Chemical and Biochemical
Silica Surface Modifications
for the Development of
Medical Analyses
Submitted by Ivo Kretzers BSc, MSc,
to the University of Exeter as a thesis for the degree of Master of Philosophy
in Biosciences, September 2009.
This thesis is available for Library use on the understanding that it is copyright
material and that no quotation from the thesis may be published without
I. M. J. Kretzers proper acknowledgement.
I certify that all material in this thesis which is not my own work has been
identified and that no material has previously been submitted and approved
for the award of a degree by this or any other University.
Abstract
This report can be divided into two parts. The first part studying the
aggregation kinetics for citrate-reduced 15 nm gold nanoparticles at the native
silica and modified silica-water interfaces. At the native, negatively charged
silica-water interfaces a two-phase adsorption is observed: a pseudoLangmurian adsorption phase and, after an acid wash to remove the citrate
ligand from the adsorbed particles, a further pseudo-Langmurian adsorption
phase. A kinetic analysis of these phases shows an average adsorption rate
constant of (2.0 ± 2.5)  105 M-1 s-1 with no measurable desorption.
The second part of this report describes protocol development for tethering
DNA onto gold nanoparticles and the development of hybridization
procedures.
1-ethyl-3-(3-dimethylaminopropyl)
carbodiimide/N-
Hydroxysuccinimide (EDC/NHS) activation of a thioctic acid monolayer on a
gold nanoparticle shows the most promising method.
2
Acknowledgement
I would like to thank my supervisor for help and guidance regarding my
research over the last 2 years. Thanks also to Rouslan Olkhov for his help
and amusement in the laboratory. I would also like to thank Andy Murray for
his SEM measurements and analyses.
3
Contents
1
Introduction ............................................................................................... 5
1.1 Evanescent Wave Cavity Ring-down Spectroscopy (e-CRDS)........... 6
1.2 Gold Nanoparticles - colloidal suspensions ........................................ 9
1.3 Surface Plasmon Resonance ........................................................... 11
1.4 Adsorption Kinetics ........................................................................... 14
1.5 References ....................................................................................... 16
2 Experimental Methods ............................................................................ 19
2.1 The e-CRDS Instrument ................................................................... 19
2.2 Cavity Design .................................................................................... 19
2.3 Flow cell ............................................................................................ 24
2.4 References ....................................................................................... 26
3 Chemically Induced Assembly of Gold Nano-aggregates on Silica Oxide
Surfaces......................................................................................................... 27
3.1 Introduction ....................................................................................... 27
3.2 Experimental Methods ...................................................................... 31
3.3 Results .............................................................................................. 33
3.4 Discussion ........................................................................................ 45
3.5 Conclusions ...................................................................................... 46
3.6 References ....................................................................................... 48
4 DNA Hybridization .................................................................................. 50
4.1 Introduction ....................................................................................... 50
4.1.1 Current trends in DNA biosensors .............................................. 50
4.1.2 Kinetics....................................................................................... 52
4.1.3 DNA............................................................................................ 52
4.1.4 Binding Strategies ...................................................................... 54
4.2 Experimental Design ......................................................................... 58
4.2.1 The sequence ............................................................................ 58
4.2.2 sDNA adsorption to gold surfaces using EDC/NHS coupling ..... 59
4.3 Results .............................................................................................. 59
4.3.1 Adsorption and hybridization of sDNA onto gold nanoparticles .. 60
4.3.2 sDNA adsorption to gold surfaces using EDC/NHS coupling ..... 63
4.4 Discussion ........................................................................................ 66
4.5 Conclusions and Future Work........................................................... 66
4.6 References ....................................................................................... 67
4
1 Introduction
A major problem in the field of chemical biology is, at the moment, the
understanding of complex systems. The human genome consists of 20,000 to
25,000 genes of which some are expressed in a single human cell. 1 Each
expressed gene produces a protein which has a role in either the metabolism
or the structure of the cell. Owing to the large number of chemical reactions
involved in the metabolism of the cell, the need for research into rapid
analyses of low concentrations of molecules has increased in recent years.
The ability to detect molecules in small quantities (attogram amounts) can
provide a useful tool in fields ranging from medical to environmental science.
The 2D Attogram Surface Plasmon Imaging project seeks to produce a new
basic technology that will enable screening of large numbers of molecules at
low concentrations and, importantly, label free.2
Many other more sensitive techniques exist for detecting molecules, including
multi-photon ionization (MPI) detection3 and fluorescence detection.4 These
techniques are capable of detecting 10 molecules per cm 3, however all these
alternatives require complex detection schemes. The confocal microscope is
one of the most powerful techniques that is available. It is able to detect single
molecules within a biological system.5
Most of these rapid analysis techniques are, at the moment, based on
fluorescence measurements of a tagged molecule. 6 The disadvantage of this
is that for all molecules that will be eventually targeted, all chemistry needs to
be
developed.
Eventually
this
project
should
produce
standardized
functionalized surfaces which have no need for fluorescence tagging.
However, these surfaces should be produced in such a way that small
quantities of molecules can be measured, perhaps even a few hundred
molecules.7
The Attogram project2 will use surface plasmon imaging as its principal
technique to produce a label-free screening technology. A surface plasmon is
5
an electromagnetic wave that propagates along a metal surface. This wave is
excited by laser light. The plasmon has an associated plasmon field which
penetrates above the surface and is therefore very sensitive to changes in
local refractive index. Therefore, molecule binding events at or near the
surface, may be detected by observing a shift in resonance frequency of the
electromagnetic wave.8, 9
This chapter will provide a theoretical background to the investigations that
were carried out during these previous two years.
1.1 Evanescent Wave Cavity Ring-down Spectroscopy (eCRDS)
Cavity Ring-Down Spectroscopy (CRDS) is an ultra-sensitive laser absorption
technique that has evolved from previous spectroscopic techniques, used to
measure mirror reflectivities.10 CRDS has a number of advantages as a direct
absorption
technique
when
comparing
it
with
traditional
absorption
techniques.11 In principle CRDS is immune to variation in laser intensity. The
technique also benefits from an immensely long effective path length, which
may be achieved in a comparatively simple bench-top instrument. The
technique relies on the fact that a molecular sample can absorb
electromagnetic radiation. This can be generated by a laser (at a particular
wavelength).
Figure 1-1 gives a schematic representation of an eCRDS setup.
High reflectivity
Cw diode
laser
PMT
mirrors
Ring down decay wave
observed
by PMT
Dove prism
prism
Wave plate
Figure 1-1: Diagram showing the basic principles of cavity ring-down spectroscopy.
6
The use of a continuous wave laser is preferred over the use of a pulsed,
tunable laser. The laser pulse yields sufficient signal levels inside the cavity to
be measured at the detector. This enables very high-reflectivity mirrors to be
used with a resulting ringdown time of 230 μs.12 The equivalent path length
that corresponds to this ringdown time is 69 km, resulting in very high
sensitivity.12 Cavity-locked cw CRDS produces the most sensitive version of
the technique, with sensitivity levels within a factor of 10 of the ultimate shotnoise limit being achieved.13, 14 Cavity-locked cw-CRDS uses two orthogonally
polarised laser beams from a single external cavity diode laser (ECDL). 13 The
first laser beam is used to continuously lock a high finesse external cavity to
the output of the ECDL. The second laser beam is used to measure the
ringdown time of the cavity.13 This experimental setup is more difficult to set
up. It does however show a significant reduction in the shot-to-shot noise
levels in the ringdown time as well as a decrease in the optical feedback to
the laser.13
Utilizing this Cavity Ringdown Spectroscopy instrument (described above)
makes it possible to perform highly sensitive measurements of molecules in
the gas phase.15,
16, 17
A recent development is to use CRDS to probe
surfaces, thin films and condensed material using a standard CRDS setup
that incorporates an intra-cavity total-internal reflection (TIR).17,
18, 19
TIR
enables us to absorb molecules directly at the interface between silica and the
medium above (see Figure 1-2).
7
Figure 1-2: Diagram showing a total internal reflection event within a stable cavity.
When liquid is introduced directly into the cavity, absorption and scattering
losses make the ringdown time too short. This is a direct result of a high
density of molecules in the liquid (typically 1019). The liquid causes a reduction
in the Q-factor but introducing the TIR element preserves the Q of the cavity.
The only limit is scattering that is a result of surface roughness, under the
assumption that non-specular losses can be neglected.20 TIR creates an
evanescent wave at the silica interface. This wave decays exponentially and
has a certain penetration depth into the medium. Any molecules present
inside this evanescent field will absorb the radiation and this will cause a
decrease in ringdown time. The use of evanescent waves for probing surface
chemistry forms the basis of attenuated total internal reflectance (ATR)
spectroscopy.21 This has also been enhanced by including metals on the
surface which produces a Surface Plasmon (called surface plasmon
resonance, SPR). 22
By utilizing TIR and the deposition of gold nanoparticles on the silica surface,
it is possible to produce an instrument that is able to observe reactions and
molecule binding regardless of whether the molecule of interest absorbs at
the laser wavelength or not due to the surface plasmon properties of the metal
nanoparticles (see section 1.3). Molecules (specifically proteins) can also be
labelled with fluorophores targeted at the wavelength of interest.
8
1.2 Gold Nanoparticles - colloidal suspensions
The particles that were used for all experiments were gold nanoparticles,
which are relatively easy to prepare. In solution these particles are finely
dispersed and the charged layer around them prevents them from
aggregating together. Derjaguin, Verway, Landau and Overbeek put a theory
forward that dealt with colloid stability in the 1940’s (DVLO). 23 The DVLO
theory states that the stability of a colloid system is determined by the sum of
the double layer repulsive and van der Waals forces that particles experience
when they approach each other. The theory speaks of an energy barrier that
is the direct result of these repulsive forces therefore keeping them finely
dispersed in the solution. However if colloid particles collide together with
sufficient energy to overcome this barrier then the attractive force will result in
pulling them together, thus forming precipitates that are irreversible (see
Figure 1-3).
Double Layer
Repulsive Force
Energy
Net
Energy
Van der Waals
Attracive Force
Particle Separation
Figure 1-3: Diagram showing relationship between attractive/repulsive forces and
particle separation.24
The electric double layer that surrounds the metal nanoparticle is the result of
the development of a net charge at the particle surface that affects the
distribution of ions in the surrounding interfacial region.25 This net charge will
result in an increased concentration of counter ions (ions of opposite charge
to that of the particle) close to the surface. This electric double layer exists
9
around each particle.25 The liquid layer around the nanoparticle consists of
two parts (see Figure 1-4):

An inner region (Stern layer) where the ions are strongly bound.

And an outer region (which is diffuse) where binding is less firm.
Figure 1-4: Diagram of charge associated with a single nanoparticle.26
The slipping plane is a notational boundary that exists in the diffuse layer. The
electric potential of this boundary is called the zeta potential, which provides
stability to the colloid.27 The zeta potential is the overall potential energy of
particles in solution.27 The higher the zeta potential is (negative or positive)
the more stable the colloid is. A high negative or positive value will result in a
higher repulsion of each particle from each other. This means that the lower
the zeta potential, the lower the repulsion from each other is and the higher
the chance that flocculation will occur. A dividing line between stable and
unstable colloids is given to be +/- 30 mV.28
History shows that gold colloids have been used for medical applications due
to their appearance.29 The red colour matched that of blood; however you
would expect a gold metallic colour. Michael Faraday 29 was the first person to
10
attempt to explain the red colour of gold colloids as well as being the first to
produce aqueous dispersions of gold colloids. The paper that he published in
1857 explains a procedure in which he was able to reduce a gold salt, HAuCl4
(yellow in colour), using a two-phase reaction mixture. The reducing agent he
used was white phosphorus in diethyl ether. He observed a colour change
from pale yellow to deep red and was also able to characterize a rough
estimate of the particle size of the gold dispersion.29 He hypothesized that the
gold particles were spherical in shape and about 5-10 nm in diameter.29
Turkevich established the first standard protocols for the preparation of metal
nanoparticles.30 The protocol he developed was the reduction of HAuCl4 using
sodium citrate, in which 20 nm gold nanoparticles were formed. This is the
method that has become a standard for making gold colloids with a narrow
size distribution.
Other protocols for the synthesis of gold nanoparticles are available as well.
These include the use of organic solvents such as tetrahydrofuran (THF) 31,
THF/MeOH
32,
organometallics
33,
surfactants
34
and long chain alcohols.35
Recently other protocols have been developed such as intracellular synthesis
by an alkalotolerant actinomycete,36 in which the result was gold nanoparticles
in the cell wall and on the cytoplasmic membrane of about 5 – 15 nm.36
1.3 Surface Plasmon Resonance
Nano-metal particles with dimensions smaller than visible wavelengths exhibit
a wealth of phenomena directly related to geometry-dependent surface
plasmon resonances that can be excited when electromagnetic fields are
directed towards them.37 The characteristic red colour associated with gold
colloid suspensions is due to large scattering cross sections of plasmon
resonances. Colours associated with colloidal suspensions of gold and silver
are due to large scattering cross sections (SCS’s) of plasmon resonances. In
a continuous metal surface, plasmons are waves that propagate along the
surface of a conductor at the interface between a metal and a dielectric
material (see Figure 1-5).
11
Figure 1-5: Diagram showing surface plasmons at the interface between a metal and a
dielectric.38
The surface plasmon (electromagnetic wave) is excited when photons are
made incident upon this metal/dielectric interface and induce a resonant
charge density oscillation at that surface, creating a propagating wave (a
surface plasmon).8,
9
For metallic nanoparticles, their small size leads to an
intense absorption in the visible/near-UV region. The conduction electrons
show a characteristic collective oscillation, which leads to a plasmon band
being observed near 530 nm for nanoparticles in the 5-20 nm range. This is
known as localized surface plasmon resonance (LSPR).9, 39
The plasmon resonance frequency of the nanoparticle depends on the local
refractive index. A change in refractive index will result in a shift in the
oscillation frequency and will therefore result in refractive index sensitivity
(RIS).40 If for instance, the surface of a nanoparticle is immersed in an
aqueous buffer and subsequently is immersed in a solution with a higher
refractive index, then this causes an increase in refractive index, which is
detected by a shift in the SPR (and extinction) of the nanoparticles (see
Figure 1-6).
12
Increase
Increaseininabsorbance
extinctionat
at
selected
selectedwavelength
wavelengthdue
duetoto
shift
shiftininplasmon
plasmonresonance
resonance
band
band
Extinction
Absorbance
Extinction at
Absorbance
at selected
wavelength
wavelength
Wavelength
Figure 1-6: Diagram showing how a shift in the surface plasmon band will affect the
particle extinction at a specific wavelength.
Figure 1-6Error! Reference source not found. can also be interpreted as a
binding event onto the nanoparticle. This sensitivity to binding events also
makes nanoparticles suitable for functionalizing them using, for example
antibodies and turning them into bio-specific sensor particles, which makes
nanoparticles suitable for use in the field of biomedical science.7, 8
The sensitivity of LSPR is dependent on particle morphology (size and
shape), dielectric environment (coating, surrounding medium, supporting
substrate), interparticle coupling (state of aggregation)41 and the sensitivity of
the evanescent cavity ring-down technique.
SPR sensors can be divided into two major classes42, those that use
wavelength interrogation and those that use angle interrogation. The
BIACORE43, 44 (made by Biacore AB) uses angle interrogation and works at a
fixed wavelength, by employing photo detectors that allow tracking of the
angle of reflectance minimum. The other type or SPR keeps the angle of
incidence fixed and monitors the spectral changes.
13
1.4 Adsorption Kinetics
Quantifying adsorption and desorption rates and constants is a major part of
this report, because this can help us to eventually investigate more complex
biochemical reactions quantitatively.
The signal change of adsorption experiments generally shows the following
trend (see Figure 1-7).
0,6
Extinction/10-2
0,5
0,4
0,3
0,2
0,1
0
0
500
1000
1500
2000
2500
Time/s
Figure 1-7: Example of experiment done that seems to follow the Langmuir adsorption
isotherm.
The levelling off of the signal shows that equilibrium is reached, meaning that
there is a balance between adsorption and desorption (according to
Langmuir’s model). Reaching this balance is a typical outcome that can be
modelled using the Langmuir adsorption equation.45, 46 (see Equation 1-1)
dθ
n1
n2
 k a Col  1  θ   k d 
dt
Equation 1-1
dθ
n1
n2
 k a Col  1  θ   k d 
The symbols in dt
Equation 1-1 mean the following: ka (M-1 s-1) is the adsorption constant,
kd (s-1) is the desorption constant, [Col] is the colloid concentration, t is time, θ
is the surface coverage, n1 and n2 (n1 and n2 are 1 in the Langmurian model)
14
are values that are used to model co-operative binding, allowing the
adsorption kinetics to depend in a non-linear way on the particles already on
the surface and in solution.46
The Langmurian model was fitted to experimental data (obtained in the lab) by
solving the differential equation numerically using the Runga Kutta method,
which gave, as a result, values for the model parameters ka and kd. The
colloid concentration is calculated using The Beer-Lambert law. The values
for surface coverage lie in between 0 and 1. When 1 is reached, then a
monolayer is formed. Previous research (performed in our laboratory) showed
that the concentration of the monolayer (for gold nanoparticles of diameter 15
nm) on the prism surface is 364 μM. This value was determined based on
particle and surface size and the assumption that one layer is a closely
packed layer of nanoparticles.
This MPhil report will focus on the results that were obtained during the past
two years. The experimental setup that was used for all experimental work will
be explained in chapter 2. Chapter 3 will present results on the aggregation of
gold nanoparticles on a silica surface to act as localised particle plasmon
sensing surfaces. Chapter 4 will also provide a literature review of the process
of DNA hybridization that was used as a measure of the sensitivity of the
plasmon surface to molecular binding events: preliminary results will be
presented in this chapter.
15
1.5 References
1. http://www.ornl.gov/sci/techresources/Human_Genome/faq/genenumber.s
html#seventh
2. http://www.projects.ex.ac.uk/atto/project.html
3. Sorokin, A., Bobashev, S., V., Tiedtke, K., Richter, M. J. Phys. B: At. Mol.
Opt. Phys. 2006, 39, 299.
4. Kovalev, V., I., Barton, J., S., Richardson, P., R., Jones, A., C. Journal of
Physics 2006, 45, 201.
5. Semwogerere, D., Weeks, E., R. Encyclopaedia of Biomaterials and
Biomedical Engineering 2005, Taylor and Francis.
6. Brock, R., Vamosi, G., Vereb, G., Joveing, T., M. Proc. Natl. Acad. Sci.
1999, 96, 10123.
7. Schuck, P., Minton, A., P. Analytical Biochemistry 1996, 240, 262.
8. Kress-Rogers, E., Phil, D. Handbook of biosensors and electronic noses.
Medicine, food and the environment. (p149-p168)
9. Riu, J., Marato, A., Rius, F., X. Talanta 2006, 69, 288.
10. Herbelin, J., M., McKay, J., A., Kwok, M., A., Uenten, R., H., Urevig, D., S.,
Spencer, D., J., Bernard, D., J. Appl. Opt. 1980, 19, 144.
11. Wheeler, M., D., Newman, S., M., Orr-Ewing, A., J., Ashfold, M., N., R. J.
Chem. Soc., Faraday Trans. 1998, 94 (3), 337.
12. Romanini, D., Lehmann, K., K. J. Chem. Phys. 1993, 99, 6287.
13. Paldus, B., A., Harb, C., C., Spence, T., G., Wilke, B., Xie, J., Harris, H.,
S., Zare, R., N. Journal of Applied Physics, 1998, 83, 3991.
14. Schultz, K., J., Simpson, W., R. Chem. Phys. Lett. 1998, 297, 523.
15. Totschnig, G., Baer, D., S., Wang, J., Winter, F., Hofbauer, H., Hanson,
R., K. Applied Optics, 2000, 39, 2009.
16. Romanini, D., Kachanov, A., A., Stoeckel, F. Chemical Physics Letters
1997, 270, 538.
17. Pipino, A., C., R., Hudgens, J., W., Huie, R., E. Rev. Sci. Instrum. 1997,
68, 2978.
16
18. Pipino, A., C., R., Hudgens, J., W., Huie, R., E. Chemical Physics Letters
1997, 280, 104.
19. Regan, J., J., Anderson, D., R. Comput. Phys. 1991, 5, 49.
20. Lui, C., Kaiser, T., Lange, S., Schweiger, G. Optics Communications 1995,
117, 521.
21. Harrick, N., H. Internal Reflection Spectroscopy, Wiley, New York, 1967.
22. Kretschmann, E., Raether, H. Z. Naturforsch 1968, 2135.
23. Adamczyc, A., Weroński, P. Advances in Colloid and Interface Science
1999, 83, 137.
24. http://www.malvern.com/LabEng/industry/colloids/dlvo_theory.htm
25. Shaw, D., J. Colloid and Surface Chemistry – 4th Edition, ButterworthHeinemann Ltd 1992.
26. http://209.85.229.132/search?q=cache:1i2EtHaMqesJ:www.nbtc.cornell.ed
u/facilities/downloads/Zeta%2520potential%2520%2520An%2520introduction%2520in%252030%2520minutes.pdf+DVLO+
theory&cd=6&hl=en&ct=clnk&gl=uk
27. Attard, P. Current Opinion in Colloid & Interface Science 2001, 6, 366.
28. Greenwood, R. Advances In Colloid And Interface Science 2003, 106, 55.
29. Faraday, M. Philos. Trans. R. Soc. London 1857, 147, 145.
30. Turkevich, J., Stevenson, P., C., Hillier, J. Discuss. Faraday Soc. 1951,
11, 55.
31. Franke, R., Rothe, J., Pollmann, J., Hormes, J., Bonnemann, H., Brijoux,
W., Hindenburg, T. J. Am. Chem. Soc. 1996, 118, 12090.
32. Vidoni, O., Philippot, K., Amiens, C., Chaudret, B., Balmes, O., Malm, J.,
O., Bovin, J., O., Senocq, F., Casanove, M., J. Angew. Chem., Int. Ed.
1999, 38, 3736.
33. Sinzig, J., De Jongh, L., J., Bonnemann, H., Brijoux, W., Koppler, R. Appl.
Organomet. Chem. 1998, 12, 387.
34. Reetz, M., T., Lohmer, G. Chem. Commun. 1996, 1921.
35. Tanori, J., Pileni, M., P. Langmuir 1997, 13, 639.
36. Ahmad, A., Senapati, S., Islam Kham, M., Kumar, R., Ramani, R., Sriniva,
V., Sastry, M. Nanotechnology 2003, 14, 824.
37. Kottmann, J., P., Martin, O., J., F., Smith, D., R., Schultz, S. Physical
Review B 2001, 64. 235402.
17
38. http://www.stanford.edu/group/cpn/research/investigators_2.html
39. Liao, H., Nehl, C., L., Hafner, J., H. Nanomedicine, 2006, 1(2), 201.
40. Chen, C., Cheng, S., Chau, L., Wang, C., R., C. Biosensors and
bioelectronics 2006, 22, 568.
41. Xu, H., Kall, M. Sensors and actuators B 2002, 87, 244.
42. Naimushin, A., N., Soelberg, S., D., Nguyen, D., K., Dunlap, L.,
Bartholomew, D., Elkind, J., Melendez, J., Furlong, C., E. Biosensors and
Bioelectronics 2002, 17, 573.
43. Homola, J., Yee, S., S., Gauglitz, G. Sensors and Actuators B 1999, 54, 3.
44. Carrick, F., E., Forbes, B., E., Wallace, J., C. The Journal of Biochemistry
2001, 276, 27120.
45. Atkins, P., de Paula, J. Elements of physical chemistry. Fourth edition.
Freeman and Company
46. Fisk, J., D., Rooth, M., Shaw, A., M. J. Phys. Chem. C, 2007, 111, 2588.
18
2 Experimental Methods
2.1 The e-CRDS Instrument
The experimental setup that was used for all experiments consists of a
continuous wave (cw) laser, two opposing mirrors and a photomultiplier tube
(PMT) connected to an oscilloscope for rapid data collection. The setup
results in a simple linear optical cavity. The cw laser has a line width of 5 nm
(at 635 nm) that will overlap some 104 cavity modes allowing light to enter the
cavity at all times without having to lock the laser to a mode of the cavity. This
is an advantage, because it makes alignment of this cavity extremely easy.
Spectra can be collected rapidly and averaged together over a period of one
second to improve the signal-to-noise ratio. The collection rate is typically
done at 6 kHz. A Dove prism is introduced into this cavity to introduce a TIR
element into the cavity allowing for biosensing to occur at the TIR surface
whilst preserving the optical alignment of the cavity, whilst maintaining that
light path allowing for a number of different measurements on top of this
surface. The free-running Dove Cavity implementation of evanescent wave
cavity ring-down spectroscopy (e-CRDS) has previously been developed in
this laboratory and has been further used as the platform for detection of
various adsorption and binding reactions in this thesis.
This chapter will describe the development of the experimental methods that
were used during all experiments
2.2 Cavity Design
An optical e-CRDS cavity was constructed by placing two high reflectivity
mirrors opposing each other (R = 0.9995, Layertec), centred at a wavelength
of 635 nm. A fibre-coupled laser transfers the radiation to the cavity and the
laser intensity may be modulated at a frequency of 6 kHz .
The rise-time of the laser is less than 10 ns and does not affect the ring-down
exponential trace and hence the determination of the ring-down time, . The
radiation that exits the fibre passes through a collimating lens at the end of the
19
fibre prior to entry into the cavity. When the laser radiation passes through the
collimating lens it results in a collimated laser beam that matches the width of
the cavity mode in the centre of the cavity. Immediately after the collimating
lens the laser radiation passes through a λ/4 quarter-wave plate to ensure the
vertical orientation of the light polarisation is optimal. This is needed because
the AR-coating on the Dove prisms is optimized for p-polarised light.
The light is launched into the back of a flat, high reflectivity mirror (R =
0.9995). The other mirror that is mounted at the end of the cavity is a
concave, high reflectivity mirror (R = 0.9995, radius of curvature = 1 m) to
collect the light and produce a stable optical cavity (see Figure 2-1).1
High reflectivity
Cw diode
laser
PMT
mirrors
Ring down decay wave
observed
by PMT
Dove prism
prism
Wave plate
Figure 2-1: Schematic drawing of e-CRDS setup.
According to literature reports a stable optical resonator is formed when the
length of the cavity (which is 86 cm in this case) is shorter than the radius of
curvature of the mirror (in this case 1 m).2 The mirrors that were used in the
experimental setup were purchased from Layertec Optical Coatings with
reflectivity equal to 0.9995 and a transmission of 0.08% at 633 nm. The
optimal ringdown time that can be obtained in the free-running cavity mode of
operation (no prism) is 7.00 ± 0.02 µs with / being 0.6% (σ is the standard
deviation of the sample, which in this case is the ring-down time). The
ringdown time can be defined according to Equation 2-1.
20
τ=
tr / 2
(1 R )
Equation 2-1
Where tr is the round trip time (ns) and R is the reflectivity of the mirrors
For the 86 cm cavity the round trip time is 5.72 ns and the mirror reflectivity is
0.9995. Substituting these numbers into equation 2.1 will give 5.72 µs as the
ring-down time for an empty cavity.
The difference between the quoted
number of 7 µs and 5.72 µs is caused by very small changes in the mirror
reflectivities.
Light must overlap with at least one mode to be trapped in the cavity. The
bandwidth of the laser is large enough to overlap more than one cavity mode
(actually about 104 modes) in the free-running configuration. Therefore light
that enters the cavity will build up to an intensity that is determined by the Qfactor (see Equation 2-2).2
Q=
2π τ
tr
Equation 2-2
Although the wavelength of the laser is considered to be 635 nm (the
bandwidth is ± 5 nm according to the manufacturer), in this report there will
always be a certain range in wavelength which is called the bandwidth. When
radiation of the laser overlaps one or more cavity modes, it will start to
interfere with itself.2 In the cavity, this can lead to standing waves between the
two mirrors. Standing waves are two waves with identical frequencies that
interfere with themselves while travelling in opposite directions inside the
same medium. These standing waves form a set of discrete frequencies
known as the longitudinal modes of the cavity.3 The result is that all other
frequencies are being suppressed by this interference and therefore, these
longitudinal modes are the only self-generating frequencies of light that are
allowed to oscillate within the resonant cavity. Overlapping of more than one
cavity mode by a cw laser requires that the laser light bandwidth should be
greater than the free spectral range (FSR). The FSR ( δν ) of a cavity can be
determined by Equation 2-3:4
21
δν =
c
2l
Equation 2-3
where l is the length of the cavity and c is the speed of light
The FSR for the cavity which was used for experimentation is 172.8 MHz.
Equation 2-4 shows how to calculate the full-width-half-maximum (FWHM) of
each cavity mode, which is related to the Finesse of the cavity.4
FWHM 
FSR
Finesse
Equation 2-4
The diode laser has a bandwidth of  5 nm centred at 635 nm and overlaps
with the longitudinal modes of the Dove cavity. The wavelength of the cavity
modes is given by Equation 2-5:5
λ =
2l
n
Equation 2-5
Where n is the mode number, λ is the wavelength of the mode and l is the
length of the cavity. When, for the experimental setup, the parameters are
substituted into Equation 2-5, the laser overlaps about 1.7  104 longitudinal
cavity modes, therefore the cavity is always in resonance and light always
enters the cavity.
Each ring-down time value corresponds to the decay of light in each of the 1.7
 104 cavity modes overlapped by the laser and so is an average. This ringdown time varies a little due to the bandwidth of the mirror reflectivity. If mirror
reflectivity is assumed to be constant, then the large number of cavity mode
values (τ) provides for a large sample set. This large sample set improves the
confidence of τ, however the practical limit of variation of / ~ 1%, may be
improved.6 The first improvement in / shows an increase with the number
of shots averaged (signal averaging). The laser is switched off at a repetition
22
rate of 6 kHz and the radiation intensity in the cavity decays with a ring-down
time, , determined similarly by the Q-factor (see Equation 2-2). The decrease
in τ is calculated using the following equation.7
τ=
tr
2 ((1 R ) + (1 T ) + (1 AR) + Lsurf + σ l )
Equation 2-6
where, T is the transmission loss through the silica, Lsurf is the diffraction
losses at the prism surface per roundtrip, σl is the absorption of light when the
beam enters the silica at the interrogation wavelength and AR is the antireflection coating losses per round trip.8 Increasing the sensitivity of the cavity
can therefore be achieved by minimizing the Lsurf and σl. If zero is taken for
these than the maximum τ will be around 600 s, when T is 0.9975, AR is
8×10-3, R is 0.9995 and tr is 6 ns.8
The penetration depth of the evanescent field into the medium above the
Dove prism may be calculated using Equation 2-7.
dp =
λ
2 π (( sin( ))
2
Equation 2-7
1
2 2
12
n
)
Calculation of the penetration depth (dp) can be done by using the ratio of the
refractive indexes (n12) of the two media and the angle of incidence of the
refracting radiation. The dp for 635 nm is 189.2 nm (in which n12 is 1.4677 at
635 nm).9
The absorption of light can be calculated based on the results obtained during
an experiment (decrease in ring down time). Equation 2-8 shows how to do
that.9
Abs =
Δτ t r
2.303
τ 0τ 2
Equation 2-8
23
Where, Δτ is the change in τ due to absorbance and τ0 is the ringdown time of
the empty cavity. The results are then converted to base 10 logarithms for
comparison with the Beer-Lambert expression for comparison (see Equation
2-9).9
Abs = ε [C ]l
Equation 2-9
The extinction and concentration profile at the interface is directly related to
the change in Δτ. The extinction coefficient (ε) can therefore be determined by
using a known concentration of nanoparticles and Δτ and counting the
particles using the SEM (scanning electron microscope). [C] is the
concentration of the adsorbed material on the prism surface and l is the length
of the sample.9
The light after the second mirror is focused onto a Hamamatsu PMT module
(H7732MOD, containing a R4632 PMT). The signal is digitised on a Lecroy
Waverunner oscilloscope (LT262, 8 bit 350 MHz sampling rate). 256 traces
are averaged on the oscilloscope and then passed to the computer interface
routine. This routine was written in-house for analysis purposes, using the
LabView interface software.
The routine fits the resulting ring-down time to a single exponential by
correcting for the base line and then fitting the logarithm of the trace to a
least-squares fit to a straight line. Real-time calculations of    and / are
derived from the fitted slope parameter and baseline correction (using a nonweighted
Levenburg-Marquart non-linear least
squares fitting routine
averaging over 4 shots).
2.3 Flow cell
On the TIR surface of the prism, a single-pass flow cell made of
polytetrafluoroethene (PTFE) was constructed. The flow channel has a width
of 10 mm when it is placed on the underside of a specially constructed PTFE
24
block. Once everything is clamped to the prism surface, the flow cell is 1 mm
thick using a nitrile ‘O’-ring, the total volume of the flow cell was 190 l.
Samples were flowed through the flow cell at a rate of 4 ml per hour using a
syringe pump (see Figure 2-2). This corresponds to a maximum linear flow
velocity of 0.14 mm s-1.
Flow in
Flow out
Volume is
approx. 0.2 ml
Cross section O-ring
is approx 8mm x 1mm
Prism
Mounted into a
holder
Figure 2-2: Schematic drawing of the flow cell.
The flow through the flow cell determines the rate of transfer of molecules to
the surface. A measure of the rate of transfer is Reynolds Number. This
number represents a measure for the flow regime present within the flow cell
and can be calculated according to Equation 2-10.10
Re =
ρ×u ×d
μ
Equation 2-10
where, ρ is the fluid density, u is the flow velocity, d is the characteristic flow
dimension and µ is the fluid viscosity. This calculation assumes that the fluid
viscosity and density are equal to those of water at 25 °C (0.8909 × 10-3 N s
m-2 and 998 kg m-3 respectively). With a cell dimension of 1 mm, the Reynolds
Number for this experimental setup is 0.16. Therefore, diffusion-limited flow
conditions should prevail.
25
2.4 References
1. Atkins, P., de Paula, J. Elements of physical chemistry. Fourth edition.
Freeman and Company.
2. Berden, G., Peeters, R., Meijer, G. Int. Reviews in Physical Chemistry
2000, 19, 565.
3. Demtröder, W. Laser Spectroscopy 3rd Edition, Springer-Verlag Berlin
Heidelberg New York 2003.
4. Gryezynski, I., Gryezynski, Z., Lakowicz, J., R. Analytical Biochemistry
1997, 247, 69.
5. O’Keefe, A., Deacon, D., A., G. Rev. Sci. Instrum. 1998, 12, 2544
6. Pipino, A., C., R. Phys. Rev. Lett. 1999, 83, 3093.
7. Chen, C., Cheng, S., Chau, L., Wang, C., R., C. Biosensors and
bioelectronics 2006, 22, 568.
8. Xu, H., Kall, M. Sensors and Actuators B 2002, 87, 244.
9. Fisk, J., D., Rooth, M., Shaw, A., M. J. Phys. Chem. C, 2007, 111, 2588.
10. Rott, N. Annual Review of Fluid Mechanics 1990, 22, 1.
26
3 Chemically Induced Assembly of Gold Nanoaggregates on Silica Oxide Surfaces
3.1 Introduction
Aggregates of metal nanoparticles on silica surfaces show enhanced optical
effects leading to surface enhanced Raman scattering (SERS) 1, 2 and surface
enhanced absorption spectroscopy (SEAS).3 Our group recently reported
non-linear protein adsorption kinetics4 for binding proteins to an 800nanoparticle cluster showing sensitivity to attogram ml-1 of protein binding to
surface sites, also demonstrating an extreme sensitivity enhancement
compared with the conventional nanoparticle surface. The conventional
surfaces previously utilized by our group are single nanoparticles adsorption
onto silica surfaces. The electric field between the surface roughness features
or aggregated nanoparticles is enhanced, coupling the radiation to the
nanoparticle optical scattering5,
6
properties, allowing the nanometre-scale
structure to act as a local optical aerial. The enhancement effect is not
however completely explained by the physical association of a molecule to the
enhanced region but there appears to be a chemical component relating to
the structure of the molecule.2, 7 The electric field enhancement appears to be
localised to a region within a few nanometres of the surface, suggesting an
ideal separation and perhaps optimum geometry of the touching particles. 8
The effects of the field enhancement have been monitored using the total
optical extinction of the nanoparticles with contributions from Rayleigh scatter
and the localised plasmon. Excitation of the localised particle plasmon has
been used as an alternative to continuous gold surfaces for detecting
biological processes bound to the gold surface. The localised plasmon field
penetrates approximately one particle radius into the medium above the
particle and is sensitive to the local refractive index (RI). Protein adsorption to
the surface and protein-antibody binding has been observed by monitoring the
change in the extinction of the metal nanoparticle using a number of
techniques including evanescent wave cavity ring-down spectroscopy9-11. The
27
nanoparticle surfaces appear to be less sensitive to the refractive index in the
plasmon, detecting changes of 10-4 in the local refractive index, although this
depends on the wavelength of interrogation. The RI sensitivity may be
compared with 10-6 routinely achieved with continuous gold surface plasmon
instruments.
Improving the refractive index sensitivity of nanoparticle sensor surfaces has
been directed towards fabricating particles of controlled geometry12 to
maximise the interaction with the interrogating radiation, leading to an optical
aerial. Nanoparticle synthesis of triangles, squares, spheres and octagons
have been reported.5,
13-15
but these show little improvement in the bulk RI
sensitivity. Growing rods with controlled aspect ratios is usually a two-step
process: the first is to prepare a seed solution of small spherical
nanoparticles, typically of 4 nm in diameter, and then to use a surface acting
ligand or capping agent to induce asymmetric growth, probably via a surface
energy specific interaction.16 The ligand, however, remains on the surface of
the particle and prevents interaction with binding proteins and so it must be
removed after synthesis and before it is adsorbed to the sensor surface.
The ligand on the surface of the nanoparticle synthesised in solution is
responsible for maintaining the colloidal phase and in the case of the citrate
reduced colloids17 the ligand provides a charged interface. Adsorption of the
nanoparticle to a sensor surface such as a native silica surface or a modified
silica surface requires the stabilised particle to preferentially bind to the
surface. Binding kinetics have been observed previously to a native silica
surface18 for the citrate-reduced gold nanoparticles, indicating a strong
attachment to the negatively charged silica surface. The native silica surface
has two Si-OH per nm2 which de-protonate as a function of the bulk pH
producing a negatively charged surface with a surface potential of -125 mV
when fully dissociated.10, 19 The charged surface attracts a bilayer of positively
charged counter-ions, balancing the negative charge and producing a large
concentration enhancement above in the interface. The charged bilayer
around the nanoparticles may be reduced by the bulk ionic strength, ultimately
leading to the instability and the aggregation of the colloid. The same
28
destabilisation appears to occur at the charged silica interface, resulting in a
nanoparticle structured biosensor surface.
Two different mechanisms for aggregation were utilized. The first mechanism
was aggregation of nanoparticles in salt-destabilised solution to generate the
fractal clusters that demonstrated the non-linear response to protein binding
observed previously. Control of the aggregation process in solution is
however, difficult as the kinetic process accelerates when nucleation around
the seed particles occurs. Some interesting structures have been trapped
during aggregation, including chains and cluster aggregates but the interface
offers a more controllable environment in which it moderates the growth of the
aggregates.
The second mechanism for aggregation was to first chemically modify the
OH-groups on the silica surface. This was done using an aminosilane. The
aminosilane
that
was
used
aminopropyltrimethoxysilane
to
modify the
(3-APM).
Figure
prism
3-1
surface
gives
a
was
3-
schematic
representation of the surface chemistry of the prism surface when it is treated
with 3-APM.
29
Ethanol
Ethanol
NH2
OHHO
OHHO
OH HO
OH HO
OH
Si
O
Si
Si
Si
O
Si
O
O
O
Si
O
O
O
O
Si
Si
O
O
O
O
O
O
Si
O
O
O
O
O
O
Si
Si
Si
Si
O
O
O
O
Si
O
O
Si
Si
O
O
NH2
NH2
O
O
O
HO
NH2
NH2
NH2
H2O
Water
+
O
O
O
O
O
O
O
O
Si
O
O
O
O
O
O
O
Si
O
O
O
O
Si
O
O
O
O
Si
Si
O
O
O
O
O
O
O
O
O
O
Si
O
O
Si
Si
Si
Si
O
O
Si
Gold Colloid
Si
NH3
NH3
O
O
O
O
Si
Si
Si
Si
Si
O
+
+
NH3
NH3
O
Si
O
O
NH3
+
+
NH3
NH3
Si
Si
+
+
NH3
O
O
+
NH3
+
+
NH3
O
Figure 3-1: Schematic representation of reaction of 3-APM with the prism surface and
gold adsorption onto this aminated surface.
In Figure 3-1, 3-APM reacts with the OH groups on the prism surface. The
hydrolysis of one or more alkoxy groups results in 3-APM covalently binding
to the prism surface (see top part of the scheme). The pKa value (pKa=log([NH2][H+]/[NH3+])) of organic -NH3+ is about 9 which means that below
pH 9 it favors the protonated form. At pH 7, which is two orders of magnitude
lower, the majority of amino groups are protonated, hence the positive surface
charge (see bottom part of the scheme). Functionalizing a prism surface like
this can effectively be used to build up interesting nano-level structures. 20
The aminated prism surface is called a functionalised surface. Aminating a
surface like this is also interesting because afterwards biomolecules can be
attached to it. This aminosilane had three ethoxy groups. One, two or three of
these groups can bind to silica oxide. The ethoxy groups that can be left after
30
binding may cross link to each other. A similar approach can be used to
absorb silver nanoparticles to the Dove prism surface.
21
In this chapter a series of aggregation experiments with citrate-reduced 15 nm
gold nanoparticles adsorbing to the native silica is presented. The extinction
of the surface has been observed in real time using evanescent wave cavity
ring-down spectroscopy and the adsorption and aggregation kinetics have
been observed directly. The aggregates have been imaged using SEM and
tested for its sensitivity using bulk refractive index changes.
The results in this chapter were used as a basis for a publication that was
recently accepted by J. Phys. Chem. C.22
3.2 Experimental Methods
The e-CRDS technique has been described in detail elsewhere 10 and in
chapter 2 and will therefore only be briefly described here. Two highreflectivity (R=99.95%) mirrors are mounted onto an optical table at a distance
of 81 cm: one mirror is planar and the other mirror has a radius of curvature of
one metre forming a stable optical cavity.
Radiation from a continuous wave diode laser (635 nm, 5 nm bandwidth) is
introduced into the cavity and passes through a wave plate to control the
plane of polarisation and enters the cavity through the back of the planar
mirror. A photomultiplier tube is mounted behind the concave mirror to collect
the ring-down signal. The diode laser is pulsed at a frequency of 6 kHz and
the bandwidth is sufficient to overlap more than one cavity mode, allowing
maximum intensity to build up within the cavity, determined by the cavity Q
factor. The radiation intensity in the cavity decays when the laser is switched
with a ring-down time, τ, determined by the Q factor.
To create a total internal reflection (TIR) event, a Dove prism (fused silica) is
introduced into the cavity with a fixed-angle 450 and anti-reflection coating
31
optimised for 635 nm on both ends to minimize the loss of light and maintain
the Q-factor of the cavity. The TIR creates an evanescent wave at the
interface between the silica surface of the prism and the rarer medium above.
The evanescent wave decays exponentially with a 1/e penetration depth of
186 nm, fixed by the angle of incidence at the reflecting surface and the
wavelength of the radiation. Molecules present within the evanescent field that
scatter or absorb at the wavelength of the radiation will cause a decrease in
the Q of the cavity and hence a decrease in τ. The evanescent wave can also
excite the localized surface plasmons (LSP) in metal nanoparticles and the
change in extinction monitored as a function of gold surface chemistry.
The quality of the prism surface is determined by the cleaning procedure. We
found out that day to day variations could be minimized by using the following
cleaning procedure; the Dove prism was washed in aqua regia (25% v/v of
HNO3 and 75% v/v of HCl) for an hour, after which, the prism was rinsed with
excess Decon90 water and a sequence of water, ethanol and isopropyl
alcohol (IPA) and dried with lens tissue. The prism was then placed into an
optical mount and the flow cell (volume of 190 μl) secured above the reflecting
surface. Further surface preparation continued in situ with more washes with
IPA, water and finally sodium citrate (8.2 mM), the colloid buffer. The surface
preparation continued until the ring-down time stabilised before starting the
aggregation experiments.
A citrate-reduced colloid was prepared using the Turkevich-method described
in detail in chapter 1.17 The protocol was as follows: 100 ml of 1 mM HAuCl4
was reduced using 10 ml of 90 mM of sodium citrate. The HAuCl4 solution
was heated to 95 oC and the citrate was injected whilst stirring vigorously. The
colloid was cooled to room temperature, after which, the absorbance
spectrum was measured, from which max = 523 nm indicates a colloid with 15
nm diameter particles.18 The colloid solution was diluted with 8.2 mM sodium
citrate to reduce the concentration to 0.35 nM before use in the aggregation
experiments.
32
The diluted solution was introduced into flow cell and the adsorption kinetics
monitored in real time as a change in the ring-down of the cavity. Once
adsorbed onto the surface, a washing cycle of sodium citrate, HCl (pH=2) and
water was used to remove the citrate ligand from the surface of the
nanoparticle. The refractive index sensitivity was measured by measuring the
change in extinction at 635 nm when the bulk solvent was changed from
water to IPA. A second colloid adsorption step was then allowed by
introducing the colloid into flow cell until the ring-down time stabilised. Wash
and adsorption cycles were repeated until the extinction exceeded the
dynamic range of the cavity i.e. the ring-down time became too short. SEM
imaging was performed at each stage of the surface aggregation.
The experiments using 3-APM used only one adsorption step and utilized a
different protocol. The first step was to coat the prism surface using a 0.34 nM
3-APM solution (dissolved in methanol) for 1 minute, before the adsorption
experiment was performed. RIS sensitivity was determined in the same way
as with the multi-phase adsorption experiments.
3.3 Results
The ring-down times are converted to extinction for each experiment and the
absorbance rate constants determined by fitting them to first order differential
rate law given in Equation 3-1:
dθ
= k a (θ m - θ )
dt
Equation 3-1
Where θ is the surface coverage at time t, θm the maximum surface coverage
and ka is the adsorption rate constant. The dissociation rate is assumed to be
negligible. The refractive index sensitivity of the surface at each stage in the
surface aggregation was monitored by determining the change in  associated
with the switch in the refractive index from IPA to water, n = 0.0445. The
refractive index sensitivity (RIS) is then determined from Equation 3-2:
33
RIS =
0.0445 σ y
Equation 3-2
ΔYwater / isopropano l
Where σy is the standard deviation in  and Ywater/isopropanol is the difference in
 between water and IPA. However at each stage there is increasing gold
nanoparticle coverage and normalized refractive index sensitivity (nRIS),
Equation 3-3, is used to remove the day-to-day variations in the cavity and
surface coverage:
nRIS =
ΔYwater / isopropano l
Equation 3-3
Ywater
Ywater is the average extinction value of water during each adsorption phase of
the experiment; Ywater/isopropanol is the difference between water and IPA.
Multiphase adsorption experiments were performed for the gold colloid
solutions (nanoparticles of diameter 15±3 nm) as can be seen in Figure 3-2.
34
100
90
80
(c)
Extinction/10-4
70
60
50
(b)
40
30
20
(a)
10
0
-10
0
2000
4000
6000
8000
Time/s
Figure 3-2: Adsorption of citrate reduced gold nanoparticles (pH 7) onto the silica
surface; (a) first phase adsorption; (b) second phase adsorption after the HCl wash at
pH 2; (c) third phase adsorption after the second HCl wash showing an increased
adsorption rate compared to the other phases.
From Figure 3-2 three distinct adsorption phases can be distinguished: (a) is
the adsorption of the initial colloid solution to the surface followed by the
extensive water, IPA and HCl washing phases; (b) is the second phase
addition associated with the formation of the nanofunctionalised surface; and
(c) shows the onset of a more rapid aggregation process. SEM images were
taken for each of the phases (a-c) and are presented in Figure 3-3.
35
(a)
(b)
500 nm
500 nm
(c)
500 nm
Figure 3-3: SEM images corresponding to phase (a), (b) and (c) presented in Figure 3-2.
Figure 3-3 (a) shows a uniform distribution of single particles adsorbed to the
surface with only occasionally pairs of particles. Figure 3-3 (b) shows a less
dense region of the surface. On this surface more pairs, triplets and
aggregates can be distinguished. Finally, Figure 3-3 (c) shows a larger
surface coverage with some well developed surface clusters and aggregates.
Figure 3-4 shows the distribution of the particles in each of the different
phases of the experiment presented in Figure 3-2. The different categories
that were identified were:

1; single particles

2; two particles attached to each other

3; three particles attached to each other

4; four particles attached to each other

linear shaped aggregates

circular shaped aggregates
36
1
Normalized values of amount of particles
0.9
A
B
C
0.8
0.7
0.6
0.5
0.4
0.3
0.2
0.1
0
1
2
3
4
Linear
shaped
aggregates
Circular
shaped
aggregates
Definition of different aggregates
Figure 3-4: Distribution graph in which, on the x-axis, 1 to 4 means the amount of
citrate reduced gold nanoparticles aggregated to each other in linear shape onto the
silica surface, larger linear shaped aggregates; more than 4 nanoparticles and circular
shaped aggregates; island formation of nanoparticles onto the silica surface; the yaxis consists of the normalized values for these different aggregates. A, B and C are
each of the phases that we identified when performing these experiments.
The experiment shown in Figure 3-2 was repeated and the results are shown
in Figure 3-5 and Figure 3-6.
37
90
80
70
(b)
Extinction/10-4
60
50
40
30
(a)
20
10
0
0
2000
4000
6000
8000
10000
12000
Time/s
Figure 3-5: Adsorption of citrate reduced gold nanoparticles onto a silica surface; (a)
multiple adsorption steps using the previously described rinsing procedure ultimately
leading to (b) linear kinetics.
In Figure 3-2 two initial adsorption phases were observed before the onset of
the aggregation phase. Figure 3-5 can be divided into two parts, part a and
part b. In part a four adsorption phases (experiments), without the onset of
linear aggregation kinetics, were performed. Part b consists of the onset of
aggregation. In this chapter a total of four experiments are presented and the
gold colloid solution that was prepared and used was, in each of these
experiments, taken from the same stock solution.
Figure 3-6 shows the third aggregation experiment done using this gold
colloid solution.
38
70
60
Adsorption 5
50
Adsorption 4
Extinction/10-4
40
30
Adsorption 3
20
Adsorption 2
Adsorption 1
10
0
0
5000
10000
15000
20000
25000
-10
Tim e/s
Figure 3-6: Five adsorption phases with the same rinsing procedure described
previously.
Figure 3-6 shows five adsorption phases without aggregation of gold
nanoparticles on the silica surface. Figure 3-6 also shows the water-IPA
switches. These switches are shown here because it was difficult to
distinguish the different adsorption phases in the trace of this experiment, also
no aggregation occurred in this experiment. The colloid solution that was used
came from the same stock solution as the previous experiments, presented in
Figure 3-2 and Figure 3-5.
Figure 3-7 shows a comparison between an experiment with an aminated
surface and the aggregation phases in Figure 3-2 and Figure 3-5.
39
6
c
5
Extinction/10-3
4
b
3
a
2
1
0
0
100
200
300
400
500
600
700
Time/s
Figure 3-7: Comparison of aggregation kinetics using different methods; (a)
aggregation kinetics taken from Figure 3-2, (b) aggregation kinetics taken from Figure
3-5 and (c) aggregation kinetics using 3-APM to chemically modify the surface of the
prism.
What can be seen in Figure 3-7 is the kinetics of these different methods are
very similar.
A kinetic analysis using Equation 3-1 was performed for each of the
adsorption steps and these data are collected in Table 3-1, Table 3-2 and
Table 3-3; a linear kinetics fit was performed on the aggregation step for all of
the data collected. At each of the deposition stages, the refractive index
sensitivity was measured for all of the surfaces, using both expressions
described by Equation 3-2 and Equation 3-3.
40
Table 3-1 Values for RIS, nRIS and ka corresponding to experimental results in Figure
3-2. All determined values using labview had an error of less than 1%.
RIS
nRIS
kax103 Mol-1.s-1
Adsorption step 1
0.61x10-3
0.51
74.80
Adsorption step 2
0.54 x10-3
0.54
20.50
Adsorption step 3
0.27 x10-3
0.31
29.60
Values
belonging
to
figure 1
Table 3-2: Values for RIS, nRIS and ka corresponding to experimental results in Figure
3-5. All determined values using labview had an error of less than 1%.
RIS
nRIS
kax103 Mol-1.s-1
Adsorption step 1
1.43x10-3
0.66
118.00
Adsorption step 2
0.84x10-3
0.55.
40.90
Adsorption step 3
0.74x10-3
0.52
894.00
Adsorption step 4
0.60x10-3
0.49
101.00
Adsorption step 5
0.23x10-3
0.35
227.00
Values belonging to
figure 2
Table 3-3: Values for RIS, nRIS and ka corresponding to experimental results in Figure
3-7. All determined values using labview had an error of less than 1%.
RIS
nRIS
kax103 Mol-1.s-1
Adsorption step 1
1.62x10-3
0.75
51.40
Adsorption step 2
0.79x10-3
0.71
451.00
Adsorption step 3
0.59x10-3
0.63
410.00
Adsorption step 4
0.62x10-3
0.54
54.20
Adsorption step 5
0.62x10-3
0.50
80.10
Values belonging to
figure 3
In Figure 3-8 the average values of ka is plotted against each of the
adsorption steps to see the reproducibility of these experiments and whether
there is any identifiable trend that can be recognized. Any overlap in standard
deviation automatically means that there is no significant difference between
the two points.
41
Figure 3-8: A plot of observed average adsorption rate, ka (mol-1 s-1) plotted against
each of the adsorption steps.
Figure 3-8 shows much overlap meaning that there is no significant difference
in ka values with increasing aggregation of particles on the silica surface.
Figure 3-9 gives the nRIS of each adsorption step for all three experiments
that are described previously. If more aggregation of nanoparticles improves
the sensitivity of the surface (so the shift in ringdown time from water to IPA)
then the values for nRIS should increase with each phase of all three
experiments shown in Figure 3-2, Figure 3-5 and Figure 3-6.
42
Figure 3-9: In this graph all nRIS values are given for each adsorption step. Each line
nRIS =
was calculated based on
Equation
ΔYwater / isopropano l
Ywater
3-3 and the results of figure 3-2 (▲), 3-5 (●) and 3-6 (■).
Figure 3-10 gives the RIS values for each adsorption step of all three
experiments.
43
Figure 3-10: In this graph all RIS values are given for each adsorption step. Each line
RIS =
was calculated based on
Equation
3-2 nRIS =
0.0445 σ y
ΔYwater / isopropano l
ΔYwater / isopropano l
Ywater
Equation 3-3 and the results of figure 3-2 (▲), 3-5 (●) and 3-6 (■).
From Figure 3-9 and Figure 3-10 the observation can be made that the
refractive index sensitivity of the first adsorption phase seems to be the
highest in all performed experiments. The nRIS values seem to decrease
when significant aggregation is observed, showing lower refractive index
sensitivities of 0.31 ± 0.12. RIS values are decreasing as well when linear
kinetics are observed at typical value of RIS is 0.6 ± 0.4  10-3. The surfaces
with increasing particle load show remarkably constant RIS values with the
same range.
44
3.4 Discussion
Adsorption onto the silica surface at neutral pH is dominated by the interaction
of the negatively charged surface with the incoming adsorbing nanoparticle.
The surface has a surface charge and hence surface potential of
approximately – 50 mV at pH 7 and this is responsible for the structure of the
charged interface.10 The gold nanoparticles are negatively charged23 with a
charge of approximately 1500 e and yet they bind to the interface. The shape
of the adsorption curve in Figure 3-2 is consistent with a Langmuir-type
adsorption isotherm reaching a maximum density which is within the dynamic
range of measurement. SEM images indicate that less than 20 % of the silica
surface is covered.22 The nanoparticles are negatively charged, as is the
surface at the pH of the colloidal solution, so the particles are attracted to the
positive counter-ion concentration in the bilayer. The particles form a pseudoionic lattice on the surface, minimizing the interaction energies by maximizing
the separation between the particles.22 The packing into the available sites on
the surface therefore depends on the surface charge, the concentration of the
counter-ions in the bilayer, the charge on the nanoparticles and the surface
morphology.22 The latter may favour adsorption and aggregation at the
surface, preferentially along a defect such as a polishing scratch.22
The second adsorption phase at the surface is then facilitated by removing
the citrate ligand from the surface of the nanoparticles by the acid, water
alcohol washes. The particles are then free from the ligand and a new ioniclattice of adsorbing particles is allowed to form with essentially the same rate
constant but with the additional mechanism of aggregation. The number of
pseudo-lattice phase adsorptions depends on the aggregation efficiency and
the random processes leading to dimer and aggregate formation. A clear
transition occurs however when the aggregation contribution dominates and
near-linear kinetics are observed as the colloid becomes unstable at the
interface and flocculates. Colloid stability in solution is controlled by the ionic
strength of the solution and hence the Debye length of the charge bilayer
around the particle which controls the interaction separation of the particles.24
45
The surface charge increases with the increasing number of nanoparticles at
the interface, the ionic strength increases, inducing the colloidal instability and
aggregation.22
The proposed mechanism of interfacial aggregation is supported by the
observation of the aminated surface aggregation kinetics. The aminated
surface is positively charged at the pH of the colloid solution which attracts the
negatively charged nanoparticle leading to rapid attraction to the interface and
the subsequent interfacial ionic strength-induced aggregation. The proposed
mechanism is also corroborated by the concentration dependence of the
surface coverage.22 At high colloid concentrations (undiluted from the stock
solution) the colloid aggregates spontaneously on the surface, forming
multilayers up to three layers in thickness (based on the extinction). 18 The
increased concentration at the interface of charged particles and counter ions
increases the ionic strength, causing aggregation of the nanoparticles.22
The refractive index sensitivity of the aggregated surface shows only a small
improvement over the low coverage surfaces and the nanoparticles present in
solution.22 This indicates that the surface does not contain pairs or aggregates
of particles that are sufficiently close to one another to produce the enhanced
field and hence the enhanced scattering properties responsible for the SERS
and SEAS effects.22 This is in contrast to the aggregates prepared in solution
and then deposited on the surface. Here larger aggregates form in solution
containing several hundred nanoparticles which are then deposited on the
surface. Washing of the aggregate may remove the ionic bilayer, resulting in
aggregate annealing decreasing the separation between the aggregate
features and increasing the sensitivity.22
3.5 Conclusions
Nanofabricated surfaces may be formed at the silica-water interface where
the interfacial concentration of counter ions increases the ionic strength and
destabilises the colloid. Aggregation at the interface is characterised by a
sharp increase in the rate of adsorption and hence the rate constant. The
46
small clusters formed in the interfacial mechanism do not show significant
non-linear plasmon sensitivity to external refractive index changes.
47
3.6 References
1. Nie, S., Emory, S. R. Science 1997, 275, 1102.
2. Xu, H., Aizpurua, J., Käll, M., Apell, P. Physical Review E 2000, 62, 4318.
3. Osawa, M., Ataka, K. –I., Yoshii, K., Nishikawa, Y. Applied Spectroscopy
1993, 47, 1497.
4. O’Reilly, J. P., Fisk, J. D., Rooth, M., Perkins, E., Shaw, A. M. PCCP 2007,
9, 344.
5. Jana, N., R., Gearheart, L., Murphy, C., J. Journal of Physical Chemistry B
2001, 105, 4065.
6. Kelly, K. L., Coronado, E., Zhao, L. L., Schatz, G. C. J. Phys. Chem. B
2003, 107, 668.
7. Garcia-Vidal, F. J., Pendry, J. B. Physical Review Letters 1996, 77, 1163.
8. Murray, C. A., Allara, D. L., Rhinewine, M. Physical Review Letters 1981,
46, 57.
9. Shaw, A. M., Hannon, T. E., Li, F., Zare, R. N. J. Phys. Chem. B 2003,
107, 7070.
10. Fisk, J. D., Batten, R., Jones, G., O’Reilly, J. P., Shaw, A. M. J. Phys.
Chem. B 2005, 109, 14475.
11. Pipino, A. C R. Physical Review Letters 1999, 83, 3093.
12. Chen, C., Cheng, S., Chau, L., Wang, C., R., C. Biosensors and
Bioelectronics 2006, 22, 568.
13. Perez-Juste, J., Pastoriza-Santos, Liz-Marzan, L., M., Mulvaney, P.
Coordination Chemistry Reviews 2005, 249, 1870.
14. Murphy, C., J., Sau, T., K., Cole A., M., Orendorff, C., J., Gao, J., Gou, L.,
Hunyadi, S., E., Li, T. Journal of Physical Chemistry B 2005, 109, 13857.
15. Huang, C., Wang, Y., Chiu, P., Shih, M., Meen, T. Materials Letters 2006,
60, 1896.
16. Jana, N. R., Gearheart, L., Murphy, C. J. Phys. Chem. B 2001, 105, 4065.
17. Turkevich, J., Stevenson, P. C., Hillier, J. J. Discussion Farady Society
1951, 11, 55.
18. Fisk, J. D., Rooth, M., Shaw, A. M. J. Phys. Chem. C 2007, 111, 2588.
48
19. O’Reilly, J. P., Butts, C. P., I’Anson, I. A., Shaw, A. M. J. Am. Chem. Soc.
2005, 127, 1632.
20. Arslan, G., Ozmen, M., Gunduz, B., Zhang, X., Ersoz, M. Turk J Chem
2006, 30, 203.
21. Frattini, A., Pellegri, N., Nicastro, D., de Sanctis, O. Materials Chemistry
and Physics 2005, 94, 148.
22. Kretzers, I., M., J., Parker, R., J., Olkhov, R., V., Shaw, A., M. J. Phys.
Chem. C. 2009, 113, 5514.
23. Schumacher, G. A., Ven, T. G. M. v. d. Faraday Discussions of the
Chemical Society 1983, 83, 75.
24. Israelachvili, J. Intermolecular and Surface Forces, Second ed., Academic
Press London, 1992.
49
4 DNA Hybridization
4.1 Introduction
Even though there has been extensive investigation of the physico-chemical
properties of DNA, much work is still needed in order to understand its
complex behaviour inside the cell.1,
2
Therefore, DNA is an important
analytical tool in molecular biology, in which most of the techniques that are
utilized depend on a hybridization event in which the target is identified by a
probe.3 Commonly used DNA techniques are based on fluorescence labelling,
which is time consuming, expensive and not sensitive enough.3
PCR is the standard method for DNA sequencing. The PCR method has
disadvantages. The first one is that during the process of multiplying DNA,
specific target bases are labelled using a fluorophore. 4, 5 It is well known that
this interferes with (de)hybridization of DNA (the melting temperature).4, 5 The
second problem is non-specific binding of fluorophores to bases. This will also
interfere with the melting temperature.4, 5 The melting temperature of DNA is
the temperature at which the double strand de-hybridizes.
The previously mentioned disadvantages of using DNA techniques based on
fluorescence labelling could make e-CRDS a good analytical tool for DNA
hybridization. Using DNA of different strand lengths tethered to nanoparticles
will result in a biological measurement of sensitivity.
4.1.1 Current trends in DNA biosensors
A major trend in the development of novel diagnostic systems is the concept
of DNA chips (or microarrays). By using different techniques, sensor surfaces
are functionalized using different printing techniques. This results in very high
microband sensor arrays coated with different DNA probes (with or without
label) on the chip.6 This paragraph will give a short review on the current
developments of optical analysis techniques that eventually should be utilized
on these arrays.
50
The optical methods that are currently being developed can be broadly
divided into four categories. These categories are; optical fibers, SPR, gold
nanoparticles and quantum dots.7
An optical fiber is generally utilized by placing a probe (in this case sDNA) to
the end of the optical fiber after which hybridization with the complementary
strain takes place.8 Measurements can be done by a change in fluorescence
intensity after hybridization.8 Previously, these measurements were performed
using ethidium bromide as a hybridization indicator. Ethidium bromide is a
carcinogenic compound and much development is focussed on finding a
replacement of this chemical.
Directing light waves to an interface between a metal and a dielectric will
result in SPR. Literature shows the utilization of a DNA detection system by
using an avidin coated surface with an immobilized biotinylated probe and
further binding of the target-DNA.9 Systems like this show high specificity
hybridization within 10 minutes at room temperature. 10 The use of
functionalized gold nanoparticles provides for another label-free optical
detection method for DNA hybridization, with the added advantage of less
background noise than with fluorescence tagging.11 According to the
literature7, gold nanoparticles are generally utilized in solution in which the
formation of complex aggregates and the associated colour change during
hybridization can be measured using photo spectroscopic methods. A recent
reduction in background signal was developed by coupling gold nanoparticles
to latex microparticles in which both particle types are linked to sDNA
probes.12
Quantum dots are dots that consist of nanoparticles for fluorescence tagging
of probe biomolecules.7 The difference compared to conventional organic
fluorophores is that quantum dots are much brighter (higher quantum yield)
and more photostable. An application of quantum dot-based DNA analyses is
a surface plasmon enhanced fluorescence microscopy detection scheme in a
microarray format.13
51
4.1.2 Kinetics
Using sDNA strains of different length and different base composition might
give us an understanding of the association/dissociation rates of DNA.
Literature shows an association constant (kon) ranging from 3.7 × 104 M-1 s-1
to 2.5 × 105 M-1 s-1.14,
15
These numbers were determined using a simple
association/dissociation model (Langmurian). Under the assumptions that the
concentration of targets is in excess compared to the surface density of
probes, the bulk concentration is uniform over the entire sensor area and that
the dissociation rate is negligible.14 DNA hybridization was performed under
room temperature condition.14,
15
The number of bases in the sDNA strains
ranged from 15-20.14, 15
4.1.3 DNA
DNA consists of 4 different bases, adenine, thymine, cytosine and guanine.
The complementary base-pairs are adenine-thymine (A-T) and cytosineguanine (C-G) (see Figure 4-1).16
52
Purines
O
NH2
N
N
NH
N
N
N
N
N
NH2
NH2
ribose
ribose
adenosine
guanine
Pyrimidines
O
CH3
CH3
HN
O
NH2
N
N
ribose
thymine
O
N
ribose
cytosine
Figure 4-1: The four bases that make up DNA.
Figure 4-2 shows the double helix formation of the different base pairs that
can be formed.
53
HO
OH
P
O
H3C
H
O
NH
N
OH
H
O
N
H
H
N
N
N
N
O
O
H
O
HO
P
O
O
NH H
N
O
O
N
N
H
O
O
P
O
OH
O
N
H
O
NH H
N
O
O
N
N
O
P
O
H
H
HO
CH3
H
O
CH3
H
O
O
P
O
OH
H
N
H
N
H
N
O
N
O
N
H
O
H
HO
O
O
P
HO
OH
Figure 4-2: Structure of DNA. The strand on the left side has the following bases (from
bottom to the top), adenine, adenine and thymine. The chain on the left side is the
complementary strand.
The DNA strand is held together with H-bridges that are formed between the
primary amine groups in the structure of the bases, in which two H-bridges
are formed between A and T and three H-bridges are formed between C and
G. Therefore the C-G bond should be about 30% stronger then the A-T bond,
which makes the A-T bond more sensitive to changes in temperature.
4.1.4 Binding Strategies
Two different binding strategies have been used to tether DNA to gold
nanoparticles.
The first method consists of a direct binding to the gold
nanoparticle by using sDNA with a thiol-group at the 3’-end of the strand. The
thiol-group will directly bind to the gold nanoparticle, after which, the
hybridization experiment could be performed (see Figure 4-3). The other
strand is of course unmodified.
54
sDNA
sDNA
sDNA
sDNA
sDNA
sDNA
sDNA
sDNA
S
S
S
S
S
S
S
S
Au
Au
Au
Au
Au
Au
Au
Au
Figure 4-3: Schematic drawing depicting thiol-sDNA binding to a gold surface.
The second strategy consisted of EDC/NHS coupling of an amine group to a
carboxylic acid group already present on the gold nanoparticle. In the
experiments that are presented in this chapter, the gold surface was modified
by a self assembled monolayer (SAM), using thioctic acid (TOA) (see Figure
4-4).
O
HO
O
O
HO
HO
O
HO
Thioctic acid
S
S
S
Au
Au
S
Au
Au
Gold surface
Figure 4-4: Schematic drawing depicting a SAM of TOA on a gold surface.
EDC is a water soluble carbodiimide. Used on its own, EDC mediates amide
bond formation between amine groups on proteins and a carboxylic acid
group on the surface. It does this by forming an O-isourea group that readily
reacts to form the amide bond (see Figure 4-5). The strains that were used
during experiments were initially sDNA strains. The risk with performing the
binding procedure using sDNA instead of DNA is that the primary amine
55
groups on the bases could just as easily bind to the activated surface. This
will result in a strand that is bound base down on the gold nanoparticle so that
hybridization will not occur.
+
R1
EDC
H3C
O
N
N
OH
CH3
N
HCl
+
H
CH3
H3C
CH3
HN
N
N
CH3
O
Unstable Amine Reactive Intermediate
O
R1
sDNA
NH2
O
O
sDNA
R1
HCl
CH3
+ H3C
NH
NH
NH
N
HCl
CH3
Figure 4-5: Mechanism of EDC – amine coupling
EDC is often used in conjunction with NHS.17 Using EDC and NHS together
produces a more stable intermediate ester that is then able to react with
amine groups to form an amide bond (see Figure 4-6).
56
EDC
H3C
NHS
O
N
+
R1
N
OH
+
CH3
O
N
HCl
N
O
HO
CH3
H3C
O
O
HN
R1
O
+
N
NH
O
O
CH3
HCl
N
CH3
NH2
sDNA
O
sDNA
R1
O
+
N
NH
O
HO
Figure 4-6: Mechanism of EDC/NHS –amine coupling.
A molecule that is bound to the surface using the EDC/NHS coupling will be
more greatly adhered as the bond strength of an amide bond is greater than
physisorption due to electrostatic forces.
57
4.2 Experimental Design
4.2.1 The sequence
The plasmon field of a gold nanoparticle is excited into its direct surroundings.
The penetration depth of this field is about 20 nm. For the first experiments
the strand was designed to stay in this field. Ten base pairs are about 3.4 nm
long so 15 base pairs are about 5.1 nm long. Strands no longer then 15 base
pairs were chosen because previous research in the literature shows that
most strains that are used for hybridization are not much longer than 15 base
pairs.14,
15
The primer sequence consists of 5 codons (or 15 bases). The
sequence being used for these experiments is:
5’-TTTTAACCCGGGGAC-3’
The complementary sequence is:
3’-AAAATTGGGCCCCTG-5’
The modifications to the 3’-end of the strand were either a thiol- or aminogroup (see Figure 4-7).
5'-end
5'-end
R
R
HO
P
O
O
NH H
N
O
HO
P
H
N
H
O
N
O
H
N
H
N
NH H
N
O
N
N
O
O
H2N
HS
3'-end
3'-end
Figure 4-7: Structure of the chemical modifications used during the experiments. The
OH group on the ribose is replace by either SH or NH2.
Figure 4-7 shows the modifications that were made. The R in the structure is
followed by the rest of the sequence.
58
The thiol group would bind to the gold surface directly (see Figure 4-3) and
the amino-group could be used after modifying the gold surface first
(EDC/NHS coupling, see Figure 4-6). All strains were ordered at MWGbiotech.
4.2.2 sDNA adsorption to gold surfaces using EDC/NHS coupling
A citrate-reduced colloid was prepared using the Turkevich method.18 This
method is described in chapter 2 in more detail. The protocol was as followed;
100 ml of 1mM HAuCl4 was reduced using 10 ml of 90 mM of Sodium citrate.
The HAuCl4 solution was heated to 95 oC and the citrate was injected whilst
stirring vigorously. The colloid was cooled to room temperature, after which,
the absorbance spectrum was measured, from which max = 523 nm indicates
a colloid with 15 nm diameter particles.19 The colloid solution was diluted with
8.2 mM sodium citrate to reduce the concentration to 0.35 nM before use in
the aggregation experiments.
Surface equilibration was carried out using ethanol for 20 minutes. After that,
a solution of 1 mM of thioctic acid (TOA) was introduced into the flow cell. To
get a self-assembled monolayer (SAM) this was done for 45 minutes.
The next step was to stabilize the SAM and to remove excess TOA using
clean ethanol. SAM activation was performed by using a solution of Nhydroxysuccinimide
(NHS)
and
1-ethyl-3-(3-dimethylaminopropyl)
carbodiimide hydrochloride (EDC). The EDC/NHS solution (0.4 M/0.1 M) was
injected across the surface for 1 hour. After 1 hour, clean water was injected
to remove any remaining EDC/NHS solution. A new baseline was established
using a phosphate buffer after which sDNA was flown across the surface.
4.3 Results
The results presented here are preliminary and form an evolving series of
experiments towards the objective of observing the hybridisation kinetics.
59
4.3.1 Adsorption and
nanoparticles
hybridization
of
sDNA
onto
gold
The first set of DNA hybridization experiments that were performed used
sDNA in which the 3’-end of the strand was modified in such a way that it
ended with a thiol group. The theory was that this thiol-group should bind to
gold nanoparticles, after which, the complementary strand could be hybridized
to the single strand already present on the gold surface. This complementary
strand did not have a thiol group at the 3’-end of its strand; therefore it would
not bind to the gold surface (see Figure 4-3 for a schematic representation).
The first step in these experiments was to create a surface of gold
nanoparticles on the silica prism surface. In Figure 4-8, the first step is shown
in these experiments.
4
3.5
Extinction/10 -3
3
2.5
B
A
2
1.5
1
0.5
0
0
100
200
300
400
500
Time/s
Figure 4-8: This graph shows the adsorption of gold nanoparticles using the standard
colloid solution that has been described previously. The experimental cut off point was
chosen at about 3.5x10-3 at a wavelength of 635 nm. A and B are duplicate experiments.
60
The first phase of these experiments took about 5 minutes and the cut-off
values for extinction was about 3.5  10-3 to allow for sufficient dynamic range
in  to perform the other experiments. The difference in kinetics that can be
seen here is due to surface and colloid effects. The silica surface and the
colloid vary slightly every day. The next step is absorbing sDNA with a thiolgroup on the gold nanoparticle surface the results of this can be seen in
Figure 4-9.
A
B
Figure 4-9: This step is adsorption of the first sDNA strand onto the gold surface
(nanoparticle). A and B are duplicate experiments.
Both experimental traces in Figure 4-9 show a Langmurian trend. The arrows
in the graph show the switch to only PBS-buffer (with no sDNA). After that
switch, the signal goes down. This is probably due to rinsing of excess sDNA.
This means that all binding sites are occupied.
61
The third and last step in these experiments was the hybridization of the
complementary strand. In Figure 4-12 the results can be seen from these
experiments.
1
0.8
B
Extinction/10 -3
0.6
0.4
A
0.2
C
0
0
2000
4000
6000
8000
10000
12000
-0.2
-0.4
Time/s
Figure 4-10: Hybridisation of the sDNA strand to the tethered DNA. A and B are the
signal changes corresponding to Figure 4-8 and C is the signal change corresponding
with the control experiment.
The signal changes in Figure 4-8 and Figure 4-9 could be attributed to
adsorption of gold nanoparticles and binding of sDNA to the gold nanoparticle.
Figure 4-10 however does not show that there is any complementary
hybridization of the two sDNA strands. Also a control experiment was done to
determine any non-specific binding of unmodified sDNA to the gold
nanoparticle.
62
4.3.2 sDNA adsorption to gold surfaces using EDC/NHS coupling
The last experiment that was performed involved absorbing a thioctic acid
layer on the gold surface and activating it by using EDC/NHS coupling; after
which sDNA was tethered to the activated surface. The results are described
in this section.
The next three graphs show the results of one single experiment. Figure 4-13
represents the signal change when gold is absorbed onto the silica surface.
Again, the dynamic range of the experimental setup has to be taken into
account or else there it will not be possible to measure the other chemical
steps.
0.80
0.70
0.60
Extinction/10-3
0.50
0.40
0.30
0.20
0.10
0.00
0
500
1000
1500
2000
2500
3000
-0.10
Time/s
Figure 4-11: Adsorption of gold nanoparticles onto the silica prism surface using the eCRDS technique to measure the increase in adsorption of light at 635 nm.
The first phase of these experiments took about 45-50 minutes. This is
significantly longer than the adsorption phase in Figure 4-8. A reason for this
63
might be that the gold colloid is not as stable as was assumed (also in chapter
3 colloid stability seems to be changing).
Figure 4-12 shows the signal change due to the deposition of TOA onto the
gold surface.
0.18
0.16
0.14
Extinction/10-3
0.12
0.10
0.08
0.06
0.04
0.02
0.00
0
500
1000
1500
2000
2500
3000
-0.02
-0.04
Time/s
Figure 4-12: Signal change during the deposition of thioctic acid onto a gold
nanoparticle surface.
The second phase of these experiments was the deposition of thioctic acid
(concentration) onto the gold nanoparticle. The deposition time was about 45
minutes as can be seen in Figure 4-12. After about 40 minutes the signal
change seemed to level off at an extinction value of 0.15x10 -3. To check if the
surface was deposited correctly pH switches were carried out. The pH shift
will result in a change of the surface charge which will show as a shift in signal
change. The pH range that was utilized to show TOA binding ranged from 3 to
64
12 using acetic acid for the acid value and sodium hydroxide for the base
value.
The next phase is activating this thioctic surface by using EDC/NHS coupling
to connect the sDNA onto the thioctic acid monolayer. This step was repeated
three times to show reproducibility. In Figure 4-13 the results of this phase are
presented.
0.60
Extinction/10 -3
0.50
0.40
0.30
0.20
0.10
0.00
0
50
100
150
200
250
300
Time/s
Figure 4-13: Attachment kinetics of sDNA onto the thioctic acid monolayer using the
EDC/NHS activation method.
The activity of this method lasts very shortly. Also the concentration of the
chemicals was very high. This might explain why the kinetics are so fast. The
chemical binding between DNA and thioctic acid reaches a stable value after
about 20 seconds.
To exclude a shift in the surface plasmon due to a stability change in the
cavity this experiment was done three times to confirm the results.
65
4.4 Discussion
The experiments that were performed were designed following two phases.
The first phase consisted of tethering sDNA onto gold nanoparticles. The
strain that had a thiol group at its 3’-end bound to the gold nanoparticles,
which is reflected in Figure 4-9. The second tethering strategy; the formation
of a thioctic acid self assembled monolayer on the gold surface followed by
the activation of the monolayer using EDC/NHS also resulted in the
adsorption of sDNA onto the gold nanoparticles. The result can be seen in
Figure 4-13.
The second phase of these experiments was supposed to consist of
complementary hybridization of the sDNA strands. This phase has been
unsuccessful. The sequence of the two sDNA strands consisted of all four
bases divided in five codons. There is a possibility that during the experiment
these strands hybridized with themselves and that therefore no hybridization
occurred. EDC/NHS activation of the thioctic acid surface makes it possible to
bind with an amine group. This could have resulted in binding of the primary
amine groups of the bases inside the sDNA strands onto the activated thioctic
acid surface, meaning that the sDNA strands will have bound base down and
that hybridization can therefore not occur.
4.5
Conclusions and Future Work
From this chapter the conclusion can be drawn that both binding strategies
seem to work. Binding sDNA directly to gold nanoparticles using a thiol group
this has been shown in a duplicate experiment and for binding sDNA to an
activated thioctic acid surface, this has been performed in triplicate.
The best way to move forward with these experiments is to first bind double
stranded DNA onto the gold nanoparticle, then de-hybridize it, followed by
hybridization of the complimentary strand. If this works, then strands with
varying G-C content can be tested to see whether or not there is a difference
in binding kinetics.
66
4.6 References
1. Zhang, Z., H., Feng, C., L. Applied Surface Science 2007, 253, 8915.
2. Ageno, M., Dore, E., Frontali, C. From the Physics Laboratory, Instituto di
Sanita, Rome, Italy.
3. Zhang, G., Chua, J., H., Chee, R., Agarwal, A., Wong, W., M., Buddharaju,
K., D., Balasubramanian, N. Biosensors and Bioelectronics 2008, 23,
1701.
4. Niu, S., Singh, G., Saraf, R., F. Biosensors and Bioelectronics 2007, 23,
714.
5. Zhang, X., Su, H., Bi, S., Li, S., Zhang, S. Biosensors and Bioelectronics
2009 (Accepted).
6. Pividori, M., I., Merkoci, A., Alegret, S. Biosensors and Bioelectronics
2000, 15, 291.
7. Teles, F., R., R., Fonseca, L., P. Talanta 2008, 77, 606.
8. Piunno, P., A., E., Krull, U., J., Hudson, R., H., E., Damha, M., J., Cohen,
H. Anal. Chem. 1995, 67, 2635.
9. Wood, S., J. Microchem. J. 1993, 47, 330.
10. McCormack, T., Keating, G., Killard, A., Manning, B., M., O’Kennedy, R.
Principles of Chemical and Biological Sensors, John Wiley & Sons 1998,
Toronto, 133.
11. Fritsche, W. Rev. Mol. Biotechnol. 2001, 82, 37.
12. Reynolds, R., A., III, Mirkin, C., A., Letsinger, R., L. Pure Appl. Chem.
2000, 72, 229.
13. Robelek, R., Niu, L., Schmid, E., L., Knoll, W. Anal. Chem. 2004, 76, 6160.
14. Schultz, E., Galland, R., Bouetiez, D., D., Flahout, T., Planat-Chretien, A.,
Lesbre, G., Hoang, A., Volland, H., Perraut, F. Biosensors and
Bioelectronics 2008, 23, 987.
15. Chu, L., Knoll, W., Forch, R. Biosensors and Bioelectronics 2008, 24, 118.
16. Watson, J., D., Crick, F., H., C. Nature 1953, 4356, 737.
67
17. K. Bonroy, F. Frederix, G. Reekmans, E. Dewolf, R. De Palma, G. Borghs,
P. Declerck, B. Goddeeris, Journal of Immunological Methods, 2006, 321,
167.
18. Turkevich, J., Stevenson, P. C., Hillier, J. J. Discussion Farady Society
1951, 11, 55.
19. Fisk, J. D., Rooth, M., Shaw, A. M. J. Phys. Chem. C 2007, 111, 2588.
68