Hydrogenase Biomimetics with Redox-Active Ligands: Synthesis, Structure, and Electrocatalytic Studies on [Fe 2 (CO) 4 ( κ 2 -dppn)( µ -edt)] (edt = Ethanedithiolate; dppn

: Addition of the bulky redox-active diphosphine 1,8-bis(diphenylphosphino)naphthalene (dppn) to [Fe 2 (CO) 6 ( µ -edt)] ( 1 ) (edt = 1,2-ethanedithiolate) affords [Fe 2 (CO) 4 ( κ 2 -dppn)( µ -edt)] ( 3 ) as the major product, together with small amounts of a P–C bond cleavage product [Fe 2 (CO) 5 { κ 1 -PPh 2 (1-C 10 H 7 )}( µ -edt)] ( 2 ). The redox properties of 3 have been examined by cyclic voltammetry and it has been tested as a proton-reduction catalyst. It undergoes a reversible reduction at E 1/2 = − 2.18 V and exhibits two overlapping reversible oxidations at E 1/2 = − 0.08 V and E 1/2 = 0.04 V. DFT calculations show that while the Highest Occupied Molecular Orbital (HOMO) is metal-centred (Fe–Fe σ -bonding), the Lowest Unoccupied Molecular Orbital (LUMO) is primarily ligand-based, but also contains an antibonding Fe–Fe contribution, highlighting the redox-active nature of the diphosphine. It is readily protonated upon addition of strong acids and catalyzes the electrochemical reduction of protons at E p = − 2.00 V in the presence of CF 3 CO 2 H. The catalytic current indicates that it is one of the most efﬁcient diiron electrocatalysts for the reduction of protons, albeit operating at quite a negative potential. [NBu 4 ][PF 6 ], scan rate 0.1 V · s − 1 , glassy carbon electrode, potential vs Fc + /Fc).

Structural characterization of both 2 and 3 was made on the basis of the crystal structure as shown in Figures 1 and 2, respectively. The former contains a diiron core coordinated by five carbonyls, a PPh2(1-C10H7), and a bridging edt ligand. The phosphine occupies an axial site, the Fe-P bond distance of 2.2442(6) Å being very similar to those reported for other [Fe2(CO)5(phosphine)(μ-edt)] complexes [26]. The Fe-Fe bond distance [2.5036(4) Å] is not affected by the phosphine substitution, being the same (within experimental error) as that in the parent hexacarbonyl 1 [2.5032(5) Å] [53]. Spectroscopic data indicate that the solid-state structure persists in solution. The carbonyl region of the IR spectrum shows a characteristic absorptions pattern for [Fe2(CO)5(phosphine)(μ-dithiolate)] complexes, while the 31 P{ 1 H} NMR spectrum displays a singlet at 61.1 ppm. The 1 H NMR spectrum is not very informative, but shows two multiplets centred at 1.58 Structural characterization of both 2 and 3 was made on the basis of the crystal structure as shown in Figures 1 and 2, respectively. The former contains a diiron core coordinated by five carbonyls, a PPh 2 (1-C 10 H 7 ), and a bridging edt ligand. The phosphine occupies an axial site, the Fe-P bond distance of 2.2442(6) Å being very similar to those reported for other [Fe 2 (CO) 5 (phosphine)(µ-edt)] complexes [26]. The Fe-Fe bond distance [2.5036(4) Å] is not affected by the phosphine substitution, being the same (within experimental error) as that in the parent hexacarbonyl 1 [2.5032(5) Å] [53]. Spectroscopic data indicate that the solid-state structure persists in solution. The carbonyl region of the IR spectrum shows a characteristic absorptions pattern for [Fe 2 (CO) 5 (phosphine)(µ-dithiolate)] complexes, while the 31 P{ 1 H} NMR spectrum displays a singlet at 61.1 ppm. The 1 H NMR spectrum is not very informative, but shows two multiplets centred at 1.58 and 0.65 ppm (each integrating to two protons) attributed to the methylene protons of the edt-bridge, in addition to naphthyl and phenyl proton resonances in the aromatic region.       The molecular structure of 3 is complicated by the existence of two independent molecules in the asymmetric unit linked via π-interactions of the naphthalene backbones [C(39)-C(40) and C(76)-C(77) C···C 3.271-3.633 Å]. Both independent molecules are similar (bond lengths and angles do not differ significantly), consisting of a diiron framework coordinated by four carbonyls, a chelating dppn, and an edt ligand, which bridges the diiron centre. The Fe(1)-Fe(2) bond distance [2.5377(8) Å] is slightly elongated as compared to that of the parent hexacarbonyl 1 [2.5032(5) Å] [53], probably to minimize the steric strain that accompanies dppn chelation. The dppn ligand is bound to Fe(1) occupying the apical and one of the basal coordination sites in the solid-state with a bite angle of 88.14(3) • . The Fe-P bond distances [Fe(1)-P(1) 2.1756 (9) and Fe(1)-P(2) 2.1970(9) Å] are slightly shorter than that observed in 2, but are within the range reported for related [Fe 2 (CO) 4 complexes [16][17][18][19][20][21][22][23][24][25][26][27]33]. Solution spectroscopic data of 3 are consistent with the solid-state structure. The IR spectrum shows three absorptions at 2021 s, 1950 m, and 1901 w cm −1 for the carbonyls, while the 31 P{ 1 H} NMR spectrum displays only a singlet at 68.4 ppm. The latter is associated with the well-studied interconversion of apical-basal and dibasal isomers [19,20,22,41]. The 1 H NMR spectrum shows two doublets at 1.87 and 1.29 (J 7.8 Hz) ppm, each of which integrated for two protons, for the methylene protons of the edt-bridge in addition to naphthyl and phenyl proton resonances in the aromatic region. The 13 C{ 1 H} NMR spectrum at room temperature shows a broad resonance at 206.8 ppm attributed to the three carbonyls on the unsubstituted iron centre, and a sharp triplet at 220.3 ppm (J P-C 19 Hz) for the fourth carbonyl. Upon cooling to −50 • C, the broad resonance splits into two sharp singlets at 215.7 and 206.6 ppm (ratio 2:1) consistent with freezing out of the trigonal rotation of the Fe(CO) 3 moiety. No change is seen to the other carbonyl resonance and the 31 P{ 1 H} NMR spectrum broadens only slightly at −50 • C showing that interconversion of axial and equatorial phosphorus sites remains rapid even at this temperature.

Protonation
Addition of a slight excess of HBF 4 ·Et 2 O to a CH 2 Cl 2 solution of 3 resulted in the immediate consumption of the latter as shown by IR spectroscopy. The appearance of new bands relating to two cationic hydrides was apparent as identified as dibasal [Fe 2 (CO) 4  . Over time, absorptions associated with 3bb diminished with concomitant growth of the υ(CO) bands for 3ab being associated with the well-known isomerization of the kinetically formed hydride isomer to that which is thermodynamically more stable [16][17][18][19][20][21][22]. The rate of isomerization was dependent on acid concentration, being slower at low concentrations. With ca. 5 stoichiometric amounts of acid, immediate removal of excess acid and volatiles followed by washing in Et 2 O allowed a clean IR spectrum of 3bb to be obtained. Attempts to observe the cationic hydrides via NMR spectroscopy were unsuccessful, but it is clear from the IR data that they are formed rapidly and cleanly.

Electrochemistry
Complex 3 has been investigated by cyclic voltammetry (CV) in MeCN, which shows a reversible reduction wave at E 1/2 = −2.18 V (∆E = 80 mV) and two overlapping reversible oxidative waves at E 1/2 = −0.08 V (∆E = 80 mV) and E 1/2 = 0.04 V (∆E = 60 mV) ( Figure 3). The CV does not show any discernable change when the scan rate is varied ( Figure S1). The reduction wave also shows good chemical reversibility (i p ox / i p red =~0.85), and plots of reductive and oxidative peak currents of this reversible process against the square root of the scan rate give straight lines in support of a diffusion-controlled process on the CV time scale ( Figure S2). The current function (i p / √ ν) associated with reduction deviates from linearity only at slow scan rates (<0.05 V/s), indicating that more than one electron may be involved in reduction at longer time scales ( Figure S3); otherwise, reduction of 3 is a one-electron process in MeCN. It is reduced at a similar potential to related [Fe 2 (CO) 4 (κ 2 -diphosphine)(µ-dithiolate)] complexes containing non redox inactive diphosphines [21,22,25], which suggests that reduction is diiron-centered (as confirmed by DFT). Overlap of the two oxidative waves indicates that either the electrons are coming from separate parts of the same molecule, otherwise, a considerable gap would be seen between the two oxidation peaks, or 3 undergoes solvolysis after first oxidation to form probably [Fe 2 (CO) 3 (MeCN)(κ 2 -dppn)(µ-edt)] + , which reduces at the second oxidation potential. We can rule out the solvolysis because it would render the first oxidation wave irreversible. The nature of the HOMO and LUMO in 3 ( Figure 4) was evaluated by DFT in order to better understand the role the dppn ligand plays, if any, in the observed reduction of proton to H2 (vide infra). The structure of 3 was optimized, and the geometry-optimized structure (not shown) revealed excellent agreement with the experimentally determined structure depicted in Figure 2. The HOMO for 3 (left) is localized over the two iron centers and is best viewed as an in-phase Fe-Fe bond. The LUMO for 3 (right) exhibits the expected antibonding Fe-Fe interaction found in related derivatives along with a significant orbital contribution from the naphthalene π system, whose π* nodal properties are evident. The composition of the LUMO is best described as ligand-based that contains an antibonding Fe-Fe contribution. The nodal pattern of the naphthalene π* in the LUMO of 3 is comparable to the LUMO computed for the cluster [Fe4(CO)10(κ 2 -dppn)(μ4-O)] [54]. The dppn π* system makes a much smaller contribution to the LUMO in the Fe4 cluster relative to the antibonding metallic core. The enhanced catalytic behavior exhibited by 3, vis-à-vis related derivatives of [Fe2(CO)4(κ 2 -diphosphine)(μ-dithiolate)] whose ancillary diphosphine ligand does not contribute to the LUMO, signals the importance of the redox-active dppn ligand in promoting effective proton reduction.

Catalysis
Complex 3 was tested as a proton reduction catalyst in the presence of CF3CO2H in MeCN solvent (Figures 5 and S4). Selected CVs recorded upon sequential addition of 1-7 molar equivalents The nature of the HOMO and LUMO in 3 ( Figure 4) was evaluated by DFT in order to better understand the role the dppn ligand plays, if any, in the observed reduction of proton to H 2 (vide infra). The structure of 3 was optimized, and the geometry-optimized structure (not shown) revealed excellent agreement with the experimentally determined structure depicted in Figure 2. The HOMO for 3 (left) is localized over the two iron centers and is best viewed as an in-phase Fe-Fe bond. The LUMO for 3 (right) exhibits the expected antibonding Fe-Fe interaction found in related derivatives along with a significant orbital contribution from the naphthalene π system, whose π* nodal properties are evident. The composition of the LUMO is best described as ligand-based that contains an antibonding Fe-Fe contribution. The nodal pattern of the naphthalene π* in the LUMO of 3 is comparable to the LUMO computed for the cluster [Fe 4 (CO) 10 (κ 2 -dppn)(µ 4 -O)] [54]. The dppn π* system makes a much smaller contribution to the LUMO in the Fe 4 cluster relative to the antibonding metallic core. The enhanced catalytic behavior exhibited by 3, vis-à-vis related derivatives of [Fe 2 (CO) 4 (κ 2 -diphosphine)(µ-dithiolate)] whose ancillary diphosphine ligand does not contribute to the LUMO, signals the importance of the redox-active dppn ligand in promoting effective proton reduction. The nature of the HOMO and LUMO in 3 ( Figure 4) was evaluated by DFT in order to better understand the role the dppn ligand plays, if any, in the observed reduction of proton to H2 (vide infra). The structure of 3 was optimized, and the geometry-optimized structure (not shown) revealed excellent agreement with the experimentally determined structure depicted in Figure 2. The HOMO for 3 (left) is localized over the two iron centers and is best viewed as an in-phase Fe-Fe bond. The LUMO for 3 (right) exhibits the expected antibonding Fe-Fe interaction found in related derivatives along with a significant orbital contribution from the naphthalene π system, whose π* nodal properties are evident. The composition of the LUMO is best described as ligand-based that contains an antibonding Fe-Fe contribution. The nodal pattern of the naphthalene π* in the LUMO of 3 is comparable to the LUMO computed for the cluster [Fe4(CO)10(κ 2 -dppn)(μ4-O)] [54]. The dppn π* system makes a much smaller contribution to the LUMO in the Fe4 cluster relative to the antibonding metallic core. The enhanced catalytic behavior exhibited by 3, vis-à-vis related derivatives of [Fe2(CO)4(κ 2 -diphosphine)(μ-dithiolate)] whose ancillary diphosphine ligand does not contribute to the LUMO, signals the importance of the redox-active dppn ligand in promoting effective proton reduction.

Catalysis
Complex 3 was tested as a proton reduction catalyst in the presence of CF3CO2H in MeCN solvent (Figures 5 and S4). Selected CVs recorded upon sequential addition of 1-7 molar equivalents

Catalysis
Complex 3 was tested as a proton reduction catalyst in the presence of CF 3 CO 2 H in MeCN solvent ( Figure 5 and Figure S4). Selected CVs recorded upon sequential addition of 1-7 molar equivalents of acid are shown in Figure 5. Two new reduction peaks at E p = −2.00 and −2.11 V are seen after addition of one molar equivalent of CF 3 CO 2 H. The reduction potential shows ca. 0.2 V positive shift after addition of CF 3 CO 2 H, suggesting the generation of cationic hydride. A small peak is also observed at the reduction potential of 3 at a very low acid concentration (≤1 equivalent), which disappears at higher acid concentrations (≥2 equivalents). Since CF 3 CO 2 H (pK a = 12.6 in MeCN) is a much weaker acid than HBF 4 ·Et 2 O (pK a = −0.1 in MeCN) [55], the latter is used for the protonation studies, which protonate 3 instantaneously, and we assume that protonation of 3 is slow at low CF 3 CO 2 H concentration. The height of the peaks at E p = −2.00 and −2.11 V increase with acid concentration and is characteristic of electrocatalytic proton reduction by this complex at these potentials. It appears that 3 enters into the catalytic cycle via a chemical step (protonation), followed by an electrochemical reduction, which generates the neutral 35-electron complex [Fe 2 (CO) 4 (µ-H)(κ 2 -dppn)(µ-edt)]. This neutral hydride can either protonate or undergo a further reduction before a second protonation to liberate hydrogen. The CVs also show curve-crossing i.e., build-up of reduction current on the return scan (ca. −1.80 V), in the presence of acid. This indicates that a more easily reducible product or intermediate is formed during catalysis, most probably via a slow chemical reaction [56][57][58][59][60]. This product or intermediate is sufficiently stable and its concentration increases as the concentration of acid is increased and it diffuses back to the electrode to undergo reduction at a more positive potential [56][57][58][59][60]. A plot of the catalytic current/noncatalytic current ratio (i cat /i p ) against the concentration of acid is shown in Figure 6 for the first catalytic wave. The i cat /i p value increases to 20 after addition of 10 equivalents of CF 3 CO 2 H. To our knowledge, very few biomimetic diiron systems developed as models of the active site of [FeFe]-hydrogenases show such a high i cat /i p value [14,30,31,61]. For example, [Fe 2 (CO) 5 (κ 1 -IMes)(µ-pdt)] [IMes = 1,3-bis(2,4,6-trimethylphenyl)imidazol-2-ylidene], which undergoes a two-electron reduction at −1.90 V vs. SCE, shows an i cat /i p value of~4 after addition of 10 molar equivalents of CH 3 CO 2 H [61,62]. The i cat /i p value serves as a measure of the catalyst efficiency [61,63], and the values observed for 3 indicate that it is very efficient for the reduction of protons to H 2 , although it operates at a very negative potential (−2.00 V). The highest i cat /i p value observed to date for electrocatalytic proton reduction is 38, shown by the nickel complex [(P Ph 2 N Ph )Ni] 2+ (P Ph 2 N Ph = 1,3,6-triphenyl-1-aza-3,6-diphosphacycloheptane) [64].
Inorganics 2018, 6, x 6 of 13 of acid are shown in Figure 5. Two new reduction peaks at Ep = −2.00 and −2.11 V are seen after addition of one molar equivalent of CF3CO2H. The reduction potential shows ca. 0.2 V positive shift after addition of CF3CO2H, suggesting the generation of cationic hydride. A small peak is also observed at the reduction potential of 3 at a very low acid concentration (≤1 equivalent), which disappears at higher acid concentrations (≥2 equivalents). Since CF3CO2H (pKa = 12.6 in MeCN) is a much weaker acid than HBF4•Et2O (pKa = −0.1 in MeCN) [55], the latter is used for the protonation studies, which protonate 3 instantaneously, and we assume that protonation of 3 is slow at low CF3CO2H concentration. The height of the peaks at Ep = −2.00 and −2.11 V increase with acid concentration and is characteristic of electrocatalytic proton reduction by this complex at these potentials. It appears that 3 enters into the catalytic cycle via a chemical step (protonation), followed by an electrochemical reduction, which generates the neutral 35-electron complex [Fe2(CO)4(μ-H)(κ 2 -dppn)(μ-edt)]. This neutral hydride can either protonate or undergo a further reduction before a second protonation to liberate hydrogen. The CVs also show curve-crossing i.e., build-up of reduction current on the return scan (ca. −1.80 V), in the presence of acid. This indicates that a more easily reducible product or intermediate is formed during catalysis, most probably via a slow chemical reaction [56][57][58][59][60]. This product or intermediate is sufficiently stable and its concentration increases as the concentration of acid is increased and it diffuses back to the electrode to undergo reduction at a more positive potential [56][57][58][59][60]. A plot of the catalytic current/noncatalytic current ratio (icat/ip) against the concentration of acid is shown in Figure 6 for the first catalytic wave.

General Procedures
All reactions were carried out under a nitrogen atmosphere using standard Schlenk techniques. Reagent grade solvents were dried by the standard procedures and were freshly distilled prior to use. [Fe2(CO)6(μ-edt)] (1) was synthesized according to the literature procedure [65]. IR spectra were recorded on a Shimadzu FTIR 8101 spectrophotometer (Shimadzu Corporation, Kyoto, Japan) while the NMR spectra were recorded on a Bruker DPX 400 instrument (Billerica, MA, USA). The chemical shifts were referenced to residual solvent resonances or external 85% H3PO4 in 1 H and 31 P spectra, respectively. Elemental analyses were performed in the Microanalytical Laboratories of Wazed Miah Science Research Centre at Jahangirnagar University (Dhaka, Bangladesh). Preparative thin layer chromatography was carried out on 1 mm plates prepared from silica gel GF254 (type 60, E. Merck) at Jahangirnagar University.

Synthesis
Me3NO (21 mg, 0.279 mmol) was added to a MeCN solution (15 mL) of 1 (100 mg, 0.269 mmol) and dppn (134 mg, 0.270 mmol), and the mixture was heated to reflux for 1.5 h. After cooling to room temperature, volatiles were removed under reduced pressure and the residue chromatographed by Thin Layer Chromatography (TLC) on silica gel. Elution with hexane/CH2Cl2

General Procedures
All reactions were carried out under a nitrogen atmosphere using standard Schlenk techniques. Reagent grade solvents were dried by the standard procedures and were freshly distilled prior to use. [Fe 2 (CO) 6 (µ-edt)] (1) was synthesized according to the literature procedure [65]. IR spectra were recorded on a Shimadzu FTIR 8101 spectrophotometer (Shimadzu Corporation, Kyoto, Japan) while the NMR spectra were recorded on a Bruker DPX 400 instrument (Billerica, MA, USA). The chemical shifts were referenced to residual solvent resonances or external 85% H 3 PO 4 in 1 H and 31 P spectra, respectively. Elemental analyses were performed in the Microanalytical Laboratories of Wazed Miah Science Research Centre at Jahangirnagar University (Dhaka, Bangladesh). Preparative thin layer chromatography was carried out on 1 mm plates prepared from silica gel GF254 (type 60, E. Merck) at Jahangirnagar University.

Protonation Experiments
To a CH 2 Cl 2 solution (ca. 2 mL) of 3 (made by dissolving 4.5 mg, 0.005 mmol), 2 molar equivalents of HBF 4 ·Et 2 O were added. The resultant acid-containing solution was immediately transferred to an IR cell and monitored over time.

X-Ray Crystallography
Single crystals of 2 and 3 suitable for X-ray diffraction were grown by slow diffusion of hexane into a CH 2 Cl 2 solution at 4 • C. All geometric and crystallographic data were collected at 150(2) K on a Bruker SMART APEX CCD diffractometer (Billerica, MA, USA) using Mo-Kα radiation (λ = 0.71073 Å) [66]. Data reduction and integration were carried out with SAINT+ [67], and absorption corrections were applied using the program SADABS [68]. The structures were solved by direct methods and refined by full-matrix least squares on F 2 [69]. All non-hydrogen atoms were refined anisotropically. The hydrogen atoms were placed in the calculated positions and their thermal parameters were linked to those of the atoms to which they were attached (riding model). The SHELXTL PLUS V6.10 program package was used for structure solution and refinement [69]. Final difference maps did not show any residual electron density of stereochemical significance. The details of the data collection and structure refinement are given in Table A1.

Electrochemical Studies
Electrochemistry was carried out in deoxygenated MeCN with 0.1 M TBAPF 6 as the supporting electrolyte. The working electrode was a 3 mm diameter glassy carbon electrode that was polished with 0.3 µm alumina slurry prior to each scan. The counter electrode was a Pt wire and the quasi-reference electrode was a silver wire. All CVs were referenced to the Fc + /Fc redox couple. An Autolab potentiostat (EcoChemie, Utrecht, Netherlands) was used for all electrochemical measurements. Catalysis studies were carried out by adding equivalents of CF 3 CO 2 H (Sigma-Aldrich, St. Louis, MO, USA).

Computational Methodology and Modeling Details
The DFT calculations were performed with the hybrid meta exchange-correlation functional M06 [70], as implemented by the Gaussian 09 program package [71]. The Fe [72] atoms were described by Stuttgart-Dresden effective core potentials (ecp) and an Stuttgart-Dresden (SDD) basis set, while a 6-31G(d') basis set was employed for all second row elements, and a 6-31G* basis set utilized for third row elements. The computed DFT structure for 3 represents a fully optimized ground state based on the positive eigenvalues displayed by the analytical Hessian.

Conclusions
The diiron-dithiolate [Fe 2 (CO) 4 (κ 2 -dppn)(µ-edt)] (3) containing a chelating dppn ligand has been synthesized from the reaction between [Fe 2 (CO) 6 (µ-edt)] (1) and dppn, together with a side product [Fe 2 (CO) 5 {κ 1 -PPh 2 (1-C 10 H 7 )}(µ-edt)] (2) resulting from P-C bond cleavage. Both 2 and 3 have been characterized by single-crystal X-ray diffraction analysis and structural features are unexceptional. DFT calculations on 3 show that while the HOMO is based exclusively at the diiron centre, the LUMO has a significant contribution from the naphthalene π-system, showing that the dppn ligand is an integral part of the redox system. Cyclic voltammetry reveals that it undergoes a reversible part ligand based reduction and displays two overlapping reversible metal-centred oxidations in MeCN. Control experiments confirm that 3 is readily protonated to give the cationic hydrides tentatively identified as dibasal [Fe 2 (CO) 4 (µ-H)(κ 2 -dppn)(µ-edt)][BF 4 ] (3bb) and apical-basal [Fe 2 (CO) 4 (µ-H)(κ 2 -dppn)(µ-edt)][BF 4 ] (3ab); the former hydride is less stable and transforms into the latter hydride over time. Complex 3 has been examined as an electrocatalyst for proton-reduction. Analysis of electrocatalytic data indicates that it operates at a very negative potential (ca. −2 V), showing that even after protonation, reduction is primarily ligand based. However, that it is an active catalyst shows that there must be electronic communication between the dppn and diiron centres. Indeed, 3 is one of the most efficient diiron biomimetics reported to date and thus intramolecular electron-transfer within the cationic dihydride must be efficient. Thus, the redox-active dppn ligand plays a critical role in the observed electrochemical proton-reduction. The exact nature of this remains unknown and in future work we will aim to better understand and exploit this electron coupling.