1534
Langmuir 2008, 24, 1534-1540
Adhesion and Stable Low Friction Provided by a Subnanometer-Thick Monolayer of a Natural Polysaccharide† Delphine Gourdon,*,‡ Qi Lin,‡ Emin Oroudjev,§ Helen Hansma,§ Yuval Golan,| Shoshana Arad,⊥ and Jacob Israelachvili‡ Department of Chemical Engineering and Department of Physics, UniVersity of California, Santa Barbara, California 93106, Department of Materials Engineering and the Ilse Katz Center for Nanoscience and Technology and Department of Biotechnology Engineering, Ben-Gurion UniVersity of the NegeV, Beer-SheVa 84105, Israel ReceiVed July 25, 2007. In Final Form: October 8, 2007 Using a surface forces apparatus, we have investigated the adhesive and lubrication forces of mica surfaces separated by a molecularly thin, subnanometer film of a high-molecular-weight (2.3 MDa) anionic polysaccharide from the algae Porphyridium sp. adsorbed from aqueous solution. The adhesion and friction forces of the confined biopolymer were monitored as a function of time, shearing distance, and driving velocity under a large range of compressive loads (pressures). Although the thickness of the dilute polysaccharide was 0), the elastically deformed (flattened) junction diameter 2r was typically 20 to 90 µm in the load range from 0.1 to 38 mN, which (18) Hansma, H. G.; Bezanilla, M.; Zenhausern, F.; Adrian, M.; Sinsheimer, R. L. Nucleic Acids Res. 1993, 21, 505-512.
means that the maximum pressure P ) F⊥/A in the contact area ranged from 1 to 10 MPa. No Bridging Ability. To establish which kind of adhesion mechanisms are involved in the good adsorption of the polysaccharide on single surfaces, we first brought two bare mica surfaces into flat adhesive contact (in air) in the SFA and then introduced the polysaccharide solution, allowing for the biopolymer eventually to adsorb simultaneously to both contacting surfaces. We reasoned that the molecules adsorbing to the surfaces at the edge of the junction might bind to both surfaces and would then act as adhesive bridges or tethers when the surfaces were separated. Surprisingly, no adhesive bridging was measured between the mica surfaces even after prolonged contact (result not shown but identical to those in Figure 2b). Pressure-Induced Adhesion. In this section, we present SFA force-distance and adhesion data measured when two mica surfaces were brought together in polysaccharide solution B after they were left far apart for 20 min to allow the polymer to adsorb. Figure 2a shows the measured force profile between two mica
Adhesion, Stable Low Friction of a Polysaccharide
Langmuir, Vol. 24, No. 4, 2008 1537
surfaces in water (no polyelectrolyte) as the applied load was max -1 increased to F⊥ ) Fmax ⊥ ) 12.9 mN (F⊥ /R ) 712 mN m , P ) 3.5 MPa). As seen on the semilog scale plot (inset of Figure 2a), the force profile can be fitted to the theoretical approximate expression for the exponentially decaying electrostatic doublelayer repulsion in monovalent electrolyte solution, given by
( )
F⊥ eψ0 -κD ) 128πnskBTκ-1 tanh2 e R 4kBT
(1)
where ns is the bulk salt concentration, kB is the Boltzmann constant, T is the temperature, ψ0 is the surface electrostatic potential, e is the electronic charge, and κ - 1 is the Debye screening length, which for monovalent electrolytes is given by
κ-1 )
[ ] 0kBT 2nse2
1/2
(2)
Equation 1 is valid for D larger than the Debye length. The thin line in Figure 2a (inset) is the best exponential fit to the measured double-layer force, which gives a Debye length of κ -1 ) 18.5 ( 0.2 nm and a surface potential of ψ0 ) -135 mV. From eq 2, this corresponds to a salt concentration of ns ) 0.26 mM. This long-range pure double-layer repulsion dominates down to ∼10 nm, below which there is an additional short-range repulsive hydration force instead of the attractive van der Waals force expected from the Derjaguin-Landau-Verwey-Overbeek (DLVO) theory.19 Figure 2b-d shows the force-distance curves obtained in were polysaccharide solution as the maximum loads Fmax ⊥ steadily increased from 1.9 mN (P ) 1.2 MPa) to 17.7 mN (P ) 4.4 MPa). Here, too, the measured forces (In) are exponentially repulsive, and assuming that eq 1 still holds, the measured (fitted) Debye length of 12.6 ( 0.1 nm is smaller than that measured in water (18.5 nm). Qualitatively, a decreased Debye length is expected from the increase in the ionic strength. On separating the surfaces (Out), an adhesion force, Fad ⊥ , was measured, which ad increased with Fmax ⊥ (Figure 2b-d), plateauing at F⊥ /R ≈ 50 mN m-1 (corresponding to W ≈ 6 mJ m-2), as summarized in the inset of Figure 2d. This measured adhesion, achieved by compressing the films (and/or shearing them as described in the next section), is likely to be due to H bonding between each polysaccharide molecule (or the sugars forming the polysaccharide) and the two confining mica surfaces, as described in the Discussion. Note that no oscillations were ever measured in the force or adhesion profile.19 Friction/Lubrication Properties of Surfaces Sheared in Polysaccharide Solution. We also used the SFA to measure both the shear (friction) forces, Fll, and simultaneously the adhesion forces, F⊥, while continuously visualizing the sheared area of the confined films. To characterize the lubrication properties of the polysaccharide and to determine whether we could orient the polymer molecules (or its constituent monosaccharides) by different means, the polysaccharide was confined and sheared under various experimental conditions. This method has recently been found to be very effective for ordering alkane chains20 and nanoparticles21 between surfaces. It involves monitoring the thickness and lateral uniformity of the films during (19) Israelachvili, J. N. Intermolecular and Surface Forces, 2nd ed.; Academic Press: San Diego, CA, 1992. (20) Drummond, C.; Alcantar, N. A.; Israelachvili, J. N. Phys. ReV. E 2002, 66, 011705. (21) Gourdon, D.; Yasa, M.; Godfrey Alig, A.; Li, Y.; Safinya, C.; Israelachvili, J. N. AdV. Funct. Mater. 2004, 14, 238-242.
Figure 3. (a) Friction force Fll versus time t recorded as two mica surfaces are sheared past each other in a dilute polysaccharide solution (solution B). (b) Steady-state friction force Fll and contact area A (×) versus normal load F⊥ measured in solution B. The data refer to loading conditions (increasing F⊥). The dashed line is the JKR fit to the area vs load data at low loads using K ) 6.5 × 109 N m-2 and W ) 7.5 mJ m-2. Diamonds and circles correspond to the friction data measured during the first and second runs (amplitude of shear, 60 µm) indicating a large drop in the friction force. Triangles are the friction data measured during the third run, when the shearing amplitude was increased to 120 µm but the driving velocity of V ) 0.6 µm s-1 was kept constant, showing that the low-friction regime persists even over shearing distances larger than the contact diameter.
the shearing while the normal (compressive) pressure and lateral (friction) forces are also being measured. Friction Behavior versus Pressure: Transition to Stable Low Friction. The friction force, which is particularly informative because friction is very sensitive to small changes in the molecular ordering of a film, was measured as a function of load, driving velocity, shearing distance, and time at different polysaccharide concentrations. Figure 3a shows typical friction traces Fll versus time t recorded for two shearing mica surfaces in contact in the dilute polysaccharide solution (solution B). The friction forces are smooth (no stick-slip) and slightly decrease with shearing time (not noticeable on the traces displayed in Figure 3a, which shows friction after it has reached steady state), indicating that the polysaccharide conformation within the film is changing, although no change in the film thickness was measured. Figure 3b shows the steady-state friction force Fll as a function of both the applied load F⊥ and the corresponding contact area A (×). Between each set of data (i.e., each run), the surfaces were separated for 30 min to allow for possible relaxation of the previously sheared polymer. During the first sliding run (i.e., the first compressive load cycle while shearing), the friction Fll is proportional to the contact area A at low loads (indicative of adhesion-controlled friction), and the area of contact A (identical for all runs) is well described by the Johnson-Kendall-Roberts (JKR) equation.22 Thus, initially Fll R A at low loads, and Fll is finite at F⊥ ) 0. In contrast, during the second run over the same area, Fll increases linearly with F⊥ (indicative of load-controlled
1538 Langmuir, Vol. 24, No. 4, 2008
friction) with a remarkably decreased friction coefficient of µ ) 0.015 and Fll ≈ 0 at F⊥ ) 0. This transition reveals the capacity of the polysaccharide to both reorient (reorder) and to carry high loads because no damage was ever observed in the experiments shown in Figure 3. A likely explanation is that, during the first run, the molecules experiencing the shearing force become aligned progressively (cooperatively); however, there is still an important adhesive (bridging) contribution that makes the friction high. During surface separation, the sheared molecules keep their inplane alignment, but all binding sites have enough space and time to point toward a single mica surface, “canceling” the adhesive contribution and making the friction very low for the second run. The third run shown in Figure 3b was for a doubling of the shearing distance (amplitude or strain) from 60 to 120 µm (∼4r) at the same shearing velocity, showing that the low friction force persists even for amplitudes larger than the contact diameter. This suggests that once oriented on one surface the polysaccharide is carried with that surface into new areas, where it continues to confer low friction and wear resistance. Also, the film thickness measured during the friction experiments shown in Figure 3 ranged from 0.5 to 1.0 nm, implying that a subnanometer-thick monolayer of the dilute, very mobile polysaccharide is enough to provide efficient lubrication and wear protection at least at pressures up to 10 MPa. Note also that, as discussed later in the Shear-Induced Adhesion section, the large reduction in the friction coefficient was accompanied by a gradual increase in the adhesion force from W ≈ 0 mJ m-2 before shearing (repulsive profile) to W ≈ 3 mJ m-2 after the first run, reaching a maximum value of ∼6 mJ m-2 after the second run (data not shown). After the second run, both Fll and W remained roughly constant even after separating the surfaces and bringing them back into contact (within 30 min), which indicates irreversible reordering, and no removal, of the molecules from within the ultrathin films during the initial shear alignment. As mentioned above, all of the data shown in Figure 3 correspond to wearless friction. No damage was ever observed, even at F⊥ ) 35 mN when sliding over a distance equal to or smaller than the contact diameter (for example, at the end of the second run). Damage was observed at F⊥ ) 40 mN (corresponding to a pressure of around 11 MPa) only when the sliding distance was extended to twice the contact diameter (i.e., at the end of the third run of Figure 3; see also Figure 4a, bottom trace), which was the last measurement performed in the wearless regime for solution B. Figure 4a shows a comparison of the friction traces Fll versus t measured in the concentrated and dilute polysaccharide solutions just before and just after the first sign of damage as visualized by the FECO images. The experimental conditions are identical (i.e., shearing amplitude of 120 µm and velocity of 0.6 µm s-1 (corresponding to the third run in each case)). Before damage occurs, the friction measured on Sol A is more than 20 times higher than the one measured on Sol B, accompanied by a higher film thickness in the case of sol A (∼2.0 nm). A possible explanation is that, because of the higher density of polymer, the reorientation (realignment) of monomers gets more difficult (over the same time scale) and the polymer is less easily dragged with the surfaces while shearing, also making it more difficult to “push” the nonaligned molecules out of the shearing junction. A higher degree of entanglements in the thicker film formed by Sol A could also explain the higher friction. In both cases, the friction traces exhibit a transition from a “smooth” to a “rough” or “noisy” profile, typical of damaged surfaces. Figure 4b displays (22) Johnson, K. L.; Kendall, K.; Roberts, A. D. Proc. R. Soc. London, Ser. A 1971, 324, 301-313.
Gourdon et al.
Figure 4. (a) Friction traces measured in concentrated solution A (top) and dilute solution B (bottom) before, during, and after damage occurs. (b) Comparison of wearless friction forces Fll vs normal load F⊥ measured in solution A (squares) and solution B (circles). Triangles show the steady-state friction forces after damage (wear) has occurred in both films.
Figure 5. Steady-state friction forces Fll vs driving velocity V in solution B. Squares are friction data measured when shearing starts from quiescent conditions at a load F⊥ of 14.1 mN. Triangles are friction data measured following a compression cycle (identical to the one shown in Figure 3b) when the load F⊥ is set back to 12.8 mN. In both cases, the friction is nearly constant (logarithmic growth) over 3 decades. The friction traces (not shown here) were always smooth (i.e., no stick-slip was observed under any of the conditions shown).
the corresponding friction forces Fll versus load F⊥ where we can see that after damage has occurred the friction laws for the two different concentrations coincide, with an average friction coefficient of µ ) 0.23, which is the typical value for damaged (rough) mica both in air and in water23 (i.e., independent of the solution conditions). Friction Behavior versus Sliding Velocity: Weak Dependency. Figure 5 shows that the friction force Fll at constant load and shearing distance is nearly independent of the sliding velocity over a range of 3 decades from 10 nm s-1 to 10 µm s-1, exhibiting
Adhesion, Stable Low Friction of a Polysaccharide
Figure 6. Schematic drawing of likely conformations of β-D-xylose, the main constituent of the alga polysaccharide, between two mica surfaces: (a) adsorbed to a single mica surface and (b) adsorbed to both mica surfaces after being confined (highly compressed) and sheared.
logarithmic behavior. These results reinforce the idea that the actual sliding velocity is not a critical parameter in the lubrication process. Neither stick-slip nor damage was observed under these conditions. Shear-Induced Adhesion. We performed detailed adhesion measurements before, during, and after shearing the polysaccharide films (results not shown here) and found that shearing readily generates adhesion that increases from ∼0 mJ m-2 (pure repulsion before shearing) to 2-4 mJ m-2 after shearing back and forth only a few cycles even though the (simultaneously measured) friction force has not yet reached steady state. The adhesion energy then plateaus at W ≈ 6 mJ m-2 after prolonged shear (i.e., when the friction finally stabilizes at its steady-state value).
Discussion and Comparison with Other Biopolymers and Polyelectrolyte Systems Pressure and/or Shear-Induced Adhesion. An important finding is that pressing two surfaces with a polysaccharide film between them can significantly alter the adhesion. Figure 2 shows that applying pressure induces adhesion between surfaces that have been previously “protected” by a layer of polysaccharide (initially, no adhesion is measured as shown by the purely repulsive force profile of Figure 2b). The pressure of confinement apparently forces the polysaccharide molecules to bridge the two surfaces. In addition, we found that shearing also alters the adhesion in a similar way (i.e., the more the surfaces are sheared, the higher the adhesion). In solution, the sugars of the polysaccharide can adopt many conformations because of the highly flexible single bonds. We attribute the high adhesion, achieved by compressing or shearing the films between two mica surfaces, to H bonding between the negative oxygens on the opposing mica surfaces via the proton donors on different parts of the sugars, as illustrated in Figure 6b. The schematics illustrate the likely changes in the H-bonding network and ionic conformations of β-D-xylose arising from confinement and shear-induced reorientations (realignment) that can allow such a molecule to act either as a good protector of a single surface (Figure 6a) or as a good “bridging agent” between two surfaces (Figure 6b) under different conditions. Complementary Stable Low Friction. Although the thickness of the biopolymer film confined between the mica surfaces was 310 atm), where damage3 and/or high friction4,26 are measured, although cross-linked HA and lubricin appear to be good wear protectors (but with high friction coefficients). These findings suggest that some biolubricating systems, such as the eyelids, which are not subjected to high pressures, may be mediated by brushlike polyelectrolyte layers whereas other systems, such as in the knee joint, which are subjected to high pressure,27 may need to be mediated by more complex mechanisms. Note that none of the biopolymers mentioned above exhibit (i) steady low friction at (ii) high pressures (up to ∼110 atm) (iii) over large shearing distances and (iv) at both high and low sliding velocities while also (v) protecting the surfaces from wear. A subnanometer-thick monolayer of the highly mobile polysaccharide investigated here is enough to fulfill all five of these requirements. In addition, the damage reported to occur at 110 atm only during extending sliding (i.e., when shearing over twice the contact diameter) is not a serious limitation of our system: 11 MPa considerably exceeds most pressures encountered in biological systems and even typical peak pressures in synovial joints that are ca. 5 MPa. The above-mentioned properties are therefore ideally suited to the purpose for which the polysaccharide could be put to work in various applications. (24) Lin, Q.; Gourdon, D.; Sun, C.; Holten-Andersen, N.; Anderson, T. H.; Waite, J. H.; Israelachvili J. N. Proc. Natl. Acad. Sci. U.S.A. 2007, 104, 37823786. (25) Moro, T.; Takatori, Y.; Ishihara, K.; Konno, T.; Takigawa, Y.; Matsushita, T.; Chung, U. I.; Nakamura, K.; Kawaguchi, H. Nat. Mater. 2004, 3, 829-836. (26) Zappone, B.; Ruths, M.; Greene, G. W.; Jay, G. D.; Israelachvili, J. N. Biophys. J. 2007, 92, 1693-1708. (27) Wright, V. Lubrication and Wear in Joints; Harrison and Sons: London, 1969.
1540 Langmuir, Vol. 24, No. 4, 2008
Conclusions We have found that high pressure and/or shear plays an important role in reordering (reorienting) polysaccharide molecules between two closely opposed surfaces and that this has a large effect on the molecules’ ability to bind the two surfaces together (bridging), to protect them from damage, and to control their friction. In addition, we found that both significant adhesion to surfaces and a stable low-friction regime can be achieved simultaneously by the polysaccharide even under contact pressures exceeding 100 atmospheres and a 1000-fold range of sliding velocities. The lubricant molecules do not get squeezed out of the junction (joint) during sliding, ensuring very low energy dissipation during shearing. High mobility of the lubricant, when ordered on one surface, allows its transfer into new areas, providing low friction and wear resistance. Furthermore, the lubricant layer shows good reordering ability without signs for the removal of the molecules during the initial shear alignment. All of these properties are manifested by only a subnanometer-
Gourdon et al.
thick monolayer of polysaccharide, which is quite different from the thick brush layers currently thought to be the best biomimetic lubricants. By further investigating this class of polysaccharides and also other natural polyelectrolytes and synthetic peptides, we expect to be able to identify the crucial chemical groups and/or sequences that determine their different functionalities, with obvious potential for medical synthesis and biomimetic applications. Acknowledgment. This work was partially supported by the MRL Program of the National Science Foundation under award no. DMR00-80034, by the McCutchen Foundation, by the U.S.Israel Binational Science Foundation under grant no. 2006032, and by the Institute for Collaborative Biotechnologies through contract DAAD19-03-D-0004 from the U.S. Army Research Office. LA702259C
