One method of peak deconvolution is to utilise a knowledge of known relative intensities and energies of an analyte’s characteristic lines in conjunction with the detector’s energy resolution. Hence in the above situation, the intensity of the MnKa peak and its profile is first evaluated. Then the predicted profile for its corresponding Kb is generated. Part of any deconvolution procedure includes the estimation of a background profile since net intensities, by definition, should be free of both background and spectral overlap effects.
The generated MnKb profile is subtracted prior to evaluating the spectral content within the FeKa ROI. This method of spectral deconvolution is a sequential procedure with increasing energy. It is assumed that the first peak in the sequence is free from interferences. The success of this method depends upon well tabulated spectral data (relative intensities and energies) and a well calibrated energy scale i.e. the measured peaks occur at their expected energy locations. Only the most sophisticated software will compensate for changes in tabulated intensity ratios due, for example, to the presence of another (major) element’s intervening absorption edge. Another, of the many, spectral deconvolution methods is to establish a “library” of pure element spectral profiles. These “standard profiles”are then “fitted” to a measured spectrum profile.
The method of peak deconvolution used will depend upon the energy resolution of the detector used. It is clear that the sequential procedure previously described would not be appropriate for data obtained using a gas proportional counter.
“Continuous” WDXRF scans represent the changes in signal output from a RATEMETER (Y-axis) with respect to changing goniometer angle (X-axis). The software can translate this angle into wavelength or energy. Even when “step-scanning”, the intensity scale (Y-axis) is invariably represented as count rates (cps). Common terminology denotes: Rp or rp count Rate Peak Rb or rb count Rate Background Rp - Rb or rp - rb net count Rate The net countrate is calculated from: The breadth of peaks (and ultimately spectral resolution) is a function of the dispersion efficiency and degree of collimation. A peak’s distribution approximates to a Gaussian distribution and its area contains the maximum signal available for an analyte line. However since it is not possible to simultaneously integrate over this area (and it would not make sense to do so), single position measurements are made at the peak position and, where necessary, also at a suitable background location. With WDXRF methods, for maximum reproducibility it is important to ensure that the spectrometer re-sets to the same peak position at all times.
The total counts contained within a standard analysis window or ‘ROI’ prior to any processing (i.e. the as-measured “raw” data) can include a significant background contribution. This background may be solely due to scattered primary radiation or a combination of this and spectral overlap. The raw data counts are often referred to as the “GROSS” counts.
The difference with WDXRF is, that the count rates “rp” & “rp” are obtained by integration over a certain area (energy range; see blue-coloured ranges in figure) rather than measuring a peak height.
For the situation illustrated above, the following relationship holds for the defined areas:
where it is clear that “rp” in EDXRF can be considered equivalent to “rp” in WDXRF.
- The background is determined after applying the fitting algorithm.
Change/include lines, release line ratio’s etc, till a perfect spectral fit is obtained.
- Background fitting can sometimes be improved by selecting a different filter and/or kV/mA and re-measure the standards.
- A separate background measurement can be done, using ‘R.O.I.’
- If a separate background correction is used, make sure there are no lines present at the selected background position.
The total counts contained within a standard analysis window or ‘ROI’ prior to any processing (i.e. the as-measured “raw” data) can include a significant background contribution. This background may be solely due to scattered primary radiation or a combination of this and spectral overlap. The raw data counts are often referred to as the “GROSS” counts. The difference with WDXRF is, that the count rates “rp” & “rp” are obtained by integration over a certain area (energy range; see blue-coloured ranges in figure) rather than measuring a peak height.
For the situation illustrated above, the following relationship holds for the defined areas:
where it is clear that “rp” in EDXRF can be considered equivalent to “rp” in WDXRF.
- The background is determined after applying the fitting algorithm. Change/include lines, release line ratio’s etc, till a perfect spectral fit is obtained. - Background fitting can sometimes be improved by selecting a different filter and/or kV/mA and re-measure the standards. - A separate background measurement can be done, using ‘R.O.I.’ - If a separate background correction is used, make sure there are no lines present at the selected background position.
我的社区
签到
