# Is this photochemical [2+2] cycloaddition enantioselective?

How is it deduced that one enantiomer is preferentially formed in the reaction? When I think of the MOs of the reactants, both enantiomers are viable as products.

If there is no source of chirality present in the reaction mixture, then there is no reason for the reaction to be enantioselective. In this case there are two possible sources of chirality: firstly the light, which could be circularly polarised (not a common thing to do), and secondly the catalyst, which can be chiral.

Although the exact catalyst is not specified, the above scheme seems to be one of the earlier reports of transition metal photoredox catalysis, from the lab of Tehshik Yoon.1 The catalyst used is $\ce{[Ru(bpy)3]Cl2}$, the cation of which is chiral (it has point group $D_3$). These cations exist in two enantiomeric forms, labelled Λ and Δ. [Thanks to Zhe for pointing this out.] Upon absorption of visible light, the excited $\ce{[Ru(bpy)3]^2+}$ complex can transfer an electron to the aryl enone, the LUMO of which is lowered by coordination to $\ce{Li+}$. This generates a radical anion which adds into methyl acrylate.

However, there is no report of the product being formed enantioselectively – only diastereoselectively. There is no indication that the authors used a single enantiomer of the catalyst. [Exactly why it is diastereoselective is a slight mystery, to me at least.] Note that this cycloaddition reaction is not concerted, so the MO-based arguments for pericyclic reactions such as the concerted Diels–Alder are not valid here.

In any case, transition metal photoredox chemistry is generally very poor at asymmetric induction, because the role of the catalyst is only to generate transient radical intermediates; the catalyst does not play any significant role in the actual bond formation. It is for this reason that photoredox catalysis is often combined with other modes of catalysis, e.g. organocatalysis, for highly enantioselective reactions to take place.

Apart from this, there are no other reports of this particular cycloaddition in the literature. All in all, I think it is safe to say that the 92% ee is not real.

Reference

1. Du, J.; Yoon, T. P. Crossed Intermolecular [2+2] Cycloadditions of Acyclic Enones via Visible Light Photocatalysis. J. Am. Chem. Soc. 2009, 131 (41), 14604–14605. DOI:10.1021/ja903732v.