Theoretical and Experimental Investigation of the Formation of E-and Z-Aldimines from the Reaction of Methylamine with Acetaldehyde

A reação entre metilamina e acetaldeído em pentano a 230 K leva à formação das aldiminas Ee Z-, sendo o isômero Zo produto principal e gerado sob controle cinético. O isômero Eé mais estável do que o Zpor 3.3 kcal/mol (∆G*). Os cálculos ab initio mostraram que a etapa bimolecular de adição ocorre por um processo concertado, com formação da ligação C-N e transferência de próton através de um estado de transição de quatro centros. O intermediário zwiteriônico, frequentemente invocado neste tipo de reação, não existe como um mínimo na superfície de energia potencial. Entretanto, a energia livre de ativação para o mecanismo concertado é muito alta, e não é capaz de explicar a cinética deste processo. O mesmo acorre com a etapa de eliminação de água para formar a imina, ou seja, possui uma barreira de ativação muito alta. Estes resultados implicam que tanto a etapa de adição, como a de eliminação, devem ocorrer por um mecanismo de catálise. O isômero Zpode se interconverter para a forma Eatravés de uma inversão sobre o nitrogênio sp. A 298.15 K, a energia livre de ativação para este processo foi calculada como sendo 27.0 kcal/mol, o que leva a um tempo de vida de quatro meses para este isômero.


Introduction
The study of the properties and the formation of imines has great interest due to its role in several important chemical processes 1 , mainly in the synthesis of polyazocycles 2 and other nitrogened heterocycle compounds 3 .In relation to its biological importance 1 , the literature reports the par-ticipation of the imine group in the synthesis of artificial oxiredutases 4 , in the formation of compounds with antiviral 5 , antimicrobial 6 , antihypertensive 7 , and antitumor 8 activities, and also in the synthesis of penicilin 9 , cefaloporfirin 10 and nocardicin 11 , and many other biologically active compounds 12 .
Although they have a large applicability, methodological studies have been addressed mainly to obtain aromatic aldimines and ketimines, as they are relatively stable under several experimental conditions 13 .On the other hand, aliphatic aldimines have been less chemically studied due to their instability in protic solvents and acidic medium 14 .In these situations, they can undergo imine-enamine tautomeric equilibrium 15 and imine-imine isomerization 16 , leading to the formation of several auto-condensation products 17 , and decreasing their yield 18 .Thus, experimental procedures for the preparation of aliphatic aldimines have become more specific in order to improve the yield 19 .
The reaction mechanism of amine addition to carbonyl compounds in aqueous solution has been studied at the experimental level [20][21][22][23][24] , and some theoretical studies have been reported 25,26 .Sayer et al has proposed a general mechanism for this reaction in aqueous solution and at variable pH.The first step can be catalyzed by H 3 O + at low pH, forming the aminoalcohol intermediate.Increasing the pH, a competitive uncatalyzed mechanism can take place, which occurs through the formation of a supposed zwitterion tetrahedral intermediate.This species could dissociate backward, or rearrange to aminoalcohol in an acid catalyzed or uncatalyzed mechanism.Support for this mechanism was provided by the pH dependent rate profile.However, ab initio studies by Yamataka et al 25 for the addition of several amines to carbonyls do not support this proposed mechanism.The zwitterion intermediate is not a minimum energy structure on the potential energy surface, and these authors have found a four-center transition state for this step, generating directly the aminoalcohol.Another ab initio study of the NH 3 addition to HCHO carried out by Williams 26 has shown that one or two water molecules can act as a catalyst through a cyclic transition state.These results raise doubts about the formation of the zwitterion intermediate, and point out that the uncatalyzed mechanism in reality is water catalyzed.
The second step of the reaction is the water elimination of the aminoalcohol, forming the imine.This step can be acid catalyzed in low pH, and presents a competitive uncatalyzed mechanism, which can become important when the pH increases.So, in the case of a reaction involving amine and aldehyde in a neutral and low polarity medium, which is the interest of this study, the overall mechanism can be represented by the Scheme 1, where we have included the possibility of the formation of both E-and Zisomers.
The interconversion from E-to Z-isomers of imines has been the object of many theoretical and experimental studies which have addressed questions on the mechanism, kinetics and thermodynamics of this process 16,[27][28][29][30][31][32][33][34] .These studies pointed out that in aprotic medium, isomerization occurs by inversion of the nitrogen atom, as indicated in Scheme 2. For the prototypical methylimine species, ab initio calculations by Pople et al 29 indicated a barrier of 30.5 kcal/mol at MP4/6-31G**//HF/6-31G* level of theory.With the increase of the groups bonded to nitrogen and carbon, this barrier can decrease by up to 15 kcal/mol, as found in experimental studies 34 .In thermodynamic terms, either E-or Z-isomers can be the more stable, depending on the substituinte groups.For aldimines, E-isomers in general present lower free energy.As an example, Wiberg 31 has combined experimental and theoretical data to obtain the heat of formation of (E)-and (Z)-N-methyl-acetaldimine, and has found that the E-isomer is more stable by 4.4 kcal/mol.Another pathway to isomerization is by enamine formation in protic medium 27,30 .For acetaldimine, the enamine form is less stable by 3.94 kcal/mol 32 , and depending on the substituinte groups, this order can be reversed 33 .The catalysis by the solvent in this pathway is evident, since a unimolecular process involves a high activation energy, 66.34 kcal/mol for acetaldimine 33 .
In order to obtain stereochemical control in aldimine synthesis, the above discussion should be taken into consideration before proceeding with the experiment.In the present work, our aim is to study the mechanism, kinetic and thermodynamics for the formation of a prototypical aldimine, N-methyl-acetaldimine.We hope that this study might help in the planning of the synthesis of aldimines with a specific isomeric form.

Experimental
The 1 H-NMR spectra were recorded on a Bruker DRX 400 using pentane and CDCl 3 as solvents and TMS as the internal reference.The spectrum was registered with signal suppression of trace of CHCl 3 , using a temperature variable program.
The reaction was conducted utilizing the following procedure: Methylamine 0.1 mol (Note: aqueous solution of methylamine 25% p/v was utilized), 100 mL of pentane and 65 g of anhydrous sodium sulfate in a 3-necked, roundbottomed flask connected to an addition funnel and a condenser with a drying tube.After the mixture has been stirred for 1 h at 0 °C, acetaldehyde 0.1 mol (purified by bidistillation at atmospheric pressure) was slowly added with vigorous stirring at 200 K. Aliquots of the reaction mixture were analyzed by 1 H-NMR at temperatures of 230 K.
The ratio of aldehyde to imine in the reaction mixtures was calculated by the integration of the signals corresponding to the aldehydic hydrogen near to δ 9.0 and the aldiminic hydrogens near to δ 7.0.An error of 2.5% in the ratio of the integrals of signals was estimated using standard solutions of butanal and chloroform.

Ab Initio Calculations
The potential energy surface for methylamine addition to acetaldehyde was investigated at the ab initio Hartree-Fock (HF) and second order Mφller-Plesset Perturbation Theory (MP2) levels, using the 6-31G*, 6-31+G* and 6-311+G(2d,p) basis sets.Full geometry optimizations and harmonic frequency calculations were performed at the HF/6-31G* level, and in order to obtain a good estimate of reaction and activation energies, single point calculations were carried out at the MP2 level using the 6-31+G* and 6-311+G(2d,p) basis sets.The ab initio data were used to determine the thermodynamic and kinetic parameters for each step by statistical mechanics and transition state theory calculations [35][36][37] .The harmonic frequencies scaled by a factor of 0.90 were used in the calculation of the vibrational partition function.The symbols ° and * were used to represent the standard states of 1 atm of pressure and 1 mol/L of concentration (free energy).The parameters H* and S* are defined by the relations: The optimized structures are in Fig. 1 and the results of the calculations in Tables 1 and 2. All ab initio calculations were performed using the GAUSSIAN 94 program 38 .

Results
The 1 H-NMR spectrum of an aliquot of the reaction mixture between methylamine and acetaldehyde after 1 hour at 230 K presents a quartet signal in δ 9.75, attributed to aldehydic hydrogen.The singlet signals at δ 7.31 and δ 7.19 are attributed to aldiminic hydrogens of the two isomers.Analyzing the integration of signals (proportion of 11.86, 1.00 and 1.74, respectively) we can note that in the time of the experimental measurements, just 19% of the aldehyde has reacted (not considering the aminoalcohol), forming both E-and Z-isomers.We have attributed the δ   The structures of minima and transition states for the present system are in Fig. 1 and a diagram of the potential energy surface is in Fig. 2. We have located two transition states (TS1a and TS1b) for the first step of methylamine addition to acetaldehyde.In accordance with Yamataka et al. 25     water molecule.The free energy barriers are 48.4 kcal/mol and 51.2 kcal/mol, respectively.In relation to aminoalcohol, the Z isomer is 0.5 kcal/mol less stable, and the E isomer is 2.8 kcal/mol lower in free energy.In these calculations, the existence of two enantiomers of the aminoalcohol was taken into account.Also, four other conformations were found for the aminoalcohol, but we have reported only the most stable one, because these conformers can interconvert easily.Thus, they will not be kinetically significant.Another possible step is the isomerization from the Zform to the E-form by the transition state TS4.This motion corresponds to inversion on the sp 2 nitrogen, where the C-N-C atoms are almost collinear.The activation free energy predicted at 298.15 K is 27.0 kcal/mol.

Discussion
Considering the competitive steps 2 and 3, the ab initio results suggest that the Z-isomer is formed faster than the E-isomer, so that under kinetic control, it will be the dominant product.The 1 H-NMR measurements confirm this prediction, with the proportion of products of 63% of Zand 37% of E-isomers.The thermodynamic data obtained from the calculations indicate that the equilibrium is shifted to the E-isomer product, with an overall ∆G* of -3.4 kcal/mol, and it is more stable than the Z isomer by 4.3 kcal/mol in ∆H* terms, in excellent agreement with the Wiberg 31 value of 4.4 kcal/mol.So, at thermodynamic equilibrium, only the E-isomer will be observed.This result only permits us to infer that the experimentally conducted reaction has not attained thermodynamic equilibrium.However, if we make the assumption that the system is in a quasi-equilibrium state with relation to Z-isomer formation, and the water molecules formed remain in pentane solution, we can calculate the ∆G* for Z-isomer formation from aldehyde and amine as 1.5 kcal/mol at 230 K, in reasonable agreement with the theoretical value of -0.1 kcal/mol at this temperature.The small difference may be due to the quality of the ab initio calculations, the assumption that water molecules remain in pentane solution, and more probably the quasi-equilibrium assumption.Indeed, progression of the reaction would lead to a smaller experimental ∆G*.
The qualitative analysis of kinetic parameters is in agreement with experimental observations i.e. the Z-isomer is the kinetically dominant product.However, the obtained barriers in all steps are very high.In order to have a quantitative comparison between theory and experiment let us make a rough estimate of the experimental free energy for the first step, the formation of the aminoalcohol.If the first step is bimolecular and rate determining for the reaction, and after one hour 20% of the reactants at 1mol/L have reacted, then we can estimate an activation free energy of 18 kcal/mol at 230 K, against 33 kcal/mol obtained by ab initio calculations.The difference of 15 kcal/mol is very large, and it can not be attributed to the imprecision of the level of theory used.If we observe the barriers for water elimination (steps 2 and 3), they are about 50 kcal/mol.The implication of these results is clear.The first step does not occur via a bimolecular process, and the water elimination step does not occur by a unimolecular mechanism.Some catalyzed mechanism is acting.In the case of reactions in aqueous solution, water molecules play the dominant role as shown by Willians 26 .In the present case, as the reaction was conducted in an aprotic solvent, others species could act as a catalyst, as the proper amine, the aminoalcohol intermediate and the water product.A definitive answer will require more investigation.
The unimolecular isomerization from the Z-to the Eisomer presents a ∆G* ≠ of 26.9 kcal/mol at 230 K.The resulting rate constant is 1.3 x 10 -13 s -1 , and this very low value indicates that the Z-isomer is kinetically stable at this temperature.So, this pathway can not lead to the E-isomer.At room temperature, we obtained ∆G* ≠ = 27.0 kcal/mol, in excellent agreement with 27.1 kcal/mol found for Nmethyl-1-(1-naphthyl)-ethylimine 27 , which also presents motion of the methyl group bonded to nitrogen.The lifetime to convert the Z-isomer to E-isomer, through TS4, is about four months.Thus, the theoretical calculations predict that the Z-isomer can be obtained and conserved without isomerization to the E-form, since catalytic species are not present and the temperature is kept low.

Conclusion
The present theoretical and experimental results show that it is possible to obtain the less stable Z-isomer under kinetic control as the dominant product.The reaction should be conducted in an aprotic solvent and at low temperature.A drying agent should be used in order to shift the equilibrium toward products.
Analysis of the reaction mechanism shows that in an aprotic and low polarity solvent, the addition step can not be a bimolecular process and neither can the elimination step be unimolecular.A catalyzed mechanism is actuating, and further study is necessary in order to discover the actual reaction pathway.
Alcântara would like to thank the Chemistry Department of UFMG.

Figure 1 .
Figure1.Minima (MS) and transition state structures (TS) located on the potential energy surface for the CH3NH2 + CH3CHO reaction.TS1 corresponds to the transition state for step 1 (Scheme 1), TS2 corresponds to the transition state for step 2, and so on.

7. 31
signal to the E-isomer, and the δ 7.19 signal to the Z-isomer.
study for other amine-aldehyde pairs, no zwitterion intermediate was located on the potential energy surface.The bimolecular process occurs by attack of the nitrogen on the carbonylic carbon and concerted transfer of proton to oxygen, forming the aminoalcohol intermediate (MS2).These transition states have an activation free energy of 35.8 kcal/mol and 35.0 kcal/mol, respectively, and the ∆G* for the formation of the aminoalcohol is -0.6 kcal/mol.The following step is the elimination of water by an unimolecular process.Depending on the aminoalcohol conformation, either Z-or E-isomer can be formed through transition states TS2 and TS3.These structures present a considerable heterolytic C-O bond cleavage, and the forming hydroxyl anion captures the proton bonded to nitrogen to form the 385 Investigation of the Formation of E-and Z-Aldimine J. Braz.Chem.Soc.

Figure 2 .
Figure 2. Potential energy surface profile for the CH3NH2 + CH3CHO reaction.Structures of minima and transition states are indicated in accordance with Scheme 1 and Fig. 1.Free energy in kcal/mol, and standard state of 1 mol/L.

7 a-
The numbering corresponds to the transition states in figure1and steps in Scheme 1. Energies are in kcal/mol.T = 298.15K.The Symbols ° and * correspond to standard state of 1 atm of pressure and 1 mol/L of concentration.b -Zero point vibrational energy contribution.Frequencies scaled by 0.90.c -MP2/6-311+G(2d, p) energy + ∆ZPE.d -The two enantiomers of the aminoalcohol and of the TS1a and TS1b transition states were considered in the calculation of the activation properties.e -Rate constant at 298.15 K calculated by transition state theory.Units of L, mol and s.

Table 1 .
Reaction energies and thermodynamics properties for methy- lamine addition to acetaldheyde determined by ab initio calculations a .
b -Zero point vibrational energy contribution.Frequencies scaled by 0.90.c -MP2/6-311+G(2d, p) energy + ∆ZPE.d -The two enantiomers of the aminoalcohol were considered in the calculation of the thermodynamics properties.e -Equilibrium constant at 298.15 K. Units of mol and L.

Table 2 .
Energy barriers, activation thermodynamic properties and rate constants for methylamine addition to acetaldehyde determined by ab initio