Symmetrization of INADEQUATE spectra
INADEQUATE spectra contain a special kind of symmetry which is difficult to exploit directly to increase S/N, as is done with proton correlation experiments. Isotope shifts also disrupt the ideal symmetry.
iNMR adopts a non-linear filter that removes the signals (and the noise) from the regions where no genuine signal can appear. The program starts from the known frequencies of the 1-D peaks. It then calculates the DQF frequency for every possible pair of nuclei. The 2-D spectrum is scanned at these frequencies. Only the peaks falling there and which appear in regular pairs are retained. The spectrum becomes much easier to inspect because of the drastic cleaning. As with any symmetrization method (and any non-linear filter in general) there are risks of false positives (artifacts created by the filter) and false negatives (genuine peaks that are suppressed). Both can be spotted through the interactive interface of iNMR.
Since this filter, among its many tricks, takes advantage of the symmetry of the experiment, we can call it “symmetrization”, though this is only part of the truth.
How to Safely Symmetrize an INADEQUATE Spectrum:
- Process the 1-D spectrum. Perform 1-D peak-picking. Do not include solvent and impurity peaks in the list of peaks.
- Process the 2-D spectrum. Add the 1-D spectrum as an external projection. Verify that the frequency scales of both spectra are correct and accurate.
- Select, from the main menu, the visualization mode Format > Contours. Adjust the intensity of the plot. Ideally, all peaks should be visible and random noise should be hidden. If this is not possible because the signals are weak, choose a compromise: show all the plausible signals, at the cost of including a little noise. The lowest contour will be the initial threshold of the filter we are going to apply. Peaks below the threshold will (initially) disappear.
- Choose the command Process > Symmetrize > Inadequate. You will see the filtered spectrum and a dialog over it. The dialog contains a checkbox called “filter”. Use this switch to compare the spectrum before and after the cleaning. It is important to verify what happens with a different threshold value: try moving the slider. Peaks that appear only at lower threshold levels could be noise. Treat them with suspicion. A comprehensive analysis of the spectrum is required to resolve any ambiguity.
-
Apart from the threshold, the filter requires three parameters.
They are less critical, yet it is worth verifying if their values are correct
for your compound and your spectrum.
To see the effect on the plot after changing a value, click the “filter” switch twice.
Splitting: this is the coupling constant. You can take an approximate value from similar known compounds. If the same molecule exhibits a wide range of couplings, set this parameter to the smallest value.
Linewidth: the average linewidth of a peak. If you cannot measure it, set this parameter to twice the digital resolution of the directly detected (single-quantum) axis.
Spread: the complete width, along the double-quantum axis, of the largest cross-peak. - Adjust the threshold until you achieve the optimal result. Close the dialog.
This filter can alter the apparent coupling constants.