Boundary lubricating properties of hydrophobically modified polyacrylamide

Abstract

Boundary lubrication has been studied in many associative bio-systems, including synovial fluids, phosphatidylcholine vesicles and mucus. However, comprehensive tribological investigations on synthetic associative polymers have not received sufficient attention. In this paper, we investigated the frictional behavior of polydimethylsiloxane (PDMS) rubber substrates lubricated with a universal and typical synthetic hydrophobically modified polyacrylamide (PAM) aqueous solution. Boundary lubrication was only observed at high concentrations above which intermolecular association was strong while adsorption mass and conformation were constant. Direct dilution with excessive water and replacement of concentrated solution with water both gave the PDMS the same boundary friction coefficient as that of lubrication only with water. Addition of surfactant that could disrupt the interchain association evidently destroyed the boundary lubrication of the hydrophobically modified PAM aqueous solution. Results revealed that interchain association, rather than the robust adsorption layer, plays a significant role in boundary lubrication of compliant PDMS–PDMS contact. This finding may provide new insight into the understanding of the boundary lubrication mechanism and developing novel boundary lubricants for artificial cartilage or consumer industry.

---

Introduction

Lubrication, in which the lubricating liquid is confined within the thin slit between two substrates, has been extensively studied for one and half a century.^1–6 Once locomotion occurs, lubricants could greatly reduce friction and minimize wear of the underlying substrates, which is highly relevant for engineering applications,⁷ biological apparatus,^8,9 consumer industry,¹⁰ etc. Lubrication behavior could be partitioned into three regions as sliding velocity or fluid film thickness increases, known as the Stribeck curve.^10,11 At high sliding velocity, a continuous and relatively thick fluid film of lubricant completely separates the opposite surfaces, named as the hydrodynamic region. In this region, the friction coefficient (μ) is directly proportional to the bulk viscosity and entrainment speed. When the sliding velocity is extremely low, the load is fully supported by the contacting asperities, which makes boundary lubrication particularly important for reducing wear. In the boundary region, μ is constant and mainly affected by the interface morphologies and interactions.¹¹ A mixed region is located between two regions wherein μ sharply decreases as sliding velocity increases.

In most cases, efficient boundary lubrication is crucial for the whole lubrication system. Over the past 20 years, efforts have been focused on several main objects of boundary lubrication in polymer field, containing synovial fluid as well as its components,^5,9,12 polymer brushes,^13–15 hydrogels^16–21 and so on.^22,23 Chen et al.¹³ showed that molecularly smooth mica surfaces in water bearing polyzwitterionic brushes that polymerized directly on the surface could exhibit extremely low μ values of 10⁻³ at physiological pressures, whereas the homopolymer solution could not provide such excellent lubricating properties. This extraordinary lubrication, particularly under boundary conditions, is attributed to the strong hydration and steric repulsion of the robustly attached polyzwitterionic brushes. Numerous researchers^8,9,12,24 have investigated the lubrication and structure of different mammalian synovial fluids. Israelachvili et al.¹² found that synovial components adsorbed on mica surfaces gradually form a homogenous gel layer and induce a strong repulsion between the substrate surfaces that prevents them from coming into contact under boundary conditions. As shearing (or sliding) continues, the gel layer gradually breaks up into discrete/individual gel particles that are extremely stable over time under drastic dilution conditions, and behave as roller bearings in the contact keeping the sheared surfaces far apart even under high compressive load. Meanwhile, Gong et al.^{16–18,20,25} proposed a repulsion–adsorption model¹⁷ to elucidate the rich and complex frictional behavior of polymer hydrogels. The low friction of hydrogel was attributed to the elastic deformation of the adsorbed polymer chain when the gel was adhesive to the substrate, and the hydrated layer of the gel surface when the gel was non-adhesive (repulsive).

Most researchers attributed the boundary lubrication to the hydration of the attached layer, either chemically or physically, which may result in a repulsive force due to different interactions.^12,13,15,26 Notably, most of the tribopairs investigated were molecularly smooth mica,^12,13,22 between which only the adsorbed or attached lubricant layer in nano scale remained under physiological pressures. However, the surfaces in vivo are usually not molecularly flat and have asperities of different scales, thus full direct contact could not reach even under extremely high loads.^18,27 In this case, the complicated rheological behavior of lubricants (synovial fluid, mucus, etc.) trapped in the cavities could not be ignored in the boundary region. Biolubricants in vivo usually are highly associative,^12,22,23 while the role of hydrophobic association remains unclear. To investigate whether generally used hydrophobically associative polymer solution could supply an efficient boundary lubrication, we synthesized hydrophobically modified polyacrylamide^28,29 by using butyl methacrylate as a hydrophobe. Basing on the determination of the rheological and tribological behavior of the copolymer aqueous solution, we attempted to figure out the relationship between boundary lubrication and hydrophobic association.

Experimental section

Materials

Acrylamide (AM), butyl methacrylate (BMA), polyoxyethylene sorbitan monolaurate (Tween 20) and hexadecyl trimethyl ammonium bromide (CTAB) from Aladdin (China) were used as received. 2,2′-Azobis(2-(2-imidazolin-2-yl)propane)dihydrochloride (VA-044) was purchased from J&K Scientific Ltd (China). Pyrene obtained from Fluka (USA) was recrystallized three times from absolute ethanol. The polydimethylsiloxane (PDMS) substrate used for the tribological test was prepared in a glass model with Sylgard® 184 (Dow Corning Corporation, USA) according to the manufacturer's instructions. Deionized water used for all experiments was supplied by a water purification system.

Synthesis

Polyacrylamide-butyl methacrylate (PAM-BMA) copolymer was synthesized through free-radical micellar copolymerization^28–30 of AM and BMA monomers. The molar fraction of BMA in the monomer feed was 3 mol%. The synthesis was performed in 1 M AM aqueous solution with 10 wt% Tween 20. A molar fraction of 0.05 mol% (to the monomers) VA-044 was used as initiator. The polymerization was allowed to proceed for 6 h at 50 °C under nitrogen flow and agitation. To remove the surfactants used in the polymerization, the product was repeatedly dissolved (in water) or precipitated (in ethanol) for at least three times. The as-obtained sample was then lyophilized and stored in a desiccator.

Element analysis

The C, H and N contents in the copolymer sample were determined with an element analysis instrument (EA1112, ThermoFinnigan Italia S.P.A), from which the hydrophobe content was calculated.

Light scattering measurements

The molecular weight of polymer samples was determined in formamide solution by static light scattering with a BI-200SM Laser Light Scattering Instruments (Brookhaven, USA) at 25 °C, with a vertically polarized incident light wavelength (λ) of 633 nm supplied by a He–Ne laser. The required specific refractive index increment dn/dc values were measured using a BI-DNDC differential refractometer. The polymer molecular weight was then calculated as Zimm plot, and the results are exhibited in Table 1.

Table 1 Characteristics of the polymers

	[BMA]^a (mol%)	Tween 20^b (wt%)	Yield (wt%)	[BMA]p^c (mol%)	MW^d (g mol^−1)
a Initial hydrophobe molar ratio in the monomer feed.b Surfactant concentration used in the synthesis.c Hydrophobe content in the copolymer.d Molecular weight determined by static light scattering.
PAM	—	—	90.9	—	2.98 × 10^7
PAM-BMA	3	10	88.6	2.77	4.69 × 10^6

---

Steady-state pyrene fluorescence measurement

PAM-BMA copolymer was dissolved in deionized water at a concentration range of 0.01–10 mg ml⁻¹. The copolymer solution was then subjected to steady-state pyrene fluorescence measurement with an LS-55 fluorescence spectrophotometer (Perkin-Elmer, USA) and a 10 mm path length quartz cell at room temperature. 1 mM pyrene in absolute methanol was prepared as stock solution, from which a certain amount of pyrene solution was then added into each sample with a micro injector achieving a final concentration of 6 × 10⁻⁷ M. After equilibration for 20 h with shaking, the sample was scanned and emission spectra (λ_ex = 335 nm) were obtained by averaging three scan results.

Rheological method

Rheological tests were carried out on a stress-controlled rheometer AR-G2 (TA Instruments, USA) equipped with a 40 mm cone plate geometry (cone angle 2°) at 25 °C. The sample was loaded on the rheometer and allowed to rest for 3 min without preshear before each specific rheological test. All oscillatory frequency sweep tests were conducted within the linear viscoelastic region determined by oscillatory strain sweep tests.

Tribological test

The friction coefficients of PDMS lubricated with PAM-BMA copolymer solutions were determined at 25 °C using an AR-G2 rotational rheometer (TA Instruments, USA). The rheometer was equipped with a plate geometry (radius R = 10 mm) and a Peltier plate. The plate geometry and Peltier plate were both glued with a PDMS rubber plate (compression modulus, 1.1 ± 0.1 MPa; thickness, 1 ± 0.05 mm; root mean square roughness, 4.61 × 10² nm; water contact angle, 120° ± 4°). For tribological measurements, approximately 0.5 ml of polymer aqueous solution was placed on the Peltier plate and then pressed using the geometry with a normal force (F_n) of 3 N (apparent pressure of 9.5 kPa). The rheometer drove the plate geometry to rotate in one direction with a rotating angle velocity (Ω) at torque (T). Friction coefficient (μ) was obtained according to μ = T/RF_n.³¹ To investigate the sliding velocity dependence of friction, Ω was changed stepwise from 0.1 to 10 rad s⁻¹ with every single velocity point lasting for 180 s, corresponding to the sliding velocity (v) of 10⁻³ to 10⁻¹ m s⁻¹, and the average torque of the last 150 s was adopted to calculate the friction coefficients.

Adsorption measurement

The adsorption of PAM-BMA solutions on hydrophobic gold surface was investigated using a quartz crystal microbalance (QCM, Resonant Probes GmbH, Goslar, Germany) at room temperature (20–25 °C). The QCM chip was coated with a monolayer of methyl-terminated thiol, according to the method in our previous study.²³ The average static water contact angle of the methyl-terminated gold surfaces was ca. 101° ± 3°, which indicates that the gold surface was hydrophobic.

For the QCM measurement, the hydrophobic gold-coated QCM chip with a fundamental resonant frequency (F₀) of 5 MHz was installed into the liquid cell. After initial equilibration of the surface at a deionic water flow rate of 50 μl min⁻¹, a stable baseline was established. Corresponding to the tribological test, the PAM-BMA solutions dissolved in water with concentrations of 0.5–50 mg ml⁻¹ was sequentially pumped into the liquid cell. Each injection step lasted for 30 min, followed by rinsing with deionic water to remove the loosely adsorbed polymers. The shifts of frequency and dissipation from the baseline were monitored.

Results and discussion

Association and adsorption of PAM-BMA copolymer

PAM-BMA copolymer was synthetized by free-radical micellar copolymerization.^28–30 In the micellar copolymerization, the hydropobes favor their incorporation as blocks randomly distributed along the backbone rather than as isolated units owing to their high local concentration in the micelles. Table 1 shows that the BMA content in copolymer is approximately 2.77 mol%, with slight difference from that in the feed. The molecular weight of PAM-BMA copolymer is 4.69 × 10⁶ g mol⁻¹, which is slightly lower than 2.98 × 10⁷ g mol⁻¹ of AM homopolymer.

Fluorescence spectra of pyrene are extremely sensitive to local environmental polarity and thus are usually used for exploring hydrophobic association in aqueous solutions.^28,32 If local environment changes to less polar media, the intensity ratio of the first (373 nm) to the third (384 nm) vibrational peak, i.e. I₁/I₃, will show a sharp drop. Fig. 1 shows the plot of I₁/I₃ in pyrene fluorescence spectra of PAM-BMA copolymer aqueous solution as a function of concentration (c). The curve shows the characteristic sigmoidal shape of three regions. When c is extremely low, pyrene senses the hydrophilic surrounding, and I₁/I₃ approaches the value recorded in the pure solvent. The decline of I₁/I₃ after a specific c (∼0.1 mg ml⁻¹) reflects the pyrene partition between the aqueous phase and the hydrophobic domains, i.e. hydrophobic interactions exist in the PAM-BMA aqueous solution at c >0.1 mg ml⁻¹.

Fig. 1 I₁/I₃ ratio of pyrene in solutions of PAM-BMA copolymer dissolved in deionized water at a series of concentration.

QCM was used to investigate the adsorption of PAM-BMA copolymer on hydrophobic substrate, i.e. methyl-terminated gold surface. Fig. 2 plots the frequency shift of the third overtone (−Δf₃/3) and the dissipation normalized by negative change in frequency (−ΔΓ₃/Δf₃) of the quartz crystal that adsorbed PAM-BMA after rinsing as functions of c. Generally, −Δf₃/3 is directly proportional to the sensed mass (m_f) according to Sauerbrey equation.³³ −ΔΓ₃/Δf₃ describes the conformation of adsorbed layer qualitatively³⁴ in which ΔΓ represents the energy loss (dissipation) of the quartz crystal caused by the presence of adsorbed layer.³⁴ As shown in Fig. 2, −Δf₃/3 increased with c, and the data could be well fitted by a Langmuir model (R-square 0.99), indicating that the mass of PAM-BMA molecules adsorbed onto the hydrophobic surface increased as c increases and followed the monolayer mechanism.^35,36 Comparatively, −ΔΓ₃/Δf₃ barely changed as c increases, implying that the concentration increase could not cause the conformation change of adsorbed macromolecules. That is, as bulk solution concentration increased, more individual macromolecules adsorbed onto the unoccupied surface sites and arranged according to almost the same conformation. Notably, when c > 10 mg ml⁻¹, surface sites were mostly occupied, and further increase of the bulk concentration could not lead to significant increase of adsorption mass or remarkable change of adsorption conformation.

Fig. 2 QCM data of methyl-terminated (hydrophobic) gold surface adsorbing PAM-BMA copolymer from aqueous solutions of different concentrations.

Rheological behavior of PAM-BMA copolymer aqueous solution

The PAM-BMA copolymer aqueous solution exhibits a classical steady shear behavior of viscoelastic polymer solution, i.e. a Newtonian plateau followed by a shear-thinning behavior above a critical shear rate (data not shown for brevity). Specific viscosity (η_sp) is shown in Fig. 3, which is derived from η_sp = (η₀ − η_s)/η_s, where η₀ is the zero shear viscosity of polymer solution and η_s is the solvent viscosity, i.e. water. η_sp increases linearly with c. A critical value exists at ca. 10 mg ml⁻¹, which is considered to be the critical overlap concentration (c*) of the copolymer solution. At c < c*, the slope of η_sp ∼ c curve is 1.50, and at c > c*, the slope is 3.88. The sharp increase of η_sp at c > c* indicates that both the entanglement of the molecular chain and intermolecular hydrophobic association greatly increase with c.

Fig. 3 Specific viscosity (η_sp) of PAM-BMA aqueous solution with various concentrations at 25 °C.

Fig. 4 depicts the steady and dynamic rheological behavior of 50 mg ml⁻¹ polyacrylamide and hydrophobically modified polyacrylamide, i.e. PAM-BMA aqueous solution, respectively. The apparent viscosity (η_a) of the two aqueous solutions at different shear rates was almost the same (the inset). This is probably because the effect from smaller molecular weight of PAM-BMA copolymer compromises with the viscosity enhancement from the hydrophobic association of the combined hydrophobic segments. Dynamic rheological test was conducted to distinguish further the difference between the two aqueous solutions; storage modulus (G′) and loss modulus (G′′) of the two solutions were investigated by oscillatory frequency (ω) sweep tests. The result G′′ > G′ over the measurement scope indicates that viscous properties dominates in both PAM and PAM-BMA copolymer aqueous solution. Interestingly, G′ of 50 mg ml⁻¹ PAM-BMA solution was 10-fold higher than that of PAM in the terminal zone, and the corresponding slopes of G′ for the PAM and PAM-BMA were 1.9 and 1.0, respectively. The former is consistent with typical nonassociative polymer solutions, and the latter indicates less sensitive to angular frequency, which implies the occurrence of hydrophobic association among PAM-BMA chains. Although the alkyl chain length of the incorporated hydrophobe (C4) was short compared with the typical hydrophobic monomers (C8–C12),²⁸ the fluorescence spectra demonstrated that hydrophobic association existed at high concentrations (c > 0.1 mg ml⁻¹).

Fig. 4 G′ and G′′ of 50 mg ml⁻¹ PAM and PAM-BMA dissolved in deionized water respectively, as functions of angular frequency during oscillatory frequency sweep tests. The inset depicts apparent viscosity of the both solutions.

Adding surfactant, particularly cationic, to hydrophobically modified polyacrylamide aqueous solution could decrease viscosity and dynamic modulus by disrupting the interchain liaisons.³⁷ Surfactant CTAB was used to destroy the hydrophobic association in PAM-BMA copolymer solution. Fig. 5 depicts the dynamic rheological behavior of 50 mg ml⁻¹ PAM-BMA copolymer dissolved in 0.02 M NaCl and 0.02 M CTAB aqueous solution, respectively. The result G′′ > G′ over the measured frequency scope prevailed in both solutions. Association in PAM-BMA was slightly influenced by the addition of 0.02 M NaCl compared with the result in deionized water (Fig. 4). However, the solution viscosity reduced in the presence of CTAB that possessed the same ionic strength with NaCl solution. Meanwhile, G′ and G′′ of PAM-BMA copolymer dissolved in 0.02 M CTAB solution were much lower than those in NaCl solution. This finding demonstrates that 0.02 M CTAB weakened the interchain hydrophobic association in PAM-BMA copolymer solution.

Fig. 5 G′ and G′′ of 50 mg ml⁻¹ PAM-BMA dissolved in 0.02 M NaCl (circle) and 0.02 M CTAB (triangle) respectively, as functions of angular frequency during oscillatory frequency sweep tests. The inset depicts apparent viscosity of the both solutions.

Boundary lubrication of PAM-BMA solution

The frictional behavior of PDMS substrates was investigated using PAM-BMA aqueous solution as lubricant. Fig. 6 presents the friction coefficient (μ) of PDMS substrates lubricated with water, 50 mg ml⁻¹ PAM and 50 mg ml⁻¹ PAM-BMA aqueous solutions respectively. All curves exhibit a typical Stribeck behavior, i.e. plateau at low sliding velocity (v, boundary region) and then a decrease with v increasing (mixed region). μ in the boundary region using PAM solution as lubricant was slightly lower than that of water. This is consistent with a previous study that non-associative polymers could not significantly reduce friction in the boundary region when an effective viscosity is considered.³⁸ However, the μ of PAM-BMA copolymer solution in the boundary region was only 0.16, which is much lower than that of water (1.16) and PAM solution (0.59). The μ of 0.16 was extremely low and made the transition between the boundary and mixed region unclear. This means that μ always maintains a low value over a relatively wide v scope. The viscosity of 50 mg ml⁻¹ PAM-BMA solution at different shear rates was almost the same as that of PAM solution as shown in Fig. 4, suggesting that the boundary lubrication function of PAM-BMA solution was not due to high viscosity. This is consistent with the results that high viscosity cannot enhance boundary lubrication of polymer solutions.^10,38 Moreover, lubrication with PAM and PAM-BMA solutions exhibited mixed region at lower v than water.

Fig. 6 Friction coefficients of PDMS lubricated with water (triangle), 50 mg ml⁻¹ PAM (square) and PAM-BMA (circle) dissolved in deionized water respectively.

The μ of PDMS substrates lubricated with PAM-BMA solutions at different concentrations were measured as functions of v to investigate the role of hydrophobic association in boundary lubrication (Fig. 7a). Each curve exhibited a typical Stribeck behavior with increasing v. All of these values were repeated more than three times (different experiments, data not shown). With increasing c, the μ in the boundary and mixed regions decreased gradually. Moreover, when c > 10 mg ml⁻¹, the mixed region moved toward low v, and the μ of the boundary region decreased more significantly. To elucidate the relationship between c and boundary lubrication better, the μ of the boundary region was obtained by averaging the data during friction tests at v of 10⁻³ m s⁻¹ for 10 min (Fig. 7b). The μ of the boundary region initially decreased slowly and then dropped sharply with increasing c. A critical c of approximately 10 mg ml⁻¹ existed; at c < 10 mg ml⁻¹, μ decreased slightly with c, and at c > 10 mg ml⁻¹, μ decreased significantly with c and μ ∼ c^−1.2 was found.

Fig. 7 (a) Friction coefficients of PDMS as functions of velocity, lubricated with PAM-BMA dissolved in deionized water at a series of concentration; (b) concentration dependence of boundary friction coefficients of PAM-BMA dissolved in deionized water. The friction tests were performed between PDMS substrates respectively at a rotating velocity of 10⁻³ m s⁻¹ (boundary region) for 10 min, and friction coefficients were derived by averaging data from each procedure.

As shown in Fig. 8, after loading 50 mg ml⁻¹ PAM-BMA solution, friction tests were performed at a v of 10⁻³ m s⁻¹ (boundary region, Ω = 0.1 rad s⁻¹, rotational period 2π/Ω ≈ 62.8 s) as the number of rotating circle (n) increased. Excessive water was added to surround the friction pair without changing the sample at the 10th circle, and the friction test was carried out continuously. By contrast, the μ of another PDMS lubricated with water under the same condition was also determined separately. Each datum plotted in the figure is derived by averaging the data during the corresponding rotating circles. PDMS lubricated with PAM-BMA solution had a μ of 0.16, which is relatively low compared with that of water (1.07). However, after dilution, μ gradually enhanced to an equilibrium value, which ultimately almost equals to that of water. The copolymer molecules non-adsorbed or loosely adsorbed onto the substrate were probably washed away as the substrates rub against each other. Only the tightly adsorbed copolymer chains were reserved, indicating that tightly adsorbed layer could not efficiently improve the boundary lubrication.

Fig. 8 Friction coefficients hydrophobic PDMS lubricated by water and 50 mg ml⁻¹ PAM-BMA aqueous solution at a rotating velocity of 0.1 rad s⁻¹ (∼10⁻³ m s⁻¹) as rotating circle increase. Excessive water was directly added surrounding the tribopair at the 10th circle (arrow pointed).

In addition, the μ of PDMS surfaces lubricated by 50 mg ml⁻¹ PAM-BMA copolymer dissolved in 0.02 M CTAB and 0.02 M NaCl are shown in Fig. 9. NaCl solution exhibited no ability of boundary lubrication, whereas CTAB aqueous solution showed a minor ability. For PAM-BMA dissolved in NaCl as lubricant (similar to Fig. 6), the μ of the boundary region was minimized to an extremely low value (0.12). While dissolved in 0.02 M CTAB, the μ of the boundary region increased to 0.33, even close to that of the CTAB solution. Taking Fig. 5 into account, disrupting interchain hydrophobic association by CTAB clearly impaired the boundary lubrication properties of PAM-BMA solution.

Fig. 9 Friction coefficients of PDMS lubricated by 50 mg ml⁻¹ PAM-BMA dissolved in 0.02 M NaCl (circle) and 0.02 M CTAB (triangle) respectively.

Role of hydrophobic association in boundary lubrication

Universal elucidation of the specific mechanism of boundary lubrication has not been achieved because of various tribopairs, different contact states as well as complicated mutual effect between lubricant–lubricant and lubricant–substrate. Besides the surface asperities, most studies attributed the boundary lubricating properties to the strong hydration of the physical adsorption or chemical attached layer, which could induce a repulsion force between the sliding surfaces because of the hydration layer and other different mechanisms. For example, Jahn et al.³⁹ proposed a hydration lubrication mechanism. In the proposed mechanism, the robustly bounded water molecules that surround the tightly attached or adsorbed layer, including polymer brushes, proteins, phospholipid bilayers and liposomes, could sustain high load while exhibiting a weak frictional stress to shear, which plays a key role in boundary lubrication. The hydration lubrication mechanism seems not work in macro scale. Poly(L-lysine)-g-poly(ethylene glycol) brushes could efficiently reduce μ of molecularly smooth mica surface from 0.3 to 10⁻³, but obviously higher μ (from 0.25 to 0.12) in a macroscopic tribometric experiment.^40,41

The boundary lubrication behavior of PAM-BMA solutions in the present paper is quite different from the reported. During the tribological test of 50 mg ml⁻¹ PAM-BMA solution, direct dilution with excessive water or replacement of solution with water both gave the PDMS the same boundary friction coefficient as that of lubrication only with water. In low concentration region (c < 10 mg ml⁻¹), μ changed slightly (Fig. 7b), while the adsorption mass of PAM-BMA on the hydrophobic surface increased quickly (Fig. 3). In relatively high concentration (c > 10 mg ml⁻¹) region, μ declined sharply with increasing c, though the adsorption mass did not increase significantly and the conformation of PAM-BMA adsorbed on the hydrophobic surface was also not changed. These results indicate that the robust adsorption layer of PAM-BMA copolymer cannot enhance boundary lubrication. This is sharply in contrast to the commonly accepted theory that boundary lubrication results from tightly adsorbed layer.^12,13,39 In addition, frictional behavior of the associative PAM-BMA solution differed from that of nonassociative PAM solution. PAM solution with high concentration (50 mg ml⁻¹) could reduce friction in mixed region because of higher viscosity, but only a slight boundary lubrication was observed. However, associative PAM-BMA solution (50 mg ml⁻¹), which has the same viscosity as PAM homopolymer solution but much higher G′ under low frequency due to hydrophobic association, could reduce boundary friction by 10-fold. Notably, the boundary lubrication of PAM-BMA copolymer solutions was observed only at concentrations above which interchain associations are enhanced (c > 10 mg ml⁻¹), despite the adsorption mass nearly approached equilibrium and conformation remained constant. Destroying the interchain association by CTAB largely impaired the boundary-lubricating ability of PAM-BMA solution. Given the above results, intermolecular association of PAM-BMA copolymer solution, rather than the robust adsorption layer, is speculated to play a significant role in boundary lubrication between two soft PDMS surfaces. This finding is comparable with our previous result that a natural complex polymer, mucin, can improve boundary lubrication due to intermolecular association rather than tightly adsorbed layer.²³ To the best of our knowledge, PAM-BMA and mucin are the first two systems that highlight the role of intermolecular association of polymer solution in boundary lubrication. Mucin and PAM-BMA possess highly different molecular structures, but their boundary lubricating behaviors are similar, indicating that enhancement of boundary lubrication by intermolecular association may be universal in hydrophilic polymer solution.

During our tribology test system, as two PDMS plates sliding or rotating against each other, microscopic and macroscopic fluctuations of the surface always existed due to surface roughness, contact area and contact state; thus, full contact could not be reached, and the copolymer solution was confined between the soft interfaces. Likewise, Gong et al. revealed that a macro-scale of water is trapped at the soft tribological contact between hydrogel and glass even under boundary conditions by direct interfacial observation.¹⁸ This finding implies that the rheological behavior of the liquid confined in the asperities is significantly related to the boundary lubricating property. The PAM-BMA copolymer solution behave a higher elasticity than PAM homopolymer solution due to hydrophobic associations (relatively high G′ in low frequency), which could sustain large load and thus favor reduction of boundary friction coefficients. In particular at high concentrations, size and quantity of hydrophobic domain rapidly increased with concentration, resulting in a sharp linear decrease of friction coefficients in the boundary region.

PAM is generally used as hydrophilic polymer, and PAM-BMA is obtained via micellar copolymerization with a hydrophobic monomer BMA; no other specific interactions existed except hydrophobic association. The boundary lubrication of associative PAM-BMA solutions reported in this study is easily accessible and represented, which may provide new instructions in developing water soluble boundary lubricants.

Conclusion

In summary, boundary lubrication was only found in associative PAM-BMA copolymer aqueous solution and highly related to concentration. At low concentration, boundary friction μ was almost constant, while the adsorption mass of PAM-BMA on the hydrophobic surface increased quickly with concentration. However, at high concentration, further increase of concentration could not significantly increase the adsorption amount or change the conformation of PAM-BMA on the hydrophobic surface, while the boundary friction μ declined sharply. This implies that only the tightly adsorbed layer on the substrate surface could not result in boundary lubrication. Boundary lubrication only prevailed in high concentrations with interchain associations and followed μ ∼ c^−1.2. Destroying the interchain associations by CTAB impaired largely the boundary-lubricating ability of PAM-BMA solution. Intermolecular association rather than the robust adsorption layer played a significant role in boundary lubrication. Considering that most natural lubricating systems are highly associative, this result may be useful in understanding the boundary lubrication mechanism and developing novel boundary lubricants for artificial cartilage or consumer industry.^8–10

Acknowledgements

We gratefully acknowledge Mr JianYuan Li for the QCM measurement. This research is supported by Zhejiang Provincial Natural Science Foundation of China (LY16B040002) and the National Natural Science Foundation of China (Grant 51173164).

References

1. J. Gao, W. D. Luedtke, D. Gourdon, M. Ruths, J. N. Israelachvili and U. Landman, J. Phys. Chem., 2004, 108, 3410–3425.
2. H. Yoshizawa, Y. L. Chen and J. N. Israelachvili, J. Phys. Chem., 1993, 97, 4128–4140.
3. W. H. Briscoe, S. Titmuss, F. Tiberg, R. K. Thomas, D. J. McGillivray and J. Klein, Nature, 2006, 444, 191–194.
4. G. Luengo, J. N. lsraelachvili and S. Granick, Wear, 1996, 200, 328–335.
5. T. A. Schmidt, N. S. Gastelum, Q. T. Nguyen, B. L. Schumacher and R. L. Sah, Arthritis Rheum., 2007, 56, 882–891.
6. G. W. Greene, A. Olszewska, M. Osterberg, H. Zhu and R. Horn, Soft Matter, 2014, 10, 374–382.
7. A. Erdemir, Tribol. Int., 2005, 38, 249–256.
8. M. P. Heuberger, M. R. Widmer, E. Zobeley, R. Glockshuber and N. D. Spencer, Biomaterials, 2005, 26, 1165–1173.
9. T. A. Schmidt and R. L. Sah, Osteoarthritis Cartilage, 2007, 15, 35–47.
10. J. de Vicente, J. R. Stokes and H. A. Spikes, Food Hydrocolloids, 2006, 20, 483–491.
11. W. Z. Wang, S. Wang, F. H. Shi, Y. C. Wang, H. B. Chen, H. Wang and Y. Z. Hu, J. Tribol., 2007, 129, 495–501.
12. X. Banquy, D. W. Lee, S. Das, J. Hogan and J. N. Israelachvili, Adv. Funct. Mater., 2014, 24, 3152–3161.
13. M. Chen, W. H. Briscoe, S. P. Armes and J. Klein, Science, 2009, 323, 1698–1701.
14. J. Klein, E. Kumacheva, D. Mahalu, D. Perahla and L. J. Fetters, Letters to Nature, 1994, 370, 634–636.
15. X. Banquy, J. Burdynska, D. W. Lee, K. Matyjaszewski and J. N. Israelachvili, J. Am. Chem. Soc., 2014, 136, 6199–6202.
16. K. Kamada, H. Furukawa, T. Kurokawa, T. Tada, T. Tominaga, Y. Nakano and J. P. Gong, J. Phys.: Condens. Matter, 2011, 23, 284107.
17. J. P. Gong, Soft Matter, 2006, 2, 544–552.
18. T. Yamamoto, T. Kurokawa, J. Ahmed, G. Kamita, S. Yashima, Y. Furukawa, Y. Ota, H. Furukawa and J. P. Gong, Soft Matter, 2014, 10, 5589–5596.
19. M. Du, Y. Zhang, Y. H. Song and Q. Zheng, J. Polym. Sci., Part B: Polym. Phys., 2014, 52, 765–772.
20. S. Yashima, N. Takase, T. Kurokawa and J. P. Gong, Soft Matter, 2014, 10, 3192–3199.
21. M. Du, Y. Maki, T. Tominaga, H. Furukawa, J. P. Gong, Y. Osada and Q. Zheng, Macromolecules, 2007, 40, 4313–4321.
22. R. Sorkin, N. Kampf, Y. Dror, E. Shimoni and J. Klein, Biomaterials, 2013, 34, 5465–5475.
23. X. Wang, M. Du, H. P. Han, Y. H. Song and Q. Zheng, Langmuir, 2015, 31, 4733–4740.
24. S. Ghosh, D. Choudhury, N. S. Das and B. P. Murphy, Lubr. Sci., 2014, 26, 387–410.
25. J. Ahmed, H. L. Guo, T. Yamamoto, T. Kurokawa, M. Takahata, T. Nakajima and J. P. Gong, Macromolecules, 2014, 47, 3101–3107.
26. J. Seror, R. Sorkin and J. Klein, Polym. Adv. Technol., 2014, 25, 468–477.
27. J. H. H. Bongaerts, K. Fourtouni and J. R. Stokes, Tribol. Int., 2007, 40, 1531–1542.
28. F. Candau and J. Selb, Adv. Colloid Interface Sci., 1999, 79, 149–172.
29. A. Hill, F. Candau and J. Selb, Prog. Colloid Polym. Sci., 1991, 84, 61–65.
30. E. Volpert, J. Selb and F. Candau, Macromolecules, 1996, 29, 1452–1463.
31. M. Wathier, B. A. Lakin, P. N. Bansal, S. S. Stoddart, B. D. Snyder and M. W. Grinstaff, J. Am. Chem. Soc., 2013, 135, 4930–4933.
32. M. Li, M. Jiang, Y. X. Zhang and Q. Fang, Macromolecules, 1996, 30, 470–478.
33. Y. P. Lu, X. H. Zhang, Z. Q. Fan and B. Y. Du, Polymer, 2012, 53, 3791–3801.
34. M. A. Plunkett, Z. H. Wang, M. W. Rutland and D. Johannsmann, Langmuir, 2003, 19, 6837–6844.
35. C. Tamerler, E. E. Oren, M. Duman, E. Venkatasubramanian and M. Sarikaya, Langmuir, 2006, 22, 7712–7718.
36. S. H. Brewer, W. R. Glomm, M. C. Johnson, M. K. Knag and S. Franzen, Langmuir, 2005, 21, 9303–9307.
37. E. Jimenez-Regalado, J. Selb and F. Candau, Langmuir, 2000, 16, 8611–8621.
38. J. de Vicente, J. R. Stokes and H. A. Spikes, Tribol. Int., 2005, 38, 515–526.
39. S. Jahn and J. Klein, Macromolecules, 2015, 48, 5059–5075.
40. S. Lee, M. Müller, M. Ratoi-Salagean, J. Vörös, S. Pasche, S. M. D. Paul, H. A. Spikes, M. Textor and N. D. Spencer, Tribol. Lett., 2003, 15, 231–239.
41. T. Drobek and N. D. Spencer, Langmuir, 2008, 24, 1484–1488.

---