High-temperature X-ray powder photography has proved of great value in many directions in the metallurgical field. Some of the successes achieved and problems encountered by the author and other investigators using an established-type camera (British Unicam) for temperatures up to 1400°C will be discussed. This will be by way of specific case histories, chosen both for their own scientific interest and to illustrate the techniques employed. The applications concern phase transformations and the measurement of coefficients of thermal expansion in iron and steels, as well as in some nonferrous metals and inorganic compounds.

A high-temperature X-ray diffractometer specimen holder that fits on a standard General Electric diffractometer is described. The thermal expansion of iridium from 23 to 1720°C is reported as 7.5 • 10−6−6 per °C and diagrams made at 2000°C are shown.

A high-temperature high-vacuum X-ray diffractometer attachment was constructed to observe formation of materials at elevated temperatures. The device consists of a conventional water-cooled compartment, a resistance-type furnace for heating samples to 800°C, and inlets and outlets for gaseous reactants and evacuation. Two sources are available for evaporation of materials which may react with or coat the sample. The sample is aligned by a two-part sample-slit system. The sample temperature is regulated by means of a current-adjusting control unit and a self-saturating reactor with suitably positioned thermocouples. Temperature distribution and constancy over the sample surface were determined.

Tests were carried out with an X-ray diffractometer furnace to determine the temperature variations from point to point in the specimen and the correction to be applied to thermocouple readings. It was found that, under the conditions of the tests, temperature corrections depend on the preparation and properties of the sample. Calibration curves for two standards are given to illustrate this point, and suggestions for improving temperature measurements are made.

Two high-temperature X-ray diffraction cameras are described which have been employed at Battelle to determine thermal expansion of metals and ceramic materials. Specimen preparation and temperature measurement and control are described. Lattice-parameter data vs. temperature are presented for uranium, uranium dioxide, and magnesium oxide.

High-temperature X-ray diffractometer measurements by the Bureau showed that alpha-alumina expands anisotropically with expansion parallel to the c axis approximately 10% greater than that parallel to the a axis. Thermal-expansion characteristics of four different alumina samples were determined up to 1200°C; the results are presented in both tabular and graphical form. The combined random and systematic errors in the measurements are considered less than 3% above 500°C and approximately 2% at 1000°C. No differences in expansion were observed between the samples or between the heating and cooling cycles.

Critical evaluation of the literature showed that published values for thermal-expansion coefficients of single crystals vary over a range of 40% (relative values) and disagreement concerning whether expansion of alpha-alumina is isotropic or anisotropic. Systematic and experimental errors were found to be the principal factor for the variation in results, rather than differences in measuring techniques or in samples.

Published values for polycrystalline aggregates also varied. Two laboratories reported values 20–30% lower than those of the Bureau of Mines, but two other laboratories agree with the Bureau within 3–5% above 500°C.

The temperature dependence of the reflection intensities from copper was examined in order to determine the mean-square vibration amplitudes of the atoms in the metal. The reflections from a prealigned single crystal were examined in a Norelco diffractometer outfitted with a high-temperature chamber designed by Chiotti. The intensities of the (400) reflection for copper were measured at temperatures from 13 to 1000°C. Difficulty was experienced in obtaining reproducible intensity measurements, and it is suggested that this was due to changes in the mosaic structure of the single crystal during thermal cycling.

The high-temperature continuous X-ray diffraction technique has been used at the University of Illinois for several years in our investigation of structural transformations and the development of phase minerals as natural materials are heated. Of particular interest are the effects which chemical impurities have on predicted structural transformations at elevated temperatures.

The formation of mullite and beta-cristobalite as they develop from the clay mineral kaolinite is examined by continuous X-ray diffraction up to 1450°C, and the influence of added impurities either to enhance or to retard nucleation of these phase minerals is shown. The 2θ range containing diagnostic peaks of both the initial material and of the transition and phase minerals which develop on heating was continually traversed, thus providing a constant record for comparing structural changes vs. temperature.

The furnace used is described in detail, and the effects of added impurities on kaolinite transformations are evaluated to show their importance in controlling the thermal history of a natural material under nonequilibrium conditions.

High-temperature X-ray diffraction techniques were employed to study the condensation reactions which occur when sodium orthophosphates are heated to 380°C. Crystalline Na4P2O7 and an amorphous phase were formed first from an equimolar mixture of Na2HPO4·NaH2PO4 and Na2HPO4 at temperatures above 150°C. Further heating resulted in the formation of Na5P3O10-I (high-temperature form) at the expense of the crystalline Na4P4O7 and amorphous phase. Crystalline Na5P3O10-II (low-temperature form) appears after Na5P3O10-I.

Conditions which affect the yield of crystalline Na4P2O7 and amorphous phase as intermediates and their effect on the yield of Na5P3O10 are also presented.

The diffraction, reflection, absorption, fluorescence, and the electronic emission that results from the interaction with ultrasoft X-rays (λ > 10 A) are presented as practical bases for microanalysis. Recent developments on sources and detectors for the ultrasoft X-radiations are described. A preliminary report on a current investigation on low-energy photo-Auger electron analysis and on a new type of low-energy electron spectrometer is also presented.

A mathematical method has been developed which reduces nuclear-reactor gamma pulse-height spectral data to the form of differential energy spectra through the use of a high-speed computer machine. In essence, the method consists of a least-squares fit of weighted multichannel analyzer data and the utilization of curve-smoothing parametrization. The least-squares approach tends to reduce the magnitude of data that must be handled, i.e., reduces the order of matrix involved. Weighting is used to obtain fractional deviations for minimization by lease squares and thus obtain a satisfactory fit throughout the entire channel range. The parametrization smoothes the reduced data by making use of the fact that reactor gamma spectra can be represented by the product of an exponential and a polynomial. Difficulties that arise when pure matrix inversion is applied have been obviated, and the advantage of high-speed data reduction is gained through the use of an IBM 704-7090 computer program. Error analyses have been undertaken, and data have been reduced for comparative purposes. Results are included in the presentation of the investigation.

Results of recent experiments on microanalyses with an X-ray projection microscope will be reviewed. As the use of monochromatic radiation is imperative, spectral analyses of the point source were carried out. A simple stationary divergent-beam-type transmission spectrograph was used. The shape and size were miniaturized to fit in the specimen holder of the Norelco projection unit. Emission spectra from clean, targets and the fluorescent radiation emerging from a 10- to 50-μ-diameter spot of contaminated targets and of two-layer targets will be shown. The spectra were recorded photographically. In addition, a proportional counter, in combination with the R.C.L. 128-channel pulse-height analyzer, was used for measuring the ratio of “white” to line radiation.

A fundamental quantitative treatment of the heterogeneity effects in X-ray fluorescence has been made. The theory predicts that the grain-size effect appears only in a limited region of grain sizes which depends on the wavelength of the primary radiation and the nature of the compounds in the mixture. With monochromatic radiation, the fluorescence intensity showed increase or decrease by a factor of a few units as grain size is decreased, A factor as large as 12, the theoretical value, has been observed in one particular experiment. Usually the grain-size effect can be eliminated by intensive grinding. For the light elements fine grinding is disastrous if long wavelengths are used. By an appropriate choice of the wavelength it is possible to eliminate the effect even without grinding. The mathematical treatment also predicts, but less rigorously, a grain-size effect in X-ray diffraction.

The effect on the fluorescence intensities by changes in the chemical composition of the grains that contain the fluorescent element is predicted by the theory.

These findings are discussed in relation to the analysis of elements when polychromatic beams are used.

A simple and accurate graphical method is described for converting the composition of ternary alloys from weight per cent to atomic per cent, and vice versa. By making certain simplifying assumptions, the principles of the method are extended to the approximate determination of ternary alloy compositions from the count fractions obtained from fluorescence X-ray spectra. An example is given for the not-so-favorable system bismuth-germanium-tellurium. On account of the simpler excitation conditions, the same procedures may also be applied to X-ray microprobe analysis.

Most considerations of excitation conditions either have ignored the target material of the X-ray tube and the resulting spectrum of the primary X-ray beam or have dealt with it in very general terms, making a few broad assumptions regarding its effect on the final result. This is quite understandable since any attempt to handle the quantitative relationship between the primary spectrum and the resulting fluorescent spectrum results in extremely complex and unwieldy mathematical expressions. However, with the number of different primary sources now available, it is possible to evaluate these complex expressions in terms of actual data and permit more reliable values to be placed on the relative contributions of the components in the primary spectrum.

A discussion is included of results obtained with X-ray tubes with chromium, molybdenum, tungsten, and platinum targets together with a few typical applications to actual X-ray spectrochemical procedures. Also discussed are the other variables which play a part in the excitation of fluorescent X-ray spectra, such as the primary X-ray tube voltage and the window of the X-ray tube.

The application of the X-ray absorption-edge technique was extended to the determination of cobalt in aqueous and alcoholic solutions containing a wide variety of impurity elements. In the procedure developed, secondary radiation from a 50% copper-nickel alloy is passed through an absorption cell filled alternately with the solvent and the sample solution. The transmitted Intensities of the Kα lines for copper and nickel are measured, and the concentration of cobalt is determined using accepted absorption principles. The K absorption edge for cobalt occurs at 1.604 A, restricting cell construction materials and solvents to those containing light elements with low X-ray absorption characteristics and also limiting the path length of the cell.

Cells of 0.16- and 0.34-cm path length were used in the analysis of aqueous and alcoholic solutions, respectively. With the 0.16-cm path-length cell, relative standard deviations of 4.6 to 0.5% were obtained for cobalt concentrations ranging from 1.00 to 10.00 mg/ml for known aqueous solutions that contain various known concentrations of nitric acid. With the longer path-length cell, relative standard deviations from 1.8 to 0.46% were obtained for cobalt concentrations in the same range in known alcoholic solutions containing various known concentrations of nitric acid. The standard deviation of determining the blank is 0.043 mg of cobalt per milliliter for the 0.16-cm cell and 0.016 mg of cobalt per milliliter for the longer cell.

A Norelco X-ray spectrograph with a three-position head was used in these analyses. Less than 5 min is required to convert this instrument from normal fluorescence operation to absorption-edge analysis. Approximately 15 to 20 analyses can be performed daily.

Techniques of mineral analysis by fluorescent X-ray spectrography can generally be classified on the basis of (1) extensive sample preparation and simple instrumental examination, (2) simple sample preparation and extensive instrumental examination, (3) extensive processes for both, or (4) simple processes for both major phases of the technique. The latter type of technique is usually the most commensurate with the concept “maximum information with minimum effort.”

Suitable analytical results with multielement calibration systems for random mineral samples can be obtained if there are valid considerations of the relevant absorption and enhancement effects or if the systems are based on general procedures to minimize absorption and enhancement effects. Preliminary investigation indicated that control of sample mass and use of thin films of the mineral sample is a convenient and simple method of minimizing absorption and enhancement effects.

Successful application of thin-film samples with known masses in a multielement calibration system is highly dependent upon details of sample preparation and comprehensive studies of these details will be discussed.

The calibration system is suitable for the determination of minor and major concentrations in the sample, but it is not recommended for determination of trace concentrations.

In X-ray spectrochemical determinations, the surface preparation of metal samples is important since the emitted secondary radiation is absorbed except for that emerging from the immediate surface. Surface preparation also has been found to be a critical factor in eliminating or at least in minimizing the effects of certain metallurgical changes among or within samples on the X-ray determinations. For example, a change of particle size of an undissolved constituent, or a change in grain size, may markedly influence the X-ray analysis, depending on the surface preparation. A metallographic polishing technique, whereby the sample surfaces were finished with ¼μ diamond dust, was reported for the determination of lead in leaded steel (Kilday and Michaelis, in press). This investigation has been extended to other materials such as hypereutectic silicon-aluminum alloy and white-cast iron. A comparison is made of the effects of different surface preparations on X-ray determinations.

Quantitative analysis of major elements (Na, Mg, Al, Si, K, Ca, Ti, and Fe) in rocks by classic wet-chemical procedures, in addition to being a costly and laborious process, has been shown to be subject to significant errors. Mineralogists and petrologists, increasingly concerned with determining small elemental differences in rock bodies, have turned to more rapid and precise methods in the field of spectral analysis. The development of X-ray spectrographic techniques in the light-element range has therefore been of particular interest. This paper summarizes the results of a test of the precision and sources of error in the X-ray method as applied to major elements in granitic rocks.

The test, using commerically available equipment and the fusion technique of specimen preparation, was designed for a five-way, completely nested variance analysis. Sources of variance are (1) different lengths of fusion time, (2) replications of a single fusion time, (3) replications of briquetting the fused specimens, (4) replications of individual runs in the spectrograph, and (5) replications of individual spectrograph readings of the Kα line intensity. A total of 64 readings per element were obtained and the significance of each factor tested for each element. Greatest sources of error are shown to be in factors 2 and 4. The total variance for each element is expressed as per cent relative deviation of counts per second. Calibration curves of natural and synthetic granitic rocks, fused with dilutions ranging from 13 to 36% rock in borax, provide conversions from relative deviation in counts per second to relative deviation in weight per cent. These values indicate that the precision of the X-ray method is directly comparable with wet chemistry for Si, probably better for Al, and distinctly superior for K, Ca, Ti, and Fe. Of the major rock-forming elements only Na and Mg are presently beyond the scope of the fusion method.

A large amount of data has been treated statistically to determine daily and long-range variations which occur when replicate determinations are made on low-alloy steel samples. The experiment reported here was designed and run to evaluate the results obtained using an RCA Siemens X-ray spectrograph. As part of a quality-control method for the analysis of low-alloy steels, it was desirable to know the confidence limits for within-day and day-to-day replication. This paper reviews the results of this study and represents an outline of future work planned.