X-ray diffraction topography : methods and applications

This thesis describes the application of the well established technique of X-ray diffraction topography to a variety of problems, and includes considerations of the optimum conditions for taking rapid topographs. Chapter I contains a brief review of the subject together with an indication of the range of applicability. Several modifications of X-ray topography exist and several are briefly described to illustrate the principles and mechanisms of image formation. Contrast is formed in one or both of two ways. Regions of crystal may be so badly misoriented from the bulk, that no beams from the source can satisfy the reflection condition formulated by Bragg. This type of contrast is known as orientation contrast. The second type of contrast arises from the point to point lattice displacements giving a different reflecting power around the defect. This second type of contrast, termed extinction contrast, is dependent on the perfect crystal diffraction. A description of the different types of extinction contrast is given in terms of the dynamical theory of diffraction in a perfect crystal and the significance of the direct, intermediate, and dynamical images stressed. The chapter concludes by sketching some of the important fields of application. The second chapter is concerned with the relative merits of scanning and wide beam X-ray topography. A description of the technique of wide beam topography using the Kandbeta; line is given and it is demonstrated that the resolution is as good as that using Lang's scanning technique. A new wide beam method using the tungsten Landalpha;1 line is proposed and demonstrated. From expressions derived for the exposure times, it is seen that the important parameter to maximise is the power per unit horizontal length of the X-ray source. The speeds of the two techniques are compared for a variety of generators and the advantage of using an Elliott GX6 is made clear. A direct imaging system is clearly preferable in principle to film recording and Chapter III is devoted to a discussion of such systems. Following a review of previously developed imaging systems, a new method of directly displaying X-ray images using a channel plate is demonstrated. The remainder of the chapter gives detailed calculations of the efficiency of a channel plate to X-rays in the wavelength region 0.5andndash;2andnbsp;andAring;. Using published data, the efficiency is calculated for several cases and compared with previous experimental work. To better than an order of magnitude, these efficiencies may be used to predict the intensity of the image produced on the image converter screen. Chapter IV falls into two sections. The first describes an experiment to determine the minimum thickness at which dislocations are visible in crystals using normal topographic techniques. It was found that the value was dependent on the fraction of the material thickness taken up by misoriented material. This explanation was found to be in qualitative agreement with experiment, the minimum thickness fell from O.4 of an extinction distance for the low order reflections to 0.2 of an extinction distance for high order reflections. The second part consists of a comparison between experimental image profiles and those simulated on a computer. Quite reasonable agreement was found and the sources of error in the method of computing and ways to avoid them are discussed in terms of practical and basic limitations. The remainder of this thesis describes application of X-ray topography to four different types of material. Chapter V is devoted to a study of defects in a silicon slice following device fabrication. Interest in the effect of crystallographic defects on the performance of integrated circuit devices has led to a large number of investigations over the last decade and these are reviewed and their conclusions summarised. An attempt is made to measure the stress at the junction edge, produced by the mismatch in the ionic radii between dopant and matrix. As a result of approximations in the theory and measurements, the force per unit length of junction, measured to be 2 andtimes; 104 dynes/cm, must be taken as order of magnitude only. Similarly, only an order of magnitude estimate of the stress is permissible. There follows a description of a combined Scanning Electron Microscope and X-ray topographic experiment to determine the cause of breakdown fingers, sometimes seen at the junction edges in the beam induced conductivity mode in the S.E.M. It is concluded that they are due to scratches in the oxide mask prior to diffusion. The rest of the chapter is given to a detailed description of dislocations and their contrast observed in a silicon slice heavily doped with boron. The slip behaviour is not analysed in detail, but seems to support the conclusion that this deformation, due to thermal shock, is independent of the doping. Interactions between 'emitter edge' dislocations are described and the anomalously narrow image widths of such dislocations explained by considering the effects of overlapping strain fields. Burgers vectors of dislocations inside the diffused regions are determined, and the contrast of these dislocations at high density is interpreted as an effect due to overlapping images. Interactions between the 'inside' and 'emitter edge' dislocations are described. In regions where the dislocation density was low, the dislocations exhibited a reversal of contrast on reversal of the diffraction vector. With the diffraction vector parallel to the dislocation line, the contrast was black-white, reversing with the diffraction vector. With the diffraction vector at 60° to the line, the contrast was either black or white, reversing with the diffraction vector. This effect is interpreted in terms of surface relaxation and computations performed using the Penning-Polder theory are in qualitative agreement with the experimental results. Chapter VI describes the application of X-ray topography to the study of defects in natural fluorite. Dislocations are identified with Burgers vectors parallel to andlang;101andrang; directions. These are nearly pure edge in character and the Burgers vector is presumed to be 1andfrasl;2 andlang;101andrang;. A direct correspondence is observed between bundles of dislocations and regions of birefringence contrast. A uniform birefringence contrast in the matrix, which can not be explained in terms of dislocation stress, is also observed. This is explained in terms of impurity atoms, sited between {111} planes during growth, distorting the crystal normal to the {111} faces. The boundaries between regions of birefringence lie along inclined andlang;110andrang; directions and on this model, no strain is associated with these boundaries, in agreement with the X-ray topographic evidence. Extensive planar faults, lying on {111} planes and with fault vector a non integral value of the lattice spacing, are postulated to be thin lamellae of material containing markedly different impurity concentration. In Chapter VII, some vapour grown layer compounds are studied. Individual dislocations are resolved in SnSe2, TiSe2, SnS2, TiS2, ZrS2 and HfS2. The Burgers vectors of the dislocations in SnS2 were determined and their origins discussed. Large area stacking faults bounded by 1andfrasl;3 andlang;1100andrang; partial dislocations are also observed in these specimens. Dislocation and defect configurations in the other compounds are described and a modified divergent beam method for taking topographs of bent crystals is demonstrated. Some preliminary experiments on pure iron are described in Chapter VIII. Contrast is observed suggestive of dislocation helices and rows of loops in these strain-anneal grown crystals. Arrays of dislocations of pure edge type and Burgers vector parallel to [100] are interpreted as being contained in a low angle boundary wall, following Futagami. The tilt angle across such boundaries is measured to be a few seconds of arc. Finally, Chapter IX suggests a few directions for further investigation.