"Deduced" baseline correction
Some baseline distortions arise from broad peaks due to exchangeable protons or a polymer component in the sample. Even though they are "real", they can be undesirable and interfere with integration. An example is shown here in the top trace.
A person can look at this spectrum and readily deduce where the "baseline" should be.
The same process can be accomplished by software. Some of the steps are similar to the procedure used by the FB routine.
The process involves automatically distinguishing regions of just baseline from regions containing peaks, interpolating where the baseline should go underneath each peak, then subtracting this deduced baseline from the spectrum. The process of determining the difference between noise and peaks requires an estimate of the noise level in the particular spectrum. NUTS automatically determines the rms noise of the spectrum. The spectrum is broken into segments, and a determination made as to whether or not each segment contains any peaks. A segment of the spectrum is determined to contain a peak if its (maximum – minimum) value exceeds a defined multiple of the rms noise. By default, Nuts will use as that multiplier the RM value (same as is used in peak-picking), but the user can specify a different value. By default, Nuts divides the spectrum into 64 segments, but the user can specify how many points each segment should contain.
This correction is implemented in NUTS as a modification of the QB command. When issued without parameters, QB applies a non-interactive polynomial correction, equivalent to that performed by FB using default parameters.
To use the "deduced" baseline correction, the user must first place Nuts into the non-2-letter command mode, by typing 2F. There are 3 optional parameters:
The first parameter specifies the number of points for each segment. The smaller the number, the more segments into which the spectrum is divided. If no other parameter is specified, the current value of the RM parameter will be used to determine which segments contain noise.
A second parameter can be specified, which is the value of the multiplier for the rms noise.
When QB is used with 2 numerical parameters, the baseline is calculated and subtracted, and the current spectrum is replaced by the corrected spectrum. (Users might want to enable Un-Do, to facilitate trial-and-error adjustment of parameters.)
The lower trace in the figure above was obtained using the QB command with the following parameters:
QB 32 8
In this case, each segment contained 32 data points, and segments were determined to contain peaks if their (maximum – minimum) data values exceeded 8 times the rms noise.
To examine the "deduced" baseline itself, the QB command is issued with the optional 3rd argument of "show". The result is shown in the lower trace. The current spectrum has been replaced by this calculated baseline. To apply the correction, the spectrum needs to be opened again and QB applied without the "show" parameter, or use Un-Do to back up one step.
Care in parameter selection is important to avoid distorting the low level lineshapes of peaks that can result in integration errors.
Last updated: 12/5/02