A Self-Assembled Electro-Active M8L4 Cage Based on Tetrathiafulvalene Ligands

Two self-assembled redox-active cages are presented. They are obtained by coordination-driven self-assembly of a tetra-pyridile tetrathiafulvalene ligand with cis-M(dppf)(OTf)2 (M = Pd or Pt; dppf = 1,1′-bis(diphenylphosphino)ferrocene; OTf = trifluoromethane-sulfonate) complexes. Both species are fully characterized and are constituted of 12 electro-active subunits that can be reversibly oxidized.


Introduction
The coordination-driven self-assembly methodology has proven during the last decades to be very efficient for the preparation of discrete polygons (2D) or polyhedra (3D) otherwise challenging to synthesize through conventional covalent multi-step synthesis [1][2][3][4][5][6][7][8][9][10][11][12][13]. This approach involves a coordination process between a rigid organic ligand and a metal center of complementary geometry, to provide thermodynamically stable discrete assemblies in high yields. In particular, nitrogen-based binding sites such as substituted pyridines are often used in presence of square planar Palladium (II) or Platinum (II) salts.
Herein, we present an original example of coordination-driven self-assembled architectures involving the parent TTF redox-active framework and cis-blocked Pd(II) and Pt(II) salts.
Single crystals of ligand L1 were obtained by slow diffusion of hexane in a solution of L1 in dichloromethane. The XRD analysis (Tables S1) revealed that the TTF skeleton and the four nitrogen atoms are coplanar ( Figure 1a) and set in a rectangle defined by a length of 13.1 Å and a width of 6.7 Å (Figure 1b). The electro-deficient pyridile plans are highly rotated related to the TTF backbone (i.e., between 31° and 77°) and are therefore poorly conjugated to the donating part of the molecule. Consequently, the π-donating ability of ligand L1 should be only moderately altered compared to the parent TTF.

Assemblies 1 and 2: Synthesis and Characterization
Ligand L1 was engaged in a self-assembly process with cis-M(dppf)(OTf) 2 (M = Pt or Pd) complexes in nitromethane at 50 °C and the reaction was monitored by 1 H NMR (Figures 2, S1). Several discrete architectures are potentially expected from this reaction, according to the number of ligand L1 involved in the final structures. In both cases, the reaction converged to one unique species 1 (Pt complex) or 2 (Pd complex) in 2 h and 5 min respectively. The self-assembled discrete structures could be isolated by precipitation from Et 2 O. . This behavior is expected and has already been reported for similar compounds [46]. One singlet is observed for each of the 19 F and 31 P NMR spectra of 1 and 2 ( Figures S4, S5, S9, S10), and the corresponding DOSY NMR spectra exhibit a single alignment of signals (Figures 2e, S11). All these data are in agreement with the formation of a unique species during the self-assembly reaction. Remarkably, both self-assemblies present the same diffusion coefficient in solution (D ≈ 2.2 × 10 −10 m² •s −1 ) extracted from the DOSY experiments. This result indicates that both species are of similar size with an estimated hydrodynamic radius of ca. 16.5 Å calculated from the Stokes-Einstein equation [54]. This value is in line with the formation of a large discrete self-assembled architecture but does not allow discriminating between a M 6 L 3 or a M 8 L 4 species.
High resolution ESI mass spectrometry experiments were carried out from dichloromethane solutions of 1 (Figure 3 and Figure S13) and 2 ( Figure S14  Depending on the orientation of the ligands inside the structure, two types of geometry are conceivable for the M 4 L 8 assemblies (i.e., the rectangular ligands are connected to the metal centers through their short (Scheme 2A) or long side (Scheme 2B). Molecular force field (MM+) studies were undertaken to determinate the most probable structure. Geometry optimization could not yield a symmetric stationary point for the B assembly (Scheme 2B) due to a high structural strain. In contrast, a highly symmetric assembly could be reached for the geometry A ( Figure 4). As expected, the structural characteristics of the tetra-pyridyl ligand within the metalla-assemblies are close to those of free L1 ( Figure 1) and a distance of 16 Å is found between two facing TTF plans. A total diameter of 36 Å was found for complex 1 in reasonable accordance with the value determined from the DOSY NMR experiment.

Redox Properties
Redox properties of ligand L1, complexes 1 and 2 were studied by cyclic voltammetry ( Figure 5). As usually observed with TTF derivatives, two reversible oxidation waves are observed for ligand L1 and are assigned to the successive generation of the cation-radical and dication states. A lower π-donor ability is found compared to the parent unsubstituted TTF, with a first redox potential which is shifted by +230 mV. This behavior is ascribed to the presence of four electro-deficient pyridile units on the periphery of the TTF backbone. Three reversible redox processes are observed in the case of the self-assembled metalla-cages 1 and 2. The first two are assigned to the TTF side walls and the third one to the corner ferrocene units. Interestingly, the latter can be used as an internal reference to address the number of electrons exchanged along the redox processes. The relative intensities collected from the deconvoluted cyclic voltammogram (Figure 5b) agree with a successive one, one, two electrons oxidation sequence. Since the assembly encompasses four TTF and eight ferrocene redox units, such observation indicates that full oxidation at a potential of 1.0 V vs Fc/Fc + leads to the reversible generation of sixteen positive charges in addition to those centered on the eight M(II) complexes. Finally, it has to be mentioned that the CV shape remains unchanged upon several scanning between 0.0 V and 1.0 V vs Fc/Fc + . Nevertheless, a repeated cycling to the upper limit of potentials results in a progressive decreasing of the redox waves intensities, in particular with the Pd derivative, which we attribute to a passivation of the electrode. Therefore, this stability and the possibility which is offered by such systems to control the charge (up to sixteen) on the periphery of the cavity, open promising perspectives in terms of guest binding and transport.

Instrumentation
The 300.3 ( 1 H), 75.5 ( 13 C), 121.6 ( 31 P) and 282.6 MHz ( 19 F) NMR spectra were recorded at room temperature using perdeuterated solvents as internal standards ( 1 H), external H 3 PO 4 solution ( 31 P) and CFCl 3 ( 19 F), on a NMR Bruker Avance III 300 spectrometer (Bruker, Rheinstetten, Germany). MALDI-TOF-MS spectra were recorded on a MALDI-TOF Bruker Biflex III instrument using a positive-ion mode. ESI-MS spectra were achieved on a Bruker MicrO-Tof-Q 2 spectrometer in CH 2 Cl 2 . Cyclic voltammetry experiments were carried out on an ALS electrochemical analyzer model 660 and the conditions were the following: 0.1 M nBu 4 NPF 6 in acetonitrile or acetronitrile/methylene chloride (1/1 v/v), Ag/Ag + reference electrode, GC or Pt working electrode, and Pt counter electrode, calibrated using internal ferrocene. Elemental analyses were achieved on a Thermo Electron analyzer.

Single-crystal X-ray Crystallography
X-ray single-crystal diffraction data were collected at low temperature on a Bruker KappaCCD diffractometer, equipped with a graphite monochromator utilizing MoKα radiation (λ = 0.71073 Å). The structure was solved by direct method and refined on F 2 by full matrix least-squares techniques using SHELX97 (Programs for Crystal Structure Analysis (Release 97-2). Sheldrick, G.M. [56]) package. All non-hydrogen atoms were refined anisotropically. Absorption was corrected by SADABS program (Sheldrick, Bruker, 2008). The H atoms were found by Fourier difference map. CCDC reference number CCDC 951195 (L1) contains the supplementary crystallographic data for this paper. These data can be obtained free of charge from The Cambridge Crystallographic Data Centre [57].

Molecular Modeling
Molecular modeling was performed by using the molecular mechanics force field MM+ method from the HyperChem 8.0.3 program (Hypercube, Inc., Waterloo, ON, Canada,) configured in vacuo, with a RMS of 10 −5 kcal/mole, a number of maximum cycles of 32500, and a Polak-Ribiere algorithm. Counter anions were omitted to simplify the calculation.

Experimental Procedure and Characterization Data
To a suspension of palladium acetate (82 mg, 0.36 mmol), tri-tert-butylphosphine tetrafluoroborate (320 mg, 1.10 mmol) and cesium carbonate (2.40 g, 7.30 mmol) stirred for 10 min at 90 °C under argon in distilled dioxane (20 mL) was added an argon degassed solution of tetrathiafulvalene (300 mg, 1.46 mmol) and 4-iodopyridine (1.50 g, 7.34 mmol) in dioxane (20 mL). The reaction was stirred under reflux for 24 h. After cooling, a large excess of dichloromethane and water were added. The aqueous phase was extracted and the organic extracts were washed with brine, dried over magnesium sulfate, filtered and concentrated. The residue was purified by chromatography on silica gel (deactivated with triethylamine 1%) eluting from dichloromethane to dichloromethane/methanol (97/3 v/v) to give a red powder (530 mg, 1.03 mmol, 71%). Crystals (dark red needles) were obtained by slow diffusion of hexanes in dichloromethane. Melting Point: >260 °C. 1

Conclusions
The synthesis and the characterization of two coordination-driven self-assembled cages is depicted. They are constructed from an electron-rich TTF based ligand L1 and an electro-active cis-M(dppf)(OTf) 2 (M = Pd or Pt) complex. They exhibit an internal cavity of 16 Å in diameter, which is surrounded by four redox-active TTF and eight ferrocene units. All of them can be oxidized, offering the possibility to reversibly generate up to sixteen positive charges on the assembly. This full control over the charge state of the cavity opens very promising perspectives, in particular for answering the key question of the triggering of the guest binding.