The thermal and boron-catalysed direct amide formation reactions: mechanistically understudied yet important processes
Hayley Charvillea, David Jacksonb, George Hodgesc and Andrew Whiting*a
aDepartment of Chemistry, Science Laboratories, Durham University, South Road, Durham, UK DH1 3LEbSyngenta AG, Process Technology, Muenchwilen, SwitzerlandcSyngenta, Jealott's Hill International Research Centre, Bracknell, Berkshire, UK RG42 6EYFirst published on the web 9th February 2010
Despite the amide formation reaction being one of the key cornerstone reactions in organic chemistry, the direct amide formation is both little used and little explored. Acceptance of the feasibility and general applicability of the reaction depends upon the ability of researchers to bring it into the mainstream by development of: (1) an understanding of the mechanism of the reaction; and (2) the design of catalysts which promote the reaction on a wide range of substrates and under ambient conditions. From the earliest report of the direct amide formation in the 19th century, there have been relatively few reports of mechanistic studies, though it is clear that there is not a simple relationship between ease of direct amide formation and the pKa of the carboxylic acid and amine, or whether salt ammonium carboxylate formation is important. Consequently, direct amide formation has historically been run under higher temperature conditions. However, more recently, stoichiometric and catalytic boron compounds have been developed that considerably reduce the reaction temperatures under which direct amide formation will proceed. Limited attempts at mechanistic studies point to the formation of acyloxyborate or boronate species acting essentially as mixed anhydrides, though the exact order of these systems remains to be categorically determined.
Introduction
Environmentally benign alternatives to standard chemical processes are increasingly sought after for a wide range of important chemical conversions.1 To this end, there is a significant demand for catalytic solutions for many reactions, especially if the desired product can be obtained in high yield, with high atom efficiency2 and hence, a reduced E-factor3 (defined as the mass ratio of waste to desired product). In particular, the use of stoichiometric reagents, such as condensing and activating agents, is atom uneconomic, and should ideally be avoided if possible and replaced by catalytic processes, resulting in lower costs and reduced impact on the environment. Condensation reactions in general are prime candidates for the development of environmentally benign solutions, and amide bond formation in particular is a key chemical transformation that needs a general environmentally benign solution.Amide bonds are typically formed from amines and pre-activated carboxylic acid derivatives, such as acid chlorides (using thionyl or oxalyl chloride, for example), anhydrides, or by using the carboxylic acid directly with coupling reagents such as carbodiimides.4–7 All of these methods possess considerable drawbacks, not only because they represent poor atom economy, but also because they all involve the use of toxic reactive reagents and can complicate product purification.5 The ubiquity of the amide bond in natural product structures and materials, combined with its importance in industrial and pharmaceutical chemistry means there is a major drive to develop clean, catalytic, ambient and efficient processes for amide bond formation. The most desirable solution to this problem of amide formation is the direct condensation, i.e. by direct reaction between an amine and a carboxylic acid. This is generally thought to be difficult due to the formation of unreactive carboxylate-ammonium salts. However, direct condensation has been known since 1858 and the reaction was even carried out under flow conditions.8 Due to the forcing conditions employed for these early reactions, it is still generally assumed that this type of procedure is severely limited, mainly due to the thermal stability of the reaction components and products under the reaction conditions.
 | (1) |
In the case of direct amide formation, full mechanistic details have yet to be elucidated, although some proposals have been made. This review provides a summary and discussion of work reported to date regarding the mechanistic aspects of the catalysed and uncatalysed direct amide formation reaction.
Amide formation via pyrolysis of amines and carboxylic acids and their ammonium carboxylate salts
Carboxylic acids and amines can react together to form salts, strong heating of these salts can lead to amide formation, which is the case for the conversion of ammonium acetate to acetamide.9 The reaction is carried out in hot acetic acid with continuous distillation to remove excess acetic acid and water, which ensures that the reaction is driven towards the amide. Other examples of this method of amide formation include the preparation of benzanilide10 by heating benzoic acid with an excess of aniline and also the preparation of succinimide from heating ammonium succinate.11Mitchell and Reid8 found that aliphatic amides could be synthesised by passing ammonia gas through a number of carboxylic acids at a temperature that allowed the removal of water to be continuous. Dimethyl amides were produced by the same method, with ammonia being replaced by dimethylamine. A drawback of this method became evident during the examination of carboxylic acids with a longer alkyl chain length than butyric acid; the rate of the reaction decreased considerably and the temperature at which the acid was heated to was sufficient to dehydrate the longer chain amides to nitriles. No amide was produced from either hexadecanoic acid or octadecanoic acid when heated to 125 °C or 190 °C for considerable lengths of time.8 This approach has been used as the basis of a commercial synthesis12,13 of dimethylacetamide from the reaction of acetic acid and dimethylamine with the product being removed as an azeotrope (84% amide, 16% acetic acid, b.p. 170.8–170.9 °C). Interestingly, if a substantial excess of dimethylamine is reacted with acetic acid under conditions of elevated temperature and pressure, a reduced amount of azeotrope is formed; optimum temperatures of 250–325 °C with pressures in excess of 6200 kPa are required.
It has been reported that the pyrolysis of monocarboxylic acids in the absence of any other reagent leads to the formation of the anhydride.14 Several carboxylic acids were shown to undergo dehydration when refluxed at temperatures between 250–350 °C to afford the corresponding anhydride. It is, therefore, not surprising that upon the addition of an amine to the heated carboxylic acid, amides can be formed since a conventional synthesis of amides is via the anhydrides,4 however, the intervention of an anhydride in these reactions has not been reported.
In 1993, the preparation of a range of amides by heating a mixture of different amines and carboxylic acids in the absence of any catalyst was reported.15 The optimum conditions for the heating of carboxylic acid–amine mixtures was found to be heating between 160–180 °C for 10–30 min. There are many advantages to preparing amides in this way: the procedure is simple; no catalysts or solvents are required; and, reaction times are short. Despite these advantages, there are several drawbacks to the method: both the amines and carboxylic acids used need to be thermally stable; have a melting point below 200 °C; be non-volatile; and have a high boiling point. Extreme heating can lead to tar formation, whereas not heating to a high enough temperature means that the reaction does not go to completion. For these reasons, this method has not developed into a widely-used method for direct amide formation and may not be amenable for small scale reactions and high value reactants. A recent communication16 has shown that neat mixtures of amines and carboxylic acid are readily combined in the presence of 3 Å molecular sieves at between 120 and 160 °C, to give generally good to high yields of amide, with the odd exception of N-methylbenzylamine reacting with benzoic acid, which provides only traces of amide after 24 h at 160 °C (compared with 75% using benzylamine).
The use of microwave irradiation has been reported to simplify and improve a number of organic reactions, often leading to higher conversions and shorter reaction times.17,18 Preparation of amides by heating of carboxylate ammonium salts obtained from the mixture of an amine and carboxylic acid has been examined under microwave irradiation conditions in the absence of a catalyst and of solvent (Scheme 1).19
 | ||
| Scheme 1 Preparation of amides from amines and carboxylic acids under microwave irradiation. | ||
Heating the acid–amine mixture to 150 °C using microwave irradiation resulted in increased yields of the amides in comparison to conventional heating for the majority of combinations tested. For example, benzylamine reacted with benzoic acid, affording the corresponding amide in high yield (80%) after 30 min and using 1.5 equivalents of amine. However, when the same reaction was heated using an oil bath, only 8% yield of the amide was isolated. The reaction of n-octylamine with phenylacetic acid (1
:1 molar ratio) provided the amide in 54% yield; an improvement on the conventionally heated reaction (23%). Significant differences in reactivity were observed according to the structure of the amine and carboxylic acid reagents. The following sequence was established for the reactivities of the carboxylic acids used: n-C9H19CO2H ≥ C6H5CH2CO2H ≫C6H5CO2H, where for the amines used, the reactivities were in the order: C6H5CH2NH2 > n-C8H17NH2 > p-CH3OC6H4NH2 > C6H5NH2. Different factors were considered to try and explain the relative reactivities, including the relationship between reactivity and pKa of each substrate. Similar behaviour between carboxylic acids with similar pKa values would be expected, along with there being a large influence depending on the amine structure, the aliphatic amines have pKb values between 3–5 and the aromatic amines have pKb values between 8–10. The three carboxylic acids used in this study had similar pKa values (4.89, 4.31 and 4.21 for n-C9H19CO2H, C6H5CH2CO2H, C6H5CO2H, respectively). The experimental results from this study showed that there was not a simple relationship between reactivity and pK values and, therefore, the influence of the position of acid–base equilibrium upon reactivity was rejected as being an important factor.19
For several of the substrate combinations, it was claimed that having either the amine or carboxylic acid in excess was favourable. This was particularly evident in the case of benzylamine reacting with benzoic acid (see Table 1).
Table 1 Reaction of benzylamine with benzoic acid at 150 °C under microwave activation
| Reaction time/min | Relative ratio (acid:amine) | Yield (%) |
|---|---|---|
| 30 | 1:1 | 10 |
| 30 | 1:1.5 | 80 |
| 30 | 1.5:1 | 75 |
The simplest explanation claimed for the above observations was that the excess carboxylic acid or amine complexed to the carboxylic acid by hydrogen bonding to the carbonyl group. This results in electronic assistance to nucleophilic attack by the amine nitrogen atom [Fig. 1(a) and (b)]. If this is indeed likely to occur, then another, perhaps more convincing possibility, is that the protonated amine could also provide assistance [as in Fig. 1(c)]. It is generally assumed that the mixing of an amine and carboxylic acid results in the formation of an ammonium carboxylate salt, which could have a role in assisting the attack of the nucleophile, i.e. by protic acid catalysis. However, this explanation fails to take into account that the pKas between (a), (b) and (c) (Fig. 1) would be quite different. Perhaps another plausible theory is that the addition of an excess of either the carboxylic acid or amine starting material would help to drive the reaction towards product formation, according to Le Chatelier's Principle (assuming the reaction is not effectively irreversible under the reaction conditions).
 | ||
| Fig. 1 Electrophilic assistance to nucleophilic attack. | ||
It has also been claimed that the addition of one equivalent of imidazole to a mixture of benzylamine and benzoic acid facilitates the amide formation reaction and leads to increased yields (from 13 to 61%) under microwave conditions.20 The imidazole is claimed to act by assisting the carbonyl group in a similar manner to the excess carboxylic acid or amine as portrayed in Fig. 1.
In the absence of solvent interactions (non-polar reaction conditions), π–π interactions might be expected to play an important role when both the carboxylic acid and amine possess an aromatic ring (Fig. 2). It has been speculated that π–π interactions might stabilise an amide formation transition state, leading to a decrease in activation energy and, therefore, a higher conversion of starting materials into product.19 However, there is no direct evidence to date that such π–π interactions do directly affect direct amide formations.
 | ||
| Fig. 2 Optimal π–π interactions | ||
If π–π interactions are important in some direct amide formation reactions, it is also possible that edge-to-face arrangements21,22 may seem more likely than that suggested in Fig. 2, however, it was reported that MP3 molecular modeling studies (a method that is adapted to describe hydrogen bonding and the interactions between non-bonding atoms) using the Hyperchem program predicted that the most stable arrangement of N-benzyl-2-phenylacetamide involves the overlapping of the π-systems. This has yet to be confirmed by more detailed modeling studies.
The use of stoichiometric boron reagents: acyloxyboron intermediates
As early as 1965, it was reported that boron containing compounds could be used as reactive entities with certain functional groups, in particular for converting carboxylic acids to amides.23 The mixing of carboxylic acids with trisdialkylaminoborane compounds [B(NR′2)3] 4 in an inert solvent produced a considerable exotherm. Depending on the carboxylic acid used, either cooling or refluxing of the reaction mixture afforded the amide products. It was found that only one of the dialkylamino groups was employed for the conversion of the carboxylic acid in the amine. Hence, for the reaction to go to completion, one equivalent of the boron reagent was required without any additional catalyst present. A suggested mechanism for this transformation is shown in Scheme 2, i.e. that the reaction proceeds via initial salt formation followed by the production of mixed anhydrides. Mixed anhydride 5 being formed following nucleophilic attack by the carboxylate group on the protonated trisdialkylaminoborane. | ||
| Scheme 2 Postulated mechanism for amide formation from the mixing of carboxylic acids with trisdialkylaminoboranes.23 | ||
This method is just one example, which demonstrates that activated carboxylic acids, such as acyl chlorides, do not need to be prepared and are not essential for direct amide formation. The conditions for this transformation are mild and the desired amide can be produced in high yield and selectivity.23,24 A disadvantage of this process is, however, that only one of the amino groups is utilised during the reaction.
Further studies demonstrated that other boron containing compounds containing the general structure 7 could facilitate amide formation.24 Notably, acyloxydialkylboranes [(RCO2B(OR′)2] were suggested to be good candidates for direct amide formation which fit the general structure 7.
 | ||
One possible route to acyloxydialkylboranes of type 7 is outlined in Scheme 3, i.e. from the addition of dimethyloxyboron chloride 9 to the sodium salt of a carboxylic acid 8. This reaction proceeded rapidly and there was infrared evidence for formation of the corresponding acyloxydialkylborane species 7.24,25
 | ||
| Scheme 3 Formation of acyloxydialkylborane 7 and the corresponding amide. | ||
Upon the addition of amine at room temperature to the acyloxyborate derivative 7 (Scheme 3), the corresponding amide 3 was generated. Several combinations of amine and alkyloxyborane species were tested and yields were consistently below 50%. This was increased to approximately 70% by heating the reaction mixture.24 The reaction conditions were mild, and when applied to peptide synthesis, racemisation was low. For these reasons, further investigations were carried out to establish why low conversions were obtained at room temperature. Possible reasons were proposed as follows: firstly, the dialkyloxyborane species was not being formed efficiently or was undergoing further reaction before the addition of the carboxylic acid; secondly, the mixed anhydride of type 7 was susceptible to decomposition or reduction; thirdly, the reaction of the mixed anhydride with an amine occurred at the boron atom or resulted in products that could not participate further in the reaction.25 Investigations into each of these options showed that the third possibility was most likely. One mole of alcohol can be liberated in the attack of an amine upon the mixed anhydride species 7 [Scheme 4(a)]. This can also go on to compete with the amine for the mixed anhydride and result in the formation of an unreactive salt 10, i.e. as outlined in Scheme 4(b).
 | ||
| Scheme 4 (a) Liberation of alcohol; (b) formation of an unreactive carboxylate ammonium salt. | ||
Addition of isopropanol to caproyloxydiisopropoxyborane rapidly produced the free acid, with no evidence of the ester being formed.
Since alcohols can attack acyloxydialkyloxyboranes 7 at boron, it is possible that nucleophiles such as amines will also attack at boron, although less selectively. This means that as well as the amide being produced, as in Scheme 4(a), the aminodialkyloxyborane species 11 could also be produced along with an unreactive carboxylate ammonium salt (Scheme 5). This presents a likely explanation for the low conversion to amide at room temperature by this method, since it is a competitive reaction with the amide formation.25
 | ||
| Scheme 5 Formation of an aminodialkyloxyborane species. | ||
The use of stoichiometric boron reagents: borane and catecholborane
Trapani et al.26 employed a borane trimethylamine complex in a 1:1:3 molar ratio for the amine:borane:carboxylic acid, resulting in good to high amide yields under refluxing xylene conditions. Though lacking direct evidence, it is claimed that triacyloxyborane species are the activated acylating species involved in the reaction.Closely related to this work, Ganem et al.27 reported that carboxylic acids and amines condense readily to form amides in the presence of stoichiometric amounts of catecholborane 12, but under much milder conditions, i.e. THF, −78 °C to room temperature as outlined in Scheme 6.27 The condensation is claimed to occur via the intermediate 2-acyloxy-1,2,3-benzodioxaborolane, for which there is some infrared evidence (1740 cm−1 C
O absorption). The reaction is carried out with two equivalents of amine, since it proceeds via nucleophilic attack of the amine to give [13·amine] (see Scheme 6).
 | ||
| Scheme 6 Ganem's amide formation using catecholborane as condensing reagent. | ||
Ganem's method for direct amide formation requires the addition stoichiometric amounts of catecholborane; however, in 2006 Yamamoto proposed a possible catalytic scheme28 based on this work (vide infra).
As mentioned previously, mixed anhydride intermediates of type 7 can be destroyed by the one equivalent of alcohol formed during the reaction (Scheme 4). For the intermediate 25, formed when using catecholborane, this does not occur. The aromatic ring in this system enhances the reactivity of the active ester 25 and can reduce side reactions. Following the addition of an amine, the corresponding amide can be produced with high yields. The reaction between nonanoic acid and benzylamine in the presence of catecholborane afforded the amide with a yield of 92%.27
Resin-bound catecholborane can also be used for direct amide formation, as a solid phase reagent (eqn (2)).29 When shaken with the carboxylic acid, the activated mixed anhydride acyloxyborane is claimed to be formed. Upon addition of an amine under ambient conditions, the amide is formed in modest yields. For example, the reaction of octanoic acid with benzylamine gives the amide in a yield of 54%.
 | (2) |
Catalytic direct amide formation
It is desirable to be able to perform direct amide condensation reactions between equimolar amounts of carboxylic acids and amines under mild, catalytic conditions. Examples of amidation catalysts in the literature are largely based around boric acid or arylboronic acids possessing electron withdrawing substituents.In 1996, Yamamoto found that benzeneboronic acids bearing electron-withdrawing groups at the meta- or para-position are highly efficient catalysts for direct amide formation in less polar solvents.30 Unlike the boron-mediated amidation mentioned previously, arylboronic acids with electron-withdrawing groups can overcome the problem of transformation into an inactive species or it is possible that they are in equilibrium with these species. Arylboronic acid catalysts are water-, acid- and base-tolerant Lewis acids that can generate acyloxyboron species, enhanced by the Lewis acidity and resulting in an increased reactivity with amines. 3,4,5-Trifluorobenzeneboronic acid 14 and 3,5-bis(trifluoromethyl)benzeneboronic acid 15 catalyse direct amide formation in toluene at reflux with the azeotropic removal of water by 4 Å molecular sieves in a Soxhlet thimble.
 | ||
Catalyst 14 (1 mol%) was found to be the most active, and for the reaction of benzylamine with 4-phenylbutyric acid, produced the amide in 96% yield in 18 h. For more demanding substrates, more forcing conditions were employed. Replacing benzylamine with aniline required refluxing mesitylene (b.p. 163–166 °C) and the yield of amide was 99% after 4 h. A proposed mechanism is shown in Scheme 7.30
 | ||
| Scheme 7 Proposed catalytic cycle for direct amide formation.30 | ||
Arylboronic acids contain varying amounts of cyclic trimeric anhydrides (boroxines). On heating 4-phenylbutyric acid with 16 (2:1 mixture) in toluene-d8 with removal of water for 2 h, the monoacyloxyboronic acid 17 was claimed to be produced. Further investigations are required to determine whether monoacyloxyboronate species are the active acylating species, since the reported 1H NMR and IR data (see ref. 23, Supplementary Information) is inconclusive without corroboration from 11B NMR. In addition, a much more activated carbonyl stretching frequency than the 1586 cm−1 absorption claimed to be an monoacyloxyboronate reported (compare with the Ganem et al.27 monoacyl species at 1740 cm−1vide supra) might be expected, since the unactivated starting carboxylic acid is reported to have an IR absorption at 1709 cm−1. No evidence for the detection of diacyloxyboronate derivatives was reported. However, upon addition of benzylamine to a toluene solution of 17, the corresponding amide was produced at room temperature, with up to 50% conversion achieved. It was suggested that the reaction stopped because the intermediate 17 can be decomposed by hydrolysis with water. The results presented suggested that the rate-determining step for this catalysed reaction was the formation of monoacyloxyboronate intermediate 17,30,31 though the intervention and requirement for more activated species cannot be ruled out.
3,4,5-Trifluorobenzeneboronic acid 14 is also an effective catalyst for the polycondensation of carboxylic acids and amines.32 Direct polycondensation is desirable both environmentally and industrially. The direct polycondensation of adipic acid and hexamethylenediame (eqn (3)) was examined for the formation of nylon-6,6. With 10 mol% of catalyst 14 at reflux in o-xylene and removal of water by 4 Å molecular sieves, the desired product was formed with a yield of 89% after 20 h. The number-average molecular weight (Mn) and the weight-average molecular weight (Mw) of the nylon obtained were estimated to be 2680 and 8330, respectively. These values were increased to 4690 and 22400, respectively on changing the solvent to a 1:3 mixture of m-cresol and o-xylene.
 | (3) |
 | ||
Although arylboronic acids 14 and 15 were more active than 18, catalyst 18 can be fully recovered by extraction with fluorous solvents. In a 1:1:1 mixture of o-xylene, xylene and perfluorodecalin under azeotropic reflux, amide formation reactions can be carried out with 3 mol% of catalyst 18. Upon cooling, the two heterogenous phases can be separated and the amide can be isolated in a quantitative yield. The catalyst can be recovered from the fluorous phase and reused with no loss of activity.33
Yamamoto et al. have also found that the compounds 14 and 15 will catalyse the condensation of carboxylic acids with ureas (eqn (4)).34N-acylcarbamates play an important role in medicinal chemistry, and the direct condensation of carboxylic acids and ureas can occur in the presence of excess strong acids, such as chlorosulfonic acid as an example. This method is environmentally undesirable and the use of excess reagents is preferably avoided. The addition of 5 mol% of 14 or 15 to an equimolar mixture of carboxylic acid and urea derivatives (which are less nucleophilic than the amines previously screened) results in the formation of N-acylureas in high yields.
 | (4) |
 | ||
| Scheme 8 Proposed scheme for direct amide formation using catalytic boric acid. | ||
A disadvantage of these arylboronic acid catalysts is that their activity is greatly reduced in polar solvents, which has restricted the scope of substrates that can be used. A polar solvent tolerant catalyst that has been demonstrated to be successful for direct amide formation is N-alkyl-4-boronopyridinium iodide 20. This compound is much more active than the other arylboronic acid catalysts when direct amidation is carried out in solvents such as anisole, acetonitrile and N-methylpyrrolidinone (NMP).35–37 Catalyst 20 can be reused through the use of ionic liquid–toluene biphasic solvents. A resin bound version of the catalyst has also been developed. N-Polystyrene resin-bound 4-boronopyridinium salts 21a–d have been produced as heterogenous catalysts for direct amidation without the need for ionic liquids.
 | ||
In 2005, Tang reported that cheap and readily available boric acid alone was an efficient catalyst for direct amide formation.38 Benzylamines and cyclic aliphatic amines, such as piperidines, reacted smoothly with a B(OH)3 loading of 5 mol% leading to excellent yields. It was proposed that boric acid reacts with the carboxylic acid to form a mixed anhydride as the actual acylating agent.38 Upon the addition of the amine, the desired amide is formed and boric acid is regenerated (Scheme 8). However, the intervention of tetraacyldiborate species such as 2239–41 cannot be ruled out in these types of reactions.
This amidation procedure has been employed in the synthesis of several active pharmaceutical ingredients.42,43
In 2006, it was reported that 4,5,6,7-tetrachlorobenzo[d][1,2,3]dioxaborol-2-ol 23b was effective as a catalyst for the amide condensation of sterically demanding carboxylic acids.28
 | ||
Taking into account the work by Ganem discussed previously,27 it was found that compounds 23a–b could be used to catalyse direct amidation with 5 mol% catalyst loading in toluene or o-xylene, with water removal achieved by azeotropic reflux. The proposed catalytic pathway for this transformation is shown in Scheme 9.
 | ||
| Scheme 9 Proposed catalytic scheme by which the Ganem procedure27 might be transformed into a catalytic process.28 | ||
For the reaction between 4-phenylbutyric acid and benzylamine in toluene with azeotropic water removal, catalyst 23b yielded the desired amide in 93%. Under the same conditions, when boric acid was used as the catalyst, the amide yield was reduced to 31%. For a more sterically demanding combination of substrates, the direct condensation of cyclohexanecarboxylic acid with benzylamine was carried out using catalyst 24b and boric acid. This afforded the amide in 62% and 2% yield, respectively.
4,5,6,7-Tetrachlorobenzo[d][1,2,3]dioxaborol-2-ol 23b displays a similar catalytic activity to 3,5-bis(trifluoromethyl)benzeneboronic acid 15. Since boric acid is readily available at a low price, 23b can be prepared relatively cheaply.28,42
Kinetic studies published in 2006 from this group44 demonstrated that direct formation of amides from amines and carboxylic acids does occur in the absence of a catalyst under relatively low temperature conditions and is highly substrate dependant. For less reactive carboxylic acids, the presence of a boronic acid (such as 14, 24 or 25) or boric acid catalyst greatly improves the yield of amide produced.44 Experiments combining different carboxylic acids and amines were carried out in refluxing toluene or fluorobenzene with water removal in the presence of 3 Å molecular sieves (Soxhlet). Importantly the thermal reaction was also compared with the addition of 1 mol% of several catalysts.
 | ||
After 22 h, the uncatalysed reaction between 4-phenylbutyric acid and benzylamine produced the corresponding amide with a yield of 60%. The performance of the catalysed reactions showed only a slight improvement, with boric acid, 14 and 25 being close to identical. In order to study only the catalytic effects, conditions that minimised the thermal reaction were employed. The reactions were carried out in refluxing fluorobenzene (85 °C) with an increased catalyst loading (10 mol%) and showed significant improvement over the thermal reaction.
For the more reactive combinations of amines and carboxylic acids, catalyst choice was not the most important factor. However, for less reactive substrates the bifunctional catalyst 25 was advantageous.45 For example, the reaction between benzoic acid and benzylamine in the presence of 25 gave the desired amide in a yield of ca. 80% over 48 h. Clear evidence of bifunctional activity assisting substrate dependent reactions, particularly at low temperatures, was demonstrated. Studies to identify the intermediate species in the catalytic pathway were also carried out, and mass spectrometry provided evidence for existence of boroxine 26, diboronate 27 and diacyloxyboronate 28 species, amongst others. On the basis of these observations, the following mechanism was proposed (Scheme 10).44 It is noteworthy that it is not clear to date whether there is an equilibrium between the amine and carboxylic acid versus ammonium carboxylate salt, or indeed if this plays a role in the amide formation reaction.
 | ||
| Scheme 10 Proposed mechanism for amide formation involving boric or arylboronic acid catalysis. | ||
It is thought that the reaction proceeds through an intermediate, which is likely to be the mono- or di-acylated species, i.e.17 or 28, respectively. For the uncatalysed, thermal reactions, the likely intermediate is the anhydride formed by thermolysis, which explains the subsequent formation of the amide. It is suggested that the formation of this anhydride is rate-limiting and, therefore, by comparison with the catalysed reactions the carboxylate activation (formation of the acyloxyboronte species, whether mono- or di-acylated) is rate determining. In addition, design of experiment (DoE) studies undertaken by Whiting et al.45 examined four factors to identify the ideal reaction conditions for catalysed direct amide formation. The results of these reactions carried out in refluxing fluorobenzene showed that the most important factors were catalyst loading, time and a two-factor interaction between time and catalyst loading. Concentration and excess acid or amine had no significant effect; this is in direct contrast to the work carried out by Loupy et al.19 who reported that amide yield is significantly increased in the presence of excess carboxylic acid or amine (vide supra). In addition, Whiting et al.45 reported that upon changing the experimental design to use toluene, catalyst loading was still the most important factor, followed by time. However, catalyst loading was involved in an interaction with concentration so that at higher catalyst loadings and lower concentration, an increased amide yield resulted. Carboxylic acid stoichiometry was still insignificant. The conclusions from the concentration effects suggested that the reaction rate was limited by the rate at which water can be azeotroped from the reaction mixture. The effect of not drying the reaction mixture led to zero order kinetics and a slower reaction, further corroborating the need to remove water and that the dehydration process is significantly rate limiting. This reinforces the idea that the acylation step is likely to be fast relative to the rate of physical water removal and that the amide formation is irreversible under the reaction conditions.
The first report of an asymmetric direct amide formation via kinetic amine resolution appeared in 2008.46 Previously there had been no reports of direct amide formation involving asymmetric induction. This is because the process is generally regarded as one that requires high temperatures, which would lead to reactant or product degradation. Also, asymmetric induction processes are generally more efficient at lower temperatures as the effects of small differences in energy between diastereoisomeric transition states are amplified. However, it is reported that asymmetric processes are possible under elevated temperatures (refluxing fluorobenzene) in the presence of a planar, chiral ferrocene based bifunctional amino–boronic acid catalyst 30. This catalyst can induce the kinetic resolution of racemic α-substituted benzylamines through direct amide condensation with achiral carboxylic acids (eqn (5)).
 | (5) |
In 2008, Hall and co-workers reported direct amide formation reactions carried out at 50 °C and at room temperature in the presence of a 2-halo-benzeneboronic acid catalyst.47 These catalysts are more active than 3,4,5-trifluorophenylboronic acid 14, which requires temperatures of 120 °C. For the reaction of phenylacetic acid and benzylamine with water removal by 4 Å molecular sieves and a reaction temperature of 25 °C, the presence of catalyst 14 (10 mol%) resulted in the amide being formed with a yield of 42% in dichloromethane. When the catalyst was changed to 2-iodobenzeneboronic acid 31 the amide yield increased impressively to 91%.
 | (6) |
2-Iodobenzeneboronic acid is the most active of the 2-halo-benzeneboronic acids. The importance of the ortho-substituent is confirmed by the poor activity of the para-isomer. When both ortho-positions are occupied the effectiveness of the catalyst is reduced and so confirms the importance of one unsubstituted ortho-position. Inductive effects do not account for the activity of 31 because of the reverse trend observed in the o-halide series (I > Br > Cl > F). The size and electron density of the iodo-substituent results in the distortion of the B–C–C bonds of the boronic acid. Therefore, it is possible that electronic or structural effects are important to the trend in activity observed here.47
Conclusions
Considering the long history of the direct amide formation, it is remarkable how little understood the process is. This review of the published literature reveals how incomplete our understanding of the mechanisms are for both the uncatalysed and catalysed direct amide formation reactions, and prompts the chemical community that detailed studies are urgently needed. A fundamental understanding of the basic kinetics and thermodynamics of direct amide formation will facilitate, and there remains the need for, the development of clean, sustainable processes for direct amidation which will work under ambient conditions. Clear steps have been taken in this direction, especially with the boron-based catalysts. However, an understanding of their mode of action is likely to be a prerequisite to developing systems which are effective on a wide range of substrates. A better understanding of the impact of amine and carboxylic acid versus ammonium carboxylate, and exactly which of these species are reactive with precisely which activating species, in the reaction is required. The uncovering of these mechanistic details will inevitably lead to major advances in catalyst design, new applications including larger scale industrial process application, and more importantly, universal acceptance that direct amide formation is a viable and important, general reaction.Acknowledgements
We thank Syngenta AG for funding and the Royal Society of Chemistry for Journals Grant (to AW).Notes and references
| 1 | D. J. C. Constable, P. J. Dunn, J. D. Hayler, G. R. Humphrey, J. L. Leazer, R. J. Linderman, K. Lorenz, J. Manley, B. A. Pearlman, A. Wells, A. Zaks and T. Y. Zhang, Green Chem., 2007, 9, 411–420 | |
| 2 | B. M. Trost, Acc. Chem. Res., 2002, 35, 695–705 | |
| 3 | R. A. Sheldon, Green Chem., 2007, 9, 1273–1283 | |
| 4 | J. March and M. B. Smith, Advanced Organic Chemistry,John Wiley & Sons, New Jersey, 6th edn, 2007 | |
| 5 | C. Montalbetti and V. Falque, Tetrahedron, 2005, 61, 10827–10852 | |
| 6 | E. Valeur and M. Bradley, Chem. Soc. Rev., 2009, 38, 606–631 | |
| 7 | S. Y. Han and Y. A. Kim, Tetrahedron, 2004, 60, 2447–2467 | |
| 8 | J. A. Mitchell and E. E. Reid, J. Am. Chem. Soc., 1931, 53, 1879–1883 | |
| 9 | G. H. Coleman and A. M. Alvarado, Org. Synth., 1923, 3, 3–5 | |
| 10 | C. N. Webb, Org. Synth., 1927, 7, 6–7 | |
| 11 | H. T. Clarke and L. D. Behr, Org. Synth., 1936, 16, 75–76 | |
| 12 | C. Downing (Chemstrand Corp.), U. S. Patent, 2667511, 1954 | |
| 13 | A. W. Campbell (Commercial Solvents Corp.), U.S. Patent, 3006956, 1961 | |
| 14 | D. Davidson and P. Newman, J. Am. Chem. Soc., 1952, 74, 1515–1516 | |
| 15 | B. S. Jursic and Z. Zdravkovski, Synth. Commun., 1993, 23, 2761–2770 | |
| 16 | L. J. Goossen, D. M. Ohlmann and P. P. Lange, Synthesis, 2009, 160–164 | |
| 17 | S. Caddick, Tetrahedron, 1995, 51, 10403–10432 | |
| 18 | R. S. Varma, Green Chem., 1999, 1, 43–55 | |
| 19 | L. Perreux, A. Loupy and F. Volatron, Tetrahedron, 2002, 58, 2155–2162 | |
| 20 | B. W. Baldwin, T. Hirose and Z. H. Wang, Chem. Commun., 1996, 2669–2670 | |
| 21 | S. L. Cockroft and C. A. Hunter, Chem. Commun., 2009, 3961–3963 | |
| 22 | S. E. Wheeler and K. N. Houk, Mol. Phys., 2009, 107, 749–760 | |
| 23 | P. Nelson and A. Pelter, J. Chem. Soc., 1965, 5142–5144 | |
| 24 | A. Pelter, T. E. Levitt and P. Nelson, Tetrahedron, 1970, 26, 1539–1544 | |
| 25 | A. Pelter and T. E. Levitt, Tetrahedron, 1970, 26, 1545–1553 | |
| 26 | G. Trapani, A. Reho and A. Latrofa, Synthesis, 1983, 1013–1014 | |
| 27 | D. B. Collum, S. C. Chen and B. Ganem, J. Org. Chem., 1978, 43, 4393–4394 | |
| 28 | T. Maki, K. Ishihara and H. Yamamoto, Org. Lett., 2006, 8, 1431–1434 | |
| 29 | W. Q. Yang, X. M. Gao, G. Springsteen and B. H. Wang, Tetrahedron Lett., 2002, 43, 6339–6342 | |
| 30 | K. Ishihara, S. Ohara and H. Yamamoto, J. Org. Chem., 1996, 61, 4196–4197 | |
| 31 | K. Ishihara, S. Ohara and H. Yamamoto, Org. Synth., 2002, 79, 176–185 | |
| 32 | K. Ishihara, S. Ohara and H. Yamamoto, Macromolecules, 2000, 33, 3511–3513 | |
| 33 | K. Ishihara, S. Kondo and H. Yamamoto, Synlett, 2001, 1371–1374 | |
| 34 | T. Maki, K. Ishihara and H. Yamamoto, Synlett, 2004, 1355–1358 | |
| 35 | T. Maki, K. Ishihara and H. Yamamoto, Org. Lett., 2005, 7, 5043–5046 | |
| 36 | T. Maki, K. Ishihara and H. Yamamoto, Org. Lett., 2005, 7, 5047–5050 | |
| 37 | T. Maki, K. Ishihara and H. Yamamoto, Tetrahedron, 2007, 63, 8645–8657 | |
| 38 | P. Tang, Org. Synth., 2005, 81, 262–267 | |
| 39 | R. G. Hayter, A. W. Laubengayer and P. G. Thompson, J. Am. Chem. Soc., 1957, 79, 4243–4244 | |
| 40 | A. D. Negro, L. Ungarett and A. Perotti, J. Chem. Soc., Dalton Trans., 1972, 1639 | |
| 41 | L. A. Duncanson, W. Gerrard, M. F. Lappert, H. Pyszora and R. Shafferman, J. Chem. Soc., 1958, 3652–3656 | |
| 42 | K. Ishihara, Tetrahedron, 2009, 65, 1085–1109 | |
| 43 | R. K. Mylavarapu, G. C. M. Kondaiah, N. Kolla, R. Veeramalla, P. Koilkonda, A. Bhattacharya and R. Bandichhor, Org. Process Res. Dev., 2007, 11, 1065–1068 | |
| 44 | K. Arnold, B. Davies, R. L. Giles, C. Grosjean, G. E. Smith and A. Whiting, Adv. Synth. Catal., 2006, 348, 813–820 | |
| 45 | K. Arnold, A. S. Batsanov, B. Davies and A. Whiting, Green Chem., 2008, 10, 124–134 | |
| 46 | K. Arnold, B. Davies, D. Herault and A. Whiting, Angew. Chem., Int. Ed., 2008, 47, 2673–2676 | |
| 47 | R. M. Al-Zoubi, O. Marion and D. G. Hall, Angew. Chem., Int. Ed., 2008, 47, 2876–2879 |
| This journal is © The Royal Society of Chemistry 2010 |