In order to obtain high-resolution spectra of dilute spins in powdered or amorphous solids, magic-angle spinning (MAS) has become almost indispensable ( 1-4). Spinning a sample with a frequency of vR results in a set of rotational echoes, spaced apart by a time interval of 1 /Q after the original free induction decay signal. Fourier transformation results in a characteristic pattern from which the centerband gives the isotropic frequency shift. The sideband intensities can be used to obtain the principal values of the chemical shielding tensor. However, for some applications it may be useful to eliminate the rotational sidebands, e.g., for analytical purposes, where the isotropic chemical shifts are used to distinguish nuclei with different chemical environments. In order to separate both the isotropic and the anisotropic shifts of powders containing chemically distinctive nuclei, two-dimensional techniques must be employed, for which it is also necessary at some stage to suppress the rotational sidebands (5). A useful method for sideband suppression has been devised by Dixon (6, 7) and has been named TOSS (total suppression of sidebands). The original sequence consists of four rotor-synchronized r pulses applied within one rotor period. The timing of the pulses is such that it leads to an in-phase alignment of the centerbands of each crystallite, while the sideband phases are such that they cancel when averaged over a powder (8, 9). Crucial for complete suppression of the sidebands are exact timing of the pulses, resonance offset, and pulse lengths. The timing of the pulses is prescribed by vR. The time interval between pulses is measured from the center of the pulses, implying that the x pulses must be kept as short as possible for the system to evolve freely between pulses. Pulse length errors and offset effects lead to incomplete cancellation of sidebands, especially when the sidebands are very narrow. This is the case in systems exhibiting a very small shift dispersion. As is well known, composite pulses can be used to compensate for these effects ( 10, 11). This, however, reduces the free evolution time of the system. Thus, the use of composite pulses in the TOSS experiment sets a limit to the spinning speed. Furthermore, the duration of the composite pulses must be kept as short as possible. It is the purpose of this Note to demonstrate that the application of composite pulses together with phase cycling in the TOSS sequence drastically reduces the artifacts of applying this sequence. A system with very narrow sidebands suitable for demonstrating
Read full abstract