In the crystal structure of (b15c5)BiI₃(I₂), a bismuth(III) iodide molecule of ψ³-octahedral structure is attached to a benzo-15-crown-5 ligand to form a neutral (b15c5)BiI₃ complex. Bi³⁺−I⁻ distances are 287.9(1) (to I1), 296.4(6) (I2) and 287.3(1) pm (I3), see Figure 1. These distances with an average of 290.5 pm correspond very well to the Bi–I(terminal) distances of 292.0(2) pm in Cs₃Bi₂I₉ [13], which contains confacial bioctahedra [Bi₂I₉]³⁻. The distance from the central Bi³⁺ to the bridging I⁻ in this compound is 324.4(2) pm; for one [BiI₆] octahedron they average 308.2 pm, a distance that matches very well the 309.0 pm average of the individual Bi–I distances of 305.41(3) and 312.53(3) pm (three times each) in solid BiI₃ [14]. In (b15c5)BiI₃(I₂), I–Bi–I angles are 88.62(2) (I1–Bi–I2), 92.90(2) (I1–Bi–I3), and 91.45(2) pm (I2–Bi–I3), hence there is a meridional configuration with respect to an octahedron. The five oxygen atoms of benzo-15-crown-5 are situated opposite to the I1–I2–I3 triangle with Bi–O distances ranging from 282.4(7) to 291.1(5) pm, with an average of 286.7 pm.

Isostructural complexes (b15c5)BiX₃ with X = Cl, Br have already been characterized [15]. In these complexes the Bi–O distances are 283 (X = Cl) and 285 (Br) pm; hence the Bi–O distances are very similar, as expected, although there seems to be a slight trend from chloride to iodide. The Bi–X distances follow, with averages of 253 (X = Cl), 268 (Br) and 291 (I) pm, the increase of ionic radii from Cl⁻ via Br⁻ to I⁻. When these (167, 182, and 206 pm) [16] are subtracted from the mentioned averaged Bi–X distances, a reasonable ionic radius for Bi³⁺ of 85–86 pm is obtained.

In (b15c5)BiI₃(I₂), the discrete complexes (b15c5)BiI₃ are connected via iodine molecules to zig-zag chains, as Figure 1 shows. The I–I distance in the I₂ molecule is, with 273.9(3) pm, only slightly longer than in solid iodine, 271.5(6) pm at 110 K [7,8]. The distances I⁻–I (from (b15c5) (I^-)₂Bi³⁺–I⁻ to the I–I molecule) are 335.0(5) (I2-I5) and 353.7(2) pm (I3-I4), respectively. Both distances are close to the sum of the ionic radius of iodide (206 pm) and the atomic radius of iodine (140 pm) [17], 346 pm. The latter distance, 354 pm, is close to I–I distances between neighboring iodine molecules in solid iodine (350 pm). Both distances are considerably shorter than twice the van der Waals radius [18] of iodine, 2 × 198 = 396 pm. In summary, iodide–iodine interactions that form the zig-zag chains, qualify as “secondary bonding” [19].

Topologically, the zig-zag chains are arranged in layers, as Figure 2 shows, and the layers are stacked parallel to [100], as Figure 3 exhibits.

Trivalent bismuth has an electronic configuration of [Xe]6s²5d¹⁰4f¹⁴, hence filled 4f and 5d subshells and a lone 6s² electron pair. One could, therefore, believe that the special architecture of the (b15c5)BiI₃ complex were due to the 6s² pair being stereochemically active, as was suggested previously for (b15c5)BiX₃ (X = Cl, Br) [15]. If this were the case, a ψ¹-tetrahedral (with I–Bi–I angles around 109°) rather than a ψ³-octahedral structure (with I–Bi–I angles around 90°) would be expected. Indeed, it has been shown recently that lone-pair ions such as Tl(I)—isoelectronic with Bi(III)—suffer from 6s²-lone-pair–oxygen-2p orbital antibonding interactions leading to structural distortions [20].

As the interactions between the molecular complex (b15c5)BiI₃ and the iodine molecules in (b15c5)BiI₃(I₂) qualify as “secondary bonding”, calculations of the electronic structure have only been carried out for the (b15c5)BiI₃ complex. This subunit can be structurally described as a BiI₃ molecule interacting with the crown ether b15c5. Albeit BiI₃ itself forms an extended structure [14], it appears that the interaction with the crown ether favors the formation of a molecular unit.

Ab initio calculations were carried out with the aid of density functional theory (DFT); for technical details see Experimental Section and Supporting Information. Figure 4 shows the optimized ground state geometry of (b15c5)BiI₃ in comparison with the corresponding structural motif as observed in the crystal structure of (b15c5)BiI₃(I₂). No geometric or symmetry restraints were applied during the optimization process.

The theoretically calculated gas phase ground state geometry compares quite well with the experimentally observed structure of the complex. With 288 pm, the calculated interatomic Bi–I distances of the complex (b15c5)BiI₃ are slightly shorter than the average of the experimentally observed distances, 291 pm. However, as the calculations are done for T = 0 K and, furthermore, two of the iodide ions are connected to iodine molecules in (b15c5)BiI₃(I₂), these small deviations are no surprise. The mean interatomic Bi–O distances for the calculated (b15c5)BiI₃ unit are somewhat larger (324 pm) than experimentally observed distances, 291 pm, which is perhaps the result of packing effects in solid (b15c5)BiI₃(I₂).

Bi—I bonding interactions as represented by the HOMOs -11, -12 and -13 (Figure 5) are found at comparatively low energies. They are formed by interaction of iodine 6p orbitals with bismuth 6p orbitals. These molecular orbitals show only small contributions from the crown ether. The empty LUMOs (LUMO, LUMO+1 and LUMO+2) correspond to the anti-bonding set of the bismuth 6p and iodine 6p interactions. Again, only a small contribution of the crown ether moiety to these orbitals is observed. Thus, it is indeed justified to speak of a molecular BiI₃ unit interacting with the crown ether b15c5. However, the attractive BiI₃—crown-ether interactions have to be mainly electrostatic as HOMO-4, which is basically Bi—I nonbonding, shows a clearly antibonding interaction of the 6s orbital of bismuth with the 2p orbitals of the oxygen atoms of the crown ether. This is a similar situation as found for the interaction of isolated Tl⁺ cations with organic macrocycles [20]. In this case the combination of the occupied 6s orbital of Tl⁺ with the oxygen 2p orbitals of the macrocycle lead to fully occupied bonding and anti-bonding levels, and a subtle counterplay between repulsive covalent interactions and attractive electrostatic forces is observed. In the case of (b15c5)BiI₃ where BiI₃ interacts with the organic macrocycle b15c5 a similar bonding scenario is found. However, here it is not an isolated cation but a molecular BiI₃ unit which interacts. Thus, the anti-bonding interaction forms by a Bi-I group orbital with strong Bi-6s contribution.

A simplified, schematic energy diagram describing the interaction between bismuth and iodine to form BiI₃ and then between the BiI₃ fragment and the crown ether yielding (b15c5)BiI₃ is depicted in Figure 6.