TribologicalInvestigations UsingFriction ForceMicroscopy

R. Overneyand E. Meyer

~

IntroductionMany attempts have been made in re-

cent centuries to investigate friction, adhe-sion, lubrication, and wear. Most of theexperimental approaches and theories werebased on macroscopic experiments, suchas tensile and indentation tests. For a longtime, only the bulk properties of the mate-rials were considered.

Late in this century a new term wascreated combining all of the above-men-tioned properties which deal with the sci-ence of interacting material interfaces inrelative motion: tribology. The state of theart of science today reveals that process-ing in nature depends strongly on inter-faces that cannot be described only by buJkproperties. Tribologists realize they muststudy the sliding surfaces by analyticalsurface-science tools. With the surface forceapparatus developed by J.N. Israelachviliand D. Tabor,40 we have a surface analysistool that provides new insight into the fieldof macroscopic sliding contact of lubri-cated systems.

After AmontOIlS' laws were established41as a first attempt to describe sIlding fric-tion analytically, theories were advancedover the course of this century. A classicdiscipline was developed: contact mechan-ics. More quantitative treatments of fric-tion were developed by various authors.The energy dissipation in most processesin tribology induced the theorists to con-sider the sliding bodies as spring modelscreating phonon-phonon interactions. Andwith modem computer facilities, theystarted to perform computational experi-ments whenever classical experimentscould not provide information on the su~micron scale.

With the inception of the field of scan-ning force microscopies by G. Binnig,

26

can adhere to the sample and undergosliding start-up effects, known in tribol-ogy as static friction, and dynamic slid-ing effects by scanning. CM. Mate andG.M. McClelland took advantage of thissituation, modified the scanning force mi-croscope, and measured the lateral insteadof the normal movement of the tip.4 Theirexperiment provided the first friction mea-surements on the atomic scale.

With the development of the laser beam-deflection method (see next section),G. Meyer et a1.5 and O. Marti et al.6 pro-duced a single detection setup capable ofproviding simultaneous information onnormal and lateral movement of the tinycantilever sliding over the sample surface.This article seeks to:. highlight the problems of crosstalk onthe topography and torsion signals andpresent a method that allows one to dis-tinguish between topographically inducedand frictionally induced torsion (frictionloap), and. describe the first successful applicationsof the friction force microscope carriedout on lubricating systems, such as organicfilms, also known as boundary lubricants.

These boundary lubricant films are ei-ther physically or chemically adsorbed onthe solid surfaces and their exact micro-scopic nature remains largely unknown.Self-assembled organic films, such asLangmuir-Blodgett (LB) films, serve asmodel boundary lubricants. These filmsare known lubricants, behave in a solidlikemanner, adhere well to substrates, andform ordered two-dimensional structures.The scanning force microscope provides aunique opportunity to study friction andwear resulting from a microscopic asper-ity on these organic compounds.

c.E Quate, and C. Gerber 1 in 1986; a sur-

face analysis technique was introduced,able to probe any surface with a resolu-tion on the atomic scale. Many groupsworldwide have used this new techniqueto investigate surfaces of crystalline andamorphous materials and to examine inor-ganic and organic systems.:I The design ofthis real-space ~ytical tool is based onthe scanning tunneling rnicroscope3(Co Binnig, H. Rohrer, Nobel Prize, 1986),which scans a very fine tip over the samplesurface. The scanning force microscopetip is located on a sensitive cantileverspring. Interactions between the tip andsample cause deflections of the cantilever.These deflections are monitored by a com-puter providing a two-dimensional imageof the sample surface in real space. De-pending on the load of the spring, the tip

Experimental DetailsMate et <;11.4 modified the atomic force

microscope (AFM) to measure the deflec-tion of the cantilever in response to lateralforces (Figure 1). The sideways bending ofthe cantilever (tungsten wire) was detectedusing a laterC\lly. positioned optical inter-ferometer. Wit~\ ~<~!/"'~ early measurementsof lateral tdctiO1;L'rces, it was demon-strated that it is pof,-:;ible to measure lateralforces on a subr:~UI~ scale on graphite4and mica7 (even though the spring con-stant of the tungsten wire was 2,500 N/mand 100 N/m, respectively), This astonish-ing resolution was possible because ofthe sensitive optical sensor (resolution<0.01 run). Since those initial experiments,several groupS8-11 have combined normaland Jateral-force detection.

Combined normal and lateral-force mea-surements can be performed by two dif-ferent methods:

SN!8

t81

Figure 1. Bidirectional AFM Method (A).Normal and lateral bending is measuredby two sensors: SN (normal) and SL(lateral). Typically, the feedback cantrolholds the normal fora! canstant. In thesetup of Mate et al.,4 only the lateralsensor is used.

MRS BULLETIN/MAY 1993

Tribologicallnvestigations Using Friction Force Microscopy

(A) bidirectional measurements withtwo sensOtS, and

(8) bidirectional measurements withone sensor.

Method (A) logically extends the normalforce measurement by bringing a second,

"'[]IA+C

N@

photo

FigllTe 2. Principle of simultJmeous measurement of /WrmJJlmd IRteraI (torsional) forces. Theintensity difference of the upper and lower segments of the photodiode is proportional to thez-bending of the cantilever. The intensity difference between the right and left segments isproportimulJ to the torsion. t, of tM forC£ sensor.

Figrue 3. Cross section of amti/evers ft>r methods (A) and (8): (A): X-llruJ z-defltction Me

measured; (8): z-defledi(Jlt and torsi(Jlt, I, are measured.

MRS BULLETIN/MAY 1993 Z7

~

additional sensor (SL) up sideways to thecantilever (Figure 1). This approach hasbeen implemented by G. McClelland et at 9

with a two-fiber optical interference ultra-high vacuum. (UHV) AFM by G. Neubaueret al.8 with bidirectional capacitance sen-

SOl, and in our group with two tunnelingdetection schemes.12

In measuring the normal deflectionsimultaneously with the torsion (methodB 111.11), only one sensor is required to ac-quire lateral (torsional) and nonnal forceinformation. The reflected beam is moni-tored with a four-quadrant photo diode(Figure 2). Normal bending of the cantile-ver is measured by the intensity difference(1MB - lc+D) of the upper and lower seg-ments of the diode. The signal difference(IMc - 18+0) of the left arui right segmentsprovides torsiooal information. These twomeasurements are performed simulta-neously (Figure 2). Within certain limits(see next section), methods. (A) and (B) areboth sensitive to lateral (frictional) forces.To increase this sensitivity, detection-sensor-specific cantilever shapes are re-quired (Figure 3).

A square or drcular cross section of thecantilever is required for method (A), toenable positioning of a second sensor. Formethod (8), a flat rectangular cross sec-tion of the lever Is required. Also, a longtip is required for method (8), whereas (A)should have a short one. The reasons for

IB+D

laser beam

B

diodethese requirements originate in the formalexpression of the spring constants of arectanguJar beam (Tables] and H).

Method (8) has the advantage of need-ing only one detection sensor, and of be-ln2 convenient in air because the sensor

zreSPoods qujdcly to the motion of the lever.In generaL with the :laser beam deflectionscheme, no feedthroughs are required,as in the optiea) interference AFM, mak-ing the method particularly adaptable to aUHV system. IS

FrictionaJ InformationThe aim in measuring lateral or tor-

sional deflections of the cantilever is toreceive information about the frictionalbehavior of the scanned sample with re-spect to the cantilever tip. In all the detec-tion methods discussed, the deflectionsensor position is fixed. This immobili-zation of the sensor with respect to thedynamics of the cantilever can cause me-chanical crosstalk between torsional, Jata3l,and nonnal movements. It is irrelevantwhether the lateral forces are measured bybending or torsion, as tong as the crosssection of the cantilever is accommodatedin the detection method (d. Figure 3). Thefollowing discussion in lateral-force mea-surement, therefore, will be based on asideways-positioned sensor SL.

As long as the cantilever is gliding overan atomically flat surface, the, bendingforce FM in the gliding direction scales withthe same magnitude as the friction forceFpo The measured Jatera1 bending direction

~~~~

Investigations

is equal to the lateral movement of thelever. In this case, the detector SL acts asa tool to monitor real lateral (frictional)forces (Figure 48).

Scanning over step sites gives rise tohigh torques on the cantilever tip. This, inturn, alters the cantilever's gliding direc-tion, and the static-positioned sensor nolonger measures parallel to the gliding di-rection (Figure 4b).

It would be possible to recalculate thelateral force FF if the exact geometry of thecantilever tip were known. In practice,however, the tip geometry and its exactorientation to the sample surface areunknown. Because of the intrinsic pr0b-lem of a static deflection sensor, frictionalmeasurements, which require force moni-toring parallel to the gliding direction,cannot be accurately performed on stepsites. Instead, indication of the sharpness

28

ForceUsing Friction Microscopy

material specific. Because the cantilever issliding. I1t is a measure for the relativesliding friction I1FSL. Static friction Fsr isgiven at the beginning (1) and (4) of eachforward and reverse scan. The resistanceagainst the sliding direction causes thehysteresis of the friction loop, which com-pares the forward with the reverse scan.The inverse torsional signals of the for-ward and reverse scan is an important in-dicating device for deciding whether atorsional feature is caused by topographyor friction.

Wear Properties ofOrganic Lubricants

In the science of tribology, seIf-assembledorganic films pla~ an important part influid lubrication. 9 Several studies have

demonstrated that the AFM can providean important new view of ultrathin. well-ordered organic OUlltilayer fiIms.2,3UI Withsma1IIoadings, on the order of 10-9-10-8 N,it has been shown to be ~ssibIe to imageLangmuir-Blodgett films.u.D Film thick-ness, mono- and multilayer steps, and thearrangement of molecules in the surfaceof the film have been recorded.14-:r1 TheAFM has also been used as a tool to inten-tionally deform the films, creating featuresof tailored dimensiml. The forces requiredto deform the films are on the order of10-6 N. More important for achieving plas-tic deformation than the norma1 load areintrinsic properties of the film (e.g., pack-ing density), feechck loop frequency, andscan velocity. Wear occurs if the feedbackloop frequency is smaller than some criti-cal value, i.e" when the applied shear stressovercomes the material-specific yield point(see Figure 6).

To decrease the feedback control fre-quency per scan line, the scan velocityand/or the scan area (under constant scan-line frequerq) can be increased. There is,however, an interesting contradiction. De-creasing the scan velocity can also causewear (Figure 7). The sliding velocity is animportant parameter for friction forces.Sokoloff,a for instance, found in his modelthat (for sliding velocities that are smallrelative to the velocity of sound in thematerial), the friction force Ffri< is inverselyproportional to the sliding velocity v, i.e.,

(steepness) of the step is deducted fromthe degree of cantilever twisting.

In order to extract frictional informationfrom the torsion, it is important to com-pare the friction loop (forward and reversescan of the torsional signal) with thetopography. In Figure 5, a forward and re-verse scan line of a stepped surface isschematically drawn. The sample surfaceis assumed to be composed of a homoge-neous material except for the lowest sur-face (e.g., an underlying substrate). Stepheights and widths of the terraces can bemeasured in the topographical mode. Atstep sites the tc:nim t shows outlines as dis-cussed above. In the hcmogeneous part (2),t remains constant on either side of steps.On part (3), where the material is differ-ent, the amount of t changes to anotherconstant value. The difference I1t - t 1 - t 3is independent of the topography and is

where k is a material-specific constant. InFigure 7, the velocity-wear dependence ona mixed L B film is documented. This figureshows that wear is strongly material spe-cific. Whereas the hydrocarbon-containingpart of the film (bright islandHke stnx:turesin Figure 7a) is conspicuously modified

BULlETIN/MAY 1993MAS

Tribological Investigations Using Friction Force Microscopy

(a)

s ..L

(]~(;/.. "-

", ~Sample

~

~ ~

FM = FF

Figure 4. Lilterrd defledion '05. lslleT/rl measurement: (a) gliding over all atomically filltsurface, and (b) gliding over II step (where k is a real constant).

Topography

Figure 5. Fridion loop and ropogrtlphy on a heterogeneouS stepped surface. Terraces (2.) tmd(3) are composed of different materials. In regions (1) and (4), the cantilever sticks to thesample surface because of static friction Fsr. The sliding friction is tt on part (2), and tJ onpart (3). In a torsional-force image, the contrast difference is caused by the relative slidingfriction, AFSL ~ t, - b.

MRS BULLETIN/MAY 1993 29

(ie., plastically deformed), the fluorocarbon-containing part of the film (darker part ofthe AFM image in Figure 7a) remainsdefect-free at any scan velocity.

In the literah1re, an increase in shearstrength with an increase in sliding veloc.ity is usuall~ observed. 8.J. Briscoe andD.c.B. Evans observed a proportional in-crease in friction force with the logarithmof the sliding velocity under constant tem-perature and pressure over the range of0.3-3.4 IJ,III/sec. Their observation agreed

(1».

s ~

cD::"

~ ~

FM:F k Frwith a fust approximation with Eyring'smodel of friction, which expresses the shearstrength as a function of temperature, slid-ing velocity, and applied normal load. TheAFM resuh of variable velocity cannot beexplained by this theory, wheIeas AFMresults of variable applied normal loadingcan be explained, in first-order approxi-mation, as shown by our group.])

In the adhesive theory of friction, fric-tion is explained by asperity junctions.JIIn these theories, the mechanics of station-ary contact is extrapolated to the moving-contact situation. The time step of funningjunctions is either infinite or zero, ie.. trereis or there is not a junction. A dynamicprocess, however, is time dependent. There-fore, the real area of contact, A, that isformed by the asperity junctions shouldbe time dependent, i.e.,

A = A(t)

with

A(t+At) > A(t);

and the kinetic friction Fkin should passinto the static friction F..ti<, i.e.

F~(A(t»~ F...The above AFM result of increasing wearby decreasing the sliding velocity sup-ports these assumptions. With this expla-nation, static friction could be re~rded asa limiting case of kinetic friction.

Inter- and intralayer interaction canbe compared on one- and two-bilayer Cd-arachicfate Lan~~ir-Blodgett films by in-tentionally modllying the film surface andalso by comparing the stacking behaviorof two- or tour-layer films. Steps in one-bilayer films ~re mostly qne monolayer inheight (-27 A), and tw~bilayer films, onebilayer in height (-54 A) (Figure 8).

The applied £onE and the feedback-loopfrequency required to disrupt the inter-faces Vij (i = 1,2,3 andj - 2,3,4), sche-matically represented in Figure 9, allow aranking of the interJayer (V:D,V3t) and sub-strate-film interaction VGt wlUch is:

V31 < V23 < VU.

Tribologicallnvestigations Using Friction Force Microscopy---

Figure 6. IntentiomUly created hole in afour-layer LB film. 2.5 X 2.5-p.ml AFMimage of Cd-arachidate film. A quadratichole hils been created by increasing thesam spml from as p.m/ se£ to 5 p.m/ se£(the load was also increased by about afactor of 10, to 10-7 N). Damage wasfound to stRrt at step edges ukre thetorsion is strong. Laterally increasedshear stress in the film can be observed inthe vie/nity of the hall, where the densityof pores is d.ecreJlStll.

This scratching experiment corresponds tonondestructive measurements on varioussample sites, which show that:. the interaction between the poW heads,V 23, is stronger than the interaction betweenthe hydrophobic tails, V 34. This follows fromthe ciomirtanre of bilayer steps in the four-layer system; ~. the substrate-film interaction, Vu, is thestrongest of all three of the mentionedinterface interactions. This statement isdeduced from the phenomenon that, inthe one-bilayer system, mostly monolayersteps and rarely uncovered substrates arefound.

Whereas it is easy to scratch the first bi-layer (counted. from the film-air interface)of the two-bilayer system, it is difficult toremove the film entirely and to uncoverthe silicon substrate. Entire sheets are re-moved in the saatching process. This in.dicates larger intralayer interactions thaninterlayer interactions.

The ranking of the interactions can beestimated by the strength of the actingpotentials. Attractive hydrophobic foo:esare acting between the tails of the mole-cules. They are much stronger than thevan der Waals attraction at small sepa1'It-tion and are of surprisingly long range. ~

Hydrophilic interactions are repulsive andform, together with the hydrophobic in-

30

matjon of strong ionic bonds is responsible.Future experiments with pure arachidicacid will clarify whether the ranking ofinterlayer forces is as pronounced as in thecase of the soap Cd-arachidate.

teractions, some nonadditive net intralayerinteraction. Cd ions- surfactants at theheads of the amphiphilic moleOlles-havethe effect of stabilizing the film. G The hy-

drophobic interaction V12 is expected to bevery strong. Considering the length of themolecules, it seems reasonable that the in-Ualayer interaction dominates the inter-layer interaction, as concluded by the AFMresults. The ranking V34 < V2'!, however,cannot be expJained by these two drivingforces. It might be possible that the saponi-fication (hydrolysis of esters into acids andalcohols) of the films that leads to the for-

Friction Measurements onOrganic Lubricants

Friction measurements on. the above-presented multilayers of Cd-arachidate LB films on silicon (for experi-mental details see Reference 3D), and. phase-separated LB films of mixtures ofhydrocarbons and fluorocarbons on silicon

b)a)

Figure 7. Structum created in mixed rnono/Jzyer film by low-speed SClinning: (Il) 2.5 X2.5-pml AFM image of II 1:1 mixture of fluorocarbons (dark) and hydrocarboNs (bright)."Smileys" were written with the AFM by deaeasing the scan speed from -1 pm/see too.D1 pm/SIX. Iltukr these conditions, only the isl4ndlike structum (hydroalrbons) could bescribed by the AFM cantileuer tip. The fluoroc:arbtm-containing ptlrt of the film could not bedestroyed by scan-speed reduction. (b) Friction-force map of the same region. Brighter areashave higher friction.

E.s..N

00 20002000x(nm) x(nm)

Figure 8. Line trllCe of the single and double bi14yer films. (II) MonoliJyer steps in tM bilayerfilm. (b) Mainly tJilayer steps are found in the double bilayer films.

MRS BULLETIN/MAY 1993

Tribologicallnvestigations Using Friction Force Mic

(a)

(b)

Figure 9. Schematic drawing cf thestepped Cd-arachidate films; (a) one-bilayer Cd-arachidate film, and (b) two.bilayer Cd-arachidate film. The relativestrength of in~riRyer potentials, V, aredetermined with the AFM (54' text). Thesilicon substrllte was hydrophobized.JJ

(for experimental details see Reference 43),provide material-specific identification.The studies were carried out on two-bilayerfilms on hydrophobized silicon wafers.

Figure lOa is a top-view image of a Cd-arachidate bilayer fihn. The image containsthree different levels: The first level (1)(dark) is the silicon substrate, the secondlevel (2) (dark grey) js the fu& liIayer system.(step height from the substrate 54 A),and the third level (3) (bright) is the sec-ond bilayer system (108 A). In the lateralimage (friction image), Figure lOb, thereis an evident contrast difference betweenlevel (1) and the other two levels. Togetherwith the topographical information of Fig-ure lOa, it is obvious that the film-coveredpart of the sample shows decreased rela-tive friction of one order of magnitude(Figure IDe). The friction loop (Figure 1Od)indicates increased lateral forces at step

MRS BULLETIN/MAY 1993

pyrase 0

able normal forces show no functionaldependence on the applied normal load(Figure 11). The investigations, however,are performed only in the range of 1-10 nNand at low speeds of less than 1 p.mJs.Wear processes start at forces above athreshold of about 10 oN, primarily at filmedges. The above observation of AFM-induced shear at film is1a1'ld; confirms theassumptioni9 that wear does not evolvecontinuously; instead, it starts at a criticalforce of 10 nN. The collective motion ofmolecules is preferred and the size of thesheared particles is detennined by the ap-plied shear stress.

At loads of about 10 nN, before the on-set of wear, increased lateral forces at step

sites, arising mainly from a crosstaIk £romtopography.

The lubricating property of such organiclayers is demonstrated with the AFM, andthe factor of 10 differen:e in lubricated andunlubricated friction agrees with macro-scopic-scale friction measurements.34

No contrast difference can be observedbetween the two levels of the film (firstand second bilayers). Within an uncertaintyof 10%, the shear strength does not dependnTl thp thiC'knp!!!; nf thp!!p fjJmR. whether

they are a two- or four-layer system. But itis expected that multilayer films of about10 layers or more should provide differ-ences in shearing.

Friction force measurement under vari-

d)

Figure 10. Bidirectional meaSUTeTnent of a faur-layer Cd-arachjdate LB film. (II) 2.0 X 2.O-pmltopographical AFM image. The silicon substrate (dark), (level 1); olle-biiayer system (darkgrey), (!evei 2); and two-bilayer system (bright), (level 3) are imaged. The step heightsart 54 A. (b) Simultaneously measured friction-force mtlp shows increased friction (lightareas) on the substrate Q$ComptlTed to that on th organic films. No diffrrence in jridion ismeasurable between single- and double-bilayer surfttces. At step si~, an increased lateralsignal caused mainly by topographical crosstalk, is documented. (c) Forward (solid line) andreverse (dotted line) scan lines of topography. (d) Friction force loop, forward (solid line) andrtVerse (dotted line) scan lines of friction. The differtlW: bef'1.o«n the upper and lower curve.divided by two, yields the fridiO1Ul1 fora. The frirtion force of the LB-covered ptlrt is 0.2 nN,

and on the uncovered silicon surface is 3 nN. Applied ncrmnlload is 3 nN.

31

Tribological Investigations Using Friction Force Microscopy

......zOJ>

0,...-."u I

0 2 4 6 8 10 12 14 16

Figure 11. Latml-.force depetulmce on loading F". Lateral forces Fl are mtaSUTtd (1) on flat,film..covered areas, (2) at step edges, and (3) on the substrate. On film-covered areas, thelateral forces remain constant. Lateral forces on the substrates and at the step edges increaseas a function of the applied normallGad.

sites are measured (Figure D). As alreadymentioned above, lateral forces are higherat step edges. Varying the applied normalload in the range of about 2-:16 nN the fric-tional forces at step sites varies betweenfour and 12 oN. This increase in frictionalre&stance is due to nonideal force controlat step edges (see the beginning of thissection) and is also due to the increase, onimpact, in contact area with the edge ofthe upper bilayer. On the silicon substrate,lateral force measurements increase as afunction of normal forces.

Figure 12a shows a 5.0 X 5.0 p,m2 top-view AFM image of a 1:1 molar mixtureof arachidic acid and partially fluorinatedcarboxylic acid bound ionically to a com-mon cationic polymer (see schematic repre-sentation of the carboxylates in Figure 13).Round islandlike structures, 100 nm to1 p.m in diameter and 1.6 om in height,above a surrounding sealike film are 0b-served. The higher circular domains inFigure 12a are assigned to the hydrocar-bons and the surrounding flat film to thefluorocarbons, based partly on their dif-ferent molecular lengths and partly ontheir different frictional behaviors (describedlater in this section). With molecular Iengthsof - 2.5 om and - 2.0 run for the arachidic

acid and the fluorocarbon-terminated

32

grc:phy ~ents (Figure 12a) roughlyindicate friction that is four times higherover the fluorinated regions than over thehydrocarbon regions. Holes in the film r~veal the silicon substrate (Figure 12b). Thesilicoo, appearing as bright areas of highcontrast in the lateral image, displays afriction force that is a factor of ten higherthan the hydrocarbon domains. This differ-ence in friction is in good agreement withother experiments performed on both themacroscopic and nanometer scales.26.Zi',36 Insum, from the AFM measurements onthese phase-separared LB films, the rela-tive friction oillie hydrocarbon, fluoro-carbon. and silicon surfaces are shown tobe 1:4:10.

The differences in friction are. not dueto changes in topography. This point isdemonstrated particularly well by the abil-ity of the friction measurements to identifythe scattered islands of extraneous materialsitting atOp hydrocarbon domains. Where-as by normal-force measurement, only thegeometry of the island is determined, bylateral-force measurement, the composi-tion of the island is determined. The fric-tional response of the island correspondsto that of the circular hydrocarbon do-mains, thereby allowing it to be identifiedas hydrocarbon in nature. This assign-ment would not be possible with the nor-mal AFM alone.

A surprising result is that the frictionon the fluorinated areas is higher thanthat on the hydrocarbons. From the per~formance of fluorin~ontaining lubri-cants, such as Teflon* (PTFE), a reductionin friction might be expected. However,it is known from surface force apparatus(SPA) experiments29 that fluorinated LBfihns have larger shear strengths than their

hydrogen-containing counterparts.These observations of friction and r0-

bustness on the scale of nanometers canbe applied to the study of tn"bology, par-ticu1arly boundary lubrication. Boundarylubrication, as established by Hardy,37 dealswith the lubricating effect: of the layers oflubricant in closest proximity to the solidsurfaces undergoing frictional contact, andis of far-reaching importance in mostsurface-on-surface sliding mechanisms.Tre particu1ar advantage of the fluorinatedfilm is its resistance to rupture, as knownfrom both the SFA29 and AFM measure-ments. Therefore, the good performance ofthe fluorocarbon films as lubricants canbe traced to an excellent stability in thepresence of applied stress and a reducedhiction (compared to unlubricated surfaces),which create a reduction in wear to lowvalues (reduction by a factor of 10,000).38

FN [10 .9N]

add, respectively, a difference in height of-1.0 nm between domaiIls of the two com-ponents in this bilayer LB system is antici-pated. Figure 12a shows a step height fromsea to island of 1..6 rnn, a slightly greaterdifference that could be due to a greatertilt angle in the fluorocarbon domains.3S

In a separate series of experiments,AFMimages have been recorded on LB bilayerfUms composed of a single carboxylic acidcomponent: one from arachidate and onefrom the partially fluorinated carboxylate.The fluorinated acid film contains fewerdefects than the arachidic acid film. Theseresults support the assignment of the cir-cular domains to the hydrocarbon c0mpo-nent, since holes are only observed in thecircular domains in the mixed films. Theseholes are -5.0 nm in depth, consistentwith the thickness of a bilayer. Further-more, increasing the force while scanningdisrupts the hydrocarbon but not the fluo-rinated areas, in both the one- and two-component films. The fluorinated sit~sshow good resistance to rupture duringsliding, which is in agreement with previ-ous tribology experiments in the litera-ture, usin% macroscopic measurementtechniques.

Lateral-force measurements (Figure 12b)performed simultaneously with the topa-

MAS BULLETIN/MAY 1993

~~~

Tribologicallnvestigations Using Friction Force Microscopy

Figure 12. Simultaneous normal and friction nreasurements on a bilayer mixed LB film.(Il) 2.5 X 2.5-p.m2 AFM image of the surface of the bilayer prepared from a mixture of thefluorocarbon and hydrocarbon carboxylates ('1:1 molar). The image represents the topography.The circular domains are assigned to the hydrocarbon component and the surrounding flatfilm to the partially fluorinated component. The difference in height surrounding flat film tothe partially fluorinated component. The difference in height between these two regions is-1.6 nm. Holes of about 5 nm in depth are oolyobserved in the islandlike circular hydro-carbon domains, whereas the fluorinated film remains fairly uniform and continuouslyunbroken. (b) Simulataneously measured friction-force map. The lateral-force image indicateshigher friction (brighter contrast) over the fluorinated regions. Highest friction is mea!JUredon the silicon !JUrface (at the bottom of the holes in the film). No difference in contrast

(friction) is observed between the hydrocarbon layers of different heights.(c) Three-dimensional display topographical image.

MRS BULLETIN/MAY 1993 33

Figure 13. Schtmlltic view of tIft LBbilayer system imaged by AFM in Figure12. The h represents the htight differenceof -1.6 nm between the hydrocarbon and

fluorocarbon species (see text).

ConclusionsThe AFM is well-suited for studying wear

and frictional behavior by simultaneous,bidirection!il normal and lateral measure-ments. It has been shown that complexphenomena of boundary lubrication canbe studied. Material-distinct identificationcan be performed and film inhomogenei-ties exposed. Wear studies on LB filmsreveal an interesting-functional behaviordepending on sliding velocities. This newtechnique allows a variety of tribologicalstudies on all kind of materials, for example,inorganic compounds, that are most likelyinvestigated under ultrahigh vacuum con-ditions to avoid ambient film formation onthe surface; ~1B the origin of the resistanceto sliding, such as elastic properties of thematerials;'13 and interaction forces.3:I

AcknowledgmentsWe wish to thank J. Frommer for help-

.tul and stimulating discussions andM. Fujihira for the preparation of someof the samples and very fruitful discus-sions. This work was supported by theSwiss National Science Foundation andthe Kommission zur Forderung delwissenschaftlichen Forschung;

References1. G. Binnig, c.E Quate, and C. Gerber, Phys.Rea LeU. 56 (1986) p. 930.

Tribologicallnvestigations Using Friction Force Microscopy

2. For recent reviews of scanning force micro-scopies, see: E. M~yer and H. Heinzelmann,Scanning Tunneling Microscopy and RelatedMethods, edited by R.J. Behm (Kluwer, Dor-<hecht, 1990) p. 443; J. Frommer and E. Meyer, j.Phys. Condo MDtt. 3 (1991) p. 51; Ultramicroscopy42-4 (entire volumes devoted to scanning tun-neling microscopy and scanning force micro-scopies) (1992); W. Heckl, Thin Solid Films210-211 (1992) p. 640; and J. Frommer, Ange-'WfJndte Chemie, Int. Ed. Eogl. 31 (1992) p. 1298.3. G. Binnig and H. Rohrer, Heln Phys. Acta 55(1982) p. 726; and G. Binning, H. Rohrer,Ch. Gerber, and E. Weibel, Phys. Rev. Left. 49(1982) po 57.4. CoM. Mate, G.M McClelland, R. Erlandsson,and S. Chiang, Phys. Reu Lett. 59 (1987) p. 1942.5. G. Meyer and N.M. Amer, Appl. Phys. Left. 57(1990) p. 2089.6. O. Marti, J. Colchero, and J. Mlynek, Nano-technol. 1 (1990) p. 141.7. R. Erlandsson, G.M. McClelland, CoM. Mate,and S. Chiang, j. Vae. Set. Technol. A. 6 (1988)

p.266.8. G. Neubauer, S.R. Cohen, G.M. McClelland,D. Horne, and C.M. Mate, Reu Sci. Ins/rum. 61{199O} p. 2296.9. G.M. McClelland, Adhesion and Friction, ed-ited by M. Grunze and H.J. Kreuzer (Kluwer,Amsterdam, 1992) p. 81.10. G. Meyer and N.M. Amer, Appl. Phys. LeU.57 (1990) p. 2089-11. O. Marti, J. Co1chero, and J. Mlynek, Nano-technol. 1 (1990) p. 141.12. R.M. Overney and E. Meyer (private com-munication).13. R.M. Overney, PhD thesis, University ofBaseL 1992.14. Park Scientific Instruments, 1171 BorregasAvenue, Sunnyvale, CA 94089-15. O. Wolter, T. Bayer, and J. Grechner,]. Vac.Sci. Technol. B 9 (1991) p. 1353; O. Wolter, Institut

34

fUr Mikrostrukturtechnik und Optoelek- D. Brodbeck, R. Luthi, and H-J. Guntherodt,

tronik, Wetzler BlankenfeldiGermany. Fundamentals of friction, edited by 1. Singer and

16. KE. Petersen, "Silicon as a Mechanical Ma- H. Pollock (I<luwer, Dordrecht, 1992) p. 427.terial,H Proc. IEEE 40 (1982) p. 420. 28. J.B. Sokoloff, Phys. Ren B 42 (1990) p. 760.17. E. Meyer, R Ovemey, R Luthi, D. Brodbeck, 29. B.J. Briscoe and D.C.B. Evans, Proc. R. Soc.L. Howald., J. Frommer, and H-J. Guntherodt, London, SeT. A 380 (1982) p. 389.Thin Solid films 220 (1992) p. 132. 30. E. Meyer, R. Overney, D. Brodbeck,18. L. Howard., E. Meyer, R. Luthi, H. Haefke, L. Howald, R. Luthi, J. Frommer, and H- J.R. Ovemey, H. Rudin, and H-J. Guntherodt, Giintherodt, Phys. Reu Lett. 69 (1992) p. 1777.submitted to Applied Phys. Lett. (1993). 31. EP. Bowden and D. Tabor, Friction and Lu-19. S;M. Hsu, MRS Bulletin XVI (10) (1991) brication of Solids, Part 1 (Clarendon Press, Ox-p. 54. ford, 1954).20. For recent reviews of thin organic films, see: 32. J.N. Israelachvili, Intermolecular and SurfaceA. Ulman, An Introduction to Ultrathin Organic Ft1t'ces, 2nd ed (Academic Press, London, 1991).Films, From Langmuir-Blodgett to Self-Assembly 33. M. Schreck, D. Schmeisser, W. Gopel,(Academic Press, San Diego 1991); and J. Swalen, H. Schier, H.U. Habermeier, S. Roth, andAnnual Rea Milter. Sci. 21 (1991) p. 373. L. Dulog, Thin Solid films 175 {1989} p. 95.21. R.M. Overney, E. Meyer, J. Frommer, H-J. 34. Al. Bailey and J.S. Courtney-Pratt, Froe. R.Giintherodt, G. Decher, J. Reibel, and U. Sohling, Soc. London, SeT. A 2Z7 (1955) p. 501.Langmuir 9 (1992) p. 341. 35. C. Naselli, J.D. Swalen, and J.F. Rabolt,22. E. Meyer, L. Howald, R. Overney, H. ]. Chern. Phys.90 (1989) p.3855-Heinzelmann, J. Frommer, H-J. Guntherodt, 36. B. Briscoe, D. Tabor, Interfacial PhenomenaT. Wagner, H. Schier, and S. Roth, Nature 349 in Apolar Media, edited by H-F. Bicke, G.D.(l99l) p. 398. Parfitt (Marcel Dekker, New York, 1987) p. 327.23. R. Overney, E. Meyer, J. Frommer, 37. WB. Hardy, Proc. R. Soc. London, Ser. A 88D. Brodbeck, R. Luthi, L. Howald, H- (1913) p. 313.J. Giintherodt, M. Fujihira, H. Takano, and 38. E. Rabinowitz, D. Tabor, Proc. R. Soc. Lon-Y. G etch, Nature 359 (1992) p. 349; and don, SeT. A 208 (1951) p.455.

. E. Meyer, R. Ovemey, R. Luthi, D. Brodbeck, 39. A. Wadas and RM. Overney (private com-L. Howald, J. Frommer, H- J. Guntherodt, munication). .

O. Wolter, M. Fujihira, H. Talamo, and Y. Gotoh, 40. J.N. Israelachvili and D. Tabor, Proc. R. Soc.Thin Solid Films 220 (1992) p. 132. London, Ser. A 331 (1972) p. 19-24. L. Bourdieu, P. Silberzan, and D. Chatenay, 41. D. Dowson, History of Tribology (Longman,Phys. Rea Lett. 6 (199l) p. 2029; and C. Alves, London, 1979),E. Smith, and M. Porter, J. Amer. Chern. Soc. 114 42. E. Meyer, R. Ovemey, R. LUthi, D. Brod-(1992) p. 1222. beck, L. Howard, J. Frommer, H-J. Giintherodt,25. H. Fuchs, L. Chi, L. Bng, and K Graf, Thin O. Wolter, M. Fujibira, H. Takano, a11d Y. Gotoh,Solid Films 210-211 (1992) p. 655. Thin Solid Films 220 (1992) p. 132.26. E. Meyer, L. Howald, R. Overney, 43. R. Overney, E. Meyer, J. Frommer,H.HeinzeJmann.J.Frommer,H.-J.Guntherodt, D. Brodbeck, R. Lilthi, L. Howard,Ultramicroscopy 42-44 (1992) p.274. B-J. Giintherodt, M. Fujihira, H. Takano, and27. E. Meyer, R. Overney, L. Howald, Y. Gotch, Nature 359 (1992) p. 349. . 0

MRS BULLETIN/MAY 1993