Search

An x-ray method for the determination of concentration profiles in thin film diffusion couples is presented. This method is based on the theory of Fourier analysis of x-ray diffraction profiles which is generalized to polycrystalline samples showing nonuniform lattice parameter. A Fourier synthesis of the concentration spectrum is possible when the influences of the particle size and the strain in the sample as well as the instrumental function are eliminated from the measured diffraction profile. This can be done by means of reference profiles obtained from layers of the diffusion components. Absorption of the radiation in the sample is negligible when diffusion couples of symmetrical sandwich structure are used. The method is tested experimentally in the system Au-Cu.

A method for the retrieval of concentration spectra of crystalline mixtures from x-ray diffraction profiles is described. Both spectra are associated with an integral equation which can be simplified to yield a set of linear equations. Solutions are readily obtained by triangular decomposition since the equations are typically described by band matrices. The method has been tested with simulated diffraction profiles to allow for a variety of options concerning the concentration spectra. The procedure is well suited for binary alloys containing crystals of different structure and lattice parameters. The time consumption of the calculations is significantly reduced as compared to hitherto available methods.

Computer simulation of superimposed lattice, grain boundary and surface diffusion, characteristic for polycrystalline thin film diffusion systems, was performed by way of discretisation of the nonlinear diffusion law. In order to give a vivid impression how such a complex process takes place under some typical conditions we have chosen pseudo three-dimensional computer plots of the spacial distribution of the concentration instead of the commonly used iso-concentration diagrams. The following cases are considered: a) Polycristal — single crystal couple with highly diffusive grain boundaries and surfaces, grain boundaries and interface, grain boundaries, surfaces and interface. b) Polycrystal — polycrystal couple with different displaced grain boundaries as well as different diffusivity ratios for the bulk and the highly conducting regions.

Limited diffusion spaces which may be easily traversed by atoms diffusing along grain boundaries generally cause thin-film diffusion to differ from common bulk diffusion. These peculiarities were studied in Ag−Au thin-film couples by means of electrical resistance measurements. Diffusion coefficients were found to decrease with annealing time mainly as a consequence of recrystallisation and recovery in the films during the diffusion anneal. It is shown that the rate of homogenisation is fairly independent of the film thickness thus giving evidence that diffusion into the crystallites occurs out of the grain boundary network rather than directly through the couple interface. Effective diffusion coefficients determined between 150 and 250° C ranged from 10−14 to 10−16 cm2 s−1 revealing an activation energy of 25 kcal mol−1.

On the basis of hypothetical thin film couples in which a grain-boundary supported diffusion process is simulated, it is shown that the concentration spectrum may contain maxima which would not occur if the process were governed by lattice diffusion only. As such maxima consequently appear in the diffraction pattern, real diffraction profiles of interdiffusion layers must be interpreted carefully, keeping in mind that diffraction peaks need not necessarily be associated with a particular phase. The thin-film diffusion system Au-Cu is chosen as an example to demonstrate experimentally the generation of diffraction maxima which must not be attributed to phase formation.

A novel microhardness-tester has been designed for loads between 10−2 N and 5 x 10−5 N. The apparatus can be operated in a scanning electron microscope without installation of additional adjusting shafts. The force is generated electromagnetically, transduced to the indenter (diamond cone or pyramid) by bending a double leaf spring, and measured by means of strain gauges fixed to the springs. The indentation cycle is fully programmable and is controlled by the strain gauge signal. The equipment can be tilted with respect to the electron beam thus making possible the observation of the tip during indentation.