Lubrication Properties of Bottle-Brush ... - ACS Publications

Feb 16, 2008 - (METAC) does not display the characteristics of a stick-slip motion, despite the ... adjacent measuring points are displayed in order t...
1 downloads 0 Views 241KB Size
3336

Langmuir 2008, 24, 3336-3347

Lubrication Properties of Bottle-Brush Polyelectrolytes: An AFM Study on the Effect of Side Chain and Charge Density Torbjo¨rn Pettersson,†,‡ Ali Naderi,*,† Ricˇardas Makusˇka,§ and Per M. Claesson†,| Department of Chemistry, Surface Chemistry, Royal Institute of Technology, Drottning Kristinas Va¨g 51, SE-10044 Stockholm, Sweden, ForceIT, MossVa¨gen 14, SE-153 37 Ja¨rna, Sweden, Department of Polymer Chemistry, Vilnius UniVersity, Naugarduko 24, LT-03225 Vilnius, Lithuania, and YKI, Institute for Surface Chemistry, Box 5607, SE-114 86 Stockholm, Sweden ReceiVed October 17, 2007. In Final Form: December 19, 2007 The effect of side chain to charge ratio on the frictional properties of adsorbed layers formed by bottle-brush polyelectrolytes with poly(ethylene oxide) side chains has been investigated. The brush polyelectrolytes were preadsorbed from 0.1 mM NaNO3 solutions onto mica and silica surfaces; the interfacial friction was then measured in polyelectrolytefree solutions via AFM (with the silica surface acting as the colloidal probe). It was concluded that the decisive factor for achieving favorable lubrication properties is the concentration of nonadsorbing poly(ethylene oxide) side chains in the interfacial region. However, contrary to what may be expected, the results showed that an ideal brush layer structure with the adsorbed polymers adopting comb-like conformation is not necessary for achieving a low coefficient of friction in the asymmetric mica-silica system. In fact, the lowest coefficient of friction ( 25.23 A combination of reflectometry and QCM-D data suggests23 that PEO45MEMA:METAC-25 layers on silica are slightly more extended than those formed by the higher charge density architectures. The polymers with the lowest charge density, X ) 2 and X ) 10 adopt extended conformations on silica. We note that the figures provided in Table 2 are appropriate indicators for the layer structure only when the backbone adopts a flat conformation on the surface. We will have these findings in mind when discussing the results. Force-Distance Profiles. As illustrated in Figure 10, PEO45MEMA:METAC-2 does not have the ability (under conditions similar to this work) to adsorb on mica, but it does adsorb on silica. It can therefore be expected that the adsorption of the polyelectrolyte will have a minor effect on the measured doublelayer force of the system, as none of the surface charges of the highly charged mica surface will be compensated for. The magnitude of the double-layer force is indeed very close to that (35) Kawaguchi, S.; Imai, G.; Suzuki, J.; Miyahara, A.; Kitano, T.; Ito, K. Polymer 1997, 38, 2885.

observed in the polyelectrolyte-free mica-silica system (compare Figures 3a and 4a), and the surface potential on silica, as determined from the force measurements, only dropped by 2 mV due to adsorption of PEO45MEMA:METAC-2 (the surface potential on mica was fixed at -70 mV). When the charge density of the polyelectrolyte is increased to 10%, the brush polyelectrolyte finds the ability to adsorb also onto the mica surface (Figure 10), which leads to compensation of more surface charges in the system. Thus, the magnitude of the double-layer force contribution is decreased. The higher adsorbed amount of the polyelectrolyte in the system (compared to PEO45MEMA:METAC-2) results in formation of thicker polymer layers and stronger steric forces (Figure 4a,b). At large separations, the steric force is rather small and thus the segment density in the overlap region is low at these separations. We also note the hysteresis in the force curves measured on approach and separation as observed under moderate compression. The origin of the hysteresis could in principle be due to either hydrodynamic forces or formation of new anchoring points between the brush polyelectrolyte and the adsorbing interface upon compression. The hydrodynamic force, FH, prior to contact between the adsorbed layers is given by36

η dD (D - 2LH) dt

FH ) - 6πR2

(3)

where R is the radius of the colloidal probe, D the distance between the surfaces, dD/dt the velocity (1 µm/s in our case), η the solvent viscosity, and LH the hydrodynamic thickness of the polymer layer. At high compression, the corresponding hydrodynamic interaction is given by36

FH ) -

()

4πR2 ηeff dD D 3 D dt ξD

2

valid for ξD , D < 2L (4)

where ηeff is the effective viscosity and ξD is the hydrodynamic screening length in the gap. If hydrodynamic forces are indeed the cause behind the hysteresis in the force curves, then both the compression and separation force curves should superimpose once the hydrodynamic interaction has been subtracted. However, attempts to fit either eq 3 or eq 4 to the results were not successful. Therefore, it can be concluded that the observed hysteresis is due to formation of new anchoring points during compression, as observed for other polymer-coated surfaces.37 The unsmooth profile of the force curve of PEO45MEMA:METAC-10 upon separation and the observed hysteresis in the interaction curves of the same system when studied by SFA25 give support to the given explanation. The same reasoning is offered for the hysteresis observed between PEO45MEMA:METAC-25 layers (Figure 5a). However, as shown in Figure 5b, after a shearing force has been applied, the hysteresis in the force curves disappears, the range of the steric force decreases, and the decay length of the longrange force becomes consistent with the Debye length. Thus, shearing has clearly resulted in the formation of more compact PEO45MEMA:METAC-25 layers. The importance of the amount of adsorbed PEO45 side chains is seen in Figures 5c and 6a,c, where minima in the compression force curves and weak adhesive forces in the separation curves (when the polyelectrolyte charge density is equal to or higher than 50%) are observed. These results differ from what has been recorded (for the same polyelectrolytes) in the symmetric micamica system.25 For poly(METAC), the long-range forces are (36) Klein, J. Annu. ReV. Mater. Sci. 1996, 26, 581. (37) Raviv, U.; Klein, J.; Witten, T. A. Eur. Phys. J. E 2002, 9, 405.

Bottle-Brush Polyelectrolytes as Lubricants

Langmuir, Vol. 24, No. 7, 2008 3345

similar in the two cases, whereas the adhesion measured on separation is much larger when two mica surfaces are used; this is consistent with an electrostatic bridging mechanism.38-41 In the case of PEO45MEMA:METAC-50 and PEO45MEMA: METAC-75, no minima in the force curve upon compression and no adhesive force upon separation were observed when two mica surfaces were used. The reason for why these features are observed in the mica-silica system is suggested to be that PEO45, which makes up the outer part of these layers, has a significant affinity for silica but no affinity for mica. The relatively small amount adsorbed on silica provides enough space for PEO45 chains from the polyelectrolyte layer on mica to reach the silica surface and form new favorable PEO45-silica anchoring points. Thus, the attractive force in the asymmetric mica-silica system is also in this case suggested to be a bridging attraction, but now mainly mediated by the PEO45 side chains rather than the charged backbone, as was the case for poly(METAC). This is suggested to explain the longer range of the attractive force in the presence of PEO45MEMA:METAC-90 compared to poly(METAC), as the PEO45 chains are expected to reside farther away from the surface than the backbone charges of the polyelectrolytes. In support of the bridging hypothesis, we note that a simple calculation (based on eq 5) on the magnitude of the nonretarded van der Waals force42 (FVDW) reveals that this force contribution away from contact is significantly smaller than the measured attractive force:

FVDW A )R 6(D - D0)2

(5)

In the calculations, D is the distance between the interacting surfaces, D0 denotes the plane of origin of FVDW (which was assumed to coincide with the hard wall, D0 ) 0), A is the Hamaker constant (which has a magnitude on the order of 1 × 10-20 J), and R is the radius of the sphere (equal to the radius of the colloidal AFM probe). Interestingly, the magnitude of the bridging and adhesive forces observed for PEO45MEMA:METAC-X polyelectrolytes with X > 25 are affected differently when a shear force is applied on the polyelectrolyte layers. As seen in Figure 5c,d, the attractive forces for PEO45MEMA:METAC-50 and PEO45MEMA:METAC75 that we attribute to PEO45 chains bridging to the silica surface disappear to a great extent after shearing. This indicates a more homogeneous coverage of the silica surface after shearing due to flattening of the adsorbed layer. In contrast, for PEO45MEMA: METAC-90 and poly(METAC), the magnitude of the adhesive force increases after shearing, and the distance from which the surfaces jump into adhesive contact also increases. Since the range of the bridging force depends on the length of the tails,38,39,41 this indicates that some chains are partly torn away from the surface during shearing, causing the layer to become more extended. This hypothesis finds support from the observation that an increase in the maximum applied load during the frictional measurements increases the attractive force measured upon separation directly after the shearing has been terminated; see Figure 11. (38) Åkesson, T.; Woodward, C.; Jo¨nsson, B. J. Chem. Phys. 1989, 91, 2461. (39) Ennis, J.; Sjo¨stro¨m, L.; Åkesson, T.; Jo¨nsson, B. Langmuir 2000, 16, 7116. (40) Claesson, P. M.; Poptoshev, E.; Blomberg, E.; Dedinaite, A. AdV. Colloid Interface Sci. 2005, 114-115, 173. (41) Dahlgren, M. A. G.; Waltermo, Å.; Blomberg, E.; Claesson, P. M.; Sjo¨stro¨m, L.; Åkesson, T.; Jo¨nsson, B. J. Phys. Chem. 1993, 97, 11769. (42) Israelachvili, J. Intermolecular and Surface Forces; Academic Press: Amsterdam, 1991.

Figure 11. Surface interactions of PEO45MEMA:METAC-90 upon separation of the surfaces, after conducting frictional measurements with varying maximum applied loads (FLoad): (squares) 0.6 mN/m, (circles) 1.5 mN/m, (triangles) 2.5 mN/m, and (stars) 4.2 mN/m. The measurements were conducted on the same spot on the PEO45MEMA:METAC-90-coated mica surface.

Frictional Properties. Instruments such as the SFA have been successful in revealing the mechanism behind the low effective coefficient of friction (µeff) between mica surfaces in aqueous solutions. It is now clear that this is due to the repulsive hydration forces, caused by the hydration layers of adsorbed ions at the surfaces.43,44 The key is that the hydration force has a high loadbearing capacity while at the same time the fluidity of the interfacial layer is retained. In view of this, it is not surprising that no measurable friction forces are encountered in this study at low loads when this load is carried by the repulsive doublelayer force. However, strong hydration and double layer forces are not present in most cases, and polymer coatings often need to be used to lubricate contacts and protect surfaces against wear. To compare the relevance of the exerted pressures (P) in this work with those in the contact area of, for example, human joints, the following set of equations based on Hertz contact mechanics can be used:45

E* )

((

) (

))

1 - νmica2 1 - νsilica2 + Emica Esilica a)

(

)

3RFLoad 4E*

P)

FLoad πa2

-1

(6)

1/3

(7) (8)

where Emica (56.5 GPa) and νmica (0.1) are the Young modulus and Poisson ratio of mica.46 The corresponding values for silica are Esilica ) 72 GPa, νsilica ) 0.17.47 R is the radius of the probe, FLoad is the normal load applied by the AFM cantilever, and a is the radius of the flat area at the contact point between the silica probe and the mica surface. Equations (6-8) are valid when no plastic deformation of the surfaces occurs during the loading process and when adhesive forces are absent or insignificant. (43) Homola, A. M.; Israelachvili, J. N.; Gee, M. L.; McGuiggan, P. M. J Tribol-T ASME 1989, 111, 675. (44) Raviv, U.; Klein, J. Science 2002, 297, 1540. (45) Meyer, E.; Overney, R. M.; Dransfeld, K.; Gyalog, T. Nanoscience: Friction and Rheology on the Nanometer Scale; World Scientific: River Edge, NJ, 1998. (46) Kopta, S.; Salmeron, M. J. Chem. Phys. 2000, 113, 8249. (47) Bamber, M. J.; Cooke, K. E.; Mann, A. B.; Derby, B. Thin Solid Films 2001, 398-399, 299.

3346 Langmuir, Vol. 24, No. 7, 2008

Pettersson et al.

These conditions are met in this work, since adhesive forces are not significant (except at the highest charge densities) and since repeated measurements on the same contact area presented similar force curves. For FLoad ≈ 4-8 nN (assuming that the coated surfaces have approximately the same E and ν as bare mica and silica), the maximum pressure (P) is in the range of 27-34 MPa. These are an order of magnitude higher than what has been measured for healthy joints, for example.20 Clearly the comb polyelectrolytes used in this work have a good load bearing capacity and they can thus find potential use in many applications. The frictional force needed to slide a body over a surface is related to the energy dissipation per unit time and unit surface area, W

FFriction )

WAeff V

(9)

where Aeff is the effective contact area and V the sliding velocity. It has been suggested36 that the main energy dissipation mechanism between sliding polymer-coated surfaces is the dragging of polymer chains through the interpenetration zone (with thickness d). Other dissipation mechanisms are the flow of solvent through the polymer layer, and the formation and breakage of anchoring points on opposite surfaces. The lower observed coefficient of friction between brush polymer layers compared to layers of adsorbed linear homopolymers has been assigned to lower chain interpenetration in the first-mentioned system.36 Since for brushes d ()sβ1/3, where s is the distance between the extended chains on the surface and β the degree of compression of the layer) varies slowly36 with compression, the interfacial region is able to maintain a high fluidity (low effective viscosity in the interface region due to the low amount of interpenetrating chains) even under significant compression. The generally low frictional forces observed in this study (for X < 90) suggest that the adsorbed layers in our study, even though they are not all formally brush layers,25 from a frictional point of view behave as brush layers. The reason for this is suggested to be the high side chain density on the (PEO45MEMA:METACX) polyelectrolytes that locally provides a high density of PEO45 side chains, which counteracts chain interpenetration and bridging. The importance of the PEO45 side chains for the frictional properties becomes evident when the friction curves between layers of PEO45MEMA:METAC-2 and layers of poly(METAC) (Figures 4c and 6e) are compared with that for the bare micasilica system (Figures 3b). The friction force of the system increases by an order of magnitude when poly(METAC) is adsorbed; this is due to electrostatic bridging, which brings an additional energy dissipative process to the system as polymersurface anchoring points are broken and re-formed. We note that coating only one of the interacting surfaces with PEO45MEMA:METAC-2 is enough to decrease the frictional force by a factor of approximately 4 compared to the bare micasilica case. This is because by adsorbing PEO45MEMA:METAC-2 on silica, the plane of shear between the mica and silica surfaces is moved to the interfacial region of the hydrated mica surface and the hydrated PEO45 chains (on the silica surface). In passing it is noted that similar results have been obtained by Yan et al.,21 who investigated the frictional properties of PLL-PEO polyelectrolytes in an asymmetric system. The frictional characteristics of polymer-brush-bearing surfaces have been addressed theoretically36,48 and through simulation49,50 (48) Sokoloff, J. B. arXiV:cond-mat 2006, 0602320. (49) Grest, G. S. AdV. Polym. Sci. 1999, 138, 149. (50) Kreer, T.; Mu¨ser, M. H.; Binder, K.; Klein, J. Langmuir 2001, 17, 7804.

and experimental work.51 It has been proposed that the effective coefficient of friction of moderately compressed brush-bearing surfaces in a good solvent can be estimated by36

(

)

FFriction (6πηeffνβ7/4)Aeff 6πηeffνs2 ) ) µeff ) FLoad sFLoad β1/2kBT

(10)

V < τzb-1 d In the above relation (derived for two flat surfaces), FFriction is the frictional force, s is the distance between the brush chains at the interface, kB and T are respectively the Boltzmann constant and temperature, d is the thickness of the interpenetration zone, and τzb is the relaxation time of the Zimm blobs, characterizing the layer structure. β, the degree of the compression of the layers, is defined as

β)

2L D

(11)

where L is the unperturbed brush layer thickness (in a symmetric system) and D is the distance between the interacting brush surfaces. The last equality in eq 10 holds under moderate compression when the normal force law follows the Alexander expression36 for the steric force between brush layers. Under the assumption that ηeff remains constant when compressing the layers (valid for moderate compressions),36 this expression allows several predictions. First, when the normal force curve increases steeply with compression (i.e., FLoad increases rapidly while β increases slowly), µeff should decrease with increasing applied load. This is indeed observed for systems with (10 < X < 90) where strong adhesive forces are absent. Further, µeff is predicted to decrease with decreasing s as qualitatively confirmed by Tsujii.16 In contrast to end-grafted polymers systems, adsorption of bottle-brush polymers (e.g., PEO45MEMA:METAC-X brush polyelectrolytes) at interfaces produces brush layers that are inhomogeneous on the molecular scale. In this picture, due to the strong lateral repulsions, the side chains are extended in all directions and exhibit different degrees of stretching perpendicular to the surface.25 The layer inhomogeneity is expected to become less pronounced as the layers become compressed, and the segment density in the gap is expected to become more homogeneous, thus approaching the constant segment density limit of the Alexander-de Gennes brush. Thus, eq 10 can be expected to be qualitatively valid and the lowest µeff should be obtained for the system with the highest amount of PEO45 side chains in the gap between the surfaces. The validity of this prediction in our polymer system is also evident in Figures 9a and 10b (see the dashed line), where it is observed that the PEO45MEMA:METAC-X polyelectrolytes with the highest concentrations of PEO45 segments (10 e X e 75) at the mica-silica interface exhibit the lowest coefficient of frictions. However, a closer investigation of the results (Figures 9a and 10) reveals that the lubrication properties depend to a significant degree on which of the surfaces has the highest concentration of adsorbed side chains. This conclusion is reached from the fact that while the total number of PEO45 segments in the interfacial region is approximately the same for brush polyelectrolytes with X ) 10-50 (see Figure 10b), PEO45MEMA:METAC-10, which adsorbs the most on silica, exhibits better lubrication properties (51) Klein, J.; Kumacheva, E.; Perahia, D.; Mahalu, D.; Warburg, S. Faraday Discuss. 1994, 98, 173.

Bottle-Brush Polyelectrolytes as Lubricants

than PEO45MEMA:METAC-50 (which adsorbed the highest on mica). This is consistent with the finding that poly(ethylene oxide) has a significantly larger affinity for silica than for mica. Thus, the silica surface must be fully covered in order to obtain the lowest possible friction against another surface exposing poly(ethylene oxide) chains.

Conclusions In this work, we used AFM to investigate the impact of the molecular structure of bottle-brush polyelectrolyes on their frictional properties; particular attention was paid to the effect of the side chain on the charge density ratio of the polymers. To this end, a series of PEO45MEMA:METAC-X brush polyelectrolytes were adsorbed from 0.1 mM NaNO3 on mica surfaces and silica spheres. It was shown that adsorption of the brush polyelectrolytes [with a density of the PEO45 side chains larger than 10% (on molar bases)] significantly reduces the already low coefficient of friction of the asymmetric bare mica-silica system. Furthermore, low friction forces are obtained already by coating only one of the interacting surfaces, but a more efficient reduction of the frictional forces is achieved when both surfaces are coated with the polyelectrolytes: coefficients of friction (µeff) as low as 0.006 under applied pressures as high as 30 MPa were achieved. Another finding of our study was that an ideal brush layer structure, with a backbone closely attached to the surface and side chains protruding away from it, is not necessary to achieve low µeff in the asymmetric mica-silica system. In fact, the lowest µeff was noted for PEO45MEMA:METAC-10, which has an

Langmuir, Vol. 24, No. 7, 2008 3347

extended layer conformation. It was therefore concluded that the concentration of nonadsorbing side chains at the interface, together with the resulting osmotic pressure, is one decisive factor for achieving low frictional forces. We also conclude that it is more important to have a high coverage on silica (to which the poly(ethylene oxide) side chains have an affinity) than to have a high coverage on mica (to which the poly(ethylene oxide) side chains have no or very limited affinity). Decreasing the number of side chains in the interfacial region resulted in an increase in the µeff. However, the good lubrication properties were upheld as long as the number of side chains on the surfaces remained sufficiently high; this was observed for systems having a graft-to-charge density ratio equal to or higher than 1:3. At even lower side chain densities, attractive electrostatic bridging forces became significant, and as a result, the friction forces increased significantly. The findings of this work suggest that the PEO45MEMA: METAC-X bottle-brush polyelectrolytes have the potential to be used in applications that demand very low coefficient of frictions. Acknowledgment. T.P., A.N., and P.C. acknowledge financial support by the Swedish Research Council (VR). The SwedishLithuanian collaboration was supported by the Marie Curie RTN “Self-organization under confinement, SOCON”. Dr. Johan Fro¨berg is thanked for providing the software for analyzing the DLVO interactions. Dr. Juan Jose´ Valle Delgado is thanked for his assistance. LA703229N