T.W. Clyne

Fracture behaviour of ceramic laminates in bending—II. Comparison of model predictions with experimental data

In this paper, experimental data are presented for SiC laminates and compared with predictions from the model described in Part 1. The first requirement is a critical strain energy release rate, G[sub ic], value for the interfaces present in the laminates under study. A brief review is first given of how such data can be obtained, including reference to the phase angle characterizing the mode mixity. Testing is then described giving values of G[sub ic] in specimens containing single interfaces produced in a similar manner to those in the laminates. For the standard method of interface production, a G[sub ic] value was obtained of 7.5J m[sup [minus]2] [plus minus]0.5 J m[sup [minus]2] over the range of [psi] values ([approximately]40-60[degree]) believed appropriate for the laminate loading situation under study. Laminates have then been loaded in three point bend and load/displacement data recorded. Tests have been repeated after changes in certain variables, including layer thickness, interfacial toughness, loading span and layer strength. Comparisons have been made between theory and experiment for the shape of load/displacement plots and the dependence of the total work done on the variables studied. Close agreement has been observed. It is deduced that, for these laminates under thismorexa0» type of loading, the most promising approaches to improvement in toughness lie in decreasing the layer thickness and raising the interfacial toughness to the limiting value at which crack deflection can just occur. In addition, any increases that can be achieved in the strength of the SiC sheets will be beneficial.«xa0less

The effect of residual stresses on the debonding of coatings. I: a model for delamination at a bimaterial interface

An analysis is presented of how a general set of residual stresses is predicted to influence the debonding behaviour of a coating. This is oriented towards relatively thick coatings, in which significant through-thickness variations of residual stresses may occur. A through-thickness deposit crack is assumed to exist and the external loading is pure bending. An expression for the interfacial strain energy release rate, Gi, comprised of three terms, due to the applied loading, to the residual stresses and to an interaction term dependent on both, is derived. An experimental procedure to characterise the residual stress distribution is described. During interfacial fracture, the crack path, and therefore the loading state at the crack tip, are pre-determined by the plane of weakness. Since the critical value of G1 for interfacial failure has been observed to be a function of this loading state, a procedure is described for determining the phase angle characterising it for a given applied load and residual stress state. The results of these analyses are presented for various generalised residual stress states. In the second part of this pair of papers, the model is applied to experimental data for thermally sprayed coatings.

The effect of biological fluids on the adhesion of diamond-like carbon films to metallic substrates

Abstract Diamond-like carbon (DLC) films deposited by standard r.f. glow discharge CVD were exposed to various fluids (distilled water, phosphate-buffered saline (PBS) solution and 40% serum in PBS) to study the effect of this exposure on the stability of DLC coatings on metallic substrates. A substrate plastic straining technique was used to study the strength and adhesion of the coatings before and after exposure to various fluids. Distilled water had no apparent effect, whereas PBS caused localized debonding and cracking of the film. The serum-PBS solution had an intermediate effect. X-Ray photoelectron spectroscopy (XPS) measurements indicated that PBS tends to penetrate through surface perforations and attack the thin transition layer of graded Si C composition between the Si-rich layer on the substrate and the DLC coating. An increase in exposure temperature increased the population of defects in samples exposed to PBS. Coatings on Ti were found to be more resistant to damage.

The effect of residual stresses on the debonding of coatings—II. An experimental study of a thermally sprayed system

Abstract Specimens have been produced by plasma spraying of boron carbide coatings about 1 mm in thickness on to titanium alloy substrates about 3 mm thick. The residual stress distributions in these specimens have been calculated using a numerical process model and also estimated from observed changes in curvature on debonding. Good agreement was observed between the two methods, with both suggesting the average substrate stress to be about + 20 MPa and the average coating stress to be about −60 MPa. In both constituents, there was a significant positive gradient of stress level through the thickness. These specimens were loaded in four point bending until cracks propagated along the interface between substrate and coating. From the load/displacement plots obtained during this testing, and taking account of the effect of relaxation of the residual stresses during debonding, the critical strain energy release rates of the interfaces, G ic , were estimated to be ∼ 0.2–0.5 kJ m −2 . Substantial errors would have resulted from neglect of the presence of the residual stresses. Also of significance is the effect of the residual stresses on the mode mixity of interfacial loading, as characterised by the phase angle, ψ, since G ic has been often observed to vary with ψ. The value of ψ for the four point bend test in the absence of residual stress is about 47°, whereas for the specimens tested here it was estimated to cover the complete range from 90° (pure shear) to 0° (pure opening) as the applied load was increased. The quoted values of G ic were obtained in a regime where ψ ∼ 30°.

The influence of process parameters on consolidation efficiency when forming composites by spraying onto monofilaments

Abstract Composites have been produced by spraying titanium powder onto parallel arrays of tungsten monofilaments, using the vacuum plasma spray technique. The main variables studied were powder particle size, chamber pressure, fibre spacing and plasma gun operating variables. Porosity levels were measured by a gravimetric technique and also studied metallographically. It was found that relatively high porosity levels, largely located in the “shadow” of the fibres, persisted in all of the composites produced. These were measured as ranging from 10–15% to over 30%. Similar trends were apparent from metallographic sections, although the void contents appeared lower, 11 was found that lower porosity levels were favoured by larger particle sizes and lower chamber pressures, although the latter trend was not very clear-cut. Porosity increased sharply as the ratio of the gap between the fibres to the droplet diameter became less than a value of about 5. These results have been considered in terms of the mechanisms by which droplet impingement leads to envelopment of an array of fibres. It is concluded that it will be very difficult to identify a combination of operating parameters which will allow production of composite material with both fibre contents of commercial interest and suitably low porosity levels. The technique does not at present appear to represent an attractive option for commercial exploitation.

Tribological characterization of diamond-like carbon films on nonledeburitic high-speed steels

Abstract Hydrogenated diamond-like carbon (DLC) films were deposited on as-tempered nonledeburitic high-speed steel (NHSS) substrates (0–5–1–3Ti–2Nb grade) by capacitively coupled RF glow discharge in a methane atmosphere. The substrates were cleaned with an argon flux before coating deposition. Coatings of 1-μm thickness, good smoothness and adhesion were successfully produced. Microscale abrasive wear tests and field emission scanning electron microscopy were employed to evaluate wear resistance and morphology of the DLC films, substrate–film boundaries and worn components. The wear coefficient of the steel substrates was less than that of M2 tool steel, whereas that for the DLC layer corresponded to the value previously reported for titanium aluminium nitride. Certain decohesion micromechanisms in the NHSS samples after microscale abrasion testing were similar to these found in fractured specimens. Structural influences on the results are discussed.

The effect of biological fluids on the response of DLC films to a novel erosion durability test

Abstract The durability of diamond-like carbon (DLC) films, deposited by an r.f. plasma-assisted glow discharge chemical vapour deposition (CVD) technique onto metallic substrates (304 steel and Ti-6Al-4V), has been measured by a novel method in which the coating is bombarded with small solid particles in a carefully controlled manner. Two different types of particle, soft rounded glass beads and hard angular silica particles, were used to examine the effect of particle type on the wear behaviour. The durability was measured both before and after exposure of the specimens to various biological fluids, including distilled water, phosphate-buffered saline (PBS) solution and 40% bovine serum in sterile PBS. The specimens were exposed for 7 and 28 days at 37 °C in a humidified atmosphere containing 5% carbon dioxide. One set of experiments was carried out at 60 °C to determine the effect of temperature on the film stability. Environmental exposure had little direct effect on the film properties, but the fluids tended to degrade the interfacial adhesion when penetration to the interface was possible. This can occur via microscopic penetrations in the film or via the edges. The buffered saline solution appeared to be the most aggressive in this respect. A correlation between the damage mechanism during solid particle erosion and the type of particle used was found.