Advances in Colloid and Interface Science
104 (2003) 53–74
Polyelectrolytes as adhesion modifiers
Per M. Claessona,b,*, Andra Dedinaitea,b, Orlando J. Rojasa,b,c
a
Department of Chemistry, Surface Chemistry, Royal Institute of Technology,
¨ 51, SE-100 44 Stockholm, Sweden
Drottning Kristinas vag
b
Institute for Surface Chemistry, Box 5607, SE-114 86 Stockholm, Sweden
c
´ Quımica,
´
´
Escuela de Ingenierıa
Lab. FIRP, Universidad de Los Andes, Merida
5101, Venezuela
Abstract
Adsorbed layers of polyelectrolytes have been studied with atomic force microscopy
(AFM) and the interferometric surface force apparatus (SFA). Particular emphasis was put
on determining the effect of the polyelectrolyte charge density on surface topography, and
the effect of the polyelectrolyte coating on the adhesive properties. The AFM was employed
to image individual polymer chains at low adsorption densities and to characterize the layer
topography and coverage at higher adsorption densities. The adhesive properties between
two polyelectrolyte-coated surfaces in air were determined as a function of the number of
contacts made at any given spot. The data provide evidence for formation of electrostatic
bridges, particularly when highly charged polyelectrolytes are used. Further, material transport
between the surfaces is observed when the polyelectrolyte is either highly charged or have a
very low charge density. For intermediate charge densities we could not observe any
indication of material transfer. The adhesion between one polyelectrolyte-coated surface and
one bare surface was initially higher than that between two polyelectrolyte-coated surfaces.
However, due to material transfer between the two surfaces the adhesion decreased
significantly with the number of times that the surfaces were driven into contact. For the
polyelectrolytes of the lowest charge density the results suggest that entanglement effects
contribute to the adhesive interaction. The modification of the adhesion by polyelectrolytes
in practical systems such as in the case of dry-strength additives to improve paper resistance
is also considered.
䊚 2003 Elsevier Science B.V. All rights reserved.
Keywords: Adhesion; Polyelectrolytes; AFM imaging; Surface forces; Dry-strength additives; Paper
strength
*Corresponding author. Tel.: q46-8790-9972; fax: q46-8208-998.
E-mail address: [email protected] (P.M. Claesson).
0001-8686/03/$ - see front matter 䊚 2003 Elsevier Science B.V. All rights reserved.
doi:10.1016/S0001-8686(03)00036-8
54
P.M. Claesson et al. / Advances in Colloid and Interface Science 104 (2003) 53–74
1. Introduction
Polyelectrolytes are commonly used as additives to control colloidal stability and
adhesive properties of surfaces. One classic example of the latter is the use of
cationic polyelectrolytes as retention aids and as dry and wet strength additives in
paper making. For example, the pretreatment of clay particles with cationic
poly(ethylene imine) increases the clay deposition on pulp fibers at pH)4 w1x. This
is rationalized in terms of attractive electrostatic interactions. The dry strength of
paper is often increased by addition of cationic polyelectrolytes such as cationic
starch to the fiber furnish, which is subsequently dried. The cationic polymer
adsorbs to the negatively charged fibers and mediates an increased fiber–fiber bond
w2x. It has been reported that the dry strength of the paper increases with decreasing
charge density of the polymer w2x, presumably due to increased polymer–polymer
interpenetration and due to increased viscoelastic losses that occur during the rupture
of the paper sheet under strain.
In recent years surface measurements have been used to investigate the interactions
between polyelectrolyte-coated surfaces in aqueous solutions w3–6x. From such
studies much have been learned about how the polyelectrolyte charge density w7,8x,
ionic strength w9–11x, substrate characteristics w12–14x and order of addition of
polyelectrolytes and other components w15,16x influence the adsorbed layer structure
and the resulting surface forces. The shear forces between polyelectrolyte-coated
surfaces have also been quantified w17x. Conversely, very little fundamental work
has been done on the adhesion between polyelectrolyte-coated surfaces in air.
The theory of adhesion is bounded to many confusions usually related to the
process of bond-forming and bond-breaking which have a very different nature w18x.
Pioneering work by Johnson et al. w19x and Derjaguin et al. established some basis
for understanding the influence of molecular forces on adhesion between solid
bodies, especially for the case of an elastic particle and a rigid substrate w20–24x.
The effect of contact deformations on the adhesion of particles has also been
considered w25x. Nevertheless, dissipation effects and rate-dependent adhesion has
not been discussed to a large extent. Note, however, the work by Plunkett et al.
w26x.
The case of adhesion between solid bodies with adsorbed layers of surfactant and
polymers is even more difficult to characterize and understand. Most of the work
aims to solve practical problems in biological and pharmaceutical systems wsee for
example, w27,28x or for instance, in mineral processing and cellulosic (papermaking)
systemsx. In many cases the subject has been viewed considering detachment of
particles under different flow conditions w29,30x.
Interaction forces as measured by the surface force apparatus and the atomic
force microscope have been used to investigate adhesion phenomena between
surfactant or polymer layers adsorbed on solid surfaces. An interesting fact is that
adhesion in air between two monolayer-coated surfaces depends on atmospheric
conditions such as the relative humidity or the presence of organic vapors w31x. It
is worth noting that Ruths and Granick have described detailed work on the rate-
P.M. Claesson et al. / Advances in Colloid and Interface Science 104 (2003) 53–74
55
Fig. 1. Molecular structure of the monomer AM (left) and MAPTAC (right).
dependent adhesion between surfaces coated with polymer and surfactant layers in
air w32,33x.
The purpose of the present study is to investigate the effect of thin polyelectrolyte
layers on the adhesion force between two polyelectrolyte-coated surfaces and
between one polyelectrolyte-coated surface and one bare surface. We are particularly
interested in the effect of the polyelectrolyte charge density. Atomic force microscopy
has been employed to characterize the preadsorbed layers and the interferometric
surface force apparatus has been utilized for the adhesion force measurements.
2. Materials and methods
The polyelectrolytes used in this investigation were random copolymers of
uncharged acrylamide (AM) and positively charged w3-(2-methylpropionamido)propylxtrimethylammonium chloride (MAPTAC) (see Fig. 1). By balancing the
ratio AMyMAPTAC in the (free radical) copolymerization process, macromolecules
with different charge densities (or percentage molar ratio of cationic monomers)
were synthesized and kindly provided by the Laboratoire de Physico-Chimie
Macromoleculaire (Paris). In Table 1 a list of the investigated polyelectrolytes along
with their charge density (or cationicity, t) and molecular weight is provided. For
convenience, the polyelectrolytes are referred to as ‘AM-MAPTAC-X’ where X is a
number that indicates the percentage of charged segments.
2.1. Substrate preparation
Muscovite mica from Reliance Co. (NY) was used as substrate. Before any
experiment, pieces of mica were cleaved several times on both sides in a laminar
56
P.M. Claesson et al. / Advances in Colloid and Interface Science 104 (2003) 53–74
Table 1
Charge density (t) and molecular weight of the investigated polyelectrolytes (data provided by Laboratoire de Physico-Chimie Macromoleculaire, Universite´ Pierre et Marie Curie, Paris)
Polyelectrolyte
AM-MAPTAC-1
AM-MAPTAC-10
AM-MAPTAC-30
AM-MAPTAC-100
t (%)
Theoretical
Elemental
analysis
Potentiometry
NMR
1
10
30
100
0.5
10
31
99
1
9
31
95
–
8.9–9.5
24.2–25.6
–
MW
(molyg)
900 000
1 000 000
780 000
480 000
flow cabinet until an adequate thickness was obtained. All the employed tools were
previously cleaned, and protective clothing and gloves were worn to minimize
contamination of the high-energy mica surfaces.
Adsorption of the studied copolymers of AM-MAPTAC onto oppositely charged
mica planar surfaces was allowed in aqueous solutions at pH 5 at 20 8C. In a
typical experiment an aqueous solution of the polymer (at 0.1 or 20 ppm
concentration) was freshly prepared by dilution with pure water of approximately
700 ppm polyelectrolyte stock solution. Freshly cleaved mica pieces (5=2 cm)
were then immersed in the solution (contained in a 25-ml glass beaker) for
approximately 6 h during which equilibrium adsorption is reached. At the end of
this period, the substrate was withdrawn from the polymer solution and immersed
in a large vessel filled with pure water in order to remove any non-adsorbed
polymer. Previous studies have demonstrated that the polyelectrolytes adsorb strongly
to most oppositely charged surfaces (e.g. silica, mica, gold) and that hardly any
desorption takes place upon immersion of the polyelectrolyte-coated substrate in
pure water w8,14,34,35x. To reduce the risk of Langmuir–Blodgett deposition at the
three-phase line, the respective liquid surface was aspirated (using a Pasteur pipette
connected to a water-jet pump) prior to any substrate transfer between the air and
the liquid phases. Finally, the substrate was placed vertically inside a glass hood in
a laminar flow cabinet and left to dry overnight. The water used in all experiments
was obtained by using a Milli-Q Plus 185 unit.
2.2. Atomic force microscopy
Imaging of adsorbed polymer on mica was accomplished using a Nanoscope III
MultiMode娃 scanning probe microscope (Digital Instruments, Santa Barbara, CA,
USA). The tapping mode imaging technique was used to study the sample’s
topography by probing the surface with an oscillating tip. In this mode a piezo
stack excites a cantilever vertically causing the tip to oscillate near its resonant
frequency w36x. Close to the sample surface the tip tends to be deflected due to its
interaction with the surface material. However, the amplitude of the oscillation is
kept constant by a feedback loop that adjusts the vertical position of the sample
(using a piezoelectric tube on which the sample is mounted). The vertical position
P.M. Claesson et al. / Advances in Colloid and Interface Science 104 (2003) 53–74
57
of the scanner (z) at each (x,y) location is thus used to reveal the topographic
image of the sample surface.
The probe used in the experiments consisted of 125 mm length, single-beam
cantilever and a tip (5–10 nm nominal radius of curvature) as an integrated
assembly of single crystal silicon produced by etching techniques (TappinngMode
etched silicon probe model TESP). The spring constant and resonant frequency are
reported by the manufacturer to be 20–100 Nym and 290–346 kHz, respectively.
Before imaging, the substrate was cut and gently pressed onto a sticky tab on a
15-mm-diameter metal disk that was then attached to a magnetic sample holder on
the piezoelectric scanner (Model AS-12V, ‘E’ vertical MultiMode SPM scanner).
The polyelectrolyte-coated surfaces were imaged in dry air since the use of ambient
conditions may introduce a layer of adsorbed water molecules on the sample surface
that gives rise to extra capillary forces between the tip and the sample (which may
interfere with the AFM experiment) w37,38x. In order to achieve controlled relative
humidity and temperature, nitrogen from a humidity generator (Model RH-100, VTI
Corporation) was flowed continuously through the instrument head (where the
sample and probe are enclosed). The temperature and relative humidity were set in
all cases at 20 8C and approximately 0%, respectively.
2.3. Surface force apparatus, SFA
A surface force apparatus, Mark II model w39x, was used for the measurement of
adhesion forces. The mica substrate surfaces were glued onto optically polished
half-cylindrical silica discs using Epon 1004. The surfaces were mounted in the
measuring chamber in a crossed cylinder configuration and the adhesion force and
the contact position were determined. The measured force in crossed-cylinder
geometry is, according to the Derjaguin approximation w40x, equivalent to the force
between a sphere (with the same radius as the geometric mean of the cylinder radii)
and a flat surface, provided that the radius of the surfaces (f2 cm in our set-up)
is much larger than the range of the surface forces. This condition is fulfilled in our
set-up.
The relative humidity in the measuring chamber was controlled to close to zero
by placing a beaker with P2O5 within the sealed chamber. Next, the surfaces were
removed from the measuring chamber and immersed in a beaker containing an
aqueous 20 mgyml polyelectrolyte solution and no added extra salt. The polyelectrolyte was allowed to adsorb for 1 h before the surfaces were transferred into a
beaker of water and then immediately removed and dried with a gentle stream of
dry nitrogen gas. The samples were remounted in the measuring cell and allowed
to equilibrate with the P2O5-dried atmosphere for 30 min prior to determination of
the adhesion force. To avoid excessive shearing a relatively stiff double cantilever
spring was employed. The spring constant was approximately 7500 Nym, and
accurately determined after each experiment. The surfaces were separated slowly
and the negative load was increased step-wise until the surfaces jumped apart. Thus,
unlike Ruths and Granick w32x in their work on surfactant and polymer coated
58
P.M. Claesson et al. / Advances in Colloid and Interface Science 104 (2003) 53–74
surfaces, we did not attempt to study the effect of the rate of separation on the
adhesion force.
The surface separation and the shape of the surfaces were determined using white
light interferometry. White light is directed perpendicularly from below towards the
surface, and multiple-reflected between the silvered backsides of the mica surfaces.
Only wavelengths that interfere constructively will be able to exit from the optical
cavity. Fringes of equal chromatic order, FECO, are generated and analyzed in a
spectrometer. From the shape and position of the FECO the surface separation and
the shape of the surfaces can be determined.
3. Results and discussion
3.1. AFM-images
Tapping mode AFM images in air of adsorbed copolymers of AM and MAPTAC
on mica from solutions with polymer concentration of 0.1 ppm are shown in Fig.
2. Typical height profiles for the cases of polyelectrolyte charge densities of 1 and
100% are also illustrated in Fig. 3. The low polymer concentration used in these
experiments was chosen so as to avoid surface saturation and allow the resolution
of ‘individual’ adsorbed polymer chains. From large scan images (not shown) it
was noticed that the AM-MAPTAC-100 polymer adsorbs uniformly onto the mica
surface. However, as the polyelectrolyte charge density is reduced the number
density of adsorbed molecules increased and some ‘aggregates’ (or ‘patches’) of
polymer molecules are formed (see Fig. 2). The increase in the adsorbed amount
for polymers of lower cationicity is explained by the fact that in such cases, more
polymer units are needed to compensate the surface charge. This observation is
corroborated by XPS quantitative measurements on similar systems w8x.
It was observed that as the polyelectrolyte charge density was reduced it became
more difficult to image the adsorbed polymer. We suggest that this is due to the
fact that high charge-density polyelectrolytes are more strongly bound to the
oppositely charged surface than the low-charge density ones.
The section profiles shown in Fig. 3 indicate an apparent chain thickness of
approximately 0.2–0.7 nm in agreement with a very flat adsorbed layer structure.
Furthermore, it is noticed that the AM-MAPTAC-1 copolymer has a smaller
‘thickness’ (approx. 0.2 nm) compared to that of AM-MAPTAC-100 (approx. 0.5–
0.7 nm). This observation is consistent with a ‘bulkier’ AM-MAPTAC-100 as
anticipated from the molecular structure of this polymer. The reported values for
the polymer thickness can, however, only be taken as approximate since under the
present conditions molecular resolution is limited by tip broadening, and this effect
also precludes the determination of the width of the adsorbed polymer chain.
Fig. 4 shows AFM height and phase images for AM-MAPTAC-100 adsorbed on
mica and scanned over an area of 400=400 nm2 in size. This figure provides some
insight of the polymer conformation and metrics upon adsorption. From the
molecular weight of AM-MAPTAC-100 a large end-to-end distance was expected,
but only some fragments of the polymer chains are often seen. Possible explanations
P.M. Claesson et al. / Advances in Colloid and Interface Science 104 (2003) 53–74
59
Fig. 2. Tapping mode image (0% relative humidity) of adsorbed polyelectrolytes on mica after equilibrium adsorption in 0.1 ppm aqueous polymer solution (scan range 1=1 mm): (a) AM-MAPTAC-1; (b)
AM-MAPTAC-10; (c) AM-MAPTAC-30; and (d) AM-MAPTAC-100.
for this is that the polymer conformation is locally coiled, lowering the end-to-end
distance, or that the AFM-imaging process in fact results in some breakage of the
polymer chain.
Tapping mode images of adsorbed polyelectrolytes on mica after equilibrium
adsorption at a higher polymer concentration (20 ppm) (as compared to the
respective cases in Fig. 2a–d) are displayed in Fig. 5a–d. It is evident that at this
concentration the surface becomes almost fully saturated and inter-chain and intrachain entanglements take place. Hence, under these circumstances the resolution of
individual polymer molecules becomes very difficult. It is, however, observed that
60
P.M. Claesson et al. / Advances in Colloid and Interface Science 104 (2003) 53–74
Fig. 3. Height profiles for a horizontal scan of the tapping mode images of AM-MAPTAC-1 (a) and
AM-MAPTAC-100 (b) (under the same conditions as in Fig. 2). Typically, the height images yielded
maximum thickness (with respect to the mica base line) of approximately 0.2 and 0.5–0.7 nm for AMMAPTAC-1 and AM-MAPTAC-100, respectively.
at high charge density the polymers adsorb more uniformly (or orderly) whereas at
lower charge densities more disordered structures are predominant. Thus, the
presence of polymer aggregates or ‘patches’ occurs more often as the charge density
is reduced. We note that the layer thickness cannot be determined from these AFM
images since no large enough uncoated patches are present on the surface.
Intermediate polymer concentrations (data not shown) demonstrated that surface
saturation is reached at rather low polymer concentrations. For example, the surface
coverage was very similar when the polyelectrolyte was adsorbed from a 1 ppm
bulk concentration as compared to the 20 ppm case. This observation confirms the
high-affinity characteristic of the adsorption isotherm as reported for similar systems
w8 x .
A difficulty in the application of AFM to imaging of adsorbed polymer layers is
illustrated in Fig. 6 obtained after scanning the tip over a smaller area. Here, the
edge of a square ‘window’ that corresponds to the accumulation of polymeric chains
in the slow-scanning direction occurs because the adsorbed polymers are partly
P.M. Claesson et al. / Advances in Colloid and Interface Science 104 (2003) 53–74
61
Fig. 4. Tapping mode height (left) and phase (right) images (0% relative humidity) of adsorbed AMMAPTAC-100 polyelectrolytes on mica after equilibrium adsorption in 0.1 ppm aqueous polymer solution (scan range 400=400 nm).
dragged by the tip in the fast-scanning direction. This problem mainly occurs when
imaging adsorbed low-charge-density polyelectrolytes. In this case, the polymer is
attached to the surface by relatively few charged segments, and in solution most of
the polymer chains extend out from the surface in the form of loops and tails w11x.
These extended structures eventually collapse on the surface during drying, but
nevertheless they provide a less strong anchoring compared to the charged segments.
3.2. The mica surface and adsorbed amount of polyelectrolyte
Muscovite mica is a layered aluminosilicate mineral. For a detailed description
w41x. The important aspect from the point of view of
¨
of the mica crystal see Guven
this work is that each aluminosilicate layer is negatively charged due to isomorphous
substitution of aluminum for silicon. This charge is compensated by small ions,
mainly potassium, located between the aluminosilicate sheets. The amount of
negative charges in the topmost lattice layer is 2.1=1018 my2. Consequently, this
is the number of exchangeable cations at the mica basal plane. The cationic
polyelectrolytes thus adsorb to the mica surface through an ion-exchange process.
For highly charged polyelectrolytes the exchange is close to stoichiometric whereas
for low charge density polyelectrolytes steric repulsion between the adsorbed
polyelectrolyte chains limits the adsorption. The adsorbed amount of the polyelectrolytes studied in this work has been quantified by means of XPS w8,42x and found
to be 2.3–2.5 mgym2 for AM-MAPTAC-1, 1.5–1.7 mgym2 for AM-MAPTAC-10,
1.1–1.2 mgym2 for AM-MAPTAC-30 and 0.6–0.8 mgym2 for AM-MAPTAC-100.
62
P.M. Claesson et al. / Advances in Colloid and Interface Science 104 (2003) 53–74
Fig. 5. Tapping mode image (0% relative humidity) of adsorbed polyelectrolyte on mica after equilibrium
adsorption in 20 ppm aqueous polymer solution (scan range 1=1 mm): (a) AM-MAPTAC-1; (b) AMMAPTAC-10; (c) AM-MAPTAC-30; and (d) AM-MAPTAC-100.
3.3. Adhesion measurements in dry air
For soft, large, homogeneous and elastic surfaces in a crossed cylinder geometry
the pull-off force (F) normalized by the geometric mean radius (R) is related to the
interfacial energy (g) as w19x:
FyRs3pg
(1)
where it has been assumed that the contact between the surfaces is ideal, i.e. not
associated with any excess free energy (e.g. due to surface roughness). This last
assumption is hardly ever valid, and even for such smooth substrates as mica the
P.M. Claesson et al. / Advances in Colloid and Interface Science 104 (2003) 53–74
63
Fig. 6. Atomic force microscope image of adsorbed AM-MAPTAC-1 (20 ppm concentration) on mica
(scan range 1=1 mm2). The edge of a ‘window’ that corresponds to a previously scanned area of smaller
size (400=400 nm2) is clearly seen.
pull-off force reported in the literature varies considerably w43x. This is partly due
not only to the fact that the relative orientation of the mica crystal in the two
surfaces plays a role w44x, but also due to adsorption of water vapor and small
amounts of other contaminants w43x. Further, Eq. (1) is valid only under equilibrium
conditions, i.e. rate-dependent viscoelastic effects invalidate the equation. We note
that the SFA set-up consists of a layered system, glass-glue-mica, which complicates
the adhesion mechanics as discussed by Sridhar et al. w45x. The normalized pull-off
forces measured between mica surfaces in dry air in our laboratory typically falls
in the range 800–1200 mNym, consistent with the values reported by Christensson
for the same conditions. During the separation process the contact radius, r, of the
flat surface region decreases, and at the point of separation the value of ryr0
decreased to a value of 0.6–0.7, where r0 is the contact radius under zero load.
This is in agreement with predictions based on the JKR-theory w19x.
The normalized pull-off forces measured between two mica surfaces coated with
a preadsorbed polyelectrolyte layer are shown in Figs. 7–10. It should be noted that
the fact that the polyelectrolytes were adsorbed outside the measuring chamber
64
P.M. Claesson et al. / Advances in Colloid and Interface Science 104 (2003) 53–74
Fig. 7. The normalized pull-off force as a function of the number of times the surfaces have been
separated from each other at a given contact spot. The measurements were carried out in dry air. Open
squares represent the situation with both mica surfaces being coated by a layer of AM-MAPTAC-100.
The filled squares represent measurements of pull-off forces between one mica surface coated with AMMAPTAC-100 and one bare mica surface. Results for two different contact positions are illustrated for
each case. The polyelectrolytes were adsorbed from a 20 ppm aqueous polymer solution.
made it difficult to determine the absolute layer thickness with high accuracy. This
is due to the fact that measurements of the zero contact between bare mica surfaces
were determined on a different spot on the surfaces than that used for measuring
the layer thickness. However, the expected trend with an increase in layer thickness
Fig. 8. The normalized pull-off force as a function of the number of times the surfaces have been
separated from each other at a given contact spot. The measurements were carried out in dry air. Open
squares represent the situation with both mica surfaces being coated by a layer of AM-MAPTAC-30.
The filled squares represent measurements of pull-off forces between one mica surface coated with AMMAPTAC-30 and one bare mica surface. Results for two different contact positions are illustrated for
each case. The polyelectrolytes were adsorbed from a 20 ppm aqueous polymer solution.
P.M. Claesson et al. / Advances in Colloid and Interface Science 104 (2003) 53–74
65
Fig. 9. The normalized pull-off force as a function of the number of times the surfaces have been
separated from each other at a given contact spot. The measurements were carried out in dry air. Open
squares represent the situation with both mica surfaces being coated by a layer of AM-MAPTAC-10.
The filled squares represent measurements of pull-off forces between one mica surface coated with AMMAPTAC-10 and one bare mica surface. Results for two different contact positions are illustrated for
each case. The polyelectrolytes were adsorbed from a 20 ppm aqueous polymer solution.
Fig. 10. The normalized pull-off force as a function of the number of times the surfaces have been
separated from each other at a given contact spot. The measurements were carried out in dry air. Open
squares represent the situation with both mica surfaces being coated by a layer of AM-MAPTAC-1. The
filled squares represent measurements of pull-off forces between one mica surface coated with AMMAPTAC-1 and one bare mica surface. Results for two different contact positions are illustrated for
each case. The polyelectrolytes were adsorbed from a 20 ppm aqueous polymer solution.
66
P.M. Claesson et al. / Advances in Colloid and Interface Science 104 (2003) 53–74
from below 1 nm for the polyelectrolyte with the highest charge density (AMMAPTAC-100) to 2–3 nm for the polyelectrolyte with the lowest charge density
(AM-MAPTAC-1) was observed. Further, changes in layer thickness during repeated
measurements on the same spot could be determined accurately to approximately
0.2 nm. The layers formed by AM-MAPTAC-100 on mica are very thin and flat,
as observed from the shape of the interference fringes and more locally with the
AFM (see Figs. 3 and 4). The pull-off force measured during the first separation
cycle was very large, 2300 mNym, i.e. approximately twice as large as between the
uncoated mica surfaces. This strong increase in adhesion cannot be due to an
increase in the van der Waals force since the refractive index of mica is higher than
that of the polymer. Instead we suggest that the increased adhesion is due to
bridging, i.e. once the surfaces are in contact the conformation of the adsorbed
polymers may change to allow the segments of the same polymer to bind to both
surfaces. This change in conformation is entropically driven. We note that the
charges on the MAPTAC-polymer are located 0.7–0.8 nm away from the backbone
and by a simple rotation the charged groups may change their position in space by
1.5 nm. This flexibility is likely to facilitate the bridging process. We note that the
AFM images show that the mica surfaces are fully packed with polyelectrolytes,
but nevertheless some bridging do occur due to the entropically driven changes in
polymer conformation.
When the pull-off force is measured repeatedly at one and the same contact
position, a decrease in adhesion with the number of measurements is observed (two
different sets of data are shown in Fig. 7). However, the adhesion force remains
high, and the lowest value obtained, after 10 measurements, was 1700 mNym. The
decrease in adhesion force indicates that the act of repeatedly bringing together and
separating the surfaces affects the adsorbed layer. This is also observed as a slight
increase in layer thickness, by some 0.5–1 nm. Hence, we conclude that some
polymer molecules are, due to being bound to both surfaces, stretched out during
the separation process. When the bridges are broken the polyelectrolytes collapse
back onto the surfaces but they are unable to find an equally flat conformation as
prior to separation. The increased layer thickness makes formation of bridges slightly
more difficult during the subsequent contact, which explains the reduction in
adhesion force. Due to the strong attractive forces between the polyelectrolytecoated surfaces a large flat contact area is observed once the surfaces are in contact.
Upon separation the radius of the flat region decreased until ryr0 was in the range
0.65–0.75, which is slightly larger than the value of 0.63 expected from JKRtheory. It was also noted that the contact area did not shrink in a continuous fashion
with increasing negative load, but rather the contact radius decreased in a stepwise
fashion related to stick–slip friction behavior and adhesion hysteresis w46x.
Some experiments were also carried out using one mica surface coated with AMMAPTAC-100 and one bare mica surface. The results from two different experiments
are illustrated in Fig. 7. Clearly, the pull-off force measured during the first
separation is very high, approximately 4000 mNym. The large value of the adhesion
force is explained by extensive bridging between the polyelectrolyte-coated surface
P.M. Claesson et al. / Advances in Colloid and Interface Science 104 (2003) 53–74
67
and the bare mica surface. The adhesion force decreased with the number of contacts
and reached a plateau value of 2600–2700 mNym. Similarly, the layer thickness
increased by 0.6–0.8 nm. These results give strong support for the view that as the
surfaces are separated polymer molecules are transferred from the polyelectrolyte
coated surface to the initially bare surface, and as a result the number of bridges
formed are reduced. We note, however, that the adhesion force remains higher than
that between two mica surfaces coated with AM-MAPTAC-100, indicating that even
after the polyelectrolytes have been distributed between the two surfaces more
bridges are formed in the, at least initially, asymmetric system as a result of a less
complete surface coverage. Thus, it is clear that the surfaces should be less than
fully covered with polyelectrolytes if the goal is to obtain maximum adhesion.
However, from these measurements the optimum coverage cannot be determined.
Finally, we note that in the majority of the measurements the decrease in contact
radius prior to separation was close to that expected from JKR-theory, but in a few
cases a significantly smaller decrease in contact radius was observed.
The results for the adhesion force measurements between AM-MAPTAC-30
coated mica surfaces are displayed in Fig. 8. The pull-off force is significantly
smaller than for AM-MAPTAC-100 coated mica, and it is independent of the
number of measurements at a given spot. This indicates that the number of bridges
formed is significantly less and thus the layer is not disturbed by the measurement.
This conclusion is supported by the observation that the layer thickness is
independent of the number of measurements. The reason why less bridges are
formed as compared to AM-MAPTAC-100 is that less charged segments are present
and that the uncharged segments constitute a steric barrier counteracting the
reconformation needed in order for a polymer chain to cross from one surface and
bind to the other. In fact, the adhesion force observed between AM-MAPTAC-30
coated mica surfaces is rather similar to that between two bare mica surfaces,
indicating that the van der Waals force is a major cause for the observed adhesion.
In order to determine whether more bridges were formed if the surfaces were left
in contact for a prolonged time we made one experiment where the surfaces were
left in contact for 16 h prior to separation. The result was a 5% increase in adhesion
compared to the previous measurement with the typical short contact time (5–10
min). This increase is within the scatter of the measurements and one may conclude
that the contact time (t)5 min) does not affect the pull-off force significantly.
Finally, we note that the contact radius of the flat region decreased until the ratio
ryr0 reached a value of 0.7–0.8 at which point the surfaces jumped apart. This is
slightly larger than predicted by JKR-theory. One may speculate that this is due to
the discontinuous decrease in contact radius during the separation process and thus
related to stick–slip friction.
For the asymmetric system, one AM-MAPTAC-30 coated mica and one bare
mica surface, the adhesion during the first separation is significantly higher, 2200–
2100 mNym, than between two AM-MAPTAC-30 coated surfaces. This demonstrates that a significant number of bridges are formed. Just as for AM-MAPTAC-100
the pull-off force decreases with the number of contacts. However, for AM-
68
P.M. Claesson et al. / Advances in Colloid and Interface Science 104 (2003) 53–74
MAPTAC-30 the pull-off force after several contacts is lower for the asymmetric
case compared to for two AM-MAPTAC-30 coated surfaces. When inspecting the
FECO fringes after the first separation it was noted that the contact region no longer
appeared absolutely flat. Clearly, the redistribution of polyelectrolytes between the
two surfaces results in formation of a less homogeneous coating, and the increased
surface roughness contributes to the relatively low value of the adhesion force found
in the asymmetric system after the first separation. Support for this conclusion
comes from the observation that by leaving the surfaces in contact for a prolonged
time resulted in a more flat contact and an increase in the pull-off force by 50%.
As a result of the increased surface roughness we observed that the value of ryr0
decreased to a value of 0.6–0.4, slightly lower than expected by the JKR theory.
The tendency was that the contact radius decreased to lower values with increasing
number of separations, i.e. with increasing roughness of the polymer layer. A smaller
than expected contact radius at the point of separation in air has also been observed
for surfaces coated with uncharged polymers w32x.
The results for AM-MAPTAC-10, displayed in Fig. 9, are qualitatively similar to
those obtained for AM-MAPTAC-30. However, the adhesion forces observed, both
between two polyelectrolyte-coated surfaces and between one AM-MAPTAC-10
coated surface and a bare mica surface, is lower than for the more highly charged
polyelectrolytes. Clearly, the bridging mechanism becomes less important as the
charge density of the polyelectrolyte is decreased. The decrease in contact radius
prior to the jump out was found to be consistent with predictions of the JKR-theory
for the symmetric system. However, for the asymmetric system it was observed,
just as for AM-MAPTAC-30, that the contact became less homogeneous with
increasing number of contacts and the value of ryr0 decreased to lower values than
expected from the JKR theory.
Some measurements were also carried out using AM-MAPTAC-1. The pull-off
force as a function of the number of contacts is illustrated in Fig. 10. For two AMMAPTAC-1 coated surfaces we observe an adhesion force that is significantly larger
than for the AM-MAPTAC-10 coated surface. Further, a slight decrease in the pulloff force with increasing number of separations was observed. The decrease in
adhesion force was accompanied by an increased inhomogeneity of the contact
region. Clearly, unlike the situation with AM-MAPTAC-30 and AM-MAPTAC-10,
but similarly to AM-MAPTAC-100, we now have material transfer between the two
surfaces. Due to the low charge density of the adsorbed polymer, this cannot be
explained by formation of electrostatic bridges. Rather, we suggest that the layer
for the low charge density AM-MAPTAC-1 is less compact than for the other
polymers. This promotes the possibility of chain interpenetration and entanglement
between the two polymer layers. During the separation process, even though it is
carried out slowly, the polymer layers do not have time to disentangle but the chain
interlocking contributes to viscoelastic losses that increases the measured adhesion.
The resulting material transfer process is facilitated by the relatively low affinity
between AM-MAPTAC-1 and the mica surface due to the low number of electrostatic
bonds between the surface and the polyelectrolyte. The relatively low polyelectro-
P.M. Claesson et al. / Advances in Colloid and Interface Science 104 (2003) 53–74
69
Fig. 11. The normalized pull-off force between two polyelectrolyte-coated surfaces measured during the
first (filled squares) and fifth (open circles) separation as a function of the charge density of the polyelectrolyte. The polyelectrolytes were adsorbed from a 20 ppm aqueous polymer solution.
lyte–surface affinity for this system was also observed during the AFM-imaging
(see Fig. 6).
For the asymmetric system, AM-MAPTAC-1 coated mica vs. bare mica, the
adhesion during the first separation is lower than for the other polyelectrolytes. This
is a consequence of the lower number of electrostatic bridges formed. However,
material transfer does still occur due to attractive interactions between the uncharged
acrylamide segments and the bare mica surface. This is clearly observed as an
increase in layer thickness and a less homogeneous contact region with increasing
number of separations. The adhesion force after the first separation is higher than
between one AM-MAPTAC-10 and one bare mica surface. We attribute this to an
increased importance of viscoelastic losses due to chain entanglement.
The effect of the charge density of the polyelectrolyte on the pull-off force
between two polyelectrolyte-coated surfaces is summarized in Fig. 11 that shows
the mean pull-off force for the first and fifth separation, respectively. First, we note
that when the charge density is decreased from 100% to 10% the adhesion force
during the first separation is reduced. This is a result of a decreasing importance of
electrostatic bridges. However, when the charge density is decreased further to 1%,
the adhesion force is increased again, an observation that strongly indicates that
another contribution comes into play. We argue that the adhesion due to chain
interpenetration and entanglement increases as the charge density of the polyelectrolyte is decreased and at this very low charge density makes a significant contribution.
For the most highly charged polyelectrolyte and the lowest charge density polyelec-
70
P.M. Claesson et al. / Advances in Colloid and Interface Science 104 (2003) 53–74
Fig. 12. The normalized pull-off force between one polyelectrolyte-coated surface and one bare mica
surface measured during the first (filled squares) and fifth (open circles) separation as a function of the
charge density of the polyelectrolyte. The polyelectrolytes were adsorbed from a 20 ppm aqueous polymer solution.
trolyte the pull-off force decreases significantly with the number of contacts. The
fact that the layer thickness increases and the contact becomes less homogeneous
strongly suggests that the decrease in pull-off force is due to the fact that the layer
is disrupted during the separation process. For the intermediately charged polyelectrolytes, no indication of disruption of the polymer layer is observed, and conversely
the adhesion force is not strongly affected by the number of separations.
The effect of the charge density on the adhesion between one polyelectrolytecoated surface and one bare mica surface is summarized in Fig. 12. In this case,
the adhesion between the two surfaces decreases with the number of contacts for
all polyelectrolytes studied, and this is due to material transfer from the polyelectrolyte-coated surface to the initially bare surface. We further suggest that the relatively
low decrease in adhesion force for AM-MAPTAC-1 is due to a relatively large
contribution of polymer entanglements to the adhesion force.
We note that in our experiments the polyelectrolyte with the highest charge
density increases the adhesion in air the most whereas the opposite trend is observed
when polyelectrolytes are used as dry-strength additives w2x. This may be due to
several effects. First, the charge density of mica is larger than that of the pulp fiber,
which leads to a decreased importance of electrostatic bridges in the latter case.
Further, in our case the polyelectrolyte layer was preadsorbed and the surfaces were
dried before they were brought into contact. On the other hand, in real papermaking
systems these additives are added to the paper furnish (fiber suspension) which is
subsequently dried in the drier section (after water removal in the forming and press
sections). Thus, when paper is made the fibers are dried in a condition of close
proximity between each other due to surface tension and capillarity effects. In this
P.M. Claesson et al. / Advances in Colloid and Interface Science 104 (2003) 53–74
71
case, there are much larger chances for the adsorbed polymer layers to interpenetrate
and interlock. As a result, the viscoelastic losses during the eventual course of paper
rupture are expected to be much larger in the latter case. Such experiments can also
be carried out with the surface force apparatus, and we note that studies with the
SFA have shown that the adhesion between model cellulose surfaces that are brought
into contact in water and subsequently dried is larger by a factor of three than the
adhesion between dried cellulose surfaces w47x. This is due to increased chain
interpenetration that increases the viscoelastic losses occurring during the separation
process.
3.4. Comparison between adhesion forces between polyelectrolyte-coated surfaces
in air and water
There are several molecular contributions to the adhesion force. The van der
Waals attraction is always present, and this force contribution is approximately a
factor of five-fold larger in air than in water. This is a consequence of the fact that
the difference in dielectric properties of the surface and air is larger than between
the surface and water.
A second contribution comes from formation of electrostatic surface–polyelectrolyte–surface bridges. This effect increases with the number of bridges formed and
the strength of each bridge. Since the electrostatic force is reduced when the
dielectric constant is increased, the strength of each electrostatic bridge will be
higher in air than in water. On the other hand, the mobility of the polyelectrolyte
chain will be higher in contact with water as compared to the case of contact with
air, which may favor formation of more bridges in water. Furthermore, the
polyelectrolyte layer swells in water and this will oppose the formation of surface–
polyelectrolyte–surface bridges.
A third contribution arises from bridges formed by uncharged segments. This
contribution is considerably smaller in water since in this case both the surface and
the acrylamide units have to be dehydrated in order to form a mica-acrylamide
bond.
A fourth contribution arises from entanglement effects. The more open layer
structure in water favors chain interpenetration, but it also makes it easier for the
chains to disentangle during the separation process. It is difficult to a priori state if
this effect will be more important in water or in air since (in an intricate way) it
will depend on the structure of the polyelectrolyte layer, chain relaxation rate and
the deformation rate during rupture of the adhesive joint. Some recent work on
uncharged polymers has shown that the solvency of the polymer chain greatly
influences the role of chain interpenetration for the adhesion between polymercoated surfaces in liquid media w26,48x.
The experimental observation is that the adhesion force is significantly stronger
in air than in water, much more so than expected from consideration of only van
der Waals forces (see Table 2). This is the case also for hydrophilic surfaces such
as mica without any polymer coating, and this is related to the hydration of the
surface that reduces the interfacial tension. The adsorption of AM-MAPTAC-100
72
P.M. Claesson et al. / Advances in Colloid and Interface Science 104 (2003) 53–74
Table 2
The pull-off force in air and in water
Surface
Pull-off force
in dry air
(mNym)
Pull-off force
in water
(mNym)
Bare mica
AM-MAPTAC-100 coated mica
AM-MAPTAC-30 coated mica
AM-MAPTAC-10 coated mica
AM-MAPTAC-1 coated mica
800–1200
2300
1000
800
1100
20–40
100–200
2–5
0.5–1
0
on mica surfaces increases the adhesion force in water significantly, and this has
been related to bridging w9x. However, the adhesion between AM-MAPTAC-100
coated surfaces in air is significantly stronger providing evidence that this force
contribution is even stronger in the low dielectric media due to the increased
strength of each electrostatic bridge.
The adhesion force in water, just as in air, decreases with decreasing polyelectrolyte charge density. This is partly due to a decreased importance of electrostatic
bridges but also due to the development of a steric repulsion between the surfaces
in water as the charge density of the polyelectrolyte is decreased. This, in turn, is
related to the hydration of the polymer segments for which water is a good solvent.
4. Conclusions
Tapping mode AFM-imaging are able to resolve adsorbed individual polyelectrolyte chains at low coverage. However, at high coverage this is not possible, but the
images demonstrate that smooth and homogeneous layers are formed. The finding
that adsorbed low-charge density polyelectrolytes (e.g. AM-MAPTAC-1) were partly
dragged along the surface during scanning suggests that the binding strength of the
polyelectrolyte to the surface is reduced when the charge density of the polyelectrolyte decreases. This is due to formation of less electrostatic binding points.
The adhesion force measurements carried out with the SFA show that thin
polyelectrolyte layers are effective adhesion modifiers in air. In particular, strong
adhesion forces were found with AM-MAPTAC-100 and this strongly suggests that
electrostatic surface–polyelectrolyte–surface bridges are the cause of the adhesion.
At low charge density another force contribution due to entanglement of polymer
chains gives rise to a significant contribution to the adhesion force. This is clearly
seen for AM-MAPTAC-1. The disruption of the contact leads to irreversible changes
in the adsorbed layers when either the number of electrostatic bridges formed are
large (AM-MAPTAC-100) or when entanglement effects are important for polymers
with a relatively low binding strength to the surface (AM-MAPTAC-1).
The adhesion force between one polyelectrolyte-coated surface and one bare mica
surface during the first separation was generally higher than that found between
two polyelectrolyte-coated surfaces. It is suggested that this is due to formation of
P.M. Claesson et al. / Advances in Colloid and Interface Science 104 (2003) 53–74
73
a larger number of bridges. The adhesion force in the asymmetric system decreased
with decreasing charge density of the polyelectrolyte, and with increasing number
of separations. The latter observation is due to material transfer between the
polyelectrolyte-coated surface and the initially bare surface.
The decrease in contact radius during the separation process was in many cases
consistent with the predictions of JKR-theory. However, a smaller than expected
decrease was observed in some cases and related to stick–slip friction and adhesion
hysteresis occurring during the shrinkage of the contact radius. For less homogeneous
layers, such as those obtained after repeated separations, the contact radius shrank
to lower values than predicted by JKR-theory prior to surface–surface separation.
Acknowledgments
O.R. acknowledges financial support from the SSF program ‘Colloid and Interface
Technology’. This work constitutes a contribution to the program of the Biofibre
Materials Center (BiMaC) at the Royal Institute of Technology, Stockholm.
References
w1x B. Alince, Colloids Surf. 39 (1989) 39–51.
w2x J. Zhang, R. Pelton, L. Wagberg,
¨ Nord. Pulp Pap. Res. J 15 (2000) 440–445.
M. Rundlof,
˚
w3x P.M. Claesson, A. Dedinaite, E. Poptoshev, in: T. Radeva (Ed.), Physical Chemistry of
Polyelectrolytes, 99, Marcell Dekker Inc, New York, 2001, pp. 447–507.
w4x M.A.G. Dahlgren, P.M. Claesson, R. Audebert, J. Colloid Interface Sci. 166 (1994) 343–349.
w5x E. Poptoshev, M.W. Rutland, P.M. Claesson, Langmuir 15 (1999) 7789–7794.
w6x E. Poptoshev, M.W. Rutland, P.M. Claesson, Langmuir 16 (2000) 1987–1992.
w7x P.M. Claesson, M.A.G. Dahlgren, L. Eriksson, Colloids Surf. 93 (1994) 293–303.
w8x O.J. Rojas, M. Ernstsson, R.D. Neuman, P.M. Claesson, Langmuir 18 (2002) 1604–1612.
˚ Waltermo, E. Blomberg, P.M. Claesson, L. Sjostrom,
˚
w9x M.A.G. Dahlgren, A.
¨ ¨
T. Akesson,
et al., J.
Phys. Chem. 97 (1993) 11769–11775.
w10x M.A.G. Dahlgren, Langmuir 10 (1994) 1580.
w11x O.J. Rojas, P.M. Claesson, D. Muller, R.D. Neuman, J. Colloid Interface Sci. 205 (1998) 77–88.
¨
w12x M. Osterberg,
J. Colloid Interface Sci. 229 (2000) 620–627.
w13x E. Poptoshev, P.M. Claesson, Langmuir 18 (2002) 2590–2594.
w14x O. Rojas, M. Ernstsson, R.D. Neuman, P.M. Claesson, J. Phys. Chem. B 104 (2000) 10032–10042.
w15x M.A.G. Dahlgren, H.C.M. Hollenberg, P.M. Claesson, Langmuir 11 (1995) 4480–4485.
w16x A. Dedinaite, P.M. Claesson, M. Bergstrom,
¨ Langmuir 16 (2000) 5257–5266.
w17x M. Ruths, S.A. Sukhishvili, S. Granick, J. Phys. Chem. B 105 (2001) 6202–6210.
w18x B.V. Derjaguin, Progr. Surface Sci. 45 (1994) 223–231.
w19x K.L. Johnson, K. Kendall, A.D. Roberts, Proc. R. Soc. Lond. A 324 (1971) 301–313.
w20x V.M. Muller, V.S. Yushchenko, B.V. Derjaguin, J. Colloid Interface Sci. 77 (1980) 91–101.
w21x B.V. Derjaguin, Y.P. Toporov, V.M. Muller, I.N. Aleinikova, J. Colloid Interface Sci. 58 (1977)
528–533.
w22x V.M. Muller, V.S. Yushchenko, B.V. Derjaguin, J. Colloid Interface Sci. 92 (1983) 92–101.
w23x V.M. Muller, B.V. Derjaguin, Y.P. Toporov, Colloids Surf. 7 (1983) 251–259.
w24x B.V. Derjaguin, V.M. Muller, Y.P. Toporov, I.N. Aleinikova, Prog. Surf. Sci. 45 (1994) 192–198.
w25x B.V. Derjaguin, V.M. Muller, Y.P. Toporov, Prog. Surf. Sci. 45 (1994) 131–143.
w26x M.A. Plunkett, M.W. Rutland, J. Adhes. Sci. Technol. 16 (2002) 983–996.
w27x Y.P. Huang, W. Leobandung, A. Foss, N.A. Peppas, J. Controlled Release 65 (2000) 63–71.
74
w28x
w29x
w30x
w31x
w32x
w33x
w34x
w35x
w36x
w37x
w38x
w39x
w40x
w41x
w42x
w43x
w44x
w45x
w46x
w47x
w48x
P.M. Claesson et al. / Advances in Colloid and Interface Science 104 (2003) 53–74
E.R. Beach, G.W. Tormoen, J. Drelich, R. Han, J. Colloid Interface Sci. 247 (2002) 84–99.
M.A. Hubbe, Colloids Surf. 25 (1987) 325–339.
M.L. Janex, R. Audebert, V. Chaplain, J.L. Counord, Colloid Polym. Sci. 275 (1997) 352–363.
Y.L. Chen, J.N. Israelachvili, J. Phys. Chem. 96 (1992) 7752–7760.
M. Ruths, S. Granick, Langmuir 14 (1998) 1804–1814.
M. Ruths, S. Granick, J. Phys. Chem. B 102 (1998) 6056–6063.
O.J. Rojas, R.D. Neuman, P.M. Claesson, J. Colloid Interface Sci. 237 (2001) 104–111.
M.A. Plunkett, P.M. Claesson, M.W. Rutland, Langmuir 18 (2002) 1274–1280.
Q. Zhong, D. Innins, K. Kjoller, V.B. Elings, Surf. Sci. Lett. L688 (1993) 290.
A.L. Weisenhorn, P.K. Hansma, T.R. Albrecht, C.F. Quate, Appl. Phys. Lett. 54 (1989) 2651–2653.
B. Drake, C.B. Prater, A.L. Weisenhorn, S.A.C. Gould, T.R. Albrecht, C.F. Quate, et al., Science
243 (1989) 1586–1589.
J.N. Israelachvili, G.E. Adams, J. Chem. Soc. Faraday Trans. 1 74 (1978) 975–1001.
B. Derjaguin, Kolloid. Zeits 69 (1934) 155–164.
¨
N. Guven,
Z. Kristallogr 134 (1971) 196–212.
M.A. Plunkett, P.M. Claesson, M. Ernstsson, M.W. Rutland, in press.
H.K. Christensson, J. Phys. Chem. 97 (1993) 12034–12041.
P.M. McGuiggan, J.N. Israelachvili, J. Mater. Res. 5 (1990) 2232–2243.
I. Sridhar, K.L. Johnson, N.A. Fleck, J. Phys. D 30 (1997) 1710–1719.
J. Israelachvili, in: B. Bhushan (Ed.), Handbook of MicryNano Tribology, CRC Press, Boca
Raton, 1995, pp. 267–319.
R. Neuman, J. Berg, P.M. Claesson, Nord. Pulp Pap. Res. J 8 (1993) 96–104.
¨
¨
M.A. Plunkett, S. Rodner,
L. Bergstrom,
M.W. Rutland, J. Adhesion Sci. Technol. 16 (2002)
965–981.