偏压对镀膜的影响.doc

Vacuum90(2013)145e150Contents lists available atSciVerse ScienceDirectVacuumjournal homepage: /locate/vacuumEffect of substrate bias voltage on the properties of CrCN and CrNcoatings deposited bycathodic arc evaporationB.Warcholinski*,A.GilewiczKoszalin University of Technology,Institute ofMechatronics, Nanotechnology and Vacuum Technique, Sniadeckich 2,75-453 Koszalin, Polanda r t i c l e i n f oa b s t r a c tArticle history:Thesubstratebiasvoltageisoneofthesignicantparametersoftechnologycontrollingthemechanicalproperties of PVD coatings.Chromium nitride and chromium carbonitride coatingswere depositedonHS6-5-2 steel substrates by cathodic arc evaporation technique. The relationships between one ofdepositionparameter,negativesubstratebiasvoltagesVs(10Ve300V),andthecoatingscharacterizationsuchasmicrohardness,stress,adhesionofthecoatings,elasticstraintofailurerelatedtoH/Eratiowereinvestigated.OnSEMphotosonecanobserveadensepolycrystallinestructurewithcolumnargrainsforthecoatingsdepositedatlowbiasvoltageandhomogenous,ne-grainedmorphologyobtainedathigherbiasvoltage.Residualstressesweremeasuredbysubstratecurvaturetechnique.TheCrNcoatingsshowacompressiveresidualstressthatincreasesfrom1.0(Vs10V)to2.1GPa(Vs70V).Athigherbiasvoltagesadecreaseofthecompressiveresidualstressesisapparent.SimilarrelationbutforhigherstressvaluesisobservedforCrCNcoatings.Nanoindentationshowedamaximumhardnessof25GPaforCrNand26GPaforCrCNcoatingsdepositedat150Vofbiasvoltage.ThecriticalloadsofCrNcoatingsinascratchtestdecreasedmonotonicallyfrom95Nto78Nwithincreasingnegativesubstratebiasvoltage.ThecriticalloadofCrCNcoatingswasnearlyconstant,about78N.Received 15October 2011Received in revised form24April 2012Accepted 27April 2012Keywords:CrCN coatingsCrN coatingsCathodic arc evaporationBiasHardnessAdhesion2012ElsevierLtd.Allrightsreserved.1. Introductionpointed outthatfriction coefcientandwearrateofCrCNcoatingsdepend on their carbon content and for nitrogen content higherTransition metal nitrides hard coatings deposited byPVD tech-nique have found widespread industrial applications because oftheir high hardness and good wear resistance. Chromium nitridethan 25 at.% the coatings show no cracks and very little delami-nation along the indent boundary after the Rockwell C tests. Choiet al. 6 showed that with increase of carbon content in CrCNcoating theinterplanar distance (d)alsoincrease. Forabout 20at.%of carbon the coatings reveal high hardness and residual stress.The friction coefcient is lower compared with CrN coating.Lugscheider etal.7investigated differentcarboncarriergasesandconcluded that depending on the gas carrier and the processparameters, reactive gaspressure andcomposition different phasestructures with different mechanical properties were formed. Inour previous works we observed decrease of microhardness andcritical loadinscratch testwithincrease ofcarbon content inCrCNcoatings 8. Further experiments, especially for CrCN/CrN multi-layercoatings 9enable ustooptimize thecarbon content inCrCNcoatings forminimum coefcientoffriction andwearrate.Pengfeiand Bailing 10 indicated in their investigations for high (above27at.%) carbon content inCrCN coatings ondecrease ofcoefcientoffriction andwearratewithraiseofcarbon content. Ontheotherhand, Tong et al. 11 in tests of CrCN coatings with low (up to5at.%)carboncontentindicated onthelowestcoefcientoffrictionforthe coating with carbon amount of1.5at.%.coatingexhibits higher corrosive and oxidation resistancecompared toother nitride coatings. Itshows good wear resistanceunder difcult tribological conditions such as high temperature,speed and applied pressure 1.Theaddition ofcarbon intochromium nitride hasanimportanteffect on the microstructure and properties of the coatings. Almeret al. 2 found that large amount of carbon in CrCN structurereduces the stress, hardness and critical load of the coatings. Forlow carbon content increase of critical load is observed. Cekadaetal.3,4indicated thatadhesion ofCrCNcoatings tothesubstrateislower than CrNandCrCcoatings andtheoxidation resistance ofCrCN coatings depends onthe deposition process parameters. Thesputter-depositedcoatings are better than the evaporation-deposited ones with respect to adhesion and worse with respecttoqualityofsurfaceofthecoatings(higherroughness). Wuetal.5* Corresponding author.Tel.:48 943486634; fax: 48 943486652.E-mail address: bogdan.warcholinskitu.koszalin.pl (B. Warcholinski).0042-207X/$ esee front matter 2012Elsevier Ltd. All rights reserved.doi:10.1016/j.vacuum.2012.04.039146B.Warcholinski,A.Gilewicz/Vacuum90(2013)145e150Thebiasvoltage isoneoftheimportant technology parametersAfterthisoperation, signicantlyreducing thesurfaceroughness ofinuencing mechanical parameters ofPVD coatings. Wanetal.1presented the evolution of CrN phase structure with the biasvoltage raise. The consequence of it was deposition rate decreaseand surface smoothing. The coatings showed maximum of thehardness and residual stress forabout 100Vofbias voltage. ThiseffectwasconrmedbyOdnetal.12whoadditionally pointedatreduction the critical load with bias increase and the minimum ofwear rate for negative bias voltage of 100 V.The similar phenom-enon for TiN was observed by Bull 13. The maximum hardnessand compressive residual stress and minimum wear rate occurredfor about 70 V. Other transition metal nitride, zirconium nitrideZrN showed maximum hardness at60Vofbias voltage 14.This paper contains the morphology of the cross section andmechanical resultsofCrCNandCrNcoatings deposited onHS6-5-2steel substrates with cathodic arc evaporation at various negativesubstrate bias voltages (10Ve300 V).the coatings, the statistic ofmeasurements washighly improved.Adhesion of the coating to the substrate was assessed usingscratch tester (Revetest CSEM). Inthis method adiamond stylus eaRockwellCdiamond conewitha200mmradiushemispherical tipwasdrivenacrossthecoating attheconstant speedof10mm/min,atthedistance of10mminlengthandaloadincreasing witharateof 100 N/min to determine the coating-to-substrate adhesion. Atleastvescratch tests,eachseparated byadistance ofabout 1mmwereperformed oneach sample. Thecritical load wasdeterminedfrom the loadescratch distance plotasthepoint atwhich the rstsudden dropintheloadoccurred duetocoating delamination. Thescratches were subsequently observed under a scanning electronmicroscope (SEM) toconrm the physical spallation/delaminationofthecoating. Inthisstudy,thecritical loadLc1wasdenedastheloadatwhich thesubstrate wasobserved forthersttime andLc2as a load at which the coating delamination is observed. Coatingstresses were determined using Stoneys formula after measuringthe curvature radius by surface roughness analyzer (HommelWerke Prolometer T8000) 16:2. Experimental details2.1. Depositionwhere:EeYoungmodulus ofthesubstrate, vePoisson coefcientofthesubstrate, tfethickness ofalayer,Reradiusofthesubstratecurvature together with the coating, Rs e radius of the substratecurvature. The stress was measured using a silicon substrate of0.3mm thick, 30mm long and 4mmwidth.E1v6tfts21R1s Rs(1)CrNandCrCNcoatings weredeposited inthemulti-source PVDsystem using cathodic arcevaporation onHS6-5-2 steel substratesand Si (100) wafers. Hardened and annealed steel substrates of32mmindiameter and3mmthickweremechanically polished toamean roughness Raofabout 0.02 mm. The samples were chemi-callydegreased andultrasonically cleaned inahotalkaline bathfor10 min and dried in warm air. The substrates were located about18cmawayfrom thechromium cathode. After thepressure inthechamber was lowered to 1 mPa it was lled with argon to about3. Results and discussion0.5 Pa. For further cleaning to remove the surface oxide layer,thesubstrates were sputter etched using chromium Cr and argon3.1. SEM morphology(Ar)ion bombardment with abias voltage of600 Vfor 20min.The CrN and CrCN coatings were deposited from a pure Cr target(99.99%) inpure nitrogen (99.99% N)atmosphere atanarccurrentof80Afor100min.TheN2gaspartialpressurewaskeptconstantat1.8Pa.IncaseofCrCNcoatings acetylene wasaddedandaowrateof 10sccm was controlled by an MKS 100 mass ow controller. Achromium interlayer of about 0.1 mm thick was rstly depositedonto the substrates to increase the adhesion of the coating to thesubstrate. Deposition was carried out with intentional substrateheating up to about 573 K. During deposition negative substratebiasvoltages (10,70,100,150,200, 250and300V)wereapplied tothe substrates.Thethickness ofCrNandCrCNcoatings aresimilar andamountto2.20.2mm.Theyarecharacterized byroughness Radependentononeofthetechnological parameters ethesubstrate biasvoltage17.Theroughness ofCrCNcoatingsishigher,about15toeven40%compared to CrN coatings. For a small negative substrate biasvoltages the roughness of both coatings is the highest, decreaseswiththegrowthofthebiasvoltagetoabout150VO200V,andnextslightlyrises.Thesameeffectofreducing theroughness ofthecoatings with increasing substrate bias voltage was also observedbyother authors 12,17,18.Roughness of the coatings is correlated with the surface mac-roparticle density 17. At the lowest bias voltage, 10 V, thenumberofmacroparticles isthelargestanddecreases permanentlywith increasing the bias voltage. For CrN coating the amount ofmacroparticles isproportionally lower 17.Cross-section SEM images of CrN and CrCN coatings (Fig. 1)show thecolumnar growth ofCrN coating deposited atthelowestbiasvoltage (10V).Thewidth ofthecolumnar grainsisrelativelybig. An increase of negative bias voltage causes a decrease ofcolumn width and formation of a more homogenous ne-grainedmicrostructure. In case of CrCN coatings this phenomenon isobserved from 70V.This effect wasearlier observed byLinetal.19.Theuseofahigher negative substrate biasvoltage resulted inthe removal of loosely bonded coating droplets and the improve-ment of coating surface. Large roughness of the surface is con-nected with columnar structure ofagrown coatings especially forlow negative substrate bias voltages. Additionally, the highsubstratebiasvoltagealsoinducesenergetic particlebombardmentduring the lmdeposition increasing the surface roughness.Thechangeincoatingsurfaceroughness andmorphology canbeexplained by the ion energy change. The increase of negative2.2. CharacterizationAnalysis of the microstructure was completed using scanningelectron microscopy SEM (JEOL JSM 5500LV). To evaluate thethickness of the coatings the ball-cratering method (Calotest) wasused.The hardness of the CrN and CrCN coatings were determinedusingFISCHERSCOPE HM2000 device. Evaluation ofmicrohardness(by nanoindentation) of coatings deposited using cathodic arcevaporation is extremely difcult to perform due to the largenumber ofmacroparticles ontheirsurface andhighroughness. Therequirement for hardness measurement correctness is on the onehanddepthofindentation nogreaterthan1/10oftotalthickness ofthecoating(inordertoeliminate theeffectofthesubstrate), ontheotherhand adepth ofindentation should benotlessthan 20Ra(arithmetic mean roughness). Therefore, to estimate the micro-hardnessofthecoatingsthemethoddescribed byRomeroetal.15was used. For this reason, the coatings were polished with nediamond powder (1mm),inorder toremove these surface defects.B.Warcholinski,A.Gilewicz/Vacuum90(2013)145e150147Fig.1. Cross-sectionSEMimagesofCrNandCrCNcoatingsdepositedatdifferentbiasvoltages.substrate biasvoltageresultsinanincreased ionsmobility tomoveatoms effectively. The highly mobile adatoms can move or diffuseintotheinter-grainvoidsunderthehighenergy ionbombardment,break down the large columnar grain growth and create morenucleation sites.Therefore, adenserstructure andgrainrenementare attained 19,20. As the next effect of increase of negativesubstratebiasvoltageisoccurrence ofresputtering whichcancauseachange inchemical composition andmorphology ofthecoatings.This is probably the reason for the formation of hexagonal Cr2Nphase characterized byanergrainsize,andthusdifferent surfacemorphology compared tothe cubic CrN phase.CrCN. The range ofhardness values issimilar tothat presented byWanetal.1.The increase in substrate bias voltage does not cause 12 orcausesonlyslightlyanincreaseofthemetal-to-nitrogen ratiointhe3.2. HardnessThe hardness of CrN and CrCN coatings is similar within therange of measurement uncertainty. The lowest hardness (about18e20 GPa) have both coatings deposited at the substrate biasvoltageof10V.ExcepttheCrNandCrCNsamples obtained atthesubstrate biasvoltageof10Vtheothercoatings arecharacterizedby similar hardness, almost unchanged with increasing the nega-tive bias voltage. But one can observe that with substrate biasvoltage increases to about 150 V, the hardness increases toapproximately 25e26.5 GPa and afterward decreases eFig. 2.Thehardness ofcoatings obtained inthesubstratebiasvoltageof70Vto300Vareintherange22.2e25.0GPaforCrNand21.3e26.4forFig.2. ThehardnessofCrNandCrCNcoatingsdepositedbycathodicarcevaporationontheHS6-5-2steelsubstrateatdifferentbiasvoltages.148B.Warcholinski,A.Gilewicz/Vacuum90(2013)145e150chemical composition of coatings 1,15,21. High substrate biasvoltage favors the forming of Cr2N hexagonal phase during depo-sition of CrN cubic phase 1,12,17,22. The X-ray tests conrm thepresence of the cubic phase of CrN for all bias voltages and Cr2Nhexagonal phase coexisting with the CrN cubic phase at a biasvoltage of 300 V 17. This hexagonal phase is characterized byahigher hardness compared toCrNapparently but does notaffectthe hardness ofthe coating.The coatings deposited atlow substrate bias voltages are char-acterized bylargeamounts ofusuallysmallmacroparticles ontheirsurface17.Theseparticleswhicharegenerallymetallicchromiumdroplets, causethehardness ofthecoatings isrelativelylow.Asthesubstrate bias (energy of ions) increases the number of macro-particles decreases and the coating structure becomes compact.Other possible effects ofbombardment byhigh-speed particles areforcing atoms in the coating to leave from their lattice sites andmove intothe interstitial positions orrelaxation and recrystalliza-tion1.Firstoftheseeffectscausesthehardnessincrease,thelatterone thehardness decrease. These same effects areassociated withchanges instress.Fig.3. ElasticstraintofailurerelatedtoH/EratioevaluatedforCrNandCrCNcoatingsdepositedatdifferentsubstratebiasvoltages.Hardness H is one of the most important material propertiesaffecting theirresistance towear.Another,equallyimportant istheYoungsmodulus E.Elasticstraintofailure,relatedtotheratioofH/E 23 reveals important information about the durability of thematerial and well describes the ability ofcoatings toelastic defor-mation. Itplays an important role in assessing the strength of thecoatings on the local dynamic loads. A little load of coatings witha low H/E causes their plastic deformation and rapid increase ofstress in the coating, frequently exceeding the local strength. Thisresults in the formation and propagation of microcracks, whichcombine and lead to local detachment of the coating fromsubstrate. Coatings with a higher H/E index need more energy(load)toinitiateplasticdeformation. Itcausesatoughness increase.Increasing the hardness of the material tends to increase of itsbrittleness, andtheenergy needed tofractures islowerthanintheplasticmaterial. ThehighervalueofH/Eindexisastrongindicationof the coating resistance toplastic deformation, but such coatingsdo not always show improvement of elastic deformation andfracture toughness. This restrains an application of the index asameasure ofwearresistance toacertain sizelimit, acharacteristicfor each material. The effect of Youngs modulus increase isadecrease inthe value ofH/E,which has abenecial effect ontheresistance ofcoatings tocracking.substrate system. Forthisreason theresults ofmaximum hardnessandstressdonotoccurforthesamesubstratebiasvoltage.Itmeansfor the same ions energies. In our earlier studies on TiC coatingsproperties dependent ontheofTi/Cratio,deposited bymagnetronsputtering on Si and steel substrates we also observed differentvalues and the position of maximum stress for TiCeSisystem andTiCesteel system 26. Position ofstress maximum for the studiedsystemsoccurredforcoatingswithdifferentTi/Cratio.Difference inthe evaluated values ofstresses corresponds tothe thermal stress.In alarge majority of papers the maximum of stress and hard-nessoccurs atthesamesubstrate biasvoltage. Butforexample, forCrN coatings, Lin et al. 19 presented the maximum of stress forbiasvoltage amounts toabout 100Vandmonotonically increaseof the hardness. Su et al. 21 observed for WCN0.75coatings0.25reverse behavior, almost linear increase of stress with increase ofsubstratevoltage. biasto200Vandmaximum hardness for120VofbiasReference CrN coatings present a similar character of changes,which is close to the data presented in Ref. 2. For single-phasecoatings macrostresses are caused by microstructural defectsproduced byastrokesofhighenergy ionsintothesubstrate duringThese indexes for the CrN and