PECVD)

PECVD)

Surface & Coatings Technology 203 (2008) 848–854 Contents lists available at ScienceDirect Surface & Coatings Technology j o u r n a l h o m e p a g...

1MB Sizes 4 Downloads 33 Views

Surface & Coatings Technology 203 (2008) 848–854

Contents lists available at ScienceDirect

Surface & Coatings Technology j o u r n a l h o m e p a g e : w w w. e l s e v i e r. c o m / l o c a t e / s u r f c o a t

Influence of plasma-nitriding and surface roughness on the wear and corrosion resistance of thin films (PVD/PECVD) O. Durst, J. Ellermeier ⁎, C. Berger Institut für Werkstoffkunde, TU Darmstadt, Grafenstraße 2, 64283 Darmstadt, Germany

A R T I C L E

I N F O

Available online 23 May 2008 Keywords: PVD Wear Corrosion Plasma-nitriding Surface roughness

A B S T R A C T The application of PVD coatings on low-alloy steels for the protection against wear and corrosion, may allow the substitution of high-alloy steel materials, which are used for applications under mechanical-corrosive loads. This could prolong the component lifetime and save costs. A pretreatment method, which may improve the corrosion resistance of PVD coated low-alloy steel, is plasma-nitriding of the substrate. So far there were no investigations to quantify the influence of plasmanitriding on the corrosion behaviour of PVD coated low-alloy steel. Hence one objective of this study was to quantify this influence by means of polarisation tests with determination of corrosion potentials and corrosion current densities. Furthermore the influence of plasma-nitriding of the substrate on the wear properties of PVD coated low-alloy steel was investigated. Moreover the influence of surface roughness on the corrosion properties of PVD coated low-alloy steel was an aim of this study. The investigated PVD coatings were new graded zirconium carbide layers (ZrCg) and graded chromium carbide layers (CrCg). The tests revealed, that higher roughness leads to worse corrosion protection by the investigated PVD coatings. This can be explained by the number of defects in the layer, which rises with increasing roughness and which allows the penetration of corrosive media to the substrate and thus accelerates corrosion underneath the film. The tests in artificial sea water in accordance with DIN 50905 did not show corrosive attacks on the tested films themselves. Plasma-nitriding of the substrate resulted in significantly higher corrosion potentials and significantly lower corrosion rates for the PVD coated specimens. The positive influence of plasma-nitriding of the substrate on the wear behaviour has been less pronounced. Only when higher normal forces were applied, which allowed the counter body to have direct frictional contact with the substrate, the wear was reduced due to plasma-nitriding. In summary, plasma-nitriding of the substrate is beneficial for achieving good corrosion and wear properties for PVD coatings on low-alloy steel. However a special focus has to be put on a proper pretreatment of the surface, which has to be as smooth as possible. © 2008 Elsevier B.V. All rights reserved.

1. Introduction In modern technical applications, components are to a high degree subjected to a complex thermo-mechanical-corrosive load. Wear and corrosion may exert substantial influence onto their functionality. The efficiency of machines may be reduced and the operating time be shortened. In such cases often special steels or cast materials are used and material costs are increasing substantially. Corrosion and wearresistant PVD coatings which are applied on low-alloy steel substrates, e.g. heat-treated steels, may contribute to a reduction of costs and to the increase of cost effectiveness. So far, the focus of research into the field of these layers was put on the wear protection [1]. The advanced coating technique, however, makes also the application under ⁎ Corresponding author. E-mail address: [email protected] (J. Ellermeier). 0257-8972/$ – see front matter © 2008 Elsevier B.V. All rights reserved. doi:10.1016/j.surfcoat.2008.05.022

corrosive conditions advisable [2,3]. Particularly when low-alloy steels are applied, the machine processing and pretreatment methods contribute towards a layer with a low number of defects [4,5]. This paper demonstrates the influence of surface roughness and of different pretreatment methods, more precisely the comparison of tempered and plasma-nitrided substrate, on the corrosion and wear behaviour of coating-substrate composites. 1.1. State-of-the-art The faultlessness of the layers plays a substantial role in the corrosion resistance of PVD coated materials [6–8]. In corrosion protection via PVD coatings, basically, the penetration of corrosive media to the substrate must be prevented. PVD coatings show, in general, always a certain penetrability which is due to micro-porosity and defects in the layers (e.g. pinholes). These may have been caused by material impurities or by dust particles which are adhering to the

O. Durst et al. / Surface & Coatings Technology 203 (2008) 848–854

surface during deposition. But also the roughness which depends on the finishing may exert substantial influence on the faultlessness of PVD coatings. The effects which a different roughness of the substrate exerts on the defect density have been tested for titanium nitride films (TiN) [9], chromium nitride films (CrN) [10,11] and chromium carbide layers (CrC) [12] on low-alloy steel. In summary the investigations revealed that the number of visible coating defects after corrosion tests increased along with the roughness and moreover good adhesion of the layers after the starting of corrosion decelerates the corrosion progress. If layer defects have already caused corrosive media to penetrate the layer, also the substrate corrosion resistance has a substantial influence on the corrosion behaviour of the entire coating-substrate composite. Plasma-nitriding is a pretreatment method, where nitrogen is implanted into the near-surface peripheral zone of a material. Thus, hardness is increased and corrosion resistance may also be increased [13]. The influence of plasma-nitriding on the corrosion behaviour of different coating-substrate composites has already been tested for some aspects. Through the plasma-nitriding of a stainless steel, the hardness and wear resistance have been improved and, on the other hand, the corrosion resistance may also be improved if nitriding is applied at low temperatures. This is caused by the development of a nitrogen-supersaturated S phase in the surface peripheral zone [14]. The nitrogen which is contained within has, in the case of pitting corrosion, alkalescent effects: N + 4 H+ + 3 e−⇒ NH+4. The steel is thus passivated and the progress of corrosion is hindered. For heattreatable steel the positive effect of plasma-nitriding was more strongly pronounced than that of a titanium nitride film alone where layer defects resulted in galvanic corrosion [15]. Plasma-nitriding, however, also exerts influence on the wear behaviour. For different PVD coatings, e.g. titanium nitride layers, on plasma-nitrided low-alloy steels investigations brought about a decrease of wear due to plasma-nitriding and an improved adhesion of the layers on plasma-nitrided substrate. This was ascribed to the chemical reaction of the titanium in the compound layer with the nitrogen in the substrate which has been tested by means of Auger electron spectroscopy at the interface of film and substrate [16,17]. In summary can be concluded that plasma-nitriding is a known pretreatment method to increase hardness and corrosion resistance of steels and also to enhance the adhesion of PVD-layers on steel substrates. But there were only few investigations on corrosion and wear properties of the combination of a PVD coating and plasmanitrided low-alloy steel so far. Especially the influence of plasmanitriding on corrosion current density and corrosion rates would be an interesting field of research to quantify the positive effect of plasmanitriding on the corrosion properties of PVD coated low-alloy steel. Furthermore the influence of surface roughness on the corrosion properties of PVD coated steel was investigated in many ways and showed an decreasing corrosion protection through PVD-layers with increasing roughness. In this case new graded, carbon containing multilayer coatings, which were discussed in this study, may have advantages and provide a better corrosion protection in combination with higher surface roughness.

849

In addition the influence of plasma-nitriding on the wear behaviour of the coating-substrate composites was determined. On the other hand an investigation was conducted to study the corrosion protection capability of the new coatings and their dependence on surface roughness. 2. Experiments 2.1. Materials 2.1.1. Substrate The substrate, heat-treatable low-alloy steel with a tempered structure, consists of proportions of martensite, ferrite and pearlite, Fig. 1b. The unetched, metallographical sections show a large number of impurities, as, e.g. linear manganese sulphide, Fig. 1a. During a coating process, such manganese sulphides at the surface may act as the starting points for layer defects. 2.1.2. Pretreatment The low-alloy steel is, via heat-treatment, settable to a certain hardness, strength and ductility. In the case under consideration, a part of the specimens (diameter 24 mm, height 8 mm) has been hardened and tempered at 300 °C, which resulted in a hardness of approx. 700 HV0.3 (mark.: no pretreatment). The other part of the specimens has been tempered at a temperature of 500 °C. To enhance the corrosion resistance and adhesion of the PVD coatings these specimen were plasma-nitrided (mark.: plasma-nitrided) without a compound layer in an industrial plasma process (temperature b480 °C, time = 10 h, voltage = 450 V, gas mixture: N2 b10%, H2 N 90%). These specimens have a hardness of approx. 800 HV0.3. The specimens were, moreover, prepared with three different categories of roughness; some were polished, Ra = 0.01 μm, some were industrially grinded and lapped, Ra = 0.3 μm, and some were lathed, Ra = 1.2 μm. Lapping is a mechanical surface finishing process, at which abrasive particles, e.g. silicon carbide, are rolling and grinding between the specimen's and a counterbody's surface. Hence asperities are cut off, but cavities remain. 2.1.3. Layer systems Five different layer systems provided by the “Institut für Oberflächentechnik der RWTH Aachen” were investigated. These layers are partially amorphous PVD coatings which have been delivered with and without reactive gas atmosphere by means of Magnetron Sputtering Ion Plating (MSIP). This method is known for its particularly low number of defects in the layers and promises, together with the amorphous structure, a good corrosion protection

1.2. Objective of this study The objective of this study is to investigate the capability of new PVD coatings to protect against corrosion and wear. These PVD coatings are graded zirconium carbide layers (ZrCg) and graded chromium carbide layers (CrCg), which were deposited on heattreatable low-alloy steel. On one hand the influence of a plasmanitriding pretreatment of the substrate on the corrosion behaviour of the coating-substrate composites was investigated by means of polarisation tests with determination of corrosion current densities.

Fig. 1. a) Low-alloy steel with manganese sulfide and impurities, b) low-alloy steel etched: martensitic–ferritic–perlitic texture.

850

O. Durst et al. / Surface & Coatings Technology 203 (2008) 848–854

effect. Connection to the substrate is made by a pure zirconium or chromium layer. The carbon content of the layer is increased along with the increasing layer thickness (grading) which results in the existence of an amorphous carbon layer at the surface. Through the grading, a smooth transition to the hard phase is achieved and thus good adhesion is supported. Furthermore the gradation results in a distribution of hardness along the coating thickness. The maximum hardness is obtained in the centre of the coating. The soft carbon-rich top layer acts as a solid lubricant and enables a running-in of a tribological system [18,19]. Among the five tested coatings were three hydrogenous graded zirconium carbide layers, Fig. 2; the first showed a simple grading (row 1 Table 1), the second had a multilayer structure with two chromium interlayers (row 2 Table 1), the third had a multilayer structure and zirconium interlayers (row 3 Table 1). Moreover, two PVD coatings with hydrogen-free carbon were tested; these were, on the one hand, a graded zirconium carbide layer with a multilayer structure consisting of three graded layers with amorphous carbon interlayers and an amorphous carbon top layer (row 4 Table 1), and, on the other hand, a graded chromium carbide layer which showed nearly the same multilayer structure (row 5 Table 1). The stated layer thickness has been determined with metallographical sections, Table 1. 2.2. Test methods 2.2.1. Electrochemical characterization The corrosion behaviour of coated specimens is determined via the measurement of the corrosion potential and the recording of polarisation curves. These methods allow the evaluation of the faultlessness of the layers and the comparison with uncoated materials. For the tests under consideration, a Gamry potentiostat PCI4/750 with a three-electrode cell has been used for a first measurement of the corrosion potential (15 min), relating to the standard hydrogen electrode [mVH]. Subsequently, a polarisation curve has been recorded. The curve started at a potential of approx. 100 mV below the corrosion potential and ended at +1200 [mVH]. The potential was increased in 2 mV steps with a change rate of 1000 mV/h. The test medium has been artificial sea water in accordance with DIN50905 which had a temperature of 30 °C. Sea water is known to be an aggressive and frequently used corrosion medium, which is of high importance, e.g. in off-shore applications. Due to its high chloride content it has a corrosive effect also on high-alloy steels, the possible substitution of which by coated standard materials is one intention of this investigation. 2.2.2. Characterization of the friction and wear behaviour Reciprocating sliding tests offer the possibility to investigate a high number of influential factors on the friction and wear behaviour with a good reproducibility (scattering b10%) and with little requirement of

Fig. 2. Schematic diagram of the principle coating composition with t = toplayer, g = graded layer, i = interlayer, a = adhesive layer and relative thickness of these layers according to the schematic diagram.

Table 1 Survey of tested PVD coatings with layer composition according to Fig. 2 Row no.

PVD-layer

1 2 3 4 5

ZrCg ZrCg–Ml–Cr ZrCg–Ml–Zr ZrCg–aC–Ml CrCg–aC–Ml

Layers (Fig. 3) a

g

i

t

Zr Zr Zr Zr Cr

1 × ZrCg 3 × ZrCg 3 × ZrCg 3 × ZrCg 1 × CrCN 2 × CrCg

– 2 × Cr 2 × Zr 2 × aC 2 × aC

– – – aC aC

Thickness [μm] 2.4 3.4 7.4 5.1 4.3

space (up to 24 tests per specimen with a diameter of 24 mm). The tests can be carried out with and without intermediates. At that, an oscillating ball (diameter 10 mm) which is subjected to overhand load is sliding over a planar specimen. For these experiments an Optimol Instruments SRV machine was used. The stroke of motion was 1.5 mm and the frequency was 20 Hz. Normal forces between 2 N and 50 N, resulting in a contact pressure of about 700 to 2000 MPa, were applied for loading the ball. The specimen temperature was 30 °C, the relative air humidity was 50%. The material used for the ball was ceramic silicon nitride, because in this way transfer layers on the ball and the PVD coating after the test could be exactly assigned to the substrate, coating or ball. During the test, the medium friction coefficient via an oscillation cycle has been constantly recorded and evaluated subsequently. By means of mechanically sampled surface profiles the volumetric wear of the specimens has been calculated after the tests. Moreover, the wear of the ball has been determined by the measurement of the flattening which was caused by friction. Both were subsumed to the volumetric total wear of the system as a specific value (total wear of specimen and ball). 3. Results 3.1. Corrosion tests 3.1.1. Influence of surface roughness on the corrosion behaviour In order to investigate the influence of roughness on the corrosion behaviour, polarisation tests have been carried out using different coatings on polished (Ra = 0.01 μm) and on lapped (Ra = 0.3 μm) tempered low-alloy steel. It applies to all tests that for the coatings on polished substrate in the entire potential region lower current densities were measured than this was the case for the coatings on the lapped substrate, Fig. 3. The light-microscopical evaluation of the appropriate specimen surface shows, in the case of a polished specimen, only a single void in the highlighted corrosion area, Fig. 4a. For the lapped substrate, single corrosion pits are no longer distinguishable. In fact a large area corrosion attack took place, Fig. 4b. A good example for the effect of roughness peaks is displayed on a scanning electron micrograph of a rough, lathed substrate surface (Ra = 1.2 μm) coated with a hydrogen-free graded chromium carbide layer, Fig. 5. It is obvious, that the corrosion pits, which developed during a polarisation test, are following the peaks of the machining scratches. The test results confirm the results from technical literature for the investigated coating-substrate composites where a higher surface roughness brings about a higher number of layer defects and also results in a worse corrosion protection by the layers [9,10,11]. At that, the mechanism is as follows: Roughness peaks and discontinuities of a rough substrate surface are inducing layer defects. These layer defects entail, during a corrosion exposure, the corrosion of the substrate under the layer which, again, entails blistering of the layer and thus increased corrosion. The layers which have been investigated did not show any corrosion of the layers themselves. 3.1.2. Influence of plasma-nitriding on the corrosion behaviour The measurement of the corrosion potential of the layers has been carried out for 15 min each on untreated and on plasma-nitrided

O. Durst et al. / Surface & Coatings Technology 203 (2008) 848–854

851

Fig. 3. Anodic polarisation curves for ZrCg coated low-alloy steel, a) polished with Ra = 0.01 μm, b) lapped with Ra = 0.3 μm. Fig. 5. SEM image of a lathed substrate surface coated with ZrCg after polarisation test.

substrate, Fig. 6. The corrosion potentials are, for both substrate conditions, clearly different, Fig. 6. The uncoated substrate is, in the plasma-nitrided condition, by approx. 480 mV higher than without plasma-nitriding (untreated). It applies to all layers on plasmanitrided substrates that the corrosion potential is, by 200 mV up to 300 mV higher than on untreated substrate. This behaviour is a first pointer to a lower corrosion current in the corrosion potential and thus also to a lower corrosion rate as a consequence of plasmanitriding. In this regard the highest corrosion potential was measured for the graded chromium carbide layer on plasma-nitrided substrate. It is, however, not possible to infer directly from a higher corrosion potential to a lower corrosion current and a low corrosion rate because this is not only depending on the anodic polarisation curve of the substrate but also on the polarisation curve of the cathodic partial reaction which is taking place at the layer surface. For a closer investigation, a hydrogen-free, graded multilayer zirconium carbide layer has been deposited on glass and, for the test, connected electroconductively. A cathodic polarisation curve has been recorded for this layer and was, together with the anodic polarisation curves of the untreated and the plasma-nitrided substrates adjusted towards a common point of intersection, Fig. 7a, b, c. This demonstrates galvanic corrosion as it occurs at a coating-substrate

composite consisting of an porous layer with a high corrosion potential on a substrate with a clearly lower corrosion potential. As far as this type of galvanic corrosion is concerned, the cathodic partial reaction which may stand for a hydrogen deposition or oxygen reduction mainly takes place on the surface of the layer. The anodic partial reaction which stands for material dissolution takes place in micro porosities and layer defects at the substrate; thus in places where the substrate has contact with the corrosion medium. For reasons of electrical neutrality, the anodic and the cathodic partial current must be equal as long as current discharge to the outside does not take place. That means that if the cathodic area is equal in size with the anodic area, the current densities must also be equal. Provided that the areas are of equal size, the intersection points of the anodic and the cathodic current density curves represent the theoretically determined corrosion potential of the layer on the according substrate (CP1 and CP2), Fig. 7. The current densities which have been read off at these points and which are approx. 10 μA/cm2 at −370 mVH in the case of the untreated substrate (CP1) and approx.. 0.24 μA/cm2 at +116 mVH in the case of the plasma-nitrided substrate (CP2) are the theoretical corrosion current densities in the corrosion potential. Thus, the theoretical corrosion rate which is proportional to the corrosion current density is, for the coated and untreated substrate approximately 40 times higher than it is for the coated plasma-nitrided substrate. The previous theoretical reflections acted on the assumption that the cathodic area, the surface of the layer, is of equal size as the anodic area, the surface of the substrate in layer defects which is in contact

Fig. 4. Specimens after polarisation tests a) polished substrate, b) lapped substrate.

Fig. 6. Open circuit potential for different substrate surface conditions.

852

O. Durst et al. / Surface & Coatings Technology 203 (2008) 848–854

caustic soda with a pH value greater than 10. A passive material behaviour with a corrosion potential which was by 300 mV higher was observed. This positive shift of the corrosion potential as a consequence of the increased pH value has also been observed in the plasma-nitrided specimens and confirms the above test results. 3.2. Reciprocating sliding tests on the influence of plasma-nitriding on wear resistance

Fig. 7. Anodic polarisation curves for uncoated low-alloy steel and cathodic for ZrCg aC–Ml coated glass (abs. value), a) no pretreatment, b) plasma-nitrided, c) glass + ZrCg–aC–Ml.

with the corrosion medium. Since, in the initial state, the voids in the layer are only very small, the cathodic area is substantially larger than the anodic area. This could be described by a parallel upwards shift of the cathodic curve. The consequence is a higher corrosion potential of the coated specimens and a higher corrosion current density in the corrosion potential than theoretically estimated. This is also confirmed by the value of the corrosion potentials of coated specimens which are higher than the theoretically estimated corrosion potentials, Fig. 8. Due to the approximately equal rising of both anodic curves, however, the ratio of approx. 40:1 of the corrosion current densities of the coated untreated substrate to coated plasmanitrided substrate will remain approximately the same. In conclusion, it is determined that the plasma-nitriding of the substrate has a positive influence on the corrosion behaviour of the PVD coated specimens. The corrosion potentials were, in the case of a plasma-nitriding of the substrate, clearly higher than those without plasma-nitriding. The graphical determination of the corrosion current densities at the corrosion potential from the intersection of anodic and cathodic polarisation curves established, that the corrosion rates in the case of the plasma-nitrided low-alloy steel were clearly lower than those of the coated untreated low-alloy steel. The reason is assumed to be the alkalescent effect of the nitrogen, in accordance with N + 4 H+ + 3 e− ⇒ NH+4 [13]. The pH value of the corrosion medium is locally increased by the effect. This effect was investigated using an untreated low-alloy steel in alkaline 0.1 N

Fig. 8. Anodic polarisation curves for ZrCg–aC–Ml coated low-alloy steel, a) no pretreatment + ZrCg–aC–Ml, b) plasma-nitrided + ZrCg–aC–Ml.

In order to investigate the influence which plasma-nitriding of the substrate exerts on the wear behaviour, reciprocating sliding tests have been carried out. The used substrate was polished low-alloy steel. The substrate was applied with and without plasma-nitriding (mark.: untreated). A hydrogen-free graded multilayer zirconium carbide layer and a hydrogen-free graded multilayer chromium carbide layer were exemplarily selected for this study. The evaluation of the reciprocating sliding tests with regard to wear shows that with the increasing normal force the volumetric total wear (total wear of ball and specimen) has clearly increased, Fig. 9. If a normal force of 10 N has been exceeded, less wear is measured for those layers which have been deposited on plasma-nitrided substrate, curves c and d, Fig. 9, than this applies to the layers on untreated substrate, curves a and b, Fig. 9. This difference in wear increased with increasing normal force. Below 10 N this kind of differentiation is no longer possible. After each test, profile recordings have been carried out in the centre of the wear tracks. These profiles are depicted besides the ideal ball profile to demonstrate the wear of the ball, which is the difference between both profiles. The wear track profiles for the test with hydrogen-free graded multilayer zirconium carbide layers with a normal force of 50 N allow to determine, that the wear track for the layer on untreated substrate is obviously deeper than that of the layer on plasma-nitrided substrate, Figs. 10 and 11. Both wear tracks are clearly deeper than the layer thickness of approx. 5 μm (dashed line). Hence after local wearthrough of the layer not only the wear of the layer but also the wear of the substrate are part of the measured total wear, Figs. 10 and 11. Furthermore the profile recordings of the wear tracks deflect strongly from the ideal ball geometry. This means, that considerable wear occurred on the very hard ceramic silicon nitride ball. Comparable results had also been determined for the other tests downwards to 10 N. For normal forces below 10 N, wear-through of the layers in the centre of the wear tracks wasn't determined. That means that with these forces the wear occurred only on the layer and not on the substrate.

Fig. 9. Volumetric total wear as a function of the normal force in a reciprocating sliding test with silicon nitride ball and at 50% relative humidity, a) no pretreatment+ZrCg–aC–Ml, b) no pretreatment+ CrCg–aC–Ml, c) plasma-nitrided+ZrCg–aC–Ml, d) plasma-nitrided + CrCg–aC–Ml.

O. Durst et al. / Surface & Coatings Technology 203 (2008) 848–854

853

Fig. 10. Profile for 50 N and substrate with no pretreatment + ZrCg–aC–Ml. Fig. 12. Average friction coefficient for 2 N to 50 N according to Fig. 9.

The determination of the adhesive strength in accordance with the “VDI Fachbericht” 39 (adhesive strength 1 very good, adhesive strength 6 very bad) showed that the adhesion of both layers was very good. On the untreated substrate (adhesive strength 2) it was slightly worse than on the plasma-nitrided substrate (adhesive strength 1). The friction coefficients for all tests were between 0.7 and 1.0, Fig. 12. Definite connections between pretreatment, normal force and friction coefficients have, however, not been established. Representative scanning electron microscope images in combination with energy dispersive X-ray analysis of a wear track at a normal force level of 2 N reveal wear grooves inside the track, Fig. 13. But no delamination of the layer is visible. At 50 N SEM images show local delamination within the layer at the edge of the wear track, where only chromium without carbon remains on the substrate, Fig. 14. Inside the wear track grooves and exposed substrate material are visible. Furthermore a transfer layer consisting of silicon, iron and oxygen was formed inside the wear track. This implies a transfer of ball material (silicon) to the wear track and an oxidation during dry friction. Hence the dominant wear mechanism at low loads is abrasive wear. At higher loads, as a result of high contact pressures, a cohesive failure of the coating takes place. This leads to adhesive wear in combination with oxidation and a transfer layer build up. These results apply to both kinds of substrate, plasma-nitrided and non-plasma-nitrided. In summary, it can be concluded that plasma-nitriding for the two presented hydrogen-free layers did not exert significant influence on the wear as long as the normal force was below 10 N. In this case of low normal force, wear occurred mainly on the layer, not on the substrate.

There was no definite positive effect of plasma-nitriding of the substrate on the volumetric wear and friction coefficients detected. In the case of higher loads, the wear also affected the substrate since the layer was worn-through during the test. These tests established that the layers on plasma-nitrided substrates showed less traces of wear. The reason lies in the higher hardness of the plasma-nitrided low-alloy steel (800 HV0.3 vs. 700 HV0.3 with no plasma-nitriding) [16]. After the wear-through of the layer, the wear of the plasma-nitrided substrate is, due to the higher hardness, lower.

Fig. 11. Profile for 50 N and plasma-nitrided substrate + ZrCg–aC–Ml.

Fig. 13. SEM image of a wear track for 2 N and untreated substrate + CrCg–aC–Ml.

4. Discussion Different coating-substrate composites which consisted of graded zirconium carbide and graded chromium carbide layers and a lowalloy steel substrate have been investigated. Firstly the influence of different roughness of the substrate which was caused by different finishing methods (polishing, lapping) was investigated on one of the coating-substrate composites. It showed that consistent with investigations of other researchers [9–12] for this new carbon based coatings a higher roughness induced a higher number of layer defects which led to worse corrosion protection by the layer. The tested PVD coating itself did not show any traces of corrosion. Plasma-nitriding of low- and high-alloy steels to improve corrosion resistance was described in several publications [13–15]. Furthermore a duplex treatment consisting of plasma-nitriding and PVD coating was realised, e.g. to improve wear resistance [20]. But so far there were no investigations to quantify the influence of such a

854

O. Durst et al. / Surface & Coatings Technology 203 (2008) 848–854

can be given: The surface of the components should, on the one hand, be as smooth as possible. This allows to expect a low number of defects in the layers. On the other hand, the plasma-nitriding of the substrate surface is recommended in order to improve the corrosion resistance of the total coating-substrate composite. This can be attributed to an improvement of the corrosion resistance of the substrate due to an alkalescent effect of the embedded nitrogen. From the wear test results could be concluded, that there is no significant influence of plasmanitriding on wear for the investigated layers, as long as the ball doesn't have direct contact to the substrate and as long as layer adhesion is not significantly higher throughout plasma-nitriding. As a result of latest polarisation tests with enhanced PVD coatings based on the layers presented in this study, a very good corrosion behaviour was achieved. During the polarisation tests no corrosion pits occurred. Furthermore there are ongoing investigations on substrates coated with these enhanced PVD coatings subjected to a simultaneous mechanical-corrosive load in a new testing rig, which realizes an erosive load in a corrosive medium. Fig. 14. SEM image with energy dispersive X-ray analysis of a wear track for 50 N and untreated substrate + CrCg–aC–Ml.

duplex treatment on corrosion current densities and hence corrosion rates. So this kind of investigation was one objective of this study by means of polarisation tests with determination of corrosion potentials and corrosion current densities. One result was that for all layers on plasma-nitrided substrates the corrosion potential was clearly higher in contrast to the layers on non-plasma-nitrided substrates. The highest corrosion potential was measured for a hydrogen-free graded chromium carbide layer. This implies a good corrosion behaviour of this coating-substrate composite, which is also confirmed by ongoing polarisation tests. By means of a graded hydrogen-free zirconium carbide layer on a glass substrate it was established that the corrosion rate for the coating-substrate composite with plasma-nitrided substrate was clearly lower than with non-plasma-nitrided substrate. The result was a ratio of approx. 40:1 for the corrosion rates of the coated untreated substrate to the coated plasma-nitrided substrate. It is assumed that one reason for this is the alkalescent effect of the implanted nitrogen which causes a local increase of the pH value of the corrosion medium in layer defects and therefore hinders substrate corrosion. For the other investigated layers could be concluded, that the higher corrosion potential for these layers on plasma-nitrided low-alloy steel comes along with lower corrosion rates. On the basis of graded, hydrogen-free zirconium carbide and graded hydrogen-free chromium carbide layers, the influence of plasma-nitriding of the substrates on the wear behaviour was investigated. At that, the positive influence of plasma-nitriding was established only when the normal force during the test was sufficiently high for allowing the substrate to make direct contact with the counter body. This was ascribed to the higher hardness of the plasma-nitrided substrate. The dominant wear mechanism at low loads was abrasive wear. At higher loads as a result of high contact pressures a cohesive failure of the coating took place leading to adhesive wear in combination with oxidation and a transfer layer build up. 5. Conclusions In summary, two recommendations for technical applications of the investigated PVD coatings as corrosion protection on low-alloy steels

Acknowledgments The research project (no. 14866) of the Forschungsvereinigung “Forschungskuratorium Maschinenbau (FKM) e.V.” has been supported financially within the scope of the program about the support of the “Industriellen Gemeinschaftsforschung (IGF)” by the “Bundesministerium für Wirtschaft und Technologie” via the AiF. We wish to express our thanks to the Bundesministerium, the AiF and also to the FKM for their research promotion and financial support. We would, moreover, like to express our thanks to the assisting study group of the “Verband Deutscher Maschinen- und Anlagenbau e.V. (VDMA)” for the friendly support and also to the “Institut für Oberflächentechnik der RWTH Aachen” for supplying the coated specimens. References [1] C.S. Lin, C.S. Ke, H. Peng, Surf. Coat. Technol. 146–147 (2001) 168. [2] T. Savisalo, D.B. Lewis, Q. Luo, M. Bolton, P. Hovsepian, Surf. Coat. Technol. 202 (2008) 1661. [3] M.A. Wall, et al., Surf. Sci. 581 (2005) L122. [4] Y. Purundare, M.M. Stack, P. Hovsepian, Wear 259 (2005) 256. [5] D. Probst, H. Hoche, Y. Zhou, R. Hauser, T. Stelzner, H. Scheerer, E. Broszeit, C. Berger, R. Riedel, H. Stafast, Surf. Coat. Technol. 200 (2005) 355. [6] Z. Zhou, et al., Wear 263 (2007) 1328. [7] W.-D. Münz, et al., Surf. Eng. 17 (2001) 15. [8] W.-D. Münz, et al., Surf. Coat. Technol. 125 (2000) 269. [9] J. Munemasa, T. Kumakiri, Surf. Coat. Technol. 49 (1991) 496. [10] C. Liu, A. Leyland, S. Lyon, A. Matthews, Surf. Coat. Technol. 76 (1995) 623. [11] Shen-Chih Lee, Wie-Yu Ho, G.D. Lai, Mater. Chem. Phys. 42 (1996) 266. [12] J.M. Soro, L. Lelait, J.C. van Duysen, G. Zacharie, J. von Stebut, Surf. Coat. Technol. 98 (1998) 1490. [13] L.C. Gontijo, R. Machado, S.E. Kuri, L.C. Casteletti, P.A.P. Nascente, Thin Solid Films 515 (2006) 1093. [14] K. Ichii, K. Fujimura, T. Takase, Technol. Rep. Kansai Univ. 27 (1986) 135. [15] Ellina Lunarska, O. Czerniajewa, A. Nakonieczny, Vacuum 63 (2001) 469. [16] J.C.A. Batista, M.C. Joseph, C. Godoy, A. Matthews, Wear 249 (2001) 971. [17] B. Podgornik, J. Vizintin, O. Wänstrand, M. Larsson, S. Hogmark, H. Ronkainen, K. Holmberg, Wear 249 (2001) 254. [18] E. Lugscheider, O. Knotek, K. Bobzin, M. Maes, K. Arntz, MRS Fall Meeting, vol. 45, 2002. [19] K. Bobzin, E. Lugscheider, M. Maes, Mat.-wiss. u. Werkstofftech. 35 (2004) 112. [20] E. Broszeit, B. Matthes, W. Herr, H.-J. Spies, K. Hoeck, Surf. Coat. Technol. 60 (1993) 441.