The precipitation hardened CuCrZr alloy has been used in the past as a structural material for actively cooled plasma facing components of nuclear reactors (Tore Supra and JET) and selected as the heat sink material for divertor parts of the future International Thermonuclear Experimental Reactor (ITER) [1]. Because of its high melting point and good thermal conductivity, W is a promising material to use as protection (armour) for thermal shields made of CuCrZr. However, W-Cu joining is really challenging for the high thermal expansion mismatch between W and Cu alloys (alpha(Cu) approximate to 4 alpha(W)), the high elastic modulus and brittleness of W. Important requirements for high quality W-armours are low porosity and low impurity content. For realizing the joints, plasma spraying (PS) has been used for its simplicity, the possibility to cover complex and extended surfaces and the relatively low cost. An appropriate interlayer was optimized to increase the adhesion of W on the Cu alloy and to provide a soft interface with intermediate thermal expansion coefficient for better thermo-mechanical compatibility. It was possible to manufacture mock-ups for ITER by depositing up to 5-mm thick W-coatings on tubular substrates of CuCrZr (Fig.1), which were able to sustain a remarkable number of thermal fatigue cycles under high heat flux (up to 5 MW/m(2)) [2-4]. Microstructural and mechanical characterization confirmed the good properties [4,5]. The chemical composition of the Cu alloy of the substrate [6] is: Cr 0.65, Zr 0.05, Cu to balance (% wt). The interface (thickness similar to 800 mu m) was realized by keeping tie deposition temperature below 157 degrees C to prevent overaging of CuCrZr alloy [7]. PS process involved the following steps: 1- the first layer (-100 mu m) of pure Ni directly on the CuCrZr alloy, 2- the second layer (similar to 350 mu m) composed of a grading mixture of Al-12%Si and Ni-20%Al powders (100% of Al-12%Si near pure Ni, 100% of Ni-20%Al at the opposite side, the relative quantity has been progressively modified with steps of 5 % in successive torch passes), 3- the third layer (similar to 350 mu m) formed of a grading mixture of Ni-20%Al and W (content varying from 0 to 100% with steps of 5%). Finally, pure W was sprayed; the total thickness of the coatings (layered interface + W) ranged from 4.5 to 5.0 mm in all the manufactured mock-ups. The investigated samples (14) have been obtained by cutting the material perpendicularly to the free surface of the coating. A set of 7 samples as been heat treated at 550 degrees C for 24 hours in ultra high vacuum (UHV) to investigate the extension of Al and Ni diffusion into W coating. SEM observations and EDS mapping have been carried out to investigate morphology and distribution of chemical elements across the layered interface, which has been then studied in more detail by X-ray photoelectron spectroscopy (XPS) and Auger electron spectroscopy (AES) [8]. The external W layer has been examined by Inductively Coupled Plasma Emission Spectrometry (ICP-ES) to assess the possible presence of other chemical element;, in particular Ni, and to measure their content. X-ray diffraction measurements have been carried out at increasing temperatures up to 425 degrees C in Ar atmosphere by means of a Anton Paar HT-16 camera with Mo-K alpha (lambda = 0,071 nm) radiation. As shown in Fig. 2, the X-ray beam was simultaneously focused on CuCrZr alloy, interlayer and W coating. Peak shifts are correlated to residual stresses. Distribution of chemical elements in interlayer and coating The SEM micrograph and EDS maps in Fig. 3 show the structure of the interface near the CuCrZr alloy. Fig.3 (a) manifests the complex shape of the interface layers: the material appears compact and no macro-porosity is observed. EDS micro-analytical maps, recorded in the same sample area, evidence the distribution of Ni (b), Al (c) and Si (d), resulting from the successive spraying of powders with different chemical composition. The sprayed powder is a mixture of particles of different metals. In general, even if some droplets may coalesce during the flight, each droplet corresponds to a single particle and has the same composition. When the droplets splat on the substrate they solidify and, in the case that the temperature of the substrate is sufficiently high, long-range solid state diffusion may occur. In the present case the chemical distribution of single elements in EDS maps well corresponds to the shape of the single droplets and no nuance effect across the boundaries between neighbour droplets is observed. The result indicates that no long-range elemental interdiffusion takes place. Therefore, the low temperature of the substrate guarantees a fast cooling of the solidified metal and suppresses the long-range diffusion. XPS data after surface cleaning by Ar ion sputtering and after sample heating in UHV are reported in Table 1. Since the area of XPS analysis (about 1 mm of diameter) is much larger than that of AES analysis (less than 1 mu m of diameter), the possible diffusion of constituent elements can be evaluated only by using AES. SEM image and multipoint AES analysis of the layered interface are reported in Fig. 4. The spectra of Al KLL (b), Ni LMM (c) and W MNN (d) were collected in the points 1-10 marked in Fig. 4 a). Auger spectra confirm that there is no evident diffusion of the chemical species between neighbour solidified droplets. The interface structure after UHV heat treatment is displayed in Fig. 5 a). Auger spectra of Ni and Al, collected in the points 1-3, are reported in Fig. 3 b). The heat treatment does not cause a long-range migration of Al and Ni into the W coating, since the intensity of Al and Ni peaks is negligible in the points 1 and 3. The ICP-ES results show that the total content of Ni in the external W layer of as-prepared material is 0.4 (wt%) whereas Al and Si are below the detection limit. As shown by AES measurements, the presence of Ni in the external layer can not be ascribed to the diffusion from the layered interface, thus it is clearly clue to the low purity (commercial grade) of W powders used in spraying operations. Therefore, it is concluded that coating process requires W powders of higher purity. Residual stresses in substrate, interlayer and coating Fig. 6 shows the X-ray diffraction spectrum of the system substrate-interlayer-coating, collected at room temperature (25 degrees C). High precision measurements have been carried out at different temperatures and experimental peak positions have been compared with those calculated from interplanar distances d(hkl) of JCPDS-ICDD database taking into account the thermal expansion due to the temperature change T: (1) Delta d(hkl)/d(hkl) = alpha Delta T being the thermal expansion coefficient (see Table 2). In Fig. 7 the spectra collected at 25 and 425 degrees C are compared; the markers indicate the positions (calculated) of the peaks at room temperature without residual stresses. The peak positions of Cu and W at 25 degrees C correspond to the calculated ones whereas remarkable shifts occur in Ni and Al peaks. This demonstrates that residual stresses are present in the interlayer but not in the substrate and coating. At 425 degrees C all the peaks shift towards lower angles. To understand whether the shift is also clue to residual stresses, in addition to thermal expansion, experimental peak positions have been compared to the calculated ones. Table 3 summarizes the results and reports the strains depending by residual stresses, which have been determined by: (2) epsilon = (d(Sper) - d(Calc)) / d(Calc) Table 3 clearly shows how the strain of Al at 25 degrees C is about two orders of magnitude higher that those of Cu and W. The strain of W is negligible, that of Cu low. Therefore, it is concluded that the soft interlayer properly serves its purpose in the temperature range examined here.
