The analysis of magnesium brines can be performed by atomic absorption and classical wet chemical techniques. These methods are subject to dilution errors and are time consuming. This paper describes a method using energy dispersive x-ray fluorescence to determine magnesium, sulfur, chlorine, potassium and calcium in concentrated magnesium brines.

Energy dispersive X-ray fluorescence (EDXRF) analysis is finding increasing application in field assay of soils and stream sediments for geochemical prospecting; in borehole logging and core analysis for formation evaluation in mine control, and in on stream analysis for process control.

The analytical capability of the method is determined by the choice of source, the energy of the exciting radiation and the energy resolution of the detector. The nature of the samples, their atomic number and concentration are also important. However, the most dominant constraints on limits of detection are often imposed by the environment at the region of measurement: by the borehole condition, the nature of the rock face or the form of the particulates and stability of the fluid in a process stream.

In 1968 Consolidated Gold Fields commenced exploration below adit level at the Wheal Jane tin prospect in south-west England. During the surface exploration program the relatively recently introduced Portable X-ray Fluorescence Analyzer (“P.I.F.“) had been used for semi-quantitative scanning of drill core and for providing preliminary assays for tin on prepared core samples. The P.I.F. was introduced to the mining industry in 1965 (Bowie et. al. (1) and had originally been equipped specifically for tin analysis due to the current interest in tin exploration, the difficulties of chemically assaying for tin and the suitability of tin to X-ray analytical techniques. The instrumentation and principals of operation have been described by Bowie (1,2). At Wheal Jane the need to obtain rapid analyses to help control underground development and difficulties in visually recognizing the tin mineralization made the prospect an excellent testing ground for the practical aspects of the instrument.

A new hand-portable, microprocessor-based, multielement X-ray fluorescence analyzer is described. The instrument is light in weight (19 1b), completely self-contained and powered by internal rechargeable batteries. The detector is a special proportional counter whose room temperature energy resolution is sufficient to enable adjacent element X-rays to be deconvoluted with the help of the microprocessor. The instrument yields concentration readout of elements, corrected for matrix effects, in groups of four at a time. A number of field and laboratory applications to ore, solution and alloy analysis are described.

Recent advances in solid state electronic circuitry and cryostat design have allowed the development of a low power consumption, fully portable XRF analyzer. The potential of such an instrument in geochemical surveying was investigated using the Inax 540 portable XRF pulse height analyzer/computer combined with the Inax 600 portable Si(Li) detector/dewar/LCD display. These two units comprise a total XRF analyzer and will determine and display the analyses of up to 36 elements simultaneously. Apart from providing a freshly exposed surface, no sample preparation is required.

Data illustrating the problems in the calibration of the unit for whole rock analysis are presented. Specifically data are presented on Uranium bearing rocks where the Uranium occurs along fracture lines in the medium. Results are compared with known assay and sampling problems are discussed.

The intensity of Compton scattering, Fe, and Ca chatacteristic radiation can be used to estimate the amount of ash in coal by Xray fluorescence spectroscopy. Mo, W, and Cr radiation were used to study a suite of New Mexico coals, and the best results were obtained with Mo and W X-ray tubes. If the actual concentrations of Fe2O3 and CaO and the mass absorption coefficient, μ*, at the Compton wavelength of scattered Mo K radiation can be determined, the regression equation:

%Ash - 24.2μ* - 6.28(%Fe2O3) - 1.96(%CaO) - 3.4

estimates the ash content with an average error of 0.5% ash at 0.71Å.

The x-ray system used was a Columbia Scientific Industries Corporation Portable X-Ray Fluorescence Analyzer Minilab 700 with a ten millicurie curium 244 source. This source was chosen instead of the available 30mCi Pu-238 or the 30mCi Cm-244 source because it could be supplied under a general radioactivity materials license. The XRF analyzer could be powered by a variety of power supplies, either AC or DC. All data were evaluated and best curves were selected utilizing the Honeywell Time Sharing System. Programs from the statistical library were used with minor modifications.

The analysis of sulfur in coal has been brought to the attention of the general public in recent years primarily as it concerns the pollution problem with coal-fired electric generation. Although our application is different, some of the results would be equally applicable to that circumstance. The Cement Division of Florida Mining and Materials Corporation produces annually about 500,000 tons of both Portland and Pozzolanic cements, using a coal-fired rotary kiln. The sulfur content of coal used in the manufacture of clinker ranges from traces to as much as 6%. Sulfur content in a dry process plant such as ours (this is the modern energy-efficient type of cement plant) is very critical, and in our case must be limited to 1.2% max.

The elemental analysis of geologic samples such as rocks, minerals and coal ash is a complicated task because of their wide, complex compositional range. Energy dispersive x-ray fluorescence (EDXRF) can provide a rapid, accurate and precise way of analyzing geologic samples. Two approaches to reducing EDXRF intensity data to elemental concentrations are the empirical approach and the fundamental parameters (theoretical) approach. Empirical methods require numerous standards within restricted compositional ranges so can become complex, time consuming and, therefore, expensive if diverse suites of samples are to be analyzed for many elements. Fundamental parameters, on the other hand, requires knowledge of physical constants such as mass absorption coefficients, jump ratios and fluorescent yields, and only one matrix independent standard to calculate a calibration constant for each element making it an ideal approach to the analysis of diverse geologic samples.

Large-scale uranium exploration programs and industrial mine development require rapid and accurate analytical determinations of uranium and associated elements, at concentration levels ranging from a few parts-per-million to the per-cent range. The analytical requirements for associated suites of elements vary with application, but quality control and production practices in U.S. uranium mines and mills commonly require determinations of calcium, iron, copper, molybdenum, and vanadium, in addition to uranium. Previous work has demonstrated the capabilities of wavelength-dispersive x-ray spectrometric methods for many exploration applications. The purpose of this investigation is to assess the accuracy of x-ray determinations of uranium at ore-grade levels, utilizing both wavelength-dispersive and energy-dispersive spectrometers, and to discuss the acquistion of multi-element data.

Energy dispersive X-ray fluorescence (EDXRF) spectrometry has been applied to the high speed analysis of uranium ore for the purpose of grading and sorting truckloads of ore at the mine. The system is able to analyze uranium ore in less than 60 seconds with uranium levels ranging from < 0.005% to > 5% U3O8. Precision at the 2σ level for 200 ppm ore is better than 10% relative. Good agreement is obtained with wet chemical results in a large variety of rock matrices.

An algorithm is proposed as a model that describes the intensity concentration relation over a wide concentration range in binary and more complex systems involving absorption and enhancement. All the influence coefficients can be determined from fundamental parameter equations. Terms containing product of two weight fractions are always negligible in pure absorption cases, are often negligible in absorption/weak enhancement cases and are necessary but small in strong absorption/strong enhancement situations.

Once the magnitude of the coefficients is known, the range of application of the one- or two-coefficient models can be established with certitude.

Fundamental-parameters calculations can be made on a laboratory microcomputer fo r automatic treatment of interelement absorption and enhancement effects in x-ray fluorescence analysis. A new software package, called XRF-11, uses an efficient combination of fundamental parameters and alpha factors to compensate for any lack of measured reference materials, while taking full advantage of whatever standards are available, even just pure elements. In many cases, one multi-element standard is enough for accurate analysis.

The new XRF-11 software uses the same data base of absorption coefficients, fluorescence yields, etc. as the big-computer program NRLXRF, and combines theory with experiment in a consis tent way that is similar to, but more efficient than, the treatment used in NRLXRF.

An X-ray fluorescence program, similar to NRLXRF2 has been written by John Criss to fit into the DEC LSI-11 32K computer with floppy disk system. EDAX has combined the new version with our own software to create a set of FORTRAN programs, called “XRAY 95” to use with the new EDAX EXAM 9500 energy dispersive X-ray fluorescence system. The XRAY 95 program employs both fundamental-parameter equations and influence coefficient equations to optimize the matrix corrections for multicomponent samples. It is an extremely versatile program, using whatever reference standards the user provides and supplementing them with physical theory. The result is a practical approach to fast, accurate analysis.

On the basis of fundamental-parameters calculations of x-ray fluorescence intensities, it is predicted that relative intensities much greater than one can be observed from thick, homogeneous samples. That is , the intensity of a particular line from a mixture of two or more elements can be much greater than the intensity from the pure element. Similarly, the intensity from a mixture of compounds can be greater than the intensity from a pure sample of the compound containing the emitting element.

The error introduced by sample scattering in EDXRF analysis is evaluated by Monte Carlo simulation. This is accomplished by deriving a Monte Carlo model capable of simulating single Compton and Rayleigh scatters from the exciting photon source and from fluorescent X rays in homogeneous samples. The model also includes primary, secondary, and tertiary fluorescence events. (1) Results are given for Ni-Fe-Cr ternary samples for various exciting energies with and without scattering and indicate that errors as large as 2% can be attributed to this effect.

One of the uncertain aspects in the analysis of x-ray spectra is the determination of the proper background to subtract from the raw data. In those cases where the background is a smoothly varying funct ion of the x-ray energy, the application of a digital filter to the raw data will effectively remove the background leaving only the filtered peak information. These filtered peaks can then be fit by using a non-linear least squares method in conjunction with a suitably chosen mathematical model of the peak structure.

A recent surge of interest in sulfur in the environment has revealed the need for improved methods of analysis for sulfate, SO4=, in rain, freshwater and sediment interstitial fluids. Ion chromatography permits the rapid determination of SO4= in the ppm range (1 ppm = 1 mgL'1 = 10 μmol L-1 sulfate) on relatively small samples with good specificity. If a suitable instrument is available, this technique is a good choice for many environmental analyses. Other approaches to sulfate analysis are based on its precipitation with organic or heavy metal cations, usually barium or lead. The amount of precipitate formed may be determined by inter alia gravimetry, turbidimetry, radiometry (using 133-Ba), atomic absorption spectrophotometry (Ba or Pb detection), potentiometry (using a Pb++ ion selective electrode) , colorimetry, or by x-ray fluorescence spectrometry (Ba, Pb or S detection)(1). Because of our experience with x-ray fluorescence analyses, we chose to develop and test a procedure using that technique.