2D porous honeycomb polymers versus discrete nanocubes from trigonal trinuclear complexes and ligands with variable topology.

Trinuclear building block {Fe(2)NiO(Piv)(6)} (Piv = pivalate), which possessed pseudo-D(3h) symmetry, was linked by two ligands, pseudo-D(3h) ligand tris-(4-pyridyl)pyridine (L1) and C(2v) ligand 4-(N,N-dimethylamino)phenyl-2,6-bis(4-pyridyl)pyridine (L2) into two products with different topologies: 2D coordination polymer Fe(2)NiO(Piv)(6)(L1) (1), and discrete molecule [{Fe(2)NiO(Piv)(6)}(8) {L2}(12)], which had a nanocube structure (2). In compound 1, trinuclear {Fe(2)NiO(Piv)(6)} blocks were linked through ligand L1 into layers with honeycomb topology. In compound 2, eight trinuclear blocks were located in the vertices of the nanocube, with each L2 ligand linked to two {Fe(2)NiO(Piv)(6)} units. In the crystal structure, these nanocubes formed infinite catenated chains. Analysis of possible structures that could be assembled from these building blocks showed that compounds 1 and 2 corresponded to their respective predicted topologies. Compound [1⋅solvent] possessed a porous structure, in which the voids were filled by solvent molecules (DMF or DMSO). This structure was retained following desolvation, and compound 1 absorbed significant quantities of N(2) and H(2) at 78 K (S(BET) = 730 m(2)  g(-1), H(2) sorption capacity: 0.9 % by weight at 865 Torr). Desolvation of [2⋅solvent] led to disorder of its crystal structure, and compound 2 only adsorbed negligible quantities of N(2) but adsorbed 0.27 % H(2) (by weight) at 855 Torr and 78 K. The magnetic properties of these complexes (temperature dependence of molar magnetic susceptibility) were governed by the magnetic properties of the trinuclear "building block".