Metallosupramolecular Polygons and Polyhedra:
Gas-Phase Chemistry and Surface Patterning
Since Fujita and coworkers have reported a metallo-supramolecular square in the late 1980s, a wide variety of polygons and polyhedra with different metal-containing and organic subunits were prepared by self-assembly a nd investigated with different, complementary methods such as NMR, IR, UV/Vis spectroscopy, cyclic voltammetry, X-ray crystal structure analysis, vapor phase osmometry, or mass spectrometry.
In our group, polygons of Fujita and Stang-type were investigated by electrospray-ionization Fourier-transform ion-cyclotron-resonance (ESI-FTICR) mass spectrometry and - together with Dr. Peter Broekmann at the University of Bonn - electrochemical scanning tunneling microscopy (EC-STM).
1. MS-Investigation of Metallo-Supramolecular Squares and Triangles
We have chosen polygons of the Stang-type for MS-investigation because they are highly soluble in organic media and they are easily charged by the loss of one or more counter-anions in the electrospray process (spray solvent: acetone). The following three examples show that MS can not only be used to detect the molecular weight and elemental compositions, but also get a profound insight into the particles reactivity in solution and in the gas phase.
Figure 1: Self-assembling metallosupramolecular squares.
1.1. Square Triangle Equilibrium in Solution
When the organic ligands are flexible enough as in the case of azopyridine (Figure 2), a square-triangle equilibrium is found in solution. This equilibrium is of course dependent on the solvent and the temperature. While different time-consuming NMR-techniques are needed to get insight into this equilibrium, it can be monitored by MS easily, because of the different masses of the involved complexes. One limitation is that the intensities observed in the mass spectra do not necessarily quantitatively reflect solution concentrations.
Figure 2: Square-triangle equilibrium in solution. The flexibility of the ligand is decisive.
1.2. Neighboring-Group Effect in the Gas Phase
The mechanism of fragmentation is studied in more detail by tandem-MS experiments: The triply or quintuply charged squares are mass-selected in the FTICR cell and irradiated with the beam of a CO2 laser in the IR range. Fragmentation reactions can then be examined without interference of ligand exchange processes prevailing in solution. To our surprise, squares of four bipyridines and 4 metal corners (4:4 complexes) do not decompose into two 2:2 complexes but exclusively into a 3:3 and a 1:1 complex. Both theoretical fragmentation pathways require the cleavage of two Pt-N-bonds and should thus be equally energy demanding.
Figure 3: Backside-attack mechanism which rationalizes the formation of 3:3- and 1:1-fragments. The square in its +5 charge state undergoes a similarly selective reaction into a +4 triangle and a +1 1:1-complex of ligand and corner.
Figure 3 illustrates schematically, how a supramolecular analogue of a “neighboring-group effects” might explain these findings. Backside attack of a non-coordinated pyridine nitrogen atom leads to the formation of a triangle c oncomitant with the loss of a 1:- fragment (top). Instead, a similar attack at the other central metal center (bottom) is geometrically not feasible.
1.3. Ligand exchange between two different squares in Solution
When two different Pd-squares are mixed, an ESI mass spectrum recorded after ca. 30 sec shows five signals for all different mixed forms in a close-to-statistical ratio of 1:4:6:4:1. Consequently, the ligand exchange within Pd squares is so fast that only the equilibrium situation can be observed. In marked contrast, the analogous Pt squares exchange ligands much slower (Figure 4) and reach the equilibrium only after 2 days.
Figure 4: Slow solution-phase exchange of ligands between two different Pt squares can be monitored by mass spectrometry.
2. Polyhedra: Anionic Tetrahedral Host-Guest Complexes
Together with Dr. Iris Oppel at the Ruhr-Universität Bochum, the tetrahedon shown in Figure 5 was investigated by mass spectrometry. The assembly of the tetrahedral cage requires the presence of a templating guest cation such as tetraethyl ammonium. Instead, also triethyl ammonium can be used. Isotopic pattern analysis revealed a water molecule to be co-encapsulated inside the metallo supramolecular tetrahedron together with the smaller triethyl ammonium cation. With this information from mass spectrometry, it could indeed be identified in the crystal structure. The presence of the water molecule exclusively in the tetrahedra filled by triethyl ammonium clearly indicates the tetrahedron a) to be structurally intact after ionization and b) to have a fully closed surface through which not even a water molecule can easily escape.
Figure 5: An M6L4 tetrahedron with completely closed surface. Each ball in the schematic representation (left) corresponds to a Cd-O-Cd-O four-membered ring connecting two faces. Structure of the ligand and coordination pattern (center), crystal structure in space filling representation (right).
3. Electrochemical Scanning Tunneling Microscopy (EC-STM)
The symmetry of a negatively charged chloride layer on top of a Cu(100) electrode is dictated by the copper substrate (first-order template effect, Figure 6). Water soluble and positively charged Fujita-type polygons were forced by strong charge-charge attractions to adsorb parallel to the surface with the cavity open towards the electrolyte (second-order template effect, Figure 6). With this technique, sub-molecular resolution STM images can be obtained at the solid liquid interface.
Figure 6: Schematic representation of the structure-determining effects controlling the deposition of positively charged metallo-supramolecular squares on a chloride-covered Cu(100) electrode.
Among others, two assemblies (Figure 7) were deposited by this technique: achiral squares and chiral rhombs.
Figure 7: Squares (left) and chiral rhombs (right), which have been deposited on chloride-covered Cu(100) surfaces.
Large perfectly ordered domains of the squares are observed (Figure 8), which are oriented parallel to the chloride-stabilized step edges and thus parallel to the main symmetry axes of the chloride layer. With enantiomerically pure metallo-supramolecular rhombs, an adlayer is formed on top of the chloride layer, again. Laterally ordered patches of bright STM spots (individual rhombs) are clearly visible in Figure 9, left. From the sub-molecular resolution presented in Figure 9, center, the rhombic shape of the metal complex becomes evident. With a racemic mixture of these rhombs, two domains were observed in the STM experiment. The presence of surface leads to a spontaneous separation of the enantiomers, resulting in enantiomerically pure domains (Figure 9, right).
 |
 |
Figure 8: Large domains of laterally ordered squares can be deposited.
Figure 9: Domains of chiral rhombs. Left: Pure (S,S,S,S) enantiomer. Lines indicate main axes of the chloride and the rhomb layers. Center: Sub-molecular resolution of rhombs. Right: Spontaneous separation of the racemate in two enantiomeric domains.