Papers by Author: P. Fielitz

Abstract: Aluminium is a key element in geological and man-made materials which has only one stable isotope and no radionuclides with half-life times suitable for standard experimental diffusion studies. Here we report on our method using the radioisotope ²⁶Al (t_1/2 = 7.4×10⁵ a) as a quasi-stable tracer for aluminium in combination with SIMS depth profiling. First, our data for the aluminium bulk diffusivity in a-alumina are discussed jointly with published oxygen bulk diffusion coefficients. They clearly show that the relation D_Al>>D₀ is valid in the temperature range 1200 °C ≤ T ≤ 1800 °C. In an analogous manner, the two rare stable isotopes ¹⁸O and ³⁰Si are used together with ²⁶Al in diffusion studies of generic examples of materials which either consist of aluminium, silicon and oxygen only, or where these three elements are key constituents of the structure. For the crystalline aluminium silicate mullite our diffusivity data for aluminium, oxygen and silicon are used to explain the kinetics of the solid state formation reaction of mullite and the segregation kinetics of alumina from mullite. Finally, the diffusivities of oxygen and aluminium in model aluminosilicate glasses are presented as a function of temperature for different Al³⁺/Na⁺ ratios. For the aluminium silicate mullite and for the aluminosilicate glasses the relation D₀>D_Al>D_Si is valid regardless of the exact composition. For the glass system the activation enthalpies of aluminium and oxygen diffusion decrease with decreasing Al³⁺/Na⁺ ratio.

Abstract: Simultaneous ¹⁸O and ²⁶Al tracer diffusion experiments were performed in nominally undoped single crystalline α-Al₂O₃. The results clearly show that the bulk diffusivity of aluminium is much higher than the bulk diffusivity of oxygen in nominally undoped alumina. Comparing the ²⁶Al tracer diffusivities of Ti doped (300-400 wt. ppm Ti) and nominally undoped single crystalline α-Al₂O₃ one finds that the aluminium bulk diffusivity is insensitive to the Ti doping.

Abstract: Exact mathematical solutions of the grain boundary diffusion equation in thin films have generally a complicated form, which is too cumbersome for the evaluation of experimental average concentration depth profiles obtained by sectioning techniques. On the other hand the accuracy of the exact solutions is not necessary for practical purposes so that it is useful to derive sufficiently accurate approximate solutions. We propose a method to derive such solutions for a thin film if the grain boundary diffusion is in the B2 regime. These solutions are derived for different diffusion sources.