X-ray-determined structure of the technetium complex [Tc2(μ-CO)2(NC5H5)2(CO)6] revisited: [Tc2(μ-OMe)2(NC5H5)2(CO)6] as the correct formulation

New processing of the raw diffraction data of a compound formulated in 2008 as [Tc2(μ-CO)2(NC5H5)2(CO)6] has revealed that the bridging ligands were actually methoxide. The formation of [Tc2(μ-OMe)2(NC5H5)2(CO)6] in the room-temperature reaction of [Tc2(CO)10] with pyridine might have been facilitated at the time by the presence of trace amounts of methanol and air in the reaction mixture.


Introduction
Some of us reported previously that the room-temperature reaction of [Tc 2 (CO) 10 ] with pyridine, using the reagent itself as solvent, yields the octacarbonyl complex [Tc 2 (NC 5 H 5 ) 2 -(CO) 8 ] (1) as the unique product, which upon heating undergoes an interesting C-H bond cleavage of a pyridine molecule (Zuhayra et al., 2008).On the basis of an X-ray analysis of the above product, a molecular structure was proposed for isomer syn-1 with two bridging carbonyls displaying several unusual geometrical features not explained at the time (Fig. 1): (i) a strong pyramidalization of the bridgehead C atoms, with unusually small displacement parameters and very large C-O separations of ca 1.45 A ˚, actually close to the reference value of 1.42 A ˚for a C(sp 3 )-O single bond (Cordero et al., 2008), and far larger than the reference value of 1.21 A ˚for a double bond between these atoms (Pyykko ¨& Atsumi, 2009); and (ii) a large intermetallic separation of ca 3.37 A ˚, far above that of the parent complex [Tc 2 (CO) 10 ] (ca 3.03 A ˚; Bailey & Dahl, 1965;Sidorenko et al., 2011), and inconsistent with the formulation of a single Tc-Tc bond, as required by application of the 18-electron rule to complex syn-1.Recently, we used density functional theory (DFT) calculations to find that the most likely structure of 1 would display only terminal carbonyls and a staggered conformation, as observed in the parent precursor, and that the crystals analysed by X-ray diffraction in 2008 would most likely correspond to either the hydroperoxide-bridged ditechnetium(I) complex syn-[Tc 2 (�-OOH) 2 (NC 5 H 5 ) 2 (CO) 6 ] (syn-2) or its methoxide-bridged analogue syn-[Tc 2 (�-OMe) 2 (NC 5 H 5 ) 2 (CO) 6 ] (syn-3) (Garcı ´a-Vivo ´& Ruiz, 2020).This prompted us to revise the structure determination of compound syn-1 by performing new refinements using the original raw diffraction data, which is the purpose of the present article.As will be shown below, the new refinements indicate beyond doubt that the crystal actually analyzed at the time was not that of compound syn-1 but that of the methoxide-bridged complex syn-[Tc 2 (�-OMe) 2 (NC 5 H 5 ) 2 (CO) 6 ] (syn-3), whereby the 'anomalous' geometrical parameters mentioned above now become 'as expected'.

Experimental
Diffraction data were collected on a Siemens Nicolet Syntex R3m/V diffractometer using graphite-monochromated Mo K� radiation (� = 0.71073 A ˚). Intensities were measured by fineslicing '-scans and corrected for background, polarization and Lorentz effects.The original structure of syn-1 was solved by direct methods and refined with the programs SHELXS86 and SHELXL93 (Sheldrick, 2008) by a full-matrix least-squares method based on F 2 (Zuhayra et al., 2008).

Refinement
Crystal data, data collection and structure refinement details for syn-1, syn-2 and syn-3 are summarized in Table 1.All carbon-bound H atoms were calculated at their optimal positions and treated as riding on their parent atoms using isotropic displacement parameters 1.2 (or 1.5 in the case of methyl groups) times larger than the U eq values of the respective parent atoms.The methyl groups in syn-3 were calculated as idealized rotating groups.We could not recover Note: (a) data taken from Zuhayra et al. (2008).Computer programs: SHELXL2017 (Sheldrick, 2015b) in WinGX (Farrugia, 2012), SHELXT2014 (Sheldrick, 2015a) and OLEX2 (Dolomanov et al., 2009).

Table 2
Selected parameters (A ˚, A ˚2) for structural determinations following formulations as syn-1 to syn-3.from the stored old data (recorded some 20 years ago) all the information currently required for standard CIF files, and this caused the appearance of some A-level alerts in the corresponding checkCIF reports for syn-2 and syn-3.

Results and discussion
The small size of the displacement ellipsoids of the bridgehead 'carbon' atoms (C4 and C5) in the original structure determination of syn-1, compared to those of the corresponding O atoms (O4 and O5; Fig. 1 and Table 2), suggested that positions C4 and C5 might actually correspond to atoms having a higher number of electrons (Stout & Jensen, 1989).Moreover, the theoretical calculations mentioned above indicated that replacing the bridging carbonyl ligands in syn-1 with either OOH (peroxide) or OMe (methoxide) groups would yield complexes with geometries matching the anomalous features of the original structural determination (Garcı ´a-Vivo ´& Ruiz, 2020).We then proceeded to make new refinements with the original raw diffraction data under both hypotheses (syn-2 and syn-3, respectively).Both refinements converged satisfactorily to give improved fitting parameters, compared to the original refinement based on the formulation syn-[Tc 2 (�-CO) 2 (N-C 5 H 5 ) 2 (CO) 6 ] (Fig. 2, and Tables 1 and 2), but there were some significant differences between them: (i) the R 1 , wR 2 and goodness-of-fit (GOF) values were better for syn-3.(ii) the U eq values for the heavy atoms at the bridging positions (OO or OC) were more similar to each other in the case of syn-3; in contrast, the U eq values for the O(H) atoms in syn-2 were almost three times the value of the corresponding bridgehead O atom.This is clearly reflected in the significantly smaller values of ca 0.01 A ˚2 in the difference between the meansquare displacement amplitudes (�MSDA) for the C4/O4 or  C5/O5 pairs in syn-3, as expected for mutually bonded atoms (Hirshfeld, 1976), which can be compared with values of ca 0.04 A ˚2 for the corresponding pairs in either syn-2 or syn-1 (Table 2).Moreover, we note that the average C-O bond length for the bridging methoxide groups in syn-3 (ca 1.42 A ˚) exactly matches the reference value for a C(sp 3 )-O single bond.In contrast, the average O-O bond length of 1.43 A ˚in the formulation as syn-2 falls below the values of 1.45-1.50A typically determined for OOR-bridged complexes (Garcı ´a-Vivo ´& Ruiz, 2020).All of this provides conclusive evidence for the presence of methoxide groups bridging the Tc atoms in the complex under discussion.It is thus concluded that the crystal analyzed at the time actually was not one of compound syn-1 but one of the methoxide-bridged complex syn-[Tc 2 (�-OMe) 2 (NC 5 H 5 ) 2 (CO) 6 ] (syn-3).We finally note that the geometrical parameters obtained for this complex are similar to those determined previously for different rhenium complexes with dimetal cores of the type syn-[Re 2 (�-OR) 2 L 2 (CO) 6 ] having bridging alkoxide or hydroxide ligands and terminal pyridine, dipyridyl and polypyridyl ligands (Garcı ´a-Vivo ´& Ruiz, 2020).The latter belong to a relatively large family of complexes which have been studied extensively because of their photophysical and chemical properties, host-guest interactions and biological activity.
After concluding that the crystal analyzed at the time, formed through crystallization from acetone/n-hexane of the bulk product obtained when reacting [Tc 2 (CO) 10 ] with pyridine at room temperature, corresponds to the methoxidebridged complex syn-3 rather than the simple substitution product syn-1, the question then to be answered is from where could the methoxide ligands possibly arise.Unfortunately, we are not in a position to reproduce the above synthetic procedure in our laboratories, so we can only speculate about its possible origin.We currently trust that complex syn-3 might just have been a very minor side product formed along with the major product, which just happened to crystallize first from the reaction mixture.Interestingly, we note that many dirhenium polypyridyl complexes with metal cores of the type syn-[Re 2 (�-OR) 2 L 2 (CO) 6 ] have been made by reacting [Re 2 -(CO) 10 ] with stoichiometric amounts of the pertinent N-donor ligand in the corresponding alcohol (ROH) or water, although high temperatures are typically required to form these products.However, a separate experiment carried out previously by us revealed that stirring [Re 2 (CO) 10 ] in pyridine at room temperature for 4 d caused no detectable transformation on the Re 2 substrate, unless air is admitted into the reaction flask (Garcı ´a-Vivo ´& Ruiz, 2020).Based on the above indirect pieces of evidence, we tend now to think that formation of the methoxide-bridged complex syn-3 during the slow reaction of [Tc 2 (CO) 10 ] with pyridine at room temperature (5 d) might have followed from the presence of trace amounts of methanol and air in the reaction mixture.

Conclusion
The raw diffraction data of the compound formulated in 2008 as syn-[Tc 2 (�-CO) 2 (NC 5 H 5 ) 2 (CO) 6 ] have now been re-processed under the hypothesis that the bridging ligands might actually be either hydroperoxide or methoxide ligands.The latter option proved to be the correct one, as it leads not only to better agreement parameters, such as R, wR or GOF, but also to chemically more sensible interatomic distances and displacement parameters for the non-H atoms of the bridging ligands.The formation of syn-[Tc 2 (�-OMe) 2 (NC 5 H 5 ) 2 (CO) 6 ] in the room-temperature reaction of [Tc 2 (CO) 10 ] with pyridine might have been facilitated at the time by the presence of unnoticed trace amounts of methanol and air in the reaction mixture.

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.

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.

Table 1
Experimental details for structural determinations of complexes syn-1 to syn-3.