Effects of Remote Ligand Substituents on the Structures, Spectroscopic, and Magnetic Properties of Two-Coordinate Transition Metal Thiolate Complexes

The first row transition metal(II) dithiolates M(SAr4)2 (Ar iPr4 = C6H3-2,6-(C6H3-2,6-iPr2)2, M = Cr (1), Mn (3), Fe (4), Co (5), Ni (6), and Zn (7)), Cr(SAr6)2 (2) (Ar Me6 = C6H3-2,6-(C6H22,4,6-Me3)2) and the ligand transfer reagent (NaSAr 4)2 (8) are described. In contrast to their M(SAr6)2 (M = Cr, Mn, Fe, Co, Ni, and Zn; Ar iPr6 = C6H3-2,6-(C6H2-2,4,6-iPr3)2) congeners, which differ from 1 and 3 – 6 in having para-isopropyl groups on the flanking aryl rings of the terphenyl substituents, compounds 1 and 4 – 6 display highly bent coordination geometries with S–M–S angles of 109.802(2) (1), 120.2828(3) (4), 91.730(3) (5), and 92.68(2)° (6) as well as relatively close metal–flanking aryl ring η interactions with metal–centroid distances of 2.11477(6) (1), 1.97188(3) (2), 2.15269(6) (4), 1.62058(9) (5), and 1.724(8) Å (6). However, the


EXPERIMENTAL SECTION
All manipulations were carried out under anaerobic and anhydrous conditions by using Schlenk techniques under a dinitrogen atmosphere or in a Vacuum Atmospheres HE-43 drybox. Solvents were dried by the method of Grubbs and coworkers, 48 stored over potassium or sodium, and then degassed by the freeze-pump-thaw method. All physical measurements were made under strictly anaerobic and anhydrous conditions. IR spectra were recorded as Nujol mulls between CsI plates on a Perkin-Elmer 1430 spectrometer. UV-visible spectra were recorded as dilute hexane or toluene solutions in 3.5 mL quartz cuvettes using an Olis 17 Modernized Cary 14 UV/Vis/NIR spectrophotometer or a HP 8452 diode array spectrophotometer. Melting points were determined on a MEL-TEMP II apparatus using glass capillaries sealed with vacuum grease. Unless otherwise stated, all materials were obtained from commercial sources and used as received.
MnCl2 and FeCl2 were dehydrated from FeCl2 . 4H2O and MnCl2 . 6H2O respectively, by following a similar dehydration procedure to that previously reported for MnCl2. 49   X-ray Crystallography. Crystals suitable for X-ray diffraction studies were removed from the Schlenk tube under a flow of nitrogen and immediately covered with a layer of hydrocarbon oil.
A single crystal was mounted on a glass fiber attached to a copper mounting pin and placed in a low-temperature nitrogen stream. 54 Data for 1, 2, 5, and 7 were collected at 90(2) K with (λ = 1.5418 Å) Cu Kα1, and data for 4 and 8 were collected at 90(2) K with (0.71073 Å) Mo Kα1 radiation using a Bruker DUO diffractometer in conjunction with a CCD detector. Data for 3 were collected at 101 K with (0.71073 Å) Mo Kα1, and data for 6 were collected at 100 K with (λ = 1.54178 Å) Cu Kα1 radiation, using a Bruker Kappa diffractometer in conjunction with a CCD detector. The collected reflections were corrected for Lorentz and polarization effects and for absorption by use of Blessing's method as incorporated into the program SADABS, 55,56 The structures were solved by direct methods and refined with the SHELXTL (2012, v.6.1) or SHELXTL (2013) software packages. 57 Refinement was by full-matrix least-squares procedures with all carbon-bound hydrogen atoms included in calculated positions and treated as riding atoms. The thermal ellipsoid plots were drawn using OLEX2 software. 58 A summary of crystallographic and data collection parameters is given in the SI. Effective magnetic moments were calculated using multireference ab initio methods. Static electron correlation was accounted using the complete active space self-consistent field (CASSCF) method. 68,69 The calculations were carried out on the model species M(SPh)2 and M(SC6H3-2,6-(C6M5))2 in different geometries (see SI for further details). An active space consisting of the five 3d orbitals and the required 3d electrons were used. All states in each possible multiplicity were solved in a single state-averaged calculation. Spin-orbit coupling (SOC) was introduced using quasi-degenerate perturbation theory (QDPT) approach, 70 where the matrix of the SOC operator is constructed in a basis of the CASSCF eigenstates, and the full Hamiltonian is then diagonalized to yield the spin-orbit coupled states and eigenvalues. Dynamic electron correlation outside the active orbital space was included using the quasi-degenerate Nelectron valence state perturbation theory at the second order (NEVPT2) in its strongly contracted formulation. [71][72][73][74] Scalar relativistic effects were treated using the standard secondorder Douglas-Kroll-Hess (DKH) transformation of the one-particle operator 75,76 along with the correct picture-change effects. The relativistically contracted DKH-def2-TZVPP basis was used for the metal ions and the DKH-def2-TZVP basis was used for other atoms. 77 All multireference calculations were carried out using the Orca, version 4.0.1, software. 78 The effective magnetic moments were calculated from magnetic susceptibility using standard expressions. The magnetic susceptibility was calculated from the magnetic field derivatives of the electronic partition function as implemented in Orca.

RESULTS AND DISCUSSION
Synthesis and Spectroscopy. Compounds 1, 2, and 4 -7 were synthesized in 9-39% yield by salt metathesis or ligand exchange. For 1 and 2, the initial pale green color of the reaction mixtures slowly deepened to a forest green upon warming to room temperature. Stirring was continued for ca. 2 -3 days to obtain better yields of the products. Crystals were grown at ca. −18 °C or −32 °C from toluene and hexane extracts of the dry reaction mixtures after filtration.  Inspection of the data in Table 1 show the spectra of the -SAr iPr 4 derivatives have a greater number of bands that are also generally more intense than those of the corresponding -SAr iPr 6 derivatives. This is consistent with their lower symmetry and the greater number of metal-ligand interactions. The data also imply that the bent geometry is retained in solution. Thus the electronic spectra allow differentiation between the bent and linear structures in solution.
Structures. The solid state structures determined by single-crystal X-ray diffraction of compounds 1 -7 are illustrated in Figures 1 -7. Structural data are provided in Table 2 with average C-C distances within the interacting and non-interacting flanking rings being given in Table 3. The structure of 1 is illustrated in Figure 1. The data show that the chromium and sulfur atoms are disordered over two positions with 50 % metal occupancy at the sites Cr(1) and Cr(1a). The disordered sulfur afford different Cr-S distances, with the longer Cr-S distance being associated with the terphenyl ligand exhibiting a close Cr-C interaction. The first coordination set may be regarded as containing S(1)-Cr (1)   a Only one set of Cr-S distances from one of the two crystallographically independent molecules is given.  (1) 100.64 (6) 124.19 (6) 109.86 (7)     Compound 7 (Figure 7) is centrosymmetric with one half of the molecule being symmetry generated. The S1-Zn1-S1a angle is strictly linear at 180°. The Zn-S distance is 2.1596(6) Å.
The long Zn-centroid distance, 3.1850(2) Å, and the C-C bond lengths of the aromatic ring indicate little or no Zn interactions with the flanking ring. The only other known monomeric two-23 coordinate zinc(II) thiolate is Zn(SAr iPr 6 )2, which has linear coordination at the zinc atom and a slightly longer Zn-S distance of 2.182(1) Å. 35 Figure 7. The X-ray crystal structure of Zn(SAr iPr 4 )2 (7). H atoms are not shown and thermal ellipsoids are drawn at 50% probability.
In summary, the structures of compounds 1, 2, and 4 -6 show that they display highly bent S-  Computations. To further investigate the reasons for the bending of the S-Cr-S unit in 1 and 2, dispersion corrected PBE0/def2-TZVPP DFT calculations were performed first for the simplified model system 9 in which the flanking rings are phenyl groups. A stable minimum geometry was achieved by breaking the symmetry and optimizing the structure without symmetry constraints.
This led to bending of the S-Cr-S angle from strictly linear geometry to 157.1°. However, despite such distortion, the energy difference between the bent minimum and linear transition state was found to be only 5 kJ mol −1 , thereby demonstrating that the energy required for bending of the S-Cr-S angle is very small. Similar results were obtained whether the calculations used the empirical dispersion correction or not. The dispersion correction was also found to have little effect on the optimized bond angles and bond lengths (see SI for further details), which is entirely expected as 9 contains the parent terphenyl ligand.
To quantify the effect of steric bulk of the terphenyl substituents on the calculated energies, the structures of 1 and 2 were optimized in both bent and linear geometries. The results show that the bent structure is always energetically preferred with the linear form residing 10 and 17 kJ mol −1 higher for 1 and 2. The empirical dispersion correction has only a minor effect on the calculated energies, and reduced the energy difference between the bent and linear forms to 6 and 10 kJ mol −1 for 1 and 2 respectively. Although the computational data indicate that the bent structure is intrinsically favored over the linear, the energy differences are small and prevent any definitive statement to be made, especially because no frequency data are available. The calculations show that dispersion interactions do not play a decisive role in determining the structures of 1 and 2.
determined metrical parameters, reveals only slight discrepancies from the experimental data.
Though the calculated Cr(1)-S(1) bond lengths (2.302 (1) and 2.307 Å (2)) are in good agreement with the experimental values, larger differences are observed in the S(1)-Cr(1)-S (2) angles that are slightly narrower in the optimized structures (104.7° for both (1) and (2)) than that observed crystallographically (109.802(2) o (1) and 108.832(1)° (2)). Furthermore, the calculated Cr(1)-S(1)-C(1) angle in 1, 103.2°, differs considerably from the experimental 121.126(2)°, which may be due to the structural disorder; the same bond angle is predicted within 1° of its experimental value for 2. We also note that the theoretically predicted orientation of the ligands in both 1 and 2, and especially that of their flanking rings, is comparable to the crystal structures.
For example, the calculated interactions between the chromium and the ipso-carbon are 2.337 (1) and 2.353 Å (2), and are very close to the distances observed in the X-ray structure ( Table 2).
The most probable reason for the observation of linear coordination for Cr(SAr iPr 6 )2 is the steric hindrance caused by the bulky Ar iPr 6 substituents that carry an isopropyl group at the para position in addition to those in the ortho positions of the flanking rings of the terphenyl group.
Some of the steric differences between −SAr iPr 4 and −SAr iPr 6 are apparent from a representation of 1 with −SAr iPr 4 ligands (Figure 8), which shows that the para position is congested in the bent structure and is unlikely to easily accommodate a para-isopropyl group. To test this hypothesis, the structure of Cr(SAr iPr 6 )2 was optimized in both bent and linear geometries with dispersion corrected DFT. In good agreement with the prediction, the linear structure was found to be lower in energy by 14 kJ mol −1 . Further computational studies for a combination of model compounds with different metals and aryloxide and arylamido ligands are currently underway to understand the factors determining their geometries in the solid state. To clarify the contribution from different electron configurations to the observed effective magnetic moments, further calculations were carried out to compare the spin-state energetics.
The calculations were performed for SAr iPr 6 complexes 1 and 3 -6, which have been characterized for all of the studied metal ions. As discussed earlier, the observed effective magnetic moment most likely arises from numerous different conformations present in the solution. The number of such structures can be extremely large and it is thus not feasible to try and optimize all of them. Instead, the magnetic properties were studied by using the crystal structure geometries and simplified model structures. In the geometries extracted from the crystal structures, the isopropyl groups were replaced by hydrogen to reduce the computational cost.
Although these groups play an important role in the energetics of the different geometries, they do not carry any spin density, and therefore their contribution to the relative energies of different spin states and electron configurations evaluated at a fixed geometry is likely to be small.
In essence, the calculations for 1 and 3 afford µeff values for the D2h, C2h, bent, and full geometry models that are in good agreement with the measured µeff values and spin-only values (see SI for further details). For the iron species 4, the calculated µeff of 5.2 µB for the full geometry produces the closest agreement with the measured value near 4.0 µB. However, the D2h, C2h, and bent geometries all produce significantly higher µeff values near 6.0 µB. For the cobalt species 5, the full geometry calculated value of 2.5 µB is less than the experimental value 3.8 µB, whereas in the nickel species 6 the calculated value for the full geometry, near 3.0 µB, exceeds the experimental value of 1.6 µB. Although the results for the late d-block derivatives 4, 5, and 6 appear to support the view that bending the geometry lowers µeff and the secondary M-aryl interactions lower it further, 8 the effective magnetic moment is lower than the spin-only value only in the case of the cobalt species 5, which also displays the strongest M-aryl interactions.

CONCLUSIONS
In summary, we have structurally and spectroscopically characterized six new transition metal(II) bis(thiolato) derivatives 1 -6, their zinc congener 7, and the synthetically useful sodium arylthiolate transfer agent 8. In sharp contrast to the derivatives of the related -SAr iPr 6 ligand, the divalent metal -SAr iPr 4 , except the d 5 (Mn 2+ ) and d 10 (Zn 2+ ) derivatives 3 and 7, feature strongly bent coordination and close metal-flanking ring interactions of varying strength; the cobalt (II) species has the shortest such interaction with a Co-centroid distance of 1.62058(9) Å. Effective magnetic moments indicate mostly quenched orbital angular momentum in the compounds, which is consistent with their bent structures and a covalent interaction between the metal ion and flanking aryl ring. Although the DFT calculations indicate that the bent structure is intrinsically favored over a linear one because of metal-ligand interactions, it is the absence of the seemingly remote para-isopropyl substituents and the ligand flexibility permitted by the larger sulfur ligating atom (cf. -NHAr iPr 4 4,20 and -OAr iPr 4 9,10 ligands) that allows the bending to occur.

Supporting Information
Selected 1 H NMR, IR, and UV-vis spectra; selected interatomic distances and angles for the second molecule of 2 in its unit cell; X-ray crystallographic data collection parameters and selected bond distances and angles; xyz-coordinates of optimized geometries; and CIFs for compounds 1 -8 (CCDC numbers 1555896 155898-1555903 and 1828738). This material is available free of charge via the Internet at http://pubs.acs.org.

Table of Contents Entry
On the basis of work with earlier linear two-coordinate meal complexes of the amido (-NAr iPr 4 and -NAr iPr 6 ; Ar iPr 4 = C6H3-2,6(C6H3-2,6-iPr2); Ar iPr 6 = C6H3-2,6(C6H3-2,4,6-iPr3) or aryloxo (-OAr iPr 4 and -OAr iPr 6 ) ligands, it was anticipated that the thiolato ligand -SAr iPr 4 would also induce linear or near-linear coordination, similar to that in M(SAr iPr 6 )2 (M = Cr, Mn, Fe, Co, and Ni) complexes. However, it was found that the M(SAr iPr 4 )2 (M = Cr, Fe, Co, and Ni) species have highly bent geometries with short metal-ligand interactions, owing to the absence of paraisopropyl groups on the flanking aryl rings due to the larger size of sulfur and consequent steric flexibility of the ligand.