Assembly and Redox-Rich Hydride Chemistry of an Asymmetric Mo2S2 Platform

Although molybdenum sulfide materials show promise as electrocatalysts for proton reduction, the hydrido species proposed as intermediates remain poorly characterized. We report herein the synthesis, reactions and spectroscopic properties of a molybdenum-hydride complex featuring an asymmetric Mo2S2 core. This molecule displays rich redox chemistry with electrochemical couples at E½ = −0.45, −0.78 and −1.99 V vs. Fc/Fc+. The corresponding hydrido-complexes for all three redox levels were isolated and characterized crystallographically. Through an analysis of solid-state bond metrics and DFT calculations, we show that the electron-transfer processes for the two more positive couples are centered predominantly on the pyridinediimine supporting ligand, whereas for the most negative couple electron-transfer is mostly Mo-localized.


Introduction
Efficient electro or photocatalytic reduction of protons to produce H 2 gas remains an important challenge [1][2][3][4][5]. Not only is H 2 a potential clean fuel, but is by far the most used chemical reductant [6]. H 2 is typically produced industrially by the steam reforming of methane [7]. Although this process utilizes a nonrenewable hydrocarbon feedstock and generates large volumes of greenhouse gases, the low efficiency of current water splitting technologies makes steam reforming by far the most economical process [5]. As a result, a tremendous amount of research has revolved around discovery and optimization of new materials for the hydrogen evolution reaction (HER) [1][2][3][4][5]. Some of the most promising heterogeneous catalysts for this reaction are molybdenum sulfides (MoS x ),which exhibit comparable overpotentials and current densities to Pd/Pt electrodes but are significantly cheaper [8][9][10]. Although the mechanism of proton reduction by MoS x materials is not fully understood, calculations implicate molybdenum-hydrido intermediates that are protonated to give H 2 [11][12][13][14]. A number of molecular analogues for the surface interface of MoS x materials have been shown to be active HER catalysts, although Mo-H intermediates, if they are formed, have not been well-characterized [15][16][17][18][19][20][21][22]. Indeed, molecular Mo-H complexes featuring sulfur-donor ligands, particularly sulfides, are uncommon [23][24][25][26][27][28][29][30][31][32][33][34][35][36][37]. Consequently, we have been interested in preparing molybdenum-hydride complexes with sulfide ligands and studying their chemistry and spectroscopy. We report, herein, preparation of hydrido complexes featuring an Mo 2 S 2 core and an examination of their redox properties, reactions and electronic structure.

Synthesis of Complexes
In approaching this chemistry, we targeted Mo 2 S 2 complexes for which each Mo site would reside in a distinct coordination environment. We hoped this would facilitate selective installation of a hydrido ligand at a single Mo, and so greatly simplify reactivity and spectroscopic studies. The Mo 2 S 2 motif is common in the literature; however, all such complexes are approximately symmetric and, typically, coordinatively saturated [38].
Molecules 2020, 25, x FOR PEER REVIEW 2 of 14 In approaching this chemistry, we targeted Mo2S2 complexes for which each Mo site would reside in a distinct coordination environment. We hoped this would facilitate selective installation of a hydrido ligand at a single Mo, and so greatly simplify reactivity and spectroscopic studies. The Mo2S2 motif is common in the literature; however, all such complexes are approximately symmetric and, typically, coordinatively saturated [38].
Molecules 2020, 25, 3090 3 of 14 A 1D-1 H-1 H NOESY experiment with selective irradiation of the hydride resonance showed a strong cross-peak to the C 5 H 5 signal, indicating that the hydrido ligand bound the Cp-substituted Mo ( Figure  S2B). Positive-mode liquid injected field desorption (LIFDI) mass spectrometry showed a peak at m/z 807.1475 (calcd. 807.1471) for the molecular ion, [4] + . As is not uncommon [41] the FT-IR spectrum of 4 contained no bands that could be unambiguously assigned to an Mo-H stretch (nor could any be identified in (Fourier transform infrared) FT-IR spectra of its reduced/oxidized congeners, vide infra; see Figures S3 and S4). Further, indirect evidence for the existence of a Mo-H moiety in 4 was provided by its rapid reaction with CDCl 3 to produce CDHCl 2 (see Figure S5) [42].
Given the potential relationship between Mo-H species and heterogenous MoS x HER catalysts (see Section 1), we were particularly interested to explore the redox chemistry of 4. The cyclic voltammogram (CV) recorded for 4 in 0.1M Na[BAr F 4 ] in THF is shown in Figure 1. On the CV timescale, complex 4 displayed rich redox chemistry, with reversible processes at  [43] we expected that it would be problematic to assign these couples to localized electron-transfer processes from CV data alone. We wondered, then, if [4]  assigned to an Mo-H stretch (nor could any be identified in (Fourier transform infrared) FT-IR spectra of its reduced/oxidized congeners, vide infra; see Figures S3 and S4). Further, indirect evidence for the existence of a Mo-H moiety in 4 was provided by its rapid reaction with CDCl3 to produce CDHCl2 (see Figure S5) [42]. Given the potential relationship between Mo-H species and heterogenous MoSx HER catalysts (see Section 1), we were particularly interested to explore the redox chemistry of 4. The cyclic voltammogram (CV) recorded for 4 in 0.1M Na[BAr F 4] in THF is shown in Figure 1 [43] we expected that it would be problematic to assign these couples to localized electron-transfer processes from CV data alone. We wondered, then, if [4]   The redox and protonation chemistry of 4 is summarized in Scheme 2. First, reduction of 4 using metallic Na in diethyl ether followed by encapsulation of Na + by [2.2.2]Cryptand (crypt-2.2.2) gave [4][Na(crypt-2.2.2)] as a brown, crystalline solid in 84% yield. Second, the dark-blue one-electron oxidized complex [4][BAr F 4] was prepared from 4 in 70% yield by addition of 1 equivalent of either [Ph3C][BAr F 4] or [Fc][BAr F 4]. In the case of the former reagent, peaks for Gomberg's dimer (the oneelectron reduction product of Ph3C + ) were observed in the 1 H-NMR spectrum of the crude reaction mixture. In the absence of air and moisture, both the one-electron oxidized and reduced clusters were stable for prolonged periods (>1 week) at room temperature (RT) in both the solid state and in solution. Both had S = 1/2 ground-states as indicated by their electron paramagnetic resonance (EPR) spectra (see Section 2.2) and solution state magnetic moments of ~2 μB.
In contrast to the syntheses of [4] +/− , isolation of the two-electron oxidized complex, [4] 2+ , proved more difficult. For example, addition of 2 equivalents of [Fc][BAr F 4] to complex 4 in a variety of solvents gave intractable mixtures; the crude samples showed at least 3 diamagnetic Mo-H species. We speculated that the presumably highly electrophilic dication [4]   The redox and protonation chemistry of 4 is summarized in Scheme 2. First, reduction of 4 using metallic Na in diethyl ether followed by encapsulation of Na + by [2.2.2]Cryptand (crypt-2.2.2) gave [4][Na(crypt-2.2.2)] as a brown, crystalline solid in 84% yield. Second, the dark-blue one-electron oxidized complex [4] . In the case of the former reagent, peaks for Gomberg's dimer (the one-electron reduction product of Ph 3 C + ) were observed in the 1 H-NMR spectrum of the crude reaction mixture. In the absence of air and moisture, both the one-electron oxidized and reduced clusters were stable for prolonged periods (>1 week) at room temperature (RT) in both the solid state and in solution. Both had S = 1/2 ground-states as indicated by their electron paramagnetic resonance (EPR) spectra (see Section 2.2) and solution state magnetic moments of~2 µB.
In contrast to the syntheses of [4]  In addition to its redox chemistry, we were also interested in examining the reaction of 4 with proton sources. Addition of excess [Ph2NH2][BF4] to a suspension of 4 in MeCN resulted in rapid evolution of H2 gas (by 1 H-NMR spectroscopy) and formation of an intensely purple solution. From this mixture the dark-purple dication [(PDI)Mo(MeCN)(μ-S)2Mo(η 5-Cp)(MeCN)][BF4]2 (5) was isolated in 90% yield. As with [4] +/− , complex 5 had an S = 1/2 ground-state as shown by EPR spectroscopy (see Section 2.2) and a solution state magnetic moment of 2.0 ± 0.1 μB. In contrast to the neutral 4, reaction of the anionic complex [4] − with [Ph2NH2][BF4] gave a complex mixture of products. Repeated recrystallizations ultimately separated a small quantity of red-orange crystals, which were identified by XRD as a Mo4S4 cubane (6), which featured a partially hydrogenated PDI ligand (see Figure S1E). The variable protic stability of the Mo2S2 core, when in different formal redox states, will be an important consideration when optimizing these types of systems as catalysts for the HER.

EPR Spectroscopy
As expected, the paramagnetic, half-integer spin [4] +/− and 5 were all EPR-active. Continuous wave (CW) X-band EPR spectra recorded on dilute, frozen glasses of these complexes at 100K showed complex envelopes of peaks with extensive hyperfine structure centered at g ≈ 2, indicative of an S = 1/2 ground state for all three species (Figure 2). The most likely origins of these hyperfine interactions are coupling to a combination of 95/97 Mo, 14    In addition to its redox chemistry, we were also interested in examining the reaction of 4 with proton sources. Addition of excess [Ph 2 NH 2 ][BF 4 ] to a suspension of 4 in MeCN resulted in rapid evolution of H 2 gas (by 1 H-NMR spectroscopy) and formation of an intensely purple solution. From this mixture the dark-purple dication [(PDI)Mo(MeCN)(µ-S) 2 Mo(η 5-Cp)(MeCN)][BF 4 ] 2 (5) was isolated in 90% yield. As with [4] +/− , complex 5 had an S = 1/2 ground-state as shown by EPR spectroscopy (see Section 2.2) and a solution state magnetic moment of 2.0 ± 0.1 µ B . In contrast to the neutral 4, reaction of the anionic complex [4] − with [Ph 2 NH 2 ][BF 4 ] gave a complex mixture of products. Repeated recrystallizations ultimately separated a small quantity of red-orange crystals, which were identified by XRD as a Mo 4 S 4 cubane (6), which featured a partially hydrogenated PDI ligand (see Figure S1E). The variable protic stability of the Mo 2 S 2 core, when in different formal redox states, will be an important consideration when optimizing these types of systems as catalysts for the HER.

EPR Spectroscopy
As expected, the paramagnetic, half-integer spin [4] +/− and 5 were all EPR-active. Continuous wave (CW) X-band EPR spectra recorded on dilute, frozen glasses of these complexes at 100 K showed complex envelopes of peaks with extensive hyperfine structure centered at g ≈ 2, indicative of an S = 1/2 ground state for all three species (Figure 2). The most likely origins of these hyperfine interactions are coupling to a combination of 95/97 Mo, 14

Crystallography
Single crystal XRD studies were performed on the four hydride complexes and also complexes 1, 3 and 5. Views from the crystal structures of the hydride complexes are presented in Figure 3; structures for 1, 3, 5 and connectivity assignments for 2 and 6 are shown in Figure S1. Selected experimental bond lengths and angles are summarized in Table 1. Electron density for the hydrido ligands was located in the difference map for all of the hydride complexes. PDIs are prototypical redox active ligands, and electron transfer processes for their complexes are frequently PDI-localized [44,45]. The extent to which the C = Nim (im = imine) and Cpy-Cim (py = pyridine) bonds lengthen and shorten, respectively, has proved a useful proxy for a qualitative assessment of the extent to which the PDI has been reduced. [46] PDI ligands are known to be able to accept up to 3 electrons to give the (formal) radical trianion, [PDI • ] 3- [46]. With this in mind, inspection of the structural data revealed several key trends. First, the average Mo-S bond distances changed only slightly (≤0.01 Å) between [4-OTf] + , [4]

Crystallography
Single crystal XRD studies were performed on the four hydride complexes and also complexes 1, 3 and 5. Views from the crystal structures of the hydride complexes are presented in Figure 3; structures for 1, 3, 5 and connectivity assignments for 2 and 6 are shown in Figure S1. Selected experimental bond lengths and angles are summarized in Table 1. Electron density for the hydrido ligands was located in the difference map for all of the hydride complexes. increased and decreased, respectively, by ~0.05 Å. Second, the C = Nim and Cpy-Cim bonds expanded and contracted, respectively, on moving from [4-OTf] + to [4] + to 4 by roughly 0.03 Å per oxidation level. In contrast, metrics within the PDI system changed little between 4 and [4] − ; indeed, the Cpy-Cim bonds lengthened very slightly. Overall, we hypothesise that upon sequential reduction of [4-OTf] + , the first two electrons are placed in MOs of mostly PDI-π* character, and the third into an MO of mostly metal character. This idea was explored further using computational methods (Section 2.4).

Density Functional Theory Calculations
In order to more thoroughly understand the electronic structure of these molecules, we turned to density functional theory (DFT). All calculations were performed on truncated structures obtained from XRD coordinates with the diisopropylphenyl (Dipp) and methyl groups of the PDI ligand replaced with protons. The unrestricted Kohn-Sham HOMOs and HOMO-1s calculated for the four hydride complexes are shown in Figure 4. The calculated orbital picture agrees well with the electronic structure assignments suggested by the solid state bond metrics. First, for [4-OTf] + the doubly occupied HOMO is of predominantly Mo(d)-character and is essentially nonbonding with respect to the entire σ framework. The low-lying PDI-π* MO (not shown) remains unfilled. For complex [4] + , the singly-occupied HOMO is strongly localized (65%) on the PDI ligand with some π*-Mo(d) mixing. The orbital picture upon one-electron reduction to [4] − remains similar with the HOMO now doubly-occupied. In contrast, the singly-occupied HOMO for the anion [4] − is of predominantly (PDI)Mo(d)-character (54%). This MO is largely nonbonding with respect to the PDI ligand with minor Mo-Nim(σ*) contributions. This would account for these bonds being slightly longer in [4] − compared to 4. Finally, the HOMO-1 for [4] -is doubly-occupied and localized mostly on the PDI-π* system (61%). PDIs are prototypical redox active ligands, and electron transfer processes for their complexes are frequently PDI-localized [44,45]. The extent to which the C = N im (im = imine) and C py -C im (py = pyridine) bonds lengthen and shorten, respectively, has proved a useful proxy for a qualitative assessment of the extent to which the PDI has been reduced. [46] PDI ligands are known to be able to accept up to 3 electrons to give the (formal) radical trianion, [PDI • ] 3− [46]. With this in mind, inspection of the structural data revealed several key trends. First, the average Mo-S bond distances changed only slightly (≤0.01 Å) between [4-OTf] + , [4] + and 4. In contrast, for the anion [4] − , the Mo-S bonds at both Mo sites were longer than for the three more oxidized complexes by~0.03 Å. In addition, upon one-electron reduction of 4 to [4] − , the Mo-N im and Mo-N py distances sharply increased and decreased, respectively, by~0.05 Å. Second, the C = N im and C py -C im bonds expanded and contracted, respectively, on moving from [4-OTf] + to [4] + to 4 by roughly 0.03 Å per oxidation level. In contrast, metrics within the PDI system changed little between 4 and [4] − ; indeed, the C py -C im bonds lengthened very slightly. Overall, we hypothesise that upon sequential reduction of [4-OTf] + , the first two electrons are placed in MOs of mostly PDI-π* character, and the third into an MO of mostly metal character. This idea was explored further using computational methods (Section 2.4).

Density Functional Theory Calculations
In order to more thoroughly understand the electronic structure of these molecules, we turned to density functional theory (DFT). All calculations were performed on truncated structures obtained from XRD coordinates with the diisopropylphenyl (Dipp) and methyl groups of the PDI ligand replaced with protons. The unrestricted Kohn-Sham HOMOs and HOMO-1s calculated for the four hydride complexes are shown in Figure 4. The calculated orbital picture agrees well with the electronic structure assignments suggested by the solid state bond metrics. First, for [4-OTf] + the doubly occupied HOMO is of predominantly Mo(d)-character and is essentially nonbonding with respect to the entire σ framework. The low-lying PDI-π* MO (not shown) remains unfilled. For complex [4] + , the singly-occupied HOMO is strongly localized (65%) on the PDI ligand with some π*-Mo(d) mixing. The orbital picture upon one-electron reduction to [4] − remains similar with the HOMO now doubly-occupied. In contrast, the singly-occupied HOMO for the anion [4] − is of predominantly (PDI)Mo(d)-character (54%). This MO is largely nonbonding with respect to the PDI ligand with minor Mo-N im (σ*) contributions. This would account for these bonds being slightly longer in [4] − compared to 4. Finally, the HOMO-1 for [4] − is doubly-occupied and localized mostly on the PDI-π* system (61%).

Discussion
This work illustrates a new approach to the synthesis of bimetallic complexes featuring an Mo2S2 core. Of note, protonation of 4 to give 5 was somewhat unexpected and warrants some comment. We had hypothesized that, given two-electron oxidation of 4 affords the isolable [4-OTf] + , reaction of 4 with [Ph2NH2][BF4] would similarly give [4-BF4] + via sequential two-electron steps; i.e., protonation of the Mo-H atom followed by protonation of the Cp-bound Mo. In contrast, H2 evolution from 4 to give 5 proceeds through, formally, one 2-electron and one single-electron step. In an effort to understand these processes, single equivalents of different anilinium acids were added to 4. Although these reactions invariably gave mixtures, 1 H-NMR analysis revealed that the major product was diamagnetic and still contained a MoH moiety. Importantly, a quartet at δ ~5.5 (J = 7 Hz) of integration 1 relative to the MoH signal was consistently observed. We attribute this signal to the methine proton of a protonated imino carbon atom, with the hyperfine pattern due to coupling with the adjacent CH3 group. Addition of further protons to these mixtures gave 5 without discernable intermediate(s). We suggest the PDI-bound Mo in 4 is first protonated and that the resulting hydrido species rapidly transfers hydrogen to the imino-carbon [39,47]. How this complex is then protonated to give 5 is not clear, although bimolecular, homolytic H2 evolution from a MoH species seems likely. It is probable that similar ligand-protonation pathways give rise to the partially hydrogenated PDI observed in the cubane 6.
Taken together, the XRD and computational data presented herein permit formulation of Lewis structure representations for the four hydride complexes. These are shown in Figure 5. In summary, on moving from [4-OTf] + to [4] + to 4 the PDI ligand is formally reduced in one-electron steps from PDI 0 , to PDI •-to the dianion PDI 2− . Further reduction to give [4] -places an electron in an MO of mostly Mo(d)-character with the PDI remaining as PDI 2− .

Discussion
This work illustrates a new approach to the synthesis of bimetallic complexes featuring an Mo 2 S 2 core. Of note, protonation of 4 to give 5 was somewhat unexpected and warrants some comment. We had hypothesized that, given two-electron oxidation of 4 affords the isolable [4-OTf] + , reaction of 4 with [Ph 2 NH 2 ][BF 4 ] would similarly give  ] + via sequential two-electron steps; i.e., protonation of the Mo-H atom followed by protonation of the Cp-bound Mo. In contrast, H 2 evolution from 4 to give 5 proceeds through, formally, one 2-electron and one single-electron step. In an effort to understand these processes, single equivalents of different anilinium acids were added to 4. Although these reactions invariably gave mixtures, 1 H-NMR analysis revealed that the major product was diamagnetic and still contained a MoH moiety. Importantly, a quartet at δ~5.5 (J = 7 Hz) of integration 1 relative to the MoH signal was consistently observed. We attribute this signal to the methine proton of a protonated imino carbon atom, with the hyperfine pattern due to coupling with the adjacent CH 3 group. Addition of further protons to these mixtures gave 5 without discernable intermediate(s). We suggest the PDI-bound Mo in 4 is first protonated and that the resulting hydrido species rapidly transfers hydrogen to the imino-carbon [39,47]. How this complex is then protonated to give 5 is not clear, although bimolecular, homolytic H 2 evolution from a MoH species seems likely. It is probable that similar ligand-protonation pathways give rise to the partially hydrogenated PDI observed in the cubane 6.
Taken together, the XRD and computational data presented herein permit formulation of Lewis structure representations for the four hydride complexes. These are shown in Figure 5. In summary, on moving from [4-OTf] + to [4] + to 4 the PDI ligand is formally reduced in one-electron steps from PDI 0 , to PDI •to the dianion PDI 2− . Further reduction to give [4] − places an electron in an MO of mostly Mo(d)-character with the PDI remaining as PDI 2− . cubane 6.
Taken together, the XRD and computational data presented herein permit formulation of Lewis structure representations for the four hydride complexes. These are shown in Figure 5. In summary, on moving from [4-OTf] + to [4] + to 4 the PDI ligand is formally reduced in one-electron steps from PDI 0 , to PDI •-to the dianion PDI 2− . Further reduction to give [4] -places an electron in an MO of mostly Mo(d)-character with the PDI remaining as PDI 2− .

Conclusions
A series of site-differentiated hydrido complexes featuring an Mo2S2 core were prepared. These molecules represent rare examples of Mo-H complexes bound by sulfido ligands and are related to

Conclusions
A series of site-differentiated hydrido complexes featuring an Mo 2 S 2 core were prepared. These molecules represent rare examples of Mo-H complexes bound by sulfido ligands and are related to species proposed as intermediates in the HER catalyzed by MoS x materials. Reactivity, XRD and computational studies showed that redox and protonation events for these complexes were mostly PDI-centered. This work further highlights the utility of PDI ligands as proton and electron-reservoirs. We are currently examining this class of complexes as electrocatalysts for proton reduction.

General Considerations
Unless stated otherwise, all compounds were purchased from commercial sources and used without further purification. Solvents were dried and deoxygenated by argon sparge followed by passage through an activated alumina column and were stored over 4 Å molecular sieves. DBU was dried over CaH 2 and distilled in vacuo. All Mo complexes, with the exception of (PDI)MoCl 3 , which could be handled in air for brief periods, were extremely air and moisture sensitive in solution and the solid state. As such, all manipulations were performed under an N 2 atmosphere either in a glovebox or using standard Schlenk techniques. [MoCl 3 (THF) 3 ] [48], (PDI)MoCl 3 [39], PDI [49], K(C 10 H 8 )(THF) 0.5 [50], Cp 2 Mo(SH) 2 [40] 4 ] to a solution of Ph 2 NH in hexane and collecting the precipitate by filtration. NMR spectra were recorded at 298 K using Varian 300 MHz, 400 or 500 MHz instruments. Chemical shifts were reported in ppm relative to tetramethylsilane using residual solvent as an internal standard. IR spectra were recorded using a PerkinElmer Spectrum One FT-IR spectrometer (Waltham, MA, USA) at 4 cm −1 resolution. EPR X-band spectra were recorded using a Bruker EMX spectrometer (Billerica, MA, USA) and analysed using Bruker Win-EPR software. Mass spectra were recorded using either an Agilent LCTOF mass spectrometer or a Waters GCT high resolution mass spectrometer operating in LIFDI mode. CV experiments were performed using a Pine AFP1 potentiostat. The cell consisted of a glassy carbon working electrode, a Pt wire auxiliary electrode and a Pt wire pseudo-reference electrode. All potentials were referenced vs. the Fc 0/+ couple measured as an internal standard. Elemental analyses were performed by Midwest Microlab. Solution-phase effective magnetic moments were determined by the method described by Evans [54] and corrected for diamagnetic contributions [55].

Crystallography
X-ray intensity data were collected on a Bruker APEX2 CCD detector (Billerica, MA, USA) employing graphite-monochromated Mo-Kα radiation (λ = 0.71073 Å) at 100(1) K. Absorption and other corrections were applied using SADABS [56]. The structures were solved by direct methods using SHELXT [57] and refined against F 2 on all data by full-matrix least squares with SHELXL-2015 [58]. Nonhydrogen atoms were refined anisotropically. Hydrogen atoms bound to Mo were permitted to refine freely; all others were refined using a riding model. CCDC 2006498-2006504 contains the supplementary crystallographic data for this paper. These data can be obtained free of charge via http://www.ccdc.cam.ac.uk/conts/retrieving.html (or from the CCDC, 12 Union Road, Cambridge CB2 1EZ, UK; Fax: +44-1223-336033; E-mail: deposit@ccdc.cam.ac.uk).

Experimental
trans-(PDI)MoCl 2 (THF) (1) (PDI)MoCl 3 (2.04 g, 2.98 mmol) and K(C 10 H 8 )(THF) 0.5 (0.667 g, 3.28 mmol) were suspended in toluene (20 mL). After stirring for 3 h at RT the emerald green mixture was filtered through a pad of Celite. Dilution with 3 volumes of hexane and cooling at −35 • C overnight produced pale green crystals which were collected by filtration. Recrystallisation by dissolving the complex in a minimum amount of THF and diluting with 3 volumes of hexane gave the product as glistening black-green crystals. Yield: 1.21 g, 56%. Layering a saturated THF solution of the complex with hexane gave crystals suitable for XRD studies. 1  trans-(PDI)MoCl 2 (µ-SH)Mo(SH)Cp 2 (2) 1 (1.504 g, 2.09 mmol) and Cp 2 Mo(SH) 2 (579 mg, 1.98 mmol) were suspended in THF (10 mL). After stirring at RT for 30 min a purple microcrystalline solid separated from the mixture. Toluene (20 mL) was added to the suspension and the mixture cooled to −35 • C overnight to complete precipitation. The red-purple microcrystalline solid was collected by filtration and washed with toluene (2 × 5 mL), twice with Et 2 O (2 × 5 mL) and dried under reduced pressure. Although not analytically pure, this material could be used in the next step without further purification. Yield: 1.640 g,~88%.
2 was soluble only in chlorinated solvents in which it decomposed over several hours. Nevertheless, layering a CH 2 Cl 2 solution of the complex with hexane at −35 • C gave crystals of sufficient quality to assign bond connectivity by XRD studies. FT-IR: cm −1 1875 (SH). 2 (950 mg, 1.01 mmol) was suspended in toluene (10 mL). A solution of DBU (314 mg, 2.07 mmol) in toluene (5 mL) was added dropwise with rapid stirring. After stirring at RT for 30 min the dark-red mixture was diluted with an equal volume of hexane and filtered through a pad of celite. The solvent was removed under reduced pressure and the residue recrystallized by dissolving the solid in a minimum of THF, diluting with several volumes of MeCN and concentrating to~50% of the original volume. The resulting black-red microcrystals were collected, washed with MeCN (3 × 5 mL) and dried under reduced pressure. Peaks for impurities were observed by 1 H NMR spectroscopy; however, this complex was sufficiently pure for the next step. Yield: 701 mg,~80%. Layering a saturated THF solution of the complex with hexane gave crystals suitable for XRD studies. 1   To a mixture of 3 (204 mg, 0.235 mmol) and Na[BH 4 ] (100 mg, 2.65 mmol) was added THF (5 mL) and the resulting mixture stirred at RT for 15 min to give a yellow-brown suspension. Prolonged reaction time gave decreased yields. The mixture was diluted with 2 volumes of hexane and filtered through a pad of Celite to remove excess Na [BH 4 ]. The solvent was removed under reduced pressure and the crude solid suspended in Et 2 O (5 mL). A stir bar was added and the mixture stirred for 15 min before the black microcrystalline solid was collected by filtration and washed with Et 2 O (3 × 5 mL).