HPPMS tool coatings: Chip formation and friction

HPPMS‐Werkzeugbeschichtungen: Spanbildung und Reibung


Introduction
During the machining of steel, mechanical and thermal loads occur. Adhesive and abrasive wear as well as oxidation appear as damage mechanisms of the tools and lead to a reduction of tool performance and lifetime. The energy required to overcome the friction between tool surface and chip as well as the workpiece during machining is almost completely converted into heat and increases the temperature of the tool surface, as described in the fundamental principles of machining [1]. Thus, as the coefficient of friction increases, the surface temperature of the cutting tool increases. The higher temperature facilitates diffusion and oxidation processes and reduces hardness and resistance against abrasive wear. These cause changes in elastic-plastic properties and microstructure and damage the tool surface during the machining process. A reduced friction is therefore advantageous for machining processes. The thermal and mechanical loads on the cutting edge depend on the coefficient of friction and the contact length between rake face and chip.
Thin hard coatings deposited by physical vapor deposition (PVD) for wear protection of cutting tools are state of the art. The low-pressure plasma during PVD allows the synthesis of metastable nitrides like TiAlN, CrAlN, and TiCrAlN. Aluminum and chromium increase the thermal resistance of the hard coatings. This is caused by the formation of a dense, firmly adhering passivating oxide layer of chromium oxide and aluminum oxide on the surface, which inhibits the diffusion of oxygen into the coating [2,3]. In the coating system CrAlN+Mo:S, molybdenum and sulfur are incorporated into a CrAlN coating. Out of these, the self-lubricating transition metal dichalcogenide molybdenum disulfide (MoS 2 ) is formed under tribological load. The

Indexable insert (left) used for planing studies and coating of indexable inserts in the industrial coating unit CC800/9 HPPMS (right).
layers in the crystallographic structure of the transition metal dichalcogenide are bonded by relatively weak van der Waals forces. This leads to a low shear strength and to a self-lubricating effect [4]. As a result, the adhesion properties against steel in technical applications such as dry cold bulk metal forming can be improved. Moreover, the coatings show a high abrasion resistance [5]. Previous work has shown that the friction against steel can be reduced for CrAlON compared to CrAlN coating in tribologicaly tests [6]. Moreover, the resistance of coated carbide tools to damage caused by adhesive and abrasive wear during machining can be improved by increasing the oxygen content in the CrAlON ZUSAMMENFASSUNG coatings [7]. A reduction of friction and adhesive wear might be attributed to the changed chemical bond character. The addition of silicon to the quaternary system of TiAlCrN leads to the formation of an amorphous silicon nitride matrix. This matrix interrupts grain growth during the deposition of the coating and leads to a nanocrystalline microstructure with enhanced properties, such as an improved indentation hardness of TiAlCrSiN compared to TiAlCrN [8,9]. Moreover, the coefficient of friction can be reduced and tool life during cutting can be improved [10].

HPPMS-Werkzeugbeschichtungen: Spanbildung und Reibung
Hard coatings can be deposited by magnetron sputtering. High power pulsed magnetron sputtering (HPPMS) has technical advantages compared to direct current magnetron sputtering (dcMS), such as a more homogeneous coating thickness distribution on flank and rake face [11]. DcMS, in contrast, shows higher deposition rates and therefore has a higher economic efficiency. Both processes can be combined in a hybrid process, taking benefit of the advantages of both processes.
The mean coefficient of friction in the contact area between tool and workpiece during machining with geometrically defined cutting edges in continuous cutting can be calculated from cutting and feed force. The measurement of the forces is usually performed by piezoelectric force transducers [1]. The forces are composed of the normal force on the rake face F Nγ and the tangential force on the rake face F Tγ . Due to tool wear as well as elastic and plastic deformation of the tool the normal force on the rake face F Nα and the tangential force on the rake face F Tα additionally occur [12]. Thus, the coefficient of friction includes therefore both, the friction, which occurs on the rake face and the friction, which occurs on the flank face. In order to determine the mechanical stress distribution on the cutting edge, it is necessary to measure the contact area between tool surface and chip. The width of the contact area corresponds to the chip width. The contact length in the direction of chip flow, on the other hand, is primarily determined by the material properties, the tool geometry, the process parameters and the friction between tool and work-  piece. The contact length on rake face and flank face affects the distribution of the normal and tangential stress distribution on the cutting edge [13]. The normal and tangential stress distribution on the cutting edge influences the effective stress distribution inside the cutting edge and are therefore of great interest for the design of cutting tools. The contact length is also one of the essential boundary conditions when calculating the stress distribution using FEM, based on the load stresses and temperatures in the contact zone.
In order to measure the contact length, the chip thickness and the coefficient of friction simultaneously during continuous cutting within a single experiment, a special experimental setup was developed. For this purpose, a planing test rig was used, which is equipped with a simultaneous measuring system for the analysis of process forces and chip formation on the basis of high-speed recordings. In preliminary studies in [14], the influence of the coating on the contact length, the chip thickness and the coefficient of friction during planing was already investigated at a specific cutting speed. The current study aims to additionally investigate the influence of the cutting speed.

Materials and experimental details
The coatings were deposited by a hybrid process consisting of dcMS (direct current magnetron sputtering) and HPPMS (high power pulsed magnetron sputtering) using the industrial coating unit CC800/9 HPPMS, CemeCon AG, Würselen, Germany. The coating unit is equipped with four dcMS cathodes and two HPPMS cathodes. Fig. 1 a) shows the target configuration for the deposition of the coating system CrAlN+Mo:S, which is schematically depicted in Fig. 1 b) in nitride CrAlN bond coat were deposited using the cathodes HPPMS1, HPPMS2, dcMS-2, dcMS-3 and dcMS-4. These targets were equipped with chromium targets plugged with 20 aluminum plugs. The CrAlN+Mo:S toplayer was deposited additionally using the cathode dcMS1, which was equipped with a molybdenum disulfide target. Since this toplayer is in direct contact with the workpiece during cutting, the entire coating system is designated shortly by the material system of the toplayer CrAlN+Mo:S. Fig. 1 c) shows the target configuration for the deposition of the coating system CrAlON, which is schematically depicted in Fig. 1d). It consists of a thin metallic Cr and a thin nitride CrN bond coat. Those were deposited using the cathodes HPPMS2 and dcMS-3, which were equipped with chromium targets. A CrN/AlN nanolayer was subsequently deposited additionally using the cathodes dcMS1 and dcMS4, which were equipped with aluminum targets. Finally, the CrAlON toplayer was deposited only using the cathodes HPPMS1 and dcMS2, which were equipped with chromium targets plugged with 20 aluminum plugs. Since the oxynitride toplayer is in direct contact with the workpiece during cutting, the coating system it is designated shortly as CrAlON.

Fig. 1 e)
shows the target configuration for the deposition of the coating system TiAlCrSiN, which is schematically depicted in Fig. 1 f). Initially, a HPPMS TiAlN interlayer was deposited using the cathodes HPPMS1 and HPPMS2, which were equipped with titanium targets plugged with 48 aluminum plugs. Subsequently, the TiAlCrSiN toplayer was deposited additionally using the cathodes dcMS-1, equipped with a titanium target plugged with 17 silicon plugs, dcMS2, equipped with an aluminum target plugged with 20 chromium plugs, dcMS-3, equipped with a titanium target and dcMS-4, equipped with a titanium target plugged with 34 aluminum plugs as well as 17 silicon plugs. Since this toplayer is in direct contact with the workpiece during cutting, the coating system it is designated shortly as TiAlCrSiN.
In order to control the oxygen content and the microstructure of CrAlON, the total pressure, the reactive gas flows and the bias voltage were varied for the different layers, see Table 1. For TiAlCrSiN, two different pressures and nitrogen flows were selected in order to achieve a suitable working point for the deposition of both, the interlayer and the toplayer.
The roughness values Ra of the coatings were measured using the confocal laser scanning microscopy VKX 210, Keyence Corporation, Osaka, Japan. The morphology of the coatings was investigated using scanning electron microscope cross-section (SEM) images. A SEM ZEISS DSM 982 Gemini, Jena, Germany, was used for this purpose. The chemical composition of the coatings was measured by electron probe microanalysis (EPMA) using a JEOL JXA8530, Jeol, Tokyo, Japan. The SEM and EPMA investigations were carried out by the Central Facility of Electron Microscopy of the RWTH Aachen University. Nanoindentation measurements without any further post-treatment were performed after deposition. The indentation hardness H IT and the indentation  The influence of the coatings and of the cutting speed on the contact length between chip and rake face, on the chip thickness and on the coefficient of friction were investigated during a planing process. Coated indexable inserts made of cemented carbide were used as tools. They had a sharp cutting edge radius of r ß < 5 µm. 42CrMo4+QT was used as workpiece material. The planing tests were carried out using a uncut chip thickness of h = 0.1 mm, a width of cut of b = 3 mm, a rake angle of γ = 6 °, a clearance angle of α = 6 ° and the three different cutting speeds of v c = 50 m/min, v c = 100 m/min as well as v c = 150 m/ min. The cutting speed of v c = 150 m/ min is a common cutting speed for machining of 42CrMo4+QT in industrial applications. The cutting length of one planing operation is L = 120 mm. The contact length at the rake face, the chip thickness and the coefficient of friction during planing were measured simultaneously and averaged from two planing operations using a specially developed experimental setup, see Fig. 2. The tool is fixated on a dynamometer during the continuous cutting operation. The linear guide system Simodrive 1FN1, Siemens, Berlin, Germany, generates the relative movement of the workpiece to the coated tool.
In order to measure the feed force F f and the cutting force F c, a dynamometer Kistler 9257B, Kistler Instrumente GmbH, Sindelfingen, Germany, was used. The coefficient of friction was calculated from those forces according to the force diagram proposed by Merchant [16]. Assuming an ideally sharp cutting edge and neglecting flank face wear, the coefficient of friction at the rake face can be calculated by Eq.1.
The high-speed camera Fastcam SA5, Photron Deutschland GmbH, Reutlingen, Germany, was used to measure the contact length between chip and rake face of the coated tool as well as the chip thickness. The cold light projector Techno light 270, Karl Storz GmbH & Co. KG, Tuttlingen, Germany, was additionally used, in order to avoid motion blur and to ensure enough brightness. To ensure a sufficient temporal resolution, the frame rate was chosen depending on the cutting speed, see Table 2.
The high-speed images were taken at a resolution of A = 512 x 512 pixels with V = 6x magnification. This results in a geometric resolution of R = 3.3 µm per pixel. To ensure plain strain conditions during machining, the tool and the workpiece were mounted against sapphire glass. Immersion oil was applied in the contact zone between tool and glass to reduce the friction between glass and tool as well as to achieve the best optical conditions.
After performing the planing tests, the coated tools were investigated using SEM. Furthermore, the chemical composition after the cutting tests was measured using energy dispersive X-ray spectroscopy (EDS). A Phenom XL, Thermo Fisher Scientific Inc., Waltham, Massachusetts, USA, was used for both the SEM and EDS analysis.

Coating characterization
Three different coating systems were investigated in the scope of this article.  Table 3 shows the coating properties.
Since the toplayer of the coating systems is in direct contact with the work piece during cutting, Table 3 only shows the chemical composition of    The toplayer showed no columns due to its even more fine crystallinity or high amorphous share caused by the high amount of oxygen. TiAlCrSiN consisted of a TiAlN bond coat and a TiAlCrSiN toplayer, see Fig. 3 c). The coating thickness was s = 2.0 µm in total. The coating showed an average line roughness of Ra = 0.09 µm and a dense morphology. There were no columns visible. The fine crystalline morphology was caused by the nanocomposite coating architecture as proven in earlier publications [17,18]. The coating showed a slightly increased indentation hardness of H IT = 23.2 ± 2.2 GPa and a slightly decreased indentation modulus of E IT = 291.5 ± 31.3 GPa compared to the CrAlN+Mo:S and CrAlON coatings.

Analysis of chip formation and friction
In order to investigate process forces and chip formation during cutting simultaneously, a special experimental setup was developed, as described in section 2. Fig. 4 shows images recorded by the high-speed camera of the CrAl-N+Mo:S coated tool during planing with v c = 100 m/min at a cutting length of l = 10 mm, l = 30 mm and l = 80 mm. At the beginning of the planing operation, at a cutting length of l = 0 mm, the chip forms out and the contact length as well as the chip thickness increase from l c = 0 mm and h' = 0 mm on.
The high-speed camera enables the measurement of the contact length and the chip thickness depending on the cutting length, see Fig. 5 a). After reaching a critical cutting length of approx. l = 100 mm, the chip hits the edge of the workpiece. The results after reaching this critical cutting length are therefore not considered for further analysis. The cutting force F c and the feed force F f were measured simultaneously depending on the cutting length using a dynamometer. The coefficient of friction µ was calculated from those forces according to the force diagram proposed by Merchant [16]. Fig. 5 b) shows the cutting force F c , feed force F f and coefficient of friction µ depending on the cutting length l.

ViP 31
The planing tests were conducted using differently coated tools at different cutting speeds. Fig. 6 shows images from the high-speed camera of the CrAlN+Mo:S (a), CrAlON (b) and TiAl-CrSiN (c) coated tools during planing with v = 100 m/min at a cutting length of l = 80 mm.
The CrAlN+Mo:S coated tool showed the highest contact length of l c = 1.01 mm and the highest chip thickness of h' = 0.29 mm. For the CrAlON coated tool contact length of l c = 0.71 mm and chip thickness of h' = 0.21 mm were the lowest.  Fig. 7 shows the contact length l c and coefficient of friction µ of the coated tools during planing at different cutting speeds v depending on the cutting length l. The results reveal that the contact length and the coefficient of friction increases during the planing operation over the cutting length. This was most pronounced for the CrAlN+Mo:S coated tool and less pronounced for the CrAlON coated tool. The contact lengt hc correlated with the coefficient of friction. The chip thickness did not depend on the cutting length, which is why it is only exemplarily shown in Fig. 5 a) for the CrAlN+Mo:S coated tool at v = 100 m/min. Fig. 8 shows the contact length l c (a), chip thickness h' (b) and coefficient of friction µ (c) averaged over the cutting length for two planing operations depending on coating and cutting speed. It was not possible to measure the contact length of the second planing operation at v = 100 m/min and at v = 150 m/min. This was due to incorrect measurements, which are e. g. caused by the chip rolling up or twisting off during the planing process.
For the CrAlON and TiAlCrSiN coated tools the contact length decreased, when the cutting speed increased. For the CrAlN+Mo:S coated tool, the con-tact length at v = 100 m/min was the lowest. Furthermore, the contact length of the CrAlN+Mo:S coated tool was higher compared to the CrAlON and TiAlCrSiN coated tools at v = 50 m/min, v = 100 m/min and v = 150 m/min. The CrAlON coated tool showed similar contact lengths compared to the TiAlCrSiN coated tool at v = 50 m/min. However, at v = 100 m/min and v = 150 m/min, the contact length of the CrAlON coated tool was significantly reduced.
An increased contact length is generally accompanied by an increase in thermal tool load due to the increased friction distance, since the friction power is largely converted into heat. In addition, as calculated in [13], the tool is subject to greater mechanical stress at higher contact lengths, especially in feed direction. A reduced contact length during cutting is therefore advantageous.
The chip thickness h' of the coated tools was similar at v = 50 m/min, see Fig. 8    coated tool was lower compared to the CrAlN+Mo:S and TiAlCrSiN coated tools. The difference between the two measurements of the CrAlN+Mo:S coated tools was relatively high at v = 100 m/min and v = 150 m/min. Therefore, no distinct statement can be made regarding the ratio of the chip thickness between the CrAlN+Mo:S and TiAlCrSiN coated tools. The coatings had a significant influence on the coefficient of friction, see Fig. 8 c). The difference between the two measurements of the coefficient of fricition were relatively low, especially compared to the measurements of the chip thickness. The coefficient of friction was higher for the CrAlN+Mo:S and TiAl-CrSiN coated tool at v c = 50 m/min compared to v c = 100 m/min. For the CrAlON coated tool it was similar at both cutting speeds. At v c = 150 m/min, the coated tools showed a reduced coefficient of friction compared to v c = 100 m/min. Overall, the CrAlON coated tool showed the lowest coefficient of friction followed by the TiAlCrSiN coated tool. The CrAlN+Mo:S coated tool showed the highest coefficient of friction. An increased friction decreased the shear angle. According to [16], the higher shear angle increased the chip thickness. The higher coefficient of friction of the TiAlCrSiN and CrAlN+Mo:S coated tools compared to the CrAlON coated tool might therefore cause the higher chip thickness of those coated tools.

Analysis of the coated tools after performing the planing tests
After performing the planing tests, the tools were investigated by SEM and EDS. Independently of the cutting speed, the coated tools showed similar damage characteristics. Fig. 9 exemplarily shows the rake face of the coated tools after performing a planing operation at v = 150 m/min.
There are darker and brighter areas near the cutting edge of the CrAl-N+Mo:S coated tool. The TiAlCrSiN coated tool showed brighter areas and the surface of the CrAlON coated tool remains unchanged compared to prior the planing test. Table 4 shows the chemical composition at the marked areas in Fig. 9 measured by EDS. The EDS measurement reveal that there was iron adherent on the CrAlN+Mo:S and TiAlCrSiN coated samples in both, the brighter and darker areas. Since light elements like carbon, oxygen and nitrogen cannot be quantified reliably by the EDS detector used for the investigation, Table 4 does not show the content of those elements. However, due to the material contrast caused by the back scattered electrons (BSE), it can be assumed that the darker areas show a higher carbon or oxygen content compared to the brighter areas. Therefore, they might contain more carbides or oxides. Due to the higher density of iron compared to the elements containing the coating material, it appeared brighter than the coating material. Only small amounts of the substrate material tungsten and cobalt were detected by EDS, indicating that no substrate is exposed due to abrasive wear. There were no steel adhesions nor abrasive wear observed for the CrAlON coated tool. The oxinitride CrAlON coating therefore seemed to show a lower adhe-