The role of neutral Rh(PONOP)H, free NMe₂H, boronium and ammonium salts in the dehydrocoupling of dimethylamine-borane using the cationic pincer [Rh(PONOP)(η²-H₂)]⁺ catalyst

Abstract

The σ-amine-borane pincer complex [Rh(PONOP)(η¹-H₃B·NMe₃)][BAr^F₄] [2, PONOP = κ³-NC₅H₃-2,6-(OP^tBu₂)₂] is prepared by addition of H₃B·NMe₃ to the dihydrogen precursor [Rh(PONOP)(η²-H₂)][BAr^F₄], 1. In a similar way the related H₃B·NMe₂H complex [Rh(PONOP)(η¹-H₃B·NMe₂H)][BAr^F₄], 3, can be made in situ, but this undergoes dehydrocoupling to reform 1 and give the aminoborane dimer [H₂BNMe₂]₂. NMR studies on this system reveal an intermediate neutral hydride forms, Rh(PONOP)H, 4, that has been prepared independently. 1 is a competent catalyst (2 mol%, ∼30 min) for the dehydrocoupling of H₃B·Me₂H. Kinetic, mechanistic and computational studies point to the role of NMe₂H in both forming the neutral hydride, via deprotonation of a σ-amine-borane complex and formation of aminoborane, and closing the catalytic cycle by reprotonation of the hydride by the thus-formed dimethyl ammonium [NMe₂H₂]⁺. Competitive processes involving the generation of boronium [H₂B(NMe₂H)₂]⁺ are also discussed, but shown to be higher in energy. Off-cycle adducts between [NMe₂H₂]⁺ or [H₂B(NMe₂H)₂]⁺ and amine-boranes are also discussed that act to modify the kinetics of dehydrocoupling.

---

Introduction

The dehydrocoupling of amine-boranes, H₃B·NRR′H (R, R′ = H, alkyl), is an efficient way to produce new molecules and materials with B–N bonds, with hydrogen as the only by-product.^1–5 Catalytic routes, mainly mediated by transition metal complexes, offer the possibility to influence kinetics and product distributions. This is especially important for the controlled dehydropolymerisation^6–11 of primary amine-boranes, prototypically H₃B·NMeH₂, that form polymeric materials with main-chain B–N units, (BHNMe)_n, that are isosteres of polyolefins and precursors to BN-containing materials.^12–14 Challenges remain,¹⁵ in determining the precise mechanism of dehydropolymerisation which, at the current level of understanding, appears to be a complex and nuanced process. These are exemplified by: (i) the low catalyst loadings required to selectively produce polymers that hamper speciation studies, (ii) polymeric material that becomes insoluble at high molecular weight or with cross-linking, (iii) complex kinetics that often involve induction periods and modification by the hydrogen co-product, (iv) apparent changes in the precise mechanism dependent on the identity of the precatalyst, and (v) a general overarching process that requires two elementary transformations that need to work in concert: dehydrogenation of amine-borane to aminoborane and subsequent controlled polymerisation (Scheme 1).

Scheme 1 Generalised dehydrocoupling of methyl amine-boranes. R = H or Me.

The dehydrocoupling of the secondary amine-borane H₃B·NMe₂H offers a more straightforward platform to study these processes as the product is a simple soluble dimer, [H₂BNMe₂]₂ (Scheme 1), and selectivity for its formation is generally less strongly influenced by catalyst loading, meaning that catalyst speciation and kinetics are easier to study using techniques such as NMR spectroscopy. This provides opportunities to study, in closer detail, the elementary processes occurring in dehydrocoupling, with the caveat that the extra N-methyl group may influence both the kinetics and speciation of the catalyst when compared with H₃B·NMeH₂.^16,17

Fundamental to any mechanism for dehydrocoupling is the initial dehydrogenation step to form an aminoborane via BH/NH activation and loss of H₂. Depending on the catalyst system a number of different routes have been proposed to operate for this, that all invoke σ-amine-borane complexes^18,19 as early intermediates (Scheme 2): (i) step-wise, or concerted, inner sphere;^17,20–22 (ii) ligand-assisted cooperation;^23–26 and (iii) hydride transfer to form a boronium²⁷ cation that reprotonates the transiently formed hydride.^28–31

Scheme 2 Generic amine-borane dehydrogenation pathways at a metal centre. The structures of the intermediates are illustrative and do not necessarily capture the order of elementary mechanistic steps for each isolated event.

The initial reports of dehydropolymerisation of H₃B·NMeH₂ used the neutral pincer catalyst Ir(POCOP)H₂^7,8,16 [A, POCOP = κ³-C₆H₃-2,6-(OP^tBu₂)₂], Scheme 3. A number of closely related pincer-based systems have since been used to catalyse amine-borane dehydrocoupling.^{21,26,32–35} However to date no σ-amine-borane complexes have been reported with such systems, despite their key role in catalysis. Related off-cycle products have been characterised.³⁶ Analogous POP ligand complexes (POP = e.g. Xantphos) have a richer coordination chemistry with amine-boranes, and η¹ (e.g.B) and η²:η²-systems have been characterised that are also relevant to the dehydrocoupling of amine-boranes^30,37,38 – although they may not actually lie on the catalytic cycle.¹¹ For POCOP or PONOP-type systems the preparation, and deployment in catalysis, of a σ-amine-borane complex would provide valuable insight into the mechanism of dehydrocoupling.

Scheme 3 Examples of pincer complexes used in amine-borane dehydrocoupling and complex 1 used in this contribution.

We now report that by use of the readily prepared cationic-precatalyst [Rh(PONOP)(η²-H₂)][BAr^F₄]³⁹ [1, Ar^F = 3,5-(CF₃)₂C₆H₃, PONOP = κ³-NC₅H₃-2,6-(OP^tBu₂)₂] a σ-amine-borane pincer complex can be prepared. 1 is also a competent catalyst for the dehydrocoupling of H₃B·NMe₂H, and detailed mechanistic studies probe the potential roles of amine (NMe₂H), boronium ([H₂B(NMe₂H)₂]⁺), ammonium ([NMe₂H₂]⁺) and a neutral Rh-hydride species in catalytic turnover. The importance of boronium and neutral hydride intermediates in the dehydrocoupling of H₃B·NMe₂H was first reported by Conejero and co-workers in cationic Pt-based systems.²⁸ Hydride transfer from amine-boranes to cationic metal centres has been reported, for example, by Peruzzini⁴⁰ and Jagirdar.⁴¹

Results and discussion

Synthesis, structures and reactivity of [Rh(PONOP)(H₃B·NMe₂R)][BAr^F₄] R = Me, 2; H, 3

Dihydrogen σ-complexes offer convenient entry-points into amine-borane coordination chemistry, as the H₂ ligand is readily displaced.^42–45 The dihydrogen complex [Rh(PONOP)(η²-H₂)][BAr^F₄], 1, is prepared as a microcrystalline powder³⁹ from addition of Na[BAr^F₄]⁴⁶ to Rh(PONOP)Cl⁴⁷ under an atmosphere of H₂ in CH₂Cl₂ solution followed by recrystallisation. Addition of one equivalent of H₃B·NMe₃ to 1 in 1,2-F₂C₆H₄ solution resulted in displacement of H₂ and the formation in quantitative yield by NMR spectroscopy of [Rh(PONOP)(η¹-H₃B·NMe₃)][BAr^F₄], 2 (Fig. 1). Recrystallisation of this solution from pentane afforded dark yellow crystals suitable for single crystal X-ray diffraction. The tertiary amine-borane H₃B·NMe₃ was used in these initial studies to stop onward dehydrocoupling by BH/NH activation.

Fig. 1 Synthesis and solid-state structure of complex 2. Displacement ellipsoids shown at the 30% level. [BAr^F₄]⁻ anion and most H-atoms are not shown for clarity. Selected bond distances (Å) and angles (°): Rh1–B1, 2.567(5); Rh1–P1, 2.2832(9); Rh1–P2, 2.2594(9); Rh1–N2, 2.026(3); B1–N1, 1.631(6); Rh1–H1, 1.72(5); P1–Rh1–P2, 161.63(3); N2–Rh1–H1, 172.6(17); Rh1–H1–B1, 121(4).

The solid-state structure of the cationic portion of 2 is shown in Fig. 1. The hydrogen atoms associated with the boron were located and refined. The structure demonstrates an η¹-bound H₃B·NMe₃ ligand, rather than bidentate η²:η², as determined by a long Rh⋯B distance [2.567(5) Å] and a rather open Rh–H–B angle [121(4)°].^48,49 There is one closer Rh–H distance [H1, 1.72(5) Å], with the other two considerably distant [2.8–2.9 Å]. In comparison, a pincer complex with a η²:η² H₃B·NMe₃ bonding mode, [Rh(NNN)(H₃B·NMe₃)][BAr^F₄] [NNN = 2,6-bis-[1-(2,6-diisopropylphenylimino)ethyl]pyridine] C, shows a closer Rh⋯B distance [2.305(5) Å] and more acute Rh–H–B angles [93(3)°], Scheme 4.³⁸ Complex 2 is more closely related to [Ir(POCOP)(H)₂(η¹-HSiEt₃)][B(C₆F₅)₄], D,⁵⁰ which has a far more pronounced η¹ binding mode [Ir⋯Si 3.346(1) Å, Ir–H–Si 157°], as well as [Pt(I^tBu’)(I^tBu)(η¹-H₃B·pyridine)][BAr^F₄], E [Pt⋯·B 2.8436(5) Å, Pt–H–B 147(4)°]⁵¹ (I^tBu = 1,3-di-tert-butylimidazolylidene).

Scheme 4 Comparison of η¹ and η²:η² amine-borane and silane complexes.

Solution NMR spectroscopic data are in agreement with the η¹-binding of the amine-borane being retained in CD₂Cl₂. In the 298 K ¹H NMR spectrum a broad signal integrating to 3 H is observed at δ −2.55 that sharpens on decoupling ¹¹B, that does not change on cooling to 173 K and no coupling to ¹⁰³Rh is observed. This is indicative of a rapid exchange between terminal and bridging B–H positions.^52,53 A single ^tBu environment was also observed. In the ¹¹B NMR spectrum a broad signal at δ −15.2 that is assigned to the H₃B·NMe₃ group is shifted 7.9 ppm upfield from free H₃B·NMe₃ (δ −7.3). The ³¹P{¹H} NMR spectrum displays a single doublet at δ 208.9 [J(RhP) = 142 Hz]. Electrospray-Ionisation Mass Spectrometry (ESI-MS) shows the correct isotope pattern for the cation (m/z = 575.26, calc. 575.26). Complex 2 is stable in 1,2-F₂C₆H₄ solution for at least 24 hours.

H₂ is a competitive ligand with H₃B·NMe₃ for coordination at the {Rh(PONOP)}⁺ fragment, and placing a sample of 2 (8.7 mM, 1,2-F₂C₆H₄) under 4 atm H₂ in a sealed NMR tube immediately results in a 4 : 1 ratio of 1 : 2 and free H₃B·NMe₃ – which is now observed as a sharp quartet in the ¹¹B NMR spectrum. Degassing returns 2 as the major component showing that these two species are in equilibrium (Fig. 1). These experimental results are consistent with a computed ΔG from DFT calculations (see later) that is close to thermoneutral, being −1.2 kcal mol⁻¹ in favour of 2. Under a D₂ atmosphere (1 atm) H/D exchange at the borane also occurs, to afford the H₂-, HD- and D₂-isotopologues of 1, as identified by their distinctive isotopically-shifted ³¹P chemical shifts (ESI), free H_xD_3−xB·NMe₃ – as shown by a loss of resolvable coupling in the ¹¹B NMR spectrum – and dissolved H₂/HD and D₂. This H/D exchange likely involves reversible oxidative cleavage of B–H or D₂ at the Rh(I)-centre followed by H/D exchange – via a σ-complex assisted metathesis (σ-CAM) mechanism.⁵⁴ Similar H/D exchange has been noted in related Rh-dihydrogen pincer complexes.^39,55

The reaction between 1 and one equivalent of H₃B·NMe₂H initially follows the same course as with H₃B·NMe₃ (Scheme 5). On time of mixing the complex [Rh(PONOP)(η¹-H₃B·NMe₂H)][BAr^F₄], 3, is observed to be the major organometallic product (greater than 95%), as identified by ¹H [δ −1.98, 3 H], ¹¹B [δ −21.3] and ³¹P{¹H} [δ 212.9, J(RhP) = 138 Hz] NMR spectroscopies in comparison with complex 2. The remainder is complex 1. A triplet observed in the ¹¹B NMR spectrum (∼10%) at δ −0.36 [J(BH) 110 Hz] is identified as H₂B(NMe₂)₂H,⁵⁶ and not boronium [H₂B(NMe₂H)₂][BAr^F₄] by comparison with an independently prepared sample of the latter in 1,2-F₂C₆H₄ (δ −2.15, J(BH) 117 Hz).²⁸ After 5 minutes aminoborane H₂B=NMe₂ [δ 37.8, t, J(BH) = 129 Hz]^43,57 is observed, indicating dehydrogenation is proceeding that then eventually forms dimeric [H₂BNMe₂]₂ [δ 5.7, J(BH) = 127 Hz]. After 35 minutes this solution has changed to return 1 as the sole organometallic complex, during which time the H₃B·NMe₂H has undergone dehydrocoupling to form [H₂BNMe₂]₂ as the major product. The diamino-borane HB(NMe₂)₂ is the other, minor (2%), product [δ 28.7, d, J(BH) 130 Hz]. This formally comes from a hetero-dehydrocoupling of NMe₂H and H₃B·NMe₂H.^51,58

Scheme 5 Dehydrogenation of H₃B·NMe₂H by complex 1.

Catalysis and catalyst speciation: induction periods, change in resting state and a neutral hydride

These stoichiometric reactions demonstrate σ-complex formation, B–H activation, and competitive H₂ binding, and thus set the scene for the catalytic studies on H₃B·NMe₂H dehydrocoupling. Initial studies using catalyst 1 in a sealed NMR tube [2 mol%, H₃B·NMe₂H 0.144 M, 298 K, 1,2-F₂C₆H₄ solution] revealed, by ¹¹B NMR spectroscopy (Fig. 2A), an induction period of approximately 300 seconds, followed by the consumption of H₃B·NMe₂H to finally give [H₂BNMe₂]₂ after 2700 s. H₂B=NMe₂ is observed to grow in and decay, with a temporal profile characteristic of an intermediate. Following the same reaction by interleaved ³¹P{¹H} NMR spectroscopy (Fig. 2B) revealed the immediate formation of the σ-amine-borane complex 3, which reforms complex 1 as H₃B·NMe₂H is consumed and H₂ builds up in the reaction head-space. A further complex is observed to grow in and then out again, which is characterised by a ³¹P{¹H} resonance at δ 225.4 [J(RhP) = 170 Hz] and a broad, high field, signal in the ¹H NMR spectrum at δ −10.13. This is identified by an independent synthesis as Rh(PONOP)H, 4. Notably, 4 appears at the early stages of reaction, post-induction period, and is then consumed as 1 grows in. A small amount (∼2%) of HB(NMe₂)₂ was also noted.

Fig. 2 Time course plots for dehydrogenation of H₃B·NMe₂H [0.144 M, 298 K, 1,2-F₂C₆H₄ solvent] using complex 1 (2 mol%) as measured by in situ NMR spectroscopy in a sealed NMR tube. (A) ¹¹B: △, H₃B·NMe₂H; ○, [H₂BNMe₂]₂; □, H₂B=NMe₂. (B) ³¹P{¹H}: ◊, 3; ○, 4; □, 1. Dotted lines to guide the eye.

Complex 4, Rh(PONOP)H, was independently synthesised by addition of the lithium amidoborane Li[NMe₂BH₃]⁵⁹ to Rh(PONOP)Cl in pentane solution, and is formed alongside [H₂BNMe₂]₂. Filtration and recrystallisation from methylcyclohexane afforded 4 as red crystals. Fig. 3 shows the solid-state structure of complex 4, which was refined successfully as a two-component twin. The structural solution shows a pseudo square-planar coordination geometry around Rh in which the hydride ligand was located. The associated structural metrics are unremarkable, and complex 4 adds to the relatively small number of structurally characterised planar pincer mono-hydrides.^60–67 In the ¹H NMR spectrum of the isolated product (C₆D₆), the hydride ligand is signalled by a doublet of triplets at δ −9.60 [J(RhH) = 19.5, J(PH) = 22.6 Hz] that collapses into a doublet in the ¹H{³¹P} NMR spectrum. In the ³¹P{¹H} NMR spectrum a doublet is observed at δ 226.2 [J(RhP) = 171 Hz]. While these data are very similar to those observed during catalysis in 1,2-F₂C₆H₄ solution, the hydride chemical shift is different for pure material compared to that observed in situ during catalysis in this solvent: δ −9.92 and δ −10.13 respectively. This suggests the possibility of secondary, [Rh]–H⋯H–X, interactions⁶⁸ that are discussed in the computational section.

Fig. 3 Independent synthesis, solid-state structure of complex 4, and possible intermediates. Displacement ellipsoids shown at the 30% level. Most H-atoms are not shown for clarity. Selected bond distances (Å) and angles (°): Rh1–H1, 1.54(5); Rh1–P1, 2.2179(17); Rh1–P2, 2.2020(17); Rh1–N1, 2.046(6); P1–Rh1–P2, 162.80(7). Free energies are DFT-calculated.

The mechanism to form 4 from Rh(PONOP)Cl could operate via an amidoborane^69–71 intermediate I (inset, Fig. 3) that undergoes β-elimination to generate 4 and H₂B=NMe₂ (experimentally observed as the dimer). Arguing against this is that β-elimination processes in group-2 amidoboranes have been found to be rather high in energy,⁷² while the ethyl analogues M(PONOP)(CH₂CH₃) (M = Rh and Ir) have been reported to be unusually stable with regard to β-elimination and formation of the corresponding hydride.^73,74 An alternative mechanism is that Li[NMe₂BH₃] acts as a simple hydride source,^75,76 avoiding I and directly eliminating LiCl and H₂B=NMe₂. DFT calculations suggest the latter scenario is more likely, with the σ-amidoborane adduct, Rh(PONOP)(H₃BNMe₂), II, computed to be 9.3 kcal mol⁻¹ more stable than its N-bound isomer, Rh(PONOP)(NMe₂BH₃), I. Moreover, II exhibits a minimal barrier of 1.1 kcal mol⁻¹ to B–H bond cleavage to form 4 and free H₂BNMe₂ (see ESI‡).

The role of boronium, [NMe₂H₂][BAr^F₄], NMe₂H and the neutral hydride in catalytic turnover

With the identity of the intermediate complex 4 determined as being a neutral hydride, we considered possible routes for its formation and consumption in the catalytic ensemble, taking into account the induction period that is observed, that also precedes significant productive turnover. We have recently reported¹¹ that for cationic [Rh(DPEphos)]⁺-based dehydropolymerisation catalysts significant induction periods can be removed by adding amine, e.g. NMeH₂, as this promotes the formation of the active catalyst. While the actual mechanism of the active catalyst being brought on cycle with the [Rh(DPEphos)]⁺-based catalysts is complex, one role of free amine (formed from B–N bond cleavage) is proposed to be attack at a cationic σ-amine-borane complex to afford a neutral hydride and boronium cation, e.g. [H₂B(NMe₂H)₂]⁺.^28,31 The role of boronium in productive turnover has also been probed in pincer-like [Rh(Xantphos-ⁱPr)]⁺ systems, B,³⁰ and Pt-based systems related to E,²⁸Scheme 2, as well as others.²⁹ We propose a similar set of fundamental processes operates here, in which amine-induced B–H bond cleavage gives a metal hydride that is then reprotonated. Two realistic scenarios presented themselves for this process: (a) the boronium route as described, or, (b) one that invoked the formation of ammonium (Scheme 6). This latter route is a deprotonation of σ-bound H₃B·NMe₂H to form an amidoborane, i.e.IIFig. 3, that then eliminates H₂B=NMe₂. Experiments to probe these and other fundamental steps in catalysis are detailed below, and also explored in the computational section.

Scheme 6 Possible mechanisms for the formation of hydride 4.

(i) H₂ evolution experiments (0.072 M H₃B·NMe₂H, 2 mol% 1, eudiometer, 298 K) confirm an induction period (∼450 s) also operates in an open system, before significant hydrogen production starts (TON = 50, max. rate (v_max) = 1.0 × 10⁻⁴ M s⁻¹), Fig. 4, which is followed by an deceleration in rate consistent with substrate depletion. [H₂BNMe₂]₂ is produced as the principal product (¹¹B NMR), alongside a small amount of HB(NMe₂)₂. Periodic sampling demonstrates a similar profile for catalyst speciation as observed in a sealed NMR tube. There was no significant change in profile when Hg or sub-stoichiometric PPh₃ (0.2 equivalents) was added – suggesting a homogeneous process.⁷⁷ The post induction period reaction profile could not be reconciled with a simple kinetic model.

Fig. 4 Time course plot for dehydrogenation of H₃B·NMe₂H [0.072 M, 298 K, 1,2-F₂C₆H₄ solvent] using complex 1 (2 mol%), plus 1 equivalent NMe₂H, [H₂B(NMe₂H)₂][BAr^F₄] or [NMe₂H₂][BAr^F₄]. H₂ as measured by H₂ eudiometer.

(ii) The induction period using catalyst 1 is removed by addition of one equivalent of NMe₂H at the start of catalysis, that also promotes a slightly faster turnover (v_max = 1.6 × 10⁻⁴ M s⁻¹), Fig. 4. Again, the temporal profile could not be reconciled with a simple kinetic model. Speciation experiments under these conditions (sealed NMR tube) indicate that 4 is now formed exclusively at the start of catalysis, with no 3 observed. The final resting state is 1. Amine thus promotes catalysis and moves the initial resting state to neutral 4.

(iii) Adding one equivalent of [H₂B(NMe₂H)₂][BAr^F₄] to catalyst 1 results in a much more pronounced induction profile (∼450 s), and once turnover starts catalysis is slightly faster (v_max = 1.8 × 10⁻⁴ M s⁻¹), and decelerates slower. However, as boronium is also a source of free NMe₂H on protonation of 4, we cannot discount that it is simply acts in this way to promote catalysis (vide infra).

(iv) Addition of one equivalent of [NMe₂H₂][BAr^F₄] to catalyst 1 increases the induction period to ∼1 hour, but once turnover starts it is comparable to NMe₂H and [H₂B(NMe₂H)₂]⁺ doped systems (v_max = 2.0 × 10⁻⁴ M s⁻¹).

(v) Complex 4 is a poor catalyst on its own (0.072 M H₃BNMe₂H, 2 mol% 1, eudiometer, 298 K), promoting slow turnover with only 25% conversion observed after ∼1 h (v_max = 0.5 × 10⁻⁴ M s⁻¹).

(vi) When complex 4 is doped with one equivalent of [NMe₂H₂][BAr^F₄] the induction period is removed and catalysis proceeds at a rate comparable to the NMe₂H doped system (v_max = 1.6 × 10⁻⁴ M s⁻¹). Doping with one equivalent of [H₂B(NMe₂H)₂][BAr^F₄] also removes the induction period, but turnover is slower (v_max = 0.5 × 10⁻⁴ M s⁻¹).

(vii) Under stoichiometric conditions addition of [H₂B(NMe₂H)₂][BAr^F₄] or [NMe₂H₂][BAr^F₄] to 4 recovers 1.⁷⁸

These observations show the important role of NMe₂H in catalysis, and that complex 1 can be regenerated by protonation of complex 4. However, the kinetic profiles are still complex, and in particular changes in induction periods on doping suggest off-cycle processes and more complex equilibria are operating, while the precise role of [NMe₂H₂][BAr^F₄] and/or [H₂B(NMe₂H)₂][BAr^F₄] remain unclear. These points are explored in more detail next.

Kinetic analysis of H₂ generation

Given non-trivial temporal profiles we have deployed a combination of maximum rates⁷⁹ and Burés’ graphical approach of Variable Time Normalisation Analysis (VTNA) to interrogate the kinetics of catalysis.^80,81 VTNA works by plotting reaction course (reactant or product) against t[cat]ⁿ for different [cat]_TOTAL. By adjusting the power value until the various plots visually overlay the order in [cat]_TOTAL can be determined independent of the complexity of the kinetic regime.⁸²Fig. 5 shows this approach to determine the order in [Rh]_TOTAL, that demonstrates an overall first order kinetic regime between 0.2 mol% and 4 mol% [Rh]_TOTAL – thus excluding any dimer/monomer equilibria in catalysis and also that [Rh]_TOTAL remains approximately constant throughout. The data was time-shifted (∼450 s) to remove the induction period that comes from the catalyst-independent B–N bond cleavage.

Fig. 5 (A) Time course plot for dehydrogenation of H₃B·NMe₂H [0.072 M, 298 K, 1,2-F₂C₆H₄ solvent] using complex 1 (2 mol%) as measured by H₂ evolution. Variable Time Normalization Analysis^80,81 for the order in [Rh]_TOT (0.2 mol% – 4 mol%, induction periods removed): (B) [Rh]¹, (C) [Rh]^0.5, (D) [Rh]².

Given the complex kinetic profile, maximum rates (v_max) were used to determine the effect of isotopic substitution on the rate, and using H₃B·NMe₂D and D₃B·NMe₂H a k_H/k_D of 1.9 and 1.1 for NH and BH respectively was measured. This suggests that NH activation is involved in the rate-determining transition state, while BH activation is not.

The order in H₃B·NMe₂H was probed by determining the maximum rate measured as [H₃B·NMe₂H] was varied, while keeping [Rh]_TOTAL fixed. Fig. 6 shows that this offers a profile that is initially positive order in amine-borane, but at higher concentrations of H₃B·NMe₂H the rate decelerates. This suggests that H₃B·NMe₂H is participating in an equilibrium that removes at least one of the reaction partners implicated in, or prior to, the turn-over limiting step. A scenario that explains these data is an off-cycle interaction between H₃B·NMe₂H and [H₂B(NMe₂H)₂]⁺ or [NMe₂H₂]⁺ that would reduce the available concentration of H₃B·NMe₂H and [H₂B(NMe₂H)₂]⁺ or [NMe₂H₂]⁺; and depending on the relative concentrations could act in an inhibitory manner. We discount a scenario where the product (i.e. [H₂BNMe₂]₂) modifies the kinetics, as doping catalysis using 1 (2 mol%) with 50 equivalents of [H₂BNMe₂]₂ leads to no change in the temporal profile (v_max = 1.0 × 10⁻⁴ M s⁻¹).

Fig. 6 Maximum rate of catalysis versus [H₃B·NMe₂H] using complex 1 (0.00144 M) as measured by H₂ evolution.

Interaction between amine-boranes and [H₂B(NMe₂H)₂][BAr^F₄] or [NMe₂H₂][BAr^F₄]: competing off-cycle equilibria

To probe the existence of off-cycle interactions, NMR titration experiments were carried out between [H₂B(NMe₂H)₂][BAr^F₄] and H₃B·NMe₃ or H₃B·NMe₂H. Monitoring the change in the chemical shift of the NH protons in [H₂B(NMe₂H)₂]⁺ as a function of amine-borane concentration generated titration isotherms. WinEQNMR2⁸³ analysis of the H₃B·NMe₃ titration data determined a 1 : 1 stoichiometric association constant (K_a = 9.3(1) M⁻¹), Fig. 7A. We suggest adducts such as III with non-classical dihydrogen bonds are formed.⁸⁵ Related bisphosphine boronium adducts have been reported which show P–H⋯X hydrogen bonds.⁸⁴ For H₃B·NMe₂H the situation is more complex. Although the titration data clearly demonstrate an interaction between the two species, 1 : 1, 2 : 1 or 1 : 2 binding models failed to provide satisfactory agreement with experimental data, implying higher stoichiometry and/or complex equilibria. Nevertheless, these experiments show that H₃B·NMe₂H and [H₂B(NMe₂H)₂][BAr^F₄] can form off-cycle adducts that attenuate the availability of both. Similar 1 : 1 equilibria are also operating with [NMe₂H₂][BAr^F₄]/H₃B·NMe₃, K_a = 28(2) M⁻¹, possibly via adducts such as IV, Fig. 7B. At high [H₃B·NMe₂H] the decrease in max rate suggests these interactions (with whichever partner) are significant enough to cause a relative reduction in rate. Similar adducts could also be involved in sequestering free amine-leading to the change in induction periods observed when [H₂B(NMe₂H)₂][BAr^F₄] or [NMe₂H₂][BAr^F₄] are doped into catalysis; and productive turnover only occurs once sufficient NMe₂H has been formed by B–N bond cleavage to overcome these off-cycle equilibria. Crystallographically characterized [R₂NH₂⋯NR₂H]⁺ complexes are known.⁸⁶

Fig. 7 (A) Titration binding curve of [H₂B(NMe₂H)₂][BAr^F₄] (298 K, 30 mM, 1,2-F₂C₆H₄) with H₃B·NMe₃. Fitted binding isotherm is indicated by line. Association Constant, K_a, 9.3(1) M⁻¹, calculated using WinEQNMR2⁸³ monitoring the chemical shift data for the NH protons in [H₂B(NMe₂H)₂][BAr^F₄]. (B) Titration binding curve of [NMe₂H₂][BAr^F₄] (298 K, 30 mM, 1,2-F₂C₆H₄) with H₃B·NMe₃. Association Constant, K_a, 28(2) M⁻¹ derived from monitoring the chemical shift data for the NMe protons in H₃B·NMe₃.

Elementary steps of the mechanism

Pulling these observations together a catalytic cycle can be proposed, Scheme 7. ①: σ-Dihydrogen complex 1 reversibly reacts with H₃B·NMe₂H to form 3, that is observed to be the major species at the very start of catalysis. ②: NMe₂H (formed from slow B–N bond cleavage of H₃B·NMe₂H) then rapidly reacts with 3 to form 4, and either [H₂B(NMe₂H)₂]⁺ or [NMe₂H₂]⁺/H₂B=NMe₂. ③: Protonation of 4 regenerates 1, free NMeH₂ and in the case of boronium, H₂B=NMe₂ that dimerises to give [H₂BNMe₂]₂. The concentration of 1/4/3 follows temporal profiles (Fig. 2) that suggest they are closely matched in relative stabilities and their concentrations thus depend on how [H₃B·NMe₂H] and [H₂] evolve throughout catalysis. Off-cycle equilibria between H₃B·NMe₂H and either [H₂B(NMe₂H)₂]⁺ or [NMe₂H₂]⁺ operate to modify the available concentration of species involved in turnover. The measured KIE indicates NH activation is involved in the turnover limiting transition state. This could come from step ② being turnover limiting in the ammonium pathway (N–H cleavage of H₃B·NMe₂H), or step ③ in both boronium or ammonium pathways (N–H cleavage to reprotonate 4). The lack of a significant KIE measured for BH argues that B–H bond cleavage is not significant in the turnover limiting transition state. As the experimental data do not allow for discrimination between an ammonium or a boronium pathway we turned to computational studies to determine the relative energies of each pathway.

Scheme 7 Suggested simplified catalytic cycle with boronium (red) ammonium (blue, italics) and pathways.

Computational studies

The details of the dehydrogenation and protonation steps (3 → 4 → 1) were also probed with DFT calculations. The model used incorporated the full PONOP ligand with geometries optimised with the BP86 functional. Free energies were then corrected for solvation, dispersion (BJD3) and basis set effects (def2TZVP, see ESI‡ for full details). 2-Hexanone (ε = 14.14) was employed as solvent as a model for 1,2-difluorobenzene (ε = 13.81) for which parameters are not currently available. Two pathways were considered for H₃B·NMe₂H dehydrogenation from 3 and free NMe₂H to form 1 (see Scheme 8): initial nucleophilic attack at B to form boronium (following the proposal of Conejero and co-workers²⁸) and N–H deprotonation of the bound H₃B·NMe₂H ligand, i.e. the ammonium route.⁸⁷ Both processes initially form 4 which is then reprotonated by either [H₂B(NMe₂H)₂]⁺ or [NMe₂H₂]⁺ to form 1. Alternative processes based on B–H and/or N–H activation from 3 without amine or boronium involvement were also considered and shown to be significantly higher in energy (see Fig. S1–S3, ESI‡).

Scheme 8 Computed free energy profiles (kcal mol⁻¹) for dehydrogenation of H₃B·NMe₂H from H-bonded adduct 3 (centre) via initial nucleophilic attack at B (left) or initial deprotonation at N (right). [Rh] = (PONOP)Rh. Energies are quoted relative to 3 + free H₃B·NMe₂H and NMe₂H set to 0.0 kcal mol⁻¹. Ball and stick representations of TS(3-4)_B and TS(3-4)_N have the tBu Me groups removed for clarity and all distances are given in Å. Atom colouring scheme: Rh (teal); C (grey); H (while); O (red); N (blue); P (orange); B (yellow). ^a TS(4-1)_N is a true stationary point on the electronic energy surface but differential zero-point energy effects cause the free energy of this species to fall below 1·NMe₂H.

Starting from 3 (Scheme 8 centre), a H-bonded adduct, 3·NMe₂H (+2.8 kcal mol⁻¹), is formed that features a short NH⋯N distance of 1.76 Å. Nucleophilic attack at B then passes through TS(3-4)_B at +24.9 kcal mol⁻¹ to form 4·HMe₂NBH₂NMe₂H⁺ with a strong Rh–H^δ−⋯H^δ+–N dihydrogen interaction (1.75 Å). This is thus set up for facile proton transfer viaTS(4-1)_B (+10.1 kcal mol⁻¹) to give [(PONOP)Rh(η²-H₂)]⁺, 1, initially as an H-bonded adduct with the Me₂NBH₂NMe₂H moiety. Release of H₂BNMe₂ and NMe₂H along with substitution of H₂ in 1 by H₃B·NMe₂H (see Scheme 9) then reforms 3 and completes the catalytic cycle. The overall process is exergonic by 5.6 kcal mol⁻¹ and has an energy span of 24.9 kcal mol⁻¹. The alternative N–H deprotonation of the H₃B·NMe₂H ligand in 3·NMe₂H proceeds through TS(3-4)_N at +17.9 kcal mol⁻¹ and occurs with concomitant B–H bond cleavage to give neutral hydride 4, as a weakly bound adduct with H₂BNMe₂ and [H₂NMe₂]⁺. Removal of H₂BNMe₂ allows formation of 4·H₂NMe₂⁺ (−1.7 kcal mol⁻¹). This again features a strong Rh–H^δ−⋯H^δ+–N interaction (1.55 Å) and allows proton transfer to give 1·NMe₂H at +6.8 kcal mol⁻¹. Loss of NMe₂H and H₂/H₃B·NMe₂H substitution again completes the cycle. The overall barrier for this N-deprotonation mechanism is 17.9 kcal mol⁻¹ and is therefore predicted to be favoured over nucleophilic attack at B.

Scheme 9 Computed free energy profiles (kcal mol⁻¹) for H₂/H₃B·NMe₂H substitution in 1 to reform 3; [Rh]⁺ = [Rh(PONOP)]⁺. Energies are quoted relative to 3 + free H₃B·NMe₂H and NMe₂H set to 0.0 kcal mol⁻¹.

Scheme 9 shows the details of the H₂/H₃B·NMe₂H substitution to reform 3 from 1. An associative transition state could not be located, but instead initial, very facile oxidative cleavage of the η²-H₂ ligand permits formation of an η¹-H₃B·NMe₂H Rh(III) adduct, Int(1-3)2, at +5.6 kcal mol⁻¹. Reductive elimination of H₂ then proceeds viaTS(1-3)3 at +15.3 kcal mol⁻¹ to form 3.⁸⁸

Returning to Scheme 8, the computed geometries of the key transition states show TS(3-4)_B is similar to that located by Conejero and co-workers in their study,²⁸ and features significant B–H stretching corresponding to hydride transfer onto the Rh centre. In contrast TS(3-4)_N exhibits a much later geometry in which hydride transfer to Rh is complete and significant H⁺ transfer to form the [H₂NMe₂]⁺ cation is evident. As a result TS(3-4)_N displays a very large dipole moment (24.7 D) making its energy sensitive to solvation effects.⁸⁹TS(3-4)_N is favoured on energetic grounds and would be expected to show a small N–H/N–D KIE (and no B–H/B–D KIE). In contrast the significant B–H bond stretching in TS(3-4)_B would imply a significant B–H/B–D KIE that is not seen experimentally.

The computed energies of 1, 3 and 4 in Scheme 8 bear comparison with the time course plots derived from experiment in Fig. 2. 1 and 3 are computed to be most stable as isolated species, however 4 is present as the dihydrogen-bonded adduct 4·H₂NMe₂⁺. This formulation is also consistent with the small shift in ¹H chemical shift associated with this species under catalytic conditions. Overall these three species are all within 1.8 kcal mol⁻¹ of each other. The barriers for the formation of 4 (as 4·H₂NMe₂⁺, ΔG^‡ = 17.9 kcal mol⁻¹viaTS(3-4)_N) and its onward reaction to reform 3 (ΔG^‡ = 17.0 kcal mol⁻¹, i.e. from 4·H₂NMe₂⁺ at −1.7 kcal mol⁻¹ to TS(1-3)3 at +15.3 kcal mol⁻¹) are also finely balanced. The computed energetics therefore reflect the observation of 1, 3 and 4 during catalysis.

Finally, H₃B·NMe₂H dehydrogenation at the neutral hydride 4 was also assessed and a novel concerted transition state, TS(4-4), involving B–H transfer onto the Rh–H ligand, with concomitant N–H transfer onto the Rh centre was characterised (Fig. 8). This releases not only H₂BNMe₂ but also H₂ in a single step, regenerating 4 with a computed barrier of 26.2 kcal mol⁻¹. This barrier is significantly higher than that for the amine-assisted pathway in Scheme 8 and so is consistent with the poor performance of isolated 4 as a dehydrocoupling catalyst.⁹⁰ We have recently reported a similar amine-borane dehydrogenation transition state at an Fe–H species.⁹¹

Fig. 8 Computed geometry of TS(4-4) for dehydrogenation of H₃B·NMe₂H at 4, with selected distances in Å. Energies is quoted relative to 4 + free H₃B·NMe₂H set to 0.0 kcal mol⁻¹. Rh (teal); C (grey); H (while); O (red); N (blue); P (orange); B (yellow).

Conclusions

By using the [Rh(PONOP)(η²-H₂)][BAr^F₄] precatalyst, which has a labile dihydrogen ligand, we have been able to map out the catalytic dehydrocoupling of H₃B·NMe₂H using a pincer complex. As is becoming increasingly apparent,^28,30,31 for cationic systems a hydride transfer/reprotonation route from a σ-bound amine-borane is a viable pathway for dehydrogenation when using cationic catalysts. While this can occur via a nucleophilic attack on B, via a boronium (i.e.Scheme 2), we show here that an alternative pathway of deprotonation of the σ-bound amine-borane to form an intermediate ammonium salt is also a viable route (Scheme 10). Central to both these processes is the generation of free amine from B–N bond cleavage to act as a nucleophile or base respectively. We,¹¹ and others,²⁴ have recently commented on the role of amine in promoting catalytic turnover in a variety of dehydrocoupling systems; and there is a very recent complementary report of the stoichiometric role of ammonium/amine in the protonation of borohydride complexes to eventually form σ-bound aminoboranes.⁹² In the system under discussion here the amine-assisted formation of the neutral hydride, i.e.4, and [NMe₂H₂]⁺ is rate limiting in catalysis, although this may not necessarily be the case for every system in a more general sense. Whatever the precise barriers of each step, the off-cycle equilibria involving [NMe₂H₂]⁺ (or [H₂B(NMe₂H)₂]⁺ in a boronium route) will likely have an additional influence on the overall kinetics. A more detailed understanding of the role of initial B–N bond cleavage, to form free amine and the resulting co-catalysts, in dehydrocoupling – and especially dehydropolymerisation – could well be important in building an overarching general mechanism for such processes.

Scheme 10 Hydride transfer and reprotonation via ammonium.

Experimental

All manipulations, unless otherwise stated, were performed under an argon atmosphere using standard Schlenk line and glovebox techniques. Glassware was oven dried at 130 °C overnight and flame dried under vacuum prior to use. Pentane, hexane, toluene, Et₂O and CH₂Cl₂ were dried using a Grubbs-type solvent purification system (MBraun SPS-800) and degassed by three successive freeze–pump–thaw cycles.⁹³ THF was dried over Na/benzophenone, vacuum distilled, degassed by three successive freeze–pump–thaw cycles and stored over 3.0 Å molecular sieves. 1,2-F₂C₆H₄ (pre-treated with alumina) and CD₂Cl₂ were dried over CaH₂, vacuum distilled, degassed by three successive freeze–pump–thaw cycles and stored over 3.0 Å molecular sieves. H₃B·NMe₃ and H₃B·NMe₂H were purchased from Sigma-Aldrich and sublimed prior to use (5.0 × 10⁻² mbar, 298 K and 303 K respectively). Hg (99.9995%) was purchased from Sigma-Aldrich, washed with 1,2-F₂C₆H₄ and dried in vacuo prior to use. PPh₃ and n-butyllithium (2.5 M in hexanes) were purchased from Sigma-Aldrich and used as received. BH₃·THF (1.0 M in THF) and NMe₂H (2.0 M in THF) were purchased from Fisher Scientific and used as received to form solutions in 1,2-F₂C₆H₄ of the desired concentrations. Na[BAr^F₄] (Ar^F = 3,5-(CF₃)₂C₆H₃),⁴⁶ [BH₂(NMe₂H)₂][BAr^F₄],^28,94 D₃B·NMe₃,⁹⁵ H₃B·NMe₂D,⁹⁶ Rh(PONOP)C⁴⁷ and 1 ^39,97 were prepared by literature methods.

NMR spectra were recorded on a Bruker AVIIIHD 500 or Bruker AVIIIHD 400 nanobay spectrometer at room temperature, unless otherwise stated. Residual protio solvent was used as a reference for ¹H NMR spectra in deuterated solvent samples. For 1,2-F₂C₆H₄ solvent the NMR spectrometer was pre-locked to a sample of C₆D₆ (25%) and 1,2-F₂C₆H₄ (75%) and referenced to the centre of the downfield solvent multiplet, δ = 7.07. ³¹P and ¹¹B NMR spectra were referenced externally against 85% H₃PO₄ and BF₃·OEt₂ respectively. Chemical shifts (δ) are quoted in ppm and coupling constants (J) in Hz. Electrospray ionization mass spectrometry (ESI-MS) of organometallic complexes were recorded using a Bruker MicrOTOF instrument directly connected to a modified Innovative Technology glovebox.⁹⁸ Samples were diluted in 1,2-F₂C₆H₄ to a concentration of approximately 1.0 × 10⁻⁶ M before analysis. Elemental microanalyses were performed by Stephen Boyer at London Metropolitan University.

Synthesis of [Rh(PONOP)(H₃B·NMe₃)][BAr^F₄] (2)

1 (50 mg, 0.036 mmol) and H₃B·NMe₃ (2.62 mg, 0.036 mmol) were dissolved in 1,2-F₂C₆H₄ (2 mL) and the reaction stirred at room temperature for 24 h. The solution was then concentrated in vacuo to ca. 1 mL, cooled to 0 °C and pentane (5 mL) was added to give a precipitate. The solid was isolated by filtration and washed with pentane (2 mL × 2) before being dried under vacuum to give the product as a yellow powder. An isolated yield 32.8 mg (63%) was obtained. Layering a 1,2-F₂C₆H₄ solution of complex 2 with pentane and storing at 5 °C overnight yielded dark yellow crystals suitable for single crystal X-ray diffraction.

¹H NMR (500 MHz, CD₂Cl₂, 298 K). δ 7.73 (br s, 8H, [BAr^F₄]-o-CH), 7.69 (obscured tr, ³J_HH 8.0, 1H, C₅H₃N), 7.56 (br s, 4H, [BAr^F₄]-p-CH), 6.67 (d, ³J_HH 8.0, 2H, C₅H₃N), 2.72 (s, 9H, NMe₃), 1.42 (vt, J_PH 8.0, 36H, P(^tBu)₂), −2.55 (br d, 3H, RhH₃B).

¹¹B NMR (160 MHz, CD₂Cl₂, 298 K). δ −6.61 (s, [BAr^F₄]), −15.24 (br s, RhH₃B).

³¹P{¹H} NMR (202 MHz, CD₂Cl₂, 298 K). δ 208.9 (d, ¹J_RhP 141.6).

ESI-MS (1,2-F₂C₆H₄, 60 °C, 4.5 kV). m/z 575.26 (calc. 575.26 for [Rh(PONOP)(H₃B·NMe₃)]⁺ fragment).

Elemental microanalysis. Calc. C₅₆H₆₃B₂F₂₄N₂O₂P₂Rh (1438.57 gmol⁻¹) C, 46.76; H, 4.41; N, 1.95. Found: C, 46.90; H, 4.27; N, 2.01.

Spectroscopic data for [Rh(PONOP)(H₃B·NMe₂H)][BAr^F₄] (3)

¹H NMR (500 MHz, 1,2-F₂C₆H₄, 298 K). δ 8.32 (br s, 8H, [BAr^F₄]-o-CH), 7.69 (br s, 4H, [BAr^F₄]-p-CH), 7.61 (tr, ³J_HH 8.0, 1H, C₅H₃N), 6.67 (d, ³J_HH 8.3, 2H, C₅H₃N), 3.25 (br s, 1H, NH), 2.76 (d, J 5.6, 6H, NMe₂), 1.42 (vt, J_PH 7.4, 36H, P(tBu)₂), −1.98 (br d, 3H, RhH₃B).

¹¹B NMR (160 MHz, 1,2-F₂C₆H₄, 298 K). δ −6.22 (s, [BAr^F₄]), −21.36 (br s, RhH₃B).

³¹P{¹H} NMR (202 MHz, 1,2-F₂C₆H₄, 298 K). δ 212.9 (d, ¹J_RhP 137.5).

Synthesis of Li[H₃B·NMe₂]⁵⁹

Hexane (20 mL) was added to H₃B·NMe₂H (400 mg, 6.79 mmol) to give a white suspension, which was then cooled to −78 °C. n-Butyllithium (2.5 M, 3 mL, 7.5 mmol) was added dropwise by syringe over 15 minutes. The reaction was stirred at −78 °C for 1 hour and then allowed to come to room temperature and stirred for an additional hour. The solution was then filtered, leaving a white solid which was washed with hexane. The solid was dried cold in vacuo for 2 hours to yield a white powder (412.9 mg, 94% yield).

¹¹B NMR (128 MHz, THF, 298 K). δ −14.45 (q, J_BH 86.9).

Synthesis of Rh(PONOP)H (4)

Rh(PONOP)Cl (150 mg, 0.279 mmol) and Li[H₃B·NMe₂] (36.2 mg, 0.558 mmol) were dissolved in pentane (5 mL) to give an orange/red solution, which was stirred at room temperature for 24 hours. The solution was then filtered and the solvent removed in vacuo and dried under vacuum overnight to yield a red powder. An isolated yield of 113.3 mg (81% yield) was obtained. A concentrated solution of 4 in methylcyclohexane was stored at 5 °C for 1 hour and then moved to −20 °C and after 3 days, red crystals suitable for single crystal X-ray diffraction were formed.

¹H NMR (400 MHz, C₆D₆, 298 K). δ 6.92 (tr, ³J_HH 8.0, 1H, C₅H₃N), 6.27 (d, ³J_HH 7.9, 2H, C₅H₃N), 1.41 (vt, J_PH 6.9, 36H, P(tBu)₂), −9.60 (td, J_PH 22.6, J_RhH 19.5, 1H, RhH).

¹H{³¹P} NMR (400 MHz, C₆D₆, 298 K). δ 6.92 (tr, ³J_HH 8.0, 1H, C₅H₃N), 6.28 (d, ³J_HH 7.9, 2H, C₅H₃N), 1.42 (s, 36H, P(tBu)₂), −9.60 (d, J_RhH 19.6, 1H, RhH).

³¹P{¹H} NMR (162 MHz, C₆D₆, 298 K). δ 226.2 (d, ¹J_RhP 171.2).

¹³C{¹H} NMR (100 MHz, C₆D₆, 298 K). δ 28.7 (vt, J_CP 5.3, 12C, P–C(CH₃)₃), 39.0 (vq, J_CPRh 3.9, 4C, P–C(CH₃)₃), 100.8 (d, J_CP 2.3, 2C, C₅H₃N), 135.9 (s, 1C, C₆H₅N), 163.4 (vt, J_CP 4.9, 2C, C₆H₅N).

Elemental microanalysis. Calc. C₂₁H₄₀NO₂P₂Rh (503.41 gmol⁻¹) C, 50.10; H, 8.01; N, 2.78. Found: C, 49.99; H, 8.09; N, 2.86.

Synthesis of D₃B·NMe₂H

D₃B·NMe₂H was prepared by a modification of the literature procedure for the synthesis of D₃B·NMeH₂.⁹⁵ D₃B·NMe₃ (200 mg, 2.6 mmol, 99% D) was added to a J. Young flask and cooled to −78 °C. NMe₂H (2.0 M in THF, 7 mL, 132.0 mmol) was added to the D₃B·NMe₃, the solution allowed to warm to room temperature and was stirred for 72 hours. The volatiles were removed in vacuo at 0 °C and the resulting solid purified by sublimation (5.0 × 10⁻² mbar, 303 K) to give a white solid (130 mg, 80% yield). NMR data as reported previously in the literature.⁹⁶

General procedure for kinetic measurements of the catalytic dehydrogenation of H₃B·NMe₂H

Under closed/NMR tube conditions. In a typical experiment (e.g. 2 mol% catalyst loading), H₃B·NMe₂H (3.4 mg, 57.7 μmol) was dissolved in 1,2-F₂C₆H₄ (0.4 mL) in a Schlenk flask and 0.2 mL of this solution was added to a J. Young's high-pressure NMR tube. In a separate flask, 1 (2 mg, 1.4 μmol) was dissolved in 1,2-F₂C₆H₄ (0.5 mL) and 0.2 mL of this solution was added to the same NMR tube, which was immediately sealed and then frozen in N_2(l). When the NMR spectrometer was set up for kinetic measurements, the NMR tube was thawed, shaken thoroughly and immediately inserted into the NMR spectrometer for analysis.

Under hydrogen evolution measurement conditions. In typical experiment (e.g. 2 mol% catalyst loading), H₃B·NMe₂H (21.2 mg, 0.36 mmol) was dissolved in 1,2-F₂C₆H₄ (4 mL) in a jacketed two-neck Schlenk flask connected to a recirculating cooler and the temperature set to 25 °C. In a separate Schlenk flask, 1 (11.8 mg, 8.6 μmol) was dissolved in 1,2-F₂C₆H₄ (1.2 mL). The flask containing amine-borane was sealed off from the argon supply and connected to a water-filled 100.0 mL gas burette. 1 mL of the catalyst solution was added to the reaction mixture and the resultant solution was stirred at 400 rpm. The volume of gas evolved was then recorded as function of time, starting from the addition of catalyst to amine-borane.

Conflicts of interest

There are no conflicts of interest.

Acknowledgements

University of Oxford, Heriot-Watt University, the EPSRC (DTP and EP/M024210/1) and SCG Chemicals Co. Ltd for funding. The ToC for this article was prepared using Catacycle.⁹⁹ David Ryan for stimulating and insightful discussions.

Notes and references

1. A. Staubitz, A. P. M. Robertson, M. E. Sloan and I. Manners, Chem. Rev., 2010, 110, 4023–4078.
2. R. Waterman, Chem. Soc. Rev., 2013, 42, 5629–5641.
3. H. C. Johnson, T. N. Hooper and A. S. Weller, Top. Organomet. Chem., 2015, 49, 153–220.
4. A. Rossin and M. Peruzzini, Chem. Rev., 2016, 116, 8848–8872.
5. S. Bhunya, T. Malakar, G. Ganguly and A. Paul, ACS Catal., 2016, 6, 7907–7934.
6. E. M. Leitao, T. Jurca and I. Manners, Nat. Chem., 2013, 5, 817–829.
7. A. Staubitz, A. Presa Soto and I. Manners, Angew. Chem., Int. Ed., 2008, 47, 6212–6215.
8. A. Staubitz, M. E. Sloan, A. P. M. Robertson, A. Friedrich, S. Schneider, P. J. Gates, J. Schmedt auf der Günne and I. Manners, J. Am. Chem. Soc., 2010, 132, 13332–13345.
9. D. Han, F. Anke, M. Trose and T. Beweries, Coord. Chem. Rev., 2019, 380, 260–286.
10. Smart Inorganic Polymers: Synthesis, Properties, and Emerging Applications in Materials and Life Sciences, ed. E. Hey-Hawkins and M. Hissler, Wiley-VCH, Weinhiem, 2019.
11. G. M. Adams, D. E. Ryan, N. A. Beattie, A. I. McKay, G. C. Lloyd-Jones and A. S. Weller, ACS Catal., 2019, 9, 3657–3666.
12. V. Du, G. Whittell and I. Manners, Dalton Trans., 2016, 45, 1055–1062.
13. D. A. Resendiz-Lara, N. E. Stubbs, M. I. Arz, N. E. Pridmore, H. A. Sparkes and I. Manners, Chem. Commun., 2017, 53, 11701–11704.
14. X. Wang, T. N. Hooper, A. Kumar, I. K. Priest, Y. Sheng, T. O. M. Samuels, S. Wang, A. W. Robertson, M. Pacios, H. Bhaskaran, A. S. Weller and J. H. Warner, CrystEngComm, 2017, 19, 285–294.
15. A. L. Colebatch and A. S. Weller, Chem. – Eur. J., 2019, 25, 1379–1390.
16. B. L. Dietrich, K. I. Goldberg, D. M. Heinekey, T. Autrey and J. C. Linehan, Inorg. Chem., 2008, 47, 8583–8585.
17. A. Kumar, H. C. Johnson, T. N. Hooper, A. S. Weller, A. G. Algarra and S. A. Macgregor, Chem. Sci., 2014, 5, 2546–2553.
18. M. Shimoi, S.-i. Nagai, M. Ichikawa, Y. Kawano, K. Katoh, M. Uruichi and H. Ogino, J. Am. Chem. Soc., 1999, 121, 11704–11712.
19. M. A. Esteruelas, A. M. López and M. Oliván, Chem. Rev., 2016, 116, 8770–8847.
20. A. Paul and C. B. Musgrave, Angew. Chem., Int. Ed., 2007, 46, 8153–8156.
21. E. M. Titova, E. S. Osipova, A. A. Pavlov, O. A. Filippov, S. V. Safronov, E. S. Shubina and N. V. Belkova, ACS Catal., 2017, 7, 2325–2333.
22. M. A. Esteruelas, P. Nolis, M. Oliván, E. Oñate, A. Vallribera and A. Vélez, Inorg. Chem., 2016, 55, 7176–7181.
23. A. N. Marziale, A. Friedrich, I. Klopsch, M. Drees, V. R. Celinski, J. Schmedt auf der Günne and S. Schneider, J. Am. Chem. Soc., 2013, 135, 13342–13355.
24. A. Glüer, M. Förster, V. R. Celinski, J. Schmedt auf der Günne, M. C. Holthausen and S. Schneider, ACS Catal., 2015, 5, 7214–7217.
25. X. Zhang, L. Kam, R. Trerise and T. J. Williams, Acc. Chem. Res., 2017, 50, 86–95.
26. P. Bhattacharya, J. A. Krause and H. Guan, J. Am. Chem. Soc., 2014, 136, 11153–11161.
27. W. E. Piers, S. C. Bourke and K. D. Conroy, Angew. Chem., Int. Ed., 2005, 44, 5016–5036.
28. M. Roselló-Merino, J. López-Serrano and S. Conejero, J. Am. Chem. Soc., 2013, 135, 10910–10913.
29. S. Pal, S. Kusumoto and K. Nozaki, Organometallics, 2018, 37, 906–914.
30. G. M. Adams, A. L. Colebatch, J. T. Skornia, A. I. McKay, H. C. Johnson, G. C. Lloyd-Jones, S. A. Macgregor, N. A. Beattie and A. S. Weller, J. Am. Chem. Soc., 2018, 140, 1481–1495.
31. A. Kumar, N. A. Beattie, S. D. Pike, S. A. Macgregor and A. S. Weller, Angew. Chem., Int. Ed., 2016, 55, 6651–6656.
32. A. St John, K. I. Goldberg and D. M. Heinekey, in Organometallic Pincer Chemistry, ed. G. van Koten and D. Milstein, Springer Berlin Heidelberg, Berlin, Heidelberg, 2013, pp. 271–287.
33. T.-P. Lin and J. C. Peters, J. Am. Chem. Soc., 2013, 135, 15310–15313.
34. A. Rossin, G. Bottari, A. M. Lozano-Vila, M. Paneque, M. Peruzzini, A. Rossi and F. Zanobini, Dalton Trans., 2013, 42, 3533–3541.
35. D. Han, M. Joksch, M. Klahn, A. Spannenberg, H. J. Drexler, W. Baumann, H. Jiao, R. Knitsch, M. R. Hansen, H. Eckert and T. Beweries, Dalton Trans., 2016, 45, 17697–17704.
36. T. J. Hebden, M. C. Denney, V. Pons, P. M. B. Piccoli, T. F. Koetzle, A. J. Schultz, W. Kaminsky, K. I. Goldberg and D. M. Heinekey, J. Am. Chem. Soc., 2008, 130, 10812–10820.
37. H. C. Johnson, C. L. McMullin, S. D. Pike, S. A. Macgregor and A. S. Weller, Angew. Chem., Int. Ed., 2013, 52, 9776–9780.
38. A. Johnson, A. J. Martínez-Martínez, S. A. Macgregor and A. S. Weller, Dalton Trans., 2019, 48, 9776–9781.
39. M. Findlater, K. M. Schultz, W. H. Bernskoetter, A. Cartwright-Sykes, D. M. Heinekey and M. Brookhart, Inorg. Chem., 2012, 51, 4672–4678.
40. A. Rossin, M. Caporali, L. Gonsalvi, A. Guerri, A. Lledós, M. Peruzzini and F. Zanobini, Eur. J. Inorg. Chem., 2009, 2009, 3055–3059.
41. R. Kumar and B. R. Jagirdar, Inorg. Chem., 2013, 52, 28–36.
42. G. Alcaraz, A. B. Chaplin, C. J. Stevens, E. Clot, L. Vendier, A. S. Weller and S. Sabo-Etienne, Organometallics, 2010, 29, 5591–5595.
43. C. J. Stevens, R. Dallanegra, A. B. Chaplin, A. S. Weller, S. A. Macgregor, B. Ward, D. McKay, G. Alcaraz and S. Sabo-Etienne, Chem. – Eur. J., 2011, 17, 3011–3020.
44. Y. Gloaguen, G. Bénac-Lestrille, L. Vendier, U. Helmstedt, E. Clot, G. Alcaraz and S. Sabo-Etienne, Organometallics, 2013, 32, 4868–4877.
45. T. M. Douglas, A. B. Chaplin, A. S. Weller, X. Yang and M. B. Hall, J. Am. Chem. Soc., 2009, 131, 15440–15456.
46. A. J. Martínez-Martínez and A. S. Weller, Dalton Trans., 2019, 48, 3551–3554.
47. W. H. Bernskoetter, C. K. Schauer, K. I. Goldberg and M. Brookhart, Science, 2009, 326, 553.
48. M. L. H. Green and G. Parkin, in The Chemical Bond III: 100 years old and getting stronger, ed. D. M. P. Mingos, Springer International Publishing, Cham, 2017, pp. 79–139.
49. A. J. Martínez-Martínez, B. E. Tegner, A. I. McKay, A. J. Bukvic, N. H. Rees, G. J. Tizzard, S. J. Coles, M. R. Warren, S. A. Macgregor and A. S. Weller, J. Am. Chem. Soc., 2018, 140, 14958–14970.
50. J. Yang, P. S. White, C. K. Schauer and M. Brookhart, Angew. Chem., Int. Ed., 2008, 47, 4141–4143.
51. M. Roselló-Merino, R. J. Rama, J. Díez and S. Conejero, Chem. Commun., 2016, 52, 8389–8392.
52. Y. Kawano, T. Kakizawa, K. Yamaguchi and M. Shimoi, Chem. Lett., 2006, 35, 568–569.
53. A. G. Algarra, L. J. Sewell, H. C. Johnson, S. A. Macgregor and A. S. Weller, Dalton Trans., 2014, 43, 11118–11128.
54. R. N. Perutz and S. Sabo-Etienne, Angew. Chem., Int. Ed., 2007, 46, 2578–2592.
55. A. B. Chaplin and A. S. Weller, Organometallics, 2011, 30, 4466–4469.
56. L. K. Krannich, C. L. Watkins, D. K. Srivastava and R. K. Kanjolia, Coord. Chem. Rev., 1992, 112, 117–129.
57. H. Nöth and H. Vahrenkamp, Chem. Ber., 1967, 100, 3353–3362.
58. P. Bellham, M. S. Hill, G. Kociok-Köhn and D. J. Liptrot, Chem. Commun., 2013, 49, 1960–1962.
59. Adapted from: G. B. Fisher, J. C. Fuller, J. Harrison, S. G. Alvarez, E. R. Burkhardt, C. T. Goralski and B. Singaram, J. Org. Chem., 1994, 59, 6378–6385.
60. S. P. Semproni, C. Milsmann and P. J. Chirik, J. Am. Chem. Soc., 2014, 136, 9211–9224.
61. L. M. Guard, T. J. Hebden, D. E. Linn and D. M. Heinekey, Organometallics, 2017, 36, 3104–3109.
62. M. Feller, U. Gellrich, A. Anaby, Y. Diskin-Posner and D. Milstein, J. Am. Chem. Soc., 2016, 138, 6445–6454.
63. W. D. Bailey, W. Kaminsky, R. A. Kemp and K. I. Goldberg, Organometallics, 2014, 33, 2503–2509.
64. M. C. Haibach, D. Y. Wang, T. J. Emge, K. Krogh-Jespersen and A. S. Goldman, Chem. Sci., 2013, 4, 3683–3692.
65. M. A. Esteruelas, M. Oliván and A. Vélez, Inorg. Chem., 2013, 52, 5339–5349.
66. L. M. Martínez-Prieto, C. Melero, D. del Río, P. Palma, J. Cámpora and E. Álvarez, Organometallics, 2012, 31, 1425–1438.
67. R. Johansson and O. F. Wendt, Organometallics, 2007, 26, 2426–2430.
68. N. V. Belkova, L. M. Epstein, O. A. Filippov and E. S. Shubina, Chem. Rev., 2016, 116, 8545–8587.
69. T. D. Forster, H. M. Tuononen, M. Parvez and R. Roesler, J. Am. Chem. Soc., 2009, 131, 6689–6691.
70. H. Helten, B. Dutta, J. R. Vance, M. E. Sloan, M. F. Haddow, S. Sproules, D. Collison, G. R. Whittell, G. C. Lloyd-Jones and I. Manners, Angew. Chem., Int. Ed., 2013, 52, 437–440.
71. J. Spielmann, D. Piesik, B. Wittkamp, G. Jansen and S. Harder, Chem. Commun., 2009, 3455–3456.
72. P. Bellham, M. D. Anker, M. S. Hill, G. Kociok-Köhn and M. F. Mahon, Dalton Trans., 2016, 45, 13969–13978.
73. M. D. Walter, P. S. White, C. K. Schauer and M. Brookhart, J. Am. Chem. Soc., 2013, 135, 15933–15947.
74. The ³¹P{¹H} NMR chemical shift of Ir(PONOP)H has been reported in ESI in ref. 73, the synthesis of which comes from β-elimination from Ir(PONOP)Et. While Rh(PONOP)Et was also structurally characterised its stability towards β-elimination was not reported.
75. C. L. Bailey, A. Y. Joh, Z. Q. Hurley, C. L. Anderson and B. Singaram, J. Org. Chem., 2016, 81, 3619–3628.
76. G. Barozzino Consiglio, P. Queval, A. Harrison-Marchand, A. Mordini, J.-F. Lohier, O. Delacroix, A.-C. Gaumont, H. Gérard, J. Maddaluno and H. Oulyadi, J. Am. Chem. Soc., 2011, 133, 6472–6480.
77. C. A. Jaska and I. Manners, J. Am. Chem. Soc., 2004, 126, 9776–9785.
78. Under these conditions a small amount (∼20%) of a new species identified as [Rh(PONOP)(NMe₂H)][BAr^F₄] was also observed, δ 198.2 [J(RhP) = 151 Hz]. This does not react with H₃B·NMe₃ and we have not observed it in any other experiments – suggesting that it does not play a role in catalysis.
79. A. A. Mikhailine, M. I. Maishan, A. J. Lough and R. H. Morris, J. Am. Chem. Soc., 2012, 134, 12266–12280.
80. J. Burés, Angew. Chem., Int. Ed., 2016, 55, 2028–2031.
81. J. Burés, Angew. Chem., Int. Ed., 2016, 55, 16084–16087.
82. V. Vasilenko, C. K. Blasius and L. H. Gade, J. Am. Chem. Soc., 2018, 140, 9244–9254.
83. M. J. Hynes, J. Chem. Soc., Dalton Trans., 1993, 311–312.
84. T. A. Shuttleworth, M. A. Huertos, I. Pernik, R. D. Young and A. S. Weller, Dalton Trans., 2013, 42, 12917–12925.
85. X. Chen, J.-C. Zhao and S. G. Shore, Acc. Chem. Res., 2013, 46, 2666–2675.
86. M. A. Fox, A. E. Goeta, A. K. Hughes and A. L. Johnson, J. Chem. Soc., Dalton Trans., 2002, 2132–2141.
87. Mechanisms involving B–H and N–H activation at the Rh centre entailed significantly higher barriers – see ESI.‡.
88. The energy of 3-coordinate [Rh(PONOP)]⁺ (+17.5 kcal mol⁻¹) suggests a purely dissociative mechanism is less favourable.
89. In comparison the computed dipole for TS(3-4)_B is 9.1 D and for 3·NMe₂H 6.3 D. TS(3-4)_N remains more stable in the absence of the solvent correction, but only by 0.5 kcal mol⁻¹.
90. An alternative pathway involving stepwise B–H and N–H activation was also characterised but this has a higher barrier of 34.2 kcal mol⁻¹.
91. M. Espinal-Viguri, S. E. Neale, N. T. Coles, S. A. Macgregor and R. L. Webster, J. Am. Chem. Soc., 2019, 141, 572–582.
92. K. Kirchner, N. Gorgas, B. Stöger and L. F. Veiros, Angew. Chem., Int. Ed., 2019 DOI:10.1002/anie.201906971.
93. A. B. Pangborn, M. A. Giardello, R. H. Grubbs, R. K. Rosen and F. J. Timmers, Organometallics, 1996, 15, 1518–1520.
94. O. J. Metters, A. M. Chapman, A. P. M. Robertson, C. H. Woodall, P. J. Gates, D. F. Wass and I. Manners, Chem. Commun., 2014, 50, 12146–12149.
95. A. L. Colebatch, B. W. Hawkey Gilder, G. R. Whittell, N. L. Oldroyd, I. Manners and A. S. Weller, Chem. – Eur. J., 2018, 24, 5450–5455.
96. M. E. Sloan, A. Staubitz, T. J. Clark, C. A. Russell, G. C. Lloyd-Jones and I. Manners, J. Am. Chem. Soc., 2010, 132, 3831–3841.
97. G. M. Adams, F. M. Chadwick, S. D. Pike and A. S. Weller, Dalton Trans., 2015, 44, 6340–6342.
98. A. T. Lubben, J. S. McIndoe and A. S. Weller, Organometallics, 2008, 27, 3303–3306.
99. ToC graphic prepared using: J. McFarlane, B. Henderson, S. Donnecke and J. S. McIndoe, An Information-Rich Graphical Representation of Catalytic Cycles, ChemRxiv, 2019 DOI:10.26434/chemrxiv.8206637.v1 . Preprint.

---

Footnotes

† Dedication: In recognition of Professor Robin N. Perutz's outstanding scientific contributions to physical organometallic chemistry, and also the mentoring and friendship RNP has generously shared with SAM and ASW over their careers.

‡ Electronic supplementary information (ESI) available: Synthesis, characterisation data, structures of 2 and 4, computational details. CCDC 1917326 and 1917160. For ESI and crystallographic data in CIF or other electronic format see DOI: 10.1039/c9dt03358k

---