Performing automatic peak fit
The first step of the deconvolution is to locate the non-background regions of the spectrum, where peaks may occur. This is done by a peak search algorithm, which requires at least a rough FWHM calibration (Full Width at Half Maximum) for its convolution function to work.
At least an approximate energy calibration is also necessary, because HyperLab switches to Variable peak width method in the vicinity of the 511keV annihilation peak – and naturally for calculating exact peak positions in energy units.
Therefore it is always advisable to check the validity of the calibrations before the deconvolution process.
Checking the FWHM calibration
To display current FWHM calibration, select the Calibration / FWHM calibration menu item. Now the FWHM calibration editor window appears.
For a quick calibration check, click Chart tab at the top.
The currently used FWHM calibration curve is displayed here together with filled green squares as FWHM calibration points. HyperLab also esti­mates the width of the suspected peaks, and displays these values as gray crosses. The figure shows the case when the FWHM calibration results too low values.
If a large deviation is found between the current calibration line (red line), and the suspected peaks' FWHM, a mis-calibrated FWHM may be suspected.
In this case, the FWHM and/or the position values of the FWHM calibration points should be modified on the Calibration points page.
The figure shows the case when the FWHM calibration results too high values.
In case of proper FWHM calibration the majority of the suspected peaks' estimated FWHM value is in the neigh­bor­hood of the calibration curve.
Please note that large deviations may occur before the deconvolution, even for valid calibration, as peaks' FWHM values are just rough estimations and not fitted values at this time.
Much better agreement is expected after a successful fit, when peaks' FWHM values are fitted and displayed on this chart as gray triangles.
For further details see section FWHM calibration editor.
Checking Energy calibration
Exact energy calibration is important for sophisticated tasks, like isotope identification. However, when only a quick peak list is required for well-known isotopes, a rough energy calibration is enough before fit.
Your energy calibration is proper for fitting if the annihilation peak is in the 509 – 513 keV range. Otherwise select Calibration / Energy calibration menu item, and edit the calibration points' position or energy values in order to correct them.
See further details in section “Energy calibration editor”.
Please note that a finer energy calibration may only be constructed when the fit is completed, because exact, fitted peak centroid positions will be then only available.
Starting the automatic peak fit
Click the Autofit button or select Fit / Start automatic fitting menu item to initiate a fit for the whole spectrum. A progress window is displayed during the fit, listing the problematic regions, where the Reduced Chi-squared of the fit exceeds the value 3.0.
After the automatic fit is completed, the counts of the first fitted spectrum region is displayed, along with the fitted curves. The list of the regions is also refreshed on the right, with red exclamation marks beside the problematic regions.
It is advisable to always revise all of the regions at least once, because sometimes unphysical fits provide you a region fit with satisfactory Chi-squared, but the fit may be improved further by a trained spectroscopist. To revise the regions, use the double-arrow button on the navigation toolbar repeatedly, in order to walk through all of the fitted regions.
Copyright © 1998-2007 by HyperLabs Software Budapest, Hungary