Friction and wear of ceramics

Friction and wear of ceramics

Wear, 100 (1984) FRICTION DONALD 333 333 - 353 AND WEAR H. BUCKLEY Lewis Research Center, OH 44135 (U.S.A.) OF CERAMICS and KAZUHISA National...

2MB Sizes 7 Downloads 124 Views

Wear, 100 (1984)

FRICTION DONALD

333

333

- 353

AND WEAR

H. BUCKLEY

Lewis Research Center, OH 44135 (U.S.A.)

OF CERAMICS

and KAZUHISA National

MIYOSHI

Aeronautics

and

Space

Administration,

Cleveland,

Summary The adhesion, friction, wear and lubricated behaviors of both oxide and non-oxide ceramics are reviewed. Ceramics are examined in contact with themselves, other harder materials and metals. Elastic, plastic and fracture behavior of ceramics in solid state contact is discussed. The contact load necessary to initiate fracture in ceramics is shown to be appreciably reduced with tangential motion. Both friction and wear of ceramics are anisotropic and relate to crystal structure as with metals. Grit size effects in two- and three-body abrasive wear are observed for ceramics. Both free energy of oxide formation and the d valence bond character of metals are related to the friction and wear characteristics for metals in contact with ceramics. Surface contaminants affect friction and adhesive wear. For example, carbon on silicon carbide and chlorine on aluminum oxide reduce friction while oxygen on metal surfaces in contact with ceramics increases friction. Lubrication increases the critical load necessary to initiate fracture of ceramics both in indentation and with sliding or rubbing.

1. Introduction Considerable effort has been put into fundamental studies of adhesion, friction and wear of metals. In recent years the increasing potential for the use of ceramics as components in lubrication systems has focused attention on these materials. Adhesion, friction and wear studies have been conducted with ceramic materials to understand better their physical and chemical properties which will affect their behavior when in contact with themselves, other ceramics or metals. In general metals readily deform plastically whereas ceramics, while having high strength, are normally brittle and fracture with little or no evidence of plastic flow. At the interface between two ceramics in solid state contact under load and relative motion, plastic flow has been observed in the surface layers of some ceramics. Plastic flow has been observed with magnesium oxide [ 11, aluminum oxide [ 21 and silicon carbide [ 31, for example, under relatively modest conditions of rubbing contact. 0043-1648/84/$3.00

0 Elsevier Sequoia/Printed

in The Netherlands

334

Those factors which influence the mechanical behavior of materials when plastic flow occurs such as dislocations, vacancies, stacking faults and crystal structure will therefore influence the friction and wear behavior of ceramics (e.g. A1,03) [4] and comparisons can be made between the anisotropic friction behavior of ceramics and metals. The presence of surface films such as adsorbates markedly influences the adhesion, friction and wear behavior of metals and ceramics as well. Further, with ceramics and other ionic solids, the presence of surface films such as water and surface-active organics can influence adhesion, friction and wear by altering the amount of plastic defo~ation that will occur during sliding or rubbing [ 5,6]. The objectives of this paper are to review the adhesion, friction and wear of ceramics, anisotropic friction and wear behavior and the effect of surface films and ceramic-metal interactions. Analogies with metals will be made where applicable. Both oxide and nonoxide ceramics will be discussed.

2. Friction Ceramics behave much like metals when they are brought into contact with solids. For example, when a silicon carbide surface is placed in contact with a diamond under relatively low contact pressure, elastic deformation can occur in both the silicon carbide and the diamond. Sliding occurs at the interface. No groove formation due to plastic flow and no cracking of silicon carbide with sliding are observed [ 3 1. Under the foregoing condition, friction is a function of the shear strength of the elastic contact area, as indicated in Fig. l(a). That is, the relation between coefficient of friction p and load W is given by an expression of the form

The exponent arises from an adhesion mechanism, the area of contact being determined by the elastic deformation. Over the entire load range, the mean contact pressure is 1.5 X 103 - 3.5 X 10’ N mm-*. The maximum pressure at the center of the contact area according to Hertz will be 2.3 X 103 - 4.9 X lo3 N mm-*. A similar friction characteristic for diamond on diamond was also found by Bowden and Tabor [ ‘71. A large increase in applied contact pressure, however, results in a complete reversal in friction characteristic with an applied load. Figure l(b) reveals an entirely different mode of deformation and energy dissipation with an estimated maximum hertzian contact stress ranging from 1.4 X lo4 to 3.0 X lo4 N mm-*. Plastic deformation occurred in silicon carbide, causing permanent grooves during sliding with very small cracks being generated in the silicon carbide. The diamond indented silicon carbide without itself suffering permanent defo~ation. The frictional energy

335

A

(b)

.050W5

(cl LOAD.N

Fig. 1. Coefficient of friction as a function of load for hemispherical diamond riders of different radii and a conical diamond rider sliding on a single-crystal silicon carbide (0001) surface (sliding velocity, 3 mm min-‘; room temperature): (a) elastic contact (diamond radius, 0.3 mm); (b) plastic contact (diamond radius, 0.02 mm); (c) sliding accompanied by gross fracture (conical diamond with an apical angle of 117”).

dissipated during sliding is due to shearing at the interface and plastic deformation of the silicon carbide, i.e. ploughing of silicon carbide by the diamond. The relation between coefficient of friction ~1and load W now takes the form cc = k W”*3-o*4[3]. The exponent depends on the crystallographic orientation of the silicon carbide. When a much higher contact pressure due to high concentrated stress in the contact area is provided, the sliding action produces gross surface and subsurface cracking as well as plastic deformation [3]. Under such conditions wear debris particles and large fracture pits due to cleavage are observed. The area of a fracture pit is a few times larger than that of the plastically deformed groove. In this case, the coefficient of friction is much higher (four times or more) than those in elastic and plastic contacts, as indicated in Fig. 1. Although fracture and plastic deformation in silicon carbide are responsible for the friction behavior observed, most of the frictional energy dissipated during sliding is due to fracturing of the silicon carbide. The coefficient of friction is strongly influenced by the bulk properties of the silicon carbide such as crystallographic orientation and the presence of grain boundaries [ 3,8]. 3. Wear In general there are two types of wear encountered with ceramics. These are adhesive and abrasive. Adhesive wear occurs when adhesion takes

336

place across an interface between two ceramic surfaces or a ceramic surface and a metal. With tangential motion, i.e. the sliding or rubbing of one surface over the other, if fracture occurs in the ceramic, adhesive wear has taken place. For this type of wear to exist, adhesion must first occur. Secondly, the fracture strength of one of the two materials in contact must be less than that of the interfacial junction. If bonding in the interfacial junctions is less than that in either of the two materials, fracture will occur at the interface with theoretically no wear occurring. When an atomically clean silicon carbide surface is brought into contact with a clean metal surface, the silicon carbide deforms mostly elastically. High pressures, however, are developed very locally in the regions where real contact occurs at the interface. The adhesive bonds formed at silicon carbide-to-metal interfaces are sufficiently strong that fracture of cohesive bonds in the metal and transfer of metal at the silicon carbide surface result [ 9, lo]. The presence of surface and subsurface defects such as dislocations, vacancies, voids, impurities and microcracks in the surficial layers of the materials in contact will generally dictate the zones from which wear damage and wear particles are generated. The extent and distribution of such defects will, to a large extent, determine the size of the wear damage and wear particles generated. This is observed in the scanning electron microscope. Figure 2 presents a scanning electron photomicro~aph of the wear track on silicon carbide generated by ten passes of sliding of a 6.66at.~W-Fe alloy pin across the silicon carbide surface. Fracture cracks develop in the silicon carbide during multipass sliding. Fracturing is primarily due to

Fig. 2. Fracture pits of single-crystal silicon carbide in contact with 6.66at.%W-Fe alloy as a result of ten passes of the rider in vacuum (10 -8 Pa) (scanning electron photomicrograph of wear tracks on &con carbide {OOOl} surface; sliding velocity, 3 mm min-‘; load, 0.2 N ; room temperature}.

337

cleavage cracking on the surface and in subsurface regions. The cracks for silicon carbide which were observed in the wear track primarily propagate along cleavage planes of the { lOi0) and { OOOl} orientations. Similar fracture pits and multiangular wear debris, having crystallographically oriented sharp edges, have been observed with ferrite-ferrite contacts, ferrite-metal contacts and silicon carbide-silicon carbide contacts [9 - 111. The fracture behavior of the Mn-Zn ferrite crystal during sliding was similarly found to be significantly dependent on the cleavage systems of the {llO} planes. Abrasive wear occurs when two surfaces, one considerably harder than the other, are brought into contact. If a hard particle or a hard asperity of approximately spherical shape is in contact with a ceramic under load, the ceramic initially deforms locally according to the well-known hertzian elastic equation [ 121. Fracture with ring cracks has been produced in flatsurfaced ceramics critically loaded either statically or dynamically with hard spheres [13 - 181. Indenting with a hemispherical diamond indenter on a silicon carbide (0001) surface also results in the formation of circular (ring) cracks as well as in elastic and plastic deformation [ 191. Figure 3(a) presents a scanning electron photomicrograph of the permanent indentation and the surrounding circular cracks generated by a hemispherical diamond indenter of 0.02 mm radius. The plastic deformation is accompanied by nearly perfect circular cracking in the silicon carbide. Figure 3(a) also reveals cracks propagating and expanding radially from the center of the contact circle. The cracks form preferentially on the planes of easy cleavage in the silicon carbide. The photomicrographs clearly show several slip lines, accompanying plastic deformation, in the indentation. The slip lines are crystallographically oriented and are in the (1120) directions. If a tangential force is applied to the normally loaded hemisphere, the additional tangential stress compresses the ceramic and opposes the tensile tangential stress due to the normal load at the leading edge of the rider. Conversely at the trailing edge the tensile tangential stress is augmented. As a result, cracking will be restricted to the rear of the contact region and suppressed at the leading edge. Cracks of this type have long been observed when a hard hemisphere slides over the surface of a brittle solid or ceramic [20 - 231. This is demonstrated in Fig. 3(b), with the presentation of a scanning electron photomicrograph of a wear track on silicon carbide generated by a hemispherical diamond rider of 0.02 mm radius. Plastic deformation occurs in the silicon carbide. Two kinds of cracks are observed in Fig. 3(b). One occurs in the wear track and propagates perpendicular to the sliding direction. The other is primarily observed on both sides of the wear track, propagating outward from the wear track. The two effects observed and of potential importance for sliding contact will be considered and these are as follows. (1) The catastrophic cracks may grow easily by application of a shearing force during sliding.

338

Fig. 3. Scanning electron photomicrographs of indentation and wear track on singlecrystal silicon carbide (0001) surfaces generated by a hemispherical diamond indenter and rider (0.02 mm radius): (a) indentation (load, 5 N); (b) wear track (load, 2 N).

(2) The deformation and fracture behavior of the ceramic material is very dependent on crystallographic orientation, as observed in friction behavior. With a diamond hemisphere of 0.02 mm radius, circular cracks formed on the silicon carbide surface when the normal load exceeded 2 N in the indentation process. However, if a tangential force was applied, cracking occurred at the rear of the contact region for a normal load of 0.4 N [3]. Further catastrophic fracturing occurred in silicon carbide at a normal load of 2 N, as indicated in Fig. 3(b). Therefore relatively small ~gent~~ stresses are very potent in producing conditions for fracture at the surface. A similar fracture characteristic for Mn-Zn ferrite oxide ceramic was also found [ 22,191.

339

RADIUS OF RIDER,MI

Fig. 4. Critical normal load to fracture of Mn-Zn ferrite as a function of radius of indenter and rider in indentation ((100) surfaces) and in sliding contact ((100) surface, (001) direction; (100) surface, (011) direction).

Figure 4 indicates the critical normal load to fracture of Mn-Zn ferrite as a function of radius for both indenter and rider in indentation and in sliding contact. The load to fracture is directly proportional to the radius of the indenter and rider. The tangential force introduced by sliding plays an important role in the generation of surface fracture. With indentation, the cracks produced in the single-crystal Mn-Zn ferrite surfaces were not circular (ring) cracks, but rather extended in radial directions from the indentation. The cracks generated in sliding propagated perpendicular to the sliding direction in the wear track. When a number of grits of hard abrasive material embedded in a resin matrix come into contact with a softer ceramic (the so-called two-body condition), the abrasive grits begin to cut or skive grooves in the ceramic surfaces [ 241. For example, with abrasive-impregnated tapes rubbing on Mn-Zn ferrite oxide ceramic the abrasive grits (e.g. silicon carbide and aluminum oxide) can cut into and remove material from the oxide ceramic, and the action results in grooves that are formed plastically primarily by the ploughing and microcutting of the abrasive grits, as indicated in the example presented in Fig. 5(a) [25]. A large number of plastically deformed grooves are formed on the ferrite surface in the sliding direction of the tape. The roughness of the wear surface measured by surface profilometer was 0.2 pm for the maximum height of the irregularities. Figure 5(b) presents a scanning electron photomicrograph of the surface of a tape impregnated with 14 E.cm A&O, abrasive after ten passes in sliding contact with Mn-Zn ferrite. Particles of A1203 are held in the non-metallic binder. Figure 5(b) clearly indicates that relatively few abrasive grits were in contact with ferrite, and ferrite was removed by a microcutting and ploughing action. Only the tips of such grits contained ferrite wear debris and were involved in the abrasive action. Most of the ferrite debris observed on the tape surface was powdery and irregular in shape, but some was curled which is indicative of a cutting action. Abrasive wear can also occur when a third particle harder than one or both of the surfaces in contact becomes trapped at the interface. It can

Fig. 5. Wear surface of Mn-Zn ferrite {llOJ surface and abrasive-impregnated tape: (a) replication electron photomicrograph of wear surface (single-pass sliding of tape impregnated with 14 p A1203; sliding surface, (110); sliding direction, (110);sliding velocity, 11.19 m s-r; laboratory air; room temperature); (b) scanning electron photomicrograph of tape impregnated with 14 pm A1203 after ten-pass sliding against Mn-Zn ferrite (sliding velocity, 0.52 m s-r ; laboratory air; room temperature).

then operate to remove material from one or both surfaces. This mode of wear is then the so-called three-body abrasion. Figure 6 presents replication electron photomicrographs and reflection electron diffraction patterns of wear surfaces of Mn-Zn ferrite abraded by 15 and 4 pm silicon carbide grits in the three-body condition. The abrasion with 15 pm silicon carbide grits results in brittle fractured facets on the Mn-Zn ferrite surface due to cleavage and quasicleavage, while 4 I.trnsilicon carbide grits mostly produced a large number of plastically deformed indentations and grooves. These were formed primarily by ploughing and microcu~ing processes. A very few brittle fractured facets were observed on the surface abraded by the 4 pm silicon carbide grits.

Fig. 6. Replication electron photomicrographs of abraded surfaces of Mn-Zn ferrite (singlecrystal Mn-Zn ferrite (100) plane; abrasion direction, (011); lapping machine; lapping disk, cast iron; abrasive Sic grit 1000 (15 pm diameter) and 4000 (4 pm diameter); lapping fluid, olive oil; sliding velocity, 0.5 m s-l ; abrasive:fluid ratio, 27 wt.%; arrows denote cracks): (a) abraded surface on 15 pm SIC grits (apparent contact pressure, 3 N cmW2); (b) abraded surface on 4 pm Sic grits (apparent contact pressure, 8 N cmV2).

Polycrystalline diffraction patterns taken on both abraded surfaces of single-crystal Mn-Zn ferrite contained continuous arcs extending over nearly a semicircle. The rest were blocked out by the shadow of the solid specimen. The arcs for the abraded surface generated by the 15 pm silicon carbide grits are much sharper than those by 4 pm silicon carbide grits. The broad arcs indicate a large extent of plastic deformation on the abraded surface by 4 pm silicon carbide grits. Thus there is a drastic change in abrasion mechanism

342

0

5

10

15 0 5 ABRASIVE GRIT SIZE, m

(4

10

15

@I

Fig. 7. Specific wear rates of polycrystalline Mn-Zn ferrite as a function of abrasive grit size: (a) two-body conditions (abrasive-impregnated tapes; sliding velocity, 0.52 m s-’ ; laboratory air; room temperature); (b) three-body conditions (lapping disk, cast iron; lapping fluid, olive oil; sliding velocity, 0.5 m 6-r ; abrasive:fluid ratio, 27 wt.%; apparent contact pressure, 4 N cmP2 ).

of ferrite oxide ceramic determined by the size of abrasive grits. With 15 pm silicon carbide grits abrasion of ferrite is p~cip~ly due to brittle fracture, while with 4 pm silicon carbide grits it is due to plastic defo~ation. Many investigators have found that the abrasive wear of metals increases with an increase in the size of the abrasive grits up to a certain critical value and, thereafter, the abrasive wear is independent of the grit size [24 3, The critical grit size is in the range 30 - 150 pm in diameter but strongly depends on abrasives, materials and conditions, such as load, velocity and environment. The grit size effect has also been found to hold for ceramic materials. Figure 7 presents specific wear rates of polycrystalline Mn-Zn ferrite as a function of abrasive grit size under two- and three-body conditions. The specific wear rates were strongly dependent on abrasive grit size in both two-body and three-body abrasion, The specific wear rates decreased rapidly with a decrease in grit size.

4. Metal-ceramic systems The removal of adsorbed films (water vapor, carbon monoxide and dioxide, and oxide layers) from the surfaces of ceramics and metals results in very strong interfacial adhesion and surface chemistry plays a very important role in the observed friction and wear behavior when two such solids are brought into contact. The shear strength of the clean metal in contact with oxide ceramics such as aluminum oxide (sapphire) and Mn-Zn and Ni-Zn ferrites directly

343

f[y<,

i_

0

(a)

:‘I*/I, .4

10 20 30405obo FREE ENERGY OF FORMATION OF LOWEST OXIDE. -A$ Ikcal Igm ATOM OXYGEN)

0

25 54 75 100 125 FREE ENERGY OF FORMATION of LOWEST OXIDE -AF 3CO.lkcal /g ATOM OXYGEN

@I

Fig. 8. Shear coefficients of clean and chlorinated metals in contact with clean {OOOl} sapphire and coefficients of friction for clean metals in contact with clean Ni-Zn ferrite polycrystalline as a function of free energy of formation of the lowest metal oxide: (a) metal-to-sapphire contact (normal load, approximately 0.98 N; vacuum pressure, 30 nPa; room temperature); (b) metal-to-Ni-Zn ferrite contact (normal load, 0.05 - 0.2 N; vacuum pressure, 30 nPa; room temperature).

correlated with the free energy of oxide formation for the lowest metal oxide, in accord with the hypothesis that a chemical bond is formed between metal cations and oxygen anions in the oxide ceramic surfaces [26,27]. Typical results are presented in Fig. 8. The strength of the metal-oxide ceramic bond is related to the oxygen-metal bond strength in the metal oxide. The data for chlorinated metal-sapphire contacts presented in Fig. 8(a) will be discussed later. The relative chemical activity of the transition metals (metals with partially filled d shells) as a group can be ascertained from their percentage d bond character, as shown by Pauling [28]. The frictional properties of metal-ceramic (such as BN, Sic and ferrites) contacts have been shown to be related to this character [29, 30,111. The greater the percentage of d bond character, the less active is the metal and the lower is the friction. Conversely, the more active the metal, the higher is the coefficient of friction. This is demonstrated in Fig. 9. The coefficients of friction for some transition metals (titanium, zirconium, vanadium, cobalt, platinum, nickel, iron, chromium, rhenium, rhodium and tungsten) in contact with the silicon carbide surface decrease with an increase in d bond character. Titanium and zirconium, which are chemically very active, exhibited very strong interfacial adhesive bonding to the silicon carbide. In contrast, rhodium and rhenium, which have a very high percentage of d bond character, have relatively low coefficients of friction. All the transition metals examined failed either in tension or in shear and transferred to the surfaces of silicon carbide because the interfacial bonds are generally stronger than the cohesive bonds in the metals. In other words the weakest bond in the interfacial region was the metal bond. There-

344

g z -

.6

"0 5 4 k

.5

g

.4

Ti

SOLID SYMBOLS DENOTE COEFFICIENT OF FRICTIONAS FUNCTION OF THEORETICALSHEAR STRENGTH

E

.3La

55

PERCENT d CHARACTER OFMETAL BOND

I 0

I I I 15 5 10 THEOREICALSHEARSTRENGTH, ~~~~~ GPa

J al

Fig, 9. Coefficient of friction as function of and theoretical shear strength of metals (- surface in vacuum (sliding direction, (1010); 0.5 N; room temperature; vacuum pressure, 10

fore

percentage metal d bond character (-) -) in contact with silicon carbide (0001) sliding velocity, 3 mm min-’ ; load, 0.05 nPa).

friction properties should correlate with the tensile and shear properties of these metals as well. There appeared to be a general correlation between the coefficient of friction and the theoretical shear strength of metals, as indicated typically with the full symbols in Fig. 9. The higher the shear strength, the lower the friction [31]. The morphology of metal transfer to the ceramic material revealed that much more transfer occurred with the metals that have low strength and a low percentage d bond character than with those having higher values [ 301. In addition to shear of the adhesive bonds at the interface or of cohesive bonds in the metal, another parameter which can influence the coefficient of friction and wear is ploughing. Ploughing is extremely important when metals contact ceramics. The marked difference in elastic and plastic deformation of ceramics and metals can result in ploughing being the principal contributor to measured friction forces and wear. The ploughing effect on friction and wear is demonstrated in Fig. 10 [32]. Figure 10 shows some results when a hemispherical silicon carbide rider slid on a number of metals in mineral oil. Oil was used to minimize adhesion effects on friction. The coefficients of friction and the groove heights corresponding to the groove wear volume as a function of the shear strength of the metal are presented in Fig. 10. The coefficients of friction and groove heights correlate with the shear strength of the metal, and the relationships are approximately linearly decreasing. They are governed by the shear strength of the bulk metal. Such a relationship between the groove height and the shear strength of the metal was also observed with a spherical silicon carbide rider in sliding contact with metals in dry argon [ 331.

345

ULTIMATE SHEAR STRENGTH OF METAL. Pa

Fig. 10. (a) Average coefficient of friction and (b) groove height as a function of shear strength for various metals as a result of the single-pass sliding of a silicon carbide rider (radius, 0.025 mm) in mineral oil (sliding velocity, 3 mm min-‘; load, 0.049 or 0.2 N; room temperature).

5. Adsorbate and environmental effects With ceramic materials, adsorbates from the env~onment are present on the surface. The adsorbates play a very large role in mechanical and chemical behavior of ceramic surfaces in tribological systems. Typical friction results of silicon carbide surfaces in contact with iron as a function of sliding temperatures are indicated in Fig. 11. The iron was argon ion sputter cleaned before the friction experiments in order to remove adsorbed films and oxides on the iron. The silicon carbide flat specimen was as received.

0-m

SLIDING TEMPERATURE.

‘C

Fig. 11. Effect of temperature on coefficient of friction for sintered polycrystalline silicon carbide surface sliding against clean iron rider (which was argon ion sputter cleaned before the experiments were carried out) (normal load, 0.1 - 0.2 N; vacuum pressure, 30 nPa).

346

In Fig. 11 the coefficient of friction remains low below 250 “C. The low friction is associated with the presence of adsorbates such as oxygen and carbon contaminants on the as-received silicon carbide specimen [S]. The coefficient of friction increased rapidly with increasing temperature and remained high in the range 400 - 800 ‘C, while the adsorbed carbon contaminant, according to X-ray photoelectron spectroscopy results, had decreased rapidly with increasing temperature to 400 “C. Above 400 “C it had disappeared from the silicon carbide surface completely [ 81. Above 400 “C silicon dioxide, which was beneath the layer of adsorbate at temperatures below 400 “C, appeared on the silicon carbide surface. Thus the rapid increase in friction and the high friction at 400 800 “C can be related to the absence of the adsorbate. This friction increase is due to increased adhesion and increased plastic flow in the area of contact between two solids. Above 800 “C the coefficient of friction decreases rapidly with an increase in temperature. The rapid decrease in friction correlated with the graphitization of the silicon carbide surface. Silicon carbide exhibits a capability for self-lubrication in vacuum at temperatures above 800 “C!,and this subject will be discussed later. As already indicated in Fig. 8, contacts of sapphire (singlecrystal aluminum oxide) with metals which were artificially exposed to chlorine and had surfaces nearly saturated with chemisorbed chlorinated species exhibited uniformly low shear strength [ 261. That is, the effect of chlorine exposure was to reduce the shear strengths of all the contacts to about the same level, except for silver which was unaffected. These results indicate a van der Waals bonding between chlorinated metal surfaces and aluminum oxide. In contrast, contacts of oxide ceramics (aluminum oxide and Mn-Zn and Ni-Zn ferrites) with metals which were exposed to oxygen exhibited enhanced shear strength and high coefficients of friction indicative of strong oxide-oxide bonding between oxidized metal surfaces and oxide ceramics [26,27]. Silicon carbide-silicon carbide and silicon carbide-titanium contacts exposed to oxygen also exhibited enhanced shear strength between two solids [34]. This is due to the enhanced bond of the silicon oxide to silicon oxide or to titanium oxide.

6. Lubrication Lubrication is extremely important to ceramic materials in tribological systems. Not only are adhesion and friction of ceramic materials reduced by the presence of lubricants but also brittle fracture during sliding. This is one of the main limitations to a wider use of ceramic materials in tribological applications. Figure 12 presents scanning electron photomicrographs of wear tracks for Mn-Zn ferrite oxide ceramic in sliding contact with a hemispherical

347

(a)

(b) Fig. 12. Scanning electron photomicrographs of wear tracks on single-crystal Mn-Zn ferrite (100) surface in sliding contact with a hemispherical diamond rider (0.1 mm radius) (a) in lubricant (olive oil) and (b) in air (sliding surface, (100); sliding direction, (001); normal load, 5 N; sliding velocity, 3 mm min-’ ; room temperature).

(100 pm radius) diamond rider in dry and under lubricated conditions. The lubricant was an olive oil, which is commonly used in fine lapping processes of ceramic materials and semiconductors. As shown in the photomicrographs, the fracture behavior of the oxide ceramic is very dependent on the lubricant. Cracks generate much more easily in the dry condition than with lubrication. As mentioned earlier (Fig. 4), the tangential force introduced by sliding plays an important role in the generation of surface fracture. The critical load to fracture in dry sliding was one-half of that observed with static indentation. However, under lubricated conditions the critical load to fracture was two times greater than that in dry sliding and is almost the same as that in indentation. The critical load to fracture in lubricated sliding and in static indentation with a 100 pm diamond was 4 N, while that in dry sliding was 1.5 N on the (100) surfaces of the Mn-Zn ferrite. As already mentioned in relation to Fig. 11, heating of silicon carbide in the range 800 - 1500 “C graphitized the silicon carbide surface as a result

34% SLIDING DIRECTION

tj 0

5

E

lad (a)

102

iw

98

Sizp XPS PEAK

96 292 290 28s 286 284 Eindins enemv. eV -&S

282

280

218

loin: < 1120i

<

jiy%r_

216

XPS PEAKS

0

0

(b)

4.63 am SLIDING TEMPERATURE, ‘C

ml

Fig, 13, Coefficient of friction for single-crystal silicon carbide {OOOI) surface preheated to 1500 “C sliding against an iron rider (normal load, 0.2 N; sliding vetocity, 3 mm min-’ ; vacuum pressure, 10 nPa): (a) surface chemistry of silicon carbide heated at 1500 “C; (b) friction behavior of graphitized silicon carbide surface.

of segregation and evaporation of the silicon in the surficial region of the silicon carbide. Thereafter, recons~uction of the remaining carbon occurred at the silicon carbide surface 135 - 371. The ~aphitization of the silicon carbide surface was demonstrated by X-ray photoelectron spectroscopy analysis (Fig. 13(a)). Figure 13(a) presents the graphitized silicon carbide surface after heating to 1500 “C. After being cooled to room temperature, sliding friction experiments from room temperature to 1200 “C in vacuum were conducted on the preheated surface of the silicon carbide. The friction properties are presented in Fig. 13(b). The silicon carbide specimens preheated to 1500 “C produced coefficients of friction lower, by one-half, than those for the surfaces cleaned by argon ion sputtering [37]. The marked difference in friction between ~phitiz~ and non-~aphitized surfaces was apparent even in air [38]. Thus the silicon carbide generates its own solid lubricant surface films in the form of the graphite.

7. Crystallographic orientation (anisotropy) Much like metals, ceramics exhibit anisotropic behavior in many of their mechanical properties. Friction and wear behaviors of ceramics are also anisotropic under adhesive or abrasive conditions. Anisotropy results can be of two kinds. First, there is the observed variation in friction and wear when the sliding surface is changed from one crystal plane to another for a given material. Secondly, there is the variation in friction and wear observed when the orientation of the sliding surface is changed with respect to a specific crystallographic direction on a given crystal plane. For example, the differences in the coefficients of friction with respect to the mating crystallographic planes and directions are significant under adhesive conditions as indicated in Table 1 ill]. The data of

349

TABLE 1 Anisotropic

friction

for Mn-Zn

Mated

plane

Effect

of crystallographic

{llO}on (111) on (100) on (110) on (110) on (110) on Effect

{llO}on

Mated

(110) (111) (100) (100) (111) {211} (110)

Load, 0.05 - 0.5 temperature.

direction

Sliding

direction

Coefficient of fric tion

plane

(110) (110) (110) (110) (110) (110)

(110) (110) (110) (110) (110) (110)

0.21 0.21 0.24 0.21 0.29 0.29

(110) on (110) (110) on (100)

(110) (110)

0.21 0.43

(110) on (110) on (110) on (110) on (110) on (1lO)on

of crystallographic

ferrite and silicon carbide under adhesive conditions

direction

N; sliding velocity,

3 mm min-’ ; vacuum pressure, lo-*

N m-2 ; room

Table 1 were obtained in vacuum with clean ferrite-ferrite oxide ceramics. The mating of preferred slip plane and highest atomic density plane and direction, such as {llO}(llO) and {111}(110) for Mn-Zn ferrite, gives the lowest coefficients of friction. In other words, the lowest coefficients of friction were obtained on the preferred slip plane when sliding in the preferred slip direction. Similar results have been obtained with silicon carbide and aluminum oxide (sapphire). Their anisotropic friction is shown in Table 2. Again, the lowest coefficients of friction with silicon carbide and aluminum oxide were obtained on the preferred basal slip plane when sliding in the preferred (1120) slip direction [ 10,391. The coefficient of friction reflects the force required to shear at the interface with the basal planes of the silicon carbide or aluminum oxide parallel to the interface. The results presented in Table 2 indicate that the force required to resist the shearing fracture of the adhesion bond at the interface is lower in the preferred crystallographic (1120) direction than it is in the direction. The frictional properties of the silicon carbide (0001) surface in contact with iron at temperatures to 800 “C in vacuum indicated that the coefficient of friction was lower in the (1120) direction than it was in the direction over the entire temperature range from room temperature to 800 “C [lo]. Under abrasion just as under adhesive conditions the differences in the coefficients of friction and wear with respect to the crystallographic plane and direction were also significant, as indicated in Fig. 1 [3, lo]. In this case, ploughing was extremely important and influenced the friction and wear when ceramics contacted hard materials. With silicon carbide the
350 TABLE

2

Anisotropic friction for sapphire-sapphire under adhesive conditions

and silicon carbide-silicon

Direction

Coefficient

Prismatic {lOlO>

(1120) (OOOll

0.93 1.00

Basal (0001)

(1lZOl QOiOl

0.50 0.96

(1120)

0.54 0.68

Plane Sapphire

carbide contacts

of friction

sliding on sapphirea

SiC on SiCb

Basal (0001)

(loio)

aLoad, 10 N; sliding velocity, 7.8 mm min-’ ; vacuum pressure, 10W8 N rnp2; room temperature. bLoad, 0.3 N; sliding velocity, 3 mm min-” ; vacuum pressure, 10e8 N rn-* ; room temperature.

TABLE 3 Anisotropic friction and wear for silicon carbide under abrasive conditions Plane

(0001)

Direcfion

Cioio)

(1120)

{ioio} {llZO]

Coefficient of friction

0.19 0.21

Width of permanent groove corresponding to plastic deformation (pm) 2.6

2.7

(0001) (1120)

0.29 0.23

3.6 2.9

(0001)

0.31

3.5

0.24

3.1

(ioio>

Conical diamond with an apical angle of 117”; load, 0.2 N; sliding velocity, 3 mm min-’ ; mineral oil; room temperature.

deformation, as indicated in Table 3. The anisotropic friction and wear on the {OOOl}, {lOiO} and (11~0) planes of silicon carbide were primarily controlled by the slip system {lOiO}(ll~O> and were explained by a resolved shear stress calculation [ lo]. The abrasive wear of Mn-Zn ferrite sliding against a tape impregnated with 7 flrn Al,Os abrasive (the two-body condition} decreased in the order (211) > (111) > (100) > (110). The slip planes (110) (most closely packed planes) of Mn-Zn ferrite exhibited the highest resistance to abrasion [ 19 1. Crys~o~phic direction effects on the wear of sapphire were also found

L401.

351

8. Concluding

remarks

On the basis of fundamental studies conducted with both oxide and non-oxide ceramics the following remarks can be made. Ceramics like metals will deform elastically in the interfacial region between two solids in contact under load. Unlike metals, however, when the elastic limit is exceeded, gross fracture as well as plastic deformation can occur. With tangential motion the forces necessary to initiate brittle fracture in the ceramic are markedly less than those associated with normal loading. The friction and wear characteristics of ceramics, again like metals, are anisotropic. In general the lowest coefficient, of friction is observed when sliding on the preferred slip plane in the preferred slip direction on that plane. Adhesive wear behaves with respect to orientation as does friction. Likewise the atomic planes of high atomic density and most resistant to deformation exhibited the greatest resistance to abrasive wear. In twoand three-body abrasion the grit size effect on abrasive wear holds for ceramics. Wear decreases with a decrease in grit size. When ceramics are in contact with metals, surface chemistry is extremely important to friction and wear behavior. For oxide ceramics the free energy of oxide formation for the lowest metal oxide is directly correlated with metal shear properties which relates to friction. With the transition metals, the d valence bond character correlates directly with the coefficient of friction for both the oxide and the non-oxide ceramics. The higher the percentage d bond character the lower is the friction coefficient. Contaminants on the surface of ceramics affect their tribological behavior. For example, desorption of carbon contaminants from a silicon carbide surface results in an appreciable increase in the coefficient of friction. With aluminum oxide in contact with metals, chlorine as a surface contaminant decreases interfacial bond strengths, and accordingly friction, while oxygen increases both adhesion and friction. Lubrication of ceramic surfaces increases the critical load required to initiate fracture of ceramics under sliding or rubbing. Heating of a ceramic such as silicon carbide to high temperatures can result, in the graphitization of the ceramic surface with the graphite functioning to reduce abrasion and friction. A lubricating film is therefore provided from the material itself.

References K. F. Dufrane and W. A. Glaeser, Study of rolling-contact phenomena in magnesium oxide, NASA Tech. Rep. CR-72295, 1967 (National Aeronautics and Space Administration). R. P. Steijn, On the wear of sapphire, J. Appl. Phys., 32 (10) (1961) 1951 - 1958. K. Miyoshi and D. H. Buckley, Friction, deformation and fracture of single-crystal silicon carbide, ASLE Tmns., 22 (1) (1979) 79 - 90. M. L. Kronberg, Plastic deformation of single crystals of sapphire: basal slip and twinning, Acta Metall., 5 (9) (1957) 507 - 524.

352 5 D. H. Buckley,

6

I 8 9

10 11 12

Influence of surface active agents on friction, deformation and fracture of lithium fluoride, NASA Tech. Rep. TN D-4716, 1968 (National Aeronautics and Space Administration). D. H. Buckley, Effect of surface active media on friction, deformation and fracture of calcium fluoride, NASA Tech. Rep. TN D-5580, 1969 (National Aeronautics and Space Administration). F. P. Bowden and D. Tabor, The Friction and Lubrication of Solids, Part II, Clarendon, Oxford, 1964, p. 173. K. Miyoshi and D. H. Buckley, Tribological properties of sintered polycrystalline and single-crystal silicon carbide, Am. Ceram. Sot. Bull., 62 (4) (1983) 494 - 500. K. Miyoshi and D. H. Buckley, Friction and fracture of single-crystal silicon carbide in contact with itself and titanium, ASLE Trans., 22 (2) (1979) 146 - 153. K. Miyoshi and D. H. Buckley, Anisotropic tribological properties of Sic, Wear, 75 (1982) 253 - 268. K. Miyoshi and D. H. Buckley, Friction and wear of single-crystal manganese-zinc ferrite, Wear, 66 (1981) 157 - 173. H. Hertz, ober die Beruhrung fester elastischer KSrper, 2. Reine Angew. Math., 92 (1881)

156

- 171.

F. Auebach, Ann. Phys. Chem., 43 (1891) 61. 14 0. W. Johnson, Damage produced in Ge at room temperature by indentation, J. Appl. Phys., 37 (7) (1966) 2521 - 2526. 15 B. D. Powell and D. Tabor, The fracture of titanium carbide under static and sliding contact, J. Phys., 3 (1970) 783 - 788. 16 F. C. Frank and B. R. Lawn, On the theory of hertzian fracture, Proc. R. Sot. London, Ser. A, 299 (1967) 291 - 306. 17 B. R. Lawn, Hertzian fracture in single crystals with the diamond structure, J. Appl. Phys., 39 (10) (1968) 4828 - 4836. 18 B. R. Lawn and M. V. Swain, Microfracture beneath point indentations in brittle solids, J. Mater. Sci., 10 (1975) 113 - 122. 19 K. Miyoshi and D. H. Buckley, Ceramic wear in indentation and sliding contact, ASLE Preprint 84-AM-4A-2, 1984 (American Society of Lubricating Engineers); NASA Tech. Rep. TM-83585, 1984 (National Aeronautics and Space Administration). 20 F. W. Preston, The structure of abraded glass surfaces, Trans. Opt. Sot. (London), 23 (3) (1922) 141- 164. 21 P. R. Billinghurst, C. A. Brookes and D. Tabor, The sliding process as a fractureinducing mechanism, Physical Basis of Yield and Fracture, Inst. Phys. Sot. Conf. Ser. 1 (1967) 253 - 258. 22 K. Tanaka, K. Miyoshi, T. Miyao and T. Murayama, Friction and deformation of Mn-Zn ferrite single crystals. In T. Sakurai (ed.), Proc. JSLE-ASLE Int. Lubrication Conf., Elsevier, Amsterdam, 1976, pp. 58 - 66. 23 T. Sugita and A. Hashikawa, The roles of brittle microfracture and plastic flow in the wear of MgO single crystals, Wear, 72 (3) (1981) 295 - 303. 24 E. Rabinowicz, Friction and Wear of Materials, Wiley, New York, 1964. 25 K. Miyoshi, Lapping of manganese-zinc ferrite by abrasive tape, J. Am. Sot. Lubr. Eng., 38 (3) (1982) 165 - 172. 26 S. V. Pepper, Shear strength of metal--sapphire contacts, J. Appl. Phys., 47 (3) (1976) 801 - 808. 27 K. Miyoshi and D. H. Buckley, Surface chemistry, friction, and wear of Ni-Zn and Mn-Zn ferrites in contact with metals, Ceram. Eng. Sci. Proc., 4 (7 - 8) (1983) 13

674

- 693.

L. Pauling, A resonating-valence-bond theory of metals and intermetallic compounds, Proc. R. Sot. London, Ser. A, 196 (1949) 343 - 362. 29 D. H. Buckley, Friction and transfer behavior of pyrolytic boron nitride in contact with various metals, ASLE Trans., 21 (2) (1978) 118 - 124. 30 K. Miyoshi and D. H. Buckley, Friction and wear behavior of single-crystal silicon carbide in sliding contact with various metals, ASLE Trans., 22 (3) (1979) 245 - 256. 28

353 31 K. Miyoshi and D. H. Buckley, Correlation of tensile and shear strengths of metals with their friction properties, ASLE Trans., 27 (1) (1984) 15 - 23. 32 K. Miyoshi and D. H. Buckley, The friction and wear of metals and binary alloys in contact with an abrasive grit of single-crystal silicon carbide, ASLE Trans., 23 (4) (1980) 460 - 472. 33 K. Miyoshi and D. H. Buckley, Friction and wear of metals with a single-crystal abrasive grit of silicon carbide - effect of shear strength of metal, NASA Tech. Publ. TP-1293, 1978 (National Aeronautics and Space Administration). 34 K. Miyoshi and D. H. Buckley, Effect of oxygen and nitrogen interactions on friction of single-crystal silicon carbide, NASA Tech. Publ. TP-1265, 1978 (National Aeronautics and Space Administration). 35 D. V. Badami, X-ray studies of graphite formed by decomposing silicon carbide, Carbon, 3 (1) (1965) 53 - 57. 36 A. J. Van Bommel, J. E. Crombeen and A. Van Tooren, LEED and Auger electron observations of the SIC (0001) surface, Surf. Sci., 48 (2) (1975) 463 - 472. 37 K. Miyoshi and D. H. Buckley, XPS, AES and friction studies of single-crystal silicon carbide, Appl. Surf. Sci., 10 (1982) 357 - 376. 38 K. Miyoshi and D. H. Buckley, Ceramics in high temperature applications - fundamental considerations in friction and wear, Proc. Int. Conf. on Tribology in the 8Os, Vol. 1, in NASA CP-2300, 1984, pp. 291 - 329 (National Aeronautics and Space Administration). 39 D. H. Buckley, Friction and wear behavior of glasses and ceramics, Mater. Sci. Res., 7 (1974) 101 - 126. 40 D. H. Buckley, Friction and wear of ceramics, Am. Ceram. Sot. Bulk, 51 (12) (1972) 884 - 891.