ADVANCES AND APPLICATIONS OF NEUTRON TEXTURE ANALYSIS

Due to the high penetration depth ofthermal neutrons for most materials, neutron dif- fraction is an efficient tool for bulk texture analysis. The main applications are investigations of coarse-grained materials, non-destructive pole figure measurements of identical samples at different states, detection ofweak textures, measurement of unpreparedsamples and texture investigations of multi-phase systems. It should be pointed out that neutron texture analysis is suitable for basic research as well as for applications.


INTRODUCTION
Texture analysis by means of neutron diffraction is one of the standard methods in modern texture analysis. Since the first neutron diffraction texture measurement carried out by Brockhouse (1953) a number of improvements have been made. This includes instrumental developments, methodical improvements and of course mathematical approaches for a quantitative texture analysis. Different authors have published these improvements from time to time with a concentration to different topics. The following list gives only a brief collection of some important papers: neutron texture analysis (Szpunar, 1976;Welch, 1986;Bunge, 1989;Brokmeier, 1994); instrumentation (Feldmann, 1989;Hansen et al., 198 l;Bunge et al., 1982;Wenk et al., 1988); quantitative texture analysis (Bunge, 1969;Matthies et al., 1987;Schaeben, 1988). Looking on diffraction methods to measure textures of a sample volume with a statistical arrangement of crystallites mainly three different radiations are in use. These are X-rays produced by a conventional X-ray tube or by a synchrotron, electrons and thermal neutrons. The differences of these radiations are based on the interaction of the radiation with matter (see Bacon, 1975;Cullity, 1956). Focusing to texture measurements the main differences between X-rays, electrons and neutrons are the penetration depth, the local sample volume and the global sample volume, (see Table I). It must be pointed out that Table I is given for a high number of conventional materials. But the enormous variety of materials having large differences in the scattering behavior cannot be taken into account completely. Thus, the applicability ofnew materials has to be estimated individually.
According to the more physical differences there are practical differences such as the availability and the costs. Methods needing a large scale facility like a neutron beam or a high energetic X-ray beam are limited in the number ofsources, in the number ofinstruments and ofcourse in the availability of beam time. Based on the properties shown in Table I electron diffraction is favored in microtexture analysis. The standard method for texture analysis is the use of conventional X-rays. Recent improvements particularly in new X-ray optics and new detector systems have enlarged the efficiency and the accuracy of the standard X-ray techniques. Nevertheless the high penetration depth is necessary for a lot of problems particularly in global texture analysis and in local texture analysis inside a bulk. Compared to high energetic X-rays neutron diffraction texture analysis is a well-established method available in most of the neutron research centers worldwide with a high number of applications on different materials. In future new developments in high energetic X-rays will overcome some neutron applications which are limited by the beam flux and the lack of a local resolution in sample volume. The principles of neutron diffraction are the same as those for X-ray diffraction. Therefore, neutron diffraction methods for texture analysis are closely parallel to well-known X-ray diffraction techniques. But owing to the high penetration depth of neutrons there are some typical advantages ofneutrons against conventional X-rays. Table II shows the penetration depth of some minerals for neutron (1.00 ,), for Cu Kc-Xrays (1.54,) and for high energetic X-rays (0.21 ,).
Compared to 1.54 X-rays, neutrons have in all cases a much higher penetration depth even for epsomite, a mineral with a high water content. Neutrons and 0.21 , X-rays have penetration depths in the same order but show some remarkable differences. Hematite, an iron oxide has a relatively high penetration for neutrons while epsomite has a relatively high penetration depth for 0.21 A X-rays. This behavior correlates to the presence of heavy atoms, light atoms and hydrogen in the material. According to this relatively high penetration a number of advantages result which will be discussed in detail later.
A second difference between neutrons and X-rays is that the atomic scattering amplitude for neutrons is independent on one hand on the scattering angle 20 and on the other hand on the atomic number. Consequently, the diffraction patterns obtained by neutrons or X-rays are different. Light atoms contribute in the same order as heavy atoms on neutron scattering. In the case of calcite and corundum the basal plane pole figure can be measured with neutrons but virtually not with conventional X-rays due to the low scattering factor for X-rays. Moreover, high order reflections ((400), (440), (444)) can be measured using neutrons with nearly the same order ofintensities as low indexed reflections Material  ((200), (110), (111)). A very special situation is given for tetragonal 7-TiA1 (see Fig. 1). According to the crystal structure P 4/mmm there are two sets of reflections detectable by X-rays. Strong reflections will be obtained in the case of adding the contributions of titanium and aluminium ( (111) (112), which cannot be measured by X-rays.

POLE FIGURE MEASUREMENT
One of the main advantages of neutron measurements is the use of the spherical sample method after Tobisch and Bunge (1972) to get complete pole figures. This method allows the measurement ofthe average texture ofsamples with some cm 3 in volume and a background correction only. Depending on the high transmission absorption corrections can be neglected not only in spherical samples but also in cylinders and cubes. Thus pole figures can be measured non-destructively of non-uniform samples as shown in Fig. 2.
The scanning routines to measure pole figures are closely parallel to Xray measurements. Figure  It must be pointed out that the availability of these methods is close connected to the instrumentation. A four circle diffractometer with a single detector belongs to the standard instrumentation of all research centers but the additional detector systems are restricted to some places worldwide.

ADVANTAGES OF NEUTRON DIFFRACTION
As pointed out above the main advantages arise from the high penetration of neutrons. The investigated sample volume in neutron diffraction texture measurement is 104-10 6 times higher than in the case of X-rays. Thus neutron diffraction is an efficient tool for global texture investigations. The measurement oflarge sample volumes in the order of some cm 3 gives on one hand the possibility to analyze coarse-grained materials and on the other hand to increase the grain statistics in finegrained materials. Moreover, in the case of inhomogeneous textures an averaging takes place and texture can be correlated very well to mechanical testing carried out at similar sample volumes.
A second advantage is the sample preparation. The main step of sample preparation is to fix the sample on a sample holder. A surface preparation such as polishing is not necessary because a surface roughness of some mm has no important influence on the accuracy of the measurement. The ideal neutron sample is a sphere, to have no anisotropic absorption. But most materials show a rather high penetration depth (Brokmeier, 1995) so that the measurement of cylindrical or of cubic samples gives the same accuracy as spherical samples. Therefore in practice cubic and spherical samples are preferred because of the simpler sample preparation. Figure 4 shows the typical sample preparation. A calcite cylinder of 20 mm in diameter was fixed in an adjustment unit to glue a vanadium pin. In order to increase the sample volume of highly deformed materials it is possible to compact sheet sections or segments of a wire. An increased sample volume reduces the total counting time particularly in thin samples. Figure 5 shows a compacted cubic sample glued on a vanadium pin already fixed in a sample holder. The high penetration allows the measurement of local textures at the sample surface as well as inside a compact sample. This method is based on the usage of two cadmium slits as described first by Choi et al. (1979) to cover a well-defined sample volume. The main restriction is given by the minimum sample volume which is necessary to get a sufficient peak to background ratio. This is due to a relatively low intense neutron beam. In contrast to a standard neutron beam a synchrotron beam is much more powerful so that most ofall local texture measurements carried out today by neutrons will be overcome in a near future by synchrotron experiments.
More efficient for the application of neutrons is the non-destructive measurement in order to analyze a sample at different states such as: as received forging or hammering as received deformation annealing NEUTRON TEXTURE ANALYSIS 23 as received low cycling (first step) low cycling (second step) I.n geoscience as well as in recent developments in material science more and more multi-phase systems have to be characterized. The main differences between single-phase and multi-phase systems are: a reduction of the volume fraction; each phase has its own texture; overlapping of different phases; textures are very often weaker in a multi-phase material; multi-phase systems are more often inhomogeneous.
Looking on these differences some practical consequences result. First is that the total counting time for an investigation of a multi-phase (a) material is much longer than for single-phase material. Due to the lower volume fraction this applies also if only the majority phase is investigated. The minimum volume fraction which is necessary to have a sufficient peak to background ratio depends strongly on the scattering behavior. In the case of copper in aluminium volume fraction of lower than 1% copper is sufficient (Brokmeier and Bunge, 1987). Second is an increasing overlapping for multi-phase materials. Particularly in rocks with more than four or five phases having low crystal symmetry overlapping is a current problem. Among others the experimental conditions can help to solve overlapping problems. Table III shows a comparison between the number of Bragg-reflections for A1 and A1203 and different wavelengths. Both materials are very common and not very complicated to measure. It must be pointed out that the number of Bragg-reflections includes total overlapping such as 113 and 11-3 for corundum. In the case of cubic aluminium two or three pole figures are sufficient for a quantitative texture analysis while trigonal A1203 needs between four and six pole figures at minimum. Cr Ka, Cu Ka, Mo Ka and Ag Ka, are standard radiations of common X-ray tubes. A short wavelength like molybdenum and silver shows rather high numbers ofreflections mainly used for crystal structure analysis and not for texture analysis. Due to high energetic X-rays with a penetration comparable to neutrons a wavelength of about 0.2 (such as W Ka) gives a very high number of Bragg-reflections. In the case ofcubic materials this is no problem, in the case of low symmetric materials it is a remarkable problem, and in the case of a multi-phase rock (amphibolite, lherzolithe etc.) a very heavy problem. The three neutron wavelengths given in Table III are available at TEX-2 (GKSS-Research Center) using different monochromators and different monochromator take-off angles. That means, the neutron beam can be optimized to help for deconvolution of overlapping.
Another point of interest is the possibility to use a containment. A number of applications have been carried out to use different types of containments such as a furnace (Juul Jensen, 1992), a cryostat (Elf et al., 1990) or even a box to save light-sensitive materials (Ewald, 1986). Even in the case of very brittle sediments an aluminium box is very helpful in order not to destroy the sample. One important application of neutron diffraction is the nondestructive texture analysis of non-uniform samples as shown in Fig. 7.   Figure 7(a) shows a standard sample for a tensile test. In the case ofsuch a sample it is possible on one hand to investigate the local texture of different parts of the sample and on the other hand to carry out a low cycle experiment measuring the texture at identical sample positions. A second example of a non-uniform sample is a spring steel shown in Fig. 7(b). Samples of 30mm thicknesses and up to 1000 mm in length were investigated to look on texture inhomogeneities.

APPLICATIONS
Since the first texture measurements using neutrons on nickel alloy in 1953 numerous investigations were carried out. This brief collection of some typical neutron applications will focus on TEX-2 measurements. It should be noticed that TEX-2 is a standard four circle diffractometer optimized for texture analysis. That means, between 200-300 samples were measured every year at TEX-2. One ofthe consequences is that sets ofup to 40 individual samples and more for one project can be measured, (see Table IV).
Example 1: YBaCuO-cylinders Among others the effective magnetic levitation force of melt-grown YBaCuO-cylinders depends strongly on the crystallographic texture (Bornemann et al., 1996). The deviation of the c-axis distribution from the cylinder axis should be within 5 with a spread of less than 5. Neutron diffraction is indicated because of the coarse-grained material to get a sufficient-grain statis-

NEUTRON TEXTURE ANALYSIS
29 cylinder was carried out to investigate the influence of the seed crystal on the texture homogeneity. The basic idea is to configure a set of seed crystals to produce an YBaCuO-ring having an optimized texture and an optimized magnetic levitation force.
Example 2: Zinc on steel The investigation of relatively large sample volumes combined with a method of minor corrections allows the measurement of low volume fractions. One of these examples was the measurement of galvanized zinc on steel. Quadratic sections of a 0.7mm steel sheet coated with 8-10tm zinc were cut. Thereafter a cubic sample of 12 12 12 mm 3 was compacted to get a 'neutron sample' of 1728 mm 3 which consists of 98.6 vol% Fe and 1.4 vol% Zn.
Simultaneously the bulk texture of zinc and iron were measured.
Figures 9(a) and 9(b) show the recalculated pole figures of Zn (002) and Fe (110) using the series expansion method. For more details of this investigation see Brokmeier (1993).
Example 3: AgBr-AgCl-Ag The basic idea of this silver-halide composite was to use the optical properties in the mid-infrared region (MIR-fibers). On one hand the technique to produce such fibers had to be developed and on the other hand a non-destructive method to control the crystallographic texture particular of the AgBr core had to be applied. In the case of neutron diffraction three main problems  must be solved. First is the relatively high attenuation of silver, second is the low volume of the thin fibers and third is the overlapping of a number of reflections. Table V gives a summary of the sample characteristics after extrusion, cold drawing and annealing. For details of the graded fiber production and of the texture measurement see B6cker et al. (1995). Figure 10 shows the inverse pole figures of the extruded material, the first step in the processing line. The calculations were performed using the iterative s_eries expansion m__ethod ISEM (Dahms and Bunge, 1989). ISEM allows on one hand to reduce the number ofpole figures which are needed for an ODF calculation and on the other hand to separate overlapping reflections.

CONCLUSIONS
It can be summarized that neutron diffraction texture analysis has been developed to the standard method in bulk texture analysis. This is due to the high penetration of neutrons. Consequently the advantages are directly connected to the high penetration such as the investigation of coarse-grained materials, the measurement ofcomplete pole figures, the neglection of surface roughness in sample preparation, the nondestructive texture measurement of non-uniform samples and the investigation of multi-phase systems. Potential candidates for neutron texture analysis are all geological materials. But also in materials science 32 H.-G. BROKMEIER a number ofapplications need, up to now, neutron diffraction using high penetration. Future developments will lead to a combination of the usage of high energetic X-rays and neutrons to study bulk materials. High energetic X-rays are excellent in the study oflocal textures ofa bulk material while neutron diffraction has still some advantages in the average texture of the bulk.
The main disadvantage of neutron diffraction as well as of high energetic X-ray diffraction against conventional techniques is the radiation source. Owing to the limited number of sources and of course to the limited number ofdedicated instruments open for texture analysis the application will be restricted.