1,3-diphenyl-prop-2-en-1-one, but mainly known as chalcone, is a compound with very rich chemistry. Due to the presence of two functional groups, C=O and C=C double bonds, it can participate in numerous transformations. The reduction potential of chalcone and related compounds, α,β-unsaturated ketones, is -1…-1.75 V vs SCE that is in the working range of common photocatalysts. Therefore, they are convenient substrates in photoredox catalysis.

Earlier we used chalcone and other α,β-unsaturated ketones in synthesis of cyclopentanoles and N-fused pyrroles.

While working on the synthesis of N-fused pyrroles, we tested different solvents – a somewhat routine procedure in catalysis. The outcome of these tests was also predictable, we obtained the expected product in slightly higher or lower yields, until we employed chloroform.

In case of chloroform, we observed formation of an unknown compound. Therefore, we concluded that chloroform is not innocent solvent, but a suspect engaged in the formation of the unknown product!

The first and the most important task was to solve the structure of the product formed by mixing chalcon, triethanolamine, carbon nitride photocatalyst in chloroform and shining light on the mixture. I must say it was not a trivial task.

First we collected ‘fingerprints’, a mass spectrum, and checked the data base – no reasonable matches were found.

The ‘composite portrait’ of the product, the ¹³C NMR spectrum, revealed one additional carbon atom with the chemical shift at 76 ppm compared to the spectrum of chalcone. While ¹H NMR spectrum revealed additional proton around 6.3 ppm. It gave us a hint that CH-group was transferred from the chloroform to chalcone.

To check this hypothesis, we used a ‘bait’, CDCl­₃ instead of CHCl₃. As a result, we observed a raise of triplet in ¹³C NMR. Our guess was confirmed by this experiment. But still there was mismatch between the expected mass of the molecular ion and that observed in the mass spectrum.

In the modern research facility equipped with highly sensitive mass- and NMR spectrometers, we sometimes forget about old, somewhat laborious, but nevertheless reliable techniques. In our story, elemental analysis for chlorine played crucial role.

Indeed, elemental analysis showed perfect agreement between the chlorine content determined experimentally and the theoretical chlorine content assuming addition of H-CHCl₂ to chalcone. In other words CH₂Cl₂ molecule. But which positions in chalcone were involved in the reaction?

With that piece of evidence in hand, we ‘interrogated’ the product again. With the multiplet analysis of the ¹H NMR spectrum, this time the product ‘fessed up’! The CHCl₂ moiety was attached to the beta-carbon atom, while H to the alpha-carbon atom of the chalcone!

It looked like the riddle was solved and we were about to close the case, but hunch suggested us that this is not all. Like in every good detective, there is always a twist. We thought that chalcone might be not the only compound working in this reaction, but it might have associates.

Therefore, we synthesized a series of substituted chalcones with F, CH₃, OMe, etc. and as later, during the peer-review, was suggested by the reviewer, also with CO₂Me, Br, Cl. It turned out that these chalcones also undergo the same reaction path of CHCl­₂ addition! Overall, a group of 15 chalcones has been exposed to participate in the photocatalytic reaction. The products, γ,γ-dichloroketones, were isolated and ‘interrogated’ by NMR and GC-MS.

There was one common detail in their stories. None of them gave the molecular ion of the γ,γ-dichloroketone, instead we always registered 2,4-disubstituted furans – the products of dichloromethyl group electrophilic attack at oxygen carbonyl atom accompanied by elimination of 2 HCl molecules. It gave us a hunch that CHCl₂-group in γ,γ-dichloroketones is highly electrophilic and may react with different nucleophiles, like amines and hydroxide ion.

The hypothesis was confirmed and we prepared few substituted furans and pyrroles.

Now the case can be closed. Or maybe there are other pathways of chalcone reactivity in carbon nitride photocatalysis?...

Read full story in Nature Communications.

Follow our ‘investigators bureau’ on Twitter!