The molecule $\ce{HC(=NH)OH}$ has been shown to derive from hydrogen cyanide and water (compound 7). Is it a stable compound?

enter image description here

source: Das, T.; Ghule, S.; Vanka, K. Insights Into the Origin of Life: Did It Begin from $\ce{HCN}$ and $\ce{H_2O}$? ACS Cent. Sci. 2019, 5, 1532–1540. https://doi.org/10.1021/acscentsci.9b00520 (open access).

  • $\begingroup$ It is not that difficult to add an illustration to the question to provide some context of the question (especially if the source in question is an open access publication). $\endgroup$
    – Buttonwood
    May 11, 2023 at 10:21
  • $\begingroup$ @GeoffHutchison The initially used SMILES string C(=N)O is understood by obabel 3.1.1, ChemDrawJS (which assigns to this structure N=CO), or molcalc.org. However the form now used by @Mihoron of HC(=NH)OH is not understood by said obabel (contrasting to the other programs). $\endgroup$
    – Buttonwood
    May 11, 2023 at 19:42

2 Answers 2


You are considering the following molecule.

formimidic acid structure and name

As stated in the referenced study:

Values have been calculated at the B3LYP-D3/TZVP+COSMO(ε = 80.0)//RI-CC2/TZVP+COSMO(ε = 80.0) and the B3LYP-D3/TZVP+ COSMO(ε = 80.0)//RIMP2/TZVP+COSMO(ε = 80.0) (values shown in parentheses) levels of theory in kcal/mol.

Regarding the accuracy of calculations, Becke, A.D.$^1$ (2014) says, "with dispersion interactions in place, the future looks bright," albeit with warnings "we cannot, quite yet forget the basics," "DFT heaven is probably unattainable," and "DFAs, local or nonlocal, will never be exact," and ends with the note "The next fifty (forty-three) years will be as interesting as the first."

However, the study (referred in the question) suggests that many compounds in the suggested mechanism have been observed experimentally:

The use of the nanoreactor produces as an output many different pathways to new species from the starting reactants. In most of the pathways, formaldehyde 3, urea 26, formaldimine 2, and glycolonitrile 4 were seen to be formed as intermediates. This suggests that these species were the key intermediates en route to the formation of the target molecules, as has also been noted by experimentalists.

Quick (and even less accurate) calculations using MolCalc shows $\Delta_f H = \mathrm{-122.76 kJ mol^{-1}}$, a negative value, suggesting it is a (somewhat) stable molecule.

$$ \ce{C(s) + \dfrac{3}{2}H2(g) + \dfrac{1}{2}N_2(g) \xrightarrow{\Delta H \approx -122.76 \mathrm{kJ\ mol^{-1}}} HOCH=NH} $$

However, I couldn't find any results for "synthesis of formimidic acid" and thus it could be a (somewhat unstable) intermediate. As has been pointed out in other answers, the tautomer (formamide) is more stable*.

formimidic acid formamide equilibrium favoring formamide formation

*Approximate heats of formation calculated using MolCalc.


  1. Axel D. Becke; Perspective: Fifty years of density-functional theory in chemical physics. J. Chem. Phys. 14 May 2014; 140 (18): 18A301. 10.1063/1.4869598
  • 2
    $\begingroup$ I wouldn't call anything with B3LYP accurate. DFT isn't accurate in general. But with proper diligence and calibration it provides quite some insight. That being said, local minima on very approximate levels tend to also be local minima on higher levels. So the basic gist still applies. $\endgroup$ May 11, 2023 at 21:41
  • 1
    $\begingroup$ In my opinion, B3LYP is (among) the worst offender(s). At least they include some dispersion and some solvent effects. To my knowledge, there still isn't a generally applicable solvent model. To sum it up: if it is right, it's basically happy coincidence of error cancellation. $\endgroup$ May 12, 2023 at 6:44
  • 1
    $\begingroup$ I have partially read the paper. It uses a terrible level of theory and old software, too. It doesn't correctly cite all the methods it is actually using. Maybe they produce a full list in the supporting information. I wouldn't be very surprised if you obtain something with a better suited method of calculation. $\endgroup$ May 12, 2023 at 7:03
  • 1
    $\begingroup$ I was a bit imprecise. DFT is of course exact. Density functionals, or density functional approximations (DFA) are not exact. There is a wonderful article by Becke about this. The error in the calculations is problematic and it is within their design. A method that isn't accurate will never become accurate. DFAs perform well, or well enough, but that has nothing to do with accuracy. And without calibration, you won't even realise whether a DFA produces garbage. Accurate methods are expensive and I don't see computational chemistry migrating towards those anytime soon. $\endgroup$ May 12, 2023 at 7:09
  • 1
    $\begingroup$ Another point of order: I have down voted this answer. This is due to the fact, that the following statement is incorrect: These are (very) accurate calculations and shouldn't be underestimated. I will gladly reverse this, when a more appropriate wording is used. $\endgroup$ May 12, 2023 at 7:14

Compound 7 is the imidic acid of formic acid (hence formimidic acid). However as indicated by $\Delta{}G$, its tautomer formamide 8 is thermodynamically favoured. Some activation ($\Delta{}G^\#$) has to be invested, though.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.