Dry sliding wear behavior of TIG welding clad WC composite coatings

Dry sliding wear behavior of TIG weldingclad WC composite coatingsSoner Buytoz a ,Mustafa Ulutan b ,*,M.Mustafa Yildirim aaUniversity of Fırat,Faculty of Technical Education,Department of Metallurgy,Elazıg˘23119,Turkey bUniversity of Osmangazi,Faculty of Engineering,Department of Mechanical Engineering,Bat Meselik,Eskisehir 26480,TurkeyReceived 20December 2004;received in revised form 14February 2005;accepted 14February 2005Available online 25March 2005AbstractIn this study,melted tungsten carbide powders on the surface of AISI 4340steel was applied by using tungsten inert gas (TIG)method.It was observed that it has been solidified in different microstructures depending on the production parameters.As a result of microstructure examinations,in the surface modified layers an eutectic and dendrite solidification was observed together with WC,W 2C phases.In the layer produced,the hardness values varied between 950and 1200HV .The minimum mass loss was observed in the sample,which was treated in 1.209mm/s production rate,0.5g/s powder feed rate and 13.9kJ/cm heat input.#2005Elsevier B.V .All rights reserved.Keywords:TIG method;Surface modification;Abrasive wear;Dendrite;WC;M 7C 31.IntroductionSurface engineering contains the surface processes,which are applied to improve the surface characteristics of the material and hard coating methods [1].The coating materials applied on different substrates have recently become widespread in the automotive industry,petrochemistry and space technology fields [2].The risk of microcrack formation and ability of low plasticdeformation in the coatings have constituted main problems [3].It is possible to reduce the risk of cracking in the main matrix by applying pure powder,wire or their alloys on the surface of the elastic material [4].Together with coating methods,such as boronizing,carburizing,carbonitriding,nitriding [5,6],plasma based physical vapor deposition (PVD)and chemical vapor deposition (CVD),applied to obtain a fine,hard,wear resistant layer;it has been manufactured with the laser [7],plasma [8],powder spray [9]and tungsten inert gas (TIG)methods [10],which are called to be semi-melting methods.The alloying method by means of TIG,which is stated as surface modification is/locate/apsuscApplied Surface Science 252(2005)1313–1323*Corresponding author.Tel.:+902222393750/3528;fax:+902222290535.E-mail address:mulutan@.tr (M.Ulutan).0169-4332/$–see front matter #2005Elsevier B.V .All rights reserved.doi:10.1016/j.apsusc.2005.02.088principally a method,which was produced by melting the alloy powder or powders having proper composition placed on the substrate material surface.Melting takes place at the same time in both,modified coating material and base metal,which bonded with each other in a metallurgical way [10,11].Fine-grained micro-structures,which solidified rapidly contain hard carbides phases [12].Tungsten carbide is particularly preferable in the wear resistant cases due to the fact that it provides the characteristics of good wear resistance,high hardness,low thermal dilution,a little amount of plasticity and good wettability by molten steels [13].In the Fe–Cr–W–C based coatings improved on low-carbon steel with the laser method depending on production conditions,hypoeutectic and hypereutecticmicrostructures were observed [14].Lo and his colleagues [15]improved coarse-grained tungsten carbide powders on UNS S31600stainless steel.WC,W 2C,M 7C 3,M 6C and M 3C carbides were observed in the microstructure,which were formed after coating the WC grains in different laser power densities and 3–28%vol.%rates.It was seen that this kind of carbides had significantly more contribution to the wear resistance compared to substrate material.However,in the thermal spray techniques like HVOF,brittle W 2C phase is formed when WC grains and metal powder grains are sprayed on the coating zone [16,17].In addition to this,when thermal spray techniques are used,it limits the work field because it is really difficult to prevent porocity development in the coatings [18,19],and also it dissolves again in corrosive environment and high temperatures such as 5008C.This paper introduces tungsten carbide coatings on AISI 4340steel produced with a tungsten inert gas torch.To avoid the oxidation of the alloyed coating,and to provide a relatively inert environment,argon gas was blown directly into the molten pool and to the processing region.Cladding layer thickness was about 1.5mm.EDX,X-ray,spectral and SEM analysis were used to assess the compositions of coating layer.Besides the microstructure,micro hardness and effects of adding tungsten carbide particles on microstructure were presented.The effects of the powder content and heat input value on abrasive wear rate of the coatings were also investigated.S.Buytoz et al./Applied Surface Science 252(2005)1313–13231314Fig.1.Schematic representation of the TIG welding clad process and the coating layer.Table 1Chemical compositions of AISI 4340steel and WC powder (wt.%)used for surface alloying Chemical compositions (wt.%)AISI 4340Cr 0.930Si 0.267C 0.415Mo 0.237Mn 0.656Ni 1.75Fe95.54WC (99%pure)WC 99.435Cr 2O 30.015FeO 0.38TiO 20.17Table 2Process parameters Current/voltage 120A/20VScan speed 1.209–1.510mm/s Shielding gas 99.9%pure Ar Gas flow rate 11l/minElectrode2%thorium tungsten electrode Electrode diameter2.4mmTable 3Spectral analysis results taken from the surfaces of the samples to which surface coated WC Sample number Chemical compositions (wt.%)C Mn Cr Ni Si W Fe S 2 1.0560.097 1.35 1.4090.53219.20Bal.S 30.6750.095 1.070.9800.75615.25Bal.S 40.7500.1041.861.3350.75011.78Bal.2.Experimental methodIn this study,tungsten carbide powder and AISI 4340steel specimens in dimensions of100mmÂ20mmÂ15mm were used for coating process. Preplaced powder layer and the coating was produced by using a tungsten inert gas torch shown schematically in Fig.1.The chemical analyses of the material and powder used in the experiment are shown in Table1, and the experimental conditions are shown in Table2. The alloying process has been carried out with the two-step method.Initially,the sample surface was polished with400grit SiC papers,then the mixed powder layers in amounts of0.5,1,2and4g were covered with a paste of the alloying powder by adding small amount of alcohol for binding of powders.Then,the substrates with this coating were dried in furnace at a temperature of508C for60min.It was enabled to keep the powder layers on the surface under theflow of argon gas during the arc melting and to vaporize the chemical binder before the treatment.Subsequently,the coating has been melted using the TIG weld method.Specimens for metallographic examinations were taken from cross section of the modified surface from the coating areas.The metallographic specimens obtained were rubbed with80–1200mesh sandpaper; as a result,their surfaces were cleaned and then cross-sectional surfaces were polished by1and6m m diamond paste and solvent.The polished specimens were etched with ferric chloride solution(25g of FeCl3,15ml of HCl and100ml of distilled water). The microstructures of the specimens were observed by scanning electron microscope(SEM),energy dispersive spectra(EDX),X-ray diffraction(XRD) and spectral analyses were benefited from for microstructure and elemental analysis of the modified surface.XRD analyses were obtained using Cu–K a radiation in30kV and15mA with Rigaku Geigerflex X-ray diffractometer.Abrasive wear experiments were done at the room temperature and normal atmosphere conditions with dry sliding pin-on-disc experiment apparatus.Prior to wear experiments,each sample was exposed to1200 mesh sandpaper in order to make wear surface to touch abrasive surface completely.In the experimental procedure,three different loads of10,20and30N were exposed to80mesh abrasive wear in four different sliding distances such as15,30,60,and 120m.After each experiment,80mesh abrasive sandpapers were changed.Abrasive wear rates were recorded by determining the weight losses of the samples before and after the abrasive tests.3.The results3.1.Microstructure resultsWhen the microstructures of the layers with tungsten carbide which were modified with tungstenS.Buytoz et al./Applied Surface Science252(2005)1313–13231315 Table4Mechanical results of coating surfacesSample number Process parameters Coating dimensions(m m)Hardness(HV) Process speed(mm/s)Powder feed rate(g/s)Heat input(kJ/cm)d c h c t c w cS1Untreated material224S2 1.2090.513.9173085025809650965S3 1.380.512.21450870232079801156S4 1.5100.311.11810550236080101184d c,max depth below the original substrate surface;h c,max height above the original substrate surface;t c,max total thickness(=d c+h c);w c,max width;heat input Q=h.U.I.60/(V.1000).Table5EDX analysis results of samples to which surface coated WCSample number Chemical compositions(wt.%)C Cr W FeS2Carbide 5.84 3.7585.15 5.260Matrix 3.757.6038.2550.40S3Carbide 2.55 4.6562.8529.95Matrix 1.98 3.8231.1063.10S4Matrix 4.398.3068.4318.88Dendrite 1.23 5.459.8283.50inert gas welding method was examined,depending on the production parameters they were seen to have been solidified in varying composition and amount of carbides and phases.The spectral analysis data taken form 0.5mm beneath the layer surfaces including WC after the surface modification were shown in Table 3,and EDX analyses were shown in Table 5.When the modification conditions such as powder feed rate and energy input values,the chemical composition of the composite upper layer also changed (Tables 3and 5).As reveals in Table 3,when the powder feed quantity on the material surface increased,the percentage rates of the W,C,Cr,and Fe elements increased compared to chemical concentration of untreated AISI 4340steel.In the S 2and S 3samples modified in the same quantities of powder rate but in different heat input values it was observed that when the production ratesS.Buytoz et al./Applied Surface Science 252(2005)1313–13231316Fig.2.Microstructure of S 2specimen produced with TIG method.(a)Upper layer (250Â),(b)coating layer (250Â),(c)SEM photo-graph of coating layer/interface layer/substrate (100Â).Fig.3.Microstructure of S 3sample produced with TIG method.(a)Coating layer (400Â),(b)upper layer (250Â).changed(Table4),the amount of the dissolved element quantities in the microstructure also changed (Table5).This resulted from different solidification rates due to changes in the energy inputs given to the modified layer.Fig.2represents the microstructure of S2specimen. Although,there was a carbide structure to be observed on the surface layer modified,it was recorded that the dendrites were dense in a small area(can be seen in Fig.2(a)).While hypereutectic structure with high W element dominates in inter-carbides areas,carbides formed having rich-W,Fe and Cr content(Table5). Carbides distributed homogeneously along the upper surface(Fig.2(b)and(c)).In the formation of carbides having rich-W,-Fe and-Cr,depending on the high-energy input and low production speed in the modified area,it was seen from the EDX results that the melting amount of lower layer increased and as a result Fe rate increased.Furthermore,in the literature,there are similar studies showing that depending on the production parameters and mixture rate,when the production speed is low,the lower plate melts more and subsequently it forms coating layer by solidifying [14,20].According to W–Fe–C phase diagram[21], WC+a-Fe+h(Fe3W3C)phases may be formed in the microstructure.In Fig.4(a)showed that WC, Fe3W3C,M2C,M7C3and g-FeCrNiW carbides and phases were formed according to XRD results,in the modified layer improved with low powder rate andS.Buytoz et al./Applied Surface Science252(2005)1313–13231317Fig.4.X-ray diffractograms of(a)S2specimen(b)S3specimen.Fig.5.Microstructure of S4specimen produced with TIG method.(a)Upper layer(250Â),(b)solidified dendrite(like the star shape)(250Â).higher energy input value.In the coatings improved with the laser method,WC,W 2C,M 6C (M =Fe,W,Cr),M 7C 3and M 23C 6complex metal carbides were seen [22].The chemical composition of the matrix is important due to dissolution of the WC grains and the reactions,which can be formed in the composite layer.In this type of coating,while the amount of C had little effect on WC dissolution and the formation of the W 2C and M 6C phases in the present material,the effect Fe,W,Cr,and Co elements was relatively high and the most effective one was molybdenum [15–17].Adding the molybdenum to the structure,hardening ability has been increased along with the formation of the M 2C and M 7C 3carbides in the microstructure [23].The microstructure of the S 3specimen produced using the WC powders with 12.2kJ/cm energy input,0.5g stable powder content (compared to S 2speci-men)and 1.38mm/s process speed was given in Fig.3.The coating material exhibited a different micro-structure after solidification,depending on the heat input value (Fig.3(a)).According to Table 5,while the carbides include 1.98%C,3.82%Cr,31.25%W and 63.10%Fe matrix carried out hypoeutectic solidifica-tion containing 2.55%C,4.65%Cr,62.85%W and 29.95%Fe concentration.The carbide grains are larger in the specimen having S 3hypoeutectic microstructure formed the same quantity of the powder melted,relatively less heat input but highS.Buytoz et al./Applied Surface Science 252(2005)1313–13231318Fig.6.Hardness variation of WC coated specimens from surface tosubstrate.Fig.7.Weight loss variation of WC coated specimens 10,20and 30N load,respectively,in contrast to wear distance.process speed(Fig.3(b)).Melted grains may becomesolidified by deforming tungsten carbide distributionat varied temperature and may cause differentmicrostructures on the coating layer.Neither anincrease in the grain dimension nor preparation of onephased coating is the structure desired[24].The X-raydiffractograph of the S3specimen was given in Fig.4b.In the microstructure of the material to which surfacemodification applied,WC and W2C types,most ofwhich were M7C3(M=W,Fe)carbides,wererecorded together with austenitic hypoeutectic struc-ture.The present results are in harmony with the otherstudies gained by means of coating[22,23].The microstructure of the S4specimen can be seenin Fig.5.The composite layer modified showed a verydifferent microstructure compared to S2and S3specimens.The low heat input and high processspeed partly caused the substrate to melt and causedthefinal structure dendritic(Fig.5(a)).Besides,due tohigh process speed some of the WC grains remainedand not dissolved in the melt bath.According toTable5,it was composed of4.39%C,8.30%Cr,68.43W and18.88Fe concentrated matrix togetherwith the dendrites having the contents of1.23%C,5.45%Cr,9.82W and83.50%Fe.It was observed thatthe increase in the amount of W in the dendrites waslow.Due to the fact that dendrites are thefirst to besolidified,the extra W in the melt may solidify in onlyS.Buytoz et al./Applied Surface Science252(2005)1313–13231319Fig.8.Relationship between applied load and total weight loss of WC coatedspecimens.Fig.9.Wear rate variation changed with wear distance with respect to(a)10N,(b)20N and(c)30N load of WC coated specimens.eutectic areas.The increase in the amount of the WC is found to be parallel to that of the dendrite formation amount [7].The added WC grains normally melt and dissolve in melt bath again.During the cooling,microstructures having solidified again were obtained.Due to the amount of WC grains or high process speed,it can be seen as precipitate in the form of the grains,which have not melted in the structure or mixed carbides [13].For that reason,in the modified area,it can be concluded that the energy input and production speed given did not have enough time to let WC grains to melt and solidify again in the melt bath.The chemical composition of the matrix is important because of the dissolution of the WC grains and the reactions to be formed on the composite layer.As a result,in the microstructure of the material with the WC modification,depending on the production parameters and powder amounts,two significant solidification characteristics were found.First,when the process speed was low and energy input was high for WC modification,g -FeCrNiW phases were obtained along with M 7C 3,M 23C 6,W 2C and Fe 3W 3C carbides distributed orderly.When the process speed changed,in the microstructure of the modificationlayer,hypoeutectic structure was formed together with WC,W 2C,M 7C 3carbides.However,when the process speed was really high and the quantity of the powder was low,dendrite and inter-dendrites eutectic,which was poor in terms of W concentration,was formed.3.2.Micro hardness resultsIn Fig.6,hardness values taken from 0.5mm beneath the S 2–S 4samples surfaces on which surface modification improved with WC powders to the center depending on the distance were given.The coating micro hardness results for different production para-meters were also shown in Table 4.Although the distribution of hardness in the specimens,the surface of which was modified,showed a homogeneous distribu-tion,a decrease observed in the interface area.Hardness distributions on the upper layer found to be varying between 1000and 1200HV .When the carbide content increased,there was an increase in the hardness.However,the thickness of the modified layer can also be determined by the change of hardness,and according to this maximum thickness depending on the production parameters was observed in the S 4specimen.S.Buytoz et al./Applied Surface Science 252(2005)1313–13231320Fig.10.Relation between (a)hardness–weight loss,(b)surface roughness–hardness and (c)surface roughness–weight loss of coated specimens.3.3.Wear resultsThe relation of the distance to the mass loss of the S 2–S 4samples which surface modification was applied to with WC powders in different amounts after abrasive wear in different load rates was given in Fig.7.In Fig.7,when the wear distance considered,the weight loss amounts of the specimens S 2–S 4modified with WC powders worn under the loads of 10,20and 30N increased,showing an approximate parabolic inclination.It was observed that S 2speci-men had less mass loss when compared to the other specimens.When the weight loss amounts under 10N load were compared (Fig.7(a)),while S 2specimen lost 0.02g for a distance of 15m,the samples S 2and S 3worn in similar amounts (0.028g)to each other.The more the distance was,the more the differences were between the weight loss amounts.In the distance of 60m,in each of the three specimens,there was a decrease observed in the weight loss amounts.The specimen S 4lost 0.19g,the specimen S 3did 0.082g while the specimen S 2lost 0.065g mass.For the 20and 30N loads,the same situation was also observed in Fig.7(b)and (c).The total weight loss amounts for 225m of the modified samples S 2–S 4in accordance with the loads applied were given in Fig.8.When the specimens were put into order considering the total weight losses under the load of 10N,although the specimen S 3had lost more weight,the minimum weight loss was measured in the specimen S 2.While the total weight loss under 30N was recorded in the specimen S 4,the weight loss amounts in the specimen S 3was less than that of S 4and stayed more than that of the sample S 2.In Fig.9,wear rates values were calculated dividing the weight loss emerged as a result of the load applied by distance.In Fig.9,the specimen S 3with respect to wear rates values in the distances of 15and 30m,drew a maximum peak with the values of 1.8Â10À5to 2.8Â10À5mm 3/Nm.Besides,it fell to the value of 1.7Â10À5mm 3/Nm in the distance of 60m.For all the samples there was a peak formation observed at the distance of 30m.However,when the distance was 120m,the wear rates (0.9Â10À5to 1.1Â10À5mm 3/Nm)were calculated very similar to each other.There was a fall with hyperbolic inclination observed in the wear rates under the load of 30N depending on the distance.In the distance ofS.Buytoz et al./Applied Surface Science 252(2005)1313–13231321Fig.11.SEM micrographs of worn surfaces of samples S 2–S 4under 20N load,respectively (250Â).120m the specimen S2showed5.5Â10À5mm3/Nm and the specimen S3did0.7Â10À5mm3/Nm while the specimen S4exhibited0.8Â10À5mm3/Nm wear rates.In Fig.10,the relation of the hardness values to weight loss(Fig.10(a)),surface roughness(Fig.10(b)) and the relation of the surface roughness to weight loss (Fig.10(c))were shown.In Fig.10(a),while specimen S4having956HV hardness lost1.38g mass in total, there was also a decrease observed in the mass loss when the hardness increased(the specimen S2). However,it was discovered that the specimen S3 had1150HV hardness and it had the maximum weight loss,which was1.45g.It was seen that having 1200HV hardness value,the specimen S2was the specimen,which had the minimum weight loss, resulting1g total.There was decrease observed in the surface roughness when the hardness values increased (Fig.10(b)).While the specimen S4having950HV hardness showed0.65m m surface roughness,the surface roughness degree of the specimen S2with 1200HV hardness was0.23m m.However,the pin-on-disc tracks of the worn surfaces are presented in Fig.11.The surfaces of sample S2narrower and regular tracks than the samples S3and S4.4.ConclusionsThe following results were obtained in the surface modification with WC by using the TIG method.In the WC modificated specimen S2,which was improved with high powder amount and high-energy input,produced a systematic solidification.When the production speed increased(which leads to a decrease in energy input),WC grains showing solidification without dissolution were observed in the microstruc-ture.Due to lower heat input value,modificated layer did not exhibit a systematic solidification and as a result, there were WC grains remained dissolved in the structure.This caused the hardness values to be high in the specimens S3and S4.Therefore,it was concluded that heat input value is the parameter directly affecting the solidification speed.In the WC modification,in the specimen S2,which was improved with high powder feed rate,low process speed and high heat input and which produced a systematic solidification,the minimum mass loss was recorded.The effects of the thermal process to be applied in this kind of coating and surface hardening methods (whether normalization or stress relaxing processes) on the microstructure and as a result on the mechanical properties are worth studying.Furthermore,the effects of the grain size of WC powders used for modification on the solidification characteristic may also be examined.AcknowledgementsThe experimental part of this study was supported by Scientific Research Unit,Firat University(FUBAP-Project No.733).References[1]J.S.Selvan,K.Subramanıan,A.K.Nath,J.Mater.Process.Technol.91(1999)29–36.[2]Y.Yang,U.H.C.Man,Surf.Coat.Technol.132(2000)130–136.[3]S.Wulin,Z.Beidi,X.Changsheng,H.Wie,C.Kun,Surf.Coat.Technol.115(1999)270–272.[4]J.Tusek,M.Suban,J.Mater.Process.Technol.(2003)207–213.[5]J.P.Modenesi,R.E.Apolina´rio,M.I.Pereira,J.Mater.Process.Technol.99(2000)260–265.[6]A.Alsaran,Mater.Charact.49(2)(2002)171–177.[7]A.Molınarı,B.Tesı,T.Baccı,T.Marcu,Surf.Coat.Technol.140(2001)251–255.[8]H.Berns,A.Fischer,Mater.Charact.39(1997)499–527.[9]L.Prchlik,S.Sampath,J.Gutleber,G.Bancke,A.W.Ruff,Wear249(2001)1103–1115.[10]M.H.Korkut,O.Yılmaz,S.Buytoz,Surf.Coat.Technol.157(2002)5–13.[11]J.Khedkar,A.S.Khanna,K.M.Gupt,Wear205(1997)220–227.[12]S.Mridha,H.S.L.S.Ong Poh,P.Cheang,J.Mater.Process.Technol.113(2001)516–520.[13]D.Lou,J.Hellman,D.Luhulima,J.Liimatainen,V.K.Lin-droos,Mater.Sci.Eng.A340(2003)155–162.[14]J.Mateos,J.M.Cuetos,E.Fernandez,R.Vijande,Wear239(2000)274–281.[15]K.H.Lo,F.T.Cheng,H.C.Man,2003.Mater.Sci.Eng.A357,1–225168–180.[16]T.Sahraoui,N.E.Fenineche,G.Montavon,C.Coddet,Mater.Des.24(5)(2003)309–313.[17]Q.Yang,T.Senda,A.Ohmori,Wear254(1–2)(2003)23–34.S.Buytoz et al./Applied Surface Science252(2005)1313–1323 1322[18]D.A.Stewart,P.H.Shipway,D.G.McCartney,Wear225–229(2)(1999)789–798.[19]D.A.Stewart,P.H.Shipway,D.G.McCartney,Surf.Coat.Technol.105(1–2)(1998)13–24.[20]K.Nagarathnam,K.Komvopoulos,Metall.Mater.Trans.A26(1995)2131–2139.[21]G.V.Raynor,V.G.Rivlin,Inst.Met.(1988).[22]R.Zhou,D.Jiang,D.Lu,Wear255(2003)134–138.[23]H.Oh,S.Lee,J.Y.Jung,S.Ahn,Metall.Mater.Trans.32A(3)(2001)515–523.[24]Y.C.Zhua,C.X.Ding,K.Yukimurac,D.Xiao,P.R.Struttd,Ceram.Int.27(2001)669–674.S.Buytoz et al./Applied Surface Science252(2005)1313–13231323。