Single-crystal to single-crystal transformations in discrete solvated metallocycles: the role of the metal ion

Abstract

Crystals of solvated cobalt- and zinc-containing metallocycles undergo single-crystal to single-crystal (SCTSCT) transformations upon desolvation to yield their close-packed forms; this is in stark contrast to the analogous cadmium-based metallocycle, which affords an empty, transiently porous phase upon desolvation.

---

In contrast to the success of porous network materials such as metal–organic frameworks¹ and zeolites,² examples of molecular crystalline materials that display permanent or transient porosity are still relatively rare since molecular compounds tend to pack very efficiently.³ Owing to the need for microporous synthetic materials for the separation, storage and sensing of gases,⁴ much effort has been dedicated to countermanding close-packing tendencies in molecular compounds to engineer voids or channels in the crystalline state. In recent years the majority of this research has been devoted to studies of crystalline coordination compounds. This can partly be ascribed to the great structural diversity arising from different transition metal coordination geometries and the variety of multidentate ligands available for building supermolecules.⁵ Indeed, such materials have already shown much promise, with potential applications including molecular recognition and catalysis.⁶ Other pathways for the construction of porous molecular compounds include classical non-reversible covalent synthesis and the utilization of labile covalent linkages for assembling macromolecules,⁷ with interest in the latter slowly gaining momentum since it generally circumvents time-consuming classic covalent synthesis.⁸

Notwithstanding recent reports of labile organic synthetic products, synthesis based on coordination-driven self-assembly still offers an easy and versatile route to discrete cyclic units and cage compounds,⁹ albeit that predicting the outcomes of rationally designed experiments is still not generally possible.¹⁰ However, our approach to constructing metallocyclic host molecules using flexible ditopic imidazole ligands have proven to be successful.¹¹ Furthermore, these cyclic molecules are awkwardly shaped, which impacts on their ability to form compact structures. Thus, even when efficiently packed in the crystalline state, these structural motifs are predisposed to yield significant solvated space or voids in the structure. We have also demonstrated that the interstitial solvent molecules can be readily removed to yield a material that is permeable to gases and/or vapours, often in a single-crystal to single-crystal fashion.¹¹ Indeed, in this regard we recently reported a new cadmium chloride based metallocyclic host that is capable of accepting two small guest molecules per discrete cavity.¹² We showed that the conformational integrity of the metallocyclic host prevails when the methanol guest molecules are removed to yield void space transiently accessible to carbon dioxide and acetylene; these guest uptake processes proceed in a single-crystal to single-crystal fashion. We have now extended our investigation of this metallocycle by combining the ligand with other metal chlorides with the objective of making similar metallocycles with tuneable void spaces.

The ligand 4,4′-bis(2-methylimidazol-1-ylmethyl)biphenyl, L was complexed with cobalt chloride and zinc chloride to yield neutral cyclic dinuclear complexes [M₂L₂Cl₄] (where M = Co²⁺ or Zn²⁺). Within each complex the two metal ions are tetrahedrally coordinated to two chloride anions and two C-shaped bridging ligands (Fig. 1). These two metallocycles are isostructural with respect to the previously reported cadmium metallocycle.

Fig. 1 Thermal ellipsoid plot (50% probability) of [Co₂L₂Cl₄]·2MeOH. The metallocycles of [Zn₂L₂Cl₄]·2MeOH and [Cd₂L₂Cl₄]·2MeOH are isostructural to that depicted here.

Fig. 2 shows how the metallocycles stack to form columns, and that this packing mode creates solvent-filled pockets between adjacent metallocycles. Each metallocycle shares cavities with both of its neighbours (i.e. a solvent-filled cavity exists between metallocycles 2 and 3, but also between metallocycles 1 and 2, and metallocycles 3 and 4). The ends of the approximately rectangular pocket are bounded by the chloride ions, while the phenyl and methyl hydrogen atoms of L form the walls of the cavity, i.e. chloride anions of metallocycles 1 and 4 (shown space-filled) stopper the top and bottom of a cavity while the sides are formed by the van der Waals surfaces of metallocycles 2 and 3. Each cavity is occupied by two methanol guest molecules, which are presumed to template the assembly of the molecular rings (which approximate hexagons). Two chloride anions of each metallocycle form O–H⋯Cl⁻ hydrogen bonds with the two solvent molecules: D⋯A = 3.203(6), 3.186(9) and 3.173(5) Å for the series cobalt, zinc and cadmium, respectively. The metallocycles and methanol guests for all of these structures are located about sites of 2/m symmetry, with a mirror plane running parallel to the projection plane of Fig. 2 and passing through the metal centres.

Fig. 2 The metallocycles encapsulating guest molecules, as viewed along [010]. The neutral complexes are shown as capped-sticks and the guest molecules and interacting chloride anion are shown in van der Waals representation. Connolly surfaces¹³ (using a probe radius of 1.4 Å) of the solvent-filled pockets are shown in semi-transparent grey. The host enclathrates two methanol molecules per cavity with mapped volumes increasing from 118 to 124 to 131 ų, respectively for the series cobalt, zinc and cadmium.

In order to study the removal of solvent molecules from the host structures, crystals were subjected to thermogravimetric analysis, which showed that desolvation of compounds [Co₂L₂Cl₄]·2MeOH and [Zn₂L₂Cl₄]·2MeOH occurs readily, even at room temperature (Fig. 3). A mass loss of 6.4% is expected for two molecules of methanol per cobalt metallocycle, but a slightly lower value of 5.6% was obtained thermogravimetrically. This can be attributed to the spontaneous desolvation process starting as soon as the crystals are removed from the mother liquor. Using the as-synthesised crystals and a heating rate of 2.5 °C per minute, weight-loss due to solvent removal is complete at 63 °C. Differential scanning calorimetry (Fig. 3) reveals an endotherm that coincides with the guest-loss process observed by thermogravimetric analysis. To determine whether the single crystals survive the desorption process intact, selected crystals of [Co₂L₂Cl₄]·2MeOH and [Zn₂L₂Cl₄]·2MeOH were heated at ca. 65 °C for several hours and then inspected under a microscope. Outwardly, the crystals appeared to maintain their level of mosaicity and single-crystal diffraction analysis showed, in both cases, that the ligands undergo significant conformational changes upon desolvation of the material. This results in the formation of an apohost phase in which the cyclic complexes are distorted to yield a more efficiently-packed structure. This is in contrast to the Cd-based metallocycle where no ligand rearrangement is observed, in fact, removal of solvent molecules results in void spaces of ca. 120 ų.

Fig. 3 Thermogravimetric analysis of [Co₂L₂Cl₄]·2MeOH showing guest loss starting at room temperature and differential scanning calorimetry, indicating a phase change concomitant with guest loss between 20 and 70 °C.

In all three cases (i.e. M = Cd, Co, Zn) the solvated structure crystallises in the monoclinic space group C2/m. A mirror plane passes through the two metal centres and the four Cl⁻ anions, and a two-fold rotation axis perpendicular to this plane bisects the two C1–C1′ biphenyl bonds (i.e. the site symmetry at the centre of the metallocycle is 2/m as shown in Fig. 4a). This requires only a quarter of the metallocycle to be in the asymmetric unit. During desorption in the cases of M = Zn or Co, one of the ligands apparently inverts its conformation such that the metallocycle loses its mirror symmetry (Fig. 4b), but retains its two-fold symmetry (conformational flexibility induced by guest-loss has previously been observed for an analogous copper-based metallocycle).¹⁴ This results in doubling of the size of the asymmetric unit to half a metallocycle (i.e. two separate halves of a ligand, a metal cation and two halide anions). The desolvated structure conforms to the monoclinic space group C2/c and the changes in its symmetry result in doubling of the crystallographic c axis. Twisting of the ligand as a result of desolvation promotes closer packing of the complexes in the apohost structures – i.e. the space previously filled by guest molecules is now mostly occupied as a result of more efficient packing of the reconformed metallocycles (Fig. 5). The conformational change is accompanied by a slight elongation of the metal⋯metal distance within the metallocycle from 15.881(2) to 16.295(4) Å for [Co₂L₂Cl₄], and from 15.964(2) to 16.365(3) Å for [Zn₂L₂Cl₄]. There is also a minor decrease in the N–metal–N angle from 115.6(2)° to 108.7(3)° for [Co₂L₂Cl₄], and from 113.0(3)° to 107.3(2)° for [ZnL₂Cl₄]. Crystals of [Co₂L₂Cl₄] and [Zn₂L₂Cl₄] were immersed in acetonitrile and methanol in an attempt to regain a solvated phase (and the expected corresponding 2/m conformation), but these attempts were unsuccessful. It is important to note that, upon ligand rearrangement, significant changes occur within the lattice and one would expect that this would result in sufficient mechanical strain to cause cracking of the crystals. Therefore, it is surprising that desolvation occurs as single-crystal to single-crystal transformations.

Fig. 4 Van der Waals projections of [M₂L₂Cl₄]·2MeOH (M = Co or Zn) showing the conformation of metallocycles (a) as grown from solution and (b) [M₂L₂Cl₄] after desolvation. Molecules are viewed perpendicular to the (100) plane and methyl groups of the 2-methylimidazole moieties are shown in dark red for clarity. Loss of the horizontal mirror plane upon desolvation is illustrated.

Fig. 5 Capped-stick representation of the desolvated rearranged metallocycles [M₂L₂Cl₄]·2MeOH (M = Co or Zn) stacked along [001]. The semi-transparent grey surface shows the small void space which has decreased to ca. 26 and 25 ų respectively for cobalt and zinc (using a probe radius of 1.175 Å).

In conclusion, we have previously reported the formation of a cadmium-based metallocycle that is robust to solvent removal, thus affording a transiently¹⁵ porous apohost phase capable of hosting various gaseous molecules. We have now extended this work to include analogous and isostructural cobalt- and zinc-based metallocycles. Interestingly, these metallocycles do not exhibit the same conformational indifference to solvent removal; the desolvated metallocycles show a marked difference in spatial arrangement of the ligand moieties, thus resulting in a nonporous apohost phase. We rationalise the isoskeletal¹⁶ stabilisation of Cd²⁺ > Zn²⁺ = Co²⁺ under the investigated conditions (where coordination number, geometry and ligands are equivalent), based on size and charge considerations. Co²⁺ and Zn²⁺ possess nearly identical ionic radii¹⁷ for the same coordination number, as well as similar stabilities for most ligands;¹⁸ they have 3d⁸ and 3d¹⁰ configurations, respectively. Cd²⁺ on the other hand has a 4d¹⁰ electronic configuration and has a slightly larger ionic radius.¹⁷ As a result of an equivalent charge, a larger radius and complete d-orbital filling, Cd²⁺ has a lower charge density (soft acid) with better covalent bonding character.¹⁸ This subtle difference in stability is in agreement with the relative binding strengths of Co²⁺, Zn²⁺ and Cd²⁺ with imidazole, as well as chloride as donor ligands in aqueous solutions, as reflected by their stability constants.¹⁹

Experimental

Synthesis

[Co₂L₂Cl₄]·2MeOH. 0.04 mmol L (13.70 mg) was dissolved in 2 ml chloroform and layered with 0.04 mmol CoCl₂·6H₂O (9.52 mg) dissolved in 2 ml methanol. After three days blue needle-shaped crystals had grown that were suitable for single-crystal diffraction studies. Yield: 60%. ATR FT-IR ν/cm⁻¹ 2365 (w), 1545 (w), 1499 (s), 1425 (m), 1363 (m), 1143 (m), 1005 (m), 755 (s), 672 (s).

[Zn₂L₂Cl₄]·2MeOH. 0.03 mmol L (10.27 mg) was dissolved in 2 ml chloroform and layered with 0.03 mmol ZnCl₂ (4.09 mg) in 2 ml methanol. After a week, colourless rods were afforded and the structure elucidated by single-crystal diffraction analysis. Yield: 58%. ATR FT-IR ν/cm⁻¹ 2365 (w), 1548, (w), 1500 (s), 1426 (m), 1364 (m), 1143 (m), 1005 (m), 756 (s), 669 (m).

Single-crystal X-ray diffraction

Crystal data for [Co₂L₂Cl₄]·2MeOH: C₄₆H₅₂Cl₄Co₂N₈O₂, M = 1008.62, blue prism, 0.30 × 0.22 × 0.11 mm³, monoclinic, space group C2/m (No. 12), a = 17.232(2), b = 15.471(2), c = 9.4730(11) Å, β = 114.150(2)°, V = 2304.3(5) ų, Z = 2, D_c = 1.454 g cm⁻³, F₀₀₀ = 1044, MoKα radiation, λ = 0.71073 Å, T = 100(2) K, 2θ_max = 56.7°, 7310 reflections collected, 2774 unique (R_int = 0.0708). Final GooF = 1.161, R₁ = 0.0826, wR₂ = 0.1660, R indices based on 2072 reflections with I > 2σ (I) (refinement on F²), 150 parameters, 0 restraints. Lp and absorption corrections applied, μ = 0.999 mm⁻¹.

Crystal data for [Zn₂L₂Cl₄]·2MeOH: C₄₆H₅₂Cl₄N₈O₂Zn₂, M = 1021.50, colourless prism, 0.29 × 0.24 × 0.13 mm³, monoclinic, space group C2/m (No. 12), a = 17.36(2), b = 15.483(20), c = 9.440(12) Å, β = 113.97(2)°, V = 2318(5) ų, Z = 2, D_c = 1.463 g cm⁻³, F₀₀₀ = 1056, MoKα radiation, λ = 0.71073 Å, T = 100(2) K, 2θ_max = 56.4°, 7139 reflections collected, 2778 unique (R_int = 0.1075). Final GooF = 1.040, R₁ = 0.0859, wR₂ = 0.1777, R indices based on 1809 reflections with I > 2σ(I) (refinement on F²), 150 parameters, 0 restraints. Lp and absorption corrections applied, μ = 1.313 mm⁻¹.

Crystal data for [Co₂L₂Cl₄]: C₄₄H₄₄Cl₄Co₂N₈, M = 944.53, blue prism, 0.30 × 0.22 × 0.11 mm³, monoclinic, space group C2/c (No. 15), a = 20.020(4), b = 14.595(3), c = 17.102(3) Å, β = 118.62(3)°, V = 4386.6(19) ų, Z = 4, D_c = 1.430 g cm⁻³, F₀₀₀ = 1944, MoKα radiation, λ = 0.71073 Å, T = 100(2) K, 2θ_max = 56.7°, 13 411 reflections collected, 5107 unique (R_int = 0.1106). Final GooF = 1.152, R₁ = 0.1304, wR₂ = 0.2309, R indices based on 3069 reflections with I > 2σ(I) (refinement on F²), 264 parameters, 0 restraints. Lp and absorption corrections applied, μ = 1.042 mm⁻¹.

Crystal data for [Zn₂L₂Cl₄]: C₄₄H₄₄Cl₄N₈Zn₂, M = 957.41, colourless prism, 0.24 × 0.20 × 0.19 mm³, monoclinic, space group C2/c (No. 15), a = 20.010(5), b = 14.658(3), c = 17.146(4) Å, β = 118.325(3)°, V = 4426.7(18) ų, Z = 4, D_c = 1.437 g cm⁻³, F₀₀₀ = 1968, MoKα radiation, λ = 0.71073 Å, T = 100(2) K, 2θ_max = 56.6°, 12 944 reflections collected, 5067 unique (R_int = 0.0908). Final GooF = 1.019, R₁ = 0.0804, wR₂ = 0.1650, R indices based on 2502 reflections with I > 2σ(I) (refinement on F²), 264 parameters, 0 restraints. Lp and absorption corrections applied, μ = 1.366 mm⁻¹.

Notes and references

1. (a) O. M. Yaghi, M. O'Keeffe, N. W. Ockwig, H. K. Chae, M. Eddaoudi and J. Kim, Nature, 2003, 423, 705; (b) Y. L. Liu, V. C. Kravtsov, R. Larsen and M. Eddaoudi, Chem. Commun., 2006, 1488; (c) S. L. James, Chem. Soc. Rev., 2003, 32, 276.
2. D. W. Breck, Zeolite Molecular Sieves, John Wiley & Sons, Inc., New York, 1974.
3. (a) J. L. Atwood, L. J. Barbour and A. Jerga, Science, 2002, 296, 2367; (b) J. R. Holst, A. Trewin and A. I. Cooper, Nat. Chem., 2010, 2, 915.
4. D. M. Rudkevich, Eur. J. Org. Chem., 2007, 3255.
5. H. Amouri, C. Desmarets and J. Moussa, Chem. Rev., 2012, 112, 2015.
6. (a) A. Kumar, S. S. Sun and A. J. Lees, Coord. Chem. Rev., 2008, 252, 922; (b) Y. Inokuma, M. Kawano and M. Fujita, Nat. Chem., 2011, 3, 349.
7. M. J. Bojdys, M. E. Briggs, J. T. A. Jones, D. J. Adams, S. Y. Chong, M. Schmidtmann and A. I. Cooper, J. Am. Chem. Soc., 2011, 133, 16566.
8. M. Mastalerz, Angew. Chem., Int. Ed., 2010, 49, 5042.
9. S. Leininger, B. Olenyuk and P. J. Stang, Chem. Rev., 2000, 100, 853.
10. B. Moulton and M. J. Zaworotko, Chem. Rev., 2001, 101, 1629.
11. (a) L. Dobrzanska, G. O. Lloyd, H. G. Raubenheimer and L. J. Barbour, J. Am. Chem. Soc., 2005, 127, 13134; (b) L. Dobrzanska, G. O. Lloyd, H. G. Raubenheimer and L. J. Barbour, J. Am. Chem. Soc., 2006, 128, 698; (c) T. Jacobs, J.-A. Gertenbach, D. Das and L. J. Barbour, Aust. J. Chem., 2010, 63, 573.
12. T. Jacobs, G. O. Lloyd, J.-A. Gertenbach, K. K. Mueller-Nedebock, C. Esterhuysen and L. J. Barbour, Angew. Chem., Int. Ed., 2012, 51, 4913.
13. M. L. Connolly, Science, 1983, 221, 709.
14. L. Dobrzanska, G. O. Lloyd, C. Esterhuysen and L. J. Barbour, Angew. Chem., Int. Ed., 2006, 45, 5856.
15. L. J. Barbour, Chem. Commun., 2006, 1163.
16. G. O. Lloyd, J. Alen, M. W. Bredenkamp, E. J. C. de Vries, C. Esterhuysen and L. J. Barbour, Angew. Chem., Int. Ed., 2006, 45, 5354.
17. R. D. Shannon, Acta Crystallogr., Sect. A: Cryst. Phys., Diffr., Theor. Gen. Crystallogr., 1976, 32, 751.
18. Encyclopedia of Inorganic Chemistry, ed. R. B. King, John Wiley & Sons Ltd., Chichester, 1994, vol. 4.
19. (a) S. Sjoberg, Pure Appl. Chem., 1997, 69, 1549; (b) Critical Stability Constants, ed. A. E. Martell and R. M. Smith, Plenum, New York, 1974, vol. 4.

---

Footnotes

† This article is included in the All Aboard 2013 themed issue.

‡ Electronic supplementary information (ESI) available: Experimental details, thermal analysis and crystallographic data. CCDC 892651–892654. For ESI and crystallographic data in CIF or other electronic format see DOI: 10.1039/c2nj40612h

---