Nadine Kammerlander
Optical Properties of metallized DNA
ISBN: 978-3-8366-2212-7
Herstellung: Diplomica® Verlag GmbH, Hamburg, 2009
Zugl. Technische Universität München, München, Deutschland, Diplomarbeit, 2007
Dieses Werk ist urheberrechtlich geschützt. Die dadurch begründeten Rechte,
insbesondere die der Übersetzung, des Nachdrucks, des Vortrags, der Entnahme von
Abbildungen und Tabellen, der Funksendung, der Mikroverfilmung oder der
Vervielfältigung auf anderen Wegen und der Speicherung in Datenverarbeitungsanlagen,
bleiben, auch bei nur auszugsweiser Verwertung, vorbehalten. Eine Vervielfältigung
dieses Werkes oder von Teilen dieses Werkes ist auch im Einzelfall nur in den Grenzen
der gesetzlichen Bestimmungen des Urheberrechtsgesetzes der Bundesrepublik
Deutschland in der jeweils geltenden Fassung zulässig. Sie ist grundsätzlich
vergütungspflichtig. Zuwiderhandlungen unterliegen den Strafbestimmungen des
Urheberrechtes.
Die Wiedergabe von Gebrauchsnamen, Handelsnamen, Warenbezeichnungen usw. in
diesem Werk berechtigt auch ohne besondere Kennzeichnung nicht zu der Annahme,
dass solche Namen im Sinne der Warenzeichen- und Markenschutz-Gesetzgebung als frei
zu betrachten wären und daher von jedermann benutzt werden dürften.
Die Informationen in diesem Werk wurden mit Sorgfalt erarbeitet. Dennoch können
Fehler nicht vollständig ausgeschlossen werden und der Verlag, die Autoren oder
Übersetzer übernehmen keine juristische Verantwortung oder irgendeine Haftung für evtl.
verbliebene fehlerhafte Angaben und deren Folgen.
© Diplomica Verlag GmbH
http://www.diplomica.de, Hamburg 2009
Abstract
Metallic nanoparticles have been studied intensively during the last decades because of
their intriguing optical properties: Due to collective oscillations of the conducting elec-
trons - the so called plasmonic oscillations - they absorb light in the visible spectrum. The
resonance frequency thereby sensitively depends on parameters such as the particle size
and shape as well as the dielectric constant of the medium. DNA exhibits outstanding
recognition properties and can be modified easily. Thus, template-directed material syn-
thesis along synthetic DNA is a promising route to grow nanoparticles of defined shape
and size and with defined interparticle-spacing.
In this study, two different methods are used to deposit silver on oligonucleotides of
different lengths, ranging from 23 to 96 basepairs, in order to synthesize metallic nanorods
of controlled aspect ratios. The first method involves the specific labeling of nucleotides
with aldehyde groups, followed by exposure to a Tollens reagent and a developer. The
second method relies on the photoinduced deposition of silver onto unmodified DNA
samples. Several preparation parameters such as the DNA sequence, buffer salt type,
silver concentration and UV illumination time are varied systematically.
The metallized DNA molecules are characterized concerning their optical and structural
properties. Absorption spectra show plasmon peaks around 420nm. Peak positions, inten-
sities and bandwidths are analyzed. Dynamic Light Scattering studies in solution provide
information about the particle sizes as well as their structural asymmetry. Both opti-
cal techniques are used to observe the temporal evolution of the nanoparticle growth in
the Tollens metallization process. Structural information is inferred from Atomic Force
Microscopy; for that purpose, the particles are deposited on single-crystalline silicon sub-
strates.
i
Zusammenfassung
Metallische Nanopartikel werden seit einigen Jahrzehnten intensiv erforscht, unter an-
derem, weil sie sich durch besondere optische Eigenschaften, sogennante Plasmonenreso-
nanzen, auszeichnen: Bei Einfall von Licht im sichtbaren Bereich zeigen die Teilchen starke
Absorption, welche durch kollektive Schwingungen der Leitungsbandelektronen bedingt
ist. Die Wellenl¨
ange dieser Plasmonenresonanz ist abh¨
angig von diversen Parametern,
darunter Eigenschaften der Nanopartikel, wie Gr¨
oße und Form, sowie die dielektrischen
Eigenschaften des umgebenden Mediums. Der Einsatz von synthetischer DNS als Vorlage,
entlang derer Nanopartikel wachsen, ist sehr vielversprechend: Aufgrund ihrer Selbsterken-
nungseigenschaften sowie der vorhandenen M¨
oglichkeit, die DNS zu modifizieren, k¨
onnen
zum einen die Gr¨
oße und Geometrie der Teilchen vorgegeben werden. Zum anderen k¨
onnen
die Nanopartikel in einem definierten Abstand zueinander erzeugt werden.
In dieser Arbeit werden zwei verschiedene Methoden angewendet um Silber an der DNS
anzulagern. Um metallische Nanost¨
abchen mit kontrollierten Achsenverh¨
altnissen herzu-
stellen, werden verschiedene DNS-L¨
angen im Bereich von 23 bis 96 Basenpaaren verwen-
det. F¨
ur die erste Metallisierungsmethode werden an bestimmte Nukleotide Aldehydgrup-
pen angebunden. Anschließend wird zur L¨
osung ein Tollensreagenz sowie eine Entwick-
lerl¨
osung zugegeben. Die zweite Metallisierungsmethode besteht in einer UV-induzierten
Reduktion von Silberionen an unmodifizierter DNS. Bei der Synthese werden verschiedene
Parameter, darunter die DNS-Sequenz, die Pufferl¨
osung, die zugegebene Silberkonzentra-
tion sowie die Bestrahlungsdauer systematisch variiert.
Die in L¨
osung metallisierte DNS wird anschließend bez¨
uglich ihrer optischen und struk-
turellen Eigenschaften charakterisiert: Absorptionsspektren zeigen die f¨
ur Silbernanopar-
tikel typischen Maxima bei 420nm, die durch die Plasmonenoszillationen hervorgerufen
werden. F¨
ur die verschiedenen Proben werden die Position des Maximums, seine In-
tensit¨
at sowie die Halbwertsbreite bestimmt. Messungen der dynamischen Lichtstreuung
geben Aufschluss ¨
uber die Teilchengr¨
oßen in der L¨
osung und ihre H¨
aufigkeit. Polarisierte
Lichtstreuungsmessungen geben den Grad der Asymmetrie f¨
ur die Teilchen an. Beide
optischen Charakterisierungsmethoden werden verwendet um die zeitliche Entwicklung
des Teilchenwachstums bei der Tollens-Metallisierung zu beobachten. Aus Untersuchun-
gen mit einem Rasterkraftmikroskop lassen sich Informationen ¨
uber die Struktur der
metallisierten DNS-Molek¨
ule ableiten; daf¨
ur werden die Teilchen auf ein einkristallines
Siliziumsubstrat aufgebracht.
ii
Contents
1
Motivation
1
2
Fundamentals
3
2.1
Metallic Nanoparticles . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
3
2.2
Optical Properties of Nanospheres . . . . . . . . . . . . . . . . . . . . . . .
4
2.2.1
Extinction of Light . . . . . . . . . . . . . . . . . . . . . . . . . . .
4
2.3
Optical Properties of Nanorods . . . . . . . . . . . . . . . . . . . . . . . .
8
2.3.1
Electrostatic Calculations
. . . . . . . . . . . . . . . . . . . . . . .
9
2.4
Deoxyribonucleic Acid . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
11
2.4.1
Composition and Properties . . . . . . . . . . . . . . . . . . . . . .
11
2.4.2
Metallized DNA . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
12
3
Synthesis of Metallized DNA
13
3.1
Hybridization of DNA
. . . . . . . . . . . . . . . . . . . . . . . . . . . . .
13
3.2
Requirements for Metallization
. . . . . . . . . . . . . . . . . . . . . . . .
15
3.3
Photochemical Metallization . . . . . . . . . . . . . . . . . . . . . . . . . .
16
3.4
Tollens Metallization of Aldehyde-Modified DNA
. . . . . . . . . . . . . .
16
3.5
Materials Section . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
18
4
Absorption Spectrometry
19
4.1
Principle of Absorption Spectrometry . . . . . . . . . . . . . . . . . . . . .
19
4.2
Absorption Measurements of UV-Metallized Samples . . . . . . . . . . . .
20
4.2.1
Influence of pH . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
21
iii
CONTENTS
4.2.2
DNA Sequence . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
22
4.2.3
Silver Amount . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
25
4.2.4
Irradiation Time
. . . . . . . . . . . . . . . . . . . . . . . . . . . .
27
4.2.5
DNA Length
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
29
4.2.6
Reference Samples
. . . . . . . . . . . . . . . . . . . . . . . . . . .
30
4.2.7
Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
31
4.3
Absorption Measurements of Tollens-Metallized Samples . . . . . . . . . .
37
4.3.1
Absorption Measurements of Tollens-Metallized dsDNA . . . . . . .
38
4.3.2
Absorption Measurements of Tollens-Metallized ssDNA . . . . . . .
42
4.3.3
Development Solution
. . . . . . . . . . . . . . . . . . . . . . . . .
43
4.3.4
Data Analysis and Discussion . . . . . . . . . . . . . . . . . . . . .
45
5
Dynamic Light Scattering
49
5.1
Principle of Dynamic Light Scattering
. . . . . . . . . . . . . . . . . . . .
49
5.1.1
Measurement Technique . . . . . . . . . . . . . . . . . . . . . . . .
49
5.1.2
Fundamentals . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
49
5.1.3
Polarized Dynamic Light Scattering . . . . . . . . . . . . . . . . . .
52
5.1.4
Apparatus . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
53
5.2
Unpolarized Measurements . . . . . . . . . . . . . . . . . . . . . . . . . . .
53
5.2.1
Experimental Results . . . . . . . . . . . . . . . . . . . . . . . . . .
53
5.2.2
CONTIN Analysis
. . . . . . . . . . . . . . . . . . . . . . . . . . .
55
5.2.3
Temporal Evolution . . . . . . . . . . . . . . . . . . . . . . . . . . .
60
5.2.4
Reference Samples
. . . . . . . . . . . . . . . . . . . . . . . . . . .
61
5.3
Polarized Measurements . . . . . . . . . . . . . . . . . . . . . . . . . . . .
63
5.3.1
Simulations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
63
5.3.2
Experimental Results . . . . . . . . . . . . . . . . . . . . . . . . . .
65
6
Stabilization of the Nanoparticles
69
6.1
Fundamentals . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
69
iv
CONTENTS
6.2
Experimental Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
70
7
Nanorings
74
8
Atomic Force Microscopy
79
8.1
Fundamentals . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
79
8.2
Sample Preparation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
79
8.2.1
Requirements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
80
8.2.2
Functionalization of Silicon Wafers . . . . . . . . . . . . . . . . . .
82
8.3
Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
83
8.3.1
Ultrasonication . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
83
8.3.2
MPTMS-Silanization . . . . . . . . . . . . . . . . . . . . . . . . . .
84
8.3.3
Spin-Coating
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
88
9
Conclusion
89
10 Outlook
91
A DNA Sequences
93
B Reference Samples
94
B.1 Absorption Spectroscopy . . . . . . . . . . . . . . . . . . . . . . . . . . . .
94
B.2 DLS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
96
B.3 AFM . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
97
C Parameters of MPTMS-Silanization
98
D Acknowledgments
100
v
Chapter 1
Motivation
Modern technologies cannot be imagined without electronic devices, as e.g. the success
stories of transistors and diodes illustrate. The performance of these electronic devices has
tremendously increased during the last decades through a process known as downscaling:
Minimizing structural units allows for extremely high densities of electronic modules. This
entails an increase of speed as well as an enhanced performance. Today's photonic devices,
however, are still rather large. The reason therefore can be found in the nature of light:
Diffraction sets a lower limit for the scales of photonic devices and therefore prevents
densely packed photonic components. The diffraction limit can be, however, overcome by
so called 'Surface Plasmon Polaritons' (SPPs): These are oscillations of the conducting
electrons in a metallic nanoparticle induced by an incoming electromagnetic field with a
suitable frequency. By this conversion of the optical mode to surface plasmons, the elec-
tromagnetic energy can be localized to regions of merely several tens of nanometers. This
property is highly desirable for the construction of future optoelectronic devices: Chains
of metallic nanoparticles, for example, are promising waveguides with lateral confinements
below the diffraction limit.
DNA is a potential pattern material for constructing nanostructures as it can be synthe-
sized in almost arbitrary lengths in the nanometer scale and also allows for modifications
that make site-specific reactions possible. Its inherent scaling limit is determined by the
distance of two neighbored basepairs: This length is only 0.34nm and therefore much be-
low the diffraction limit. Moreover, it seems very appealing to exploit the self-recognition
properties of DNA to build functional nanostructures by self-assembly processes. Among
all possible template-directed material syntheses, metallization of DNA is particularly in-
teresting for photonic devices as it makes a controlled growth of metal nanoparticles with
defined size and shape along the DNA template possible. Furthermore the use of DNA
as a pattern promises to achieve a defined arrangement of the particles relative to each
other.
Experiments concerning metallization of DNA have been done for approximately a decade
now. Most experiments, however, have concentrated on investigating the electrical con-
ductance of the DNA. Absorption spectra have merely been recorded for verifying the
1
CHAPTER 1. MOTIVATION
existence of metallized DNA in the solution. The optical properties of metallic nanopar-
ticles grown on DNA templates have remained almost completely unexplored up to now.
Thus it is the main topic of this thesis to investigate the optical properties of metal-
lized DNA structures, i.e. their absorption as well as scattering characteristics. Therefore
two different metallization methods are applied and the resulting metallized particles are
compared. Varying synthesis parameters, such as concentrations of the reagent or the re-
action time systematically and analyzing the optical data, provides an understanding on
what variables the metallization depends on. Changes in optical properties are related to
changes in the size or structure of the nanoparticles. This information is extracted from po-
larized Dynamic Light Scattering experiments as well as Atomic Force Microscopy. More-
over Dynamic Light Scattering allows for direct monitoring of the nanoparticle growth;
thereby the particle sizes as well as intensities are observed. These results are necessary
to understand the DNA metallization process in detail and to be able to optimize the
synthesis conditions on the way to a defined growth of metallic nanoparticles. Using var-
ious seed densities for site-specific DNA metallization and analyzing the resulting optical
properties is a fist step to a defined arrangement of nanoparticles.
This thesis is structured as follows: After a brief introduction into theory about the prop-
erties of metallic nanoparticles and about DNA in chapter 2, chapter 3 explains the
two different methods of metallization that are used for synthesis. Chapter 4 addresses
absorption spectrometry: After a short description of the working principle, absorption
spectra of metallized DNA samples are shown for numerous varied parameters and ana-
lyzed. The results are finally discussed and compared with two different theories about
nanoparticle absorption. Chapter 5 is about Dynamic Light Scattering: In a theoretical
part, simulations are done for a later comparison with experimental data. The second half
of the chapter presents first results for the various metallized DNA samples. In chapter 6
the necessity for stabilizers and their working principle is explained before some experi-
mental data is shown. Chapter 6 handles with metallization of ring-like DNA structures.
In chapter 8 micrographs recorded with an Atomic Force Microscope are presented and
analyzed.Afterward the obtained results of this thesis are concluded and an outlook for
future investigations is given.
2
Chapter 2
Fundamentals
2.1
Metallic Nanoparticles
Metallic nanoparticles have been used for more than a millennium, e.g. gold nanoparticles
were utilized for staining glass. 150 years ago Lord Faraday was one of the first to do
scientific work about metallic colloidal solutions. He recognized that the color of the
colloids originates from the small size of the particles.
Nowadays it is possible to produce nanoparticles in defined sizes and to a certain ex-
tent also with defined shapes. Important experiments have been done by Link, El Sayed
[LES99], and J. J. Mock et al. [MBS
+
02]. Transmission electron microscopy (TEM) im-
ages of pentagon- and triangle-shaped nanoparticles produced by Mock are shown in fig.
2.1.
Figure 2.1: TEM images of nanoparticles with a pentagon and a triangle shape, respectively.
This work was done by Mock in 2002, [MBS
+
02]
Production processes range from wet chemical synthesis to e-beam lithography. Whereas
the first method provides for highly monocrystalline particles, the latter allows for an
defined arrangement of the particles on a surface.
Fields of application are widely spread and include medicine as well as materials science.
Metallic nanoparticles are promising nanosensors due to their optical properties, the so
called 'surface plasmons', which will be explained later in 2.2.
As the name already tells, nanoparticles have sizes between a few and hundreds of nanome-
ters. The clusters typically consist of several up to thousands of atoms and differ a lot in
3
CHAPTER 2. FUNDAMENTALS
their properties from the bulk material.
One reason are surface effects, which become much more important as a larger percentage
of atoms is on the surface. For very small particles, e.g. silver particles with a diameter
smaller than about 52nm [SGHURS
+
05] the mean free path of the electrons is limited
mainly by collision with the surface. Quantum effects due to the confinement of space,
however, do not play an important role for the nanoparticles regarded here: The splitting
of energy levels can be approximated as E =
E
F
N
with the Fermi-energy E
F
=5.49eV
[AN05] and N the number of levels in the band [KV83]. To see quantum effects, the energy
splitting must be larger than the thermal energy k
B
T , 25meV at room temperature. This
is only fulfilled for clusters containing less than approximately 200 atoms.
2.2
Optical Properties of Nanospheres
As mentioned above metallic nanoparticles are characterized by particular optical prop-
erties. In the following an overview about the underlying physical fundamentals is given:
2.2.1
Extinction of Light
When light transverses a medium other than vacuum, dipoles are induced due to the
electromagnetic field. The nature of these dipoles depends on the medium and can be
either molecular, atomic or the displacement of the free electron gas in a conductor. The
'polarizability' indicates the dipole strength in dependence of the incoming field. It is
a frequency dependent material constant, that determines the response strength of the
material to an incoming field.
The induced dipoles themselves emit so called 'secondary radiation' in the toroidal radi-
ation pattern that is typical for dipoles. On a macroscopic scale, this secondary radiation
appears as scattered light. It is also possible that the medium releases the energy taken up
from the electromagnetic wave in a non-radiant way; this is absorption. Hence absorption
and scattering are not independent of each other. The sum of absorption and scattering
is called 'extinction'.
In electrodynamics all optical properties are characterized by the complex dielectric func-
tion , that depends on the medium as well as the frequency of the incoming field. It
describes Starting from this function, one can calculate further important physical values
such as the refractive index, the transmission or the absorption.
Scattering of Light
As mentioned before, the induced dipoles in the medium act as secondary scatterers for
an incoming electromagnetic wave. The interference of this scattered light determines the
4
2.2. OPTICAL PROPERTIES OF NANOSPHERES
Figure 2.2: The real and imaginary part of the dielectric function of bulk silver according to
[JC72]
angle-dependence of the outcoming light. Very small particles with a radius smaller than
/10 scatter unpolarized light isotropically as all the secondary waves are approximately
in phase. The intensity of the scattered light can be approximated as I
d
6
4
, the so called
Rayleigh approximation [BH83]. Huge particles larger than roughly 10, however, scatter
mainly in the forward direction. The scattered intensity is inversely proportional to the
scattering angle.
Absorption of Light by Plasmonic Oscillations
The Rayleigh formula in the previous subsection has already shown that scattering in-
creases with the particle size. For small nanoparticles absorption therefore predominates.
Absorption of light can be traced back to atomic or molecular transitions or to collective
phenomena. One of these - the plasmonic oscillations - is described in more detail as it is
important for nanoparticle analysis.
Plasmonic Oscillations
When the free electrons of a metallic conduction band are moved away from their equi-
librium position, the electromagnetic attraction between the negatively charged electrons
and the positively charged atomic cores counteracts the external force of the electro-
magnetic wave. This is the situation of a damped, driven harmonic oscillator. When the
external frequency matches the eigenfrequency of the electron system, the amplitude of
the electron movement increases, the electrons undergo so called plasmonic oscillations
and energy is absorbed. In general there are several methods for releasing the energy, such
as single electron excitation or heating up.
In bulk material, the dielectric function according to the Drude theory is () = 1-
2
p
2
-i
.
This term includes the damping constant =
v
F
l
0
with v
F
the Fermi velocity and l
0
the free
5
CHAPTER 2. FUNDAMENTALS
Figure 2.3: An incoming electromagnetic wave, whose electric field vectors are indicated with
black arrows, moves the electronic cloud of the metallic spheres out of equilibrium. The positively
charged atomic cores stay at their original position and therefore an attractive Coulomb force
acts on the electronic cloud. This induces oscillations. When the frequency of the incoming light
matches the eigenfrequency of the particles, light can be absorbed. This can be seen as a so
called plasmon peak in the absorption spectrum.
path length [AN05].
P
=
N e
2
0
m
e
is the plasmon frequency, that occurs at a wavelength
of 138nm for silver. Light with a wavelength longer than the plasmon peak wavelength is
reflected which leads the shiny glance that is typical for metals. These volume plasmon
waves are - like all oscillations in free gases - uniquely longitudinal waves and therefore
they cannot be excited by light, which is merely transverse. Thus, although plasmonic
oscillations are an important optical property of the bulk material, no incoming light can
be absorbed by exciting these plasmons. The dielectric function of bulk silver, determining
its optical properties, can be seen in fig. 2.2. Despite the plasmon wavelength, which
separates the region of reflection and transmission/absorption, there is also an important
feature at 318nm: At this wavelength the interband transitions set in.
Surface Plasmons
For surface plasmons the influence of the induced dipoles, that are located on the surface
of the particle and undergo oscillations, is important. Reflections of the electron cloud
on the surface change the boundary conditions of the problem: Shear forces induced by
such reflections entail a transverse component of the plasmon oscillation. Thus the surface
plasmons can, in contrast to the volume plasmons, indeed, be excited by light. For bulk
material, surface plasmons are negligible due to the small percentage of surface atoms. But
they gain importance with decreasing medium size and dominate the optical properties
of metallic nanoparticles. Fig. 2.3 shows schematically how plasmonic oscillations are
induced in a small metallic sphere.
6
2.2. OPTICAL PROPERTIES OF NANOSPHERES
The task to calculate the absorption of metallic nanoparticles is an electromagnetic one
and can therefore be done by solving the Maxwell equations in matter. Although this
is possible in general, there are indeed only few cases were exact solutions are known.
Mie solved the problem for spheres by expanding the incoming, internal and scattered
electromagnetic waves in an infinite series of vector spherical harmonics. These spherical
harmonics for the electric and the magnetic field are set in the Maxwell equations. The
boundary conditions determine the scattering coefficients for the specific problem. For
particles, not smaller than about 15nm, the size-independent dielectric function of the
bulk material can be used. A detailed derivation of the Mie theory is given by Bohren
and Huffman [BH83].
Figure 2.4: Plasmon peaks for commercial silver nanospheres with pM concentrations measured
with absorption spectroscopy. The spectra show that with increasing size, the peak shifts to the
red and becomes broader. The graph for r=40nm shows a quadrupole mode at 350nm.
Electrostatic Approximation
For small spherical particles the electrostatic approximation (also known as 'quasistatic
approximation') can be applied. The extinction cross section is thereby expressed by
dip
ext
R
3
3/2
m
Im
()
(
Re
() + 2
m
)
2
+
Im
()
2
with R radius of the particle,
m
dielectric constant of the medium,
Im
and
Re
imaginary
and real part of the frequency dependent dielectric function of the nanoparticle respec-
tively. This electrostatic approximation estimates the position of the plasmon peak. It
shows that the intensity of the peak increases by the power of three with the particle
radius (i.e. it is proportional to the particle volume). The peak position is determined by
7
CHAPTER 2. FUNDAMENTALS
the zero of the denominator, i.e. by the equation
Re
() = -2
m
. Thus the approximation
only depends on the material and the environment, but neglects the influence of size on
the exact peak position. The electrostatic approximation only accounts for the first terms
in the series of harmonics, the electrical dipole term. For particles with a radius larger
than 40nm, the quadrupole term is not negligible any more. An additional peak at a
higher energy occurs.
Although Mie theory is only valid for spherical particles and neglects any interaction
among the individual particles, it provides an understanding about what conditions the
plasmon peak position and shape depends on: The dielectric functions of the particle and
the medium are very important. There is also a dependence on the size: For small particles
the dielectric function itself is size-dependent. Theories about the effect of this intrinsic
feature differ. Larger particles show a red-shift with increasing size due to retardation
effects. Due to various damping-mechanisms, the peak width also depends on the particle
size. The width as well as the asymmetry of the peak are also influenced by the particle size
distribution when considering an ensemble of particles. Extinction spectra of commercially
available silver nanospheres provided by BBI/Tedpella Inc.
1
, Redding, CA is shown in fig.
2.4.
2.3
Optical Properties of Nanorods
Theory presented so far applies only to spherical particles. For non-spherical particles the
position as well as the number of absorption peaks in the spectrum changes. Ellipsoids
are the simplest case of non-spherical particles and there exists an extension of the Mie
theory to treat them. For ellipsoidal particles two cases can be imagined: Depending on the
relative orientation of the incoming field and the particle, a displacement of the electrons
along the short or the long axis of the particle is induced. In the first case one speaks of
'transverse plasmons', in the second case one speaks of 'longitudinal plasmons', its energy
is lower due to a longer wavelength. Fig. 2.5 schematically shows such a metallic nanorod
with the length L and the diameter d. For a collection of randomly orientated ellipsoids
two absorption peaks are expected in the spectrum corresponding to the longitudinal and
transversal plasmonic oscillation, respectively.
Figure 2.5: Schematic of a metallic nanorod with length L and diameter d.
1
Ag nanospheres were produced by reduction of silver citrate and therefore have a negative surface
charge. Depending on their size, their concentration lies between 20pM and 1.2nM.
8
2.3. OPTICAL PROPERTIES OF NANORODS
2.3.1
Electrostatic Calculations
To get a first idea about the plasmon peak positions of silver nanorods, I have done an
electrostatic calculation. This was based on Mie's electrostatic approximation for pro-
late spheroids [BH83]. For an electromagnetic field parallel to the principal axis i, the
absorption coefficient is thereby given as
i
() =
0
() -
m
m
+ [ () -
m
]P
i
V
whereas
0
is the dielectric constant of the vacuum,
m
is the dielectric constant of the
embedding medium, () is the (complex) dielectric function of silver [JC72] and V is
the volume of the nanoparticles. P
i
are the geometrical depolarization factors, which are
given by the nanoparticle shape. Details can be found in [BH83]. They follow the sum
rule
i
P
i
= 1. For prolate ellipsoids the geometrical factor for one of the (identical) short
axes is
P
i
=
1 - e
2
e
2
-1 +
1
2e
ln
1 + e
1 - e
with the eccentricity
e
2
= 1 -
b
2
a
2
whereby a is the long and b is the short axis of the ellipsoid.
For
m
was set to 1.77 which equals the dielectric constant of pure water in the visible
spectrum
2
.
For randomly orientated ellipsoids only the average cross sections for absorption and
scattering, that are independent of the polarization of the incoming light, are of interest
[BH83]:
C
abs
= kIm
1
3
1
+
1
3
2
+
1
3
3
C
sca
=
k
4
6
1
3
|
1
|
2
+
1
3
|
2
|
2
+
1
3
|
3
|
2
These cross sections were calculated and are shown in fig. 2.6. The formulas already show
that maxima occur when the denominator of the absorption coefficient of any principal
axis becomes zero.
According to the calculation two peaks appear, that shift relatively to each other. This
shift, that is more pronounced for the longitudinal plasmon, is also visualized in fig. 2.7.
2
The value
m
= 80 that is frequently quoted for water is only valid for low frequencies. At about
10GHz the dielectric constant starts to decrease, as it is shown for example in [Fli00]
9
CHAPTER 2. FUNDAMENTALS
Figure 2.6: Electrostatic calculation of the plasmonic absorption of ellipsoids. Two peaks are
visible whereby the long-wavelength peak shifts to the red with increasing axis ratio and becomes
more intense.
Figure 2.7: According to electrostatic calculations the plasmon peaks are centered around 400nm.
The long-wavelength peak significantly shifts to the red with increasing non-sphericity whereas
the short-wavelength peak undergoes only a slight shift to the blue.
10
2.4. DEOXYRIBONUCLEIC ACID
2.4
Deoxyribonucleic Acid
Deoxyribonucleic acid (DNA) is not only important for biological matters but can be also
used for template-directed material synthesis. Thereby DNA is chosen as a pattern as it
can be easily synthesized with arbitrary length and sequence by solid state synthesis and
can even build up complex networks.
2.4.1
Composition and Properties
DNA is composed of a negatively charged phosphate backbone, pentose sugar molecules
as well as the four bases adenine (A), guanine (G), cytosine (C) and thymine (T) whose
sequence determine the code of the DNA. Natural DNA occurs in the so called 'B-form',
a right-handed rotated double helix with a base-to-base distance along the axis of 3.4nm.
For undergoing hybridization, i.e. forming a double helix, the bases have to match: A
and T are complementary bases, which can be bound via two hydrogen bonds. Another
complementary base pair is G and C, whose bond strength is slightly stronger due to
three hydrogen bonds. A schematical configuration of DNA is shown in fig. 2.8.
Figure 2.8: Configuration of double stranded DNA consisting of a phosphate backbone, sugars
and bases [Wik06]
Because of this hybridization, that takes place under certain conditions discussed in
[Man78], DNA possesses self-assembly properties that can be exploited for many ap-
plications in nanotechnology.
11
CHAPTER 2. FUNDAMENTALS
2.4.2
Metallized DNA
One possible application in nanotechnology is the growth of metallic nanoparticles on a
DNA pattern. First DNA metallization experiments were done by the group of Braun et
Keren [BESBY98]. They started coating DNA with metals by several chemical reactions
to provide for conductive nanowires. According to Eichhorn et al. [ES68] silver ions, albeit
they prefer binding to the bases instead of the backbone, do not unwind the double helix
and therefore do not destabilize the DNA. Various methods and metals (Pt, Ag, ...) have
been used by many groups working in this field. Richter presents a good summary about
experiments and results in this research area in his review paper [Ric03]. Most experi-
ments concentrate on µm-long DNA, e.g. -phage-DNA
3
, and result in several, randomly
distributed metal nanoclusters on the DNA, while homogeneous metal deposition in the
surrounding medium is suppressed. To yield a complete cover of the DNA, as intended by
Braun et al., an additional metallization step is necessary. This results in a wire diameter
of roughly 50nm that is 25 times larger than the diameter of bare double-stranded DNA.
To avoid a random placement of the nanoclusters, Braun invented a method to prevent
specific DNA parts from being metallized by incorporating RecA enzymes [KB04].
Silver clusters on DNA are typically in the nanometer size and therefore exhibit the
optical properties described above in 2.2 and 2.3. These optical properties and their de-
pendence on various synthesis as well as DNA parameters are researched in this work.
Moreover a method is applied to metallize only labeled parts of the DNA and the growth
of nanoclusters on short oligonucleotides is investigated.
3
DNA of the bacteriophage '-phage' with a length of approximately 50.000 basepairs
12
Chapter 3
Synthesis of Metallized DNA
In this chapter the metallization processes are explained in detail. Moreover the challenges
in synthesis are discussed.
The metallized nanoparticles are in general produced as colloids. Colloidal solutions can
be classified between real solutions and suspensions, meaning they consist of small clusters
dispersed in a solution: In contrast to suspensions, the clusters will not sediment after a
certain time as the buoyancy compensates for gravity. It is important to avoid aggregation
of the clusters caused by attractive van der Waals forces; this aspect is discussed in detail
in chapter 6.
3.1
Hybridization of DNA
To synthesize metallic rods on DNA instead of nanoparticles with arbitrary shape, it is
necessary to use double stranded DNA. The reason for this is, that the persistence length
1
of double-stranded DNA is roughly 50nm, whereas that of single stranded DNA is only
approximately 1nm, depending on the buffer solution [MVH99], [SCB96]. To achieve hy-
bridization of complementary DNA single strands, one has to provide a sufficient amount
of counter ions, such as sodium ions, in solution [Man78]. Furthermore, a moderate pH
value is necessary. Hybridization experiments were conducted in different buffer solutions
at room temperature.
The success of hybridization can be detected by measuring the absorption at 260nm: When
the DNA hybridizes, the light absorption of the DNA bases at 260nm decreases. This
can be attributed to Coulombic interactions of the stacked bases in the double stranded
DNA: The dipole transitions moments of the bases interact via Coulombic forces. The
outcome of this interaction depends on the relative orientation of the bases. By forming
a double helix secondary structure the dipole transition moments change from random
distribution to a parallel orientation as schematically shown in fig. 3.1. The square of the
1
The persistence length is a mechanical quantity that gives the stiffness of a of a polymer[Wik07]
13
CHAPTER 3. SYNTHESIS OF METALLIZED DNA
Figure 3.1: Principle of hypochromism for DNA: During hybridization the random distribution
of the DNA bases changes to a parallel alignment. The absorption of the molecule, which is
dependent on the relative orientation of the bases, thereby changes.
Figure 3.2: Absorption of 2µM of complementary 48mer DNA strands in 10mM Tris buffer
before and after hybridization: One spectrum is taken 10min after mixing the complementary
strands. The second spectrum is taken 35min later. A decrease in the absorption at 260nm can
be seen. This is attributed to the hypochromism effect and indicates hybridization.
dipole transition moment is directly proportional to the absorption of light. Therefore the
absorption intensity changes during hybridization. In 1960 Tinoco was the first to explain
this effect and to calculate it by perturbation theory calculations [Tin60].
For the concentrations (1µM) and lengths (ca. 10 up to 30nm) of DNA used for this work,
I typically observed changes between 5% and 15% in the absorption spectrum when the
DNA hybridized. A typical spectrum is shown in fig. 3.2.
Fig. 3.3 shows how the absorption at 260nm for DNA in 3mM phosphate buffer changes
with time. The time spans for hybridization ranged from 30min (10mM Tris buffer
2
with 50mM NaCl) to several hours (3mM phosphate buffer
3
). All hybridizations were
conducted at room temperature.
2
=2-amino-2-hydroxymethyl-1,3-propanediol, trivial name: trishydroxymethylaminomethane, molecu-
lar formula: C
4
H
1
1NO
3
. At 20
o
C the pK
S
value is 8.3, the pH is titrated to its final value by addition of
HCl
3
Phosphate buffer was prepared with monosodium phosphate, monohydrate (NaH
2
PO
4
·H
2
O) and
disodium phosphate, heptahydrate (NaHPO
4
·7H
2
O)
14
3.2. REQUIREMENTS FOR METALLIZATION
Figure 3.3: Hybridization of 48bp DNA in 3mM phosphate buffer. The change in absorption at
260nm with time indicates the transition from single to double stranded DNA.
3.2
Requirements for Metallization
Besides the presence of counter ions for the hybridization of the DNA, another requirement
for synthesis of metallized DNA structures is the appropriate choice of the pH value around
7: On the one hand DNA only hybridizes at moderate pH values, on the other hand the
reduction rate for silver ions strongly depends on the environment and increases in an
alkaline environment.
Repeating metallization experiments in various solutions for several times, indicates the
necessity of a buffer solution: Only in buffer solutions the peak positions, intensities and
widths are reproducible. This might also be related to the pH-dependence of the silver re-
duction, because in non-buffered solutions the precipitation of AgOH leads to fluctuations
of the pH value.
Furthermore, one has to avoid precipitates of insoluble silver compounds: First, these
precipitates diminish the concentration of silver ions available for the metallization and
therefore prevents reproducibility concerning the silver amount. Second, these precipitates
decrease the transmission of light due to scattering; for silver concentrations larger than
about 1mM this decrease is so pronounced, that recording a transmission spectrum has
become impossible.
For the photochemical method I tested different solutions, among them 10mM Tris buffer
with 50mM sodium chloride (pH 7.3), 3mM phosphate buffer, aqueous sodium nitrate
solution at pH 6 and aqueous sodium nitrate solution with sodium hydroxide at pH 8.3.
Tris buffer provided best results regarding reproducibility and the peak shapes.
For the so called Tollens reaction method, that is explained in 3.4, ammonium and sodium
hydroxide being present in the solution guarantee a basic environment. As the silver ions
are bound in a silver diamine complex, no precipitation occurs. Thus, an aqueous solution
15
CHAPTER 3. SYNTHESIS OF METALLIZED DNA
containing 5mM sodium nitrate for hybridization is sufficient.
3.3
Photochemical Metallization
Photochemical Metallization is a fast method for coating commercial DNA with silver.
This technique, which was first introduced by Berti for -phage-DNA [BAF05] exploits
the fact that UV light of 253nm oxidizes DNA bases.
Silver ions form complex bonds with DNA bases, when silver nitrate is added to a DNA so-
lution. When irradiating the sample with appropriate UV light, a redox reaction proceeds
in which the excess silver ions in solution are reduced to elementary silver. The silver-base
complexes act as nucleation seeds for this reduction and cluster growth along the DNA
starts. This process is schematically shown in fig. 3.4. According to [PWLW02] the bonds
between metals and DNA after irradiation are covalent. The photochemical DNA met-
allization is a fast and simple method. Its major drawback is, that no site-specific DNA
metallization is possible.
Figure 3.4: Principle of UV Metallization of DNA
For the experiments DNA with a length from 23 up to 96 basepairs, corresponding to
a length between 8nm and 33nm, and different sequences was provided by IBA GmbH,
Germany. The DNA was diluted to concentrations of 1µM in buffer solution (total volume
100µl) and hybridized at room temperature for several hours, depending on the buffer
solution. An amount of typically 100µM to 1mM silver nitrate, provided by Sigma-Aldrich,
Germany, was added. For comparison, the silver cluster concentration in the commercial
20nm silver sphere solution, mentioned in chapter 2.2.1 is 1.2nM. This corresponds to
300µmol silver atoms per liter. After several minutes the samples were put in front of an
UV hand lamp (Benda, 253nm, ca. 6W), the distance between the cuvettes and the lamp
was about 1cm. Typical irradiation time was 45min.
3.4
Tollens Metallization of Aldehyde-Modified DNA
The Tollens reaction is a redox reaction that is sensitive to reducing functional groups
such as aldehydes. By modifying DNA bases with aldehydes, this method allows for a site
16
3.4. TOLLENS METALLIZATION OF ALDEHYDE-MODIFIED DNA
selective metallization of the DNA as schematically shown in fig. 3.5.
Figure 3.5: Schematic of site selectively metallized DNA: Metallic nanoclusters are expected to
grow at certain aldehyde-modified sites only.
Glenn Burley from Thomas Carell's group at Ludwig-Maximilians-University, Munich,
provided different oligonucleotides with a length from 23 up to 38 basepairs, whereby a
fraction of thymine bases in the DNA sequence was modified with aldehyde groups.
The modified DNA can be either synthesized by solid-phase synthesis or by a polymerase
chain reaction (PCR). In a certain reaction, described in detail in a paper of G. Burley
[BGM
+
06], aldehyde groups are added to the T bases. Via an enzymatic incorporation
of modified nucleosides, acetylene reporter groups (see fig. 3.6) are inserted into certain
Figure 3.6: Acetylene molecule
positions of the oligonucleotide strands. These acetylene reporter groups react with alde-
hyde azides (see figure 3.7 by an Huisgen
4
1,3-cycloaddition click reaction[HBSG64]. This
Figure 3.7: Aldehyde azide
reaction is catalyzed by Cu(I). The product of this reduction is an oligonucleotide with
oxidizable aldehyde groups added to some pyrimidine bases as schematically shown in fig.
3.8.
Figure 3.8: Schematic of the modified DNA: Aldehyde groups are clicked to several bases of one
strand of the DNA
After the modified DNA had been hybridized in an aqueous solution containing a suffi-
cient concentration of sodium ions, Tollens reagent was added. Tollens reagent contains
silver nitrate and sodium hydroxide. A brownish precipitate, Ag
2
O, immediately falls
out. Adding ammonium dissolves this precipitate again and binds the silver ions in so
called silver diamine complexes, [Ag(NH
3
)
2
]
+
. Aldehyde groups as well as other reducing
sugar groups can reduce the silver ions in this complex to elementary silver. Thereby the
aldehyde groups are oxidized to a carboxylic acid as shown in fig. 3.9.
4
Two unsaturated systems can react by building up a ring structure. This class of reaction was mainly
investigated by Rolf Huisgen
17
CHAPTER 3. SYNTHESIS OF METALLIZED DNA
Figure 3.9: Chemical equation of the Tollens reaction: Silver ions are reduced to elementary
silver whereas aldehyde groups are oxidized
The alkaline environment is necessary so that the carboxylic acids are present in a depro-
tonated form; this shifts the equilibrium position to the side of the products [Blu03].
As can be seen in the formula each aldehyde group can reduce two silver ions. The excess
silver ions can only be reduced in presence of a reducing agent, such as a development
solution containing formaldehyde. This development process will not produce new silver
clusters, but enlarges the already available seeds.
3.5
Materials Section
All chemicals were provided by Sigma-Aldrich. DI Water was provided by Millipore. The
DNA sequences are listed in the Appendix A. All hybridization and metallization was
done in disposable Brand UV cuvettes at room temperature.
Particular attention should be paid to the used cuvettes: When conducting the classical
silver mirror reaction, where Tollens reagent is added to a glucose solution, the elementary
silver sticks to the borders of a glass vial as it acts as an ion exchanger and attracts the
silver ions. In the experiments of this work, however, it is assumed that the positively
charged silver complexes are attracted by the negatively charged DNA molecules and not
by the glass. This is confirmed by Braun, who mentions that DNA can work as an ion
exchanger [BESBY98].
To achieve certainty that the reaction is carried out on DNA instead of the cuvette
window, a metallization process was done in a Hellma QS cuvette as well as a Brand
plastic UV cuvette: The spectra for the two samples were almost identical indicating that
the cuvette does not influence the reaction. Afterward the solution with the metallized
DNA was removed from the glass cuvette and it was dried with nitrogen gas: Now the
UV/vis spectrum showed no plasmon peak any more. This is a strong indication that the
silver precipitates on the DNA in solution and not on the glass surface.
Nevertheless, to avoid any competing processes plastic cuvettes were used instead of glass
ware.
18
Chapter 4
Absorption Spectrometry
In this chapter the principle of absorption spectrometry will be shown. Plasmon peaks
of metallized DNA will be investigated for UV- as well as Tollens-metallized samples.
Therefor several synthesis parameters were systematically varied. The experimental results
are compared to theoretical calculation.
4.1
Principle of Absorption Spectrometry
Absorption spectrometry, also known as UV/vis spectrometry, is a widely used tool for
metallic nanoparticle characterization as it shows the plasmon absorption peaks.
The working principle of this method is: A light source consisting of a tungsten halogen
lamp and a deuterium lamp irradiates the cuvette. A grating in the beam path adjusts
the wavelength and provides for monochromacity, the wavelength of the irradiating beam
can thereby be tuned over a wide range, from the UV to the near infrared spectral region.
Inside the solution the light intensity is attenuated according to the Lambert-Beer relation
I
trans
= I
0
e
-D
with the path length D, the extinction coefficient = N
ext
. N is the
density,
ext
is the extinction cross section, introduced in 2.2.1. A detector measures the
transmitted intensity. A schematic is shown in fig. 4.1.
By using the relation 'T+E=1', whereby 'T' is the transmission and 'E' is the extinction,
Figure 4.1: Principle of UV/vis spectrometry
19
CHAPTER 4. ABSORPTION SPECTROMETRY
the latter can be calculated. Extinction is the sum of absorption and scattering. As the
metallized DNA molecules are only a few nanometers in size, scattering is negligible for
the low light intensities in the spectrometer due to the Rayleigh approximation, described
in chapter 2.2.1. Thus in the following the term 'absorption spectrum' will be mostly used,
although some contribution to the extinction might come from scattering, i.e. by larger
particles available in the solutions. This notation points up that absorption of light is the
predominant effect we are looking for when applying this technique.
For the experiments a Lambda 900 two-beam spectrometer of Perkin Elmer was used.
Contrary to the simplified description above, now two beams are employed: The first beam
transverses the cuvette filled with the solution. The second beam, that goes just through
air and an aperture, compensates for any fluctuation. All spectra were normalized by the
absorption of a cuvette filled with water/buffer-solution. The reason for this normalization
is, that the plastic cuvettes partly absorb in the UV and therefore overlap the DNA
absorption.
4.2
Absorption Measurements of UV-Metallized Sam-
ples
Different short oligonucleotides were metallized using the UV method as described in
chapter 3.3. Thereby various parameters such as lengths and sequences of the oligonu-
cleotides, the irradiation time and the silver concentration were systematically varied.
Afterward the samples were characterized by UV/vis spectrometry.
Before metallization, the UV/vis spectra look like the following: The DNA in solution
exhibits a peak around 260nm due to the absorption of the bases. When adding silver
nitrate, the peak intensity increases and shifts to the red
1
. When ions, that form insoluble
silver compounds, such as chloride, phosphate or hydroxide, are present in the solution,
the extinction increases due to scattering of light by the precipitate. As discussed in 2.2.1
this extinction is proportional to
-4
. These spectra before metallization are shown in fig.
4.2.
Metallization of the DNA, i.e. the growth of nanoclusters on the DNA-template can be
monitored by the appearance of a plasmon peak between 400 and 450nm, such as shown
for commercial silver spheres in fig. 2.4.
Unless otherwise noted, the following measurements are conducted in 10mM Tris buffer
with 50mM NaCl at pH 7.3. Hybridization at room temperature was completed after 90
minutes at the latest. The oligonucleotides are always brought to final concentrations of
1µM. Unless otherwise noted the amount of added silver was 100µM, corresponding to a
silver-to-base of 1:1. Typical irradiation times range from 30 to 45 minutes.
1
The reason is that nitrate absorbs light around 300nm. The absorption peaks of the DNA and the
nitrate cannot be resolved in the spectra.
20
4.2. ABSORPTION MEASUREMENTS OF UV-METALLIZED
SAMPLES
Figure 4.2: Absorption spectrum of a Tris buffer solution with 1µM 48bp dsDNA before and
after addition of silver. Details about the different contributions to the absorption can be found
in the text.
4.2.1
Influence of pH
To find out optimum synthesis conditions, metallization measurements were done in so-
lutions with varying pH values.
Fig. 4.3 shows the UV/vis spectra of UV metallized DNA in pure water titrated to certain
pH values by adding NaOH
2
.
As the pH measurements were done with litmus paper, the values are just rough estima-
tions. The DNA is present as single strands, because the aqueous solution contains too
few counter ions. The ratio of silver ions to DNA bases is 1:1, the irradiation time in this
measurement is approximately 30 minutes.
Only in an alkaline environment, a plasmon peak appears. The peak intensity increases
with the pH value, the peak position shifts to the blue suggesting a smaller average particle
size. There is also a significant rise in the background extinction when increasing the pH.
This is caused by precipitated silverhydroxide, AgOH.
The reduction potential of silver is pH-dependent, therefore the reaction rate is also pH-
dependent. This was experimentally shown here: A pH above 7 is necessary to provide a
plasmon peak. As silver ions can bind OH
-
ions by forming silver hydroxide and therefore
lower the pH-value, a buffer solution is recommended for the UV-metallization. These
results support the use of Tris buffer at 7.3 as a solution for the UV-metallization synthesis.
2
The pH value of pure water was only 6. This probably originates from carbonic acids in water that
occur when carbon dioxide from air is taken up in solution.
21
CHAPTER 4. ABSORPTION SPECTROMETRY
Figure 4.3: Influence of the plasmon peaks of UV-metallized DNA on the pH value in solution.
Only for alkaline environments plasmon peaks show up.
4.2.2
DNA Sequence
Hybridization
As the persistence length of single-stranded DNA (ssDNA) is much shorter than that of
double-stranded DNA (dsDNA), different cluster shapes, especially different asphericities
of the nanoparticles, are expected. Therefore UV metallization experiments were carried
out on ss as well as dsDNA.
Silver was added to 48mer DNA in an Ag-to-base ratio of 1:1. The samples were irradiated
for 30min. The spectra can be seen in fig. 4.4, the results are summed up in the following
table:
Table 4.1: Absorption measurements of UV-metallized ss and dsDNA
ssDNA
dsDNA
plasm
404nm
409nm
I
plasm
25.9%
33.4 %
FWHM
80nm
86nm
I
background
16%
18 %
DN A
262nm
262 nm
I
DN A
77.3%
68.1%
plasm
is thereby the position of the plasmon peak. Its intensity is denoted by I
plasm
,
22
4.2. ABSORPTION MEASUREMENTS OF UV-METALLIZED
SAMPLES
Figure 4.4: Comparison of UV-metallization for single- and double-stranded DNA
given relative to the intensity of the incoming beam. FWHM, the 'Full Width at Half
Maximum' is a characterization for the broadness, it is given in nm. I
background
is the
background absorption, read off as absorption at long wavelengths, i.e. 700nm. It is given
relative to the incoming intensity.
DN A
gives the peak position for the absorption of the
bases, its intensity I
DN A
is again relative to the incoming beam. These denotations will
be also used for the following analysis.
The intensity of the plasmon peak in the spectrum is 1.5 times higher in the case of the
dsDNA. The concentration of DNA molecules in the single-stranded solution is, however,
twice as large as that in the double-stranded solution
3
The single-stranded sample might
therefor contain more clusters, whereas the clusters in the double-stranded solution might
be larger, assuming that all DNA molecules are involved in the metallization process and
all silver atoms become reduced. Thus the intensity is not applicable for comparing the two
systems. The plasmon peak for the dsDNA is slightly red-shifted compared to the ssDNA,
indicating larger particles. This originates either from the lower DNA concentration as
discussed above or from the larger dimensions of dsDNA due to its long persistence length.
The absorption of the DNA bases at 262nm is higher for the dsDNA than for the ssDNA.
4
The reason of this is the composition of the bases: The special sequence of the used
3
The reason is simply that the amount of bases was kept constant in solution. The double-stranded
DNA molecules contain twice as much bases as the single-stranded DNA molecules.
4
This is not a contradiction to Tinoco's theory presented in 3.1 as the DNA sequences in the two
solutions are not identical: The ssDNA consists of 2µM of the 48mer DNA sequence listed in B, whereas
the dsDNA consists of 1µM of this sequence and 1µM of the complementary sequence. In general, DNA
sequences do not have the same absorption coefficients as their complementary sequences.
23
CHAPTER 4. ABSORPTION SPECTROMETRY
48 basepair (=48mer, 48bp) absorbs less than its complementary strand. There is no
difference in the background extinction. The peak width is slightly larger for double-
stranded DNA.
Thus, there are slight differences between metallized ss and dsDNA. The red-shift of the
plasmon peak indicates that metallized dsDNA is larger than metallized ssDNA. This
accords to the expectations.
Base Sequence
To investigate the influence of the base sequence on the outcome of the metallization
process, three different 48mer dsDNA samples were metallized with fractions of G- and
C-bases of 0%, 48.5% and 75%. Fig. 4.5 shows the specra, the table lists the data:
Figure 4.5: UV/vis spectra for different 48mer oligonucletide sequences. All bases can be metal-
lized. The exact peak position as well as the intensity depend on the specific sequence, however.
Table 4.2: Absorption measurements of UV-metallized DNA with different GC-fractions
GC content
0%
48.5 %
75%
plasm
415nm
410nm
425nm
I
plasm
54%
34%
55%
FWHM
99nm
87nm
126nm
I
background
26%
17%
22%
DN A
263.5nm
262nm
259.6nm
I
DN A
86%
77%
83%
In contrast to low and medium GC-contents, the 75%GC-peak is quite asymmetric, in-
24
4.2. ABSORPTION MEASUREMENTS OF UV-METALLIZED
SAMPLES
dicating a very broad particle size dispersion according to Henglein [HG99]. The peak
intensity of the 48.5% GC-sample is significantly lower than for the other samples. This
is a reproducible effect that must be attributed to the special sequence of this sample.
The background intensities are approximately proportional to the peak intensities. The
positions and intensities for the DNA-nitrate-peak cannot be compared in this case, as
different bases show different absorption spectra. The peak position seems to depend on
the base sequence, but there is no general trend observable.
The measurements showed, that all bases can in general be metallized by the UV-method.
The peak intensities, positions and shapes, however, depend crucially on the DNA se-
quence. These observed differences indicate different particle sizes for the different base
sequences. As there is no general relation between the peak position or bandwidth and
the GC-content, the metallization probably does not only depend on the bases themselves
but also on the surrounding sequence.
4.2.3
Silver Amount
To observe the size and shape dependence of the particles on the silver concentration, the
amount of silver ions was varied between 10µM and 0.5mM, corresponding to silver-to-
base ratios between 0.1 and 5. The samples with 48mer DNA and silver were irradiated
for 45min.
Figure 4.6: UV/vis spectra of UV metallized DNA with various Ag concentrations
25
CHAPTER 4. ABSORPTION SPECTROMETRY
Figure 4.7: Blue shift of the plasmon peak when increasing the silver amount
Table 4.3: Absorption measurements of UV-metallized DNA for various silver concentrations
ratio Ag/bp
0.1
0.5
1
5
plasm
416nm
413nm
408nm
405nm
I
plasm
6.4%
22.6%
43.3%
90.8%
FWHM
-
94nm
96nm
110nm
I
background
2%
8%
17%
30%
DN A
262nm
262nm
262nm
262nm
I
DN A
77.4%
79.4%
80.8%
86.4%
The spectra show a distinguishable peak around 410nm for silver-to-base-ratios larger
than roughly 0.5. For silver-to-base ratios larger than about 10, a significant precipitation
of silver chloride reduces the transmittance of the sample and thereby makes the extinction
spectrum less reproducible. Thus only low silver-to-base-ratios shall be analyzed here.
The position of the nanoparticle peak shifts to the blue when increasing the silver-to-base
ratio. That is also shown in fig. 4.7. This trend still holds good after deconvoluting the
spectrum to eliminate to influence of AgCl scattering which is proportional to
-45
. The
interpretation of this blue-shift is given in 4.2.7.
One would expect the DNA peaks to shift to the red due to an overlap with the increasing
nitrate peak. This cannot be detected. A possible explanation is that more silver entails
more complexation. Silver-DNA-complexes, however, cause a blue shift of the base absorp-
tion peak [ES68]. Thus, the two effects might compensate for each other. The increasing
intensity of the DNA peak, that can be detected for large silver concentrations, traces
back to the increasing amount of nitrate ions in the solution.
5
The exact deconvolution method will be explained in 4.2.7. Deconvolution is necessary for the peak
determination as the plasmon spectrum is overlapped by the AgCl scattering spectrum, that depends on
the amount of added silver.
26
4.2. ABSORPTION MEASUREMENTS OF UV-METALLIZED
SAMPLES
With increasing silver amount, the intensity of the nanoparticle peak increases from 6.7%
at ratio 0.1 to 90.8% for ratio 5. This indicates either a greater number of metallized
particles or larger clusters. There is also an increase in the extinction background from
3% to 30%. This might originate from unspecific silver deposition in the sample, that
increases with the silver amount. For silver-to-base ratios not larger than 1 the FWHM
is roughly 70nm. When increasing the silver concentration by a factor of five, the peak
width grows 37% to a value of 96%, indicating a huge change in the particle size or a
bigger polydispersity of the particle sizes.
Summarizing it can be stated that plasmon peaks occur already for very small silver-to-
base-ratios. The peak intensity increases with the silver amount, a saturation value was
not observed. Thus more silver ions means either more or larger particles. The discussion
about the tendency of the peak position to shift to the blue with increasing silver-to-base
ratios will be discussed later, in 4.2.7.
4.2.4
Irradiation Time
To gain information about the reaction kinetics, the irradiation time for a 48mer sam-
ple with a silver-to-base ratio 1:1 was varied from 5min to 150min. Fig. 4.8 shows the
corresponding spectra.
Table 4.4: Absorption measurements of UV-metallized DNA as a function of the irradiation time
irradiation time
5min
15min
30min
150min
plasm
420nm
414 nm
410nm
404nm
I
plasm
55.5%
52.5%
49.5%
31.7%
FWHM
126nm
96nm
94nm
86nm
I
background
25%
22%
22%
8%
DN A
260nm
261.5nm
216.7nm
262nm
I
DN A
84%
80.7%
80%
71%
For all irradiation times a plasmon peak between 400 and 420nm is observable in the
UV/vis spectrum. With increasing time, the peak position shifts from 420nm to 404 nm
as shown in fig. 4.9. The discussion about this will be delayed to 4.2.7. In 4.2.6 it is shown
that the UV-Metallization does not induce breakage of DNA strands, thus this cannot be
the origin for the blue shift.
The peak intensity decreases with time from 55.3% to 31.5%. The background i.e. the
constant absorption at long wavelengths, remains relatively constant during the measure-
ment. The broadness of the peak (FWHM) decreases significantly from 126nm at 5min
to 86nm at 150min.
The intensity of the DNA-nitrate-peak decreases with time. The main reason for this is
that the bases get more and more covered by silver which partially screens them from
absorbing UV-light. A second contribution is a reduced overlap with the plasmon peak
27
CHAPTER 4. ABSORPTION SPECTROMETRY
Figure 4.8: UV/vis spectra of UV metallized DNA with various irradiation times
Figure 4.9: Blue shift of the plasmon peak when increasing the irradiation time at UV metal-
lization
28
4.2. ABSORPTION MEASUREMENTS OF UV-METALLIZED
SAMPLES
due to narrower peaks.
The metallization reaction seems to set in immediately with UV-irradiation as a pro-
nounced plasmon peak already occurs after short irradiating times. This is also confirmed
by the observation that a change in color from achromatic to light yellow instantaneously
sets in when irradiating the sample. In the first 60 to 90hours (due to clarity not all
measured graphs are shown) the peak positions, bandwidths and intensities change with
time. Spectra taken after 90min and 150min are similar indicating that the reaction has
been completed after approximately one hour.
4.2.5
DNA Length
As the absorption spectrum of a metallic nanorod depends on its length, the influence of
the DNA length on the plasmon peaks shall be investigated. Therefor the DNA length
was varied between 23bp and 96bp, corresponding to 7.8nm and 3.2.6nm.
Figure 4.10: UV/vis spectra for UV-metallized oligonucleotides of varying length
29
CHAPTER 4. ABSORPTION SPECTROMETRY
Table 4.5: Absorption measurements of metallized DNA with various lengths
length
23bp
38bp
48bp
96bp
GC-content
65.2%
57.9%
48.5%
47.9%
plasm
413nm
413nm
408nm
421nm
I
plasm
47.2
52
33.4
49.7
FWHM
101
101
96
124
I
background
17%
17 %
17%
21%
DN A
261nm
261nm
262nm
261nm
I
DN A
68.2%
70.8%
77%
94.6%
There are slight, reproducible differences between the individual spectra, listed in the
table above. The spectrum of the 23mer oligonucleotides is almost identical to that of the
38mer. This can be easily understood regarding the sequences as listed in Appendix A:
They both have the same starting and ending sequence. The intermediate part consisting
of 15 basepairs occurs once in the 23mer whereas it is repeated in the 38mer. The 48mer
spectrum does not fit to the other sequences as it is (reproducibly) much lower in intensity
and slightly red-shifted. This might trace back to the specific sequence of the 48mer as the
same observations were made when comparing the 48mer sequence to 48mer sequences
with different GC-contents in chapter 4.2. For the other sequences, there is a slight red-
shift of the peak with the DNA length, indicating larger particles. The fact, that a very
small quadrupole shoulder at 350nm can be discerned in the 96mer spectrum confirms
this conclusion.
In summary, samples with different length show slight differences in the absorption spec-
tra. A possible influence of the DNA length on the spectrum is, however, overlapped
by the sequence-dependence of the UV-metallization. Thus DNA molecules of various
lengths with similar DNA sequences have to be produced and analyzed to draw further
conclusions about the influence of the DNA length.
4.2.6
Reference Samples
Irradiation of samples with 10mM Tris buffer solution and silver nitrate shows a charac-
teristic structure: Below 320nm the transmission is lowered due to the absorption of the
nitrate present in the solution. Above 320nm the transmission is lowered due to unspecific
precipitation of silver chloride that occasionally occurs. A UV/vis spectrum of this can be
seen in the appendix B. This interpretation is confirmed by the way, that the spectrum
changes when varying the amount of added silver.
When a solution with DNA but without silver is irradiated, the spectrum does not change.
This is also an indication that the UV lamp does not break DNA strands - otherwise the
absorption would change according to 3.1.
The UV-irradiation is implicitly necessary for the particle synthesis, without it no plas-
mon peak occurs, not even after some hours. According to manufacturer information the
30
4.2. ABSORPTION MEASUREMENTS OF UV-METALLIZED
SAMPLES
Figure 4.11: Theoretical calculation of silver plasmon peaks in water, done by Slistan-Grijalva
et al. [SGHURS
+
05] Small particles mainly exhibit an enhancement of the absorbance as well
as a narrowing of the peak with increasing size. Larger particles show a red shift with increasing
size because of retardation effects.
intensity of the UV/vis spectrometer is not sufficient to reduce silver ions and induce
nanoparticle growth.
Thus, plasmon peaks only occur if silver as well as DNA is present in solution and the
sample is irradiated.
4.2.7
Discussion
Comparison with Theory
The experimental values shall now be compared to theoretical calculations done by Slistan-
Grijalva [SGHURS
+
05] et al.. They applied Mie theory up to the tenth multipole for
silver nanospheres in water. Although the metallized DNA structures are not assumed to
be spherical particles, the experimental and theoretical values shall be compared to get a
first insight about the processes happening during metallization. The theoretical values,
that are shown in fig. 4.11 and 4.12, agree with experiments I did with commercial silver
spheres in a range from 20 to 80nm in diameter, shown in fig. 4.13.
6
For small particles, the peak position remains approximately constant. Small red shifts
with increasing size are due to a radius-dependent dielectric function. Very small parti-
cles show broad peaks: In bulk medium the mean free path of electrons is 52nm. Thus
for particles with a radius shorter than 26nm an additional damping mechanism, surface
scattering, becomes important and increases the peak width. According to the electro-
static calculation in 2.2.1, the peak intensity is proportional to the volume. This is at least
quantitatively in accordance with Slistan-Grijalva's results for small particles. Theory pre-
dicts a significant red shift in the plasmon spectrum for larger particles since retardation
effects, such as screening of the incoming wave, shift the plasmon peak to the red. The
6
Experimental results for the maximum absorbance in dependence on the particle size are not shown as
the concentrations of the commercial particles are size-dependent. Thus the intensities are not comparable.
31
CHAPTER 4. ABSORPTION SPECTROMETRY
(a) maximum absorbance
(b) bandwidth
Figure 4.12: According to Slistan-Grijalva's calculations [SGHURS
+
05], particles in water with
a radius of 20nm show a maximum plasmon absorbance twice as large as 3nm particles. The
bandwidth remains almost constant for particles with sizes from 10 to 25nm. For small and large
particles, several damping mechanisms broaden the peak. Only the graphs for water, not the
graphs for ethylene glycol are relevant for the comparison!
(a) plasmon peak position
(b) FWHM
Figure 4.13: Experimental values for the plasmon peak position and the bandwidth in depen-
dence on the particle size. extracted from 2.4 Measurements were done with commercial silver
spheres described in the text.
32
4.2. ABSORPTION MEASUREMENTS OF UV-METALLIZED
SAMPLES
excitation of higher order modes, such as the quadrupole mode, broadens the spectrum.
One aspect Slistan-Grijalva does not discuss, is polydispersity of the particle sizes. Thus
experimental spectra will certainly show broader spectra than Slistan-Grijalva predicts.
Another difference between my measurement conditions and Slistan-Grijalvas calculations
is the refractive index of the surrounding medium. Slistan-Grijalva chose water and there-
fore a refractive index of 1.33. In the UV-metallization experiments, the aqueous solution
contains various amounts of salt; the refractive index is therefore slightly increased. More-
over the influence of the DNA, namely its refractive index and the core-shell character
of the resulting nanoparticles must not be ignored. The refractive index influences the
absolute values of the peak position, the maximum absorbance as well as the bandwidth.
The result of the UV metallization is, that the metallization process worked for all DNA
samples. All spectra showed plasmon peaks in the same range and with similar shapes
and widths. There were slight, reproducible differences, indeed, that shall be concluded
and discussed in the following:
All measurements show plasmon peaks in the range between 405 and 425nm. As the re-
fractive indexes of neither the buffer solution nor the DNA are known, the values cannot
be quantitatively compared to the theoretical values. It can be only concluded that the
particle radius is probably below 20nm. Taking the peak position as the sole indicator for
particle size, the smallest particles, corresponding to the shortest peak wavelengths, oc-
curs for ssDNA, for high silver concentrations and long irradiation times. The former can
be understood as ssDNA has a shorter persistence length than dsDNA and therefor forms
bundles that are at least in one direction smaller than the dsDNA. The latter cannot be
understood with this simple theory and will be explained below. A possible explanation
for the third one is that the DNA strands break during irradiation and therefore the par-
ticle size decreases. An alternative explanation will be given below. The largest particle
sizes occur for long DNA, short irradiation times and very high GC-contents. The large
particle size for long DNA matches the expectations as long DNA means more seed posi-
tions and therefore larger nanoparticles. The large size for short irradiation times cannot
be explained with this theory, only conjectures can be made about its origin. The large
particle size for high GC-contents shows that the metallization is, indeed, sequence depen-
dent. The FWHM is mostly between 80 and 100nm. Assuming complete monodispersity,
which certainly does not hold in reality, this means diameters below 9 or above 70nm
according to Slistan-Grijalva. The smaller size is more in accordance with the results of
the peak positions and therefore more probable. In reality, however, polydispersity, the
deviations of the refractive index from that of water as well as shape effects may influ-
ence the FWHM. Most peaks were quite symmetric, exceptions are the 75%GC-sample,
samples that were irradiated for a very short time and samples with long DNA. Accord-
ing to Henglein [HG99] an asymmetric peak shape indicates a broad size distribution of
particles. The reverse, however, is not always true. A slight shoulder is detectable for the
75% GC sample which might be indicative of a quadrupole mode as predicted for large
(60nm)particles. On the other hand it might reflect the particle asymmetry; a definite
assignment is not possible at this stage. The background extinction, that can be assigned
33
CHAPTER 4. ABSORPTION SPECTROMETRY
to extinction by unspecific silver deposition is between 2 and 30% and depends e.g. on
the amount of added silver ions. There might be a tendency for a red-shift and a broad-
ening of the plasmon peaks with the length of the DNA. To prove this, it is necessary to
build up similar DNA sequences of various lengths, as the sequence strongly influences the
metallization process. When increasing the amount of silver the absorption intensity as
well as the bandwidth increases, when increasing the irradiation time, the absorption in-
tensity as well as the bandwidth increases. Regarding the calculations of Slistan-Grijalva,
nor of them can be explained by simple Mie theory as the absorption intensity and the
bandwidth always show opposite trends. Hence a more advanced theory concerning about
particle shapes, as will be discussed later, has to be applied.
As a conclusion, the comparison of the measured values and the Mie theory calculated
by Slistan-Grijalva show that small nanoparticles have been produces. The peak position
depends on the length and sequence of the DNA as well as the measurement conditions.
The results cannot be completely explained in a quantitative way and thus another theory
has to be applied.
Rodlike Character
As we assume the metallized DNA to have either rod-like (whole DNA is metallized)
or chain-like (nanoclusters have grown on DNA) character, two extinction peaks should
appear in the spectrum. The electrostatic calculation for nanorods from 2.3.1 predicts a
low intensity peak at short wavelengths around 400nm as well as a high intensity peak
at long wavelengths. The latter one can be anywhere in the visible or near infrared range
and strongly depends on the axis ratio. A similar behavior is also expected for chains of
nanoparticles [RABdA06], as they might have grown on the DNA strands. Experiments
with nanorods in solution, as for example done by C. Soennichsen's group [BKRS06] are
in great accordance with this theory.
Recorded absorption spectra of the metallized DNA, however, have not shown a dominant
long-wavelength peak, even when the scan range was extended up to 1200nm.
Literature research, however, shows that not all plasmon peaks obtained for nanorods are
in the wavelength range specified by the electrostatic calculation as mentioned above:
E.g. in 1989, Goudonnet [GBP89] analyzed silver spheroids with a length of 12nm and a
diameter of 3.6nm on a quartz substrate. He identified two peaks in the spectrum, one at
468nm and one at 338nm. Qu [QD05] found a 350nm plasmon peak for silver nanoparticles
grown on a copper foil by silver mirror reaction. In his PhD thesis, M. Noyong [Noy05]
calculated the plasmon spectrum for different aggregates of nine silver spheres with a
diameter of 40nm. He found one peak around 350nm and another one between 450nm
and 500nm.
Possible low-intensity absorption peaks of metallized DNA in a range below 400nm might
be covered by the absorption of the DNA and the nitrate. Thus the spectra were decon-
voluted
7
to eliminate all peaks that are not related to plasmon oscillations. The result of
7
The transmission spectrum after the irradiation was divided by the spectrum before irradiation.
Therefore the influence of DNA, salts and silver chloride precipitation is eliminated. Afterward the spectra
34
4.2. ABSORPTION MEASUREMENTS OF UV-METALLIZED
SAMPLES
Figure 4.14: Relative absorption of UV-metallized DNA of various lengths. The spectrum is
a deconvolution of the measured data with the data just before nanoparticle growth. As this
spectrum shows just the relative absorption, also negative values are possible: As explained
already above, the absorption of the DNA bases decreases during the metallization process,
therefore the relative absorption is below zero.
such a deconvolution can be seen in fig. 4.14. Indeed, there are two peaks visible: One
at 300nm and one at 420nm, which can be assigned to the transverse and longitudinal
plasmon respectively.
In the spectra the long wavelength peak is more intense and more sensitive to changes in
the synthesis parameters than the short wavelength peak. This is in accordance with the
electrostatic calculation. The following table lists the peak positions found for deconvo-
luted data.
was plotted as an absorption spectrum. The correction was done with the transmission spectra to avoid
division by zero
35
CHAPTER 4. ABSORPTION SPECTROMETRY
Table 4.6: Peak positions after deconvolution
length
23bp
313nm
416nm
38bp
313nm
420nm
48bp
314nm
415nm
96bp
307nm
423nm
sqeuence
0% GC
302nm
419nm
48.5% GC
313nm
415nm
75% GC
306nm
428nm
irr. time
5min
302nm
423nm
15min
306nm
415nm
30min
308nm
413nm
45min
312nm
411nm
90min
314nm
407nm
150min
313nm
407nm
Ag-base ratio
0.5:1
307nm
416nm
1:1
312nm
412nm
5:1
312nm
404nm
The longer the DNA, the more separated the peaks are, i.e. the more elongated the
particle is. This is again not true for the 48mer DNA. The 48mer sequences with different
GC-values also exhibit two peaks. There is, however, no trend observable. The longer
the samples are irradiated, the closer the peaks approach. This indicates a decrease of
the particles' axis ratio: The maximum length of the nanoparticles is determined by the
length of the DNA and therefore does not change during metallization. With increasing
irradiation time, more silver is deposited on the DNA, the particles become rounder. A
similar explanation can be used to explain the peak shifts dependent on the silver-to-base
amount: The more silver is added, the more spherical the particle becomes.
Thus, both peaks could be identified in the spectrum but contrary to the electrostatic
calculation, they are tremendously blue-shifted: One possible reason for the blue shift
of the long wavelength plasmon is the charge of the nanoparticles. According to Mul-
vaney [MPJG
+
06] the injection of electrons to the surface shifts the longitudinal plasmon
50nm/V to the red. This effect is more pronounced for rods than for spheres, as the en-
hanced oscillator strength makes the particle be more prone to surface perturbations. It
also leads to a damping of the peak.
Concerning the literature mentioned above, the blue shifts of both plasmon peaks mainly
occurs for particles in contact with a surface. The 'surface' in this experiment is the DNA.
Unfortunately there is no theory in literature explaining this effect up to now.
Summing up it can be said, that small rod- or chainlike metallic nanoparticles have grown
36
4.3. ABSORPTION MEASUREMENTS OF TOLLENS-METALLIZED
SAMPLES
on DNA templates. Their axis ratio can be tuned by varying the length, by the added
amount of silver and by the irradiation time. Unfortunately no quantitative theory is
applicable to extract the axis ratios from the optical data.
4.3
Absorption Measurements of Tollens-Metallized
Samples
The following section covers the optical properties of Tollens-metallized DNA. Aldehyde-
modified oligonucleotides are therefor metallized in solution by the Tollens method de-
scribed in 3.4 and investigated. The analysis method, UV/vis absorption spectrometry,
is the same as for UV-metallized DNA. Parameters such as the length, the modification
density and the silver amount are varied systematically. Moreover the temporal evolution
of the growth as well as the influence of a development solution are investigated.
In contrary to the UV metallization process, where a buffer is necessary to guarantee
for good reproducibility, the solution for the Tollens metallization must not buffer at all.
This is, because the Tollens reaction works best in alkaline environments and therefore the
Tollens reagent itself has a pH value above 10, depending on its dilution. DNA however
cannot be hybridized under extreme pH conditions.
8
. If DNA was hybridized in a buffered
solution with a pH significantly lower than 9, the environment would not be basic enough
for the reaction, even after the addition of the Tollens solution. Thus, to fulfill all these
requirements, the modified DNA was hybridized in an aqueous solution containing 10mM
NaNO
3
. The samples were kept at room temperature overnight to provide enough time
for complete hybridization. Afterward an aliquot of a Tollens reagent was added to the
solution and raised the pH to the required pH above 9.
For metallization, three different oligonucleotides were provided: A 38 basepair and two
different 23 basepair oligonucleotides. The DNA consisted of 30 basepair and 15 base-
pair long middle parts, respectively, of which every third or fifth base is modified. The
denotation for the oligonucleotides is first the length of the middle part and second the
density: 15-5 means a 23 basepair DNA containing a middle part where every fifth atom
is aldehyde modified. The available oligonucleotides therefore are denoted as '30-3', '15-
5', '15-3'. It is important to notice, that only one strand of the dsDNA is modified. A
schematic of the different oligonucleotides is shown in fig. 4.15.
Composition of Tollens Reagent
The Tollens reagent used for the experiments, unless otherwise noted, contains 317µM
AgNO
3
, 37.5µM NaOH and 200µl of a 28% ammonium hydroxide solution and was filled
up with water to reach a final volume of 1ml.
The outcome of the experiment depends crucially on the amount of the ammonium: Too
8
This was tested with 1µM 48mer DNA in an aqueous solution with 10mM NaNO
3
to which NaOH
was added until a pH value of 9.5 was reached: No hypochromism was observed indicating that the DNA
was not hybridized
37
CHAPTER 4. ABSORPTION SPECTROMETRY
Figure 4.15: Simple schematic of aldehyde modified dsDNA, for the the utilized samples 15-3,
15-5 and 30-3. Each black vertical line indicates a basepair, red disc indicates an aldehyde-
modification.
little ammonium cannot dissolve all the silver oxide precipitation. Too much ammonium,
however, drastically slows the reaction.
When using a Tollens reagent containing 450µl of the ammonium solution instead of
200µl, it lasts between 24 and 72 hours until a reaction takes place. The peak that occurs
is not intense, only 0.06% of the light is absorbed. The spectra show differences between
samples containing modified and that containing unmodified DNA, but these differences
are not very pronounced, indeed. This is shown in fig. 4.16. When decreasing the ammo-
nium amount to the 200µl mentioned already above, a significant plasmonic peak already
appears after several minutes for the modified DNA. After some hours the velocity of the
reaction slows down. Various spectra are shown in the next paragraphs.
Reference Samples
There is also a very small peak in the spectrum, when Tollens reagent is added to un-
modified DNA, that can be attributed to electrostatic attraction of the silver ions to the
negatively charged phosphate backbone. A 10mM NaNO
3
solution with an aliquot of the
Tollens reagent containing no DNA did not show any plasmonic peak at all. Reference
samples such as B.4 can be found in Appendix B.
4.3.1
Absorption Measurements of Tollens-Metallized dsDNA
Modified dsDNA shall now be Tollens-metallized and afterward be characterized with
respect to its optical properties.
As described above, modified oligonucleotides were hybridized in an aqueous NaNO
3
so-
lution and afterwards provided with aliquots of a Tollens solution poor in ammonium.
Figure 4.17 shows the temporal evolution of a plasmon peak after addition of Tollens.
The peak absorbance increases with time and finally approaches a maximum value (not
shown in the figure). The peak position shifts to the blue with time. There is no general
trend for the bandwidth: Whereas the 30-3 samples broaden with time from around 100nm
38
4.3. ABSORPTION MEASUREMENTS OF TOLLENS-METALLIZED
SAMPLES
Figure 4.16: UV/vis spectra 72 hours after addition of an ammonium-rich Tollens reagent to
three different samples. For modified 30-3 DNA a weak peak has developed. The sample with
unmodified DNA shows a slight increase in absorption at the same wavelength as can be seen in
the zoom. The absorption of the sample containing no DNA is also enhanced without showing
any peak structure.
39
CHAPTER 4. ABSORPTION SPECTROMETRY
Figure 4.17: Temporal evolution of the plasmon peak for Tollens metallization. The silver amount
(32mM) was chosen in the saturation regime, the dsDNA center parts lengths was 15 basepairs;
every third basepair was aldehyde-modified.
to 113nm, the 15-3 sample seems to decrease its HWHM
9
from 95nm to 91nm.
The increasing intensity indicates either more or larger particles. This matches the expec-
tations as the particles are monitored during their growth process. This is discussed in
more detail in 4.3.4. The blue shift with time reminds of the blue shift of UV-metallized
DNA with increasing irradiation time. Contrary to the UV metallization, a deconvolu-
tion of the spectrum some time after the Tollens addition with the spectra directly after
the Tollens addition does not reveal a second peak. For wavelengths shorter than 350nm
absorption can be attributed to the Tollens reagent. The inexplicable fluctuation in the
FWHM may be attributed to inaccuracies in evaluating the data.
After keeping the samples overnight at room temperature, the mentioned trends are not
valid any more: The peak shifts to the red and gains disproportionately high intensity.
This tendency suggests the aggregation of the nanoparticles in solution. A further growth
of the clusters is not probable, as a saturation was already observed some hours after
preparation.
9
Contrary to the UV-metallized samples now the Half-Width-Half-Maximum (HWHM) is given: Due
to an overlap with the DNA/nitrate peak the FWHM cannot be read off for the broad peaks of Tollens-
metallized samples. As one cannot assume a symmetric particle shape either, the HWHM is given instead
of the FWHM.
40
4.3. ABSORPTION MEASUREMENTS OF TOLLENS-METALLIZED
SAMPLES
Fig. 4.18 shows the UV/vis spectra for different modified oligonucleotides five hours after
addition of two different Tollens concentrations.
Figure 4.18: The three oligonucleotide types described in the text 5 hours after addition of two
different amounts of Tollens solution
The maximum absorbance increases with the amount of added silver as well as the con-
centration of modified bases. The more seeds are in solution, i.e. the higher the amount
of modified bases, the more the peak shifts to the longer wavelengths, from 417nm for
the 15-5 oligonucleotides to 430nm for the 30-3 oligonucleotides. For the 30-3 sample the
addition of more silver leads to a red-shift of the peak, for the 15-3 and 15-5 basepair
samples, the peak position seems to be independent of the silver amount. The peaks are
broad compared with the UV-metallized samples, the HWHM values range from 86nm
for short DNA with low modification density to 112nm for long DNA with many modi-
fications. The background, an indication for unspecific silver precipitation, is below 10%
and therefore comparatively low.
The increasing intensity is according to the expectations and indicates larger effective
particle sizes caused by either cluster-cluster-interactions or a coalescence of the single
clusters: The single clusters grown on DNA, that are only separated by few nanometers,
may finally grow together to one particle. The red-shift with increasing modification den-
sity, i.e. seed amount is also understandable: More silver can be deposited, the particles
therefore become larger.
Thus, modified DNA can be metallized by adding a Tollens reagent. The maximum ab-
sorbance depends on the amount of modified bases and increases with time. The samples
are probably not stable for more than one afternoon. The plasmonic peaks are red-shifted
41
CHAPTER 4. ABSORPTION SPECTROMETRY
Figure 4.19: Temporal evolution of a Plasmon peak for ssDNA in water. The oligonucleotides
are modified 30-3 samples, Tollens with 32mM silver was added.
and broadened in comparison to UV-metallized samples. However, the absorption spec-
troscopy results cannot directly be compared for the UV- and the Tollens-metallized data.
The reason is that two different solutions with different refractive indices, ionic strength,
etc. were used.
4.3.2
Absorption Measurements of Tollens-Metallized ssDNA
The results of double stranded oligonucleotides shall now be compared to measurements
of single stranded oligonucleotides. Single stranded 30-3 DNA was metallized in water and
analyzed by spectrometric methods up to 4hours later. The added silver concentration
was 32mM.
According to expectations, the Tollens metallization also works for single-stranded DNA.
The plasmon peak is red-shifted in comparison to the measurements with dsDNA and
differs in its temporal evolution: Within several hours, the plasmon peak for 30-3 oligonu-
cleotides shifts from 436 to 447nm. The HWHM is above 100nm and therefore in the same
regime as for comparable dsDNA. The maximum absorbance is 62% after 5 hours. This
can be seen in fig. 4.19.
When conducting the experiments with single-stranded DNA in 10mM NaNO
3
instead
of water, the results hardly differ: The HWHM is even 110nm, the maximum intensity
reaches 80% after 2 hours.
After storing the sample at room temperature overnight, the maximum absorption is
100%, peak wavelength as well as the width of the peak cannot be determined any more.
42
4.3. ABSORPTION MEASUREMENTS OF TOLLENS-METALLIZED
SAMPLES
The background has also gained intensity. This suggests that the DNA is not stable and
aggregates.
Tollens metallized ssDNA samples therefore hardly differ from metallized dsDNA despite
that the plasmon peak is slightly red-shifted.
4.3.3
Development Solution
Principles
For the reduction of silver in the Tollens reagent, a reducing substance, such as aldehyde
groups, is necessary. Each aldehyde group can only reduce two silver ions, that then form a
negatively charged Ag
2
cluster[BGM
+
06]. The excess silver ions in solution are attracted
to these clusters due to electrostatic forces but they cannot be reduced any more. To
achieve a growth of larger clusters, the addition of a development solution (DS) which
reduces the excess silver ions is necessary. The development solution, that I used for my
experiments, consisted of 200µl formaldehyde (35%) and 250µl citric acid (1%) in 100ml
water.
Concentrated Development Solution
20µl of this development solution were added to the samples, where no peak has developed
yet. After some minutes all the samples showed a huge plasmon peak regardless whether
they contained DNA and whether the bases were modified or not. After some hours, for
the sample without DNA the spectra has turned to a broad absorption band without any
peaks, indicating a large number of unspecifically reduced silver clusters sizes, not only
in the nanometer range, present in the solution. The samples containing DNA, however,
show a huge plasmon peak. As the samples were not transmitting any more due to the
strong reaction, the samples had to be diluted by a factor of 20 before measuring. This is
shown in fig. 4.20
There is hardly any difference between the modified and the non-modified DNA. A possible
explanation is, that the development solution has also reduced silver ions, that were
electrostatically attached to the unmodified DNA backbone.
Diluted Development Solution
When adding an amount of 2µl development solution to the samples, only the samples
with modified DNA exhibit a plasmon peak, as shown in fig. 4.21.
The addition of 2µl development solution enhanced the maximum absorption by a fac-
tor of seven (3.2mM Tollens) or four (32mM Tollens) respectively. The addition of DS
slightly increases the HWHM (98nm instead of 91nm). But as the use of development
solution allows for lower silver concentrations, smaller peaks with a HWHM of 74nm can
be produced. The plasmon peak is blue-shifted (from 417 to 410nm).
Duration of the Development Process
For a controlled nanoparticle growth it is necessary to stop the development process
deliberately.
43
CHAPTER 4. ABSORPTION SPECTROMETRY
Figure 4.20: Absorption two hours after adding 20µl of the development solution.
Figure 4.21: Comparison of absorption spectra with and without development solution
44
4.3. ABSORPTION MEASUREMENTS OF TOLLENS-METALLIZED
SAMPLES
One idea is to filter out the development solution after a certain time. Thus the sample
was filtered for 30min with a Millipore filter membrane with a pore size of 0.2µm. Taking
a UV/vis spectrum afterward showed that the amount of nitrate ions was significantly
reduced in comparison with the original solution, which is indicated by the steep drop of
absorption below 300nm. The UV/vis spectra for two identically prepared samples, one
with and one without filtering, show almost identical plasmon peaks indicating that the
development process has not been stopped. A slightly higher maximum absorbance for the
filtered sample even indicates that the filtration might accelerate the development process.
This might be due to catalysis effect on the membrane surface. Using concentrators as an
alternative filtering method is impossible due to its long duration that makes investigations
of the temporal evolution impossible.
Another possibility to stop the cluster growth is acidifying the solution as the reduction
potential of silver is pH dependent. Addition of most acids, such as HCl or HSO
4
leads to
the precipitation of insoluble silver salts. According to [Blu03] the addition of nitric acid
destroys the existing silver clusters by oxidizing the atoms:
Ag + HN O
3
+ H
+
Ag
+
+ H
2
O + N O
2
. Thus an optimum method to stop the development process is still to be found. A promis-
ing method is acidifying the solution with some different chemicals, such as carboxylic
acid.
In conclusion, development solutions enhance the plasmon intensities and therefore make
it possible to work with only few modified bases and low silver concentrations. The chal-
lenge is to control the development process in such a way that it can be stopped again.
4.3.4
Data Analysis and Discussion
Affinity constants and kinetic rates
The experimental results presented in the previous chapters shall now be discussed with
regard to their affinity constants and kinetic rates. The first one addresses to the 'binding
strength' of the silver to the DNA, the second one covers the temporal evolutions.
The peak intensities for the three types of modified oligonucleotides one hour after adding
the Tollens reagent are plotted in fig. 4.22 in dependence of the silver concentration.
The measured data, absorption vs. concentration, are fit with a function
A = Const. ·
cK
A
1 + cK
A
according to the Langmuir model, where I is the plasmon peak intensity, c the silver
concentration in M and K
A
the affinity constant, see [LKSH00] for reference. The inverse
of the affinity 1/K
A
is the half-saturation concentration, where half of the binding sites are
occupied. One hour after complexation, the half saturation value
1
K
A
is 7mM for the 30-3,
45
CHAPTER 4. ABSORPTION SPECTROMETRY
Figure 4.22: Concentration dependence of the plasmon peak intensity for different DNA length
and aldehyde density. The spectrum was taken one hour after addition of the Tollens reagent.
The solid lines are the fits described in the text.
5mM for the 15-3 and 3mM for the 15-5 DNA. The average half saturation constant per
modified base is 1mM. As the saturation has probably not been completed after one hour,
the calculated half saturation values might exhibit large errors. The short time span of one
hour was chosen to avoid any influence of processes such as cluster-cluster-agglomeration.
The temporal evolution of the plasmon peak intensity for a silver concentration in the
saturation regime is shown in graph 4.23. The reaction is fast in the beginning and then
slows down. An exact saturation value cannot be extracted from the experimental data
because the clusters aggregate overnight and this falsifies the data. A normalization of
the curves in fig. 4.23 (not presented here) shows that the temporal evolution is fastest
for the 30-3 sample. 15-5 and 15-3 samples have similar development velocities, the 15-3
sample is slightly faster.
Fig. 4.24 shows the temporal evolution of the peak intensity after 2µl of the development
solution have been added. Two samples containing 30-3 oligonucleotides were therefore
kept at room temperature for 90min after 3.2mM and 32mM of Tollens, respectively, had
been added. Then the development solution was added.
The more silver is present in solution, the faster the development solution works. A sat-
uration of the plasmon peak's growth is reached after one to three hours.
46
4.3. ABSORPTION MEASUREMENTS OF TOLLENS-METALLIZED
SAMPLES
Figure 4.23: Temporal evolution of the plasmon peak intensity for different DNA lengths and
modification densities
Figure 4.24: Temporal evolution of the plasmon peak intensity for two different silver concen-
trations after adding 2µl of a development solution
47
CHAPTER 4. ABSORPTION SPECTROMETRY
Discussion
Tollens-Metallization exclusively metallizes modified oligonucleotides. With absorption
spectroscopy the growth of nanoparticles can be monitored. The reaction speed as well
as the maximum absorbance depend on the number of modified bases. Addition of a
developing solution speeds up the kinetics and enhances the intensity. There are only
slight differences between metallized ssDNA and metallized dsDNA. A second plasmon
peak could not be found be deconvoluting the data. This might be an indication that the
particles are less rod-like. Another possible explanation is, that the second peak cannot be
resolved due to the large bandwidth: The peaks are broad - the HWHM is about 100nm -
and red-shifted - typical peaks occur at 440nm - compared to UV-metallized samples. To
analyze whether this originates from larger particles sizes or from extrinsic effects such as
the refractive index in solution or the ionic strength, a different measurement technique
must be applied. An optical method to determine particle sizes down to the nanometer
range, Dynamic Light Scattering, will be introduced in the next chapter.
48
Chapter 5
Dynamic Light Scattering
In the following chapter the principle of Dynamic Light Scattering (DLS) will be explained.
Simulations about the DLS results of rodlike nanoparticles will be presented. Furthermore,
experimental results obtained for UV- as well as for Tollens-metallized samples will be
shown and discussed.
5.1
Principle of Dynamic Light Scattering
Dynamic Light Scattering, DLS, is a time-resolved measurement technique, that provides
information about the size of small particles with diameters ranging from nano- to mi-
crometers.
5.1.1
Measurement Technique
A scheme of the setup for this measurement is the following: A cw laser beam irradiates
the sample at a fixed wavelength that has to be chosen so that the sample shows sufficient
scattering intensity at this wavelength. At a fixed angle, typically 90 degrees, a photode-
tector measures the scattered intensity and sends the signal to an autocorrelator whose
output is processed by a computer program. A scheme of this is shown in fig. 5.1.
5.1.2
Fundamentals
The physical principle of DLS is a decorrelation of the scattered electrical field or the
scattered intensity, respectively, due to Brownian motion of the particles.
When the sample is irradiated, all the particles in the illumination volume act as sec-
ondary sources and scatter the light as explained in 2.2.1. Depending on the inter-particle
49
CHAPTER 5. DYNAMIC LIGHT SCATTERING
Figure 5.1: Simplified illustration of the DLS setup: A laser beam irradiates a sample at a fixed
wavelength. The scattered intensity is measured by a photodetector and sent to an autocorrelator
Figure 5.2: Random fluctuations of the scattered intensity due to Brownian motion. Only during
a very small time span, the signal is correlated.
distances, the scattered electromagnetic waves of these secondary particles interfere con-
structively as well as destructively and sum up to an effective scattered wave that is seen
by the detector. Since the particles move due to Brownian motion, intensity fluctuations
are measured with respect to the average scattering intensity.
When the signal is measured at two times that are close together, hardly anything will
have changed in the illumination volume and the outcome signals are almost the same.
The correlation for these two measurements is therefore close to 1. When the signal is
measured again after a certain span of time, the particles will have moved in respect
to each other due to Brownian motion. Because of this movement, the phase relations
between the scattered waves have changed and the effective sum signal differs from the
original one. There is no correlation any more between the initial signal and the signal
measured after a certain time span. This principle is sketched in fig. 5.2 and 5.3.
Thus the measurement provides a correlation function as shown in fig. 5.4, defined as
ACF=
t=0
I
0
· I(t)dt, that starts at value 1 (=fully correlated) and falls to zero for long
intervals. This decrease is due to the random fluctuations of the detected signal as the
particles undergo random Brownian motion. The Cauchy-Schwartz inequality shows that
the signal at an arbitrary time can never exceed the maximum value which is reached at
the origin of time.
50
5.1. PRINCIPLE OF DYNAMIC LIGHT SCATTERING
Figure 5.3: Physical principle of the decorrelation of the scattered signal due to Brownian motion:
In short time spans the particles hardy move with respect to each other and thus the effective
output signal remains approximately constant. This is not valid any more for longer time periods,
when the particles have randomly moved: The output signals are not correlated any more.
Figure 5.4: Measured autocorrelation function for a UV-metallized sample (1µM 48mer dsDNA
in 10mM Tris buffer, Ag-to-base 1:1, 10min irradiation)
51
CHAPTER 5. DYNAMIC LIGHT SCATTERING
From this autocorrelation function one can yield the decorrelation time , that is defined
as the time when the signal has dropped to half of its initial value. The decorrelation time
is inverse proportional to the diffusion coefficient D:
=
1
q
2
· D
where q is the wave vector of the scattered light that is defined as
q =
4n
sin/2
with the refractive index n, the wavelength and the scattering angle . The diffusion
coefficient is related to the size of the particle by the Stokes-Einstein-equation:
D =
k
B
T
6R
H
with k
B
the Boltzmann constant, T the temperature and the viscosity of the solution.
R
H
is the hydrodynamic radius of the particle which is assumed to be spherical. For
particles without sharp borders the hydrodynamic radius can differ enormously from the
real radius.
An illustrative example for these equations is the following: When the particle size in-
creases, the motion becomes slower and therefore the diffusion coefficient decreases. We
can see this as an increased decorrelation time in the spectrum.
5.1.3
Polarized Dynamic Light Scattering
More information about the particles can be extracted from polarized measurements: One
linear polarizer is placed between the laser source and the sample and lets only light pass
through that is polarized perpendicular to the scattering plane spanned by the incoming
and the scattered beam. This polarizer is necessary as the laser beam is not completely
polarized from the beginning. A second polarizer - the so called 'analyzer' - is inserted
between the sample and the detector.
For optically inactive spherical particles there is no polarization rotating in scattering
[KV83] meaning that if the two polarizers are perpendicular (so called VH mode), the
detected intensity is zero. This is not the case for elongated particles, such as cylinders or
spheroids, even if they are randomly orientated.
Now two processes are responsible for the decorrelation of the light: The translation of
the particles - as described above - as well as the rotation of the particles. One can
distinguish between two diffusion coefficients, the 'translational diffusion' D
t
and the
'rotational diffusion' D
r
, that depend among others on the axis ratio of the particle.
52
5.2. UNPOLARIZED MEASUREMENTS
5.1.4
Apparatus
First measurements were done with a Triton Hellas Axios 150 instrument (Triton Hellas,
Greece). This setup consists of a diode laser emitting at 658nm with a power of 35mW
that could be attenuated with different filters. A photodiode was positioned to detect the
scattered light at 90 degrees. Between the sample holder and the photodiode an aperture
guaranteed for a narrow laser beam. The autocorrelator used hat a dead time of 11.6ns.
For polarized measurements, linear polarizers of Thorlabs GmbH, Germany, were used. In
the depolarized measurements the incoming beam was not perfectly linearized. The major
drawbacks of this setup are the insufficient laser power, the disability to change the scat-
tering angle as well as a suboptimal software, that did not allow to detect low scattering
intensities. Therefore further measurements were conducted with an ALV-Langen setup
(ALV Vertriebsgesellschaft mbH, Germany) consisting of a 150mW HeNe laser. The angle
could be automatically varied between less than 30
o
and 155
o
. Polarizers were already
integrated in this setup.
5.2
Unpolarized Measurements
Several UV- as well as Tollens-metallized oligonucleotide solutions were analyzed by un-
polarized Dynamic Light Scattering. These measurements were intended to give first in-
formation about the particle sizes available in solution.
5.2.1
Experimental Results
for the analysis of DLS data, several methods of varying complexity are available. Two of
them shall be discussed in the following and be applied to the measured data.
Single Exponential Analysis
Assuming a monodisperse sample, the decorrelation is caused by only one correlation
time. The correlation graphs can therefore be fit with single exponential decay functions:
I(t) = f ·
1
1 +
2
/t
2
+ C
I is the intensity measured by the detector, f and C are constants and is the correlation
time (for reference, see e.g. [Ran05], page 153. The fitting procedure was done with the
program Origin that executed a Levenberg-Marquardt iteration. The hydrodynamic radius
can then be calculated as
R
H
=
k · T · q
2
·
6 · ·
53
CHAPTER 5. DYNAMIC LIGHT SCATTERING
(a) 10min UV
(b) 45min UV
Figure 5.5: Measured VV DLS data for a 96bp dsDNA, UV-metallized, with an Ag-to-base ratio
of 1:1. One can see that the deviations from the single exponential decay fit increase with time.
The hydrodynamic radii, that can be calculated from the decay time of the fit function, are in
the range of several tens of nanometers and therefore much larger than expected
UV-Metallized DNA
UV-Metallized samples with a silver-to-base ratio of 1:1 and various lengths and irradia-
tion times were investigated in the Triton-Hellas setup, the data was analyzed by single
exponential analysis.
The resulting hydrodynamic radii range from 33nm for 23mer DNA that was irradiated
for 10min up to 46nm for 96mer DNA, also irradiated for 10min. When increasing the ir-
radiation time from 10 to 45min, the radius for the 96mer DNA remains almost constant,
whereas the radius for 23mer DNA increased to 56.5nm. The deviation of the measured
data from a single exponential fit function increases with irradiation time. Typical mea-
sured data as well as the single exponential decay fits are shown in fig. 5.5.
Both parameters varied, the DNA length as well as the irradiation time, seem to have
an influence on the hydrodynamic radius. Comparing the outcome with the UV/vis spec-
trometry results that are given in 4.2.7, it can be seen that the particle sizes are much
larger than expected. However, it is known that large particles contribute much stronger
to the scattered intensity (d
6
) than small particles. Therefore, the influence of a few large
particles which are present in solution might dominate over the contribution from small
particles. Hence, a more complex analysis method, that accounts for more than one par-
ticle size, is necessary.
Tollens-Metallized DNA
Several Tollens-metallized DNA samples of various lengths, modification densities and
silver concentrations were prepared and stored overnight before the light scattering was
measured and analyzed by the single exponential decay method.
The measurement, done with the Triton-Hellas-setup, showed significant scattering in-
54
5.2. UNPOLARIZED MEASUREMENTS
Figure 5.6: VV DLS data of a Tollens-metallized 30-3 sample, 1.5mM Ag, stored overnight. A
single exponential decay fit yields a hydrodynamic radius of several tens of nanometers which is
much higher than expected.
tensity only for medium silver concentrations between about 0.32mM and 16mM, also
depending on the DNA length. The hydrodynamic radii calculated from the fit single ex-
ponential decay, range from 56nm to 104nm without any observable tendency. One typical
spectrum as well as the single exponential decay fit is shown in fig. 5.6.
These results can be explained as follows: For silver concentrations below 0.32mM the
produced particles are too small to scatter light efficiently. For silver concentrations above
16mM, the particles might have aggregated to particles larger than 200nm. These particles,
however, are filtered out before the measurement. Thus, a sufficient scattering intensity
is only reached for an intermediate silver concentration. Thus the results achieved by
single exponential analysis of the DLS data support the spectrometric result that Tollens
metallized samples aggregate overnight. The results for the radii are of the same order of
magnitude as the single exponential decay results for UV-metallized samples. They are,
however, roughly a factor of two larger. This agrees with the spectrometric results, where
plasmon peaks of Tollens-metallized samples were red-shifted compared to UV-metallized
samples, corresponding to an increase in size.
Like for the UV-metallized samples, the hydrodynamic radii for Tollens-metallized samples
are, however, much larger than expected. Therefore a more sophisticated analysis method
must be applied to account for effects like polydispersity and various distinct particle
sizes. This is explained in the next paragraphs.
5.2.2
CONTIN Analysis
The previous analysis has shown, that single exponential decay analysis of the results
is not sufficient. In principle, it is possible to extend the fit to include a sum of several
exponents, but this makes the fit unstable. Furthermore, the width of the size distributions
55
CHAPTER 5. DYNAMIC LIGHT SCATTERING
cannot be analyzed by this simple method. Therefore a commercial DLS analysis program
of ALV, Germany, was used.
For measurements done with the ALV setup, a CONTIN analysis of the particle sizes is
possible. The CONTIN algorithm, that was invented by S. Provencher [Pro82], uses an
inverse Laplace transformation for determining the particle sizes. For the calculated size
distributions, three different types of weighting can be chosen: The unweighted signal, also
known as intensity-weighted signal, is proportional to d
6
, according to Rayleigh, as shown
in 2.2.1. In the mass-weighted spectrum, the intensity is proportional to d
3
. This weighting
type is important for comparing the DLS data with the UV/vis data, as the plasmonic
absorption intensities of nanoparticles are also proportional to d
3
in a first approximation.
The number-weighted spectrum shows the number of particles in solution. All spectra are
normalized to the peak of maximum intensity.
UV-metallized samples of various parameters were freshly prepared and investigated with
DLS. The CONTIN analysis detected up to three or four particle sizes: A 23mer DNA
sample with an Ag-to-base ratio of 2:1 and 10min irradiation shows peaks at 0.5nm, 1-
2nm, 8nm and a small but broad distribution for larger particle radii. The normalized,
mass-weighted spectrum can be seen in fig. 5.7a), the number weighted spectrum can be
seen in 5.7b).
When the irradiation time is increased to 90min the peak below 1nm shifts to larger radii
as shown in fig. 5.8. The amount of particles having radii larger than 10nm, however,
increases.
Increasing the silver amount to a silver-to-base ratio of 20:1 increases also the particle
sizes, as shown in 5.9. The amount of particles having radii larger than 10nm, however,
increases. The 0.5nm particles increase to 0.7nm, the 1-2nm particles grow to a radius of
2-3nm. The peak around 10nm vanishes completely, another peak occurs at 40nm. The
intensities of larger particles have significantly decreased in comparison to the peak below
1nm. The second-smallest peak has become much narrower.
When longer oligonucleotides with 96 instead of 23 basepairs are metallized, the average
radius of particles smaller than 1nm shifts to 0.7 and drastically loses intensity, whereas
the 1-2nm peak has gained intensity. Moreover, peaks at 5nm and 30nm appear. This is
shown in fig. 5.10
Experiments with an unbuffered solution were done with 48mer DNA. The DNA with
a silver-to-base ratio of 10:1 was metallized in an aqueous solution with 10mM NaNO
3
at pH 8 - and afterward examined with DLS. The result of a CONTIN analysis can be
seen in fig. 5.11. Due to a different DNA length the CONTIN results cannot be directly
compared to one of the samples metallized in 10mM Tris buffer. In the unbuffered sample
no peak occurs for radii below 1nm - such a component exists, however for 23bp DNA with
a Ag-to-base ratio 20:1 as well as for the 96bp sample. The unbuffered sample exhibits
two more particle radii, at 9nm and at 50nm. These peaks are red-shifted compared to
the samples metallized in Tris buffer. This increase in particle size is in accordance with
UV/vis absorption spectra: When DNA was metallized in an unbuffered solution, the
56
5.2. UNPOLARIZED MEASUREMENTS
(a) mass-weighted spectrum
(b) number-weighted spectrum
Figure 5.7: Mass-weighted and number-weighted size distributions of metallized 23mer DNA
with a silver-to-base ratio of 2:1 and 10min irradiation
57
CHAPTER 5. DYNAMIC LIGHT SCATTERING
Figure 5.8: Comparison of mass weighted size distributions of metallized 23mer DNA with a
silver-to-base ratio of 2:1 and 10min and 90min irradiation
Figure 5.9: Comparison of mass weighted size distributions of metallized 23mer DNA with silver-
to-base ratio of 2:1 and 20:1 and 10min irradiation.
58
5.2. UNPOLARIZED MEASUREMENTS
Figure 5.10: Comparison of mass weighted size distributions of metallized 23mer and 96mer
DNA
Figure 5.11: Mass-weighted size distribution in an unbuffered NaNO
3
solution at pH 8 containing
metallized DNA. Three particle sizes are available in the sample.
59
CHAPTER 5. DYNAMIC LIGHT SCATTERING
plasmon peak shifted to long wavelengths indicating an increase in size. The assumption,
that this is accompanied by an increase in the size distribution, indicated by a broad
peak in the UV/vis spectrum, is not confirmed by the DLS data: The widths of the size
distributions is similar to the Tris buffer samples.
All spectra discussed were recorded at 90
o
. Varying the angle between 30
o
and 155
o
changed the results only slightly.
The CONTIN analysis showed that at least three different particle sizes are present in
solution, whereby the smaller ones are the most frequent. The particle sizes broaden and
shift to bigger radii when the irradiation time is prolonged. This is in accordance with
absorption spectrometry results that the photoinduced reduction is not finished before
60min. In that time longer irradiation time means more silver deposition (and thereby
decreasing the axis ratio of the prolate particles). A larger amount of silver ions induces a
slight shift to larger radii and significantly narrows the spectrum. A possible explanation
is that the silver saturation value has still not be reached as it is also indicated by the
absorption spectrometry results: If it is further assumed that photoinduced silver deposi-
tion is a self-limited process leading to stable nanoclusters, more silver means that more
clusters with the final size can be built. This agrees with the fact that the second smallest
peak has not only been narrowed but also shifted to a larger radius. Using longer DNA,
the smallest particle size becomes less frequent, whereas the other peaks gain intensity.
This meets the expectations that larger DNA come along with larger nanoparticles.
5.2.3
Temporal Evolution
DLS measurements can be used to monitor the growth of nanoparticles with the Tollens
method. A sample containing 1µM of 15-3 DNA modifications was prepared and filtered
and 16mM silver in a Tollens reagent were added. Every five minutes DLS a spectrum
was taken with the ALV setup. They are shown in fig. 5.13
The spectra show a large-radius contribution that can be attributed to unidentified back-
ground in the unfiltered Tollens solution and that provides for enough scattering intensity
to record correlation graphs. After 5min a very small contribution of 10 to 20nm particles
develops. Five more minutes later, the peak intensity has increased and the asymmetric
peak has shift to 10nm. 15min after addition of the Tollens aliquot the peak position is
at 7nm. In the following time it fluctuates between 3 and 8nm. Presumably, these fluc-
tuations reflect the limited measurement accuracy. The scattering intensity increases by
a factor of two during the growth process. Fig. 5.12 shows the development of the peak
intensity and the radius with time. The CONTIN analyzed spectra of the growth process
are shown in fig. 5.13.
In comparison to the UV-metallized samples, only one peak size below 100nm is present in
the solution. The width of the size distribution is much larger than for the UV-metalized
sample, it ranges from 2nm up to several tens of nanometers. This is in accordance with
the broad plasmon peaks in the absorption spectra.
60
5.2. UNPOLARIZED MEASUREMENTS
Figure 5.12: Temporal evolution of the peak intensity and the peak position for 15-3 DNA to
which 16mM of silver in a Tollens reagent were added
In conclusion, DLS measurements can be used to monitor the growth of particles in
solution. Already 5min after addition of the Tollens reagent first clusters have been built.
The particle size decreases with time. One possible explanation is that ions and silver
diamine complexes in solution are electrostatically attracted to the negatively charged
DNA backbone and increase its hydrodynamic radius. The more silver is reduced, the
more compensated the charge of the DNA is and therefore less particles are attracted: the
large diffuse ionic shell is replaced by smaller silver clusters. Silver clusters scatter visible
light efficiently, this makes the intensity increase. Further measurements have to be done
to confirm or confute this argument.
5.2.4
Reference Samples
A reference solution containing just Tollens but no DNA exhibited almost the same scat-
tering intensity as the sample with DNA. Particle sizes of 200nm and 2µm were also
present in this sample but smaller particles did not develop within 60 minutes.
A solution containing only DNA without silver clusters does not scatter at all in the
Triton-Hellas setup, even when its concentration is increased by a factor 50 to 50µM.
Triton-Hellas measurements on freshly prepared Tollens-metallized samples do not provide
enough scattering intensity for analysis.
The Triton-Hellas setup could not detect scattering of irradiated reference samples con-
61
CHAPTER 5. DYNAMIC LIGHT SCATTERING
Figure 5.13: Temporal evolution of nanoparticles during Tollens-metallization of 15-3 DNA to
which 16mM of silver in a Tollens reagent were added
taining buffer solution and silver but no DNA. The more sensitive ALV setup could not
detect more than about 3 counts per second (cps), which is three times the background
scattering signal. This small signal can be attributed to particles with a radius of 30 to
40nm, probably unspecified silver deposition due to dirt in the solution. Smaller particles,
as they were present in all metallized DNA samples, could not be detected. For Tollens
reference samples the situation is more complicated: A fresh undiluted Tollens solution
does not scatter when analyzed by the Triton-Hellas setup. When ammonium evaporates
from the solution, the silver diamine complexes are dissolved and unsoluble silver oxide
falls out again. These silver oxide particles are able to scatter light. In samples used for
metallization this effect is less important as the Tollens solution is diluted by a factor
10 to 100. This problem can be easily overcome by preparing the Tollens reagent freshly
before the measurement.
Sensitivity of the apparatus
Silver spheres with a radius of 10nm ±2.5nm were provided by BBI/Tedpella at a con-
centration of 1.2nM. Without any filter, the scattering intensity was 7 to 11cps for the
Triton-Hellas- and 20cps for the ALV-setup at a laser attenuation of roughly 50%. A
CONTIN analyzed spectrum can be found in the appendix B; it shows that the evaluated
diameter corresponds to the expected one. At the Triton-Hellas setup an intensity of 7cps
or larger is necessary to record data, at the ALV-setup 1cps is sufficient.
62
5.3. POLARIZED MEASUREMENTS
5.3
Polarized Measurements
For polarized measurements the theory is more complex and there is no widely accepted
algorithm to calculate the dimensions of the nanorods. Thus I have performed simulations
with the program Matlab: Using the formulas described below I calculated the correlation
times for various axes' lengths for both orientations of the polarizers. The experimental
values can then be compared to the theoretical ones and the length and diameter can be
determined in principle.
5.3.1
Simulations
Garcia de la Torre and Tirado developed formulas for the translational and rotational
diffusion coefficients of prolate ellipsoids [TG80]:
D
t
=
k
B
T
3a
(log(a/b) + 0.312 + 0.565 · (b/a) + 0.100 · (b/a)
2
)
D
r
=
3k
B
T
a
3
(log(a/b) - 0.662 + 0.917 · (b/a) - 0.05 · (b/a)
2
)
whereas a is the long and b the short axis of the ellipsoid. D
t
has the unit m
2
/s, D
r
is
given in 1/s.
The prefactors show that the order of magnitude of the coefficients as well as the ratio
D
r
/D
t
mainly depends on the length of the long axis a.
Van der Zande [vdZDBP00] suggests and performs DLS measurements with parallel
polarizers (so called vertical-vertical -VV- mode) and with crossed polarizers (vertical-
horizontal -VH- mode). According to van der Zande, the diffusion coefficients are related
to the decay times as follows:
-1
V V
= 2q
2
D
t
-1
V H
= 2q
2
D
t
+ 12D
r
For data acquired by the Triton-Hellas software, the right sides of these equations have
to be divided by 2 as the decorrelation of the electric field was analyzed instead of that
of the intensity. With these formulas for the electric field decorrelation I calculated the
diffusion coefficients in dependence of the long axis a and the short axis b in Matlab. The
result is shown in fig. 5.14 und 5.15.
Some general trends can be seen from the formulas above: The longer the particle, the
smaller the difference between
V V
and
V H
and the larger are
V V
and
V H
. The longer
63
CHAPTER 5. DYNAMIC LIGHT SCATTERING
Figure 5.14: Calculated correlation time in the VV mode in sec. The x and the y-axis show the
long and the short axis of the ellipsoid in m.
Figure 5.15: Calculated correlation time in the VH mode in sec. The x and the y-axis show the
long and the short axis of the ellipsoid in m.
64
5.3. POLARIZED MEASUREMENTS
(a) VV mode
(b) VH mode
Figure 5.16: Correlation time in the VV and VH mode dependent on the axis ratio. A particle
length of 10nm was assumed
the short axis b, the bigger the correlation times are as shown in fig. 5.16. The steepness
of this increase is different for the two modi, however.
Fig. and 5.16 shows the correlation time for different axis ratios. Therefore a particle
length of 10nm was assumed. The correlation time in the VH mode is more than two
orders of magnitude smaller than that of the VV mode.
5.3.2
Experimental Results
Only aspherical particles should possess a polarized contribution, for spherical particles
the scattering intensity for crossed polarizers should be zero. This expectation was checked
with commercial spherical Latex spheres with a radius of 40nm: The signal in the VV
mode is at least a factor 175 more intense than the signal in the VH mode, that is only
marginally above the background signal.
The metallized DNA samples, however have a significant polarized contribution, the VV
and the VH signals differ by a factor of 4 to 8. This indicates significantly aspherical
particles.
Single and Double Exponential Fits
To get a first idea about the decay times, the spectra were again analyzed in Origin. This
time also a double exponential decay is applied.
Single exponential fits for samples with UV-metallized DNA of various lengths provided
correlation times
V H
around 20µsec for 10min UV irradiation. As this value is compara-
tively high and differs from the VV-time by less then one order of magnitude, the results
cannot be quantitatively compared with the theoretical calculations. For the depolarized
measurements a CONTIN analysis has pointed out the fact that two particle sizes are
65
CHAPTER 5. DYNAMIC LIGHT SCATTERING
(a) 10min UV
(b) 45min UV
Figure 5.17: Measured VH DLS data for a 96bp dsDNA, UV-metallized, with an Ag-to-base ratio
of 1:1. One can see that the deviations from the single exponential decay fit increase with time.
For 45min irradiation two distinct correlation times can be calculated with a double exponential
fit function.
present in solution. Thus a single exponential decay function is not appropriate although
it matches well at first sight.
The VH-spectra of samples irradiated for 45min show a curve shape that deviates from
a single exponential decay. A double exponential decay function shows two correlation
times, that are separated by at least one order of magnitude: 23mer DNA has 180µs
and 7.1µs, 96mer DNA has 450µs and 20µs. These two correlation modes, both in the
VH-mode, indicate that at least two particle sizes are available in solution.
Typical VH-spectra for 10min and 45min irradiation time as well as their exponential fits
can be seen in fig. 5.17.
For Tollens metallized DNA stored overnight, the VH correlation times are between 88µs
and 130µs for single exponential analysis and therefore exceed the UV-values again. Fig.
5.18 shows typical measured data as well as the single exponential decay fit.
This simple analysis showed the presence of an aspherical term. Again, the calculated
decay times are much too long. For quantitative information, a CONTIN analysis is
necessary.
CONTIN Analysis
UV-metallized samples were measured with DLS in the VV and the VH mode. The CON-
TIN analysis of the decay times in both modes is shown in fig. 5.19.
For the 23mer and the 96mer with low silver concentrations and short irradiation time,
there is a dominant peak at 390µs in the VV mode. In the VH mode a peak at 17µs for the
23mer and 38µs for the 96mer becomes more dominant. The peaks in the VH mode are
broadened. For the 23mer a peak at 1.2µs, for the 96mer a peak at 4.2µs occurs in the VH
66
5.3. POLARIZED MEASUREMENTS
Figure 5.18: VH DLS data of a Tollens-metallized 30-3 sample, 1.5mM Ag, stored overnight.
The time-constant yielded is 2.3 times faster than in the VV mode.
mode. When the irradiation time is longer, the peaks in the VH mode are also broadened:
Only one peak at 300µs - instead of 380µs in the VV mode - and some shoulders are
distinguishable. For the sample with larger amount of silver the long decay time does
not depend on the mode. The peaks at shorter decay times, however, are broadened and
shifted to shorter times. A short decorrelation time, 0.2-0.3µs occurs. The intensity of the
short time peaks is enhanced in comparison to the VV spectrum.
The CONTIN analysis showed non-vanishing VH spectra indicating aspherical particles.
At least three different decorrelation times can be read off the VH spectra correspond-
ing to various particle sizes or shapes. Compared to VV spectra, the intensity of short
decorrelation times is enhanced; this is in accordance with the results of the simulations.
Before the data can be compared to the simulations and before conclusions can be drawn,
the peaks of one particle size in the VV and VH mode must be correctly to each other.
To do so, measurements with HPLC
1
purified samples have to be done.
1
High Performance Liquid Chromatography
67
CHAPTER 5. DYNAMIC LIGHT SCATTERING
(a) 23mer, Ag-base 2:1, 10min UV,
(b) 23mer, Ag-base 2:1, 90min UV
(c) 23mer, Ag-base 20:1, 10min UV
(d) 96mer, Ag-base, 1:2, 10min UV
Figure 5.19: CONTIN analysis of the VV and VH decay times for various UV-metallized samples.
68
Chapter 6
Stabilization of the Nanoparticles
This chapter presents the fundamentals of nanoparticle stabilization and presents first
results achieved with two different stabilizers.
6.1
Fundamentals
In nanoparticle synthesis it is very important to stabilize the colloids against aggregation
due to attractive forces, e.g. van der Waals forces. This is especially valid for particles
made of silver, as its tendency for aggregation is larger than for many other metals, such
as gold. The stabilization can be realized either sterically or electrostatically.
For steric stabilization polymers such as PVP
1
are adsorbed or chemically bound to
the surface of the nanoparticle, as shown in fig. 6.1. Because of these organic shells the
Figure 6.1: Scheme of steric stabilization: The nanoparticles are covered by a shell of polymers.
particles cannot get close enough to be attracted by the short ranged van der Waals forces.
1
Polyvinylpyrrolidone, for the experiments done in this work polymers with an average weight of
10.000 Dalton, provided by Sigma-Aldrich, Germany, were used
69
CHAPTER 6. STABILIZATION OF THE NANOPARTICLES
When polymers, which are soluble in the solution, are used, no aggregation occurs among
the covered nanoparticles. A drawback of this method is, however, that the polymers,
that have weights of more than 10.000 Dalton
2
and are tens to hundreds of nanometer
long, immensely increase the hydrodynamic radius; this makes it difficult to identify the
'real' particle diameter by techniques like, for instance, DLS. According to [Blu03] organic
substances, i.e. PVP in this case, act as inhibitors for crystallization as they bind to the
growing clusters. Therefore the final size of clusters is reduced whereas the number of
clusters is increased.
In the case of electrostatic stabilization, the surfaces of the nanoparticles are charged.
These surface charges yield an arrangement of ions in solution around the particle to
build up a counter-field. A scheme of this is shown in fig. 6.2: The repulsive Coulombic
Figure 6.2: Electrostatic stabilization: The nanoparticles carry a surface charge, here drawn as
negative, counter ions in solution arrange around the particles and partly screen the charges.
The Coulomb forces among the particles prevent the particles from agglomerating.
forces stabilize the particles. The advantage of this stabilization method is, that it does
not change the radius of the DNA. One drawback is its sensitivity to the pH value in
the solution as well as the ionic strength. Most frequently sodium citrate is used as an
electrostatic stabilizer: This substance also acts as a reducing agent for silver ions. Therefor
synthesis conditions must be carefully chosen so that the clusters really grow on the DNA
pattern instead of being reduced homogeneously in solution.
6.2
Experimental Results
Measurements with platinum done by Mertig and Pompe [MP04] showed that a stabiliza-
tion of the metallized DNA is not necessary as no aggregation occurs. In papers dealing
with silver-DNA-structures, stabilization of the particles is simply not mentioned. The
authors typically just write, that clusters were heterogeneously grown on DNA while
homogeneous deposition in the surrounding medium is suppressed.
The charge of my DNA-silver-structures and thus their intrinsic electrostatic stabilization
of the nanostructures is unknown. DLS data indicate small, probably not agglomerated,
2
1 Dalton = 1 atomic mass unit = 1.66E-27 kg
70
6.2. EXPERIMENTAL RESULTS
particles as well as larger ones. AFM micrographs, presented in chapter 8, show linearly
agglomerated structures. It is not known, whether the particles already agglomerate in
solution or when being brought on substrate. UV/vis data do not allow for quantita-
tive information about the particle size. The large bandwidth of the plasmonic peaks of
Tollens-metallized samples indicate a very broad size distribution, the temporal evolution
of these peaks implies that the solutions are not stable for more than some hours. Thus a
stabilizer is recommendable. First experiments were done with the UV-metallized samples
as the measurement conditions are better controllable in this case. PVP with an average
molecular weight 10.000 Dalton and sodium citrate (all provided by Sigma-Aldrich, Ger-
many) were used as stabilizers.
When adding the stabilizers to the sample before the UV-irradiation, plasmon peaks at
414nm (sodium citrate) or 425nm (PVP), respectively occur even if no DNA is present in
solution. This is because of the reducing character of the two stabilizers. This problem does
not occur when the substances are added directly after the irradiation. The corresponding
spectra for these reference measurements are shown in the appendix B.
Measurements were carried out in aqueous solutions containing 5mM NaNO
3
, that were
titrated to pH 8.3 by NaOH. The sample contained 1µM 48mer dsDNA. This solution was
chosen as pre-measurements without stabilizers showed broad peaks at long wavelengths.
Thus the effect of a stabilizer might pronounced as for Tris buffer solutions with narrow
peaks at short wavelengths. The final concentration of sodium citrate in the solution was
0.01mM. PVP was used in a concentration of 5.1mg/l.
Fig. 6.3 shows the plasmon peaks for three different samples: One was unstabilized, to the
others sodium citrate or PVP, respectively were added after irradiation. The amount of
silver was 1mM, corresponding to a silver-to-base ratio of 10:1, the irradiation time was
10min.
For the unstabilized sample a plasmon peak at 454nm occurs. For the sodium-citrate-
stabilized sample, this peak has shifted to 437nm, for the PVP-stabilized sample it has
only slightly shift to 454nm. The unstabilized sample shows the least, the sodium-citrate-
stabilized sample shows the most intense peak. All spectra are so broad, that the band-
width cannot be determined. Comparing the peak shapes in the range below 700nm pro-
vides that the unstabilized and the sodium citrate stabilized samples have almost the same
shape, whereas the PVP-stabilized sample has a broader peak. At 350nm a quadrupole
shoulder can be seen, this is less pronounced for the sodium-citrate stabilized sample.
The reference samples shown in B and discussed above showed that also in presence of
stabilizers the nanoparticles are grown on DNA. The high peak intensities for the stabilized
samples indicate that this growth process is enhanced in the presence of sodium citrate
or PVP. The fact that the spectra change when stabilizers are added confirm that these
substances do interact with the silver nanoparticles. The blue shift as well as the less
pronounced quadrupole shoulder indicate that smaller particles are grown in presence
of sodium citrate. No such effect is seen for PVP. The red-shift can be attributed to a
raised refractive index around the particles due to the organic shell. As the spectrum even
broadens, indicating either larger particles or more polydispersity, no more experiments
71
CHAPTER 6. STABILIZATION OF THE NANOPARTICLES
Figure 6.3: Plasmonic peaks for unstabilized and stabilized samples. When sodium citrate is used
as a stabilizer the peak is shifted to the blue and the quadrupole shoulder is less pronounced.
PVP does not have any improvement compared with the unstabilized sample. The measurement
was carried out in a non-buffered alkaline aqueous solution, details can be found in the text.
were carried out with PVP. Further measurements concentrated on sodium citrate as a
stabilizer.
For DLS analysis 23mer DNA was metallized, whereby the silver-to-base ratio was 2:1, the
irradiation time was again 10min. For this measurement, again 10mM Tris buffer was used
to provide a better comparison with the samples discussed in 5.2.2. The mass-weighted
as well as the number-weighted CONTIN analyzed spectra are shown in fig. 6.4.
The 0.7nm peak and the 2nm peak in the mass-weighted spectrum of non-stabilized DNA
have merged to one broad peak. The 8nm has shifted to 20nm. The number-weighted
spectrum also shows one broad peak instead of two. This indicates that in Tris buffer the
sizes cannot be reduced sodium citrate. The size distribution even broadens. Although
there is no improvement for the particle size and distribution in this case, this experiments
show that sodium citrate interacts with the metallized DNA and influences the particle
size distributions.
Thus sodium citrate might be nevertheless a working stabilizer for the non buffered solu-
tion of the Tollens-metallized samples. The UV/vis data of the metallization in alkaline
water solution confirms that assumption. As sodium citrate can also act as a reducing
agent, the development solution containing formaldehyde might be replaced by one con-
taining sodium citrate. Sodium citrate will then act as a developer as well as a stabilizer.
More experiments have to be done.
72
6.2. EXPERIMENTAL RESULTS
(a) mass-weighted spectrum
(b) number-weighted spectrum
Figure 6.4: CONTIN analysis for citrate stabilized 23mer DNA, Ag-to-base ratio 2:1, 10min irra-
diation time. The mass-weighted and the number-weighted spectrum are shown. For comparison
the results for unstabilized metallized DNA are also included.
73
Chapter 7
Nanorings
The previous chapters deal with metallic nanospheres (from BBI/Tedpella) and nanorods
(metallized dsDNA). But the recognition properties of DNA allow for even more complex
particle shapes.
Figure 7.1: Base sequence of Molecular Beacons [Zuk03]: The complementary parts of the stem
hybridize while the intermediate sequence forms a loop-structure.
Nanorings can be produced by metallizing molecular beacons. These are single DNA
strands with a special base sequence that forms a stem loop structure [TK96] as shown
in fig. 7.1 [Zuk03]: One end of the single strand is complementary to the other one and
74
therefore the two ends hybridize while the strand sequence in between is 'bent' and forms
a loop.
Hybridization
To produce ring-like structures the stem must have hybridized.
The molecular beacons were labeled with the fluorophore Cy3 at one end. This makes it
possible to observe the hybridization status
1
: When the stem hybridizes, the fluorescence
increases by approximately 70%. This is a special feature of the used dye and linker system
and has been characterized previously [Ran05]. Thus the fluorescence of molecular bea-
cons in 10mM Tris buffer/50mM NaCl was measured in dependence of the temperature
(not shown here). The main contribution to the temperature dependence comes from the
intrinsic temperature-dependence of the fluorophore, whose signal decreases with temper-
ature. A kink in the curve indicates the melting temperature. The experimental values
63.6
o
C for the 50mer and 60
o
C for the 80mer differ by about 13
o
C from the theoretical
values 77
o
C for the 50mer and 73.7
o
for the 80mer calculated by M. Zuker's program
Mfold [Zuk03]. This might be due to the special experiment conditions, such as the buffer
or the concentration. As the effect is overlapped by the fluorophores intrinsic temperature
dependence, errors in evaluation of the values might also play a role. Nevertheless, theory
as well as experiments predict a stem-loop form for the used DNA at room temperature.
According to literature [MKT03] the intramolecular hybridization strength is larger than
the intermolecular one. Therefore the molecular beacon DNA prefers to assemble into
rings instead of hybridizing with a second strand to form a linear molecule.
Thus, molecular beacons in 10mM Tris-buffer/50mM NaCl are hybridized at room tem-
perature and form ring-like structures.
UV-Metallization
The UV-metallization was conducted in the same way as for the linear DNA. UV/vis
spectra showed a plasmon peak, see fig. 7.2.
For the 50bp molecular beacon a symmetric peak with FWHM of 108nm occurred at
412nm. The 80bp molecular beacon showed a 114nm broad peak at 416nm. The red shift
indicates larger particles, as it is expected for the larger molecule.
The Cy3 absorption peaks at 522nm and 557nm vanish during metallization, even before
the nanoparticle peak has developed. At the same time, the color of the solution changes
from red to yellow, indicating that the Cy3 fluorophore reacted with the added Ag. The
immense decrease of the base absorption at 260nm proves that silver is also deposited on
the DNA itself.
Unpolarized DLS measurements, analyzed with CONTIN as shown in fig. 7.3
result in three different particle radii, 2nm, 7nm and 40nm with the relative (mass
weighted) intensities 10:5:1. A low VV:VH intensity ratio of 5.4 points at asymmetric
particle shapes. This asymmetry in the ring-like particles arise from the stem-part of the
1
The hypochromism method is not applicable here: Because of the short stem length, the difference
in absorption between double and single stranded DNA is hardly detectable
75
CHAPTER 7. NANORINGS
molecule. The polarized data, however, differ significantly from that of comparable linear
metallized DNA: Only two correlation times, 10
-4
s and 10
-6
s, can be detected and they
are both extremely broads.
As a conclusion, the applied UV-metallization method also works for more complex DNA
structures. More measurements have to be done to yield information about the shape of
the metallized particles.
76
(a) 50mer
(b) 80mer
Figure 7.2: UV-Metallization spectra of a 50bp and a 80bp long molecular beacon for two
different silver concentrations and 10min irradiation time. Note: The concentration of DNA and
silver was only one fifth of the concentrations typically used for UV-metallization of linear DNA.
77
CHAPTER 7. NANORINGS
(a) Particle sizes
(b) Decay times
Figure 7.3: CONTIN analysis of DLS data yielded from UV-metallized 80bases ringlike DNA.
a) Comparison between ringlike and linear DNA b) Correlation times of ring-DNA in the VV
and the VH mode
78
Chapter 8
Atomic Force Microscopy
8.1
Fundamentals
Atomic Force Microscopy (AFM) is a technique to analyze the topography of areas up to
several micrometers. It is possible to visualize structure that have dimensions of only a
few nanometers.
For the investigations the tapping mode was used: A cantilever containing a silicon tip with
a radius of 10nm ('normal tip') or 2nm ('sharp tip') at its end, oscillates at roughly 300kHz,
a frequency near its resonance frequency, with a certain, fixed amplitude. As interactions
with the surface damp the amplitude, a control loop has to correct the distance between
the cantilever and the surface to provide for the constancy. The required readjustments
give information about the distance between the surface and the tip. Scanning of the
surface with the help of a piezoelectric crystal provides the surface topology. The principle
of AFM measurements is shown in fig. 8.1:
Important parameters to adjust are the proportional and integral gain, that regulate the
control loop, the amplitude setpoint, which determines the oscillation amplitude, and the
scan velocity.
Atomic Force Micrographs were taken with a Multimode and Dimension V instrument
(Digital Instruments Inc., Santa Barbara, USA). Tips were provided by Budget Sensors.
8.2
Sample Preparation
An essential point of AFM is sample preparation. As one of the requirements is a good flat-
ness of the surface, silicon wafers were used as substrates. There were two main challenges
for bringing the metallized DNA from solution to substrate:
79
CHAPTER 8. ATOMIC FORCE MICROSCOPY
Figure 8.1: Schematic of the working principle of an AFM from [Lin06]. The cantilever oscillates
while the surface is scanned with a piezoelectric crystal. Forces between the tip atoms and the
surface atoms change the cantilever's amplitude. A feedback loop compensates for these changes.
Its regulations give information about the surface topography.
8.2.1
Requirements
Salt Removal
The first challenge is the removal of salt from solution. As the salt concentration (typi-
cally more than 60mM) exceeded the DNA concentration by a factor of 60000, it would be
otherwise difficult to find and distinguish the DNA between the large salt crystals. Scan-
ning Electron Microscopy (SEM) images of metallized DNA deposited on a gold surface
without removing the salt can be seen in fig. 8.2.
Due to several reasons discussed in the previous chapters, the DNA metallization has to
be carried out in solutions containing at least some mM of salt. The salt has therefore to
be removed afterward. For this purpose several methods were applied: One was filtering
the solution before dispersing it on surface. This can be done with concentrators provided
by Millipore. These devices, designed mainly for buffer-exchange work as follows: The
solution is filled into a vial that is separated from a larger reservoir by a vertical mem-
brane, whose pore size has to be chosen in advance according to the molecular weight
of the metallized DNA. This concentrator device is centrifugated at 13.000 g for 45min.
Meanwhile the water with the salt molecules is pushed through the membrane. Due to its
size the metallized DNA cannot pass the membrane. A very small rest volume of about
30µl containing the metallized DNA remains. When diluting this with distilled water to
the initial DNA concentration, the salt concentration has decreased. After repeating this
procedure typically five times, the salt concentration was lowered by a factor of 100 000.
The advantage of this method is that the salt concentration can be reduced at will by
repeating the concentration cycles. The drawbacks are that the method is expensive as
well as time-consuming and that there is a risk of agglomeration of the DNA clusters when
they are close together in a small volume in absence of salt. A much faster method of
filtering is to use filter membranes provided by Millipore. This 25mm diameter membrane
80
8.2. SAMPLE PREPARATION
(a) SEM image of salt crystals
(b) Zoom in SEM image of salt crystals
Figure 8.2: SEM images of 2µl of a solution of metallized DNA on a gold surface without removal
of the salt: The surface is dominated by large salt crystals. The sample consists of an aqueous
solution with 10mM NaNO
3
0.2µM 48bp dsDNA, 1mM AgNO
3
. The irradiation time was 45min.
These micrographs indicate the necessity for salt removal.
with pore sizes of 200µm floats on the surface of a water reservoir. Droplets of the solution
with a volume of about 10µl are distributed on the membrane. Due to osmotic pressure
the salt molecules pass the membrane and diffuse into the much larger water reservoir. Be-
cause of their much larger size, the metallic DNA particles remain mostly in the droplets.
After about 25min the droplets are taken off the membrane again using pipettes. This
method is much faster than the concentrator centrifugation. But the main disadvantages
are, that the final salt concentration remains unknown and there is a higher risk to lose
the metallized DNA because of a larger pore size compared to the concentrators.
First dispersing the solution on a silicon wafer and then rinsing it with ultrapure water is
not applicable as the attractive forces between the metallized DNA and the silicon surface
are so weak, that the metallized DNA is also washed off.
Homogeneous Dispersion
The second challenge is to avoid agglomeration: First measurements using commercially
available silver colloids of BBI/Tedpella showed that silver particles have a strong ten-
dency to form large three-dimensional clusters as can be seen in Scanning Electron Mi-
croscopy (SEM) images like in fig. 8.3.
Experiments with the BBI silver spheres also indicated that simply diluting the solution
does not solve the problem as it only reduces the number of colloidal aggregates but not
their size.
Several deposition techniques tried were spin-coating and ultrasonication. For spin coat-
ing approximately 2µl of a desalted solution are dropped onto a wafer. The wafer is then
rotated on the spinner for about 15min until the water has evaporated. The silicon wafer,
81
CHAPTER 8. ATOMIC FORCE MICROSCOPY
Figure 8.3: Commercial 80nm silver spheres dispersed onto a silicon waver imaged by SEM,
magnification: 30000. The spheres aggregate to large three-dimensional clusters.
that is originally hydrophobic, has to be treated in advance to become hydrophilic: Oth-
erwise the drop does not cover the whole wafer and the risk to slide off the wafer during
duration is increased. For the second technique, 2µl of a desalted solution are dropped
onto a hydrophobic silicon wafer floating on a ultrasonic bath that induces movement
of the clusters in the solution and therefore decreases the agglomeration tendency. Both
methods yielded good results for commercial 20nm silver spheres.
8.2.2
Functionalization of Silicon Wafers
A different method to overcome the challenges described above is functionalization of the
wafer with molecules with sticky ends. A drop of the solution is then brought in contact
with the functionalized surface. Due to Brownian motion the clusters randomly hit the
sticky surface and get bound. Afterward the sample is rinsed with water to remove salt
residues.
The silicon surfaces are functionalized with (3-Mercaptopropyl)trimethoxysilane (MPTMS,
linear chemical formula HS(CH
2
)
3
Si(OCH
3
)
3
, provided by Aldrich) so that the silane
groups stick to the wafer surface whereas the sulfide groups can bind the silver clusters.
The fabrication follows in principle the procedure described by [PQGO02a], but we opti-
mized the parameters for our setup: After 15min of cleaning in a piranha solution (15ml
H
2
O
2
, 35ml H
2
SO
4
), a silicon wafer is dried by an argon flow and immediately positioned
on a grid in a 50ml polyproylene reaction tube with a 4mm hole on top. Due to an ar-
gon flow of 2l/min the evaporation rate of 20µl MPTMS, placed just below the wafer, is
enhanced. The molecules are directed upward by the Argon and a turbulent flow makes
some of the molecules settle down on the wafer surface (facing upward, i.e., away from
the MPTMS droplet) forming a self-assembled layer. This very fast and easy technique
to functionalize silicon wafers is schematically shown in fig. 8.4.
82
8.3. RESULTS
Figure 8.4: a) MPTMS molecules are evaporated and hit the cleaned silicon surface. Some of
the methoxy-groups dissolve from the rest of the molecule.
b) The Silicon atoms can bind to oxygen atoms in the water layer on the surface.
After 30min of functionalization, the wafers are put in an ultrasonic bath for 30min
to remove unspecifically bound MPTMS molecules and to break weak hydrogen bonds
between the MPTMS-molecules. The molecules then rearrange and form a covalently
bound network.
AFM images of the functionalized wafers are shown in Appendix C.
Solution containing metallized DNA is placed on a functionalized wafer such that it covers
the whole surface. When a silver cluster on the DNA, that undergoes diffusion, randomly
hits a sulfur atom, it gets bound to it. After some hours, the whole sample is washed with
water to remove salt. It is dried with nitrogen and then investigated by AFM. Afterward
the sample is stored in vacuum.
8.3
Results
8.3.1
Ultrasonication
Fig. 8.5 shows a silicon waver onto which UV-metallized 23mer dsDNA was dispersed.
The dispersion was done by putting a 2µl drop of metallized DNA solution onto a waver
and keeping the waver in an ultrasonic bath until the water was evaporated.
Wires with a length of several hundred nanometers and a diameter of approximately 15
to 20nm can be seen all over the AFM image 8.5. The green arrows point to two very
pronounced structures. Section analysis (not shown here) indicates a substructure of the
wires: They consist of 10 to 16nm long subunit as shown schematically in fig. 8.8. As they
83
CHAPTER 8. ATOMIC FORCE MICROSCOPY
Figure 8.5: Metallized 23mer dsDNA dispersed onto a silicon waver, showing parallel agglomer-
ation
do only appear in solutions containing metallized DNA but not in reference solutions, they
can be assigned to metallized DNA molecules that are agglomerated. The agglomeration
is not random but linear which might be traced back to dipole interactions among the
nanoparticles. The structures are not distributed homogeneously on the waver, they are
located to a very small area. An explanation for this is, that due to surface tension the
drop carries the DNA along when shrinking while evaporation.
8.3.2
MPTMS-Silanization
The silanization parameters were optimized to a MPTMS amount of 20µl, an Argon flow
of 2l/min and a silanization time of 30min and 30min in an ultrasonic bath, see Appendix
C. Afterward 20µl of the solution with the metallized DNA was put on the surface and
removed again several hours later. A measurement series, varying the application time,
showed that already after 2min DNA has bound to the silane layer.
Fig. 8.6 shows typical AFM images for UV-metallized 48mer dsDNA: Therefore 1µM
dsDNA was hybridized in aqueous 50mM sodium nitrate solution, that was titrated to
a pH between 7 and 8 by addition of NaOH. The silver-to-base ratio was 100:1, the
irradiation time was 10min. UV/vis spectra taken before the AFM investigation showed
a plasmon peak at 440nm.
The 5µm x 5µm scan, fig. 8.6 a), shows large clusters in the range of some hundreds
of nanometers. Next to them some lower, chain-like structures can be seen. As these
structures are hardly detectable due to the high z-range, the interesting part with many
chain-like structures is encircled with green color. These long, thin structures can be found
all over the samples, the images b) and c) show regions with high concentration. Fig. c) is
a zoom in these structures. Noise limits the maximum achieved resolution. Therefor the
composition of the long structures cannot be resolved. Thus, all AFM images show wire-
like structures with a diameter of ca. 14nm and lengths of several hundreds of nanometers.
The heights vary between 1.5 and 5nm. Some single particles are observable.
84
8.3. RESULTS
(a) Typical 5µm x 5µm scan
(b) Typical 1µm x 1µm scan
(c) Typical 1µm x 1µm scan
(d) Typical 340.5nm x 340.5nm scan
Figure 8.6: Typical AFM images of metallized 48mer dsDNA on functionalized silicon wafers.
Detailed information about the sample can be found in the text. The metallized DNA particles
agglomerate in one dimension and form long wires. Several single particles are observable.
These chain-like structures can be attributed to one-dimensional agglomerations of the
metallized DNA as shown in fig. 8.8, that can be induced by dipole-dipole-forces among
the nanoparticles. The few non-agglomerated particles have a length of 14 to 16nm and
a diameter of 7 to 9nm. As the AFM tip broadens the structures by roughly 2nm the
real dimensions are 12 to 14nm in length and 5 to 7nm in diameter. The length of the
unmetallized DNA is 16nm, its diameter is 2nm. Thus the metallized particles appear
slightly shorter. The broadening of the diameter during metallization meets the expecta-
tions. In fig. 8.6 c) structures with a length of 16nm and a width of 14nm appear. These
can be attributed to two metallized oligonucleotides, that are stacked together, see fig.
8.8b. Due to its rod-like geometry the height of the DNA structures is expected to equal
its width. This is true for the maximum heights measured as 5nm. As the MPTMS layers
are neither tight nor exactly smooth, it is possible that the DNA molecules sink in the
layer. This decreases the measured heights.
Fig. 8.7 shows an AFM image of DNA metallized in 10mM Tris buffer. As oligonucleotides
85
CHAPTER 8. ATOMIC FORCE MICROSCOPY
Figure 8.7: Typical AFM images of metallized 48mer dsDNA on functionalized silicon wafers.
Detailed information about the sample can be found in the text. Single particles, arranged as a
ring, can be seen, they are indicated by a green arrow. The height as well as the phase image
are shown.
again 1µM 48mer dsDNA was used, the silver-to-base-ratio was 5:1, the sample was irra-
diated for 10min.
The height as well as the phase image are shown. In phase imaging the phase of of
the cantilever oscillation during a scan is recorded. This provides information about the
composition of the material, its hardness, adhesion and many more properties and yields
images with an enlarged resolution compared to normal Tapping mode images. Details
about this powerful technique can be found in [BP04].
Whereas in the height image only one large homogeneous ring-like structure, indicated by
the green error, is visible, in the phase image single nanoparticles can be resolved: The
structure consists of about 50 nanoparticles with 10nm length and 6nm width.
The dimensions of these particles are in accordance with the values found in fig. 8.6. One
difference between the two samples is the way of agglomeration: Now the DNA sample
are agglomerated along their long axis, as can be seen in fig. 8.8c.
Fig. 8.9 shows another typical AFM image of the same sample. Now the agglomeration is
again parallel.
The section analysis shows that the elongated structures are built of several clusters. The
width of these single clusters is between 5 and 10nm. This confirms the statement, that
these chain-like structures are built of metallized DNA molecules.
86
8.3. RESULTS
Figure 8.8: Schematic of different possibilities for one-dimensional agglomeration of metallized
DNA:
a) shows parallel agglomeration. This seems to be the most frequent type. Next to the strands
with lengths of hundreds of nanometers, some individual particles are observable.
b) shows clusters built of 2 individual parallel particles.
c) shows agglomeration along the particle axis
Figure 8.9: In the left part a typical AFM image of UV-metallized DNA (48mer, in Tris buffer,
silver-to-base 5:1, 10min irradiation) is shown. Many typical chain-like structures can be seen. In
the right part two section analysis along one of the chainlike structures are shown: The structure
is built of several smaller units with a width of 5 to 10nm
87
CHAPTER 8. ATOMIC FORCE MICROSCOPY
Figure 8.10: Height and amplitude AFM images of UV-metallized DNA that was deposited on
a piranha-cleaned silicon wafer by spin-coating. Large round particles as well as small elongated
particles can be seen.
8.3.3
Spin-Coating
2µl of UV-metallized DNA were deposited onto a rotating piranha-cleaned silicon wafer.
Fig. 8.10 shows a typical AFM image.
Large round particles with a diameter of 40 to 50nm are present on the surface as well as
small particles with a length of 12 to 14nm and a width of 7nm. Particle heights of 5nm,
7nm and 14nm can be found for the small particles.
The two particle sizes are in great accordance with the CONTIN analysis done for UV-
metallized DNA as presented in 5.2.2. The dimensions of the smaller particles are the same
as found for the MPTMS-functionalized samples. The height of the particles is, however,
increased, because there is no soft molecule layer where the particles can cave in.
Thus various UV-metallized DNA samples were deposited on silicon wafers. Three differ-
ent techniques - the ultrasonic bath method, spin coating and the use of functionalized
surfaces - were applied, whereby the last one achieved best reproducible results. Besides
some large spherical particle in the range of some tens to some hundred of nanometers,
small elongated particles can be found in solution. The larger particles are assigned to
unspecific silver depositions and agglomerations, whereas the smaller particles are met-
allized DNA molecules: They are 10 to 16nm long and 5 to 7nm broad. The measured
heights differ depending on the method of deposition. Assuming that the nanoparticles
can sink in soft molecular layers, all heights are in the range between 5 and 14nm. The
metallized DNA molecules tend to aggregate by building chain-like structures.
88
Chapter 9
Conclusion
Metallized DNA structures were produced by two different methods, photoinduced silver
deposition and aldehyde-modified Tollens metallization. Afterward they were character-
ized by optical techniques as well as atomic force microscopy.
All metallized DNA samples show characteristic plasmonic peaks between 405 and 440nm
in the absorption spectrum indicating silver particles in the nanometer range. These peaks
only occur in samples containing DNA (for the unspecific UV -metallization) or modified
DNA (for the Tollens-metallization) confirming that the nanoclusters are grown on the
template and not randomly in solution.
Conditions for a reproducible and efficient synthesis of nanoparticles have been found: The
use of a buffer solution provides reproducibility; and an alkaline environment enhances
the silver reduction rate.
Comparison of the peak positions and bandwidths of UV-metallized samples with calcu-
lated values from Mie theory indicate that the grown silver nanoparticles have radii of
only a few nanometers. Moreover, the spectra show that dsDNA can be metallized more
efficiently than ssDNA and yields slightly larger particles. The resulting particles, i.e. their
size distribution and intensity, also depend on the specific base sequence as experiments
with several samples with varying GC-content show. Therefore not only the fraction of
GC-bases influences the nanoparticle growth but also the neighboring bases. By decon-
voluting the absorption spectra to eliminate overlapping effects caused by the DNA and
nitrate absorption, two peaks can be determined. The low-intensity peak at short wave-
lengths is attributed to transverse plasmons, the high-intensity peak at long wavelengths is
attributed to longitudinal plasmons. This indicates an either rod-like or chain-like shape of
the metallized DNA molecules: This is in accordance with electrostatic calculations except
that the measured peaks occur at much shorter wavelengths than expected. Deconvoluted
absorption spectra show that the axis ratio of the particles can be in principle tuned
by varying one of the following parameters: length of the used DNA, amount of added
silver or irradiation time. An applicable theory allowing for quantitative determination
of the axis ratios is missing. Polarized Dynamic Light Scattering measurements quali-
tatively confirm an asymmetric particle shape. Unpolarized measurements show that at
89
CHAPTER 9. CONCLUSION
least three different particle sizes - hydrodynamic radii around 1nm, 8nm and 30nm - are
present in solution, whereby the smaller particles are more frequent. With increased DNA
length or extended irradiation time the particles become larger. The addition of more
silver produces slightly larger particles with a much narrower size-distribution indicating
that a stable cluster size has been reached.
The reaction rate of the Tollens-metallization process crucially depends on the amount of
ammonium present in the reagent: By increasing this amount, the reaction is slowed and
the plasmonic peak intensity, i.e. the number of particles or their size, is reduced. More-
over, the reaction speed depends on the silver concentration as well as the density of the
modified bases. The reaction can be accelerated and its maximum absorbance enhanced
by adding a development solution. Absorption spectra as well as DLS measurements in-
dicate that the solution is not stable and an aggregation of the individual clusters sets in
after some hours. Proof-of-principle-experiments with sodium citrate showed that it can
be possibly applied as a developer as well as a stabilizing agent. Absorption spectra of
Tollens-metallized DNA samples show a plasmonic peak that is broadened and red-shifted
compared to the spectra taken after UV-metallization. DLS spectra indicate a broadened
particle distribution compared to the UV/vis spectra. Despite from large clusters in the
micrometer range, which can be attributed to dirt in the solution, only one particle size
is present in the solution. Monitoring the growth process temporally with DLS, one can
see that first clusters build within few minutes. The average particle size decreases with
time whereas the intensity increases. Contrary to UV-metallized DNA, no second peak
appears after deconvolution. The reason might be that the plasmonic peak width is too
broad to resolve another peak nearby. DLS studies, however, confirm the rod- or chainlike
character.
Structural characterization was done by Atomic Force Microscopy: Particles with a length
of 14nm and a diameter of 5nm were found and attributed to the metallized DNA. The
length is shorter than the nominal DNA length of 16nm, the diameter is broadened from
2nm to about 5nm. Most particles were aggregated parallel to one-dimensional structures
with lengths of several hundreds of nanometers
90
Chapter 10
Outlook
Metallization Environment
The requirements for the solution in the UV-metallization are the ability to keep the
pH constant, a slightly basic pH value and a preferably low salt concentration. Tris
buffer/NaCl matches most of these claims, but nevertheless the amount of added silver is
limited by precipitation of silver chloride. Thus a buffer that fulfills the requirements and
that does not contain any compounds building precipitations with silver has to be found.
Stabilization
First experiments showed that stabilization with sodium citrate works in principle. In
further experiments the optimum parameters for the stabilization have to be found out.
Development Solution
Adding a development solution to Tollens-metallized samples turned out to be a promising
method for enhancing the absorption peaks. The outcome are blue-shifted, comparatively
narrow peaks that stand for small particles. DLS measurements with Development Solu-
tions are still pending. Moreover, different acids to stop the reaction in a controlled way
have to be tested. Sodium citrate might be used as a stabilizer as well as a developing
agent at the same time.
Separating the Particles
Recent DLS measurements confirmed the existence of up to three different particle sizes
in solution. The largest ones, that have radii of typically 30 to 50nm, are also present
in reference samples, whereas the smaller ones can be attributed to nanoclusters grown
on DNA. Although the small clusters are much more frequent, the effective signal is very
sensitive to the larger particles, as the absorption intensity is proportional to r
3
and the
scattering intensity even to r
6
. Therefore the particles shall be separated by HPLC and
afterward being analyzed again. The DLS correlation times in both polarization modes
can then be compared to the calculated values presented in chapter 5.3.1.
Structural Studies
AFM images gave first information about metallized DNA. Even better resolution can,
however, be achieved with High Resolution Transmission Electron Microscopy (HRTEM).
91
CHAPTER 10. OUTLOOK
The challenge to overcome in this method is the homogeneous distribution of the particles
on the TEM-grid.
More Complex Structures
Linearized DNA has extensively been studied in this thesis. Also, first experiments with
circular DNA have been made and analyzed by DLS as well as absorption spectrometry
showing commonalities as well as differences. In future, more complex DNA structures
shall be metallized.
92
Appendix A
DNA Sequences
Used DNA sequences:
23mer:
GCC GGT CCT GTT ACT TGT GGC GC
38mer: GCC GGT CCT GTT ACT TGT GGT CCT GTT ACT TGT GGC GC
48mer, seq1: TAG TCG TAA GCT GAT ATG GCT GAT TAG TCG GAA GCA TCG AAC
GCT GAT
48mer, seq2: TTT TTT TTT TTT TTT TTT TTT TTT TTT TTT TTT TTT TTT TTT
TTT TTT
48mer, seq3: GGG TGG GTG GGT GGG TGG GTG GGT GGG TGG GTG GGT GGG TGG
GTG GGT
96mer: TAG TCG GAA GCA TCG AAC GCT GAT TAG TCG TGA GCA CAT GGA CCT
GAT TAG TCG TAA GCT GAT ATG GCT GAT TAG TCG GAA GCA TCG AAC GCT
GAT
50mer MB: GGC CGC CGC CTT TTT TTT TTT TTT TTT TTT TTT TTT TTT TGG
CGG CGG CC
80mer MB: GGC CGC CGC CTT TTT TTT TTT TTT TTT TTT TTT TTT TTT TTT
TTT TTT TTT TTT TTT TTT TTT TTT TTT TGG CGG CGG CC
93
Appendix B
Reference Samples
B.1
Absorption Spectroscopy
Figure B.1: Absporption spectrum of 10mM Trisbuffer/HCl with 1mM AgNO
3
after 45min of
irradiation. The absorption is in the same range as the background of the samples with metallized
DNA. The spectrum is an overlap of two features: Around 300nm nitrate absorbs. Above ca.
300nm the absorption is raised due to unspecific silver deposition. In contrast to the metallized
DNA samples there is no plasmonic peak between 400 and 420nm.
94
B.1. ABSORPTION SPECTROSCOPY
Figure B.2: UV/vis spectra of 10mM Tris buffer solutions containing 1mM silver nitrate and
sodium citrate/PVP. The stabilizer was added before the 10min of irradiation. The spectra show
plasmonic peaks at 414/425nm.
Figure B.3: UV/vis spectra of 10mM Tris buffer solutions containing 1mM silver nitrate and
sodium citrate. The stabilizer was added directly after the 10min of irradiation. The spectra
show no plasmonic peak. The peak seen around 300nm can be attributed to the absorption of
nitrate.
95
APPENDIX B. REFERENCE SAMPLES
Figure B.4: UV/vis spectra of unmodified 38bp DNA in aqueous solution to which an aliquot
of a Tollens reagent containing 3.2mM Ag is added. Within 45min no peak occurs. Adding a
Development solution slightly increases the absorption.
B.2
DLS
[H]
Figure B.5: Commercial 20nm silver spheres in a concentration of 1.2nM were analyzed with
the ALV DLS setup. The graph shows a CONTIN calculation of the number weighted radius.
According to the manufacturer the deviations from the mean diameter are ±5nm.
96
B.3. AFM
B.3
AFM
(a) image1
(b) image1
Figure B.6: 5µm x 5µm AFM micrographs of a reference sample on DNA, z-range: 10nm
100µM of silver nitrate was added to standard buffer. Afterward 20µl of this solution were put
on a silanized silicon wafer and let react there for 30min. The sample was rinsed with pure
water, dried with nitrogen gas and analyzed by AFM. The micrograph shows some unspecific
silver deposition and maybe salt residues but no elongated structures as on the metallized DNA
samples.
97
Appendix C
Parameters of MPTMS-Silanization
Fig. C.1 shows a silicon wafer after 15min Piranha (35ml H
2
SO
4
, 15ml H
2
O
2
) cleaning.
Figure C.1: Piranha-cleaned silicon wafer, 1µm × 1µm scan, z-range 3nm
The MPTMS amount was varied between 10µl and 100µl. The higher the amount of MPTMS,
the more heterogeneous the surface of the sample is: The molecules form large aggregates on the
substrate. Thus the amount of MPTMS was set to 20µl.
As already mentioned by Pavlovic [PQGO02b] the number of molecules on the wafer surface is
independent of the silanization time. For the optimization of the parameters, the silanization
time was varied between 10min and 60min. For the experiments a medium time of 30min was
chosen. The argon flow was set to 2l/min.
Typical images recorded after silanization are shown in fig. C.2: Large cluster, that appear
on the relatively flat surface, can be identified as excess MPTMS molecules that have formed
aggregates. Holes with diameters of some tens of nanometers up to a few micrometers confirm
that self-assembled layers have been deposited on the wafer.
Measuring the depths of the holes in comparison to the upper layer one finds steps of roughly
2.0nm, 2.5nm, 3.5nm and 5.2nm. The theoretical length of the MPTMS-molecule is 0.77nm.
Granted that 5.2nm is the maximum layer thickness on the analyzed surfaces, one finds layers
of two (step = 2.0nm), four (step = 3.5nm) and six (step = 5.2nm). The holes might originate
98
(a) after silanization, 1µm ×
1µm scan, z-range 10nm
(b) after silanization, 5µm ×
5µm scan, z-range 10nm
Figure C.2: Surfaces directly after silanization. Functionalization parameters are20µl MPTMS,
2l/min Argon for 60min
from holes in water layer. Assuming that the water layer is about 0.4nm, the theoretical and
experimental values match quite well.
After silanization the wafers were put into the ultrasonic bath for 30min. Thereby the large
aggregates were removed from the surface. An image of a wafer after the ultrasonic bath can be
seen in fig. C.3.
Figure C.3: Silanized silicon wafer after 30min ultrasonication. Functionalization parameters
are 20µl MPTMS, 2l/min Argon flow for 30min. The image represents a 5µm × 5µm scan, the
z-range is 3nm.
99
Appendix D
Acknowledgments
An diese Stelle m¨
ochte ich noch all jenen danken, die mich w¨
ahrend meiner Diplomarbeit un-
terst¨
utzt haben und zum Gelingen dieser Arbeit beigetragen haben.
Prof. Dr. Gerhard Abstreiter danke ich f¨
ur die M¨
oglichkeit, an diesem großartigen Institut
arbeiten und meine Diplomarbeit an seinem Lehrstuhl anfertigen zu k¨
onnen.
Dr. Ulrich Rant gilt mein besonderer Dank: f¨
ur die ¨
außerst interessante Themenstellung, die
gute Arbeitsatmosph¨
are sowie den großen Freiraum, den ich bereits w¨
ahrend meiner Diplomar-
beit genießen durfte.
Den Verantwortlichen vom Lehrstuhl E25 danke ich daf¨
ur, dass ich sowohl ihr AFM als auch das
UV/vis Absorptionsspektrometer f¨
ur teilweise sehr lange Zeit belegen durfte. Dr. Paul Berberich
vom Lehrstuhl E10 danke ich f¨
ur die Bereitstellung des REM sowie f¨
ur die große Hilfsbere-
itschaft, die ich bei allen Fragen und Problemen erfahren durfte. Dr. Nikolaus Neumaier,
vom Chemie-Department der TU M¨
unchen, danke ich f¨
ur die Einf¨
uhrung am DLS-Ger¨
at. Dr.
Marie-Sousai Appavou und Dr. Wolfgang Doster vom Lehrstuhl E13 gilt mein besonderer
Dank daf¨
ur, dass ich kurzfristig ihren DLS-Aufbau verwenden konnte und f¨
ur die vielen netten
Diskussionen.
Martina Hofmann f¨
ur die große Hilfe mit der Silanisierung und den AFM-Aufnahmen sowie
die netten Gespr¨
ache zwischendurch.
Anna und Erika f¨
ur ihre geduldige Hilfe bei all meinen chemischen Fragen. Sebastian, Daniel,
Rocio, Jelena, Wolfgang und Kenji f¨
ur die tolle Arbeitsatmosph¨
are.
Ade Ziegltrum f¨
ur den Bau der Silanisierungsanlage sowie die Hilfe bei vielen Kleinigkeiten.
Irmgard f¨
ur ihre stetige Hilfsbereitschaft und deren Arbeit im Hintergrund viel zur guten
Arbeitsatmosph¨
are beitr¨
agt.
Christian Wirges vom Chemie-Department der LMU danke ich f¨
ur die vielen fruchtbaren
Diskussionen ¨
uber unsere Arbeit.
Konrad f¨
ur die vielen Gespr¨
ache fachlicher und nichtfachlicher Art.
Zu guter Letzt m¨
ochte ich meiner Familie sowie Friedrich f¨
ur all ihre Geduld und Un-
terst¨
utzung herzlich danken.
100
Bibliography
[AN05]
N.W. Ashcroft and Mermin N.D. Festk¨
orperphysik. Oldenbourg, 2005.
[BAF05]
L. Berti, A. Alessandrini, and P. Facci.
Dna-templated photoinduced silver
deposition. Journal of the American Chemical Society, 127(32):1121611217,
2005.
[BESBY98]
E. Braun, Y. Eichen, U. Sivan, and G. Ben-Yoseph. Dna-templated assembly
and electrode attachment of a conducting silver wire. Nature, 391(6669):775778,
1998. 30.
[BGM
+
06]
G.A. Burley, J. Gierlich, M.R. Mofid, H. Nir, S. Tal, Y. Eichen, and T. Carell.
Directed dna metallization.
Journal of the American Chemical Society,
128(5):13981399, 2006.
[BH83]
C.F. Bohren and D.R. Huffman. Absorption and Scattering of Light by Small
Particles. WILEY-VCH, 1983.
[BKRS06]
J. Boleininger, A. Kurz, V. Reuss, and C. Sonnichsen. Microfluidic continuous
flow synthesis of rod-shaped gold and silver nanocrystals. Physical Chemistry
Chemical Physics, 8(33):38243827, 2006. Boleininger, Johann Kurz, Andreas
Reuss, Valerie Soennichsen, Carsten 42.
[Blu03]
R. Blume.
Prof. blumes bildungsserver f¨
ur chemie.
http://dc2.uni-
bielefeld.de/dc2/tip/12 00.htm, 2003. [Online; accessed 14-June-2007].
[BP04]
K.I. Babcock and C.B. Prater. Phase imaging: Beyond topography. Veeco Ap-
plication Note, 2004. [AN11, RevA1].
[ES68]
Gunther L. Eichhorn and Yong Ae Shin. Interaction of metal ions with polynu-
cleotides and related compounds. xii. the relative effect of various metal ions on
dna helicity. Journal of the American Chemical Society, 90(26):73237328, 1968.
[Fli00]
T. Fliessbach. chapter 6. Spektrum, third edition, 2000.
[GBP89]
J. P. Goudonnet, J. L. Bijeon, and M. Pauty. A method for determining shape
and volume of ellipsoidal silver particles from absorbance measurements. Thin
Solid Films, 177:4957, 1989. 13.
[HBSG64]
R. Huisgen, H. O. Bayer, F. C. Schaefer, and H. Gotthardt.
New type of
mesoionic aromatic compound + its 1,3-dipolar cycloaddition reactions with
acetylene derivatives.
Angewandte Chemie-International Edition, 3(2):136,
1964. 10.
101
BIBLIOGRAPHY
[HG99]
A. Henglein and M. Giersig. Formation of colloidal silver nanoparticles: Capping
action of citrate. Journal of Physical Chemistry B, 103(44):95339539, 1999. 16.
[JC72]
P. B. Johnson and R. W. Christy. Optical constants of the noble metals. Phys.
Rev. B, 6(12):43704379, Dec 1972.
[KB04]
K. Keren and E. Braun. Sequence-specific molecular lithography towards dna-
templated electronics.
Chemical Engineering and Technology, 27(4):447452,
2004. 12.
[KV83]
U. Kreibig and M. Vollmer. Optical Properties of Metal Clusters. Springer, 1983.
[LES99]
S. Link and M. A. El-Sayed. Spectral properties and relaxation dynamics of
surface plasmon electronic oscillations in gold and silver nanodots and nanorods.
Journal of Physical Chemistry B, 103(40):84108426, 1999. 133.
[Lin06]
S. Lingitz. Electronic transport investigations on conjugated molecules based on
nanogap electrode devices, 2006.
[LKSH00]
T. Liebermann, W. Knoll, P. Sluka, and R. Herrmann.
Complement hy-
bridization from solution to surface-attached probe-oligonucleotides observed by
surface-plasmon-field-eahanced fluorescence spectroscopy. Colloids and Surfaces
a-Physicochemical and Engineering Aspects, 169(1-3):337350, 2000. 37.
[Man78]
G. S. Manning. Molecular theory of polyelectrolyte solutions with applications
to electrostatic properties of polynucleotides. Quarterly Reviews of Biophysics,
11(2):179246, 1978. 92.
[MBS
+
02]
J. J. Mock, M. Barbic, D. R. Smith, D. A. Schultz, and S. Schultz. Shape
effects in plasmon resonance of individual colloidal silver nanoparticles. Journal
of Chemical Physics, 116(15):67556759, 2002. 18.
[MKT03]
S.A.E Marras, F.R. Kramer, and s. Tyagi. Genotyping single nucleotide poly-
morphisms with molecular beacons, volume 212. The Humana Press Inc., 2003.
[MP04]
M. Mertig and W. Pompe.
Biomimetic Fabrication of DNA-based Metallic
Nanowires and Networks, chapter 17. WILEY-VCH, first edition, 2004.
[MPJG
+
06]
P. Mulvaney, J. P´
erez-Juste, M. Giersig, L.M. Liz-Marz´
an, and Pecharrom´
an C.
Drastic surface plasmon mode shifts in gold nanorods due to electron charging.
Plasmonics, 1(1):15571955, March 2006.
[MVH99]
J. B. Mills, E. Vacano, and P. J. Hagerman. Flexibility of single-stranded dna:
Use of gapped duplex helices to determine the persistence lengths of poly(dt)
and poly(da). Journal of Molecular Biology, 285(1):245257, 1999. 62.
[Noy05]
M. Noyong. Synthese und Organisation von Gold-Nanopartikeln mittels DNA.
PhD thesis, TU Dreseden, 2005.
[PQGO02a]
E. Pavlovic, A. P. Quist, U. Gelius, and S. Oscarsson. Surface functionalization of
silicon oxide at room temperature and atmospheric pressure. Journal of Colloid
and Interface Science, 254(1):200203, 2002. 18.
102
BIBLIOGRAPHY
[PQGO02b]
E. Pavlovic, A. P. Quist, U. Gelius, and S. Oscarsson. Surface functionalization of
silicon oxide at room temperature and atmospheric pressure. Journal of Colloid
and Interface Science, 254(1):200203, 2002. 18.
[Pro82]
S. W. Provencher. Contin - a general-purpose constrained regularization program
for inverting noisy linear algebraic and integral-equations. Computer Physics
Communications, 27(3):229242, 1982. 6.
[PWLW02]
F. Patolsky, Y. Weizmann, O. Lioubashevski, and I. Willner. Au-nanoparticle
nanowires based on dna and polylysine templates.
Angewandte Chemie-
International Edition, 41(13):23232327, 2002. 29.
[QD05]
L. T. Qu and L. M. Dai. Novel silver nanostructures from silver mirror reaction
on reactive substrates. Journal of Physical Chemistry B, 109(29):1398513990,
2005. 18.
[RABdA06]
I. Romero, J. Aizpurua, G. W. Bryant, and F. J. G. de Abajo. Plasmons in
nearly touching metallic nanoparticles: singular response in the limit of touch-
ing dimers. Optics Express, 14(21):99889999, 2006. Romero, Isabel Aizpurua,
Javier Bryant, Garnett W. Garcia de Abajo, F. Javier 44.
[Ran05]
U. Rant. Electrical manipulation of dna-layers on gold surfaces. PhD thesis, TU
Munich, 2005.
[Ric03]
J. Richter.
Metallization of dna.
Physica E-Low-Dimensional Systems and
Nanostructures, 16(2):157173, 2003. 73.
[SCB96]
S. B. Smith, Y. J. Cui, and C. Bustamante. Overstretching b-dna: The elas-
tic response of individual double-stranded and single-stranded dna molecules.
Science, 271(5250):795799, 1996. 29.
[SGHURS
+
05] A. Slistan-Grijalva, R. Herrera-Urbina, J. F. Rivas-Silva, M. Avalos-Borja, F. F.
Castillon-Barraza, and A. Posada-Amarillas. Classical theoretical characteriza-
tion of the surface plasmon absorption band for silver spherical nanoparticles
suspended in water and ethylene glycol. Physica E-Low-Dimensional Systems
and Nanostructures, 27(1-2):104112, 2005. 38.
[TG80]
M. M. Tirado and J. Garciadelatorre. Rotational-dynamics of rigid, symmet-
ric top macromolecules - application to circular-cylinders. Journal of Chemical
Physics, 73(4):19861993, 1980. 27.
[Tin60]
Ignacio Tinoco. Hypochromism in polynucleotides. Journal of the American
Chemical Society, 82(18):47854790, 1960.
[TK96]
S. Tyagi and F. R. Kramer. Molecular beacons: Probes that fluoresce upon
hybridization. Nature Biotechnology, 14(3):303308, 1996. 23.
[vdZDBP00]
B.M.I. van der Zande, J.K.G. Dhont, M.R. Bohmer, and A.P. Philipse. Col-
loidal dispersions of gold rods characterized by dynamic light scattering and
electrophoresis. Langmuir, 16(2):459464, 2000.
103
BIBLIOGRAPHY
[Wik06]
Wikipedia.
Dna
--
wikipedia,
the
free
encyclopedia.
http://en.wikipedia.org/w/index.php?title=DNAoldid=138027697,
2006.
[Online; accessed 26-October-2006].
[Wik07]
Wikipedia.
Persistence
length
--
wikipedia,
the
free
encyclopedia.
http://en.wikipedia.org/w/index.php?title=Persistencelen gth&oldid=134452654,
2007. [Online; accessed 15-June-2007].
[Zuk03]
M. Zuker. Mfold web server for nucleic acid folding and hybridization prediction.
Nucleic Acids Research, 31(13):34063415, 2003. 42.
104
Wissensquellen gewinnbringend nutzen
Qualität, Praxisrelevanz und Aktualität zeichnen unsere Studien aus.
Wir bieten Ihnen im Auftrag unserer Autorinnen und Autoren Diplom-,
Bachelor-, Master-, Magister- und Staatsexamensarbeiten sowie
Dissertationen, Habilitationen und andere wissenschaftliche Studien und
Forschungsarbeiten zum Kauf an. Die Studien wurden an
Universitäten, Fachhochschulen, Akademien oder vergleichbaren
Institutionen im In- und Ausland verfasst. Der Notendurchschnitt liegt im
Prädikatsbereich bei 1,5.
Wettbewerbsvorteile verschaffen Vergleichen Sie den Preis
unserer Studien mit den Honoraren externer Berater. Um dieses Wissen
selbst zusammenzutragen, müssten Sie viel Zeit und Geld aufbringen.
http://www.diplom.de
bietet Ihnen unser vollständiges
Lieferprogramm mit mehreren tausend Studien im Internet. Neben
dem Online-Katalog und der Online-Suchmaschine für Ihre Recherche steht
Ihnen auch eine Online-Bestellfunktion zur Verfügung. Eine inhaltliche
Zusammenfassung und ein Inhaltsverzeichnis zu jeder Studie sind im
Internet einsehbar.
Individueller Service Für Fragen und Anregungen stehen wir Ihnen
gerne zur Verfügung. Wir freuen uns auf eine gute Zusammenarbeit.
Ihr Team der Diplomarbeiten Agentur
Diplomica Verlag GmbH
Hermannstal 119k
22119 Hamburg
Fon 040-655 99 20
Fax 040-655 99 222
agentur@diplom.de
www.diplom.de
112 von 112 Seiten X- nach oben
Details
- Arbeit zitieren
- , 2007, Optical Properties of metallized DNA, Hamburg, Bedey Media GmbH, https://www.diplom.de/document/226331
Ähnliche Arbeiten
Deutschlands größter Sharing Community
Jetzt Zusammenfassungen, Skripte und Klausuren kostenlos downloaden!
Uniturm.de
Ihre Vorteile als Autor
Geld verdienen Für jedes verkaufte Exemplar erhalten Sie Autorenhonorar - bis zu 40%.Kostenlose Buchveröffentlichung
Als eBook-Studie im Original und / oder auf Wunsch zusätzlich als hochwertiges Fachbuch in unserem renommierten Buchverlag.
Renommee als Fachbuchautor
Präsentieren Sie sich als Fachfrau oder Fachmann in Ihrem Fachgebiet und machen Sie sich bekannt.
Persönliche Betreuung
Unser Lektorat wird Sie persönlich betreuen und ist auch telefonisch unter 040-655992-14 erreichbar.
Weltweit im Buchhandel
Ihr Buch ist weltweit im Buchhandel und Online-Buchhandel wie z.B. amazon erhältlich.
Jetzt Autor werden!
Kommentare