Publications Repository - Helmholtz-Zentrum Dresden-Rossendorf

To check the general quality of our analytical results, we started an intercomparison between four more or less common non-destructive X-ray based methods:

• low energy particle induced X-ray emission with protons of 2 MeV (PIXE)
• high energy particle induced X-ray emission with protons of 68 MeV (HE-PIXE)
• X-ray fluorescence with a portable device (µ-XRF)
• synchrotron-induced X-ray fluorescence (SY-XRF)
As test objects we selected six Roman coins with corrosion layers of different occurrence. Each coin was analyzed at three locations: original surface, surface with fully removed and partially removed corrosion layer, respectively.
We used two different external beam set-ups for PIXE: A “classical” one at a 2 MV tandem accelerator (BAM) and a high-energy one at the Helmholtz-Zentrum Berlin. The low-energy proton beam was extracted into air through a thin polyimide window (8 µm thick) and focused by a magnetic quadrupole doublet followed by a carbon aperture (Ø=0.7 mm). X-rays were energy dispersively measured by a Si(Li) detector [1]. There are some advantages in using higher energy protons [2] over low-energy ones: The protons can penetrate more deeply the material and the excitation probability for K-lines of heavy elements is bigger, resulting in better detection limits for those elements. At our set-up the beam is extracted into air via a thin Kapton foil and the X-ray signals are collected by a HPGe (resolution: 180 eV at 5.9 keV). Measurement times were 200 s. The data evaluation was done using the Guelph PIXE software GUPIX.
SY-XRF was performed at the hard X-ray beamline BAMline at BESSY. We analysed the Roman coins by our typical set-up [3]: A Si(111) Double-Crystal-Monochromator (DCM), and a W/Si Double-Multilayer-Monochromator (DMM) were used to produce a monochromatic X-ray beam of 32 keV. Signals were collected by a Si(Li) detector.
Last but not least, a mobile XRF-device equipped with a Mo-X-ray tube (30 W) and a polycapillary (spot size: 70 µm) was used to check to which extent a method that could in principal analyse objects in the field or in the collections can keep up with the three stationary (and more expensive) methods.
The four different methods produced tolerable to horrendous differences of quantitative results. Some discrepancies can be e.g. explained by variations of the analysed volume. For instance, a possible inhomogeneity of the sample will differently influence analytical results, if one changes the spot-size and penetration depth. The influence of the corrosion layer on the obtained concentrations also depends on the analytical depth of each method. Quantification procedures for each method should be adapted to these effects.
References: [1] I. Reiche et al., X-Ray Spectrom. 34 (2005) 42. [2] A. Denker et al., X-Ray Spectrom. 34 (2005) 376. [3] H. Riesemeier et al., X-Ray Spectrom. 34 (2005) 160.