Comparing Neutral (Monometallic) and Anionic (Bimetallic) Aluminum Complexes in Hydroboration Catalysis: Influences of Lithium Cooperation and Ligand Set

Abstract Bimetallic lithium aluminates and neutral aluminum counterparts are compared as catalysts in hydroboration reactions with aldehydes, ketones, imines and alkynes. Possessing Li–Al cooperativity, ate catalysts are found to be generally superior. Catalytic activity is also influenced by the ligand set, alkyl and/or amido. Devoid of an Al−H bond, iBu2Al(TMP) operates as a masked hydride reducing benzophenone through a β‐Η transfer process. This catalyst library therefore provides an entry point into the future design of Al catalysts targeting substrate specific transformations.

The synthetic value of main-group metal complexes aside from the highly reactive and versatile organolithium and organomagnesium reagents have,f rom ah istorical perspective,b een overshadowed by the illustrious reputation of transition-metal (notably precious metals) and lanthanidemetal counterparts especially in catalysis. [1] To al arge extent main-group research has been driven by fundamental curiosity and the understanding of the nature of chemical bonding and structure.Astep-change occurred when it was realised that such main-group-metal species can act in homogeneous catalytic roles,previously the exclusive province of transitionmetal and lanthanide complexes.E mulating the high reactivity,s electivity and versatility of the often toxic and scarce precious metal complexes is atantalising challenge that needs addressing.I nt his regard, the pioneering work of Harder, Hill, Jones,Okuda, Power,Roesky,Wright among others,are expanding the vistas of main-group complexes in homoge-neous catalysis. [1,2] Since aluminum is the most abundant metal in the earthscrust, and also benefits from low toxicity, harnessing its reactivity is given high prominence in this main group uprising with longer term sustainability being ak ey issue.T hus,recently aluminum complexes have made significant strides forward in important stoichiometric and catalytic transformations. [3] Forexample,they are utilised in CÀCcross coupling chemistries,a nd in deprotonative metalation. [4] Catalytic hydroelementation reactions have also witnessed impressive progress in the past few years.R oesky and coworkers demonstrated that a b-diketiminato stabilised aluminium hydride complex is an excellent catalyst for hydroboration of alkynes and carbonyl groups. [5] More recently, Cowley,T homas and Bismuto revealed that DIBAL(H), and Et 3 Al·DABCO can catalyse hydroboration of alkynes. [6] Our groupsi nterests lie in exploiting the synergistic reactivity imparted by two distinct metal centres [7,8] installed within ab imetallic complex. In this regard we introduced ate complexes (Figure 1), detailing that heteroleptic lithium diamido-dihydridoaluminates and lithium monoamido-monohydrido-dialkylaluminates implicate that the alkali metal influences the ensuing "aluminum reactivity" in the hydroboration of aldehydes,k etones and terminal alkynes. [8] Further,t he catalytic chemistry of LiAlH 4 has recently been explored by Cowley,T homas and Bismuto in the challenging hydroboration of alkenes,h owever the role of the alkali metal was not elaborated. [9] Thus,the current state of the field dictates that asystematic analysis of the secondary metal cooperative effects and various ligand factors that contribute to efficient hydroboration, is required in order to establish empirical rules for ap osteriori design of future catalysts.
Hydroboration of unsaturated substrates under aluminum catalysis is gaining afoothold in the literature,and avariety of neutral aluminum complexes are displaying excellent potential in this role. [2a, 3,10] Previously,w er eported that bimetallic lithium [iBu 2 AlTMP(H)Li] 2 (1)a nd [(HMDS) 2 AlH(m-H)Li·3 THF)] (2)a re both efficient bimetallic (pre)catalysts in the hydroboration of aldehydes and ketones. [8] However, any synthetic advantages/disadvantageso fu sing ate complexes are yet to be fully uncovered, despite their potential. Thus,here,for the first time ate complexes are compared with their neutral aluminum counterparts to fully quantify their value in synthesis,a nd to glean understanding of their mode of action. Moreover,t he complexes chosen vary in their ligand constitution, that is,a lkyl versus amido constituents, providing further comparison. Mechanistically af requently postulated two-step reaction pathway is:1 )insertion of an unsaturated substrate into an AlÀHb ond;2 )s-bond metathesis with ab orane,r egenerating an active species and liberating product (Scheme 1).
Catalytic activities were screened with aldehydes,ketones, imines and alkynes,p roviding reaction scope to determine key divergences in catalyst reactivity.W epreviously reported 1 and 2 in catalytic hydroboration and these are compared with the neutral analogues 4 and 5,w hich differ by formal removal of LiH. [8] We prepared new complex 3,a na ll alkyl variant of 1 (established via NMR characterisation, including DOSY) by as imple co-complexation procedure (see Supporting Information). Compound 3 can be considered an ate version of DIBAL(H), 6.O ur results from comparative studies (reaction conditions are identical between different catalysts) are summarized in Table 1. Complexes 1-6 (5 mol %) were all tested in hydroboration reactions of benzophenone with pinacolborane (HBpin) at room temperature in J. Youngst ubes in C 6 D 6 .E ach bimetallic complex exhibits superior activity to its monometallic counterpart, affording quantitative conversion after 30 mins,apart from 4, which is 94 %complete after 30 mins.This is surprising since 4 does not possess an AlÀHb ond. Rationalising that an AlÀH bond must form in situ during the catalysis we performed as toichiometric reaction between 4 and benzophenone in hexane and C 6 D 6 ,w here clear, facile quantitative reaction occurs rapidly at room temperature (isobutene,the coproduct of b-hydride elimination, is seen in the 1 HNMR spectra). Xray diffraction studies of colourless crystals grown from the hexane solution revealed formation of [(TMP){Ph 2 -(H)CO}Al{m-OC(H)Ph 2 }] 2 (7)i na4 5% isolated yield (Scheme 2). It is germane to note that Et 3 Al·DABCO can catalyse hydroboration of alkynes due to ar edistribution reaction with HBpin generating the active Et 2 AlH species. [6] Thes tructure of 7 ( Figure 2, left) reveals ad imer wherein both iBu À groups of 4 have been replaced, by Ph 2 (H)CO À Scheme 1. Postulated insertion mechanism in Al-catalysed hydroboration.  ligands,f ormed by an apparent b-hydride process from the parent complex. b-Hydride elimination is known in alkylaluminum chemistry with carbonyls, [11] but to our knowledge this is the first example in hydroboration catalysis used to generate at ransient aluminum hydride.T hus 4 may be considered am asked hydride complex in hydroboration of ketones.E laborating this step further, it is pertinent to consider the Meerwein-Ponndorf-Verley (MPV) reduction, [11a,12, 13] employing aluminum alkoxides as the hydride source to reduce ketones.T wo competing mechanisms have been studied in silico. [13] Thefirst involves b-hydride transfer from the alkoxide ligand giving ah igh energy Al-H intermediate,w hich can then follow the pathway represented in Scheme 1. Thes econd pathway is much lower in energy and describes ac oncerted process containing a6 -membered transition state,f acilitating direct hydride transfer to the substrate.
Compound 7 (2.5 mol %) is shown to be catalytically active in ar eaction with benzophenone and HBpin, where quantitative hydroboration occurs after 3hours.S ince 4 seems ar eactivity outlier,s howing comparable reactivity to 1,t hey were both screened catalytically with one aldehyde and two further ketones.Ineach case the bimetallic complex 1 showed far superior activity.
Furthermore,acontrol reaction employing LiH as acatalyst (5 mol %) for hydroboration of benzophenone gave ay ield of only 10 %a fter 4h.T his illustrates that, in this regard, the neutral aluminum or lithium reagents in isolation deliver markedly reduced reactivities compared with the bimetallic formulations.I mportantly,f or the first time direct competition experiments reveal the synthetic superiority of lithium aluminate complexes in the context of hydroboration.
Hypothesising that any "ate effect" would be magnified with more challenging substrates we turned our attention to imines,w hich hitherto have not been catalytically hydroborated with Al complexes.That said, examples exist of main group complexes catalysing this transformation, and of Al complexes catalysing hydrosilylation or hydrogenation of imines, [2c, 10e,14] suggesting that imine hydroboration is aviable synthetic target. Catalytic hydroboration reactions of N-benzylidenemethylamine,using 1-6 showed lower reactivity at room temperature than with aldehydes and ketones,h owever the same reactivity pattern emerges,i nt hat the bimetallic complexes are superior to monometallic counterparts.A fter two hours, conversions are with 1 (42 %), 2 (35 %), 3 (53 %), 4 (3 %), 5 (22 %) and 6 (5 %). Nevertheless,t hese results with 1-3 constitute the first use of Al complexes in imine hydroboration. Stoichiometric reactions between 1, 3, 4 and 6 with the imine provide further insight. Compound 4 forms only ac oordination adduct with the imine in contrast to the bhydride elimination product with benzophenone,w hereas, 1, 3 and 6 add across the C=Nd ouble bond, with 6 displaying higher insertion reactivity (see Supporting Information). Notably Stephan and co-workers reported adimeric structure of an analogous reaction between 6 and ar elated imine. [14a] However,f aster substrate insertion does not translate into fast catalytic transformation. Thus we infer that the s-bond metathesis step with HBpin is greatly facilitated by the additional polarity imposed by the bimetallic ate constitution. Reinforcing this hypothesis,H arder and co-workers imine hydrogenation using catalytic LiAlH 4 illuminates the important role of the alkali metal, via DFT studies,wherein Al-H-Li interactions are retained throughout the proposed catalytic cycle. [14d] We next screened benzophenone imine in the catalysis with 1, 3, 4 and 6 (10 mol %), since this substrate has an acidic N-H atom amenable to deprotonation and therefore provides the possibility of reaction proceeding via an alternative deprotonation pathway.F urthermore,a mido groups in 1 and 4 can be directly compared with alkyl groups in 3 and 6. 1 and 3 achieve 73 %a nd 80 %c onversion after 2h or 30 minutes,respectively.Compounds 4 and 6 perform poorly, showing no catalytic activity at room temperature,prompting further consideration. Tw os toichiometric reactions between benzophenone imine and 1,a nd 4 were conducted, wherein both exhibit amido basicity.Inthe reaction with 4 [iBu 2 Al(m-N=CPh 2 )] 2 (8;F igure 2, right), was isolated as single crystals in a2 4% yield ( 1 HNMR yield of 86 %a gainst hexamethylcyclotrisiloxane as internal standard). In contrast to the benzophenone case where catalysis proceeds after a bhydride process step,the reactivity here ceases after an initial deprotonation by the TMP basicity.Interestingly,both 3 and 6 display trace amounts of H 2 evolution in their catalytic reactions as evidenced by alow intensity singlet resonance in the respective 1 HNMR spectra at d 4.47 ppm.
Thec atalytic results with benzophenone imine merit further comment. Both 1 and 4 exhibit deprotonation, suggesting that in ac atalytic regime,r eaction (using 1)m ay proceed in the pathway outlined in Scheme 3, that is, deprotonation followed by hydroboration then protonolysis to liberate product and generate acatalytically active species.
That 1 is active and 4 is not, may be assigned to the nature of deprotonation products,w hich clearly demonstrates the key role of bimetallic (Li-Al) cooperativity. I is the proposed deprotonation intermediate using 1 and I' ' using 4,w hich corresponds to the crystallographically authenticated 8.I nI the alkali metal would instil ad ifferent molecular charge distribution to that in I' '.T his scenario clearly facilitates the hydroboration step,w hich is not the case with I' '.Afinal comment on imine hydroboration is that in both cases 1 (73 % 0.5 h) offers marginally less reactivity than 3 (80 %0.5 h). This difference may describe as ubtle alkyl versus amido effect, whereby the replacement of one TMP anion for an iBu anion imparts greater nucleophilicity onto the hydride,p riming it for addition across the unsaturated substrate.A lternatively, the increased steric demand of TMP may slow reactivity. Moreover,i ti sa pparent that even when the deprotonation pathway is available (catalyst 1 with benzophenone imine), the pathway that follows,i nsertion (catalyst 3 with benzophenone imine) is favoured, albeit marginally.
Finally,w eturned to acetylene hydroboration comparing reactivity once more between 1, 3, 4 and 6.S toichiometric reactions of TMP-containing 1 and 4 with terminal alkyne phenylacetylene (PhCCH) in C 6 D 6 ,r eveal deprotonation of PhCCH at room temperature,inagreement with the fact that hydroboration of PhCCH with 1 implicated deprotonation as akey step. [8b] Alternatively 3 is unreactive with PhCCH, and 6 only very slowly hydroaluminates PhCCH, at room temperature.C atalysis,u sing 10 mol %l oadings in [D 8 ]toluene at 110 8 8C, in line with the reported reaction conditions using 6 (85 %conversion after 2hours), [6] reveal that 1 and 3 catalyse the transformation to the anti-Markovnikov vinylboronate ester in yields of 71 %and 83 %respectively.Conversely, 4 as expected, does not function as ac atalyst. Thus 3 is comparable to 6 however, for the first time we note that aclear ate effect is not in operation. Furthermore, 3 is ab etter catalyst than 1 underlying that increased hydride nucleophilicity is more important, mechanistically,than deprotonation, though reduced steric effects may also be afactor.
As imilar picture is seen with the internal alkyne diphenylacetylene. 6 (10 mol %) is reported to convert diphenylacetylene to the boronic ester in 40 %y ield after 2hours at 110 8 8Ci n[ D 8 ]toluene, [6] whereas 1 is completely inactive,a nd 3 only reaches conversions of approximately 10 %a fter 2hours,w hich is surprising given our preceding observations.O ne potential rationale for this marked reduc-tion in ate reactivity with diphenylacetylene may be attributed to asteric effect (Scheme 4).
Considering the required initial insertion step at the sp-C of diphenylacetylene,i nsertion into the Al À Hb ond of 3 (three iBu groups,one hydride) is likely to be slower than for 6 (two iBu groups,o ne hydride) due to the inherently more sterically demanding ate constitution, even given the trimeric solution constitution of 6 (via DOSY NMR spectroscopy,see Supporting Information). Clearly,w ith ketones and imines any insertion step at the sp 2 O/N would be considerably less congested, thus facile insertion would occur, thereby facilitating the ate enhancement seen in the ensuing hydroboration catalysis.E laborating further, we attempted one further substrate in comparative catalytic experiments with 6 and 3. With 6,1 -phenyl-propyne is only hydroborated in trace amounts,d espite the intrinsically smaller CH 3 group with respect to diphenylacetylene. [6] On the other hand, 3 catalyses the transformation to am ixture of regio-isomers (60 % conversion overall) in favour of borylation at the least sterically hindered alkyne carbon atom, demonstrating once more the advantage of ate complexes in these catalytic transformations.
This study into hydroboration of aldehydes,k etones and imines reveals that anionic ate complexes are important additions to the main-group catalyst toolbox, providing higher conversions in shorter timescales.W ea ttribute this superiority to the greater polarisation of key reaction intermediates induced by the heterobimetallic complexes. Moreover,anovel new catalytic activation pathway was elucidated for ketone hydroboration involving a b-hydride process.W ith internal alkynes the scenario is different and mononuclear species are the catalysts of choice when steric constraints override the ate effect. Overall this study illuminated that while ate complexes are beneficial in most cases, the mononuclear species are more effective in others.Thus,in the field of aluminum-catalysed hydroelementation, there is ah igh degree of substrate dependence,g overning the appropriate choice of catalyst.