Isoprene-mediated lithiation of imidazole derivatives: mechanistic considerations

The isoprene-mediated lithiation, with lithium metal, of different imidazole derivatives is an interesting methodology for their functionalization. Studies of different possible intermediates involved in the reaction employing density functional theory calculations, at the B3LYP/6-311++G(d,p) level are considered. A plausible mechanism is described, in which isoprene is reduced, to the corresponding radical anion, in the presence of Li (s) , acting then as a base deprotonating N -methylimidazole (NMI) and producing the 1,1-dimethylallyl radical. This radical is further reduced by the excess of lithium proceeding once more as a base. This final step produces stable final products that compensate the previous equilibriums, making favourable the whole process.


Introduction
Compounds classified as heterocyclic probably constitute the largest and the most varied family of organic compounds, and often display important bioactivities.Actually, more than 67% of the compounds listed in the Comprehensive Medicinal Chemistry (CMC) database contain heterocyclic rings. 1 Indeed, the remarkable ability of heterocyclic nuclei to serve both as biomimetics and reactive pharmacophores has largely contributed to their unique value as traditional key elements of numerous drugs. 2 Lithium metal in combination with a substoichiometric amount of an arene, as an electron carrier, has become a very versatile methodology in the preparation of organolithium intermediates. 3A significant variety of functionalized organolithium reagents have been prepared by means of this protocol. 4Additionally the arene-promoted lithiation has been employed in the preparation of polylithiated synthons 5 and in the generation of active nanoparticles by activation of other transition metals. 6In the last years, interesting mechanistic studies on the arene-catalyzed lithiation process have been taken into consideration, providing interesting information regarding this well-established methodology. 7ecently, we also found that lithium metal, without any additive, is able to generate 2-lithio-N-methylimidazole (2-Li-NMI) from N-methylimidazole (NMI, 1), although the use of isoprene during the lithiation process gave an improvement in the results. 8Thus, in the last five years, we have been working on the use of isoprene as a promoting lithiation agent for the generation of different lithio-imidazole derivatives.Isoprene-mediated lithiation of 1-alkyl-1-4, 8,9 1-phenyl-5 10 and 1-(diethoxymethyl)imidazole 6 11 has been reported (Figure 1).Thus, concerning the role of isoprene in the reaction, we have considered, from the beginning, that it can act as an electron carrier agent 12 and/or as a hydrogen acceptor. 13Lithium metal, under certain reaction conditions, can produced the ring opening of the THF, 7 employed as solvent, generating the corresponding alcoholate which could act as a base. 14However, the mixture Li(s)/isoprene did not produced any traces of 1-butanol after hydrolysis. 15In this work we consider some interesting studies towards understanding the isoprene-mediated lithiation of imidazole derivatives.

Results and Discussion
We considered the study of the generation of 2-Li-NMI from NMI 1 employing isoprene as an additive during the process.Firstly, in order to obtain a theoretical qualitative description of the experimental reactivity for the isoprene/NMI/Li(s) system, semiempirical calculations 16 for NMI, s-cis-and s-trans-isoprene, both as neutral and anionic species, were performed.Soon it was evident that this protocol provided largely defective data when quantitative analyses of the energies involved were sought after.Density functional theory calculations 17 of neutral and anionic structures, energies, and harmonic vibrational analysis of NMI (-1) , and s-cis-and s-transisoprene (-1) , as well as related species, were carried out thereafter using the Becke-Lee-Yang-Parr (B3LYP) exchange-correlation functional, 18 in an attempt to attain larger accuracies in the calculations and set the foundations for a theoretical description of the reactivity and the processes involved.The geometries of the isolated neutral and anionic species have been optimized both in the gas phase using the 6-311++G (d,p) basis set. 19,20e substrate NMI was found to occur in only one conformation 1, being a stationary point as confirmed by vibrational analysis.Additionally, the corresponding methyl rotamer 1' appears as a saddle point some 1.17 kcal mol -1 above (ΔGº ‡ ) conformers of type 1 and was not considered in the calculations (Equation 1).Regarding the anionic NMI 1 -, the geometry of the minimum is well defined and resembles that of the neutral NMI with the corresponding bond lengthening within the ring (Equation 2).The relevant calculated energies and thermodynamic functions are described later.
On the other hand, isoprene 7 appears as a mixture of s-trans, gauche and s-cis conformers in 95.3% (s-trans) and the rest (ca.4.7%) corresponding to the gauche (or skew), the s-cis being too scarcely populated to be detected, according to vibrational spectroscopy and gas electron diffraction studies. 21According to our calculations for the equilibrium depicted in Equation 3, three conformers were also detected, corresponding to the s-trans-7 (180º dihedral), 22 skew-7 (38º dihedral) 23 and s-cis-7 (0º dihedral), 24 in relative amounts of 97.1:2.7:0.2 respectively, being in excellent agreement with the experimental observations.For the anionic species, the s-transisoprene radical anion (s-trans-7 -) 25 retains the antiperiplanar geometry of the s-trans-7 after the electron capture, and both the skew-7 and the s-cis-7 converge to the same synperiplanar anion (s-cis-7 -) 26 after the electron capture.This latter conformer is the energetic minimum, with a meager difference of Gºisom = -1.1 kcal mol -1 for the equilibrium between s-trans-7 -and s-cis-7 -(Kº = 6.4).In both cases, similar bond lengthening respect to the neutral molecule was observed.
Electron affinity (EA) calculations.In the context of the quantum mechanics, the experimental values of adiabatic electron affinity (AEA) of several small molecules have been well correlated by the use of the functional hybrid B3LYP with different basis sets, such as 6-31+G(d), 27 which has been considered to provide satisfactory reliable results.The B3LYP/6-311++G(d,p) method, with vibrational energy corrections at 0 K (ZPE) has been recently employed in the calculation of adiabatic electron affinities of small-medium size molecules (up to 6 heavy atoms). 28For bigger molecules, up to 15-20 heavy atoms, theoretical calculations in the level of B3LYP/6-31G(d) with ZPE corrections have been used in the calculation of AEA. 29 In this work, AEAs have been calculated employing the above mentioned method with the incorporation of the ZPE corrections in all the cases. 30The calculated EA for NMI 1 31 is -0.39 eV, and for isoprene are -0.43 eV (for s-trans-isomer) and -0.29 eV (for skew-isomer). 32The small difference between these EAs, which is within the ±0.21 eV of average absolute deviation calculated for this method in comparison with experimental values, 33 did not provide any significant evidence regarding the possible mechanism.Proton affinity calculations: isolated anions.In order to set out a plausible reaction mechanism, different calculations in gas phase were carried out to obtain the estimation of the proton affinities for some species involved in the process.The considered process, which is represented in Equation 4, shows the deprotonation of NMI by the isoprene radical-anion 7 -, 26 generating the corresponding radical 8 34 and the deprotonated NMI 1 -. 35This preliminary result, employing isolated anions, showed that the deprotonation between the isoprene radical-anion 7 - and the NMI 1 is a highly unfavorable process, having an activation free energy of Gº(4) = 14.5 kcal mol -1 (Kº = 2.510 -11 ) for the s-trans and s-cis-7 -equilibrium mixture.
Proton affinity calculations: stabilized anions with Li + .The previous unfavorable results, for Equation 4, can be rationalized taking into consideration that the calculations were performed for the isolated anions in gas phase without the stabilizing effect of the counterions.Thus, the observed energetic balance is principally due to the contribution of NMI (as the only neutral and stable species in the equilibrium).For that reason we considered the deprotonation reaction having the Li + as counterion of the anionic species, as depicted in Equation 5.
From the data analysis, we observed that the cationic stabilization modified the equilibrium in the expected direction. 36The free energy balance is far more equilibrated, with a Gº(5) = 1.8 kcal mol -1 (Kº = 4.610 -2 ).The structure of the species Li + 7 •-and Li + 1 -are depicted in Figure 2.These results predicted a deprotonation equilibrium between the isoprene radical-anion Li + 7 •- and the NMI 1, providing the reaction intermediates Li + 1 -and 1,1-dimethylallyl radical 8.The slightly endergonic balance of this first step, can be made up for a subsequent reaction, essentially irreversible, where radical 8 • would react.Reaction of dimethylallyl radical.The 1,1-dimethylallyl radical 8 • is an unstable reaction intermediate and it can progress in various manners.Without dismissing the typical behavior of the radicals (such as coupling processes, dismutations, etc.), we considered at this point that radical 8 • mainly would be transformed into the corresponding anion stabilized with Li + (1,1dimethylallyllithium, Li + 8 -) by reaction with the excess of Li(s) in the reaction medium (Equation 6).
In Equation 6, it is involved a condensed phase, so a strict analysis of the process employing calculations is hampered.However, due to the fundamentally irreversible nature of the process, the thermodynamic parameters of the process were estimated employing some approximations and the experimental values of the reduction potentials, knowing full well that these approximations did not affect significantly the energetic balance of the global process, which is highly exergonic.For such purpose, © ARKAT-USA, Inc.
a) the values of the half-reactions of reduction depicted in Equations 6a and 6b were used, considering the EA for the radical 8 • in gas phase 34 and the experimental redox potential for Li + (THF)/Li(s), to obtain an approximated value of Gº(6); b) the energy of solvation of Li + in THF and the energy of formation of the Li + 8 -ion pair were equated (Equation 6c).This implies that Gº(6c) = 0, what simplify the thermodynamic estimations for the overall process.These estimations, which are without doubt conservative (vide infra), avoid including solvation energies, what would make an exceedingly tedious calculation, as well as unnecessary.
Strictly, the solvation energies for intermediates 8 • and 8 -were not estimated in the approximation done in 6c, as well as the energy of the ion pair Li + 8 -(THF) formation from the corresponding THF solvated ions, being all of them exergonic contributions. 37The free energy variations in standard conditions for both half-reaction redox were calculated: Gº(6a) = -7.7 kcal mol -1 , 38 and Gº(6b) = 75.9kcal mol -1 . 39As expected, the standard free energy variation for Equation ( 6), Gº(6) = -83.6 kcal mol -1 , showed an irreversible process ruled by the high redox potential of Li(s) in THF.Thus, the reduction of the radical 8 • by metal lithium means adding an essentially irreversible step to the initial proposed step (slightly endergonic), what moves the different equilibriums to the formation of the corresponding products.Final step.Finally, the corresponding calculations for deprotonation of NMI by 1,1dimethylallyllithium Li + 8 -(Figure 2) were considered (Equation 7). 40The variation of free energy for this last step, Gº(7) = -12.0kcal mol -1 (Kº = 6.810 8 ), shows an exergonic favorable process.With this final step, the reagents were transformed to the stable final products, with the mediation of isoprene in the deprotonation step of NMI.

Conclusions
The deprotonation step in Equation 4, which uses the radical anion of isoprene Li + 7 -as base, is slightly endergonic.Thus, this step can be considered the rate determinant step, among the postulated steps in homogenous phase.The protonation of 7 -occurs, in principle, at the C(1) of isoprene.This carbon is the one supporting the highest charge density and has the highest contribution to the HOMO, both as a free anion as well as forming an ionic pair with Li + .The methylenic character of C(1) [as well as C( 4)] in 7 -can facilitate the approximation to the most acidic proton of NMI (the one in position 2), and the corresponding radical 1,1-dimethylallyl 8 • is formed.The following irreversible reduction of 8 • with Li(s), depicted in Equation 4, moves the previous equilibriums to the products.Finally, the formed dimethylallyllithium Li + 8 -in that reduction can act also as a base deprotonating another molecule of NMI, so the isoprene can participate in substoichiometric amount compare to NMI (Scheme 1). 41Scheme 1. Proposed mechanism for the isoprene-mediated lithiation of N-methylimidazole employing lithium metal.
Although, there is no contribution regarding calculations of specific kinetic studies yet, the information provided by those would be, in any case, an incomplete vision of the process.Many of the heterogeneous processes involved, such as the initial isoprene reduction with Li(s) giving Li + 8 -(which was deliberately omitted in the whole treatment) are not readily accessible from a theoretical point of view.Indeed, they are likely to be controlled by a mass-transport phenomenon, which is associated with complex, heterogeneous kinetics.To conclude, we have to say that different mechanisms, as well as alternative mechanisms running in parallel to this one cannot be ruled out, but they have not been further considered.41.After performing the reaction employing the standard conditions (excess of lithium, 20 mol% of isoprene in THF at 25 ºC), the presence of 2-methyl-2-butene was detected in the reaction medium by GC-MS (EI at 70 eV on an Agilent 5973 spectrometer).