Strength determination for rough substrate-coating interfaces with three-dimensional defect structure

Highlights

• Substrate-coating interface strength determined for CVD diamond-coated hard metal by new materials testing technique.

• First-time strength determination for rough substrate-coating interfaces with three-dimensional defect structure.

• Specimen geometry variation facilitates previously not possible investigation of interface defect texture.

Abstract

Quantitative information on the strength of interfaces between thin ceramic coatings and their substrates is crucial when aiming to understand their failure behavior. Many technically relevant substrate-coating interfaces are not ideally planar, but rather rough and contain stress concentrators forming three-dimensional defect structures. An example for such a realistic case, a polycrystalline diamond coating that partially penetrates voids and cavities in its Co binder-depleted WC-Co hard metal substrate, was investigated within the current work with respect to its unknown interface strength behavior. Special focus was laid on the influence of the applied load direction on the observed fracture behavior. Micromechanical specimens were produced via focused ion beam milling as geometry variants of a micro shear compression specimen with their loaded areas' relative inclination towards the substrate-coating interface varied from 0° to 88°. Specimen loading was performed until fracture with a flat punch indenter in a scanning electron microscope. The recorded fracture loads were associated with the spatial stress distributions at fracture via finite element-based analysis. A plateau of the determined maximum principal stress triggering fracture in the ceramic-ceramic interfaces was found for inclination angles ≥45°. This plateau value was identified as the interface strength by observation of the crack path at the substrate-coating interface via scanning electron microscopy and analysis of the effectively loaded interface area values. The presented novel material testing technique gives first and previously not accessible insight into the fracture behavior of rough substrate-coating interfaces with complex defect structure.

Introduction

In many applications the performance of components exposed to tribological loads is enhanced by thin ceramic coatings. For example, protective hard coatings are an established solution to enhance the productivity of metalworking tools in production processes such as milling, turning or drilling [1]. The loss of the coating in application is highly detrimental to the performance of the coated structure, e.g. a cutting edge of a milling tool that experiences accelerated wear when deprived of its coating [2]. Coating loss may be caused by wear or formation of cracks, e.g. those that propagate along the substrate-coating interface and lead to coating decohesion. Gradual coating wear is often the preferred failure mode of coated structures, since it does not occur suddenly as the formation of cracks. Coating decohesion is influenced by numerous factors such as the magnitude of load stresses compared to the substrate's yield strength [3], the load stress ratio [4] as well as residual stresses in coating [5,6] and substrate [7], respectively. Tensile load components are most critical with respect to crack initiation and growth in homogeneous materials, as well as at interfaces between metallic phases [8] and even more so between ceramic materials [9]. This is especially relevant when interfaces themselves are not planar but three-dimensionally rough. A very general definition of three-dimensional (3D) interfaces was suggested in [10] as “interfaces with hetero-phase character that extend out-of-plane into the two phases they join, and are chemically, crystallographically, and / or topologically divergent in three dimensions”. A good example for such a case are diamond thin films on WC-Co hard metal substrates [11]. In this system, the metallic Co binder acts as a catalyst for the formation of graphite during chemical vapor deposition of the coating. Therefore, the Co is etched out of the substrate material to facilitate the growth of the coating in diamond structure and avoid the formation of graphite [12]. The etching gives rise to a binder-depleted zone in the WC-Co substrate with a rough surface and complex-shaped partially interconnected cavities that are in part filled by the coating during its deposition [13].

Coating decohesion was reported to hinder the further propagation of the use of the described coating-substrate example system in machining of parts made of non-carbide forming metals, in which it has shown great potential to improve machining productivity [14]. A thorough understanding of the strength of interfaces between coatings and their substrates is essential when trying to understand and avoid coating decohesion. Most methods available for the characterization of the strength of interfaces only yield qualitative rankings for certain loading scenarios, but no quantitative interface strength information [15]. Typically, they also do not take account of the possible influence of a variation of the orientation of the external load on the determined strength value, i.e. a possible texture of a 3D interface defect structure. An example for the mentioned shortcomings are indentation techniques that lack the quantitative and inter-situative comparability of their results and currently do not allow for a direct control of the directions of the stresses that lead to interface failure [15,16]. Also scratching-based methods lack quantitative result comparability due to the complex and transient stress fields arising at the tested interfaces [17,18]. Micromechanical testing and interpretation of the stress-strain evolution during testing, combined with electron microscopy-based investigation of the μm-sized specimens can be used to address the discussed issues [19]. Some of the proposed methods enable to study interface strength by bending to impose a tensile load at interfaces [19,20], use a push-to-shear device to impose shear stresses [21], or load interfaces that are inclined with respect to the direction of a compressive load [22,23]. The last-mentioned techniques are interesting since in many applications, tension, compression and shear load components act simultaneously [24] and should thus be induced in laboratory test setups. Some documented efforts to do so have fallen short of demonstrating fracture at the studied strong interfaces [25], or induced strong plastification before plastic shear-off close to the studied metal/metal [22] and metal-ceramic interface [23] instead of interface fracture. Many proposed micromechanical test methods evaluate fracture properties of structures notched by focused ion beam (FIB) milling, also under multiaxial loading, but are not directly applicable for determination of substrate-coating interface strength [26,27].

Another micromechanical test method uses a micro shear compression (MSC) specimen to induce fracture at ceramic-ceramic interfaces by the sought combination of tension, compression and shear load [28]. In this approach, the tensile load component, determined quantitatively as the maximum principal stress pointing in an out-of-interface-plane direction, is referred to as the interface strength [28]. The MSC testing test setup was applied to determine interface strength between the microstructural constituents of high speed steel and a TiN hard coating [28]. The used MSC specimen geometry [28], however, is not yet useable for varying the applied load direction. Such possibility would be useful when studying the strength behavior of interfaces with a 3D defect structure and the possible influence of defect-orientation effects, i.e. defect texture, on the observed strength.

The current work presents a material testing method involving modifications of an MSC specimen geometry. Its already demonstrated capability to induce interface fracture and to deliver quantitative interface strength values will be advanced to study the strength of larger interface areas with a 3D defect structure under variable external loading directions.