In case of an improper linear energy calibration, as it is shown in this picture, the user forgot to use the high-energy lines of 152Eu for the linear calibration.
These lines possibly will not be found by the nonlinearity analyzer, thus 'drawing off' the nonlinearity curve.
|
This figure shows a much better linear energy calibration for the same situation.
You can see that the calibration line placed such way that no any library lines is located too far away from the actual peaks.
This way the nonlinearity analyzer will probably be able to find those peaks and incorporate them onto the nonlinearity curve.
|
This is because the fitting algorithm has the freedom to rotate the curve part freely, in order to get a joint with minimum overall nonlinearity.
Please take a look at thisfaultymulti-isotope nonlinearity curve constructed from the two overlapping 152Eu and 56Co measurements.
We took that curve, and just for the sake of the demonstration, manually removed the 56Co lines which overlapped the 152Eu ones, this way created two measurements with disjoint regions of peaks. These two measurements cannot be used for a nonlinearity analysis.
In order to reduce the overall nonlinearity, the fitting algorithm applied a 'rotation' to the 56Co measurement, and flattened out the whole curve significantly, giving an improper result.
You should always avoid such situations, by adding another measurement or missing peaks to the nonlinearity analysis.
A possible solution would be to use a third measurement, which covers the middle region, while overlaps a bit with the first and the second measurements.
|