# What causes the deshielding in carbon - carbon bonds?

I thought of this question while reading through a chapter on Carbon 13 NMR. It's easy to see that in a bond with two different atoms, such as in a bond involving $\mathrm C$ and $\mathrm O$, that the difference in electronegativity causes a deshielding effect. It's also easy to imagine that the deshielding effect is proportional to the number of electrons involved in the bond. These two facts are consistent with the Carbon 13 NMR spectra for single and double bonds between oxygen and carbon atoms. A single bond between carbon and oxygen generally causes a chemical shift of $50 - 100$ p.p.m. A double bond between these two atoms is close to the $150 - 200$ p.p.m range.

However, there is a discrepancy with this approach when we look at carbon carbon bonds. Firstly, there is no difference in electronegativity between these two atoms, so we would not expect a significant chemical shift, however there is a significant observed chemical shift in the NMR spectra. Let's ignore this for a moment. We have for atoms involved in carbon carbon single bonds a chemical shift in the $0 - 50$ p.p.m. range. In carbon carbon double bonds, a shift of $100 - 150$ p.p.m. is observed. However, we have for carbon carbon triple bonds, shifts of $50 - 100$ p.p.m. observed.

Clearly, electronegativity is not the only factor affecting deshielding. Also, we cannot rely on the assumption that the chemical shift is directly proportional to the amount of bonds between the atoms. So, my question is this: What causes the deshielding between atoms involved in bonding? How can we explain the observed spectra theoretically? Our current theory is clearly inadequate.

• Carbon-13 shifts are determined primarily by paramagnetic shielding $\sigma_\mathrm{p}$ - whereas electronegativity affects diamagnetic shielding $\sigma_\mathrm{d}$ (and is a relatively small effect for C-13). $\sigma_\mathrm{p}$ is inversely proportional to $\Delta E$, an averaged excitation energy, and hence the shifts can be related to the energies of $n \to \pi^*, \sigma \to \pi^*, \cdots$ transitions. (Note that $\pi \to \pi^*$ transitions don't count due to symmetry considerations.) More information - Gunther NMR Spectroscopy 3rd ed.; pp 409 onwards – orthocresol Mar 4 '17 at 0:48