Chemistry - Can non-deuterated solvents be used for 13C-NMR (and in fact other nuclei)?

Solution 1:

If using a deuterium lock, then normally you would require, as minimum, about 10% deuterated solvent. Without any deuterated solvent, it is not possible to obtain a spectrometer lock (unless using a 19F lock channel, but that is unlikely, and is a hardware requirement).

It is, however, perfectly feasible to acquire spectral data for all nuclei without any deuterated solvent, and hence, without a spectrometer lock. When running without a lock, you need to turn the lock sweep off, to hold the spectrometer field constant. This does not provide the same level of field stability as a lock, but is more than adequate for short-medium 1D experiments that do not require very high resolution. Using gradient shimming, it is also possible to shim on non-deuterated solvents and achieve the same lineshape specifications. Running long experiments (such as an overnight 13C{1H} experiment), without a lock will give you linewidths that are slightly broader than usual; the consequences of this are that (i) your resolution is reduced (rarely a problem with 13C), and (ii) your signal-to-noise will be reduced (and therefore longer experiment times may be required...leading to further broadening and longer experiment times...)

Not claiming to be the first to report about this technique, the Hoye group published two papers in Organic Letters (2004OL953 and 2004OL2567) with examples of application of no-D $\ce{^1H}$-NMR spectroscopy, covering sample preparation and data acquisition, too. Especially the second one, outlining how to determine the concentration of LDA or Grignard reagents by addition of 1,5-cyclooctadiene as internal standard may be a reliable and faster alternative to the titrations once you are used shim manually for samples of this type. But speak and obtain clearance in advance with your NMR manager before running your no-D sample; some interfaces to the spectrometer tend to be less puzzled to record NMR with a lock after your sample (e.g., elder Varian), than the more modern (e.g., Topspin Bruker) to prevent unnecessary delays in the queue of samples following yours.

Solution 2:

It is indeed possible and you will indeed have problem with locking. My solution is to use sealed capillaries with a deuterated solvent inside. This procedure allows you to lock without having the solvent in your mixture.