AmineBorane Dehydropolymerization Using Rh-Based Precatalysts: Resting State, Chain Control, and Efficient Polymer Synthesis

eprints@whiterose.ac.uk https://eprints.whiterose.ac.uk/ Reuse Items deposited in White Rose Research Online are protected by copyright, with all rights reserved unless indicated otherwise. They may be downloaded and/or printed for private study, or other acts as permitted by national copyright laws. The publisher or other rights holders may allow further reproduction and re-use of the full text version. This is indicated by the licence information on the White Rose Research Online record for the item.

The catalyzed dehydropolymerization of amine-boranes, archetypically H3B·NMeH2, is an atom-efficient methodology for the synthesis of polyaminoboranes (H2BNRH)n (Scheme 1A), forming H2 as the only by-product. [1][2][3][4] This new class of main-group polymer 5 is based upon BN main-chain units, and is isosteric with technologically-mature polyolefins. These main-chain B-N units suggest, in addition to unexplored material and chemical properties, potential applications as piezoelectric materials, 6,7 or as precursors to boron-based ceramics and h-BN. 1,8,9

Scheme 2: Exemplar Complex 1 and Prior Observations.
We have also reported on the key role of NMeH2, formed by B-N bond cleavage in H3B·NMeH2. 21 Despite these advances, the precise details of initiation, propagation and termination remain to be determined for these diverse catalyst systems, 3 while identification of resting states is rare 14,16 and challenging. 18 Herein we report on an investigation of the [Rh(DPEphos)] + pre-catalyst system, 1, in which a study of the kinetics, speciation and synthesis has allowed identification of the active catalyst, as well as allowing for the polymer-growth/termination processes to be interrogated. These insights are then harnessed in the design of a new, efficient, Rh-based catalyst that produces polyaminoborane on scale. A simple protocol is also described to significantly reduce the levels of residual catalyst in the isolated polymer.
We have previously reported that when 1 is employed as pre-catalyst, the monocationic hydrido-aminoborane dimer [Rh2(DPEphos)2(μ-H)(μ-H2B=NHMe)][BAr F 4] 2 is formed during the early stages of reaction. 23 We propose this arises via an amine-promoted B-H hydride transfer 26  respectively. Based on these observations, a simple kinetics model was constructed for the induction process, involving generation of 2 by rapid trapping of Rh(DPEphos)H with unreacted E, followed by a slow, amine-dependent, fragmentation to form the active catalyst. This telescopes the elementary steps of the induction process, 29 allows H2 evolution to be used as proxy for transient H2B=NMeH, and successfully reproduces the temporal concentration profiles, 23  With an effective model for the induction process determined, we then focussed on identification of the catalyst resting state. Based on our model, and the work of Fryzuk 32,33 F 4] is added. 10 Thus, when using a cationic pre-catalyst (i.e. 1 or 5), persistent NMeH2 will favor soluble neutral hydride via equilibration with complex 2 (k3, Scheme 3). When using neutral precatalyst 4, a high initial concentration of amine-borane, e.g.
[H3B·NMeH2]0 = 0.446 M in THF, inhibits the formation of a precipitate. Presumably, the amine-borane intercepts Rh(DPEphos)H before it oligomerizes. Thus, dimeric, neutral, hydride 3 is observed as the common resting state, irrespective of the pre-catalyst, or solvent. The half order dependence in [Rh]TOTAL points to a rapid endergonic equilibrium between dimer and monomer, prior to the turn-over limiting step. This has been noted in other Rh2Hx systems, 32,36,37 and the data are thus consistent with the resting state being dimeric 3. An important difference between neutral versus cationic pre-catalysts, is that the latter generate a boronium co-product, which has important implications for the dehydropolymerization, as discussed next. Neutral pre-catalyst 4 was deployed in the dehydropolymerization of H3B·NMeH2 at a variety of catalyst loadings, Table 1. Using 1,2-F2C6H4 as the solvent, kinetics measurements were hampered by formation of the insoluble precipitate. In THF, eudiometric measurements on H2 production were less reliable due to solvent volatility. Nevertheless,   Figure 1A. 38 A plot of conversion versus Mn ( Figure 1B, relative to polystyrene standards) 3,11,16 is characteristic of a non-living chain-growth polymerization: at low conversions the polymer is formed with high Mn and H3B·NMeH2 dominates. Variations in catalyst loading did not affect the degree of polymerization of the resulting polyaminoborane, in either 1,2-F2C6H4 (Figure 1C, Mn = 15000 g mol -1 ) or THF solutions (Mn = 17000 g mol -1 ), under 'open conditions' with a slow Ar-flow. This is different to cationic pre-catalysts, such as 1, where Mn scales inversely with [Rh]TOTAL: e.g. 6400 (1 mol%), 34900 g mol -1 (0.2 mol%). 23 However, 'closed conditions' that allow for build-up of H2 result in very low molecular weight oligomers being formed (1 mol % 4, less than 1000 gmol -1 by GPC, 11 B NMR spectroscopy 31 ). The cationic pre-catalyst 1 behaves analogously. 22 The neutral and cationic pre-catalyst systems differ by the presence of a boronium co-product with the latter, the relative concentration of which will scale with [Rh]TOTAL. 39 F 4] was doped (0.25 to 1 mol%) into 1 mol% 4 / H3B·NMeH2, to selectively form polyaminoborane ( 11 B NMR). Although GPC analysis of the resulting polymer using refractive index detection is affected by the co-eluting [BAr F 4] -masking the lower molecular weight region ( Figure 1D), 15   shown to be useful where there is overlap between distributions of polymer and transfer agents, such as that noted here, allowing for chain control processes to be probed. 40 A Mayo-type plot of [boronium]/[H3B·NMeH2] versus the ln-CLD slope indicates an inversely linear relationship ( Figure  1E and F), further supporting the conclusion that the boronium functions as a rapid chain control agent in the dehydropolymerization. Collectively, the analysis above facilitates the construction of a mechanistic landscape for dehydropolymerization, Scheme 5, that is consistent not only with the results herein, but also with our previous observations on cationic Rhbased systems. 15,22,23,28,41 Thus, dehydrogenation of amineborane to give the reactive monomer, H2B=NMeH, occurs at a neutral [Rh-H] species, in an H2-mediated equilibrium with dimer 3. Dehydrogenation to form H2B=NMeH, via BH/NH activation (Scheme 5A) could be facilitated by a hemilabile DPEphos ligand (e.g. k 2 and k 3 coordination 42 ) as previously suggested. 43 Initiated by a formal hydride-transfer from the rhodium hydride, 44 that is now playing a dual role in both dehydrogenation and initiation, 11,14 H2B=NMeH then undergoes rapid head-to-tail end-chain nucleophilic B-N bond formation, as proposed previously (Scheme 5B). 12,13,15,16 Chain-control by protonation of the terminal nucleophilic amine of the polymeryl group by boronium returns a cationic pre-catalyst, aminoborane and NMeH2, that are rapidly recycled (Scheme 5C). 25 A related intramolecular proton transfer has been proposed by Paul and co-workers for Ir(POCOP)H2 systems. 13 We speculate that, in the absence of boronium, chain transfer to pre-monomer H3B·NMeH2 controls chain-length, Scheme 5C. Whatever the precise mechanism for these chain-control processes, they result in relatively narrow dispersities of the final isolated polymer, as a result of the constant degrees of polymerization during the entire reaction ( Figure 1B). 45 H2 loss from 3, and related systems, 32,35 occurs readily on degassing. The position of the initial monomer/dimer equilibrium is thus expected to be sensitive to [H2], impacting on the rate of dehydrogenation as well initiator concentration. This, we suggest, is the origin of the low degrees of polymerization observed under 'closed conditions'. In support of this, for a system where hydride-bridged dimer formation is disfavoured due to sterics, e.g. Rh(Xantphosi Pr)H, H2 does not act to modify the degree of polymerization. 15 The precise gearing of all of these interconnected relationships is therefore pre-catalyst, co-product (e.g. boronium) and solvent specific. The use of new pre-catalysts based upon neutral 4 demonstrates wider applicability, and also signal the opportunity for the exploitation of structure/activity relationships (Figure 2, Table 1). For example, the Xantphos benzyl complex, 6, is an effective pre-catalyst for dehydropolymerization (1 mol%, 88000 g mol -1 , Ð 1.6), and the degree of polymerization can be controlled by [H2B(NMeH2)2][BAr F 4], e.g. 1 mol% Mn = 21000 g mol -1 . Complex 6 can be used at low loadings and high [H3B·NMeH2] (0.1 mol%, 5 M in THF, using commercially sourced amine-borane), to produce high 3 molecular weight polyaminoborane on gram scale (85000 g mol -1 , 1.1 g). Use of activated charcoal in the polymer work-up reduces the [Rh]-content from 195 ppm (no workup) to 81 ppm. This is considerably lower than reported for other Rh and Co dehydropolymerization systems. 15,21,46 The simple benzyl-dppp-catalyst 7 also promotes formation of high molecular weight polyaminoborane (98000 g mol -1 ). In summary, the identification of the catalyst resting state, the events that lead to its formation, and thus the role that co-products such as boronium and H2 likely play in chaincontrol, have provided important insights into the complex and nuanced set of interconnected processes that are required for selective amine-borane dehydropolymerization using Rh(bisphosphine)-based catalysts. While the detailed elucidation of the elementary steps awaits further study, Scheme 5 provides a testable framework for the analysis and design of catalyst systems for controlled amine-borane dehydropolymerization.

Supporting Information
Full experimental, structural, kinetics data and details of the simulated model. The Supporting Information is available free of charge on the ACS Publications website.

AUTHOR INFORMATION
Corresponding Author * andrew.weller@york.ac.uk

Author Contributions
The manuscript was written through contributions of all authors ACKNOWLEDGMENT Drs Romaeo Dallanegra