η2-Alkene Complexes of [Rh(PONOP-iPr)(L)]+ Cations (L = COD, NBD, Ethene). Intramolecular Alkene-Assisted Hydrogenation and Dihydrogen Complex [Rh(PONOP-iPr)(η-H2)]+

Rhodium-alkene complexes of the pincer ligand κ3-C5H3N-2,6-(OPiPr2)2 (PONOP-iPr) have been prepared and structurally characterized: [Rh(PONOP-iPr)(η2-alkene)][BArF4] [alkene = cyclooctadiene (COD), norbornadiene (NBD), ethene; ArF = 3,5-(CF3)2C6H3]. Only one of these, alkene = COD, undergoes a reaction with H2 (1 bar), to form [Rh(PONOP-iPr)(η2-COE)][BArF4] (COE = cyclooctene), while the others show no significant reactivity. This COE complex does not undergo further hydrogenation. This difference in reactivity between COD and the other alkenes is proposed to be due to intramolecular alkene-assisted reductive elimination in the COD complex, in which the η2-bound diene can engage in bonding with its additional alkene unit. H/D exchange experiments on the ethene complex show that reductive elimination from a reversibly formed alkyl hydride intermediate is likely rate-limiting and with a high barrier. The proposed final product of alkene hydrogenation would be the dihydrogen complex [Rh(PONOP-iPr)(η2-H2)][BArF4], which has been independently synthesized and undergoes exchange with free H2 on the NMR time scale, as well as with D2 to form free HD. When the H2 addition to [Rh(PONOP-iPr)(η2-ethene)][BArF4] is interrogated using pH2 at higher pressure (3 bar), this produces the dihydrogen complex as a transient product, for which enhancements in the 1H NMR signal for the bound H2 ligand, as well as that for free H2, are observed. This is a unique example of the partially negative line-shape effect, with the enhanced signals that are observed for the dihydrogen complex being explained by the exchange processes already noted.


INTRODUCTION
Pincer complexes of the group 9 metals (Co, Rh, and Ir) are used widely in catalysis. 1 One important process that they are used in is the catalytic dehydrogenation of alkanes to form the corresponding alkene, often (although not exclusively 2 ) using a sacrificial alkene to drive the reaction thermodynamics. 3,4 An elementary step in this overall process when using a sacrificial alkene is the reductive elimination of an alkyl hydride at a metal(III) center. Reductive elimination of C−C and C−H bonds at d 6 metal centers is often considered to operate through a five-coordinate intermediate, where a ligand dissociates prior to the reductive coupling event, if necessary. 5−8 However, there are reports of reductive elimination being promoted by association of an external ligand, when steric and electronic factors allow. 9−13 For example, Goldman has demonstrated that, for the neutral iridium(III), 16-electron, carbazolide-based pincer complex Ir(carb−PNP)H 2 , 14 reaction with excess ethene ultimately gives ethane and Ir(carb−PNP)(η 2 -H 2 CCH 2 ). Here reductive elimination from an intermediate ethyl hydride is promoted by coordination of exogenous ethene (Scheme 1A), or even more strongly by H 2 which then returns Ir(carb− PNP)H 2 instead.
In this contribution, we report the synthesis of new cationic [Rh(PONOP-i Pr)(η 2 -alkene)][BAr F 4 ] complexes [PONO-P-i Pr = κ 3 -C 5 H 3 N-2,6-(OP i Pr 2 ) 2 ; alkene = cyclooctadiene (COD), norbornadiene (NBD), ethene; Ar F = 3,5-(CF 3 ) 2 C 6 H 3 ]. When they are exposed to H 2 , a proposed intramolecular alkene-assisted reductive elimination leads to marked differences in both reactivity and selectivity in hydrogenation reactions, the outcome of which is dependent on the identity of the η 2 -bound alkene and, in particular, its ability to engage in bonding to the rhodium center with its additional alkene unit. We also report the independent synthesis of the dihydrogen complex [Rh(PONOP-i Pr)(η 2 -H 2 )][BAr F 4 ] and some highly unusual observations during monitoring of the reaction of the corresponding ethene complex with paradihydrogen (pH 2 ). This produces the dihydrogen complex as a transient product in which we, remarkably, observe enhanced 1 H NMR signals for the bound H 2 ligand, as well as for free H 2 , a consequence of the partially negative line-shape (PNL) effect that is in operation. 15 2. RESULTS AND DISCUSSION 2.1. Synthesis of [Rh(PONOP-i Pr)(η-L)][BAr F 4 ]. η 2 Binding of Alkenes. We have recently reported on the mechanism of amine−borane dehydrocoupling using the cationic rhodium- 16 This dihydrogen complex has previously been reported to be made by H 2 addition to the corresponding ethene adduct 17 or to the equilibrium mixture of [Rh(PONOP-t Bu) 2 (μ-η 2 ,η 2 -COD)][BAr F 4 ]/COD and [Rh-(PONOP-t Bu)(η 2 -COD)][BAr F 4 ]. 18 Our initial aim was to explore how the now well-documented 19,20 different steric demands of P i Pr 2 versus P t Bu 2 pincer arms could be harnessed in the synthesis and reactivity of pincer−dihydrogen complexes. We thus targeted synthesis of the precursor alkene complexes [Rh(PONOP-i Pr)(η-L)][BAr F 4 ] (L = COD, NBD, ethene; Scheme 2). As we show, this was not an appropriate route to afford a dihydrogen complex. Alkene (or alkyne) adducts of group 9 pincer complexes are useful precursors in synthesis and catalysis, often activated by hydrogenation of the alkene 17,18,21−23 ligand, and are also intermediates in alkane dehydrogenation reactions. 4,24 More generally, while complexes of PNP-i Pr [PNP = κ 3 -NC 6 H 3 -2,6-(P i Pr 2 ) 2 ] 25−27 and PONOP-i Pr 28−31 complexes are well established, the only group 9 example of the latter is a cobalt complex. 32 The ligand PONOP-i Pr (L1) was prepared in good yield (92%) and excellent purity (by 1 H NMR spectroscopy) as a colorless oil by a slight modification of the published method. 29  These three new PONOP-i Pr complexes were characterized by single-crystal X-ray diffraction and solution NMR spectroscopy. Figure 1 shows the solid-state structures of the cations. Despite containing dienes, both 1 and 2 display η 2 -alkene bonding at the 16-electron pseudo-square-planar rhodium(I) centers. The C−C bond lengths in the coordinating, and noncoordinating, alkene groups are fully consistent with this description. Although the three structures are broadly similar, they differ in the orientation of the alkene ligand with respect to the RhP 2 N plane. The COD ligand in complex 1 lies toward being upright (u), while for NBD (2) and ethene (3), an inplane (ip) conformation is seen. Both ip and u orientations of bound ethene have been observed for {ML 3 }-type group 9 PCP, POCOP, and PCNCN pincer systems, reflecting the interplay of steric and electronic factors at the metal center. 26,33−36 Within the consistent set of complexes reported here, we suggest that steric effects dominate because the more locally compact ethene and bicyclic NBD ligands adopt a different orientation (ip) compared with the larger cyclic COD (u). The u orientation of the COD ligand is the same as that observed for [{Rh(PONOP-t Bu)} 2 (μ-η 2 ,η 2 -COD)][BAr F 4 ]. 18 It is interesting to note that COD and NBD ligands bind almost exclusively in a bidentate η 2 ,η 2 -coordination mode to single metal centers, 37 although rare examples of η 2 binding are reported. 38−41 The NBD ligand binds through its exo face.
Solution NMR data are broadly consistent with these solidstate structures. For complexes 1 and 2, signals due to the unbound and η 2 -bound alkene units are observed (the latter resonances are upfield-shifted relative to the former) in the 1  in the room temperature NMR spectra, as shown by signals indicating equivalent (but individually diastereotopic) i Pr groups, a symmetric pyridine ligand, and a single 31 P environment [with coupling indicative of a rhodium(I) center in such a pincer complex 17,18,27,42 ]. Given the orientation of the COD (1) and NBD (2) ligands observed in the solid state, this suggests that a fluxional process operates in solution. A simple rotation 34 of the bound alkene best explains the observed symmetry in solution rather than a dissociative process because there is no exchange between bound COD and free COD signals on the NMR time scale as determined by exchange spectroscopy (EXSY) NMR experiments; both bound and free alkene environments are observed, and diastereotopic NBD-bridged methylene signals are observed for 2 (δ 1.72 and 1.22). This lack of observable diene dissociation in solution is in contrast with the recently reported F 4 ] in solution via dissociation of COD, which preferentially crystallizes as the COD-bridged dimer. 18 These differences are likely due to the different steric requirements of PONOP-t Bu versus PONOP-i Pr. 19,20 Such differences carry over into the reactivity with H 2 as discussed next.

Stoichiometric Hydrogenation in Solution. Role of the Distal Alkene as an Intramolecular Assisting
Ligand. The addition of H 2 (1 bar) to a CD 2 Cl 2 solution of complex 1 resulted in the relatively slow (2 h) but clean conversion to a complex in which partial, but selective, hydrogenation of COD to cyclooctene (COE) had occurred, to give [Rh(PONOP-i Pr)(η 2 -COE)][BAr F 4 ] (5; Figure 2). 43 Selective hydrogenation of a single CC unit is shown by a single, relative integral 2 H, alkene resonance being observed in the 1 H NMR spectrum (δ 4.65). The solid-state structure, as determined by single-crystal X-ray diffraction, shows a well-ordered COE ligand with bond lengths fully consistent with this semihydrogenation [i.e., C1−C2, 1.360(4) Å; C5−C6, 1.529(5) Å]. All other NMR and metrical data are very similar to that of the COD precursor, complex 1. Complex 5 can also be formed by the slow (24 h) reaction of complex 1 with a large excess of COE, displacing COD. We have not determined the mechanism, but the absence of products that signal a dissociative process, such as a bridged COD complex, suggests that an associative process operates. Associative substitutions at d 8 pincer complexes are well established. 21,44,45 In contrast to complex 1, exposure of the COE complex 5 to H 2 (1 bar, CD 2 Cl 2 ) over an extended period (48 h) returned complex 5 unchanged (Scheme 3). Identical behavior is observed for the NBD adduct 2. For the ethene complex 3, a small amount of decomposition (∼5%) to multiple unidenti-    17,18 is not observed under these conditions (vide infra).
Selective hydrogenation of the COD complex, but lack of observed reactivity of the NBD complex, is noteworthy because in cationic systems based upon [Rh(chelating diphosphine)(diene)] + NBD is well established to undergo stoichiometric hydrogenation faster than COD. 46,47 Informed by Goldman's studies on the reactivity of Ir(carb−PNP)H 2 with ethene, 14 where reductive elimination from an intermediate ethyl hydride was found to be high in energy (Scheme 1) in the absence of external alkene, we studied the reactivity of the ethene complex 3 with D 2 to probe the possibility of a similar situation occurring here (Scheme 4A). Over the course of 5 days, this resulted in the incorporation of a deuterium label into the bound ethene to give a mixture of isotopologues, 3-d n . This was shown by 1 H NMR spectroscopy (a ∼25% reduction in the relative integral of the η 2 -ethene resonance at δ 3.09 and a concomitant broadening due to H/D coupling), with a signal at δ 3.09 being observed in the 2 H NMR spectrum and assigned to bound ethene, and electrospray ionization mass spectrometry (ESI-MS), 48 which shows a mixture of isotopologues (3-d n , where n = 0−4; see the Supporting Information). These observations are consistent with the reversible oxidative addition of D 2 to 3 and hydride migration to give an (unobserved) ethyl deuteride. Subsequent β-H transfer, followed by the reductive elimination of HD [observed, δ 4.55, t, J(HD) = 43 Hz], returns isotopically enriched 3-d (Scheme 4B). Decomposition to a number of unidentified species also occurs over 5 days of exposure to D 2 , as noted for reactivity with H 2 .
We propose that, for the COD complex, where alkene hydrogenation is observed, reductive bond formation from the alkyl hydride intermediate is promoted by intramolecular coordination of the distal alkene. This is related to the Ir(carb−PNP)H 2 system, where coordination of exogenous ethene acts to promote reductive elimination, 14 or the similar role suggested for intramolecular C−H agostic interactions in H 2 reductive elimination from Ir(P t Bu 2 Ph) 2 (CCPh)(H). 49 50 and crystallographically characterized as Ir(σ,π-C 8 H 13 )(CO) 2 (AsPh 3 ). 51 This motif is unavailable for the ethene or COE complexes because of the lack of an additional alkene group. For the NBD complex 2, geometric constraints must mean that such a σ,π-intermediate is not accessible. The initial exo coordination of the alkene ( Figure  1) would result in an intermediate, exo-Rh-IV (Scheme 6) that is unable to partake in a σ,π-coordination mode, compared to endo-Rh-IV which we propose is not accessible. Exo addition of D 2 to norbornene (NBE) analogues of 2 is well established. 52−54 Attempts to promote this reactivity with another,    17,18 with associated formation of the corresponding alkane. We suggest that this points to a different, dissociative, mechanism operating for t Bu variants, consistent with the formation of η 2 η 2 -COD-bridged dimers, 18 and previous observations on preferred coordination motifs in Ir(POCOP-R)-based systems (R = t Bu, i Pr). 19 2.3. Catalysis. While the monoalkene complexes do not undergo appreciable reaction with H 2 on their own, in the presence of excess alkene catalytic turnover, albeit slow, can occur (Scheme 7). These are unoptimized conditions and simply demonstrate turnover rather than catalytic efficiency. For complex 1, at 5 mol % after 24 h, 78% conversion of the COD substrate to give a mixture of COE and COA (2.5:1 ratio) is observed, with an overall turnover number (TON) ∼ 20. Using the ethene complex 3, now at lower loadings (0.55 mol %), a similar TON for ethene hydrogenation is achieved after 24 h (TON = 27). For the NBD catalyst 2, there is no detectable hydrogenation of NBD. These observations are fully consistent with the role of additional alkene in promoting hydrogenation. Now, rather than intramolecular assistance, as seen for complex 1, the exogenous alkenes COD and ethene are promoters. With NBD as a substrate, we suggest that coordination of an additional NBD ligand at the metal center is disfavored because of its globular three-dimensional structure, and thus there is no overall hydrogenation. NBD is slowly hydrogenated to NBE in the presence of 1 equiv of COD per NBD using catalyst 3 (2.5 mol %). COD or COE can presumably coordinate to promote reductive elimination from a norbornyl intermediate, exo-Rh-IV. COE is also formed in this reaction from the competitive hydrogenation of COD.
2.4. Formation of the Dihydrogen Adduct 7. The lack of onward reactivity of the monoalkene complexes with excess H 2 under stoichiometric conditions might suggest that the expected product of such a reaction, 7, is not accessible. We show that this is not the case. Using an in situ halide abstraction route, 17  itate, presumed to be NaCl, is also observed to be formed. Complex 7 is unstable in solution, decomposing at 298 K under a H 2 atmosphere over 48 h to a mixture of products, likely those arising from reaction with the CD 2 Cl 2 solvent under these conditions. 56 In contrast, the t Bu analogue is reported to be more stable and can be prepared directly from the ethene precursor. 17 The independent synthesis of 7 suggests that the lack of reactivity of the monoalkene complexes with H 2 is a kinetic phenomenon. The small amount of decomposition observed for the alkene complex 3 under a H 2 atmosphere (ca. 5% after 48 h) may point to the slow formation of complex 7, vide infra, which then decomposes at a rate faster than its formation. Under these conditions, at 298 K, complex 7 is characterized by a broad, relative integral 2 H, whose signal is observed at δ −8.4 [full width at half-maximum (fwhm) = 150 Hz] in the 1 H NMR spectrum. Dissolved H 2 (δ 4.59) is also observed as a broad signal, suggesting exchange between free and bound H 2 that is likely to be both temperature-and pressure-dependent. Upon cooling, both of these signals sharpen and move further apart in frequency, supporting such a process. At 235 K (500 MHz), a T 1 (min) of 48 ± 6 ms is measured on the signal now at δ −9.20 (fwhm = 90 Hz) that clearly identifies 7 as a dihydrogen complex. However, this removal of the H 2 atmosphere also results in the decomposition of complex 7 to a range of products (50% decomposition in 20 min), an observation that supports the H 2 ligand being rather labile. Rapid freeze−pump−thawing of a CD 2 Cl 2 solution of 7 and then the addition of excess COE reforms complex 5 almost quantitively. When 7 is exposed to a D 2 atmosphere (10 min, 1 bar), HD(dissolved) is observed as a broad 1  57 That this dihydrogen complex 7 is not observed when H 2 is added to the alkene complexes, but there is H/D exchange into the bound ethene complex 3, indicates that reductive elimination of an alkyl hydride is the rate-limiting step, but this intermediate must be in endergonic equilibrium with the observed starting complex 3, i.e., Scheme 4. Interested in probing this further, we turned to the use of pH 2 to help identify any intermediates in this process. 58 2.5. Observation of the Dihydrogen Adduct under Catalytically Relevant Conditions Using pH 2 via the PNL Effect. When 3 is exposed to pH 2 at 298 K (3 bar) in a 5 mm J. Young NMR tube, 1 H NMR observation at 298 K reveals a broad H 2 signal at δ 4.6 (Figure 3a), which exhibits a PNL effect 15,59 when additionally observed with a 45°pulse. A relatively strong H 2 peak can also be observed at this position through the observe parahydrogen only spectroscopy (OPSY 60 ) sequence that selectively detects two spin-order terms while filtering out the thermal single-spin components. These two observations together indicate the presence of slow exchange between free and a rhodium-bound H 2 species. At this early temporal stage in the reaction, this is consistent with the reversible formation of [Rh(PONOP-i Pr)(η 2 -C 2 H 4 )(η 2 -H 2 )][BAr F 4 ] from complex 3 and/or its dihydride isomer [Rh(PONOP-i Pr)(η 2 -C 2 H 4 )(H) 2 ][BAr F 4 ] (e.g., Scheme 4b). 61 No complex 7 is observed that this point.
Repeating these measurements, at 3 bar of pH 2 , but then subsequent removal of the sample from the magnet, shaking under ambient laboratory conditions and return, resulted in the additional observation, at very low concentration, of transient 7, as indicated by the appearance of a broad signal at δ −8.3 (Figure 3b) at 248 K. This signal sharpens partially into a doublet of ∼24 Hz under 31 P decoupling conditions. Upon further cooling to 238 K, the size of the PNL signal of free H 2 increases, and, remarkably, after the application of a 45°pulse under 31 P decoupling conditions, the original signal at δ −8.3 that is assigned to 7 is now also observed as two antiphase signals, with frequency separations of 25 and 10.4 Hz 62 ( Figure  3c). The former is consistent with a J(RhH) splitting. This process of removal of the sample from the spectrometer, shaking, and return to observe both the antiphase free H 2 and high-field signals can be repeated a number of times. These data confirm that there is a substantial signal gain associated with these resonances from a PHIP effect (p-hydrogen-induced polarization), and we thus conclude that a metal complex of a dihydrogen ligand exhibits a clear PNL effect in a manner directly analogous to that of the free H 2 signal. For this PHIPenhanced dihydrogen complex 7, the high-field signal in the 31 P-coupled 1 H NMR spectrum exhibits the same intrinsic behavior but is now observed to be much broader, suffering from signal cancellation. 63 The detection of such character is indicative of the observation of a two-spin-order term, like that associated with the PASADENA 58 detection of an AX spin system formed by pH 2 addition. This leads to the creation of a longitudinal two-spin-order term that reflects the involvement of both pH 2 -derived proton spins and results in a flip-angle dependence to the observed response: zero with a 90°e xcitation pulse and maximized for the 45°pulse. We tested this hypothesis, and, as expected, when the same measurement was completed with a 90°excitation pulse, the two signals were found to lose their positive peak contributions but, rather than disappear, remained with negative line intensity (Figure 3d). This suggests that the competitive creation of net single-spin polarization in H 2 also occurs as a result of differential relaxation experienced by the two protons in what was pH 2 after binding and prior to the re-formation of free H 2 , as described by Aime under ALTADENA conditions. 64 The resulting single spin order is then visible through the action of the 90°excitation pulse in the normal way, although it appears with negative amplitude relative to the thermal Zeemanderived magnetization.
The pH 2 -enhanced NMR signal seen for both free and, remarkably, bound H 2 in this study reflect unusual, but Inorganic Chemistry pubs.acs.org/IC Forum Article theoretically predicted, behavior for a system that is under slow exchange. 15 pH 2 is simply NMR-silent molecular H 2 with singlet nuclear spin order. Three other spin orders that are associated with H 2 (T 0 ) are possible, also known as oH 2 , and are NMR-active. Normally, the NMR signal-intensity-enhancing effect PHIP is associated with the observation of stable pairwise H 2 oxidative addition products with magnetically distinct hydride ligands. 65,66 However, more complex behavior reflective of the one-proton PHIP effect, 67 and more recently the PNL effect, 15 can occur. The latter effect leads to the observation of a negative signal for dissolved molecular hydrogen and has been explained by singlet−triplet state conversion caused by exchange between H 2 and a transient dihydrogen complex. Such a scenario is shown in Scheme 9, in which there are a number of steps in the reaction manifold that forms complex 7 that can also account for the generation of H 2 (T 0 ), as supported by the H/D exchange experiments already described. In the specific case here, an additional exchange between the dissolved H 2 (T 0 ) and bound, but labile, H 2 in 7 can also lead to the remarkable signal enhancement seen in the high-field Rh(η 2 -H 2 ) signal of 7. Our observations also show that while 7 is formed slowly upon hydrogenation of 3, because it also exchanges very rapidly at 298 K with free H 2 , we fail to see an enhanced signal at this temperature, although it is visible under standard conditions. This is compounded by the decomposition of 7 at 298 K, which is competitive with its slow formation.

CONCLUSIONS
While hydrogenation of the simple alkene adducts of {Rh(pincer)(alkene)} is an attractive and expedient methodology to remove the alkene and generate a reactive dihydride/ dihydrogen complex, our observations here indicate that the identity of the alkene in such complexes with the PONOP-i Pr ligand can be crucial in onward reactivity and thus needs to be considered in complex or catalyst design. The lack of reactivity of the monoalkene complexes with H 2 is in contrast with the bulkier t Bu analogues, which react readily to form the corresponding dihydrogen complexes. Such observations are related, in a more general sense, to [Rh(phosphine)(diene)] + complexes, which are important precatalysts for a wide variety of transformations, and often activated by hydrogenation of the bound alkene ligand, the identity of which can also be important. 5,46,52 Moreover, the specific reactivity that the [Rh(PONOP-i Pr)(η 2 -ethene)] + complex presents (i.e., reversible H 2 addition and the slow formation of a dihydrogen complex that undergoes exchange with dissolved H 2 ) leads to an unusual situation where a signal enhancement due to the PNL effect 15 in the bound dihydrogen ligand can be observed as a consequence of the initial addition of pH 2 . While the original PHIP effect has been known since 1987 58 and is used widely to detect dihydride oxidative addition products at transition metal centers, to our knowledge, this is the first example where a dihydrogen ligand has been directly observed as an enhanced signal using pH 2 . While the precise reaction manifold that leads to this is still to be determined, the observation of this effect is noteworthy in itself and encourages further investigation.
■ ASSOCIATED CONTENT