Design and self-assembly of wide and robust coordination cages

Abstract

The self-assembly of a new class of coordination cages formed from two tetrapyridyl-substituted cavitands connected through four square-planar palladium or platinum complexes is reported. The shape of the internal cavity resembles an ellipsoid with a calculated volume of 840 ų. The four lateral portals have a diameter of about 6 Å, large enough to allow the fast entrance/egress of counterions in solution. The platinum cages 3a,e cannot be disassembled using triethylamine as competitive ligand and they are kinetically stable at room temperature, whereas the palladium cages 3b-d, 3f-h are disassembled and kinetically labile under the same conditions. The different solubility properties of the cage components have allowed the extension of this self-assembly protocol to the liquid–liquid interface.

Metal-directed self-assembly has been widely used to construct three-dimensional structures presenting internal cavities of molecular dimensions, capable of trapping ions and neutral molecules (1, 2). The ability of such container molecules to confine (3) and stabilize (4) their guests makes them particularly attractive for many potential applications ranging from catalysis (5, 6) to nanotechnology (for self-assembly in nanotechnology, see ref. 7; ref. 8). Specifically, surface-controlled self-assembly is rapidly emerging as a valuable tool for the generation of complex structures directly on solid supports (9–11). Unlike the covalent approach used for carcerands synthesis (12), the thermodynamic control of the process conveys interesting features, among which reversibility (13), selection (14), and self-repairing properties are the most interesting.

Cavitand-based coordination cages are receiving increasing attention because of the versatility of cavitand platforms in terms of preorganization and synthetic modularity (15–20). Strong metal–ligand interactions are necessary to operate cage self-assembly (CSA) in a wide range of solvents, to extend further these self-assembly protocols to solid–liquid and liquid–liquid interfaces.

For this purpose we designed and synthesized new cavitands bearing pyridines (for references on cages derived from multidentate pyridine ligands, see ref. 21; refs. 22 and 23) instead of nitriles (24) at the upper rim as ligands for the coordination to the metal centers. CSA requires a cavitand with four pyridines preorganized in a diverging spatial orientation, pointing outward (o) the cavity, being designated the oooo isomer (for in/out isomerism in cavitands, see ref. 25).

Materials and Methods

General.

All commercial reagents were ACS reagent grade and used as received. All solvents were dried over 3- and 4-Å molecular sieves. Resorcinarenes 1a,b were prepared according to the literature (26). Metal precursors were prepared from the corresponding dichlorobis derivatives following established procedures (27–29). ¹H NMR were recorded on Bruker (Billerica, MA) AC300 (300 MHz) and AMX400 (400 MHz) spectrometers and all chemical shifts (δ) were reported in parts per million (ppm) relative to the proton resonances resulting from incomplete deuteration of the NMR solvents. ¹³C NMR spectra were recorded on a Bruker AC300 (75 MHz) spectrometer. ³¹P NMR spectra were recorded on a Bruker XP200 (81 MHz), and all chemical shifts were reported in ppm relative to external 85% H₃PO₄ at 0.00 ppm. ¹⁹F NMR spectra were recorded on a Bruker XP200 (188 MHz) spectrometer, and all chemical shifts were reported relative to external CFCl₃ at 0.00 ppm. Melting points were obtained with an electrothermal capillary melting points apparatus and are uncorrected. Mass spectra of the organic compounds were measured with a Finnigan-MAT (San Jose, CA) SSQ 710 spectrometer, using the CI (chemical ionization) technique. Electrospray ionization (ESI)-MS experiments were performed on a Perkin–Elmer API 100 SCIEX spectrometer equipped with an ionospray interface. Matrix-assisted laser desorption ionization–time-of-flight (MALDI-TOF) mass spectra were obtained on a PerSeptive Biosystems (Framingham, MA) Voyager DE-RP spectrometer equipped with delayed extraction. HPLC chromatography was performed using a semipreparative HPLC S10 Nitrile column on a Perkin–Elmer Series II instrument.

Cavitand 2a.

A mixture of resorcinarene 1a (R = C₁₁H₂₃; 0.707 g, 0.64 mmol), K₂CO₃ (1.06 g, 7.67 mmol) and 4-(dibromomethyl)pyridine (see Supporting Text, which is published as supporting information on the PNAS web site, www.pnas.org) (1.61 g, 6.40 mmol) in 30 ml of dry N,N,-dimethylacetamide (DMA) was stirred in a sealed tube at 80°C for 15 h. The reaction mixture was then cooled, poured into a saturated Na₂CO₃ solution, and extracted with CH₂Cl₂. Removal of the CH₂Cl₂ gave a black solid that was purified in two steps: at first by column chromatography (SiO₂, CH₂Cl₂/EtOH from 9:1 to 1:1), then on silica gel plates (CH₂Cl₂/EtOH 9:1) to give cavitand 2a in 10% yield (0.090 g, 0.06 mmol); mp 106–108°C.

¹H NMR (300 MHz, CDCl₃, 300 K): δ 0.88 (bt, 12H, CH₃), 1.28 [m, 72H (CH₂)₉], 2.32 (m, 8H, CHCH₂), 4.91 (t, 4H, CHCH₂, J = 7.9 Hz), 5.46 (s, 4H, CHPy), 6.71 (s, 4H, ArH), 7.28 (s, 4H, ArH), 7.71 (d, 8 H, m-Py, AA′ part of an AA′XX′ system, J = 5.5 Hz), 8.72 (d, 8 H, o-Py, XX′ part of an AA′XX′ system, J = 5.5 Hz); CI-MS: m/z: 1462 (M⁻,100).

Cavitand 2b.

A mixture of resorcinarene 1b (R = CH₂CH₂Ph; 1.01 g, 1.11 mmol), K₂CO₃ (3.28 g, 23.73 mmol) and 4-(dibromomethyl)pyridine (4.16 g, 16.64 mmol) in 30 ml of dry DMA was stirred in a sealed tube at 80°C for 15 h. The reaction mixture was then cooled, poured into a saturated Na₂CO₃ solution, and extracted with CH₂Cl₂. Removal of the CH₂Cl₂ gave a black solid that was purified in two steps: at first by column chromatography (SiO₂, CH₂Cl₂/EtOH from 9:1 to 1:1), then by HPLC (S10 Nitrile, EtOH/CH₂Cl₂ 9:1) to give cavitand 2b in 12% yield (0.168 g, 0.13 mmol); mp 230°C (with decomposition).

¹H NMR (300 MHz, CDCl₃, 300 K): δ 2.62 (m, 8H, CH₂CH₂Ph), 2.75 (m, 8H, CH₂CH₂Ph), 5.04 (t, 4H, CHCH₂, J = 7.8 Hz), 5.45 (s, 4H, CHPy), 6.73 (s, 4H, ArH), 7.17–7.27 (m, 20H, CH₂C₆H₅), 7.34 (s, 4H, ArH), 7.60 (d, 8 H, m-Py, AA′ part of a AA′XX′ system, J = 4.7 Hz), 8.74 (d, 8 H, o-Py, XX′ part of an AA′XX′ system, J = 4.7 Hz); ¹³C NMR (75 MHz, CDCl₃, 300 K): δ 32.1 (CH₂), 34.3 (CH₂), 36.5 (CH), 106.0 (CH), 116.7 (CH), 120.8 (CH), 120.9 (CH), 126.2 (CH), 128.3 (CH), 128.6 (CH), 139.0 (Cq), 141.3 (Cq), 146.1 (Cq), 150.3 (CH), 154.2 (Cq); CI-MS: m/z: 1261 (M⁻, 100).

Crystallographic Data for 2b.

C₈₄H₆₈N₄O₈, fw = 1261.43, orthorombic space group Pca2₁, a = 20.608 (5), b = 33.315 (5), c = 21.863 (5) Å, V = 15010.2 (5) ų, ρ_calcd = g/cm⁻³, Z = 8, λ(Cu_Kα) = 1.54184 Å, T = 293 (2) K, θ/2θ scan technique, absorption coefficient = 0.57 mm⁻¹. A total of 14,476 unique data with 3,878 reflections having I > 4σ(I) were collected. Index range: −12 ≤ h ≤ 25, −12 ≤ k ≤ 26, −1 ≤ l ≤ 40, range from 3.23 to 70.27°. Refinement on F² converged to R₁ = 0.098 and wR₂ = 0.239.

The diffraction experiment was carried out on an Enraf Nonius CAD4 diffractometer. Data were corrected for Lorentz and polarization effects but not for absorption effects. The structure was solved by direct methods in the SIR92 program and refined by full-matrix least-squares procedures, using the SHELXL-97 system of crystallographic computer programs. Two symmetry independent molecules were found in the asymmetric unit. All non-hydrogen atoms were refined anisotropically with the exception of those belonging to the terminal phenyl rings; all hydrogen atoms were taken in their calculated positions and refined “riding” on the corresponding parent atoms; the weighting scheme used in the last cycle of refinement was w = 1/[σ²F + (0.0159 P)²], where P = (F + 2F)/3. The highest peak and the deepest hole in the final difference Fourier ΔF map were 1.081 and −0.339 eÅ⁻³. Molecular geometry calculations were carried out using the PARST97 program.

General Procedure for Cage Formation.

Cages 3a-h were assembled by simply mixing cavitands 2a and 2b with different metal precursors MLX₂ [M = Pd, Pt; L = 1,3-bis(diphenylphosphino)-propane (dppp), ethylenediamine (en); X = CF₃SO₃, NO₃] in a 1:2 molar ratio at room temperature in solvents such as CH₂Cl₂, CHCl₃, acetone, or dimethyl sulfoxide (DMSO). In all cases removal of the solvent in vacuo gave the desired cage in quantitative yields.

3a.

(M = Pt, L = dppp, X = CF₃SO₃, R = C₁₁H₂₃): ¹H NMR (300 MHz, CDCl₃, 300 K): δ = 0.85 (t, 24 H, CH₃, J = 6.8 Hz), 1.24–1.28 [m, 128 H (CH₂)₈], 1.41 (m, 16 H, CHCH₂CH₂), 2.25 (m, 24 H, CHCH₂CH₂ and Ph₂PCH₂CH₂), 3.39 (m, 16 H, Ph₂PCH₂CH₂), 4.68 (t, 8 H, CHCH₂, J = 8.0 Hz), 4.96 (s, 8 H, CHPy), 6.50 (s, 8 H, ArH), 7.10 (s, 8 H, ArH), 7.37 (d, 16 H, m-Py, AA′ part of a AA′XX′ system, J = 6.0 Hz), 7.42 [m, 48 H, P(C₆H₅)₂], 7.59 [m, 32 H, P(C₆H₅)₂], 8.87 (d, 16 H, o-Py, XX′ part of a AA′XX′ system, J = 6.0 Hz); ³¹P-NMR (81 MHz, CDCl₃) δ = −12.1, J (Pt-P) = 3120 Hz; ¹⁹F NMR (188.3 MHz, CD₂Cl₂) δ = −77.5; ESI-MS: m/z 3122.9 [M-2CF₃SO]²⁺, 2032.1 [M-3CF₃SO]³⁺, 1486.9 [M-4CF₃SO]⁴⁺ (M = C₃₀₈H₃₅₂F₂₄N₈O₄₀S₈P₈Pt₄).

3b.

(M = Pd, L = dppp, X = CF₃SO₃, R = C₁₁H₂₃): ¹H NMR (300 MHz, CDCl₃, 300 K): δ = 0.85 (t, 24 H, CH₃, J = 7.2 Hz), 1.22–1.25 [m, 128 H (CH₂)₈], 1.39 (m, 16 H, CHCH₂CH₂), 2.24 (m, 16 H, CHCH₂CH₂), 2.45 (m, 8 H, Ph₂PCH₂CH₂), 3.27 (m, 16 H, Ph₂PCH₂CH₂), 4.65 (t, 8 H, CHCH₂, J = 7.9 Hz), 4.99 (s, 8 H, CHPy), 6.46 (s, 8 H, ArH), 7.10 (s, 8 H, ArH), 7.31–7.43 [m, 64 H, m-Py, AA′ part of a AA′XX′ system and P(C₆H₅)₂], 7.56 [m, 32 H, P(C₆H₅)₂], 8.91 (bd, 16 H, o-Py, XX′ part of a AA′XX′ system); ³¹P-NMR (81 MHz, CDCl₃) δ = 8.8; ¹⁹F NMR (188.3 MHz, CD₂Cl₂) δ = −79.8; ESI-MS: m/z 2946.2 [M-2CF₃SO]²⁺, 1913.8 [M-3CF₃SO]³⁺ (M = C₃₀₈H₃₅₂F₂₄N₈O₄₀S₈P₈Pd₄).

3c.

(M = Pd, L = en, X = CF₃SO₃, R = C₁₁H₂₃): ¹H NMR (300 MHz, acetone-d₆, 300 K): δ = 0.86 (bt, 24 H, CH₃), 1.26 [m, 128 H (CH₂)₈], 1.38 (m, 16 H, CHCH₂CH₂), 2.41 (m, 16 H, CHCH₂CH₂), 3.12 (s, 16 H, H₂NCH₂CH₂NH₂), 4.78 (t, 8 H, CHCH₂, J = 7.8 Hz), 5.25 (bs, 16 H, H₂NCH₂CH₂NH₂), 5.37 (s, 8 H, CHPy), 7.17 (s, 8 H, ArH), 7.68 (s, 8 H, ArH), 8.05 (d, 16 H, m-Py, AA′ part of a AA′XX′ system, J = 6.7 Hz), 9.19 (d, 16 H, o-Py, XX′ part of a AA′XX′ system, J = 6.7 Hz); ESI-MS: m/z 2242.6 [M-2CF₃SO]²⁺, 1445.5 [M-3CF₃SO]³⁺, 1046.9 [M-4CF₃SO]⁴⁺ (M = C₂₀₈H₂₈₀F₂₄N₁₆O₄₀S₈Pd₄).

3d.

(M = Pd, L = en, X = NO₃, R = C₁₁H₂₃): ¹H NMR (300 MHz, DMSO-d₆, 300 K): δ = 0.86 (t, 24 H, CH₃, J = 6.9 Hz), 1.22 [m, 128 H (CH₂)₈], 1.37 (m, 16 H, CHCH₂CH₂), 2.39 (m, 16 H, CHCH₂CH₂), 2.73 (s, 16 H, H₂NCH₂CH₂NH₂), 4.59 (bt, 8 H, CHCH₂), 5.26 (s, 8 H, CHPy), 5.53 (bs, 16 H, H₂NCH₂CH₂NH₂), 7.20 (s, 8 H, ArH), 7.62 (s, 8 H, ArH), 8.05 (d, 16 H, m-Py, AA′ part of a AA′XX′ system, J = 5.9 Hz), 9.14 (d, 16 H, o-Py, XX′ part of a AA′XX′ system, J = 5.9 Hz).

3e.

(M = Pt, L = dppp, X = CF₃SO₃, R = CH₂CH₂Ph): ¹H NMR (300 MHz, CD₂Cl₂, 300 K): δ = 2.27 (m, 8 H, Ph₂PCH₂CH₂), 2.61 (m, 16 H, CH₂CH₂Ph), 2.68 (m, 16 H, CH₂CH₂Ph), 3.33 (m, 16 H, Ph₂PCH₂CH₂), 4.69 (t, 8 H, CHCH₂, J = 7.6 Hz), 4.84 (s, 8 H, CHPy), 6.54 (s, 8 H, ArH), 7.11–7.23 (s, 48 H, CH₂CH₂Ph and ArH), 7.36–7.40 [m, 64 H, m-Py, AA′ part of a AA′XX′ system and P(C₆H₅)₂], 7.57 [m, 32 H, P(C₆H₅)₂], 8.80 (d, 16 H, o-Py, XX′ part of a AA′XX′ system, J = 5.3 Hz); ¹³C NMR (75 MHz, CD₂Cl₂, 300 K): δ 30.0 (dppp CH₂), 30.9 (dppp CH₂), 32.0 (CH₂), 34.2 (CH₂), 36.2 (CH), 104.1 (CH), 117.0 (CH), 121.2 (q, CF₃, J = 320 Hz), 121.5 (CH), 123.6 (dppp Cq), 125.8 (CH), 126.5 (CH), 128.8 (CH), 128.9 (CH), 130.0 (dppp CH), 133.0 (dppp CH), 133.2 (dppp CH), 138.8 (Cq), 141.9 (Cq), 146.1 (Cq), 150.3 (CH), 154.2 (Cq); ³¹P-NMR (81 MHz, CD₂Cl₂) δ = −12.9, J (Pt-P) = 30,96 Hz; ESI-MS: m/z 2921.9 [M-2CF₃SO]²⁺, 1898.2 [M-3CF₃SO]³⁺, 1386.6 [M-4CF₃SO]⁴⁺ (M = C₂₈₄H₂₄₀F₂₄N₈O₄₀S₈P₈Pt₄).

3f.

(M = Pd, L = dppp, X = CF₃SO₃, R = CH₂CH₂Ph): ¹H NMR (300 MHz, CD₂Cl₂, 300 K): δ = 2.40 (m, 8 H, Ph₂PCH₂CH₂), 2.65 (m, 16 H, CH₂CH₂Ph), 2.74 (m, 16 H, CH₂CH₂Ph), 3.27 (m, 16 H, Ph₂PCH₂CH₂), 4.75 (t, 8 H, CHCH₂, J = 7.7 Hz), 4.93 (s, 8 H, CHPy), 6.57 (s, 8 H, ArH), 7.15–7.25 (s, 48 H, CH₂CH₂Ph and ArH), 7.38–7.43 [m, 64 H, m-Py, AA′ part of a AA′XX′ system and P(C₆H₅)₂], 7.59 [m, 32 H, P(C₆H₅)₂], 8.85 (d, 16 H, o-Py, XX′ part of a AA′XX′ system, J = 6.0 Hz); ³¹P-NMR (81 MHz, CD₂Cl₂) δ = 9.4; ¹⁹F NMR (188.3 MHz, CD₂Cl₂) δ = −77.3; ESI-MS: m/z 2745.3 [M-2CF₃SO]²⁺, 1780.6 [M-3CF₃SO]³⁺, 1298.1 [M-4CF₃SO]⁴⁺ (M = C₂₈₄H₂₄₀F₂₄N₈O₄₀S₈P₈Pd₄).

3g.

(M = Pd, L = en, X = CF₃SO₃, R = CH₂CH₂Ph): NMR (300 MHz, acetone-d₆, 300 K): δ = 2.82 (m, 16 H, CH₂CH₂Ph), 2.85 (m, 16 H, CH₂CH₂Ph), 3.17 (m, 16 H. H₂NCH₂CH₂NH₂), 4.89 (bt, 8 H, CHCH₂), 5.30 (bs, 16 H, H₂NCH₂CH₂NH₂), 5.45 (s, 8 H, CHPy), 7.20–7.24 (m, 40 H, CH₂CH₂Ph), 7.30 (s, 8 H, ArH), 7.83 (s, 8 H, ArH), 8.12 (d, 16 H, m-Py, AA′ part of a AA′XX′ system, J = 6.6 Hz), 9.25 (d, 16 H, o-Py, XX′ part of a AA′XX′ system, J = 6.6 Hz); ESI-MS: m/z 2042.2 [M-2CF₃SO]²⁺, 1312.1 [M-3CF₃SO]³⁺, 946.8 [M-4CF₃SO]⁴⁺ (M = C₁₈₄H₁₆₈F₂₄N₁₆O₄₀S₈Pd₄).

3h.

(M = Pd, L = en, X = NO₃, R = CH₂CH₂Ph): NMR (300 MHz, DMSO-d₆, 300 K): δ = 2.67 (m, 16 H, CH₂CH₂Ph), 2.71–2.76 (m, 32 H, CH₂CH₂Ph and H₂NCH₂CH₂NH₂), 4.71 (t, 8 H, CHCH₂, J = 6.85 Hz), 5.32 (s, 8 H, CHPy), 5.52 (bs, 16 H, H₂NCH₂CH₂NH₂), 7.15–7.22 (m, 40 H, CH₂CH₂Ph), 7.23 (s, 8 H, ArH), 7.71 (s, 8 H, ArH), 8.07 (d, 16 H, m-Py, AA′ part of a AA′XX′ system, J = 6.3 Hz), 9.14 (d, 16 H, o-Py, XX′ part of a AA′XX′ system, J = 6.3 Hz). MALDI-MS: [M-NO]⁺ (M = C₁₇₆H₁₆₈N₂₄O₄₀Pd₄) calculated 3623.09 m/z, found 3623.5 m/z.

Results

The synthesis of these deep-cavity cavitands (30–32), of general structure 2a-b, was performed by bridging the corresponding resorcinarenes with 4-(α,α′-dibromomethyl)pyridine (Scheme S1). In all cases the low yields (10–12%) of the desired (oooo) isomer are due to the competitive formation of useless isomers having one (iooo) or two (iioo, ioio) pyridyl groups pointing inward (i) the cavity (see Fig. 7, which is published as supporting information on the PNAS web site, www.pnas.org). The correct stereochemical assignment of the oooo isomer has been confirmed by the x-ray crystal structure of cavitand 2b. X-ray structure analysis reveals that the cavitand exists in two slightly different conformations. The dihedral angles between the least-squares plane through the bridging CH groups and the planes of the benzene rings of the resorcinarene skeleton [109.3(3), 123.9(3), 129.3(3), and 109.5(2) in one molecule, and 118.4(3), 116(4(3), 118.5(3), and 118.0(3) in the other] indicate that the former is in the “pinched cone” conformation and the latter is in a more regular “cone” conformation (Fig. 1 Left). In both cavitands the terminal phenethyl chains are affected by high thermal motion and static disorder.

Scheme 1.

Synthesis of the cavitand ligands 2a-b and self-assembly of cages 3a-h.

Figure 1.

(Left) X-ray crystal structure of the phenethyl cavitand 2b. (Right) The average α angle is 42°, close to the ideal 45° value for cage formation through four square planar complexes; the gray ellipsoid represents the internal cavity volume whose principal axes are a = 11.69 Å, b = 5.86 Å, c = 11.69 Å (calculated from the spartan minimized structure of Fig. 5).

From the crystal structure of the tetradentate cavitand it is possible to evaluate the α angle of the pyridyl ligands relative to the C_4v symmetry axis (Fig. 1 Right): the estimated value is 42°, close to the ideal value of 45° required for the formation of strainless square-planar complexes. Therefore the cavitand ligand is correctly preorganized for CSA.

The typical procedure for cage formation is shown in the case of cavitands 2a and 2b and various metal precursors MLX₂ (M = Pd, Pt; L = dppp, en; X = CF₃SO₃, NO₃; Scheme S1): by mixing the two components in a 1:2 molar ratio at room temperature in solvents like CH₂Cl₂, CHCl₃, acetone, or DMSO, cages 3a-h were obtained in quantitative yields. In all cases ¹H NMR spectra showed the formation of a new set of signals, indicative of the presence of a single highly symmetric compound (D_4h symmetry). The downfield shift of the pyridine protons ortho to the N is indicative of coordination to the metal, whereas up field shift of the CH pointing inside the cavity is diagnostic of cage formation. ³¹P NMR spectra exhibited sharp singlets, with appropriate Pt satellites for Pt complexes 3a and 3e, indicating the equivalence of the eight dppp phosphorus atoms, thus confirming the high symmetry of the cages. No permanent inclusion of triflate counterions was detected by ¹⁹F NMR. Additional evidence of cage formation was obtained through ESI-MS, which showed prominent [M-2X⁻]²⁺, [M-3X⁻]³⁺ for most of the cages formed; the molecular ions could not be detected because their molecular weight exceeded the limit of the instrument. In the case of 3h the MALDI-TOF spectrum has been recorded, showing the [M-NO]⁺ molecular ion at 3623.5 m/z.

Strong cooperativity in CSA has been evidenced via ¹H NMR (Fig. 2): adding 1 eq of Pt(dppp)(CF₃SO₃)₂ to a solution of cavitand 2a led to the formation of a 1:1 mixture of Pt-cage and free cavitand. After addition of the second equivalent of metal precursor the ¹H NMR spectrum showed the signals of the cage as the only product. The same behavior has been observed in the case of cavitand 2b plus Pd(en)(NO₃)₂ in DMSO-d₆ as solvent (Fig. 3).

Figure 2.

Self-assembly of cage 3a monitored by means of ¹H NMR. (a) spectral window of cavitand 2a in CDCl₃. (b) 1:1 mixture of free cavitand ligand (L) and cage (C) after addition of 1 eq Pt(dppp)(CF₃SO₃)₂. (c) Cage 3a as single product after addition of 2 eq of Pt(dppp)(CF₃SO₃)₂.

Figure 3.

Self-assembly of cage 3h monitored by means of ¹H NMR. (a) spectrum of cavitand 2b in DMSO-d₆. (b) Mixture of free cavitand ligand (L) and cage (C) after addition of slightly less than 1 eq of Pd(en)(NO₃)₂. (c) Cage 3h as single product after addition of 2 eq of Pd(en)(NO₃)₂.

The robustness of the M-Py coordination allowed us to perform CSA in polar and competitive solvents like DMSO or acetonitrile, where the nitrile-based coordination cages immediately decompose (24). The solubility properties of cages 3a-h can be tuned by appropriately choosing counterion, chelating ligand and R substituents at the lower rim of the cavitand ligand. In particular the solubility of the cages in DMSO or even in DMSO/H₂O mixtures (up to 3:1 ratio) has been achieved, introducing the highly hydrophilic NO counterion as in the case of 3d and 3h. The different solubility properties of the cage components have been exploited to perform CSA in a liquid–liquid two-phase system. In a typical experiment cavitand 2b (1 eq) dissolved in 1,1,2,2-tetrachloroethane was exposed to an aqueous solution of Pd(en)(NO₃)₂ (2.5 eq) in a test tube (Fig. 4). Immediately, a solid formed at the interface, which became more abundant over time. ¹H NMR analysis of the recovered solid indicated the formation of cage 3h in pure form. A MALDI-TOF experiment performed directly on the solid sample confirmed the attribution, ruling out the possibility of the formation of a sheet-like network at the interface.§ This last result is particularly interesting because liquid–liquid interfaces represent another potentially useful field of action for self-assembly protocols, which has barely been explored so far (33).

Figure 4.

Cage formation at the water/organic solvent interface.

Several attempts of determining the crystal structure of cages 3e,g failed repeatedly because of the total absence of any diffraction pattern, although the crystals were well formed and shaped. This led us to undertake a computational approach to study the molecular properties of the cage.

Theoretical calculations [the calculations were carried in vacuo by using spartan (PC Spartan Pro, Version 1.05, Wavefunction, Irvine, CA)] performed on 3g indicate that the cage is squeezed in the polar regions (Fig. 5), embracing a wide cavity that can be fitted by an ellipsoid of rotation whose principal axes (at the van der Waals limits) are a = 11.69 Å, b = 5.86 Å, and c = 11.69 Å (Fig. 1 Right), corresponding to a total available volume of nearly 840 ų.¶ This estimate is conservative because it does not take into account the residual space of the resorcinarene caps. The cavity can be accessed through the four large circular lateral portals, of approximate diameter of 6 Å, large enough to allow the entrance/egress of solvent molecules and triflate anions, in accordance with ¹⁹F NMR experiments. The calculated Pd-N(Py) bond distances range from 2.028 to 2.035 Å, those of Pd-N(en) from 2.017 to 2.020 Å, in agreement with the values observed in the solid state (35, 36). The torsion angles C-C-N-Pd involving the pyridine atoms (from 165 to 170°), as well as the N(Py)-Pd-N(Py) bond angles (from 85.6 to 88.8°) and those of N(en)-Pd-N(en) (from 92.7 to 95.4°), indicate a distortion in the coordination sphere at the Pd metal centers.

Figure 5.

Computer modeling of cage 3g (the phenethyl chains at the lower rim are omitted).

Several attempts to disassemble Pt and Pd cages were performed. Pd cage 3b was disassembled into its cavitand component and [Pd(dppp)(NEt₃)₂(CF₃SO₃)₂] in the presence of an excess of NEt₃ as competitive ligand (32 eq) at 323 K (Fig. 6), whereas, under the same conditions, the ¹H NMR spectrum of the corresponding Pt cage 3a remained unchanged. This finding is consistent with the greater Pt-N bond strength with respect to the Pd-N one.

Figure 6.

(a) Cage 3b in the presence of an excess of NEt₃ at 300 K. (b) Partial disassembly of the cage (C) into its cavitand component (L) after heating at 323 K for a few minutes.

The kinetic stability of the Pt cages has been assessed by means of ESI-MS.‖ A ligand-exchange experiment was performed starting from two Pt-cages 3a and 3e different for the chains at the lower rim (R = C₁₁H₂₃ and CH₂CH₂Ph): both cages were separately analyzed by means of ESI-MS, showing [M-2CF₃SO₃]²⁺, [M-3CF₃SO₃]³⁺, and [M-4CF₃SO₃]⁴⁺ ions. Then they were mixed in a 1:1 molar ratio at room temperature, giving a spectrum resulting from the simple addition of the signals owing to the two homocages. On heating the mixture, the ESI-MS spectrum exhibited a third set of peaks belonging to the heterocage in a statistic 1:2:1 ratio with the signals of the homocages (see Figs. 8–10, which are published as supporting information on the PNAS web site).

In summary the introduction of four pyridine groups at the upper rim of a cavitand ligand in a preorganized diverging orientation has allowed the self-assembly of a series of new coordination cages, embracing an internal cavity of approximately 840 ų. The robustness of the pyridine–metal coordinative bond has been exploited to perform CSA in highly polar solvents, precluded to the first generation of cavitand-based coordination cages bearing nitrile moieties (24). Kinetically the platinum cages are stable at room temperature both in polar and apolar solvents. The extension of the CSA protocol to liquid–liquid interfaces is particularly appealing in view of the possible applications to surface patterning for microanalytical devices (38).

Abbreviations

CSA

cage self-assembly

DMSO

dimethyl sulfoxide

ESI

electrospray ionization

dppp

1,3-bis(diphenylphosphino)-propane

en

ethylenediamine

MALDI-TOF

matrix-assisted laser desorption ionization–time-of-flight