X-rays are a form of electromagnetic radiation having wavelengths in the range of approximately 0.1 to 100 angstroms, which correspond to energies of 124 to 0.124 keV. X-ray tubes are almost exclusively used in modern XRF spectrometers as the source of the primary x-ray beam. When this primary beam of polychromatic x-rays is directed onto a test specimen, the x-rays interact with electrons of the constituent atoms present in the sample. One of the key interactions is the removal of inner shell orbital electrons of the atoms when the energy of the incident x-rays sufficiently exceeds the binding energy of these electrons. When this occurs, electrons from outer orbital shells of the atom transition to fill the vacancies created in the inner shells. These transitions result in the emission of secondary x-rays by the atom with energies corresponding to the difference in binding energy between the shell where the vacancy occurred and the shell where the electron transition originated. The discrete energy or wavelength of these secondary x-rays is essentially unique to (characteristic of) the element which emitted them. The characteristic x-rays thus serve as a signature of the elements that are present in a sample, and the measurement of these x-ray energies or wavelengths by the XRF spectrometer is the basis for qualitative analysis. Measurement of the intensity or count rate (number of x-ray photons detected by a counter per unit time) of the various characteristic x-rays emitted by a sample can be correlated with the concentration of each element present, so the count rate data provide the basis for quantitative analysis.

In addition to the characteristic x-rays emitted by the sample under investigation, a certain portion of the primary x-rays from the x-ray tube is also scattered from the surface of the test specimen, mainly from interactions with loosely bound outer shell electrons of the constituent atoms. Like the primary beam itself, these scattered x-rays are made up of a continuous band of wavelengths. This scattered primary radiation is the main contributor to the background of the observed x-ray spectrum. The intensity of the scattered radiation varies in accordance with the average atomic number of the test sample, such that the spectrum of samples containing a large portion of heavy elements, which are more absorbing of primary x-rays, exhibits a relatively insignificant background level while the background of samples composed mainly of light elements will be quite high and deteriorate the ability to measure components of low concentration in the sample due to poor peak-to-background ratio.

The field of XRF spectrometry is divided into two major categories based on instrument design, wavelength-dispersive x-ray fluorescence (WDXRF) and energy-dispersive x-ray fluorescence (EDXRF). The two differ primarily with respect to the manner in which x-rays emitted by or scattered from the test specimen are detected. In the WDXRF spectrometer, x-rays are collimated with a finely spaced parallel plate slit assembly to yield a near parallel beam of x-rays that then strikes the surface of a flat analyzing crystal (i.e., a single crystal or a layered synthetic multi-structure that replicates the behavior of single crystal). In accordance with Bragg’s Law, x-rays of different wavelength striking the crystal will diffract (“reflect”) at different angles when certain geometric conditions are satisfied. The crystal thus serves as a means of monochromatization of the x-rays, dispersing or separating them into their discrete wavelength components so that they can be identified and quantified, somewhat akin to the separation of white light (visible light) by a prism. A detector can be placed at a certain angle(s) to measure a specific wavelength (i.e., a specific element) or the detector can be attached to a goniometer and scanned through a desired angular range to produce the x-ray spectrum of several different wavelengths. In the latter case, the x-ray spectrum is typically plotted as x-ray intensity in kilocounts per second versus diffraction angle, and the spectrometer is referred to as a sequential type since the x-ray wavelengths are measured one after the other. Once the diffraction angles of the characteristic x-rays are measured, their wavelengths are calculated through the Bragg equation (taking into consideration the known interplanar spacing of the analyzing crystal) and compared to a published database of x-ray wavelengths of the elements in an automated fashion to reveal which elements are present in the sample, with the measured intensity for each wavelength proportional to the concentration of the element. In practice, the sequential spectrometer is outfitted with multiple analyzing crystals on a turret-like assembly in order to cover various wavelength ranges (various elements) with optimum performance. The rotation of the analyzing crystal to vary the angle of incidence (theta) is coupled to that of the detector (two-theta) to provide the proper conditions for diffraction for all possible x-ray wavelengths. Sequential WDXRF spectrometers are typically equipped with two detectors, a gas-filled proportional counter that is most efficient for x-rays of lower energy (longer wavelength) and a scintillation counter that is useful for x-rays of medium to high energies. Both detectors operate at room temperature and can accommodate high count rates without loss of linearity. Although their physics of operation differs significantly, each detector outputs voltage pulses that are proportional in amplitude to the energy of the incoming x-ray photons.

 Simultaneous WDXRF spectrometers differ from sequential models in that they are equipped with a set of “fixed channels” positioned annularly to the sample that are dedicated to measuring a specific, predetermined group of elements of interest. Each individual channel is configured with a curved crystal, slits and detector to measure one particular element in optimal fashion utilizing focusing beam geometry, which provides higher intensity, higher resolution and lower background than equivalent measurements in a sequential WDXRF system with parallel beam, flat crystal geometry. These systems are less flexible than sequential models and are generally more costly due to the extra hardware associated with multiple fixed channels, but the systems have the key advantage of high speed of analysis since all elements of interest are measured at the same time rather than one at a time. For this reason they are typically used in time-sensitive testing applications where it is sufficient to measure a fixed array of elements for purposes of process control such may be needed in a steel mill, aluminum foundry, mining operation or cement plant. A scanning goniometer (moving detector) is sometimes added to these systems to increase flexibility in terms of which elements can be measured. Since each fixed channel measures a discrete wavelength (a particular peak in the x-ray spectrum), it is not possible to make off-peak measurements of background intensity to obtain the net intensity of the peak of interest as sometimes required for accurate quantitative analysis. This can be overcome by adding additional fixed channels dedicated to measurement of background for selected elements, but these extra channels further increase the cost of the systems. Taking advantage of focusing beam geometry, some systems are innovatively designed with movable slit assemblies, which allow both the peak and the background intensity to be measured using the same channel for reduced hardware costs; however, this approach adds to the time of analysis since separate measurements have to be taken with the slits in the peak and the background positions.

Detection of x-rays in EDXRF instruments is accomplished using a solid state semiconductor detector without pre-separation of the x-rays from the sample by crystal diffraction. These detectors and associated processing electronics have the ability to directly determine the distinct energies of the incoming x-rays, and thus provide the speed advantage of simultaneous measurement of the “full” x-ray spectrum (or a very wide range of elements), not just a pre-defined set of elements as measured using simultaneous WDXRF spectrometers. Commonly used detectors in EDXRF include the lithium-drifted silicon or Si(Li) detector, the PIN diode detector, and the silicon drift detector or SDD. A high-purity germanium detector or HPGe is also used in selected models. Cryogenic cooling using liquid nitrogen or thermoelectric cooling is required for these detectors to reduce electronic noise, and the cryostats must be vacuum-isolated. Particularly when thermoelectric coolers are employed, EDXRF systems can be constructed in considerably more compact fashion than WDXRF units. Their smaller size and simpler design lower the cost of manufacturing these systems and make them more “field portable” than their WDXRF counterparts and more amenable to operation in harsh environments. Compared to WDXRF systems, EDXRF systems have lower spectral resolution except in the case of higher energy x-rays, have poorer sensitivity for x-rays of low energy (i.e., poorer lighter element sensitivity), and have more pronounced count rate limitations. Thus, they do not provide as high of precision and accuracy as WDXRF on average, but are still extremely valuable and flexible for a myriad of multi-element analysis needs. The thin beryllium entrance window (8–25 mm) that provides the vacuum and light seal for EDXRF detectors preferentially absorbs x-rays of lower energy emitted by the lighter elements, precluding analysis of elements lower in atomic no. than 11 (sodium). By contrast, the ultra-thin entrance window (e.g., 0.6 mm polyester film) used in the WDXRF gas flow-proportional counter allows far better transmission of low energy x-rays, extending the useful range of WDXRF spectrometry down to atomic no. 4 (beryllium) when equipped with suitable analyzing crystals and high performance x-ray tube.

The stationary arrangement of components used in EDXRF spectrometers is ideally suited for geometric configurations that exploit polarization phenomena to reduce background. One unique feature of certain EDXRF models is an indirect excitation mode where the primary x-ray beam from the x-ray tube irradiates a secondary target, often a pure metal, and the characteristic x-rays emitted by the target material are used to irradiate the test sample. Used in tri-axial, orthogonal “Cartesian” geometry, this approach eliminates scattered primary x-rays through polarization effects for significant reduction in spectral background and associated improvements in sensitivity and detection thresholds. Employing a strategic selection of secondary targets of different materials allows optimization of excitation efficiency across the Periodic Table. Barkla scatterers (polychromatic) and Bragg reflectors (monochromatic) are also employed as polarizers in systems of this design for further flexibility in optimizing excitation for various energy ranges.

EDXRF and WDXRF are each commonly thought of as a “bulk” method of analysis rather than small-area or microanalysis techniques, with measurement diameters from about a millimeter to a few centimeters routinely employed. In more sophisticated systems, the measured area can be varied to fit the application using computer selectable apertures or collimators of different sizes in the optical path. These accept x-rays from a limited region of the sample and block acceptance of x-rays from the remaining areas. The sizable decreases in x-ray intensity encountered with progressively smaller “pinhole” collimation makes measurements in the micro or near-micro realm impractical. Innovations in the use of monolithic polycapillary devices as x-ray optics with micro-focus x-ray tubes have accelerated the field of micro x-ray fluorescence spectrometry (micro-XRF). These devices channel x-rays via total external reflection at shallow grazing angles through a bundle of subtlety curved and tapered, small, hollow glass tubes, providing a focusing effect. The focusing of x-rays by the polycapillary yields high gains in intensity compared to pinhole collimators and provides beams as narrow as 10–25 mm. This enables analysis of individual particles or fragments of fine size, characterization of small defects or inclusions in a larger sample, and automated multi-point analysis and 2-D spatial mapping for systems outfitted with movable, high precision sample stages and video control systems. Micro-XRF systems with confocal geometry use separate polycapillary focusing optics, one for the incident beam (excitation optic) and one to collect x-rays emitted by the sample (collection optic). The confocal systems are capable of analyzing either the surface of the sample or the interior by moving the confocal analysis volume along the Z-axis, and thus are unique in their usefulness for elemental depth profiling (3-D spatial distribution). Micro-XRF spectrometers classically use energy-dispersive detection systems given the need for fast, simultaneous acquisition of the x-ray spectrum in producing detailed elemental maps, thus the elemental range of these systems, as in conventional EDXRF spectrometry, is sodium through uranium. The micro-XRF technique often requires minimal sample preparation, and there is no need to apply a conductive coating to samples as in scanning electron microscopy (SEM) since no sample charging effects are encountered when inducing fluorescence with x-rays (XRF) versus electrons (SEM). Micro-XRF analysis is better suited than SEM to samples with larger-scale features or larger areas of interest and can provide better quantitation in many instances.

References

J. Heckel, R.W. Ryon, Handbook of X-Ray Spectrometry, 2^nd Ed., CRC Press (2001), pp. 603-630.

C. MacDonald, Focusing Polycapillary Optics and Their Applications, X-Ray Optics and Instrumentation, Hindawi Publishing Corp. (2010)

P. Brouwer, Theory of XRF, 3^rd Ed., PAN. B.V. (2010)