Bis(catecholato-κ2 O,O′)bis(dimethyl sulfoxide-κO)titanium(IV)

Bis(catecholato)bis(DMSO)titanium crystallizes with two crystallographically independent molecules related by pseudo-glide symmetry in a distorted octahedral O6 donor-atom geometry around titanium.

Titanium catecholate complexes have found various uses: Titanium triscatecholate complexes of the alkaline earth metals were utilized as molecular precursors to a number of  M II TiO 6 -type perovskites (Ali & Milne, 1987;Marteel-Parrish et al., 2008). Titanium catecholates have been exploited as catalysts in acetylene hydrogenation (Bazhenova et al., 2016) while three-dimensional titanium catecholate frameworks of high proton conductivity (Nguyen et al., 2015) and titanium catecholate-based MOFs have been described (Cao et al., 2020). Metal catecholates have been suggested as adsorbents for toxic gases (Bobbitt & Snurr, 2018). Titanium triscatecholates were also used to self-assemble a potential bimodal contrast agent (Dehaen et al., 2012). A number of heteroleptic mono-and multi-nuclear titanium complexes do contain titanium catechol units (Sakata et al., 2010;Bazhenova et al., 2016;Sonströ m et al., 2019;Passadis et al., 2020) (for further examples, see Database survey below). Most prominently, titanium tris-catecholates have been used as versatile building blocks in a range of supramolecular, oligonuclear homo-and hetero-metal-ligand cluster assemblies (Brü ckner et al., 1998;Caulder et al., 2001;Albrecht et al., 2008Albrecht et al., , 2019. Reaction of TiCl 4 under anhydrous conditions in toluene generates a brick-colored amorphous powder of the diprotonated titanium tris-catecholate complex 2 H (Davies & Dutremez, 1990). We found that this compound dissolves sufficiently enough in ambient-temperature DMSO-d 6 to record a simple 1 H NMR spectrum, showing only two multiplets of equal integration (at 6.84 and 6.75 ppm), corresponding to the ortho-and meta-hydrogens on three nearidentical catecholate moieties (the counter-cations -protonsare believed to be dynamically associated with the trigonal faces of the pseudo-octahedral coordination sphere formed by the six catecholate oxygens) ( Fig. 2A). Upon heating 2 H in DMSO (or DMSO-d 6 ), its solubility increases drastically. When followed by 1 H NMR spectroscopy, the formation of a new species with two catecholate signals (m at 6.42 and 6.12 ppm, in 2:2 intensity) and 1 equivalent of free catechol (m at 6.73 and 6.60, br s at 8.1 ppm, all 1:1:1) can be observed (Fig. 2B). Upon cooling, dark red-orange crystals of the title compound 3 formed. Isolated and analyzed by NMR spectroscopy, they exhibit only the signals for the new species formed (Fig. 2C) and the signals for two slightly high-fieldshifted DMSO molecules (not shown).
The material was also analyzed by single crystal X-ray diffraction (Fig. 3). Evidently, one protonated catecholate ligand (i.e., catechol 1) of the starting triscatcholate complex 2 H was exchanged for two DMSO molecules, coordinating through their oxygen atoms in adjacent positions, thus forming a neutral, heteroleptic, mononuclear octahedral complex. Details of the structural arrangement will be discussed in the Structural commentary section below.
The UV-vis spectrum of the orange solution of 3 is overall similar to that of the starting material 2 H ; both spectra are dominated in the visible range by broad, little-structured catecholate ligand-to-metal charge-transfer bands (for 3, l max = 441 nm; half-height width > 150 nm; Fig. 4). In comparison to the spectrum of 2 H , all bands for 3 are bathochromically shifted.

Structural commentary
The title complex 3, having solution C 2 symmetry, crystallizes as a racemic mixture with two crystallographically independent molecules in the monoclinic space group P2 1 /c (Fig. 3  H NMR spectra (400 MHz, DMSO-d 6 ) of (A) complex 2 H dissolved at ambient temperature; (B) of complex 2 H at $373 K, showing the presence of complex 3 and free catechol 1; (C) of isolated crystals of 3 precipitated from DMSO at ambient temperature (* indicates the presence of residual 1).

Figure 3
The two crystallographically independent molecules of 3. View along the a axis. Molecules A and B are related by pseudo-glide operations (see discussion for details).
For both molecules, the solution C 2 symmetry is broken in the solid state, and the Ti atoms are each bonded to two chelating catecholate and two monodentate O-coordinated dimethylsulfoxide ligands. The Ti-O Cat bond distances range from 1.9113 (19) (Borgias et al., 1984). The four Ti-O DMSO bond distances in 3 are at 2.0214 (19) to 2.0416 (18) Å significantly longer than the Ti-O Cat bond lengths, as would be expected for neutral and uncharged DMSO ligands.
The bond lengths in 3 also compare well with those of the [biscatecholate-bis-DMF]titanium complex 6, the DMF analogue to the title compound (Bazhenova et al., 2016) and the only other reported heteroleptic mono-nuclear and uncharged bis-catecholate titanium complex with MO 6 metal coordination (see Database survey). The Ti-O cat bond lengths in 6 are 1.9003 (12) and 1.9181 (12) Å for the oxygen atoms trans to the DMF molecules, 1.9408 (11) and 1.9483 (11) Å when trans to another O cat atom, and 2.0396 (12) and 2.0736 (12) Å for the Ti-O DMF bond lengths. They thus closely mirror those found in 3.
The small bite angles of the chelating catecholate anions induce substantial distortions from idealized octahedral symmetry. The catecholate O-Ti-O angles in 3 are 80.73 (7) and 81.00 (8)  This more pronounced deviation from ideal octahedral symmetry for molecule B of 3 is also confirmed by a more holistic analysis, using a normalized root-mean-square deviation algorithm to calculate the distortion from octahedral symmetry as implemented in the program SHAPE (Pinsky & Avnir, 1998;Alvarez et al., 2002;Casanova et al., 2004). The calculated continuous shape measures (CShM's) relative to ideal reference octahedral symmetry are 1.434 for 2 Et3NH , 1.513 for 6, 1.491 for less distorted molecule A of 3, and 1.854 for molecule B (Table 1). Shape measures may be between 0 and 100 where zero represents a perfect fit for the selected shape, and CShM values of less than 1.0 are usually interpreted as only minor distortions from the reference shape. Values between 1 and 3 indicate substantial distortions, but the reference shape still provides a good stereochemical description (Cirera et al., 2005). For the four cases analyzed here, the CShM's for the next best fit, trigonal prismatic, are all around 10 ( Table 1). The geometries of 3, 6 and 2 Et3NH are thus best described as distorted octahedral, being far removed from fitting any other polygon.
The two independent molecules in compound 3 are related to each other by crystallographic pseudosymmetry. Complex 3 crystallized in a pseudo-orthorhombic setting with a refined angle of 90.0445 (9) , and emulates space group Pbca with additional b-and a-glide operations along the a-and c-axis directions. Exact translational symmetry is broken by modulation of one of the catecholate and one of the DMSO ligands, as discussed below. The metric pseudosymmetry allows for the Normalized UV-vis spectrum of title compound 3 (DMSO) in comparison to that of the starting triscatecholate 1 (H 2 O). Table 1 Continuous shape measures (CShM's) relative to ideal reference octahedral symmetry for 2 Et3NH , 3 and 6. possibility of twinning. Indeed, the crystal investigated was found to be pseudo-merohedrically twinned by symmetry elements of the emulated orthorhombic symmetry. Application of the twin transformation matrix 1 0 0, 0 À1 0, 0 0 À1 yielded close to equal twin components with a refined twin ratio of 0.5499 (7) to 0.4401 (7). A root-mean-square overlay of the two molecules yields an r.m.s. deviation of 0.459 Å , indicating substantial variation between the geometries of molecules A and B (Fig. 5). A similar overlay based on only the titanium and oxygen atoms gives a much smaller value of only 0.056 Å , indicating that the main differences between the two complexes is rooted in the ligands, even though there are small and noticeable differences for the TiO 6 cores as well (with molecule B deviating more from ideal octahedral symmetry than molecule A, as discussed above). The main distinction between the two molecules is, however, associated with substantial twists and torsions of the catecholate and DMSO ligands. The r.m.s. overlay reveals a close match of one of the catecholate ligands and one of the DMSO ligands. The other catecholate and DMSO ligands, on the other hand, do show substantial variation between the two molecules. The catecholate of C7-C12 undergoes a twist-motion by a slight rotation around the O cat -O cat axis. In molecule A (blue in the overlay), the catecholate ligand is close to coplanar with the titanium atom, while in molecule B (red in the overlay) the catecholate and the TiO(Cat) 2 plane are clearly angled against each other. The deviations of the Ti atoms from the mean catecholate planes are 0.049 (2) and 0.349 (2) Å for molecules A and B, respectively. The angle between the mean catecholate and Ti(OCat) 2 planes is 1.97 (9) for complex A, but a much larger value of 13.6 (8) for complex B.
The other main difference between the two molecules is a rotation of about 14 for one of the two DMSO ligands around the Ti-O bond, which can be expressed via the torsion angle O5-Ti1-O6-S2 (S2 is the sulfur atom of the rotated DMSO ligand, O5 the oxygen atom of the other invariant DMSO ligand). These torsion angles are 165.13 (16) for molecule A, and 151.10 (14) for molecule B. The largest overall motion is observed for the DMSO methyl groups of C15 [1.774 (3) Å in the r.m.s. overlay based on the titanium and oxygen atoms].
The differences in molecular geometry between molecules A and B and the modulation by pseudo-orthorhombic symmetry are closely related, showing molecules A and B as they are related to each other by a pseudo b-glide perpendicular to (100) (Fig. 6). In addition to the variations in molecular geometry seen in the molecule overlay, a very slight rotation of the entire complex is also observed.

Supramolecular features
The  View of 3 down the c axis, showing the modulation by pseudo-Pbca symmetry. Molecules color coded by symmetry equivalence (red: molecule A, blue molecule B). Atom labels included for Ti, DMSO S and O atoms, and for C atoms with the largest modulation. 50% probability ellipsoids. Table 2 Hydrogen-bond geometry (Å , ). methyl groups as hydrogen-bond donors, and catecholate and DMSO oxygen atoms as the respective acceptors. Hydrogen bonds with CÁ Á ÁO distances below 3.50 Å are given in Table 2. Some of these HÁ Á ÁO distances are unusually short for C-HÁ Á ÁO interactions, with HÁ Á ÁO distances as short as 2.27 and 2.34 Å , approaching distances usually only observed for classical hydrogen bonds involving acidic hydrogens. This might indicate stronger than usual interactions with a possibly larger influence on the packing and molecular arrangement in the solid state than usually observed for C-HÁ Á ÁO interactions.
When plotting the C-HÁ Á ÁO hydrogen bonds (Fig. 7), it becomes evident that the interactions are different for the two molecules, despite their close relationship by a pseudo-glide operation. Interactions involving the less modulated fragments of 3A and 3B exhibit similar hydrogen-bonding environments. C13 and C14 of the less-modulated DMSO molecule exhibit the same type of hydrogen bonds to O1, O4 and O5 in neighboring molecules (see Table 2 for symmetry operators and numerical values). The exact bond lengths for C14 vary slightly, a bond to O5 is broken and one to O4 significantly elongated for molecule B, but the overall hydrogen-bonding pattern for this DMSO fragment is very similar for both molecules A and B. This is not the case for the other significantly modulated DMSO molecule. For 3B, three significant C-HÁ Á ÁO interactions are observed, towards O1A, O3A and O4A of neighboring entities. None of these are found for 3A. All hydrogen-to-oxygen distances are beyond what could be still regarded as attractive and stabilizing. Methyl carbon atom C16A is at a distance of 3.128 Å from O1B, close enough for a C-HÁ Á ÁO hydrogen bond to be suspected, but its hydrogen atoms are rotated such that the HÁ Á ÁO distances are > 2.8 Å , and the C-HÁ Á ÁO angles are unfavorable at 97.9 and 99.5 (H-atom positions were clearly resolved in difference-density maps and were allowed to rotate to fit the experimental electron density). The shortest distance involving the H atoms of C16A is instead towards C1B of a neighboring catecholate ring (2.734 Å , shown as a green dashed line in Fig. 7), and C15A does not exhibit any HÁ Á ÁX contacts < 2.8 Å . This clear difference between the hydrogen-bonding interactions of C15 and C16 in the two molecules is clearly related to the modulation that breaks the exact Pbca glide symmetry in the structure of 3. It is not clear whether the ability to form stronger interactions is the cause for the modulation, or whether the modulation causes the differences in intermolecular interactions and the modulation itself is caused by other less-directional forces such as dispersive interactions. Most likely the concerted effects of both modulation and intermolecular interactions reinforcing and stabilizing each other lead to the observed packing of the molecules.
The conversion of triscatecholate complexes to heteroleptic complexes has been observed previously, as the hydroxideinduced displacement of a catecholate ligand from 2 Et3NH to form the bis-(-oxo-bridged) biscatecholate 4 exemplifies (Borgias et al., 1984). Direct syntheses are also known (Sakata et al., 2010). For example, treatment of titanium methoxide with a methanolic solution of catechol 1 under ambient conditions resulted in the formation of a dinuclear heteroleptic complex 5 with a mixture of catechol/catecholate and methanol/methanolate ligands (Bazhenova et al., 2016 Directional interactions in 3, viewed down the c axis. Red: molecule A, blue: molecule B. Lighter colored molecules are generated by crystal symmetry. Dark-blue dashed lines: C-HÁ Á ÁO bonds with HÁ Á ÁO distance < 2.5 Å ; light-blue dashed lines: C-HÁ Á ÁO bonds with HÁ Á ÁO distance between 2.5 and 2.62 A; green dashed lines C-HÁ Á Á contacts. 50% probability ellipsoids. complex 6, the DMF analogue of the title compound (CCDC 1489371; Bazhenova et al., 2016). Neutral and monometallic complexes of this kind are exceedingly rare. A search of the CSD yielded complex 6 as the only other heteroleptic mononuclear, neutral bis-catecholate complex with TiO 6 metal coordination; complex 3 is only the second such complex.

Refinement
Crystal data, data collection and structure refinement details are summarized in Table 3. The structure exhibits pseudoorthorhombic symmetry (Pbca) and is twinned by a 180 rotation around the a-or c-axis. Application of the transformation matrix 1 0 0, 0 À 1 0, 0 0 À 1 yielded a twin ratio of 0.5399 (7):0.4401 (7). The pseudo-orthorhombic symmetry is broken by modulation of one of the catecholate rings by up to 1.4 Å , and one of the DMSO ligands by over 1.7 Å .
C-H bond distances were constrained to 0.95 Å for aromatic C-H and to 0.98 Å for aliphatic CH 3 moieties, respectively. U iso (H) values were set to a multiple of U eq (C) with 1.5 for CH 3 and 1.2 for C-H units. Reflections 112, 112 and 013 were affected by the beam stop and were omitted from the refinement.

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. Refinement. The structure exhibits pseudo-orthorhombic symmetry (Pbca) and is twinned by a 180 degree rotation around the a or c-axis. Application of the twin matrix 1 0 0, 0 -1 0, 0 0 -1 yielded a twin ratio of 0.5399 (7) to 0.4401 (7). The pseudo-orthorhombic symmetry is broken by modulation of the phenylene rings by up to 1.4 Angstrom.