Geometrical influence on the non-biomimetic heterolytic splitting of H₂ by bio-inspired [FeFe]-hydrogenase complexes: a rare example of inverted frustrated Lewis pair based reactivity

Abstract

Despite the high levels of interest in the synthesis of bio-inspired [FeFe]-hydrogenase complexes, H₂ oxidation, which is one specific aspect of hydrogenase enzymatic activity, is not observed for most reported complexes. To attempt H–H bond cleavage, two disubstituted diiron dithiolate complexes in the form of [Fe₂(μ-pdt)L₂(CO)₄] (L: PMe₃, dmpe) have been used to play the non-biomimetic role of a Lewis base, with frustrated Lewis pairs (FLPs) formed in the presence of B(C₆F₅)₃ Lewis acid. These unprecedented FLPs, based on the bimetallic Lewis base partner, allow the heterolytic splitting of the H₂ molecule, forming a protonated diiron cation and hydrido-borate anion. The substitution, symmetrical or asymmetrical, of two phosphine ligands at the diiron dithiolate core induces a strong difference in the H₂ bond cleavage abilities, with the FLP based on the first complex being more efficient than the second. DFT investigations examined the different mechanistic pathways involving each accessible isomer and rationalized the experimental findings. One of the main DFT results highlights that the iron site acting as a Lewis base for the asymmetrical complex is the {Fe(CO)₃} subunit, which is less electron-rich than the {FeL(CO)₂} site of the symmetrical complex, diminishing the reactivity towards H₂. Calculations relating to the different mechanistic pathways revealed the presence of a terminal hydride intermediate at the apical site of a rotated {Fe(CO)₃} site, which is experimentally observed, and a semi-bridging hydride intermediate from H₂ activation at the Fe–Fe site; these are responsible for a favourable back-reaction, reducing the conversion yield observed in the case of the asymmetrical complex. The use of two equivalents of Lewis acid allows for more complete and faster H₂ bond cleavage due to the encapsulation of the hydrido-borate species by a second borane, favouring the reactivity of each FLP, in agreement with DFT calculations.

---

Introduction

Metalloenzymes have, over the years, shown high catalytic efficacy for the conversion of small and inert molecules (N₂, H₂, O₂, CO, etc.) into valuable products without requiring noble metals for their active sites.¹ Among them, [FeFe]-hydrogenase enzymes have received considerable attention.² This enzyme family demonstrates efficient catalytic activity for the reversible conversion of protons into H₂ gas, which is recognised as a key driver of future power and energy sources.² However this reactivity is only supported under limited working conditions owing to the enzymatic sensitivity to exogenous substrates/oxygen.^3–10 That is why synthetic chemists are tackling the challenging task of designing more robust catalysts using bio-inspired approaches.¹¹ A multitude of mimics of the H-cluster, the active site of [FeFe]-hydrogenases, have been developed containing two iron metals bridged by a dithiolate ligand, allowing proton reduction,^12–14 but only a handful of these complexes are able to activate H₂.¹⁵ Of the rare examples supporting the arduous process of H₂ activation, some complexes involve classical reactions, such as oxidative addition¹⁶ and reduction,¹⁷ or permit H₂/D₂ exchange.^18–21 Inspired by the structure of the H-cluster, mixed-valent Fe^IFe^II diiron cores were envisioned to support an active redox state for H₂ binding and splitting, but the stability of such Fe^IFe^II species limits this approach.^22–30 Associated with amine functionality localized at the bridgehead of the dithiolate bridging ligand, some diiron complexes, e.g., {Fe^IFe^II(μ-azadithiolate)}, permit the stoichiometric heterolytic splitting of H₂;^31,32 to date only three reported examples have achieved catalytic H₂ oxidation.^29,30,33 The limited interaction between the amine function of the bridgehead and the distal oxidized Fe^II metallic centre of the active site results in what has been described as a frustrated Lewis pair (FLP),^34–36 where the amine in the second coordination sphere acts as a Lewis base and the iron metal plays the role of a Lewis acid.

Concomitantly to this bioinspired strategy, the heterolytic activation of H₂ using FLPs has emerged quickly in the past 15 years, starting from main group elements^37–41 and then being extended to most elements of the periodic table. Metallic elements, such as transition metals, lanthanides, and even actinides, are alternatively used as supporting complexes,^42–48 Lewis acids,^49–54 or even Lewis base partners.^55–65 To the best of our knowledge, no heterolytic splitting of H₂ by a FLP combining a dinuclear moiety with metal–metal interactions as a Lewis base with a borane Lewis acid has been reported. One rare example, reminiscent of FLP chemistry, suggests that the compound [Fe₂(μ-edt)(PMe₃)₂(CO)₄] (edt²⁻: 1,2-ethanedithiolate) acts as a Lewis base when associated with an external borane Lewis acid, and this induces H₂ cleavage.¹⁶ Considering the very limited number of models of [FeFe]-hydrogenase active sites able to activate the H–H bond, we have explored a non-biomimetic strategy to develop H₂ activation using diiron dithiolate complexes.

In this work, we propose a unique study of the influence of symmetrical or unsymmetrical substitution at a diiron core, inspired by the H-cluster, on stoichiometric H₂ activation based on two rare FLP examples involving a bimetallic moiety and B(C₆F₅)₃ as the Lewis base and acid, respectively. The activities towards H₂ of two low-valent Fe^IFe^I diiron complexes [Fe₂(μ-pdt)(PMe₃)₂(CO)₄] (1) (pdt²⁻: 1,3-propanedithiolate)¹⁸ and [Fe₂(μ-pdt)(κ²-dmpe)(CO)₄] (2) (dmpe: 1,2-bis(dimethylphosphino)ethane)⁶⁶ (Scheme 1) have been explored. Mechanistic investigations were performed via DFT to rationalize the different behaviours.

Scheme 1 The [FeFe]-complexes studied as Lewis bases in the studied FLPs.

Results and discussion

Formation of the FLPs

The addition of one equivalent of B(C₆F₅)₃ to the [Fe₂(μ-pdt)(PMe₃)₂(CO)₄] (1) or [Fe₂(μ-pdt)(κ²-dmpe)(CO)₄] (2) complex (1 equiv.) yielded a red solution. Multinuclear (¹H, ³¹P, ¹¹B) NMR and IR analysis revealed specific spectroscopic fingerprints of each separated molecule, indicating the absence of any interaction between the diiron species and borane in solution (see the ESI†). These findings differ from reported examples of [FeFe]/[NiFe] complexes in which the addition of a borane molecule resulted in its binding to an available nitrogen-based ligand, such as cyanide^67,68 or azadithiolate.⁶⁹ Interactions between the carbonyl group and Lewis acid were also reported for the complex [Fe₂(μ-pdt)(κ²-dppv)₂(CO)₂] (dppv = cis-1,2-C₂H₂(PPh₂)₂), leading to the diiron site adopting a rotated conformation.⁷⁰ The different behaviours shown by 1 and 2 under these reaction conditions may arise from the less-electron-rich nature of disubstituted diiron complexes compared to related tetrasubstituted bis-dppv diiron compounds. On the basis of these observations, both the Lewis acid and base components in these complexes (B(C₆F₅)₃ and 1/2, respectively) coexist in solution. According to FLP chemistry, these pairs may be predisposed to activating H₂ through the heterolytic splitting of the H–H bond. Consequently, the exposure of 1 and 2 to H₂ in the presence of B(C₆F₅)₃ was performed.

Activities of 1 and 2 towards H₂ in the presence of B(C₆F₅)₃

The bishydride salt [Fe₂(μ-pdt)(μ-H)(PMe₃)₂(CO)₄][HB(C₆F₅)₃] (3) was isolated from an equimolar CH₂Cl₂ solution of [Fe₂(μ-pdt)(PMe₃)₂(CO)₄] (1) and B(C₆F₅)₃ exposed to H₂ (1 atm.) (Scheme 2). The slow formation of this complex was observed through spectroscopic analysis after 2 h. In solution, a maximum yield of 77% for the conversion of 1 into 3 was reached after 5 days. This bishydride salt was isolated, upon the addition of hexane to the CH₂Cl₂ solution, as a red powder in only 53% yield due to its partial solubility in a non-polar organic solvent. Spectroscopic characterization of 3 matched the previously reported data for the cation [Fe₂(μ-pdt)(μ-H)(PMe₃)₂(CO)₄]⁺ (ref. 18 and 19) and the anionic [HB(C₆F₅)₃]⁻ species (see the ESI†).⁷¹ Notably, a triplet at −15.3 ppm in the ¹H-NMR spectrum with a ²J_H–P coupling constant of 22 Hz is characteristic of a bridging hydride between the two iron atoms of the [Fe₂(μ-pdt)(μ-H)(PMe₃)₂(CO)₄]⁺ complex, while the borohydride [HB(C₆F₅)₃]⁻ was characterized based on a broad quadruplet at 3.6 ppm in the ¹H-NMR spectrum, which is associated with a doublet in the ¹¹B-NMR spectrum at −25.7 ppm (¹J_H–B = 82 Hz). IR spectroscopy revealed two strong bands at 2032 and 1992 cm⁻¹ in CH₂Cl₂ solution in the carbonyl region of the IR spectrum, corresponding to the protonated iron complex [Fe₂(μ-pdt)(μ-H)(PMe₃)₂(CO)₄]⁺. In order to ascertain the origin of both hydride ligands, isotopic labelling experiments were performed using D₂ gas. The salt [Fe₂(μ-pdt)(μ-D)(PMe₃)₂(CO)₄][DB(C₆F₅)₃] (3_D) was characterized unambiguously based on ²H-NMR analysis (see the ESI†).

Scheme 2 The reactivities of complexes 1 and 2 with H₂ in the presence of B(C₆F₅)₃.

Similarly, the exposure to H₂ of a solution containing 1 equivalent of [Fe₂(μ-pdt)(κ²-dmpe)(CO)₄] (2) and 1 equivalent of B(C₆F₅)₃ in CH₂Cl₂ resulted in the slow formation of [Fe₂(μ-pdt)(μ-H)(κ²-dmpe)(CO)₄][HB(C₆F₅)₃] (4) (Scheme 2). This salt was isolated after 5 days in 33% yield upon the addition of diethylether to the reaction mixture. The bridging hydride in the cationic complex [Fe₂(μ-pdt)(μ-H)(κ²-dmpe)(CO)₄]⁺ was characterized based on a triplet at −14.6 ppm (²J_P–H = 21 Hz) in the ¹H-NMR spectrum, and the IR spectrum showed three strong bands at 2095 (s), 2039 (s) and 1963 (s) cm⁻¹ in the carbonyl region, which is in close agreement with the reported data for a sample of [Fe₂(μ-pdt)(μ-H)(κ²-dmpe)(CO)₄]⁺.⁶⁶ A characteristic doublet in the ¹¹B-NMR spectrum at −25.7 ppm corroborates the formation of the borohydride anion (see the ESI†).⁷¹

Additional unexpected resonances were observed in the NMR spectra of the crude reaction mixture. The triplet at −10.5 ppm (J = 50 Hz) and singlet at 67.8 ppm in the ¹H- and ³¹P-NMR spectra, respectively, were assigned to the iron hydride complex [FeH(κ²-dmpe)(CO)₃]⁺,^72,73 while two resonances (at 0 ppm and −14 ppm) and the broad singlet (at 10.6 ppm) in the ¹¹B- and ¹H-NMR spectra, respectively, were attributed to the formyl-borate species [B(C₆F₅)₃–CHO–B(C₆F₅)₃]⁻.^64,74

NMR monitoring showed that the species [FeH(κ²-dmpe)(CO)₃]⁺ appears after at least 3 days of H₂ exposure, and it is present in solution with 4 in a 1 : 3 ratio. The formyl borate anion is observed as soon as H₂ is added, suggesting ready CO release, and this then decomposed over time, as reported in the literature.^64,74

In order to investigate the possibility of the intermediate formation of a terminal hydride species, the pair formed from the complex 2 and B(C₆F₅)₃ was exposed to H₂ (1 atm.) at −70 °C (see the ESI†). A characteristic singlet at −4.5 ppm in the ¹H-NMR spectrum revealed the presence of the cationic species [Fe₂(μ-pdt)(H_t)(κ²-dmpe)(CO)₄]⁺, featuring a terminal hydride ligand on the {Fe(CO)₃} unit, in agreement with the literature.⁶⁶ The sample was then allowed to warm to room temperature and above 253 K the terminal hydride complex converted into both apical-basal and dibasal bridging hydride isomers [Fe₂(μ-pdt)(μ-H)(κ²-dmpe)(CO)₄]⁺, with the dibasal isomer being the thermodynamically preferred species after several hours at room temperature.

Experiments carried out with D₂ gas definitively established the origin of the hydride ligands in [Fe₂(μ-pdt)(μ-H)(κ²-dmpe)(CO)₄][HB(C₆F₅)₃] (4). Moreover, the formation of the bis-deuteride salt [Fe₂(μ-pdt)(μ-D)(κ²-dmpe)(CO)₄][DB(C₆F₅)₃] was accompanied by that of the mononuclear complex [FeD(κ²-dmpe)(CO)₃]⁺. The detection of this latter species and the detection of the formyl-borate species in the reaction mixtures revealed the cleavage of 2 upon reaction with H₂ in the presence of borane.

It should be noted that similar experiments with less acidic B(C₆H₅)₃ borane do not allow H₂ heterolytic splitting.

Influence of borane stoichiometry on heterolytic H₂ splitting

The exposure to H₂ of a 1 : 2 solution of [Fe₂(μ-pdt)(PMe₃)₂(CO)₄] (1) and B(C₆F₅)₃ in CH₂Cl₂ leads to the complete conversion of the diiron complex into the protonated [Fe₂(μ-pdt)(μ-H)(PMe₃)₂(CO)₄]⁺ cation after 40 h (Scheme 2). Faster and more complete conversion is thus observed when using 2 equivalents of borane, as a maximum conversion yield of 77% was observed in the presence of one equivalent of B(C₆F₅)₃. In the ¹¹B{¹H}-NMR spectrum, however, no signal was detected at room temperature. Reducing the temperature to 213 K allowed for the observation of a strong broad signal at −25 ppm (see the ESI†). The ¹H-NMR spectrum displayed one broad resonance at 3.5 ppm without a well-resolved coupling constant with the boron atom. These observations strongly suggest the formation of the diborohydride anion [(C₆F₅)₃B-(μ-H)-B(C₆F₅)₃]⁻, whose NMR characterization requires low-temperature recording.⁷⁵ The related diborohydride species [(C₆H₃(CF₃)₂)₃B-(μ-H)-B(C₆H₃(CF₃)₂)₃]⁻ has already been reported, for which no ¹H–¹¹B coupling was detected until a donor solvent, such as pyridine or acetonitrile, was added, inducing the cleavage of the dimeric species and yielding the known [HB(C₆H₃(CF₃)₂)₃]⁻ anion.^76,77 To highlight the formation of the diborohydride anion [(C₆F₅)₃B-(μ-H)-B(C₆F₅)₃]⁻, acetonitrile was added to the reaction mixture, resulting in the appearance of the characteristic resonances of [HB(C₆F₅)₃]⁻ together with a broad singlet at −10.8 ppm in the ¹¹B-NMR spectrum, which is attributed to the Lewis adduct MeCN–B(C₆F₅)₃.⁷⁸

These observations were also made in relation to the dissymmetrical complex 2; the exposure to H₂ of a solution of B(C₆F₅)₃ (2 equiv.) and [Fe₂(μ-pdt)(κ²-dmpe)(CO)₄] (1 equiv.) in CH₂Cl₂ quantitatively yields a mixture of [Fe₂(μ-pdt)(μ-H)(κ²-dmpe)(CO)₄]⁺ and [FeH(κ²-dmpe)(CO)₃]⁺ cations in a 2 : 1 ratio. ¹¹B{¹H}-NMR analysis revealed the presence of the formyl-borate [B(C₆F₅)₃-CHO-B(C₆F₅)₃]⁻, which degrades over time,^64,74 and the diborohydride anion [(C₆F₅)₃B-(μ-H)-B(C₆F₅)₃]⁻ in the anionic counterpart.

The difference in the conversion efficiencies of the reactions of 1 and 2 with one or two equivalents of borane may arise from the involvement of the hydride ligand as a bridge to the dimeric species [(C₆F₅)₃B-(μ-H)-B(C₆F₅)₃]⁻, thus preventing the reverse reaction between [HB(C₆F₅)₃]⁻ and the protonated diiron complexes.

Discussion of the different activities of 1 and 2 towards H₂ in the presence of borane

1 and 2 display similar electronic properties.^79,80 However, they behave differently towards the activation of H₂ in the presence of B(C₆F₅)₃. The differences, highlighted in this study, may arise from the symmetrical or dissymmetrical arrangement of the phosphine/diphosphine ligands at the diiron centre of the complex. In the symmetrically disubstituted complex 1, both iron metallic {Fe(PMe)₃(CO)₂} moieties have the same ligand environments, while in 2, the two iron moieties, {Fe(CO)₃} and {Fe(κ²-dmpe)(CO)}, are electronically and sterically very different. The {Fe(CO)₃} fragment is indeed less basic but also less sterically encumbered than the {Fe(κ²-dmpe)(CO)} part; thus, there are two different sites where the reaction can take place. DFT calculations have been performed in order to rationalize the different activities of 1 and 2 towards H₂/B(C₆F₅)₃. It is worth noting that in the presence of HBF₄, it has previously been reported that the dissymmetry of 2 orientates the reaction path toward the formation of a transient terminal hydride species, while this was not observed in the case of 1 under similar reaction conditions.^66,81

Mechanistic investigations via theoretical calculations

In order to shed light on the mechanistic aspects of the heterolytic splitting of the H–H bond by 1 and 2 in the presence of B(C₆F₅)₃ and on the experimental discrepancies observed between those two systems, DFT investigations using the BP86-D3/TZVP functional were performed. The theoretical rationalization of the experimental findings involves different challenges. The Lewis base component of the FLP formed between 1 or 2 and the borane Lewis acid has a bimetallic nature, which itself induces two possible iron sites for H₂ interactions. Furthermore, the fluxionality of the coordination sphere around each iron center of such species is known, and various isomers should be taken into account when considering the reactivity pathway. The most stable isomers of 1 and 2 feature the phosphine ligands in a dibasal position, but the apical-basal isomers are only 0.7 kcal mol⁻¹ and 3.0 kcal mol⁻¹ higher in energy, respectively (see the ESI†). This is consistent with the isomers reported for 1 and 2 in the literature.^18,66,81 In addition, isomers featuring a structure close to that of the active site of [FeFe]-hydrogenases were also considered, especially for 2. In such isomers, the square pyramidal coordination geometry of one Fe center is “inverted” compared to the other Fe center (see the ESI†). Considering these various geometries, the dibasal, basal-apical, and rotated diiron isomers for each FLP were examined.

H₂ cleavage by the FLP 1/B(C₆F₅)₃. A symmetrical disubstituted complex has a poor tendency to rotate at a {FeL(CO)₂} unit,^82,83 and indeed the rotated geometry was not found to be an energy minimum structure on the potential energy surface of 1. Therefore, the most likely mechanism involves the Fe–Fe region instead of an apical approach at one inverted iron center. The mechanism of H₂ splitting by the pair formed between 1 and the borane molecule is presented in Fig. 1. Two mechanisms, using the most stable dibasal or slightly less stable apical-basal isomer, by which 1 can activate H₂ have been investigated. The approach of one H₂ molecule in between the Fe–Fe region of 1 and B(C₆F₅)₃ is a low energy process, requiring 3.4 kcal mol⁻¹ and 6.1 kcal mol⁻¹ for the dibasal and apical-basal isomers, respectively. At this point, the crucial event of FLP-based H–H heterolytic cleavage occurs and the H₂ molecule is split, forming two hydrides, one installed at the diiron core and the other at the borane molecule. This process is exergonic (−14.3/−15.5 kcal mol⁻¹) and characterized by overall activation barriers of 16.8 (1-TS_bb) and 11.0 kcal mol⁻¹ (1-TS_ab) for the dibasal and the apical-basal isomers, respectively (see Fig. 1 for the optimized TS structures and the ESI† for those of other relevant intermediates). These differences are conceivably due to reduced steric hindrance in the Fe–Fe region with the PMe₃ ligand in the apical position, which reduces unfavorable repulsion between the aromatic rings of the Lewis acid. The H₂ activation that proceeds after the isomerization of the dibasal isomer of 1 to its slightly less stable apical-basal form is thus kinetically preferred. As a note, the slight discrepancy between the relative stabilities of the two rotamers in the absence of H₂ and borane and in their presence does not affect the general conclusions presented relating to reactivity.

Fig. 1 The H₂ splitting mechanisms of 1 and B(C₆F₅)₃. The optimized TS structures for H₂ splitting of 1 basal–basal (1-TS_bb) and 1 apical-basal (1-TS_ab) structures are represented in the insets. Selected distances are given in Å. NBO charges, calculated for the H1 and H2 atoms of 1-TS_bb (blue values) and 1-TS_ab (red values), calculated at the BP86-D3 and B3LYP-D3 (in parentheses) levels, are reported in the respective insets. Colors: grey = carbon; red = oxygen; light blue = iron; purple = phosphorus; yellow = sulfur; orange = boron; and green = fluorine.

H₂ cleavage by the FLP 2/B(C₆F₅)₃ at the apical vacant site. Similar considerations to those relating to 1 have been made for the FLP based on 2/B(C₆F₅)₃. Complex 2 is asymmetrical and contains two differentiated subunits: {Fe(CO)₃} and {Fe(κ²-dmpe)(CO)}. All viable mechanisms examined via DFT calculations suggest that the reaction takes place at the less basic iron part {Fe(CO)₃} rather than at the more basic {Fe(κ²-dmpe)(CO)} side, which is also more sterically crowded. Three energetically equivalent mechanisms can be invoked for H₂ cleavage by 2 and B(C₆F₅)₃. Two of them are centered at the Fe–Fe site, similar to 1, while the third involves the formation of a transient rotated isomer, directing the interaction with H₂ towards the apical position of the {Fe(CO)₃} unit. This last proposition is supported by the observation of a terminal hydride complex during low temperature experiments and is presented below.

The mechanisms shown in Fig. 2 involve the preliminary rotation of the {Fe(CO)₃} moiety of 2. Indeed, 2 has a higher tendency, compared to 1, to rotate at the {Fe(CO)₃} unit, which is consistent with the asymmetric coordination environments of the two iron ions.^82,83 The isomer of 2 featuring a vacant coordination site in the apical position of a single iron is only 3.3 kcal mol⁻¹ higher in energy with respect to the unrotated example. The dibasal form has been considered as the isomer that interacts first with H₂, because the apical-basal isomer, at this stage, is predicted to be less stable than the dibasal one (see Fig. S20 in the ESI†). Only after H₂ heterolysis has occurred do the two isomers become energetically equivalent (vide infra).

Fig. 2 The H₂ splitting mechanisms of 2 and B(C₆F₅)₃ at the apical vacant site of rotated iron. The blue pathway shows the reactivity of the dibasal rotamer of 2, whereas the red one is related to the apical-basal rotamer. Selected distances given in Å and NBO charges, calculated for the H1 and H2 atoms of 2-Int_bb (blue values) and 2-Int_ab (red values), calculated at the BP86-D3 and B3LYP-D3 (in parentheses) levels, are reported in the respective insets. Colors: grey = carbon; red = oxygen; light blue = iron; purple = phosphorus; yellow = sulfur; orange = boron; and green = fluorine.

The H₂ uptake between the diiron site and the borane molecule is a slightly exergonic process (−3.7 kcal mol⁻¹), with the lengthening of the H–H bond between the diiron center and the borane moiety to around 1.28 Å. This step is predicted to be extremely fast, considering that our simulations show the immediate formation (i.e., with no detectable barrier of activation) of 2-Int_bb once H₂ is added to the 2/Lewis acid system. A terminal hydride cationic species is then formed after the loss of the anion [HB(C₆F₅)₃]⁻. DFT calculations show that, at this stage, the two chelate orientations (dibasal and apical-basal) are mutually close to equilibrium, with quite a low barrier of interconversion. Such quasi-equilibrium results are basically unaffected by the presence of the R₃BH anion, which implies that it equally involves both 2-Int_bb/2-Int_ab and 2-t-H_bb/2-t-H_ab.

The rate determining step of the entire mechanism of H₂ cleavage mediated by the rotated structure of the diiron dithiolate is predicted to be the final step of the double Ray–Dutt twist (a pseudo-rhombic octahedral rearrangement)⁸⁴ that converts the reactive terminally coordinated hydride into its thermodynamically more stable bridging counterpart with an overall energy barrier of around 17 kcal mol⁻¹ (roughly independent of the chelate rotational isomer). A Bailar twist (trigonal-prismatic rearrangement) leads from the apical-basal μ-hydride to the thermodynamically more stable dibasal rotational isomer, as previously reported for strict analogues of 2.⁸⁴ This DFT picture of the facile and kinetically accessible conversion of apical-basal μ-H is in full agreement with the experimental observation that dibasal μ-H is the thermodynamic product at room temperature.

H₂ cleavage by the FLP 2/B(C₆F₅)₃ at the Fe–Fe region. As already mentioned, an alternative mechanism (compared to that described above) revealed that both dibasal and apical-basal forms of 2 can cleave H₂ in the presence of the Lewis acid at the Fe–Fe site. Indeed, H₂ uptake by 2 may also occur, according to the energetic criteria (and analogously to 1) between the diiron site and the borane molecule, as shown in Fig. 3. In this case, a slightly endergonic process occurs, requiring 3.9 kcal mol⁻¹ or 8.4 kcal mol⁻¹ for the two rotational isomers. An overall energy barrier of 16.8 kcal mol⁻¹ (2-TS_bb) or 15.9 kcal mol⁻¹ (2-TS_ab) is calculated, and both transition states show the lengthening of the H–H bond to around 1.21–1.22 Å between the diiron center and the borane moiety. An intermediate, which was not found for 1, was located along the reaction profile featuring an asymmetric semi-bridging hydride that interacts electrostatically with [HB(C₆F₅)₃]⁻ at the {Fe(CO)₃} unit The newly formed almost-terminal Fe–H is stabilized by a long-range interaction with the boron hydride. This intermediate is less stable than the final bridging hydride, but it can be considered as the first species that is formed upon H₂ splitting by 2 at the Fe–Fe site.

Fig. 3 The H₂ splitting mechanisms by 2 and B(C₆F₅)₃; the optimized structures of the H₂-splitting TSs for 2 basal–basal (2-TS_bb) and 2 apical-basal (2-TS_ab) isomers are represented in the insets. Selected distances are given in Å. NBO charges, calculated for the H1 and H2 atoms of 2-TS_bb (blue values) and 2-TS_ab (red values), calculated at the BP86-D3 and B3LYP-D3 (in parentheses) levels, are also reported. Colors: grey = carbon; red = oxygen; light blue = iron; purple = phosphorus; yellow = sulfur; orange = boron; and green = fluorine.

The occurrence of H₂ cleavage at the non-substituted iron site makes this reactive event essentially unsensitive to the nature (dibasal or basal-apical) of the isomer that performs the reaction, as shown by the close energies of both transition states during the rate-determining step. The three mechanisms revealed significantly close energy barriers, but the one involving a terminal hydride isomer, observed experimentally, may be the most probable. Indeed, from the DFT viewpoint, the heights of the rate-determining step of each H₂ cleavage pathway are very similar (differences within a 1 kcal mol⁻¹ range), which makes them all theoretically possible. However, the experimental evidence points to a slight preference for the “enzyme-like mechanism”. This statement is further supported by the extremely facile H₂ cleavage step at the apical site of a rotated iron atom (Fig. 2), in close analogy to what occurs in the H-cluster within the hydrogenase active site. The energy bottleneck of H₂ splitting via the “rotated iron center mechanism” comes, in fact, after H₂ undergoes heterolysis in the FLP system and consists of a two-step rearrangement that is impeded in the biological catalyst.⁸⁵

Comparison of H₂ cleavage mechanisms by the FLPs 1 and 2/B(C₆F₅)₃. Considering the mechanisms at the Fe–Fe site, H₂ activation is much faster for the apical-basal isomer of 1 than the dibasal example, while in the case of 2, similar activation barriers have been calculated for both isomers. Indeed, the basal-apical isomer of 1 displays a less crowded equatorial side at one iron center, significantly reducing the steric hindrance in the Fe–Fe region and favoring the approach of both the Lewis acid and, afterwards, H₂ closer to the diiron Lewis base, granting faster reactivity; however, the reaction takes place at the unsubstituted {Fe(CO)₃} site of 2, which is less impacted by the orientation of the dmpe ligand.

The basal–apical forms of 1 and 2 and of the hydride species involved are relatively less stable than their dibasal isomers. This is due to steric bumps between the phosphine ligand and the propanedithiolate bridge, which are more emphasized with the more-rigid chelating dmpe ligand. Comparing the lowest energy profiles calculated for 1 and 2, it can be noticed that the former is preferred from a kinetic (by ca. 5.8 kcal mol⁻¹) standpoint while the thermodynamics of both processes are similar. The steric hindrance at the Fe–Fe site was roughly similar for the apical-basal forms of both 1 and 2, and the experimentally observed difference can be ascribed to difference in the electron-rich natures of the iron sites directly involved in H₂ splitting ({Fe(PMe₃)(CO)₂} for 1 and {Fe(CO)₃} for 2). In the case of 1, the PMe₃ ligand adopts a trans orientation with respect to the incoming hydrogen atom, which has protic character (that forming the B–H bond is hydridic). Analysis of the electron density distribution over the atomic centers based on classical partition schemes, such as NBO (see Fig. 1–3 and the ESI†) and Löwdin (see the ESI†), calculated at both the BP86-D3 and B3LYP-D3 levels on the H₂ molecule with respect to 1-TS_bb, 1-TS_ab, 2-TS_bb, and 2-TS_ab corroborates the aforementioned considerations. Indeed, the extent of H₂ heterolysis (corresponding to the polarization level of the H–H bond) is slightly higher for the apical-basal isomers than the dibasal examples. More relevant is that H^δ+–H^δ− polarization is more pronounced in 1 than in 2 as a result of the increased basicity of the metal center involved in splitting. Similar partial-charge-related arguments, based on the NBO data, have to be considered with some caution in the case of the splitting mechanism reported in Fig. 2 due to the actual absence of a transition state associated with such a step. However, considering the two intermediates arising just after H₂ heterolysis (2-Int_bb and 2-Int_ab, see the insets of Fig. 2), it can be concluded that the highest levels of H^δ+–H^δ− polarization is predicted to occur exactly when H₂ splits in an enzyme-like fashion, which further supports such an activation pathway compared with the others.

The formation of the asymmetric semi-bridging hydride intermediate (not observed for 1) or the terminal hydride intermediate from 2 (Fig. 2 and 3) enhanced the reversibility of H₂ splitting with respect to 1, and the backward reaction (namely H₂ formation) could be easily triggered in the case of the former complex. This can also explain the similar behaviors of 2 and 1 in the presence of two Lewis acid equivalents, since [HB(C₆F₅)₃]⁻, which is necessary to stabilize the higher-energy product in the case of 2, can be easily trapped by a second B(C₆F₅)₃ molecule. Indeed, coupling between B(C₆F₅)₃ and [HB(C₆F₅)₃]⁻ to form a [(C₆F₅)₃B-(μ-H)-B(C₆F₅)₃]⁻ bridging hydride was calculated to be exergonic by −10.2 kcal mol⁻¹ (see the ESI† for details).

Conclusions

In this work, we have established that FLPs consisting of a reduced Fe^IFe^I complex and a borane molecule are able to activate H₂, forming two hydride species, one connected to the diiron core and the other to the borane molecule. The higher reactivity of symmetrical [Fe₂(μ-pdt)(PMe₃)₂(CO)₄] (1) compared to that of asymmetrical chelated [Fe₂(μ-pdt)(κ²-dmpe)(CO)₄] (2) towards H₂ in the presence of B(C₆F₅)₃ has been elucidated based on theoretical calculations both in steric and electronic terms. The different symmetries of the two complexes lead to distinctive steric factors determining the selection of the iron center for performing the heterolytic cleavage of H₂. For both complexes, the reaction takes place at the less sterically hindered iron center in order to favor the approach of both H₂ and the Lewis acid. The reaction involving the symmetrical diiron complex 1 takes place at one of the two {Fe(PMe₃)(CO)₂} units, which are more basic than the {Fe(CO)₃} part of the asymmetrical complex 2, thus favoring faster reactivity. In the case of 2, the direct observation of a terminal hydride intermediate in low-temperature experiments highlights a preferential mechanism that leads to [FeFe]-hydrogenase-like reactivity, occurring through the binding-activation of H₂ at a free apical/terminal iron site via accessible rotated geometry at the {Fe(CO)₃} site. The mechanistic differences emerging between 1 and 2 corroborate the importance of the coordination symmetry/asymmetry pattern when it comes to driving the regiochemistry of H₂ heterolysis during hydrogenase biomimicry.

Finally, this study may open up novel routes to H₂ oxidation, and it offers a previously unexplored non-biomimetic strategy for H₂ activation using diiron dithiolate complexes. In general, the use of transition-metal systems as Lewis acid/base components is less common than the use of main-group elements within the vast realm of FLP chemistry.⁴⁵ Further experiments relating to the transfer of hydrogen atoms from H₂ to unsaturated substrates and the oxidation of hydrido-borate are ongoing.

Experimental

Materials and methods

All the experiments were carried out under an inert atmosphere using Schlenk or glovebox techniques for synthesis. Solvents were deoxygenated and dried according to standard procedures. [Fe₂(μ-pdt)(PMe₃)₂(CO)₄] (1) and [Fe₂(μ-pdt)(κ²-dmpe)(CO)₄] (2) were prepared according to reported procedures.^19,66 B(C₆F₅)₃ was prepared according to the literature and sublimated.⁸⁶¹H-, ³¹P-, ¹¹B-, ¹⁹F-, and ¹³C-NMR spectra were respectively recorded at room temperature in deuterated solvents using a Bruker Avance 400 and Avance III HD 500 spectrometer at “Service RMN-SM de l’Université de Bretagne Occidentale”. They were referenced to SiMe₄ (¹H, ¹³C), H₃PO₄ (³¹P), BF₃·OEt₂ (¹¹B), and CFCl₃ (¹⁹F). Infrared spectra were recorded with a PerkinElmer spectrometer. Chemical analyses were carried out by the “Service d'Analyses – Chromato – Masse BioCIS – UMR 8076”, Châtenay-Malabry (France).

Synthetic procedures

Activity of 1 towards H₂ in the presence of B(C₆F₅)₃. The complex [Fe₂(μ-pdt)(PMe₃)₂(CO)₄] (1) (163 mg, 0.338 mmol, 1 eq.) and B(C₆F₅)₃ (173 mg, 0.338 mmol, 1 eq.) were dissolved in CH₂Cl₂ (15 mL). The resulting red solution was placed in a J. Young flask and degassed over two freeze/vacuum/thaw cycles. The headspace was then filled with H₂ (1 atm). The solution was stirred for five days at room temperature. The solution was concentrated to 1 mL of CH₂Cl₂, and 5 mL of hexane was added, precipitating the [Fe₂(μ-pdt)(μ-H)(PMe₃)₂(CO)₄][HB(C₆F₅)₃] salt (3) as a red solid. This solid was washed five times with hexane (2 mL) until no more [Fe₂(μ-pdt)(PMe₃)₂(CO)₄] (1) was present in the remaining solid. Volatiles were removed under vacuum, yielding 180 mg of [Fe₂(μ-pdt)(μ-H)(PMe₃)₂(CO)₄][HB(C₆F₅)₃] (3) as a red solid (53%). Some of this compound dissolved in hexane, reducing the yield of the isolated product. 3:¹H NMR (500 MHz, 298 K, CD₂Cl₂): δ (ppm) 3.60 (br q., 1H, ¹J_H–B = 90 Hz, B–H), 2.60 (m, 4H, –S–CH₂–), 2.20 (m, 2H, –CH₂–), 1.58 (d, ²J_H–P = 10 Hz, P–CH₃), −15.36 (t, ²J_H–P = 22 Hz, 1H, μ-H). ³¹P{¹H} NMR (202.46 MHz, 298 K, CD₂Cl₂): δ (ppm) 21.33 (s, PMe₃), 21.29 (s, PMe₃). ¹¹B{¹H} NMR (160.46 MHz, 298 K, CD₂Cl₂): δ (ppm) −25.74 (br s, B–H). ¹¹B NMR (160.46 MHz, 298 K, CD₂Cl₂): δ (ppm) −25.74 (d, ¹J_B–H = 90 Hz, B–H). ¹⁹F{¹H} NMR (470.59 MHz, 298 K, CD₂Cl₂): δ (ppm) −134.15 (br d, 6F, ³J_F–F = 16 Hz, ortho-C₆F₅), −164.98 (t, 3F, ³J_F–F = 21 Hz, para-C₆F₅), −167.80 (br t, 6F, ³J_F–F = 18 Hz, meta-C₆F₅). C, H, N anal. (%) calc. for C₃₁H₂₆BF₁₅Fe₂O₄P₂S₂ (996.09): C 37.38, H 2.63; found: C 37.39, H 2.29. IR (CH₂Cl₂, cm⁻¹): ν(CO) 2032 (s), 1992 (s).

Activity of 2 towards H₂ in the presence of B(C₆F₅)₃. The complex [Fe₂(μ-pdt)(κ²-dmpe)(CO)₄] (2) (208 mg, 0.433 mmol, 1 eq.) and B(C₆F₅)₃ (222 mg, 0.433 mmol, 1 eq.) were dissolved in CH₂Cl₂ (15 mL). The resulting red solution was placed in a J. Young flask and degassed over two freeze/vacuum/thaw cycles. The headspace was then filled with H₂ (1 atm). The solution was stirred for four days at room temperature. The solution was concentrated to 1 mL of CH₂Cl₂, and 5 mL of Et₂O was added, precipitating the [Fe₂(μ-pdt)(μ-H)(κ²-dmpe)(CO)₄][HB(C₆F₅)₃] salt (4). This solid was washed with 2 mL of Et₂O and dried under vacuum, yielding 146 mg of the brown solid 4 (33%). 4: ¹H NMR (500 MHz, 298 K, CD₂Cl₂): δ (ppm) 2.83 (d, 13 Hz, 2H), 2.60 (d, 15 Hz, 1H), 2.50 (t, 13 Hz, 2H), 2.25 (m, 4H), 2.00 (m, 1H), 1.67 (d, 9 Hz, 6H, CH₃), 1.33 (d, 9 Hz, 6H, CH₃), −14.67 (t, ²J_H–P = 21 Hz, 1H, μ-H), B–H resonance not detected with ¹H NMR. ³¹P{¹H} NMR (202.46 MHz, 298 K, CD₂Cl₂): δ (ppm) 67.38 (s). ¹¹B{¹H} NMR (160.46 MHz, 298 K, CD₂Cl₂): δ (ppm) −25.50 (br s, B–H). ¹¹B NMR (160.46 MHz, 298 K, CD₂Cl₂): δ (ppm) −25.74 (d, ¹J_B–H = 92 Hz, B–H). ¹⁹F{¹H} NMR (470.59 MHz, 298 K, CD₂Cl₂): δ (ppm) −133.74 (br, 6F, ortho-C₆F₅), −159.83 (br m, 3F, para-C₆F₅), −165.86 (br m, 6F, meta-C₆F₅). C, H, N anal. (%) calc. for C₃₁H₂₄BF₁₅Fe₂O₄P₂S₂ (994.05): C 37.46, H 2.43; found: C 37.28, H 1.76. IR (CH₂Cl₂, cm⁻¹): ν(CO) 2095 (s), 2039 (s), 1963 (s).

[FeH(κ²-dmpe)(CO)₃]⁺: ¹H NMR (500 MHz, 298 K, CD₂Cl₂): δ (ppm) 2.25 (m), 1.79 (m), 1.58 (m), −10.5 (t, ²J_H–P = 50 Hz, 1H), integrals were not perfectly defined as the product was not isolated and resonances superimposed with the resonances of the complex [Fe₂(μ-pdt)(μ-H)(κ²-dmpe)(CO)₄]⁺. ³¹P{¹H} NMR (202.46 MHz, 298 K, CD₂Cl₂): δ (ppm) 67.80 (s).

When the reaction was carried out in C₆D₆, no spectroscopic changes were observed in the ¹H-, ³¹P-, and ¹¹B-NMR spectra, even after 8 days at room temperature. However, a lot of solid precipitated in the NMR tube. After removing the supernatant and dissolving the residue in CD₂Cl₂, NMR spectra revealed the presence of the same two hydride complexes in a 1 : 3 ratio, together with residual traces of the starting complex.

Reactivity of [Fe₂(μ-pdt)(PMe₃)₂(CO)₄] (1) and [Fe₂(μ-pdt)(κ²-dmpe)(CO)₄] (2) with B(C₆F₅)₃ and D₂. The diiron complex (20 mg, 1 eq.) and B(C₆F₅)₃ (1 eq.) were dissolved in 0.5 mL of CD₂Cl₂ and placed in an NMR tube equipped with a J. Young valve. After degassing, the red solution was exposed to D₂ (1 atm). The evolution of the reaction was followed via NMR spectroscopy, and this allowed the characterization of the deuterium atoms as bridging deuteride bound to the diiron core and as deuteride in [DB(C₆F₅)₃]⁻.

²H NMR (76.73 MHz, 298 K, CD₂Cl₂): δ (ppm) 3.60 (br, B–D), −15.45 (t, ²J_D–P = 3.1 Hz, μ-D) for [Fe₂(μ-pdt)(μ-D)(PMe₃)₂(CO)₄][DB(C₆F₅)₃] and −14.7 (t, ²J_D–P = 2.6 Hz, μ-D) for [Fe₂(μ-pdt)(μ-D)(κ²-dmpe)(CO)₄]⁺, −10.5 ppm (²J_D–P = 7.0 Hz) for [Fe(κ²-dmpe)(CO)₃D]⁺. ¹¹B{¹H} NMR (160.46 MHz, 298 K, CD₂Cl₂): δ (ppm) −25.74 (br s, B–D). ¹¹B NMR (160.46 MHz, 298 K, CD₂Cl₂): δ (ppm) −25.74 (br s, B–D).

Reactivity of [Fe₂(μ-pdt)(PMe₃)₂(CO)₄] (1) and [Fe₂(μ-pdt)(κ²-dmpe)(CO)₄] (2) with 2 eq. of B(C₆F₅)₃ and H₂. The diiron complex (20 mg, 1 eq.) and B(C₆F₅)₃ (2 eq.) were dissolved in 0.5 mL of CD₂Cl₂ and placed in an NMR tube equipped with a J. Young valve. After degassing, the red solution was exposed to H₂ (1 atm). The evolution of the reaction was followed via NMR spectroscopy. Complete conversion into 3′ or 4′ (with the characteristic decomposition product [Fe(κ²-dmpe)(CO)₃H]⁺) was observed after 40 h.
NMR characterization of [Fe₂(μ-pdt)(μ-H)(PMe₃)₂(CO)₄][(C₆F₅)₃B-(μ-H)-B(C₆F₅)₃] (3′). ¹H NMR (400 MHz, 298 K, CD₂Cl₂): δ (ppm) 3.58 (br s., 1H, B–H–B), 2.58 (m, 4H, –S–CH₂–), 2.19 (q, 2H, –CH₂–), 1.56 (d, ²J_H–P = 10 Hz, P–CH₃), −15.37 (t, ²J_H–P = 22 Hz, 1H, μ-H). ³¹P{¹H} NMR (161.916 MHz, 298 K, CD₂Cl₂): δ (ppm) 21.26 (s, PMe₃). ¹¹B NMR (160.46 MHz, 213 K, CD₂Cl₂): δ (ppm) −26.4 (br s, B–H).
NMR characterization of [Fe₂(μ-pdt)(μ-H)(κ²-dmpe)₂(CO)₄][(C₆F₅)₃B-(μ-H)-B(C₆F₅)₃] (4′). ¹H NMR (400 MHz, 298 K, CD₂Cl₂): δ (ppm) 3.47 (br s, 1H, B–H–B), 2.83 (d, 13 Hz, 2H), 2.57 (d, 15 Hz, 1H), 2.47 (t, 13 Hz, 2H), 2.24 (m, 4H), 1.99 (m, 1H), 1.65 (d, 5 Hz, 6H, CH₃), 1.30 (d, 5 Hz, 6H, CH₃), −14.67 (t, ²J_H–P = 21 Hz, 1H, μ-H). ³¹P{¹H} NMR (161.916 MHz, 298 K, CD₂Cl₂): δ (ppm) 67.48 (s). ¹¹B NMR (160.46 MHz, 213 K, CD₂Cl₂): δ (ppm) −26.4 (br s, B–H).

The addition of one drop of MeCN to a solution of 3′ or 4′ can provide spectroscopic data for the respective salt 3 or 4 and the adduct MeCN–B(C₆F₅)₃ (¹¹B{¹H} NMR (160.46 MHz, 298 K, CD₂Cl₂): δ (ppm) −10.8 (br s)).

DFT

The TURBOMOLE 7.2 suite of programs⁸⁷ was used to perform all calculations within the density functional theory (DFT) framework at the BP86/TZVP level.^88–90 BP86 is a typical choice for the investigation of bioinorganic systems, particularly hydrogenase-related models, given its satisfactory performance in the reproduction of structural and energetic features.^85,91–93 Dispersive interactions were accounted for using the Grimme D3 scheme.⁹⁴ The solvent (CH₂Cl₂) was modelled implicitly according to the conductor-like screening model (COSMO) approach,^95,96 by setting a dielectric constant of 8.9. We took advantage of the Resolution-of-Identity (RI) technique to speed up calculations.⁹⁷ All species have been modeled as low-spin states, in light of the characteristics of the Fe coordination sphere. Full vibrational analysis has been performed with the aim of characterizing the nature of each stationary point (zero imaginary frequencies for pure minima and a single imaginary frequency for saddle points). A transition-state search was performed following a pseudo-Newton–Raphson procedure, as illustrated in detail in ref. 98.⁹⁸ Free energy (G) values have been obtained from the electronic self-consistent field (SCF) energies via considering three contributions to the total partition function (Q): q_{translational}, q_rotational, and q_vibrational, assuming that it can be written as a product of these. To evaluate enthalpy and entropy contributions, the values for temperature and pressure, and the scaling factor for the SCF wavenumbers were set to 298.15 K, 1 atm, and 0.9144 (as implemented in TURBOMOLE), respectively. Natural bond orbitals (NBO)⁹⁹ and Löwdin¹⁰⁰ population analysis were performed on selected intermediates and transition states, both at the BP86-D3 and B3LYP-D3 (single point calculations on BP86-D3 optimized structures) levels.

Data availability

Experimental data have been provided in ESI.†

Author contributions

The research project was defined by L. C. within the JCJC ANR program OxySplit-H2. J. B. B. and L. C. carried out the synthetic experiments and analysed the experimental data. F. A. and G. Z. carried out and analysed the DFT calculations. The manuscript was prepared through contributions from F. A., L. C., P. S., and G. Z.

Conflicts of interest

There are no conflicts to declare.

Acknowledgements

NMR experiments were performed at the ‘Service Général des Plateformes, Brest’ (SGPLAT). JBB is funded by the Region Bretagne. CNRS (Centre National de la Recherche Scientifique) and the Université de Bretagne Occidentale (Brest) are acknowledged for financial support. LC is thankful for the support of the Région Bretagne relating to the project ConvcatH (Stratégie d’Attractivité Durable) and the ANR (Agence Nationale de Recherche) for funding the project JCJC OxySplit-H2 (ANR-20-CE07-0027-01).

References

1. F. Möller, S. Piontek, R. G. Miller and U.-P. Apfel, Chem. –Eur. J., 2018, 24, 1471–1493.
2. J. T. Kleinhaus, F. Wittkamp, S. Yadav, D. Siegmund and U.-P. Apfel, Chem. Soc. Rev., 2021, 50, 1668–1784.
3. C. Baffert, M. Demuez, L. Cournac, B. Burlat, B. Guigliarelli, P. Bertrand, L. Girbal and C. Léger, Angew. Chem., Int. Ed., 2008, 47, 2052–2054.
4. G. Goldet, C. Brandmayr, S. T. Stripp, T. Happe, C. Cavazza, J. C. Fontecilla-Camps and F. A. Armstrong, J. Am. Chem. Soc., 2009, 131, 14979–14989.
5. S. T. Stripp, G. Goldet, C. Brandmayr, O. Sanganas, K. A. Vincent, M. Haumann, F. A. Armstrong and T. Happe, Proc. Natl. Acad. Sci. U. S. A., 2009, 106, 17331–17336.
6. A. F. Wait, C. Brandmayr, S. T. Stripp, C. Cavazza, J. C. Fontecilla-Camps, T. Happe and F. A. Armstrong, J. Am. Chem. Soc., 2011, 133, 1282–1285.
7. A. Bachmeier, J. Esselborn, S. V. Hexter, T. Krämer, K. Klein, T. Happe, J. E. McGrady, W. K. Myers and F. A. Armstrong, J. Am. Chem. Soc., 2015, 137, 5381–5389.
8. A. Kubas, C. Orain, D. De Sancho, L. Saujet, M. Sensi, C. Gauquelin, I. Meynial-Salles, P. Soucaille, H. Bottin, C. Baffert, V. Fourmond, R. B. Best, J. Blumberger and C. Léger, Nat. Chem., 2017, 9, 88–95.
9. M. del Barrio, M. Sensi, L. Fradale, M. Bruschi, C. Greco, L. de Gioia, L. Bertini, V. Fourmond and C. Léger, J. Am. Chem. Soc., 2018, 140, 5485–5492.
10. M. Winkler, J. Duan, A. Rutz, C. Felbek, L. Scholtysek, O. Lampret, J. Jaenecke, U.-P. Apfel, G. Gilardi, F. Valetti, V. Fourmond, E. Hofmann, C. Léger and T. Happe, Nat. Commun., 2021, 12, 756.
11. Z. Thammavongsy, I. P. Mercer and J. Y. Yang, Chem. Commun., 2019, 55, 10342–10358.
12. C. Elleouet, F. Y. Pétillon and P. Schollhammer, in Advances in Bioorganometallic Chemistry, ed. T. Hirao and T. Moriuchi, Elsevier, 2019, pp. 347–364.
13. D. Schilter, J. M. Camara, M. T. Huynh, S. Hammes-Schiffer and T. B. Rauchfuss, Chem. Rev., 2016, 116, 8693–8749.
14. F. Wittkamp, M. Senger, S. T. Stripp and U.-P. Apfel, Chem. Commun., 2018, 54, 5934–5942.
15. T. Xu, D. Chen and X. Hu, Coord. Chem. Rev., 2015, 303, 32–41.
16. Z. M. Heiden, G. Zampella, L. De Gioia and T. B. Rauchfuss, Angew. Chem., Int. Ed., 2008, 47, 9756–9759.
17. S. L. Matthews and D. M. Heinekey, Inorg. Chem., 2011, 50, 7925–7927.
18. X. Zhao, I. P. Georgakaki, M. L. Miller, J. C. Yarbrough and M. Y. Darensbourg, J. Am. Chem. Soc., 2001, 123, 9710–9711.
19. X. Zhao, I. P. Georgakaki, M. L. Miller, R. Mejia-Rodriguez, C.-Y. Chiang and M. Y. Darensbourg, Inorg. Chem., 2002, 41, 3917–3928.
20. I. P. Georgakaki, M. L. Miller and M. Y. Darensbourg, Inorg. Chem., 2003, 42, 2489–2494.
21. X. Yu, M. Pang, S. Zhang, X. Hu, C.-H. Tung and W. Wang, J. Am. Chem. Soc., 2018, 140, 11454–11463.
22. M. Razavet, S. J. Borg, S. J. George, S. P. Best, S. A. Fairhurst and C. J. Pickett, Chem. Commun., 2002, 700–701.
23. Ö. F. Erdem, L. Schwartz, M. Stein, A. Silakov, S. Kaur-Ghumaan, P. Huang, S. Ott, E. J. Reijerse and W. Lubitz, Angew. Chem., Int. Ed., 2011, 50, 1439–1443.
24. T. Liu and M. Y. Darensbourg, J. Am. Chem. Soc., 2007, 129, 7008–7009.
25. A. K. Justice, T. B. Rauchfuss and S. R. Wilson, Angew. Chem., Int. Ed., 2007, 46, 6152–6154.
26. C. M. Thomas, T. Liu, M. B. Hall and M. Y. Darensbourg, Inorg. Chem., 2008, 47, 7009–7024.
27. A. K. Justice, L. De Gioia, M. J. Nilges, T. B. Rauchfuss, S. R. Wilson and G. Zampella, Inorg. Chem., 2008, 47, 7405–7414.
28. M. Karnahl, S. Tschierlei, Ö. F. Erdem, S. Pullen, M.-P. Santoni, E. J. Reijerse, W. Lubitz and S. Ott, Dalton Trans., 2012, 41, 12468–12477.
29. J. M. Camara and T. B. Rauchfuss, Nat. Chem., 2012, 4, 26–30.
30. N. Wang, M. Wang, Y. Wang, D. Zheng, H. Han, M. S. G. Ahlquist and L. Sun, J. Am. Chem. Soc., 2013, 135, 13688–13691.
31. M. T. Olsen, B. E. Barton and T. B. Rauchfuss, Inorg. Chem., 2009, 48, 7507–7509.
32. J. M. Camara and T. B. Rauchfuss, J. Am. Chem. Soc., 2011, 133, 8098–8101.
33. S. Ghosh, G. Hogarth, N. Hollingsworth, K. B. Holt, S. E. Kabir and B. E. Sanchez, Chem. Commun., 2013, 50, 945–947.
34. R. M. Bullock and G. M. Chambers, Philos. Trans. R. Soc., A, 2017, 375, 20170002.
35. F.-G. Fontaine and D. W. Stephan, Philos. Trans. R. Soc., A, 2017, 375, 20170004.
36. D. W. Stephan, J. Am. Chem. Soc., 2021, 143, 20002–20014.
37. G. C. Welch, R. R. S. Juan, J. D. Masuda and D. W. Stephan, Science, 2006, 314, 1124–1126.
38. D. W. Stephan and G. Erker, Angew. Chem., Int. Ed., 2015, 54, 6400–6441.
39. D. W. Stephan, J. Am. Chem. Soc., 2015, 137, 10018–10032.
40. D. W. Stephan, Acc. Chem. Res., 2015, 48, 306–316.
41. D. W. Stephan, Science, 2016, 354, aaf7229.
42. D. P. Huber, G. Kehr, K. Bergander, R. Fröhlich, G. Erker, S. Tanino, Y. Ohki and K. Tatsumi, Organometallics, 2008, 27, 5279–5284.
43. A. J. M. Miller, J. A. Labinger and J. E. Bercaw, J. Am. Chem. Soc., 2010, 132, 3301–3303.
44. M. P. Boone and D. W. Stephan, J. Am. Chem. Soc., 2013, 135, 8508–8511.
45. S. R. Flynn and D. F. Wass, ACS Catal., 2013, 3, 2574–2581.
46. N. S. Lambic, R. D. Sommer and E. A. Ison, J. Am. Chem. Soc., 2016, 138, 4832–4842.
47. L. Chatelain, E. Louyriac, I. Douair, E. Lu, F. Tuna, A. J. Wooles, B. M. Gardner, L. Maron and S. T. Liddle, Nat. Commun., 2020, 11, 1–12.
48. A. Paparo, A. L. P. Nguyen, J. S. Silvia, T. P. Spaniol, L. Maron, C. C. Cummins and J. Okuda, Dalton Trans., 2021, 50, 10692–10695.
49. A. M. Chapman, M. F. Haddow and D. F. Wass, J. Am. Chem. Soc., 2011, 133, 8826–8829.
50. A. T. Normand, C. G. Daniliuc, B. Wibbeling, G. Kehr, P. Le Gendre and G. Erker, J. Am. Chem. Soc., 2015, 137, 10796–10808.
51. H. B. Hamilton, A. M. King, H. A. Sparkes, N. E. Pridmore and D. F. Wass, Inorg. Chem., 2019, 58, 6399–6409.
52. M. Carmona, J. Ferrer, R. Rodríguez, V. Passarelli, F. J. Lahoz, P. García-Orduña, L. Cañadillas-Delgado and D. Carmona, Chem. –Eur. J., 2019, 25, 13665–13670.
53. P. H. M. Budzelaar, D. L. Hughes, M. Bochmann, A. Macchioni and L. Rocchigiani, Chem. Commun., 2020, 56, 2542–2545.
54. M. Isegawa, T. Matsumoto and S. Ogo, RSC Adv., 2021, 11, 28420–28432.
55. J. B. Bonanno, T. P. Henry, P. T. Wolczanski, A. W. Pierpont and T. R. Cundari, Inorg. Chem., 2007, 46, 1222–1232.
56. N. Tsoureas, Y.-Y. Kuo, M. F. Haddow and G. R. Owen, Chem. Commun., 2010, 47, 484–486.
57. W. H. Harman and J. C. Peters, J. Am. Chem. Soc., 2012, 134, 5080–5082.
58. H. Fong, M.-E. Moret, Y. Lee and J. C. Peters, Organometallics, 2013, 32, 3053–3062.
59. B. E. Cowie and D. J. H. Emslie, Chem. –Eur. J., 2014, 20, 16899–16912.
60. B. R. Barnett, C. E. Moore, A. L. Rheingold and J. S. Figueroa, J. Am. Chem. Soc., 2014, 136, 10262–10265.
61. S. J. K. Forrest, J. Clifton, N. Fey, P. G. Pringle, H. A. Sparkes and D. F. Wass, Angew. Chem., Int. Ed., 2015, 54, 2223–2227.
62. J. Campos, J. Am. Chem. Soc., 2017, 139, 2944–2947.
63. N. Hidalgo, J. J. Moreno, M. Pérez-Jiménez, C. Maya, J. López-Serrano and J. Campos, Chem. –Eur. J., 2020, 26, 5982–5993.
64. K. Mistry, P. G. Pringle, H. A. Sparkes and D. F. Wass, Organometallics, 2020, 39, 468–477.
65. H. Tinnermann, C. Fraser and R. D. Young, Dalton Trans., 2020, 49, 15184–15189.
66. S. Ezzaher, J.-F. Capon, F. Gloaguen, N. Kervarec, F. Y. Pétillon, R. Pichon, P. Schollhammer and J. Talarmin, C. R. Chimie, 2008, 11, 906–914.
67. B. C. Manor and T. B. Rauchfuss, J. Am. Chem. Soc., 2013, 135, 11895–11900.
68. B. C. Manor, M. R. Ringenberg and T. B. Rauchfuss, Inorg. Chem., 2014, 53, 7241–7247.
69. N. Lalaoui, T. Woods, T. B. Rauchfuss and G. Zampella, Organometallics, 2017, 36, 2054–2057.
70. A. K. Justice, G. Zampella, L. D. Gioia and T. B. Rauchfuss, Chem. Commun., 2007, 2019–2021.
71. E. J. Lawrence, V. S. Oganesyan, D. L. Hughes, A. E. Ashley and G. G. Wildgoose, J. Am. Chem. Soc., 2014, 136, 6031–6036.
72. J. R. S. Jr, V. Zanotti, G. Facchin and R. J. Angelici, J. Am. Chem. Soc., 1992, 114, 160–165.
73. O. R. Allen, S. J. Dalgarno and L. D. Field, Organometallics, 2008, 27, 3328–3330.
74. R. Dobrovetsky and D. W. Stephan, J. Am. Chem. Soc., 2013, 135, 4974–4977.
75. R. González-Hernández, J. Chai, R. Charles, O. Pérez-Camacho, S. Kniajanski and S. Collins, Organometallics, 2006, 25, 5366–5373.
76. T. J. Herrington, A. J. W. Thom, A. J. P. White and A. E. Ashley, Dalton Trans., 2012, 41, 9019–9022.
77. E. L. Kolychev, T. Bannenberg, M. Freytag, C. G. Daniliuc, P. G. Jones and M. Tamm, Chem. –Eur. J., 2012, 18, 16938–16946.
78. H. Jacobsen, H. Berke, S. Döring, G. Kehr, G. Erker, R. Fröhlich and O. Meyer, Organometallics, 1999, 18, 1724–1735.
79. A. Jablonskytė, L. R. Webster, T. R. Simmons, J. A. Wright and C. J. Pickett, J. Am. Chem. Soc., 2014, 136, 13038–13044.
80. F. Arrigoni, S. Mohamed Bouh, L. De Gioia, C. Elleouet, F. Y. Pétillon, P. Schollhammer and G. Zampella, Chem. –Eur. J., 2017, 23, 4364–4372.
81. J. A. Wright and C. J. Pickett, Chem. Commun., 2009, 5719.
82. J. W. Tye, M. Y. Darensbourg and M. B. Hall, Inorg. Chem., 2006, 45, 1552–1559.
83. F. Arrigoni, L. Bertini, R. Breglia, C. Greco, L. De Gioia and G. Zampella, New J. Chem., 2020, 44, 17596–17615.
84. G. Zampella, P. Fantucci and L. De Gioia, Chem. Commun., 2010, 46, 8824.
85. A. R. Finkelmann, M. T. Stiebritz and M. Reiher, Chem. Sci., 2014, 5, 215–221.
86. D. S. Lancaster, Alkylation of boron trifluoride with pentafluorophenyl Grignard reagent; Tris(pentafluorophenyl)boron, http://cssp.chemspider.com/215, accessed 14 October 2021.
87. R. Ahlrichs, M. Bär, M. Häser, H. Horn and C. Kölmel, Chem. Phys. Lett., 1989, 162, 165–169.
88. A. D. Becke, Phys. Rev. A, 1988, 38, 3098–3100.
89. J. P. Perdew, Phys. Rev. B, 1986, 33, 8822–8824.
90. A. Schäfer, C. Huber and R. Ahlrichs, J. Chem. Phys., 1994, 100, 5829–5835.
91. M. Bruschi, R. Breglia, F. Arrigoni, P. Fantucci and L. De Gioia, Int. J. Quantum Chem., 2016, 116, 1695–1705.
92. G. M. Chambers, T. B. Rauchfuss, F. Arrigoni and G. Zampella, Organometallics, 2016, 35, 836–846.
93. S. Sicolo, M. Bruschi, L. Bertini, G. Zampella, G. Filippi, F. Arrigoni, L. De Gioia and C. Greco, Int. J. Hydrog. Energy, 2014, 39, 18565–18573.
94. S. Grimme, J. Antony, S. Ehrlich and H. Krieg, J. Chem. Phys., 2010, 132, 154104–154122.
95. A. Klamt, J. Phys. Chem., 1995, 99, 2224–2235.
96. A. Klamt, J. Phys. Chem., 1996, 100, 3349–3353.
97. K. Eichkorn, F. Weigend, O. Treutler and R. Ahlrichs, Theor. Chem. Acc., 1997, 97, 119–124.
98. F. Arrigoni, S. Mohamed Bouh, C. Elleouet, F. Y. Pétillon, P. Schollhammer, L. De Gioia and G. Zampella, Chem. –Eur. J., 2018, 24, 15036–15051.
99. A. E. Reed, R. B. Weinstock and F. Weinhold, J. Chem. Phys., 1985, 83, 735–746.
100. P. Löwdin, J. Chem. Phys., 1950, 18, 365–375.

---

Footnote

† Electronic supplementary information (ESI) available: Spectroscopic, DFT and additional experimental data. See DOI: 10.1039/d1sc06975f

---