Labile coodination dendrimers

Abstract

Addition of anionic benzylsulfate dendrons to dynamic mixtures of Ag⁺ and triphosphine ligands results in the assembly of loosely-bonded cage–core dendrimers.

---

We recently reported a T-symmetry cage [Ag₆(triphos)₄(anion)₄]²⁺ {triphos = CH₃C(CH₂PPh₂)₃} which assembles from six Ag⁺ ions, four triphosphines and four templating oxo anions such as NO₃⁻, ClO₄⁻ or OTf⁻.¹ Due to the ever increasing attention given to dendrimers,² we have investigated this cage as a novel dendritic core, by using dendritic benzylsulfates, ROSO₃⁻, given that their anionic potentially tripodal OSO₃⁻ groups could act as organising core templates in the presence of dynamic mixtures of Ag⁺ cations and triphosphine ligands. Here we report on these complexes, which are unusually labile examples of metallo-dendrimers, undergoing reversible dendron exchange for available SbF₆ anions (Scheme 1).

Scheme 1

G0, G1 and G2 alcohol-functionalised dendritic wedges³ were readily converted to the corresponding benzylsulfate salts [C₅H₅NH][GnOSO₃] (n = 0, 1, or 2) by reaction with pyridine sulfur trioxide. Titrations of these wedges into 6∶4 AgSbF₆/triphos mixtures in CDCl₃–CD₃CN solutions were monitored by ³¹P NMR spectroscopy, and the results of the G1OSO₃⁻ titration are shown in Fig. 1. Clearly, successive additions of one equivalent of G1OSO₃⁻ transform the dynamic mixture into a single symmetrical species with well resolved Ag–P coupling. Parallel results were obtained for G0OSO₃⁻ and G2OSO₃⁻. ³¹P NMR data for the resulting cage complexes were very similar,‡ and consistent with the previously reported Ag₆ cages with inorganic oxo anions.¹ However, for both G0OSO₃⁻ and G1OSO₃⁻, full conversion to their corresponding complexes required five rather than the stoichiometric four equivalents of the anion, and for G2OSO₃⁻ fifteen equivalents were required. The requirement for more than four equivalents suggests that the weak coordinating ability of the sulfate groups, combined with steric effects for G2OSO₃⁻, could be causing reversible coordination of the wedges at the cage core, as depicted in Scheme 1. Unfortunately, ¹⁹F NMR signals were very broad due to the quadrupolar Sb nuclei, and gave no information on whether the ′vacant′ core sites generated were, in fact, occupied by the available SbF₆ anions. Solid samples were obtained by diethyl ether precipitation of 6∶4∶6 Ag–triphos–GnOSO₃⁻ solutions.‡ The crystalline products were washed with methanol to remove any excess of GnOSO₃⁻ and pyridinium salt by-products. For G1OSO₃⁻, elemental analysis of the dried solid material, 1, suggested the overall composition [Ag₆(triphos)₄(G1OSO₃)₃(SbF₆)][SbF₆]₂·1.5CHCl₃ ,‡ i.e. containing only three wedges. Therefore, even though ³¹P NMR spectra showed the main species present in solution to be the fully saturated cage, there was clearly dissociation of G1OSO₃⁻ which led to precipitation of the less soluble ′unsaturated′ complex. Once redissolved, ³¹P and ¹H NMR spectroscopies confirmed that only three G1OSO₃⁻ units were present per cage in this material (the spectra were identical to those for the in situ prepared 6∶4∶3 mixture). ESMS spectra showed, in addition to the saturated cage, peaks for the SbF₆-substituted species [Ag₆(triphos)₄(G1SO₃)₃(SbF₆)]²⁺ and [Ag₆(triphos)₄(G1OSO₃)₂(SbF₆)₂]²⁺ (see ESI†).

Fig. 1 ³¹P NMR titration of [C₅H₅NH][G1OSO₃] into a 6:4 AgSbF₆–triphos mixture in 4∶1 CDCl₃–CD₃CN.

Crystals suitable for X-ray analysis were obtained by slow diethyl ether diffusion into a 6∶4∶4 AgSbF₆–triphos–G1OSO₃ ⁻ solution in CHCl₃–CH₃NO₂, and the molecular structure determined§ is shown in Fig. 2. The Ag₆(triphos)₄ core was well resolved, with bond lengths and angles similar to those for the OTf⁻ cage.¹ However, only two G1OSO₃⁻ units had full occupancy, each being unsymmetrically tripodally coordinated to three Ag⁺ centres with Ag–O distances in the range 2.44(2)–2.78(2) Å. Another core site unexpectedly contained a tripodal hydrogensulfate anion, HOSO₃⁻ (Ag⋯O distances in the range 2.57–2.66 Å), which apparently arose from slow hydrolysis of G1OSO₃⁻(no detectable hydrolysis occurred in wet CDCl₃–CH₃CN solutions of G1OSO₃⁻, or 6∶4∶6 AgSbF₆–triphos–G1OSO₃⁻ mixtures, but aqueous suspensions of G1OSO₃⁻ did hydrolyse to the corresponding alcohol up to 5% after 3 days). Although the hydrogen atom could not be located it is suspected to form a strong hydrogen bond to a co-crystallised diethyl ether molecule (O⋯O separation 2.29(4) Å).

Fig. 2 Molecular structures of [Ag₆(triphos)₄(G1OSO₃)₃(HOSO₃)]²⁺ (a) and [Ag₆(triphos)₄(G1OSO₃)₂(SbF₆)(HOSO₃)]²⁺ (b) showing cage cores (cylinder), dendritic groups (space filling) and triphos phenyl substituents (lines). Ag = brown, P = pink, SbF₆ = green, S = yellow, O = red.

The fourth core site contained both G1OSO₃⁻, at 50% occupancy, and SbF₆⁻, at 25% occupancy, as shown in Fig. 2 (a and b, respectively). An additional SbF₆⁻ at 25% occupancy was located some distance from the cage core, but in a position which blocked any coordination of G1OSO₃⁻ at the fourth site. The ′coordinated′ SbF₆ anion is very weakly bonded in an unsymmetrical tripodal fashion to three silver centres with the Ag–F distances being 2.46(3), 2.72(2) and 2.72(3) Å (sum of F and Ag covalent radii = 2.24 Å, sum of van der Waals radii = 3.19 Å). There are very few structurally characterised SbF₆⁻ metal complexes,⁴ and none in the Cambridge Crystallographic Database that involve coordination (even weak) to three metals. The crystal structure suggests that the large central core separates the G1OSO₃⁻ wedges sufficiently that they are unlikely to repel each other sterically, but that G2OSO₃⁻ could indeed result in steric crowding, consistent with the NMR titration described above. Although the elemental analysis, NMR spectra and the ESMS results of material 1 show that this crystal structure does not accurately represent the bulk material, it does provide a model for the dendritic complexes observed in solution, and confirms the novel aspect of dendron exchange for SbF₆⁻. Although the connectivity in the core is unambiguous, care should be taken making detailed conclusions from the rest of the structure.

Polymetallic cages previously used as dendrimer cores are Fe₄S₄(SR)₄, Ru₆Se₈L₆ and Mo₆Cl₈(OR)₆.⁵ Due to the essentially irreversible metal–ligand coordination in these cages, the resulting dendrimers are non-labile with respect to dendron binding.

We acknowledge the provision of time on DARTS, the UK national synchrotron radiation service at the CLRC Daresbury laboratory, through funding by the EPSRC, and the funding of X. X. through an EPSRC Fast Stream grant and the McClay Trust.

Notes and references

1. S. L. James, D. M. P. Mingos, A. J. P. White and D. J. Williams, Chem. Commun., 1998, 2323.
2. F. Vogtle, S. Gestermann, R. Hesse, H. Schwierz and B. Windisch, Prog. Polym. Sci., 2000, 25, 987–1041; A. W. Bosman, H. M. Janssen and E. W. Meijer, Chem. Rev., 1999, 99, 1665–1688; F. Zeng and S. C. Zimmerman, Chem. Rev., 1997, 97, 1682–1712.
3. C. Hawker and J. M. J. Frechet, J. Am. Chem. Soc., 1990, 112, 7638.
4. X-Ray crystal structures with short Ag⋯SbF₆ contacts are reported in: T. Nishinaga, T. Kawamura and K. Komatsu, Chem. Commun., 1998, 2263; M. Bertelli, L. Carlucci, G. Ciani, D. M. Proserpio and A. Sironi, J. Mater. Chem., 1997, 7, 1271; Acta Crystallogr. Sect. C, 1996, 52, 1380;they are all monodentate with Ag⋯F distances of 2.43–2.67 Å.
5. C. B. Gorman, B. L. Parkhurst, W. Y. Su and K. Y. Chen, J. Am. Chem. Soc., 1997, 119, 1141; B. R. Wang and Z. Zheng, J. Am. Chem. Soc., 1999, 121, 3549; C. B. Gorman, W. Y. Su, H. W. Jiang, C. M. Watson and P. Boyle, Chem. Commun., 1999, 877.
6. R. J. Cernik, W. Clegg, C. R. A. Catlow, G. Bushnell-Wye, J. V. Flaherty, G. N. Greaves, I. Burrows, D. J. Taylor, S. J. Teat and M. Hamichi, J. Synchrotron Radiat., 1997, 4, 279–286; W. Clegg, M. R. J. Elsegood, S. J. Teat, C. Redshaw and V. C. Gibson, J. Chem. Soc., Dalton Trans., 1998, 3037 ; SMART version 5.054, Bruker AXS Inc., Madison, WI, USA, 1998; J. Cosier and A. M. Glazer, J. Appl. Crystallogr., 1986, 19, 105; SAINT version 6.02a, Bruker AXS Inc. Madison, WI, USA, 2000; G.M. Sheldrick, SADABS, program for scaling and correction of area detector data, University of Gottingen, 1997; G.M. Sheldrick, SHELXTL, version 5.10, Bruker AXS Inc., Madison, WI, USA, 2000.

---

Footnotes

† Electronic supplementary information (ESI) available: ESMS spectrum of 1 and dendron NMR data. See http://www.rsc.org/suppdata/cc/b1/b107658m/

‡ ³¹ P NMR (121.47 MHz, CDCl₃–CD₃NO₂, 298 K), δ_P/ppm and ¹J(¹⁰⁹Ag–³¹P)/Hz: [Ag₆(triphos)₄(G0OSO₃)₄]²⁺ −6.56, 540; [Ag₆(triphos)₄(G1OSO₃)₄]²⁺; −6.47, 541; [Ag₆(triphos)₄(G2OSO₃)₄]²⁺ −6.40, 540. Preparation and characterising data for {Ag₆(triphos)₄(G1OSO₃)₃(SbF₆)₃} 1: to a clear solution of AgSbF₆ (103 mg, 0.3 mmol) and triphos (125 mg, 0.2 mmol) mixture in 1∶1 CH₃NO₂–CHCl₃ (2 ml) was added a solution of [C₅H₅NH][G1OSO₃] (144 mg, 0.3 mmol) in CHCl₃ (1.5 ml). The product was crystallised by vapour diffusion of diethyl ether into this mixture over 2 days. The off-white crystalline solid was collected, washed with methanol and dried in vacuo overnight. Yield: 185 mg, 73%. Microanalysis for C₂₂₇H₂₁₃F₁₈O₁₈P₁₂S₃Ag₆Sb₃·1.5CHCl₃: found (calc.) C 52.4 (52.5), H 4.0 (4.1), S 2.4 (1.8), Cl 2.9 (3.0)%. ³¹P{H} NMR (121.47 MHz, CDCl₃, 298 K): δ −5.72 (dd, ¹J(¹⁰⁹Ag–³¹P) 541 Hz. Positive FAB (m/z): [Ag₆(triphos)₄(G1SO₃) ₃(SbF₆)]²⁺, 2290. Crystals for X-ray analysis were grown in a similar way over 3 weeks.

§ Crystal data for 2: C_219.45H_213.05Ag₆F_15.30O_{1 9.70}P₁₂S_3.45Sb_2.55, M = 4896.18, monoclinic, space group P2₁/n, a = 27.4440(12), b = 25.3852(11), c = 34.8363(15) Å, U = 22722.7(17) ų, T = 150(2) K, Z = 4, μ = 9.91 mm⁻¹, Synchrotron radiation (CLRC Daresbury Laboratory Station 9.8, λ = 0.6923 Å) Bruker AXS SMART CCD diffractometer,⁶ narrow frame ω rotations, and were corrected semiempirically for absorption and incident beam decay (transmission 0.72–0.87), 53647 reflections collected, 19642 unique (R_int = 0.0905). The final R₁ = 0.1305 (for 11319 reflections I > 2σ(I)), wR₂ = 0.3440 for 1209 parameters. The core of the molecule was well resolved, allowing anisotropic refinement of the Ag, P, S, Sb, O (attached to S) and C (linking the P) atoms. All other Sb atoms were also refined anisotropically. The isotropic displacement parameters of the rest of the molecule reflect the inherent disorder in the system, which lead to the data only being significant to 2θ_max = 38°, even with synchrotron radiation. In addition one peripheral phenyl group in the partial G1SO₃⁻ could not be found. CCDC reference number 166377. See http://www.rsc.org/suppdata/cc/b1/b107658m/ for crystallographic data in CIF or other electronic format.

---