Aminopotassiation by Mixed Potassium/Lithium Amides: A Synthetic Path to Difficult to Access Phenethylamine Derivates

Abstract Insights gained from a comparison of aminometalation reactions with lithium amides, potassium amides and mixed lithium/potassium amides are presented. A combination of structural characterization, DFT calculations and electrophile reactions of aminometalated intermediates has shown the advantages of using a mixed metal strategy. While potassium amides fail to add, the lithium amides are uncontrollable and eliminated, yet the mixed K/Li amides deliver the best of both systems. Aminopotassiation proceeds to form the alkylpotassium species which has enhanced stability over its lithium counterpart allowing for its isolation and thereby its further characterization.

The transition metal catalysed hydroamination is an important reaction in synthetic chemistry. [1] The related catalytic reaction of alkene derivatives with lithium amides has been widely investigated (Scheme 1). [2] Major limitations remain for the use of polar lithium metal amides for alkene addition reactions as uncontrollable polymerization is often an occurring reaction. [3] Presumably, this is due to a combination of reversible b-elimination from addition product A (higher stability of B compared to A) and carbolithiation by A of the starting alkene substrate (higher nucleophilicity of LiCR 3 than LiNR 2 ) (Scheme 1). [2e,4] Yet research for new approaches is a necessary topic of modern chemistry as if these undesirable features could be controlled, the synthetic scope of A (beyond protonation as in hydroamination) would become available via reactions with electrophiles producing H. [5] To date, reactive intermediates such as A have not been isolated, making progress in addressing these issue challenging and slow. [6] To access new synthetic strategies and influence the reaction pathway, these two limiting components must be overcome. At first glance, finding a means of preventing both the polymerization and b-elimination reactions of A may appear contradictory. To prohibit these undesirable pathways, the reaction barrier for addition should be lowered and the carbanionic centre formed needs to be stabilized. In this account, we report our efforts to achieve this by exploiting the characteristics of different alkali metals (Li and K) in combination with stabilizing groups and allowing the thermodynamics of the reaction to select the preferred metal from a mixture of both. [6] Within this paper, the reaction pathway too difficult to access building blocks is presented on the basis of synthesizing b-metalated amines. At the outset of this study, the advantage that aminometalated intermediates of type A can be accessed by either a deprotonation of G or aminometalation reaction with C was recognized as a unique approach to investigating this challenging problem. [7] The inaccessibility of A via an alkene aminometalation route has restricted studies which may shine light on why this route is so challenging to control. As such, we first chose to access derivatives of A via deprotonation using either tBuLi or Schlossers base mixture of tBuOK and nBuLi which would allow a comparison of metallic reactions containing either lithium alone or both lithium and potassium. Previous work has shown the value of mixed K/Li amides for selective deprotonations which indicated that they had potential for the development of a new aminometalation strategy. [8] The first substrate chosen for investigation was N,Ndimethyl-2,2-diphenylethan-1-amine (2 a; Scheme 2) as the inclusion of a geminal diphenyl group should limit undesirable amide eliminations through stabilization of the metalated intermediates. [9] Scheme 1. Schematic sequence and side reactions of the catalytic aminolithiation shown for the example of the addition of lithium dimethylamide (B) to styrene. tBuLi and Schlossers base (tBuOK/nBuLi) were used to achieve an irreversible deprotonation of 2 a, and the subsequent metalated mixtures were subjected to crystallization studies. As expected, reaction of 2 a with tBuLi did give deprotonation, although the aminolithiated 1 a was not observed. In all attempts, the subsequent b-elimination product lithium dimethylamide [4 a·2 THF] 2 was obtained. The lithium dimethylamide crystallizes in a mixture of THF and n-pentane in the triclinic crystal system, space group P1 (Figure 1). The asymmetric unit contains one half of an inversion symmetric dimer of the lithium amide. The central structural motif is a Li-N rhombus. Each of the lithium cations is coordinated by two THF molecules as well as two dimethylamide groups. Attempts to in situ react 1 a with electrophiles failed to provide products (see later for discussion). These results indicate that under the typical reaction conditions used the equilibrium lies towards the aminolithiation starting materials 3 and 4 a, which is consistent with the known failure of this reaction.
Next, deprotonation with Schlossers mixed metal system of tBuOK/nBuLi was explored. Again, as expected, an irreversible deprotonation occurs, but in contrast to the above example, this reacting system can self-select from either lithium or potassium, allowing either amino-metalated species 1 a or 1 b to be formed. From crystallization studies of this reaction, single crystals of the metalated intermediate in which potassium was the metal of choice as in [1 b·4 THF] 2 were obtained. The potassiated species crystallizes at À80 8C in THF in the monoclinic crystal system, space group P2 1 /n. The asymmetric unit contains one monomer of the metalated species. The potassium is coordinated by four THF molecules, the dimethylamino group and C2, C9, C14 and C19 of the diphenyl group. The C-K distances range from 3.065(2) to 3.389 (2) . To the best of our knowledge, this is the first monomeric potassiated structure only coordinated by the solvent THF and one nitrogen since other reported mono-meric structures utilize chelating nitrogen-based ligands. By warming up the crystals of the monomeric species on the microscope slide in perfluorinated oil to À20 8C and recrystallisation, a polymeric potassiated species [1 b·2 THF] 1 is formed, which crystallized in the monoclinic crystal system, space group P2 1 /c ( Figure 2). [10] This transformation shows that THF can easily be removed from a potassium cation in favour of forming a polymeric potassium network. This can also be done selectively by using the reaction mixture of [1 b·4 THF] 2 , removing the solvent in vaccuo, resolve the residue in n-pentane and crystallize the compound [1 b·2 THF] 1 at À79 8C. This favouring of potassium over lithium at the carbanion centre is consistent with previous work in which we have observed that the deprotonation of toluene with a mixed K/Li base results in the formation of benzyl potassium. [11] To gain a more in-depth understanding of the reaction mechanisms involved, quantum chemical calculations of the deprotonation and metal amide elimination reactions utilizing tBuLi and Schlossers base have been performed (Scheme 3). [12] The obtained structures [1 b·4 THF] and [4 a·2 THF] 2 served as basis for the calculations and a mixed Na/Li system (tBuLi/tBuONa) was included for additional comparison (Table 1). Calculations predicted that the most favourable deprotonation conditions are with Schlossers base (63 kJ mol À1 ) followed by tBuLi (73 kJ mol À1 ) being the next best. Other alkyllithiums such as MeLi or iPrLi were less efficient and the combination of tBuLi with a sodium source further decreased the reactivity (86 kJ mol À1 ). Comparable results were obtained for the deprotonation of N,N-dimethyl-2-phenylethan-1-amine and N,N-dimethyl-2-phenyl-2-(trimethyl-silyl)ethan-1-amine (see SI).   The subsequent elimination reaction has the highest barrier with potassium (activation barrier: 88 kJ mol À1 , thermodynamics: 70 kJ mol À1 , free energy: 36 kJ mol À1 ) and is most likely to happen with lithium compounds (activation barrier: 51 kJ mol À1 , thermodynamics: 17 kJ mol À1 , free energy: À15 kJ mol À1 ) (Scheme 3, Table 2).
Comparable results were obtained for lithium, sodium and potassium metalated N,N-dimethyl-2-phenylethan-1amine and N,N-dimethyl-2-phenyl-2-(trimethylsilyl)ethan-1amine (see SI). Taking together the experimental and computational results strongly indicates that the undesired b-elimination observed for aminolithiation reactions should be experimentally controllable if potassium amides are used for an addition to styrene derivatives (Scheme 4). Using 4methoxystyrene and alkali metal dimethyl-amides (Li, Na, K) as substrate, quantum chemical calculations of the aminometalation reaction show that a reaction with potassium amides should be kinetically (lower addition barrier/higher elimination barrier) and thermodynamically possible (Scheme 4). In contrast, the lithium was the least favourable with sodium in between. Calculations also revealed that an aminolithiation should be kinetically hindered and, depending on the styrene derivative, could also be endothermic (see SI).
While computational studies showed the advantage of potassium over lithium and sodium, experimental evidence was obtained to show that potassium alone was insufficient and a mixed K/Li amide is essential for a positive reaction outcome. Using potassium piperidide on its own and in combination with potassium-t-butoxide and 5 or 3 as substrates, no aminometalation product was obtained, with starting material recovered (Scheme 5).
This illustrates that a more complex and synergistic mixed metal species is necessary to facilitate conditions for a stoichiometric aminometalation. Schlossers base provides the two metal components and an alkoxide to in situ produce a potassium amide more suited to our needs. By mixing piperidine with tBuOK and nBuLi a more effective, synergistic mixed metal system is formed. After adding of 3, the aminometalation with subsequent aqueous work-up could be performed in an isolated yield of 85 % (Scheme 6). By changing the solvent to a more nonpolar 1:1 mixture of THF and n-pentane we herein report the first functionalization of an aminometalated species with different electrophiles (MeOD, nBuBr, Me 3 SiCl).
By adding a second stabilizing phenyl group to the molecule as well as utilizing the effect of the potassium, the aminometalation reaction could be performed. Is the barrier lowering effect of the potassium high enough that no second stabilizing phenyl group is needed? Also, the reaction with 4methoxystyrene (5) has been performed. An isolated yield of 88 % could be obtained (Scheme 7). Unfortunately, the omission of a second stabilizing phenyl group leads to a more complicated reaction kinetic, being more sensitive towards changes of the reaction parameters and thereby Scheme 3. Quantum chemical calculation of the b-elimination of different alkali metal dimethylamides (DA) using dimethylether (DME) as solvent. Table 1: Results of the quantum chemical calculations of the deprotonation of diphenylethylamine 2 a in kJ mol À1 with different organometallic bases; for further information regarding the modulated reaction schemes see Supporting Information; M062X/6-31 + G(d). [12,13] tBuLi MeLi iPrLi tBuOK/ nBuLi tBuONa/ nBuLi DDH TS 73 58 52 63 86 DDH product À99 À79 À94 À99 À78 Table 2: Results of the quantum chemical calculations of the belimination of metal dimethylamides from 1 in kJ mol À1 ; for further information regarding the modulated reaction schemes see Supporting Information; M062X/6-31 + G(d). [12,13] Li + 2 DME K + 3 DME Na + 2 DME Na + 3 DME Scheme 4. Quantum chemical calculation of the aminometalation of 4methoxystyrene with different alkali metal amides using DME as solvent; M062X/6-31 + G(d). [9,13] Scheme 5. Failed aminopotassiation with potassium piperidide with and without potassium-tert-butoxide.

Angewandte Chemie
Communications 22502 www.angewandte.org hindering quenching with electrophiles. For example, as recently shown by Hevia et al., moisture plays a significant role in hydroamination reactions. [14] Furthermore, a source from which the metalated 4-methoxystyrene abstracts a proton could not be identified. [15] To prove this reaction, also the intermediate of the aminometalation reaction with a potassium amide should be isolated. Crystals of the aminometalated 1,1-diphenylethene could be obtained (Figure 3). The species crystallizes in THF in the monoclinic crystal system, space group P2 1 /n. The structure demonstrates that an aminometalation is possible and the potassium, as already assumed in the calculations, is significantly better stabilized in the reactive intermediate than in the corresponding potassium amide because of interactions with p-electrons.
Moreover, another species could be obtained, which is generated during the aminometalation (Figure 4). This aggregate contains deprotonated piperidine, potassium, lithium, tert-butoxide, enolate and THF as ligand. The species crystallizes in THF in the orthorhombic crystal system, space group Pnma. The mixed lithium/potassium structure 10 shows that the extraordinary reactivity might be increased by using this special mixture of an organolithium compound, a potassium compound and an amine. Proof of this synergistic effect was also given by using a reaction mixture without lithium (Scheme 5). [16] The molecular structure in the crystal in combination with the failed reactions of the potassiated piperidide on its own and in combination with potassium-tert-butoxide with both styrene derivatives show that the situation of the reactive potassium amide is much more complex. Considerations are needed whether parts of the structure observed in the crystal are also involved in the reaction mechanism and influence the reaction mechanism. Further anions such as the alkoxide anion seem to be necessary in addition to the amide. Additionally, alkoxides might also increase the solubility and by this increase the reactivity. Also, two different or even more alkali metal ions must be present. However, structure 10 represents only the thermodynamic minimum of a decomposition product of THF and does not show the desired reactivity.
In conclusion, highly reactive intermediates can be accessed either by deprotonation reactions of phenethylamine derivates but also by an alternative pathway: the addition of alkali metal amides to the double bond. A stoichiometric aminometalation reaction of styrene derivatives with potassium amides at low temperatures without   The hydrogen atoms and disorder in the THF molecules are omitted for clarity. [10]