Five readily available fundamental parameters x-ray spectrometry programs were compared for their performance in a more or less routine laboratory environment. The programs were used to determine a variety of elements in cement, geological samples, low-alloy steels, nickel alloys and plant material. All data were acquired using energydispersive x-ray spectrometry. Criteria used in the comparison include accuracy, versatility and ease of use. Performance is given in statistical summaries.
LNTRODUCTIONAlthough quantitative x-ray spectrometry is a well established method, improved techniques for accurate analyses are still being sought. Measured x-ray intensities from a given element in a sample are frequently not linearly related to composition because of the influence of other elements in the matrix. It is necessary to correct analyte intensities for both absorption and enhancement by these concomitant elements. Matrix effects have traditionally been handled by multiple regression methods to obtain correction coefficients.'-' These correction coefficients are called alpha or influence coefficients and are often obtained from multiple regression of intensities against concentration of analyte and matrix elements in a suite of standards in which analyte concentrations bracket the range of interest. Because this approach is pragmatic, the resulting coefficients are frequently called empirical coefficients. Two common algorithms employed are known as the intensity correction model4 and the concentration correction model.' The concentration correction model uses the concentration of the analyte and matrix elements in the standards as the basis for the regression. If concentrations of non-analyte matrix elements are not known in the standards, the measured intensity of the matrix ele-
Oil was chosen as a new sample type for total reflection XRF. A thin-film sample procedure is described which requires minimal sample preparation. This involves first bringing the oil into solution with a volatile solvent, then pipetting microliter amounts on to the reflector substrate followed by evaporation of the solvent. The linear dynamic range was found to extend to at least 5000 ppm. Detection limit measurements show that minute sample quantities are necessary for optimal detection limits of nanogram amounts. A maximum sample loading is required for the best detection limits in ppm. Using an Mo or Cu x-ray tube, limits were found to be a few ppm or sub-nanogram amounts. The spectrometer was calibrated by an internal standardization method and the results of determinations on sample solutions were typically within 10% relative accuracy.
In recent years there has been substantial development of computer programs which permit the computation of elemental concentration in a variety of samples from basic principles of X-ray absorption and emission. These prog rates are generally called "fundamental parameter" programs. For many years these programs required large main frame computers for execution. An example is the well known NRLXRF program. For the past few years, programs have been available for minicomputers such as the PDP/11 OR LSI/11 systems. The XRF-11 program from Criss Software, the SAP3 program from Batelle Pacific Northwest Laboratory. and a variety of programs from the X-ray instrumentation vendors are examples. Some fundamental parameter programs have been available for several years that can be executed on microcomputers or the so-called personal computers. Some programs use an "effective wavelength" approximation of the tube output to simplify computations.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.