Michael T.
Keßler
,
Silas
Robke
,
Sebastian
Sahler
and
Martin H. G.
Prechtl
*
Department of Chemistry, Institute of Inorganic Chemistry, University of Cologne, Greinstraße 6, 50939 Cologne, Germany. E-mail: martin.prechtl@uni-koeln.de; Web: http://catalysis.uni-koeln.de Fax: +49 221 4701788
First published on 8th October 2013
In the following, we present a simple and feasible methodology for a C–N coupling reaction using nanoscale Cu2O catalysts incorporated in n-Bu4POAc ionic liquid media. It is shown that a wide range of amines and aryl halides can be coupled selectively in high yields, without the use of ligands or additives (bases) and without precautions against water or air. All catalyses can be carried out with a nanoparticle catalyst loading as low as 5 mol%, based on the used precursor.
However, up to now nanoscale copper(I) catalysed cross coupling reactions, especially C–N bond formation, have rarely been reported in scientific literature13,20–22or show drawbacks in versatility and catalyst loadings.20,22–25 Furthermore, it would be more convenient to conduct catalysis at temperatures right below the boiling point of water without the necessity for ligands, additives and inert reaction conditions.26 Nonetheless, we want to establish a Cu2O-nanocatalyst which combines the advantages of classical homogeneous and heterogeneous catalyses. On the one hand, Cu(I) nanocatalysts show good yields and high TONs under moderate reaction conditions; on the other hand, excellent reaction workup and low product contamination are secured. Apart from a few publications,20,23 Cu2O nanoparticles (Cu2O-NPs) were considered as rather unreactive and non-versatile catalysts, which was why they seemed to be useless in C–N coupling reactions. As a rather toxic alternative with adequate activity and versatility, CuI is widely employed.21,27
Usually solvents can alter the chemo-physical properties of catalysts.28 Ionic liquids (ILs) are widely known as so-called “designer solvents”. They can act as promoters and activators,29,30 as protecting agents and stabilisers,28,30–35 as reducing agents11,30,35 and certainly as (co-)solvents themselves for molecular and nanoscale species.36–39 Only limited reports about the use and influence of ionic liquids as solvents for nanocatalytic C–N cross coupling reactions are known.40,41 To the best of our knowledge, publications about the incorporation and stabilisation of Cu2O nanoparticles in ionic liquid media and their use for catalytic amination reaction have not been reported yet.
Herein we established a new approach to catalytic C–N couplings with Cu2O-NP in ionic liquid media. With regard to future applications for a recyclable catalytic system, we tried in particular to simplify the reaction process. The influence of the reaction parameters such as reaction time, temperature and catalyst loading as well as the influence of additives, solvents or inert atmosphere were investigated. Finally, we could establish a versatile methodology for the coupling of iodo- and bromobenzenes with a broad scope of amines and ammonia with a remarkable selectivity. It has to be pointed out that the synthesis of the Cu2O nanoparticles proceeds directly in the used ionic liquid from cheap precursors like basic CuCO3 with subsequent C–N coupling reactions. The as-synthesised nanoparticles are not separated or purified in an additional step but can be directly used as catalysts for amination reactions. In contrast to many other homogeneous and heterogeneous catalyses, the presented coupling reaction in ionic liquid medium turns out to have no necessity for ligands, additional base, additives or an inert atmosphere and can even proceed in the presence of water at temperatures below 80 °C.
Scheme 1 Synthesis of Cu(I) oxide nanoparticles by thermal reduction of copper(II) carbonate in ionic liquid medium. |
In our case, we were pleased to see that iodobenzene couples smoothly even with ammonia in aqueous solution, in the presence of a ligand-free Cu2O nanocatalyst in n-Bu4POAc (Scheme 2). It is quite uncommon that the yields were satisfactory or high at temperatures of about 75–85 °C within 24 h even in the absence of an additional base (see Table 1, entries 7 and 10–13).
Scheme 2 Amination of iodobenzene with aqueous ammonia solution as a test reaction for the activity of the Cu2O NPs in IL. |
Entry | IL | T (°C) | t (h) | Cat.-loadingb (mol%) | Additives | Conv.c (%) |
---|---|---|---|---|---|---|
Reaction conditions: 1 mmol iodobenzene, 1 ml NH3 (aq., 20%, 12 mmol), Cu-cat in 1 g IL.a No Cu2O-NPs detectable.b Based on used precursor amount (CuCO3).c Conversions were determined using 1H NMR with hexamethyldisilane as the internal standard. | ||||||
1 | bmim NTf2 | 85 | 24 | 100a | 0 | |
2 | bmmim NTf2 | 85 | 24 | 100a | 0 | |
3 | bmim OAc | 85 | 24 | 100 | 38 | |
4 | bmpyrr OAc | 85 | 24 | 100 | <1 | |
5 | bmim Cl | 85 | 24 | 100a | 34 | |
6 | (n-Bu4P)2SO4 | 85 | 24 | 100 | 0 | |
7 | n-Bu4POAc | 85 | 24 | 100 | 61 | |
8 | n-Bu4POAc | 85 | 24 | 100 | K2CO3 | 63 |
9 | n-Bu4POAc | 85 | 24 | 100 | KI | 59 |
10 | n-Bu4POAc | 85 | 16 | 100 | 64 | |
11 | n-Bu4POAc | 85 | 16 | 10 | 83 | |
12 | n-Bu4POAc | 100 | 24 | 100 | 40 | |
13 | n-Bu4POAc | 75 | 16 | 10 | 92 |
Different ionic liquids as reaction media were investigated by varying the anion as well as the cation to change the polarity of the ionic liquid. Polarity, basicity and acidity, which are crucial attributes for application in catalysis, are summarized in the so-called Kamlet–Taft parameters. We tried other polar ionic liquids with different Kamlet–Taft parameters50 for the synthesis of Cu2O from CuCO3, pointing out that the particles obtained in n-Bu4POAc were the most active. Rather apolar ionic liquids such as 1-butyl-3-methylimidazolium N,N-bistrifluoromethylsulfonylimide (bmim NTf2) and 1-butyl-2,3-dimethylimidazolium N,N-bistrifluoromethylsulfonylimide (bmmim NTf2) are not capable of reducing the Cu(II) species.11 Therefore, we were not surprised that we could not detect any nanoparticles in the ionic liquids. Furthermore, Cu(II) salts do not support the amination of aryl halides very well (Table 1, entries 1 and 2). However, highly polar ionic liquids often suffer from high melting points ((n-Bu4P)2SO4) and seem to be impractical as reaction media (Table 1, entry 6). The obtained reaction mixtures show high viscosities, leading to a rather inhomogeneous intermixture of educts and the catalytic IL-phase. Nevertheless, acetate-based ionic liquids combine both low melting points and the ability to form Cu2O. The catalytic activities of the formed copper(I) species, 1-butyl-3-methylimidazolium acetate (bmim OAc) and butyl-methylpyrrolidinium acetate (bmpyrr OAc), are rather low (Table 1, entries 3 and 4); only the high polarity of the reaction medium seems to support the amination reaction (Table 1, entry 5). Interestingly, only the nanoparticles generated in n-Bu4POAc show good results in the amination of iodobenzene with a yield of about 61% (Table 1, entries 7 and 10–13).
With n-Bu4POAc as our reaction medium of choice, we investigated the influence on the reaction time towards the yield. We observed that the conversion of iodobenzene to aniline increases drastically with time and a maximum conversion could be reached after 16 h (64%) (Fig. 1, Table 1, entry 10).
Fig. 1 Plot of reaction time and corresponding yield of aniline. The maximum yield is reached within 16 h at 85 °C. |
Investigation of the catalyst loading showed that a decrease in the amount of Cu2O raised the yield drastically. The maximum value of about 83% was achieved using only 10 mol% of the catalyst (Table 1, entry 11). The catalyst loading is referred to as the amount of catalyst precursor applied. In fact, due to the high surface/volume ratio of the particles, the amount of real active catalyst is even lower.11,41 Loadings of the catalyst below 10 mol% as well as loadings higher than 10 mol% lowered the yield of aniline. Moreover, with high catalyst loadings, the system's miscibility probably suffers due to the higher viscosity of the liquid phase. Analogous to several other publications dealing with Pd- or Ni-NPs, it is possible that the Cu2O-NPs act as a catalyst reservoir for Cu(I), and a molecular species might act as the active species.32,51 There are different ways of how a molecular species can theoretically be leached from the particle surface as Beletskaya and co-workers have revealed. Usually there is an equilibrium between leaching and the Ostwald-ripening of the particles. In this context, the synonym “nanosalt” is introduced for metal–chalcogen nanoparticles.51 Up to now, the mechanism for the leaching effect has not been intensively investigated, but usually the Ar–X species is believed to generate molecular metal species by oxidative addition.26 During the extraction of the reaction mixture using n-pentane, copper species can be leached out of the catalytic phase and contaminate the final product. In order to investigate the amount of copper in the organic phase, ICP-OES measurements revealed that the extract contains only a maximum of 0.041 μmol (lowest value found: 0.00029 μmol) copper. Based on the utilised amount of catalyst precursor (CuCO3 0.05 mmol), the leached copper amount is equivalent to a maximum of 0.082%, which is remarkably low.
Moreover, the outcome of the amination reaction catalysed by Cu(I) oxide nanoparticles could be improved further by adjusting the reaction temperature. The yields can effectively be increased to an excellent yield of 92% at temperatures as low as 75 °C (Table 1, entry 13). The higher conversions at lower temperatures are most likely related to the low boiling point of ammonia; thus, the concentration in the liquid phase is higher at lower temperature. It is a remarkable fact that a ligand- and additive-free approach shows best results far below the boiling point of water, which makes the reaction also possible in water. Most other ligand-free approaches need temperatures between 100 °C and 135 °C.16 A further increase in the reaction temperature affects the yield of aniline negatively (Fig. 2).
Fig. 2 Correlation between reaction temperature and corresponding yield of aniline. The best results could be achieved at 75 °C within 16 h and a catalyst loading of 10 mol%. |
The addition of additives like a base (K2CO3) or KI (Table 1, entries 8 and 9) to change the pH-value or to simulate a different copper(I)-species did not improve the yield of aniline measurably. Obviously it is not necessary to neutralise the HX species,16,19 which is produced during the reaction process, by an additional base. On the one hand, the acetate ionic liquid may act as a buffer reagent and keep the pH-value constant, and on the other hand, ammonia can act as a proton scavenger as well.
In comparison to the used copper(I) oxide nanoparticles, we investigated different (bulk) copper salts in oxidation states (I) and (II) (Fig. 3). In contrast to ref. 17, we could not show activity for Cu2+ species in our reaction sequence. All used Cu(II) salts (Cu(NO3)2, CuSO4, CuCl2, Cu(OAc)2 and CuF2) failed completely in catalysing the amination of iodobenzene with ammonia under the same conditions. A competing reaction might be the formation of a stable Cu(NH3)42+ complex which colours the reaction mixture immediately. CuI (as used in other manuscripts)22 and bulk Cu2O applied in ionic liquid media showed very poor activity of <1% and 3% yields, respectively.
As a starting point, N-arylation reactions of iodobenzene with primary amines were chosen with a catalyst loading of 10 mol%. The results of the arylation experiments are listed in Table 2, showing remarkable results concerning yields and catalyst loadings. We were pleased to find out that we could reduce the loading of the copper-catalyst to 5 mol% without a significant loss in activity (Table 2, entry 2).
Entry | Amine | Cat.-loadingb (mol%) | Conversion (sec. amine)a (%) |
---|---|---|---|
Reaction conditions: 1 mmol iodobenzene, 1 ml prim. amine (CH3-NH2 (40 wt%) 12.9 mmol, n-Bu-NH2 10.1 mmol, iso-Bu-NH2 10.6 mmol, cHex-NH2 8.7 mmol, n-Oct-NH2 6.0 mmol, iso-Pr-NH2 11.6 mmol, tert-Bu-NH2 9.5 mmol, C6H5-NH2 10.7 mmol), Cu-cat in 1 g (3.1 mmol) IL, reaction temperature: 75 °C, reaction time: 16 h.a Conversions were determined using 1H NMR with hexamethyldisilane as the internal standard.b Based on used precursor amount (CuCO3). No by-products could be detected. | |||
1 | CH3-NH2 | 10 | 65 |
2 | CH3-NH2 | 5 | 65 |
3 | n-Bu-NH2 | 5 | 71 |
4 | iso-Bu-NH2 | 5 | 95 |
5 | cHex-NH2 | 5 | 77 |
6 | n-Oct-NH2 | 5 | 90 |
7 | iso-Pr-NH2 | 5 | 99 |
8 | tert-Bu-NH2 | 5 | 57 |
9 | C6H5-NH2 | 5 | 65 |
The best results could be obtained with linear and branched alkyl amines (65%–99%, Table 2, entries 2–4 and 6–8), followed by cyclic (77%, Table 2, entry 5) and aryl amines (65%, Table 2, entry 9). The low yields of the coupling between iodobenzene and methyl amine can be explained by the rather low electron density (e.g. compared to n-butyl amine). No further arylation of the desired secondary amine was detectable, which might be in correlation with the high excess of primary amine in the reaction mixture. A similar screening with secondary amines is shown in Table 3.
Entry | Amine | Cat.-loadingb (mol%) | Conversion (tert. amine)a (%) |
---|---|---|---|
Reaction conditions: 1 mmol iodobenzene, 1 ml sec. amine (Et2NH 9.6 mmol, n-Bu2NH 5.9 mmol, piperidine 10.1 mmol, Ph2NH 6.9 mmol, morpholine 11.5 mmol, L-Prolin 11.7 mmol), Cu-cat in 1 g (3.1 mmol) IL, reaction temperature: 75 °C, reaction time: 16 h.a Conversions were determined using 1H NMR with hexamethyldisilane as the internal standard.b Based on used precursor amount (CuCO3). No by-products could be detected. | |||
1 | Et2NH | 5 | 65 |
2 | n-Bu2NH | 5 | 16 |
3 | Piperidine | 5 | 99 |
4 | Ph2NH | 5 | 0 |
5 | Morpholine | 5 | 99 |
6 | L-Prolin | 5 | 33 |
In summary, the conversions of the coupling between secondary amines and aryl halides are slightly lower in comparison to the conversions of the coupling between primary amines and aryl halides. This might be due to emerging steric effects. The starting materials (secondary amines) and products (tertiary amines) are much bulkier and thus slightly disfavoured in the transmetallation steps. Diethylamine gave a satisfactory yield of 65% (Table 3, entry 1). There was on the contrary no conversion detectable when diphenylamine was used as the amination reagent (Table 3, entry 4). Nevertheless, there are also some examples which verify the versatility of this method with good to excellent results (e.g. morpholine and piperidine, both 99%, Table 3, entries 3 and 5). These heterocyclic compounds are less bulky than common secondary alkyl amines due to their fixed ring structure. No by-products, such as perarylated ammonium derivatives, could be detected in the reaction mixture.
Entry | Aryl–X | Amine | Cat.-loadingb (mol%) | Conversiona (%) |
---|---|---|---|---|
Reaction conditions: 1 mmol aryl halide, 1 ml piperidine (10.1 mmol), diethylamine (9.6 mmol) or ammonia solution (20%, 12 mmol), Cu-cat in 1 g (3.1 mmol) IL, reaction temperature: 75 °C, reaction time: 16 h.a Conversions were determined using 1H NMR with hexamethyldisilane as the internal standard.b Based on used precursor amount (CuCO3). No by-products could be detected. | ||||
1 | C6H5Br | NH3 | 10 | 12 |
2 | 3-Iodoanisole | NH3 | 5 | 92 |
3 | 4-Iodotoluene | NH3 | 10 | 79 |
4 | 1,2-Dichlorobenzene | Et2NH | 5 | 0 |
5 | C6H5Br | Et2NH | 5 | 10 |
6 | 2-Iodophenol | Et2NH | 5 | 0 |
7 | 2-Iodoanisole | Et2NH | 5 | 0 |
8 | 3-Iodoanisole | Et2NH | 5 | 24 |
9 | 2-Iodotoluene | Et2NH | 5 | 0 |
10 | 4-Iodotoluene | Et2NH | 5 | 33 |
11 | 3-Iodoanisole | Piperidine | 5 | 93 |
12 | 4-Iodotoluene | Piperidine | 5 | 85 |
This journal is © The Royal Society of Chemistry 2014 |