1‐Alkali‐metal‐2‐alkyl‐1,2‐dihydropyridines: Soluble Hydride Surrogates for Catalytic Dehydrogenative Coupling and Hydroboration Applications

Abstract Equipped with excellent hydrocarbon solubility, the lithium hydride surrogate 1‐lithium‐2‐tert‐butyl‐1,2‐dihydropyridine (1tLi) functions as a precatalyst to convert Me2NH⋅BH3 to [NMe2BH2]2 (89 % conversion) under competitive conditions (2.5 mol %, 60 h, 80 °C, toluene solvent) to that of previously reported LiN(SiMe3)2. Sodium and potassium dihydropyridine congeners produce similar high yields of [NMe2BH2]2 but require longer times. Switching the solvent to pyridine induces a remarkable change in the dehydrocoupling product ratio, with (NMe2)2BH favoured over [NMe2BH2]2 (e.g., 94 %:2 % for 1tLi). Demonstrating its versatility, precatalyst 1tLi was also successful in promoting hydroboration reactions between pinacolborane and a selection of aldehydes and ketones. Most reactions gave near quantitative conversion to the hydroborated products in 15 minutes, though sterically demanding carbonyl substrates require longer times. The mechanisms of these rare examples of Group 1 metal‐catalysed processes are discussed.

With as eries of soluble alkali metal hydride surrogate congeners in hand we sought to examinet heir application in two distinct catalytic processes, namely dehydrocoupling of amine boranes and hydroboration of aldehydes and ketones. In each reactionm etal hydride species have been found to either catalyse or have been identified as key intermediates in the process. The controlled formation of boron-nitrogen bonds by dehydrocoupling of amine boranes, HNR 2 ·BH 3 (R = H, alkyl) is ar eaction that attracts widespread attention in the synthesis of novel polymers and ceramics, [11] and in the arena of hydrogen storagem aterials. [12] Thus over the past two decades, much activity has been directed at transition-metal-catalysed dehydrocoupling of ammonia borane and amine boranes, and moreover much insight has been garnered regarding mechanistic aspects of the various catalytic pathways. [13] Recent insightful work from the groups of Harder, [14] Hill [15] and Wright,[16] among others, [17] demonstrated that main group (d 0 )c omplexes are active in both stoichiometric and catalytic dehydrocoupling of main group element-H bonds. Furthermore, Bertrand demonstrated that cross-dehydrocouplingo fs econdary boranes with alcohols, thiols and amines can be accomplished without acatalyst. [18] It is also noteworthy that precatalysts discussed in these reportst ended to be more economically viable and environmentally innocuous than their invariably expensive and toxic noble transition metal counterparts, albeit at this point they do not (yet) match the best catalytic efficiencies. Among the most studied main group precatalysts are those from Group 2a nd Group 13 whicht ypically contain bulky b-diketiminato or (silyl)amide ligands. Similarly these Group 2c omplexes and related speciesh ave been found to catalyset he hydroboration of ar ange of substrates, including pyridines, [6] aldehydesa nd ketones, [19] nitriles, [20] isonitriles, [21] and esters. [22] Impressively the hydrosilylation of alkenes using ap otassium hydridec atalyst was reported by Harder. [23] More recently Okudah as provided mechanistic evidence for potassium catalysed hydrosilylation of ar ange of alkenes using aK (18-crown-6)(SiPh 3 )c atalyst. [24a ] Further,t he Okuda group hasr ecently demonstrated that alkali metal hydridotriphenylborates can catalyset he hydroboration of benzophenone. [24b] These transformations, for example converting an aldehydei nto an alcohol, are of central importance within organic chemistry and have historically been accomplishedu sing stoichiometric metal hydride species, for example LiAlH 4 ,w hich can suffer from poor functionalg roup selectivity and low solubility in hydrocarbon solvents. [25] Thus utilisation of milder hydride sources (e.g.,b oranes) in tandem with as uitable catalyst remains a tantalising synthetic strategy.B reakthroughsr eportedh erein will extend the versatility of hydrocarbon soluble Group 1 DHPs as metal hydride surrogates in the catalytic dehydrocoupling of amine boranes andi nh ydroboration of aldehydesa nd ketones.W ea lso disclose the crucial importance of reaction solvento nc atalytic efficiency.
Hill recently noted the first example of Group 1s ilylamide precatalysts [MN(SiMe 3 ) 2 ,M= Li, Na, K] for dehydrocoupling of dimethylamine borane. [27] 5mol %o fL iN(SiMe 3 ) 2 in toluene gave the best conversion,d etermined by 11 BNMR, to 72 % [NMe 2 BH 2 ] 2 and 5% (NMe 2 ) 2 BH after heatinga t8 08Cf or 124 h. In this study an intermediate potassium [NMe 2 BH 2 NMe 2 BH 3 ] À complex was isolated, indicating that the catalysis likely follows that suggested for Group 2a nd 13 precatalysts ( Figure 2A). These important results are more impressive given that the catalytically activem etal hydride species are reported to form insoluble aggregates during the experiments, slowing down the process, particularly for the heavier alkali metal silylamides NaN(SiMe 3 ) 2 and KN(SiMe 3 ) 2 .S olubility problemsh ave also been encountered by Wright on employing LiAlH 4 as ac atalyst in as imilarr eactionw ith HNMe 2 ·BH 3 ,a nd by Panda in the LiN-(SiMe 3 ) 2 catalysed cross-dehydrocoupling of HBpin or 9-BBN (9borabicyclo[3.3.1]nonane) with ar ange of amines, anotherr are example of Group 1c atalysis. [28] It is therefore apparent that effective solubility of key metal hydrides is criticalf or high catalytic efficiency.G iven that 1tLi represents as oluble sourceo f lithium hydride in hexane, we reasoned that the in situ generated metal hydride would exist as as oluble dihydropyridine species, thus enhancing the catalytic process.
Reaction between 2.5 mol % 1tLi and HNMe 2 ·BH 3 in [D 8 ]toluene at 80 8Cr esultsi nc onversion (determined via 11 BNMR integrals) to 89 %o f[ NMe 2 BH 2 ]a nd 4% of (NMe 2 ) 2 BH after 60 h( Ta ble1 entry 2). Significantly this reaction proceeded faster than that of 5mol %[ Mg{CH(SiMe 3 ) 2 } 2 (THF) 2 ]w ith dimethylamine borane in [D 6 ]benzene (72 ha t6 08C), indicating that 1tLi is ac ompetitive precatalyst. [15a] The in situ 1tLi induced reactionw as monitored by 11 BNMR spectroscopy (Figure 2B) revealing the presence of severals pecies (identified by comparison with literature data where appropriate). Initial mixing of the reagents in aJ.Y oung's NMR tube resulted in immediate H 2 gas evolution. This observation may be tentatively ascribed to the initial reaction between 1tLi and HNMe 2 ·BH 3 formingL i[NMe 2 BH 3 ]( I), 2-tert-butylpyridine and H 2 (Scheme 2A). To be consistent with our hypothesis we expect that 2-tert-butylpyridine will act as aL iH storage/release vessel during the process, by forming dihydropyridines as ar esult of interaction with Li[amidoborane] species (Scheme 2B). At the initial time point the 11 BNMR spectrum displays two resonances:atripleta td = 3.4 ppm ( 1 J BH = 100.1 Hz) corresponding to Li[NMe 2 BH 2 NMe 2 BH 3 ]( II) [27] and aq uartet composed of the mutually coincident signals [16c, 27] The last named is formed by polar insertiono fh ighly reactive NMe 2 BH 2 into Li[NMe 2 BH 3 ], in line with the literature mechanism. Analysis of the 11 BNMR spectrum after heating the solution at 80 8Cf or 24 hours reveals the presence of several new species:adoublet at d = 28.9 ppm ( 1 J BH = 129.9 Hz) confirmed as (NMe 2 ) 2 BH (III); [29] at riplet at d = 5.4 ppm ( 1 J BH = 113.1 Hz) assigned to cyclic dimer [NMe 2 BH 2 ] 2 (IV); [30] ap artially obscured quartet centred around d = À11.0 ppm ( 1 J BH = 91.1 Hz) assigned to Li[NMe 2 (BH 3 ) 2 ]( V); [16c] and aq uintet at d = À40.9 ppm ( 4 ]a nd has been noted before by Wright, who rationally synthesized and structurally characterisedt he compound. [16c] As the reaction progressesi ti sa pparent from 11 BNMR data that the metallated amidoboranes are consumed. In the case of Li[N-Me 2 BH 2 NMe 2 BH 3 ]i ti sc lear that the major process is d-hydride elimination to produce [NMe 2 BH 2 ] 2 (IV). We propose that Li[N-Me 2 (BH 3 ) 2 ]i sc onsumed via one (or both) of two similar routes. The first scenario involves ah ydride transfer which would reform Li[BH 4 ]a nd also generate BH 2 NMe 2 (Scheme 3A). Both compounds could then re-enter the catalytic cycle, or in the latter case an off-metal dimerization pathway is conceivable. Alternatively,am olecule of NMe 2 BH 2 could insert into Li[N-Me 2 (BH 3 ) 2 ]g iving [NMe 2 BH 2 ] 2 and Li[BH 4 ]d irectly (Scheme3B). Although ad efinitive pathway has not been discovered it is clear that Li[NMe 2 (BH 3 ) 2 ]i sa ni mportantp roduct-forming inter- Af urther important observation from 11 BNMR data is that at high conversions to products, that is, low concentrations of HNMe 2 BH 3 /Li[NMe 2 BH 3 ]atriplet of very low intensity is observed at d = 38.1 ppm ( 1 J BH = 132.8 Hz) correspondingt o NMe 2 BH 2 .T he presenceo ft his intermediate is somewhat surprising since it reacts/inserts very rapidlya te arly stages in the reaction. The inference is that the off-metal dimerization step is likely to be very slow and thus insertion is preferred for NMe 2 BH 2 giving credence to the amidoborane insertion path proposed in Scheme 3B. Altogether,t he higher conversion, lower catalystl oading and shortert imescale found with 1tLi, compared to the current state of the art, suggests that the presenceo fD HP speciesi si mportant in the enhancement observed in these reactions.
Dehydrogenative coupling with the sodium and potassium dihydropyridyl precatalysts Next we assessed the role of alkali metal on the reaction. Sodium (1tNa) and potassium (1tK) variants were prepared via as imple and high yielding metathetical approach. [9] Employing 1tNa or 1tKi nc atalytic reactions (Table 1e ntries 3, 4) under analogous conditions used for 1tLi resulted in similar conversions in both cases. All three reactions appear to proceed via similar routes since the analogousi ntermediates are observed in each case in the 11 BNMR spectra (see Supporting Information). Notably these results comparev ery favourably with literature values( conversions to > 85 %[ NMe 2 BH 2 ] 2 with 1tNa or 1tKc ompared with approximately 43 %w ith NaN(SiMe 3 ) 2 or KN(SiMe 3 ) 2 ). [27] Reaction timescales were comparatively long (72 hf or 1tNa and 144 hf or 1tK) with respect to 1tLi (60 h), albeit considerably shorter than the reported values for NaN-(SiMe 3 ) 2 and KN(SiMe 3 ) 2 (both 172 h). Thus,i ts eems clear that the issues with modest conversion in previousN aa nd Kb ased catalysis, which was attributed to poorly soluble MÀHs pecies, has been somewhat resolved via use of "MÀHs olubilising" alkali metal alkyl-dihydropyridine precatalysts. That 1tLi outperforms the Na andKprecatalysts agreesw ith both the enhanced solubility and the trend observed previously in main group dehydrocoupling systems, [27] in which slower activity may be attributed to:i ncreasing cation radius which promotes al onger,l ooser M···HÀBc ontact and slows down hydride elimination;o rt he more dispersed charge density at the d 0 metal which affects steps involving polar insertion of unsaturated fragments or s-bond metathesis leading to product formation.
The influence of reaction solventw as also investigated using 1tLi as ar epresentative precatalyst (Table1 entries 5-8). Conductingt he reaction in [D 12 ]cyclohexane resultsi nh igh conversion (94 %) to [NMe 2 BH 2 ] 2 ,a lbeit only after heating at 75 8Cf or 168 h. This comparatively long timescale is attributed to poor solubility of the dimethylamine borane startingm aterial in cyclohexane slowing down the reaction. By moving to am ore polar reaction medium, [D 8 ]tetrahydrofuran,t he reaction slowed considerably more, only reaching ac onversion of 88 % [NMe 2 BH 2 ] 2 after 360 h. Presumably efficient stabilising Lewis base solvationo fl ithiated amidoboranes inhibits the polar insertiono fN Me 2 BH 2 into Li[NMe 2 BH 3 ]a nd/or the hydride elimination steps. Moreover,itsuggests that in this case fast catalytic turnover is reliant on the level of alkali-metals olvation.T he solvente ffect here is in contrast to that reported by Wright,[16c] where both toluene and THF gave similarr esults with LiAlH 4 as catalyst, albeit the poor solubility of LiAlH 4 in hydrocarbon solvents mayb eafactor in this report. To assess the donore ffect more thoroughly,t he reaction was repeated with ad onor solvated complex of 1tLi in [D 8 ]toluene, therebyd ifferentiating any effect from bulk donor solvent (Table 1e ntry 7). We selected previously reported chelate complex 1tLi·Me 4 AEE, [7b] where two Na nd one Od onor sites of the tridentatel igand fill three Li coordinations ites. Reactionu sing 1tLi·Me 4 AEE in toluene is faster than 1tLi in bulk THF (120 vs. 360 h) although it is still much slower than unsolvated 1tLi in toluene. Therefore it is clear that the level of solvation of the alkali metal is pivotal in this process.
Surprisingly,m ovingt ob ulk pyridine (Table 1e ntry 8) results in ar emarkable acceleration of the reaction. Even more unexpected is the ratio of products dramatically switches such that near quantitative conversionso ft he diamine borane (94 %i n 5h)t o( NMe 2 ) 2 BH rather than [NMe 2 BH 2 ] 2 are obtained (note   [b] 1tLi (2.5 %) pyridine 58 02 94 9 [b] 2 (  that since (NMe 2 ) 2 BH is the major product, as toichiometric quantity of boron remains unaccounted for by analysingt he products observed in the 11 BNMR spectrum.T he identity of the "missing" boron has not been proven, however it is unlikely to be lost as B 2 H 6 ,s ince diborane was not identified in NMR reactionm onitoring). At this point it is unclear why the presence of bulk pyridine results in such ap ronounced switch in reactivity.A nalysis of 11 BNMR data reveals the presence of Li[NMe 2 BH 2 NMe 2 BH 3 ]( II)a nd Li[NMe 2 (BH 3 ) 2 ]( V), the same intermediates observed in the catalysis conducted in [D 8 ]toluene, alongside an additional overlapping quartet resonance. Therefore, the main catalytic process may be considered to proceed via as imilar route as in toluene, except that the product formation step is b-H eliminationf rom Li[NMe 2 BH 2 NMe 2 BH 3 ]( vide supra), which can be tentatively explained by somep yridine "induced" change in charge polarisation over the intermediate, that is, coordination of pyridinet oaboron atom in the intermediate would lead to ac hange in the charge distribution across the molecule. Sicilia previously disclosed that the in silico energetics of the (NMe 2 ) 2 BH product forming steps are very high in energy for ar elatedM g II system. [26] Clearly the solvation effect of excess pyridine in some wayp romotes the hydride transfer from Li[NMe 2 BH 2 NMe 2 BH 3 ]g iving (NMe 2 ) 2 BH. An alternative explanation for preferential (NMe 2 ) 2 BH formation is that in as econdary competing process, aB H 3 group is transferred to pyridine at some stage in the process formingt he Py·BH 3 adduct, which is in line with the additional low intensity quartet present in the 11 BNMR spectrum. Ac ontrol reaction of HNMe 2 ·BH 3 in [D 5 ]pyridine at 80 8Cf or 20 h. confirms that BH 3 transfer from HNMe 2 ·BH 3 to pyridine does not occurt oa ny significant extent (ca. 15 %i sp resent at d = À11.2 ppm after prolonged heating). An alternative proposed reactions equence; accounting for the unexpected reactivity in pyridine is given in Scheme 4.
The initial deprotonation and insertion steps remain the same. However,t he intermediate Li[NMe 2 BH 2 NMe 2 BH 3 ]h as been depicted in an alternative conformation, ideally suited to transfer BH 3 to am olecule of pyridine (Scheme 4A). From here, elimination of LiH (possibly as ad ihydropyridine species), and reactionw ith the pyridine borane adduct would account for the formation of LiBH 4 (Scheme 4B). It is also important to state that the identityo ft he precatalyst in pyridine solution is likely to be differentf rom 1tLi. Reactiono ft he n-butyli somer of 1tLi with excess pyridine resultsi na1,4-dihydropyridyl bridged lithium dimer,[ py 2 Li(Àm-1,4-DHP)] 2 (2), with each Li atom solvated by two pyridine molecules (Scheme 5). [31] Therefore it is likely that the active catalytic species more closely resembles 2 than 1tLi. 2 was synthesised and tested as ap recatalyst (1.25 mol %) in [D 5 ]pyridine and in [D 8 ]toluene (entries 9 and 10). In [D 5 ]pyridine the reaction is complete in 5h ours, essentially replicating the reactivity observed using 1tLi, reinforcing the idea that in pyridine 1tLi converts to as pecies resembling 2.
In [D 8 ]toluene the catalysis is much slower.I nitiallyt he product ratio is only approximately 3:1i nf avour of IV over III,h ighlightingt he influence of pyridine in product determination (here there are two equivalents of pyridine for each LiDHP). However,a st he reaction proceeds the ratio changes to approximately 9:1a fter 146 h. Exploring the concept of solvent controlf urther we elected to employ LiAlH 4 as ac atalyst in [D 5 ]pyridine (i.e.,acatalytic amount of the usually stoichiometrically employed Lansbury's reagent). Further,W right demonstrated that LiAlH 4 is an effective catalyst in dehydrocoupling of dimethylamine borane in THFa nd toluene. Once more, the use of pyridine as reaction solvent results in high consumption of HNMe 2 ·BH 3 ,a fter 9hat 80 8C, forming III as the major product (entry 11). Together these findings outline the importance of reactions olventa nd suggest that ac ontrol of variousd ehydrocoupling reactions can be achieved with careful selection of precatalyst/solvent combinations.I nterestingly,i ne ach case where pyridine was used as ar eactions olvent, prolonged heatingo ft he reaction, after consumption of startingm aterial results in the appearance of ap artially obscured singlet resonance at about d = 26 ppm, alongside that corresponding to (NMe 2 ) 2 BH in the 11 BNMR spectra. The similarity of (NMe 2 ) 2 BH to the commonly used hydroboration reagents pinacol or catechol borane, prompted us to consider whether,o nce formed, could then III hydroborate pyridine in the presence of al ithium DHP catalyst. As toichiometricr eaction between LiAlH 4 and HNMe 2 ·BH 3 at 80 8Ci nb ulk pyridine wasc onductedt ot est this hypothesis (Scheme 6). After removal of solvent, the crude Scheme4.Alternative proposed reaction sequence accounting for the role of pyridine.
Scheme6.Synthesis of VI,formed by hydroboration with III. Chem. Eur.J.2017, 23,16853 -16861 www.chemeurj.org 2017 The Authors. Published by Wiley-VCH Verlag GmbH &Co. KGaA, Weinheim solid, identified as primarily Lansbury's reagent, was washed with hexane and the hexane washings were subsequently analysed by NMR spectroscopy.C rucially the 11 BNMR spectrum revealed the expected singlet at d = 26.4 ppm. The 1 HNMR spectrum displayed three equal intensity multipletsa t5 .96, 4.53 and 2.95 ppm, characteristic of a1 ,4 dihydropyridine species. A singlet at 2.31 ppm can be assigned as the methyl hydrogens of an NMe 2 group. The ratio of the peaks are in agreement with those of (DHP) 2 B(NMe 2 )( VI), indicating HNMe 2 has been lost from III during the reaction.
Importantly this result indicates that aD HP based catalyst is still active in pyridine after expected product formation, and further, provingt he hypothesis providedu sw ith an impetus to test 1tLi as ah ydroboration precatalyst under more controlled conditions. Finalising ouri nvestigations in [D 5 ]pyridine the reaction was repeated using precatalysts 1tNa and 1tKu nder analogousc onditions (Table 1, entries 12 and 13). In both cases conversion of HNMe 2 ·BH 3 to (NMe 2 ) 2 BH was rapid (ca > 90 %i n 8h), albeit again slower than for 1tLi, and interestingly the product resonances were clean with no presence of the hydroboration product.

Hydroboration of aldehydes and ketones
Seekingt oa chieve our aim of extending the versatility of 1tLi (the best performing precatalyst from the preceding section) in ac atalytic regime we next attempted as eries of hydroboration reactions with as election of aldehydesa nd ketones using pinacol borane (HBpin). Traditionally HBpin is employedi nh ydroboration due to its hydridic hydrogen and electrophilic boron,h owever ar ecent break-through has demonstrated it can also be employed as an easily accessed sourceo fn ucleophilic boron. [32] These hydroboration products are important intermediates in the synthesis of alcohols from aldehydes and ketones,a nd remove the necessity to use as toichiometric amount of metal reducing agent. Hill reported that DIPPnacnac-MgnBu is an excellent precatalyst for this reaction, which proceeds with low catalystl oadings, high conversionsa nd mild conditions. [19] MoreoveraMgÀHs peciesw as pinpointed as the active catalyst, involved in the first step of at wo-step process.T he first step is insertion of the unsaturated carbonyl compound into the MgÀHb ond. The second step, am etathesis with HBpin, affords hydroborated product and regenerates the activec atalyst. We have already disclosed that alkali-metal DHPs can efficiently transfer LiÀHt ob enzophenone, [7,9] ar eaction that mirrors the first step in the catalytic process since LiÀ Hf rom 1tLi adds across the C=Ob ond. Provided that the subsequent metathetical reaction with HBpin, in the presenceo f 2-tert-butylpyridine, regenerates an active 1-lithio-DHP then catalysis should proceed as described. Testing the hypothesis, benzaldehyde and HBpin were placed in aJ .Y oung's NMR tube in [D 6 ]benzene and the 1 Ha nd 11 BNMR spectra were monitored over time after addition of 5mol % 1tLi. After 15 min at room temperature the 1 Ha nd 11 BNMR spectra indicate essentially clean quantitative conversion to the hydroborated product, (Table 2, entry 1). 13 [b] 24 89 14 0.259 6 15 [b] 24 69 [a] Yield determined by formation of RR'CHOBpinr elative to internals tandard hexamethylcyclotrisiloxane. [b] Heateda t7 08C.
[c]1%c atalyst loading. Importantly the result demonstrates the versatility of Group 1D HP based precatalysts since they can effectively catalyse both dehydrocoupling and hydroboration reactions. Next we turned our attention to extending the scopeo fa ldehydes and ketones employedi nh ydroboration reactions using the same conditions. 2-Methoxybenzaldehyde, 2-naphthaldehyde and ferrocene carboxaldehyde (entries 2-4) are all cleanly converted into the corresponding protected alcohols after only 15 min at room temperature in high NMR yields (ca. 95 %) versus an internal standard. Notably the analogous reactiono f2 -methoxybenzaldehyde using DIPPnacnac-MgnBu (0.5 mol %) is complete in one hour. [19] Further,t he hydroboration of 2-naphthaldehyde is faster than that catalysed by the ruthenium complex [Ru(p-cymene)Cl 2 ] 2 (0.1 mol %, 4h), [33] albeit lower catalyst loadingsw ere used in each case. Hydroboration of 4-bromobenzaldehyde (entry 5) is also complete within 15 min, indicating at olerance to Li/halogen exchange under the reaction conditions, therebyi ncreasing the range of useful substrates able to participate in these reactions. Furthermore this reaction occurs quicker than those using either 0.05 mol %A r*N-(Si(iPr) 3 )SnOtBu, [34] in 4.5h (Ar* = (C 6 H 2 {C(H)Ph 2 } 2 iPr-2,6,4), (IPr)-CuOtBu, [35] (0.1 mol %, 1h)o r[ Ru(p-cymene)Cl 2 ] 2 (0.1 mol %, 3h), although again 1tLi has ah igher loading (5 mol %). [33] Interestingly,h ydroboration of mesitaldehyde (entry 6) takes longer for complete conversion (24 ha t7 0 8C). We attribute this to the steric hindrance of two ortho-mesityl methyl groups,w hich slows down the process, presumably by either inhibiting the hydrometallation step and/orb yp reventing efficient reformation of the putative active DHP catalyst. Moving to ketones, the hydroboration potential of 1tLi was examined with benzophenonea ss ubstrate (entry 7). Under the same conditions outlined above, clean conversion wasa chieved albeit after 30 min at room temperature. 4-Iodoacetophenone and trifluoroacetophenone (entries8and 9) both react in high yields and with short reaction times (ca. > 95 %i n1 5min). In the latter case, Jones reports Ar*N(Si(iPr) 3 )GeOtBu (2.5 mol %, 15 min) and Ar*N(Si(iPr) 3 )SnOtBu (0.5 mol %, < 15 min) precatalysts that perform the reactionw ith lower loadings or are slightly faster in the Sn case. [34] Hydroboration of 2-phenylacetophenone, 2-acetylferrocene and 2-benzoylpyridine (entries [10][11][12] are also complete in 15 minutes at room temperature, with in the third case efficient hydroboration occurring only at the carbonyl functionality.O nce more the increased sterics of am esityl substituted carbonyl (entry 13) necessitates al onger reaction (24 hours) and increased temperature (70 8C) to achieve full conversion. Dialkylketonesa re smoothly hydroborated, with 2-butanone taking1 5minutes at room temperature (entry 14). Like the aryl systems, increased stericb ulk necessitates longert imes and highert emperatures, with di-tertbutyl ketone requiring 24 ha t7 08Ct og ive almost 70 %c onversion (entry 15). To assess whether the reaction may proceed via an alternative reaction pathway to that postulated for other main group systems (vide supra) [19] as eries of controlr eactions were performed. As dihydropyridines and their parent aromatic counterparts would be present in the reaction mixture, the reactivity between HBpin and 1tLi and with pyridine (as am odel variant of 2-tert-butylpyridine) were probed. The stoichiometric reaction between HBpin and 1tLi in toluene at room temperature (Scheme 7A)r esults in complete trans-elementation giving in situ generated 1tBpin as evidenced by 1 HNMR studies (Figure 3). Here the five proton resonances from the dihydropyridyl ring 1tLi are replaced by five new dihydropyridyl resonances, consistent with replacemento fl ithium with aB pin unit and presumably generating LiH as ac oproduct.F urthermore the 11 BNMR displays as inglet resonance at d = 24.5 ppm corresponding to the newly installed BÀN bond.
Scheme7.Control reaction of A) 1tLi with HBping iving 1tBpinand B) Pyridine with HBpin giving 3. Potentially 1tBpin could act as an active catalytic entity in the hydroborationp rocess, therefore benzophenone was added to ar eaction mixture containing 1tBpin to investigate whether it would convert to hydroboratedp roduct, and the reaction was monitored by 11 BNMR spectroscopy.T he emergence of as inglet at d = 23 ppm corresponds to the hydroborated product.F or 1tBpin to act as av iable catalytic intermediate, conversion of the parent pyridine into ad ihydropyridine speciesm ust occur by some mechanism. It is long established that commercialL iH, owing to its insolubility in organic media (originating from its considerable lattice energy), on its own does not add across pyridine, indicating this pathway is unlikely,a lbeit in situ generated LiH may exhibit higherr eactivity in this regard. [31] As econd possibility is the direct addition of HBpin across the parentpyridine.
Direct reactionb etween HBpin and pyridine (Scheme 7B) suggestst hat hydroboration and concomitantd earomatisation of the pyridine does not readily occur.T his was duly confirmed with an X-ray crystallographic study,r evealing the major product as the simple donor-acceptor adduct HBpin·py (3)i na 58 %y ield. This structure represents the 'pyridine-activated HBpin'i ntermediate postulated by Wright and co-workersi n their very recently reportedboronium cation initiated hydroboration of pyridine. [36] In 3,B1i si nadistorted tetrahedral geometry [range of angles 103.4(9)-116.7 (9) BNMR studies of the decompositionp roduct revealt hat as expectedt he major resonance is that of 3,adoublet at d = 28.3 ppm accountingf or about 80 %o ft he material via integration of the boronN MR spectrum. The remainder of the material is represented by as inglet at d = 23.9 ppm indicating a minor amount of hydroborated pyridine. In agreement the 1 HNMR displays resonances potentially attributable to aD HP species, alongside the expected HBpin andp yridine resonances.
Scheme8 displays two potential routes for catalysis to proceed. PathwayA follows one commonly accepted mechanism of main group hydroboration catalysis (insertion/metathesis), [19] albeit in this case pyridine/dihydropyridine plays an active role as am etal hydride storage/release vehicle. Alternatively pathway Bd escribes ac oncerted process between 1tBpin, the car-bonyl substrate, and the in situ generated LiH, explainingb oth hydroboration and catalyst reformation. It may be significant that in pathway B, LiH is generated in as tep prior to aromatic pyridine formation. Due to the poor hydrocarbon solubility of LiH, polymeric LiH aggregates are likely to precipitate. Therefore one may expect pathway At ob et he favoured catalytic manifold since LiH is generated in the presence of the aromatic pyridine andc an therefore add across it in this regime. A second consideration in pathway Bi st hat the incipient LiH may simply associatew ith excess HBpin giving as ubstituted borohydride specieso ft he form Li[H 2 Bpin],a nd therebyr emaining solubilized. Howeverw es ee no spectroscopice vidence to support such as cenario.

Experimental Section
Full details of experimental procedures are provided in the electronic Supporting Information.