Introduction

X-ray fluorescence spectroscopy (XRF) is a non-destructive, fast and versatile method for elemental analysis. It is based on characteristic X-ray emission of the elements that has been bombarded with high energy radiation, such as X-rays or gamma-rays. XRF can be applied over a wide range of analyte concentration from 0.1 ppm and 100%, meaning it is applicable for different types of sample. XRF analysis covers the atomic numbers between 9-92, which means that nearly the whole periodic table is available for analysis (Figure 1). Technique is able to analyze up to 28 pre-selected elements simultaneously. The sample may also be in any shape or form, e.g. a powder, a metal disk, a shoe, a spring, oil or painting. For these reasons, it is one of the most widely used analytical technique in industry today.^[1]

Figure 1. Periodic table for XRF. Term LLD means the Lower Limit of Detection. Data from^[1]. Original periodic table template from Wikipedia (License: Public Domain). (Figure: Nea Möttönen.)

Theory

XRF is based on the excitation of electronic states that is caused by high-energy X-ray beam (primary radiation). Any electrons in inner cores can be ejected, and there are various electrons in the outer shells that can “drop” to fill the void. This orbital drop of electron releases a photon with an energy corresponding to the orbital energy difference (secondary radiation).^[2]

Each electronic transition inside an atom have its own specific energy or line. The three main types of transitions or spectral series are labeled K, L, or M, corresponding to the shell from which the electron was initially removed (e.g. an excitation from K-shell is labeled as K). The K series lines are of the highest energy, followed by L and M. The specific transitions are marked by the subscripts α, β, γ, etc. to point which upper energy shell was involved in the relaxation.^[2] For instance, CsKα - transition means that electron from L-shell was transitioned to K-shell. The characteristic nature of secondary radiation is based on unique transition energies between different elements.

Primary radiation is needed to excite the sample atoms, and this is typically obtained with conventional X-ray tubes - a vacuum tube that converts electrical input power into X-rays by bombarding a metal anode with electrons due to applied high electric field.^[3] Electrons excite the metal anode, thus awakening the secondary radiation of a bombarded metal - along with the continuum spectra due to velocity-slowing interaction between electron and anode-material (Bremsstrahlung radiation). Both ratiation types are seen in Figure 2.

Figure 2. Schematic representation of a Rh tube spectrum at 60 kV divided into discrete wavelength intervals. The peaks corresponds to the secondary radiation (Rh K-lines) and the continuous part is due to bremsstrahlung radiation. Figure from Wikipedia. (License: Public Domain.)

As presented earlier in Figure 1, the elements 1-8 are not easily determinable. Light elements have small transition energies approximately between 1-10 keV, so they require a primary radiation of correspondent energy: 120-1200 pm. Let's compare these values to Figure 2, and it is seen, that at least Rh-tube could not provide high enough intensity in the range of 120-1200 pm. The primary radiation of this energy range must be provided, if one wants to determine light elemnts of 4-8.^[4]

Energy- and wavelength-dispersive XRF

In energy-dispersive X-ray fluorescence (EDXRF, left side of Figure 3), secondary radiation of the sample is directly let to interact with the detector able to differentiate incoming photons by their energy. The incoming, secondary X-ray photon ionizes gas atoms inside the detector chamber, so the amount of charge produced (and measured) is proportional to the energy of the incoming photon - thus, photons are differentiated according to their energies. This type of detector is called as a proportional detector.^[5]

Figure 3. The schematic representations of EDXRF (left) and WDXRF (right). Figure: Jussi Nieminen

Wavelength dispersive X-ray fluorescence (WDXRF, right side of Figure 3) relies on the wavelength differentiation of secondary radiation (in analytic crystal) before entering to the detector. Basically, the secondary photons with different energies tend to diffract uniquely from analytic crystal - reaching the detector at different angles. The software calculates the specific photon energies based on Bragg Law of diffraction. The WDXRF-detectors work similarly to Geiger counters – in other words it does not sense the photon energy but only their presence.^[1] Photomultipliers are one of the common choices for the detector. These detectors multiply the current produced by incident light by as much as 100 million times in multiple dynode stages.^[6]

The main advantage of EDXRF over WDXRF is shorter analysis time, because all secondary radiation reaches the proportional-type detector simultaneously. EDXRF is also suitable for miniaturization (even to hand-held size). This is mainly possible due to the development of thermoelectrically cooled detectors (like XR100 Si-PIN).^[7] Miniatyrized devices are great for large and non-portable samples, e.g industrial steel products. For instance, Bruker S1 Titan XRF is a hand-held, pistol-size product that is able to separate steel alloys in under 3 seconds.^[8]

WDXRF systems have better resolution over EDXRF, because it can routinely provide working resolutions between 5 eV and 20 eV, depending on their set up, whereas EDXRF systems typically provide resolutions ranging from 150 eV to 300 eV. This clear difference is presented in Figure 4. However, additional reflection from the analytic crystal is a major drawback due to significant intensity-drop in secondary radiation that reaches the detector. This naturally leads to higher intensity requirements what comes to primary radiation source. Larger radiation source along with the analytic crystal makes WDXRF more expensive and certainly not a hand-held type device.^[9]

Figure 4. Resolution differences between Wavelength dispersive- and energy dispersive XRF. (Figure: Nea Möttönen)

A typical WDXRF measurement for 28 elements takes approximately 2-3 minutes,^[1] instead of seconds spent with EDXRF.

Chemical characterization

XRF is used in both qualitative and quantitative chemical analyses. Every element provide unique secondary radiation, so qualitative analysis is easy to perform. Unfortunately, the different matrix effects in variety of samples makes the usage of calibration-curves difficult, because one has to have external standards containing the analytes in a similar chemical environment compared to one's sample - in order to propagate similar matrix effects. The most important matrix effects are presented in Figure 5 – elements are not necessarily excited only by primary excitation originated from radiation source, but also because of other elements present in sample. Therefore, analytic signals might have a contribution from other elements. In the example, secondary radiation of Cr is awakened by the primary radiation and secondary radiation of Ni. Without the presence on nickel, Cr would excite less. In conclusion, one can have two samples containing same amount of chromium, but different Cr-signals due to chemical environment.

Figure 5. Diagram illustrating primary excitation of atom (elements) by absorption of tube radiation; secondary radiation; and tertiary excitation. Thick grey solid arrows represent primary radiation. (Figure: Nea Möttönen, inspired by reference^[1])

As mentioned before, XRF is very versatile tool. It is used e.g. in art research, where the selection of paint is tied on to history. For instance, Ti- and Pb- oxides have been used as a white pigment in oil paintings. However, titanium oxide was commercialized in 1920 to completely replace the harmful lead oxides. Therefore, the forgeries may be revealed if titanium is found from e.g. 1800 era painting.^[10]

In Figure 6^[11], microscopic paint samples have been collected from different parts of Rembrandt's Homer-painting. Samples have been analysed using Macroscopic X-ray Fluorescence imaging (MA-XRF). The results showed the composition of used paints, and the distribution of different elements found in the painting (Figure 7). However, it is not untypical to use portable EDXRF devices, because taking samples from the valuable painting could be ethically questionable.^[12]

Figure 6. Paint samples taken from different parts of Rembrandt's Homer-painting^[11]. (License CC BY 4.0)

Figure 7. MA-XRF has been used to analyse paint samples from Rembrandt's Homer. Visible light image (a), and corresponding MA-XRF maps: cobalt (b), nickel (c), lead (d), tin (e), copper (f), iron (g), manganese (h), and calcium (i)^[11]. (License CC BY 4.0)

WDXRF is used in applications, where taking a sample from the object is possible. For instance, ZSX Primus III+ device can be used for quantitative analysis of cement raw meal by the pressed powder method. Pulverized cement samples are pressed into aluminum rings at 120 kN to form pressed pellet specimens. These are positioned into this bench-scale WDXRF-device, and measured.^[13]

References