Reversible cooperative dihydrogen binding and transfer with a bis-phosphenium complex of chromium†

The reversible reaction of H2 with a bis-phosphenium complex of chromium provides a rare example of 3d transition metal/phosphenium cooperativity. Photolysis induces the activation of H2 and yields a spectroscopically detectable phosphenium-stabilized (σ–H2)-complex, readily showing exchange with gaseous H2 and D2. Further reaction of this complex affords a phosphine-functionalized metal hydride, representing a unique example of reversible H2 cleavage across a 3d M 
<svg xmlns="http://www.w3.org/2000/svg" version="1.0" width="13.200000pt" height="16.000000pt" viewBox="0 0 13.200000 16.000000" preserveAspectRatio="xMidYMid meet"><metadata>
Created by potrace 1.16, written by Peter Selinger 2001-2019
</metadata><g transform="translate(1.000000,15.000000) scale(0.017500,-0.017500)" fill="currentColor" stroke="none"><path d="M0 440 l0 -40 320 0 320 0 0 40 0 40 -320 0 -320 0 0 -40z M0 280 l0 -40 320 0 320 0 0 40 0 40 -320 0 -320 0 0 -40z"/></g></svg>
 P bond. The same species is also accessible via stepwise H+/H− transfer to the bis-phosphenium complex, and releases H2 upon heating or irradiation. Dihydrogen transfer from the H2-complex to styrene is exploited to demonstrate the first example of promoting hydrogenation with a phosphenium complex.


Characterization of Complex 3b and isolation of co-crystals of 2/3b.
For spectroscopic characterization of 3b, a solution of complex 2 (20 mg, 20 μmol) in CD 3 CN/C 6 D 6 (6:1, total volume 0.7 mL) was sonicated in an ultrasonic bath for 15 min. Analysis of the 1 H NMR spectrum indicated the presence of a 70:30 mixture of 3a and 2. An incomplete set of NMR data of 3b could be determined by analysis of 2D NMR spectra of the equilibrium mixture. Crystals containing an 1:1 mixture of both species suitable for a single-crystal X-ray diffraction study were grown from a solution of 2 (40 mg, 40 μmol) in toluene/MeCN (4:14, 18 mL total volume). 3b: 1

Thermolysis of 2.
A solution of 2 (13 mg, 13 μmol) in C 6 D 6 (0.5 mL) in a valved NMR tube was homogenized by sonification for 10 min and then tempered at 80 °C for 12 h. NMR spectra recorded before and after thermolysis revealed conversion of 2 into 4 and H 2 ( Figure S1). Figure S1. 1 H (top) and 31 P{ 1 H} NMR spectra recorded before (red traces) and after (blue traces) tempering a solution of 2 in C 6 D 6 for 12 h at 80 °C. The 1 H NMR signal at 4.5 ppm is attributable to molecular H 2 .

Thermal reaction of 4 with H 2
A solution of 4 (7 mg, 7 μmol) in THF-D 8 (0.5 mL) was transferred to a medium-walled high pressure NMR tube and degassed by completing three freeze-pump-thaw cycles. The NMR tube was then pressurized with H 2 (8 bar) and kept at 60° C for 300h. NMR spectra recorded before and after heating showed that trace amounts of 2 had formed ( Figure S2). Figure S2. 1 H NMR spectra of a solution of 4 in C 6 D 6 before (blue trace) and after (red trace) heating for 300 h at 60 °C under 8 bar of H 2 . The signals attributable to the NCH signals of 4 and 2 are labelled by triangles and stars, respectively.

Photolysis of 4 under H 2 , D 2 , or a mixture of H 2 /HD/D 2 .
A solution of 4 (7 mg, 7 μmol) in THF-D 8 (0.5 mL) was transferred to a medium-walled high pressure NMR tube and degassed by completing three freeze-pump-thaw cycles. The NMR tube was then pressurized with the appropriate gas (8 bar H 2 , D 2 , or a mixture of all isotopomers, respectively) and subsequently irradiated for 7.5 h with a medium pressure mercury lamp. The reaction was monitored by recording NMR spectra before and after the irradiation period ( Figures S3 -S9).

Kinetic study of the photolysis of 2 and 4 under H 2 -pressure
A solution of 2 or 4 (7 mg, 7 μmol) in THF-D 8 (0.5 mL) was transferred to a medium-walled high pressure NMR tube and degassed by completing three freeze-pump-thaw cycles. The NMR tube was pressurized with H 2 , followed by acquisition of an initial 1 H NMR spectrum. The solution was then irradiated for different intervals with a medium-pressure mercury lamp (up to a total irradiation time of 270 min), and a 1 H NMR spectrum recorded after each interval. Evaluation of the product distribution at each point in time was performed by spectral integration of suitable signals using the residual signal of the deuterated solvent as reference. Results obtained under different H 2 -pressures are summarized in Tables S1 to S4. Evaluation of the data revealed that the consumption of starting material 2 during the initial reaction stages follows a pseudo-first-order rate law. The rate constants decrease with increasing H 2 -pressure without that a quantitative relation becomes apparent ( Figure S10). Further evaluation of the data was hampered, among others, by the appearance of an unidentified product arising most likely from decomposition.    [a] normalized in % of the total integral.

Reaction of 6 with D 2
A solution of 6-H 2 was generated photochemically from 4 and H 2 (8 bar) in a high pressure NMR tube as described above. The solution was then frozen in liq. N 2 and the NMR tube evacuated to remove gaseous H 2 and, after the solution had been allowed to warm up to ambient temperature, pressurized with D 2 (8 bar). The sample was agitated to ensure dissolution of D 2 in the liquid phase. A 1 H NMR spectrum run immediately afterwards showed that the NCH signals of 2 and 6 as well as the hydride signal of 2 were still visible whereas the hydride signal of 6 had disappeared, pointing out that 6 had undergone rapid H 2 /D 2 -exchange whereas 2 was obviously inert ( Figure S11).

Thermal conversion of 6 under H 2 atmosphere
Photolysis of 4 under H 2 (8 bar) in a high pressure NMR tube was carried out as described above. A 1 H NMR spectrum was recorded and the sample then kept at 40 °C for 336 h without further irradiation, with small interruptions in regular intervals to acquire NMR spectra for monitoring the progress of the reaction. Analysis of the composition of the mixture at each point in time was performed by evaluating the integrals of the signals of NCH-units, using the resonance of the deuterated solvent as internal reference. The results (Table S5, Figure S12) confirm that eventual conversion of 6 to form 2 and, to a minor extent, 4, took place. Table S5. Evolution of product distribution in a solution containing a mixture of 2/4/6 at 40 °C under 8 bar of H 2 .
[a] normalized in % of the total integral.

Thermal conversion of 6 under vacuum
A solution containing 6 was generated photochemically from 4 and H 2 (8 bar) in a high pressure NMR tube as described above. A 1 H NMR spectrum was recorded and the sample then frozen in liq. N 2 . The NMR tube was evacuated and then allowed to warm to ambient temperature, heated to 40 °C under autogenous pressure of the solvent, and kept at this temperature for 516 h without further irradiation with small interruptions to acquire NMR spectra for monitoring the progress of the reaction. The composition of the mixture was analyzed at each point in time by evaluating the integrals of the signals of NCH-units, using the resonance of the deuterated solvent as internal reference. The results (Table S6, Figure S13) confirm the eventual conversion 6→ 4. Table S6. Evolution of product distribution in a solution containing a mixture of 2/4/6 at 40 °C in the absence of gaseous H 2 .
[a] normalized in % of the total integral.  The sample was transferred to a medium-walled high pressure NMR tube and degassed by completing with three freeze-pumpthaw cycles. The NMR tube was then pressurized with H 2 (8 bar initial pressure). After acquisition of a reference 1 H NMR spectrum, the solution was irradiated for 11.5 h with a medium pressure Hg lamp, with short interruptions to acquire 1 H NMR spectra for reaction monitoring. Analysis of the composition of the mixture at each point in time was performed by evaluating the integrals of the signals of NCH-units and the CH 2 -signal of ethylbenzene, using a silicone resonance as internal reference (see Table S7). Near quantitative conversion of styrene to ethylbenzene (95% based on the original amount of styrene) occurred within 11.5 h of irradiation time ( Figure S14). Analysis of the phosphorus-containing species at the end of the reaction revealed the presence of 4 (45% based on the amount originally present), 6 (10%), secondary diazaphospholene 1 (24%) and a species X (21%) tentatively identified as P-phenethyl-substituted diazaphospholene arising from phosphination of styrene by 1 (Table S7, Figure S15, S18).

Hydrogenation of Styrene
(b) Hydrogenation of styrene (3 µL, 2.7 mg, 26 μmol) in THF-D 8 (0.5 mL) in the presence of 4 (1 mg, 1 μmol, 4 mol-% based on styrene) was carried out as described above to result in 71% conversion ( Figure S16). Analysis of the phosphorus-containing species at the end of the reaction indicated that only minor amounts of phosphenium complexes (8% of 4 and traces of 6) were still present and that extensive deactivation to give 1 (30%) and X (62%) was observable (Table S8, Figure S17). (c) To try identifying the catalytically active species in the hydrogenation, a sample containing styrene and 4 (16 mol-%) under H 2 (8 bar) was prepared as described above and irradiated for 20 min with a medium-pressure Hg lamp. An NMR-spectroscopic analysis confirmed that a significant amount of 6 (7%) and a small amount of ethylbenzene (16%; both values based on the amount of 4 initially present, see Figure S19) had formed. The solution was then stored at 20 °C for 18 h without further irradiation. Subsequent NMR spectroscopic analysis revealed that 4 had been completely consumed while the amount of ethylbenzene had increased ( Figure S19). Signals of 1 and X were not detectable, suggesting negligible decay of the NHP complex during the rather short photolysis time. Quantitative evaluation of the changes in signal integrals allowed us to calculate a turnover number TON = Δn(ethylbenzene)/Δn(6) = 9.6 which suggests that the dihydrogen complex transfers H 2 to the substrate in a thermal reaction and can perform several turnovers without additional photochemical activation.
(d) Control experiments were carried out using the same procedure as described above. Employing Cr(CO) 3 (naphthalene) (2 mg, 7.6 µmol, 16 mol-% based on styrene) as pre-catalyst resulted in 18 % conversion of styrene to ethylbenzene after 17 h of irradiation ( Figure S20), implying an essentially stoichiometric reaction. No formation of ethylbenzene was observed in the presence of 1 (2 mg, 4.9 µmol, 16 mol-% based on styrene, Figure S21) or in the absence of any catalyst ( Figure S22).

Crystallographic Studies
X-ray diffraction data were collected on a Bruker diffractometer equipped with a Kappa Apex II Duo CCD-detector and a KRYO-FLEX cooling device with Moradiation ( = 0.71073 Å) at 130(2) K for 2/3b (co-crystal containing both molecules in 1:1 ratio) and with Cu-radiation ( = 1.5406 Å) at 135(2) K for 5[BAr f 4 ]. The structures were solved with direct methods (SHELXS-97 7 ) and refined with a full-matrix least squares scheme on F 2 (SHELXL-2014 and SHELXL-97 7 ). Semi-empirical absorption corrections were applied. Non-hydrogen atoms were refined anisotropically and hydrogens atoms using a riding model. The three carbonyl and one hydride ligands in 2 are disordered on four positions around the chromium atoms. Refinement under application of SIMU and SUMP with free occupation on all four positions at Cr gave three CO ligands displaced on four positions. The position of the hydrogen atom on the metal could not be refined. However, the presence of significantly different P-Cr distances, different coordination environments on the phosphorus atoms and a freely refined hydrogen atom on the pyramidal phosphorous atom, as well as conclusive NMR data provide proof for a phosphenium phosphane hydride complex with a metal bound hydrogen atom. In general, the data of the 2/3b co-crystal are very weak (R int = 27.9 %). In the crystal of 5[BAr f 4 ], the fluorine atoms on three CF 3substituents in the anion displayed large anisotropic displacement parameters, and a refinement of disorder was performed by using split positions for each atom and refining the occupation of both positions with isotropic displacement parameters. The occupancies gained were then fixed and used for the anisotropic refinement of the disordered CF 3 groups. SADI was applied for C-F, F-F and 1,3-C-F distances. RIGU was applied for the displacement parameters, and ISOR for the disordered F atoms. CCDC-2005812 (2/3b) and CCDC-2005814 (5[BAr f 4 ]) contain the crystallographic data for this paper, which can be obtained free of charge from the Cambridge Crystallographic Data Centre via www.ccdc.cam.ac.uk/data_request/cif.

Computational Studies
General remarks. All computations were performed with the Gaussian 16 program package. 8 DFT calculations were carried out using the ωB97xD 9 functional that had previously been successfully applied for the study of hydrogenation of NHP complexes, 10 using an ultrafine grid for numerical integration and Weigend's and Ahlrichs' def2-tzvp basis sets. 11 The molecular structures were established by full energy optimization. Subsequent stability tests (keyword stable=opt) confirmed that the electronic states are not compromised by singlet/triplet or RHF/UHF instabilities. Harmonic vibrational frequency calculations were finally carried out at the same level to establish the nature of the stationary points obtained as local minima (only positive normal modes) or transition states (one imaginary normal mode), and to calculate standard Gibbs free energies ΔG 0 (referring to p=1 bar and T=298K). NBO population analyses were carried out using the NBO module implemented in the Gaussian package. The electronic spectrum of 4 Me was computed at the TD-ωB97xD/def2-tzvp level of theory. The relaxed molecular structure of the excited state was located by energy optimization of the first excited state identified in the TD-DFT calculation at the ωB97xD/def2-svp level of theory, and its energy recalculated at the final geometry at the ωB97xD/def2-tzvp level.