An improved sequence for broadband decoupling: WALTZ-16

Abstract

Progressive improvements in broadband decoupling performance have recently been achieved with the pulse sequences known as MLEV-4, MLEV-16, MLEV-64, etc. (Z-5). Applied to carbon-13 spectroscopy, such sequences permit operation with a much lower radiofrequency power. A common feature of these and related experiments (6) is that their effectiveness can be improved by combining different versions of the primitive cycle into extended “supercycles” in which some of the residual pulse imperfections are compensated in a manner reminiscent of the folklore of solid state NMR. The original treatment of these experiments was based on average Hamiltonian theory (7) which, although it provides insight into the mechanism of error compensation, can be rather cumbersome in its application (5). An elegant new theory has recently been proposed (6, 8) which represents the effects of the proton irradiation sequence by means of a train of spin rotation operators, the overall effect at the end of the cycle being calculated by explicit matrix multiplication. The offset dependence of this proton response then determines the residual splitting of the carbon-13 resonance and hence the effectiveness of the decoupling. A particular virtue of this treatment is that it provides a simple mechanism for testing new decoupling sequences by computer simulation, and it acts as a guide to the intuitive approach. The principal criteria for decoupling performance are (a) wide effective proton bandwidth for a given power dissipation, (b) residual splittings of carbon-13 small compared with the line width, (c) insensitivity to pulse length error or