C 2-isomer of [Pd(tfd)]6 [tfd is S2C2(CF3)2] as its benzene solvate: a new member of the small but growing class of homoleptic palladium(II) monodithiolenes in the form of hexameric cubes

S 6- and C 2-symmetric structure types are possible for homoleptic palladium monodithiolenes in the form of hexameric cubes. The C 2-isomer of [Pd(S2C2(CF3)2)]6 is described here and the question of whether additional isomers are possible is answered.

The title compound, hexakis[ 3 -1,2-bis(trifluoromethyl)ethene-1,2-dithiolato]octahedro-hexapalladium(II), [Pd(C 4 F 6 S 2 )] 6 , crystallizes as its benzene solvate, [Pd(tfd)] 6 Á2.5C 6 H 6 , where tfd is the dithiolene S 2 C 2 (CF 3 ) 2 . The molecular structure of [Pd(tfd)] 6 is of the hexametallic cube type seen previously in three examples of hexameric homoleptic palladium monodithiolene structures. All structures have in common: (a) the cluster closely approximates a cube containing six Pd II atoms, one at the centre of each cube face; (b) 12 S atoms occupy the mid-points of all 12 cube edges, providing for each Pd II atom an approximately square-planar S 4 environment; (c) each S atom is part of a dithiolene molecule, where the size of the dithiolene ligand necessitates that only sulfur atoms on adjacent cube edges can be part of the same dithiolene. This general cube-type framework has so far given rise to two isomeric types: an S 6 -symmetric isomer and a C 2 -chiral type (two isomers that are enantiomers of each other). The structure of [Pd(tfd)] 6 is of the C 2 -type. Out of the 12 CF 3 groups, three are rotationally disordered over two positions. Further, we answer the question of whether additional, previously undiscovered, isomers could follow from the cube rules (a) through (c) above. An exhaustive analysis shows that no additional isomers are possible and that the list of isomers (one S 6 isomer, two C 2 enantiomers) is complete. Each isomer type could give rise to an unlimited number of compounds if the specific dithiolene used is varied.

Structural commentary
The molecular structure of [Pd(tfd)] 6 is shown in Fig. 1, where the second position for the rotationally disordered trifluoromethyl groups (attached to C1, C13, and C14), the second position for disordered atom C15, as well as the benzene solvate molecules are not displayed. The structure has approximate, non-crystallographic, C 2 symmetry (C 2 through Pd2 and Pd4). The gross features of this cube-like structure will be discussed first, followed by details such as bond lengths. The structure is of the hexametallic cube type seen previously in the hexameric homoleptic palladium monodithiolenes characterized by Stiefel and co-workers (Beswick et al., 2002), Stibrany (Stibrany, 2012), and Rawson and co-workers (Wrixon et al. 2015). All structures have in common: (a) the cluster closely approximates a cube containing six Pd II atoms, one at the centre of each cube face; (b) 12 S atoms occupy the midpoint of all 12 cube edges, providing for each Pd II atom an approximately square-planar S 4 environment; (c) each S atom is part of a dithiolene molecule, where the size of the dithiolene ligand requires that only sulfurs on adjacent cube edges can be part of the same dithiolene. This general cubetype framework has so far given rise to two isomer types: one S 6 -symmetric isomer seen for the charge-neutral palladium complex of S 2 C 2 (COOMe) 2 (Beswick et al., 2002) and for a partially reduced complex involving the same ligand (Stibrany, 2012). The charge-neutral complex Pd(dithiolene) 6 with the dithiolene S 2 C 6 H 2 (OMe) 2 was found by Rawson and coworkers (Wrixon et al., 2015) to have a different, C 2symmetric, structure. The two isomeric types are shown here schematically, inscribed into a cube (Fig. 2).
The question of whether additional isomers are possible remained open and is answered here (Figs. 3 and 4). The starting point is the constraint that 12 donor atoms (such as sulfur) reside at the midpoint of the 12 cube edges and that metal atoms (such as Pd) occupy the centres of all cube faces. The ligand length constraint forbids trans-spanning placement of a chelate bridge, and chelates can only bridge the short distance between adjacent edge positions. All possible isomers following from this framework are explored in an exhaustive fashion (Figs. 3 and 4), following the initially arbitrary placement of the first bridge. It results that no additional isomers are possible and that the list of isomers (one S 6 isomer, two C 2 enantiomers) is complete. It is worthwhile noting that the S 6 isomer cannot have a dipole moment, by virtue of its inversion centre, while the C 2 isomer has a dipole.
With these general results for homoleptic square-planar metal-based short-span chelates in the fom of hexameric cubes [M(L 2 )] 6 in hand, we return to discussing the details of the structure of [Pd(tfd)] 6 ( Fig. 1). It is notable that not all Pd-S distances are the same. Shorter distances, on average 2.294 Å (12 values; standard deviation = 0.008 Å ), are found for the Pd-S distances within an approximately planar C 2 S 2 Pd five-    Graphical proof that only one S 6 isomer and two (enantiomeric) C 2 isomers are possible for homoleptic palladium monodithiolenes in the form of hexameric cubes, given the 'cube rules' discussed in the text. Continues in Fig. 4. research communications Figure 4 Continued from Fig. 3. membered ring. Longer distances, on average 2.364 Å (12 values; standard deviation = 0.01 Å ) are found for coordination of an S atom outside its own C 2 S 2 Pd ring onto a different Pd II atom at an approximately right angle to the fivemembered ring. All bonds to Pd2, the Pd atom at the bottom of the Pd 5 (tfd) 4 C 4 -symmetric 'box' (Fig. 2) are long. All bonds to Pd 4 , the Pd II atom in the Pd(tfd) 2 'lid' are short. The charge on the tfd ligand can be seen from the C-C distance within the chelate ring, which is short (double bond) for the dianion (ene-dithiolate; C-C distance of 1.35 Å or shorter expected) and long for monoanionic tfd (C-C bond order = 1.5; C-C distance of 1.38 Å expected), as is known from Tang et al. (2009) and Kogut et al. (2006) (see analysis in Hosking et al., 2009). The chelate C-C bond distances in the current structure of [Pd(tfd)] 6 average to 1.339 Å (six values; standard deviation = 0.006 Å ) and indicate a dianionic chelate. Charge balance necessitates that all palladium atoms are in the oxidation state 2+, which is also supported by the coordination geometry around each Pd II atom, which is approximately square-planar, as expected for a d 8 metal centre. The structure may thus be described as a charge-neutral C 4 -symmetric Pd 4 (dithiolene) 4 tiara capped on one side with a Pd 2+ dication and on the other side with a Pd(tfd) 2 2À dianion. While the structure is likely more charge balanced than this zwitterionic description implies, this description suggests a direction of the dipole moment.

Supramolecular features
Molecules of [Pd(tfd)] 6 and benzene solvate molecules pack via contacting van der Waals surfaces. There are no particularly short intermolecular distances (such as hydrogen bonds).

Synthesis and crystallization
General specifications: All manipulations were carried out under an inert (N 2 ) atmosphere using standard glove box (M. Braun UniLab) and Schlenk techniques. NMR solvents were obtained from Cambridge Isotope Laboratories. Solvents were purified prior to use by vacuum distillation from purple sodium benzophenone ketyl. NMR data were obtained on a Bruker Avance III 400 MHz spectrometer. Pd 2 dba 3 was obtained from Sigma-Aldrich. S 2 C 2 (CF 3 ) 2 (tfd) was synthesized as in Harrison et al. (2006).
Synthesis: A pyrex reaction vessel containing 10 ml of toluene, 80 mg of Pd 2 dba 3 (175 mmol of Pd) and 80 ml (350 mmol) of tfd was heated to 353 K overnight. At the end of the reaction, the dark-red solution had turned an intense brown. Solvent and volatiles were removed under vacuum at room temperature, followed by heating to 383 K for 4 h, also under vacuum. NMR spectroscopy (C 6 D 6 solvent) showed a complex mixture, as indicated by multiple 19 F resonances, chiefly two intense quartets (J F-F = 14.5 Hz) at À58.7 ppm and À59.3 ppm but also a large number of overlapping signals between À56 and À58 ppm. C 2 -[Pd(tfd)] 6 is clearly not the only species in solution, as it would give rise to six distinct fluorine signals (quartets) in equal intensity. It seems likely that multiple species are in equilibrium in the reaction mixture, and C 2 -[Pd(tfd)] 6 might crystallize from non-polar solvents relatively early due to its dipole moment, which makes it less soluble in non-polar solvents compared to nonpolar species. Dissolving the sample in a minimal amount of benzene, followed by storage at 285 K for one week, led to the formation of crystals suitable for X-ray crystallography.

Refinement
Crystal data, data collection and structure refinement details are summarized in Table 1. The F atoms of three of the -CF 3 groups were refined as disordered over two sets of sites with the ratios of refined occupancies being 0.898 (6):0.102 (6) for F1/F2/F3, 0.784 (7):0.216 (7) for F19/F20/F21 and 0.749 (9): 0.251 (9) for F22/F23/F24. Both the major and minor compo- 2.95, À1.37 nents were refined with anisotropic displacement parameters. In the -CF 3 group containing F19/F20/F21, the attached atom C15 was also refined over two sets of sites with occupancies of 0.784 (7) and 0.216 (7). The SIMU command in SHELXL (Sheldrick, 2015) was used to restrain the anisotropic displacement parameters of the disordered atoms. The asymmetric unit contains 2.5 benzene solvent molecules. One benzene molecule is disordered about an inversion centre and hence has 0.5 occupancy. The RIGU command in SHELXL was used to restrain the anisotropic displacement parameters of the 0.5 occupancy benzene molecule. The H atoms bonded to C atoms were placed in calculated positions C-H = 0.95 Å ) and included in the refinement in a riding-motion approximation with U iso (H) = 1.2U eq (C). Data collection: APEX2 (Bruker, 2014); cell refinement: APEX2; data reduction: SAINT (Bruker, 2014); program(s) used to solve structure: SHELXS97 (Sheldrick, 2008); program(s) used to refine structure: SHELXL2014 (Sheldrick, 2015); molecular graphics: ORTEP-3 (Farrugia, 2012); software used to prepare material for publication: SHELXTL (Sheldrick, 2008).

Hexakis[µ 3 -1,2-bis(trifluoromethyl)ethene-1,2-dithiolato]-octahedro-hexapalladium(II) benzene 2.5-solvate
Crystal data [Pd 6 (C 4 F 6 S 2 ) 6 ]·2. 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.