Amine–Boranes as Transfer Hydrogenation and Hydrogenation Reagents: A Mechanistic Perspective

Abstract Transfer hydrogenation (TH) has historically been dominated by Meerwein–Ponndorf–Verley (MPV) reactions. However, with growing interest in amine–boranes, not least ammonia–borane (H3N⋅BH3), as potential hydrogen storage materials, these compounds have also started to emerge as an alternative reagent in TH reactions. In this Review we discuss TH chemistry using H3N⋅BH3 and their analogues (amine–boranes and metal amidoboranes) as sacrificial hydrogen donors. Three distinct pathways were considered: 1) classical TH, 2) nonclassical TH, and 3) hydrogenation. Simple experimental mechanistic probes can be employed to distinguish which pathway is operating and computational analysis can corroborate or discount mechanisms. We find that the pathway in operation can be perturbed by changing the temperature, solvent, amine–borane, or even the substrate used in the system, and subsequently assignment of the mechanism can become nontrivial.


Introduction
Hydrogenation is one of the most important and fundamental transformations used in chemistry.The direct addition of dihydrogen gas (H 2 )across an unsaturated moiety is awelldeveloped area of reduction chemistry which has resulted in H 2 as the preferred hydrogen source in many of these transformations. [1] However,t ransfer hydrogenation (TH) offers an alternative pathway that avoids the use of highly pressurized gas and potentially proffers greater control in the level of reduction. Here,as acrificial TH agent is used to donate hydrogen, whereby the TH agent is usually cheap, abundant, and easily manipulated. Early examples include Meerwein-Ponndorf-Verley (MPV) reductions using secondary alcohols as the TH agent to reduce aldehydes and ketones to their respective alcohols. [2] Furthermore,p rogression into asymmetric transfer hydrogenation (ATH) reactions,mediated by ruthenium catalysts,were pioneered in the 1980s using isopropanol or an azeotropic mixture of formic acid and triethylamine as the TH agent;t his first report has since spawned ag reat interest in this area alone. [3] Beyond isopropanol and formic acid, where the byproducts formed, acetone and CO 2 ,r espectively,a re usually trivial to separate from the reaction mixture,a dditional TH agents reported in literature includes Hantzch esters, [4] dimethylformamide with ab ase as additive, [5] sodium hypophosphite, [6] benzothiazoline, [7] and hydrazine, [8] although more commonly in the form of hydrazine hydrate. [9] More recently,a mmonia-borane (H 3 N·BH 3 )h as been studied and shown to be apromising addition to the numerous TH agents reported to date.T he low molecular weight (30.87 gmol À1 ), high hydrogen content (19.6 %wt%), and ease of handling as abench-stable crystalline solid makes H 3 N·BH 3 an attractive TH agent. Furthermore,d erivatives of H 3 N·BH 3 ,a mineboranes (R 3Àn H n N·BH n R' 3Àn )a nd metal amidoboranes (MAB) [10] have also been studied as hydrogen donors but so far are less developed in the area of TH chemistry.T he advantages of using amine-boranes over H 3 N·BH 3 can be found in their: 1) greater solubility in common analytical solvents such as benzene and chloroform, 2) ease of identification of by-products in TH reactions as some amineboranes are less likely to form insoluble polymeric substances, and 3) greater control of reduction by altering Rg roups on both the nitrogen and boron atom and by virtue of fewer hydrogen atoms available to transfer to the acceptor molecule.
In this Review we will present recent publications that use H 3 N·BH 3 and amine-borane TH agents with an emphasis on understanding the mechanism operating in these reactions. [11] Pertinently,this Review is not an evaluation of the chemistry of dehydrogenation/dehydrocoupling (DHC) of amine-boranes,f or which there are numerous reviews, [12] but instead focuses exclusively on the related tandem dehydrogenation TH process.T herefore,t wo fundamental questions present themselves in order to clarify the following discussion:What is the origin of the hydrogen atoms and how are they transferred to the substrate?W ef ind that the literature studies reviewed here can be classed as:1 )classical TH processes whereby the double hydrogen transfer comes from both the amine and borane (Sections 2and 3), 2) nonclassical TH processes whereby hydroboration from the amine-borane is followed by solvolysis (Section 4), and 3) hydrogenation via H 2 released from dehydrogenation of amine-borane (Sec-Transfer hydrogenation (TH) has historically been dominated by Meerwein-Ponndorf-Verley (MPV) reactions.However,with growing interest in amine-boranes,n ot least ammonia-borane (H 3 N·BH 3 ), as potential hydrogen storage materials,t hese compounds have also started to emerge as an alternative reagent in TH reactions.I nt his Review we discuss TH chemistry using H 3 N·BH 3 and their analogues (amine-boranes and metal amidoboranes) as sacrificial hydrogen donors.T hree distinct pathways were considered:1 )classical TH, 2) nonclassical TH, and 3) hydrogenation. Simple experimental mechanistic probes can be employed to distinguish whichpathwayi s operating and computational analysis can corroborate or discount mechanisms.W efind that the pathwayino peration can be perturbed by changing the temperature,s olvent, amine-borane,o re ven the substrate used in the system, and subsequently assignment of the mechanism can become nontrivial.
From the Contents 1. Introduction 14273 tion 5). Thel atter process is therefore not strictly aT H process as described by Wang and Astruc:" TH reaction, referring to the addition of hydrogen to am olecule from an on-H 2 hydrogen source,i saconvenient and powerful method to access various hydrogenated compounds". [13] However,w eb elieve summarizing the different reduction pathways that can occur is important for ar ound understanding of the topic. By targeting the mechanism of these reactions,t his will aid the future design of catalytic processes using H 3 N·BH 3 , amine-boranes,a nd MAB as TH agents.M oreover,i t provides as imple framework into the methodology one can apply to probe the mechanism of reduction chemistry involving amine-boranes and confirm whether classical TH, nonclassical TH, or hydrogenation mechanism is in operation.

Catalyst-Free Classical TH of Preactivated Substrates
TheTHofsubstrates without acatalyst has been achieved for molecules containing apolarized unsaturated bond. These reactions therefore are not applicable to ag reat range of substrates but still provide vital mechanistic understanding into this elementary reaction that can be informative for catalyzed processes (Section 3). In addition to lowering the activation barrier using preactivated substrates,the formation of by-products from dimerization and cyclisation of amineboranes could provide the entropic and enthalpic driving forces of the forward reactions.Inthis section we will review notable examples from the literature that have pioneered uncatalyzed TH using amine-boranes but also have an emphasis on mechanistic investigation in their studies.
In 2010, Berke and co-workers reported the TH of imines with 1-2 equiv of H 3 N·BH 3 in tetrahydrofuran (THF) to generate the corresponding amines in good to excellent yields with concomitant formation of borazine (BZ) or polyborazylene (PBZ) as the by-product (Scheme 1). [14] Thew orking hypothesis for this reaction was that the polarity match between the protic H N and hydridic H B of H 3 N·BH 3 with the polarized N dÀ =C d+ moiety of the substrate would allow for spontaneous double Ht ransfer. To probe the mechanism of this reaction and confirm this hypothesis,t he reaction temperature was kept below 60 8 8Ce nsuring that no thermal decomposition of H 3 N·BH 3 occurred and avoiding H 2 release, therefore simple hydrogenation was omitted as ar eaction pathway. [15] Additionally,h eating am ixture of H 3 N·BH 3 and D 3 N·BD 3 at 60 8 8Ci nT HF resulted in no deuterium scrambling, indicating that the adduct does not dissociate and therefore Lewis acid (BH 3 )o rb ase mediated (NH 3 )t ransfer hydrogenation could also be discounted.
Deuterium labeling experiments were performed using benzylidene aniline as the model substrate.U sing H 3 N·BD 3 , deuterium incorporation was found solely at the Ca tom of the imine moiety;c onversely,u sing D 3 N·BH 3 resulted in deuterium incorporation exclusively at the Na tom. Using D 3 N·BD 3 gave deuterium incorporation both at the Ca nd N atoms.T hese experiments corroborated the polarity match mechanism, which was found to be feasible based on quantum mechanics calculations performed. As ix-membered transition state (TS) whereby the double Htransfer of NÀH···N and the BÀH···C was found to be 23.5 kcal mol À1 more favorable than the polarity mismatch configuration. In order to decipher whether the mechanism was via ac oncerted or stepwise process,p rimary deuterium kinetic isotope effect (DKIE) experiments were undertaken, revealing:1 )inverse DKIE (0.87) when H 3 N·BD 3 was used, 2) an ormal DKIE (1.93) when D 3 N·BH 3 was used, and 3) asmall positive DKIE (1.39) when D 3 N·BD 3 was used. Hammett correlations also revealed positive values of the sensitivity constants (1)f or para-substituted benzylidene anilines (substitution at aniline side, 1 = 1.61 and substitution at benzylidene side, 1 = 0.69). All these results indicate an asynchronous concerted double Ht ransfer,w hereby the breaking of the NÀHb ond was the rate-determining step (RDS) of the transformation.
In related studies,B erke and co-workers expanded this methodology for polarized olefins (Scheme 2). [16] Ther eactions were too quick to be monitored by NMR spectroscopy in THF,s oa cetonitrile was used as slower reactivity was observed in this solvent. Deuterium labeling experiments confirmed the polarity match of the hydridic H B and protic H N transfer to the Catom with aryl/alky groups and Catom with electron-withdrawing substituents,respectively.However,the measured DKIE using 2-cyclohexylidenemalononitrile as the model substrate revealed 1) no DKIE (1.00) when H 3 N·BD 3 was used, 2) anormal DKIE (1.55) when D 3 N·BH 3 was used, and 3) an ormal DKIE (1.61) when D 3 N·BD 3 was used. This indicated that this was as tepwise process whereby the RDS involved cleavage of the N À Hbond. Further experimentation of a1:3 mixture of H 3 N·BH 3 with 2-cyclohexylidenemalononitrile at À40 8 8Cin[D 8 ]THF (and also in CD 3 CN) allowed the authors to observe the hydroboration intermediate INT1 by in situ multinuclear NMR spectroscopy,s uggesting the mechanism involved af acile hydroboration step preceding the RDS.T his mechanism was distinctly different to that observed with imines (vide supra). Continuing on from this work, Berke and co-workers were also able to effect the TH of polarized olefins using MeH 2 N·BH 3 , tBuH 2 N·BH 3 ,a nd Me 2 HN·BH 3 as well as H 3 N·BH 3 as the hydrogen donor. [17] Mechanistic investigation showed similar results of astepwise double Ht ransfer.
In 2011, Manners and co-workers reported the metal-free TH between several amine-boranes and the aminoborane iPr 2 N=BH 2 in THF at 20 8 8C( Scheme 3a). [18] Experimental and computational investigation into this reaction followed. [19] Me 2 HN·BH 3 was chosen as the model substrate in these studies.E xperimentally the reaction proceeded more cleanly than the reaction with MeH 2 N·BH 3 and H 3 N·BH 3 , with the only side-product being cyclodiborazane [Me 2 N-BH 2 ] 2 .H owever,t he reaction reached equilibrium at % 50 % conversion.
Overall, am echanism analogous to that reported by Berke for the TH of imines [14] was proposed, involving an asynchronous concerted double Ht ransfer (Scheme 3b). However,t he thermodynamic accessibility of the two reactions was vastly different. When the TH of iPr 2 N = BH 2 with MeH 2 N·BH 3 was monitored by multinuclear NMR spectroscopy at varying temperatures,t he calculated thermodynamic parameters showed the TH from Me 2 HN·BH 3 to iPr 2 N=BH 2 was endergonic (DG8 8 (295) = 10 AE 7kJmol À1 )but the dimerization of the transient [Me 2 N = BH 2 ]s pecies was more exergonic (DG8 8 (295) = À28 AE 14 kJ mol À1 )and therefore driving the reaction in the forward direction. Themeasured large entropy and small enthalpy of activation for the forward TH reaction (DS°( 295) = À210 AE 11 kJ mol À1 and DH°( 295) = 29 AE 5kJmol À1 ) were consistent with ah ighly ordered bimolecular TS, suggesting ac oncerted TS with values similar to those previously reported for Diels-Alder reactions. [20] DKIE experiments with iPr 2 N = BH 2 showed al arge positive DKIE (k H/ k D = 6.7 AE 0.9) when Me 2 DN·BH 3 was used, but as mall inverse DKIE (k H/ k D = 0.7 AE 0.1) with Me 2 HN·BD 3 and al arge positive DKIE (k H/ k D = 5.2 AE 0.8) with Me 2 DN·BD 3 .M anners and co-workers rationalized the Scheme 2. TH of polarized olefins via stepwise double Ht ransfer.
Scheme 3. a) Reversible TH of iPr 2 N=BH 2 with Me 2 HN·BH 3 and b) simplified reaction profile of the forward reaction. [18] small inverse DKIE obtained for the hydride transfer as the result of asecondary kinetic isotope effect and the change in the geometry around the boron atom at the TS.D ensity functional theory (DFT) calculations of these thermodynamic parameters gave ag ood match to the experimental values. Furthermore,DFT calculations showed that alternative pathways,s uch as stepwise addition with BÀH···B transfer first, stepwise addition with NÀH···N transfer first, or adissociative process were energetically unfeasible and did not align with the experimental evidence.
It is worth noting that this study focused on Me 2 HN·BH 3 as the TH partner.H owever,w hen RR'HN·BH 3 was used (where R = Ha nd R' = Me or H), an additional by-product was observed:[ H 2 B(m-H)(m-NRR')BH 2 ]. This would suggest that under these reaction conditions astepwise or dissociative pathway could be operating, [19] and highlights the sensitivity of these reaction pathways and how they could be perturbed by simply changing the substituents on the amine-boranes used.
More recently,B raunschweig and co-workers reported the TH of three iminoboranes with bulky Rs ubstituents (R-N B-R 1 ,w here R = tBu and R 1 = tBu, mesityl, or 2,3,5,6tetramethylphenyl) with H 3 N·BH 3 . [21] They calculated that the formation of two aminoboranes as the products (more accurately the cyclization products from H 2 N=BH 2 )w ould be thermodynamically favorable to drive the reaction forward. Placing tBu-NB-tBu under high H 2 pressure led to no hydrogenation even in the presence of Pd/C catalyst, indicative that ac lassical TH process was occurring. Although the multinuclear NMR and FTIR spectroscopic data supported the formation of tBuHN=BtBuH as the product, the cis/trans stereochemistry of this aminoborane was not clear.I solation of the products for X-ray diffraction analysis from the subsequent reaction of tBuHN = BtBuH with HCl or the Nheterocyclic carbene 1,3-diisopropylimidizol-2-ylidene (IPr) suggested the aminoborane carried trans stereochemistry.
Probing the mechanism further, deuterium labeling experiments using D 3 N·BH 3 and H 3 N·BD 3 confirmed the polarity matching of the substrates.H owever,n oD KIE experiments were reported to substantiate the DFT calculations,w hich supported ac oncerted double Ht ransfer through av ery low-energy TS (5.4 kcal mol À1 )( Scheme 4). This concerted addition would lead to the cis-aminoborane, which was 8.4 kcal mol À1 higher in energy than the transaminoborane.O bservation of the cis conformation would align with the proposed mechanism, but the trans-aminoborane as the final product was inferred experimentally (vide supra) and could indicate as tepwise pathway instead. However,arotation around the N = Bb ond to allow the isomerization from cis to trans was found through arelatively high barrier of 17.8 kcal mol À1 at room temperature.T his isomerization step would therefore be the RDS and in theory the cis-aminoborane should be observed by in situ multinuclear NMR spectroscopy prior to isomerization, but this was not reported by the authors.T his could indicate that an alternative isomerization pathway with amuch lower barrier could be involved than the one calculated and reported;lowtemperature studies could give insight.
In 2012, Chen and co-workers reported an umber of studies using lithium amidoborane (LiH 2 N·BH 3 )and calcium amidoborane (Ca(H 2 N·BH 3 ) 2 )t oc hemoselectively reduce a,b-unsaturated ketones to allylic alcohols under ambient temperature (Scheme 5a,b). [10c,22] Using MABs circumvented the use of conventional reducing agents such as NaBH 4 ,which often has poor selectivity from over reduction of the substrate,o ru sing lithium aminoborohydrides (LiR 2 N·BH 3 , where R ¼ 6 H), which requires as ubsequent hydrolysis step. Optimization of the reaction found THF to be the best solvent, as MeOH resulted in solvolysis of the MAB.Keeping the reactions at ambient temperature negated dehydrogenation of the MAB,with these processes occurring at elevated temperatures (LiH 2 N·BH 3 , % 40 8 8C; [23] Ca(H 2 N·BH 3 ) 2 , % 80 8 8C [24] ). Deuterium labeling experiments using [M n+ -(D 2 N·BH 3 ) n À ]( M= Li or Ca) showed deuterium incorporation only at the oxygen, and when [M n+ (H 2 N·BD 3 ) n À ]w as used only at the Ca tom of the carbonyl moiety,c onfirming the double Ht ransfer process and that the hydrogen came from the respective MAB.
Afurther comparison of the two different MABs revealed that Ca(H 2 N·BH 3 ) 2 was more competent at the TH of a,bunsaturated aldehydes to the allylic alcohol (Scheme 5c). When the Chen group reacted LiH 2 N·BH 3 with the substrate, they observed full conversion of the starting material but only % 50 %o ft he allylic product was formed according to 1 HNMR spectroscopy.Awhite precipitate observed in the reaction mixture was assigned to al ithium aminoarylborate species,w hich upon hydrolysis with aqueous HCl gave the desired alcohol. This lower reactivity was postulated as ar esult of poor solubility of the intermediate in THF and the potentially higher enthalpic penalty of losing as econd (Li)N À Hb ond versus (Ca)N À Hb ond. [10b,c] Chen and co-workers followed up this work by reporting the TH of ketones and imines with LiH 2 N·BH 3 ,C a-(H 2 N·BH 3 ) 2 ,a nd also sodium amidoboranes (Na(H 2 N·BH 3 ) with high conversion to secondary alcohols and amines, respectively,a cross all MABs used. [25] Higher reactivities were displayed by the MABs in comparison to H 3 N·BH 3 in these TH reactions.T his was attributed to the weaker B À H bond of the former due to am ore electron-rich Bc enter [10a] and also M···HÀBinteractions. [26] Thereaction mechanism of these TH reaction kinetic studies was probed using LiH 2 N·BH 3 with benzophenone and N-benzylideneaniline, and af irst order dependence with respect to LiAB and az eroth order dependence on the substrate was found.
Although MABs were reported to be superior TH agents than H 3 N·BH 3 in these studies,the addition of the alkali and alkaline earth metals complicates the mechanism operating in these reactions.C hen et al. reported ac omplementary DFT investigation of the TH of N-benzylideneaniline with LiH 2 N·BH 3 (Scheme 6). [25] Ac alculated pathway was found which involved the initial elimination of LiH from LiH 2 N·BH 3 ,r epresenting the RDS of the reaction at DG°= 17.2 kcal mol À1 .T his RDS agreed with the kinetic and DKIE experiment showing zeroth order dependence on the substrate and the breaking of the BÀHbond in this step.Chen et al. also attributed the small DKIE observed when LiD 2 N·BH 3 was used as ac onsequence of the small difference in energy between the two TSs, DDG°= 3.2 kcal mol À1 ,i nvolving both the N À Ha nd B À Hb ond-breaking steps.F urthermore,t hey found ah igher RDS (DG°= 28.0 kcal mol À1 )w hen H 3 N·BH 3 was used as the TH agent which matched the higher reactivity displayed by LiH 2 N·BH 3 in the reduction reactions.

Catalyzed Classical TH Reactions
In this section we highlight the chronological development of catalyzed, along with some stoichiometric, TH reactions where amine-boranes are required as hydrogen source and precatalyst activator;itisvital to comprehend that in aclassical TH the amine-borane assumes this double role, allowing the formation of an active species/catalyst and also furnishing the H + /H À critical to reduction. Classifying the following reactions as classical TH therefore makes it possible to distinguish them from nonclassical TH (Section 4) and hydrogenation reactions (Section 5). Following the aim of this Review,t he focus will be given to studies where the mechanistic investigations are detailed. It is worth noticing that some reactions cannot be classified exactly in the three main categories that we have chosen to investigate,a nd that grey areas exist with mechanistic changes occurring with varying substrates and/or reaction conditions.T herefore,w e carefully comment and propose ar ationale for these unclear points to allow ac omplete description of the topic and provide at houghtful analysis of classical TH with amineboranes.

Metal Catalysis
One of the first examples of metal-catalyzed transfer hydrogenation of olefins with amine-boranes was reported by Berke and co-workers (Scheme 7). [27] Thea uthors used 1mol %o f[Re(Br) 2 (NO)(PCy 3 ) 2 (h 2 -H 2 )] 1 to convert octene into octane with Me 2 HN·BH 3 .T he reaction allowed quantitative conversion in 1h,i rrespective of whether the reaction was carried out in an open or closed vessel, suggesting that H 2 was not responsible for the reduction. Initial stoichiometric studies hinted to the importance of transient phosphine dissociation from the metal precursor to allow oxidative addition of the amine-borane to Re I ,with an excess of phosphine found to decrease reactivity drastically. Ther eaction mechanism proceeded stepwise,w ith initial BÀ H s-bond activation and oxidative addition to form aR e III species,f ollowed by hydride insertion and Re-alkyl bond formation. After a b-hydride shift to liberate cycloborazine, the reductive elimination step ensured product formation and catalyst regeneration.
In 2013, Lin and Peters explored the reduction of olefins by Co-boryl complex 2 (Scheme 8a); [29] the formation of aC o-hydridoborane complex 4 was found when 2 was subjected to an excess of Me 2 HN·BH 3 (Scheme 8b), and the structure was confirmed by X-ray analysis and NMR spectroscopy.I tw as noted that the reduction of styrene was dramatically faster under an atmosphere of H 2 with full conversion to ethylbenzene in 1h versus the 24 hn eeded when Me 2 HN·BH 3 was used. Paul and co-workers analyzed the computational details of this reduction; [30] the authors calculated that in the presence of amine-borane,c omplex 2 can form 3 via an associative mechanism with an activation energy of 24.7 kcal mol À1 .H owever,t he generation of active catalyst 3 was found to be more energetically demanding with Me 2 HN·BH 3 than its formation in the presence of molecular H 2 (17.3 kcal mol À1 ), which confirms the results observed experimentally.Co-hydridodiborane 4,which formed only in the presence of amine-borane,was found to be off-cycle and was described as resulting from the decomposition of 3. Further experimental and computational details from Paul and co-workers showed that an excess of base NEt 3 can convert 4 back to 2.
Studies from Cazin and co-workers showed an efficient [Pd(NHC)(PR 3 )]-catalyzed TH of alkenes and alkynes with H 3 N·BH 3 (Scheme 9). [31] Thea ctive intermediate [Pd(H) 2 -(IPr)(PCy 3 )] 5 was isolated, [32] and its formation and role in the hydrogenation was computationally clarified by Yi and co-workers. [33] Intermediate 5 formed via sequential ligandassisted N-H followed by B-H activation of H 3 N·BH 3 (DG°= 23.8 kcal mol À1 ). Stepwise TH from 5,i nstead of molecular H 2 ,was found to be kinetically and thermodynamic favorable with an energy barrier of 22.3 kcal mol À1 ,t hus highlighting TH not hydrogenation is in place in this Pdcatalyzed system. Ther ole of iPrOH in the reaction mechanism was not analyzed in detail, thus the possibility of solvolysis for this TH process cannot be ruled out (Section 4.).
An ickel version of alkyne TH was performed by Garcia and Barrios-Francisco using [Ni(dppe)(h 2 -dpa)] (dpa = diphenylacetylene);t he authors showed the selective semi-TH of alkynes with H 3 N·BH 3 . [34] Theobserved stereoselective divergencytocis or trans olefin was dependent on the solvent, with THF favoring the former while MeOH furnished trans selection. It is worth highlighting that anonclassical TH might be in action when polar protic solvent is used, following our considerations on solvolysis-mediatedT Hr eactions (Section 4). No trace of H 2 was evident by GC-MS in the catalytic tests,w hich highlights that the TH mechanism might be the preferred pathway for this transformation.
Thefirst example of semi-TH of alkynes with H 3 N·BH 3 in the presence of copper was reported in 2017; [35] the authors used air-stable [Cu(IPr)(OH)] and an excess of H 3 N·BH 3 to selectively reduce alkynes to (Z)-alkenes in THF at 50 8 8C, and further expanded the substrate scope to the full reduction of propiolates.Ablank reaction under 1bar of H 2 with acatalytic amount of H 3 N·BH 3 (20 mol %) allowed only 20 %c onversion into product, which suggested ad irect hydride transfer mechanism to be in place.
Wang, Liao,a nd co-workers reported an elegant sequential dimerization/semihydrogenation reaction of alkynes into (E,Z)-1,3-dienes with Co II complex 6 and H 3 N·BH 3 (Scheme 10 a). [36] Theauthors distinguished the two sequential catalytic cycles and studied the full reaction profile both experimentally and by DFT analysis.Initial dimerization of the alkyne to a1 ,3-enyne proceeded through metal-ligand cooperative Scheme 7. a) General TH of olefins with Me 2 HN·BH 3 and with Re I precatalyst 1;b )range of Re I precatalysts and TOF results for the TH of octene after 1hreaction at 85 8 8C.
Scheme 8. a) General scheme of the Co I -catalyzed hydrogenation of styrene with Me 2 HN·BH 3 ;b)major reaction intermediates with details on the RDS for active catalyst formation (kcal mol À1 ).
Scheme 9. NHC-Pd catalyzed TH of alkenes and alkynes with H 3 N·BH 3. activation of the alkyne by 6 to form an alkylidene-Co complex 7 (Scheme 10 b);t he latter could further react through an anti-Markovnikov addition to asecond equivalent of alkyne,release the 1,3-enyne,and reform precatalyst 7.The subsequent addition of H 3 N·BH 3 to the reaction mixture allowed the pyridonate ligand mediated formation of a[ Co-H] intermediate (Scheme 10 c);t his step was proposed to proceed via borane activation by nucleophilic attack of the pyridonate ligand with ammonia release (DG°= 22.1 kcal mol À1 ), which further re-enters the cycle and attacks the new borate formed to allow CoÀHb ond formation (RDS with DG°= 23.9 kcal mol À1 ). Thel atter hydride species further reacts with the enyne for the transfer reduction;t he hydride transfer to the a-carbon versus the b-carbon atoms of the enyne differed by 0.5 kcal mol À1 ,w hich did not allow to distinguish the fate of H À incorporation. This hypothesis was further corroborated by deuterium labeling studies,w hich showed that there was no distinction between deuterium incorporation in the 1,3-diene with H 3 N·BD 3 or D 3 N·BH 3 . Final facile intramolecular protonolysis by the amino group releases the product with a cis configuration of the reduced triple bond.
Thek inetic profiles for the dehydrogenation reaction showed as econd order rate in catalyst 8 (Scheme 12), suggesting the formation of adinuclear [Co-H] active species, with catalyst deactivation observed at higher conversions;Hg drop test and P(OMe) 3 poisoning experiment did not affect conversions.However,when the strongly coordinating ligand dibenzo[a,e]cyclooctatetraene (dct) was used, the reaction slowed down and traces of dct partial hydrogenation were found, which suggested the process to be homogeneous. Moreover,a ni nduction period was observed at low catalyst loading that was found to be related to partial hydrogenation of 1,5-cyclooctadiene through poisoning experiments.T he mechanism for TH was derived from these initial findings, with an observed reaction rate similar to that of the dehydrogenation reaction, which highlights that the dinuclear [Co-H] species is the common species formed in both the TH and dehydrogenation reactions.WhenTHofa-methylstyrene was performed in aD 2 atmosphere (1.1 bar), 20 %o f deuterium incorporation into the cumene product was found, which supported aH 3 N·BH 3 -mediated reduction. In contrast, when bulky olefins were subjected to the optimized reaction conditions,1 0bar of H 2 was necessary to ensure product formation. This highlights that hydrogenation is in action with hindered substrates.
Liu et al. presented acatalyzed TH of nitriles by H 2 N·BH 3 (Scheme 12). [38] TheNNP-type cobalt pincer complexes 9 and 10 were active for achemodivergent nitrile hydrogenation to primary,s econdary,o rt ertiary amines,d epending on the solvent. When hexane was used and benzonitrile was subjected to the reaction conditions,b enzylamine was found to be the predominant product, while switching to hexafluoro-2-propanol (HFIP) led to the selective formation of secondary dibenzylamine.T he reaction worked well with N-H or Scheme 10. a) Sequential dimerization/semi-TH of alkynes catalyzed by PN-Co II complex 6;b)details of dimerization mechanism;c)Gibbs free energy diagram for the TH reaction of 1,3-enynes into 1,3-dienes.

Angewandte Chemie
Reviews N-Me complexes,w hich highlights that an outer-sphere mechanism with ligand cooperativity might not be involved in this reaction, while an inner-sphere mechanism should be considered. Moreover,t he involvement of HFIP should not be discarded, as the authors reported when studying olefin TH (Section 4.2). [39] Webster and co-workers presented arare example of Fe IIcatalyzed TH and semi-TH of olefins. [40] The b-diketiminate iron alkyl precursor 11 in the presence of sacrificial amines and borane selectively reduced alkenes and alkynes (Scheme 13 a). Thec atalytic system was found to be extremely robust when amino substituents were included in the substrate,b y eliminating the use of as acrificial external amine.L abeling experiments showed that selective anti-Markovnikov monodeuteration occurred with deuterated aniline,w hile incorporation of deuterium was found to be predominant in the internal positions (Markovnikov product) when DBpin was used (Scheme 13 b).
Gelation was observed when the TH of olefins was performed, which, together with the lack of competitive hydroboration of substrates,prompted the authors to analyze the involvement of oligomeric species in the reaction mechanism. Theformation of dimeric and tetrameric species of [nBuH 2 N·BHPin] x (x = 1: AB2; x = 2: AB4)w as proven based on the evident shift of the B-H signal versus free pinacol borane in the 11 BNMR spectrum;these species were found to be entropically favored compared to DHC adducts. Subsequent DFT calculation of the catalytic cycle highlighted the formation of an [Fe-H] active species and 1,2-alkene insertion followed by rate-limiting protonolysis with a DG°= 17.3 kcal mol À1 (Scheme 13 c). Therefore,t he formation of oligomers in this transformation was found to be crucial to decrease the concentration of free borane and amine in solution and to allow the TH process to be the preferred pathway.
Punji and Sharma studied the TH of nitriles to secondary amines using [Co(Xantphos)Cl 2 ]. [41] Thet ransformation was found to be dependent on the amine-borane used;w ith H 3 N·BH 3 ,t he selective formation of symmetrical secondary amine was observed. When Me 2 HN·BH 3 was used and the reaction was performed in the presence of as econd equivalent of amine,t he synthesis of unsymmetrical amines resulted instead.
Wang, Liao,and co-workers also described acatalytic TH of nitriles using am olybdenum-thiolate complex to synthesize primary amines in the presence of H 3 N·BH 3 (Scheme temperature (Scheme 15 b). Thel atter species 13 could catalyze the TH of nitriles but only in the presence of added H 3 N·BH 3 .Kinetic studies showed afirst order dependence in H 3 N·BH 3 and precatalyst 12,w hile the reaction was zeroth order in substrates,which ruled out nitrile activation as the RDS.Further DKIE values were determined, (k D 3 NÁBH 3 = 1.4, k H 3 NÁBD 3 = 2.6, and k D 3 NÁBD 3 = 3.3) which could suggest that both NÀHand BÀHbond activation are involved in the RDS. However,f urther computational calculations to describe the catalytic cycle indicate that, instead, Mo-H insertion into the C Nb ond with a DG°of 22 kcal mol À1 is the RDS (Scheme 14 c). Intermediate 13 was found to be the resting state of the catalytic cycle.P rotonolysis via H 3 N·BH 3 releases the product, while protonation by free NH 3 was discarded.

Metal-Free TH
In the metal-free reduction reactions with amine-boranes, most of the examples follow concerted TH pathways with the formation of as ix-membered-ring transition state analogous to the work published by Berkesgroup on the TH of imines with H 3 N·BH 3 , [14] with the exception of ar eported stepwise pathway for CO 2 reduction (vide infra).
Thedevelopment of pnictogen-catalyzed TH of azoarenes with H 3 N·BH 3 was initiated by the Radosevich group (Scheme 15 a). [43] Thea uthors reported the synthesis of the T-shaped strained P III compound 14,w hich was able to oxidatively add H 2 from H 3 N·BH 3 and form the active P V compound 16.T he latter was isolated and proposed to be the active species which enabled this catalytic transformation. However,l ater computational studies on the reaction mech-anism suggested that aP III -ligand cooperativity might be in action instead of the initially hypothesized active P III /P V redox cycle (Scheme 15 b). [44] Concerted activation of H 3 N·BH 3 with ab arrier of DG°= 27.1 kcal mol À1 allowed formation of species 15 (DG8 8 = 13.5 kcal mol À1 ), the active species for the TH of azobenzene to 1,2-diphenylhydrazine through ac oncerted six-membered-ring TS (DG°= 28.1 kcal mol À1 ). This step can therefore be denoted as the RDS of the reaction.
Dimerization of 15 resulting in formation of the P Vhydride species 16,originally isolated by Radosevichsgroup, was found to be energetically feasible (DG°= 8.9 kcal mol À1 ). However,i tw as determined that 16 was an off-cycle species for the TH of azobenzene,a nd other potential mechanistic pathways,f or example,i nsertion of the N=Nb ond into PÀH bond (43.2 kcal mol À1 )o rr eduction of azobenzene through ion-pair interaction with P À H( 29.4 kcal mol À1 ), were discarded because of the high energy requirements.
Following these findings,H irao and Kinjo reported the catalytic reduction of azoarenes with H 3 N·BH 3 and 5mol % 1,3,2-diazaphospholenes (DAPs) (Scheme 16 a). [45] This process was more efficient than Radosevichsp rocedure,w here catalysis was generally faster. Aproposed catalytic cycle was presented which starts with insertion of the P À Hb ond into N = N, followed by hydrogenolysis of the exocyclic P À Nb ond via H 3 N·BH 3 .T he latter hydrogen transfer was found to proceed through ac oncerted six-membered-ring TS with protic and hydridic hydrogen transfer to Na nd Pa toms, respectively.T his pathway was energetically feasible (DG°= 25.2 AE 4.2 kcal mol À1 , DH°= 21.8 AE 2.2 kcal mol À1 ,a nd DS°= À11.6 AE 6.8 e.u.), with oligomerization of H 3 N·BH 3 accounting for the slightly endergonic nature of the reaction. Additional stepwise pathways were analyzed but discarded as they were found to be more energetically demanding.D KIE analysis for the reduction of azobenzene to 1,2-diphenylhydrazine was carried out using isotopically labeled H 3 N·BH 3 . These experiments revealed normal KIEs with D 3 N·BH 3 (3.05), H 3 N·BD 3 (1.44), and D 3 N·BD 3 (4.67). These values demonstrated that B-H and N-H activation of H 3 N·BH 3 are involved in the RDS.WhenH 3 N·BD 3 was used, incorporation of deuterium was found to be selective for 1,3,2-diazaphospholene recovered at the end of the reaction, with traces of PH 3 formation. Therole of phosphane in the reaction was not investigated further.
Recent findings of Landaeta and co-workers on asimilar reduction reaction with an acyclicp hosphite precatalyst (Scheme 16 b) [46] substantiated that the reduction of azoarenes with pnictogenide precatalysts is prone to an associative mechanism. Thea uthors presented ad etailed mechanistic study of the latter confirming,through DKIE values following the same trend as that in Hirao and Kinjosw ork (k H 3 NÁBD 3 < k D 3 NÁBH 3 < k D 3 NÁBH 3 ), that concerted BÀHa nd NÀHb ond breaking was involved in the slowest reaction step. [47] With the rise of frustrated Lewis pair (FLP) chemistry and following the finding that H 2 is released from H 3 N·BH 3 initiated by B(C 6 H 5 ) 3 (BCF), [48] TH reactions were performed using Lewis acid activation. In one of the first examples,D u and co-workers [49] reported the activation of H 3 N·BH 3 with BCF (10 mol %) to obtain as tereoselective reduction of pyridines to piperidines (Scheme 17). Thea uthors proposed the formation of azwitterion species of type 17 resulting from hydride abstraction of H 3 N·BH 3 from the Lewis base/Lewis acid adduct. Asimilar hypothesis was proposed by Shi and coworkers in the development of N-heteroarene reduction, [50] by Xiao and co-workers for the deoxygenation of amides and lactams, [51] and by Zhong and co-workers for the reductive amination of ketones. [52] Du, Meng, and co-workers reported an asymmetric and stereoselective TH of imines using H 3 N·BH 3 (Scheme 18 a). [53] Theauthors exploited the idea of zwitterion ion pair by using chiral tert-butyl sulfonamide and Piers borane HB(C 6 H 5 ) 2 , originally finding success in the stoichiometric reaction and then transposing it into ac atalytic process (10 mol %). It is worth noting that H 2 (20 bar) was not an efficient reductant, with low conversion found compared to the reaction performed with H 3 N·BH 3 (10 %v ersus 99 %c onversion after 20 h). TheB -O isomer 18 was proposed to be the active intermediate and 11 BNMR analysis of the catalytic reaction allowed identification of the species,w hich showed ab road signal at À4.8 ppm (Scheme 18 b). Formation of this isomer, from addition of Piers borane and the sulfonamide,was also found to be the most likely by DFT calculations.F rom this active species TH occurred via the eight-membered TS1-(S), which is responsible for the enantioinductive step.Release of chiral amine and formation of dehydrated species 19 were confirmed by NMR spectroscopy,with 19 having acharacteristic 11 BNMR signal at 1.3 ppm. Stoichiometric experiments showed that 19 could quickly regenerate active catalyst 18 in the presence of H 3 N·BH 3 via an energetically viable con-certed pathway (DG°= 14.4 kcal mol À1 ). Thea uthors hypothesized the role of 19 as aB rønsted acid initiator for the reaction, although the barrier found for this process (DG°= 29.0 kcal mol À1 )w as not comparable to that of the FLP mechanism. Moreover, 20,adimer of 19,w as isolated and proven to be unable to perform catalysis and described as an off-cycle species for the TH of imines.
Further advancement of the asymmetric TH of imines and ketones with H 3 N·BH 3 was described by Du and co-workers when they used enantioenriched phosphoric acid (Scheme 19 a,b). [54] Thec hiral ammonia-borane complex 21 was isolated but, while it was found to be off-cyclef or ketone reduction (Scheme 19 b), it was proven to be an active intermediate in the TH of imines by stoichiometric studies and DFT calculations (Scheme 19 a). Interestingly for both processes,DFT calculations supported the formation of asixmembered-ring TS which accounted for substrate activation by amine-borane chiral complex 21 for imine reduction, while it involves substrate activation by phosphoric acid to form species 22,f ollowed by TH from H 3 N·BH 3 for ketone reduction.
Ar ecent development of catalysis by pnictogens was reported by Cornella and co-workers who used awell-defined Bi I compound to deliver TH of azo-and nitroarenes (Scheme 20 a,b). [55] Thea uthors suggested an elusive bismuthine(III) hydride species might form by oxidative addition to Bi I (Scheme 20 c);t he evidence for such species was given by analyzing the catalytic reaction by highresolution mass spectrometry (HRMS), which showed an adduct at 453.1738 gmol À1 assigned to ac ationic Bi III monohydride complex 24.R elease of H 2 was evident when the dehydrogenation of H 3 N·BH 3 was performed with bismuthine Scheme 17. TH of pyridine with ammonia-borane using BCF precatalyst.
Scheme 18. a) Imine TH with FLP pair using H 3 N·BH 3 as asacrificial reductant;b )detailed mechanistic studies with calculated free Gibbs energies in toluene in parenthesis.

Angewandte Chemie
Reviews 14282 www.angewandte.org species.H owever,n of urther description of the role of H 2 in the reaction was made,leaving an open question whether the reduction performed is classical TH or hydrogenation. The role of H 3 N·BH 3 was found to be crucial in the reaction with 57 %c onversion found after 16 h, while switching to amineboranes,f or example,M e 3 N·BH 3 or H 3 N·BEt 3 ,r esulted in lower conversion (10 %) or no conversion after the same reaction time.H 2 Oplayed an important, if undefined, role in the TH of azoarenes,a nd 1equiv was added to the reaction mixture in the reduction of azoarenes to decrease the reaction time (from 16 to 2h)a nd the amount of reductant (1 equiv instead of 2). Thecombination of astoichiometric amount of H 2 Oa nd H 3 N·BH 3 increased conversion to 99 %a fter 2h from 57 %w hen only 1equiv of reductant was used and to 86 %after 16 hwhen 2equiv of H 3 N·BH 3 was used. Reasonably,t he authors could not further discriminate the role of H 2 Ot hrough isotope labeling experiments,b ecause of the potential fast exchange with H 3 N·BH 3 .H owever,D KIE analysis obtained studying the initial conversion of azobenzene into 1,2-diphenylhydrazine showed al arge primary kinetic isotope effect, with k D 3 NÁBH 3 = 1.63, k H 3 NÁBD 3 = 3.94 and k D 3 NÁBD 3 = 7.05 which indicated ac oncerted TS as RDS, reminiscent of the results found for DPAs and phosphitecatalyzed TH (Scheme 16).
Ac omputational exploration of the potential of SCS-Ni pincer complexes in the TH of acetone,a cetophenone,a nd methanamine with H 3 N·BH 3 has also been reported. [56] The calculations reveal that ap roton-coupled hydride transfer is the more energetically demanding step for the reduction of ketones,w hile as tepwise hydride and proton transfer might occur in the TH of imines (Scheme 21). Thek ey to this transformation was the imidazolium substituent on the SCS ligand that acted as proton shuttle.T he reactivity of these complexes was compared to the reactivity of lactate racemate, [57] where the potential role of the metal center might be solely to stabilize the molecular entity which needs to perform the transformation, as observed by others. [58] Future experimental and mechanistic details would be of great interest to clarify and test the potential of this theoretical exploration.
Thef inal example of TH is that of CO 2 and it differs slightly from the non-metal-catalyzed TH examples reported so far. Initially reported by MØnard and Stephan in 2010 (Scheme 22 a), [59] the stoichiometric reduction of CO 2 to MeOH was performed by tris(2,4,6-trimethylphenyl)phosphine (PMes 3 )a nd an excess of AlX 3 (X = Cl, Br) FLP species;t he formation of FLP-CO 2 adducts 26 and 27 was confirmed by NMR spectroscopy and X-ray analysis,and they could be further reduced with H 3 N·BH 3 and quenched with H 2 O. Ther eduction reaction was found to be facile allowing am oderate yield (37-51 %) of MeOH after 15 min at room temperature.T he mechanism of this TH was further studied by computational methods by Paul and co-workers (Scheme 22 b). [60] Ther eduction of CO 2 to liquid fuel was found to be initiated by interaction of the hydridic B-H with the C1 atom of FLP-CO 2 and PMes 3 displacement. Thee nergy barrier to TS1 of 15.1 kcal mol À1 was in line with the mild experimental conditions.Subsequent reduction steps were calculated to be driven by B-H activation, with the exception of the hydrolysis step which allows final CÀOb ond cleavage,w hich was postulated to be facilitated by H 3 N·BH 3 or dehydrated oligomers,a sf ound by Webster and co-workers (Scheme 13, Section 3.1). Interestingly,t he authors compared this mechanism to the uncatalyzed reduction of CO 2 to formic acid, [61] for which they could locate as ix-membered-ring TS,w hich favors the hypothesis of ac oncerted mechanism.  In acreative theoretical experiment, Maeda, Sakaki, and co-workers described the use of computationally designed ONO and NNN pincer P III compounds for the activation of CO 2 (Scheme 23). [62] Theauthorscalculations were found to be in line with literature findings,with ligand-P III dehydration of H 3 N·BH 3 being the RDS (19.7 kcal mol À1 ), followed by CO 2 reduction. Thelatter proceeded via aconcerted pathway when optimizations were performed with ONO-P pincer ligands.W hen calculations were focused on NNN-P pincer ligands,astepwise coordination of formate to P-H followed by reduction to formic acid was most likely to be in action.

Supramolecular and Heterogeneous Examples
Reaction mechanisms for TH reactions with supramolecular and heterogeneous systems are somewhat less studied and the lack of mechanistic investigation does not allow discrimination between these systems as being either TH or standard hydrogenation reactions.T herefore,w ew ill highlight in this section only the reactions where the mechanism has been proven to be ac lassical TH by amine-boranes.
Initial reports of supramolecular systems of amineboranes to perform reduction reactions were published in 1984. [63] Allwood and co-workers explored the formation of as upramolecular adduct formed between substituted chiral 18-crown-6-ethers and H 3 N·BH 3 ;t he adducts were isolated and characterized by X-ray diffraction, and were found to be active in the enantioselective reduction of ketones with selectivities up to 67 % ee.
Arare example of heterogeneous TH was reported by Li and co-workers in 2015 (Scheme 24 a). [64] Thea uthors built cobalt nanoparticles on graphitic carbon nitride dyad (Co/CN or Co/g-C 3 N 4 )onamesoporous carbon nitride as the catalyst support which resulted in ah ybrid structure of amorphous shells (Co 2+ )a nd metallic core (Co 0 )a sc onfirmed by X-ray photoelectron spectroscopy (XPS). TheC o/CN material was highly active in the TH of nitroarenes,o lefins,k etones,a nd aldehydes with H 3 N·BH 3 as the hydrogen transfer agent. The TH was efficiently performed at room temperature in less than 1h.T he reaction supported small-scale application with 20 mg of catalyst per 0.5 mmol of substrate,but could also be scaled up by afactor of 10. It is important to note that when the reaction was performed in an atmosphere of H 2 (1 bar), there was no conversion after 12 h. Ap otential Co 2+ /Co 0 redox pair was postulated to be involved in the reaction mechanism (Scheme 24 b);h owever, the formation of CoH x and/or Co-amidoborane intermediates could not be excluded.
Non-supported commercially available CuO was used for the TH of nitro compounds (Scheme 25). [65] H 3 N·BH 3 was the only reducing agent capable of performing the reaction, while NaBH 4 ,h ydrazine,a cetic acid, and H 2 showed limited conversion (< 10 %). Ther eaction was run in alcoholic solvents,w ith MeOH being optimum;w hen CD 3 OD was used, no deuterium incorporation was found in the final product, which discounted the solvolysis of H 3 N·BH 3 in the reaction mixture.R educed intermediates,s uch as azoxybenzene and diazobenzene,w hich were found by 1 HNMR analysis,favored as tepwise TH mechanism. Scheme 23. Simplified reaction mechanism for the reduction of CO 2 with H 3 N·BH 3 by pincer-P III complexes.

Reviews
In ad etailed computational exploration, Ankan and coworkers described the potential fixation of N 2 onto tantalum atoms supported on asilica surface,and its further reduction using iPr 2 HN·BH 3 (Scheme 26). [66] Thea uthors analyzed the reduction in the presence of iPr 2 HN·BH 3 as the reducing agent, and noticed that the latter,i nc ontrast to H 3 N·BH 3 , does not oligomerize once dehydrogenated;t his could allow the aminoborane product iPr 2 N=BH 2 to be rehydrogenated to the amine-borane equivalent. Concerning the fixation/reduction of nitrogen on tantalum, the author proposed the formation of at antalum amido imido intermediate [( SiO) 2 Ta( = NH)(NH 2 )] (28)asthe end-product of the reaction, as found by others when analyzing the fate of N 2 reduction with molecular H 2 . [67] N 2 could approach the supported Ta atoms and be activated through ar elatively low activation barrier of 14.5 kcal mol À1 ,resulting in elongation of the NÀN bond length (1.20 )compared to free N 2 (1.09 ), indicating activation. Further stepwise proton and hydride transfer to the N À Ta bond allowed formation of ad iazenido species, which was further reduced to [( SiO) 2 TaH(NHNH)] by hydride migration from Ta to N. As econd equivalent of amine-borane could further activate species 28 and form [( SiO) 2 TaH 2 (NH 2 NH)];t he following second hydride migration from the Ta centre was predicted to be RDS with an activation energy DG°= 33.8 kcal mol À1 to form [( SiO) 2 Ta-( = NH)(NH 2 )] 28.T his event was found to be energetically favored at 13.5 kcal mol À1 compared to other models using molecular H 2 as reductant with an activation barrier of 43 kcal mol À1 . [68] Furthermore,t he author predicted that the reaction could be implemented experimentally at low temper-ature (160-170 8 8C) and suggested away to circumvent catalyst decomposition via exposing the surface to N 2 at high pressure in order to maximize fixation and subsequently allow more facile hydrogenation to occur. However, even though the results looked promising,nofurther experimental evidence to disprove the authors findings have been reported yet;this is certainly an encouragement to expand on the topic.
Recently Jiang and co-workers synthesized ac ore-shell CuPd@ZIF-8 composite,with acubic CuPd core and aMOF shell, and applied the system to the selective semi-TH of alkynes with H 3 N·BH 3 . [69] This system is notable because of the synergistic behavior of the Cu and Pd centres,w hich allows selective absorption of H 3 N·BH 3 (À1.38 eV on Cu versus À1.49 eV on Pd) and phenylacetylene (À2.43 on Pd versus À0.61 eV on Cu), respectively.The MOF shell protects the core,d ecreasing potential chemical etching of Cu nanocubes,e ven after five consecutive runs.D euterium labeling experiments allowed assessment of the role of H 3 N·BH 3 in the system;afirst order rate dependence on the reductant was observed under catalytic conditions for the reduction of phenylacetylene,w ith aK IE of 4.08 using H 3 N·BD 3 ,i ndicating that the B-H activation is rate determining.Incorporation of deuterium occurs also in the presence of D 3 N·BH 3 ,w hile solvolysis was not in action with no deuterium incorporation into styrene when MeOD was used instead. Then egligible capacity of H 2 to perform the reduction was also described and further DFT calculations on the catalytic system could define clearly that ac lassical TH reaction is in effect.

Solvolysis of Amine-Boranes in Nonclassical TH Reactions
In this section, we are concerned with the parallel/ alternative route that can occur once H 3 N·BH 3 dissociates into free NH 3 and the solvent adduct of BH 3 -that is the reduction of an unsaturated bond by an initial hydroboration step followed by protic solvent work-up.This route should be categorized as nonclassical TH as the protic hydrogen is donated from the solvent and not the amine counterpart, but importantly,n or are the hydrogens due to H 2 released from the solvolysis of H 3 N·BH 3 . [15] Similar to Section 3, the literature reviewed here,o pens up some ambiguity into the precise mechanism operating.W eh ave inferred as olvolysis pathway occurring where the authors themselves have not classified whether classical or nonclassical TH is undergoing in their systems.
Early examples of solvolysis of amine-boranes in reduction reactions have been reported by Jones using Me 3 N·BH 3 to reduce 4-tert-butylcyclohexanone in benzene followed by aqueous work up. [70] In 1971, Borsch and Levitan investigated the asymmetric reduction of ketones with optically active phenethylamine-borane with high conversion but very poor optical purity of the final product. [71] These early examples, although they did not provide ag reat deal of mechanistic insight, showed the potential of amine-boranes as reducing agents in combination with aqueous work-up or action of solvolysis.

Scheme 25. CuO-catalyzed TH of nitroarenes.
Scheme 26. Details of the N 2 splitting at Ta supported on silica with iPr 2 HN·BH 3 reducing agent;free energies are given in kcal mol À1 .

Uncatalyzed Solvolysis
Continuing on their previous work of uncatalyzed TH of polarized bonds (Section 2), and on metal catalyzed TH of olefin (Section 3.1) Berke and co-workers explored the reactivity of aldehydes and ketones with H 3 N·BH 3 (Scheme 27). [72] Rather than observing analogous TH reactions in THF they found only hydroboration of the C=O moiety to form borate esters.T he presence of free NH 3 was also observed in situ by NMR spectroscopy (d H = 0.4 ppm), suggesting an alternative mechanism was operating in contrast to those previously reported in Section 2. When benzophenone was used as the model substrate,deuterium labeling experiments showed only deuterium incorporation at the carbon position of the C = Ounit when H 3 N·BD 3 or D 3 N·BD 3 was used, confirming the "spectator" role of NH 3 (e.g. D 3 N·BH 3 led to no deuterium incorporation into the product). Furthermore,s imilar reactivities were observed with H 3 B·THF as the hydrogen source.DKIE experiments showed normal DKIEs (k D 3 NÁBH 3 /k H 3 NÁBH 3 = 1.74 and k D 3 NÁBH 3 / k D 3 NÁBD 3 = 1.10) with D 3 N·BH 3 and normal DKIEs (k H 3 NÁBD 3 / k H 3 NÁBH 3 = 1.28 and k H 3 NÁBD 3 /k D 3 NÁBD 3 = 1.49) with H 3 N·BD 3 . These experiments suggest the dissociation of H 3 N·BH 3 is the RDS and the values are indicative of asecondary KIE due to changing geometry at the Na nd Ba toms.T he H 3 B·THF species can then undergo standard hydroboration reactions with aldehydes and ketones to form borate esters.
Performing the reaction in MeOH, Berke found formation of the desired primary and secondary alcohols along with B(OMe) 3 and free NH 3 as the by-products.Itwas postulated that dissociation of H 3 N·BH 3 would be the RDS to form free NH 3 as as pectator molecule and BH 3 as the reagent. BH 3 could immediately form an adduct with the C=Om oiety of the substrate and hydroboration would form the borate ester intermediate which then undergoes methanolysis to give the products.A lternatively,aMeOH·BH 3 adduct could form after dissociation and undergo direct hydrogenation via ad ouble Ht ransfer with the protic hydrogen coming from the alcohol. [28] Deuterium labeling experiments were unable to distinguish between these two pathways.H owever,u sing MeOD confirmed that the deuterium incorporation at the O atom of the carbonyl moiety was solely from the solvent and again indicating that this reaction does not undergo aclassical TH process.
It is worth noting that in countering studies,Chen and coworkers found the formation of the primary alcohols when reacting H 3 N·BH 3 with an umber of aromatic aldehydes in THF (Scheme 28), [10b] and not formation of the borate ester. Direct comparisons with Berkeswork, [72] where both studies used the same substrates (benzaldehyde and 4-methoxybenzaldehyde) under same conditions,r evealed the different results from the two groups.F ollowing the reaction by multinuclear NMR and FTIR spectroscopy,C hen found no evidence of NH 3 formation and deuterium labeling experiments confirmed participation of both the protic and hydridic hydrogens from H 3 N·BH 3 .
Considering the divergent reactivities displayed in these two investigations in reactions of H 3 N·BH 3 with aldehydes in THF,itw ould be pertinent to probe the energetic difference between the hydroboration pathway and the classical TH pathway.A dditionally,t he energetic difference between dissociation of H 3 N·BH 3 in THF compared to that in MeOH would also provide greater insight into solvent effects. It is worth recalling that Berke and co-workers experimentally showed that no deuterium scrambling occurred when H 3 N·BH 3 was heated with D 3 N·BD 3 at 60 8 8Cfor several hours or at room temperature for several days in THF,s uggesting ahigh barrier for dissociation. [14] In 2020, Zhang,Ma, and coworkers published aD FT study on the reduction of benzaldehyde with H 3 N·BH 3 . [73] Thea uthors first examined S N 1versus S N 2-type processes for the dissociation of H 3 N·BH 3 ,for areaction involving THF,MeOH, and benzaldehyde.The S N 1 pathway was the most favorable route to the common adduct, PhCHO·BH 3 ,w ith minor differences in the energies in THF (DG°= 23.5 kcal mol À1 )a nd MeOH (DG°= 24.5 kcal mol À1 ). Furthermore,t he RDS in all the S N 2r outes was after the initial H 3 N·BH 3 dissociation and involved as econd dissociation of the BH 3 ·solvent adduct to form PhCHO·BH 3 . Therefore,only the boron counterpart, and not NH 3 ,appears to be involved, which would contradict the observed normal DKIE effects with D 3 N·BH 3 or D 3 N·BD 3 reported by Berke and co-workers. [72] From the PhCHO·BH 3 species,ahydroboration step (THF, DG°= 31.8 kcal mol À1 ;MeOH, DG°= 32.8 kcal mol À1 ) was found representing the RDS of the pathway (Scheme 30 a). In comparison, the direct TH route from H 3 N·BH 3 was found to be more kinetically favorable,w ith the RDS involving the concerted double Ht ransfer (DG°= 27.1 kcal Scheme 27. Solvolysis of H 3 N·BH 3 to effect hydroboration of aldehydes and ketones in THF (top) and hydrogenation of aldehydes and ketones in MeOH (bottom).

Scheme 28.
Products reported by Chen and co-workers [10b] versus those reported by Berke and co-workers [72] for the reduction of benzaldehyde and 4-methoxybenzaldehyde in THF with H 3 N·BH 3 . mol À1 )( Scheme 29 b). Thes mall difference of 4.7 kcal mol À1 between the TH route and hydroboration may suggest that there is some interchangeability between the two pathways depending on the product-determining step,a nd potentially could explain the difference in reactivity observed by Berkes group and Chensgroup for different aldehydes used in their investigations (Scheme 29).

Homogeneous Mediated Solvolysis
In 2016 Liu, Luo,a nd co-workers published the TH of alkynes to cis-a nd trans-alkenes selectively using PNP-and NNP-type Co-pincer complexes. [39] Controlling the steric profile around the cobalt center by altering the groups on the pincer ligands allowed them to access good chemo-and stereoselective transformation of numerous alkenes (Scheme 30). Ther ole of H 3 N·BH 3 was seemingly just as the borohydride source.C ontrol and optimization reactions confirmed:1)ACo catalyst was necessary for the conversion of the alkyne;2 )The reaction was most likely homogeneous under Hg poisoning testing;3 )Itw as important to use H 3 N·BH 3 as the boron source rather than other conventional borohydrides (NaBHEt 3 ,N aBH 3 CN,M e 2 S·BH 3 ,N aBH-(OAc) 3 ,M e 2 HN·BH 3 ); 4) Ther eaction in alcohols had the higher activity than that in THF or toluene,w ith MeOH chosen as the preferred solvent.
In order to identify the hydrogen source,d euterium labeling experiments were performed. When diphenylacetylene was used as the model substrate,C D 3 OH showed no deuterium incorporation into the product, but CD 3 OD allowed the isolation of the monodeuterated trans-1,2-diphenylethene.T he reaction of H 3 N·BH 3 with 1mol % 31 under standard conditions with and without 1equiv of diphenylacetylene always gave B(OMe) 3 as the product with formation of H 2 observed. Without any catalyst or substrate,f ormation of B(OMe) 3 in only 5%yield was observed after 16 hat508 8C, suggesting the methanolysis of H 3 N·BH 3 is acatalytic process in this system. Theg roup also demonstrated that without H 3 N·BH 3 no product was observed, so MeOH alone could not act as the hydrogen source.
Aplausible mechanism was proposed (Scheme 31) based on all the experimental evidence,s uggesting the role of H 3 N·BH 3 was to generate the active [Co-H] species,w hich hydrometalates the alkyne across the triple bond to generate an alkenyl cobalt complex. Methanolysis of the CoÀCb ond releases the cis-alkene product and forms a[ Co-OMe] complex, observed by NMR spectroscopy.R egeneration of the active [Co-H] species is enabled by H 3 N·BH 3 and after 3 turnovers can give B(OMe) 3 as the by-product. Thea uthors also propose the competitive isomerization cycle to give the trans-alkene product from the common [Co-H] species.I ti s worth noting that the use of NaBH 4 also gave successful results in the optimization reaction and would further corroborate the spectator role of the amine counterpart in H 3 N·BH 3 .However,the lack of success when using NaHBEt 3 , as tronger hydride donor, to generate the [Co-H] species is somewhat surprising given the precedence [74] and may allude to amore complex process or alternative process operating to form the active [Co-H] species.U sing D 3 N·BD 3 may help to confirm the formation of a[ Co-D] species,a dding more weight to the mechanism. Furthermore,acontrol experiment under an atmosphere of H 2 would be informative and allow further scrutiny whether hydrogenation from H 2 participates in the mechanism.
Thesolvolysis of H 3 N·BH 3 to effect reduction reviewed in this section can therefore be viewed as nonclassical TH reactions.T he role of the H 3 N·BH 3 is akin to that of borohydride reagents to reduce and activate the catalyst and to regenerate the active metal hydride complex during the cycle,with the NH 3 component not partaking in the active cycle.I nstead, proton transfer is from the protic solvent, namely MeOH, and formation of B(OMe) 3 or H 3 N·B(OMe) 3 is observed as the by-product. Deuterium labeling experiments,u sing alternative borohydride sources and using tertiary ammonia boranes (R 3 N·BH 3 ,R¼ 6 H), are simple methods in the chemistst oolbox that could be used to determine whether classical TH is taking place or whether hydride transfer and solvolysis is occurring instead.
What is interesting and less understood is the mechanism of activation of the precatalyst by H 3 N·BH 3 .T hese transition metal hydride species are often invoked based on the precedence of related hydrogenation reactions but not further scrutinized within these systems.Parallel DKIE experiments, kinetic experiments,a nd initial rates would have provided additional invaluable data towards understanding this preactivation step.I nference from transition metal mediated dehydrogenation/dehydrocoupling of H 3 N·BH 3 may be pertinent in this instance. [12e,f,15c, 75]

Heterogeneous Mediated Solvolysis
Thel iterature around the reduction of unsaturated substrates by methanolysis of amine-boranes under heterogenous conditions is scarce.T his may be due in part to the dearth of mechanistic data available in order to determine whether the reactions are simply dehydrogenation of amineboranes with molecular H 2 transferred to as urface to participate in subsequent hydrogenolysis.H owever,i n2 001 Couturier and co-workers reported the methanolysis of primary,s econdary,t ertiary,a nd aromatic amine-boranes with Pd/C and Raney Ni at room temperature in MeOH. [76] Theabsence of protic hydrogens on the tertiary and aromatic amines would confirm that the hydrogen release is due to methanolysis of the amine borane.I nf ollow-up studies they envisioned the reduction of nitroaryls using amine-boranes, provided the rate of reduction was faster than the rate of H 2 release. [77] Me 3 N·BH 3 was chosen in this study.Reaction times for the reduction of nitroaryls varied from 0.7-22 hfor room temperature reactions (Scheme 32). Ac ontrol reaction with Me 3 N·BH 3 and 10 mol %Pd(OH) 2 /C showed areaction time of 20 hf or complete methanolysis,w hich was determined by monitoring the amount of H 2 released. This provided good evidence that reduction was occurring faster than H 2 release.
In 2013, Stratakis and co-workers reported the reduction of nitroarenes and nitroalkanes into anilines and alkylhydroxylamines,r espectively,u sing H 3 N·BH 3 as the reductant and Au NPs supported on TiO 2 as the catalyst (Scheme 33). [78] Optimization reactions using p-nitrotoluene showed the reaction performed best in EtOH and H 2 Oa st he solvent with less than 5% conversion observed with polar aprotic and nonpolar solvents.Acontrol reaction without any Au NPs in EtOH showed no conversion to product. Stratakis et al. noted that the reaction was unlikely to involve H 2 gas as in related studies by Corma and co-workers,w here high temperatures (100-140 8 8C) and high pressures of H 2 (9-25 bar) were required to mediate the chemoselective reduction of nitroarenes by the same catalyst system. [79] Instead, the authors suggested involvement of Au-H species without further scrutiny of the mechanism and they were unable to identify the fate of H 3 N·BH 3 after the reaction. However, based on the solvent optimization reactions,t he effect of EtOH indicates that solvolysis pathway might be in operation, but without further experimental evidence,t his cannot be substantiated.
Following up on this study,S tratakis and co-workers expanded the scope of their reaction to report the stereoselective cis-semihydrogenation of alkynes to alkenes (Scheme 34). [80] Solvent optimization of their system found again that aprotic polar solvents and nonpolar solvents resulted in poor conversion, with EtOH again showing the best results.I nterestingly adding 5% v/v H 2 Oi nT HF improved reduction from 11 to > 99 %c onversion when compared to just using THF.I nvestigating the reductant source,t hey found using H 3 N·BH 3, Me 2 HN·BH 3 ,a nd MeH 2 N·BH 3 gave full conversion, but tBuH 2 N·BH 3 and Me 3 N·BH 3 resulted in 15 %a nd no conversion, respectively. Furthermore,using H 3 B·SMe 2 or HBpin also gave no product, which suggested the amine counterpart is important but also both the B-H and N-H are involved with the reduction. Reaction of 0.5 equiv of H 3 N·BH 3 with deuterium-labeled pmethoxyphenylacetylene to afford the stereoselective cisaddition product indicated ap otential concerted addition of the hydrogens.I nf urther studies of the reaction mechanism, 11 BNMR spectroscopy provided information on the destination of the H 3 N·BH 3 after the reaction. Analysis of the liquid phase using CD 3 OD showed ap eak at d B = 9.0 ppm, which was assigned as NH 4 B(OCD 3 ) 4 ;t his was the only additional peak observed in the 11 BNMR spectra at the end of the reaction. Based on their experimental data, they proposed involvement of Au-H species generated from insertion of the B À Hb ond from H 3 N·BH 3 .I mportantly,t he first double H transfer to the triple bond would therefore arise from N-H and Au-H moieties to release H 2 N=BH 2 as the by-product. This would explain the inadequacy of using Me 3 N·BH 3 , H 3 B·SMe 2 ,a nd HBpin in the reaction. However,H 2 N=BH 2 can quickly react with the protic solvent (ROH) to form the ammonia alkoxyborane complex, (RO)H 2 B·NH 3 ,w hich is anticipated to be more reactive than the parent H 3 N·BH 3 . This complex can then undergo an additional round of reduction, with the double Ht ransfer originating from the borane moiety of the complex and proton from the solvent, to finally give the borate salt NH 4 B(OR) 4 as the by-product in the reaction. When the reaction was performed using CD 3 OD or THF/D 2 Oand p-methoxyphenylacetylene as the substrate, there was 60-65 %d euterium incorporation on both carbon atoms of the styrene moiety,corroborating with the proposed mechanism. This study represents involvement of both classical TH and nonclassical TH (solvolysis) processes at different stages of the reaction with the choice of starting amine-borane salient to the success of the reduction reactions.T his further highlights the difficulty distinguishing the "true" mechanism in operation of these reactions with amineboranes as the reductant, as the easy interchangeability of pathways that can be undertaken by the amine-borane based on reaction conditions can cloud the mechanism.
In 2017 Fu and co-workers reported the reduction of nitrile and nitro groups to primary amines using Ni 2 PN Ps with H 3 N·BH 3 in amixed ethanol/water solvent system (1/4,v/ v) (Scheme 35 a). [81] Control reactions using H 2 (1 atm) as the hydrogen source showed no formation of product, suggesting that the dehydrogenation of H 3 N·BH 3 is not operating in this system. Thereaction mechanism was further probed by DFT calculations using 4-methoxybenzonitrile as the model substrate with Ni 2 PN Ps as the catalyst in water (Scheme 35 b). Theinitial hydrolysis of H 3 N·BH 3 mediated by Ni 2 PNPs was previously reported by the group and was shown to be exothermic to give INT1. [82] Subsequent steps involve the interaction between INT1 with another H 2 Om olecule and the substrate at the NiP 2 surface.This orientation allowed the transfer of two Hatoms from the H 2 Omolecule and the BH 3 moiety in INT1 to the C Ng roup,r espectively,t of orm benzylamine.Asecond transfer of two Ha toms from the -BH 2 (OH) moiety and another H 2 Om olecule to the C=N moiety represented the kinetic key step (3.17 eV) and resulted in the formation of benzylamine as the product.
Very recently,Glorius and co-workers reported the TH of benzene derivatives and heteroarenes using H 3 N·BH 3 mediated by [{Rh(cod)(m-Cl) 2 }]. [83] Moderate to excellent yields were achieved along with good diastereomeric ratio for numerous substrates (Scheme 36). Optimization found that the reaction performed best in fluorinated alcohols,with TFE (2,2,2-trifluoroethanol) giving the best yield and d.r. values; reactions performed in hexane,T HF and EtOH resulting in no yield of product. Theformation of ablack suspension over the course of the reaction indicated the involvement of heterogeneous complexes,w hich was supported by Hg drop test experiment resulting in no product formation. In addition, Rh nanoparticles (60-100 nm), boron clusters,a nd aluminum impurities were observed by SEM analysis of the black suspension, supporting aheterogenous mediated catalysis.P robing the reaction further,t he average deuterium incorporation into the model substrate (tert-butyldimethyl(ptolyloxy)silane) using different deuterated H(D) 3 N·BH(D) 3 and TFE indicated that the protic hydrogen was from the solvent and not the NH 3 counterpart. Furthermore,t he reaction was also successful using Me 3 N·BH 3 or HBpin as the boron source.Totest whether hydrogenation played arole in the mechanism, the reaction using the model substrate was performed under 1bar H 2 without any H 3 N·BH 3 and gave no conversion to the desired product even at 2bar H 2 .Moreover, letting [{Rh(cod)(m-Cl) 2 }] react with H 3 N·BH 3 for % 3hthen adding the substrate under 1bar H 2 resulted in only 8% conversion, suggesting that H 2 is deleterious to the reaction. Cumulatively,these experiments indicated anonclassical TH mechanism in operation mediated by Rh nanoparticles.

Hydrogenation Reactions
To conclude,i nt his section we examine examples of hydrogenation using amine-boranes.T his route differs from the classical and nonclassical TH reactions presented so far, because the real reducing agent is the H 2 released in situ. When screening the literature,w eo bserved ap aucity of hydrogenation reaction using amine-boranes in homogeneous catalytic systems (Section 3.1). [37] We rationalize this finding,a sd ifferentiating whether ah omogeneous system is undergoing classical TH or hydrogenation is not trivial. However,w ec annot be certain that there have not been examples of the use of alkene traps to monitor gas release in investigations into the DHC of amine-boranes-most literature on this area of chemistry has monitored the direct release of H 2 . [12] What has been reported is the use of cyclohexene to trap H 2 N=BH 2 ,w ith the formation of Cy 2 BNH 2 as the product but no mention of the formation of cyclohexane. [84] Experimental control reactions can help elucidate whether TH or hydrogenation is occurring.P rimarily if no reduction occurs in ahomogeneous system when the reaction is performed with H 2 instead of amine-boranes-this indicates classical TH. If reduction is observed but at ad ifferent rate to that observed using amine-borane,then it would also indicate aclassical TH process.However,ifreduction occurs in the system with H 2 at the same rate as that using amineboranes,then the identity of the mechanism is ambiguous and computational insight could be helpful.
Thek ey question in Section 3i sw hether the amineboranesrole is specific to forming the active catalytic species to mediate the reduction process as well as providing the hydrogen source?T he complication in answering this question is that acommon catalytic species is often associated with both classical TH and hydrogenation pathways.H owever,i f the direct release of H 2 from the amine-borane results in the formation of the active species,then the role of amine-borane is no different to just using H 2 in the reaction and therefore we classify this as standard hydrogenation.
In contrast, we find that ap lethora of examples using heterogeneous catalysts have been reported, [85] and we highlight only those which present productive mechanistic studies for the understanding of the reaction.

Heterogeneous Hydrogenation Reactions with Amine-Boranes
An otable example of ah ydrogenation reaction performed with amine-boranes was reported by Manners and coworkers,w ho analyzed the formation of catalytically active Rh colloids when reacting [{Rh(cod)(m-Cl) 2 }] with H 3 N·BH 3 . [86] Then ew system was able to dehydrogenate H 3 N·BH 3 and sequentially hydrogenate cyclohexene with molecular H 2 in ac losed vessel. When the reaction was performed in an open vessel, no alkene reduction was observed, clearly demonstrating the direct hydrogen addition was taking place in this transformation. Further studies from the same research group on heterogeneous hydrogenation showed that the air-stable Rh/Al 2 O 3 system in the presence of Me 2 HN·BH 3 could perform the reduction of alkenes without external H 2 . [87] However, these reactions were still performed in closed vessels and no further evidence of indirect hydrogen transfer was furnished.
Nanoparticulate systems (NPs) based on different metals, alloys,a nd sizes have been developed and tested in the catalytic reduction of nitriles and nitroarenes,f or example, Pd@MIL-101, Pd NPs enclosed in amesoporous MOF, [88] and g-Cu 36 Ni 64 ,CuNi NPs grafted on graphite. [89] When H 3 N·BH 3 was simply replaced with H 2 ,c omparable yields could be found, indicating ac lear involvement of the gaseous source. Moreover,t ests conducted with open vessels gave lower conversion (< 20 %) than experiments with higher pressurized closed vessels,a lso highlighting that gas evolution and solubility is paramount for the reduction to occur.
Xu and co-workers elegantly described at andem dehydrogenation/hydrogenation of alkenes by H 3 N·BH 3 using Pickering emulsions, [90] which are emulsions stabilized by solid particles instead of surfactants.T he authors chose Pd NPs coated onto g-C 3 N 4 and carefully analyzed the efficiency and behavior of these microreactors (Figure 1). Importantly, when H 3 N·BH 3 was replaced with gaseous H 2 ,l ow reactivity was observed mainly due to mass transfer effects from the gas to the liquid phase.M oreover,l imited emulsification could decrease the reaction efficiency and stirring was found to be Scheme 36. Nonclassical TH of benzene derivatives and heteroarenes mediated by heterogeneous Rh complexes. beneficial to increase interface area (H 2 interaction with Pd NPs). These control reactions showed that the Pickering emulsions also function as transient H 2 storage materials,with potential chemisorbed gas onto Pd and/or formation of gas microbubbles.M ore importantly,t hese results highlight that simple interchange between H 3 N·BH 3 and H 2 might not be sufficient to define the nature of the reduction mechanism, while gas solubility and mass transfer effects need to be considered and tested.

Summary and Outlook
In this Review we have examined the use of amineboranes as TH agents.W eh ave carefully analyzed and classified the reduction reactions following three major mechanistic pathways;1 )classical TH, 2) nonclassical TH or solvolysis,a nd (3)hydrogenation with amine-boranes.I n each of these contexts we have defined the role of the amineborane species as the reducing agent and/or precatalyst activator.W eh ave further highlighted the major types of amine-borane derivatives,s uch as H 3 N·BH 3 ,m ore elaborate amine-boranes,and MAB,and categorized these according to their role in reduction reactions.
We have examined the experimental and theoretical characterization techniques which have been used to allow such am echanistic portrayal. Thel eading actor is as imple "H 2 test", which allows the initial classification of TH versus hydrogenation. However,t his experiment is underrated, and it is apparent that this simple test is not used routinely. Moreover,i ti sc rucial to study the role of the solvent to identify nonclassical TH, which is still ill-defined. Thelack of differentiation between solvent-mediated protonolysis and amine-mediated protonolysis thereby places some reactions along the continuum between classical and solvolysis-mediated TH. Thei ndisputable leading techniques are isotope labeling experiments and kinetic analysis for uncatalyzed and catalyzed TH reactions that can allow differentiation between stepwise versus concerted routes.
Each class of TH presented carries its own merits.There is not asingle best route but aplethora of different options that can be selected in order to meet the usersr equirements.A s described, classical TH reactions have proven to be an efficient method to access selectively deuterate substrates with relatively cheap and easy-to-handle amine-boranes. Solvolysis TH reactions further allow the introduction of benign and green solvents,w hilst classical hydrogenation reactions using amine-boranes allow the use of avery precise quantity of an easy-to-handle "drop-in" source of H 2 gas and are still the best way to reduce challenging substrates,f or example,h indered alkenes.A ss tated, the area of amineborane DHC is ab uoyant one,b ut there are virtually no papers which clearly set out to undertake the type of consecutive dual catalysis that is necessary to dehydrogenate an amine-borane then use the H 2 released to reduce an unsaturated bond, particularly in the field of homogeneous catalysis.T his presents au nique opportunity:W ith aw elldefined catalyst that can undertake these dual roles and with full mechanistic understanding, researchers could be in as trong position to develop further divergent, asynchronous reactions.
We have highlighted that the stepwise transfer hydrogenation mechanism is kinetically and thermodynamically more favorable than dehydrogenation of amine-boranes when homogeneous systems are used. Vice versa, ap redominant hydrogenation mechanism is in place when reduction reactions are performed with heterogeneous catalysts.I t would be useful to be able to pinpoint why these differences exist:D oes it depend on the rate of dehydrogenation versus the rate of stepwise B-H/N-H activation?D oes this effect depend on the substrate affinity of the catalyst (e.g.b etter substrate affinity for styrene) versus amine-borane dehydrogenated side-products?I nh ighlighting this,w eh ope to fuel discussion and research into this area.
As discussed, most TH substrates are azoarenes and unsaturated hydrocarbons;s trikingly few examples of smallmolecule activation have been reported. Rauchfuss et al. [91] have undertaken ah ighly relevant study on the reduction of O 2 to H 2 O, ak ey transformation in fuel cell research.
Although not the focus of their study,t he authors do report on the ability of H 3 N·BH 3 ,along with other hydrogen donors, to undertake the reaction in the presence of their Ir catalyst. Beyond this,itappears that the literature is very limited, with only one example of CO 2 reduction reported by Stephan (Section 3.2) along with atheoretical study of N 2 reduction on aT as urface from Paul. [66] It is evident that the TH of small molecules is an area of unmet need:can small molecules such as N 2 ,NO 2 ,and N 2 Oundergo activation and reduction using an amine borane in al aboratory setting and what is the mechanism of these processes?C an we expand on the chemistry of TH of CO 2 and O 2 ,d evelop new catalysts,a nd obtain deeper mechanistic understanding?
Another area that is underdeveloped is that of enzymatic TH. Many of the examples reported are deracemizations in the presence of H 3 N·BH 3 and an oxidase. [92] However,itcould be argued that the substrates presented are more challenging, or at least more complex, than those tackled using transition metal catalysis.A lthough mechanistic elucidation is likely to be demanding,ac ertain level of detail is needed to clearly understand the precise nature of the bond-breaking and bond-making process in biological media;the latter aspect is ahighly attractive feature of enzymatic chemistry,harmonizing with the benign nature of aT Ha gent such as H 3 N·BH 3 .
Finally,a lthough we have covered several different amine-boranes and their role in TH, great opportunities must exist beyond these standard hydrogen sources.I ndeed, elegant studies into the hydrogen-release properties of amino complex borane encapsulated in metal organic frameworks (MOFs), [93] ethyldiaminoboranes (EDABs), [94] and metal ethyldiaminoboranes (MEDABs) [95] and theoretical studies into boron nitride nanotubes (BNNTs) [96] have been undertaken, but detailed synthetic investigations into their ability to reduce classes of substrate are lacking. Although the TH reagents may not be suitable over ab road spectrum of substrates,they do present an opportunity to probe the extent of reactivity and functional group tolerance,which may allow key or unique targets to be met. As an example,the structures of these unusual amine boranes may lend themselves well to selective reduction of targeted sets of double bonds in am ultiply double-bonded system, for example,i nt erpene feedstocks.
In conclusion, TH presents great opportunities in mechanistic studies,organic synthesis,and catalyst design. We have identified aplethora of variables associated with TH reactions from:1 )complex substrates through to traditionally inert small molecules,2 )catalyst design ranging from homogeneous,h eterogeneous,e nzymatic,o re ven catalyst-free transformations,3 )the variability of the TH reagent itself,w ith many possible reagents still to undergo comprehensive testing and opportunities for regio-and stereoselectivity,a nd 4) the nuanced means by which TH takes place that fall into three broad categories.A ll factors combined indicate that amineborane-mediated TH is an area set for growth and likely to be arich topic of research for years to come.