Nano surface chemistry
Published on: Mar 3, 2016
Transcripts - Nano surface chemistry
This book is printed on acid-free paper.
Marcel Dekker, Inc.
270 Madison Avenue, New York, NY 10016
tel: 212-696-9000; fax: 212-685-4540
Eastern Hemisphere Distribution
Marcel Dekker AG
Hutgasse 4, Postfach 812, CH-4001 Basel, Switzerland
tel: 41-61-261-8482; fax: 41-61-261-8896
World Wide Web
The publisher offers discounts on this book when ordered in bulk quantities. For more information,
write to Special Sales/Professional Marketing at the headquarters address above.
Copyright © 2002 by Marcel Dekker, Inc. All Rights Reserved.
Neither this book nor any part may be reproduced or transmitted in any form or by any means,
electronic or mechanical, including photocopying, microfilming, and recording, or by any
information storage and retrieval system, without permission in writing from the publisher.
Current printing (last digit):
10 9 8 7 6 5 4 3 2 1
PRINTED IN THE UNITED STATES OF AMERICA
Tools shape how we think; when the only tool you have is an axe, everything resembles a
tree or a log. The rapid advances in instrumentation in the last decade, which allow us to
measure and manipulate individual molecules and structures on the nanoscale, have caused
a paradigm shift in the way we view molecular behavior and surfaces. The microscopic de-
tails underlying interfacial phenomena have customarily been inferred from in situ mea-
surements of macroscopic quantities. Now we can see and “finger” physical and chemical
processes at interfaces.
The reviews collected in this book convey some of the themes recurrent in nano-col-
loid science: self-assembly, construction of supramolecular architecture, nanoconfinement
and compartmentalization, measurement and control of interfacial forces, novel synthetic
materials, and computer simulation. They also reveal the interaction of a spectrum of dis-
ciplines in which physics, chemistry, biology, and materials science intersect. Not only is
the vast range of industrial and technological applications depicted, but it is also shown
how this new way of thinking has generated exciting developments in fundamental science.
Some of the chapters also skirt the frontiers, where there are still unanswered questions.
The book should be of value to scientific readers who wish to become acquainted
with the field as well as to experienced researchers in the many areas, both basic and tech-
nological, of nanoscience.
The lengthy maturation of a multiauthored book of this nature is subject to life’s con-
tingencies. Hopefully, its structure is sound and has survived the bumps of “outrageous for-
tune.” I wish to thank all the contributors for their courage in writing. It is their work and
commitment that have made this book possible.
1. Molecular Architectures at Solid–Liquid Interfaces Studied by Surface Forces
2. Adhesion on the Nanoscale 17
Suzanne P. Jarvis
3. Langmuir Monolayers: Fundamentals and Relevance to Nanotechnology 59
Keith J. Stine and Brian G. Moore
4. Supramolecular Organic Layer Engineering for Industrial Nanotechnology 141
Claudio Nicolini, V. Erokhin, and M. K. Ram
5. Mono- and Multilayers of Spherical Polymer Particles Prepared by
Langmuir–Blodgett and Self-Assembly Techniques 213
Bernd Tieke, Karl-Ulrich Fulda, and Achim Kampes
6. Studies of Wetting and Capillary Phenomena at Nanometer Scale with Scanning
Polarization Force Microscopy 243
Lei Xu and Miquel Salmeron
7. Nanometric Solid Deformation of Soft Materials in Capillary Phenomena 289
Martin E. R. Shanahan and Alain Carré
8. Two-Dimensional and Three-Dimensional Superlattices: Syntheses and Collective
Physical Properties 315
9. Molecular Nanotechnology and Nanobiotechnology with Two-Dimensional Protein
Crystals (S-Layers) 333
Uwe B. Sleytr, Margit Sára, Dietmar Pum, and Bernhard Schuster
10. DNA as a Material for Nanobiotechnology 391
Christof M. Niemeyer
11. Self-Assembled DNA/Polymer Complexes 431
Vladimir S. Trubetskoy and Jon A. Wolff
12. Supramolecular Assemblies Made of Biological Macromolecules 461
Nir Dotan, Noa Cohen, Ori Kalid, and Amihay Freeman
13. Reversed Micelles as Nanometer-Size Solvent Media 473
Vincenzo Turco Liveri
14. Engineering of Core-Shell Particles and Hollow Capsules 505
15. Electro-Transport in Hydrophilic Nanostructured Materials 527
Bruce R. Locke
16. Electrolytes in Nanostructures 625
17. Polymer–Clay Nanocomposites: Synthesis and Properties 653
Syed Qutubuddin and Xiaoan Fu
Alain Carré Fontainebleau Research Center, Corning S.A., Avon, France
Frank Caruso Max-Planck-Institute of Colloids and Interfaces, Potsdam, Germany
Kwong-Yu Chan Department of Chemistry, The University of Hong Kong, Hong Kong
Noa Cohen Department of Molecular Microbiology and Biotechnology, Faculty of Life
Sciences, Tel Aviv University, Tel Aviv, Israel
Nir Dotan Glycominds Ltd., Maccabim, Israel
V. Erokhin Department of Biophysical M&O Science and Technologies, University of
Genoa, Genoa, Italy
Amihay Freeman Department of Molecular Microbiology and Biotechnology, Faculty
of Life Sciences, Tel Aviv University, Tel Aviv, Israel
Xiaoan Fu Department of Chemical Engineering, Case Western Reserve University,
Karl-Ulrich Fulda Institute of Physical Chemistry, University of Cologne, Cologne,
Suzanne P. Jarvis Nanotechnology Research Institute, National Institute of Advanced
Industrial Science and Technology, Ibaraki, Japan
Ori Kalid Department of Molecular Microbiology and Biotechnology, Faculty of Life
Sciences, Tel Aviv University, Tel Aviv, Israel
Achim Kampes Institute for Physical Chemistry, University of Cologne, Cologne,
Kazue Kurihara Institute for Chemical Reaction Science, Tohoku University, Sendai,
Bruce R. Locke Department of Chemical Engineering, Florida State University,
Brian G. Moore School of Science, Penn State Erie–The Behrend College, Erie,
Claudio Nicolini Department of Biophysical M&O Science and Technologies,
University of Genoa, Genoa, Italy
Christof M. Niemeyer Department of Biotechnology, University of Bremen, Bremen,
Marie-Paule Pileni Université Pierre et Marie Curie, LM2N, Paris, France
Dietmar Pum Center for Ultrastructure Research, Universität für Bodenkultur Wien,
Syed Qutubuddin Department of Chemical Engineering, Case Western Reserve
University, Cleveland, Ohio
M. K. Ram Department of Biophysical M&O Science and Technologies, University of
Genoa, Genoa, Italy
Miquel Salmeron Materials Sciences Division, Lawrence Berkeley National
Laboratory, Berkeley, California
Margit Sára Center for Ultrastructure Research, Universität für Bodenkultur Wien,
Bernhard Schuster Center for Ultrastructure Research, Universität für Bodenkultur
Wien, Vienna, Austria
Martin E. R. Shanahan Adhesion, Wetting, and Bonding, National Centre for Scientific
Research/School of Mines Paris, Evry, France
Uwe B. Sleytr Center for Ultrastructure Research, Universität für Bodenkultur Wien,
Keith J. Stine Department of Chemistry and Center for Molecular Electronics,
University of Missouri–St. Louis, St. Louis, Missouri
Bernd Tieke Institute for Physical Chemistry, University of Cologne, Cologne,
Vladimir S. Trubetskoy Mirus Corporation, Madison, Wisconsin
Vincenzo Turco Liveri Department of Physical Chemistry, University of Palermo,
Jon A. Wolff Departments of Pediatrics and Medical Genetics, University of
Wisconsin–Madison, Madison, Wisconsin
Lei Xu Materials Sciences Division, Lawrence Berkeley National Laboratory, Berkeley,
The problems of chemistry and biology can be greatly helped if our ability to see what we are
doing, and to do things on an atomic level is ultimately developed—a development which I
think can’t be avoided.
God created all matter—but the surfaces are the work of the Devil.
The prefix nano-, derived from the Greek word meaning “dwarf,” has been applied most of-
ten to systems whose functions and characteristics are determined by their tiny size. Struc-
tures less than 100 nanometers in length (i.e., one-ten-millionth of a meter) are typical in
nano-technology, which emphasizes the approach of building up from molecules and nano-
structures (“bottom-up”) versus the “top-down,” or miniaturization, approach. Nano- actually
refers not so much to the size of the object as to the resolution at the molecular scale. At such
small scales, about half of the atoms are in the surface layer, the surface energy dominates,
and the surface layer can be considered a new material with properties different from those of
bulk. The hierarchy of scales, both spatial and temporal, is represented in the following table:
Quantum Atom/nano Mesoscopic Macroscopic
Length (meters) 10Ϫ11
Time (seconds) 10Ϫ16
Classical surface and colloid chemistry generally treats systems experimentally in a
statistical fashion, with phenomenological theories that are applicable only to building sim-
plified microstructural models. In recent years scientists have learned not only to observe
individual atoms or molecules but also to manipulate them with subangstrom precision.
The characterization of surfaces and interfaces on nanoscopic and mesoscopic length scales
is important both for a basic understanding of colloidal phenomena and for the creation and
mastery of a multitude of industrial applications.
The self-organization or assembly of units at the nanoscale to form supramolecular
ensembles on mesoscopic length scales comprises the range of colloidal systems. There is
a need to understand the connection between structure and properties, the evolution and dy-
namics of these structures at the different levels—supramolecular, molecular, and sub-
molecular—by “learning from below.”
When interaction and physical phenomena length scales become comparable to or
larger than the size of the structure, as, for example, with polymer contour chain length, the
system may exhibit unusual behavior and generate novel arrangements not accessible in
It is also at these levels (10–500 nm) that nature utilizes hierarchical assemblies in bi-
ology, and biological processes almost invariably take place at the nanoscale, across mem-
branes and at interfaces. Biomolecular materials with unique properties may be developed
by mimicking biological processes or modifying them. There is still much to discover about
improving periodic arrays of biomolecules, biological templating, and how to exploit the
differences between biological and nonbiological self-assembly.
The linkage of microscopic and macroscopic properties is not without challenges,
both theoretical and experimental. Statistical mechanics and thermodynamics provide the
connection between molecular properties and the behavior of macroscopic matter. Coupled
with statistical mechanics, computer simulation of the structure, properties, and dynamics
of mesoscale models is now feasible and can handle the increase in length and time scales.
Scanning proble techniques (SPM)—i.e., scanning tunneling microscopy (STM) and
atomic force microscopy (AFM), as well as their variations—have the power to visualize
nanoscale surface phenomena in three dimensions, manipulate and modify individual
molecules, and measure such properties as adhesion, stiffness, and friction as well as mag-
netic and electric fields. The use of chemically modified tips extends the technique to in-
clude chemical imaging and measurement of specific molecular interactions. Improved op-
tical methods complement probe images and are capable of imaging films a single molecule
thick. Optical traps, laser tweezers, and “nano-pokers” have been developed to measure
forces and manipulate single molecules. In addition, there is a vast range of experimental
tools that cross different length and time scales and provide important information (x-ray,
neutrons, surface plasmon resonance). Nevertheless, there is a further need for instrumen-
tation of higher resolution, for example, in the decreased ranged of space and time en-
countered when exploring the dynamics and kinetics of surface films.
Chapter 1 is a view of the potential of surface forces apparatus (SFA) measurements
of two-dimensional organized ensembles at solid–liquid interfaces. At this level, informa-
tion is acquired that is not available at the scale of single molecules. Chapter 2 describes
the measurement of surface interactions that occur between and within nanosized surface
structures—interfacial forces responsible for adhesion, friction, and recognition.
In Chapter 3, Langmuir–Blodgett films of varying organizational complexity are dis-
cussed, as well as nanoparticles and fullerenes. Molecular dynamic simulation of mono-
layers and multilayers of surfactants is also reviewed. Chapter 4 presents those aspects of
supramolecular layer assemblies related to the development of nanotechnological applica-
tions. Problems of preparing particle films with long-range two-dimensional and three-di-
mensional order by Langmuir–Blodgett and self-assembly techniques are dealt with in
The next two chapters are concerned with wetting and capillarity. Wetting phenom-
ena are still poorly understood; contact angles, for example, are simply an empirical pa-
rameter to quantify wettability. Chapter 6 reviews the use of scanning polarization force
microscopy (SPFM), a new application of AFM using electrostatic forces, to study the
nanostructure of liquid films and droplets. The effect of solid nanometric deformation on
the kinetics of wetting and dewetting and capillary flow in soft materials, such as some
polymers and gels, is treated in Chapter 7.
Chapter 8 presents evidence on how the physical properties of colloidal crystals or-
ganized by self-assembly in two-dimensional and three-dimensional superlattices differ
from those of the free nanoparticles in dispersion.
A biomolecular system of glycoproteins derived from bacterial cell envelopes that
spontaneously aggregates to form crystalline arrays in the mesoscopic range is reviewed in
Chapter 9. The structure and features of these S-layers that can be applied in biotechnol-
ogy, membrane biomimetics, sensors, and vaccine development are discussed.
DNA is ideally suited as a structural material in supramolecular chemistry. It has
sticky ends and simple rules of assembly, arbitrary sequences can be obtained, and there is
a profusion of enzymes for modification. The molecule is stiff and stable and encodes in-
formation. Chapter 10 surveys its varied applications in nanobiotechnology. The emphasis
of Chapter 11 is on DNA nanoensembles, condensed by polymer interactions and electro-
static forces for gene transfer. Chapter 12 focuses on proteins as building blocks for nano-
The next two chapters concern nanostructured core particles. Chapter 13 provides ex-
amples of nano-fabrication of cored colloidal particles and hollow capsules. These systems
and the synthetic methods used to prepare them are exceptionally adaptable for applications
in physical and biological fields. Chapter 14, discusses reversed micelles from the theoret-
ical viewpoint, as well as their use as nano-hosts for solvents and drugs and as carriers and
Chapter 15 gives an extensive and detailed review of theoretical and practical aspects
of macromolecular transport in nanostructured media. Chapter 16 examines the change in
transport properties of electrolytes confined in nanostructures, such as pores of membranes.
The confinment effect is also analyzed by molecular dynamic simulation.
Nanolayers of clay interacting with polymers to form nanocomposites with improved
material properties relative to the untreated polymer are discussed in Chapter 17.
Molecular Architectures at Solid–Liquid
Interfaces Studied by Surface Forces
KAZUE KURIHARA Tohoku University, Sendai, Japan
Molecular and surface interactions are ubiquitous in molecular science, including biology.
Surface forces measurement and atomic force microscopy (AFM) have made it possible to
directly measure, with high sensitivity, molecular and surface interactions in liquids as a
function of the surface separation. Naturally, they have become powerful tools for study-
ing the origins of forces (van der Waals, electrostatic, steric, etc.) operating between
molecules and/or surfaces of interest [1–4]. They also offer a unique, novel surface char-
acterization method that “monitors surface properties changing from the surface to the bulk
(depth profiles)” and provides new insights into surface phenomena. This method is direct
and simple. It is difficult to obtain a similar depth profile by other methods; x-ray and neu-
tron scattering measurements can provide similar information but require extensive instru-
mentation and appropriate analytical models .
Molecular architectures are self-organized polymolecular systems where molecular
interactions play important roles . They exhibit specific and unique functions that could
not be afforded by single molecules. Molecular architecture chemistry beyond molecules
is not only gaining a central position in chemistry but becoming an important interdisci-
plinary field of science. Investigations of molecular architectures by surface forces mea-
surement is important for the following reasons.
1. It is essential to elucidate intermolecular interactions involved in self-organization,
whose significance is not limited to material science but extends to the ingenuity of bi-
ological systems .
2. The importance of surface characterization in molecular architecture chemistry and en-
gineering is obvious. Solid surfaces are becoming essential building blocks for con-
structing molecular architectures, as demonstrated in self-assembled monolayer for-
mation  and alternate layer-by-layer adsorption . Surface-induced structuring of
liquids is also well-known [8,9], which has implications for micro- and nano-tech-
nologies (i.e., liquid crystal displays and micromachines). The virtue of the force mea-
surement has been demonstrated, for example, in our report on novel molecular archi-
tectures (alcohol clusters) at solid–liquid interfaces .
3. Two-dimensionally organized molecular architectures can be used to simplify the
complexities of three-dimensional solutions and allow surface forces measurement. By
employing this approach, we can study complex systems such as polypeptides and
polyelectrolytes in solutions. For example, it is possible to obtain essential information
such as the length and the compressibility of these polymers in solutions by systemat-
ically varying their chemical structures and the solution conditions .
Earlier studies of surface forces measurement were concerned mainly with surface
interactions determining the colloidal stability, including surfactant assemblies. It has
been demonstrated, however, that a “force–distance” curve can provide much richer in-
formation on surface molecules; thus it should be utilized for studying a wider range of
phenomena . Practically, the preparation of well-defined surfaces, mostly modified
by two-dimensional organized molecules, and the characterization of the surfaces by
complementary techniques are keys to this approach. A similar concept is “force spec-
troscopy” , coined to address force as a new parameter for monitoring the properties
of materials. A major interest in force spectroscopy is the single molecular measurement
generally employing an atomic force microscope. This measurement treats relatively
strong forces, such as adhesion, and discusses the binding of biotin-streptavidin  and
complementary strands of DNA  as well as the unfolding and folding of proteins
. On the other hand, the forces measurement of two-dimensionally organized
molecules has advantages complementary to those of single molecule force spectroscopy.
It can monitor many molecules at the same time and thus is better suited for studying
long-range weaker forces. The measurement should bear a close relevance to real systems
that consist of many molecules, because interactions between multiple molecules and/or
macroscopic surfaces in solvents may exhibit characteristics different from those between
The aim of this review is to demonstrate the potential of surface forces measurement
as a novel means for investigating surfaces and complex soft systems by describing our re-
cent studies, which include cluster formation of alcohol, polyion adsorption, and polyelec-
II. SURFACE FORCES MEASUREMENT
Surface forces measurement directly determines interaction forces between two surfaces as
a function of the surface separation (D) using a simple spring balance. Instruments em-
ployed are a surface forces apparatus (SFA), developed by Israelachivili and Tabor ,
and a colloidal probe atomic force microscope introduced by Ducker et al.  (Fig. 1). The
former utilizes crossed cylinder geometry, and the latter uses the sphere-plate geometry.
For both geometries, the measured force (F) normalized by the mean radius (R) of cylin-
ders or a sphere, F/R, is known to be proportional to the interaction energy, Gƒ, between
flat plates (Derjaguin approximation),
ᎏ ϭ 2Gƒ (1)
This enables us to quantitatively evaluate the measured forces, e.g., by comparing them
with a theoretical model.
Sample surfaces are atomically smooth surfaces of cleaved mica sheets for SFA, and
various colloidal spheres and plates for a colloidal probe AFM. These surfaces can be mod-
ified using various chemical modification techniques, such as Langmuir–Blodgett (LB) de-
position [12,19] and silanization reactions [20,21]. For more detailed information, see the
original papers and references texts.
III. ALCOHOL CLUSTER FORMATION ON SILICA SURFACES
Surface forces measurement is a unique tool for surface characterization. It can directly
monitor the distance (D) dependence of surface properties, which is difficult to obtain by
other techniques. One of the simplest examples is the case of the electric double-layer force.
The repulsion observed between charged surfaces describes the counterion distribution in
the vicinity of surfaces and is known as the electric double-layer force (repulsion). In a sim-
ilar manner, we should be able to study various, more complex surface phenomena and ob-
tain new insight into them. Indeed, based on observation by surface forces measurement
and Fourier transform infrared (FTIR) spectroscopy, we have found the formation of a
novel molecular architecture, an alcohol macrocluster, at the solid–liquid interface.
Adsorption phenomena from solutions onto solid surfaces have been one of the im-
portant subjects in colloid and surface chemistry. Sophisticated application of adsorption
has been demonstrated recently in the formation of self-assembling monolayers and multi-
layers on various substrates [4,7]. However, only a limited number of researchers have
been devoted to the study of adsorption in binary liquid systems. The adsorption isotherm
and colloidal stability measurement have been the main tools for these studies. The molec-
ular level of characterization is needed to elucidate the phenomenon. We have employed
the combination of surface forces measurement and Fourier transform infrared spec-
troscopy in attenuated total reflection (FTIR-ATR) to study the preferential (selective) ad-
sorption of alcohol (methanol, ethanol, and propanol) onto glass surfaces from their binary
mixtures with cyclohexane. Our studies have demonstrated the cluster formation of alco-
hol adsorbed on the surfaces and the long-range attraction associated with such adsorption.
We may call these clusters macroclusters, because the thickness of the adsorbed alcohol
layer is about 15 nm, which is quite large compared to the size of the alcohol. The follow-
ing describes the results for the ethanol–cycohexane mixtures .
Typical forces profiles measured between glass surfaces in ethanol–cyclohexane
mixtures are shown in Fig. 2. Colloidal probe atomic force microscopy has been employed.
In pure cyclohexane, the observed force agrees well with the conventional van der Waals
attraction calculated with the nonretarded Hamaker constant for glass/cyclohexane/glass,
Surface Forces Measurement 3
FIG. 1 Schematic drawings of (a) the surface forces apparatus and (b) the colloidal probe atomic
3.1 ϫ 10Ϫ21
J. At an ethanol concentration of 0.1 mol%, the interaction changes remark-
ably: The long-range attraction appears at a distance of 35 nm, shows a maximum around
10 nm, and turns into repulsion at distances shorter than 5 nm. The pull-off force of the con-
tacting surfaces is 140 Ϯ 19 mN/m, which is much higher than that in pure cyclohexane,
10 Ϯ 7 mN/m. Similar force profiles have been obtained on increasing the ethanol con-
centration to 0.4 mol%. A further increase in the concentration results in a decrease in the
long-range attraction. At an ethanol concentration of 1.4 mol%, the interaction becomes
identical to that in pure cyclohexane. When the ethanol concentration is increased, the
range where the long-range attraction extends changes in parallel to the value of the pull-
off force, indicating that both forces are associated with the identical phenomenon, most
likely the adsorption of ethanol. Separation force profiles after the surfaces are in contact
shows the presence of a concentrated ethanol layer near and on the surfaces (see Ref. 10a).
The short-range repulsion is ascribable to steric force due to structure formation of ethanol
molecules adjacent to the glass surfaces.
In order to understand the conditions better, we determined the adsorption isotherm
by measuring the concentration changes in the alcohol upon adsorption onto glass particles
using a differential refractometer. Figure 3 plots the range of the attraction vs. the ethanol
concentration, together with the apparent adsorption layer thickness estimated from the ad-
sorption isotherm, assuming that only ethanol is present in the adsorption layer . For
0.1 mol% ethanol, half the distance where the long-range attraction appears, 18 Ϯ 2 nm, is
close to the apparent layer thickness of the adsorbed ethanol, 13 Ϯ 1 nm. This supports our
interpretation that the attraction is caused by contact of opposed ethanol adsorption layers.
Half the attraction range is constant up to ~0.4 mol% ethanol and decreases with increas-
ing ethanol concentration, while the apparent adsorption layer thickness remains constant
at all concentration ranges studied. The discrepancy between the two quantities indicates a
change in the structure of the ethanol adsorption layer at concentrations higher than ~0.4
FIG. 2 Interaction forces between glass surfaces upon compression in ethanol–cyclohexane mix-
tures. The dashed and solid lines represent the van der Waals force calculated using the nonretarded
Hamarker constants of 3 ϫ 10Ϫ21
J for glass/cyclohexane/glass and 6 ϫ 10Ϫ21
J for glass/ethanol
mol%. The structures of the adsorbed ethanol turned out to be hydrogen-bonded clusters,
via the study employing FTIR-ATR spectroscopy.
FTIR-ATR spectra were recorded on a Perkin Elmer FTIR system 2000 using a TGS
detector and the ATR attachment from Grasby Specac. The ATR prism made of an oxi-
dized silicon crystal was used as a solid adsorbent surface because of its similarity to glass
surfaces. Immediately prior to each experiment, the silicon crystal was treated with water
vapor plasma in order to ensure the formation of silanol groups on the surfaces. Obtained
spectra have been examined by referring to well-established, general spectral characteris-
tics of hydrogen-bonded alcohols in the fundamental OH stretching region, because ethanol
is known to form hydrogen-bonded dimers and polymers (clusters) in nonpolar liquids .
We have also experimentally examined hydrogen-bonded ethanol cluster formation in bulk
cyclohexane–ethanol mixtures using transmission infrared spectroscopy.
FTIR-ATR spectra of ethanol in cyclohexane at various ethanol concentrations
(0.0–3.0 mol%) are presented in Figure 4. At 0.1 mol% ethanol, a narrow negative band at
, a weak absorption at 3640 cmϪ1
(free OH), and a broad strong absorption
) with shoulders at 3530 cmϪ1
(cyclic dimer or donor end OH), 3450, and
are observed. It is known that the isolated silanol group exhibits an absorption
band at 3675–3690 cmϪ1
in a nonpolar liquid, e.g., CCl4 and when the silanol groups hy-
drogen bond with esters, the absorption band shifts to a lower wavenumber (3425–3440
) . Thus, the negative absorption at 3680 cmϪ1
and the positive shoulder at 3450
should correspond to the decrease in the isolated silanol groups and the appearance
of the silanol groups hydrogen bonded with the adsorbed ethanol, respectively. The strong
broad band ascribed to the polymer OH appeared at 3600–3000 cmϪ1
together with the rel-
atively weak monomer OH band at 3640 cmϪ1
. This demonstrated the cluster formation of
ethanol adsorbed on the silicon oxide surface even at 0.1 mol% ethanol, where no polymer
peak appeared in the spectrum of the bulk solution at 0.1 mol% ethanol. With increasing
ethanol concentration, the free monomer OH (3640 cmϪ1
) and the polymer OH peak (3330
) increased, while the peaks at 3530, 3450, and 3180 cmϪ1
remained the same.
Surface Forces Measurement 5
FIG. 3 Plots of half the range of attraction (see Fig. 2) and the apparent thickness of the ethanol ad-
sorption layer vs. the ethanol concentration.
At higher ethanol concentrations, ATR spectra should contain the contribution from
bulk species, because of the long penetration depth of the evanescent wave, 250 nm. To ex-
amine the bulk contribution, the integrated peak intensities of polymer OH peaks of trans-
mission (ATS) and ATR (AATR) spectra are plotted as a function of the ethanol concentration
in Figure 5. The former monitors cluster formation in the bulk liquid, and the latter contains
contributions of clusters both on the surface and in the bulk. A sharp increase is seen in AATR
FIG. 4 FTIR-ATR spectra of ethanol on a silicon oxide surface in ethanol–cyclohexane binary liq-
uids at various ethanol concentrations: 0.0, 0.1, 0.3, 0.5, 1.0, and 2.0 mol%.
FIG. 5 Plots of integrated peak intensities of polymer OH (3600–3000 cmϪ1
) as a function of the
ethanol concentration. Filled circles represent the value obtained from the transmission spectra (ATS),
while filled squares represent those from ATR (AATR).
even at 0.1 mol% ethanol, but no significant increase is seen in ATS at ethanol concentra-
tions lower than 0.5 mol%. A comparison of ATS and AATR clearly indicated that ethanol
clusters formed locally on the surface at concentrations of ethanol lower than ~0.5 mol%,
where practically only a negligible number of clusters exist in the bulk. The thick adsorp-
tion layer of ethanol most likely consists of ethanol clusters formed through hydrogen
bonding of surface silanol groups and ethanol as well as those between ethanol molecules.
A plausible structure of the ethanol adsorption layer is presented in Figure 6.
The contact of adsorbed ethanol layers should bring about the long-range attraction
observed between glass surfaces in ethanol–cyclohexane mixtures. The attraction starts to
decrease at ~0.5 mol% ethanol, where ethanol starts to form clusters in the bulk phase. It is
conceivable that the cluster formation in the bulk influences the structure of the adsorbed al-
cohol cluster layer, thus modulating the attraction. We think that the decrease in the attrac-
tion is due to the exchange of alcohol molecules between the surface and the bulk clusters.
A similar long-range attraction associated with cluster formation has been found for
cyclohexane–carboxylic acid mixtures and is under active investigation in our laboratory.
Such knowledge should be important for understanding various surface-treatment pro-
cesses performed in solvent mixtures and for designing new materials with the use of
molecular assembling at the solid–liquid interfaces. For the latter, we have prepared poly-
mer thin films by in situ polymerization of acrylic acid preferentially adsorbed on glass sur-
IV. ADSORPTION OF POLYELECTROLYTES ONTO
OPPOSITELY CHARGED SURFACES
The process of adsorption of polyelectrolytes on solid surfaces has been intensively stud-
ied because of its importance in technology, including steric stabilization of colloid parti-
cles [3,4]. This process has attracted increasing attention because of the recently developed,
sophisticated use of polyelectrolyte adsorption: alternate layer-by-layer adsorption  and
stabilization of surfactant monolayers at the air–water interface . Surface forces mea-
surement has been performed to study the adsorption process of a negatively charged poly-
mer, poly(styrene sulfonate) (PSS), on a cationic monolayer of fluorocarbon ammonium
amphiphille 1 (Fig. 7) .
A force–distance curve between layers of the ammonium amphiphiles in water is
shown in Figure 8. The interaction is repulsive and is attributed to the electric double-layer
Surface Forces Measurement 7
FIG. 6 Plausible structure of the adsorption layer composed of ethanol clusters.
force. The addition of 0.7 mg/L PSS (1.4 ϫ 10Ϫ9
M, equivalent to the addition of 0.7 nmol
of PSS, which is close to the amount of the amphiphile on the surface) into the aqueous
phase drastically alters the interaction. Here, the molecular weight (Mw) of PSS is 5 ϫ 105
Over the whole range of separations from 5 to 100 nm, the force decreases more than one
order of magnitude and does not exceed 100 N/m. The analysis of the force profile has
shown that more than 99% of the initial surface charges are shielded by PSAS binding. The
FIG. 7 Chemical structures of fluorocarbon ammonium amphiphile 1 and poly(styrene sulfonate)
FIG. 8 Force–distance dependence for surfaces covered with fluorocarbon amphiphile 1 in pure
water (1) and in aqueous solutions containing 0.7 mg/L poly (styrenesulfonate) (2) and 7.0 g/L poly
(styrenesulfonate) (3). The molecular weight of the polymer is 5 ϫ 105
. Lines are drawn as a visual
thickness of the adsorbed layer of PSS is in the range of 1.5–2.5 nm (it is less than 1 nm in
the case of PSS of 1 ϫ 104
Mw). These data indicate flat and stoichiometric adsorption of
the polyelectrolytes onto the monolayer surface (Fig. 9a).
Increased concentration of PSS at 7.0 g/L (1.4 ϫ 10Ϫ5
M) leads to an increase in the
force to value seven higher than that between the surfaces of fluorocarbon monolayers
alone. The origin of this force is electrostatic in nature. Recharging of the surface by addi-
tional adsorption of PSS should occur as shown in Figure 9b.
Our results demonstrate well the complexities of polyelectrolyte adsorption and pro-
vide a basis for various surface treatments utilizing polyelectrolytes. They especially afford
physical-chemical support for alternate layer-by-layer film formation of polyelectrolytes,
which is becoming a standard tool for building composite polymer nano-films in advanced
V. POLYPEPTIDE AND POLYELECTROLYTE BRUSHES
Polypeptides and polyelectrolytes are essential classes of substances because of their im-
portance in such areas as advanced materials science (functionalized gel) and biology (pro-
teins, living cells, and DNA). Being polymers with charges and counterions and/or hydro-
gen bonding, they exhibit interesting, albeit complicated, properties. Two-dimensionally
organized brush structures of polymers can simplify the complexities of the polyelectrolyte
solutions. Attempts to investigate polyelectrolyte brushes have been carried out experi-
mentally [11,28–32] and theoretically [33,34]. Direct measurement of surface forces has
been proven useful in obtaining information about the concrete structures of polypeptide
and polyelectrolyte brush layers. Taking advantage of the LB method, we prepared well-
defined brush layers of chain-end-anchored polypeptides and polyelectrolytes [11,28–30].
Surface Forces Measurement 9
FIG. 9 Schematic illustration of adsorption of poly(styrenesulfonate) on an oppositely charged sur-
face. For an amphiphile surface in pure water or in simple electrolyte solutions, dissociation of
charged groups leads to buildup of a classical double layer. (a) In the initial stage of adsorption, the
polymer forms stoichiometric ion pairs and the layer becomes electroneutral. (b) At higher polyion
concentrations, a process of restructuring of the adsorbed polymer builds a new double layer by ad-
ditional binding of the polymer.
We then investigated them based on the force profiles, together with FTIR spectra and sur-
face pressure–area isotherms by systematically varying the polymer chain length, chemical
structure, brush density, and solution conditions (pH, salt concentrations, etc). When the
surfaces of the opposed polymer layers approach to a separation distance of molecular di-
mensions, the steric repulsion becomes predominant and hence measurable. By analyzing
them, it is possible to obtain key parameters, such as thickness (length) and compressibil-
ity of polyelectrolyte layers, which are difficult to obtain by other methods, and to corre-
late them with polymer structures. Obtained information should form a basis for elucidat-
ing their properties and developing physical models. Moreover, it is more likely to discover
new phenomena via a novel approach: We have found the density-dependent transition of
polyelectrolyte brushes, which we have accounted for in terms of the change in the binding
modes of counterions to polyelectrolytes .
A. Brush Layers of Poly(glutamic acid) and Poly(lysine)
Polypeptides form various secondary structures (␣-helix, ␤-sheet, etc.), depending on so-
lution pHs. We have investigated end-anchored poly(L-glutamic acid) and poly(L-lysine) in
various secondary structures [11,29,35,36], using the analytical method for the steric force
FIG. 10 Schematic drawing of surface forces measurement on charged polypeptide brushes pre-
pared by LB deposition of amphiphiles 2 and 3.
in order to examine more quantitatively the structures and structural changes in polyelec-
trolyte layers. The elastic compressibility modulus of polypeptide brushes was obtained, to
our knowledge, as the first quantitative determination of the mechanical modulus of an ori-
ented, monomolecular polymer layer in solvents.
Poly(L-glutamic acid) and poly(L-lysine) brush layers were prepared using am-
phiphiles 2 and 3 carrying the poly(L-glutamic acid) (2C18PLGA(n), degree of polymer-
ization, n ϭ 21, 44, 48) and the ply(L-lysine) segment (2C18PLL(n), n ϭ 41), respectively
(Fig. 10). They formed a stable monolayer at the air–water interface in which different sec-
ondary structures, such as ␣-helix and ␤-structures, were formed through intra- and inter-
molecular hydrogen bonding, depending on surface pressure and subphase pH. They were
deposited onto mica surfaces and subjected to surface forces measurement. We used FTIR
spectroscopy to study the formation and orientation of their secondary structures.
Figure 11a shows a force–distance profile measured for poly(L-glutamic acid) brushes
(2C18PLGA(44)) in water (pH ϭ 3.0, 10Ϫ3
M HNO3) deposited at 40 mN/m from the wa-
ter subphase at pH ϭ 3.0. The majority of peptides are in the forms of an ␣-helix (38% de-
termined from the amide I band) and a random coil. Two major regions are clearly seen in
Surface Forces Measurement 11
FIG. 11 Force profiles between poly(glutamic acid), 2C18PLGA(44), brushes in water (a) at pH
ϭ 3.0 (HNO3), (b) at pH 10 (KOH) 1/ represents the decay length of the double-layer force. The
brush layers were deposited at ϭ 40 mN/m from the water subphase at pH ϭ 3.0 and 10, respec-
the force–distance profiles. At surface separations longer than 35 nm, the interaction is a typ-
ical double-layer electrostatic force, with a decay length of 10 Ϯ 1 nm, which agrees well
with the Debye length (9.6 nm) for 10Ϫ3
M HNO3, due to ionized carboxyl groups. At sep-
arations shorter than ~20 nm, the repulsion is steric in origin and varies depending on the
secondary structures existing in the surface layer. In order to examine detailed changes in
the interactions, a force–distance profile is converted to a stress–distance (P-D) profile by
differentiating the free energy of interaction Gƒ [Eq. (1)] between two flat surfaces as
P ϭ Ϫᎏ
ᎏ ϭ Ϫᎏ
The stress curve sharply increases when the steric component appears upon compression.
The initial thickness of a deformed layer is equal to be half the distance D0 obtained by ex-
trapolating the sharpest initial increase to stress zero. The value D0 is 21 Ϯ 1 nm, which is
close the thickness of two molecular layers (19.2 nm) of the ␣-helix brush, calculated us-
ing the CPK model and the orientation angles obtained by FTIR analysis. We have calcu-
lated the elastic compressibility modulus Y,
Y ϭ Ϫᎏ
to be 38 Ϯ 8 MPa from the steepest slope of the stress–distance curve. This value is one to
two orders of magnitude larger than the elasticity measured for a typical rubber (1 MPa).
Figure 11b shows a profile at pH 10, measured between the 2C18PLGA(44) LB sur-
faces prepared at 40 mN/m from the aqueous KOH subphase (pH 10). In this sample, two-
thirds of the carboxylic acid groups dissociate; therefore, it behaves as a simple polyelec-
trolyte. The initial thickness of the deformed layer is 35 Ϯ 2 nm, which is close to twice the
length of 2C18PLGA(44) in the extended form, 37 nm. The elastic compressibility modu-
lus is 0.2 Ϯ 0.1, which is even smaller than the value for a typical rubber. Unexpectedly,
the ionized layers are easily compressed. Counterion binding to the ionized chain should
play an important role in decreasing the stress for compression by reducing the effective
charges through the shielding and charge-recombination mechanisms.
Similar measurements have been done on poly(L-lysine) brushes. Table 1 lists a part
of our data, which display specific features: (1) The value D0 depends on the polymer chain
TABLE 1 Effective Length and Compressibility Modulus of Polypeptide Brushes Determined
by SFA in Water
␣-Helix content D0 Compressibility modulus, Y
Peptide pH R␣ (%) (nm) (MPa)
PLL (n ϭ 41) 10 34 16 Ϯ 1 1.2 Ϯ 0.6
(ionized chain) 11 47 19 Ϯ 1 3.1 Ϯ 0.8
12 54 14 Ϯ 1 3.3 Ϯ 0.8
4 0 32 Ϯ 1 0.14 Ϯ 0.05
PLGA (n ϭ 44) 3 38 21 Ϯ 1 38 Ϯ 8
(ionized chain) 5.6 32 22 Ϯ 1 22 Ϯ 5
10 0 35 Ϯ 2 0.2 Ϯ 0.1
PLGA (n ϭ 21) 9.6 0 25 Ϯ 2 0.2 Ϯ 0.1
The length D0 corresponds to twice the thickness of the brush layers.
length as well as the secondary structures; thus it is a good measure for determining the
thickness (length) of the polypeptide (or polyelectrolyte brush); (2) the compressibility
modulus is sensitive to changes in the kind of secondary structures; (3) the moduli of ␣-he-
lix brushes are one order of magnitude larger for poly(L-glutamic acid) than for poly(L-ly-
sine), which is likely due to interchain hydrogen bonding between the carboxylic acid
groups of neighboring poly(glutamic acid) chains; (4) the moduli of ionized chains are
identical for poly(L-glutamic acid) and poly(L-lysine).
The stress–distance profile measured by the surface forces apparatus thus provides
information on structural changes in polymers and polyelectrolytes in solvents. One ad-
vantage of our approach is that a model calculation is not necessary to extract physical pa-
rameters involved in the structural changes. One may note that the mechanical properties
discussed here reflect not only the intrinsic flexibility of polypeptide chains but also other
effects, such as the osmotic pressure of counterions present within charged brush layers.
Such knowledge is essential for the theoretical understanding of polyelectrolytes and
polypeptides. Our work employing surface forces measurement opens the door to studies
on a wide range of structural changes of polymers in solvents, including proteins and poly-
electrolyte networks in water. The complexities of their solution properties can be reduced
by aligning them in a two-dimensional manner. Very recently, polyelectrolyte brushes have
also begun to attract attention as a novel molecular architecture for nanotechnology .
The forces measurement should also provide valuable information for effectively design-
ing such materials.
B. Density-Dependent Transition of Polyelectrolyte Layers
The ionized forms of polypeptides exhibit many characteristics in common; therefore, we
have studied them under various conditions. The most interesting observation is the transi-
tion of a polyelectrolyte brush found by changing the polyelectrolyte chain density. The
brush layers have been prepared by means of the LB film deposition of an amphiphile,
2C18PLGA(48), at pH 10. Mixed monolayers of 2C18PLGA(48) and dioctadecylphos-
phoric acid, DOP, were used in order to vary the 2C18PLGA(48) content in the monolayer.
Surface force profiles between these polyelectrolyte brush layers have consisted of a
long-range electrostatic repulsion and a short-range steric repulsion, as described earlier.
Short-range steric repulsion has been analyzed quantitatively to provide the compressibility
12a). The modulus Y decreases linearly with a decrease in the chain density ⌫, and suddenly
increases beyond the critical density. The maximum value lies at ⌫ ϭ 0.13 chain/nm2
we have decreased the chain density further, the modulus again linearly decreased relative to
the chain density, which is natural for chains in the same state. The linear dependence of Y on
⌫ in both the low- and the high-density regions indicates that the jump in the compressibility
modulus should be correlated with a kind of transition between the two different states.
To examine this peculiar behavior, we have converted the elastic compressibility
modulus, per unit area, Y (Fig. 12a), to the modulus per chain, YЈ ϭ Y/1018
⌫ (Fig. 12b).
The elastic compressibility modulus per chain is practically constant, 0.6 Ϯ 0.1 pN/chain,
at high densities and jumps to another constant value, 4.4 Ϯ 0.7 pN/chain, when the den-
sity decreases below the critical value. The ionization degree, ␣, of the carboxylic acid de-
termined by FTIR spectroscopy gradually decreases with increasing chain density due to
the charge regulation mechanism (also plotted in Fig. 12b). This shows that ␣ does not ac-
count for the abrupt change in the elastic compressibility modulus.
Surface Forces Measurement 13
The density-dependent jump in the properties of polyelectrolyte brushes has also
been found in the transfer ratio and the surface potential of the brushes , establishing
the existence of the density (interchain distance)-dependent transition of polyelectrolytes in
The transition of the compressibility, and other properties of the polyelectrolyte
brushes, is most likely accounted for in terms of the transition in the binding mode of the
counterion to the polyelectrolytes, from the loosely bound state to the tightly bound one,
which reduces inter- and intrachain repulsive interactions. The following supports this ac-
count: (1) At the critical density, ⌫c ϭ 0.20 chain/nm2
, the separation distance between poly-
electrolyte chains, d, is 2.4 nm. This distance is close to the sum, 2.6 nm, of the chain diam-
eter, 1.3 nm, and the size of two hydrated Kϩ
counterions, 1.32 nm, indicating that the abrupt
change in the compressibility modulus should be closely related to the counterion binding
mode. (2) The critical distance satisfies the energy requirement for the tight binding of coun-
terions (coulombic interaction between two unscreened elementary charges is equal to the
thermal energy). (3) The stress profiles can be fitted to the theoretical equation derived based
on the assumption that the stress of deformation arises from the osmotic pressure of the coun-
terions. The analysis revealed that the osmotic coefficient in the high-brush-density region
is one order of magnitude larger than that in the low-density region. (4) At the critical chain
density ⌫c ϭ 0.2 chain/nm2
, we have found that the distance between the ionized charges
becomes close to twice the Bjerrum length . Therefore, counterions must bind strongly
to the polyelectrolytes at densities greater than the critical density.
In polyelectrolyte solutions, the counterion condensation on linear polyelectrolyte
chains is known to occur when the charge density along the chain exceeds the critical value
. Our work indicates the existence of a critical value for the separation distance between
chains, where the interchain interaction changes drastically, most likely due to the transi-
tion in the binding mode of the counterions (see Fig. 13). Many peculiar forms of behav-
ior, which are often interpreted by the cluster formation or the interchain organization of
polyelectrolytes, have been reported for high concentrations of aqueous polyelectrolytes
FIG. 12 Plots of elastic compressibility modulus (a) per unit area, Y; and (b) that per chain, YЈ, of
2C18PLGA(48) brushes as a function of chain density ⌫. The ionization degree of the carboxylic acid
group, ␣, is also plotted in part b.
. Our observation should be important in understanding these properties of polyelec-
trolytes in solutions and perhaps in gels.
VI. CONCLUDING REMARKS
The nanometer level of characterization is necessary for nanochemistry. We have learned
from the history of once-new disciplines such as polymer science that progress in synthe-
sis (production method) and in physical and chemical characterization methods are essen-
tial to establish a new chemistry. They should be made simultaneously by exchanging de-
velopments in the two areas. Surface forces measurement is certainly unique and powerful
and will make a great contribution to nanochemistry, especially as a technique for the char-
acterization of solid–liquid interfaces, though its potential has not yet been fully exploited.
Another important application of measurement in nanochemistry should be the characteri-
zation of liquids confined in a nanometer-level gap between two solid surfaces, for which
this review cites only Refs. 42–43.
1. JN Israelachivili. Intermolecular and Surface Forces. 2nd ed. London: Academic Press, 1992.
2. F Ohnesorge, G Binnig. Science 260:1451, 1993.
3. PC Heimenz, R Rajagopalan. Principles of Colloid and Surface Chemistry. New York: Marcel
4. AW Adamson, AP Gast. Physical Chemistry of Surfaces. 6th ed. New York: Wiley, 1997.
5. J-M Lehn. Supramolecular Chemistry. Weinheim, Germany: VCH, 1995.
6. B Alberts, D Bray, J Lewis, M Raff, K Roberts, JD Watson. Molecular Biology of the Cell. 3rd
ed. New York: Garlaud, 1994.
7. G Decher. Science 277:1232, 1997.
8. J Israelachvili, H Wennerstrom. Nature 379:219, 1996.
9. G Reiter, AL Demiral, S Granick. Science 263:1741, 1994.
10. (a) M Mizukami, K Kurihara. Chem Lett: 1005–1006, 1999; (b) M Mizukami, K Kurihara.
Chem Lett 248, 2000.
11. T Abe, K Kurihara, N Higashi, M Niwa. J Phys Chem 99:1820, 1995.
12. K Kurihara. Adv Colloid Sci 71–72:243, 1997.
Surface Forces Measurement 15
FIG. 13 Schematic drawing of possible binding modes of counterions to polyelectrolyte chains.
Counterions loosely bind and form a cloud around the polyelectrolyte chains when the interchain dis-
tance (d) is greater than 2.4 Ϯ 0.5 nm, while they strongly bind to form nearly neutral polyelectrolytes
at smaller distances (d Ͻ 2.4 Ϯ 0.5 nm).
13. NA Burnham, RJ Colton. In: DA Bonnel, ed. Scanning Tunneling Microscopy and Spec-
troscopy. New York: VCH, 1993, p 191.
14. VT Moy, EL Florin, HE Gaub. Science 266:257, 1994.
15. GV Lee, LA Chrisey, RJ Colton. Science 266:771, 1994.
16. M Rief, J Pascual, M Saraste, HE Gaub. J Mol Biol 186:553, 1999.
17. JN Israelachivili, GE Adams. J Chem Soc: Faraday Trans 74:975, 1978.
18. WA Ducker, TJ Senden, RM Pashley. Langmuir 8:1831, 1992.
19. K. Kurihara, T Kunitake, N Higashi, M. Niwa. Thin Solid Films 210/211:681, 1992.
20. JL Parker, P Claesson, P Attard. J Phys Chem 98:8468, 1994.
21. H Okusa, K Kurihara, T Kunitake. Langmuir 10:3577, 1994.
22. M Mizukami, K Kurihara, manuscript in preparation.
23. U Liddel, ED Becker. Spectrochim Acta 10:70, 1957.
24. AK Mills, JA Hockey. J Chem Soc, Faraday Trans 71:2398, 1975.
25. S Nakasone, M Mizukami, K Kurihara. 78th JCS Spring Annual Meeting: 2PA175, 2000.
26. M Shimomura, T Kunitake. Thin Solid Films 132:243, 1985.
27. P Berndt, K Kurihara, T Kunitake. Langmuir 8:2486, 1992.
28. K Kurihara, T Kunitake, N Higashi, M Niwa. Langmuir 8:2087, 1992.
29. K Kurihara, T Abe, N Higashi, M Niwa. Colloids Surfaces A 103:265, 1995.
30. T Abe, N Higashi, M Niwa, K Kurihara. Langmuir 15:7725, 1999.
31. Y Mir, P Auroy, L Auvray. Phys Rev Lett 75:2863, 1995.
32. P Goenoum, A Schlachli, D Sentenac, JW Mays, J Benattar. Phys Rev Lett 75:3628, 1995.
33. SJ Miklavic, SJ Marcelja. J Phys Chem 92:6718, 1988.
34. P Pincus. Macromolecules 24:2912, 1991.
35. T Abe. PhD thesis, Nagoya University, Nagoya, 1997.
36. S Hayashi, T Abe, K Kurihara. manuscript in preparation.
37. RR Shah, D Merreceyos, M Husemann, I Rees, NL Abbott, CJ Hawker, JL Hedrick. Macro-
molecules 33:597, 2000.
38. S Hayashi, T Abe, N Higashi, M Niwa, K Kurihara. MCLC: in press.
39. T Abe, S Hayashi, N Higashi, M Niwa, K Kurihara. Colloids Surfaces A 169:351, 2000.
40. GS Manning, Ber Bunsen-Ges. Phys Chem 100:909, 1996.
41. H Dautzenberg, W Jaeger, J Kotz, B Philipp, Ch Seidel, D Stscherbina. Polyelectrolytes. New
York: Hanser, 1994.
42. JN Israelachvili, PM McGuiggan, AM Homola. Science 240:189, 1988.
43. H-W Hu, S Granick. Science 258:1339, 1992.
44. C Dushkin, K Kurihara. Rev Sci Inst 69:2095, 1998.
Adhesion on the Nanoscale
SUZANNE P. JARVIS Nanotechnology Research Institute, National
Institute of Advanced Industrial Science and Technology, Ibaraki, Japan
I. ADHESION OVERVIEW
Sticky, one of our earliest childhood experiences and probably one of the first words to en-
ter our vocabulary, is familiar to scientists and nonscientists alike. However, does our di-
rect experience of stickiness, or scientifically speaking, adhesion, have any relevance at the
nanoscale? How can adhesion be measured, how can it be manipulated, and what role does
it play both in technological applications and intrinsically in nature? These are the ques-
tions that I will try to address in this chapter.
The adhesion of surfaces on a macroscopic scale is usually associated with specially
designed glues or tapes, which are a prerequisite for holding two dry, solid surfaces to-
gether. Exceptions to this tend to be very smooth surfaces with small amounts of moisture
between them, such as two sheets of glass or a rubber sucker on a bathroom tile. Intuitively
then, even on a macroscopic scale, it is apparent that surface roughness and environment
play a critical role in adhesion. Similarly, for many years scientists have realized that as the
surfaces approach nanoscale dimensions, the surface roughness and the area of contact
reach comparable dimensions, such that the apparent and true contact areas become ap-
proximately equal, as shown in Figure 1. This significantly increases the importance of ad-
hesion in the interaction between the two surfaces. Ultimately it becomes necessary to con-
sider the materials as ideal systems with properties no longer limited by defects, impurities,
and contamination, which dominate for bulk materials. When considering the material on
a near-atomic level, there are a number of attractive forces that can act between two sur-
faces brought into contact that can cause them to adhere to each other. The force required
to separate the two surfaces is then known as the adhesive force or pull-off force. The mag-
nitude of this force depends on the true contact area and the nature of the attractive forces
holding the surfaces together. These forces could include, for example, van der Waals, cap-
illary, or electrostatic forces. An excellent text explaining intermolecular and surface forces
in detail is that of Israelachvili .
There are a number of industrial and technological areas in which nanoscale adhesion
is important. One of the earliest fields concerned with adhesion on this scale was colloid
science. Colloid particles lie in the intermediate region between macro and nano, with di-
mensions typically of the order of hundreds of nanometers up to a few microns. This means
that their true contact areas lie well within the nano-domain and are influenced by interac-
tions on this length scale. Adhesion between such particles is important, due to its influence
on mineral separation processes and on the aggregation of powders, for example, on the
walls of machinery or in the forming of medical tablets. In an extraterrestrial context, such
processes are also important, because cosmic dust aggregation plays a role in planet for-
As physical structures used in technological applications have been reduced in size,
there has been an increasing need to understand the limiting processes of adhesion and to
try to minimize them. For example, adhesion due to humidity is known to have a major ef-
fect on the durability and friction forces experienced at the recording head/disk interface.
Microelectromechanical systems (MEMS) are also detrimentally affected by nanoscale ad-
hesion, with their motion being perturbed or prevented.
On a molecular level there are a number of aspects of adhesion that are important.
Preventing the infection of biocompatible materials by preventing bacterial adhesion is
very important in the medical industry, particularly for artificial heart valves as well as the
more commonly used contact lenses and dentures. A wider understanding of adhesion will
be required to support the current boom in biotechnology, with particular regard to molec-
ular motors and drug delivery systems. Adhesion and its manipulation may also lie at the
heart of many biological functions and recognition processes.
II. THEORETICAL PERSPECTIVE
A. Continuum Mechanics
Various continuum models have been developed to describe contact phenomena between
solids. Over the years there has been much disagreement as to the appropriateness of these
models (Derjaguin et al. [2–4] and Tabor [5–7]). Experimental verification can be complex
due to uncertainties over the effects of contaminants and asperities dominating the contact.
Models trying to include these effects are no longer solvable analytically. A range of mod-
els describing contact between both nondeformable and deformable solids in various envi-
ronments are discussed in more detail later. In all cases, the system of a sphere on a plane
is considered, for this is the most relevant to the experimental techniques used to measure
1. Nondeformable Solids
(a) In Vacuum. For smooth, ideal, rigid solids, the Derjaguin approximation  relat-
ing the force law between a sphere of radius R and a flat surface to the energy per unit area
W(D) between two planar surfaces separated by a distance D gives:
F(D)sphere ϭ 2RW(D)plane (1)
FIG. 1 Real and apparent contact areas. (a) macroscopic surfaces; (b) at the nanoscale.
This equation is useful in that it is applicable to any type of force law so long as the range
of interaction and the separation are much less than the radius of the sphere. Thus the force
to overcome the work of adhesion between a rigid sphere and a flat surface written in terms
of the surface energy ⌬␥ is:
Fpull-off ϭ 2R ⌬␥ where ⌬␥ ϭ ␥sphere ϩ ␥flat Ϫ ␥interface (Dupré) (2)
If the sphere and the flat surface are the same material, then:
␥interface ϭ 0 and ␥sphere ϭ ␥flat ϭ ␥s ⇒ Fpull-off ϭ 4R␥s (3)
This assumes that the only source of adhesion is the solid–solid contact.
(b) Forces Due to Capillary Condensation. For experiments conducted in air, the ad-
hesive force acting between the two bodies may be dominated by the presence of capillary
condensed water. These additional forces due to capillary condensation may be calculated
for smooth, ideal, rigid solids. For a sphere and a flat surface joined by a liquid bridge the
force F due to the Laplace pressure within the meniscus is given by :
F ϭ 4R␥LV cos (4)
where ␥LV is the surface tension of the liquid in the condensate and is the contact angle
of this liquid on the solid.
(c) Nondeformable Solids in Condensable Vapor. The capillary forces just discussed
act as an additional force; thus the force needed to separate a rigid sphere and a flat surface
of the same material joined by a liquid bridge is given by:
Fpull-off ϭ 4R␥LV cos ϩ 4R␥SL (5)
2. Deformable Solids
The foregoing models considered incompressible bodies; however, this is never the case in
practice. The following section discusses models that specifically consider contact between
(a) Hertz. For deforming solids, Hertzian analysis  gives the simplest approxima-
tion, for adhesive forces are ignored, i.e., no pull-off force and zero contact area for zero
applied load. Given an applied force, P, and a tip radius, R, the contact diameter, 2a, is
2a ϭ 2 ᎏ
E* ϭ ᎏ
ᎏ ϩ ᎏ
and E1, 1 and E2, 2 are Young’s modulus and Poisson’s ratio of the sphere and plane, re-
spectively. Because surface forces are neglected in this model, it is not possible to apply it
to adhesion measurements. However, it is included here because it is used as the basis for
some of the other models.
(b) Johnson, Kendall, and Roberts (JKR). The theory of Johnson, Kendall, and Roberts
 incorporates adhesion via the change in surface energy only where the surfaces are in
contact. This gives:
2a ϭ 2 Άᎏ
ᎏ (P ϩ 3R ⌬␥ ϩ ͙6ෆෆRෆPෆ ⌬ෆ␥ෆ ϩෆ (ෆ3ෆෆRෆ ⌬ෆ␥ෆ)2
Adhesion on the Nanoscale 19
where ⌬␥ is the work of adhesion. This is related to the force required to separate the sphere
and the flat surface after contact (known as the adhesive force) Fpull-off by the following
Fpull-off ϭ ᎏ
ᎏ R ⌬␥ (8)
Note from Eq.  that even at zero applied load there is a finite contact area due to
the adhesive forces alone, which is given by:
a0 ϭ ᎏ
Although this model seems to reflect well some experimental observations of contact
and separation [6,7] the assumptions made in its formulation are in fact unphysical. They
assume that the solids do not interact outside the contact region, whereas in reality electro-
static and van der Waals forces are nonzero at separations of several nanometers. The as-
sumptions made by JKR lead to infinite values of stress around the perimeter of the con-
necting neck between sphere and plane.
The model is found to be most appropriate for contact between low-elastic-moduli
materials with large radii when the work of adhesion is high. In comparison, the following
model assumes that the surface forces extend over a finite range and act in the region just
outside the contact. It is found to be more appropriate for systems with small radii of cur-
vature, low work of adhesion, and high modulus .
(c) Derjaguin, Muller, and Toporov (DMT). Derjaguin, Muller, and Toporov  as-
sume that under the influence of surface forces, the sphere will deform in the contact region
in accordance with the Hertzian model. Since the deformation is taken as Hertzian, the sur-
faces do not separate until the contact area is reduced to zero. At this instant, the pull-off
force is predicted to be:
Fpull-off ϭ 2R ⌬␥ (10)
This is the same as Eq. , the value for nondeformable solids in vacuum. However, in the
case of deformable solids, DMT theory gives a finite contact radius at zero applied load:
a0 ϭ ᎏ
Although the DMT theory attempts to incorporate distance-dependent surface inter-
actions into the adhesion problem, it does not take into account the effect surface forces
have on the elastic deformation. In other words, it does not predict the neck formation pre-
dicted by JKR.
(d) Maugis, Dugdale. Maugis  included a surface force, which acts both inside and
outside the contact perimeter. The attractive interaction is assumed to be constant up to a
separation of ho, at which point it falls to zero abruptly. The value of ho is defined such that
the maximum attractive force and the work of adhesion correspond to a Lennard–Jones po-
tential, the Dugdale approximation . The error of this somewhat arbitrary approxima-
tion is only apparent at low values of the elastic parameter lambda, ,
where oho ϭ w, the work of adhesion, R is the radius, and E* is the combined modulus as
defined in Eq. (6). Even for low values of the discrepancy in the elastic compression is
less than the atomic spacing. The net force acting on the contact is composed of a Hertzian
pressure associated with the contact radius and an adhesive Dugdale component extending
beyond the contact region up to a second radius. This radius can be found by solving two
(e) Muller, Yushchenko, and Derjaguin (MYD)/Burgess, Hughes, and White (BHW).
Muller et al.  and Burgess et al. [in 15] have formulated more complete descriptions for
the adhesion between a sphere and a plane by allowing the solid–solid interaction to be a
prescribed function of the local separation between the surfaces. The complete solution is
mathematically complex and cannot be solved analytically. As a consequence it is rarely
applied to experimental data. In any case, the uncertainties inherent in experiments of this
type mean that a more precise model often gives very little advantage over the JKR and
From the previous sections it is clear that there are a number of different possible
models that can be applied to the contact of an elastic sphere and a flat surface. Depending
on the scale of the objects, their elasticity and the load to which they are subjected, one par-
ticular model can be more suitably applied than the others. The evaluation of the combina-
tion of relevant parameters can be made via two nondimensional coordinates and P .
The former can be interpreted as the ratio of elastic deformation resulting from adhesion to
the effective range of the surface forces. The second parameter, P, is the load parameter and
corresponds to the ratio of the applied load to the adhesive pull-off force. An adhesion map
of model zones can be seen in Figure 2.
As discussed by Johnson and Greenwood , in principle both a full numerical
analysis using the Lennard–Jones potential and the Maugis analysis using the Dugdale ap-
proximation apply throughout the map with the Hertz, JKR, DMT and rigid zones being re-
gions where some simplification is possible. However, in practice both analyses become
less appropriate at high values of (in the JKR zone).
Unfortunately, to determine the appropriate zone of an adhesion experiment requires
knowledge of a number of parameters, which may not be easily accessible, particularly for
Adhesion on the Nanoscale 21
FIG. 2 Map of the elastic behavior of bodies. (Reprinted from Ref. 16.) Copyright 1997 by Aca-
nanoscale experiments. Such parameters include the radius of the probe or the combined
modulus, which can be affected even by traces of contamination. In the literature, the most
popular models tend to be DMT and JKR. The induced errors from the incorrect use of such
models are usually less than the errors encountered in the experimental determination of the
relevant parameters on the nanoscale.
(f) Liquid Bridge. When working in the presence of condensable vapors (i.e., water un-
der ambient conditions), it is possible for an annulus of capillary condensate to form as the
tip approaches the surface of the sample, as shown in Figure 3. Capillary forces can then
dominate the interaction between the sphere and plane. The general formula for the adhe-
sive force between the two due to capillary condensation is given by :
F ϭ F(⌬P) ϩ Fsolid–solid ϩ F(␥LV) (13)
where F(⌬P) is the force due to the Laplace pressure ⌬P within the meniscus, Fsolid–solid is
the force due to direct solid–solid interaction, and F(␥LV) is the resolved force due to the
liquid/vapor surface tension. Theory gives [5,18]:
F(⌬P) ϭ 4R␥LV cos (14)
Fsolid–solid ϭ 4R␥SL (15)
F(␥LV) ϭ 2R␥LV sin sin( ϩ ) (16)
where R is the radius of the sphere, is the solid/liquid contact angle, ␥SL is the solid/liq-
uid interfacial free energy, ␥LV is the liquid/vapor interfacial free energy, and is as shown
in Figure 3. In the case of deformable solids, Eq.  may be replaced by Fsolid–solid ϭ
3R␥SL. The preceding equations rely on the liquid’s acting in a macroscopic, hydrody-
namic manner. If the contact angle is assumed to be small and the radius of the sphere
greatly exceeds the neck radius of the liquid condensate such that is also small, then sin
sin( ϩ ) Ӷ 1. Thus the direct surface tension contribution is negligible when compared
to the Laplace pressure contribution.
When the sphere and plane are separated by a small distance D, as shown in Figure
4, then the force due to the Laplace pressure in the liquid bridge may be calculated by con-
sidering how the total surface free energy of the system changes with separation :
Utotal ϭ ␥SV (A1 ϩ A2) Ϫ ␥SL (A1 ϩ A2) Ϫ ␥LVA3 ϩ ␥SVAsolids
FIG. 3 Schematic of a liquid meniscus around the sphere.
where ␥SV is the solid/vapor interfacial free energy and A corresponds to the area of each
interface as defined in Figure 3.
If A1 Ϸ A2 Ϸ R2
, then for small ,
Utotal ϭ 2R2
(␥SV Ϫ ␥SL) ϩ constant ϩ smaller terms
But ␥SV Ϫ ␥SL ϭ Ϫ␥L cos (Young equation), so
Utotal Ϸ Ϫ2R2
␥L cos ϩ constant
F ϭ Ϫᎏ
ᎏ ϭ 4R2
␥L cos ᎏ
Calculating the volume by considering a straight-sided meniscus and assuming that the liq-
uid volume remains constant, Eq.  then becomes:
F ϭ ᎏ
Use of the condition of constant meniscus volume is most appropriate when growth and
dissolution of the meniscus is comparatively slow. An alternative is to consider the Kelvin
equilibrium condition. The Kelvin equation relates the equilibrium meniscus curvature
(also known as the Kelvin radius) to the relative vapor pressure; and if Kelvin equilibrium
is maintained during the separation process, then the adhesive force becomes :
F ϭ Ϫ4R␥LV cos 1 Ϫ ᎏ
ᎏ (Kelvin condition of constant rm) (19)
where the meniscus curvature,
rm ϭ ᎏ
ᎏ ϩ ᎏ
and r1 and r2 are as defined in Figure 4.
Adhesion on the Nanoscale 23
FIG. 4 Schematic of a liquid meniscus between the sphere and plane at finite separation.
In the limit of small surface separations, the adhesive force and its gradient tend to
the same value for both the constant-meniscus-volume and Kelvin-equilibrium conditions.
The derivations of the foregoing equations have been based on the principles of ther-
modynamics and the macroscopic concepts of density, surface tension, and radius of cur-
vature. They may therefore cease to be appropriate as the mean radius of curvature ap-
proaches molecular dimensions.
Including capillary condensation with the Hertz approximation, as considered by
Fogden and White , introduces pressure outside the contact area; i.e., adhesion enters
the problem nonenergetically through the tensile normal stress exerted by the condensate
in an annulus around the contact circle. The resulting equations cannot be solved analyti-
cally; however, their asymptotic analysis may be summarized as follows.
For nonadhering bodies in contact in the presence of capillary condensation, the pre-
vious result for rigid solids is found to apply more generally to systems of small, hard, but
deformable spheres in contact in vapor near saturation:
Fpull-off ϭ 4R␥LV for ϭ 0 (20)
In the limit of large, softer solids in vapor pressure closer to the value marking the
onset of capillary condensation, the generalized Hertz and the original JKR theories are
found to be qualitatively identical. However, the contact area for zero applied load will in
general be different, since it is dependent upon the nature of the source of adhesion:
Fpull-off ϭ 3R␥LV and a0 ϭ ΄9␥LV ᎏ
Thus (1/2)⌬␥ in the JKR approximation is replaced by ␥LV. The elastic modulus affects the
contact area but not the adhesion force.
B. Molecular Dynamics and First-Principles Calculations
Molecular dynamics (MD) permits the nature of contact formation, indentation, and adhe-
sion to be examined on the nanometer scale. These are computer experiments in which the
equations of motion of each constituent particle are considered. The evolution of the sys-
tem of interacting particles can thus be tracked with high spatial and temporal resolution.
As computer speeds increase, so do the number of constituent particles that can be consid-
ered within realistic time frames. To enable experimental comparison, many MD simula-
tions take the form of a tip-substrate geometry corresponding to scanning probe methods
of investigating single-asperity contacts (see Section III.A).
One of the earliest molecular dynamics simulations to be compared directly to
atomic-scale measurements was the work of Landman et al. . They used MD simu-
lations to investigate the intermetallic adhesive interactions of a nickel tip with a gold
surface. The interatomic interactions of the nickel–gold system were described by many-
body potentials, which were obtained using the embedded-atom method [22,23]. Long-
range interactions, such as van der Waals forces, were neglected in the calculations. The
theoretically calculated force versus tip–sample separation curves showed hysteresis,
which was related to the formation, stretching, and breaking of bonds due to adhesion,
cohesion, and decohesion. This was characterized by adhesive wetting of the nickel tip
by the gold atoms. The wetting was instigated by the spontaneous jump to contact of the
sample and the tip across a distance of approximately 4 Å. In this case, mechanical in-
stabilities of the materials themselves caused the jump (dependent on the cohesive
strength of the material), and it should not be confused with the jump-to-contact phe-
nomenon often observed experimentally due to compliant measurement systems. Subse-
quent indentation resulted in plastic yielding, adhesion-induced atomic flow, and the gen-
eration of slip in the surface region of the gold substrate. The whole cycle of approach
and retraction can be seen in Figure 5. A more detailed evaluation of the wetting phe-
nomenon was obtained by considering pressure contours. In this way the atoms at the pe-
riphery of the contact were found to be under an extreme tensile stress of 10 GPa. In fact,
both the structural deformation profile and the pressure distribution found in the MD sim-
ulations were similar to those described by contact theories that include adhesion (see
Section II.A). On retraction of the tip, a conductive bridge of gold atoms was formed, the
elongation of which involved a series of elastic and plastic yielding events, accompanied
by atomic structural rearrangements. The eventual fracture of the neck resulted in gold
atoms remaining on the bottom of the nickel tip.
Subsequently, Landman and Luedtke inverted the system to look at a gold (001) tip
adhering to a nickel (001) surface  and noted a number of differences. In this case the
tip atoms jumped to the surface, instead of the surface atoms coming up to meet the tip.
Continuing the approach beyond this point resulted in the compression of the tip, wetting
of the surface, and the formation and annealing of an interstitial layer dislocation in the core
of the tip. Also, on retraction of the tip, a much more extended neck was drawn out between
tip and sample, in this case, consisting solely of gold-tip atoms. This occurred regardless of
whether or not additional force was applied to the contact region after the initial jump to
The study of adhesive contacts was further extended by Landman et al. to include in-
terionic CaF2 interactions and intermetallic interactions mediated by thin alkane films .
In the case of CaF2 the bonding is significantly different from that in the metallic systems
studied previously. The authors treated the long-range coulombic interactions via the
Ewald summation method and controlled the temperature at 300 K. As the tip and surface
were brought together, the tip elongated to contact the surface in an analogous way to the
Adhesion on the Nanoscale 25
(a) (b) (c)
(d) (e) (f)
FIG. 5 Atomic configurations generated by the MD simulation. (Courtesy David Luedtke.)
jump to contact observed in metallic systems but involving smaller displacements of ap-
proximately 0.35 Å. Retracting the tip after a small net repulsive interaction resulted in
clear hysteresis in the force–distance curve, as seen in Figure 6. This reflected the plastic
deformation of the crystalline tip leading to eventual fracture. After the final pull-off, part
of the tip remained fixed to the substrate surface. For their investigation of the effect of thin
films, the authors considered n-hexadecane (C16H34), which they modeled using the inter-
action potentials developed by Ryckaert and Bellmans . This was used as the mediat-
ing layer between a nickel tip and a gold (001) substrate. Before the onset of the interac-
tion, the molecular film was layered on the gold substrate, with the layer closest to the
surface exhibiting a high degree of orientational order parallel to the surface plane. The ini-
tial stages of the interaction involved some of the alkane molecules adhering to the tip, fol-
lowed by flattening of the film and partial wetting of the sides of the tip. Continued reduc-
tion of the tip–sample distance induced drainage of the second layer of molecules under the
tip and the pinning of the final layer. Subsequently, drainage of these molecules from un-
der the tip was assisted by transient local inward deformations of the gold, which would
seem to lower for the relaxation of unfavorable conformations of the alkane molecules.
Finally, true tip–sample contact occured via displacement of gold-surface atoms toward
Harrison et al. use MD  simulations to investigate diamond–diamond adhesion,
which occurs due to covalent bond formation (they neglect long-range interactions such
as van der Waals). Typical MD interaction times are orders of magnitude faster than
those typically used experimentally. However, they are still sufficiently slow to permit
equilibration of the system. In the case of hydrogen-terminated diamond (111) tip and
surface, a nonadhesive interaction was observed for an indentation with an applied load
of 200 nN. This ceased to be the case if the applied load was increased to 250 nN, for
plastic yielding of the diamond occurred and, consequently, adhesion between the tip and
sample. For a non-hydrogen-terminated diamond surface and a hydrogen-terminated tip,
adhesion was observed at both applied loads. Further, precise positions of the tip and sur-
face atoms could be evaluated during the transition to contact and subsequent retraction,
exposing different indentation and fracture mechanisms in the two cases. The main pre-
dictions of the simulation were that breaking the symmetry of the system and increasing
the applied load both had the effect of facilitating adhesion. In addition, removing the hy-
FIG. 6 Calculated force on the tip atoms for a CaF2 tip approaching () and subsequently retract-
ing (ϫ) from a CaF2 (111) surface. The distance from the bottom layer of the tip to the topmost sur-
face layer for the points marked by letters is: A (8.6 Å), B (3.8 Å), C (3.0 Å), D (2.3 Å), E (1.43 Å),
F (2.54 Å), G (2.7 Å), and H (3.3 Å). (Reprinted from Ref. 25, copyright 1992, with permission from
drogen termination from the surface caused adhesive behavior to set in at smaller values
of applied load, a feature they attributed to the increased surface energy of the diamond
crystal when the hydrogen was removed. So far, it has proved extremely difficult to
mimic such precision experimentally.
However, it has been observed experimentally that even low levels of impurities can
have a strong influence on the adhesion of free surfaces and that they usually have the ef-
fect of lowering adhesion. Zhang and Smith  performed fully self-consistent, first-prin-
ciples calculations to investigate the cause of such impurity influence on an Fe–Al inter-
face. They found an asymmetry in the adhesion modification, depending on whether the
impurities originated on the Fe or the Al surface, as shown in Figure 7. Clearly, the effects
on the work of adhesion were much larger when the impurity was added to the Fe surface,
due to the much larger adhesive energies of the impurity monolayers to the Fe surface. The
effect of the impurity atoms were twofold: On one hand they pushed the surfaces apart, thus
weakening the Fe/Al bonds; on the other, they formed bonds across the interface. Of these
two competing effects, the former was always found to overwhelm the latter for their cho-
sen systems of five different contaminant layers. This impurity lowering of adhesion has
been found to be typical, both experimentally and theoretically.
Adhesion on the Nanoscale 27
FIG. 7 Total energy per cross-sectional area as a function of interfacial separation between Fe and
Al surfaces for the clean interface and for monolayer interfacial impurity concentrations of B, C, N,
O, and S. Graph (a) is for the case where the impurity monolayer is applied to the free Al surface prior
to adhesion, while graph (b) has the impurity monolayer applied to the free Fe surface prior to adhe-
sion. The curves fitted to the computed points are from the universal binding energy relation. (From
Ref. 28. Copyright 1999 by the American Physical Society.)
III. EXPERIMENTAL TECHNIQUES
A. Atomic Force Microscopy
One of the most popular methods of measuring adhesion on the nanoscale is to operate an
atomic force microscope in “spectroscopy mode.” The microscope consists of a small
lever, of submillimeter dimensions, that is usually microfabricated. At the end of this can-
tilever is a sharp tip used to probe the interactions with a given sample. The lever on which
the tip is mounted will change its position and also its apparent resonant frequency and
stiffness according to the interaction with the surface. These changes can be sensitively
measured by detecting either the static or dynamic motion of the lever with subangstrom
sensitivity. This is usually done via either optical interferometry or the more popular
method of optical beam deflection. The latter method, although inferior to the former in
terms of sensitivity and ease of calibration, dominates, due to its ease of use and its
widespread popularity among commercial AFM manufacturers. With this system, light
from a laser diode or light-emitting diode is reflected from the back of the cantilever onto
a four-segment photodiode. As the cantilever bends, the reflected beam moves across the
photodiode, thus altering the relative voltages from each segment. It is then possible to an-
alyze this to give normal and torsional motion of the lever, as seen in Figure 8.
The final component of the AFM is the tip–sample approach mechanism. Bringing
the probe tip and the sample within interaction range in a nondestructive manner involves
mounting one surface on an approach mechanism with subangstrom sensitivity. Most ap-
proach devices normally combine some form of stepper or slider with single step sizes of
the order of microns, with a piezo device providing displacements of the order of one tenth
of an angstrom up to a maximum of several microns. In microscopy mode, the piezo device
FIG. 7 Continued.
makes relative lateral displacements between the probe and the sample in order to build up
an image of the surface as a function of a particular value of applied force. However, it is
the “spectroscopy mode” that is of most interest for measuring adhesion. In this case the tip
and the sample are brought into intimate contact and the surfaces separated by applying
small, known displacements to one of the surfaces. Simultaneously, the motion of the lever
is recorded, resulting in a curve of the form shown in Figure 9. The attractive force in the
Adhesion on the Nanoscale 29
FIG. 8 Schematic of an atomic force microscope with optical beam deflection detection showing a
typical angle of 10° between lever and sample.
FIG. 9 Curve of sample motion versus lever motion. The experimental starting point for generat-
ing a force versus distance curve.
unloading part of the force–distance curve reflects the adhesion between the tip and the
sample. From this it is possible to extract the adhesion force from the pull-off distance us-
ing Hooke’s law. Thus, the force is derived by multiplying together the displacement and
the cantilever stiffness.
There are some exceptions to this method. For example, it is possible to apply forces
directly to the end of the cantilever rather than displacements to the sample in order to con-
trol the approach and separation of the two surfaces [29,30]. This more direct method re-
duces unwanted relative lateral motion between the tip and the surface. The application of
direct forces in this way also has alternative uses, such as enabling sensitive dynamic mea-
surements to be made .
One point, which is often disregarded when using AFM, is that accurate cantilever
stiffness calibration is essential, in order to calculate accurate pull-off forces from mea-
sured displacements. Although many researchers take values quoted by cantilever manu-
facturers, which are usually calculated from approximate dimensions, more accurate meth-
ods include direct measurement with known springs , thermal resonant frequency curve
fitting , temporary addition of known masses , and finite element analysis .
The lever deflection must also be calibrated. In the case of interferometric detection,
the calibration is readily at hand from the wavelength of the light used. In the case of the
optical beam method, the best calibration technique is less obvious. The most usual method
is to apply a large displacement to the sample via the piezo, with the tip and sample in con-
tact. If the lever is very compliant and the contact very stiff, then the lever and sample can
be assumed to be moving together, and thus the deflection sensor can be calibrated from
the known sample displacement. There are clearly a number of errors with this method, in-
cluding deformation in the contact region and unintentional lateral sliding of the tip on the
sample. An alternative calibration method for small displacements relies on the high-reso-
lution imaging of materials exhibiting a stepped structure of known height.
1. Relationship Between Pull-Off and Adhesion
The magnitude of the pull-off force depends on the nature of the tip–sample interaction dur-
ing contact. Adhesion depends on the deformation of the tip and the sample, because at-
tractive forces are proportional to the contact area. Quantifying the work of adhesion is dif-
ficult. The measured magnitude of ⌬␥ is strongly dependent on environment, surface
roughness, the rate of pull-off, and inelastic deformation surrounding the contact.
An important consideration for the direct physical measurement of adhesion via pull-
off measurements is the influence of the precise direction of the applied force. In AFM the
cantilever does not usually lie parallel to the surface, due to the risk that another part of the
cantilever chip or chip holder will make contact with the surface before the tip. Another
problem relates to the fact that the spot size in the optical beam deflection method is usu-
ally larger than the width of the lever. This can result in an interference effect between the
reflection from the sample and the reflection from the cantilever. This is reduced if the can-
tilever and sample are not parallel. Most commercial AFM systems use an angle in the
range of 10°–15° between the sample and the cantilever. Depending on this angle and the
extent to which the cantilever is bent away from its equilibrium position, there can be a sig-
nificant fraction of unintentional lateral forces applied to the contact.
Chang and Hammer  investigated this issue theoretically by simulating the de-
tachment of receptor-coated hard spheres from ligand-coated surfaces using normal, tan-
gential, and shear forces after allowing the particles to bind to steady state. They point out
that we should not expect different types of force to be equally effective at removing the
spherical particles and attempt to quantify this by an adhesive dynamics computer simula-
tion method. The interacting molecules were regarded as linear springs, with their rate of re-
action dependent on the distance from the substrate. They found that tangential forces were
transmitted to the bonds at the interface much more efficiently than normal forces, due to
the small ratio of bond length to bead radius, causing a large axial strain in the former case.
Tangential forces were found to be 20 times more effective for detaching a single bond and
56 times more effective when there were many bonds, as demonstrated in Figure 10. Were
the radius of the spherical particle to be reduced, the difference between the normal and tan-
gential forces to break the adhesive bonds would also become smaller. The authors make
some harsh conclusions related to AFM. They show that if the force is to be exerted normal
to the substrate, then it must be controlled precisely, for deviations as small as 10° can sig-
nificantly alter the measured value of the adhesion. They also recommend that for systems
where the sphere–substrate separation is small, the calculation of average force per bond
made by dividing the pull-off force by the number of bonds in the interface should be abol-
ished, because the method produces very misleading results. Actually, a related problem was
also observed with surface forces apparatus (see Section III.B). Christenson  found that
the use of double cantilevers to reduce relative torsional movement of the two surfaces re-
sulted in larger measured adhesion values than with a single cantilever structure.
Stuart and Hlady  found that unintentional lateral forces influenced their mea-
surements of adhesion between surface-bound protein molecules and colloid probe–bound
ligands. They noted a greatly exaggerated separation distance and a stick-slip behavior in
their adhesion curves, which they attribute to rolling and buckling of the cantilever under
the influence of lateral forces as the sample was retracted with the probe stuck to it.
A similar concern due to lateral motion also applies to purely qualitative imaging.
Sugisaki et al.  observed an artifact in their simultaneously topographic and adhesion
Adhesion on the Nanoscale 31
FIG. 10 Critical force needed to rupture all the bonds as a function of ␤, the angle at which the force
is exerted. Simulation results are given by solid circles. Tangential critical force TCF/cos ␤ is also
plotted for comparison (solid line). (Reprinted with permission from Ref. 35. Copyright 1996 Amer-
ican Chemical Society.)
images as a consequence of lateral sliding of the tip on the surface during the retraction of
the surface. They showed schematically how the two images, although obtained almost in-
stantaneously, actually correspond to different points on the surface due to this lateral slid-
ing of the tip. The shift between the two points depended on the amount of cantilever de-
flection. The deflection depends on the stiffness of the cantilever and the strength of the
adhesion. Thus, for compliant levers on highly adhesive samples, the lateral sliding during
retraction is likely to be large. They do not discuss how this lateral sliding could also in-
fluence the quantitative measurement of adhesion.
2. Adhesive Force Measurements with AFM
Within the first five years after its invention, many adhesion measurements with AFM
were motivated by the fact that adhesion was seen to be limiting the imaging resolution
of the instrument. High adhesive forces were often observed that totally dominated the
interaction between the tip and the sample. The additional load from these surface forces
was sufficient in some cases to prevent the study of delicate biological samples. Weisen-
horn et al.  investigated adhesion between a microfabricated tip and a mica surface
in air and water and found adhesive forces larger than 100 nN while operating in air.
These could be reduced by two orders of magnitude by operating in water. They also saw
a two-stage pull-off in air, which they attributed to tip–sample separation followed by
A subsequent and more detailed study by Weisenhorn et al.  investigated a Si3N4
tip on mica and a tungsten carbide tip on metallic foils in various liquids. While in ethanol
the adhesion force of both systems was sub-nN and reproducible, results in water were
much more erratic. The metallic surfaces showed high adhesion in water, more than an or-
der of magnitude larger than that measured in ethanol, which could be due to the adsorp-
tion of alkane contaminants in air prior to the experiment. These would be removed by the
solvent action of ethanol but not by water. The influence of applied load on the adhesion
force was also investigated for a Si3N4 tip on mica in water. Increasing the applied load
from 0.35 nN to 5.3 nN increased the measured pull-off force from 0.2 nN to 0.7 nN. Due
to the compliance of the levers in these measurements (k ϭ 0.035 N/m), applying such
loads involved large displacements of the sample in contact with the tip. Due to the can-
tilever’s lying at an angle to the surface (typically 10°–15° in most commercial systems),
it is likely that a significant displacement would also have occurred in the lateral direction
in the form of sliding of the tip across the surface. Such effects were not usually considered
in early AFM measurements.
Operating in dry nitrogen, Burnham et al.  measured the adhesive forces between
a large tungsten tip (R ϭ 2000 nm) and a variety of surfaces, including mica, graphite, and
two thin films of CH3(CH2)16COOH and CF3(CH2)16COOH. Using a double cross-can-
tilever designed to limit the motion of the tip to the direction perpendicular to the sample,
they compared measured adhesive forces with sample surface energies as determined from
advancing contact angle measurements. The high-surface-energy samples showed the ex-
pected square-root dependence on surface energy, but with lower magnitude, probably due
to the influence of microasperities on the tip. The lower-surface-energy materials, such as
the stearic acid films, exhibited significant deviation from the square-root dependence,
which may be related to mechanical-property differences. Interestingly, a large difference
in adhesive force was reported between the –CH3 and –CF3 terminated thin films, even
without any special treatment of the tip, suggesting chemical sensitivity without chemical
modification of the probe (see Section III.A.6).