In designing a radioisotope, energy-dispersive (proportional counter) x-ray fluorescence spectrometer for elemental analysis of geologic specimens, a theoretical framework has been developed to model instrument response. This model is based upon the fundamental physics of x-ray excitation, absorption, and scattering, and employs the most modern available values of the applicable physical constants. The model includes matrix absorption and enhancement effects. By explicitly including scattering in the model and the measurements, element concentrations can be calculated from the shape alone of the x-ray spectrum and the presence of elements having non-observable fluorescences can be detected.

The computerized fluorescence radiation induced energy dispersive analyzer (FRIEDA) (1) described earlier uses an x-ray beam with a well defined energy for the excitation of fluorescence radiation, and an Si(Li) detector to measure the total x-ray spectra emitted. Such a system can also simultaneously provide supplemental data for the determination of the dry mass and the sample mass absorption which is necessary for accurate quantitation of the results. This instrumental capabillty has been utilized in the measurement of the trace elements iron, copper, and zinc in serum.

Known amounts of two elements are thoroughly mixed with the sample. One element has a ‘high energy’ K line, the other a ‘low energy’ K line. The ratio of these intensities, in the absence of absorption, is a known constant and dependent only on the relative amounts of the respective elements, and on the energy of the exciting radiation. Whenever absorption is present, the ratio will change in a manner directly related to the mass absorption of the sample for these radiations.

Resin-loaded papers composed of approximately 50% cellulose and 50% ion-exchange or chelating resin provide an ideal matrix for many X-ray spectrographic analyses. Standards are prepared by multiple filtration of solutions of known composition through the paper to achieve quantitative collection or by the use of a radiotracer as a monitor for nonquantitative collection. Solutions prepared fram unknown samples are processed in the same manner as the standards.

Advantages of the resin-loaded papers are: reduction of interelement effects because standards and unknowns are present in a similar low X-ray absorbing matrix; physical parameters such as metallurgical history, grain size, and surface preparation are eliminated; and sampling errors are significantly reduced and sensitivity greatly increased by concentrating trace elements separated from large samples.

Application of these papers to a variety of metallurgical, geological, and water samples will be summarized. The possible use of resin-loaded papers as standards for air pollution monitoring will be examined.

Functional groups which form chelate compounds with metal ions have long been of analytical usefulness. Such compounds have served as titrants, precipitating reagents, extractants, and as ionogenic groups on ion exchange resins. These ion exchange resins often have high distribution coefficients which enable the quantitative extraction of metal ions from solution in a single batch treatment with the resin. This report describes the application of chelating resins to sampling and determination of certain trace metals using x-ray fluorescence. Two commercial resins will be discussed. Chelex- 100 (Bio Rad) is a resin containing iminodiacetic acid functional groups and has chelation chemistry quite similar to EDTA. NMRR (Srafion) is a resin containing a functional site selective for gold, platinum metals, mercuric ions and methyl mercury.

The application of the null-determinant technique to eliminate the inconsistency of x-ray empirical equations is investigated. Experiments with a series of stainless steel standards show that, in general, calculated compositions obtained by the null-determinant technique agree more closely with chemical analysis than do those obtained by simple averaging. Since it is usually errors in measured intensities which lead to inconsistent estimates of concentrations, the advantages of the null-determinant technique in adjusting the intensities become evident.

The current experimental work extends our calibration concept of separating the effects of absorption and fluorescence to the high - temperature superalloys. The new calibration procedure produces calibration equations which are valid over wide ranges of composition — a feature which is useful in the analysis of high - temperature superalloys . For the specimens considered, the elements iron, nickel, chromium, cobalt and molybdenum can be present at levels greater than 10%; while tantalum, aluminum, titanium, manganese, silicon and vanadium may be present at levels between 1 and 6%. The calibration for this group of alloys has required, in the past, a large number of standards; the number is reduced by judicious application of the given correction equations. Analytical errors can be limited to 1 to 2% in the use of this method.

The following study describes the experimental results obtained by energy dispersive x-ray analysis on a series of representative samples, concentrated by precipitation, to spots 1.5 mm in diameter on Millipore (R) membranes.

Using modified detector collimation, the contributions of air scatter and membrane (Substrate) scatter to the background under a peak, can be minimized to achieve limits of detection approaching a nanogram of the element present in the precipitate.

The equipment used here is composed of an 80 mm2 Si(Li) detector connected to a computer based data handling unit. Excitation was achieved with a Rhodium target x-ray tube having a maximum output of 50 kV at 5.0 ma. Even with the high detection efficiency of the Si(Li) detector and the capability of minimizing the source to sample and sample to detector pauh lengths, it was found with these samples that at least 3.0 ma of tube current is needed in the excitation source.

A mechanism has been developed for this work that allows reproducible sample positioning for standards and unknowns. This mechanism is used to produce element profiles on larger precipitated standards to measure radial uniformity.

A rapid direct x-ray fluorescence (XRF) method has been developed for simultaneously determining brass composition and plating weight for brass -plated-steel tire cord wires. On a manual XRF instrument, a complete analysis can be done in ten minutes including calculations. A complete analysis can be done in only two or three minutes on a multi-channel Quantometer-type instrument. The method can easily be adapted to “on line” use.

Analysis is done directly on the “as received wires. Sample preparation is simple and non-critical with no surface pre-treatment of any kind required. Sample weighing and dissolution of the brass plating are entirely eliminated. Wire samples may be either stranded or unstranded. Rather large variations in sample size can be tolerated because of the ratio-type calculation system used. Analytical precision is high.

Calibration curves were established by wet chemical, atomic absorption, and solution-XRF methods. No serious calibration problems were encountered because the plating is an alloy, or because count data were taken from round, non-flat samples. Low impurity levels of iron in the brass plating are shown to have an insignificant effect on the plating weight calibration curve.

Fluorescent x-ray energy spectra were successfully acquired from uranium and plutonium in 400:1 ratios in samples containing 2 Ci/gram of mixed fission products. The analytical system consists of a silver transmission anode x-ray tube, a low-Z scattering chamber, a magnetic ß-ray trap, a beam monitor probe, a commercial Si(Li) detector, a set of modified electronics to handle the large γ-ray overload rate, and a Computer analyzer using a higher-level language to handle data reduction.

The Computer programs used to obtain peak areas from the closely spaced uranium and plutonium (Lα 1+2) peaks were constructed to make use of knowledge of upper-edge tailing gained in this experiment. Programs are being developed to properly remove background under the uranium and plutonium peaks. Absorption effects in the larger samples have been measured using the ratio of uranium (Lℓ) to uranium (Lα) peak area and are incorporated in the data-analysis schemes. A titanium monitor probe, consisting of a fixed titanium plate near the sample, introduces a constant-area titanium K x-ray line into the spectrum. The program uses the area of this peak to correct for effects of total exciting flux, geometry, and system dead-time losses. Standard samples of various types are used to generate calibration curves from which quantitative results are obtained.

Samples are taken from dissolved high-burnup power-reactor fuel rods. The liquid sample is acidic and has a radiation level at one foot of approximately 2 R/hr ß and 300 mR/hr γ. Sample preparation involves only the evaporation of the liquid sample on a 1/2-mil polycarbonate substrate and subsequent sealing with another layer of polycarbonate film. The samples are then mounted in standard 35-mm slide-holders.

Preliminary testing on a limited number of prepared uranium and plutonium samples indicates a precision of about 1% and an accuracy of about 2%, over a range of 1 to 58 μg total mass. The samples have not yet been verified by independent chemical analysis. The system has been installed at the AEC Savannah River facility for extensive testing.

The results of the fourth year of observations on the effects of self-irradiation on the lattice of 238(80%)PuO2 confirm the predictions derived from previous observations. Namely, the lattice constant continues to decrease with time but at a slower rate indicating an approach to equilibrium. Very small reductions are also noted in the half-width of the diffraction lines.

The effects of various combinations of divergence, receiving and Soller slits on x-ray measurements were investigated for Siemens-Halske and General Electric diffractometers. Influences of the following factors which also affect accuracy and precision of x-ray R.S. results were determined in addition: (a) parafocus versus stationary detector focusing geometry, (b) method of peak location, (c) LPA intensity correction, (d) diffractometer electronic stability and (e) elastic constants.

The optimum choiees of beam optics and factors (a-e) were defined with regard to aecuraey, precision and minimurn time for stress deterniination, on sharp and broad line speeimens of soft (annealed) and hardened steel and of annealed Cr-powder.

The standard three-point parabola technique has been used extensively in residual stress analysis for peak location with precisions (standard deviation) of peak shift measurements in the range of 0.01° to 0.03° (20). While this level of precision produces useful data for peaks at high diffraction angles, it becomes untenable when a material with high elastic modulus is encountered whose only suitable peak falls below approximately 140° 20. Uranium-0.75 weight percent titanium alloy, which is the material of interest in this investigation, falls into this category.

A technique has been developed by which diffraction peak centroid shifts are determined with a precision of 0.002° of diffraction angle (Δ20). The method consists of step-scanning over the peak of interest to acquire data in digital form. A small Computer is then used to locate the diffraction line by calculating its centroid from digital data.

The use of a special probe consisting of four distance transdueers (LVDT) to achieve precise alignment of the specimen surface on a Siemens back-reflection stress diffractometer will also be described.

A calibration experiment was performed on uranium-0.75 titanium tensile specimens in which the X-ray diffraction response was measured as a function of various levels of applied stress in order to determine the experimental elastic moduli. It was found that repetitive analyses on a single area of a cylinder had a standard deviation on the measured diffraction centroid shift of 0.002 degree which is approximately an order of magnitude smaller than that achieved by the standard parabola fitting technique. Thus, the residual stress can be defined to a precision of ∼ 2,000 psi, as opposed to 30,000 psi, at best, were the measurements obtained by the conventional procedure.

Structural modifications occurring in corroded potassium silicate glasses were studied using x-ray amorphous scattering techniques. Pair function distribution curves are compared and the differential method is used to follow the structural changes. The structure appears to approach that of vitreous silica after long leaching times.

X-ray diffraction methods for studying solid-state diffusion are reviewed. Because of the lack of penetration, such methods are suited for diffusion zones spanning only a few microns. Most involve analyses of one, or more, (hkl) “band” of intensities spread in 2θ as a result of the corresponding lattice parameter spread associated with compositional inhomogeneities. Further, many are non-destructive, making it possible to follow the progression of diffusion with time in the same specimen.

This paper describes a method for automatic quantitative pole figure plotting up to 70°, using only one sample and Schulz reflection technique. Random intensities are calculated for the usual planes of iron, using the ratios of calculated povder integrated intensities of these planes to the intensity of a high multiplicity factor plane, the random intensity of which had been obtained by integration up to 70° over the pole figure, using Bragg and Packer's method; the latter integration shows a decreasing error when the multiplicity factor increases. With this method it is possible to normalize the pole figures without using a physical standard and to reduce greatly the time to obtain a quantitative pole figure.

Intense x-rays are emitted by plasmas formed when sub-nanosecond laser pulses are focused onto materials, Plasmas produced by pulses containing up to 100 J can re-emit over ten percent of the energy as x-rays above about 1.0 keV. These plasmas may be useful flash x-ray sources.

A method for calculation of the integral reflection coefficient of crystals of interrnediate perfection is introduced. This method can greatly reduce experimental effort for the selection and calibration of crystals, It also serves as a conceptual framework for studies of mosaic block structure and of crystal modification. Good agreement between calculated and experimental values of the integral reflection coefficient is shown for, (a) LiF crystals of two degrees of perfection, (b) elastically bent quartz, and (c) 001, 005, 006, and 007 diffraction from KAP. Zachariasen's division of crystals into two types is extended. It is concluded that the integral reflection coefficients for 200 LiF cannot be raised to the ideally imperfect limiting values.

Absolute K-shell ionization cross sections were measured for Ti, Co, Ge, Rb, and Sn for incident oxygen ions from 16-44 MeV. The x-rays were measured with a high resolution Si(Li) detector (166 eV at 5.9 keV). All of the data represents cross section measurements for thin targets. The measured cross sections for these elements are compared to the theoretical predictions of the Binary Encounter Approximation (BEA). Kα/Kβ ratios and energy shifts were also extracted from the data. The experimental data are compared to measured cross sections for other elements to give an overview of the systematics for oxygen ion induced x-ray production cross sections in this energy range. Some comment will also be given in regard to the use of oxygen ions to measure the parameters associated with ion implanted semiconductors.

X-ray fluorescence induced by charged particles has been employed in trace element analysis of both animal and human blood, tissue and bone samples. Preparation techniques included microtome slicing and wet digestion in nitric acid, internal chemical standards being used in the latter case.

Most of the specimens arose from a study of interactions between the toxic elements lead and zinc in growing foals; this was motivated by reports of sickness and death in foals raised near lead-zinc smelters. The cause of toxicity in animals from environmental pollution is often attributed to Single factors, whereas in reality interactions among many factors, including a variety of toxic and nutrient trace elements, should be considered.

A variety of spectra are presented and elemental concentrations derived. Agreement between the X-ray data and atomic absorption spectrophotometry is encouraging. The results demonstrate the potential of particle-excited X-ray fluorescenee as a broad-range analytical technique for the study of trace element interactions.

The value of the divergent beam method as a powerful tool for investigating crystal perfection over rather wide limits has been known since the analytical work of K, Lonsdale. This method has been treated experimentally and theoretically by numerous authors and was popularized by Sigmund Weissmann. Although not nearly as useful as the reflection and transmission topographic techniques of Berg-Barrett and Lang for crystal characterization, it is potentially very useful for three dimensional quantitative strain analysis. The “technique phase” as Harvey Yakowitz remarked, is essentially over; however, refinements and optical improvements are certain to evolve if quantitative measurements are to become useful. Very short exposures and the non-destructive nature of the reflection divergent beara - Kossei method is a useful survey tool that has gained limited acceptance.

Presently, the reflection mode is useful in the qualitative survey of numerous semiconductor materials and processes, some of which have been considered from time to time by various authors. A range of crystal perfection, as well as some surface modifications, are surveyed. Kossel reflection pattern artifacts and irregularities are discussed.