In the course of the past few years, X-ray and spectroscopic methods of analysis have found an increasing usefulness at the Dow Chemical Company. There are a large number of different types of problems on which information can be obtained by the variations of apparatus and technic which are possible in these two fields. It is not the purpose of this paper, however, to discuss these methods or applications in general, but to describe in some detail a scheme of classifying and using X-ray diffraction patterns which has been found very helpful in one particular application of X-rays — namely, that of identifying unknown substances by means of their Hull powder diffraction patterns.

The inherent power of X-ray diffraction as a practical means of chemical analysis was pointed out a good many years ago. Having a different theoretical basis and depending upon an entirely different technic than other methods, it would be expected to supplement the information to be obtained from other methods and, at times, to be applicable where other methods are not suitable. It appears, however, that the use of this method has not increased at a rate commensurate with its unique and valuable features, and that it is used by relatively few academic and industrial laboratories.

The objectives of this paper are: 1. to up-date the manual search system and index book described in 1936 and 1938; 2. to introduce a “work form” which serves to guide the analyst through the procedural steps involved in using the manual search index book and 3. to review briefly the literature on manual search/match systems.

Two of the basic features of the 1936 design of the search index book were first, to divide the “d” range into arbitrarily sized “groups” and “subgroups” rather than using a scale of continuously decreasing “d” values, and second, to use the “d” values of the strongest lines of the diffraction pattern in order of decreasing intensity in making the entries of the standard patterns in the search index book. These two features are still the basis of the design of the 1986 Hanawalt Search Manual published by the JCPD S International Centre for Diffraction Data.

The original use of the search index book at Dow Chemical was to lead the analyst to corresponding patterns among the thousands of Debye films which had been produced and placed on file. The “d” values and intensities of only the three or four strongest lines of the pattern were measured. Comparisons and indentifications were then made visually simply by holding the concerned films together in juxtaposition. Unfortunately, attempts to reproduce usable copies of such Debye films for general reference have not been satisfactory. Therefore, for general usage the diffraction data carried by the film negative or also the data from a diffractometer trace are recorded using the numerical values of the “d” spacings and intensities. These tables of numerical data are then used to represent the diffraction patterns. The search index book for general usage must therefore deal entirely with these numerical values.

I have made use of X-ray powder diffraction patterns for over sixty years. In the summer of 1922, in anticipation of my becoming a graduate student in chemistry, I read the book “X-Rays and Crystal Structure,” by W. H. and W. L. Bragg. Then in September 1922 I arrived in Pasadena, and immediately began to learn how to determine the structure of a crystal by a study of the X-ray diffraction pattern from Roscoe Gilkey Dickinson, who was the first person to have received a Ph.D. degree from the California Institute of Technology (1920). The procedure in use in Pasadena started with the preparation of a photograph showing lines obtained by Bragg reflection from a developed face of a large crystal with monochromatic radiation, usually molybdenum K alpha and beta. Measurement of the angle of reflection gave a set of possible values for the length of the edges of the unit of structure, usually of a cubic, hexagonal, or tetragonal crystal, since the methods were not powerful enough to permit the evaluation of more than two or three parameters. The next step was the preparation of Laue photographs, and their analysis. This was a powerful method, which often led to the correct structures.

A new, convenient program has been designed and implemented on a VAX computer to facilitate the use of the Johnson/Vand program (Version 21) for identifying components in crystalline mixtures by X-ray diffraction. (The data base of references is distributed solely by the JCPDS – International Centre for Diffraction Data.) This new program uses an easy-to-follow conversational mode of communication for setting up the input file for the identification program from a remote terminal. The program is menu driven with screens for input of sample information, for change of default computational parameters, and for handling the experimental diffraction data. Many of the input screens can be readily bypassed when the default parameters are acceptable. An editor feature is provided for viewing the final input file and for correcting the diffraction data.

Crystal data for RuIn3 are reported. The material is tetragonal, , with a = 6.9983 (3), c = 7.2440 (4) Å, Vol = 354.78 Å3, Z = 4, Dc = 8.34 gm/cm3. It is isostructural with the compound CoGa3. The crystals of RuIn3, which form long needles, display electrical resistivity characteristics of a typical metal.

Intensity data were obtained from a diffractometer trace taken with CuKα radiation, λ = 1.54178 Å. The structure was refined by least-squares calculations to R1 = 0.116 with 53 observed reflections. Powder pattern data are also given.

The Semi-manual Structure-sensitive SEARCH-MATCH procedure [Frevel, (1982), Anal. Chem. 54, 691–697] has been automated for the Hewlett-Packard HP 3354 Computer. The software is coded in the interpretive HP LAB BASIC II language. A core file of 1026 selected powder diffraction standards has been compiled covering the common crystalline phases encountered in industry, the ubiquitous crystalline minerals, and prototypes of the more frequently found crystal structures. Solid solutions can be readily identified.

An improved method of alignment of a Philips diffractometer equipped with a theta-compensating slit is described. The method employs a special fluorescent screen which is flat to better than 10 microns. The procedure results in accurate alignment of the theta-compensating slit at low angles and assures reliable intensity data down to one degree two theta or less.

μPDSM, a powder diffraction search/match system, is derived from the original interactive time-sharing system written for the NIH/EPA Chemical Information System. Transformed and essentially rewritten, it still searches the entire JCPDS database, but on an inexpensive IBM PC microcomputer. In the transition from mainframe to micro, μPDSM has lost none of its speed or performance. Indeed, the basic discriminating power of the search/match algorithms had to be improved, since the Powder Diffraction File is some 50% larger than it was when The CIS version was in operation. While μPDSM will solve typical problems on a “push-button” mode, it provides an environment that promotes optimization of search parameters by the diffractionist, and provides all of the subfile and chemistry functions associated with larger systems.

Patterns useful for identification are obtained by automated diffractometer methods. The lattice constants from the experimental work are refined by least-squares methods; reflections are assigned hkℓ indices consistent with space group extinctions. Relative intensities, calculated densities, literature references, and other relevant data are included.

Features of the powder diffraction patterns of known polytype-like modifications of Ca3GeO5, such as 2H, 9R and 24R types, and a monoclinic polymorph have been studied by means of the patterns calculated based on their crystal structures. The result has provided keys to identify them when they coexist in the same synthetic product. Accounts have been given on the crystallographic features of the building layer and rules for layer stackings of the modifications in general. Any modification may then be constructed to predict its powder diffraction pattern.

A strategy for qualitative X-ray phase analysis of complex multiphase solids is illustrated through the analysis of eight fixed-bed gasification ashes. Coal from the same lignite formation was gasified in three different reactors. Despite the differences in gasifiers and their operating conditions, all of the ash specimens had the same basic crystalline phase assemblage: silicates and aluminosilicates (merwinite, dicalcium silicates, melilite, nepheline, carnegieite and a sodalite structure phase), oxides (ferrite spinels, hematite and periclase), calcite and residual lignite minerals (quartz, plagioclase). The identification strategy relied on a bulk chemical analysis of each ash and used manual methods with the JCPDS-ICDD search manuals for the Powder Diffraction File. Grain picking, size separations, magnetic enrichment, separation by hardness, dissolution of water soluble phases, and further crystallization of the ash through heat treatment all figured in the identification strategy. Identification of more than a dozen minor and trace phases resulted from these additional methods. Several trials with an automated search/match system demonstrated the difficulties of analyses of such complex phase assemblages. The need for SEM characterization of the ashes to accompany the XRD analyses is discussed.

Crystals of green to bluish green rosasite from Tsumeb, Namibia are monoclinic, space group P21/a, with refined unit-cell parameters a = 12.873 (3), b = 9.354 (3), c = 3.156 (2) A, β = 110.36 (3)°, V = 356.3 (5) A3 and a:b:c = 1.3762:1:0.3374, with Z = 4 and Dx = 4.15 g./cm.3. Crystals are twinned by 180° rotation about a*. A fully indexed powder pattern is presented. Electron-microprobe analyses, average (wt. %): Cu 32.9, Zn 23.7; this suggests an analytical formula of (Cu1·18 Zn0·82) (CO3) (OH)2. The rosasite is optically biaxial negative, α = 1.673 (1), β = 1.796 (3), γ = 1.811 (3), 2Vx = 33 (5)° (calc. 36.4°), with X = c, Y = a* and Z = b.