Synthetic Designs and Structural Investigations of Biomimetic Ni-Fe Thiolates

Abstract

Described are the syntheses of several Ni(μ-SR)₂Fe complexes, including hydride derivatives, in a search for improved models for the active site of [NiFe]-hydrogenases. The nickel(II) precursors include (i) nickel with tripodal ligands: Ni(PS₃)⁻ and Ni(NS₃)⁻ (PS₃^3- = tris(phenyl-2-thiolato)phosphine, NS₃^3- = tris(benzyl-2-thiolato)amine) (ii) traditional diphosphine-dithiolates, including chiral diphosphine R,R-DIPAMP, (iii) cationic Ni(phosphine-imine/amine) complexes, (iv) organonickel precursors Ni(o-tolyl)Cl(tmeda) and Ni(C₆F₅)₂. The following new nickel precursor complexes were characterized: PPh₄[Ni(NS₃)] and the dimeric imino/amino-phosphine complexes [NiCl₂(PCH=N^An)]₂ and [NiCl₂(PCH₂NH^An)]₂ (P = Ph₂PC₆H₄-2-). The iron(II) reagents include [CpFe(CO)₂(thf)]BF₄, [Cp*Fe(CO)(MeCN)₂]BF₄, FeI₂(CO)₄, FeCl₂(diphos)(CO)₂, and Fe(pdt)(CO)₂(diphos) (diphos = chelating diphosphines). Reactions of the nickel and iron complexes gave the following new Ni-Fe compounds: Cp*Fe(CO)Ni(NS₃), [Cp(CO)Fe(μ-pdt)Ni(dppbz)]BF₄, [(R,R-DIPAMP)Ni(μ-pdt)(H)Fe(CO)₃]BAr^F₄, [(PCH=N^An)Ni(μ-pdt)(Cl)Fe(dppbz)(CO)]BF₄, [(PCH₂NH^An)Ni(μ-pdt)(Cl)Fe(dppbz)(CO)]BF₄, [(PCH=N^An)Ni(μ-pdt)(H)Fe(dppbz)(CO)]BF₄, [(dppv)(CO)Fe(μ-pdt)]₂Ni, {H[(dppv)(CO)Fe(μ-pdt)]₂Ni]}BF₄, and (C₆F₅)₂Ni(μ-pdt)Fe(CO)₂(dppv) (DIPAMP = (CH₂P(C₆H₄-2-OMe)₂)₂; BAr^F₄⁻ = [B(C₆H₃-3,5-(CF₃)₂]₄⁻)) Within the context of Ni-(SR)₂-Fe complexes, these new complexes feature new microenvironments for the nickel center: tetrahedral Ni, chirality, imine and amine coligands, and Ni-C bonds. In the case of {H[(dppv)(CO)Fe(μ-pdt)]₂Ni}⁺, four low-energy isomers are separated by ≤ 3 kcal/mol, one of which features a biomimetic HNi(SR)₄ site, as supported by DFT calculations.

Table of Contents Synopsis and Graphic

The nickel complex of tris(benzy-2-thiolato)amine (NS₃^3-) is a robust air-stable tetrahedral nickel thiolate that failed to give suitably reactive active site models. A more traditional [NiFe]-models is the chiral [(R,R-DIPAMP)Ni(pdt)(H)Fe(CO)₃]⁺. The iminophosphine and amino-phosphine [NiCl₂(P-N)] complexes are precursors to the bimetallic derivatives [(P-N)Ni(pdt)(Cl)Fe(dppbz)(CO)]⁺. The organonickel precursors Ni(o-tolyl)Cl(tmeda) was used to prepare the basic mixed valence [(dppv)(CO)Fe(μ-pdt)]₂Ni, featuring a tetrahedral NiS₄ core. From Ni(C₆F₅)₂(dioxane)₂ was obtained (C₆F₅)₂Ni(pdt)Fe(CO)₂(dppv).

Introduction

[NiFe] hydrogenases are enzymes that catalyze the reversible oxidation of H₂, a process that is relevant to the metabolism of many microorganisms.¹ Motivated by the implications of H₂ redox and related reactions,² much effort has been expended on replicating the active site of these enzymes.³ Progress in modeling the active site can be measured by the structural similarity of models and the various catalytically significant states. These states include Ni-SI, Ni-R, Ni-L, and Ni-C. Structural models for the first three states have been described although they lack biomimetic terminal ligands. Analysis of models for the Ni(II)-Fe(II)-hydride state, Ni-R, is presented in Figure 1. The figure highlights aspects where synthetic models are deficient or differ strongly from the natural system.

Figure 1.

Active site of [NiFe]-hydrogenase core in the Ni-R state.

From the synthetic perspective, models either manipulate the Ni site or the Fe site. In this report we continue this approach, emphasizing expanding the range of ligands on nickel. We address, sometimes with only incremental progress, the following deficiencies in modeling: (1) nonplanar Ni(II) sites, (2) the rarity of Ni(SR)_x sites (x >2), (3) the asymmetry of the Ni…H-Fe interactions, and (4) the introduction of exchangeable terminal ligand sites on Ni, specifically those based on Ni-C sigma bonds.

Problems 1 and 2: Nonplanar Ni(II) sites and nonbridging NiSR sites.

The Ni-center in the Ni-SI_a is a distorted tetrahedron with four cysteine-thiolate ligands, two terminal and two bridging. The terminal thiolates are implicated as bases in proton transfer reactions,^4,5 such as the heterolysis of H₂. In the two models with terminal Ni-SR centers, the corresponding hydride derivatives could not be prepared.^6,7 All Ni(II)-containing models feature square planar or five-coordinate Ni. Thus, one goal in this work focused on incorporating Ni centers with rigidly nonplanar coordination sphere and at least one terminal thiolate.

Problem 3: the Asymmetry of the Ni⋯H-Fe Interactions.

Several hydride-containing NiFe complexes are known.³ They are all of the type L₂Ni(SR)₂FeHL₃. In these models the Fe-H bond is usually⁸ shorter than the Ni-H bond,^9–12 whereas in the enzyme, the d_Ni-H < d_Fe-H. ¹³ The shorter Ni-H distance underscores the role of Ni as the hydride carrier, whereas Fe center serves as a redox-inactive Lewis acid. In this paper, the Ni coordination sphere is manipulated in an effort to strengthen the Ni-H interaction.

Results and Discussion

1. Ni(II) complexes of tripodal PS₃ and NS₃ ligand and their [NiFe] derivatives.

Nonplanar Ni(II) thiolate complexes were targeted several years ago by Liaw, Koch and Millar, and their coworkers.^14–19 They described several complexes of tripodal phosphine trithiolates^20,21 (Figure S1), but not the corresponding Ni-Fe derivatives. Our efforts initially focused on the phosphine trithiolates, but quickly turned to amine trithiolates.

Ni(PS₃)- and Ni(NS₃)-Based Models.

Using P(C₆H₄-2-SH)₃, Liaw, Millar, and their coworkers described promising nickel precursors with terminal thiolates. Our experiments showed however that in its reactions with iron reagents, [Ni(PS₃H)PPh₃]¹⁶ (Scheme S1, Figures S2–11) serves as a source of PPh₃ or transfers the PS₃^3- chelate to Fe. One solution to the transfer of the PPh₃ involves use of [Ni(PS₃)]⁻, but this species does not exist per se.

In view of the challenges with the Ni(PS₃)-based models, we turned to complexes derived from N(CH₂C₆H₄-2-SH)₃, (abbreviated N(SH)₃). This tripodal ligand was developed by Koch et al., who characterized the tetrahedral complexes [M(NS₃)]⁻ (M = Ga(III), Zn(II), Fe(II)).^22–24 Considerable progress has also been reported on the synthesis of N(SH)₃. ^25,26 As shown schematically in eq 1, the NS₃ ligand envelops the metal center because it features six-membered chelate rings. In contrast the PS₃^3- ligands, forming only five-membered chelate rings, leave one site open for coordination of a fifth ligand.²⁰

The complex [Ni(NS₃)]⁻ was synthesized by combining [N(SNa)₃] with NiCl₂^.6H₂O in ethanol solution (eq 1).

(1)

The salt PPh₄[Ni(NS₃)] was isolated as green crystals. The solid is soluble in polar organic solvents to give air-stable solutions that are intensely green owing to strong absorptions at 477 (ε = 3800) and 637 nm (ε = 1200 cm⁻¹M⁻¹, Figure S12), consistent with tetrahedral Ni(II). This salt was characterized by X-ray crystallography, ESI-MS, ¹H-NMR spectroscopy, and elemental analyses. Its ¹H-NMR spectrum in CD₂Cl₂ solution spanned from d300 to −20 (Figure S13), consistent with the paramagnetic nature of this complex. ESI-MS in negative mode displayed the molecular ion peak (Figure S14).

According to X-ray crystallography, PPh₄[Ni(NS₃)] consists of a pseudo-tetrahedral Ni complex (Figure 2). The S-Ni-S bond angles range from 105 to 136°, perhaps due to the Jahn-Teller distortion arising from the (e₂)⁴(t₂)⁴ configuration. A related [Ga(NS₃)] complex exhibited average S-Ga-S and N-Ga-S angles of 115 and 102°, respectively.²² The Ni-N bond length is ~2.03 Å, whereas the Ni-S bond-lengths range from ~2.26–2.28 Å. Solutions of Ph₄P[Ni(NS₃)] are unreactive toward CO and PPh₃. In contrast, adducts of the type [Ni(PS₃)CO]⁻ have been characterized crystallographically.²⁰

Figure 2.

ORTEP (50 % probability) of the anion in PPh₄[Ni(NS₃)]. Ni(1)-N(1), 2.0310(15) Å; Ni(1)-S(1), 2.2621(5) Å; Ni(1)-S(3), 2.2705(5) Å; Ni(1)-S(2), 2.2829(5) Å. Selected angles: N(1)-Ni(1)-S(1): 100.49(4)°; N(1)-Ni(1)-S(3), 101.02(4)°; N(1)-Ni(1)-S(2), 100.83(4)°; S(1)-Ni(1)-S(3), 105.03(2)°; S(1)-Ni(1)-S(2), 136.60(2)°; S(3)-Ni(1)-S(2), 107.36(2)°.

To generate Fe-Ni complexes, reactions of [Ni(NS₃)]⁻ with cyclopentadienyl iron complexes were investigated.^27,28 Treatment of Na[Ni(NS₃] or Ph₄P[Ni(NS₃)] with [CpFe(CO)₂(thf)]BF₄ gave an immediate reaction. The deep brown product exhibited a pair of v_CO bands, shifted by 25–30 cm⁻¹ to lower energy with respect to the cationic precursor (Figure S15). This initial product is assigned as the 1:1 adduct Ni(NS₃)Fe(CO)₂Cp. Allowing solutions of this adduct to stand in ambient light, the pair of v_CO bands is replaced by a new broad peak at 1952 cm⁻¹. Similar shifts are observed for other L_xNi(SR)₂Fe(CO)Cp complexes from (planar) Ni-thiolates and Cp(CO)₂Fe-based precursors.²⁶ The positive-ion ESI-mass spectrum of the decarbonylated product (despite its being charge-neutral) shows an intense peak envelope centered at m/Z = 1173.9, which corresponds to [Ni(NS₃)Fe(CO)Cp]₂ (Figure S16). The same product was obtained from the reaction of [Ni(NS₃)]⁻ with [CpFe(CO)(MeCN)₂]BF₄ (v_CO = 2013 cm⁻¹, Figure S17–19). The experimental peak envelope was well simulated with theoretical values. The ¹H NMR spectrum of the product was broad.

Experiments were also conducted on related Cp*Fe(CO)-based complexes (Cp* = C₅Me₅). Being bulky, the Cp*Fe-based reagent was expected to favor formation of a 1:1 Ni-Fe derivative.²⁹ Indeed, treating a CH₂Cl₂ solution of [Ni(NS₃)]⁻ salts with [Cp*Fe(CO)(MeCN)₂]BF₄ generated a new complex with a single ν_CO band shifted to lower energy vs the cationic iron precursor (Figure S20). Based on the relative positions of CO and Cp* ligands, two stereoisomers are possible (Scheme 1), as have been observed for other Ni(μ-SR)₂Fe(CO)Cp* complexes.^30,31 The positive-ESI-mass spectrum shows intense peaks corresponding to M⁺ and [M-CO]⁺ (Figure S21). The isotopic distributions for both peak envelopes match simulated spectra (Figure S22, S23).

Scheme 1.

Structures proposed for isomers of Ni(NS₃)Fe(CO)Cp*.

2. Ni(P₂S₂) + Fe(II) Reagents

The addition of electrophilic iron reagents to nickel-(diphos)-dithiolates is a traditional route to Ni-(SR)₂-Fe complexes (diphos = chelating diphosphine ligand).^3,27 Analogous methods using planar NiN₂S₂ complexes have been well investigated, where N 2S₂^2- is a tetradentate diaminodithiolate.^32,33 Relative to NiN₂S₂, Ni(dithiolate)(diphos) complexes are advantageous because the geometry at Ni is less restrained, which allows the generation of tetrahedral Ni(I) derivatives.³

(a). [Cp(CO)Fe(μ-pdt)Ni(diphosphine)]⁺.

Reinvestigation of Schroeder’s synthesis of [Cp(CO)Fe(pdt)Ni(dppe)]⁺, ³⁴ although mainly confirmatory, is briefly described because the stepwise refinement of methodology is instructive. According to Schroeder, the reaction of Ni(pdt)(dppe) and CpFe(CO)₂I gives [Cp(CO)Fe(μ- pdt)Ni(dppe)]⁺ as well as a considerable amount of [Ni₂(μ-pdt)(μ-I)(dppe)₂]⁺. The salt [CpFe(CO)₂(thf)]BF₄ was found to react with Ni(pdt)(dppe) in a less complicated manner (Scheme 2, Figure S26–27). As revealed by in situ ³¹P NMR spectroscopy, however [Ni₂(μ-pdt)(μ-Cl)(dppe)₂]⁺ was a byproduct. When performed in THF (Figure S28), the reaction produced [CpFe(CO)(dppe)]⁺ as a side product. The lability of the diphosphine was suppressed by replacing Ni(pdt)(dppe) with Ni(pdt)(dppv) (dppv = cis-1,2-C 2H₂(PPh₂)₂, Figure S29–32). With this optimized protocol, the reaction Ni(pdt)(dppv) + [CpFe(CO)₂(thf)]⁺ was found to produce a previously undetected intermediate. With two equally intense ³¹P NMR signals, this new species is assigned as [Cp(CO)₂Fe(μ- pdt)Ni(dppv)]⁺, wherein only one thiolate bridges Ni and Fe.²⁰ The related reaction Ni(pdt)(dppbz) + [CpFe(CO)(NCMe)₂]⁺ produced high yields of [Cp(CO)Fe(μ- pdt)Ni(dppbz)]⁺, which was characterized by X-ray crystallography as its BF₄⁻ salt (Figure 3, S33–S35).

Scheme 2.

Reaction of Ni(pdt)(diphos) and [CpFe(CO)₂(thf)]⁺.

Figure 3.

ORTEP diagram (50 % probability) of the cation in the salt [Cp(CO)Fe(μ-pdt)Ni (dppbz)]BF₄ (H atoms were removed for clarity). Selected distances (Å): Ni(1)-P(2), 2.1508(7); Ni(1)-P(1), 2.1577(7); Ni(1)-S(2), 2.2104(7); Ni(1)-S(1), 2.2221(7); Ni(1)-Fe(1), 2.8066(5); Fe(1)-S(1), 2.2852(9); Fe(1)-S(2), 2.2784(7) Å.

[(CO)₃Fe(H)(μ-pdt)Ni(R,R-DIPAMP)]⁺.

Another route from Ni(dithiolate)(diphos) to Ni-(SR)₂-Fe complexes involves use of FeI₂(CO)₄ as a source of [FeI(CO)₃]⁺. ¹⁰ This approach was adopted to prepare a complex containing R,R-DIPAMP (DIPAMP = 1,2-bis[(2-methoxyphenyl)(phenyl)phosphino)]ethane). In addition to being chiral, R,RDIPAMP features 2-methoxyphenyl groups, which are suitably positioned to coordinate³⁵ to nickel, although this aspect was not manifested. As summarized in Scheme 3, NiCl₂(R,R-DIPAMP) was converted to Ni(pdt)(R,R-DIPAMP). ³¹P and ¹H NMR spectra confirm the diamagnetic nature of the complex (Figure S36–S37). Treatment of Ni(pdt)(R,R-DIPAMP) with FeI₂(CO)₄ yielded the corresponding iodo-bridged [NiFe] complex [(R,R-DIPAMP)Ni(μ-pdt)(μ-I)Fe(CO)₃]⁺, which was characterized in situ by FTIR spectrum. Reduction of this cationic iodo complex with Cp₂Co generated (R,RDIPAMP)Ni(μ-pdt)Fe(CO)₃, which was also characterized in situ by its FT-IR spectrum. Reaction of this reduced species with H(OEt₂)₂BAr^F₄ gave [(R,R-DIPAMP)Ni(μ- pdt)(H)Fe(CO)₃]BAr^F₄, which was fully characterized (BAr^F₄⁻ = [B(C₆H₃-3,5-(CF₃)₂]_4-Scheme 3, Figure S36–S41). The pattern of the v_CO bands in the FT-IR spectrum product and the intermediates (Figure 4, Figure S38–39) match those observed for related (achiral) complexes.¹⁰

Scheme 3.

Synthesis of [(R,R-DIPAMP)Ni(μ-pdt)(H)Fe(CO)₃]⁺.

Figure 4.

IR spectra (CH₂Cl₂) of [(R,R-DIPAMP)Ni(μ-pdt)(I)Fe(CO)₃]I (top), [(R,R-DIPAMP)Ni(μ-pdt)Fe(CO)₃] (middle), and [(R,R-DIPAMP)Ni(μ-pdt)(H)Fe(CO)₃]BAr^F₄ (bottom).

The crystallographically determined structure of [(R,R-DIPAMP)Ni(μ-pdt)(H)Fe(CO)₃]BAr^F₄ is shown in Figure 5. The Ni-Fe distance is 2.65 Å. The Ni and Fe centers adopt square-pyramidal and pseudo-octahedral geometries, respectively. The Ni-H and Fe-H distances are 1.94(4) and 1.58(4) Å, respectively. ³¹P-NMR spectrum of [(R,R-DIPAMP)Ni(μ-pdt)(H)Fe(CO)₃]BAr^F₄ exhibited two slightly broad peaks (δ65.0, 70.2), indicating the P’s on Ni are nonequivalent as required by the presence of the chiral diphosphine (Figure S40–S41).

Figure 5.

ORTEP diagram (50 % probability) of [(R,R-DIPAMP)Ni(μ-pdt)(H)Fe(CO)₃]BAr^F₄ (H atoms were removed for clarity. Selected distances (Å): Ni(1)-H(1): 1.94(4); Fe(1)-H(1): 1.58(4); Ni(1)-Fe(1): 2.6589(5); Ni(1)-S(1): 2.2153(8); Ni(1)-S(2): 2.2174(9); Ni(1)-P(1): 2.1629(8); Ni(1)-P(2): 2.1669(9); Fe(1)-S(1): 2.3119(9); Fe(1)-S(2): 2.3136(8).

3. NiCl₂(N-chelate) + Fe(pdt)(CO)₂(diphosphine)

An established route to nickel-iron dithiolates involves the reaction NiCl₂L₂ + Fe(pdt)(CO)₂L’₂. This method is useful when the corresponding Ni(SR)₂L₂ building block is not available, as illustrated by Song’s synthesis of models containing Ni(dppp)²⁺ centers (dppp = Ph₂PCH₂CH₂CH₂PPh₂).³⁶ We applied this approach to the preparation of Ni-Fe complexes with Ni(phosphine-amine/imine) centers. The majority of Ni-Fe-H complexes are derived from nickel diphosphine centers.³ Since NiCl₂ derivatives of P-N ligands tend to be pentacoordinate,³⁷ Ni(PN)-based models were anticipated to adopt novel structures.

[Ni(P-imine)]-Based Models.

This effort focused on complexes of the iminephosphine PCH=N^An derived from 2-diphenylphosphinobenzaldehyde and anisidine (H₂NC₆H₄OMe).³⁸ The reaction of equimolar NiCl₂(DME) and PCH=N^An gave a yellow paramagnetic 1:1 complex (DME = 1,2-dimethoxyethane). As shown by X-ray crystallography, the complex is the dimer [NiCl₂(PCH=N^An)]₂ (Figure 6). It features a pair of pentacoordinate Ni centers with two bridging chloride ligands. This dimeric motif is

(2)

well-precedented in other [NiCl₂(P-N)]₂ complexes.^39–41 The 5-coordinated paramagnetic high-spin nickel(II) centers adopt pseudo square pyramidal geometry (τ = 0.26⁴²) with apical phosphines.

Figure 6.

ORTEP diagram (50 % probability) of [Ni Cl₂(PCH=N^An)]₂; H atoms were removed for clarity Selected distances (Å): Ni(1)-Cl(1): 2.3638(5); Ni(1)-Cl(2): 2.3128(5); Ni(1)-Cl(1): 2.4033(6); Ni(1)-P(1): 2.2842(3); Ni(1)-N(1): 2.074(1).

The targeted Ni(SR)₂Fe complex was prepared by combining Fe(pdt)(CO)₂(dppbz) and [NiCl₂(PCH=N^An)]₂. When conducted in acetone solution at 0 °C in the presence of NaBF₄, the reaction gave [(PCH=N^An)Ni(μ-pdt)ClFe(dppbz)(CO)]BF₄ as the exclusive product (eq 2). With v_CO = 1946 cm⁻¹ (Figure S42), the cation resembles related complexes such as [(dppe)Ni(μ-pdt)ClFe(dppe)(CO)]⁺. ⁴³ The (+)-ESI-mass spectrum displayed peak envelopes for both the molecular ion ([M]⁺) and [M-CO]⁺ (Figure S43). The ³¹P-NMR spectrum exhibits three doublet of doublets (Figure S44). Signals at δ67.7 and 72.6 are assigned to the FeP centers, and that at δ19.4 is assigned to the NiP centers. Three nonequivalent phosphine sites indicates that the diphosphine on Fe is apical-basal, as is observed in other complexes of the type [(diphos)Ni(μ-pdt)ClFe(diphos)(CO)]⁺. ⁸ The ¹H-NMR spectrum is well resolved (Figure S45).

The reduced complex (PCH=N^An)Ni(μ-pdt)Fe(dppbz)(CO was obtained by treating [(PCH=N^An)Ni(μ-pdt)ClFe(dppbz)(CO)]BF₄ with Cp*₂Co. Yellow solids that precipitated from the reaction solution are presumed to be salts of [Cp*₂Co]⁺. The targeted (PCH=N^An)Ni(μ-pdt)Fe(dppbz)(CO) was isolated as a brown solid, which is soluble in THF, toluene, and CH₂Cl₂. In the IR spectrum of this [Ni^IFe^I] derivative, v_CO appears at 1919 cm⁻¹ vs 1946 cm⁻¹ for [(PCH=N^An)Ni(μ-pdt)ClFe(dppbz)(CO)]BF₄ (Figure S46). The ³¹P-NMR spectrum revealed three new sets of peaks, indicating that the diphosphine on Fe is apical-basal (Figure S47).

(3)

Protonation of (PCH=N^An)Ni(μ-pdt)Fe(dppbz)(CO) with HBF₄^.Et₂O yielded the corresponding hydride. This product is characterized by a v_CO band at 1945 cm⁻¹ (Figure S49), which is shifted by 26 cm⁻¹ vs its precursor (eq 3). The ³¹P-NMR spectrum consisted of two sets of three equi-intensity peaks, indicating the presence of two species (Figure S50). ESI-MS in the positive mode displayed peak envelopes for both the molecular ion ([M]⁺) and [M-CO]⁺ for [(PCH=N^An)Ni(pdt)(H)Fe(dppbz)(CO)]⁺ (Figure S52–S53). The ¹HNMR spectrum revealed one hydride peak (triplet) at δ−0.7 (Figure S54). Although the complex is stable in solid state, the hydride signal disappears over a period of days in CD₂Cl₂ solution. We tentatively assigned the product as [(PCH=N^An)Ni(μ-pdt)(H)Fe(dppbz)(CO)]BF₄; whereas the two set of peaks in ³¹P-NMR can be attributed to (i) the hydride complex (δ26.8, 91.21, 94.6), and (ii) the decomposition product (δ19.8, 72.9, 75.8) in CH₂Cl₂.

[Ni(P-amine)]-Based Models.

Using method completely analogous to the preceding section, Ni-Fe complexes were pursued with amino-phosphines. The required nickel amino-phosphine precursor was prepared from the phosphine-amine chelate PCH₂NH^An, which in turn was obtained by borohydride reduction of PCH=N^An (Figure S55–S56). Treating PCH₂NH^An with NiCl₂(DME) gave [NiCl₂(PCH₂NH^An)]₂, isolated in analytical purity as a yellow paramagnetic solid. X-ray crystallography established that the complex is dimeric, consisting of a pair of pentacoordinate Ni centers with bridging chloride ligands (Figure 7).²⁴ The 5-coordinated nickel(II) centers adopt pseudo square pyramidal geometry (T = 0.13 – 0.17).

Figure 7.

ORTEP diagram (50 % probability) of [NiCl₂(PCH₂NH^An)]₂. H atoms were removed for clarity except N-H. Selected distances (Å): Ni(1)-N(1), 2.1173(11); Ni(1)-Cl(2), 2.3065(3); Ni(1)-P(1), 2.3271(4); Ni(1)-Cl(1), 2.3544(3); Ni(1)-Cl(1)1’, 2.3584(3).

The preparation of the amine-supported Ni-Fe complex followed the method for the aforementioned (PCH=N^An)Ni-Fe complex. Combining Fe(pdt)(CO)₂(dppbz) and [NiCl₂(PCH₂NH^An)]₂ in acetone in the presence of NaBF₄ gave [(PCH₂NH^An)Ni(μ-pdt)ClFe(dppbz)(CO)]BF₄ in excellent yield (eq 4). The (+)-ESI-mass spectrum displayed peak envelops for both the molecular ion ([M]⁺) and [M-CO]⁺ (Figure S57). Judging from its IR spectrum (v_CO = 1949 cm⁻¹), the cation resembles the analogous PCH=N^An complex (Figure S58). The ³¹P NMR spectrum revealed a single species (3 peaks), showing that the secondary amine coordinates diastereoselectively (Figure S59).

Attempted reduction of [(PCH₂NH^An)Ni(μ-pdt)ClFe(dppbz)(CO)]BF₄ with Cp*₂Co produced [Cp*₂Co]⁺ as well as a large amount of free PCH₂NH^An, as verified respectively by ¹H and ³¹P NMR spectroscopy (Figure S61). The formation of free phosphine indicates that the PCH₂NH^An does not stabilize the reduced complex. Attempted dehydrohalogenation of [(PCH₂NH^An)Ni(μ-pdt)ClFe(dppbz)(CO)]⁺ using with NaOMe also yielded a large amount of free PCH₂NH^An (Figure S62).

(4)

4. NiR₂ + Fe(pdt)(CO)₂(diphosphine)

The modeling of the active site would be facilitated if it were possible to introduce ligands at nickel after formation of the Ni(μ-SR)₂Fe or Ni(μ-SR)₂(μ-H)Fe core. One specific idea was that terminal thiolates could be introduced by protonolysis of Ni-carbon sigma bonds using thiols.^44,45 Ni-(SR)₂-Fe complexes with Ni-carbon sigma-bonds are however unknown. To simplify the synthetic chemistry, we chose to focus on nickel aryl complexes.

(a). Ni(tolyl)Cl(tmeda) + Fe(pdt)(CO)₂(diphos).

Experiments aimed at generating Ni-aryl substituents entailed reaction of Fe(pdt)(CO)₂(dppv) and NiCl(2-tolyl)(tmeda) (tmeda = tetramethylethylenediamine).^46,47 With a labile diamine ligand, this Ni complex was expected to serve as a source of Ni(C₆H₄-2-Me)Cl. In practice, the reaction proved more complicated. Chromatographic work-up reliably gave yields of ~30% of [(dppv)(CO)Fe(μ-pdt)]₂Ni, isolated as a chocolate brown, air-stable solid (eq 5).

(5)

This complex is one of several with the Fe(SR)₂Ni(SR)₂Fe core.^27,48 The distinctive aspect of our Fe(SR)₂Ni(SR)₂Fe complex is that the constituent metals are not all divalent. The metal centers have a collective formal charge of 4+. The tetrahedral geometry of the Ni center is appropriate for Ni(0) or Ni(I), leading to descriptions of the complex as Fe(1.5)Ni(I)Fe(1.5) or Fe(II)Ni(0)Fe(II). The Fe(1.5)Ni(I)Fe(1.5) description is favored because the ν_CO band at 1914 cm⁻¹ is within 3 cm⁻¹ of that for (dppv)(CO)Fe(μ-pdt)Ni(dppv), which we had previously assigned as Fe(I)-Ni(I).⁸ The formation of a reduced complex is consistent with NiCl(2-tolyl)(tmeda) functioning as a source of Ni(0) via ligand redistribution and reductive elimination of bitolyl.⁴⁷ The complex undergoes reversible oxidation at −1.2 V.

The complex [(dppv)(CO)Fe(μ-pdt)]₂Ni is C₂-symmetric, consisting of a central [Ni(pdt)₂]^2n- center attached to a pair of [(dppv)(CO)Fe]ⁿ⁺ modules (Figure 8). The S-Ni-S angles range from 88 to 138°, with an average of 114°. The two Fe centers are equivalent, as imposed by the crystallographic site symmetry. Counting Ni-Fe bond as a sixth ligand, the Fe centers are pseudo-octahedral, which is also consistent with the SFe-P and P-Fe-C angles. The dppv ligands span apical and basal coordination site on Fe, which is typical for Fe(I)Ni(I) derivatives such as (dppv)(CO)Fe(μ-pdt)Ni(dppv).⁸

Figure 8.

ORTEP diagram (50 % probability) of [(dppv)(CO)Fe(μ-pdt)]₂Ni. H atoms were removed for clarity. Selected distances (Å) [DFT-calculated values]: Fe(1)-Ni(1), 2.4532(3) [2.523]; Fe(1)-P(1), 2.2134(7); Fe(1)-P(2), 2.1871(6); Fe(1)-S(1), 2.2589(6) [2.314]; Fe(1)-S(2), 2.2925(7) [2.271]; Ni(1)-S(1), 2.2459 [2.290]; Ni(1)-S(2), 2.2612 [2.235].

The ³¹P NMR spectrum [(dppv)(CO)Fe(μ-pdt)]₂Ni features two broad singlets, assigned to phosphorus centers attached to apical and basal sites on each Fe as observed also for (dppv)(CO)Fe(μ-pdt)Ni(dppv).⁸ Upon warming the sample to 80 °C, the two ³¹P NMR singlets broadened into the baseline, but coalescence was not observed.

The crystallographic structure of [(dppv)(CO)Fe(μ-pdt)]₂Ni matches the low-energy DFT result, although the computed bond distances are about 5% longer than those observed experimentally. The stereochemistry of the (dppv)(CO)Fe(pdt)NiL₂ core for the low energy structures follow structural patterns for related NiFe complexes:⁸ lower energy isomers differ with respect to apical-basal dppv vs basal-basal dppv (6.7 kcal/mol). A structure with planar NiS₄ center is observed only at 12.7 kcal/mol. This planar nickel structure can also be described with a dihedral angle (θ) of 0° between the S1-Ni1-S2 and S1’-Ni1-S2’ planes. High spin (triplet) states are high energy, especially for larger values of θ. Attempts to antiferromagnetically localize spins on metal centers (through the Broken Symmetry approximation) failed. NBO charges on Fe1, Ni, and Fe2 are −0.20, +0.61 and −0.20, respectively in 1A and −0.18, +0.53 and −0.18 in 1B. In other words, Fe centers carry more negative charge than Ni independent of θ. The frontier molecular orbitals are localized almost entirely on metal and sulfur centers. The HOMO is predominantly Fe1-Ni1-Fe2 σ bonding, whereas the LUMO is Fe1-Ni1 and Ni1-Fe2 σ* antibonding (Figure 9).

Figure 9.

Frontier MOs of [(dppv)(CO)Fe(μ-pdt)]₂Ni (isosurface cutoff = 0.05 a.u).

With oxidation states assigned as Fe(1.5)Ni(I)Fe(1.5) or Fe(I)Ni(II)Fe(I), [(dppv)(CO)Fe(μ-pdt)]₂Ni has two reducing equivalents and may therefore be expected to protonate. Indeed, as a CH₂Cl₂ solution, the complex reacts with HBF₄ to give the dark purple cation H[(dppv)(CO)Fe(μ-pdt)]₂Ni]⁺ (eq 6). This hydride complex was initially identified by its ESI-MS spectrum. The new species features v_CO bands at 1943 and 1961 cm⁻¹ in CH₂Cl₂ solution assigned to unprotonated and protonated FeCO centers, respectively. Like other Ni(SR)₂Fe hydrides, {H[(dppv)(CO)Fe(μ-pdt)]₂Ni]}⁺ is an active catalyst for hydrogen evolution from HBF₄, as demonstrated by cyclic voltammetry (Fig. S67–68).

The structure of {H[(dppv)(CO)Fe(μ-pdt)]₂Ni]}BF₄ was elucidated by single crystal X-ray diffraction (Figure 9). The connectivity in [(dppv)(CO)Fe(μ-pdt)]₂Ni is retained, however protonation induces Ni to convert from tetrahedral to square planar geometry.⁴⁹ The Fe-H and Ni-H distances are 1.65(4) and 1.82(4) Å, respectively. Such unsymmetrically bridging hydrides are typical for other Ni-Fe hydrides.⁹ The HFe-Ni distance is only slightly elongated (2.5653(9) Å) compared with the neighboring unprotonated Ni-Fe fragment (2.5327(9) Å). The Fe(1)-P and Fe(2)-P distances are almost the same, but the Fe-S distances for unprotonated Fe are 0.05 A shorter than the ~2.28 Å HFe-S distances.

According to ¹H NMR spectra, solutions of {H[(dppv)(CO)Fe(μ-pdt)]₂Ni]}⁺ exist as a mixture of three isomers as revealed by three sets of signals in the hydride region (Figure 10). These signals are centered at δ −4.9 (t, J(H,P) = 36 Hz), −6.6 (broad), and −15.5 (s) with respective integrated intensities of 1:1.5:2.5 at room temperature. The broad signal at δ −6.6 sharpens to a singlet at elevated temperatures. The ³¹P NMR spectrum of {H[(dppv)(CO)Fe(μ-pdt)]₂Ni]}BF₄ consists of five peaks, which can be assigned to the three isomeric hydrides (Figure S64). The relative intensities of the ¹H and ³¹P NMR signals are mutually consistent. In addition to {H[(dppv)(CO)Fe(μ-pdt)]₂Ni]}BF₄, the salt {H[(dppv)(CO)Fe(μ-pdt)]₂Ni]}BAr^F₄ was also prepared. The change in counteranion did not noticeably affect the NMR results, indicating that the isomer distribution is unaffected by ion pairing.

Figure 10.

ORTEP diagram (50 % probability) of the cation in {H[(dppv)(CO)Fe(μ-pdt)]₂Ni]}BF₄. H atoms were removed for clarity. Selected distances (Å): Fe(1)-Ni(1), 2.5327(9) [2.572]; Fe(2)-Ni(1), 2.5653(9) [2.598]; Fe(2)-H, 1.65(4) [1.635]; Ni(1)-H, 1.82(4) [1.745]; Fe(2)-S(3), 2.285(1) [2.264]; Fe(2)-S(4), 2.274(1) [2.262]; Ni(1)-S(1), 2.197(1) [2.249]; Ni(1)-S(2), 2.203(1) [2.248]; Ni(1)-S(3), 2.223(1) [2.263]; Ni(1)-S(4), 2.217(1) [2.264].

Additional insight into the structures of and relationships between the isomeric hydrides was provided by low temperature NMR studies (Figure S65). Solutions prepared by dissolving crystals of {H[(dppv)(CO)Fe(μ-pdt)]₂Ni]}BF₄ in CD₂Cl₂ at −80 °C contain exclusively the isomer with a hydride signal at δ −4.9. At −80 °C, the ³¹P NMR spectrum consists of four equally intense signals, which are decoalesced daughters of the δ 86 and 72 signals observed in the thermally equilibrated room temperature spectrum. The relative frequencies of these four signals indicate that the splitting does not arise from spin-spin coupling; instead, these signals are assigned to four nonequivalent phosphine sites. Because the cation {H[(dppv)(CO)Fe(μ-pdt)]₂Ni]}⁺ is C_s-symmetric, the observation of four signals may be attributed to frozen conformations of the pdt^2- backbones. Upon warming the solution to −45 °C, the four ³¹P NMR signals coalesce into two broad singlets, which continue to sharpen as the solution is warmed to room temperature. These two singlets are assigned to protonated and unprotonated Fe(dppv)(CO) sites. These data provide some insight into the slowness of the degenerate hydride (or proton) transfer between two equivalent FeNi sites. Also appearing in the −45 °C spectrum are three signals in intensity 1:1:2 (the latter intensity from accidental degeneracy), assigned to second isomer. Near room temperature these three signals coalesce into the singlet at δ 86.8, indicating again that two Fe(dppv) sites are equilibrating rapidly on the NMR timescale. At −15 °C two broad singlets appear at low field, assigned to a third isomer. These signals appear concomitant with the appearance of the hydride signal at − δ15.5.

The crystallographic structure for {H[(dppv)(CO)Fe(μ-pdt)]₂Ni}⁺ matches the low-energy DFT result. The DFT structure is one of four isomers that are separated by ≤ 3 kcal/mol. These isomeric hydrides differ in terms of the stereochemistry of the Fe(dppv)(CO) site and θ, the dihedral or “twist” angle defined by the S1NiS2 and S3NiS4 planes. All stable isomers feature bridging hydrides. The two lowest-energy isomers are depicted in eq 6 and Figure 12.

(6)

Figure 12.

Optimized structures of the low-energy isomers of {H[(dppv)(CO)Fe(μ-pdt)]₂Ni}⁺. Selected distances (Å): Fe(2)-H, 1.635 in [H1A]⁺ and 1.654 in [H1B]⁺; Ni(1)-H, 1.745 in [H1A]⁺ and 1.643 in [H1B]⁺.

The DFT-calculated chemical shifts of the hydride ligand, δ_H, correlate with the twist angle. When θ = 0, ~δ_H-2 (e.g., in [1A]⁺) vs the observed chemical shift of δ−4.9. For the isomers with θ ~ 60°, δ_H ranges from ~ δ −9.1 to −10.2.

(b). Ni(C₆F₅)₂(dioxane)₂ + Fe(pdt)(CO)₂(diphos).

The previous experiment using NiCl(2-tolyl)(tmeda) failed to deliver Ni-Fe complexes with Ni-C sigma bonds. The methodology was revised with a focus on Ni(C₆F₅)₂(dioxane)₂, which is known to have labile ether ligands and robust Ni-C bonds.⁵⁰ Indeed, this complex was found to react readily with Fe(pdt)(CO)₂(dppv) to form a ruby-red 1:1 adduct (C₆F₅)₂Ni(μ-pdt)Fe(CO)₂(dppv) (eq 7, Figure S66 −69).

(7)

The ³¹P and ¹⁹F NMR data are consistent with a C_s-symmetric Ni-Fe complex with a square planar Ni center and octahedral Fe site (Figure S68–69). X-ray crystallographic characterization confirms these findings (Figure 13).

Figure 13.

ORTEP diagram (50 % probability) of (C₆F₅)₂Ni(μ-pdt)Fe(CO)₂(dppv). H atoms were removed for clarity. Selected distances (Å): Ni(1)-S(1), 2.2596(8); Ni(1)-S(2), 2.2335(8); Fe(1)-S(1), 2.3313(8); Fe(1)-S(2), 2.3313(7); Fe(1)-P(1), 2.2246(8); Fe(1)-P(2), 2.2183(9).

Summary and Conclusions

Many properties of the active site of the [NiFe]-hydrogenases remain to be simulated with models (Fig. 1). Addressing these gaps requires the development of new and the evolution of established synthetic methods. This report describes both approaches to complexes with Ni(μ-SR)₂Fe cores.

The boldest approach to active site models entailed the use of the tetrahedral [Ni(NS₃)]⁻ building block. The starting complex [Ni(NS₃)]⁻ is a remarkable example of a robust, air-stable tetrahedral nickel(II) thiolate. Many nickel tri- and tetrathiolates are known, including the iconic [Ni(SPh)₄]^2-, ⁵¹ but they are either exchange-labile, oligomerize, or both.⁵² The complex [Ni(NS₃)]⁻ was deployed in an attempt to approximate the geometry of the Ni(SR)₄ site, as seen in the enzyme. One barrier to producing Ni-(SR)₂-Fe ensembles, suggested by the formation of tetrametallic complexes with [Ni(NS₃)]₂Fe₂ cores, is the wide S-Ni-S angle in [Ni(NS₃)]⁻. An associated deficiency in this Ni-containing building block is that the Ni center exhibits no Lewis acidity, which is a requirement for Ni(II)-Fe(II) models.

The active sites of the hydrogenases are chiral owing to the asymmetric distribution of ligands around the metal centers. Using chiral ancillary ligands, e.g. R,R-DIPAMP, represents one method to impose asymmetry. The twisting introduced by the C₂-symmetric ligand is not however transmitted to the geometry at the Ni cite, which adopts a planar NiS₂P₂ coordination sphere. Furthermore, neither the Ni center nor the hydride exhibited any affinity for the methoxy substituent. In the enzyme, water is assumed to bind to Ni(II) and especially Ni(III).⁴ The work does however demonstrate the utility of chiral ligands to probe the stereodynamics of Ni-(SR)₂-Fe models. In a related modelling effort, phosphine-imine and phosphine-amine ligands were installed in rare examples of models with Ni-N bonds.

The final phase of this study examined nickel-aryl reagents. The original idea was to introduce alkyl/aryl ligands that would be susceptible to protonolysis by thiols. In one case, this approach afforded the first Ni(SR)₄-hydrides. As active site models, these NiFe₂ complexes have incorrect metal stoichiometry. Nonetheless DFT calculations support the existence of Fe-H-Ni(SR)₄ centers where the geometry at Ni approaches that in the enzyme.¹³ The preparation of Fe-H-Ni(SR)₄ centers with terminal thiolates remains an unsolved challenge.

Experimental

General.

Literature procedures were followed for the preparation of P(SH)₃ from thiophenol.⁵³ For the preparation of N(SH)₃, N(CH₂C₆H₄-2-Br)₃ was prepared using aqueous ammonia method,²⁶ followed by lithiation and sulfidation.^54,55 PCH=N^An was prepared by a modification of the published route,⁵⁶ as described in the Supporting Information. The complexes Ni(pdt)(dppe), Ni(pdt)(dppv), and Ni(pdt)(dppbz) were prepared from the corresponding NiCl₂(diphosphine) by treatment with propanedithiol and two equiv of NaOMe in a mixture of CH₂Cl₂ and MeOH. The following organometallic reagents were prepared according to literature procedures: [CpFe(CO)₂(thf)]BF₄,⁵⁷ [CpFe(CO)(MeCN)₂]BF₄ and [Cp*Fe(CO)(MeCN)₂]BF_4,⁵⁸ NiCl₂(DME),⁵⁹ Fe(dppv)(CO)₂Cl₂,⁶⁰ FeI₂(CO)_4.,⁶¹ and Ni(C₆F₅)₂(dioxane)₂. ⁵⁰ Fe(pdt)(CO)₂(dppv) and Fe(pdt)(CO)₂(dppbz) were prepared as we previously described.⁶² Efforts to improve the yields were fruitless. Other methods have been described recently.⁶³

Ph₄P[Ni(NS₃)].

A solution of NaOMe (0.216 g, 4 mmol) in MeOH (10 mL) was added to a slurry of [NH(SH)₃]⁺Cl⁻ (0.420 g, 1mmol), also in MeOH (20 mL). The resulting homogeneous solution of N(SNa)₃ was allowed to stir for 30 min. before being evaporated. A solution of this white residue in 10 mL of EtOH was treated with a solution of NiCl₂^.6H₂O (0.240 g, 1 mmol) in 10 mL of EtOH. The color of the solution immediately changed to dark green. After 1 h, solvent was removed, leaving a green-colored residue of Na[Ni(NS₃)]. Alternatively, addition of Ph₄PBr (0.630 g, 1.5 mmol) to the dark green Na[Ni(NS₃) solution precipitated a green solid. Yield: 662 mg (85%). X-ray quality crystals of Ph₄P[Ni(NS₃)] were obtained from vapor diffusion of pentane into CH₂Cl₂ solution of the complex at −30 °C. ¹H NMR (500 MHz, CD₂Cl₂): δ −11.50, −10.85, 7.70, 7.95, 8.05, 13.33, 19.40, 96.99, 287.43. ³¹P NMR (202 MHz, CH₂Cl₂): δ 23.58. ESI-MS (negative, CH₂Cl₂): 437.9965 ([M]⁻, 100%). UV-Vis (CH₂Cl₂): λ = 373 nm (ε = 6367 cm-¹M⁻¹), 435 (ε = 3681), 477 (ε = 3766), 637 (ε = 1185). Anal. Calcd for C₄₆H₄₀Cl₂NNiPS₃ [Ph₄P[Ni(NS₃)] CH₂Cl₂ (found): C, 63.98 (64.08); H, 4.67 (4.71); N, 1.62 (1.81).

[Ni(NS₃)Fe(CO)₂Cp] and [Ni(NS₃)Fe(CO)Cp]₂.

A solution of Na[Ni(NS₃)] (0.462 g, 1 mmol) (or Ph₄P[Ni(NS₃)] (0.779 g, 1 mmol)) in 15 mL of THF was treated with a solution of [CpFe(CO)₂(thf)]BF₄ (0.336 g, 1 mmol) in 15 mL of THF. The resulting dark brownish solution was stirred for 2 h. Addition of pentane to the reaction mixture precipitated [Ni(NS₃)Fe(CO)₂Cp] as a brown solid (IR (THF): v_CO = 2046, 2002 cm⁻¹). A THF solution of this crude product was stirred for a few days in daylight. Vapor diffusion of pentane into the THF solution of the complex yielded crude [Ni(NS₃)Fe(CO)Cp]₂. This complex was purified by extraction into CH₂Cl₂. The filtered extract was evaporated to dryness to yield the desired product. This [Ni(NS₃)Fe(CO)Cp]₂ complex can also be synthesized upon reaction between Na[Ni(NS₃)] (0.462 g, 1 mmol) (or Ph₄P[Ni(NS₃)] (0.779 g, 1 mmol)) and [CpFe(CO)(MeCN)₂]BF₄ (0.318 g, 1 mmol) in 20 mL of CH₂Cl₂. After 1 day, the solution was filtered, and the filtrate was evaporated, leaving a residue of [Ni(NS₃)Fe(CO)Cp]₂. Low quality crystals could be obtained by layering a CH₂Cl₂ solution of the complex with pentane in an NMR tube at room temperature. Yield: 441 mg (75%). IR (CH₂Cl₂): v_CO = 1957 cm⁻¹. ESI-MS (positive, CH₂Cl₂): 1173.9290 ([M]⁺, 100%). Anal. Calcd for C₅₇H₅₂Cl₆Fe₂N₂Ni₂O₂S₆ [[Ni(NS₃)Fe(CO)Cp]₂·3CH₂Cl₂] (found): C, 47.84 (47.52); H, 3.66 (3.46); N, 1.96 (1.67).

Ni(NS₃)Fe(CO)Cp*.

A solution of Na[Ni(NS₃)] (0.462 g, 1 mmol) (or Ph₄P[Ni(NS₃)] (0.779 g, 1 mmol)) in 15 mL of CH₂Cl₂ was treated with a solution of [Cp*Fe(CO)(MeCN)₂]BF₄ (0.388 g, 1 mmol) in 15 mL of CH₂Cl₂. After being stirred for 24 h, the reaction solution was filtered, and the filtrate was evaporated to leave a brown residue. Yield: 461 mg (70%). IR (CH₂Cl₂): v_CO = 1954, 1925 cm⁻¹. ESI-MS (CH₂Cl₂): 657.0421 ([M]⁺, 100%), 629.0474 ([M-CO]⁺, 77%). Anal. Calcd for C₃₆H₄₁Cl₈FeNNiOS₃ [[Ni(NS₃)Fe(CO)Cp*]·4CH₂Cl₂] (found): C, 43.32 (43.28); H, 4.14 (4.34); N, 1.40 (1.92).

Reaction of [CpFe(CO)₂(thf)]BF₄ with Ni(pdt)(dppe).

A solution of [CpFe(CO)₂(thf)]BF₄ (0.336 g, 1 mmol) in 15 mL of CH₂Cl₂ was added to a solution of Ni(pdt)(dppe) (0.563 g, 1 mmol) in 25 mL of CH₂Cl₂. The reaction mixture was stirred for 72 h and then filtered. Dilution of the filtrate with pentane precipitated a mixture of [Cp(CO)Fe(μ-pdt)Ni(dppe)]BF₄ and [Ni₂(μ-pdt)(μ-Cl)(dppe)₂]BF₄, which were identified spectroscopically. ³¹P NMR (202 MHz, CH₂Cl₂): δ56.58, 56.95, 57.81; IR (CH₂Cl₂): v_CO = 1953 cm⁻¹. X-ray quality crystals of [Cp(CO)Fe(μ-pdt)Ni(dppe)]BF₄ were obtained by vapor diffusion of pentane into CH₂Cl₂ solution of the mixture. When the reaction was repeated using THF in place of CH₂Cl₂, the pentane-precipitate was a mixture of [Cp(CO)Fe(μ-pdt)Ni(dppe)]BF₄ and [CpFe(CO)(dppe)]BF₄, which were identified spectroscopically. ¹H NMR (500 MHz, CH₂Cl₂): δ2.51 – 2.67 (2H, m), 2.67 – 2.82 (2H, m), 4.75 (5H, s), 7.11 – 7.22 (4H, m), 7.45 (4H, d, J = 6.71), 7.48 – 7.64 (12H, m). ³¹P NMR (202 MHz, THF): δ92.93. IR (CH₂Cl₂): v_CO = 1980 cm⁻¹.

Reaction of [CpFe(CO)₂(thf)]BF₄ with Ni(pdt)(dppv).

A solution of [CpFe(CO)₂(thf)]BF₄ (0.336 g, 1 mmol) in 15 mL of CH₂Cl₂ was added into a solution of Ni(dppv)(pdt) (0.561 g, 1 mmol) in 25 mL of CH₂Cl₂. After stirring for 3 days, the mixture was filtered, and the filtrate was diluted with pentane to precipitate a mixture of [Cp(CO)Fe(μ-pdt)Ni(dppv)]BF₄ and [Ni₂(pdt)(μ-Cl)(dppv)₂]BF₄, which were identified spectroscopically. ³¹P NMR (202 MHz, CH₂Cl₂): δ64.61, 65.97, 66.58, 59.07. The last two signals are assigned to [Cp(CO)Fe(μ-pdt)Ni(dppv)]BF₄ and [Ni₂(pdt)(μ-Cl)(dppv)₂]BF₄, respectively. When the same reaction was repeated in THF, the exclusive product precipitated with pentane was [Cp(CO)Fe(μ-pdt)Ni(dppv)]BF₄. Yield: 558 mg (70%). ³¹P NMR (202 MHz, THF): δ 64.6 (major), 66.0, 66.6. (weak) IR (THF): v_CO = 1946 cm⁻¹.

[Cp(CO)Fe(μ-pdt)Ni(dppbz)]BF₄.

A solution of [CpFe(CO)(MeCN)₂]BF₄ (0.032 g, 0.1 mmol) in 5 mL of THF was added into a solution of Ni(pdt)(dppbz) (0.061 g, 0.1 mmol) in 5 mL of THF. As the reaction mixture stirred for 24 h, green solid precipitated. X-ray quality crystals were grown by vapor diffusion of pentane into CH₂Cl₂ solution of the salt at −29 °C. Yield: 72 mg (85%). ³¹P NMR (202 MHz, CH₂Cl₂): δ56.4 (major), 57.3 (small peak, may be due to a stereoisomer based on the position of Cp and CO) IR (CH₂Cl₂): v_CO = 1953 cm⁻¹. ESI-MS (positive, CH₂Cl₂): 759.0300 ([M]⁺), 731.0234 ([M-CO]⁺). Anal. Calcd for C_39.5H₃₆BClF₄FeNiOP₂S₂ [[Cp(CO)Fe(μ-pdt)Ni(dppbz)]BF₄]·0.5CH₂Cl₂] (found): C, 53.33 (53.8); H, 4.08 (4.52).

NiCl₂(R,R-DIPAMP).

A slurry of NiCl₂·6H₂O (0.65 mmol, 85 mg) and R,R-DIPAMP (0.65 mmol, 300 mg) in 5 mL of EtOH was heated at reflux for 15 min. After cooling the mixture to 0 °C, an orange solid precipitated. The solid product was washed with EtOH and Et₂O. Yield: 360 mg (94%). Anal. Calcd for C_28.2H_28.4Cl_2.4NiO₂P₂ [NiCl₂(R,R DIPAMP).0.2 CH₂Cl₂] (found): C, 55.98 (55.97); H, 4.73 (4.6).

Ni(pdt)(R,R-DIPAMP).

A suspension of NiCl₂(R,R-DIPAMP) (0.61 mmol, 360 mg) and 1,3-propanedithiol (0.61 mmol) in CH₂Cl₂ was treated dropwise with a solution of NaOMe (1.22 mmol) in 1 mL of MeOH. The resulting mixture was stirred for 30 min. Solvent was removed, and the residue was extracted into CH₂Cl₂. The extract was filtered through a pad of Celite and then evaporated to afford Ni(pdt)(R,R-DIPAMP) as a brown-red solid. Yield: 350 mg (92%). ¹H NMR (500 MHz, CD₂Cl₂) δ8.33 (s, 4H), 7.58 (s, 6H), 7.38 (t, J = 7.7 Hz, 2H), 6.99 – 6.79 (m, 6H), 3.39 (s, 6H), 2.41–2.35 (m, 4H), 2.30–2.25 m, 2H), 2.24–2.20 (m, 2H), 1.89 (q, J = 6.0 Hz, 2H). ³¹P NMR (202 MHz, CD₂Cl₂): δ 53.4 (s).

[(R,R-DIPAMP)Ni(μ-pdt)(H)Fe(CO)₃]BAr^F₄.

At −80 °C, a stirred mixture of FeI₂(CO)₄ (236 mg, 0.56 mmol) and Ni(pdt)(R,R-DIPAMP) (350 mg, 0.56 mmol) was treated with 7 mL of CH₂Cl₂, resulting in a dark brown homogeneous solution. After 2 min, the IR spectrum of the mixture showed v_CO = 2094, 2076, 2050 and 2021 cm⁻¹, consistent with the formation of the μ-iodo intermediate [(R,R-DIPAMP)Ni(μ-pdt)(I)Fe(CO)₃]⁺. The cold reaction mixture was treated with 2 equiv (212 mg, 1.12 mmol) of Cp₂Co. The reaction solution became dark green. After stirring for 1h at room temperature, the reaction mixture was evaporated, and the resulting green solid was washed with MeCN. The solid was extracted into 5 mL of CH₂Cl₂, and the extract was filtered through Celite. The CH₂Cl₂ solution was evaporated, leaving (R,R-DIPAMP)Ni(μ-pdt)Fe(CO)₃ as a green solid. Yield: 215 mg (50%). ¹H NMR (500 MHz, CD₂Cl₂): δ 7.78 (s, 6H), 7.37 (s, 8H), 7.13 – 6.73 (m, 4H), 3.67 (s, 6H), 2.62 (s, 2H), 2.43–2.27 (m, 4H), 2.05 – 1.61 (m, 4H). ³¹P NMR (202 MHz, CD₂Cl₂): δ 60.0 (s). IR (CH₂Cl₂): v_CO = 2024, 1950 cm⁻¹. A stirred solution of (R,R-DIPAMP)Ni(μ-pdt)Fe(CO)₃ (100 mg, 0.13 mmol) in 3 mL of CH₂Cl₂ was treated with H(OEt₂)₂BAr^F₄ (121 mg, 0.12 mmol). An immediate color change from green to red was observed. The solution was concentrated to ~ 1 mL. The concentrate was layered with pentane and cooled to −20 °C to give [(R,R-DIPAMP)Ni(μ-pdt)(H)Fe(CO)₃]BAr^F₄ as red colored solid. Yield: 82 mg (42%). ³¹P NMR (202 MHz, CD₂Cl₂): δ 66.8 (br. s). ¹H NMR (500 MHz, CD₂Cl₂): δ −3.06 (s, 1H, hydride). IR (CH₂Cl₂): v_CO = 2078, 2016 cm⁻¹. Anal. Calcd for C₆₇H₄₉Cl₂BF₂₄FeNiO₅P₂S₂ [(R,R-DIPAMP)Ni(μ-pdt)(H)Fe(CO)₃]BAr^F₄^·CH₂Cl₂] (found): C, 46.99 (47.31); H, 2.88 (2.89).

[NiCl₂(PCH=N^An)]₂.

A slurry of NiCl₂(DME) (0.220 g, 1 mmol) in 15 mL of CH₂Cl₂ was added into a stirred solution of PCH=N^An (0.395 g, 1 mmol) in 20 mL of CH₂Cl₂. After 2 h, the solution was concentrated to 10 mL. Dilution with pentane gave yellow-brown crystals. X-ray quality crystals were obtained upon layering a CH₂Cl₂ solution of the complex with pentane. Yield: 420 mg (80%). Anal. Calcd for C₅₃H₄₆Cl₆N₂Ni₂O₂P₂ [[NiCl₂(PCH=N^An)]₂·CH₂Cl₂] (found): C, 56.09 (55.87); H, 4.09 (3.81); N, 2.47 (2.49).

PCH₂NH^An.

To a solution of PCH=N^An (0.490 g, 1.24 mmol) in 30 mL of MeOH at 0 °C was added NaBH₄ (0.180 g, 4.76 mmol) over the course of a few minutes. After the addition was complete, the reaction mixture was stirred for 15 min. The reaction mixture was treated with aqueous 1M NaOH solution (5 mL). The mixture was extracted into CH₂Cl₂. After drying the extract over MgSO₄, solvent was removed leaving a colorless oil. Yield: 443 mg (90%). ³¹P NMR (202 MHz, CD₂Cl₂): δ −15.88 (s). ¹H NMR (500 MHz, CD₂Cl₂): 3.74 (s, 3H), 4.48 (d, 2H), 6.41 (d, 2H), 6.72 (d, 2H), 6.95 (t, 1H), 7.21 (t, 1H), 7.29 – 7.38 (m, 12 H), 7.53 (q, 1H).

[NiCl₂(PCH₂NH^An)]₂.

A slurry of NiCl₂(DME) (0.110 g, 0.5 mmol) in 15 mL of CH₂Cl₂ was added into a solution of PCH₂NH^An (0.199 g, 0.5 mmol) in 20 mL of CH₂Cl₂. From the initially homogeneous solution precipitated a yellow solid. The reaction was continued for 2 h before the solid was collected by filtration. This compound is insoluble in CH₂Cl₂ but soluble in THF. X-ray quality crystals were obtained upon vapor diffusion of pentane into THF solution of the complex at room temp. Yield: 210 mg (80%). Anal. Calcd for C₅₃H₅₀Cl₆N₂Ni₂O₂P₂ [[NiCl₂(PCH₂NH^An)]₂·CH₂Cl₂] (found): C, 55.89 (56.33); H, 4.42 (5.00); N, 2.46 (2.44).

[(PCH=N^An)Ni(μ-pdt)ClFe(dppbz)(CO)]BF₄.

A slurry of [NiCl₂(PCH=N^An)]₂ (0.050 g, 0.05 mmol) in 10 mL of acetone was treated with a solution of NaBF₄ (0.017 g, 0.15 mmol) in 10 mL of acetone followed by a slurry of Fe(pdt)(CO)₂(dppbz) (0.066 g, 0.1 mmol) also in 10 mL of acetone. The mixture immediately became dark brown. After stirring the mixture for 2 h at 0 °C, solvent was removed under vacuum. The solid residue was extracted into CH₂Cl₂, and this extract was filtered through glass wool. The filtrate was evaporated to give a brown-colored solid. Crystals were obtained from vapor diffusion of pentane into a CH₂Cl₂ solution of the salt at −30 °C. Yield: 115 mg (95%). IR (CH₂Cl₂): v_CO = 1946 cm⁻¹. ESI-MS (positive, CH₂Cl₂): 1124.1044 [M]⁺, 1096.1278 [MCO]⁺. ³¹P NMR (202 MHz, CH₂Cl₂): δ72.66 (dd, 1P, Fe-P), 67.80 (dd, 1P, Fe-P), 19.43 (dd, 1P, Ni-P). ¹H-NMR (500 MHz, CD₂Cl₂): δ7.76–6.54 (overlapping signals, 43H), 3.80 (s, 3H), 3.62 (dt, J = 13.73, 4.58 Hz, 1 H), 2.89 (d, J = 13.73 Hz, 1 H), 2.27 (d, J = 13.73 Hz, 1 H), 2.01 – 2.10 (m, 1 H), 1.92 – 2.01 (m, 1 H), 1.86 (d, J = 7.63 Hz, 1 H). Anal. Calcd for C_62.5H₅₇BCl₆F₄FeNNiO₂P₃S₂ [[(PCH=N^An)Ni(μ-pdt)ClFe(dppbz)(CO)]BF₄^.2.5 CH₂Cl₂] (found): C, 52.67 (52.39); H, 4.03 (4.2); N, 0.98 (1.0).

[(PCH=N^An)Ni(μ-pdt)(H)Fe(dppbz)(CO)]BF₄.

A solution of Cp*₂Co (0.066 g, 0.2 mmol) in 5 mL of THF was added into a slurry of [(PCH=N^An)Ni(μ-pdt)ClFe(dppbz)(CO)]BF₄ (0.121 g, 0.1 mmol) in 5 mL of THF. The resulting mixture was stirred for 2 h at room temperature. The yellow solid precipitate ([Cp*₂Co]Cl/[Cp*₂Co]BF₄) was filtered off. The filtrate was evaporated to dryness to obtain the (PCH=N^An)Ni(μ-pdt)Fe(dppbz)(CO) as a brown solid. This compound is soluble in THF, toluene, and CH₂Cl₂. IR (CH₂Cl₂): v_CO = 1919 cm⁻¹. ³¹P NMR (202 MHz, THF): δ88.35 (dd, 1P, Fe-P), 80.74 (dd, 1P, Fe-P), 36.13 (q, 1P, Ni-P). ¹H-NMR (500 MHz, THF-d⁸): δ8.22 – 6.16 (overlapping signals, 43H), 3.49 (s, 3H), 2.17 (m, 1H), 1.92 (m, 1H), 1.84 (m, 1H), 1.66 (m, 1H), 1.13 (m, 1H), 1.04 (m, 1H). Crude (PCH=N^An)Ni(μ-pdt)Fe(dppbz)(CO) (0.087 g, 0.08 mmol) was dissolved in 10 mL of THF, and HBF₄·Et₂O (11 μL, 0.08 mmol) was added dropwise to this solution. The resulting dark brown mixture was stirred for 2 h. The solvent was evaporated to dryness to obtain [(PCH=N^An)Ni(μ-pdt)(H)Fe(dppbz)(CO)]BF₄. Crystals were obtained upon vapor diffusion of pentane into CH₂Cl₂ solution of the salt. Yield: 85 mg (90%). IR (CH₂Cl₂): v_CO = 1945 cm⁻¹. ESI-MS (positive, CH₂Cl₂): 1090.1417 [M]⁺, 1062.1503 [M-CO]⁺. ³¹P NMR (202 MHz, THF): δ94.47 (1P, Fe-P), 91.58 (1P, Fe-P), 76.78 – 77.30 (2P, Fe-P), 28.01 (d, 1P, Ni-P), 21.50 (t, 1P, Ni-P). ³¹P NMR (202 MHz, CD₂Cl₂): δ 94.55 (1P, Fe-P), 91.01 (1P, Fe-P), 75.83 (0.5P, Fe-P), 72.92 (0.5P, Fe-P), 26.75 (1P, Ni-P), 19.83 (0.5P,Ni-P). ¹H NMR (CD₂Cl₂): −0.66 (t, hydride, 1H). Anal Calcd for C₆₁H₅₅BCl₂F₄FeNNiO₂P₃S₂ [[(PCH=N^An)Ni(μ-pdt)(H)Fe(dppbz)(CO)]BF₄. CH₂Cl₂] (found): C, 57.99 (58.13); H, 4.39 (5.05); N, 1.11 (1.12).

[(PCH₂NH^An)Ni(μ-pdt)ClFe(dppbz)(CO)]BF₄.

A slurry of Fe(pdt)(CO)₂(dppbz) (0.066 g, 0.1 mmol) in 10 mL of acetone was treated with a slurry of [NiCl₂(PCH₂NH^An)]₂ (0.050 g, 0.05 mmol) in 10 mL of acetone and a solution of NaBF₄ (0.017 g, 0.15 mmol) also in 10 mL of acetone. The solution immediately turned dark brown. The resulting mixture was stirred for 2 h at 0 °C. Solvent was removed, and the solid residue was extracted with CH₂Cl₂. After filtering through glass wool, this extract was evaporated, leaving brown-black solid product. Yield: 110 mg (91%). IR (CH₂Cl₂): v_CO = 1949 cm⁻¹. ESI-MS (positive, CH₂Cl₂): 1126.1201 [M]⁺, 1098.1071 [M-CO]⁺. ³¹P NMR (202 MHz, CH₂Cl₂): δ74.07 (dd, 1P, Fe-P), 71.43 (dd, 1P, Fe-P), −3.25 (dd, 1P, Ni-P). ¹H-NMR (500 MHz, CD₂Cl₂): δ7.82–6.50 (overlapping signals, 42H), 3.89 – 4.05 (m, 2 H), 3.78 (s, 3H), 3.50 (m, 1 H), 3.04 (m, 1 H), 2.49 (d, J = 9.16 Hz, 1 H), 2.04 (m, 1 H), 1.90 (m, 1 H), 1.44 (m, 1 H), 0.62 (d, J = 7.63 Hz, 1 H). Anal. Calcd for C₆₁H₅₆BCl₃F₄FeNNiO₂P₃S₂ [[(PCH₂NH^An)Ni(μ-pdt)ClFe(dppbz) (CO)]BF₄. CH₂Cl₂] (found): C, 56.37 (56.32); H, 4.34 (4.56); N, 1.08 (1.18).

[(dppv)(CO)Fe(μ-pdt)]₂Ni.

To a solution of Fe(pdt)(CO)₂(dppv) (1.02 g, 1.66 mmol, 1 equiv) in 10 mL of THF was added a solution of Ni(o-tol)Cl(tmeda) (0.50 g, 1.66 mmol, 1 equiv) in 5 mL of THF. The reaction was stirred at room temperature for 8 h. A dark brown solution developed. The solvent was removed under vacuum. The solid residue was extracted into 3 mL of CH₂Cl₂, and this extract was filtered through a 5 × 15 cm column of silica gel using additional CH₂Cl₂ as an eluent. The brown-colored band was collected and reduced in volume. Addition of 60 mL of pentane to the concentrated CH₂Cl₂ solution precipitated [(dppv)(CO)Fe(μ-pdt)]₂Ni as a brown solid. Yield: 0.362 g (35%). Vapor diffusion of pentane into a concentrated solution of [(dppv)(CO)Fe(μ-pdt)]₂Ni in CH₂Cl₂ gave red-orange prisms suitable for X-ray crystallography. ¹H NMR (500 MHz, C₆D₆): δ 8.34 (br s, 8H), 7.74–7.55 (m, 12H), 7.32 (br s, 8H), 7.11 (m, 4H), 7.03 (m, 8H), 6.97 (m, 4H), 2.16–1.77 (m, 8H), 1.32–1.07 (m, 4H). ³¹P{1H} NMR (202 MHz, C 6D₆): δ100.19 (br s, 2P), 81.51 (br s, 2P). IR (THF): v_CO = 1914 (s) cm⁻¹. Anal. Calcd for C₆₀H₅₈Fe₂NiO₂P₄S₄: C, 58.42; H, 4.74. Found: C, 57.65; H, 4.44.

{H[(dppv)(CO)Fe(μ-pdt)]₂Ni]}BF₄.

A solution of [(dppv)(CO)Fe(μ-pdt)]₂Ni (200 mg, 1.62 × 10⁻⁴ mmol, 1 equiv) in 4 mL of CH₂Cl₂ was treated with HBF₄•Et₂O (22 μL, 26 mg, 1.62 × 10⁻⁴ mol, 1 equiv). The mixture was stirred at room temperature for 30 min. A dark red solution developed. The solvent was removed under vacuum. The solid residue was washed with dry Et₂O (3 × 5 mL) and dried under vacuum. Crystals of this were grown by vapor diffusion of pentane into CH₂Cl₂ solutions of the salt at room temperature. Yield: 152 mg (71%). ¹H NMR (500 MHz, CD₂Cl₂, basal-basal isomer, −60 °C): δ 8.27 (ddd, J = 85, 45, 12 Hz, 4H), 7.73 (s, 4H), 7.32 (br s, 8H), 7.64–7.05 (m, 36H), 2.52–1.22 (overlapping signals, 6H), 1.10 (s, 2H), 0.31 (br s, 1H), 0.25 (br s, 2H), −0.28 (br s, 1H), −5.10 (br s, 1H). ¹H NMR (500 MHz, CD₂Cl₂, mixture of isomers, 20 °C): δ 8.56–7.05 (overlapping signals, 40H), 2.48 (m, 0.74 H), 2.04 (t, J = 13 Hz, 0.91H). 1.97 (m, 0.44H), 1.75 (m, 2.75H), 1.59 (m, 1.31H), 1.27 (m, 0.8H), 1.15 (m, 0.75H), 1.08 (m, 1.36H), 0.82 (m, 1.63H), 0.51 (m, 0.49H), 0.27 (m, 0.77H), −0.15 (q, J = 13 Hz), −4.87 (t, J = 38 Hz, 0.15H), −6.67 (br s, 0.23H), −15.50 (s, 0.36H). ³¹P{1H} NMR (202 MHz, CD₂Cl₂, 20 °C): δ95.23 (br s, 2P, apical-basal isomer), 91.49 (br s, 2P, apical-basal isomer), 87.43 (s, 4P, Ni-H), 86.72 (s, 2P, basal-basal isomer), 71.70 (s, 2P, basal-basal isomer). IR (THF): v_CO = 1965, 1943 cm⁻¹. Anal. Calcd for C₆₁H₅₉B Cl₂F₄Fe₂NiO₂P₄S₄: C, 52.17; H, 4.23. Found: C 52.09; H. 4.35 {H[(dppv)(CO)Fe(μ-pdt)]₂Ni]}BAr^F₄ was prepared similarly. ¹H NMR (500 MHz, CD₂Cl₂, mixture of isomers, 20 °C): δ 8.56–7.05 (overlapping signals, 48H), 2.48 (m, 0.74 H), 2.04 (t, J = 13 Hz, 0.91H). 1.97 (m, 0.44H), 1.75 (m, 2.75H), 1.59 (m, 1.31H), 1.27 (m, 0.8H), 1.15 (m, 0.75H), 1.08 (m, 1.36H), 0.82 (m, 1.63H), 0.51 (m, 0.49H), 0.27 (m, 0.77H), −0.15 (q, J = 13 Hz), −4.87 (t, J = 38 Hz, 0.15H), −6.67 (br s, 0.23H), −15.50 (s, 0.36H). ³¹P{1H} NMR (202 MHz, CD₂Cl₂, 20 °C): 95.23 (br s, 2P, apical-basal isomer), 91.49 (br s, 2P, apical-basal isomer), 87.43 (s, 4P, Ni-H), 86.72 (s, 2P, basal-basal isomer), 71.70 (s, 2P, basal-basal isomer). IR (THF): v_CO = 1965, 1943 cm⁻¹.

(C₆F₅)₂Ni(μ-pdt)Fe(CO)₂(dppv).

In the glove box a solution of Fe(pdt)(CO)₂(dppv) (370.0 mg, 0.602 mmol) in 30 mL of THF was added dropwise to a stirred solution of Ni(C₆F₅)₂(dioxane)₂ (342.5 mg, 0.602 mmol) in 10 mL of THF over 20 min at room temperature. The solution quickly changed from orange to a very dark red. The reaction was judged complete based on ³¹P NMR measurements. The volume of the reaction mixture was reduced to ~10 mL under vacuum. Addition of 20 mL of pentane to this filtrate produced a bright red powder. Slow diffusion of pentane into a THF solution of the complex resulted in long ruby-red needles. Yield: 480 mg (79%). ¹H NMR (500 MHz, d_8-THF, 20 °C): δ8.56–7.05 (overlapping signals, 48H), 2.48 (m, 0.74 H), 2.04 (t, J = 13 Hz, 0.91H). 1.97 (m, 0.44H), 1.75 (m, 2.75H), 1.59 (m, 1.31H), 1.27 (m, 0.8H), 1.15 (m, 0.75H), 1.08 (m, 1.36H), 0.82 (m, 1.63H), 0.51 (m, 0.49H), 0.27 (m, 0.77H), −0.15 (q, J = 13 Hz), −4.87 (t, J = 38 Hz, 0.15H), −6.67 (br s, 0.23H), −15.50 (s, 0.36H). ¹⁹F{1H} NMR (470.4 MHz, d₈-THF, 20 °C): d119.2 (d, J= 30.8 Hz, 2F), 120.8 (d, J= 30.8 Hz, 2F), 169.75 (t, J= 20.2 Hz, 2F), 171.0 (br, 2F), 171.2 (br, 2F). ³¹P{1H} NMR (202 MHz, d₈-THF, 20 °C): δ70.5 (s, 2P). IR (THF): v_CO = 2059, 1990 cm⁻¹. Anal. Calcd for C₄₃H₂₈F₁₀FeNiO₂P₂S₂: C, 51.27; H, 2.80. Found: C 50.95; H. 3.00.

X-ray Crystallographic Determinations.

Crystallographic Data for Ph₄P[Ni(NS₃)], [Cp(CO)Fe(μ-pdt)Ni(dppbz)]BF₄, [(R,R-DIPAMP)Ni(μ-pdt)(H)Fe(CO)₃]BAr^F₄, [NiCl₂(PCH=N^An)]₂, [NiCl₂(PCH₂NH^An)]_2, [(dppv)(CO)Fe(μ-pdt)]₂Ni, {H[(dppv)(CO)Fe(μ-pdt)]₂Ni]}BF₄, and (C₆F₅)₂Ni(μ-pdt)Fe(CO)₂(dppv) were collected on a Bruker D8 Venture instrument equipped with a four-circle kappa diffractometer and Photon 100 detector. An Iμs microfocus Mo (λ = 0.71073 Å) source was supplied the multi-mirror monochromated incident beam. The samples were mounted on a 0.3 mm loop with the minimal amount of Paratone-N oil. Data were collected as a series of φ and/or ω scans. Data were collected at 100K and integrated and filtered for statistical outliers using SAINT,⁶⁴ and corrected for absorption by multi-scan methods using SADABS⁶⁵ v2014/7. The structures were phased using direct methods or intrinsic phasing methods and then refined with the SHELX software package SHELX-2014–7.⁶⁶

Density Functional Theory (DFT).

Computations were carried out according to the TZVP/B97-D scheme, as implemented in the TURBOMOLE 7.2 programs suite.^67–69 The pure B97-D functional, which intrinsically accounts for dispersive effects, has been chosen since it proved to reliably reproduce both geometries and reactive behavior of hydrogenase-like models.⁷⁰ The Resolution-of-the-Identity (RI) technique has been used to reduce the computational costs deriving by the treatment of the classical electron-electron repulsive contribution to the total energy.⁷¹ Both low (S = 0) and high (S = 1) spin forms of [(dppv)(CO)Fe(μ-pdt)]₂Ni have been characterized. With the attempt of localizing opposite spins of the monodeterminant wave function in different parts of the molecule, the Broken Symmetry approximation has been adopted, to detect (eventual) antiferromagnetic spin couplings.⁷² Nuclear magnetic shielding tensors for ¹H NMR chemical shifts were computed according to the Gauge Including Atomic Orbital (GIAO) method^73,74 on TZVP/BP86 geometries, a very popular scheme for the reproduction of spectroscopic and stereoelectronic features of transition metal complexes relevant to hydrogenases.^63,75,76

Figure 11.

¹H-NMR spectra of {H[(dppv)(CO)Fe(μ-pdt)]₂Ni]}BF₄ (CD₂Cl₂ solution) providing evidence for three isomers.