Introduction

When liquid molecules are confined in a nanometer-scale gap between two solid surfaces, their dynamic properties are extremely different from their properties in the bulk. One of the typical examples of this difference is the slowing down of the molecular motions of liquid in confined geometries; the effective viscosity increases by many orders of magnitudes [1,2,3,4]. The increase in the effective viscosity of confined liquids is a key issue in tribology. Lubricant oils commonly used in industry generally have low viscosity because oil viscosity directly determines the friction force between sliding surfaces at low load/high sliding velocity conditions (hydrodynamic lubrication). However, such low-viscosity lubricant oils are expelled from the sliding interface easily under high load/low sliding velocity conditions (boundary lubrication), and the thickness of the lubricant films reaches molecular dimensions. In this case, the intervening lubricant liquids no longer behave as bulk-like liquids but behave as high-viscosity liquids or essentially as solids, resulting in high friction. Therefore, obtaining a low-friction interface using low-viscosity lubricant oils both in hydrodynamic and boundary lubrication is one of the most important issues for oil-based lubrication.

Adding suitable friction modifiers to low-viscosity lubricant oils is an effective solution to this problem, and various friction modifier additives have been developed and are used in industrial systems [5]. Representative additives can be divided into three categories: organic friction modifiers [5,6,7,8,9,10,11], soluble organo-molybdenum additives [5, 12, 13], and functionalized polymers [5, 14,15,16,17]. In this study, we focus on functionalized polymers.

Early application of polymeric materials as lubricant additives started more than half a century ago [14, 18, 19]; polymers without functional groups were dissolved in lubricant base oils. These polymers are referred to as “viscosity index improver”, and their main function is to reduce the temperature dependence of the bulk viscosity of base oils (generally, oil viscosity decreases with temperature). Recently, functionalized polymers have attracted attention because sliding surfaces are expected to be covered by adsorbed polymers that should prevent the decrease in lubricant film thickness, particularly in boundary lubrication. Low-friction systems using surface-adsorbed polymer layers in solutions have been extensively studied for water-based lubrication systems because these systems are important models for biological lubrication [20,21,22,23]. For oil-based lubrication systems, macroscopic friction properties of block copolymer additives and their effectiveness in reducing friction in boundary lubrication have been reported [14,15,16,17]. However, the relationship between the molecular structures of the functionalized polymers, confined/sliding film structures between surfaces, and molecular mechanisms of friction have not been studied in detail.

In our previous work [24], the friction properties of a dilute solution of a triblock copolymer, poly(11-acrylamidoundecanoic acid)-block-poly(stearyl methacrylate)-block-poly(11-acrylamidoundecanoic acid) (henceforth A5S992A5, see Fig. 1 and Table 1) in a lubricant base oil (poly-α-olefin, PAO) were investigated. The A5S992A5/PAO solution was injected and molecularly confined between smooth mica surfaces, and the friction was measured as a function of applied load and sliding velocity using the surface forces apparatus (SFA). The results showed a gradual decrease in the adsorbed layer thickness and friction with time after surface preparation (sample injection between mica surfaces), and 2 days after preparation, the confined film exhibited an excellent friction reduction ability (the friction coefficient μ was well below 0.01, which is one or two orders of magnitude lower than that of PAO without the additive). The time-dependent adsorbed structures and extremely low-friction mechanism were discussed from the viewpoint of the relatively low affinity of A5S992A5 to PAO and the shrunken adsorbed conformations in the confined films.

Fig. 1
figure 1

Molecular structure of block copolymer additives: n = 11 (m = 946) and 17 (m = 992) were A5L946A5 and A5S992A5, respectively

Full size image
Table 1 Composition and molecular weight data of the block copolymer additives
Full size table

In the present study, a triblock copolymer with a molecular structure similar to that of A5S992A5, poly(11-acrylamidoundecanoic acid)-block-poly(lauryl methacrylate)-block-poly(11-acrylamidoundecanoic acid) (henceforth A5L946A5, see Fig. 1 and Table 1) was examined using the SFA, and the results were compared to those of the confined PAO (without additive) and A5S992A5/PAO systems. The μ value of the dilute solution of A5L946A5 in PAO (A5L946A5/PAO system) was approximately 0.5, which was larger than the μ of PAO without the additive (between 0.04 and 0.47 [11, 24]). Even though the alkyl chain length in the poly(alkyl methacrylate) blocks is the only structural difference between the two additives (A5L946A5 and A5S992A5), the friction reduction abilities of the two systems were very different (a difference of more than two orders of magnitude in μ). The relationship between the additive molecular structures (and the resulting affinity to the base oil PAO), confined/sliding film structures between surfaces, and friction properties were investigated in detail, shedding new light on the design of low-friction surfaces with block copolymer additives in oil-based lubrication.

Experimental section

Materials

Poly(11-acrylamidoundecanoic acid) (PAaU) was synthesized according to the methods reported previously [24]. PAaU can be used as a bifunctional macro chain transfer agent (CTA). Number-average molecular weight (Mn (GPC)) and molecular weight distribution (Mw/Mn) for PAaU were 3.77 × 103 and 1.05, respectively, as estimated by gel-permeation chromatography (GPC). The number-average degree of polymerization (DP) and number-average molecular weight (Mn (NMR)) were 10 and 2.84 × 103, respectively, as estimated using 1H NMR. Lauryl methacrylate (LMA, >95.0%) was purchased from Wako Pure Chemical Industries and purified using a disposable inhibitor remover column (Sigma-Aldrich). 2,2′-Azobis(isobutyronitrile) (AIBN, 98%) was purchased from Wako Pure Chemical Industries, and was recrystallized from methanol. Tetrahydrofuran (THF) and ethanol were dried using molecular sieves 4A and distilled. Water was purified using a Millipore Milli-Q system. PAO (hydrogenated oligomers of 1-decene, industrial grade) was used as a base lubricant oil, and had a bulk viscosity of 76 mPa s.

Synthesis of ABA triblock copolymer (A5L946A5)

PAaU (24.4 mg, 8.60 μmol), LMA (2.71 g, 10.7 mmol), and AIBN (0.57 mg, 3.47 μmol) were dissolved in a mixture of THF (8.6 mL) and ethanol (2.2 mL). The solution was deoxygenated by Ar bubbling for 0.5 h. Polymerization was carried out for 50 h at 60 °C. After the polymerization, using 1H NMR, the conversion was estimated to be 76.4%. The polymer solution was poured into a large excess of acetone to precipitate the polymer. Reprecipitation was performed three times. The triblock copolymer (A5L946A5) was dried at 30 °C under reduced pressure for 24 h (1.71 g, 63.1%). The preparation route of A5L946A5 is shown in Scheme S1 (Supplementary Information). Mn (GPC) and Mw/Mn of A5L946A5 estimated by GPC were 1.81 × 105 and 1.62, respectively (Figure S1, Supplementary Information). Mn (theory) for A5L946A5 and DP for the poly(lauryl methacrylate) (PLMA) block estimated from conversion were 2.44 × 105 and 946, respectively. 1H NMR signals for the PAaU block cannot be detected, because DP of PAaU was too small compared to that of PLMA (Figure S2, Supplementary Information). Figure 1 and Table 1 present the molecular structure and the important molecular parameters, respectively.

NMR and GPC measurements

1H NMR measurements were performed using a Bruker DRX-500 spectrometer. GPC measurements were carried out using three Shodex KF-803L columns and a Shodex KF-805L column equipped with a refractive index (RI) detector at 40 °C. THF was used as the eluent with a flow rate of 1.0 mL/min. Mn (GPC) and Mw/Mn were calibrated using standard polystyrene samples.

Preparation of additive solution

The block copolymer additive A5L946A5 was dissolved in PAO (additive concentration: 0.5 wt%) by stirring with a magnetic stirrer for 3 to 5 h at 45 °C. The A5L946A5/PAO solution was clear and stable at room temperature (23 °C).

Friction measurements using the SFA

Friction measurements were performed for the smooth mica surfaces separated by the thin films of A5L946A5/PAO solution using the SFA RSM-1 (Advance Riko, Inc., Japan) [25] equipped with shear attachments. Two opposed mica surfaces were installed into the SFA chamber in a crossed cylinder configuration, and a droplet of the sample solution (volume of approximately 0.1 ml) was injected between the surfaces. The friction experiment was performed using a home-built bimorph slider [24, 26]. The lower mica surface was moved horizontally at a constant sliding velocity V, that was varied within the range of 0.0056–5.6 μm/s. The maximum sliding distance was 14 μm. The resonance shear unit [27] of the RSM-1 was used to measure the friction force F between the surfaces. The upper mica surface was supported by a friction measuring spring and the deflection of the spring was detected by a capacitance probe. The applied load L was varied up to 20 mN, which was controlled by a normal force spring (spring constant k = 1810 N/m) of the bimorph slider. The thickness of the confined film and the contact geometry were evaluated by an optical technique called fringes of equal chromatic order (FECO) [28]. FECO observation enables us to measure the size of the contact area A during friction measurements and the film thickness D with the accuracy of 0.2 nm. Figure 2 schematically illustrates the contact interface and friction measurement using the SFA. The atmosphere in the SFA chamber was kept in completely dry conditions, and some P2O5 was placed in the sealed chamber. The experimental room was kept at a fixed temperature of 23 °C ± 1 °C

Fig. 2
figure 2

Schematic illustration describing the contact interface in the SFA friction measurement. Smooth mica surfaces are separated by the molecularly thin films of the A5L946A5/PAO solution. The contact interface is compressed under an applied load L. The lower surface is moved laterally at a constant sliding velocity V. The upper surface is supported by a friction measuring spring and friction force F is measured by the deflection of the spring. Simultaneous observation of FECO fringes enables us to measure the film thickness D and real contact area A during the friction measurement

Full size image

Results

Contact geometry and dynamic thickness

The contact geometry and dynamic thickness (thickness during sliding) of the A5L946A5/PAO system continuously changed with repetitive back and forth sliding motions and applied load, which were monitored by the observation of the FECO fringes. A typical example of the FECO images is shown in Fig. 3. The fringe shape, which directly reflects the contact geometry, was rounded, and the film thickness was minimum (21.6 nm) at the center of the contact area before the sliding motion was applied (static hard-wall state, Fig. 3(i)). This fringe shape indicates the Hertz contact [1] observed for a soft and deformable interface; the contact pressure is the maximum at the center of the contact area. When lateral sliding motions were applied, whole fringes moved to longer wavelengths (upward direction in Fig. 3), which suggests an increase in thickness. As shown in Fig. 3(ii), the fringes had a small bump near the center of the contact area, where the confined film thickness was the maximum. Increasing the applied load and repetitive sliding motions further increased the thickness, particularly at (or near) the center of the contact area, and a large deformation of the fringe shape was observed. At L = 15 mN (Fig. 3(iii)), the fringes had a concave shape, and the thickness at the center was 202.3 nm, while the thickness at the edge was 127.4 nm. A further increase in L (=20.4 mN, Fig. 3(iv)) increased the difference in the thickness between the maximum (near the center) and minimum (at the edge) of the contact area (also see Fig. 4). We note that the scale ratio between the contact diameter (horizontal direction) and film thickness (vertical direction) in Fig. 3 is of the order of 1000:1 (see the scale bars in Fig. 3(iv)).

Fig. 3
figure 3

Typical FECO fringes of the A5L946A5/PAO system during a sliding experiment observed in the SFA. The fringe position in the vertical direction (measure of the thickness) and fringe shape (reflection of the contact geometry) changed with repetitive back and forth sliding motions and the increase of L. The static hard-wall state (L = 2.6 mN, before sliding (i)) had a rounded fringe shape. The fringe position shifted to longer wavelength (upward direction in the figure) as soon as sliding was commenced (ii), indicating an increase in the dynamic thickness. Repetitive sliding motions and increasing L increased the thickness particularly at (or near) the center of the contact area, while the fringes deformed to a concave shape ((iii) and (iv))

Full size image
Fig. 4
figure 4

Minimum thickness Dmin (thickness at the edge of the contact area) and maximum thickness Dmax (thickness at or near the center of the contact area) during sliding as a function of the applied load L for the A5L946A5/PAO system. The static hard-wall thickness (thickness before sliding was applied, shown by the arrow) was 21.6 nm. When sliding motions were applied, the thickness increased particularly at (or near) the center of the contact area. The Dmin did not exhibit an apparent load dependence, but the Dmax gradually increased with increasing L

Full size image

Figure 4 shows the dynamic thickness of the A5L946A5/PAO system as a function of the applied load. Because the contact geometry was altered along with the load and repetitive sliding motions as shown in Fig. 3, the maximum thickness Dmax (obtained at or near the center of the contact area) and the minimum thickness Dmin (obtained at the edge of the contact area) were plotted. The static hard-wall thickness (thickness before the sliding motions were applied) was 21.6 nm (shown by the arrow, corresponds to the fringe shown in Fig. 3(i)). When sliding was commenced, the thickness immediately increased and then the deformation of the contact interface gradually proceeded with repetitive sliding motions. Both thicknesses (Dmax and Dmin) were unstable and the data at each applied load condition were rather scattered. Multiple data points at a given applied load represent the variation in the thickness with the repetitive sliding motions and sliding velocities under the fixed load conditions. Although a precise evaluation was difficult due to the scatter, it was observed that Dmax gradually increased with the increase in the applied load and repetitive sliding motions. On the other hand, little applied load dependence was observed for Dmin.

Friction forces

The relationship between the kinetic friction force and sliding velocity for the A5L946A5/PAO system is shown in Fig. 5. Friction force increased with the increase of applied load, and increased slightly with the increase of sliding velocity.

Fig. 5
figure 5

Kinetic friction force as a function of the sliding velocity at the different applied load conditions for the A5L946A5/PAO system

Full size image

Figure 6 shows the kinetic friction force as a function of the applied load for the A5L946A5/PAO system at two sliding velocities. Friction coefficient μ was obtained from the slope of the linear fitting to the data (assuming Amontons’ law of friction [1]), and was 0.54 for V = 0.011 μm/s and 0.56 for V = 1.1 μm/s.

Fig. 6
figure 6

Kinetic friction force as a function of the applied load for the A5L946A5/PAO system at two different sliding velocities. Friction coefficient μ was obtained from the slope of the linear fitting (see Table 2)

Full size image

In our previous study of the A5S992A5/PAO system [24], we reported that the contact geometry and dynamic thickness varied with time after surface preparation (sample injection between mica surfaces). The resulting friction properties also varied with time, and an extremely low-friction surface (with μ well below 0.01) was obtained 2 days after surface preparation. However, we did not observe such a time dependence of the structures/properties for the A5L946A5/PAO system studied here. The results of the time effects on the dynamic thickness and friction properties are briefly summarized in Figures S3 and S4 in the Supplementary Information.

Discussion

Confined film structure and contact geometry

The static hard-wall thickness of the A5L946A5/PAO system was 21.6 nm (Figs. 3 and 4), which was much larger than that of the confined PAO without the additives measured in our previous study (1.4 nm or less [11, 24]). The A5L946A5 molecules directly attached to the mica surface constitute an adsorbed layer due to the adsorption mainly by the carboxylic acid groups in the molecule [6, 7, 11, 24]. There may be some non-adsorbed A5L946A5 molecules (molecules not directly attached to the mica surfaces) in the hard-wall film because a chain entanglement effect is expected between the adsorbed and non-adsorbed A5L946A5 molecules that could prevent the non-adsorbed molecules from being completely squeezed out by the normal compression alone. In this measurement, we typically took approximately 2 h to approach the opposite mica surfaces into a static hard-wall state from the surface distance of approximately a few microns, and a faster approach generally gave a larger hard-wall thickness [1, 29, 30]. The hydrodynamic radius Rh of the A5L946A5 molecule in the bulk PAO solution was evaluated from the dynamic light scattering analysis performed at 25 °C (data not shown) and was approximately 25 nm. Therefore, the A5L946A5 molecules in the hard-wall film should have a highly compressed (flattened) conformation.

Repetitive back and forth sliding motions and increasing applied pressure increased the dynamic thickness, as shown in Figs. 3 and 4. Changes in the liquid film thickness by sliding motions and/or applied load are often observed for a variety of confined liquid systems in SFA friction measurements. However, in most cases, the thickness does not increase but rather decreases due to the squeezing of molecules from the sliding interface and/or the shear-induced conformational rearrangements of the intervening molecules [29, 31, 32]. Figure 7 shows the schematic illustration of the “sliding front” of the contact area to explain the possible scenario of the thickness increase. Prior to the application of the sliding motions, the contact shape is rounded (the contact pressure is the highest at the center of the contact area for a Hertzian contact, see Fig. 3i) and the minimum thickness is found at the center. Therefore, the A5L946A5 molecules at the center are the most compressed and flattened, and the surrounding molecules should have more expanded conformations. When sliding motions are applied, the A5L946A5 molecules at the “sliding front” of the contact area (having expanded conformations) newly enter into the area, resulting in the increase in the sliding film thickness.

Fig. 7
figure 7

Schematic illustrations of the contact geometry (a) and the “sliding front” of the contact area (b) for the A5L946A5/PAO system. Typical values of the contact geometry (Dmax and Dmin) and sliding distance are included in a. The A5L946A5 molecules confined in the contact area are highly compressed and flattened because of the applied pressure, whereas the A5L946A5 molecules at the “sliding front” have more expanded conformations. An increase of the thickness upon shearing may occur due to the entry of new expanded (non-adsorbed) A5L946A5 molecules at the sliding front into the contact area. Accumulation of the newly trapped molecules from both sliding directions leads to a concave contact geometry. Note that the scales of the thickness and the contact diameter differ by approximately three orders of magnitude

Full size image

In this study, repetitive back and forth sliding motions were applied, which induced the entry of new A5L946A5 molecules from both sliding directions. The maximum sliding distance in our experimental setting was 14 μm as mentioned above, which was smaller than the contact diameter (typical diameter was approximately 50 μm, see Figs. 3 and 7). Therefore, repetitive back and forth motions gradually gathered the newly entered (non-adsorbed) A5L946A5 molecules into the contact area, and then these molecules were accumulated and kinetically trapped by the high normal pressure at or near the center of the contact area. This is a plausible mechanism of the formation of the concave contact shape during sliding.

Friction properties of the A5L946A5/PAO system

Table 2 lists the friction coefficients μ of the A5L946A5/PAO system; the results of the confined PAO (without additives) [11, 24] and the A5S992A5/PAO system [24] are also included for comparison. The μ of the A5L946A5/PAO system was in the range of 0.54–0.56 (Fig. 6), which is larger than that of the confined PAO. On the other hand, the μ of the A5S992A5/PAO system indicates an excellent friction reduction ability. The two block copolymer additives compared here have similar molecular structures (the only difference between them is the alkyl chain length in the poly(alkyl methacrylate) blocks) and similar degree of polymerization (see Table 1), but their μ values are different by about one or two orders of magnitude. Now we discuss the molecular shear mechanisms that generate the large difference in the friction properties.

Table 2 Comparison of the friction coefficient μ between the A5L946A5/PAO system (this study), PAO (without lubricant) [11, 24], and the A5S992A5/PAO system [24]
Full size table

To discuss the molecular mechanism of the high friction of the A5L946A5/PAO system, we examine the friction trace (friction force versus time plot) that includes an abundance of information regarding the molecular shear behaviors. The A5L946A5/PAO system exhibited stick-slip friction at the low load (5 mN or less) and low sliding velocity (0.1 μm/s or less) conditions. Figure 8 shows the friction trace obtained at L = 2.6 mN and V = 0.003 μm/s. Stick-slip friction was observed (Region (I) in Fig. 8) whose amplitude (the difference between Fs and Fk, inset) was roughly constant and equal to approximately 0.4 mN. However, the stick-slip behavior did not show a regular saw-tooth pattern but rather exhibited an irregular profile (inset), suggesting that multiple small stick-slip events were involved in one “large” spike. The dynamic thickness of the film at the sliding condition was in the range of 50–100 nm (Fig. 4), and the confined film should consist of adsorbed and non-adsorbed A5L946A5 molecules (as was already discussed). Entanglement/disentanglement between the adjacent non-adsorbed/adsorbed and non-adsorbed/non-adsorbed A5L946A5 molecules is expected during sliding, which generates small/multiple stick and slip events in one large stick-slip spike. Disappearance of the large stick-slip spikes was observed at t ≈ 1700 s (Region (II)). This implies that the repetitive back and forth sliding motions could gradually stretch and disentangle the PLMA blocks between the molecules and/or rearrange the packing structure of the non-adsorbing molecules. However, the friction trace in Region (II) still exhibits a noisy force output, implying that stick-slip events between the adjacent molecular chains may occur. This extensive chain entanglement effect for the A5L946A5/PAO system gives rise to the high-viscosity properties of the confined film and results in high friction. We note that the force plotted in Fig. 5 represents the kinetic friction force (Fk) obtained from the minimum force in the stick-slip friction regime (low L and V) or the average force in the smooth sliding regime.

Fig. 8
figure 8

Friction trace for the A5L946A5/PAO system obtained under the sliding condition of L = 2.6 mN and V = 0.003 μm/s. Stick-slip friction (I) was observed, and the stick-slip did not have a regular saw-tooth pattern but rather exhibited a complicated profile (inset). The Fs and Fk represent the maximum and minimum force in the irregular stick-slip spikes. The stick-slip disappeared along with repetitive sliding motions (II)

Full size image

As discussed above, the contact geometry of the A5L946A5/PAO system has a concave shape and the thickness at the center is larger than that at the edge (Figs. 3 and 4). Back and forth sliding motions continuously trap new non-adsorbed A5L946A5 molecules from the sliding front edge in both sliding directions, and the newly trapped molecules are gradually accumulated at (or near) the center of the contact area. This unique contact shape is accompanied by the elastic deformation of the glue layer under each mica sheet. During sliding, while some PAO molecules are squeezed out at the edge, the A5L946A5 molecules cannot be expelled due to the chain entanglement effect and their large size. Then, the A5L946A5 molecules become concentrated in the confined film, making an additional contribution to the increase in the effective viscosity and high friction of the system.

The friction behaviors of the A5S992A5/PAO system studied previously [24] are briefly summarized as follows. The solubility of A5S992A5 in PAO is relatively low and the bulk A5S992A5/PAO solution prepared at 45 °C becomes cloudy after it is left to stand at room temperature for 5–8 h. Reflecting this low solubility (affinity), the dynamic thickness of the A5S992A5/PAO system confined between the mica surfaces gradually decreases with time. The dynamic thickness 1 day after the surface preparation (sample injection, Day 1 film in Table 2) was approximately 40 ∼ 70 nm and the thickness 2 days after preparation (Day 2 film) was approximately 20 nm, indicating the gradual shrinkage of the A5S992A5 molecules in the confined film. The μ value for the A5S992A5/PAO system was lower than the values for PAO and decreases with time; the μ for the Day 2 film was well below 0.01. The friction trace for the Day 2 film exhibits smooth sliding, and no apparent chain entanglement effect is expected at the sliding interface. Overall, the results suggest that the film consists of adsorbed A5S992A5 molecules and PAO; no non-adsorbed A5S992A5 molecules are present in the confined film. The adsorbed A5S992A5 layers (with a shrunken conformation) constitute a rigid slip plane at the contact interface, and extremely low friction is attained by the slipping of the intervening PAO layer on the slip plane.

The two block copolymer additives, A5L946A5 and A5S992A5, have similar molecular structures (the only structural difference between them is the alkyl chain length in the poly(alkyl methacrylate) blocks) and similar degree of polymerization (Fig. 1 and Table 1). By contrast, the friction reduction ability of the two additives is quite different. The major reason for this difference is attributed to the affinity (solubility) of the additive in PAO. A5L946A5 is highly soluble in PAO while A5S992A5 has a smaller solubility, as mentioned above. The PLMA blocks in the A5L946A5 molecule have expanded conformation in PAO giving rise to the extensive chain entanglement effect between the adjacent molecules. On the other hand, the poly(stearyl methacrylate) blocks in the A5S992A5 molecule have shrunken conformation in PAO and little chain entanglement is expected. As a result, the friction properties of the A5L946A5/PAO system are strongly affected by the continuous trapping of the non-adsorbed additive molecules into the contact area during sliding, generating a thick highly viscous film and resulting in high friction. By contrast, the confined film of the A5S992A5/PAO system consists of two opposed adsorbed A5S992A5 layers with shrunken conformations; intervening slippery PAO layer results in the extremely low friction.

To evaluate the possible practical performance of the additives for the engineering systems, we performed ball-on-disk friction measurements using a Bruker UMNT-1 tribometer with a rotary drive. Friction force was measured between a steel ball (diameter of 3 mm) and a steel disk (surface roughness Ra of approximately 6 nm). The applied load was below 5 mN and the sliding velocity was 3 μm/s; these values are comparable to the SFA experiment. The obtained friction coefficient μ was 0.15 ± 0.02 for PAO (without additive), 0.06 ± 0.01 for A5L946A5/PAO, and 0.04 ± 0.00 for A5S992A5/PAO; the A5L946A5 additive exhibited approximately 60% friction reduction when dissolved in PAO. Because of the surface roughness, surface properties of the materials (e.g., hardness), and different experimental setup/conditions, the friction mechanisms that give rise to the results of the ball-on-disk measurements should be very different from those in the SFA. However, it was found that the friction reduction ability of A5S992A5 is superior to that of A5L946A5; the addition of A5S992A5 to PAO showed the friction reduction of approximately 75%. Therefore, different extent of chain entanglement and different adsorbed/non-adsorbed film structures observed in the SFA should provide insights into controlling and reducing friction also for the engineering systems.

Conclusions

The friction properties of a dilute solution of a triblock copolymer additive in PAO confined between mica surfaces were investigated using the SFA. Friction measurements were performed for a triblock copolymer, poly(11-acrylamidoundecanoic acid)-block-poly(alkyl methacrylate)-block-poly(11-acrylamidoundecanoic acid) with an alkyl chain length in the poly(alkyl methacrylate) blocks of C12 (A5L946A5), and the results were compared to the results obtained in a previous work for a similar triblock copolymer with an alkyl chain length of C18 (A5S992A5) [24]. The friction coefficient μ of the A5L946A5/PAO system is 0.54−0.56, which is larger than that of the confined PAO without the additive (0.04−0.47). The μ of the A5S992A5/PAO system examined previously is well below 0.01; the friction reduction ability of the two additives is very different, despite their similar molecular structures and similar degree of polymerization. The large difference in the friction reduction ability can be explained based on the differences in the affinity (solubility) of the additives to PAO and the resulting additive conformation/confined film structure. The molecular conformation of the A5L946A5 molecules in PAO is expanded because of the high affinity, giving rise to an extensive chain entanglement effect between additive molecules. This entanglement effect results in the accumulation of non-adsorbed additive molecules from the sliding front into the contact area during repetitive sliding motions, and the formation of a thick, highly viscous interfacial film results in high friction. We concluded in our previous study [24] that the low affinity to PAO and resulting shrunken adsorbed conformation of A5S992A5 on sliding surfaces are the essential features for the extremely low friction of the system. The approach of designing low-friction interfacial structures by optimizing the affinity between the additives and a base oil is clearly supported by the systematic comparison of additive molecules with similar molecular structures.