In the mass spectrum of niacin (3-pyridine-carboxylic acid), what ion corresponds to the peak at m/z = 105?

Question

The mass spectrum can be found here.

I am interested to know what ion this is and why the peak at $m/z = 105$ is higher than the peak at $m/z = 106$ (which I believe corresponds to $\ce{C5H4NCO+}$).

Answer 1

From the general interpretation of electron ionization mass spectra, it would seem logical to report the $m/z$ 105 ion as a loss of $\ce{H_{2}O}$, thus $\ce{[C_{6}H_{3}NO]^{\bullet +}}$. As correctly suggested in the question, the peak at $m/z$ 106 is likely arising from a loss of $\ce{OH^{\bullet}}$ through an $\alpha$-cleavage of the $\ce{C-O}$ bond.

So the issue is: why would niacin (or nicotinic acid) undergo this $\ce{H_{2}O}$ loss? It is not observed for other similar molecules such as benzoic acid, which starts by a $\ce{OH^{\bullet}}$ loss. I did find two relevant publications in the literature on this point. Neeter and Nibbering^[1] performed $\ce{D}$ labelling experiments, showing that the transferred hydrogen atom originates mostly from the $\alpha$ position (ortho from both the carboxylic acid and the nitrogen atom). Opitz^[2] measured the appearance energy (AE) for $m/z$ 105 at 10.94 eV, more than 0.6 eV lower than the AE for $m/z$ 106 at 11.58 eV. He also suggests, based on metastable dissociation data showing loss of CO to $m/z$ 77, two possible structures for the $m/z$ 105 ion. So to answer the question: (a) formation of the $m/z$ 105 ion is thermodynamically favored compared to the formation of the $m/z$ 106 ion.

Based on this (scarce) literature evidence, I would suggest, based on labelling experiments, that some proton transfer occurs between the $\alpha$ cycle hydrogen and the hydroxyl group. This exchange can either move back and forth, or lead to a loss of water through formation of a cyclic structure.

---

References

1. Neeter, R.; Nibbering, N. M. M. Mass spectrometry of pyridine compounds. II: Hydrogen exchange in the molecular ions of isonicotinic and nicotinic acid as a function of internal energy. Org. Mass Spectrom. 1971, 5 (6), 735–742. DOI: 10.1002/oms.1210050612.

2. Opitz, J. Electron-impact ionization of benzoic acid, nicotinic acid and their n-butyl esters: an approach to regioselective proton affinities derived from ionization and appearance energy data. Int. J. Mass Spectrom. 2007, 265 (1), 1–14. DOI: 10.1016/j.ijms.2007.04.014.

Answer 2

I did quick simulation with MMass (open source, available on Windows, MacOS, Linux) for the isotopic distribution in this region (Tools > Mass Calculator), and it looks like there are two $[\ce{M}-\ce{1e}]^+$ fragments:

$$ \begin{array}{rlr} \hline \text{avg.}~m/z & \text{Fragment} & \text{a.i.}\\ \hline 105.09 & \ce{C6H3NO+} & \approx45\%\\ 106.10 & \ce{C6H4NO+} & \approx21\%\\ \hline \end{array} $$

Addition of the signals from both fragments (blue and green colors) would give nearly ideal peaks distribution reflecting experimental data (red color).

Technical note: MMass can import ASCII data, but NIST provides EI-MS spectra in unsupported JCAMP-DX format. To convert JDX to TXT, I used JDXview (free, Windows only).