Wet sliding friction of elastomer compounds on a rough surface under varied lubrication conditions

Wear262(2007)707–717Wet sliding friction of elastomer compounds on a roughsurface under varied lubrication conditionsXiao-Dong Pan∗Bridgestone Americas Center for Research and Technology,1200Firestone Parkway,Akron,OH44317-0001,United StatesReceived28February2006;received in revised form1August2006;accepted3August2006Available online18September2006Abstract‘Green’tire bearing tread rubber reinforced with precipitated silica can exhibit improved wet traction performance.The underlying mechanism is currently not well understood.To improve our capability of rational material design for enhanced driving safety,wet sliding friction for various rubber compounds is tested on a Portland cement concrete surface under varied lubrication conditions.The wetting liquid is either ethanol or water,and the initial amount of liquid on the concrete surface is adjusted.Sliding friction is detected to alter with lubrication condition.Under ethanol lubrication,the sliding friction is markedly lower than that under water lubrication.Additionally the benefit in wet traction from silica is significantly diminished or eliminated in ethanol.Such observations cannot be rationalized with simple considerations of rubber bulk viscoelasticity, liquid viscosity,cavitation,or capillary effect.We believe that these observations strongly indicate the significance of interfacial interactions in determining the wet sliding friction of elastomer compounds.The potential relevance of capillary porosity in concrete and Marangoni drying effect to wet traction is also introduced.©2006Elsevier B.V.All rights reserved.Keywords:Elastomer;Wet sliding;Friction;Silica;Water;Ethanol1.IntroductionIn order to stop a moving automobile on a highway wetted by rain,a maximal level of sliding friction from tires is obviously desired upon brake application.At low sliding speed before the emergence of elastohydrodynamic lubrication,other than the tread pattern,material properties of tread rubber compounds play a significant role in determining the wet traction performance of a pneumatic tire.Tread rubber compounds are made of crosslinked elastomers reinforced withfinefiller particles.The conventionalfiller has been carbon black derived from fossil fuel.Starting from the early1990s,precipitated silica used along with silane has become popular in preparing the so-called‘green’tire[1].In comparison to carbon black-filled rubber,silica-reinforced rub-ber can lead to simultaneous reduction in fuel consumption and enhancement in wet grip while maintaining wear resis-∗Tel.:+13303797339;fax:+13303797530.E-mail address:panxiaodong@.tance.Such unexpected gain in the combination of performancecharacteristics has spurred renewed interests in fundamentalunderstanding on mechanisms of wet sliding friction.For tribological systems consisting of rubber sliding on arough road surface with water in-between,factors affecting the friction characteristics have been recognized[2–4].During slid-ing,the multi-scale asperities on road surface induce high fre-quency dynamic deformation of rubber.Viscoelastic dissipationof energy inside rubber is considered as the primary contributorto sliding friction[5].Coefficient of frictionμdepends on theso-called reduced variable a T·v.Here the temperature dependent a T is the well-known Williams–Landel–Ferry time–temperatureshifting factor[6]while v is the sliding speed.In comparison,theload exerts much less sensitive impact on friction[4].Quantita-tive analysis of rubber friction certainly requires proper accountof the typically nonlinear viscoelasticity for rubber compounds,the specific roughness characters for the road surface,as wellas effects from friction heating at high v.Though not yet wellexplained,testing of‘veneered’rubber compounds indicates theimportance of the thin surface layer of rubber in determining wetsliding friction[7,8].0043-1648/$–see front matter©2006Elsevier B.V.All rights reserved. doi:10.1016/j.wear.2006.08.003708X.-D.Pan/Wear262(2007)707–717Without interference from surface charge[4],remaining contribution from adhesion(or adhesion hysteresis[9–11])is detectable(embodied as the presence of a hump along a friction master curve).Possible impact from the chemical nature of road surface has rarely been assessed[3].In the presence of sharp surface asperities,tearing of rubber can further enhance friction [12].Even though intuitive concern on possible change in rhe-ological behavior of extremely thin waterfilm was raised[13], viscous dissipation within the remaining waterfilm is generally considered insignificant.Some detailed studies have been performed for the case of smooth rubber on a wet smooth glass surface.Summarizing variation of friction with dynamic lubrication conditions,some Stribeck curves were established[14].Thefine details of extru-sion of liquid between the rubber and the glass surface were examined with the method of optical interferometry[15].Under proper conditions during sliding,waves of rubber detachment (Schallamach waves)were observed[16].Under both non-sliding and sliding conditions,the process of interfacial energy-driven dewetting has also been studied[17–20],including the situation where there exists single defect on the glass surface.A smooth glass surface is certainly much less complicated than a realistic highway surface.Other than the multi-scale asperities on the road surface,a piece of typical Portland cement concrete can have porosity of10–20%[21].Thus the road sur-face can be leaky.Even though possible implications of capillary porosity to liquid extrusion,viscous energy dissipation,and rewetting of surface are completely ignored in recent literature, there is an earlier investigation of dynamic permeability of pave-ments[22].Facing such multitude of factors,the difficulty in predicting the likely level of friction of rubber components in engineering practice has been realized[23].Recently,quantitative modeling of rubber friction on a fractal surface has been presented based on bulk viscoelastic description of material behavior[24,25]. Among the many proposals attempting to rationalize the benefit in wet traction from silica,the existence of a softer skin at the sliding interface for silica-filled rubber appears plausible[26]. The stress-softening effect can be stronger for silica-filled rubber than that for carbon black-filled rubber.However,practitioners also reported discrete singular observation that the benefit in wet traction from silica disappears on a surface lubricated with ethanol[27].Obviously establishing a complete set of phenomenology is essential for achieving a true understanding on friction mecha-nisms,which is the main purpose for this study.Here wet sliding friction for various rubber compounds is tested under different lubrication conditions with a portable British Pendulum Skid Tester(BPST)[28,29].A Portland cement concrete patio block provides the road surface.The concrete surface is wetted with either ethanol or water.Due to the current absence of efficient method for accurate quantification of waterfilm thickness on a rough surface,the lubrication condition is further adjusted by generating visually different amount of liquid on the surface before any sliding test.With the BPST,a rubber slider attached to the end of the swinging arm slides on the road surface for afixed length of path.During sliding,kinetic energy of the swinging arm is con-sumed to overcome the sliding friction.Consequently after the slider comes off contact with the road surface,the initial releas-ing height of the arm cannot be regained.The instrument reading, British Pendulum(Tester)Number or BPN,corresponds to such lost in height and hence indicates the level of sliding friction. Approximately BPN is of order of100×μ,hereμis the coef-ficient of friction[30].Hence it is more sensitive to use BPN to display the variation of friction with lubrication condition.It is difficult to quantify the dynamic contact condition for BPST.However,effects from a variety of factors on wet skid resistance of rubber compounds have been revealed with BPST [30–34].The designed average contact pressure during sliding for BPST is of order of0.21MPa[28],similar to the nominal pressure between a passenger tire and road.However,the local contact pressure between rubber and sharp surface asperities can reach well beyond10MPa[35].The rubber slider moves at about 2.7m/s(or9.7km/h)when itfirst makes contact with the surface. At such sliding speed,the shear rate in a thin waterfilm of1nm thickness on a smooth surface is of order of109s−1.The relax-ation rate of bulk water molecules is of order of1012s−1[36].The lubrication behavior of extremely thin liquidfilms con-fined between molecularly smooth surfaces has been quan-titatively examined with various Surface Force Apparatuses (SFA)[37–39].The contact pressure encountered in SFA can be3–5MPa[36].Current experimental evidences indicate that when confined between hydrophilic surfaces,thefluidity and lubricity of water remains down to one single layer of molecules [40,41].With a different experimental technique,ordering of liquid alkanes confined between an elastomer and a sapphire substrate has recently been reported[42].Examination of sliding friction under ethanol lubrication is also important for other reasons.As displayed in Fig.1,rapid formation of a‘dry’patch can be induced by dropping a small bead of ethanol onto water residing on top of the concrete sur-face.This arises from the significant difference in surface tension between the two liquids.At20◦C,the surface tension is22.3 and72.8mN/m for ethanol and water,respectively[43].Can such Marangoni drying effect[44,45]be applied to improve tire wet traction performance?Such expectation may befurther Fig.1.Marangoni drying on the water-covered concrete surface caused by a drop of ethanol(diameter∼2mm).X.-D.Pan/Wear262(2007)707–717709enhanced by the following observation[3,14].For a smooth rubber sliding(at small v)on a wet smooth glass surface,in the boundary lubrication regime,the friction under alcohol lubri-cation is higher than that under water lubrication.This was attributed to the absorption of alcohol into the rubber.A larger contact area was resulted from the softened swollen rubber.For paraffin alcohols of varied hydrocarbon chain length,friction maximize at octanol.In contrast,on a Perspex(polymethyl-methacrylate)surface,ethanol becomes a better lubricant[46].2.Experimental methods2.1.MaterialsEven though rubber compounds prepared with elastomers exhibiting quite different glass transition temperature T g have been tested,in this report we focus primarily on results obtained for four compounds made of a tin-coupled styrene–butadiene rubber(SBR,random copolymer)with T g at−29.5◦C(Duradene739from Firestone Polymers).Here the midpoint T g is identified via the differential scanning calorime-try(DSC)characterization at the heating rate of10◦C/min.The microstructure of polymer molecules was revealed with nuclear magnetic resonance(NMR)test:21.0%bound styrene(by weight),47.3%1,2-butadiene content,and31.7%1,4-butadiene content.The gel permeation chromatography(GPC)tests indi-cated a weight-average molecular weight of411.9×103g/mol, a polydispersity index of1.59,and a total coupling of60.3% including a thermal coupling of11.8%.The four compounds include a gum compound containing no filler,a carbon black-filled compound,a silica-filled compound prepared in the absence of silane,and a silica-filled compound prepared in the presence of silane Si69.In the following they are denoted as SBR-G,SBR-C,SBR-SiO2,and SBR-SiO2-Si69, respectively.The compounding formula is shown in Table1. Thefilled compounds contain50phr(part per hundred rubber, by weight)offiller.Carbon black of grade N339was used in SBR-C while amorphous precipitated silica(Hi-sil190G from PPG Indus-tries)was used in the silica-filled compounds.The typical pri-mary particle size according to electron microscopy is between 26and30nm for N339,and is averaged at17nm for Hi-Sil 190G.The nitrogen-adsorption surface area(according to the Brunauer–Emmett–Teller equation)for N339and Hi-Sil190G is96and215m2/g,respectively.In preparing SBR-SiO2-Si69,5phr of liquid organosilane bis(3-triethoxysilylpropyl)tetrasulfane(Si69from Degussa)was added during remill.In the absence of silane,to compensate for potential loss of sulfur due to its adsorption onto silica surface, a larger amount of sulfur was incorporated into SBR-SiO2to achieve a sufficient level of crosslinking.This,in addition to the strong inter-particle hydrogen bonding(arising from hydroxyl groups on silica surface),makes SBR-SiO2the hardest com-pound.Considering the essential difference in surface chemistry between carbon black and silica,as well as the role of coupling from the silane Si69,the bonding between matrix polymer and filler particle surface is very different among the threefilled compounds.Some typical characteristics for the four compounds are listed in Table2.The tensile testing was performed with a model4501 Instron Universal Testing Instrument.Viscoelastic behavior for the compounds was examined with two Advanced Rheomet-ric Expansion Systems(from Rheometric Scientific)[34].To examine the extent of absorption of ethanol into the rubber com-pounds,for each compound,four rectangular specimens were immersed into ethanol.Each specimen was about1.9mm thick and weighted about1.0g.The ethanol was replaced after24h of immersion.The wet weight was obtained for two specimens after72h of immersion and for the other two specimens after 144h of immersion.Each specimen was subsequently dried for24h in a vacuum oven heated to about60◦C and its dry weight was acquired.The total swelling is calculated as the dif-ference in wet weight and dry weight normalized by the dry weight.Table1Compounding formulasMixing stage In phr SBR-G SBR-C SBR-SiO2SBR-SiO2-Si69 Master batch SBR100100100100Carbon black05000Silica005050Stearic acid22226PPD a1111Remill Si69b N/A005Final batch Zinc oxide3333DPG c0.50.50.70.7MBTS c1111TBBS c0011Sulfur 1.3 1.3 3.0 1.15Total108.8158.8161.7164.85a The antidegradant6PPD is N-(1,3-dimethylbutyl)-N -phenyl-p-phenylenediamine.b The bifunctional,sulfur-containing organosilane Si69is bis(triethoxysilylpropyl)tetrasulfane.c The vulcanization accelerators DPG,MBTS,and TBBS represent diphenyl guanidine,dibenzothiazyl disulfide,and N-tert-butyl-2-benzothiazyl sulfenamide, respectively.710X.-D.Pan/Wear262(2007)707–717Table2General characteristics for the cured compoundsSBR-G SBR-C SBR-SiO2SBR-SiO2-Si69 Filler volume fraction a(%)0.019.717.917.4Midpoint T g(◦C)−25.8−24.7−21.6−24.7 Temperature at peak tanδb(◦C)−16.0−16.0−14.0−16.0 Tensile test at RT Strain rate during testing at0.282s−1Stress at break(MPa) 1.9322.116.620.1Strain at break(%)381.2437.3325.6424.7Stress at200%strain(MPa) 1.147.729.887.52Modulus at200%strain(MPa)0.350 5.18 5.02 4.33 Dynamic strain sweep test c Tested at1Hz,values below taken at the strain amplitude of11.8%(1)At RT(1.1)G at RT(MPa)0.627/0.643 2.52/2.56 5.38/5.43 2.68/2.79(1.2)tanδat RT0.0720/0.07250.141/0.1440.190/0.1880.153/0.156(2)Temperature d◦C−14.2−13.4−11.1−13.0(2.1)G (MPa) 3.86/4.1012.6/12.920.3/21.112.5/12.9(2.2)tanδ 1.82/1.83 1.32/1.32 1.11/1.11 1.33/1.31(3)Temperature e(◦C)−14.7−14.0−11.7−13.5(3.1)G (MPa) 4.0313.419.712.7(3.2)tanδ 1.88 1.36 1.21 1.40(4)At−14.2◦C f(4.1)G (MPa) 3.57/3.8013.523.413.3(4.2)tanδ 1.81/1.78 1.37 1.30 1.44Total swelling g(wt.%)Sample weight∼1.0g,thickness∼1.9mm(1)After72h in ethanol 2.73 1.83 2.78 2.37(2)After144h in ethanol 3.50 2.28 3.10 3.26a Calculated according to the formula in Table1,taking1.8and2.0g/cm3as the density for carbon black and silica,respectively.b From dynamic temperature step test at1Hz with the strain amplitude at0.20%.c Two samples tested under some conditions to check reproducibility.d Determined via the method of time–temperature shifting,corresponding to BPST testing at20.9◦C[34].e Corresponding to BPST testing at19.2◦C for demonstration of sensitivity.f Tested after completion of all previous dynamic strain sweep tests.g Two rectangular samples tested for each compound.By itself,the rubber grade carbon black is hydrophobic while the precipitated silica is hydrophilic.The matrix elastomer is hydrophobic while the concrete surface is hydrophilic.The silanized silica becomes hydrophobic and thus compatible with the matrix elastomer.Both the rubber compound surface and the road surface are physicochemically heterogeneous.2.2.Testing of wet sliding frictionResults reported here were obtained under ambient condi-tions with the room temperature(RT)in the range between23.0 and24.8◦C.To remove possible residual mold release agent, the sliding edge of the rubber sliders was wiped with acetone at least1h before testing.BPST testing was performed for all the compounds in ethanol(anhydrous from AAPER)first.With-out being moved,the concrete block went through air-drying and extensive washing with water before further BPST testing under water lubrication.De-ionized water was then added for testing of wet sliding friction.Videos and pictures related to BPST testing were recorded with a digital camera(Canon Pow-erShot A620).For each compound,two sliders were tested under ethanol lubrication and two different sliders were tested under water lubrication.The concrete block has the dimensions of approximately 20.2cm×20.3cm×4.4cm.By examining the difference in weight between the wet block soaked with water and the block after thorough air-drying,the volume fraction of capillary porosity is estimated to be12.6%.Prior to this study,the testing surface has been exposed to a substantial amount of testing of wet friction.As displayed in Fig.2,the varied lubrication conditions investigated consist of the visually different initial amount of liquid present on the concrete surface before any sliding test. Corresponding to Fig.2a–c,in the following,these different lubrication conditions are referred to as Condition A,Condition B,and Condition C,respectively.A large portion of the con-crete body is immersed in the liquid.To avoid accumulation of possible wear debris of rubber on the testing surface,the sur-face was refreshed in the presence of sufficient amount of liquid with a brush bearing soft stainless steel hair.The diameter of an individual bristle is about120␮m.Unless otherwise noted,such surface agitation was consistently applied before every sliding test and the hair did become gradually worn away from brushing.Concerning Condition A,the relatively thick layer of liquid on the surface was formed by adding extra liquid after brushing. Previous BPST testing reported was all performed under thisX.-D.Pan/Wear262(2007)707–717711Fig.2.Different wet state of the concrete surface before friction testing.(a)Con-dition A:covered with a relatively thick layer of water.(b)Condition B:covered with a relatively thin layer of water.(c)Condition C:only the residual amount of water remaining.(d)Almost dried after overnight air-drying as a comparison.condition[32–34]for closer simulation of realistic concern.The obviously thinner layer of liquid on the surface for Condition B was formed by gently brushing off the extra liquid from the surface.After one sliding test under Condition B,a few sliding tests were performed consecutively without any other action to the surface.Such surface state corresponds to Condition C.For comparison purpose,a partially dried surface after air-drying for overnight is shown in Fig.2d.3.Results and discussionUnder an imposed sinusoidal deformation of small amplitude at the frequency of f0,the variation of elastic modulus G and loss tangent tanδwith temperature T was measured(Fig.3a).Loss tangent,characterizing the viscoelastic loss property,exhibits a distinct peak at the center of the transitional zone between the glassy state and the rubbery state.The temperature at the peak of tanδ,T Peak tanδ,varies with f0.Wet sliding friction for compounds very similar to SBR-G and SBR-C were previously tested for0<T<40◦C(Fig.3b)[32].Comparing the variation of BPN with T to the variation of tanδwith T,it appears that during BPST testing between0and40◦C,the dynamic state of the rubber compounds made of this SBR stays close to the center of the transitional zone.When the rubber slider slides across the surface bearing dif-ferent initial amount of water,the extent of water splashing is obviously different(Fig.4).Under Condition A,after the rubber slider leaves the surface,the patch swept out by the rubberslider Fig.3.Similarity in temperature dependency between bulk viscoelastic loss and wet sliding friction.(a)Variation of the elastic modulus G and the loss tangent tanδwith temperature.These are obtained under an imposed sinusoidal strain at the frequency of f0=0.01Hz and with the strain amplitude at0.20%.The temperature at the peak of tanδ,T Peak tanδ,is−23.3and−23.0◦C for SBR-G and SBR-C,respectively.(b)Variation of BPN with temperature reported previously for similar compounds[32].Reinforcing carbon black of different grade(N339vs.N121)was employed for preparing SBR-C in the two plots.712X.-D.Pan/Wear262(2007)707–717Fig.4.The rubber slider(for SBR-C)striking the concrete surface during thependulum skid testing.(a)A schematic plot of the portion of the pendulumswinging arm bearing the rubber slider.(b)Corresponding to Fig.2a.(c)Cor-responding to Fig.2b.(d)Corresponding to Fig.2c.Fig.5.Rewetting of the patch on the concrete surface swept out by the rubberslider(for SBR-C).The starting condition is that shown in Fig.2a.(a)Rightafter the slider comes off contact with the concrete.(b)A short moment later.(c)The shrinking‘dry’patch.(d)Complete disappearance of the‘dry’patch.The time interval from(a)to(c)is about1s.The‘dry’patch almost completelydisappeared in2s from(a).X.-D.Pan/Wear262(2007)707–717713Fig.6.Wet sliding friction detected under varied lubrication conditions.(a)For the gum compound with the surface wetted with either water or ethanol.(b)For all the reinforced compounds with the surface wetted with water.(c)For all the reinforced compounds with the surface wetted with ethanol.becomes quickly rewetted in just a few seconds(Fig.5).After repeated testing under Condition C,the surface remains visu-ally wet.Discussion on thefine difference between wet and dry surface was made long ago[47].It was noted that even a‘dry’surface may still be covered with adsorbed water up to the limit of invisibility at about500˚A.The possibility of boiling away of such thin waterfilm from friction heating at large sliding speed was recognized.In Fig.6,the time sequence of BPN collected for each com-pound under varied lubrication conditions is displayed.Here two samples were tested for each compound with good reproducibil-ity.Thefirst12and thefinal4sliding tests were performed under Condition A.In between there were four sliding tests under Con-dition B and four sliding tests under Condition C.For testing under Condition C,the time interval between two consecutive tests was about8s.Even though it is not dramatic,the impact on sliding friction from the initial amount of liquid on the concrete surface is always detected.It has been pointed out that at low sliding speed(less than about30km/h)the hydrodynamic buildup of liquid between rubber and road surface is negligible[48].However,we hope that the detected variation can offer clues to the understanding of energy dissipation mechanisms during wet sliding.When the lubricating liquid is changed from water to ethanol, however,the sliding friction becomes markedly reduced for all the compounds.Consequently,the attempt to enhance friction under water lubrication via the Marangoni drying effect is unsuc-cessful.The sliding edge for the slider was smeared with a paper towel soaked with ethanol or the more absorbing acetone.No increase in BPN was observed under Condition A.In fact,small negative impact was observed.If considering that ethanol may be absorbed into the com-pound and results in a softened compound,the observed reduced friction is just opposite to such reasoning.In fact,after immer-sion in ethanol for over144h,the absorption of ethanol into the compounds is most for SBR-G at3.5%by weight(Table2).Other compounds prepared with polymers of widely different T g(≤−29◦C,not including butyl rubber)were also tested under ethanol lubrication.Under ambient testing condition,BPN in water increases with increasing T g.Such a trend is retained in ethanol.During BPST testing the system is in a highly non-equilibrium state.The constant agitation to the liquid during testing increases evaporation of liquid and consequently grad-ually lowers the liquid temperature.For one series of testing under water lubrication,RT was between24.1and24.8◦C, the initial water temperature was at22.6◦C.After over2h of testing,temperature for water surrounding the concrete block became20.0◦C.In contrast,for one series of testing under ethanol lubrication,RT was between23.4and23.5◦C,the initial ethanol temperature was at20.4◦C.After about2h of testing,the ethanol temperature became15.0◦C.Temperature on the con-crete surface may be even lower.Under water lubrication,BPN for compounds made of a low-cis polybutadiene(T g=−90.4◦C) increases gradually with decreasing T[33].However,BPN for such compounds still becomes markedly lower when tested in ethanol under ambient condition.Thus the reduced friction from ethanol lubrication is unlikely explained from bulk viscoelastic consideration.The viscosity of water is1002and1793␮Pa s at20and 0◦C,respectively.The viscosity of ethanol is1201,1317and 1786␮Pa s at19.2,14.5and0◦C,respectively[49].As dis-played in Fig.3b for the compound SBR-G,when T is decreased from20to about0◦C,the wet sliding friction is only reduced by about5BPN units.Thus the reduction of about10BPN units in sliding friction for SBR-G under ethanol lubrication cannot come from the higher ethanol viscosity under the testing condi-tions employed.For consideration of possible cavitation based on the simplis-tic conventional criterion,the vapor pressure for water is2.34and 6.63kPa at20and38◦C,respectively.The vapor pressure for ethanol is5.81and4.27kPa at20and15◦C,respectively.Under water lubrication,when T is increased from20to about40◦C, BPN for the compound very similar to SBR-C remains essen-tially unchanged(Fig.3b).As reported earlier[33],the benefit in wet traction from silica under water lubrication persists in the same temperature range.Therefore,the vapor pressure at15◦C714X.-D.Pan/Wear262(2007)707–717for ethanol cannot explain the significantly reduced sliding fric-tion under ethanol lubrication.Under the lubrication of the same initial amount of water (Fig.6b),the sliding friction for both silica-filled compounds is markedly higher than that for the carbon black-filled compound. Even comparing across testing under the lubrication of different initial amount of water on the concrete surface,on average,BPN for SBR-C under Condition B is still lower than BPN for SBR-SiO2and SBR-SiO2-Si69under Condition A by1.5and3.8, respectively.In contrast,under the lubrication of the same initial amount of ethanol(Fig.6c),the enhancement in wet sliding friction from silica is much reduced for SBR-SiO2-Si69and eliminated for the harder SBR-SiO2.From tensile testing at RT(Table2),the modulus at200% strain for SBR-C is higher than that for SBR-SiO2-Si69.This is partially due to the fact that the density for silica is slightly higher than that for carbon black,thus at the samefiller loading of50phr(by weight),the actual volume fraction offiller is lower for SBR-SiO2-Si69.A compound was prepared with the same SBR,in the presence of Si69,but with the amount of silica increased to match thefiller volume fraction for SBR-C.The modulus at200%strain for this compound is higher than that for SBR-C.Under water lubrication,BPN for this compound under Condition A is slightly lower than BPN for SBR-SiO2-Si69under Condition A(by1.1),but still markedly higher than BPN for SBR-C under Condition B(by2.7).Comparing SBR-SiO2to SBR-SiO2-Si69,there exists obvi-ous difference in modulus(Table2),some quantitative difference in wet sliding friction is also detected(Fig.6).This indicates that for compoundsfilled with the same kind offiller,change in viscoelastic characters is expected to result in corresponding variation in wet sliding friction.Values of the conventional vis-coelastic predictor for wet traction,tanδ,are listed in Table2 for all the compounds.The dynamic strain sweep tests were performed at1Hz and at low temperature determined with the method of time–temperature shifting,assuming that the domi-nant frequency of material deformation during BPST testing is at104Hz[34].Value of tanδis somewhat arbitrarily taken at the specific strain amplitude of11.8%for all the compounds.From Table2,the predictor from bulk viscoelastic measurement can-not account for the difference in wet sliding friction under water lubrication that arises from the usage of differentfiller particles.After completion of the testing displayed in Fig.6,the wear scar at the sliding edge of the rubber sliders was examined under a stereo optical microscope.Pictures of the wear scar are shown in Fig.7for all the compounds.Clearly,the wear scar on the sliders tested under ethanol lubrication is less severe than that on the corresponding sliders tested under water lubrication.Thus the lower wet sliding friction detected under ethanol lubrication is accompanied with a less severe wear at the sliding edge of the slider.Based on the discussions above,the eliminated benefit in wet traction from silica under ethanol lubrication cannot be explained with simple considerations of liquid viscosity,cav-itation,capillary effect,or rubber bulk viscoelasticity.However, such effect from ethanol lubrication does resemble strongly fea-tures of interfacial interactions in a liquid mediumrevealed Fig.7.Wear scar at the tested edge for the rubber sliders subjected to the testing displayed in Fig.6.Relative to the slider,the road surface moves along the arrow direction.The division on the ruler indicates1mm.。