HARDPVDCOATINGSANDTHEIRPERSPECTIVES IN FORMING …€¦ · VANADIS 4 steel against austenitic stainless steel the initial friction co-efﬁcient varied between 0.30 and 0.35. The

HARD PVD COATINGS AND THEIR PERSPECTIVESIN FORMING TOOL APPLICATIONS

B. Podgornik and S. HogmarkThe Tribomaterials group

Ångstrom Laboratory

Uppsala University

Box 534

SE- 751 21 Uppsala

Sweden

O. SandbergUddeholm Tooling AB

SE-683 85 Hagfors

Sweden

Abstract The aim of the present work was to investigate the potential of using hardPVD coatings on forming tools. Tribological evaluation of TiN, TiB2, TaCand DLC coatings deposited on a cold work tool steel was carried out in a load-scanning test rig and compared to the behaviour of differentuncoated formingtool steels. The special test configuration, where austenitic stainless steel wasused as a counter-material, makes it possible to gradually increase the normalload during forward sliding strokes and to correspondinglydecrease the loadduring postreversed ones. In this investigation, the load range was 100 to1300 N (1 to 5.1 GPa).

Experimental results indicate that introduction of a proper hard coatingwill lead to an improved wear resistance and a longer lifetime of the formingtool. Furthermore, by using hard low-friction coatings excellent anti-stickingproperties can be obtained.

Keywords: forming tools, PVD coatings, adhesion, wear, friction

1053

1054 6TH INTERNATIONAL TOOLING CONFERENCE

INTRODUCTION

Hard and corrosion-resistant coatings are frequently usedto protect andenhance the lifetime of tools under high and constant wear loads [1]. Al-though introduced more than two decades ago, TiN still dominates amongthe hard coatings employed in the industry. However, requirements to with-stand aggressive environments and to improve oxidation andwear resistanceunder extreme conditions constantly lead to a development and introductionof new coatings [2].

In contrast to cutting tools, the majority of forming tools are still uncoated.This is due to the larger size and a complex shape of most forming tools,which makes it difficult to apply a coating and to obtain a goodadhesionbetween the coating and the substrate material [3]. Although hard ceramiccoatings are routinely deposited with excellent adhesion,there is alwaysthe risk of depositing a coating with poor adhesion [4]. Evenif the latteris most undesirable for cutting tools, it is not a disaster. However, if acoating fails on a forming tool, coating fragments can constitute a sourceof abrasive particles within the system, which can lead to impairment inproduct surface quality and destruction of a very expensivetool. There arealso other reasons why the typical hard coatings are not usedmore widely informing tool applications. One is the relatively high coefficient of frictiongenerated by most of the commercial ceramic coatings used incutting toolapplications [1], which lead to a high tendency to galling when slid againstsoft metals [5]. However, in the last couple of years, tremendous progresshas been seen in the field of coating deposition as well as in introducing newcarbon-based coatings with excellent frictional properties [6, 7, 8, 9].

The aim of the present work was to investigate the possibilities of usinghard PVD coatings on forming tools. Tribological evaluation of TiN, TiB2,TaC and DLC coatings deposited on a cold work tool steel was carried out ina load-scanning test rig and compared to the behaviour of different uncoatedforming tool steels, using soft stainless steel as a countermaterial.

EXPERIMENTAL

Four different PVD coatings with a thickness of about 2 µm were usedin this investigation; TiN, TiB2, TaC and DLC. The investigated coatingswere deposited on a hardened and tempered powder metallurgycold worktool steel, VANADIS 4 (Uddeholm Tooling AB designation, seeTable 2),

Hard PVD coatings and their perspectives in forming tool applications 1055

using commercial PVD processes. Process parameters and properties of thecoatings are listed in Table 1. The DLC coatings, which were WC doped

Table 1. Deposition parameters and resulting coating properties

Coating ProcessTemperature

[℃ ]Substratebias [V]

Hardness[GPa]

Young’smodulus[GPa]

Residualstress[GPa]

TiNReactivee-beam

320–420 −110 30 ± 2 500 ± 50 −3.8 ± 0.4

TiB2 Sputtering 300 50 54 ± 9 600 ± 85 −0.5 ± 0.2

TaC Sputtering 70 −50 15 ± 2 230 ± 20 NA

DLCReactive

sputtering230 NA 12 ± 1 130 ± 7 −0.3 ± 0.1

hydrogenated diamond like carbon coatings with a multilayer structure ofWC and C (DLC), were deposited at a substrate temperature of∼ 230℃. Forthe refractoryhard coatings of TiN, TiB2 andTaC, the deposition temperaturewas in the range between 70 and 420℃. Prior to the coating deposition athin (∼ 0.1 µm) Ti intermediate layer was deposited for the TiN, TiB2 andTaC coatings, and a Cr layer for the DLC coating, to improve adhesion ofthe coatings.

The adhesion of the coatings deposited on polished flat samples (Ra≃0.02 µm) was evaluated with a Scratch tester equipped with a 200 µmradiusRockwell-C diamond stylus. The loading rate used was 10 N/mmand themaximum load 100 N. The critical load at which first failure ofthe coat-ing occurred as cracking or spallation was determined by post-test opticalmicroscopy (OM).

Tribological properties of a coated VANADIS 4 steel were investigatedin the load-scanning test rig and compared to uncoated hardened or plasmanitrided VANADIS 4 steel as well as to four different formingtool steelsproduced at Uddeholm Tooling AB, see Table 2. Heat treatments and hard-ness values of the forming tool steels included in this investigation are givenin Table 3.

As counter material in the load scanning tests, a soft (350 HV) austeniticstainless steel (AISI 304) was used for friction tests and a hardened andtempered (850 HV) ball bearing steel (AISI 52100) for the wear resistanceassessment.


Table 2. Production process and nominal chemical composition of theinvestigated formingtool steels

Nominal Chemical CompositionSteel Process∗%C %Si %Mn %Cr %Mo %V %W

VANADIS 4 PM 1.5 1.0 0.4 8.0 1.5 4.0 —VANADIS 6 PM 2.1 1.0 0.4 6.8 1.5 5.4 —VANADIS 23 PM 1.3 0.5 0.3 4.2 5.0 3.1 6.4WEARTEC SF 2.8 0.8 0.7 7.0 2.3 8.9 —

∗PM - powder metallurgy, SF - spray forming

Table 3. Process, heat treatments and resulting hardness values of the investigated formingtool steels

Steel Treatment Treatment parametersHardness[HRC]

VANADIS 4 AH Hardening 1050℃/30min/air +2 × 525℃/2h 62

VANADIS 4 ANPlasmanitriding

500℃/9h/95%H2-5%N2 70

VANADIS 6 B Hardening 1050℃/30min/air +2 × 525℃/2h 62VANADIS 23 C Hardening 1050℃/30min/air +3×560℃/1h 62WEARTEC D Hardening 1020℃/30min/air +2 × 525℃/2h 62

In the load-scanning test rig, which involves two crossed, elongated cylin-drical test specimens of∅ = 10 mm (Ra≃ 0.2 µm) that are forced to slideagainst each other under a constant speed, the normal load isallowed togradually increase during the forward stroke and to correspondingly de-crease during reverse stroke [10, 11]. Thus, each point along the contactpath of both specimens will experience a unique load and display a uniquetribological history after test completion.

For the purpose of this investigation the range of normal load was of theorder of 100 – 1300 N. However, depending on the tribologicalpropertyinvestigated, different modes of testing were used. For thepurpose of anti-sticking tests, where the ability of investigated materials and coatings toprevent transfer of a soft austenitic stainless steel to thetool surface wasevaluated, the test equipment was set to a single, forward stroke mode. Drysliding conditions with a sliding speed fixed to 0.01 m/s wereused.


Todetermine frictional behaviour of investigated materials against austeniticstainless steel under starved lubricated conditions, the load-scanning test rigwas set to multicycle mode. An approximately 10 µm thick film of purepoly-alpha-olefin oil (PAO,ν40 = 46.6 mm2/s) was applied on the austeniticstainless steel sample before each test. A fully formulatedforming oil (Cas-trol Iloform TDN 81, ν40 = 120 mm2/s) was used in one of the tests ofnitrided steel for comparison The sliding speed was set to 0.1 m/s and thehighest number of test cycles was 50.

The same test procedure, with the sliding speed of 0.1 m/s, multicyclemode and usage of lubricant was used to determine the wear resistance ofdifferent materials and coatings. However, a hardened ballbearing steel hadto be used as counter material to provoke wear of the investigated materialsand coatings. The maximum number of test cycles was 200.

During testing the coefficient of friction was monitored as afunction ofloadand time andafter the completion of the test, critical loads correspondingto the appearance of material transfer and wear of the investigated materialswere determined by post-test optical microscopy (OM) and optical surfaceprofilometry, respectively.

RESULTS AND DISCUSSION

Figure 1 shows critical loads for the investigated coatings, correspondingto the appearance of the first visible failure of the coating during scratchingand determined by OM. In the case of TiN, TiB2 and DLC coatings depositedon VANADIS 4 steel, the first failure of the coating as cracking and spallationon either side of the scratch, Fig. 2 (a), was detected in the load range 10to 25 N. The TiN coating displayed the best results, followedby the muchsofter DLC, and the very hard and brittle TiB2 coating, which started tofail at ∼ 10 N load, as shown in Fig. 1. However, the TaC coatings flakedinstantaneously at loads below 5 N, Figs. 1 and 2 (b), which indicates verypoor adhesion properties of the TaC coating.

Figure 3 reveals the anti-sticking properties as the friction coefficientis monitored versus load in the dry sliding test. In the case of hardenedVANADIS 4 steel against austenitic stainless steel the initial friction co-efficient varied between 0.30 and 0.35. The first sign of adhesion of workmaterial to the tool steel surface, as indicated by a sudden increase in frictionand confirmed by post-test microscopic observation was detected at about200 N load. Similar results with only marginal differences in frictional be-


Figure 1. Scratch test results of investigated coatings.

(a) TiN (b) TaC

Figure 2. Coating failure mechanisms observed in scratch testing of coatings depositedon VANADIS 4 steel.

haviour were observed for all forming tool steels investigated, as shown inFig. 3 (a). However, depending on the load, at which a layer ofstainless steelstarts to build-up on the tool surface, the investigated forming tool steels canbe classified into two groups, see Fig. 4 (a). For the first group with hard-ened VANADIS 4 and VANADIS 6 steel, transfer of work materialstartedat approximately 200 N load, while VANADIS 23 and WEARTEC steels


(a) Forming tool steels.

(b) Surface engineered VANADIS 4 steel.

Figure 3. Friction coefficient vs. normal load recorded during sliding against stainlesssteel.

displayed adhesion of the austenitic stainless steel in theload range 250 –300 N, see Fig. 4 (a).


(a) (b)

Figure 4. Beginning of transfer of stainless steel to (a) forming toolsteels and (b) surfaceengineered VANADIS 4 steel.

Figures 3 (b) and 4 (b) show coefficient of friction curves andcriticalloads of material transfer, respectively, for surface engineered VANADIS 4steel. A sudden increase in friction was found to correspondto a beginning ofmaterial transfer for the nitrided VANADIS 4 and VANADIS 4 supplied withTiN, TiB2 and TaC coatings. Plasma nitriding improved the anti-stickingproperties of VANADIS 4 (Lc ≃ 300 N), which then outperformed all otherforming tool steels investigated. However, plasma nitrided surfaces wereunable to reach the very good properties obtained by the TaC and DLCcoatings, as shown in Fig. 3 (b) and 4 (b).

The TaC and DLC coatings considerably reduced the initial friction coef-ficient against austenitic stainless steel (µ≃ 0.15, see Fig. 3b) and gave thelowest ability to material transfer. For the TaC coating, transfer of stainlesssteel started arround 700 N load, while virtually no transfer of work mate-rial could be detected for DLC coated VANADIS 4 steel up to a maximumload of 1300 N, as shown in Fig. 5 (a). On the other hand, the TiB2 coatedsteel showed by far the highest friction coefficient (0.5 – 0.8) and an almostinstantaneous transfer of stainless steel to the coated surface, Fig.5 (b). Ap-plication of a TiN coating reduced the initial friction coefficient to about0.25, which, however, did not have any influence on the process of materialtransfer in comparison to uncoated VANADIS 4 steel, see Figs. 4 (a) and4 (b).

Monitoring of the friction coefficient as a function of load and time makesit possible to prepare friction maps, which show transitionpoints in the


(a) DLC coating at 1300 N load. (b) TiB2 coating at 150 N load.

Figure 5. Typical appearance of the contact surfaces of sliding test specimens at thebeggining of stainless steel transfer (light contrast). The arrows indicate the direction ofsliding.

tribological behavior of investigated materials. Friction maps for plasmasurface treated VANADIS 4 steel loaded against austenitic stainless steelunder starved lubrication conditions are shown in Fig. 6.

An increase in friction was detected already during the second stroke at≃ 400 N load for the plasma nitrided steel and the test had to be stopped dueto extensive transfer of stainless steel to the tool steel surface after the thirdstroke, as indicated in Fig. 6 (a). These results indicate, that as the reciprocalsliding proceeds, the initial regime of boundary lubrication moves towardsa mixture of boundary lubrication and dry sliding. Similar results, with theinitial friction in the range of 0.15 and 0.20 and transfer ofwork materialstarting already during the second stroke, were observed for all forming toolsteels investigated. However, the use of a fully formulatedforming oil gavea very smooth sliding of the nitrided steel (µ≃ 0.1) and complete protectionagainst material transfer, see Fig. 6 (b).

Figures 6 (c) and 6 (d) show friction maps for TiN and TiB2 coated steelloaded against austenitic stainless steel, respectively.In the case of TiNcoated steel a rapid increase in friction corresponding to arapid transfer fromboundary lubricated to dry sliding started already during the first stroke atapproximately 1100 N load. The TiB2 coating showed the highest increaserate in friction under starved lubricated conditions (0.4 –0.6), and an im-mediate transfer of stainless steel, Fig. 6 (d). On the otherhand, TaC and


(a) plasma nitrided steel + PAO. (b) plasma nitrided steel + fully formulatedforming oil.

(c) TiN coated steel + PAO. (d) TiB2 coated steel + PAO.

(e) TaC coated steel + PAO. (f) DLC coated steel + PAO.

Figure 6. Friction maps for surface engineered VANADIS 4 steel, sliding against softaustenitic stainless steel.


DLC coated samples showed improved frictional properties under starvedlubrication conditions, as compared to uncoated steel, Figs 6 (e) and 6 (f).For the TaC coating, an increase in friction was also detected during thesecond stroke. However, it was more load dependant,see Fig.6 (e), with theadhesion of the work material limited to high loads. Due to the poor adhe-sion, the TaC coating may fail under high loads leading to exposure of thesubstrate material, and accelerated material transfer. Byfar, the best resultwas obtained for the DLC coated steel, which during the whole50 cycle testdisplayed a uniform frictional behavior with a friction coefficient of ∼ 0.1,see Fig. 6 (f).

The differences in wear resistance among the test materialswere not asdramatic in sliding wear under starved lubrication as they were in friction,see Fig. 7. It shows the wear of the investigated materials measured at a

Figure 7. Wear rate of investigated materials loaded against ball bearing steel under starvedlubrication conditions (POA, FN = 700 N, 200 cycles).

position along the contact path corresponding to 700 N load (≃ 4.2 GPamaximum Hertzian contact pressure). Similar results were observed for thewhole load range. The general observations are that plasma nitriding andcoating improve the surface wear resistance. However, it isnot at all straightforward to interpret the sliding wear test results since several mechanismsare operating simultaneously.

Generally, a high hardness in combination with low frictionshould givea low wear rate. The counter material (ball bearing steel) contains somesmall volume content of hard particles in the form of µm sizedCr and Fe


carbides (about 1200 – 1500 HV), which could wear some of the testedsurfaces abrasively. Thus, a hard coating would act beneficially. On the otherhand, wear fragments form the coatings and treated tool steel surfaces couldpossibly be embedded in the counter material and act as abrasives againstthe test materials. As to the friction, a high friction promotes adhesion of thecounter material to the wearing surface, which may prevent further wear.

As in the case of the friction against stainless steel, see Fig. 3 (a), allforming tool steels were rather difficult to separate when comparing wearin the sliding test against ball bearing steel, see Fig. 7. VANADIS 23 andWEARTEC did display a slightly better wear resistance than the others. Onthe other hand, plasma nitriding gave up to 15% higher wear resistance ofVANADIS 4 steel.

In the case of coated tool steel, the TaC and DLC coatings, giving thelowest friction, see Fig. 3 (b) and 6, were outperformed by the TiN and TiB2

coatings. This is likely to be explained by the protective action of adheredwork material, which appeared most frequently on the lattercoatings. Thereason why TaC is inferior to the other coatings, could be therelatively pooradhesion of the TaC coating to the substrate, compare Figs. 1and 2 (b). Withthe TiB2 coating, the wear tests had to be stopped after approximately 15cycles due to extensive material transfer and building up ofa thick layer ofcounter material on the coated surface.

In the case of forming tools, the ability of the surface to prevent adhesionof work material is often more important than its wear resistance. Therefore,hard wear resistant ceramic coatings of TiN and TiB2 with relatively highcoefficient of friction and high tendency to material transfer do not representthe best solution. In addition, a poor coating adhesion may lead to coatingspallation, causing a deterioration in forming tool performance instead ofan expected improvement. Since any possible change in forming tool steelcomposition and/or structure gives only minor improvementin tool perfor-mance, plasma nitriding represents the most reliable way ofimproving thetribological properties of forming tools. On the other hand, the DLC coat-ing was found to prevent any transfer of work material to the coated surfaceeven under starved lubrication by non-additivated PAO, compare. Fig. 6 (f).Therefore, among the tested coatings, DLC seem to be the bestsolutionfor improving the tribological properties of forming tools, provided that thecoating-to-substrate adhesion is sufficient.


CONCLUSIONS

All forming tool steels investigated give comparable friction and wearproperties when tested in a load-scanning test rig against soft austeniticstainless steel and ball bearing steel, respectively. However, VANADIS 23and WEARTEC show a slight advantage over the rest.

After plasma nitriding, the VANADIS 4 steel outperformed all other form-ing tool steels investigated with regard to anti-sticking properties as well aswear resistance. Therefore, plasma nitriding represents the most reliableway of improving the tribological properties and performance of formingtools.

Although the hard TiN and TiB2 coatings showed the best wear resis-tance, they posses a high tendency to pick up work material. On the otherhand, the softer DLC coating with its excellent anti-sticking properties andsufficiently good wear resistance shows a high potential foruse in formingtool applications. On the condition that adequate coating-to-substrate adhe-sion is obtained, DLC coated forming tools, lubricated withonly PAO oil,may compete with the combination of uncoated forming tool steel and fullyformulated forming oil.

ACKNOWLEDGMENTS

Uddeholm Toolong AB and The Swedish Research Council are greatlyacknowledged for financial support. The supply of test materials and DLCcoatings from Uddeholm Tooling and Balzers Sandvik CoatingAB, respec-tively, is much appreciated. Many thanks go also to Urban Wiklund andDaniel Nilsson for preparation of the TiN, TaC and TiB2 coatings and toVojteh Leskovsšek for preparation of plasma nitrided samples.

REFERENCES

[1] B. BHUSHAN, "Modern Tribology Handbook" (CRC Press, NY,2000).

[2] V. IMBENI, C. MARTINI, E. LANZONI, G. POLI and I.M. HUTCHINGS, Wear 251(2001) 997.

[3] S. HOGMARK, S. JACOBSON, M. LARSSON and U. WIKLUND, in "Modern Tri-bology Handbook" (CRC Press, NY, 2000) p. 931.

[4] N.M. RENEVIER, J. HAMPHIRE, V.C. FOX, J. WITTS, T. ALLEN and D.G. TEER,Surf.Coat.Technol.142-144 (2001) 67.


[5] K. HOLMBERG and A. MATTHEWS, "Coatings Tribology" (Elsevier, Amsterdam,1994).

[6] A. ERDEMIR, F.A. NICHOLS, X.Z. PAN, R. WEII and P. WILBUR,Diamond andRelated Materials 3 (1993) 119.

[7] P. KODALI, K.C. WALTER and M. NASTASI, Trib.Int.30 (1997) 591.

[8] O. WANSTRAND, N. AXEN and R. FELLA, Surf.Coat.Technol.94-95 (1997) 469.

[9] C. RINCON, G. ZAMBRANO, A. CARVAJAL, P. PRIETO, H. GALINDO, E. MAR-TINEZ, A. LOUSA and J. ESTEVE, Surf.Coat.Technol.148 (2001) 277.

[10] S. HOGMARK, S. JACOBSON and O. WANSTRAND, in Proceedings of the 21stIRG-OECD Meeting, Amsterdam, March 1999, edited by D.J. Scipper.

[11] S. HOGMARK, S. JACOBSON and O. WANSTRAND, in Proceedings of the 22ndIRG-OECD Meeting, Cambridge, September 2000, edited by D.J. Scipper.

Documents

HARDPVDCOATINGSANDTHEIRPERSPECTIVES IN FORMING …€¦ · VANADIS 4 steel against austenitic stainless steel the initial friction co-efﬁcient varied between 0.30 and 0.35. The