Collecting multi-dimensional spectra requires recording chemical shift information via the use of an incremented delay. Since the length of the delay increases over the course of the experiment the signal is recorded at different times for each increment. This allows coupling in the indirect dimension to evolve, which broadens the peaks and decreases resolution. The "constant time" method was developed to eliminate this problem.
All multi-dimensional NMR experiments use a delay (Δ) that is incremented to sample the evolving chemical shift one point at a time. This approach means that the time between the first pulse and when the signal is recorded gradually increases as the incremented delay gets longer and longer, as shown on the left side of the figure below. As the length of the pulse sequence increases, the 1H-1H coupling is sampled along with the chemical shift. This leads to the formation of a multiplet in the indirect dimension. In some cases this may not matter much, but where resolution in the indirect dimension is important eliminating this homonuclear coupling can be very helpful.
The constant time protocol prevents the 1H-1H coupling from evolving by adding an extra delay to the pulse sequence. This extra delay is reduced as the incremented delay increases so that the sum of the two delays (τ) remains constant, as shown on the right of the figure above.
In heteronuclear experiments the heteronuclear coupling (1H-13C or 1H-15N) will evolve during the extra delay and complicate the spectra. To remove this coupling a 180o pulse on the heteronucleus (not shown in the figure above) is applied in the middle of the constant time period to refocus the coupling.
Incorporating a constant time period into the HMBC experiment is particularly useful when trying to resolve overlapped peaks. The figure below (taken from Claridge and Perez-Victoria) shows the carbonyl region of several HMBC spectra recorded on a peptide. The top panel shows the non-constant time version, recorded with a 180 ppm 13C sweep width. Here it is difficult to tell how many 13C resonances are present. The middle panel shows a non-constant time HMBC spectrum recorded with a 6 ppm sweep width. Now the peaks are better resolved, but the peaks are twisted because of the 1H-1H coupling. The bottom panel shows the constant time HMBC spectrum. With this spectrum the width of the resonances in the 13C dimension is reduced, making it easy to count the 13C resonances and identify the correlations.
Incorporating a constant time period comes with a few drawbacks. The added length of the pulse sequence will reduce the signal due to relaxation during the extra time, but for slowly relaxing small molecules this is not very significant. The interferograms produced in a constant time experiment show very little decay, which affects the performance of NUS reconstruction algorithms. Thus, NUS does not work well with constant time experiments. The major drawback, however, is that the number of increments that can be used is restricted by the number that will fit within the constant time period, τ. Fortunately, in most cases sufficient increments are possible. The standard HMBC parameters on the Skaggs NMRs use the constant time method.
ReferencesEnhanced 13C resolution in semi-selective HMBC: a band-selective, constant-time HMBC for complex organic structure elucidation by NMR
Tim D. W. Claridge and Ignacio Pérez-Victoria
Org. Biomol. Chem. 2003 3632-3634
Really interesting, thanks for sharing! I wondered if you could elaborate on NUS not working well with constant time experiments? My experience with processing 3D constant time experiments in Topspin is that it works ok, but is it different for 2D spectra?
ReplyDeleteHi Alex, NUS probably works OK with constant time experiments, but not as well as with non-constant time experiments. The NUS reconstruction algorithms assume that the signals they are reconstructing are decaying. If they are not decaying, like in a constant time dimension, then they will likely not work as well.
DeleteThis comment has been removed by the author.
ReplyDelete