6-2005-Radial fretting behaviours of dental feldspathic ceramics

Wear259(2005)996–1004Radial fretting behaviours of dental feldspathic ceramics againstdifferent counterbodiesM.H.Zhu a,H.Y.Yu a,b,Z.B.Cai a,Z.R.Zhou a,∗a Key Laboratory of Advanced Technologies of Materials(Ministry of Education),Tribology Research Institute,Southwest Jiaotong University,Chengdu610031,Chinab West China College of Stomatology,Sichuan University,Chengdu610041,ChinaReceived30August2004;received in revised form20January2005;accepted1February2005Available online10May2005AbstractRadial fretting is a fretting phenomenon mainly induced by oscillating normal loads or thermal cycling,where the two surfaces keep in contact and no impact action occurs.In fact,such fretting mode is frequently encountered in oral environment after prosthesis inser-tion.In this paper,radial fretting tests on two types of ball specimens(52100steel and Si3N4ceramics)with diameter of12mm against the dental feldspathic ceramicsflats(Cerec Vita Mark II)have been carried out.Maximum normal load(F max)was varied from100N to 800N with a constant speed of12mm/min,and the number of cycles from1to105.In order to maintain the surfaces offlat and ball specimens in contact and avoid an impact effect,a minimum normal load was set at50N.The kinetics behaviours of the feldspathic ce-ramics under different normal loads were analyzed in details.The fretting scars,which were examined by optical microscope and laser confocal scanning microscope(LCSM),were displayed as the worn zones in shape of annular.The critical normal load of cracks initi-ation was verified.Below the critical normal load,the damage mechanisms were the combination of adhesion and delamination.Above the critical normal load,four damage zones were analyzed.To clarify the experimental results,the analysis of classic Hertzian theory has been discussed.It was also found that the radial fretting damages greatly decreased when a drop of oil was added into the contact inter-faces.©2005Elsevier B.V.All rights reserved.Keywords:Radial fretting;Wear;Cracking;Dental feldspathic ceramics1.IntroductionFretting is small amplitude oscillatory relative movement which occur between contacting surfaces[1,2].Fretting con-ditions are found virtually in any technical system where contact vibrations are induced by fatigue stress,cyclic loads, thermal cycling,acoustical noise or electromagnetic shak-ing.This small oscillatory displacement along the contact surfaces can result in the damages of materials losing by wear or/and cracks initiation by cyclic contact stress[1–3]. In general,the fretting damages are recognized as one of modern plagues in industrial ually,the fretting ∗Corresponding author.Tel.:+862887600971;fax:+862887600971.E-mail address:zrzhou@(Z.R.Zhou).phenomena are very complex in practical industrial applica-tions.For example in nuclear reactors,sometimes the relative movement is not simply tangential but is complicated by the normal force also oscillating,to the extent that the surfaces lose contact in each cycle.This leads to a hammering ef-fect which is termed“impact fretting”[4–6].Therefore,the relationships of the directions of relative movements along the contact surfaces are important factors for scientific re-searches.According to the directions of relative movements of ball-on-flat contact,the fretting modes can usually be di-vided into four basic types,i.e.,tangential,radial,rotational and torsional fretting.Also,some researchers characterized the fretting modes into three types,i.e.,linear(tangential),ra-dial and circumferential(torsional),and termed them Mode I,Mode II and Mode III,respectively[7].0043-1648/$–see front matter©2005Elsevier B.V.All rights reserved. doi:10.1016/j.wear.2005.02.010M.H.Zhu et al./Wear 259(2005)996–1004997Fig.1.The scheme of force direction of mastication:(a)inciser teeth;(b)molar teeth.Up to now,investigations of fretting are mostly focused on tangential mode,and the main theories of fretting wear and fretting fatigue are built up under this mode.In fact,the phe-nomena of radial fretting are usually presented at the actual industrial components,such as “false brinelling”occurring in rolling bearings,fuel rod/spring contacts in pressurized water reactors,leaf springs in automobiles and railway lo-comotives,jet and diesel engine components,electrical con-tacts,power plants and orthopedic implants.Fretting under oscillating normal loads or thermal cycling,classified as ra-dial fretting,occurs where the two surfaces are kept in con-tact and no impact effect occurs [8].Under radial fretting,micro-slip occurs in a contact zone delimited by the mini-mum (inner)and maximum (outer)contact radius.In buccal condition,the movement of radial fretting also frequently encountered in the process of mastication.Fig.1shows the schematic of the forces in radial direction for inciser and mo-lar teeth.It should be noted that Kennedy [9],Johnson [10]and Tyler et al.[11]had earlier used oscillatory normal loading in contact fatigue studies before almost half a century.Al-though the radial fretting phenomena had been observed in the earlier investigations of contact fatigue,the friction be-haviours between contact interfaces were ignored.Recently,the research groups of Belgian [12]and ours [8]had designed and constructed two new radial fretting test systems,respec-tively.Some experimental results had been obtained from the contacts of TiN coating,high-speed steel (Ti,Al)N coat-ing and soft tin coating against Al 2O 3ball [7,12],steel,alu-minium alloy,TiN coating and MoS 2coating against 52100bearing steel ball [8,13].However,it is not enough to estab-lish the theory system of radial fretting.For dental ceram-ics,the tests of single-cycle or multi-cycle Hertzian contact [14,15]were usually used to detect the toughness,but no researchers paid attention to the radial fretting phenomenon occurred in the contact interfaces up to now.Detail investi-gations of ceramics materials under radial fretting condition have not been reported yet.Therefore,the purpose of this paper is to understand the radial fretting behaviours of brittlematerials.Fig.2.Schematic diagram of radial fretting test rig.2.Experimental procedureThe test apparatus,which was developed from a tension-compression hydraulic machine and had been described in detail in reference [8],was designed for a ball-on-flat contact configuration,as seen in Fig.2.The flat specimen was fixed on the upper holder linked to the load cell.The ball specimen was fixed at the lower holder and was mounted on the piston,which subjected to oscillatory movement with a given speed of 12mm/min.Cyclic deformation between two contact sur-faces (i.e.,the relative displacement of contact system)was measured by the extensometer.In order to maintain the two surfaces in contact all the time and avoid an impact effect,a minimum normal load (F min )was set at 50N.Maximum nor-mal load (F max )was chosen as 100N,200N,300N,400N and 800N,respectively.In the radial fretting test,the num-ber of cycles varied from 1to 105.The whole fretting process was computerized and variations of deformation (indentation depth)versus normal load can be recorded as a function of cycles.52100bearing steel and silicon nitride (Si 3N 4)ceram-ics with diameter of 12mm were used as spheric specimen.The dental ceramics flats of the Cerec Vita Mark II (ce-ramic solid,modified feldspar frits and inorganic pigments molten in matrix.)were machined as the flat specimen in the size of 12mm ×12mm ×15mm.The main mechanical properties for three materials are shown in Table 1.For the flat specimen,the elastic modulus and hardness were de-tected by nano-indentometer (CSEM)under maximum load of 100mN at loading rate of 200mN/min,and the fracture toughness was obtained by three-point bending examina-tion.The flat surfaces were carefully polished to a final rough-ness (R a )of about 0.12␮m.The samples were ultrasonically cleaned in hexane.The fretting tests were performed under temperature of 23◦C and relative humidity of 60%±10.Af-ter the fretting tests,the morophologies of fretting scars were998M.H.Zhu et al./Wear259(2005)996–1004Table1The main mechanical properties of ball andflat specimensMaterials Elastic modulus(GPa)Hardness(HV)Fracture toughness(MPa m1/2)Poisson’s ratio 52100steel208790250.3Si3N430017006.50.26Cerec Vita Mark II64.28710.920.3examined by optical microscopy and laser confocal scanning microscopy(LCSM).3.Experimental results3.1.Kinetic behaviours3.1.1.F–D curvesThe curves of normal load(F)against displacement(D), which can be recorded as a function of cycles,supplied the most important information of deformation behaviours of the contact interfaces during the radial fretting test processes [8,13].In thefirst cycle of radial fretting,the load was applied from0to F max,and then unloaded to F min.The important in-formation about materials deformation was obtained through the F–D curves of thefirst cycle.Fig.3shows the F–D curves of Cerec Vita Mark II ceramics against52,100steel ball ob-tained after thefirst radial fretting cycle.It can be founded that the loading and unloading curves of the F–D curve were superposed under the normal load of F max=100N(Fig.3(a)), suggesting that only elastic deformation occurred.When the maximum normal load increased to200N,the F–D curve was open(Fig.3(b)),corresponding with the plastic defor-mation occurring between the surfaces of two specimens. The residual deformations of the open F–D curves,which in-creased with the increasing in the normal loads,reflected the extent of plastic deformation.Under the higher normal loads (F max=400N and F max=800N,as seen in Fig.3(d)and(e)), some zigzags can be observed in the loading curve,resulting from some cracks of ceramics initiated and propagated under higher contact stress.For two contact pairs,under lower normal loads (F max=100N and F max=200N),all F–D curves closed near to a line in the whole subsequent fretting cycles(Fig.4).The relative movement between two contact pairs was fully ac-commodated by the elastic deformation.From Fig.3(c),it can be found that the loading curves departure from the line after the normal load reached250N.It might be inferred that some energy dissipated due to the plastic deformation of con-tact pairs during the radial fretting cycles when the maximum normal load exceeded250N.Fig.5displayed that the open F–D curves last to no more than100cycles under the con-dition of F max=300N,i.e.,the deformation accommodated by the elastic deformation after100cycles.With the increase of the normal load,the duration of the open F–D curves in-creased significantly.For example,under F max=400N,the number of the cycle which the F–D curve transformed from the open to the closed was5000(Fig.6).While the F max in-creased to800N,all the F–D curves were open before the number of cycles was equaled to105,as shown in Fig.7.Fig.6shows that the F–D curves were partitioned to two parts which corresponded with different contact stiffness after 5000cycles,resulting from the maro-cracks propagated due to the accumulation of cyclic deformation.In Fig.6(d),the bended F–D curve was perfectly replaced by line AB and CD with different slopes.In general,the contact stiffness,which can be simply calculated from the formula of K= F/ D, reflected the elastic properties of the counterfaces.Therefore, different elastic behaviours have occurred in radialfretting Fig.3.The F–D curves of thefirst cycle under different normal loads:Cerec Vita Mark II vs.52100contact.M.H.Zhu et al./Wear259(2005)996–1004999Fig.4.The F–D curves of subsequent cycles:Cerec Vita Mark II vs.52100contact,F max=200N.Fig.5.The F–D curves of subsequent cycles:Cerec Vita Mark II vs.52100contact,F max=300N.Fig.6.The F–D curves of subsequent cycles:Cerec Vita Mark II vs.52100contact,F max=400N.1000M.H.Zhu et al./Wear 259(2005)996–1004Fig.7.The F –D curves of subsequent cycles:Cerec Vita Mark II vs.52100contact,F max =800N.Fig.8.Variations of the displacement as a function of the number of cycles.under different levels of normal load.Under higher maximum normal load of 800N,this situation can be observed in all radial fretting cycles (Fig.7).3.1.2.D–N curvesFig.8shows the displacement (D )as a function of the number of cycles (N )for two contacts.It can be found that a lower radial displacement corresponded to a lower normal load.However,when the normal load increased to a higher level,the higher displacement corresponded to severe dam-ages and large difference on displacement levels for two con-tacts.Due to the smaller elastic modulus for 52100steel ball than that of Si 3N 4ceramics,the displacement for the CerecVita Mark II against 52100contact is always lower than that of Cerec Vita Mark II against Si 3N 4contact (Fig.9).The higher the normal load was,the larger difference on displacement existed.As shown in Fig.10,no obvious discrepancy on the ratio of relative displacement of two contacts was presented under lower normal loads (F max =100N and F max =200N)owing to the only elastic deformation on the dental ceram-ics.When the normal loads reached and exceeded 300N,the discrepancy quickly increased with the rising of the normal loads (Fig.10).In general,the displacement decreased with the increase of the number of cycles in radial fretting test,which was ob-tained under the test materials of metals [8,13].In ourtests,Fig.9.Variations of the displacement as a function of the number of cycles under different maximum normal loads.M.H.Zhu et al./Wear259(2005)996–10041001Fig.10.The ratio of relative displacement of two contact pairs under the different maximum normal loads:r=D52100/D Si3N4.an abnormal phenomenon can be observed under the condi-tion of higher normal loads(F max=400N and F max=800N) as seen in Fig.9(d)and(e).This was maybe the results of the rapid propagating of cracks.3.2.Surface damage observationsConcerning to radial fretting,the micro-slip occurred in an annular zone which formed between the maximum and min-imum contact circle.Under lower normal loads,the fretting scars exhibited the worn zones in shape of annulus as shown in Fig.11where bright area was the worn zones and the black spots were the defects(micro-cavity)of dental feldspathic ceramics.In Fig.11(b),nonhomogeneous wear was observed in whole slip zone due to nonhomogeneous stress distribu-tion between the contact interfaces.It is very clear thatno Fig.12.The profile of fretting scar:Cerec Vita Mark II vs.52100contact, F max=300N,105cycles(corresponding to Fig.11(b)).macro-cracks can be formed in the contact zones after105 cycles under the conditions of F max≤300N(Fig.11).With the increase of the maximum normal load,the extent of dam-age and the width of the annulus increased obviously,as seen in Fig.11.In Fig.12,the profile of the micro-slip zone which corresponded to the image of Fig.11(b),was obviously over-topped the surface of theflat specimen.This maybe the results of the materials transferring from its counterbody by adhe-sion wear mechanism.Obviously,the occurrence of material transferring was easier for the contact pairs of steel ball than that of Si3N4ball.Under higher normal loads(F max≥400N),the cracks formed in shape of homocentric circularity at the boundary of the maximum contact zone,as shown in Fig.13.Because the cracks initiated atfirst cycle under these test conditions, the cracks can be detected on the surfaces of specimens after several cycles(Fig.14).Although obvious residual deforma-tion occurred at thefirst cycle under F max=300N,no cracks had been observed after105cycles,as shown in Fig.11.To Fig.11.LCSM photomicrographs of the fretting scars:(a)Cerec Vita Mark II vs.52100contact,F max=200N,105cycles;(b)Cerec Vita Mark II vs.52100 contact,F max=300N,105cycles;(c)Cerec Vita Mark II vs.Si3N4contact,F max=100N,105cycles;(d)Cerec Vita Mark II vs.Si3N4contact,F max=300N, 105cycles.1002M.H.Zhu et al./Wear 259(2005)996–1004Fig.13.LCSM photomicrographs of the fretting scars:(a)Cerec Vita Mark II vs.52100contact,F max =400N,105cycles;(b)Cerec Vita Mark II vs.52100contact,F max =800N,105cycles;(c)Cerec Vita Mark II vs.Si 3N 4contact,F max =400N,105cycles;(d)Cerec Vita Mark II vs.Si 3N 4contact,F max =800N,105cycles.Fig.14.LCSM photomicrographs of the fretting scars:Cerec Vita Mark II vs.Si 3N 4contact,F max =800N,5cycles;(a)lower magnification;(b)higher magnification.compare Fig.11with Fig.13,it was assumed that there was a critical normal load resulting in homocentric circular cracks initiation under N =105.Thus,the critical normal load clearlyintervened between 300N and 400N,which means that the highest stresses appeared at the outer area of contact.Figs.11(b)and 13indicated that the damage zones of the fretting scars could be plotted into four regions,i.e.,from cen-ter to exterior:contact center,slight wear zone,severe wear zone and cracking zone,as shown schematicly in Fig.15.In the center of contact,the interface of two counterparts were kept in contact all the time,where no damage was detected.In severe worn zone,a great deal of particles in shape of flake usually in thickness about 0.5␮m which detached by delami-nation [16]was founded in Fig.16.The coexistence of cracks and wear was the essential feature in cracking zone as seen in Fig.17.The feature of wear was substantially different from the configuration spalled by high contact stress (Fig.14(b)).Fig.15.Schematic diagram of damage zones of ceramics under condition of radial fretting.M.H.Zhu et al./Wear 259(2005)996–10041003Fig.16.The morphography of delamination in the fretting scars:Cerec Vita Mark II vs.52100contact,F max =400N,105cycles;(a)in mode of brightness;(b)in mode of height;(c)profile of lineAB.Fig.17.The morphography of cracking zone in the fretting scars:(a)Cerec Vita Mark II vs.52100contact,F max =400N,105cycles;(b)Cerec Vita Mark II vs.Si 3N 4contact,F max =800N,105cycles.Obviously,the wear mechanism during the cracking zone was also delamination.To compare Fig.14with 17,it is very easy to deduce that the formation of cracking was earlier than that of wear under high normal loads.From the boundary to the contact center,the interspaces of cracks increased gradually from several micrometers to several decades micrometers as shown in Fig.17(b).4.DiscussionsAccording to the experimental results mentioned above,the damage mechanism of dental ceramics under radial fret-ting can be marked into two cases of different normal load levels.When F max ≤300N only wear occurred through the mechanisms of combination of adhesion and delamination.But when F max ≥400N the cracking formed prior to wear and accelerated the detachment of particles.Generally,the formation of cracks is quite dangerous for engineering ap-plications of ceramics.Therefore,the ceramics must be used below the critical contact stress under radial fretting.Never-theless,up to now there is no method to detect the fatigue properties of engineering ceramics under dynamic contact conditions.According to the above results,the critical load or stress for cracking initiation can be obtained by testing ra-dial fretting under different parameters.Also,the fatigue life curves can be obtained under nonhomogeneous deformation.Therefore,the radial fretting test will be proposed as a new and potential method to evaluate fatigue life of ceramics un-der dynamic contact conditions.In the compressive stress distribution given by classic Hertzian theory [17],the maximum value lies in the cen-ter of contact zone under the contact of ball-on-flat.As a balance,there is a tensile stress distribution at the boundary of the outer contact consequentially.This is the reason why the cracking zone formed at the outer contact region.Surface shear stresses were calculated from the principal stresses of Hertzian theory by Tyler et al.[11].The octahedral shear stresses of surface of flat specimen at different contact radii were calculated in terms of the principal stresses.The oc-tahedral shear stress was maximum at the outer rather than inner contact radius.During a cycle of oscillatory normal loading,the stresses variation formed the higher peaks just in the middle between the inner and outer contact radius.In other words,it was revealed that the maximum shear stresses appeared in an intermediate position of the micro-slip zone,i.e.,corresponded to the severe wear zone.No friction is one of basic hypotheses in Hertzian theory.Since the worn of materials were the basic phenomena under radial fretting,the friction force can not be neglected during the damage processes,while it is balanced by its symmetrical distribution.Obviously,our results are in good agreement with the stresses analysis provided by Hertzian theory.In order to understand the effect of friction force on damages,a drop of lubrication oil was put between the interfaces of contact pairs.The damage extent was much slighter than that of dry condition under the same test parameters,as shown in Fig.18(compared with Fig.13(a)).It is indicated that the material damages were weakened greatly with the decrease of friction force by lubrication.1004M.H.Zhu et al./Wear259(2005)996–1004Fig.18.LCSM photomicrograph of the fretting scars under oil lubrication: Cerec Vita Mark II vs.52100contact,F max=400N,105cycles.Therefore,the friction force is impossible to change the es-sential structure of the stresses distribution.The actual stress distribution appended the effect of friction force is difficult to get an analytic solution.To expect,a new model needs to be built up to calculate the stress distribution of radial fretting by numerical calculation methods in our future work.5.Conclusions1)For the dental ceramics of Cerec Vita Mark II,the defor-mation behaviours were strongly depended on the levels of maximum normal loads under radial fretting.When the maximum normal load was equaled to100N,only elastic deformation occurred in all cycles.When the maximum normal load increased to200N,residual deformation can only be detected in thefirst cycle.The plastic deforma-tion could be observed in thefirst and subsequent fretting cycles under higher normal loads.When the maximum normal load reached and exceeded400N,the propaga-tion of cracks can be reflected in the D–N curves.2)The critical normal load of cracks initiation under our testparameters was verified to be between300N and400N.Below the critical normal load,no macro-cracks can be detected in the fretting scars.It is suggested that the ra-dial fretting test is a potential method to evaluate fatigue properties of ceramics under dynamic contact conditions.3)Under lower normal loads(below the critical normalload),the damage mechanisms were the combination of adhesion and delamination.Under the higher normal loads(above the critical normal load),the fretting scars can be plotted into four regions,i.e.,contact center,slightwear zone,severe wear zone and cracking zone.The de-lamination was the main damage mechanism in the wear zone,and the combination of cracking and delamination in the cracking zone.4)Under oil lubrication,the extent of radial fretting damagesgreatly reduced.AcknowledgementsThis work was supported by National Natural Science Foundation of China(No.50105017and No.50305030)and Sichuan Youth Science&Technology Foundation of China (No.03ZQ026-062).References[1]R.B.Waterhouse,Fretting Fatigue,Appl.Sci.,London,1981.[2]R.B.Waterhouse,Fretting wear,in:ASM Handbook:Friction,Lubri-cation and Wear Technology,vol.18,ASM International,Clevelaud, OH,1992,p.233.[3]Z.R.Zhou,L.Vincent,Cracking induced by fretting of aluminiumalloys,J.Tribol.,ASME119(1997)36.[4]P.L.Ko,The significance of shear and normal load force componentson tube wear due to fretting and periodic impacting,Wear106(1985) 261.[5]D.H.Jones,A.Y.Nehru,J.Skinner,The impact fretting wear of anuclear reactor component,Wear106(1985)139.[6]G.Levy,J.Morri,Impact fretting wear in CO2-based environments,Wear106(1985)97.[7]M.Z.Huq,C.Butaye,J.P.Celis,An innovative system for frettingwear under oscillating normal force,J.Mater.Res.15(2000)1591.[8]M.H.Zhu,Z.R.Zhou,An experimental study on radial fretting be-haviour,Tribol.Int.34(2001)321.[9]N.G.Kennedy,Fatigue of curved surface in contact under repeatedload cycles,in:Proceedings of the International Conference on Fa-tigue of Metals,London,New York,1956,p.282.[10]K.L.Johnson,Energy dissipation at spherical surfaces in contacttransmitting oscillating forces,J.Mech.Eng.Sci.3(1961)362. [11]J.C.Tyler,R.A.Burton,P.M.Ku,Contact fatigue under oscillatorynormal load,ASLE Trans.6(1963)255.[12]H.Mohrbacher,J.P.Celis,J.R.Roos,Laboratory testing of displace-ment and load induced fretting,Tribol.Int.28(5)(1995)269. [13]M.H.Zhu,Z.R.Zhou,Ph.Kapsa,L.Vincent,Radial fretting fatiguedamage of surface coatings,Wear250(1–12)(2001)650.[14]F.Guiberteau,N.P.Padture, wn,Effect of grain size onHertzian contact in alumina,J.Am.Ceram.Soc.77(1994)1825.[15]I.M.Peterson,A.Pajares,wn,V.P.Thompson,E.D.Rekow,Mechanical characterization of dental ceramics using Hertzian con-tacts,J.Dent.Res.77(1998)589.[16]K.P.Suh,An overview of the delamination theory of wear,Wear44(1977)1.[17]K.L.Johnson,Contact Mechanics,Cambridge University Press,Great Britain,1985.。