No abstract

The determination of residual stresses via X-ray diffraction is briefly reviewed, with particular emphasis on the triaxial stress state. A new method is proposed for determining the general stress tensor, which considerably reduces the variances of the stresses due to counting statistics and gradients. The procedure involves a generalised least-squares solution of strains measured at various tilts of the X-ray beam to the sample, and a new set of tilts is recommended to minimise these errors.

Short-range order parameters were obtained for a single crystal of 62Ni3Fe quenched from above T c. These are the first absolute measurements of diffuse elastic neutron scattering complete enough to separate the effects of local order and atomic displacements (up to quadratic terms) and indicate that neutrons are ideally suited for such studies. The results are compared to those obtained with X-rays. The short-range order intensity of Ni3Fe is similar in shape to that for Cu3Au but in this alloy it is due to the plate-like nature of ordered domains, not the antiphase domain boundaries which are present in such regions in Cu3Au.

No abstract

No abstract

In order to convert residual strains measured by x-ray diffraction techniques into residual stresses, appropriate x-ray elastic constants have to be measured. Since these x-ray elastic constants may depend on the metallurgical state, deformation, and entire specimen history, errors in stress values may result if the constants are not measured for representative material states. In the present work, it is shown that in same cases these errors may be large.

Equations are derived for errors due to counting statistics in the determination of local order parameters and interatomic displacements from the diffuse X-ray scattering from solid solutions or nonstoichiometric compounds. The Georgopoulous-Cohen procedure is employed, which permits the determination of displacements for individual (AA and BB) pairs. It is shown that certain analysis procedures and a large variation in the ratio of scattering factors through the measured volume in reciprocal space improve the accuracy of the separation of the various contributions to the intensity with this procedure, particularly with respect to the terms due to displacements. Therefore, a short-wavelength radiation (Mo K~, for example) is sometimes desirable. Weighting the measured intensities with respect to the counting errors also improves the separation. The new procedures also facilitate the analysis of high-temperature measurements and data from ternary systems.

No abstract