- Email: [email protected]
Surface and friction forces between grafted polysaccharide layers in the absence and presence of surfactant
Journal of Colloid and Interface Science 364 (2011) 351–358
Contents lists available at SciVerse ScienceDirect
Journal of Colloid and Interface Science www.elsevier.com/locate/jcis
Surface and friction forces between grafted polysaccharide layers in the absence and presence of surfactant Cathy E. McNamee a,⇑, Shinpei Yamamoto b, Michael Kappl c, Hans-Jürgen Butt c, Ko Higashitani d, Andra De˙dinaite˙ e,f, Per M. Claesson e,f a
Shinshu University, Tokida 3-15-1, Ueda-shi, Nagano-ken 386-8567, Japan Institute for Integrated Cell-Material Sciences, Kyoto University, Yoshida-Ushinomiyacho, Sakyo-ku, Kyoto 606-8501, Japan Max Planck Institute for Polymer Research, Ackermannweg 10, 55128 Mainz, Germany d Department of Chemical Engineering, Kyoto University – Katsura, Nishikyo-ku, Kyoto 615-8510, Japan e Department of Chemistry, Surface and Corrosion Science, Royal Institute of Technology, Drottning Kristinas väg 51, SE-100 44 Stockholm, Sweden f Institute for Surface Chemistry, YKI, P.O. Box 5607, SE-114 86 Stockholm, Sweden b c
a r t i c l e
i n f o
Article history: Received 1 July 2011 Accepted 25 August 2011 Available online 1 September 2011 Keywords: Dextran Silane coupling agent Polymer layer Silica Friction Surface forces Adhesion Tethered polysaccharide AFM Ellipsometry
a b s t r a c t We analyzed the interaction between chemically grafted polysaccharide layers in aqueous solutions. To fabricate such layers, an end-terminated dextran silane coupling agent was synthesized and the polydextran was grafted to oxidized silicon wafers and to silica particles. This resulted in the formation of a 28 nm thick layer (in air) and a grafted amount of 40 mg/m2 as determined by ellipsometry. The physical properties of the grafted layer were investigated in aqueous solutions by atomic force microscope imaging and colloidal probe force measurements. Surface and friction forces were measured between one bare and one polydextran coated silica surface. A notable feature was a bridging attraction due to affinity between dextran and the silica surface. Surface interactions and friction forces were also investigated between two surfaces coated with grafted polydextran. Repulsive forces were predominant, but nevertheless a high friction force was observed. The repulsive forces were enhanced by addition of sodium dodecyl sulfate (SDS) that associates with the tethered polydextran layers. SDS also decreased the friction force. Our data suggests that energy dissipation due to shear-induced structural changes within the grafted layer is of prime importance for the high friction forces observed, in particular deformation of protrusions in the surface layer. Ó 2011 Elsevier Inc. All rights reserved.
1. Introduction Knowledge and understanding of physico-chemical and mechanical properties of surfaces are needed for designing materials for optimum performance, e.g. for uses as biomaterials. An implant surface should be non-toxic, display biocompatibility to neighboring cells, and possess pseudo-physiological mechanical properties [1]. For example, materials to be used in orthopaedic/ periodontal applications as hard tissue substitutes should be tough but flexible [2]. Depending on the intended use, other requirements also need to be fulfilled. Biomaterials to be used in tissue engineering applications, such as the scaffolds used in regenerative medicine to promote cell growth on implants, only serve as a temporary matrix and should therefore be biodegradable [1,6]. The coating of a material that is to be used in a joint area must be resistant enough that it will not be eroded by the combined action of shear and load [3,4]. It should also display enough lubricity to
⇑ Corresponding author. E-mail address: [email protected] (C.E. McNamee). 0021-9797/$ - see front matter Ó 2011 Elsevier Inc. All rights reserved. doi:10.1016/j.jcis.2011.08.065
allow the two surfaces to move against one another without a large friction [5,4], which would otherwise impair the relative motion of the constituents in the joint. Thus, in this case important material properties are hardness, elasticity, low wear and lubricity. Polysaccharides are becoming increasingly popular as biomaterials for use in the body. This is due to their low toxicity, low cost, and large-scale availability [1,6–8]. Physically adsorbed layers of polysaccharides have been shown to provide a low friction coefficient when two such modified surfaces are sheared against each other [9–11]. This suggests that polysaccharides can be used for producing films with favorable lubricating properties. However, physically adsorbed polymers most often have a limited load bearing capacity since they are relatively easily removed under the combined action of load and shear. This limitation may be overcome by chemically attaching the polysaccharide, which has the potential to achieve high lubrication ability and high load bearing capacity even though wear of the film can be a problem during prolonged sliding. A film consisting of chemically attached polysaccharides would allow the thickness and density of the film to be controlled by choosing the polysaccharide type, molecular weight, and reactant
352
C.E. McNamee et al. / Journal of Colloid and Interface Science 364 (2011) 351–358
Dextran-Lactone was obtained. Next, 3-aminopropyltriethoxysilane (0.2 g) was added to a solution of Dextran-Lactone (1.5 g) in DMSO (15 mL) and ethanol (7.5 mL), corresponding to concentrations of 0.9 mM and 0.075 mM for 3-aminopropyltriethoxysilane and Dextran-Lactone, respectively. The mixture was then stirred for 20 h at 60 °C. As the reaction solution was concentrated, a large amount of acetone was subsequently added. Finally, the white precipitate was collected and dried in vacuo to give the dextransilane-coupling agent (DEX-si). The structure of the compound is shown in Fig. 1. The silica wafers, cleaned by solvents washing (acetone, ethanol, and then water) and then by plasma treatment (PDC002, Harrick Plama, Ithaka, NY)), and silica particles were rendered fully hydroxylated [16] by placing them in a glass container and adding 3 mL of a solution containing 17.5 mL methanol (high purity, Wako, Japan) and 1.39 g of 28% ammonium hydroxide solution (Wako, Japan) for 2 h, after which a freshly prepared solution of 0.1 g DEX-si in 1 mL water was added drop-wise to the wafer/particle containing solution. This solution was then left for 16 h, after which the substrate/particles were washed with copious amounts of water and ethanol to remove non-chemically adsorbed material. The presence of the silane coupling agent facilitates chemical grafting to the surface and also results in chemical binding between the Dex-si molecules. The end result is a firmly attached and partly cross-linked dextran layer on the surface. The grafted layer on flat silica surfaces was characterized by ellipsometry (M-2000U, J.A. Woolam, USA) measurements in air. The colloid probes for the force and friction measurements were prepared by evaporating the solvent from a small volume of the particle dispersion, and then attaching a single particle to a cantilever by using an XYZ micromanipulator and an epoxy resin (rapid araldite, Vantico, Showa polymer company, Tokyo, Japan). Tipless V-shaped gold-plated Si3N4 cantilevers (normal spring constant 0.12 N m1 and 0.58 N m1, NP-S, Veeco NanoProbe™ Tips, USA) and rectangular cantilevers (tip-less, non-Al coated, nominal spring constant 0.08 N m1, CSC12/tipless/noAl, MikroMasch, USA) were used for the force and friction measurements, respectively.
concentration. The lubricating ability of a film is affected by the presence of other species in the surrounding fluid. For instance, the addition of a surfactant, sodium dodecyl sulfate, between two surfaces coated with a mucin layer has been shown to decrease the friction significantly [9]. Therefore, naturally occurring compounds, such as phospholipids and proteins, may affect the lubrication of a grafted polysaccharide layer and enhance the lubrication ability of the manufactured biomaterial. For other applications, this may also be achieved by adding man-made surfactants. We have synthesized an end-terminated dextran silane coupling agent (DEX-si) and chemically attached it to silica surfaces. The physical properties of DEX-si modified silica surfaces were studied with an atomic force microscope (AFM) in order to obtain information on layer morphology, and surface and friction forces. The effect of introducing a surfactant into the surrounding aqueous solution on the physico-chemical properties of the surface bound polymer layer was studied by using different concentrations of sodium dodecyl sulfate (SDS). 2. Experimental 2.1. Materials Dextran (DEX, molecular weight (Mw = 20 kg/mol), GR grade, Nacalai Tesque, Japan) was used without further purification. The polydispersity (Mw/Mn) of the dextran sample was determined to be 1.42 by employing poly(ethyleneglycol)-calibrated GPC (Shodex GPC-101, SHOWA DENKO, Japan) using water as the eluent. The radius of gyration (Rg) of a typical polymer in our DEX sample in water was estimated to be about 5 nm, using literature data on Rg versus Mw [12]. The contour length of a typical dextran chain can be estimated to be 32 nm from the Mw and the length of one maltose unit, 0.52 ± 0.03 nm [13]. Potassium hydroxide (GR grade, Nacalai Tesque, Japan), ethanol (GR grade, Nacalai Tesque, Japan), dimethylsulfoxide (DMSO, GR grade, Nacalai Tesque, Japan), 3-aminopropyltriethoxysilane (GR grade, Nacalai Tesque, Japan), iodine (99.9% purity, Wako Pure Chemical, Japan), and sodium dodecyl sulfate (SDS, 95% purity, Wako, Japan) were used without further purification. Naturally oxidized silicon wafers (Silicon Quest INT., USA) were used as the substrates. The water used was distilled and de-ionized to give a resistivity of 18.2 MX cm and a total organic content of <5 ppm. The non-modified and DEX-si modified silica particles had a radius of 3.4 lm (Bangs Laboratory, USA) and 3 lm (micromer, micromod Partikeltechnologie GmbH, Rostock, Germany), respectively. Dextran having a lactonized end (Dextran-Lactone) was prepared according to reported methods (Fig. 1) [14,15]. Briefly, 0.1 M iodine in water and 0.8 M potassium hydroxide in water were added slowly to a 0.01 M dextran aqueous solution. The reaction mixture was stirred at ambient temperature for 24 h. It was then placed in a cellulose dialyzer tubing (Mw cut-off, 14 000 g/mol) and purified by dialysis. The solution was subsequently passed through a column packed with a cation-exchange resin (Amberlite IR-120), lyophilized, and a white powder of
OH OH
O
I2, KOH
OH O OH
OH
93-123
Amberlite IR 120
O OH
OH OH
OH
2.2.1. Atomic force microscopy The substrates were imaged in aqueous solutions with an AFM (Digital Instruments NanoScope III Multimode, Santa Barbara, USA) in tapping mode by using a tapping mode fluid cell and Si3N4 cantilevers (NP-S, Veeco NanoProbe™ Tips, USA) with reflective gold coating. The V-shaped cantilevers were 196 lm long and 23 lm wide, and had a nominal spring constant of 0.06 m1, a resonant frequency of 31 kHz (in air), and square pyramidal tips with an end-tip radius of curvature in the range 10–40 nm. The imaging scan rate was 2 Hz, and the scan size was 1 1 lm2. Both the resolution of data points per line and the number of lines were 512. All images are unmodified (i.e., no filtering was applied), except for a 1st order flattening along the scan lines. The images shown for each condition are representative of the images made from at OH
OH
OH
2.2. Methods
O OHO OH OH
-H2O 93-123
OH OHO
OH OH
OH O OHO
+H 2O OH
OH OH
O
OH DMSO/EtOH OH
OHO
93-123
O OH O
NH2
Si(OEt)3
OH OH
93-123
OH H N OH O
DEX -si dextran Fig. 1. Schematic diagram showing the synthesis of dextran-triethoxy silane (DEX-si).
Si(O Et) 3
C.E. McNamee et al. / Journal of Colloid and Interface Science 364 (2011) 351–358
least two different samples and at a minimum of two different places on each sample. The topography and physical features were the same, regardless of the sample and imaging position. The images were reproducible upon multiple scans, and did not change if they were scanned from top to bottom or from bottom to top. Imaging over a wider scale of 4 4 lm2 gave smaller but similar topographical features as that obtained in the 1 1 lm2 scale image, indicating the absence of tip artefacts. Surface forces were measured between a substrate and a colloid probe in aqueous solution as a function of separation relative to hard wall contact by using the AFM and a contact mode fluid cell. A syringe was used to inject solutions into the fluid cell. The force measurements were carried out by bringing the cantilever into contact with the substrate, during which time the change in the deflection of the cantilever (Dx) was measured as a function of the piezo displacement with a split photodiode. The detector signal versus piezo position curves were converted to force versus distance curves using the conventional procedure. First, a linear fit of the zero force line (i.e. the signal at large cantilever-substrate separations where no surface forces were acting) was subtracted. The zero force position was defined at large cantilever-substrate separations, where no surface forces were acting on the cantilever. The conversion factor between the detector signal in volts and the cantilever deflection in nm was then determined from the slope of the linear compliance region, which occurred when the probe and the substrate were moving in parallel, i.e. when they were in hard wall contact. The force curves were then obtained by subtracting the cantilever deflection (Dx) from the piezo position and using Hooke’s law, F = kDx. Here, F and k are the force and the nominal spring constant of the cantilever, respectively. The lateral (friction) force was measured using a Digital Instruments NanoScope III multimode microscope in the ‘‘friction mode’’ and a liquid fluid cell. Briefly, the colloid probe was pressed against the substrate in solution at a constant applied load, while the substrate slid horizontally underneath the cantilever at a speed of 3.60 lm/s. The magnitude of the lateral frictional force (FF) was determined using
FF ¼
V L SL K L 2H
ð1Þ
where VL, SL, KL, and H are the difference in the lateral force detector signal in one complete scan, the lateral detector sensitivity, the lateral spring constant, and the distance from the bottom of the sphere to the mid-point of the cantilever, respectively. The value of SL was determined to be 3.1 104 rad/V from the method of Meurk and others [17]. The value of KL was determined from
KL ¼
2K N L2 3ð1 þ mÞ
ð2Þ
Here, KN and L are the normal spring constant and the length of the cantilever, respectively. The symbol m is the Poisson’s ratio. A typical value [18] of 0.27 was used. The values of FF for each system and condition were averaged from four separate experiments using different substrates and probes. The error in each FF measurement is estimated to be 0.6 nN. Each friction measurement was made at a different position on the substrate. The expressions used to relate FF to the applied normal load (FN) depend on whether an intermolecular attraction between the surfaces or the external load dominates [19]. In the non-adhesive case the following equation was used:
F F ¼ ðF N F 0 Þl; when F n > F 0
ð3Þ
Here, F0 is the load at which Amontons’ law starts to apply, and l is the friction coefficient. The effective coefficient of friction (leff) is defined as
leff ¼
FF FN
353
ð4Þ
and when F0 = 0, l = leff. This load-dependent model is similar to the one used by Ecke and others in an earlier study [20]. That model also gave a roughly linear increase with load, but showed some deviation for low loads. The following relation can be used when FF is adhesion controlled,
F F ¼ Sc A þ F N l
ð5Þ
Here, Sc is the critical shear stress, A is the contact area, and FN is the normal load. The JKR theory [21] can be used to calculate A,
qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi2=3 R A ¼ p F N 6pRc þ 12pRcF N þ ð6pRcÞ2 E
ð6Þ
Here, R is the particle radius, c is the interfacial tension that can be calculated using the adhesion force obtained from the normal force curve,
c¼
F adh 3pR
ð7Þ
E⁄ is the effective elasticity of the system. We estimate it from the Young’s modulus (E) and Poisson ratio (m) of silica using
E ¼
1 2ð1 m2silica Þ Esilica
ð8Þ
with Esilica = 72 GPa and msilica = 0.17 [22]. 3. Analysis The total interaction free energy (Etot) per unit area can have several contributions, which can be considered as being approximately additive. In this work we encounter three contributions due to (i) repulsive double layer interactions (EDL), (ii) attractive van der Waals interactions (EvdW), and (iii) polymer induced interactions that include steric repulsion and attractive bridging interactions (Epol). Thus,
Etot ¼ EDL þ Ev dw þ Epol
ð9Þ
The repulsive double layer interaction at a distance x was calculated using constant surface potential boundary conditions, assuming equal surface potentials, or using constant surface charge densities boundary conditions, assuming equal charge densities on the two surfaces. We used the approximate equations, valid for small midpoint potentials [23]. The van der Waals interaction was described using a non-retarded Hamaker constant. The force (F) between the spherical particle and the substrate can be related to Etot using the Derjaguin approximation (Eq. (10)), provided the particle radius (R) is much larger than the separation distance between the particle and the substrate. This condition is fulfilled.
F ¼ 2pEtot R
ð10Þ
4. Results and discussion 4.1. Layer characteristics and surface morphology Ellipsometry measurements in air showed that the DEX-si film thickness, t, was 27.9 ± 2.8 nm, evaluated using a typical refractive index of 1.54 for polysaccharides. This rather large thickness indicates that the trifunctional silane not only binds DEX-si to the surface, but also allows DEX-si polymers to chemically bind to each
354
C.E. McNamee et al. / Journal of Colloid and Interface Science 364 (2011) 351–358
Fig. 2. 1 1 lm2 tapping mode image of an unmodified SiO2 wafer in 30 mM NaNO3. (A) Height image. (B) Phase image.
other. The corresponding mass of the grafted layer, C, was calculated from the density, q, of the polymer (q = 1.43 g/cm3 for dextran [24]) and the layer thickness as:
ð11Þ
It was found that C = 39.9 ± 4.0 mg/m2, which is of the order of a factor of 10 larger than typically found for physisorbed polymer layers. The surface morphology of bare silica wafers (Fig. 2) and silica wafers modified with DEX-si (Fig. 3A and B) were characterized in 30 mM NaNO3 using tapping mode AFM. The images of the bare silica surface are featureless and demonstrate the flat nature of the substrate. In contrast, surface features are seen on the DEX-si modified surfaces, and most clearly in the phase images. Surface features with widths 25 nm and heights between 10 and 15 nm were found to be characteristic of the layer (see height profile of Fig. 4A). From the images, we conclude that the surfaces coverage is non-homogeneous. We note that the dry thickness of the grafted DEX-si layer as measured by ellipsometry was about 28 nm. The polymer layer swells when placed in water, a good solvent for dextran, and the total polymer thickness should be larger than the dry thickness. Thus, the relatively small height variations observed by AFM in aqueous solutions demonstrates that the AFM tip does not reach the bare silica surface. We conclude that the silica surfaces are fully covered by DEX-si, but that the thickness of the layer is not homogeneous. The height and phase images of silica surfaces carrying grafted DEX-si changed when SDS was added to the NaNO3 solution (Fig. 3C–H). The lateral area of the protrusions increased as the concentration of SDS increased from below the critical micelle concentration (cmc) (1 mM, Fig. 3C and D), to a value just above the cmc (5 mM, Fig. 3E and F), and to a value significantly above the cmc (10 mM, Fig. 3G and H); the cmc of SDS in the presence of 30 mM NaNO3 is reported to be 3.3 mM [25]. The height profiles (Fig. 4B–D) corresponding to the lines in the height images demonstrate that the surface features become smeared out in the presence of SDS. A similar effect of SDS on the morphology of physisorbed mucin–chitosan complexes has previously been noted [26]. These results indicate that SDS adsorbs to the DEX-si coated substrate at concentrations below its cmc and whereby expand the polymer chains. We suggest that the adsorption is driven by the hydrophobic tails of the SDS, which by association with the surface bound dextran chains can reduce its exposure to bulk water. Typically, anionic surfactants bind cooperatively to non-ionic polymers at about 0.5 cmc. This has for example been shown for linear poly(ethylene oxide) [27–29], bottle-brush polymers with poly(ethylene oxide) side chains [30,31], poly(methoxyhexa(ethylene glycol) methacrylates), and non-ionic triblock copolymers [32,33]. 4.2. Surface forces 4.2.1. Effect of SDS on interactions between bare silica surfaces Forces between a silica particle and a silica substrate across 30 mM NaNO3 in the absence and presence of SDS were found to
Fig. 3. 1 1 lm2 tapping mode images of DEX-si modified SiO2 wafers in solutions. Height images are provided in the left column and phase images in the right column. The black lines show the position of the height profiles provided in Fig. 4. (A and B) 30 mM NaNO3; (C and D) 30 mM NaNO3 + 1 mM SDS; (E and F) 30 mM NaNO3 + 5 mM SDS; (G and H): 30 mM NaNO3 + 10 mM SDS. The z-scale in all the height and phase images covers the range 0–30 nm and 0–5°, respectively.
(A)
10 0
Height (nm)
C ¼ tq
10
(B)
0
(C)
10 0
(D)
10 0 0
100
200
300
400
500
x-profile (nm) Fig. 4. Height profiles of the lines shown in the 1 1 lm2 tapping mode height images of DEX-si modified SiO2 wafer in solutions. (A) 30 mM NaNO3; (B) 30 mM NaNO3 + 1 mM SDS; (C) 30 mM NaNO3 + 5 mM SDS; (D) 30 mM NaNO3 + 10 mM SDS.
be repulsive, having a roughly exponential decay at large separations (Fig. 5). The addition of SDS to the solution did not affect the measured forces significantly. Comparison of the theoretical force curve calculated within the DLVO framework and using a Hamaker constant for the SiO2–water–SiO2 system of 0.46 1020 J
C.E. McNamee et al. / Journal of Colloid and Interface Science 364 (2011) 351–358
[34], with the experimental force data showed that the magnitude of the surface potential and Debye length were 15 mV and 1.8 nm, respectively. For comparison, the expected Debye-length is 1.8 nm and 1.5 nm for a solution of 30 mM NaNO3 in the absence and presence of 10 mM SDS. The surface potential of silica in 30 mM NaNO3 (15 mV) corresponds to a surface charge density of 6.0 mC/m2. The corresponding values reported for silica in 1 mM NaNO3 are 42 mV and 3.4 mC/m2, respectively [35] Thus, as expected for a charge regulating surface the surface potential decreases and the surface charge density increases as the inert salt concentration is increased [36]. The fact that the force curves were not significantly affected by SDS indicates that this surfactant does not adsorb to any significant degree to the bare silica surface, as also concluded from adsorption studies [37,38]. DLVO theory predicts an attractive net interaction at short separations. However, this was not observed in our experiments, in agreement with numerous other reports. The reason for this is that the silica surface is highly hydrated and a repulsion due to the dehydration of the surface layer overcomes the van der Waals force at short separations [39,40]. The measured force curves were independent of the approaching/retracting speed in the range investigated, 300–14 000 nm/s, and no hysteresis was detected between the approach and retraction curves. 4.2.2. Forces between one bare silica surface and one DEX-si modified silica surface in the absence and presence of SDS Force curves measured between a DEX-si coated silica plate and bare silica probe on approach demonstrate an attraction commencing at approximately 40 nm in 30 mM NaNO3 in the absence and presence of 1 mM SDS (Fig. 6). We interpret the attraction to be caused by bridging of polymer chains to the approaching silica particle, and the large range of this force as being a consequence of formation of DEX-si polymers. The DEX-si molecule has several – OH and –O– sites that may interact favorably with the silanol groups on the silica surface. Dextran has indeed previously been reported to bind to silica surfaces in water. The normalized energy as determined from the pull-off force (Fad/2pR) for dextran (Mw = 10 kDa) from a silica colloidal tip in water was reported to be 0.026 1018 J/nm2, i.e, 26 mN/m [41]. This strong affinity was suggested to be due to hydrogen-bond formation between dextran and silica, and the presence of such bonds was confirmed by infrared spectroscopy [41]. We note, however, that to state that adsorption is driven by dextran-silanol hydrogen bonding is a sim-
F/R (mN/m)
1
0.1
0.01
0
2
4
6
Separation (nm) Fig. 5. Examples of force curves measured between a SiO2 particle and a SiO2 wafer in solutions of 30 mM NaNO3 in the absence and presence of SDS. The measurements were carried out at a frequency of 1 Hz (speed = 1993 nm/s). The experimental data are given by h, j: 30 mM NaNO3; s, d: 30 mM NaNO3 + 1 mM SDS; D, : 30 mM NaNO3 + 5 mM SDS; e, : 30 mM NaNO3 + 10 mM SDS. The symbols represent the forces measured on approach and the solid symbols show the forces measured on separation. The theoretical force curves were calculated using constant potential (- - -) and constant charge (–) boundary conditions. The parameters used for the calculations were: w0 = 15 mV, 1/j = 1.8 nm, A = 0.46 1020 J, R = 3.42 lm.
355
plification as discussed for the case of poly(ethylene oxide)–silanol interactions [42]. The maximum normalized adhesion force in the 30 mM NaNO3 solution observed in the retraction force curves in our experiment was found to be 6 ± 0.3 mN/m. The strength of the attraction on approach was observed to decrease with increasing driving velocity (Fig. 6). This could be due to either hydrodynamic effects or chain relaxation effects. A precise calculation of the hydrodynamic force is complicated since the viscosity of the liquid in the polymer layer is higher than in bulk water. However, a simple estimation can be made by assuming the same viscosity as bulk water and the no-slip boundary condition. In this case, the normalized hydrodynamic force of a sphere approaching a planar surface [43,44] is:
F 6pgmR ¼ R x
ð12Þ
Here, g is the viscosity of the liquid (103 Pa s for water) and v is the approaching velocity. At a velocity of v = 10 lm/s and a radius of 3 lm at a distance x = 10 nm, the normalized hydrodynamic force is 0.06 mN/m. This is much lower than the forces observed. Thus, as any effect of hydrodynamic forces is negligible, we attribute the reduction of the attractive force to the finite time required for bridge formation. Taking the time the silica sphere requires to approach the dextran-coated surface to be the range of the attraction divided by the approaching speed, we conclude that more than 0.1 s is needed for formation of all bridges, and the relatively long relaxation time is suggested to be a consequence the presence of DEX-si– DEX-si cross-links within the layer. The hypothesis of bridge formation in the DEX-si–silica system is supported by the force curves measured on retraction (Fig. 6B, D and F). The retraction force curves recorded in 30 mM NaNO3 and in 30 mM NaNO3 with 1 mM SDS showed a strong attractive force with a range of 40–80 nm. Several smaller attractive maxima were observed at large distances, before the two surfaces were completely detached. Such multiple adhesion maxima are due to individual polymer chains detaching from the retracting silica surface [45]. Increasing the concentration of SDS in the solution to 5 mM and 10 mM resulted in removal of the attractive force observed on approach (Fig. 6A, C and E). Instead, a repulsive force was noted at separations below 20–30 nm. This distance is much longer than the Debye length associated with a double layer force in a 30 mM NaNO3 solution in the presence of 5 and 10 mM SDS, which is calculated to be 1.6 nm and 1.5 nm, respectively. Clearly, binding of SDS to DEX-si impairs the possibility of bridge formation and instead a repulsive force of electrosteric nature is encountered. However, on separation, a long-range attraction, extending to a separation of 50–70 nm, was observed both in the absence and presence of SDS, and the adhesion was somewhat larger in the presence of SDS. This suggests that when the two surfaces are in close contact SDS is partly desorbed from the layer, and when this occurs bridges can again be formed between DEX-si and the silica particle. The slight increased adhesion upon SDS addition is attributed to hydrophobic interactions between the tails of the SDS molecules that remain in the layer also under compression. Removal of SDS by compressing polymer-surfactant layers has also been observed in other systems [46]. The magnitude of the adhesion decreased marginally with increasing measurement speed, from 6 ± 0.3 mN/m at 0.3 Hz to 5±0.3 mN/m at 14 Hz. 4.2.3. Forces between two DEX-si modified silica surface in the absence and presence of SDS In most cases the forces measured between a DEX-si silica substrate and a DEX-si silica particle in 30 mM NaNO3 in the absence and presence of SDS were found to be repulsive both on approach and separation (Fig. 7). However, in a few cases a small attraction was noted on separation. Some hysteresis was observed in the
356
C.E. McNamee et al. / Journal of Colloid and Interface Science 364 (2011) 351–358
4
(A) 0.3 Hz
(C) 1 Hz
(E) 14 Hz
(B) 0.3 Hz
(D) 1 Hz
(F) 14 Hz
F/R (mN/m)
2 0 -2 4 2 0 -2 -4 -6 0
40
80
120
0
40
80
120
0
40
80
120
Separation (nm) Fig. 6. Examples of force curves measured between a SiO2 particle (R = 3.42 lm) and a DEX-si wafer in solutions of 30 mM NaNO3 with and without SDS. The measurements were performed at frequencies of 0.3 Hz (speed = 296 nm/s, A and B), 1 Hz (speed = 1993 nm/s, C and D), and 14 Hz (speed = 13 951 nm/s, E and F). h: 30 mM NaNO3; s: 30 mM NaNO3 + 1 mM SDS; D: 3 30 mM NaNO3 + 5 mM SDS; e: 30 mM NaNO3 + 10 mM SDS. The upper row displays forces measured on approach and the lower row illustrates forces measured on separation.
FN (nN) 20
0
10
20
30
(A) SiO2-SiO2
15 10 5 0 20
FF (nN)
force curves, where the forces measured on separation were lower than those encountered on approach. The hysteresis is assigned to slow chain relaxation processes in the layer. The somewhat uneven appearance of the force curve is attributed to the non-homogeneous nature of the dextran-coated surfaces (see Fig. 3). The distance dependence of the force is inconsistent with that of a double-layer force at the given ionic strength. Thus, it is concluded that it has a steric or electrosteric origin. The range and magnitude of the repulsion increased with increasing SDS concentration due to some binding of SDS to the dextran coated surfaces, which causes the polymer chains to become stretched for electrostatic reasons. We note that the range of the force is significantly less than twice the thickness of the layer as determined by ellipsometry. This is a consequence of the fact that the zero separation in the force curves determined with the AFM is defined as the ‘‘hard wall’’ contact between the compressed polymer layers and not as the distance from the bare surfaces.
(B) SiO2 - DEX-si
15 10 5 0
4.3. Friction forces
40
The friction force (FF) as a function of load (FN, FN/R) for the SiO2–SiO2, SiO2–DEX-si, and DEX-si–DEX-si systems in 30 mM NaNO3 in the absence and presence of SDS are shown in Fig. 8.
20
F/R (mN/m)
0
(C) DEX-si - DEX-si
0
2
4
6
8
10
FN/R (mN/m) 1
0.1
0
5
10
15
20
Separation (nm) Fig. 7. Typical force curves measured between a DEX-si particle (R = 3.0 lm) and a DEX-si wafer in solutions of 30 mM NaNO3 in the absence and presence of SDS. The measurements were carried out at a frequency of 1 Hz (speed = 1993 nm/s). The experimental data are given by h: 30 mM NaNO3; s: 30 mM NaNO3 + 1 mM SDS; D: 30 mM NaNO3 + 5 mM SDS; and e: 30 mM NaNO3 + 10 mM SDS.
Fig. 8. Friction force versus load measured between (A) two SiO2 surfaces, (B) SiO2 and DEX-si, and (C) two DEX-si surfaces in solutions of 30 mM NaNO3 in the absence and presence of SDS. No SDS: h; 1 mM SDS: ; 5 mM SDS: ; 10 mM SDS: e. The load controlled fits to the experimental data are given by –: no SDS; : 1 mM SDS; : 5 mM SDS; : 10 mM SDS. The adhesion controlled fits are shown as and , obtained using l = 0, Esilica = 72 Gpa, msilica = 0.17, Fad/R = 0.10 and R = 3 lm. The upper and lower x-axes show the FN and FN/R data, respectively. (For color interpretation in this figure legend the reader is referred to see the web version of this article.)
The plotted FF data represents the average FF values obtained from at least four independent experiments, where different substrates, solutions and probes were used each time. The average variation in the FF data set for the SiO2–SiO2, SiO2–DEX-si, and DEX-si–DEX-si
357
C.E. McNamee et al. / Journal of Colloid and Interface Science 364 (2011) 351–358
Table 1 Friction parameters obtained from the load controlled fit of the FF versus FL/R data of Fig. 8. F0 and l are the load at which Amontons’ law starts to apply and the coefficient of friction, respectively. SDS conc. (mM)
0 1 5 10 a
SiO2–SiO2
SiO2-DEX-si
DEX-si–DEX-si
F0/R (mN/m)
l
F0/R (mN/m)
l
F0/R (mN/m)
l
0 0 0 0
0.42 ± 0.00 0.14 ± 0.01 0.05 ± 0.00 0.02 ± 0.01
0.19 ± 0.29 0 0 0
0.42 ± 0.03 0.28 ± 0.02 0.18 ± 0.02 0.24 ± 0.02
–a –a 0 0
–a –a 0.19 ± 0.01 0.20 ± 0.02
The friction force was always found to be high as indicated in Fig. 8, but the scatter in the data was large.
Table 2 Friction parameters obtained from the Adhesion controlled fit of the FF versus FL/R data of Fig. 8. Sc and l are the critical shear stress and the coefficient of friction, respectively. SDS conc. (mM)
0 1 5 10
SiO2–DEX-si Sc (Mpa)
l
12 8 8 8
0 0 0 0
systems were 0.8, 4.4, and 2.6 nN, and in each experimental series the same qualitative changes in the FF values with SDS concentration were observed. No hysteresis was observed when the load was increased and then subsequently decreased, indicating the absence of significant wear of the surfaces. The FF versus FL/R data was fitted using Eq. (3) (Fig. 8) for all cases, and also by Eq. (5) (Fig. 8) for SiO2–DEX-si case, as adhesion was observed for this system. The F0/R and l values obtained from the load-controlled model are shown in Table 1, whereas the corresponding values obtained from the adhesion-controlled model are shown in Table 2. In these latter calculations, the value of Fad/R determined from the force curves was used. 4.3.1. The silica–silica case As expected, the value of FF increased with load for the SiO2 substrate–SiO2 particle case in 30 mM NaNO3 in the absence and presence of SDS. The results are consistent with the load-controlled model. We note that at low forces (F/R < 0.7 mN/m) the doublelayer force carries the load. In this case a highly fluid layer separates the surfaces and the expected friction force is below the detection limit of our instrument [19]. The low load bearing capacity of the double-layer force means that the off-set value, F0 in Eq. (3), is expected to be small. This was also found when fitting the load-controlled model to the data (Table 1 reports F0/R). The interaction between the silica surfaces in 30 mM NaNO3 were repulsive on approach and retraction, with a magnitude, range and shape that did not change with the addition of SDS. Nevertheless, the frictional forces decreased when SDS was added. This effect, which was observed in all experiments, is surprising considering that SDS does not adsorb to silica. It may be due to some dodecanol adsorption, the hydrolysis product of SDS, but this should be regarded as speculation and the observation is not yet understood. 4.3.2. The silica–DEX-si case The friction forces between the DEX-si substrate and the bare SiO2 particle in the absence of SDS increased linearly with the applied loads for low to intermediate loads (FL/R < 6 mN/m). Thus, the load-controlled model, despite the adhesion, represents the data satisfactory. The use of the model for adhesion-controlled friction did not provide a better fit to the data (see Fig. 8). We note that
the friction increased considerably for FL/R greater than 6 mN/m. This indicates the appearance of a new friction mechanism when the water content of the layer is reduced below a critical value. A similar result with a low friction at low and medium loads and high friction at high loads has been reported for PLL–PEG brushes [47]. In that case, the low friction was explained as being a consequence of the osmotic pressure resulting from the compression of the polymer brushes, which lead to strong repulsive forces that limited the interpenetration of opposing polymer layers. At higher loads the water between the polymer brushes were partly squeezed out, resulting in a reduced fluidity of the end segments of the polymer chains, explaining the significantly higher friction forces at higher loads. We suggest the same mechanism in our case. The friction in presence of SDS, and in absence of SDS at high loads, for the SiO2–DEX-si system was higher than that for the SiO2–SiO2 case. This is attributed primarily to energy dissipative shear-induced changes within the polymer layer, e.g. deformation of the protrusions observed in the AFM images. The importance of energy dissipative mechanisms within a single polymer layer for the magnitude of the friction force has also been noted for a grafted polyacrylic acid layer sliding against a silica surface [48], but in that case the energy dissipative mechanism was suggested to be the breakage and formation of physical intra-layer bonds. The adhesion present between the Dex-si surface and the bare silica surface should also contribute to the friction, but this mechanism is not dominant as seen by the poor fit of the adhesioncontrolled model to the data in Fig. 8b. Addition of SDS caused FF to decrease in magnitude, particularly at high loads. This is consistent with the AFM images, Fig. 3, that show how the protrusions are smeared out due to the incorporation of SDS in the DEX-si layer, and thus shearing in presence of SDS is expected to lead to less deformation of the protrusions and less energy dissipation.
4.3.3. The DEX-si–DEX-si case The friction between the DEX-si substrate–the DEX-si particle in the absence of SDS and after the addition of 1 mM SDS (Fig. 8c) was found to be high. The high value of the friction force in these cases arises from the energy dissipative shear-induced structural changes occurring within each layer, amplified by the interlocking of the protrusions present on the two opposing layers and the interpenetration of the DEX-si chains upon the application of a load and a sliding motion. In this situation neither the load-controlled nor the adhesion-controlled model provided a good fit to the data. We note that Kampf et al. found a significant increase in friction between layers of the cationic polysaccharide chitosan when these layers were cross-linked [11]. The large grafted amount in our case can only be achieved due to formation of DEX-si, polymers, and the trifunctional silane group allows for cross-linking reactions. Thus, it appears that the presence of chemical cross-links in the layer provides for additional energy dissipative mechanisms associated with the shear-induced strain on these cross-links and the
358
C.E. McNamee et al. / Journal of Colloid and Interface Science 364 (2011) 351–358
subsequent relaxation to the equilibrium state. An increase in the SDS concentration to 5 mM and 10 mM resulted in a decrease in the value of FF, giving values in the order of those obtained for the SiO2–DEX-si case. Hence, the presence of SDS on the DEX-si surface decreases the friction. This is suggested to be due to smearing out of the surface protrusions and reduced chain interpenetration, both being facilitated by the electrostatic repulsion due to the presence of SDS. The latter mechanism has also been suggested to explain the lubricating effect of SDS for mucin coated surfaces [9]. 5. Conclusions Silica surfaces have been modified with a polydextrane-silane, DEX-si. Ellipsometry measurements demonstrate a high amount of surface bound DEX-si, about 40 mg/m2, which demonstrate the formation of DEX-si polymers. The trifunctional silane coupling agent allows for cross-linking reactions within the layer. AFM imaging of the resulting surfaces provides evidence for the formation of a non-homogeneous surface layer containing protrusions with heights up to about 10 nm. Measurements of the forces acting between one DEX-si modified surface and a bare silica surface demonstrates the presence of a long-range attractive force and a strong adhesion. This force is attributed to bridging. The attraction measured on approach was found to decrease with increasing approach rate. From this observation the relaxation time corresponding to bridge formation was estimated to be around 0.1 s. This relatively long time is suggested to a consequence of the presence of crosslinks in the layers formed. Addition of SDS decreased the attraction observed on approach and lead to repulsion at concentrations above the cmc. This effect is due to binding of SDS to the surface attached dextran layer, which counteracts bridge formation. The interactions between two DEX-si coated silica surfaces are repulsive due to steric forces. Addition of SDS increases the range of the repulsion somewhat due to incorporation of anionic surfactants in the layer. The friction force for the SiO2ADEX-si system in the absence of SDS was found to be relatively high, which is explained by shearinduced energy-dissipative structural changes within the DEX-si layer, e.g. deformation of the protrusions observed by AFM imaging. SDS addition reduced the friction force, and this effect was attributed to smoothening of the surface protrusions due to the incorporation of SDS in the DEX-si layer. The friction force between two DEX-si coated surfaces was found to be high in the absence and presence of small (1 mM) concentrations of SDS. This is also rationalized by energy-dissipative structural changes within each layer due to the action of shear and load, but the friction is amplified by interlocking of the protrusions and chain interpenetration. Addition of SDS reduces the friction force also in this case due to smoothening of the surface protrusions. The presence of SDS also counteracts chain interpenetration for electrostatic reasons. Acknowledgments This research was partly supported by the Core-to-Core Program promoted by Japan Society for the Promotion of Science (Project No. 18004). This study was also performed through the Program for Dissemination of Tenure-Track System funded by the Ministry of Education and Science, Japan. P.C. and A.D. acknowledge financial support from the SSF program ‘‘Microstructure, Corrosion and Friction Control’’.
References [1] F. Couet, N. Rajan, D. Mantovani, Macromol. Biosci. 7 (2007) 701–718. [2] E. Khor, L.Y. Lim, Biomaterials 24 (2003) 2339–2349. [3] S. Affatato, M. Spinelli, M. Zavalloni, C. Mazzega-Fabbro, M. Viceconti, Med. Eng. Phys. 30 (2008) 1305–1317. [4] J. Katta, Z. Jin, E. Ingham, J. Fisher, Med. Eng. Phys. 30 (2008) 1349–1363. [5] R. Crockett, Tribol. Lett. 35 (2009) 77–84. [6] M. Rinaudo, Polym. Int. 57 (2008) 397–430. [7] M. Rinaudo, Prog. Polym. Sci. 31 (2006) 603–632. [8] V.R. Sinha, A.K. Singla, S. Wadhawan, R. Kaushik, R. Kumria, K. Bansal, S. Dhawan, Int. J. Pharm. 274 (2004) 1–33. [9] T. Pettersson, A. De˙dinaite˙, J. Colloid Interface Sci. 324 (2008) 246–256. [10] D. Gourdon, Q. Lin, E. Oeoudjev, Y. Golan, J. Arad, J. Israelachvili, Langmuir 224 (2008) 1534–1540. [11] N. Kampf, U. Raviv, J. Klein, Macromolecules 37 (2004) 1134–1142. [12] C.E. Ioan, T. Aberle, W. Burchard, Macromolecules 33 (2000) 5730–5739. [13] M.S. Motawia, I. Damager, C.E. Olsen, B.L. Møller, S.B. Engelsen, S. Hansen, L.H. Øgendal, R. Bauer, Biomacromolecules 6 (2005) 143–151. [14] K. Hashimoto, S. Imanishi, M. Okada, H. Sumitomo, J. Polym. Sci., Pol. Chem. 29 (1991) 1271. [15] T. Zhang, R.E. Marchant, Macromolecules 27 (1994) 7302. [16] K. Ohno, T. Morinaga, K. Koh, Y. Tsujii, T. Fukuda, Macromolecules 38 (2005) 2137–2142. [17] A. Meurk, I. Larson, L. Bergström, Mater. Res. Soc. Symp. Proc. 552 (1998) 427. [18] P. Hess, Appl. Surf. Sci. 106 (1996) 433. [19] A. De˙dinaite˙, E. Thormann, G. Olanya, P.M. Claesson, B. Nyström, K. Zhu, Soft Matter 6 (2010) 2489–2498. [20] S. Ecke, H.-J. Butt, J. Colloid Interface Sci. 244 (2001) 432–435. [21] K.L. Johnson, K. Kendall, A.D. Roberts, Proc. Roy. Soc. London, Ser A 1324 (1971) 301. [22] M.J. Bamber, K.E. Cooke, A.B. Mann, B. Derby, Thin Solid Films 398–399 (2001) 299–305. [23] J. Israelachvili, Intermolecular and Surface Forces, third ed., Academic Press, London, 2010. [24] E.C. Gil, A.I. Colarte, B. Bataille, I. Caraballo, A.A.P.S. Pharm, Sci. Technol. 8 (2007) 281–288. [25] N. Lundin, L. Macakova, A. De˙dinaite˙, P. Claesson, Langmuir 24 (2008) 3814– 3827. [26] A. De˙dinaite˙, M. Lundin, L. Macakova, T. Auletta, Langmuir 21 (2005) 9502– 9509. [27] B. Cabane, R. Duplessix, J. Phys. (Paris) 43 (1982) 1529–1542. [28] E. Pettersson, D. Topgaard, P. Stilbs, O. Söderman, Langmuir 20 (2004) 1138– 1143. [29] R. Meszaros, I. Varga, T. Gilanyi, J. Phys. Chem. B 109 (2005) 13538–13544. [30] L.A. Bastardo, J. Iruthayaraj, M. Lundin, A. De˙dinaite˙, A. Vareikis, R. Makuška, A. van der Wal, I. Furo, V.M. Garamus, P.M. Claesson, J. Colloid Interface Sci. 312 (2007) 21–33. [31] N. Peron, R.A. Campbell, T. Nylander, A. Vareikis, R. Makuška, T. Gilyani, R. Meszaros, J. Phys. Chem. B 112 (2008) 7410–7419. [32] S. Couderc-Azouani, J. Sidhu, T. Thurn, R. Xu, D.M. Bloor, J. Penfold, J.F. Holzwarth, E. Wyn-Jones, Langmuir 21 (2005) 10197–10208. [33] S. Couderc-Azouani, J. Sidhu, T.K. Georgiou, D.C. Charalambous, M. Vamvakaki, C.S. Patrickios, D.M. Bloor, J. Penfold, J.F. Holzwarth, E. Wyn-Jones, Langmuir 20 (2004) 6458–6469. [34] H.-J. Butt, M. Kappl, Surface and Interfacial Forces, Wiley-VCH Verlag GmbH& Co., Weinheim, 2010. [35] P.G. Hartley, I. Larson, P.J. Scales, Langmuir 13 (1997) 2207–2214. [36] G. Olanya, J. Iruthayaraj, E. Poptoshev, R. Makuška, A. Vareikis, P.M. Claesson, Langmuir 24 (2008) 5341–5349. [37] F. Joabsson, K. Thuresson, B. Lindman, Langmuir 17 (2001) 1499–1505. [38] A. De˙dinaite˙, L. Bastardo, Langmuir 18 (2002) 9383–9392. [39] R.G. Horn, D.T. Smith, W. Haller, Chem. Phys. Lett. 162 (1989) 404–408. [40] P.M. Claesson, J.L. Parker, J.C. Fröberg, J. Disp. Sci. Technol. 15 (1994) 375–397. [41] K.D. Kwon, V. Vadillo-Rodriguez, B.E. Logan, J.D. Kubicki, Geochim. Cosmochim. Acta 70 (2006) 3808–3819. [42] J. Iruthayaraj, E. Poptoshev, A. Vareikis, R. Makuska, A. van der Wal, P.M. Claesson, Macromolecules 38 (2005) 6152–6160. [43] R.G. Cox, Int. J. Multiph. Flow 1 (1974) 343–371. [44] D.Y.C. Chan, R.G. Horn, J. Chem. Phys. 83 (1985) 5311–5324. [45] E. Thormann, D.R. Evans, V.S.J. Craig, Macromolecules 39 (2006) 6180–6185. [46] M.U.R. Kjellin, P.M. Claesson, R. Audebert, J. Colloid Interface Sci 190 (1997) 476–484. [47] T. Drobek, N.D. Spencer, Langmuir 24 (2008) 1484–1488. [48] G. Dunér, Disp. Sci. Technol. 31 (2010) 1285.
Contents lists available at SciVerse ScienceDirect
Journal of Colloid and Interface Science www.elsevier.com/locate/jcis
Surface and friction forces between grafted polysaccharide layers in the absence and presence of surfactant Cathy E. McNamee a,⇑, Shinpei Yamamoto b, Michael Kappl c, Hans-Jürgen Butt c, Ko Higashitani d, Andra De˙dinaite˙ e,f, Per M. Claesson e,f a
Shinshu University, Tokida 3-15-1, Ueda-shi, Nagano-ken 386-8567, Japan Institute for Integrated Cell-Material Sciences, Kyoto University, Yoshida-Ushinomiyacho, Sakyo-ku, Kyoto 606-8501, Japan Max Planck Institute for Polymer Research, Ackermannweg 10, 55128 Mainz, Germany d Department of Chemical Engineering, Kyoto University – Katsura, Nishikyo-ku, Kyoto 615-8510, Japan e Department of Chemistry, Surface and Corrosion Science, Royal Institute of Technology, Drottning Kristinas väg 51, SE-100 44 Stockholm, Sweden f Institute for Surface Chemistry, YKI, P.O. Box 5607, SE-114 86 Stockholm, Sweden b c
a r t i c l e
i n f o
Article history: Received 1 July 2011 Accepted 25 August 2011 Available online 1 September 2011 Keywords: Dextran Silane coupling agent Polymer layer Silica Friction Surface forces Adhesion Tethered polysaccharide AFM Ellipsometry
a b s t r a c t We analyzed the interaction between chemically grafted polysaccharide layers in aqueous solutions. To fabricate such layers, an end-terminated dextran silane coupling agent was synthesized and the polydextran was grafted to oxidized silicon wafers and to silica particles. This resulted in the formation of a 28 nm thick layer (in air) and a grafted amount of 40 mg/m2 as determined by ellipsometry. The physical properties of the grafted layer were investigated in aqueous solutions by atomic force microscope imaging and colloidal probe force measurements. Surface and friction forces were measured between one bare and one polydextran coated silica surface. A notable feature was a bridging attraction due to affinity between dextran and the silica surface. Surface interactions and friction forces were also investigated between two surfaces coated with grafted polydextran. Repulsive forces were predominant, but nevertheless a high friction force was observed. The repulsive forces were enhanced by addition of sodium dodecyl sulfate (SDS) that associates with the tethered polydextran layers. SDS also decreased the friction force. Our data suggests that energy dissipation due to shear-induced structural changes within the grafted layer is of prime importance for the high friction forces observed, in particular deformation of protrusions in the surface layer. Ó 2011 Elsevier Inc. All rights reserved.
1. Introduction Knowledge and understanding of physico-chemical and mechanical properties of surfaces are needed for designing materials for optimum performance, e.g. for uses as biomaterials. An implant surface should be non-toxic, display biocompatibility to neighboring cells, and possess pseudo-physiological mechanical properties [1]. For example, materials to be used in orthopaedic/ periodontal applications as hard tissue substitutes should be tough but flexible [2]. Depending on the intended use, other requirements also need to be fulfilled. Biomaterials to be used in tissue engineering applications, such as the scaffolds used in regenerative medicine to promote cell growth on implants, only serve as a temporary matrix and should therefore be biodegradable [1,6]. The coating of a material that is to be used in a joint area must be resistant enough that it will not be eroded by the combined action of shear and load [3,4]. It should also display enough lubricity to
⇑ Corresponding author. E-mail address: [email protected] (C.E. McNamee). 0021-9797/$ - see front matter Ó 2011 Elsevier Inc. All rights reserved. doi:10.1016/j.jcis.2011.08.065
allow the two surfaces to move against one another without a large friction [5,4], which would otherwise impair the relative motion of the constituents in the joint. Thus, in this case important material properties are hardness, elasticity, low wear and lubricity. Polysaccharides are becoming increasingly popular as biomaterials for use in the body. This is due to their low toxicity, low cost, and large-scale availability [1,6–8]. Physically adsorbed layers of polysaccharides have been shown to provide a low friction coefficient when two such modified surfaces are sheared against each other [9–11]. This suggests that polysaccharides can be used for producing films with favorable lubricating properties. However, physically adsorbed polymers most often have a limited load bearing capacity since they are relatively easily removed under the combined action of load and shear. This limitation may be overcome by chemically attaching the polysaccharide, which has the potential to achieve high lubrication ability and high load bearing capacity even though wear of the film can be a problem during prolonged sliding. A film consisting of chemically attached polysaccharides would allow the thickness and density of the film to be controlled by choosing the polysaccharide type, molecular weight, and reactant
352
C.E. McNamee et al. / Journal of Colloid and Interface Science 364 (2011) 351–358
Dextran-Lactone was obtained. Next, 3-aminopropyltriethoxysilane (0.2 g) was added to a solution of Dextran-Lactone (1.5 g) in DMSO (15 mL) and ethanol (7.5 mL), corresponding to concentrations of 0.9 mM and 0.075 mM for 3-aminopropyltriethoxysilane and Dextran-Lactone, respectively. The mixture was then stirred for 20 h at 60 °C. As the reaction solution was concentrated, a large amount of acetone was subsequently added. Finally, the white precipitate was collected and dried in vacuo to give the dextransilane-coupling agent (DEX-si). The structure of the compound is shown in Fig. 1. The silica wafers, cleaned by solvents washing (acetone, ethanol, and then water) and then by plasma treatment (PDC002, Harrick Plama, Ithaka, NY)), and silica particles were rendered fully hydroxylated [16] by placing them in a glass container and adding 3 mL of a solution containing 17.5 mL methanol (high purity, Wako, Japan) and 1.39 g of 28% ammonium hydroxide solution (Wako, Japan) for 2 h, after which a freshly prepared solution of 0.1 g DEX-si in 1 mL water was added drop-wise to the wafer/particle containing solution. This solution was then left for 16 h, after which the substrate/particles were washed with copious amounts of water and ethanol to remove non-chemically adsorbed material. The presence of the silane coupling agent facilitates chemical grafting to the surface and also results in chemical binding between the Dex-si molecules. The end result is a firmly attached and partly cross-linked dextran layer on the surface. The grafted layer on flat silica surfaces was characterized by ellipsometry (M-2000U, J.A. Woolam, USA) measurements in air. The colloid probes for the force and friction measurements were prepared by evaporating the solvent from a small volume of the particle dispersion, and then attaching a single particle to a cantilever by using an XYZ micromanipulator and an epoxy resin (rapid araldite, Vantico, Showa polymer company, Tokyo, Japan). Tipless V-shaped gold-plated Si3N4 cantilevers (normal spring constant 0.12 N m1 and 0.58 N m1, NP-S, Veeco NanoProbe™ Tips, USA) and rectangular cantilevers (tip-less, non-Al coated, nominal spring constant 0.08 N m1, CSC12/tipless/noAl, MikroMasch, USA) were used for the force and friction measurements, respectively.
concentration. The lubricating ability of a film is affected by the presence of other species in the surrounding fluid. For instance, the addition of a surfactant, sodium dodecyl sulfate, between two surfaces coated with a mucin layer has been shown to decrease the friction significantly [9]. Therefore, naturally occurring compounds, such as phospholipids and proteins, may affect the lubrication of a grafted polysaccharide layer and enhance the lubrication ability of the manufactured biomaterial. For other applications, this may also be achieved by adding man-made surfactants. We have synthesized an end-terminated dextran silane coupling agent (DEX-si) and chemically attached it to silica surfaces. The physical properties of DEX-si modified silica surfaces were studied with an atomic force microscope (AFM) in order to obtain information on layer morphology, and surface and friction forces. The effect of introducing a surfactant into the surrounding aqueous solution on the physico-chemical properties of the surface bound polymer layer was studied by using different concentrations of sodium dodecyl sulfate (SDS). 2. Experimental 2.1. Materials Dextran (DEX, molecular weight (Mw = 20 kg/mol), GR grade, Nacalai Tesque, Japan) was used without further purification. The polydispersity (Mw/Mn) of the dextran sample was determined to be 1.42 by employing poly(ethyleneglycol)-calibrated GPC (Shodex GPC-101, SHOWA DENKO, Japan) using water as the eluent. The radius of gyration (Rg) of a typical polymer in our DEX sample in water was estimated to be about 5 nm, using literature data on Rg versus Mw [12]. The contour length of a typical dextran chain can be estimated to be 32 nm from the Mw and the length of one maltose unit, 0.52 ± 0.03 nm [13]. Potassium hydroxide (GR grade, Nacalai Tesque, Japan), ethanol (GR grade, Nacalai Tesque, Japan), dimethylsulfoxide (DMSO, GR grade, Nacalai Tesque, Japan), 3-aminopropyltriethoxysilane (GR grade, Nacalai Tesque, Japan), iodine (99.9% purity, Wako Pure Chemical, Japan), and sodium dodecyl sulfate (SDS, 95% purity, Wako, Japan) were used without further purification. Naturally oxidized silicon wafers (Silicon Quest INT., USA) were used as the substrates. The water used was distilled and de-ionized to give a resistivity of 18.2 MX cm and a total organic content of <5 ppm. The non-modified and DEX-si modified silica particles had a radius of 3.4 lm (Bangs Laboratory, USA) and 3 lm (micromer, micromod Partikeltechnologie GmbH, Rostock, Germany), respectively. Dextran having a lactonized end (Dextran-Lactone) was prepared according to reported methods (Fig. 1) [14,15]. Briefly, 0.1 M iodine in water and 0.8 M potassium hydroxide in water were added slowly to a 0.01 M dextran aqueous solution. The reaction mixture was stirred at ambient temperature for 24 h. It was then placed in a cellulose dialyzer tubing (Mw cut-off, 14 000 g/mol) and purified by dialysis. The solution was subsequently passed through a column packed with a cation-exchange resin (Amberlite IR-120), lyophilized, and a white powder of
OH OH
O
I2, KOH
OH O OH
OH
93-123
Amberlite IR 120
O OH
OH OH
OH
2.2.1. Atomic force microscopy The substrates were imaged in aqueous solutions with an AFM (Digital Instruments NanoScope III Multimode, Santa Barbara, USA) in tapping mode by using a tapping mode fluid cell and Si3N4 cantilevers (NP-S, Veeco NanoProbe™ Tips, USA) with reflective gold coating. The V-shaped cantilevers were 196 lm long and 23 lm wide, and had a nominal spring constant of 0.06 m1, a resonant frequency of 31 kHz (in air), and square pyramidal tips with an end-tip radius of curvature in the range 10–40 nm. The imaging scan rate was 2 Hz, and the scan size was 1 1 lm2. Both the resolution of data points per line and the number of lines were 512. All images are unmodified (i.e., no filtering was applied), except for a 1st order flattening along the scan lines. The images shown for each condition are representative of the images made from at OH
OH
OH
2.2. Methods
O OHO OH OH
-H2O 93-123
OH OHO
OH OH
OH O OHO
+H 2O OH
OH OH
O
OH DMSO/EtOH OH
OHO
93-123
O OH O
NH2
Si(OEt)3
OH OH
93-123
OH H N OH O
DEX -si dextran Fig. 1. Schematic diagram showing the synthesis of dextran-triethoxy silane (DEX-si).
Si(O Et) 3
C.E. McNamee et al. / Journal of Colloid and Interface Science 364 (2011) 351–358
least two different samples and at a minimum of two different places on each sample. The topography and physical features were the same, regardless of the sample and imaging position. The images were reproducible upon multiple scans, and did not change if they were scanned from top to bottom or from bottom to top. Imaging over a wider scale of 4 4 lm2 gave smaller but similar topographical features as that obtained in the 1 1 lm2 scale image, indicating the absence of tip artefacts. Surface forces were measured between a substrate and a colloid probe in aqueous solution as a function of separation relative to hard wall contact by using the AFM and a contact mode fluid cell. A syringe was used to inject solutions into the fluid cell. The force measurements were carried out by bringing the cantilever into contact with the substrate, during which time the change in the deflection of the cantilever (Dx) was measured as a function of the piezo displacement with a split photodiode. The detector signal versus piezo position curves were converted to force versus distance curves using the conventional procedure. First, a linear fit of the zero force line (i.e. the signal at large cantilever-substrate separations where no surface forces were acting) was subtracted. The zero force position was defined at large cantilever-substrate separations, where no surface forces were acting on the cantilever. The conversion factor between the detector signal in volts and the cantilever deflection in nm was then determined from the slope of the linear compliance region, which occurred when the probe and the substrate were moving in parallel, i.e. when they were in hard wall contact. The force curves were then obtained by subtracting the cantilever deflection (Dx) from the piezo position and using Hooke’s law, F = kDx. Here, F and k are the force and the nominal spring constant of the cantilever, respectively. The lateral (friction) force was measured using a Digital Instruments NanoScope III multimode microscope in the ‘‘friction mode’’ and a liquid fluid cell. Briefly, the colloid probe was pressed against the substrate in solution at a constant applied load, while the substrate slid horizontally underneath the cantilever at a speed of 3.60 lm/s. The magnitude of the lateral frictional force (FF) was determined using
FF ¼
V L SL K L 2H
ð1Þ
where VL, SL, KL, and H are the difference in the lateral force detector signal in one complete scan, the lateral detector sensitivity, the lateral spring constant, and the distance from the bottom of the sphere to the mid-point of the cantilever, respectively. The value of SL was determined to be 3.1 104 rad/V from the method of Meurk and others [17]. The value of KL was determined from
KL ¼
2K N L2 3ð1 þ mÞ
ð2Þ
Here, KN and L are the normal spring constant and the length of the cantilever, respectively. The symbol m is the Poisson’s ratio. A typical value [18] of 0.27 was used. The values of FF for each system and condition were averaged from four separate experiments using different substrates and probes. The error in each FF measurement is estimated to be 0.6 nN. Each friction measurement was made at a different position on the substrate. The expressions used to relate FF to the applied normal load (FN) depend on whether an intermolecular attraction between the surfaces or the external load dominates [19]. In the non-adhesive case the following equation was used:
F F ¼ ðF N F 0 Þl; when F n > F 0
ð3Þ
Here, F0 is the load at which Amontons’ law starts to apply, and l is the friction coefficient. The effective coefficient of friction (leff) is defined as
leff ¼
FF FN
353
ð4Þ
and when F0 = 0, l = leff. This load-dependent model is similar to the one used by Ecke and others in an earlier study [20]. That model also gave a roughly linear increase with load, but showed some deviation for low loads. The following relation can be used when FF is adhesion controlled,
F F ¼ Sc A þ F N l
ð5Þ
Here, Sc is the critical shear stress, A is the contact area, and FN is the normal load. The JKR theory [21] can be used to calculate A,
qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi2=3 R A ¼ p F N 6pRc þ 12pRcF N þ ð6pRcÞ2 E
ð6Þ
Here, R is the particle radius, c is the interfacial tension that can be calculated using the adhesion force obtained from the normal force curve,
c¼
F adh 3pR
ð7Þ
E⁄ is the effective elasticity of the system. We estimate it from the Young’s modulus (E) and Poisson ratio (m) of silica using
E ¼
1 2ð1 m2silica Þ Esilica
ð8Þ
with Esilica = 72 GPa and msilica = 0.17 [22]. 3. Analysis The total interaction free energy (Etot) per unit area can have several contributions, which can be considered as being approximately additive. In this work we encounter three contributions due to (i) repulsive double layer interactions (EDL), (ii) attractive van der Waals interactions (EvdW), and (iii) polymer induced interactions that include steric repulsion and attractive bridging interactions (Epol). Thus,
Etot ¼ EDL þ Ev dw þ Epol
ð9Þ
The repulsive double layer interaction at a distance x was calculated using constant surface potential boundary conditions, assuming equal surface potentials, or using constant surface charge densities boundary conditions, assuming equal charge densities on the two surfaces. We used the approximate equations, valid for small midpoint potentials [23]. The van der Waals interaction was described using a non-retarded Hamaker constant. The force (F) between the spherical particle and the substrate can be related to Etot using the Derjaguin approximation (Eq. (10)), provided the particle radius (R) is much larger than the separation distance between the particle and the substrate. This condition is fulfilled.
F ¼ 2pEtot R
ð10Þ
4. Results and discussion 4.1. Layer characteristics and surface morphology Ellipsometry measurements in air showed that the DEX-si film thickness, t, was 27.9 ± 2.8 nm, evaluated using a typical refractive index of 1.54 for polysaccharides. This rather large thickness indicates that the trifunctional silane not only binds DEX-si to the surface, but also allows DEX-si polymers to chemically bind to each
354
C.E. McNamee et al. / Journal of Colloid and Interface Science 364 (2011) 351–358
Fig. 2. 1 1 lm2 tapping mode image of an unmodified SiO2 wafer in 30 mM NaNO3. (A) Height image. (B) Phase image.
other. The corresponding mass of the grafted layer, C, was calculated from the density, q, of the polymer (q = 1.43 g/cm3 for dextran [24]) and the layer thickness as:
ð11Þ
It was found that C = 39.9 ± 4.0 mg/m2, which is of the order of a factor of 10 larger than typically found for physisorbed polymer layers. The surface morphology of bare silica wafers (Fig. 2) and silica wafers modified with DEX-si (Fig. 3A and B) were characterized in 30 mM NaNO3 using tapping mode AFM. The images of the bare silica surface are featureless and demonstrate the flat nature of the substrate. In contrast, surface features are seen on the DEX-si modified surfaces, and most clearly in the phase images. Surface features with widths 25 nm and heights between 10 and 15 nm were found to be characteristic of the layer (see height profile of Fig. 4A). From the images, we conclude that the surfaces coverage is non-homogeneous. We note that the dry thickness of the grafted DEX-si layer as measured by ellipsometry was about 28 nm. The polymer layer swells when placed in water, a good solvent for dextran, and the total polymer thickness should be larger than the dry thickness. Thus, the relatively small height variations observed by AFM in aqueous solutions demonstrates that the AFM tip does not reach the bare silica surface. We conclude that the silica surfaces are fully covered by DEX-si, but that the thickness of the layer is not homogeneous. The height and phase images of silica surfaces carrying grafted DEX-si changed when SDS was added to the NaNO3 solution (Fig. 3C–H). The lateral area of the protrusions increased as the concentration of SDS increased from below the critical micelle concentration (cmc) (1 mM, Fig. 3C and D), to a value just above the cmc (5 mM, Fig. 3E and F), and to a value significantly above the cmc (10 mM, Fig. 3G and H); the cmc of SDS in the presence of 30 mM NaNO3 is reported to be 3.3 mM [25]. The height profiles (Fig. 4B–D) corresponding to the lines in the height images demonstrate that the surface features become smeared out in the presence of SDS. A similar effect of SDS on the morphology of physisorbed mucin–chitosan complexes has previously been noted [26]. These results indicate that SDS adsorbs to the DEX-si coated substrate at concentrations below its cmc and whereby expand the polymer chains. We suggest that the adsorption is driven by the hydrophobic tails of the SDS, which by association with the surface bound dextran chains can reduce its exposure to bulk water. Typically, anionic surfactants bind cooperatively to non-ionic polymers at about 0.5 cmc. This has for example been shown for linear poly(ethylene oxide) [27–29], bottle-brush polymers with poly(ethylene oxide) side chains [30,31], poly(methoxyhexa(ethylene glycol) methacrylates), and non-ionic triblock copolymers [32,33]. 4.2. Surface forces 4.2.1. Effect of SDS on interactions between bare silica surfaces Forces between a silica particle and a silica substrate across 30 mM NaNO3 in the absence and presence of SDS were found to
Fig. 3. 1 1 lm2 tapping mode images of DEX-si modified SiO2 wafers in solutions. Height images are provided in the left column and phase images in the right column. The black lines show the position of the height profiles provided in Fig. 4. (A and B) 30 mM NaNO3; (C and D) 30 mM NaNO3 + 1 mM SDS; (E and F) 30 mM NaNO3 + 5 mM SDS; (G and H): 30 mM NaNO3 + 10 mM SDS. The z-scale in all the height and phase images covers the range 0–30 nm and 0–5°, respectively.
(A)
10 0
Height (nm)
C ¼ tq
10
(B)
0
(C)
10 0
(D)
10 0 0
100
200
300
400
500
x-profile (nm) Fig. 4. Height profiles of the lines shown in the 1 1 lm2 tapping mode height images of DEX-si modified SiO2 wafer in solutions. (A) 30 mM NaNO3; (B) 30 mM NaNO3 + 1 mM SDS; (C) 30 mM NaNO3 + 5 mM SDS; (D) 30 mM NaNO3 + 10 mM SDS.
be repulsive, having a roughly exponential decay at large separations (Fig. 5). The addition of SDS to the solution did not affect the measured forces significantly. Comparison of the theoretical force curve calculated within the DLVO framework and using a Hamaker constant for the SiO2–water–SiO2 system of 0.46 1020 J
C.E. McNamee et al. / Journal of Colloid and Interface Science 364 (2011) 351–358
[34], with the experimental force data showed that the magnitude of the surface potential and Debye length were 15 mV and 1.8 nm, respectively. For comparison, the expected Debye-length is 1.8 nm and 1.5 nm for a solution of 30 mM NaNO3 in the absence and presence of 10 mM SDS. The surface potential of silica in 30 mM NaNO3 (15 mV) corresponds to a surface charge density of 6.0 mC/m2. The corresponding values reported for silica in 1 mM NaNO3 are 42 mV and 3.4 mC/m2, respectively [35] Thus, as expected for a charge regulating surface the surface potential decreases and the surface charge density increases as the inert salt concentration is increased [36]. The fact that the force curves were not significantly affected by SDS indicates that this surfactant does not adsorb to any significant degree to the bare silica surface, as also concluded from adsorption studies [37,38]. DLVO theory predicts an attractive net interaction at short separations. However, this was not observed in our experiments, in agreement with numerous other reports. The reason for this is that the silica surface is highly hydrated and a repulsion due to the dehydration of the surface layer overcomes the van der Waals force at short separations [39,40]. The measured force curves were independent of the approaching/retracting speed in the range investigated, 300–14 000 nm/s, and no hysteresis was detected between the approach and retraction curves. 4.2.2. Forces between one bare silica surface and one DEX-si modified silica surface in the absence and presence of SDS Force curves measured between a DEX-si coated silica plate and bare silica probe on approach demonstrate an attraction commencing at approximately 40 nm in 30 mM NaNO3 in the absence and presence of 1 mM SDS (Fig. 6). We interpret the attraction to be caused by bridging of polymer chains to the approaching silica particle, and the large range of this force as being a consequence of formation of DEX-si polymers. The DEX-si molecule has several – OH and –O– sites that may interact favorably with the silanol groups on the silica surface. Dextran has indeed previously been reported to bind to silica surfaces in water. The normalized energy as determined from the pull-off force (Fad/2pR) for dextran (Mw = 10 kDa) from a silica colloidal tip in water was reported to be 0.026 1018 J/nm2, i.e, 26 mN/m [41]. This strong affinity was suggested to be due to hydrogen-bond formation between dextran and silica, and the presence of such bonds was confirmed by infrared spectroscopy [41]. We note, however, that to state that adsorption is driven by dextran-silanol hydrogen bonding is a sim-
F/R (mN/m)
1
0.1
0.01
0
2
4
6
Separation (nm) Fig. 5. Examples of force curves measured between a SiO2 particle and a SiO2 wafer in solutions of 30 mM NaNO3 in the absence and presence of SDS. The measurements were carried out at a frequency of 1 Hz (speed = 1993 nm/s). The experimental data are given by h, j: 30 mM NaNO3; s, d: 30 mM NaNO3 + 1 mM SDS; D, : 30 mM NaNO3 + 5 mM SDS; e, : 30 mM NaNO3 + 10 mM SDS. The symbols represent the forces measured on approach and the solid symbols show the forces measured on separation. The theoretical force curves were calculated using constant potential (- - -) and constant charge (–) boundary conditions. The parameters used for the calculations were: w0 = 15 mV, 1/j = 1.8 nm, A = 0.46 1020 J, R = 3.42 lm.
355
plification as discussed for the case of poly(ethylene oxide)–silanol interactions [42]. The maximum normalized adhesion force in the 30 mM NaNO3 solution observed in the retraction force curves in our experiment was found to be 6 ± 0.3 mN/m. The strength of the attraction on approach was observed to decrease with increasing driving velocity (Fig. 6). This could be due to either hydrodynamic effects or chain relaxation effects. A precise calculation of the hydrodynamic force is complicated since the viscosity of the liquid in the polymer layer is higher than in bulk water. However, a simple estimation can be made by assuming the same viscosity as bulk water and the no-slip boundary condition. In this case, the normalized hydrodynamic force of a sphere approaching a planar surface [43,44] is:
F 6pgmR ¼ R x
ð12Þ
Here, g is the viscosity of the liquid (103 Pa s for water) and v is the approaching velocity. At a velocity of v = 10 lm/s and a radius of 3 lm at a distance x = 10 nm, the normalized hydrodynamic force is 0.06 mN/m. This is much lower than the forces observed. Thus, as any effect of hydrodynamic forces is negligible, we attribute the reduction of the attractive force to the finite time required for bridge formation. Taking the time the silica sphere requires to approach the dextran-coated surface to be the range of the attraction divided by the approaching speed, we conclude that more than 0.1 s is needed for formation of all bridges, and the relatively long relaxation time is suggested to be a consequence the presence of DEX-si– DEX-si cross-links within the layer. The hypothesis of bridge formation in the DEX-si–silica system is supported by the force curves measured on retraction (Fig. 6B, D and F). The retraction force curves recorded in 30 mM NaNO3 and in 30 mM NaNO3 with 1 mM SDS showed a strong attractive force with a range of 40–80 nm. Several smaller attractive maxima were observed at large distances, before the two surfaces were completely detached. Such multiple adhesion maxima are due to individual polymer chains detaching from the retracting silica surface [45]. Increasing the concentration of SDS in the solution to 5 mM and 10 mM resulted in removal of the attractive force observed on approach (Fig. 6A, C and E). Instead, a repulsive force was noted at separations below 20–30 nm. This distance is much longer than the Debye length associated with a double layer force in a 30 mM NaNO3 solution in the presence of 5 and 10 mM SDS, which is calculated to be 1.6 nm and 1.5 nm, respectively. Clearly, binding of SDS to DEX-si impairs the possibility of bridge formation and instead a repulsive force of electrosteric nature is encountered. However, on separation, a long-range attraction, extending to a separation of 50–70 nm, was observed both in the absence and presence of SDS, and the adhesion was somewhat larger in the presence of SDS. This suggests that when the two surfaces are in close contact SDS is partly desorbed from the layer, and when this occurs bridges can again be formed between DEX-si and the silica particle. The slight increased adhesion upon SDS addition is attributed to hydrophobic interactions between the tails of the SDS molecules that remain in the layer also under compression. Removal of SDS by compressing polymer-surfactant layers has also been observed in other systems [46]. The magnitude of the adhesion decreased marginally with increasing measurement speed, from 6 ± 0.3 mN/m at 0.3 Hz to 5±0.3 mN/m at 14 Hz. 4.2.3. Forces between two DEX-si modified silica surface in the absence and presence of SDS In most cases the forces measured between a DEX-si silica substrate and a DEX-si silica particle in 30 mM NaNO3 in the absence and presence of SDS were found to be repulsive both on approach and separation (Fig. 7). However, in a few cases a small attraction was noted on separation. Some hysteresis was observed in the
356
C.E. McNamee et al. / Journal of Colloid and Interface Science 364 (2011) 351–358
4
(A) 0.3 Hz
(C) 1 Hz
(E) 14 Hz
(B) 0.3 Hz
(D) 1 Hz
(F) 14 Hz
F/R (mN/m)
2 0 -2 4 2 0 -2 -4 -6 0
40
80
120
0
40
80
120
0
40
80
120
Separation (nm) Fig. 6. Examples of force curves measured between a SiO2 particle (R = 3.42 lm) and a DEX-si wafer in solutions of 30 mM NaNO3 with and without SDS. The measurements were performed at frequencies of 0.3 Hz (speed = 296 nm/s, A and B), 1 Hz (speed = 1993 nm/s, C and D), and 14 Hz (speed = 13 951 nm/s, E and F). h: 30 mM NaNO3; s: 30 mM NaNO3 + 1 mM SDS; D: 3 30 mM NaNO3 + 5 mM SDS; e: 30 mM NaNO3 + 10 mM SDS. The upper row displays forces measured on approach and the lower row illustrates forces measured on separation.
FN (nN) 20
0
10
20
30
(A) SiO2-SiO2
15 10 5 0 20
FF (nN)
force curves, where the forces measured on separation were lower than those encountered on approach. The hysteresis is assigned to slow chain relaxation processes in the layer. The somewhat uneven appearance of the force curve is attributed to the non-homogeneous nature of the dextran-coated surfaces (see Fig. 3). The distance dependence of the force is inconsistent with that of a double-layer force at the given ionic strength. Thus, it is concluded that it has a steric or electrosteric origin. The range and magnitude of the repulsion increased with increasing SDS concentration due to some binding of SDS to the dextran coated surfaces, which causes the polymer chains to become stretched for electrostatic reasons. We note that the range of the force is significantly less than twice the thickness of the layer as determined by ellipsometry. This is a consequence of the fact that the zero separation in the force curves determined with the AFM is defined as the ‘‘hard wall’’ contact between the compressed polymer layers and not as the distance from the bare surfaces.
(B) SiO2 - DEX-si
15 10 5 0
4.3. Friction forces
40
The friction force (FF) as a function of load (FN, FN/R) for the SiO2–SiO2, SiO2–DEX-si, and DEX-si–DEX-si systems in 30 mM NaNO3 in the absence and presence of SDS are shown in Fig. 8.
20
F/R (mN/m)
0
(C) DEX-si - DEX-si
0
2
4
6
8
10
FN/R (mN/m) 1
0.1
0
5
10
15
20
Separation (nm) Fig. 7. Typical force curves measured between a DEX-si particle (R = 3.0 lm) and a DEX-si wafer in solutions of 30 mM NaNO3 in the absence and presence of SDS. The measurements were carried out at a frequency of 1 Hz (speed = 1993 nm/s). The experimental data are given by h: 30 mM NaNO3; s: 30 mM NaNO3 + 1 mM SDS; D: 30 mM NaNO3 + 5 mM SDS; and e: 30 mM NaNO3 + 10 mM SDS.
Fig. 8. Friction force versus load measured between (A) two SiO2 surfaces, (B) SiO2 and DEX-si, and (C) two DEX-si surfaces in solutions of 30 mM NaNO3 in the absence and presence of SDS. No SDS: h; 1 mM SDS: ; 5 mM SDS: ; 10 mM SDS: e. The load controlled fits to the experimental data are given by –: no SDS; : 1 mM SDS; : 5 mM SDS; : 10 mM SDS. The adhesion controlled fits are shown as and , obtained using l = 0, Esilica = 72 Gpa, msilica = 0.17, Fad/R = 0.10 and R = 3 lm. The upper and lower x-axes show the FN and FN/R data, respectively. (For color interpretation in this figure legend the reader is referred to see the web version of this article.)
The plotted FF data represents the average FF values obtained from at least four independent experiments, where different substrates, solutions and probes were used each time. The average variation in the FF data set for the SiO2–SiO2, SiO2–DEX-si, and DEX-si–DEX-si
357
C.E. McNamee et al. / Journal of Colloid and Interface Science 364 (2011) 351–358
Table 1 Friction parameters obtained from the load controlled fit of the FF versus FL/R data of Fig. 8. F0 and l are the load at which Amontons’ law starts to apply and the coefficient of friction, respectively. SDS conc. (mM)
0 1 5 10 a
SiO2–SiO2
SiO2-DEX-si
DEX-si–DEX-si
F0/R (mN/m)
l
F0/R (mN/m)
l
F0/R (mN/m)
l
0 0 0 0
0.42 ± 0.00 0.14 ± 0.01 0.05 ± 0.00 0.02 ± 0.01
0.19 ± 0.29 0 0 0
0.42 ± 0.03 0.28 ± 0.02 0.18 ± 0.02 0.24 ± 0.02
–a –a 0 0
–a –a 0.19 ± 0.01 0.20 ± 0.02
The friction force was always found to be high as indicated in Fig. 8, but the scatter in the data was large.
Table 2 Friction parameters obtained from the Adhesion controlled fit of the FF versus FL/R data of Fig. 8. Sc and l are the critical shear stress and the coefficient of friction, respectively. SDS conc. (mM)
0 1 5 10
SiO2–DEX-si Sc (Mpa)
l
12 8 8 8
0 0 0 0
systems were 0.8, 4.4, and 2.6 nN, and in each experimental series the same qualitative changes in the FF values with SDS concentration were observed. No hysteresis was observed when the load was increased and then subsequently decreased, indicating the absence of significant wear of the surfaces. The FF versus FL/R data was fitted using Eq. (3) (Fig. 8) for all cases, and also by Eq. (5) (Fig. 8) for SiO2–DEX-si case, as adhesion was observed for this system. The F0/R and l values obtained from the load-controlled model are shown in Table 1, whereas the corresponding values obtained from the adhesion-controlled model are shown in Table 2. In these latter calculations, the value of Fad/R determined from the force curves was used. 4.3.1. The silica–silica case As expected, the value of FF increased with load for the SiO2 substrate–SiO2 particle case in 30 mM NaNO3 in the absence and presence of SDS. The results are consistent with the load-controlled model. We note that at low forces (F/R < 0.7 mN/m) the doublelayer force carries the load. In this case a highly fluid layer separates the surfaces and the expected friction force is below the detection limit of our instrument [19]. The low load bearing capacity of the double-layer force means that the off-set value, F0 in Eq. (3), is expected to be small. This was also found when fitting the load-controlled model to the data (Table 1 reports F0/R). The interaction between the silica surfaces in 30 mM NaNO3 were repulsive on approach and retraction, with a magnitude, range and shape that did not change with the addition of SDS. Nevertheless, the frictional forces decreased when SDS was added. This effect, which was observed in all experiments, is surprising considering that SDS does not adsorb to silica. It may be due to some dodecanol adsorption, the hydrolysis product of SDS, but this should be regarded as speculation and the observation is not yet understood. 4.3.2. The silica–DEX-si case The friction forces between the DEX-si substrate and the bare SiO2 particle in the absence of SDS increased linearly with the applied loads for low to intermediate loads (FL/R < 6 mN/m). Thus, the load-controlled model, despite the adhesion, represents the data satisfactory. The use of the model for adhesion-controlled friction did not provide a better fit to the data (see Fig. 8). We note that
the friction increased considerably for FL/R greater than 6 mN/m. This indicates the appearance of a new friction mechanism when the water content of the layer is reduced below a critical value. A similar result with a low friction at low and medium loads and high friction at high loads has been reported for PLL–PEG brushes [47]. In that case, the low friction was explained as being a consequence of the osmotic pressure resulting from the compression of the polymer brushes, which lead to strong repulsive forces that limited the interpenetration of opposing polymer layers. At higher loads the water between the polymer brushes were partly squeezed out, resulting in a reduced fluidity of the end segments of the polymer chains, explaining the significantly higher friction forces at higher loads. We suggest the same mechanism in our case. The friction in presence of SDS, and in absence of SDS at high loads, for the SiO2–DEX-si system was higher than that for the SiO2–SiO2 case. This is attributed primarily to energy dissipative shear-induced changes within the polymer layer, e.g. deformation of the protrusions observed in the AFM images. The importance of energy dissipative mechanisms within a single polymer layer for the magnitude of the friction force has also been noted for a grafted polyacrylic acid layer sliding against a silica surface [48], but in that case the energy dissipative mechanism was suggested to be the breakage and formation of physical intra-layer bonds. The adhesion present between the Dex-si surface and the bare silica surface should also contribute to the friction, but this mechanism is not dominant as seen by the poor fit of the adhesioncontrolled model to the data in Fig. 8b. Addition of SDS caused FF to decrease in magnitude, particularly at high loads. This is consistent with the AFM images, Fig. 3, that show how the protrusions are smeared out due to the incorporation of SDS in the DEX-si layer, and thus shearing in presence of SDS is expected to lead to less deformation of the protrusions and less energy dissipation.
4.3.3. The DEX-si–DEX-si case The friction between the DEX-si substrate–the DEX-si particle in the absence of SDS and after the addition of 1 mM SDS (Fig. 8c) was found to be high. The high value of the friction force in these cases arises from the energy dissipative shear-induced structural changes occurring within each layer, amplified by the interlocking of the protrusions present on the two opposing layers and the interpenetration of the DEX-si chains upon the application of a load and a sliding motion. In this situation neither the load-controlled nor the adhesion-controlled model provided a good fit to the data. We note that Kampf et al. found a significant increase in friction between layers of the cationic polysaccharide chitosan when these layers were cross-linked [11]. The large grafted amount in our case can only be achieved due to formation of DEX-si, polymers, and the trifunctional silane group allows for cross-linking reactions. Thus, it appears that the presence of chemical cross-links in the layer provides for additional energy dissipative mechanisms associated with the shear-induced strain on these cross-links and the
358
C.E. McNamee et al. / Journal of Colloid and Interface Science 364 (2011) 351–358
subsequent relaxation to the equilibrium state. An increase in the SDS concentration to 5 mM and 10 mM resulted in a decrease in the value of FF, giving values in the order of those obtained for the SiO2–DEX-si case. Hence, the presence of SDS on the DEX-si surface decreases the friction. This is suggested to be due to smearing out of the surface protrusions and reduced chain interpenetration, both being facilitated by the electrostatic repulsion due to the presence of SDS. The latter mechanism has also been suggested to explain the lubricating effect of SDS for mucin coated surfaces [9]. 5. Conclusions Silica surfaces have been modified with a polydextrane-silane, DEX-si. Ellipsometry measurements demonstrate a high amount of surface bound DEX-si, about 40 mg/m2, which demonstrate the formation of DEX-si polymers. The trifunctional silane coupling agent allows for cross-linking reactions within the layer. AFM imaging of the resulting surfaces provides evidence for the formation of a non-homogeneous surface layer containing protrusions with heights up to about 10 nm. Measurements of the forces acting between one DEX-si modified surface and a bare silica surface demonstrates the presence of a long-range attractive force and a strong adhesion. This force is attributed to bridging. The attraction measured on approach was found to decrease with increasing approach rate. From this observation the relaxation time corresponding to bridge formation was estimated to be around 0.1 s. This relatively long time is suggested to a consequence of the presence of crosslinks in the layers formed. Addition of SDS decreased the attraction observed on approach and lead to repulsion at concentrations above the cmc. This effect is due to binding of SDS to the surface attached dextran layer, which counteracts bridge formation. The interactions between two DEX-si coated silica surfaces are repulsive due to steric forces. Addition of SDS increases the range of the repulsion somewhat due to incorporation of anionic surfactants in the layer. The friction force for the SiO2ADEX-si system in the absence of SDS was found to be relatively high, which is explained by shearinduced energy-dissipative structural changes within the DEX-si layer, e.g. deformation of the protrusions observed by AFM imaging. SDS addition reduced the friction force, and this effect was attributed to smoothening of the surface protrusions due to the incorporation of SDS in the DEX-si layer. The friction force between two DEX-si coated surfaces was found to be high in the absence and presence of small (1 mM) concentrations of SDS. This is also rationalized by energy-dissipative structural changes within each layer due to the action of shear and load, but the friction is amplified by interlocking of the protrusions and chain interpenetration. Addition of SDS reduces the friction force also in this case due to smoothening of the surface protrusions. The presence of SDS also counteracts chain interpenetration for electrostatic reasons. Acknowledgments This research was partly supported by the Core-to-Core Program promoted by Japan Society for the Promotion of Science (Project No. 18004). This study was also performed through the Program for Dissemination of Tenure-Track System funded by the Ministry of Education and Science, Japan. P.C. and A.D. acknowledge financial support from the SSF program ‘‘Microstructure, Corrosion and Friction Control’’.
References [1] F. Couet, N. Rajan, D. Mantovani, Macromol. Biosci. 7 (2007) 701–718. [2] E. Khor, L.Y. Lim, Biomaterials 24 (2003) 2339–2349. [3] S. Affatato, M. Spinelli, M. Zavalloni, C. Mazzega-Fabbro, M. Viceconti, Med. Eng. Phys. 30 (2008) 1305–1317. [4] J. Katta, Z. Jin, E. Ingham, J. Fisher, Med. Eng. Phys. 30 (2008) 1349–1363. [5] R. Crockett, Tribol. Lett. 35 (2009) 77–84. [6] M. Rinaudo, Polym. Int. 57 (2008) 397–430. [7] M. Rinaudo, Prog. Polym. Sci. 31 (2006) 603–632. [8] V.R. Sinha, A.K. Singla, S. Wadhawan, R. Kaushik, R. Kumria, K. Bansal, S. Dhawan, Int. J. Pharm. 274 (2004) 1–33. [9] T. Pettersson, A. De˙dinaite˙, J. Colloid Interface Sci. 324 (2008) 246–256. [10] D. Gourdon, Q. Lin, E. Oeoudjev, Y. Golan, J. Arad, J. Israelachvili, Langmuir 224 (2008) 1534–1540. [11] N. Kampf, U. Raviv, J. Klein, Macromolecules 37 (2004) 1134–1142. [12] C.E. Ioan, T. Aberle, W. Burchard, Macromolecules 33 (2000) 5730–5739. [13] M.S. Motawia, I. Damager, C.E. Olsen, B.L. Møller, S.B. Engelsen, S. Hansen, L.H. Øgendal, R. Bauer, Biomacromolecules 6 (2005) 143–151. [14] K. Hashimoto, S. Imanishi, M. Okada, H. Sumitomo, J. Polym. Sci., Pol. Chem. 29 (1991) 1271. [15] T. Zhang, R.E. Marchant, Macromolecules 27 (1994) 7302. [16] K. Ohno, T. Morinaga, K. Koh, Y. Tsujii, T. Fukuda, Macromolecules 38 (2005) 2137–2142. [17] A. Meurk, I. Larson, L. Bergström, Mater. Res. Soc. Symp. Proc. 552 (1998) 427. [18] P. Hess, Appl. Surf. Sci. 106 (1996) 433. [19] A. De˙dinaite˙, E. Thormann, G. Olanya, P.M. Claesson, B. Nyström, K. Zhu, Soft Matter 6 (2010) 2489–2498. [20] S. Ecke, H.-J. Butt, J. Colloid Interface Sci. 244 (2001) 432–435. [21] K.L. Johnson, K. Kendall, A.D. Roberts, Proc. Roy. Soc. London, Ser A 1324 (1971) 301. [22] M.J. Bamber, K.E. Cooke, A.B. Mann, B. Derby, Thin Solid Films 398–399 (2001) 299–305. [23] J. Israelachvili, Intermolecular and Surface Forces, third ed., Academic Press, London, 2010. [24] E.C. Gil, A.I. Colarte, B. Bataille, I. Caraballo, A.A.P.S. Pharm, Sci. Technol. 8 (2007) 281–288. [25] N. Lundin, L. Macakova, A. De˙dinaite˙, P. Claesson, Langmuir 24 (2008) 3814– 3827. [26] A. De˙dinaite˙, M. Lundin, L. Macakova, T. Auletta, Langmuir 21 (2005) 9502– 9509. [27] B. Cabane, R. Duplessix, J. Phys. (Paris) 43 (1982) 1529–1542. [28] E. Pettersson, D. Topgaard, P. Stilbs, O. Söderman, Langmuir 20 (2004) 1138– 1143. [29] R. Meszaros, I. Varga, T. Gilanyi, J. Phys. Chem. B 109 (2005) 13538–13544. [30] L.A. Bastardo, J. Iruthayaraj, M. Lundin, A. De˙dinaite˙, A. Vareikis, R. Makuška, A. van der Wal, I. Furo, V.M. Garamus, P.M. Claesson, J. Colloid Interface Sci. 312 (2007) 21–33. [31] N. Peron, R.A. Campbell, T. Nylander, A. Vareikis, R. Makuška, T. Gilyani, R. Meszaros, J. Phys. Chem. B 112 (2008) 7410–7419. [32] S. Couderc-Azouani, J. Sidhu, T. Thurn, R. Xu, D.M. Bloor, J. Penfold, J.F. Holzwarth, E. Wyn-Jones, Langmuir 21 (2005) 10197–10208. [33] S. Couderc-Azouani, J. Sidhu, T.K. Georgiou, D.C. Charalambous, M. Vamvakaki, C.S. Patrickios, D.M. Bloor, J. Penfold, J.F. Holzwarth, E. Wyn-Jones, Langmuir 20 (2004) 6458–6469. [34] H.-J. Butt, M. Kappl, Surface and Interfacial Forces, Wiley-VCH Verlag GmbH& Co., Weinheim, 2010. [35] P.G. Hartley, I. Larson, P.J. Scales, Langmuir 13 (1997) 2207–2214. [36] G. Olanya, J. Iruthayaraj, E. Poptoshev, R. Makuška, A. Vareikis, P.M. Claesson, Langmuir 24 (2008) 5341–5349. [37] F. Joabsson, K. Thuresson, B. Lindman, Langmuir 17 (2001) 1499–1505. [38] A. De˙dinaite˙, L. Bastardo, Langmuir 18 (2002) 9383–9392. [39] R.G. Horn, D.T. Smith, W. Haller, Chem. Phys. Lett. 162 (1989) 404–408. [40] P.M. Claesson, J.L. Parker, J.C. Fröberg, J. Disp. Sci. Technol. 15 (1994) 375–397. [41] K.D. Kwon, V. Vadillo-Rodriguez, B.E. Logan, J.D. Kubicki, Geochim. Cosmochim. Acta 70 (2006) 3808–3819. [42] J. Iruthayaraj, E. Poptoshev, A. Vareikis, R. Makuska, A. van der Wal, P.M. Claesson, Macromolecules 38 (2005) 6152–6160. [43] R.G. Cox, Int. J. Multiph. Flow 1 (1974) 343–371. [44] D.Y.C. Chan, R.G. Horn, J. Chem. Phys. 83 (1985) 5311–5324. [45] E. Thormann, D.R. Evans, V.S.J. Craig, Macromolecules 39 (2006) 6180–6185. [46] M.U.R. Kjellin, P.M. Claesson, R. Audebert, J. Colloid Interface Sci 190 (1997) 476–484. [47] T. Drobek, N.D. Spencer, Langmuir 24 (2008) 1484–1488. [48] G. Dunér, Disp. Sci. Technol. 31 (2010) 1285.