IBPT - IPS - Industry Services

Polymers – Examples of services


Thin film reference samples for micro XRF
Figure 1: Spectrum of S10 standard (25keV exc)

An ideal reference sample for the calibration of the sensitivity of micro-XRF set-ups has negligible absorption of exciting and fluorescent radiation, high degree of uniformity and homogeneity and emits a multitude of non overlapping x-ray fluorescence lines over a broad spectral range with suitable intensity. Free standing thin films (FSTF) can meet these requirements; unfortunately after the discontinuation of NIST SRM 1832/1833 suitable reference materials are no longer commercially available. Therefore a project was started to develop FTSF into functional reference materials and first test samples consisting of Ca, Fe, Cu, Pb, Mo, Pd and La have been produced by stacking thin metal films using PVD at AXO Dresden.
As support we used “Ultralene” foil and recently Si3N4 membranes. The deposition technique assures very homogeneous layers and a large degree of flexibility regarding the choice of elements and mass densities. These samples show graded mass area densities in the (μg/cm²)-range.
Different reference sample foils (RS1, RS2, SF1, SF2, S10) containing S, Ca, Fe, Cu, Mo, Pd, La, Pb in different concentrations have been analyzed for homogeneity of layer graduation on macroscopic scale and with sampling using a micro-probe beam. The large area maps covered an area of 15 mm x 15 mm using a beamsize of 400 μm x 400 μm. For the micro probe beam a CRL (FZK-IMT) was used to produce a 2.3 μm x 4.5 μm spot, which yielded a quantity of the probed masses in the order of 5fg- 500fg. By using a beam intensity monitor (pin diode) placed after the sample for triggering spectra recording times a measurement reproducibility (after subtraction of statistical noise) of typically 0.2% was obtained for large area scans and 0.5% for micro probe beam.
The time per spectrum ranged between 10s to 500s depending on sample and conditions.
For area scans and micro probe measurements the variance caused by sample and experiment was calculated by subtracting the count statistic variance from the total variance. The results show a good uniformity without marked long range modulations of the fluorescence signal. No degradation over time has been observed (5 month). The observed density gradient of lead is probably an artefact caused by a provisional collimator, which contained lead. This assumption is corroborate by tests with a lead free collimator. The micro probe homogeneity data are in general very close to the reproducibility. For the S10 Si3N4 sample hot spots of Cu and Fe were observed, but this is probably caused by improper sample handling (Scotch tape). For subsequent measurements on theses samples the frame holder is being improved.
We observed that a 10keV beam focussed with a XOS polycapillary into an high intensity microbeam (18μm spotsize, ~1011ph/s) rapidly caused mechanical alterations of the foil. This was not observed when the beam was focussed with a CRL (2.3 μm x 4.5 μm spotsize, 2*109ph/s. As a consequence a new sequence of test samples was started based on submicrometer thick Si3N4 Membranes as film supports.
The films showed very good suitability for the characterisation of a confocal setup[1] in terms of elemental sensitivity and energy dependent depth resolution. Furthermore they are very well suited to obtain minimum
sampling masses.
XRF and ICP-OEMS measurements showed that multimetal standards were successfully produced via PVD on thin polymer foils and Si3N4 membranes with a high degree of homogeneity and reproducibility.
SF1, SF2, S10 cover large spectral range without line overlap, emission lines have similar intensities and
the absorption of exciting and fluorescent x-rays is very small (below 2% in standard SF1 for E>3keV).