Computer-Optimized Decoupling Scheme for Wideband Applications and Low-Level Operation

Traditional modulation methods for broadband proton decoupling (1-4) have now been superseded by modern pulse techniques (5) using sequences of composite 180” pulses (6). Initial progress was achieved with the MLEV decoupling sequences (7-9) and guided by coherent averaging theory (10). This approximate analysis showed that longer and more effective decoupling “supercycles” could be constructed from primitive decoupling sequences according to an inductive “expansion procedure.” Waugh’s exact theory of periodic decoupling (II) further clarified the role of expansion procedures (12) in terms of the net rotation induced by the sequence of pulses. The particular merits of the exact analysis were that it provided a simple criterion for good decoupling, allowed prospective sequences to be evaluated by computer simulation, and isolated the effects of instrumental shortcomings. The WALTZ decoupling sequences (13, 14) resulted from careful attention to the influence of these imperfections, in particular the accuracy of the radiofrequency phase shifts. Applied to carbon13 spectroscopy, WALTZ16 decoupling produces residual splittings of less than 0.1 Hz over a range of decoupler offsets aB = +&. In routine applications splittings as large as 1 Hz can go unnoticed due to coarse digitization in the frequency domain and line broadening produced by sensitivity enhancement functions. However, for more demanding applications, including the investigation of small isotope shifts, of couplings to other nuclei, or of carbon-carbon couplings in natural abundance (15), WALTZ16 has definite advantages. A number of modified broadband decoupling sequences, all broadly based on the WALTZ16 scheme, have been introduced recently (14-19). At 2 Hz carbon-l 3 linewidths, the best of these promises to deliver a 40% improvement in bandwidth over WALTZ-16. In fact this sequence, dubbed PAR-75, gives a bandwidth comparable to an earlier sequence based on composite pulses incorporating short periods of free precession (20). These sequences do not restrict the pulsewidths to multiples of 90” and therefore require more sophisticated pulse programming, but the additional complexity may be justified for certain applications, such as 19F decoupling, where extremely wideband operation is required. The present communication demonstrates that sequences with extremely wide bandwidths may be devised by further relaxing the requirement of programming simplicity. For windowless decoupling sequences a figure of merit (20)