Peak width issues with generalised 2D correlation NMR spectroscopy

Two-dimensional spectral correlation analysis is shown to be sensitive to fluctuations in spectral peak width as a function of perturbation variable. This is particularly significant where peak width fluctuations are of similar order of magnitude as the peak width values themselves and where changes in peak width are not random but are, for example, proportional to intensity. In such cases these trends appear in the asynchronous matrix as false peaks that serve to interfere with interpretation of the data. Complex, narrow band spectra such as provided by 1H NMR spectroscopy are demonstrated to be prone to such interference. 2D correlation analysis was applied to a series of NMR spectra corresponding to a commercial wine fermentation, in which the samples collected over a period of several days exhibit dramatic changes in concentration of minor and major components. The interference due to changing peak width effects is eliminated by synthesizing the recorded spectra using a constant peak width value prior to performing 2D correlation analysis.