It is with a profound sense of the historical context of this morning that I share this podium with three people who are very much responsible for our ability to rapidly identify and characterize solids. In order to do such solids analysis, we need an instrument that provides the necessary diffraction data and a set of reference patterns for comparison to those we measure, along with a means of rapidly accessing those patterns to make our identifications. I do not need to tell this audience that I am alluding to the diffractometer developed by our Session Chairman, Bill Parrish, in the 1940's and to the milestone publications by Hanawalt, Rinn and Frevel in the 1930's that gave us the first 1000 patterns of the database that was to become the Powder Diffraction File, and the search method later called the Hanawalt Method.

A condensed chronology of computer SEARCH/MATCH programs for the analysis of multiphase crystalline powders is presented covering the years 1965 to 1981. For this period the various algorithms for searching a data base of standard powder diffraction patterns, {ds, (I/I1)s}, are predicated on empirical “fingerprint” matching of the experimental powder data, {dν,Iν}, with one or more of the standard patterns. Within the past year an interactive computer SEARCH/MATCH program has been developed based on a structure-sensitive SEARCH procedure.

The evolution of computer SEARCH/MATCH procedures undoubtedly will continue and will be influenced by marked improvements in the quality of digitized, machine-readable, powder diffraction data. Three conjectures are offered on future developments in x-ray diffractometry.

First, let me say that I am deeply honored by the action of my friends of the Joint Committee in establishing the Hanawalt Award. I have greatly enjoyed my association over the years with the members of the Joint Committee, So many brilliant and talented people have contributed so much to the work of the Joint Committee, By comparison, my contributions have been minor and I feel very humble to be so honored. I want to thank you all for your part in the proceedings today. It is a very high point in my life.

This paper describes the procedures used to retrieve JCPDS powder diffraction data by certain characteristics. These characteristics may include chemistry, mineral name, highest intensity dspacing, largest dspacing, PDF number, etc. The storage scheme used for the powder data and the procedures used to enhance the retrieval speed are described.

This paper describes a new approach to the interpretation of spectra with ‘fuzzy sets’ and a fast search algorithm for spectrum libraries, using a combination of this theory and the inverse search method. The concepts are applied in a computer program named CIF (Compound identification with Inverted search and Fuzzy sets) which is used in the current application for the interpretation of X-ray powder diffraction signals. The methods are readily applicable to other kinds of spectra such as IR, UV, chromatrography etc.

The CYBER 205 has a number of features applicable to the manipulation of the powder diffraction file (PDF). The FORTRAN language has been extended to include many vector procedures and calls to internal system procedures (e.g., the system clock, is accessible and timing of routines is accurate to a microsecond). A VECTOR is specified by its address and its length. A pointer to a vector, a DESCRIPTOR, can be assigned and then used in arithmetic and relational expressions. A BIT vector is a binary switch such that a bit value (ON or OFF) can control the disposition of a corresponding element of a vector. The maximum size of a vector is 65535 elements.

This method for precisely determining the background level of broadened x-ray profiles assumes that the background under the peak can be approximated by a straight line, and the decay in the tails several profile widths from the peak follows a Cauchy-like law. It uses a plot of I(s)s2 vs. s2 where I(s) is the intensity and the scattering vector s = 2cosθo sin(θ-θo)/ƛ where θo is the centroid, θ the Bragg angle and X the x-ray wavelength. The true background is determined by the slope of the linear portion of the plot and its extension to s2=0 gives the rate of decay of the tails. Results on synthesized and experimental profiles show that the method is useful for Warran-Averbach analysis, because it avoids spurious oscillations and the “hook” effect in the plot of the corrected Fourier coefficients, and makes it possible to correct for overlapping tails of adjacent reflections.

A comprehensive study of derivative methods for the peak search analysis of X-ray diffraction data was made to determine the relative merits of the methods. The peak positions were best determined by the cubic first derivative method which had an intrinsic error ≤ 0.001°, and random error ∼ ± 0.003 ° to 0.02 ° depending on the counting statistical noise. The quadratic/cubic second derivative method had the highest resolution with a separation limit ≥ 1/2w (w = full width at half maximum). An effective algorithm combining the cubic first derivative and the quadratic/cubic second derivative methods was developed for high precision and resolution. The method uses a full screen menu for parameter selection, and the entire peak search analysis including peak identification and position determination, and graphic and numeric display of results at the color terminal is completed in a few seconds using a time sharing mode on an IBM 3083 central processing unit. The combined derivative method should be also applicable to other spectra such as gamma-rays, X-ray fluorescence, optical, infrared, ESCA, Mossbauer, etc.

The combined derivative method (accompanying paper) was tested with a large number of experimental patterns to illustrate its use in various difficult problems commonly arising in peak search analysis of X-ray diffraction data. Patterns obtained with various step sizes, resolution, counting statistical noise, and profile widths were used. The precision in 2θ determination and overlap resolution are in good agreement with those previously obtained from calculated profiles, raise identification of noise as diffraction peaks was eliminated by using a convolution range proportional to the full width at half maximum. Peak search results (both 2θ and intensity) were also compared to those obtained by profile fitting to illustrate the different characteristics of these two methods.

Of the roughly 3,000 known mineral species over 290 or 10 percent are phosphates. This report deals with a recently initiated project, the purpose of which is to produce better X-ray powder diffraction standards than currently exist for selected phosphate minerals. The project involves evaluation of existing X-ray diffraction standards for phosphate minerals, and collection of new data from species for which current data are judged unsatisfactory.

Among powder diffractionists, there is a concern about establishing criteria for testing existing and new computer methods for searching the Powder Diffraction File (PDF) data base. A large number of papers on this subject have been presented at recent Denver Conferences and at other meetings, such as the American Crystallographic Association.

The PDF now exceeds 42,000 patterns, with an average of about thirty lines per pattern. Thus, it contains approximately 2.5 million numerical values. At its current rate of growth, the PDF will double in less than twenty years.

A method for computer simulation of X-ray powder diffraction patterns which are identical to those obtained experimentally is described. The calculated pattern is generated directly from the d's (or 2θs) and intensities of the phase(s) and is based on a profile fitting algorithm which uses the instrument function to form the profile shapes at all reflection angles. Examples of simulated patterns of mixtures, line broadening, linear and amorphous backgrounds, and counting noise are given.

This paper summarizes experiences available for the measurement of lattice strains in different materials with different wavelengths to evaluate stresses by means of X-rays. The recommendations given are based on previous statements. Some principles of fundamentals of X-ray physics for the recording of interference lines with Ω and ψ-diffractometers are dealt with. Methods applicable for the determination of the peak position of the interference lines, the assessment of linear and non-linear lattice strain distributions, and tine calculation of stresses are outlined. For iron, aluminium, copper, nickel and titanium the constants for practical X-ray stress evaluation (XSE) and the parameters of measurement are tabled.

The state of the art of stress evaluation on materials having non-linear lattice strain distributions is presented.

New results on heterogeneous materials with measurements conducted on both phases of the material show compensation of the shear stress components σ13 in ground surface layers of (α+β) brass. There is only the compensation of normal components σ11 of (α+γ) steel after plastic straining.

The fundamental aspects and the evaluation of macro- and microresidual stresses on materials having preferred orientation are broadened. The use of Mo-Ka-radiation shows linear lattice strain distributions, as a result of minimizing the influence of micro residual stresses causing oscillations. The interplanar distance- or strain-polefigure shows similarities with the intensity polefigure.

The knowledge of the theoretical influence of stress distribution with depth from the surface of the material is extended. The experimental procedure should use either different radiations having different penetration depths or a low-penetrating radiation in combination with removal of surface layers.

Since the method of X-ray multiaxial stress analysis (ψ-splitting problem) was proposed by Dolle and Hauk, their method is often utilized for residual stress evaluation on a processed surface. However, as pointed out by thanselves, a theoretical problem still remains on the assumption of the stress state. Namely, the effect of ψ-splitting is impossible unless stress gradients with respect to the direction of the depth in σ13 or σ23 are present, because these components and σ33 have to vanish at the outer surface. Actually, we often find the 2θ vs. sin2ψ relations which do not agree with their theory. In this paper, we proposed a new method for X-ray multiaxial stress analysis in which the effect of stress gradient was considered. The basic equation of this method was solved by means of the integral method proposed by Lode and Peiter. The validity of the present method was proved through a numerical simulation and an experiment.

The physical meaning of non-linearities in “d” vs. sin2ψ lines, encountered in X-ray measurements of surface residual stresses in polycrystalline materials is investigated. It is shown that when oscillations are present in any one reflection, switching to another reflection to obtain a straight line in “d“ vs. sin2.ψ is feasible only under very special conditions. We also discuss the effect of “quasi-homogeneous” strain distributions and investigate the effects of ψ-range on the accuracy of X-ray residual stress measurements when ψ-splitting” is present. A new geometric error is also discussed that can not be detected by the “annealed powder” method often used for alignment.

The study of the mechanical behaviour of materials by X-ray measurements in a classical laboratory is limited by the possibilities of the X-ray tubes used. Some crystallographic planes are not conducive to good diffraction conditions and the radiation characteristics' may not be optimum. The use of synchrotron radiation resolves many of these problems by providing a continuously variable wavelength which allows measurements of stress and stress gradients to be carried out in very good conditions. The high intensity and perfectly monochromated radiation with a small beam divergence are very helpful for microstrain measurements.

Residual and applied stresses (σij) are often measured via X-ray diffraction, by calculating the resultant elastic strains (ϵij) from the measured change in interplanar spacing (“d”). This method is non-destructive, reasonably reproducible (typically ±14 MPa), can be carried out in the field, and is readily automated to give values to an operator-specified precision , Let Li represent the axes of the measuring system with L3 normal to the diffracting planes, and Pi represent the sample axes. These axes are illustrated in Figure 1. In what follows, primed stresses and strains are in the laboratory system, while unprimed values are in the sample system.

Multilayers composed of many thin films of PbTe and Pb1-xSnxTe on BaF2 substrates were grown epitaxially by hot-wall-vapor deposition. In order to investigate the fraction of the total misfit (2.5x10-3 at x=O, 12) accommodated by misfit strain we have performed strain measurements on these superlattices by two different X-ray diffractometer techniques. We also report on substrate induced strain due to different thermal expansion coefficients of films and substrate. For film thicknesses smaller than 300 nm there is clear evidence for almost complete accommodation of lattice mismatch by misfit strain. Below room temperature the substrate induces a tensile strain which is comparable to that of the misfit strain.