“Size-selectivity” in the template-directed assembly of dinuclear triple-stranded helicates

Abstract

The self-assembly of supramolecular structures depends on a subtle interplay of a series of different control mechanisms. The geometric as well as electronic complementarity of the molecular building blocks is crucial for the specific formation of defined supramolecular species. In addition, secondary effects, like templating, also have an important function. The templating ability of different cations in the formation of triple-stranded helicate-type complexes from alkyl-bridged di(8-hydroxyquinoline) ligands is investigated by introduction of alkyl chains of different length as ligand spacers. Hereby a “size-selectivity” between the cations and the dinuclear helicate-type complexes {(ligand)₃M₂} is observed. Large cations support the formation of big dinuclear complexes, whereas small cations are able to template the formation of small complexes.

To design and synthesize artificial molecular machines from molecular building blocks we need a deeper understanding of fundamental processes of the self-assembly of supramolecular structures. Stereochemical and regiochemical aspects are of importance as well as the specificity of the formation of supramolecular aggregates out of a mixture of molecular components (1, 2).

We are interested in the chemistry of helicates and of related complexes because those simple supramolecular coordination compounds allow the detailed study of various mechanistic aspects of self-assembly processes (3, 4). Hereby we found a template control in the formation of triple-stranded helicates (or meso-helicates; ref. 5), which can be discussed as a dynamic combinatorial process (6). First a mixture (i.e., library) of metallo-supramolecular species is formed (i.e., generation of molecular diversity), which in the presence of an appropriate template is transformed into only one receptor/substrate complex (i.e., selection; refs. 7, 8, and 9–13).

In this paper we discuss the template-directed self-assembly of dinuclear triple-stranded helicate-type complexes from di(8-hydroxyquinoline) derivatives 1-4-H₂ (Fig. 1) and show a “size-selectivity” between the templating alkali metal cations and the dinuclear coordination compounds.

Figure 1.

8-Hydroxyquinoline derivatives 1–4-H₂.

Materials and Methods

Instrumental Techniques.

¹H NMR and ¹³C NMR spectra were recorded on a Bruker (Billerica, MA) DRX 500 or AM 400 spectrometer by using DEPT techniques for the assignment of the multiplicity of carbon atoms. Fourier transform (FT)-IR spectra were recorded by diffuse reflection (KBr) on a Bruker IFS spectrometer. Mass spectra [EI, 70 eV (1 eV = 1.602 × 10⁻¹⁹ J) or fast atom bombardment (FAB)-(+/−), 3-NBA as matrix] were taken on a Finnigan (San Jose, CA) MAT 90 mass spectrometer. Molar peaks are often observed in low intensity; however, no signals of other species are observed at high masses. Elemental analyses were obtained with a Heraeus CHN-O-Rapid analyzer. Solvents were purified by standard methods.

Synthesis.

The alkyl-bridged di(8-hydroxyquinoline) derivatives 1-4-H₂ were prepared according to literature procedures (14–16).

[Ga₂(1)₃] ⋅ x[M′Cl] (M′ = Li, Na).

Di(3-n-decyl-8-hydroxyquinolin-7-yl)methane (1-H₂, 15 mg, 0.03 mmol) and Ga(NO₃)₃ ⋅ H₂O (7.1 mg, 0.02 mmol) are refluxed for 15 h in 15 ml of a methanol/dichloromethane mixture (1:1). The precipitate is collected, washed with water, and dried in vacuum. An excess of LiCl or NaCl, respectively, is added and the mixture is heated in chloroform/methanol/water (5:1:0.1) for 15 h to 35°C. Solvent is removed and the product is isolated by extraction with chloroform.

[Ga₂(1)₃] ⋅ x[LiCl].

Yield: 12 mg [Ga₂(1)₃] ⋅ x[LiCl] as a yellow-green solid. ¹H NMR (CDCl₃): δ = 8.19 (s, 6H), 7.79 (d, J = 8.4 Hz, 6H), 7.58 (s, 6H), 7.15 (d, J = 8.4 Hz, 6H), 5.04 (d, J = 12.4 Hz, 3H), 3.21 (d, J = 12.4 Hz, 3H), 2.69 (m, 12H), 1.63 (m, 12H), 1.23 (m, 84H), 0.87 (m, 18H). ¹³C NMR (CDCl₃): δ = 154.4 (C), 143.9 (CH), 139.5 (CH), 136.5 (C), 135.2 (C), 133.1 (CH), 128.9 (C), 126.2 (C), 112.3 (CH), 32.8 (CH₂), 31.9 (CH₂), 30.8 (CH₂), 29.7 (2 CH₂), 29.6 (CH₂), 29.5 (CH₂), 29.4 (CH₂), 28.9 (CH₂), 22.7 (CH₂), 14.2 (CH₃). FAB(+)-MS: m/z (%) = 1887 (70, [{Ga₂(1)₃}Li]⁺). IR (KBr): ν̃ = 3414, 2925, 2853, 1583, 1490, 1466, 1379, 1260, 1112, 758, 724 cm⁻¹. UV-vis (CHCl₃): λ = 226, 274, 398 nm.

[Ga₂(1)₃] ⋅ x[NaCl].

Yield: 14 mg of [Ga₂(1)₃] ⋅ x[NaCl] as a yellow-green solid. ¹H NMR (CDCl₃): δ = 8.19 (s, 6H), 7.80 (d, J = 8.4 Hz, 6H), 7.60 (s, 6H), 7.15 (d, J = 8.4 Hz, 6H), 5.04 (d, J = 12.4 Hz, 3H), 3.21 (d, J = 12.4 Hz, 3H), 2.70 (m, 12H), 1.70 (m, 12H), 1.24 (m, 84H), 0.87 (m, 18H). ¹³C NMR (CDCl₃): δ = 154.5 (C), 143.9 (CH), 139.5 (CH), 136.5 (C), 135.2 (C), 133.1 (CH), 128.9 (C), 126.2 (C), 112.3 (CH), 32.7 (CH₂), 31.9 (CH₂), 30.8 (CH₂), 30.1 (CH₂), 29.7 (CH₂), 29.6 (CH₂), 29.5 (CH₂), 29.4 (CH₂), 28.9 (CH₂), 22.7 (CH₂), 14.2 (CH₃). FAB(+)-MS: m/z (%) = 1903 (100, [{Ga₂(1)₃}Na]⁺). IR (KBr): ν̃ = 3311, 2925, 2922, 2853, 1528, 1466, 1378, 1349, 1113, 758, 731 cm⁻¹. UV-vis (CHCl₃): λ = 227, 274, 400 nm.

[M′⊂(Ga₂(3)₃)]Cl (M′ = K, NH₄, Rb, Cs).

Ligand 3-H₂ (25 mg, 0.07 mmol) together with Ga(NO₃)₃ (11.2 mg, 0.05 mmol) in 10 ml of methanol is refluxed for 15 h in the presence of 50 mg of M′Cl (M′ = K, NH₄, Rb, Cs). The precipitate is collected, washed with cold water, and dried in vacuum.

[K⊂(Ga₂(3)₃)]Cl.

For characterization see ref. 17.

[NH₄⊂(Ga₂(3)₃)]Cl.

Yield: 18.0 mg as a yellow solid. ¹H NMR (DMSO-d₆): δ = 8.65 (d, J = 8.2 Hz, 6H), 7.54 (dd, J = 4.7, 8.2 Hz, 6H), 7.46 (d, J = 8.4 Hz, 6H), 7.37 (d, J = 4.7 Hz, 6H), 7.28 (d, J = 8.4 Hz, 6H), 6.80 (br s, 4H), 4.18 (s, 6H), 3.60 (d, J = 16.7 Hz, 6H), 3.17 (d, J = 16.7 Hz, 6H). ¹³C NMR (DMSO-d₆): δ = 154.8 (C), 148.2 (C), 143.7 (CH), 140.9 (CH), 136.0 (C), 133.2 (CH), 128.4 (C), 122.8 (C), 121.7 (CH), 112.2 (CH), 107.0 (CH₂), 37.5 (CH₂). FAB(+)-MS: m/z (%) = 1178 (30, [{Ga₂(3)₃}(NH₄)]⁺). IR (KBr): ν̃ = 3055, 2889, 1504, 1461, 1377, 1114, 826, 738, 671 cm⁻¹. UV-vis (DMSO): λ = 265, 335, 396 nm.

[Rb⊂(Ga₂(3)₃)]Cl.

Yield: 26.0 mg (0.01 mmol, 75%) as a yellow solid. ¹H NMR (DMSO-d₆): δ = 8.65 (dd, J = 0.9, 8.3 Hz, 6H), 7.54 (dd, J = 4.4, 8.3 Hz, 6H), 7.46 (d, J = 8.4 Hz, 6H), 7.38 (dd, J = 0.9, 4.4 Hz, 6H), 7.27 (d, J = 8.4 Hz, 6H), 4.13 (s, 6H), 3.62 (d, J = 16.0 Hz, 6H), 3.34 (d, J = 16.0 Hz, 6H). ¹³C NMR (DMSO-d₆): δ = 155.1 (C), 148.7 (C), 143.7 (CH), 140.9 (CH), 136.0 (C), 133.3 (CH), 128.5 (C), 122.6 (C), 121.7 (CH), 112.1 (CH), 106.6 (CH₂), 37.7 (CH₂). FAB(+)-MS: m/z (%) = 1245 (35, [{Ga₂(3)₃}Rb]⁺). Elemental analysis calcd for Ga₂C₆₆H₄₈N₆O₆ ⋅ RbCl ⋅ 8H₂O: C 55.61, H 4.53, N 5.89; found: C 55.40, H 4.35, N 6.25. IR (KBr): ν̃ = 3380, 2889, 1580, 1504, 1462, 1377, 1318, 1115, 826, 739, 671 cm⁻¹. UV-vis (DMSO): λ = 265, 319, 333, 392 nm.

[Cs⊂(Ga₂(3)₃)]Cl.

Yield: 26.0 mg (0.01 mmol, 72%) as a green solid. For the ¹H NMR spectrum see ref. 17. ¹³C NMR (DMSO-d₆): δ = 155.1 (C), 150.6 (C), 143.9 (CH), 140.9 (CH), 136.0 (C), 133.4 (CH), 128.5 (C), 122.1 (C), 121.8 (CH), 112.0 (CH), 108.1 (CH₂), 37.4 (CH₂). FAB(+)-MS: m/z (%) = 1293 (19, [{Ga₂(3)₃}Cs]⁺). Elemental analysis calcd for Ga₂C₆₆H₄₈N₆O₆ ⋅ CsCl ⋅ 9H₂O: C 53.16, H 4.46, N 5.64; found: C 53.17, H 4.17, N 6.08. IR (KBr): ν̃ = 3385, 2886, 1580, 1504, 1461, 1377, 1318, 1115, 825, 739, 671 cm⁻¹. UV-vis (DMSO): λ = 266, 319, 334, 392 nm.

[Cs⊂(Ga₂(4)₃)]Cl.

Ligand 4-H₂ (15 mg, 0.04 mmol) and Ga(NO₃)₃ ⋅ H₂O (12.1 mg, 0.03 mmol) in the presence of 100 mg of CsCl are refluxed in methanol for 15 h. The precipitate is filtered off, washed with cold water, and dried in vacuum.

Yield: 14 mg (0.01 mmol, 71%) of a yellow solid. ¹H NMR (DMSO-d₆): δ = 8.81 (d, J = 4.3 Hz, 6H), 8.28 (d, J = 8.2, 6H), 7.49 (dd, J = 4.3, 8.2 Hz, 6H), 7.39 (d, J = 8.4 Hz, 6H), 7.32 (d, J = 8.4 Hz, 6H), 1.50 (m, 12H). Further signals are hidden under the peak of the solvent. FAB(+)-MS: m/z (%) = 1299 (1, [{Ga₂(4)₃}Cs]⁺). Elemental analysis calcd for Ga₂C₆₆H₅₄N₆O₆ ⋅ CsCl ⋅ H₂O: C 58.59, H 4.17, N 6.21; found: C 58.66, H 4.82, N 5.86. IR (KBr): ν̃ = 3332, 2929, 1503, 1460, 1399, 1376, 1283, 1117, 832, 824, 674 cm⁻¹.

Preparation of Aluminum(III), Iron(III), and Chromium(III) Complexes.

The complexes are prepared according to the procedure described for the corresponding gallium complexes, using AlCl₃ or FeCl₃ ⋅ 6H₂O in methanol/water (9:1)

[Na⊂(Al₂(2)₃)]Cl.

Yield: 12.0 mg (0.01 mmol, 68%) as a yellow solid. ¹H NMR (DMSO-d₆): δ = 8.63 (d, J = 8.3 Hz, 6H), 7.66 (d, J = 8.3 Hz, 6H), 7.50 (dd, J = 4.5, 8.3 Hz, 6H), 7.32 (d, J = 8.3 Hz, 6H), 7.15 (d, J = 4.5 Hz, 6H). The signal of the spacer is hidden under the solvent peak. FAB(+)-MS: m/z (%) = 1019 (30, [{Al₂(2)₃}Na]⁺). Elemental analysis calcd for Al₂C₆₀H₄₂N₆O₆ ⋅ NaCl ⋅ 5H₂O: C 62.91, H 4.58, N 7.34; found: C 62.86, H 5.29, N 7.10. IR (KBr): ν̃ = 3529, 3057, 2860, 1503, 1466, 1380, 773, 683 cm⁻¹. UV-vis (DMSO): λ = 265, 418 nm.

[K⊂(Al₂(2)₃)]Cl.

Yield: 9.0 mg (0.01 mmol, 51%) as a yellow solid. ¹H NMR (DMSO-d₆): δ = 8.65 (dd, J = 8.5 Hz, 6H), 7.75 (d, J = 4.9 Hz, 6H), 7.73 (d, J = 8.5 Hz, 6H), 7.59 (dd, J = 4.9, 8.5 Hz, 6H), 7.33 (d, J = 8.5 Hz, 6H), 3.02 (m, 6H), 2.69 (m, 6H). ¹³C NMR (DMSO-d₆): δ = 153.6 (C), 145.2 (CH), 140.7 (CH), 138.2 (C), 131.7 (CH), 128.0 (C), 125.4 (C), 122.0 (CH), 112.8 (CH); one CH₂ is not observed. FAB(+)-MS: m/z (%) = 1035 (40, [{Al₂(2)₃}K]⁺). Elemental analysis calcd for Al₂C₆₀H₄₂N₆O₆ ⋅ KCl ⋅ 4.5H₂O: C 62.52, H 4.46, N 7.29; found: C 62.77, H 4.95, N 7.18. IR (KBr): ν̃ = 3380, 3056, 2939, 2859, 1503, 1467, 1380, 1117, 722, 682 cm⁻¹. UV-vis (DMSO): λ = 265, 394 nm.

[Na⊂(Fe₂(2)₃)]Cl.

Yield: 14 mg (0.02 mmol, 75%) as a black solid. FAB(+)-MS: m/z (%) = 1077 (1, [{Fe₂(2)₃}Na]⁺). Elemental analysis calcd for Fe₂C₆₀H₄₂N₆O₆ ⋅ NaCl ⋅ 6H₂O: C 59.01, H 4.46, N 6.88; found: C 58.60, H 5.24, N 6.74. IR (KBr): ν̃ = 3404, 3050, 2913, 2855, 1501, 1458, 1373, 1312, 1104, 826, 731, 681 cm⁻¹. UV-vis (DMSO): λ = 263, 478 nm.

[K⊂(Fe₂(2)₃)]Cl.

Yield: 14 mg (0.01 mmol, 76%) as a black solid. FAB(+)-MS: m/z (%) = 1094 (low intensity, [{Fe₂(2)₃}K]⁺). Elemental analysis calcd for Fe₂C₆₀H₄₂N₆O₆ ⋅ KCl ⋅ 4H₂O: C 59.99, H 4.20, N 7.00; found: C 60.34, H 4.97, N 6.87. IR (KBr): ν̃ = 3567, 1501, 1457, 1372, 1311, 1104, 826, 731, 682 cm⁻¹. UV-vis (DMSO): λ = 263, 666 nm.

[Na⊂(Cr₂(2)₃)]Cl.

The reaction of 2-H₂ is performed in tetrahydrofuran (THF) with CrCl₃ ⋅ THF. Yield: 13 mg (0.01 mmol, 63%) of a green solid. FAB(+)-MS: m/z (%) = 1069 (1, [{Cr₂(2)₃}Na]⁺). Elemental analysis calcd for Cr₂C₆₀H₄₂N₆O₆ ⋅ NaCl ⋅ 14H₂O: C 53.08, H 5.20, N 6.19; found: C 53.04, H 4.63, N 5.84. IR (KBr): ν̃ = 3052, 1598, 1545, 1520, 1458, 1371, 1313, 1112, 830, 758, 678 cm⁻¹. UV-vis (methanol): λ = 197, 262, 319, 381 nm.

[K⊂(Cr₂(2)₃)]Cl.

The reaction of 2-H₂ is performed in THF with CrCl₃ ⋅ THF. Yield: 14 mg (0.01 mmol, 78%) as a green solid. FAB(+)-MS: m/z (%) = 1086 (0.4, [{Cr₂(2)₃}K]⁺). Elemental analysis calcd for Cr₂C₆₀H₄₂N₆O₆ ⋅ KCl ⋅ 3H₂O: C 61.30, H 4.12, N 7.15; found: C 61.21, H 4.62, N 6.69. IR (KBr): ν̃ = 3059, 1545, 1502, 1459, 1371, 1313, 1112, 829, 758, 678 cm⁻¹. UV-vis (methanol): λ = 197, 247, 266, 412 nm.

[K⊂(Al₂(3)₃)]Cl.

Yield: 22 mg (0.01 mmol, 67%) as a yellow solid. ¹H NMR (DMSO-d₆): δ = 8.62 (d, J = 8.4 Hz, 6H), 7.50 (dd, J = 4.8, 8.4 Hz, 6H), 7.43 (d, J = 8.3 Hz, 6H), 7.28 (d, J = 8.3 Hz, 6H), 7.15 (d, J = 4.8 Hz, 6H), 4.22 (s, 6H), 3.55 (d, J = 16.5 Hz, 6H), 3.15 (d, J = 16.5 Hz, 6H). ¹³C NMR (DMSO-d₆): δ = 155.1 (C), 148.2 (C), 144.1 (CH), 140.5 (CH), 138.1 (C), 133.0 (CH), 128.1 (C), 122.9 (C), 121.8 (CH), 112.3 (CH), 107.2 (CH₂), 37.0 (CH₂). FAB(+)-MS: m/z (%) = 1113 (25, [{Al₂(3)₃}K]⁺). Elemental analysis calcd for Al₂C₆₆H₄₈N₆O₆ ⋅ KCl ⋅ 11H₂O: C 58.82, H 5.23, N 6.24; found: C 58.66, H 4.63, N 6.03. IR (KBr): ν̃ = 3383, 1502, 1467, 1380, 1319, 1116, 827, 763, 746, 675 cm⁻¹. UV-vis (DMSO): λ = 264, 317, 382 nm.

[NH₄⊂(Al₂(3)₃)]Cl.

Yield: 21 mg (0.01 mmol, 74%) of a yellow solid. ¹H NMR (DMSO-d₆): δ = 8.64 (d, J = 8.3 Hz, 6H), 7.52 (dd, J = 4.8, 8.3 Hz, 6H), 7.45 (d, J = 8.4 Hz, 6H), 7.30 (d, J = 8.4 Hz, 6H), 7.16 (d, J = 4.8 Hz, 6H), 6.65 (br s, 4H), 4.18 (s, 6H), 3.57 (d, J = 16.8 Hz, 6H), 3.15 (d, J = 16.8 Hz, 6H). ¹³C NMR (DMSO-d₆): δ = 154.9 (C), 148.6 (C), 144.1 (CH), 140.5 (CH), 138.0 (C), 133.3 (CH), 128.1 (C), 122.5 (CH), 121.9 (C), 112.5 (CH), 107.5 (CH₂), 37.2 (CH₂). FAB(+)-MS: m/z (%) = 1092 (50, [{Al₂(3)₃}(NH₄)]⁺). Elemental analysis calcd for Al₂C₆₆H₄₈N₆O₆ ⋅ NH₄Cl ⋅ 2H₂O: C 68.07, H 4.85, N 8.42; found: C 67.96, H 5.12, N 8.07. IR (KBr): ν̃ = 3061, 2890, 1502, 1466, 1380, 1320, 1115, 827, 764, 746, 673 cm⁻¹. UV-vis (DMSO): λ = 266, 321, 393 nm.

[Rb⊂(Al₂(3)₃)]Cl.

Yield: 20 mg (0.01 mmol, 71%) of a yellow solid. ¹H NMR (DMSO-d₆): δ = 8.63 (dd, J = 1.1, 8.4 Hz, 6H), 7.51 (dd, J = 4.9, 8.4 Hz, 6H), 7.44 (d, J = 8.4 Hz, 6H), 7.29 (d, J = 8.4 Hz, 6H), 7.15 (dd, J = 1.1, 4.9 Hz, 6H), 4.13 (s, 6H), 3.57 (d, J = 17.2 Hz, 6H), 3.13 (d, J = 17.2 Hz, 6H). ¹³C NMR (DMSO-d₆): δ = 155.1 (C), 149.1 (C), 144.1 (CH), 140.5 (CH), 138.0 (C), 133.3 (CH), 128.2 (C), 122.3 (CH), 122.0 (C), 112.4 (CH), 106.8 (CH₂), 37.3 (CH₂). FAB(+)-MS: m/z (%) = 1159 (100, [{Al₂(3)₃}Rb]⁺). Elemental analysis calcd for Al₂C₆₆H₄₈N₆O₆ ⋅ RbCl ⋅ 4H₂O: C 62.51, H 4.45, N 6.63; found: C 62.57, H 4.96, N 6.53. IR (KBr): ν̃ = 3382, 2887, 1583, 1504, 1467, 1379, 1319, 1116, 831, 772, 747, 673, 648 cm⁻¹. UV-vis (DMSO): λ = 264, 389 nm.

[Cs⊂(Al₂(3)₃)]Cl.

Yield: 22 mg (0.01 mmol, 68%) of a yellow solid. ¹H NMR (DMSO-d₆): δ = 8.64 (dd, J = 1.2, 8.4 Hz, 6H), 7.55 (dd, J = 4.9, 8.4 Hz, 6H), 7.44 (d, J = 8.4 Hz, 6H), 7.38 (dd, J = 1.2, 4.9 Hz, 6H), 7.29 (d, J = 8.4 Hz, 6H), 4.13 (s, 6H), 3.53 (d, J = 17.5 Hz, 6H), 3.13 (d, J = 17.5 Hz, 6H). ¹³C NMR (DMSO-d₆): δ = 155.2 (C), 151.1 (C), 144.4 (CH), 140.5 (CH), 138.0 (C), 133.5 (CH), 128.3 (C), 122.1 (C), 121.8 (CH), 112.3 (CH), 108.5 (CH₂), 37.0 (CH₂). FAB(+)-MS: m/z (%) = 1207 (50, [{Al₂(3)₃}Cs]⁺). Elemental analysis calcd for Al₂C₆₆H₄₈N₆O₆ ⋅ CsCl ⋅ 5H₂O: C 59.45, H 4.38, N 6.30; found: C 59.52, H 4.61, N 6.07. IR (KBr): ν̃ = 3383, 2888, 1583, 1504, 1466, 1379, 1321, 1116, 826, 776, 767, 673 cm⁻¹. UV-vis (DMSO): λ = 264, 388 nm.

[K⊂(Fe₂(3)₃)]Cl.

Yield: 20 mg (0.01 mmol, 62%) of a black solid. FAB(+)-MS: m/z (%) = 1171 (40, [{Fe₂(3)₃}K]⁺). Elemental analysis calcd for Fe₂C₆₆H₄₈N₆O₆ ⋅ KCl ⋅ 6H₂O: C 60.26, H 4.60, N 6.39; found: C 60.26, H 4.81 N 6.30. IR (KBr): ν̃ = 3352, 3050, 1501, 1457, 1372, 1314, 1110, 826, 736, 631 cm⁻¹. UV-vis (DMSO): λ = 261, 374, 475, 605 nm.

[NH₄⊂(Fe₂(3)₃)]Cl.

Yield: 24 mg (0.01 mmol, 79%) as a black solid. FAB(+)-MS: m/z (%) = 1150 (4, [{Fe₂(3)₃}(NH₄)]⁺). Elemental analysis calcd for Fe₂C₆₆H₄₈N₆O₆ ⋅ NH₄Cl ⋅ 3H₂O: C 63.91, H 4.71, N 7.90; found: C 64.20, H 4.82, N: 7.63. IR (KBr): ν̃ = 3064, 2899, 1570, 1501, 1457, 1372, 1314, 1110, 826, 736 cm⁻¹. UV-vis (DMSO): λ = 262, 375, 475, 599 nm.

[Rb⊂(Fe₂(3)₃)]Cl.

Yield: 22 mg (0.01 mmol, 68%) as a black solid. FAB(+)-MS: m/z (%) = 1217 (18, [{Fe₂(3)₃}Rb]⁺). Elemental analysis calcd for Fe₂C₆₆H₄₈N₆O₆ ⋅ RbCl ⋅ 4H₂O: C 59.79, H 4.26, N 6.34; found: C 59.52, H 4.06, N: 6.25. IR (KBr): ν̃ = 3329, 3050, 2888, 1571, 1501, 1458, 1372, 1314, 1109, 825, 736 cm⁻¹. UV-vis (DMSO): λ = 260, 372, 473, 599 nm.

[Cs⊂(Fe₂(3)₃)]Cl.

Yield: 22 mg (0.02 mmol, 67%) of a black solid. FAB(+)-MS: m/z (%) = 1265 (40, [M]⁺). Elemental analysis calcd for Fe₂C₆₆H₄₈N₆O₆ ⋅ CsCl ⋅ 2.5H₂O: C 58.88, H 3.97, N 6.24; found: C 58.91, H 3.74, N 6.05. IR (KBr): ν̃ = 3350, 2888, 1501, 1457, 1372, 1315, 1110, 826, 737, 670 cm⁻¹. UV-vis (DMSO): λ = 260, 368, 469, 607 nm.

X-Ray Structural Analyses.

Data sets were collected with a Nonius KappaCCD diffractometer, equipped with a rotating anode generator Nonius FR591. Programs used: data collection, COLLECT (Nonius B.V., 1998); data reduction, DENZO-SMN (18); absorption correction, SORTAV (19, 20); structure solution, SHELXS-97; structure refinement, SHELXL-97 (21, 22); and graphics SCHAKAL (23).

Crystallographic data (excluding structure factors) for the structures reported in this paper have been deposited with the Cambridge Crystallographic Data Centre as supplementary publication CCDC 173321, 173322, 173323, 173324, 173325, and 173326. Copies of the data can be obtained free of charge on application to The Director, CCDC, 12 Union Road, CambridgeCB2 1EZ, U.K. [fax: + 44(1223)336–033; e-mail: deposit@ccdc.cam.ac.uk].

The three structures [M⊂(Ga₂(3)₃)]NO₃ (M = NH₄, K, Cs) are isostructural. The quality of the structures is heavily suffering from disordered nitrate and missing hydrogens on NH₄⁺ and H₃O⁺ (light cations and anions are only refined with isotropic thermal parameters). Remaining electron density could not be assigned in a chemically meaningful way.

[NH₄⊂(Ga₂ (3)₃)]NO_3.

C₆₆H₄₈N₆O₆Ga₂ ⋅ NH₄NO₃ ⋅ 2 H₃ONO₃, M = 1402.66, yellow crystal 0.50 × 0.25 × 0.20 mm, a = 14.633(1), c = 19.418(1) Å, V = 3600.8(4) ų, ρ_calc = 1.294 g⋅cm⁻³, μ = 8.19 cm⁻¹, empirical absorption correction by SORTAV (0.685 ≤ T ≤ 0.853), Z = 2, hexagonal, space group P6₃/m (No. 176), λ = 0.71073 Å, T = 198 K, ω and ϕ scans, 5,898 reflections collected (+h, +k, ±l), [(sinθ)/λ] = 0.67 Å⁻¹, 3,058 independent (R_int = 0.034) and 1,801 observed reflections [I ≥ 2 σ(I)], 137 refined parameters, R = 0.122, wR (2) = 0.334, max residual electron density 1.59 (−0.76) e⋅Å⁻³, hydrogens calculated and refined as riding atoms.

[K⊂(Ga₂(3)₃)]NO₃.

C₆₆H₄₈N₆O₆Ga₂ ⋅ KNO₃ ⋅ 2 H₃ONO₃, M = 1423.72, yellow crystal 0.50 × 0.45 × 0.40 mm, a = 14.756(1), c = 19.194(1) Å, V = 3619.4(4) ų, ρ_calc = 1.306 g⋅cm⁻³, μ = 8.72 cm⁻¹, empirical absorption correction by SORTAV (0.670 ≤ T ≤ 0.722), Z = 2, hexagonal, space group P6₃/m (No. 176), λ = 0.71073 Å, T = 198 K, ω and ϕ scans, 5,967 reflections collected (+h, −k, ±l), [(sinθ)/λ] = 0.67 Å⁻¹, 3,074 independent (R_int = 0.016) and 2,102 observed reflections [I ≥ 2 σ(I)], 138 refined parameters, R = 0.145, wR (2) = 0.422, max residual electron density 3.13 (−1.98) e⋅Å⁻³, hydrogens calculated and refined as riding atoms.

[Cs⊂(Ga₂(3)₃)]NO₃.

C₆₆H₄₈N₆O₆Ga₂ ⋅ CsNO₃ ⋅ 2 H₃ONO₃, M = 1517.53, yellow crystal 0.35 × 0.25 × 0.10 mm, a = 14.767(1), c = 19.578(1) Å, V = 3697.3(4) ų, ρ_calc = 1.363 g⋅cm⁻³, μ = 12.82 cm⁻¹, empirical absorption correction by SORTAV (0.663 ≤ T ≤ 0.883), Z = 2, hexagonal, space group P6₃/m (No. 176), λ = 0.71073 Å, T = 198 K, ω and ϕ scans, 1,824 reflections collected (+h, -k, +l), [(sinθ)/λ] = 0.56 Å⁻¹, 1,824 independent (R_int = 0.000) and 1,441 observed reflections [I ≥ 2 σ(I)], 138 refined parameters, R = 0.149, wR (2) = 0.448, max residual electron density 3.45 (−2.56) e⋅Å⁻³, hydrogens calculated and refined as riding atoms.

[NH₄⊂(Al₂(3)₃)]Cl.

C₆₆H₄₈N₆O₆Al₂ ⋅ NH₄Cl ⋅ 1.5 NH₄NO₃, M = 1248.63, yellow crystal 0.20 × 0.20 × 0.10 mm, a = 14.531(1), c = 19.324(1) Å, V = 3533.6(4) ų, ρ_calc = 1.174 g⋅cm⁻³, μ = 1.40 cm⁻¹, empirical absorption correction by SORTAV (0.986 ≤ T ≤ 0.973), Z = 2, hexagonal, space group P6₃/m (No. 176), λ = 0.71073 Å, T = 198 K, ω and ϕ scans, 14,470 reflections collected (±h, ±k, ±l), [(sinθ)/λ] = 0.62 Å⁻¹, 2,468 independent (R_int = 0.071) and 1,547 observed reflections [I ≥ 2 σ(I)], 149 refined parameters, R = 0.130, wR (2) = 0.327, max residual electron density 1.46 (−0.43) e⋅Å⁻³, hydrogens calculated and refined as riding atoms.

The ammonium ion inside the complex occupies three symmetry equivalent positions because of the size of the cavity. The additional ammonium nitrates show some disorder and are refined with isotropic thermal parameters. Hydrogen atoms could not be localized. Remaining electron density could not be assigned in a chemically meaningful way.

The two structures [M(H₂O)⊂(Fe₂(3)₃)]Cl (M = K, Rb) are isostructural. The quality of the structures is heavily suffering from disordered solvent molecules (for two of them occupancies are set to 0.5 and 0.25), which are refined with common isotropic thermal parameters and restraints. Remaining electron density could not be assigned in a chemically meaningful way.

[K(H₂O)⊂(Fe₂(3)₃)]Cl.

C₆₆H₄₈N₆O₆Fe₂ ⋅ KCl ⋅ 3.5 C₃H₇NO ⋅ H₂O, M = 1481.21, red crystal 0.30 × 0.25 × 0.10 mm, a = 12.341(1), b = 20.529(1), c = 15.003(1) Å, β = 100.55(1)°, V = 3736.7(4) ų, ρ_calc = 1.316 g⋅cm⁻³, μ = 5.44 cm⁻¹, empirical absorption correction by SORTAV (0.854 ≤ T ≤ 0.948), Z = 2, monoclinic, space group P2₁/m (No. 11), λ = 0.71073 Å, T = 198 K, ω and ϕ scans, 11,257 reflections collected (±h, ±k, ±l), [(sinθ)/λ] = 0.59 Å⁻¹, 6,603 independent (R_int = 0.045) and 2,217 observed reflections [I ≥ 2 σ(I)], 467 refined parameters, R = 0.095, wR (2) = 0.233, max residual electron density 2.31 (−0.81) e⋅Å⁻³, hydrogens calculated and refined as riding atoms.

[Rb(H₂O)⊂(Fe₂(3)₃)]Cl.

C₆₆H₄₈N₆O₆Fe₂ ⋅ KCl ⋅ 3.5 C₃H₇NO ⋅ H₂O, M = 1527.58, dark-red crystal 0.50 × 0.30 × 0.05 mm, a = 12.290(1), b = 20.528(1), c = 14.966(1) Å, β = 100.07(1)°, V = 3717.6(4) ų, ρ_calc = 1.365 g⋅cm⁻³, μ = 11.41 cm⁻¹, empirical absorption correction by SORTAV (0.599 ≤ T ≤ 0.945), Z = 2, monoclinic, space group P2₁/m (No. 11), λ = 0.71073 Å, T = 198 K, ω and ϕ scans, 11,453 reflections collected (±h, ±k, ±l), [(sinθ)/λ] = 0.59 Å⁻¹, 6,743 independent (R_int = 0.045) and 5,091 observed reflections [I ≥ 2 σ(I)], 467 refined parameters, R = 0.146, wR (2) = 0.403, max residual electron density 5.28 (−2.17) e⋅Å⁻³, hydrogens calculated and refined as riding atoms.

Results and Discussion

Formation of the Metalla-Cryptates [M′⊂{(ligand)₃M₂}]⁺ (M = Al, Ga, Fe, Cr; M′ = Li, Na, K, Rb, Cs, NH₄).

There are two possible routes to obtain metalla-cryptates [M′⊂{(1-4)₃M₂}]⁺ (i) They are formed if a suspension of the preformed mixtures of neutral complexes “[(1-4)₃M₂]” is heated in the presence of an excess of alkali metal or ammonium salts (Scheme S1 b and c); or (ii) if alkali metal or ammonium salts are added during the complex formation (Scheme S1a; ref. 6).

Scheme 1.

Schematic representation of the formation of helicate-type metalla-cryptates [M′⊂{(1-4)₃M₂}]⁺ in either a one-step template-directed self-assembly process (a) or in a dynamic combinatorial approach generating first molecular diversity (b) followed by a selection step (c).

Dinuclear Gallium(III) Metalla-Cryptates.

Dinuclear gallium(III) complexes (24) of ligand 1 are obtained by reaction of the ligand 1-H₂ (3 eq) with 2 eq of Ga(NO₃)₃. The precipitate is filtered off, washed, and heated together with an excess of LiCl or NaCl in a mixture of chloroform, methanol, and water (5:1:0.1) to yield an unstoichiometric adduct of the dinuclear complex [(1)₃Ga₂] ⋅ nM′Cl (M′ = Li, Na). The ¹H NMR spectrum of the lithium chloride adduct of [(1)₃Ga₂] (CDCl₃) shows the signals of the aromatic moieties at δ = 8.19 (s, 6H), 7.79 (d, J = 8.4 Hz, 6H), 7.58 (s, 6H), and 7.15 (d, J = 8.4 Hz, 6H). The diastereotopic protons of the spacer occur as doublets at δ = 5.04 and 3.21 (J = 12.4 Hz, 3H each) and the resonances of the decyl substituent are detected at δ = 2.69 (m, 12H), 1.63 (m, 24H), 1.23 (m, 84H), and 0.87 (m, 18H). The sodium compound leads to an analogous spectrum.

The fact that an excess of lithium or sodium cations have to be present in solution indicates that the internal cavity of [(1)₃Ga₂] is too small to encapsulate one cation in its interior but two or three cations bind to the “outside” of this small metalla-cryptand (Fig. 2). Similar binding behavior was observed earlier for related dinuclear titanium(IV) complexes of a methylene-bridged di(catecholate) ligand (5, 25). Larger cations than Na⁺ are not able to act as template to stabilize the dinuclear gallium cryptate [(1)₃Ga₂]. Therefore no defined complexes are obtained in the presence of, for example, potassium or ammonium cations (Scheme S2).

Figure 2.

Proposed stabilization of the dinuclear gallium(III) complex [(1)₃Ga₂] by binding of more than one templating lithium or sodium cation (M) (5, 25).

Scheme 2.

Template-directed formation of triple-stranded dinuclear gallium(III) cryptates. The templating ability of different alkali metal cations depending on the length of the carbon chain in the ligand spacer is shown (the radius of the cations is taken from ref. 28, and is given for hexacoordinated species with a charge of +1).

We described earlier that the ethylene-bridged ligands 2 in the presence of sodium, potassium, rubidium, or ammonium cations form dinuclear metalla-cryptates [M′{(2)₃Ga₂}]⁺ (M′ = Na, K, NH₄, Rb) with gallium(III) ions. The small Li⁺ cation does not support the selective formation of a dinuclear complex neither does Cs⁺, which is too big (Scheme S2; ref. 6).

The di(8-hydroxyquinoline) ligand 3-H₂ was introduced by Hiratani (14). It bears a C₃ spacer with a central sp² carbon atom to which a methylene unit is attached. Reaction of this ligand 3-H₂ with Ga(NO₃)₃, in the presence of KCl, NH₄Cl, RbCl, or CsCl, leads to the metalla-cryptates [M′{(3)₃Ga₂}]Cl (M′ = K, NH₄, Rb, Cs) (17). The dinuclear complexes cannot be obtained in the presence of the small lithium or sodium cations.

All complexes [M′{(3)₃Ga₂}]Cl (M′ = K, NH₄, Rb, Cs) lead to similar characteristic NMR spectra. [K{(3)₃Ga₂}]Cl for example shows signals in the ¹H NMR spectrum in DMSO-d₆ for the aromatic moiety at δ = 8.65 (dd, J = 1.2, 8.3 Hz), 7.54 (dd, J = 4.8, 8.3 Hz), 7.45–7.43 (m, double intensity), and 7.26 (d, J = 8.4 Hz). The olefinic protons of the spacer are observed at δ = 4.24 (s) and the diastereotopic protons in benzylic position at δ = 3.57 (d) and 3.18 (d, J = 16.6 Hz). The olefinic protons are shifted to high field compared with related ⩵CH₂ units. This anisotropic shift is due to “C”-type conformation of the ligand in this meso-helicate (26) [“side-by-side complex” (3), “mesocate” (27)], resulting in a location of the alkene protons close to the aromatic ligand moieties.

The dinuclear gallium(III) complex of ligand 4, with relatively long (CH₂)₄ spacers, possesses a large internal cavity. Therefore a dinuclear gallium complex can be obtained in the presence of a large template only. In this case the template-directed self-assembly is only successful if cesium chloride is present. No corresponding cryptates are formed with smaller templates (e.g., Li⁺, Na⁺, K⁺, NH₄⁺, Rb⁺). In the positive FAB MS the cationic cryptate [Cs{(4)₃Ga₂}]⁺ is observed at m/z = 1299. The ¹H NMR spectrum in DMSO-d₆ shows characteristic signals at δ = 8.81 (d, J = 4.3 Hz), 8.28 (d, J = 8.2), 7.49 (dd, J = 4.3, 8.2 Hz), 7.39 (d, J = 8.4 Hz), 7.32 (d, J = 8.4 Hz), and 1.50 (m, double intensity). A further signal of the spacer is hidden under the solvent peak.

Dinuclear Aluminum(III), Iron(III), and Chromium(III) Metalla-Cryptates.

For comparison we prepared some selected dinuclear complexes of aluminum(III), iron(III), and chromium(III). The cryptates [M′{(2)₃M₂}]⁺ (M = Al, Fe, Cr; M′ = Na, K) and [M′{(3)₃M₂}]⁺ (M = Al, Fe; M′ = K, NH₄, Rb, Cs) are obtained by reaction of the ligands 2-H₂ or 3-H₂ with metal chlorides (AlCl₃, FeCl₃ ⋅ 6H₂O, CrCl₃ ⋅ THF) in the presence of an excess of the corresponding alkali metal or ammonium salt.

The complexes were characterized by positive FAB MS, elemental analysis, UV-vis, and in case of the diamagnetic aluminum complexes by NMR spectroscopy.

Solid-State Structures of Metalla-Cryptates.

Several x-ray structures were obtained for metal complexes of the ligands 2 or 3 although many of them were of poor quality because of some disorder of solvent and anions. However, the overall structure of the cationic metalla-cryptates can be observed nicely. In total eight structures were found that show only three different types (Fig. 3).

Figure 3.

Solid-state structures of the metalla-cryptates [K(dmf)₂{(2)₃Ga₂}]⁺, (6) [Cs{(3)₃Ga₂}]⁺, and [(H₂O)KCl{(3)₃Fe₂}] as representative examples for the three different structural types of complexes (hydrogen atoms are omitted for clarity, dmf molecules are only indicated; further representations of the new x-ray structures can be found as supporting information on the PNAS web site, www.pnas.org).

The structures of the sodium as well as of the potassium containing helicates [M′(dmf)₂{(2)₃Ga₂}]⁺ were published earlier (6). In [M′(dmf)₂{(2)₃Ga₂}]⁺ the neutral cryptand [(2)₃Ga₃] to some extent is able to adjust to the size of the guest cations. Gallium–gallium separations of 6.287 Å (M′ = Na) and 6.565 Å (M′ = K) are observed showing the bigger size of the potassium cryptate compared with the one with sodium.

The dinuclear gallium complexes [M′{(3)₃Ga₂}]⁺ (M′ = NH₄, K, Cs) all crystallize isostructural in the space group P6₃/m and could be refined to relatively poor R values only. (However, three similar poor structures unequivocally show the structure of the cationic metalla-cryptate.) In the crystal, nitrate is present as anion that was introduced during the preparation of the coordination compounds from gallium nitrate. The dinuclear gallium complex [(3)₃Ga₂] forms a C₃-symmetric cavity that encapsulates the cations. Potassium and cesium bind to the internal oxygen atoms (K–O = 3.095 Å, Cs–O = 3.197) and ammonium is also encapsulated (the hydrogen atoms could not be located). Because of the size of the cavity the positions of the cations are not well defined. The electron density map indicates a high mobility of the cations in the cavity. In the case of the ammonium and potassium derivative of [M′{(3)₃Ga₂}]⁺ a gallium–gallium separation of 7.389 and 7.382 Å is found, showing the similar size of the bound monocations. For [Cs{(3)₃Ga₂}]⁺ the distance Ga–Ga = 7.651 Å is somewhat longer because of the bigger size of the cation. The aluminum complex [(NH₄){(3)₃Al₂}]⁺ possesses the same structure as observed for the corresponding gallium compound with an Al–Al separation of 7.267 Å.

The complex [KCl(H₂O){(3)₃Fe₂}] crystallizes in the space group P2₁/m. In the solid-state structure the alkali metal cation is encapsulated in the interior of [(3)₃Fe₂] in an unsymmetric fashion. Potassium possesses four short contacts to the oxygen atoms of two of the bis-8-oxyquinolinate strands. In addition, a chloride is bound to the potassium from the outside of the cavity (K–Cl = 2.987 Å) and a water molecule, which is located in the interior, is fixed trans to the chloride (K–O = 2.619 Å). For the rubidium derivative [RbCl(H₂O){(3)₃Fe₂}] a similar structure is found in the crystal with elongated distances due to the larger size of the cation.

The discussed x-ray structural results show that the dinuclear cryptand-type complexes [(2)₃M₂] and [(3)₃M₂] possess different sizes. This results in a different separation of the two M(III) metal ions with 6.3–6.6 Å for [(2)₃M₂] and 7.2–7.7 Å for [(3)₃M₂]. To some extent, the dinuclear triple-stranded metal complexes are able to adjust to the size (and the geometry) of the cationic guest that is bound in the cavity.

Conclusion

The results reported in this paper show that the self-assembly of dinuclear metal complexes from alkyl-bridged di(8-hydroxyquinoline) ligands highly depends on the size of the template that directs the formation of the complexes. It was described earlier that the alkyl-bridged helicate-type dinuclear complexes are able to adjust to the size of different templates by using the octahedral complex units as molecular hinges (6). However, this adjustment seems to be possible to a limited extend only. In the case of ligands with short spacers small cationic templates are needed, whereas large cavities, which are obtained from ligands with long spacers, prefer bigger templates. To study this effect we used the four ligands 1-4-H₂ where the spacer successively is extended by one carbon atom. The influence of different alkali metal cations as templates for the formation of differently sized triple-stranded dinuclear gallium complexes was systematically investigated.

As a result, defined complexes of the small ligand 1 are only obtained in the presence of an excess of lithium or sodium cations. The somewhat larger metalla-cryptates of ligand 2 are formed in the presence of sodium, potassium, ammonium, and rubidium cations, whereas potassium, ammonium, rubidium, and cesium support the formation of complexes of ligand 3. Dinuclear coordination compounds of the biggest ligand 4 are only observed in the presence of the large cesium cation.

Thus, size-selectivity is observed in the template-directed self-assembly of metalla-cryptates [M′{(1-4)₃M₂}]⁺. Only cations M′ of a specific size are able to support the formation of well defined complexes. If there is no size-complementarity between the template M′⁺ and the cryptand [(1-4)₃M₂] (which is a component of a dynamic supramolecular combinatorial library), no selection by the template can take place and no specific product is obtained.

Herein we presented a very simple example where we could show how subtle the influences can be that control the outcome of a self-recognition process. In our case this depends on the molecular recognition between ligands and gallium(III) [or aluminum(III), iron(III), chromium(III)] to form dinuclear coordination compounds and in addition on the control of this process by the selective recognition of templates. The different metal ions, gallium(III), aluminum(III), iron(III), or chromium(III), do not influence the self-assembly process significantly. They all form related triple-stranded helicate-type complexes.

The understanding of fundamental reaction steps as discussed in this paper are essential for a rational design of large supramolecular architectures as found, for example, in nature. Therefore we will proceed to study such simple but important mechanistic aspects of supramolecular as well as metallosupramolecular chemistry.

Abbreviations

FAB

fast atom bombardment

THF

tetrahydrofuran