SERS using Plasmonic Nanoparticles and
J-aggregate Dyes
By
Colin R. Zamecnik
A thesis submitted in conformity with the requirements
for the degree of Masters of Science
Graduate Department of Chemistry
University of Toronto
© Copyright by Colin R. Zamecnik (2013)
ii
SERS using Plasmonic Nanoparticles and
J-aggregate Dyes
Colin R. Zamecnik
Master of Science.
Department of Chemistry
University of Toronto
2013
Abstract
Semiconductor-metal composite nanoparticles offer optical properties that are superior to those
of pure materials. In this study, we exploit the epsilon near zero (ENZ) phenomenon using gold
and silver nanoparticles functionalized with cyanine dyes, which form distinct J-aggregates on
the surface of the particle. These J-aggregates show a distinct nonlinear optical response, and a
wavelength matching approach was used to couple the plasmonic particle to the J-aggregate.
When the appropriate wavelength is utilized, these particles showed an enhanced SERS signal
as a result of the excitonic resonance of the J-aggregate species. The theoretical properties of the
particles were simulated using FDTD techniques, and these were experimentally verified by
varying the dye/particle distance and testing different molecular and plasmonic resonators.
Experimental SERS spectra had excellent agreement with simulations. These composite
particles were then encapsulated in a lipid bilayer, forming a multi-shell structure with
properties suitable for biosensing and other detection applications.
iii
Acknowledgments
“The most incomprehensible part of the world is that it is all comprehensible.” Albert Einstein.
(who was clearly not a surface chemist.)
I want to firstly acknowledge my supervisor, Prof. Gilbert Walker, who has given me
tremendous support, mentorship, and beyond generous funding. More than that, though, I’m
truly grateful for his trust – questioning my hypotheses in the right places and encouraging
fledgling ideas in others. I have learned a lot about good management and good science under
his guidance, and lessons I will carry with me for the rest of my life.
The success of this project would never have been possible without a very important undergrad,
Chris Walters. Working in parallel with a talented and motivated researcher such as Chris was
integral to propelling this project forward when things were getting bogged down. I cannot
stress enough how important it is to have someone you can just trust to be resourceful and get
the job done. This is an uncommon trait; combined with his work ethic, he has been
indispensable.
I am of the strong belief that everyone is fueled by competition. I don’t mean that negatively – I
mean that having smarter/faster/harder working people around you is fuel in the tank, for your
quite literal engine of production – whether the output is bricks laid or academic papers
published. It’s just how we’re wired. Being in the presence of sharp, intuitive people who ask
insightful questions – it keeps you from getting lazy, complacent. I want to especially thank
Duncan – his deceivingly simple questions have helped me grow and think critically, not just
accept spout the zero order answer we were given during class. He has a perspective all his own,
and I have been humbled on many occasions after rethinking and debating something that had a
deceivingly obvious answer.
And lastly but not least, thanks to my wonderful Jimena; for your undying patience and
understanding. I wouldn’t be here without you.
iv
Table of Contents
Acknowledgments ......................................................................................................................... iii
List of Tables .................................................................................................................................. vi
List of Figures ................................................................................................................................vii
List of Appendices .......................................................................................................................... ix
Overview.......................................................................................................................................... 1
Chapter 1 – Silver Nanoparticles and Model Thiacyanine J-Aggregate Dye ............................. 3
1.1 Technology Introduction ..................................................................................................... 3
1.2 Experimental Methods ......................................................................................................... 6
1.2.1 Instrumentation and Measurement .......................................................................... 8
1.3 Theoretical Methods ............................................................................................................ 9
1.4 Results and Discussion ........................................................................................................ 9
1.4.1 Theoretical Results .................................................................................................. 9
1.4.2 Experimental Results ............................................................................................. 13
1.4.3 Lipid Encapsulation ............................................................................................... 17
Chapter 2 – Optimization for Commercialization .................................................................... 19
2.1 Introduction and Design Flow ........................................................................................... 19
2.2 Materials ............................................................................................................................ 21
2.3 Methods ............................................................................................................................. 21
2.4 Lipid Bilayer Optimization ................................................................................................ 24
2.5 Bilayer Effects on Optical Properties ................................................................................ 28
2.6 Alternate Assemblies and Future Work ............................................................................. 31
Conclusions ................................................................................................................................... 33
Appendix 1 – NMR Data, Temporal Stability, Peak Assignments ............................................... 34
v
A1. 1H NMR data for TMAT ....................................................................................... 34
A2. UV-Vis spectra of lipid encapsulated J-aggregate AgNPs demonstrating
temporal stability. .................................................................................................. 35
A3. Raman Peak Identification for TC dye .................................................................. 36
References...................................................................................................................................... 38
Copyright and Contributing Work ................................................................................................. 43
vi
List of Tables
Table 1: Wavelengths and design requirements of J-Aggregate hybrid nanoparticles ............... 18
vii
List of Figures
Figure 1: Top - Modeled dielectric constant of the J-aggregate layer. Bottom - Field as a function
of wavelength for 40nm diameter particle, 0.5nm linker and 1nm J-aggregate using FDTD
method. Field is plotted at r=21.5, 23.5 and 25.5 nm, respectively. ............................................ 10
Figure 2: Field intensity as a function of distance from the center of the particle. Left side of
each plot with 0.5 nm TMAT linker and right side is naked (no J-aggregate) for reference. ..... 12
Figure 3: a) Naked particle, b,c) with TMAT, TMA linker layers respectively coated in TC J-
aggregate. Irradiated at 407nm. d-f) Irradiated at 514nm. The electric field is strongly confined
within the J-aggregate layer at the 407nm. The electric field is measured in (V/m)2, and the
distance from the center of the particle in nm. ............................................................................ 13
Figure 4: Top - UV-Vis spectra of J-aggregate and TC. Inset: Molecular structure of dye. Bottom
- UV-Vis spectra of J-aggregate AgNPs with TMAT and TMA spacer layers. Inset: Molecular
structures of cationic linkers. ....................................................................................................... 14
Figure 5: Surface-Enhanced (resonance) Raman Spectra of the J-aggregate functionalized
AgNPs. No significant background generated by the lipid bilayer. ............................................ 16
Figure 6: Typical negative-stain TEM images. Top - Linker/J-aggregate functionalized AgNPs,
bottom - functionalized particles after lipid encapsulation. Lipid bilayer appears larger than
expected due to partial fusing with grid upon sample drying. Scale bar is 20nm. ...................... 18
Figure 7: Molecular structures of candidate dyes for Au nanoparticles. Top left: S0046, top
right: S0440, bottom: S0271. ....................................................................................................... 22
Figure 8: UV-Vis of Au compatible dyes. ................................................................................... 22
Figure 9: Raman of Au compatible dyes. Each plot was baselined with cubic spline background
fitting............................................................................................................................................ 24
viii
Figure 10: Biostability study of S0271/Ag particles encapsulated in various lipid formulations.
Top left: in distilled water, Top Right: in PBS, Bottom Left: in PBS + 20% Serum, Bottom
Right: DMM serum media. .......................................................................................................... 27
Figure 11: Fluorescent vs. non-fluorescent S0271 formulations. Raman on left is 25%
DOPC:PEG, Raman on right is 15% DOPC:PEG. Bottom: UV-Vis of both formulations and
molecular structure of S0271 dye. ............................................................................................... 29
Figure 12: Zoomed-in spectra around J-aggregate peak for S0271/Ag particles. ....................... 30
ix
List of Appendices
Appendix 1: NMR data, temporal stability, Raman peak assignments…………………………34
1
Overview
This year, cancer will kill approximately 1500 people per day in America alone, and will be the
third leading cause of death.1 Early screening of the most common cancers is key to leveraging
existing treatment options to their utmost potential. In lung cancer patients early stage
intervention, i.e. resection, can increase median survival rates from 5.0 to 16.0 years.2
Furthermore, screening biases such as high false positive rates mitigate the statistical
significance of these figures. However, that same study shows that a screening benefit for more
aggressive cancers is limited when it is performed in intervals of greater than just three years.
This drives home the need for frequent, accurate screening for individuals, which will only be
realized with simple assays that can be routinely performed for low cost.
The economic relevancy of the emerging cancer clinical environment has not been lost on
nanotechnology researchers and entrepreneurs in biotech.3,4
As of 2006, in vitro diagnostics was
the second most published and patented nanomedicine category – outstripped only by therapeutic
drug delivery – in what was estimated to be a $12 billion market in 2012 and growing steadily.5
Instead of costly, clinically cumbersome techniques such as chest x-rays and CT scans to
perform screening, nanomedicine has taken advantage of a newer phenomenological approach to
cancer development, known as network models.1 This is where molecular distribution pathways
link together many possible biochemical event processes to disease progression. This presents a
promising future for oncologists. The ability to take a snapshot of a patient and their protein
breakdown in various bodily fluids, paired with a clear background known a priori from
previous healthy screens, can give information not only progression of cancer, but its type and
most relevant clinical treatment pathway.
Practically speaking, extracting relevant information from complex bodily fluid sample, most
notably the blood, is difficult. Clinicians need to be able to run assays to identify proteins, cells
and tissue routinely and incorporate these measurements into a diagnostic picture. While
microfluidics and other nanotechnology-enabled platforms have the potential to drastically
improve the scale at which these assays can be administered,6,7
the actual development of
brighter, more sensitive diagnostic labels which would be used in such platforms remains a topic
of great research focus8.
2
Traditionally, histopathology and immunophenotyping have been the preferred techniques for
determining cellular disease state. Histopathology, or examination of tissue or cells for
abnormalities in shape or structure, requires trained technicians and is relatively subjective. It
also requires more extensive sample preparation and is therefore innately more expensive for
hospitals. In light of this, immunophenotyping, especially when used in conjunction with flow
cytometry, has emerged as the dominant diagnostic platform for detection of cancerous cells.
This is typically done with fluorescent labels that indicate the presence – or lack thereof – of
particular cell surface markers, which in turn indicate disease state of the cell in question.9
Unfortunately, the use of fluorescent labels to indicate cell surface markers introduces several
constraints to the immunophenotyping diagnostic platform. The most important of these is the
limit of surface proteins that can be imaged due to broad bandwidths of fluorescent labels.10
A
major thrust in nanomedicine in vitro diagnostics has been to use fluorescent quantum dots to
address this issue.11–13
However, their potential is still limited, as their emission bandwidths are
still tens of nanometers and cannot be overlapped.11
They can also be difficult to synthesize in
large quantities, and those with the relevant optical properties are made from toxic materials such
as cadmium and lead, limiting their clinical utility.
To compete with simple fluorescent labels, nanoparticle based labels need to excel in several
areas: they must be shelf stable, easy to synthesize, compatible with biological fluids, as well as
reasonably non-toxic to achieve FDA approval. Maybe most importantly, with the emergence of
massively-parallel processing techniques such as microfluidics, is the ability to multiplex several
tests in tandem to give a clearer picture to the relevant medical expert. In a market dominated by
fluorescence-based imaging for the past 50 years, it seems a new technology is finally emerging
in the field of nanomedicine and in vitro diagnostics that may address some of these issues faced
by fluorescent labels.
3
Chapter 1 – Silver Nanoparticles and Model Thiacyanine J-
Aggregate Dye
1.1 Technology Introduction
Plasmonic nanomaterials, and their utility in surface enhanced chemical and biological sensors,
have garnered immense interest in the past two decades due to their size-dependent optical
properties.14–17
Surface-enhanced Raman spectroscopy (SERS) is the most widely studied and
offers the possibility of single molecule detection.18
Previously, it was mentioned that fluorescence based contrast agents are limited by their
emission bandwidth overlap – mainly due to their inherent width in the range of tens of
nanometers. SERS nanoparticles inherently address this issue; Raman bandwidths of 1-3nm have
the potential to significantly increase the number of labels that can be detected concurrently on a
cell surface.19
In terms of machine complexity, a SERS based system has the advantage of
detecting several labels using the same excitation wavelength.20,21
The multiplexing capabilities
and reasonably simple integration with existing hardware make SERS nanoparticle labels an
attractive choice for in vitro labeling.22,23
However, due to complex design criteria such as
binding specificity, robust colloidal stability, and optical sensitivity, few commercially viable
sensor systems have been generated as a result of widespread research.
Active plasmonics has generated significant interest, especially in the past few years, as a next
generation platform to address the need for brighter SERS signals.24–28
Rhodamine 6G,29
cytochrome C,25
porphyrin derivatives,30,31
and host-guest charge transfer complexes32
have all
been successfully coupled to plasmonic nanostructures to elicit a much brighter surface-
enhanced resonance Raman (SERRS) signal.33
Even greater enhancement has been achieved
when a wavelength matching approach is employed to couple the molecular and plasmonic
resonances together with the excitation field.32,34,35
Appropriate coordination of multiple
resonating entities offers a powerful tool for brighter SERS-based sensing platforms.36
J-aggregates are one of the most well-studied resonant excitonic species.37
Under sufficiently
high concentration,38,39
or otherwise suitable conditions such as an appropriately charged
surface,40,41
particular cyanine dye molecules will self-assemble into J-aggregates.42,43
J-
aggregates generate a absorption response that is red shifted from the respective monomeric
4
band, which generates a collective exciton within the aggregate.37
When these dyes are adsorbed
onto noble metal nanoparticles in the J-aggregate assembly, the collective exciton can couple to
the surface plasmon,44–48
and interfere both destructively and constructively depending on their
relative energy level positions.49–51
Ultrafast transient state absorption state spectroscopies51,52
as
well as theoretical quantum mechanical treatment53
have verified the presence of both weak- and
strong-type coupling states between the exciton and plasmon.54
The resonances of these
molecules with various plasmonic nanostructures55,56
have been demonstrated previously, and
have shown that predictable spectral overlap and energetic coupling is possible.57
The
orientation58
and adsorption kinetics59
have also been studied in detail, with both thiol-metal
bonded60
and electrostatic-type adsorption.61
3,3′-disulfopropyl-5,5′-dichloro-9-ethyl-thiacarbocyanine (TC) is the most widely studied of
these J-aggregate forming dyes in conjunction with nanoparticles, due to the fact that its
resonance is between that of pure gold and silver. It is photostable, very soluble in water, readily
available, and forms J-aggregates relatively easily. It has been shown in the past that this dye,
and ones of similar structure, will self-assemble into the J-aggregate structure when in the
presence of a strongly cationic surface. This surface can be generated on silver and gold
nanoparticles by use of modified alkanethiols, TMAT and TMA, which contain a cationic
headgroup extending into solution and thiol group bonded to the nanoparticle surface. The
subsequent adsorption and self-assembly of the TC dye does not require external energy (i.e.
heating) or further ionic strength to initiate, and this simple charge-directed electrostatic
adsorption technique was our basis for assembling the anionic TC J-aggregates onto the surface
of metallic nanoparticles in this work.
The epsilon-near-zero (ENZ) effect has recently garnered interest in the field of metamaterials,62
as a way to tailor various sub-wavelength optical properties.63,64
This effect has been predicted to
give very large enhancement factors for SERS generating nanoparticles.65
In the paper on our
work in this area, we presented what we believe is the first use of ENZ J-aggregate materials
used in coordination with metal nanoparticles to prepare a SERS sensory platform.66
While
explained in greater detail later in this thesis, the ENZ effect allows for extraordinary field
enhancement in a material which has a dielectric constant that is near zero. Given SERS intensity
is related to the field enhancement, this opens up potential for very bright particles when careful
wavelength matching is employed.
5
For practical use of nanoparticles as a detection modality, stability in various solutions and shelf
life are of paramount importance. While in water, these particles may exhibit desirable properties
– but when placed in biological media or buffer solution, they tend to aggregate very quickly and
optical properties are greatly diminished as a result. Often, smaller particles are employed during
complicated surface chemistry,40,60
as they are more resistant to aggregation due to electrostatics,
i.e. when charged species or proteins are introduced into solution and around the particle.
However, smaller particles have much weaker plasmon resonances, so there is an effective trade-
off between particle stability and effective optical brightness.17
Several ways to increase the long term stability of larger SERS-active nanoparticles have been
proposed,8 the most popular being the co-adsorption of various PEG chain lengths to the surface
of the particle along with the Raman dye.19,20
Recently, encapsulation of the Raman dye within a
vesicle that surrounds particles has been employed for relatively large particles of 60nm in
diameter.21
Our lab has previously used a 2:2:1 molar ratio of egg sphingomyelin,
dioleoylphosphatidylcholine, and cholesterol dubbed DEC221 to encapsulate Raman-active
nanoparticles. These have shown great promise as shelf-stable, biologically compatible
biosensors, as the lipid bilayer prevents particle aggregation in biological media while offering a
versatile platform for targeting functionality and other surface chemistry.67
The time is right, then, to incorporate J-aggregate forming dyes into an architecture that can
serve as a model SERRS-based detection platform which is both optically bright and stable in
non-ideal conditions. We introduce a highly engineered multi-shell nanoparticle design that
allows facile self-assembly of a J-aggregate forming dye on its surface, and simultaneously
encapsulation within a lipid bilayer to prevent dye leakage, and prevent J-aggregate monolayer
rupture. The synthetic scheme for this process is shown in Scheme 1.
6
Scheme 1: Synthesis of multi-shell J-aggregate plasmonic nanoparticles. Dye and linkers are conjugated in
one step, and subsequently lipid encapsulated with DEC221 lipid formulation. Particles are then washed via
centrifugation before analysis.
These particles are large enough to elicit a bright SERS response under the correct optical
conditions and their optical properties are stable. We employ 3-point wavelength matching to
exploit the ENZ properties of the architecture, aligning the excitation source with the resonances
of the silver particle and J-aggregate chromophore to generate an ultrabright SERS response. We
explore the strong coupling between the dye and particle as a function of their separation
distance, in addition to excitation at co-resonant and non-resonant wavelengths.
1.2 Experimental Methods
Materials: Acetylthiocholine, silver nitrate, sodium citrate (99.0%) and hydrogen
tetracholoroaurate were purchased from Sigma-Aldrich Co. (Canada). Uranyl acetate dehydrate
was purchased from Ted Pella, Inc. (USA). N,N,N-trimethyl-(11-mercaptoundecyl) ammonium
chloride (TMA) was purchased from ProChimia Surfaces sp z o.o. (Poland). 3,3′-disulfopropyl-
7
5,5′-dichloro-9-ethyl-thiacarbocyanine sodium salt (TC) was ordered from Hayashibara
Biochemical Laboratories, Inc. (Japan). Dioleoylphosphatidylcholine (DOPC), egg
sphingomyelin (ESM), and ovine cholesterol (Chol) were received from Avanti Polar Lipids
(USA). All chemicals were used as received. Water was purified with a Millipore Milli-Q water
system to 18.2 MΩ•cm. All glassware was piranha cleaned.a
Synthesis of Silver Nanoparticles: The synthesis of the silver nanoparticles (AgNPs) followed
the reported procedure by Meisel and Lee.68
Briefly, 6 mg of AgNO3 was dissolved in 33.3mL of
H2O and brought to boil at which point 666μL of 1% sodium citrate was added. The solution was
left to reflux for one hour before removing from heat. The final product was a murky greenish
yellow. These were diluted 1:3 in water and stored at 4°C until use.
Thiocholine (TMAT) Synthesis: TMAT was synthesized by simple acid hydrolysis of
commercially available acetylthiocholine in a manner similar to Peng et al.69
500 mg of
acetylthiocholine was dissolved in 15 mL of absolute ethanol and 4 mL of 37% HCl. While
stirring, the solution was refluxed at 100°C for seven hours, after which it was allowed to cool
for 30 min. Excess solvent was removed by rotovap. Recrystallization of thiocholine was carried
out in a H2O/isopropanol/ether (0.5mL/5mL/25mL) solvent system. The solution was
subsequently chilled in an ice bath for 20 min, recovered by filtration and washed with 25mL of
ether, then allowed to dry. The resultant white product, herein referred to as TMAT, was dried in
a desiccator overnight and stored under argon at -20°C. The product’s structure was confirmed
by NMR (see Appendix 1) with purity of approximately 90%. 1H NMR (400MHz, D2O) δ 3.76
(t, J = 8Hz, 2H) 3.23 (m, 11H).
Preparation of Ag/Linker/TC NPs. While stirring, 62.5 μL of 1mM TC was added to 800 μL of
AgNP’s. Separately, 62.5μL of 1mM TMAT or TMA, respectively, was added to 325μL of
deionized water and stirred. The linker solution was then added to the stirring Au/TC solution
and allowed to stir overnight.
a Piranha safety statement: piranha, a mixture of 3-4 parts sulfuric acid to 1 part 30% hydrogen peroxide, is an
extremely potent oxidizer and will corrode through typical safety gear. Polycarbonate face shield, neoprene gloves
and apron as well as full containment in fumed environment is necessary for handling. Active temperatures can
exceed 80°C. Piranha is known to detonate in the presence of organic solvents, so area must be clear of these
compounds. Once handled, it should be diluted 10x with distilled water and left overnight to finish bubbling and
cool, then disposed of separately from other materials.
8
Lipid Preparation. Lipids were prepared as per the protocol reported by Ip et al.21
Briefly, in a
3:1 chloroform/methanol solution, DOPC, ESM and Chol were mixed in a 2:2:1 molar ratio
(DEC221), respectively, to a final mass of 10.7mg. One milligram aliquots were placed in glass
vials and dried under a stream of argon gas until the solvent evaporated, and a film of lipid was
visible on the bottom of each vial. The vials were left to dry under vacuum overnight to remove
any remaining solvent, then backfilled with argon and capped. DEC221 lipids were stored at -
20˚C until use.
Lipid encapsulation of J-aggregate/nanoparticle complex. Prior to encapsulation, DEC221 was
thawed and hydrated with water to a concentration of 1mg/mL. The lipids were then warmed in a
50°C water bath. The lipids were agitated with vortex mixing every 10 min for 30 min until a
multilamellar vesicle suspension was formed. The functionalized nanoparticles were then
encapsulated by adding 1 mL of particles to 1 mL of the DEC221, and sonicating at 50°C for 60
min or until clear. The transition from a cloudy suspension to a clear one was visually indicative
of the formation of unilamellar vesicles around the particles. Once encapsulated, the particles
were washed at 5 kRPM for 10 min to remove excess lipid and dye, and suspended in water to
volume.
1.2.1 Instrumentation and Measurement
UV-Vis Spectroscopy was performed on a Varian Cary 5000 UV-Vis-NIR spectrophotometer.
Particles were placed in a 1-cm-path-length black wall cuvette and spectra were collected at a
scan speed of 240nm/s using 18.2 MΩ•cm water as a blank. Spectra were used to confirm the
plasmon shift and J-aggregate dip or peak on silver nanoparticles, respectively.
Transmission Electron Microscopy (TEM) was performed on a Hitachi H-7000 TEM instrument
operating at 100 kV. Samples were prepared by placing a droplet of aqueous solution containing
the particles on the grid. A small droplet of 2% uranyl acetate solution was added to the larger
one on the grid, which was then allowed to air dry.
Raman measurements were carried out on a Renishaw InVia Confocal Raman Spectrometer,
equipped with a research grade Leica microscope and 50x long-range objective. Coherent 407nm
and 514nm lasers were used. Data was collected with WiRe 2.0 and analyzed with GRAMS
Suite software.
9
1.3 Theoretical Methods
The optical response of the hybrid nanostructure was investigated using finite difference time
domain calculations using FDTD solutions (Lumerical Inc). The dielectric function of the TC
dye was obtained via extinction cross section of the TC dye in the J-aggregate state. This was
easily accomplished by titrating a divalent cationic salt into a 10 µm TC aqueous solution, which
gave rise to almost complete J-aggregate formation once the salt concentration reached 1 mM.
For accurate modeling of the thin spacer and J-aggregate layer, a mesh size of 0.3 nm is used.
The simulation domain is terminated by perfectly matched layer (PML) for minimal reflections.
To calculate the absorption and scattering cross sections of the hybrid structure, we employed the
formalism of the total field scattered field (TFSF).
1.4 Results and Discussion
1.4.1 Theoretical Results
In the FDTD calculations, the silver core is modeled by a fit to the experimental data of Palik70
whereas the exciton resonance of the J-aggregate is modeled using a Lorentz line shape given as
follows
2
0
2 2
0
jagg
jagg
f
i
(1)
where f is the reduced oscillator strength, 0 represents the resonant angular frequency and
jagg is the line width. The permittivity of the modeled J-aggregate layer producing the best fit to
the experimental extinction data is shown in Figure 1 below. The following parameters resulted
in the best fit to the experimental data: 0.33f , 15
0 4.023 10 , 142.4 10jagg and
1.769 . The anisotropy of the J-aggregate response is neglected in this analysis; while this
is not rigorous, the behavior is dominated by the radial permittivity component in the quasi-static
regime. This is close to the present regime of operation and so we expect little deviation from
these calculations for a full anisotropic model.71
10
In this work, we exploit the epsilon near zero (ENZ) phenomenon.65
It is noted that the real part
of jagg approaches zero at a wavelength of 435 nm. This effect can result in strong local field
enhancement at the interface between two different materials, as dictated by the boundary
condition1 1 2 2E E , where
i and iE ( 1,2i ) are the permittivity and normal component of
the electric field at the interface. Thus as the permittivity of one medium approaches zero, the
field strength in that medium is expected to increase significantly. However, it should be noted
that the maximum enhancement achievable is limited by the non-zero imaginary part ofjagg .
Given that the SERS scattering intensity varies according to the electric field around the particle,
the wavelength at which to irradiate the particle was chosen carefully to optimize the ratio
between the dielectric constants of the metal and dye. The calculations optimized for the
maximum field intensity, and the results are shown in Figure 1. The maximum field intensity for
the J-aggregate around the silver particle is 420 nm, and is sharply resonant around that
wavelength.
Figure 1: Top - Modeled dielectric constant of the J-aggregate layer. Bottom - Field as a function of
wavelength for 40nm diameter particle, 0.5nm linker and 1nm J-aggregate using FDTD method. Field is
plotted at r=21.5, 23.5 and 25.5 nm, respectively.
11
In practice, the peak is too narrow to accommodate all Stokes-shifted peaks for the TC dye. We
are interested mostly in enhancement of the pump electromagnetic radiation with regards to the
plasmonic particle. We therefore chose to irradiate the particle at a wavelength of 407 nm, which
falls very close to the plasmon resonance for this particle size, which lies around 410 nm. The
brighter Stokes-shifted peaks then fall very close to the field intensity peak. This wavelength
matching approach is what gives rise to the large enhancement factors in terms of
electromagnetic field around the particle.
Figure 2 shows how the interaction distance between the particle and the J-aggregate dye layer
affects the field intensity in the particle and dye layer. This was done in 1-D fashion across the
hybrid particle and beyond. As expected, the field drop-off is approximately exponential as the
layer is further removed from the particle surface. However, we found that the relationship of
field strength with respect to distance from the particle is heavily dependent on the wavelength.
Figure 3 demonstrates that at the resonant wavelength, the field intensity is greatly attenuated
with respect to distance from the surface. However, at the non-resonant wavelength, the
difference is negligible across the distances considered, as the dielectric constant of the J-
aggregate dye becomes relatively large at wavelengths red of ~450 nm. These 2-D plots show the
dipolar plasmon resonance of the hybrid and naked nanoparticles, and are not spherically
symmetric due to a polarized excitation source.
12
Figure 2: Field intensity as a function of distance from the center of the particle. Left side of each plot with
0.5 nm TMAT linker and right side is naked (no J-aggregate) for reference.
13
Figure 3: a) Naked particle, b,c) with TMAT, TMA linker layers respectively coated in TC J-aggregate.
Irradiated at 407nm. d-f) Irradiated at 514nm. The electric field is strongly confined within the J-aggregate
layer at the 407nm. The electric field is measured in (V/m)2, and the distance from the center of the particle in
nm.
This generates significant field confinement within the J-aggregate dye when irradiated at
407nm, giving values of |E| of over 800 – almost twice the field generated by the surface of the
particle with no dye layer present. Figure 3 shows, however, that there is negligible field
enhancement at the non-resonant wavelength of 514nm, where the dielectric of the dye is
comparable to that of the metal. It is the strong confinement within the dye layer as a result of the
ENZ phenomenon that gives rise to the large increase in SERS scattering intensity as shown in
the following section.
1.4.2 Experimental Results
We attempted to engineer a particle which is able to exploit this unusual optical behaviour and
generate a strong resonance Raman signature from the J-aggregate dye monolayer that surrounds
it. We first synthesized silver nanoparticles which were approximately 40 +/- 5nm in diameter,
confirmed by TEM. These were then functionalized with the J-aggregate layer. Given the lack of
chemistry available to bind this particular dye to the surface of the particle, we employed a
cationic linker similar to Kometani et al,40
which directed the TC dye to self-assemble on the
surface of the particle in the J-aggregate state. This strategy is applicable to most anionic J-
aggregate forming dyes. In order to improve particle stability, the functionalized particles were
then encapsulated within a lipid bilayer. An outline of the synthetic route to generate these
particles is shown in Scheme 1. The plasmon peaks and generation of the J-aggregates were
confirmed via UV-Vis, as shown in Figure 4. We tested two different linker molecule lengths in
order to validate the relationship between field enhancement and distance of the J-aggregate
from the silver particle surface.
14
Figure 4: Top - UV-Vis spectra of J-aggregate and TC. Inset: Molecular structure of dye. Bottom - UV-Vis
spectra of J-aggregate AgNPs with TMAT and TMA spacer layers. Inset: Molecular structures of cationic
linkers.
The shortest linker, TMAT,72
was not commercially available, and was instead synthesized by
acid hydrolysis of an acetylthiocholine precursor. The multi-layer particles were functionalized
by the dye and spacer layer without particle aggregation. In addition to the characteristic ‘peak’
of the J-aggregate in the UV-Vis spectra, its presence was confirmed by ζ-potential
measurements and visually identified via HRTEM. J-aggregates did not form on as-synthesized
citrate-capped particles that were not functionalized with a cationic linker, in agreement with
previous work.61
The particles functionalized with the TMA linker had a somewhat sharper J-aggregate peak in
the UV-Vis spectra compared to the TMAT functionalized particles. Given the more favourable
15
enthalpic interactions between the aliphatic portions of the longer TMA molecules in comparison
to the very short TMAT molecule,73
we postulate that there was greater SAM density when using
the longer linker. Given a discrete charge per linker molecule, the TMA functionalized AgNPs
were better able to accommodate J-aggregate formation of the anionic TC dye, which in turn
generates a larger peak in the UV-Vis spectrum for these particles.
The SERRS spectra of the two types of particles are shown in Figure 5. TC has been previously
shown to give a rather weak Raman signature at traditional pump wavelengths74
with some
improvement as the excitation approaches the J-aggregate resonance located at 458 nm. The
strong peaks in the 400-1000 cm-1
region are indicative of J-aggregate formation within the
proximity of the particle. This is due to enhancement of out of plane vibrational modes of the
Albrecht A term through the resonance Raman effect.75,76
Figure 5 demonstrates SERRS signal at
the resonant wavelength of 407 nm is over an order of magnitude greater in intensity as
compared to the particles irradiated at the off resonant wavelength, and considerably brighter
than when irradiated at wavelengths of 458 and 488 nm.74
16
Figure 5: Surface-Enhanced (resonance) Raman Spectra of the J-aggregate functionalized AgNPs. No
significant background generated by the lipid bilayer.
The SERRS intensity difference between different separation distances from the surface of the
particle also closely followed simulations. The TMAT and TMA molecules were approximated
to be 0.5 and 2 nm in length, respectively, and the J-aggregate monolayer thickness is
approximately 1nm. The relationship between Raman scattering intensity, I, and field intensity,
|E|, is often approximated to be I∝|E|4.17,77,78
The ratio of the field strengths at the two distances
as calculated by the FDTD simulations is 1.15. We can calculate the experimental value for the
ratio of the field intensity using the following equation,
( ) ( (2)
17
Using the average ratio of the peak height between the two thicknesses was taken over eight
major peaks across the spectrum, the ratio of the field strengths was found to be approximately
1.14, which agrees well with the values generated by theory. However, at 514nm, we see a
negligible difference in terms of optical brightness between the two spacer layers, once again in
line with the FDTD simulations.
It is interesting that while there may have been a greater amount of dye on the surface of
particles functionalized with the TMA in comparison to the TMAT, the signal generated seems
to be much more strongly related to the distance between the dye layer and the particle, as seen
in the significantly brighter SERS spectrum for the TMAT functionalized particles at 407nm.
However, all other things constant, the height of the peak was a reasonably accurate diagnostic
for the SERS intensity of a particular sample.
1.4.3 Lipid Encapsulation
The lipid encapsulation of these particles utilized a sphingolipid-cholesterol mixture, which we
have utilized previously in this lab for encapsulation of Raman active nanoparticles.21
These
produced robust particles that proved mechanically and temporally stable over the course of
several months. TEM images of the particles with and without the lipid bilayer around the
particle are shown in Figure 6. The particles settled somewhat over the course of several weeks,
but were easily resuspended with simple shaking of the vial they were contained in.
18
Figure 6: Typical negative-stain TEM images. Top - Linker/J-aggregate functionalized AgNPs, bottom -
functionalized particles after lipid encapsulation. Lipid bilayer appears larger than expected due to partial
fusing with grid upon sample drying. Scale bar is 20nm.
The vesicle confers several advantages to this platform over ‘naked’ particles. Firstly, the
particles can be washed and excess dye can be removed from solution – something that is
impossible to do without the presence of the vesicle to prevent rupture of the J-aggregate
monolayer. This allows facile removal of excess dye from solution, reducing toxicity in
biological systems. Likewise, the presence of the bilayer also ‘traps’ the J-aggregate dye within it
– preventing interaction with other species which may induce undesired chemical or biological
effects, while inhibiting degradation of optical properties due to breakup of the multi-shell
nanoparticle structure.
The lipid vesicle surrounding the particle is an ideal platform for further modification, be it other
Raman dyes or targeting moieties. These can easily be accommodated within the vesicle either
by covalent attachment to the lipid anchor or physical incorporation into the aliphatic portion of
the bilayer.
19
Chapter 2 – Optimization for Commercialization
2.1 Introduction and Design Flow
The first attempt at creating these co-resonant hybrid nanoparticles was successful and
interesting from a scientific perspective, and/or as a proof of concept. The thiacyanine dye used
is the most commonly studied because it is water soluble, photostable, and forms J-aggregates
relatively easily. While a useful prototype of the physical phenomenon that underpins these
ultrabright particles, it does not appropriately take into account the requirements for a clinically
relevant medical diagnostic label.
Consider the technology presented in the first chapter as a set of resonating entities that need to
be aligned, and each wavelength has its own economic constraints that inform the design
process. This is outlined in Table 1.
Table 1: Wavelengths and design requirements of J-Aggregate hybrid nanoparticles
Wavelength Constraints
λ1 – Plasmon resonance of
particle
Solid noble metal nanoparticles have resonances only in the
range of 405-435nm for Ag, 520-550nm for Au.
Alloyed (Au/Ag) particles – tunable depending on mass
fractions between resonances of pure constituents
Nanorods, nanoshells – tunable into the IR, range from 600-
1000nm
λ2 – Exciton resonance of J-
aggregate dye
Limited to commercial availability of dyes, which range in
resonances of 420-600nm.
Dye engineering is possible, if prohibitively expensive.
λ3 – Excitation source What is available in currently compatible technology, i.e. flow
cytometry setups or Raman microscopes. This limits one to
532nm, 633nm, and 785nm, with a small subset of flow
20
cytometers having laser sources in the 458-514nm range.
It’s clear from the table that development of more relevant active-wavelength particles is
constrained mostly by the availability of excitation wavelengths in existing commercial systems
– infringement upon existing intellectual property notwithstanding. The next limiting factor is
the type of plasmonic species, including its ease of synthesis (or purchase), and its compatibility
with the surface chemistry regime that was presented in the first chapter.
There are a massive array of J-aggregate forming dyes,37,79,80
and their electronic structure as a
result of their molecular makeup has been well studied.81–83
More exotic supramolecular designs
such as bilayers and tubes have been demonstrated,43,84,85
and so the potential for coupling with
other nanostructures, at least the in range of reasonably available wavelengths as mentioned in
Table 1, is great. With this knowledge, one could design a dye with desirable resonance and
aggregation properties,42
though purchasing commercially available dyes is obviously more
economical, and allows for faster screening of potential labels.
The lipid encapsulation of the particles – and their eventual conjugation to targeting entities –
isn’t dependent on the dye that is placed in the particle, as it is a physical encapsulation process,
and only loosely dependent on the internal species. It is well known that within a few water
layers, the charge from the particle is mostly screened – and this is enhanced by the presence of
counterions. Therefore the lipid encapsulation isn’t a contributing criteria in the design of the
particles.
It becomes apparent, then, that the optimal design process flows in the order of λ3 λ2 λ1, and
not λ1 λ2 λ3 as we originally followed with the prototypical Ag/TC particles. Practically
speaking, this means that further development routes should first look at what wavelengths are
currently available in existing equipment, then what plasmonic species have relevant resonances
around these wavelengths, and then coordinate with the appropriate dyes that are available
commercially.
21
2.2 Materials
S0046, S0271, and S0440 were ordered from FEW Chemicals GmBh (Bitterfeld-Wolfen,
Germany) and were used as received. All dye stocks were 3mM in methanol and stored at 2°C.
Phosphate buffered saline (PBS) and Dulbecco’s Modified Eagle Medium (DMM) were used as
received from Lonza (Basel, Switzerland). Biostability studies were conducted in flame-sterile
conditions. Dioleoylphosphatidylcholine : polyethylene glycol 2000 (DOPC:PEG) was
purchased from Avanti Polar Lipids (USA).
2.3 Methods
The next stage of the development process then took a more economically relevant turn. We
decided to utilize gold nanoparticles, which were resonant at a readily available wavelength that
many Raman microscopes had equipped – 532nm. This opens up more potential collaboration
with biomedical researchers, as uses for cell and tissue staining become possible.
As per the design flow described earlier, the search began for J-aggregating dyes with resonances
in the 560-590nm range. This would allow for a composite particle that aligned the ENZ point of
the J-aggregate dye with the plasmon resonance of the nanoparticle and excitation source. We
were able to procure several dyes that were screened as potential candidates for use at this
wavelength range. Their structures and spectra are shown in Figures 7 and 8, respectively.
22
Figure 7: Molecular structures of candidate dyes for Au nanoparticles. Top left: S0046, top right: S0440,
bottom: S0271.
Figure 8: UV-Vis of Au compatible dyes.
23
Each of these candidates was adsorbed onto the surface of the gold nanoparticles using the same
chemistry and protocol as the Ag/TC model particles discussed in Chapter 1. Given the clearly
favourable properties of TMAT functionalized particles compared to TMA, the shorter linker
was used in all future formulations for optimal optical performance.
We see clearly that the free J-aggregates in solution, as with the more common TC dye used
earlier, are further red shifted when adsorbed onto the particle surface. This is due to some
constraint in the packing arrangement compared to their free structure in solution as a result of
adsorbing onto a curved surface. This was a useful diagnostic to determine if J-aggregates were
indeed being formed on particles or just simply in solution.
As a general formulation rule, the quality of the monolayer of J-aggregates saturated once the
TMAT concentration reached approximately 1x10-4
M. The Ted Pella nanoparticle final
concentration was roughly 1x1010
particles/mL. The dye concentration was found to have a
similar relationship, as there was saturation in Raman signal when the dye concentration was
greater than 5x10-5
M. The baselined SERRS spectra of each of the three dyes screened are
shown in Figure 9. Unprocessed spectra can be found in Appendix 1-A4.
24
Figure 9: Raman of Au compatible dyes. Each plot was baselined with cubic spline background fitting.
Clearly the S0271 formulation with the gold nanoparticles presents the highest enhancement of
the group, and was the focus of further optimization studies. In accordance with the FDTD
simulations and theory, the closer the dye resonance to the excitation and underlying plasmon
resonance, the stronger the resulting electric field and thus SERS signal intensity.66
These
particles exhibited similar stability and shelf life as the Ag/TC particles that were produced
previously, and were able to be washed and resuspended in water without damage to the J-
aggregate monolayer.
2.4 Lipid Bilayer Optimization
Our system originally utilized the DEC lipid formulation – this was an internal formulation
recommended and used by Dr. Shell Ip and others that resulted in a stable, robust bilayer that
25
allowed us to handle the particles without significant aggregation. Data on utilization of SERS
particles encapsulated with these bilayers is available21
and was therefore our first choice as a
model lipid encapsulation strategy for these J-aggregate particles.
While the Ag/TC particles were not conjugated to any targeting agent, it would be desirable to
have a lipid makeup that would conducive to effective targeting, as well as display appropriate
biocompatibility, stability in biological fluids, and shelf life while preserving optical properties.
While there is limited data on lipid encapsulation of SERS nanoparticles, concurrently,
significant research has been done into lipid encapsulation systems for drug delivery.86–89
Given
the tremendous investment in both the private and public sector into this area, it makes sense to
borrow lipid formulations from this neighboring field, while keeping in mind specific end point
requirements of SERS particles compared to those designed to carry a therapeutic payload.
In particular, we were interested in lipid bilayers that were utilized in a supported structure, i.e.
on a solid, inorganic, and preferably negatively charged particle surface of similar size. Brinker
et. al., for example, have extensively studied lipid encapsulation strategies of mesoporous silica
nanoparticles for drug delivery applications.90–93
This group studied the effect of the fluid
transition temperature on targeting efficiency of their encapsulated particles. The gel-to-fluid
transition temperature, Tm, of the constituent lipids was found to have a pronounced effect on the
ability of a lipid encapsulated particle to bind to a host protein on a cell surface.94
With the
targeting agent anchored to a lipid inside the bilayer, fluid bilayers allow said agent to ‘roam’
freely on the surface. This is one mechanism which allows for better binding probability of these
particles to cells.
Another proposed mechanism takes advantage of the tendency of cell surface proteins to appear
in clusters. Given the diffusion within the bilayer, once one targeting molecule interacts with a
surface protein on a cell, others will be ‘recruited’ to the area of interest probabilistically and
enable multiple interactions to mediate targeting instead of one. With a typical gel-phase bilayer,
i.e. one that is below Tm, the targeting molecule density must be very high in order to obtain
sufficient interaction with the desired surface. This is difficult partly because antibodies are
expensive, but more importantly permits greater non-specific binding of the particles with the
ubiquitous protein environment that surrounds these particles in vivo. Utilizing lipids that are
well above the Tm at physiological temperatures enables the targeting molecule surface density to
26
remain low, and yet still experience multiple-affinity events with the cell surface by means of
this recruiting mechanism. This leaves a high number of remaining ‘empty’ lipids to be
conjugated to a stealth-enhancing molecule such as PEG to further reduce non-specific
interactions.
We then aimed to take advantage of this new information in the design of a more optimized lipid
system for future targeting and stability. The formulation used in the Ag/TC prototype particles
includes DOPC and egg sphingomyelin (ESM) as its constituent structural lipids, as well as
cholesterol. Cholesterol was found to improve fluidity of the lipids and reduce non-specific
binding.94
DOPC has a Tm of -20°C, indicating it is well into the fluid phase at physiological and
benchtop temperatures.
However, egg sphingomyelin is a natural source lipid that contains 86 mol% hexadecanoyl
sphingomyelin (16:0), the rest being a mixture of other saturated lipids with longer fatty acid
tails. The downsides of the SM is its high cost – over 4x that of the DOPC – as well as its innate
batch-to-batch variation due to its natural origin. Synthetic derivatives are over 10x the cost, and
therefore not economically viable.
Upon further inspection the properties of this lipid subset also lent the resulting particles some
non-ideal characteristics. The most prevalent species in the ESM extract has a Tm of
approximately 40°C.95
This means that when combined with a more fluid composition such as
DOPC, one can expect it to form rafts that stick together and ‘float’ among its more mobile
counterpart.96
In our opinion this would lead to negative properties in the final vesicle.
For one, the sphingomyelin is not available conjugated to a PEG molecule to help shield it from
surrounding environment, whereas DOPC is readily available with such chemistry. This means
that if we were to substitute a portion of DOPC for DOPC:PEG, large patches of the particle
may be left uncovered by any PEG – up to half, in fact, given the previous molar equivalent
formulations – exposing it to the surrounding environment and possible undesired non-specific
adsorption. Another reason is the reduced effective fluidity of the bilayer and as a result, the
respective diffusion capability of the targeting species to be anchored to it in future work. As
previous evidence suggests, this may hamper the targeting efficiency of the particle as a whole.
27
As a result of this analysis, we decided to proceed with the testing of several different lipid
formulations. Two variables were tested: first, the amount of DOPC:PEG (Mw = 2000 g/mol)
that was incorporated into the lipid formulation, from 15% to 30% by mol ratio to other
structural lipids. DOPC:PEG was chosen because it was thought to float freely among the non-
conjugated DOPC lipids, allowing for maximal coverage and particle shielding. Second, whether
removing the sphingomyelin from the mixture would improve the stability of the particles in
solution. These formulations were first made according to the procedure outlined above, and then
placed into an equal volume of distilled water, PBS, 20% serum, and serum media with
antibiotic. The stability of the particles is judged by the height of the J-aggregate peak and
amount of particle aggregation, as indicated by absorbance in the IR region. The results are
summarized in Figure 10.
Figure 10: Biostability study of S0271/Ag particles encapsulated in various lipid formulations. Top left: in
distilled water, Top Right: in PBS, Bottom Left: in PBS + 20% Serum, Bottom Right: DMM serum media.
28
The media treated particles were nearly identical, and showed a general broadening of the J-
aggregate and nanoparticle plasmon peak. This is likely due to the protein ‘corona’ that loosely
surrounds all particles in protein rich environments, which forms their so called ‘physiological
identity’.97
However, the PBS and PBS serum data suggests that the DOPC:PEG had lower
aggregation as evidenced by the lower absorption across the IR region, as well as preservation of
the J-aggregate as indicated by peak quality at 560nm. With this data and the potential value of
having a completely fluid bilayer for targeting, it was concluded that the ESM-free, high
PEG:DOPC was a more optimal formulation for coating these J-aggregate SERS particles.
2.5 Bilayer Effects on Optical Properties
While evaluating these new optimized lipid formulations, it quickly became apparent that there
was a significant difference in the optical properties of the S0271/Au particles. With previous
DEC lipid formulations, the particles are highly fluorescent; this was hypothesized to be due to
the close proximity of the excitation wavelength to the fluorescence excitation maxima of the
S0271 dyes.b
However, we noticed that in ESM-free formulations that specifically contained DOPC:PEG mol
ratios of 10-15%, some of these particles lacked this fluorescent background. In formulations
with DOPC:PEG mol ratios greater than 20%, this effect diminished, and a fluorescent
background was almost always present. The unprocessed SERS spectra and UV-Vis are shown
in Figure 11.
b J-aggregates experience a non-Stokes shifted fluorescence at the J-aggregate absorption peak.
37 For the S0271
particles, according to the UV-Vis data, this would be located at approximately 560nm.
29
Figure 11: Fluorescent vs. non-fluorescent S0271 formulations. Raman on left is 25% DOPC:PEG, Raman on
right is 15% DOPC:PEG. Bottom: UV-Vis of both formulations and molecular structure of S0271 dye.
The extremely bright Raman peaks of these dyes are still visible in the fluorescent samples, and
can be extracted relatively easily with baseline removal. It is interesting to note that after
removal, the intensity of the Raman peaks are roughly the same, indicating that in these lipid
formulations, only the fluorescence is mitigated, but the Raman effect is not.
While this was not seen in every sample, we did notice it was a statistically significant trend.
Other variables were also changed to explore more fully whether or not the lipid was the only
agent responsible, such as ionic strength, concentrations of TMAT and dye, cleaning method of
the glassware, and temperature of sonication during encapsulation. While a stochastic element
clearly remains in the fluorescence quenching, the lipid type and composition seems to correlate
most closely to the level of background in the sample.
30
Our hypothesis for this phenomenon is as follows: first, the functionalized particles are likely
oriented with the sulfonyl groups facing outwards into solution.49,59
This means that when the
particles are encapsulated with the DEC formulations, the positively charged headgroup of the
DOPC interacts favourably with the particles. However, with the introduction of some
DOPC:PEG into the system, this effect is mitigated, as the PEG is slightly negatively charged as
indicated by zeta potential.19
This could cause a rearrangement in the packing structure of the J-
aggregate dyes, namely to one that no longer favours fluorescence as a decay pathway for
exciton within the dye superstructure. Lending credence to this idea is a specific J-aggregate
peak shift in the UV-Vis as shown in the zoomed in spectra in Figure 12.
Figure 12: Zoomed-in spectra around J-aggregate peak for S0271/Ag particles.
It is clear that the fluorescent particles with DOPC:PEG mol ratio of 30% have their J-aggregate
peak shifted blue relative to the non-fluorescent, 15% DOPC:PEG formulations. As established
earlier, the J-aggregate peak shifts slightly when the supramolecular structure is altered. This
31
suggests that the PEG environment contributes to the alteration of the packing arrangement in the
J-aggregate dye, likely through electrostatic interactions (or lack thereof).98–100
It is worth noting that the particles with ESM (data not shown) were indeed fluorescent. This
could be interpreted as further proof of the existence of a ‘raft’ of gel phase sphingomyelin, as
we wouldn’t expect this raft of PEG-free lipid to contribute to the rearrangement of the dyes into
a non-fluorescent pattern.
2.6 Alternate Assemblies and Future Work
There is a significant push to develop in vitro imaging and diagnostic platforms that are active in
the IR region, as this limits autofluorescence of native species in intra- and intercellular
environments. If one takes a look at the design requirements in Table 1, it becomes apparent that
the extension of the J-aggregate SERS platform is possible only through using alternative
plasmonic structures that can achieve resonances in the relevant IR region of 700-950nm, where
the fluorescence of native proteins is greatly attenuated. Utilizing particles such as these may be
of more physiological relevance to the medical community. While outside of the scope of this
thesis, it is worth discussing in principle how these alternative nanostructures could developed,
taking advantage of the same ENZ wavelength matching approach as the visible-wavelength
particles outlined here.
The most notable plasmonic particles that fit this profile are nanorods and nanoshells. Nanorods
have been widely investigated as a potential platform for SERS biosensors.101,102
They exhibit
both transverse and longitudinal plasmon peaks that are located near 520nm and 800nm
respectively. The longitudinal peak can be tuned with relative ease over the range of +/- 100nm
from this position by changing the aspect ratio and absolute length of the particles.103
There is
little discussion about functionalization with J-aggregate forming dyes,41,57
likely due to the
propensity of these structures to aggregate without the presence of high concentrations of CTAB
or other similar stabilizers. The surface chemistry that exists natively on these particles is very
different from that of spherical gold nanoparticles, as a result of the synthetic conditions under
which they experience anisotropic growth.104
Nanoshells, or metal nanoparticles with a dielectric core typically made of silica or polystyrene,
have resonances that are even more widely tunable than nanorods. Instead of changing absolute
32
size, it is the thickness of the shell that dictates the plasmon resonance.105
These can be varied
from approximately 600-1100nm, or even further into the red with larger particles.106
Their
absorption in the IR, and the relative lack of absorption of tissue in the region in which they are
active, has driven significant research and investment in these particles as potential photothermal
treatments for cancer.107–109
There has been some limited work done on utilizing these particles
in conjunction with J-aggregates55,56,110
but nanoshells analogues to those shown here have been
experimentally demonstrated. In principle, as these particles follow similar surface chemical
routes as the ones outlined in the scope of this thesis, they make ideal candidates to extend the
library into the IR.
A major hurdle to develop particles is the availability of λ2, i.e. appropriately resonant J-
aggregates in the IR range. These exist, even commercially so; however not in nearly the same
quantity as those in the visible region (550-700nm). The issues extend beyond availability. To
generate lower energy resonances, the size or electron runway of the dye molecule needs to
increase drastically. Given that the J-aggregation phenomenon is a charge-directed self-
assembly,111
increasing the size of the dye molecule or introducing electron sinks/donors to
change the resonance may interfere with its ability to aggregate. While not rigorous, this may be
the reason why there are so few IR resonant J-aggregate dyes discussed in literature in
comparison with visible ones, despite their apparent utility.
33
Conclusions
We have demonstrated a multi-shell nanostructure that has novel nonlinear optical properties that
can be exploited as a uniquely stable and optically bright sensing platform. It is possible to
employ a wavelength matching approach to improve SERS response when used in conjunction
with another resonating entity such as a J-aggregate. We’ve shown that the electric field can be
strongly confined within the J-aggregate monolayer when irradiated at the appropriate
wavelength as a result of the ENZ phenomenon. We experimentally validated the SERRS signal
as a result of field intensity by varying the effective distance between the resonating entities. The
effect of distance with respect to the particle is demonstrated experimentally at both resonant and
non-resonant wavelengths. The effect was further validated by using a different two different
plasmonic substrates and molecular entities.
The lipid encapsulation of these J-aggregate functionalized particles entraps the dye monolayer
within the vesicle, and prevents dye leakage from the particle (see Appendix 1). This allows for
limited interaction of the dye with other species in solution, and simultaneously promotes
mechanical stability of the complex. An optimal lipid formulation for stability in various
conditions, while preserving optical properties, was found.
These model SERRS particles serve as a platform to generate solution-stable biosensors for cell
surface marker detection and other applications. Logical next steps include extension to
alternative nanostructures for potentially expanding these biosensors’ utility in the IR region, as
well as functionalization with targeting agents for cellular identification.
34
Appendix 1 – NMR Data, Temporal Stability, Peak
Assignments
A1. 1H NMR data for TMAT
Figure S1: 1H NMR spectrum of TMAT.
1H NMR(400MHz, D2O) δ 3.76 (t, J = 8Hz, 2H) 3.23 (m, 11H)
35
A2. UV-Vis spectra of lipid encapsulated J-aggregate AgNPs
demonstrating temporal stability.
Figure S2: UV-Vis spectra of washed, lipid encapsulated J-Aggregate AgNPs with a TMA spacer layer. The
spectrum of Ag TMA t=0 d was taken immediately after adsorption of the dye onto the particle. The
spectrum Ag TMA t=21 d was taken 21 days after absorption of the dye onto the particle to confirm absence
of dye leakage into solution. The latter particles had settled to the bottom of the vial, and were gently shaken
to resuspend.
36
A3. Raman Peak Identification for TC dye
Figure S3: Molecular structure of J-Aggregate forming dye, 3,3′-disulfopropyl-5,5′-dichloro-9-ethyl-
thiacarbocyanine sodium salt (TC).
Table S1: Stokes-shifted peaks in SERS spectrum of TC dye J-Aggregates on AgNPs and their corresponding
origin. Peak locations are similar to that identified previously.74,76,112
Peak, cm-1
Description
344 δ(CC aliphatic chains
413
621 υ(C-Cl)
654 υ(C-S) aliphatic
900 υ(CC aliphatic chain vibrations
1244
1463 δ(CH2
1492 υ(CC aromatic ring vibrations
1585 υ(C=C
37
A4. Unprocessed S0440 and S0046 Red Dye Spectra
Figure S4: Unprocessed Raman spectra of S0046 and S0440 dyes in water.
38
References
(1) Heath, J. R.; Davis, M. E. Annu. Rev. Med. 2008, 59, 251–65.
(2) Marcus, P. M.; Bergstralh, E. J.; Fagerstrom, R. M.; Williams, D. E.; Fontana, R.; Taylor,
W. F.; Prorok, P. C. Journal of the National Cancer Institute 2000, 92, 1308–16.
(3) Deura, V. In Nanotechnology in Cancer and Drug Therapy; 2008; p. Chapter 6.
(4) West, J. L.; Halas, N. J. Annu. Rev. Biomed. Eng. 2003, 5, 285–92.
(5) Wagner, V.; Dullaart, A.; Bock, A.-K.; Zweck, A. Nat. Biotechnol. 2006, 24, 1211–7.
(6) Kling, J. Nat. Biotechnol. 2006, 24, 891–893.
(7) Viegers, T.; Pelssers, E.; Lenders, E. Medical Device Technology 2008, 19, 23–25.
(8) Ip, S.; MacLaughlin, C. .; Nguyen, C. .; Walker, G. . In Comprehensive Nanoscience and
Technology; 2011; pp. 59–104.
(9) Naeim, F.; Rao, P. N.; Grody, W. W. Hematopathology: morphology, immunophenotype,
cytogenetics, and molecular approaches; Academic Press, 2009.
(10) Faulds, K.; Barbagallo, R. P.; Keer, J. T.; Smith, W. E.; Graham, D. The Analyst 2004,
129, 567–568.
(11) Xing, Y.; Rao, J. Cancer Biomarkers 2008, 4, 307–319.
(12) Han, M.; Gao, X.; Su, J. Z.; Nie, S. Nat. Biotechnol. 2001, 19, 631–635.
(13) Gao, X.; Cui, Y.; Levenson, R. M.; Chung, L. W. K.; Nie, S. Nat. Biotechnol. 2004, 22,
969–76.
(14) Halas, N. J.; Lal, S.; Chang, W.-S.; Link, S.; Nordlander, P. Chem. Rev. 2011, 111, 3913–
61.
(15) Haes, A. J.; Haynes, C. L.; McFarland, A. D.; Schatz, G. C.; Van Duyne, R. P.; Zou, S.
MRS Bull. 2005, 30.
(16) Halas, N. J.; Lal, S.; Link, S.; Chang, W.-S.; Natelson, D.; Hafner, J. H.; Nordlander, P.
Adv. Mater. 2012, 24, 4842–77.
(17) Campion, A.; Kambhampati, P.; Harris, C. Chem. Soc. Rev. 1998, 27, 241–250.
(18) Kneipp, K.; Wang, Y.; Kneipp, H.; Perelman, L.; Itzkan, I.; Dasari, R.; Feld, M. Phys.
Rev. Lett. 1997, 78, 1667–1670.
(19) Maclaughlin, C. M.; Mullaithilaga, N.; Yang, G.; Ip, S. Y.; Wang, C.; Walker, G. C.
Langmuir 2013, 29, 1908–19.
(20) Maclaughlin, C. M.; Parker, E. P. K.; Walker, G. C.; Wang, C. Nanomed-Nanotechnol
2012, DOI: 10.1016/j.nano.2012.04.003.
(21) Ip, S.; Maclaughlin, C. M.; Gunari, N.; Walker, G. C. Langmuir 2011, 27, 7024–7033.
(22) Abdulhalim, I.; Zourob, M.; Lakhtakia, A. Electromagnetics 2008, 28, 214–242.
(23) Tripp, R.; Dluhy, R.; Zhao, Y. Nano Today 2008, 3, 31–37.
39
(24) Fahrenbach, A. C.; Sampath, S.; Late, D. J.; Barnes, J. C.; Kleinman, S. L.; Valley, N.;
Hartlieb, K. J.; Liu, Z.; Dravid, V. P.; Schatz, G. C. ACS Nano 2012, DOI:
10.1021/nn303553z.
(25) Sivanesan, A.; Kalaivani, G.; Fischer, A.; Stiba, K.; Leimkühler, S.; Weidinger, I. M.
Anal. Chem. 2012, 84, 5759–64.
(26) Witlicki, E. H.; Johnsen, C.; Hansen, S. W.; Silverstein, D. W.; Bottomley, V. J.;
Jeppesen, J. O.; Wong, E. W.; Jensen, L.; Flood, A. H. J. Am. Chem. Soc. 2011, 133,
7288–91.
(27) Zhou, H.; Zhang, Z.; Jiang, C.; Guan, G.; Zhang, K.; Mei, Q.; Liu, R.; Wang, S. Anal.
Chem. 2011, 83, 6913–6917.
(28) Yuan, Q.; Zhang, Y.; Chen, Y.; Wang, R.; Du, C.; Yasun, E.; Tan, W. P. Natl. A. Sci. USA
2011, 108, 9331–6.
(29) Zhao, J.; Jensen, L.; Sung, J.; Zou, S.; Schatz, G. C.; Duyne, R. P. Van J. Am. Chem. Soc.
2007, 129, 7647–56.
(30) Murphy, S.; Huang, L.; Kamat, P. V. J. Phys. Chem. C 2011, 115, 22761–22769.
(31) Haes, A. J.; Zou, S.; Zhao, J.; Schatz, G. C.; Van Duyne, R. P. J. Am. Chem. Soc. 2006,
128, 10905–14.
(32) Witlicki, E. H.; Andersen, S. S.; Hansen, S. W.; Jeppesen, J. O.; Wong, E. W.; Jensen, L.;
Flood, A. H. J. Am. Chem. Soc. 2010, 132, 6099–107.
(33) Rycenga, M.; Cobley, C. M.; Zeng, J.; Li, W.; Moran, C. H.; Zhang, Q.; Qin, D.; Xia, Y.
Chem. Rev. 2011, 111, 3669–712.
(34) Haynes, C. L.; Van Duyne, R. P. J. Phys. Chem. B 2003, 107, 7426–7433.
(35) z elik, S. z elik, I. Akins, D. L. Appl. Phys. Lett. 1998, 73, 1949.
(36) Hao, F.; Nordlander, P.; Sonnefraud, Y.; Van Dorpe, P.; Maier, S. ACS Nano 2009, 3,
643–52.
(37) Würthner, F.; Kaiser, T. E.; Saha-Möller, C. R. Ang. Chem. Int. Ed. 2011, 50, 3376–410.
(38) Jelley, E. E. Nature 1936, 1009–1010.
(39) Scheibe, G. Ang. Chem. Int. Ed. 1937, 50, 212–219.
(40) Yoshida, A.; Yonezawa, Y.; Kometani, N. Langmuir 2009, 25, 6683–9.
(41) Yoshida, A.; Uchida, N.; Kometani, N. Langmuir 2009, 25, 11802–7.
(42) Rossi, U. De; Moll, J.; Spieles, M. J. Prak. Chem-Chem. Ztg. 1995, 337, 203–208.
(43) Eisele, D. M.; Knoester, J.; Kirstein, S.; Rabe, J. P.; Vanden Bout, D. a Nature
Nanotechnology 2009, 4, 658–63.
(44) Hranisavljevic, J.; Dimitrijevic, N. M.; Wurtz, G. a; Wiederrecht, G. P. J. Am. Chem. Soc.
2002, 124, 4536–7.
(45) Sato, T.; Tsugawa, F.; Tomita, T.; Kawasaki, M. Chem. Lett. 2001, 402–403.
(46) Yoshida, A.; Kometani, N. J. Phys. Chem. C 2010, 114, 2867–2872.
(47) Kometani, N.; Tsubonishi, M.; Fujita, T. Langmuir 2001, 578–580.
40
(48) Saito, K. J. Phys. Chem. B 1999, 103, 6579–6583.
(49) Wiederrecht, G. P.; Wurtz, G. a.; Hranisavljevic, J. Nano Lett. 2004, 4, 2121–2125.
(50) Wiederrecht, G.; Hall, J.; Bouhelier, A. Phys. Rev. Lett. 2007, 98, 1–4.
(51) Wiederrecht, G. P.; Wurtz, G. a.; Bouhelier, A. Chem. Phys. Lett. 2008, 461, 171–179.
(52) Uwada, T.; Toyota, R.; Masuhara, H.; Asahi, T. J. Phys. Chem. C 2007, 111, 1549–1552.
(53) Kelley, A. Nano Lett. 2007, 7, 3235–3240.
(54) Bellessa, J.; Bonnand, C.; Plenet, J.; Mugnier, J. Phys. Rev. Lett. 2004, 93, 3–6.
(55) Fofang, N. T.; Grady, N. K.; Fan, Z.; Govorov, A. O.; Halas, N. J. Nano Lett. 2011, 11,
1556–1560.
(56) Fofang, N. T.; Park, T.-H.; Neumann, O.; Mirin, N. a; Nordlander, P.; Halas, N. J. Nano
Lett. 2008, 8, 3481–7.
(57) Wurtz, G. a; Evans, P. R.; Hendren, W.; Atkinson, R.; Dickson, W.; Pollard, R. J.; Zayats,
A. V; Harrison, W.; Bower, C. Nano Lett. 2007, 7, 1297–303.
(58) Owens, R. W.; Smith, D. A. Langmuir 2000, 16, 562–567.
(59) Vujacic, A.; Vasic, V.; Dramicanin, M.; Sovilj, S. P.; Bibic, N.; Hranisavljevic, J.;
Wiederrecht, G. P. J. Phys. Chem. C 2012, 116, 4655–4661.
(60) Menéndez, G. O.; Cortés, E.; Grumelli, D.; Méndez De Leo, L. P.; Williams, F. J.;
Tognalli, N. G.; Fainstein, A.; Vela, M. E.; Jares-Erijman, E. a; Salvarezza, R. C.
Nanoscale 2012, 4, 531–40.
(61) Lim, I.-I. S.; Goroleski, F.; Mott, D.; Kariuki, N.; Ip, W.; Luo, J.; Zhong, C.-J. J. Phys.
Chem. B 2006, 110, 6673–82.
(62) Silveirinha, M.; Engheta, N. Phys. Rev. Lett. 2006, 97, 157403.
(63) Edwards, B.; Alù, A.; Young, M. E.; Silveirinha, M.; Engheta, N. Phys. Rev. Lett. 2008,
100, 33903.
(64) Alù, A.; Silveirinha, M. G.; Salandrino, A.; Engheta, N. Phys. Rev. B 2007, 75, 155410.
(65) Simovski, C. Phys. Rev. B 2009, 79, 214303.
(66) Zamecnik, C. R.; Ahmed, A.; Walters, C. M.; Gordon, R.; Walker, G. C. J. Phys. Chem. C
2013, 117, 1879–1886.
(67) Tam, N. C. M.; Scott, B. M. T.; Voicu, D.; Wilson, B. C.; Zheng, G. Bioconjugate Chem.
2010, 21, 2178–82.
(68) Lee, P. C.; Meisel, D. J. Phys. Chem. 1982, 86, 3391–3395.
(69) Peng, L.; Zhang, G.; Zhang, D.; Xiang, J.; Zhao, R.; Wang, Y.; Zhu, D. Org. Lett. 2009,
11, 4014–7.
(70) Palik, E. D. Handbook of Optical Constants of Solids; Academic press, 1998; Vol. 3.
(71) Gao, L.; Fung, T.; Yu, K.; Qiu, C. Phys. Rev. E 2008, 78, 046609.
(72) Harper, S. L.; Carriere, J. L.; Miller, J. M.; Hutchison, J. E.; Maddux, B. L. S.; Tanguay,
R. L. ACS Nano 2011, 5, 4688–97.
41
(73) Weisbecker, C. S.; Merritt, M. V.; Whitesides, G. M. Langmuir 1996, 12, 3763–3772.
(74) Kitahama, Y.; Tanaka, Y.; Itoh, T.; Ishikawa, M.; Ozaki, Y. Bull. Chem. Soc. Jpn. 2009,
82, 1126–1132.
(75) Akins, D. L. J. Phys. Chem. 1986, 90, 1530–1534.
(76) Aydin, M.; Jean-Mary, F.; Stevens, N.; Akins, D. L. J. Phys. Chem. B 2004, 108, 9695–
9702.
(77) Gersten, J.; Nitzan, A. J. Chem. Phys 1980, 73, 3023.
(78) Bosnick, K.; Maillard, M.; Brus, L. J. Phys. Chem. B 2003, 107, 9964–9972.
(79) Mustroph, H.; Stollenwerk, M.; Bressau, V. Angewandte Chemie (International ed. in
English) 2006, 45, 2016–2035.
(80) Kirstein, S.; Daehne, S. International Journal of Photoenergy 2006, 2006, 1–21.
(81) Von Berlepsch, H.; Kirstein, S.; Hania, R.; Pugzlys, A.; Böttcher, C. J. Phys. Chem. B
2007, 111, 1701–11.
(82) Berlepsch, H. Von; Kirstein, S.; Bottcher, C. J. Phys. Chem. B 2004, 108, 18725–18733.
(83) Mustroph, H.; Reiner, K.; Mistol, J.; Ernst, S.; Keil, D.; Hennig, L. ChemPhysChem 2009,
10, 835–840.
(84) Lyon, J. L.; Eisele, D. M.; Kirstein, S.; Rabe, J. P.; Vanden Bout, D. A.; Stevenson, K. J.
J. Phys. Chem. B 2008, 112, 1260–1268.
(85) Von Berlepsch, H.; Brandenburg, E.; Koksch, B.; Böttcher, C. Langmuir 2010, 26,
11452–60.
(86) Wissing, S. a; Kayser, O.; Müller, R. H. Advanced drug delivery reviews 2004, 56, 1257–
72.
(87) Lukyanov, A. N.; Torchilin, V. P. Advanced drug delivery reviews 2004, 56, 1273–89.
(88) Müller, R. H.; Mäder, K.; Gohla, S. European Journal of Pharmaceutics and
Biopharmaceutics 2000, 50, 161–77.
(89) Guo, X.; Szoka, F. C. Acc. Chem. Res. 2003, 36, 335–41.
(90) Tarn, D.; Ashley, C. E.; Xue, M. I. N.; Carnes, E. C.; Zink, J. I.; Brinker, C. J. Acc. Chem.
Res. 2013, 46, 792–801.
(91) Dengler, E. C.; Liu, J.; Kerwin, A.; Torres, S.; Olcott, C. M.; Bowman, B. N.; Armijo, L.;
Gentry, K.; Wilkerson, J.; Wallace, J.; Jiang, X.; Carnes, E. C.; Brinker, C. J.; Milligan, E.
D. J. Controlled Release 2013, 168, 209–24.
(92) Ashley, C. E.; Carnes, E. C.; Epler, K. E.; Padilla, D. P.; Phillips, G. K.; Castillo, R. E.;
Wilkinson, D. C.; Wilkinson, B. S.; Burgard, C. a; Kalinich, R. M.; Townson, J. L.;
Chackerian, B.; Willman, C. L.; Peabody, D. S.; Wharton, W.; Brinker, C. J. ACS Nano
2012, 6, 2174–88.
(93) Liu, J.; Stace-Naughton, A.; Jiang, X.; Brinker, C. J. J. Am. Chem. Soc. 2009, 131, 1354–
5.
(94) Ashley, C. E.; Carnes, E. C.; Phillips, G. K.; Padilla, D.; Durfee, P. N.; Brown, P. A.;
Hanna, T. N.; Liu, J.; Phillips, B.; Carter, M. B.; Carroll, N. J.; Jiang, X.; Dunphy, D. R.;
42
Willman, C. L.; Petsev, D. N.; Evans, D. G.; Parikh, A. N.; Chackerian, B.; Wharton, W.;
Peabody, D. S.; Brinker, C. J. Nat. Mater. 2011, 10, 389–397.
(95) Barenholz, Y.; Thompson, T. E. Chemistry and physics of lipids 1999, 102, 29–34.
(96) Ip, S.; Li, J. K.; Walker, G. C. Langmuir 2010, 26, 11060–70.
(97) Walkey, C. D.; Chan, W. C. W. Chemical Society reviews 2012, 41, 2780–99.
(98) Kato, N.; Prime, J.; Katagiri, K.; Caruso, F. Langmuir 2004, 20, 5718–23.
(99) Mo, G. C. H.; Yip, C. M. Langmuir 2009, 25, 10719–29.
(100) Tatikolov, A. S.; Costa, S. M. B. Chem. Phys. Lett. 2001, 346, 233–240.
(101) Huang, X.; El-Sayed, I. H.; Qian, W.; El-Sayed, M. a J. Am. Chem. Soc. 2006, 128, 2115–
20.
(102) El-Sayed, M. a Acc. Chem. Res. 2001, 34, 257–64.
(103) Link, S.; Mohamed, M. B.; El-Sayed, M. a. J. Phys. Chem. B 1999, 103, 3073–3077.
(104) Murphy, C. J.; Sau, T. K.; Gole, A. M.; Orendorff, C. J.; Gao, J.; Gou, L.; Hunyadi, S. E.;
Li, T. J. Phys. Chem. B 2005, 109, 13857–70.
(105) Oldenburg, S. .; Averitt, R. .; Westcott, S. .; Halas, N. . Chem. Phys. Lett. 1998, 288, 243–
247.
(106) Westcott, S.; Jackson, J.; Radloff, C.; Halas, N. Phys. Rev. B 2002, 66, 155431.
(107) Loo, C.; Lowery, A.; Halas, N.; West, J.; Drezek, R. Nano letters 2005, 5, 709–11.
(108) Gobin, A. M.; Lee, M. H.; Halas, N. J.; James, W. D.; Drezek, R. a; West, J. L. Nano Lett.
2007, 7, 1929–34.
(109) Loo, C.; Lin, A.; Hirsch, L.; Lee, M.-H.; Barton, J.; Halas, N.; West, J.; Drezek, R.
Technology in Cancer Research & Treatment 2004, 3, 33–40.
(110) Djoumessi Lekeufack, D.; Brioude, A.; Coleman, A. W.; Miele, P.; Bellessa, J.; De Zeng,
L.; Stadelmann, P. Appl. Phys. Lett. 2010, 96, 253107.
(111) Fukumoto, H.; Yonezawa, Y. Thin Solid Films 1998, 327-329, 748–751.
(112) Kitahama, Y.; Ogawa, A.; Tanaka, Y.; Obeidat, S.; Itoh, T.; Ishikawa, M.; Ozaki, Y.
Chem. Phys. Lett. 2010, 493, 309–313.
43
Copyright and Contributing Work
Portions of Chapter 1 are taken from previously authored work, from the citation:
C. Zamecnik, A. Ahmed, C. Walters, R. Gordon, G. C. Walker. “SERS using Lipid
Encapsulated Plasmonic Nanoparticles and J-Aggregates to Create Locally Enhanced
Electric Fields”, J. Phys. Chem. C, Vol. 117, No. 4, pp. 1879-1886, 2013.
Permission in writing from the Journal of Physical Chemistry was given to use the figures and
text from this paper for use in the preceding thesis.
Theoretical methods and FDTD simulations were written by Dr. Aftab Ahmed and Dr. Reuven
Gordon, Department of Electrical and Computer Engineering, University of Victoria, Victoria,
BC, Canada. The theoretical results section of Chapter 1 was written in collaboration with these
authors.
Funding for this project was provided through NSERC Strategic Network for Bioplasmonic
Systems (BiopSys). Valentin Sereda at SUNY Albany facilitated collection of SERS
measurements of the Ag/TC nanoparticles at various wavelengths. Duncan Smith-Halverson
performed the NMR analysis.