Natural-abundance C nuclear magnetic resonance spectra of medium-molecular-weight organic compounds.

The difficulties in detection of C13 magnetic resonance signals in organic compounds which arise from the low natural abundance and the small magnetic moment of C'3 are well known.' For complex substances of biochemical interest, further reduction in signal-to-noise ratio occurs because of extensive carbonproton spin-spin interactions. Considerable relief from these difficulties is possible in principle with complete proton decoupling that enhances the signals by removing the proton splittings and contributing a favorable Overhauser effect. 2 The problem with proton decoupling as ordinarily practiced is the necessity for extensive variations in the decoupling frequency to locate the proper values for sharp decoupling of all of the carbon resonances. If the C'3 signals are already so weak as to require enhancement by time averaging with repetitive scans,3 excessively long scanning times may be necessary to obtain satisfactory spectra. Some idea of what is involved, in a rather favorable case, may be seen from Figure 1, which shows a series of C13 spectra of a mixture of aand 0-pinenes run with four different proton decoupling frequencies with the use of the DFS spectrometer previously described.3' 4 Although the power input at the decoupling frequencies chosen was close to the maximum that could be supplied to the oscillator coil, no one frequency gave sharp proton decouplings for all the carbons. Furthermore, if the intensities of the resonances are to be maximized, quite fine adjustments of the decoupling frequencies are required; this may well lead to prohibitive scanning times. Much more general proton decoupling can be achieved with noise modulation of an average proton frequency with a random-noise generator,' and this technique has now been found to be highly useful for C13 spectra. The hookup is shown in Figure 2, and sample noise-decoupled spectra for the same pinene sample used to obtain Figure 1 are given in Figure 3. It will be seen that spectral quality is relatively independent of the center-band frequency. Other spectra which illustrate the value of noise decoupling for C13 spectra of medium-molecular-weight organic compounds have been obtained. One example is cholesterol C27H460 which has such a variety of carbons and so many proton couplings as to provide little hope for detailed structural information by ordinary C13 spectroscopy. Even with 1000 scans the undecoupled spectrum shows only few easily discernible peaks (Fig. 4). With noise decoupling, a dramatic change is evident (Fig. 5); with narrower sweeps, fully 26 of the 27 carbon atoms of cholesterol have large enough chemical-shift differences to be easily distinguishable. The assignment of these resonances to specific carbons will not be an easy task but is straightforward in principle at least Two further