JOURNAL OF MAGNETIC RESONANCE 89, I-9 ( 1990) The Effect of Experimental Imperfections on TOSS Spectra DANIEL P. RALEIGH, * ANDREW C. KOLBERT, AND ROBERT G. GRIFFIN Francis Bitter National Magnet Laboratory and Department of Chemistry, Massachusetts Institute of Technology, Cambridge, Massachusetts 02139 Received October 25. 1989 The effect of imperfections on the TOSS experiment, used to suppress rotational side- bands in magic-angle-spinning NMR, is examined. We consider the effects of errors in pulse lengths, phases, and timing as well as resonance offset effects. Errors in the phases of the r pulses have no effect on the efficiency of sideband suppression, and only introduce a phase shift into the spectrum. In contrast, errors in the length of the P pukes or in their timing lead to an attenuated centerband as well as the reintroduction of residual side- bands. Our results show that phase alternation of the * pulses does not necessarily com- pensate for pulse length errors and can, in fact, exacerbate their effects. The use of com- posite x pulses in TOSS is briefly examined and is shown to yield superior performance for spectral lines far from resonance. 0 1990 Academic RUSS. IX. The TOSS experiment ( 1, 2), which eliminates all rotational sidebands leaving a spectrum composed of centerbands alone, has proven a useful method for simplifying magic-angle-spinning (MAS) NMR spectra. Although the sideband intensities in a MAS NMR spectrum can be used to determine the principal values of the chemical shielding tensors (3, 4)) a sideband-free spectrum is often useful as a first step in performing spectral assignments. During the last few years, a number of papers have appeared offering theoretical descriptions of the TOSS experiment and describing variations and modifications to the basic method (5-10) ; however, no general a&y- sis of experimental imperfections in TOSS has been given. In this paper we ana- lyze the effect of pulse width and phase errors, as well as timing errors, on the TOSS experiment. The TOSS sequence, shown in Fig. 1, uses four ?r pulses applied at precise times after the establishment of transverse magnetization to eliminate rotational sidebands. The experiment has been analyzed using the magnetization vector model of MAS (5, 9)) and this analysis shows that the four H pulses and interpulse delays twist the paths of the magnetization vectors of different crystallites, so that the average magnetiza- tion vector of each is aligned along a common axis in the x-y plane of the rotating frame. This alignment causes the centerbands of the subspectra associated with each crystallite to be in (or ?r out of) phase, while introducing phase shifts in the sidebands * Present address: Inorganic Chemistry Laboratory and Laboratory of Molecular Biophysics. Oxford University. Oxford OX 1 3QR, United Kingdom. I 0022-2364190 $3.00 Copyright 0 1990 by Academic Press. Inc. Al1 rights of reproduction I” any form resewed