Crystal engineering of zinc(II) and copper(II) complexes containing 3,5-dimethylisoxazole-4-carboxylate ligand via O–H⋯N, C–H⋯A (A = N, O and π) and bifurcated C–H⋯N/O interactions

Abstract

Supramolecular frameworks of Zn₃(DMIC)₆L₂ [L = H₂O (1), CH₃CH₂OH (2) or 1/2 4,4′-bipyridine (3)], Cu(DMIC)₂(4,4′-bipyridine) (4) and Cu(DMIC)₂ (5) (DMIC = 3,5-dimethylisoxazole-4-carboxylate), are self-assembled by hydro- or solvothermal methods and characterized structurally, in which a rich family of hydrogen bonds have been observed including O–H⋯N (strong), C–H⋯A (A = N, O and π) (weak and very weak) and unusual bifurcated C–H⋯N/O.

---

With tremendous success in the preparation of diverse and specifically customized materials by crystal engineering,^1–3 significant efforts have focused on the development of related methods for the efficient and controllable assembly of metal complexes. It is apparent that the assembly based on hydrogen bond interactions is a versatile route to organically modified structures due to their strength and directionality.^1,2 The topology and function of materials could be rationally controlled if the hydrogen bonding pattern bestowed by solvent molecules could be deployed in an effective assembly for metal-containing discrete units or polymers. Noticeably, although a variety of frameworks built up from metal complexes by strong hydrogen bonds D–H⋯A (D, A = O, N, etc.), and weak ones C–H⋯A (A = O, N, etc.) have been reported,^2,4 there are fewer examples about very weak C–H⋯π interactions.⁵ Therefore, it is a great challenge for the same ligand system to make a full screen of various hydrogen bonds changing from strong (O–H⋯N, etc.), through weak (C–H⋯O), to very weak (C–H⋯π).⁶ Recently, it was found that C–H can form unusual bifurcated C–H⋯O1/O2 hydrogen bonds with –NO₂ and C₆H₄(OMe)₂-1,2 in an organic system,^6b,c,7 raising the intriguing question of whether the corresponding bifurcated C–H⋯N/O hydrogen bond is feasible.

With the aim of establishing whether the presence of the isoxazole ring imparted to carboxylate complexes an enhanced hydrogen bonding pattern analogous to that observed in the case of 4,5-imidazoledicarboxylate,⁸ we are interested to explore the coordination chemistry of DMIC ligand and crystal engineering based on the related metal complexes owing to its following characteristics: (i) the N and O atoms of the isoxazole ring have large potential as hydrogen bond acceptors or as coordination sites to construct polymers; (ii) the deprotonated DMIC may facilitate access to observing weak C–H⋯A contacts (A = N, O or π) due to the lack of strong donors except C–H group; (iii) the adjacent arrangement of N and O could lead to unusual bifurcated C–H⋯A1/A2 interactions. However, to our knowledge, no example of frameworks based on this ligand has been reported so far. We herein report the syntheses and crystal structures of a series of zinc and copper frameworks with the DMIC ligand, representing a good ligand system that can map out the change of hydrogen bonds from O–H⋯N (strong), through C–H⋯A (A = N, O) and/or C–H⋯O/N (weak) to C–H⋯π (very weak).

Hydro- or solvothermal treatments of DMIC acid and zinc salts with the absence or presence of 4,4′-bipyridine, produced complexes 1–3.†Fig. 1 (left) shows the crystal structure of the linear trinuclear Zn₃(DMIC)₆(H₂O)₂ of complex 1.‡ This molecule contains three zinc, six DMIC ligands, and two terminal waters. Three zinc atoms, with an inversion center at Zn1, are coordinated into two types of environments, in which the center Zn1 is bonded into octahedral ZnO₆ by uniform carboxylate oxygens while the Zn2 and Zn2A into tetrahedral ZnO₄ by three carboxylate oxygens and one water oxygen. Between the Zn1 and Zn2(Zn2A) there are three individual DMIC ligands acting as a bidentate bridge in syn–syn mode. The bond distances of Zn1–O are between 2.069–2.090 Å. The Zn2–O(carboxylate) distances are between 1.931 and 1.941 Å but 1.998 Å for water oxygen. The Zn–O distances are comparable with those of reported similar trinuclear cores.⁹ The linear trinuclear units are reticulated into one 2-D supramolecular (4, 4) net by four pairs of O10–H10A⋯N1 and O10–H10B⋯N2 in distances of 2.824 and 2.829 Å (O⋯N), as shown in Fig. 1 (right).

Fig. 1 The crystal structure (left) of 1 is shown as the representative for the trinuclear unit Zn₃(DMIC)₆ and (4, 4) net of 1 constructed by hydrogen bonds (right). Selected bond distances (Å) and angles (°) for 1 (2/3): Zn1–O1, 2.090(2) (2.117(6)/2.113(4)); Zn1–O4, 2.070(2) (2.030(8)/2.019(4)); Zn1–O7, 2.078(2) (2.054(8)/2.071(4)); Zn2–O2, 1.940(2) (1.936(6)/1.912(3)); Zn2–O5, 1.942(2) (1.920(7)/1.925(4)); Zn2–O8, 1.931 (2) (1.948(8)/1.912(4)); Zn2–O10, 1.998(2) (2.018(6)); Zn2–N4 2.043(4) 3. Hydrogen atoms (a) are omitted for clarity.

Interestingly, when the coordinated water molecule is replaced by ethanol, the resulting complex 2 Zn₃(DMIC)₆(C₂H₅OH)₂ shows a different supramolecular (4, 4) net formed by four hydrogen contacts between the terminated ethanol oxygen and N1 distanced in 2.755 Å of O⋯N (Fig. 2), although the structure of trinuclear unit Zn₃(DMIC)₆ in 2 is similar to that in 1 (Fig. 1, left). This difference may be attributed to the less O–H group of ethanol compared with H₂O.

Fig. 2 The (4, 4) network in 2 constructed by hydrogen bonds formed between the hydroxyl group of ethanol and N atoms. Color representatives: green, Zn; blue, N; gray, C; red, O.

To explore the factor that influences the pattern of hydrogen bonds, Zn₃(DMIC)₆(C₁₀H₈N₂) (3) was synthesized. In contrast to 1 and 2, the introduction of 4,4′-bipyridine coligand leads to a significant change in the hydrogen bond pattern due to the absence of a strong O–H donor (Fig. 3). The 1-D coordination polymer chain in 3 consists of a trinuclear Zn₃(DMIC)₆ unit and a bidentate 4,4′-bipyridine linker. In the trinuclear unit, the center Zn1 is bonded into octahedral ZnO₆ while the Zn2 and Zn2A into different tetrahedral ZnO₃N instead of ZnO₄. All the Zn–O(carboxylate) bond distances (Zn1–O, 2.019–2.113 Å and Zn2–O 1.912–1.925 Å) are comparable to the corresponding values in 1 and 2. The Zn2–N distance of 2.043 Å is comparable to those found in other trinuclear cores.⁹ All the carboxylate groups in each of the present zinc complexes are bonded to Zn in a uniform syn–syn bridging mode, which is different from the observation that there are two carboxylates in the monodentate bridging mode in Zn₃(benz)₆(nia)₂.^9a This difference might result from the nature of ligands. As shown in Fig. 3a, the polymeric chains are interconnected into a 2-D network by weak hydrogen bonds of C4–H4B⋯N2 and C4–H4C⋯N3 (D⋯A 3.485 and 3.472 Å, respectively). Then, the resulting 2-D nets are further interlinked into 3-D network by bifurcated hydrogen bonds C20–H⋯N/O between pyridine C–H and the adjacent N and O on the isoxazole ring with distances of 3.091 Å for C⋯N1 and 3.285 Å for C⋯O3 (Fig. 3b).^6a,b,d The H⋯O3 (2.401 Å) distance of the bifurcated C–H⋯A1/A2 interaction is significantly shorter than the reported values in 1,3-dibromo-2,4,6-trinitrobenzene (H⋯O 2.60 Å, 155.2°).^6b,c Although metal complexes containing heterocycle ligands have been investigated extensively, to our knowledge, the asymmetric bifurcated C–H⋯N/O hydrogen bond is unprecedented.

Fig. 3 The supramolecular structures in complex 3: (a) the network constructed by hydrogen bonds C4–H4B⋯N2 (D⋯A 3.485 Å) and C4–H4C⋯N3 (D⋯A 3.472 Å); (b) the bifurcated hydrogen bonds in the third direction C20–H20A⋯N1 (D⋯A 3.091 Å) and C20–H20A⋯O3 (D⋯A 3.285 Å).

Encouraged by the intetresting change of hydrogen bonds, we examine further the influences of the nature of metal ions on the structures and hydrogen bonding patterns of DMIC complexes. For this target two new copper complexes Cu(DMIC)₂(4,4′-bipyridine) 4 and Cu(DMIC)₂5 were prepared under the same conditions.† As shown in Fig. 4a, the Cu(II) ions in 4 are trans-tetracoordinated into CuN₂O₂ in square planar geometry by two O from two individual carboxylate groups and two N of two 4,4′-bipyridine to form a 1-D polymer chain. Interestingly, the polymeric chains are joined into 2-D network by C5–H⋯π (3.680 Å C⋯π) interaction^6a,c between the methyl group and the isoxazole ring. Consequently, the 2-D nets are further stabilized by another C5–H⋯π (3.687 Å C⋯π) interaction between the methyl group and the pyridine ring (Fig. 4b). Despite the fact that C–H⋯π interactions are ubiquitous in biologic, organic and organometallics systems and play an important role in molecular recognition and reactivity control,⁶ they are observed in limited capacity in inorganic complexes.

Fig. 4 The structure of 1-D chain in 4 and the packing motif of two chains. Selected bond distances (Å) and angles (°): Cu1–O1, 1.898(3); Cu1–N1, 2.014(3). Color representatives: pale blue, Cu; blue, N; gray, C; red, O; green, H.

As shown in Fig. 5, a 2-D (4, 4) net in 5 is constructed from a 4-connected node Cu₂(COO)₄ shaped in a quadruple paddle-wheel. The Cu(II) ions in the unit are coordinated into a square pyramid CuNO₄ with four O atoms from four different carboxylate groups and one N from the fifth ligand of another unit. The ligands behave in two types of coordination modes, namely, one as bidentate bridge in syn–syn mode to clip the dinuclear unit, while the other as a tridentate linker not only to clip the dinuclear unit, but further to link another copper from an adjacent dinuclear unit by the N atom. Then the coordination 2-D nets are interconnected into 3-D framework by C–H⋯π interaction (3.741 Å C6⋯π) between DMIC ligands that were oriented in perpendicular directions to the (4, 4) net.

Fig. 5 (4, 4) Network in 5 constructed by 4-connected paddle-wheel dinuclear units Cu₂(COO)₄. Selected bond distances (Å) and angles (°): Cu1–O1, 1.992(3); Cu1–O2A, 1.985(3); Cu1–O3B, 1.978(3); Cu1–O4C, 1.976(3); Cu1–N2, 2.202(4). Color representatives: pale blue, Cu; blue, N; gray, C; red, O; green, H.

For all the zinc complexes, the trinuclear Zn₃(DMIC)₆ unit is not affected by co-ligand and coordinated solvents, but the change of hydrogen bonding patterns can be driven by controlling the nature and number of hydrogen donors. DMIC could be expected to allow the formation of weak hydrogen bonding interactions in the absence of strong donors, giving a succeeding along with the adjacent arrangement of N and O to drive the formation of unusual bifurcated hydrogen bonds. In contrast to zinc complexes, the coordination modes of DMIC to the copper ion depend on the co-ligand which may be attributed to the stronger nitrophilic characteristics of the copper ion. Furthermore, the formation of unusual C–H⋯π interactions between the methyl and the isoxazole/pyridine rings in 4 and 5 may result from their favorable orientation to each other.

In conclusion, a series of zinc and copper complexes involving DMIC ligand have been prepared by self-assembly and structurally characterized. The results indicate that the structure of the Zn₃(DMIC)₆ unit is independent of co-ligands in the zinc system, whereas the bonding mode of DMIC to metal in the copper series varies as a function of the co-ligand. Significantly, these results represent a good example that the change of hydrogen bonds from strong to weak can be finely tuned simply by changing the nature of solvents, co-ligands and metals.

This work is supported by the NNSF of China, NSF of Shanghai, the Research Funds for the New Century Distinguished Scientist and the Doctoral Program of National Education Ministry of China.

Notes and references

1. (a) J. M. Lehn, Supramol. Chem., VCH, Weinheim, 1995; (b) K. Biradha, CrystEngComm, 2003, 5, 374; (c) B. Moulton and M. J. Zaworotko, Chem. Rev., 2001, 101, 1629; (d) M. J. Zaworotko, Chem. Commun., 2001, 1; (e) S. Furukawa, H. Uji-i, K. Tahara, T. Ichikawa, M. Sonoda, F. C. D. Schryver, Y. Tobe and S. D. Feyter, J. Am. Chem. Soc., 2006, 128, 3502; (f) O. R. Evans and W. B. Lin, Acc. Chem. Res., 2002, 35, 511.
2. (a) G. R. Desiraju, Angew. Chem., Int. Ed. Engl., 1995, 34, 2311; (b) X. H. Bu, M. L. Tong, J. R. Li, H. C. Chang, L. J. Li and S. Kitagawa, CrystEngComm, 2005, 7, 411.
3. (a) J. D. Wuest, Chem. Commun., 2005, 5830 and references therein.
4. (a) J. H. He, J. H. Yu, Y. T. Zhang, Q. H. Pan and R. R. Xu, Inorg. Chem., 2005, 44, 9279; (b) M. Lee, M. H. Park, N. K. Oh, W. C. Zin, H. T. Jung and D. K. Yoon, Angew. Chem., Int. Ed., 2004, 43, 6466; (c) B. F. Hoskins and R. Robson, J. Am. Chem. Soc., 1989, 111, 5962; (d) H. W. Roesky and M. Andruh, Coord. Chem. Rev., 2003, 236, 91; (e) A. M. Beatty, Coord. Chem. Rev., 2003, 246, 131; (f) M. Tadokoro and K. Nakasuji, Coord. Chem. Rev., 2000, 198, 205; (g) S. Kawata, K. Adachi, Y. Sugiyama, M. K. Kabirb and S. Kaizaki, CrystEngComm, 2002, 4, 496.
5. (a) B. Manimaran, L. J. Lai, P. Thanasekaran, J. Y. Wu, R. T. Liao, T. W. Tseng, Y. H. Liu, G. H. Lee, S. M. Peng and K. L. Lu, Inorg. Chem., 2006, 45, 8070; (b) G. A. Bogdanovic, A. S. Bire and S. D. Zaric, Eur. J. Inorg. Chem., 2002, 1599; (c) D. Braga and F. Grepioni, Acc. Chem. Res., 2000, 33, 601; (d) J. C. Ma and D. A. Dougherty, Chem. Rev., 1997, 97, 1303.
6. (a) G. R. Desiraju and T. Steiner, The Weak Hydrogen Bond in Structural Chemistry and Biology, Oxford University Press, Oxford, 1999; (b) P. K. Thallapally, G. R. Desiraju, M. Bagieu-Beucher, R. Masse, C. Bourgogne and J. F. Nicoud, Chem. Commun., 2002, 1052; (c) T. Steiner and G. Koellner, J. Mol. Biol., 2001, 305, 535 and references therein; (d) G. R. Desiraju, Chem. Commun., 2005, 2995; (e) M. Nishio, CrystEngComm, 2004, 6, 130.
7. M. J. Hardie and C. L. Raston, Chem. Commun., 1999, 1153.
8. C. F. Wang, E. Q. Gao, H. Zheng and C. H. Yan, Chem. Commun., 2004, 720.
9. (a) V. Zelenak, M. Sabo, W. Massa and P. Llewellyn, Inorg. Chim. Acta, 2004, 357, 2049; (b) Q. R. Fang, X. Shi, G. Wu, G. Tian, G. s. Zhu, R. W. Wang and S. L. Qiu, J. Solid State Chem., 2003, 176, 1.

---

Footnotes

† Preparation and analytical data: To a mixture of DMIC acid (0.056 g, 0.4 mmol) and Zn(O₂CCH₃)₂·2H₂O (0.044 g, 0.2 mmol) in a 25 mL Teflon-lined bomb, 8 mL H₂O was added. Then the bomb was sealed and kept at 120 °C for 4 d. The final pH value of the reaction mixture was 4. 1 as colorless crystals crystallized from the filtrate, yields in 39%. Similar to 1, the colorless crystal 2 was directly obtained in the bomb after 7 d reaction of just ethanol instead of water, 56% yield. For 3, the reaction mixture of DMIC acid, Zn(O₂CCH₃)₂·2H₂O and 4,4′-bipy in ratio of 6 : 3 : 1 in ethanol was kept at 120 °C for 10 d, pyramid crystals of 3 were obtained, yield 79%. A Mixture of DMIC and Cu(NO₃)₂·3H₂O and 4,4′-bipy in molar ratio of 2 : 1 : 1 in water was kept at 140 °C for 7 d producing block bluish purple crystals of 4, 36% yield. Using the same method for preparation of 4, the mixture of DMIC and Cu(ClO₄)₂·6H₂O in a molar ratio of 2 : 1 in 10 mL ethanol produced blue square plate crystals of 5 (50% yield). Elemental Anal. for 1: Calcd: C, 40.30; H, 3.76; N, 7.83%. Found: C, 40.28; H, 3.638; N, 7.699%. For 2 Calcd: C, 42.55; H, 4.29; N, 7.44%. Found: C, 42.08; H, 4.004; N, 7.053%. For 3 Calcd: C, 46.31; H, 3.72; N, 9.39%. Found: C, 46.87; H, 3.499; N, 9.062%. For 4 Calcd: C, 52.85; H, 4.03; N, 11.21%. Found: C, 52.87; H, 3.90; N, 11.32%. For 5 Calcd: C, 41.93; H, 3.52; N, 8.15%. Found: C, 42.38; H, 3.650; N, 8.27%. IR adsorptions (cm⁻¹) of carboxylate group (ν_as and ν_s): 1 1652 and 1455; 2 1624 and 1400; 3 1617 and 1398; 4 1627 and 1381; 5 1636 and 1430.

‡ Crystal data for1: C₃₆H₄₀N₆O₂₀Zn₃, M_r = 1072.85, triclinic, space group P-1, a = 10.122(7) Å, b = 10.124(7) Å, c = 11.446(8) Å, α = 88.961(9)°, β = 69.655(9)°, γ = 89.692(9)°, U = 1099.5(14) ų, T = 293 K, Z = 1, ρ_calcd = 1.620 Mg m⁻³, μ(Mo Kα) = 1.709 mm⁻¹, S = 1.050, 4591 reflections measured, 3790 unique (R_int = 0.0225), R₁ = 0.0297, wR₂ = 0.0864 (all data). 2: C₄₀H₄₈N₆O₂₀Zn₃, M_r = 1128.95, monoclinic, space group P2₁/n, a = 11.500(5) Å, b = 17.654(8) Å, c = 12.048(5) Å, β = 95.909(10)°, U = 2432.9(18) ų, T = 293 K, Z = 2, ρ_calcd = 1.541 Mg m⁻³, μ(Mo-Kα) = 1.549 mm⁻¹, S = 0.842, 5370 reflections measured, 3572 unique (R_int = 0.0443), R₁= 0.0794, wR₂ = 0.2013 (all data). 3: C₄₆H₄₄N₈O₁₈Zn₃, M_r = 1193, monoclinic, space group P2₁/n, a = 11.377(7) Å, b = 20.360(12) Å, c = 11.665(7) Å, β = 103.981(8)°, U = 2622(3) ų, T = 293 K, Z = 2, ρ_calcd = 1.511 Mg m⁻³, μ(Mo Kα) = 1.440 mm⁻¹, S = 0.904, 10796 reflections measured, 4603 unique (R_int = 0.0483), R₁= 0.0548, wR₂ = 0.1454 (all data). 4: C₁₁H₁₀N₂O₃Cu_0.5, M_r = 249.98, ticlinic, space group P-1, a = 5.738(3) Å, b = 9.471(5) Å, c = 10.852(7) Å, α = 64.246(7)°, β = 80.464(12)°, γ = 89.857(9)°, U = 522.2(5) ų, T = 293 K, Z = 2, ρ_calcd = 1.590 Mg m⁻³, μ(Mo Kα) = 1.095 mm⁻¹, S = 0.777, 2213 reflections measured, 1819 unique (R_int = 0.0320), R₁= 0.0508, wR₂ = 0.0932 (all data). 5: C₁₂H₁₂N₂O₆Cu, M_r = 343.78, monoclinic, space group P2₁/n, a = 10.028(4) Å, b = 12.693(5) Å, c = 11.881(5) Å, β = 97.236(6)°, U = 1500.4(11) ų, T = 293 K, Z = 4, ρ_calcd = 1.522 Mg m⁻³, μ(Mo Kα) = 1.482 mm⁻¹, S = 0.642, 5993 reflections measured, 2630 unique (R_int = 0.0682), R₁= 0.0405, wR₂ = 0.0853 (all data). CCDC reference numbers 624328–624332. For crystallographic data in CIF or other electronic format see DOI: 10.1039/b615091h

---