X-ray fluorescence is well established as an analytical method for the determination of multi-element systems. It is a technique which provides high accuracy and precision. However, it has serious limitations in the lower limit of detection for elemental analysis in environmental samples. In order to overcome this limitation some method of preconcentration or enrichment of the trace elements in water samples must be provided. In recent years many techniques have been developed for this purpose. Some of these techniques involve simple chemical or physical manipulations of the sample. In all cases, the technique should be rapid, simple, and not contaminate the sample. Because it is more convenient to use solid samples for X-ray fluorescence analysis, the ideal method of preconcentration results in a solid sample.

Methods are presented for the determination of uranium, or uranium and neptunium, in plutonium metal and plutonium alloys. Anion exchange or a combination of anion exchange - solvent extraction is used to concentrate the elements for x-ray fluorescence analysis, depending upon the impurities present and the elements to be determined. The precision for determining between 3 and 250 μg of uranium or neptunium ranges “between 30 and 2%.

Use of Chelex-100 impregnated filter membranes for preconcentration of uranium in ground-water samples for XRF analysis leads to lower processing time, sample size, and detection limit than achieved in a batch extraction process. Data is presented, for potential effects of iron and organic matter in natural waters upon the recovery of uranium when using Chelex-100.

When a sample is fused with a flux to produce a glass button for X-ray fluorescence analysis, some of the constituents are lost by decomposition (H2O, CO2) or by evaporation (Br, Cl, S) . Therefore the glass button is smaller than if no “Loss on Ignition” had occurred, and the X-ray intensities appear stronger. It is shown that “L.O.I.” can be handled as a regular element and that corrections can be made by means of influence coefficients. Since “L.O.I.” cannot be measured by X-ray fluorescence, one should normally know the “L.O.I.” to make the appropriate corrections, but under certain conditions, X-ray intensities yield both composition and “L.O.I.” as shown by Tertian and as confirmed by our experiments.

The measurement of “chemical shift,” that is, the change in energy of an element's x-ray emission lines with the state of its chemical combination, has been carried out for some years. Of the three major aspects of the technique, two have received major attention. Nagel (1) has an excellent treatise on the interpretation of valence band x-ray spectra, while such workers as Fischer (2) and Koffman and Moll (3) have attempted to correlate the data with structure. The third area, convenient data collection, has not been so well investigated. Much, but not all, of the effort has been toward direct electron excitation with its attendant problems of sample damage due to high vacuum and electron bombardment effects.

We describe a new and reliable experimental technique for the measurement of mass absorption coefficients in the 0.1 to 1 keV energy range. In this technique, the absorbing film is supported directly on a substrate which under proton bombardment will generate the x-rays whose absorption will be measured. Results are given for thirteen different metals at the C Kα (277 eV) line.

Much attention has been given to the treatment of energy-dispersive X-ray spectra. Descriptions of mathematical approaches (and their implementation in computer programs) have dealt with the major areas of separating peaks from background and from each other. There has been an unspoken but widely accepted underlying assumption that the more complex the approach, and the more general it seems to be, the bette r. This assumption appeals to the mathematicians and programmers but should be sharply challenged by users , whose focus is more properly the analytical results that can be obtained in spec ific circumstances.

Two methods are described for computing multielement x-ray absorption corrections for aerosol samples collected in IPC-1478 and Whatman 4l filters. The first relies on scatter peak intensities and scattering cross sections to estimate the mass of light elements (Z < l4) in the sample. This mass is used -with the measured heavy element (Z ≥l4) masses to iteratively compute sample absorption corrections. The second method utilizes a linear function of ln(μ) vs. ln(E) determined from the scatter peak ratios and estimates sample mass from the scatter peak intensities. Both methods assume a homogeneous depth distribution of aerosol in a fraction of the front of the filters, and the assumption is evaluated with respect to an exponential aerosol depth distribution. Penetration depths for various real, synthetic and liquid aerosols were measured.

Computer controlled X-ray spectrometers have been available since the mid 1950's and there are many hundreds of these systems in use today. The degree of sophistication of these machines has increased gradually over the past several years, but in all cases the flexibility of the analytical system is directly relatable to the available computer hardware. In the design of any computer controlled X-ray spectrometer, it is highly desirable to keep the cost of the computer, plus its associated interfaces and software, to within a reasonable fraction of the total cost of the whole system. A figure of perhaps 30% or so is a good figure to aim for in this context. A compromise must, therefore, always be sought between cost and computational power--which generally means flexibility.

A comprehensive Fortran IV program designed to perform the matrix correction in x-ray fluorescence analysis is described. Specimens and standards can be in bulk or film form. All necessary fundamental parameters are provided by internal routines thereby requiring a minimum of input data.

A new proportional counter x-ray detector with application for powder diffractometry has been developed. The new detector collects virtually the entire diffraction spectrum simultaneously with good efficiency and angular resolution. Thus, the powder spectrum of a small sample can be obtained much faster than with film or a conventional powder diffractometer. The detector read-out is digital and is interfaced to a dedicated minicomputer. A description of the detector system is discussed and preliminary results are presented.

Energy - dispersive x-ray fluorescence systems are characterized by limited count rate capability, which is an important practical limitation on their application. This arises from two factors: 1) The amplifier time constant is many times greater than needed in wavelength-dispersive systems, in order to permit accurate measurement of pulse height, and reducing the time constant degrades system resolution; 2) x-ray pulses from ail the major and minor elements, as welt as scattered background and/or peaks from the incident radiation, must all be processed, reducing further the effective count rate for a single element of interest. Hence there has been a strong desire to find ways to increase the count rate capability of energy-dispersive systems.

The wall effect present in the gas-filled proportional counters and the low-energy background caused by it are well-known disadvantages. In this paper a method is presented to define the radial position of a single X-ray absorption in ordinary proportional counters. The method utilizes the variation in detector pulse collection times due to the electron diffusion. This variation can be determined with a specially constructed risetime analyzer which effectively forms a dead zone of cylindrical shape inside the detector. The properties and performance of the method are demonstrated. In the experiment performed the signal to background ratio was improved by a factor of twenty.

The relative intensities of the Kα characteristic radiation obtained from copper-target X-ray diffraction tubes have been calculated for a range of tube accelerating voltages and take-off angles. The calculations employ an over-voltage function, and absorption and atomic number corrections similar to those used in electron microprobe analysis. They apply only to constant potential X-ray generators. Measurements of actual intensities obtained on a Picker diffractometer using a sodium chloride monochromator gave relative intensities in close agreement with those calculated. The calculations and measurements show that there is an optimum tube voltage, with respect to intensity, for each take-off angle. This voltage increases with increasing take-off angle. The application of these results to the consideration of the relative intensities obtainable from broad, standard and fine focus copper-target X-ray diffraction tubes is discussed.

Polarized x-radiation produced by scatter at 90° from boron carbide is shown to be a superior excitation source for the measurement of trace elements by energy dispersive x-ray fluorescence. In a close-coupled system which uses a high power x-ray tube as the primary source, the losses due to the geometric requirements and scattering efficiency of the polarizer can be more than compensated. With the system described here which uses a Mo anode x-ray tube, detection limits for the element between K and Sr in NBS orchard leaves are approximately 2 to 4.5 times lower using polarized excitation in comparison to direct excitation, and about <1 to 3 times lower in comparison to secondary excitation.