In steam-generating systems of all types, producing and testing pure water is of utmost importance for the life of the system. A 1000-MWe (megawatt electric) power plant generates 6 million pounds of steam per hour if fossil fueled and 11 million pounds per hour if nuclear. Should an impurity have a concentration of only 10 ppb, in a year's time, 550 to 1000 pounds of solids can accumulate in the power-generating cycle. These solids may initiate numerous problems, including preboiler-cycle corrosion, boiler-tube failure, and turbine damage.

At the Babcock & Wilcox (B&W) Alliance (Ohio) Research Center and throughout the industry, three principal methods are used to monitor water chemistry. On-line continuous analyzers measure parameters such as pH, conductivity, chloride, and sodium. Grab sampling (followed by laboratory analysis) is used for parameters that are measured less frequently; for example, sulfate. For some parameters, such as corrosion products, special sampling techniques are required. This is because no on-line instruments are available, and grab-samples are not chemically stable. Corrosion products are collected near the sample tap using a 0.45-micron filter-disc followed by a stack of resin-impregnated membranes.

A new fundamental-parameter program has been developed which corrects for light element absorption based on the mean atomic number of the sample. The mean atomic number, in turn, is determined from the Compton/Rayleigh scatter intensity ratio. The program is quite flexible with regard to the number and the type of standards which may be used. The accuracy and precision of the method has been evaluated with several geological and biological standards. The results are in good agreement with those obtained by some other methods (CEMAS, XRF-11).

One method of dealing with unanalyzed light elements in x-ray fluorescence analyses is to add a relatively large quantity of a diluent to the sample. The interelement effects on the measured radiation are then essentially all due to the diluent, and the effects of the unanalyzed elements are ignored. The general practice has been to use a diluent compound composed of light elements. The rationale for this is that light elements absorb the analyte radiation less strongly than heavy ones, leaving higher net intensity available after dilution. However, since absorption effects are greater with heavy elements, a smaller dilution factor can be used to reach the analyte concentration at which the effects of unanalyzed elements on the analyte radiation are negligible compared to diluent effects. In many cases, the smaller dilution factor almost exactly compensates for the increased absorption. There is essentially no intensity penalty for using a heavy element diluent. In fact, there are some advantages to doing so. The lower dilution factor with heavy elements can avoid problems with multiple dilutions or accurate weighing and homogeneous blending of very small quantities necessary to achieve high dilution factors. Significantly less continuum radiation is scattered from heavy element matrices, and signal-to-background ratio can be higher in some parts of the spectrum. With high energy analyte lines, heavy element dilution results in infinitely thick samples, simplifying sample preparation. Theoretical intensity calculations using the XRFll matrix correction program are used to predict the appropriate dilution factor for a range of analyte x-ray emission wavelengths with a variety of diluents. The results are verified experimentally.

Layered Synthetic Microstructures (LSMs) were first reported as commercially available diffracting structures for routine x-ray spectrochemical analysis in 1984 (Nicolosi et al.). Innitially, these devices (PX-1 and PX-2) were intended for use as alternatives for Thallium Acid Phthlate (TAP) and Lead Octadeconate (LOD) for improved stability and diffracted intensity. Improvements in the analytical performance of X-Ray Spectrometers for the elements carbon to magnesium has met, and in some cases exceeded, innitial expectations. Having captured the interest of many new possible applications, developments for LSMs with larger 2d-spacing and greater reflectivity have continued.

Previous instruments employing LSMs have had difficulty in measuring X-Ray energies below the carbon K-series characteristic. Below the carbon K-edge energy, several significant phenomenon combine to reduce sensitivity. Among these are: 1) Reduced reflectivity of conventional LSMs, 2) Increased absorption at the polypropylene window of the detector, and 3) Greater electronic noise.

Layered Synthetic Microstructures (LSM) for XRF instruments became commercially available in the early 1980's. It was quickly recognized that these devices improved the efficiency of diffraction for the elements boron through magnesium considerably over the LOD soap multilayers and single crystals available at that time. LSM devices of 40, 50, 80, and 120 ang-strom (2D) spacing were offered at first. Gradually, many different 2D spacings and compositions have become available, each with its own special area of application, usually limited to one or two elements.

For many years Layered Synthetic Microstructures (LSM's) have been used successfully as dispersive devices. The commercial variety introduced in 1984 offered significantly greater stability and diffracted intensity characteristics over Thallium Acid Phthalate (TAP) crystals. These devices offer superior sensitivity for fluorine, sodium and magnesium in all but a few applications where resolution is very critical. Subsequently, a variety of these devices have been optimised for analysis of the lower atomic number elements.

Recently, several industrial applications have required the development of LSM's with 2d spacings optimised to enable quantitative analysis of oxygen, nitrogen, carbon and even boron. In these applications, the large incremental difference in characteristic wavelength per atomic number step limits the dynamic range of elements over which the LSM can be optimised. These considerations are discussed as they relate to routine analyses of very low atomic number elements by sequential Wavelength Dispersive Spectrometry (WDS).

Newly developed fabrication techniques are used to produce curved LSM's for use with the focusing optics of simultaneous WDS. Application of these curved devices for the analysis of Mg in steel, Na in slags and F in glass will be discussed.

Finally, an application which requires the analysis of thin films on semiconductor devices will be described. In this method, characteristic K-spectra and L-spectra of low and high atomic number elements, respectively, are selected for optimum sensitivity to the surface composition.

The development of X-ray spectrographic analysis of light elements, which are O, C and B, has bee n performed for many applications using an end-window type X-ray tube with Rh-target and thin Be-window, wavelength dispersing devices, which are synthetic multilayers or total reflection mirror (with a specific filter) and a gas flow proportional counter with a thin film window. In Fig. 1 factors related to the intensity measurements in X-ray fluorescence analysis are shown. The excitation efficiency in the soft and ultrasoft X-ray region is very low because of the lower intensity of primary X-rays and low fluorescence yield of light elements. Instead of the wavelength dispersive method of Bragg reflection, having high resolution and low reflectivity, monochromatization combining total reflection by a selected mirror and an appropriate filter offered an alternate approach in order to increase measured intensity with reasonable optical resolution. Synthetic multilayers which have higher resolution and lower intensity compared with the performance of the mirror method have become popular for the detection of soft and ultrasoft X-ray region.

Over the past 30 years, X-ray fluorescence spectrometry has become one of the more valuable methods for the qualitative and quantitative analysis of materials. Today, there are many methods of instrumental analysis available, and among the factors that will be taken into account in the method of selection are:

1. - Accuracy

2. - Range of application

3. - Speed

4. - Sensitivity

5. - Reliability

No one technique can provide all of the features that a given analyst requires. XRF does offer good overall performance over the widest range of elements. Speed, accuracy, and versatility are among the features that have made XRF the method of choice for over 15,000 laboratories worldwide.

The efficiency of the recessed source geometry for the analysis of the content of wear metals in engine lubricant oils by Energy Dispersive X-Ray Fluorescence (EDXRF) methods is calculated by a Monte Carlo simulation considering X-ray detectors of the Gas Proportional Scintillation Counter type and excitation by Cd-109, Cm-244 and Fe-55 radioactive sources. Calculated spectra for the case of a typical aircraft oil with Cu, Fe, Cr, Ti, Sn and Ag impurities in the p.p.m. range were obtained and results are presented that allow the prediction of the counting rate per mCi and p.p.m. of the impurity content. The performance of the system is discussed concerning its potential applications to in-line monitoring of the wear metals.

A novel windowless sample preparation method is described which allows for X-ray fluorescence analysis of light elements (Z = 13 to 20) in coal or oil-tar pitches, asphalts and other bituminous compounds. The article also discusses blending and spiking of pitches with suitable organometallic compounds in order to prepare a range of trace-element calibration standards. The general instrumental conditions, recommended peak counting times and detection limits are also discussed.

Fly ash is a finely divided particulate material that is removed from the flue gases of pulverized coal burning power plants. Recent emphasis on the utilization of fossil fuel (i.e., coal) has created a serious waste disposal problem in that, of the 60 million tons of fly ash produced annually in the United States, only about 20% is being utilized [1], The remaining fly ash (about 48 million tons per year) must be disposed of in some environmentally safe manner.

High-resolution measurements of the molybdenum L heavy-ion-induced X-ray satellite emission (HIXSE) spectra of a series of Mo alloys and compounds have been obtained with a new, high-efficiency, high-resolution Bragg crystal spectrometer in the von Hamos geometry. The improved resolution (5 eV at 2.3 keV) is sufficient to reveal the LnMm configuration lines in the Lα and Lβ hypersatellite bands. Both sets of lines exhibit the same trend in the variation of the relative yield distribution with the chemical environment as was observed for KLn lines of lower Z targets. Difference spectra, using elemental molybdenum as a subtrahend, enhance the systematic variation. These results confirm the analysis and conclusions of lower resolution studies. They also indicate a potential for even greater sensitivity to the chemical environment.

Traditionally, aluminum has been the most important raw metal material for the aerospace industry. Since the major part of aluminum production is consumed for aircraft, military, space, and other programs, sponsored or regulated by the U. S. Government, quality of various Al alloys is under governmental control. The representatives of the governmental agency are placed at each aluminum plant in order to assure a high quality of production on-site. A fast and reliable analytical method is essential to control the production process and to keep chemical composition of the alloys inside specified ranges.

A preconcentration procedure is reported for rare earth elements suitable for concentration levels generally found in rocks. This method is of geochemical interest because the abundance of rare earth elements gives important clues to the origin of rock formations.

The procedure is based on an anion-exchange separation step followed by a precipitation of rare-earth elements chelates onto a membrane filter. For the investigation of recovery a combined radionuclide tracer and X-ray sensitivity method was used. Detection limits in the sub-μg/g-range are reported.

A previous paper by this author (1984) described the evolution of XRF analysis in the British Geological Survey. More recent developments are described here, with particular reference to wide-area networking. Since the Survey was founded in 1835 it has been to a large extent concerned with the amassing of data. In most cases the information is permanently retained, in contrast to some other applications where the data are stored only temporarily. The result of this permanent storage is that a very large database has been created and continues to grow at an increasing rate, partly as the result of extension of the terms of reference to include the continental shelf. Another reason is that geochemical, geophysical and geotechnical information is now included, and automated methods of measurement and data capture now exist.

Energy-dispersive X-ray techniques can be employed in two different ways for the accurate determination of element concentrations in specimens: (1) spectrometry of fluoresced characteristic X-rays as widely applied in the various modes of the traditional XRF analysis technique, and (2) spectrometry of the energy-differential transmittance of an X-ray continuum at the element-specific absorption-edge energies.

Quantitative utilization of energy dispersive X-ray fluorescence spectrometry has been made for many years, and the development of instruments for this purpose have arisen from needs which can be met by the sensitivity of the technique. In a need to get information which could be obtained only by scanning single strands of hair to obtain the profile of the mercury levels along the hair, a unique X-ray fluorescence energy dispersive spectrometer was built. The prototype of this instrument was described at the Denver Conference of 1977, and published in the subsequent publication (1). The present instrument is described in a later publication (2).

Many manufacturing processes in the chemical industry require frequent monitoring of the composition of the product. Ideally, an analysis of this type should be simple to carry out, rapid, reliable, and not require a seperate analytical laboratory. Energy Dispersive X-Ray Fluorescence (EDXRF), due to its relative simplicity of instrumentation, speed of analysis, and non-destructive nature, is well suited to on-line applications, when used in conjunction with a flow cell.

A question that is frequently raised in this type of analysis is: How do the precision and accuracy of these measurements compare with measurements made upon the same solution in sample cups (static). The purpose of this study was to compare the precision and accuracy of on-line with static analyses of the same types of solutions.

A method for the direct determination of trace elements in light element matrices is described. The intensity of Compton scatter of the incident X-rays from the specimen was used to evaluate the apparent absorption factor for the direct correction for matrix effects. This permits the use of single element standards for samples of varied and complicated matrix composition. Relative errors of approximately 5% were obtained for results in the concentration range of 10-200 ppm.

The developments in the analysis of low atomic number elements or the measurements of soft and ultrasoft X-rays have been carried out. Using this technology, a measuring method for thinner layered materials has been developed recently. This type of analysis requires a high excitation and wavelength dispersive devices with high reflectivity.

On the other side, the film technologies related to electronic and optical devices in the field of semiconductor industry have made a remarkable progress. A variety of layered material thickness in the multilayered structure of LSI is in the range of 100Å and 2μm. It is too thick to be measured by electron methods and too thin to be measured by a regular X-ray method using the X-ray wavelength of 0.2-3Å. Accordingly, it is generally said that the soft and ultrasoft X-ray method should be applied to measure the thickness and composition of layered materials on silicon wafers.