The Influence of Bias Voltage and Gas Pressure on Edge Covering during the Arc-PVD Deposition of Hard Coatings

Abstract

The edge area is especially essential for cutting tools, since this is the contact zone between the work piece and the tool. Hard coatings (PVD or CVD coatings) can protect the edge against wear and they are commonly used. The geometries of the cutting edges change during the coating process, with the edge radius increasing. Therefore, the film thickness is limited and the initial radius of the uncoated tool must be smaller than the target radius of the coated edge. A new coating process based on vacuum arc PVD was developed to overcome this limitation. The film growth at the edges can be properly controlled by means of selected coating materials and process conditions. Thus, it is possible to grow edges sharper than the initial edge geometry. Different substrates were coated with different coating systems. Parameters such as the bias voltage, coating pressure, and initial radius were varied within this work. It was found that the application of a bias voltage is crucial for the generation of sharp edges. It was also found that the edge sharpening caused by coatings works best on samples with an initial radius of around 15 µm.

1. Introduction

In recent years, physical vapor deposition (PVD) has established itself as one of the most important coating deposition methods for tools such as milling cutters, drills, and indexable inserts [1,2]. Coating with hard materials leads to a significant improvement in the performance and lifetime of such cutting tools [1,2,3]. However, the uniform coating of such complex shapes is challenging. Geometry-related shadowing effects or the formation of strong rounding on sharp edges due to the large range of incidence angles for coating-forming particles counteract a contour-accurate coating [4]. PVD processes often use negative bias potentials on the substrates in order to force the positively charged metal ions towards the substrate. The electric field lines of such electric fields are distorted depending on the shape of the substrate and can lead to local field enhancement at peaks or edges and to field weakening at holes or cavities. On the other hand, the electric field arranges itself homogeneously on flat surfaces [5].

Tool edges have to fulfil various requirements, including material removal, ensuring a high surface quality, avoiding sticking of the machined material on the tool edge surface, minimizing the load applied to the tool, the formation of homogenous flakes, and a high productivity of the cutting process [6,7].

The edges suitable for cutting processes need careful preparation to avoid the uneven edges that may stem from tool processing. Tikal et al. [6] provided an overview of the different edge preparation procedures for commonly used tools. Usually, edge rounding or chamfers are applied. However, edge preparation results in an increase in the cutting edge radius. If a PVD coating is applied to the prepared edges, the initial edge radius increases further by the amount of the coating thickness, following Equation (1):

$$r_C=r_A+d_{coating}$$

(1)

with the coated edge radius being r_C, the initial edge radius r_A, and the applied coating thickness d_coating. A layer thickness of only a few µm is usually applied in order to limit the rounding. A thicker layer would be desirable for a sufficient load reserve, but this would cause massive rounding of the edges. Coating processes that can produce sharp cutting edges are therefore in demand. Up to now, such processes have usually involved subsequent treatment of the coated substrates. The aim of this work is to present a coating process with an intrinsic sharpening effect.

Table 1. Important studies of coating deposition on complex-shaped substrates by the means of PVD and CVD deposition processes. Abbreviations: magnetron sputtering (MS); reactive magnetron sputtering (RMS); dual magnetron sputtering (DMS); unbalanced magnetron sputtering (UBMS); cathodic arc evaporation (CAE); hollow cathode deposition (HCD); and filtered cathodic vacuum arc deposition (FCVAD).

Coating Material	Deposition Process	Main Results	References
TiN	RMS	There was a TiN layer thickness variation at different bias potentials: between −350 and −500 V, no coating could be deposited on the sample edges, while at a lower bias potential of −200 V, TiN coatings could be deposited. Even thicker layers could be achieved at bias potentials between −32 and −150 V, due to lower back sputtering. Increasing the Ar gas pressure weakened the edge effect; the back sputtering was suppressed; and the coating thickness increased.	[8]
TiN	MS	There was a determination of geometry-dependent hardness curves. Denser and harder layers were formed near the edges. A larger edge angle yielded a harder deposited layer.	[9]
TiN	DMS + UBMS	There was a determination of the geometry-dependent layer thickness distribution. Sharper edges yielded a lower layer thickness; the hardness increased towards the edge; and a sharper edge yielded a lower hardness.	[10]
TiN	RMS, CAE, HCD	The coating thickness on the edges decreased due to back sputtering, and the coating thickness decreased with a decreasing wedge angle. With CAE coating, there was a better depth coatability due to higher ionization of the plasma; a higher gas pressure led to a higher coating thickness at the edge; the hardness at the edges was lower than on the flat specimen. Furthermore, an analysis of the coatability of edges with edge angles of 30°, 54°, and 90°, as well as the recesses, was carried out. Compared to other coating processes, CAE provided the highest coating rates, as well as a relatively uniform coating thickness distribution along the edges. At an edge angle of 30°, the coating thickness decreased with the increasing distance from the edge, while at 54° and 90°, the coating thickness initially decreased but increased again with an increasing distance from the edge. When determining the coating hardness, it was found that coatings produced with CAE have a higher coating hardness than coatings deposited by RMS.	[4]
CrN	CAE	The performance of coatings on samples with different edge angles and a flat sample was considered. When applying a bias voltage, an increased accumulation of macro particles and thus a roughening of the edge was observed with a decreasing edge angle and an increasing voltage between −100 V and −400 V. In addition, an increase in the coating rate with an increasing edge angle was observed in the edge region compared to a flat sample. The accumulation of macro particles could be reduced by a pulsed bias voltage.	[11]
TiN	CAE, RMS	Two geometries (cutting edges with right-angled and acute-angled edges) were coated at bias voltages of 0 V, −30 V, and −60 V. As the bias voltage increased, the coating rate decreased and the residual compressive stresses increased. During the bias voltage coating process, increased ion bombardment occurred at the edges and tips of the curved samples.	[12]
TiAlCrYN, TiAlN/VN,	CAE + UBMS	Coatings deposited on the edges under increased ion flux density showed different properties than coatings on flat substrates. For this purpose, a comparative study of TiAlCrYN layers on different samples (angles of 30°, 45°, and 60° and radii of 2 µm and 100 µm) was carried out. The applied bias voltage was −75 V. The Al/Ti ratio increased slightly with increasing distance from the edge and an increasing edge angle, indicating partial back sputtering back Al at the edge. While the Al/Ti ratio was almost the same for all samples at the edge, it diverged with an increasing distance from the edge. For the 60° edge, the Al/Ti ratio reached the value of the flat sample with an increasing edge distance. The influence of the edge radius was more pronounced for the 30° edge than for the 60° edge. Smaller radius yielded a smaller Al/Ti ratio, due to the stronger sputtering effect. All edges had a thicker layer compared to the flat samples, with the effect decreasing with an increasing edge angle. This was in contrast to the investigations of TiAlN/VN coatings. The coatings were deposited at −75 V to −150 V. An influence of the edge (angles of 30°, 45°, and 90°) on the angle of incidence of the ions was found, resulting in changes of the morphology, thickness, composition, and structure of the layers. Directly at the edge (up to 0.3 mm), the layer was completely absent; and in the area near the edge (1–2 mm from the edge), the layer started to partially delaminate, which was accompanied by a decrease in the Al/Ti ratio and an increase in the ion incidence angle and ion flux density. Undisturbed regions far from the edge were characterized by a dense layer, a constant Al/Ti ratio, a constant back sputtering rate of Al, a constant ion flux density, and an ion incidence angle of 0°. Directly at the edge, the Al back sputter rate was an order of magnitude higher than in the undisturbed region far from the edge, and the incidence angle increased from 0° to almost 90°. In addition, there was a strong increase in the ion flux density. The following influencing variables were observed. The edge effects were related to the shape of the plasma boundary layer, which was influenced by the edge geometry. The re-sputtering rate depended on the energy of the ion flux (bias-dependent) and the type of ions (Ar ions lead to a higher sputtering yield than N ions). In summary, a zone diagram was developed which describes the relationship between the bias voltage and edge angle. There was a division into three zones: Zone 1:, where the edges have a homogeneous and closed layer even away from the edge; Zone 2. Where the layer along the edge is partially missing; and Zone 3, where the layer along the edges is completely missing.	[13,14,15,16]
Diamond	CVD	There was the development of a plasma sharpening process with Ar ions for edge sharpening. The cutting edge became sharper than the initial tool.	[17,18,19]
TiAlCrSiN, TiAlN	CAE	Re-sputtering on the edges led to a sharpening effect, as the sputtering yield increased. A bias voltage of −80 V led to a slight sharpening of the edge and a high residual stress; a bias voltage of −40 V, on the other hand, led to less wear during machining.	[20]
AlCrSiN/TiN	CAE	This was an investigation of the influence of the bias voltage and the initial edge geometry on the edge radius and the sharpening of the cutting edge during coating. The wedge angle changed only slightly due to the coating.	[21,22]
TiAlN	Simulation, PVD	The fatigue and wear behavior of coated inserts with different cutting radii was investigated experimentally. The inserts with cutting radii from 8 µm to 35 µm were manufactured by honing and microblasting. It was found that the wear behavior of the cutting radii produced by honing was significantly improved compared to those produced by microblasting. In addition, an increase in the cutting radius resulted in an increase in the tool life.	[23]
Ti	Simulation, PVD	The deposition behavior of Ti films on a sharp tip was investigated by simulation for conventional and ionized PVD coatings. In conventional PVD, it was found that the morphology at the sharp tip was difficult to control, because the energy and flux of the sputtered particles could not be properly controlled. With ionized PVD, the morphology at the tip could be controlled. It was found that with a smaller edge angle, the deposition rate was lower but the uniformity of the layer was improved.	[24]
TiAlN based multi layers	FCVAD	There as the development of an approach for the design of multilayer composite coatings on a nanoscale for cutting tools. Various analyses were carried out to map the effects of the developed composite coatings on tool wear rates.	[25]
TiN	CAE	This was a characterization of the stress state in the edge area of a coated cutting tool made of a WC-Co carbide. The coating was carried out at a bias voltage of −100 V and a coating pressure of 4 Pa. At the flank face, an increase in compressive stress from the substrate surface to the coating surface from −1.2 GPa to −2.5 GPa could be seen, which was due to the development of microstructures during the coating process. At the cutting edge directly, the residual stresses were low but increased to −3 GPa towards the outer surface. The maximum coating thickness was also found in this area. At the rake face, scattering stresses of up to −3 GPa were observed. A hypothesis for this effect was a nanoporous microstructure.	[26]
TiAlN	CAE	This was an investigation of layer growth on sharp tool edges. The layer thickness was constant over most of the free surfaces but increased significantly on all analyzed edges (by a factor of 1.5 to 1.8). The increase in the layer thickness was successive and not abrupt. In the area where the two edges met, the effect of the layer thickness increase multiplied.	[27]

summarizes significant studies on edge-related effects, shows the most important results, and provides information on the material systems and deposition processes. Based on this literature review (Table 1) where only low bias voltages were utilized, a more general and systematic approach (e.g., higher bias voltages) for solving the problem of edge rounding and thus the edge radius design was chosen.

2. Materials and Methods

The nitride hard coatings were produced by means of CAE using an industrial MR 313 coating system from Metaplas Ionon, Hohenlockstedt, Germany (now Oerlikon Balzers Coating Germany GmbH), shown in Figure 1.

The coating system allows the simultaneous operation of two evaporators arranged at 90° with a magnetic field-controlled rectangular evaporator and a circular evaporator. Two different models of circular evaporators were used. The first one was a standard circular evaporator with a cathode diameter of 63 mm. In this device, the magnetic field is controlled by permanent magnetic disks placed behind the cathode. The second one was an APA (Advanced Plasma-Assisted) round evaporator. This was equipped with a 105 mm diameter cathode. Here, the arc spot behavior is influenced both by permanent magnets, which are arranged around the cathode, and by a solenoid electromagnetic coil behind the cathode. Both circular evaporators were used alternatively. All experiments were performed in a two-axis rotation of the substrates under a nitrogen atmosphere. The bias voltage was set to 0 V, −100 V, −300 V, −500 V, and −800 V. The coating pressures chosen were 1 Pa, 2 Pa, and 5 Pa (in a nitrogen atmosphere). The cathode materials listed in

Table 2. Utilized cathode materials.

Cathode Material	Cathode Dimension [mm]	Cathode Composition [at. %]
Ti	440 × 170 × 20	100
AlCrSi	Ø105 × 15	66/29/5
AlTi	Ø63 × 420	50/50
AlTi	Ø105 × 15	66/34
AlTi	Ø105 × 15	70/30
AlTi	Ø63 × 420	80/20
Cr	Ø63 × 420	100
Ti	Ø105 × 15	100

were produced by powder metallurgy, with the exception of the Ti rectangular cathode, which was produced by melting metallurgy (GfE Metalle und Materialien GmbH, Nuernberg, Germany was the supplier of the metals).

The substrates used consisted of cuboid samples (WC-Co 90 wt %–10 wt %, dimensions 20 mm × 12 mm × 2 mm), a testing cutting tool (WC-Co 90 wt %–10 wt %, Gesau GmbH, Glauchau, Germany) and sharp-edged samples (chromium steel, dimensions 20 mm × 12 mm × 2 mm) machined from so-called cutter knives. The surface of the cuboid samples was used for basic investigations such as the measurement of the coating thickness. Special attention was paid to the film growth at the edges. The coating of the cutter knife samples was used to investigate the film properties at various edge geometries.

In order to round the edges of the cuboid samples to a defined initial radius, they were manually machined with a corundum grinding stone. Initial radii of r_y = 10 μm, r_y = 15 μm, and r_y = 20 μm were aimed for. For this purpose, the edges of two samples per batch were rounded. The third edge remained unmachined. The results of a measurement of the rounded edges using a confocal microscope (see below for the type) are shown in

Table 3. Nominal versus measured radii determined by confocal microscopy.

Nominal Initial Radii ry	Measured Initial Radii rA
ry = 10 µm	rA,10µm = (15.6 ± 1.9) µm
ry = 15 µm	rA,15µm = (19.4 ± 4.1) µm
ry = 20 µm	rA,20µm = (25.3 ± 2.2) µm

.

It can be seen that the measured initial radii r_A are almost 5 µm larger than the targeted initial radii r_y. In addition, rounding was detected on all analyzed edges. The measurement of the edge radii was carried out by the company Zecha Hartmetall-Fabrikation GmbH, Königsbach-Stein, Germany, using the μsurf Expert confocal microscope from NanoFocus AG, Oberhausen, Germany. The radii were measured by placing semicircles tangentially touching the tips of the generated profiles using µsoft analysis imaging software (https://www.nanofocus.com/products/usoft/analysis/ (accessed on 4 June 2024)). The edge radii could then be determined on the basis of the created semicircles (see Figure 2).

Prior to the coating process, wet chemical cleaning was carried out in ultrasound-assisted baths containing cleaning solution Galvex 17.30 SUP, (Producer NGL Cleaning Technology SA, Nyon, Switzerland), and in deionized water for several minutes. The subsequent drying step was completed under a dry air flow. In the vacuum chamber, the samples were heated by electron bombardment and then subjected to a further cleaning step by Ti ion sputtering and Ar ion etching for 30 min. After cleaning, a TiN adhesion layer (working pressure: 0.8 Pa; evaporation current: 100 A; bias voltage: −100 V; and coating thickness: 200 nm) was deposited to ensure adhesion between the base material and the functional layer.

After deposition, the samples were removed from the vacuum chamber and analyzed using a Nikon H550S light microscope (producer Nikon Corporation, Tokyo, Japan). Particular attention was paid to the surface of the coating and the condition of the edges. A JEOL 6610 scanning electron microscope (producer JEOL Ltd., Tokyo, Japan) was used to analyze the surface morphology and layer structure in cross sections of the edge area. A KSG110 calotte grinder (Inovap HEF Group, Andrézieux-Bouthéon, France) was used to determine the coating thickness in accordance with DIN EN ISO 26423:2016-11 [28] and for stability testing of the coated sample edges (see Figure 8). A micro-scratch tester AE CPX NHT² MCT from Anton Paar Germany GmbH, Ostfildern, Germany, was also used for the edge stability testing. The indenter used was a sphero-conical scratch diamond with a 90° opening angle and a target radius of 50 µm, with a linear load progression of 5.0 to 10.0 N and a load progression rate of 10 N/min. The total scratch length was 0.5 mm with a maximum scratch width of approximately 30 μm. The scratch test was performed at a distance of approx. 50 µm from the edge. The measurement was carried out by Fritz Stepper GmbH & Co. KG, Pforzheim, Germany.

3. Results

3.1. Application of Coating Material to the Edges

In order to investigate the influence of the coating material on the sharpening effect, the rounded cuboid samples were coated with different layer systems at a coating pressure of p = 5 Pa and a bias voltage- of U_B = −500 V and U_B = −800 V and then analyzed with the optical microscope. The parameters were based on former investigations of the authors, where a sharpening on the edges only occurred using multilayer coating systems [21]. In particular, the coating properties in the edge region were observed. Coated, defective as well as coating-free areas were found at the edges (see Section 3.2).

Subsequently, the edges of the coated cuboid samples were photographed and precisely measured using a confocal microscope. Only those edge radii were measured that showed coating adhesion in the edge area. Subsequently, the relative radius change ∆r_C of the samples was determined (Equation (2)). This corresponds to the quotient of the absolute radius change ∆r and the initial radius r_A, multiplied by 100%, where the absolute radius change ∆r results from the difference between the coated radii r_C and the corresponding initial radii r_A listed in Table 3.

$$∆r_C=\frac{∆r}{r_A}\times 100\mathrm{\%}=\frac{r_C-r_A}{r_A}\times 100\mathrm{\%}$$

(2)

The following diagrams in Figure 3 and Figure 4 show the results of this test series.

For the monolayer coatings that were deposited with a bias voltage of U_B = −500 V (Figure 3a), it can be seen that only the AlTiN 50/50 and AlTiN 70/30 coating systems were able to form adherent coatings in the edge region. As a result, the relative radius changes were determined exclusively for these two systems. It can also be seen that edge sharpening occurs at initial radii of 15 and 20 µm. Contrary to the edges rounded to 25 µm, a radius increase of almost 30% can be observed. In the monolayer coatings deposited with U_B = −800 V (Figure 3b), edge sharpening was achieved everywhere. It can be seen that the edges rounded to 25 µm experienced a less pronounced sharpening compared to the other edges. For the multilayer coatings deposited with U_B = −500 V (Figure 4a), it can again be seen that each edge that has an adhesive strength in the edge region was sharpened by the coating. In particular, a significant sharpening can be observed for the coatings deposited with CrN/TiN, AlTiN/TiN 66/34, and AlTiN/TiN 70/30.

The multilayer coatings deposited with U_B = −800 V are shown in Figure 4b. Edge sharpening was also observed on all measured samples. In addition, it was found that the relative radius change increases with an increasing initial radius for the samples coated with AlTiN/TiN 70/30. In contrast, the relative change in the radius decreases with increasing initial radii for the cuboid specimens coated with AlTiN/TiN 80/20 and AlTiN/TiN 50/50.

Table 4. SEM side views of the cutting edge of cutting blades coated with TiN, AlTiN, and ATiN/TiN coatings without bias potential and with a bias potential of −800 V (see [29]).

UB = 0 V	UB = −800 V
Uncoated substrate (cutting blade)
TiN-coated cutting blades
AlTiN-coated cutting blades
AlTiN/TiN-coated cutting blades

compiles SEM side views of the cutting edges of cutter knives with TiN, AlTiN, and AlTiN/TiN coatings deposited without (U_B = 0 V) and with (U_B = −800 V) bias potential. In these images, one can see the effect of the applied bias potential, leading to a significant sharpening of the cutting edge area compared to the uncoated edge.

3.2. Layer Properties in Edge Area

In a further test series, the coating properties in the edge region were analyzed in detail. For this purpose, the cuboid specimens were coated with the TiN, AlTiN 70/30, and AlTiN 70/30/TiN coating systems. The coating pressures were chosen with p = 1, 2, and 5 Pa. Subsequently, the coating quality in the edge region was classified into three categories. This classification correlates with the zone model of Macak et al. [14,15,16]. An exemplary overview of the classifications per category is shown in

Table 5. Examples for the classification of the layer quality in the edge area according to the zone model of Macak et al. [13,14,15,16]; see also [29].

Coating Adhesion (Zone 1)	Coating Defects (Zone 2)	Coating-Free Edge (Zone 3)
AlTiN(70/30)/TiN; p = 5 Pa; and UB = −500 V	AlTiN(70/30)/TiN; p = 5 Pa; and UB = −300 V	TiN; p = 1 Pa; and UB = −500 V

. No correlation of the layer quality with the initial radius was found. The coating qualities listed in

Table 6. Quality of different coating materials near the edges.

Coating System	N2 Pressure [Pa]	UB = 0	UB = −100 V	UB = −300 V	UB = −500 V	UB = −800 V
TiN	1	coating adhesion	defects	coating adhesion	free edge	free edge
	2	coating adhesion	coating adhesion	coating adhesion	free edge	free edge
	5	coating adhesion	coating adhesion	coating adhesion	coating adhesion	coating adhesion
AlTiN	1	coating adhesion	defects	defects	coating adhesion	no coating
	2	coating adhesion	defects	defects	coating adhesion	coating adhesion
	5	coating adhesion	defects	defects	coating adhesion	coating adhesion
AlTiN/TiN	1	coating adhesion	defects	defects	coating adhesion	coating adhesion
	2	coating adhesion	defects	defects	coating adhesion	coating adhesion
	5	coating adhesion	defects	defects	coating adhesion	coating adhesion

are representative of the three initial radii.

The specimens characterized by an adhesive layer without defects at the edge can be seen in the left column. This corresponds to Zone 1 according to Macak et al. [13,14,15,16]. The second category is plotted in the middle column. The layer defects at the edges are characterized by spalling and chipping. This characteristic correlates with Zone 2. The third category of layer-free edges is shown in the right column. Its characteristic is the absence or partial presence of the deposited layer in the edge area, which reflects Zone 3 in the zone model. In the case of the AlTiN coating system, no layer could be produced on the specimen at an applied bias voltage of U_B = −800 V and a coating pressure of p = 1 Pa.

It can be seen that the edge area of the TiN-coated cuboid specimens (Table 6) remained uncoated at low coating pressures and high bias voltages. Otherwise, adhesion near the edges could be demonstrated on almost all other coatings produced with this coating system.

The edges of the samples coated with AlTiN and AlTiN/TiN (Table 6) show comparable properties. For both coating systems, this region is characterized almost entirely by a uniform coating adhesion for the samples coated with U_B = 0 V, U_B = −500 V, and U_B = −800 V. The coatings deposited at bias voltages of U_B = −100 V and U_B = −300 V are characterized by defects in the edge region for both coating systems.

Coating defects and coating-free edges could be attributed to ion etching, due to low coating pressures and high bias voltages [4,14,15,16], to a different phase formation, to residual stress [26], or to combined effects. Reduced adhesion forces, especially in the edge region, could also be responsible for damaged coating areas. Furthermore, high inherent stress of the coating in the edge area could cause spalling in this region.

3.3. Pressure Variation

Based on the quality classification, the edges of the coated cuboid specimens were measured again by confocal microscopy. Only those samples were analyzed that showed coating adhesion in the edge area. The relative radius changes were repeatedly determined via Equation (2). The results are shown in Figure 5, Figure 6 and Figure 7.

The starting point of this investigation is the TiN coating system (Figure 5). The relative radius changes are plotted in relation to the determined initial radii and the applied bias voltage. It can be seen that all TiN layers deposited at a bias voltage of U_B = 0 V lead to a rounding of the edge (represented by the positive relative radius changes). This is independent of the initial radius or the coating pressure. With the application of a bias voltage, edge sharpening could be achieved on the TiN-coated samples. The only exception is the sample with the edge rounded to 25.3 µm, which was coated at p = 2 Pa and U_B = −100 V. An edge radius increase of 88.9 % was observed with this sample.

Figure 6 lists the results of the profile section analyses of the AlTiN coatings. Once again, it can be seen that edge rounding was achieved at all three coating pressures for the coatings deposited with U_B = 0 V. The results of the profile section analyses are also shown in Figure 6. With the application of a bias voltage, a clear sharpening of the edges can be achieved. Only the samples with the initial radii of 19.4 µm and 25.3 µm, which were deposited with U_B = −500 V and a coating pressure of p = 5 Pa, also show rounded edges. The radius coated with a relative radius change of 225.8% deviates significantly from the others. This value is to be considered in more detail in further analyses and is therefore marked separately. This behavior is still unclear and needs further investigation.

Figure 7 shows the results of the cuboid specimens coated with the AlTiN/TiN multilayer system. With this coating system, it was possible to produce adhesive layers in the edge region at bias voltages of U_B = 0 V, U_B = −500 V, and U_B = −800 V. While the coated radii increase compared to the initial radii at a bias voltage of U_B = 0 V, the cutting edge radii are significantly reduced for the coatings deposited with an applied bias voltage. Only the coating produced at a coating pressure of p = 2 Pa and a bias voltage of U_B = −800 V leads to a radius increase of 8.3% at an initial radius r_A = 25.3 µm.

Based on the results, it can be seen that the application of a bias voltage is the decisive factor for the generation of sharp edges. In all the layers examined that were deposited without the application of a bias voltage, an increase in the radius of the edges was observed. This is in agreement with the results of Bohlmark et al. [20].

Varying the initial radius, the coating pressure and the coating system leads to different findings. If a radius increase occurs, the cuboid specimens rounded to 25.3 μm tend to possess about the largest coated radii. If a sharpening of the edge is achieved, particularly sharp edges are found in the cuboid specimens rounded to 15.6 μm. When considering the coating pressure variation, it can be seen that sharp-edged coatings could be produced at both low and high pressure. Likewise, coating deposition without applying a bias voltage at all the pressures used in this series of tests leads to edge rounding. In the evaluation of the coating systems, it was also possible to achieve both edge rounding and edge sharpening with the coatings used. Thus, the coating pressure and the coating system can only be assigned a subordinate role with respect to the generation of sharp edges.

3.4. Stability Testing of the Edges

Two different methods were used to test the coating stability in the edge area. The direct loading of the edge was carried out by means of a calotte grinding method. Within this series of tests, the edges of the coated cuboid specimens were subjected to the force of a rotating carbide ball (WC-Co, 90 wt%–10 wt%). The abrasive loading as a result of the rotational movement leads to a coating removal in the area of the edge.

During loading, the cuboid specimens were positioned in such a way that the ball rotated transversely to the edge on the one hand and longitudinally to the edge on the other (Figure 8). The ball rotation was performed at a speed of 500 rpm and a specimen angle of 80°, as well as a total of 1500 rotations per stability test. To ensure uniform removal of the ground-out particles, the stability test was carried out with ethanol added.

Subsequent evaluation of the damage pattern using optical microscopy allowed a qualitative statement to be made on the edge stability of the coating. Cuboid specimens coated with p = 5 Pa and bias voltages of U_B = 0 V, U_B = −500 V, and U_B = −800 V were tested. The damage patterns of the stressed edge areas were divided into two categories (

Table 7. Categorization of the damage recordings after the stability test; see [29].

Layer Adhesion	Layer Delamination
TiN, UB = −500 V, and transverse grinding	AlTiN/TiN, UB = −500 V, and longitudinal grinding

): first, the uniform abrasion of the coating without chipping, the coating adhesion; and second, the failure of the coating, the coating delamination, as a result of the abrasive load.

Subsequently, the samples tested for longitudinal sectioning (

Table 8. Condition of the coatings after edge stability testing in longitudinal grinding.

	UB = 0	UB = −500 V	UB = −800 V
TiN	coating adhesion	coating adhesion	coating adhesion
AlTiN	coating adhesion	delamination	coating adhesion
AlTiN/TiN	coating adhesion	delamination	coating adhesion

) and transverse sectioning (

Table 9. Condition of the coatings after edge stability testing in transverse grinding.

	UB = 0	UB = −500 V	UB = −800 V
TiN	coating adhesion	coating adhesion	coating adhesion
AlTiN	delamination	delamination	coating adhesion
AlTiN/TiN	coating adhesion	delamination	coating adhesion

) were categorized.

As a result of this series of tests, it was found that both longitudinal grinding and transverse grinding could be carried out on the edges of the coated cuboid samples by means of calotte grinding. In addition, it was possible to demonstrate different levels of stability in response to abrasive stress on the coatings investigated. The edges coated with TiN and the coatings deposited at U_B = −800 V exhibited the best edge stability against the abrasive load.

In a second series of tests, so-called micro-scratch tests were carried out in the edge area. For this purpose, the edges of the cuboid specimens were tested, which were coated with AlTiN 50/50 and AlTiN/TiN 50/50 at bias voltages of U_B = −500 V and U_B = −800 V. The SEM images of the micro-scratch tests are listed in

Table 10. SEM images of the micro-scratch tests near to the edge of samples with initial radii of r_A = 15.6 µm and r_A = 19.4 µm.

	rA = 15.6 µm	rA = 19.4 µm
AlTiN UB = −500 V
AlTiN UB = −800 V
AlTiN/TiN UB = −500 V
AlTiN/TiN UB = −800 V

.

The scratch marks on the cuboid specimen coated with AlTiN and a bias voltage of U_B = −500 V show no damage over the entire length. Only at the end of the track are chippings noticeable, which are slightly more pronounced on the edge rounded to r_A = 19.4 µm. The behavior of the specimen coated with AlTiN and a bias voltage of −800 V is clearly different from that of the other scratched tracks. The side with the initial radius r_A = 15.6 µm exhibits significantly stronger damage in the form of semi-circular chipping. This damage pattern can already be seen after one-third of the load and continues to the end of the scratch, indicating a critical load of about 6.8 N. In this area, it can be clearly seen that the substrate has been uncovered. The semi-circular arcs can be attributed to compression of the layer by the diamond tip. The side with the initial radius of r_A = 19.4 µm shows a local adhesive failure pattern after about half of the scratch length, which does not cover the residual scratch compared to the r_A = 15.6 µm sample. The corresponding critical load is 7.0 N and comparable to the critical load of the r_A = 15.6 µm sample. Here, most of the breakout area can be seen in the direction of the edge.

The scratch marks of the specimen coated with AlTiN/TiN at U_B = −500 V show damage patterns like those of the specimens coated with AlTiN at U_B = −500 V. Once again, chipping can only be observed in the area of maximum load (10 N) at the end of the track (r_A = 19.4 µm). The cuboid specimens coated with AlTiN/TiN at U_B = −800 V show different damage patterns. The scratch mark in the edge area of the specimen side, which is rounded to r_A = 15.6 µm, exhibits a first failure event with a critical load of around 7.9 N, and an adhesive breakout shortly before the end of the scratch. The brighter substrate is clearly visible. On the specimen side with the initial radius of r_A = 19.4 µm, several smaller breakouts can be observed over the entire scratch length. The corresponding critical load is estimated to be around 9.4 N. The brighter area within the scratch is not considered a critical defect; therefore, further investigation is required (marked with an arrow). Compared to the AlTiN monolayer, the critical loads are higher on the AlTiN/TiN multilayer coatings (Table 10). This effect may be due to a better stability of the multilayered coatings and the ability to stop growing cracks, which may be caused by the indenter movement. Based on the scratch test results, the coatings deposited at U_B = −500 V exhibit the best coating stability within this test series. In a comparison of the two initial radii, the edge areas rounded to r_A = 19.4 µm have fewer damaged scratch marks.

4. Summary and Conclusions

A technological approach for the adjustment of the final radius of coated edges by film growth is introduced in this work. A wide range of relevant parameters—the initial edge radius of the coated samples, the coating material, and the bias voltage and gas pressure during the coating process—are evaluated. It is shown that the bias potential in particular has an important impact on the formation of the radius of the coated edges. A massive edge rounding (an increased edge radius) by the grown coating is typically observed at a 0 V bias potential. A medium bias range (−100 … −300 V) yields more film defects and delamination. At a higher bias potential (−500 … −800 V), a sharpening effect (a decreased edge radius) can be shown. Furthermore, the stability of the grown coatings in the edge zone are the subject of investigation. The edge stability is investigated by means of calotte grinding tests and micro-scratching. Significant differences between the types of coatings are found. It can be shown that stable sharp edges can be produced by this coating approach.

As a possible sharpening mechanism, it is assumed that the bias voltage changes the incidence of the ions from the plasma and forms the edge geometry during the deposition process. This could be used to design the edge geometry using different bias voltages or different material systems. As the results have shown, the initial radius is of special importance to the results, as well as the coating thickness.

Nevertheless, there are a couple of open questions. A possible change in the element concentration in the coating near the edge has not yet been investigated. Furthermore, the limits of the sharpening effect (the minimum/ maximum film thickness, sample size, geometry, etc.) should be the subject of future investigation combined with material cutting tests.