The x-ray laser has recently been described (1) as a potentially attractive tool for chemical analysis, both as an improved source for ESCA and also for radiochemistry. The intensity (∼1016 photons/pulse) and associated short exposure time (10-15-10-12 sec) promised are the most attractive features, as is the narrow line width (0.1-0.01 eV). [Such photon fluxes would be equivalent to 1-12 days of exposure with projected electron storage rings (1).] Improved resolution in chemical shift determinations by ESCA are promised as are observations of short-lived excited states and rapid reaction rates in highly-concentrated chemical species. For general utility and optimum analytical capabilities, a photon energy in the 1 keV range from a 10 Å laser would be adequate and is projected (2) to be within the realm of presently conceived technology. Neither a high degree of collimation nor coherence is considered essential although the latter is associated with particularly narrow lines, so that plasma-type noncavity amplified spontaneous emission (ASE) devices (2) will prove adequate.

Current versions of the J.C.P.D.S. Powder Diffraction File Include some 26,000 standard patterns and this number is Increasing at the rate of around 1500 patterns per year. Although the actual number of patterns still represents a relatively small fraction of the total number of phases possible, it does include the majority of well known phases. A variety of search/match schemes have been designed under the auspices of the Joint Committee on Powder Diffraction Standards and these include data retrieval methods based both on hand and computer searching.

Since the majority of analytical problems involve multiphase materials, one of the greatest problems in qualitative identification is the “unscrambling” of the various phase patterns to allow matches to be made with standard patterns of pure materials.

A rapid analytical procedure for simultaneous determination of Ca, Mn and Fe in agglomerate and slag has been developed.

A 100 mCi annular source of 238Pu has been used for excitation and Si(Li) detection system with ADC coupled with minicomputer for data acquisition and processing. A modified Heinrich-Rasberry procedure has been applied for analytical results evaluation. The measurement of each sample consumes about 100 s and the data processing time is 20 s approximately. The peaks were integrated in a fixed number of channels and linear background was subtracted. Seven standard samples both of slag and agglomerate have been used for system calibration.

X.R.F. analysis of steelmaking slags has been done routinely since the 1960s. Fused buttons or pressed powdered pellets are used.

Under certain conditions steelmaking slags may contain appreciable amounts of free lime and metallic iron droplets. These can cause gross inaccuracies in the result of the X.R.F. analysis as the normal correction procedures used in X-ray fluorescence techniques can not compensate for them.

A special sample preparation technique as well as appropriate analytical procedures will be discussed. Examples of slags with different levels of free lime and metallic iron will be shown.

The casting of iron is largely a chemical process: that is, the chemistry of molten iron greatly influences the properties of finished castings. Other parameters - casting design, mold integrity, inoculation practice, pouring temperature, and shakeout procedure - are also vital to consistent production. But if the chemistry is wrong, the probability of defective castings is high. For this reason, control of chemical composition has long been a major objective of foundrymen.

To be useful for either process control or quality assurance checking, analytical work must be reliable and accurate. Data should also be available in a minimal period of time to make possible the on-line adjustments necessary to achieve proper chemistry, and thus the desired properties, in the final castings.

The reliability of energy dispersive x-ray fluorescence data for the analysis of low-alloy steels is reported for chromium, manganese, nickel, and molybdenum. The accuracy and precision of the analysis were evaluated for two mathematical models, linear and multiple regression, using twelve NBS standard reference samples. Data on results are furnished.

Comparable uranium sensitivities and detection limits can be obtained by either energy- or wavelength-dispersive x-ray fluorescence (XRF) analysis of L x-rays, and by wavelength-dispersive analysis of M x-rays. XRF methods are more sensitive than 252Cf neutron-induced γ-ray analysis or direct photon analysis of rapid-equilibrium uranium daughters. A laboratory method is demonstrated for quantitative uranium determination in unknown geological matrices without using ore standards.

Silica gel treated with a commercial silylation reagent (Dow-Corning Z-6020) has been found to extract uranium from carbonate solutions. The materials used are simple to prepare and once the uranium is extracted, x-ray fluorescence may be employed to determine the uranium directly on the solid. A dependence of the extraction upon carbonate concentration is observed. A lower limit of detection of 2.46 μg of 0.12 ppm uranium from a 20 ml sample is obtained.

A program of ore reserve and grade determination on a major bauxite deposit has recently been completed. More than 2,000 samples were chemically analyzed for available Al2O3, reactive SiO2, loss on ignition; and by energy dispersive XRF for a total Al2O3, SiO2, TiO2, and Fe2O3. As compared to standard wet chemical analytical methods, the XRF analyses proved to be cost and time effective, and the observed analytical precision was superior.

An X-ray fluorescence (XRF) method has been devised for the rapid quantitative determination of 16 minor and trace elements in geological materials. This method, a modification of a direct dilution method (1, 2, 3), uses a sample-to-binder ratio of 85:15 for sample preparation. Pellets prepared by this method are durable and do not deteriorate rapidly when exposed to high X-ray irradiation. Interferences and matrix effects are successfully corrected in actual analyses by employment of multiple linear regression equations.

Accuracy and precision have been improved over the method previously used in this laboratory. Detection limits have been lowered for Zn, Rb, Y and Zn approximately by the factor of 2. When interelement corrections are made, the XRF values are found to be in good agreement with the preferred chemical values for the 19 international silicate-rock standards.

Because the cost of the large number of standards required for quantitative X-ray fluorescence analysis is prohibitive, a mathematical procedure, the “fundamental parameters method” is gaining popularity. A major problem in using the method is the limited knowledge of the spectral distribution of the primary radiation from the X-ray tube, the need for a large computer memory which makes it difficult to use with minicomputers and uncertainties in the basic parameters. A monochromatic X-ray source was developed to avoid the problem of the spectral distribution. This permitted simple calculations for the calibration and a simple algorithm to transform the intensities to mass fraction or coating thickness. The calculations were made with a slide rule calculator.

The industrial use of X-ray fluorescence (XRF) coating gauges is a well established practice. For homogeneous coatings such as zinc or tin on steel, on-line XRF gauges are in use around the world. In continuous coating lines coils of uncoated metal are welded to one another and are coated at speeds between 100 and 2000 ft/min. Gauges are used to read cross-sheet coating profiles and determine the average coating thickness through the length of the coil. The motivation for the use of these on-line X-ray instruments is quality control, particularly coating uniformity, and raw materials savings.

The excitation efficiencies for K- and L- x-ray fluorescence by a Mo transmission tube operated at 30 kV have been measured. K x-ray excitation efficiencies for 14 elements ranging from A1 to Y and L x-ray excitation efficiencies for 12 elements ranging from Br to U are presented. The K x-ray excitation efficiencies are compared to those predicted from theoretical photoionization cross sections and empirical fluorescence yields, and the L x-ray excitation efficiencies are used in conjunction with theoretical photoionization cross sections to derive average L-shell fluorescence yields.

Exact expressions are given for the K and L x-ray fluorescence cross sections. Hole transfer between the L subshells is taken into account and shown to lead to a substantial enhancement of the cross sections for x rays of the L2 and L3 emission series. Absolute values of Kα and Lα cross sections are calculated for all elements and four excitation energies, namely GeK ≙ 10.0 keV, MoKα ≙ 17.44 keV, AgK ≙ 22.5 keV and Am 241 ≙ 59.6 keV.

The Monte Carlo simulation method that has been previously developed and demonstrated for EDXRF analysis with annular radioisotope excitation sources is extended to systems using secondary fluorescer X-ray machines for excitation. Comparisons of the Monte Carlo predictions with experimental results indicate that the modification is valid.

Traditionally, in X-ray fluorescence analysis for the determination of elemental compositions, the fluorescence is measured from the irradiated side of the sample. The composition measurements obtained by this method are sensitive to a gradient in composition as a function of depth of the element being measured. This report presents a simple method for measuring a mean composition for an element segregated in layers applicable to thin samples so that the fluorescence can be measured on the side opposite to that being irradiated. It is shown that for a particular relation between the angle of incidence of the primary radiation and of the detection angle on the opposite side of the sample that a mean concentration can be measured for an element which is independent of the composition in the separate layers.

Results of measurements of the mass attenuation coefficients (μen/ρ) and the dispersion corrections f’ and f” have been given in two recent papers. The data cited in these papers were gained using both a conventional XRF spectrometer and a multiple reflection double crystal spectrometer based on the Bonse-Hart low angle scattering camera. Both of these techniques have deficiencies which limit their usefulness in operation. The XRF technique has been shown to become inaccurate for thick specimens. The multiple reflection spectrometer experiences alignment and stability problems, because, for each wavelength, three different axes must be adjusted. Since the angle of acceptance of the grooved multiply-reflecting elements is less than 10 seconds of arc, any misalignment of one element of the spectrometer with respect to the other causes large variations in the intensity of the beam transmitted through the spectrometer.

Errors pertinent in dual beam absorptiometry have been studied. Five areas are given in detail: 1. Scattering, in which a computer analysis of multiple scattering shows little error. 2. Geometrical configuration effects, in which the shape of the sample influences accuracy. 3. Poisson variations, in which dosage and statistical error are minimized. 4. Absorption coefficients, in which variations in compilations are examined. 5. Filtering, wherein is shown the need for Kβ filtering. A zero filter system is outlined.