i-HMBC : Identifying two bond correlations

The HMBC experiment provides ¹H-¹³C correlations over multiple bonds and is essential for assigning small molecules. The main drawback to the HMBC is that it does not identify the length of the correlation and in some cases may be ambiguous. A recent publication describes a method, called i-HMBC, for distinguishing two bond HMBC correlations from longer ones.

The i-HMBC is essentially a standard HMBC experiment recorded with vastly increased resolution in the ¹H dimension. The increased resolution allows small differences in the chemical shifts of peaks to be measured. Wang et al report that peaks from two bond correlations are shifted upfield by 0.3-1.6 ppb relative to three bond or longer correlations. 

The Facility's standard HMBC parameters acquire 4096 points over 12 ppm at 600 MHz. During processing zero filling is not used for this dimension, so the processed data contains 2048 points in the ¹H dimension with a digital resolution of 3.5 Hz/point. For the i-HMBC the number of acquired points was increased to 32768 and zero filling was applied to give 65536 points in the ¹H dimension and a digital resolution of 0.11 Hz/point. This increased number of points does not greatly increase the experiment time as the relaxation delay can be reduced to keep the time taken for a single scan the same as typically used. It does, however, produce a much larger processed data matrix and this can be slow to work with.

With this greatly increased resolution, distortions in the peak shape become much more apparent and make it difficult to measure chemical shifts. Wang et al report that distortions can be reduced by turning the lock off and on during the pulse sequence and by reducing the lock power. I modified the pulse sequence to turn the lock off and on at the appropriate times and found this improved the line shape. I haven't tried reducing the lock power yet.

The test sample was a 100 mg/ml solution of cholesteryl acetate in chloroform-d. The structure and the atom numbering is shown below.

The left panels in the figures below show expansions of selected peaks from the 2D i-HMBC. The vertical line indicates the chemical shift of the ¹H involved in these peaks. Close inspection of the peaks shows that though most line up well, some are shifted upfield (to the right). For H18, the C13 peak is offset. For H21 the C20 peak is shifted. The shifted peaks correspond to the two bond correlations. Extracting 1D slices through these peaks allows the shifts to perhaps be seen more clearly, and to be measured. In both cases the two-bond correlations are shifted upfield by about 0.5ppb.

These two examples were the clearest in the i-HMBC I recorded. Other sets of correlations involved more couplings and more complex peak shapes making measuring the chemical shift differences difficult. Perhaps reducing the lock power during acquisition would have helped. Happily, the i-HMBC gives better and simpler lineshapes in proton deficient systems, which is perhaps where it is most needed. Standard parameters for the i-HMBC are available at the Facility.

References

Unequivocal identification of two-bond heteronuclear correlations in natural products at nanomole scale by i-HMBC
Yunyi Wang, Aili Fan, Ryan D Cohen, Guilherme Dal Poggetto, Zheng Huang, Haifeng Yang, Gary E Martin, Edward C Sherer, Mikhail Reibarkh, Xiao Wang
Nat Commun 2023 Apr 3;14(1):1842.