Effect of coating thickness on the tool wear performance of low stress TiAlN PVD coating during turning of compacted graphite iron (CGI)

Abstract

Compacted graphite iron (CGI), with its superior mechanical properties, is a promising candidate to replace grey cast iron in the automotive industry. However, the low machinability of CGI compared to grey cast iron, has made this transition difficult. Built-up edge formation, combined with abrasive and adhesive wear is the main problem of CGI turning at moderate cutting speeds. In this study, a low compressive residual stress PVD coating was developed using newly introduced super fine cathode (SFC) technology. The main advantage of low compressive residual stress SFC coating, is the possibility of increasing its thickness compared to the commercial range of arc coatings (usually with a thickness within 1–5 µm) without any process induced spallation of the coating layer. Therefore, three different low compressive residual stress Ti40Al60N coatings with thicknesses of around 5, 11 and 17 µm were deposited. The coatings were characterized by X-Ray diffraction, scratch test, ball crater test and nanoindentation. Furthermore, the cutting performance of the coated inserts was investigated in finish turning of CGI. An improvement of around 35% tool life is achieved for the cutting tool with the low residual stress thick coating compared to the commercial benchmark. Progression of flank wear was studied by means of TEM, SEM-EDS, optical microscopy and 3D wear measurement. Chip undersurface morphology as well as cross-sectional studies of the chip structure were performed. This was combined with analysis of the workpiece surface. Evaluation of the results obtained would help to achieve a better understanding of the wear mechanism and built-up edge formation of the studied coatings. According to the recorded data, coating thickness significantly affects cutting tool wear behavior and the mechanical properties of the coatings. A certain thickness range, specifically within 10 µm was found to be optimum.