How the Metal Ion Affects the 1H NMR Chemical Shift Values of Schiff Base Metal Complexes: Rationalization by DFT Calculations

The chemical shift (CS) values obtained by 1H NMR spectroscopy for the hydrogen atoms of a tetradentate N2O2-substituted Salphen ligand (H2L1) are differently shifted in its complexes of nickel(II), palladium(II), platinum(II), and zinc(II), all bearing the same charge on the metal ions. To rationalize the observed trends, DFT calculations have been performed in the implicit d6-DMSO solvent in terms of the electronic effects induced by the metal ion and of the nature and strength of the metal-N and metal-O bonds. Overall, the results obtained point out that, in the complexes involving group 10 elements, the CS values show the greater shift when considering the two hydrogen atoms at a shorter distance from the coordinated metal center and follow the decreasing metal charge in the order Ni > Pd > Pt. This trend suggests a more covalent character of the ligand–metal bonds with the increase of the metal atomic number. Furthermore, a slightly poorer agreement between experimental and calculated data is observed in the presence of the nickel(II) ion. Such discrepancy is explained by the formation of stacked oligomers, aimed at minimizing the repulsive interactions with the polar DMSO solvent.


Preliminary tests on ZnL1 to select the computational method
In order to select the best combination of DFT functional and basis set which lead to the highest linear correlation coefficient, we performed several tests on the ZnL1 complex.We first optimized the complex structure at the M06L level of theory with 6311G(d,p) basis set for all atoms but Zn, for which the LANL2DZ ECP was employed.We then calculate the values of the proton chemical shifts using the same protocol, obtaining an experimental vs calculated linear correlation coefficient of 0.9674.We checked if different combination of functionals and basis sets could result in an improvement of the correlation.Therefore, starting from the structure optimized with M06L functional, the 1 H-NMR CS values were calculated employing the all-electron 6-311G** basis sets with M06L, PBE0, B3LYP and ω-B97XD, whose linear correlation coefficients are reported in Table S1.Interestingly, the extension of all-electron basis sets to the metal cation leads to a worsening of the R value for both M06L and PBE0, while the obtained values at the B3LYP level of theory are almost identical.On the other hand, a better linear correlation is obtained when ω-B97XD is used, underlying the notable improving contribution of the dispersion forces included in this functional.
We tested the performance of the same functionals in combination with the Ahlrichs def2-TZVP basis set for the optimization of the structures.In this case, the ORCA software was used, to include the possibility of using RI integrals. 1H-NMR CS values were also calculated using the PCSSEG-2 basis set, which is suited for NMR calculations, as reported in ORCA manual.The corresponding linear correlation coefficients are listed in Table S2, and clearly show the better performance of the ω-B97XD functional.We eventually checked the employment of def2-TZVP to further improve the experimental/calculated linear correlation.The structure of H2L1 (Figure S10) is not planar, with the torsion angle involving the two dichlorophenol moieties rotated of ~50° with respect to the plan of the third diimine bearing ring.On the other hand, the optimized structure of all the four substituted Salphen metal complexes confirm the expected N2O2 tetra coordination in a roughly square planar geometry (Figure S11).The most stable optimized structures of H2L1, (Figure S10), reveals the formation of two hydrogen bonds between the hydrogen atoms of the hydroxyl groups and the N atoms, whose calculated distance is 1.71 Å.Table S3 reports the main geometric parameters calculated for complex NiL1 (Figure S11) and the experimental ones reported in ref. 20 for a similar Ni-MeOSalphen complex (Figure S12).
The very good agreement between the calculated and experimental values nicely supports the reliability of the selected computational protocol.The calculated O1-N1-N2-O2 dihedral angle is -0.86° while the C1-Ni-C2 is 172.0°.When Ni is replaced by Zn, the calculated O1-N1-N2-O2 and C1-Ni-C2 angles become 0.00° and 164.9°, respectively, which results in a reduced planarity of the aryl halides substituents while the Zn cation reaches a perfect squared planar coordination.Similarly to ZnL1, for both PdL1 and PtL1 the calculated O1-N1-N2-O2 angle is 0.00°, while the C1-Ni-C2 angle values are 176.6°and 176.8°, respectively.Therefore, the presence of different metal ions affects the bending of the two aryl halide groups inferring a distortion from the planarity to the overall structure.Such an effect is more pronounced in the case of ZnL1, while it is almost negligible in the presence of the heavier Pt and Pd metals.Interestingly, only for the NiL1 the calculated O1-N1-N2-O2 dihedral angle is different from zero.The M-N distances decrease from 2.06 Å in ZnL1 to 1.87, 1.97 and 1.96 Å in NiL1, PtL1 and PdL1, respectively.On the other hand, the shortest M-O distances of 1.85 Å is calculated for the NiL1, which elongates to 1.95 Å in ZnL1 and 2.00 Å in both PdL1 and PtL1.Therefore, while very small difference of 0.02, 0.03 and 0.04 Å between the M-N and M-O are found for NiL1, PdL1 and PtL1, respectively, the significant larger value of 0.11 Å obtained for ZnL1 implies a higher elongation of the Zn-O bonds with respect to the Zn-N, which might also explain the higher distortion of the corresponding structure.9.33391203097145 11.46781160938520 14.46184491871365 C 11.07838047905355 7.08862550651153 8.68023309324471 C 10.65955126143341 6.19776449371482 7.69514408022154 C 11.59614537746145 5.56796259583588 6.88279073547818 C 12.94218634474201 5.82932920423827 7.06275806341452 C 13.35949955957181 6.72332543287425 8.04161506995455 C 12.43250909283605 7.35679805789020 8.84909862862710 H 9.36594111423017 4.57606310555549 6.18937084573280 H 7.66370739788777 3.42651916049502 5.17681461394192 H 3.56201538310409 4.29833427001485 6.01569924576131 H 11.29463751726606 4.88067164856956 6.10441940629303 H 13.66937964427793 5.33956298540019 6.42727656236168 H 12.77795473968521 8.06305026080485 9.59161397033241 H 14.41333453774018 6.93480307464848 8.17261860432092 H 11.28379273734934 8.59105836879546 10.74435907271483 H 10.76415614288866 9.91454003144680 12.53816398591155 H 6.71925221883135 10.95657660553257 13.45580869471386 Ni 8.33032030095052 7.15318095842730 8.80658644357578 H 5.92323877658680 9.60039258419780 11.56137329616277 H 4.24264442829402 6.02430938903571 7.63307752480118

Figure S10 .
Figure S10.Top and side view of H2L1

Figure S13 .
Figure S13.Natural bond order (NBO) representation of the Pt-N and Pt-O bonds in PtL1.

Table S1 .
Calculated 1 H-NMR CS values (ppm) and linear correlation coefficients of ZnL1 using different computational protocols with Gaussian06.The isotropic shielding of the TMS reference calculated at each level of theory is also reported (au)

Table S2 .
Calculated 1 H-NMR CS values (ppm) and linear correlation coefficients of ZnL1 using different computational protocols with ORCA.The isotropic shielding of the TMS reference calculated at each level of theory is also reported (au).

Table S3 .
Main calculated geometric parameters of NiL1 (distances, Å and angles, °) and corresponding experimental values of the Ni-MeOSalphen complex reported in ref. 20.