Homonuclear Hartmann-Hahn effects in broadband decoupling

Abstract

The technique of noise decoupling is gradually being replaced by new deterministic pulse sequences (I-4) for broadband heteronuclear decoupling, allowing very high resolution with minimal radiofrequency heating. For example, the WALTZ-16 sequence (3, 4) can be used to achieve decoupled carbon-l 3 linewidths of the order of 0.0 1 Hz provided that there is good control of the probe temperature to avoid broadening by temperature-dependent chemical shifts (5). WALTZ-16 has also been used to decouple carbon13 from protons (6), a particularly demanding application in view of the large chemical-shift range of the carbon13 nucleus. In applications where decoupling efficiency is of prime importance, for instance, in vivo carbon13 spectroscopy, where even kilowatt transmitters may only produce & fields of the order of I kHz at the region of interest, the GARP decoupling sequence (7) may prove to be useful; it has an effective decoupling bandwidth ABIB = k2.4. Although few applications require the very high resolution mentioned above, it is disconcerting to see two recent reports (8,5) of instances where WALTZ16 decoupling falls short of its expected performance with regard to the residual carbon-l 3 linewidth. For instance, certain ethyl groups (8) give unexpectedly large residual linewidths, even when other carbon13 resonances in the same molecule decouple as expected. As an example, Fig. 1 shows the carbon-13 spectrum of a mixture of alcohols; the ethyl resonances, and the methine carbon of isopropanol are particularly broad. This effect is relatively independent of the decoupler offset and becomes more pronounced as the decoupler power is decreased. Is all well with decoupling theory? HOHAHA! We believe that these “anomalous” linewidths are due to magnetization transfer between coupled protons through the homonuclear Hartmann-Hahn “HOHAHA” effect (9, 10). Modern decoupling sequences are based on the use of a composite 180” pulse (R) that provides good I spin inversion over a wide range of resonance offsets. This is combined with its phaseinverted counterpart R in a “magic cycle” R R l? 2 and any remaining imperfections are minimized by an expansion into a “supercycle” (I 1). This prescription assumes that the I-spin offset remains constant throughout the decoupling cycle (12). This will not be the case if there is transfer of I spin magnetization from one chemical site to another, a process which can occur when there is scalar coupling between I spins and magnetization transfer in the rotating reference frame-the homonuclear analogue of the classic Hartmann-Hahn experiment (9). There are well-documented instances of Hartmann-Hahn transfer in liquid-phase experiments (13, 14). The condition for efficient transfer is that the two sites have essentially equal precession frequencies in their respective rotating reference frames. For the heteronuclear case, this normally