Heterobimetallic [ NiFe ] Complexes Containing Mixed CO / CN-Ligands : Analogs of the Active Site of the [ NiFe ] Hydrogenases

The development of synthetic analogs of the active sites of [NiFe] hydrogenases remains challenging, and, in spite of the number of complexes featuring a [NiFe] center, those featuring CO and CN- ligands at the Fe center are under-represented. We report herein the synthesis of three bimetallic [NiFe] complexes [Ni( N2 S2)Fe(CO)2(CN)2], [Ni( S4)Fe(CO)2(CN)2], and [Ni( N2 S3)Fe(CO)2(CN)2] that each contain a Ni center that bridges through two thiolato S donors to a {Fe(CO)2(CN)2} unit. X-ray crystallographic studies on [Ni( N2 S3)Fe(CO)2(CN)2], supported by DFT calculations, are consistent with a solid-state structure containing distinct molecules in the singlet ( S = 0) and triplet ( S = 1) states. Each cluster exhibits irreversible reduction processes between -1.45 and -1.67 V vs Fc+/Fc and [Ni( N2 S3)Fe(CO)2(CN)2] possesses a reversible oxidation process at 0.17 V vs Fc+/Fc. Spectroelectrochemical infrared (IR) and electron paramagnetic resonance (EPR) studies, supported by density functional theory (DFT) calculations, are consistent with a NiIIIFeII formulation for [Ni( N2 S3)Fe(CO)2(CN)2]+. The singly occupied molecular orbital (SOMO) in [Ni( N2 S3)Fe(CO)2(CN)2]+ is based on Ni 3dz2 and 3p S with the S contributions deriving principally from the apical S-donor. The nature of the SOMO corresponds to that proposed for the Ni-C state of the [NiFe] hydrogenases for which a NiIIIFeII formulation has also been proposed. A comparison of the experimental structures, and the electrochemical and spectroscopic properties of [Ni( N2 S3)Fe(CO)2(CN)2] and its [Ni( N2 S3)] precursor, together with calculations on the oxidized [Ni( N2 S3)Fe(CO)2(CN)2]+ and [Ni( N2 S3)]+ forms suggests that the binding of the {Fe(CO)(CN)2} unit to the {Ni(CysS)4} center at the active site of the [NiFe] hydrogenases suppresses thiolate-based oxidative chemistry involving the bridging thiolate S donors. This is in addition to the role of the Fe center in modulating the redox potential and geometry and supporting a bridging hydride species between the Ni and Fe centers in the Ni-C state.


Introduction
The [NiFe] hydrogenases catalyze the two-electron inter-conversion of two protons and molecular H2, reactions that are relevant for the development of new clean energy technologies. 1 The active site of the [NiFe] hydrogenases consists of a heterobimetallic Ni-Fe cluster in which the Ni center is bound by two terminal cysteine S-donors and two cysteine S-donors that bridge to an Fe center that is also co-ordinated by one carbonyl and two cyanato ligands. Depending on the state of the enzyme, a third bridging ligand, X (X = OH − or H − ), may also be found at the active site. In these states, the geometry about the Fe center is pseudo-octahedral and that about the Ni center is distorted square-based pyramidal, with X occupying a basal position. At least ten different states of the enzyme have been identified with formulations dependent on the oxidation state of the Ni center (i.e., Ni III , Ni II and Ni I ) and on the nature of X. 1c,1d,2 The importance of each state in the catalytic cycle continues to be debated, 1c,1d,3 but it is widely accepted that catalytic turnover is associated with changes of the formal oxidation state of the Ni center while the Fe center remains as low spin d 6 Fe II . 2 The mechanisms proposed for catalysis 1c,1d,4 involve three key states, Ni-SI, Ni-R and Ni-C (Scheme 1) that involve the formal Ni II (Ni-SI and Ni-R) and Ni III  A range of analogs of the active site of the [NiFe] hydrogenases has been synthesized and characterized in attempts to replicate the principal structural and functional features of the active sites. 3,6 Complexes that possess a mixed CN − and CO co-ordination about the Fe center, such as [Fe(fac-CO)3(CN)3] − , 7 and Ni complexes that possess a Ni III/II couple with a relatively low reduction potential, such as [Ni(ema)] 2− [ema 4− = N,N'-ethylenebis (2-(acetylthio)acetamide] 8  In these complexes the redox events are associated principally with the Fe center rather than the Ni unit. 9c There has been considerable success in reproducing key aspects of the proposed catalytic cycle for the [NiFe] hydrogenases and in the development of complexes containing a [Ni II Fe II ] center. 6e,10 However, the syntheses of relevant paramagnetic analogs have proven more challenging. [ (0.2 M); the cell consisted of a Pt/Rh gauze basket working electrode separated by a glass frit from a Pt/Rh gauze secondary electrode. The saturated calomel reference electrode was placed at the center of the working electrode and the solution stirred rapidly during electrolysis using a magnetic stirring bar.
UV/vis spectroelectrochemical experiments were carried out at 273 K or 243 K using an optically transparent electrode mounted in a modified quartz cuvette with an optical pathlength of 0.5 mm. A three-electrode configuration consisting of a Pt/Rh gauze working electrode, a Pt wire secondary electrode (in a fritted PTFE sleeve) and a saturated calomel electrode, chemically isolated from the test solution via a bridge tube containing electrolyte solution and terminated in a porous frit, was used in the cell. The potential at the working electrode was controlled by a Sycopel Scientific Ltd. DD10M potentiostat. UV/vis spectra were recorded on a Perkin Elmer Lambda 16 spectrophotometer. The spectrometer cavity was purged with N2 and temperature control at the sample was achieved by flowing cooled N2 across the surface of the cell. X-band EPR spectra were recorded on a Bruker EMX spectrometer. The simulations of the EPR spectra were performed using the Bruker WINEPR SimFonia package.
X-ray crystallography. Crystals of 1A and 1B were collected on a Bruker SMART APEX diffractometer with graphite-monochromated MoKα radiation ( = 0.71073 Å). Crystal of 2 and 3 were examined on a Rigaku Oxford Diffraction SuperNova diffractometer using mirror-monochromated CuKα radiation ( = 1.5418 Å). Intensities were integrated from data recorded on 0.3° (1A and 1B) or 1° (2 and 3) frames by ω rotation. Cell parameters were refined from the observed positions of all strong reflections in each data set. Either a multiscan absorption correction 13 (1A and 1B) or Gaussian grid face-indexed absorption correction with a beam profile correction 14 (2 and 3) was applied. The structures were solved by direct (1A, 1B and 2) 15 or charge flipping 16 (3) methods and were refined by full-matrix least-squares on all unique F 2 values. 17 Anisotropic displacement parameters were refined for all non-hydrogen atoms; hydrogen atoms were geometrically constrained with Uiso(H) set at 1.2 (1.5 for methyl groups) times Ueq of the parent atom. During the structure analysis, the crystal of 1A was discovered to be a pseudo-merohedral twin. The TWINROTMAT routine in PLATON 18 was used to deconvolute the twin components and produce a file suitable for subsequent twin refinement. There are two possible orientations for the carbon atoms of the S2N2 ligand. The occupancies of the two components were refined competitively, converging at a ratio of 0.67:0.33. The lengths of chemically equivalent bonds of the disordered atoms were restrained to be approximately equal. Enhanced rigid bond and similarity restraints 17 were applied to the displacement parameters of all non-hydrogen atoms.
Computational details. All DFT calculations were performed using Gaussian 03. 19 Geometry optimizations and IR spectra were calculated using the BP86 functional. 20 Single point electronic calculations were performed using the three-parameter hybrid exchange functional 21 and the Lee-Yang-Parr correlation function 22 (B3LYP). For geometry optimizations, the basis set was an adapted version of that used by Hall et al.. 23 The Ni and Fe atoms were described by the Hay and Wadt basis set 24 with an effective core potential (ECP); the 4p orbitals in the ECP basis set were replaced by optimized (41) split valence functions from Couty and Hall 25 and augmented by an f-polarization function. 26 The standard LANL2DZ basis set was augmented with a d-polarization and p-diffuse function for S. 27 The 6-31G(d,p) 28   were made to maximize the yield of 3.

X-ray Crystallography of 1 -3
The X-ray crystal structures of 1 -3 are shown in Figure 2 and selected bond distances and angles are presented in Table 1.      (Table 1). Thus, it appears that 3 displays a phase in the solid state where high-spin (3A and 3B) and essentially low-spin (3C) species coexist. 40 Penta-co-ordinate Ni II complexes can exhibit temperature-dependent spin crossover that may be    The variation in reduction potential for 1 -3 with the nature of the co-ordination about the Ni center also suggests that the reduction process is associated with the Ni center via a Ni II Fe II /Ni I Fe II couple. This is supported by DFT calculations that show the LUMO is localized at the NiN2S2, NiS4 and NiN2S3 centers in 1 -3, respectively (see Supplementary Information). However, the chemically irreversible nature of this process precludes further study and a definitive assignment of the products of reduction cannot be confirmed. 3 also exhibits an oxidation process at 0.17 V vs Fc + /Fc, which is electrochemically reversible over the 20 -300 mVs −1 range of scan rates (Figure 3) Table   4). The relatively high g-anisotropy with gav ≠ ge is consistent with an unpaired electron that is localized principally on a transition metal, i.e., [Ni(L 3 )] + may be formulated as a formal Ni III center. 44  supporting electrolyte results in a frozen solution X-band EPR spectrum that may be simulated as an S = ½ center with spin Hamiltonian parameters g33 = 2.163, g22 = 2.129 and g11 = 2.014 ( Figure 5, Table 4).
The magnitude of the smallest g value (g33 ≈ ge) suggests that [3] + contains a formal Ni III center in which one unpaired electron resides primarily in a 3dz² orbital. 1c, 46 The frozen solution EPR spectrum of  (Table 4) that have been assigned to a (3dz²) 1 ground state. [46][47] Preliminary studies suggest that 3 and electrochemically prepared [3] + show no activity with respect to H2 evolution or H2 consumption following treatment with CF3COOH and H2, respectively.
We ascribe this to the saturated co-ordination sphere about the Fe center in 3 and [3] + .

DFT calculations
We undertook DFT calculations models of [3]  Selected metrical parameters for gas-phase, geometry optimized models of 3 in the S = 0 and S = 1 states are shown in Table 1 Figure 4(b) and Table 5].  The geometry of the Ni site in Ni-C may be described as approximately square-based pyramidal in which the base plane is formed from three CysS donors and the bridging hydride, and the apical position is occupied by the remaining CysS donor. 46 An analysis of the g-tensor for the Ni

Conclusions
We  hydrogenases through co-ordination of a second CO ligand to the Fe II center at the active site. 53a Our results appear to support this mechanism in which two CO and two CNligands co-ordinate the Fe(1) center in 1 -3, with no interactions of these ligands with the adjacent Ni(1) atom being observed.

Associated content
The Supporting Information contains Figures S1 -S8 and Tables S1 -S18. CCDC 1584099-1584102 contain the supplementary crystallographic data for this paper. These data can be obtained free