Construction of Coordination Driven Self-Assembled [5+5] Pentagons using Metal-Carbonyl Dipyridine Ligands

Pentagonal molecular architectures possessing five-fold (C₅) symmetry exist throughout the chemical world from a wealth of inorganic species with pentagonal, pyramidal, bipyramidal, and prismatic geometries,¹ to all-carbon frameworks such as fullerenes and carbon nanotubes bearing curvature-inducing five-membered rings.² The unique C₅ symmetry has also been identified in nanoscale materials³ as well as DNA nanostructures.⁴ Moreover, a 2D arrangement of pentagonal structures with C₅ or quasi-C₅ symmetry, distinct from the significantly more common C₂, C₃, C₄ and C₆ periodic symmetries, has been a long-term target pursued by crystal engineers that has been meet with limited success.⁵ Pentagonal architectures are also attractive because of their potential applications in functional materials such as quasicrystals⁶ and discotic liquid crystals.⁷

In light of their potentials, the synthesis of discrete pentagonal architectures has remained a formidable challenge. Only a few discrete C₅-symmetrical pentagonal organic molecules have been synthesized, generally in low yield and through arduous synthetic work.⁸ Coordination-driven self-assembly has been extensively explored in the past few decades and shown to be a powerful synthetic strategy for the construction of metallosupramolecular architectures.⁹ By combining specifically designed organic donor building blocks with directional metal acceptors, a plethora of two-dimensional supramolecular structures, including molecular loops,¹⁰ triangles,¹¹ squares,¹² and hexagons,¹³ have been synthesized in high yield. This synthetic methodology provides an efficient and viable means to construct discrete pentagonal structures.

Our group has long endeavored to establish a “molecular library” of metallosupramolecular structures built from the coordination-driven self-assembly of appropriately designed Pt(II) or Pd(II) acceptors and specifically angled donor units in a controllable manner.^9a–d According to this design concept, discrete pentagonal entities may be exclusively assembled by the incorporation of five 108° building units with five complementary linear units. However, the scarcity of suitable 108° subunits has complicated the realization of such a design concept. In the few reported examples of supramolecular metal-ligand pentagonal architectures,¹⁴ polydentate flexible ligands were used since their coordination to metal centers may lead to a 108° bonding conformation, though an encapsulated anion of specific size must be included to template the assembly process. Furthermore, the difficulty to develop a common methodology to construct a metallosupramolecular pentagon also arises from the internal turning angle of a regular pentagon, 108°, which is close to that of a regular hexagon (120°).¹⁵ The small 12° difference between the 120° angle needed for a hexagonal assembly and the 108° angle needed for an analogous pentagon assembly often leads to an equilibrium mixture of pentagon and hexagon suprastructures.

Acetylene units (C≡C) are extensively incorporated into many donor and acceptor building blocks due to their rigid linear conformation. In view of the ready reactivity¹⁶ of a wide range of metal carbonyl cluster complexes with acetylene moieties, we envisioned that the steric bulk of a metal carbonyl cluster species adhered to the acetylene moiety may be used as a control factor to adjust the bonding angle of the building block in order to exclusively form a pentagon self-assembly. Two metal carbonyl dipyridine adduct ligands, (4-C₅H₄N)₂C≡CCo₂(CO)₆ (1) and (4-C₅H₄N)₂C≡CMo₂Cp₂(CO)₄ (2), were synthesized and were combined with a linear acceptor ligand bis[1,4-(trans-Pt(PEt₃)₂OTf)]ethynylbenzene (3) to investigate the possibility of constructing [5+5] pentagonal metallosupramolecules (Scheme 1).

Scheme 1.

Molecular structures of donor (red) and acceptor (blue) building blocks and their self-assembly into metallacyclic supramolecules.

Crystallographic studies have shown that the acetylene moiety adducted by Co₂(CO)₆ can form a tetrahedral Co₂C₂ core,^16b thus making an angle of 120° between the two pyridine rings in 1. Self-assembly between this 120° donor with the complementary linear acceptor 3 is assumed to construct a [6+6] hexagon. The reaction of 1 with 3 in a 1:1 ratio in CD₂Cl₂ gave a wine-colored homogeneous solution of 4, whose ³¹P{¹H} NMR spectrum showed a single peak at 16.5 ppm with concomitant ¹⁹⁵Pt satellites, upfield shifted by roughly 6.4 ppm compared with 3 (δ = 23.0 ppm) as a result of the coordination of the pyridine rings (Figure S2 in Supporting Information). However, the proton NMR of 4 displayed broad signals in contrast to the sharp peaks of previously reported for discrete hexagon structures,¹⁷ implying the possible existence of several species in the mixture (Figure S2). The ESI-TOF mass spectrum indicated that two self-assembled polygons – [5+5] pentagon and [6+6] hexagon – do indeed co-exist in self-assembly 4. Two charge states at m/z = 2040.0 and 1310.3 corresponding to [pentagon – 4CF₃SO₃]⁴⁺ and [pentagon – 6CF₃SO₃]⁶⁺, respectively, were observed and were in good agreement with their theoretical isotopic distributions. The isotropically well-resolved mass peak at m/z = 1952.8, resulting from [hexagon – 5CF₃SO₃]⁵⁺, was found in the mass spectrum as well (Figure 1a).

Figure 1.

ESI-TOF-MS spectra of (a) self assembly 4 containing both pentagon and hexagon, and (b) two charge states of [5+5] pentagon 5. Red vertical lines are the theoretical abundances.

Mixing molybdenum cluster donor ligand 2 in a 1:1 stoichiometric ratio with 3 generated a homogeneous dark-red solution of 5. A single sharp peak at 16.7 ppm with two ¹⁹⁵Pt flanking satellites were observed in the ³¹P{¹H} NMR spectrum of 5 (Figure S3). The proton NMR spectrum of 5 displayed sharp signals with differentiable coupling constants (Figure S3). The signals of the pyridine ring α-protons experienced a small upfield shift of 0.03 ppm, but the β-protons and the hydrogen nuclei on the Cp ring undergo approximately 0.2–0.3 ppm downfield shifts, suggestive of the strong back-donation effect of the molybdenum carbonyl cluster to the pyridine ring.

The ESI-TOF mass spectrum of 5 displayed four peaks corresponding to four charge states of the [5+5] pentagon, including [M – 3CF₃SO₃]³⁺ (m/z 3016.6), [M – 4CF₃SO₃]⁴⁺ (m/z 2225.0), [M – 5CF₃SO₃]⁵⁺ (m/z 1750.2, overlapping with the 1+ fragment) and [M – 6CF₃SO₃]⁶⁺ (m/z 1433.5), which were all isotopically well-resolved and agree very well with their respective theoretical distributions (the 4+ and 6+ charge state are illustrated in Figure 1b and the full spectrum is shown in Figure S5). No evidence for any other species such as a [4+4] square, [6+6] hexagonal, or [7+7] heptagonal assembly was found. The exclusive formation of a [5+5] pentagon is also supported by comparing the proton NMR spectrum of pentagon-hexagon mixture 4 and that of 5, wherein the peaks of the former are much broader than the latter (Figure S6).

Our attempts to crystallize the polygonal structures 4 (pentagon and hexagon mixture) and 5 (exclusively pentagon) have so far been unsuccessful. We have therefore used molecular force field simulations to investigate the structural details of the supramolecular pentagon and hexagon composed of cobalt donor 1 and linear di-Pt(II) acceptor 3, as well as the [5+5] pentagon formed by the self-assembly of molybdenum donor 2 with 3. In the case of the pentagonal and hexagonal supramolecules that incorporate 1 with 3, the energies of the two different polygon structures are nearly identical with the [6+6] hexagon being slightly more stable. In the self assembly between 2 and 3, modeling suggests the pentagonal structure is more stable than the hexagon. The modeled suprastructures show that the linear acceptor units in the hexagonal structure must distort away from a 180° orientation in order to fit the complementarity requirement of a [6+6] hexagon whereas the acceptors retain their 180° geometry in the modeled [5+5] pentagonal structure (Figure 2). The formation of a discrete [5+5] pentagon in 5, derived from 2, can also be rationalized by the fact that the primary steric effect of the Cp rings in building block 2 forces the two pyridine rings closer to each other, thus forming a smaller bonding angle than in the related donor 1.

Figure 2.

Pentagonal and hexagonal structures composed of molybdenum donor ligand 2 and linear acceptor 3 as obtained from molecular force field modeling.

We also investigated the self-assembly of 1 and 2 with the 120° ferrocenyl acceptor 6 in order to construct a Fe₃-Co₆-Pt₆ and a Fe₃-Mo₆-Pt₆ trimetal [3+3] hexagons, respectively, and substantiate the effects of the bonding angle difference between 1 and 2 in self-assembly(Scheme 1). Preparation of the novel trimetal Fe₃-Co₆-Pt₆ [3+3] hexagon 7 was successfully achieved by mixing 1 with ferrocenyl acceptor 6 in a 1:1 ratio in dichloromethane. The resulting red solution was characterized by ¹H and ³¹P{¹H} NMR spectroscopy, with the latter displaying a single sharp peak (Figure S4). The ESI mass spectrum exhibited two charge states of the Fe₃-Co₆-Pt₆ hexagon 7 (m/z = 1335.1 and 1038.3 for +4 and +5, respectively), which were isotropically resolved and were in good agreement with theoretical isotopic distributions (Figure S7). However, our attempt to produce a Fe₃-Mo₆-Pt₆ trimetal [3+3] hexagon using a similar synthetic protocol employing 2 instead of 1 was unsuccessful. This differing reactivity of related donor ligands 1 and 2 further confirms that the bonding angles of the two pyridine rings of each metal carbonyl donor ligand are measurably dissimilar in coordination-driven self assembly because of the difference in the steric bulk of Co₂(CO)₆ and Mo₂Cp₂(CO)₄.

In conclusion, we have successfully prepared a [5+5] supramolecular pentagon by the self-assembly of a molybdenum carbonyl cluster dipyridyl donor ligand (2) with a linear di-Pt(II) acceptor (3). The roughly 108° bonding angle encoded within 2 directs this coordination-driven self-assembly process to form a single pentagonal metallosupramolecule rather than a pentagon-hexagon mixture, implying that the generality of this method can be extend to construct other pentagonal structures with a variety of functional groups for even more advanced multifunctional materials.