Manchester NMR Methodology Group

Reference deconvolution [G. A. Morris, J.Magn. Reson. 80, 547 (1988)], in which the experimental lineshape of a reference signal is used to deconvolve a spectrum in order to correct instrumental errors, normally requires a well-resolved singlet as the reference signal, The presence of zeroes in the time-domain signal of a multiplet makes it unsuitable for use in the normal reference deconvolution algorithm, since the zeroes would give rise to discontinuities in the correction function. The problem may be circumvented if the algorithm is adapted to include interpolation over the gaps in the correction function caused by the zeroes, Practical details of the interpolation and of the choice of reference signal are discussed, and experimental results are shown for spectra deconvolved using a doublet signal as reference. (C) 1995 Academic Press, Inc.