7. Comments

The sequence is in part a standard gradient-selected HMBC with a low-pass filter; thus the pulses p1, p4, p5, p7 and p8 have exactly the identical meaning as already described in Experiments 10.16 and 12.5; delay d2 is the low-pass filter delay.

The sequence employs a constant time period with moving 180° pulses, and this constant time period starts after pS and ends with p8. It consists of three parts: the first part (2-d3), which is decremented during the experiment, the second part (2-d4), which is incremented and the third part (t,), both of which are incremented. The total duration of this constant time period should be set to about 500 ms; much more will be problematic due to relaxation, much less will not give very good resolution in F}.

Homonuclear proton spin couplings will not contribute to a modulation in F), since the total time for their evolution is fixed and the 180° pulses p2, p3 and p4 will always lead to a complete refocusing of the proton chemical shift. On the other hand the C,H spin couplings are decoupled in the first part of the constant time period by pulse p2, but will be effective in the second part, since two 180° pulses p3 and pé6 are applied at the same time. During ¢,, the C,H spin couplings are again decoupled by pulse p4. The incrementation steps during the first two periods are m/[2-sw1], where m is a scaling factor usually set to 30; thus the modulation by the C,H spin coupling during the sec- ond period will be multiplied by this factor and this has to be known when recalculating the spin coupling constants from the spectral splittings in F). Note that insufficiently suppressed 'J(C,H) correlations will also be scaled in F; by the factor m and this therefore leads to a splitting of several kHz.

The method is considerably less sensitive than the gradient-selected HETLOC ex- periment, and a compromise has to be found between the average length of the active HMBC period (2-d4), the total length of the constant time period, the number of f, in- crements, and the chosen scaling factor m for the achievable resolution. As an advan- tage and in contrast to HETLOC, it also gives spin coupling constants to quaternary carbon nuclei as demonstrated in the figure.

8. Own Observations