A μ‐Phosphido Diiron Dumbbell in Multiple Oxidation States

Abstract The reaction of the ferrous complex [LFe(NCMe)2](OTf)2 (1), which contains a macrocyclic tetracarbene as ligand (L), with Na(OCP) generates the OCP−‐ligated complex [LFe(PCO)(CO)]OTf (2) together with the dinuclear μ‐phosphido complex [(LFe)2P](OTf)3 (3), which features an unprecedented linear Fe‐(μ‐P)‐Fe motif and a “naked” P‐atom bridge that appears at δ=+1480 ppm in the 31P NMR spectrum. 3 exhibits rich redox chemistry, and both the singly and doubly oxidized species 4 and 5 could be isolated and fully characterized. X‐ray crystallography, spectroscopic studies, in combination with DFT computations provide a comprehensive electronic structure description and show that the Fe‐(μ‐P)‐Fe core is highly covalent and structurally invariant over the series of oxidation states that are formally described as ranging from FeIIIFeIII to FeIVFeIV. 3–5 now add a higher homologue set of complexes to the many systems with Fe‐(μ‐O)‐Fe and Fe‐(μ‐N)‐Fe core structures that are prominent in bioinorganic chemistry and catalysis.


General Considerations and Equipment Used
All air-sensitive experiments were carried out under a dry nitrogen atmosphere using either a MBraun LabMaster glovebox or standard Schlenk techniques. Solvents were dried and degassed according to standard procedures before use. Deuterated solvents were dried over activated molecular sieves (4 Å for DMF-d7, 3 Å for all other solvents) and degassed. [LFe(NCMe)2](OTf)2 (1) [1] and NaOCP [2] were synthesized according to literature procedures. All other reagents were purchased from commercial sources and used without further purification unless mentioned otherwise.
NMR spectra were recorded on a Bruker Avance III HD 500, a Bruker Avance III HD 400 and a Bruker Avance III 300. If not stated otherwise, all spectra were measured at 298 K. 1 H NMR and 13 C{ 1 H} NMR chemical shifts are reported in parts per million in relation to tetramethylsilane. 1 H NMR chemical shifts were referenced to the residual hydrogen atom signals of the used deuterated solvents and 13 C NMR chemical shifts were calibrated with the solvent's natural abundant 13 C resonances. [3] UV-vis spectra were recorded on a Varian Cary 5000 instrument and a Varian Cary 8454.
IR spectra were measured with a Jasco FT/IR-4100 PikeGladi ATR spectrometer or with an Agilent Technologies Cary 630 FTIR spectrometer with Dial Path Technology and analyzed with FTIR MicroLab software.
Elemental analyses were carried out using an Elementar Vario EL III instrument by the analytical laboratory of the Institute of Inorganic Chemistry at the Georg-August-University Göttingen.
ESI mass spectrometry was performed on a Bruker HCT Ultra connected to an argon glovebox, or on a Bruker maXis ESI-QTOF.
Cyclic voltammograms were recorded inside a N2-filled glovebox with a Gamry Reference 600 potentiostat using a gas-tight CV cell with glassy carbon as working electrode, a platinum wire as counter electrode and a silver wire as pseudo reference electrode. Cyclic voltammograms were referenced to the internal standard Fc/Fc + (Fc = ferrocene) that was added at the end of the experiment. All electrochemical measurements were performed in dry and degassed MeCN/0.1 M [N n Bu4]PF6 solutions at room temperature if not stated otherwise.
X-band EPR spectra were measured on a Bruker E500 ELEXSYS spectrometer equipped with a standard cavity (ER4102ST, 9.45 GHz) and simulated using the program EasySpin. [4] Mößbauer spectra were recorded with a 57 Co source in a Rh matrix using an alternating constant acceleration Wissel Mößbauer spectrometer operated in the transmission mode and equipped with a Janis closed-cycle helium cryostat. Isomer shifts are given relative to iron metal at ambient temperature. Simulation of the experimental data was performed with the Mfit program using Lorentzian line doublets. [5] Magnetic susceptibility measurements were carried out with a Quantum-Design MPMS-XL-5 SQUID magnetometer equipped with a 5 Tesla magnet in the range from 295 or 250 to 2.0 K at a magnetic field of 0.5 T. The powdered samples were contained in a Teflon bucket and fixed in a non-magnetic sample holder. Each raw data file for the measured magnetic moment was corrected for the diamagnetic contribution of the Teflon bucket according to M dia (bucket) = χg•m•H, with an experimentally obtained gram susceptibility of the Teflon bucket. The molar susceptibility data were corrected for the diamagnetic contribution. Experimental data for were modelled by using a fitting procedure to the appropriate Heisenberg-Dirac-van-Vleck (HDvV) spin Hamiltonian for isotropic exchange coupling and Zeeman splitting: Temperature-independent paramagnetism (TIP) and a Curie-behaved paramagnetic impurity (PI) with spin S = 5/2 were included according to χcalc = (1  PI)·χ + PI·χmono + TIP. Experimental data were modelled with the julX program. [6]  was suspended in THF and kept at -35 o C. A precooled THF solution of Na(OCP)(dioxane)2.5 (8.5 mg, 30.6 µmol, 0.6 equivalent) was added dropwise. After few seconds a green precipitate started to appear, and the reaction mixture was kept at -35 o C for 12 hours and then stirred for another 12 hours at room temperature. The green precipitate was separated by filtration and the THF solution was layered with Et2O (12 mL) to give yellow crystals of the product 2 (15 mg, 23.3 µmol, 45.9% with respect to 1). 2 is soluble in CH3CN and stable at room temperature.

NMR Spectra for 3
Numbering Scheme for 3 (for NMR assignment); the ligands L are symmetry related (D2d) Three isomers are possible if one considers that the macrocyclic ligands adopt a saddle shaped conformation. Hence, there are two isomers possible in which the methylene bridges of one ligand are eclipsed with the ethylene bridge of the other ligand (effective D2d symmetry at 298 K in solution): in one isomer the methylene bridges point inside towards the other ligand, and the ethylene bridges point outwards (like in the crystal structure), while in another isomer (in the manuscript termed 3 rd isomer due to its low population of about 2%) it is vice versa, equivalent to both macrocycles being flipped. The third isomer (in the manuscript termed 2 nd isomer, about 25% population) has the methylene groups eclipsed, equivalent to a 90° rotation and simultaneous flip of just one macrocycle with respect to the 1 st isomer. Here, the two ligands become inequivalent (effective C2v symmetry in solution) as the methylene groups of the unrotated ligand continue to point inwards, while those of the flipped ligand points outwards.
The successive macrocyclic ring flips can be observed in EXSY spectra of 3 at 298 K in both MeCN-d3 and DMF-d7 in the form of exchange peaks between axial and equatorial protons of different isomers, with rate constants between 0.1 and 0.5 s -1 . Hence, coalescence for this isomerization would be expected only above 100°C. Additional peaks observed at 238 K originate from a (much faster) second dynamic process, namely torsional motion within the ethylene bridges: This process slows down to about 1 s -1 at 238 K (exchange peaks between the two CH2 units of the ethylene bridges) and reduces the symmetry from D2d to D2. The additional peaks exchange with the main peaks at the same rate and are assigned to a (overall C2 symmetric) species with opposite N-C-C-N torsion angles of the two macrocycles, while the main D2 symmetric species has four similar angles (as observed in the crystal structure). Variable temperature 1 H and 31 P NMR spectra in different solvents (MeCN-d3 and DMF-d7) are shown in Figures S11 -S14. 1 H and 31 P EXSY spectra showing the expected cross peaks indicating exchange between all three isomers in both MeCN-d3 and DMF-d7 are shown in Figures S16 -S19.   Figure S12: Variable temperature 31 P NMR spectra of complex 3 in the range from 238 K to 328 K in CD3CN Figure S13: Variable temperature 1 H NMR spectra of complex 3 in the range from 223 K to 323 K in DMF-d7. S14 Figure S14: Variable temperature 31 P NMR spectra of complex 3 in the range from 238 K to 308 K in DMF-d7.

X-ray Crystallography
Crystal data and details of the data collections are given in Table S1, molecular structures are shown in Figures S48 -S51. X-ray data were collected on a STOE IPDS II diffractometer (graphite monochromated Mo-Kα radiation, λ = 0.71073 Å) by use of w scans at -140 °C. The structures were solved with SHELXT and refined on F 2 using all reflections with SHELXL-2018. [7] Non-hydrogen atoms were refined anisotropically. Hydrogen atoms were placed in calculated positions and assigned to an isotropic displacement parameter of 1.5/1.2 Ueq(C). Crystals of 3 were found to be twinned (twin law: 1 0.021 0.003, 0 -1 0, 0 0 -1; BASF: 0.126 (2)). Acetonitrile molecules and CF3SO3anions were found to be disordered  (4)) and were refined using SAME and RIGU restraints and EADP constraints. In case of 4 only the cationic part could be identified from the crystal structure (monoclinic, C2/c, a=42.62 Å, b=15.75 Å, c=23.15 Å, β=119°). The quality of the diffraction data, however, was not sufficient for anisotropic refinement or discussion of bonding parameters. Faceindexed absorption corrections were performed numerically with the program X-RED. [8] Table S1. Crystal data and refinement details for 2, 3, and 5.

DFT Calculations Computational Setup
Geometry optimizations were performed with the B3LYP [9] density functional. The respective def2-TZVP [10] for the first coordination sphere and def2-SV(P) basis sets [11] for the remaining atoms were applied in combination with the auxiliary basis sets def2/J. [12] The RI [13] and RIJCOSX [14] approximations were used to accelerate the calculations.
The Mössbauer spectroscopic parameters were computed using the B3LYP density functional. The CP(PPP) [15] basis set for Fe and the def2-TZVP basis set for remaining atoms were used.
Isomer shifts  were calculated from the electron densities 0 at the Fe nuclei by employing the linear regression: (1) Here, C is a prefixed value, and  and  are the fit parameters. Their values for different combinations of the density functionals and basis sets can be found in our earlier work ( = -0.366,  = 2.852, C = 11810). [16] Quadrupole splittings EQ were obtained from electric field gradients Vij (i = x, y, z; Vii are the eigenvalues of the electric field gradient tensor) by using a nuclear quadrupole moment Q( 57 Fe) = 0.16 barn: [17] Here, is the asymmetry parameter.
TDDFT calculations were also performed with B3LYP in conjunction with the def2-TZVP basis sets for all atoms.
All computations in this work were carried out with the ORCA program package. [18] We took complex 3 as example to test closed-and open-shell singlet calculations. Our results showed that key metric parameters calculated by the restricted B3LYP calculations are in reasonable agreement with those found experimentally. In contrast, the unrestricted B3LYP computations indeed delivered an open-shell singlet solution, but the computed geometry of 3 having a substantially bent Fe-P-Fe core (159.3˚) is distinct from that determined by X-ray crystallography. Despite of this, the three non-bonding molecular orbitals in the spin-up manifold involving the Fe-dz2, -dxz and -dyz orbitals are only slightly different from those in the spin-down set ( Figure S53), and the difference can be readily attributed to the bent Fe-P-Fe core. Thus, the broken-symmetry solution is, in fact, qualitatively the same as that predicted by the restricted computations. On the basis of these findings, we did not employ broken-symmetry calculations further for complexes 3 and 5, and the results discussed in the main text were obtained by the restricted closed-shell calculations. More importantly, the restricted calculations also successfully reproduced the Mössbauer parameters within the uncertainty of the computations.  As shown in Figure S53, because of the bent Fe-P-Fe core, the three non-bonding orbitals (vide infra) involving the Fe-dz2, -dxz and -dyz orbitals of each iron center in the spin-up manifold are slightly different from those in the spin-down set. Thus, the B3LYP broken symmetry calculations in fact deliver a similar bonding picture as those predicted by BP86 and TPSSh.