St

Ea

Kb

c

d

h

•

•

•

•

a

ARRAA

KTCMfLAS

1

tpspe

0h

Colloids and Surfaces A: Physicochem. Eng. Aspects 441 (2014) 701– 708

Contents lists available at ScienceDirect

Colloids and Surfaces A: Physicochemical andEngineering Aspects

jo ur nal ho me page: www.elsev ier .com/ locate /co lsur fa

urface forces and friction between non-polar surfaces coated byemperature-responsive methylcellulose

sben Thormanna,b,∗, Rasmus Bodvika, Leif Karlsonc, Per M. Claessona,d

KTH Royal Institute of Technology, School of Chemical Science and Engineering, Department of Chemistry, Surface and Corrosion Science, Drottningristinas väg 51, SE-100 44 Stockholm, SwedenTechnical University of Denmark, Department of Chemistry, Kemitorvet 207, DK-2800 Kgs. Lyngby, DenmarkAkzo Nobel Functional Chemicals AB, SE-444 85 Stenungsund, SwedenSP Technical Research Institute of Sweden, Chemistry, Materials and Surfaces, Box 5607, SE-114 86 Stockholm, Sweden

i g h l i g h t s

Studies of modified celluloseadsorbed at hydrophobic surfaces.Normal and frictional forces betweenadsorbed layers.Weak hydrophobic anchor and shear-induced wear.Self-healing properties with freepolymers in solution.

g r a p h i c a l a b s t r a c t

r t i c l e i n f o

rticle history:eceived 21 August 2013eceived in revised form 10 October 2013ccepted 17 October 2013vailable online 26 October 2013

eywords:

a b s t r a c t

Methylcellulose is a heterogeneous polymer that exposes both methyl groups and –OH-groups to thesolution, and the solvent quality of water for methylcellulose deceases with increasing temperature. Inbulk solution this leads to aggregation into fibrils at high temperatures. In this report we address howtemperature affects adsorbed layers of methylcellulose on hydrophobized silica surfaces in contact withan aqueous methylcellulose solution. The layers were imaged using PeakForce tapping mode atomic forcemicroscopy, in order to determine how the additional adsorption that occurs with increasing temperature

emperature-responsive polymerellulose ethersethylcellulose

rictionoad bearing capacity

affects the layer structure. Surface force and friction measurements were carried out using the AFMcolloidal probe method. The data demonstrate that the normal surface forces were rather insensitive totemperature, whereas the friction forces changed significantly with increasing temperature. At low loadsthe friction increases with increasing temperature, whereas at high loads the reverse is observed. Thesefindings are discussed in terms of how the worsening of the solvent condition affects the aggregation

r, an

tomic force microscopyurface forces

state in the adsorbed laye

. Introduction

The colloidal stability of dispersions can be controlled by elec-rostatic forces or due to steric interactions between adsorbedolymer layers, where the latter is preferred in high ionic strengtholutions. Of particular interest is to be able to control the dis-

ersion stability by using polymers that are responsive to thenvironmental conditions, such as temperature [1,2].

∗ Corresponding author. Tel.: +45 53828118; fax: +46 8208284.E-mail address: esth@kemi.dtu.dk (E. Thormann).

927-7757/$ – see front matter © 2013 Elsevier B.V. All rights reserved.ttp://dx.doi.org/10.1016/j.colsurfa.2013.10.038

d the polymer-surface affinity.© 2013 Elsevier B.V. All rights reserved.

For many types of polymer, water becomes a less good sol-vent with increasing temperature. This is the case, for instance,for polymers like poly(N-isopropylacrylamide) (PNIPAAm) [3–5],poly(2-isopropyl-2-oxazoline) [6,7], poly(2-(dimethylamino)ethylmethacrylate [1], poly(ethylene oxide) (PEO) [8,9], poly(propyleneoxide) (PPO) [10–13], ethyl(hydroxyethyl)cellulose (EHEC) [14],methylcellulose (MC) [15] and hydroxypropylmethylcellulose(HPMC) [16]. The responsive properties of these types of polymershave inspired attempts to use them in controlled delivery appli-

cations. For instance, the cellulose ethers have received interestfrom this perspective [16,17], even though their main applica-tion may be as temperature-responsive viscosity modifiers in e.g.paints.

7 Physi

iatabwemoe

itpHoeeAcocdw

2

Aiomfiacmrlcwswmpm

mtt

2

i

02 E. Thormann et al. / Colloids and Surfaces A:

Surfaces are present in all of the above application areas, and its thus of high interest to understand adsorption properties as wells interactions between surfaces coated with cellulose ethers, ashey approach each other and as they slide past each other. Only

few reports can be found on this topic. The surface forces actingetween hydrophilic [18] and hydrophobic surfaces [19,20] coatedith EHEC has been described in a few studies, but to our knowl-

dge no investigation on interactions between surfaces coated withethylcellulose has been reported. Likewise, no report of the nan-

tribological properties of surfaces coated with cellulose ethersxists.

This study has been initiated to gain better understanding ofnteractions within and between methylcellulose layers adsorbedo non-polar surfaces. To this end we have used the AFM colloidalrobe technique to investigate surface forces and friction forces.ere we aimed at determining if attractive surface forces werebserved at the aggregation temperature observed in solution, or ifnhanced steric stabilization was facilitated due to preferential ori-ntations of hydrophobic and hydrophilic groups within the layer.

further aim was to elucidate the lubricating ability of the methyl-ellulose for non-polar surfaces in water, with particular emphasisn temperature effects on the friction force and the load bearingapacity. We have further used AFM PeakForce imaging to eluci-ate if aggregation occurs within adsorbed methylcellulose layersith increasing temperature.

. Materials and methods

Freeze-dried samples of methylcellulose were obtained fromkzo Nobel, Stenungsund, Sweden. It will be referred to as M1.6C

n this report in order to highlight that the methoxy degreef substitution per anhydroglucose unit is 1.6. The structure ofethylcellulose is illustrated in Fig. 1. The polymer was puri-

ed from eventual by-products by first dissolving it in water to concentration of around 1 wt%. The solution was subsequentlyentrifuged at ca. 7000 × g for 60 min to remove water insolubleaterial. Water soluble impurities such as salt and glycols were

emoved through dialysis of the supernatants against excess of Mil-ipore water for 7 days. A Spectra/Por® membrane tubing with Mw

ut-off of 6–8000 g/mole was used for the dialysis. The Milliporeater in the water tank was exchanged once a day. After the dialy-

is the polymers were freeze-dried. The molecular weight, 530 kDa,as obtained from size exclusion chromatography at 30 ◦C with aobile phase of 0.04 M sodium acetate and 0.02% sodium azide at

H 6. Refractive index, light scattering (at 7◦ and 90◦) and viscosityeasurements were employed in the analysis.Solutions were prepared by dispersing the freeze-dried poly-

ers in water at high temperature (80–90 ◦C) under stirring, andhen cooling to room temperature under continued stirring. Finally,he solutions were stirred in an ice bath for at least 30 min.

.1. Substrates

A silicon substrate with a 100-nm thick thermally grown sil-ca layer (WaferNet GmbH, Eching, Germany) with a typical rms

Fig. 1. Structure of methylcellul

cochem. Eng. Aspects 441 (2014) 701– 708

roughness below 1 nm was used as the flat surface in AFM colloidalprobe force measurements. The colloidal probe was a sphericalsilica bead (SS06N, Bangs Laboratories Inc.) with a diameter of7 �m. All surfaces, except for the colloidal probe, were cleaned for5 min in 80 ◦C 5:1:1 (w/w/w) H2O:NH3:H2O2, rinsed extensivelywith Milli-Q water, then cleaned for 5 min in 80 ◦C 5:1:1 (w/w/w)H2O:HCl:H2O2, and rinsed with Milli-Q water again.

Hydrophobic substrates were obtained by silanizationby exposing the surfaces to the vapour of (3,3-dimethylbutyl)dimethylchlorosilane (DDS) in a desiccatorovernight. The hydrophobized surfaces were then rinsed withwater and stored in ethanol. The contact angle of water wasapproximately 100◦.

2.2. Atomic force microscopy—Peak force imaging

Topographical images of hydrophobized surfaces with anadsorbed polymer layer and polymers present in solution wererecorded using an atomic force microscope (AFM) Nanoscope Mul-timode V (Bruker) operating in PeakForce mode [21–24] usingScanAsyst, silicon nitride cantilevers (ScanAsyst Air, Bruker probes)and a fluid cell with in- and outlets allowing for solvent exchange.The temperature was set by use of a thermal application controllerattached to a Bioheater element (Bruker) mounted under the sam-ple. The temperature on the sample surface was calibrated by anexternal thermocouple and controlled with an accuracy of ±1 ◦C.Hydrophobic substrates were prepared as described above and theexperiments were conducted as follows: The cell was filled with40 ppm aqueous M1.6C solution and the system was left for 30 minto allow adsorption at 25 ◦C before the surface was imaged. Next,the temperature was increased to first 40 ◦C and then 50 ◦C. At bothtemperatures images were acquired 30 min after the temperaturehad stabilized to allow the system to equilibrate. The imaging wascarried out using a peak force setpoint of 1 nN.

2.3. Atomic force microscope—Colloidal probe measurements

Friction and normal force measurements were performed in afused silica liquid cell using an atomic force microscope, MultimodeNanoscope III Pico Force (Bruker) and the colloidal probe tech-nique [25,26]. A silica particle with diameter of approximately 7 �mwere attached to the end of a tipless cantilever (CSC12, F-lever,Mikromasch, Estonia) with the aid of an Ependorf Micromanip-ulator 5171, a Nikon Optiphot 100S reflection microscope, anda small amount of high-temperature melting epoxy glue (ShellEpikote 1009). The size of the particles was determined using aNikon Optiphot 100S reflection microscope, employing image anal-ysis with National Instrument Vision Assistant 8.0. Before particleattachment, the values of the normal and torsional spring constantswere determined by the Sader method [27,28]. The type of can-tilever was chosen such that the maximum force of interest was

not leading to a cantilever deflection outside the linear region ofthe detector [29]. The lateral photodetector sensitivity was cali-brated using the method of tilting the AFM head as suggested byPettersson et al. [30].

ose, where R=H or R=CH3.

E. Thormann et al. / Colloids and Surfaces A: Physi

Table 1Adsorbed layer properties of M1.6C on hydrophobized silica surfaces.

Temperature(◦C)

Adsorbedamounta

(mg/m2)

Sensed massb

(mg/m2)Layerthicknessc

(nm)

Watercontentc

(%)

25 1.1 20 19 94.540 1.5* 20* 19 92.6*

50 1.95 21 20 91

*Values interpolated from data at 35 ◦C and 45 ◦C reported in Ref. [33].

ssoafn53

rtw1asmavv

3

femWtittwhmm

rl

3

iotFmitrwt

substructure at the molecular length scale. In apparent disagree-ment with the adsorption data in Table 1, no significant changesare observed as the temperature is increased to first 40 ◦C and

a Based on ellipsometry measurements reported in Ref. [33].b Based on QCM-D measurements reported in Ref. [33].c Based on QCM-D and ellipsometry measurements reported in Ref. [33].

The AFM-based force versus distance measurements weretarted by measuring the normal forces between the pure silicaurface and the silica probe in water. Hereafter, a 40 ppm solutionf M1.6C in water was introduced and the polymer was allowed todsorb for 45 min. Next, the normal forces were measured at 25 ◦C,ollowed by friction measurements, and again measurements oformal forces. This procedure was then repeated at 40 ◦C and at0 ◦C. At each temperature the system was allowed to stabilize for0 min before the first force measurement.

The normal forces were measured with a constant approach andetraction speed of 200 nm/s, which is sufficiently slow to allow uso neglect hydrodynamic forces [31]. The friction measurementsere performed by sliding the surfaces backwards and forwards

0 times at each normal load and registering the cantilever twistngle. The sliding distance was 1 �m in each direction and thecan rate was 1 Hz, giving a sliding speed of 2 �m/s. For both nor-al force and friction results of several consecutive measurements

t different surface position were obtained. Further, no apparentelocity dependence was observed and results obtained at differentelocities are not presented in this work.

. Results

Before we describe the results on layer morphology, surfaceorces and friction obtained in this study, we recapitulate somessential features of the solution and adsorption properties ofethylcellulose, as described in previous publications [32–34].ater becomes a progressively worse solvent for M1.6C as the

emperature increases, and by use of dynamic light scatteringt has been found that a slow aggregation starts to occur atemperatures just above 50 ◦C, which closely corresponds to theemperature where the viscosity starts to increase significantlyhen the temperature sweep rate is extrapolated to zero [33]. Atigher temperatures the aggregation process, which leads to for-ation of fibrils as visualized by CRYO-TEM, proceeds considerablyore rapidly.The adsorption of M1.6C onto hydrophobized silica has been

eported previously [33], and some information about the adsorbedayer at the temperatures used in this study is collected in Table 1.

.1. Peak force imaging

The structure of the adsorbed layer visualized by AFM peak forcemaging at 25, 40 and 50 ◦C is shown in Fig. 2. These images werebtained using a peak force setpoint of 1 nN and the radius of theip used was about 2 nm. Thus, the peak force normalized by radius,/R, amounts to about 500 mN/m, which is about two orders ofagnitude larger than the maximum normalized force applied dur-

ng a typical surface force measurement with the colloidal probe

echnique. This is important to remember since the AFM imagesepresent the structure of the layer under significant compression,hereas the surface force measurements probe the outer region of

he polymer layer at much lower compression.

cochem. Eng. Aspects 441 (2014) 701– 708 703

At 25 ◦C the surface appear rather smooth with some degree of

Fig. 2. Height images of an adsorbed methylcellulose layer at 25 ◦C, 40 ◦C and 50 ◦Cobtained by PeakForce tapping mode AFM operating with a constant feedback peakforce of 1 nN.

7 Physi

ttafidirmviTtvlscae

3

hcrtaFfbtWt1

snltar[ttiwaaabonasfAmbi

cmted

04 E. Thormann et al. / Colloids and Surfaces A:

hen 50 ◦C. We however note two minor observations. Firstly, whenhe temperature in increased above 25 ◦C the images starts toppear more blurry which is an indication of a less compliant sur-ace structure that hamper good image feedback. This observations suggested to be an effect of the increased adsorbed amountescribed in relation to Table 1. Secondly, as the temperature is

ncreased, the topographical variations are increased and distinctound objects starts to form. We suggest that these objects areolecular aggregates which are formed due to the worsened sol-

ent condition. We note that in bulk solution M1.6C aggregatesnto fibrillar structures above about 50 ◦C, as visualized by Cryo-EM [33], and similar aggregation could possibly be induced withinhe adsorbed layer at somewhat lower temperatures. The obser-ation that aggregation might occur at lower temperatures in theayer compared to in bulk is suggested to be an entropy effectince the molecules already has lost a considerable fraction of theironformational (and translational) entropy upon adsorption. Thus,ggregation in the adsorbed layer is associated with lower loss ofntropy than for the molecules present in bulk solution.

.2. Surface forces

The forces measured between two M1.6C layers adsorbed toydrophobized silica surfaces at 25 ◦C, 40 ◦C and 50 ◦C using theolloidal probe technique are shown in Fig. 3. On approach a long-anged roughly exponential decaying repulsion is observed at allemperatures. The decay length is found to be about 22 nm at 25nd 50 ◦C and slightly lower at 40 ◦C. As seen from the insets inig. 3, these numbers does not vary significantly between repeatedorce measurements. An exponential decaying force can in principlee caused by either a double-layer interaction or a steric interac-ion due to overlap of polymer tails extending from the surfaces.

e assign this to a steric interaction. The decay-length is too shorto be a double-layer force in pure water, where a decay length of24–961 nm should be expected in the pH range 5.6–7.

If the repulsive force is of steric origin, the interaction rangehould be related to (two times) the layer thickness while the mag-itude of the interaction should be related to the stiffness of the

ayer. In our case the range of the steric force is thus affected byhe solvent quality in two opposite ways. In a situation where thedsorbed amount is constant, a worsening of the solvent qualityesults in a more compact layer and a less long-ranged steric force19]. However, in our case the adsorbed amount also increases dueo additional adsorption, and this may lead to an increased range ofhe steric force [20]. Apparently the former effect dominates whenncreasing the temperature from 25 ◦C to 40 ◦C, and the latter effect

hen the temperature is increased further from 40 ◦C to 50 ◦C. Atll three temperatures the onset of the repulsive force occurs atn apparent separation which is larger than two times the aver-ge layer thickness. This is often seen in studies of surface forcesetween polymer surfaces and is interpreted to be due to weakverlap between dangling chains. Since such dangling chains areot included in a model where the layer is described as film withn even density profile and a sharp cut-off the offset of the repul-ive force have to occur at a larger surface separation than expectedrom the data in Table 1. It should, however, also be noted that theFM in contrast to e.g. the surface force apparatus (SFA) does noteasures the absolute distances but only the relative distances set

y the constant compliance region. This fast could also bias thenterpretation of the interaction ranges.

On retraction a long-ranged bridging attraction, having aharacteristic saw-tooth profile due to stretching of a few poly-

er chains [31,35–37] is observed at all temperatures. Due

o the high adsorbed amount (Table 1) and the absence ofmpty areas on the surface (Fig. 2) we interpret this as beingue to attraction between M1.6C polymers residing on the

cochem. Eng. Aspects 441 (2014) 701– 708

opposing surfaces rather than due to polymers binding directlyto two surfaces. Neither the magnitude nor the range of theattraction is significantly affected by an increase in tempera-ture.

3.3. Friction forces

According to Amontons’ classical first rule, the friction force, Ff,should be proportional to the load, Fn:

Ff = �Fn (1)

where � is the friction coefficient. Clearly, this rule is not followedat 25 ◦C and 40 ◦C in the present case (Fig. 3), indicating that themain energy dissipative mechanism changes with the applied load[38,39]. At 25 ◦C, the friction force is relatively low at low loadsand in this regime the friction coefficient is about 0.17. At thesesmall forces the load is carried by the steric repulsion in the dilutetail region, and the energy dissipation associated with sliding thepolymer tails over each other is low. However, at higher loads thefriction force increases dramatically, and it remains high as theload is decreased to zero and even negative values. Clearly, a newenergy dissipative mechanism has come into play. The forces mea-sured after the friction experiment is completely different to thosemeasured before the surfaces were sheared against each other (seeFig. 4). Immediately after the friction measurements the long-rangeforce is attractive and the adhesion is very strong, and similar towhat has been reported for these types of silanated surfaces inabsence of an adsorbed polymer layer in aqueous solutions [40–45].Thus, it is clear that the M1.6C layer is reorganized at the surfacesduring shearing under a high load, and thus the load bearing capac-ity, defined as the load where the friction force suddenly increases,is relatively low. It occurs in the load interval 2.6–5 nN in differentsets of measurements at 25 ◦C. However, if the surfaces were leftapart for 5–10 min, the polymer reabsorbed and the initial forceversus separation curve was observed again (see Fig. 4). Thus, theadsorbed layer has a self-healing capacity due to readsorption ofM1.6C, and possibly also due to surface diffusion of the polymerto the bare surface region introduced during shearing. Such a self-healing affect is common when adsorbed polymers or surfactantsare used to lower boundary friction with the lubricants present inbulk solution [33,34].

An increase in temperature to 40 ◦C results in an increase inthe friction force at low loads, characterized by a friction coeffi-cient of about 1 up to a load of about 2.3 nN. Thus, the friction forceobserved at low loads is significantly higher at 40 ◦C than at 25 ◦C,which is a consequence of the worsening of the solvent quality.At higher loads the friction increases sharply, but less dramaticallythan observed at 25 ◦C. In contrast to what was observed at 25 ◦C,the friction force decreases again as the load is released, and it islower on unloading than on loading. We suggest that this is due toflattening of the M1.6C layer. A further increase in temperature to50 ◦C results in minor changes in the friction force observed at lowloads, but in this case the friction coefficient remains constant up tothe highest load explored (12 nN). This latter observation is a con-sequence of the increase in load bearing capacity that counteractsremoval of the polymer from the surface during shearing at highloads. The increased load bearing capacity, in turn, is a direct con-sequence of the worsening of the solvent quality that result in bothhigher surface affinity, and increased cohesion within the layer. Atall three temperatures the trends and magnitudes but not the exactshapes of the friction loops are reproduced in repeated measure-

ments and at different surfaces positions. The variations betweenrepeated runs are related to the stochastic nature of layer failureand variations in the exact structure of the reorganized layer afterfailure.

E. Thormann et al. / Colloids and Surfaces A: Physicochem. Eng. Aspects 441 (2014) 701– 708 705

Fig. 3. Representative approach and retraction force curves obtained before friction force measurements at 25 ◦C, 40 ◦C and 50 ◦C, respectively (left column). The insertss plotte( pen tt

4

4a

phat

how 10 consecutive approach curves, obtained at each of the three temperatures,right column). The open circles represent data obtained on increasing load, while ohe friction force at low loads in more detail.

. Discussion

.1. Shielding of hydrophobic groups due to aggregation anddsorption

Methylcellulose can be viewed as a heterogeneous polymer

resenting both hydrophobic regions (the methyl groups) andydrophilic regions (unreacted –OH-groups) towards solution. Themphiphilic character of the polymer leads to extensive aggrega-ion at high temperatures, and the onset of aggregation depends

d on a semi-log scale. Friction versus normal load at the same three temperaturesriangles represent friction during decreasing normal load. The insets at 25 ◦C show

on the molar substitution of the methyl groups and the molec-ular weight. For M1.6C, the critical temperature is close to 50 ◦C[33]. Upon aggregation the hydrophobic groups become signif-icantly shielded from water in the fibrillar aggregates that areformed. This results in significantly decreased mobility of themethyl groups as evidenced by NMR measurements [46], and also a

slow aggregation kinetics as compared to other modified celluloseethers that form less compact aggregation structures, e.g. hydrox-ypropylcellulose (HPMC) and ethyl(hydroxyethyl)cellulose (EHEC)[33].

706 E. Thormann et al. / Colloids and Surfaces A: Physicochem. Eng. Aspects 441 (2014) 701– 708

Fig. 4. Force and friction measurements between layers of M1.6C adsorbed to hydrophobized silica at 25 ◦C. (A) The normal forces measured before friction measurements,where the blue curve corresponds to the approach curve and the red one to the retraction curve. (B) Friction force as a function of applied load. The open cycles and triangles,r hen dm ayed i

latwrapmsiitaout

4

ndptTfft

espectively, indicate how the friction changes as the load first is increased and teasurement. (D) The normal force curve measured 5–10 min after the curve displ

For the same reason the hydrophobic groups present in the cel-ulose ethers are preferentially oriented towards the inside of andsorbed layer, whereas the more hydrophilic regions are exposedowards solution. This has previously been demonstrated for EHEC,here surface force measurements showed that the transition from

epulsive interactions at low temperature to attractive interactionst high temperature was shifted upwards by several degrees com-ared to the cloud point temperature in bulk solution [19,20]. Forethylcellulose the same behaviour is expected, and it is thus no

urprise that the surface interaction between M1.6C layers probedn this study up to 50 ◦C, similar to the aggregation temperaturen bulk solution, show only weak temperature dependence. In par-icular, we observe no increase in the magnitude of the bridgingttraction with temperature, but the trend is rather reversed. Thisbservation is rationalized by formation of self-assembled molec-lar structures in the adsorbed layer that allow the methyl groupso be efficiently shielded.

.2. Interactions during shearing

The interactions during shearing show, in sharp contrast to theormal force measurements, very strong temperature dependenceue to the worsening of the solvent quality of water at higher tem-eratures. At low loads, when the adsorbed layer remains intact,he friction coefficient increases as the solvent quality is decreased.

his is in line with numerous other studies [38,47–49], and resultsrom the increased attraction between the polymer segments. Theriction coefficient obtained between intact layers at high tempera-ures is five times higher than at 25 ◦C. This high friction is observed

ecreased again. (C) A normal force curve obtained immediately after the frictionn (C).

despite very low attraction between the two opposing surfaces.Thus, we conclude that the main energy dissipation occurs withinthe adsorbed layers (rather than between these), where the com-bination of shear and load results in breakage and reformation ofintralayer physical bonds. The importance of this type of energydissipative mechanism was also recently reported for the case of agrafted poly(acrylic acid) layer sliding against a bare silica surface.In this case, addition of divalent ions at slightly alkaline pH resultedin layer compaction due to physical cross-links mediated by thepresence of divalent cations, but no adhesion between the two slid-ing surfaces. Nevertheless, the friction force increased substantiallyby addition of the divalent ions [50].

At 25 ◦C and high loads, we observe a significantly higher frictionforce compared to at a more elevated temperature. We interpretthis as a consequence of an increase in the load bearing capac-ity with increasing temperature, due to higher polymer-surfaceaffinity and higher cohesion in the layer due to more extensive poly-mer aggregation into self-assembled structures. To achieve botha low friction force and high load bearing capacity by adsorptionto non-polar surfaces was not possible with methylcellulose, butwould require another polymer structure that provides strongeranchoring to the substrate surface and at the same time expo-sure of hydrophilic groups towards the solution side. This suggeststhat diblock copolymers may be suitable. Diblock copolymers andbottle-brush polymers have indeed successfully been used for

achieving low friction forces and high load bearing capacity onhydrophilic surfaces [51–54]. The key has been to anchor the poly-mer strongly to the surface by electrostatic adsorption, and atthe same time expose hydrophilic uncharged blocks towards the

Physi

shmfiwehbfiTaohahitsitsHntcms

5

hitepfatadfaasi

atmstowpmlaitoTw

[

[

[

[

[

[

[

[

[

[

[

E. Thormann et al. / Colloids and Surfaces A:

olution side. It may, however, be more problematic to achieve aigh load bearing capacity using diblock or bottle-brush copoly-ers on an uncharged non-polar surface. The fundamental reason

or this is that the electrostatic anchoring relies on site specificnteractions where e.g. negative surface sites interacts strongly

ith positively charged polymer segments, which means that thenergy barrier for sliding the polymer along the surface will beigh. On a non-polar surface, strong adsorption due to hydropho-ic interactions can be achieved, but in this case the energy barrieror sliding the polymer along the surface is expected to be signif-cantly lower since the surface is chemically more homogeneous.his has been confirmed by single polymer desorption [31,55–58]nd by single polymer friction [59] experiments where it has beenbserved that a correlation between high desorption energy andigh friction is dependent of the nature of the adhesive inter-ction between the polymer and the substrate. The pronouncedysteresis between the friction measured upon loading and unload-

ng as see in Fig. 3 (top, right) and 4B is an interesting effect ofhe changed adhesive contact during shearing. Hysteresis in adhe-ive contact due to reorganization of the layer or due to a changen the balance between layer–layer and layer–substrate interac-ion is a common observation for friction measurements betweenoft structured layers when employing a loading–unloading cycle.owever, in extreme cases, as in Fig. 4B, it can lead to a curve with aegative slope in the plot of friction versus applied load, a situationhat one could define as a system with “a negative friction coeffi-ient”. The understanding of such systems has recently attracteduch attention [60–63], and in this particular case we assign it to

uccessive depletion of polymer from the sheared contact region.

. Conclusions

Adsorbed layers of methylcellulose on silanized silica surfacesave been probed as a function of temperature by AFM Peak Force

maging, as well as surface force and friction measurements usinghe AFM colloidal probe method. The images of the adsorbed lay-rs at 25 ◦C, obtained using a small peak force setpoint, reveal theresence of topographical heights on top of a relatively flat sur-ace. These heights are assigned to individual polymers or smallggregates extending from the surface. With increasing tempera-ure these protrusions becomes more abundant due to increaseddsorption. Clearly, aggregation occurs within the adsorbed layerue to the worsening of the solvent condition, and the structuresormed are probably similar to those that form in bulk solutiont and above 50 ◦C. The lower aggregation temperature in thedsorbed layer is suggested to be due to lower entropic penaltyince a large portion of the conformational entropy is lost alreadyn the adsorption process.

The steric forces acting between two methylcellulose layerscross aqueous polymer solution are relatively insensitive to theemperature in the range 25–50 ◦C. This is rationalized by the accu-

ulation of the hydrophobic groups within the layer, and furtherhielding of these groups due to intralayer aggregation at higheremperatures. In particular, no strong attractive forces are devel-ped due to the worsening of the solvent conditions, but onlyeak bridging interactions are observed on retraction at all tem-eratures investigated. In contrast, the friction forces between theethylcellulose layers changes significantly with temperature. At

ow loads the friction force increases with increasing temperatures a result of increasing segment–segment attraction. Upon shear-ng such physical bonds are broken and reformed mainly within

he adsorbed layers, explaining why a strong temperature effect isbserved for the friction force but not for the normal surface forces.he load bearing capacity of the methylcellulose layer increasesith increasing temperature as a consequence of the worsening of

[

cochem. Eng. Aspects 441 (2014) 701– 708 707

the solvent quality that result in stronger polymer-surface affinityand higher cohesion due to stronger polymer–polymer interac-tions.

Acknowledgements

This work was supported by the institute excellent centreCODIRECT, financed by VINNOVA and the Swedish Foundation forStrategic Research (SSF) at the Institute for Surface Chemistry. ETacknowledges financial support from the program “Microstructure,Corrosion and Friction Control” supported by SSF.

References

[1] T. Saigal, H. Dong, K. Matyjaszewski, R.D. Tilton, Pickering emulsions stabilizedby nanoparticles with thermally responsive grafted polymer brushes, Langmuir26 (19) (2010) 15200–15209.

[2] J.-P. O’Shea, G.G. Qiao, G.V. Franks, Solid–liquid separations with a temperature-responsive polymeric flocculant: effect of temperature and molecular weighton polymer adsorption and deposition, Journal of Colloid and Interface Science348 (1) (2010) 9–23.

[3] H.G. Schild, Poly(N-isopropylacrylamide)—experiment theory and application,Progress in Polymer Science 17 (2) (1992) 163–249.

[4] E.S. Gil, S.M. Hudson, Stimuli-reponsive polymers and their bioconjugates,Progress in Polymer Science 29 (12) (2004) 1173–1222.

[5] Z.M.O. Rzaev, S. Dincer, E. Piskin, Functional copolymers of N-isopropylacrylamide for bioengineering applications, Progress in PolymerScience 32 (5) (2007) 534–595.

[6] N. Morimoto, R. Obeid, S. Yamane, F.M. Winnik, K. Akiyoshi, Compositenanomaterials by self-assembly and controlled crystallization of poly(2-isopropyl-2-oxazoline)-grafted polysaccharides, Soft Matter 5 (8) (2009)1597–1600.

[7] J.S. Park, Y. Akiyama, F.M. Winnik, K. Kataoka, Versatile synthesis ofend-functionalized thermosensitive poly(2-isopropyl-2-oxazolines), Macro-molecules 37 (18) (2004) 6786–6792.

[8] R. Kjellander, E. Florin, Water-structure and changes in thermal-stability of thesystem poly(ethylene oxide)-water, Journal of the Chemical Society, FaradayTransactions I 77 (1981) 2053–2060.

[9] P. Alexandridis, J.F. Holzwarth, Differential scanning calorimetry investigationof the effect of salts on aqueous solution properties of an amphiphilic blockcopolymer (poloxamer), Langmuir 13 (23) (1997) 6074–6082.

10] S.S. Naik, J.G. Ray, D.A. Savin, Temperature- and pH-responsive self-assemblyof poly(propylene oxide)-b–poly(lysine) block copolymers in aqueous solution,Langmuir 27 (11) (2011) 7231–7240.

11] P. Alexandridis, J.F. Holzwarth, T.A. Hatton, Micellization of poly(ethyleneoxide)–poly(propylene oxide)–poly(ethylene oxide) triblock copolymersin aqueous-solutions—thermodynamics of copolymer association, Macro-molecules 27 (9) (1994) 2414–2425.

12] P. Alexandridis, T.A. Hatton, Poly(ethylene oxide)–poly(propyleneoxide)–poly(ethylene oxide) block-copolymer surfactants in aqueous-solutions and at interfaces—thermodynamics, structure, dynamics andmodeling, Colloids and Surfaces A—Physicochemical and Engineering Aspects96 (1–2) (1995) 1–46.

13] E. Thormann, On understanding of the Hofmeister effect: how addition of saltalters the stability of temperature responsive polymers in aqueous solutions,RCS Advances 2 (22) (2012) 8297–8305.

14] M. Malmsten, B. Lindman, Ellipsometry studies of the adsorption of celluloseethers, Langmuir 6 (2) (1990) 357–364.

15] N. Rasenack, H. Hartenhauer, B.W. Muller, Microcrystals for dissolution rateenhancement of poorly water-soluble drugs, International Journal of Pharma-ceutics 254 (2) (2003) 137–145.

16] S. Baumgartner, O. Planinsekc, S. Srcic, J. Kristl, Analysis of surface properties ofcellulose ethers and drug release from their matrix tablets, European Journalof Pharmaceutical Sciences 27 (4) (2006) 375–383.

17] M. Scherlund, A. Brodin, M. Malmsten, Nonionic cellulose ethers as poten-tial drug delivery systems for periodontal anesthesia, Journal of Colloid andInterface Science 229 (2) (2000) 365–374.

18] E. Freyssingeas, K. Thuresson, T. Nylander, F. Joabsson, B. Lindman, A surfaceforce, light scattering, and osmotic pressure study of semidilute aqueous solu-tions of ethyl(hydroxyethyl)cellulose-long-range attractive force between twopolymer-coated surfaces, Langmuir 14 (20) (1998) 5877–5889.

19] M. Malmsten, P.M. Claesson, E. Pezron, I. Pezron, Temperature-dependent forces between hydrophobic surfaces coated withethyl(hydroxyethyl)cellulose, Langmuir 6 (10) (1990) 1572–1578.

20] M. Malmsten, P.M. Claesson, Temperature-dependent adsorption and sur-face forces in aqueous ethyl(hydroxyethyl)cellulose solutions, Langmuir 7 (5)

(1991) 988–994.

21] M. Sababi, J. Kettle, H. Rautkoski, P.M. Claesson, E. Thormann, Structural andnanomechanical properties of paperboard coatings studied by peak force tapp-ing atomic force microscopy, ACS Applied Materials & Interfaces 4 (10) (2012)5534–5541.

7 Physi

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

08 E. Thormann et al. / Colloids and Surfaces A:

22] T.J. Young, M.A. Monclus, T.L. Burnett, W.R. Broughton, S.L. Ogin, P.A. Smith,The use of the PeakForce (TM) quantitative nanomechanical mapping AFM-based method for high-resolution Young’s modulus measurement of polymers,Measurement Science & Technology 22 (12) (2011).

23] K. Sweers, K. van der Werf, M. Bennink, V. Subramaniam, Nanomechani-cal properties of alpha-synuclein amyloid fibrils: a comparative study bynanoindentation, harmonic force microscopy, and Peakforce QNM, NanoscaleResearch Letters (2011) 6.

24] M.E. Dokukin, I. Sokolov, Quantitative mapping of the elastic modulus of softmaterials with HarmoniX and Peak Force QNM AFM modes, Langmuir 28 (46)(2012) 16060–16071.

25] W.A. Ducker, T.J. Senden, R.M. Pashley, Measurement of forces in liquids usinga force microscope, Langmuir 8 (7) (1992) 1831–1836.

26] W.A. Ducker, T.J. Senden, R.M. Pashley, Direct measurement of colloidal forcesusing an atomic force microscope, Nature 353 (6341) (1991) 239–241.

27] C.P. Green, H. Lioe, J.P. Cleveland, R. Proksch, P. Mulvaney, J.E. Sader, Normaland torsional spring constants of atomic force microscope cantilevers, Reviewof Scientific Instruments 75 (6) (2004) 1988–1996.

28] J.E. Sader, J.W.M. Chon, P. Mulvaney, Calibration of rectangular atomic forcemicroscope cantilevers, Review of Scientific Instruments 70 (10) (1999)3967–3969.

29] E. Thormann, T. Pettersson, P.M. Claesson, How to measure forces with atomicforce microscopy without significant influence from nonlinear optical leversensitivity, Review of Scientific Instruments 80 (9) (2009).

30] T. Pettersson, N. Nordgren, M.W. Rutland, Comparison of different meth-ods to calibrate torsional spring constant and photodetector for atomic forcemicroscopy friction measurements in air and liquid, Review of Scientific Instru-ments 78 (9) (2007).

31] E. Thormann, A.C. Simonsen, P.L. Hansen, O.G. Mouritsen, Interactions betweena polystyrene particle and hydrophilic and hydrophobic surfaces in aqueoussolutions, Langmuir 24 (14) (2008) 7278–7284.

32] R. Bodvik, A. Dedinaite, L. Karlson, M. Bergstrom, P. Baverback, J.S. Pedersen, K.Edwards, G. Karlsson, I. Varga, P.M. Claesson, Aggregation and network forma-tion of aqueous methylcellulose and hydroxypropylmethylcellulose solutions,Colloids and Surfaces A—Physicochemical and Engineering Aspects 354 (1–3)(2010) 162–171.

33] R. Bodvik, E. Thormann, L. Karlson, P.M. Claesson, Temperature-dependentadsorption of cellulose ethers on silica and hydrophobized silica immersed inaqueous polymer solution, RCS Advances 1 (2) (2011) 305–314.

34] R. Bodvik, E. Thormann, L. Karlson, P.M. Claesson, Temperature responsive sur-face layers of modified celluloses, Physical Chemistry Chemical Physics 13 (10)(2011) 4260–4268.

35] E. Thormann, D.R. Evans, V.S.J. Craig, Experimental studies of the dynamicmechanical response of a single polymer chain, Macromolecules 39 (18) (2006)6180–6185.

36] E. Thormann, H. Mizuno, K. Jansson, N. Hedin, M. Soledad Fernandez, J. LuisArias, M.W. Rutland, R.K. Pai, L. Bergstrom, Embedded proteins and sacrificialbonds provide the strong adhesive properties of gastroliths, Nanoscale 4 (13)(2012) 3910–3916.

37] E. Thormann, P.L. Hansen, A.C. Simonsen, O.G. Mouritsen, Dynamic force spec-troscopy on soft molecular systems: improved analysis of unbinding spectrawith varying linker compliance, Colloids and Surfaces B—Biointerfaces 53 (2)(2006) 149–156.

38] A. Dedinaite, E. Thormann, G. Olanya, P.M. Claesson, B. Nystrom, A.-L. Kjoniksen,K. Zhu, Friction in aqueous media tuned by temperature-responsive polymerlayers, Soft Matter 6 (11) (2010) 2489–2498.

39] E. Thormann, S.H. Yun, P.M. Claesson, J. Linnros, Amontonian friction induced

by flexible surface features on microstructured silicon, ACS Applied Materials& Interfaces 3 (9) (2011) 3432–3439.

40] E. Thormann, A.C. Simonsen, P.L. Hansen, O.G. Mouritsen, Force trace hysteresisand temperature dependence of bridging nanobubble induced forces betweenhydrophobic surfaces, ACS Nano 2 (9) (2008) 1817–1824.

[

[

cochem. Eng. Aspects 441 (2014) 701– 708

41] P.M. Hansson, A. Swerin, J. Schoelkopf, P.A.C. Gane, E. Thormann, Influence ofsurface topography on the interactions between nanostructured hydrophobicsurfaces, Langmuir 28 (21) (2012) 8026–8034.

42] H.J. Butt, B. Cappella, M. Kappl, Force measurements with the atomicforce microscope: technique, interpretation and applications, Surface ScienceReports 59 (1–6) (2005) 1–152.

43] A. Carambassis, L.C. Jonker, P. Attard, M.W. Rutland, Forces measured betweenhydrophobic surfaces due to a submicroscopic bridging bubble, PhysicalReview Letters 80 (24) (1998) 5357–5360.

44] J.W.G. Tyrrell, P. Attard, Atomic force microscope images of nanobubbles ona hydrophobic surface and corresponding force-separation data, Langmuir 18(1) (2002) 160–167.

45] P. Attard, Nanobubbles and the hydrophobic attraction, Advances in Colloidand Interface Science 104 (2003) 75–91.

46] R.N. Ibbett, K. Philp, D.M. Price, C-13 NMR-studies of the thermal-behavior ofaqueous-solutions of cellulose ethers, Polymer 33 (19) (1992) 4087–4094.

47] M.K. Vyas, B. Nandan, K. Schneider, M. Stamm, Nanowear studies in chemicallyheterogeneous responsive polymeric brushes by surface force microscopy,European Polymer Journal 45 (5) (2009) 1367–1376.

48] D.P. Chang, J.E. Dolbow, S. Zauscher, Switchable friction of stimulus-responsivehydrogels, Langmuir 23 (1) (2007) 250–257.

49] N. Nordgren, M.W. Rutland, Tunable nanolubrication between dual-responsivepolyionic grafts, Nano Letters 9 (8) (2009) 2984–2990.

50] G. Duner, E. Thormann, O. Ramstrom, A. Dedinaite, Letter to the Editor:friction between surfaces polyacrylic acid brush and silica mediated by cal-cium ions, Journal of Dispersion Science and Technology 31 (10) (2010)1285–1287.

51] K. Chawla, S. Lee, B.P. Lee, J.L. Dalsin, P.B. Messersmith, N.D. Spencer, Anovel low-friction surface for biomedical applications: modification ofpoly(dimethylsiloxane) (PDMS) with polyethylene glycol (PEG)–DOPA–lysine,Journal of Biomedical Materials Research Part A 90A (3) (2009)742–749.

52] T. Drobek, N.D. Spencer, Nanotribology of surface-grafted PEG layers in anaqueous environment, Langmuir 24 (4) (2008) 1484–1488.

53] W. Hartung, T. Drobek, S. Lee, S. Zuercher, N.D. Spencer, The influence ofanchoring-group structure on the lubricating properties of brush-forminggraft copolymers in an aqueous medium, Tribology Letters 31 (2) (2008)119–128.

54] M. Muller, S. Lee, H.A. Spikes, N.D. Spencer, The influence of moleculararchitecture on the macroscopic lubrication properties of the brush-like co-polyelectrolyte poly(l-lysine)-g-poly(ethylene glycol) (PLL-g-PEG) adsorbedon oxide surfaces, Tribology Letters 15 (4) (2003) 395–405.

55] C. Friedsam, A.D. Becares, U. Jonas, M. Seitz, H.E. Gaub, Adsorption of polyacrylicacid on self-assembled monolayers investigated by single-molecule force spec-troscopy, New Journal of Physics (2004) 6.

56] C. Friedsam, H.E. Gaub, R.R. Netz, Adsorption energies of single charged poly-mers, Europhysics Letters 72 (5) (2005) 844–850.

57] C. Friedsam, M. Seitz, H.E. Gaub, Investigation of polyelectrolyte desorption bysingle molecule force spectroscopy, Journal of Physics—Condensed Matter 16(26) (2004) S2369–S2382.

58] M. Seitz, C. Friedsam, W. Jostl, T. Hugel, H.E. Gaub, Probing solid surfaces withsingle polymers, ChemPhysChem 4 (9) (2003) 986–990.

59] F. Kuehner, M. Erdmann, L. Sonnenberg, A. Serr, J. Morfill, H.E. Gaub, Friction ofsingle polymers at surfaces, Langmuir 22 (26) (2006) 11180–11186.

60] N. Canter, Negative friction coefficient, Tribology & Lubrication Technology 69(1) (2013) 10–11.

61] Z. Deng, A. Smolyanitsky, Q. Li, X.-Q. Feng, R.J. Cannara, Adhesion-dependent

negative friction coefficient on chemically modified graphite at the nanoscale,Nature Materials 11 (12) (2012) 1032–1037.

62] J. Njoroge, Chemically modified graphite yields adhesion-dependent negativefriction coefficient, MRS Bulletin 38 (2) (2013) 118.

63] E. Thormann, Negative friction coefficients, Nature Materials 12 (6) (2013) 468.