Jumping in the Chiral Pool: Asymmetric Hydroaminations with Early Metals

The application of early-metal-based catalysts featuring natural chiral pool motifs, such as amino acids, terpenes and alkaloids, in hydroamination reactions is discussed and compared to those beyond the chiral pool. In particular, alkaline (Li), alkaline earth (Mg, Ca), rare earth (Y, La, Nd, Sm, Lu), group IV (Ti, Zr, Hf) metal-, and tantalum-based catalytic systems are described, which in recent years improved considerably and have become more practical in their usability. Additional emphasis is directed towards their catalytic performance including yields and regio- as well as stereoselectivity in comparison with the group IV and V transition metals and more widely used rare earth metal-based catalysts.


Introduction
The coupling of carbon and nitrogen bonds is of great importance to organic chemistry [1,2]. The thusly formed nitrogen-containing compounds including N-heterocycles offer diverse applications not only in material sciences, but also in natural product synthesis and pharmaceutical chemistry. One synthetic concept in the mostly applied preparation of such molecules is the hydroamination reaction [3][4][5][6][7][8][9][10][11].
Hydroamination is the addition of an N-H bond of a primary or secondary amine across a carbon-carbon double or triple bond of, for example, alkenes, alkynes, dienes or allenes, resulting in an optimal atomic economy of 100% [5,6,12,13]. However, asymmetric C,N coupling processes including the Aza-Wacker [14], Buchwald-Hartwig [14], aminoacetoxylation [14] and photoredox (aminium radicals) [15] reactions are less atomically efficient than hydroaminations. Depending on the substrates used, hydroamination reactions occur either intermolecularly, in which the relevant functional groups are part of the separated starting materials, or intramolecularly, wherein the substrates combine both the amine and unsaturated C=C or C≡C building blocks in a single molecule [5].
Commonly, the intermolecular hydroamination of alkenes and alkynes results in the formation of Markovnikov and/or anti-Markovnikov regioisomers [5]. In the case of allenes and alkynes, E/Z isomers are produced [6,7,16]. Intramolecular hydroamination favors the Markovnikov product giving α-alkyl N-heterocycles for alkene substrates [5]. In addition, substituents in the β-position to the amino unit of the nitrogen-bonded unsaturated organic carbon-hydrogen substrate affect the reaction rate, which is known as Thorpe-Ingold effect [17].
Hydroamination reactions are thermodynamically neutral [5,6,18,19]. Due to the electrostatic repulsion between the nitrogen lone pair, the C,C π-system and the orbital symmetryforbidden [2 + 2] cycloaddition, hydroamination reactions possess a high reaction barrier despite being kinetically favored as caused by the increase in the total bonds. Therefore it is necessary to catalyze or run the respective reactions at a high temperature [5,6].
To the best of our knowledge, the first hydroamination in solution, the C,N coupling reaction of p-toluidine with cyclohexene, was reported by Hickinbottom in 1932 [20]. reaction barrier despite being kinetically favored as caused by the increase in the total bonds. Therefore it is necessary to catalyze or run the respective reactions at a high temperature [5,6].
To the best of our knowledge, the first hydroamination in solution, the C,N coupling reaction of p-toluidine with cyclohexene, was reported by Hickinbottom in 1932 [20]. Shortly after this, Kozlov et al., published the catalytically controlled hydroamination of an amine with an alkyne in the presence of mercury (II), copper (II) or silver (I) halides as catalysts [21,22]. In 1971, Coulson described the reaction of amines with alkenes by using catalytic active Rh and Ir species [23]. Since then, the field of C,N coupling via hydroamination has been expanding [6,12,[24][25][26][27][28][29][30]. Early transition metals of group IV and V from the periodic table of elements were introduced by Bergman and Livinghouse [31,32]. Rare earth metal-based catalysts were launched by Marks dating back to 1989 [33], and early main-group elements as catalytic systems were established at the start of the new millennium [34].
Main-group or lanthanide-element-based catalysts are generally less tolerant towards amines and alkenes featuring polar functional groups, e.g., esters, ketones and alcohols and are more sensitive towards air and moisture as compared to late transition metal complexes. However, they exhibit an overall higher reactivity, a better regioselectivity and are more ecologically friendly compared with late transition metal hydroamination catalysts. Hence, early-metal-based catalysts are the preferred catalysts over expensive and often toxic late transition metal ones, especially in intramolecular hydroamination reactions [5,6,13].
Since intermolecular alkene hydroaminations and intramolecular cyclization reactions of aminoalkenes may form stereocenters, chiral catalysts are required to obtain enantiopure isomers. For this to occur, ligands such as 1,1′-bi-2-naphthol(=BINOL), 2,2'-diamino-1,1'-binaphthaline(=DABN) or 2,2'-bis(diphenylphosphino)-1,1'-binaphthyl(=BINAP) derivatives are best suited, due to their bulk and (excellent) enantioselectivity [35,36]. In addition to these synthetic, accessible but hard to purify and difficult to up-scale biaryls, a series of enantiopure building blocks provided by nature are also of great benefit. The so called chiral pool-based ligands are readily available, ecologically friendly and hence "green" [37][38][39][40]. Due to their low cost, high abundance and general sustainability, the chiral pool has been extensively utilized by synthetic chemists in the preparation of ligand systems in enantioselective catalysis of natural products as well as pharmaceutical agents, with an extensive literature available on these topics [39,[41][42][43][44][45][46]. The most relevant chiral pool motifs in hydroamination reactions are amino acids, both proteinogenic and non-proteinogenic, alkaloids and terpenes ( Figure 1).  While originally only naturally occurring, enantiopure compounds were considered to be part of the chiral pool, modern definitions tend to include on a significant scale industrially produced, enantiomerically pure compounds, which can be obtained either by racemate cleavage, enantioselective synthesis or the derivatization of enantiomerically pure natural products [47].
Herein, we focus on asymmetric homogeneous metal-catalyzed hydroamination coupling reactions using early group IV and V transition, rare earth and main-group metals as catalysts featuring ligands originating from the chiral pool. The regio-and stereoselectivities, activities, conversions and yields towards the formation of the corresponding hydroamination products will be discussed in dependence of the metals, chiral pool motifs and the appropriate catalysis conditions.

Chiral Pool-Based Catalysts for Asymmetric Hydroamination Reactions
In the first catalytic hydroamination reaction dating back to the early nineteen thirties, group XI and XII metal halides were applied as catalysts [21,22]. Since then, a multitude of metal compounds have been researched for their suitability as catalytic active systems in inter-and intramolecular hydroamination reactions [3,5,48]. The catalysts can be differentiated into late and early transition metals, rare earth metals and early main-group elements. In the following, the application of hydroamination catalysts especially featuring ligands originating from the natural chiral pool will be discussed in detail and compared to non-chiral pool ligands.

Early Transition Metals
Early transition metals and rare earth metals have been extensively studied in intramolecular hydroamination catalysis [24]. The HSAB principle states that "hard metals" bind to "hard ligands". Therefore, early metals have been combined with ligands, such as amines, alcohols and ethers. For transition metals of group IV of the periodic table of elements and rare earth metals, cyclopentadienyls have also been proven to be excellent ligands for the catalytic active center [33,[53][54][55][56][57][58].

Rare Earth Metals
The intramolecular hydroamination of non-activated olefins using rare-earth metal complexes was pioneered by the group of Marks in the 1990s [33,54]. The majority of the catalysts featuring cyclopentadienyl entities allows the efficient generation of racemic or enantio-enriched N-heterocycles [33,54,[59][60][61][62][63]. Using these systems, mechanistic studies were undertaken and two mechanisms were proposed [64][65][66][67]. The σ-insertion mechanism suggested by Marks et al., (Figure 2a) postulates a rapid, reversible migratory olefin insertion of the metal amide followed by a slower, irreversible rate-determining metal alkyl bond protonolysis by a further substrate molecule [29,[53][54][55]61,[68][69][70]. The turnoverlimiting M-C σ-bond aminolysis occurs by a substrate molecule which is followed by a kinetically favored displacement of the N-heterocycle, as confirmed by deuterium-labelling experiments [29,54,61,68,69]. Exemplary NH/ND kinetic isotope effect (=KIE) and isotopic perturbation studies on Cp* 2 LnR (Ln = La, Nd, Sm, Y, Lu; R = H, CH(TMS) 2 , η 3 -C 3 H 5 , N(TMS) 2 ; TMS = SiMe 3 ) complexes were carried out to define the stereochemistry of the corresponding heterocycles [71]. These studies found that the NH/ND KIE cannot be derived from protonolysis of a previously formed Ln-C bond. In order to explain this finding a non-insertive catalytic cycle was proposed (Figure 2b), involving a second coordinated amine substrate, partially transferring one of its two NH protons to the terminal alkene carbon atom to form the pyrrolidine product by insertion (Figure 2b) [29,68,71]. Finally, the coordinated pyrrolidine is released by a new substrate molecule [29,68,71]. The two discussed mechanisms for intramolecular hydroamination reactions using rare earth (or main-group) metal-based catalysts compete with each other [71]. stereochemistry of the corresponding heterocycles [71]. These studies found NH/ND KIE cannot be derived from protonolysis of a previously formed Ln-C order to explain this finding a non-insertive catalytic cycle was proposed (Fi involving a second coordinated amine substrate, partially transferring one of it protons to the terminal alkene carbon atom to form the pyrrolidine product by (Figure 2b) [29,68,71]. Finally, the coordinated pyrrolidine is released by a new molecule [29,68,71]. The two discussed mechanisms for intramolecular hydroa reactions using rare earth (or main-group) metal-based catalysts compete with e [71].

Entry
Cat Substrate screening was later extended by the group of Marks et al., towards the conjugated 1,3-aminodienes 11 and 12a,b using (S)-5a,b, (S)-5h and (S)-6b as organolanthanide catalysts (Table 3) [61]. The reaction rate is higher for the aminodienes 11 and 12a,b than for the corresponding aminoalkenes 1a and 2a,b, despite increased steric hindrance of the cyclization transition state [25,61]. However, the enantioselectivity is generally lower, with the exception of the formation of N-heterocycle 14a with (S)-6b as a catalyst showing up to 71% ee (Table 3, entry 10) [25,61]. The authors also show the high stereoselectivity of the intramolecularly proceeding aminodiene hydroamination by concise synthesis of naturally occurring alkaloids (±)-pinidine and (+)-coniine from easily accessible diene substrates [25,61].
In 2003, Marks et al., published a series of C 2 -symmetric bis(oxazolinato)lanthanum complexes and discussed their use as efficient catalysts for the intramolecular hydroamination of aminoalkenes and aminodienes [75]. Two complexes out of the reported series possess L-valinol-(15a) ( Table 4, entry 1) and L-tert-leucinol-derived (15b) ( Table 4, entry 2) chiral pool ligands for the cyclization of 1b (Scheme 2) [75]. However, the observed enantioselectivities were with 6% (15a) and 39% (15b) at 25 • C lower than those for the non-chiral-pool-based systems with aryl functionalities in the α-position to the nitrogen atom, which result in up to 67% ee for substrate 1b. Generally, it can be stated that lanthanides possessing the largest ionic radii display the highest turnover frequencies and enantioselectivities in the hydroamination for these systems [75]. Table 3. Catalytic asymmetric hydroamination of 11 and 12a,b using chiral rare earth metal complexes (S)-5a,b, (S)-5h and (S)-6b a .
In 2007, Carpentier et al., reported on the successful application of the yttrium catalyst 16d (Scheme 2; Table 4, entries [26][27][28], comprising a C 2 -symmetric chiral tetradentate diamine-diamide ligand with two L-proline-derived building blocks attached to the N,N'-dimethylethylenediamine backbone, in the intramolecular hydroamination of aminoalkenes 1b,c and 2c. Despite the high activities, only ee values as high as 11% could be reached for 3b,c [78]. In addition, 16d is suited for the rac-lactide ring-opening polymerization at ambient temperatures, whereby isotactic-enriched polylactides were formed [78]. The Hultzsch group published the synthesis, chemical and physical properties of (+)-neomenthyl-functionalized cyclopentadienyl and indenyl yttrocene complexes 17a and 18 (Scheme 2, Table 4 and Table 11) [79]. The synthetic methodology to prepare 17a includes a facile arene elimination starting from [Y(o-C 6 H 4 CH 2 NMe 2 ) 3 ], while 18 was accessible by salt metathesis from the lithium species and YCl 3 . for comparison, the (−)-phenylmenthyl derivative 17b was also prepared. Complexes 17a and 18 displayed moderate to good catalytic activity in the tested asymmetric hydroamination reactions (Table 4, entries 9-1, and Table 11, entries 1 and 2), but only low to moderate enantioselectivities of up to 22% (Table 4, entry 29) for 17a and 11% ee ( Table 4, entry 31) for the sterically more hindered catalyst 18 were observed in the cyclization of 1a,b [79]. The catalytic activity and enantioselectivity of non-chiral-pool-derived 17b was comparable to 17a. Furthermore, the authors indicated that the protolytic loss of an indenyl ligand in 18 occurs at low catalyst loading ( 0.5 mol-%), when applying the sterically undemanding substrate 1a [79].
In 2011, Manna et al., introduced a highly enantioselective bis(amido)yttrium complex based on chiral cyclopentadienylbis(oxazolinyl)borates(19a,b), in which the chirality is induced by L-valinol-(19a) and L-tert-leucinol-derived(19b) moieties (Scheme 2) [56]. The catalyst 19b in the intramolecular hydroaminations of primary aminoalkenes 1c-e ( Table 4, entries [12][13][14] and aminodialkenes 20a-d (Table 11, entries 3-6) showed excellent activities and yielded the corresponding pyrrolidines with high optical purities ranging from 89% to 94% ee (Table 4, entries [34][35][36] in the synthesis of 3c-e or from 92% to 96% (Table 11, entries 3-6) for the transformations of 20a-d. The achieved values for the enantiomeric excess are comparable to those obtained for the isostructural zirconium complex (Tables 10 and 11) [56]. However, the (R)-configuration of the generated stereocenter is opposite to the pyrrolidines 3c-e formed with the yttrium analog 19b. Furthermore, the authors report on mechanistic studies, indicating that 19b reacts by concerted C-N and C-H bond formations, which is maintained by the kinetic rate law for conversion, saturation of the respective substrate under initial rate conditions, isotopic enantioselectivity disruption and kinetic isotope effects [56]. By carrying out N-H/N-D kinetic studies, Manna et al., were able to show that the stereochemistry determining step for both Y and Zr catalysts involves an N-H (or N-D) bond. They demonstrated that the (S)-diastereomeric pathway is slowed down to greater extent than the (R)-pathway for both metal centers. Based on these results, they conclude that the catalysts have similar transition states but are of opposite energetic favorability, resulting in the observed difference in stereoselectivity [56].
Chai et al., reported on a tridentate-linked amido-indenyl yttrium complex on the basis of 1,2-diaminocyclohexane, which transforms amino-olefins 1b-d and 2b-d into the corresponding N-heterocycles with ee values of up to 97% [89]. Those systems show a similar enantioselectivity as the 3,3 -bis(arylalkylsilyl)-substituted binaphtholate complexes towards aminoalkenes 1b-d and a higher enantioselectivity towards 2d, which increases the difference to the chiral-pool-derived catalyst even further.
Chai et al., reported on a tridentate-linked amido-indenyl yttrium complex on the basis of 1,2-diaminocyclohexane, which transforms amino-olefins 1b-d and 2b-d into the corresponding N-heterocycles with ee values of up to 97% [89]. Those systems show a similar enantioselectivity as the 3,3′-bis(arylalkylsilyl)-substituted binaphtholate complexes towards aminoalkenes 1b-d and a higher enantioselectivity towards 2d, which increases the difference to the chiral-pool-derived catalyst even further.
[%]  pared by a consecutive two-step synthetic procedure, whereas catalysts 23a-i and 25a-i were generated in situ. Intramolecular hydroamination of aminoallene 34b exclusively results in pyrrolidine 37 with enantiomeric excesses of 16% (max.) at 135 °C (Table 6, entries 30 and 33) with quantitative conversions. No correlation between the steric bulk of the ligands and the ee values could be identified [91]. Table 6. Catalytic asymmetric hydroamination of 34b using chiral titanium 21a-c, 22a-c, 23a-i and 25a-i and tantalum catalysts 24a-l and 26a-l a .
In contrast, the cyclization of the more sterically hindered 6-methylhepta-4,5-dienylamine 34b afforded exclusively five-membered 2-(2-methylpropenyl)pyrrolidine 37 with high conversions (Table 6). Nevertheless, the enantiomeric excesses of 37 are with a maximum of 15% ee (Table 6, entry 6) [90]. A significantly higher rate acceleration when using 21a-c and 22a-c as catalysts in comparison to the titanium complex Ti(NMe 2 ) 4 was observed. It is still an open question if either isolated or in situ-generated metal imidos, which are common for group IV catalysts, are the catalytic active species [90]. Comparative experiments with phenylglycine-derived ligands (= Phg) were carried out showing similar activities towards 34a,b as for 22a-c [90].
In 2009, Johnson et al., extended the series of aminoalcohol-based titanium catalysts 21a-c and 22a-c towards the more bulky chiral compounds 23a-i and 25a-i by replacing R = H by R = methyl, n butyl or phenyl groups [91]. The corresponding ligands were prepared by a consecutive two-step synthetic procedure, whereas catalysts 23a-i and 25a-i were generated in situ. Intramolecular hydroamination of aminoallene 34b exclusively results in pyrrolidine 37 with enantiomeric excesses of 16% (max.) at 135 • C ( Table 6, entries 30 and 33) with quantitative conversions. No correlation between the steric bulk of the ligands and the ee values could be identified [91].
The Johnson group later used the previously discussed aminoalcohols (vide supra) for the preparation of the respective tantalum complexes (catalysts 24a-l and 26a-l) [92]. In comparison with titanium complexes 21a-c, 22a-c, 23a-i and 25a-i, which are dimeric in the solid state, tantalum compounds 24a-l and 26a-l are monomeric possessing a somewhat distorted trigonal-bipyramidal structure as confirmed by single crystal X-ray structure analysis. Next to the chiral pool motifs derived from L-valine and L-phenylalanine, nonnatural D-valine and D-phenylalanine were also studied. The best results in the cyclization of aminoallene 34b to pyrrolidine 37 were obtained by catalysts containing R' = Ph as substituents (24d,h,l and 26d,h,l). Enantioselectivities ≤ 80% ee were obtained with a 5 mol-% catalyst loading (Table 6, entries 19, 23, 27, 40, 44 and 48) [92]. Generally, the tantalum derivatives show better ee values than those of the respective titanium catalysts at the cost of higher reaction times and a greater variance in conversion rates.
For the intramolecular hydroamination of aminoalkanes using chiral-pool-derived catalysts, Sadow et al., published the highly enantioselective bis(amido)zirconium complex 30b possessing a chiral cyclopentadienylbis(oxazolinyl)borate in which chirality is induced by the incorporation of L-valinol into the ligand (Scheme 3) [56,57,96]. The addition of catalytic amounts of 30b to primary aminoalkenes 1a-f, 2c,d and 38-40 yielded the  corresponding N-heterocycles 3a-f, 4c,d and 41-43 with enantiomeric excesses ranging from 31% for 2c (Table 10; entry 22) to 98% for 1d (Table 10; entries 14; 18). It was proposed that the observed reactivity and high enantioselectivity of 30b may relate to the ability of the relevant intermediate to stabilize the proposed six-center transition state [56,57]. Curiously, complex 30b and its yttrium derivative 18 (vide supra) gave pyrrolidines 3c-e and 44d with an opposite absolute configuration, despite using the same ligand system as the (R)-derivative of 30b using D-valine as chiral building block. In addition, the L-tert-leucine derivative 31 was prepared in a multiple-step synthetic procedure [56]. The catalytic performance of 31 on the cyclization of aminoalkenes 1c-e and 2c corresponds to L-valine-derived 30b, resulting in similar conversions with a maximum of 93% (Table 10; entry 36) and 29% ee (Table 10; entry 38) with generally high conversions. The existence of a kinetic rate dependence was further shown, evolving from a 1st order at a low substrate concentration to zero-order at a high concentration, which is representative of a reversible catalyst/substrate interaction preceding the N-H bond cleavage in the turnover-limiting and irreversible step of the catalytic cycle [56].  [56,57,96]. (For more details concerning catalysis data see Table  10).  [56,57,96]. (For more details concerning catalysis data see Table 10).
Exchanging zirconium in 30b by titanium (30a) or hafnium (30c), the latter two species catalyze the cyclization of amino-olefins 1b-f, 2c,d and 38-40 to result in the corresponding N-heterocycles 3b-f, 4c,d and 41-43 in enantiomeric excesses of 76-82% (30a) or 18-97% (30c) with moderate to high conversions (for more details see Table 10) [58]. This work was extended to aminodialkenes 20a-h and aminodialkynes 45a-c using 19b, 30b,c and 31 as catalysts as depicted in Scheme 4 [56][57][58]96]. Diastereomers 44a-h (Scheme 4) of five-to seven-membered N-heterocycles were obtained when aminodialkenes 20a-h were used as substrates, while in the case of aminodialkynes 45a-c, the respective imines 46a-c were produced in an enantioselective reaction in high to moderate yields. Depending on the cyclization conditions applied, diastereo-and enantioselectivities of max. 99% could be reached using zirconium catalyst 30b (Table 11, entries [13][14][15] [96]. In comparison, yttriumbased systems 17a and 18 reached lower ee values of up to 38% (Table 11, entries 1 and 2), while 19b showed similar enantioselectivities to 30b. It was found that catalytically generated stereocenters in cyclized 44a-h can be independently controlled by the catalyst's properties and reaction conditions (Table 11). At low concentrations Z-44b is favored, and at high concentrations combined with lower temperatures, E-44b (Table 11, entries 7-9) is favored. It could be further demonstrated that isotopic substitution of hydrogen by deuterium (H 2 NR/D 2 NR in 20b) significantly improved the diastereoselectivity from the ratio of 8:1 to a maximum of 43:1 and increased the optical purity to 99% ee [96]. As demonstrated for 30b, experimental studies on aminodialkene ring-closing reactions to examine the effects of the catalyst-to-substrate ratio, the absolute catalyst concentration and the absolute original substrate concentration show that the latter parameter greatly influences the stereoselectivity, whereas the absolute configuration of the α-amino stereocenter created by the C-N bond generation is not influenced by any parameters of the concerted protontriggered cyclization mechanism (Figure 3) [96]. Coordination of a primary amine changes the ring conformation in the transition state to place the cis group axial to avoid unfavorable interactions between the bulkier substituent and the cyclizing substrate resulting in the formation of the trans diastereomer. With decreasing concentrations of the primary amine, pathway B becomes more unlikely, while cycle A is more favored, resulting in the increased formation of the cis diastereomer [96]. With amine deuteration, the coordination becomes more hindered, resulting in the observed increase in the respective cis product.
was found that catalytically generated stereocenters in cyclized 44a-h can b pendently controlled by the catalyst's properties and reaction conditions (Table  low concentrations Z-44b is favored, and at high concentrations combined wit temperatures, E-44b (Table 11, entries 7-9) is favored. It could be further demo that isotopic substitution of hydrogen by deuterium (H2NR/D2NR in 20b) sign improved the diastereoselectivity from the ratio of 8:1 to a maximum of 43:1 creased the optical purity to 99% ee [96]. As demonstrated for 30b, experimental on aminodialkene ring-closing reactions to examine the effects of the catalyst-to-s ratio, the absolute catalyst concentration and the absolute original substrate co tion show that the latter parameter greatly influences the stereoselectivity, whe absolute configuration of the α-amino stereocenter created by the C-N bond gene not influenced by any parameters of the concerted proton-triggered cyclization nism (Figure 3) [96]. Coordination of a primary amine changes the ring conform the transition state to place the cis group axial to avoid unfavorable interactions b the bulkier substituent and the cyclizing substrate resulting in the formation of t diastereomer. With decreasing concentrations of the primary amine, pathway B b more unlikely, while cycle A is more favored, resulting in the increased formatio cis diastereomer [96]. With amine deuteration, the coordination becomes more h resulting in the observed increase in the respective cis product. (For more details concerning catalysis data and assignments of R and R' see Table 11). Furthermore, dibenzyl zirconium complexes 32a-g and 33a-e (Scheme 5, Table 12) have been applied in the cyclization of primary aminoalkenes 1b-e, 2d and 20a,b [98]. The chirality of the appropriate catalyst was introduced by the L-valine-(32a-g), L-proline-(33a-d) or L-pipecolic acid-derived (33e) backbone of the tridentate dianionic amino-diol ligand with variation possibilities being at the ether functionality (32a-g) or Furthermore, dibenzyl zirconium complexes 32a-g and 33a-e (Scheme 5, Table 12) have been applied in the cyclization of primary aminoalkenes 1b-e, 2d and 20a,b [98]. The chirality of the appropriate catalyst was introduced by the L-valine-(32a-g), L-proline-(33a-d) or L-pipecolic acid-derived (33e) backbone of the tridentate dianionic amino-diol ligand with variation possibilities being at the ether functionality (32a-g) or the substituents of the α-position to the alcohol functionality (33a-e) and the aromatic substituent R' in the ligand system (Scheme 5). These catalysts show satisfactory catalytic activities in the C-N bond formation of aminopentenes 1b-e and aminohexene 2b. Conversions as high as 97% and high enantiomeric excesses (32d, max. 56% ee for 3d (Table 12, entry 4); 33b, up to 94% for 3d (Table 12, entries 15 and 16)) were observed in the catalytic synthesis of five-membered pyrrolidines 3b-e and E/Z-44a,b [98]. The authors also proposed a mechanism involving a highly ordered transition state and a concerted bond formation pathway. Variations in the temperature for the hydroamination of 1d using 33b as catalyst resulted only in minor changes in conversion and ee values (Table 12). Furthermore, dibenzyl zirconium complexes 32a-g and 33a-e (Scheme 5, Table 12) have been applied in the cyclization of primary aminoalkenes 1b-e, 2d and 20a,b [98]. The chirality of the appropriate catalyst was introduced by the L-valine-(32a-g), L-proline-(33a-d) or L-pipecolic acid-derived (33e) backbone of the tridentate dianionic amino-diol ligand with variation possibilities being at the ether functionality (32a-g) or the substituents of the α-position to the alcohol functionality (33a-e) and the aromatic substituent R' in the ligand system (Scheme 5). These catalysts show satisfactory catalytic activities in the C-N bond formation of aminopentenes 1b-e and aminohexene 2b. Conversions as high as 97% and high enantiomeric excesses (32d, max. 56% ee for 3d (Table  12, entry 4); 33b, up to 94% for 3d (Table 12, entries 15 and 16)) were observed in the catalytic synthesis of five-membered pyrrolidines 3b-e and E/Z-44a,b [98]. The authors also proposed a mechanism involving a highly ordered transition state and a concerted bond formation pathway. Variations in the temperature for the hydroamination of 1d using 33b as catalyst resulted only in minor changes in conversion and ee values (Table 12).
Overall, group IV metal catalysts 30a-c and 33a-e show high quantitative conversions with enatiomeric excesses as high as 98% for aminoalkenes 1a-f, 2c,d and 38-40 and up to 99% for aminodialkenes 20a-h and aminodialkynes 45a-c. Both values are comparable for aminohexene substrates 2c,d and better for aminopentenes 1a-f, 38, 39 than those obtained by non-chiral-pool-derived catalysts which are mainly based on bisaryl-derived or salen-type ligands [81,84,99,100]. Comparisons of the hydroamination of aminoallenes 34a,b using Ti and Ta catalysts 21a-c-29a-h with those applying non-chiral-pool-derived catalytic systems, which are mainly based on bisaryl-derived ligands, are more complicated due to differences in substrate screenings [86,[101][102][103].

Early Main-Group Elements
In contrast to transition metals, which can appear in different oxidation states defining their reactivity by d-electrons, early main-group elements of group I and II are primarily characterized by mono-(alkaline) or dicationic (alkaline earth) ions depending on their outer shell s-electrons. Hence, main-group elements cannot easily switch between oxidation states, and therefore, catalysis with those metals is solely based on polar reaction mechanisms and Lewis-acid activations [104].

Alkaline Metals
Group I elements can be used as pre-catalysts both in their elementary as well as ionic form [8,[105][106][107][108]. In enantioselective hydroamination reactions, solely lithium-based catalysts have been reported (intermolecular: 47, 48, Scheme 6, Table 13; 49a; 50a-f, Table 14; intramolecular: 49a,b, Table 15) [109][110][111][112][113][114]. catalyst loadings and/or by the addition of coordinating solvents such as tetrahydrofuran. In contrast, these variations influenced the formation of the N-heterocyclic molecules more significantly. When, instead of the DABN backbone in 47, naphthyl was introduced in chiral 48, almost no enantiomeric excess and significant lower conversions in the formation of the respective N-heterocyclic compounds 3b,d was observed. No enantioselectivities were obtained by using the combination (−)-sparteine/LiN(SiMe3)2 (49a) as a pre-catalyst, albeit the observed conversion of 98% is comparable to 47 [110]. Scheme 6. Catalytic asymmetric intramolecular hydroamination of aminoalkenes 1b-d, 20a and 51 using lithium-based catalysts 47 and 48 [110]. (For more details concerning catalysis data see Table  13).  [110]. (For more details concerning catalysis data see Table 13).   In 2007, Tomioka and his group discussed the intramolecular hydroamination of aminoalkenes 53a,b at −60 °C in toluene by applying in situ-produced catalytic systems 50a-f, containing diverse chiral bisoxazoline (= BOX) ligands (Table 14) [113]. In kinetically controlled catalytic reactions, almost quantitative yields and good ee values for the formation of N-heterocycles 54a,b was observed (Table 14). Within the catalytic active system, diisopropylamine acts as coordinating and proton-donating reagent. Out of the nine studied catalysts, 50a-e contain amino acid-based chiral pool ligands, of which 50d converted substrate 53a into the corresponding six-membered tetrahydroisoquinoline 54a with 84% ee (Table 14, entry 11), while catalysts 50a,c produced five-membered isoindoline 54b with high regioselectivity and an ee of 84% (Table 14, entries 2 and 10) using 54b as substrate. In none of the cases the formation of endo-cyclized 55 was observed. On the other hand, the more rigid terpene camphor-modified catalyst 50f resulted in lower activities and ee values for the cyclization of aminopentene 53b, while for aminohexene 53a comparable results to 50d could be reached. However, both synthesized N-heterocycles 54a,b using 50f as catalyst possess the (R)-configuration instead of the (S)-products favored by 50a-e. The best catalytic performance for the hydroamination of 53b was found for 50g having the non-chiral-pool D-isoleucine-derived groups attached to the BOX ligand (91% ee). Exchanging the solvent from toluene to tetrahydrofuran for catalysis resulted in the formation of both 54a,b as the kinetic and endo-cyclized 55a,b as the thermodynamic product [113].
Catalyst 49a (vide supra) can be successfully used in the intermolecular hydroam nation of olefins 58a,b with amines 59a,b resulting in ee values of up to 14% (Table 15 entry 14) and conversions from 38-71% [109], which contrasts the earlier discussed in tramolecular hydroamination reactions showing no enantioselectivity. No enantiomeri excess was observed for 49b with (−)-α-isosparteine as ligand [109]. Outside of the chiral pool, Deschamp et al., reported a non-chiral-pool-based d amidobinaphthyl building block, allowing the variation in alkyl and methylene-ary substituents at the amino functionalities [111,112]. Addition of LiCH2SiMe3 to the re spective N,N'-disubstituted binaphthyldiamine resulted in the corresponding i Ates et al., first described the suitability of n BuLi (16 mol-%) in the catalytic high-yield synthesis of five-and six-membered N-heterocycles via the intramolecular hydroamination of non-activated aminoalkenes such as 1a-c [105]. Shortly after, Hultzsch et al., reported the first Li-catalyzed enantioselective ring-closing reaction of 2,2-substituted pent-4-en-1amines 1b-d, 20a and 51 (Scheme 6, Table 13), providing the corresponding pyrrolidine derivatives 3b-d, 44a and 52 [110]. As a catalyst, they used the dimeric, tetranuclear (S,S,S)-N,N'-dimethylpyrrolidinediamidobinaphthyl dilithium complex 47 (Scheme 6). The catalytic reactions succeeded with almost quantitative conversions and an enantiomeric excess of a max. 75% (Table 13, entry 5). The binaphthyl-centered chelate ligand in 47 is based on a DABN backbone to which two L-proline-derived moieties are attached. In the solid state, each of the four lithium ions possess different coordination environments, which exhibit a similar structure in solution [110]. It was found that only minor differences in enantiomeric excesses exist by various catalyst loadings and/or by the addition of coordinating solvents such as tetrahydrofuran. In contrast, these variations influenced the formation of the N-heterocyclic molecules more significantly. When, instead of the DABN backbone in 47, naphthyl was introduced in chiral 48, almost no enantiomeric excess and significant lower conversions in the formation of the respective N-heterocyclic compounds 3b,d was observed. No enantioselectivities were obtained by using the combination (−)-sparteine/LiN(SiMe 3 ) 2 (49a) as a pre-catalyst, albeit the observed conversion of 98% is comparable to 47 [110].
In 2007, Tomioka and his group discussed the intramolecular hydroamination of aminoalkenes 53a,b at −60 • C in toluene by applying in situ-produced catalytic systems 50a-f, containing diverse chiral bisoxazoline (= BOX) ligands (Table 14) [113]. In kinetically controlled catalytic reactions, almost quantitative yields and good ee values for the formation of N-heterocycles 54a,b was observed (Table 14). Within the catalytic active system, diisopropylamine acts as coordinating and proton-donating reagent. Out of the nine studied catalysts, 50a-e contain amino acid-based chiral pool ligands, of which 50d converted substrate 53a into the corresponding six-membered tetrahydroisoquinoline 54a with 84% ee (Table 14, entry 11), while catalysts 50a,c produced five-membered isoindoline 54b with high regioselectivity and an ee of 84% (Table 14, entries 2 and 10) using 54b as substrate. In none of the cases the formation of endo-cyclized 55 was observed. On the other hand, the more rigid terpene camphor-modified catalyst 50f resulted in lower activities and ee values for the cyclization of aminopentene 53b, while for aminohexene 53a comparable results to 50d could be reached. However, both synthesized N-heterocycles 54a,b using 50f as catalyst possess the (R)-configuration instead of the (S)-products favored by 50a-e. The best catalytic performance for the hydroamination of 53b was found for 50g having the non-chiral-pool D-isoleucine-derived groups attached to the BOX ligand (91% ee). Exchanging the solvent from toluene to tetrahydrofuran for catalysis resulted in the formation of both 54a,b as the kinetic and endo-cyclized 55a,b as the thermodynamic product [113].
Outside of the chiral pool, Deschamp et al., reported a non-chiral-pool-based diamidobinaphthyl building block, allowing the variation in alkyl and methylene-aryl substituents at the amino functionalities [111,112]. Addition of LiCH 2 SiMe 3 to the respective N,N'-disubstituted binaphthyldiamine resulted in the corresponding in situ-generated chiral lithium catalysts. Their use in the cyclization of conjugated 1,3-aminodienes 11 and 12a results in 13 and 14a with E/Z selectivities and ee values of up to 72%, while aminopentenes including 1b-d, 2c and 51 provided 3b-d, 4c and 52 with a maximum of 58% ee [111,112]. The enantioselectivities of the latter catalysts are for 1b (∆ee = −61%), 1c (∆ee = −63%) and 51 (∆ee = −15%), significantly lower, and for 1d (∆ee = 27%), higher, than for 47 [109]. To the best of our knowledge, no other chiral lithium catalysts were so far reported for the discussed substrates.
In  Table 16) [116]. In contrast to the lithium derivative 47 (vide supra), magnesium complexes (S,S,S)-61 and (R,S,S)-61 with their L-proline-derived axial chiral tetraamine ligands show moderate to high catalytic activities, but only limited enantiomeric excesses, with a maximum of 14% (Table 16, entry 6), due to the protolytic ligand exchange processes as typical for heteroleptic alkaline earth metal complexes. This solution-based phenomenon is known as the Schlenk equilibrium [117][118][119]. Within reference [116], the zinc derivatives of (S,S,S)-61 and (R,S,S)-61 were prepared. It was found that they are active hydroamination catalysts, yielding higher ee values (up to 29%) as their magnesium homologs [116]. Scheme 8. Catalytic asymmetric intramolecular hydroamination of aminoalkenes 1b-d using magnesium-based catalysts (S,S,S)-61, (R,S,S)-61 and 62a and calcium complex 62b [116,120]. (For more details concerning catalysis data see Table 16).   [116,120]. (For more details concerning catalysis data see Table 16).  In 2011, Sadow et al., described the synthesis of the magnesium complex 62a comprising a chiral, pseudo C 3 -symmetric, mono-anionic tris(oxazolinyl)borato ligand (Scheme 8) [120]. Its use in hydroamination reactions was also reported. The chirality of 62a results from three L-tert-leucine moieties. Due to the bulkiness of the respective ligand, the Schlenk equilibrium is hindered. Catalyst 62a produced good to excellent conversions in the intramolecular hydroamination of 1b-d (Table 16, entries 7-12). The enantiomeric excesses, as compared with structurally similar complexes 31 and 32 (vide supra), were with a max. of 36% ee lower [120].
In addition to 62a, the isostructural optically active calcium complex 62b was synthesized (Scheme 8) [120]. It was observed that within this species the Schlenk equilibrium is hindered in solution, as evidenced by NMR and IR studies. Catalyst 62b showed increased activities and quantitative conversions after minutes in comparison to 62a, but the stereoselectivity decreased to 18% ee for 3b,c (Table 16, entries 13 and 14) [120].
The first chiral-pool-derived (L-valine) calcium catalysts 64a-d (Scheme 9) for enantioselective hydroamination reactions of aminoalkenes 1b,d originate from the Ward group, showing for 64a,b (Table 17, entries 1-4) similar activities and conversions (>90%) when compared to 62b (Table 16, entries 13 and 14) [26]. Nevertheless, only an enantiomeric excess of 0-12% was observed for 3b,d. It should be mentioned that the para-fluorophenyl derivative 64d displayed no activity for substrates 1b,d, even after several weeks. Catalyst 64c (R = Ph) also revealed no activity when using 1b as substrate, while in the cyclization of 1d an 80% conversion occurred with 26% ee (Table 17, entry 6) [26]. This enantioselectivity signifies a notable increase in ee as compared with 64a,b (vide supra). It is also higher than the values reported for the calcium complex 62b and other non-chiral-pool-based BOX-containing calcium systems published by Buch and Harder [117,120].  [26,122,123]. (For more details concerning c talysis data see Table 17).  64a-d, 65a-c and 66a,b [26,122,123]. (For more details concerning catalysis data see Table 17). In 2011, Wixey and Ward described the use of chiral-pool-based bisimidazoline calcium complexes 65a-c in the catalytic cyclization of aminoalkenes 1b,d (Scheme 9) [122]. Like 64a-d, complexes 65a-c are derived from L-valine as a chirality inducing motif. It was shown that the ligand redistribution through the Schlenk equilibrium depends on the substituents R [122]. The measured ee values are within <12% low, however, they compare well to those for 64a-c and the complexes containing other non-chiral-pool-derived BOX ligands [117,120].
In 2012, Nixon and Ward extended the series of bisoxazoline calcium complexes by bis(oxazolinylphenyl)amines(=BOPA), of which two of the three introduced BOPA-based catalysts (66a,b) (Scheme 9) derive from the chiral pool (L-valinol, L-phenylalaninol) [123]. In the enantioselective hydroamination of 1b, quantitative conversions and ee values of up to 26% ee could be achieved (Table 17, entry 22). The conversion for aminohexene 2d was determined to be 0-83% with enantiomeric excesses as high as 16% at 80 • C (Table 17, entry 26). A major improvement in stereoselectivity (as high as 50% ee for 1d) could be reached by employing BOPA ligands based on the non-natural, non-protogenic amino acid L-α-phenylglycine [123]. This significant improvement is due to the relatively slow ligand redistribution rate. A further increase in enantioselectivity to 56% ee for substrate 1d was reported by Harder et al., using non-chiral pool BINAM derivatives as bulky dianionic ligands [28].
While the use of free alcohols as ligands is rather common for early transition metals such as titanium or tantalum, their application in alkaline-earth-metal-based catalysts is rather limited, with phenoxyamine 63 from the Hultzsch group being the most prominent one in the case of magnesium [121]. However, no system is currently used which incorporates structural motifs derived from the chiral pool. In the case of calcium, alcoholates have, up to now, not been used at all. Therefore, we expanded on this type of binding site with the synthesis of an amino acid-derived tertiary alcohol (Scheme 10). This tridentate proto-ligand is accessible from L-isoleucine via a cascade of reductive aminations followed by a Grignard reaction. The transformation of aminoalkene 1d to the respective pyrrolidine 3d in a yield of >99% with an enantiomeric excess of 67% could be obtained by in situ formation of the catalyst 67 (Scheme 10) [124]. To the best of our knowledge, this enantioselectivity is the highest observed one for calcium-based species, including non-chiral-pool-derived catalysts, which greatly illustrates the potential of such compounds in the area of intramolecular hydroamination reactions.
In 2011, Wixey and Ward described the use of chiral-pool-based bisimidazoline ca cium complexes 65a-c in the catalytic cyclization of aminoalkenes 1b,d (Scheme 9) [122 Like 64a-d, complexes 65a-c are derived from L-valine as a chirality inducing motif. was shown that the ligand redistribution through the Schlenk equilibrium depends o the substituents R [122]. The measured ee values are within <12% low, however, the compare well to those for 64a-c and the complexes containing oth non-chiral-pool-derived BOX ligands [117,120].
In 2012, Nixon and Ward extended the series of bisoxazoline calcium complexes b bis(oxazolinylphenyl)amines(=BOPA), of which two of the three introduced BOPA-base catalysts (66a,b) (Scheme 9) derive from the chiral pool (L-valinol, L-phenylalanino [123]. In the enantioselective hydroamination of 1b, quantitative conversions and ee va ues of up to 26% ee could be achieved (Table 17, entry 22). The conversion for am nohexene 2d was determined to be 0-83% with enantiomeric excesses as high as 16% 80 °C (Table 17, entry 26). A major improvement in stereoselectivity (as high as 50% ee f 1d) could be reached by employing BOPA ligands based on the non-natur non-protogenic amino acid L-α-phenylglycine [123]. This significant improvement is du to the relatively slow ligand redistribution rate. A further increase in enantioselectivity 56% ee for substrate 1d was reported by Harder et al., using non-chiral pool BINAM d rivatives as bulky dianionic ligands [28].
While the use of free alcohols as ligands is rather common for early transition meta such as titanium or tantalum, their application in alkaline-earth-metal-based catalysts rather limited, with phenoxyamine 63 from the Hultzsch group being the most prom nent one in the case of magnesium [121]. However, no system is currently used whi incorporates structural motifs derived from the chiral pool. In the case of calcium, alc holates have, up to now, not been used at all. Therefore, we expanded on this type binding site with the synthesis of an amino acid-derived tertiary alcohol (Scheme 10 This tridentate proto-ligand is accessible from L-isoleucine via a cascade of reductiv aminations followed by a Grignard reaction. The transformation of aminoalkene 1d to th respective pyrrolidine 3d in a yield of >99% with an enantiomeric excess of 67% could obtained by in situ formation of the catalyst 67 (Scheme 10) [124]. To the best of o knowledge, this enantioselectivity is the highest observed one for calcium-based specie including non-chiral-pool-derived catalysts, which greatly illustrates the potential such compounds in the area of intramolecular hydroamination reactions. Scheme 10. Catalytic asymmetric intramolecular hydroamination of aminopentene 1d using t calcium-based catalytic system 67 [124].
Another type of catalysts for the enantioselective intramolecular hydroamination based on the application of alkaline earth metals as pure Lewis-acidic metal centers. F example, non-basic calcium iodide as a Lewis acid and an external base for deprotonatio can be used [27,125]. In general, it was found that the activities of alkaline earth met iodides decrease in the series Ca > Sr >> Mg > Ba [27]. The proposed mechanism is show in Figure 4. Coordination of the amino alkene to CaI2 acidifies one of the two NH2 pr tons. Deprotonation occurs by the tBu P4 phosphazene base, followed by the cyclization the amino-olefin at the calcium metal center. After protonation of the forme N-heterocycle by [ tBu P4H] + , pyrrolidine is released [125]. As chiral catalys (−)-fenchone-based 68 and non-chiral-pool-based BINOL-modified 69 were applied (T ble 18). Catalyst 68 along with tBu P4 gave in the enantioselective hydroamination Scheme 10. Catalytic asymmetric intramolecular hydroamination of aminopentene 1d using the calcium-based catalytic system 67 [124].
Another type of catalysts for the enantioselective intramolecular hydroamination is based on the application of alkaline earth metals as pure Lewis-acidic metal centers. For example, non-basic calcium iodide as a Lewis acid and an external base for deprotonation can be used [27,125]. In general, it was found that the activities of alkaline earth metal iodides decrease in the series Ca > Sr >> Mg > Ba [27]. The proposed mechanism is shown in Figure 4. Coordination of the amino alkene to CaI 2 acidifies one of the two NH 2 protons. Deprotonation occurs by the tBu P4 phosphazene base, followed by the cyclization of the amino-olefin at the calcium metal center. After protonation of the formed N-heterocycle by [ tBu P4H] + , pyrrolidine is released [125]. As chiral catalysts, (−)-fenchone-based 68 and non-chiral-pool-based BINOL-modified 69 were applied (Table 18). Catalyst 68 along with tBu P4 gave in the enantioselective hydroamination of aminoalkenes 1b-d pyrrolidines 3b-d with almost quantitative conversions and ee values reaching 15% (Table 18, entries 3 and 4). The experimentally determined ee values are generally lower than those for (S)-69, which achieves enantioselectivities of a max. 33% (Table 18,   In general, amido or benzyl strontium and barium complexes are also active in hy droamination reactions. Their overall activity is, however, lower than that of calcium and no chiral catalyst based on the natural chiral pool have yet been reported [69,118,126,127].

Conclusions
The hydroamination reaction is an atom economical possibility for C-N bond for mation, starting from common functional groups such as an amino functionality togethe with unsaturated C,C bonds. One of the main challenges arises from the high reaction barrier, which is attributed to the strong electronic repulsion of the participating group   [27] a Reaction conditions: 10 mol-% catalyst, benzene-d 6 , Ar atm. b Determined from 1 H NMR spectroscopy. c Enantiomeric excess (= ee) determined by 1 H and/or 19 F NMR spectroscopy after derivatization with Mosher's acid. The absolute configuration was not determined for the reaction products. Figure 4. Catalytic cycle of the intramolecular hydroamination of aminoalkenes using CaI2 as catalyst and tBu P4 as external base [27].
for such transformations in case of asymmetric reaction products. Alkaline (Li), alkaline earth (Mg, Ca), rare earth (Y, La, Nd, Sm, Lu), group IV (Ti, Zr, Hf) metals, and tantalum are heavily applied in this field of research. With the rising demand for cheap and easily accessible catalysts, a promising strategy for the induction of chirality is the use of moieties obtainable from the chiral pool. In this case, the majority of ligand systems is derived from amino acids, while terpenes and alkaloids are only applied scarcely.
The resulting complexes are often of equal reactivity and selectivity than their nonchiral-pool-based, often bisaryl-derived counterparts. Therefore, they are a good alternative to established catalytic systems, with the exception of magnesium catalysts, which show significantly lower enantioselectivity compared to non-chiral-pool-derived ones, such as the phenoxyamine-based system 63. However, comparison between chiral-pool-and non-chiral-pool-derived titanium and tantalum catalysts is complicated due to the differences in substrate screening and the low amount of chiral catalytic systems found in the literature [101,102].
While a variety of different substrates, such as aminoalkenes, aminodialkenes and aminoallenes are investigated with great success, applications on higher functionalized substrates are only viewed scarcely and are often limited to a narrow number of model systems.
In summary, a range of different catalysts based on early metals is nowadays available for the application in hydroamination reactions, greatly enhanced by motifs originating from the chiral pool. Progress has been made towards high performant systems accompanied by a detailed understanding of their reaction behavior. Today, those catalysts are comparable to their more expensive heavy-transition-metal-based counterparts, often using cheaper and more accessible ligand systems. Their main limitation resides in the scarce substrate scope against which those catalysts were tested, greatly diminishing the possibilities arising from those catalysts. Therefore, upcoming challenges for early-metal-based hydroamination reactions need to shift from a pure catalyst development stage towards applications on more complex targets relevant for, e.g., pharmaceuticals or fine chemicals. By doing so, the extensive knowledge on early-metal-based catalysts can be harnessed and tailored to further enhance the toolkit in organic chemistry towards more (atom)economic and sustainable synthetic routes.
Author Contributions: All three authors contributed equally to the conceptualization, original draft preparation, as well as review and editing of the article. All authors have read and agreed to the published version of the manuscript.