Structural and Electronic Studies of Substituted m-Terphenyl Group 12 Complexes

The effects of para-substitution on the structural and electronic properties of four series of two-coordinate m-terphenyl Group 12 complexes (R-Ar#)2M (M = Zn, Cd, Hg; R = t-Bu 1–3, SiMe34–6, Cl 7–9, CF310–12, where R-Ar# = 2,6-{2,6-Xyl}2-4-R-C6H2 and 2,6-Xyl = 2,6-Me2C6H3) have been investigated. X-ray crystallography shows little structural variation across the series, with no significant change in the C–M–C bond distances and angles. However, considerable electronic differences are revealed by heteronuclear nuclear magnetic resonance (NMR) spectroscopy; a linear correlation is observed between the 113Cd, 199Hg, and 1H (2,6-Xyl methyl protons) NMR chemical shifts of the para-substituted complexes and the Hammett constants for the R-substituents. Specifically, an upfield shift in the NMR signal is observed with increasingly electron-withdrawing R-substituents. Density functional theory (DFT) calculations are employed to attempt to rationalize these trends.

2.2. Solid-State Characterization. The crystal structures of 1−12 confirm that all complexes are monomeric in the solid state, owing to the steric demands of the m-terphenyl ligands, with no intermolecular interactions between the metal centers. In all cases, the complexes are two-coordinate and quasi-linear, featuring a single metal center coordinated by two σ-bonded m-terphenyl ligands. Unlike the 3,5-Xyl complexes [2,6-{3,5-Xyl} 2 C 6 H 3 ] 2 M (M = Zn, Cd, Hg), no M···H contacts are formed to the flanking aryl rings. 29 The crystal structure of 1 is presented in Figure 1, with key measurements about the metal center for 1−12 provided in Table 1 . It should be noted that the crystal data for 4 are of low quality due to weak diffraction from a small crystal. Despite repeated attempts, it was not possible to grow highquality crystals of 4. However, the data are sufficient to demonstrate the connectivity of the molecule and are included here for completeness.
For each Group 12 metal, the corresponding series of parasubstituted complexes show no significant change in the M−C bond distances as the functional group is varied. The Zn−C bond distances for 1, 4, 7, and 10 fall within a narrow range    9,10 In summary, the crystal structures of 1−12 show little structural variation as the para-substituent of the m-terphenyl ligand is varied. This suggests that the geometries of these complexes are dominated by steric and crystal packing effects, rather than the electronic structure of the ligand.
2.3. Solution-State Characterization. The electronic structures of 1−12 were studied by 1 H, 13 C{ 1 H}, 113 Cd, and 199 Hg NMR spectroscopies in d 6 -benzene and compared to those of the unsubstituted analogues (H-Ar # ) 2 M (M = Zn, Cd, Hg). 29 Here, a numbering scheme has been assigned to the mterphenyl unit, as shown in Figure 2. The electronic strengths of different para-substituents are quantified using Hammett constants, σ para . 50 A comparison of the 1 H NMR spectra for complexes 1−12 reveals three noteworthy features (Table 2). First, the meta-protons (H-3) on the central aryl rings exhibit notable peak shifts as the para-substituent is changed, although no overall trend is evident. There is, however, a clear downfield shift in H-3 when varying the metal from Zn (6.76−7.14 ppm) to Cd (6.87−7.22 ppm) to Hg (6.92−7.30 ppm). Second, the 2,6-Xyl aryl protons (H-7 and H-8) for 1−12 remain relatively unshifted by changing the para-substituent or the metal, suggesting there is minimal electronic communication with the flanking aryl rings. Third, the 2,6-Xyl methyl protons (H-9) shift upfield with increased electron-withdrawing strength of the para-substituent. A plot of the chemical shifts, δ, against the Hammett constants, σ para , reveals a linear correlation (Figures 3 and S1). 50 A similar trend was observed in recent studies of the analogous lithium complexes [R-Ar # -Li] 2 (R = t-Bu, SiMe 3 , H, Cl, CF 3 ). 38 We note that the chemical shifts for H-9 are largely unaffected by the identity of the metal ( Table  2).
The 13 (Table 2). This can again be attributed to poor electronic communication between the central and flanking aryl rings. However, the 13 C{ 1 H} NMR signals for the central aryl ring shift considerably with the notable exception of C-2 ( Table 2). We note that the largest shifts are for the ipso-carbon atoms (C-1) where, in addition to a downfield shift in δ C with increasing σ para of the substituent, large downfield shifts of ca. 10 ppm are observed as the metal varies from Zn (148.5−156.8 ppm) to Cd (158.3−167.0 ppm) to Hg (169.1−176.0 ppm). For similar complexes in the literature, this downfield trend has been ascribed to the increasing Pauling electronegativity as Group 12 is descended (1.65, 1.69, and 2.00 for Zn, Cd, and Hg, respectively). 10,16,29,46,51−53 The 113 Cd and 199 Hg NMR spectra of 2, 5, 8, 11 and 3, 6, 9, 12 were also recorded. Multiple NMR measurements revealed no change in chemical shift with varying analyte concentration, most likely due to the steric bulk of the ligands preventing interaction of the metal with the surrounding solvent. 54−56 In all cases, the 113 Cd and 199 Hg NMR spectra show a single peak indicating one metal environment in solution, in the same region as other literature metal diaryl complexes (see Table  3). 37  For 3, C(1) = C(23) due to symmetry (Z′ = 0.5). b Crystal data for 4 are of low quality due to weak diffraction from a very small crystal. Data are included here for completeness. c Measurements for the second molecule in asymmetric unit given in square brackets. Organometallics pubs.acs.org/Organometallics Article increasing the steric bulk of the flanking groups was found to cause an upfield shift in their 113 Cd and 199 Hg NMR spectra. 29 However, since complexes 1−12 all feature the same flanking groups (2,6-Xyl) and are crystallographically similar, we suggest that steric effects are unlikely to have a major influence on their 113 Cd and 199 Hg NMR shifts. A plot of the 113 Cd and 199 Hg NMR chemical shifts (δ) for each of the para-substituted complexes, vs their corresponding Hammett constant (σ para ) is shown in Figure 3. 50 Linear correlations can be fitted to the 113 Cd (blue line; R 2 = 0.96) and 199 Hg (red line; R 2 = 0.95) NMR data, both with a negative gradient, indicating that more electron-withdrawing substituents shift the NMR peak of the Cd and Hg centers further upfield. This trend is somewhat counterintuitive, as electron-withdrawing groups might be expected to deshield the nuclei and cause a downfield shift. However, similar findings were reported for a series of para-substituted mercury diaryls (4-R-C 6 H 4 ) 2 Hg (R = OMe, Me, H, F, Cl, CF 3 ), 60−62 suggesting that these chemical shifts depend on more than simple σ donor effects. One hypothesis suggests that the bonding in organomercury compounds mainly involves the valence 6s orbital 63,64 since the 6p orbital is too high in energy to overlap. However, by incorporating electron-donating groups onto the ligand, the ligand orbitals increase in energy and overlap better with the 6p orbitals. 58,65 This populates the more diffuse 6p orbitals and depopulates the less diffuse 6s. Hence, the electron density around the metal center moves away from the nucleus and becomes more diffuse, resulting in less shielding and a downfield NMR shift. 58 Cyclic voltammetry studies were also carried out on the mercury complexes 3 and 12 (R = t-Bu and CF 3 ) in THF solution (Supporting Information, Section S4). However, no redox events were observed upon scanning from −0.5 to −2.5 V (vs Fc + /Fc) in either case (Figure S44), suggesting a large HOMO−LUMO gap for these complexes.
2.4. Computational Analysis. Density functional theory (DFT) calculations were employed to attempt to rationalize the trends in the NMR spectroscopic parameters. Full geometry optimizations (BP86/TZVP, see Supporting In- The flanking aryl atoms remain unshifted and thus have been omitted. b Literature NMR data for the unsubstituted complexes (H-Ar # ) 2 M (M = Zn, Cd, Hg), original data re-referenced to C 6 D 6 . 29 Figure 3. Plot of the 1 H (for flanking methyl protons, H-9), 113 Cd, and 199 Hg NMR chemical shifts, δ, for the metal diaryls (R-Ar # ) 2 M (1−12, plus R = H) 29 vs their Hammett constants, σ para . 50 For clarity, the 1 H NMR (H-9) trend is given only for the Zn series; plots for the Cd and Hg series are provided in Supporting Information Figure S1. Organometallics pubs.acs.org/Organometallics Article formation Section S5.1 for full details) were performed on 1− 12, as well as the unsubstituted analogues. All optimized structures displayed near-linear bond angles in a very narrow range (Table S5) 38 However, for the Group 12 complexes, bond paths corresponding to C−H···C arene interactions were observed between the H-9 protons and aromatic carbons of the flanking aryl rings situated opposite to them ( Figure S48). Properties of the electron density at the bond critical points for these interactions are provided in Supporting Information  Table S6.
Subsequently, the 1 H, 113 Cd, and 199 Hg NMR chemical shift parameters for 1−12 and the unsubstituted analogues were calculated using the ReSpect program. 66−71 These calculations were carried out on both the fully optimized structures used above, as well as the structures taken directly from the crystallographic data in which only the H atom positions had been optimized (see Supporting Information Section S5.1 for details). NMR shielding constants were calculated using the KT2 density functional approximation, 72 which was specifically designed for the calculation of NMR shielding constants. The calculations were carried out at two levels of theory: dyallvdz 73,74 basis set for Zn/Cd/Hg and pcS-1 75 for all other atoms (vdz/pcS-1) or dyall-vtz 73 for Zn/Cd and pcS-2 75 for all other atoms (vtz/pcS-2). Calculations for the mercury complexes at the vtz/pcS-2 level could not be completed due to technical limitations of the ReSpect program. 66−71 A summary of the calculated 1 H, 113 Cd, and 119 Hg NMR chemical shifts for the H-9 protons of 1−12 (in both the fully optimized and H-atom optimized geometries) are provided in Supporting Information Tables S9 and S10. Plots of the computed vs experimental shifts are shown in Supporting Information Figures S49−S56. In these, a weak positive correlation is observed between calculated and experimental shifts for the H-9 protons of all complexes (Figures S49−S53). This trend is evident in both the fully optimized and H-atom optimized structures and at both the vdz/pcS-1 and vtz/pcS-2 levels. However, the correlation is not particularly strong, and some computed results [particularly (H-Ar # ) 2 Zn] deviate significantly from the experimental values. The experimental trend in 1 H NMR shifts for the H-9 protons occurs over such a narrow chemical shift range (ca. 0.3 ppm) that the accuracy of the DFT calculations may not be sufficient to reliably reproduce this behavior. Despite the lack of C−H···M (M = Zn, Cd, Hg) close contacts, the H-9 chemical shifts feature large paramagnetic contributions to the shielding constant (Tables S7 and S8), much like the analogous lithium complexes [R-Ar # -Li] 2 (R = t-Bu, SiMe 3 , H, Cl, CF 3 ). 38 It is known that when the paramagnetic components are dominant, density functional methods often fail to achieve high accuracy, as appears to be the case here.
The computed 113 Cd and 199 Hg NMR chemical shifts (vdz/ pcS-1) show relatively poor agreement with the experimental values. While the 113 Cd NMR shifts for the H-atom optimized structures appear to roughly correlate with the experimental values ( Figure S54), this correlation is lost in the fully geometry optimized structures. No convincing correlation is observed for the 199 Hg shifts in either geometry ( Figure S56). In addition, the computed chemical shifts differ significantly (by >100 ppm) from the experimental shifts in all cases. At the vtz/pcS-2 level, the computed 113 Cd shifts follow a similar trend relative to the experimental shifts as at the vdz/pcS-1 level ( Figure S55), but the absolute values of the computed chemical shifts are closer to the experimental values.
These results suggest that the computed chemical shifts are strongly dependent on geometry, with small changes in the coordination environment of the metal resulting in dramatic changes in the computed shift. We propose that to model the NMR properties of these complexes more accurately, it may be necessary to perform dynamics calculations and account for conformational flexibility.

CONCLUSIONS
Four series of para-substituted m-terphenyl Group 12 complexes (R-Ar # ) 2 M (M = Zn, Cd, Hg; R = t-Bu 1−3, SiMe 3 4−6, Cl 7−9, CF 3 10−12) have been reported. While negligible structural differences are observed by X-ray crystallography, NMR spectroscopic studies reveal considerable electronic differences within the ligand framework and at the metal center. A linear correlation of the 113 Cd and 199 Hg NMR chemical shifts is observed with the Hammett constants of the para-groups. Moreover, the flanking methyl protons, H-9, exhibit similar shifts in their 1 H NMR spectra. In all cases, an upfield shift is observed with increasingly electron-withdrawing substituents. DFT modeling suggests that the H-9 1 H NMR chemical shifts, as well as the 113 Cd and 199 Hg chemical shifts, all feature large paramagnetic contributions to the shielding constants. As a result, the experimental trends could not be reproduced by our computational analysis.
Full experimental details for the synthesis, characterization, and crystallographic data (PDF) Coordinates (XYZ) Accession Codes CCDC 2163371−2163382 contain the supplementary crystallographic data for this paper. These data can be obtained free of charge via www.ccdc.cam.ac.uk/data_request/cif, or by emailing data_request@ccdc.cam.ac.uk, or by contacting The Organometallics pubs.acs.org/Organometallics Article