Homonuclear and heteronuclear complexes of a four-armed octadentate ligand: synthetic control based on matching ligand denticity with metal ion coordination preferences

Abstract

The octadentate ligand 1,2,4,5-tetrakis-[3-(2-pyridyl)pyrazol-1-yl]benzene (L), with four bidentate arms radiating from a central phenyl ring, combines with 6-coordinate and 4-coordinate metal ions (as their tetrafluoroborate salts) in different ways according to the coordination number preferences of the metal ions. The four bidentate arms are not ideally matched to the requirement of octahedral metal ions, such that complexes with Cd(II), such as tetranuclear [Cd₄L₂(μ-F)₂(solv)₄](BF₄)₆ (solv = MeOH or MeCN), and Co(II)/Ni(II), such as hexanuclear [Co₆L₄(μ-F)₂][BF₄]₁₀, require monodentate ancillary ligands (solvent molecules or fluoride ions) to provide coordinative saturation. In contrast, a mixture of octahedral [M = Co(II) or Ni(II)] and tetrahedral [Ag(I)] metal ions reacts with L to afford the simpler trinuclear heterometallic complexes [Co₂AgL₂](BF₄)₅ in which the requirements of the metal ions (for 3 + 3 + 2 bidentate arms) are matched by two four-armed ligands. The maximum site occupancy principle can accordingly be used to direct self-assembly of heterometallic complexes.

---

Introduction

The self-assembly of structurally elaborate polynuclear metal complexes from relatively simple bridging ligands and labile metal ions has been a major feature of research in inorganic chemistry over the last two decades,¹ since Lehn described the first systematic studies into double helical complexes of Cu(I) and Ag(I) with oligo-bipyridine ligands.² More recent examples of such complexes include grids,³ metallomacrocycles,⁴ and a wide variety of coordination cages based on polyhedral arrays of metal ions.⁵

Implicit in all of this work is the principle that a match between the number of donor atoms available in the ligand array, and the coordination number of the metal ions used, is an essential part of the design of the system. Ligands containing a series of bipyridyl-type binding sites give two-stranded double helicates with four coordinate metal ions such as Cu(I) and Ag(I),² but give three-stranded triple helicates with six-coordinate metal ions.⁶ Piguet's triple helical lanthanide(III) arrays rely on the metal ions being nine coordinate and hence binding three tridentate ligand fragments from separate ligand strands.⁷ In all of the examples in ref. 1–6, and many others, in cases where the complexes are homoleptic—that is, they contain only one type of ligand—the final assembly takes account of the principle of ‘maximum site occupancy’.^1d This favours, for simple thermodynamic reasons, formation of coordinatively saturated complexes in which all ligand binding sites are used: as mentioned above this requires a match between the number of binding sites in the ligand donor array and in the metal ions’ (acceptor) array.

This principle provides interesting opportunities to prepare mixed-metal systems in which the donor set provided by the ligand array cannot be exactly matched by one type of metal ion alone, but is better matched by a combination of two metal ions having different coordination number preferences (e.g. six-coordinate and four-coordinate). This results in a self-assembly process that is one step more sophisticated than those described earlier, in that it is based on three components rather than two. Examples of this include assembly of double helicates in which a ligand contains distinct tridentate and bidentate binding sites, such that the pair of ligands in the helical array provides binding pockets that are optimal for six-coordinate and four-coordinate metal ions respectively;⁸ in these cases, combination of the ligands with two types of metal ion leads to specific formation of the mixed-metal helicate in which the principle of maximum site occupancy is perfectly obeyed without needing ancillary ligands such as solvent molecules or counter ions. Additional examples are provided by some dinuclear triple helicates which contain a lanthanide(III) ion in a 9-coordinate compartment and a (kinetically inert) transition metal ion in a six-coordinate environment.⁹

The potential value of this is that careful matching of polydentate ligands with two types of metal ions could, in principle, allow two types of metal-centred function to be combined in one assembly: for example octahedral and tetrahedral metal ions may have quite different magnetic or luminescence properties, and a polynuclear assembly containing both may be a more appealing target from a functional point of view than an assembly containing just one type of functional centre.

In this paper, we report studies into the coordination chemistry of the ligand L (Scheme 1), which contains four bidentate chelating pyrazolyl-pyridine sites and is therefore octadentate. As such it is not a simple match for metal ions that adopt a six-coordinate geometry as it would require an M₄L₃assembly on the basis of ‘maximum site occupancy’. We show how complexes of L with six-coordinate metal ions result in polynuclear assemblies that require additional ancillary ligands to complete the coordination around the array of metal ions, but reaction of L with a combination of six-coordinate and four-coordinate metal ions results in coordinatively saturated trinuclear mixed-metal complexes based on the principal of maximum site occupancy. The ligand L has been reported before by McMorran and Steel who described the dinuclear complex [{Zn(O₂CMe)}₂(μ-L)](BPh₄)₄, in which the ligand donates a pair of bidentate arms to each of two Zn(II) ions, whose six-fold coordination spheres are completed by a bidentate acetate ligand in each case.¹⁰

Results and discussion

Homometallic complexes

Since the ligand L has eight donor atoms, a perfect match with a six-coordinate metal ion would require 1.333 metal ions per ligand, i.e. a M₄L₃ ratio if the complex is homoleptic. The reaction of L with Cd(BF₄)₂ in these proportions under solvothermal conditions in MeOH produced X-ray quality crystals which proved, on crystallographic analysis, to be the tetranuclear complex [Cd₄L₂(MeOH)₄(μ-F)₂](BF₄)₆ (Fig. 1 and 2). The complex has inversion symmetry such that only half of it is crystallographically unique. The structure consists of a near-rectangular array of four Cd(II) ions, capped above and below by two ligands L, each of which donates one bidentate arm to each metal ion. Both ligands therefore bind to all four Cd(II) ions, with a configuration in which two of the pyrazolyl-pyridine arms are approximately parallel to one another and almost perfectly eclipsed, with an separation between them varying from 3.3–3.8 Å, and the other two pyrazolyl-pyridine arms are diverging from each other. The two ligands L accordingly provide between them four donor atoms to each Cd(II) ion, whose octahedral coordination sphere is completed by a terminal MeOH ligand and a bridging fluoride ion which links two Cd(II) ions along the short edge of the Cd₄ rectangle, 4.01 Å apart.

Fig. 1 Structure of the complex cation of [Cd₄L₂(MeOH)₄(μ–F)₂](BF₄)₆·6MeOH showing the atomic labelling scheme. One ligand L is shown with hollow bonds for clarity. H atoms are omitted for clarity.

Fig. 2 Alternate view of the complex cation [Cd₄L₂(MeOH)₄(μ-F)₂](BF₄)₆·6MeOH with the two (crystallographically equivalent) ligands coloured separately for clarity. H atoms are omitted for clarity.

The methylene protons of the ligands are involved in some non-covalent interactions, which are facilitated by the rather folded conformation of the complex. One H atom from each of C(26) and C(46) is directed over the face of the central phenyl ring of the other ligand [C(91)–C(96)], giving CH⋯π interactions¹¹ with separations between the H atoms and the mean plane of the aromatic ring being 2.78 Å [for H(26B)] and 2.81 Å [for H(46A)]. Similarly, one H atom from each of the methylene groups C(66) and C(86) is directed towards the bridging fluoride ion F(1), with H⋯F separations of 2.49 Å [from H(86A)] and 2.59 Å [from H(66B)]. The corresponding C⋯F separations are 3.18 and 3.25 Å, respectively, with C–H⋯F angles of 126.1° and 124.2°, and these interactions constitute CH⋯F hydrogen bonds.¹² Finally, the H atoms H(16) and H(36), ortho to the pyridine N atoms at the H⁶ positions of the pyridyl rings, are directed towards the oxygen atoms of the adjacent MeOH ligands, giving H⋯O separations of 2.62 and 2.63 Å for H(16)⋯O(101) and H(36)⋯O(111), respectively. The corresponding C–H⋯O angles are 122.5 and 123.4° respectively, and the C⋯O separations are 3.24 and 3.25 Å, respectively, consistent with the presence of weak CH⋯O hydrogen bonding interactions.¹²

From the point of view of stoichiometry it is obvious that the two octadentate ligands, providing 16 donor atoms, do not match the requirement of the four Cd(II) ions for a set of 24 donor atoms: the coordination spheres of the metal ions are completed by the four terminal MeOH ligands and two bridging fluoride ligands, between them occupying the remaining eight coordination sites.

The structural integrity of this complex in solution is confirmed by electrospray mass spectrometry and NMR spectroscopy. The ES mass spectrum shows a series of peak clusters centred at 1124.3, 720.9, 518.6 and 317.1 corresponding to {[Cd₄L₂(μ-F)₂](BF₄)_6−n}ⁿ⁺ (n = 2, 3, 4, 6), i.e. the intact tetranuclear complex cation (minus the labile MeOH ligands) with varying numbers of associated tetrafluorborate anions. In every case the isotope pattern is consistent with the formulation, and the separation between adjacent peaks in the isotope cluster is commensurate with the charge (1/n units for an n⁺ species).

The ¹H NMR spectrum (Fig. 3) shows 17 signals, each corresponding to 4 H, that can be ascribed to the ligand L, consistent with both ligands being equivalent (as per the crystal structure) and also with each ligand having bilateral symmetry in solution such that only half of each one is magnetically unique. This is easy to understand from examination of the structure: each ligand will have in solution a plane of symmetry bisecting positions C¹ and C², and C⁴ and C⁵, of the central phenyl spacer, with the ‘inward pointing’ and ‘outward pointing’ pyrazolyl-pyridine arms being inequivalent. The two independent sets of four protons from the two pyridyl rings (labelled 3/4/5/6 and 3′/4′/5′/6′ corresponding to the proton positions), and two pairs of pyrazolyl protons (labelled ‘A’ and ‘B’), are obvious from a COSY spectrum, as are the four CH₂protons which split into two coupled pairs (labelled ‘a’ and ‘b’). It is noticeable that one of the methylene protons of the ‘a’ pair is at a significantly higher chemical shift (6.85 ppm) than the other three (between 4.4 and 5.7 ppm), and we ascribe this to the CH⋯F hydrogen-bonded proton which will be deshielded more than the others. Similarly, the two resonances from the inequivalent pyridyl H⁶protons are in very different positions, consistent with the crystal structure which shows that one type of H⁶proton lies above the aromatic ring of a cis-related pyridyl ligand, whereas the other is directed towards a coordinated solvent molecule (cf.Fig. 2).

Fig. 3 Structure of the complex cation of [Cd₄L₂(MeCN)₄(μ-F)₂](BF₄)₆·8MeOH showing the atomic labelling scheme. One ligand L is shown with hollow bonds for clarity. H atoms are omitted for clarity.

¹⁹F NMR spectroscopy (Fig. 4) is also helpful, showing two well-separated signals, corresponding to the six free (BF₄)⁻ ions (sharp signal at −152 ppm) and the two coordinated, bridging fluorides (broad signal at −212 ppm), with the expected integral ratio of ca. 12:1. Actually the sharp ¹⁹F signal for the free (BF₄)⁻ ions displays two very closely spaced components corresponding to the two isotopes of boron in their natural abundance. Finally the ¹³³Cd NMR spectrum shows only a single peak at −567 ppm, consistent with the fourfold symmetry of the complex in solution as shown by the ¹H NMR data. Overall, the NMR and mass spectrometric data convincingly confirm that the structure of the molecule as seen in the solid state is retained in solution, apart from possible exchange of the labile coordinated solvent molecules. In fact, a repeat of the synthesis using MeCN in the solvothermal synthesis instead of MeOH afforded [Cd₄L₂(MeCN)₄(μ-F)₂](BF₄)₆, an essentially identical structure in which the MeOH ligands are replaced by MeCN (see Fig. 5).

Fig. 4 500 MHz ¹H NMR spectrum of [Cd₄L₂(MeOH)₄(μ-F)₂](BF₄)₆ in CD₃NO₂. The lower case labels ‘a’ and ‘b’ denote the coupled pairs of diastereotopic methylene protons; the upper case labels ‘A’ and ‘B’ denote the coupled pairs of pyrazolyl ring protons; and the labels 3/4/5/6 and 3′/4′/5′/6′ denote protons H³–H⁶, respectively, on the two inequivalent types of pyridyl ring. The inset shows an expansion of the three overlapping signals close to 8.0 ppm.

Fig. 5 400 MHz ¹⁹F NMR spectrum of [Cd₄L₂(MeOH)₄(μ-F)₂](BF₄)₆ in CD₃NO₂ showing (left) the free tetrafluoroborate anions (with two components corresponding to the ¹¹B and ¹⁰B isotopomers), and (right) the broad signal associated with the two bridging fluorides each bridging two Cd(II) centres. The relative integrals of the two signals are ca. 12 : 1, as required. Note that the resonance at −212 ppm has been magnified in intensity for clarity.

The reaction of L with M(BF₄)₂ (M = Co, Ni) in the same proportions under solvothermal conditions afforded crystals of what proved to be the isostructural hexanuclear complexes [M₆L₄(μ-F)₂](BF₄)₁₀ (Fig. 6 and 7). The Ni(II) complex gave a better refinement so the discussion is focused on that, although the Co(II) complex is essentially identical.

Fig. 6 Structure of the complex cation of [Ni₆L₄(μ-F)₂](BF₄)₁₀ showing the atomic labelling scheme. Two of the four ligands L are shown with hollow bonds for clarity. H atoms are omitted for clarity.

Fig. 7 Two additional views of the complex cation of [Ni₆L₄(μ-F)₂](BF₄)₁₀ with the ligands coloured separately for clarity. View (a) shows half of the complex cation, i.e. a {Ni₃L₂} with the associated fluoride bridges to the other half of the hexanuclear complex (similar but not crystallographically equivalent); view (b) shows the entire hexanuclear complex cation. H atoms are omitted for clarity.

The structure of [Ni₆L₄(μ-F)₂](BF₄)₁₀ consists of two very similar (but not crystallographically identical) {Ni₃L₂}⁶⁺ halves connected by a pair of bridging fluoride ions [emphasised in Fig. 7(a)], which we assume to have arisen from decomposition of some of the tetrafluoroborate anions. In each {Ni₃L₂}⁶⁺ unit, one ligand (shown in red) coordinates its four bidentate arms to two Ni(II) centres, with two adjacent arms to each Ni(II): thus it coordinates in a ‘bis-tetradentate’ manner. The pairs of metal ions concerned are separated by 8.65 Å [Ni(1) and Ni(2)] and 8.83 Å [Ni(5) and Ni(6)]. The other ligand (shown in blue) donates one bidentate arm to each of these two Ni(II) ions, thereby completing the octahedral coordination around each, and coordinates the remaining two arms to the third Ni(II) ion. The remaining two coordination sites on this Ni(II) ion are occupied by two fluoride bridges, which connect two such units together to give the complete hexanuclear structure.

A significant feature of the structure is the interleaving of planar aromatic sections of the two ligands, which results in the phenyl ring of the blue ligand being sandwiched between two coordinated pyridyl rings of the red ligand (Fig. 7). The stacking distance between adjacent parallel rings is ca. 3.5 Å, and the stacked rings are offset such that some of the H atoms of one ring lie above the centre of the other, which is the optimum arrangement for maximising the electrostatic contribution to π-stacking.¹¹ The three-component stack (pyridyl-phenyl-pyridyl) has alternating electron poor/electron rich/electron poor character, with the pyridyl rings (coordinated to 2+ metal ions) on the outside of the stack being considerably more electron deficient than the central phenyl ring with four alkyl substituents. Thus the stack is an acceptor/donor/acceptor array, as we have commonly observed in polyhedral cages based on ligands of this general class.^5a,13

The stoichiometry of this complex is therefore simple to rationalise: in each half, three metal ions require 18 donor atoms but two octadentate ligands provide only 16, and the consequence of this is that two ancillary ligands are required to complete the coordination around all of the Ni(II) ions. These could be monodentate terminal ligands (e.g.solvent molecules) but happen in this case to be fluorides, which act as bridges, connecting two trinuclear units together; the Ni(3)–Ni(4) separation across the fluoride bridges is 3.10 Å.

NMR and ESMS studies confirmed the structural integrity of these complexes in solution. For both [M₆L₄(μ-F)₂](BF₄)₁₀ (M = Co, Ni) the highest m/z peak in the ES mass spectrum corresponded to {[M₆L₄(μ-F)₂](BF₄)₇}³⁺ (M = Ni, observed at m/z 1274.8, calculated 1274.7; M = Co, observed at m/z 1275.5, calculated 1275.3) arising from loss of three tetrafluorborate anions. Again, the isotope pattern and 1/3 mass unit spacing between components confirmed the assignments of these peaks as belonging to the hexanuclear cluster with the two fluoride bridges intact. More intense peaks at lower m/z values were also observed corresponding to cleavage of the cluster in half across its weakest point by breaking the fluoride bridges, generating trinuclear species {[M₃L₂F](BF₄)_5−n}ⁿ⁺ (n = 2, 3) in both cases (see Experimental).

We also examined the ¹H NMR spectrum of [Co₆L₄(μ-F)₂](BF₄)₁₀ (Fig. 8). Despite its paramagnetism, high-spin Co(II) gives very informative ¹H NMR spectra with peaks spread over a wide range; the width of the peaks correlates nicely with the sum of the r⁻⁶ distances of each proton from the different Co(II) centres and this can be used as a basis for assignment.¹⁴ From the crystal structure of [Co₆L₄(μ-F)₂](BF₄)₁₀ it appears that each of the two ligand types (i.e. bis-bidentate, spanning two metals; and 2 + 2 + 4 dentate, spanning three metals) has no internal symmetry, which would result in 68 magnetically inequivalent proton environments. In fact in the ¹H NMR spectrum there are 61 signals clearly identifiable spread out over the range +165 to −122 ppm (Fig. 8). It is likely that a few more signals are obscured by the intense and broad peaks for water and residual protonated solvent in the 0–5 ppm region, and undetected overlap of two or more signals is also possible. On this basis it is clear that both ligands have no internal symmetry in solution, as the presence of a single symmetry element would reduce the number of independent ¹H signals to 34; but (in contrast to the crystal structure) it does confirm that the complex has overall twofold symmetry in solution, with each trinuclear half being equivalent. We also note that appearance of this many ¹H signals does not preclude the complex splitting in half across the fluoride bridges to give trinuclear species, but on the basis of the ESMS data the complex remains intact in solution. It is clear that several of the signals are broad and of low intensity, due to short T₁ relaxation times, and these correspond to the protons that lie closest to the Co(II) centres: the pyridyl H⁶protons (ca. 3.17 Å), and some of the diastereotopic methylene protons which are pointed towards the Co(II) centres (ca. 3.05 Å).¹⁴

Fig. 8 ¹H NMR spectrum of [Co₆L₄(μ-F)₂](BF₄)₁₀ in nitromethane. The signals marked * in the 0–5 ppm region belong to the complex but are closely overlapping with more intense residual solvent and water signals.

We also attempted to prepare a simple Ag(I) complex with L, reasoning that the presence of 8 donor atoms would result in a dinuclear complex [Ag₂L]²⁺ in which each Ag(I) is four coordinate from an adjacent pair of pyrazolyl-pyridine chelates. Isolation of a microcrystalline solid in the usual way was straightforward but we could not obtain X-ray quality crystals. However, the ES mass spectrum of the product confirmed its identity as [Ag₂L](BF₄)₂, with peaks at m/z 1009 and 461 corresponding to {[Ag₂L](BF₄)_2−n}ⁿ⁺ (n = 1,2), with appropriate isotopic patterns and spacings between adjacent peaks in the isotope clusters. Elemental analysis was also consistent with this formulation.

Mixed-metal complexes

The structures of [M₆L₄(μ-F)₂](BF₄)₁₀ (M = Co, Ni), with the central two metal ions only achieving a coordination number of six by using two fluoride ligands in addition to the two pyrazolyl-pyridine units, implies that a mixed-metal complex could be prepared based on a mixture of 6-coordinate and 4-coordinate metal ions with no ancillary ligands required.

The four-coordinate metal ion would be accommodated in the four-coordinate binding pocket arising from two bidentate pyrazolyl-pyridine units. This suggested that use of a mixture of metal ions—two octahedral M(II) ions (six coordinate) and one Ag(I) ion (four coordinate)—would result in a trinuclear M₂Ag assembly whose coordination requirements (6 + 6 + 4 = 16 donors, or 8 bidentate units) would be exactly matched by two equivalents of the ligand L.

This proved to be the case: reaction of L, M(BF₄)₂ and Ag(BF₄) in a 2 : 2 : 1 ratio in nitromethane solution, followed by crystallisation, afforded batches of X-ray quality crystalline products in high yields. Solvothermal methods were unsuccessful in these cases, affording only dark-coloured solids which we assume to be contaminated with silver metal due to reduction of the Ag(BF₄) under the relatively harsh reaction conditions.

The crystalline products proved to be the desired mixed-metal complexes [M₂AgL₂](BF₄)₅ (M = Co, Ni; Fig. 9 and 10). These are isostructural; the discussion of the structure will be based on [Ni₂AgL₂](BF₄)₅. The way in which the binding sites of the two ligands are partitioned between the three metal ions is identical to what we observed in each half of [M₆L₄(μ-F)₂](BF₄)₁₀ (Fig. 6 and 7), with one ligand presenting two bidentate arms to each Ni(II) ion—which are 8.88 Å apart—and the other presenting one arm to each of the Ni(II) ions, and the remaining two arms to the four-coordinate Ag(I) ion, which is in a highly irregular ‘pseudo tetrahedral’ geometry, with N–Ag–N angles lying in the range 70.6°–148.5°. Again the central phenyl ring of one ligand [C(291)–C(296)] lies sandwiched between two of the coordinated pyridyl rings of the other ligand, forming an acceptor/donor/acceptor π-stack with an offset geometry; the H atoms attached to C(133) and C(173) of the pyridyl rings lie ca. 3.6 Å above the π-cloud of the central phenyl ring.

Fig. 9 Structure of the complex cation of [Ni₂AgL₂](BF₄)₅·4MeCN·H₂O showing the atomic labelling scheme. H atoms are omitted for clarity.

Fig. 10 Alternative view of the complex cation of [Ni₂AgL₂](BF₄)₅·4MeCN·H₂O with the ligands coloured separately for clarity. H atoms are omitted for clarity.

ES mass spectrometry confirms the integrity of the complexes in solution, with a clear sequence of peaks corresponding to {[Ni₂AgL₂](BF₄)_5−n}ⁿ⁺ (n = 2, 3, 4), i.e. the intact trinuclear cation associated with varying numbers of tetrafluoroborate anions. The mass spectrum of [Co₂Ag(L)₂][BF₄]₅ likewise shows peaks assigned to {[Co₂AgL₂](BF₄)_5−n}ⁿ⁺ (n = 2, 3, 4) but also shows evidence for formation of some [Ag₂L][BF₄]₂, indicated by peaks corresponding to {[Ag₂L][BF₄]_2−n}ⁿ⁺ (where n = 1, 2). The lability of [Co₂Ag(L)₂][BF₄]₅ in solution was confirmed by its ¹H NMR spectrum. The intact complex would be expected to display 68 proton environments in its paramagnetically-shifted spectrum {cf.Fig. 7, for [Co₆L₄(μ-F)₂](BF₄)₁₀} but we observed a much higher number of signals than this with a large number of overlapping resonances in the region of 5 to 10 ppm consistent with either free ligand or the existence of a diamagnetic Ag(I) complex in the equilibrium mixture.

Conclusion

In conclusion, we have demonstrated how matching the known stereochemical and coordination number preferences of two types of metal ion with the total number of coordination sites available from an octadentate ligand can result in controlled assembly of mixed-metal complexes.

Experimental

Materials and methods

NMR spectra were recorded using Bruker AV1–250, AV3–400 or DRX–500 spectrometers; ES mass spectra were recorded with a Waters LCT instrument, using a low cone voltage (5V) for the metal complexes. 3(2-Pyridyl)pyrazole was prepared as described previously,¹⁵ and its reaction with 1,2,4,5-tetrakis(bromomethyl)benzene to afford the ligand L followed the method of McMorran and Steel.¹⁰ Other organic reagents and metal salts were obtained from Aldrich and used as received.

Syntheses of metal complexes

All homometallic clusters could be prepared by either solvothermal or conventional synthetic methods; solvothermal synthesis provided the benefits of higher yield and isolation of X-ray quality crystals directly from the slowly-cooled reaction mixture. The heteronuclear Ni₂Ag or Co₂Ag complexes could only be obtained by synthesis under ambient conditions; typical examples of both types of synthesis are outlined below:

[M₆L₄F₂][BF₄]₁₀ (M = Co, Ni) by solvothermal method

A Teflon lined autoclave was charged with Co(BF₄)₂·xH₂O (0.036 g, 0.110 mmol), L (0.050 g, 0.071 mmol) and methanol (9 cm³). Heating to 100 °C for 12 h followed by slow cooling to room temperature yielded large orange prismatic crystals directly from the cooled reaction mixture. The crystals were separated from the reaction mixture by decantation and dried (0.034 g, 47%). ESMS: m/z 1956.7, {[Co₆L₄F₂](BF₄)₈}²⁺; 1275.5, {[Co₆L₄F₂](BF₄)₇}³⁺; 934.8, {[Co₃L₂F](BF₄)₃}²⁺; 594.2, {[Co₃L₂F](BF₄)₂}³⁺. Calcd for [Co₆L₄F₂](BF₄)₁₀·4H₂O (C₁₆₈H₁₄₄B₁₀Co₆F₄₂N₄₈O₄): C 48.5, H 3.5, N 16.2. Found: C 48.6, H 3.4, N 16.1%.

Data for [Ni₆L₄F₂](BF₄)₁₀ (prepared in the same way). Yield: 0.041 g, 57%. ESMS: m/z 1274.8, {[Ni₆L₄F₂](BF₄)₇}³⁺; 934.3, {[NiL₂F₂](BF₄)₃}²⁺; 730.1, {[Ni₆L₄F₂](BF₄)₅}⁵⁺; 593.9, {[Ni₃L₂F](BF₄)₂}³⁺, 423.7, {[Ni₃L₂F₂](BF₄)}⁴⁺. Calcd. for [Ni₆L₄F₂](BF₄)₁₀·5H₂O (C₁₆₈H₁₄₆B₁₀F₄₂N₄₈Ni₆O₅): C 48.3, H 3.5, N 16.1. Found: C 48.4, H 3.5, N 16.0%.

Synthesis of [Co₂AgL₂][BF₄]₅ by conventional method

A solution of Co(BF₄)₂·xH₂O (0.024 g, 0.071 mmol) and Ag(BF₄)·xH₂O (0.007 g, 0.035 mmol) in MeOH (8 cm³) was combined with a solution of L (0.050 g, 0.071 mmol) in chloroform (8 cm³). The resulting solution was vigorously stirred for 24 h. The solvent was subsequently removed under reduced pressure, and the resulting crude powder was washed with methanol and then chloroform to remove any unreacted starting materials. Diethyl ether was allowed to slowly diffuse into a solution of the dried orange powder in nitromethane. X-Ray quality orange prismatic crystals grew on standing within one week (0.024 g, 33%). Data for [Co₂Ag(L)₂](BF₄)₅. ESMS: m/z 949.4, {[Co₂Ag(L)₂](BF₄)₅}²⁺; 604.6, {[Ni₂Ag(L)₂](BF₄)₅}³⁺, 431.1, {[Ni₂Ag(L)₂](BF₄)₅}⁴⁺. Calcd. for [Co₂Ag(L)₂](BF₄)₅.5H₂O (C₈₄H₇₈AgB₅Co₂F₂₀N₂₄O₅): C 46.6, H 3.6, N 15.5. Found: C 46.1, H 3.4, N 15.2%.

Data for [Ni₂AgL₂](BF₄)₅. Yield: 0.030 g, 41%. ESMS: m/z 949.2, {[Ni₂Ag)₂](BF₄)₅}²⁺; 603.8, {[Ni₂AgL₂](BF₄)₅}³⁺; 431.1, {[Ni₂AgL₂](BF₄)₅}⁴⁺. Calcd for [Ni₂Ag(L)₂](BF₄)₅·3H₂O (C₈₄H₇₄AgB₅F₂₀N₂₄Ni₂O₃): C 47.4, H 3.5, N 15.8. Found: C 47.4, H 3.5, N 15.6%.

Solvothermal synthesis of [Cd₄L₂F₂(MeOH)₄](BF₄)₆

A Teflon lined autoclave was charged with Cd(BF₄)₂·xH₂O (0.049 g, 0.142 mmol), L (0.050 g, 0.071 mmol) and methanol (9 cm³). The solution was subjected to solvothermal treatment as outlined above. The crystals were separated from the reaction mixture by decantation and dried (yield: 0.034 g, 47%). ESMS: m/z 1124.3, {[Cd₄L₂F₂](BF₄)₄}²⁺, 720.9, {[Cd₄L₂F₂](BF₄)₃}³⁺, 518.6, {[Cd₄L₂F₂](BF₄)₂}⁴⁺, 317.1, {[Cd₄L₂F₂]}⁶⁺. Calcd for [Cd₄L₂F₂](BF₄)₆·4H₂O (C₈₄H₇₆B₆Cd₄F₂₆N₂₄O₄): C 40. 5, H 3.1, N 13.5. Found: C 40.5, H 2.9, N 13.3%.

Solvothermal synthesis of [Cd₄L₂F₂(MeCN)₄](BF₄)₆

This was prepared in the same way as [Cd₄L₂F₂(MeOH)₄](BF₄)₆ (above) except that acetonitrile was used as the reaction solvent (yield: 0.045 g, 49%). ESMS: m/z 1124.3, {[Cd₄L₂F₂](BF₄)₄}²⁺; 720.2, {[Cd₄L₂F₂](BF₄)₃}³⁺, 519.1, {[Cd₄L₂F₂](BF₄)₂}⁴⁺, 397.7, {[Cd₄L₂F₂](BF₄)}⁵⁺, 317.3, {[Cd₄L₂F₂]}⁶⁺. Calcd for [Cd₄L₂F₂(MeCN)₄](BF₄)₆ (C₉₂H₈₀B₆Cd₄F₂₆N₂₈): C 42.7, H 3.1, N 15.2. Found: C 42.6, H 3.2, N 15.3%.

X-Ray crystallography

In each case a suitable crystal was mounted in a stream of cold N₂ on a Bruker APEX-2 CCD diffractometer equipped with graphite-monochromated Mo Kα radiation from a sealed-tube source. Details of the crystal, data collection and refinement parameters are summarized in Table 7, and selected structural parameters are provided in Tables 1–6. Data were corrected for absorption using empirical methods (SADABS)¹⁶ based upon symmetry-equivalent reflections combined with measurements at different azimuthal angles. The structures were solved by direct methods and refined by full-matrix least squares on weighted F² values for all reflections using the SHELX suite of programs.¹⁷ Non-hydrogen atoms were refined anisotropically. Hydrogen atoms were placed in calculated positions, refined using idealized geometries (riding model) and assigned fixed isotropic displacement parameters.

Table 1 Selected bond lengths (Å) and angles (°) for the structure of [Cd₄L₂F₂(MeOH)₄][BF₄]₆·6MeOH

Bond lengths/Å
Cd(1)–F(1)	2.175(4)	Cd(2)–F(1)	2.197(5)
Cd(1)–O(101)	2.303(6)	Cd(2)–N(42)	2.310(7)
Cd(1)–N(62)	2.336(7)	Cd(2)–N(82)	2.315(7)
Cd(1)–N(22)	2.338(7)	Cd(2)–O(111)	2.339(6)
Cd(1)–N(11)	2.358(7)	Cd(2)–N(31)	2.354(7)
Cd(1)–N(51)	2.416(7)	Cd(2)–N(71)	2.434(7)

---

Bond angles/°
F(1)–Cd(1)–O(101)	84.2(2)	F(1)–Cd(2)–N(42)	88.9(2)
F(1)–Cd(1)–N(62)	95.3(2)	F(1)–Cd(2)–N(82)	95.3(2)
O(101)–Cd(1)–N(62)	93.3(2)	N(42)–Cd(2)–N(82)	117.4(2)
F(1)–Cd(1)–N(22)	96.0(2)	F(1)–Cd(2)–O(111)	85.7(2)
O(101)–Cd(1)–N(22)	156.4(2)	N(42)–Cd(2)–O(111)	153.4(2)
N(62)–Cd(1)–N(22)	110.1(2)	N(82)–Cd(2)–O(111)	89.0(2)
F(1)–Cd(1)–N(11)	104.6(2)	F(1)–Cd(2)–N(31)	109.7(2)
O(101)–Cd(1)–N(11)	86.3(2)	N(42)–Cd(2)–N(31)	71.4(2)
N(62)–Cd(1)–N(11)	159.9(2)	N(82)–Cd(2)–N(31)	154.0(2)
N(22)–Cd(1)–N(11)	70.8(2)	O(111)–Cd(2)–N(31)	86.0(3)
F(1)–Cd(1)–N(51)	165.3(2)	F(1)–Cd(2)–N(71)	165.5(2)
O(101)–Cd(1)–N(51)	98.3(3)	N(42)–Cd(2)–N(71)	94.7(2)
N(62)–Cd(1)–N(51)	70.1(2)	N(82)–Cd(2)–N(71)	70.6(2)
N(22)–Cd(1)–N(51)	87.5(2)	O(111)–Cd(2)–N(71)	97.0(2)
N(11)–Cd(1)–N(51)	90.1(2)	N(31)–Cd(2)–N(71)	84.7(2)

---

Table 2 Selected bond lengths (Å) and angles (°) for the structure of [Cd₄L₂F₂(MeCN)₄][BF₄]₆·8MeCN

Bond lengths/Å
Cd(1)–F(1)	2.2407(18)	Cd(2)–F(1)	2.1545(18)
Cd(1)–N(22)	2.282(3)	Cd(2)–N(31)	2.321(3)
Cd(1)–N(101)	2.314(3)	Cd(2)–N(82)	2.321(3)
Cd(1)–N(62)	2.323(3)	Cd(2)–N(111)	2.357(3)
Cd(1)–N(11)	2.395(3)	Cd(2)–N(71)	2.380(3)
Cd(1)–N(51)	2.412(3)	Cd(2)–N(42)	2.424(3)

---

Bond angles/°
F(1)–Cd(1)–N(22)	86.20(8)	F(1)–Cd(2)–N(31)	98.20(8)
F(1)–Cd(1)–N(101)	79.70(9)	F(1)–Cd(2)–N(82)	97.70(8)
N(22)–Cd(1)–N(101)	141.98(10)	N(31)–Cd(2)–N(82)	159.69(9)
F(1)–Cd(1)–N(62)	91.41(8)	F(1)–Cd(2)–N(111)	102.35(9)
N(22)–Cd(1)–N(62)	120.73(9)	N(31)–Cd(2)–N(111)	89.03(10)
N(101)–Cd(1)–N(62)	94.93(10)	N(82)–Cd(2)–N(111)	99.81(10)
F(1)–Cd(1)–N(11)	117.18(8)	F(1)–Cd(2)–N(71)	165.05(8)
N(22)–Cd(1)–N(11)	70.56(9)	N(31)–Cd(2)–N(71)	91.10(9)
N(101)–Cd(1)–N(11)	84.82(10)	N(82)–Cd(2)–N(71)	70.91(9)
N(62)–Cd(1)–N(11)	150.71(9)	N(111)–Cd(2)–N(71)	89.41(10)
F(1)–Cd(1)–N(51)	159.53(8)	F(1)–Cd(2)–N(42)	92.40(8)
N(22)–Cd(1)–N(51)	95.53(9)	N(31)–Cd(2)–N(42)	71.09(9)
N(101)–Cd(1)–N(51)	109.79(10)	N(82)–Cd(2)–N(42)	95.71(9)
N(62)–Cd(1)–N(51)	70.13(9)	N(111)–Cd(2)–N(42)	156.85(10)
N(11)–Cd(1)–N(51)	82.36(9)	N(71)–Cd(2)–N(42)	79.53(9)

---

Table 3 Selected bond lengths (Å) for the structure of [Ni₆L₄F₂][BF₄]₁₀

Bond lengths/Å
Ni(1)–N(122)	2.072(7)	Ni(4)–F(2)	2.026(5)
Ni(1)–N(131)	2.085(7)	Ni(4)–F(1)	2.030(4)
Ni(1)–N(211)	2.107(8)	Ni(4)–N(482)	2.066(8)
Ni(1)–N(111)	2.108(8)	Ni(4)–N(451)	2.072(7)
Ni(1)–N(142)	2.109(8)	Ni(4)–N(462)	2.092(7)
Ni(1)–N(222)	2.127(8)	Ni(4)–N(471)	2.121(8)
Ni(2)–N(162)	2.069(7)	Ni(5)–N(322)	2.069(8)
Ni(2)–N(151)	2.073(8)	Ni(5)–N(331)	2.084(8)
Ni(2)–N(171)	2.082(7)	Ni(5)–N(311)	2.102(9)
Ni(2)–N(242)	2.111(8)	Ni(5)–N(411)	2.106(5)
Ni(2)–N(231)	2.127(5)	Ni(5)–N(342)	2.116(8)
Ni(2)–N(182)	2.140(7)	Ni(5)–N(422)	2.122(8)
Ni(3)–F(1)	2.007(5)	Ni(6)–N(362)	2.056(8)
Ni(3)–F(2)	2.034(5)	Ni(6)–N(371)	2.077(8)
Ni(3)–N(262)	2.048(7)	Ni(6)–N(431)	2.084(7)
Ni(3)–N(271)	2.062(4)	Ni(6)–N(351)	2.096(7)
Ni(3)–N(282)	2.116(7)	Ni(6)–N(382)	2.118(8)
Ni(3)–N(251)	2.133(7)	Ni(6)–N(442)	2.138(8)

---

Table 4 Selected bond lengths (Å) for the structure of [Co₆L₄F₂][BF₄]₁₀

Bond lengths/Å
a This corresponds to the major disorder component of one pyrdiyl/pyrazole chelate unit.
Co(1)–N(122)	2.128(10)	Co(4)–F(1)	2.035(6)
Co(1)–N(131)	2.133(10)	Co(4)–F(2)	2.046(6)
Co(1)–N(211)	2.150(11)	Co(4)–N(482)	2.093(12)
Co(1)–N(111)	2.163(11)	Co(4)–N(462)	2.120(11)
Co(1)–N(142)	2.167(11)	Co(4)–N(471)	2.167(9)
Co(1)–N(222)	2.180(11)	Co(4)–N(451)	2.208(13)
Co(2)–N(162)	2.102(9)	Co(5)–N(422)	2.128(11)
Co(2)–N(171)	2.123(8)	Co(5)–N(322)	2.137(10)
Co(2)–N(242)	2.135(11)	Co(5)–N(331)	2.150(10)
Co(2)–N(231)	2.149(6)	Co(5)–N(342)	2.153(10)
Co(2)–N(151)	2.158(10)	Co(5)–N(411)	2.160(7)
Co(2)–N(182)	2.171(8)	Co(5)–N(311)	2.182(11)
Co(3)–F(2)	2.052(6)	Co(6)–N(362)	2.103(10)
Co(3)–F(1)	2.058(5)	Co(6)–N(8Y)^a	2.111(6)
Co(3)–N(262)	2.096(8)	Co(6)–N(351)	2.126(10)
Co(3)–N(271)	2.127(6)	Co(6)–N(431)	2.134(9)
Co(3)–N(251)	2.184(9)	Co(6)–N(1Y)^a	2.147(6)
Co(3)–N(282)	2.190(9)	Co(6)–N(442)	2.186(11)

---

Table 5 Selected bond lengths (Å) and angles (°) for the structure of [Ni₂AgL₂][BF₄]₅·4MeCN·H₂O

Bond lengths/Å
Ni(1)–N(122)	2.052(11)	Ni(2)–N(162)	2.040(10)
Ni(1)–N(131)	2.070(10)	Ni(2)–N(171)	2.051(11)
Ni(1)–N(211)	2.091(9)	Ni(2)–N(182)	2.102(9)
Ni(1)–N(111)	2.096(9)	Ni(2)–N(151)	2.107(10)
Ni(1)–N(142)	2.125(9)	Ni(2)–N(242)	2.116(9)
Ni(1)–N(222)	2.139(9)	Ni(2)–N(231)	2.127(9)
Ag(1)–N(251)	2.280(10)	Ag(1)–N(282)	2.392(10)
Ag(1)–N(271)	2.317(9)	Ag(1)–N(262)	2.505(11)

---

Bond angles/°
N(122)–Ni(1)–N(131)	169.5(4)	N(162)–Ni(2)–N(171)	168.6(4)
N(122)–Ni(1)–N(211)	93.6(4)	N(162)–Ni(2)–N(182)	91.3(4)
N(131)–Ni(1)–N(211)	95.3(4)	N(171)–Ni(2)–N(182)	78.9(4)
N(122)–Ni(1)–N(111)	77.8(4)	N(162)–Ni(2)–N(151)	77.8(4)
N(131)–Ni(1)–N(111)	96.7(4)	N(171)–Ni(2)–N(151)	96.6(4)
N(211)–Ni(1)–N(111)	89.7(4)	N(182)–Ni(2)–N(151)	91.3(4)
N(122)–Ni(1)–N(142)	92.4(4)	N(162)–Ni(2)–N(242)	100.9(4)
N(131)–Ni(1)–N(142)	78.8(4)	N(171)–Ni(2)–N(242)	86.6(4)
N(211)–Ni(1)–N(142)	174.0(4)	N(182)–Ni(2)–N(242)	100.3(4)
N(111)–Ni(1)–N(142)	92.1(4)	N(151)–Ni(2)–N(242)	168.4(4)
N(122)–Ni(1)–N(222)	100.2(4)	N(162)–Ni(2)–N(231)	89.9(4)
N(131)–Ni(1)–N(222)	87.2(3)	N(171)–Ni(2)–N(231)	100.1(4)
N(211)–Ni(1)–N(222)	77.8(4)	N(182)–Ni(2)–N(231)	177.9(4)
N(111)–Ni(1)–N(222)	167.2(4)	N(151)–Ni(2)–N(231)	90.7(4)
N(142)–Ni(1)–N(222)	100.7(3)	N(242)–Ni(2)–N(231)	77.7(4)
N(251)–Ag(1)–N(271)	147.2(4)	N(251)–Ag(1)–N(262)	71.5(4)
N(251)–Ag(1)–N(282)	140.2(4)	N(271)–Ag(1)–N(262)	133.5(3)
N(271)–Ag(1)–N(282)	71.4(3)	N(282)–Ag(1)–N(262)	83.7(3)

---

Table 6 Selected bond lengths (Å) and angles (°) for the structure of [Co₂AgL₂][BF₄]₅·4MeCN

Bond lengths/Å
Co(1)–N(122)	2.089(5)	Co(2)–N(171)	2.103(5)
Co(1)–N(131)	2.116(5)	Co(2)–N(162)	2.110(5)
Co(1)–N(111)	2.142(5)	Co(2)–N(242)	2.146(5)
Co(1)–N(222)	2.150(5)	Co(2)–N(151)	2.149(5)
Co(1)–N(211)	2.188(5)	Co(2)–N(231)	2.173(5)
Co(1)–N(142)	2.226(5)	Co(2)–N(182)	2.190(5)
Ag(1)–N(271)	2.253(5)	Ag(1)–N(262)	2.377(5)
Ag(1)–N(251)	2.318(5)	Ag(1)–N(282)	2.548(5)

---

Bond angles/°
N(122)–Co(1)–N(131)	163.6(2)	N(171)–Co(2)–N(162)	163.39(19)
N(122)–Co(1)–N(111)	77.1(2)	N(171)–Co(2)–N(242)	88.32(19)
N(131)–Co(1)–N(111)	95.8(2)	N(162)–Co(2)–N(242)	101.85(19)
N(122)–Co(1)–N(222)	102.82(19)	N(171)–Co(2)–N(151)	95.1(2)
N(131)–Co(1)–N(222)	87.95(19)	N(162)–Co(2)–N(151)	77.7(2)
N(111)–Co(1)–N(222)	164.69(19)	N(242)–Co(2)–N(151)	167.91(19)
N(122)–Co(1)–N(211)	93.3(2)	N(171)–Co(2)–N(231)	102.42(19)
N(131)–Co(1)–N(211)	101.4(2)	N(162)–Co(2)–N(231)	92.84(19)
N(111)–Co(1)–N(211)	88.34(19)	N(242)–Co(2)–N(231)	76.09(18)
N(222)–Co(1)–N(211)	76.37(19)	N(151)–Co(2)–N(231)	91.84(19)
N(122)–Co(1)–N(142)	88.71(19)	N(171)–Co(2)–N(182)	76.97(19)
N(131)–Co(1)–N(142)	76.6(2)	N(162)–Co(2)–N(182)	87.94(19)
N(111)–Co(1)–N(142)	91.57(19)	N(242)–Co(2)–N(182)	102.41(18)
N(222)–Co(1)–N(142)	103.73(19)	N(151)–Co(2)–N(182)	89.66(18)
N(211)–Co(1)–N(142)	177.9(2)	N(231)–Co(2)–N(182)	178.43(19)
N(271)–Ag(1)–N(251)	140.33(18)	N(271)–Ag(1)–N(282)	69.97(17)
N(271)–Ag(1)–N(262)	136.37(18)	N(251)–Ag(1)–N(282)	149.12(18)
N(251)–Ag(1)–N(262)	71.54(18)	N(262)–Ag(1)–N(282)	88.15(16)

---

Table 7 Summary of crystal data, data collection and refinement parameters for the new structures

Structure	[Ni6L4F2][BF4]10^a	[Co6L4F2][BF4]10^a	[Cd4L2F2(MeOH)4][BF4]6·6MeOH	[Cd4L2F2(MeCN)4][BF4]6·8MeCN	[Ni2AgL2][BF4]5·4MeCN·H2O	[Co2AgL2][BF4]5·4MeCN^a
a Due to extensive disorder of anions/solvent molecules that could not be adequately modelled, the SQUEEZE command in PLATON was used in the final refinement. b structure twinned so Rint value not defined. c The value of R1 is based on ‘observed’ data with I > 2σ(I); the value of wR2 is based on all data.
Formula	C168H136B10F42N48Ni6	C168H136B10Co6F42N48	C94H108B6Cd4F26N24O10	C108H104B6Cd4F26N36	C92H82AgB5F20N28Ni2O	C96H86AgB5Co2F20N30
Molecular weight	4085.6	8445.1	2742.5	2914.7	2255.2	2319.7
T/K	150(2)	150(2)	150(2)	120(2)	100(2)	100(2)
Crystal system	Orthorhombic	Orthorhombic	Monoclinic	Triclinic	Triclinic	Triclinic
Space group	Pccn	Pccn	P21/n	P1̄	P1̄	P1̄
a/Å	29.9778(7)	29.8251(6)	13.8873(11)	13.4883(14)	11.5875(4)	15.2802(12)
b/Å	50.1664(12)	50.3863(10)	27.175(2)	13.9814(15)	18.7782(7)	16.0017(13)
c/Å	26.2454(6)	26.4444(6)	14.4995(11)	16.7171(18)	23.2260(8)	24.297(2)
α/°	90	90	90	95.870(6)	77.510(2)	81.203(5)
β/°	90	90	99.468(4)	106.613(5)	83.373(3)	86.810(5)
γ/°	90	90	90	91.114(5)	86.226(3)	62.775(3)
V/Å^3	39 469.9(16)	39 740.0(14)	5397.4(7)	3001.2(6)	4896.9(3)	5219.8(7)
Z	8	8	2	1	2	2
D c/g cm^−3	1.375	1.412	1.687	1.613	1.529	1.476
Crystal size/mm^3	0.60 × 0.4 × 0.4	0.50 × 0.40 × 0.30	0.30 × 0.30 × 0.05	0.70 × 0.40 × 0.40	0.12 × 0.11 × 0.08	0.20 × 0.20 × 0.20
μ/mm^−1	0.661	0.595	0.891	0.803	0.68	0.597
Data/restraints/parameters	25 811/544/2392	25 964/2574/2347	11 038/51/729	12 929/210/768	15 344/218/1215	18 372/220/1257
R int for independent data	0.0911	0.0884	0.0551	0.0308	—	0.0756
Final R1, wR2^c	0.1151, 0.3509	0.1603, 0.4590	0.0796, 0.1830	0.0347, 0.0996	0.0900, 0.2935	0.0726, 0.1987

---

Several of the structures presented problems during data collection and refinement due to rapid solvent loss and/or disorder, which resulted in weak scattering. These issues, and their treatments, are discussed in detail in the individual CIFs. In brief: all crystal structures suffered from highly disordered tetrafluoroborate anions. Where possible these were modeled with geometric and displacement restraints. Disordered anions and solvents have been left isotropic. A Platon SQUEEZE function was used to eliminate large volumes containing highly disordered counter ions and lattice solvent molecules in the refinements of [M₆L₄F₂][BF₄]₁₀ (M = Co, Ni) and [Co₂AgL₂][BF₄]₅. The diffraction images for these structures displayed a lot of diffuse and weak reflectance, symptomatic of high disorder, which limited the quality of the data. The data for [Ni₂AgL₂][BF₄]₅ has been treated for rotational disorder (twinning). Only the major component has been used to solve the structure, the minor component being rejected due to poor intensity statistics.

Acknowledgements

We thank the EPSRC for financial support.

References

1. (a) C. Piguet, G. Bernardinelli and G. Hopfgartner, Chem. Rev., 1997, 97, 2005; (b) M. Albrecht, Chem. Rev., 2001, 101, 3457; (c) G. F. Swiegers and T. J. Malefetse, Coord. Chem. Rev., 2002, 225, 91; (d) M. Fujita, M. Tominaga, A. Hori and B. Therrien, Acc. Chem. Res., 2005, 38, 369; (e) D. Fiedler, D. H. Leung, R. G. Bergman and K. N. Raymond, Acc. Chem. Res., 2005, 38, 349; (f) D. Philp and J. F. Stoddart, Angew. Chem., Int. Ed. Engl., 1996, 35, 1155.
2. (a) J.-M. Lehn, A. Rigault, J. Siegel, J. Harrowfield, B. Chevrier and D. Moras, Proc. Natl. Acad. Sci. U. S. A., 1987, 84, 2565; (b) J.-M. Lehn and A. Rigault, Angew. Chem., Int. Ed. Engl., 1988, 27, 1095; (c) T. M. Garrett, U. Koert, J.-M. Lehn, A. Rigault, D. Meyer and J. Fisher, J. Chem. Soc., Chem. Commun., 1990, 557; (d) R. Kramer, J.-M. Lehn and A. Marquis-Rigault, Proc. Natl. Acad. Sci. U. S. A., 1993, 90, 5394.
3. (a) Reviews: M. Ruben, J. Rojo, F. J. Romero-Salguero, L. H. Uppadine and J.-M. Lehn, Angew. Chem., Int. Ed., 2004, 43, 3644; (b) L. K. Thompson, Coord. Chem. Rev., 2002, 233, 193.
4. (a) Representative recent examples: W.-L. Liu, K.-J. Lee, M. Park, R. P. John, D. Moon, Y. Zou, X.-F. Liu, H.-C. Ri, G.-H. Kim and M.-S. Lah, Inorg. Chem., 2008, 47, 8807; (b) E. Coronado, J. R. Galan-Mascaros, P. Gavina, C. Marti-Gastaldo, F. M. Romero and S. Tatay, Inorg. Chem., 2008, 47, 5197; (c) S. P. Argent, H. Adams, T. Riis-Johannessen, J. C. Jeffery, L. P. Harding, O. Mamula and M. D. Ward, Inorg. Chem., 2006, 45, 3905; (d) O. Mamula, A. von Zelewsky, P. Broadard, C. W. Schlapfer, G. Bernardinelli and H. Stoeckli-Evans, Chem.–Eur. J., 2005, 11, 3049; (e) C. S. Campos-Fernandez, B. L. Schottel, H. T. Chifotides, J. K. Bera, J. Bacsa, J. M. Koomen, D. H. Russell and K. R. Dunbar, J. Am. Chem. Soc., 2005, 127, 12909.
5. (a) Representative recent examples: I. S. Tidmarsh, T. B. Faust, H. Adams, L. P. Harding, W. Clegg and M. D. Ward, J. Am. Chem. Soc., 2008, 130, 15167; (b) A. V. Davis, D. Fiedler, G. Seeber, A. Zahl, R. van Eldik and K. N. Raymond, J. Am. Chem. Soc., 2006, 128, 1324; (c) T. Brassey, R. Scopelliti and K. Severin, Chem. Commun., 2006, 3308; (d) J. K. Clegg, L. F. Lindoy, B. Moubaraki, K. S. Murray and J. C. McMurtrie, Dalton Trans., 2004, 2417; (e) M. Fujita, M. Tominaga, A. Hori and B. Therrien, Acc. Chem. Res., 2005, 38, 369; (f) Q. H. Yuan, L. J. Wan, H. Jude and P. J. Stang, J. Am. Chem. Soc., 2005, 127, 16279; (g) I. M. Muller and D. Moller, Angew. Chem., Int. Ed., 2005, 44, 2969.
6. (a) R. Kramer, J.-M. Lehn, A. de Cian and J. Fischer, Angew. Chem., Int. Ed. Engl., 1993, 32, 703; (b) L. J. Charbonniere, A. F. Williams, C. Piguet, G. Bernardinelli and E. Rivara-Minten, Chem.–Eur. J., 1998, 4, 485; (c) L. P. Harding, J. C. Jeffery, T. Riis-Johannessen, C. R. Rice and Z. T. Zeng, Chem. Commun., 2004, 654.
7. (a) N. André, R. Scopelliti, G. Hopfgartner, C. Piguet and J.-C. G. Bünzli, Chem. Commun., 2002, 214; (b) N. Martin, J.-C. G. Bünzli, V. McKee, C. Piguet and G. Hopfgartner, Inorg. Chem., 1998, 37, 577.
8. (a) S. P. Argent, H. Adams, L. P. Harding, T. Riis-Johannessen, J. C. Jeffery and M. D. Ward, New J. Chem., 2005, 29, 904; E. C. Constable, A. J. Edwards, P. R. Raithby and J. V. Walker, Angew. Chem., Int. Ed. Engl., 1993, 32, 1465; C. Piguet, G. Hopfgartner, B. Bocquet, O. Schaad and A. F. Williams, J. Am. Chem. Soc., 1994, 116, 9092.
9. (a) T. Riis-Johannessen, N. Dupont, G. Canard, G. Bernardinelli, A. Hauser and C. Piguet, Dalton Trans., 2008, 3661; (b) M. Cantuel, F. Gumy, J.-C. G. Bünzli and C. Piguet, Dalton Trans., 2006, 2647; (c) M. Cantuel, G. Bernardinelli, G. Muller, J. P. Riehl and C. Piguet, Inorg. Chem., 2004, 23, 1840.
10. D. A. McMorran and P. J. Steel, Inorg. Chem. Commun., 2003, 6, 43.
11. C. A. Hunter and J. K. M. Sanders, J. Am. Chem. Soc., 1990, 112, 5525.
12. G. R. Desiraju, and T. Steiner, The Weak Hydrogen Bond in Structural Chemistry and Biology, Oxford University Press, Oxford, 1999.
13. (a) N. K. Al-Rasbi, C. Sabatini, F. Barigelletti and M. D. Ward, Dalton Trans., 2006, 4769; (b) S. P. Argent, H. Adams, T. Riis-Johannessen, J. C. Jeffery, L. P. Harding, O. Mamula and M. D. Ward, Inorg. Chem., 2006, 45, 3905; (c) I. S. Tidmarsh, T. B. Faust, H. Adams, L. P. Harding, L. Russo, W. Clegg and M. D. Ward, J. Am. Chem. Soc., 2008, 130, 15167.
14. (a) I. S. Tidmarsh, B. F. Taylor, M. J. Hardie, L. Russo, W. Clegg and M. D. Ward, New J. Chem., 2009, 33, 366; (b) H. Amouri, L. Mimassi, M. N. Rager, B. E. Mann, C. Guyard-Duhayon and L. Raehm, Angew. Chem., Int. Ed., 2005, 44, 4543.
15. (a) A. J. Amoroso, A. M. W. Cargill Thompson, J. C. Jeffery, P. L. Jones, J. A. McCleverty and M. D. Ward, J. Chem. Soc., Chem. Commun., 1994, 2751; (b) H. Brunner and T. Scheck, Chem. Ber., 1992, 125, 701; (c) Y. Lin and S. A. Lang, J. Heterocycl. Chem., 1977, 14, 345.
16. G. M. Sheldrick, SADABS: A program for absorption correction with the Siemens SMART system, University of Göttingen, Germany, 1996.
17. G. M. Sheldrick, Acta Crystallogr., Sect. A: Found. Crystallogr., 2008, 64, 112.

---

Footnote

† CCDC reference numbers 718438–718443. For crystallographic data in CIF or other electronic format see DOI: 10.1039/b901891c

---