Issue link: https://iconnect007.uberflip.com/i/239245
determining phosphorus content in en plating using xrf spectroscopy continues One aspect of the distance dependence often surprises the user: Newer model instruments with high-resolution semiconductor detectors are not at all suited for this approach, because the distances between sample and detector are significantly smaller. For this reason, distance uncertainties of the same magnitude, i.e., <50 μm, are significantly more serious (by a factor of 2–3!) than with proportional counter tube instruments. This must be taken into account when evaluating spectra with a direct phosphorus analysis (Section 2.3). 2.3 Direct Phosphorus Measurement (analysis of P-K- Peak) SDDs can achieve today what was impossible with even Si-PIN detectors of the last decade. The P-K radiation component can be detected reliably. Figure 4 graphically illustrates the measurement effect. Samples with different phosphorus contents show P-K peaks of different magnitudes. Their intensities are (nearly) proportional to the phosphorus content. The energy of the characteristic P-K radiation is only 2 keV. Due to the very critical dependence of the absorption coefficient on the photon energy, the information depth is just under 1 µm (Figure 5). As shown in Figure 6, the spectral signal of the phosphorus is not very large. For commonly used excitation conditions (W- Anode, 50 kV, 10 µm Ni-filter) and a typical sample of about 10% P, the P-K peak is 2000 times smaller than that of Ni-K. This combined with a rather poor signal/background ratio of only two provides unfavourable conditions for routine analyses. The signal/background ratio, which is so important for the measurement sensitivity of phosphorus, can be increased by a factor of 4 through "soft" excitation, as demonstrated by the yellow 10 kV-spectrum shown in Figure 6. Unfortunately, lowering the excitation voltage of the X-ray tube, which is favourable for the detection of phosphorus, is not effective for thickness measurement. The solution is to combine both types of excitation in one mea- Figure 3: Measured distance dependence of the %P measurement using a Fischerscope X-Ray XDLM-C4 for the sample from Table 1; the correct measuring distance is z = 0.5. The measurements are repeated several times, and the scatter provides a measure for the random (statistical) measurement error. 42 The PCB Magazine • January 2014