Langmuir 2005, 21, 2199-2208
2199
Atomic Force Microscopy Study of the Interaction between Adsorbed Poly(ethylene oxide) Layers: Effects of Surface Modification and Approach Velocity Scott C. McLean,† Hadi Lioe,† Laurence Meagher,‡ Vincent S. J. Craig,§ and Michelle L. Gee*,† School of Chemistry, University of Melbourne, Melbourne, Victoria 3010, Australia, CSIRO Molecular Science, Bag 10, Clayton South, Melbourne, Victoria 3169, Australia, and Department of Applied Mathematics, Research School of Physical Sciences, Australian National University, Canberra, ACT 0200, Australia Received August 17, 2004. In Final Form: December 13, 2004 The interaction forces between layers of the triblock copolymer Pluronic F108 adsorbed onto hydrophobic radio frequency glow discharge (RFGD) thin film surfaces and hydrophilic silica, in polymer-free 0.15 M NaCl solution, have been measured using the atomic force microscope (AFM) colloid probe technique. Compression of Pluronic F108 layers adsorbed on the hydrophobic RFGD surfaces results in a purely repulsive force due to the steric overlap of the layers, the form of which suggests that the PEO chains adopt a brush conformation. Subsequent fitting of these data to the polymer brush models of Alexander-de Gennes and Milner, Witten, and Cates confirms that the adsorbed Pluronic F108 adsorbs onto hydrophobic surfaces as a polymer brush with a parabolic segment density profile. In comparison, the interaction between Pluronic F108 layers adsorbed on silica exhibits a long ranged shallow attractive force and a weaker steric repulsion. The attractive component is reasonably well described by van der Waals forces, but polymer bridging cannot be ruled out. The weaker steric component of the force suggests that the polymer is less densely packed on the surface and is less extended into solution, existing as polymeric isolated mushrooms. When the surfaces are driven together at high piezo ramp velocities, an additional repulsive force is measured, attributable to hydrodynamic drainage forces between the surfaces. In comparing theoretical predictions of the hydrodynamic force to the experimentally obtained data, agreement could only be obtained if the flow profile of the aqueous solution penetrated significantly into the polymer brush. This finding is in line with the theoretical predictions of Milner and provides further evidence that the segment density profile of the adsorbed polymer brush is parabolic. A velocity dependent additional stepped repulsive force, reminiscent of a solvation oscillatory force, is also observed when the adsorbed layers are compressed under high loads. This additional force is presumably a result of hindered drainage of water due to the presence of a high volume fraction of polymer chains between the surfaces.
Introduction Poly(ethylene oxide) (PEO) is a polymer which, in various guises, has been used to enhance and improve material properties in numerous technological areas. It has been utilized within the industrial colloidal area because of its ability to act as a steric stabilizer (e.g., as a dispersant) and in flocculant formations.1-3 In recent years, there has been extensive interest in the incorporation of PEO into synthetic biomaterial systems to improve their biocompatibility. For example, the surface of many materials introduced into the body must often be modified to mask undesirable surface properties and avoid possibly severe clinical complications. Often, the first step in a cascade of events, which may end in thrombosis in a bloodcontacting application, is protein adsorption onto the surface of a biomaterial. Consequently, one may wish to markedly reduce or even eliminate adsorption of the first layer of protein molecules.4-8 * Corresponding author. † University of Melbourne. ‡ CSIRO Molecular Science. § Australian National University. (1) Baker, J. A.; Berg, J. C. Langmuir 1988, 4, 1055. (2) Elliott, S. L.; Russel, W. B. J. Rheol. (N. Y.) 1998, 42, 361. (3) Stenkamp, V. S.; Berg, J. C. Langmuir 1997, 13, 3827. (4) Espadas-Torre, C.; Meyerhoff, M. E. Anal. Chem. 1995, 67, 3108. (5) Kenworthy, A. K.; Hristova, K.; Needham, D.; McIntosh, T. J. Biophys. J. 1995, 68, 1921. (6) Lee, J. H.; Lee, H. B.; Andrade, J. D. Prog. Polym. Sci. 1995, 20, 1043.
Some studies have concluded that surface-grafted or adsorbed PEO layers must be densely packed, preferably with the molecules in a brush-type conformation, to effectively resist protein adsorption.9,10 Surface modification using PEO is commonly achieved either by a covalent grafting reaction or through the adsorption of a PEOcontaining block copolymer system.11,12 When cloud point grafting is used to covalently attach PEO to a surface, it is possible to form densely packed polymer layers.13 However this can also be achieved using PEO-containing block copolymers.14 These copolymer systems are generally either di- or triblock copolymers. One of the most widely studied of these is the commercially available Pluronic triblock copolymer.15-18 The Pluronics are of the form (7) Razatos, A.; Ong, Y.-L.; Boulay, F.; Elbert, D. L.; Hubbell, J. A.; Sharma, M. M.; Georgiou, G. Langmuir 2000, 16, 9155. (8) Snellings, G. M. B. F.; Vansteenkiste, S. O.; Corneillie, S. I.; Davies, M. C.; Schacht, E. H. Adv. Mater. 2000, 12, 1959. (9) Leckband, D.; Sheth, S.; Halperin, A. J. Biomat. Sci., Polym. Ed. 1999, 10, 1125. (10) Halperin, A. Langmuir 1999, 15, 2525. (11) Sofia, S. J.; Premnath, V.; Merrill, E. W. Macromolecules 1998, 31, 5059. (12) Kidane, A.; McPherson, T.; Shim, H. S.; Park, K. Colloids Surf., B 2000, 18, 347. (13) Kingshott, P.; Thissen, H.; Griesser, H. J. Biomaterials 2002, 23, 2043. (14) Marsh, L. H.; Coke, M.; Dettmar, P. W.; Ewen, R. J.; Havler, M.; Nevell, T. G.; Smart, J. D.; Smith, J. R.; Timmins, B.; Tsibouklis, J.; Alexander, C. J. Biomed. Mater. Res. 2002, 61, 641. (15) Green, R. J.; Tasker, S.; Davies, J.; Davies, M. C.; Roberts, C. J.; Tendler, S. J. B. Langmuir 1997, 13, 6510.
10.1021/la047942s CCC: $30.25 © 2005 American Chemical Society Published on Web 02/11/2005
2200
Langmuir, Vol. 21, No. 6, 2005
PEOn-PPOm-PEOn; where PPO is poly(propylene oxide) and PEO is poly(ethylene oxide). The usefulness of the Pluronics derives from the differing solubilities of the PEO and PPO blocks. The hydrated PEO blocks are hydrophilic and are thought to extend into the aqueous solvent phase to provide a protein resistant barrier. The PPO blocks, on the other hand, are hydrophobic and provide the driving force to adsorption of the polymer at an interface, particularly when the interface is hydrophobic.17,19,20 The surfaces of synthetic biomaterials are often hydrophobic and so it is important to understand the behavior of Pluronic molecules upon adsorption to such surfaces and the resulting adsorbed layer structure. To this end, numerous studies1,3,16,18,20,21 have monitored the adsorption of Pluronic species onto model hydrophobic surfaces. Such investigations have employed a variety of experimental techniques such as total internal reflectance microscopy,18 ellipsometry,22 reflectometry,23 photon correlation spectroscopy,1 and surface plasmon resonance.17,24,25 It should be noted that the majority of the studies mentioned above have also considered the protein resistance capabilities of adsorbed layers containing Pluronic molecules. Atomic force microscopy is another technique that can also be applied to the study of adsorbed layers. The atomic force microscope (AFM) was originally developed in 1986 as a means to image the topography of insulating surfaces using a sharp tip.26 Further development of AFM has led to its use as a tool for the measurement of surface, that is, the “colloid probe technique”, in a systematic manner.27,28 In this method, the sharp tip used in imaging is replaced with a spherical colloidal particle. There have been a number of studies which have used this technique to obtain direct force measurements between two surfaces with attached polymers.7,29-31 However, few of these have been carried out using the AFM to investigate the structure of adsorbed Pluronic layers. In addition, a number of recent publications have reported the use of the AFM in the investigation of hydrodynamic drainage forces.32-34 However, there is little work in the literature that has considered this phenomenon with the addition of an adsorbed polymer to the system. (16) Li, J. Q.; Caldwell, K. D. Colloids Surf., B 1996, 7, 9. (17) McGurk, S. L.; Green, R. J.; Sanders, G. H. W.; Davies, M. C.; Roberts, C. J.; Tendler, S. J. B.; Williams, P. M. Langmuir 1999, 15, 5136. (18) Bevan, M. A.; Prieve, D. C. Langmuir 2000, 16, 9274. (19) O’Connor, S. M.; DeAnglis, A. P.; Gehrke, S. H.; Retzinger, G. S. Biotechnol. Appl. Biochem. 2000, 31, 185. (20) Shar, J. A.; Obey, T. M.; Cosgrove, T. Colloids Surf., A 1998, 136, 21. (21) O’Connor, S. M.; Gehrke, S. H.; Retzinger, G. S. Langmuir 1999, 15, 2580. (22) Tiberg, F.; Malmsten, M.; Linse, P.; Lindman, B. Langmuir 1991, 7, 2723. (23) Schroe¨n, C. G. P. H.; Stuart, M. A. C.; van der Voort Maarschalk, K.; van der Padt, A.; van’t Riet, K. Langmuir 1995, 11, 3068. (24) Green, R. J.; Croneillie, S.; Davies, J.; Davies, M. C.; Roberts, C. J.; Schacht, E.; Tendler, J. B.; Williams, P. M. Langmuir 2000, 16, 2744. (25) Green, R. J.; Frazier, R. A.; Shakesheff, K. M.; Davies, M. C.; Roberts, C. J.; Tendler, S. J. B. Biomaterials 2000, 21, 1823. (26) Binnig, C.; Quate, C. F.; Gerber, C. Phys. Rev. Lett. 1986, 56, 930. (27) Ducker, W. A.; Senden, T. J.; Pashley, R. M. Langmuir 1992, 8, 1831. (28) Butt, H. J. Biophys. J. 1991, 60, 1438. (29) Meagher, L.; Maurdev, G.; Gee, M. L. Langmuir 2002, 18, 2649. (30) Jimenez, J.; Rajagopalan, R. Langmuir 1998, 14, 2598. (31) Kelley, T. W.; Schorr, P. A.; Johnson, K. D.; Tirrell, M.; Frisbie, C. D. Macromolecules 1998, 31, 4297. (32) Bonaccurso, E.; Kappl, M.; Butt, H.-J. Phys. Rev. Lett. 2002, 88, 076103/1. (33) Craig, V. S. J.; Neto, C.; Williams, D. R. M. Phys. Rev. Lett. 2001, 87, 054504. (34) Vinogradova, O. I.; Butt, H.-J.; Yakubov, G. E.; Feuillebois, F. Rev. Sci. Instrum. 2001, 72, 2330.
McLean et al.
Figure 1. Chemical structure of the radio frequency glow discharge thin film monomer units: (a) heptylamine monomer and (b) HMDSO monomer.
The primary aim of the present study was to examine the adsorption characteristics and layer structure of the Pluronic triblock copolymer F108 by measuring the surface forces between layers of this polymer adsorbed on the three types of surfaces, one hydrophilic and two hydrophobic. Specifically, we have used the AFM in colloidal probe mode to study the adsorption of Pluronic F108 onto silica and two types of hydrophobically modified silica. The two hydrophobic surfaces were plasma polymer (pp) thin films of heptylamine (n-HApp) and hexamethlydisiloxane (HMDSOpp) made by radio frequency glow discharge (RFGD), described below. One of the uses of PEO is in drug delivery where, in some situations, the polymeric or PEGylated liposome drug carriers are required to squeeze through narrow capillaries under high shear forces.33 It is therefore important that the behavior of PEO grafted layers under these conditions is better understood. Hence, we also investigated how varying degrees of shear force effect the interaction forces between and layer structure of adsorbed Pluronic F108. This was achieved by varying the approach velocity between the two surfaces containing adsorbed Pluronic and monitoring the affect of this on the surface interactions. Experimental Section Materials. The silica spheres used in this study were a gift from Allied Signal (Illinois, U.S.A.) prepared by a modified Sto¨ber process35 with a diameter between 4 and 6 µm, measured by high magnification optical microscopy. The AFM cantilevers were obtained from Digital Instruments (Santa Barbara, CA) and the EPON 1004 adhesive, used to attach the silica particles to the cantilevers, was obtained from Shell. Silicon wafer substrates used in the radio frequency glow discharge (RFGD) thin film deposition were obtained from Quality Semiconductor Australia (Homebush, NSW). Silica (Suprasil) was used as a flat substrate. Highly polished, flat silica samples were obtained from H.A Groiss, Ltd. (Wantirna, VIC, Australia). Water used in both cleaning and solution preparation was purified using a Milli-Q Plus water purification system. This produces water with a resistivity of 18.2 MΩ‚cm-1. RFGD thin films were prepared from the vapor of heptylamine (n-HA) and hexamethlydisiloxane (HMDSO; Aldrich, Milwaukee, U.S.A.). These liquids were used as received. The chemical structure of these “monomer” units is shown in Figure 1. All other chemicals used in solution preparation and cleaning were of AR grade and were used as received. Cleaning operations were carried out using a surfactant solution (1% RBS 35) supplied by PIERCE (Illinois, U.S.A.). Nitrogen gas (BOC Gases) was passed through a hydrocarbon and moisture trap and filtered using a 0.22 µm pore size membrane before use. The Pluronic F108 sample was a gift from BASF (Wyandotte, U.S.A.) and used without further purification. As shown in Figure 2a, it is a triblock copolymer comprised of two PEO buoy segments and one PPO anchor segment, with the PEO and PPO chemical structures illustrated in Figure 2b. Pluronic F108 has an average molecular weight of around 14 000. The PEO segments have a combined molecular weight of 10 750 and the PPO segment a molecular weight of 3250. These values correspond to the PPO anchor block containing 56 monomer units and each PEO segment containing 122 monomer units.3,36 However it should be noted (35) Sto¨ber, W.; Fink, A.; Bohn, E. J. Colloid Interface Sci. 1968, 26, 62. (36) Alexandridis, P.; Hatton, T. A. Colloids Surf., A 1995, 96, 1.
AFM of the Interaction between Adsorbed PEO Layers
Figure 2. (a) Schematic representation of the layer structure that Pluronic F108 molecules might be expected to adopt upon adsorption to a hydrophobic surface: PEO, poly(ethylene oxide); PPO, poly(proplyene oxide); RFGD, radio frequency glow discharge. (b) Chemical structure of the PEO and PPO segments. that because of the synthesis method employed, F108 is somewhat polydisperse.37 RFGD Thin Film Deposition. The RFGD thin film deposition was carried out in a cylindrical glass reactor chamber (diameter 175 mm; height 350 mm) containing two capacitively coupled, horizontal disk electrodes (diameter 150 mm, separation 90 mm). The “monomer” liquid, n-HA or HMDSO, was placed in a round-bottom flask connected to the reactor chamber by a stainless steel line and a manual flow control valve and was degassed prior to ignition of the plasma. The substrate materials were placed on the lower electrode for the deposition, with the plasma discharge powered by a commercial radio frequency generator (EN1 HPG-2). The RFGD thin film deposition parameters were different for formation of the polymer films n-HApp and HMDSOpp. The notation “pp” denotes plasma polymer. For n-HApp, the parameters were as follows: monomer pressure 0.127-0.130 Torr, frequency 200 kHz, load power 20 W, deposition time 20-30 s. Whereas for HMDSOpp the film deposition parameters were: 0.443 Torr, frequency 225 kHz, load power 10 W, deposition time 21 s. Although RFGD thin films are sometimes referred to as plasma polymers, they are not polymers in the traditional sense. Rather they consist of a random, tightly cross-linked three-dimensional network structure, incorporating various fragments which are derived from the initial “monomer”.38 It is known that within the first few days following deposition, hydrophobic n-HApp films rapidly incorporate oxygen within their network structure. Consequently, the film becomes progressively more polar and thus increasingly hydrophilic.39 Thus, to retain the hydrophobicity of the n-HApp thin films, all experiments were carried out 1 day after deposition. In contrast, HMDSOpp films remain stable for the first 1-2 weeks before any decrease in hydrophobicity is observed.40 For consistency, the HMDSOpp experiments were also carried out 1 day after deposition. Cleaning Protocol. The substrates (polished silicon wafers) were cleaned prior to coating with the RFGD thin film. They were first placed in an ultrasonic bath in a 1% (v/v) RBS 35 detergent solution for 1 h, rinsed with Milli-Q water, and finally soaked in AR grade ethanol. Immediately prior to placement in the reactor, the substrates were dried with a high velocity stream of high purity nitrogen gas. All glassware was cleaned with a 10% (w/v) solution of NaOH, followed by thorough rinsing with Milli-Q water. The AFM fluid cell, tube fittings, O-ring, tweezers and syringes were soaked in RBS 35 surfactant solution overnight and then rinsed with copious amounts of Milli-Q water. They were then (37) Chu, B. Langmuir 1995, 11, 414. (38) Griesser, H. J. Mater. Forum 1990, 14, 192. (39) Gengenbach, T. R.; Vasic, Z. R.; Chatelier, R. C.; Griesser, H. J. Plasmas Polym. 1997, 2, 91. (40) Gengenbach, T. R.; Griesser, H. J. Polymer 1999, 40, 5079.
Langmuir, Vol. 21, No. 6, 2005 2201 soaked in AR ethanol overnight. Immediately prior to each experiment, ethanol was removed by drying the equipment under a stream of high purity nitrogen gas. To ensure that the AFM fluid cell, hydrophilic silica substrates and cantilevers with attached silica spheres were rendered clean and free of surface impurities, they were ozone scrubbed, using a UV light source according to the method of Vig,41 for an hour just prior to the beginning of an experiment. Where silica was the substrate, both the silica colloid probe and fused silica substrates (cleaned using RBS 35 surfactant as above) were UV treated in the same manner. Surface Force Measurements. All surface force measurements obtained in this study were measured with a Nanoscope Multimode atomic force microscope (Digital Instruments, Santa Barbara, CA), using the colloid probe method. This involves the attachment of a silica sphere to the end of the cantilever tip using EPON 1004, as detailed by Ducker et al.27 For experiments where hydrophobic surfaces were required, RFGD thin films, formed from either n-HA (n-HApp) or HMDSO (HMDSOpp), were deposited onto both the probe and silicon wafer substrates. New probes were made before each deposition. Experiments conducted in this study employed cantilevers with spring constants of 0.076 N m-1, 0.11 N m-1, and 0.292 N m-1. The cantilever spring constants were determined using the method proposed by Cleveland et al.42 The average spring constant was calculated based on measurements on at least 10 cantilevers of each type. All force measurements presented here were carried out in polymer-free 0.15 M NaCl solution at natural pH (approximately pH 5.6). At the beginning of each experiment a series of force measurements were obtained between the two bare, polymerfree surfaces in 0.15 M NaCl solution to determine the slope of the constant compliance region (where probe and flat substrate are in contact). Subsequently, a 1000 ppm (∼1 mg/mL) solution of Pluronic F108 was introduced into the system and allowed to adsorb from a 0.15 M NaCl aqueous solution at natural pH for 5 h. The polymer solution was then flushed out with a salt solution (0.15 M NaCl, natural pH) and the system allowed to equilibrate for 15-20 min, whereupon a series of force measurements were carried out at various approach rates (piezo ramp rates). At least five force runs were obtained at any particular piezo ramp rate or contact position.
Results and Discussion As the absolute separation between the two interacting surfaces cannot be measured independently during an AFM force experiment, the zero surface separation is defined as the point at which the two adsorbed polymer layers are no longer further compressible on increasing the applied load. In this region of zero compressibility, there is a linear relationship between the cantilever deflection and the piezo travel. All force profiles presented in this section are semilog plots of the scaled force (F/R) versus surface separation, where R is the radius of curvature of the colloidal probe. This allows the decay length of exponentially decaying force profiles to be clearly identified, as well as providing some qualitative information about the conformation of the adsorbed layers. Note that the Pluronic F108 polymer segments are uncharged. Therefore any double layer force will emanate solely from the substrate materials (n-HApp, HMDSOpp and silica). Note, however, that, under the solution conditions used in this study (0.15 M NaCl and natural pH), the decay length of a double layer force is expected to be only 0.8 nm. Therefore any double layer interaction is negligible. The plots in Figure 3 are the measured force profiles for the interaction between two n-HApp surfaces after incubation in a 0.15 M salt solution containing 1000 ppm Pluronic F108, as described in the experimental section. (41) Vig, J. R. J. Vac. Sci. Technol. A 1983, 3, 1027. (42) Cleveland, J. P.; Manne, S.; Bocek, D.; Hansma, P. K. Rev. Sci. Instrum. 1993, 64, 403.
2202
Langmuir, Vol. 21, No. 6, 2005
Figure 3. Measured force/radius versus apparent surface separation for seven sequentially obtained force profiles of the interaction between Pluronic F108 adsorbed onto n-HApp. Pluronic F108 was adsorbed from a 0.15 M NaCl containing 1000 ppm of the polymer. Prior to the force measurements excess polymer was removed from solution. The force curves were obtained at a piezo ramp velocity of 598 nm s-1.
Note that we can state with certainty that adsorption of Pluronic F108 has occurred since these force profiles are very different to those measured before addition of the polymer (not shown); that is, no interaction was observed before hard wall contact between the two surfaces. The interaction between two n-HApp-adsorbed Pluronic F108 layers (Figure 3) is purely repulsive and decays exponentially, but the range of the interaction is much greater than expected for a double-layer repulsion in 0.15 M electrolyte. The onset of the repulsive force occurs at an apparent surface separation of approximately 19 nm and is most likely due to a polymer steric force, arising from the repulsion between the polymer chains as they undergo compression (resulting in a reduction in conformational entropy of the polymer chains and an increase in osmotic pressure as the local monomer concentration is increased).43 Interestingly, the force profile is typical in form to that expected for the interaction between two neutral polymer brush layers.44,45 The adsorbed conformation of Pluronic F108 is addressed in detail below. A piezo ramp rate of 598 nm s-1 was used in obtaining this set of data; therefore, hydrodynamic forces present within the system are small46 but not completely negligible (see later discussion). Note that the data obtained from seven consecutive force runs were reproducible and typical of that obtained throughout an experiment. This consistency is also observed when these force curves are obtained from a number of different contact positions, and as such implies surface homogeneity. Effect of Surface Hydrophobicity. In Figure 4, the force profiles obtained from the interaction between (43) Israelachvili, J. N. Intermolecular and Surface Forces with applications to colloidal and biological systems; Academic Press: Sydney, 1991. (44) de Gennes, P. G. Adv. Colloid Interface Sci. 1987, 27, 189. (45) Milner, S. T. Science 1991, 251, 905. (46) Chan, D. Y. C.; Horn, R. G. J. Phys. Chem. 1985, 83, 5311.
McLean et al.
Figure 4. Measured force/radius versus apparent surface separation for the interaction of Pluronic F108 adsorbed onto (A) n-HApp, (B) HMDSOpp, and (C) silica. For each force curve, Pluronic F108 was adsorbed from a 0.15 M NaCl containing 1000 ppm of the polymer. Prior to the force measurements excess polymer was removed from solution. The ramp velocities used to drive the surfaces together in the force measurement were (A) 598 nm s-1, (B) 797 nm s-1, and (C) 648 nm s-1. The inset contains force curve C, plotted linearly to highlight the presence of the weak, shallow attractive force. A theoretical prediction of the van der Waals force is also shown in the inset (line), offset to 9 nm from zero apparent surface separation to coincide with the perimeter of the adsorbed layers. Table 1. Contact Angle Data for Water on n-HApp39 and HMDSOpp Thin Films40 contact angle (deg) RFDG thin film
advancing
receding
n-HApp HMDSOpp
90 115
60 95
Pluronic F108 adsorbed onto three substrate pairs, each with different surface chemistries, are compared. Two of the substrate pairs are the hydrophobic n-HApp and HMDSOpp RFGD thin films and the third substrate pair is hydrophilic silica. Of these surfaces, HMDSOpp is expected to be the most hydrophobic (refer to Table 1).39,40 Force profiles, labeled A-C, correspond to the interaction between Pluronic F108 layers adsorbed onto n-HApp, HMDSOpp, and silica, respectively. The piezo ramp rates used in obtaining the force profiles for these three systems are comparable, that is, A, 598 nm s-1; B, 797 nm s-1; and C, 648 nm s-1. Note that the steric repulsive forces observed in force profiles A and B extend out to surface separations of around 19 and 22 nm, respectively. No attractive interactions arising, for example, from van der Waals forces, were measurable. As the two adsorbed polymer layers undergo compression, the steric repulsion increases monotonically until the surfaces reach a “hard wall” separation, at which point the layers are effectively incompressible. If we assume that this position equals a surface separation of zero, then an approximate estimate of the apparent thickness of an adsorbed Pluronic F108 layer is around 9.50 nm on n-HApp and 11.0 nm on HMDSOpp. It should be born in mind that there still remains a highly
AFM of the Interaction between Adsorbed PEO Layers
compressed polymer layer between the two surfaces at the hard wall position, so these thicknesses are lower estimates. The compressed layer thickness should be included for accurate fitting of the force profiles, thus offsetting the data to slightly higher surface separations (see later discussion). Further comparison of force profiles A and B indicates that when HMDSOpp is the substrate, the repulsive force is higher in magnitude at large surface separations and the compressive load required to reach the hard wall separation is slightly less than when n-HApp is the substrate. Nonetheless, force profiles A and B are remarkably similar, despite the fact that HMDSOpp and n-HApp are two chemically very different hydrophobic surfaces. Perhaps the small discrepancies are indicative of slight differences in the configuration adopted by the F108 molecules on adsorption to these two substrates. In contrast to force profiles A and B, force profile C, a measure of the interaction between two hydrophilic silica surfaces with adsorbed Pluronic F108, contains a short range repulsive force extending out to around 6 nm surface separation and a weak, shallow attractive force at larger separations (see inset in Figure 4). The short ranged repulsive force is most likely due to steric overlap of the adsorbed polymer layers, as discussed above. However, the range of this steric repulsion is less than that obtained when the underlying substrate is hydrophobic. This is a sensible result since F108 molecules are expected to have a lower affinity for hydrophilic surfaces leading to a less dense, more compressible polymer layer. The origin of the weak attractive force is not completely certain. A van der Waals force is a possibility so we have also included in the inset of Figure 4C a theoretical prediction of the van der Waals interaction (solid line) calculated using the Lifshitz theory.47,48 The oscillator strengths and frequencies for various dielectric relaxations were obtained from optical data for silica47 and water.49 Retardation of the interaction as a function of separation distance is implicitly included in the calculations as is the effect of salt screening.50 Note that, as a first approximation, we have used the data for silica, not silica plus an adsorbed layer of Pluronic. The predicted force has been shifted by 9 nm to zero apparent surface separation to coincide with the perimeter of the adsorbed layers. This comparison serves to demonstrate that van der Waals interactions can explain the attractive component of the forces present in the silica-Pluronic F108 system. Based on the work of Bevan and Prieve18 who measured van der Waals forces between Pluronic F108 layers adsorbed onto polystyrene, we expect that the calculation presented in Figure 4C could represent a lower limit for the expected force. Bevan observed experimentally that the presence of the adsorbed polymer layer increased the attractive van der Waals force slightly. While the presence of van der Waals forces can be used to explain the weak attraction between Pluronic F108 adsorbed onto silica, there are no obvious attractive forces present for Pluronic F108 layers adsorbed onto either hydrophobic nHApp or HMDSOpp surfaces. This might seem counterintuitive since van der Waals attraction should be present in all systems. However, one must bear in mind that direct force measurements yield the net force. A weak attractive van der Waals interaction could easily be offset by a stronger, longer ranged steric repulsion (47) Hunter, R. J. Foundations of Colloid Science; Clarendon Press: Oxford, 1987. (48) Pashley, R. M. J. Colloid Interface Sci. 1977, 62, 344. (49) Parsegian, V. A.; Weiss, G. H. J. Colloid Interface Sci. 1981, 81, 285. (50) Parsegian, V. A.; Gingell, D. Biophys. J. 1972, 12, 1192.
Langmuir, Vol. 21, No. 6, 2005 2203
yielding a net repulsive interaction, such as obtained when the substrate is hydrophobic. It is also possible that the weak attraction is due to attractive polymer bridging related to the relatively low surface coverage of Pluronic F108 when adsorbed onto silica. Since most of the theoretical studies of bridging interactions have used very idealized systems, it is not easy to compare them with results obtained here. It is believed that the adsorption of Pluronic F108 occurs via the hydrophobic pendant methyl groups of the PPO blocks, as shown schematically in Figure 2. Hence, F108 should have a greater affinity for the hydrophobic n-HApp and HMDSOpp surfaces than hydrophilic silica. The form and range of the force profiles that result from adsorption of Pluronic F108 on either n-HApp or HMDSOpp (Figure 4) are consistent with the formation of a polymer brush layer. This generally requires that the polymer adsorb with a high packing density and the tethered chains extend in a direction normal to the substrate surface.44,45 In contrast, adsorption of Pluronic F108 on silica results in a force profile with a form suggestive of adsorbed polymeric noninteracting or isolated mushrooms.51,52 To compare the above results with theoretical models, two expressions that predict the interaction between end grafted polymer layers on compression were used. The first of these is the Alexander-de Gennes theory for polymer brushes, namely,
F(D) 16KTπL 2L ≈ 7 R D 35s3
5/4
[( )
+
(2LD )
7/4
- 12
]
(1)
which is valid when D , R, D < 2L. The theory predicts the scaled force, F(D)/R, as a function of surface separation, D, for compression of two polymer brushes attached to crossed cylindrical surfaces of geometric radius R,44,52,53 with the assumption that the density of polymer segments is uniform normal to the surface. K and T are constants in this expression. L is the uncompressed brush layer thickness and s is the average distance between grafting points of the polymer on the surface. The second expression is derived from the Milner, Witten, and Cates (MWC) theory54,55 as used by Kenworthy et al.5 to express the distance dependence of the pressure, P, between bilayers exposing grafted PEO chains. Therefore to estimate the repulsive force between a sphere and a flat surface as a function of surface separation,56 the following equation was used:
F(D) ) 2πE(D) ) -2π R
∫ P(D) dD )
[ ( ) ( ) ]
4πP0
2L0 D + D 2L0
2
-
D 2L0
5
-
9 (2) 5
where
P0 )
( )
kTN π2 2 12
1/3
a4/3 s10/3
L0 represents the equilibrium brush thickness used in the MWC model. As eq 2 is derived for two interacting (51) Dolan, A.; Edwards, F. Proc. R. Soc. London 1974, 337, 509. (52) Efremova, N. V.; Bondurant, B.; O’Brien, D. F.; Leckband, D. E. Biochemistry 2000, 39, 3441. (53) Alexander, S. J. Phys. (Paris) 1977, 38, 983. (54) Milner, S. T. Europhys. Lett. 1988, 7, 695. (55) Milner, S. T.; Witten, T. A.; Cates, M. E. Macromolecules 1988, 21, 2610. (56) Derjaguin, B. V. Kolloid Z. 1934, 69, 155.
2204
Langmuir, Vol. 21, No. 6, 2005
McLean et al.
Table 2. Thickness of the Uncompressed Pluronic F108 Layer Adjusted for the Existence of a Compressed Polymer Layer between the Surfaces under High Loads surface
compressed layer thickness (nm)
adjusted layer thickness (nm)
n-HApp HMDSOpp silica
1.10 1.80 0.55
10.05 11.90 3.33
impenetrable brushes, rather than for a polymer brush confined by hard wall, a simple correlation of L0 ) 1.3L exists to equate the equilibrium brush thicknesses of the AdG and MWC models.52 The average distance, s, between grafting points on the surface can be calculated using the expression, s ) 2(σ/π)1/2, where σ is the surface area per polymer chain, directly related to the surface excess (mass per unit area) of adsorbed polymer. The number of segments in a polymer chain is N and a is the segment length. The segment length used for PEO in this study was 0.35 nm.5 The force profiles in Figure 4 were fitted to the theories of Alexander-de Gennes (AdG) and Milner, Witten, and Cates (MWC), above, using an iterative process for which it is necessary to nominate starting values for the fitting parameters, s and L. The starting value of s was obtained from previously reported experimental data for the surface excess and, hence, σ, of Pluronic F108 on polystyrene.57 A starting value of L was selected based on half the range of the steric repulsion measured in the force profiles of Figure 4, assuming that the total range is the sum of the polymer layer thickness on each interacting surface. In fitting the experimental data, it was also necessary to estimate a value for the fully compressed layer thickness as an offset to the experimental data. These values are 1.1, 1.8, and 0.55 nm for the n-HApp, HMDSOpp, and silica systems, respectively. They are given in Table 2 together with the adjusted layer thickness that takes into account the thickness of the compressed layer. The offset effectively shifts the experimental force profile along the x axis, thus increasing the experimental layer thickness by the offset amount. Note that Bevan and Prieve18 used total internal reflection microscopy (TIRM) to estimate a thickness 18.5 nm for adsorbed Pluronic F108. This thickness is comparable to those obtained in the present work. The discrepancy can be explained since Bevan and Prieve used a much higher salt concentration, that is, 0.4 M NaCl, and assumed a larger compressed layer thickness, that is, 3 nm. The adjusted layer thickness for Pluronic F108 on silica (Table 2) is very close to the calculated, unperturbed PEO molecule radius of gyration of around 3 nm3. Therefore the PEO segments are unlikely to be extending out into solution, as they would if adopting a brush conformation. They are more likely isolated mushrooms, as mentioned above. Finally, a numerical prefactor (see Table 3) was used to scale the theoretical profile to an appropriate F/R fit to the experimental data. Table 3 contains the parameters that yield the best fit to the experimental data for both the AdG and MWC theories. When compared to the experimental modified layer thickness given in Table 2, both theories underestimate the brush layer thickness for adsorption of Pluronic F108 on n-HApp. The theoretical estimate of the brush layer thickness is very good for adsorption on HMDSOpp. Interestingly, in both these systems, the AdG model yields a better estimate of the brush layer thickness than the MWC model. Both AdG and MWC yield a brush layer (57) Li, J.; Carlsson, J.; Huang, S.; Caldwell, K. D. In Hydrophilic Polymers: Performance with Environmental Acceptance; J. E. Glass, Ed.; American Chemical Society: Washington, DC, 1996.
thickness of 4.00 nm for adsorption of Pluronic F108 on silica. This is surprisingly close to the experimental value of L (Table 2) considering the unlikelihood of the polymer adopting a brush conformation on the silica surface, as discussed above. An attempt was also made to fit the silica data to the Dolan-Edwards equation,51,52 which is used to fit force profiles of the interaction between two surfaces with polymer adsorbed in a mushroom conformation. However an acceptable fit could not be obtained (results not shown). Figure 5a-c shows experimental force profiles of Figure 4, for the interaction of Pluronic F108 adsorbed on n-HApp, HMDSOpp, and silica, respectively, but now offset to take into account the compressed layer thickness, as described above. The AdG and MWC lines of best fit to these data are also shown. Interestingly, although the AdG model yields the best values for the brush layer thickness, as discussed above, it is clear that the MWC theory actually provides a much better fit to the data. The AdG theory predicts an interaction force that decreases too rapidly with surface separation compared to the experimental data, thus underestimating the contribution of the polymer tails to the net force. This was also observed in a recent study by Efremova et al.52 In contrast, the MWC theory gives a very good fit to the tail portion of the n-HApp data and only slightly overestimates the tail in the HMDSOpp and silica data. At small surface separations, the MWC theory is in excellent agreement with data for each of the three systems studied, whereas the AdG theory is not as good a fit. Note that both theories only model the repulsive section of the silica data and not the weak and shallow attractive interaction observed before the onset of repulsion. Efremova et al.52 suggested that the parabolic profile used in the MWC theory provides a better description of segment distribution within a brush layer than with the step function used in the AdG theory. Our findings support this conclusion. As detailed above, the average distance, s, between grafting points of the polymer is a fitting parameter in both the AdG and MWC theories, where s ) 2(σ/π)1/2. σ is the surface area per polymer chain, which is directly related to the surface excess of adsorbed polymer. The values of s and surface excess that best fit the experimental data for adsorption of Pluronic F108 on n-HApp, HMDSOpp, and silica are given in Table 3. In a previous study, Li et al.57 measured a surface excess of 2.1 mg m-2 for the adsorption of Pluronic on polystyrene from a 0.15 M phosphate-buffered saline solution. This converts to s ) 3.75 nm. These values compare favorably with both the AdG and MWC predictions for adsorption of Pluronic F108 on n-HApp and HMDSOpp, but again, the MWC model performs best. A number of groups have investigated the adsorption of F108 onto silica.58,59 Braem et al.58 measured a surface excess of 0.73 mg m-2, for the adsorption of Pluronic F108 on silica, this value being independent of the background NaCl concentration. This converts to s ) 6.4 nm. This is in fair agreement with the AdG and MWC predictions in Table 3. Effect of Approach Velocity. Figure 6 contains data for the interaction between Pluronic F108 adsorbed onto HMDSOpp from 0.15 M NaCl at three different piezo ramp velocities, that is, 0.797, 7440, and 44 600 nm s-1. Note here that we have now plotted F on the y axis not F/R. This makes no difference to the interpretation since R is the same in each case. The ramp velocity is equivalent to (58) Braem, A. D.; Prieve, D. C.; Tilton, R. D. Langmuir 2001, 17, 883. (59) Malmsten, M.; Linse, P.; Cosgrove, T. Macromolecules 1992, 25, 2474.
AFM of the Interaction between Adsorbed PEO Layers
Langmuir, Vol. 21, No. 6, 2005 2205
Table 3. Parameters Used in Fitting the Measured Force Profiles for the Interaction between Adsorbed Pluronic F108 to the Alexander-de Gennes (AdG) and Milner, Witten, and Cates (MWC) Models of Adsorbed Polymer Brushesa L (nm)
adsorbed amount (mg/m2)
s (nm)
prefactor
surface
AdG
MWC
AdG
MWC
AdG
MWC
AdG
MWC
n-HApp HMDSOpp silica
9.10 12.10 4.00
7.90 11.10 4.00
4.23 3.90 7.00
3.60 3.90 7.00
1.66 1.95 0.60
2.28 1.95 0.60
8.10 × 10-2 3.99 × 10-2 0
4.80 × 10-2 4.10 × 10-2 8.80 × 10-2
a
L is the uncompressed brush layer thickness, and s is the average distance between grafting points of the polymer on the surface.
Figure 5. (a-c) Measured force/radius versus apparent surface separation for the interaction of Pluronic F108 adsorbed onto (a) n-HApp, (b) HMDSOpp, and (c) silica, as in Figure 4, adjusted using an offset to take into account the compressed layer thickness of the adsorbed polymer (see Table 2). The experimental force profiles (O) are fitted to the Alexander-de Gennes model (solid line) and the Milner, Witten, and Cates model (dashed line) for the interaction of polymer brush layers.
the surface approach velocity at large surface separations, that is, before the AFM cantilever spring deflects. It is calculated from the z ramp size and the frequency of approach and retract runs in an AFM experiment. As the surfaces approach each other and the cantilever deflects in response to a force, the approach velocity is either increased or decreased depending on whether the inter-
action is attractive or repulsive, respectively. The interaction between Pluronic F108 adsorbed on HMDSOpp is always repulsive as a function of surface separation (see Figure 4), thus reducing the true approach velocity of the surfaces in comparison to the ramp rate. The rate of change in deflection of the cantilever is easily calculated;60 thus, the true approach velocity is known for the entire range
2206
Langmuir, Vol. 21, No. 6, 2005
Figure 6. Interaction force between Pluronic F108 layers adsorbed on HMDSOpp in 0.15 M NaCl, for apparent surface separations grater than 10 nm. The force was measured at piezo ramp rates of 797 nm s-1 (b), 7440 nm s-1 (O), and 44 600 nm s-1 (9). The difference in these force profiles is due to a rate dependent hydrodynamic force, which increases as ramp rate is increased. Note that the high compressive force data is shown in Figure 9.
of the force versus data. It is the true approach velocity that is used in all the hydrodynamic calculations presented here. The major point of interest with respect to Figure 6 is that, as the piezo ramp rate is increased, an additional repulsive force is detected. Given the ramp rate dependency of this repulsion, it must be due to the presence of hydrodynamic forces that result from hindered drainage of water from within the gap between the surfaces, as surface separation is decreased. Note that at the high ramp rates used, there is a possibility that the brush layer may also undergo nonequilibrium deformations. However, in this study it has not been possible to determine this, since no influence of the hydrodynamic flow on the configuration of the brush layer could be detected in the measurements. At the lowest piezo ramp rate the hydrodynamic forces are small, therefore the interaction observed in this case is dominated by the steric interactions of the polymer layers, as expected. One approach for evaluating the hydrodynamic component of the total force measured for the higher piezo ramp rates is to subtract the steric component of the force from the total force. However this is not easily achieved since, even at the lowest rate of approach, some hydrodynamic force is present. A further problem with a point by point subtraction approach, is that both points need to be measured at the same surface separation. The complication arises because once the data has been adjusted to account for the cantilever deflection, and since each deflection is different, the apparent separation values will be different for every curve. Consequently, a simple point by point subtraction is impossible. Another alternative would be to fit the curves to the experimental data and (60) Craig, V. S. J.; Neto, C. Langmuir 2001, 17, 6018.
McLean et al.
Figure 7. Interaction force between Pluronic F108 layers adsorbed on HMDSOpp in 0.15 M NaCl, for apparent surface separations greater than 10 nm, measured at a piezo ramp of 797 nm s-1 (b). The interaction measured using piezo ramp rates of 7440 nm s-1 (O) and 44 600 nm s-1 (9) have the calculated hydrodynamic force subtracted. A shear plane at a surface separation of 10 nm (5 nm per surface) was required to fit the high piezo ramp velocity data minus the hydrodynamic repulsion to the low velocity data. A viscosity of 1.00 mPa‚s and the instantaneous approach velocity was used to calculate the hydrodynamic force.
then subtract the values of the fits. However, this is not only more complicated than the approach that we have adopted (as described below), but also more prone to error given the level of noise within the experimental data. Therefore, in the approach we have adopted, we calculate the hydrodynamic force using the velocity difference between the systems and subtract the calculated force from the measured interaction at the higher approach rates (7440 and 44 600 nm s-1). So in effect we treat the piezo ramp velocity as 6643 nm s-1 and 43 803 nm s-1, respectively, and then correct the approach velocity for the influence of the cantilever deflection. When the correct hydrodynamic interaction is subtracted, the manipulated data should coincide with the data obtained at the lowest approach velocity (797 nm s-1). When calculating the hydrodynamic force the only parameter that is not known is the position of the shear plane. 33 By systematically varying the position of the shear plane for the calculated hydrodynamic force and comparing the predicted force to the experimentally obtained force data, the position of the shear plane for our experimental system can be determined. It was found that subtraction of the hydrodynamic force calculated using a shear plane at a separation of 5 ( 1 nm from the hard contact (or compressed polymer separation) from the higher velocity force curves brought them into excellent agreement with the low velocity data (see Figure 7). Note that a fit is not obtained at separations of less than ∼10 nm as the hydrodynamic data is meaningless at separations inside the shear plane. Thus we can conclude that the flow profile of the aqueous solution extends approximately 7 nm into the approximately 12 nm thick brush layer. This is
AFM of the Interaction between Adsorbed PEO Layers
Langmuir, Vol. 21, No. 6, 2005 2207
Figure 8. Schematic representation of the flow velocity of water during approach of the sphere (colloidal probe) to the flat surface in the region of the polymer brush. The dotted line indicates the extrapolated velocity at the boundary of the brush that determines the location of the shear plane (no slip boundary). The arrows indicate the flow velocity of the solvent and indicate that the flow velocity decreases more gradually in the brush layer, as described by Milner.61
depicted schematically in Figure 8. Milner61 has shown that the fluid velocity profile extends further into a polymer brush consisting of a parabolic segment density profile (MWC model) than for a step-function brush (AdG model) and consequently the slip plane will exist at a greater depth within the polymer brush for the former. Further, for the step-function brush the slip plane is located very close to the outer edge of the brush; therefore, the hydrodynamic interactions we have inferred from our measurements strongly suggest that the segment chain density profile of the brush formed by Pluronic F108 adsorbed onto a hydrophobic surface such as the HMDSOpp thin film is parabolic in nature. This is in agreement with our observation that the MWC model gives superior fits to the experimental force data and with the previous work by Efremova et al.52 It is noted in the work of Bevan and Prieve18 that the hydrodynamic shear plane and onset of repulsive steric forces were not at exactly the same position, with the shear plane being located approximately 1-3 nm inside the adsorbed F108 layers. Our work confirms this finding and suggests that the approach velocities used here, being so much larger than those expected in TIRM measurements, result in greater penetration of the flow field into the polymer brush. We also observe a velocity dependent force for surface separations of less than 10 nm (see Figure 9). The highest piezo ramp velocity employed (i.e., 44 600 nm s-1) results in an increased force over that obtained at the slower piezo ramp velocities (see Figure 9). In this surface separation region, the polymer brush layers are strongly compressed, forcing the removal of water from between the polymer chains and increasing the polymer density in the gap between the HMDSOpp substrates. The flow of water is highly constrained due to the density of polymer chains and therefore at high approach rates the hindered flow of water from within the compressing polymer layers results in an additional repulsion. Attempts were made to fit this component with a hydrodynamic force, locating the shear plane at the zero apparent surface separation. However, we found that the viscosity values required to obtain an acceptable fit were very low (approximately 4 mPa‚s), compared to those one might expect given the local concentration of confined polymer and the fact that the adsorbed, compressed polymer chains have reduced (61) Milner, S. T. Macromolecules 1991, 24, 3704.
Figure 9. Interaction force between Pluronic F108 layers adsorbed on HMDSOpp in 0.15 M NaCl, for apparent surface separations less than 10 nm. The force was measured at piezo ramp rates of 797 nm s-1 (b), 7440 nm s-1 (O), and 44 600 nm s-1 (9). The presence of a rate dependent force component is apparent for the data captured using a ramp rate of 44 600 nm s-1. This additional component may be attributed to the egress of water being squeezed from within the polymer brush layers upon compression. The inset highlights the steps obtained in the measured force with arrows used to denote the position of the steps.
degrees of motion compared to those in solution. We therefore propose that a percolation calculation under squeezing would be more appropriate. At separations below 1 nm, steps are apparent in the force data obtained at all three piezo ramp velocities. These are highlighted in the inset in Figure 9. Heuberger et al.62 have also observed steps in the interaction force data measured between PEO layers of about 1.25 Å, which is a little smaller than what has been meausered in this study. Data of this form is often associated with an oscillatory solvation force resulting from the expulsion of layered molecules between surfaces,63,64 though it is unlikely that such solvent layering exists within a highly compressed polymer layer. Finally, since the force profiles at all three piezo ramp rates are similar under higher loads, it appears that the structure of the layers are not markedly affected by the higher shear conditions being imposed. This finding agrees with the results of neutron reflectivity measurements of the segment density distribution of polymer layers under shear65 and has implications for drug delivery vehicles, for example, PEGylated liposomes in high shear conditions, such as those present in vivo. Conclusions The interaction forces obtained between Pluronic F108 layers adsorbed onto hydrophobic RFGD surfaces in 0.15 (62) Heuberger, M.; Drobek, T.; Voeroes, J. Langmuir 2004, 20, 9445. (63) Christenson, H. K.; Gruen, D. W. R.; Horn, R. G.; Israelachvili, J. N. J. Chem. Phys. 1987, 87, 1834. (64) McGuiggan, P. M.; Pashley, R. M. J. Phys. Chem. 1988, 92, 1235. (65) Ivkov, R.; Butler, P. D.; Satija, S. K.; Fetters, L. J. Langmuir 2001, 17, 2999.
2208
Langmuir, Vol. 21, No. 6, 2005
M NaCl solutions were repulsive in nature due to the steric overlap between the adsorbed polymer layers. The range and form of the force profiles suggest that the PEO chains of adsorbed Pluronic F108 extend normal to the surface forming a brush layer. In contrast, the interaction between Pluronic F108 adsorbed onto silica consists of a longer ranged, shallow attractive force and a shorter ranged steric repulsion. The attractive component is reasonably described by a van der Waals surface interaction, but a polymer bridging force cannot be ruled out. The range of the steric repulsion suggests that the PEO blocks of the polymer molecules are less extended that when adsorbed onto a hydrophobic surface, existing as noninteracting, isolated polymeric mushrooms. The force profiles could be fitted to the theories of Alexander-de Gennes (AdG) and Milner, Witten, and Cates (MWC), with the MWC theory giving consistently superior results. This finding confirms that Pluronic F108 adsorbs onto hydrophobic surfaces as a polymer brush with a parabolic segment density profile in line with the assumptions of the MWC model. We found that, for adsorption of Pluronic F108 onto silica, the force profiles were also best fitted with the WMC model, despite literature suggestions that the PEO blocks adsorb onto silica surfaces. To obtain good agreement between the experimental data and the theoretical fits it was necessary to offset the experimental data to slightly higher apparent surface separations to account for the presence of compressed polymer layers between the surfaces. From the fitting parameters used in these expressions it was possible to estimate an adsorbed amount for Pluronic F108 on each surface. These estimates compare favorably with literature
McLean et al.
values for the adsorption of Pluronic F108 on polystyrene and silica. When the surfaces with adsorbed polymer are driven together a high piezo ramp velocities, an additional repulsive force is measured, attributable to the presence of hydrodynamic drainage forces between the surfaces. Comparison of theoretical predictions of the hydrodynamic force to the experimentally obtained data suggests that the position of the hydrodynamic shear plane is approximately 7 nm inside the polymer brush. This implies that the flow profile of the aqueous solution penetrates significantly into the polymer brush, in line with the suggestions of Milner. This provides additional evidence that the segment density profile of the adsorbed polymer layer is parabolic in nature. When the adsorbed layers are highly compressed and so at small surface separations, that is, less than 10 nm apart, there exists an additional velocity-dependent repulsive force. It is most likely that this repulsion arises from the hindered flow of water through the compressed polymer layers when high piezo ramp rates are used. Interestingly, steps in the measured force are observed that are reminiscent of an oscillatory solvation force. Acknowledgment. M.L.G. would like to thank the Australian Research Council (ARC) and the University of Melbourne for their financial support of this project. The Particulate Fluids Processing Centre at the University of Melbourne, a Special Research Centre of the ARC, is also gratefully acknowledged for the infrastructure provided. V.S.J.C. would like to thank the Australian Research Council for the provision of a Research Fellowship. LA047942S
