Frictional response of fatty acids on steel

Journal of Colloid and Interface Science333(2009)707–718Contents lists available at ScienceDirectJournal of Colloid and Interface Science/locate/jcisFrictional response of fatty acids on steelRashmi R.Sahoo,S.K.Biswas∗Department of Mechanical Engineering,Indian Institute of Science,Bangalore560012,Indiaa r t i c l e i n f o ab s t r ac tArticle history:Received13October2008 Accepted24January2009 Available online20February2009Keywords:Self-assemblyAnti-frictionFatty acidsLateral force microscopy Nanotribology Self-assembled monolayers of fatty acids were formed on stainless steel by room-temperature solution deposition.The acids are covalently bound to the surface as carboxylate in a bidentate manner.To explore the effect of saturation in the carbon backbone on friction in sliding tribology,we study the response of saturated stearic acid(SA)and unsaturated linoleic acid(LA)as self-assembled monolayers using lateral force microscopy and nanotribometry and when the molecules are dispersed in hexadecane,using pin-on-disc tribometry.Over a very wide range(10MPa–2.5GPa)of contact pressures it is consistently demonstrated that the unsaturated linoleic acid molecules yield friction which is significantly lower than that of the saturated stearic acid.It is argued,using density functional theory predictions and XPS of slid track,that when the molecular backbone of unsaturated fatty acids are tilted and pressed strongly by a probe,in tribological contact,the high charge density of the double bond region of the backbone allows coupling with the steel substrate.The interaction yields a low friction carboxylate soapfilm on the substrate.The saturated fatty acid does not show this effect.©2009Elsevier Inc.All rights reserved.1.IntroductionUltrathin organic monolayers deposited on solid substrates by spontaneous organization of molecules(self-assembled mono-layers,(SAM)),because they provide substantial mechanical and chemical protection to the substrate.Such protection has found many practical applications in lubrication,catalysis,biological sen-sors,optoelectronic device and information storage systems[1–6]. Steel an important engineering material oxidizes easily in air form-ing a reactive hydrophilic surface.This generally leads to poor tribological performance.Passivation of a steel surface can be achieved with self-assembled monolayers as they bond to render the surface hydrophobic and chemically inert.In this carboxylic acid monolayers have been found to be particularly effective[7–10]as they chemiadsorb on metallic substrates as monocarboxylate anions[1–3].When such an assembly is slid against a solid counterface,a ‘low friction’thinfilm under certain conditions imposed by ex-ternal variables may be generated in situ.This has been observed when the acid molecules are adsorbed from being dispersed in oil[11]as well as when they are self-assembled on metallic sub-strates[7–10].This phenomenon has been exploited to design friction modifiers for engine lubrication[11],lubrication of micro-electro mechanical system and mechanisms[12,13],to improve the dispersion capabilities of functionalized carbon nanotubes in sol-vent and composites[14]and lubrication of information storage*Corresponding author.Fax:+918022932963.E-mail address:skbis@mecheng.iisc.ernet.in(S.K.Biswas).systems[15–17].Thefirst and so far the most authoritative expla-nation of the phenomenon emerged from the work of Bowden and Tabor[18].They underlined the importance of chemisorption as a prelude to good boundary lubrication tribology and showed that the adsorbed molecules are transformed under tribological condi-tion to a layer of fatty soaps such as stearate and palmitates by chemical reaction with the substrates.The anti-friction property is thus limited by the melting point temperature of the soap.While this has been validated as a generic phenomenological mechanism, optimization of the conformational structure to select a molecule from a range including many biological species[19]has not made major strides to date.We believe that,that is because afirm phys-ical rationality for adsorption,related to the electronic and spa-tial interaction between the molecules and the substrates,has not been established.Kondo[15]argues that the lack of packing order in unsaturated molecules assembled on aluminum is responsible for their poor tribology compared to that of saturated molecules. Adhvaryu et al.[19]based on their study of multicomponent bio-based lubricants also observed that increasing unsaturation low-ers the ability of the molecules to be chemisorbed.The work of Smith et al.[11]using a reciprocating tribometer however showed that the effect of unsaturation at low loads and temperatures is marginal while at higher loads and temperatures unsaturation re-duces friction substantially.Double bonds in an organic molecule localize electronic density and this localized charge density leads to a greater polarizability for the carbon atoms in the double bonds.We proceed in this pa-per with a hypothesis that if such high-density localized charges are allowed to interact with a metallic solid substrate,it would aid0021-9797/$–see front matter©2009Elsevier Inc.All rights reserved. doi:10.1016/j.jcis.2009.01.046708R.R.Sahoo,S.K.Biswas /Journal of Colloid and Interface Science 333(2009)707–718Fig.1.(a)The molecular conformation and the energy minimized structures of the stearic (SA)and linoleic acid (LA),estimated from DFT calculations with B3LYP correlation and basis Set LANL2DZ.The carbon atoms are depicted in white,the hydrogen atoms in black and the oxygen atoms in gray.The decorated carbon atoms in LA are the positions of the double bonds.(b)The calculated electron density distribution on stearic and linoleic acid.the formation of a strongly bonded beneficial tribofilm.If this can be demonstrated it would provide a basis for distinguishing the tri-bology of saturated and unsaturated fatty acids.The test molecules chosen to explore this hypothesis are saturated stearic acid (CH 3–(CH 2)16–COOH)and unsaturated linoleic acid (CH 3–(CH 2)4–C =C–(CH 2)–C =C–(CH 2)7–COOH).The molecules (Fig.1)have the same stoichiometry except that linoleic acid carries two C =C double bond in the alkyl backbone between the 9th and 10th &12th and 13th carbon atoms.These molecules are self-assembled on stain-less steel and Si(111)substrates and their monolayer status and crystallinity characterized using grazing angle FTIR.The frictional properties of the dry monolayers are recorded in atomic force microscope lateral mode and nanotribometer.The latter allows a modulation of contact pressure to vary the physical proximity of the pressed molecular backbone to the steel substrate.To enable chemical analysis of the tribological residue we disperse the fatty acids in hexadecane and use the mixture to lubricate the contact in a pin-on-disc experiment.The track generated was clear enough and large enough to examine the residue using X-ray photoelec-tron spectroscopy (XPS).2.Materials and methods 2.1.MaterialsPolycrystalline EN-31steel (carbon 1%,manganese 1.1%,silicon 0.1%,phosphorus 0.05%,sulfur 0.05%,chromium 1%,rest iron)is used as substrates in our experiments.The fatty acid monolay-ers were formed from solutions of stearic acid (C18:0,octade-canoic acid,>99%,Lancaster),and linoleic acid (C18:2,cis ,cis -9,12-octadecadienoic acid,99%,Sigma-Aldrich)in n -hexadecane (>99%,Sigma-Aldrich).All the compounds were used as received.Hexade-cane was further purified by percolating through neutral alumina (s.d.fine-chem.,India)twice.2.2.Pretreatment of stainless steel and preparation of samples for AFM and nanotribometry testsThe EN-31steel was machined and ground to give sample di-mensions,30×10×5mm.The samples were then emery polishedusing 220,320,400,600,800,1000,1200and 1500grit silicon carbide papers.After emery polish the samples were rinsed in ace-tone (AR Grade,s.d.fine-chem,India)and then disk polished using diamond paste of grade 1–3μm followed by 0.5μm.The polished samples were cleaned by ultrasonication in acetone to remove all polishing debris.Finally,the substrates were dried in a stream of dry nitrogen gas and preserved in a desiccator.Before SAM deposi-tion,steel substrates were kept in a UV cleaning chamber (Bioforce Nanosciences,USA)for 30min to burn all carbonaceous contami-nants which may block adsorption sites.The polished samples were characterized for their roughness using a profilometer (Talysurf,Taylor Hobson,Leicester,UK)and then atomic force microscopy.The average roughness of these samples after disk polishing was found to be ∼9nm in the pro-filometer (scan length =4.8mm)and ∼3nm in AFM (scan area 2μm ×2μm).The steel substrates prepared as above,were im-mersed in a freshly prepared additive solution (1mM)of fatty acid in n -hexadecane for 24h at room temperature.After deposition the samples were taken out,rinsed,and washed with hexadecane followed by ethanol (AR Grade,Les Alcools De Commerce Inc.,On-tario,Canada)to remove excess and physisorbed acid molecules.Finally the samples were stored in a desiccator prior to subjecting them to spectroscopic and tribological characterization.2.3.Surface functionalization of Si substrate with fatty acid monolayers p-Type Si(111)wafers (Microelectronics Inc.,CA,USA)were washed with acetone,methanol and ethanol and then sonicated in acetone for 5min.The samples were then placed in 3:1(v/v)concentrated H 2SO 4/30%H 2O 2(aq)(Piranha solution)for 30min at 90◦C to clean the surfaces.Warning:Piranha solution is exceed-ingly dangerous and should be kept from organic materials and treated with care .The wafers were removed from the Piranha solution and washed with copious amounts of HPLC grade water.The wafers were partially hydrophilic after this treatment.The Si wafer was immediately immersed in a neat hexadecane solution of fatty acid for 36h at room temperature.The excess physisorbed reagent on the planar Si was removed by rinsing,at room temperature,with hexadecane and ethanol,whereupon the samples were dried in aR.R.Sahoo,S.K.Biswas/Journal of Colloid and Interface Science333(2009)707–718709stream of nitrogen gas.Finally the samples were stored in a des-iccator prior to subjecting them to spectroscopic and tribological characterization.2.4.Contact angle measurementsAdvancing and receding contact angle measurements on the SAM surfaces were performed using the sessile drop method to record the motion of the three-phase contact line between water and SAM surface.The measuring device was an OCA30commer-cial goniometer(Dataphysics,Germany)equipped with a stepper motor for controlling the volume of the liquid supplied from a mi-crosyringe.2.5.FTIR analysisAll spectra were acquired by infrared reflection absorption spec-troscopy(IRRAS)in a Perkin-Elmer Spectrum GX spectrometer equipped with liquid nitrogen cooled mercury cadmium telluride (MCT)detector.All IR spectra reported here were referenced to a bare steel or Si(111)substrate and acquired over1024scans at 4cm−1resolution with a p-polarized beam.The sample chamber and the detector were purged with dry nitrogen gas before starting the experiments(in air)and at regular intervals during the experi-ment.The IR beam was reflected from the surface with an incident angle of75◦from the surface normal.The spectral analyses were carried out with Spectrum v3.02software(Perkin-Elmer).2.6.Tribological testsThree types of tribological tests were carried out:(1)Lateral force microscopy(Explorer AFM,Thermo Microscope,Santa Bar-bara,USA)—Si3N4tips of30nm tip radius mounted on a can-tilever of stiffness0.15N/m were used to generate a mean Hertzian pressure range of970–2500MPa;(2)Nanotribometer(CSM Instru-ments,Switzerland)—2mm diameter100Cr6steel ball mounted at the end of a normally loaded cantilever was used to generate a mean Hertzian pressure range of200–400MPa;and(3)Pin-on-disc(POD)tribometer(DUCOM,Bangalore,India)—3mm diam-eter steel pin was used to generate a mean Hertzian pressure of 10MPa.2.7.Atomic force microscopyAFM experiments were performed using an Explorer AFM (Thermo Microscopes,Santa Barbara,USA)with Si3N4cantilevers (Veeco,USA)that have pyramidal tips of a nominal end radius of30nm,with a single tube scanner under ambient condition (21◦C,relative humidity35–45%).A V-shaped cantilever(normal force constant,0.15N m−1)was used to measure normal and lat-eral forces.All the tips were cleaned in a UV chamber(Bioforce Nanosciences,USA)for20min before use.The lateral force was recorded using2μm×2μm scan area.For roughness measure-ment,the scanner was allowed to attain stability before imaging in order to avoid any thermal drift.2.7.1.AFM cantilever calibrationWe used a straight forwardfinite element(FEM)based tech-nique[20]to estimate the torsional or lateral stiffness of the‘V’-shaped cantilever that is used here.This method does not require a ‘multiple cantilever’(where one cantilever is of rectangular geom-etry)(Green et al.[21]),additional mass(Cleveland et al.[22])or a well-defined scanning geometry(wedge calibration method[23]). The normal sensor(NR)response(A/m)is recorded from the repul-sive part of the force distance curve.Known the normal cantilever stiffness(N/m)from the manufacturer and the current geometry during an experiment we estimate the normal force.Writing an angular response AR=NR×L,where L is the length of the cantilever,the twist angleφis recorded as LRA/AR, where LRA is the twist current.Known the real geometry of the cantilever(from SEM images)and the material properties of the cantilever,the FEM(ABAQUS)is used to determine the torsional stiffness T/φ.The torque T is estimated for knownφ.For the present cantilever the manufacturer specifies the normal stiffness =0.15N/m.The FEM estimation gives T/φ=1.6×10−9N m/rad. The ratio of(T/φ)/normal stiffness=10−8m2/rad compares with those reported by Green et al.[21]who used the Sader method (2.2×10−8m2/rad)and that formulated by Cleveland et al.(2.3×10−8m2/rad)to calibrate‘V’-shaped cantilevers.2.8.NanotribometerTribological experiments were carried out in the range25–100mN using a nanotribometer(CSM Instruments,Switzerland). The nanotribometer is composed of three stepper motors(two in X-and Y-axis linked to pin-on-disc module and one in Z-axis linked to measuring head).The cantilever was mounted on the measuring head.A2mm diameter steel ball(rms(root mean square)roughness∼1–2nm)is attached to the end of the can-tilever.Before attachment,the steel ball was cleaned in acetone using an ultrasonic bath.Two optical sensing mirrors placed near to the cantilever head,perpendicular to each other(X-and Z-axis),measure the displacement of the cantilever during sliding against a SAM-coated substrate.The friction coefficient was deter-mined during sliding by noting the deflection of the cantilever in both horizontal and vertical planes.All measurements were carried out in reciprocating mode under ambient conditions(relative hu-midity:42%,temperature:295K)without using any external liquid lubricant.The sliding speed was kept at2mm/s.2.9.Pin-on-discMacrotribological tests were carried out using a pin-on disc machine,procured from DUCOM(Bangalore,India).Theflat face (rms roughness∼200nm)of a high speed steel pin(diameter 3mm)was loaded normally(60N)and pressed against theflat (rms roughness∼500nm)surface of a rotating disc.The disc was slid against the pin situated at6cm radial distance at0.3m/s sur-face speed.The friction force was measured by a load cell attached to the pin holder(resolution0.1N)and the displacement of the pin was measured using a Syscon(Bangalore,India)displacement sensor(LVDT,resolution1μm,range=±500μm).n-Hexadecane (SpectroChem,99%,India)was used as the base oil.5%(v/v)of the fatty acids were dispersed in the base oil.Prior to the actual ex-periment a full pin-on-disc contact was established by running in; load5N speed60rpm.For the lubricated test,drops of oil dis-persed with the additives was continuously added with a burette at the sliding interface.The tests were conducted for90min.2.10.XPS analysisComplementary information on the state of wear track on the metal surfaces was obtained by X-ray photoelectron spectroscopy (XPS).After the tribological tests,the disc was subjected to EDM cutting,where the disc was machined to the dimension of1×1cm with the3mm wear track in the middle.Then the specimens were cleaned with acetone and ethanol and stored in a desicca-tor for XPS analysis.X-ray photoelectron spectroscopy(XPS)mea-surements were carried out on a VG Microtech(UK Instruments) spectrometer equipped with a monochromatic Mg KαX-ray source (hν=1253.6eV)and X-rayflux of0.28nA at10−10Torr.A take710R.R.Sahoo,S.K.Biswas /Journal of Colloid and Interface Science 333(2009)707–718off angle and angle of incidence of 45◦from the surface normal was employed for each sample.The sample chamber had a 5-axis sample manipulator,X ,Y ,Z ,rotate and tilt,and was equipped with a cylindrical analyzer with E =10meV.The high-resolution spectra were recorded with 11.71eV pass energy,1×1mm spot size and were referenced to C1s at 284.6eV.The area intensities were obtained following the spectral fitting and integration with Voigt-like functions.3.Results3.1.Contact angles on SAMsTable 1shows the advancing and receding contact angles of water as measured on fatty acid layers formed on steel and Si sub-strate.Low contact angles suggest a low coverage of the acids on the substrate and loose packing of molecules in the monolayer.The advancing contact angles are lower than the values (>100◦)gen-erally reported for close-packed structures such as that of stearic acid adsorbed from hexadecane on Ag,Cu,and Al surfaces [24]or of palmitic acid on stainless steel where a voltage was applied to facilitate the adsorption of a close-packed layer [25].A contact an-gle of 80◦is reported for stearic acid on Cu [26]where it was noted that the monolayer thickness of 2.3nm does not correspond to optimum close-packing.Water contact angles of ∼82◦and 85◦have been reported for stearic and linoleic acids on steel deposited on Si wafers by sputtering method,the authors [27]claim that fatty acids adsorbed from n -hexadecane on a model steel surface form loose-packed layers where no distinct domains are observed.3.2.FTIR analysis3.2.1.Steel surfaceThe C–H stretching region of the infrared spectrum,obtained using a grazing angle FTIR spectrometer,of stearic and linoleic acid monolayers on stainless steel substrate is shown in Fig.2a.The assignments of the observed frequencies to vibrational modes of these spectra are listed in Table 2.An “ordered”aliphatic monolayer is one with chains in an all-trans configuration [28–30].The FTIR spectrum of such a monolayer is characterized by (νCH 2)anti-sym ∼2918cm −1and (νCH 2)sym ∼2849cm −1.Table 2shows that when the monolayers are adsorbed on steel there is a 1cm −1peak frequency shift away from an ideally crystalline monolayer.The monolayer is however more crystalline when ad-sorbed on silicon wafer.The FWHM of the CH 2antisymmetric peaks however have values of 21.7and 23.1cm −1for LA and SA on steel,respectively.Generally a FWHM value of 18cm −1or less is associated with a very well packed monolayer [31].The filmsTable 1Contact angle of water on adsorbed fatty acid layers.System Deposition time (h)Advancing angle [θa ],(◦)Receding angle [θr ],(◦)Bare steel 48±340±2SA on steel 2482±274±3LA on steel 2485±378±2Silicon (Si)20±216±2SA on Si 3694±386±2LA on Si3682±276±1Table 2IR vibration frequencies of fatty acid SAMs deposited on stainless steel and silicon and their vibrational assignment.The values in the brackets are CH 2antisymmetric peak FWHM.Peak frequencies (cm −1)AssignmentLinoleic acid Stearic acid Steel⎧⎪⎪⎪⎪⎪⎪⎪⎪⎪⎪⎨⎪⎪⎪⎪⎪⎪⎪⎪⎪⎪⎩2980.62981.6CH 3antisymmetric stretching 2919.1(21.7)2919.3(23.1)CH 2antisymmetric stretching 2870.72870.44CH 3symmetric stretching 2848.52848.8CH 2symmetric stretching 1735.941735.83C =O stretching 15401539.18CO −2antisymmetric stretching 1466.51466.12CH 2scissoring 1439.141439.1CO −2symmetric stretching Silicon2917.1(23.4)2917.4(22.8)CH 2antisymmetric stretching 2847.12847.3CH 2symmetric stretchingformed here are therefore definitely loose-packed.Adsorption of fatty acids from non-aqueous solution has been shown to be a fraction of what is possible for a close-packed monolayer [32].Increasing the degree of unsaturation has been found to make the adsorbed layer more close-packed.In the present study the latter effect of unsaturation is not seen.The broadening of a ‘crys-talline’peak observed for the fatty acids adsorbed on steel from hexadecane may also be due to the coadsorption of hexadecane and the resulting multilayer status of the molecules on the steel surface [33].Fig.2b shows the low-frequency region IR spectra of stearic and linoleic acid self-assembled monolayers.Three vi-bration modes associated withthe carboxylic group are seen:a symmetric and an antisymmetric –CO −2stretching at 1439and1542cm −1,respectively,and a C–O stretching mode in a –COOH moiety at 1737cm −1.The presence of these peaks in conjunction with the lack of a C–O stretching mode,characteristic of a hydrogen bonded car-boxylic moiety at 1703cm −1,suggests that the carboxylic head group undergoes partial dissociation to form surface carboxylate species [9].The presence of a symmetric –CO −2mode suggests that the carboxylate headgroup is attached to the surface in a bidentateFig.2.FTIR spectra of the thin film of SA and LA on stainless steel substrate;(a)CH 2region,(b)C–O region.R.R.Sahoo,S.K.Biswas/Journal of Colloid and Interface Science333(2009)707–718711Fig.3.FTIR spectra of the of SA and LA SAM on Si(111)substrate;(a)CH2region,(b)C–O region.configuration.On the other hand,the existence of an antisymmet-ric mode alludes to a monobinding configuration via a single oxy-gen atom[3,34].The relative intensities of these two bands provide a qualitative measure of the surface configuration of the carboxy-late headgroup.Yet,a small fraction of the headgroups does not undergo proton dissociation and are entrapped at the iron oxide–monolayer interface as carboxylic species,evidenced by the peak at1739cm−1.3.2.2.Si surfaceThe C–H stretching region of the infrared spectrum of stearic and linoleic acid monolayers on Si(111)substrate is shown in Fig.3a.SA and LA formed well-orderedfilms with chains in an all-trans configuration(Fig.3a)on Si,as indicated by the CH2sym-metric and antisymmetric peaks at(νCH2)antisym∼2917cm−1and (νCH2)sym∼2847cm−1in the IR spectra.Fig.3b shows the low-frequency region IR spectra of stearic and linoleic acid monolayers on silicon.The presence of C=O stretching in a–COOH moiety at1744cm−1in conjunction with the lack of a C–O stretching mode,a characteristic of the hydrogen bonded carboxylic moi-ety at1703cm−1,suggests that the fatty acids(head group) may undergo chemical bonding to the Si through the carboxylic group.Atomic force microscopy was used to map the topography of SA and LA modified steel surface.The rms(root mean square)rough-ness value of the bare steel substrate measured using the AFM was found to be∼3.5nm.The corresponding values of SA and LA SAM on steel are3.8±0.2nm.There was no evidence of micelle or island formation in thefilm.3.3.Lubricated tribologyWefirst explore lubricated tribology of the test molecules dis-persed in oil,in a pin-on-disc(POD)machine at a mean nomi-nal contact pressure(normal/apparent area of contact)of10MPa. Fig.4shows the coefficient of friction of the stearic acid dispersed oil to decrease with time initially(up to400s)and then to stabi-lized to a time invariant value of0.09.The recorded value of coeffi-cient of friction matches well with those reported in literature[11]. The coefficient of friction corresponding to the linoleic acid dis-persed in oil decreases steadily with sliding time till(4000s)it appears to reach a steady average value of about0.04.At times more than4000s the frictional responsefluctuates over1000s time scale,although the average friction coefficient is at a level of 0.04or less.The reason for this unstable behavior is not clear at thisstage.Fig.4.Coefficient of friction of steel against steel pin,lubricated:hexadecane(HD) and SA and LA additives(5%v/v)dispersed in n-hexadecane,as a function of sliding time;normal load=60N(mean contact pressure p m=10MPa),sliding velocity∼0.35m/s,pin-on-disc tribometer.Error bars shows±σ(standard deviation)from5repeat experiments.3.4.Dry tribologyFig.5shows that when the tribological tests are done on the self-assembled test molecules(without oil)in a nanotribometer, in a pressure range of250–400MPa the friction coefficient is re-duced from that obtained with bare steel.Increasing the mean contact pressure from217to370MPa(load range20–100mN) makes marginal difference to the friction of stearic acid.This is followed by a jump to high coefficient of friction approaching that of bare steel at specific sliding times governed by the normal load. Increasing the load decreases the coefficient of friction of linoleic acid from0.45to0.15.The significant change in friction(LA)occurs in the293to370MPa mean pressure(normal load50–100mN) range where,as in the lubricated pin-on-disc experiment,the co-efficient of friction declines steadily with time to a steady state value of0.15.These pressures are calculated from Hertzian consid-erations.These are realistic estimates as both the ball and the sub-strate were polished to be very smooth(rms roughness∼1–2nm).The lateral force microscopy carried out on the self-assembled films in the970–2500MPa mean pressure range shows(Fig.6) that the friction force corresponding to the stearic acid is higher than that corresponding to the linoleic acid at all contact test pres-sures,when the experiments are done at35%RH ambient but is lower than that of the linoleic acid when the experiments are done712R.R.Sahoo,S.K.Biswas /Journal of Colloid and Interface Science 333(2009)707–718Fig.5.Coefficient of friction vs time for (a)SA and (b)LA monolayers self-assembled on steel (roughness:5–10nm,rms)reciprocating sliding against a steel ball of 2mm diameter and 4nm rms roughness.Measurements were taken at normal loads of 100mN (Hertzian mean contact pressure p m =370MPa),50mN (p m =293MPa)and 20mN (p m =217MPa),2mm/s slidingspeed.Fig.6.LFM friction-load characteristics of the SAMs formed using the (a)SA and (b)LA on steel substrate at ambient and 0%humidity condition.Taking an adhesion overload of 6nN (the average pull off force obtained from the force curves in AFM experiments on the test monolayers)the mean contact pressure is estimated to vary from 970MPa to 2500MPa (normal load range 0–40nN).at 0%RH.The stearic acid friction coefficient is however quite in-sensitive to ambient humidity.3.5.XPS analysisWhen a monolayer is slid against a rigid rough surface,it can dissipate energy by activating vibrational,rotational and defor-mation modes such as bending and stretching.For the additive boundary film to be protective the energy associated with such ac-tivation has to be less than the energy of binding of the molecules to the substrate.An alternate method of protection is for the monolayer to use the work input to chemically transform itself into a stable film on the substrate such that the plane of shear is now internal to the film.The work of Tabor [18]has suggested for ex-ample that fatty acids under traction react with metal substrate to form low friction soaps.We examine the friction results presented here within this perspective.As it was not possible to investigate the surfaces slid in the AFM and the nanotribometer using FTIR or XPS,we examine below the large tracks made on the disc in the pin-on-disc machine experiments,using the XPS.Figs.7and 8show the curve fitted Fe2p,C1s,O1s spectra of the bare steel,the slid tracks generated with for SA and LA addi-tives dispersed in oil and unslid SA and LA SAM on the steel sub-strate.The bare steel surface is characterized by two Fe2p 3/2peaks (Fig.7.1a)at 709.6and 715.6eV assigned to Fe 2+and Fe 3+,respec-tively [9].The Fe2p 1/2peak appears at a higher binding energy of 720.4eV.It is evident that the iron surface is mainly composed of iron oxides such as Fe 2O 3and FeO.The curve fitted high-resolution C1s spectrum (Fig.7.1b)for the bare steel surface is composed of two components:a major one at 283.9eV and another smaller peak at 287.3eV.The former peak is ascribed to the impurities such as carbides and hydrocarbon fragments present on the steel surface.The high binding energy peak is attributed to the carbon involved in bonding with the electronegative oxygen atoms.Two peaks are evident from the curve fitted O1s spectrum (Fig.7.1c)at 528.7eV and 530.8eV.These are associated with the O1s electron binding energy for the bare steel surface.However,while the peak at 528.7eV is attributed to lattice oxygen,O 2−,in the metal oxide,that at 530.8eV is assigned to the hydroxyl anion,OH −.The latter can originate from surface Fe–OH groups.With reference to Fig.7,the spectra obtained from the slid track generated using stearic acid additive is almost identical to that ob-tained from the bare substrate except that the C1s spectrum of the slid track with stearic acid show a peak at 284.9eV (Fig.7.2b).The latter may be ascribed to some residue from the hexadecane。