Catalyst Speciation

With evidence of a molecular Fe-based catalyst for the fixation of N2 in hand,

initial studies on the speciation of the (TPB)Fe complexes in solution and potential causes of termination of catalysis were undertaken. A previous study on the (TPB)Fe scaffold showed that exposure of (TPB)Fe(N2) to H2 resulted in formation of (TPB)(μ-

H)Fe(H)(N2).18 Because H2 was formed as a by-product of the reaction, it was suspected

that (TPB)(μ-H)Fe(H)(N2) could be formed under the catalytic conditions. In this

context, 4.1 was reacted with 10 equivalents of HBArF4 · 2 Et2O and 12 equivalents of

KC8. The presence of (TPB)(μ-H)Fe(H)(N2) in a ~ 30% yield was confirmed by both 31P

NMR and IR spectroscopy, where the strong N-N stretching vibration can be observed at νNN = 2073 cm-1. These results indicate that formation of (TPB)(μ-H)Fe(H)(N2) under

the catalytic conditions is likely. To test whether (TPB)(μ-H)Fe(H)(N2) would terminate

a catalytic cycle, this complex was submitted to the standard catalytic conditions, and it was found that only 0.5 equivalents of NH3 were formed per Fe center. This result

suggests that if (TPB)(μ-H)Fe(H)(N2) is formed under catalysis, it likely is unable to re-

enter the catalytic cycle. Further evidence supporting this conclusion comes from the observation that (TPB)(μ-H)Fe(H)(N2) does not show appreciable reactivity with either

HBArF4 · 2 Et2O or KC8 at RT in Et2O.

When considering the comparably low yields of NH3 provided when (TPB)Fe(N2)

(Table 1, entry 5) is used as a pre-catalyst the observation that (TPB)(μ-H)Fe(H)(N2) can

be formed from reaction of (TPB)Fe(N2) with H2 suggests that starting with (TPB)Fe(N2)

may allow for a greater percentage of the (TPB)Fe centers to be trapped as the catalytically inactive (TPB)(μ-H)Fe(H)(N2), thus decreasing catalysis. Alternately, a

low-spin state such as that observed in (TPB)Fe(CO)13 could dominate in (TPB)Fe(N2) at

low temperature, significantly decreasing its catalytic activity. Avoidance of this oxidation state by adding KC8 initially to reduce all of the (TPB)Fe(N2) to

[(TPB)Fe(N2)]- matches with both of these hypotheses.

Scheme 4.5. Formation of (TPB)(μ-H)Fe(H)(N2) and its reactivity with acid and

reductant indicating that it is not a competent pre-catalyst.

Another possible termination pathway arises from the presence of Cl- _{in the}

reaction mixture. Since HBArF4 · 2 Et2O is made via loss of NaCl, Cl- contamination is a

reasonable possibility. The most conclusive test for the effect of Cl- is use of the S = 3/2 complex (TPB)FeCl, which was synthesized and crystallized in an analogous manner to the previously reported (TPB)FeBr.13a Catalytic runs using this pre-catalyst give, on average, 3.2 equivalents of NH3 per Fe center, suggesting that while Cl- does not

terminate catalysis, it does hinder the reactivity of the (TPB)Fe species relative to using 4.1 as a pre-catalyst. Similarly, addition of exogenous Cl- in the form of excess KCl to the standard catalytic conditions with 4.1 also results in 3.0 equivalents of NH3 being

formed. These result suggests that any residual Cl- may lower the yield of NH3, but

should not terminate the catalytic cycle by formation of (TPB)FeCl.

In addition to termination processes, it was of great interest to determine whether N2H4 was involved in the catalytic cycle, as any presence of N2H4 would imply the

involvement of an “alternating” type mechanism. A UV-Visible spectroscopic assay similar to the indophenol method for NH3 quantification was employed to detect the

presence of N2H4 in reaction mixtures.19 No N2H4 was detected from any catalytic

mixture. As an additional test, two equivalents of N2H4 were added to 4.1 before

performing the standard catalytic protocol. In these runs, only trace N2H4 was observed,

suggesting that if N2H4 is formed during catalysis, it would likely be consumed and

should not be observed.

4.3 Conclusion

The anionic N2 complex 4.1 generates NH3 in the form of complex 4.4 upon

exposure to excess acid, likely via some disproportionation pathway. Use of HBArF4 · 2

Et2O and KC8 in Et2O enables the catalytic reduction of N2 to NH3 with 7 equivalents of

NH3 being generated per Fe center. Multiple other (TPB)Fe species also serve as pre-

catalysts, but simple Fe salts and complexes do not, suggesting the agency of a molecular (TPB)Fe catalyst. Analysis of reaction mixtures indicates the formation of (TPB)(μ- H)Fe(H)(N2) as a likely byproduct of catalysis. This complex is also not competent as a

pre-catalyst and hence its formation indicates a possible pathway for the termination of catalysis. The presence of chloride appears to inhibit, but not terminate, catalysis and (TPB)FeCl still serves as a competent pre-catalyst. Finally, no N2H4 is detected in

catalytic reaction mixtures, but N2H4 is also consumed under catalysis, as determined by

the addition of two equivalents of N2H4 to catalytic mixtures. This result suggests that if

N2H4 were formed during catalysis it would likely not be detectable.

Due to the dearth of synthetic Fe species that can form appreciable quantities of NH3 from N2, Mo complexes have received a great deal of attention as the likely binding

site of N2 in the FeMoco active site of nitrogenase.20 The results disclosed herein show

for the first time that a molecular Fe species can also catalyze nitrogen fixation with protons and electrons in comparable yields to the Mo systems. This work validates the hypothesis that N2 binding and reduction may occur at a single Fe site in the FeMoco.21

Furthermore, the flexible nature of the Fe-B interaction in the (TPB)Fe scaffold further motivates the consideration of a similar hemi-labile role for the central C atom of the FeMoco.22