Crystal structures of two cross-bridged chromium(III) tetraazamacrocycles

Chromium(III) complexes of two ethylene cross-bridged tetraazamacrocycles were prepared and structurally characterized in order to extend the coordination chemistry of this ligand type farther towards the early transition metals.


Structural commentary
Each of the title compounds crystallizes with a single positively-charged metal complex and one PF 6 À anion in the asymmetric unit. The metal ion in each complex adopts a distorted octahedral geometry. The N atoms of each macrocycle occupy four coordination sites, while two chloride ions in a cis arrangement complete the coordination of Cr III . This socalled cis-V conformation, expected to be dictated by the ligand cross-bridge, is apparent for both of the complexes structurally characterized here. Figs. 1 and 2 illustrate the local geometry about Cr III in (I) (dimethyl bridged-cyclen complex) and (II) (dimethyl bridged-cyclam complex), respectively. Apparently, neither the identity of the metal ion, nor that of the alkyl substituents affects this conformation. This same conformation has been seen in all known metal complexes with ethylene cross-bridged cyclam and cyclen ligands. The ring size of the parent macrocycle alters the degree to which the metal ion is engulfed by the bridged macrocycle. This is most clearly evident in the N2-Cr1-N4 bond angle between two axially bound nitrogen atoms. This bond angle is 161.62 (11) in the case of the smaller macrocycle, cylcen, while it is 171.44 (14) for the cyclam complex. A larger bond angle, closer to linearity, indicates a better fit, or complementarity, between the ligand and the preferred octahedral geometry of the Cr III ion. A more subtle difference in the N1-Cr1-N3 bond angles, viz. the equatorially bound N atoms, shows the same trend: this angle is 83.23 (10) for the cyclen complex and 84.18 (13) for the cyclam complex. Finally, the Cr-N bond lengths are somewhat affected by the ligand size as well. The mean of the four Cr-N bond lengths is 2.08 Å in (I), while this average is 2.12 Å in (II). The mean value for a number of Cr-NR 3 bonds in the literature is 2.093 Å ( = 0.044 Å ) (Orpen et al., 1989). The molecular entities of (I), with atoms shown as displacement ellipsoids at the 50% probability level.

Supramolecular features
There are no classical hydrogen bonds present in either (I) and (II) but each structure contains a great many C-HÁ Á ÁF and C-HÁ Á ÁCl interactions which generate three-dimensional arrays. These interactions were identified from the standard criterion that the distance from the hydrogen atom to the hydrogen-bond acceptor should not exceed the sum of the radius of the acceptor plus 2 Å . Tables 1 and 2 contain full details of these interactions for (I) and (II), respectively. For (I), each PF 6 À anion resides in a pocket between six metal complexes and there are C-HÁ Á ÁF interactions to each of them. The mean CÁ Á ÁF distance of those in Table 1 is 3.35 Å . Supplementary C-HÁ Á ÁCl intramolecular contacts are present and intermolecular interactions between neighbouring metal complexes are also observed. The overall effect of these intermolecular interactions is to generate an extended network. One way to describe this is in terms of puckered sheets of the cationic complex and PF 6 À anions that extend in the bc plane. Between these sheets further C-HÁ Á ÁF and C-HÁ Á ÁCl interactions assemble these layers in an ABAB fashion along a to generate a densely packed three-dimensional array as shown in Fig. 3.
For (II), the arrangement is rather similar and again a threedimensional array is constructed from nonclassical hydrogen bonds between the cations and anions. The PF 6 À anion is located in a pocket formed from four metal complexes in a distorted tetrahedral arrangement and forms C-HÁ Á ÁF interactions to each of them, with a mean CÁ Á ÁF distance of 3.23 Å . Further C-HÁ Á ÁCl interactions are also present. In a similar fashion to (I), these nonclassical interactions assemble the cations and anions into puckered sheets that extend in the bc plane. The sheets are then ABAB stacked along a as shown in Fig. 4.

Database survey
The structures of three complexes that are directly analogous to (I) have been reported. These are the manganese (  Crystal packing of (I), viewed perpendicular to the bc plane. Dashed lines represent halideÁ Á ÁH-C interactions.
lengths. The Co III analogue is rather different because it is in a low spin state. The axial N-Co-N bond angle is 168.8 (4) and the equatorial bond angle is 87.2 (4) . As expected, the mean bond length is shorter for the Co case at 1.978 Å . The smaller, low-spin Co III ion fits into the pocket of the macrocyle better than Cr III . Chromium(III) complexes similar to (I) and (II) but crystallized with different anions have been reported before (Maples et al., 2009). The chloride analogue of (I) has bond angles of 160.83 (19) and 83.50 (18) about the chromium ion and a mean Cr-N bond length of 2.08 Å , which are in good agreement with (I), demonstrating the counter-anion has very little effect on the coordination about the metal. A cyclenbased macrocycle with benzyl groups replacing the methyl groups in (I), has key bond angles 160.35 (19) and 83.6 (2) and a mean Cr-N bond length of 2.09 Å (Maples et al., 2009). The pocket in the macrocycle is of a similar shape in this example but slightly enlarged because of the pendant benzyl groups.
The chloride analogue of (II) (Maples et al., 2009) displays a similarly sized pocket; the N-Cr-N axial bond angle is 172.46 (11) and the equatorial angle is 84.63 (11) , while mean Cr-N bond length is 2.12 Å . In line with the observation in (I) and (II), the pocket of the cyclam-derived ligand is better able to accomodate the octahedrally surrounded Cr III ion and displays larger bond lengths than the cyclen equivalent.

Synthesis and crystallization
The cross-bridged ligands were prepared according to literature procedures (Weisman et al., 1990;Wong et al., 2000). The title complexes were prepared by a procedure slightly modified from those found in Hubin et al. (2001) for other trivalent metal ions. In an inert atmosphere glove-box, 1 mmol of the respective ligand was dissolved in 20 ml of anhydrous dimethylformamide in a 50 ml Erlenmeyer flask. 1 mmol of anhydrous chromium(II) chloride was added to the stirring ligand solution. The reaction was allowed to stir at room temperature overnight. The solution was then filtered through filter paper and the solvent was removed under vacuum to give blue-violet solids. In the glove-box, this divalent complex was dissolved in 20 ml of methanol in a round-bottomed flask. Five equivalents of NH 4 PF 6 (5 mmol, 0.815 g) were dissolved in the solution. The flask was removed from the glove-box with a stopper to protect it from air. In a fume hood, a stream of nitrogen gas was directed over the surface of the solution. Four to six drops of Br 2 were added and the reaction was stirred for 15 min. Bright purple precipitates formed immediately. The nitrogen gas was then allowed to bubble through the solution for 15 min to remove excess Br 2 . The flask was then stoppered and placed in a freezer for 30 min to complete the precipitation. The purple solid product was collected by vacuum filtration on a glass frit and washed with methanol and then ether. Crystals suitable for X-ray diffraction (purple blocks) were grown from the slow evaporation of aqueous solutions of the product.

Refinement
Crystal data, data collection and structure refinement details are summarized in Table 3. H atoms were placed in idealised positions and refined using a riding model, with C-H = 0.98 and 0.99 Å for -CH 3 and -CH 2 -groups, respectively, and with U iso (H) values of, respectively, 1.5 and 1.2 times U eq of the carrier atom. In (I), there is evidence for a very small degree of disorder (10%) in the position of the PF 6 À anions. Refinement with a second orientation for this anion did not lead to a substantial improve in the fit. A model with a single orientation was therefore retained.