Three phenanthroline–metal complexes with topologically similar but geometrically different conformations

Two out of the three very similar complexes described present twofold symmetry but not the third one, probably by way of a strong intramolecular C—H⋯O hydrogen bond disrupting the symmetry, a fact which is analysed in detail.


Chemical context
In the last fifteen years we have made several contributions to the structural chemistry of group XII divalent cations, in particular Cd and Zn, complexed by the peroxodisulfate anion S 2 O 8 2À (pds) and several nitrogen-containing aromatic bases (nab). In all these cases, the basic general formula appeared to be M(pds)(nab) 2 , plus the possible inclusion of some water   (Harvey et al., 2001b;Marsh, 2004, and this work); (VI): Zn(pds)(TMPhen) 2 , P1 (this work); (VII): Cd(pds)(TMPhen) 2 , Pbcn (this work). Ligand code: Bpy = 2,2 0bipyridine; Phen = 1,10-phenanthroline, DMPhen = 2,9-dimethyl-1,10phenanthroline, TMPhen = 3,4,7,8-tetramethyl-1,10-phenanthroline. molecules, either coordinating or as a solvate (details of these complexes, including the ones to be described in the present work, are summarized in Fig. 1). Even if too few structures are reported to make any confident statistical analysis, the results suggest some kind of a trend between the identity of the nitrogen-containing base and the way the pds anion performs in coordination. Thus, for the smallest one, nab = 2,2 0 -bipyridine (Bpy), the structures obtained [(I) and (II)] show two coordinating pds units in a bridging -O-S-O-mode. For the intermediate nab = 1,10-phenanthroline (Phen), one of these pds appears to be replaced by a (smaller) coordinating water molecule, while the bound pds acts as a pendant monocoordinating ligand [(III) and (IV)]. Finally, at the beginning of this work we had at hand only one single example of a relatively larger nab species, represented by nab = 2,9-dimethyl-1,10-phenanthroline, [DMPhen, (V)], where the single coordinating pds folds into itself to bind through both ends, acting in a chelating fashion. Furthermore, in both compounds of each pair of homologues (I)-(II) and (III)-(IV), the anion displays very similar conformations, defined by selected dihedral angles (Harvey et al., 2011).
In order to go further in this analysis we synthesized two new complexes of this sort, with M = Zn (VI) and M = Cd (VII), and with a common, tetra-methylated nab ligand, 3,4,7,8-tetramethyl-1,10-phenanthroline (TMPhen). We shall see that they present the same topology as compound (V), but with subtle, interesting differences regarding intramolecular interactions which will be discussed in detail. Unlike what happens in the (I)-(II) and (III)-(IV) homologues, in (VI) and (VII) the anion displays remarkably different conformations (Harvey et al., 2011). Since a comparison with the DMPhen structure (V) will be an important part of the discussion, and taking into account that the available data in the correct space group C2/c [as disclosed by Marsh (2004)] come from an averaging process (without further refinement) of previous results in Cc by our team (Harvey et al., 2001b), we indulge in including herein, for completeness, a fresh refinement in C2/c based on the original data for this structure, in addition to the synthesis and crystal structure of the two new complexes, (VI) and (VII). Even though we shall restrict this discussion to the pds anion, it is pertinent to state that the tetrathionate anion (S 4 O 6 2À ) behaves in a rather similar way, and that the tetrathionate Zn and Cd complexes with nab = DMBpy = 4,4 0dimethyl-2,2 0 -bipyridine (Harvey et al., 2013) have a similar coordination disposition to (V), (VI) and (VII).

Structural commentary
The Zn complex (VI) crystallizes in space group Pbca, and the complete molecule is bisected by a twofold symmetry axis, hence only half of the molecule is independent (Z 0 = 1/2); even if in a different space group, these properties are shared by structure (V). The Cd counterpart (VII), in turn, crystallizes in space group P1 with a full molecule in the asymmetric unit.
All three compounds present a topologically similar molecular configuration (Fig. 2), consisting in a central cation to which three bidentate chelating ligands bind, viz. two N,N 0nab and one O,O 0 -pds units. In particular, the 'close' character of the pds anion is in line with the trend so far observed, that methylated bases favour the chelating behaviour of pds.
Coordination distances in all three compounds are basically featureless, and agree with the expected values for each cation-ligand pair. However, a difference arises in the asym-metric way in which the ligands bind in (VII), contrasting dramatically with the twofold arrangement in (V) and (VI).
The chelating character of the ligands involved induces highly distorted coordination polyhedra. Proof of this is presented in Table 1, which shows the departure of the 'trans' angles in (V), (VI) and (VII) from their expected values of 180 for a regular octahedron. This makes the polyhedra difficult to classify, and impairs the description of coordination in terms of any regular model. In this regard, all three compounds are suitable for the analysis via the Vectorial Bond Valence Model (VBVM) suggested by Harvey et al. (2006), an approach tending to a simpler description of multidentate binding, in which the action of each ligand is integrated into a single interaction vector, or VBV (Vectorial Bond Valence), derived from the individual bond valences of the coordinating atoms. VBVM predicts a nil resultant of the vectorial sum of all the VBV vectors and, as a consequence, in this particular case of three-ligand coordination geometry, their disposition in a planar array. The first condition is complied satisfactorily with a very short resultant for the Bond Valence Vectors [0.08, 0.03 and 0.08 valence units for (V), (VI) and (VII), respectively]. The second requirement (planar array of vectors), applies sensu stricto in (V) and (VI), due to the intrinsic twofold symmetry around the cation, and it falls well within experimental error in (VII), where the calculated angles between Bond Valence Vectors add up to 359.5 (3) and the plane defined by their extremes leave the Cd II atom only 0.09 (2) Å aside.
As an unwitting bonus of this description, these planes appear as a natural reference frame for describing ligand orientations in the polyhedra, evidencing in (V) and (VI) their adherence to twofold symmetry and in (VII) significant departures from a symmetric arrangement. This can be visualized in Fig. 3, where a schematic representation (with an exaggerated perspective) is made of the ligand bites (open bonds) as well as the VBV representing their joint effect as a ligand (solid lines). At the left, the explanation of a group of angles helping to describe the orientation of the coordination planes is provided: angles labeled give account of the angular separation in the plane between vectors, while those labeled ! measure the out-of-plane rotation of the coordination planes around the corresponding VBV vectors. It is apparent, either by visual inspection of Fig. 3 or through the analysis of the ! values (Table 2), that the coordination polyhedron in (VII) is abnormally distorted. Since this could be the result of packing strain (intermolecular interactions) or just due to genuine intramolecular forces, we shall analyze and compare the three packing arrangements for (V), (VI) and (VII).

Supramolecular features
The most relevant, non-covalent interactions involved are presented in Table 3 (hydrogen bonds) and Table 4 (contacts). The second column includes a code, which labels each interaction for easy reference; in the last column, the role the interaction plays in packing is listed. Fig. 4 presents packing views of all three structures: it is apparent that in spite of crystallizing in different space groups, with different symmetry environments, the leitmotifs are strictly the same, viz.bound chains running along [101] in (V) and [001] in (VI) and (VII), the link being the stacking interaction appearing in Table 4, which in all cases connect inversion-related moieties. Except for the rather strong #2c in (VII), the remaining intermolecular interactions are weak and serve either to strengthen the link within the chains (marked as 'intrachain' in the tables) or to weakly connect parallel chains with each other ('interchain') to end up defining weakly bound three-dimensional structures. This description is valid for all three structures, and there is nothing special about the packing interactions in (VII) so as to ascribe to them the responsibility for the coordination 'anomaly'. In fact, interaction #2c, which due to its outstanding character might be thought of as a candidate to blame, involves the 'well behaved'    N21,N22-TMPhen and not the one departing from geometrical regularity (N1,N2-TMPhen). This fact can be clearly appreciated in Fig. 4 (bottom). As far as intramolecular interactions are concerned, the symmetric cases (V) and (VI) present different behaviours regarding these contacts. Methyl groups at the 2,9 positions inhibit structure (V) from entering into any significant (C-H) arom Á Á ÁO pds intramolecular contact, as suggested in Fig. 2 and disclosed in Table 3, where only weak, intermolecular interactions are to be found. Structure (VI), in turn, having sites 2 and 9 free, is amenable of a closer approach of (C-H) arom donors and O pds acceptors, and in fact a pair of weak bonds set up (#1b and #2b, Fig. 2 and Table 3). However, it is in structure (VII) where things depart from normal, with a second unusually short and almost straight C-HÁ Á ÁO bond internally linking the 'offending' N1,N2-TMPhen ligand and the pds anion in the same coordination sphere (interaction #1c in Table 3). In order to evaluate, at least in comparative terms the real significance of this bond (and, by extension, the similar #2c), we made some CSD (Version 5.37; Groom et al., 2016) data mining and statistical comparisons.
When comparing interaction #1c with its peers in the database, we looked for (C-H) arom Á Á ÁO intramolecular bonds with almost no restrictions (viz. 2 Å < HÁ Á ÁO < 3.0 Å ; 120 < C-HÁ Á ÁO < 180 ). The results (from ca 30000 entries analysed) are quoted in Fig. 5, where the distance (a) and angle (b) histograms, as well as the combined scatterplot (c) are presented. The two hydrogen bonds in (VI), marked in cyan, appear to be absolutely average, as are their structural consequences. The one in (VII) (marked in red), instead appears endowed with a rather unique character, in particular its nearly straight C-HÁ Á ÁO configuration. We tried to eval-  Table 3 Hydrogen-bond geometry (Å , ) for (V), (VI) and (VII).
Cg3Á Á ÁCg3 v 3.717 (3) 0 21.5 3.4577 (9) Intrachain #10c C g 4Á Á ÁCg6 vi 3.700 (2) 0.6 (2) 21.8 (3) 3.43 (2) Intrachain #11c C g 6Á Á ÁCg6 vi 3.669 (2) 0 20.9 3.4269 (9) Intrachain Symmetry code for (V): (v) 1 2 À x, 1 uate how frequent this kind of disrupting behaviour was (in terms of molecular distortions) among comparable C-HÁ Á ÁO interactions. Inspection of the occurrences found showed that they tended to appear either in monocoordinating ligands or pendant groups, in all cases with free rotations at some point in the chain, which made the C-HÁ Á ÁO contact almost irrelevant in terms of configurational energy. What makes the case in (VII) unusual is the chelating character of the ligands involved, with the concomitant deformation of the coordination polyhedron. Summarizing, there are in principle two possible reasons for the molecular geometry in (VII): either the (packing-assisted) asymmetry with which ligand (N1,N2)TMPhen binds Cd1 is the reason allowing for an unusual closeness between C1-H1 and O7, giving room to a strong hydrogen bond, or (the other way round) it is this hydrogen bond that is the cause, and the asymmetric coordination its concomitant consequence. The lack of significant intermolecular packing interactions which may justify the distortion in (VII), in addition to the outstanding character of the #1c C-HÁ Á ÁO bond seem to sustain the latter hypothesis, viz. that it is the presence of this hydrogen bond ('weak' among 'strong' but 'strong' among 'weak') which disrupts the expected symmetrical geometry in the Cd(pds)(TMPhen) 2 unit, constituting thus a rare case of a non-conventional C-HÁ Á ÁO bond being responsible for a surprising molecular configuration.

Synthesis and crystallization
Compounds (VI) and (VII) were synthesized in a similar fashion: a solution (4 ml) containing 0.050 mmol (13.5 mg) of potassium peroxodisulfate and 0.100 mmol (23.6 mg) of 3,4,7,8-tetramethyl-1,10-phenanthroline (in a 3:1 v/v methanol:water mixture) were added to 0.050 mmol of the corresponding metal acetate [Zn(OAc) 2 : 11.0 mg; Cd(OAc) 2 : 13.3 mg). An initial precipitate of extremely small needles was readily digested, but in a few days a crop of single crystals suitable for X-ray diffraction were obtained, in the form of colorless blocks. For the synthesis of (V), see Harvey et al. (2001b).

Refinement details
Data collection details and refinement results for (V), (VI) and (VII) are summarized in Table 5. The data set for (V) is the same used in the original publication (Harvey et al., 2001b) reporting the structure refined in the Cc space group. All hydrogen atoms were found in a difference Fourier map, but further idealized and allowed to ride on their parent atoms with C-H = 0.93-0.98 Å , and U iso (H) = 1.2U eq (C) or 1.5U eq (C) for methyl H atoms. A rotating model was used for The -bonded one-dimensional leitmotifs in all three structures. Stacking interactions labeled as in Table 4. H atoms have been omitted for clarity.

Figure 5
Statistical analysis of intramolecular (C-H) arom Á Á ÁO bonds as found in the literature. In cyan, those found in (VI); in red, the one in (VII). the methyl groups. For (V), a soft restraint in displacement factors was applied (RIGU in SHELXL2014).  (Sheldrick, 2008); program(s) used to refine structure: SHELXL2014 (Sheldrick, 2015); molecular graphics: SHELXTL (Sheldrick, 2008); software used to prepare material for publication:

research communications
SHELXL2014 (Sheldrick, 2015) and PLATON (Spek, 2009). where P = (F o 2 + 2F c 2 )/3 (Δ/σ) max < 0.001 Δρ max = 0.47 e Å −3 Δρ min = −0.41 e Å −3 Special details Geometry. All esds (except the esd in the dihedral angle between two l.s. planes) are estimated using the full covariance matrix. The cell esds are taken into account individually in the estimation of esds in distances, angles and torsion angles; correlations between esds in cell parameters are only used when they are defined by crystal symmetry. An approximate (isotropic) treatment of cell esds is used for estimating esds involving l.s. planes.

(VI) Bis(3,4,7,8-tetramethy-1,10-phenanthroline-κ 2 N,N′)(peroxodisulfato-κ 2 O,O′)zinc(II)
Crystal data Special details Geometry. All esds (except the esd in the dihedral angle between two l.s. planes) are estimated using the full covariance matrix. The cell esds are taken into account individually in the estimation of esds in distances, angles and torsion angles; correlations between esds in cell parameters are only used when they are defined by crystal symmetry. An approximate (isotropic) treatment of cell esds is used for estimating esds involving l.s. planes.