Plasmonic Enhancement of Nonlinear Fluorescence in Quantum Dots

shootceaselessΠολεοδομικά Έργα

16 Νοε 2013 (πριν από 3 χρόνια και 7 μήνες)

213 εμφανίσεις

UNM NANOPHOTONICS REU PROGRAM: SUMMER 2012




Plasmonic Enhancement of Nonlinear
Fluorescence in Quantum Dots


Drew Morrill

Department of Physics, Department of Chemistry, Brown
University, Providence, RI

Advisors:
Dr. Ravi Jain (CHTM
-
UNM), Dr. Bernadette Hernandez
-
Sanchez (Sandia Nat’l Labs),
Mike Klopfer (CHTM
-
UNM)




Contents



§1: Nanoparticle Optics

1.1 Introduction

................................
................................
..........................

2


1.2 Semiconductor Quantum Dots

................................
..........................

2


1.3
Fluorescence

in Semiconductor Quantum Dots

.............................

3


1.4 Two Photon Absorption Fluorescence

................................
............

4


1.5 Plasmonics

................................
................................
.............................

5


1.6 Plasmon Resonance

................................
................................
..............

6


§2: Plasmonic Enhancement of TPAF in QD/Spacer/Shell Nanoparticles

2.1 QD/Spacer/Shell Nanoparticles

................................
.......................

8


2.2
TPA Fluorescence Enhanceme
nt

................................
.......................

9


§3: Synthesis
and Characterization

3.1 Introduction

................................
................................
..........................

9


3.2 Synthesis of Quantum Dots

................................
.............................

1
0


3.3 Quantum Dot Surface Preparation

................................
..................

1
1


3.4 Spacer Growth

................................
................................
....................

12

3
.5

Gold Growth

................................
................................
......................

13


3
.6

Characterization Techniques

................................
............................

14


§4:
Results

4.1 Mesoporous

Silica Coating and Encapsulation

..............................

15


4.2 Gold Coating

................................
................................
.......................

16

§5: Conclusions

................................
................................
................................
.....

17

Acknowledgements

................................
................................
............................

18

Works Cited

................................
................................
................................
...........

1
9


Appendix: Synthesis Procedure

................................
................................
.......

20



Morrill
2

§1:
Nanoparticle Optics

1.1 Introduction

The optical properties of metallic nanoparticles

have captivated artisans for centuries, long before
the fascinating microscopic explanation for their unique behavior was made clear by modern
science. Many of the brilliant colors which adorn stained glass windows in the great cathedrals of
Europe can tr
ac
e their origin to “plasmonically
-
active


gold
and silve
r n
anoparticles
.

Scientific
investigations of this phenomenon

most likely began with Faraday’s investigations into colloidal gold
in mid
-
19
th

century (Kelly, 2003). More recently, i
n the last several decades, the field of optics,
enabled by advances in synthetic chemistry, has opened up many more avenues of study in
nanophotonics. This paper will report on the accomplishments of a 10
-
week study geared at
exploring one such avenue: th
e
plasmonic enhancement of nonlinear fluorescence in semiconductor quantum dots
.
Before such an explanation can be properly made, it is critical that sufficient background be
presented to make clear what is meant by the combination of technical terms in th
e title of this
work. Therefore, the remainder of this section will be devoted to relevant scientific background on
the topic.

1.2 Semiconductor Quantum Dots

Semiconductors are materials that exhibit electrical conductivity
whose

magnitude

is somewhere
b
etwe
en
that of
a conductor and an insulator. This corresponds to a ‘moderate’ (generally no greater
than 4 eV) energy gap, known as a
band gap
, between the valence and conduction bands of the
semiconductor. The magnitude of the band gap is the amount of en
ergy required to excite electrons
in the valence band to the conduction band. Thus, materials with a greater band gap tend to
demonstrate more ‘insulator
-
like’ properties, while materials with a lesser band gap tend to behave
more
like

metals.

When an ele
ctron is given enough energy to jump from the valence
band
to the conduction band
(for the purpose of this report, the source of energy for fluorescence is generally one or two
photons), the electron leaves behind a so
-
called
hole
in the now empty state. A


hole


is a conceptual
entity
that

represents the absence, or opposite, of an electron.

The electrostatically
-
bo
und state of an
electron and a

hole is known as an
exciton
,
and is characterized by a length constant known as the
exciton

Bohr radius. This length is critical for determining the size scales in which quantum
confinement becomes a significant factor.

Due to the aforementioned electronic structure, semiconductors behave in exceptionally interesting
ways when fabricated in the

form of quantum dots. A semiconductor
quantum dot
is a nanoscaled
portion of semiconductor which is so small that its excitons are quantum mechanically confined in
all three spatial directions. These confinement effects become strong when size of the part
icle is at
or below the
exciton Bohr radius (typically around 10 nm).

While complicated in nature, quantum confinement arises quite readily out of even the most basic
textbook problem in quantum mechanics.
The ‘particle in a box’ problem describes a parti
cle
Morrill
3

confined to a
specific

spatial dimension,
L
, the b
oundaries of which are dictated by (in
finite
or
finite
)
sharp

potential barriers. Like all systems in quantum mechanics, a wavefunction
Ψ

can be
writt
en which
can be used to deduce a lot of critical

information about the particle. With knowledg
e
of the wavefunction, the time
-
independent Schr
ö
dinger equation can be solved for its energy
eigenvalues
E
(the quantum mechanically allowed energy values of the system):


For large potential barriers, t
he s
olution
of the time
-
independent Schrödinger equation
yields a series
of energy values that are inversely proportional to the square of the dimension
L.
This confinement
energy, when combined with the remaining energy terms for an exciton system, defines th
e possible
energy
values for excitonic transitions
. The implications are immediately clear: the

allowed energy
values for excitonic transitions

are dependent on the size of the quantum dot.


1.3
Fluorescence in Semiconductor Quantum Dots

The size depende
nce of the allowed excitonic energy levels can be

readily observed by measuring the
wavelengths of
the
fluorescence output from adequately
-
excited
semiconductor quantum dots.
(Note:
Fluorescence is the emission of light upon the absorption of electromagne
tic radiation. It
occurs after an electron is exc
ited to a higher energy state leaving behind a hole
. When
the electron
recombines “radiatively” with a

hole and falls back

to a ground state, the emitted photon is usually
in the form of “spontaneous emission”, also called fluorescence)
.

Fluorescence emission can be characterized
by several important quantities and spectra.
The

quantum
yield

quantifies

the efficiency of the fluorescence process, an
d is defined as the following ratio:



A second key way to characterize fluorescence is
through absorption and emission spec
tra, which
give wavelength dependent intensities of the
absorption and emission of light

by a fluorophore.
Since some no
n
-
radiativ
e relaxation of the excited
electron generally occurs (Figure 1), it should be
noted that the emission peak is typically red
-
shifted
from the absorption peak. The relationship between
the absorption and emission spectra differs
significantly for the case o
f
two
-
photon absorption
fluorescence
(TPAF), as will be elucidat
ed in the
following section.

Currently, the use of fluorescent materials for the
Figure 1: A Jablonski diagram demonstrating that
absorption of electromagnetic energy occurs both
vibrationally and electronically, le
ading to non
-
radiative relaxation pathways (figure from Creative
Commons, Public Domain).




E
ˆ
H 




num
be
r
of
phot
ons
emitted
num
be
r
of
phot
ons
abs
or
be
d
Morrill
4

purpose of biological labeling and imaging is dominated by organic fluorophores and dyes. Suc
h
dyes have greatly assisted
biological imaging for research purposes and biomedical applications alike.
However, the use of these dyes
has been limit
ed by a few inadequacies, including their tendency to
photobleach (be photochemically destroyed, eliminati
ng fluorescence)

and their
relatively low

brightness. The use of quantum dots mitigates both of these concerns, in addition to providing
additional benefits, such as easy surface functionalization and easily tunable absorption and emission
frequencies.

1.
4
Two Photon Absorption Fluorescence

Two photon absorption fluorescence (TPAF) occurs when the excitation
of an electron is caused by
the absorption of two photons. Typically, this results in the production of a single photon. For the
purpose of semicondu
ctor quantum dots, the energy of an individual excitation photon is not
sufficient to cause fluorescence; rather, the energy of two photons is combined to generate
fluorescence, resulting in an emitted photon with energy approaching twice that of the absor
bance
frequency.

TPAF is only observable under extreme excitation
intensities
,
as the likelihood of near simul
taneous
absorption of two photons is very low
. Such intensities
are presently only achievable with ultrashort laser pulses
(generally on the order of femtoseconds in temporal
le
n
gth); in this way,
if only for an extremely brief period,
the fluorophore may
be exposed to extraordinarily
intense electromagnetic fields (for the Ti:sapphire laser
used in the reported experiments,
energies on the order
of 100 microjoules are delivered by the

beam in a pulse
lasting approximately

100 femtoseconds).

A key

characte
ristic of

TPAF is that

it

is a nonlinear process
.

Because of this feature, the
intensity of the emi
ssion is
not linearly proportional to the excitation intensity (as is
the case for fluorescence caused by

singl
e photon
absorption), but increases quadratica
lly with the
excitation intensity.

TPAF is often employed in two
-
photon excitation
microscopy, which holds several advantages over single
photon fluorescence spectroscopy. Sinc
e the excitation
source for two
-
photon microscopy
is often in the near
i
nfrared
region, the light required to excite fluorescence
penetrates biological tissue significantly more easily than
the visible light used in single photon microscopy. This
Figure 2: (a)
top
, a multiplexed two
-
photon
excitation microscopy image of a 16
μ
m
cryostat section of a mouse intestine stained
with several different organic dyes (Diaspro,
2006) (b)
bottom
, a comparison of single
photon
fluorescence and TPAF, demonstrating
the enhanced spatial resolution and
decreased output signal amplitude of the later
(Kable Group, web).

(a)

(
b
)

Morrill
5

allows for very deep imaging (typically up to 1 mm). The dramatic difference in wavelengt
h between
emission and excitation frequencies also allows for dramatically easier filtering of the imaging signal
from the excitation signal
, improving the resolution of two
-
photon microscope images (Figure 2a)
.
Furthermore,
the need for very high light in
tensity for TPAF can be exploited for dramatically
improved spatial resolution. Since the TPAF output scales quadratical
ly with the intensity of the
“excitation”

light, emission from the focal waist of the excitation beam may be significantly larger
than e
mission in any
other part of the focused beam. This can be seen as a significant improvement
over single photon fluorescence microscopy, in which, in order to get a usable signal, enough light
must be used so as to cause fluorescence in a broad spatial reg
ion of the sample being imaged
(Figure 2b).

This dramatic increase in spatial resolution can be used to ‘section’ a sample, imaging
different planes at different times, to reconstruct a 3
-
dimension
al

image of the sample

(Larson,
2003)
.
Two photon microscop
y is in need of more specialized fluorophores with greater quantum
efficiencies and high TPA coefficients, a niche which plasmonically enhanced semiconductor
quantum d
ots

may fill very well. Another application of TPAF that can benefit from

brighter
nonlin
ear fluorophores include photodynamic therapy, which employs the relatively high energy
photons created by TPAF to generate cytotoxic reactive oxygen species. These reactive compounds
are capable of causing cell death in cancerous tissue

(Yaghini, 2009)
.

1.5 Plasmonics

Gold and silver, as well as other metals, such as aluminum and copper, are all part of a family of
metals that
exhibit unique

optical properties.

One such property
--

the same property
that

make
s
such metals particularly well
-
suited

to be used as mirrors
--

is the abundance of free conduction
electrons.

The free conduction electrons in a metal are generally modeled as a
free
-
electron plasma,
or a
solid
-
state
plasma.
Modeling the interaction of this plasma with the local electromagne
tic environment is a
challenging problem in electrodynamics. Such challenges arise from the many different factors
affecting

the dynamics of such a system,
including
:
the periodic structure of the metal ions, the
quantum mechanical considerations arising f
rom the fermionic nature of electrons, the correlation
functions describing electron
-
electron interactions, the interaction
of the plasma with impurities
and
phonons (collective excitations in a periodic arr
angement of atoms)
.

Despite the many complicatio
ns, a surprisingly effective method of modeling the free
-
electron
plasma is the
relatively simple

Drude mode
l. The Drude model is based on
the Lorentz model
for
atomic polarizability, which describes the optical response of a metal by modeling the electron
s as
being bound to atoms or molecules with a spring (such that the intera
ction obeys Hooke’s law). For
a
free
-
electron plasma, the conduction electrons are not considered bound to the metal, therefore
the Lorentz model is adjusted by removing the restorin
g force
.
It is important to note that this
approximation
holds only for
intra
-
band transitions.
However, the expression for the dielectric
function of the metal, from which all relevant optical properties can be deduced, is easily adjusted to
account for
i
nter
-
band transitions
by the simple addition of a term.

Morrill
6

Although the use of the Drude model to solve for the dielectric function will yield a good
approximation for the physics related to plasmonic

enhancement of fluorescence, it is nonetheless
valuable to consider
what is meant by the term
plasmon.
[Note that there is some confusion

in the
literature about the actual nature of a plasmon,
in part due to

the differing perspectives of classical
and qu
antum mechanics

(Ru, 2008)
]
.

The term
plasmon
was first introduced in the literature by Pines in 1956. In the introduction to a
review article about the energy losses of electrons moving through metals, Pines records the
following:

"
The valence electron
collective oscillations resemble closely the electronic plasma
oscillations observed in gaseous discharges. We introduce the term ‘plasmon’ to
describe the quantum of elementary excitation associated with this high
-
frequency
collective mot
ion
"

(Pines 1956)
.


In the current literature, it is typical to see a plasmon defined as
the quasiparticle representing a
quantum of collective charge oscillation of the free electrons in a metal. Considering these
defintions, it seems somewhat peculiar at first that qua
ntum mechanics may be unnecessary to study
the optical behavior of plasmons, yet this is the case.

“The dynamics of the plasma is therefore
intricately linked to its optical properties and both can be entirely described using the d
ielectric
function of the

metal,


write
Ru and Etchegoin
in their book
Principles of Surface Enhanced Raman
Spectroscopy.

In elucidating some of the confusion surrounding the nature of plasmons, Ru and
Etchegoin employ an analogy with photons, which are defined as the
quanta of free electromagnetic
field oscillations.
Similarly, plasmons are the quanta of charge density

oscillations in a plasma. Two
important notes must be added to this analogy. The first is somewhat semantic: authors in the
literature often refer to pl
asmons in both quantum and classical situations, whereas the term photon
is generally reserved purely for discussion of the quantum nature of light. The second difference is
more significant: a plasmon is a ‘
quasi
particle,’ not a true quantum particle like

a photon. The
difference

is that a true photon may propa
gate indefinitely, whereas a plasmon is
more susceptible

to
interaction with the medium, and thus
to
decay due to various loss mechanisms, such as collisions.
This difference allows the entire physic
s of the optical behavior a plasmon to be housed in its wave
description under Maxwell’s equations.

With an understanding of the term plasmon, an additional modification to the terminology should
be not
ed for the sake of completeness,

a necessity which un
fortunately may add to the confusion
around plasmons. Most practical discussions of plasmons refer to what is actually a
plasmon
-
polariton.

When referencing plasmons, this paper will refer also to plasmon
-
polaritons. A plasmon
-
polariton is
a mixed electrom
agnetic mode arising from the interaction

between a photon and a plasmon; in
other words,

a photon coupled to the internal degrees of freedom of the plasmon. For the sake of
convenience, the term photon is very often used approximately; the technical defin
ition of a photon
is reserved for a particle of light traveling through a vacuum. Light, travelling even through air,
becomes a polariton due to its interaction with the surrounding media

(Ru, 2008)
.

Morrill
7

1.6
Plasmon Resonance

Plasmons can exist as
propogating surface plasmon
-
polaritons
or as
localized surface plasmon
-
polaritons

(LSP)
.

Solving the dielectric function is generally an easier task for the former, in which a planar
approximation can be made for the interface between the metal and the die
lectric. For the purpose
of plasmonic enhancement of fluorescence, it is most relevant to discuss LSP’s.
The idea of LSP
resonance is perhaps best understood in terms of the simple example of a metal sphere with a
dielectric function
put in the presence of

light and surrounded by a medium with relative
dielectric constant
. In this example, the sphere is much smaller than the wavelength of the light,
whose intensity can be characterized by the

incident field
E
0
.
Using the electrostatic approximation,
the el
ectric field inside the sphere can represented as
E
In
:





The potential for this system to exhibit
resonance behavior is immediately apparent.
As the denominator approa
ches zero (i.e.,



(

)


2

M
), the magnitude of the electric
field on the inside of the metallic sphere
increases dramatically. This condition is
made slightly more complicated by the
complex nature of the dielectric function.
Nonetheless, for c
ertain metals (gold and
silver being prime examples), the imaginary
component of the dielectric function
approaches zero
while
the real component of
the dielectric function approaches

at
roughly the same wavelength

(note, of course,
that the dielectric i
s a function of
). The
contribution of the imaginary component of
the dielectric function cannot be neglected,
as it is proportional to the optical
absorption of the metal, a property
that

damps the desired resonance.

An illustrative example of
an
LSP
resonance
is shown in

Figure 3
, which plots the
theoretical absorption coefficient of a silver
sphere of 25 nm radius in air as a function
of wavelength. Note that the absorption



(

)



M



2

M





E
In

3

M

(

)

2

M
E
0
Figure 4: A simplified illustration of the interaction of a varying
electromagnetic field and the displaced charge density of a cloud
of conduction electrons (green) relative to the nucleus (grey)
(figure from Kelly, 2003).

Figure 3: Absorption coeffi
cienct of a silver sphere
(radius = 25 nm) in air as a function of wavelength under
the electrostatic approximation (figure from Ru, 2008).


Morrill
8

coefficient is proportional to
.

Figure 3

also demonstrates the importance o
f the imaginary
component of the dielectric function, which is the limiting factor in the absorption coefficient,
acting to damp the resonance.
A

diagrammatic
illustration of the mechanics of LSP resonance on a
metallic
nan
oparticle is shown in Figure 4.

§2:
Plasmonic
Enhancement of TPAF
in QD/Spacer/Shell
Nanoparticles

2
.1
QD/Spacer/Shell Nanoparticles

In the previous section, it was shown that plasmonic resonance
s

can significantly enhance the
magnitude of the electric field inside a metallic sphere.
In fact, p
lasmons are
capable of arising in a
variety of nanoparticle geometries, including spherical shells. Shells offer similar electric field
enhancement inside of the cavity, with the added advantage that the intense electromagnetic
environment within

can be utilized for the purpose of fluorescence enhancement (the
electrodynamics of this enhancement will be discussed briefly in the following sub
-
section).
To make
such a fluorescence enhancement possible, a few considerations must be first addressed.

Gold is particularly well suited for the formation of plasmons, yet its exceptional conductive
properties also make it particu
larly effective at
quenching

(that is, decreasing the emiss
ion intensity) of
fluorescence.
When the distance between a very conductive material, such as a gold nanoshell, and a
fluorophore approaches zero, the quantum efficiency of the
fluorophore

also app
roaches zero. This
is due to non
-
radiative energy transfer from the fluorophore to the metal. Since the fluorescence
brightness can be thought of as the product of the quantum efficiency and absorption, this effect
eliminates fluorescence output
(
Norton, 2
011). This type of quenching is, fortunately, preventable
by the addition of a dielectric layer (a layer permeable to electromagnetic radiation but impermeable
to the flow of electrons
) between the fluorophore and the
metal. Such materials include silica (
silicon dioxide), titania
(titanium dioxide), some organic polymers (plastics), and
others. This paper will focus on the use of silica as a dielectric
insulator. The use of a dielectric layer between the gold and the
quantum dot serves an additional purpos
e, as well.
By
controlling the spacing distance, it is possible to further
optimize the electric field enhancement at the site of the
fluorophore, effectively increasing the absorption term in the
fluorescence brightness.

This project sought to fabricate a nanoparticle that could
demonstrate plasmonically enhanced quantum dot fluorescence
through a core/spacer/shell geometry. Figure 5 illustrates the

desired nanoparticle geometry.
Beyond the aforementioned
plasmonic enhance
ment which could theoretically be achieved


E
In
2
Figure 5: a cartoon showing the
core/spacer/shell nanoparticle design.
The total particle radius should ideally
be 20
-
30 nm

Morrill
9

by such a composite nanoparticle, the multi
-
layer design may also mitigate the significant challenge
presented by the toxicity of certain semiconductor materials well suited for fluorescence. Cadmium
sulphide and
selenide “
alloys


make particularly effective
semiconductor
quantum dots, but their use
in biomedical applications has been plagued by concerns of heavy metal poisoning. In a composite
particle of the type illustrated in Figure 5, the cadmium would be well

contained in

layers of
relati
vely non
-
toxic silica and gold, possibly mitigating toxicity concerns.


2.2

Fluorescence Enhancement


The
physics of the spatial dependence of the electromagnetic fields within a metallic nanoshell

is
fundamentally fairly complex. Nonetheless, the application of Mie scattering theory using vector
spherical harmonics produces an effective model of idealized systems.
Klopfer and Jain
(2011) have
report
ed

the application of this method to determine the

effect of various parameters on the electric
field enhancement (EFE) at the center of a multilayered nanostructure

with a 6 nm diameter
quantum dot in the center
. In the report, the effect of several key parameters are highlighted: the
choice of metal (go
ld or silver), excitation wavelength (800 nm or 950 nm), the value of the relative
permittivity of the dielectric, the thickness of the noble metal layer, and the thickness of the
dielectric layer. The local EFE maxima were calculated using a
gradient asce
nt search method
, allowing for the optimization of
the physical specifications for an ultrabright TPA fluorescent
nanoparticle.

Using this model
, it was predicted that an ideal TPAF
multishelled NP
fluorophore would have a gold or silver shell
thickness o
n the order of 2
-
3 nm in thickness (Figure
6),
allowing

for both relative optical transparency and the
facilitation of a strong surface plasmon resonance. Additionally,
using the relatively common dielectric material silicon dioxide
(or silica), it was sho
wn that for thin gold shells (2
-
4 nm), the
optimal thickness of the silica spacer layer is between 15
-
25
nm. These theoretical specifications thus set the design criteria
for the final product, and were fundamental in shaping the
synthesis and characteriza
tion routes.

§
3
:
Synthesis and Characterization

Note: the full synthesis proced
ure can be found in the Appendix (p. 20)
.

3.1 Introduction

The application of the Mie solutions to Maxwell’s equations
has allowed for the optimization of various design parameters
to enhance the TPAF in multishelled nanoparticles. The
Figure 7: “EFE as a function of silica
radius and gold thickness for
QD/silica/gold NP” (figure from
Klopfer, 2011)


Figure 6: “EFE plot with/without QD
[as a function of gold shell thickness],
silica radius = 20 nm” (figure from
Klopfer, 2011).

Morrill
10

fabrication of such a particle presents a dramatically different type of challe
nge in the field of
synthetic nanochemistry.
This section reports on the development of a colloidal synthesis procedure
to fabricate nanoparticles with: (1) a CdSe quantum dot

(QD)

core, (2) a silica shell 15
-
25 nm in
thickness, and (2) a relatively unifor
m 2
-
3 nm gold coating.
It is difficult to avoid a certain degree of
trial and error in the development of a synthesis; therefore, the presented result is informed by many
previous attempts
-

some more successful than others.
In brief, the final synthesis pr
ocedure includes the
following steps:

(1)
Synthesis of quantum dots:
commercial QD’s were
generally employed in the composite NP synthesis due
to their greater monodispersivity and the presence of a
ZnS passivation layer, though a solution
-
pre
cipitation
CdSe QD synthesis was still performed.

(2)

QD surface
preparation
:
the hexadecylamine (HDA)
used as the surface ligand of most commercial QD’s
had to be displaced by TOPO (
trioctylphosphine
oxide
), which binds more strongly to the CdSe surface.


(3)
Spacer growth:
an approx. 20 nm shell of mesoporous
silica was grown
around the QD’s from hydrolyzed
TEOS (tetraethyl orthosilicate). The QD’s were
prepared
in a microemulsion

formed by organic soluble
quantum dots inside of a CTAB stabilized micelle. The
organic solvent and the CTAB are removed, leaving
fluorescent, water soluble silica coated QD’s.

(4)
Gold growth:
poly
-
L
-
histidine is electrostatically
adhered to the silica
surface, providing a structural
‘scaffolding’ for the biomineralization of a thin and
uniform layer of gold, grown out of a reduced solution
of chloroauric acid.


3.2 Synthesis of Quantum Dots

The CdSe quantum dots which comprise the core of
the multishe
ll TPAF nanoparticle were primarily
attained through commercial sources, including Evident
Technologies and Ocean NanoTech. Despite the
availability of these commercial particles, a synthesis of
CdSe QD’s was completed using the solution
-
precipitation meth
od adapted from Bunge
et al.
The
rigorous exclusion of air and water required for this
synthesis provided invaluable experience working with
TEOS (tetraethyl orthosilicate)


HDA (hexadecylamine)


PLH (poly
-
L
-
histidine) hydrochloride


CTAB (cetyltrimethylammonium bromide)


Chloroauric acid


TOPO (trioctylphosphine oxide)


Figure 8: Key molecules used in the
synthesis.


Morrill
11

standard Schlenk line and glove
box techniques. Cadmium solution was prepared under argon
through the pyrolysis (the

anoxic thermochemical decomposition) of Cd(OAc)
2

in the presence of
the solvent TOPO at 320˚C. An anhydrous selenium stock solution was produced in the glovebox
by mixing solid selenium powder with anhydrous toluene and TOP (
trioctylphosphine
).
Approximately 3.0 mL of this solution was pulled into a syringe and rapidly injected into the heated
cadmium solution, initiating the temporally dependent aggregation of CdSe quantum dots. Time
dependent aliquots were removed at various times (approxima
tely 10 samples between 5s
-

3600s)
and rapidly mixed with a solution of methanol and acetone, quenching the growth of the quantum
dots (Bunge, 2003). This produced TOPO stabilized (with TOPO acting as the surface ligand)
QD’s of a variety of sizes; the si
ze dispersion was readily apparent in the spectrum of fluorescence
emission wavelengths produced by the different samples.

3.3

Quantum Dot Surface Preparation


Many aspects of classical synthetic chemistry can be employed in the
synthesis of nanoparticles
-

for instance, understanding the resp
onse of
a product to different solvent environments is of great concern. The
vital role that surface chemistry plays in the synthesis of composite
nanoparticles provides a challenge less commonly found in traditional
synthesis problems. The properties of a

nanoparticle are in many ways
intermediate to both bulk and atomic/molecular systems, and a good
example of this intermediacy is in the surface chemistry of
nanoparticles. The surface of a colloidal quantum dot will interact with
the solvent in which it i
s suspended due to the dramatically different
electronic properties of the two materials. If the QD surface is not
capped
or adequately prepared with a surface ligand, it is likely that this
electronic dissonance will favor the aggregation of the nanoparti
cles in order to lower the net surface
energy of the system
-

the result of which eliminates many of the useful properties of QD’s.
Therefore, a surface ligand must be chosen which
not only preve
nts aggregation, but also solubi
lizes
the pa
rticle in the desi
red solvent. Additionally, i
n this synthesis, the
strength of the interaction
which joins the QD and ligand must be of the correct magnitude, so as to make the QD an effective
precursor. Surface ligands are illustrated in Figure 9.

In addition to the choi
ce of surface ligand, the presence of a
passivation layer
is also critical. This layer
is generally comprised of several monolayers of ZnS (or some other semiconductor material) coating
the CdSe QD. Such a layer has several useful effects. If an organic su
rface ligand binds directly to
the surface of the fluorescent material (in this case, CdSe), surface tra
p states are readily formed.
The
s
e

traps can cause decreased quantum yields, and are linked to blinking. Added advantages
include protection against pot
entially damaging environments and photo
-
degradation.

The method of silica coating was adapted from a procedure reported by Hu
et al
. QD’s with a
hydrophobic surface ligand are dissolved in the organic and hydrophobic solvent chloroform. A
small quantity
of this solution is added to an aqueous solution of CTAB and stirred vigorously to
Figure 9: a quantum dot (red)
with surface ligands. Note the
blue interacting head
(generally a functional group)
and the non
-
in
teraction tail,
which determines the
solubility of the QD
(Guerrero
-
Martinez, 2010).

Morrill
12

obtain a homogenous microemulsion (Hu, 2009). This is allowed by the formation of a small
spherical shell of CTAB molecules, all oriented in such a way that their hydrophobi
c tails are pointed
to the interior of the shell (containing chloroform) while the polar head points to the aqueous
exterior. This configuration is known as a
micelle.
The organic cavity within the micelle acts as an
energy well, which draws in the hydroph
obic quantum dots. At this stage, the solution is heated to
50˚C, which dramatically accelerates the evaporation of the chloroform from within the micelle.

When all the
chloroform has been evaporated, the inward facing carbon chains from the CTAB are
forced to interdigitate with the carbon chains of the surface ligand, an interaction strengthened by
Van der Waals forces. The result is illustrated in the ce
nter diagram of Figure 10.

The particles have now been made soluble in water
without
dramatic alteration of the surface
chemistry of the QD. This aspect of the synthesis is critical for maintaining high fluorescent
quantum yield, which can easily be lost

d
uring encapsulation procedures through the formation of
surface trap states.

3.
4

Spacer Growth

When TEOS is dissolved in an aqueous solution, the ethyl group leaves the molecule and is replaced
by a hydrogen, forming what is known as a silanol

(characterized by Si
-
OH). It is ultimately this
hydrolyzed form of TEOS
that

acts as the precursor for silica growth, not the original ethyl ester.
Following the water solubilization of the nanoparticles, the CTAB encapsulated, TOPO coated
QD’s have a rel
atively positive zeta potential
1

of 50 mV, according to Hu
et al.
Not only does this
ultimately provide the electronic basis for the solubility of these particles in polar solvents, but it also
facilitates the interaction with the relatively negative, now
-
hydrolyzed TEOS dissolved in the
reaction mixture.

While the first reaction TEOS must undergo in the formation of the silica shell is considered
hydrolysis, characterized by the addition of water, the next reaction it must undergo is the opposite:



1

The zeta
-
potential, so named for the Greek letter commonly used to denote its value, is a way to
measure electrical potential of a colloidal system. It represents the potential difference between the
particle and the bulk fluid in which it resides.

Figure 10: An overview of the method of silica encapsulation used in this procedure (Hu, 2009).

Morrill
13

condens
ation, or the removal of water. In the presence of a base (in this synthesis, sodium
hydroxide is used), the silanol group deprotonates, leaving behind an Si
-
O
-

group. The free floating
proton,

doing all that it can to allow the
mixture

to satisfy Le C
hatelier's principle

and reach
equilibrium
, then pulls a hydroxyl group from another silanol, leaving behind Si
+

and forming water
in the process. The Si
-
O
-

then
bonds to the Si
+

of another hydrolyzed TEOS. Over and over again,
Si
-
O
-
Si linkages are formed
around the relatively positive quantum dot, encapsulating the
semiconductor nanoparticle in a mineral
-
like solid
, represented in Figure 11.

Over the course of approximately three hours, a silica shell
is grown to approximately 20 nm in thi
ckness. However, at
this stage, a significant number of CTAB molecules are
threaded through the siloxane matrix encapsulating the
QD. The electronic properties of CTAB make it
particularly effective at quenching fluorescence. Therefore,
in order to keep th
e QD fluoresc
ent, it is necessary to
rapidly

and completely remove all of the CTAB from the
reaction solution, including that which is intertwined in the
siloxane matrix. This is done through a series of ethanol
washes. After the washes, the product has ch
anged in only
two ways: (1) the fluorescence has been preserved, and (2)
the siloxane matrix is no
w

riddled with nanometer scale

holes, in which the CTAB previously resided, thus making
the silica coating
mesoporous.
It is relevant to note that at this
poi
nt, the outermost functional groups of the nanoparticle
are now silanol groups. Since this leaves hydroxyl groups
exposed to the aqueous solution, the surface of the
nanoparticle
obtains again

a negative zeta potential.

3.
5

Gold Growth

In order for the gold shell to be both optically transparent and be able
to support a surface plasmon
, it must be relatively thin (approx. 3 nm);
for this plasmon to cause the desired EFE, the particle must also be
relatively small (approx. 50 nm diameter). Thicker gold coatings on
larger nanoparticles has previously been achieved, typically using
particl
es greater than 100 nm in diameter (Oldenburg, 1998). The
challenge arises in creating a
thin

gold coating on a small particle. In
2009, such a feat was
reported

in the literature by
Jin

and Gao

(see
Figure 12)
. This was achieved through the use of the bio
molecule poly
-
L
-
histidine (PLH), a poly amino acid containing an imidazole
functional group and characterized by a relatively positive charge. The
affinity of histidines for metal ions, including Au
3+
, is well established
Figure 11: A schematic illustration (not to scale)
of the encapsulation of the QD with a siloxane
shell through

condensation. The blue shell
around the red QD represents the ZnS
passivation layer, and the green shading
represents the surface ligands (see Figure 10).

Figure 12: A transmission el
ectron
microscopy image of
QD/spacer/Au
-
shell NP’s
synthesized by Jin and Gao (figure
from Jin, 2009).

Morrill
14

in the literature and is commonly
employed in protein purification. Jin and Gao employed PLH as a
‘scaffolding’ on which biomineralization of the gold can occur (Jin, 2009).

Jin and Gao deposited PLH onto a layer of polyethylene glycol functionalized with carboxylic acid;
after deprotonat
ion of the acid, the surface obtains a negative zeta potential, allowing for the
electrostatic adsorption of the PLH onto the functionalized QD surface. Inspired by this work, the
growth of gold on the silica spacer similarly occurs through PLH mediated bi
omineralization
-

the
primary difference being the substrate onto which the PLH is adsorbed. Fortunately, the surface of
the silica nanoparticle is also negative, allowing for a similar electrostatic interaction.

Once the silica coated particle
s

are functi
onalized with PLH, they are then
mixed into a chloroauric
acid solution (basified with sodium hydroxide to pH 9
-
10). After a 10 min incubation period, the
gold is reduced with hydroxylamine, stimulating the deposition of gold on the imidazole rich surface
of the nanoparticle. The thickness of the gold layer can be adjusted by changing the quantity of gold
and reducing agent added. To prevent aggregation of the particles, methoxy polyethylene glycol thiol
(mPEG
-
thiol) may be used to stabilize the particles.
Polyethylene glycol (PEG) derivatives are
commonly used as a surfactant due to their water solubility. The addition of the thiol group allows
for the formation of strong gold
-
thiol linkages, binding the PEG to the surface of the nanoparticle.

3.5

Characte
rization Techniques

Synthetic nanochemistry is frequently impeded by challenges presented in analytical nanochemistry
-

that is, the characterization of nanoparticles. Despite challenges, several techniques prove
exceptionally useful in characterization. T
hose techniques relevant to this project will be briefly
highlighted, so as to clarify the nature of certain results presented
in §4
.

Transmission Electron Microscopy (TEM):
A beam of electrons is passed through a very thin sample, and
an image results from the interaction of the electrons with the sample. This image is magnified and
focused onto a detector. Exceptionally high spatial resolution is allowed by the very small de

Broglie
wavelength of an electron, which dramatically reduces the diffraction effects
that

limit optical
microscopy. This technique is useful in characterizing the spatial dependence of the composition of
a sample (for instance, in determining the distanc
e between a QD and a gold shell).

Energy Dispersive X
-
ray Spectroscopy (EDS):

Often incorporated into a TEM machine, EDS employs a
beam of high energy electrons (or protons or X
-
rays) to eject ground state electrons from a sample.
Electrons from a higher e
nergy state fall down a considerable potential difference and fill the hole
left
by

the ejected electron, but in the process generate X
-
rays with energies characteristic of the
element which generated them. The quantity and energy of these X
-
rays is measur
ed, and an energy
spectrum is produced
,

identifying the elements present in a sample (a spectrum which could be used,
for instance, in determining whether a dark spot in a TEM image is a gold NP or CdSe QD).

UV
-
Vis Absorbance Spectroscopy:

UV
-
Vis measures

the amount of absorbance of a material at various
wavelengths. This is achieved by shining white light (relatively consistent intensities across different
wavelengths in the visible and near
-
visible region) into a sample. Certain wavelengths will be more
Morrill
15

readily absorbed, as they correspond to allowed electronic transition (generally, from some ground
state to an excited state). The energies of such electronic transitions, and therefore the features on an
absorbance spectrum, are characteristic of the mate
rial being measured (for instance, the absorbance
peak changes with the degree of quantum confinement in a QD).

Spectrofluorometry:
While UV
-
Vis measures the amount of energy associated with electronic transitions
from a ground state to an excited state,
spectrofluorometry measures the energy which is released
when fluorescence occurs and an electron falls to a ground state and emits light. While absorption
can occur across a wide range of vibrational states, fluorescence emission primarily occurs from the

lowest energy vibrational state of the excited electronic state (see the Jablonski diagram in Figure 1),
yielding a much sharper peak. For the case of nonlinear fluorescence, spectrofluorometric
measurements are even more critical, as absorption and emiss
ion wavelengths vary dramatically.

§
4
:
Results

4.1 Mesoporous Silica Coating and Encapsulation

The commercial QD’s used in this study are prepared with HDA as a
surface ligand. It was found that the affinity of this ligand for the QD
surface was insufficient for the purposes of mesoporous silica coating.
When the procedure was run with HDA as a liga
nd, the silica particles
produced, while relatively monodisperse and generally of the correct
dimensions and geometry, were found not to contain QD’s at their cores.
Figure 13 shows the TEM resulting from silica encapsulation with HDA.
In addition to the e
vidence provided by the TEM images, the bulk sample
demonstrated no noticeable fluorescence, further indicating the absence
(or quenching) of QD’s.

The exact mechanism which prevented encapsulation is difficult to
determine. However, the amine
-
QD interact
ion is relatively weak, making
HDA a somewhat labile capping agent.
Guerrero
-
Martinez
et al
report that
such labile capping ligands can be easily replaced by more aggressive
surfactants, such as CTAB (2010). If such a ligand exchange took place, it
is unli
kely that the QD’s would have developed the desired solubility.
Without the correct solubility, it is unlikely that the QD’s would be
incorporated at all in the micelles.

As a result of this hypothesis, a ligand exchange reaction was run to replace the HD
A with TOPO, a
much stronger binding capping agent. Unfortunately, due to insufficient washing, the final product
of this ligand exchange reaction contained toluene, an organic solvent with insufficient volatility to
facilitate the microemulsion technique
described in Section 3.3. While the apparent ‘product’ of this
reaction did indeed demonstrate fluorescence, it seemed to primarily result from an oily residue
adhered to the side of the reaction flask. Since the boiling point of toluene is well above the
50˚C
Figure 13: TEM of the result of
attempted encapsulation of CdSe
QD’s in silica. Note the absence of
QD’s and the ‘texture’ indicating
mesopores.

Morrill
16

Silicon

Gold

Figure 14:
Left
, EDS spectrum of the inset TEM sample. The dramatic difference in electron transmission properties of the dark
spots and the larger particle, as indic
ated in the contrast of the image, suggest that the speckling may be caused by the growth of
small gold clusters upon a silica particle. The EDS spectrum agrees with this hypothesis, demonstrating peaks for both gold a
nd
silica.
Right
, another TEM image of

gold speckled silica nanospheres (scale bar = 20 nm).

that the encapsulation reaction is run at, there is no reason that the toluene would have evaporated
off in the manner of chlorofom, the desired organic solvent. It is likely, then, that instead of
incorporating into micelles, the QD’s remained in tol
uene throughout the encapsulation reaction.

4.2 Gold Coating

Although the QD core is absolutely critical for the fabrication of the desired final product,
encapsulation of a QD is not required to investigate the gold coating of a silica particle. The
biominera
lization technique proved more successful than the attempted
encapsulation.












Figure 14 shows both TEM images and an EDS spec
trum

of an
attempt to apply a thin gold coating to silica nanospheres using PLH as
a ‘scaffolding’ for biomineralization. While no QD’s are present in the
core, gold nanocrystals are clearly seen growing in vi
rtually every region
of the silica nanosphere surface. This result is particularly promising,
considering previous attempts to gold coat silica spheres using
alternative gold
-
silica linkers (two different silicates, one functionalized
with an amine and the

other employing a thiol) resulted in not uniform
seeding of very small particles across the surface of the sphere, but
rather relatively large gold crystals growing unevenly on the surface
2
.


This previous result is shown in Figure 15 for comparison.




2

It
should be noted that the arrangement shown in Figure 15, although different from the intended
core/spacer/shell product, may also be an effective system for plasmonic enhancement of
fluorescence and is worthy of further study.

Figure 15: TEM image of an attempted
gold coating of silica nanospheres
using amino
-

and mercapto
-
silicates
(scale bar = 10 nm).

Morrill
17


!
(
a
)
$
(
c
)
$
F
i
g
u
r
e
$
1
6
:
$
(
a
)
$
g
o
l
d
$
c
o
a
t
e
d
$
s
i
l
i
c
a
$
p
a
r
t
i
c
l
e
s
$
w
i
t
h
$
i
n
c
r
e
a
s
i
n
g
$
(
l
e
f
t
$
t
o
$
r
i
g
h
t
)
$
q
u
a
n
t
i
t
i
e
s
$
o
f
$
H
A
u
C
l
4
/
N
H
2
O
H
,
$
(
b
)
$
T
E
M
$
i
m
a
g
e
$
o
f
$
g
o
l
d
$
c
o
a
t
e
d
$
s
i
l
i
c
a
$
N
P

s
$
f
r
o
m
$
c
e
n
t
r
a
l
$
v
i
a
l
$
i
n
$
(
a
)
,
$
i
n
d
i
c
a
t
i
n
g
$
a
g
g
r
e
g
a
t
i
o
n
$
a
n
d
$
t
h
i
c
k
$
s
h
e
l
l
s
$
(
s
c
a
l
e
$
b
a
r
$
=
$
2
0
0
$
n
m
)
,
$
(
c
)
$
c
l
o
s
e
r
$
l
o
o
k
$
a
t
$
(
b
)
%
(
b
a
r
$
=
$
1
0
$
n
m
)
.
$
$
(
b
)
$

While the TEM images in Figure 14 seem promising, the
gold ‘layer’ generated was less a layer and more a speckling
of gold nanocrystals. To achieve the uniform layer desired,
the procedure was modified through the addition of a
greater quantity of chl
oroauric acid (with a proportional
increase in the amount of reducing agent). The photograph
in Figure 16(a) shows three different quantities of reduced
gold solution (increasing left to right) added to equivalent
silica nanosphere solutions. The crystalli
zation of the gold is
immediately apparent from the increase in optical thickness
of the solutions, and it is interesting to note the subtle
bluish tint adopted by the solution, likely due to gold’s
strong plasmonic absorption in the red. Due to time
limit
ations on the TEM, only the sample in the central vial
in Figure 4(a) could be analyzed with TEM, and the
resulting images are presented in Figures 4(b) and 4(c).
Furthermore, due to the strength of the gold signal, the
EDS spectrum on Figure 4(b) was inco
nclusive on whether
or not silicon was present. However, should Figure’s 4(b)
and 4(c) be gold coated silica particles, it is particularly
interesting to note the transition in gold coatings from a
small speckling to what appears to be a dramatically thick
er
and smoother coating.

The differences in appearance between the gold coatings in

Figures 14 and 16 may

be
due to
an arti
fact of the
experiment, instead of
the intended control variable (the
amount of gold added in the shell growth procedure). Notably,
the precursor silica nanospheres in
Figure 16 were likely formed in the presence of toluene (see Section 4.1), unlike the particles in
Figure 16. This artifact may indeed have changed the gold biomineralization process in ways difficult
to predict.

§
5
:
Co
nclusions

At present, there is a critical need for high efficiency and ultra
-
bright fluorescent nanoparticles.
Applications can readily be found in opto
-
electronics, biotechnology,
fundamental biological
research,
and even in cancer therapies. With quantum yields approaching maximum efficiency, high
photochemical stability, and the many unique benefits that arise from nanoparticles, such as size
-
tunable properties and easy surface functionalization, CdSe/ZnS quantum

dots are particularly well
suited for many of these demanding applications. Yet despite these impressive properties,
electrodynamic models clearly indicate that by utilizing plasmonics and modern methods in
Morrill
18

nanofabrication, it should be possible to synthe
size a core/spacer/shell particle with significantly
brighter fluorescence, with the added advantage of improved physical isolation of toxic
semiconductor materials. In the case of TPAF, due to the nonlinear dependence of fluorescence
intensity on the loca
l electromagnetic environment, models suggest even more dramatic
enhancements. Using vector spherical harmonics and the Mie solutions to Maxwell’s equations, it is
predicted that by placing a fluorophore within an optimized plasmonic nanostructure, TPAF si
gnal
intensities may be enhanced by a factor of
>160,000 from the
unmodified fluorophore (Klopfer, 2011).

This report seeks to introduce and elucidate some o
f the key points in
the physics and chemistry behind the design and synthesis of a
core/spacer/shell plasmonically enhanced fluorescent nanoparticle.
Additionally, it reports on attempts at the fabrication of such a particle
using the following three compo
nents: (1) a fluorescent semiconductor
CdSe/ZnS core/shell quantum dot as the core, (2) a 15
-
25 nm silica
-
based dielectric spacer to insulate the semiconductor and the metal and
optimize EFE, and (3) a uniform gold coating of ~2
-
3 nm to facilitate a
surfac
e plasmon and give rise to dramatic EFE. While significant
ground was made on engineering a synthesis, still more must be done
before dramatic fluorescence enhancements are realized. The future outlook of this project may
indeed look much different than it

does now. Due to potential constraints placed on the physics of
the plasmonic system by the very real demands of chemical synthesis, it may, for instance, be
necessary to consider not a thin and uniform gold shell, but rather a
system of small gold
nanocr
ystals deposited on the surface of the dielectric (see Figure 17).

Once a successful fab
rication technique is found, numerous

applications
are anticipated for these

bright and uniqu
e fluorophores
. Like any experimental science, biology is moved forward b
y
improvements in our ability to observe previously unclear processes. Plasmonic enhancement of
fluorescent nanoparticles, once realized, stands in a ready position to focus our perspective on some
of life’s fundamental questions, and indeed it may do so i
n unexpected ways. This report has focused
primarily on fluorescence as a standalone imaging technique, yet much of the research surrounding
plasmonics
instead
considers biosensors that
not only employ fluorescence, but also the extreme

sensitivity of the
plasmonic resonance peak to the local dielectric environment. Such imaging and
detection schemes
exploiting these

nano
particle
s

may someday revolutionize our ability to probe
biolo
gy and cure illness; nevertheless, it is very likely that
the most exciting
applications
have not yet
been anticipated.

Acknowledgments

I would like to thank
Dr. Ravi Jain
, for taking me on as his student for the summer and making this
work possible, as well as his graduate student,
Mike Klopfer
, for his constant and ready assist
ance
along the way.
Linda Bugge

was a better REU coordinator than anyone could have asked for, and
Figure 17: a 120 nm silica
nanosphere decorated with
small gold particles
(Oldenburg, 1998).

Morrill
19

Dr. Marek O
sinski

deserves high praise for bringing the REU in Nanophotonics program to the
Center for High Tech Materials at UNM.
Through opening her lab and offering her expertise,
Dr. Bernadette Hernandez
-
Sanchez

made the experimental chemistry possible, while her group


Doug,
Matt, Allison, Lee, Danny, Cory, Sarah
(who prepared exceptional TEM images), and
Lesly



were always rea
dy with a helping hand.
Dr. Tim Boyle
graciously offered his equipment at
Sandia’s Advanced Materials Lab
. Finally,
Aaron So
offered great assistance not just in the lab
but in thinking through many challenging problems.

References

[Bunge, 2003]: Bunge
, S.D., Krueger, K.M., Boyle, T.J., Rodriguez, M.A., Headleya, T.J., Colvin,


V.L.,
J. Mater. Chem.
2003
, 13, 1705
-
1709.


[Diaspro, 2006]: Diaspro, A., Bianchini, P., Vicidomini, G., Faretta, M., Ramoino, P., Usai, C.,
BioMed. Eng.


OnLine

2006
, 5, 36.


[Guerrero
-
Martinez, 2010]: Guerrero
-
Martinez, A., Perez
-
Juste, J., Liz
-
Marzan, L.M.,
Adv. Mat.

2010
, 22,


1182
-
1195.


[Hu, 2009]: Hu, X., Zrazhevskiy, P., Gao, X.,
Ann. of Biomed. Eng.
2009
, 37 (10), 1960
-
1966.


[Jin, 2009]: Jin, Y., Gao, X.
Nature Nan
otechnology
2009
, 4, 571.


[Kable Group, web]: http://sydney.ed
u.au/science/chemistry/~kable_s/
projects/photon/index.html


[Kelly, 2003]: Kelly, K.L., Coronado, E., Zhao, L.L., Schatz, G.C.
J. Phys. Chem. B.
2003
, 107, 668
-
677.


[Klopfer, 2011]: Klopfer,

M., and Jain, R.K.,
Opt. Mater. Express

2011
, 1, 1353.


[Larson, 2003]:
Larson, D. R; Zipfel, W. R.; Williams, R. M.; Clark, S. W.; Bruchez, M. P.; Wise, F. W., and


Webb W. W.,
Science

2003
,

300
,

1434.

[Norton, 2011]: Norton, S.J., Vo
-
Dinh, T.,
IEEE Trans. On Nanotech.
2011
, 10 (6), 1264.


[Oldenburg, 1998]: Oldenburg, S.J., Averitt, R.D., Westcott, S.L., Halas, N.J.
Chem. Phys. Lett.
1998
, 288, 243
-


247.


[Pines, 1956]: Pines, D.
Rev. of Mod. Phys.
1956
, 28, 184
-
198


[Ru, 2008]: Ru, E.L.,
Etchegoin, P.E.
Principles of Surface Enhanced Raman Spectroscopy: and related plasmonic effects



(Elsevier Science, 2008), 1
st

Ed., Web.


[Schneider, 2006]: Schneider, G., Decher, G.,
Nano. Lett.
2006
, 6 (3), 530
-
536.


[Yaghini, 2009]: Yaghini, E.; Sei
falian, A. M., and MacRobert, A. J.,
Nanomedicine

2009
,
4,

353.

Morrill
20

Appendix:
Synthesis of Plasmonic Quantum Dots


Employing a Mesoporous Silica Spacer: for 20 nm silica shell thickness (silica coating adapted from
Hu, Zrazhevskiy, and Gao,
Annals of
Biomedical Engineering
, 2009, and gold coating adapted from Jin
and Gao,
Nature Nanotechnology,
2009).

Updated: 2 Aug 2012

Spacer growth

1.

Add

5

mL cetyltrimethylammonium bromide (CTAB) aqueous solution (55 mM)

to 100 mL
roundbottom

flask on heating mantel with stir bar.

2.

Add 0.5 mL of 3.0 micromolar
CdSe/ZnS QD

in chloroform solution to flask.

3.

Stir vigorously to obtain a homogenous microemulsion

(should appear milky); keep stirring on for at
least 10 min.

4.

Heat mixture at 50°C for
approximately 15 minutes to evaporate the chloroform (this should result
in a clear aqueous solution of CTAB stabilized quantum dots
).

a.

Some suds may appear; this is due to the present of CTAB (a soap).

b.

At this point, all organic solvents should be out of the solution; this is why it is critical that
no TOPO remain in solution after QD surface prep, as the TOPO will not evaporate at
these low temperatures.

5.

Dilute the mixture with 45

mL NaOH

(13 mM)

solu
tion preheated to 50°C

6.

Follow immediately with the addition of 0.5 mL tetr
aethylorthosilicate (TEOS) and 3

mL of ethyl

acetate.


Surface preparation for gold growth

7.

Keep the reaction mixture stirring

at
50°C for 3 h
ours. After the 3 hour period has passe
d,
allow
the
reaction mixture
to slowly cool down to room temperature.

8.

As soon as the flask is at room temperature, r
inse the silica encapsulated quantum dots repeatedly
with ethanol to remove the unreacted precursors and surfactants.

a.

Note that the longe
r the CTAB stays in the silica as a porogen, the more the fluorescence of
the silica coated QD’s is quenched (by 48 hours, may be totally quenched)

b.

Check fluorescence of reaction flask with UV lamp before and after emptying for washing

c.

Wash very thoroughly
; do at least 3 spin cycles (finished product should not form soapy
bubbles when shaken, as all CTAB should be out)

9.

During the final rinse, do not add additional ethanol to the pellet; instead, allow ethanol to evaporate
(if necessary, heat wet pellet to
80°C and let sit until dry
; can be done in drying oven
).

10.

Redisperse nanoparticles in 10

mL of DI water.

a.

These are now stable, mesoporous silica coated quantum dots

b.

Check fluorescence with UV lamp
-

a good thing to periodically do.

11.

Remove 200 microliters
o
f dispersed NP’s and transfer to a small
vial filled previously with 800
microliters of DI water
.

a.

Assuming a 50% yield at this point, this quantity should roughly correspond to 15 picomoles
of QD’s.

12.

Add 2

mg of
poly
-
L
-
histidine hydrochloride

(molecular
weight greater than 5,000)
and gently mix in
to flask. Allow the NP’s and the PLH to incubate at room temperature for ~1 hr to allow for
electrostatic adsorption of the PLH.

a.

This may be an excessive quantity of PLH. It would be useful to experiment with h
igher
QD/PLH ratios (this should correspond to about 27,000 PLH molecules per silica QD).

Morrill
21

b.

Note that the surface area of the silica particle is at least 10 times greater than the particles
Gao used in his synthesis, so the amount of PLH per QD was stepped
up in this procedure
from the amount he used.

13.

Remov
e excess PLH by centrifugation,
leaving behind a PLH functionalized pellet ready for gold
growth.

a.

Gao used ultracentrifugation (30,000 RPM) at this stage, so rapid centrifugation should be
fine, however
it is worth considering whether the high speed centrifugation strips PLH from
the silica surface.

14.

Disperse pellet in
1 mL of DI water. Check fluorescence.


An alternative approach to surface preparation using
MPS

(begin af
t
er step 6)
:

7.

After 10 minutes, add 50 microliters of
mercaptopropyltris(methyloxy)silane (MPS)

8.

Keep the reaction mixture stirring for 3 hours and allow to slowly cool down to room temperature.

9.

Rinse the silica encapsulated quantum dots repeatedly with ethanol to remo
ve the unreacted
precursors and surfactants.

10.

During the final rinse, do not add additional ethanol to the pellet; instead, allow ethanol to evaporate
(if necessary, heat wet pellet to 80°C and let sit until dry).

11.

Redisperse pellet in about 5 mL DI water.
Check fluorescence.


Gold growth

12.

Prepare HAuCl
4

aqueous stock solution (w/w 1%); 10 mL is a good amount, although much less is
needed for the synthesis.

13.

Prepare approximately 1
0

mL of 20 millimolar NH
2
OH s
tock s
olution with DI water
. Once again,
much l
ess is needed, but its good to have a stock solution on hand. This is the reducing agent.

14.


Prepare a solution of 50 microliters of the chloroauric acid solution brought to a total volume of
1800 microliters with a solution of sodium hydroxide and DI water
. Add an appropriate amount of
NaOH to bring this solution to pH 9
-
10.

a.

If experimenting with different amounts of gold for different thickness of coatings, try
preparing multiple vials with 10, 25, 50, 100, etc microliters of chloroauric acid stock
soluti
on.

15.

Remove 900 microliters of NaOH/
HAuCl
4

aqueous solution and add to a small vial. To this vial add
100 microliters of surface prepared silica coated quantum dots.

16.

Allow mixture
s

to incubate for ~ 10 min.

17.

Add a volume of
NH
2
OH
stock solution equivalent

to the volume of chloroauric acid solution added
in step 14 to the NaOH/
HAuCl
4
/silica QD aqueous solution from step 15. The volume of reducing
agent used should be double the volume of chloroauric acid stock solution present in the solution
from step 15.

a.

Upon mixing, the colloidal solution may change from
colorless to brownish blue. This may
indicate the formation of plasmonic nanostructures (either shells or standalone particles)

18.

Gold/silica/QD nanoparticles can be stabilized with mPEG
-
thiol

a.

Note that t
his step was not performed by the author, but he would recommend it. An
appropriate amount may be about 0.2 mg, but this is not a verified quantity.