Multivariate Data Analysis on (Ti,Al)N Arc-PVD coating process : MVDA of the growth parameters thickness, stress,composition, and cutting performance

Abstract: This diploma work was done at Seco Tools AB (SECO) in Fagersta and aimed to evaluate the possibility to model the relationship between deposition data, deposition properties and, cutting performance of a (Ti,Al)N coating on cutting inserts by applying the Multivariate Data Analysis (MVDA) modeling technique Partial Least Squares Projection to Latent Structures Modeling (PLS). Cathodic Arc Deposition (Arc-PVD) was the PVD technique focused on this study. The deposition technique that was focused on in this study was Cathodic Arc Deposition (Arc-PVD). For this purpose, two series of Arc-PVD coatings were manufactured. The first series aimed to generate a supervised explorative model for the deposition process. The second manufactured series was aimed to generate a batch-to-batch variation model of a deposition process. In the first supervised explorative model, the deposition parameters were set by a Design of Experiment (DOE) setup using a quarter factorial design with resolution III. In the second batch-to-batch model, the non-fixed deposition parameters and the cathode wear were monitored, and all other parameters were kept the same for every run. The results demonstrate good possibilities to model Arc-PVD coating properties and its performance in metal cutting with respect to the applied deposition parameters. The supervised explorative model confirmed previously established relationships, while the batch-to-batch model shows that variations between batches could be related to the wear of the cathode. This wear was shown to have a negative influence on the properties of the deposited coating.