CALIBRATION OF STUD DECOUPLING TO ACHIEVE SELECTED SIDE-BAND AMPLITUDES

Experimental calibration of the amplitude of sidebands resulting from STUD decoupling over a range of decoupled bandwidths from 2 to 120 kHz shows that sideband amplitude depends on the length of each single se ch/tanh pulse, T-p, and the bandwidth divided by the square of the max imum RF amplitude during the pulse, bwdth\(RF(max))(2). Plots of sideb and amplitude versus bwdth/(RF(max))(2) are independent of bandwidth a nd so provide convenient calibration curves enabling choice of T-p. Su ch calibration curves will vary modestly between NMR probes because of the differing homogeneity of the RF field produced by the probe acros s the sample, The dependence on bwdth/(RF(max))(2) arises from the inv ersion efficiency of a single sech/tanh pulse, This dependence corresp onds to the principle that the optimal RF power required for STUD deco upling is linearly proportional to the chosen decoupled bandwidth, Thi s behavior, which is common to adiabatic decoupling methods, is differ ent from composite-pulse decoupling schemes where chosen bandwidths an d corresponding optimal RF power settings are more limited, and effect ive bandwidths are proportional to the applied RF amplitude rather tha n RF power, The STUD method achieves the widest effective decoupled ba ndwidths of any broadband technique to date when decoupling schemes ar e compared at the same average power. This is particularly relevant to applications where sample heating must be minimized. Additional crite ria for good decoupling performance are considered in further detail. The accurate measurement of sideband amplitudes is shown to provide a convenient means for evaluating the efficiency of different decoupling schemes, Adiabatic decoupling methods are especially interesting sinc e, in contrast to composite-pulse methods, the RF amplitude and pulse length can be varied separately, providing additional opportunity for comparison with theoretical models of decoupling. In particular, excep tions to a frequently cited condition for minimizing sideband intensit y, T(c)J much less than 1, where T-c is the cycle time to return the i rradiated spins to their initial orientation, are discussed. (C) 1996 Academic Press, Inc.