Fundamentals of Energy Dispersive X-ray Fluorescence Technology

The traditional use of X-ray fluorescence analysis (XRF) has its roots in geology. Solid samples were the first sample types to be analyzed using X-rays. Over the years, applications have expanded to include the analysis of alloys, various types of powders, liquid samples and filter materials.

Principle of XRF

The effect of X-ray fluorescence is based on the excitation of atoms in the sample. Unlike optical spectroscopy, the excitation involves interaction with the inner shell electrons rather than valence electrons, as indicated in the image of the Bohr model of the atom below.

X-ray Fluorescence begins with an excitation (or primary) X-ray, typically generated using an X-ray tube.

This excitation X-ray hits an inner shell electron of the atom and ejects the electron from the atom. An electron from a further outer shell fills the open position, and fluorescence radiation is emitted. The energy of this radiation is characteristic of the specific atom and indicates what atom is present in the sample.

As many atoms may be present in the sample, X-rays with different energies will be emitted (below). In an energy-dispersive XRF (ED-XRF) instrument, the fluorescence radiation is collected by a semiconductor detector.

The X-rays create signals in the detector, which depend on the incoming radiation's energy. The signals are collected using a multi-channel analyzer. The signals are converted into a spectrum (below). The y-axis represents the intensity of the peaks in counts per second, and the x-axis represents the emission energies.

The energies of the peaks can be used to identify the elements present in the sample.

The process handles each X-ray one by one at high speed. Modern detectors can handle 1 million counts per second or more. The spectrum can be recorded quasi-simultaneously. Even with a short measurement time, the spectrum can give sufficient information to calculate intensities to determine the sample's composition. A longer measurement time allows for better statistics, resulting in better precision and peak-to-background, thus improving detection limits. The intensities follow a "Poisson"-statistics.

A minimum of a few million counts should be collected for a highly precise analysis of an element's content. This is relatively easy if the sample contains a high concentration of an element and the detector can handle a high count rate. It will be more difficult if concentrations are low and the detection system can only handle a low count rate.

A combination of high sensitivity with low background is essential to obtain low detection limits.

Optimization of Excitation

Many applications for XRF require only a fundamental setup with a tube-sample-detector.

Optimizing excitation and detection systems is crucial for more challenging applications requiring high sensitivity and/or low detection limits.

High sensitivity can be achieved with a carefully selected X-ray tube. The key characteristics are the tube design (side-window, end-window, transmission window, etc.), tube power, and anode material.

The selection of the anode material is essential when high sensitivity for a specific group of elements is required. The following schematics show the effect of using different colors to represent different excitation energies.

Detection

A critical aspect of the detector is its resolution. This is usually given for a fixed energy to compare the performance data. The reference energy is that of Mn Kα as for testing detectors, typically an Fe-55 source is used, and this emits Mn Kα radiation.

Different detectors provide different resolution. Typical ED-XRF detectors include proportional counter detectors, Si-PIN detectors, and Si Drift Detections (SDDs). SDDs offer the best resolution. The advantage of a good resolution becomes evident if low concentration levels of an element with a fluorescence signal next to that of a high-concentration element must be determined.

Sample preparation

Traditionally, XRF is known to be non-destructive, but this is not always the case, and the sample prep has to be selected according to the analytical goal. The selected sample prep depends on the sample type and is certainly different for alloys, granulates, powders, or liquid samples.

Sample preparations are important because, depending on the energy of the X-rays, the depth from which we can collect the fluorescence radiation can be relatively small. In addition, this effect is also matrix-specific. Generally speaking, the heavier the sample matrix, the lower the information depth. To get an estimate of this effect, the "attenuation length" is calculated. This is the thickness from which the fluorescence signal is suppressed with a factor of 1/e.

Quantification

Depending on the complexity of the sample composition, matrix effects can make it challenging to determine the elemental content. Matrix effects are, for example, that the primary X-rays will be absorbed on their way into the sample. The same is true for the fluorescence radiation on its way out of the sample, as shown in the little image. In addition, other effects like secondary excitation must be considered.

Depending on the sample matrix, this will lead to different intensities and calibration curves.

Instrumentation

ED-XRF instruments are available in various configurations. And there is no unique answer as to which instrument is the most suitable.

---------------------------------

For a more comprehensive review of XRF detection technology, get our 13-page white paper, "The XRF Principle: The Fundamentals of Energy Dispersive X-ray Fluorescence Technology," featuring in-depth discussions, detailed reference figures, practical insights, and valuable criteria of 19 key application examples.