Authors: Ayumi Takaoka, Neal P. Mankad, and Jonas C. Peters
Journal: Journal of the American Chemical Society
Affiliation: Division of Chemistry and Chemical Engineering, California Institute of Technology
One great mystery in bioinorganic chemistry is how the FeMo nitrogenase cofactor, called FeMoco, is able to convert nitrogen gas into ammonia (the structure of FeMoco is on the left side of the above figure). There is a debate of where N2 binds to the FeMoco cluster. While some believe dinitrogen binds the molybdenum atom, recent studies suggest dinitrogen binding at an iron (see figure above). This iron center is rich with coordinating sulfur atoms. Previously, synthetic inorganic chemists have only found one N2 adduct of a iron complex containing a sulfur atom donor. Here, the authors report the use of thioether substituted tetradentate tri(phosphino)silyl ligands whose iron complexes can bind a nitrogen molecule. These complexes model the S-Fe-N2 feature thought to occur with FeMoco.
Using the hybrid ligands 5 and 6, reaction with iron(II) chloride and the grignard reagent of chloromethane resulted in the methyl complexes 8 and 9. Subsequent protonation using the acid HBArF4, released methane which was replaced by dinitrogen (compound 10) or by a molecule of diethyl ether which was used as the solvent (compound 11). The reason 11 formed rather than a dinitrogen adduct likely lies in the slight reduction of electron density in 6 rather than 5. This is because 6 has one less phosphine donor (replaced by a thioether donor). The authors considered if adding a hydride donor would increase the electron density at the iron enough to enable a nitrogen bound complex using 6. Indeed, reacting 10 and 11 with NaEt3BH afforded hydride-nitrogen complexes 12 and 13 respectively.
Next, the authors explored the reduction of the iron(II) complexes 10 and 11. Using the strong reductants potassium graphite and sodium amalgam on 10 produced a mixture of products including a thiolate bridged diiron complex (see scheme above) where S-C bonds in the ligand have been cleaved. Reaction of 10 with the reductant cobaltocene (CoCp2) displaced the thioether donor and added a Cp (cyclopentadienide) ligand. In both cases, N2 was no longer bound. These results show some of the challenges associated with using thiolates and thioethers in trying to create stable N2 complexes.
Reducing the solvent adduct, 11, with Cr(C6H6)2 led to a mixed-valent Fe(II)/Fe(I) complex with a dinitrogen bridge (see scheme above). Although no X-ray crystal structure was yet reported, spectroscopic evidence strongly supported their proposed structure. A relatively weak IR N2 stretching frequency (1881 cm-1), with comparison to previously reported adducts, suggested the influence of two metals and thus supports the N2 bridged structure. The observation of an IR-active N2 stretch shows that the molecule does not have an inversion center at the center of the N2. Thus, the relative orientation of the ligands must be asymmetric. A peak at 1360 nm in the near-IR spectrum was assigned as an intervalence charge-transfer band which suggests a mixed-valent species. Solution and solid-state magnetic measurement determined a spin state of S = 3/2 from ferromagnetic coupling between a S = 1 Fe(II) and a S = 1/2 Fe(I).
In summary, this communication reported the development of a new class of N2 complexes of iron with sulfur donor ligands. These complexes are models for the possibility of dinitrogen binding at an iron center in FeMoco.