Nickel-Iron Dithiolates Related to the Deactivated [NiFe]-Hydrogenases

Abstract

Described herein are preparations of synthetic models for the deactivated Ni(II)Fe(II) states of the [NiFe]-hydrogenases. Iodination of the S = ½ species [(dppe)Ni(pdt)Fe(CO)₃]⁺ afforded the diamagnetic iodo complex [(dppe)Ni(pdt)IFe(CO)₃]⁺. Crystallographic analysis of this species confirmed the presence of square-pyramidal Ni linked to an octahedral Fe centre. The Ni···Fe separation of 3.018 Å indicated the absence of metal-metal bonding. This complex could be reduced to give (dppe)Ni(pdt)Fe(CO)₃ and, in the presence of iodide, decarbonylated to afford (dppe)Ni(pdt)FeI₂. Derivatives of the type [(diphosphine)Ni(dithiolate)XFe(CO)₂L]⁺ (X = Cl, Br, I) were prepared by halogenation of mixed-valence precursors [(diphosphine)Ni(dithiolate)Fe(CO)₂L]⁺ (diphosphine = dppe, dcpe; L = tertiary phosphine or CO). The Fe(CO)₂(PR₃)-containing derivatives are more robust than the related tricarbonyl derivatives. Exploiting this greater stability, we characterised examples of chloride and bromide derivatives. Related fluorides could be prepared by F⁻ abstraction from BF₄⁻. Spectroscopic evidence is presented for the hydroperoxide [(diphosphine)Ni(dithiolate)(OOH)Fe(CO)₂L]⁺, which was prepared by oxidation of a model for Ni-SU and Ni-SI_r states.

Introduction

Hydrogen processing in Nature is mediated by the hydrogenases (H₂ases), a class of enzyme prevalent in anaerobic bacteria.¹ A major subclass of these enzymes, the nickel-iron hydrogenases, is expressed to catalyse the redox reaction 2H⁺ + 2e⁻ ⇌H₂, which is central to the metabolism of certain bacteria and archaea.² The [NiFe]-H₂ase active site features a Ni(cysteinato)₄ centre, two thiolate ligands of which bridge to a Fe(CN)₂(CO) fragment (Figure 1). The states of the enzyme are classified according to the identity (or absence) of the bridging ligand X, the oxidation states of the metal centres, and the protonation of two of the four thiolate ligands. Of relevance to this paper, aerobic deactivation of the enzyme affords states in which oxygenic ligands bridge the two metal ions.³ For example, a Ni(II)(OH)Fe(II) core is observed for (Ni-SI_r)_I,⁴ whereas in the Ni-SU state the divalent metal ions are thought to be bridged by a hydroperoxo ligand.⁵ Our attention was thus drawn to the rarity of model compounds featuring Fe(μ-SR)₂(μ-X)Ni cores.⁶

Figure 1.

Schematic of the [NiFe]-H₂ase active site (left) and some of the enzyme states (right).

Much recent work has focused on synthetic modelling of the [NiFe]-H₂ase active sites^,7 no doubt motivated by the increased understanding of the natural systems and hydrogen processing catalysts that such work promises.⁸ The Ni(I)Fe(I) models of type (diphosphine)Ni(dithiolate)Fe(CO)₃ are of particular interest,9 as their conjugate acids, the only synthetic nickel-iron dithiolate hydrides, represent structural and functional models for Ni-R.¹⁰ We have reported a promising procedure for the neutral complexes from the precursors (diphosphine)Ni(dithiolate) and cis-[FeI₂(CO)₄].¹¹ The Ni(II)Fe(II) intermediate thus generated, tentatively formulated as [(diphosphine)Ni(dithiolate)IFe(CO)₃]I, is reduced with CoCp₂ to afford the targeted Ni(I)Fe(I) complex. Attempts to isolate the metastable intermediate have been unfruitful, as exemplified by the crystallization of the CO-free species (dcpe)Ni(2,2-dimethyl-1,3-propanedithiolate)FeI₂ (dcpe = 1,2-bis(dicyclohexylphosphino)ethane), thought to arise from nucleophilic attack of I⁻ and concomitant decarbonylation.¹¹ According to IR spectroscopic data, solutions of the putative iodide salt can also be prepared by treatment of (diphosphine)Ni(dithiolate)Fe(CO)₃ with I₂, the ensuing heterolysis resulting in the bridging iodo species as its iodide salt. Presented here are syntheses and characterization of bridging halide complexes of type [(diphosphine)Ni(pdt)XFe(CO)₃]⁺ (pdt²⁻ = 1,3-propanedithiolate) as well as substituted derivatives [(diphosphine)Ni(pdt)XFe(CO)₂(monophosphine)]⁺. The work illustrates the radical character of [(diphosphine)Ni(pdt)Fe(CO)₃]⁺ and its participation in atom transfer reactions.

Results

Tricarbonyl Halides [(diphosphine)Ni(pdt)XFe(CO)₃]⁺

As mentioned in the Introduction, halide salts of type [(diphosphine)Ni(pdt)XFe(CO)₃]X are unstable towards decarbonylation. In CH₂Cl₂ solution, the tricarbonyls [(dppe)Ni(pdt)IFe(CO)₃]Br([1I]I, dppe = 1,2- bis(diphenylphosphino)ethane) and [(dppe)Ni(pdt)BrFe(CO)₃]Br [1Br]Br were found to convert to (dppe)Ni(pdt)FeI₂ and (dppe)Ni(pdt)FeBr₂, respectively. IR analysis of these paramagnetic materials indicated the absence of ν_CO bands, with elemental analyses also being consistent with these formulations (see Experimental section). Consequently, our focus shifted to salts of [(diphosphine)Ni(pdt)XFe(CO)₃]⁺ with non-nucleophilic anions. Oxidation of (diphosphine)Ni(pdt)Fe(CO)₃ with ferrocenium tetrafluoroborate (FcBF₄) furnishes the paramagnetic Ni(II)Fe(I) derivatives [(diphosphine)Ni(pdt)Fe(CO)₃]BF₄.¹² The salt [(dppe)Ni(pdt)Fe(CO)₃]BF₄ ([1]BF₄) reacts with I₂ (0.5 equiv.) in CH₂Cl₂, resulting in a colour change from dark brown to red. The solutions thus formed have ν_CO bands identical to those of [(diphosphine)Ni(dithiolate)Fe(CO)₃]/I₂ mixtures. Notably, the stretching frequencies (ν_CO = 2097, 2055, 2026 cm⁻¹) are higher than those for the hydride [1H]BF₄ (ν_CO = 2082, 2024 cm⁻¹), reflecting the poorer σ-donating ability of the iodo ligand vs the hydride ligand (Figure 2). Furthermore, three ν_CO bands are easily resolved in the IR spectra for these halides, whereas the hydrides exhibit an A-E pattern in the ν_CO region typical of a more symmetrical coordination environment.

Figure 2.

FT-IR spectrum (ν_CO region, CH₂Cl₂) of [1I]BF₄ (top) and [2I]BF₄ (bottom).

The new compound was obtained as an orange powder, analytical data for which were consistent with its formulation as the diamagnetic salt [(dppe)Ni(pdt)IFe(CO)₃]BF₄ ([1I]BF₄). The complex cation could be identified by ESI-MS (m/z 828.9) and ³¹P NMR spectroscopy (δ 49.2), the latter confirming its high symmetry. Analogously, iodination of the dcpe-containing species [(dcpe)Ni(pdt)Fe(CO)₃]BF₄ ([2]BF₄)¹³ afforded the salt [(dcpe)Ni(pdt)IFe(CO)₃]BF₄ ([2I]BF₄). The transformations are summarised in Scheme 1.

Scheme 1.

Reactions of Ni(I)Fe(I), Ni(II)Fe(I) and Ni(II)XFe(II) complexes.

Structure of (diphosphine)Ni(pdt)IFe(CO)₃]⁺

The solid state structure of [1I]BF₄ was determined by X-ray crystallography; an ORTEP of the cation is presented in Figure 3. As expected, the complex features Fe and Ni centres bridged by dithiolate and iodo ligands. The intermetallic separation (3.018 Å) is consistent with the absence of metal-metal bonding, which is expected for a Ni(II)Fe(II) system. The Fe coordination environment is approximately octahedral, with the Ni centre being best described as square-pyramidal, although the long Ni···I distance (2.931 Å) suggests that this apical ligand interacts only weakly with Ni. The iodo ligand is indeed nearer to the Fe centre (2.682 Å) and the situation is similar in the hydride [1H]⁺ (Fe–H, 1.460 Å; Ni–H, 1.637 Å), with which [1I]⁺ is almost isostructural. The related chloro complex [(dppe)Ni(pdt)ClFe(CO)(dppe)]⁺ has a similar Ni–Fe distance (3.076 Å), with the Cl⁻ ligand being much closer to the metal centres (Fe–Cl, 2.362 Å; Ni–Cl, 2.668Å) than I⁻ is in [1I]^+.10 Notably, [(dppe)Ni(pdt)ClFe(CO)(dppe)]⁺ is prepared from the divalent metal precursors (dppe)NiCl₂ and (pdt)Fe(CO)₂(dppe), whereas [1I]⁺ is formed by a pathway probably involving I₂ homolysis.

Figure 3.

ORTEP of [1I]BF₄·2CH₂Cl₂ with ellipsoids drawn at the 50% probability level. The H atoms, disordered CH₂Cl₂ solvate molecules and BF₄⁻ anion are omitted for clarity. Selected distances (Å): Ni1–Fe1, 3.018; Ni1–P1, 2.189; Ni1–P2, 2.187; Ni1–S1, 2.240; Ni1–S2, 2.227; Ni1–I1, 2.931; Fe1–S1, 2.318; Fe1–S2, 2.311; Fe1–I1, 2.682; Fe1–C30, 1.815; Fe1–C31, 1.775; Fe1–C32, 1.836. Selected angles (°): Ni1–I1–Fe1, 64.9.

Reactivity of [(diphosphine)Ni(pdt)XFe(CO)₃]⁺

Solid [1I]BF₄ is air-stable, although in CH₂Cl₂ solution it decomposes over the course of hours even in the absence of air. The ν_CO bands for [1I]BF₄ rapidly disappear when CH₂Cl₂ solutions of the compound are treated with NBu₄I (1 equiv.), consistent with the formation of the CO-free product (dppe)Ni(pdt)FeI₂. This halide-induced decarbonylation is consistent with the instability of the iodide salt [1I]I (vide supra) and is depicted in Scheme 2. Bromination of [1]BF₄ using Br₂ (0.5 equiv.) or N-bromosuccinimide (1 equiv.) affords the bromide derivative [(dppe)Ni(pdt)BrFe(CO)₃]BF₄ ([1Br]BF₄). However, in comparison to the iodide complex, the bromide is less stable, decomposition to insoluble CO-free materials being apparent within minutes in CH₂Cl₂. Also unstable is the chloro species [(dppe)Ni(pdt)ClFe(CO)₃]BF₄ ([1Cl]BF₄), which was prepared from [1]BF₄ and PhICl₂ (0.5 equiv.). This trend in the stability of the halide complexes mirrors that observed for the FeX₂(CO)₄ compounds (X = I, Br, Cl).¹⁴

Scheme 2.

Reactions of the μ-iodo cation [1I]⁺.

Conversion of [1Br]BF₄ to [1I]BF₄ (as well as the decomposition product) proceeds in the presence of I₂, reflecting the better donor ability of I⁻ versus Br⁻. Similarly, the hydride [1H]BF₄ is completely consumed upon treatment with excess I₂, [1I]BF₄ being the only CO-containing product. A CO-free byproduct is also likely given the lower intensity of the CO bands and the detection of free CO in the headspace. Like [(dppe)Ni(pdt)ClFe(CO)(dppe)]⁺, [1Br]⁺ and [1I]⁺ are unreactive toward Ag⁺ salts, behaviour that is often characteristic of bridging halide ligands.¹⁵ The bromo complex [1Br]⁺ also forms from [1]⁺ upon abstraction of Br^· from benzyl bromide.

Dicarbonyl-Phosphine Iodides [(diphosphine)Ni(pdt)IFe(CO)₂(PR₃)]⁺

Given the Fe(CN)₂(CO) fragment present in the enzyme, it was of interest to prepare derivatives in which of one or more of the carbonyl ligands was replaced by stronger σ-donor ligands. Similar approaches were previously used by us to obtain the mixed-valence species [(diphosphine)Ni(pdt)Fe(CO)₂(PR₃)]⁺, and the hydrides [(diphosphine)Ni(pdt)HFe(CO)₂(PR₃)]⁺, models for the Ni-L¹² and Ni-R states,¹⁶ respectively. The tricarbonyl iodide [1I]⁺ reacts only slowly with tertiary phosphines. A more efficient route to the derivatives [(diphosphine)Ni(pdt)IFe(CO)₂(PR₃)]⁺ involves radical iodination of the substituted Ni-L models. For example, treatment of [(dppe)Ni(pdt)Fe(CO)₂(PPh₃)]BF₄ (1a]BF₄)¹² with I₂ (0.5 equiv.) in CH₂Cl₂ afforded [(dppe)Ni(pdt)IFe(CO)₂(PPh₃)]BF₄ ([1aI]BF₄). The phosphine-substituted iodide exhibits enhanced stability relative to its unsubstituted parent [1I]BF₄. The product was characterised according to FT-IR, NMR, ESI-MS and analytical data. In particular, ³¹P NMR spectra of [1aI]BF₄ in CD₂Cl₂ feature three resonances, consistent with a basal orientation of the PPh₃ ligand. This stereochemistry is also found in the hydride [1aH]BF₄,¹⁶ but it contrasts that of the precursor [1a]BF₄, in which the monophosphine is apical.¹² The PCy₃-iodo complex salt [(dppe)Ni(pdt)IFe(CO)₂(PCy₃)]BF₄ ([1bI]BF₄) could be prepared in a similar fashion; selected data for new compounds are presented below (Table 1). Cationic diiron dithiolato halides, exemplified by [Fe₂(pdt)Cl(CO)₆]⁺,¹⁵ are also stabilised by phosphine ligands.¹⁷

Table 1.

IR data for complexes of type [(dxpe)Ni(pdt)XFe(CO)₂L]BF₄ in CH₂Cl₂ solution.

Species	dxpe	X	L	νCO/cm−1	Ered
[1I]BF4	dppe	I	CO	2097, 2055, 2026	−0.99 (irrev.)
[1Br]BF4	dppe	Br	CO	2106, 2065, 2028
[1Cl]BF4	dppe	Cl	CO	2111, 2072, 2031	−1.00 (irrev.)
[1aI]BF4	dppe	I	PPh3	2031, 1978
[1bI]BF4	dppe	I	PCy3	2024, 1965	−1.43 (irrev.)
[1aCl]BF4	dppe	Cl	PPh3	2042, 1981	−1.34 (irrev.)
[1bCl]BF4	dppe	Cl	Pcy3	2033, 2068	−1.44 (irrev.)
[2I]BF4	dcpe	I	CO	2096, 2054, 2023	−0.86 (ipc/ipa = 1.87)

The well-resolved ³¹P NMR spectrum indicates that [1aI]BF₄ is stereochemically rigid. In contrast, the corresponding hydride is dynamic, resulting from rotation at the Ni(diphosphine) centre. This difference is consistent with Fe-I forming a stronger bond to Ni than does Fe-H.

Cyclic Voltammetry Studies

The salts [1I]BF₄, [1bI]BF₄ and [2I]BF₄ were studied by cyclic voltammetry. Whereas the oxidation waves for the compounds are complicated, the reductions are more obvious. In the case of [1I]BF₄, irreversible reduction occurs at −0.99 V, whereas for [2I]BF₄ the process is partially reversible and occurs at −0.86 V. The hydride analogues [1H]BF₄ and [2H]BF₄ are irreversibly reduced at −1.34 and −1.56 V, respectively.¹¹ Overall, trends in reduction potentials are consistent with the stronger donor ability of H⁻ versus I⁻. The ease of reduction of the tricarbonyls [1I]⁺ and [2I]⁺ is consistent with their reduction by cobaltocene (CoCp₂), exploited by our group to prepare the neutral species 1 and 2 (Scheme 2).

The PCy₃-substituted derivative [1bI]BF₄ is reduced at −1.43 V, considerably more cathodic that the tricarbonyl parent [1I]BF₄. Similar shifts (~ −0.4 V) are observed for the mixed-valence salt [1b]BF₄ relative to [1]BF₄.¹³ As evidenced by ³¹P NMR spectroscopy, chemical reduction of [1I]BF₄ and [2I]BF₄ with CoCp₂ (2 equiv.) affords the expected crude mixtures containing the neutral species 1 (δ 63.0) and 2 (δ 80.9 ppm) respectively.

Dicarbonyl-Phosphine Fluorides [(diphosphine)Ni(pdt)FFe(CO)₂(PR₃)]⁺

Attempts were made to oxidise the substituted mixed-valence derivatives [1a]⁺ and [1b]⁺ to the 32e⁻ clusters [1a]²⁺ and [1b]²⁺, these latter species representing models for the highly electrophilic Ni-SI_a state responsible for H₂ heterolysis. However, treatment of [1a]BF₄ with FcBF₄ in CH₂Cl₂ instead afforded the μ-fluorido species [(dppe)Ni(pdt)FFe(CO)₂(PPh₃)]BF₄ ([1aF]BF₄), presumably from the abstraction of F⁻ from BF₄⁻. FT-IR analysis revealed two sets of overlapping ν_CO bands (2063, 2040, 1984, 1977 cm⁻¹), consistent with the presence of two isomers (not uncommon for such pdt²⁻-bridged systems).¹² The mean ν_CO frequency (2016 cm⁻¹) is slightly higher than that for [1aCl]BF₄ (2011 cm⁻¹), which is in line with the poorer donor ability of the lighter halido ligands. The identity of [1aF]BF₄ was further confirmed by ESI-MS (m/z 955.3) and analytical data. The PCy₃ congener [1bF]BF₄ could be prepared in a similar fashion from [1b]BF₄.

Attempted Synthesis of Hydroxo- and Hydroperoxo-Bridged Complexes

Several sets of experiments were conducted in an effort to install bridging oxygenic ligands, as seen in the Ni-SI and Ni-SU states of the enzyme. The mixed-valence Ni(II)Fe(I) complexes, e.g. [1]⁺ react with water but not in a rationalizable manner, and no hydroxides could be detected. Thus treatment of a CH₂Cl₂ solution of [1]BF₄ with water afforded 1 in ~50% yield according to IR analysis. Sensitive qualitative tests failed to indicate the formation of hydrogen peroxide in the reaction solutions.

Attempts were also made to install peroxo bridging groups. Treatment of a CH₂Cl₂ solution of the hydride [1aH]BF₄ with aqueous H₂O₂ resulting in a colour change from brown to red. The proposed stoichiometry is shown in Scheme 3.

Scheme 3.

Formation of a hydroperoxo complex from a hydride precursor.

Although we were unable to purify the product, ESI-MS analysis of the red solution allowed for the detection of [(dppe)Ni(pdt)(OOH)Fe(CO)₂PPh₃]BF₄ ([1aOOH]⁺, m/z 971.3). The IR spectrum exhibited two ν_CO bands (2042, 1981 cm⁻¹) at energies similar to those of the iodo salt [1aI]BF₄ (2031, 1978 cm⁻¹).

Discussion

Complexes of the type (diphosphine)Ni(dithiolate)Fe(CO)₂L represent the most versatile models for the active site of the [NiFe]-hydrogenases, despite their abiological terminal ligation. In previous reports, we have described models for the hydride-containing state Ni-R⁹ and the S = ½ state Ni-L.¹² This paper summarises efforts to prepare synthetic models for the Ni-SU and Ni-SI_r states. These enzyme states feature Ni(II)Fe(II) centres bridged by hydroxide and hydroperoxide.¹⁸

As for other complexes in this series of models, substitution of one CO ligand on Fe by a phosphine ligand enhances the stability of these compounds considerably. The halide-bridged species are not only important precursors to functional models¹⁹ but also represent the closest approach to synthetic analogues of these inactive states. Halide complexes incorporating NiFe nuclei have also been prepared by Bouwman and co-workers.²⁰ Our attempts to purify hydroxide-, hydroperoxide- and alkoxide-bridged NiFe models have been unsuccessful, although we did detect a hydroperoxide derivative in one case. Derivatives incorporating Ni(II)(OR)(OH)Fe(II), Ni(II)(O)₂Fe(II)²¹ and Ni(II)(OH)Fe(II) cores²² have been reported by the Driess and Holm groups, respectively. Metal carbonyl complexes containing hydroperoxide ligands are rare, probably owing to the facility of intramolecular reactions involving CO and the peroxide. As highlighted by Stiebritz and Reiher,²³ the importance of studying O₂-induced hydrogenase inhibition cannot be understated, as robust enzyme variants (and their models) represent promising catalysts for hydrogen processing.

Experimental

Unless otherwise stated, all chemicals were purchased from commercial sources and used as received. N-Bromosuccinimide was recrystallized from H₂O. CD₂Cl₂ was distilled from CaH₂. PhICl₂,²⁴ 1²⁵ and 2¹¹ were prepared according to the literature methods. All reactions were conducted in an MBraun glovebox equipped with a solvent purification system; the concentrations of O₂ and H₂O in the N₂ atmosphere were typically 1 and 0.2 ppm, respectively. IR spectra of complexes (in CH₂Cl₂) were recorded on a Perkin-Elmer Spectrum 100 FTIR spectrometer. A Micromass Quattro II spectrometer was used to acquire ESI-MS data for analytes as dilute CH₂Cl₂ solutions. Analytical data were acquired using an Exeter Analytical CE-440 elemental analyser. NMR spectra were recorded at room temperature on a Varian Unity 500 spectrometer. ¹H and ³¹P{¹H} spectra were collected at 500 and 202 MHz, respectively, and chemical shifts are referenced to residual CHDCl₂ (5.32 ppm vs TMS for ¹H) and external 85% H₃PO₄ (for 31P). Cyclic voltammetry experiments were carried out in a single compartment glass cell using a CH Instruments CHI630D electrochemical analyser. The working, counter and pseudo-reference electrodes were glassy carbon, platinum and silver, respectively. The analyte (1 mM) and NBu₄PF₆ (100 mM) were dissolved in CH₂Cl₂, and potentials (reported here relative to internal Fc/Fc⁺) were swept at 0.1 Vs⁻¹. Crystallographic data were collected using a Bruker X8 diffractometer equipped with a Cu K_α source (λ = 1.54178 Å) and an Apex II detector.

(dppe)Ni(pdt)FeI₂

Compound 1 (21.1 mg, 30 μmol) was partially dissolved in CH₂Cl₂ (0.5 mL) and treated with I₂ (3.8 mg, 15 μmol) in CH₂Cl₂ (1 mL). The solution was stirred for a further 3 days, after which IR analysis indicated complete decarbonylation. Pentane (−28 °C, 15 mL) was added and the mixture allowed to stand at −28 °C for 1 h. The solids were isolated by filtration, washed with pentane (−28 °C, 2 × 2 mL) and dried briefly to afford the title compound as a red-brown powder (24.6 mg, 28 μmol, 94%).

Anal. calcd for C₂₉H₃₀FeI₂NiP₂S₂: C, 39.90; H, 3.46; N, 0.00. Found: C, 40.11; H, 3.47; N, 0.50.

(dppe)Ni(pdt)FeBr₂

This compound was prepared analogously to the iodo derivative, using Br₂ in place of I₂. Yield: 91%, red-brown powder.

Anal. calcd for C₂₉H₃₀FeBr₂NiP₂S₂: C, 44.71; H, 3.88; N, 0.00. Found: C, 44.68; H, 3.98; N, 0.49.

[(dppe)Ni(pdt)IFe(CO)₃]BF₄ ([1I]BF₄)

Compound 1 (21.1 mg, 30 μmol) and FcBF₄ (8.2 mg, 30 μmol) were dissolved in CH₂Cl₂ (1 mL) with rapid stirring. After 1 min the solution was treated with I₂ (3.8 mg, 15 μmol) in CH₂Cl₂ (1 mL). The solution was stirred for a further 0.5 min and pentane (−28 °C, 15 mL) was added and the mixture allowed to stand at −28 °C for 1 h. The solids were isolated by filtration, washed with pentane (−28 °C, 2 × 2 mL) and dried briefly to afford the title compound as an orange powder (26.9 mg, 29 μmol, 98%). ³¹P NMR (CD₂Cl₂, 202 MHz): δ 49.2. Anal. calcd for C₃₂H₃₀O₃S₂P₂NiFeIBF₄·CH₂Cl₂: C, 39.56; H, 3.22; N, 0.00. Found: C, 39.71; H, 3.09; N, 0.00. ESI-MS: m/z 828.9 [M – BF₄⁻]⁺. Orange prismatic single crystals of [1I]BF₄·2CH₂Cl₂ formed upon slow diffusion of pentane vapour into a concentrated CH₂Cl₂ solution of the title compound at −28 °C. One crystal (0.366 × 0.102 × 0.098 mm³) was subjected to X-ray diffraction at 193 K. Its space group was determined to be monoclinic C2/c with cell parameters: a 22.916Å, b 17.958 Å, c 21.500 Å, α 90.00°, β 107.87°, γ 90.00°. Integration of 5113 reflections and solution by direct methods using SHELXTL V6.12²⁶ afforded a model with R1 = 0.0548 and wR2 = 0.0920.

[(dppe)Ni(pdt)BrFe(CO)₃]BF₄ ([1Br]BF₄)

This compound was prepared analogously to [1I]BF₄, using Br₂ in place of I₂. Yield: 98%, orange powder. 31P{1H} NMR (CD₂Cl₂, 202 MHz): δ 59.6 (~18% impurity), 53.4. Anal. calcd for C₃₂H₃₀O₃S₂P₂NiFeBrBF₄: C, 44.18; H, 3.48; N, 0.00. Found: C, 44.45; H, 3.40; N, 0.00. ESI-MS: m/z 782.8 [M – BF₄⁻]⁺. N-Bromosuccinimide (5.3 mg, 30 μmol) can be used in place of Br₂. Treatment with CoCp₂ (2 equiv.) in CH₂Cl₂ gave a green solution and IR bands at 2028 and 1952 cm⁻¹, consistent with formation of the neutral species.

[(dppe)Ni(pdt)ClFe(CO)₃]BF₄ ([1Cl]BF₄)

This compound was prepared analogously to [1I]BF₄, using PhICl₂ in place of I₂. Yield: 93%, orange powder.

³¹P{¹H} NMR (CD₂Cl₂, 202 MHz): δ 54.5. Anal. calcd for C₃₂H₃₀O₃S₂P₂NiFeClBF₄: C, 46.56; H, 3.66; N, 0.00. Found: C, 46.81; H, 3.52; N, 0.00. ESI-MS: m/z 736.8 [M – BF₄⁻]⁺.

[(dppe)Ni(pdt)IFe(CO)₂(PR₃)]BF₄

Compound 1 (14.1 mg, 20 μmol) and FcBF₄ (5.5 mg, 20 μmol) were dissolved in CH₂Cl₂ (1 mL) with rapid stirring. After 1 min the solution was added dropwise to PR₃ (25 μmol) in CH₂Cl₂ (0.5 mL). The solution was stirred for a further 5 min before being treated with I₂ (2.5 mg, 10 μmol) in CH₂Cl₂ (1 mL) with rapid stirring. After 5 min, pentane (−28 °C, 15 mL) was added and the mixture allowed to stand at −28 °C for 1 h. The solids were isolated by filtration, washed with pentane (−28 °C, 2 × 2 mL) and dried briefly to afford the product.

[(dppe)Ni(pdt)IFe(CO)₂(PPh₃)]BF₄ ([1aI]BF₄)

Yield: 72%, orange powder. ³¹P NMR (CD₂Cl₂, 202 MHz): δ 48.6 (dd, J_PP = 36, 30 Hz), 45.1 (d, J_PP = 36 Hz), 42.6 (d, J_PP = 30 Hz). Anal. calcd for C₄₉H₄₅BF₄FeINiO₂P₃S₂·0.3CH₂Cl₂: C, 50.32; H, 3.91; N, 0.00. Found: C, 50.22; H, 3.91; N, 0.00. ESI-MS: m/z 1063.0 [M – BF₄⁻]⁺.

[(dppe)Ni(pdt)IFe(CO)₂(PCy₃)]BF₄ ([1bI]BF₄)

Yield: 68%, orange powder. ³¹P{¹H} NMR (CD₂Cl₂, 202 MHz): δ 44.9 (dd, J_PP = 36, 24 Hz), 44.2 (d, J_PP = 36 Hz), 42.8 (d, J_PP = 24Hz). Anal. calcd for C₄₉H₆₃BF₄FeINiO₂P₃S₂: C, 50.33; H, 5.43; N, 0.00. Found: C, 50.09; H, 5.54; N, 0.00. ESI-MS: m/z 1081.2 [M – BF₄⁻]⁺.

[(dppe)Ni(pdt)ClFe(CO)₂(PR₃)]BF₄

Compound 1 (14.1 mg, 20 μmol) and FcBF₄ (5.5 mg, 20 μmol) were dissolved in CH₂Cl₂ (1 mL) with rapid stirring. After 1 min the solution was added dropwise to PR₃ (25 μmol) in CH₂Cl₂ (0.5 mL). The solution was stirred for a further 5 min before being treated with PhICl₂ (4.1 mg, 15 μmol) in CH₂Cl₂ (1 mL) with rapid stirring. After 20 min, pentane (−28 °C, 15 mL) was added and the mixture allowed to stand at −28 °C for 1 h. The solids were isolated by filtration, washed with pentane (−28 °C, 2 × 2 mL) and dried briefly to afford the products.

[(dppe)Ni(pdt)ClFe(CO)₂(PPh₃)]BF₄ ([1aCl]BF₄)

Yield: 95%, orange powder. ³¹P NMR (CD₂Cl₂, 202 MHz): δ 53.1 (m), 44.7 (m) ppm. Anal. calcd for C₄₉H₄₅BF₄FeClNiO₂P₃S₂: C, 55.54; H, 4.34; N, 0.00. Found: C, 55.59; H, 4.49; N, 0.00. ESI-MS: m/z 1054.9 [M + CH₂Cl₂ –BF₄⁻]⁺, 970.9 [M – BF₄⁻]⁺.

[(dppe)Ni(pdt)ClFe(CO)₂(PCy₃)]BF₄ ([1bCl]BF₄)

Yield: 92%, orange powder. ³¹P NMR (CD₂Cl₂, 202 MHz): δ 53.7 (m), 51.6 (m) ppm. Anal. calcd for C₄₉H₆₃BF₄FeClNiO₂P₃S₂: C, 54.60; H, 5.89; N, 0.00. Found: C, 54.48; H, 5.99; N, 0.00. ESI-MS: m/z 989.2 [M – BF₄⁻]⁺.

[(dppe)Ni(pdt)FFe(CO)₂(PPh₃)]BF₄ ([1aF]BF₄)

[1a]BF₄ (21.2 mg, 20 μmol) in CH₂Cl₂ (1 mL) was treated with FcBF₄ (5.5 mg, 20 μmol) in CH₂Cl₂ (1 mL) with stirring. After 5 min, pentane (−28 °C, 15 mL) was added and the mixture allowed to stand at −28 °C for 1 h. The solids were isolated by filtration, washed with pentane (2 × 2 mL) and dried briefly to afford the title compound as an orange powder (18.9 mg, 18 μmol, 90%). FT-IR: ν_CO = 2063, 2040, 1984, 1977 cm⁻¹. Anal. calcd for C₄₉H₄₅BF₅FeNiO₂P₃S₂·CH₂Cl₂: C, 53.23; H, 4.20; N, 0.00. Found: C, 53.27; H, 4.20; N, 0.09. ESI-MS: m/z 955.3 [M –BF₄⁻]⁺.

[(dppe)Ni(pdt)FFe(CO)₂(PCy₃)]BF₄ ([1bF]BF₄)

This compound was prepared analogously to [1aF]BF₄, instead using the precursor [1b]BF₄. Yield = 86%, orange powder. FT-IR: ν_CO = 2051, 2030, 1975, 1961 cm⁻¹. Anal. calcd for C₄₉H₆₃BF₅FeNiO₂P₃S₂·1.2CH₂Cl₂: C, 51.83; H, 5.67; N, 0.00. Found: C, 51.90; H, 5.76; N, 0.09. ESI-MS: m/z 972.6 [M –BF₄⁻]⁺.

Generation of [(dppe)Ni(pdt)OOHFe(CO)₂(PPh₃)]BF₄ ([1aOOH]BF₄)

[1aH]BF₄ (20.5 mg, 20 μmol) in CH₂Cl₂ (2 mL) was treated with H₂O₂ (30% aqueous solution, 100 μL, ~900 μmol) and stirred in the absence of light. The dark brown solution became pale red, with ESI-MS and IR analyses indicating conversion to the new salt. ESI-MS: m/z 971.3 [M – BF₄⁻]⁺.

[(dcpe)Ni(pdt)IFe(CO)₃]BF₄ ([2I]BF₄)

This compound was prepared analogously to [1I]BF₄, instead using 2 as the precursor. Yield: 72%, orange powder. ³¹P NMR (CD₂Cl₂, 202 MHz) : δ 63.6. Anal. calcd for C₃₂H₅₄O₃S₂P₂NiFeIBF₄: C, 40.84; H, 5.78; N, 0.00. Found: C, p40.78; H, 5.96; N, 0.00. ESI-MS: m/z 853.0 [M – BF₄⁻]⁺.

Figure 4.

Positive ion ESI mass spectrum (top) and FT-IR spectrum (ν_CO region, CH₂Cl₂) of [1aOOH]BF₄ (bottom).