Caged water

When fullerene chemistry took off in the 1990s, one enticing use of the carbon-cage molecules was as a protective wrapping for trapped molecules, isolating them from their environment. This, it was suggested, might offer a kind of matrix-isolation technique for studying sensitive chemical species under ambient conditions — a notion explored previously using other cage complexes¹. Trapped metal atoms — the first class of endohedral fullerene compounds reported — might be used as contrast agents for magnetic resonance imaging, or perhaps even as qubits for quantum computing^2,3. If encapsulation could be made reversible, maybe fullerenes could be used to chaperone drug molecules to their target cells.

Some of these possibilities continue to show promise; no doubt others will turn out to have been over-optimistic. The enterprise of trapping molecules to study in isolation has now been furthered significantly by Kurotobi and Murata, who have used inventive chemistry to open up a hole in C₆₀, insert a single water molecule, and then seal up the cage again⁴. They report a crystal structure of H₂O@C₆₀ (the @ sign denoting encapsulation) complexed to a nickel porphyrin, with the lone H₂O localized in the cage centre.

Kurotobi and Murata suggest that H₂O@C₆₀ might offer a rare opportunity to study the water molecule in a situation where it is not coordinated to other species via either hydrogen bonds or lone-pair dative bonding. Well, perhaps. The nominally hydrophobic interior of the fullerene cage might certainly lead one to imagine that the water molecule behaves almost as if in vacuo. And the ultraviolet-visible spectrum of the complex is almost identical to that of empty C₆₀, suggesting almost no interaction between the H₂O and the carbon cage, while the ¹³C NMR spectrum shows all the carbons to be identical, indicating that the water molecule rotates rapidly on the NMR timescale. Yet it would be surprising if indeed there was no interaction between the hydrogen atoms and the aromatic rings of the cage, as hydrogen bonding to the π electrons of benzene rings has been reported for other encapsulated water molecules, both in synthetic supramolecular complexes⁵ and in a hydrophobic protein cavity⁶.

It's in the latter regard that H₂O@C₆₀ might prove to be a useful model compound. Several proteins have hydrophobic interior spaces in which a few water molecules might lodge. In the cytokine interleukin-1β, for example, such a cavity hosts between one and four water molecules on average⁶. There is predominantly a water dimer in the cavity itself, but it is not totally isolated: lone water molecules in two channels opening onto the cavity sometimes connect the dimer to the exterior⁷.

Such trapped water molecules seem to have significantly slowed dynamics relative to the bulk⁸, although it's not clear whether this depends on there being several water molecules present so that the dynamics become cooperative. If that's so, the trapped water might be expected to have a different dielectric constant from the bulk, altering its solvent properties⁹. If the same chemistry will work for larger fullerenes, Kurotobi and Murata might have devised a valuable method for exploring the characteristics of such small, hydrophobically encapsulated water clusters.