[nmrlearner]

NMR Verification and Information overload

NMR is probably one of the most important tools that are at the chemist’s disposal for, structure elucidation studies. In the particular case of 1H-NMR, we usually rely in the analysis of all detectable resonances, typically in the FT spectrum, though time domain based approaches are also possible. Accurate knowledge of all the parameters associated to each resonance is key in the processes of structure confirmation and elucidation.
In this sense, it is important to highlight the point that this knowledge is usually limited by the resolution power of the data evaluation technique employed. Obviously, there is always the intrinsic resolution limit due to instrumental considerations or even to molecular specific attributes (e.g. relaxation properties).
For instance, traditional frequency domain data evaluation methods consisted in the determination of peaks maxima (or minima). A simple example of the performance of such algorithm is shown in the picture below.





Armed just with those peaks detected, confirmation of the chemical structure showed in the picture should be plausible: One peak corresponding to the OCH3 group and then the 3 aromatic protons. Integration and couplings also support this hypothesis. Actually, automatic verification with Mnova under these conditions yields a very class pass.
However, a closer look at this spectrum shows that things are not so simple. First of all, if we consider the multiplet at 8.21, it looks quite evident that it might have a finer structure and not just 4 peaks, but the peak picking algorithm is not able to resolve them.





If instead of a basic traditional peak picking algorithm, a more sophisticated method is used, like GSD, the following result is obtained. Notice that this time, for the sake of clarity, the individual peaks represented as Lorentzian-like lines are displayed in blue.







At first glance, it appears that the number of peaks has now been doubled, a fact which clearly jeopardizes the confirmation of the putative structure. Those extra peaks detected by GSD are not processing artifacts, they are real peaks which were not pinpointed by the traditional peak picking algorithm.

Doubling of the peaks can also be noticed not only in the main peaks but also in the 13C satellites:







Mnova Verify now does not give such a clear pass, indicating that incompatibility between the proposed structure and the spectrum might exist.
Starting with the plain 2D molecule, conformational search using GMMX shows that there are two major conformers with relative populations of 1:0.77.












This is compatible with the relative integrals of each peak in the spectrum, 1:0.72, as depicted in the figure below:





Conclusions:

Reliable data evaluation is a critical feature when trying to assess the degree of compatibility between a chemical structure and its corresponding NMR spectrum. Traditional peak picking methods might miss very relevant spectral components that will not tell the complete story and could result in misleading conclusions, as shown in this simple example.
GSD yields superior results, although it should also be mentioned that very often, many potential NMR transitions will not be resolved, even using the most advanced processing algorithm.


More...

Source: NMR-analysis blog