Sensitivity
As a consequence of low receptivity, 13C NMR spectroscopy suffers from complications not encountered in proton NMR spectroscopy. Many measures can be implemented to compensate for the low receptivity of this nucleus. For example, high field magnets with internal bores are capable of accepting larger sample tubes (typically 10 mm in diameter for 13C NMR versus 5 mm for 1H NMR). Relaxation reagents allow more rapid pulsing.[5] A typical relaxation agent is chromium(III) acetylacetonate. For a typical sample, recording a 13C NMR spectrum may require several hours, compared to 15–30 minutes for 1H NMR. The nuclear dipole is weaker, the difference in energy between alpha and beta states is one-quarter that of proton NMR, and the Boltzmann population difference is correspondingly less.[6] One final measure to compensate for low receptivity is isotopic enrichment.
Some NMR probes, called cryoprobes, offer 20x signal enhancement relative to ordinary NMR probes. In cryoprobes, the RF generating and receiving electronics are maintained at ~ 25K using helium gas, which greatly enhances their sensitivity.[7] The trade-off is that cryoprobes are costly.
Coupling modes
Another potential complication results from the presence of large one bond J-coupling constants between carbon and hydrogen (typically from 100 to 250 Hz). While potentially informative, these couplings can complicate the spectra and reduce sensitivity. For these reasons, 13C-NMR spectra are usually recorded with proton NMR decoupling. Couplings between carbons can be ignored due to the low natural abundance of 13C. Hence in contrast to typical proton NMR spectra, which show multiplets for each proton position, carbon NMR spectra show a single peak for each chemically non-equivalent carbon atom.[8]
In further contrast to 1H NMR, the intensities of the signals are often not proportional to the number of equivalent 13C atoms. Instead, signal intensity is strongly influenced by (and proportional to) the number of surrounding spins (typically 1H). Integrations are more quantitative if the delay times are long, i.e. if the delay times greatly exceed relaxation times.
The most common modes of recording 13C spectra are proton-noise decoupling (also known as noise-, proton-, or broadband- decoupling), off-resonance decoupling, and gated decoupling. These modes are meant to address the large J values for 13C - H (110–320 Hz), 13C - C - H (5–60 Hz), and 13C - C - C - H (5–25 Hz) which otherwise make completely proton coupled 13C spectra difficult to interpret.[9]
With proton-noise decoupling, in which most spectra are run, a noise decoupler strongly irradiates the sample with a broad (approximately 1000 Hz) range of radio frequencies covering the range (such as 100 MHz for a 23,486 gauss field) at which protons change their nuclear spin. The rapid changes in proton spin create an effective heteronuclear decoupling, increasing carbon signal strength on account of the nuclear Overhauser effect (NOE) and simplifying the spectrum so that each non-equivalent carbon produces a singlet peak. The relative intensities are unreliable because some carbons have a larger spin-lattice relaxation time and others have weaker NOE enhancement.[9]
In gated decoupling, the noise decoupler is gated on early in the free induction delay but gated off for the pulse delay. This largely prevents NOE enhancement, allowing the strength of individual 13C peaks to be meaningfully compared by integration, at a cost of half to two-thirds of the overall sensitivity.[9]
With off-resonance decoupling, the noise decoupler irradiates the sample at 1000–2000 Hz upfield or 2000–3000 Hz downfield of the proton resonance frequency. This retains couplings between protons immediately adjacent to 13C atoms but most often removes the others, allowing narrow multiplets to be visualized with one extra peak per bound proton (unless bound methylene protons are non-equivalent, in which case a pair of doublets may be observed).[9]