X-ray fluorescence (analysis) – XRF(A)
X-ray fluorescence analysis is based on experiments of Richard Glocker and Hans-
Wilhelm Schreiber. Today, XRF is one of the most used methods for qualitative and
quantitative determination of the elemental composition mainly on account of the
non-destructive analysis of solid samples. Special application can be found in metal
industry, in geology and mineralogy and in the examination of glass, ceramic and
building materials as well.
X-ray fluorescence analysis
In XRF samples are excited to fluorescence by irradiation with characteristic X-rays
and by bremsstrahlung. Therewith, electrons are transferred from lower levels of an
atom to higher levels. By filling up of the produced holes by electrons from outer
spheres electromagnetic radiation (⇒ characteristic secondary radiation) is emitted,
which is recorded by a suitable detection system. The lower limit of detection (LLD) is
in the range of 1-100 µg/g (ppm) and a function of various parameters (e.g. atomic
number of the element, energy of the excitation radiation, sample properties).
Production of X-rays
As sources of the primary X-radiation are used:
• X-ray tubes,
• radioactive nuclides,
• electron accelerators (synchrotron radiation).
The device in the laboratory is equipped with an X-ray tube, why these are described
more in detail: Electrons produced in a glow cathode are accelerated in a high
voltage (about 1 kV) and impinge on an anode (or anticathode) with a high-melting
material (e.g., Cr, Cu, Mo, Ag, Rh, Pd, Ta, W). The impinging electrons generate two
kinds of X-rays:
• Bremsstrahlung: it is produced in the stopping procedure of the projectile
electrons in the anticathode (broad band).
• Characteristic X-radiation: will be emitted in case of the transfer of the excited
atoms of the anticathode material into the ground state (selective energy
levels produce a line spectrum). The excitation can be performed in various
electron levels and therefore the lines are named respectively (Fig. 1).
Fig. 1: Generation of X-rays.
For generation of X-rays an electron vacancy in an atom have to be produced. This
excitation process can be performed by X-rays and by charged particles (e.g.,
electrons, protons), as well. Therefore, a threshold energy is necessary. In measuring
the absorption efficiency as function of the energy so-called absorption edges are
observed (Fig. 2). At energies below the absorption edge the probability of absorption
is low and increases drastically at the threshold energy value. With the increase of
the absorption probability the emission of characteristic X-rays is increasing as well.
Increasing the primary energy furthermore, the probability of absorption and the
intensity of emission is decreasing again.
As a result, the excitation is most effective at energies slightly above the absorption
edge. At increasing differences between the excitation energy and the absorption
edge the excitation probability and the emission intensity are decreasing.
Consequently, for each element a specific energy is given for a most effective
excitation (resonance excitation).
Fig. 2: Absorption edge (schematic) for manganese.
In practice it would take experimental expenditure and would be expensive, to use for
each element an optimal X-ray tube. Therefore, often secondary targets are used,
which are positioned between the X-ray tube and the sample. The periodic system of
the elements (PSE) can be divided into three regions: heavy, average, and light
elements. In each region the excitation is most effective by one of the secondary
targets mentioned below. According to the material a secondary target can
• scatter the primary radiation onto the sample. In that case the sample is
excited by radiation from the X-ray tube (for example: Al2O3 as secondary
target – heavy elements).
• be excited to fluorescence. In that case the sample is excited by the
fluorescence radiation from target (for example: Mo as secondary target –
• filter a special energy region of the X-ray tube spectrum. In that case the
sample is excited by radiation of this energy region (for example: High-
oriented pyrolytic graphite (HOPG) as secondary target – light elements).
Emission of electromagnetic radiation after absorption/excitation of/by
electromagnetic radiation is called fluorescence. By excitation with particles (e.g.,
electrons, protons) the production of electromagnetic radiation is just called emission.
The atom with an electron vacancy is an excited system, which deexcite by emission
of an X-ray or an Auger electron.
The fluorescence yield (w) is defined as the ratio of emitted X-rays (nf) to the total of
electron transfers (n):
w = nf/n.
The fluorescence yield (w) and the probability for emission of Auger electrons (a) are
added to unity:
w + a = 1.
For the energy of the Kα-lines a relation is given to the atomic number of the excited
element (Moseley´s law):
λ (Å ) =
∆E ∝ Z 2 or ∝Z
λ ∆E (keV )
(∆E = energy of the emitted X-ray photon, λ = wavelength of the emitted X-ray
photon, Z = atomic number)
Using Moseley´s law in the spectra after an energy calibration the elements in the
sample can be identified. For heavy elements, additionally, the L-lines can be used
taking the intensity ratios into account:
Kα : Kβ ≈ 4 :1 Lα : Lβ ≈ 1 :1.
The intensity of a fluorescence line is proportional to the amount of the element in the
sample (excited region of the sample). This intensity is influenced by other effects:
• the excitation efficiency changes with the difference of the energies of the
excitation and of the fluorescence radiation (resonance effect),
• as the exciting and the emission radiation are absorbed in the sample material
(matrix effect: absorption), the reference samples have to be elementally
composed most identically with the samples to measure,
• fluorescence radiation can excite elements with atomic numbers less than two
units (matrix effect: secondary enhancement).
Therefore, it is necessary to calibrate the system with reference samples (samples
with known elemental concentrations) and to plot calibration curves (intensity in
relation to the concentration) for each element.
There are existing to different technologies to obtain XRF spectra
Energy-dispersive X-ray fluorescence analysis (EDXRF)
In case of the absorption of fluorescence radiation in an energy-dispersive detector
(in most cases semiconductor material as Si or Ge) electron-vacancy pairs are
produced. The number of them is proportional to the photon energy. The counting
rate is measured as function of the energy. In the actual case a Si(Li) detector is
used. For optimal excitation of special elements or for decreasing the background
intensity filters are inserted between the X-ray source and the sample.
Wavelength-dispersive X-ray fluorescence analysis (WDXRF)
In this case the fluorescence radiation is guided through a collimator and after
spectrally dispersion by diffraction in a mono-crystal detected by a proportional or
scintillation counter. The counting rate is measured as function of the diffraction
Equipment technology: EDXRF
Look in sect. Fundamentals: Production of X-rays
Selection of the filter material
The radiation produced in an X-ray tube is composed of the characteristic radiation of
the anode material and of a broad band of bremsstrahlung. This spectrum is modified
by filters to absorb the characteristic lines (optimal absorption efficiency by a material
with an atomic number which is less by one or two units in comparison to the X-ray
tube anode material).
In the used Spectro XLAB 2000 spectrometer the main parts (tube, secondary target,
sample, detector) are positioned in a cartesian geometry to minimize the scattering
background. Tube, secondary target and sample are placed in a plane (2-
dimensional) and the angle tube-target-sample is fixed to 90° (Fig. 3: zy-plane).
Therefore, the sample is excited by polarized radiation. The detector is positioned in
90° to this plane (Fig. 3, x-direction), the backgr ound from the X-ray tube is now
minimized because of the two time polarization and the recorded spectra show
optimized fluorescence intensity.
Fig. 3: Cartesian geometry
Energy dispersive detectors
• In a Si(Li) detector the intrinsic zone is produced by drifting of n-type Li in a p-
type Si. This detector is applied by a voltage of about 1 kV. The electron-
vacancy pairs are separated in the i-zone by the high electrical field and
produces an electrical signal. For stabilisation of this pin-system the detector
(3 – 5 mm thick) is cooled with liquid N2. The detector itself is separated from
the atmosphere by a thin foil (e.g., Be).
• In silicon drift detectors (0.3 – 0.5 mm thick) the signal is collected in the
middle of the system on a small anode. Therefore, the electrical capacity is
smaller with the consequence that higher counting rates can be processed,
that the energy resolution is better and that shorter collection times are
Equipment technology: WDXRF
Look in sect. Fundamentals: Production of X-rays
The collimators which are necessary to guarantee good spectral resolution are thin
tubes or lamella (soller slits).
For spectral separation of the fluorescence radiation Bragg diffraction at mono-
crystals or multilayers are used. Varying the angle by rotation of the diffraction device
the spectra are recorded. In an efficient system up to 6 mono-crystals and multilayers
n⋅λ = 2d⋅sin(θ)
n = order
d = lattice space
θ = angle of relection
λ = wavelength
Detectors for WDXRF
• Scintillation counters are used for detection of small wavelength (high energy)
emitted from elements with atomic numbers ≥ 26 (Fe). The detector material
usually is NaI doped with Tl. The absorbed X-rays are transmitted to light in
the optical spectral range. The light quanta produce electrons in a
photomultiplier diode which are multiplied to a factor of about 106. This current
pulse is recorded as a single event.
• Proportional counters (filled with Ar or a mixture Ar/CH4) are used for the
detection of long-waved radiation, which is emitted from elements with atomic
numbers from 9 (F) to 25 (Mn). The photoelectrons are guided to an anode
where an electrical pulse is produced.
The differences between these two methods originate from the different detection
In EDXRF the spectra are collected simultaneously in the whole energy scale. In
WDXRF the spectra are build up sequentially with the rotation of the diffraction
This parameter usually is described by the “full width at half maximum = FWHM” of
the Mn Kα-line at 5.9 keV. For EDXRF typical values are 150 eV, for WDXRF values
down to about 10 eV are possible. From that point of view WDXRF is advantageous.
Because of the intensity loss in collimators and diffraction devices in WDXRF the
intensity input of the X-ray tube have to be high. Usually, the power of X-ray tubes in
WDXRF are about 2 orders of magnitude higher in comparison to those used in
EDXRF usually is the less expensive mode in comparison to WDXRF.
Manual for the experiment
1) Seminar/lecture: Fundamentals of XRF – 1.5 h
2) Discussion about some fundamentals – 0.5 h
3) Accomplishment of the experiment, measurements and evaluation – 2 h
4) Short report (≈ 2-3 pages) of the experiment, of the measurements and of the
results – 1 h
5) Short discussion