By what mechanism might the decarboxylation of sodium benzoate in the presence of heat and sodium lye occur?

Question

I am trying to get into chemistry and I saw a video by Nile Red in which powdered sodium benzoate and sodium hydroxide are mixed and heated and a reaction proceeds as follows:

$\ce{NaOH (s) + C6H5COONa (s) ->[200^\circ C] C6H6 (g) + Na2CO3}$

I don't understand how the bond between the benzene ring and the carboxyl group is being broken because a benzene anion seems very unstable, and I also don't get how the hydroxide is being ripped apart considering the OH bond is very strong. Can someone explain a plausible mechanism for how this reaction might take place?

Answer 1

I have viewed the presumed Nile Red video to which the OP refers on the thermal decomposition of sodium benzoate in the presence of sodium hydroxide. An intimate, powdered mixture of sodium benzoate (0.69 mole) and sodium hydroxide (1.5 moles) contained in a metal paint can and fitted with a distillation apparatus was heated over a camp burner.^[1] Although the OP indicates that the temperature of the reaction was 200^oC, there is no evidence that the temperature of the reaction was determined. At these elevated temperatures, it is clear that the entropy (TΔS) outweighs enthalpy (ΔH).

Two previous Answers to this question offered variations on an ionic fragmentation of the carbon-carbon bond in question of sodium benzoate to afford benzene and carbon dioxide. The prospect of a radical fragmentation may also be considered. One source of hydrogen atoms is surreptitious water with a bond dissociation energy (BDE) of 119 kcal/mol. Alternatively, hydroxide has a BDE of 110 kcal/mol. There is the possibility that, under the vigorous conditions of the reaction, addition of hydroxide to the carbonyl group of the benzoate anion 1 occurs to form adduct 2. Bond dissociation of 2 leads to phenyl radical 4 and radical 3. continued...

The BDE of the O-H bond in 3 is expected to be much weaker than that in the other candidates. The BDE of the C-H bond in formic acid 7, although not formate, is 96 kcal/mol. However, the BDE for the same bond in the radical 8 is only 12 kcal/mol. Presumably, a radical cage effect would be operable prior to diffusion.

BDE's are taken from S. J. Blanksby and G. B. Ellison, Acct. Chem. Res., 2003, 36, 255.

1. My advice, don't try this at home.

Answer 2

An alternative possibility is proposed here, which does not require formation of a free phenyl anion. Instead the phenyl-anion moiety is incorporated into a delocalized bonding in an intermediate state.

In this alternative, the hydroxide ion first adds to the carboxyl carbon to from the orthocarboxylate salt on the left. Ordinarily, we would not expect a carboxylate ion to act as an electrophile, but here the addition of heat energy and the strong basicity of the hydroxide ion may enable such an addition. The orthocarboxylate then undergoes an electronic rearrangement to separate the carbonate portion and leave the decarboxylated hydrocarbon behind.

Answer 3

Decarboxylations tend to follow the mechanistic path I've drawn below. You're right in that this is a bold transformation--partly this is driven by the high heating. Decarboxylations are also driven by entropic and enthalpic forces.

Entropically, you are starting with one compounded (sodium benzoate) and producing two (phenyl carbanion and CO₂ gas). Entropy is frequently described as "disorder" which you can quantify as how many microstates are possible. If you have a loaf of bread and a four-shelf book case, you can put in on the first, second, third or fourth shelf. If you tear the loaf of bread in half, you can for example, put one loaf half in the third shelf and the other loaf in the first shelf—you can see there is many more combinations of options or "microstates". Replace "bread" with molecules and "book shelf" with energy level and you have entropy. Breaking the molecule in two increases entropy, especially because the CO₂ is a gas, which has more entropy than a solid.

Enthalpic means, put simply, how strong the bonds are. While you do raise legitimate enthalpic concerns, this will be offset in part by the strong double bond you form making CO₂ gas (a double bond is stronger than a single bond). If you make more strong bonds than you break, this is enthalpically favoured.

CO₂ becomes NaCO₃H and Na₂CO₃ readily. In fact this happens in your blood and is why people used to recommend breathing into a brown bag—increase the CO₂ concentration which pushes towards NaCO₃H. On that same point, decarboxylations are also an extremely common reaction in the body, albeit not with sodium benzoate.

Finally, the proposed reaction proceeds through phenyl carbanion (all other proposed mechanisms involve the same), though the mechanism may be concerted. This is a bit strange, but I'll just reference one or two previous posts on this species and point out that the aromaticity is not disturbed, as shown in the second figure. Typically you may see phenyl carbanions stabilised as lithiated species (i.e., using extremely strong organolithium bases to generate), which I think points to why such extreme heating was required by NileRed. Also, knowing his channel is sort of "garage chemistry", heating is much accessible than unusual reagents.

Answer 4

I know this is probably a bit late to mention but ill mention it anyways. I preformed this exact reaction to maybe shed a little light on the mechanism by looking at the by-products formed in the reaction.

I mustn't lie, this method of benzene preparation is actually really cool. Reagents are easily accessible and cheap, reaction continuous very nicely, and there isn't anything too dangerous + its very self driven, and requires practically no involvement, just throw it in a can, heat it until vapour stops coming over and your done.

Anyways, there seems to be a very orange liquid that forms during the reaction, that necessitates further purification, and the benzene directly out of the decarboxylation doesn't have the best purity. I wasn't able to characterise it, all i know is that its fluorescent (possible many conjugations), as well as there isnt too much of it produced. 1 other by-products I actually managed to characterise was Biphenyl

This was mainly characterised through FT-IR,

There was a very close match to Biphenyl, the obtained product was a white waxy solid. Once again, there wasn't very much of it. Now I don't want to attempt to make any prediction on theoretical mechanisms, though I feel the formation of a phenyl ion is very likely, as it would be unlikely we would get the formation of biphenyl in other major way.

Do with this information what you wish.