Peak fits using multiple spectral regions: An example

 

Consider a dataset containing a number of spectral regions where the same set of regions are acquired repeatedly as part of a time-elapse experiment. The spectra recorded at a given time are subject to identical charging effects, however, each temporally dispersed set of spectra exhibits differing energy shifts due to sample charging effects. One means of charge compensating the data is to use the Al 2p peak shown in Figure 1 as a reference for the three peaks shown in the O 1s spectrum and shift each spectrum recorded at the same time by a common offset so that the Al 2p spectra are aligned throughout the time sequence. Alternatively, the positional relationship between the Al 2p and O 1s peaks can be maintained using a peak model that includes both the Al 2p and the O 1s peaks and then use the peak position constraint to guide the O 1s peak positions with respect to the Al 2p peak position during the course of optimisation.

 

Figure 1: Synthetic peaks defined for the O 1s region, which includes both O 1s and Al 2p peak, displayed by overlaying the O 1s and Al 2p spectra in a single tile. Inset Tiles display the same data but zoomed into each of the respective spectral regions.

 

Constructing the Peak Model

 

When two or more spectral regions are displayed in a single tile, the active spectrum within that tile is the spectrum for which the block name appears in the top text-field on the Quantification Parameters dialog window (Figure 1). The combined peak model for the O 1s and the Al 2p spectra must be defined on the active spectrum, which for the situation shown in Figure 1 is the O 1s spectrum in each row. Once the four peaks, three O 1s and one Al 2p, are defined on the O 1s spectrum and the two spectral regions for the O 1s and Al 2p data are overlaid in a single display tile, pressing the Fit Components pushbutton causes a dialog window to warn that multiple sets of data will be used to optimise the fitting parameters (Figure 2). Press the Yes pushbutton to continue with the optimisation based upon both sets of data.

 

Figure 2: Dialog window that indicated more than one spectral region will be used to perform peak parameter optimisation.

 

Viewing the Extended Peak Model on Overlaid Spectra

 

The synthetic peaks and corresponding residual displayed with the data is determined by a check box on the Tile Parameters dialog window shown in Figure 3. If the Display Comps from First Spectrum check box is ticked then the display shows only the components defined on the active spectrum. Note, a synthetic component is only displayed in energy intervals for which spectral data exists. Selecting the Display Comps from First Spectrum check box causes the overlaid spectral regions to be applied to the components defined on the active spectrum and therefore the components and residual for the extended data become visible in the overlay tile.

 

Figure 3: Display Comps from First Spectrum check box.

 

Copying Peak Models

 

The auto-fit option uses the display tile to determine when the peak model should derive from the active spectrum. The Copy and Fit pushbutton on the Components property page of the Quantification Parameters dialog window will copy the peak model from the active spectrum in the active tile and attempt to fit this model to all spectra loaded into the current list of display tiles. Note that the background regions must be defined before the Copy and Fit pushbutton is invoked since the copy action only concerns the synthetic peaks. To transfer and fit the peaks in the current example:

 

  1. Propagate any regions required to define the background for the data involved in the peak-fit. In this case both the O 1s and the Al 2p regions should be transferred to the other members of the experiment using the propagate mechanism.
  2. Select the two columns of O 1s and Al 2p spectra as shown in Figure 4.
  3. Press the toolbar display button . A scrolled-list of display tiles is created where each tile contains the spectra overlaid from each row of selected tiles in the right-hand browser view.
  4. Now press the Copy and Fit pushbutton on the Components property page of the Quantification Parameters dialog window. A dialog window offers information about how many tiles will be affected by the operation and provides a chance to abort the action. If the Yes pushbutton is chosen on the popup dialog window then the peak-fit from the active tile will be copied to each active spectra in the scrolled-list of display tiles and a fit to the data is performed.

 

Figure 4: List of display tiles showing each row of selected O 1s/Al 2p spectra display overlaid in preparation for the Copy and Fit operation.