WEAR Influence of whisker reinforcement on the abrasive wear behavior of silicon nitride- and alumina-based composites CP.Do总an,JA.Hawk Albany Research Ceater Deparment d Energy Ahu. OR 97521054 behavior is atunbuted to the r Keyword: Composites: Abrasve wear Amund, MIcowucTure nction of the ha ecton [9- ng crack. Thus, while the long- crack toughness of the bul whisker t en noted that ine have addressed the relative perfo forced behavior of several silicon nitride. and al mina-based ceramic materials by companing ue microst c043-1648/97/SI7 00c 1997 Elsevier SclenceS A. All righes resend
ELSEVIER WEAR Wear 203-204 ( 1997) 267-277 Influence of whisker reinforcement on the abrasive wear behavior of silicon nitride- and alumina-based composites C.P. Dogan, J.A. Hawk Albany Research Cenrrr. Deportmmr of Energy Albany. OR 97321. USA The abrasive wear of brittle materials, while not a tme property of materials, is generally modeled as an inverse limction of both the bulk hardness and frashue toughness. According to these models. an increase in the hardness and/or fmcture tott@ of a material will thereforr enhance its wear resistance. In ceramic mat&Is, the addition of whisker reinforcement is a pmven method ofe&ancing long-crack fmcmm toughness via such mechanisms as crack bridging attd wbiskerdcbondittg. However, less is known about bow wbiskerreinforcementingtzmes the properties that arc dependent upon the shott-ctack oughttess. such as abrasive wear. The results of this study indicate th8t while the addition of randomly oriented SE whiskers can dramatically improve the abrasive wear resistance of an abmdna-based ceramic. the arkiition of Sic reinforcement to silicon nitride ceramics does not always result in improved wear resistance. This wtriation in the intluence of whisker ninforcetnent on wear behavior is attributed to the residual stress state created in the composites as a result of the addition of the second phase whiskers. Keywordc Composites; Abrasive wear. Silicon nitride; Alumina; Micmsmxture 1. lntroductioa Because of the complexity of wear processes, a detailed understanding of how ceramic materials react in trihological environments continues to elude both materials and design engineers. For the abrasive wear of brittle materials. mathematical models generally express volume wear as an inverse function of the hardness and the fracture toughness of the material [ 1,2]. However, these. models fail to adequately describe the abrasive wear behavior of most advanced ceramic materials in at least one important way: they assume that bulk hardness and fracture toughness measurements are sufficient to describe the deformation and fracturecharacteristics of the test material in an abrasive wear environment. A number of studies on a wide variety of ceramic materials have indicated that this is not the case [3-S]. In particular, it has been noted that in ceramic materials which exhibit increasing fracture toughness with increasing crack length (called Rcurve or T-curve behavior), such as alumina and zirconia ceramics and many ceramic-based composites, the measured bulk fracture toughness may not describe the fracture tougbness of the ceramic at the micmstructural scale where abrasive wearmechanismsareactive [6-g]. Ceramic-reinforced ceramic composites represent one of the best developments so far in the race to produce tough, yet 0043.1648/97/$17.OO Q1997 Elsevier Science .%A AU tights reserved PllSOO43.1648(96)07348-6 mechanically reliable, ceramic materials for advanced sttttcturalapplications,andp~cular)yforapplicationsatelevated temperatures. Over the course of the last decade, much research has gone into the tmderstandhtg of precisely how the addition of reinforcement phases can intloence the strength and toughness of the bulk ceramic material. For whisker-reinforced ceramic materials, enhanced toughness can occur through any one. or several, of the followhtgmechanisms: whisker &bonding, whii pullout, crack bridging. and/or crack deflection [S-14]. However, in almost eveq case, :arge-scaie toughening requires that these mechanisms be activated over some distance behind the tip of a propagating crack. Thus, while the long-crack toughness of the bulk material may be enhanced by the pmsence of the whisker reinforcement, the short-crack toughness of the bulk material may remain ttnchanged, or even ba degraded, by the preaenca of these whiskers. Studies of how the preseneo of whisker reinforcement influences the ttibological pmpetties of ceramic materials are largely absent, although sevaral studies have addressed the nlative performance of whiia-reinforced ceramic matrix composites in sling and abrasive wear environments [45.15-111. In thii study. we begin to examine how randomly oriented SE whiskers ittflucnce the abrasive wear behavior of several silicon nitride- and ahtmina-based ceramic materials by comparing the micmsbuc-
ures, hardness, fracture toughness and wear behavior of both ies and ceramic composites of this study were identically the reinforced and unreinforced materials. 2 Experimental particle sizes of 37, 58 an Three commercially available, SiC-whisker reinforced tests the The htst pair of sitcom nitnoe-base matrix. Because no similarly processed monolithic alumina 99.5% Al,on of constant was calculated according to the equation [221 In these equations, Am was the mass loss of the test specimen secondary clectron imaging mode. Sample preparation for normal load(N)on the specimen; and e pre by coating with a Au-Pd alloy to prevent charging in the wear rate or the wear constar st results of greater than +5.5% attributed to real differences in the wear behavior of the materials. easurements were taken for each material, with the results rements were made for each 99.8% A1-O, can be found in simulate abrasion bonded by a continuous crystalline ver, it is important to recognize that during this test the boundary phase. The composition sample was contimuallyexposed to fresh abrasive. Tbecetam- boundary phas end upon the additives and impurities
268 C.P. Do&m. J.A. Hawk/ Wear 203-204 (1997) 267-277 mres, hardness, fracture toughness and wear behavior of both the reinforced and unreinforced materials. 2. Experimental Three commercially available, Sic-whisker reinforced composite materials were selected for this study, along with several chemically similar but unreinforced matix materials for comparison. The first pair of silicon nitride-based materials, S&N,-A and S&N,-A + SIC,, were processed in an identical manner, except for the addition of 15 vol.% SIC whiskers to the composite material. The second series of silimn _..L...C nrAA^-5ised Y ceramics, S&N.-B and S&N.,-B +SiC,, were processed somewhat differently than the hrst, in order to produce a different matrix microstructure; the composite material also contained i5 vol.% SE whiskers. For the alumina-based materials, the composite, A120~+SiC,,., consisted of 34 vol.% SE whiskers in a high-purity alumina matrix. Because no similarly processed monolithic alumina was available, two high purity ahuninas-a 99.8% AlzO, of relatively high hardness, and a 99.5% A1203 of intermediate hardness-were selected for comparison with tbe composite. Microstructural characterization of these materials was accomplished primarily by transmission electron microscopy (TEM), in combination with chemical analysis by X-ray energy dispersive spectroscopy (XBDS). Analysis of the materials’ microstructural response to the various abrasive wear tests was by scanning electron microscopy (SBM) in secondary electron imaging mode. Sample preparation for TBM analysis followed traditional ceramographic techniques, including ion milling to electron transparency. Sample preparation for SEM analysis was limited to ultrasonic removal of the wear debris from the wear surfaces, followed by coating with u Au-Pd alloy to prevent charging in the microscope Hardness and fracture toughness were measured for each material on surfaces mechanically polished to a 1 km diamond 8nish. Vickers hardness was determined under a load of 1 kg, with a dwell time of 15 s. Ten separate hardness measurements were taken for each material, with the results averaged. Fracture toughness was measured utilizing the indentation technique described by Anstis et al. [ 191 with the indenting load varied between 10 and 20 kg. depending on what was required to develop a well-defined crack pattern with a measured crack length that was at least three times the diagonal diameter. A minimum of five fracture toughness me‘asurements were made for each material and the results averaged. Material response to two-body abrasive wear was measured utilizing a pin abrasion test designed at me Albany Research Center to simulate abrasion mat occurs during crushing and grinding operations. This test is described in detail elsewhere [ 201, and will not be discussed here-; however, it is important to recognize that during this test the sample was continually exposed to fresh abrasive. Theceram- 3. Results 3.1. Microstructure Micrographs illustrating the general microstructural features of the three composite materials, and &he monolithic 99.8% A&O,, can be found in Fig. 1, and an outline of the microstructural characteristics of all of the ceramics examined in this study can be. found in Table I. With the exception of the 99.8% Al,O,, all ceramics and ceramic composites are liquid-phase sintered materials in which the matrix grams are bonded by a continuous crystalline and/or amorphous grain boundary phase. The composition and nature of these gram boundary phases depend upon the additives and impurities
CP Dofan, J.A Hawk/wear 203-204(1997)267-277 由9肠A0 g and (d Werestrucneral cha Random, intergrator Randoe imer and inmagraul 10, The matrix m amorphous phase is obsened at all homo- and heteropl a thin amorphous phase also wets all S C, -Si N, and SiCw. interfaces(including all whisker interfaces)in both
Fig. I. General microruuetunl features oE (a) the Si,N,-A+SiCL eompositc; (b) dae Si,N,-B +Si composk: (c) the 99.11% Al&; aad (d) tbz AI,O, + SiC, compxite. Table I Microstructural characteristics of the ceramics and ceramic composites Material primary mabix phare Primwy grain boundary phase Si,N.-A .‘%Si,N. Si,N,-A + SiC, /3-SijNI S&N.-B p&N, S&N,-B + Sic, PW% 99.8% AlzO, a-AlzO, 99.5% Al,03 a!-ALO, AI,O, + Sic, a-AlrO, Ciyst;llline a-Y&% Crysclllii a-Y&O, Amor@ous Y.Al silicate Amqhous Y. Al silicate Graphite. (3+&O, Amorphous Ca. Al sdicate Amorphous Mg. Al silicate introducedintothematerialsduringprocessing,andtheirtotal thermal history. ;n the S&N,-A ceramic, this ,&n boundary phase is primarily crystalline a-Y&O,; however, a narrow amorphous phase is also observed at all home- and heterophase boundaries, and is expected 10 be a continuous phase in these materials. The Si,N,-A +SiC, composite [Fig. 1 (a) ] is microstructurally identical except for the distribution of randomly oriented, intergranular SIC whiskers, which are. large relative to the matrix grains. Although not apparent at the magnification of the micrograph in Fig. 1 (a), a thin amorphous phase also wets all Sic, -S&N4 and SE,- Y&O, boundaries, and likely plays a key role in determining the fracture characteristics of the whisker-matrix interface in this composite. The matrix microstructure of the second series of silicon nitride ceramics, Si,N,-B, consists of a distrktion of SiaNa grains which are larger than those in the A-series (average matrix grain size is 0.75 Frn as opposed to 0.4 pm in S&N,- A), bonded by an amorphous yttrium atuminosilicate phase that is also believed to be continuous in these materials. This amorphous phase is obscrvcd at all home- and hetcrophase interfaces (Including all whisker-matrix interfaces) in both
CP Dof JA Hawk/wear 20J-20(/992)267-277 monotic and composite m yy. 8 Aly those provided by the manufacturer to identical matenals, N-a+sic Sic whiskersresults in norealchange in the toughness of Si,Na-B. 3.3 Ab this 99 5% material is 3 der than in 99.8% AL-O,, and grains as large as and also a glass-bonded ceramic ness, the 150-grit alumina abrasive is the least aggressive for mately 2 wol. of a -matrix [23] is ohserve Such aa size within this composite ave Measured values for hardness and fracture toughness, 日9g monolithic SiN NN-B an increase In the hardness of the composite in all monolith SiN4 mater
270 C.P. Do&n, J.A. Hawk/ Wear203-204 (1997) 267-277 the monolithic and composite Si,N,-B ceramics The microstructure of the composite is similar [Fig. 1 (b) 1, except for the presence of randomly oriented Sic whiskers. which have an average diameter of 0.75 pm and a variable aspect ratio. In this case, the whiskers are locatedboth intra8ranuldy (i.e. either partially orcntrrely encapse!a~~withinasihcon nitride grain) and inrergranularly within the microstructure. For the alumina-based ceramics, the microstructums vary quite a lot between the two monolithic materials and between the monolithic and composite materials. 99.8% AlsOs [Fig. 1 (c) ] is carbon bonded, with graphite detected at most alumina-aluminaboundarics. Inaddition,elongated “whiskers” of a potassium-modified &Also, phase are also occasionally observed at the grain boundaries. Significant stresses are apparent at the a-AlsO&AlsOs interfaces; however, because the population of such interfaces is relatively small within this material, the presence of this stress is unlikely to influence the tribological properties of the bulk material. The alumina grain size in this material averages around 2 pm, although there arc pockets of much smaller. sub-micron, grains within the microstructure. In monolithic 99.5% AlsOs, on the other hand, tha microstructum is typical of that of a liquid-phase sintered ceramic, with the alumina grains bonded by an amorphous calcium aluminosilicate phase that is continuous in this material. The average alumina grabt size in this 99.5% material is 3 p.m. but the grain sire distribution is much wider than in 99.8% A120sr and grains as large as 10 Pm are not unusual. The alumina-based composite, AlsOs+SiC,, is also a glass-bonded ceramic [Fig. l(d)], containirtg approximately 2 vol.% of an amorphous magnesium altinosilicate phase that is located at all three- and four-grain junctions and along most two-gram boundaries. This amorphous phase is also observed as a thin layer ( <SO nm) at whisker-matrix interfaces. The Sic whiskers. with an average diameter of 0.75 pm and a vartablc aspect ratio, are distributed randomly throughout the alumina matrix and occur both inter- and intragranularly. In amorphous pockets adjacent to the SE whiskers, small crystals of graphite and an iron-nickel intermetallic are often observed [as in Fig. 1 (d) 1. Alumina grain size within this composite averages around 4 pm. 3.2. Hardness and fiactam toughness Measured values for hardness and fracture toughness, along withreportedvaluesforYoung’smodulus (asprovided by the manufacturers), are listed in Table 2 for all of the ceramic materials examined in this study. The alumina-based materials, Also, + SiC, and 99.8% Also,, tend to have the highest hardness, with values of 23.8 and 19.2 GPa, respectively: whereas unreinforced S&N.-B and 99.5% AlsOs have the lowest hardness, with values of 15.0 and 15.2GPa, respectively. As expected, the addition of silicon carbide whiskers to the silicon nitride and alumina matrices results in an increase in the hardness of the composite in all cases. Indentation fracture toughness measurements do not always provide the most accurate measure of the bulk fracture toughness of a ceramic, often resulting in values lower than those obtained by other measurement ~&uilques 1 i9 ] ; howewr, the indentation technique is selected here because it is believed to provide the value most representative of the nearsurface regions of a material exposed to an abrasive wear environment. As predicted, the measured values for fracture toughness obtained in this study are somewhat lower than thosa provided by the manufacturer for identical materials, or quoted in the literature for similar materials (theexception being the 99.5% A1203). Nonetheless, it is apparent from Table 2 that as a class of materials, the alumina-based ceramics are not as tough as the silicon nitride.-based ceramics, and that 99.8% AlsOs has the lowest toughness of all of the matcvials tested in this study. The addition of SE whiskers to S&N.-A results in a 29% increase in toughness, making S&N.-A + SE, the highest toughness material of this study, whemas the addition of SE whiskers results in no teal change in the toughness of S&N.-B. 3.3. Abrasive wear behavior The measured specific wear rates and wear constants for all of the ceramics and ceramic composites tested against alumina and silicon carbide abrasives are listed in Table 3. and the wear constants for the Si&A and S&N.-A + Sic, materials tested against ISO-grit SE as a function of load are listed in Table 4. As expected from its relatively lower hardness, the 150-grit alumbra abrasive is the least aggressive for all of the ceramics tested except for the finest SE abrasive against the 99.5% Also,. Similarly. the expected increase in abrasive wear rate with increase in Sic abrasive particle size [ 231 is observed for all of tha ceramics and ceramic composites over the abrasive size range of 37-100 pm (Fig. 2). An interesting aspect of Lis data is the large increase ( lSO- 388%) in wear rate with increase in abrasive particle size from 37 pm (400-g&) to 58 pm (240-g&). Such a large difference in wear rate over a relatively small change in abrasive particle size suggests a change in wear mechanisms, and in fact such a change is observed. Following wear against the 4OOgrit Sic abrasive, examination of the wear surfaces indicates that the response of these materials is primarily ona of plastic deformation, with only minimal fracture observed. Bxamination of the wear surfaces following the test against the 24~grit Sic, on the other hand, indicates that fracture is now a signi8cant material response to the wear environment. Under the various abrasive wear conditions of thls study, monolithic SisN,-A is consistently a better performer than is monolithic SisN,-B, particularly against the “softer” alumina abrasive. This result is interesting since examination of the wear surfaces of these materials after abrasion against 15Ogrit alumina suggests that both materials are in the mild wear regime 1241, where fracture toughness is expected to dominate wear behavior. Yet there is no real difference in the fracture toughness of the two monolithic SisN, materials
P. Dodn IA. Hawk/ Wew 203-204(1997 )26y--177 ( cpay 136 54 by the respective manufacturers 100谥mA0 Ss am SiC AHO,+ Q7(09) (Table 2). For the monolithic aluminas, 99.8% AL-O, is asistant than is the 99.5% Al O, uter all study. In fact, with its high hardness and conditions tested. This large varia re toughness, al-o,+SiC is by far the m the two high purity 学 which, with its abrasive wearenvi- in alumina ceramics [25-291), although the difference in 3. 4. Microstrmcrurad response to the wear environment he wear surfaces immediately following ar rate for all SiC abrasive give hw址物邮9场AQ时的邮如可自的此址A again result of this study is the and examples of wear su dbmm减pire and dela of i hand, the addition of Sic whiskers leads to a dramane particles can penetrate more deeply into the surface of the
C.P. Do&m. JA. Hawk/ Wear203-204 (1997) 267- 277 Young’s modulus (GPa)’ Hardners (GPa) Sifl.-A 310 15.6 5.4 SiIN.-A + SiC. 335 19.0 6.4 S&N,-B 303 15.0 5.5 Si,N,-B + SiC, 335 16.5 5.4 99.8% ~~0~ 400 19.3 3.4 99.5% Ai@> 386 15.2 3.9 AlzO, + SiC. 395 23.8 4.6 ’ Dntn provided by the rcspcaive manufactums. TPble 3 Abrasive war rawa for ceramics and ceramic composites (66.7 N nomnl load; 16.0 m sliding distance) 100 pm SIC 58 pm Sic 37pmsiC SisN.-A 0.9 (1.8) 9.1 (19.1) 6.6 (13.9) 2.5 (5.2) Si,N,-A + SiC, 1.0 (2.2) 9.9 (20.8) 6.8 (142) 1.8 (3.9) Z&N.-B 1.7 (3.7) 12.4 (26.1) 8.8 (18.6) 2.9 (6.2) Si3N.-B + SiC, 1.9 (4.0) 14.0 (Z9.4) 9.6 (20.2) 3.4 (7.1) 99.8% Al,o, 1.5 (3.2) 8.6 (18.0) 5.9 (i2.4) 2.1 (4.5) 99.5% Al,o, 6.4 (13.4) 23.6 (49.7) 22.1 (46.6) 5.3 (11.2) Al,O,+SiC_ 0.7 (0.9) 4.8(101) 3.3 (6.9) 0.9 (1.8) (Table 2). For the monolithic aluminas, 99.8% A&O, is much more V.VP -&ant than is the !XW& A!$, urder all conditions tested. This large variation in abrasive wearbehavior between the two high purity ahuninas is most likely the result of the difference in grain size and grain s&distribution between the two materials (a smaller grain size and narrower grain sizedistribution is known to enhance wear performance in alumina ceramics [ 25-291). although the difference in grain boundary microstructure also plays a contributing role by influencing the residual stress state of Ihe ceramic [ 6,301. A comparison of the top performing silicon nitride (S&N.- A) with the top performing alumina (99.8% A&O,), shows that the alumina, with the higher ha&e88 but lower fracture tOughIh?SS, possesses a lower wear rate for all SIC abrasive environments. Against the softer aluminabrasive, however, Si,N4-A has a 75% lower wear rate than the 99.8% A1203. thanks to its higher fracture toughness. Perhaps the most interesting result of this study is the observation that tbe addition of 15 vol.% SE whiskers to a silicon nitride matrix either does not affect the abrasive wear behavior of the bulk material, or degrades it slightly in some instances (Figs. 2 and 3). This cccur8 in spite of the fact that the addition of SiC whiskers increases the hardness of both Si,N,materials and increases the fracture toughness of S&N,- A. The only exception to this rule is in the SisN,-A ceramics tested against 400-grit Sic, where the composite outperforms the monolith In the alumina-based ceramics, on the other hand, the addition of Sic whiskers leads to a dramatic improvement io the abrasive wear resistance under all test conditions of this study. III fact, with it8 high ha&e88 and respectable fracture toughness, A&O, + Sii is by far the most wear resistant material examined in thii study. This is in stark contrast to the 99.5% A1203 which, with it8 low hardness and propensity for fracture in all abrasive wearenvironments. is clearly the worst performer. 3.4. Microstructural response to the wear environment In addition to the measured response to various abrasive wear environmenta listed in Tablea 3 and 4. examination of the wear sorf8ces immediately following the wear tests can give clues to the influence of micro8hucture and whisker reinforcement on the wear behavior of a ceramic material. A8 an example, micrograph of the wear surface8 following abrasion against tbe h8rder U&grit SE are provided in Pig. 4, and examples of wear surfaces produced by tests against tbe softer HO-grit Al,Os arc given in Fig. 5. Comparison of tbe micrograpbs in Figs. 4 and 5 clearly indicate varhtions in the microstructural response ofthedifferentccram&totheabrasive wear environments. For the A-silicon nitride materials, the principal re8ponse to the Sic wear en vironment [Fig. 4(a) and (b)] is plastic deformation, although fmcture at wear groove peripheries and delamination within the grooves is also rexlily rppraent Because of its relatively lower hardness. tbe Sii abrasive particles can penetrate more deeply into the surfa~x of the
CP Dogan, J.. Hawk/wew 203-204(1997)257-277 abled 100 umm Sic: 160m sliding finance) 42N IN A+Sic a)and(b)l, plastic deformation is grooving caused by ab Si,,A, fracture h叫s nese materials, subsurface frac Abrasive Patil Sice (uu) 2. Po of the wear constant vs SC abrasive panick size. nitrides In Si NrB+SiC,, evidence of whisker debonding (ar more Grooves particularly at the surface of the Si, N B composite. response to th 产A site are rela- against 150.grit SiC. Si,NcA monolith than Al,O,+SiC. materials [Fig 5(e) with compos its higher measured wear rate. Against the soter IDU-gnt erogeneities is the dominant mode of material removal. In the
272 Table 4 C.P. Do&n, J.A. Hawk/ Wear 203-204 (1997) 267-277 Influence of applied load on abrasive wear t 100 pm Sic: 16.0 m sliding distance) MathI Wesrconstsnt()rmm-‘) 103 N 91 N 19 N 67 N 54N 42 N 30N Si,N,-A 25.5 23.6 21.6 Si,N.-A + Sic, 26.6 24.4 22.3 19.1 16.1 13.0 9.4 20.8 16.9 13.8 10.4 i 40- g 30- I 8 20 - 3 10 - 0-r Fig. 2 Plot of the wear EcmsliUI? VI. Sic abrasive panicle size. Abrasive Particle Size (pm) Applied Load (fin?) Fig. 3. Plot of tk wear cOnSrant mhuerials against 150~grit Sic. the applied load for tb-. S&N,-A ceramic Si3NrA monolith than into that of the composite, leaving deeper grooves in the wear surface of the monolithic material. However, in spite of the signbicant increase in both the hardness and toughness of the Si,N.-A + Sic, composite. there are no obvious differences in tb- wear mechanisms in the two materials. In fact, the wear surface of the comnosite [Fig. 4(b)] showsmoreevidenceoffracture,consistentwith its higher measured wear rate. Against the softer I50-grit alumina abrasive [Fig. S(a) and (b) 1, plastic deformation is again the primary microstructural wear mechanism, although in this case it is deformation created as the sample surface and drum surface move across one another, with only minimal grooving caused by abrasive particles. In the monolithic Si3N4-A, fracture occurs primarily within the wear sheet (i.e. the damaged surface layer created primarily by plastic deformation during the wear process), with little subsurface material removal. In the A-silicon nitride composite, fracture penetrates the wear sheet, leading to more subsurface material removal. In the B-silicon nitride materials, plastic deformation is again tbe primary response to the UO-grit SE wear envlronment [Fig. 4(c) and (d)], although therelativelylowerhardness of this SisN, series results in deeper penetration by the abrasive particles, zud t!rercforti. a deeper surface damage layer. Although delamination fracture is not as obvious within the wear grooves created in these materials, subsurface fracture at the groove peripheries is more extensive, leading to relatively more material removal than in tbe series A silicon nitrides. In S&N*-B + SE,, evidence of whisker debonding and removal is apparent within the fractured regions. Response to tbe 150-grit alumina environment [Fig. 5(c) and (d)] is similar to that described for the first series of silicon nitride ceramics, except that there are more grooves (a result of the lower hardness) and there is more fracture, particularly at the surface of the S&N.-B composite. For the alumina-based ceramics worn against the l50-grit SIC [Fig. 4(e)-(g) 1, plastic deformation plays much less of a role in the materialsresponse ’ to the wear environment. Because of their relatively high hardness, the wear sheets (primarily due to plastic deformation) produced at the surfaces of 99.8% Al203 and the alumina composite are relatively thin, and the grooves created by the abrasive particles are correspondingly shallow. Delamination fracture of the wear sheet extends into the bulk of the material, resulting in subsurface material removal. The mode of subsurface fracture in both ceramics is almost entirely intergranular. In the lower hardness 99.5% AlrOB [Fig. 4(g)], intergranular fracture dominates the wear response, and any plastic wear sheet produced at the surface is removed almost as soon as it forms. Against the softer ISO-grit alumina abrasive, the response of the 99.8% Al,O, and Al,O, + SiC, materials [Fig. S(e) and (f)] is again primarily plastic deformation, although almost no wear grooves are produced at the surface of the harder composite. However, fracture initiated atmicrostructural heterogencities is the dominant mode of material removal. In the
CP Dofn JA 23-2049267-27 273 岛A+翻cA apainst 100 um SCpartices: (a)SN A: ()Si, NaA+SC (c)Sl, B; (d)S NrB+SC:(e)99 8 Al-O:
C.P. Do&m. J.A. Hawk/ Wear203-204 119547) 267-277 213 fig. 4. Wear surfaces following abrasion against 100 tmt Sii patticks: (a) Si,N& (b) Si,N,-A + Sic; (c) S&N.-B; (d) S&N,-B + Sii; (e) 99.8%40,: (f) A1203 + SC,,; and (g) 99.5% A1203
274 C P. Doda Hawk/War203-204199257-27 AG:0A+s删购1mA0(队A()5AC(8别,(d5B+C(努8
C.P. Do&n. LA. Hawk/ Wear 203-204 (1997) 267-277 Fig. 5. Wear surfaces following abrasion agains4 ALO,: (0 Al,O, + SiC.; and (g) 99.5% AI,O,. Si,N.-a: (b) Si,N,-A+SiC: (c, Si,N.-9: (d) Si,N.-B + Sic,:
C.P. DoBan, /A Hawk/Wew 20j-204459971267-27 the prin. the local fracture toughness (31 cipal microstructural response to the wear environm Si,N,B. The addition of SiC whiskers to these matrices 4. Discussion in the case of S: N A, an increase in bulk fracture toug sker-matrix interface that a whisk parent at the wear surfaces of both Si incrense in the hardness and a 35% increas n than dod of whisker reinforcer mics [6, 7]. In the composite ear environment have on the wear behavior of the debonding and pullout of the SiC 4. 1. influence of microstructure retained the SiC whiskers with their higher hardiess are mate. composit t the s race to enhance t the wear resistance of the ence the tribological per this ds to a difference 1a havior. In Si,N A, the grain boundary hig d in the previous secton Becaise aterial performs equally well under bot ough the ther- tred. it is believed that sults in an increase in the coefficient of thermal expansion compo ot the boundary regons, leading to a concomitant increase n particle sIze, see Hg. 2), t is only hen the abrasive partcle
C.P. Uo&n, J.A. Hawk/ Wear 203-204 (I 997) 267-277 99.5% AIrOr [Fig. 5(g) 1, fracture continues to be the principal microstructural response to the wear environment. 4. Discussion It is clear from the results of this study that a simple increase in the hardness and/or toughness of aceramic-basedmaterial through the addition of randomly oriented whiskerreinforccment does not necessarily translate into improved perfonnante in an abrasive wear environment. For example, the addition of IS vol.% silicon carbide whiskers to the S&N.,-A matrix results in a 22% increase in the hardness of the material and a 20% increase in the fracture toughness. Yet, for the most part, the abrasive wear behavior is degraded by the presence of these whiskers. On the other hand, the addition of SIC whiskers to a high-Al,O, matrix resuhs in a 23% increase in the hardness and a 35% increase in the fracture toughness (in comparison to the 99.8% AI*O,) In this case, the addition of SIC whiskers provides a dramatic improvement in the abrasive wear resistance of the ceramic. A clue to this variation in the effect of whisker reinforcement on the wear behavior of ceramic-based comeosites resides in the micr~structures of these materials, and in the residual stresses created by the addition of second phases to the ceramic. A second important aspect is the influence that the variabics of the wear environment have on the wear behavior of the composite. 4. I. hjluence of microstructure Variations in the microstructures of the monolithic materials. as well as in the microstructures of the composites, can be directly linked to variations in the abrasive wear behavior, primarily through their role in the creation of internal stress. In ceramics it is well known that differences in the thermal expansion between phases of a multiphase material, as well as anisotropic thermal expansion in a single phase material, can result in the creation of residual tensile and compressive stresses at hetero- and homophase boundaries. The magnitude of such residual stresses, which are proportional to both the expansion mismatch (A a) and the temperature range over which the stresses develop (AT). can be sufficient todirectly influence the tribological performance of the bulk ceramic [6-8,25-291. In the unreinforced silicon nitrides of this study, a difference in the grain boundary microstructures leads to a difference in the residual stress state at the grain boundaries of the two materials, and therefore, to a difference in abrasive wear behavior. In S&N,-A, the grain boundary regions have nearly completely crystallized to a-Y,SizO,. whereas in S&X.+-B the grain noundary regions consist of an amorphous yttrium aluminosilicate phase. Although the thermal expansion coefficients of the boundary phases have not heen measured, it is believed that ctystallizationofa-Y&O, results in an increase in the coefficient of thermal expansion of the boundary regions, leading to a concomitant increase in the local fracture toughness [ 311. As a result, the abrasive wear behavior of S&N.-A is improved relative to that of Si3N4-B. The addition of SIC whiskers to these matrices results in an increase in the hardness of the composites and, in the case of Si$$-A, au increase m bulk fracture toughness. However, the narrow glass phase which wets the SE whiskers creates a relatively weak whisker-matrix interface that leads to easy debonding of the whisker during fracturetype events (the origin of enhanced long-crack fracture toughness m this composite). As a result of this debonding, the SIC whiskers are easily pulled from the surface of the composite in an abrasive wear environment, leading to enhanced mabrial removal, and degradation of the wear properties when compared to the monolithic SiJ$. Regions of whisker debonding are apparent at the wear surfaces of both S&N,- based composite materials. In the alumina-baaed ceramics, matrix gram size plays a decisive role in determining the abrasive wear behavior [6- 8.25-291. Thus, the 99.8% AlaOs. with a smaller gram size and narrower grain size distribution, has a much lower wear rate than does the 99.5% AlaOs. Enhancing the grain size effectistheinRuenceofgrainbcumdaryphasesontheresidual stress state of the alumina ceramics [6,7] _ In the composite material, tbe addition of the relatively lower expansion Sic whiskers to the alumina matrix creates residual compressive stresses at the whisker-matrix interfaces. As a result, the whisker-matrix interfaces are locally tougher, resisting the debonding and pullout of the SE whiskets in the ahtasive wearenvironmentunderalltestconditionsofthisstudy.Thus retained, the Sic whiskers with their higher hardness are avai!abb at me surtace to enhance the wear resistance. of the composite material. 4.2. Injluence of the wear environment Variables in the wear environment combine with variables in the material properties to determine the wear resistance of a composite material. Under the relatively “softer” abrasive wearconditionsofthealuminaabrasive,theprimaryresponse of all composites is plastic deformation with only limited fracture. However, the presence of whisker reinforcement in the silicon nitrides tends to increase the amount of fracture observed at the wear surface relative to that of the monolithic materials. This results in a higher measured wear rate and suggests that the relatively weaker SisN&iC interfaces are the origin of this increased surface hnchtre. Under the more aggressive condttions of the harder SE abrasive, SIC whisker debonding and pullout in the S&N.+ composites results in a higher wear rate, as described in the previous w&m. Bccatts~ of tbc stronger matrix-whisker interfaces in the AlaO, + SIC, composite, this material performs equally well under both “soft” and “hard” abrasive conditions. While abrasive particle size does influence the measured wear rate of both the monolithic ceramics and the ceramic composites (increasing the abrasive wear rate withincreasing particle size, see Fig. 2), it is only when the abrasive particle
CP DoMn J.A. Hawk/Wew 203-204(/)257-277 aerial than are the SiaN4+SiC, composites. are smaller, a similar effect is not observed in that material provided by Cercom Inc. We greatly appreciate the provision mic and ceramic composite References U MA Moere and FS KGg Abrasive wea of brite sods, Wear, 60 2】A [3].A, Libet. P.C. Ehaolurtc valuc (eds, ) Adanced Cerumin r per unit are resulted in the power law relationship. w=1072c07 The relationship derived for the Asilicon nitride materials is ed on 60and 180-grit SiC. In tied fro m o. 6i for ahras t. Cerat Soe.. tween those obtained by moore and King mle nerin eempeoiles. [12]G H Campbell M. Ruh study of the influence of randomly oriented Sic m 13] M. Bengisu, O.T. Inal IP. sainik, I,Dusra dition of Sic whiskers to a Si, N4 matrix sets up ten 1cHMm(门比一水m noting casler 117]H, Uiu M face, producing a relatively stronger bond that can better soF419122M4223
C.P. Do&n. J.A. Hawk/ Weor203-204 11997) 267-277 size approaches that of the reinforcement phase in the silicon nitride-based materials that the composite becomes more wear resistant than the matrix material. In this study, this phenomenon is observed only for the S&N.,-A materials in wear tests against 400-grit SIC. where the composite is 25% more wear resistant than the monolithic material. Under these conditions, the Sic abrasive particles arc no longer large enough to fracture and/or scoop the Sic whiskers from the surface of the composite, and the wear rate is correspondingly reduced. Because the SE whiskers in the S&N.,-B composite are smaller, a similar effect is not observed in that material for the range of abrasive sizes used. 4.3. Effect of load on ceramic and ceramic composite wear The S&N.,-A and the Si,N,-A + Sic, materials wereexamined to determine the dependency of wear on load. The wear was measured against 150-g& SIC for applied loads per unit area (u) of between 0.93 and 3.27 MN m-* (see Table 4). The graph of the wear constant (w) vs. applied load per unit area is shown in Fig. 3. For the two A-silicon nihide materials, there is very little difference between the absolute value of rhe w for the monolith and the composite. It is also observed that the value- of w for S&N,-A at each load increment is slightly less than that of the composite. In performing a regression analysis of the combined datacontained in Fig. 3, the calculated wear constant (w) as a function of applied load per unit are resulted in the power law miationship: w= 10.72~~” (3) The relationship derived for the A-silicon nitride materials is similar to that found by Moore and King [ I ] in their analysis of 97.5% A&O3 alumina abraded on 60- and HO-grit SIC. In the Moore and King investigation, the exponent for the applied load per unit area varied from 0.61 for abrasion on the I SO-grit SE to 1.04 for abrasion on the 60-g& Sic. The value obtained in this study for silicon nitride abr; -‘ad on IX-grit Sic falls between those obtained by Moore and King for 97.5% A1203. 5. Conclusions A detailed study of the influence of randomly oriented Sic whisker reinforcement on the abrasive wear behavior of silicon nitride- and alumina-based ceramicr suggests that it is the residual stress state of the whisker-matiix inretiace that determmes the wear r&stance of the composite material. The addition of SIC whiskers to a S&N4 matrix sets up tensile stresses at the whisker-matrix interfaces, enhancing the bulk toughness of the composite, but degrading the abrasive wear properties by promoting easier whisker debonding and removal by the abrasive particles. The addition of Sic whiskerr to an alumina matrix, on the other hand, results in the creation of compressive stresses at the whisker-matrix interface, producing a relatively stronger bond that can better withstand the rigors of an abrasive wear environment. As a result, the Al,03 + SE, composite is consistently a more wear resistant material than are the S&N, + Sic, composites. Acknowledgements The 99.5% A1203 was provided by the Coors Ceramic Company. All other ceramic and ceramic composites were p:ovidcd by Cercom Inc. We greatly appreciate the provision of these materials by the aforementioned companies. References [II MA. Moore and F.S. King. Abrasive weat of brittle solids, Wear. 60 (1980) 123-140. I21 A.G. Evansand D.B. Marshall. in D.A. Rigoey fed.). Fun&men&& of Friction and Wear of Maferials, ASM, Metals Park. OH, 1980. pp. 439-452. [31 T.A. Libsch. P.C. Becker and SK. Rbee. Dry friction and wear of toughened zircottiar end toughened alumhw against steel. Wear. 111 ( 1986) 263-268. 141 J.A. Hawk and C.P. Do&o. in H.M. Hawthorne and T. Tmczyoskt (cd%). Advanced Ceramics for Strucmral and Trib&gical Apphfions. CIM. Monaal. Qwhec. 1995. pp. 139-150. [S] T. Yamamoto. M. Olsson and S. Hogmark. Three-body nh&v::eacai of ceramic !!tzte?ids. Tear. ii4 (1994) 21-31. 161 C.P.Do@mndJ.A. Hawk.Effe~ofgrainbound~glasseomposition end devittilication on tile e%esive wear of AIzO,, Wear, 181-183 (1995) 129-137. 171 C.P. Do&o and J.A. Hawk. in H.M. Hwthome ettd T. Ttoczynski (e&h Advanced Ceramics for Shucnwal and Tribological Apph?ions. CIM. Montreal. Qoebec. 1995. pp. 181-192. [S] S-J. Cho. B.J. Hockey, B.R. Lawn end S.J. Bermsion. Chain size attd R-curve effects in the abrasive wear of elomhw 1. Am Ceram. Sot., 72 (1989) 1249-1252. [91 M. RUhle. B.J. Dalgkish and A.G. &tts. On tbc touglwting of ceramics by whiskers, kripfa Metoll.. 21 ( 1987) 681486. [ 101 PP. Becher. C-H. Hseuh. P. Angeiini and T.N. Tiegs. Toughening behavior in whisker-reinforced caamic matrix composites. _I. Am Cerom. Sot.. 71(1988) 1050-1061. [II] P.G. Chamlambides and A.G. Evans, Debonding pmpetties of residually sttessed btitt&nattix composites. 1. Am. Ccranr. Sot.. 72 ( 1989) 746-753. [I21 G.H. Campbell, M. RUhle. 8.1. Delgleish and A.G. Evens. Whisker toughening: a comparison b&vex% ahnmna oxide and silicon nitride toughened with silicon carbide. 1. Am Cerom. Sot.. 73 (1990) 521- 530. [ I31 M. Bengisu. O.T. lnel and 0. Tosyeli. Oo w!ttker tougixoing to e+voic zxwiais. Acta Met&. Mater, 39 (1991) 2sO9-2517. !I41 P. &tjgaiik. J. Dosw and M.J. Hoffman. Relationship between micfostmcture. tooghentng mechattisrr~. attd ftactuve toughness of ninforccd silicon nitride ceramics. 1. Am. Gram. Sot.. 78 (1595) 2619-26240. II51 C.S.Yost.J.M.L&ttakereedC.E.Devote.Wearofattalumine-silicon carbide whiskercomposits. Wear. 122 (1988) 151-164. [ 161 SC. Farmer. C. Dcllacortc and P.O. Book, Sliding wear of self-mated Al,O,-SIC whisker-reinforced composites at 23-1200 “C, 1. Mater. Sci.. 28 (1993) 1147-l 154. 1171 H. Liu. ME. Fine end H.S. Chettg. Ttibological behavior of Sic whisker/&O, composites against catburized 8620 steel in lubricated sliding, 1. Am. Ceram. Sot.. 74 (1991) 2224-2233
