o-Aminomethylderivatives of phenols. Part 3. Mechanistic investigation of a Mannich reaction of phenols with N-methylenealkylamines

The mechanism of the Mannich reaction of phenols ( 1 ) with 1,3,5-trialkyl-hexahydro-1,3,5-triazines ( 2 ) as a source of N -methylenealkylamines ( 3 ) in organic solvent has been investigated. It has been determined that the formation of o -hydroxybenzylamine 4 and its subsequent quick transformation into benzoxazine 5 and primary amine in the reversible reaction with 3 takes place in the first stage of the process. Then, the yield of 4 slowly increases and that of 5 appropriately decreases as a result of the reaction of 3 with the so far unreacted 1 and the slow aminolysis of 5 . In addition, benzyldiamine 6 is formed as a by-product (Scheme 1).


Introduction
o-Aminomethylderivatives of phenols exhibit many interesting chemical properties and are the subject of growing interest.Some of them were tested as potential metal complexing agents including metal ions extractants, catalysts or their components, semi-products in organic synthesis, biologically active agents or components of polymeric materials. 1n the search for a simple and efficient synthetic method of 4, the application of a well known Mannich reaction 2 was spotted.The practical advantage of this "one-pot" process is the high availability of substrates -phenols, primary amines and formaldehyde.Its weak points are the possible side-reactions involving not only substrates but also products resulting in low yields of benzylamines synthesis. 1

Scheme 1
To determine the reaction course and to correctly interpret the results, a convenient analytical method based on 1 H NMR spectroscopy has been elaborated.All the products originating from 1a contain a methyl group connected to the phenol ring.The integration of the appropriate group of signals (2.25-2.20 ppm) was used as an "internal standard", corresponding to the initial amount of 1a.The conversion of 1a at every stage of the process was determined basing on the integration of its isolated methyl signal at 2.25 ppm versus the internal standard.The specific signals coming from methylene linkages at 3.86-4.95ppm for 4a-6a, 3.54 ppm for 2a and at 7.21 ppm for 3a allowed the determination of the yield or conversion of the appropriate products (Table 1).The cyclohexyl methylene groups' signals were observed in the range of 1.00-2.00ppm and did not overlap with any of the considered ones.An example of the reaction mixture spectrum is presented (Figure 1).Yields of products 4a-6a as well as conversion of 1a and 2a were evaluated basing on 1 H NMR spectra after 1, 5 and 10 h.After the first hour 2a was totally consumed, whereas the conversion of 1a was only about 55% and benzoxazine 5a was the main product (36% yield).In subsequent hours the yield of 5a gradually decreased whereas the conversion of 1a and yields of benzylamine 4a and benzyldiamine 6a increased (Figure 2).No significant influence of the solvent used (methanol, dioxane, chloroform and hexane) on the reaction course has been observed.Similarly, in the reaction of other phenols (1b-e) with 1,3,5-trialkyl-hexahydro-1,3,5-triazines (2b-e) in methanol or hexane, appropriate benzylamines 4b-e were isolated as the main products.The 1 H NMR control revealed changes analogous to that observed in the model system.The mutual reactivity of 4a and 2a has additionally been examined.Thus, after heating of 4a and 2a at 3:1 ratio in refluxing methanol for 1 h, 5a was formed with 66% yield, whereas the conversion of 4a was 75%.2a, 3a and traces of side products but no sign of 6a have been observed.No change of the reaction mixture composition after 5 and 10 h suggests that the reaction is reversible and reaches its equilibrium after maximum 1 h.The presence of cyclohexylamine (10% of the initial amount of 4a) caused the formation of a significant amount of product 6a (25% yield) in the multicomponent mixture.
Based on the results obtained, the mechanistic proposal of the investigated Mannich reaction is as follows.The first stage of the process is the addition of phenol 1 and Nmethylenealkylamine 3 with the formation of benzylamine 4, according to the aromatic electrophilic substitution mechanism (Scheme 1).Subsequently, 4 quickly reacts with the second molecule of 3 resulting in benzoxazine 5 and primary amine, therefore, the conversion of 1 is limited at this stage.The reaction of 4 with 3 is reversible, so, due to the addition of 3 with the so far unreacted 1, a further amount of 4 is being formed and 5 -dominating at the beginning of the process -undergoes slow aminolysis with regeneration of 4. This is why in the last stage of the process, while 3 is being consumed, the concentration of 4 slowly increases and that of 5 appropriately decreases.Benzyldiamine 6 results from a side reaction of 4 with 3 favoured in the presence of primary amine, which probably causes deprotonation of 4, increasing its reactivity at the same time.

Conclusions
The reported variant of the Mannich reaction is a simple convenient and practical method of ohydroxybenzylamines synthesis, 8 highly competitive towards the widely used multi-step procedure: o-hydroxybenzaldehyde → o-hydroxybenzylimine → o-hydroxybenzylamine, starting from expensive or not easily obtainable aldehydes.The application of 1 H NMR enables analytical control of the process at every stage.The practical conclusions of the presented mechanistic proposal are as follows: 1. Reaction time reasonably influences the yields of products.If it is sufficiently short, benzoxazine 5 is the dominant product that can possibly be isolated.Only after appropriately long reaction time the maximum yield of benzylamine 4 is observed.Various isolated products of the reaction reported previously 5,6 result probably from underestimation of the time factor.2. The highest yield of 4 can be obtained by application of an equimolar or almost equimolar ratio of phenol 1 and N-methylenealkylamine 3. The application of an excess of 3, which is traditionally expected to result in higher yield of 4 and conversion of 1, causes an increase in the yield of 5, especially after short reaction time.Even after relatively long, optimized reaction time, the excess of 3 results in lower yields of 4 and contamination of this product in the reaction mixture with 5 and 6.

Experimental Section
General Procedures.Other 1,3,5-trialkyl-hexahydro-1,3,5-triazines 2c-e were obtained according to the procedure described for 2a as high-boiling oils of a characteristic smell and were used without purification.
To the vigorously stirred suspension of 4a

Figure 1 . 1 H
Figure 1. 1 H NMR spectrum of the reaction mixture.