Micro-chemical characterization of fibermatrix interface in TI6ALAV-SICF composite [Caratterizzazione microchimica dell interfaccia fibra-matrice nel composito TI6AL4V-SICf]

The examined composite has the Ti6Al4V matrix reinforced by unidirectional SiC fibers (SCS-6) whose structure is shown in Fig.1. Under thermomechanical stresses diffusion phenomena and chemical reactions may occur at the fiber-matrix interface leading to instability and rupture [1-4], The material has been fabricated at C.S.M. Spa (Centro Sviluppo Materials by Hot Isostatic Pressing process (HIP); details of its preparation and characterization are reported in [5]. The samples have been subjected to the heat treatments reported in the Tab. 1. The interface has been examined by energy dispersion spectrometry (EDS), X-ray photoelectron spectroscopy (XPS) and Auger electron spectroscopy (AES) to assess chemical profiles in as-fabricated condition and after different heat treatments in vacuum. A detailed description of XPS and AES techniques can be found in ref. [6-8]. The concentration of the chemical elements has been determined by examining metallographic sections. Analyses have been performed in several positions of fiber, fiber-carbon interface, carbon coating and matrix. To get observation zones of greater extension, surfaces forming a small angle (<=2°) with the major axis of fibers have been prepared by mechanical polishing (sketch in Fig.2). Fig. 3 shows the XPS chemical image of a cross-section in as-fabricated state. The image and its detail in Fig.4 indicate that the chemical status of carbon is of two different types: graphite (BE of 284.7 eV) and carbide (BE of 283.0 eV). Fig.5 shows the SEM image of a sample prepared as described in Fig.2. Ten EDS measurements, from point P1 to P10, have been carried out along the direction indicated by the arrow; the distance between each test position is 10 μm. The results, plotted in Fig.6, can be summarized as follows: 1- Si is found in graphite (P3-P9), 2- Si signal disappears at the boundary graphite-matrix, 3- there is a transition region where progressively the signal of carbon decreases while that of titanium increases (P8-P10). In EDS measurements the diameter of interaction volume is 1,5 μm for matrix and 4 μm for carbon. To get an higher spatial resolution (∼ 200 nm), AES examinations have been performed on all the samples. To simplify only the results about the most different samples (as-fabricated and treated 1000 h at 600 °C) are reported in this paper (Fig.7 and Fig.8 respectively). Fig.7 shows that the thickness of carbon layer is not homogeneous and three different zones (points 1, 2 and 3) are observed. AES spectra indicate the presence of oxygen in the interfaces (points 1 and 2) and in minor quantity also in the carbon layer (point 3). Potassium has been found too. In the heat treated sample (Fig.8) the carbon layer is thinner and more uniform than in as-fabricated one. The signal of oxygen is high only at the interface carbon-matrix where titanium is also present. XPS quantitative analyses on cross sections of 1 mm2 (Tab. 2) show the higher content of oxygen in the heat treated samples but are not able to find evidence of TiC. Fig.9 shows the image of a sample heat treated at 600°C for 1000 h and prepared as shown in Fig.2. In fact, the preparation of the surface allows to examine in the best conditions the diffusion of elements through the interface carbon-matrix. The AES spectra evidence the presence of titanium inside the carbon coating near the interface (point 4) and of carbon inside the matrix (point 5). Therefore, with respect the as-fabricated condition, carbon is penetrated in titanium and titanium in carbon. This result is in agreement with those from EDS micro-analysis (Fig. 6) and with literature data [1-3, 91. The aforesaid analyses show that the interface after heat treatment do not undergo relevant changes with respect the as-fabricated state. To explain experimental data theoretical models [9], which allow to calculate the reaction zone growth, have been considered. According to these models the values of the thickness x of the reaction zone have been obtained for fabrication condition (T1 = 890 °C, t1 = 1,5 h) and for heat-treatment parameters (T2 = 600°C, t2 = 1000h). The values of Q and h introduced in eqs.(1) and (2) have been taken from ref. [9] and were determined for a (α+β) structure. In the second case the calculated data lead to overestimate the thickness of the reaction zone because the values of Q and k0 introduced in eqs.(1) and (2) refer to a biphasic structure while at 600 °C the matrix is made exclusively of a phase. It is worth to recall that the carbon diffusion is larger in β than in a. Moreover, the heat treated material already has a zone of reaction, formed during fabrication process, thus carbon should cross first the TiC layer where diffusion is slower than in the matrix. In spite of that, the calculated data indicate that the reaction zone in the heat treated material (x = 0.26 μm) is lower than in the as-fabricated one (x = 0,78 μm). The result indicates that the reactions in the carbon-matrix interface during the fabrication of composite, are much more important than those occurring after prolonged exposure at 600 °C, working temperature foreseen for the material application in aeromobile engines. That is confirmed also by mechanical tests performed by some of us on the same materials [10]