Adsorption of Surfactants at the
Solid-Liquid Interface:
A Quartz Crystal Microbalance study
Johan J.R. Stålgren
Doctoral Thesis 2002
Department of Chemistry, Surface Chemistry
Royal Institute of Technology
Stockholm, Sweden
Akademisk avhandling som med tillstånd av Kungliga Tekniska Högskolan framlägges tilloffentlig granskning för avläggande av filosofie doktorsexamen, tisdagen den 29 januari,2002, kl. 09.00 i Kollegiesalen, Valhallvägen 79, KTH, Stockholm.
Address to the author:Johan J.R. Stålgren
Department of Chemistry, Surface ChemistryRoyal Institute of Technology
SE-100 44 StockholmSweden
ISSN 1650-0490ISBN 91-7283-238-XTRITA YTK-0201
Copyright 2002 by Johan J.R. Stålgren. All rights reserved. No part of this thesis may bereproduced without permission from the author.Other copyrighted material is used with permissionPaper I 2001 by the American Chemical SocietyPaper II 2001 by the Elsevier Science
Abstract
This thesis is concerned with the interfacial behaviour of surfactants at solid-liquid interfaces.
The main technique used for the adsorption measurements during this thesis work was the
Quartz Crystal Microbalance-Dissipation (QCM-DTM). This technique allows both the
adsorbed amount, as evaluated as a change in frequency (∆f), and the change in the dissipation
factor (∆D) that is a measure of the energy dissipated in the system, to be determined
simultaneously. Methods like null ellipsometry already exist, and they measure the amount
adsorbed to a planar, reflecting surface accurately, and the thickness of the adsorbed layer
may also be determined. The QCM-DTM technique has, however, some advantages. For
instance, the quartz crystal can be coated (physical/chemical) in a large number of ways. In
addition, simultaneous measurement of the dissipation factor allows another parameter to be
determined, this parameter is a measure of the interaction between the adsorbed surfactant
layers and the bulk solution. Further, opaque or even non-transparent solutions can be studied
with the QCM-DTM, which is not possible with the ellipsometer.
When the project began, an aim was to investigate the visco-elastic properties of
polyelectrolytes at different surfaces. This turned out to be more complex than we expected so
I decided to use a less complex systems in order to more fully understand the results. Hence,
the choice became to study surfactant adsorption, a topic which is well documented before by
several other techniques. The choice was based on the surfactants low molecular weight, and
the relatively simple distribution of a polar (hydrophilic) and non-polar (hydrophobic) part,
and a significant general knowledge about their interfacial behaviour.
The methodology for adsorption studies in liquid for the QCM-DTM was only in its infancy, so
parameters like temperature dependence, surface roughness, surface modification and
cleaning had to be kept under control or developed at the same time. Systematic surfactant
adsorption studies from liquids with the QCM technique do not exist. Hence, the aim of this
thesis was to achieve an understanding of the information provided by measured shifts in
frequency and dissipation factor for such systems, and from this draw conclusions about the
interfacial behaviour of both non-ionic and cationic surfactants. Further I aimed to learn how
valuable the QCM-DTM technique was for these systems and what pitfalls there are in
evaluating the results observed with this technique.
Sammanfattning
Den här avhandlingen handlar om det beteende som uppträder hos tensider, vid gränsytan
mellan fast fas och en vätska. Vid adsorptions mätningarna i denna avhandling har
huvudsakligen en teknik använts, en kvarts kristall mikrobalans (QCM-DTM). Denna teknik
tillåter både den adsorberade mängden, utvärderat genom skiftet i frekvens (∆f), samt
förändringen i dissipations faktorn (∆D), som är ett mått på systemets energi dämpning, att
uppmätas. Båda dessa faktorer kan uppmätas samtidigt. Metoder som noll ellipsometri
existerar redan och de mäter adsorptionen vid en plan, reflekterande yta. Denna teknik kan
också bestämma tjockleken hos det adsorberade skiktet. QCM-DTM tekniken har trots allt
några fördelar gentemot detta. Till exempel, kan kvarts kristallerna beläggas (fysiskt/kemiskt)
på flera olika sätt. En annan fördel är den samtidigt uppmätta dämpnings faktorn som tillåter
en parameter till att bestämmas. Dämpnings faktorn, är en parameter som gör att man kan
mäta styrkan hos interaktionerna mellan det adsorberade tensid skiktet och bulk lösningen.
Även lösningar som ej är genomskinliga kan studeras med QCM-DTM tekniken, detta är inte
möjligt med en teknik som till exempel ellipsometri.
När projektet började, var vårt mål att studera de viskoelastiska egenskaperna hos laddade
polymerer vid olika ytor. Detta visade sig vara mycket mer komplext än vad vi hade förväntat
oss, så vi bestämde oss för att använda ett mindre komplext system för att fullständigt förstå
våra resultat. Vårt val föll på olika tensid lösningar, ett ämne som är väl dokumenterat sedan
innan med flera olika tekniker. Valet grundade sig på tensidernas låga molekylära vikt samt
deras relativt enkla distribution av polära (hydrofila) och o-polära (hydrofoba) delar.
Dessutom fanns sedan innan en signifikant kunskapsbas om deras beteende vid gränsytan
mellan en fast fas och en vätska.
Eftersom metodologin för adsorptions studier i vätskor för QCM-DTM var bara i början på sin
utveckling behövdes faktorer som temperatur beroendet, ytråhet, ytmodifikation samt
rengöring kontrolleras samt utvecklas under tiden. Systematiska studier av tensid adsorption i
en vätska finns inte för QCM tekniken. Därav valet på avhandlingens innehåll, en ökad
förståelse av informationen för dessa system, genom studier av ändringen i frekvens samt
förändringen i dämpning. Samtidigt siktade jag på att lära mig hur värdefull QCM-DT M
tekniken var för dessa system, och vilka fallgropar det finns när man utvärderar resultaten från
denna teknik.
There are two kinds of scientists,
physicists and stamp collectors.
Ernest Rutherford (1871-1937)
Contents
1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11.1. List of paper . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .31.2. Summary of papers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
2. Surfaces . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 92.1. Gold . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .112.2. Self-Assembled Monolayers (SAMs) . . . . . . . . . . . . . . . . . . . . . . . . . 132.3. Silica . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 152.4. Silan-coated surfaces . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .16
3. Methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 183.1. Profilometer . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .183.2. Contact Angle . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 203.3. X-ray Photoelectron Spectroscopy (XPS) . . . . . . . . . . . . . . . . . . . . . .223.4. The Quartz Crystal Microbalance-Dissipation (QCM-DTM) . . . . . . . . 23
4. Surfactants . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 344.1. Lipids . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . .344.2. Cationic surfactans . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 374.3. Non-ionic surfactants . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38
5. Results and Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 395.1. Adsorption of emulsion studied with the QCM-DTM . . . . . . . . . . . . . . 395.2. Polymer adsorption on phospholipid coated surfaces . . . . . . . . . . . . . 415.3. Adsorption of surfactants studied with the QCM-DTM . . . . . . . . . . . . .475.4. Counterion effects on sensed mass and energy dissipation . . . . . . . . . 495.5. Bound / trapped water . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .52
6. Concluding remarks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 55
Acknowledgements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 56
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .57
0
1
1. Introduction
The Quartz Crystal Microbalance-Dissipation (QCM-DTM) technique is an ultra sensitive
weighing device based on the piezoelectric, electromechanical oscillator principle. It consists
of a thin single-crystal quartz disk, with one metal electrode deposited on each side. When the
electrodes are connected to an electric oscillator, the crystal can be made to oscillate in a very
stable manner at its resonance frequency, f. When a mass is adsorbed on one or both of the
electrodes, then this leads to a change in the resonance frequency of the quartz crystal, ∆f. If
the adsorbed mass is small compared to the mass of the quartz crystal and there is no slip or
deformation due to the oscillatory motion, then the resonance frequency decreases
proportionally to the mass of the adsorbed film according to the Sauerbrey relation. It is
possible to determine very small changes of the resonance frequency and hence very small
mass changes. This is possible since the QCM generally has very stable oscillations. In
addition to the adsorbed mass, simultaneous measurements of the change in dissipation factor,
(∆D), which is a measure of the energy dissipated in the system, is possible. Hence, this
parameter is a measure of the interaction between the adsorbed layer and the bulk solution.
This thesis is concerned with the interfacial behaviour of surfactants at solid-liquid interfaces.
Emphasis is placed on the adsorption / desorption of three different groups of surfactants;
cationic, non-ionic, and phospholipid surfactant.
To choose the surfactants to study was not easy, but it had to be surfactants with properties
which were well documented before by several different techniques. Further, to be able to
systematically vary the surfactant structure was regarded as important. I have used different
model surfaces to study the effect of the underlying substrate. The model surfaces also had to
be well characterized with several different techniques. The different model surfaces we
decided to work with were a metal (gold), silica, methylated silica, and several different self-
assembly monolayers on gold substrates. These surfaces had to be thoroughly evaluated to be
valuable for the QCM experiments, without adding more unknown parameters to the
interfacial study. In chapter 3 all the techniques used for evaluating both the surfaces and the
surfactants are discussed.
Systematic adsorption studies of surfactants from liquids using the QCM technique do not
exist. Hence, the aim of this thesis was to achieve an understanding of the information
2
provided by the measured shifts in frequency and dissipation factor for such system, and from
this draw conclusions about the interfacial behaviour of both non-ionic, cationic and
phospholipid surfactants. Last in the summary the main findings during these experiments,
and a hopefully valuable discussion of the different results obtained, is presented. More
details can be found in the manuscripts that constitute the second part of this thesis.
3
List of papers
This thesis consists of a summary and six papers. The papers are listed below and are in the
summary referred to by their Roman numerals (I to VI).
I Adsorption of Liposomes and Emulsions Studied with a Quartz Crystal
Microbalance.
Johan J.R. Stålgren, Per M. Claesson and Torbjörn Wärnheim
Advances in Colloid and Interface Science, 2001, 89-90, 383.
II Adsorption of a PEO-PPO-PEO Triblock Copolymer on Small Unilamellar
Vesicles: Equilibrium and Kinetic Properties and Correlation with
Membrane Permeability.
Markus Johnsson, Nill Bergstrand, Katarina Edwards and Johan J.R. Stålgren
Langmuir, 2001,17, 3902.
III Cationic and Non-ionic Surfactant Adsorption on Thiol Surfaces with
Controlled Wettability.
Katrin Boschkova and Johan J.R. Stålgren
Submitted to Langmuir.
IV A Correlation between Adsorbed Amount and Frictional Properties of Thin
Gemini Surfactant Films - CPP in Relation to Friction.
Katrin Boschkova, Adam Feiler, Bengt Kronberg and Johan J.R. Stålgren
Submitted to Langmuir.
V A Comparative Study of Surfactant Adsorption on Model Surfaces using
the Quartz Crystal Microbalance and the Ellipsometer.
Johan J.R. Stålgren, Jonny Eriksson and Katrin Boschkova
Submitted to Journal of Colloid and Interface Science.
4
VI Lubrication in Aqueous Solutions Using Cationic Surfactant –
a Study of Static and Dynamic Forces.
Katrin Boschkova, Bengt Kronberg, Johan J.R. Stålgren, Karin Persson, and
Monica Ratoi-Salagean
Accepted in Langmuir.
The papers are reproduced with permission from the publishers.
The author’s contribution to the papers is as follows:
I Major part of planning, experiments and evaluation.
II Part of planning, experiments and evaluation.
III Major part of planning and experiments, part of evaluation.
IV Part of planning, experiments and evaluation.
V Major part of planning, experiments and evaluation.
VI Part of planning, experiments and evaluation.
In all papers, I have been the main responsible for the QCM work and evaluation.
5
1.2 Summary of papers
Phospholipid adsorption at the solid-liquid interface.
The first two papers deal with adsorption of phospholipids at a gold surface, and effects of
additives. The results are important for comprehending the data obtained with the Quartz
Crystal Microbalance-Dissipation (QCM-DTM), in a useful way.
Paper I deals with adsorption from phospholipid liposome solutions (1.2%) and phospholipid
stabilised oil-in-water emulsions (20% purified soybean oil) with the same phospholipid
concentration. The main attention in the paper was given to the adsorption process at a gold
surface and the effect of repeated injections of the same solution. The second aim was to learn
how the dilution of the bulk solution affected the adsorbed layer and to determine what
remained on the surface after the dilution step was completed. The adsorption from the
liposome solution resulted in formation of a phospholipid bilayer with an additional and
incomplete outer layer of liposomes. The outer layer was removed by dilution leaving a
bilayer of phospholipids on the surface. The adsorption process observed from the
concentrated emulsion solution was considerably more complex. A slow spreading process
that also resulted in some expulsion of material from the interface followed the rapid initial
adsorption of emulsion droplets. After rinsing with water a phospholipid monolayer was
retained on the surface.
Paper II is devoted to the adsorption of the triblock copolymer F127, poly(ethylene oxide)-
poly(propylene oxide)-poly(ethylene oxide), EO98PO67EO98, onto immobilized small
unilamellar vesicles (SUVs) of egg phosphatidylcholine (EPC). With the QCM-DTM technique
we first showed that SUVs of EPC adsorb on gold to form a monolayer of vesicles. This
supported monolayer of vesicles was then used to follow the adsorption of the F127 polymer
onto the lipid vesicle membrane surface. The adsorption of F127 was found to be a rapid
process and the measured polymer binding isotherm was fitted to a Freundlich type of
isotherm. The maximum, or plateau, adsorbed amount was determined to be of a magnitude
similar to that found for adsorption of F127 on hydrophobic surfaces. Furthermore, the
desorption of the triblock copolymers from the membrane surface was followed after rinsing
the SUV monolayer with pure buffer. It was found that the desorption process displayed
essentially the same rapid kinetics as the adsorption process, indicating a weak interaction
between the polymers and the lipid membrane. The determined polymer binding isotherm was
6
used to correlate the adsorbed amount of polymer with the polymer-induced leakage of
carboxy fluorescein (CF) from the SUVs. It was found that the membrane permeability was
increased severalfold already at low surface coverage, and that the maximum magnitude of
the CF release rate was obtained at, or close to, the F127 concentrations giving rise to
maximum adsorbed amount of polymer. In addition, the increased membrane permeability
induced by the triblock copolymers was compared with the effect of adding a conventional
ethylene oxide (EO)-surfactant, Triton X-100, to the SUVs. The result emphasizes the
dramatic effect of F127 on the bilayer permeability. Another interesting result was that the
stability of the liposomes used in this study was considerably higher compared to those
formed by the phospholipids mixtures employed in paper I.
Surfactant adsorption at model surfaces.
In papers III-VI we started to modify our surfaces with silica, methylated silica and several
different self-assembled monolayers (SAMs). The ionic surfactants used were the cationic,
DTAB (dodecyltrimethylammonium bromide), DDAB (didodecyldimethylammonium
bromide), and gemini surfactants having the same headgroup and chain length as DTAB but
with the additional feature that two headgroups were chemically connected with a spacer of
different length. The non-ionic surfactants used were the poly(ethylene oxide) monoalkyl
ethers C14EO6 and C12EO8(Octa-(ethylene oxide) mono n-dodecyl ether).
In paper III we showed that thiolated surfaces work very well as model substrates in
adsorption measurements using the QCM-DTM. Functionalised SAMs were prepared from
mixtures of hydrophobic, SH-C16 (thiohexadecane) and hydrophilic, SH-C16OH
(thiohexadecanol) terminated thiols, which allowed the interfacial energy of the surfaces to be
changed in a systematic way. The prepared thiol surfaces were used as substrates for
adsorption of a cationic, DTAB, and a non-ionic, C12EO8, surfactant. The experiments showed
that when the fraction of methyl groups at the surfaces was increased, the adsorption of both
DTAB and C12EO8 is increased. In particular, there is a transition from a micellar surfactant
layer to a surfactant monolayer at 25% to 50% surface coverage of SH-C16 groups with
monolayers being formed at higher coverage of SH-C16. In addition, the role of the counterion
in the adsorbed surfactant layer for the charged surfactant was discussed in terms of its
contribution to the mass and visco-elastic response determined by the quartz crystal
microbalance.
7
With Paper IV where we used the Gemini surfactants, we showed that by changing the length
of the spacer group from 3 to 12 a systematic change in the molecular packing at a gold
surface was obtained. Furthermore, the molecular packing was shown to correlate to the
frictional behavior of the surfactant film. An increasing length of the spacer group resulted in
lower, adsorbed amount and less good frictional properties. This is discussed in terms of the
critical packing parameter (CPP) of the surfactant and a relation between CPP and frictional
behavior is proposed. The results can be viewed upon either as controlled by the rigidity of
the surfactant layer or as a result of defects, holes, in the lubricating film. No correlation
between spacer length and viscoelasticity of the adsorbed surfactant layer was detected using
the QCM-DTM. This indicates that the resolution of the dissipation factor from QCM-DTM
measurements is not sufficient to describe the viscoelastic character of the thin surfactant
film. The degree of counterion-binding to charged surfactant films is a difficulty encountered
when converting the frequency response of the crystal to packing density. This problem is
again highlighted and discussed (see also paper III).
In Paper V we investigated the adsorption behaviour of hexa-ethylene oxide mono n-
tetradecyl ether (C14EO6), on different model surfaces. This investigation was conducted with
two different techniques, the QCM-DTM and the ellipsometer. The adsorbed amount of the
non-ionic surfactant was determined both at hydrophilic and hydrophobic surfaces. In
particular, the substrates employed were; hydrophilic silica, hydrophobized silica (using
dimethyldichlorosilane), hydrophobized gold surfaces (using 10-thiodecane and 16-
thiohexadecane). We showed that the frequency shift obtained from the QCM-DT M
experiments results in an overestimation of the adsorbed mass. This is attributed to two
different effects, viz, hydrodynamic coupling of water to the adsorbed surfactant layer and
secondly, trapped water within the adsorbed surfactant layer. Furthermore, from the
ellipsometry data the adsorbed layer thickness was determined. By combining the thickness
information and the dissipation parameter (obtained from the QCM-DTM experiments), we
again noted that the dissipation parameter was insufficient in describing the visco-elastic
character of thin surfactant films.
Paper VI is devoted to lubrication in aqueous surfactant systems where the surfactants adsorb
at surfaces in relative motion forming either a surfactant monolayer or a multi (liquid
crystalline) layer. The surfactants were of two kinds, viz., a double chain cationic surfactant,
8
didodecyldimethylammonium bromide, DDAB, and a single chain cationic surfactant,
dodecyltrimethylammonium bromide, DTAB. Excellent film forming capability was shown
for DDAB. We interpret this as being due to good packing of the surfactant molecules at the
surfaces, i.e. the inherent ability of these surfactant molecules to form liquid crystalline
structures at the surface, results in good load carrying capability. This is also reflected in the
bulk properties of the surfactants, where DDAB shows lamellar liquid crystalline phases at
concentrations much lower than DTAB, which does not show good lubrication properties.
The results were discussed in terms of film stability of a surfactant layer adsorbed at the
surface, which in turn is correlated to the critical packing parameter of the surfactant. The
systems were characterized using (i) the surface force apparatus determining the interaction
forces between the adsorbed layers at the surfaces, (ii) the EHD-rig (Elastohydrodynamic-rig)
determining film formation under shear. The adsorption kinetics and composition at the
surface were determined by QCM-DTM and X-ray photoelectron spectroscopy.
9
2. Surfaces
A controlled surface/environment is required when studying interfacial properties of
surfactant molecules. Otherwise it is impossible to interpret the experimental results.
Model surfaces, where chemical and structural properties can be controlled, are not readily
found and in most cases one has to prepare and characterize them oneself.
Surfaces can be characterized and classified in many different ways. The surface topology
allows classification into “rough” and “smooth” as quantified by e.g. the root mean square
roughness, Rq. Other classifications can be based on the chemical composition of the surface,
the surface energy, or the wetting properties. The latter classification is particularly suitable
when studying surfactant adsorption from aqueous media. The wetting properties can also
very conveniently be quantified by the contact angle, θ, of the liquid on the solid, i.e. cosθ =
(γSV- γSL) / γLV. The cosine of the contact angle is thus given by the difference in surface
energy between the solid-vapour (γSV) and solid-liquid (γSL) interface, normalized by the
liquid-vapour interfacial tension (γLV). Generally, high energy solids have by definition a high
value of γSV, and in most cases, a much lower interfacial tension against water due to the
hydrogen bonding capability and high dipolar moment of the water molecule. The contact
angle of water on such surfaces is low. Based on the contact angle one can classify a given
liquid on a given surface as completely wetting θ = 0, partly wetting (0 < θ ≤ 90°) or non-
wetting (θ ≥ 90°). When the liquid is water one normally talks about hydrophilic and
hydrophobic surfaces, but there is no general agreement about what contact angle the surface
is required to have in order to be classified as “hydrophilic” or “hydrophobic”. In this thesis
we use the general term “hydrophilic” for surfaces having a low contact angle, and
“hydrophobic” for surfaces with high contact angle. The quantitative measure of the wetting
behaviour is provided by the contact angle. We note that the contact angle is extremely
sensitive to the surface composition and sub monolayer adsorption of hydrophobic
compounds is easily detected. In fact, in many cases simple contact angle measurement is a
more sensitive probe of adsorption than sophisticated XPS (X-ray Photoelectron
Spectroscopy) analysis. However, of course, the contact angle does not give the same
chemical information as the XPS-spectra.
10
In this work we have varied the wetting properties from that of hydrophilic silica to that of
hydrophobic alkane thiol SAM on gold surfaces. Some of the properties of these surfaces are
described more extensively below. The wetting properties of the surfaces have been
characterized using contact angle measurement; the surface composition has been determined
employing x-ray photoelectron spectroscopy (XPS, ESCA). The topological character of the
surfaces has been determined by the profilometer for surface roughness effects and for some
surfaces, the scanning electron microscope (SEM) looking for eventual defects. Some data
can be found in table 1, where we have summarized the characteristics of our model surfaces.
Surface Ra [nm] Rq / Ra θθθθ
Silicaellipsometer 1.3 ± 0.1 1.5 ± 0.3 < 20°
Dimetyldichlorosilaneellipsometer 1.1 ± 0.2 1.5 ± 0.5 101° ± 1°
SilicaQCM 1.3 ± 0.2 1.5 ± 0.3 < 20°
DimetyldichlorosilaneQCM 1.0 ± 0.1 1.4 ± 0.3 101° ± 1°
GoldQCM 1.4 ± 0.1 6.0 ± 3 ≈ 30° ± 5°
ThiohexadecaneQCM 1.2 ± 0.1 3.0 ± 1.0 103° ± 3°
ThiodecaneQCM 1.2 ± 0.1 3.0 ± 1.0 91° ± 3°
ThiohexadecanolQCM 1.2 ± 0.1 3.0 ± 1.0 20° ± 2°
Table 1: Characteristics of our model surfaces, where silicaellipsometer means a silica surface for ellisometry studies
and silicaQCM means a quartz crystal coated with silica for measurements with the QCM.
11
2.1 Gold.
The electrodes on the quartz crystal are made of gold, and in many cases this has been a
suitable surface to conduct some of our adsorption studies with. The water contact angle on
the gold surface obtained after cleaning (see table 1) was low indicating a low degree of
contamination, and this surface will henceforth be classified as “ hydrophilic gold” to
distinguish it from a gold surface that has been exposed to air for a prolonged time and
appears hydrophobic due to adsorption of contaminants. Gold is a material that has a partly
filled electron band in its ground state. This means that it has both empty states and electrons
in the valence band. The electronegativity1 for gold is 2.41, which is very high for being a
metal. The electronic configuration for gold is:
1s22s22p63s23p63d104s24p64d105s25p64f145d106s1
The 6s electrons move around freely in the gold crystal, whereas the 5d electrons are tightly
bound to the nucleus. These 6s electrons play a major role for the chemical bonding to gold
atoms in the surface layer. In the ideal crystalline structure of gold, the atoms are located
themselves in a face centered cubic (fcc)1 lattice, which means that the unit cell consists of a
cube with one atom in each corner and one at the centre of each side. Each atom is thus in
contact with 12 others. The fcc structure provides maximum number of nearest neighbours
and is thus the preferred structure of crystalline materials of spherical molecules or individual
atoms. However, the bulk order has to end somewhere near the surface, for gold it prefers to
end in a structure called the (111) surface1 of an fcc crystal. It gives the closest packing of
atoms in the surface layer. Each surface atom has 6 neighbours on the surface. Gold are in
practice a polycrystalline material2, where a lot of small single crystals are joined together,
and the borders between the crystals are very far from the perfect (111) surface. At a
hydrophilic gold surface, there are a lot of unpaired electrons, and they are highly reactive2.
This leads to a fast contamination of a clean surface exposed to air. To keep the surface clean
you can either store it in ultra-high vacuum and never let it come in contact with
contaminations (solid-gas), or you could clean it in-situ under clean solvent conditions (solid-
liquid). Since we are doing all our experiments at the solid-liquid interface, we have chosen to
clean our surfaces in-situ, and keep the exposure to air contaminants to a minimum This is
very demanding since you have to keep all other surfaces in the experimental setup equally
12
clean and exposed to a liquid that is as pure as the liquid in contact with the cleaned gold
surface to prevent a contamination transport from the other surfaces to the clean (highly
reactive) gold surface. It exists several other techniques to clean surfaces, but most of them
are still dependent on the environment you do your cleaning and experiments in.
13
2.2 Self-Assembled Monolayers (SAMs).
The self-assembly of molecules at the solid-liquid interface has been an area of growing
interest. There are different combinations of surfaces and molecules that form SAMs and
many of them rely on the strong interaction between one part of the assembling molecules and
the surface, in addition to the interaction between the molecules within the monolayer. Long-
chain organosulphur compounds on noble metals, such as silver, platinum and in our case
gold, can be used for forming SAM coated surfaces with stable monomolecular (monolayer)
films. It is a versatile preparation technique, and SAM coated surfaces have served as model
systems in a number of applications, such as for biosensors, biomaterials, anti-corrosion
agents and lubrication1-2. In 1983 Nuzzo and Allara3 published the first observations of
organic disulphides that formed monolayers on gold from solutions as studied with infrared
spectroscopy. The general picture for the chemisorption of alkane thiols on gold is that the
thiol moiety adsorbs in a three-fold hollow site at the Au(111) lattice whereupon it loses its
hydrogen atom to become a thiolate. This gives an area of 21.4 Å2 per thiol molecule. SAM
coated surfaces have been studied with a number of techniques, including x-ray photoelectron
spectroscopy (XPS, ESCA), infrared spectroscopy, scanning tunnelling microscope (STM),
electrochemistry, ellipsometry, contact angles and various diffraction methods, see for
instance the review by Ulman1. The experimental findings4-5 strongly support the model
proposed for the structure of SAMs, and so do theoretical calculations6. However, the detailed
quantum mechanical processes are not completely understood. The structure of SAMs has
been determined by infrared reflection absorption spectroscopy (IRAS). The distance between
the sulphur atoms on the gold surface is slightly larger than the closest possible separation
between two alkyl chains, allowing an approximately 28-30° tilt of the chains in the layer to
increase van der Waals interaction between the chains, as confirmed both experimentally7 and
theoretically8.
Whereas the adsorption of molecules on the surface is fast (minutes) the self-assembling
process into an ordered monolayer is quite slow (hours)9. After a few minutes alkane thiol
molecules forming an almost fully covered layer have been adsorbed, but the order in the
layer is low. For one of the thiols used by us, the SH-(CH2)15CH3 thiol, different groups have
obtained very different kinetics for forming a well ordered SAM, ranging from seconds to
hours10. The concentration of the thiol in the liquid from which the SAM is formed is, of
course, important to take into account when investigating SAM kinetics. At low
14
concentrations a diffusion limited adsorption kinetics has been observed4, whereas at high
concentrations the time limiting step in the process is the actual surface attachment1.
The stability of alkane thiol SAMs when immersed in a solution containing the SAM
molecules is excellent. It is a stationary system, i.e. there is a continuous exchange of
molecules from the monolayer at the surface to the bulk solution. If the bulk solution contains
another alkane thiol, the surface monolayer will be exchanged by the new alkane thiol from
the bulk solution. This exchange could take from hours to days depending on which SAMs
that are involved10. Even though molecules in the SAM can be exchanged for other SAMs, the
desorption of the SAM in contact with a SAM-free solution is very slow due to the interaction
with the surface and within the tightly packed layer. It is very easy to prepare mixed SAMs by
just mixing the adsorbing species in the solution (see paper III). The surface composition will
be highly dependent on properties like chain length, terminal functionality and solubility11.
The preparation procedure of thin monolayers of alkanethiols on surfaces is relatively easy.
One dissolves the film forming molecules in an appropriate solvent (in our case ethanol) to a
rather low concentration, for our SH-C16 a 1mM concentration is enough. A previously
cleaned gold surface is then immersed in the solution for approximately 24 hours. When the
surface is withdrawn from the solution it is rinsed and then put into pure solvent for
approximately a week in order to dissolve any loosely bound, physisorbed, molecules that
might be attached to the monolayer.
15
2.3 Silica.
The gold coated quartz crystals used in the QCM can be further modified with an evaporated
100 nm thick layer of SiO2. The roughness and contact angle is well defined, as described in
table 1. Silica has been widely used the last decade as a model for a hydrophilic surface. The
chemistry of silica is rather complex, and a more detailed description can be found
elsewhere1. SiO2 surfaces consist of two very different surface groups, relatively hydrophobic
siloxane (Si2O), and more hydrophilic silanol groups (SiOH). The silanol group is
amphoteric1, which means that it can act both as a base and an acid. Hence, when the silica
surface is exposed to water solution the surface charge is determined by the density of silanol
groups on the surface and both the ionic strength of the solution and its pH. This forced us to
perform all our experiments with SiO2 under controlled pH and electrolyte concentration, and
preferably keeping them constant.
In order to increase the number of silanol groups at the surface (making it more hydrophilic)
and to remove eventual contaminations, we treated the surface with surfactants (Hellman
ExTM) followed by plasma cleaning as described in detail in paper III.
Yaminsky et al2 have an explanation to the instability of SiO2 surfaces in water. The surface
decomposition of SiO2, into polysilicic acids may result in the formation of a diffuse silica gel
layer. This gel is probably the main reason for instability of our SiO2 surfaces in water, which
are shown as a small drift in the frequency, but not in the dissipation factor. This is contrary
to ellipsometric studies using silica surfaces where instabilities could be seen at the gas-solid
interface but not at the liquid-solid interface3.
16
2.4 Silan coated silica.
The silanol groups present at the surface of silica crystals allow a surface modification by
reactions with different types of silanes. Surfaces can for instance be made more or less
hydrophobic by reaction with different alkylchlorosilanes1-2. For all our measurements (QCM
and ellipsometry) we used a dimethyldichlorosilane to modify our silica crystals/wafers in
order to obtain a small surface roughness and a reproducible hydrophobicity. This silane does
not form large hydrophobic islands as probed by profilometry, whereas this was found to
occur when for instance dimethyloctylchlorosilane (DMOS) was used. When surfaces coated
with DMOS were used in surface force experiments, a long-range attractive force was
observed2. This may be a direct consequence of these silane islands present on the surface.
When two surfaces of DMOS are brought together for the first time, capillary condensation
immediately starts3 and drops of DMOS are formed. The reason for the long-range attraction
seen in the surface forces experiment was suggest to be is the coalescence of these drops
between the surfaces. The dimethyldichlorosilane does not show this behaviour, probably due
to the strong cross-linking between the dimethyldichlorosilanes in the hydrophobic layer and
no hydrophobic islands are formed. We note however, that a long-range attraction has also
been observed for silica surfaces coated with dimethyldichlorosilane6 despite that we do not
see any islands of silane on such surfaces. The reason may be that air-bubbles are attached to
the surface and it is the coalescence of these bubbles that gives rise to the attractive force.
This mechanism was first suggested by Parker et al. for other silane coated surfaces7. For a
further discussion on long-range attractive forces between non-polar surfaces in water we
refer the reader to a recent review9. It has been shown that correctly prepared
dimethyldichlorosilane coated surface are surprisingly stabile over several days, as long as
they are kept in Milli-Q water or in clean organic solvents4. Such good stability has not been
found for DMOS, which likely is due to the lack of crosslinking in the hydrophobic layer,
leading to a hydrolyse of the silanol-silane bond in water. This indicates that the silanes are
kept at the surface partly by the low solubility, and that they are only partly stabilized by
chemical reaction with silanol groups5.
The preparation of the silane surface is a process that easily can go wrong, because of that a
precise experimental protocol needs to be used. The method described below has shown to
have the highest success rate.
17
The SiO2 surfaces were cleaned in a mixture of 25% NH4OH, 30% H2O2 and H2O (1:1:5, by
volume) at 80° C for 10 min, followed by cleaning in a mixture of 25% HCl, 30% H2O2 and
H2O (1:1:5, by volume) at 80° C for 10 min. In between and after the cleaning procedures by
the two mixtures, the substrates were rinsed in water. The substrates were then immediately
put into a reactor and exposed to vapours of dimethyldichlorosilane for 24 hours. Afterwards
these substrates were rinsed in toluene, ethanol and water, followed by heating to 200°C for 1
hour. All substrates were stored in ethanol until use.
18
3. Methods
When positioned at one of the interfaces between chemistry and physics, called surface
science, you relatively soon realize that surface chemists are lacking in the knowledge of how
to characterize physical (mechanical), properties and surface physics are lacking in
knowledge about the basic chemical properties. Hence, in this chapter a number of analytical
methods that can be used to study thin surface layers and surface structures are described. I
will mainly discuss the Quartz Crystal Microbalance-Dissipation (QCM-DTM) technique.
However, also some of the other techniques that I have been using, and which I consider
being the most valuable ones for the characterization of my model surfaces, will be briefly
presented.
3.1 Profilometer.
The surface roughness analysis was carried out using a Zygo View 5010TM, which is a non-
contacting technique (see figure 1 for a schematic illustration). It is a precision vertical
scanning transducer and a camera put together to generate a three dimensional interferogram
of the model surface. This is processed by the software (Metro Pro PCTM) in the computer and
transformed using frequency domain analysis to give a quantitative 3-D image.
The vertical resolution is 1 Å, independent of microscope magnification and the lateral
resolution is at best 0.3 µm. The model surfaces are characterized using the average surface
roughness, Ra, which is the average deviation of all points from a plane fit to the test surface.
The standard deviation of the profile heights Rq (rms) is also given. For a gaussian surface the
ratio between Ra and Rq is close to 1.3. Both Ra and Rq (rms) can be found in table 1 (see
chapter 2) for all our model surfaces. Ten measurements were made on random spots on each
model surface. The measurement area was 0.18 mm * 0.13 mm, which gives an area of
0.0234 mm2. All measurements were carried out in ambient atmosphere.
19
Sample
ReferenceSurface
PZTStack
InterferenceMicroscopeObjective
LightSource
Camera
Figure 1: The Zygo View 5010TM, a precision vertical scanning transducer and a camera put together.
20
3.2 Contact angle.
Contact angle measurements were conducted with a Fibro DAT 1100 system. This instrument
is used for fast absorption and wetting studies for contact angles above 20°. The application of
the droplet from the syringe onto the test surface was computer controlled; giving a controlled
drop volume (4 µl). The syringe used was a Teflon syringe in order to avoid any liquid
remaining onto the tip. The spreading process was recorded using a CCD camera connected to
an image analyser. The images were analysed with respect to base width and height in terms
of contact angle and drop volume. The drop volume starts to decrease due to evaporation after
10 s of spreading time. These data are discarded in the evaluation of contact angles. As this
simple method is a sensitive measure of the interfacial properties it is widely used to
characterize surfaces. The surfaces tension γ [J/m2] are only one of many properties, but also
estimates of surface roughness and chemical heterogeneity can be obtained from spreading
experiment. It is a simple and reliable method to use as a quality control of the self-assembly
monolayer formation process and general cleanliness of for instance silica and gold or other
hydrophilic materials (in general contaminations are of hydrophobic nature).
The contact angel α is related to the involved surface tensions, in this case there are three, the
solid-vapour, γSV , solid-liquid, γSL , and finally liquid-vapour, γLV , surface tension. This
relation is described by the Young’s expression (see equation 1).
cos( )αγ γ
γ=
−SV SL
LV
(1)
Whereas the Young-Dupre equation (see equation 2) relates to the adhesion energy per unit
areas of the solid (S) and liquid (L) adhering in the medium gas/liquid in our case vapour (V),
∆WSLV [J/m2] .
γ αLV SLVW( cos )1+ = ∆ (2)
Different chemical or structural components on a surface produces a heterogeneous surface.
Cassie suggested a way to calculate the contact angle of a heterogeneous multicomponent
model surface (see equation 3).
21
cos cosα χ α= Σ i i (3)
where, χi is the fraction of the i:th component on the surface and cosαi is its contact angle on
that type of surface1. In 1989, Israelachvili2, derived a revised version for the contact angle on
a heterogeneous surface (see equation 4).
( cos ) ( cos )1 12 2+ = +α χ αΣ i i (4)
Israelachvili’s assumptions are that the work of adhesion/cohesion is proportional to the
square root of the interaction forces involved, and again adding the works of
adhesion/cohesion to give the overall work of adhesion/cohesion. This equation claims to
account for the heterogeneity that is likely to occur in patches of molecular/atomic
dimensions.
22
3.3 X-ray Photoelectron Spectroscopy (XPS).
XPS is in the chemistry society often called, Electron Spectroscopy for Chemical Analysis
(ESCA). It is described in detail in references 1-5. Photons with energy hv from an X-ray
source are irradiated at the surface under study and adsorbed by the atoms. As a direct
consequence of this irradiation, emissions of electrons with lower binding energy, the
ionisation energy (Eb), than the energy of the incoming photons will occur. As a consequence
of the law of energy conservation, the emitted photoelectron obtains a kinetic energy (EK) that
is characteristic for the type of atom, the shell of the electron and its chemical environment.
The photoelectrons are separated by their kinetic energy before they reach the detector. By the
uniqueness of the kinetic energies of the photoelectrons emitted from the atoms an elemental
analysis can be conducted. In Albert Einstein’s equation (see equation 1) for the photoelectric
effect all this is described, for his work in this area Einstein got the Nobel Prize in 1921.
E hv EK b= − − φ (1)
where, φ is a correction for the spectrometer work function.
The photoelectrons, having kinetic energy up to around 1500 eV (if the AlKα electrode is
used), do not move more than a few nanometers in the solid material until they collide and
loose all or part of their kinetic energy. The average distance that photoelectrons move within
the solid material before they collide inelastically is mainly a function of the density of the
material and the kinetic energy. The inelastic mean free path λ(EK) describes this process. A
fraction of 1/e, about 37 %, of the photoelectrons move the distance λ before being scattered,
about 5 % moves as far as 3λ before they get scattered inelastically. The mean free path is
often referred to as sampling escape depth or sampling depth. It is around 0.5-2 nm for a
metal and 1.5-4 nm for oxides, these values are typical mean free paths2 for photoelectrons
having a kinetic energy of 1000 eV. Hence, XPS is truly a surface sensitive technique for
chemical analysis.
The equipment employed is a Kratos, AXIS HS X-ray photoelectron spectrometer (Kratos
Analytical, Manchester, UK). The X-ray emitted in our case comes from an MgKα (1253.6
eV) source. There are other sources available, one of them is the energetic AlKα (1486.6 eV)
source that produces more energetic photoelectrons, and due to this in some cases is more
useful.
23
3.4 The Quartz Crystal Microbalance-Dissipation (QCM-DTM).
The Quartz Crystal Microbalance (QCM) is by no means a new technique1-2. In vacuum
physics for instance it has existed for decades, providing film thickness measurements for
metal film deposition3. However, lately numerous advancements have been made in the
measurement of the frequency factor of the QCM. Among these is the new Quartz Crystal
Microbalance-Dissipation (QCM-DTM) instrument from Q-Sense, Gothenburg, Sweden (see
figure 2), which we have used for our experiments4. This new setup has two advantages. First,
an improved resolution of the frequency factor in aqueous solutions. In fact, it is hard to find
any comparable non-vacuum QCM setup5-6. Secondly, this instrument also measures the so-
called dissipation factor, which is a measure of the damping of the crystal as will be discussed
later. Hence, the QCM technique has only recently become a potentially useful tool for the
surface scientists concerned with “wet” surface chemistry. The possibility to monitor the
interfacial processes quantitatively in real time opens up new windows of opportunity. The
QCM principle is based on evaluating a change in frequency of the oscillating crystal, ∆f
[Hz], due to the change occurring on or adjacent to its electrodes. This frequency change is
most often interpreted as being due to the change in surface mass loading. In this work we
have used the QCM as a “probe”, in order to characterize physical changes at interfaces
occurring as a result of surfactant addition7-8. Such effects may arise due to adsorption or
different phase changes9-11. This chapter will describe the basic operational parameters of the
QCM, and the focus will be on its operation in liquids, even though its use is more widely
documented for the gas phase system. We note that not so many studies have explored the use
of the QCM for studying solid-liquid interfaces in surfactant solutions12-14. However, a much
more extensive literature on surfactant films deposited onto surfaces via the air-solution
interface with monolayers of insoluble surfactants is available. A discussion of this extensive
literature is beyond the scope of this thesis. The interested reader is referred to reference 12.
24
Figure 2: The Quartz Crystal Microbalance-Dissipation (QCM-DTM).
The Quartz crystal.
Most QCMs consist of a thin wafer of piezoelectric material, usually quartz, sandwiched
between a pair of thin metal electrodes (see figure 3), usually gold. Quartz is a piezoelectric
crystalline form of silicon dioxide (SiO2). For oscillating systems the α-quartz is the preferred
choice because of its thermodynamic stability at temperatures up to 846° K. The other form,
the β-quartz, is metastable at room temperature and it is not piezoelectric15. Piezoelectricity1 is
literally “pressure electricity”, the prefix piezo- being derived from the Greek word “to press”.
The direct piezoelectric effect refers to the electric polarisation of certain materials by
mechanical stress. The converse effect refers to the deformation of the same material by an
electric field. Electrostriction is a property of all dielectric materials; it means that when they
are placed in an electric field they deform. The difference between piezoelectric materials and
purely electrostriction materials is that the piezoelectric deformation is much larger than the
ordinary electrostriction deformation and the piezoelectric deformation is reversible. As an
example, a rod of quartz is cut in such a way that an applied field causes an elongation of the
rod. Reversing the direction of the field will cause the rod to contract in a piezoelectric
material, whereas in a non-piezoelectric material, whatever deformation is caused will be
independent of the direction of the field.
The piezoelectric effect being reversible gives that it is also anisotropic, which means that the
mechanical deformation and the electric field (see below) in the material depends on the
25
direction within the material. Such materials cannot have a centre of symmetry; with a centre
of symmetry the reversal of an applied field would have no effect on the materials internal
structure. Lord Kelvin gave the first explanation of the origin of the piezoelectric effect1 in
terms of molecular structure; the assumption he made is useful as a qualitative and heuristic
guide to understand piezoelectricity.
The electrode.
The electric field is in almost all quartz crystals applied via electrodes deposited at the quartz
surface in a key hole pattern, as shown in figure 3. In general gold electrodes give a
considerably more chemically stable surface compared to other electrode materials16 such as
silver (Ag) and aluminium (Al), which both tend to oxidize in aqueous solutions. Although it
has been suggested that the gold electrodes of the QCM may also be subject to minor
oxidation. This is certainly the case when they are treated with UV/ozone (AuO3, is the
product from the UV/ozone treatment17). For the gold and silver electrodes a thin adherent
layer (2.5-5 nm) made of chromium (Cr) or titanium (Ti) are used to improve the adhesion of
the electrodes to the quartz crystal. The disadvantage of this thin adherent layer is the
increased stress in the electrodes, which can influence the output frequency.
Figure 3: The Quartz crystal, with its gold electrode in a characteristic “keyhole” pattern.
26
Frequency.
A quartz crystal used in the QCM is normally cut at an angle θ ≈ 35° from the ZX-plane, this
is known as an AT-cut18, this cut angle makes the quartz crystal less sensitive to temperature
drifts as compared to the large temperature drifts in the original X-cut quartz crystal (quartz
crystals cut normal to the x-axis) where the temperature drifts are as large as 30 ppm / °K. In
the AT-cut case the temperature drift could be as small as 2 ppm. / °K. Another improvement
from the original X-cut is that when applying an electrical field across an AT-cut crystal, a
shear strain will be induced instead of the induced strain in the thickness direction in the X-
cut case.
Consequently, an alternating electric field onto an AT-cut quartz crystal will induce shear
waves. Of all the vibrational modes that may exist in a quartz crystal, only those that can be
driven by an alternating electrical field are relevant in the context of this thesis. Mechanical
resonance begins when the thickness of the quartz crystal contains an integral number n of
half wavelengths of the extensional wave or longitudinal waves. The quartz crystal’s surfaces
will be the anti-nodes of vibration from a standing wave within the plate. When n is even the
vibrational modes of the two surfaces are in phase (destructive), and in anti-phase
(constructive) when n is odd (n=1 being the fundamental mode, n = 3 is called the first
overtone, and so forth). The resonance frequency condition is (equation 1):
fnv
tq
=2
(1)
Where v is the velocity of the extensional waves, (v/f) is the wavelength, tq is the thickness of
the quartz crystal and n is an odd integer (1,3,5,…).
Sauerbrey1 9 was the first to show that any mass, ∆ m , deposited on one or both of the
electrodes of a QCM crystal, induces a shift in the frequency, ∆ f, that is proportional to the
added mass. If the mass is deposited evenly over the electrode(s), and ∆ f is much smaller
than f, then the frequency shift versus mass relationship is:
∆∆ ∆ ∆
mt f
nf
f
nf
C f
nC
t
fq q q q q q= − = − = − ⇒ =
ρ ρ ν ρ
0 02
02(2)
27
Where ρ q and ν q are the specific density and the shear-wave velocity in quartz, respectively,
tq is the thickness of the quartz crystal, f0 the fundamental resonant frequency and n is the
shear wave number. With ρ q = 2648 kg/m3, ν q = 3340 m/s, tq = 0.33 mm, and f0 = 5 MHz, C
is 17.7 ng cm-2 Hz-1. For the relation to be valid, Sauerbrey assumed that the added mass
should be much smaller than the mass of the quartz crystal, and it should be rigidly attached to
the electrode(s), with no slip or inelastic deformation in the added mass due to the oscillatory
motion. Pulker later confirmed equation 2 by experimental data up to mass loadings (madsorbed /
mcrystal ) of approximately 2 %. There are various models or converting the frequency shift to
mass loadings up to approximately 5 %, and they all behave similarly21. Another property,
probably the most important to have under control, is that the surface area should be smooth.
The QCM surface area is approximately the same as the projected geometrical area for low Ra
values, and this roughness effect will be discussed later.
The use of QCM in liquid media.
In 1980, Nomura showed that a quartz crystal could be completely immersed in a liquid and
still be excited to stable oscillations22, after this theories had to be worked out. In 1985
Kanazawa and Gordon published a theory23 on the QCM behaviour in the liquid phase. They
were totally unaware of the theory that Stockbridge24 had published already in 1966. This
paper was concerned with the gas pressure effect on the QCM oscillations, and it turned out to
be exactly the same as the Kanazawa and Gordon theory. The relationship derived describes
the change in oscillation frequency of the quartz crystal in contact with a fluid in terms of
material parameters of the fluid and the quartz. This relationship is shown below (equation 3).
∆ff
t nq q
f f= − 0
2 ρ πρ η (3)
Where ρ f and ηf are the specific density and the absolute viscosity of the film, respectively, tq
is the thickness of the quartz crystal, f0 the fundamental resonant frequency of the dry crystal,
ρq is the specific density of quartz and n is the shear wave number. With ρ q = 2648 kg/m3, tq
= 0.33 mm, and f0 = 5 MHz. This relation is obtained from a simple physical model, which
28
couples the shear wave in the quartz crystal, to a damped shear wave in the fluid. The shear
wave extension or, as it is more commonly called, the decay length, is given by equation 4.
δπη
ρ=
4 f
n ff(4)
where, δ is the decay length of the shear wave, ηf is the absolute viscosity of the film, fn the
resonant frequency of the dry crystal in mode n and ρq is the specific density of quartz. The
decay length is the distance into the liquid where the amplitude of the shear wave has fallen
by a factor of e, and for a 5 MHz quartz crystal oscillating in water this decay length is
approximately 250 nm at 20° C.
Roughness properties.
Martin, Frye and Wessendorf examined the frequency response of smooth (low surface
roughness, Ra < 10 nm) and textured surfaces (high surface roughness, Ra > 100 nm) on
quartz crystals in liquids in 199425. Smooth quartz crystals, which viscously entrain a layer of
contacting liquid, exhibited a response that depends on the square root of the product of liquid
density and viscosity. Textured-surface quartz crystals, which also trap liquid in surface
crevices, pores, etc., exhibit an additional response that depends linearly on liquid density
alone. The resulting modification to the Stockbridge and Kanazawa equation 4, is shown in
equation 5.
∆ ∆ ∆f f ff
t n
f
ttv t
q q
f fn
q qf f= + = − −0
2 ρ πρ η
ρρ (5)
Where ∆fv is the induced frequency shift due to the liquids viscosity and density over a
uniform crystal. ∆ft is the induced frequency shift due to trapped liquid with an average
thickness of tf , ρ f and ηf are the specific density and the absolute viscosity of the film,
respectively, tq is the thickness of the quartz crystal, fn the resonant frequency of the dry
crystal, ρq is the specific density of quartz and n is the shear wave number. The liquid
29
entrained by the oscillating smooth surface is described as viscously coupled. This liquid does
not move synchronously with the surface, but undergoes a progressive phase lag with
increasingly distance from the surface. The textured-surface also traps a quantity of fluid in
excess of that viscously entrained by a smooth surface. The perpendicular character of the
texture-surface constrains this trapped liquid to move synchronously with the surface, rather
than undergoing a progressive phase lag. This trapped liquid can be viewed as an added mass,
contributing to an areal mass density, ρtf, where ρ is the absolute density of the liquid and tf is
the effective thickness of the perpendicular features of the surface.
The frequency shift measured in an adsorption experiment is relative to the frequency of the
crystal immersed in water. Under the conditions we have used the instrument, i.e. relatively
low solute concentrations, no measurable effects due to changes in bulk viscosity or density is
expected. Hence, the measured frequency shift in our experiments is due to changes occurring
close to the surface. Most importantly adsorption including bound water.
Martins addition to Gordon’s and Kanazawa’s equation is valid under the condition that the
effective thickness h, of trapped liquid is small compared to the liquid decay length δ. In such
a case the relative response due to liquid trapping is small, compared to the frequency shift
due to viscous entrainment and may be neglected. This defines a criterion for hydrodynamic
smoothness29: h << δ. If h is comparable to or larger than δ, then a significant additional
response arises due to liquid trapping. Martins definition for a smooth surface are that the
roughness should be less than 10 nm, but he was working with frequency shifts (∆f ) in the
kHz region. We are working within a lot more sensitive regime, the frequency shifts in our
experiments are in the Hz region, where this effect is significant, this is in addition to the
viscous entrainment.
Longitudinal waves.
The generation of longitudinal waves in liquids has been largely ignored, except in a few
reports30-33. The occurrence of the longitudinal wave component is usually demonstrated by a
movable plate, parallel to the quartz crystal, and starting from a remote distance (much larger
than the decay of the shear wave). The presence of longitudinal waves can be observed so far
away as centimetres34-35. There presence results in a periodicity of the resonant frequency,
inductance, capacitance, and resistance as the distance between the quartz crystal and the
movable plate is varied36-38. The wavelength, of the standing longitudinal waves that are
30
observed agree with that expected for the resonant frequency of the quartz crystal and the
properties of the fluid medium, and the periodicity of the quantities mentioned above is
approximately half the longitudinal wavelength. For a 5 MHz quartz crystal in a water
medium at 20° C the periodicity is approximately 150 µm.
Pc
f= =
λ2 2 0
(6)
where, P is the periodicity [m] , λ is the wave length [m], c is the phase velocity of the waves
in water, c = 1465 m/s at 20° C and f0 is the fundamental frequency [Hz].
Various measurements of the standing wave frequency and amplitude have revealed effects of
crystal contour, liquid properties, interface reflection coefficients, and the radial dependence
on the standing wave amplitude. Clearly it is important to design the experiments so that
contributions from such longitudinal waves are avoided, mainly taking the phase velocity of
the waves in the medium into consideration.
Amplitude of vibration.
From classical driven oscillator theory, it is clear that the values of the shear wave amplitude
depends on both the drive voltage applied to the quartz crystal and the quality factor, Q (the
inverse of the dissipation factor, D), of the system. This has to be remembered than
comparing data from different authors and theoretical calculations. In view of recent
theoretical work by Kanazawa39, it is now possible to compare experimental measurements of
the vibration amplitude itself to detailed calculations for quartz. Kanazawa calculates an
amplitude of 133 nm using a peak drive voltage of 1.0 V, and a Q of 100 000 (D = 10*10-6).
Just as with classical oscillator theory, the amplitude of vibration is expected to be
proportional to the drive voltage and the quality factor also in Kanazawa’s model39:
A C Q VC V
Dav av dav d= =* *
*(7)
31
where Aav is the average amplitude of vibration (average = 1/2 maximum), [m], Cav is found to
be 1.3 pm/V, Q is the quality factor and Vd is the drive voltage, [V].
This value for Cav is rather close to the piezoelectric strain coefficient for an AT-cut quartz
crystal, 3.1 pm/V, which Martin and Hager40 concluded was the same as the Cav. Borovsky,
Mason and Krim41 on the other hand are suggesting an Cav = 0.7 pm/V, so there are still open
questions about the amplitude of vibrations in quartz crystals. Both theoretical and
experimental results put the amplitude of vibration (Cav) for a 5 MHz quartz crystal, with a Q
factor of 100 000, and a drive voltage of 1.0 V, in air in the interval 40-200 nm. Since the Q-
value for the same system in water are around 3 000, this would shrink the amplitude of
vibration to the interval 1-6 nm for the aqueous system. In our case with a drive voltage of 0.7
V, and a dissipation factor equal to 20*10-6 in air and 310*10-6 in water, the resulting average
amplitude is 45 nm in air and 3 nm in water, using Cav.
The Dissipation factor.
There is, beside the resonant frequency, another important parameter that characterises the
oscillatory system, namely the dissipation factor, D or the inverse of Q, the quality factor26, Q
= 1/D. In 1966, Spencer and Smith27 studied the amplitude of an oscillating quartz crystal and
its decay. The amplitude was decaying as an exponential sinusoid.
A t A e ft ct( ) sin( )/= + +−0 2τ π ϕ (8)
where τ is the decay time that depends on f, ϕ is the phase angle and the constant, c, is the dc
offset. The total dissipation factor for the fundamental frequency Dtotal is related to the decay
time, τ , according to:
Dftotal =1
0π τ(9)
where f0 is the fundamental frequency. The dissipation factor is proportional to the power
dissipation in the oscillatory system:
DE
Edissipated
stored
=2π
(10)
32
Where Edissipated is the energy dissipated during one period of oscillation, and Estored is the
energy stored in the oscillating system. Consequently, D accounts for all mechanisms, that
dissipate energy in the oscillatory system. Which can be summarized as Dtotal:
D Dtotal ii
= ∑ (11)
where Di is the dissipation factor of the i:th mechanism. For instance, an adsorbed film
dissipates energy through its coupling with the solvent10-11. This is probably the largest effect
in thin films, and it will be discussed later. Internal friction in the quartz crystal is another
large factor and so are the losses due to mounting. Further, if the film is not very thin and
viscous, energy is dissipated due to the oscillatory motion induced in the film28. Hence, if the
dissipation factor can be measured correctly it may be possible to obtain additional
information about the visco-elastic properties of adsorbed layers. Other studies indicate that
phase changes within the film can be related to changes in the dissipation factor7-8, which
probably can be related back to the adsorbed films coupling to the solvent liquid. Stockbridge
not only related the frequency shift but also the damping of the oscillations for the quartz
crystal in liquids to liquid properties24:
∆Df
nl l
q q
= 2 02 2
ν ρπυ ρ
(12)
where, ∆D is the total dissipation shift, fn is the frequency, n is the shear wave number
(1,3,5,…), νq is the wave velocity in quartz, ν f is the wave velocity in the film, ρq is the
density of quartz, and finally, ρf is the density of the film. νq = 3340 m/s, ρq = 2650kg/m3, and
for water at 20° C, vl = 1465 m/s.
Martin, Frye and Wessendorf25 used smooth quartz crystals that viscously entrain a layer of
contacting liquid. They observed a response that depends on the square root of the product of
liquid density and viscosity, as predicted by the Stockbridge equation (equation 5). Since the
added liquid is fully coupled to the surface texture, there is no damping of the quartz crystal
oscillations due to the presence of trapped liquid.
33
The change in dissipation reported in our experiments is the difference between the
dissipation rate after and before addition of solute. Again, under the dilute solution conditions
studied the effects due to changes in bulk viscosity and density can be ignored. Hence, the
values reported are due to changes in dissipation occurring close to the surface. This includes
energy dissipation within the adsorbed layer and differences in coupling between the surface
with the adsorbed layer and the solution compared to between the bare surface and the
solution.
34
4. Surfactants
4.1 Lipids.
In living systems large molecules, such as proteins, polysaccharides and complex lipids, build
up the structure of the cells and tissues. The word lipid covers an astounding variety of
compounds. One of the most widely used classifications is that of Bloor and later by Deuel1:
i) Simple lipids, which consists of neutral fats (glycerol esters of largely long-chain fatty
acids) and waxes (solid esters of long-chain monohydric alcohols) and ii) Compound lipids
which consists of Phospholipids (lipids containing a phosphate residue), Cerebrosides,
Gangliosides (lipids containing a carbohydrate residue) and Sulphatides (lipids containing a
sulphate residue). The compound lipids are esters of fatty acids with alcohols, which contain
also an additional group, see above.
In this thesis work we will limit ourselves to study the phospholipids, phosphatidylcholine
(PC) and phosphatidylethanolamine (PE) and an emulsion stabilized mainly by such lipids.
Phospholipids are zwitter-ionic. The cationic group in PCs consists of a quaternary
ammonium group whereas in PEs it consists of a primary amine. For this reason PEs become
anionic at high pH, whereas they are cationic at very low pH. In the ionic state, these lipids
with monovalent counterions behave in a similar way to ionic phospholipids. In general, the
acyl chains of the phospholipids are comparably long and the monomer solubility is very low,
approximately 10-11 M, in water for a typical hydrocarbon chain length ∼ C16. As a
consequence of this the phospholipids will self-assemble at low concentrations. Vesicles, or
as they also are called in the phospholipid case, liposomes, are composed of lipid bilayers
enclosing a water core. Liposomes are in general not thermodynamically stable, i.e. they are
not an equilibrium structure. These structures are used in so diverse areas ranging from
mimicking cell membranes to delivering of drugs in the human body.
Phosphatidylcholine (PC).
The last recommendation of a nomenclature for PC is that one from IUPAC-IUB Commission
on Biochemical Nomenclature and that is 1,2-diacyl-sn-glycero-3-phosphocholine; we will
35
hereafter only call it PC. The molecular structure consists of a zwitter-ionic head-group and
two fatty acyl chains see below.
RICOOCH2 - RIICOOCH - CH2O –POO- –OCH2CH2N
+(CH3)3
Where, RI- and RII- are the fatty acyl substituents. An extensive literature exists about their
aqueous interaction and phase properties 3. For a typical PC the lamellar Lα phase dominates
the phase diagram and it is this phase that is in equilibrium with excess water. At low water
content, and as the temperature is increased it is possible to obtain the transition Lα ⇒ cubic
⇒ H11, where H11 is the reversed hexagonal phase. The driving force for this transition is the
tendency for increased chain divergence when the thermal mobility increases. When the Lα-
phase is cooled, it forms a gel-phase. The transition temperature from the Lα-phase to the gel-
phase on cooling decreases from 41° C to 23° C when the saturated chain length is reduces
from C16 to C142.
Phosphatidylethanolamines (PE).
The last recommendation of a nomenclature for PE is that one from IUPAC-IUB Commission
on Biochemical Nomenclature and that is 3-sn-phosphatidylethanolamine; we will hereafter
only call it PE. The molecular structure consists of a zwitter-ionic head-group and two fatty
acyl chains see below.
RICOOCH2 - RIICOOCH - CH2O –POO- –OCH2CH2N
+H3
Where, RI- and RII- are the fatty acyl substituents. An extensive literature exists about their
aqueous interaction and phase properties 3. Crystals of didodecyl –PE heated in excess of
water are transformed into the Lα-phase (+H2O) at about 40° C. At about 100° C, a cubic
phase (+H2O) is formed and above 120° C, the reversed hexagonal phase H11 (+H2O) is
obtained3. There are also reports of different PEs with unsaturated chains that exhibit cubic
phases and H11-phases4.
Emulsions.
An emulsion, i.e. a liquid dispersed in another liquid, is not a thermodynamically stable
system. Two processes that can destabilize oil-in-water emulsions are of interest to us. The
36
first being flocculation of emulsion droplets, which is an aggregation phenomena with the
droplet interface remaining intact and a thin liquid film remaining in between the aggregating
droplets. This normally leads to an increased viscosity and sometimes a gelation. The second
process is the coalescence of flocculated droplets. After coalescence the emulsion droplets
lose their identity. With time the density difference between the coalesced drops and the
aqueous phase leads to accumulation of an oil rich layer at the top of the container. This
phenomenon is called creaming5.
37
4.2 Cationic surfactants .
The polar moiety in a cationic surfactant has a positive charge, thus they adsorb strongly onto
most solid surfaces (which are usually negatively charged), and can impart special
characteristics to the substrate. Some examples are, softeners, anti-static agents, corrosion
inhibitors, hair conditioners, lubricants and flotation agents1-4. The adsorption of surfactant
ions from aqueous solutions onto hydrophilic negatively charged surfaces is based on two
main interactions: electrostatic and hydrophobic5-7. The electrostatic interaction is involved in
the first step of a two-step adsorption mechanism and its contribution to adsorption depends
largely on the charge of the surfactant ions, surface charge density of the adsorbent,
electrolyte concentration and pH. Hydrophobic interactions are involved in the second step of
the two-step adsorption mechanism where additional surfactants are associated with
electrostatically anchored surfactants. This interaction is mainly influenced by the surfactant
structure and particularly the size of the hydrophobic part8-11. The hydrophobic interaction acts
between the non-polar parts of the surfactant ions (hydrocarbon chain) and it is due to release
of water molecules forming a dynamic cage around the non-polar moiety10. van der Waals
forces also contribute to the association, but this contribution is small compared to the
hydrophobic interaction.
For a deeper discussion on hydrophobic interactions see for instance the book by Tanford and
refrences4, 10, 11. The corresponding attractive forces cause aggregation into micelles in the bulk
phase. At the solid-liquid interface they are responsible for the surfactant aggregation into
surface aggregates. For a further detailed discussion of the structure of the surfactant
adsorption layer, two different mechanisms must be considered: i) Surfactant ions interacting
with hydrophobic surface sites on non-polar or polar/non-polar surfaces11. ii) Surfactant ions
interacting with each other and with those primarily adsorbed by hydrophobic and
electrostatic forces12-13. Obviously on non-polar surfaces with hydrophobic groups (e.g.
methylated silica14), polar/non-polar surfaces containing both hydrophilic (charged) groups
and hydrophobic entities (e.g. low density of chemisorbed alkylsilanes15), the direct
attachment of hydrophobic entities to the surface can be expected. However, on fully hydrated
adsorbents, no significant direct interaction between the surface and the hydrophobic part of
the surfactant is expected.
38
4.3 Non-ionic surfactants .
The hydrophilic (polar) part of a non-ionic surfactant is usually a poly(ethylene oxide) chain
even though sugar-based surfactants also are becoming common. The hydrophobic (non-
polar) part is, just as for cationics, anionics and zwitter-ionics, most often a hydrocarbon
chain. The interfacial behaviour of ethylene oxide based non-ionic surfactants CnEOm, is
strongly affected by the size of the polar and non-polar parts of the molecules1-6. The adsorbed
amount is found to increase with an increasing number of methylene groups in the
hydrocarbon chain (m), whereas a decrease in the adsorbed amount with increasing number of
ethylene oxide groups in the hydrophilic chain (n) is commonly observed. Polyethylene oxide
based surfactants display a variety of different phases in aqueous solution, depending on the
surfactant structure, temperature, and concentration7-10. Observed phases include micellar
solutions, lamellar, hexagonal, and cubic phases. Depending on whether the polar or the non-
polar part of the surfactant interacts most favourably with the surface, different structures of
the adsorbed surfactant layer is formed. Generally, a monolayer structure is found on
hydrophobic surfaces, and at hydrophilic surfaces different surface aggregate forms such as
micells or bilayers11-14.
39
5. Results and Discussion
5.1 Adsorption of Emulsions studied with the QCM-DTM.
Adsorption and spreading of emulsions on solid surfaces are of importance in a range of
applications, e.g. drug delivery, lubrication, and during coating of the road surface with
asphalt emulsions, just to mention a few. Hence, understanding interfacial properties of
emulsions and solutions are relevant to a large number of technologies. There are, however,
not very many suitable experimental methods that readily can be applied to this area of
research, particularly for concentrated emulsions at solid surfaces. Hence, optical methods
such as ellipsometry and interferometric surface force measurements are difficult to apply due
to the large scattering of light that characterise these concentrated systems. However, despite
this some rather recent progress has been made. Ellipsometric1 studies of adsorption of dilute
emulsions have been carried out and surface force techniques have successfully been applied2.
In paper I we showed that the QCM-DTM technique, allowing the simultaneous measurement
of changes in resonance frequency and energy dissipation rate, is very suitable for studying
emulsion adsorption and spreading. For the study we used a model oil-in-water emulsion
consisting of purified soybean oil (20 wt.%) dispersed in water. The emulsifier (1.2 wt.%)
used was fractionated egg phosphatides with the major components being
phosphatidylcholines (≈ 70%) and phosphatidylethanolamines. In figure 1 the adsorption
behaviour of the 20 wt.% oil-in-water emulsion is illustrated. The initial adsorption is very
rapid and results in a significant lowering of the resonance frequency and an increase in the
dissipation factor. We note that the initial change in resonance frequency is smaller than for
the liposomes that also were studied (see paper I) whereas the change in dissipation is larger.
This tells us that the adsorbed amount is smaller but that the size of the adsorbed emulsion
droplets (Dav in solution is 300 nm) is larger than that of the adsorbed vesicle (Dav in solution
70 nm). After the initial rapid adsorption a slow increase in resonance frequency and a
decrease in dissipation are observed until equilibrium values are obtained. We suggest that
this is due to a spreading of the emulsion droplets on the surface accompanied by some
material desorption. After 15 minutes we injected another portion of the same emulsion
solution in the measuring chamber again. This exchange of the solution does not result in any
change in the bulk solution composition but nevertheless an increased adsorption, resulting in
40
a declining frequency shift and a rising change in dissipation, is observed. This indicates
adsorption controlled by hydrodynamic factors. Hence, the injection of the solution provides
the necessary kinetic energy to the emulsion droplets to let them overcome the energy barrier
for further adsorption.
-100
-80
-60
-40
-20
0
20
-20
0
20
40
60
80
100
0 50 100 150 200
f5 (
Hz)
D5 (10-6)
Time (min)
Figure 1: Emulsion adsorption at a gold surface, injections of solution at t = 10, 25, 55, 85 and 125 min and
rinsing with water at 160 min. The frequency shift is represented by () , and the change in dissipation factor
by (- -).
Again, after the initial adsorption a slow spreading and desorption follow. It is worth noting
that the spreading occurs slower at this stage compared to during the preceding stage, and the
probable reason for this being the increased packing density of the emulsion droplets that
makes the spreading process more difficult. Repeated exchanges of the emulsion solution (at t
= 55, 85, and 125 min) give similar results as described above. It is, however, worth noting
that the spreading process becomes slower for each successive exchange of the solution. The
final frequency shift is about –90 Hz, which correspond to an adsorbed amount of
approximately 16 mg/m2 as calculated according to the Sauerbrey relation. Considering the
high dissipation factor it seems likely that the layer on the gold surface consists of deformed
emulsion droplets. We ended the experiment by replacing the emulsion with pure water (at t
=160 min). This resulted in a rapid desorption and the dissipation returned to close to zero
whereas the final frequency shift was about –10 Hz. The latter quantity corresponds to an
adsorbed mass of 1.8 mg/m2, or an area per molecule of 70 Å2. This means that a monolayer
remains on the surface.
41
5.2 Polymer adsorption on phospholipid coated surfaces.
In paper II, the adsorption of the triblock copolymer F127, poly(ethylene oxide)-
poly(propylene oxide)-poly(ethylene oxide), PEO98PPO67EO98, onto immobilized small
unilamellar vesicles (SUVs) of egg phosphatidylcholine (EPC) was studied by the QCM-DTM.
In figure 2, the adsorption of SUVs onto the hydrophilic gold surface is illustrated. The
liposome solution was introduced into the measuring chamber at t = 10 min. Immediately
after the introduction, the resonance frequency dropped and the dissipation increased. This
shows that the liposomes adsorb to the gold surface and the large changes in frequency and
dissipation indicate that they adsorb as intact liposomes. After the initial rapid adsorption we
can see a slower increase in adsorption. The liposome solution was exchanged for a identical
liposome solution twice at t = 70 min and t = 100 min. This resulted in transient peaks in the
frequency curve due to the temperature (the reservoir is not perfect) and pressure effects due
to the increased volume resting on the oscillating quartz crystal. However, there was no
significant change of the frequency after the solution exchanges were finished, and
temperature and pressure were restored. The slow adsorption of the liposomes at longer times
indicates that the surface was almost completely saturated with liposomes already after the
first introduction of liposomes, probably due to hydrodynamic transport to the surface. At t =
130 min, the liposome solution was exchanged for pure HEPES buffer. This was done to
remove eventual loosely adsorbed liposomes from the gold surface. Evidently there was no or
little desorption of SUVs after rinsing and thus the interaction between the liposomes and the
gold surface is sufficiently strong to make the adsorption irreversible over the experimental
timescale. Further rinsing with pure buffer, at t = 160 min and t = 190 min, did not cause any
desorption of liposomes. Following the above qualitative description of the SUV adsorption
process we shall make some quantitative considerations. The final change in resonance
frequency, ∆fSUV, can be converted into the adsorbed liposome (SUV) mass, mSUV, by means
of the Sauerbrey relation. The mean value of ∆fSUV, obtained from 8 runs, was 343 (± 63) Hz.
Using the Sauerbrey relation this corresponds to an adsorbed mass of 20.35 (± 3.7) mg/m2.
Theoretically, we can estimate the adsorbed mass of the liposomes, including the mass of the
“entrapped” buffer; by assuming that the liposomes have a radius of 15 nm (Dav = 15 ± 5 nm)
and that they form a monolayer of close-packed spheres at the surface. With these
assumptions we get approximately 18 mg / m2 of adsorbed liposomes. In this calculation we
have not accounted for the buffer entrapped between the liposomes and the gold surface
42
which may also oscillate with the crystal. Nevertheless, the estimate is close to the
experimental value and we conclude that a monolayer of SUVs is formed at the gold surface
in accordance with previously published results1.
After the last rinsing of the SUV monolayer, the polymer solution was introduced at t = 220
min. The frequency dropped immediately after the introduction of the polymer solution
indicating a rapid adsorption of the polymer on the SUV monolayer. Besides the frequency
drop due to the polymer adsorption there was also a transient peak due to temperature and
pressure effects, as discussed above. Furthermore, the dissipation increased as the polymers
adsorbed. The polymer solution was exchanged for an identical polymer solution twice at t =
250 min and t = 280 min. The final frequency drop, ∆fF127, was determined after the third
addition of polymer solution. Using the Sauerbrey relation, this value was converted into the
adsorbed polymer mass, mF127. To investigate the desorption process, we exchanged the
polymer solution for a pure buffer solution at t = 310 min. As can be seen in Figure 2, the
frequency increased and the dissipation decreased, indicating desorption of F127.
Interestingly, the kinetics of the adsorption and desorption processes seem to be about the
same. Evidently, some of the adsorbed polymers are rapidly desorbed during rinsing and the
desorption of the F127 polymers causes little or none desorption of SUVs.
-600
-500
-400
-300
-200
-100
0
100
-5
0
5
10
15
20
25
30
0 50 100 150 200 250 300 350
f15
(Hz)
D15 (10-6)
Time (min)
Figure 2: The adsorption of liposomes (gold surface) after injecting the liposome solution at t = 10, 70 and 100
min, followed by rinsing with HEPES at t = 130, 160 and 190 min. The injection of F127 occurred at t = 220,
250 and 280 min, followed by rinsing with HEPES at t = 310 min. The frequency shift is represented by () ,
and the change in dissipation factor by ( - - -).
43
It is important to emphasize that there was no measurable adsorption of F127 onto the bare
gold surface. Thus, it is clear that the adsorption of F127 occurs onto the preadsorbed SUVs,
and that the adsorbed amount of F127, follows a Freundlich isotherm at low concentrations,
illustrated in figure 3. The determined polymer-binding isotherm was used to correlate the
adsorbed amount of polymer with the polymer-induced leakage of carboxy fluorescein (CF)
from the SUVs in bulk solution. It was found that the membrane permeability was increased
several fold already at low surface coverage and that the maximum magnitude of the CF
release rate was obtained at, or close to, the F127 concentration needed to reach the maximum
adsorbed amount of polymer.
0
0.05
0.1
0.15
0.2
0 0.2 0.4 0.6 0.8 1 1.2 1.4
mF1
27 /
mS
UV
CF127
/ mg mL-1
Figure 3: Adsorption isotherm for the Pluronic F127. The adsorbed amount was determined as the adsorbed
mass of F127, mF127, divided by the adsorbed mass of SUVs, mSUV. The solid line going through the data points is
only drawn to guide the eye. The horizontal solid line indicates the plateau adsorption value, Γp. The dashed line
is the calculated Freundlich isotherm.
In a follow up experiment (not yet published) we determined the adsorption of the triblock
copolymer F127, poly(ethylene oxide)-poly(propylene oxide)-poly(ethylene oxide),
PEO98PPO67PEO98, onto a bilayer of egg phosphatidylcholine (EPC). With the QCM-DTM
technique we first showed that the same vesicles used above break down to bilayers upon
contact with the silica surface, this is illustrated in figure 4. The concentration of F127 used
was the same as in the SUVs experiment (corresponding to the concentration needed to reach
the plateau value of adsorbed amount) and the normalized adsorbed amount was determined
44
to be the same as that found for adsorption of F127 on SUVs. Furthermore, the desorption of
the triblock copolymers from the bilayer surface was followed after rinsing the SUV
monolayer with pure buffer. It was found that the desorption process displayed rapid kinetics,
again indicating a weak interaction between the polymers and the lipid membrane. The
desorption of F127 was more complete from the phospholipid bilayer coated surfaces, as
compared to from SUV coated surfaces.
-200
-150
-100
-50
0
50
-1
0
1
2
3
4
5
0 20 40 60 80 100 120 140
f15
(Hz)
D15 (10-6)
Time (min)
Figure 4: Time evolution of the frequency ( ) and dissipation shift (- - ) using silica surfaces. In this
experiment SUVs of EPC was introduced in the measuring chamber at t = 10 and 30 min, followed by rinsing
with HEPES at t = 50 and 70 min. The block copolymer F127 was injected at t = 90 and 110 min. Finally the
polymer solution was replaced by HEPES at t = 130 min.
We suggest that this is due to the different geometries of the vesicle and bilayer coated
surface i.e. some polymers are trapped between the vesicles and the surface. The change in
dissipation factor with polymer adsorption (∆D) was larger when the F127 adsorption
occurred on the vesicle coated surface as compared to on the bilayer. However, when
normalizing the change in ∆D with the change in ∆f occurring upon adsorption of F127, we
reach the same value for the bilayer case and the SUV case. This will be discussed further
below.
45
We continued our experiments with different tri-block copolymers both at surfaces coated
with SUV and surfaces coated with a bilayer. Our data, that are not yet published,
demonstrate that it is the length of the hydrophilic tail that is responsible for the change in the
dissipation factor occurring upon polymer adsorption. For F88 (PEO103PPO39PEO103), F87
(PEO61PPO40PEO61) and P85 (PEO26PPO40PEO26), the dissipation factor is increasing with
increased PEO chain length, whereas the adsorbed amount is increasing with increasing chain
length of the hydrophobic group (PPO). Further, with the same hydrophilic chain length we
obtain almost similar changes in dissipation value, independent of the hydrophobic chain
length (see table 1). F127 (PEO98P P O50PEO98) ⇒ ∆D = 1.6 ± 0.3*10-6 and F88
(PEO103PO39PEO103) ⇒ ∆D = 1.75 ± 0.3*10-6. Though, the adsorbed amount was considerably
different the observed ∆f (F127) corresponds to 0.95 mg/m2 and ∆f (F88) corresponds to 0.65
mg/m2.
Triblock copolymer ∆∆∆∆f [[[[Hz]]]] ∆∆∆∆D [[[[10-6]]]] ∆∆∆∆D/∆∆∆∆f [[[[10-6/Hz]]]]
F127 (PEO98PPO67EO98,) 16 1.6 0.1
F88 (PEO103PPO39PEO103) 11 1.75 0.16
F87 (PEO61PPO40PEO61) 17 0.8 0.05
P85 (PEO26PPO40PEO26) 23 0.5 0.02
Table 1: A comparison between different triblock copolymers adsorbed onto a phospholipid bilayer.
Interpretation of the dissipation factor.
The energy dissipation factor describe the rate with which energy is dissipated in the system
and the instrument can only measure the total energy dissipation rate. As discussed in the
methods section, many different molecular mechanisms contribute to the measured effect. The
much larger energy dissipation rate observed for adsorbed vesicles compared to adsorbed
bilayers highlights that the coupling to the liquid outside the surface is increased when the
thickness, and roughness, of the surface coating increases. More interesting is that adsorption
of the same polymer on the vesicle coated and the bilayer coated surface gives rise to the
same change in energy dissipation per unit adsorbed polymer mass. This is a strong indication
that the contributions to the energy dissipation from the underlying adsorbed layer
(phospholipid vesicles or bilayers) and from the polymer layer adsorbed on top of it are
additive. A consequence of this result is that adsorption of the polymers to the vesicles does
46
not affect the vesicles visco-elastic properties as determined at a frequency of 15 MHz. This is
remarkable since the polymer adsorption clearly induces structural changes in the vesicles as
probed by fluorescein leakage.
We further note that it is the PPO block of the polymer that adsorbs to the phospholipid
coated surface, but the length of this block has only a limited effect of the energy dissipation.
On the other, hand the hydrophilic block (PEO) that extends into the solution provides an
efficient pathway to energy dissipation due to its strong interaction with the surrounding
water.
The effect of the packing of surfactant layers, above cmc, for a series of gemini surfactants
were investigated in paper IV. We noted that even though the packing was significantly
different for the different surfactants (see table 3) the change in energy dissipation was very
similar ∆D = 0.5 ± 0.1*10-6. This lack of correlation between layer structure and dissipation is
interesting and indicates that the resolution of the dissipation factor from QCM-DT M
measurements is not sufficient to describe the viscoelastic character of the thin surfactant
film. The reason being that for thin surfactant films other mechanisms dominate the energy
dissipation.
47
5.3 Adsorption of surfactants studied with the QCM-DTM.
The study of a range of gemini surfactants as described in paper IV, also gave interesting and
unexpected results. In this study, a series of alkanediyl α ω, -bis(alkyldimethylammonium
bromide) or (CnH2n+1)[N+(CH3)2](CH2)s[N
+(CH3)2] (CmH2m+1) 2Br-, (m=n), was used, which will
be referred to as m-s-m in the following. A series consisting of four different spacers, namely
12-3-12, 12-6-12, 12-8-12 and 12-12-12 was investigated.
The study was concerned with adsorption and frictional properties of gemini surfactants at
hydrophilic gold surfaces, using the QCM-DTM, and the Atomic Force Microscopy (AFM)
technique. The molecular packing of a series of gemini surfactants was determined from
QCM-DTM measurements (see Table 1) and the frictional behaviour of the surfactant films
was characterized by employing the AFM.
The results show that by changing the length of the spacer group from 3 to 12 a systematic
change in the molecular packing at the surface is obtained (see table 2). It was found that an
increase in the molecular packing resulted in a lower frictional force between the surfactant
coated surfaces. The frictional results can be viewed upon either as controlled by the rigidity
of the surfactant layer or as a result of defects, holes, in the lubricating film.
gemini Area [[[[Å2]]]]
12-3-12 106 ± 2
12-6-12 131 ± 8
12-8-12 140 ± 4
12-12-12 171 ± 6
Table 2: Area per surfactant versus spacer length for a series of gemini surfactants at a gold surface, assuming a
bilayer model.
In figure 5, the friction versus load measurements between a tungsten probe/sphere and a gold
surface in the presence of gemini surfactant solutions (2 mM) of varying spacer length is
shown. For applied forces over 20 nN all the gemini surfactant coated surfaces show an
almost linear dependence on load. The difference between the surfactants is visible and a
trend is observed between the friction force and spacer length. The observation is that the
48
friction force at any given load increases with spacer length. Converting the frictional data (in
the 10-70 nN regime) to friction coefficient gives that the gemini with the shortest spacer, 12-
3-12, displays the lowest friction coefficient in this study. Adsorption of the surfactants with
spacer, s = 6 and s = 8 results in similar friction coefficients, and the highest friction
coefficient is observed between layers of the surfactant with the longest spacer 12-12-12.
0
5
10
15
20
0 20 40 60 80
Applied Force nN
Fri
ctio
n nN
Figure 5: Friction-load measurements between a tungsten particle and a gold surface in a 2 mM dimeric
surfactant aqueous solution. Mean values of friction force as a function of spacer length for a series of gemini
surfactants 12-s-12, where results for s = 3 (o), 6 (∆), 8 (-) and 12 (+) are displayed.
A clear trend between the spacer length and thereby the surfactant structure in relation to
friction force or friction coefficient is observed. This gives us reason to believe that it is
possible to predict the friction properties of a surfactant film from the geometry of the
surfactant. The prediction being that surfactants that pack efficiently on the surface, critical
packing parameter close to 1, should provide the lowest friction coefficient.
49
5.4 Counterion effects on sensed mass and energy dissipation.
In papers III, IV and paper VI, adsorption of cationic surfactants at solid surfaces (gold) and
chemically modified surfaces (SAMs) has been studied, and compared to adsorption of a non-
ionic surfactant (paper III). One of the purposes of these studies was to determine if
adsorption measurements using the QCM-DTM (Quartz Crystal Microbalance-Dissipation)
could reveal complementary information to that revealed by ellipsometry. Of particular
interest was to learn if there was a measurable visco-elastic effect (as obtained from the
dissipation parameter measured with the QCM-DTM) of the surfactant films. This has been
discussed in a previous section; in addition, we also address the role of the counterion for
ionic surfactant adsorption as quantified by the use of the QCM-DTM.
In paper III this is done indirectly by interpreting dissipative changes upon changing the
surface hydrophobicity as a result of associated counterions within the adsorbed layer
structure. Depending on the hydrophobicity of the surface, the Br- ion will be more or less
incorporated in the adsorbed layer structure and thereby changing the rigidity of the layer as
interpreted from changes of the dissipation values. In this study the use of a combination of
hydrophobic and hydrophillic thiols enabled a systematic change of the surface
hydrophobicity of the substrate. Thiohexadecane, HS(CH2)15CH3 and thiohexadecanol
HS(CH2)16OH were mixed in different ratios to obtain variations in surface hydrophobicity. It
has previously been shown that mixtures of molecules with equal chain length have almost
the same SAM (self-assembled monolayers) composition as the composition in the solution.
For this reason we assume that the mole fraction of the SAM forming molecules at the surface
is the same as in bulk solution. For model surfactants we used the cationic DTAB,
(dodecyltrimethylammonium bromide) and the non-ionic C12EO8, (octa-(ethylene oxide)
mono n-dodecyl ether), both having the same hydrophobic chain length, but very different
hydrophilic headgroup and molecular weight.
In table 3 and 4, the contact angle (θ), the adsorbed mass (Γ), and the dissipation factor (∆D),
for the five different mixtures of thiols and both surfactants are displayed. The surfactant
concentration was just above cmc (1.2 cmc). The mixture is defined as the mole percentage of
thiohexadecane, in a mixture of thiohexadecane and thiohexadecanol.
50
Mixture θθθθ ΓΓΓΓDTAB [[[[mg/m2]]]] ∆∆∆∆DDTAB [[[[10-6]]]]
0% 23 1.40 ± 0.06 1.17 ± 0.61
25% 40 1.04 ± 0.07 1.70 ± 0.38
50% 60 2.17 ± 0.19 2.23 ± 0.52
75% 83 2.01 ± 0.09 2.26 ± 0.25
100% 105 2.19 ± 0.08 2.85 ± 0.16
Table 3: Adsorbed mass ( ΓDTAB ), and the dissipation factor ( ∆DDTAB ) versus contact angle ( θ ), for a 1.2*cmc
solution of DTAB.
Mixture θθθθ ΓΓΓΓC12EO8 [[[[mg/m2]]]] ∆∆∆∆DC12EO8 [[[[*10-6]]]]
0% 23 1.32 ± 0.10 0.31 ± 0.08
25% 40 1.87 ± 0.06 0.17 ± 0.05
50% 60 2.19 ± 0.05 0.40 ± 0.22
75% 83 2.08 ± 0.07 0.60 ± 0.25
100% 105 2.07 ± 0.08 0.22 ± 0.17
Table 4: Adsorbed mass ( ΓC12EO8 ), and the dissipation factor ( ∆DC12EO8 ) versus contact angle ( θ ), for a
1.2*cmc solution of C12EO8.
In determining the area per molecule for DTAB, the counterion, Br-, was assumed to be
incorporated in the surfactant layer and contributing to the frequency response. With this
assumption an area per molecule of about 25 Å2 was observed for the 100% SH-C16 surface.
This estimate considers a perfect degree of counterion association, β , to the monolayer
structure. From surface force measurements it has been observed that between 80-90 % of the
bromide and chloride counterions appear to be bound to similar cationic surfactants layers on
mica surfaces. This makes us estimate the actual degree of counterion binding of the Br- to a
maximum of 90%. Converting the first overtone frequency, 15 MHz to, oscillating period
gives 7*10-8 s, which is in the same order as the residence time for a counterion at the micellar
surface. In addition, the electrostatic potential outside a flat surface is larger than outside
51
micelles, which should support an even longer residence time. It is also more difficult for ions
to diffuse from a flat surface than from a spherical micelle, simply for geometrical reasons.
All together, this makes the assumption regarding the Br- ion contributing to the adsorbed
mass sensed by the crystal reasonable. We note that even so the mass registered by the QCM-
device is significantly larger than that obtained by ellipsometry (corresponding to about 40-45
Å2 per molecule). We attribute this to the fact that the QCM-device also registers bound and
trapped water as discussed in chapter 5.5. The large dissipation change occurring as a result of
DTAB adsorption, as compared to C12EO8, adsorption is surprising since the hydrophilic part
of the C12EO8 is large and rather strongly hydrated. We suggest that the effect is due to the
large interaction between the charged adsorbed DTAB layer and the oppositely charged
double layer outside the surface. This coupling increases both the sensed mass and the
dissipation factor.
52
5.5 Bound / trapped water.
It is shown that the frequency shift as obtained from the QCM-DTM experiments results in an
overestimation of the adsorbed mass (paper III-V) if it, erroneously, is interpreted as being
due to the surfactant only This is illustrated by table 5 that shows the adsorbed mass
determined by ellipsometry and the mass sensed by the QCM-DTM device for two surfactant
systems (slightly above their cmc) on two different surfaces. We suggest that this is due to
two different effects, viz., water that is coupled to the adsorbed layer due to hydration of the
polar region of the surfactant, and secondly water that for other reason are trapped within the
adsorbed layer. Both trapped water and water of hydration will contribute to the frequency
shift provided that it moves with the oscillating crystal over the time scale of the experiment,
where one oscillating period for the fundamental mode is of the order of 0.2 µs. T o
understand the hydration of hydrophilic surfactant headgroups it is worthwhile to recapitulate
some of the properties of water.
Water is far from being a simple liquid if not unique1. The complexity of liquid water is due
to a combination of the small size and distinct polar charge distribution of the water
molecule2. The hydrogen bond is caused by a combination of electrostatic, charge transfer,
dispersion forces and exchange forces1. It is nowadays often assumed that the electrostatic
part is the most important one. The charge distribution of the water molecules can be
modelled by four charges being located in the form of a tetrahedron3, which allows each water
molecule to participate in four strong interactions4-5 with a high degree of spatial
directionality. The energy of a hydrogen bond between two water molecules depends on the
oxygen-oxygen distance and the hydrogen-oxygen angle1. This strong (water-water)
interaction results in a large cohesive energy, a high boiling point, a high surface tension, and
a reluctance to dissolve inert (non-polar, hydrophobic) solutes with which water cannot
interact through similarly strong forces. However, water can also bind to and dissolve polar
and hydrophilic substrates. The binding of water to hydrophilic groups can conveniently be
studied on macroscopic surfaces using various surface force techniques.
By adsorbing surfactants onto the substrate surfaces in such a way that the polar part of the
surfactant is directed towards solution, the hydration of the surfactant layer can be
investigated. Such studies have shown that in addition to attractive van der Waals forces and
53
repulsive electrostatic double – layer forces, an additional repulsive force is present at short
separations. This force is due to a combination of dehydration of polar groups (a hydration
force)7 and restriction of the perpendicular motion of adsorbed molecules in the gap between
the surfaces, a steric/protrusion force8. The measurable range of this force is typically 1-3 nm,
and it decays roughly exponentially with surface separation having a decay length of about
0.1 - 0.3 nm. Let us for the moment assume that the major cause of this force is dehydration.
The water molecules adjacent to the polar surface, or a surfactant headgroup, are strongly
affected and can be regarded as bound, In e.g. NMR-studies of ethylene oxide-based
surfactants, the amount of bound water is estimated to be about 2 - 8 per ethylene oxide unit9.
However, also water molecules further away will be affected by the presence of a surface or a
polar surfactant headgroup, but less strongly so. It is not seen as “bound” in a NMR
experiment, but it is sufficiently affected to give rise to a “hydration force” in a surface force
experiment. How much of the “hydration water” that is sensed in a QCM-measurement
remains an open question.
Surface ΓΓΓΓ [[[[mg/m2]]]] Ra [[[[nm]]]] θθθθ
Silicaellip 2.12 ± 0.10 1.3 ± 0.1 < 20°
Silanellip. 1.58 ± 0.10 1.1 ± 0.2 98° ± 1°
SilicaQCM 4.25 ± 0.24 1.3 ± 0.2 < 20°
SilanQCM 2.73 ± 0.06 1.0 ± 0.1 98° ± 1°
Table 5: Adsorbed amount ( Γ ), roughness average ( Ra ), and the Contact angle ( θ ), for the silica and
silanised silica model surfaces.
Trapped water on the other hand, involves water molecules that are confined within a limited
geometry such as inside a vesicle, beneath a bilayer, or between tightly packed
micellar/vesicle structures and the surface. This trapped water will also be contributing to the
frequency shift of the QCM-DTM. The experimental observation is that quite a lot of water
contributes to the adsorbed mass sensed by the QCM. For instance, when we compare results
for C14EO6 adsorption as elucidated from the QCM measurements with that obtained from
ellipsometry (paper V) we found that approximately 73 % of the mass sensed by the QCM-
DTM was due to water for silanated silica substrates. For the hydrophilic silica surface this
54
overestimation was approximately 100 %. Hence, this effect is very large and it can certainly
not be ignored. In order to verify the similarity of the substrates employed using the two
different techniques, surface properties such as surface roughness and contact angles were
also determined. This comparison is elucidated in Table 2. Clearly, the surfaces are similar
enough to expect similar adsorption. Hence, the only possible interpretation is that water
contributes significantly to the mass sensed by the QCM-technique.
55
6. Concluding remarks
This thesis deals with the adsorption of surfactants, vesicles and emulsions at the solid-liquid
interface. A large part of the work has been concerned with the problems and ambiguities
concerning the interpretations of the frequency shift and the dissipation factor as determined
with a QCM-DTM device. Hence, methods to prepare and clean various solid surfaces are
described. We could use the fact that the electrodes on the piezoelectric quartz crystal are
made of gold, and the gold surfaces makes it possible to use thiol chemistry to make chemical
modifications. Thiol chemistry is probably the best surface modification technique available
at the moment; the possibilities are endless with different functionalised end groups being
available. Studies of adsorption at these well documented, very stable substrates is a dream
that came through for the surface scientist. Studies of concentrated emulsions are possible
with the QCM-DTM technique, simply because there are no need for a transparent solution,
which is a prerequisite for many adsorption techniques today. Another strength of the QCM-
technique is that adsorption onto previously adsorbed layers can easily be followed, which
also is a problem for most light-based techniques, since the modelling parameters increase
rapidly. Such adsorption upon adsorption experiments have to be carefully interpret since the
chemistry can be really complex, and the results ambiguous. One of the big “problems” with
the QCM-technique is that it also registers water associated with the adsorbed layer. This
results in an overestimation of the adsorption of the “pure” adsorbate. This “problem” can
easily be turned into an advantage by comparing with results obtained by light based
techniques, which gives the adsorbed amount without any hydration effects or effects of
trapped water. The result of this correlation gives an estimation of the amount of water
associated with the adsorbed layer. Considering all the literature based on hydration, and
hydration effects this is actually not a “problem”, instead it could be turned into something
really valuable for the scientific community. So, like all other techniques the QCM-DTM has
its drawbacks and advantages, the only real problem is when you are not aware of them.
56
Acknowledgements
It is now 2 years, 8 months and 10 days since I started my Ph.D study in the Department ofChemistry, Surface Chemistry. Hence, I am truly indebted to more people than I thought that Iwould be. So, this is for the assistance through the oscillations of my life and work.
Per Claesson, for giving me the scientific freedom, that I needed, and for being my supervisor during this time.
Katrin Boschkova, one of the best.
Jan-Christer Eriksson, for all words of wisdom.
The happily graduated surface chemists:Eva B, Magnus B, Thomas E and Andra D.
The unhappily ungraduated surface chemists:Marcus P, Mikael K, Atte K, Jonny E, Torbjörn P, Marie E and Brita R.
Personnel at YKI:Martin M, Thomas A, Lennart B and Britt N.
The so close, and so far away people:Markus J, Nill B, Anna N, Mattias Ö, Ulla J and Pontus E.
YKI is acknowledged for the cookies, copies, printouts and stamps; these things havecertainly made my life easier during this time.
At Q-Sense, Michael Rodahl and Ralf Richter.
And finally The SSF-programme, Colloid and Interface Technology, for its financial support.
57
References.
Gold
(1) Ibach, H., “Solid-State Physics”, Springer-Verlag, Heidelberg, 1996.
(2) Hayward, D. O., “Chemisorption and Reactions on a Metallic Films”, Anderson, J. R.,
Ed., Academic Press, London, 1971
Self-Assembled Monolayers (SAMs)
1) Ulman, A., “An Introduction to ULTRATHIN ORGANIC FILMS From Langmuir-
Blodgettto Self-Assembly”, Academic Press Inc. San Diego, CA, 1991.
(2) Swalen, J. D.; Allara, D. L.; Andrade, J. D.; Chandross, E. A.; Garoff, S.; Israelachvili, J.;
McCarthy, T. J.; Murray, R.; Pease, R. F.; Rabolt, J. F.; Wynne, K. J.; Yu, H., Langmuir,
1987, 3, 932.
(3) Nuzzo, R. G.; Allara, D. L. J., J. Am. Chem. Soc., 1983, 105, 4481.
(4) Alves, C. A; Smith, E. L.; Porter, M. D., J. Am. Chem. Soc., 1992, 114, 1222.
(5) Bain, C. D.; Troughton, E. B.; Tao, Y. T.; Evall, J.; Whitesides, G. M.; Nuzzo, R. G., J.
Am. Chem. Soc., 1989, 111, 321.
(6) Sellers, H.; Ulman, A.; Schnidman, Y.; Eilers, J. E., J. Am. Chem. Soc., 1993, 115, 9389.
(7) Samant, M. G.; Brown, C. A.; Gordon, J. G. I., Langmuir, 1991, 7, 437.
(8) Siepman, J. I.; McDonald, I., R., Molecular Physics, 1993, 79, 457.
(9) Allara, D. L.; Nuzzo, R. G., Langmuir, 1985, 1, 45.
(10) Bensebaa, F.; Voicu, R.; Huron, L.; Ellis, T. H., Langmuir, 1987, 13, 5335.
(11) Bertilsson, L.; Liedberg, B., Langmuir, 1993, 9, 141.
Silica
(1) Iler, R. K., “The Chemistry of Silica”, John Wiley and Sons, New York, 1979.
(2) Yaminsky, V. V.; Ninham, B. W.; Pashley, R. M., Langmuir, 1998, 14, 3223.
(3) Eskilsson, K.; Yaminsky, V. V., Langmuir, 1998, 14, 2444.
58
Silan coated silica
1) Eskilsson, K.; Tiberg, F., Macromolecules, 1997, 30, 6323.
(2) Parker, J. L.; Claesson, P. M.; Attard, P., J. Phys. Chem., 1994, 98, 8468.
(3) Yaminsky, V. V., Colloids and Surfaces A, 1998.
(4) Yaminsky, V. V.; Claesson, P. M.; Eriksson, J. C., Journal of Colloid and Interface
Science, 1993, 161, 91.
(5) Christenson, H. K.; Claesson, P. M., Advances in colloid and Interface Science, 2001, 91,
391.
(6) Rabinovich, Y. I; Derjaugin, B. V., Colloids and Surfaces, 1988, 30, 243.
(7) Parker, J. L., “ Progress in Surface Science”, 1994, 47, 205.
(8) Claesson, P. M.; Kjellin, U. R. M., “ Modern Characterization Methods of Surfactant
Systems”, ed Binks, B. P., Surfactant Science Series, Marcel Dekker, New York, 1999, 83,
255.
Contact angle
(1) Cassie, A. B. D., Faraday Disc. Chem. Soc., 1948, 3, 11.
(2) Israelachvili, J. N.; Gee, M. L., Langmuir, 1989, 5, 288.
XPS
X-ray Photoelectron Spectroscopy
(1) Siegbahn, K., Science, 1982, 217, 111.
(2) Czanderna, A. W. (Ed.), “Methods of Surface Analysis”, Vol. 1 of ”Methods and
Phenomena. Their Applications in Science and Technology”, Elsevier Publ. Co., Amsterdam,
1975.
(3) Briggs, D., (Ed.), “Handbook of X-ray and Ultraviolet Photoelectron Spectroscopy”,
Heyden, London, 1977.
(4) Carlson, T. A., “Photoelectron and Auger Spectroscopy”, Plenum Press, New York, 1975.
(5) Briggs, D.; Seah, M. P., (Ed), “Practical Surface Analysis”, John Wiley & Sons,
Chichester, 1983.
59
QCM-DTM
(1) Cady, W. G., “Piezoelectricity”, 2nd ed, Dover, New York, 1964.
(2) Lu, C.; Czanderna, A. W. (eds), “Applications of Piezoelectric Quartz Crystal
Microbalances”, Elsevier, Amsterdam, 1984.
(3) Warner, W. W.; Stockbridge, C. D., “Vacuum Microbalance Techniques”, Plenum Press,
New York, 1963, 3, 55.
(4) Rodahl, M.; Höök, F.; Krozer, A.; Brzezinski, P.; Kasemo, B., Rev. Sci. Instr., 1995, 66,
3924.
(5) Kasemo, B.; Törnqquist, E., Surface Science, 1978, 77, 209.
(6) Kasemo, B.; Törnqquist, E., Phys. Rev. Lett., 1980, 44, 1555.
(7) Bishop, D. J.; Reppy. J. D., Phys. Rew. Lett., 1978, 40, 1727.
(8) Mecea, V. M., Sensors and Actuators A, 1993, 40, 1.
(9) Kleinman, R. N.; Kaminsky, G. K.; Reppy, J. D.; Pindak, R.; Bishop, D. J., Rev. Sci.
Instr., 1985, 56, 2088.
(10) Widom, A.; Krim, J., Phys. Rev. B, 1986, 34,1403.
(11) Krim, J.; Chiarello, R., J. Vac. Sci. Tech. B, 1991, 9, 1343.
(12) Caruso, F.; Rinia, H. A.; Furlong, D. N., Langmuir, 1996, 12, 2145.
(13) Caruso, F.; Serizawa, T.; Furlong, D.; Okahata, Y., Langmuir, 1995, 11, 1546.
(14) Furlong, D.; Aston, J. R., Colloids and Surfaces A, 1982, 4, 121.
(15) Salt, D., “Hy-Q Handbook of Quartz Crystal Devices”, van Nostrand Reinhold Co.
Padstow, Cornwall, 1987.
(16) Brice, J. C., Rew. Mod. Phys., 1985, 57, 105.
(17) Rodahl, M., Thesis, “On the frequency and Q factor response of the quartz crystal
microbalance to liquid overlayers”, Chalmers University of Technology, Gothenburg, 1995.
(18) Lack, F. R.; Willard, G. W.; Fair, I. E., Bell System Technical Journal, 1934, 13, 453.
(19) Sauerbrey, G., Z. Phys. (german), 1959, 155, 206.
60
(20) Pulker, H. K. Z., Angew, Phys, 1966, 20, 537.
(21) Mecea, V.; Bucur, R. V., Thin Solid Films, 1979, 60, 73.
(22) Nomura, T.; Hattori, O., Analytica Chimica Acta, 1980, 115, 323.
(23) Kanazawa, B. K.; Gordon, J. G., Analytica, Chimica, Acta, 1985, 175, 99.
(24) Stockbridge, C. D., “Vacuum Microbalance Techniques”, Plenum Press, New York,
1966, 5, 147.
(25) Martin, S. J.; Frye, G. C.; Wessendorf, K. O., Sensors and Actuators A, 1994, 44, 209.
(26) Smith, K. L., Electronics & Wireless World, 1986, july, 51.
(27) Spencer, W. J.; Smith, K. L., J. Appl. Phys., 1966, 37, 2557.
(28) Barnes, C.; D’Silva, C.; Jones, J. P.; Lewis, T. J., Sensors and Actuators B, 1991, 3, 295.
(29) Martin, S. J.; Frye, G. C.; Ricco, A. J.; Senturia, S. D., Anal. Chem., 1993, 65, 2910.
(30) Lin, Z.; Ward, M. D., Anal. Chem., 1995, 67, 685.
(31) Lucklum, R.; Schranz, S.; Behling, C.; Eichelbaum, F.; Hauptmann, P., Sensors &
Actuators A, 1997, 60, 40.
(32) Schneider, T. W.; Martin, S. J., Anal. Chem., 1995, 67, 3324.
(33) Bund, A.; Schwitzgebel, G., Analytica Chimica Acta, 1998, 364, 189.
(34) Tessier, L.; Patat, F.; Schmitt, N.; Feuillard, G., Anal. Chem., 1994, 66, 3569.
(35) Mecea, V. M.; Carlson, J. O.; Bucur, R. V., Sensors and & Actuators A, 1996, 53, 371.
(36) Martin, B. A.; Hager, H. E., J. Appl. Phys., 1989, 65, 2627.
(37) Ward, M. D.; Buttry, D. A., Science, 1990, 249, 1000.
(38) Frye, G. C.; Martin, S. J., J. Appl. Spec. Rev., 1991, 26, 73.
(39) Kanazawa, K., Faraday Discussions, 1997, 107, 77.
(40) Martin, B. A.; Hager, H. E., J. Appl. Phys., 1989, 65, 2630.
(41) Borovsky, B.; Mason, B. L.; Krim, J., J. Appl. Phys., 2000, 88, 4017.
61
Lipids
(1) The whole issue of,“Chemistry and Physics of Lipids”, 1991, 57, 1.
(2) Lewis, R. N. A. H.; Mak, N.; Edhaney, R. N. M., Biochemistry, 1987, 26, 6118.
(3) Seddon, J. M.; Cevc, G.; Kayl, R. D.; Marsh, D., Biochemistry, 1984, 23, 2634.
(4) Lindblom, G.; Rilfors, L., Biochim. Biophys. Acta, 1989, 988, 221.
(5) “Lipids-Molecular Organization, Physical Functions and Technical Applications”,
Larsson, K, The Oily Press Ltd, Dundee, Scotland, 1994.
Cationic surfactants
(1) Rosen, M. J., “ Surfactants and Interfacial Phenomena”, John Wiley, New York, 1989, 33.
(2) Mittal, K. L., “ Solution Chemistry of Surfactants” Pharm Press, New York, 1979, 1.
(3) Florence, A. T., Attwood, D., “ Physiochemical Principles of Pharmacy”, Macmillan,
Houndsmill, 1985, 174.
(4) Yeskie, M. A.; Harwell, J. H., J. Phys. Chem., 1988, 92, 2346.
(5) Koopal, L. K.; Keltjens, L., Colloids and Surfaces, 1986, 17, 371.
(6) Gao, Y.; Du, J.; Gu, T., J. Chem. Soc., Faraday Trans. I, 1987, 83, 2671.
(7) Gu, T.; Gao, Y.; He, L., J. Chem. Soc., Faraday Trans. I, 1988, 84, 4471.
(8) Gu, T.; Huang, Z., Colloids and Surfaces, 1989, 40, 71.
(9) Gu, T.; Zhu, B.Y., Colloids and Surfaces, 1990, 44, 81.
(10) Koopal, L. A.; Ralston, J., J. Colloid Interface Sci., 1986, 112, 362.
(11) Rupprecht, H.; Kindl, G., Progr. Colloid Polymer Sci., 1976, 60, 194.
(12) Hunter, R. J., “ Foundation of Colloid Sciences”, Oxford Sciences, Oxford, 1987, 1, 591.
(13) Koopal, L. K.; Wilkinson, G. T.; Ralstone, J., J. Colloid Interface Sci., 1988, 126, 493.
(14) Zhu, B. Y.; Gu, T., J. Chem. Soc. Faraday Trans. I, 1989, 85, 3813.
(15) Lee, G.; Rupprecht, H., J. Colloid Interface Sci., 1985, 105, 257.
Non-ionic surfactants
(1) Desbene, P. L.; Portet, F.; Treiner, C. J., J. Colloid Interface Sci., 1997, 190, 350.
(2) Böhmer, M. R.; Koopal, L. K.; Janssen, R.; Lee, E. M.; Thomas, R. K.; Rennie, A. R.,
Langmuir, 1992, 8, 2228.
62
(3) Partyka, S.; Zaini, S.; Lindheimer, M.; Brun, B., Colloids Surf., 1984, 12, 255.
(4) Penfold, J.; Staples, E.; Tucker, I.; Cummins, P., J. Phys. Chem., 1996, 100, 18133.
(5) Portet, F.; Desbene, P. L.; Treiner, C., J. Colloid Interface Sci., 1996, 184, 216.
(6) Tiberg, F.; Jönsson, B.; Tang, J.; Lindman, B., Langmuir, 1994, 10, 2294.
(7) Mandell, L.; Fontell, K.; Ekwall, P., “ Ordered Fluids and Liquid Crystals”, Gould, R. F.
ed., American Chemical Society, Washington D. C., 1967, 63, 89.
(8) Ekwall, P., “ Advances in Liquid Crystals”, Brown, G. H. ed., Academic Press, New York,
1975, 1, 1.
(9) Luzzati, V.; Mustacchi, H.; Skoulios, A.; Husson, F., Acta Cryst., 1960, 13, 660.
(10) Luzzati, V.; Mustacchi, H.; Skoulios, A., Disc. Farad. Soc., 1958, 25, 43.
(11) Tiberg, F., J. Chem. Soc, Faraday Trans., 1996, 92, 531.
(12) Levitz, P., Langmuir, 1992, 7, 1595.
(13) Levitz, P.; Van Damme, H.; Keravis, D., J. Phys. Chem., 1984, 88, 2228.
(14) Levitz, P.; Van Damme H., J. Phys. Chem., 1986, 90, 1302.
Adsorption of emulsions studied with the QCM-DTM
(1) Malmsten, M.; Lindström, T.; Wärnheim, T., J. Colloidal Interface Sci., 1995, 173, 297.
(2) Blomberg, E.; Claesson, P. M.; Wärnheim, T., Colloids Surf, (in press).
Polymer adsorption on phospholipid coated surfaces
(1) Keller, C.; Kasemo, B.; Biophys. J., 1998, 75, 1397.
Counterion effects on sensed mass and energy dissipation
(1) Bertilsson, L.; Liedberg, B., Langmuir, 1993, 9, 141.
Bound /trapped water
(1) Frank, F., “ Water: A Comprehensive Treatise”, Plenum Press, New York, 1972-1982, 1-
7.
(2) Marcelja, S., “ Liquids at Interfaces”, Elsevier Science, Amsterdam, 1990, 2,99.
(3) Coulson, C. A., “ Valence”, Oxford Uni. Press, London, 1961, 12.
63
(4) Tanford, C., “ The hydrophobic Effect”, Wiley, New York, 1980.
(5) Hunter, R. J., “ Foundations of Colloid Science”, Claredon, Oxford, 1989, 1.
(6) Israelachvili, J. N., “ Intermolecular and Surface Forces with Applications to Colloidal
and Biological Systems”, Academic Press, London, 1992.
(7) Rand, R. P.; Parsegian, V. A., Biochim. Biophys. Acta, 1989, 988, 351.
(8) Israelachvili, J.; Wennerström, H., Nature, 1996, 379, 219.
(9) Derjaugin, B. V.; Churaev, N. V., “ Fluid Interfacial Phenomena”, Wiley, Chichester,
1986, 15, 663.
(10) Nilsson, P. G.; Wennerström, H.; Lindman, B., J. Phys. Chem., 1983, 87, 1377.
(11) Drost Hansen, W., “ Biophysics in Water”, Wiley, Chichester, 1982, 163.
(12) Israelachvili, J. N., Chemica Scripta, 1985, 25, 7.
(13) Israelachvili, J. N.; Pashley, R. M., “ Biophysics of Water”, Wiley, Chichester, 1982,
183.