Molecular assembly of two [Co(II)₄] linear arrays

Abstract

Molecular chains of four Co(II) ions stabilized by a bis-β-diketone/pyridyl ligand may be isolated or linked into molecular pairs of two semi-independent such units.

---

The design of Coordination Chemistry assemblies represents a promising avenue for the construction of functional molecular devices. Future information technologies represent one of the areas that will benefit from this methodology. For example, single molecule magnets (SMMs) are coordination complexes that exhibit hysteresis of their magnetization, offering the prospect of storing magnetic information into individual molecules.^1,2 Along these lines, magnetic coordination clusters have been proposed as possible hardware for implementation of quantum information processing (QIP).³ Under such an approach, universal logic operations (C-NOT or SWAP operations) would be carried out by pairs of weakly coupled well defined electronic spins. In this context, we have been engaged for years in the design of ligands for the assembly of magnetic metal ions in form of molecules displaying two weakly coupled spins (as single ions or aggregates of these).^4–7 The ultimate goal is developing synthetic methodologies for accessing molecular entities that fulfil the requirements to behave as 2qubit quantum gates. We have recently prepared a novel multidentate ligand that alternates three pyridyl donors and two β-diketone groups within its structure (H₂L, 1,3-bis-(3-oxo-3-(2-pyridyl)-propionyl)-pyridine, Scheme 1). This donor is structurally related to 2,6-di(acetoacetyl)pyridine, known to produce heterometallic linear complexes.⁸ The capacity of H₂L to facilitate the assembly of molecular chains of transition metal ions was now demonstrated with the synthesis of the complex [Co₄L₂(MeOH)₈](NO₃)₄ (1).⁹ With the goal of elaborating molecular pairs of clusters, we now have used this linear polynuclear complex as the building block for the assembly of one such architecture comprising two clearly identifiable [Co₄] units. Complex 1 seems appropriate for this purpose since the [Co₄L₂]⁴⁺ moiety appears as a robust assembly, which in turn possesses labile axial donors properly disposed for their substitution by difunctional ligands that could act as linkers of tetranuclear chains.

Scheme 1 H₂L.

Initial attempts to link two tetranuclear chains involved the use of difunctional aromatic spacers, such as pyrazine, aminopyrazine, imidazole, terephthalic acid, etc., in reactions with complex 1 (the latter usually generated in situ). These trials did never furnish any indication that the assembly of [Co₄] molecules was occurring. In the one case where crystals for X-ray diffraction could be obtained, the complex was found to be the tetranuclear aggregate [Co₄L₂(H₂O)₆(MeOH)₂](NO₃)₄ (2), which is a close derivative of 1, with some of the methanol axial ligands replaced by molecules of water (Fig. 1). The structure of 2‡ † shows a complex cation of four Co(II) ions linked together as a zig-zag chain and chelated on all the equatorial positions by two L²⁻ ligands located opposite to each other, sandwiching the string of metals. This ensemble (of 86 atoms) features a remarkable flat configuration (Fig. 1). The axial positions of the metals are occupied by six molecules of H₂O and two of MeOH. The metals exhibit either heptacoordination (as a pentagonal bipyramid; Co2 and its equivalent) or distorted octahedral geometry (Co1 ions). The distances between adjacent metals (3.475 and 3.616 Å) preclude the possibility of any metal–metal bond by direct overlap (only expected for distances around 2.3 Å).¹⁰

Fig. 1 PovRay representation of the complex cation of [Co₄L₂(H₂O)₆(MeOH)₂](NO₃)₄ (2). Only unique heteroatoms are labelled. Only hydrogen atoms bound to O are shown (in yellow). At the bottom is a representation emphasizing the flatness of the equatorial fragment. Ranges of selected distances (Å) and angles (°): Co–O, 2.0179(11) to 2.2778(10); Co–N, 2.0840(14) to 2.1831(12); Co1–O3–Co2, 109.90(6); Co1–O1#–Co2, 107.76(5); Co2–O2–Co2#, 109.23(5); Co1⋯Co2, 3.4754(4); Co2⋯Co2#, 3.6164(4); Co1⋯Co2#, 6.8721(6); Co1⋯Co1#, 10.2728(8). # = 1 − x, 1 − y, −z.

The charge +4 of this assembly is compensated by four NO₃⁻ counterions, located within the crystal lattice, helping to connect the cations to each other through a network of hydrogen bonds.† This molecule joins a small group of structurally characterized tetranuclear complexes of Co(II) with a chain-like topology.^5,9,11–13 The structure of 2 corroborates the hypothesis that the [Co₄L₂]⁴⁺ unit is a robust moiety and remains intact after substitution of MeOH ligands by H₂O. In fact, it was inferred previously that the all methanol derivative 1 replaces completely its MeOH ligands by water upon exposure to air, while maintaining the molecular backbone.⁹ The magnetic properties of 2 in air are thus expected to be exactly the same as those of 1. New attempts to link chemically [Co₄L₂]⁴⁺ building blocks into larger molecules were carried out by the use of 4,4′-bipyridine (bpy), which contains two donor rings. Thus, the reaction mixture of Co(AcO)₂/Co(NO₃)₂, H₂L and bpy in methanol using the 1 ∶ 1 ∶ 1 ∶ 2 molar ratio, generated an orange precipitate and a deep red solution, which upon layering with toluene produced crystals of the targeted cluster [Co₈L₄(OH)₂(H₂O)₄(NO₃)₂(bpy)₄](NO₃)₄ (3).§

The structure of 3‡ † (Fig. 2) reveals two [Co₄L₂] units as these discussed above, linked covalently, face-to-face by four difunctional 4,4′-bipyridine ligands bound to axial positions of all Co(II) metals. The external apical sites are occupied by two OH⁻ ligands, two NO₃⁻groups and four molecules of water (the hydrogen atoms of these ligands were found in difference Fourier maps). The metric parameters around the metals are very similar to these for complexes 1 and 2, leading here to [Co₄L₂] moieties slightly more removed from planarity than seen previously. Indeed, the inequality between both sides of this plane pushes the rings from ligand L²⁻ away from the molecular core, perhaps as a result of steric repulsions with the bpy ligands (Fig. 2). It is unclear why complex 3 can be isolated and not the analogs with pyrazine or other monocyclic ligands. Presumably, the larger bridging ligands offer more flexibility for the establishment of the eight coordination bonds yielding the assembly. In addition, the presence of two aromatic rings per ligand results in the presence of six π⋯π stacking interactions† (with centroid-to-centroid distances of 3.601 to 3.662 Å) adding stability to the assembly. It is indeed remarkable that the use of the bridging bpy ligand has led in fact to the sought molecular entity rather than to a polymeric system. Indeed, of more than 2000 coordination complexes registered with 4,4′-bipyridine (CCDC, version 5.31, Aug2010), less than 15% corresponds to discrete compounds, of which, many are cyclic molecules featuring bpy as the edges.¹⁴ With respect to this, it is likely that the orange solid that precipitates from the initial reaction mixture corresponds to a polymeric species and the crystals of 3 originate from the species remaining in the solution after this. That would be in line with the low yield observed in the obtention of this compound.

In the crystal, the tetranuclear moieties are arranged as infinite chains, by linking their sides to each other through a network of hydrogen bonds involving their axial ligands and some additional lattice water molecules.† The chains organize as sheets via π⋯π connections involving the rings of L²⁻.

Fig. 2 PovRay representation of the complex cation of [Co₈L₄(OH)₂(H₂O)₄(NO₃)₂(bpy)₄](NO₃)₄ (3). Only unique heteroatoms are labelled. Only hydrogen atoms bound to O are shown. At the bottom is a side view emphasizing the offset superposition of aromatic rings from adjacent bpy ligands. Ranges of selected distances (Å) and angles (°): Co–O, 2.012(4) to 2.259(4); Co–N, 2.084(5) to 2.183(4); Co1–O1–Co2, 108.17(18); Co1–O7–Co2, 109.41(18); Co2–O2–Co3, 109.09(16); Co2–O6–Co3, 108.77(16); Co3–O3–Co4, 109.43(18); Co3–O5–Co4, 107.82(18); Co1⋯Co2, 3.4707(14); Co3⋯Co4, 3.4666(14); Co2⋯Co3, 3.6139(13); Co1⋯Co3, 6.847(2); Co2⋯Co4, 6.8475(19); Co1⋯Co4, 10.236(3); Co1⋯Co4#, 11.358(3).

The magnetic properties of complex 3 were investigated through variable temperature bulk magnetization measurements. The results are shown in Fig. 3 as a χ_MT vs. T curve (χ_M is the paramagnetic molar susceptibility) for the 2–300 K temperature range with an applied magnetic field of 0.5 T. The χ_MT product at 300 K is 18.8 cm³ K mol⁻¹, which shows that the Co(II) ions are high-spin (S = 3/2) subject to only a weak effect of spin–orbit coupling (the spin-only calculated value for g = 2.2 is 18.15 cm³ K mol⁻¹). Contrary to systems with stronger ligand fields (which could be high or low spin), here there is no ambiguity about the spin state. Consequently, the decline of the curve with decreasing temperature to almost zero at T = 2 K is dominated by antiferromagnetic interactions possibly with a small contribution from single ion zero-field splitting (ZFS). The strongest interactions are expected to occur between adjacent metals, therefore, the effects of the exchange between both [Co₄L₂] parts of the assembly through the bpy ligands cannot be detected. In addition, the effect of ZFS is expected to be correlated with that from the antiferromagnetic coupling. Thus, including ZFS in the model, aside from being beyond the scope of this paper, would only slightly modify the values obtained for the magnetic exchange. The experimental data were thus, as an approximation, modeled through diagonalisation of the isotropic Heisenberg spin Hamiltonian in eqn (1).

H = 2 × [−2J₁(S₁S₂ + S₃S₄) − 2J₂S₂S₃] (1)

Fig. 3 χ _M T vs. T curve of [Co₈L₄(OH)₂(H₂O)₄(NO₃)₂(bpy)₄](NO₃)₄ (3). The solid curve is a fit to the experimental data using a spin-only Hamiltonian (see text).

In this equation both tetranuclear moieties in 3 are indeed considered as independent, with S₁, S₂, S₃ and S₄ (all equal to 3/2) being the spin operators for Co1, Co2, Co3 and Co4 (or symmetry equivalents, Fig. 2). The modelling was performed by full matrix diagonalization using the program CLUMAG,¹⁵ which produced the following best fit parameters: g = 2.44, J₁ = −10.87 cm⁻¹ and J₂ = −6.44 cm⁻¹. A larger distortion of the coordination geometry, favored by replacement of some O-donors by N-based ligands is the reason for the decrease of the influence of spin–orbit coupling in 3 as compared with 1.¶ This is reflected in a smaller g value (now much closer to values expected for spin-only Co(II) ions).¹⁶ These changes also have resulted into smaller exchange coupling constants.

With the preparation of 3, we have shown that ligand synthesis and coordination chemistry is a convenient strategy for the design of molecular architectures with the appropriate structure for the exploitation of their functional properties. In this case, a molecular pair of linear clusters has been created. This will serve as the basis for the preparation of related entities exhibiting non-zero spin ground states on each side of the molecular pair, in order to explore their potential as 2qubit quantum gates.

GA thanks the Generalitat de Catalunya for the prize ICREA Academia 2008. The authors thank the Spanish MCI through CTQ2009-06959 (GA, LB and DA). The Advanced Light Source (SJT) is supported by the U.S. Department of Energy under Contract No. DE-AC02-05CH11231.

Notes and references

1. G. Aromí and E. K. Brechin, Struct. Bonding, 2006, 122, 1–67.
2. A. K. Powell, Nat. Chem., 2010, 2, 351–352.
3. M. Affronte, J. Mater. Chem., 2009, 19, 1731–1737.
4. D. Aguilà, L. A. Barrios, F. Luis, A. Repollés, O. Roubeau, S. J. Teat and G. Aromí, Inorg. Chem., 2010, 49, 6784–6786.
5. L. A. Barrios, D. Aguilà, O. Roubeau, P. Gamez, J. Ribas-Ariño, S. J. Teat and G. Aromí, Chem.–Eur. J., 2009, 15, 11235–11243.
6. L. A. Barrios, D. Aguilà, S. Mellat, O. Roubeau, S. J. Teat, P. Gamez and G. Aromí, C. R. Chim., 2008, 11, 1117–1120.
7. E. C. Sañudo, T. Cauchy, E. Ruiz, R. H. Laye, O. Roubeau, S. J. Teat and G. Aromí, Inorg. Chem., 2007, 46, 9045–9047.
8. T. Shiga, M. Ohba and H. Okawa, Inorg. Chem., 2004, 43, 4435–4446.
9. L. A. Barrios, D. Aguilà, O. Roubeau, K. S. Murray and G. Aromí, Aust. J. Chem., 2009, 62, 1130–1136.
10. A. F. Cotton, C. A. Murillo and Q. Wang, Inorg. Chim. Acta, 2010 DOI:10.1016/j.ica.2010.04.045.
11. F. A. Cotton and R. C. Elder, Inorg. Chem., 1965, 4, 1145–1151.
12. R. J. Irving, M. L. Post and R. W. Baker, J. Chem. Soc., Dalton Trans., 1975, 1898–1902.
13. T. Konno, Y. Gotoh and K. Okamoto, Inorg. Chem., 1997, 36, 4992–4997.
14. M. Fujita, O. Sasaki, T. Mitsuhashi, T. Fujita, J. Yazaki, K. Yamaguchi and K. Ogura, Chem. Commun., 1996, 1535–1536.
15. D. Gatteschi and L. Pardi, Gazz. Chim. Ital., 1993, 123, 231.
16. F. Lloret, M. Julve, J. Cano, R. Ruiz-Garcia and E. Pardo, Inorg. Chim. Acta, 2008, 361, 3432–3445.
17. G. M. Sheldrick, Acta Crystallogr., Sect. A: Found. Crystallogr., 2008, 64, 112–122.

---

Footnotes

† Electronic supplementary information (ESI) available: Syntheses, microanalysis, crystallographic data and figures. CCDC 791562 and 791563. For ESI and crystallographic data in CIF or other electronic format see DOI: 10.1039/c0cc03646c

‡ Crystal data for 2: [C₄₄H₄₆Co₄N₆O₁₆](NO₃)₄·4(CH₄O)·2(H₂O). Data were collected on a red block (0.04 × 0.04 × 0.04 mm), triclinic, space groupP1̄ (no. 2) with a = 10.5822(7), b = 12.3023(8), c = 13.8898(10) Å, α = 64.4450(10)°, β = 89.4460(10)°, γ = 72.1670(10)°, V = 1537.11(18) ų, Z = 1, ρ_calcd = 1.688 g cm⁻³, μ = 1.467 mm⁻¹. 24 109 reflections were measured, 9091 of which were independent (R_int = 0.0267, R = 0.0302), on station 11.3.1 of the Advanced Light Source synchrotron facility (T = 100 K, 3.05° < θ < 33.57°, λ = 0.7749 Å). Refinement converged at final wR2 = 0.0937, R1 = 0.0342 and S = 1.019 (for 8030 reflections with I > 2σ(I)). Crystal data for 3: [C₁₂₄H₉₄Co₈N₂₂O₂₈](NO₃)₄·2(C₇H₈)·6(H₂O). Data were collected on a red plate (0.18 × 0.15 × 0.02 mm), triclinic, space groupP1̄ (no. 2) with a = 12.775(3), b = 16.956(3), c = 18.838(4) Å, α = 65.59(3)°, β = 70.26(3)°, γ = 87.57(3)°, V = 3474.9(17) ų, Z = 1, ρ_calcd = 1.602 g cm⁻³, μ = 1.190 mm⁻¹. 55 394 reflections were measured, 13 157 of which were independent (R = 0.0456), on station BM16 of the ESRF synchrotron facility (T = 150 K, 1.83° < θ < 28.03°, λ = 0.7515 Å). Refinement converged at final wR2 = 0.2113, R1 = 0.0792 and S = 1.040 (for 9567 reflections with I > 2σ(I)). Both structures were solved by direct methods and refined on F² with SHELXTL suite.¹⁷

§ Synthesis of 3: a grey suspension of H₂L (45 mg, 0.12 mmol) and 4,4′-bipyridine (37.5 mg, 0.24 mmol) in methanol (10 ml) was added to a pink solution of Co(AcO)₂·4H₂O (30 mg, 0.12 mmol) and Co(NO₃)₂·6H₂O (35 mg, 0.12 mmol) in methanol (15 ml). The mixture was brought to reflux for 45 minutes and then cooled down to room temperature. An orange solid was removed by filtration and the red solution was layered with toluene. After one month, red crystals of 3 were obtained in 12% yield. Anal. Calcd (found) for 3·12H₂O: C 45.46 (45.96), H 3.63 (3.32), N 11.12 (10.62). IR (KBr pellet, cm⁻¹): 1608 m, 1578 s, 1563 s, 1545 s, 1517 m, 1457 s, 1384 s, 1318 m, 1294 m, 1257 w, 1224 w, 1170 w, 1126 w, 1077 w, 1049 w, 1017 w, 944 w, 868 w, 810 w, 777 w, 748 w, 709 w, 684 w, 657 w, 634 w, 517 w.

¶ Parameters previously obtained for complex 1: g = 2.57, J₁ = J₂ = −12.2 cm⁻¹.

---