Stoichiometric and Catalytic Solid − Gas Reactivity of Rhodium Bis-phosphine Complexes

.


INTRODUCTION
−7 Potential advantages of reactivity in the solid-state lead from the constrained local environment that may influence selectivity and the lack of possibly deleterious solvent interactions.Indeed, for the isolation of highly reactive organometallic species, the solid state might be considered as the "perfect solvent" in that it offers a potentially noncoordinating environment.Demonstrating these advantages, we have recently reported the synthesis of σ-alkane complexes by a solid−gas route, e.g.[Rh-( i Bu 2 PCH 2 CH 2 P i Bu 2 )(NBA)][BAr F 4 ] (2) [Ar F 4 = 3,5-C 6 H 3 (CF 3 ) 2 ], in which a saturated norbornane (NBA) fragment coordinates with the metal center through two 3center 2-electron interactions. 8,9Complex 2 was generated in situ in a single-crystal to single-crystal (SC−SC) transformation by simple addition of H 2 to a crystalline norbornadiene (NBD) precursor, [Rh( i Bu 2 PCH 2 CH 2 P i Bu 2 )(NBD)][BAr F 4 ] (1), Scheme 1.
−21 For the transformation of complex 1 to 2, that retains crystallinity and only has a 1.5% change in unit-cell volume, we postulated that packing in the lattice is dominated by bulky [BAr F 4 ] − anions which also form a defined octahedral cavity around 1 that is retained almost exactly in 2. The anions thus have a dual role to play by defining the lattice and also allowing for structural change in the organometallic cation (i.e., NBD to NBA).Complex 2 is unstable at room temperature in the solid s t a t e , e v e n t u a l l y f o r m i n g t h e z w i t t e r i o n [ R h -( i Bu 2 PCH 2 CH 2 P i Bu 2 ){η 6 -C 6 H 3 (CF 3 ) 2 }BAr F 3 ], 3, and free NBA, the thermodynamic products of the reaction. 8However, we posited that freshly prepared 2 could act as a suitable precursor for solid−gas reactivity, as the weakly bound alkane ligand could be easily displaced by an incoming ligand.In this contribution, we detail such reactivity with the exogenous ligands butadiene, ethene, CO, and NH 3 to form the corresponding adducts in the solid state and their onward reactivity.Some of these reactions are also SC−SC processes.We also document reactivity (H/D exchange, dehydrocoupling) and catalysis (hydrogenation of ethene) in these welldefined solid materials.
Addition of ethene gas to crystalline 2 forms a red, oily product assigned to the bis-ethene adduct [Rh- , and free NBA as measured by solution NMR spectroscopy (Scheme 3).
Analytically pure 6 can be generated by recrystallization from CH 2 Cl 2 /pentane while under 1 atm of ethene.Complex 6 is not stable at room temperature in solution in the absence of an ethene atmosphere.For example, in CH 2 Cl 2 solution, decomposition to 3 occurs, whereas in 1,2-C 6 H 4 F 2 complex 5 forms, making solution-based routes less than ideal.By contrast, solid 6 does not lose ethene even when placed under a vacuum, as indicated by elemental microanalysis of a crystalline sample.That 6 is stable to vacuum in the solid-state suggests that ethene loss might proceed via an associative mechanism within the crystal and not by loss of ethene followed by coordination of an exogenous ligand.Similar observations have been made previously for the substitution of a dinitrogen ligand at a {Ir(POCOP)} fragment in the solid-state [POCOP = 2,6bis(di-tert-butylphosphinito)benzene]. 5he 1 H NMR spectrum of complex 6 at 240 K in CD 2 Cl 2 reveals a sharp signal assigned to coordinated ethene (δ 3.98, fwhm = 7 Hz, relative integral 8 H), 24 while the 31 P{ 1 H} NMR spectrum at 240 K shows a doublet signal (δ 54.7, J RhP = 144 Hz).At room temperature under four atmospheres of ethene, the 1 H NMR spectrum shows a broad signal assigned to free ethene (δ 5.31, fwhm = 49 Hz), with no separate peak observed for coordinated ethene, suggesting rapid exchange between coordinated and free ethene in solution.The 31 P{ 1 H} NMR spectrum displays a single environment essentially unchanged from that at 240 K. Warming to 298 K in the absence of ethene results in loss of ethene to form 3. The solid-state 31 P{ 1 H} NMR spectrum shows two very broad environments, consistent with a disordered cation in the solid state (vide infra).
2.2.Synthesis of a Bis-carbonyl Complex by a Gas− Solid Reaction.Complexes such as 4 and 6 present ideal opportunities for solid-state reactivity, as the alkene ligands are likely to be relatively labile.Reaction with hydrogen in the solid state rapidly (less than 1 min) releases butane/ethane (observed by 1 H gas-phase NMR spectroscopy, see Supporting Information) with the concomitant formation of amorphous 3 as determined by 31 P{ 1 H} SSNMR spectroscopy, Scheme 4.
Presumably, the first formed alkane is lost from the metal center to result in [Rh( i Bu 2 PCH 2 CH 2 P i Bu 2 )] + coordinated with the [BAr F 4 ] − anion, with the associated large structural change leading to the loss of crystallinity.By contrast, addition of CO gas (1 atm) to solid 4 or 6 forms bright yellow [Rh( i Bu 2 PCH 2 CH 2 P i Bu 2 )(CO) 2 ][BAr F 4 ], 7, with release of 1,3-butadiene or ethene into the head-space of the reaction vessel (a sealed NMR tube) and no loss of crystallinity (Figure 1).Complex 7 can also be produced by addition of CO gas to 5 in solution (1,2-C 6 H 4 F 2 ).The solution 31 P{ 1 H} NMR spectrum of 7 shows a single resonance (δ 56.1, J RhP = 116 Hz).The 31 P{ 1 H} SSNMR spectrum also shows one environment, which is consistent with its solid-state structure (vide infra).The CO stretching frequencies in the IR spectra of 7 shows the anticipated two CO stretching bands (CH 2 Cl 2 , 2093.4 and 2049.1 cm −1 ; ATR InfraRed, 2099.0 and 2056.9 cm −1 ).These data are similar to closely related cationic Rh bis-  25 This addition of CO gas (1 atm) to a single crystalline sample is slow, with full conversion to yellow 7 requiring ∼3 days (for crystals of various sizes up to ∼1 mm 3 ).Interestingly, the reaction occurs from the surface of the crystal inward, and an even layer of the yellow carbonyl complex forms on the faces of the crystal, as shown by optical microscopy (Figure 1).If the CO atmosphere is removed after 3 h, well-defined crystals result of CO-Passivated-4 in which surface and near-surface sites are transformed into 7.The ratio of 4:7 after 3 h, as determined by 31 P{ 1 H} SSNMR spectroscopy, was approximately 1:1, although this is dependent on the crystal size/ surface area from batch to batch.Similar surface passivation using CO has been reported by Brookhart and co-workers, for solid−gas reactivity using the [Ir(POCOP)L] (L = N 2 , CO). 5 A single-crystal X-ray diffraction experiment was conducted on the final crystalline product of this solid-state reaction of 4 with CO, i.e., 7.These crystals appeared entirely yellow, retained their ability to rotate polarized light (which is often indicative of single crystallinity), and had unit cell parameters very similar to those of independently prepared 7 (vide infra).However, the mosaicity determined from the diffraction pattern was large and we were unable to solve the structure of the material produced via this solid-state route.The structure of 7, as determined by crystals obtained by solution routes, is discussed in section 2.4.The bis-ethene complex 6 also reacts in the solid-state with CO to form 7. In this case, the reaction takes only 3 h.The product of this reaction appears crystalline and diffracts with a single pattern, however, high mosaicity and cation disorder also did not allow the structure to be solved.
It is noteworthy that for a SC−SC transition to occur from 4 (or 6) to 7, not only must CO penetrate the crystal lattice but the gaseous butadiene (ethene) co-product must be able to escape without destroying the lattice.This suggests porosity within the crystalline network, which in some instances can come from loosely associated solvent that resides in channels in the crystal. 5In the materials discussed herein, no co-crystallized solvent is present.It might well be that the CF 3 groups of the [BAr F 4 ] − anions undergo a concerted motion that facilitates transport of CO and butadiene through the crystal, as has been recently suggested for reversible SC−SC transitions in Ag + coordination polymers which uptake alcohols. 213.Synthesis of a Bis-ammonia Complex.Addition of ammonia gas to freshly prepared 2, in the solid state, results in t h e f o r m a t i o n o f b i s -a m m o n i a c o m p l e x [ R h -( i Bu 2 PCH 2 CH 2 P i Bu 2 )(NH 3 ) 2 ][BAr F 4 ], 8, alongside free NBA (Scheme 5).Crystallinity is lost in this reaction.Complex 8 can be recrystallized from 1,2-C 6 H 4 F 2 /pentane solutions to afford crystals that were suitable for X-ray diffraction.To avoid traces of 3 in the final product, the direct addition of a mixture of NH 3 and H 2 to crystalline 1 was found to be the most effective way of synthesizing 8. Ammonia complexes similar to 8 have been previously reported, 26−28 in particular the closely related [Rh(dppe)(NH 3 ) 2 ] + .29 Solution methods using 1,2-C 6 H 4 F 2 solvent were unsuccessful for the synthesis of 8 with decomposition to unidentified species occurring, even though complex 8 is stable in 1,2-C 6 H 4 F 2 solvent once formed by the solid-state route.Complex 8 can also be synthesized by the relatively rapid reaction (1 h) of solid 4 with NH 3 gas (with release of butadiene gas).Neither of these routes are SC−SC reactions, with the product forming as an amorphous solid, in contrast to the analogous reaction of 4 with CO.The 31 P{ 1 H} NMR spectrum of 8 solvated in 1,2-C 6 H 4 F 2 displays a doublet (δ 65.5, J RhP = 177 Hz), similar to [Rh(dppe)(NH 3 ) 2 ] + .29 The 1 H NMR spectrum of 8 contains a sharp peak at δ 1.93 (relative integral 6 H) assigned to the coordinated ammine ligands.The proton chemical shift of ammine ligands is dependent upon hydrogen-bonding interactions with solvent or anion, 26 and in the case of 8 with a weakly coordinating anion in the weakly coordinating 1,2-C 6 H 4 F 2 solvent, the chemical shift observed is upfield of other reported values in similar complexes.26 Complex 8 does not react with ethene in the solid state but does react with D 2 , undergoing H/D exchange at the NH 3 ligand (Section 2.5).
2.4.Solid-State Structures.Single crystals of 4, 6, 7, and 8 were prepared by solution recrystallization (see Supporting Information).For complexes 4, 6, and 7, all three structures show significant disorder of the cation, which sits within an octahedral environment described by six, well-refined, [BAr F 4 ] − anions.The cations in these structures were refined by use of similarity restraints to model the disorder components.In the case of complex 6, the cation is disordered over two positions (∼2:1 occupancy ratio) in which the corresponding two RhP 2 planes lie approximately perpendicular to each other (angle between planes 81°), and only the heavy atom positions (Rh and P) could be reliably resolved.The cation of 7 is also disordered over two positions (1:1 occupancy ratio), adopting a structure in which one cation is inverted relative to the other, and the RhP 2 planes are again perpendicular to each other.Each cation straddles a crystallographic mirror plane, and so the CO ligands in each cation are crystallographically equivalent.Figures 2A and B show this in detail.For 4, refinement reveals four cation positions overlaid, with one rotated, one inverted, and one rotated and inverted relative to the original.These structures demonstrate the impact of rigid bulky anions upon relatively small cations: the anions dominate the packing structure and the cations may lie in a variety of thermodynami- Scheme 5. Synthesis of Complex 8 by Solid/Gas Reactions cally similar positions within the enclosed cavities. 30Although the anions refine well in these structures the cation disorder means that comments on bond lengths and angles are not appropriate.
By contrast, the solid-state structure of 8 is well-ordered.However, unlike 4, 6, or 7, the [BAr F 4 ] − anions do not form a well-defined octahedron around the cation.Instead a distorted octahedron is observed in which one axial [BAr F 4 ] − anion sits closer to the rhodium center than the others.This is demonstrated by consideration of the Rh−B vectors that lie approximately along crystallographic axes: ↑b 6.759(7) Å, ↓b 9.622(7) Å; a→ 9.593(1) Å, a← 9.593(1) Å; c→ 10.234(3) Å, c← 10.234(3) Å, Figure 3.This distortion along the b-axis may be influenced by hydrogen bonding between the ammine ligands and CF 3 groups on a proximal anion, as indicated by 31 Examples of hydrogen bonding in solid-state crystal structures of ammonia complexes have been reported, often between N−H and an oxygen atom, and N•••O distances vary from 2.84 to 3.19 Å (H•••O, 2.00−2.37Å). 26,27 The change in anion packing structure to accommodate N−H•••F hydrogen bonds could also cause the loss of crystallinity upon transition from 4 to 8.−34 and, most recently, within frozen argon matrices using Ir(Cp*)(H) 4 , however, in this case, a bound ammonia complex, Ir(Cp*)(H) 2 (NH 3 ), was only implied by low temperature (10 K) IR spectroscopic studies. 35o elucidate the mechanism of H/D exchange in 8, DFT calculations were employed to study H/H exchange between the ammine ligands of the isolated [( i Bu 2 PCH 2 CH 2 P i Bu 2 )Rh-(NH 3 ) 2 ] + cation, A, with added H 2 .The most accessible computed pathway is outlined in Figure 4, which shows free energies computed both in vacuo and corrected for 1,2-C 6 H 4 Cl 2 solvent (this being used in the absence of data for 1,2-C 6 H 4 F 2 ).Initial H 2 oxidative addition in solution proceeds with a computed barrier of 18.7 kcal/mol to give a Rh III Figure 4 shows the overall computed barrier for H/H exchange between cation A and H 2 is 22.8 kcal/mol, and it is apparent that the solvent correction reduces the barrier compared to the result in vacuo (ΔG ⧧ vac = 26.5 kcal/mol).This suggests increased charge separation in TS(C−D) compared to A, consistent with the involvement of a nearintact NH 4 + cation in the former.Experimentally, H/D exchange occurs more rapidly in solution than in the solid state.While we have not attempted to model the solid-state environment here, the proximity of several fluorinated anions may produce a similar environment to that in solution, with a resultant reduction in barrier height.Alternative mechanisms for H/D exchange were also considered, including those based on initial oxidative addition of ammonia 36,37 or heterolytic   2.6.Solid−State Ethene Dehydrocoupling.Over a number of weeks at 298 K, orange/red single crystals of freshly prepared 6 change to a darker red/purple, becoming similar in color to freshly prepared complex 4. When this material is dissolved in 1,2-C 6 H 4 F 2 solvent complex 4 is indeed observed in the 31 P{ 1 H} NMR spectrum, alongside the difluorobenzene adduct 5 (confirmed by ESI-MS).Complex 5 forms on dissolving 6 in 1,2-C 6 H 4 F 2 solvent (vide supra) and is thus a marker for this unstable bis-ethene complex or closely related complexes (Scheme 7).The transformation of 6 to 4 is formally a dehydrocoupling of ethene.This transformation in the solid state was initially monitored over time by periodic sampling via dissolution in 1,2-C 6 H 4 F 2 of a solid sample kept 298 K.This showed a gradual increase in the proportion of complex 4 present and a decrease in 5, with 40% conversion after 7 weeks (Figure 5).A sample kept for four months revealed ∼50% conversion to 4, this likely representing an upper limit for conversion.The color change of 6 to dark-red/ purple occurs without apparent loss of crystallinity (by inspection of the sample using a polarizing microscope), and a single X-ray diffraction pattern could be collected.However, we were unable to solve the structure fully due to a highly disordered cation.
A sealed sample of 6 (NMR tube) releases ethane gas over several weeks as observed by gas-phase NMR spectroscopy.This observation is consistent with the conversion of 6 to 4, liberating one equivalent of hydrogen, which would then react with unreacted 6 to produce ethane gas (i.e., Scheme 4).It thus appears 6 converts to 4, accompanied by sacrificial hydrogenation of one equivalent of ethene from a second molecule of 6, also accounting for the apparent 50% maximum reaction conversion.Interestingly, crystallinity is retained during this process, which further suggests the sacrificial 6 does not lose ethene completely, as the resulting 12-electron fragment would likely form amorphous-3 (Section 2.1).We suggest the formation of a 14-electron, 3-coordinate species [Rh- , in which only one ethene ligand is retained (Scheme 7) to account for the production of ethane.This species could be stabilized by an agostic interaction from the i Bu groups, 38 or by an interaction with the CF 3 groups from proximate anions in the lattice, and would likely form 5 on dissolution in 1,2-C 6 H 4 F 2 .Budzelaar and co-workers have previously reported a related threecoordinate rhodium complex with a bidentate β-diiminate ligand and one alkene ligand. 39Evidence for the formation of complex 9 comes from solvation (at 200 K, CD 2 Cl 2 ) of a solidsample sample of 6 that has been aged for two months.The 31 P{ 1 H} NMR spectrum recorded at this temperature reveals 4 to be the major product but less than 50% of the total integral, alongside signals for starting material 6 and its solution decomposition product 3. Two new signals are also observed, which are assigned to complex 9, which couple to each other ( 31 P− 31 P COSY experiment at 200 K), at δ 47.0 (J RhP = 162 Hz, J PP = 28 Hz); δ 71.4 (J RhP = 186 Hz, J PP = 28 Hz), Figure 6.
One signal has a considerably larger J RhP coupling constant, suggesting a weakly-bound trans-ligand (solvent or agostic interaction). 40On warming to 250 K, complex 3 grows in at the expense of complex 9.
SSNMR spectroscopy was also used to follow the solid-state transformation of 6 in situ.The 13 C{ 1 H} and 31 P{ 1 H} SSNMR spectra of freshly prepared complexes 4 and 6 were initially collected, and in keeping with their highly disordered solid-state structures, these spectra revealed multiple peaks for the signals related to the cation.For example, the 31 P{ 1 H} SSNMR spectrum of 4 shows a variety of closely spaced broad peaks, while that of 6 displays at least three broad phosphorus environments in the solid state, Figure 7.The 13 C{ 1 H} SSNMR spectra of 4 and 6 also show multiple peaks for the butadiene and ethene environments in their respective spectra, between 60−100 ppm.Monitoring a sample of 6 over the course of 6 weeks revealed the formation of complex 4 in both the 31 P{ 1 H} and 13 C{ 1 H} NMR spectra.In addition, a new signal (δ 82.4) slightly downfield to the ethene resonances in 6 was also observed in the 13 C{ 1 H} NMR spectrum, while the 31 P{ 1 H} SSNMR spectra show additional broad signals centered at δ 77 (Figure 7D).These signals are assigned to [Rh- 4 ], 9 (L = vacant site, agostic or anion interaction).On exposure to ethene gas, these additional signals disappear and 6 is reformed, consistent with the formulation of 9 as having a vacant, or masked, site (Figure 7E).When a sample of 6 was left at 323 K under an atmosphere of ethene (sealed NMR tube, 6 days), a small amount of 2-butene was observed.This could arise from isomerization of 1-butene, which would be formed by dimerization of ethene. 41Under these conditions of temperature and excess ethene, no 4 is formed from the sample of 6, with orange 6 recovered and a trace (∼10%) of a complex characterized as a hexadiene adduct, 10.Complex 10 was characterized by an independent synthesis from hexadiene and complex 5 (Supporting Information).Attempts to make the reaction turnover in a catalytic sense by increasing the temperature (e.g., 363 K) led to a partial melting of the crystalline material and no significant increase in the formation of coupled products.
−44 Such processes are also related to alkene oligomerization reactions. 45Interestingly, the complex RhTp*(C 2 H 4 ) 2 [Tp* = tris(3,5-dimethyl-1-pyrazol-1yl)hydroborate] undergoes dehydrocoupling to give a butadiene complex in solution (with the concomitant release of ethane), but in the solid state, a hydrido-allyl species is formed. 44FT calculations on isolated cations were again used to assess these different mechanistic possibilities and our favored pathway is outlined in Figure 8A.Starting from E, [( i Bu 2 PCH 2 CH 2 P i Bu 2 )Rh(C 2 H 4 ) 2 ] + (i.e., the cation of 6), oxidative coupling of ethene occurs with a barrier of 22.9 kcal/ mol to give intermediate F (+10.8 kcal/mol).F exhibits a seesaw structure for which there are precedents with isoelectronic d 6 Ru(II) and Ir(III) centers. 46,47β-H transfer then generates G from which facile C−H reductive coupling yields 1-butene complex H (−8.1 kcal/mol).Rearrangement and γ-H transfer then leads to hydrido-methylallyl intermediate   I at −11.6 kcal/mol. 48I features a weak Rh•••H−C agostic interaction and so could undergo H-transfer to form a trans dihydride intermediate (an isomer of J, see below, and Supporting Information, Figure S20).Instead a lower energy process involving facile rearrangement of the methylallyl ligand places the agostic cis to the hydride in isomer I′ (−6.2 kcal/ mol) from which H-transfer then gives J with cis hydrides and a cis η-C 4 H 6 ligand.Dehydrocoupling is then completed via reductive elimination of H 2 to give K (the cation of 4) at −3.7 kcal/mol.
Figure 8A shows the formation of I is thermodynamically favored over K, and so the facile, exergonic hydrogenation of E to N (modeling the cation of species 9 postulated experimentally) is required to drive the dehydrocoupling to completion (see Figure 8B: overall barrier = 13.8 kcal/mol; ΔG = −16.0kcal/mol).The optimized structure of N features an agostic interaction between the Rh center and one of the i Bu methyl groups (see Supporting Information, Figure S11).The barrier energy span 49 of the dehydrocoupling profile (corresponding to I to TS(M-N)) is 25.4 kcal/mol, a value that is consistent with a slow dehydrocoupling process.The intermediacy of H is also consistent with the observation of 2-butene formed (after isomerization) in the presence of excess ethene.
Alternative dehydrocoupling mechanisms based on initial C− H activation to hydrido-alkenyl intermediates were also assessed.The C−H activation step (ΔG ⧧ = 24.1 kcal/mol; ΔG = +19.3kcal/mol) proved only slightly less accessible than the oxidative coupling; however, the onward reaction, either via ethene insertion into the Rh−alkenyl bond or insertion into the Rh−H bond involved transition states at 31.7 or 31.0 kcal/mol.Thus, these pathways have considerably higher barriers than the process based on oxidative coupling in Figure 8A.Full details of all alternative pathways are provided in the Supporting Information.
2.7.Solid-State Catalysis: Hydrogenation of Ethene.The complexes 1−8 were tested as solid-state catalysts for the hydrogenation of ethene.Such solid−gas reactivity has been reported before for organometallic catalysts in the solid state. 5,50−52 Gas-phase NMR spectroscopy was used to measure the hydrogenation of ethene in situ by the conversion of ethene to ethane.T 1 relaxation times for ethane and ethene in the gas phase were found to be similar for both species (ca.0.6 s), as were their relative integrals with different acquisition delay times, suggesting that the relative integration of the two signals in to the gas-phase species is reliable.In a typical experiment, a high pressure NMR tube was loaded with (3 ± 0.3) mg (e.g., 2.2 × 10 −3 mmol for 1) of solid catalyst, and to this 1 atm (∼0.08 mmol) of ethene was added followed by ∼4 atm (∼0.31 mmol) of H 2 .The progress of the reaction was monitored directly by integration of the gas-phase 1 H NMR spectra at 298 K.
Under these conditions of excess hydrogen, precatalysts containing alkene ligands (1, 4, and 6) all catalyze the reaction rapidly; for example, 1 requires less than 2 min to effect complete conversion (Figure 9).At the end of catalysis, when the ethene is fully consumed and hydrogen is still in excess, complex 3 is observed to be formed as characterized by 31 P{ 1 H} NMR spectra in CD 2 Cl 2 solution, as there is no exogenous ligand to stabilize the metal center, consistent with the stoichiometric studies (section 2.1).This zwitterionic complex 3 is a very slow catalyst itself, only turning over ∼5% after 25 min, suggesting that dissociation, or η 6 -to η 2 -ring slippage, of the anion to reveal a vacant site is a disfavored process in the solid state.This is a solid-state phenomenon, as in CH 2 Cl 2 solution under an atmosphere of ethene 3 forms 6, and hydrogenation of the bound alkene is then rapid (upon mixing).This also suggests that the resting state during catalysis in the solid state when using precatalysts such as 1 is unlikely to be 3. Complex 7 is a very slow catalyst, showing essentially no turnover after 30 min, presumably as the carbonyl ligands are strongly bound.Complex 8 does show some activity (15% conversion after 30 min), perhaps suggesting that the ammine ligands are somewhat labile, consistent with the computational studies.In both of these cases, the organometallic species observed after completion of catalysis to the detection limit of solution 31 P{ 1 H} NMR spectroscopy is the same as the starting material, i.e., unchanged 7 and 8. Interestingly, for precatalysts 2, 5, and 6, the rate of hydrogenation, which is at first rapid, drops after around 2 min.This similar slower rate for all may be indicative of transformation of any surface organometallic species into a common species (possibly related to 3) that turns over slowly.Also noteworthy is that this simple hydrogenation reaction occurs rapidly at ambient temperature (∼5 atm total pressure), while previous reports of the catalytic hydrogenation of ethene using solid-state organometallic catalysts typically require longer time scales and/or higher temperatures (albeit with lower pressure reaction conditions). 5,52,53s the surface area of a solid catalyst is likely to be important for the rate of catalysis, microcrystalline 1 was ground and separated into different particle size fractions using microsieves to give the following particle size regimes: less than 50 μm, 50− 71 μm, and 71−150 μm.These materials were then subjected to catalytic hydrogenation of ethene using the standard conditions previously defined (i.e., 3.0 ± 0.3 mg of catalyst).Figure 10 shows that all the different size particles rapidly hydrogenate ethene with broadly similar temporal profiles over the first 10 min, promoting between 75 and 90% conversation.After a much longer time (16 h), all reached 100% conversion but at a much slower rate.Interestingly, the smaller particle sized catalysts appear to operate slightly faster as an ensemble, suggesting that the surfaces of the crystals are particularly active compared to the interior, or that smaller particle sizes allow for more rapid hydrogenation of the precatalyst.An alternative explanation is that the larger particle sizes of catalyst become deactivated sooner as there are relatively fewer active sites.As it might well be that all these factors are operating, possibly with a different temporal dependence depending on crystal size, and given the differences are relatively small, we are reluctant to speculate further with the current data.
2.8.Solid-State Catalysis: Passivation Studies.The data presented in the previous section suggest that catalysis might occur fastest upon the surface sites of the organometallic crystals.However, it is also likely that the metal centers located inside the interior of the crystal could be active by gas diffusion through the crystalline lattice.To probe this, we used CO-Passivated-4, an approach encouraged by the passivation techniques used by Brookhart and co-workers in studying alkene hydrogenation using [Ir(POCOP)(H) 2 ] systems. 5As complex 7 essentially does not catalyze the hydrogenation of ethene, while 4 catalyzes the reaction rapidly (Figure 9), the CO-Passivated-4 probes the ability of these interior metal sites to take part in catalysis.However, as complex 4 reacts with H 2 in the absence of ethene to, deleteriously, form 3, in order to maintain single crystallinity, and thus the surface passivation of CO-Passivated-4, ethene is kept in excess with H 2 the limiting substrate in contrast to previously discussed catalysis in which hydrogen is in excess.
To baseline studies, complexes 4 and 7 were subjected to these conditions of catalysis in which hydrogen is the limiting substrate.A high pressure NMR tube containing (3 ± 0.3) mg of crystalline catalyst 7 was evacuated and refilled with ∼1 atm of ethene and ∼0.3 atm of H 2 .As expected, single crystals of complex 7 are very poor catalysts for the hydrogenation of ethene under these conditions and only 3% conversion occurs within 1 h and little further progress is observed over the next 24 h (Figure 11).By contrast, single crystals of 4 hydrogenate ethene, until all hydrogen is consumed, in only ∼9 min (i.e., 29% of ethene hydrogenated).These crystals retain their crystallinity and their ability to rotate polarized light.A second pressurization of hydrogen can be applied, and catalysis continues until the hydrogen is used up a second time (42% of ethene now being hydrogenated).Analysis of the resulting crystals by low temperature solution 31 P{ 1 H}NMR spectroscopy (220 K, CD 2 Cl 2 ) afterward showed that only 4 was present, and no 6 or 3 were resolved to the detection limit of 31 P NMR spectrocopy (∼5%).This suggests that catalysis in the crystalline state occurs without significant involvement of the bulk crystal when ethene is in excess, with presumably only a small number of sites, likely on the surface, active for catalysis.
Under these conditions, CO-Passivated-4 also catalyzes the hydrogenation of ethene but at a considerbly slower rate than pure crystalline 4, with only 14% conversion reached after 33 min and the reaction only reaching completion after 12 h compared with the 10 min for 4 (Figure 11).After catalysis, the crystals appear intact and in good condition without significant cracking or change.A small trace of butane was observed in the gas phase 1 H NMR spectrum (∼1% relative to the ethene signal), suggesting that a small proportion of the interior sites of 4 have been hydrogenated.A second amount of hydrogen can be subsequently added, and catalysis restarts.Low temperature 31 P{ 1 H} NMR spectroscopic analysis (220 K, CD 2 Cl 2 ) of these crystals postcatalysis resolved only 4 and 7 to be present.This much slower rate of catalysis suggests that catalysis is mainly a surface process in these systems.However, that catalysis still does occur suggests that there is some penetration of ethene and hydrogen to access active interior sites of a crystal.These sites must be particularly active as no significant spectroscopic markers for their formation, i.e., 6 or 3 are observed.We cannot discount the possibility of crystal microcracking occurring (due to reaction with hydrogen and formation of 3), which may expose the interior of the crystals. 54nfortunately, the remarkable selective hydrogenation of ethene in the presence of propene, demonstrated by Brookhart and co-workers using passivated single crystals, was not able to be reproduced with crystals of CO-Passivated-4. 5Using a mixture of ethene, butene, and hydrogen (approximate 2:2:1 ratio), an approximately equal ratio of butane and ethane was produced in each case with crystalline catalysts of 4, CO-Passivated-4 and 7, although the reaction rates varied dramatically as expected (4 > CO-Passivated-4 > 7).

CONCLUSIONS
A number of solid-state reactions of well-defined organometallic complexes with gases are reported based upon the fragment [Rh( i Bu 2 PCH 2 CH 2 P i Bu 2 )][BAr F 4 ].In particular, we have shown that the σ-alkane complex 2 is a useful synthon for such solid-state synthesis, acting via a labile alkane ligand.This allows for the addition of gases such as ammonia and ethene to form the corresponding adducts, some of which, for example with ethene or NH 3 , are not readily accessible via solution techniques.The corresponding butadiene complex, 4, acts in a similar way by loss of alkene.Crystallinity may be maintained during some of these solid−gas ligand exchange reactions, when the crystal packing geometry remains similar in the starting materials and products (e.g., the displacement of ethene or butadiene ligands by CO).The crystal packing environments, which incorporate large [BAr F 4 ] − anions that  Under these conditions, a maximum conversion of ∼30% of ethane is achievable.
provide a pseudo octahedral cavity in which the cations reside, can thus allow for subtle rearrangements in the solid state without disruption of the crystallinity or onward reactivity with the metal cation, 55 as we have noted previously. 8,9In the solid state, these complexes can be used as precatalysts for the hydrogenation of ethene, show H/D exchange at bound NH 3 ligands, and promote ethene dehydrocoupling.Mechanisms for these last two processes are proposed on the basis of DFT calculations using molecular model systems, and modeling such reactivity in the solid state is the focus of current work.For catalysis, particle size and surface passivation experiments suggest that species on the surface act as the most active sites for catalysis rather than within the crystal.These observations encourage the use of solid-state organometallic chemistry in both synthesis and catalysis.Of course, in the latter, the identification of the actual active species/sites will likely be challenging as these might only represent a relatively small proportion of the bulk crystal.

EXPERIMENTAL SECTION
4.1.General Details.All manipulations, unless otherwise stated, were performed under an atmosphere of argon, using standard Schlenk-line and glovebox techniques.Glassware was oven-dried at 403 K overnight and flamed under vacuum prior to use.CH 2 Cl 2 , Et 2 O, and pentane were dried using a Grubbs-type solvent purification system (MBraun and degassed by successive freeze−pump− thaw cycles. 56CD 2 Cl 2 and 1,2-C 6 H 4 F 2 were distilled under vacuum from CaH 2 and stored over 3 Å molecular sieves.Complexes 1, 2, 3, and 5 were prepared according to previously described methods. 8All other reagents were used as received from suppliers.Solution and gasphase NMR spectra were recorded on Varian Unity 500 MHz, Bruker AVD 500 MHz, or Varian Mercury 300 MHz spectrometers at room temperature unless otherwise stated.Nondeuterated solvents were locked to a standard C 6 D 6 solution.Residual protio solvent was used as reference for 1 H, 2 H, and 13 C NMR spectra in deuterated solvent samples.In 1,2-C 6 H 4 F 2 , 1 H NMR spectra were referenced to the center of the downfield solvent multiplet (δ 7.07). 31P NMR spectra were externally referenced to 85% H 3 PO 4 .All chemical shifts (δ) are quoted in ppm and coupling constants in Hz.Gas-phase samples were separately referenced to the reported values for hydrogen, ethane, ethene, and butadiene (all data were fully consistent with reported values). 52,57SSNMR spectra were recorded on a Bruker Avance 3HD 400 MHz (University of Oxford) and Varian VNMRS 400 MHz (EPSRC National SSNMR Service, Durham University), using the cross-polarization pulse sequence (contact time 3−10 ms and recycle delay 1−2 s), with Spinal 64 or TPPM decoupling.A magic-angle spinning rate of 10 kHz was employed (under dry N 2 ).Spectral referencing is with respect to external neat tetramethylsilane for 13 C and to 85% H 3 PO 4 for 31 P. Fourier Transform infrared (IR) spectroscopy was carried out using a Thermo-Scientific Nicolet iS5 with iD3 ATR and iD1 transmission attachments.Electrospray ionization mass spectrometry was recorded using a Bruker MicroTOF instrument directly connected to a modified Innovative Technology glovebox. 58 F 4 ] (5) (250 mg, 0.179 mmol) was dissolved in 1,2-C 6 H 4 F 2 and the solution frozen and placed under vacuum.Butadiene gas was used to refill the flask and upon warming a color change to a dark-red/purple solution occurs, the flask was left sealed for 2 h.Because reducing the pressure inside the flask drives the backward reaction, releasing butadiene gas and the formation of 5, complex 4 was recrystallized by adding pentane via syringe directly to the solution under an atmosphere of butadiene.Dark-red/purple crystalline material forms, 66% yield.

Figure 1 .
Figure 1.Reaction of 4 with CO in the solid state to form complex 7. Optical microscope pictures of single crystals of 4, CO-Passivated-4 and 7 (bottom).Grid size = 0.5 mm.
2.5.H/D Exchange in Complex 8. Microcrystalline samples of 8 react with D 2 in the solid state (48 h), or in 1,2-C 6 H 4 F 2 solution (24 h), to form d 6 -8 with H/D exchange occurring exclusively at the ammine ligands (Scheme 6).When dissolved in 1,2-C 6 H 4 F 2 , the 1 H NMR spectrum of d 6 -8 reveals that the protio ammine signal observed for 8 [δ 1.93] is absent, while the 2 H NMR spectra contains a signal corresponding to the ND 3 ligands (δ 1.94).No signals indicating H/D exchange at the phosphine ligands were observed.No degradation of the crystal was observed during this process, as confirmed by X-ray crystallography.H/D exchange of bound ammonia has been previously reported in solution dihydride, B (G solv = +3.3kcal/mol).The NH 3 ligand trans to hydride in B is labilized (Rh•••N = 2.27 Å) and so readily dissociates with a barrier of only 13.0 kcal/mol to give an isomer C (G solv = +9.8kcal/mol) in which it now H-bonds to the remaining ammine ligand.Rh−H/N−H exchange then proceeds through TS(C− D) at +22.8 kcal/mol, in which the outer-sphere NH 3 deprotonates the axial hydride

Figure 2 .
Figure 2. (A) Detailed view of the anion-cage and disordered components in complex 7 (ball and stick representation for simplicity, anions displayed as translucent, hydrogens atoms omitted).(B) Disordered cation in 7 (displacement ellipsoids at 30%, one disorder component displayed with pale atoms and bonds, hydrogens omitted).(C) Simplified packing diagrams of 6, 7, and 4, showing disorder of the cation position.Cations are represented by {RhP 2 } only, anions represented as {BC 4 } only.

Scheme 6 .
Scheme 6. H/D Exchange upon Ammine Ligands in Solution or the Solid State

Figure 4 .Figure 5 .
Figure 4. (A) Computed profile for H/H exchange via the reaction of [( i Bu 2 PCH 2 CH 2 P i Bu 2 )Rh(NH 3 ) 2 ] + , A, with H 2 .Relative free energies (kcal/ mol) are reported in vacuo (plain text) and corrected for 1,2-C 6 H 4 Cl 2 solvent (in italics).H centers derived from the added H 2 molecule are highlighted in blue.(B) Computed structure of the Rh−H/N−H exchange transition state TS(C−D) with selected distances in Å. i Pr groups and nonparticipating H atoms on the chelating phosphine are omitted for clarity.

Figure 6 .
Figure 6.Low temperature 31 P{ 1 H} NMR spectrum (CD 2 Cl 2 , 200 K) of a solid sample of 6 that has been aged for 2 months.

Figure 7 .
Figure 7. SSNMR spectra ( 13 C{ 1 H} left, 31 P{ 1 H} right) displaying the transformation of 6 over time (A−D) followed by the reaction of the products with ethene gas (E) at 298 K and a comparison with a pure sample of 6 (F).

Figure 8 .
Figure 8. Free energy profiles (kcal/mol) for (A) the dehydrocoupling of ethene in [( i Bu 2 PCH 2 CH 2 P i Bu 2 )Rh(C 2 H 4 ) 2 ] + , E, and (B) hydrogenation of E. Several processes involve more than one step and these are indicated with a double arrow, and only the energy of the highest lying transition state between the two key minima is indicated.Full details of these and alternative mechanisms are available in the Supporting Information (see Figure S18−S24).

Figure 9 .
Figure 9. Hydrogenation of ethene to ethane using a variety of solidstate catalysts, 298 K. %Ethane conversion calculated from integrals of ethane and ethene signals by gas phase NMR spectroscopy.

Figure 11 .
Figure 11.Hydrogenation of ethene using crystalline precatalysts.Ethene added at 1 atm, hydrogen added at less than 1 atm (∼0.3 atm).Under these conditions, a maximum conversion of ∼30% of ethane is achievable.