The use of non-radioactive tracers for measuring certain clinically significant parameters by x-ray fluorescence analysis (XRFA) is well established at our institution (1,2). Among the technique's advantages are high accuracy, simplicity, cost effectiveness and reduced or eliminated radiation exposure to patients. One of the more versatile tracers is iodine, which has a low body toxicity and desirable chemical properties, e.g. it is easily bound to organic molecules. Other elements of interest are cesium (3), a potassium analogue of low toxicity; and xenon, the highest atomic number, stable, noble gas.

This paper describes the development of a radioisotope x-ray fluorescence gauge built to continuously measure and control the coating weight of wet silver emulsion. The method and apparatus for correcting the basic measurement for variations in water content and base film thickness are described. Design considerations, including source selection and arrangement to minimize radiation scattering and emulsion exposure while providing acceptable measurement precision or accuracy, are discussed.

Two methods of obtaining polarized x-rays for fluorescence experiments are discussed. Compton scattering from a low-Z scatterer is the usual method used in such experiments. The polarization of x-rays undergoing anomalous Borrmann transmission in a dislocation-free crystal is also described and preliminary results are presented. Approximate expressions, useful for comparing scatter-polarizing systems, are derived for the dependence of scatter rejection and fluorescent efficiency on two scattering-system parameters: the thickness of the scattering polarizer and the geometric limit to solid angles and angular divergences in the system.

Todays demand for rapid trace element analysis in pollution control and resource materials has led to the development of a completely automatic, very sensitive and stable Si(Li) X-ray analyzer. The key element, a transmission target tube, which has inherently a very clean monochromatic X-ray output, has been studied in view of efficiency, sensitivity and stability. The transmission target tube requires different operating criteria than conventional X-ray tubes. An analysis was made to “explain” these differences using fundamental X-ray physics. Studies included various experiments directly applied to practical problems in analysis of pollutants and mineral resource materials.

Measurements of peak-to-background ratios for K x-ray production by 1.7 MeV/amu protons, alpha particles, carbon ions, and neon ions have been carried out for the purpose of determining the optimum projectile atomic number for particle induced x-ray emission analysis. The feasibility of applying the method to the trace element analysis of thick samples of tissue and blood serum has been explored.

The production of characteristic x rays by 100-keV protons has been studied to determine the feasibility of using a Cockcroft- Walton accelerator as an excitation source for x ray fluorescence analysis. Because of the low cross sections at this proton energy, it was necessary to concentrate on characteristic x rays below 6 keV and to use beam currents of 10 to 100 μA. The x rays were measured by a high resolution Si(Li) detector.

The preparation of various types of samples and the problems that arise when working at low proton energies and high beam currents are discussed.

It was found that the technique is useful for low atomic number elements down to less than 1 ppm in a water sample when there are no other significant impurities. When other elements are present, however, proton absorption becomes an important factor and quantitative determinations become very difficult. Proton absorption and its effect upon the usefulness of the technique are discussed.

The precision for the technique was determined and found to be better than 20% for most measurements. This includes errors in sample preparation which was considered to be a significant portion of the uncertainty.

An X-ray fluorescence analysis unit has been automated with a multi-position sample changer, a stepping motor to position the spectrometer, and computer addressable switches to control the selection of crystal, detector, collimator, and beam filter. The unit can be controlled off-line through a Teletype or on-line with a computer. This computer utilizes a multi-user program for the simultaneous operation of the fluorescence analysis unit and two diffractometers. Programming the system for any desired analytical or research procedure is accomplished using an expanded version of BASIC.

In a refractory manufacturing industry with widely diversified products from silica (SiO2 ≥ 90%) to high alumina materials (Al2O3 ≥ 90%) and basic refractory from periclase (MgO ≥ 95%) to chrome-periclase (Cr2O3 ≤ 40%) mixtures, introduction of a single method of analysis is impractical both economically and precision-wise. Three different methods are proposed with a view to realistic needs: (a) Direct ratio method, (b) calibration curve ratio method, and (c) matrix normalization by fusing with Li2B4O7 and La2O3 into a glass bead followed by grinding and pelletizing. Statistical regression equations were used for the calibration curves from available NBS standards 15-20 samples previously analysed by wet chemical methods by ratioing with relevant reference samples. In most cases, the relative standard deviation was 2 - 8%.

An analytical quality control procedure using x-ray Secondary Emission Spectrometry has been developed to monitor iron, yttrium, and gadolinium metal ion ratios in feed stock solutions. Binary and ternary laboratory standard solutions were analyzed to determine absorption and enhancement effects using approaches suggested by Rasberry-Heinrich, Beattie-Brissey, and Claisse-Quintin. A computation scheme combining these methods is presented which enables the user to improve his ability to predict composition. This approach has yielded relative errors of less than 1.0% for each of the three metals studied.

Alpha particles and protons from charged particle accelerators and photons from both x-ray tubes and radioactive sources have been shown to be useful for the excitation of characteristic x-rays for multi-element energy dispersive trace analysis of environmental samples to the few ppm range. We have investigated the use of 4.5 MeV alpha particles from a thin window Po-210 source of 5 mCi effective strength to directly excite x-rays from trace elements in 1 cc water samples evaporated on 1.75 mg/cm2 thick mylar backings in a helium atmosphere in a lucite enclosure. Minimum detectable amounts (MDA's) were established for 19 elements (22 ≤ Z ≤ 82) using K-, L- and M- radiation and 50 minute counting times. The smallest MDA determined was 0.11 μg for vanadium. Other representative MDA's, in μg, are Fe-0.54, Mo-0.31 and Pb-0.43. MDA's lower by an average factor of about 10 over an eight month source life would result from the use of 1 Ci of Po-210 for 50 minutes per sample. Thinner sample backings and improved source encapsulation will reduce background radiation and further improve sensitivity. Comparison of our MDA's (5 mCi-Po-210) with those of Blasius et al., who used radiophoton sources and 40,000 sec. counting times to determine trace metal pollutants in water samples, shows radioalpha excitation to have comparable sensitivity in the worst case, arsenic and better, by more than two orders of magnitude, sensitivity in the best case, vanadium.

Radioalpha induced x-ray trace element analysis offers the same advantages of portability, ease of operation, low maintenance and cost, and "in house" availability as radiophoton induced analysis. Because of the availability of more intense sources (up to 10 Ci), the fact that the detected radiation (x-rays) differs from the excitation radiation (a-particles) and that K, L, and M x-ray emission cross-sections depend essentially only on the emitted x-ray energy, lower MDA's are obtainable for many elements in thin samples for comparable counting times, as well.

Blood samples from normal and Plasmodium berghei infected mice are being analyzed for trace elements by charged particle induced x-rays. Approximately 0.25 ml of the sample (whole blood, washed red cells, or plasma) is dry ashed. The ashes are mounted on a 0.003 inch Kapton foil. The analysis is performed by bombardment of the samples by a beam of 2.0 MeV protons and detection of the characteristic x-rays by a 175-eV-resolution lithium-drifted silicon detector. The data are analyzed by an on-line PDP-9 computer-based data acquisition system. Results indicate an increase in the K, Ca, Cu, and Zn per unit volume of the red cells of the malaria infected mice relative to the amounts measured for uninfected blood, and a decrease in the K, Ca, and Fe and an increase in the Cu per unit volume in the plasma of the infected mice.

A Fortran program for analysis of spectra obtained in proton induced x-ray emission spectroscopy is under development. The first version is working and some results are shown. The program uses a model which describes the physics involved in the analysis. The continuum is described by an exponentially decreasing function representing secondary electron induced bremsstrahlung and a constant approximating different Compton scattering processes. Absorption effects are also included. All parameters in the model are fitted simultaneously. About 100 test runs have been made with good results. The continuum was accurately fitted for each spectrum. Other advantages are reliability, speed, and the fact that the program is easily adaptable to a small computer with 16K memory.

A system for the calculation of first-order influence coefficients for any element in any matrix is described in detail. Influence coefficients for either K or L spectra can be evaluated by the method described.

At the present stage of development of mathematical correction procedures, the so-called empirical methods are well established and reasonably reliable, require relatively small computers for their application, but require large numbers of calibration standards for the determination of the (alpha) correction constants. On the other hand, the so-called fundamental methods are less well proven and require more sophisticated computing facilities. One possible means of achieving the advantages of both techniques is to use a large computer to calculate the alpha coefficients from fundamental constants, then to use a small online computer for their application in an "empirical type" algorithm. de Jongh has proposed such a program called "ALPHAS" for the calculation of interelement correction constants and the present paper elaborates on this approach using as examples data collected in our own laboratory during the past year or so.

An analytical system has been defined and developed which provides the x-ray analysis of iron base alloys (metal samples) corrected for an absorption, secondary fluorescence and interference effects. The analytical system provides for the transfer of analytical data and curves from basic or reference instrument to each new installation. One of the specifications for this system is the stringent accuracy requirements of the production analytical laboratory. Two major advantages of this system are:

• (1) The reduction of operator error because no decision is required for alloy identification,

• (2) A minimum of time required for new instrument calibration by the use of the analytical transfer procedure.

Using a multiple regression procedure, X-ray fluorescence analysis has been applied to a 15-element system associated with ceramic supported automobile emissions catalysts. An equation taking inter-element effects into account was used to relate the concentrations and X-ray intensities of the active elements and contaminants deposited on the ceramic substrate. The major elements present in the substrate were intentionally excluded from the regression. The regression coefficients were determined from a set of 240 calibration standards using a multiple linear least-square regression. Elemental concentrations in the standards ranged from 0.05 to ten weight percent. The ceramic matrix was the same in both the standards and the unknowns to be analyzed and constituted at least 85 percent of the total weight of each sample. The regression coefficients were then used to solve for the contents of the elements in unknown samples. Limits of accuracy were determined by using the set of available standards. The performance of the regression was improved in certain cases by imposing constraints on the regression.

Chemical shifts of SKα, CrKβ1,3 and SnLβ2 were precisely measured with a two-crystal X-ray spectrometer (Toshiba AFV 701). Sulfur in the coke for iron smelting takes two oxidation states, i.e., S2− and S6+. Quantitative analyses of S2− and S6+ were performed by precisely measuring the profile of SKα and fitting the standard S2− of S6+ curve to the observed SKα profile with the non-linear least square method. Chemical shifts of CrKβ1,3 and SnLβ2 were utilized for the determinations of Cr6+, Cr3+, Sn4+ and Sn2+ in electro-plating solutions. Intensity ratio of IA to Ig was measured, where B means the point which showed an intensity independent of oxidation states. A means the point which showed the most dependent. Linear relations were found between intensity ratio IA/IB and Cr3+/Cr or Sn2+/Sn.

When the excitation energy lies near the ionization potential of many core electrons, core-hole relaxation effects strongly affect x-ray emissions. Under near-threshold electron excitation of the cerium 4d subshell both the incident electron and a 4d-core electron are promoted into 4f-orbital excitation states which exist only during the life of and are bound to the 4d vacancy. The exchange interaction between the 4d hole and the excited electrons in the unfilled 4f subshe11 spreads the energies of the discrete excitation states over a 20eV range sending some levels well below the 4d threshold and others into the continuum. Exchange interactions and potential barrier effects have a dominant influence upon the cerium 4d and 3d soft x-ray appearance potential spectra reported in this paper. We observe sub-threshold x rays extending to 8.6eV below the binding energy of the 4d level as well as a maximum in the yield at 12eV above the 4d onset. As the electron excitation energy is swept through the thresholds of the two 3d levels we observe resonant x-ray emissions, while at 28eV above these levels we see plasmon loss structures.

An x-ray source is described that can have high brilliance, monochromaticity and spatial coherence. This source filters, by successive diffraction, spontaneous x-ray radiation produced by electron impact in a Bonse-Hart monolithic crystal monochromater. A germanium single crystal was grooved to form a monolithic monochromater and subsequently was fitted with a filament. Field emitted electrons from the filament impact on the crystal side walls. Thus, the passive Bonse-Hart monochromater is converted into an active x-ray source. The detailed diffraction geometry within the cavity, in the limit of kinematic diffraction theory, is described and compared to the experimentally observed radiation pattern. Since this radiation source achieves its monochromaticity and spatial coherence by filtering it is highly inefficient. However, the diffraction process is highly efficient when the radiation is in the proper geometric direction to satisfy Bragg's law. Thus, the stationary radiation source can have a high brilliance although the input of electrical energy could be well in excess of a rotating anode x-ray tube.

The beryllium diamond-anvil pressure cell described by Weir, Piermarini and Block has been mounted on a Bond diffractometer equipped with an orienter of the fixed-x type. Molybdenum radiation is used to penetrate the diamonds and beryllium of which the cell is constructed, and special techniques are required to retain adequate precision in measuring cell parameters using diffraction angles in the low 2θ range. The method described provides high sensitivity in determining peak positions and eliminates the effect of centering errors on measured values of 2θ. Under favorable conditions, diffraction angles are measured with an accuracy of ±0.001° in 2θ. The method has been tested by measuring the lattice parameter of vacuum float zone refined silicon. Measurements of the compressibilities of silicon and of α-Pb(N3)2 (orthorhombic) have been carried out using the method of Barnett, Block, and Piermarini to determine pressure by measuring the shift in the R-line fluorescence spectrum of ruby.