R. de Ridder

Diffraction effects due to shear structures: A new method for determining the shear vector

Abstract A survey is given of the diffraction effects associated with the shear structures and a theoretical analysis is made of these effects. From the main results a method is deduced which enables a full crystallographic analysis of a shear structure from purely geometrical features of the diffraction pattern. The method is applied to shear structures associated with deviations from stoichiometry in rutile. The displacement vector associated with the shear process is determined.

The structure of defect manganese silicides

Mn-Si alloys of atomic composition nearby 4–7 are studied by electron diffraction. It is found that there exist a variety of phases MnSi1.75−x (x is a quasi-continuous defect fraction in the range 0.00 < x < 0.05). The four phases, already studied by X-ray analysis, are identified. A general structural model is proposed to explain the observed patterns.

Jahn-teller domains in MnxFe3−xO4 as observed by electron microscopy☆

Electron microscopy and electron diffraction is used for investigating the microstructure of MnxFe3−xO4 spinels. Special attention is paid to the composition range near the cubic to tetragonal phase transition and heating experiments in situ were performed to study the dynamical behaviour at the transition. Domains due to the cooperative Jahn-Teller effect have been observed in samples with a composition corresponding with x=1.95 and x=2.00. The effect of composition is considered and the domains are characterized crystallographically from diffraction evidence. The domains are limited by (110)-planes of the spinel structure. Heating experiments also allowed us to exclude the magnetic transition as the origin for the observed domain structure.

The most storage economical Runge-Kutta methods for the solution of large systems of coupled first-order differential equations☆

Abstract It is shown how the attainable minimum for the memory requirements of Runge-Kutta methods can be realised for methods of the third order. These economisable third order methods belong to a one parameter sub-family from which two particular members with low error bound are selected.