Abstract
A new test method based on multipass scratch testing has been developed for evaluating the mechanical and tribological properties of thin, hard coatings. The proposed test method uses a pin-on-disc tribometer and during testing a Rockwell C diamond stylus is used as the “pin” and loaded against the rotating coated sample. The influence of normal load on the number of cycles to coating damage is investigated and the resulting coating damage mechanisms are evaluated by posttest scanning electron microscopy. The present study presents the test method by evaluating the performance of Ti0.86Si0.14N, Ti0.34Al0.66N, and (Al0.7Cr0.3)2O3 coatings deposited by cathodic arc evaporation on cemented carbide inserts. The results show that the test method is quick, simple, and reproducible and can preferably be used to obtain relevant data concerning the fatigue, wear, chipping, and spalling characteristics of different coating-substrate composites. The test method can be used as a virtually nondestructive test and, for example, be used to evaluate the fatigue and wear resistance as well as the cohesive and adhesive interfacial strength of coated cemented carbide inserts prior to cutting tests.
1. Introduction
Although CVD (chemical vapour deposition) and PVD (physical vapour deposition) coatings have been used in order to increase the lifetime and performance of cutting tools for more than 30 years there is still an intensive research interest to improve the performance of existing coating materials and to develop new coating materials and coating processes. Even though many test methods have been developed to simulate the contact conditions prevailing during metal cutting conditions, cutting tests are still the most reliable methods in order to characterize the properties and performance of coatings-substrate systems aimed for metal cutting operations [1, 2]. However, cutting tests including in-depth post-test characterization of the worn cutting tools can be both costly and time-consuming. Consequently, laboratory tests, although performed at room temperature, can preferably be used in order to evaluate specific properties such as the cohesive and adhesive strength, intrinsic abrasion resistance, and so forth of specific coating systems. As a result, a wide range of laboratory techniques are available to assess the mechanical and tribological characteristics of CVD and PVD coatings for metal cutting tools [3–13].
Today, the scratch test is probably the most common and used test to increase the understanding of the mechanical response of a coating-substrate system exposed to an external load [12]. In the conventional scratch test a Rockwell C diamond stylus (120° cone with 200 μm radius) is drawn across the coated substrate surface under an increasing normal load until some kind of well-defined coating failure occurs, frequently associated with localized chipping or spalling [3]. The corresponding normal load is termed the critical load and is often used as a measure of the coating mechanical strength or “practical adhesion.” The critical load can be detected by on-line monitoring of the friction force and acoustic emission signals and/or by post-test examination using optical or scanning electron microscopy [12–18].
One drawback with the conventional scratch test is the need for such high normal loads that the deformation behaviour of the substrate material in many cases controls the mechanical response of the coating-substrate system, including the adhesion of the coating to the underlying substrate. This is to a large extent due to the fact that the coating deposition processes including substrate cleaning and pre-treatment of today result in excellent adhesion.
An alternative test, able to evaluate the mechanical properties including the fatigue resistance of CVD and PVD coatings is the repetitive impact test [5, 19–21]. In this test the coated substrate is exposed to an oscillating indenter, usually a ceramic or a cemented carbide ball, impacting the surface at a constant maximum force during a large number of impacts, typically 106. The major drawbacks with this test are the large contact area and the fact that compressive stresses dominates while in a typical tribological contact coating failure is mainly due to the presence of tensile stresses [19, 20].
Consequently, a more realistic testing approach would be a test based on multi-pass scratch testing performed at lower loads, that is, loads significantly lower than the critical normal loads obtained in the conventional scratch test. In the multi-pass scratch test, a constant normal load lower than the critical load, is applied to the same scratch track for a large number of repeated cycles [22–27].
The objectives of the present study are to:(a) propose a new test concept for multi-pass scratch testing using a pin-on-disc tribometer;(b) evaluate the new test concept by studying the multi-pass fatigue andwear properties of three different PVD coatings used for metal cutting applications.
The new test approach is based on a three-step procedure aimed at evaluating the mechanical and tribological properties of CVD and PVD coatings under repeated scratching contact conditions. In the first step, conventional single pass scratch testing using a scratch tester is carried out to determine the critical normal load(s) for coating failure. In the second step, the data obtained from the conventional scratch test are used to perform multi-pass scratch testing at several sub-critical normal loads using a pin-on-disc tribometer. Finally, post-test examination of the tested samples is performed using scanning electron microscopy and energy dispersive X-ray spectroscopy in order to determine the major coating failure mechanisms.
2. Experimental
2.1. Materials
Commercial cemented carbide inserts with a composition of 94 wt% WC and 6 wt% Co and hardness 1700 HV0.020 were used as substrate material in the present study. Prior to coating deposition the substrate surfaces were polished to a surface finish of . Three different PVD coatings, Ti0.86Si0.14N, Ti0.34Al0.66N and (Al0.7Cr0.3)2O3, representing state of the art PVD coatings for metal cutting applications were included in the study. The coatings were deposited using cathodic arc evaporation with a coating growth rate of approximately 2 μm/h using the deposition process parameters according to Table 1. When depositing the Ti0.86Si0.14N and (Al0.7Cr0.3)2O3 coatings a thin Ti0.34Al0.66N, layer (thickness 0.5 μm) was deposited in order to promote the adhesion to the underlying substrate. It should be noted that the chemical composition of the coatings has been determined by combining results from energy dispersive X-ray spectroscopy (EDX), Rutherford backscattering spectrometry (RBS) and elastic recoil detection analysis (ERDA).
2.2. Coating Characterization
The surface morphology, microstructure, and element composition of the as-deposited coatings were investigated using a 3D optical surface profilometer (WYKO NT-9100) and a ZEISS Ultra 55 field emission Gun Scanning Electron Microscope (FEG-SEM) equipped with an Oxford INCA Energy Dispersive X-Ray spectroscopy (EDX) system. The presence of defects (surface irregulaties), that is, droplets and craters in the as-deposited coating surfaces was evaluated using stereological methods (disector counting rule and point grid method) [28]. Only defects larger than 0.4 μm were considered in five representative areas () in the centre of each sample.
2.3. Mechanical Testing
The Vickers hardness values of the coatings and the coating-substrate composites were measured using a Micro Combi Tester (CSM Instruments), using a load of 200 mN and 4000 mN, respectively. The indents using the lower load were all done in flat areas free from visible defects, resulting in hardness values for the defect free coating material. Loading and unloading were performed for 30 s, respectively, with a holding period of 15 s at the maximum load. By using the obtained indentation curves (load versus displacement) the hardness and Young’s modulus were calculated according to Oliver-Pharr [29].
The Micro Combi Tester was also used in order to evaluate the resistance to scratch induced failure of the coatings. Single pass scratch testing under increasing normal load (0–30 N) with a scratch length of 10 mm and a scratching velocity of 10 mm/min was performed using a 50 μm radius Rockwell C diamond stylus. Besides, conventional single pass scratch testing, using the CSM Instruments Revetest, was performed using a 200 μm radius Rockwell C diamond stylus. A normal load range of 0–100 N, a scratch length of 10 mm and a scratching velocity of 10 mm/min were used in the experiments. During scratch testing the friction force and acoustic emission (A.E.) signals were continuously recorded.
In the scratch tests, two critical normal loads ( and , resp.) were used as a measure of the coating cohesive and interfacial adhesive strength and were taken as the normal load corresponding to the first local coating failure () and continuous substrate exposure () as determined by post-test characterization using SEM.
2.4. Tribological Testing
Circular multi-pass scratch testing was performed in order to evaluate the fatigue and wear behaviour of the coatings under repeated scratching contact conditions. The experiments were performed using a pin-on-disc test equipment (CSM High Temperature Tribometer), see Figure 1, but instead of a pin or a ball commonly used in standard sliding wear tests a Rockwell C diamond stylus used in conventional scratch testing (using a scratch tester) was used. The pin-on-disc test set-up makes it possible to significantly reduce the extensive time associated with multi pass scratch testing for a large number of cycles (>10) using conventional scratch testing equipment.
The circular multi-pass tests were carried out using a 50 μm radius Rockwell C diamond stylus, a speed of 90 mm/min and with a stepwise (loading step 2 N) increasing normal load from 3 N up to the critical normal load, , as obtained from the conventional single pass scratch tests. The maximum number of cycles was set to 2000. The radius of the circular scratch track was 1.5, 2.0, or 2.5 mm. It should be noted that as long as the diamond stylus used in the experiment is not damaged no influence from a variation in circular scratch track radius on coating failure tendency could be detected within the radius range and test parameters used.
All tests were performed in ambient air (20–22°C, 25-26% RH) and repeated two times. During testing the friction coefficient was continuously recorded and, in combination with post-test characterisation of the wear track using optical microscopy, used in order to determine the number of cycles to coating failure. After testing all wear tracks were carefully characterized by SEM and EDX in order to evaluate the influence of normal load and number of scratching cycles on the prevailing coating failure mechanisms.
3. Results
3.1. Coating Morphology and Microstructure
Figures 2(a) and 2(b) show the surface morphology and microstructure of the as-deposited Ti0.34Al0.66N coating investigated. As can be seen, the as-deposited coating surface shows a significant roughness due to the generation of μm- and sub-μm-sized droplets during the deposition process. Of these, the larger droplets (>3–5 μm in diameter) typically show an interdroplet distance in the range of 15–25 μm, that is, they constitute a nonnegligible area fraction of the coating. Also, μm and sub-μm-sized craters contribute to the observed roughness. EDX-analysis of the droplets shows that these display a higher metal content, especially in the core, as compared with the surrounding coating matrix. The above observations are common for all coatings investigated. However, when comparing the three coatings, see Table 2, it is observed that the (Al0.7Cr0.3)2O3 coating shows a significantly rougher surface as compared with the Ti0.86Si0.14N and Ti0.34Al0.66N coatings. The reason for this is mainly due to a large number of large, up to 5-6 μm in diameter, droplets as compared to the other two coatings. The surface roughness and defect density are about the same for the Ti0.86Si0.14N and Ti0.34Al0.66N coatings, with slightly higher values for the Ti0.86Si0.14N coating.
(a)
(b)
3.2. Mechanical Properties
3.2.1. Hardness
Figure 3 shows representative load-displacement curves for the coatings. From these curves, the corresponding Young’s modulus, E, and hardness values can be calculated, see Table 2. As can be seen, the Ti0.86Si0.14N coating displays the highest hardness while the (Al0.7Cr0.3)2O3 coating displays the lowest hardness. The (Al0.7Cr0.3)2O3 coating also displays the lowest E-modulus while the Ti0.34Al0.66N coating displays the highest E-modulus.
3.2.2. Single Pass Scratch Test Behaviour
Table 3 shows the critical normal loads, and , as obtained in the conventional single pass scratch test of the coatings investigated. As expected, the larger 200 μm radius diamond stylus results in significantly higher critical normal loads. As can be seen, the Ti0.34Al0.66N coating displays the highest critical normal loads and the (Al0.7Cr0.3)2O3 coating the lowest critical normal loads. Also, the difference between the two critical normal load values, that is, and , is for all coatings relatively small.
Examination of the scratch tested samples revealed that a number of coating failure and detachment mechanisms could be identified. Minor cracking and chipping of small, sub-μm, coating fragments in connection to droplets and craters were sporadically observed already at very low normal loads (≈1-2 N) dependent on the size and type of defect. However, these failures were all very limited in size and restricted to the local defects. With increasing normal load the coating-substrate composites displayed the following major scratch induced damage mechanisms; plastic deformation and minor cracking (no coating material detached) within the scratch, chipping (cohesive failure) in combination with local interfacial spalling(adhesive failure) outside the scratch, and continuous interfacial spalling(adhesive failure) outside the scratch, see Figure 4. The characteristics of the spalled off areas reveal that the observed adhesive failure is associated with the elastic recovery which occurs upon unloading behind the diamond stylus as it scratches the surface [30]. Consequently, plastic and elastic deformation of the cemented carbide substrate plays an important role for coating damage and as a result the tendency to chipping and spalling will decrease with decreasing normal load, that is, decreasing plastic deformation of the cemented carbide substrate. For normal loads above , extensive plastic deformation within the scratch will occur resulting in coating fragments being pressed or squeezed into the cemented carbide substrate surface.
(a)
(b)
(c)
(d)
3.2.3. Multipass Scratch Test Behaviour
Figure 5 shows the results from the multi-pass scratch test where the influence of number of scratching cycles (N) and load (normalised with respect to as obtained in the single pass scratch test) on the onset of continuous interfacial spalling along the circular wear track is presented. As can be seen, although minor differences exist, all coatings display a similar appearance with an increasing number of cycles to coating failure with decreasing normal load until a critical load, corresponding to a “fatigue limit”, is reached below which no major coating failure such as chipping and spalling is obtained up to 2000 cycles. Further, it should be mentioned that the values of the critical loads are obtained by combining detailed characterisation of the scratch tracks with detection of changes in the acquired friction curves. An increase in friction, see Figure 6, indicates that coating failure occurs.
(a)
(b)
Examination of the circular wear tracks after different combinations of normal load and number of cycles shows that the major failure mechanisms observed in the conventional single pass scratch tests also are present in the multi-pass scratch tests, see Figures 7 and 8. Further, Figure 8(c) shows a crack obtained in the substrate/coating interface at the rim of the scratch. However, the multi-pass test also result in two new types of coating failure controlling the coating life time at a large number of cycles.
(a)
(b)
(c)
(a)
(b)
(c)
For normal loads below the “fatigue limit” gradual wear of the coating occurs. Two different wear mechanisms can be distinguished both acting on a sub-μm scale and typically after a large number of cycles. First, chipping of fine coating fragments is observed in connection to fatigue cracks obtained after a critical number of cycles, see Figure 9(a). Secondly, delamination of thin plate-like fragments, originating from a thin tribo film formed by compacting and sintering of sub-μm coating fragments at the diamond tip-coating interface, see Figure 9(b), may also take place.
(a)
(b)
Consequently, for a combination of low loads, not resulting in pronounced deformation of the cemented carbide substrate, and large number of cycles the intrinsic tribological properties of the coating controls the life of the coating. It should be noted that the defects (droplets and craters) present in the coating do not necessarily result in any extensive coating wear at low normal loads. Also, the observed wear rate is relatively low; that is, the maximum reduction in coating thickness after 2000 cycles does not exceed 15% of the coating thickness for the coatings investigated.
For normal loads slightly above the fatigue limit large scale interfacial spalling resulting in complete exposure of the cemented carbide substrate within the scratch is observed after a large number of repeated contacts (), see Figure 10. Consequently, fatigue controlled crack propagation at the coating-substrate interface has a significant influence on coating failure.
(a)
(b)
Figure 11 shows a coating failure map where the influence of normalised normal load and number of scratching cycles on the major damage mechanisms for the Ti0.34Al0.66N coating is illustrated. In the map, the transition from a region (grey) corresponding to plastic deformation, minor cracking, and wear to a region corresponding to continuous interfacial spalling (dashed) is represented by a relatively narrow band corresponding to chipping and local interfacial spalling. The narrow band signifies a quick transition from plastic deformation, minor cracking and wear via local interfacial spalling to continuous interfacial spalling with increasing normal load.
Region A (). For very low normalised normal loads the coating displays just a mild plastic deformation as revealed by the flattening of the as-deposited surface morphology in the scratch. With increasing normal load the plastic deformation extends through the coating into the substrate and with increasing number of cycles the coating starts to display fatigue induced cracking and mild wear caused by chipping or delamination of fine coating fragments.
Region B (). For normalised normal loads below 0.20 complete interfacial spalling will occur at a high number of cycles (). For normalised normal loads slightly above 0.20 coating damage by chipping and local interfacial spalling starts to occur at a critical number of cycles (). Consequently, coating damage is affected by crack propagation within the coating and along the coating-substrate interface and high cycle fatigue controls the damage. With increasing normalised normal load, the critical number of cycles to chipping and spalling will rapidly decrease (dN/dF ) until a normalised normal load around 0.25 is reached where a transition to low cycle fatigue is obtained.
Region C. (): For normalised normal loads above 0.30 the critical number of cycles to chipping and spalling is relatively low, typically less than 250–350. In this region the normal load is high enough causing extensive deformation of the cemented carbide substrate and the critical number of cycles to chipping and spalling is to a large extent controlled by the stress state and defect density at the coating-substrate interface promoting crack propagation during repeated cyclic loading.
Region D. Coating damage controlled chipping and interfacial spalling mainly due to elastic recovery caused by pronounced plastic deformation of the cemented carbide substrate during the scratching process. This failure occurs for normalised normal loads and is restricted to a low number of cycles, typically less than 10. The decrease in number of cycles to local and interfacial spalling with increasing normal load is very low (dN/dF ) in this region.
Further, retesting of the Ti0.34Al0.66N coating after long term testing shows equal results as compared to former test, see Figure 12. The differences are most evident at higher loads but are subtle which indicates that the technique is reproducible.
4. Discussion
4.1. Proposed Test Method
In the present study, circular multi-pass scratch testing using a pin-on-disc equipment is proposed as a test method to evaluate the fatigue, wear, chipping, and spalling behaviour of thin hard coatings under repeated/cyclic sliding contact conditions. The results show that the test method is quick, simple and reproducible and can preferably be used to obtain relevant data concerning the mechanical and tribological properties of different coating-substrate composites. Other advantages of the proposed test method are.(i) The test method utilizes a commercially available tribometer, an equipment widely used for tribological testing and present at many material laboratories. (ii) The test method uses small, simple flat samples making it possible to easily evaluate for example, the effect of different types of pre- and post-coating deposition surface treatment processes, and so forth. Also, the small sample size facilitates post-test microscopy and surface analysis.
Besides, the use of a small diamond stylus as in the present study offers the possibility to characterise the coating at a very small, close to single asperity, scale. Also, the small size of the test region (the radius of the circular scratch track can be as small as 0.5 mm) makes it possible to perform tests within small restricted areas. Consequently, the test method can be used as a virtually non-destructive test if the active region of the investigated sample is avoided. For example, it can be used to evaluate the cohesive and adhesive interfacial strength, fatigue, and wear resistance of coated cemented carbide inserts prior to cutting tests in order to correlate the obtained data to the tool life, see Figure 13.
Consequently, circular multi-pass scratch testing is a fast and powerful complement to conventional single pass scratch testing being one of the most common tests for the mechanical characterization of thin hard coatings. In combination with post-test scanning electron microscopy the proposed test method will give information about not only the critical normal loads and coating damage mechanisms at single pass scratch testing and a low number of loading cycles but also the fatigue and wear properties at a high number of loading cycles, . Compared with the common used impact tests used for evaluating the fatigue properties of thin hard coatings the proposed test is believed to more closely simulate the contact conditions and prevailing stress states in sliding contact tribosystems although the number of cycles are significantly lower than in the impact tests [5, 19–21].
It may be questioned whether the test approach, which comprises a large number of loading cycles, will result in extensive wear of the diamond stylus. Steinmann et al. found that unless the diamond tip shows extensive chipping due to “half-crown” shaped cracking the influence of diamond tip wear on the critical normal load is not significant [31]. However, as stated by the authors regular microscopic observations of the diamond tips are of utmost importance. Also, the use of high quality natural single-crystal diamond tips with the [100] direction oriented along the -axis can substantially increase their life time. Furthermore, in order to optimize the life time the diamond stylus should be rotated 180° before catastrophic damage is reached. Besides the above mechanical wear, the diamond tip may, depending on the contact conditions and the surrounding environment, suffer from tribo-oxidation and graphitization [32], which will influence on the wear resistance as well as the friction at the scratching (sliding) interface. Consequently, high scratching (sliding) speeds should preferably be avoided. In the present study, two high quality diamond styli with guaranteed spherical geometry were used and the wear/damage of these was restricted to mild polishing and initial “half-crown” shaped cracking after completing the test series.
4.2. Coating Damage Mechanisms
When comparing the results from the conventional single pass scratch test with the circular multi-pass scratch test it can be concluded that both tests display similar major coating failure mechanisms as long as the normal loads in the multi-pass test is relatively high, typically , and is restricted to a low number of cycles, typically less than 10. Within this region coating failure is due to: plastic deformation and minor cracking (no coating material detached), chipping (cohesive failure) in combination with local interfacial spalling(adhesive failure), and continuous interfacial spalling(adhesive failure) outside the scratch. Also, the characteristics of the spalled off areas reveals that the adhesive failures in both tests are associated with the elastic recovery which occurs behind the diamond stylus as it scratches the surface [30]. However, while the conventional single pass scratch test mainly gives information of the mechanical response of the coating-substrate composite, the circular multi-pass scratch test also gives information of the tendency to surface fatigue and wear when exposed to a large number of contact cycles. Besides, it also makes it possible to determine a “fatigue limit” for the coating-substrate composite. In Table 4 and Figure 14 the observed mechanisms are illustrated by simplified schematic drawings and discussed in some detail.
The ranking of the coatings investigated is, keeping in mind that all coatings were deposited with identical interfacial properties, believed to be controlled by the differences in mechanical properties and defect density. Thus, the ranking with respect to spalling resistance obtained at low normal loads, that is, Ti0.86Si0.14N > Ti0.34Al0.66N > (Al0.7Cr0.3)2O3, can be explained by the fact that the Ti0.86Si0.14N coating displays the highest hardness and (Al0.7Cr0.3)2O3 the lowest hardness, a high hardness reducing the tendency to plastic deformation of the substrate material and thus the stresses at the coating-substrate interface. Increasing normal load will increase the stresses at the interface and consequently the strength of the interface region will have an increased impact on the tendency to spalling. Thus the ranking displayed at normalized normal load above 0.5, that is, Ti0.34Al0.66N> Ti0.86Si0.14N > (Al0.7Cr0.3)2O3, can be explained by the fact that the Ti0.34Al0.66N displays the lowest defect density and (Al0.7Cr0.3)2O3 displays the largest defect density of the coatings investigated, see Figure 5 and Table 2. At even higher normalized normal loads the tendency to plastic deformation of the cemented carbide substrate will increase and consequently have a strong impact on the tendency to spalling. As a result the different coatings will display similar spalling characteristics, see Figure 5.
5. Conclusions
In the present study a new test method based on multi-pass scratch testing is presented and used in order to evaluate the mechanical and tribological properties of arc evaporated coatings of Ti0.86Si0.14N, Ti0.34Al0.66N and (Al0.7Cr0.3)2O3 deposited on cemented carbide inserts. Based on the obtained results the following conclusions can be drawn.(i)The multi-pass test method utilizes a commercially available pin-on-disc tribometer, an instrument widely used for evaluating the friction and sliding wear characteristics of materials, present in many tribology laboratories making the test easily accessible.(ii)The multi-pass scratch test method is quick, simple, and reproducible and can preferably be used to obtain relevant information concerning the fatigue, wear, chipping, and spalling characteristics of different coating-substrate composites.(iii)The simple and small sample size makes the test cheap and easy to combine with post-test microscopy and surface analysis. Also, the small size of the test region (the radius of the circular scratch track can be as small as 0.5 mm) makes it possible to perform tests within small restricted areas; that is, the test can be regarded as a virtually non-destructive test.(iv)As long as the wear/damage of the diamond tip is restricted to mild polishing and initial “half-crown” shaped cracking the effect on the scratching process and the resulting critical normal loads is negligible.(v)Of the coatings investigated, the Ti0.34Al0.66N coating displays the best performance under low cycle fatigue conditions while the harder Ti0.86Si0.14N shows the best performance under high cycle fatigue conditions.(vi)The intrinsic coating properties, including the defect density, have a significant impact on the fatigue properties of the coatings. Consequently, the presence of defects in the coating and the coating-substrate interface decreases the low cycle fatigue properties while the combination of coating defects and a low coating hardness decreases the high cycle fatigue.
Acknowledgment
The financial support for this work from the National Graduate School in Materials Science is gratefully acknowledged.