Heterolytic H₂ Cleavage and Catalytic Hydrogenation by an Iron Metallaboratrane

Abstract

Reversible, heterolytic addition of H₂ across an iron-boron bond in a ferraboratrane with formal hydride transfer to the boron gives iron-borohydrido-hydride complexes. These compounds catalyze the hydrogenation of alkenes and alkynes to the respective alkanes. Notably, the boron is capable of acting as a shuttle for hydride transfer to substrates. The results are interesting in the context of heterolytic substrate addition across metal-boron bonds in metallaboratranes and related systems, as well as metal-ligand bifunctional catalysis.

INTRODUCTION

Transition metal-catalyzed bond forming reactions often involve formal two-electron redox steps (e.g., oxidative addition and reductive elimination). Noble metal catalysts are commonly used in these reactions due, in part, to their propensity to facilitate multi-electron processes.¹ There is growing interest in developing catalysts for bond forming/cleavage reactions based on earth abundant mid-to-late first-row transition metals, a goal that presents a unique set of challenges.² First-row transition metal catalysts that can circumvent undesirable one electron processes in favor of concerted two electron reaction steps present one plausible design criterion.³

Cooperative catalysis strategies that utilize ligands that operate in tandem with a coordinated metal center to activate substrates have shown promise in addressing this issue.^4,5 First-row metallaboratranes and related compounds that contain a retrodative M→B σ-interaction⁶ are appealing as catalysts⁷ because of the boron center’s ability to stabilize low valent metals.⁸ Akin to frustrated Lewis pairs,⁹ it has also been recently demonstrated that the metal-boron interaction can cooperatively facilitate the activation of H₂.^5,7 For instance, our lab recently reported that the diphosphine-borane nickel (^MesDPB^Ph)Ni complex undergoes reversible oxidative addition of H₂ to afford a nickel-borohydrido-hydride complex (Scheme 1).¹⁰ This nickel system is an efficient catalyst for olefin hydrogenation. Kameo and Nakazawa have also reported on the transfer hydrogenation of ketones catalyzed by a rhodium diphosphine-borane complex.¹¹

Scheme 1.

Related H₂ Activation Across a Ni-B Bond

Herein we report on studies of heterolytic H-H bond cleavage at a ferraboratrane complex¹² of a triphosphine-borane (TPB) ligand.¹³ Dihydrogen is shown to add reversibly across the Fe-B bond of (TPB)Fe(L) complexes to form corresponding iron-borohydrido-hydride complexes of the type (TPB)(μ-H)Fe(L)(H). Like the nickel system reported previously, olefin hydrogenation catalysis is accessible, albeit much slower, thereby facilitating detailed studies. As discussed below, other E-H bonds are also activated by the ferraboratanes described, including the terminal C-H bonds of arylacetylenes and the C-H bonds of formaldehyde.

RESULTS AND DISCUSSION

Reversible H₂ Addition

Exposing the previously reported (TPB)Fe(N₂) complex^12a (1) in d₆-benzene to H₂ (1.2 equiv) at room temperature results in H₂ addition across the Fe-B bond to give a yellow solution of the six coordinate borohydride-hydride-N₂ complex (TPB)(μ-H)Fe(N₂)(H) (2) (Scheme 2, Figure 1). The XRD structure of 2 shows that the Fe-B distance is significantly elongated relative to 1 (2.604(3) Å in 2 versus the calculated distance of 2.2 Å in 1^12b) (Figure 1). A terminal hydride ligand and a bridging hydride ligand located between the B and Fe atoms can be assigned from the electron density difference map. This structure also prevails in solution. In the ¹H NMR spectrum (d₆-benzene), the terminal hydride ligand on iron is observed as a triplet-of-doublets at −9.6 ppm, and the bridging hydride is observed as a broad singlet at −30.4 ppm. Replacing H₂ with D₂ in the reaction gives the corresponding isotopologue (TPB)(μ-D)Fe(N₂)(D). Along with the expected deuteride signals in the ²H NMR spectrum, deuterium signals from the methine and terminal methyl positions of the isopropyl groups of the TPB ligand are observed. This observation establishes that facile scrambling of the hydridic ligands into the TPB isopropyl groups occurs, presumably via a reversible C-H metalation process.

Scheme 2.

H-H and E-H Bond Activations Across Fe-B Bonds

Figure 1.

XRD structures of 2 (left) and 5 (right). Ellipsoids shown at 50% probability. Selected bond distances (Å): 2, Fe1-H1 = 1.42(2), Fe1-H2 = 1.49(2), B1-H2 = 1.17(2), Fe1-B1 = 2.604(3); 5, Fe1-H42 = 1.35(2), Fe1-H43 = 1.52(2), B1-H43 = 1.20(2), Fe1-B1 = 2.673(2).

The dinitrogen ligand (ν_NN = 2070 cm⁻¹) in 2 is labile, and it can be substituted under excess H₂ to give the dihydrogen analogue (TPB)(μ-H)Fe(H₂)(H) (3) (Scheme 2). Exposing 1 to excess H₂ (1 atm) also generates 3. Its XRD structure again confirms the presence of a bridging hydride (Figure 2). Although electron density can be located in the difference map between a widened P-Fe-P angle (136.54(2)°) and in the apical position trans to the borohydride unit, the data do not allow us to reliably distinguish between classical and non-classical hydrides. We therefore turned to NMR spectroscopy to aid in the formulation of 3.

Figure 2.

XRD (left) and DFT optimized (right) structures of 3. Ellipsoids shown at 50% probability. Selected bond distances (Å) and angles (°): XRD, Fe1-H42 = 1.56(2), B1-H42 = 1.21(2), Fe1-B1, 2.63(2), P2-Fe-P3 = 136.54(2); DFT predicted, Fe1-H42 = 1.51, Fe1-H43 = 1.50, Fe1-H44 = 1.62, Fe1-H45 = 1.59, B1-H42 = 1.24, Fe1-B1 = 2.63, H44-H45 = 0.84, P2-Fe1-P3 = 140.63.

The 20 °C ¹H NMR spectrum (d₈-toluene) of 3 (Figure 3) shows a broad singlet resonance at −15.1 ppm, indicative of hydridic protons. A broad deuteride signal is observed in the ²H NMR spectrum when D₂ is used in place of H₂, and, like in 2, deuterium signals are also observed in the methyl and methine positions of the isopropyl groups of the TPB ligand. Cooling a d₈-toluene solution of 3 under an H₂ atmosphere to −20 °C leads to sharpening of the resonance at −15.1 ppm, which integrates to three protons (3H). A second, broad hydridic resonance integrating to one proton is also observed at −24.9 ppm, and in analogy to 2, this resonance is assigned to the bridging-borohydride (μ-H). Cooling down to −90 °C leads to broadening of the 3H resonance without reaching decoalescence, suggesting that exchange of three hydrogenic ligands is fast on the NMR timescale. Compound 3 can be heated to 50 °C before significant sample decomposition is observed (vide infra). At 50 °C, the 3H and μ-H signals both broaden into the baseline, suggesting that exchange of all four hydrogens is facile at this temperature.

Figure 3.

¹H VT-NMR spectra of 3 in d₈-toluene under 1 atm of H₂.

To further assign the 3H unit in 3 (i.e. dihydrogen-hydride versus trihydride), we turned to minimal longitudinal relaxation (T_1min) measurements.^14,15 The T_1min is 35 ms at −32 °C for the 3H resonance, which suggests that the 3H unit is best described as dihydrogen-hydride and implying the following assignment for 3: (TPB)(μ-H)Fe(H₂)(H). This interpretation is additionally supported by DFT calculations (RB3LYP/6-31G(d)) that identify a dihydrogen-hydride structure (Figure 2 and Supporting Information) as the lowest energy isomer of 3. A stereoisomer in which the H₂ ligand occupies the equatorial position and the hydride ligand is in the axial position trans to boron is calculated to be 6.1 kcal/mol higher in energy. This geometric preference parallels that of the well-characterized iron-dihydrogen-dihydride complex mer-Fe(H₂)(H)₂(PEt₂Ph)₃.¹⁶ The transition state for conversion of 3 into this higher energy stereoisomer involving H-H scission is calculated to be 6.7 kcal/mol above the most stable isomer and is in line with the observed exchange behavior on the NMR timescale.

Analogous (TPB)(μ-H)Fe(L)(H) complexes (L = CN^tBu (5), CO (7)) can be synthesized (Scheme 2, Figure 1). To explore the effect of the apical ligand L on the H-H bond activation process, (TPB)Fe(L) complexes (L = CN^tBu, 4; L = CO, 6) were prepared. The previously reported carbonyl complex 6 displays an η³-interaction with a P-C_Ar-C_Ar unit of the TPB ligand^12a whereas isocyanide adduct 4, whose structure has been determined, does not: its Fe center is rigorously 5-coordinate. Complex 6 is diamagnetic whereas 4 gives rise to a solution magnetic susceptibility (μ_eff = 1.7 μ_B) at room temperature in C₆D₆. The temperature dependence of the solution susceptibility suggests an S = 0 ground state with a thermally accessible S = 1 state. No reaction occurs between (TPB)Fe(CN^tBu) (4, μ_eff = 1.7 μ_B) or the previously reported (TPB)Fe(CO)^12a (6) with H₂ (1 atm) at room temperature over a period hours. Compound 4 is fully consumed by H₂ (1 atm) over the course of 3 days at 40 °C to generate 5, while compound 6 is fully consumed by H₂ (1 atm) over the course of 5 days at 80 °C to give 7. The increase in temperature and reaction time compared to the facile room temperature reaction between 1 and H₂ is consistent with a scenario in which H₂ substitution for L occurs prior to H₂ addition across the Fe-B bond. Complex 7 can be alternatively synthesized from 1 or 2 and formaldehyde (Scheme 2).

Dihydrogen addition across the Fe-B bond of 1 is reversible. Conversion of 3 to 2 and subsequently back to 1 can be effected by exposing 3 to dynamic vacuum and then N₂, or by repeated freeze-pump-thaw-N₂ cycles. Reformation of the Fe-B bond can also occur through hydride transfer to unsaturated substrates (vide infra). Dihydrogen elimination from 5 and 7 does not occur when treated similarly.

Worth underscoring is that cleavage of the Fe-B bond in the present ferraboratrane system is distinct from the H₂ chemistry observed for a structurally related (SiP^iPr₃)Fe silatrane system (SiP^iPr₃= [Si(o-C₆H₄P^iPr₂)₃]⁻,) we have introduced elsewhere.^17,18,19 For instance, the reaction between (SiP^iPr₃)Fe(N₂) and H₂ affords (SiP^iPr₃)Fe(H₂), and that between [(SiP^iPr₃)Fe(N₂)]⁺ and H₂ affords [(SiP^iPr₃)Fe(H₂)]⁺. No disruption of the Fe-Si bond is observed in either case, even if for instance isolated [(SiP^iPr₃)Fe(N₂)]⁺ or [(SiP^iPr₃)Fe(H₂)]⁺ is exposed to excess H₂. Ligand substitution instead occurs. This sharply contrasts isoelectronic (TPB)Fe(N₂) 1, where H₂ addition readily affords the cleavage product 2 or 3. We also find (this report, Scheme 3) that hydrogenolysis of the iron(II) methyl complex (SiP^iPr₃)Fe(Me) occurs slowly at 60 °C to give an H₂/H product, but once again without disruption of the Fe-Si bond. Addition of exogenous donor ligands such as H₂, N₂ and CO effects substitution of the coordinated H₂ ligand, but the Fe-Si bond is maintained. One factor contributing to the difference between the two systems is likely the more flexible Fe-B bond in the (TPB)Fe system, as reflected in the variable Fe-B bond distances (varying by ca. 0.5 Å)^12b versus more rigid Fe-Si bond distances (varying by ca. 0.2 Å)^17,18,19 in (SiP^iPr₃)Fe that have been observed over several formal iron oxidation states. The rigidity of the Fe-Si interaction presumably reflects an appreciably stronger Fe-Si bond relative to Fe-B. One can additionally consider the relative Lewis acidity of the Ar₃B versus the Ar₃Si⁺ subunit¹⁸ in these respective systems, and their propensity to serve as H⁻ acceptors, but one might then predict the Ar₃Si⁺ to be the better acceptor, in contrast to the experimental observations. Indeed, this latter point may be the reason the Fe-Si interaction is stronger than the Fe-B interaction.

Scheme 3.

Chemistry of the Related Silatrane System

Reaction with Unsaturated Substrates

The ability of the (TPB)Fe scaffold to reversibly cleave H₂ prompted us to study if the transfer of hydrogen from (TPB)(μ-H)Fe(L)(H) to substrates is possible. The reaction of 1 with ethylene, styrene and arylacetylenes was probed. A degassed d₆-benzene solution of 1 reacts with ethylene to give a light brown solution of the paramagnetic iron-ethylene adduct (TPB)Fe(C₂H₄) (8; μ_eff = 3.2 μ_B, S = 1) (Scheme 4). Brown XRD quality crystals of 8 can be grown under an atmosphere of ethylene at 0 °C. Two molecules of 8 are found in the asymmetric unit cell. The iron center is bound η² to ethylene ((Fe-C)_avg = 2.108(1) Å) and lies above the plane defined by the phosphine donors by an average distance of 0.641 Å, with a corresponding elongation of the average Fe-B distance to 2.491(2) Å (compared to 2.2 Å in 1). The average C-C bond distance ((C-C)_avg = 1.397(2) Å) of the η²-coordinated ethylene molecule is significantly elongated from that in free ethylene (1.337 Å). ²⁰ The data are consistent with π-backbonding from iron to ethylene, which confers significant ferracyclopropane character to 8.²¹ Storing 8 for two days under an atmosphere of N₂ fully regenerates 1.

Scheme 4.

Ethylene Coordination and Arylacetylene C-H Bond Activation by 1

Compound 1 does not afford a detectable styrene adduct but reacts with both phenyl- and tolylacetylene with formal hydride transfer from the terminal C(sp)-H of the arylacetylene to the boron, forming S = 2 iron-borohydrido-arylacetylide complexes (TPBH)Fe(C₂Ar) (Ar = Ph, 9, μ_eff = 5.1 μ_B; Ar = Tol, 10, μ_eff = 5.2 μ_B) (Scheme 4). Two molecules of 10 are found in the asymmetric unit, and the XRD structure shows the presence of a tolylacetylide ligand coordinated to a pyramidalized iron center (Figure 4). While the hydride on the boron cannot be reliably located by XRD, the IR spectra for both 9 and 10 show B-H stretches at 2490 cm⁻¹ and 2500 cm⁻¹, respectively, most consistent with a non-bridging B-H unit. The vibrational bands shift to 1826 cm⁻¹ (predicted 1834 cm⁻¹) for 9 and 1824 cm⁻¹ (predicted 1841 cm⁻¹) for 10 upon labeling with the monodeuterated arylacetylene (ArC≡CD).

Figure 4.

XRD structures of 8 (left) and 10 (right). Ellipsoids shown at 50% probability. Selected bond distances (Å) and angles (°): 8 (average for two molecules in the asymmetric unit cell), Fe1-B1 = 2.491(1), Fe1-C37 = 2.103(1), Fe1-C38 = 2.113(1), C37-C38 = 1.397(2), Σ(P-Fe-P) = 338.76(3); 10 (average for two molecules in the asymmetric unit cell), Fe1-B1 = 2.761(2), Fe1-C37 = 1.918(2) C37-C38 = 1.169(3), Σ(P-Fe-P) = 345.07(2).

The activation of the arylacetylene C(sp)-H bond by 1 is reversible. Mixing a d₆-benzene solution of 9 with tolylacetylene (4 equiv) and, conversely, mixing a d₆-benzene solution of 10 with phenylacetylene (4 equiv) both result in a mixture of 9 and 10 (Scheme 5A). The corresponding exchange reactions with B-D labelled isotopologues of 9 or 10 ((TPBD)Fe(C₂Ar)) and a different all-protio arylacetylene (Ar’C≡CH) result in the exclusive formation of free ArC≡CD, indicating that the arylacetylene unit is reductively eliminated from the iron-borohydrido-arylacetylide complexes prior to activation of an incoming acetylene substrate, presumably by reversible hydride transfer from the boron to the arylacetylide to form intermediate π-adducts akin to 8 (Scheme 5B).

Scheme 5.

Reversible Arylacetylene C-H Bond Activation

Stoichiometric Hydrogenations

We also explored whether the transfer of hydrogen from (TPB)(μ-H)Fe(L)(H) species to unsaturated substrates might be possible. Exposing 8 to excess H₂ (1 atm) results in complete conversion of ethylene to ethane and 3 as the iron-containing product (Scheme 6-i) in less than 12 h. A paramagnetic intermediate (A) and 8 can be observed by in situ ¹H NMR spectroscopy. The same paramagnetic intermediate A and 8 can also be observed by in situ ¹H NMR spectroscopy if the same reaction is run under ethylene (1 atm) and H₂ (1 atm). The IR spectrum of the reaction mixture shows a diagnostic terminal B-H vibration at 2470 cm⁻¹ that is attributed to A. Akin to the iron-borohydrido-alkynyl complexes 9 and 10, A is assigned to the iron-borohydrido-ethyl complex (Scheme 6-i).²² Using D₂ in place of H₂ in the ethylene hydrogenation reaction and monitoring leads to the observation of both B-D and B-H stretches in the IR spectrum, which is consistent with facile insertion/β-hydride elimination processes prior to ethane elimination from A. We note the similarity of A (and 9 and 10) to the well-characterized zwitterionic, tris(phosphino)borate-iron-ethyl complex (PhBP₃)Fe(Et) that was observed as an intermediate in ethylene hydrogenation with the previously reported iron (PhBP₃)Fe system.^3e

Scheme 6.

Stoichiometric Hydrogenation Reactions

Complexes 1, 2, and 3 hydrogenate both phenylacetylene (1 equiv) and styrene (1 equiv) to ethylbenzene (1 atm H₂) with 3 as the observable iron-containing product. The in situ ¹H NMR spectrum of styrene hydrogenation reactions using 1, 2, or 3 show styrene, ethylbenzene and 3 in solution during the reaction course. Furthermore, for styrene hydrogenation with D₂ the ²H NMR spectrum and GC-MS data of the reaction mixture show incorporation of deuterium onto both olefinic carbon atoms of free styrene, indicating that styrene coordination to the iron-center and insertion/β-hydride elimination processes are reversible.

Under stoichiometric conditions, the addition of 1 equiv of styrene to a solution of 2 in d₆-benzene and under N₂ (1 atm) cleanly generates 1 equiv of ethylbenzene and 1 (Scheme 6-ii). In contrast, running the same reaction under a static vacuum yields a mixture of styrene, ethylbenzene, 1, and 2 (Scheme 6-iii). These observations suggest that excess H₂ or N₂ is required for the hydrogenations to proceed to completion. Moreover, the bridging monohydride appears competent for transfer to a substrate.

Substrates including trans-stilbene, N-benzylideneaniline, acetone, and acetophenone were not hydrogenated under similar conditions. Compounds 5 and 7 also do not hydrogenate ethylene, styrene, or phenylacetylene under the same conditions.

Catalytic Hydrogenations

Under the catalytic conditions of 0.01 M 1, 1 atm of H₂, and 30 equiv of the substrate in d₆-benzene at room temperature, ethylene, styrene, and phenylacetylene are hydrogenated to ethane and ethylbenzene, respectively (Table 1). Compounds 2 and 3 can also be used as precatalysts. Ambient laboratory light does not affect the reaction and the catalysis is not inhibited by elemental mercury; it thus appears to be a homogeneous process. Norbornene is hydrogenated to norbornane, and with an atmosphere of D₂ in place of H₂ the cis-addition product exo,exo-2,3-d₂-norbornane²³ is exclusively observed, indicating the syn-addition of hydrogen and arguing against radical processes.

Table 1.

Catalytic Hydrogenations by 1 with H₂

Precatalyst	Substrate	Product	TOF (h−1)
1	Ethylenea	Ethane	15
1	Styrenea	Ethylbenzene	0.27
1	Phenylacetyleneb	Ethylbenzene	0.16

Conditions: Room temperature, 0.01 M 1, 1 atm H₂, and 0.01 M ferrocene as an internal integration standard in d₆-benzene.

^a0.3 M substrate,

^b0.29 M substrate.

^cAs determined by ¹H NMR spectroscopy at > 95% product.

The hydrogenation catalysis was monitored by ¹H NMR spectroscopy with ferrocene as an internal integration standard. As with stoichiometric ethylene hydrogenation, the in situ ¹H NMR spectra of the catalytic ethylene hydrogenation reaction indicate the presence of ethylene adduct 8 and the putative ethyl-borohydride intermediate A as the iron-containing species during the reaction course. Complex 3 is the iron-containing product at the completion of the reaction. For styrene hydrogenation, styrene, ethylbenzene, and 3 are observed during catalysis, and scrambling of deuterium into the vinylic positions of styrene is observed under D₂. In contrast, for phenylacetylene hydrogenation complex 9 is the only observed iron-species early in the reaction when the phenylacetylene concentration is high. As phenylacetylene is consumed, styrene and 3 form, and ethylbenzene begins to develop slowly thereafter.

Attempting to increase the rate of catalysis by elevating the reaction temperature results in catalyst decomposition. The decomposition product (11) can be synthesized independently in near quantitative yields by heating 3 (80 °C) under H₂ (1 atm) for 2 h. The XRD structure of 11 indicates that a B-C_Ar bond is cleaved from the TPB ligand fragment (Figure 5). This result offers the cautionary note that B-C_Ar bond cleavage to give metal-borohydride products is a viable catalyst decomposition pathway.

Figure 5.

Chemical line representation and XRD structure of 11. Ellipsoids shown at 50% probability. Selected bond distances (Å): Fe1-H1 = 1.59(1), Fe1-H2 = 1.63(1), Fe1-B1 = 2.0900(6), Fe1-P1 = 2.2481(2).

Based on the results from the stoichiometric and catalytic experiments, we propose a plausible mechanistic scenario to account for the observed catalytic styrene hydrogenation by 1 (Scheme 7A), and in doing so underscore interesting aspects of the mechanism that remain unanswered. Starting from precatalyst 1, addition of H₂ generates 3, a species that can be observed by ¹H NMR spectroscopy during catalytic runs (resting state). Subsequent substitution of the apical H₂ ligand for styrene forms the unobserved iron-styrene-hydride-borohydride species B, and insertion into the terminal hydride affords the iron-alkyl intermediate A′. Intermediate A′ is analogous to the ethyl species A (Scheme 6-i), and is also related to the structurally characterized acetylide-borohydride complex 10 (Scheme 4). Olefin coordination and insertion appears to be reversible, as labeling studies show deuterium is exchanged into the vinylic positions of free styrene under a D₂ atmosphere. Elimination of ethylbenzene in the presence of H₂ regenerates the catalyst resting state.

Scheme 7.

Mechanism and Alkane Elimination Pathways

The conversion of 1 to 3 likely proceeds via the H₂-adduct intermediate C depicted in Scheme 7B by H₂-for-N₂ ligand exchange. While we have not detected such a species in the present (TPB)Fe system, its cobalt analogue (TPB)Co(H₂) can be isolated and has been thoroughly characterized,²⁴ as has the isoelectronic iron complex [(SiP^iPr₃)Fe(H₂)]^+.17 N₂/H₂ exchange is facile in these well-defined Co and Fe systems, and by extension we infer it would also be facile for (TPB)Fe(N₂) 1 to afford (TPB)Fe(H₂) C before additional reactions ensue. Since H₂ reacts with 1, but not with 5 and 7 at room temperature, we think that facile H₂ substitution for N₂ most likely occurs prior to H-H bond cleavage. We appreciate that while this scenario is consistent with the data available, it is not demanded by the available data. For instance, it is alternatively possible that styrene substitution for the N₂ ligand in 1 precedes H₂ addition to form intermediate B. No direct evidence rules out this possibility. We prefer suggesting that the H₂/dihydride species 3 precedes styrene binding because of the observation that H₂ addition/activation by other (TPB)Fe(L) adducts, for example 4 and 6, is very slow, and also because N₂, and presumably therefore also H₂, displaces ethylene from 8 in solution (regenerating 1 or (TPB)Fe(H₂)).

The final ethylbenzene elimination step can be envisioned to occur through two plausible routes (Scheme 7B). One such pathway involves a reductive elimination step where hydride transfer directly from the borohydride subunit generates the alkane product to form 3, likely via the dihydrogen adduct intermediate C. The other pathway proceeds through alkane elimination by hydrogenolysis of the phenylethyl group without hydride transfer from the borohydride subunit.

While the available data do not firmly distinguish between the two product elimination pathways shown in Scheme 7B, the stoichiometric hydrogenation studies described above show that 1 equiv of styrene is completely hydrogenated to ethylbenzene by 2 under an N₂ atm (Scheme 6-ii). This observation implies that 2 can serve as the source of the two H-atom equivalents delivered to styrene. From complex 2, substitution of the apical N₂ ligand for styrene in 2 would generate the styrene adduct intermediate B. Subsequent insertion of the bound styrene into the cis Fe-H would afford intermediate A′, which in the absence of H₂ at least, eliminates ethylbenzene concomitant with N₂ binding to reform 1. The presence of exogenous N₂ (or alternatively H₂) facilitates the generation of ethylbenzene. As noted in Scheme 6-iii, the stoichiometric reaction under static vacuum between 2 and styrene is quite slow compared with the same reaction under N₂ (Scheme 6-ii). This observation can be explained by presuming the conversion of intermediate A′ to 1 requires N₂ association prior to elimination to generate ethylbenzene.

CONCLUSIONS

In summary, we have demonstrated that the Fe-B bond in ferraboratranes (TPB)Fe(L) (where L = N₂, 1; CN^tBu, 4; CO, 6) can facilitate heterolytic cleavage of H₂ and of the C(sp)-H and C(sp²)-H bonds of arylacetylenes and formaldehyde, respectively, resulting in Fe-B bond rupture and formal hydride transfer to the boron of the ligand scaffold. The formal hydride transfer from the C(sp)-H of arylacetylenes to give iron-acetylide complexes is distinct from traditional syntheses of metal-acetylide complexes in that a hydride equivalent is formally abstracted by the Lewis acidic borane unit in 1 from a C(sp)-H hydrogen.²⁵

Dihydrogen addition across the Fe-B bond is reversible, and the boron is also capable of shuttling the hydride equivalent derived from H₂ to unsaturated substrates under stoichiometric hydrogenation conditions. The hydrogen chemistry of this (TPB)Fe(L) system contrasts with the related nickel²⁶ and cobalt²⁴ complexes of TPB, where the metal-boron bond remains intact under an H₂ atmosphere.

An understanding of the factors that govern metal-boron bond cleavage will aid in the development of cooperative catalytic reactions in metallaboratranes. The direct role, if any, of the borane ligand in assisting the H₂ cleavage step is an interesting question in this context for the present iron and recently reported diphosphine-borane-iron systems,²⁷ and also conceptually related to the nickel system (Scheme 1).¹⁰ Determining whether there is a cooperative interaction between the coordinated H₂ ligand, the iron center, and the borane subunit en route to H-H cleavage (and its microscopic reverse), akin to H-H cleavage by frustrated Lewis pairs,⁹ calls for detailed theoretical studies that are the subject of ongoing research.

EXPERIMENTAL

General Considerations

All manipulations were carried out using standard glovebox or Schlenk techniques under an N₂ atmosphere. Unless otherwise noted, solvents were deoxygenated and dried by thoroughly sparging with N₂ gas followed by passage through an activated alumina column in the solvent purification system by SG Water, USA LLC. Deuterated solvents and D₂ gas were purchased from Cambridge Isotope Laboratories, INC. The deuterated solvents were degassed and dried over activated 3 Å sieves prior to use. Unless otherwise noted, all compounds were purchased commercially and used without further purification. TPB,¹³ (SiP^iPr₃)Fe(Me),¹⁹ (SiP^iPr₃)Fe(N₂)(H),¹⁷ and monodeuterated phenyl- and tolylacetylene (PhC₂D and TolC₂D)²⁸ were synthesized by literature procedures. Elemental analyses were performed by Midwest Microlab, LLC., Indianapolis, IN.

NMR spectra were recorded on Varian 300 MHz, 400 MHz, and 500MHz spectrometers. ¹H and ¹³C chemical shifts are reported in ppm relative to residual solvent as internal standards. ³¹P and ¹¹B chemical shifts are reported in ppm relative to 85% aqueous H₃PO₄ and BF₃·Et₂O, respectively. Multiplicities are indicated by br (broad), s (singlet), d (doublet), t (triplet), quart (quart), quin (quintet), multiplet (m), d-d (doublet-of-doublets), and t-d (triplet-of-doublets).

FT-IR measurements were obtained on samples prepared as KBr pellets or in solution using a Bio-Rad Excalibur FTS 300 spectrometer with Varian Resolutions Pro software at 4 cm⁻¹ resolution. The ATR-IR measurements were measured on a thin film of the complex obtained from evaporating a drop of the solution on the surface of a Bruker APLHA ATR-IR Spectrometer probe (Platinum Sampling Module, diamond, OPUS software package) at 2 cm⁻¹ resolution. IR intensities indicated by s (strong), m (medium), and w (weak).

X-ray Crystallography

X-ray diffraction was measured on the Bruker Kappa Apex II diffractometer with Mo Kα radiation. Structures were solved using the SHELXS software and refined against F² on all data sets by full matrix least squares with SHELXL. The crystals were mounted on a glass fiber with Paratone oil.

Computational Methods

Geometry optimizations were performed using the Gaussian03 package. B3LYP exchange-correlation functional was employed with a 6-31G(d) basis set. The GDIIS algorithm was used. A full frequency calculation was performed on each structure to ensure that they were the true minima. A single negative vibrational frequency was observed for the transition state between 3 and its equatorial-H₂ isomer, confirming that this structure was the transition state. See SI for full a description of the computational method.

HD Gas Generation

D₂O (1mL) was added to an evacuated, cooled sample (−78 °C) of solid lithium aluminum hydride (316 mg, 8.2 mmol) in a Schlenk flask. An evacuated Schlenk line was filled with the resulting HD gas (ca. 1 atm) as the Schlenk flask was warmed to room temperature. A J-Young NMR tube containing a freeze-pump-thawed solution was the respective complex was exposed to the HD gas.

Synthesis of (TPB)(μ-H)Fe(N₂)(H) (2)

A J-Young NMR tube containing a brown-red solution 1 (20.3 mg, 31.1 mmol) in C₆H₆ (0.8 mL) was freeze-pump-thawed (3x) and, with the J-Young tube frozen with liquid nitrogen, exposed to H₂ (1.2 equiv). The solution was thawed and mixed, giving a yellow solution. An atmosphere of N₂ was subsequently introduced and the reaction was mixed for 2 h to yield 2 (100% yield by ¹H NMR spectroscopy with a ferrocene integration standard). Alternatively, 2 could be synthesized from 3 by degassing a solution of 3 of free H₂ by freeze-pump-thaw (3x), exposing it to an N₂ atmosphere, and mixing the solution overnight (100% yield by ¹H NMR spectroscopy with a ferrocene integration standard). Yields could not be determined by mass because 2 was unstable to prolonged exposure to dynamic vacuum. Yellow-orange XRD quality crystals were grown in a concentrated solution of pentane:THF (10:1) at −30 °C. ¹H NMR (C₆D₆, 300 MHz): δ 7.9 (2H, d, ³J_H-H = 6 Hz, Ar-H), δ 7.7 (1H, br s, Ar-H), δ 7.3 (3H, d, ³J_H-H = 6 Hz, Ar-H), δ 7.0 (3H, t, ³J_H-H = 9 Hz, Ar-H), δ 2.7 (4H, d, ²J_P-H = 18 Hz, PCH), δ 2.4 (2H, br s, PCH), δ 1.4 (6H, d,³J_H-H = 6 Hz, CH₃), δ 1.3 (12H, br s, CH₃), δ 1.1 (6H, d-d, ³J_P-H = 15 Hz, ³J_H-H = 6 Hz, CH₃), δ 0.8 (6H, d, ³J_P-H = 9 Hz, CH₃), δ − 9.6 (1H, d-t, ²J_H-Pcis = 81 Hz, ²J_H-Ptrans = 36 Hz, Fe-H), δ −30.4 (1H, s, Fe-(μ-H)-B). ²H NMR (C₆H₆/C₆D₆, 76 MHz): δ −9.5 (1D, br s), δ −30.3 (1D, br s). ³¹P NMR (C₆D₆, 121 MHz): δ 73.6 (2P, s), δ 64.2 (1P, s). ¹³C NMR (C₆D₆, 125 MHz): δ 161.9 (s, C^Ar), δ 143.5 (s, C^Ar), δ 141.0 (s, C^Ar), δ 132.3 (s, C^Ar), δ 131.8 (s, C^Ar), δ 131.3 (s, C^Ar), δ 130.3 (s, C^Ar), δ 124.6 (s, C^Ar), δ 124.0 (s, C^Ar), δ 32.0 (s, PCH), δ 29.4 (s, PCH), δ 28.5 (s, PCH), δ 22.8 (s, CH₃), δ 20.1 (s, CH₃), δ 19.7 (s, CH₃), δ 18.9 (s, CH₃). ¹¹B NMR (C₆D₆, 128 MHz): δ 8.2 (br). IR (KBr, cm⁻¹): 2071 (s, N≡N), 1960 (w) 1934 (w). UV-vis (THF, nm {M⁻¹ cm⁻¹}): 328 {shoulder, 500}, 280 {shoulder, 11250}. Anal.: Elemental analysis could not be obtained because of the instability of the compound under dynamic vacuum.

Synthesis of (TPB)(μ-H)Fe(H₂)(H) (3)

A J-Young NMR tube containing a brown-red solution 1 (21 mg, 31.1 mmol) in C₆H₆ (0.8 mL) was freeze-pump-thawed (3x). Upon warming to room temperature, the sample was exposed to H₂ (1 atm), resulting in a clear yellow solution. The reaction was mixed for 24 h to give 3 (100% by ¹H NMR spectroscopy with a ferrocene integration standard). The yield could not be determined by weight because 3 was unstable to prolonged exposure to dynamic vacuum. Yellow-orange XRD quality crystals were grown under 1 atm of H₂ in a concentrated solution of pentane:THF (10:1) at −78 °C. ¹H NMR (C₆D₆, 300 MHz): δ 8.0 (3H, d, ³J_H-H = 6 Hz, Ar-H), δ 7.3 (3H, t, ³J_H-H = 9 Hz, Ar-H), δ 7.2 (3H, t, ³J_H-H = 9 Hz, Ar-H), δ 7.0 (3H, d, ³J_H-H = 6 Hz, Ar-H), δ 4.47 (s, free H₂), δ 2.3 (6H, m, PCH), δ 1.0 (18H, d-d, ³J_H-P = 15 Hz, ³J_H-H = 6 Hz, CH₃), δ 0.8 (18H, d-d, ³J_H-P = 15 Hz, ³J_H-H = 6 Hz, CH₃), δ −15.1 (br s, 2H). T_1min (d₈-toluene): 35 ms (δ −15.1, −32 °C). ²H NMR (C₆H₆/C₆D₆, 76 MHz): δ −15.4 (1D, br s). ³¹P NMR (C₆D₆, 121 MHz): δ 90.0 (3P, s). ¹³C NMR (C₆D₆, 125 MHz): δ 163.5 (s, C^Ar), δ 144.6 (s, C^Ar), δ 144.1 (s, C^Ar), δ 130.9 (d, J_P-C = 23 Hz, C^Ar), δ 124.5 (s, C^Ar), δ 123.9 (s, C^Ar), δ 28.6 (s, PCH), δ 21.2 (s, CH₃), δ 19.9 (s, CH₃). ¹¹B NMR (C₆D₆, 128 MHz): δ 7.5 (br). IR (KBr, cm⁻¹): 2278 (w), 19618 (w), 1845 (w). UV-vis (THF, nm {M⁻¹ cm⁻¹}): 377 {shoulder, 1532}, 275 {14532}. Anal.: Elemental analysis could not be obtained because of the instability of the compound under dynamic vacuum.

Synthesis of (TPB)Fe(CN^tBu) (4)

Tert-butyl isocyanide (20 mg, 0.24 mmol) was added to a brown solution of (TPB)Fe(N₂) (40 mg, 59 μmol) in benzene (2 mL), causing an instantaneous darkening upon gentle shaking. The volatiles were removed by lyophilization and the residue was extracted with tetramethylsilane (2 mL). The resulting dark brown solution was slowly concentrated down to ca. 0.2 mL by vapor diffusion into hexamethyldisiloxane. Removal of the mother liquor by decantation, washing with cold tetramethylsilane (2 × 0.1 mL) and drying in vacuo afforded (TPB)Fe(CN^tBu) as brown crystals (33 mg, 77%). ¹H NMR (C₆D₆, 300 MHz): δ 11.2 (3H), δ 9.2 (3H), δ 8.6 (3H), δ 8.5 (9H), δ 6.4 (3H), δ 5.2 (9H), δ 3.7 (12H), δ 2.9 (9H), δ −1.5 (9H), δ −2.3 (3H). IR (KBr, cm⁻¹): 1972 (C≡N). UV-Vis (THF, nm {cm⁻¹M⁻¹}): 600 {shoulder, 428}, 910 {70}. μ_eff (C₆D₆, method of Evans, 20 °C): 1.7 μ_B. Anal.: Calc’d for C₄₁H₆₄BFeNP₃: C 67.50, H 8.70, N 1.92; found: C 67.20, H 8.54, N 1.72.

Synthesis of (TPB)(μ-H)Fe(CN^tBu)(H) (5)

A heavy-walled Schlenk tube containing a yellow-brown solution of 4 (16.4 mg, 22.4 mmol) in C₆H₆ (10 mL) was freeze-pump-thawed (3x). Upon warming to room temperature, the sample was exposed to H₂ (1 atm) for a few minutes. The Schlenk tube was sealed and heated under vigorous mixing at 40 °C for 85 h. Removal of the solvent in vacuo, extraction with C₆H₆, and lyophilization yielded a solid of 5 (17.3 mg, 98 %). Room temperature evaporation of a solution of 5 in a diethyl ether:pentane (2 mL to 1 mL) mixture yielded yellow crystals suitable for XRD analysis. ¹H NMR (C₆D₆, 300 MHz): δ 8.1 (1H, d, ³J_H-H = 9 Hz, Ar-H), δ 7.4 (3H, d, ³J_H-H = 9 Hz, Ar-H), δ 7.2 (3H, d, ³J_H-H = 6 Hz, Ar-H), δ 7.1 (3H, d, ³J_H-H = 9 Hz, Ar-H), δ 2.7 (2H, br s, PCH), δ 2.5 (2H, br s, PCH), δ 2.2 (2H, t, ²J_H-P = 9 Hz, PCH), δ 1.3 (16H, m, CH₃), δ 1.1 (9H, s, C(CH₃)₃), δ 1.0 (6H, d, ³J_H-H = 6 Hz, CH₃), δ 0.8 (6H, d-d, ³J_H-P = 15 Hz, ³J_H-H = 6 Hz, CH₃), δ 0.7 (6H, br s, CH₃), δ −11.7 (1H, t-d, ²J_H-Pcis = 87 Hz, ²J_H-Ptrans = 27 Hz, Fe-H), δ −23.9 (1H, br s, Fe-(μ-H)-B). ¹³C NMR (C₆D₆, 125 MHz): δ 176.9 (quart, ²J_C-P = 8 Hz, CN^tBu), δ 163.5 (br s, C^Ar), δ 163.0 (br s, C^Ar), δ 144.9 (m, C^Ar), δ 143.0 (d, J_C-P = 20 Hz, C^Ar), δ 131.9 (d, ²J_C-P = 7.5 Hz, C^Ar), δ 130.6 (d, J_C-P = 3.2 H, C^Ar), δ 130.5 (d, J_C-P = 2.5 H, C^Ar), δ 128.6 (s, C^Ar), δ 127.5 (s, C^Ar), δ 124.2 (s, C^Ar), δ 123.6 (s, C^Ar), δ 55.3 (s, C(CH₃)₃), δ 31.6 (s, PCH), δ 30.9 (s, PCH), δ 29.2 (m, PCH), δ 28.5 (d, ²J_C-P = 7.5 Hz, PCH), δ 23.8 (s, CH₃), δ 20.4 (s, CH₃), δ 20.3 (m, CH₃), δ 20.2 (s, CH₃), δ 19.9 (s, CH₃), δ 19.7 (s, CH₃). ³¹P NMR (C₆D₆, 121 MHz): δ 81.3 (2P, d, ²J_P-P = 63 Hz), δ 72.4 (1P, s). IR (KBr, cm⁻¹): 2027 (s, C≡N), 1942 (w, Fe-H). UV-Vis (THF, nm {cm⁻¹ M⁻¹}): 205 {5530}, 224 {15437}, 245 {17142}, 255 {16635}, 285 {4117}, 335 {shoulder, 2166}, 400 {1830}. Anal: Calc’d for C₄₁H₆₆BFeNP₃: C 67.32, H 8.96, N 1.91; found: C 66.59, H 8.61, N 1.30.

Synthesis of (TPB)(μ-H)Fe(CO)(H) (7) from 6 and H₂

In a J-Young NMR tube, 6 (6.0 mg, 8.9 μmol) was dissolved in C₆H₆ (0.7 mL) to give a brown-red solution. The solution was freeze-pump-thawed (3x) and subsequently exposed to H₂ (1 atm) for ca. 5 min. The reaction was then heated at 80 °C for 5 days, during which time a clear yellow solution developed. Removal of the solvent in vacuo, extraction with C₆H₆, and lyophilization yielded a yellow solid of 7 (5.9 mg, 98 %). Room temperature evaporation of a solution of 7 in a diethyl ether:pentane (1 mL to 0.5 mL) mixture yielded yellow analytically pure 7. ¹H NMR (C₆D₆, 300 MHz): δ 8.1 (2H, d, ³J_H-H = 6 Hz, Ar-H), δ 7.9 (1H, d, ²J_P-H = 9 Hz, Ar-H), δ 7.2 (4H, m, Ar-H), δ 7.0 (4H, m, Ar-H), δ 2.6 (2H, t, ²J_H-P = 3 Hz, PCH), δ 2.4 (2H, q, ²J_H-P = 6 Hz, PCH), δ 2.2 (2H, t, ²J_H-P = 6 Hz, PCH), δ 1.4 (6H, d, ³J_H-P = 6 Hz, CH₃), δ 1.2 (12H, m, CH₃), δ 0.9 (6H, d, ³J_H-P = 6 Hz, CH₃), δ 0.8 (6H, d-d, ³J_H-P = 8 Hz, ³J_H-H = 6 Hz, CH₃), δ 0.6 (6H, d, ³J_H-P = 6 Hz, CH₃). δ −11.6 (1H, t-d, ²J_H-Pcis = 81 Hz, ²J_H-Ptrans= 21 Hz, Fe-H), δ −20.0 (1H, br s, Fe-(μ-H)-B). ²H NMR (C₆H₆, 76 Hz): δ −12.3 (1D, t, ²J_P-D = 10 Hz), δ −20.8 (1D, br s). ¹³C NMR (THF with 1 drop of C₆D₆, 125 MHz): δ 222.7 (br s, CO), δ 161.8 (br s, C^Ar), δ 142.9 (br s, ¹J_C-P = 19 Hz, C^Ar), δ 140.8 (br s, ²J_C-P = 16 Hz, C^Ar), δ 131.0 (s, C^Ar), δ 129.8 (s, C^Ar), δ 128.5 (s, C^Ar), δ 128.0 (s, C^Ar), δ 124.0 (s, C^Ar), δ 123.4 (s, C^Ar), δ 30.8 (s, PCH), δ 28.3 (s, PCH), δ 27.7 (s, PCH), δ 22.3 (s, CH₃), δ 19.1 (s, CH₃), δ 18.8 (s, CH₃), δ 18.1 (s, CH₃). ³¹P NMR (C₆D₆, 121 MHz): δ 83.4 (2P, d, ²J_P-H = 21 Hz), δ 72.8 (1P, s). IR (KBr, cm⁻¹): 1898 (s, C≡O), 1967 (w, Fe-H). UV-Vis (THF, nm {cm⁻¹M⁻¹}): 270 {4333}, 280 {4111}, 390 {1400}. Anal.: Calc’d for C₃₇H₅₆BFeOP₃: C 65.70, H 8.34; found: C 65.64, H 8.08.

Synthesis of (TPB)(μ-H)Fe(CO)(H) (7) from Formaldehyde

Compound 1 (8 mg, 11.8 μmol) or 2 (6 mg, 8.9 μmol) was mixed with excess paraformaldehyde in C₆H₆ for 3 h to give a turbid, light yellow solution. The excess paraformaldehyde was filtered away and the solution was pumped down to give 7 as a yellow solid (from 1, 8 mg, 100%; from 2, 7 mg, 100%). Spectroscopic data is identical to those listed above.

Synthesis of (TPB)Fe(C₂H₄) (8)

A J-Young NMR tube containing an orange solution of 3 (8.4 mg, 12.5 μmol) in C₆H₆ (0.8 mL) was freeze-pump-thawed (3x) and exposed to ethylene gas (1 atm) for ca. 1 minute. Mixing immediately gave a brown solution of 8. Removal of solvent in vacuo yielded a brown solid (8.3 mg, 99%) of 8. Dissolution of this solid under N₂ atmosphere gave mostly 8 and small amounts of 1. Over time, 8 in solution converted to 1. Crystals suitable for XRD were grown in a saturated, cold pentane:diethyl ether (2:1) solution under an ethylene atmosphere. ¹H NMR (C₆D₆, 300 MHz): δ 33.1 (1H), δ 28.8 (1H), δ 18.7 (4H), δ 13.3 (1H), δ 5.25 (s, free C₂H₄), δ 4.9 (3H), δ 4.1 (1H), δ 1.9 (2H), δ −3.0 (10H), δ −6.2 (11H), δ −9.2 (11H), δ −10.0 (3H). UV-Vis (THF, nm {cm⁻¹M⁻¹}): 309 {shoulder, 6032}, 553 {1804}, 938 {418}. μ_eff (C₆D₆, method of Evans, 20 °C): 3.2 μ_B (S = 1). Anal.: Elemental analysis could not be obtained because of the instability of the compound under dynamic vacuum.

Synthesis of (TPBH)Fe(C₂Ph) (9)

Phenylacetylene (40.8 mg, 400 μmol) was added to a C₆H₆ solution (5 mL) of 3 (9.0 mg, 13 μmol), immediately giving a gray solution. Removal of the solvent in vacuo yielded a black powder of 9 (10 mg, 100%). ¹H NMR (C₆D₆, 300 MHz): δ 28.8 (3H), δ 13.1 (4H), δ 4.7 (18H), δ 2.6 (2H), δ 2.3 (4H), δ 1.2 (2H), δ −29.3 (1H), δ −30.9 (5H). μ_eff (C₆D₆, method of Evans, 20 °C): 5.1 μ_B (S = 2). UV-Vis (THF, nm {cm⁻¹M⁻¹}): 325 {20670}, 439 {1051}, 479 {971}, 522 {955}, 601 (br abs extending from 400 to 700 nm, 930}, 883 {1466}. IR (KBr, cm⁻¹): 2040 (s, C≡C), 2490 (m, B-H). Anal.: Calc’d for C₄₄H₆₀FeP₃B: C, 70.60; H, 8.08. Found: C, 70.39; H, 7.89.

Synthesis of (TPBD)Fe(C₂Ph)

The B-D labeled complex of 9 was generated by the same method described for 9, except that PhC₂H was replaced with PhC₂D. The ¹H NMR spectrum was identical to 10. IR (thin film, cm⁻¹): 1826 (br m, B-D; predicted 1832).

Synthesis of (TPBH)Fe(C₂Tol) (10)

Tolylacetylene (16.1 mg, 138 μmol) was added to a C₆H₆ solution (5 mL) of 3 (22.9 mg, 33.0 μmol), immediately giving a gray solution. Removal of the solvent in vacuo yielded a black powder of 10 (24.0 mg 100%). Black XRD quality crystals of 10 were grown by diffusion of hexamethyldisiloxane into a concentrated THF solution of 10. ¹H NMR (C₆D₆, 300 MHz): δ 44.1 (1H), δ 29.4 (1H), δ 13.3 (1H), δ 4.6 (2H), δ 3.3 (1H), δ 2.7 (2H), δ 2.3 (4H), δ 1.8 (2H), δ −32.3 (1H). μ_eff (C₆D₆, method of Evans, 20 °C): 5.2 μ_B (S = 2). UV-Vis (THF, nm {cm⁻¹M⁻¹}): 326 {24476}, 444 {1243}, 486 {1138}, 527 {shoulder, 975}, 624 (915}, 887 {1575}. IR (KBr, cm⁻¹): 2039 (s, C≡C), 2500 (m, B-H). Anal.: Calc’d for C₄₅H₆₂FeP₃B: C, 70.88; H, 8.20. Found: C, 70.44; H, 7.80.

Synthesis of (TPBD)Fe(C₂Tol)

The B-D labeled complex of 10 was generated by the same method described for 10, except that TolC₂H was replaced with TolC₂D. The ¹H NMR spectrum was identical to 10. IR (thin film, cm⁻¹): 1824 (br m, B-D; predicted 1841).

Synthesis of 11

A yellow, C₆H₆ solution of 1 (18.2 mg, 27 μmol) was heated in a J-Young NMR tube under H₂ (1 atm) at 80 °C for 2 h, giving a turbid red-purple solution. The solvent was removed in vacuo, and the dark crude material was redissolved in hexamethyldisiloxane (3 mL) and filtered through a glass frit to remove a black solid (presumably iron metal). Removal of the solvent in vacuo gave a purple solid that is a mixture of 11 (1 equiv) and diisopropyl-phosphino-benzene (^iPr2PPh, 1 equiv). Orange XRD quality crystals of 11 can be grown from slow evaporation of a concentrated hexamethyldisiloxane solution of 11 at room temperature (5.1 mg, 18%). Dissolution of these crystals by heating in benzene or THF results in decomposition. Therefore, spectral data is reported on the mixture of 11 with ^iPr2PPh. Compound 11 appears to be fluxional at RT. ¹H NMR (C₆D₆, 300 MHz): δ 8.4 (4H, s, Ar-H), δ 7.6 (4H, s, Ar-H), δ 7.5 (3H, d, ²J_H-P = 6 Hz, Ar-H), δ 2.7 (1H, d, ²J_H-P = 6 Hz, PCH), δ 2.5 (1H, s, PCH), δ 1.9 (1H, quart, ³J_H-H = 6 Hz, PCH), δ 1.3 (6H, d-d, ³J_P-H = 4 Hz, ³J_H-H = 2 Hz, CH₃), δ 1.2 (6H, d, ³J_H-H = 3 Hz, CH₃), δ 1.1 (6H, quart, ³J_P-H = 4 Hz, ³J_H-H = 2 Hz, CH₃), 0.9 (12H, m, CH₃), −17.0 (0.25H, t, ²J_H-P = 36 Hz, B-H). ¹³C NMR (C₆D₆, 125 MHz): δ 157.4 (br s, C^Ar), δ 144.2 (br s, C^Ar), δ 134.9 (d, ²J_C-P = 19 Hz, C^Ar), δ 130.5 (s, C^Ar), δ 129.3 (s, C^Ar), δ 125.7 (d, ²J_C-P = 23 Hz, C^Ar), δ 25.5 (br s, PCH), δ 24.9 (br s, PCH), δ 24.4 (br s, PCH), δ 23.7 (br s, PCH), δ 23.1 (br s, PCH), δ 20.0 (m, CH₃), δ 19.4 (m, CH₃), δ 18.2 (m, CH₃). ³¹P NMR (C₆D₆, 121 MHz): δ 99.1 (4P, s), δ 9.5 (2P, s). ¹¹B NMR (C₆D₆, 128 MHz): δ 41.0 (br). UV-Vis (THF, nm {cm⁻¹M⁻¹}): 264 {shoulder, 31650}, 520 {826}. IR (KBr, cm⁻¹): 2088 (s, B-H), 1838 (m, B-H). Anal.: Calc’d for C₅₁H₈₆FeP₄B₂Si (i.e., 11 + Me₃SiH): C, 65.96; H, 9.33. Found: C, 65.70; H, 9.16.

Synthesis of (SiP^iPr₃)Fe(H₂)(H)

In a 100 mL Schlenk tube a red solution of (SiP^iPr₃)Fe(Me) (1.05 g, 1.547 mmol) in C₆H₆ (50 mL) was degassed by freeze-pump-thaw (3x). H₂ gas (1 atm) was charged into the reaction mixture. The reaction was heated at 60 °C for over a week. The reaction solution was then quickly filtered through Celite and volatiles were removed in vacuo to give a light yellow powder. The solid was collected on a glass-frit and washed with pentane (3 mL x 2). The resulting product (SiP^iPr₃)Fe(H₂)(H) (950 mg, 1.425 mmol, 92%) was obtained as a light yellow powder after drying under vacuum. ¹H NMR (C₆D₆, 300 MHz): δ 8.3 (3H, d, ³J_H-P = 6.8 Hz, Ar-H), δ 7.3 (3H, m, Ar-H), δ 7.2 (3H, t, ³J_H-H = 7.2 Hz, Ar-H), δ 7.1 (3H, t, ³J_H-H = 7 Hz, Ar-H), δ 4.5 (s, free H₂), δ 2.2 (6H, m, PCH), δ 1.0 (18H, m, CH₃), δ 0.8 (18H, br s, CH₃), δ 0.16 (s, free CH₄), δ −10.0 (3H, quin, ²J_P-H = 18.4 Hz, Fe-H). T_1min (d₈-toluene): 32 ms (δ −10.0, −30 °C). ³¹P NMR (C₆D₆, 121 MHz): δ 100 (br); (d₈-Tol, 300 MHz, −80 °C): δ 117.9 (d, ³J_P-P = 43.5 Hz), δ 94.7 (bs), δ 84.7 (d, ³J_P-P = 43.5 Hz). ¹³C NMR (THF with 1 drop of C₆D₆, 125 MHz): δ 157.3 (d, J_C-P = 22.5 Hz, C^Ar), δ 150.5 (d, J_C-P = 21.3 Hz, C^Ar), δ 130.1 (d, J_C-P = 9.3 Hz, C^Ar), δ 128.6 (s, C^Ar), δ 127.4 (s C^Ar), δ 126.0 (d, J_C-P = 2.5 Hz, C^Ar), δ 29.0 (br s, PCH), δ 20.4 (br s, CH₃), δ 19.2 (br s, CH₃), UV-Vis (THF, nm {cm⁻¹M⁻¹}): 353 {3040}. IR (KBr pellet; cm⁻¹): 1941 (w, Fe-H). Anal.: Elemental analysis could not be obtained because of the instability of the compound under prolonged exposure to N₂.

Isotopomers of (SiP^iPr₃)Fe(H₂)(H); (SiP^iPr₃)Fe(H₂)(H), (SiP^iPr₃)Fe(HD)(H), (SiP^iPr₃)Fe(H₂)(D), (SiP^iPr₃)Fe(D₂)(H), (SiP^iPr₃)Fe(HD)(D), and (SiP^iPr₃)Fe(D₂)(D)

HD gas was generated by the method described above and charged into a J-young tube containing a degassed solution of SiP^iPr₃)Fe(Me) in C₆D₆. The solution was heated at 60 °C for over a week. Monitoring the progress of the reaction by ¹H NMR revealed the gradual disappearance of 1 and formation of diamagnetic isotopomers (SiP^iPr₃)Fe(H₂)(H). Spectroscopic features in the ¹H NMR spectrum were identical to (SiP^iPr₃)Fe(H₂)(H) except for the hydridic proton resonances, where isotopologues were observed. ¹H{³¹P} NMR (C₆D₆, 300 MHz): δ −10.0 (3H, quart, ²J_P-H = 18.4 Hz, H₃), δ −10.0 (t, ¹J_H-D = 9.5 Hz, H₂D), δ −10.23 (quin, ¹J_H-D = 9.3 Hz, HD₂).

Synthesis of (SiP^iPr₃)Fe(CO)(H)

In a 50 mL Schlenk tube a yellow solution of (SiP^iPr₃)Fe(H₂)(H) (330 mg, 0.495 mmol) in C₆H₆ (20 mL) was freeze-pump-thawed (3x). The solution was charged with CO (1 atm). The reaction was mixed overnight at RT and then for 1 h at 60 °C, resulting in a light yellow solution. After completion, the solution was degassed by freeze-pump-thaw (3x). The solution was filtered through Celite, and the volatiles were removed in vacuo to give a light yellow powder. The solid was collected on a glass-frit and washed with pentane (5 mL x 3). Removal of the solvent in vacuo yields (SiP^iPr₃)Fe(CO)(H) (266 mg, 0.384 mmol, 78%) as a light yellow powder. Crystals suitable for X-ray diffraction were obtained by RT evaporation of a pentane solution of (SiP^iPr₃)Fe(CO)(H). ¹H NMR (d₈-toluene, 300 MHz): δ 8.2 (2H, d, ³J_H-H = 7.2 Hz), δ 8.1 (1H, d, ³J_H-H = 7.2 Hz), δ 7.3 (3H, m), δ 7. 2 (3H, m), δ 7.1 (3H, m), δ 2.7 (2H, m), δ 2.4 (2H, m), δ 2.2 (2H, m), δ 1.5 (6H, d-d, ³J_H-P = 15.2 Hz, ³J_H-P = 6.8 Hz), δ 1.3 (6H, m), δ 1.1 (6H, d-d, ³J_H-P = 12.6 Hz, ³J_H-H = 6.6 Hz), δ 0.8 (12H, m), δ 0.5 (6H, s), δ −14.9 (1H, t-d, ³J_H-Pcis = 81.0 Hz, ³J_H-Ptrans = 14.1 Hz). ³¹P NMR (C₆D₆, 121 MHz, RT): δ 88.0 (br), δ 90.0 (s); (d₈-toluene, 121 MHz, −80 °C): δ 109.9 (t, ³J_P-P = 65 Hz, ³J_P-H = 80 Hz), δ 90.0 (s), δ 77.9 (t, ³J_P-P = 65 Hz, ³J_P-H = 80 Hz). ¹³C NMR (THF with 1 drop of C₆D₆, 125 MHz): δ 223.0 (d, ²J_C-P = 6.3 Hz, CO), δ 157.2 (d, J_C-P = 8.8 Hz, C^Ar), δ 155.2 (d, J_C-P = 10.6 Hz, C^Ar), δ 150.5 (d, J_C-P = 11.3 Hz, C^Ar), δ 150.2 (d, J_C-P = 8.8 Hz, C^Ar), δ 148.6 (d, J_C-P = 18.1 Hz, C^Ar), δ 132.4 (d, J_C-P = 9.4 Hz, C^Ar), δ 131.9 (d, J_C-P = 8.8 Hz, C^Ar), δ 130.9 (s, C^Ar), δ 128.3 (s, C^Ar), δ 128.5 (s, C^Ar), δ 127.1 (s, C^Ar), δ 126.8 (s, C^Ar), δ 126.4 (s, C^Ar), δ 125.8 (s, C^Ar), δ 124.7 (s, C^Ar), δ 32.4 (s, PCH), δ 30.7 (s, PCH), δ 30.0 (s, PCH), δ 29.6 (s, PCH), δ 28.7 (s, PCH), δ 22.8 (s, CH₃), δ 22.1 (s, CH₃), δ 20.2 (s, CH₃), δ 19.5 (s, CH₃), δ 19.2 (s, CH₃), δ 19.1 (s, CH₃), δ 18.9 (s, CH₃), δ 18.0 (s, CH₃). UV-Vis (THF, nm {cm⁻¹M⁻¹}): 340 {2,050}, 400 {1,500}. IR (KBr pellet; cm⁻¹): 1882 (s, C≡O), 1944 (m, Fe-H). Anal.: Calc’d for C₃₇H₅₇FeOP₃Si: C, 64.16; H, 8.00. Found: C, 64.15; H, 8.13.

Synthesis of (SiP^iPr₃)Fe(¹³CO)(H)

In a J-Young NMR tube an orange C₆D₆ solution of (SiP^iPr₃)Fe(H₂)(H) was degassed by three freeze-pump-thaw (3x). Subsequently, ¹³CO (1 atm) was added and the reaction was allowed to mix overnight at RT. The ¹H NMR spectrum was identical to (SiP^iPr₃)Fe(CO)(H). IR (KBr; cm¹): 1836 (s, ¹³C≡O).

Generation of (TPBH)Fe(Et) (A)

Compound A was observed as an intermediate of catalytic ethylene hydrogenation under the reaction conditions described in section “Catalytic Hydrogenation Studies.” We were also able to generate A starting from complex 8. This procedure described below is more amendable to observing A spectroscopically. A dark yellow C₆D₆ solution (0.5 mL) of 8 (2.8 mg, 4.2 μmol) under ethylene (1 atm) in a J-Young tube was frozen (−196 °C) and H₂ was added (1 atm). The reaction was thawed and quickly mixed only immediately prior to measuring the ¹H NMR spectrum, revealing a mixture of 8 and A. Further mixing of the solution for ca. 45 min yielded a purple solution of A with a small residual amount of 8 by ¹H NMR spectroscopy. Free C₂H₄, C₂H₆, and H₂ were also observed in the ¹H NMR spectrum. Compound A could not be isolated as a solid due its instability. For example, an ATR IR spectrum of a thin film of the reaction mixture obtained by solvent evaporation under an N₂ atmosphere over a period less than 30 sec gave vibrational bands diagnostic of 1, 2, and A. ¹H NMR (C₆D₆, 300 MHz): δ 17.3 (1H), δ 6.6 (1H), δ 5.26 (s, free C₂H₄), δ 4.4 (1H), δ 4.47 (br s, free H₂), δ 3.3 (2H), 0.80 (s, free C₂H₆), δ −1.9 (8H), δ −5.5 (6H). IR (thin film; cm⁻¹): 2470 (br s, B-H of A), 2069 (m, N≡N of (TPB)(μ-H)Fe(N₂)(H) 2), 2009 (s, N≡N of (TPB)Fe(N₂) 1). UV-Vis, obtained 45 min after exposing A under 1 atm of ethylene to 1 atm of H₂ (THF, nm {cm⁻¹M⁻¹}): 325 {20670}, 439 {1051}, 479 {971}, 522 {955}, 601 (br abs extending from 400 to 600 nm, 930}, 883 {1466}. Magnetic data could not be obtained due to residual 10 in the reaction mixture. Anal.: Elemental analysis could not be obtained because of the instability of the compound under dynamic vacuum.

Catalytic Hydrogenation Studies

Compound 1 (0.045 g, 0.07 mmol) and ferrocene (0.012 g, 0.07 mmol) were dissolved in 1.5 mL of C₆D₆, giving a 0.045 M precatalyst stock solution. Ferrocene was used as an internal ¹H NMR integration standard and did not affect the rates of hydrogenation. For a catalytic run, 0.1 mL of the stock solution was taken and mixed with 0.35 mL of C₆D₆, and 30 equiv of substrate in a J-Young NMR tube (3.2 mL capacity). For styrene hydrogenation, this equates to 0.01 M 1 and 0.3 M styrene. For phenylacetylene hydrogenation, this equates to 0.01 M 1 and 0.29 M phenylacetylene. For ethylene hydrogenation, this equates to 0.01 M 1 and 0.30 M ethylene. The sample in the J-Young NMR tube was subsequently degassed by freeze-pump-thaw (3x) and backfilled with 1 atm H₂ (0.11 mmol). The J-Young NMR tube was continually inverted (12 min⁻¹) to ensure adequate mass transfer. The tube was periodically refilled with H₂ to maintain 1 atm of H₂. All reactions were monitored periodically by ¹H NMR spectroscopy until >95% completion. All reactions resulted in clean conversion of the substrate to the corresponding product. Catalytic runs in the presence of a drop of mercury or in the absence of ambient laboratory light had no affect on the reactions. Catalytic hydrogenations could also be cleanly effected by pre-generating 2 or 3 before the addition of the substrate.