In order to corroborate the conclusions from structural and spectroscopic findings and to gain insight into the nature of the secondary bonding interactions in 26O and 26S, DFT calculations were performed for complexes 25S, 26O and 26S. The pure BP86 functional (which for open-shell systems usually favors the low-spin state) has been used for both the antiferromagnetically coupled 1X as well as the ferromagnetically coupled 11X states, and the hybrid B3LYP functional (which usually predicts the high-spin state) has been tested for the ferromagnetically coupled state for comparison (technical details are provided in Chapter 9.4). In accordance with the experimental findings, the BP86 results confirm that the singlet state is lower in energy (by 136, 110 and 66 kJ/mol for 25S, 26O and 26S, respectively) than the high-spin state for all three models (Table 5.2). Calculated spin densities on the ether-O and thioether-S atoms are considered for evaluating the secondary interactions in26O and 26S, in comparison to25S where no such interaction is present. The results collected in Table 5.2 show that there is no spin density on the pendent thioether groups for the25Smodel, which confirms the expectation that there is no bonding interaction between those atoms. This is also validated by the atoms-in-molecules (AIM) analysis, which cannot detect any Fe-thioether bond in 25S.
Table 5.4: Calculated eigenvalues of the field gradient tensor for the singlet states of 25S, 26O and 26S at the BP86/SVP level of theory, and calculated and experimental ∆EQ values.
(a) The three eigenvalues of the field gradient tensor given in atomic units (1 a.u. = 9.72·1021V/m2). (b)
∆EQcalculated according to ∆EQ=12eQVzz·(1 +η2/3)1/2, where the quadrupole momentQis 0.16 barn (0.16·10−28m2) for 57Fe, Vzz is the main value of the EFG,η= (Vxx − Vyy)/Vzz (with|Vxx| < |Vyy| <
|Vzz|) and 1 mm/s = 4.8075·10−8eV. (c) Data from Table 5.3.
On the other hand, for the26Smodel, significant spin density is found on the two thioether-S atoms (Figure 5.7), and non-negligible spin density is also found on the ether-O atoms of the 26O model. While the spin density on the thioether-S atoms ('0.04e) is much lower than that on the thiolate (0.10e) atoms, suggesting that the thioether bonds are weaker than the bonds to the other two groups, the density is still large enough to indicate a connection between the ferric ions and the thioether-S. This is also confirmed by the AIM analysis, which clearly detects a bond between the Fe ions and the thioether groups.
The electronic density in the middle of these bonds (at the bond critical point) amounts
5.6. DFT Calculations 61
Figure 5.7: Spin densities (0.0035 a.u. level) for the 25S (top), 26O (middle) and 26S (bottom) models, calculated at the BP86/def2-SVP level.
to 0.03e, which again is slightly lower than that of the Fe-sulfide and Fe-thiolate bonds (0.09 and 0.07e, respectively). For the 26O model, the spin density on the ether-O atoms (0.01e) is appreciably smaller than on the thioether atoms in the 26S model, but still significant. Likewise, the AIM analysis identifies a bond between the Fe ions and the O atoms, with an electronic density (0.02e) that is slightly lower than for the 26S model.
Thus, the calculations unambiguously confirm the existence of a Fe-thioether interaction
62 Chapter 5. Secondary bonding interactions
in the 26S model, albeit this is a relatively weak bond, and an even weaker bond in the 26O model. In order to rationalize the trend in the quadrupole splittings ∆EQ observed in the M¨ossbauer spectra, eigenvalues of the electric field gradient (EFG) tensor have been calculated for the singlet states of the 25S, 26O and 26S models. Quadrupole splittings
∆EQ derived from those values are compared with experimental data in Table 5.4. While the calculated values appear to be systematically too low by '0.12 mm/s, the overall agreement with experimental values is quite satisfying, and most importantly the trend for
∆EQ (25S <26O < 26S) is almost quantitatively reproduced.
5.7 Conclusions
Secondary interactions between the ferric ions and added ether or thioether moieties do occur in oxidized [2Fe–2S] clusters if the additional O or S donor atoms are suitably po-sitioned in proximity to the cluster core. In the case of [2Fe–2S] clusters with capping thiophenolate ligands this situation has to be enforced by a confined chelate arrangement since no bonding interaction is observed when the tethered ether or thioether groups are free to rotate away from the metal. Due to the secondary interaction, which is clearly more pronounced for a thioether-S compared to an ether-O, the Fe atoms approach a tri-gonal bipyramidal coordination geometry with the additional donor atom and one of the bridging sulfides in apical positions. This gives rise to significant structural distortion of the cluster core with increasing Fe· · ·Fe distances and widened Fe-(µ-S)-Fe angles, which is reflected by marked changes in the spectroscopic and magnetic properties, in particular a distinct decrease in antiferromagnetic coupling and an increase in the M¨ossbauer quad-rupole splitting. Considerable spin density is found on the fifth donor atom, and reduction is facilitated for the system with additional thioether-Fe bonds. Taken together, these findings show that secondary bonding interactions can modulate the electronic properties of biological [2Fe–2S] clusters, which may well play a role for, e.g., the unique [2Fe–2S]
cluster in biotin synthase with its unusual (and potentially chelating) arginine ligand.
Chapter 6
Switching the Spin State in {S 4 X 2 }-Coordinated Iron(III) Complexes by Variation of X = N, O, P, S
Abstract
A series of {S4X2}-coordinate iron complexes (NEt4)[(1,10-X-(o-C6H4S)2)2Fe] (X = NMe, O, PPh, S) was prepared and comprehensively characterized. A correlation between the experimental spin state and the tethered neutral donor atom (S= 5/2 for X = NMe, O and S= 1/2 for X = PPh, S) is evident from magnetic susceptibility measurements. In contrast to the low spin complexes, incomplete spin relaxation is observed for both high spin complexes, as indicated by broadened M¨ossbauer absorptions at 80 K (magnetic subspectra detected at 7 K). DFT calculations agree well with the experimental findings.
63
64 Chapter 6. Spin States of {S4X2}-Coordinated Iron Complexes
6.1 Introduction
Compounds (NEt4)[(1,10-X-(o-C6H4S)2)2Fe]27(X = NMe, O, PPh, S) were initially obser-ved as byproducts in the synthesis of the [2Fe–2S] clusters26(see Chapter 5),[143]where the potentially tridentate bis-(benzenethiols) ligands 1,10-X-(o-C6H4SH)2 (X = NMe,[210] O,[203]
PPh,[211] S[204]) XIV were applied as capping terminal chelates. Separation of these mo-nomeric compounds 27 from the [2Fe–2S] cluster species could not be performed by a standard protocol since solubilities of 27 are strongly dependent on the tethered donor-functionality: Complexes 27N (X = NMe) and 27O (X = O) are readily soluble in MeCN forming intensive green-blue solutions, whereas 27P (X = PPh) and 27S (X = S) are only sparingly soluble in MeCN, but dissolve readily in DMF forming intensive purple-red so-lutions. As convincing explanations for this observation were missing a priori, a more detailed study of these complexes and their properties was conducted.