Direct synthesis of amides from nonactivated carboxylic acids using urea as nitrogen source and Mg(NO3)2 or imidazole as catalysts

This methodology is particularly useful for the direct synthesis of primary and N-methyl amides using urea as a stable and easy to manipulate nitrogen source.


Introduction
The importance of the amide functional group emerges from their presence in many crucial compounds such as proteins, fabrics, fertilizers, insecticides, plastics, drugs, and in a vast number of synthetic structures. For this reason, it is very relevant to develop new methods for the efficient synthesis of amides.
in the absence of a catalyst have been previously reported [67][68][69][70][71][72] and are favoured by the use of apolar solvents such as toluene. 71 The direct synthesis of primary amides by this methodology is more challenging due to the low nucleophilic nature of the nitrogen source, and the use of coupling reagents is oen required. The use of catalysts is an attractive approach for the direct formation of primary amides. The most relevant methodologies reported in this regard involve the use of enzymes such as lipases, [73][74][75][76][77][78][79][80][81][82] boric acids, [83][84][85][86][87][88][89] Group IV metals such as zirconium, titanium, 90,91 and heterogeneous catalyst as ZrOCl 2 $8H 2 O and CAN combined with microwave radiation (Scheme 1). Although, the latter methodologies have been reported to be difficult to reproduce. 90 The number of catalytic protocols reported for the synthesis of primary amides is still limited. 43,51,[92][93][94][95][96][97][98] In this work, we present a new protocol for the synthesis of amides from nonactivated carboxylic acids by direct coupling using a lowcost, readily available, and easy to manipulate catalyst and nitrogen source.

Results and discussion
Our initial investigation identied urea as able to transform phenylacetic acid (1) into 2-phenylacetamide (3) (Table S1, see ESI ‡). A catalyst screen was carried out to meet the abovementioned requirements (Table 1). Compared to the control, Group IV metals such as titanium and zirconium (Table 1, entries 2 and 3) improved the conversion, 90 whilst others such as Ni(NO 3 ) 2 , ZnCl 2 , iodide salts and protic acids showed little effect. Interestingly, Mg(OAc) 2 , imidazole and DMAP presented similar activities to those found with Zr(IV) and Ti(IV). Considering these results, and the low cost, availability and stability of magnesium salts and imidazole, the reaction conditions were further optimised using these two catalysts.

Magnesium salts as catalyst
Applying the conditions used in the initial catalyst screen, the most suitable solvent was determined (Table S2, see ESI ‡). Dipolar aprotic solvents, DMF and DMSO, showed low conversions, 71,99 whilst polar solvents such as CPME (cyclopentylmethylether), isoamyl alcohol and butyronitrile revealed reasonable conversions. The use of non-polar solvents such as toluene and octane showed the highest conversions into the corresponding amide. 100 These solvents enabled higher temperatures, and melting of the starting materials (1 ¼ 76 C and 2 ¼ 133 C) was observed to give a second liquid phase. This polar dispersed phase might dissolve the magnesium salt    catalyst, and the high concentrations in the droplets are expected to facilitate the reaction. [101][102][103] Octane was selected to screen other magnesium salts ( Table 2). All those tested showed a good catalytic activity with the best conversions achieved using Mg(OAc) 2 $4H 2 O, MgCl 2 and Mg(NO 3 ) 2 $6H 2 O. When magnesium acetate was used, detailed analysis of the crude reaction revealed the formation of acetamide as a by-product. 104 Mg(NO 3 ) 2 $6H 2 O was chosen as the most appropriate catalyst for further optimisation of the reaction conditions. The use of 2 equivalents of urea at 120 C were found to give an optimal 93% conversion to the amide (Table 3, entries 3). Three equivalents of urea had a detrimental effect on the formation of 2-phenylacetamide ( Table 3, entries 3 and 4). Varying the catalyst loading and reaction concentration, the best conditions iden-tied were using 10 mol% of Mg(NO 3 ) 2 $6H 2 O with a 1 M concentration in octane (Tables S6 and S7, see ESI ‡). The scope of the reaction was subjected to study. This methodology turned out to be effective for a wide range of aliphatic and phenylacetic acids (Table 4). Phenylacetic acids bearing electron-withdrawing and electron-donating groups were converted into the corresponding amides (3, 4, 5, and 6) in 81-87% isolated yields. The sterically demanding substrate diphenylacetic acid was also successfully transformed into diphenylacetamide (7) in 92% yield, and aliphatic acids (8, 9, 13,  14, and 15), with internal (16) and terminal (17) double bonds were also well-tolerated. Aliphatic acids containing conjugated double bonds (18) were more challenging substrates and lower conversions were observed even at elevated temperatures. The precursor benzoic acid (10) and substituted acids 11 and 12 showed very low reactivity under the reaction conditions, perhaps due to the delocalisation of electrons and subsequent decrease of electrophilicity of the carboxyl group. Surprisingly, heterocycles such as 2-picolinic acid and benzothiophene-2carboxylic acid gave the amides 19 and 20 in excellent yields. Furthermore, the hydroxyl group in glycolic acid was also tolerated to give 21 in 68% yield.
We envisaged that our methodology could be also applied to the synthesis of secondary amides. N-Methyl amides are commonly obtained by direct coupling with methylamine gas and few alternatives methods are available. [105][106][107][108] The use of N,N 0 -dimethylurea could be particularly useful, due to its availability and simpler handling. The scope of the reaction with this and N,N 0 -diphenylurea was tested with aliphatic and phenylacetic acids (Table 5). A wide range of aliphatic and phenylacetic acids was converted into the corresponding amides. For instance, N-methyl amides 22, 23, 24 and 25 were isolated in 89-96% yield. The method could be extended to aliphatic acids, and amides 26, 27, 29, 30 and 31 were obtained in good yields. On the other hand, less electrophilic benzoic acid (28) was unreactive under these conditions. Three acids were also tested with N,N 0 -diphenylurea giving satisfactory yields of N-phenylacetamides 33, 34 and 35.

Imidazole as catalyst
The initial screen showed the potential of imidazole and DMAP as catalysts for this transformation (Table 1). So, our attention was turned into the inexpensive and readily available imidazole organocatalyst. Using the previously determined conditions, phenylacetic acid and urea were reacted to give phenylacetamide (3) in 78% conversion (Table 6, entry 4). Further improvement was achieved with 1.5 equivalents of urea, but higher loadings failed to obtain better conversions. Increasing the temperature to 120 C led to the optimal reaction conditions ( Table 6, entry 5). 109,110 Using the optimised conditions, the substrate scope of the carboxylic acid was investigated (Table 6). Phenylacetic acid and hydrocinnamic acid proceed to their corresponding amides 3 and 8 in 91% and 97%, respectively. The presence of electrondonating and -withdrawing groups at the para position had little effect on the conversion, and amides 4, 5, 6, and 9 were obtained in high yields. Aliphatic groups were also welltolerated and hexanamide (13) was produced in 89% yield. The presence of bulky substituents in the aromatic or aliphatic chain had a detrimental effect, and diphenylacetamide (7) was produced in 65% yield, 27% less than with Mg(NO 2 ) 2 $6H 2 O. On the other hand, pivalamide (15) was obtained in 60% yield. Oleic acid gave 16 in 91% yield with no alteration of the double bond. Similar to the observations in Tables 4 and 5, the conjugated carboxylic acids did not perform well, with benzoic and cinnamic acids giving 10, 11, 12 and 18 in low conversions. 2-Picolinic acid and benzothiophene-2-carboxylic acid showed conversions into amides 19 and 20 of 65% and 100%, respectively, whilst glycolic acid was also converted into 21 in 61% yield, showing this catalyst also tolerates hydroxyl groups.
N,N 0 -Dimethylurea and N,N 0 -diphenylurea were also explored to synthesise secondary amides using imidazole (Table 7). In this case, 2 equivalents of urea and 130 C were required to drive the reaction towards the formation of the amide (Table S13, see ESI ‡). Both phenylacetic acid and hydrocinnamic acid gave 22 and 23 in 84% and 80% yields, respectively. In contrast, amidation of hexanoic acid into N-methylhexamide (29) gave only 70% yield, 19% less than the metal catalyst. Pivalic acid was converted into the amide 30 in only 38% yield, again indicating a steric problem. As with the other reactions, benzoic acid did not perform well in these conditions. Imidazole was an effective catalyst for making 2-picolinamide (32), obtained in 76% yield.

Mechanistic insights
A slow uncatalysed reaction between phenylacetic acid and urea was observed (Tables 1 and 2, entries 1), however, the addition of a suitable Lewis acid or an organocatalyst considerably improves the rates and conversions. Since the reaction mechanism was unclear, three models were proposed (Scheme 2): (1) decomposition of urea and direct amidation; (2) magnesium salt facilitates the formation of an N-acylurea intermediate which can be hydrolysed to produce the amide, and an unstable carbamic acid, followed by decarboxylation of the later. To explain the observed products, reaction with the more substituted urea nitrogen is required; (3) condensation of the carboxylic acid with imidazole to form N-acyl imidazolium, this activated amide would then react with urea to produce an N- According to mechanism (1), the decomposition of urea into ammonia and CO 2 is reported to take place at temperatures above 152 C; 111 even though the temperatures in our reactions are not as high as these, the possibility of a catalysed urea degradation was investigated. Urea was treated with either imidazole or Mg(NO 3 ) 2 $6H 2 O under the reaction conditions for 24 hours. The analysis of the reaction crudes by 1 H NMR showed the presence of urea indicating no degradation. Furthermore, gravimetric analysis, before and aer reaction, gave 96% urea recovery, suggesting no degradation had taken place. The reaction was repeated with N-methylurea, N,N 0dimethylurea and N,N 0 -diphenylurea, and in each case >92% of the starting material was recovered. To investigate this further, phenylacetic acid (1) was exposed to the reaction conditions using aniline as the nitrogen source, which might be formed during the degradation of N,N 0 -diphenylurea (Scheme 3, eqn (1)). Similar conversions were observed when the reaction was carried out with and without imidazole or magnesium catalyst, suggesting it is not involved in the direct coupling. 112 However, when phenylacetic acid (1) was reacted with N,N 0 -diphenylurea, the uncatalysed reaction gave no product, but in the catalysed reactions conversions of 40-51% were observed (Scheme 3, eqn (2)); notably less than direct amidation (70-77% conversion) but expected, as anilines are well known to undergo direct amidation. 72 A previous report describes the CO 2 evolved in the acidolysis of ureas using 13 C-labelled carboxylic acids derives from the urea, and support other research that invokes a carbamic acid intermediate. 113,114 Mechanisms (2) and (3) were tested with unsymmetrical substituted ureas. The highly hindered N,N,N 0 N 0 -tetramethylurea did not form the tertiary amide 40 even at 130 C, indicating that the steric bulk interferes (Scheme 4, eqn (1)). However, the reaction between phenylacetic acid and N-methylurea or N,N-dimethylurea gave the secondary amides 22 an 40 in 66% conversion with MgNO 3 $6H 2 O and 77% with imidazole, along with traces of the primary amide (Scheme 4, eqn (2) and (3), Table S14, see ESI ‡). In order to check the formation of ammonia during the reaction a litmus paper test was conducted. A colour change from yellow to blue was observed conrming the formation of a basic gas. These results might indicate that unsymmetrical ureas react to give the most Table 7 Substrate scope for the formation of primary and secondary amides from carboxylic acids using imidazole a,b 91 (100 substituted amide, and presumably carbamic acid which decomposes liberating carbon dioxide and ammonia. 113,114 Water is required for this reaction and may come from the MgNO 3 $6H 2 O, or during condensation of acid with the imidazole. In either case a thermally unstable (N-alkyl)carbamic acid is implicated, 115 (Scheme 2, eqn (2) and (3)). MgNO 3 $6H 2 O might coordinate to the 1,3-dicarbonyl, activating it to urea condensation. 116 With imidazole, the direct reaction with urea was discounted, however its reaction with carboxylic acids and esters, in solvent under thermal conditions is reported. [117][118][119] Protonated N-acyl imidazoles are known to react with amines and thiols, so it is reasonable that urea, may react to form the Nacylurea intermediate. 115,120 The reaction mixture was analysed by HRMS aer 6 hours and a species with m/z of 201.0636 was found, which corresponds with the sodium adduct of the N-acylurea intermediate (theoretical m/z 201.0640 [M + Na] + ). To test the reactivity of the N-acylurea intermediate, N-pivaloylurea was synthesised and subjected to different reaction conditions, 121 and conversion to pivalamide was analysed by 1 H NMR (Table 8).
In the absence of the catalyst and water, only the starting material was recovered ( Table 8, entry 1). The addition of two equivalents of water produced the pivalamide with 6% conversion (Table 8, entry 2). The presence of the catalyst did not improve the reaction outcome (Table 8, entries 3 and 4), whilst either catalyst alone failed to improve the conversions (Table 8, entries 5 and 6). However, when the catalyst and water were combined, 25% and 14% of amide were detected (Table 8, entries 5 and 6). The conversions are less than pivalic acid and urea, (55% and 60% to 15, Tables 4 and 7). To dismiss a substrate effect, N-phenylacetylurea was synthesised and subjected to the same study, and similar behaviour was observed. 121,122 Since N-alkylcarbamic acids decompose readily in acidic media, 114 1 equivalent of hydrocinnamic acid was added, but also gave similar low conversions. These results suggested that other physico-chemical effects may be playing a role. Solubility studies indicated that phenylacetic acid is soluble in octane at 120 C while urea and 2-phenylacetamide (3) are not. When all the starting materials were mixed together in octane at 120 C a biphasic system was obtained, and at the end of the reaction a white solid corresponding to 2-phenylacetamide (3) was formed. In these conditions an onsolvent system may be occurring in which the urea, carboxylic acid and catalyst are at high concentration and form a polar, hydrogen bonded structure that facilitates the reaction. 101,102 Besides, the liberation of ammonia and CO 2 along with the precipitation of 2-phenylacetamide (3) might be the driving force of this transformation. The lack of solubility of the Nacylurea intermediate in octane could explain the slow reactivity observed in Table 8.
To investigate further the role of imidazole in this transformation, N-1-methylimidazole and 2-methylimidazole were used as catalysts (Table 9). A signicant decrease in conversion, from 86% to 37%, was observed when 2-methylimidazole was employed (Table 9, entry 4). Whereas, N-1-methylimidazole gave 66% conversion, only slightly less than imidazole (86%) ( Table  9, entry 3). These results suggest that the mode of activation is through the nitrogen lone pair, as the 2-methyl group blocks this position. This is supported by the similar catalytic activity of DMAP (Table 9, entry 5), that also has a sp 2 nitrogen with a lone pair of electrons. The mechanism by which the reaction proceeds is still unclear and further studies are still undergoing to understand the reaction pathway.

Conclusions
A new method for the direct synthesis of primary amides from nonactivated carboxylic acids has been described, in which wasteful activating reagents are avoided. The methodology reports the use of cheap and readily available catalysts such as Mg(NO 3 ) 2 $6H 2 O and imidazole, with urea as an atom-efficient nitrogen source. The process has been shown to produce not only primary, but secondary and a tertiary amides from readily available ureas. The method shows a broad scope of reaction, although conjugated carboxylic acids do not perform well. The reaction mechanism has been studied, and initial results point to the involvement of an N-acylurea intermediate, although, the pathway of its decomposition to the product remains unclear. Further studies are still undergoing to propose a more plausible mechanism.

Conflicts of interest
There are no conicts to declare.