Donor‐influenced Structure–Activity Correlations in Stoichiometric and Catalytic Reactions of Lithium Monoamido‐Monohydrido‐Dialkylaluminates

Abstract A series of heteroleptic monoamido‐monohydrido‐dialkylaluminate complexes of general formula [iBu2AlTMPHLi⋅donor] were synthesized and characterised in solution and in the solid state. Applying these complexes in catalytic hydroboration reactions with representative aldehydes and ketones reveals that all are competent, however a definite donor substituent effect is discernible. The bifunctional nature of the complexes is also probed by assessing their performance in metallation of a triazole and phenylacetylene and addition across pyrazine. These results lead to an example of phenylacetylene hydroboration, which likely proceeds via deprotonation, rather than insertion as observed with the aldehydes and ketones. Collectively, the results emphasise that reactivity is strongly influenced by both the mixed‐metal constitution and mixed‐ligand constitution of the new aluminates.


Introduction
Well-defined main-groupm etal complexes are currently the subjectofburgeoning synthetic interest, in both stoichiometric and catalytic transformations. [1] This attention stems from the realisation that, as chemists, we need to develop new sustainable solutionsw ithoutr ecourse to scarce and toxic noble metals,w hilst at the same time attempting to emulatetheir renownedr eactivity.F urthermore, the plethorao fe arth abundant main group metals requires that we need to develop a more fundamental understanding of the potentiala nd the limits of mono-and bimetallic main group metal systems. In this regard aluminium, the most abundant metal in the earth's crust, fits the requirement. Reports of aluminium complexes in important stoichiometric and catalytic processes are becoming increasingly common in the literature. For example aluminium reagents are emerging as important reagents in (catalyst free) cross-coupling protocols [2] and in metallation. [3] In the latter case, we recently reported that heteroleptic iBu 2 AlTMP in tandem with LiTMP can metallate ar ange of sp 2 -a nd sp 3 -CÀH bonds, albeit the presence of the bulky TMP anion bound to lithium is crucial in the CÀHb ond activation. [3f-k] In fact it is only in rare cases with relatively acidic hydrogen atoms that aluminium reagents in isolation have demonstrated utility in deprotonative metallation of organic CÀHs ubstrates. [1a] In the catalytic arena vis-à-vish ydroboration, the use of aluminium complexes is gaining momentum. [4] Importantly,R oeskye tal. utilized a b-diketiminato stabiliseda luminium hydride complex in hydroboration of alkynes andc arbonyl groups. [4a] Recently, the groups of Cowley and Thomas demonstrated that the commerically availableD IBAL(H)o rE t 3 Al·DABCO are capable of catalysing the hydroboration of alkynes. [4c] At this point, most aluminium based catalysts have been neutral complexes, thoughr ecent reports have implicated boratesa si mportant species in hydroboration. [5] Since our groups interest lie in the synergistically beneficial interplay of two distinct metal centres in ab imetallic "ate" complex, this prompted the question whether alkali-metal aluminates would demonstrably imparte xciting reactivity to hydroboration chemistry.T herefore, in ar ecent communication we reported ar ange of Lewis donors olvated lithium aluminates bearing two HMDS (1,1,1,3,3,3-hexamethyldisilazide) and two hydride functionalities and established that lithium diamidodihydridoaluminates were able to function efficiently in hydroboration catalysis and metallation applications. [6] Very recently,f ollowing their aforementioned alkyneh ydroboration advances, Cowley and Thomas have been successful in hydroboration of alkenesu sing the commercial ates, LiAlH 4 and sodium bis(2-methoxyethoxy)aluminum hydride. [7] In our lithium diamidodihydridoaluminates cases the nature of the supporting Lewis donor (level of solvation, N-donor versusOdonor) played an influential role in catalytic performance, where easily displaced monodentate donors performed better than the polydentate donors,p resumably because the chelating effect proved to be deleterious, blockingt he active metal sites. Moreover,w ealso observed similarp henomenai narelated catalytic dehydrocoupling process using the metal hydride surrogate, dihydropyridine precatalyst, 1-Li-2-tBu-1,2-dihydropyridine. [8] The Okuda group also recentlyr eported a series of alkali metal hydridotriphenylborates, [9] that showed the nature of the Lewis donor (including flexibility of coordination and consequential Lewisa cidity of the metal) impacted hydroboration performance in an unpredictable manner.T hus, in main group (bi)metallic catalysis, even small changes in the nature of ancillary ligand(s) impart large differences in the ensuing reactivity.G iven this, here we have examined for the first time structure-activity relationshipso fd onors olvated heteroleptic dialkyl-monoamido-monohydrido complexes using TMP (2,2,6,6-tetramethylpiperidide) as the amide of choice. TMP is a superiorb ase to its HMDS counterpart and is arguably the most important utility amide having taken over from diisopropylamide, [10] on account of its widespread employment not only in LiTMP but in as eries of bimetallic formulations such as Knochel's salt-supported magnesium and zinc reagents [11] and the organometallic ate type reagents introduced by Kondo/ Uchiyama/Wheatley,M ongin, and ourselves. [3,12] Thus, we report here the synthesis of as eries of these new aluminates, investigate their solid state and solution structuresa nd compare their reactivity in hydroboration of aldehydes andk etones, and assess their ambi-utility in metallation and addition reactions.

Synthesis and characterisation
Cocomplexation between the commercially availablea luminium hydride DIBAL(H) and LiTMPi nn-hexane at room temperature resulted in the immediate precipitation of aw hite powder 1 that so far has resisted recrystallisation thusr uling out X-ray crystallographic authentication. However, NMR spectroscopic studies have revealed its general constitution. 1 Ha nd 13 Cs pectra reveal the presence of two isobutyl groups ando ne TMP group and the hydride ligand.T he 7 Li spectrum displays a sharp resonance at d = 0.4 ppm albeit the 27 Al spectrum does not contain any identifiable resonance, ac ommon problem with quadrupolar aluminium centres in broadly unsymmetrical environments. The presence of an aluminium hydride was confirmed by performinga 1 H{ 27 Al} experiment whichr evealed a sharpening of the broad resonance at 3.18 ppm, and suggested formation of the expected cocomplex iBu 2 AlTMP(H)Li 1 in a 75 %i solated yield. Structural insight of the solution phase constitution of 1 was gainedb yp erforming aD OSY experiment in C 6 D 6 solution, [13] revealing the likely aggregation state of the solvent-free 1 as dimericw ith the formula [iBu 2 AlTMP(H)Li] 2 .S imilars olvent-frees tructures are knowni n the literaturea nd contain the common [Li 2 H 2 ]c entral square core depicted in 1. [14] Next, as eries of donor-solvated derivatives of 1 were obtained in high yield by addition of the appropriate ligand (THF, PMDETA, diglyme or DABCO) into toluene solutions of 1,f ollowed by introductiono fn-hexane and crystallisation at À30 8C( Scheme1). In each case, single-crystals suitable for Xray diffractions tudies wereo btained and revealed ar emarkable variance in the molecular arrangements. THF adduct 2 and PMDETAa dduct 3 in effect exhibit the same structure, wherein ad istorted tetrahedral iBu 2 AlTMP(H) fragment is bonded, via the m-hydride, to aL i·donor fragment (donor = three THF in 2 and one PMDETA in 3). Both the AlÀHa nd LiÀH distances are similar in each molecule [AlÀH1 .61(3) in 2, 1.66(4) in 3;L i ÀH1 .83(3) in 2, 1 .76(4) in 3], with those to the group 13 metal systematically shorter by an average of 0.16 .C hanging the Lewis donor from PMDETA to diglyme gives 4,which adoptsamarkedlydifferent solid-state structure. The main differences between these tridentate chelating ligands is ac hange from Nt oOdonors, and importantly the latter only contains one terminal methyl substituent, reducing steric congestion when it ligatesametal atom. Charge-separated ion pair structure 4 is best described as al ithium lithium-dialuminate, since both cationic and anionic moieties contain lithium.T he cationic moiety is ad istorted octahedrall ithium cation supported by two diglyme ligands. The anionic moiety is comprised of two peripheral iBu 2 AlTMP(H)u nits in distorted tetrahedral environments, that bond to ac entral lithiumv ia m-TMP and m-hydride ligands. Furthermore the lithium ion displays an ear square planarg eometry (t 4 = 0.04), [15] albeit it is also likely to be further stabilised by electrostatic interactions from TMP methyl substituents [LiÀC Me range:2 .946(7)-3.249 (7) ]. This arrangementc an be described as an inverse Weiss motif. The Weiss motif is as urprisingly common structure found in dialkali-metal "ate" complexes where the central, non-alkali-metal exists in at etrahedral arrangement, with respect to four bridging ligands. [16] Here in contrast we have a centrals quare planar alkali-metal that bridges through ligands to two peripherala luminium atoms.Asearch for this structural motif in the Cambridge StructuralD atabase (CSD) resulted in zero entries, confirming the structural rarity of 4.R eaction of 1 with half an equivalent of DABCO affords 5 in which two symmetry equivalent iBu 2 AlTMP(H)Li subunits are connected by the bicyclic,b initrogen Lewis donor.Aswith 4,here the lithium ion coordination sphere is completed by m-TMP and m-hydride ligands. Employing af ull equivalent of DABCO affords the stoichiometric variant 6,w hich differs from 5 by at erminal DABCO ligand,i nstead of the bridging mode observed in 5.I n2-6,d espite the diversity of donor ligands, the AlÀHa nd LiÀHd istances do not display any systematic differences across the series (AlÀHr ange 1.61(3)-1.69(3) ;L i ÀHr ange 1.76(4)-1.88 (3) ). The structures of 2-6 and the related lithium diamidodihydridoaluminates we reported previously, [6] can be regarded as well-defined modifications of LiAlH 4 containing ar ange of either alkyl or amido groups in place of either two or three hydrogen atoms. Furthermore, these more substituted LiAlH 4 modifications benefit from the synthetic advantages of enhanced solubility( soluble in hydrocarbon solvents), and easier to accurately weigh low loadings in catalytic applications.
In the knowledge that alkali metal effects can profoundly influence structural morphology,r eactivity and physicalp roperties, [12,17] we next attempted to synthesise as odium variant by reactiono fN aTMP with DIBAL(H) in n-hexane. After addition of as mall amounto fT HF to the reaction mixture we were able to isolate af ew crystals of (iBu 2 Alm-TMPm 3 -HNa) 2 ·2THF, 7,t hat were found to be suitable for diffraction studies ( Figure 1). Crystallographic characterisationo f7 revealed as tructure markedlyd ifferent from the lithium congeners, highlighting the impact of replacing lithium with its larger congener. Figure 1s hows an arrangement wherein a( NaHNaH) nearplanar kite shaped ring lies between two iBu 2 AlTMP in an asymmetricf ashion. One sodium is supported by two m-TMP and two m 3 -hydride ligands( Na2ÀH1H 2.40(2) and Na2ÀH2H 2.38(2) ). The second sodium is also bonded to the hydride ligands [Na1ÀH1H 2. 22(2) and Na1ÀH2H 2. 19(2) ]a nd two solvating THFm olecules. Unfortunately,w ew ere unable to reproducibly prepare 7 in acceptable yields and as such its characterisation remained restricted to this structural analysis.
Complexes 2-6 were also characterisedi ns olution by 1 H, 13 C, 7 Li and 27 Al NMR spectroscopy and revealed the expected resonances, albeit as with 1,t he 27 Al spectra were of little diagnostic value. 1 Ha nd 13 Cs pectra confirmed the presence of 2 iBu, 1T MP,a nd an aluminiumb ound hydride ligand, as well as the appropriate donor resonance(s).I nterestingly the 1 HNMR spectrumso f5 and 6 are neari dentical, except for the chemical shift resonances of the DABCO CH 2 protons. Moreover,i n6 the terminally bound DABCO ligand only contains one broad singlet insteado ft wo triplets (vide infra). The 7 Li NMR spectrum of 4 displayed one noticeably broad resonance at d = À0.32 ppm, instead of the expectedt wo resonances in accordance with the charged-separated molecules in the crystal. Thus, av ariable temperature 7 Li NMR experiment was performed to investigate whether an exchange process is occurring ( Figure 2). Measuring a 7 Li NMR spectrum of 4 in a [D 8 ]toluene solution at 0 8Cr esults in as ignificant broadening compared with the room temperature collection. Further cooling at temperatures down to À60 8Cr esults in the appearance of two broad and one sharp resonance at d = 1.68, À0.36 and À1.96 ppm, indicating the likelihood of an exchange process occurring at room temperature. The identity of the three resonances can be tentatively assigned. Those at d = 1.68 and À1.96 can be assigned to as tructure that resembles that in the solid state, with the latter sharper resonance corresponding to the approximately octahedral Li·(diglyme) 2 cation. The resonance at À0.36 ppm may be assigned to ac ontact ion pair resembling PMDETAa nalogue 3,w hiche xhibits as harp resonance at d = 0.38 ppm. We probed this exchange processf urther via aD OSY NMR study of 4 at room temperature in C 6 D 6 . The estimated molecular weights in this instancea re 509 and 615 gmol À1] (note the Lewis donor resonances diffuse separately from the remaining resonances), which are intermediate between thoseo fe ither the expected mass in the crystal (846 gmol À1 )o rastructure resembling 3 (424 gmol À1 ). In contrast, DOSYN MR studies of 3 indicate its greater robustness as no such exchange process occurs on the NMR timescale, where an estimated molecular weight( 463 gmol À1 )i si ne xcellent agreement with the theoretical value (446 gmol À1 ).
In summary,c omplexes 1-6 represent as eries of new monoamido-monohydrido lithium aluminates, welld efined in the solid and/or solution states. Key distinctions are 1 is donor  free; 3 contains as olvating PMDETAl igand in the solid state and retains this arrangement in solution (via DOSY studies); 4 adopts an entirely novel solid-state arrangementv ide supra, and exhibits exchange of the diglyme ligands on the NMR timescale.

Hydroboration of aldehydes and ketones
At the inception of these studies we sought to discover the answer to two key questions: (i)can these monoamido monohydrido lithium aluminates act as efficient catalysts in hydroboration applications;a nd (ii)does the structure, determined by the donor ligand play arole in any such catalytic performance? In this regardw ei nitiallys elected 1, 3,a nd 4,s ince 1 is donorfree, while 3 and 4 contain broadly similard onor ligands, except for the donor atom identity (N or O), and the slightly reduced steric profile of diglyme whichg ives rise to af luxional system in solution. Hydroboration catalysis using HBpinw as selected to trial 1, 3 and 4,s ince it is an area currently attracting increasing interesti nt he main-group arena, andt he wider chemicala udience. [18] Importantly,s imple aluminium reagents (AlEt 3 ·DABCO, DIBAL(H),L iAlH 4 ) [4c, 7] have recently been discovered as excellent catalysts in this regard, although occasionally the presence of as econd metal has been overlooked when considering the elementary steps of the catalytic profile. Given our groups' longstanding association with bimetallic "ate" complexes,w ep ondered whether these systems might also shed further light on the reactionp athways and any significant effect of the second metal. Results of catalytic hydroboration of aldehydes and ketonesare presented in Ta ble 1.
Reaction of 1 (2.5 mol %b ased upon ad imeric formula), 3 (5 mol %) or 4 (2.5 mol %) with benzaldehyde and pinacol borane in C 6 D 6 solution resulted in fast and quantitative hydroboration in 15 minutes at room temperature in every case. Lowering the precatalyst loadings to 0.5 mol %o f1 or 4 and 1mol %o f3 resulted in no loss of catalytic performance. Importantly these initial results demonstrate that 1, 3 and 4 are all suitable candidates for aldehyde hydroboration. Expanding the substrate scope to include dual-functional cinnamaldehyde demonstrates catalysts electivity with smooth and quantitative hydroboration occurring only at the carbonyl functionality. 1 and 4 achieve this transformation inside 30 min, whereas with 3 quantitative hydroboration occurs after 3h.T his demonstrates that subtle structurald ifferences within theses ystems play arole in catalystperformance. Thisperformance is broadly in line with previous aluminium-based hydroboration catalysts. Roesky et al. achieved quantitiative hydroboration of cinnamaldehydew ith his nacnacAlH 2 complex in 6h with 1% loading, whereas we previously demonstrated 76 %c onversion in 2h with (HMDS) 2 AlH(m-H)Li(THF) 3 .C ompleting our studies into aldehydeh ydroboration we screenedc omplexes 2, 5 and 6 with cinnamaldehyde as ar epresentative example. 2 containsafour coordinate lithium atom albeit bonded to three monodentate THF ligands, which are more likely to desolvate during the reaction, compared to tridentate PMDETAi n3,t hus substrate access to lithium may occur more readily. 5 and 6 both have coordinatelyu nsaturated three coordinate lithium atoms, and as such are readily accessible. Complexes 2, 5 and 6 all efficiently catalyse hydroboration of cinnamaldehyde in 30 min.
Encouraged by these findings we extended our substrates to include ketones, rationalising the increased intrinsic steric bulk compared to that of aldehydes would magnify any ligand effects in these systems. Hydroboration of acetophenone with 1 (2.5 mol %) reaches 97 %c onversion after 2.5 ha tr oom temperature in C 6 D 6 .U sing 3 as precatalyst,t he reactionr equires 65 hf or quantitative hydroboration at room temperature, or 22 ha t7 08C. Using 4 this reaction takes 2h for 90 %c onversion, similart ot he value using 1.E xploring the comparatively poor performance of 3 more thoroughly,2 ,2,2-trimethylaceto- 1 [d] 3 [d] 4 [d] 0 phenonea nd 2,2,2-trifluoroacetophenone were selected as substrates. The former shows enhanced steric features with respect to acetophenone, whereas the latter,a pproximately isosteric, is considerably more electron withdrawing. Hydroboration of 2,2,2-trimethylacetophenonew ith 3 (5 mol %) occurs in 2.5 ha tr oomt emperature, implying that ligand steric features are not the only factor here in determining hydroboration efficiency.2 ,2,2-trifluoroacetophenone is quantitativelyh ydroborated in 15 min at room temperature. This fast reactivity may be expected since the presenceo fC F 3 would significantly deplete the charge present at the ketonec arbon atom makingi t more electrophilic andt hus facilitatef aster nucleophilic hydride insertion. One possible contribution towards the slow reactivity of 3 with acetophenone is that the presence of a-hydrogen atoms may promote keto-enol tautomerism under the reactionc onditions. In this scenario, one may envisage ac oordination of the substrate to the lithium ion, resulting in as terically congested 5-coordinate lithium ion, followed by tautomerisation and stabilisation of the enol form by H-bonding to one nitrogen atom of the PMDETAligand (Scheme 2). However, it is expected that the keto-form would predominate in solution, and we detect no spectroscopic evidence of its enol tautomer.N ote no such tautomerisation and H-bond stabilisation is possible with the CF 3 analogue.Asecond suggestion fort he slow catalytic transformation of acetophenone with 3 is that a methyl hydrogen atom may be deprotonated by the TMP group. However,s ince this reactione ventually resultsi nq uantitative hydroboration, any deprotonation must be under equilibrium. Hydroboration catalysis of 1, 3 and 4 with benzophenone again reveals that 3 performsl ess efficiently( 3h for quantitative formation) than either 1 or 4 (30 min). Using cyclohexanone as substrate reveals that 3 is again slowest (3 h);w hereas 4 performsm arginally better than 1 (1.5 hv ersus 2h)f or this reaction. Finally,b utan-2-one, an aliphatic ketone, is fastest using 1 as ac atalyst( 1.5 h), while both 3 and 4 are complete in 3.5 hatr oom temperature. Butan-2-one contains two sets of a-hydrogensw hich may explain the slower reactivity with respect to 1.H ydroboration of acetophenone was also performed using 2, 5 and 6.A ll three complexes demonstrate essentially the same reactivity (2,3 0min, 93 %c onversion, 5 and 6,4 5min, quantitative). This can be rationalised by considering the structurald istinctions of the complexes.I ne ach case the DABCO ligand is monodentate, rendering the lithium atom 3coordinate, instead of the fully saturated 4-coordinate in 2-4, albeit the high labilityo ft he solvating THF in 2 is likelyt o lower the coordination number as requiredd uring reaction. Therefore, one may anticipate that initial coordination of substrate to lithium, and thuss ubsequent insertion into the AlÀH bond would occur more readily in these cases. Importantly this scenariov alidates the use andf uture exploration of bimetallic systems.
The generalt rend observedb etween 1, 3,a nd 4 in these reactions reveal that 1,t he donor free complex is the best catalyst and marginally outperforms 4. 3 in all cases is the least efficient catalyst in these processes. This reactivityc an be related to the solid and solutionp hase structures. The structure of 3 in solution,via DOSY NMR studies, resemblest hat in the crystal structure. Thus, one expects the PMDETAg roup with its three donor atoms to remaint ightly bound to lithium during the catalytic process. On the other hand, 4 can be considered a halfway-houseb etween 1 and 3.S pectroscopics tudies indicate that part of the time in solution the complex resembles that of the solid-states tructure, which is donor free, akin to 1, (both diglyme ligandsa re involved in bonding to the separated lithium cation). The remainder of the time the complex likely resembles 3,a nd thusi ng eneral, delivers reactivity intermediate between that of 1 and 3.Alower lithium coordination number also plays ar olei nc atalytic performance as indicated by reactivity of 2, 5 and 6 which hydroborate acetophenone faster than the other donor solvates studied.O nce again,w e turned to DOSY NMR studies to illuminate the solutionc onstitution of these complexes to gain insighti nto the catalytic process. The estimated molecular weight of 2 is 486 gmol À1 ,i n good agreement with the theoretical value (506 gmol À1 ), suggesting that 2 largely remains intact in solution. However, careful inspection of the spectrum reveals that the THF resonances diffuse slightly faster than the remaining resonances of 2 (MW 444 gmol À1 ), detailing that in solutions ome level of THF desolvation is in operation, lowering the metal coordination number.W ith 5 and 6 the solutionc onstitution is ambiguous. Bridging DABCO complex 5 has an estimated molecular weight of 620 gmol À1 ,l ower than the expected value of 691 gmol À1 . Similarly terminally boundD ABCO complex 6 has an estimated molecular weight of 514 gmol À1 ,h igher than the expected value of 402 gmol À1 .T hese intermediate values indicatet hat the solution constitutions may be in equilibria between the two crystallographically observed extremes, and moreover,a ccount for the similarity in 1 HNMR spectra of 5 and 6 (vide supra).I ne ach case the DABCO resonances diffuse with the same coefficient as the remaining complex resonances, ruling out complete ligand desolvation,a ffording 1,o nt he NMR experiment timescale. Extending this argumento ne step further, we performed two representativer eactions using bulk THF as reactions olvent. We selected 1 and 3 for these reactions and discoveredt hat in each case quantitative hydroboration occurs within 30 min. These fast reactions can be tentatively attributed to, in the case of 1,b reaking up of the dimeric aggregate into am onomeric solvent separated species, and in the case of 3 displacingt he PMDETA ligand resulting in as imilarm ore labile solution species, an arrangement that favours faster catalytic transformation. Corroborating this, the 1 HNMR spectrum of 3 in [D 8 ]THF solution reveals the presence of free PMDETA.
Following from the hydroboration results we sought to discover some mechanistic evidence by performing as eries of stoichiometric reactions between complexes 1 or 3 with various substrate molecules. Unfortunately,w ew ere unable to crystallographically characterise such as pecies, despite repeated attempts. However,t he different hydroborationr esults, that is, the effect of the donor ligand leads us to surmise that hydroboration of aldehydes and ketones in C 6 D 6 follows three basic elementary steps:( i) coordination of carbonyl substrate to lithium. Here, tentative evidence for the coordinations tep originates from the slow performance of 3 with acetophenone, (see Scheme2); (ii)insertion of carbonyl into aluminiumh ydride bond,a lready primedf or addition;( iii)transelementation with HBpin to afford hydroborated product and regenerate an active aluminium hydride catalyst. Scheme 3d etailst he proposed elementary steps in the reaction.

Metallation
Complexes 1-6 all contain ab asic TMP and an ucleophilic hydride ligand and all are competent, in varying degrees, in catalytic hydroboration transformations. Well-defined PMDETA complex 3 was tested, as ar epresentative complex in metallation reactions with substrates containing an acidic hydrogen atom. Metallation of 1,2,4-triazoles in general can be problematic and often requires low reaction temperatures, to prevent unwelcome fragmentation reactions of metallated (commonly lithiated) intermediate species. [19] Reaction of 3 with 1-methyl-1,2,4-triazolea tr oom temperature for 1h in a n-hexane/toluene mixture, followed by cooling at À30 8Cr esulted in formation of colourless crystalss uitable for X-ray diffraction studies (Scheme 4).
In agreement with the spectroscopic data (see Supporting Information) the X-ray diffraction data revealed that the substrate hasbeen metallated at the 5-position, via amido basicity, giving 8 in ar easonable isolated yield (42 %). Heterotrileptic 8 ( Figure 3) crystallises with two independentm olecules in the asymmetricu nit. In each case an iBu 2 AlH unit bonds to the 5positiono ft he triazole ring. Furthermore, the AlÀHd istances are the same within experimental error in each independent molecule 1.57(2) and 1.58(2) .Alithium·PMDETAm oiety is held in place by coordination from the triazolyl nitrogen atom of the aluminate moiety placed adjacent to the metallated carbon atom. Further,i no ne independent molecule (RHS Figure 3) the hydride ligand bridges to Li [LiÀH2 .23(2) ], while in the second (LHS Figure 3) the LiÀHd istance is significantly longer[ LiÀH2 .40 (3) ]. Further inspection of the structure reveals ak ey factor for this drastic difference. In each case the PMDETAl igand coordinates to the lithium in ad ifferent manner.I no ne molecule (RHS) the methyl group attachedt o the central nitrogen atom lies approximately parallel to the LiÀ Hi nteraction, whereas in the other (LHS) this same methyl group points in the opposite direction, af act we attributet o packing effects in the crystalline lattice.
It is important to note that the aluminium reagent iBu 2 AlTMP on its own can also metallate 1-methyl-1,2,4-triazole. This reactiona ffords (iBu 2 AlC 3 H 4 N 3 ) 2 , 9 in 30 %i solated yield, and is ar are example of an eutral bis-alkylamidoaluminium compound acting as an amido base towards an aromatic CÀH bond, albeit with hydrogen atoms in relatively acidic environments. [20] Though X-ray crystallographicd ata for 9 were collected, these were of insufficientq uality to justify publication (see Supporting Information).
Phenylacetylene was selected as ac andidate metallation substrate because of its acidic CCH hydrogen and because of precedence from Roesky's research thati mplicated deprotonation as ak ey step in terminal alkyne hydroboration. [4b] In that Scheme3.Representation of possible elementary steps using 3:coordination of substrate;insertion into AlÀHb ondand s-bond metathesis with pinacol borane.
Scheme4.TMP basicity of 3 with 1-methyl-1,2,4-triazole, affording 8. report the basicity derivesf rom ah ydride of a b-diketiminato stabilised AlH 2 unit. Using 3 (as ar epresentativee xample, and the PMDETA ligand to try and induce crystallisation of the formed product) we soughtt od iscover whether deprotonation or addition would occur,a nd whether any deprotonation would derive from the more basic TMPu nit or weaker hydride, and whether these complexes could then be implicated in subsequent catalytic hydroboration reactions. Reaction between 3 and phenylacetylene in aJ .Y oung's NMR tube was carriedo ut at room temperature. 1 HNMR monitoring revealed the appearance of resonances corresponding to TMPH. Furthermore, the resonance corresponding to the acidic hydrogen of phenylacetylene disappeared, confirming that 3 reacts, once again,a sa n amido base (Scheme 5). The reactionw as repeated in at oluene/n-hexane mixture at room temperature. After stirring the solution for 2h at room temperature 10 was isolated in 57 % yield as an off-white solid. Spectroscopic characterisation of isolated 10 corresponds to am ono(alkynyl)m onohydrido lithium aluminate, iBu 2 AlHCCPhLi·PMDETA. Its structure closely resembles thato f3 via formal replacement of aT MP anion with ap henylalkynyl anion as was the case with the triazolyl anion in 8.N ote similar structures are knowni nt he literature, and Uhl reported dialkylaluminiuma lkynides adopting the bridging motif shown in 10,m oreover with less bulky alkyl groups,a crystallographically authenticated pi interaction between aluminium and the triple bond was identified.
By extension to the presents ystem,s uch an interaction between the triple bond and the carbophilic aluminium centre may prime the complex for insertion. [21] Importantly, 10 is similar to an in silico modelled activec atalytic intermediate by Roesky et al.,( an eutral b-diketiminato Al-alkynyl hydride) from which the BÀHb ond of pinacol borane adds across the triple bond, followed by addition of as econd equivalent of phenyl acetylene to regenerate the active component. Since Roesky's neutrali ntermediate complexw as found to be an excellent catalysti na lkyne hydroboration, the isolation of the bimetallic complex 7,b ys imple deprotonation, afforded the opportunity to learn about the role of anionic aluminates in this context (note that in this report we have two iBu ligandsi nstead of a bulky b-diketiminate).
In aJ .Y oung's NMR tube 1 (2.5 mol %) was dissolvedi nC 6 D 6 to which phenylacetylenew as added, followed by HBPin and the reaction monitored by 1 HNMR spectroscopy.A fter heating for 18 ha t7 08Cc lean formation of the anti-Markovnikov vinylboronate ester has occurredi n7 6% yield, as referenced against an internal standard (Scheme 6). The experiment was repeated, with 3 instead of 1 as catalyst. In this case hydroboration only occurs very slowly,a ffordingo nly around 5% yield after heating at 70 8Cfor 18 hours.
This desirable outcomet herefore suggestst hat well-defined "ate" complexes can play an important role in hydroboration of av ariety of unsaturated molecules. Scheme 7o utlinesv arious potential reaction pathways in this process. Pathwaya ,i n Scheme7describes an insertion pathway,t hat is ah ydroalumination followed by a s-bond metathesis, analogous to that discussed( vide supra) for aldehyde and ketone hydroboration. Moreover,arecent report from Cowley and Thomas discuss a Scheme5.TMP basicity of 3 with phenylacetylene, affording 10.
Scheme7.Alternative potential reaction pathways for lithium aluminate catalysedh ydroboration of phenylacetylene. Chem. Eur.J.2018, 24,9940 -9948 www.chemeurj.org 2018 The Authors. Published by Wiley-VCH Verlag GmbH &Co. KGaA, Weinheim similar mechanism for the DIBAL(H) catalysed hydroboration of alkynes. Since deprotonation was observed in stoichiometric reactions between phenylacetylene and either 1 or 3,w er ationalise that this is ak ey step in any hydroboration catalysis of phenylacetylene, using the complexes herein. Interestingly, attempts to corroboratet his observation proved unsuccessful. Reactions between 10,o rt he PMDETA-free variant, generated in situ, with HBpin both afforded 1 Ha nd 11 BNMR spectrat hat were broadly uninterpretable, and indicative of decomposition of the reaction mixtures in stoichiometricr egimes. As econd proposal (Pathway b) proceeds in broad agreementt ot hat of Roesky using an eutralN acNacAlH 2 catalyst. Here, lithium aluminate 1,p erformsa ni nitial deprotonation, then hydroboration occurs, followed by ad eprotonation of as econd phenylacetylene molecule, inducing catalytic turnover.
Next, we decided to attemptt he hydroboration of an internal alkyne.R easoning that since di-phenylacetylene did not contain any hydrogen atom "primed" for deprotonation, that any hydroboration activity would, by necessity,o ccur via an insertion of the alkynei nto the aluminium hydride bond, and furthers hed light on the likely reaction pathway.T hus, diphenylacetylene wasa dded to aJ .Y oung's NMR tube in C 6 D 6 ,f ollowed by 1 (2.5 mol %) and HBpin. After heating the solution at 70 8Cf or 18 ht he 1 HNMR spectrum remainedu nchanged revealing that the addition of an AlÀHa cross diphenylacetylene is not favoured, and gives credence to the reaction mechanism depictedp athway bi nS cheme 7, although based on the resultso fs toichiometric reactions of 10 with HBpin,w e cannotcompletely discount pathway a.

Addition reactions
Despite the fact that addition of the AlÀHb ond did not freely occur across diphenylacetylene in the attempted hydroboration catalysis, we wanted to test whether addition reactions were av iable utility of thesel ithiuma luminate complexes, thus we elected to react 3 with pyrazine. Both metallation and addition across pyrazine has previously been observed depending on the specificr eagent(s) employed, [22] prompting a consideration of whether the basicity of 3,a rising from the TMP ligand,o rt he nucleophilicity of the AlÀHf unctionality would win out. At oluene/n-hexane solution of 3 was stirred with one equivalent of pyrazinea tr oom temperature for 2h resultingi nf ormationo fapale brown oil. 1 HNMR spectra of the oil revealed clean formation of as ingle product, displaying four resonancesb etween d = 7.3-4.0 ppm, in a1 :1:1:2 ratio consistentw ith hydride addition at the a-carbon atom (Scheme 8). Importantly this result indicates that as well as being competent hydroboration catalysts, these complexes demonstrate great versatility as either metallating agentsf or acidic hydrogen atoms, or as H À sourcesi na ddition reactions across pyrazine. Furthermore, it detailst he fine balance in these systems on the ensuing reactivity,w hich is highly substrate dependent.

Conclusions
This contribution has demonstrated the synthesis, spectroscopic ands tructural characterisation of an ew family of donor-solvated heteroleptic dialkyl-monoamido-monohydrido complexes. The presence andnature of the Lewis donor imparts interesting structuralc haracteristics that influence the catalytic activity in hydroboration reactions of aldehydes and ketones with pinacolborane.P olydentate donors (PMDETA) that remain boundt ot he lithium atom in solution slow down hydroboration. On the other hand, the related tridentate ligand diglyme, displays exchange on the NMR timescale, and performs similarly to the donor free species. We suggest this is, at least in part due to the coordination saturation of the lithium ion. Using the Lewis donor ligand DABCO leads to ar eduction of the solvation at lithium and leads to faster hydroboration in ar epresentativeh ydroborationo fak etone.I nt hese catalytic transformations, insertion of the polar carbonyl group into the AlÀH bond is mooted as ak ey step. Heterolepticd ialkyl-monoamido-monohydrido complexes are also revealed to be capable of hydroborating phenylacetylene, however this activity is suggestedt op roceed via deprotonation of the substrate. The bifunctional activity of the complexes is also demonstrateds toichiometrically in metallation of substrates containing an acidic hydrogen atom, and in an addition reactionwith pyrazine.

Experimental Section
Full experimental characterisation and synthetic procedures are described in the supporting information.