Curve fitting of spectroscopic data is a powerful technique that can be used to facilitate data analysis and interpretation. Especially peak analysis of overlapping peaks can be done in case the real peak area needs to be determined. A set of ideal peaks is fit to match the overall shape of the real spectrum as shown in the following:
The command Curve Fitting is applied to one 2D spectral data object and is a tool to model peaks as a sum of individual spectral contributions.
The User adds the required number peaks first, e.g. repeated times pressing the Plus
button, either for a desired area or the whole spectrum. A peak added by mistake can be simply removed with the Minus button. As a special feature, each peak can be fitted using its own Response Curve Shape.
With the sufficient number of peaks the estimation is initiated pressing the Auto Fit button, applying one of the Levenberg-Marquardt algorithm for optimization.
In a final step, a fitted spectrum of estimated peaks will be created and derived information like peak area, etc. is calculated too.
Clicking the Reset
button erases all selections.
Pressing then Calculate
button creates a peak table based on actual estimation status and shows the peak areas accordingly.
A curve fit is performed as follows:
Curve Fit menu command
To perform a curve fit using the menu command, please follow the instructions below:
Activate the 2D data object you would like to fit.
From the Mathematics menu, select the Curve Fit command.
Click Plus button to add a new peak to the list of estimated peaks.
Adjust all necessary parameters in the Mathematics tab.
Repeat from step 3 until the optimal number of peaks is defined.
Press the Auto Fitbutton.
Clicking the Calculatebuttoncreates a new peak table with detailed peak information of the estimated peaks.
Note: in some cases it is recommended to perform a baseline correction before applying the curve fit.
