Detector A standard air cooled silicon charge-coupled device (CCD) detector is equipped in the
Raman system. The resolution of the detector is 1024 x 256 pixels of 26 microns.
Figure 2.3: Schematic of the LabRam HR 800 Raman system.14
18
The setup of the Raman system is illustrated in Figure 2.3. The control and operation of the main
system components are mainly done through the provided software called LabSpec. When
acquiring a Raman spectrum, the excitation laser is first reflected by a mirror to the density filter
wheel for power adjustment. The output laser is then reflected by two more mirrors to the notch
filter at an angle that enables all of the excitation light to be reflected and focused onto the
sample through a microscope objective. The scattered light from the sample will then be
collected back by the same objective and directed through the notch filter to remove all the
excitation light and Rayleigh scattered light. The remaining Raman scattered light will then be
reflected with the forth mirror to the spectrograph through a confocal hole, two focusing lens and
an entrance slit. The Raman scattered light will be separated into the different frequency
components in the spectrograph and the Raman spectrum will be outputted through the CCD
detector.
(2.5) Raman Scattering Efficiency
When light is incident on a material, various interactions can take place simultaneously as
discussed in section 2.1. Some of the light may be absorbed by the material; others may be
Rayleigh scattered. Some of it will even cause fluorescence, which emits light after absorbing the
incident light. In most cases, only a very small portion of the incident light will be Raman
scattered. In order to compare the efficiency of the Raman scattering process with other optical
processes or among the different molecules of interest, a common measure is often desired.
Commonly, the cross-section, σ, or the differential cross-section, dσ/dΩ, is used.13, 15
The cross-
section of an optical process is defined as the ratio of the rate of photon in the optical process in
all directions to the rate of photons striking a molecule. The unit of σ is in m2 or cm
2; therefore, it
19
can be interpreted as the area of molecule that produces a particular optical process. For
example, in Raman scattering, we can define the Raman cross-section (σRS) as follows:
234 = 5345
= *34
(2.9)
where IRS and PRS are the total Raman scattered light average all over random molecular
orientations in photons/s and in W respectively, and I0 and E0 are the flux (photons/s/m2) and
irradiance (W/m2) of the incoming photons. The Raman cross-section, σRS, can be described as
the area of a molecule that produces Raman scattering when the incident light has a photon flux
of Io. The differential Raman scattering cross-section will simply be the variations of the Raman
cross-section with respect to the solid angle. Theoretically, the Raman cross-section is also
proportional to the polarizability derivative for vibrational transition from state m to state n and
the scattering frequency as follows:
234 = 6789|′-| (2.10)
where < = , the polarizability derivative derived in section 2.2. ωs is the frequency of the
scattered light and C is a numerical constant.
Typical values for cross-sections of the different optical processes are shown in Table 2.5. A
comparison of the cross-section for Raman scattering with absorption, fluorescence and Rayleigh
scattering indicates that Raman scattering is the least efficient among the different optical
processes. This shows that the Raman technique is inherently insensitive. Moreover, the cross-
section for Raman scattering is 10 orders of magnitude weaker than fluorescence and 3 orders of
magnitude weaker than Rayleigh scattering. Significant interference from these processes
prevents the Raman technique from being useful in many applications. Nevertheless, significant
enhancement in the Raman cross-section can be achieved if the excitation laser resonates with
20
the electronic transition state of the molecule, or surface enhancement is provided by metallic
particles.
Table 2.5: Approximate order of magnitude for cross-sections (per molecule), σ, of the different
optical processes.15
Process Cross-section of σ (cm2)
Absorption Ultraviolet 10-18
Absorption Infrared 10-21
Emission Fluorescence 10-19
Scattering Rayleigh Scattering 10-26
Scattering Raman Scattering 10-29
Scattering Resonance Raman 10-24
Scattering Surface Enhanced Raman Scattering 10-16
21
Chapter 3
Photonic Crystal Fibers
In the last few decades, a new frontier of research has been conducted to explore ways of
engineering optical properties of materials to control light propagation. One solution to the
problem is found to be photonic crystal, which is a periodic structure of different dielectric
constants or refractive indices on the scale of the wavelength of light. This periodic structure
produces a so-called photonic bandgap, in analogy to the electronic bandgap, in which some
frequencies of light are completely reflected in a certain direction and cannot propagate through
the structure.16
This phenomenon is caused by the constructive and destructive interferences of
the incident and reflected waves from the periodic layers of the structure.17
At each layer of the
periodic structure, light is partially reflected and partially transmitted. If the spacing between
adjacent layers is periodic at the wavelength of light, the multiple reflections can interfere
destructively to eliminate the forward propagating wave and constructively to form the reflected
wave (Figure 3.1). Hence, this structure can act like a mirror which reflects light at the desired
wavelengths by controlling the refractive index of materials in the structure and the spacing
22
between each layer. In two and three dimensional systems, a similar photonic crystal can also be
designed with the appropriate lattice structure and refractive index contrast to achieve a photonic
bandgap.
Figure 3.1: Wave interferences in the photonic bandgap of a one dimensional periodic material.
(1) Wave incident on the material, (2) wave partially reflected off each layer of the structure, (3)
the reflected wave combined with the incident wave to produce a standing wave that does not
travel through the material.17
In 1996, Russell et al. utilized this concept and demonstrated the first fiber with a photonic
crystal cladding.18
This new type of fiber is named the photonic crystal fiber (PCF) (Figure 3.2).
Often, they are also referred to as the micro-structured fiber (MSF), micro-structured optical
fiber (MOF) or holey fiber. Unlike conventional optical fiber which guides light by the TIR
effect of light, PCF can guide light using the photonic bandgap formed in the cladding that
features a two-dimensional periodic triangular structure of large air holes. The core of the PCF in
which light can propagate in is usually a defect formed by removing several air holes in the
center of the periodic structure that runs along the entire fiber length. This defect can be a solid
material like silica, or a large air core that allows gas or liquid to be filled inside while light is
propagating through. In the past decade, many different configurations of the PCF have been
23
designed for various applications of light delivery such as high power, high non-linearity, large
mode area, high birefringent, and high NA. In this chapter, we will first classify the different
types of PCF that has been designed since the first demonstration of PCF. Then, we will discuss
the two possible guiding mechanisms in these PCFs and the different regimes that they can
operate in. Finally, we will focus on a particular type of PCF, named the Hollow-core PCF (HC-
PCF), and discuss how it can be utilized to enhance the Raman scattering signal from aqueous
and biological solutions, which forms the basis of this thesis.
Figure 3.2: Cross-section of a HC-PCF taken by scanning electron microscope (SEM). The
cladding is composed of a holey region which is a two-dimensional photonic crystal surrounded
by a normal silica coating. The core is a defect formed by removing seven holes in the center of
the photonic crystal.
(3.1) Classifications of Photonic Crystal Fibers
PCF is a general term used to describe fibers that utilize a periodic distribution of holes or
dielectric rods in the cladding to form the waveguiding properties of fiber. Although the term
“photonic crystal” refers to structures that form a photonic bandgap in which certain wavelength
of light cannot propagate through, the main guiding mechanism of a PCF does not necessarily
utilize this photonic bandgap effect. If the refractive index of the fiber core is designed to be
24
greater than that of the cladding, TIR is still mainly responsible for the guidance of light even
though the cladding is composed of a periodic structure. At certain propagation directions where
light is not guided by TIR, the photonic bandgap effect may still guide the light if the photonic
bandgap exist. For example, a solid-core PCF, in which the core of the fiber is made out of silica
and the cladding is formed by a periodic distribution of air holes, would guide light in the silica
core with TIR because the average refractive index of the air hole cladding is less than that of the
silica core. In contrast, a HC-PCF in which the same cladding structure is used as the solid-core
PCF but the central core is changed to air would guide light in the central air core due to the
existence of photonic bandgaps in the fiber cladding. However, if the central air core of a HC-
PCF is filled with some liquid or gas, light can also be guided by both TIR and photonic bandgap
effects. Figure 3.3 classifies some of the different designs of PCF based on their main guiding
mechanism.
In the next two sections, we will describe the physical mechanisms of both TIR and photonic
bandgap effect using the wave theory of light in PCF. The different regimes of light propagation
will be discussed in detail.
25
Figure 3.3: Classification of PCFs based on their main light guiding mechanism.19
(3.2) Index Guiding Mechanism
Similar to conventional optical fibers, PCF can guide light using TIR when the refractive index
of the core is higher than the average refractive index of the cladding. This phenomenon is
usually described using Snell’s law in ray optics (Figure 3.4):
= sin @ = = sin @ (2.1)
26
where = is the refractive index of the material where the incident ray is in and @ is the angle the
incident ray makes with the plane normal to the interface. = is the refractive index of the
material where the transmitted ray is in and @ is the angle that the transmitted ray makes with
the normal to the interface. When a ray of light is incident at an interface between two materials,
the angle of the transmitted ray follows the relation of Snell’s law. As @ increases, @ also
increases. At the critical angle (@A, @ = @A = sinC= =⁄ , @ reaches a 90o angle in which
the transmitted ray parallels the interface if = < =. At this angle or greater (i.e. @ ≥ @A), the
incident ray is interpreted to be totally reflected at the interface. This means that light can
propagate through a high index medium, such as water or silica, when it is surrounded by a lower
index medium, such as webs of air holes.
Figure 3.4: A ray of light incident at the interface between two dielectric materials and refracted
according to Snell’s law.
Although Snell’s law is able to describe how light can confine and propagate inside a fiber, its
description is limited in the ray optics regime in which it is only valid when light is propagating
through and around materials in the length scale much greater than the wavelength of light.
Hence, it is unable to explain phenomenon that are resulted from the wave properties of light,
27
such as interference. Since the photonic bandgap effect of a periodic material is a result of the
constructive and destructive interferences of the incident and reflected waves from each layer of
the structure, Snell’s law is not sufficient to explain its effects. In order to realize the different
regimes in which TIR and photonic bandgap guide light in a PCF structure, we will need to
generalize Snell’s law to explain both of the effects.20
One way to accomplish this is to utilize the
wave theory of light which is modeled by Maxwell’s equations.
In the wave model of light, Snell’s law is simply the conservation of both the light frequency, ω,
and the component of the wave vector that is parallel to the interface of the two materials, k||, as a
result of phase matching the incident, reflected and transmitted electric field of the wave due to
boundary condition (Appendix A). This implies that we can study band structures of EM modes
(i.e. the dispersion diagram or band diagram), relating ω and k|| (the two conserved quantities), to
determine regimes that TIR can operate in PCF. At the same time, energy gaps resulted from the
periodic structure of the photonic crystal cladding can also be determined. Let us consider a PCF
composed of a high-index material, nh, and a low-index material, nl, such as silica glass and air.
The dispersion relation of the two materials, 7 = A-G
H|| and 7 = A-I
H||, divides the dispersion plot
into three regions, independent of the geometry of the core and cladding, as shown in figure
3.5.21
The wave properties in each of the three regions can be determined by the transverse wave
vector (i.e. the component of the wave vector that is perpendicular to the interface), HJ, where
HJ = KLA = − H||.
Figure 3.5: Dispersion diagram of PCF showing regions of light guidance by TIR and photonic
bandgap effect. H|| is the component of the wave vector in the invariant direction.
of the periodic air holes in the photonic crystal.
For H|| M LA =N, the only solutions exist are those with
OPKLA = − H|| ). This means that the wave is evanescent in both high and low materials; thus,
no guiding mode exists and light cannot be guided inside the fiber.
For H|| < LA =Q, HJ can take any real values which means that waves can propagate freely in both
materials. Thus, light is not confined by either material unless there is a photonic bandgap in
which certain energy cannot propagate in the material.
For LA =N M H|| M LA =Q, HJ can take any real values in the high index material but is imaginary in
the low index material. This means that the wave can propagate inside the high index material
28
Dispersion diagram of PCF showing regions of light guidance by TIR and photonic
is the component of the wave vector in the invariant direction.
of the periodic air holes in the photonic crystal.
, the only solutions exist are those with HJ imaginary (i.e HJ =
). This means that the wave is evanescent in both high and low materials; thus,
no guiding mode exists and light cannot be guided inside the fiber.
an take any real values which means that waves can propagate freely in both
materials. Thus, light is not confined by either material unless there is a photonic bandgap in
which certain energy cannot propagate in the material.
can take any real values in the high index material but is imaginary in
the low index material. This means that the wave can propagate inside the high index material
Dispersion diagram of PCF showing regions of light guidance by TIR and photonic
is the component of the wave vector in the invariant direction. a is the period
=
). This means that the wave is evanescent in both high and low materials; thus,
an take any real values which means that waves can propagate freely in both
materials. Thus, light is not confined by either material unless there is a photonic bandgap in
can take any real values in the high index material but is imaginary in
the low index material. This means that the wave can propagate inside the high index material
29
while it is evanescent in the low index material. Therefore, if we have a PCF or any fiber where
the index of the core is higher than the average index of the cladding, light has to be confined in
the high index core where the light wave can propagate. The wave would not be able to
propagate through the cladding as it becomes evanescent in the low index material. Hence, light
is guided by TIR within this region. However, photonic bandgap can also exist in this region as
well; thus, photonic bandgap confinement is also possible in this regime.
Figure 3.6 shows a more concrete example of a dispersion diagram for the case of a liquid filled
HC-PCF. In this example, the HC-PCF has a cladding of periodic air holes in silica and a central
core that is filled with water. The dispersion relations of the four different refractive indices in
the fiber (i.e. air, silica, photonic crystal, and water) divides the plot into five different regimes in
which light can propagate in different set of materials or structure. In regime 1 and 2, light can
propagate in both the photonic crystal and water; thus, light can only be guided in the water core
by photonic bandgaps. On the contrary, in regime 3, light can only be guided in water but not in
the photonic crystal cladding; thus, light can be guided by total internal reflections due to the
refractive index difference between water and the photonic crystal cladding. In regime 4 and 5,
light cannot propagate in both the water core and the photonic crystal cladding. Therefore, no
guiding mode is possible in these two regimes. If a narrowband light wave with frequency
centered at the white line in Figure 3.6 is launch into the water-core PCF, this light would
propagate through the water core in three different ways: 1) through translation symmetry along
the fiber axis, 2) through photonic bandgap guidance at the red dot and 3) through TIR along the
red line. This demonstrates the different modes of operation in a liquid core PCF.
30
Figure 3.6: Dispersion diagram of a water-core PCF. In regime 1 and 2, light is free to
propagate in both the water core and the photonic crystal cladding except at the photonic
bandgaps indicated by black lines. In regime 3, light is free to propagate in the water core but
evanescent in the photonic crystal cladding; thus, light can be guided in the water core by TIR. In
regime 4 and 5, light is evanescent in both the water core and the photonic crystal cladding;
therefore, no guidance mode is possible.
(3.3) Photonic Bandgap Guiding Mechanism
In addition to the conventional way of light guidance by TIR, a new type of guiding mechanism
is introduced by the PCF, namely the photonic bandgap guidance. This mechanism is based on
the use of a photonic crystal to prohibit the existence of a certain energy in the cladding of the
PCF; thus, preventing a certain wavelength of light from escaping the core of the fiber. The
region where light cannot exist in the dispersion plot of a photonic crystal is called the photonic
bandgap. If we operate within the photonic bandgap regime of the photonic crystal cladding,
light can propagate through the central core of the PCF even if the core index is less than that of
31
the cladding (i.e. when light guidance with TIR is not possible). This implies that light can now
be guided in an air, gas, or liquid core of a PCF with proper design of the PCF structure.
Theoretically, the photonic bandgap of a photonic crystal can be studied through the dispersion
diagram discussed in section 3.2. The dispersion relation can be determined by first deriving a
translational relation between electric field of two adjacent unit cells through imposing
continuity at the interfaces. According to the Bloch theorem, the electric field vector of a normal
mode of propagation in a periodic medium (i.e. medium periodic in the z-direction by
convention) takes the form
R = RSTUC(VWU(X,CVYCVZ[ (2.2)
where K is the Bloch wavenumber and EK(z) is a periodic function with period Λ such that EK(z)
= EK(z + Λ). This periodic condition yields an eigenvalue problem with the unit-cell translation
of the electric field. The dispersion relation of the photonic crystal can then be obtained through
the eigenvalue of the unit-cell relation. To derive the dispersion relation of a two dimensional
photonic crystal, such as the PCF, a numerical technique would generally be required because a
closed-form relation does not generally exist in most cases. Nowadays, many numerical methods
have been developed to determine dispersion relation of photonic crystals with complex
geometries such as plane wave expansion method, finite-difference time-domain (FDTD)
method, finite element method, order-n spectral method, and Korringa-Kohn-Rostoker (KKR)
method. The details of these numerical techniques are out of the scope of this thesis and
therefore will not be discussed here.
For a photonic crystal with a triangular lattice of air holes in silica, which is typically used in the
cladding of a PCF, the in-plane and out-of-plane dispersion diagram are shown in Figure 3.7.
32
The white-coloured “fingers” in the out-of-plane dispersion diagram are the bandgaps of the
photonic crystal. The large white triangular area under the red air dispersion line is the region
where light is evanescent (i.e. propagation is turned off in the photonic crystal). This domain
corresponds to the guidance of light by TIR. Note that in the out-of-plane propagation, a
complete bandgap does not exist because of the low index contrast between air and silica.22
This
means that light propagating at some angles to the photonic crystal plane is not supported.
Nevertheless, a complete bandgap is not necessary in the PCF because fibers have a continuous
translational symmetry in the z direction. Any propagation along the fiber axis (i.e. kz ≈ 0) will
be conserved as a result of the symmetry. In fact, as a result of the symmetry, the kz component
of the wave vector, is significantly higher than the transverse components (i.e. kx and ky). This
means that only a small portion of the modes need to be guided by the structure. Nevertheless,
light propagating in the plane of the photonic crystal cladding (i.e. in-plane propagation at kz = 0)
or at some angle to the photonic crystal plane will also be guided by the bandgap of the photonic
crystal as demonstrated by the two shaded rings in Figure 3.8. The ring at the waist of the sphere
indicates the in-plane photonic bandgap at kz = 0. The ring at the upper pole of the sphere
denotes the out-of-plane bandgap shown as the white “fingers” in Figure 3.7a which only guide
light for certain values of kz for a particular frequency. In sum, the photonic bandgap guidance of
light in PCF is achieved by a complex solid angle of photonic bandgaps in the cladding of the
fiber. The translational symmetry of the fiber provides light propagation in the central core
without an omni-directional bandgap, which eliminates the need for a high index material to
open up an omni-directional photonic bandgap.
Figure 3.7: Dispersion diagram of a photonic crystal as a function of (a) out
vector, kz, and (b) in-plane wave vector at k
for the triangular lattice of air holes with period
Figure 3.8: Schematic illustration of solid angles falling within the photonic band
photonic crystal with a triangular lattice operating at a fixed frequency. k
wave vectors in the plane of the photonic crystal in which the structure varies periodically. k
denotes the wave vector in the direction with translational symmetry.
33
Dispersion diagram of a photonic crystal as a function of (a) out
plane wave vector at kza/2π = 1.7 in the irreducible Brillouin zone (inset)
for the triangular lattice of air holes with period a and radius 0.47a in ε = 2.1.
Schematic illustration of solid angles falling within the photonic band
photonic crystal with a triangular lattice operating at a fixed frequency. kx and k
wave vectors in the plane of the photonic crystal in which the structure varies periodically. k
denotes the wave vector in the direction with translational symmetry.22
Dispersion diagram of a photonic crystal as a function of (a) out-of-plane wave
a/2π = 1.7 in the irreducible Brillouin zone (inset)
in ε = 2.1.23
Schematic illustration of solid angles falling within the photonic bandgap of a
and ky denote the
wave vectors in the plane of the photonic crystal in which the structure varies periodically. kz
34
(3.4) Hollow-Core Photonic Crystal Fiber for Efficient Raman Scattering
In nature, Raman scattering is an extremely rare event (roughly only 1 photon in 10 million
incident photons is Raman scattered). In fact, Raman scattering is much weaker than most other
optical processes without surface enhancement or being in resonance with the pump laser as
discussed in section 2.5, which makes it difficult to detect in most analytes. Although many
commercialized Raman systems built with low loss optical components and highly sensitive
detectors are readily available in the market now, their conventional way of inducing and
collecting Raman scatterings are still very inefficient which highly limits their system sensitivity,
particularly for aqueous and diluted solutions.
PCF can be a means of enhancing the detected Raman signals from aqueous or diluted solutions,
without the need of additional metallic nanoparticles for surface enhancements.24, 25
In particular,
HC-PCF, whose PCF featuring a central air core and a high air-filling fraction cladding formed
by a periodic air hole array in silica, enables the pump laser to be confined inside the central core
of the PCF through both photonic bandgap effects and total internal reflections. This is a
distinguishing feature of our technique as we can utilize both guiding mechanisms by selectively
filling the central air core with analyte to maintain the photonic bandgap effects of the cladding.
By confining both the pump laser and the analyte along the length of the PCF, a larger volume
for light-matter interaction is achieved compared to a conventional Raman spectroscopy scheme.
Figure 3.9 compares the conventional Raman spectroscopy scheme with the one using HC-PCF
as an interaction medium for Raman spectroscopy. In conventional Raman spectroscopy (Figure
3.9a), the pump laser is focused directly into the analyte to generate Raman scattering. Most of
the Raman signal detected is scattered from the beam waist of the pump laser in which the pump
laser is most intense. In this case, the pump laser and the analytes are only interacting in the
35
volume limited by the spot size of the pump laser in the lateral direction and the depth of field in
the axial direction. Since Raman signals are scattered in omni directions, only a fraction of the
scattered signal is collected by the objective and detected by the detector due to the limited NA
of the objective. In the case of an aqueous QD solution, the amount of scattered signals detected
from the small quantity of QDs dispersed in water is very minimal.
Figure 3.9: Raman scattering and signal collection in (a) conventional Raman spectroscopy and
(b) Raman spectroscopy using HC-PCF.
By employing HC-PCF as the medium for light-matter interactions (Figure 3.9b), both the pump
laser and the analyte can be confined within the central core of the HC-PCF. Subsequently, the
confined laser power within the core induces more interaction with the analyte that is filled
inside. Thus, Raman signals are scattered throughout the entire length of the fiber as opposed to
just the depth of field of the objective in the conventional scheme. Since Raman signals are
mostly shifted by less than tenths of nanometer from the wavelength of the pump laser, they will
also be confined inside the central core of the PCF and collected throughout the fiber. As a
36
result, the output signal from the fiber end will be collected more efficiently by the objective for
detection. With the use of HC-PCF, the detected Raman signal can be about 2 to 3 orders of
magnitude greater than that of the conventional Raman technique in practice.
Although it is not trivial to completely decouple the enhancement due to increased collection
efficiency throughout the fiber from that due to increased interaction length, the order of
magnitude increased in the detected Raman signal due to the two different components was
studied. This was carried by measuring the enhancement of the detected Raman signal with
increasing PCF and Teflon capillary tube (TCT) length (Figure 3.10). The same pump power,
coupling objective, and setup were used for all experimental points in the figure. Hence, detected
signal enhancement with increasing fiber length at a short distance is mostly contributed by the
increased light-matter interactions. For an increased PCF length from 4 cm to 11 cm, the Raman
signal is enhanced by about 3 to 4 times. The experimental points are also fitted to a model for
calculating Raman intensity in a liquid core waveguide in a backscattering configuration as
follows:26
5\]^ = *2 _2`a1 − UC\ (2.3)
where Po is the fraction of laser power that is coupled into the waveguide, α is the loss
coefficient of the waveguide, 4A is the numerical aperture of the waveguide, and L is the length
of the waveguide with solution filled inside. ρ and σ are the Raman cross-section and the number
of scattering centers per unit volume of the sample solution respectively. The experimental data
collected were fitted to the model using α and κ ≡ Poρσπ4A.27
According to our model, if we
further increase the length of the PCF and TCT beyond the experimental data points, the
37
enhancement will reach a plateau where it will not increase further because of the absorption and
attenuation loss of the fibers.
If we compare the normalized Raman intensity between PCF and TCT at a fixed fiber length,
PCF achieved a greater enhancement because of the greater collection efficiency in PCF. At a
fiber length of ~10 cm, the normalized Raman intensity from PCF is about 20 times greater than
that from TCT. This signal increase is contributed to the improved collection efficiency of the
PCF relative to the TCT. The greater collection efficiency in PCF can be understood
mathematically from the solid angle that collects Raman scattering inside a liquid core fiber at a
fiber section according to:26
bA-c = =Adc − =cee (2.4)
where bcone is the solid angle that collects Raman scattering. ncore and neff are the refractive index
of the fiber core and cladding respectively. Consider a water core fiber where ncore = 1.33. Ωcone
would be 0.33 in the case of TCT since neff of Teflon is 1.29. In the case PCF whose cladding is
made up of air holes with over 90% air filling fraction, neff would be close to 1; thus, Ωcone would
be greater than 0.33 (i.e. larger collection angle than that of the TCT). In fact, neff of the HC-
PCFs used for this thesis (i.e. HC-800 and HC-1060 from NKT Photonics) are 1.14 and 1.17
respectively according to Eftekhari et al.27
They correspond to Ωcone of value 1.47 and 1.25,
which are about 4 times larger than that of the TCT. This shows that a greater collection angle is
obtained with PCF which resulted in an increased Raman signal in comparison to the TCT.
38
Figure 3.10: Raman intensity of water mode centered at around 3400 cm-1
for varying lengths
of (a) HC-800 PCF and (b) TCT. The intensity is normalized with respect to the intensity
obtained with a cuvette. The coupling objective, pump power and setup were kept the same for
all experimental points. Insets: Predicted plateau of intensity with length up to 500 cm.
In addition to the detected Raman signal enhancement, it is important to note that the pump laser
is confined inside the “liquid” core of the PCF; thus, interaction between the pump laser and the
glass wall of the fiber is minimal and the resulted interference is very limited in the presence of
other vibrational modes and backgrounds. Figure 3.11 compares the Raman signal of silica with
that of water and a thiol-capping agent that is filled into the central core of a HC-PCF. The silica
Raman modes are very weak in the water spectrum when there is no Raman mode or background
signal present between 300 and 500 cm-1
. In the spectrum of the thiol-capping agent, no silica
Raman modes are observed when there are vibrational modes present between 300 and 500 cm-1
.
This shows that the Raman signal interference from the silica wall of the HC-PCF is insignificant
when the central core is filled with analyte.
39
Figure 3.11: Comparison of silica Raman signals between water and thiol-capping agent when
they are filled into the central core of a HC-PCF. Weak silica modes are observed in the water
spectrum when no Raman mode is present between 300 and 500 cm-1
. When Raman modes are
present as shown in the spectrum of the thiol-capping agent, no silica mode is observed.
Furthermore, unlike SERS, whose detected Raman signal enhancements require metallic
nanoparticles to be absorbed or in close proximity with the analyte, the detected signal
enhancement by efficiently inducing and collecting Raman scatterings with HC-PCF completely
eliminates the possibility of altering analyte structure or properties due to conformational
changes when interacting with the metallic nanoparticles. However, if conformational changes of
analyte are not a concern, surface enhancements can also be obtained in addition to the efficient
Raman scattering of HC-PCF by incorporating metallic nanoparticles into the analyte or central
core of the HC-PCF.
In the essence of combining HC-PCF with Raman spectroscopy, the practical advantages of
optical fibers are also added to the conventional scheme. This includes the high flexibility and
compactness of optical fibers for remote sensing and in-situ detections, sampling versatility, and
potential for low-cost analysis as the cost of HC-PCF drops dramatically through economies of
40
scale. In addition, HC-PCF requires only a very small sample volume for detection, and it is
capable of detecting multiple compounds in parallel. This is crucial in many biological,
environmental and forensic applications where sample analysis with only nanoliter volume is
required.
It is important to mention that this novel PCF platform for Raman spectroscopy is capable of
enhancing the detected Raman signal for solutions with many particles (hundreds or more) as
well as solutions with a single particle theoretically. In solutions with many particles, the
confinement of both the pump laser and the sample solution in the PCF core enable more
particles to generate detectable Raman scattering signals. Moreover, these confinements enable
more photons to interact with each particle to generate more signals due to the high photon
density in the PCF core as a result of the strong light confinements. Although there are only one
Raman scatterer in single particle solutions, more Raman scattering signal can still be generated
compared to conventional Raman spectroscopy due to the increased photon density in the PCF
core which allows a greater number of photons to interact with the one particle to generate more
Raman scattering signals. However, there are also other technique that are developed to provide
greater Raman scattering signal for single particle applications such as SERS and Raman
tweezers.28, 29
The development of the PCF platform is currently focused on the multi-particle
systems in which the ensemble averaged interactions and structures of the nanoparticles can be
monitored before, during and after their synthesis.
In the next three chapters, we will discuss how the HC-PCF can be used in practice to detect
molecular bindings in nanostructures including semiconductor QDs and metallic nanoparticles.
Experimental details and results from each of these studies will be presented and discussed.
41
Chapter 4
Raman Characterization of
Colloidal CdTe "anoparticles
(4.1) Introduction
Colloidal QDs are suspensions of semiconductor nanoparticles that offer both the intriguing
optical properties of quantum-confined particles, and the practical advantages of solution-based
processing.30-32
In the past two decades, aqueous synthesis of colloidal QDs has gained much
popularity and evolved tremedously.33-37
This simple and cost-effective technique allows very
small (2-6 nm), monodisperse, and highly water-soluble QDs to be synthesized in gram
quantities. QD capping plays a pivotal role in the properties and utility of the material; the use of
short-chain thiols, such as thioglycolic acid (TGA), as capping agents have shown to greatly
improve the PL quantum efficiency of as-synthesized QDs to values of 40-60%.33
3-
mercaptopropionic acid (MPA) has also shown to offer a larger range of size and PL tunability,
42
and a longer emission decay time.33
Other thiols, such as 1-thioglycerol (TG), are found to be
more suitable for synthesizing stable core-shell QDs.36
The unique properties of the different
colloidal QDs have potentials for a wide field of novel applications ranging from photovoltaics38
and optoelectronics,39
to biosensing,40
bioimaging41
and even cancer treatments.42
One of the most important technological challenges in QD advancement is the development of a
cost-effective, reliable and sensitive optical monitoring system to control the physical, chemical
and size-dependent properties of QDs before, during and after their fabrication on a nanometer
scale. To measure and compare these properties between the different QDs, many analytical
techniques have been employed including but not limited to PL, EL, UV-vis, TEM or HRTEM,
CZE, XRD, and XPS. These techniques provide valuable information on the composition and
properties of the QDs; yet, none of them describes how the capping agents binds with the core of
the QDs. Their impact on the overall molecular structure, molecular complex, and different QD
properties also remain unclear. Consequently, this limits our capability to improve the quantum
efficiency, stability, and bioconjugating ability further from what has been achieved today.
Complex QD designs for increasing performance and functionalities in different applications
remain very challenging.
FT-IR can be an alternative for determining the binding interactions between the QDs and their
capping agents. However, the strong and broad absorption bands of water often overlap with
those from the QDs and stabilizing agents. This limits the number of vibrational modes that can
be resolved.
A complementary technique to FT-IR is Raman spectroscopy, which is a rapid and non-
destructive means of probing molecular vibrations optically through inelastic scattering. Raman
43
provides the unique vibrational modes or “fingerprints” of the molecules that enable similar
molecular structures to be unambiguously distinguished and identified. More importantly, the
vibrational modes of water are inherently weaker in Raman scattering than in FT-IR. Distinctive
Raman modes of colloidal QDs in their native and dilute aqueous environment can be readily
obtained without significant interference. However, Raman scattering has not been successful in
characterizing QDs because it is generally very weak compared to elastic scattering, and it is
further weakened; therefore, it has only been used in characterizing QDs coated on films or in
powder form.37, 43-46
SERS has also been used to obtain an enhanced Raman signal from
QDsolutions.24, 25, 47
However, metallic nanoparticles have to be mixed with the QD solution,
which could alter the structure and properties of the QDs from its as-synthesized state due to
conformational changes when interacting with metallic nanoparticles.
In this chapter, we demonstrated the use of the novel HC-PCF as a means of enhancing the
detected Raman signals from aqueous or diluted solutions, without the need of additional
metallic nanoparticles.9, 10
In previous works of our research group, HC-PCF demonstrated
potentials for low-cost in-situ monitoring of QD structure during the synthesis process. The
usefulness of HC-PCF for Raman spectroscopy has recently been demonstrated in the studies of
ZnO nanoparticles synthesized through base hydrolysis method,10
and colloidal CdTe
nanoparticles post-synthesized using rapid thermal annealing.9 In these studies, Raman modes of
the ZnO and CdTe nanoparticles were obtained owing to the detected signal enhancements
provided by the HC-PCF.
In this chapter, we report the enhanced detected Raman signals from aqueous solutions of thiol-
stabilized CdTe QDs through the use of HC-PCF. Strong and well-resolved vibrational modes
44
ascribed to the CdTe semiconductor core, capping agents, and their interface were obtained for
the first time in the aqueous medium. The observed Raman vibrational modes are divided into
three regimes. The modes in the low Raman shift regime reveal the low crystallinity of CdTe
QDs capped with TGA due to a large inclusion of Te compounds and surface defects. The modes
in the mid and high Raman shift regime reveal the formation of thiolates, and Cd-S bonds
between the thiolate and the CdTe core, which stabilize the QD through structured CdSxTe1-x
ternary compound. MPA is also found to promote the formation of unidentate and chelating
bidentate complexes with the CdTe core. These complexes further passivate the QD and
potentially improve its stability at the expense of its solubility and bioconjugating ability. This
experiment substantiates the promise of HC-PCF in enhancing the Raman scattering signal of
nanoparticles in aqueous solutions and enables possible studies of molecular structures relating
to the different properties of QDs.
(4.2) Experimental Details
The thiol-capped CdTe QDs used in this experiment were synthesized with minor modification
to literature procedure.48, 49
The resulting solid CdTe QD samples were then weighted and
dispersed in deionised water to prepare aqueous solutions of 2 mg/mL for Raman measurements,
0.04 mg/mL for UV-vis and 0.4 mg/mL for PL. For detailed synthesis methods and materials of
the CdTe QDs, please refer to Appendix B.
The structure of the highly fluorescent CdTe QDs and the three different ligands used for
stabilization are shown in Figure 4.1. Room temperature PL measurements were performed using
a Perkin-Elmer Luminescence Spectrophotometer in aqueous solution. UV-vis measurements
were performed using an Infinite M1000 TECAN system also in aqueous solution.
45
Figure 4.1: (a) Core-shell structure of CdTe QDs. (b) Molecular structure of capping ligands.
Raman spectra were acquired using a Horiba Jobin Yvon HR800 micro-Raman system equipped
with a CW 632.8 nm HeNe laser in the range of 2 – 5 mW. The spectral resolution of the
spectrometer was about 1 cm-1
. To incorporate HC-PCF onto the stage of the Raman system for
measurements, the fiber setup shown in Figure 4.2 was used. In this setup, HC-PCF is placed
inside a fiber chuck which is held with a chuck holder on an aluminum block that is placed on
the stage of the Raman system. To measure a Raman spectrum, the laser light was focused into
the central core of the HC-PCFs using a 100x objective to maximize the Raman signal from the
analyte while minimizing those from the fiber background. Figure 4.3 compares the Raman
spectra of water that is filled inside the central core of a HC-PCF when the objective is focused
on different spots of the PCF cross-section. The strongest water signal and the weakest
background are obtained when the laser light is focused into the central core of the PCF where
water is filled.
46
Figure 4.2: Experimental setup of HC-PCF in HR800 micro-Raman system.
Figure 4.3: Raman spectrum obtained when the laser light is focused onto the cladding, holey
region and the center of the central core of the HC-PCF when it is filled with water.
0 500 1000 1500 2000 2500 3000 3500 4000
0
1000
2000
3000
4000
5000
6000
Raman Intensity (a.u.)
Raman Shift (cm-1)
Cladding
Holey Region
Core Center
47
Each spectrum of the QDs was averaged over 40 measurements with an exposure time of 30 s.
This long averaging and exposure time were to maximize the signal-to-noise and signal-to-
background of the QDs. An increased laser exposure on the QDs sample allowed some water to
be evaporated. This reduced the water signal to a level lower than that of the fluorescent signal
which eliminated Raman mode interferences. Since CdTe nanoparticles do not evaporate, the
CdTe QD signals also increased relative to that of the water.
The HC-PCFs used in this experiment were obtained from NKT Photonics, HC-800-01. Each
piece was segmented, stripped and cleaved into ~6 cm long. The core of the HC-PCF was
selectively filled, using a technique developed by Irizar et al.,10
to allow enhancement of the
detected Raman signal through bandgap confinement of the pump laser. This technique involved
splicing one end of the HC-PCF with a conventional fiber splicer to collapse all the cladding
holes while leaving the central core open (Figure 4.4a). Then, the spliced end of the PCF was
submerged into the analyte to fill the PCF’s central air core by capillary effect. Figure 4.4b
shows an optical image of the opposite end of PCF after its entire core is filled with analyte.
The intensity of the OH modes of water between different techniques is compared in Figure 4.5.
By selectively filling the entire central air core with water, the intensity of the OH stretching
mode at 3427 cm-1
is about 101 times greater than that of the conventional Raman spectroscopy
with a cover slide. This enhancement is attributed to the combination of the TIR and photonic
bandgap guidance inside the liquid core. Compared to non-selective filling (i.e. filling both the
central air core and the cladding holes with water) in which there is only the photonic bandgap
effect with the bandgap shifted to the lower wavelength due to the filling of the air cladding
holes, the intensity of the OH stretching mode at 3427 cm-1
is 73 times greater in the selective
48
filling case. This shows that our selective filling technique of the PCF gives the greatest
enhancement of the detected Raman signal compared to conventional techniques and other
analyte filling methods of the PCF.
Figure 4.4: Optical images of HC-PCF cross-section at the (a) fused end and (b) non-fused end
after the entire central air core is filled with water.
Figure 4.5: Raman intensity comparison of the OH modes of water between conventional
Raman spectroscopy using a cover slide, selective filling of PCF and non-selective filling of
PCF. Inset: OH stretching mode of water obtained from non-selectively filled PCF and
conventional Raman spectroscopy.
49
Raman spectra obtained were baseline removed and fitted with Gaussian or Gaussian-Lorentzian
mixed functions to determine their peak positions, amplitudes, and full-width at half-maximums
(FWHMs).
(4.3) CdTe Core and Te Defects
Raman spectra of aqueous CdTe QDs from 100 to 1750 cm-1
were obtained. The spectra are
divided into three regimes in which the peaks are correlated to the vibrational modes in the CdTe
semiconductor core, capping ligands, and their interfacial structures. Some of the modes are
broad because they are overlapped with multiple weak Raman modes that are located in
proximity. These modes might be originated from different functional groups, different
conformations of the same functional groups, or both. The limited sensitivity of the system
prevented them from being resolved.
In the low Raman shift regime, between 100 and 200 cm-1
as depicted in Figure 4.6, Raman
modes corresponding to the crystalline core and Te defects of the CdTe QDs are observed. The
peaks at ~120, ~140 and ~160 cm-1
are attributed to the Te A1, Te E and CdTe TO, and CdTe LO
modes of the QDs, respectively.50-54
These modes are observed in all three spectra indicating that
both the crystalline CdTe core and Te defects are presented in the QDs regardless of the type and
structure of the capping agent used. In particular, the CdTe LO modes are shifted to the lower
wavenumber by ~10 cm-1
compared to the LO mode of bulk CdTe crystal reported at ~170
cm-1
.51, 55
These shifts indicate the presence of phonon localization due to quantum confinement
effects, which suggests that the expected zero-dimensional CdTe structures are indeed in place.
In addition, the similar shifts of the CdTe LO mode among the three spectra show that similar
crystalline sizes are formed with the three different capping ligands used.
50
Figure 4.6: Raman spectra of CdTe QDs with different thiol capping agents in the lower Raman
shift regime between 100 and 200 cm-1
.
Moving toward the lower Raman shifts at ~120 and ~140 cm-1
, two Te modes are observed in the
three spectra. The two Te modes indicate the presence of Te crystals, either within the center of
the QD or the surface defects at the boundary of the QDs.53, 54, 56
QDs with surface defects are
known to have lower crystallinity and introduce trap states which reduces the PL quantum
efficiency.57
However, the amount of Te inclusions or surface defects cannot be quantified from
the spectra as the Raman cross-sections of Te modes are 75 times larger than that of the
crystalline CdTe modes.54
Because of the close proximity and broadness of the Te E and CdTe
TO vibrational mode at 140 cm-1
, precise contribution of the Te E modes from QDs with
different caps cannot be determined from their spectra. Nevertheless, we can calculate the
intensity ratios of the Te A1 mode to CdTe LO mode to compare the crystallinity of these QDs.
As shown in Table 4.1, the ratio is highest for TGA-capped QDs, followed by those for TG and
then MPA. In our previous study, we have shown that an increase in the ratio indicates an
51
increased in Te inclusion or surface defects relative to the amount of CdTe crystals formed in the
QDs.9 This suggests that crystallinity improved from TGA-capped QDs to TG-capped QDs and
is greatest with MPA-capped QDs. The improved crystallinity contributes to a greater PL
efficiency with fewer trap states that leads to non-radiative processes; thus, a brighter QD is
synthesized with MPA. This result correlates to the relative PL quantum efficiency estimated
from the UV-vis and PL spectra shown in Figure 4.7. The PL quantum efficiency can be
calculated as follows:
gh1 = gdceadceah1
=h1=dce
ih1idce
(2.1)
where g is the PL quantum efficiency, A is the absorbance at the excitation wavelength, D is the
integral area under the PL curve. The subscripts QD and ref correspond to the QD and the
reference sample (typically Rhodamine 6G with a refractive index of 1.359 and a quantum
efficiency of 95% excited at 400 nm), respectively. Since the QD concentrations are the same for
the three different capping agents, their refractive indices are also the same. However, MPA-
capped QDs give weaker absorbance compared to those capped with TGA and TA in the UV-vis
spectra at 400 nm. MPA-capped QDs also have the largest integral area in the PL spectra out of
the three QD types. The weaker absorbance and larger integral area of MPA-capped QDs
correspond to a higher PL quantum efficiency which agrees with the Raman result obtained.
Table 4.1: Intensity ratios of the Te A1 mode to CdTe LO mode for CdTe QDs capped with
different thiol agents.
Capping Agents TGA TG MPA
Te A1 / CdTe LO 2.949 0.901 0.499
52
Figure 4.7: (a) UV-vis spectra of the CdTe QDs (inset: fluorescent images of the QD solutions).
(b) PL profile of the QDs.
Overall, both the crystalline CdTe core and defects are observed in QDs capped with TGA, MPA
and TG. However, different crystalline cores are obtained with MPA-capped QDs being the most
crystalline among the three solutions.
(4.4) Core-Thiol Interface
In the mid-Raman shift regime, between 200 and 310 cm-1
, Raman modes are attributed to the
binding interactions between the QD core and the thiol capping agents as shown in Figure 4.8.
The two peaks at ~291 and ~260 cm-1
shown in this regime are ascribed to the surface optical
phonon (SO) mode of the CdS compound43, 58
and CdS-like LO phonon mode of the ternary
CdSxTe1-x compound, respectively.59, 60
The presence of the two CdS related modes denotes the
formation of CdS compounds on the surface of the QDs. This suggests that Cd ions on the
surface of the QD core are bonded with the S ions in the thiol terminus of the capping agents for
stabilization. This hypothesis is supported by the absence of the S-H stretching modes, in the
range of 2560 - 2590 cm-1
, in all three spectra (not shown), which shows that S-H bonds in the
thiol groups are broken and have then formed thiolates. These thiolates have perhaps binded with
53
Cd ions and formed CdS compounds around the CdTe core. Several other studies have also
proposed the same binding interaction between the thiol group and the core of the QDs.37, 61, 62
Since no visible CdS LO mode at ~300 cm-1
is observed, a CdS shell has not formed around the
CdTe cores in the three solutions. Instead, a CdSxTe1-x ternary structure might have formed
around the CdTe core, which is commonly formed upon heating processes.63, 64
It is known that
the position of CdS-like LO mode shifts from 305.6 (at x = 1) to 258.7 cm-1
(at x = 0) as S
content decreases in CdSxTe1-x compounds.59
Using the experimentally observed wavenumbers
and their corresponding CdSxTe1-x alloy compositions from Fischer et al.,
59 a composition around
CdS0.7Te0.3 is estimated for all three QD samples capped with the different thiol agents. The rich
S content in the alloys indicate that many thiolates are coupled with the CdTe cores and formed a
system close to the CdS shell. This strongly indicates that thiolates are strongly passivating the
surface of the QDs and thereby preventing the CdTe core from exposing and interacting with the
environment. Strong surface passivation also contributes to a greater PL efficiency through
elimination of surface defect states.
Figure 4.8: Raman spectra of CdTe QDs with different thiol capping agents in the mid-Raman
shift regime between 200 and 310 cm-1
.
54
It is important to note that both the CdS SO and the CdS-like LO modes are observed in all three
spectra with the different capping agents used. The peak position of the two modes are similar
among the three spectra, suggesting that the QD core binds similarly with the three thiol agents,
thus achieving the same alloy composition. Since the three thiol agents have different chain
lengths, this suggests that the degree of surface passivation might be unrelated to the length of
the thiol chain bonded to the QD surface. This is in agreement with the finding from Algar et al.,
in which a reduction of non-radiative decay rate with increasing alkyl thiol chain length was not
observed; thus, there was no improvement on the surface passivation of the QDs with lengthy
thiol agents.65, 66
In fact, it was suggested that the longer alkyl thiol chain of MPA and TG
excludes water from the QD surface, resulting in a higher quantum yield than the ones capped
with TGA (Figure 4.7). Previously, it was reported that photo-oxidation of the surface thiol
agents could catalyze the formation of disulfides through breaking the Cd-S bonds and reacting
with a neighbouring thiol group.67, 68
However, no S-S vibrational modes in the region between
480 and 510 cm-1
were observed in any of the three spectra. This indicates the absence of
disulfide bonds and confirms that photo-oxidization has not occurred in our solutions. The strong
surface passivation and the absence of photo-oxidization observed from the Raman spectra
further demonstrate that the CdTe QDs capped with TGA, MPA and TG are all highly stable in
their native aqueous environment.
(4.5) Carboxylate-Metal Complexes
In the high Raman shift regime, between 310 and 1750 cm-1
, vibrational modes of the thiol
capping agents are observed as shown in Figure 4.9. Notably, the peaks between 1330 and 1500
cm-1
and between 1570 and 1580 cm-1
are attributed to the symmetric and asymmetric stretching
modes of the carboxylates, respectively.37, 62, 69
These modes are only observed in TGA, and
55
MPA-capped QDs, in which one end of the thiol chain is terminated with a carboxylic acid
group. These modes indicate that the carboxylic acid groups are absorbed onto the QD surfaces
as carboxylates, most likely by forming bonds with the Cd ions on the CdTe core. Though one-
phonon Raman modes of CdO are not allowed due to the selection rules, no well defined
vibrational modes can be extracted from the Raman spectra.70
Nevertheless, it is known that the
carboxylic acid group can be absorbed on metals as carboxylates after deprotonation.37, 62, 71, 72
Multiple shifted modes of carboxylate symmetric stretches might also have risen if multiple
complexes of the carboxylate would have been formed. As such, the two symmetric stretching
modes observed are a result of two carboxylate-metal formations.
Figure 4.9: Raman spectra of CdTe QDs with different thiol capping agents in the high Raman
shift regime between 310 and 1750 cm-1
.
On the basis of previous work, carboxylate ions may coordinate to a metal in one of the three
modes: unidentate, bridging bidentate, and chelating bidentate.69, 71
These complexes can be
identified through the wavenumber separations, ∆, between the symmetric and asymmetric
56
stretches of the carboxylate. Unidentate complexes exhibit ∆ values (200 – 470 cm-1
) that are
much greater than the ionic complexes. Conversely, chelating bidentate complexes exhibit ∆
values (<110 cm-1
) that are much less than the ionic complexes. Bridging bidentate complexes
exhibit ∆ values (140 – 190 cm-1
) that lie in between the chelating bidentate and ionic complexes.
As shown in Table 4.2, both TGA and MPA-capped QDs have two carboxylate-metal
complexes. Both TGA and MPA show the formation of chelating bidentate complexes. In
addition, TGA formed the bridging bidentate complexes while MPA formed the unidentate
complexes. It was recently reported that the carboxylate-metal complex changes from unidentate
to bridging bidentate when the average size of the QDs increases from 8 to 20 nm.37
However,
the average particle sizes of our TGA and MPA-capped QDs calculated are very similar (i.e. 3.20
nm for TGA-capped QDs and 3.07 nm for MPA-capped QDs using their first absorption
maximum at 546 and 536 nm, respectively).73
The small size difference between the two QD
solutions is unlikely to cause a difference in the carboxylate-metal complex. This suggests that
the length of the alkyl thiol chain might have influenced not only the type of complexes formed
but also the amount of complexes formed. To compare the quantity of carboxylate-metal
complex formed between the different bindings interactions and QD solutions, the intensity
ratios of the COO symmetric stretches to the CdS 2SO mode are calculated. As shown in Table
4.3, the two ratios are much greater than 1 for MPA-capped QDs and less than 1 for TGA-
capped QDs. The large ratio implies that a large quantity of carboxylate ions is coordinated with
the Cd ions. This is not only suggesting that the formation of carboxylate-metal complexes is
much more favourable with MPA than TGA, but also that the surface of the QD is more
passivated by carboxylates in the MPA chain.74, 75
As a result, fewer defect sites are present in
MPA-capped QDs, which corresponds to a higher PL quantum efficiency, however, in the
57
sacrifice of losing its solubility potentially due to the reduction of free carboxylate ions.
Bioconjugation to the QDs might also be limited with a reduced number of thiol terminuses on
its surface. Since the chelating bidentate bindings of the carboxylates-metal complex are stronger
than the unidentate bindings of the thiolates-metal complex, the larger quantity of chelating
bidentate complexes formed in MPA-capped makes it more stable than the TGA-capped ones.
Altogether, both TGA and MPA-capped QDs form carboxylate-metal complexes with the CdTe
core, but the longer alkyl chain in MPA enabled a larger quantity of the complexes to be formed
with it rather than with TGA; thus potentially permitting a higher stabilized QD to be
synthesized with MPA.
Table 4.2: The wavenumber separations, ∆, between the symmetric and asymmetric stretching
modes of the carboxylate, and their corresponding carboxylate-metal.
Capping
Agent
Separation
between
νs(COO)1 and
νas(COO), ∆1 Structure
Separation
between
νs(COO)2 and
νas(COO), ∆2 Structure
TGA 187 cm-1
bridging bidentate 70 cm-1
chelating bidentate
MPA 215 cm-1
unidentate 102 cm-1
chelating bidentate
1 symmetric stretches between 1300 – 1400 cm-1
2 symmetric stretches between 1400 – 1500 cm-1
Table 4.3: Intensity ratios of the symmetric carboxylate stretch to CdS 2SO mode.
Intensity Ratio
Capping
Agent
νs(COO)1 /
CdS 2SO
νs(COO)2 /
CdS 2SO
TGA 0.205 0.547
MPA 12.803 5.390
1 symmetric stretches between 1300 – 1400 cm-1
2 symmetric stretches between 1400 – 1500 cm-1
58
(4.6) Summary
In summary, we demonstrated the use of HC-PCF to obtain efficient Raman scattering of the
different thiol-capped CdTe QDs in aqueous environment. Strong and well-resolved Raman
modes of the CdTe semiconductor core, capping ligands, and their interfacial structures were
successfully observed and compared without integrating any metallic nanoparticles for
enhancement. These modes were also correlated to the QD’s PL quantum efficiency, stability,
water solubility and bio-conjugation capability. To the best of our knowledge, this is the first
time that such strong Raman modes of the thiol-capped CdTe QDs in aqueous solution have been
reported. The enhanced detected Raman signals were achieved through increased light-matter
interaction and efficient accumulation of the Raman scattering signal along the entire length of
the HC-PCF.
59
Chapter 5
Raman Characterization of
Metallic "anoparticles
(5.1) Introduction
Metallic nanoparticles (of size 1 – 100 nm in diameter) synthesized with noble metal, such as
silver and gold, are of great interest in the past decade across a number of disciplines in science,
engineering, and medicine due to their unique and remarkable capability of enhancing optical
properties in the visible and near-infrared range (~400 – 1000 nm).76-79
Optical properties,
including absorption, Rayleigh scattering and Raman scattering, are enhanced because metallic
nanoparticles produce a plasmon resonance in the presence of an external EM wave of light. This
plasmon resonance introduces a range of wavelength in which the optical effects are increased by
orders of magnitude. In particular, Raman scattering enhancement can be induced by an
enhanced EM fields at or near the surface of the metallic nanoparticles. This local EM field
60
enhancement is introduced by an oscillating cloud of electrons at the particle surface, namely,
surface plasmon resonance, when the particle is irradiated by light (Figure 5.1). In the absence of
an external EM field, electrons at the surface of the metallic nanoparticles are free to oscillate
and move around due to the conductive nature of noble metals. However, in the presence of an
external EM wave of light, electron cloud around the particle surface oscillates coherently with
the external EM field, producing an overall EM field outside the particle that is stronger than
incident one. To particles or molecules that are adsorbed or in close proximity (< 10 nm) to the
particle surface, this is seen as one strong EM field with amplitude summing the applied field
and that resulted from the surface plasmon resonance. This strong EM wave resulted from the
surface plasmon resonance is called surface plasmon polariton (SPP). In medical diagnostic and
sensing, SPP from metallic nanoparticles, particularly gold nanoparticles, significantly enhances
the sensitivity of many conventional spectroscopic techniques such as fluorescence and Raman
spectroscopy, enabling a new world of biological materials and systems to be explored.
Figure 5.1: Schematic of plasmon resonance for a metal sphere. Electron clouds of metal
spheres are displaced relative to the nuclei in the presence of an electric field, inducing coherent
oscillations with the incident electric field.76
In the field of cancer diagnostics, metallic nanoparticles are utilized to target and locate
malignant cells in the human body.80-83
When functional metallic nanoparticles are bind to cancer
61
cells in the tissues or organs, they enhance optical signals from cancer indicators adsorbed on the
particle surface to indicate the presence of cell abnormality. The large signal enhancement from
the metallic nanoparticles enables cancers at their early stages to be diagnosed accurately and
quickly, allowing cancer treatments to be much more effective. In fact, metallic nanoparticles
can also be used simultaneously as agents for cancer therapy through ablation of tumors with
rapid conversion (~1ps) of absorbed light into heat.84-90
Typically, gold nanoparticles are used for diagnostics instead of silver or other cadmium–
containing semiconductor quantum dots because gold has little or no long term toxicity or other
adverse effects that can damage or destroy cells during diagnostics.91, 92
Gold is also more
chemically stable compared to copper. Furthermore, spherical gold colloidal nanoparticles can
easily be synthesized in a wide range of sizes (2 – 100 nm) through reduction of gold ions in
solutions.93-96
Nevertheless, previous research has shown that gold nanoparticles with a core size
of ~60 – 80 nm are most efficient for Raman signal enhancements in the red (630 – 650 nm) and
near-infrared (785 nm) those light absorption and autofluorescence is minimal.82, 97, 98
To locate and detect malignant cells, gold nanoparticles are tagged with dyes to give a strong
spectroscopic signal when attached to a malignant cell. Fluorescence dyes can be used to give
bright spectral signals when enhanced by the gold nanoparticles; however, fluorescent peaks
have a large spectral width. The FWHM of a fluorescent peak is typically around 50 – 100 nm,
which limits the number of different dyes that can be multiplexed together for simultaneous
detection of different cells.99-101
Fluorescent signals are also often bleached rapidly by the optical
probe.102-104
Moreover, autofluorescence from biological species interferes with the actual signal
from the dye, making signal detection more difficult.105
Conversely, Raman dyes give narrow
62
spectral features with FWHM of just a few nm, allowing multiple different dyes to be
multiplexed spectrally.106, 107
The narrow linewidth of the Raman signal also differentiates itself
easily from the broad autofluorescence. Further, Raman signals do not photobleach.108
However,
Raman scattering signal are typically much weaker than fluorescence, making cancer cells in
small quantities difficult to detect. Nevertheless, recent experiments demonstrated that SERS
signal can be comparative to fluorescence.80, 81, 109
In addition to dye adsorption, nanoparticles are often coated with a protective layer of polymer or
lipid for improved stability, biocompatibility and circulation in the biological system. Two types
of protective layers are commonly used, particularly for Raman imaging:83
(1) thiol-modified
polyethyleneglycol (PEG) and (2) phospholipid. Thiol-modified PEG coated nanoparticles have
shown to greatly improved the stability of the SERS signal under strong acid, strong base,
concentrated salts and different organic solvents.83
They also prevent biofouling (undesired
accumulation of biological matters on nanoparticles) and allow efficient conjugation of tumor
targeting antibodies.110-112
Recently, phospholipid coated nanoparticles incorporating a Raman
dye is demonstrated.113, 114
Phospholipids maximize the biocompatibility of the nanoparticles as
they mimic membranes of the human cells. They also provide greater stability on the adsorbed
Raman dyes than the thiol-modified PEG and more resistant to aggregation than citrate coated
gold nanoparticles.114
Lastly, for cancer detection, nanoparticles have to be conjugated with antibodies for identifying
and binding to receptors or antigens overexpressed on cancer cells.
Functional gold nanoparticles, with Raman dye and protective coating, for cancer diagnostics
have been demonstrated for different types of cancers in the literature including leukemia,112, 115,
63
116 cervical cancer,
117 head and neck cancer,
83 breast and lung cancer
115 and kidney cancer.
118
These nanoparticle systems are commonly characterized by several techniques to determine their
optical characteristics and to confirm the addition of the functions throughout and after the
synthesis:
1. UV–vis spectroscopy
UV–vis spectroscopy is utilized to confirm the stability of the nanoparticles after
adsorbing the Raman dye and coating the nanoparticle surface. The absorption of
the nanoparticles should remain unchanged or with a minor shift of only a few
nanometers after the synthesis process. A minor shift might be induced by the
addition of a Raman dye or coating material whose absorption peak is higher or
lower than the ones of the nanoparticles. However, if a larger shift (~10 – 100
nm) of the absorption peak is observed, it would indicate that the size of the
metallic nanoparticle is changed. An increase in the particle size could mean that
the surface chemistry of the nanoparticle is changed during the synthesis process
causing further growth or aggregation of the nanoparticles. Broadening or
emerging of new peaks in the spectrum further indicates that the nanoparticles are
aggregated, possibility due to changes in the ion quantity (i.e. Na+ and Ca
2+) and
acidity of the colloidal solution.
2. DLS
DLS provides size distribution of the nanoparticles. It is used to confirm that the
protective layer of the nanoparticle is coated properly in all or most of the
nanoparticles. The addition of a protective layer around the nanoparticle causes an
64
up shift of the size distribution in DLS. Due to the small size of the dye, the
measured particle size normally remains unchanged after dye adsorption.
3. TEM
TEM is used to provide physical images of the nanoparticle structure through the
use of an electron beam. TEM allows a more accurate conformation of the
nanoparticle size and coating of the particles with polymer and phospholipid. A
bright diffuse corona can be observed around the dark coloured nanoparticles in
the TEM image after the nanoparticles are successfully coated with PEG or
phospholipid.
4. Raman Spectroscopy
Raman spectroscopy is used to ensure that Raman dyes are adsorbed on the
nanoparticle surface and generate a SERS signal. In most cases, a strong Raman
dye signal obtained from the nanoparticle solution would consider that the dye
adsorption is successful as the enhanced Raman signal can only be achieved
through the gold nanoparticles.
Although these techniques provide logical evidences of the nanoparticle functionalization, direct
evidences on the actual molecular binding interaction as well as the type of bindings are not
available. For example, UV–vis spectroscopy, DLS and TEM only provide information on the
nanoparticle stability, size and size distributions. Although, shifting of the particle size
distribution after coating the particle is logically reasonable, the structure of the coating layer and
its binding interaction with nanoparticle core is undetermined. Moreover, these techniques do not
describe the stability of the bare nanoparticle (i.e. nanoparticle without any conjugation and
functionalization) which is governed by the binding interaction of the stabilizer and the
65
nanoparticle core. Raman scattering contains bonding specific information between the
stabilizer, Raman dye, polymer or lipid, and the nanoparticles, but the low sensitivity of
conventional Raman spectroscopy prevents these information to be observed. Currently there is
no conventionally technique available to confirm the binding of the functional molecules to the
nanoparticle core at the different synthesis stages. Since the nanoparticle interface provides the
nanoparticle its sensitivity, stability, solubility and capability of interacting with the
biomolecules, it is important that we understand the actual binding between the particle core and
the different functional molecules at the molecular level. The lack of understanding in the actual
particle-molecule binding limits the way which we can diagnosis any practical issues and further
improve the nanoparticle systems to the extent that it can be fully realized for cancer detection.
In the cost perspective, it is very expensive to acquire all four systems for nanoparticle
characterization if a research group or company is to purchase them or utilized facilities from
labs that have them. Therefore, there is an emerging need to development a technique which
enables monitoring of the molecular bindings between the nanoparticles and the functional
molecules.
In this chapter, effort to achieve monitoring of molecular binding interaction at the nanoparticle
surface is established using Raman spectroscopy. We will demonstrate the use of the PCF-
enhanced Raman spectroscopy for characterizing two gold nanoparticles during the different
stages of the synthesis process, namely, the pegylated nanoparticle system and the phospholipid
nanoparticle system. Binding interactions between the nanoparticle core, and the different
functional molecules essential for target-based cancer diagnostics, including the stabilizer,
Raman dye and PEG coating, were observed. In addition, no binding was observed between the
phospholipid coating and gold nanoparticle core which supported that lipid encapsulations are
66
electrostatically governed. Furthermore, the stability of the two nanoparticle systems was tested
under different laser operation for Raman spectroscopy. Degradation of the nanoparticle system
with increasing laser power and exposure time was observed. Carbonization of the nanoparticle
solution was detected at 8.5 – 17 mW of power with a 633 nm laser for exposure duration longer
than 9 minutes. This suggests that localized heating of the gold nanoparticle can alter its binding
interactions with functional molecules which destroy the nanoparticle system, rendering it
nonfunctional for cancer detections or lead to faulty information in medical diagnostics,
particularly in Raman imaging.
(5.2) Experimental Details
Two different gold nanoparticle systems were synthesized for this study, namely, the pegylated
gold nanoparticles tagged with malachite green isothiocyanate (MGITC) dye and the
phospholipid gold nanoparticles tagged with crystal violet (CV). The schematic of the
nanoparticles and the molecular structures of the stabilizer, dye, polymer and lipids are shown in
Figure 5.2. Both nanoparticle systems utilized citrate stabilized colloidal gold particles with a
core diameter of 60 nm. The pegylated and phospholipid gold nanoparticles were provided by
Dr. Gilbert Walker’s group in the Chemistry Department at the University of Toronto and Dr.
Gang Zheng’s group in the Medical Biophysics Department also at the University of Toronto,
respectively. These pegylated and phospholipid nanoparticles were designed and synthesized for
the detection of leukemia and lung cancer cells using the imaging technique with Raman
spectroscopy, respectively. The general scheme for synthesizing the two nanoparticle systems is
shown in Figure 5.3. In general, gold nanoparticles stabilized with citrate were first coated with
the Raman dye. Then, they were coated with either the PEG polymer or phospholipid layer for
further stability. The synthesis details for the nanoparticles systems can be found in literature of
67
Nguyen et al.112
for the pegylated gold nanoparticles and Tam et al.113
for the phospholipid gold
nanoparticles.
Figure 5.2: Molecular structure of (a) pegylated gold nanoparticle system (b) phospholipid gold
nanoparticle system.
68
Figure 5.3: Scheme of the nanoparticle synthesis process for (a) pegylated and (b) phospholipid
gold nanoparticles.
To study the binding interaction between the Raman dye, PEG, phospholipid and the gold
nanoparticle core, nanoparticles from different stages of the synthesis process were obtained and
studied using Raman spectroscopy. Table 5.1 shows the nanoparticle samples obtained for
studying the two nanoparticle systems. Raman characterizations were done using the new HC-
PCF modeled HC-1060 purchased from NKT Photonics. The HC-1060 has approximately the
same central hollow core size (~10 µm) as the HC-800 used for the CdTe study in Chapter 4. The
new HC-1060 was used because it provides greater Raman signal enhancement than the HC-800
due to the larger effective index of the PCF cladding.27
Figure 5.4 shows the Raman spectra of
69
water obtained using HC-800, HC-1060 and conventionally with a cover slide. Intensity of the
OH stretching mode at 3427 cm-1
is about 10 % greater with the use of HC-1060 than HC-800.
Preparation and setup of the HC-1060 for Raman experiments were the same as HC-800
discussed in Chapter 4. However, the fiber splicing program was modified to ensure all cladding
holes were collapsed while leaving the central core open to its maximum for solution filling. For
all measurements, continuous wave (CW) 633 nm HeNe was used with 0.17 mW of power and
averaged over 2 spectra unless otherwise specified.
Table 5.1: Nanoparticle samples studied for pegylated nanoparticle system and phospholipid
nanoparticle system.
Samples
Pegylated
Nanoparticles Citrate Stabilized Gold Nanoparticles
Pegylated Gold Nanoparticles with
MGITC Adsorbed
Pegylated Gold Nanoparticles without
MGITC Adsorbed
Phospholipid
Nanoparticles Citrate Stabilized Gold Nanoparticles
Citrate Stabilized Gold Nanoparticles
with CV Adsorbed
Phospholipid Coated Gold
Nanoparticles with CV Adsorbed
70
Figure 5.4: Raman spectra of water obtained using cover slide, HC-800 and HC-1060.
Enhancement of the OH stretching mode at 3421 cm-1
is about 101 times and 112 times relative
to the conventional technique respectively. Spectra are obtained with 0.17 mW of power with a
HeNe (633 nm) laser.
(5.3) Molecular Interactions in Pegylated Gold "anoparticles
(5.3.1) Citrate Stabilized Gold 4anoparticles
Raman spectra of the pegylated nanoparticles were obtained at the different stages of the
synthesis process. In the first stage, gold nanoparticles were synthesized and stabilized by
citrates to maintain their size and shape for a controlled surface plasmon effect for Raman
enhancements. Although gold nanoparticles used in this experiment, and that of the phospholipid
nanoparticles, were purchased commercially from Ted Pella Inc., these nanoparticles degrade
after a period of storage time, which cause failures of dye adsorption if they are used for
functionalization. Figure 5.5 shows the Raman spectrum obtained from a fresh batch of citrate
stabilized gold nanoparticles. In particular, stretching modes of the carboxylates between 1380
and 1600 cm-1
are of interest here. These Raman modes are associated with the three carboxylic
71
acid groups of the citrate. These carboxylic acid groups are known to form complexes with the
gold surface which stabilizes the nanoparticle core after deprotonation.119, 120
The Raman modes
at 1381, 1445 and 1544 cm-1
are assigned to the symmetric stretches of the carboxylates. The
mode at 1597 cm-1
is assigned to the asymmetric stretch of the carboxylates. Similar to the CdTe
nanoparticles discussed in section 4.3, the wavenumber difference between the symmetric and
asymmetric stretch of the carboxylates mode indicates any complexes formed between the metal
and the carboxylates (i.e. the gold-carboxylate complex in this case).
Figure 5.5: Raman spectrum of a fresh batch of gold nanoparticles stabilized by citrate. Inset:
enlarged spectrum showing carboxylate stretching modes of citrate.
As a reminder to the reader, if the wavenumber difference between the symmetric and
asymmetric modes of the carboxylate is greater than those of the ionic values, a unidentate
complex is formed. Conversely, if the wavenumber difference is much lower than the ionic
values, then a bridging bidentate complex is formed. Lastly, if the wavenumber difference is
greater than that of the bridging bidentate and closes to the ionic values, then a chelating
72
bidentate complex is formed. For liquid solutions similar to citrate, the symmetric stretches of
carboxylate anion are located around 1414 – 1425 cm-1
where the asymmetric stretch of the
carboxylate anion is located around 1560 – 1580 cm-1
.121
Therefore, the wavenumber difference
between the ionic carboxylate symmetric and asymmetric stretching mode is about 145 – 164
cm-1
. With 100 mW of power using an Argon ion laser (514.5 nm), Munro et al. also observed
carboxylate symmetric stretching modes between 1400 and 1575 cm-1
from citrate stabilized
silver nanoparticles.119
Table 5.2 shows the wavenumber position of the symmetric and asymmetric stretching modes of
carboxylate, their respective intensity ratio to the carboxylate asymmetric stretch, wavenumber
differences and proposed complexes according to the discussion above. The table shows that one
carboxyl group of the citrate is ionic which means that it did not form a complex with the
nanoparticle and was exposed on the surface. In addition to the carboxylate anions, citrate
formed unidentate and bridging bidentate complexes with the nanoparticle core. This finding
correlates with the proposed interactions of Munro et al. in which two of the three carboxyl
groups interact with the citrate reduced colloid and the third group is exposed on the surface
observed from his surface enhanced resonant Raman spectroscopy.119
Therefore, fresh gold
nanoparticles used in this experiment were stabilized strongly by citrates through both bidentate
and unidentate bindings. In addition, Table 5.2 shows that the intensity ratio of the symmetric
mode at 1544 cm-1
to the asymmetric mode at 1597 cm-1
is greater than one while the ratio of the
other two symmetric modes are less than one. This shows that more bridging bidentate
complexes were formed than unidentate and ionic complexes which suggests that the bridging
bidentate complex is more favourable in stabilizing the nanoparticle core, probably due to the
stronger bonding strength of the bidentate complex than the unidentate.
73
Table 5.2: Wavenumber position of the symmetric and asymmetric stretching carboxylate
modes of citrate, their intensity ratio to νas(COO), respective wavenumber differences and
proposed complexes for a fresh batch of citrate stabilized gold nanoparticles.
νs(COO) νas(COO)
Intensity Ratio
of νs(COO) to
νas(COO)
Wavenumber
Separation between
νs(COO) and νas(COO)
Proposed
Structure
1381 1597 0.474 216 cm-1
Unidentate
1445 1597 0.867 152 cm-1
Ionic
1544 1597 1.703 53 cm-1
Bridging Bidentate
To compare the structural difference between the degraded nanoparticles and the fresh ones,
Raman spectrum of an old batch (~ 6 months old) of citrate stabilized nanoparticles stored in the
dark at room temperature was obtained (Figure 5.6). The Raman modes at 1386 and 1454 cm-1
are assigned to the symmetric stretches of the carboxylates. Although the water mode at ~1652
cm-1
is strong, the carboxylate stretching modes at 1553 cm-1
and 1597 cm-1
were determined
through fitting three Gaussian functions to the broad peak at ~1600 cm-1
. The Raman mode at
1553 cm-1
is assigned to the symmetric stretch while the 1597 cm-1
is assigned to the asymmetric
stretch of the carboxylates. Table 5.3 shows the carboxylate-metal complexes determined from
the carboxylate stretching modes. The same three complexes were found in the degraded
nanoparticle solution. However, the intensity ratio of the unidentate and ionic complexes are
greater than one instead of less than one in the fresh nanoparticles. The ratio of the bidentate
complexes is also reduced to less than one. This suggests that some of the bidentate complexes in
the nanoparticle solution were changed to unidentate complexes, ionic complexes and probably
even detached from the gold core. Due to the lost of surface enhancement from the gold
nanoparticles, free flowing citrates could not be detected for comparison. Nevertheless, the
reduction of bidentate complexes suggests that stabilization of the gold cores was weakened
74
through long storage time at room temperature. Furthermore, the lost of citrate anions might
reduce the adsorption strength of the Raman dye as well as the phospholipid coating (in the case
of the phospholipid nanoparticles system) through electrostatic interactions.
Figure 5.6: Raman spectrum of six-month-old gold nanoparticles stabilized by citrate. Inset:
enlarge spectrum showing carboxylate stretching modes of citrate.
Table 5.3: Wavenumber position of the symmetric and asymmetric stretching carboxylate
modes of citrate, their intensity ratio to νas(COO), respective wavenumber differences and
proposed complexes for a six-month old batch of citrate stabilized nanoparticles.
νs(COO) νas(COO)
Intensity Ratio
of νs(COO) to
νas(COO)
Wavenumber
Separation between
νs(COO) and νas(COO)
Proposed
Structure
1386 1597 1.157 211 cm-1
Unidentate
1454 1597 1.209 143 cm-1
Ionic
1553 1597 0.658 44 cm-1
Bridging Bidentate
75
(5.3.2) Pegylated Gold 4anoparticles with MGITC Adsorbed
The next step in the synthesis process was to adsorb the Raman dye, MGITC, on to the gold
nanoparticle. However, without attaching PEG polymers to the nanoparticles, MGITC might
detach quickly. Therefore, we obtained Raman spectrum of the pegylated gold nanoparticles with
MGITC adsorbed instead (Figure 5.7). The nanoparticle spectrum shows many intense and
narrow (small FWHM) Raman peaks between 200 – 1630 cm-1
. These peaks correspond to the
vibrational modes of MGITC adsorbed on to the gold nanoparticle surface. Figure 5.8 compares
the Raman spectrum of pure MGITC solution with the pegylated nanoparticles before
background subtraction. The MGTIC spectrum before background subtraction not only contains
a broad background from the fluorescence, the intensity of the Raman modes is also much lower
than those obtained from the pegylated nanoparticles. This is because the strong fluorescence of
MGITC is quenched by the enhanced local EM field at the gold surface in the pegylated
nanoparticles. Since Raman modes do not photobleach (i.e. cannot be quenched by strong EM
field), they retained in the spectrum. In fact, the surface enhancement of the gold nanoparticles
increased the Raman scattering signal from MGITC, resulted in a much intensified Raman
spectrum of MGITC. The matching Raman modes between the two spectra further confirms that
the intense Raman modes in the nanoparticles spectrum are those of MGITC. Therefore, the
intensified Raman signal of MGITC in the nanoparticle spectrum was due to surface
enhancement of the gold nanoparticles; hence indicating that MGITC was adsorbed on to the
nanoparticles surface.
76
Figure 5.7: Raman spectrum of pegylated gold nanoparticles adsorbed with MGITC obtained
with 0.17 mW.
Figure 5.8: Raman spectrum of pegylated gold nanoparticles with MGITC adsorbed and pure
MGITC solution before background subtraction (a) and after background subtraction (b). Spectra
obtained with HeNe laser (633 nm) at 0.17 mW.
77
(5.3.3) Pegylated Gold 4anoparticles
It is believed that the sulphur atom in the thiol-modified PEG binds to the gold nanoparticles
core after replacing the citrate as the stabilizer because of the strong binding affinity of the
sulphur atom.122-125
However, Raman modes of citrate are much weaker than MGITC (as shown
when we compare the intensity counts of the MGITC modes in Figure 5.7 and the citrate modes
in Figure 5.5) due to the strong Raman cross-section of the phenyl rings in MGITC.126
Therefore,
the absence of citrate modes in the pegylated nanoparticle spectrum in Figure 5.7 does not
indicate that the citrates were desorbed from the nanoparticles surface. If citrates were still
stabilizing the gold nanoparticle core, the strong MGITC modes would most likely be overlapped
by the carboxylate stretching modes of citrate since they overlap between the 1390 – 1600 cm-1
regions of the spectrum. Similarly, Au-S stretching modes at the lower Raman shift regime
around 300 cm-1
would also be overlapped by the strong MGITC modes. Thus, it is difficult to
determine the bindings between the PEG, citrate and the nanoparticle core after the Raman dye is
adsorbed.
To determine the binding interaction between thiol-modified PEG and the nanoparticles, we
obtained Raman spectrum of the pegylated gold nanoparticles without any Raman dye adsorbed
(Figure 5.9). At a higher pump power (i.e. 8.5 mW), the spectrum shows stretching mode of Au-
S bonds at 305 cm-1
in addition to the carboxylate stretching modes from gold-citrate bindings.
Stretching modes of Au-S are typically reported to be between 200 – 240 cm-1
.127
However,
when the sulphur atom is attached to a long polymer chain, this mode is shifted to the Raman
shift region between 300 – 310 cm-1
.128
Nevertheless, this mode would have been overlapped by
the MGITC mode at 233 cm-1
in the MGITC adsorbed nanoparticle sample; thus, it was not
observed in the spectrum of the nanoparticles with MGITC adsorbed. Nevertheless, the Raman
78
mode at 305 cm-1
observed with the pegylated nanoparticles without MGITC indicates that thiol-
modified PEG polymers were bind to the surface of the gold nanoparticles. Moreover, we have
confirmed that the binding was made through the gold-sulphur bond. Figure 5.10 shows the UV–
vis spectra, TEM and DLS results of the same pegylated nanoparticles obtained from a different
batch of synthesis. The UV–vis spectra show that absorption of the nanoparticles remained the
same after pegylation which means that the size of the gold core did not changed and the desired
surface plasmon effects were maintained after pegylation. The bright “white” ring around the
nanoparticle in the TEM image shows that PEG polymers were functionalized on the
nanoparticle surface, indicating that pegylation was successful. Moreover, the size up-shift in the
DLS histogram shows that the average nanoparticle sizes were increased in the system. This
further confirms that the nanoparticles were functionalized with PEG polymers as we have
observed with the Raman spectra. Although these results were obtained from a different batch of
nanoparticles, they indicate that the pegylation method used in this experiment for Raman
characterization works and that our analysis of the Raman results are promising.
Figure 5.9: Raman spectrum of pegylated gold nanoparticles. Inset left: Raman mode of Au–Cl
at 259 cm-1
and Au–S at 305 cm-1
. Inset right: Stretching modes of carboxylates.
79
Figure 5.10: UV–vis (a), TEM (b) and DLS (c) results obtained from a pegylated nanoparticle
system synthesized using the same method and materials.112
To our surprise, the carboxylate stretching modes are still observed in addition to the Au-S
stretching mode. This suggests that only some of the citrates were replaced with the Au-S bonds
from pegylation. The wavenumber separations between the symmetric and asymmetric mode are
shown in Table 5.4. The carboxylate-gold complexes determined in the table show that citrates
are still stabilizing the gold nanoparticles through both the unidentate and bridging bidentate
complexes after pegylation. Thus, the strong binding affinity of the sulphur atom did not
completely replace the citrate as the stabilizer for the gold nanoparticles in this case, might be
80
due to the lack of PEG polymers. The intensity ratios of the complexes show that most of the
carboxylate–gold bindings are unidentate instead of the stronger bridging bidentate, which
suggests that many of the bridging bidentate complexes were removed or changed to unidentate
complexes through breaking one of the Au-O bonds as a result of the strong affinity of the Au–S
bindings.
Table 5.4: Wavenumber position of the symmetric and asymmetric stretching carboxylate
modes, their respective wavenumber differences and proposed complexes in pegylated gold
nanoparticles.
νs(COO) νas(COO)
Intensity Ratio
of νs(COO) to
νas(COO)
Wavenumber
Separation between
νs(COO) and νas(COO)
Proposed
Structure
1375 1595 1.366 220 cm-1
Unidentate
1465 1595 0.750 130 cm-1
Ionic
1539 1595 0.525 56 cm-1
Bridging Bidentate
It is important to note that the carboxylate stretching modes of the citrate stabilized nanoparticles
and the Au-S stretching mode of the pegylated nanoparticles were not observed using
conventional Raman spectroscopy when the nanoparticle solutions were studied on a cover slide
(see Appendix C.1 and C.2). The background-to-signal ratio of the MGITC modes from MGITC
adsorbed nanoparticles was also enhanced with HC-PCF compared to those obtained using the
conventional Raman techniques (Appendix C.3). Approximate enhancements of the detected
Raman scattering signal from the different nanoparticle solutions using the HC-PCF platform is
calculated in Appendix C. The enhanced Raman signal can be attributed to the efficient Raman
scattering of the particles using the HC-PCF platform. Furthermore, the short exposure time (2 s)
and low power exposure (hundreds of microwatts to a few milliwatts) of the pump laser ensured
81
that the characteristics of the nanoparticle system were most representative to its as-synthesis
state due to the highly non-destructive nature of the PCF platform.
(5.4) Molecular Interactions in Phospholipid Gold "anoparticles
(5.4.1) Citrate Stabilized Gold 4anoparticles
Citrate stabilized gold nanoparticles and CV adsorbed gold nanoparticles, before and after
coating with phospholipids, were obtained for Raman characterizations. First, commercially
synthesized 60 nm gold nanoparticles stabilized by citrate was characterized using Raman
spectroscopy with HC-PCF (Figure 5.11). Similar to the citrate stabilized gold nanoparticles used
for synthesizing the pegylated nanoparticles, carboxylate stretching modes of the citrates is
observed between 1390 and 1600 cm-1
. The Raman modes at 1337, 1446, and 1573 cm-1
are
attributed to the symmetric stretching modes of carboxylate. The Raman mode at 1608 cm-1
is
attributed to the asymmetric stretching modes of carboxylate. Table 5.5 shows the proposed
carboxylate-metal complexes derived from the symmetric and asymmetric stretches of the
carboxylate Raman modes. The same three complexes (unidentate, bridging bidentate and ionic)
were formed as the citrate stabilized nanoparticles studied in the pegylated nanoparticle system.
This is reasonable as both citrate stabilized gold nanoparticles are 60 nm in diameter and they
were both purchased from Ted Pella Inc. However, the intensity ratios of the three complexes
measured in this experiment are similar which differs from the pegylated nanoparticle system in
which more bidentate complexes were presented. This difference could be related to the different
batches of the nanoparticle purchased or the different duration and condition of the storage time.
Nevertheless, the similar structures formed between the citrate and the gold core shows that both
nanoparticle systems were stabilized similarly at the time of synthesis.
82
Figure 5.11: Raman spectrum of 60 nm gold nanoparticles stabilized by citrate. Inset: enlarged
spectrum showing carboxylate stretching modes of citrate.
Table 5.5: Wavenumber position of the symmetric and asymmetric stretching carboxylate
modes, their respective wavenumber differences and proposed complexes for a fresh batch of
citrate stabilized gold nanoparticles for synthesizing the phospholipid nanoparticle system.
νs(COO) νas(COO)
Intensity Ratio
of νs(COO) to
νas(COO)
Wavenumber
Separation between
νs(COO) and νas(COO)
Proposed
Structure
1337 1608 0.717 271 cm-1
Unidentate
1446 1608 0.392 162 cm-1
Ionic
1573 1608 0.556 35 cm-1
Bridging Bidentate
(5.4.2) CV Adsorbed Gold 4anoparticles
The next step in the synthesis process is to attach the Raman dye (i.e. CV) onto the nanoparticle
surface to provide a Raman signal. Raman spectrum of the CV adsorbed nanoparticles is shown
in Figure 5.12 along with the spectrum of CV solution before and after background subtraction.
Similar to MGITC solution, Raman spectrum of the CV solution contains a broad fluorescent
signal overlapping Raman modes of CV as shown in the raw Raman spectrum. In comparison,
83
Raman spectrum of the CV adsorbed gold nanoparticles shows enhanced Raman modes of CV.
This enhancement can be attributed to the surface enhanced Raman scattering due to the gold
nanoparticles after CV was adsorbed on it. Therefore, this enhancement confirms that the CV
dye was successfully adsorbed onto the gold nanoparticles surface. After the removal of
fluorescent signal (Figure 5.12b), it can be seen that the Raman modes from CV adsorbed gold
nanoparticles match closely with those from the CV solution. This further confirms that the
enhanced Raman modes are CV Raman modes; thus, CV dyes were adsorbed onto the gold
surface successfully.
Figure 5.12: Raman spectrum of CV adsorbed gold nanoparticles and CV solution without
background subtraction (a) and with background subtraction (b).
84
(5.4.3) Phospholipid Gold 4anoparticles Adsorbed with CV
The final step in preparing the phospholipid nanoparticles is to coat the CV adsorbed gold
nanoparticles with phospholipid layers for strong biocompatibility. Typically, TEM images and
DLS are used to confirm the effectiveness of the coating after the synthesis process. Raman
spectroscopy is difficult to obtain much information about the coating as phospholipid layers do
not actually bond with the gold nanoparticles core, citrate or the Raman dye. Therefore, we
expect to see no new emerging or disappearances of Raman modes to show that no bonding has
actually formed. However, in conventional Raman spectroscopy, even if Raman spectra of the
phospholipid nanoparticles resemble those of the CV adsorbed nanoparticles without the
phospholipid, the low sensitivity of conventional systems might have prevented weak changes
from being observed. With the use of HC-PCF, the results hold more promises as it provides
higher sensitivity than conventional techniques. It is important to note that although an extra C-C
bond is formed between DMPC (C36H72NO8P) and MHPC (C22H46NO7P) lipids that forms the
phospholipid coating, the addition of one C-C stretch mode is difficult to observe in the
spectrum. This is mainly because of two reasons. First, C-C stretching mode is weak relative to
the phenyl rings of CV. Second, the phospholipid layer is isolated from the gold nanoparticle
surface by the citrate due to their electrostatic interactions; therefore, Raman signals of the lipids
are not enhanced more than the Raman dye. As a result, the strong CV modes would have
overlapped the C-C stretches of the lipids, preventing it from being observed.
Figure 5.13 compares the Raman spectrum of CV adsorbed gold nanoparticles with and without
the phospholipid coating. Both of the Raman spectra contain the same Raman modes which are
those of the CV dye as we expect due its strong Raman cross-sections. More importantly,
additional C-C stretching mode or mode shifting, broadening and relative intensity changes of
85
the CV modes are not observed. This provides supporting evidence that the phospholipid layer
did not form bonds with the gold core or the Raman dye and potentially binding with the gold
nanoparticles core through electrostatic interaction. In addition, intensity of the CV modes from
phospholipid nanoparticles is found to be reduced compared to that without the coating. This is
probably due to the higher light absorption of the phospholipid at the 633 nm laser line and the
scattered Raman signal which confirms the presence of the phospholipid. UV–vis spectra
throughout the synthesis process are shown in Figure 5.14. These spectra were obtained from a
different batch of nanoparticles synthesized using the same recipes as this experiment. UV–vis
spectra show that the absorption peak of the nanoparticles remained unchanged after the addition
of CV and phospholipid coating. This confirms that the surface plasmon resonance of the
nanoparticles was retained. In addition, it is important to note that the absorption peak of the
nanoparticles after coating with the phospholipid is shifted slightly to the higher wavelength.
This correlates with our speculation that the absorption of the phospholipid is higher in the red
part of the spectrum which reduced the Raman scattering of the Raman dye. Thus, this confirms
the presence of the phospholipid layer and that it did not bond with the nanoparticles core to
form the lipid coat.
Figure 5.13: Raman spectrum of CV adsorbed gold nanoparticles before and after coating of
phospholipid layer. 0.17 mW of CW HeNe laser was used with 2 s exposure time.
86
Figure 5.14: UV–vis spectra of the phospholipid nanoparticles throughout the synthesis
process.113
The use of HC-PCF enhanced the detected Raman signal enabling Raman modes of citrates to be
determined. Raman modes of CV from the nanoparticles were also enhanced compared to
conventional technique. A comparison of the Raman signals between the use of a HC-PCF and a
cuvette from phospholipid nanoparticles at the different stages are shown in Appendix C.4 – C.6.
(5.5) Pump Power and Duration Dependence of Pegylated Gold "anoparticles
The second part of this characterization chapter is to study real time changes of the nanoparticle
systems under different pump power and exposure durations. Although Raman spectroscopy is
known to provide non-destructive characterization of the sample species, high power exposure
and long exposure durations on stabilized nanoparticles could still destroy their bindings with the
stabilizer, Raman dye, and coating; thus, deteriorating the nanoparticles system and rendering it
not functional.
To study effects of the laser power and exposure time on nanoparticles, Raman spectra of the
pegylated gold nanoparticles with MGITC adsorbed were obtained continuously under different
87
laser powers (i.e. 0.17, 1.7, 4.3, 8.5 and 17 mW). Beginning with 0.17 mW, Raman spectra were
obtained continuously with 2 s exposure time (averaging over 2 spectra) using CW HeNe (633
nm) laser. Due to the wide scanning range of the spectra, the total laser exposure time on the
nanoparticles was about 30 s per measurement. A 15 s gap with no laser exposure was in place
after each measurements to allow the grating to move back to its initial position before the next
measurement begun. Spectra were taken continuously at the same laser power until the spectrum
was stable (i.e. no significant broadening, intensity drop, or mode shifting) for 2 – 3 spectra.
Figure 5.15 shows the Raman spectra of MGITC adsorbed pegylated gold nanoparticles obtained
with power ranging from 0.17 mW to 17 mW (i.e. the maximum power output of the pump
laser). These spectra show intense Raman modes of MGITC between 200 and 1650 cm-1
as
discussed in section 5.3.2. MGITC modes obtained at 0.17 mW and 1.7 mW are identical
although their intensities are reduced in subsequent spectra at the same laser power (i.e. with
prolonged exposure time at the same power). With power increased to 4.25 mW, all MGITC
modes still retain in the spectra except that the intensity of the modes fluctuate as shown in
spectrum 10 and 11. Mode intensity in spectrum 10 is reduced relative to the previous spectrum
(i.e. spectrum 9). However, the intensity improved again in subsequent spectrum (i.e. spectrum
11). In addition, there are also some changes in the relative intensities of the peaks around 1400
and 1500 cm-1
. These intensity fluctuations and relative intensity changes could be caused by the
ensemble average orientations of the MGITC dye that were changing dynamically relative to the
gold surface during the measurements.83, 129
Since SERS enhancement is greatest when
molecular bonds are perpendicular to the gold surface, orientation changes in the molecular
bonds would lead to changes in the mode intensities and relative intensities. Nevertheless, all
Raman modes of MGITC are present in the spectra (compared to the spectra obtained with
88
Figure 5.15: Raman spectra of MGITC absorbed pegylated gold nanoparticles with increasing
power from 0.17 mW (a) to 1.7 mW (b), 4.3 mW (c), 8.5 mW (d) and 17 mW (e). Spectra
obtained with the same power were taken with increasing numerical value. The pump laser was
exposed for a total of 30 s for each measurement.
89
0.17 mW); therefore, no major structural changes of the nanoparticles system are observed at this
power level. However, at about 8.5 mW, Raman modes at 1447, 1486, 1512 and 1529 cm-1
(the
weak modes between the two strong ones around ~1500 cm-1
) begun to broaden. At the
maximum power of our pump source (i.e. 17 mW), intensities of the MGITC modes begin to
reduce and Raman modes between 1183 and 1627 cm-1
broaden further until only two broad
features, with FWHM of about 150 – 300 cm-1
, center at 1353 and 1583 cm-1
were left after
prolonged measurements. The intensity reduction of the MGITC Raman modes indicates that
MGITC molecules were detaching from the gold nanoparticle core. Since the intense MGITC
Raman modes were result of the surface enhancement of the gold nanoparticles, mode intensities
of the MGITC dye would significantly reduce when some MGITC dyes are detached from the
nanoparticle surface. MGITC detachment could be caused by excessive heating of the
nanoparticles upon long laser exposure time at high power. Metallic nanoparticles can be
efficiently heated by laser illumination at or near their plasmon resonance.130
Govorov et al. has
also reported that the surface of the nanoparticles could even heat up by almost 25 degrees
Kelvin with an array of nanoparticles.131
Furthermore, it was observed that an increased quantity
of nanoparticles would cause a stronger increase in the temperature of the system. Since
nanoparticles tends to cluster around the inner side wall of the PCF core (Figure 5.16), probably
due to electrostatic effect, significant temperature increase might be experienced by the
nanoparticles system causing decomposition and detachment of the MGTIC dye. On the other
hand, MGITC itself might have also decomposed due to strong absorption of the 633 nm pump
laser since the pump wavelength locate at the absorption peak of MGITC solution (Figure 5.17).
90
Figure 5.16: Optical image of nanoparticle clusters around the inner wall of the PCF core after
about 1 minute of selective filling. Bright dots around the PCF core at the inner wall are the
nanoparticles.
Figure 5.17: Absorption of MGITC.132
Red line shows the excitation of the HeNe laser at 633
nm which locates at the absorption peak of the MGITC dye.
The cause of mode broadenings at 1353 and 1583 cm-1
is not as clear. However, this broadening
occurs at the spectral region where the carboxylate stretching modes were observed in the citrate
stabilized nanoparticles. Therefore, we further investigated the molecular changes of the citrate
stabilized gold nanoparticles with increasing laser power and exposure time (Figure 5.18). With
pump laser power between 0.17 mW and 8.5 mW, citrate modes were observed although
91
intensity of the modes around 1600 cm-1
were much more enhanced than the other modes.
Nevertheless, the three symmetric stretching modes and one asymmetric stretching modes of
carboxylate could be observed between 1390 and 1600 cm-1
in these spectra. However, the
Raman mode at 1585 cm-1
broadened when power was increased from 4.3 mW to 8.5 mW. At
maximum power of the pump laser (i.e. 17 mW), Raman modes between 1100 and 1700 cm-1
broadened just like the spectra obtained from the MGITC adsorbed pegylated gold nanoparticles
shown in Figure 5.15e. After prolonged laser exposure, two similar broad peaks at 1353 and
1583 cm-1
were also observed. This suggests that the two broad peaks observed from the MGITC
adsorbed pegylated gold nanoparticle spectra were caused by changes in citrate-gold bindings.
The position and broadness of two peaks at 1353 and 1583 cm-1
find similarities to those
reported for carbon related structures including amorphous carbon film,133
graphite,134, 135
and
carbon nanotubes.136-138
In fact, these signals have been observed for small quantity of ‘graphitic’
carbon in silver (i.e. less than a monolayer of amorphous carbon).139, 140
In these literatures, two
carbon related peaks were reported: (1) between 1280 – 1450 cm-1
and (2) between 1520 – 1600
cm-1
. The former is referred to as the D peak (D for disordered) in which the feature correlates to
the bond angle disorder between carbon atoms. A highly disordered carbon structure, such as
benzene clusters, contributes to an intense D peak in the Raman spectrum.133
The latter carbon
peak is referred to as the G peak (G for graphite) as it is correlated to graphite-like structures.
The G peak may also arise from C=C stretch vibrations of olefinic (unsaturated hydrocarbon
chain having a general formula CnH2n which involves a carbon double bond) or conjugated
carbon chains and aromatic rings such as benzene rings.141-143
Both the D peak and G peak might
have linewidth ranging from 10 to 160 cm-1
depending on the carbon cluster size, distribution,
stress and environment.
92
Figure 5.18: Raman spectra of citrate stabilized gold nanoparticles with increasing power from
0.17 mW (a) to 1.7 mW (b), 4.3 mW (c), 8.5 mW (d) and 17 mW (e). Spectra obtained with the
same power were taken with increasing numerical value. Spectrum 25 and 26 are not shown due
to saturation of the detector over the whole spectrum. The pump laser was exposed for a total of
30 s for each measurement.
93
According to Schwan et al.,133
hydrogenated amorphous carbon film with G linewidth broader
than 45 cm-1
contains mostly small aromatic clusters with cluster sizes less than 1 nm. Although
this result was concluded from data obtained with 514.5 nm laser line, the broad G line in the
MGITC adsorbed gold nanoparticles (FWHM of 124 cm-1
) suggests that the actual structure
formed should not differ by a large extent.
The source of carbon for these aromatic clusters is not straight forward to us. Formation of
amorphous carbon structures were observed in varies conditions with different carbon sources
involved. For example, Raman signal of amorphous carbon was observed in silver island films
prepared by physical vapour deposition at room temperature on sapphire in vacuum below 10-9
Torr in which the only possible source of carbon is CO from the rest gas.144
Carbon
intermediates were also observed by Raman spectroscopy on surface of the silver electrode in
which intermediates were believed to be formed from dissolved CO2.145
Moreover, Otto et al.
reported that G and D peak at 1350 and 1595 cm-1
were observed when a freshly polished silver
sample was dipped into cyanide solution exposed to air.146
Tsang et al. proposed that adsorbed
amorphous carbon with short range order of graphite was formed.139
Tsang et al. further showed
that Raman signal of amorphous carbon signal increased under irradiation with 514.5 nm
radiation of 10 W/cm2 while signal of a monolayer of benzoic acid on a rough silver film
(evaporated on a rough CaF2 substrate film) at room temperature was decreased. This suggested
that amorphous carbon layers were formed at the expense of the benzoic acid monolayer.
Recently, Hore et al. reported that carbon layers could be formed on gold by simply blending
gold oxide with inert aliphatic polymers at 250 oC in which the carbon nanolayers were
stabilized by anionic gold after melting of the polymer.147
In addition, we have also observed the
94
D and G peaks at 1375 and 1580 cm-1
on polymer encapsulated gold and silver nanoparticles
obtained from Vivo Nano Inc. (Figure 5.19).
Figure 5.19: Raman spectra of polymer encapsulated gold and silver nanoparticles obtained
from Vivo Nano Inc.
In our experiment, the possible sources of carbon can be contaminants in the nanoparticle
solutions resulted from its synthesis process or carbon species formed from decompositions of
the citrate ligands during the Raman spectra acquisitions. Decompositions could be possible in
citrate stabilized gold nanoparticles if surface of the gold nanoparticles were oxidized. With
excessive heating, thermal denaturation in citrate could leads to formations of short polymer
chains28
and then amorphous carbon structures. Raman signal of small quantities of amorphous
carbon due to thermal denaturation of free flowing citrates might also be possible since
amorphous carbon signals are very strong in Raman. However, further investigation is required
to confirm the source of carbonization in these citrate stabilized nanoparticles.
95
(5.6) Pump Power and Duration Dependence of Phospholipid Gold
"anoparticles
The effects of pump power and exposure time on CV adsorbed phospholipid nanoparticles are
found similar to the pegylated nanoparticle system. Figure 5.20 shows the Raman spectra of CV
adsorbed phospholipid gold nanoparticles obtained with power ranging from 0.17 mW to 17
mW. The phospholipid layers were formed by SHPC (C26H54NO7P) and DMPC instead of
MHPC and DMPC; however, the absorption and Raman spectra of the nanoparticle system are
the same between the uses of the two different lipid layers. Similar to the pegylated nanoparticle
study, spectra were obtained continuously until spectrum remained stable (i.e. no significant
broadening, intensity drop or mode shifting) for 2 – 3 spectra.
At 0.17 mW and 1.7 mW, strong CV modes are observed indicating that CV molecules were
successfully adsorbed onto the gold nanoparticle surface. For 8 to 10 spectra, CV Raman modes
did not shift nor disappeared; however, intensity of the CV Raman modes drops by about 50 –
60% relative to those obtained in the very first spectrum at each power. This suggests that
prolonged laser exposure might cause minor detachment of the CV molecules, probably due to
increased local temperature from clustering and heating of the gold nanoparticles or
photodecomposition of the CV due to absorption of the 633 nm laser line. In addition, the
relative intensity of the mode at 1625 cm-1
reduced significantly with prolonged laser exposure.
The mode at 1625 cm-1
is attributed to the stretching mode between one of the three phenyl rings
and a nitrogen atom. The intensity reduction of this mode suggests that there is a photo induced
decomposition of the CV dye at the phenyl ring – nitrogen bond. Nevertheless, strong CV
Raman modes retained and stabilized after 8 – 10 measurements.
96
Figure 5.20: Raman spectra of CV adsorbed gold nanoparticles with increasing power from
0.17 mW (a) to 1.7 mW (b), 4.3 mW (c), and 17 mW (d). The pump laser was exposed for a total
of 30 s for each measurement.
97
At 4.25 mW, however, intensity of the Raman modes not only reduced further, spectral region
between 1000 and 1760 cm-1
broadened and transformed into two broad peaks around 1377 and
1583 cm-1
. At maximum power of the pump laser (i.e. 17 mW), the spectrum is further
converged into two broad peaks, at 1353 and 1594 cm-1
, similar to that observed from the
pegylated nanoparticles. This suggests that clusters of carbon structures were formed, most likely
for the same reason as the pegylated nanoparticles. Moreover, narrow CV features have
disappeared which indicate that CV molecules were detached from the gold surface as surface
enhanced signals from the gold are not observed anymore. Furthermore, we have observed
further relative intensity reduction of the phenyl ring–nitrogen bond at 1625 cm-1
which indicates
that more CV dyes were photodecomposed. The decomposed CV dyes might have contributed to
the formation of the amorphous carbon clusters through forming phenyl ring clusters.
To confirm that the two broad peaks at 1353 and 1594 cm-1
were also contributed by citrates
stabilizing the gold nanoparticles through carbonizations as determined from the pegylated
nanoparticle experiment in section 5.5, molecular changes of the citrate stabilized gold
nanoparticles used in this experiment were investigated using Raman spectroscopy with
increasing laser power and exposure time (Figure 5.21). At 4.3 mW, the spectrum of citrate
stabilized gold nanoparticles converged into two peaks at 1353 and 1576 cm-1
. The peak at 1353
cm-1
is attributed to the D peak of carbon structure whereas the peaks at 1575 and 1594 cm-1
are
attributed to the G peak. These pairs of carbon peaks indicate that carbon structures were formed
in the nanoparticle system probably upon thermal denaturation of citrates as discussed at the end
of section 5.5. It is to our surprise that the intensity of these two carbon bands was reduced at 8.5
mW. A possible explanation is that carbon structures were only formed near the beam waist of
98
the laser where the intensity was the greatest. Slight shifting of the carbon structures or the laser
spot might have led to reduced signal of the carbon structures.
It is important to note that although carbonization effect in phospholipid nanoparticles is
observed at 4.25 mW whereas the same effect is observed at 17 mW in pegylated nanoparticles,
the pegylated nanoparticle was exposed to the pump laser for 4 times longer in duration (~120 s)
at 0.17 mW and almost twice longer (~120 s) at 1.7 mW. Therefore, excessive heating of the
nanoparticles at the lower pump power could still lead to carbonization in the nanoparticles
Figure 5.21: Raman spectra of citrate stabilized gold nanoparticles, used for phospholipid
nanoparticle synthesis, with increasing power from 0.17 mW (a) to 1.7 mW (b), 4.3 mW (c),
and 8.5 mW (d). The pump laser was exposed for a total of 30 s for each measurement.
99
solution as observed in the phospholipid nanoparticles. Nevertheless, the appears of the two
broad peaks around 1350 and 1580 cm-1
in the two nanoparticle systems show that both systems
are subject to carbonization at high laser power in short exposure time or low laser power at
prolonged laser exposure time. The detail changes of the nanoparticles system obtained in real
time was enabled by the HC-PCF with its nanoliter sampling volume and high sensitivity.
(5.7) Summary
In this chapter, pegylated and phospholipid gold nanoparticle systems were characterized at the
different synthesis stages using Raman spectroscopy. Bindings between citrate, PEG and gold
nanoparticle surface as well as adsorption of the MGITC and CV dyes were observed. The
complexes formed between carboxylate ions in the citrate and the gold surface as well as
bonding between sulphur atoms in the PEG and the gold core were observed nondestructively
only with the enhanced detected Raman signal using HC-PCF. In addition, no bonding was
observed between the phospholipid coating and the gold core even with the more sensitive HC-
PCF platform. This further supported the electrostatic interaction hypnotized between the two.
Furthermore, the effects of the pump laser power and prolonged exposure time on the two
nanoparticle systems were investigated in real time using HC-PCF. Decomposition of the Raman
dyes and formation of carbon structures were observed in both nanoparticle systems at high
pump power and prolonged exposure time. As these nanoparticles were designed for imaging of
cancer cells using Raman spectroscopy, these results suggest that appropriate laser parameters
should be considered for in vivo imaging studies. Particle stability under different thermal
conditions and laser operations should be characterized for the different nanoparticle designed.
Otherwise, high power density of the pump laser could deteriorate the nanoparticle system,
which not only renders it not functional, but also potentially lead to false imaging information.
100
Raman characterization of functional nanoparticles demonstrated in this chapter is most
promising with HC-PCF for its simplicity, superior sensitivity and truly nondestructive nature
that are unachievable conventionally.
101
Chapter 6
Conclusion
(6.1) Summary
Colloidal nanoparticles have proven to be a promising candidate for a wide variety of
applications including photovoltaics, optoelectronics, medical diagnostics, medical therapy and
many more. The complex requirements of these applications require design and synthesis of the
nanoparticles to be controlled and monitored down to the molecular level where binding
interactions can be observed. However, conventional characterization techniques have been
limited to certain properties of the nanoparticles, their compositions, or their physical structure in
the micrometer scale in which molecular bindings and detailed molecular complex cannot be
obtained. In this work, a new Raman spectroscopy technique utilizing PCF has been
demonstrated to enable sensitive detection of different colloidal nanoparticle systems down to
their molecular world.
102
First, in chapter 3, we demonstrated that the detected Raman signal obtained with HC-PCF is
enhanced by about 2 to 3 orders of magnitude compared to conventional Raman spectroscopy
with a cuvette. In addition, we showed both experimentally and theoretically by adapting to the
theory of liquid core waveguides that the enhancement increases with increasing length of the
PCF. The theoretical results further showed that the enhancement reaches a plateau at about 500
cm. By comparing the detected Raman signal enhancement of the PCF with that from TCT, we
showed that the enhancement in PCF is due to both the increased light–matter interactions and
increased collection efficiency of the signal. Nevertheless, it is not trivial to completely decouple
the enhancement due to each of the two components. However, for PCF lengths that are less than
approximately 10 cm, we roughly estimated that the increased light-matter interactions enhances
the Raman signal by about 3 to 4 times while the increased collection efficiency enhances the
signal by about 20 times.
In chapter 4, we further demonstrated that the detected Raman signal is enhanced by about 1 to 2
orders of magnitudes by selectively filling the PCF to enable light guidance by both TIR and
photonic bandgap effects. By selectively filling the PCF with thiol-capped CdTe nanoparticles,
we demonstrated that strong and well-resolved Raman modes of the CdTe semiconductor core,
capping ligands, and their interfacial structures were successfully observed and compared
without integrating any metallic nanoparticles for enhancement. To the best of our knowledge,
this is the first time that such strong Raman modes of the thiol-capped CdTe QDs in aqueous
solution have been reported. These strong Raman modes enabled many binding interactions and
properties of the QDs to be compared between the uses of different capping ligands. First,
intensity ratios of the Te A1 mode to CdTe LO mode showed that MPA-capped QDs were more
crystalline than that stabilized by TGA and TG which agreed with the higher PL quantum
103
efficiency of MPA estimated from the measured absorbance and PL. Second, the sulphur heavy
CdS0.7Te0.3 interfacial layer observed from the Raman spectra showed that thiol capping agents
stabilized the QDs through their sulphur atom as speculated in many literatures. Moreover, the
strong surface passivation by the CdS0.7Te0.3 component and the absence of photo-oxidization
observed from the Raman spectra demonstrated strong stability of these thiol-capped
nanoparticles. Third, the carboxylate stretching modes observed in the Raman spectra
demonstrated the co-existence of carboxylate–metal complexes. In particular, TGA was found to
form bridging bidentate complex and chelating bidentate complex while MPA was found to form
unidentate and chelating bidentate complexes. Moreover, the longer alkyl chain MPA was found
to enable a larger quantity of the complexes to be formed compared to TGA which stabilized the
CdTe core further but potentially weakened its ability for strong solubility and bioconjugation.
Finally, in chapter 5, we presented the characterization results of a pegylated gold nanoparticle
system and a phospholipid gold nanoparticle system for cancer detection. In both nanoparticle
systems, we observed bindings between the citrate ligand, Raman dye and the gold nanoparticle
core through efficient Raman scattering with the PCF platform. Carboxylate groups in the citrate
stabilizers were found to form ionic, unidentate and strong bridging bidentate complexes with the
gold core which correlates with proposed binding interactions in the literature. Through
comparing the Raman spectrum between fresh and degraded citrate stabilized nanoparticles, we
found that nanoparticle stability were weakened in the degraded nanoparticles as the amount of
strong bridging bidentate complexes were reduced compared to ionic and unidenate complexes.
In the pegylated nanoparticle system, PEG–Au bonds were observed in the Raman spectrum
which confirms the pegylation of the nanoparticles. In the phospholipid system, no modes
correspond to any bond between the phospholipid and gold core was observed which supported
104
the hypothesis that the bindings were governed by electrostatics. These results demonstrated that
efficient Raman scattering of the PCF enables metallic nanoparticle systems to be characterized
through the different stages of the synthesis process; thus, allowing nanoparticles to be
monitored in situ for quality control. In addition, we demonstrated that high pump power or
prolonged measurement time can cause Raman dye decomposition and formation of amorphous
carbon structures in both nanoparticle systems. These observations provided insights on the
power sustainability of the nanoparticles and their limits on the acquisition power and time for
Raman imaging.
In summary, this thesis presents an ultra-sensitive platform for Raman spectroscopy to enable
studies of molecular bindings in nanoparticle systems with low pump power.
(6.2) Future Work
In the future, the use of HC-PCF can be extended to the in situ studies of colloidal nanoparticles
using Raman spectroscopy. Physical changes of the QD structure, such as crystallinity,
interfacial modes and surface passivation, can be monitored and controlled dynamically during
the various synthesis processes for property optimizations. For example, surface chemistry
between QDs and electrode surface can be optimized to achieve efficient sensitized
photocurrents for practical photovoltaic devices. Furthermore, structural dynamics of the QDs
can be determined experimentally in different biological systems to reveal the possible cause of
some undesirable effects (i.e. carbonization, cytotoxicity and photobleaching). Systematic study
of molecular structures will enable us to better understand the basis of different QD properties
which further optimizes our QD designs for different applications. Ultimately, HC-PCF can be
served as a platform for studying the vibrational modes of complex aqueous solutions of relevant
105
biological samples such as DNA, proteins and potentially peripheral blood too. In fact, targeted
detection of lung cancer cells using the pegylated and phospholipid nanoparticles are currently
underway. A suitable PCF model has been chosen and its splicing program to achieve selective
filling has been optimized for the large cell sizes (Appendix F).
The PCF platform can be further developed to achieve greater sensitivities, improved
functionalities and practical usages. For example, the platform can be developed to collect both
the forward and backward scattering to improve the system sensitivity. An enclosed cell can also
be incorporated to achieve a sensitive gas sensor. In many practical applications, selective filling
from the side of the PCF might be required because both ends of the PCF are required for signal
collection and fluidic control. However, mechanical modification of the PCF is not trivial,
therefore further investigation is required.
106
Appendix A
Snell’s Law in Wave Optics
Using Maxwell’s equation, one can describe a plane wave propagating in space using:
R = Rj cos7 − k ∙ m (A.1)
where E is the electric field with magnitude |Eo|, ω is the frequency of light, k is the wave vector,
and r is the position vector. If we consider a transverse electric (TE) field incident at the interface
between two materials as shown in Figure A.1, we can write the following three equations to
describe the electric field that is parallel to the material interface:
Rn = ( coso7( − H(p + H([q rst (A.2)
Rm = d coso7d − Hdp − Hd[qrst (A.3)
Ru = , coso7, − H,p + H,[qrst (A.4)
where the subscripts i, r, t refer to the incident, reflected and transmitted wave respectively. In
order to satisfy boundary conditions, the component of the electric fields that is parallel to the
material interface must be equal and continuous at the interface. This implies that
107
Rnq = 0 + Rmq = 0 = Ruq = 0 (A.5)
or ( cos7( − H(p |[v + d cos7d − Hdp |[v
= , cos7, − H,p|[v (A.6)
This can only be satisfied when the phase of the two terms are matched; hence,
7( − H(p = 7d − Hdp = 7, − H,p (A.7)
which can only be satisfied for all t when
7( = 7d = 7, = 7 (A.8)
and for all x when
H( = Hd = H, (A.9)
Equation (A.8) and (A.9) form the generalized rules for Snell’s law which requires the
conservation of the frequency of light and the component of the wave vector that is parallel to
the interface of the two materials.
Figure A.1: Directions of the electric field, magnetic field and wave vector when a TE field is
incident at the interface between two dielectric materials.
108
Appendix B
Synthesis Method and Materials for
Thiol-Capped CdTe Quantum Dots
The thiol-capped CdTe QDs used in this study were synthesized with minor modifications to
literature procedures.48, 49
Sodium borohydride (NaBH4, 99%), tellurium powder (~200 mesh,
99.8%), cadmium chloride (CdCl2, 99%), TGA (98%), MPA (99%), and TG (99%) purchased
from Aldrich chemicals. All chemicals were used as received. Millipore Q water (18 sm) were
used throughout the nanocrystal synthesis.
Synthesis of TGA-capped CdTe QDs: 300 mg (7.90 mmol) of NaBH4 was dissolved in 10 ml of
water under argon environment and cooled in an ice bath. 400 mg (3.14 mmol) of tellurium
powder was added and the reaction mixture was left stirring for 2 hours to get a deep pinkish-
purple clear solution. The resulting NaHTe solution was kept under argon before use. 1.15 g
(6.28 mmol) of CdCl2 was then dissolved in 70 ml of Millipore water and bubbled with argon for
109
20 mins. 1 mL (15.08 mmol) of TGA was slowly added to this solution which formed white
precipitates due to the formation of Cd-TGA complex in the solution. The pH value of the
solution was adjusted to 11.0 by dropwise addition of 2.0 M NaOH solution with stirring. The
freshly prepared NaHTe solution is then rapidly added to the Cd precursor solution at room
temperature. After continuous stirring at room temperature for 10 mins, the resulting orange
solution is refluxed at predetermined time (2 – 16 hrs) and then cooled to room temperature. The
obtained CdTe NCs were precipitated by the addition of reagent grade acetone and were isolated
and purified by repeated precipitation/centrifugation cycles with acetone/water and dried in a
vacuum overnight. Solid CdTe QD samples were then weighted and dispersed in deionised water
to prepare aqueous solutions of 2 mg/mL for Raman measurements, 0.04 mg/mL for UV-vis and
0.4mg/mL for PL measurements respectively.
MPA-capped CdTe QDs were obtained using 400 mg of tellurium powder, 297 mg of NaBH4,
1.15 g of CdCl2 and 1.15 mL of MPA as described above.
TG-capped CdTe QDs were obtained using 565 mg of tellurium powder, 420 mg of NaBH4,
1.0 g of CdCl2 and 1.62 mL of TG as described above.
110
Appendix C
"anoparticle Signal Enhancement
using HC-PCF
(C.1) Citrate Stabilized Gold "anoparticles for Pegylated System
Figure C.1 compares the Raman spectra of citrate stabilized gold nanoparticles obtained using
conventional Raman spectroscopy with a cover slide and HC-PCF with a CW HeNe (633 nm)
laser. Carboxylate stretching modes between 1370 – 1600 cm-1
are not observed using cover
slide with 0.17 mW and 2 s of exposure time (averaging over 2 spectra). In contrast, carboxylate
stretching modes are observed in the spectrum obtained using HC-PCF with the same power and
exposure time.
111
Figure C.1: Raman spectrum of citrate stabilized gold nanoparticles obtained using cover slide
and HC-PCF with 633 nm laser.
(C.2) Pegylated Gold "anoparticles
Figure C.2 compares the Raman spectra of pegylated gold nanoparticles (without Raman dye
attached) obtained using a cover slide and HC-PCF with a 633 nm laser. Gold-sulphur bond
between thiol-modified PEG and gold nanoparticles core (Au–S stretch at 305 cm-1
) as well as
carboxylate stretching modes between 1380 – 1600 cm-1
are not observed using cover slide with
0.17 mW and 2 s of laser exposure (averaging over 2 spectra). However, both modes are
observed in the spectrum obtained using HC-PCF with the pump power and exposure time as
shown in Figure C.2.
112
Figure C.2: Raman spectrum of pegylated gold nanoparticles obtained using cover slide and
HC-PCF with 633 nm laser. Inset: graph focused on spectrum obtained from direct measurement.
(C.3) Pegylated Gold "anoparticles with MGITC Adsorbed
Figure C.3 compares the Raman spectra of MGITC adsorbed gold nanoparticles obtained using a
cover slide and HC-PCF with 2 s laser exposure using 633 nm laser (averaging over 2 spectra).
MGITC modes are observed using both the cover slide and HC-PCF. Nevertheless, the detected
Raman signal of the MGITC modes are enhanced in the spectrum obtained uisng HC-PCF. Table
C.1 shows the enhancement factor using HC-PCF relative to conventional technique. HC-PCF
enhanced the MGITC modes by about 15-19 times.
113
Figure C.3: Raman spectra of MGITC adsorbed gold nanoparticles obtained using cover slide
and HC-PCF with 633 nm laser. Inset: spectra focused on region between 900 and 1600 cm-1
.
Table C.1: Enhancement factor of MGITC modes using HC-PCF.
Mode
(cm-1
)
Intensity
(Cover
Slide)
Intensity
(PCF) Enhan.
Power
(Cover
Slide)
Power
(PCF) Enhan.
Total
Enhancement
927 1771 5664 3.2 1 mW 0.17 mW 5.9 ~ 19 times
1616 8130 20474 2.5 1 mW 0.17 mW 5.9 ~ 15 times
(C.4) Citrate Stabilized Gold "anoparticles for Phospholipid System
Figure C.4 compares the Raman spectrum of citrate stabilized gold nanoparticles obtained using
a cover slide and HC-PCF with 2 s laser exposure (averaging over 2 spectra). Citrate modes
between 1300 and 1610 cm-1
are only observed with enhancements from the HC-PCF.
114
Figure C.4: Raman spectra of citrate gold nanoparticles obtained using cover slide and HC-PCF
with 633 nm laser.
(C.5) Gold "anoparticles with CV Adsorbed
Figure C.5 compares the Raman spectrum of CV adsorbed gold nanoparticles obtained using a
cuvette using 785 nm laser, cover slide and HC-PCF using 633 nm laser with 2 s laser exposure
(averaging over 2 spectra). Only CV modes are observed in all spectra. However, mode
intensities were enhanced with the use of HC-PCF. Using Raman mode at 1623 cm-1
as a
reference, mode intensity was enhanced about 64 times with HC-PCF relative to the use of a
cover slide.
115
Figure C.5: Raman spectrum of CV adsorbed gold nanoparticles obtained using cuvette with
785 nm laser, cover slide and HC-PCF with 633 nm laser. Inset: spectra focused on region
between 2000 and 2500 cm-1
.
(C.6) Phospholipid Gold "anoparticles with CV Adsorbed
Figure C.6 compares the Raman spectrum of CV adsorbed gold nanoparticles obtained using a
cuvette using 785 nm laser, cover slide and HC-PCF using 633 nm laser with 2 s laser exposure
(averaging over 2 spectra). Similar to CV adsorbed nanoparticles without any coating, only CV
modes are observed on all spectra. However, unlike those without the phospholipid coating,
some CV modes were enhanced with HC-PCF while others were weaker than those obtained
with 785 nm laser. This is probably because of the stronger light guidance of the PCF at the
higher Raman shift regime as it gets closer to the bandgap of the PCF. Nevertheless, all CV
modes were enhanced with HC-PCF relative to those obtained with a cover slide. Using Raman
mode at 1623 cm-1
as a reference, mode intensity was enhanced about ~ 20 and ~ 32 times with
HC-PCF relative to the use of a cover slide in phospholipid nanoparticles using MHPC and
SHPC respectively.
116
Figure C.6: Raman spectrum of phospholipid gold nanoparticles obtained using cuvette with
785 nm, cover slide and HC-PCF with 633 nm laser. Phospholipid was formed using (a) MHPC
and DMPC and (b) SHPC and DMPC. Inset: spectra focused on region between 2000 and 2500
cm-1
.
117
Appendix D
Raman Mode Assignments of Thiol-
Capped CdTe "anoparticles
Abbreviations: δ – in-plane-deformation γ – out-of-plane-deformation ρ – rocking vibration
ν – stretching vibration τ – twisting vibration def. – deformation [subscripts:
s – symmetric as – asymmetric]
Table D.1: Proposed assignment of thiol-capped CdTe nanoparticles. Spectra are shown in
Figure 4.6 for regime between 100 and 200 cm-1
, Figure 4.8 for regime between 200 and 310
cm-1
and Figure 4.9 for regime between 310 and 1750 cm-1
.
Raman shift (cm-1
)
proposed assignment TGA MPA TG
121 124 122 Te A1
140 141 142 Te E / CdTe TO
160 160 159 CdTe LO
263 261 261 CdS-Like LO
(Continued ext Page)
118
Raman shift (cm-1
)
proposed assignment TGA MPA TG
291 293 291 CdS SO
332 335 337 C-S def.
584 586 581 CdS 2SO
447 432 443 δ(C-C-O)
535 544 532 τ(O-H)
629 γ(C-C=O)
680 δ(O=C-O)
683 660 ν(C-S) gauche conformer
718 ρ(CH2)
755 ν(C-S)
752 747 ν(C-S) trans conformer
893 892 895 ν(C-OH)
994 993 ν(C-C)
1055 1055 1059 ν(C-C)
1136 τ(CH2)
1215 1215 1215 τ(CH2)
1363 1356 νs(COO)
1496 1469 νs(COO)
1525 ν(COS-)
1571 νas(COO)
1685 1685 ν(C=O)
119
Appendix E
Raman Mode Assignments of Gold
"anoparticle Systems
Abbreviations: δ – in-plane-deformation γ – out-of-plane-deformation ρ – rocking vibration
ν – stretching vibration I.p. – in-plane O.o.p. – out-of-plane ϕ – phenyl ring
[subscripts: s – symmetric as – asymmetric]
(E.1) Citrate Stabilized Gold "anoparticles for Pegylated System
Table E.1: Proposed assignment of citrate stabilized Gold nanoparticles for pegylated
nanoparticle system. Spectrum is shown in Figure 5.5. Assignments are based on Munro et al.,119
Kerker et al.148
and De Melo et al.149
Raman shift
(cm-1
)
Proposed
Assignment
266 Au Surface Mode
(Au-Cl- or AuO)
422 γ(COO−)
508 δ(CCC)
Raman shift
(cm-1
)
Proposed
Assignment
725 ρ(CH2)
889 νs(CC)
(Continued
ext Page)
120
Raman shift
(cm-1
)
Proposed
Assignment
1020 νs(CC)
1135 νas(CCO)
1291 ν(CO) + δ(OH)
1381 νs(COO−)
1445 νs(COO−)
1544 νs(COO−)
1597 νas(COO−)
Raman shift
(cm-1
)
Proposed
Assignment
2721 Overtones and
Combinations
2824
2855 νs(CH)
2921 νas(CH)
3240 νs(OH)
3401 νas(OH)
(E.2) Pegylated Gold "anoparticles
Table E.2: Proposed assignment of pegylated gold nanoparticles. Spectrum is shown in Figure
5.9. Assignments are based on Munro et al.,119
Kerker et al.,148
De Melo et al.,149
and Al-Sa'ady
et al.128
Raman shift
(cm-1
) Proposed
Assignment
261 Au Surface Mode
(Au-Cl- or AuO)
304 ν(AuS)
443 γ(COO−) + δ(CCC)
887 νs(CC)
940 ν(C–COO) + ρ(CH2)
1071 νs(CC)
1159 νas(CCO)
1246
1292 ν(CO) + δ(OH)
1375 νs(COO-)
1465 νs(COO-)
Raman shift
(cm-1
)
Proposed
Assignment
1539 νs(COO-)
1595 νas(COO-)
2136
2165
2718 Overtones and
Combinations
2827
2859 νs(CH)
2919 νas(CH)
3257 νs(OH)
3407 νas(OH)
121
(E.3) Pegylated Gold "anoparticles with MGITC Adsorbed
Table E.3: Proposed assignment of MGITC adsorbed pegylated gold nanoparticles. Spectrum is
shown in Figure 5.7. Assignments are based on Lueck et al.150
Raman shift
(cm-1
)
Proposed
Assignment
231 C–ϕ3 breathing
354 ϕ–C–ϕ bend
434
450
472
535 I.p. benzene ring
deformation
575
592
632
675
733 I.p. benzene ring
bend, stretch NC
bend, torsion
743
771
811 O.o.p. C–H
845
928
950 N(CH3)2 stretch,
bend
Raman shift
(cm-1
)
Proposed
Assignment
987 I.p. benzene
1001 I.p. benzene
1182 I.p. C–H bend
1197
1228 N–C stretch, NC
bend
1306 I.p. C–C, C–C–H
1347 Combination
1375 N–ϕ stretch
1398 I.p. C–C, C–H
1454 NC bend and rock
1489 NC bend and rock
1510
1530 ϕ=N+ stretch
1593 I.p. ring stretch,
bend
1625 N–ϕ, C–C stretch
3254 νs(OH)
3410 νas(OH)
122
(E.4) Citrate Stabilized Gold "anoparticles for Phospholipid System
Table E.4: Proposed Assignment of citrate stabilized Gold nanoparticles for phospholipid
nanoparticle system. Spectrum is shown in Figure 5.11. Assignments are based on Munro et
al.,119
Kerker et al.148
and De Melo et al.149
Raman shift
(cm-1
)
Proposed
Assignment
1006
1337 νs(COO-)
1446 νs(COO-)
1573 νs(COO-)
1608 νas(COO-)
2130
Raman shift
(cm-1
)
Proposed
Assignment
2722 Overtones and
Combinations
2824
2856 νs(CH)
2926 νas(CH)
3260 νs(OH)
3408 νas(OH)
(E.5) Gold "anoparticles with CV Adsorbed
Table E.5: Proposed Assignment of CV adsorbed gold nanoparticles. Spectrum is shown in
Figure 5.12b. Assignments are based on Lueck et al.150
Raman shift
(cm-1
) Proposed
Assignment
208 C–ϕ3 breathing
217
260
349 ϕ–C–ϕ bend
433
451
Raman shift
(cm-1
)
Proposed
Assignment
475
537 I.p. benzene ring
deformation
570
616
676
(Continued
ext Page)
123
Raman shift
(cm-1
)
Proposed
Assignment
734 I.p. benzene ring
bend, stretch NC
bend, torsion
754
770
780
811 O.o.p. C–H
926 I.p. benzene
953 N(CH3)2 stretch,
bend
984 I.p. benzene
1002 I.p. benzene
1131
1181 I.p. C–H bend
1232 N–C stretch, NC
bend
Raman shift
(cm-1
)
Proposed
Assignment
1308 I.p. C–C, C–C–H
1344 Combination
1378 N–ϕ stretch
1397 I.p. C–C, C–H
1454 NC bend and rock
1488 NC bend and rock
1537 ϕ=N+ stretch
1592 I.p. ring stretch,
bend
1625 N–ϕ, C–C stretch
3248 νs(OH)
3418 νas(OH)
(E.6) Phospholipid Gold "anoparticles with CV Adsorbed
Same as CV adsorbed gold nanoparticles (Table E.5).
124
Appendix F
Fiber Splicing Program Optimization
for Large Biological Molecules
PCFs utilized in the nanoparticle studies are not suitable for studying large biological molecules
with sizes that are in the range of micrometers or tenths of micrometers, such as human cells.
This is because the central hollow core of the PCF is too small by design (only about ~10 µm in
diameter). Moreover, after collapsing the cladding holes of the PCF for selective filling, the
diameter of the central core at the fiber end is reduced to only about 7 µm (Figure F.1). The large
size of many biological molecules would not be able to fill into the central core of the PCF for
measurements.
125
Figure F.1: SEM image of the HC-1060 before (a) and after (b) fiber splicing to collapse the
cladding holes. The central core of the PCF is reduced to about 7 µm after the splicing process.
To enable efficient Raman scattering of large biological molecules with PCF, we obtained PCF
modeled HC19-1550 in which its central hollow core is ~20 µm in diameter. Its core diameter is
twice of the HC-800 and HC1060 utilized in the nanoparticle experiments. HC19-1550 PCF
features the largest central hollow core for PCF commercially available at the time of
experiment. However, the trade off for the large core fiber is the lower enhancement of the
Raman modes with a 633 nm pump laser compared to HC-800 and HC1060 due to their different
cladding designs for different photonic bandgaps (Figure F.2).
Figure F.2: Comparison of water modes with HC-800, HC-1060, HC19-1550 and cover slide.
126
To selectively fill biological molecules into the central hollow core of the HC19-1550, the fiber
splicing program of the HC-800 was modified to collapse all cladding holes of the HC19-1550
PCF while maximizing its central core size after splicing. As human cells are typically 10 – 20
µm in diameter, the central core of the PCF would be best to remain as close to 20 µm as
possible (i.e. leaving the size of the central core almost unchanged). Initial trial and error in
tuning the splice power index led to a core size of about 17 – 18 µm in diameter (Figure F.3). All
of the cladding holes were closed except the three corner holes.
Figure F.3: SEM image of PCF facet after splicing with program modified from that for HC-
800. The central core of the PCF after splicing is about 17 – 18 µm and all of the cladding holes
are closed except the three corner holes.
Further improvements of the splicing program were achieved through varying the distance of the
fiber facet from the electrode. We chose to tune this parameter because we had already
determined a set of parameters that could collapse almost all the cladding holes. Varying the
distance between the fiber and the electrode would allow fine tuning of the hole collapsation.
127
Figure F.4 shows SEM images of the PCF spliced with electrode–fiber distance varying from 30
to 90 µm with precision of ± 5 µm. The core diameter was maximized (core size of ~19 – 20 µm)
with all the cladding holes collapsing at about 50 – 70 µm from the electrode. At a shorter
distance (i.e. 30 µm), all cladding holes were closed tightly but the central core was only about
14 – 15 µm in diameter. At a further distance (i.e. 80 µm), the central hole was enlarged to about
30 – 33 µm. The large core size will cause tapering along the fiber length which affects coupling
of the pump laser to the PCF. Further increasing the distance caused enlargement of the corner
holes around the central core which is also undesirable for light coupling.
To test the variability of the central core size using the determined program, the same parameters
were used to splice three different stings of the HC19-1550 PCF fibers (Figure F.5). The core
diameter fluctuates between 17 to 22 µm. This fluctuation could be caused by the focusing
differences of the PCF as focusing was done through observing the sharpest fiber edge along the
fiber axis with the naked eyes. As a result, large focusing differences could be made between
different fibers. However, this could be improved with practices or with a fiber splicer that has
an automatic focusing feature.
128
Core Diameter: 14 –15 µm Core Diameter: 20 – 21 µm
…
Core Diameter: 19 – 20 µm Core Diameter: 30 – 33 µm
Core Diameter: 21 – 22 µm
Figure F.4: SEM image of PCF fused with fiber–electrode distance at (a) 30, (b) 50, (c) 70, (d)
80, and (e) 90 µm.
129
Finally, a HC19-1550 PCF fused with the determined parameter was submerged into a solution
of lung cancer cells, with sizes ranging approximately between 10 to 15 µm in diameter. Figure
F.6 shows the optical image of the PCF facet from the opposite end. The brightness at the central
core shows the cancer cell solution. The solution was only filled in the central core and not the
claddings; therefore, all the cladding holes were closed successfully. The modified fiber splicing
program for the large core PCF enables large biological molecules, such as cancer cells, to be
studied using Raman spectroscopy with enhanced sensitivity provided by the PCF.
Core Diameter: 21 –22 µm Core Diameter: 17 – 18 µm
Core Diameter: 18 – 19 µm
Figure F.5: SEM image of three PCFs fused HC15-1990 PCFs at 50 µm from electrode.
130
Figure F.6: Optical image of HC19-1550 PCF selectively filled with lung cancer cells. The
bright light at the central core shows the cancer cell solution.
131
References
1. Bera, D.; Qian, L.; Tseng, T. K.; Holloway, P. H., Quantum Dots and Their Multimodal
Applications: A Review. Materials 2010, 3 (4), 2260-2345.
2. Tanikkul, S.; Jakmunee, J.; Rayanakorn, M.; Grudpan, K.; Marquardt, B. J.; Gross, G.
M.; Prazen, B. J.; Burgess, L. W.; Christian, G. D.; Synovec, R. E., Characterization and
use of a Raman liquid-core waveguide sensor using preconcentration principles. Talanta
2003, 59, 809-816.
3. Qi, D.; Berger, A. J., Quantitative concentration measurements of creatinine dissolved in
water and urine using Raman spectroscopy and a liquid core optical fiber. 2005, 10,
031115.
4. Jin, S.; Feng, C.; Jin, Y.; Zhang, Z., Research on Raman spectrum in liquid core optical
fiber of CCl4 solution. Chin. Opt. Lett. 2005, 3, 666-668.
5. Xie, Z.; Lu, Y.; Wei, H.; Yan, J.; Wang, P.; Ming, H., Broad spectral photonic crystal
fiber surface enhanced Raman scattering probe. Appl. Phys. B-Lasers and Optics 2009,
95, 751-755.
132
6. Pristinski, D.; Du, H., Solid-core photonic crystal fiber as a Raman spectroscopy platform
with a silica core as an internal reference. Opt. Lett. 2006, 31, 3246-3248.
7. Kyaw, K. O. M.; Han, Y.; Kanka, J.; Sukhishvili, S.; Du, H., Structure fits the purpose:
photonic crystal fibers for evanescent-field surface-enhanced Raman spectroscopy. Opt.
Lett. 2010, 35, 466-468.
8. Yan, H.; Gu, C.; Yang, C.; Liu, J.; Jin, G.; Zhang, J.; Hou, L.; Yao, Y., Hollow core
photonic crystal fiber surface-enhanced Raman probe. 2006, 89, 204101.
9. Rutledge, S. A.; Farah, A. A.; Dinglasan, J.; Anderson, D. J.; Das, A.; Goh, J.; Goh, C.;
Helmy, A. S., Postsynthesis Crystallinity Improvement of Colloidal CdTe Nanoparticles
using Rapid Thermal Annealing. J. Phys. Chem. C 2009, 113, 20208-20213.
10. Irizar, J.; Dinglasan, J.; Goh, J. B.; Khetani, A.; Anis, H.; Anderson, D.; Goh, C.; Helmy,
A. S., Raman Spectroscopy of Nanoparticles using Hollow-Core Photonic Crystal Fibers.
IEEE J. Sel. Top. Quant. 2008, 14, 1214-1222.
11. Amer, M. S., Chapter 2: Raman Spectroscopy; the Diagnostic Tool. In Raman
spectroscopy, Fullerenes and 4anotechnology, UK: Royal Society of Chemistry:
Cambridge, 2010; pp 43-108.
12. Ferraro, J. R.; Nakamoto, K.; Brown, C. W., Introductory Raman Spectroscopy. 2nd ed.;
Academic Press: San Diego, 2003.
13. Stevenson, C. L.; Vo-Dinh, T., Signal Expressions in Raman Spectroscopy. In Modern
Techniques in Raman Spectroscopy, John Wiley: New York, 1996.
133
14. Rutledge, S. Post Synthesis Rapid Thermal Annealing and Characterization of Colloidal
Nanoparticles. University of Toronto, Toronto, 2009.
15. Aroca, R., Vibrational Intensities. In Surface-Enhanced Vibrational Spectroscopy, Wiley:
New Jersey, 2006.
16. Joannopoulos, J. D., Photonic Crystals: Molding the Flow of Light. 2nd ed.; Princeton
University Press: New Jersey, 2008.
17. Yablonovitch, E., Photonic Crystals: Semiconductors of Light. Sci. Am. 2001, 285 (6),
47-51.
18. Knight, J. C.; Birks, T. A.; Russell, P. S. J.; Atkin, D. M., All-Silica Single-Mode Optical
Fiber with Photonic Crystal Cladding. Opt. Lett. 1996, 21 (19), 1547-1549.
19. Bjarklev, A.; Bjarklev, A. S.; Broeng, J., Photonic Crystal Fibres. Kluwer Academic
Publishers: Boston, 2003.
20. Joannopoulos, J. D., Chapter 3: Photonic Crystals. 2nd ed.; Princeton University Press:
New Jersey, 2008.
21. Fini, J. M., Microstructure Fibres for Optical Sensing in Gases and Liquids. Meas. Sci.
Technol. 2004, 15, 1120.
22. Bjarklev, A.; Bjarklev, A. S.; Broeng, J., Chapter 2.2.2 Two- and Three- Dimensional
Photonic Crystals. In Photonic Crystal Fibers, Kluwer Academic Publishers: Boston,
2003.
134
23. Joannopoulos, J. D., Chapter 9: Photonic-Crystal Fibers. In Photonic Crystals: Molding
the Flow of Light, 2nd ed.; Princeton University Press: New Jersey, 2008.
24. Castro, J. L.; López-Ramírez, M. R.; Arenas, J. F.; Otero, J. C., Surface-Enhanced Raman
Scattering of 3-Mercaptopropionic Acid Adsorbed on a Colloidal Silver Surface. J.
Raman Spectrosc. 2004, 35, 997-1000.
25. Wang, Y.; Zhang, J.; Jia, H.; Li, M.; Zeng, J.; Yang, B.; Zhao, B.; Xu, W.; Lombardi, J.
R., Mercaptopyridine Surface-Functionalized CdTe Quantum Dots with Enhanced
Raman Scattering Properties. J. Phys. Chem. C 2008, 112, 996-1000.
26. Altkorn, R.; Malinsky, M. D.; Van Duyne, R. P.; Koev, I., Intensity Consideration in
Liquid Core Optical Fiber. Appl. Spectrosc. 2001, 55, 373-381.
27. Eftekhari, F.; Irizar, J.; Hulbert, L.; Helmy, A. S., A Comparative Study of Raman
Enhancement in Capillaries. J. Appl. Phys. 2011, 109, 113104-1.
28. Qian, X.-M.; Nie, S. M., Single-molecule and Single-nanoparticlesers: from Fundamental
Mechanisms to Biomedical Applications. Chem. Soc. Rev. 2008, 37, 912-920.
29. Rao, S.; Raj, S.; Balint, S.; Fons, C. B.; Campoy, S.; Llagostera, M.; Petrov, D., Single
D4A molecule detection in an optical trap using surface-enhanced Raman scattering.
Appl. Phys. Lett.: 2010; Vol. 96, p 213701.
30. Medintz, I. L.; Uyeda, H. T.; Goldman, E. R.; Mattoussi, H., Quantum Dot Bioconjugates
for Imaging, Labelling and Sensing. 4at. Mater. 2005, 4, 435-446.
135
31. Wood, V.; Bulović, V., Colloidal Quantum Dot Light-Emitting Devices. 4ano Rev. 2010,
1, 5202–5209.
32. Sargent, E. H., Infrared Photovoltaics Made by Solution Processing. 4at. Photon. 2009,
3, 325-331.
33. Rogach, A. L.; Franzl, T.; Klar, T. A.; Feldmann, J.; Gaponik, N.; Lesnyak, V.; Shavel,
A.; Eychmüller, A.; Rakovich, Y. P.; Donegan, J. F., Aqueous Synthesis of Thiol-Capped
CdTe Nanocrystals: State-of-the-Art. J. Phys. Chem. C 2007, 111, 14628-14637.
34. Zhang, H.; Sun, P.; Liu, C.; Gao, H.; Xu, L.; Fang, J.; Wang, M.; Liu, J.; Xu, S., L-
Cysteine Capped CdTe-CdS Core-Shell Quantum Dots: Preparation, Characterization and
Immuno-Labeling of HeLa Cells. Luminescence 2009, n/a.
35. Liu, J.; Shi, Z.; Yu, Y.; Yang, R.; Zuo, S., Water-Soluble Multicolored Fluorescent CdTe
Quantum Dots: Synthesis and Application for Fingerprint Developing. J. Colloid. Interf.
Sci. 2010, 342, 278-82. Epub 2009 Oct 29.
36. Wang, C.; Zhang, H.; Zhang, J.; Li, M.; Sun, H.; Yang, B., Application of Ultrasonic
Irradiation in Aqueous Synthesis of Highly Fluorescent CdTe/CdS Core−Shell
Nanocrystals. J. Phys. Chem. C 2007, 111, 2465-2469.
37. Abd El-sadek, M. S.; Ram Kumar, J.; Moorthy Babu, S., The Role of Potassium Tellurite
as Tellurium Source in Mercaptoacetic Acid-Capped CdTe Nanoparticles. Curr. Appl.
Phys. 2010, 10, 317-322.
136
38. Sargent, E. H., Solar Cells, Photodetectors, and Optical Sources from Infrared Colloidal
Quantum Dots. Adv. Mater. 2008, 20, 3958-3964.
39. Cho, K.-S.; Lee, E. K.; Joo, W.-J.; Jang, E.; Kim, T.-H.; Lee, S. J.; Kwon, S.-J.; Han, J.
Y.; Kim, B.-K.; Choi, B. L.; al., e., High-Performance Crosslinked Colloidal Quantum-
Dot Light-Emitting Diodes. 4at. Photon. 2009, 3, 341-345.
40. Qian, X.; Peng, X.-H.; Ansari, D. O.; Yin-Goen, Q.; Chen, G. Z.; Shin, D. M.; Yang, L.;
Young, A. N.; Wang, M. D.; Nie, S., In Vivo Tumor Targeting and Spectroscopic
Detection with Surface-Enhanced Raman Nanoparticle Tags. 4at. Biotechnol. 2008, 26,
83-90.
41. Kim, S.; Lim, Y. T.; Soltesz, E. G.; De Grand, A. M.; Lee, J.; Nakayama, A.; Parker, J.
A.; Mihaljevic, T.; Laurence, R. G.; Dor, D. M.; al., e., Near-Infrared Fluorescent Type II
Quantum Dots for Sentinel Lymph Node Mapping. 4at. Biotechnol. 2004, 22, 93-97.
42. Yaghini, E.; Seifalian, A. M.; MacRobert, A. J., Quantum Dots and Their Potential
Biomedical Applications in Photosensitization for Photodynamic Therapy. 4anomed.
2009, 4, 353-363.
43. Schreder, B.; Schmidt, T.; Ptatschek, V.; Spanhel, L.; Materny, A.; Kiefer, W.,
Raman Characterization of CdTe/CdS-"Core-Shell"-Clusters in Colloids and Films. J.
Cryst. Growth 2000, 214-215, 782-786.
44. Byrne, S. J.; Corr, S. A.; Rakovich, T. Y.; Gun'ko, Y. K.; Rakovich, Y. P.; Donegan, J.
F.; Mitchell, S.; Volkov, Y., Optimisation of the Synthesis and Modification of CdTe
Quantum Dots for Enhanced Live Cell Imaging. J. Mater. Chem. 2006, 16, 2896-2902.
137
45. Zhang, J.-Y.; Wang, X.-Y.; Xiao, M.; Qu, L.; Peng, X., Lattice Contraction in Free-
Standing CdSe Nanocrystals. Appl. Phys. Lett. 2002, 81, 2076-2078.
46. Baranov, A. V.; Rakovich, Y. P.; Donegan, J. F.; Perova, T. S.; Moore, R. A.; Talapin, D.
V.; Rogach, A. L.; Masumoto, Y.; Nabiev, I., Effect of ZnS Shell Thickness on the
Phonon Spectra in CdSe Quantum Dots. Phys. Rev. B 2003, 68, 165306.
47. Castro, J. L.; López-Ramírez, M. R.; Centeno, S. P.; Otero, J. C., Adsorption of
Mercaptoacetic Acid on a Colloidal Silver Surface as Investigated by Raman
Spectroscopy. Biopolymers 2004, 74, 141-145.
48. Klayman, D. L.; Griffin, T. S., Reaction of Selenium with Sodium Borohydride in Protic
Solvents. A Facile Method for the Introduction of Selenium into Organic Molecules.
J. Am. Chem. Soc. 1973, 95, 197-199.
49. Zhenyu, G.; Zou, L.; Fang, Z.; Zhu, W.; Zhong, X., One-Pot Synthesis of Highly
Luminescent CdTe/CdS Core/Shell Nanocrystals in Aqueous Phase. 4anotechnology
2008, 19, 135604.
50. Ochoa, O. R.; Witkowski Iii, E. J.; Colajacomo, C.; Simmons, J. H.; Potter Jr, B. G.,
Raman Characterization of CdTe Nanocrystallites Embedded in a Glass Matrix. J. Mater.
Sci. Lett. 1997, 16, 613-616.
51. Vinogradov, V.; Karczewski, G.; Kucherenko, I.; Mel’nik, N.; Fernandez, P., Raman
Spectra of Structures with CdTe-, ZnTe-, and CdSe-Based Quantum Dots and Their
Relation to the Fabrication Technology. Phys. Solid State 2008, 50, 164-167.
138
52. Amirtharaj, P. M.; Pollak, F. H., Raman Scattering Study of the Properties and Removal
of Excess Te on CdTe Surfaces. Appl. Phys. Lett. 1984, 45, 789-791.
53. Shin, S. H.; Bajaj, J.; Moudy, L. A.; Cheung, D. T., Characterization of Te Precipitates in
CdTe Crystals. Appl. Phys. Lett. 1983, 43, 68-70.
54. Morell, G.; Reynés-Figueroa, A.; Katiyar, R. S.; Farías, M. H.; Espinoza-Beltran, F. J.;
Zelaya-Angel, O.; Sánchez-Sinencio, F., Raman Spectroscopy of Oxygenated
Amorphous CdTe Films. J. Raman Spectrosc. 1994, 25, 203-207.
55. Islam, S. S.; Rath, S.; Jain, K. P.; Abbi, S. C.; Julien, C.; Balkanski, M., Forbidden One-
LO-Phonon Resonant Raman Scattering and Multiphonon Scattering in Pure CdTe
Crystals. Phys. Rev. B 1992, 46, 4982.
56. Na, H.-K.; Shon, P.-K., A Raman Scattering Study of CdTe/p-(1 1 1)InSb
Heterointerfaces Grown by R.F. Sputtering in High Vacuum. Solid State Commun. 1993,
85, 609-612.
57. Gilbert, B.; Huang, F.; Lin, Z.; Goodell, C.; Zhang, H.; Banfield, J. F., Surface Chemistry
Controls Crystallinity of ZnS Nanoparticles. 4ano Lett. 2006, 6, 605-610.
58. Schreder, B.; Schmidt, T.; Ptatschek, V.; Winkler, U.; Materny, A.; Umbach, E.; Lerch,
M.; Muller, G.; Kiefer, W.; Spanhel, L., CdTe/CdS Clusters with “Core−Shell” Structure
in Colloids and Films: The Path of Formation and Thermal Breakup. J. Phys. Chem. B
2000, 104, 1677-1685.
139
59. Fischer, A.; Anthony, L.; Compaan, A. D., Raman Analysis of Short-Range Clustering in
Laser-Deposited CdSxTe1-x Films. Appl. Phys. Let. 1998, 72, 2559-2561.
60. Pal, R.; Dutta, J.; Chaudhuri, S.; Pal, A. K., CdSxTe1-x Films: Preparation and Properties.
J. Phys. D: Appl. Phys. 1993, 26, 704.
61. Kim, J. E., Synthesis and Surface Characterization by Raman Spectroscopy of Water-
Dispersible ZnS: Mn Nanocrystals Capped with Mercaptoacetic Acid. 4otes 2008, 29,
1247.
62. Zhang, H.; Zhou, Z.; Yang, B.; Gao, M., The Influence of Carboxyl Groups on the
Photoluminescence of Mercaptocarboxylic Acid-Stabilized CdTe Nanoparticles. J. Phys.
Chem. B 2002, 107, 8-13.
63. Dhere, R.; Wu, X.; Albin, D.; Perkins, C.; Moutinho, H.; Gessert, T. In Formation and
Characterization of CdSxTe1-x Alloys Prepared from Thin Film Couples of CdS and
CdTe, 2002; pp 484-487.
64. Murali, K.; Palanisamy, T., Electrical and Optical Properties of Pulse Plated CdSxTe1-x
Films. Ionics 2010, 16, 613-619.
65. Blum, A. S.; Moore, M. H.; Ratna, B. R., Quantum Dot Fluorescence as a Function of
Alkyl Chain Length in Aqueous Environments. Langmuir 2008, 24, 9194-9197.
66. Fadi, A.; Balan, L.; Lambert, J.; Schneider, R., The Influence of Capping Thioalkyl Acid
on the Growth and Photoluminescence Efficiency of CdTe and CdSe Quantum Dots.
4anotechnology 2008, 19, 475401.
140
67. Kloepfer, J. A.; Bradforth, S. E.; Nadeau, J. L., Photophysical Properties of Biologically
Compatible CdSe Quantum Dot Structures. J. Phys. Chem. B 2005, 109, 9996-10003.
68. Aldana, J.; Wang, Y. A.; Peng, X., Photochemical Instability of CdSe Nanocrystals
Coated by Hydrophilic Thiols. J. Am. Chem. Soc. 2001, 123, 8844-8850.
69. Nakamoto, K., Infrared and Raman spectra of inorganic and coordination compounds:
Part B. Wiley: Hoboken, 2009.
70. Popović, Z. V.; Stanišić, G.; Stojanović, D.; Kostić, R., Infrared and Raman Spectra of
CdO. Phys. Status Solidi B 1991, 165, K109-K112.
71. Socrates, G., Infrared characteristic group frequencies. Wiley: New York, 1980.
72. Shevchenko, L. L., Infrared Spectra of Salts and Complexes of Carboxylic Acids and
Some of Their Derivatives. Russ. Chem. Rev. 1963, 32, 201.
73. Yu, W. W.; Qu, L.; Guo, W.; Peng, X., Experimental Determination of the Extinction
Coefficient of CdTe, CdSe, and CdS Nanocrystals. Chem. Mater. 2003, 15, 2854-2860.
74. Gao, M.; Kirstein, S.; Mohwald, H.; Rogach, A. L.; Kornowski, A.; Eychmuller, A.;
Weller, H., Strongly Photoluminescent CdTe Nanocrystals by Proper Surface
Modification. J. Phys. Chem. B 1998, 102, 8360-8363.
75. Zhang, H.; Wang, D.; Möhwald, H., Ligand-Selective Aqueous Synthesis of One-
Dimensional CdTe Nanostructures. Angew. Chem. Int. Edit. 2006, 45, 748-751.
141
76. Kelly, K. L.; Coronado, E.; Zhao, L. L.; C., S. G., The Optical Properties of Metal
Nanoparticles: The Influence of Size, Shape, and Dielectric Environment. J. Phys. Chem.
B 2002, 107, 60208-3113.
77. Barnes, W. L.; Dereux, A.; Ebesen, T. W., Surface Plasmon Subwavelength Optics.
4ature 2003, 424, 824-830.
78. Kreibig, U.; Vollmer, M., Optical Properties of Metal Clusters. Springer: Berlin, 1995;
Vol. 25.
79. Hummel, R. E.; Wissmann, P., In Handbook of Optical Properties; Vol. II: Optics of
Small Particles,Interfaces and Surfaces. CRC Press: Boca Raton, 1997; Vol. 1.
80. Jain, P. K.; El-Sayed, I. H.; El-Sayed, M. A., Au Nanoparticles Target Cancer. 4ano
Today 2007, 2, 16.
81. Faulds, K.; Barbagallo, R. P.; Keer, J. T.; Smith, W. E.; Graham, D., SERRS as a More
Sensitive Technique for the Detection of Labelled Oligonucleotides Compared to
Fluroesecece. Analyst 2004, 129, 567-8.
82. Weissleder, R., A Clearer Vision for In Vivo Imaging. 4at. Biotech. 2001, 19, 316-317.
83. Qian, X.; Peng, X. H.; Ansari, D.; Yin-Goen, Q.; Chen, G. Z.; Shin, D. M.; Yang, L.;
Young, A. N.; May, D. W.; Nie, S., In Vivo Tumor Targeting and Spectroscopic
Detection with Surface Enhanced Raman Nanoparticle Tags. 4at. Biotech. 2008, 26, 83-
90.
142
84. P., N. C.; Torp-Pedersen, S.; Burcharth, F.; Horn, T.; Pedersen, S.; Chrisensen, N. E.;
Olldag, E. S.; Andersen, P. H.; Karstrup, S.; Lorentzen, T.; al., e., Interstitial
Hyperthermia of Colorectal Liver Metastases with a US-Guided Nd-YAG Laser with a
Diffuser Tip: a Pilot Clinical Study. Radiol. 1993, 187, 333-7.
85. Amin, Z.; Donald, J. J.; Masters, A.; Kant, R.; Steger, A. C.; Bown, S. G.; Lees, W. R.,
Hepatic Metastases: Interstitial Laser Photocoagulation with Real-Time US Monitoring
and Dynamic CT Evaluation of Treatment. Radiol. 1993, 187, 339-347.
86. El-Sayed, M. A., Some Interesting Properties of Metals Confined in Time and Nanometer
Space of Different Shapes. Acc. Chem. Res. 2001, 34, 257-264.
87. Link, S.; El-Sayed, M. A., Spectral Properties and Relaxation Dynamics of Surface
Plasmon Electronic Oscillations in Gold and Silver Nanodots and Nanorods. J. Phys.
Chem. B 1999, 103, 8410-8426.
88. Link, S.; El-Sayed, M. A., Shape and Size Dependence of Radiative, Nonradiative, and
Photothermal Properties of Gold Nanocrystals. Int. Rev. Phys. Chem. 2000, 19, 409.
89. Link, S.; El-Sayed, M. A., Optical Properties and Ultrafast Dynamics of Metallic
Nanocrystals. Annu. Rev. Phys. Chem. 2003, 54, 331-66.
90. Lakowicz, J. R.; Chowdury, M. H.; Ray, K.; Zhang, J.; Fu, Y.; Badugu, R.;
Sabanayagam, C. R.; Nowaczyk, K.; Szmacinski, H. In Plasmon-controlled
Fluorescence: a 4ew Detection Technology, Plasmonics in Biology and Medicine III,
San Jose, Proc. SPIE: San Jose, 2006; p 609909.
143
91. Connor, E. E.; Mwamuka, J.; Gole, A.; Murphy, C. J.; Wyatt, M. D., Gold Nanoparticles
Are Taken Up by Human Cells but Do Not Cause Acute Cytotoxicity. Small 2005, 1,
325-327.
92. Shukla, R.; Bansal, V.; Chaudhary, M.; Basu, A.; Bhonde, R. R.; Sastry, M.,
Biocompatibility of Gold Nanoparticles and Their Endocytotic Fate Inside the Cellular
Compartment: A Microscopic Overview. Langmuir 2005, 21, 10644-10654.
93. Turkevich, J.; Stevenson, P. C.; Jillier, J., A Study of the Nucleation and Growth
Processes in the Synthesis of Colloidal Gold. Discuss. Faraday Soc. 1951, 11, 55-57.
94. Brown, K. R.; Walter, D. G.; Natan, M. J., Seeding of Colloidal Au Nanoparticle
Solutions. 2. Improved Control of Particle Size and Shape. Chem. Mater. 2000, 12, 306-
313.
95. Brown, K. R.; Natan, M. J., Hydroxylamine Seeding of Colloidal Au Nanoparticles in
Solution and on Surfaces. Langmuir 1998, 14, 726-728.
96. Brust, M.; Walker, M.; Bethell, D.; Schiffrin, D. J.; Whyman, R., Synthesis of Thiol-
Derivatized Gold Nanoparticles in a Two-phase Liquid-liquid System. J. Chem. Soc.
Chem. Commun. 1994, 7, 801-802.
97. Krug, J. T.; Wang, G. D.; Emory, S. R.; Nie, S., Efficient Raman Enhancement and
Intermittent Light Emission Observed in Single Gold Nanocrystals. J. Am. Chem. Soc.
1999 121, 9208-9214.
144
98. Smithpeter, C.; Dunn, A.; Drezek, R.; Collier, T.; Richards-Kortum, R., Near Real Time
Confocal Microscopy of Cultured Amelanotic Cells: Sources of Signal, Contrast Agents
and Limits of Contrast. J. Biomed. Opt. 1998, 3, 429.
99. Alivisatos, A. P.; Gu, W.; Larabell, C., Quantum Dots as Cellular Probes. Ammu. Rev.
Biomed. Eng. 2005, 7, 55-76.
100. Xing, Y.; Chaudry, Q.; Shen, C.; Kong, K. Y.; Zhau, H. E.; Chung, L. W.; Petros, J. A.;
O'Regan, R. M.; Yezhelyev, M. V.; Jonathan, W. S.; al., e., Bioconjugated Quantum Dots
for Multiplexed and Quantitative Immunohistochemistry. 4at. Protoc. 2007, 2, 1152-
1165.
101. Fountaine, T. J.; Wincovitch, S. M.; Geho, D. H.; Garfield, S. H.; Pittaluga, S.,
Multispectral Imaging of Clinically Relevant Cellular Targets in Tonsil and Lymphoid
Tissue Using Semiconductor Quantum Dots. Mod. Pathol. 2006, 19, 1181-1191.
102. Thomas, K. G.; Kamat, P. V., Chromophore-Functionalized Gold Nanoparticles. Acc.
Chem. Res. 2003, 80, 5951-5957.
103. Song, L.; Hennink, E. J.; Young, I. T.; Tanke, H. J., Photobleaching Kinetics of
Fluorescein in Quantitative Fluorescence Microscopy. Biophys. J. 1995, 68, 2588-2600.
104. Landsman, M. L.; Kwant, G.; Mook, G. A.; Zijlstra, W. G., Light-absorbing Properties,
Stability, and Spectral Stabilization of Indocyanine Green. J. Appl. Physiol. 1976, 40,
575-583.
145
105. Aubin, J. E., Autofluorescence of Viable Ultured Mammalian Cells. J. Histochem.
Cytochem. 1979, 27, 36-43.
106. Liu, Z.; Tabakman, S.; Sherlock, S.; Li, X.; Chen, Z.; Jiang, K.; Fan, S.; Dai, H.,
Multiplexed Five-color Molecular Imaging of Cancer Cells and Tumor Tissues with
Carbon Nanotube Raman Tags in the Near-Infrared. 4ano Res. 2010, 3, 222-223.
107. Keren, S.; Zavaleta, C.; Cheng, Z.; Zerda, A. D. L.; Gheysens, O.; Gambhir, S. S.,
Noninvasive Molecular Imaging of Small Living Subjects Using Raman Spectroscopy.
Proc. 4atl. Acad. Sci. 2008, 5, 5844-5849.
108. Nie, S.; Emory, S. R., Probing Single Molecules and Single Nanoparticles by Surface-
Enhanced Raman Scattering. Science 1997, 275, 1102-1106.
109. Kneipp, K.; Kneipp, H., Single Molecule Raman Scattering. Appl. Spectrosc. 2006, 60,
322A-334A.
110. Chen, A. M.; Scott, M. D., Current and Future Applications of Immunological
Attenuation via Pegylation of Cells and Tissue. BioDrugs 2001, 15, 833-847.
111. Liu, Y.; Shipton, M. K.; Ryan, J.; Kaufman, E. D.; Franzen, S.; Feldheim, D. L.,
Synthesis, Stability, and Cellular Internalization of Gold Nanoparticles Containing Mixed
Peptide-poly(Ethylene Glycol) Monolayers. Anal. Chem. 2007, 79, 2221-2229.
112. Nguyen, C. T.; Niguyen, J. T.; Rutledge, S.; Zhang, J.; Wang, C.; Walker, G., Detection
of Chronic Lymphocytic Leukemi Cell Surface Markers Using Surface Enhanced Raman
Scattering Gold Nanoparticles. Cancer Lett. 2010, 292, 91-97.
146
113. Tam, N. C. M.; Scott, B. M. T.; Voicu, D.; Wilson, B. C.; Zheng, G., Facile Synthesis of
Raman Active Phospholipid Gold Nanoparticles. Bioconjugate Chem. 2010, 21, 2178-
2182.
114. Ip, S.; Maclaughlin, C. M.; Gunari, N.; Walker, G. C., Phospholipid Memrane
Encapsulation of Nanoparticles for Surface-Enhanced Raman Scattering. Langmuir 2011,
27, 7024-7033.
115. Kim, J. H.; Kim, J. S.; Choi, H.; Lee, S., M.; Jun, B. H.; Yu, K. N.; Kuk, E.; Kim, Y. K.;
Jeong, D. H.; Cho, M. H., Nanoparticle Probes with Surface Enhanced Raman
Spectroscopic Tags for Cellular Cancer Targeting. Anal. Chem. 2006, 78, 6967-6973.
116. Wang, X.; Qian, X.; Beitler, J. J.; Chen, G.; Khuri, F. R.; Lewis, M. M.; Shin, H. J.; Nie,
S.; Shin, D. M., Detection of Circulating Tumor Cells in Human Peripheral Blood using
Surface-Enhanced Raman Scattering Nanoparticles. Cancer Res. 2010, 71, 1526-1532.
117. Sokolov, K.; Follen, M.; Aaron, J.; Pavlova, I.; Malpica, A.; Lotan, R.; Richards-Kortum,
R., Real-Time Vital Optical Imaging of Precancer Using Anti-Epidermal Growth Factor
Receptor Antibodies Conjugated to Gold Nanoparticles. Cancer Res. 2003, 63, 1999-
2004.
118. Lee, S.; Kim, S.; Choo, J.; Shin, S. Y.; Lee, Y. H.; Choi, H. Y.; Ha, S.; Kang, K.; Oh, C.
H., Biological Imaging of HEK293 Cells Expressing PLCç1 Using Surface-Enhanced
Raman Microscopy. Anal. Chem. 2007, 79, 916-922.
147
119. Munro, C. H.; Smith, W. E.; Garner, M.; Clarkson, J.; White, P. C., Characterization of
the Surface of a Citrate-Reduced Colloid Optimized for Use as a Substrate for Surface
Enhanced Resonance Raman Scattering. Langmuir 1995, 11, 3712-3720.
120. Jiang, L.; Guan, J.; Zhano, L.; Li, J.; Yang, W., ph-dependent Aggregation of Citrate-
capped Au Nanoparticles Induced by Cu2+ Ions: the Competition Effect of Hydroxyl
Groups with the Carboxyl Groups. Colloid Surface A 2009, 346, 216-220.
121. Nakamoto, K., Infrared and Raman Spectra of Inorganic and Coordination Compounds:
Part B. Wiley: Hoboken, 2009.
122. Karakoti, A. S.; Das, S.; Thevuthasan, S.; Seal, S., PEGylated Inorganic Nanoparticles.
Angew. Chem. Int. Edit. 2011, 50, 1980-1994.
123. Whitesides, G. M.; Kaibinis, P. E., Wet Chemical Approaches to the Characterization of
Organic Surfaces: Self-assembled Monolayers, Wetting, and the Physical-organic
Chemistry of the Solid-liquid Interface. Langmuir 1990, 6, 87-96.
124. Arvizo, R.; hattacharya, R.; Mukherjee, P., Gold Nanoparticles: Opportunities and
Challenges in Nanomedicine. Expert Opin. Drug Deliv. 2010, 7 (753-763).
125. Zhang, Z.; Jia, J.; Lai, Y.; Ma, Y.; Weng, J.; Sun, L., Conjugating Folic Acid to Gold
Nanoparticles Through Glutathione for Targeting and Detecting Cancer Cells. Bioorgan.
Med. Chem. 2010, 18, 5528-5534.
126. Colthup, N. B.; Daly, L. H.; Wiberley, S. E., Introduction to Infrared and Raman
Spectroscopy. Academic Press: New York, 1990.
148
127. Petroski, J.; Chou, M.; Creutz, C., The Coordination Chemistry of Gold Surfaces:
Formation and Far-Infrared Spectra of Alkanethiolate-Capped Gold Nanoparticles. J.
Org. Chem. 2009, 694, 1138-1143.
128. Al-Salady, A. K. H.; Moss, K.; McAuliffe, C. A.; Parish, R. V., Mossbauer and Nuclear
Magnetic Resonance Spectroscopic Studies on 'Myocrisin', 'Solganol', 'Auranofin', and
Related Gold(I) Thiolates. J. Chem. Soc., Dalton Trans. 1984, 1609-1616.
129. Jiang, J. D.; Burstein, E.; Kobayashi, H., Resonant Raman Scattering by Crystal-Violet
Molecules Adsorbed on a Smooth Gold Surface: Evidence for a Charge-Transfer
Excitation. Phys. Rev. Lett. 1986, 57, 1793-1796.
130. Govorov, A. O.; Richardson, H. H., Generating Heat with Metal Nanoparticles. 4ano
Today 2007, 2, 30-38.
131. Govorov, A. O.; Zhang, W.; Skeini, T.; Richardson, H.; Lee, J.; Kotov, N. A., Gold
Nanoparticle Ensembles as Heaters and Actuators: Melting and Collective Plasmon
Resonances. 4anoscale Res. Lett. 2006, 1, 84-90.
132. Invitrogen Malachite Green Isothiocyanate Data Sheet. (accessed July 7).
133. Schwan, J.; Ulrich, S.; Batori, V.; Ehrhardt, H.; Silva, S. R. P., Raman Spectroscopy on
Amorphous Carbon Films. J. Appl. Phys. 1996, 80, 440.
134. Hanfland, M.; Beister, H.; Syassen, K., Graphite Under Pressure: Equation of State and
First-Order Raman Modes. Phys. Rev. B 1989, 39, 12598-12603.
149
135. Reich, S.; Thomsen, C., Raman Spectroscopy of Graphite. Philos. Transact. A Math.
Phys. Eng. Sci. 2004, 362, 2271-2288.
136. Dresselhaus, M. S.; Dresselhaus, G.; Saito, R.; Jorio, A., Raman Spectroscopy of Carbon
Nanotubes. Phys. Rep. 2004, 409, 47-99.
137. Niu, Z.; Fang, Y., Surface-Enhanced Raman Scattering of Single-Walled Carbon
Nanotubes on Silver-Coated and Gold-Coated Filter Paper. J. Colloid Interf. Sci. 2006,
303, 224-228.
138. Thomsen, C.; Reich, S., Raman Scattering in Carbon Nanotubes. Topics Appl. Phys.
2007, 108, 115-232.
139. Tsang, J. C.; Demuth, J. E.; Sanda, P. N.; Kirtley, J. R., Enhanced Raman Scattering from
Carbon Layers on Silver. Chem. Phys. Lett. 1980, 76, 54-57.
140. Goncher, G. M.; Parsons, C. A.; Harris, C. B., Photochemistry on Rough Metal Surfaces.
J. Phys. Chem. 1984, 88, 4200-4209.
141. Kudelski, A.; Pettinger, B., SERS on Carbon Chain Segments: Monitoring Locally
Surface Chemistry. Chem. Phys. Lett. 2000, 321, 356-362.
142. Pettinger, B.; Picardi, G.; Schuster, R.; Ertl, G., Surface Enhanced Raman Spectroscopy:
Towards Single Molecule Spectroscopy. Electrochemistry (Japan) 2000, 68, 942-949.
143. Moyer, P. J.; Schmidt, J.; Eng, L. M.; Meixner, A. J., Surface-Enhanced Raman
Scattering Spectroscopy of Single Carbon Domains on Individual Ag Nanoparticles on a
25 ms Time Scale. J. Am. Chem. Soc. 2000, 122, 5409-5410.
150
144. Mrozek, I.; Pettenkofer, C.; Otto, A., Raman Spectroscopy of Carbon Monoxide
Adsorbed on Silver Island Films. Surf. Sci. 1990, 238, 192-198.
145. Mahojey, M. R.; Howard, M. W.; Cooney, R. P., Carbon Dioxide Conversion to
Hydrocarbons at Silver Electrode Surfaces: Raman Spectroscpic Evidence for Surface
Carbon Intermediates. Chem. Phys. Lett. 1980, 71, 59-63.
146. Otto, A., Raman Spectra of (CN)- Adsorbed at a Silver Surface. Surf. Sci. 1978, 75,
L392-L396.
147. Hore, S.; Kaiser, G.; Hu, Y. S.; Schulz, A., Carbonization of Polyethylene on Gold
Oxide. J. Mater. Chem. 2008, 18, 5589-5591.
148. Kerker, M.; Siiman, O.; Bumm, A.; Wang, D. S., Surface Enhanced Raman Scattering
(SERS) of Citrate Ion Adsorbed on Colloidal Silver. Appl. Opt. 1980, 19, 4137-4137.
149. De Melo, V. H. S.; Zamarion, V. M.; Araki, K.; Toma, H. E., New Insights on Surface-
Enhanced Raman Scattering Based on Controlled Aggregation and Spectroscopic
Studies, DFT Calculations and Symmetry Analysis for 3,6-Bi-2-Pyridyl-1,2,4,5-Tetrazine
Adsorbed onto Citrate-Stabilized Gold Nanoparticles. J. Raman Spectrosc. 2010, 42,
644-652.
150. Lueck, H. B.; Daniel, D. C.; McHale, J. L., Resonance Raman Study of Solvent Effects
on a Series of Triarylmethane Dyes. J. Raman Spectrosc. 2005, 24, 363-370.