Printed in the United Kingdom Neutron Diffraction Texture Analysis of 1 Vol . % Cu in Aluminium

Neutron diffraction texture determination was carried out in two-phase composites, AICul%, prepared by powder metallurgical methods and extruded 84 and 89% respectively. The texture of the minor Cu-phase can be measured with a sufficient degree of accuracy within reasonable time. This texture is weaker than the texture of the Al-matrix phase corresponding to the lower internal deformation degree of the former one. It is estimated that the method can be extended to even lower volume fractions of the second phase.


INTRODUCTION
Phase analysis and texture analysis in polyphase materials can be carried out by diffraction methods measuring relative intensities of Bragg-peaks. If one of the phases is present only in a small volume fraction then this method is limited by the relative accuracy with which intensity measurements of Bragg-peaks can be carried out. If the strongest Bragg-peak of a minor phase is in the order of magnitude of the background scattering then this method does not work. Hence, the lower detection limit of a minor phase depends on the peak to background ratio of the considered diffraction method. This is generally better in the focussing methods compared to nonfocussing methods, it is better using photographic methods compared to diffractometer methods since a line on a photographic film can be easier distinguished from the background than a peak in an intensity plot and it is better in neutron diffraction compared with X-ray diffraction since all intensity correction factors in neutron diffraction are smaller than those in X-ray diffraction and can be calculated with a better accuracy (Bacon, 1975;Szpunar, 1976;Welch, 1986).
As a rule, the lower limit of texture analysis using X-ray diffraction is assumed to be at about 5 Vol. % of a minor phase in a polyphased material. In the present investigation it will be shown that, using neutron diffraction this lower limit is much smaller. The texture of 1 Vol. % copper in an A1-Cu composite could "easily" be measured and it turns out that this is not yet the lowest possible volume fraction.

EXPERIMENTAL PROCEDURE
Two A1-Cu composite samples were prepared by powder metallurgical methods. They were extruded 84% and 89% respectively. The characteristic data of the samples are given in Table 1. Because of the axial symmetry of the extrusion process the obtained textures of the copper and the aluminium phase are axially symmetric too, i.e. they are fibre textures.
The texture measurements were carried out at the research reactor FRG-1 at Geesthacht using the texture diffractometer TEX-1. A wavelength of , 0.1618 nm was obtained by an A1 (111) monochromator. A description in detail of the instrumental layout is given by a GKSS-Report (Brokmeier, Bunge, Brehler, Wagner, Wille, 1987). Due to the low absorption coefficient, samples of the dimensions given in Table 1 can be considered as being "spheres" such that no absorption correction is needed (Tobisch and Bunge, 1972). The limiting 20-angle of the diffractometer being 70 , the peaks given in Table 2 were within the available 20-range. Of these peaks, the Cu (111) peak was slightly overlapped within AI (200) peak as is to be seen from to-20-scans as shown in Figure 1.
Since the textures are fibre textures, only one-dimensional pole figures are to be measured. They are obtained by ;t-scans of the Eulerian cradle corresponding to the pole figure angle tr, i.e. the angle towards the rod axis. For each pole figure the Bragg-angle 20 was chosen according to the peak-position of the 20-scan with an angular divergence of 1.1. Additionally several 20-scans were measured in the 20-range of the A1 (200) and Cu (111) in angular steps of Ate 5 in order to better separate these two pole figures. The results are shown in Figure 2 for the two samples and three aluminium pole figures and two copper pole figures respectively. From these curves also the peak to background ratio can be estimated. The measuring times for the whole texture determination are shown in Table 3.

RESULTS
As is seen from Figure 2, the volume fraction of 1% Cu in 99% AI corresponding to total amounts of 106 or 66 mg Cu respectively was sufficient to determine two pole figures of the minor phase. The total deformation degrees of the two samples were 84% and 89% respectively. This is also the deformation degree of the softer Al-matrix phase, whereas the harder Cu-phase was much less deformed (Brokmeier, B6cker and Bunge, 1988). The estimated internal deformation degree of this phase was only 45% and 55% respectively. Correspondingly, the texture of the Cu-phase was much weaker than the texture of the Al-phase.
Orientation distribution functions (ODF) were calculated from the pole figures using three and two pole figures respectively and series expansions up to l 22 and 20 for the two phases (e.g. Bunge, 1982;Dahms, 1987). The results are shown in Figure 3.
Both textures are double fibre textures (111) and (200) with peak intensities given also in Figure 3. NEUTRON    necessary on the expense of still longer measuring times). It was estimated that in the system A1-Cu texture measurements of 0.1% Cu should be possible. (Preliminary experimental results obtained in 0.1 Vol.% samples seem to corroborate this estimation.) This opens up the possibility of texture studies in phases occurring in small volume fractions such as for example small precipitates in matrix materials which are virtually not possible by X-ray diffraction. This will be interesting with respect to the precipitation process itself as well as to the deformation of precipitation hardened materials.