hydrophobicity and its applications by a dissertation...
TRANSCRIPT
HYDROPHOBICITY AND ITS APPLICATIONS
BY
FABIAN RIOS, B.S., M.A.
A dissertation submitted to the Graduate School
in partial fulfillment of the requirements
for the degree
Doctor of Philosophy
Major Subject: Chemistry
New Mexico State University
Las Cruces New Mexico
November 2011
“Hydrophobicity and its applications,” a dissertation prepared by Fabian Rios in
partial fulfillment of the requirements for the degree, Doctor of Philosophy, has
been approved and accepted by the following:
Linda Lacey
Dean of the Graduate School
Sergei Smirnov
Chair of the Examining Committee
Date
Committee in charge:
Dr. Sergei Smirnov, Chair
Dr. David Smith
Dr. Jeremy Smith
Dr. Igor Sevostianov
ii
ACKNOWLEDGMENTS
I want to thank my advisor Dr Sergei Smirnov for his scientific guidance and
almost limitless patience. Without his constant generation of ideas and solutions
to difficult problems this work could have not been completed. I am also very
grateful to my graduate committee members, Dr David Smith, Dr Jeremy Smith
and Dr Igor Sevostianov for reading and evaluating this project.
I also want to acknowledge the help of all the members of the Smirnov’s group,
past and present. An especial thanks to Ivan Vlassiouk and Jessica Curtiss for all
their collaboration during the preparation of many of the experiments described
in the next chapters.
Finally, I want to thank the Chemistry Department at NMSU for all the sup-
port received during these years.
iii
VITA
May , 1976 Born at Bucaramanga, Santander, Colombia
1993-1999 B.S., Universidad Industrial de Santander, Bucaramanga
2001-2003 M.A., University of Texas, Austin
2005-2001 Research and Teaching Assistant, Department of Chemistry,
New Mexico State University.
Major Publications
1. Fabian Rios and Sergei N. Smirnov. ph valve based on hydrophobicity
switching. Chem. Mater., 23(16):3601–3605, 2011.
2. Sergei Smirnov, Ivan Vlassiouk, Pavel Takmakov, and Fabian Rios. Water
confinement in hydrophobic nanopores. pressure-induced wetting
and drying. ACSnano, 4(9):5069–5075, 2010.
3. Fabian Rios and Sergei Smirnov. Biochemically responsive smart
surface. ACS Appl. Mater. Interfaces, 1(4):768–744, 2009.
4. Ivan Vlassiouk, Fabian Rios, Sean A. Vail, Devens Gust, and Sergei
Smirnov. Electrical conductance of hydrophobic membranes or
what happens below the surface. Langmuir, 23(14):7784–7792, 2007.
FIELD OF STUDY
Major Field: Physical Chemistry
iv
ABSTRACT
HYDROPHOBICITY AND ITS APPLICATIONS
FABIAN RIOS, B.S., M.A.
Doctor of Philosophy
New Mexico State University
Las Cruces, New Mexico, 1997
Dr. Sergei Smirnov, Chair
Two different types of smart surfaces that are able to change their hydropho-
bicity by different stimuli are presented. First, wetting changes induced by the
presence of the protein streptavidin on biotinylated surfaces are evaluated. Sec-
ond, a mixed layer of aliphatic and aminated silanes is used to induced wettability
changes by variation on pH.
The wetting of hydrophobic porous substrates was also studied. The wetting
of nanopores induced by pressure, surfactants and pH was analyzed. Conditions
for spontaneous dewetting were determined in pores of 20-200 nm of diameter.
The concepts on smart surfaces, wetting of nanopores and hydrophobicity
switching are applied on the design of a dual release controlled hydrophobic
nanocontainer. This nanocontainer holds a cargo and release it by the presence
of amphiphilic molecules or by pH changes. The potential of this nanocontainer
as a drug delivery system is explored.
v
Contents
1 Introduction 2
1.1 Basics of wettability . . . . . . . . . . . . . . . . . . . . . . . . . 4
1.1.1 Chemical Modification . . . . . . . . . . . . . . . . . . . . 8
1.2 Smart Surfaces . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9
1.2.1 Smart Surfaces with Special Wettability . . . . . . . . . . 9
1.3 Nanoporous substrates . . . . . . . . . . . . . . . . . . . . . . . . 15
1.3.1 Capillarity . . . . . . . . . . . . . . . . . . . . . . . . . . . 16
1.4 Chemical Modification with Silanes . . . . . . . . . . . . . . . . . 18
1.4.1 Formation of silane monolayers . . . . . . . . . . . . . . . 18
2 Biochemically Responsive Smart Surface 31
2.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31
2.2 Experimental section . . . . . . . . . . . . . . . . . . . . . . . . . 33
2.3 Results and discussion . . . . . . . . . . . . . . . . . . . . . . . . 37
2.4 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45
3 Wetting of hydrophobic nanopores induced by pressure and am-
vi
phiphile adsorption 50
3.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 50
3.2 Experimental . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 52
3.2.1 Pressure induced wetting . . . . . . . . . . . . . . . . . . . 52
3.2.2 Amphiphile induced wetting . . . . . . . . . . . . . . . . . 54
3.3 Results and Analysis . . . . . . . . . . . . . . . . . . . . . . . . . 55
3.3.1 Pressure induced wetting . . . . . . . . . . . . . . . . . . . 55
3.3.2 Amphiphile induced wetting . . . . . . . . . . . . . . . . . 66
3.4 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 72
4 pH Valve Based on Hydrophobicity Switching 77
4.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 77
4.2 Experimental Section . . . . . . . . . . . . . . . . . . . . . . . . . 79
4.3 Results and Discussion . . . . . . . . . . . . . . . . . . . . . . . . 81
4.3.1 IR Absorption . . . . . . . . . . . . . . . . . . . . . . . . . 81
4.3.2 pH Sensitive Wetting of Flat Surfaces . . . . . . . . . . . 83
4.4 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 92
5 Hydrophobic nanocontainer with dual release mechanism 96
5.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 96
5.2 Experimental section . . . . . . . . . . . . . . . . . . . . . . . . . 98
5.3 Results and discussion . . . . . . . . . . . . . . . . . . . . . . . . 99
5.4 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 105
vii
List of Figures
1.1 Representation of the CA . . . . . . . . . . . . . . . . . . . . . . 4
1.2 Wetting of rough hydrophobic surfaces . . . . . . . . . . . . . . . 7
1.3 Capillarity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16
1.4 Reaction of silanes with hydroxilated surfaces . . . . . . . . . . . 18
2.1 Representation of the surface modified with a mixed monolayer of
biotin-LC-LC and perfluoric acid. . . . . . . . . . . . . . . . . . . 36
2.2 CA variations for droplets with SA solutions of different concentra-
tions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39
2.3 Evaporation of droplets with and without streptavidin on biotini-
lated surfaces . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 40
2.4 Variation of the CA for capped and free SA on biotinilated surface 41
2.5 Illustration of the mechanism of CA variation . . . . . . . . . . . 42
2.6 CA of PBS droplet on SA treated surfaces . . . . . . . . . . . . . 43
3.1 Scheme of different hydrophobic surface modifications . . . . . . . 56
3.2 Illustration of different pore morphologies in a membrane with hy-
drophobic modification and their response to hydrostatic pressure
of water . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 59
viii
3.3 Variation of the impedance at 100 Hz with hydrostatic pressure for
commercial 0.2 and 0.02 μm membranes modified with SiNH2F8
and SiH16 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 60
3.4 Pressure dependence of the conductance for hydrophobic membranes 63
3.5 Variation of the ionic impedance through the hydrophobically mod-
ified alumina membrane . . . . . . . . . . . . . . . . . . . . . . . 66
3.6 Variation of ionic impedance at different concentration of amphiphiles
through a hydrophic nanoporous membrane . . . . . . . . . . . . 67
3.7 CA and ionic impedance on flat surface and nanoporous membranes
modified with C16 silane . . . . . . . . . . . . . . . . . . . . . . . 68
3.8 Wetting of hydrophobic pores by amphiphiles . . . . . . . . . . . 70
3.9 Wetting of a hydrophobic capillary in contact with a GUV . . . . 71
4.1 FTIR spectra of alumina membranes modified with different mix-
tures of butyl and aminopropyl silanes . . . . . . . . . . . . . . . 82
4.2 SEM images of membranes with 0.2 nm pores . . . . . . . . . . . 84
4.3 Variation of the contact angle with time at different pH . . . . . . 85
4.4 Contact angles for droplets of different pH on surfaces modified
with mixtures of APTS and BTS . . . . . . . . . . . . . . . . . . 86
4.5 CA images of equilibrium sessile drops on surfaces modified with
mixtures of APTS and BTS . . . . . . . . . . . . . . . . . . . . . 88
4.6 Kinetics of safranin dye flux through nanoporous membranes with
different surface modifications and at different pH . . . . . . . . . 91
5.1 Nanoconatiner with dual release mechanism . . . . . . . . . . . . 100
5.2 Controlled release from a hydrophobic pH sensitive nanocontainer 101
ix
5.3 Release of safranine from modified free silica nanotubes at different
pH. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 104
x
List of Tables
1.1 Relating hydrophobicity and the contact angle with water . . . . 6
1.2 Examples of smart coatings . . . . . . . . . . . . . . . . . . . . . 10
1
Chapter 1
Introduction
Men have known about coating of materials since ancient times. Initially the main
purpose of such coatings was merely decorative such as those in the cave paintings
of Chauvet, Lacaux, Niaux, Cussac which are almost 35000 years old and many
others.[1]
As the industry evolved so did the chemistry and it was noticed that the useful
life of materials could be extended by applying a protective film.[2] Such films had
different mechanisms of action. Some of them were inhibitive, releasing ions slowly
to passivate the metal surfaces.[3] Others created a barrier to humidity such as
epoxides and urethanes. There were also sacrificial ones based on zinc compounds
that oxidized in preference to the metallic substrate. However, all of these coatings
had a static nature. They sit on the surface and serve mainly as a barrier to avoid
the deterioration of the bulk of the material.[4]
A new trend in coating technology has appeared recently. It has been devised
that surfaces can interact actively with their environment. If, as the result of
2
this interaction, the surface can produce a change in at least one of its physical
properties, the surface is called “smart”.[5] Application of these types of coatings is
extensive in medicine, aerospace, environmental protection, sensors, and personal
safety, among others.
This dissertation will focus on the design and application of smart surfaces that
can alter their wettability in response to their environment. In this chapter the
general features of such smart surfaces will be described. Also the basic principles
of wettability, definitions and ways of analyzing it will be discussed. In the end
the design and assembly of the different coatings on the surface of the materials
used during the development of this work will be explained.
In chapter 2 smart surfaces able to respond selectively to the presence of protein
Streptavidin will be discussed. The design, test and application of such surface
as a sensor will be demonstrated.
Chapter 3 describes the wetting of hydrophobic nanoporous by pressure and
amphiphiles. Conditions for the wetting and dewetting of this nanodevices will
be analyzed and their possible use as drug delivery systems will be explored.
In chapter 4 pH sensitive surfaces will be described. Such surfaces show a
hydrophobic behavior when in contact with solution of high pH and hydrophilic
behavior at low pH. If used as a coating in a porous support such surface can act
as a pH sensitive valve which will be presented.
Chapter 5 applies the result of the two previous chapters into the design of
a dual response smart nanocontainer. This mesoporous device can hold a molec-
ular cargo inside and deliver it in response to a pH change or the presence of
amphiphilic molecules.
3
1.1 Basics of wettability
Wetting, in a broad sense, describes the displacement of a fluid (liquid or gas) by
another fluid on a given surface (solid or liquid). Examples of where wettability
is important include the spreading of water droplets on glass, penetration of ink
in paper, liquid absorbency or repellency on fabrics, the enhanced oil recovery
by displacement, and many others. All of these applications have sparkled and
sustained the interest in studying wettability for more than a century.[6]
The most common way characterize wettability is using the contact angle (CA)
for a drop of liquid deposited on a surface, as shown in Figure 1.1. The contact
angle θ is the angle between the base line and the tangent to the liquid/gas
interface at the point where the three phases overlap.
Figure 1.1: Representation of the CA
The CA is related to the three interfacial surface tensions through the Young’s
equation:
cosθY =γsg − γsl
γlg(1.1)
4
where θY Young’s equilibrium contact angle and γij is the surface tension at
the interface between the i and j phases with the subscripts s, l and g representing
solid, liquid and gas, respectively.
Young’s equation is a very simplistic approximation that can be used only with
ideally flat surfaces.[7] The only parameters accessible through experimentation
are the CA and γlg, hence an additional equation is needed to determine the values
of γsl and γvg. More sophisticated models are required to obtain that equation
but development of such models has been only partially successful.[8, 9]
Real surfaces often do not follow Young’s equation and present a range of
CA values. Because of the surface inhomogeneity the contact line between three
interfaces can go through a number of metastable states having different contact
angles.[10] The measurable macroscopic CA can vary between the maximum value,
called the advancing CA (θa) and the minimum value called the receding CA (θr
). The two can be measured using a moving droplet on a tilted surface as the
front and the rear angle, respectively. Alternatively, θa appears as a stabilized CA
during addition of liquid to a drop. Likewise, during the removal of liquid from a
drop the CA stabilizes at θr. The difference between θa and θr is called hysteresis.
θY lies at some value between θa and θr.
The identification of the θY on a surface is still a matter of debate. Some
authors assume that in a smooth surface the advancing CA is the real θY . [8]For
rough surfaces it seems that an average between θa and θr will produce a better
approximation to θY . [11]
Although the use of CA to calculate accurate surface energies of inhomoge-
neous surfaces has been deemed futile, the measurements can be used to estimate
the wettability of a system. Surfaces with different wettability with respect to wa-
5
Surface Contact Angle Figure
Superhydrophobic >160
Hydrophobic >90
Hydrophilic <90
Superhydrophilic ~0
Table 1.1: Relating hydrophobicity and the contact angle with water
ter are conventionally referred to as either hydrophobic or hydrophillic as shown
in table 1.1 according to the values of the CA.
So far we have described the wettability on flat homogeneous surfaces. However
a surface can be altered to modify its wettability. The two main factors that can
produce such changes are:
• roughness of the surface
• chemical composition
6
Wenzel’s model Cassies Model
Figure 1.2: Wetting of rough hydrophobic surfaces
Surface Roughness
Surface roughness strongly affects the wetting character of a surface. Methods
such as etching, crystal growth and nanolithography can create topographic “de-
fects” and alter the surface roughness.[12] Such structures can emulate interesting
biological systems based on hydrophobicity such as the legs of the water strider
or self-cleaning of lotus leaves.[13, 14]
There are two main models describing wetting of a rough hydrophobic surface
(see figure 1.2).[15, 16, 17]
The Wenzel’s model describes the experimental contact angle (θW ) on a rough
surface by relating it to the θY of the flat surface according to the equation:
cosθW = rcosθ =r(γsg − γsl)
γlg(1.2)
Where r corresponds to the roughness factor defined as the ratio between the
real surface area and the geometric surface area. Based on this a smooth surface
with CA > 90o will be more hydrophobic when the roughness factor is increased.
The opposite will happen if the CA < 90o, meaning that the actual CA on a rough
surface will be smaller. Actually, only the former is true because the contact angles
always inrease on rough surfaces .[18]
7
In the Cassie’s model the liquid does not penetrate into the grooves made on
the surface. Hence, the surface can be divided into two zones one solid and one
vapor. Then, the effective CA will be given by:
cosθC = fscosθs + fvcosθv (1.3)
where fsand fvrepresent the area fractions of solid and vapor on the surface
under the drop. Opposite to the Wenzel case, Cassie’s theory predicts that the
apparent CA is always greater than that of the smooth surface.
1.1.1 Chemical Modification
Chemical modification is a versatile method to alter surface wettability. Using self-
assembled monolayers (SAM’s) it is possible to introduce virtually any chemical
group on a surface and hence modify its wetting character. Currently there are
two well developed routes to synthesize SAM’s, using alkanethiols on metallic
surfaces or silanes on oxides.[19]
Thiols used for SAM’s preparation are molecules possessing a head group (-
SH). Typically they have a methylene chain (-CH2-) and a terminal functional
group. Because of a long methylene chain, alkanethiols form on many metallic
substrates closely packed assemblies. The most common metallic substrates are
gold, silver, copper and palladium.
Preparation of SAM’s via alkanethiols is very simple and can be performed
from the vapor or liquid phase. The monolayer formation is driven by the strong
affinity of sulfur for metals and by the van der Waals interaction of the aliphatic
chains. In general, SAM’s formed by alkanethiols are stable and allow post-
8
functionalization reactions using terminal functional groups.[20]
The use of silanes in the modification of oxide surfaces has also received a lot of
attention in the literature.[21, 22]Since its introduction by Sagiv,[23] the method
has become common in the functionalization of SiO2 and Al2O3. Their silanization
will be discussed in more detail in section1.4 as this procedure is important for
the experiments performed in this thesis.
1.2 Smart Surfaces
The design of smart surfaces has attracted a lot of attention in recent years. [4,
24, 25] To be considered “smart” a surface must possess two main characteristics:
• It must detect a change in its environment.
• It must respond to such a change in a reproducible, specific and unambigous
manner.[24]
The ways these two features are realized by various system are extremely diverse.
Some representative examples are shown in table 1.2
1.2.1 Smart Surfaces with Special Wettability
Design of smart surfaces using wettability switching is a very active field of
research.[14, 31, 15, 32] Wettability is a property with significant technological
applications such as self-cleaning of surfaces, micro-fluidics, tunable optics, lab-
on-chip systems, and others. Some literature examples presented below illustrate
how hydrophobic character of a surface alters in response to different stimuli.
9
Coating Type OperatingGoal
Stimulus ResponsePrinciple
Bactericidal[26] Repeling orkilling
microbial orcellular agents
temperature Changing of thepolymer
conformation andexposing differentchemical groups
Anticorrosive[27] Release of selfhealing
material thathelps to
recover cracks
pH Release ofchemical fromnanocontainersdisperse in the
filmAntifouling[28] Removal of
accumulatedbiomaterials
temperature ChangingPolymer
solubility in thefilm at low
temperature toease detachment
Impact Sensitive[29] Nanocapsuleswith varied
wall strengthstore dyes of
differentcolors
pressure Zones subjectedto high enoughpressure get
colored
Temperature Sensitive[30] Dyesencapsulatedin hydrogels
exchangeprotons
temperature Changes intemperatureinduce colorchange in the
hydrogel
Table 1.2: Examples of smart coatings
10
Light
Light activation is one of the most popular stimuli for smart surfaces.[33]
There are two types of light-sensitive surfaces: intrinsic photoinduced charge sep-
aration response in semiconducting metal oxides and from photochromic organic
molecules attached to surfaces. Metal oxides such as TiO2[34], ZnO[35], WO2[36],
VaO2[37] can switch their wetting character from hydrophobic (CA>90) to hy-
drophilic (CA<90) by means of light irradiation multiple times. Formation of
hole-electron pairs at the surface upon irradiation with photon energy greater
than the band gap induces surface charging in these semiconductors. The holes
react with the surface oxygen creating vacancies that can induce dissociative ad-
sorption of water which turns the surface hydrophilic.[36]. The surface regenerates
its hydrophobic character when left in the dark.
Light induced changes in wettability can also be obtained by the use of organic
photochromic molecules. Azobenzenes,[38] spiropyran,[39], dipyridylethylenes,[40]
stilbenes,[41] and others can be assembled in films on different surfaces and upon
irradiation change the wetting character of the surface. The change is produced
by isomerization of the light sensitive group. Depending on the isomer structure,
a change in surface energy is induced and the wettability altered. Although the
differences in CA after irradiation are not as dramatic as in semiconductors (∼
20-30o), they can be enhanced through the use of rough surfaces.[42]
Temperature
Temperature has also been used as a stimulus to control wettability in films.
A common approach is to use of a thermo-responsive polymer that presents con-
formational changes upon temperature variation.[43, 44] For example Poly(N-
isopropylacrylamide) (PNIPAAm) can show transitions between hydrophobic and
11
hydrophylic states in a range of 10 oC. [45] At low temperatures the internal
carbonyl and amino groups tend to form hydrogen bonds with water. Once the
temperature is increased the polymer chains collapse on themselves due to the for-
mation of internal hydrogen bonds between the same carbonyl and amino groups.
This produces an increase in the surface hydrophobicity that is reversible over
numerous cycles.
Solvents
Mixtures of polymer brushes that contain hydrophobic and hydrophilic groups
can also be used for the fabrication of smart surfaces. [46, 47] The solvent response
effect can be achieved in different ways. One case is the use of mixtures of non-
miscible polymer brushes anchored to a solid surface. [48]When the film is exposed
to solvents with different polarity the brushes with greater affinity for the solvent
will emerge and the others will collapse changing the effective surface energy.
Variations of almost 30 degrees in the CA can be obtained by this method.
A second approach is to have a combined chain with high and low surface
energy groups. [47]Although the high energy groups would prefer to stay in the
bulk of the material they are maintained close to the surface by the low energy
groups to which they are attached. Once the surface is in contact with water the
high energy groups will migrate to the surface to decrease the interfacial surface
tension.[49]
A final approach is the synthesis of triblock polymer brushes of the type ABA.
In this case the middle block has different physicochemical properties than the
end blocks. When the film is exposed to a good solvent for all the blocks the
brushes will adopt an extended configuration. If the solvent has affinity only for
the middle block the brushes will fold. These changes in the configuration can
12
create effective changes in the CA of ~40o depending on the blocks used.[50]
Biological Molecules
Molecules of biological interest have been used to alter the wettability of sub-
strates as well. In a very smart approach Wang et al. [51], have demonstrated
the use of DNA to control surface wetting. To accomplish this an i-motif DNA
sequence was modified on one end with a hydrophobic group. In the folded state,
the DNA quadruplex will maintain the hydrophobic end far from the interface
creating a hydrophilic surface. When the complementary DNA oligomer is in-
troduced, the quadruplex will unfold exposing the hydrophobic group creating a
hydrophobic surface.
Bergese et al.[52] have also used DNA to alter the CA of water on a modified
surface. Thiol modified oligomers were attached to a gold surface and solutions
containing 100%, 50% and 0% complementary DNA sequences were deposited on
top. The CA results showed that the surface was able to identify the DNA present
in every solution. Although changes in CA are not dramatic there is almost a 20
degree difference between totally and non-complementary DNA.
The use of proteins and lipids for hydrophobicity switching have been described
by D’Andrea et al.[53] On a hydrophobic surface, lipids and proteins can adsorb
reversible turning it hydrophilic. The initial hydrophobic state can be recovered
by washing with solutions containing enzymes such as trypsin or lipases.
Electric Field
Electrowetting is the electric potential-induced spreading of a droplet on a
surface. Although first described by Lipmann at the end of the 19th century, it
was not until the early 90’s that it took its modern form application with the
work of Berge.[54, 55] In its most simple form a potential difference is applied
13
between a drop of liquid and the surface on which it is deposited. To avoid
electrolysis the surface electrode is isolated by a dielectric. This approach has
proved useful on interesting applications such as microfluidics,[56] microlenses,[57]
display technology[58] and others.[59]
Lahann et al.[60] have used electric fields to control wettability by a different
approach. They grafted long aliphatic thiols with carboxylic ends to a gold surface.
The carboxylic ends are hydrophilic but if a positive potential is applied to the gold
surface they bend and expose the aliphatic chain turning the surface hydrophobic.
For the effect to take place the thiols must be distanced from each other in a way
they have enough space to bend towards the surface.
pH
Carboxylic and amino groups have been known to alter a surface wettability
when in contact with solutions of different pH.[61, 62, 63] At low pH a carboxylic
group will be neutral and hence it will show a more hydrophobic character. Once
pH is increased the carboxylic group will be protonated and it will become more
hydrophilic producing a change in the CA. The opposite situation takes places with
the amino groups which are more hydrophobic at high pH and more hydrophilic
at low pH.
It is possible to adjust the pH at which the surface switches from hydrophobic
to hydrophilic in a number of ways. Depending on the structure of the aminated
molecule the pH of inflexion can be adjusted as was explained by Farley et al.[62] In
this work it was shown that primary amines are not very sensitive to pH. Secondary
amines, on the other hand, can produce larger changes in the wettability of the
surface.
A second option is to use the roughness of the surface as can be done in the
14
other types of films. In a rough surface the CA changes become more dramatic
as was explained before.[64, 65]
Finally, using mixtures of polymers, the change in the CA can also be cus-
tomized. Busquet et al. demonstrated this using a mixture of polystyrene and a
diblock copolymer with a polyacrylic segment.[66] This mixture was able to re-
spond not only to the nature of the gas phase but also to a solution pH showing
large CA’s at low pH and low CA’s at high pH.
1.3 Nanoporous substrates
Porous materials have a number of applications in different fields such as heteroge-
nous catalysis, adsorptive separations, filtration and others.[67, 68] However, with
the development of new processes that allow control of pore dimensions and their
surface chemistry, new applications have emerged, such as sensing and drug de-
livery systems (DDS)[69].
The use of nanopores in sensors and DDS requires ability to control molecular
transport through them. Martin’s group has demonstrated numerous examples.[70,
71, 72]The basic principle of these examples and others found in the literature,
is based on how pores respond to different stimuli such as an analyte. In some
cases trasport of molecules and ions is interrupted in response to the presence of
an analyte. The reverse case is also possible, a pore that is normally closed, gets
open when the analyte is present.
Similarly, DDS must have the means of carrying the cargo and releasing it
when necessary. The use of porous materials as nanocontainers for drug delivery
has been around for almost a decade,[73] and the controlled cargo release has
15
Hydrophilic capilar Hydrophobic capilar
Figure 1.3: Capillarity
become one of the most active areas of research in nanotechnology with more and
more inventive methods presented every day.
Wettability can be used to control mass flow through nanoporess as hydropho-
bic pores are normally closed (dry) in water but be made to reduce hydrophobicity
in response to stimuli and get the pores open. This principle can be applied in
sensors and DDS. This mechanism is closely related to capillarity which will be
explained in the next section.
1.3.1 Capillarity
The phenomenon of capillarity deals with interfaces that are sufficiently mobile to
assume an equilibrium shape.[74] Across a curved interface a pressure difference
(∆P ) is present. This ∆P is defined by the Young-Laplace equation as:
∆P = γ(1
R1
+1
R2
) (1.4)
Where R1 and R2 correspond to the principal radius of curvature that defines
an arbitrary surface.
Thus, a capillary in contact with a fluid can be found in two cases shown in
figure 1.3.
16
If the fluid wets the capillary walls, it rises above the surface until the force of
gravity of the column of liquid equals the force that develops accross the curved
meniscus. If on the contrary, the fluid is repelled by the surface, the negative ∆P
will maintain the fluid below the surface level. For both situation the value of h
is given by:
h =∆P
g∆ρ=
2γcosθ
rg∆ρ(1.5)
According to equation 1.5 there are at least 3 ways to wet a capillary. The first
one is to increase the pressure above ∆P .[75] This method has been explored for its
application in water-repellent fabrics and separation processes using hydrophobic
membranes.[76, 77]In those studies the authors where able to identify the liquid
entry pressure in different systems when the mass flux had no opposition. Smirnov
et al. have shown that pressures as high as 12 bar were required to wet hydrophobic
nanopores, in agreement with equation 1.4.[78]
Wetting of a hydrophobic capillary can be achieved also by altering the surface
tension by amphiphiles, as it has been demonstrated.[79, 80, 70]
Electrowetting can also generate a similar effect, decreasing the effective value
of the interfacial tension at the solid interface when applying a difference of po-
tential as described by Lu et al.[81]
A third method and a less explored one is the use of a smart surface that
can change its wetting character in response to a stimulus. Vlassiouk et al have
demonstrated this method using a light sensitive membrane prepared by surface
modification with a mixture of photocromic spiropyran molecules and hydrophobic
decanoic tails[82]. Such a membrane remains dry until exposed to UV light. Light
17
irradiation photoisomerized spiropyran into the more polar merocyanin form which
switched the surface into hydrophilic allowing for ionic current to go through.
1.4 Chemical Modification with Silanes
Since its introduction in 1980, chemical modification using silanes have become
ubiquitous in surface science.[23] Their popularity in the field of nanotechnology
arises due to their high reactivity with hydroxylated surfaces such as silica or
alumina. An additional advantage is their greater stability than that of thiol
monolayers on metal surfaces.[21]
1.4.1 Formation of silane monolayers
Reaction of silanes with metal oxide surfaces can take place from the liquid or the
gas phase under mild conditions. This is due to a high reactivity of silanes with
hydroxyls covering the surface of such oxides. The general reaction of different
types of silanes with hydroxilated surfaces is shown in figure
Figure 1.4: Reaction of silanes with hydroxilated surfaces
The high reactivity of silanes can be of disadvantage when trying to control
the reaction. The reaction rate and its extent depend on several factors such
18
as temperature, solvent, type of silane, water content, time of reaction etc. and
the outcome can be very sensitive to these conditions. This makes the matter of
reproducibility a challenge as shown by the variety of protocols reported in the
literature.[20]
Water concentration on the surface and in solution was identified from the
beginning as one of the key factors in the formation of the silane layer. Sa-
giv proposed that silanes were hydrolized with water at the surface and then
a condensation would proceed with the hydroxyl groups available at the solid
interface.[23] Rye et al tried to explain further the role of water in the formation
of the reaction of chlorosilanes. They identified that the reaction in the absence
of water will not proceed to completion since there was not possibility of lateral
polymerization.[83]In some cases silanization will not proceed at all in the absence
of water.[84] McGovern et al identified 0.15 mg of moisture/100 mL of solvent as
ideal to obtain complete coverage without overpolymerization.[85]Double silaniza-
tion with an intermediate hydrolysis step has been also proposed as a good method
to obtain complete coverage[86].
The role of solvent is also intimately related to water content. McGovern et
al. compared different solvents during adsorption of octadecyl thrichlorosilane
(OTS).[85]They found that the longer the aliphatic hydrocarbon the better the
coverage obtained. They also analyzed the capacity of the solvent for the extrac-
tion of water from the surface to the bulk. In this case benzene and cyclohexane
produce much better coverage than hexane. This also demonstrated that the
hydrolysis of the silane takes place in solution and not close to the surface.
The deposition of silanes from the gas phase has also been used although
much less than from the liquid phase.[87, 88, 89] Liquid phase deposition is easier
19
to carry out but monolayers from gas phase present certain advantages such as
smother films,[90, 91]easy control of unwanted polymerization[92] and elimination
of solvent waste.[93]
The type of silane is another important factor to take into account since their
reactivity varies depending on their chemical nature. In general a silane molecule
is composed of a head group (-SiX3where X is usually -Cl, -OMe or -OEt), an alkyl
chain of -CH2- and the terminal group.[94]In the case of the head group it is known
that chlorosilanes are more reactive than alkoxysilanes and form denser layers.
However their reactivity makes them very sensitive to the amount of water, which
can lead to problems with reproducibility. Silanization with alkoxysilanes takes
longer times but in general is more reproducible, their reactivity depending on the
size of the alkoxygroup (-OCH3> -OC2H5> -OC4H7).[95]The number of reactive
substituents at the Si can affect also the monolayer formation. For example,
Moon et. al have demonstrated that in the case of aminopropyl silanes in toluene,
the trimethoxy silane produces multilayers, while dimethyl methoxy silane cannot
generate complete coverage; methyl dimethoxy silane, on the other hand, does
not produce multilayers and its monolayer coverage is denser than for dimethyl
methoxy silane.[96, 97]
The alkyl chain length of the silane also affects packing of the silane layer. It
has been shown through ellipsometry measurements that short alkyl chains tend
to form disordered multilayers while longer alkyl chains form ordered monolayers.
Bierbaum et al compared alkylchains of 3, 18 and 30 carbons long and found that
the optimal length for a well formed monolayer was 18.[98] Similarly, Wasserman
et al. found that short length alkyl chains do not pack very well and have smaller
CA than those of longer silanes and for the chains longer than 7 carbons the CA
20
become independent of the silane used. [99]
The nature of the terminal group also influences the monolayer formation.
This is especially true for terminal amino silanes. Wasserman et al. identified
that aminosilanes with a long alkyl chain (n=18) form a more disordered layer
than a similar aliphatic silane which was attributed to interaction of the amino
group with the surface hydroxyls during assembly.[99] Wang et al. also pointed
out that an aminopropyl trimethoxysilane (APTS) layer is less stable than the one
formed with propyl trimethoxysilane (PTS).[100] The same group also suggested
that the instability can be reduced by using longer alkyl chains.
21
Bibliography
[1] Balter, M. Science 2008, 321, 904–905.
[2] Hare, C. H. Corrosion and Its Control by Coatings; Tracton, A. A., Ed.;
CRC Press, 2006; Vol. 0, pp 1–9.
[3] Pokhmurskyi, V. I.; Zin, I. M.; Lyon, S. B. Mater. Sci. 2004, 41, 383–390.
[4] Brady, R. F. Twenty-First Century Materials: Coatings That Interact with
Their Environment ; Provder, J., Theodore; Baghdachi, Ed.; Oxford Uni-
versity Press, 2007; Vol. 957, pp 3–11.
[5] Yoshida, M.; Lahann, J. ACSnano 2008, 2, 1101–1107.
[6] Bose, A. Wetting by solutions; Mittal, K. L., Ed.; Marcel Dekker Inc., 1993;
Vol. 49, pp 149–181.
[7] Kwok, D. Y.; Neumann, A. W. Contact Angle measurements and criteria
for surface energetic interpretation; Mittal, K. L., Ed.; Brill, 2003; Vol. 3,
pp 117–159.
[8] Kwok, D. Y.; Neumann, A. W. J. Phys. Chem. B 2000, 104, 741–746.
[9] Bahramian, A.; Danesh, A. J. Colloid Interface Sci. 2004, 279, 206–212.
22
[10] Johnson, R. E.; Dettre, R. H. Wetting of Low-Energy Surfaces; Berg, J. C.,
Ed.; Marcel Dekker Inc., 1993; Vol. 46, pp 1–73.
[11] Fabretto, M.; Sedev, R.; Ralston, J. Advancing, receding and vibrated con-
tact angles on rough hydrophobic surfaces; Mittal, K. L., Ed.; Brill, 2003;
Vol. 3, pp 161–173.
[12] Roach, P.; Shirtcliffe, N. J.; Newton, M. I. Soft Matter 2008, 4, 224–240.
[13] Xia, F.; Jiang, L. Adv. Mater. 2008, 20, 2842–2858.
[14] Sun, T.; Feng, L.; Gao, X.; Jiang, L. Acc. Chem. Res. 2005, 38, 644–652.
[15] Feng, X.; Jiang, L. Adv. Mater. 2006, 18, 3063–3078.
[16] Quere, D. Physica A 2002, 313, 32–46.
[17] Marmur, A. Langmuir 2003, 19, 8343–8348.
[18] Li, D.; Neumann, A. W. Thermodynamic Status of Contact Angles; Neu-
mann, J. K., A. W.; Spelt, Ed.; Marcel Dekker Inc., 1996; Vol. 63, pp
109–168.
[19] Schreiber, F. Prog. Surf. Sci. 2000, 65, 151–256.
[20] Ulman, A. Chem. Rev. 1996, 96, 1533–1554.
[21] Haensch, C.; Hoeppener, S.; Schubert, U. S. Chem. Soc. Rev. 2010, 39,
2323–2334.
[22] Onclin, S.; Ravoo, B. J.; Reinhoudt, D. N. Angew. Chem. Int. Ed. 2005,
44, 6282–6304.
23
[23] Sagiv, J. J. Am. Chem. Soc. 1980, 102, 92–98.
[24] Baghdachi, J. Smart Coatings ; Baghdachi, T., Jamil; Provder, Ed.; Ameri-
can Chemical Society, 2009; Vol. 1002, pp 3–24.
[25] Russell, T. P. Science 2002, 297, 964–967.
[26] Laloyaux, X.; Fautre, E.; Blin, T.; Purohit, V.; Leprince, J.; Jouenne, T.;
Jonas, A. M.; Glinel, K. Adv. Mater. 2010, 22, 5024–5028.
[27] Shchukin, D. G.; Mohwald, H. Chem. Comm. 2011, 47, 8730–8739.
[28] Ista, L. K.; Lopez, G. P. J. Ind. Microbiol. Biotech. 1998, 20, 121–125.
[29] Liu, T.; Campbell, B. T.; Burns, S. P.; Sullivan, J. P. Appl. Mech. Rev.
1997, 50, 227–246.
[30] Seeboth, A.; Kriwanek, J.; Vetter, R. J. Mater. Chem. 1999, 9, 2277–2278.
[31] Xin, B.; Hao, J. Chem. Soc. Rev. 2010, 39, 769–782.
[32] Gras, S. L.; Mahmud, T.; Rosengarten, G.; Mitchell, A.; Kalantar-zadeh,
K. ChemPhysChem 2007, 8, 2036–2050.
[33] Wang, S.; Songa, Y.; Jiang, L. J. Photoch. Photobio. C 2007, 8, 18–29.
[34] Wang, R.; Hashimoto, K.; Fujishima, A.; Chikuni, M.; Kojima, E.; Kita-
mura, A.; Shimohigoshi, M.; Watanabe, T. Nature 1997, 388, 431–432.
[35] Feng, X.; Feng, L.; Jin, M.; Zhai, J.; Jiang, L.; Zhu, D. J. Am. Chem. Soc.
2004, 126, 62–63.
24
[36] Wang, S.; Feng, X.; Yao, J.; Jiang, L. Angew. Chem. Int. Ed. 2006, 45,
1264–1267.
[37] Lim, H. S.; Kwak, D.; Lee, D. Y.; Lee, S. G.; Cho, K. J. Am. Chem. Soc.
2007, 129, 4128–4129.
[38] Delorme, N.; Bardeau, J. F.; Bulou, A.; Poncin-Epaillard, F. Langmuir
2005, 21, 12278–12282.
[39] Rosario, R.; Gust, D.; Hayes, M.; Jahnke, F.; Springer, J.; Garcia, A. A.
Langmuir 2002, 18, 8062–8069.
[40] Cooper, C. G. F.; MacDonald, J. C.; Soto, E.; McGimpsey, W. G. J. Am.
Chem. Soc. 2004, 126, 1032–1033.
[41] Driscoll, P. F.; Purohit, N.; Wanichacheva, N.; Lambert, C. R.; McGimpsey,
W. G. Langmuir 2007, 23, 13181–13187.
[42] Jiang, W.; Wang, G.; He, Y.; Wang, X.; An, Y.; Song, Y.; Jiang, L. Chem.
Comm. 2005, 41, 3550–3552.
[43] Liang, L.; Feng, X.; Liu, J.; Rieke, P. C.; Fryxell, G. E. Macromolecules
1998, 31, 7845–7850.
[44] Lutz, J.-F.; Akdemir, O.; Hoth, A. J. Am. Chem. Soc. 2006, 128, 13046–
13047.
[45] Sun, T.; Wang, G.; Feng, L.; Liu, B.; Ma, Y.; Jiang, L.; Zhu, D. Angew.
Chem. Int. Ed. 2004, 43, 357–360.
25
[46] Tan, H.; Sun, T.; Li, J.; Guo, M.; Xie, X.; Zhong, Y.; Fu, Q.; Jiang, L.
Macromol. Rapid. Comm. 2005, 26, 1418–1422.
[47] Vaidya, A.; Chaudhury, M. K. J. Colloid Interface Sci. 2002, 249, 235–245.
[48] Draper, J.; Luzinov, I.; Minko, S.; Tokarev, I.; Stamm, M. Langmuir 2004,
20, 4046–4075.
[49] Thanawala, S. K.; Chaudhury, M. K. Langmuir 2000, 16, 1256–1260.
[50] Jennings, G. K.; Brantley, E. L. Adv. Mater. 2004, 16, 1983–1994.
[51] Wang, S.; Liu, H.; Liu, D.; Ma, X.; Fang, X.; Jiang, L. Angew. Chem. Int.
Ed. 2007, 46, 3915–3917.
[52] Bergese, P.; Oliviero, G.; Colombo, I.; Depero, L. E. Langmuir 2009, 25,
4271–4273.
[53] D’Andrea, S. C.; Fadeev, A. Y. Langmuir 2006, 22, 3962–3963.
[54] Mugele, F.; Baret, J.-C. J. Phys.: Condens. Matter. 2005, 17, 705–774.
[55] Quilliet, C.; Berge, B. Curr. Opin. Colloid Interf. Sci. 2001, 6, 34–39.
[56] Srinivasan, V.; Pamula, V. K.; Fair, R. B. Lab. Chip. 2004, 4, 310–315.
[57] Berge, B.; Peseux, J. Eur. Phys. J. E 2000, 3, 159–163.
[58] Hayes, R. A.; Feenstra, B. J. Nature 2003, 425, 383–385.
[59] Wheeler, A. R. Science 2008, 322, 539–540.
[60] Lahann, J.; Mitragotri, S.; Tran, T.-N.; Kaido, H.; Sundaram, J.; Choi,
I. S.; Hoffer, S.; Somorjai, G. A.; Langer, R. Science 2003, 299, 371–374.
26
[61] Bain, C. D.; Whitesides, G. M. Langmuir 1989, 5, 1370–1378.
[62] Holmes-Farley, S. R.; Bain, C. D.; Whitesides, G. M. Langmuir 1988, 4,
921–937.
[63] Holmes-Farley, S. R.; Reamey, R. H.; MCarthy, T. J.; Deutch, J.; White-
sides, G. M. Langmuir 1985, 1, 725–740.
[64] Yu, X.; Wang, Z.; Jiang, Y.; Shi, F.; Zhang, X. Adv. Mater. 2005, 17,
1289–1293.
[65] Jiang, Y.; Wang, Z.; Yu, X.; Shi, F.; Xu, H.; Zhang, X. Langmuir 2005, 21,
1986–1990.
[66] Bousquet, A.; Ibarboure, E.; Teran, F. J.; Ruiz, L.; Garay, M. T.; Laza,
J. M.; Vilas, J. L.; Papon, E.; Rodriguez-Hernandez, J. J. Polym. Sci. Pol.
Chem. 2010, 48, 2982–2990.
[67] Bhatia, S. K.; Bonilla, M. R.; Nicholson, D. Phys. Chem. Chem. Phys. 2011,
13, 15350–15383.
[68] Lu, G. Q.; Zhao, X. S. Nanoporous Materials-An Overview ; Lu, X. S., G.
Q.; Zhao, Ed.; Imperial College Press, 2004; Vol. 4, pp 1–12.
[69] Kohli, P.; Wirtz, M.; Martin, C. R. Electroanalysis 2004, 16, 9–18.
[70] Steinle, E. D.; Mitchell, D. T.; Wirtz, M.; Lee, S. B.; Young, V. Y.; Martin,
C. R. Anal. Chem. 2002, 74, 2416–2422.
[71] Wirtz, M.; Martin, C. R. Sensors Update 2002, 11, 35–64.
[72] Kobayashi, Y.; Martin, C. R. Anal. Chem. 1999, 71, 3665–3672.
27
[73] Vallet-Regi, M.; Ramila, A.; Real, R. P. d.; Perez-Pariente, J. Chem. Mater.
2001, 13, 308–311.
[74] Adamson, A. W.; Gast, A. P. Physical Chemistry of Surfaces; Wiley-
Interscience: New York, 1997.
[75] WASHBURN, E. W. Phys. Rev. 1921, 17, 273–283.
[76] Kim, B.-S.; Harriott, P. J. Colloid Interface Sci. 1986, 115, 1–8.
[77] Garcia-Payo, M. C.; Izquierdo-Gil, M. A.; Fernandez-Pineda, C. J. Colloid
Interface Sci. 2000, 230, 420–431.
[78] Smirnov, S.; Vlassiouk, I.; Takmakov, P.; Rios, F. ACSnano 2010, 4, 5069–
5075.
[79] Okamoto, K.; Shook, C. J.; Bivona, L.; Lee, S. B.; English, D. S. Nano Lett.
2004, 4, 233–239.
[80] Jayaraman, K.; Okamoto, K.; Son, S. J.; Luckett, C.; Gopalani, A. H.; Lee,
S. B.; English, D. S. J. Am. Chem. Soc. 2005, 127, 17385–17392.
[81] Lu, W.; Kim, T.; Han, A.; Chen, X.; Qiao, Y. Langmuir 2009, 25, 9463–
9466.
[82] Vlassiouk, I.; Park, C.-D.; Vail, S. A.; Gust, D.; Smirnov, S. Nano Lett.
2006, 6, 1013–1017.
[83] Rye, R. R.; Nelson, G. C.; Dugger, M. T. Langmuir 1997, 13, 2965–2972.
[84] Tripp, C. P.; Hair, M. L. Langmuir 1992, 8, 1120–1126.
28
[85] McGovern, M. E.; Kallury, K. M. R.; Thompson, M. Langmuir 1994, 10,
3607–3614.
[86] Krasnoslobodtsev, A. V.; Smirnov, S. N. Langmuir 2002, 18, 3181–3184.
[87] Hoffmann, P. W.; Stelzle, M.; Rabolt, J. F. Langmuir 1997, 13, 1877–1880.
[88] Sugimura, H.; Hozumi, A.; Kameyama, T.; Takai, O. Surf. Interface Anal.
2002, 34, 550–554.
[89] Chu-jiang, C.; Zhi-gang1, S.; Yu-shan, X.; Shu-lin, M. J. Phys. D. Appl.
Phys. 2006, 39, 4829–4837.
[90] Duchet, J.; Chabert, B.; Chapel, J. P.; Gerard, J. F.; Chovelon, J. M.;
Jaffrezic-Renault, N. Langmuir 1997, 13, 2271–2278.
[91] Pavlovic, E.; Quist, A. P.; Gelius, U.; Oscarsson, S. J. Colloid Interface Sci.
2002, 254, 200–203.
[92] WANG, W.; VAUGHN, M. W. Scanning 2008, 30, 65–77.
[93] WANG, Y.; FERRARI, M. J. Mater. Sci. 2000, 35, 4923–4930.
[94] Aswal, D.; Lenfant, S.; Guerin, D.; Yakhmi, J.; Vuillaume, D. Anal. Chim.
Acta 568, 568, 84–108.
[95] Arkles, B.; Steinmetz, J. R.; Zazyczny, J.; Mehta, P. Factors contributing to
the stability of alkoxysilanes in aqueous solution; Mittal, K. L., Ed.; VSP,
1992; Vol. 3, pp 91–104.
[96] Moon, J. H.; Shin, J. W.; Kim, S. Y.; Park, J. W. Langmuir 1996, 12,
4621–4624.
29
[97] Moon, J. H.; Kim, J. H.; Kim, K.-j.; Kang, T.-H.; Kim, B.; Kim, C.-H.;
Hahn, J. H.; Park, J. W. Langmuir 1997, 13, 4305–4310.
[98] Bierbaum, K.; Kinzler, M.; Woll, C.; Grunze, M.; Hahner, G.; Heid, S.;
Effenberger, F. Langmuir 1995, 11, 512–518.
[99] Wasserman, S. R.; Tao, Y.-T. Langmuir 1989, 5, 1074–1087.
[100] Wang, A.; Tang, H.; Cao, T.; Salley, S. O.; Ng, K. S. J. Colloid Interface
Sci. 2005, 291, 438–447.
30
Chapter 2
Biochemically Responsive Smart
Surface
2.1 Introduction
Materials that can respond selectively to their environment or to a certain stimuli
by switching one or more of their critical properties are called “smart” [1]. One
of the properties that has received special attention is wettability. A number of
groups have demonstrated the fabrication of smart surfaces that display wettabil-
ity changes induced by temperature [2], light [3, 4, 5, 6], electrical potential [7],
fumes of solvents [8], and pH [9]. A contact angle (CA) of a liquid drop on such
surfaces can be altered by the above-mentioned stimuli and, for example, switch
it from hydrophobic to hydrophilic.
The CA is the most common measure of wettability that describes the angle,
31
θ, at the three-phase contact line formed by a drop of liquid resting on a surface.
The Young’s equation relates it to surface energies at the solid/vapor (γsv), the
solid/ liquid (γsl), and the liquid/vapor, (γlv), interfaces.
The CA can vary between two values: the maximum, called advancing CA,
and the lowest, called receding CA. The former can be measured using the Sessile
drop method by increasing the size of the drop until no variations in the CA
are observed. The receding CA is measured as a minimum angle while gradually
removing liquid from the drop until the contact line begins to move backward. If
manipulation with the droplet is problematic, the receding angle can be measured
by observing droplet evaporation. When the volume of a droplet shrinks, its
shape changes while the contact line stays initially the same but eventually, upon
reaching the minimum possible receding angle, the contact line detaches and the
droplet continues shrinking without further changes in the shape.
The hysteresis between advancing and receding CA can vary significantly and
is due to metastable states at the solid/ liquid/vapor interface [10]. The surface
roughness, chemical heterogeneity, molecular reorientation, and penetration of the
small-sized liquid molecules into the voids of the solid surface have been identi-
fied among the numerous causes for metastable states on surfaces modified with
organic molecules [11]. Notably, even small amounts of impurities on the surface
(chemical heterogeneities) can lead to a large hysteresis [10].
Amino acid residues in the polypeptide chain of a protein vary in their hy-
drophobicity, and the hydrophobic interactions between them as well as with the
surrounding water are the driving force for proteins folding into their native state.
Except for the membrane proteins, the outer surface of a typical protein is usu-
ally enriched with hydrophilic residues, which make the protein water-soluble.
32
Hydrophobic surfaces can be rendered hydrophilic by covering them with pro-
teins. Such coverage can be achieved by relying on the amphipathic properties of
proteins: after a prolonged contact with a hydrophobic surface, they can change
conformation to “bind” in a nonnative form via exposure of their hydrophobic
residues to the surface. Such binding is weak and is efficient at a very high pro-
tein concentration; it is nondiscriminative and forms a protein film that switches
a hydrophobic surface into a hydrophilic one [12]. Different proteins have varying
tendencies of binding to hydrophobic surfaces, with bovine serum albumin (BSA)
being one example with a strong conformationally induced adsorption [13] that is
often used for hydrophobic surface modification [14].
To the best of our knowledge, there are no reports of surfaces that could
be switched from hydrophobic to hydrophilic by specific interactions between the
analyte proteins and their ligands on the surface. In the present work, we illustrate
a design of such smart surfaces using the well known couple biotin-streptavidin
(SA), which is commonly used for protein micropatterning. We show here that
the mixed hydrophobic-biotin surfaces respond specifically to the presence of the
SA analyte by lowering the CA at the surface.
2.2 Experimental section
Materials. (3-Aminopropyl)trimethoxysilane (APTS) was obtained from Aldrich.
2H,2H,3H,3H-Perfluoroundecanoic acid was obtained from Fluorous Technologies.
Biotin-LCLC- COOH and N-[3-(dimethylamino)propyl]-N-ethylcarbodiimide (EDC)
were obtained from Anaspec. Streptavidin (SA), a 53 kDa protein, was obtained
from Invitrogen. Bovine serum albumin (BSA), a 69 kDa protein, was received
33
from Sigma. They have similar pI ) 5 and 4.7, respectively. Methanol and ethanol,
both of absolute grade from Aldrich, were used as received. Glass slides were
cleaned with Piranha solution (30% H2O2 and 70% H2SO4) for 20 min at 70 C,
washed with copious amounts of distilled water, and dried in an oven for 30 min
at 115 C. Caution! Piranha solution is explosive.
Preparation of Aminated Surfaces. The first step for surface modification
is silanization of cleaned glass slides with an ethanol solution of APTS [15, 16]
for 12 h at room temperature and with constant shaking. This reaction produces
amino groups for further steps and is prone to multilayer growth in low-polarity
solvents, but if a proper solvent is used, this problem can be minimized and
practically avoided. In our experience, silanization using a 2% solution of APTS in
ethanol results in monolayer coverage, as judged by the surface density of amines.
Evaluation of the amino group surface density was performed using the method of
Moon et al. [16], and it was established that at least 6 h was necessary to attain
monolayer coverage, ca. 3 x 1014/cm2 [15]. No further increase beyond monolayer
coverage was detected for up to 24 h of treatment (in contrast with nonpolar
solvents). Silanization for all surfaces reported in this paper was performed using
12 h of treatment in ethanol, ensuring close to monolayer coverage. Afterwards,
slides were washed with ethanol and methanol and finally cured for over 1 h at
115 C.
Mixed Biotinylated and Fluorinated Surfaces. To prepare mixed mono-
layers, mixtures of two carboxylic acids in different proportions were reacted with
the amino groups of the aminosilane layer using an EDC coupling reagent. Per-
fluoric acid was chosen to minimize the passive adsorption of proteins [7, 17]. The
biotinylated acid consisted of D-biotin attached to a long linker (LC-LC-COOH).
34
The purpose of the long linker is to extend the biotin moiety above the fluorinated
monolayer to ensure its interaction with SA. Solutions of biotinylated and fluo-
rinated carboxylic acids, both of 50 mM concentration, were mixed in a desired
proportion to make 100 µL and diluted with ethanol to the final volume of 2 mL.
The cleaned glass slides were deposited inside, and after the addition of 30 mg of
EDC, the entire solution was agitated at room temperature for at least 6 h. The
slides were finally washed with ethanol and methanol and dried by purging with
N2. We will refer to these mixed monolayers in accordance with the mole fraction
of the carboxylated biotin in the carboxylic acid mixture used during the prepara-
tion; i.e., B20 was prepared using a mixture of 0.2 mole fraction of biotin and 0.8
of the fluorinated acid. Note that the actual fraction of the biotin moiety in the
surface monolayer can differ from that because of the possible reactivity differ-
ence for different carboxylic acids. The surface F100 (B0) corresponds to a solely
fluorinated monolayer. Figure 2.1 shows a schematic of the expected monolayer
configuration.
CA Measurements. The CA measurements were carried out using a home-
made apparatus consisting of a microscope connected to a digital camera, a hori-
zontal beam holding the microscope parallel to the surface, a light dispersive plate,
and a three-axis moving platform. Most experiments were conducted without
control of the humidity or temperature; the latter was typically within the range
between 20 and 25 C. Each series of experiments represented by a graph was con-
ducted on the same day to ensure the same humidity. The water-vapor-saturated
environment used in some experiments was achieved by placing a modified glass
substrate on a support inside a glass rectangular cuvette, the bottom of which
was filled with deionized water. When used, this condition almost eliminated
35
Figure 2.1: Representation of the surface modified with a mixed monolayer ofbiotin-LC-LC and perfluoric acid. The biotin moiety must extend at least threebond lengths above the top of the fluorinated monolayer to ensure SA binding.
droplet evaporation (a very slow Kelvin evaporation due to a small droplet size
still took place). Drops of approximately 1 µL were deposited on the surface using
a Hamilton microsyringe. Movies and pictures of the drops’ profiles were recorded
every 2 min in most of the experiments. The images were analyzed using the CA
plug-in (written by Brugnara (19)) in the ImageJ software. All measurements
were done in triplicate. Cleaning of fouled surfaces by sonication (see the text)
was performed in a Branson 1200 ultrasonic cleaner.
36
2.3 Results and discussion
Our motivation was to investigate whether mixed hydrophobic surfaces can be trig-
gered into hydrophilic ones by interaction with biochemical analytes. There are
various possible applications for such a phenomenon, including electrical biosen-
sors. We have demonstrated before [18, 19, 20, 21, 22] that nanoporous mem-
branes, the surface of which is modified by organic monolayers, can be made
responsive to physical and (bio)chemical stimuli. When the surface is modified by
a mixed layer of hydrophobic molecules and hydrophobicity switching triggering
elements, not only the ionic conductance [19, 22] but also the whole solution flow
through the membrane can be switched by the stimulus [21]. Whether or not it is
possible to realize such a switching with biochemical analytes is the driving force
behind this investigation, where SA is used as a representative protein.
SA is a tetrameric protein that has four sites for binding biotin [23]. The
binding of biotin to SA is among the strongest noncovalent interactions known.
It has a very small dissociation constant, estimated between 1 x 10−15[24] and 4 x
10−14 M [25], and a long dissociation time, up to 3 days, making it an essentially
irreversible reaction. The biotin binding site is buried quite deep in SA [26, 27].
Thus, an effective coupling between the two can be achieved only when the linker,
by which biotin is attached to the surface, extends sufficiently enough, at least by
8 Å as measured from the carboxylate carbon of biotin [26, 27]. To ensure this
and, at the same time, provide enough hydrophobicity to the surface, the above-
described procedure for mixed monolayer formation was chosen, where the surface
was first aminated by APTS and then linked to carboxyl-terminated molecules
using the EDC coupling reagent. The LC-LC linker on biotin-LC-LC-COOH is
37
sufficiently long to bind SA effectively [27].
To make the switching specific to the analyte (SA in our case), one needs to
minimize the effect of a well-known phenomenon of passive adsorption of pro-
teins, which occurs even on hydrophobic surfaces [7, 12, 17]. Among the various
options for handling this effect [28], fluorination, i.e., surface modification using
fluorinated molecules, was chosen for its relative simplicity. When compared to
aliphatic surfaces, fluorinated surfaces show lower fouling by proteins, but even
they eventually succumb to fouling at high concentrations of proteins, especially
after prolonged exposures.
The effect of passive adsorption (physisorption) can be evaluated by measuring
the CA of the droplets with different concentrations of a protein (e.g., SA) on the
fully fluorinated surface B0 (F100). The simplest approach is to monitor the free-
standing Sessile droplet shapes in time upon their slow evaporation rather than to
measure the advancing and receding angles. Besides providing more reproducible
data for the receding angle, this approach also allows identification whether there
is any delayed spreading of the droplets due to SA binding to biotin, similar to
what happens with solutions of small amphiphile molecules.
Figure 2.2A demonstrates that for SA concentrations of 100 mg/L (~2 µM) or
higher there is significant nonspecific adsorption of SA to the fluorinated surface
while the solution of a lower concentration, e.g., 10 mg/L or lower, presents very
minimal adsorption and almost matches the behavior of a plain phosphate-buffered
saline (PBS) buffer. At the same time, mixed-monolayer modification, B30, has
visibly changed the receding CA down to SA concentrations of 100 μg/L, as shown
in Figure 2.2B.
Figure 2.3 provides the time snapshots of PBS and SA containing droplets
38
Figure 2.2: CA variations for droplets with SA solutions of different concentrationson B0 (A) and B30 (B) surfaces.
slowly evaporating on the surfaces modified with different solitary and mixed
monolayers, B0 (F100), B25, B50, B75, and B100. The analysis of their CAs,
given in Figure 2.4, illustrates that all surface modifications (except for fully fluo-
rinated F100) have dramatically different evolution of the PBS and SA containing
droplets. The surfaces with a higher content of biotin demonstrate lower initial
CAs, as was expected because of the more hydrophilic character of biotin. The
CA values on each surface are almost identical for the two droplets at first, but
with time, the PBS droplet starts to shrink upon reaching the corresponding re-
ceding CA. The SA-containing solutions, on the other hand, show a continuous
decrease of the CA within this time frame, which correlates well with the receding
angles measured at different times after placement of the droplet on the surface.
Obviously, this behavior is due to the specific interaction between SA in solution
39
and the surface-bound biotin.
Figure 2.3: Time evolution of the slowly evaporating droplets without (left) andwith 10 mg/L SA solutions on the surfaces with different percentages of biotin(fluorocarbon is the remaining component of the surface modification).
As Figure 2.5 illustrates, upon evaporation of the droplet, biotin moieties act
as “anchors” for SA binding and thus pin the contact line to its original position.
When the volume of the droplet shrinks, the resulting CA decreases. With no
biotin on the surface, the droplet decreases in size while maintaining the same
shape as soon the CA reaches the receding angle value.
The specificity of biotin-SA interaction as being responsible for the observed
CA hysteresis can be confirmed by the lack of such CA changes with solutions of
other proteins. To minimize the number of variables and to have a better compar-
40
Figure 2.4: Variation of the CA with time for Sessile drops with PBS and 10 mg/LSA solutions on surfaces with different amounts of biotin moieties (A), solutionswith 10 mg/L free SA and 10 mg/L SA capped with biotin, both on the B30surface (B). The capped SA variation is practically indistinguishable from that ofthe PBS buffer.
ison, we performed experiments with SA whose active sites were “capped” with
biotin. Capping was achieved by the addition of a slightly above the stoichio-
metric amount (5:1) of D-Biotin to the SA solution, thus rendering the protein
inactive to binding with biotin on the surface. Figure 2.4B confirms that the vari-
ation of CA for the capped SA on B30 is practically indistinguishable from that
of the PBS buffer in a dramatic distinction from the uncapped SA. Again, the
distinction emerges as soon as the shrinking due to evaporation droplets reached
their corresponding minimum receding CAs. The solution with capped SA does
not demonstrate hydrophobicity switching and the droplet starts shrinking early
with a large CA, while the uncapped SA binds specifically to the biotin on the
41
Figure 2.5: Illustration of the mechanism of CA variation. The advancing CA isnot affected by SA binding onto biotin because neither biotin nor SA can swayacross the contact line and the thermal oscillations of the contact line are notsufficient either.
surface, lowers its surface tension, and pins the contact line until a much lower
CA is realized.
Alternatively, the CA with a dull PBS buffer (free of any protein) can be used
to study binding of proteins to the same surfaces after their prolonged exposure
to the protein solutions. This approach allows a convenient way of discriminating
between specific binding and passive adsorption (physisorption). The fully fluo-
rinated surface (B0) gets fouled after 30 min of exposure to either SA or BSA
solutions. Because of a longer exposure time, the CA drops from 108 ± 2 before
to 78 ± 6 and 78 ± 7, respectively, for the two proteins (see Figure ) due to their
physisorption. The hydrophobic property fully recovers after 1 min of sonication
in 50% (v/v) methanol. This treatment is mild enough not to denature the pro-
tein, at least not SA. As seen in Figure 2.6, a partially biotinylated B30 surface
experiences a significant drop of the CA after exposure to either SA or BSA solu-
tion. The CA decreases from 100 ± 9 to 54 ± 6 and 71 ± 10 after SA and BSA,
respectively. Remarkably, sonication of the BSA-fouled B30 surface for 1 min in
42
50% methanol completely recovers, while it has an insignificant change, to 56 ±
7, for the SA-treated B30 surface. More harsh conditions of 30 min of sonication
in pure methanol are likely to denature proteins more significantly, as is observed
in the recovery of the CA for the SA-treated B30 surface back to the original value
of greater than 100. Multiple uses of this procedure eventually deteriorate the
surface properties, which is first revealed in a lowering of the receding angle.
Figure 2.6: Variation of the drop with the PBS buffer on B0 (F100) and B30surfaces after different treatments.
43
There are two questions worthy of further discussion. First, nonspecific bind-
ing of proteins (physisorption) occurs on all surfaces, with and without the biotin
ligand, but it is a much slower process. When the kinetics of the CA variation
with evaporation are measured after allowing a SA-containing droplet to soak
onto that surface in saturated vapor, changes in the receding angle are observed
with 10 mg/L SA as well. It requires at least an additional 10 min to observe a
significant effect for that concentration on the F100 surface. During this time, bi-
otinylated surfaces demonstrate a strong binding even with lower concentrations.
In applying this method for sensing SA, one can eliminate the nonspecifically
bound proteins by sonication, as explained above. Whether or not physisorption
is a cooperative effect would require additional studies and probably a more ap-
propriate technique. The second question that motivated this work is about the
contact-line movement as a result of specific protein binding. We observe no such
movement for either specific or nonspecific protein interaction with the surface.
Despite the large CA hysteresis upon SA binding, which exceeds 70 for B50, we
do not observe the contact-line movement outward; i.e., there is no delayed droplet
spreading. Even on the B100 surface, i.e., when only biotin is present on the sur-
face, no movement of the contact line is observed in a 100% humid atmosphere
over a 12 h period with SA concentrations of 10 mg/L or lower. This behavior
is different from that of small amphiphile molecules. For the latter, it has been
established that the process, which primarily determines the spreading of surfac-
tant solutions over hydrophobic substrates, is the transfer of surfactant molecules
onto a bare hydrophobic substrate in front of the moving three-phase contact line
[29]. This process results in a partial hydrophilization of the hydrophobic surface
in front of the drop and determines the delayed spontaneous spreading. Indeed,
44
it is easy to see from eq 1 that the decrease of only γsl and γlv resulting from
the relatively fast adsorption of amphiphiles on solid/liquid and liquid/vapor in-
terfaces cannot explain the switch from hydrophobic (θ > 90) to hydrophilic (θ
< 90) behavior. Obviously, it can be realized only when γsv becomes greater
than γsl, i.e., when γsv increases in the vicinity of the three-phase contact line
as well. Transfer of surfactants from the solution onto the solid/ vapor inter-
face just in front of the drop increases the local free energy, but the total free
energy of the system decreases. The process goes via a relatively high potential
barrier and hence is considerably slower than the adsorption at liquid/solid and
liquid/vapor interfaces; i.e., the time scale for the droplet spreading is defined by
the characteristic time of surfactant transfer from the drop onto the solid/vapor
interface. If the latter is slow, the system can “get stuck” in the metastable state
for a very long time. Large proteins such as SA or BSA cannot follow this route
directly because of the size; the only option left for them to affect the advancing
CA is if the contact line can fluctuate itself, thus exposing SA to the possibility of
binding with biotin on the surface (see Figure 2.5). The alternative version would
be with biotin ahead of the contact line fluctuating in and out of the droplet and
occasionally “fishing-out” SA from an aqueous solution into the dry region. Both
of these options are apparently too much of an uphill process and are not realized
to a sufficient degree.
2.4 Conclusions
Mixed fluorinated surfaces with covalently bound biotin demonstrate smart active
hydrophobicity switching, where specific binding of SA from a low-concentration
45
solution can decrease the CA with water from being greater that 90 down to
less than 60. The effect is clearly visible in the receding CA for solutions of SA,
while the advancing angle remains identical for the buffer and SA solutions, thus
proving that advancement of large protein molecules onto hydrophobic surfaces
ahead of the contact line even via the help of specific interactions with ligands is
a highly unfavorable process.
SA with blocked biotin binding sites and BSA lack active hydrophobicity
switching but do show nonspecific binding by physisorption that can be elimi-
nated by sonication of the exposed surface in a 50% methanol solution for 1 min.
A harsher treatment of 30 min of sonication in pure methanol denatures these
proteins and recovers the hydrophobicity of these surface.
46
Bibliography
[1] Yoshida, M.; Lahann, J. ACSnano 2008, 2, 1101–1107.
[2] Sun, T.; Wang, G.; Feng, L.; Liu, B.; Ma, Y.; Jiang, L.; Zhu, D. Acc. Chem.
Res. 2004, 43, 357–360.
[3] Wang, R.; Hashimoto, K.; Fujishima, A.; Chikuni, M.; Kojima, E.; Kitamura,
A.; Shimohigoshi, M.; Watanabe, T. Nature 1997, 388, 431–432.
[4] Feng, X.; Feng, L.; Jin, M.; Zhai, J.; Jiang, L.; Zhu, D. J. Am. Chem. Soc.
2004, 126, 62–63.
[5] Ichimura, K.; Oh, S.-K.; Nakagawa, M. Science 2000, 288, 1624–1626.
[6] Rosario, R.; Gust, D.; Hayes, M.; Jahnke, F.; Springer, J.; Garcia, A. A.
Langmuir 2002, 18, 8062–8069.
[7] Yoon, J.-Y.; Garrell, R. L. Anal. Chem. 2003, 75, 5097–5102.
[8] Heng, L.; Dong, Y.; Zhai, J.; Tang, B.; Jiang, L. Langmuir 2008, 24, 2157–
2161.
[9] Bain, C. D.; Whitesides, G. M. Langmuir 1989, 5, 1370–1378.
47
[10] Johnson, R. E.; Dettre, R. H. Wetting of Low-Energy Surfaces; Berg, J. C.,
Ed.; Marcel Dekker Inc., 1993; Vol. 46, pp 1–73.
[11] Erbil, H. Y.; McHale, G.; Rowan, S. M.; Newton, M. I. Langmuir 1999, 15,
7378–7385.
[12] D’Andrea, S. C.; Fadeev, A. Y. Langmuir 2006, 22, 3962–3963.
[13] Anzai, J.-i.; Guo, B.; Osa, T. Bioelectrochem. Bioenerg. 1996, 40, 35–40.
[14] Huang, T. T.; Sturgis, J.; Gomez, R.; Geng, T.; Bashir, R.; Bhunia, A. K.;
Robinson, J. P.; Ladisch, M. R. Biotechnol. Bioeng. 2002, 81, 618–624.
[15] Krasnoslobodtsev, A. V.; Smirnov, S. N. Langmuir 2002, 18, 3181–3184.
[16] Moon, J. H.; Kim, J. H.; Kim, K.-j.; Kang, T.-H.; Kim, B.; Kim, C.-H.;
Hahn, J. H.; Park, J. W. Langmuir 1997, 13, 4305–4310.
[17] Lee, S.-H.; Lee, C.-S.; Shin, D.-S.; Kim, B.-G.; Lee, Y.-S.; Kim, Y.-K. Sens.
Actuators B 2004, 99, 623–632.
[18] Vlassiouk, I.; Krasnoslobodtsev, A.; Smirnov, S.; Germann, M. Langmuir
2004, 20, 9913–9915.
[19] Vlassiouk, I.; Park, C.-D.; Vail, S. A.; Gust, D.; Smirnov, S. Nano Lett. 2006,
6, 1013–1017.
[20] Takmakov, P.; Vlassiouk, I.; Smirnov, S. Anal. Bioanal.Chem. 2006, 385,
954–958.
[21] Takmakov, P.; Vlassiouk, I.; Smirnov, S. Analyst 2006, 131, 1248–1253.
48
[22] Vlassiouk, I.; Takmakov, P.; Smirnov, S. Langmuir 2005, 21, 4776–4778.
[23] Weber, P. C.; Ohlendorf, D. H.; Wendoloski, J. J.; Salemme, F. R. Science
1989, 243, 85–88.
[24] Green, N. M. Biochem. J. 1963, 89, 585–591.
[25] Green, N. M. Methods Enzymol. 1990, 184, 51–67.
[26] Green, N. M.; Konieczny, L.; Toms, E. J.; Valentine, R. C. Biochem. J. 1971,
125, 781–791.
[27] Wilbur, D. S.; Pathare, P. M.; Hamlin, D. K.; Weerawarna, S. A. Bioconjugate
Chem. 1997, 8, 819–832.
[28] Zhang, G.-J.; Tanii, T.; Zako, T.; Hosaka, T.; Miyake, T.; Kanari, Y.; Fu-
natsu, T.; Ohdomari, I. Small 2005, 1, 833–837.
[29] Starov, V. M. Adv. Colloid Interface Sci. 2004, 111, 3–27.
49
Chapter 3
Wetting of hydrophobic nanopores
induced by pressure and amphiphile
adsorption
3.1 Introduction
Investigation and utilization of hydrophobic surfaces for various applications has
gained a new momentum recently. Hydrophobicity is a fundamental property
that controls interactions between nonpolar substances and water. These interac-
tions in turn are responsible for numerous physical and biophysical phenomena.
Hydrophobicity has been studied extensively, but many aspects are still not well
understood. Strong attraction between water molecules due to hydrogen bonding
makes their interaction with nonpolar substances unfavorable. Poor wetting of a
50
hydrophobic surface by water can be observed experimentally as a large contact
angle between a water droplet and the surface.
Recent development of nanometer scale systems and their applications in the
biological, chemical, and physical sciences increasingly emphasizes the importance
of interfaces: the smaller the object size, the greater its surface-to-volume ratio.
What could have been an insignificant annoyance in the macroscopic and even
microscopic systems can no longer be ignored on the nanoscale. Behavior of a
liquid near the solid surface is substantially different from that in the bulk and
is affected by confinement of liquid in nanosized voids. Water at a hydrophilic
surface was predicted by computer simulation to have a higher density than in
the bulk,[1] while near hydrophobic surfaces, a thin layer of low-density water is
expected.[1, 2, 3, 4, 5] Another striking theoretical prediction is that water confined
between two hydrophobic surfaces or in a hydrophobic pore is supposed to sponta-
neously evaporate when the size of the pore is sufficiently small.[3, 4, 5] Since the
phenomenon is important fundamentally as well as to various applications such
as electrowetting[6] and sensors,[7] it is essential to identify the conditions when
spontaneous evaporation can occur.
Similarly to the interior of biological membranes, artificial membranes can be
made hydrophobic and impermeable to water. Moreover, the hydrophobic surface
can be also made responsive to various stimuli that switch its surface tension[7,
8, 9] and turn the membrane into an artificial mediator for transport of ions and
other species. In living organisms, such transfer across the membrane is selective
and controls a variety of metabolic and signaling purposes, such as nerve impulses
generated by the controlled release of ions across the membranes. Mimicking
biological channels using synthetic nanopores is a challenging scientific problem
51
with possible applications in medicine, materials science, fuel cells, analytical
chemistry, and sensors.
In the present chapter the induced wetting of hydrophobic membranes will
be investigated. Pressure and surfactant solutions will be used to identify the
conditions at which the wetting and dewetting of hydrophobic pores take place.
The possible application of these phenomena in the design of a drug delivery
system (DDS) will be explored.
3.2 Experimental
All chemicals used in this work were obtained from Sigma- Aldrich (St Louis, MO)
and were used as received.
3.2.1 Pressure induced wetting
Three types of free-standing nanoporous alumina membranes were used for this ex-
periments: commercial “Anodisc” from Whatman (Whatman, Florham Park, NJ)
with the nominal 0.2 or 0.02 m diameter pores (60 µm thick) and the homemade
60 µm thick membranes with 70 nm diameter pores. The latter were prepared
using a previously described procedure that consists of anodization of cleaned
and electropolished Al foil (99.9%, Alfa Aesar, Ward Hill, MA) in oxalic acid at
5 C and 40 V followed by dissolution of Al substrate in CuCl2 and pore open-
ing/ widening in 1 M phosphoric acid at room temperature for 30 min.[10] The
membranes were rendered highly hydrophobic using four different modifications
(shown in Figure 3.1), as previously described.[11] In the first scheme, the mem-
52
brane surface was directly silanized with hexadecyltriethoxysilane from toluene
solution (overnight); in the second scheme, the membrane surface was also di-
rectly silanized with 1H,1H,2H,2Hperfluorooctyltrichlorosilane from toluene solu-
tion, also overnight. We will label such membranes as SiH16 and SiH2F6, respec-
tively (see Figure 3.1). In both cases, the treatment was concluded by thorough
washing in ethanol and overnight baking at 120 C. In the other two schemes, the
membrane was first aminated using 3-aminopropyl trimethoxysilane and then car-
boxylic acid ends of either decanoic acid or 2H,2H,3H,3Hperfluoroundecanoic acid
were coupled to the surface-bound amino groups using EDC coupling reagent N-[3-
(dimethylamino)propyl]-N-ethylcarbodiimide.[11] We will label such membranes
as SiNH9 and SiNH2F8, respectively (see Figure 3.1). The density of bound to
the membrane surface monolayers was monitored by IR absorbance. The data will
be presented only for SiH16 and SiNH2F8 modifications because others showed
inferior electrical resistance properties compared to these two.[11] The electrical
impedance due to ionic conductance through the membranes was measured in a
homemade two-electrode electrochemical cell[11] using a CH 604B electrochemi-
cal workstation (from CH Instruments Inc., Austin, TX). The membrane forms a
barrier between the two halves of the cell containing degassed 1.0 M potassium
chloride in water at pH 7. Two Ag/AgCl electrodes were in close proximity to
the membrane, and a low voltage (5 mV) was employed for AC impedance mea-
surements. The open area of the membranes was 0.25 cm2 that resulted in the
resistance of ∼ 17Ω with unmodified membrane in 1.0 M KCl; that is, this is the
minimum measurable resistance or the “cell resistance” under the experimental
conditions. The impedance variation with pressure was monitored at 100 Hz. At
this frequency, the capacitive contribution to the impedance is minimal. In order
53
to maintain reproducible environment, the cell was first degassed using a water
pump and a similarly degassed solution was introduced independently into the
both cell compartments by suction. Two 1/8 in. o.d. Tygon tubes were left con-
nected to each side of the cell and filled with electrolyte up to the length of 50
cm. Since gas diffusion through such a long distance is incredibly slow, apply-
ing pressure to this assembly hydrostatically with a gas precludes its penetration
into the membrane, leaving only electrolyte and water vapor. This assembly was
placed inside a homemade stainless steel high-pressure chamber with electrical
feed-through contacts for connecting the electrodes to the workstation. The pres-
sure inside the chamber was supplied by nitrogen gas from a tank connected via a
high-pressure manometer. The maximum controlled pressure was limited by the
pressure reductor not to exceed 23 bar above the atmospheric one.
3.2.2 Amphiphile induced wetting
SiH16 membranes were placed in a U tube cell as described previously.[12] Impedance
measurements were performed using Ag/AgCl electrodes. Two kind of amphiphiles
were used in this experiment: sodium dodecyl sulphate (SDS) and dodecyl tert-
buty ammonium bromide (DTAB). Solutions with concentrations between 0.2 and
1.5 CMC of the amphiphiles were place into contact with the membrane on both
sides of the U tube and the impedance followed with time.
Glass capillaries were prepared on a standard puller and modified hydropho-
bically with CH16, as described above.
Giant unilamellar vesicles (GUVs) were prepared by electroformation as de-
scribed elsewhere.[13] Briefly, two platinum wires (0.5 mm) placed 3 mm apart in a
54
Teflon chamber wires were connected to a wave generator producing at 1.5 V pulses
at 10 Hz . Solution of 1 mg of 1-palmitoyl-2-oleoyl-sn-glycero-3-phosphocholine
(POPC) in 1 mL solution of 9:1 CHCl3/MeOH was evenly spread over the wires
and dried. After introduction of DI water in the chamber to cover the wires, the
amplitude of voltage pulses was increased to 5 V and formation of GUVs was
observed after approximately 10 min.
3.3 Results and Analysis
3.3.1 Pressure induced wetting
The membranes were modified with aliphatic and fluorinated silanes, as shown
in Figure 3.1, to make them highly hydrophobic. An apparent contact angle
for a small water drop (in the sessile technique) is greater than 140 for all of
them. Despite no electrolyte intrusion into the membrane at ambient external
pressure, membranes have very low but measurable conductance that depends on
the modifier and not on the electrolyte concentration or its pH.[11] Membranes
with these hydrophobic modifiers show very high resistance, greater than 1 MΩ,
and remain superhydrophobic for indistinguishably long time if left in electrolyte.
Their resistance varies in the following order: SiH16 > SiNH2F8 > SiNH9, which
was discussed previously,[11] and is due to surface conductance from ionizable
groups below the hydrophobic monolayer. Residual hydroxyls on alumina surfaces
and on silanes as well as amines and amides of the linkers contribute to that
surface conductance. Their hydration and the resulting conductance were found
to slightly increase over time, which is also related to hysteresis of the contact
55
angle. Autoionization of residual water bound on the surface contributes to the
effect, as well.
Figure 3.1: Four types of hydrophobic surface modifications used in thisstudy with their labels. The first two (SiH16 and SiH2F6) were ob-tained using toluene solutions of hexadecyltrimethoxysilane and 1H,1H,2H,2H-perfluorooctyltrichlorosilane, respectively. The last two (SiNH9 and SiNH2F8)were obtained in two steps: first “amination” using aminopropyl trimethoxysi-lane in toluene, and then reaction with either decanoic acid or fluoroundecanoicacid using EDC coupling reagent in ethanol. For simplicity, Si atoms are drawnconnected to the surface via a single Si-O bond; the remaining two bonds are pre-sented as hydroxyls, but at high densities, most neighboring silanes form lateralSi-O-Si bonds.
The fact that the resistance is not infinite allows one to use it for monitoring
the extent of water/electrolyte intrusion into the pores. In a simple model that we
previously discussed,[11] the resistance of a single pore with constant diameter,
D, and length L, is given by:
Rpore = Rs
L
πD(3.1)
where Rs is the sheet resistance of the hydrophobic monolayer. If the pore is
partially filled with electrolyte, the resistance of that portion is insignificant (by
56
many orders of magnitude) as compared to the surface wall resistance of the “dry”
portion. Thus one can still use eq 3.1 in this case with L referring to the length
of the dry portion.
The equilibrium of forces at a water-gas interface in a nanopore can be de-
scribed by balancing the pressure difference with the capillary force:
∆P = Pext − Pin = −γ(1
r1+
1
r2) (3.2)
where r1 and r2 are the curvatures of meniscus at the pore mouth, Pext is the
external pressure on the outside of the membrane, and Pin is the pressure inside
the pore. By virtue of how the cell is prepared in our experiments, Pin always
equals the vapor pressure of water, Pin= Pvap (~23 Torr at 25 C).
For a hydrophilic surface ( θ <90), the equilibrium in eq 3.2 cannot be sus-
tained at any external pressure (since P ext≥P vap) and water fills up the pore. For
a hydrophobic surface ( θ >90), the pore remains dry until the external pressure
reaches its critical value, which is dependent on γ and the pore cross section. The
latter can be mimicked as an ellipse with the two diameters, D1 and D2, which
become identical for a perfectly cylindrical pore D = D1 = D2. The critical
pressure for nanometer-sized pores is usually much higher than the water vapor
pressure, which allows neglecting P in in eq 3.2. The pressure, Po, at which water
is capable of intruding into the pores is quite large even for the largest pores in
our study.[14] Indeed, for uniform diameter D=0.2 µm and hydrophobic modifi-
cation of a modest advancing contact angle of θa~ 105, the critical pressure is Po
3.8 bar. Simple electrolytes, such as KCl used here, have minimal effect on the
surface tension. For example 1.0 M KCl solution has only an insignificant increase
57
from 72 to 74 mN/m. Thus, the outcome of measurements with 1.0 M KCl can
be presumed to hold almost identically to such with pure water.
The pores in the membrane are not identical; they vary in the quality of
coverage and the pore diameter that fluctuates not only between the pores but
likely within the pore as well, as is illustrated by Figure 3.2. From SEM images,
“0.2 m” Whatman membranes have an average 213 nm diameter on one side and
114 nm on the other side. From porometry measurements, one side of these
membranes has pore diameters, D, ranging from 122 to 256 nm (with 143 nm at
50% level) and another side from 130 to 178 nm (with 152 nm at 50% level). As
a result, the water intrusion into the membrane happens not at a single pressure
but over a range of pressures.
Figure 3.3 demonstrates the variation of the resistance as a function of pressure
for the two types of commercial membranes, 0.2 and 0.02 µm, with two types of hy-
drophobic surface modifications. Comparison of 0.2 m membranes with SiH16 and
SiNH2F8 modifications reveals that both demonstrate a broad range of pressures
when water intrudes into the membrane and causes the resistance to decline. The
ratio of the highest to the lowest pressure is roughly a factor of 2, in agreement
with the pore diameter distribution. At the same time, the SiNH2F8-modified
membrane requires almost twice as much pressure to become totally open, more
than 12 bar versus just 7 bar for the SiH16 membrane. These values are in a
good agreement with the contact angles on flat surfaces of θa~105 and 120 for
surfaces modified with SiH16 and SiNH2F8, respectively, and the pore diameter
ca. 120 nm is the smallest pore diameter for such membranes. Incidentally, the
low limit pressures in both cases are roughly half the highest value, in agreement
with the higher end values in the pore diameter distribution (ca. 250 nm).
58
Figure 3.2: Illustration of different pore morphologies in a membrane with hy-drophobic modification and their response to hydrostatic pressure of water (blue)that is insufficient to intrude into pores of a too small diameter (I) but pene-trates through large pores (II). Pores with variable diameter (III and IV) haveincomplete water penetration. Type III has uncontrolled (unintentional) diame-ter variation, while type IV is often realized for alumina membranes that are not“perfectly open” after their construction using anodization of Al. Type V repre-sents the geometry of commercial 0.02 μm membranes from Whatman, where the0.02 μm portion is only 1 μm deep on one side of the membrane and its remaining59 μm of thickness has pores with the nominal 0.2 μm diameter.
After reaching the critical pressure, when the membrane resistance equals the
value corresponding to it being filled with electrolyte (ca. 17Ω ), the resistance
remains this low even after the pressure is reduced to the atmospheric one. This
is a relatively well understood situation,[5] where spontaneous dewetting of hy-
drophobically modified nanopores is kinetically unfeasible despite the significant
thermodynamic advantage. A high activation barrier makes such a transition ki-
netically impossible for the nanopore diameters greater than D > 10 nm. The
transition state requires formation of a bubble inside the pore, which defines a
large activation barrier.
59
Figure 3.3: Variation of the impedance at 100 Hz with hydrostatic pressure forcommercial 0.2 and 0.02 μm membranes modified with SiNH2F8 and SiH16. Ar-rows indicate the points when pressure was increased (up) to a designated value inbars and decreased (down) to the atmospheric pressure, respectively. (A) 0.2 μmmembrane modified with SiH16. Pressure steps are 0.7 bar starting with 4.3 bar.(B) 0.02 μm membrane modified with SiH16. Pressure steps are 0.7 bar (from 2.9to 8.5 bar) and 1.5 bar (from 8.5 to 16.0 bar). Every step of the pressure increaseis followed by discharge to the atmospheric pressure (0.9 bar). (C,D) 0.2 and 0.02μm membranes modified with SiNH2F8.
The situation at intermediate pressures is less obvious. When the applied pres-
sure is restored down to the atmospheric one, the membrane resistance partially
recovers. The extent of this recovery is negligible with SiH16 modifier (Figure
3.3A), but it can reach as much as 15% with SiNH2F8 (Figure 3.3B) at pressures
near 89 bar, which correspond to the maximum in the pore distribution by diam-
eters (150 nm). Distribution of pore diameters causes water intrusion at different
pressures, and the type of this distribution affects the manner in how it proceeds.
Uneven pore diameter also affects the recovery after the external pressure drops to
the atmospheric one. If the pores are perfect cylinders but of different diameters,
60
then they should be entirely filled with water as soon as the critical pressure is
reached. Large pores are filled first and the smallest pores last, as illustrated by
cases I and II in Figure 3.2. Pores entirely filled with water do not dry out after
the pressure is dropped back to the atmospheric one due to the above mentioned
large activation barrier.
A pore with variable diameter (such as in the cases III and IV of Figure 3.2),
say from Dmin to Dmax, may end up at a pressure that exceeds the critical Po
for Dmax but too low for penetrating into the narrowing of Dmin. The electrical
resistance of the pore at such pressure would significantly decrease in accordance
with a shorter vapor gap but would still be large since the conductance through
the gap portion is very small. This vapor gap helps in spontaneous pore dewetting
after releasing the pressure. Water is pushed out of the pore by the surface tension
at the interface with vapor and the walls as long as the contact angle with the
surface stays above 90. For high-quality surface modifications, the contact angle
remains almost unaffected by the history of its exposure, but for rough and/or
inhomogeneous surfaces, the hysteresis of the contact angle (the difference between
the advancing, a, and receding, r, angles) can be as high as 10-15. For aliphatic
surface modifiers, it can cause the receding contact angle to fall below 90, that
is, when spontaneous dewetting would not be possible for cylindrical pores. The
receding contact angle for fluorinated surfaces, even with large hysteresis, remains
above 90, which can explain the difference in behavior of SiNH2F8 and SiH16 0.2
µm membranes. Inhomogeneity of hydrophobic surface modification broadens the
distribution of critical pressures Po and the extent of recovery after the pressure
release.
Besides the narrowing in the pores, such as in the case III, other imperfections
61
can also appear during membrane preparation. The pores grown by anodization
of metallic Al foil are sealed on one side and need additional chemical (with
phosphoric acid), electrochemical (changing the anodization voltage at the end),
mechanical (polishing), or a combination of such means to eliminate the oxide
barrier on one side, that is, to make them opened. Depending on the process, the
resulting diameters can vary significantly, and for Whatman 0.2 m membranes,
it was shown to have different distributions of diameters on the two sides of the
membranes.
To confirm the interpretation and further illustrate the importance of bubbles
for spontaneous dewetting of pores filled with water, we repeated experiments with
the homemade membranes of the nominal pore diameter 70 nm and with 0.02 m
membranes from Whatman (Figures 3.3B,D and 3.4). Smaller pore diameters
translate into larger pressures required for water intrusion, and a narrower distri-
bution by diameters should similarly correspond to a narrower range of critical
pressures.
Homemade membranes have been prepared under much better controlled con-
ditions, which results in a more circular shape of the pores and a narrower distribu-
tion of their diameters, around 70 nm. However, the way the pores were “opened”
(by the phosphoric acid treatment) could likely cause their uneven widening from
the original 50 nm diameter (before the treatment) to 70 nm.
The smaller pore diameters of 70 nm indeed result in a larger pressure for com-
plete water intrusion, almost by a factor of 2 (see Figure 3.4), in agreement with
the factor of 2 in the diameter decrease as compared with the 0.2 µm membrane.
The narrower distribution of pore diameters in the homemade 70 nm membrane
modified with SiNH2F8 also illustrates a significant narrowing of the critical pres-
62
Figure 3.4: Pressure dependence of the conductance for hydrophobic membranesof Figure 3.3 and for the fluorinated 70 nm pore diameter membrane. Two pointsfor each pressure represent values at that pressure (filled) and after its releasingback to the atmospheric one (empty). Squares illustrate the 0.2 μm membrane,triangles the 0.02 μm membrane, and circles 70 nm. The bottom inset shows thesame graph in linear scale. The top inset illustrates the labeling of resistancesusing a portion of the graph in of Figure 3D: when the pressure is applied (10.6bar in this case), Ron, and upon recovery to the atmospheric pressure, Roff .
sure range, as compared with 0.2 µm membranes. As seen in the logarithmic plot
of the inset of Figure 3.4, less than 0.1% of the pores are filled with water before
the last step of pressure increase, which is a remarkably narrow distribution as
compared to that of the 0.2 µm membrane. Additional breadth to both distribu-
tions is brought by the nonuniformity of surface modification, which results in a
distribution of surface energies (contact angles). Note that the resistance recovery
upon pressure dropping is similarly smaller for the homemade membrane.
The so-called 0.02 m membranes from Whatman actually have the same 150
nm (0.2 µm nominal) diameter throughout 59 m of their total 60 m thickness,
and only the remaining 1 µm on one side has the nominal diameter 0.02 µm, as
was previously described and is sketched as the case V in Figure 3.2.[15] It is
63
not as important in this case that the 0.02 µm side has a broad distribution of
the pore diameters with the average diameter exceeding 20 nm, but it is essential
that these pores are noticeably smaller and connected with 0.2 µm pores. The
difference in diameters assures the existence of a wide range of pressures when
water can intrude only into the 0.2 µm part and leave the 0.02 µm side dry. It
also means that the condition when pores should be able to spontaneously dry
out upon release of excess pressure can be satisfied.
Figures 3.3 and 3.4 confirm that the 0.02 µm membrane with SiNH2F8 mod-
ification does require very high pressure for complete water intrusion. Even the
highest experimentally available pressure of 23 bar is not sufficient to achieve to-
tal wetting of the membrane. A noticeable change of resistances is observed in a
broader range of pressures, as expected for a much broader distribution of pore di-
ameters. More noticeable is significantly improved resistance recovery, especially
in the range of pressures below 11 bar, where it reaches up to 70%. Since now the
pore diameter variation resembles that of the case V in Figure 3.2, water does not
occupy the pores completely upon intrusion, and the remaining bubbles ensure
the expulsion of water (i.e., spontaneous dewetting).
Even in these 0.02 µm membranes, the resistance recovery is never 100% be-
cause of a number of reasons. First, the contact angle hysteresis is always present,
as discussed above: the receding angle can be smaller than the advancing one.
Some pores with non-uniform diameter and varying quality of the surface mod-
ification do not have sufficient surface tension to effectively expel water. It is
unlikely that this effect contributes significantly because of a dramatic change in
the pressure between the intrusion and extrusion events and almost an order of
magnitude difference in the pore diameters. Second, a hysteresis of surface con-
64
ductance is similarly present: the sheet resistance, Rs, in eq 3.1 strongly depends
on the density of residual ionizable groups on the surface beneath the hydrophobic
layer and was shown to be affected by exposure to water.[11] Third, the processes
of water intrusion and expulsion are not isothermal under our experimental condi-
tions. The membrane resistance in both processes approaches a saturation value
very slowly (Figure 3.3). Water intrusion is an exothermal process because of
the work done against the surface tension and condensation of excess water va-
por, but more significantly, pressurization of the vessel by gas is also increasing
temperature in the chamber. The temperature change needs a significant time to
equilibrate with surroundings, which would be seen as a slow component in the
resistance decline following the initial sharp drop. The duration of this slow com-
ponent obviously depends on the change in temperature and the geometry of the
cell and the pressurizing chamber. Similarly, the dewetting process is endothermic
and is accompanied by the local temperature decline, but the temperature drop
due to the nitrogen gas expulsion from the pressure chamber leads to a stronger
cooling. As a result, the resistance recovery via warming up is slow, as well.
Phenomena similar to those of spontaneous dewetting discussed here are also
of importance in designing superhydrophobic/oleophobic surfaces. Combination
of the effects of microscopic pockets of air trapped beneath the liquid droplets
and the texture of specially engineered surfaces can provide high contact angles
with low hysteresis for liquids with greatly varying surface tension.[16, 17] In such
engineered textures, it is similarly important to have incomplete surface wetting
to efficiently support metastable composite solid/liquid/air interfaces.
65
3.3.2 Amphiphile induced wetting
Steinle et al. introduced the idea of using hydrophobic membranes as sensors.[12]
Since such membranes remain dry in contact with water, the electrical resistance
due to ionic movement through the pores is very high. If a solution of amphiphile
wets the pores, the resistance drops thus offering a mechanism of sensing this
amphiphile. For example, Steinle et al. claimed that 1 µM solution of dodecyl-
benzene sulfonate (DBS) opens a SiH18 modified alumina membrane similar to
the ones used in the present work.
We were unable to reproduce their results but found instead that amphiphilic
molecules wet the hydrophobic membranes at concentrations significantly higher
and close to the critical micelle concentration (cmc). For example, Figure 3.5
shows the variation of impedance for hydrohobic SiH16 membranes in response
to the amphiphile, sodium dodecyl sulfate (SDS,an anionic surfactant). Measure-
ments were done in salt concentration of 0.1 M NaCl.
Figure 3.5: Variation of the ionic impedance through the hydrophobically modifiedalumina membrane (hexadecyl silane) as a function of amphiphile concentrations:
Increasing concentration of either surfactant eventually leads to opening of the
66
membrane pores to water intrusion seen as a significant drop of the impedance.
The concentrations required to induce wetting are much higher than those of
Steinle et al., as is illustrated in Figure 3.6 by direct comparison. Also noticeable
in the figure is a much higher resistance (~109 Ω) of our membranes, which is an
indicator of a better quality of the hydrophobic surface modification.[11]
Figure 3.6: Variation of ionic impedance at different concentration of amphiphilesthrough a hydrophic nanoporous membrane. Comparison between the results ofMartin and the data obtained in this work.
We had to conclude that the discrepancy is due to a poor quality of the mod-
ification in Steinle’s membranes. It formally agrees with a poorer performance of
imperfect modifications in pressure induced wetting but wetting here is different.
The amphiphiles alter the surface tension at the interfaces and lower the contact
angle as a result. If the contact angle gets lower than 90o, the critical pressure for
intrusion drops to zero and and lets water in at ambient pressure. This is demon-
strated in figure 3.7 where it can be observed that for solutions of amphiphile with
67
CA < 90o there is a drop in the resistance of the membrane of almost 4 orders of
magnitude.
Figure 3.7: Comparison between the ionic impedance through membrane (A) andthe contact angle on a flat surface (B) both hydrophobically modified by hexadecylsilane (C16)
On one side, the monolayers that have lower contact angle with water need
less of its drop to get below 90o. On the other side, such monolayers are not
well packed and should better accommodate amphihiles into the voids between
the hydrophobic tails leading to a greater slope in the concentration dependence
of the contact angle. Instead of investigating the effect of the quality of surface
modification, we have focused on more reproducible ’perfect’ modifications and
how the contact angle on such surfaces changes with the amphohile concentration.
The detailed mechanism of such wetting of hydrophobic surfaces by surfac-
68
tants is still a matter of debate.[18] Two mechanisms have been suggested that
can account for change in the contact angle from above 90o to below that and
thus allow for movement of the contact line. (a) In the first model, adsorption
of a surfactant happens only at the LV and LS interfaces and changes the cor-
responding surface tensions. In this case, spreading is controlled by surfactant
diffusion to these surfaces.[19] (b) In the second model, surfactant adsorption in
front of the triple contact line is viewed as the limiting factor. This adsorption
increases gSV making it more hydrophilic and causes water to spread.[20] There
are arguments in favor and against of both mechanisms which demonstrates a
complicated nature of the phenomenon. [21, 22]
Wetting of large (∼ 100µm ) hydrophobic capillaries can be fitted well by both
models either mechanism. [23, 24, 25] Experimental results in large capillaries
(∼ 100µm) are well fitted by both models. The movement of the contact line is
found to be a function of the square root of time which correspond to a diffusion
controlled process. Although our curves show a clear dependance of wetting with
time it is not possible to adjust our data to a preestablished model since our
experimental conditions do not guarantee some of the asumptions such as the
constant concentration at the entrance of the pore.
Realization that switching off the hydrophobic appearance by amphiphiles re-
quires their significant concentration close cmc brought us to an idea of using
hydrophobic pores in drug delivery.[26] In this new approach, hydrophobic nan-
pores loaded with drug remain closed in aqueous solutions but open up only in
contact with amphiphilic cell membranes.
Phospholipids and other amphiphiles present in biological membranes have
extremely low cmc and thus are almost exclusively assembled in the bilayer mem-
69
brane. They can trigger the hydrophobicity switching of the hydrophobic interior
pore walls of a nanocontainer but it requires for the bilayer membrane to be in
contact with the pore entrance, as sketched in figure 3.8.
Figure 3.8: Wetting of hydrophobic pores by amphiphiles with ‘high’ critical mi-celle concentration, cmc, proceeds via amphiphile partitioning at the meniscusfrom which they decorate the nanopore walls (A→D). Assembling of amphiphileswith low cmc, (e.g. long phospholipids) at the nanopore walls can proceed onlyvia slow diffusion of micelles and vesicles (B). Direct contact with a bilayer isdifferent from both (C→E) and is experimentally illustrated in Figure 3.9.
The process can be mimicked with giant unilamellar vesicles (GUV) prepared
by electro formation of POPC phospholipids.[13] Figure 3.9 presents the snapshots
of a GUV interacting with a submicron glass capillary, the internal surface of
which is hydrophobically modified with CH16. In these phase contrast images,
the capillary is not loaded with a dye and appears dark at first because it is dry. It
remains indefinitely dry in the proximity of GUV and even if the capillary pokes
through its phospholipid membrane. The situation changes dramatically when
the tip opening touches the membrane. If done correctly, in a matter of minutes,
the phospholipids slide in and climb up the interior walls along with water that
70
brightens the tip.
Figure 3.9: Time lapses (~5 min total) of phase contrast images for hydropho-bically modified submicron glass pipette (green arrow) and a giant unilamellarvesicle (GUV) produced by electroforming of POPC (yellow arrow). The pipetteremains dry in water indefinitely but when its opening touches the phospholipidsof the GUV membrane, the latter decorate the pipette’s walls and allow water tointrude.
GUVs represent a convenient model of cellular membranes that are ‘free’ of
membrane proteins and/or intracellular machinery, i.e., where the endocytosis
mechanism for drug internalization cannot be utilized. The hydrophobicity switch-
ing mechanism, on the other hand, does not require endocytosis and should ef-
ficiently translocate any cargo from the hydrophobic nanopores through the cell
membrane.[26] The phospholipids from the bilayer membrane, when in contact
with the pore entrance, decorate the pore walls to make an opening for cargo
release directly from inside the pores into cytosol, as in figure 3.8E.
71
3.4 Conclusions
We experimentally confirm that long hydrophobic nanopores allow water intrusion
under a sufficiently high hydrostatic pressure, the critical value of which depends
on the pore diameter and the type/quality of the hydrophobic modification. At
the same time, restoring the pressure to the atmospheric one results in sponta-
neous dewetting only when a bubble of vapor is left inside the pore. Such bubbles
appear at the regions of narrowing cross section and/or varying quality of the
hydrophobic modification and thus can be engineered to control water expulsion.
The ionic resistance through the membranes correspondingly demonstrates dra-
matic changes accompanying these events of electrolyte entering and leaving the
pores. The total resistance change spans in excess of 6 orders of magnitude. Re-
covery of the resistance to the original high value is always less than 100%, which,
in addition to the mentioned effects, is due to hysteresis in the conductance of
hydrophobic walls after wetting and drying.
Wetting of hydrophobic nanopores can be induced also by surfactants. In-
trusion of solution into the nanopores takes place only when the concentration
of amphiphile is high enough to decrease the contact angle below 90o. For am-
phiphilies with moderate cmc it happens by assembly of monomers at the contact
line and requires their concentration close to cmc. Hydrophobic nanopores used
during our experiments were able to withstand solutions with concentrations of
the amphiphile SDS of the order of 100 µM without wetting, almost two order of
magnitude higher than previous similar work.[12]
Amphiphiles with extremely low cmc, such as phospholipids do not have suf-
ficient contration of monomers to perform the wetting by any other approach
72
but only in contact with hydrophobic pores. We have demonstrated using GUV
made of POPC that hydrophobic capillary gets wetted by POPC only when in
contact with GUV membrane. This supports the proposed new concept of using
hydrophobic pores as a cell transfecting delivery system, the decoration of the
hydrophobic interior of the pore surface by the phospholipids of cell membrane
induces its wetting and should release any cargo loaded into the pores directly
into the cytosol. Since endocytosis is not required in this case, it is anticipated
that innate cytotoxicity of the delivery system should be minimal.
73
Bibliography
[1] Lum, K.; Chandler, D.; Weeks, J. D. J. Phys. Chem. B 1999, 103, 4570–
4577.
[2] Helmy, R.; Kazakevich, Y.; Ni, C.; Fadeev, A. Y. J. Am. Chem. Soc. 2005,
127, 12446–12447.
[3] Lum, K.; Chandler, D. Int. J. Thermophys. 1998, 19, 845–855.
[4] Lum, K.; Luzar, A. Phys. Rev. E 1997, 56, 6283–6283.
[5] Luzar, A. J. Phys. Chem. B 2004, 108, 19859–19866.
[6] Krupenkin, T. N.; Taylor, J. A.; Schneider, T. M.; Yang, S. Langmuir 2004,
20, 3824–3827.
[7] Vlassiouk, I.; Smirnov, S. Biosensing with Nanopores; Merkoci, A., Ed.; Wi-
ley, 2009; Vol. 0, pp 459–490.
[8] Vlassiouk, I.; Park, C.-D.; Vail, S. A.; Gust, D.; Smirnov, S. Nano Lett. 2006,
6, 1013–1017.
[9] Rios, F.; Smirnov, S. ACS Appl. Mater. Interfaces 2009, 1, 768–744.
74
[10] Takmakov, P.; Vlassiouk, I.; Smirnov, S. Anal. Bioanal.Chem. 2006, 385,
954–958.
[11] Vlassiouk, I.; Rios, F.; Vail, S. A.; Gust, D.; Smirnov, S. Langmuir 2007,
23, 7784–7792.
[12] Steinle, E. D.; Mitchell, D. T.; Wirtz, M.; Lee, S. B.; Young, V. Y.; Martin,
C. R. Anal. Chem. 2002, 74, 2416–2422.
[13] MENGER, F. M.; ANGELOVA, M. I. Acc. Chem. Res. 1998, 71, 789–797.
[14] WASHBURN, E. W. Phys. Rev. 1921, 17, 273–283.
[15] Vlassiouk, I.; Takmakov, P.; Smirnov, S. Langmuir 2005, 21, 4776–4778.
[16] Tuteja, A.; Choi, W.; Mabry, J. M.; McKinley, G. H.; Cohen, R. E. PNAS
2008, 105, 18200–18205.
[17] Tuteja, A.; Choi, W.; Ma, M.; Mabry, J. M.; Mazzella, S. A.; Rutledge, G. C.;
McKinley, G. H.; Cohen, R. E. Science 2007, 318, 1618–1622.
[18] Lee, K.; Ivanova, N.; Starov, V.; Hilal, N.; Dutschk, V. Adv. Colloid Interface
Sci. 2008, 144, 54–65.
[19] Bahr, M. v.; Tiberg, F.; Yaminsky, V. Colloid Surface A 2001, 193, 85–96.
[20] Starov, V. M.; Kosvintsev, S. R.; Velarde, M. G. J. Colloid Interface Sci.
2000, 227, 185–190.
[21] Dutschk, V.; Sabbatovskiy, K. G.; Stolz, M.; Grundke, K.; Rudoy, V. M. J.
Colloid Interface Sci. 264, 267, 456–462.
75
[22] Milne, A. J. B.; Amirfazli, A. Langmuir 2010, 26, 4668–4674.
[23] Starov, V. M. J. Colloid Interface Sci. 2004, 270, 180–186.
[24] Bain, C. D. Phys. Chem. Chem. Phys. 2005, 7, 3048–3051.
[25] Tiberg, F.; Zhmud, B.; Hallstensson, K.; Bahr, M. v. Phys. Chem. Chem.
Phys. 2000, 2, 5189–5196.
[26] Curtiss, J.; Khattabi, A.; Rios, F.; Smirnov, S. Hydrophobicity Switching as
a Drug Delivery Mechanism; 2011 In preparation.
76
Chapter 4
pH Valve Based on Hydrophobicity
Switching
4.1 Introduction
One of the rapidly growing branches in material science deals with development
of hybrid nanomaterials for applications in sensors, drug delivery, smart mem-
branes, and others. Smart gates represent a type of functional porous materials
that control molecular transport by changing their properties in response to ex-
ternal stimuli. Reports on such nanostructures can be found in the literature,
which include materials that are sensitive to such stimuli as light,[1] pressure,[2]
biochemical analytes,[3, 4, 5, 6, 7] pH,[7, 8, 9, 10, 11, 12, 13] temperature,[14] and
electrical potential.[15]
A number of studies have investigated pH sensitive surfaces and membranes.
77
Among them, the most common approach employs variation of the pore size in
membranes because of swelling/shrinking triggered by pH change. For example, in
membranes made of polymer mixtures and copolymers, variation in the pore size
because of pH change affects the mass flow through such pores.[8] Such pH sensi-
tivity is achieved by the acidic or basic functional group in the polymer chains con-
formable to swelling because of electrostatic repulsion at the pH range where they
are charged. By choosing appropriate groups, the flux enhancement at high or low
pH can be achieved. A similar effect can be achieved when polymeric molecules are
grafted on the surface of inert membranes acting as support.[8, 9] In systems with
small enough nanopores such as zeolites, more intricate supramolecular complexes
with conformational changes induced by pH change were also employed.[11, 12, 13]
A less common approach is based on controlling the hydrophobicity of nanopores.
Steinle et al.[10] reported that ionic current through nanoporous membranes mod-
ified by organic silanes with carboxylic ends dropped by 4 orders of magnitude
when pH was raised from pH 7 to above pH 8. That was assigned to transition
from the hydrophobic “closed” state at low pH to the hydrophilic “open” state
brought about by deprotonation (ionization) of the carboxylic groups at high
pHs. It remained unclear how sensitive the effect was to the density of carboxylic
groups, and the dynamics of changes was not addressed.
In this paper we utilize a similar concept but proceed further in designing a
pH sensitive valve, where not only ions but also the whole solution flow is altered
by pH in the range suitable for applications in drug delivery. We achieve it by
modifying the surface of inorganic nanoporous membranes with a mixed monolayer
that changes surface wettability from hydrophobic to hydrophilic by lowering pH
and thus allowing dramatic control of mass transport through the pores. We
78
illustrate how such modifications can be optimized by evaluating contact angles
on the similarly modified flat surfaces. This is an extension of the previously
introduced concept of hydrophobicity switching, where wettability of hydrophobic
surfaces of nanopores changes in response to various stimuli such as light and
biochemical analytes.[1, 5] Because the switching is induced by lowering the pH
from neutral to slightly acidic, this platform can be employed in intracellular drug
delivery systems where cargo is released from the hydrophobic pores because of
pH drop in endosomes and lysosomes.
4.2 Experimental Section
Materials
All chemicals were of analytical grade and purchased from commercial sup-
pliers: (3-aminopropyl) trimethoxysilane (APTS) and safranin O from Aldrich;
n-butyl trimethoxysilane (BTS) from Polyscience. The buffers were either pur-
chased, phosphate buffered saline (PBS) kit from Antibodies Incorporated (pH
7.4, 0.14 M NaCl), or prepared at 0.05 M concentration: pH 10.4 from triethy-
lamine (Aldrich), pH 4.2 from succinic anhydride (Aldrich), pH 5.0 from acetic
acid (Aldrich), and pH 7.4 from ethylenediamine (Fisher); 0.1 M NaCl was added
to all of them. The microscope glass slides were from Fisher Scientific, and alu-
mina nanoporous membranes Anopore (with a nominal diameter 0.2 μm) from
Whatman.
Surface Modification
Glass slides were first cleaned using piranha solution (35% H2O2-65% H2SO4.
Caution! piranha solution is explosive) for 20 min at 70 C, washed with copi-
79
ous amount of deionized (DI) water and dried in an oven for 30 min at 115 C.
Silanization was performed overnight in 2% v/v toluene solution of silane mixtures
(with different percentages of amino and butyl silanes).[1, 16] Modified slides were
washed with ethanol and methanol and cured for 3 h at 115 C.
Anodized aluminum oxide (AAO) membranes were first cleaned by boiling in
water for 20 min. After drying at 110 C for 20 min, they were silanized using
the same procedure as described above.
FTIR Characterization
Quality of modification of alumina membranes was evaluated with FTIR, for
which the spectra were recorded using a Perkin-Elmer Spectrum One FT-IR spec-
trometer (0.1 cm–1 resolution); unmodified membrane was used as a reference.
Contact Angle Measurements
Contact angles, θ, were measured with an Attension Theta optical tensiometer
from KSV Instruments LTD using the Sessile drop method. Droplets of approx-
imately 1.5 μL volume were deposited on the surface enclosed in an environment
chamber designed to minimize evaporation. Humidity close to 100% was main-
tained in the chamber by a container with liquid water. All measurements were
performed at 25 C. The value of θ was obtained by fitting the shape of the droplet
using the circular approximation algorithm of the Attension Theta software.
Diffusion Experiments
Diffusion experiments were carried out at 25 C in a diffusion cell made of two
optical cuvettes connected via a membrane through complementary holes on each.
The effective area of the membrane connecting the two compartments, reservoir
and sink, was 12 mm2. Solutions on both sides were buffered at the same pH and
ionic strength (0.1 M), but only the reservoir compartment initially contained
80
safranin O dye at 1.4 mM concentration. The solutions were constantly stirred
using magnetic stirrers, and the dye concentration in the sink was monitored using
a CHEM2000 UV–vis spectrometer from Ocean Optics by measuring absorption
at the wavelength of maximum (514 nm).
4.3 Results and Discussion
4.3.1 IR Absorption
There is not enough sensitivity for measuring IR absorption of monolayers on
glass slides, but FTIR spectra of alumina membranes modified with silanes of
different proportions can be recorded because of a 3 orders of magnitude increase
in the effective path length.[17] These spectra presented in Figure 4.1 have char-
acteristic features that can be assigned to aminopropyl and butyl silanes. Mem-
branes themselves present some artifacts such as lack of transmittance below 1300
cm−1 and the interference pattern due to variable membrane thickness that ap-
pears as oscillations of up to 1% in transmittance. A broad peak between 2700
and 3700 cm−1 because of hydroxyls appears as enhanced transmittance for all
samples when compared to an untreated membrane.[16, 18] Amine stretching ap-
pears as a broad absorption at lower frequencies, between 2300 and 3400 cm−1,
which increases with the concentration of aminated silane.[19] Both sharp peaks
of methylene stretching, asymmetric near νa(CH2) 2920 cm−1 and symmetric near
νs(CH2) 2860 cm−1,[16, 20, 21] practically do not change among the samples, as
would be expected since aminopropyl and butyl silanes have the same number of
methylenes. The effect is best seen in Figure 4.1B showing the spectra normalized
81
to purely butyl silane modified membrane (0% of APTS). The intensity hand, de-
clines with increasing proportion of APTS as the latter lacks the methyl. In Figure
4.1B it is seen as an increasing sharp feature at 2965 cm−1. The range 1300–1700
cm−1 has multiple bending vibrations of CH, amines, and hydroxyls. The latter
appear as enhanced transmittance near 1650 cm−1 in the modified membranes
while the former two are easier to recognize by comparison with membrane of 0%
APTS (Figure 4.1B). The increasing absorption near 1560 cm−1 with raising the
amount of APTS agrees with the corresponding peak of NH bending,[22] while
CHx bending vibration near 1370 cm−1 appears as a depression.
Figure 4.1: FTIR spectra of alumina membranes modified with different mixturesof butyl and aminopropyl silanes. A. The spectra are referenced with respectto an unmodified membrane. B. The spectra are referenced with respect to 0%of aminosilane in the modifying mixture. The inset illustrates that the amineconcentration on the surface is close to that in the modifying solution.
Plotting the absorbance at 2910 cm−1 versus the percentage of APTS in the
82
modification mixture, shown in the inset of Figure 4.1B, illustrates a straightfor-
ward correlation, which suggests that the percentage of amines in the monolayer
on the surface of alumina pores is the same as in the solution used for modifica-
tion. Such a correlation is not universal as we have seen it to differ dramatically
with longer aliphatic chains and longer amines on fused quartz as a substrate.[23]
Nevertheless, at least for similarly short butyl and aminopropyl silanes, the per-
centage of amines in the monolayer on the walls of alumina nanoporous membrane
is almost the same as in a modifying solution.
Since the thickness of a monolayer is almost 2 orders of magnitude less than the
pore diameter, the pore clearance remains the same after the modification. SEM
images confirm that (see figure 4.2 ). Because of such significant size dissimilarity,
the monolayer behavior on the pore walls is not much different from that on a
flat surface. Thus, if some segregation between aliphatic and aminated silanes
takes place, its characteristic dimension on the pore walls should also be similar
to such on flat surfaces. Hence, we can use pH sensitivity analysis of flat surfaces
modified with different mixtures of the two silanes to mimic anticipated behavior
inside the nanopores.
4.3.2 pH Sensitive Wetting of Flat Surfaces
Holmes-Farley et al.[24] have demonstrated that for some surfaces contact angles
of aqueous solutions vary with pH. Ionizable groups such as carboxyls and amines,
can make surfaces responsive to pH. For example, amines at low pH are charged
because of protonation and thus are more hydrophilic. At high pH they are neutral
and thus less hydrophilic.
83
Figure 4.2: SEM images of membranes with 0.2 nm pores before (A, C) andafter (B, D) surface modification with a mixture of butyl- and aminopropyl-trimethoxysilanes. The images are given for different magnifications and illustratethat the surface modification does not affect the pore clearance as is expected fora monolayer.
According to Holmes-Farley et al.,[24] wettability of surfaces modified with
only primary amines does not demonstrate significant variation with pH. We see
some variation, which is not as essential because such surface remains hydrophilic
(θ < 90) for both acidic and basic pH (see Figure 4.4). However, if amines are
mixed with aliphatic molecules in an appropriate ratio, lowering pH can cause
the surface switch from hydrophobic to hydrophilic. Figure 4.3 represents the
typical kinetics of the contact angles for Sessile droplets of different pH on purely
aliphatic surface and on a partially aminated surface prepared using a mixture of
aminated (APTS) and aliphatic (BTS) silanes. On the aliphatic surface, contact
angles have similar behavior for all pH: the angle initially decreases because of
mechanical equilibration (first few seconds of which are not shown) and settles at
an equilibrium value, θ 97–98. Mixed partially aminated surfaces demonstrate a
84
clear switch from hydrophobic (θ > 90) to hydrophilic (θ < 90) behavior upon
decrease of pH. The equilibrium contact angle in Figure 4.3 for droplets with pH
> 7 resembles that of the aliphatic surface with θ 93, while the drop of pH 4.2
shows a smaller contact angle, θ 82. The contact angle decrease is accompanied
by spreading of the drop which illustrates that the wettability change in this case
is not instantaneous.
Figure 4.3: Variation of the contact angle θ with time at different pH. Top:purely aliphatic butyl silane modifier. Bottom: 5% of amino silane (APTS) in themixture with butylsilane (BTS).
Figure 4.4 represents the variation of θ with pH as a function of the amino
silane portion (APTS) in the mixture used for surface modification. The graph
confirms that a surface modified with only aliphatic silane (BTS) is not pH sen-
sitive, while mixtures are. Because the droplet spreading on partially aminated
surfaces is relatively slow and is obscured by water evaporation at longer times,
the contact angles are given for a particular time, 50 s after droplet deposition,
85
when evaporation is not noticeable yet. Partially aminated surfaces all have lower
contact angles at pH = 4.2 than at pH = 7.4 and pH = 10.2 indicating their
enhanced wettability by water at low pH. With increasing percentage of amines
in the monolayer the contact angle decreases at both high and low pH, but the
latter decreases faster and thus increases the extent of pH induced variation. The
trend of pH dependent contact angle variation remains even for 100% amine, but
the contact angles indicate good wettability of this surface at all pHs.
Figure 4.4: Contact angles for droplets of different pH on surfaces modified withdifferent proportions of APTS in mixtures with BTS. The values were taken at50 s after depositing the drop on the surface. The lines are guides to the eyes.
It may seem strange at first that we observe switching for pH < 7 while pKa
of primary amines is near 10.5 but it should not be surprising. It is well-known
that charge repulsion from proximal cations in polymers like polylysine can result
in pKa reduction to pKa < 7.[25, 26] The range for pH switching should depend
on the density of amines. The pKa value is also affected by the proximity of
hydrophobic tails in the monolayer, which are of a similar length as the amine
and provide enough amine access to solution. When silanes with longer aliphatic
tails are used, they hide amines from access to water and thus prevent them
86
from ionization and significant altering of the surface tension. For example, flat
surfaces modified by 10% mixture of APTS with hexadecyl silane do not show
noticeable pH dependence of the contact angle (see Figure 4.5) . Different relative
lengths of aliphatic and amino chains and the density of amines in the hydrophobic
monolayer are the key factors that allow variation in the desired range of pH
switching. One can recognize the effect of density in Figure 3, where the switching
between hydrophobic and hydrophilic behavior (θ dropping below 90) for pH =
7.4 is close to 8% of amines and for pH = 4.2 it happens at a much lower density
3% of amines. These percentages may not coincide with those for the surface
bound amines, as was reported before for other silane mixtures of amine/aliphatic
tails,[21] but they are likely very similar for the short silanes used here, as was
confirmed above using FTIR on alumina membranes.
Nevertheless, because of the difference between silica and alumina surfaces,
some deviations should be anticipated. One should also appreciate the two-
dimensional nature of amines’ assembly on the surface, which should lead to a
much stronger variation of the resulting pKa with their surface density than in a
one-dimensional case like polylysine. Surfaces modified by carboxylic acids simi-
larly show significantly altered pKa > 8 greater than its pKa 4.5 in solution.[10]
In the pH sensitive valve application, as described below, it is imperative that
at the pH of the desired closed state the contact angle is greater than 90 while
it drops below 90 in the pH induced open state. One particular application of
such an effect is in drug delivery, where release of cargo can be induced by a
lower pH in lysosomes but should remain hydrophobic at pH 7.4 to retain the
cargo outside the region of low pH. Figure 2 shows that such a condition can be
realized using pH induced hydrophobicity switching in nanopores, the surface of
87
Figure 4.5: Illustration of equilibrium sessile drops (after 50 s) on flat surfacesmodified with different percentage of aminosilane (APTS) in mixtures with butyl-silane (BTS)
which is modified with a mixture of aliphatic and aminated silanes. The optimum
percentages of the latter should not exceed 7.5% for which we see that the contact
angle exceeds 90 at pH 7.4 but drops below 90 for acidic pH.
pH Gated Membranes
Hydrophobic pores of a small enough diameter remain dry in water and can
withstand a high pressure difference, ∆P , defined by the Laplace equation:
∆P >2γcosθ
r(4.1)
88
where r is the pore radius, γ is the liquid-vapor surface tension and θ is the
contact angle of the liquid in the flat surface.
The critical pressure exceeds 1 bar even for the very modest contact angles,
∼ 93, in quite large pores with Dpore 0.2 μm, creating an effective natural plug
against water intrusion.[2, 4] When the contact angle becomes smaller than 90,
this plug disappears and opens the valve.
On the basis of the data from the flat surface, 0.2 μm AAO membranes were
modified with 7.5% mixture of aminated (APTS) silane along with aliphatic (BTS)
silanes and investigated their performance as pH triggered valves. The transport
behavior was evaluated using safranin O dye that has a higher solubility in water
with a strong absorbance that does not significantly depend on pH.[27] Placing
concentrated safranin solution into the reservoir allows convenient optical mon-
itoring of its diffusion into the sink solution in real time. Both reservoirs had
solutions buffered at the same pH and with high concentration of salt (>0.1 M
NaCl) to minimize possible surface wall charge effects on transport of charged
molecules through the nanopores.[4, 6, 28]
Figure 4.6 shows that at pH 7.4 only the unmodified membrane is open to
flux of safranin molecules while the BTS modified membrane and the partially
aminated membrane prevent the dye transport because of their hydrophobicity. At
pH 4.2 the partially aminated membrane also opens up, as expected, allowing dye
to diffuse through to the sink solution. Note that the solely aliphatic modification
retains membrane hydrophobicity and blocks the dye transport. The rates of
transport for both, unmodified and modified, membranes are very similar as can
be seen from the slopes of their accumulation curves, with a short delay for the
partially aminated membrane because of a slow wetting of the pores by the solution
89
( 9 min in Figure 4.6). Changing pH from 4.2 back to neutral pH 7.4 does not
automatically regain the blockage because water filled hydrophobic pores do not
spontaneously dewet.[1, 2, 4] Formation of a bubble is the first step of dewetting
and is associated with a large activation barrier (>106kT for the 0.2 µm pore
diameter) that makes it kinetically unfeasible, as was discussed previously.[29, 30]
Drying off the membrane is required for the recovery of its water repelling property
at higher pHs.
Figure 4.6: Kinetics of safranin dye flux through nanoporous membranes with dif-ferent surface modifications and at different pH: (A) At pH = 7.4 only unmodifiedmembrane is open while purely hydrophobic and partially aminated membranesare closed. (B, C) Partially aminated membrane opens up at pH = 4.2 with theflux similar to that of the unmodified membrane. (D) An expanded version of Cillustrating a delay (9 min) required for the membrane wetting (dashed line). Theinset in B sketches the setup consisting of two optical cuvettes with complemen-tary holes on the sides connected via the modified membrane. Only the reservoir(pink) side has safranin at the beginning. Solutions on both sides are have beenstirred with magnetic bars (white).
The described switching effect is due to pH and not due to sensitivity to other
90
ions as the same behavior was observed with other buffers of similar pH, for
example, pH 7.4 with ethylenediamine buffer behaves as PBS and pH 5.0 with
acetic acid buffer behaves similar to pH 4.2 succinic anhydride buffer.
We wish to emphasize the benefits of our pH induced hydrophobicity switching
membranes in performing as valves when compared to other pH sensitive mem-
branes such as those employing pH induced swelling. These benefits are as follows:
(i) dry pores do not allow water intrusion and thus prevent any diffusion through
in the closed state, and (ii) in the open state, the flux is very large and is the
same as in the pores with no modification. The overall large contrast in control-
ling the flux should be particularly useful in drug delivery systems as well as in
other possible applications.
4.4 Conclusions
Hydrophobically modified nanoporous membranes prevent aqueous solutions from
passing through in both acidic and basic conditions. Mixed modifications with
5–8% of aminosilane (APTS) along with aliphatic (BTS) silane make surfaces pH
responsive, and the corresponding membranes perform as pH switchable valves.
These membranes retain hydrophobicity at neutral and basic conditions but open
to passage of molecules at slightly acidic conditions, which makes this platform
suitable for drug delivery application. The high contrast between the open and
closed states, as well as high fluxes in the open state because of large pore size
can probably be useful in other applications for pH switchable valves.
91
Bibliography
[1] Vlassiouk, I.; Park, C.-D.; Vail, S. A.; Gust, D.; Smirnov, S. Nano Lett. 2006,
6, 1013–1017.
[2] Smirnov, S.; Vlassiouk, I.; Takmakov, P.; Rios, F. ACSnano 2010, 4, 5069–
5075.
[3] Vlassiouk, I.; Takmakov, P.; Smirnov, S. Langmuir 2005, 21, 4776–4778.
[4] Vlassiouk, I.; Smirnov, S. Biosensing with Nanopores; Merkoci, A., Ed.; Wi-
ley, 2009; Vol. 0, pp 459–490.
[5] Rios, F.; Smirnov, S. ACS Appl. Mater. Interfaces 2009, 1, 768–744.
[6] Wang, X.; Smirnov, S. ACSnano 2009, 3, 1004–1010.
[7] Wandera, D.; Wickramasinghe, S. R.; Husson, S. M. J. Memb. Sci. 2010,
357, 6–35.
[8] Ying, L.; Wang, P.; Kang, E. T.; Neoh, K. G. Macromolecules 2002, 35,
673–679.
[9] Ito, Y.; Park, Y. S.; Imanishi, Y. Langmuir 2000, 16, 5376–5381.
92
[10] Steinle, E. D.; Mitchell, D. T.; Wirtz, M.; Lee, S. B.; Young, V. Y.; Martin,
C. R. Anal. Chem. 2002, 74, 2416–2422.
[11] Zhang, H.; Ito, Y. Langmuir 2001, 17, 8336–8340.
[12] Casasus, R.; Climent, E.; Marcos, M. D.; Martinez-Martinez, R.; Sanceno,
F.; Soto, J.; Amoros, P.; Cano, J.; Ruiz, E. J. Am. Chem. Soc. 2008, 130,
1903–1917.
[13] Nguyen, T. D.; Leung, K. C.-F.; Liong, M.; Pentecost, C. D.; Stoddart, J. F.;
Zink, J. I. Org. Lett. 2006, 8, 3363–3366.
[14] Jang, J. H.; In, I. Chem. Lett. 2010, 39, 1190–1191.
[15] Lahann, J.; Mitragotri, S.; Tran, T.-N.; Kaido, H.; Sundaram, J.; Choi, I. S.;
Hoffer, S.; Somorjai, G. A.; Langer, R. Science 2003, 299, 371–374.
[16] Vlassiouk, I.; Rios, F.; Vail, S. A.; Gust, D.; Smirnov, S. Langmuir 2007,
23, 7784–7792.
[17] Vlassiouk, I.; Krasnoslobodtsev, A.; Smirnov, S.; Germann, M. Langmuir
2004, 20, 9913–9915.
[18] Ott, A. W.; Klaus, J. W.; Johnson, J. M.; George, S. M. Acc. Chem. Res.
1997, 9, 707–714.
[19] Matinlinna, J. P.; Areva, S.; Lassila, L. V. J.; Vallittu, P. K. Surf. Interface
Anal. 2004, 36, 1314–1322.
[20] Porter, M. D.; Bright, T. B.; Allara, D. L.; Chidsey, C. E. D. J. Am. Chem.
Soc. 1987, 109, 3559–3568.
93
[21] Grange, J. D. L.; Markham, J. L.; Kurkjian, C. R. Langmuir 1993, 9, 1749–
1753.
[22] Kanan, S. M.; Tze, W. T. Y.; Tripp, C. P. Langmuir 2002, 18, 6623–6627.
[23] Krasnoslobodtsev, A.; Smirnov, S. Langmuir 2001, 17, 7593–7599.
[24] Holmes-Farley, S. R.; Bain, C. D.; Whitesides, G. M. Langmuir 1988, 4,
921–937.
[25] Sakai, N.; Futaki, S.; Matile, S. Soft Matter 2006, 2, 636–641.
[26] Bugg, T. Introduction to Enzyme and Coenzyme Chemistry ; Blackwell: Ox-
ford, 1997.
[27] Aznar, E.; Marcos, M. D.; Martinez-Manez, R.; Sancenon, F.; Soto, J.;
Amoros, P.; Guillem, C. J. Am. Chem. Soc. 2009, 131, 6833–6843.
[28] Vlassiouk, I.; Smirnov, S.; Siwy, Z. Nano Lett. 2008, 8, 1978–1985.
[29] Lum, K.; Chandler, D.; Weeks, J. D. J. Phys. Chem. B 1999, 103, 4570–
4577.
[30] Luzar, A. J. Phys. Chem. B 2004, 108, 19859–19866.
94
Chapter 5
Hydrophobic nanocontainer with
dual release mechanism
5.1 Introduction
The design of smart nanocontainers for drug delivery and other applications have
become an area of very active research in the last years.[1, 2] The ability to
hold molecular cargo and deliver it only when needed makes such nanocontainers
especially useful in drug delivery[3], sensing, imaging, self-healing coatings[4] and
other applications.[2]
The common approaches in fabricating nanocontainers are based of employing
vesicles, micelles, virus capsids and nanoporous inorganic materials.[1] The latter
have attracted a lot of attention since they were first described as a potential
drug delivery system (DDS).[5] These porous solids posses high area to volume
95
ratio, often have a network of ordered pores and their surface can be chemically
modified.
The modification is usually achieved by various silanes readily reacting with
the surface hydroxyl groups. Tethered on the surface, chemical moieties provide
the necessary surface functionality that can utilize a desired mechanism of drug
release. There are several examples in the literature where porous solids with
different triggering mechanisms for cargo delivery were employed including the
effects of light[6, 7], pH[8, 9, 10], temperature[11], enzymes[12], ions[13], reducing
agents[14, 15, 16, 17] etc. More recently, dual controlled release systems have been
reported.[7, 18, 19, 20]
We have demonstrated that hydrophobically modified nanopores even with
relatively large diameters (~0.2 μm), can effectively hold both hydrophobic and
hydrophilic cargo in aqueous solutions due to an effective hydrophobic plug but
promptly release the load when wetted by amphiphiles.[21] When the surface
modification combines hydrophobic molecules and amines, the surface becomes
pH sensitive and, when used in nanoporous membranes, the hydrophobic surface
character can be switched off by acids to result in a pH triggered valve.[22]
In the present work we illustrate that such mixed surface modifications can
utilize both mechanisms of cargo release, by amphiphiles and by pH, and propose
to employ such hydrophobic nanocontainers for drug delivery with the dual release
mechanism. The motivation for developing such system is based on the desire of
ensuring drug delivery using hydrophobic nanocontainers. Even though the above
described mechanism of delivery from hydrophobicity switching by phospholipids
of the cell membrane does not require endocytosis, for some small enough nanocon-
tainers it may be impossible to eliminate it completely. To make sure that such
96
internalized nanocontainers have additional means of cargo release we suggest that
typically employed pH triggering release would be the most logical choice.
5.2 Experimental section
All chemicals used in this work were obtained from Sigma- Aldrich (St Louis, MO)
and were used as received. Two types of nanocontainers were used: (a) nanoporous
alumina membranes and (b) silica nanotubes prepared by sol-gel using alimina
membranes as templates.
The surface modification of both types of nanocontainers was carried out as
described in the experimental section of chapter 4, with 5% relative portion of
APTS in the mixture with BTS, the previously optimization mixture for the
desired pH response.[22] The quality of membrane modification was characterized
by FT-IR and SEM.
The nanocontainers were loaded with safranin dye by immersion for 10 s into
ethanol solution (~1mM), drying and washing in PBS (phosphate buffer saline, pH
7.2) in order to remove safranin from the exterior and to recover the hydrophobic
entrance into the pores.
Kinetics of the dye release were studied using more convenient nanoporous
alumina membranes. Optical absorption of safranin at 530 nm was monitored for
loaded membranes submerged into 3.5 mL of discharging solution, which was con-
tinuously stirred. Amphiphiles, dodecyltrimethylammonium bromide (DTAB),
and sodium dodecyl sulfate (SDS), and the solutions of different pH, phosphate
buffer saline (PBS, pH=7.2 0.1 M NaCl, 10 mM phosphate) and acetate buffer
(AC, pH=5.0, 0.1M NaCl, 10 mM acetate) were used to observed the effects of
97
amphiphile and pH, respectively.
Free silica nanotubes were fabricated using the sol-gel method in alumina
membrane templates as outline elsewhere.[23, 24] Briefly, the sol-gel solution was
prepared by mixing absolute ethanol, tetraethyl orthosilicate (TEOS), and 1 M
HCl (50:5:1 vol/vol). Alumina template membranes were then immersed in the
sol-gel and sonicated for 30 min, air dried and the resulting sol-gel coated template
was cured overnight at 150 C. Afterwards, the silica surface was modified with the
pH sensitive coating as described above, then the membrane was polished on both
sides to remove the top silica layers and the tubes were released by dissolving the
alumina membrane in 1M NaOH. Free nanotubes were centrifuged and separated
from the supernatant solution by filtration through 0.2 mm membrane.
5.3 Results and discussion
Figure 5.1 sketches the modes of operation for such hydrophobic nanocontainers
with dual type release. The first mechanism is based on wetting the nanopores
with amphiphilic molecules (i.e. SDS or DTAB), while the second mechanism
changes wetting by a pH decrease in endosomes and lysosomes.
Figure 5.2 presents typical kinetic profiles for dye release into the acetate buffer
(pH=5.0) and the two amphiphile solutions, 4.6 g/L DTAB in PBS and 2.4 g/L
SDS PBS. All solutions had 0.1 M KCl added to eliminate the surface charge
effects in movement of charged molecules through the nanopores with charged
walls.[25]
The same Figure illustrates that no release occurs into PBS buffer, in agree-
ment with the mixed surface retaining its high contact angle at pH > 7 and thus
98
Figure 5.1: Nanocontainer with dual release mechanism. At pH=7.2 the nanocon-tainer is dry and the cargo remains inside. If put into contact with a solutionsof amphiphiles, the change in surface tensions at the interfaces will trigger therelease of cargo. In case of a decrease in pH (<5.0) the amines on the surfacewill protonate, creating a more hydrophilic environment and releasing the cargoas well.
maintaining the hydrophobic plug. Lowering the pH to 5, drops the contact angle
below 90o by ionizing the amines and thus wets the pores to induce the cargo re-
lease, as was previously demonstrated in construction of pH controlled valves.[22]
Here we confirm that such a pH triggered release is potentially attractive in drug
delivery systems, in which the pH drop to pH of ~5.1 occurs inside the endosomes
and lysosomes formed inside the cell upon digesting the foreign objects by cells
through endocytosis.[26] The amount of released dye reaches ~30% of the total
load in 15 min and gradually rises to ~40% within an hour. After 1.5 hours each
99
Figure 5.2: Controlled release from the nanocontainer with surfactantants (DTABand SDS) and acetate buffer (pH=5). PBS buffer (pH=7.2) does not trigger therelease
membrane was taken out of the discharging solution and immersed into ethanol,
which perfectly cleaned the membrane out of the remaining dye within minutes
and allowed evaluating the remaining amount loaded in the membrane.
The release by amphiphile induced hydrophobicity switching is also possible in
this system as is illustrated in Figure 5.2. Both amphiphilic solutions, DTAB and
SDS, with concentrations near their cmc (critical micellar concentration) do re-
lease safranin and more effectively (close to 100%) than via pH induced hydropho-
bicity switching. These concentrations are exceeding the necessary to decreases
the contact angle of C16 hydrophobic surfaces below 90o.[27]
One can notice that DTAB unloads the cargo much faster than SDS. This effect
was previously observed for purely hydrophobic containers where electrostatic
attraction between the oppositely charged cargo molecule and the head groups of
100
a surfactant always causes a slower rate of release than with neutral amphiphiles
(zwitterionic head group) or for the same charge of amphiphile molecules. There
are published reports confirming that positively charged safranin forms 1 to 1
complexes with SDS.[28]
A smaller amount of the released load by lowering pH can be related as to
at least two reasons. First of all, the range of pH triggering, i.e. where the
contact angle with water drops from above 90o to below 90o, is dependent on the
relative portion of amines in the monolayer. Their proximity and the hydrophobic
neighbors in the monolayer is what cause the normal pKa > 9 of primary amines
to effectively protonate at much lower pH.[29] We use ca 5% of amines (APTS) in
the monolayer to ensure lack of leakage at neutral pH, which may be insufficient
for complete opening of 100% pores at pH 5 because of slight variation in the
distribution of amines. Indeed, lowering pH to 4 slightly increases the amount of
released cargo. The second reason is due to lipophilicity of safranin, i.e. its ability
to assemble onto the hydrophobic monolayer. Such lipophilicity also reveals in
an increased amount of the cargo loaded as compared with lipophobic molecules.
[30]
We wish to emphasize the uniqueness of the proposed here delivery system
with dual mechanism of release when compared to other dual delivery systems.
Even though nanocontainers with hydrophobic coatings have been described in
the literature [31, 32, 33, 30, 34], they were inferior in their ability of retaining
the load and no triggering mechanism for active cargo release was intended for
them – the delivery was viewed as a hindered diffusion of molecules out of the
nanocontainers. Aminated surface modification has been also explored by others
but the mechanism for cargo holding and release was based on electrostatic in-
101
teraction with the surface, i.e. dedicated to negatively charged molecules such as
DNA.[35, 36]
In contrast, our nanontainers with relatively large pore diameters i) have al-
most zero leakage in the dry state at neutral pH since they do not allow water
intrusion inside the pores due to the hydrophobic plug, ii) in the open state in-
duced by hydrophobicity switching triggered either by pH or by amphiphiles, the
cargo release is fairly fast. These features allow considering this delivery system as
a truly triggered delivery. The amphiphile (phospholipid) triggering mechanism
is expected to induce the cargo translocation without need for endocytosis of the
nanocontainers themselves, which upon attaching to the cell membrane can be
coated inside by intruding of phospholipids and thus allow the payload to escape
into the opening gap in the membrane. Nevertheless, since there are no simple
means for complete elimination of endocytosis when small nanocontainers used as
delivery system, some of them will end up inside endosomes and lysosome. The
pH drop in the latter should trigger a complementary route of cargo release and
thus maximize the overall performance through such a double assurance.
Finally, we have tested the modification outlined in this chapter in free silica
nanotubes. Figure 5.3 shows solution of buffers at different pH were loaded silica
nanotubes have been deposited.
Although the tubes show pH sensitivity, such sensitivity has a shifted pH
threshold when compared with the alumina nanocontainers. The free nanotubes
modified with 5%amino silane in butyl silane are less stable in keeping their cargo
at pH 7.4 but do it much better for pH 8.5. A possible explanation for this results
is the deterioration of the mixed sensitive layer due to the relatively harsh treat-
ment to which they were exposed during release with 1M NaOH. An alternative
102
Figure 5.3: Release of safranine from modified free silica nanotubes at differentpH.
approach would be to perform surface modification after nanotube release or use
alternative formulations for surface modification to have more stable monolayer; it
has been illustrated before that APTS monolayers are not particularly stable.[37]
5.4 Conclusions
A hydrophobic nanocontainer with dual release mechanism has been proposed
and realized using alumina nanoporous membranes and silica nanotubes. Both
mechanism aim toward drug delivery into live cells through previously undescribed
methods. The release induced by amphiphiles decorating the hydrophobic pore
surface is expected when nanocontainers are in contact with the cell membrane,
while the pH triggering of the same nanopores is expected to be triggered by
103
endocytosis, where the decrease of pH to ~pH 5 occurs inside the endosomes and
liposomes.
104
Bibliography
[1] Trewyn, B. G.; Giri, S.; Slowing, I. I.; Lin, V. S.-Y. Chem. Comm. 2007, 43,
3236–3245.
[2] Kim, K. T.; Meeuwissen, S. A.; Nolte, R. J. M.; Hest, J. C. M. v. Nanoscale
2010, 2, 844–858.
[3] Taylor-Pashow, K. M. L.; Rocca, J. D.; Huxforda, R. C.; Lin, W. Chem.
Comm. 2010, 46, 5832–5849.
[4] Shchukin, D. G.; Mohwald, H. Chem. Comm. 2011, 47, 8730–8739.
[5] Vallet-Regi, M.; Ramila, A.; Real, R. P. d.; Perez-Pariente, J. Chem. Mater.
2001, 13, 308–311.
[6] Mal, N. K.; Fujiwara, M.; Tanaka, Y. Nature 2003, 421, 350–353.
[7] Aznar, E.; Casasus, R.; Garcia-Acosta, B.; Marcos, M. D.; Martinez-Manez,
R.; Sancenon, F.; Soto, J.; Amoros, P. Adv. Mater. 2007, 19, 2228–2231.
[8] Nguyen, T. D.; Leung, K. C.-F.; Liong, M.; Pentecost, C. D.; Stoddart, J. F.;
Zink, J. I. Org. Lett. 2006, 8, 3363–3366.
105
[9] Yang, Q.; Wang, S.; Fan, P.; Wang, L.; Di, Y.; Lin, K.; Xiao, F.-S. Chem.
Mater. 2005, 17, 5999–6003.
[10] Kim, T.-W.; Slowing, I. I.; Chung, P.-W.; Lin, V. S.-Y. ACSnano 2011, 5,
360–366.
[11] Chang, J. H.; Shim, C. H.; Kim, B. J.; Shin, Y.; Exarhos, G. J.; Kim, K. J.
Adv. Mater. 2005, 17, 634–637.
[12] Bernardos, A.; Mondragon, L.; Aznar, E.; Marcos, M. D.; Martinez-Manez,
R.; Sancenon, F.; Soto, J.; Barat, J. M.; Perez-Paya, E.; Guillem, C.; Amoros,
P. ACSnano 2010, 4, 6353–6368.
[13] Casasus, R.; Marcos, M. D.; Martinez-Manez, R.; Ros-Lis, J. V.; Soto, J.;
Villaescusa, L. A.; Amoros, P.; Beltran, D.; Guillem, C.; Latorre, J. J. Am.
Chem. Soc. 2004, 126, 8612–8613.
[14] Lai, C.-Y.; Trewyn, B. G.; Jeftinija, D. M.; Jeftinija, K.; Xu, S.; Jeftinija, S.;
Lin, V. S.-Y. J. Am. Chem. Soc. 2003, 125, 4451–4459.
[15] Giri, S.; Trewyn, B. G.; Stellmaker, M. P.; Lin, V. S.-Y. Angew. Chem. Int.
Ed. 2005, 44, 5038–5044.
[16] Nguyen, T. D.; Tseng, H.-R.; Celestre, P. C.; Flood, A. H.; Liu, Y.; Stoddart,
J. F.; Zink, J. I. PNAS 2005, 102, 10029–10034.
[17] Liu, R.; Zhao, X.; Wu, T.; Feng, P. J. Am. Chem. Soc. 2008, 130, 14418–
14419.
106
[18] Casasus, R.; Climent, E.; Marcos, M. D.; Martinez-Martinez, R.; Sanceno,
F.; Soto, J.; Amoros, P.; Cano, J.; Ruiz, E. J. Am. Chem. Soc. 2008, 130,
1903–1917.
[19] Aznar, E.; Marcos, M. D.; Martinez-Manez, R.; Sancenon, F.; Soto, J.;
Amoros, P.; Guillem, C. J. Am. Chem. Soc. 2009, 131, 6833–6843.
[20] Chen, C.; Geng, J.; Pu, F.; Yang, X.; Ren, J.; Qu, X. Angew. Chem. Int. Ed.
2011, 50, 882–886.
[21] Curtiss, J.; Khattabi, A.; Rios, F.; Smirnov, S. Hydrophobicity Switching as
a Drug Delivery Mechanism; 2011 In preparation.
[22] Rios, F.; Smirnov, S. N. Chem. Mater. 2011, 23, 3601–3605.
[23] Lakshmi, B. B.; Patrissi, C. J.; Martin, C. R. Chem. Mater. 1997, 9, 2544–
2550.
[24] Okamoto, K.; Shook, C. J.; Bivona, L.; Lee, S. B.; English, D. S. Nano Lett.
2004, 4, 233–239.
[25] Vlassiouk, I.; Park, C.-D.; Vail, S. A.; Gust, D.; Smirnov, S. Nano Lett. 2006,
6, 1013–1017.
[26] Mellman, I.; Fuchs, R.; Helenius, A. Ann. Rev. Biochem. 1986, 55, 663–700.
[27] Dutschk, V.; Sabbatovskiy, K. G.; Stolz, M.; Grundke, K.; Rudoy, V. M. J.
Colloid Interface Sci. 264, 267, 456–462.
[28] Jana, A. K.; Parameswari, S. Colloid Polym. Sci. 2005, 283, 1056–1062.
107
[29] Holmes-Farley, S. R.; Bain, C. D.; Whitesides, G. M. Langmuir 1988, 4,
921–937.
[30] Doadrio, J. C.; Sousa, E. M. B.; Izquierdo-Barba, I.; Doadrio, A. L.; Perez-
Pariente, J.; Vallet-Regi, M. J. Mater. Chem. 2006, 16, 462–466.
[31] Tang, Q.; Chen, Y.; Chen, J.; Li, J.; Xu, Y.; Wu, D.; Sun, Y. J. Solid State
Chem. 2010, 183, 76–83.
[32] Tang, Q.; Xu, Y.; Wu, D.; Sun, Y. Chem. Lett. 2006, 35, 474–475.
[33] Kapoor, S.; Hegde, R.; Bhattacharyya, A. J. Acc. Chem. Res. 2009, 140,
34–39.
[34] Zhu, S.; Zhang, D.; Yang, N. J. Nanopart. Res. 2009, 11, 561–568.
[35] Manzano, M.; Aina, V.; Arean, C.; Balas, F.; Cauda, V.; Colilla, M.; Delgado,
M.; Vallet-Regi, M. Chem. Eng. J. 2008, 137, 30–37.
[36] Szegedi, A.; Popova, M.; Goshev, I.; Mihaly, J. J. Solid State Chem. 2011,
184, 1201–1207.
[37] Szczepanski, V.; Vlassiouk, I.; Smirnov, S. J. Memb. Sci. 2006, 281, 587–591.
108