2020 Volume 68 Issue 8 Pages 683-693
In this review, we summarize our recent progress on functionalization of the ammonium C–N bond through a transition-metal-catalysed cross-coupling process. By synergistic utilization of computational and experimental methods, we have successfully developed several new C–N bond cleavage protocols and established new reaction mechanisms. These findings provide new possibilities for transforming naturally abundant chemicals into useful functional molecules in an efficient and selective manner.
For more than a century, chemical synthesis based on the utilization of fossil fuels has created an enormous range of materials for human use. Fossil fuels contain highly reduced molecules (C–H bond), which must be oxidized (C–X bond, X = halogen or hetero-atom) during most chemical processes. In recent years, in response to environmental issues, new strategies based on the reductive transformation of highly oxidized molecules (e.g. C–O, C–N, etc.) in biomass feedstocks, as well as in other renewable resources and functional materials, have attracted great interest.
Many of these naturally abundant chemicals and functional molecules contain C–O or C–N bonds. For example, amine groups occur widely in natural products, and are also found in many pharmaceuticals, dyes, and other functional molecules (Chart 1). A large variety of amines are commercially available, mostly at reasonable cost. Therefore, efficient C−N bond conversion methods suitable for late-stage functionalization would greatly expand their utility as versatile synthetic feedstocks.1) However, transformation of the NR2 group is generally difficult, due to the high stability of the C−N bond. Existing methods for C–N bond cleavage are still very limited, most of which require harsh conditions, highly active reagents and/or a noble metal catalyst having complex structures.
On the other hand, quaternary organo-ammonium salts can be easily prepared from various amines/anilines in quantitative yield and on a gram scale.2) Ammonium salts also show high stability toward air and moisture, enabling long-term storage at ambient temperature in open vessels. These advances indicate that ammonium salt can be used as an easy-handling and robust pre-activated form of amine/aniline to improve the reactivity of its inert C–N bonds. In recent years, by synergistic utilization of computational and experimental methods, we have successfully developed several new protocols for inert C–N functionalization, which have also been applied for the syntheses of many functional molecules and/or polymers. Here, we would like to introduce some of our results on the development of transition-metal-catalysed cross-coupling through the ammonium C–N bond cleavage process.
In 1988, Wenkert et al. described the first cross-coupling of aryl ammonium salt with Grignard reagents catalysed by NiCl2(dppp) [dppp: 1,3-bis(diphenylphosphino)propane].3) Although this observation revealed a new and potentially useful reactivity of aniline, as well as a new horizon for cross-coupling reactions, for decades it has been long and largely neglected,2) and the reaction mechanism remains unclear.
To understand this reaction in-depth, so as to design and develop new and efficient protocols, we focused our attention on clarifying the mechanism of the Ni-mediated C–N bond cleavage process. Density functional theory (DFT) computation was performed, and the results show that ammonium salt shows much higher reactivity (Chart 2-a) than aniline (Chart 2-b) toward Ni(0) species, as reflected in quite low activation barriers, indicating that the Ni-mediated C–N bond cleavage process may proceed under mild conditions.4) Indeed, when ammonium salt was treated with Ni(0) complex with 1,3-dicyclohexylimidazol-2-ylidene (ICy) as the ligand, at room temperature, the product of C–N bond cleavage was observed in moderate yield (Chart 3), and its structure was characterized by single crystal X-ray diffraction.4) These mechanistic studies present useful knowledge for the further development of C–N bond functionalization reactions. For example, according to the high reactivity of ammonium salt, it would be possible to use less reactive nucleophiles in the cross-coupling process, such as organo-stannane, which may provide wide compatibility for various functional groups and establish a robust reaction process.
Based on our 2016 computational and experimental investigations, we reported the first Ni-catalysed Stille coupling reaction through C–N bond cleavage of quaternary organo-ammonium salts4) (Chart 4). Although Pd-catalysed Stille coupling between stannanes and organic halides is currently the second most widely used cross-coupling method for the synthesis of functional molecules and polymers,5,6) both in laboratory research and in industry, Stille coupling using C–N electrophiles (amine derivatives) has hardly been investigated, and very few Ni-catalysed Stille coupling reactions have been reported so far.7,8) We found that by using a Ni(cod)2 catalyst and imidazole ligand, aryl trimethylammonium salts 1 reacted smoothly with arylstannanes 2 in a 1 : 1 M ratio in the presence of CsF, affording the corresponding biaryl 3 with broad functional group compatibility (Chart 4).
This reaction shows broad functional group compatibility, and is applicable for a variety of substrates: 1) tolyl ammonium reacted very smoothly and selectively with arylstannane to give the corresponding coupling products, and the location (ortho-, meta- or para-) of the methyl substituent did not greatly affect the reaction; 2) the reaction of highly sterically demanding mesityl ammonium still took place, albeit with a dropped yield; 3) ammonium salts bearing electron-donating groups (OMe, Me) were less reactive than those with electron-withdrawing groups such as fluorine, ester, ketone, nitrile, silyl, and sulfone, which are generally very active and sensitive to bases and organometallics; 4) ammoniums with biphenyl or naphthalene also reacted efficiently, indicating that molecules with expanded π-conjugated systems might be prepared by employing this method; 5) the reactions of various stannanes also proceeded smoothly to afford biaryl products with broad functional group tolerance; 6) heterocyclic substrates reacted without difficulty, and the desired products were obtained in high yields.
The excellent reactivity and high selectivity of this reaction showed further interesting synthetic applicability, as exemplified by sequential cross-coupling for the regio-controlled synthesis of p-terphenyl derivatives (Charts 5-a and b) and selective phenylation of the NMe2 group in Padimate A (Chart 5-c).
Furthermore, the reaction pathway was fully characterized based on experimental and computational methods. The results of DFT calculations are summarized in Chart 6. First, the Ni(0)-π complex CP0 is formed with −3.0 kcal mol−1 exothermicity from Ni(ICy)2 (generated from Ni[cod]2 and ICy) and [PhNMe3]+F− (generated via anion metathesis of [PhNMe3]+[OTf]− and CsF; the reaction route, starting from [PhNMe3]+[OTf]−, was also calculated, but there was no marked difference in geometric structure or energy profile compared with the results shown in Chart 6. From CP0, Ni(0) can migrate on the phenyl ring to the proximal position of the C–N bond via TS0, with an energy loss of only 10.2 kcal mol−1, to form the more stable CP1. Cleavage of the C–N bond then takes place very smoothly (TS1, −2.0 kcal mol−1), with the release of NMe3, affording intermediate CP2-1 with large exothermicity (−45.5 kcal mol−1). PhSnMe3 then approaches the Ni(II) center in CP2-1 after the loss of one ICy ligand, and rotation of the Ni–F bond from the vertical to the horizontal position, to generate CP2-2 with an overall energy loss of 18.4 kcal mol−1. Next, transmetalation takes place from CP2-2 through TS2 (−23.0 kcal mol−1) to give CP3-1 (−27.2 kcal mol−1) with a small activation energy (4.1 kcal mol−1). CP3-1 then ejects FSnMe3 to afford the precursor for the reductive elimination, CP3-2 (−19.0 kcal mol−1). Finally, C–C bond formation proceeds smoothly through TS3, with an energy loss of only 2.3 kcal mol−1, to produce the final product, Ph–Ph, and the Ni(ICy)2 catalyst is regenerated with a large energy gain. These results are in good agreement with the experimental findings.
The successful use of ammonium salts instead of halides as electrophiles indicates that some side reactions of organic halides could be efficiently avoided, such as the lithium–halogen exchange. Organolithium chemistry is employed in many areas of science.9–13) However, the applicability of Murahashi coupling14,15) is sometimes restricted due to the competing lithium–halogen exchange (dehalogenation), which is usually quite fast and thus lowers the selectivity. In 2017, we reported the first example of a Murahashi coupling reaction through ammonium C–N bond cleavage16) (Chart 7). We found that the cross-coupling reaction between aryl ammonium salts 1 and aryl-lithiums 4 proceeded under mild conditions, with C–N bond cleavage in the presence of Pd(PPh3)Cl2 catalyst, giving the desired products 3 in moderate to high yields (Chart 7). Both electron-donating and -deficient groups on the phenyl ring of aryl-lithiums were found to be generally suitable for this coupling. A sterically demanding aryl lithium (o-Me–C6H4–Li) did not shut down the reaction. Normally, naphthyl ammonium salts show higher C–N cleavage reactivity than their phenyl counterparts. Interestingly, this reaction is also applicable to less reactive phenyl ammonium salts to give the desired biphenyl or p-terphenyl compounds in high to excellent yield.
Meanwhile, we also found that in the presence of Ni(cod)2 catalyst and imidazole ligand, aryl-lithiums 4 react smoothly with aryl ethers 5, giving the corresponding biaryl 3 in high yield at room temperature, or with slight heating (Chart 8). For organolithium, it is noteworthy that 1) organolithiums with bulky aryl rings, such as o-tolyl and even mesityl, reacted smoothly under the current condition; 2) OR groups (OMe, OTBS, OMOM) on the Ph ring (or even at the benzylic position) survived intact in the reaction, indicating that cleavage of the ethereal C–O bond is chemoselective; 3) functional groups such as NMe2 and trimethylsilyl (TMS), as well as several heterocycles, can be well tolerated; 4) in addition, alkenyl- and alkyl-lithiums reacted smoothly, albeit the yield was somewhat decreased. On the other hand, for aryl methyl ethers, it is noteworthy that 1) both 2-methoxy- and 1-methoxynaphthalenes with different substituents were efficiently converted to the corresponding biaryl products in good to excellent yield; 2) with 2,7-dimethoxynaphthalene bearing two C–O moieties, di-phenylation proceeded smoothly; 3) under current conditions, most anisole derivatives showed quite high reactivity, in sharp contrast to the many known protocols for Ni-catalyzed C–O bond cleavage, where 1-/2-methoxynaphthalenes were usually more reactive than anisole derivatives; 4) strikingly, the C–C bond formation can be achieved efficiently without racemization at the sensitive benzylic/α-amino position; 5) the reactions of electron-deficient/rich heteroaromatic methyl ethers also could take place, albeit with dropped yields, even at high temperatures.
These results demonstrate the potential applicability of the current C–N/C–O cleavage type Murahashi coupling reaction for late-stage derivatization of functional molecules. By combination of these two newly developed methods, selective and sequential functionalization of arenes that bear both C–N and C–O bonds in one pot has been established (Chart 9).
Organoaluminum reagents offer many synthetic advantages, including low cost, ready availability, low toxicity and, in particular, their exceptional Lewis acidity.17) As one of the initial examples of transition-metal-catalysed cross-coupling, the reaction between organic halide and organoaluminum with Ni or Pd-catalyst was first reported by Negishi and Baba as early as 1976.18,19) However, in the following years, the utilization of organoaluminum compounds in modern transition-metal-catalysed cross-coupling reactions has been largely neglected in favour of organoboron (Suzuki–Miyaura reaction), organozinc (Negishi reaction), organomagnesium (Kumada–Tamao reaction), and organostannane (Stille reaction) reagents. In 2015, we presented a direct cross-coupling reaction between arylaluminum reagents and organic halides without the need for any external catalysts20) (Chart 10). In this reaction, the low halogen-metal exchange ability of aluminium reagents efficiently minimized side dehalogenation reactions, hence enabling the use of a wide variety of electrophiles, including aryl/alkenyl/alkynyl iodide/bromide/iodide with broad functional group compatibility. Further, this direct cross-coupling reaction showed interesting chemo-selectivity that is very different from transition-metal-catalysed chemo-selectivity, as reflected in the competitive reactions between OTf and Br (Chart 11).
With continuous interest in the unique reactivity of organoaluminiums, in 2017 we further found that heating aryl ammonium salts 1 in tetrahydrofuran (THF) in the presence of a commercially available NiCl2(PCy3)3 catalyst provided satisfactory reactivity with arylaluminum reagents 621) (Chart 12). Alkyl ammonium salts, as well as alkenyl aluminum reagents, were also available for this reaction.
Moreover, we also extended the scope of aluminum-mediated cross-coupling, from using ammonium salts (C–N cleavage) to employing phenol derivatives (C–O cleavage) and organic fluorides (C–F cleavage). First, we found that the Ni-catalysed cross-coupling reaction of aluminium reagents 6 is highly applicable to a wide variety of phenol derivatives, including esters (7a), carbamate (7b), carbonate (7c), tosylate (8a), mesylate (8b), triflate (8c), phosphate (9), and even the very unreactive ether (5) (Chart 13). As a representative example, in the presence of NiCl2(PCy3)3 catalyst, aluminium reagents 6 react smoothly with aryl carbamate (7b) in toluene solution, providing products in high yield with broad functional group compatibility (Chart 14). It is noteworthy that 1) many less reactive phenol carbamates showed desirable reactivity under the current condition, albeit with heating at higher temperatures; 2) a variety of functional groups, such as methoxy, trifluoromethyl, ester, and amide could be well tolerated; 3) the steric bulkiness of both aryl aluminums and aryl carbamates did not greatly affect the reactivity, though high temperature was needed to achieve a good yield in the reaction with very bulky aryl carbamate; 4) reactions of a variety of heterocyclic substrates proceeded smoothly, providing the coupling product in good to high yield.
Next, the reaction between organoaluminum and organic fluoride was also investigated (Chart 15). As a result, using NiCl2(PCy3)3 as the catalyst, cross-coupling between aluminium reagents 6 and various organic fluoride 10 proceeded smoothly in THF solution under mild conditions. It is known that electron-rich aryl fluorides are normally less reactive electrophiles for transition-metal catalyzed cross-coupling reactions. However, using the current protocol, the reaction with aryl fluoride bearing an electron-donating OMe and NMe2 group took place without any difficulty, even with the sterically demanding 2,6-dimethylphenyl aluminum reagent (albeit at higher temperature). Furthermore, the reaction was also compatible with many functional groups, such as CF3, amide, and ester, as well as heterocycles, some of which are sensitive to metal reagents and bases.
We also found that the chemo-selectivity of the present coupling system (C–N vs. C–O vs. C–F) was dependent upon the reaction solvent (Chart 16). In toluene, both the C–O bond in carbamate and the C–N bond in ammonium salt showed much higher reactivity than the C–F bond. Compared with the case of the carbamate C–O bond, the decrease in reactivity of ammonium salt in toluene is probably due to low solubility. On the other hand, in THF, the C–N bond in ammonium salt showed slightly higher reactivity than the carbamate C–O bond, and is preferentially cleaved over the C–F bond. Thus, C–F bond cleavage was the most sluggish process, both in toluene and in THF. Although the selectivity is poor in THF, it is noteworthy that the competitive reactions in toluene for C–O vs. C–F and C–F vs. C–N were highly selective.
Taking the C–O bond cleavage reaction as a representative case, we performed a systemic mechanistic study using DFT computation22) (Chart 17). As a result, the cross-coupling basically proceeded according to a catalytic cycle consisting of three elementary steps, namely, oxidative addition, transmetalation, and reductive elimination. In the initial step, the Ni(0) catalyst and aromatic substrate form a π-arene complex CP-a, in which carbonyl oxygen also coordinates to Ni. Then, oxidative addition of the C(aryl)–O bond to Ni(0) catalyst takes place via TS-ab to generate CP-b1, giving the C(aryl)–Ni(II)–O species with a total activation barrier of 24.3 kcal/mol. The OAc moiety on Ni(II) in CP-b1 then coordinates to the aluminum reagent to form CP-b2, which undergoes transmetalation through TS-bc→CP-c→TS-cd to afford Ar2Ni(II) species CP-d along the intrinsic reaction coordinate, with release of the Me2AlOAc moiety. Transmetalation proceeded with a quite low activation barrier (< 12 kcal/mol), without much energy gain. Finally, reductive elimination takes place via TS-dP, (ΔGa = 7.0 kcal/mol), leading to the coupling product PD, and regeneration of the Ni(0) catalyst with large exothermicity. The whole process did not involve any energetically unfavorable CPs/TSs with a large overall energy gain. The calculation also indicates that the strong Lewis acidity of aluminums efficiently promotes the transmetalation, rather than the oxidative addition step.22)
We further examined the applicability of the ammonium C–N bond cleavage reaction in polymer synthesis.23) Treating di-ammonium salts with Grignard reagents in the presence of 1 mol% PdCl2(PPh3)2, the polycondensation proceeded smoothly in THF at room temperature, giving the corresponding π-conjugated polymers with high molecular weight and high yield (Chart 18). At the same time, we also established the first cross-coupling polycondensation between aryl-di-ether and Grignard reagents with equally high efficiency (Chart 19). π-Conjugated polymers are important materials in the field of optoelectronics. These methods open the possibility of transforming a range of naturally abundant chemicals into useful functional polymers.
In recent years, we have successfully developed several new reactions for C–N functionalization through transition-metal-catalysed cross-coupling protocols. By synergistic utilization of computational and experimental methods, we have performed a comprehensive mechanistic study to understand these new reactions in depth. These reactions describe the superior reactivity and versatile utility of ammonium salt as a valuable analogue for organic halides, providing new synthetic applicability of naturally abundant chemicals as clean and renewable resources for future chemical synthesis. More recently, we have established a new reaction route for C–N bond cleavage without the need for transition metals, such as nucleophilic aromatic substitution24,25) and photo-catalysis.26) Furthermore, in addition to C–N cleavage reactions, we also focused on several challenging topics, including inert C–O bond activation,16,21–23,27–29) single-electron-catalysed reaction,20,30,31) and group 14 element chemistry.32,33) Further studies to clarify the reaction mechanisms and to expand the scope of these reactions and their synthetic utility, including their applicability to preparing functional molecules, are in progress.
I am deeply grateful to Professor Masanobu Uchiyama (The University of Tokyo) for his solid support, warm encouragement, and great help with both my research and my life in Japan.
I also sincerely thank all my students: Mr. Takashi Ozaki, Mr. Hiroshi Minami, Mr. Hiroyuki Ogawa, Dr. Dong-Yu Wang, Mr. Ze-Kun Yang, Ms. Ning-Xin Xu, and Ms. Bi-Xiao Li, for their remarkable endeavours and contributions.
The work described herein was supported by JSPS Grants-in-Aid for Scientific Research No. 24790035, No. 14F04908, and No. 17H06173, and also by Grants from the RIKEN-SPDR program, the Kobayashi Foundation, and the Society for the Promotion of Pharmaceutical Science.
The author declares no conflict of interest.
This review of the author’s work was written by the author upon receiving the 2018 Pharmaceutical Society of Japan Award for Young Scientists.
