We have been trying to make very large molecules for some time, inspired by gigantic polyoxometalate cages from Achim Müller’s group and by the beautiful supramolecular assemblies from the Fujita group and others. One idea we had was to functionalise our heterometallic Cr7Ni rings with the extended bipyridyl head-groups, borrowing the idea directly from Fujita. We designed a route to a rotaxane where we could grow a Cr7Ni ring around an organic thread from which an extended bipyridyl could be made. This works very nicely giving the rotaxane shown. The bipyridyl head-group is similar to those used by Fujita’s group to make Pd12 nanocapsules, where the Pd sites sit on the vertices of a cuboctahedron.
When reacted with palladium(II) tetrafluoroborate the result was unexpected. The compound formed was a rotaxane and only six Pd sites are found, making an octahedron. There were also six dethreaded Cr7Ni rings present. The yield was atrocious, so we carried out the reaction again, but now adding an additional Cr7Ni ring that we know dethreads easily. The yield improved substantially.
The structure determined by X-ray diffraction is beautiful. An octahedron of Pd metals bridged on each edge by a rotaxane – making twelve in total. The six Cr7Ni anions sit above the Pd sites and are shown striped in the video below. The other Cr7Ni rings are in green and the Pd are silver.
The crystallography was immensely challenging. The Fo-Fc electron density maps are really beautiful as art objects, but a pain to refine; for example, the structure contains 288 t-butyl groups. The crystallographers may never forgive us.
We wanted to prove that these assemblies exist in solution and we found that small angle X-ray scattering is an ideal technique. We made models of the whole structure and fragments of the structure to show that the structure was maintained in solution. It is; the Cr7Ni anion rings appear to be held in place in the structure when dissolved in THF. The blue and red traces shown are the experimental and calculated trace for the whole structure respectively. The other colours are calculated traces for fragments.
The first question we have, which we cannot answer, is why are we getting the Pd6 nanocapsule rather than the Pd12 nanocapsule we had expected? The rotaxane is a very bulky ligand so one might expect a bigger than expected nanocapsule. The compound formed is very tightly packed and we wonder whether very weak -C-H….H-C- interactions between the many t-butyl groups present are providing enough energy to drive us towards this structure rather than another.
This compound contains one hundred and fifty metal sites and nearly three thousand non-hydrogen atoms. We believe we can make even bigger supramolecules by this route: we must be able to make the rotaxane and maybe even a rotaxane by using a rotaxane as our linker.
It is possible the crystallographers may give up on us at that point!