Haloform review reviewed

28 March 2025 - Research update 00004

Noticed the "200 Years of The Haloform Reaction" article by Albert C. Rowett et al. recently. The review (DOI) stands out because it is open access (!) and covers the reaction from start to finish(!) instead of the usual "recent developments in topic X" autopilot nonsense. Good excuse as any to query the Chemical Reaction Database to detect if any haloform reactions are out there and if the results are consistent with this review.

Starting with the most common incarnation, the conversion of a methyl ketone to an carboxylic acid via a perhalogenated intermediate, the original inventor of the reaction in 1822, Georges–Simon Serullas, may be pleased to learn that his work still finds it way into modern organic reaction databases but the number is modest, just 88 in total which on a total of 1.3 million reaction is just 0.006%. For comparison the number of direct carbonylations with carbon dioxide is 190 and the number of nitrile hydrolysis reactions to the same carboxylic acid is 1675.
To make matters worse, a "reverse haloform", the reaction of a carboxylic acid with methyllithium back to the methylketone outperforms the haloform with 102 entries. As for the other haloform flavors, the conversion to an ester has just 8 entries, the amide variation is absent and the dataset did not yield any perhalogenated intermediates. As to the origin of the methylcarbonyl group the dataset is too small to allow detailed analysis (aside from computational challenges) but a quick scan reveals a number of Friedel-Crafts acylations with acetyl chloride in the ancestral lines. The popularity of the haloform reaction is impossible to ferret out but if atom economy is still a thing in 2025 it certainly does not have any.
Then to the haloform reaction specifics. The review mentions that the reaction would love to just have water as a solvent but methyl ketones do not sit well with this solvent. The collective mind of the chemists that have contributed to the database seem to have settled this issue: without exception the solvent system is fortified with dioxane. As halogen donors we see either Br2/NaOH or the bromates or chlorates (NaBrO3, NaClO3, CaClO3) with or without additional base. As mentioned in the review with chlorine the reaction is slower and chlorine reactions appear to run at higher temperatures. In one entry the a dichloromethane/TFA system was employed. The halogen excess is a predictable 3 to 4 equivalents as per review.
Of course these data are based on a 1.37M database which is puny compared to the fabled databases with magical names such as CAS or REAXYS which must be 200 times larger. How would the same analysis on these systems pan out for the haloform reaction. Should not any review have a section of this type? The reaction distribution as it is? Since the early beginnings of organic chemistry? A challenge? A new challenge!