ICMCTF 2025 Session PP1-2-MoA: PVD Coatings and Technologies II

High power impulse magnetron sputtering is an ionized physical vapor deposition technique in which the sputtered metal flux from the target is partially ionized. This enhances film properties like density and adhesion. At the same time, some of the produced metal ions are back-attracted to the target and therefore lost from the deposition process. We show that this loss in deposition rate is largely dependent on the sputter yield of the target material. Here, two extremes can be distinguished: 1) Discharges with low sputter yield targets are dominated by argon, and 2) discharges with high sputter yield targets are metal-rich. In the first case, the electron temperature must be significantly higher to enable sufficient ionization of predominantly the working gas to generate the experimentally observed high discharge currents. In those discharges we find strong electron energization by Ohmic heating in the ionization region extending beyond the cathode sheath. In the second case, we find that Ohmic heating is considerably weaker compared to the low sputter yield discharges. At the same time, frequent collisions with metal atom cool the electron population, which leads to a decrease in electron temperature. By examining a range of different target materials using the Ionization Region Model (IRM) we find a consistent trend of decreasing back-attraction probability and electron temperature as the sputter yield of the target material increases. A lower electron temperature increases the mean free path of ionization of sputtered species, shifting the average location of ionization away from the target. The much weaker electric fields at those locations compared to the target vicinity, ultimately facilitates ion escape toward the substrate, which thus explains the observed reduction in back-attraction.

Metallic substrates can be treated by a combination of nitriding and subsequent deposition of a physical vapor deposition (PVD) coating to improve coating adhesion, wear/abrasion resistance, and corrosion resistance. The combination of the two processes is known as duplex treatment. In general, conventional nitriding treatment and PVD coating deposition are typically performed as two separate processes in distinct facilities and environments. Consequently, the lead time and production cost are not optimized. We present a duplex coating process by integrating plasma nitriding and magnetron sputtering using hot filament assisted and plasma enhanced magnetron sputtering (PEMS) within the same facility. In the process, a global nitrogen plasma is generated by impact ionization from electrons emitted from the hot filaments and attracted towards the substrate surface for nitriding. In this study, the effects of the PEMS process on the structure and properties of the nitrided stainless steel have been studied. The PEMS treated stainless steel exhibited greatly improved surface hardness, wear resistance, and hydrophobicity in oil formula. In addition, duplex TiSiCN and DLC coatings deposited using the integrated process showed improved mechanical properties and adhesion as compared to the coatings deposited without the duplex treatment.

Depositing Bismuth-based thin films by the sputtering technique often results in a non-stoichiometric excess of Bi across various materials, including the ferroelectric piezoelectric Bi_0.5Na_0.5TiO₃. This phenomenon is attributed to the lower scattering of heavier sputtered species in the plasma phase. Common mitigation strategies include multi-target sputtering to control Bi flux and promoting Bi re-evaporation at elevated growth temperatures by exploiting its temperature-sensitive sticking coefficient. Herein, we systematically investigate this issue, focusing on BNT thin film deposition without in-situ substrate heating and using a mixed-powder target by single-cathode RF Magnetron sputtering in Ar plasma. Compositional analysis of the films via EDX and RBS reveals a 25-30% excess of Bi by sputtering a stoichiometric Bi_0.5Na_0.5TiO₃ (BNT50/50) target. Simulations indicate a relatively unhindered transfer of Bi towards the substrate while other species are impeded by the background gas, as shown by the target sputtering combined with species transport using TRIM and SiMTRA codes, respectively. Reducing the sputtering yield of Bi by adjusting the target composition to Bi_0.35Na_0.5TiO_2.8 (BNT35/50) eliminates the excess Bi from the BNT films. This study provides a clear insight into the origin of bismuth excess and a route map for its regulation inside the Bi-based thin films deposited via the sputtering technique.

Keywords: “Bi_0.5Na_0.5TiO₃”, “Thin Films”, “Bi Excess”, “Magnetron Sputtering”, “Powder Targets”, “Ar Plasma”.

In this work, aluminum (Al) thin films were deposited onto glass substrates using a two-plasma system. A standard 2” MAK DC magnetron sputtering and an additional water-cooled anodic plasma. The second anodic plasma was generated using a graphite electrode placed approximately 5 cm from the magnetron. We found that the anodic plasma generated additional argon ions, which were incorporated in the magnetron discharge. These additional ions changed the characteristic of the magnetron discharge and the deposition of the aluminum atoms. Here we report the dependence of the changes as a function of the position of the anode and the temperature of the graphite electrode. Similarly, we report the changes in the properties of the films, hardness, and wear resistance as a function of the experimental parameters, gas pressure, MS voltage, and the voltage applied to the graphite anode, as well as the deposition rates measured by optical profilometry.

Silicon nitride films have drawn increasing attention due to their high demand in various engineering applications. Considering many disadvantages of chemical vapor deposition (CVD) such as high process temperature and hydrogen contamination, reactive RF diode sputtering is an alternative method to produce high quality SiN_x films. In this work, we report synthesis of SiNx films at low temperature by reactive RF diode sputtering.

Hard physical vapor deposition (PVD) coatings are widely used as protective layers on cemented carbide tools due to their exceptional mechanical properties. However, these coatings can be susceptible to damage and cracking. Gaining a deeper understanding of how the coating microstructure influences the cracking behavior is essential. Moreover, most tool wear prediction does not include the effect of the cracking behavior. Thus, crack initiation and propagation under external loads and its contribution to tool wear should be investigated. A precise micromechanical simulation of cracks could enhance the accuracy of tool wear simulations in cutting applications.

This study combines experimental and mesoscale simulation to investigate the cracking behavior of a TiAlCrN PVD coated cemented carbide tool. Initially, nanoindentations coupled with inverse FEM simulations are conducted to determine mechanical properties of coatings, specifically Young’s modulus and parameters for the Ludwik-Hollomon model. These properties are then applied to simulate high-load nanoindentation at a macroscopic scale. Subsequently, scanning electron microscopy (SEM) is applied to characterize the grain morphology. Using this data, a representative finite element model is developed. Numerical simulations of the local crack initiation and growth are performed based on the implemented model in combination with the extended finite element method (XFEM). SEM micrographs taken after indentation are analyzed to study crack behavior, enabling a correlation between experimental results and numerical simulations.

The combined experimental and detailed numerical modelling approach facilitates insights into how microstructural parameters including grain size and orientation influence crack growth in the coating system. This study presents a combined analysis using experimental nanoindentation and a mesoscopic simulation model of the nanoindentation to investigate the crack growth in PVD coated cemented carbide. The correlation of experimental and simulative results allows a detailed study of the interaction of microstructure and crack growth in PVD coatings. Models comparable to the one here presented may be used in future work for optimization of coated cemented carbide tools.

The wear resistance of physical vapor deposition (PVD) coatings is heavily influenced by their elastic and plastic properties. These properties serve as essential inputs for finite element method (FEM) simulations to predict tool wear, including the elastic modulus for the characterization of elastic properties and parameters of the Johnson-Cook model for the description of the plastic behavior. A precise determination of these parameters is required for simulation of tool wear. In this study, machine learning models are developed to directly map load-depth curves from nanoindentations on TiAlSiN and TiAlCrN coatings to parameters of coatings in FEM. An FEM simulation model of nanoindentation is employed to generate a dataset comprising load-depth curves resulting from a wide range of input material properties. Several machine learning models including support vector regression (SVR), multilayer perceptron (MLP), long short-term memory (LSTM) and gated recurrent unit (GRU) are then trained, validated, and compared using this dataset. The input to these models consists of simulated load-depth curves, with the target being parameters required for the definition of the material in commercial FEM softwares. Among these machine learning models, GRU achieves the best prediction performance. Ultimately, GRU is used to predict material parameters of TiAlSiN and TiAlCrN coatings based on experimental load-depth curves. FEM simulations using the GRU-predicted material parameters show excellent alignment with experimental measurements, achieving accurate results in a single iteration without further parameter adjustments. The determined parameters can be directly used as reasonable inputs for further FEM simulations as parts of a Greybox model to predict tool wear during cutting.