Selective formylation or methylation of amines using carbon dioxide catalysed by a rhodium perimidine-based NHC complex

Raphael H. Lam a, Caitlin M. A. McQueen b, Indrek Pernik *a, Roy T. McBurney a, Anthony F. Hill *b and Barbara A. Messerle *a
aDepartment of Molecular Sciences, Macquarie University, NSW 2109, Australia. E-mail: barbara.messerle@mq.edu.au; indrek.pernik@mq.edu.au
bResearch School of Chemistry, The Australian National University, Canberra, ACT 0200, Australia. E-mail: a.hill@anu.edu.au

Received 3rd October 2018 , Accepted 8th December 2018

First published on 12th December 2018


Carbon dioxide can play a vital role as a sustainable feedstock for chemical synthesis. To be viable, the employed protocol should be as mild as possible. Herein we report a methodology to incorporate CO2 into primary, secondary, aromatic or alkyl amines catalysed by a Rh(I) complex bearing a perimidine-based NHC/phosphine pincer ligand. The periminide-based ligand belongs to a class of 6-membered NHC ligand accessed through chelate-assisted double C–H activation. N-Formylation and -methylation of amines were performed using a balloon of CO2, and phenylsilane as the reducing agent. Product selectivity between formylated and methylated products was tuned by changing the solvent, reaction temperature and the quantity of phenylsilane used. Medium to excellent conversions, as well as tolerance to a range of functional groups, were achieved. Stoichiometric reactions with reactants employed in catalysis and time course studies suggested that formylation and methylation reactions of interest begin with hydrosilylation of CO2 followed by reaction with amine substrates.


Introduction

Carbon dioxide is potentially a sustainable feedstock for the chemical industry due to its non-toxic nature, low cost and increasing abundance in the atmosphere.1–3 The key challenge of using CO2 as a building block in the chemical industry is to overcome the inertness of CO2, such that mild reaction conditions could be employed.1 Industrial examples of using CO2 as a feedstock are comparatively limited, but include the synthesis of poly-/cyclic carbonates,1,4 urea2,3,5 and salicylic acid.2,5–7

The formylation or methylation of amines using CO2 are potentially important processes.3,8–11 Formamides are commonly used in the production of drug molecules and agrochemicals.12–15 Methylated amines can improve lipophilicity16–20 and potency18–21 of drug molecules, or sometimes control the conformations of molecules.20,22–25 Current methylation protocols rely heavily on the stoichiometric use of toxic reagents such as formaldehyde, methyl iodide and dimethylsulfate amongst others.26–29 Thus, CO2 would represent a safer and more benign substitute. The first example of amine formylation with CO2 was reported by Cantat and co-workers in 2012, using organic bases as catalysts.30,31 The methylation of amines with CO2 was later achieved independently by the Cantat and Beller groups in 2013 using Zn32 and Ru10 complexes as catalysts. Several organometallic catalysts that promote the methylation of amines have been reported since then by other groups.32–48 Among these, only two reports featured a rhodium catalyst, and all NHC-based catalysts featured the traditional 5-membered NHC motif (Fig. 1a (left) and b).39,40


image file: c8gc03094d-f1.tif
Fig. 1 (a) N-Heterocyclic carbene ligands with imidazole-based (left) and perimidine-based (right) structures. (b) Previous work on the addition of CO2 to amines by Kobayashi.39,40 (c) This work – catalysed addition of CO2 to amines. Respective pre-catalysts in insets.

N-Heterocyclic carbenes (NHCs) and pincer-type ligands are popular because of their capacity to stabilise metal complexes against decomposition during catalysis. Typically, NHCs are based on five-membered heterocyclic rings, such as the 5-membered imidazole-based structure (Fig. 1a). There are only a handful of examples of 6-membered NHCs, using a perimidine-based scaffold (Fig. 1a, right).49–59 We are particularly interested in using 6-membered NHC ligands, as they place the carbene centre in an electron-rich, annulated 14 π-electron aromatic system, as in a perimidine molecule.60 Furthermore, expanding the NHC ring from five- to six-membered influences the steric impact of the ligand, decreasing the angle α (Fig. 1a) and directing the amine substituents closer to the coordinated metal.61 Studies by Richeson and co-workers on a range of free and metallated perimidinylidene-based NHCs (per-NHCs)58,59 have highlighted 6-membered NHC systems. Their reports included experimental data indicating enhanced σ-basicity for the per-NHC ligands relative to their five-membered ring analogues, as did a later report from Herrmann et al.50 Furthermore, studies by the Özdemir49 and Dötz51 groups have found palladium per-NHC complexes to be effective catalysts, and the latter of these reports noted that their per-NHC system, which incorporates the per-NHCs as the axial donors in a pincer framework, was more efficient in Heck coupling reactions than its imidazole and benzimidazole analogues.

We wanted to determine the catalytic activity of rhodium complexes featuring the novel 6-membered perimidine-based NHC ligand in CO2 incorporation reactions (Fig. 1c). Herein, we report the selective formylation and methylation of primary and secondary amines with CO2, catalysed by per-NHC rhodium complexes [RhCl{C(NCH2PPh2)2C10H6}] (2a) and [RhCl{C(NCH2PCy2)2C10H6}] (2b), using mild reaction temperatures (25–90 °C) and low pressures of CO2 (balloon). Switchable product selectivity (formylation vs. methylation) was achieved by controlling the quantity of hydrosilane and by changing the reaction solvent and temperature.

Results and discussion

Catalyst preparation

Synthesis of rhodium pincer per-NHC complexes. As described in an earlier communication,54 double chelate-assisted aminal C–H activation of 1a and 1b occurred upon direct reaction with [RhCl(PPh3)3] to give the per-NHC pincer complexes [RhCl{C(NCH2PR2)2C10H6}] (R = Ph 2a, R = Cy 2b, Scheme 1).54 Pro-ligands 1a and 1b were prepared by combining 1,8-diaminonaphthalene, paraformaldehyde and a secondary phosphine HPR2 in a remarkably convenient one-pot synthesis (Scheme 1).
image file: c8gc03094d-s1.tif
Scheme 1 Preparation of 2,3-dihydroperimidine pro-ligands and their reactions with [RhCl(PPh3)3] to form per-NHC pincer complexes.

The reactions proceeded within 21 hours and 10 minutes for 2a and 2b, respectively, and were monitored by 31P NMR spectroscopy. The only side product observed in the 31P NMR spectra was triphenylphosphine (δP ∼ −5), and in the latter rapid reaction the evolution of dihydrogen was also evident in the 1H NMR spectrum (δH = 4.47) and the observation of gas evolution. Thus, this direct preparation with rhodium54 and also ruthenium55 complexes have provided rare preliminary reports of atom-efficient instances of NHC ligand installation via double geminal C–H activation.62–64 This preparation of an NHC complex avoids the typical use of either base or silver transmetallation to activate cationic azolium precursors.65

Both rhodium complexes could be precipitated from the reaction mixtures to afford high isolated yields (2a, 88%; 2b, 92%). The formulations of 2a and 2b as NHC pincer complexes are based on the spectroscopic data, which each comprise a rhodium-103-coupled doublet resonance in the 31P{1H} NMR spectra (2a, δP = 22.9; 2b, δP = 38.3), and one 1H NMR resonance for the PCH2 protons (2a, δH = 4.14; 2b, δH = 3.58) with no additional CH2 protons evident. Virtual triplet coupling of PCH2 protons to the phosphorus nuclei is observed for 2a, with a small JH–P coupling value of 2 Hz, though it could not be discerned in the 1H NMR spectrum of 2b. Though the low solubilities of the complexes precluded acquisition of adequate 1D 13C NMR spectra, the carbene carbon resonance of 2b was detected in a 1H–13C HMBC experiment at δC = 206.7 within the range observed for previously reported per-NHC rhodium(I) complexes.56,59

Molecular structure of 2b. Previously only the molecular structure of 2a had been determined; in this work the X-ray structure of 2b is also reported which allows comparisons to be made. The space-filling representations of both 2a and 2b are shown (Fig. 2), illustrating the enhanced steric crowding of the metal centre by the cyclohexyl groups in 2b relative to the phenyl groups in 2a. The structure of 2b straddles a crystallographic C2 axis that passes through C1 and Rh1, with the ring system of the pincer ligand lying within the coordination plane. However, broadness of the PCH2 resonance in the 1H NMR spectrum is consistent with twisting of the ring system occurring in solution. The Rh1–C1 and Rh1–Cl1 distances are comparable to those observed in 2a.
image file: c8gc03094d-f2.tif
Fig. 2 Molecular structure of 2b (aryl and cylohexyl hydrogen atoms omitted, 50% displacement ellipsoids, asterisks denote symmetry-generated atoms (crystallographic C2 axis through C1 and Rh1), and space-filling representations of 2a and 2b viewed along the coordination planes. Selected bond lengths (Å) and angles (°) for 2b: Rh1–C1 = 1.930(6), C1–N1 = 1.390(5), Rh1–P1 = 2.2338(12), Rh1–Cl1 = 2.3998(16), C1–Rh1–P1 = 84.90(3), C1–Rh1–Cl1 = 180, P1–Rh1–P1* = 169.79(6).
Investigation on the formation of 2a. Though the complete formation of 2a required many hours, the starting materials appeared to have been completely consumed within one hour to give a major intermediate species that gradually converted to the final product 2a. The 31P{1H} NMR spectrum of this intermediate species showed a doublet of doublet resonance at δP = 61.3 (1JRh–P = 133 Hz, 2JP–P = 24 Hz) and a doublet of triplet resonance at δP = 17.9 (1JRh–P = 83 Hz, 2JP–P = 24 Hz), suggesting that both the pincer ligand and a triphenylphosphine ligand were coordinated to rhodium. The 1H NMR spectrum of this species showed a multiplet hydride resonance at δH = −17.6 (C6D6), and two mutually coupled resonances at δH = 3.95 and 4.83 (2JH–H = 13 Hz) that could be attributed to PCH2 protons, indicating diastereotopic protons on these methylene groups due to a low symmetry molecule. When attempting to cleanly isolate this intermediate, small quantities of complex 2a were always present, despite not being detected in the reaction mixture, precluding complete characterization of this species. All crystallisation attempts yielded only crystals of 2a.

Two possible formulations for this observed intermediate were considered, and are shown in Scheme 2. Oxidative addition of 1a to [RhCl(PPh3)3] may yield a six-coordinate σ-2-perimidinyl rhodium(III) species (3a), which subsequently eliminates dihydrogen and triphenylphosphine to form 2a. There are certainly examples of double C–H activation via such a two-step process.66 Furthermore, single C–H activation to form σ-2-perimidinyl complex [IrHCl(CO){CH(NCH2PCy2)2C10H6}] was observed in the reaction between 1b and [IrCl(CO)(PPh3)2].54


image file: c8gc03094d-s2.tif
Scheme 2 Reactions of H2C(NCH2PPh2)2C10H6 with [RhCl(PPh3)3], showing proposed intermediates in blue.

Alternatively, the reaction between 1a and [RhCl(PPh3)3] may release HCl to form the five-coordinate carbene complex [RhH(PPh3){C(NCH2PPh2)2C10H6}] 4a, which could subsequently react with the free HCl and lose H2 and PPh3 to afford 2a. A number of observations support 4a as the likely formulation of the intermediate species. Firstly, the formation of a pentacoordinate intermediate is supported by the 1JRh–P coupling of the diphenylphosphino 31P resonance at δP = 61, which shows a large value of 133 Hz. This is slightly reduced relative to that in the four-coordinate 2a (1JRh–P = 153 Hz), though significantly larger than those observed for a related hexacoordinate rhodium(III) complex (vide infra), which gives a value of around 100 Hz. Secondly, the mass spectra of the intermediate showed strong peaks consistent with complex 4a. Furthermore, a later experiment was performed in which 1a was treated with n-butyllithium, in order to deprotonate the central methylene group, prior to addition of [RhCl(PPh3)3]. The 31P{1H} and 1H NMR spectra of the resulting product are very similar to those of the observed intermediate in the preparation of 2a, but again attempts to isolate a pure sample of this product were plagued by the inevitable formation of 2a. Nonetheless, this result suggests that 4a is the product of this reaction, as well as the tenuous intermediate.

As noted, the formation of 2b (10 min) is much more rapid than that of 2a (21 h), and no intermediate was observed by NMR spectroscopy. The persistence of intermediate species 4a in the latter reaction is presumably a consequence of both the reduced steric profile and σ-basicity of the -PPh2 groups relative to their cyclohexyl analogues, which decreases the steric and electronic incentive for expulsion of a σ-donor ligand such as PPh3.

CO2 incorporation into amines

Catalyst screening for formylation reactions. The per-NHC rhodium complexes 2a and 2b, together with a small selection of locally available rhodium and iridium complexes—some possessing strong σ-donor ligands—were screened for their catalytic activity in the formylation of aniline using CO2 with phenylsilane as the reducing agent in THF at 50 °C (Chart 1). Wilkinson's catalyst (7) achieved 15% conversion to the formanilide product. No desired product was found when using Vaska's complex (8) as catalyst. Rhodium complexes bearing NHC-triazole bidentate ligand (9a and 9b),67 or N,N-bidentate ligands (1068 and 1169–71 were also found to be inactive as catalysts. Interestingly, very different conversions were achieved when testing the per-NHC complexes 2a and 2b. Complex 2a was inactive as a catalyst; whereas an excellent 93% conversion was achieved using complex 2b as the catalyst. Similar ligand effect for the formylation reaction was recently reported by He.72 An isolated yield of 86% was obtained after purification by column chromatography. X-ray structures of 2a and 2b displayed similar bond lengths and bite angles about the Rh centres in these two complexes. The markedly different conversions are therefore most likely a result of different σ-donating ability of the phosphine ligand motifs (-PPh2vs. -PCy2), given that the shape of the cavity through which reactants approach the Rh centre during catalysis would be expected to be more size-exclusive for 2b than for 2a. When no catalyst was added to the reaction mixture, no product formation was observed (Scheme 1).
image file: c8gc03094d-c1.tif
Chart 1 Catalyst screen for the formylation of aniline (5a) with CO2. Conversions were determined with 1H NMR spectroscopy using 1,1,2,2-tetrachloroethane as internal standard.
Conditions screening for formylation reactions. With a suitable catalyst identified, we proceeded to optimise the reaction conditions for the formylation of amines. Aniline was used as a model amine substrate. Highest conversion (93% conv.; 86% isolated yield) was obtained using THF as solvent, 1 equiv. of PhSiH3 as reductant, and the reaction temperature of 50 °C. Decreasing the reaction temperature to 25 °C resulted in a slight drop in conversion to the desired product to 83%. Replacing THF (50 °C) with toluene (90 °C) maintained high conversion (91%). Lowering the reaction temperature of toluene to 50 °C, however, led to a significant drop in conversion to 56%. Acetonitrile was also tested as solvent for the reaction but did not yield any desired product. The tests above are summarised in the ESI, Tables S1 and 2. Chlorinated solvents were obviated as complex 2b was found to lead to side reactions (vide infra).

Different hydrosilanes were also tested in THF at 50 °C. Replacing phenylsilane with diphenylsilane decreased the conversion from 93% to 36%. Polymethylhydrosiloxane (PMHS)—a by-product in the chemical industry—was an attractive candidate which unfortunately did not react.

A catalyst loading evaluation demonstrated that 5 mol% was the optimal catalyst loading, as lower loadings led to high silane to rhodium ratios and thus reduced reaction rates (see the stoichiometric studies section, and ESI).

The effect of using chlorinated solvent. Given that the rhodium(III) complex [RhCl2(CH2Cl){C(NCH2CH2PPh2)2C2H2}] (12a) has previously been reported to result from oxidative addition of dichloromethane to an (inferred) 16-electron rhodium(I) species,73 we anticipated that our 16-electron rhodium(I) complexes might similarly activate a C–Cl bond. Though complex 2b appeared to be stable in dichloromethane for several hours at ambient temperature, at reflux 2b cleaves a dichloromethane C–Cl bond to afford the Rh(III) complex [RhCl2(CH2Cl){C(NCH2PCy2)2C10H6}] (12b, Scheme 3). The reaction took place over 20 hours, product formation being evident from the development of a doublet resonance in the 31P{1H} NMR spectrum at δP = 22.1, with the coupling constant of 1JRh–P = 99 Hz, cf. 148 Hz for 2b, reflecting the octahedral geometry at rhodium.
image file: c8gc03094d-s3.tif
Scheme 3 Oxidative addition of dichloromethane to 2b.

The 1H NMR spectrum of 12b in CDCl3 showed one multiplet peak at δH = 4.45, which was attributed to both the PCH2 groups and the CH2Cl protons based on relative integrals. In C6D6, however, these resonances separated to give two doublets at δH = 4.24 and 4.31 for the diastereotopic PCH2 protons, indicating cis coordination of the two chlorides, and a triplet of doublets at δH = 4.93 corresponding to CH2Cl, with 3JP–H and 2JRh–H coupling constants of 7 and 2 Hz, respectively (cf. δH = 3.61, 3JP–H = 8 Hz, 2JRh–H = 3 Hz in CDCl3 for the reported complex 12a). The CH2Cl moiety appears in the 13C{1H} NMR spectrum as a doublet of triplets at δC = 37.2 (1JRh–C = 30 Hz, 2JP–C = 8 Hz), while the carbene resonance, also a doublet of triplets, is located at δC = 205.8 (1JRh–C = 50 Hz, 2JP–C = 5 Hz), close to that of the precursor 2b.

Formylation reaction scope. Having established that aniline (5a) undergoes formylation under the optimised reaction conditions, a range of substrates was tested to gauge the impact of different groups (Chart 2). Halogens (F to I) were tolerated giving good to excellent yields (6b–e). Extremely electron withdrawing 4-trifluoromethyl aniline (5f) gave a moderate conversion of 66%. 4-Nitroaniline (5g) resulted in no reaction under the reducing condition. Benzamide gave 20% conversion to the desired product (6h), and no side-reactions were observed. High or near complete conversions were obtained using substrates with electron donating groups (5i–j). Primary aliphatic and benzyl amine substrates (5k–n) led to low to medium conversions. Heterocycles (5o & 5p) were unreactive.
image file: c8gc03094d-c2.tif
Chart 2 N-Formylation of amines with CO2. Amine substrates (0.5 mmol), PhSiH3 (0.5 mmol, 1 equiv.) and catalyst 2b (25 μmol, 5 mol%) were heated with stirring in THF (2 mL, 50 °C) under CO2 atmosphere (balloon). Reactions were quenched by cooling to r.t. and adding deionised water (40 μl, 2.2 mmol). 1,1,2,2-Tetrachloroethane (53 μL, 0.5 mmol) was then added as internal standard and the reaction mixtures were analysed using 1H NMR spectroscopy. Isolated yield in parenthesis.

Secondary amine substrates were considered next. High conversion was obtained using N-methylaniline (13a) as substrate, and medium conversions were observed for piperidine and morpholine substrates (5r–s). N,N-Diphenylamine, N,N-diisopropylamine and 2,2,6,6-tetramethylpiperidine (5q, 5t–u) were unreactive, likely due to the steric congestion about the nitrogen atom. N-Benzylethylamine and N,N-dibenzylamine (13n and 5v) afforded the desirable products with similar conversions to those observed using benzylamine (5n).

Conditions screening for methylation reactions. With the success in formylation reactions established, extension of the scope of catalysed reactions to include methylation of amines was explored (Scheme 4), with the expectation that this would require stronger reducing conditions to convert the formamide group into a methylamino group. Starting with the optimal conditions for formylation of aniline (THF, 50 °C), the amount of phenylsilane was increased to 4 equivalents with respect to the aniline substrate (Table 1). This resulted in a substantial improvement in the formation of methylanilines (13a and 15a) vs. formanilides (6a and 14a). At the same time, this increase in phenylsilane resulted in a significant reduction in the overall conversion to the products from 93% (1 equiv. PhSiH3) to 31% (4 equiv. PhSiH3). When 4 molar equivalents of Ph2SiH2 was used instead of 4 equiv. of PhSiH3, no substantial change in the formation of methylanilines was observed.
image file: c8gc03094d-s4.tif
Scheme 4 Methylation of amines.
Table 1 Optimisation of reaction conditions for N-methylation of amines, using aniline as test substrate (cf.Scheme 4). The benchmark reaction condition for subsequent amine methylation is highlighted in green
Entry Solvent Hydrosilane Equiv. Temp. (°C) Conv. (%) (6a[thin space (1/6-em)]:[thin space (1/6-em)]13a[thin space (1/6-em)]:[thin space (1/6-em)]14a[thin space (1/6-em)]:[thin space (1/6-em)]15a) Total Conv. (%)
1 THF PhSiH3 1 50 93[thin space (1/6-em)]:[thin space (1/6-em)]<2[thin space (1/6-em)]:[thin space (1/6-em)]<2[thin space (1/6-em)]:[thin space (1/6-em)]<2 93
2 4 50 <2[thin space (1/6-em)]:[thin space (1/6-em)]19[thin space (1/6-em)]:[thin space (1/6-em)]<2[thin space (1/6-em)]:[thin space (1/6-em)]12 31
3 Ph2SiH2 1 50 36[thin space (1/6-em)]:[thin space (1/6-em)]<2[thin space (1/6-em)]:[thin space (1/6-em)]0[thin space (1/6-em)]:[thin space (1/6-em)]0 36
4 4 50 65[thin space (1/6-em)]:[thin space (1/6-em)]4[thin space (1/6-em)]:[thin space (1/6-em)]0[thin space (1/6-em)]:[thin space (1/6-em)]0 69
5 Tol PhSiH 3 1 90 91[thin space (1/6-em)]:[thin space (1/6-em)]0[thin space (1/6-em)]:[thin space (1/6-em)]7[thin space (1/6-em)]:[thin space (1/6-em)]0 98
6 2 90 55[thin space (1/6-em)]:[thin space (1/6-em)]2[thin space (1/6-em)]:[thin space (1/6-em)]10[thin space (1/6-em)]:[thin space (1/6-em)]3 70
7 4 90 28[thin space (1/6-em)]:[thin space (1/6-em)]14[thin space (1/6-em)]:[thin space (1/6-em)]<[thin space (1/6-em)]2:[thin space (1/6-em)]47 89
8 8 90 0[thin space (1/6-em)]:[thin space (1/6-em)]36[thin space (1/6-em)]:[thin space (1/6-em)]0[thin space (1/6-em)]:[thin space (1/6-em)]37 73
10 Ph2SiH2 4 90 12[thin space (1/6-em)]:[thin space (1/6-em)]29[thin space (1/6-em)]:[thin space (1/6-em)]3[thin space (1/6-em)]:[thin space (1/6-em)]36 80
11 Ph3SiH 90 32[thin space (1/6-em)]:[thin space (1/6-em)]0[thin space (1/6-em)]:[thin space (1/6-em)]<2[thin space (1/6-em)]:[thin space (1/6-em)]3 35


To improve the overall conversion, toluene was tested as the reaction solvent in order to raise the reaction temperature to 90 °C. Phenylsilane was added in three separate experiments (2, 4 and 8 equiv.) and the product distribution was found to be affected by the amount of PhSiH3 added. Highest conversion with formation of the largest amount of N,N-dimethylaniline product (15a) was obtained when using 4 equivalents of PhSiH3 (Table 1, entry 7, used as benchmark conditions for subsequent methylation reactions of amines). The highest selectivity for the methylation products 13a and 15a was achieved when using 8 equivalents of PhSiH3, but this came at the cost of overall conversion, which was significantly reduced (Table 1, entry 8).

The use of 4 equiv. of Ph2SiH2 (instead of 4 equiv. of PhSiH3) resulted in a similar product distribution to that observed for using PhSiH3, but with a lower conversion (Table 1, entry 10); triphenylsilane afforded 32% of formanilide and the reaction only formed methylated products to a minimal extent.

Methylation reaction scope. The scope of the methylation protocol was investigated for a range of amine substrates using 4 equiv. of PhSiH3, toluene solvent (90 °C) and CO2 atmosphere (balloon). Methylamine products were favoured over formamide products in most cases (Chart 3). Anilines with halogen substituents at the para-position (5b–f) resulted in lower yields. Apart from 4-fluoroaniline (5b) which afforded formamide and methylanilines (6b, 13b and 15b), these substrates gave methylated products only in low to medium conversions. The nitro group on 4-nitroaniline (5g) was subjected to reduction, and further reacted under the methylation reaction conditions as observed in the 1H NMR spectra. The main products for this entry could not be identified. Interestingly, benzamide (5h) did not react to give the desired product or decompose when used as a substrate under the same reaction conditions.
image file: c8gc03094d-c3.tif
Chart 3 N-Methylation of amines with CO2. Amine substrates (0.5 mmol), PhSiH3 (2 mmol, 4 equiv.) and catalyst 2b (25 μmol, 5 mol%) were heated with stirring in toluene (2 mL, 90 °C) under CO2 atmosphere (balloon). Reactions were quenched by cooling to r.t. and adding deionised water (40 μl, 2.2 mmol). 1,1,2,2-Tetrachloroethane (53 μL, 0.5 mmol) was then added as internal standard and the reaction mixtures were analysed using 1H NMR spectroscopy (CDCl3). Isolated yield in parenthesis. Results marked with asterisks (*) were obtained using 8 equiv. of PhSiH3 (4 mmol). The products 6, 13 and 14 are only added to the chart if they formed. The unaccounted material was either starting material (major) or unidentified side-products.

Aromatic substrates with electron donating substituents were then explored. Using 2,4,6-trimethylaniline (5i) as substrate afforded the dimethylated product cleanly with high conversion. 4-Methoxyaniline (5j), however, led to a mixture of formamide and methylaniline products. Hexylamine (5k) afforded N,N-dimethylhexylamine in high conversion. A mixture of products (13–15n) was obtained on using benzylamine (5n) as substrate. Indole (5p) was inactive as a substrate, as in the catalysed formylation reaction.

Secondary amine substrates were also investigated. N-Methylaniline (13a) led to N,N-dimethylaniline (15a) in high conversion. Diphenylamine (5q), however resulted in a significant drop in product conversion. Similar observations were also made on comparing the results of using piperidine (5r) and 2,2,6,6-tetramethylpiperidine (5u) substrates. This is, as before, likely due to steric hindrance about the nitrogen atom of the amine substrates impeding catalysis. Morpholine (5s) gave 33% of the target product. N-Benzylmethylamine (13n) and dibenzylamine (5v) led to a mixture to formylated and methylated products in medium to high conversions.

For the substrates that led to the methylation products, as well as the formylation products, the reactions were repeated using 8 equiv. of PhSiH3. Although the benchmark optimisation reactions (Table 1) showed that this leads to lower conversions, it also showed that higher selectivity towards the methylation products could be achieved. Indeed the reaction with 4-fluoroaniline (5b) and 4-methoxyaniline (5j) led to substantial improvements in the selectivity for the corresponding methylated products (13b, 13j, 15b and 15j). Benzylamine (5n) resulted selectively in the monomethylated product 13n. Using piperidine (5r) and N-benzylmethylamine (13n) as substrates led to improvement in selectivity towards the methylated products (13r and 15n) but still some formylation products formed. The dibenzyl substrate, N,N-dibenzylamine 5v resulted in clean conversion to the methylated product 13v. As expected, some reagents led to lower conversions when higher amounts of PhSiH3 were used, but interestingly in the case of 13n and 5v no substantial decrease was observed and in the case of 5r a significant increase in the overall conversion was obtained.

Stochiometric studies

Reactions of 2b with CO2, CS2 and CO. To gain insight to what processes may be in operation during catalysis, 2b was exposed to carbon dioxide. In addition, CS2 and CO were investigated to better understand the interactions of small molecules with the rhodium centre of 2b. In all cases, quantitative conversion to new products was observed as shown by 31P NMR spectroscopy. Removal of solvent under reduced pressure, however, resulted in reversion to the starting complex 2b.

Exposure of a benzene solution of 2b to CO2 atmosphere did not lead to noticeable colour change. The 31P{1H} NMR spectrum, however, showed a new sharp doublet at δP = 30.0 with 1JRh–P = 96 Hz; whereas the 1H NMR spectrum (500 MHz) displayed three new distinct doublet peaks at δH = 3.83, 4.14, and 6.10, and two broad signals at δH = 2.53–2.62, and 2.72–2.88 which could indicate dynamic exchange of coordinated and free CO2 in the solution. Poor solubility of 2b precluded the acquisition of informative 13C{1H} NMR data.

To better understand how CO2 might interact with 2b during catalysis, the analogous CS2 molecule was used to gain insights. In contrast to reaction with CO2, the formation of the CS2 adduct [RhCl(η2-CS2){C(NCH2PCy2)2C10H6}] (16b) from 2b was well supported by NMR spectroscopic and X-ray crystallographic data. When a suspension of 2b in C6D6 was treated with carbon disulfide, a rapid colour change from bright orange to pale orange was observed. The 31P{1H} NMR spectrum showed a sharp doublet at δP = 31.3 with 1JRh–P = 103 Hz, while the 1H spectrum displayed two distinct doublet peaks at δH = 4.32 and 4.40. The 13C{1H} NMR spectrum confirmed CS2 coordination, displaying a rhodium-coupled doublet peak at δC = 260.3 (1JRh–C = 12 Hz), close to those of literature examples of 18-electron Rh(I)(η2-CS2) complexes which give resonances at δC = 251.1[thin space (1/6-em)]74 and 256.9.75 The small Rh–C coupling (cf. 24.1–38.8 Hz in reported complexes) may be a result of the weak binding of the CS2 ligand to rhodium, as evidenced by its ready dissociation. A doublet carbene resonance was also observed at δC = 212.8 (1JRh–C = 60 Hz). The anticipated carbene 13C–31P coupling was too small to be resolved. Infrared data for 16b in solution were obtained using neat carbon disulfide as a solvent. However, the region in which the ν(C[double bond, length as m-dash]S) and ν(RhCS) modes of an η2-CS2 complex should be observed (ca. 990–1280 cm−1 and 630–650 cm−1, respectively76) includes numerous bands of the pincer ligand, as observed in the IR spectrum of the precursor 2b and the free ligand 1b, such that no unambiguous assignments could be made.

Although 16b was highly air sensitive and readily lost CS2, crystals of the complex that formed slowly in the reaction mixture proved stable enough for X-ray structural characterisation. The crystal structure (Fig. 3) confirms the η2-coordination of the CS2 ligand through one of the C[double bond, length as m-dash]S bonds. The adopted stereochemistry places the π-acidic C70 pseudo-trans to the sole π-donor (Cl). The ring system of the pincer ligand exhibits a twist angle of 24.8° relative to the coordination plane, larger than observed in the structures of the square-planar complexes 2a (19.8°)54 and 2b (essentially no twist) presumably to accommodate the additional ligand. The Rh–C, Rh–Cl and Rh–P bond distances are significantly larger than those in 2a and 2b, as expected due to the increased coordination number. The Rh–C70 bond length is within the range observed for the few structurally characterised examples of Rh(η2-CS2) complexes,74,77,78 though the Rh–S1 bond length is significantly longer than the range of 2.3662(11)–2.387(5) Å observed in these cases, perhaps reflecting the aforementioned weak binding.


image file: c8gc03094d-f3.tif
Fig. 3 Molecular structure of 16b in a crystal of 16b·2(C6H6) (aryl and cyclohexyl hydrogen atoms omitted, 50% displacement ellipsoids). Positional disorder was observed between the CS2 and Cl ligands, and this was modelled as two distinct positions for both ligands with partial occupancies of 0.9 and 0.1. The CS2 and Cl ligands with 0.9 occupancies are shown. Selected bond lengths (Å) and angles (°): Rh1–C1 = 1.999(5), C1–N1 = 1.359(6), C1–N2 = 1.353(6), Rh1–P1 = 2.3074(13), Rh1–P2 = 2.3233(13), Rh1–Cl1 = 2.588(2), Rh1–C70 = 1.998(7), Rh1–S1 = 2.4437(15), S1–C70 = 1.704(6), S2–C70 = 1.569 (7), C1–Rh1–P1 = 81.94(14), C1–Rh1–P2 = 82.43(14), P1–Rh1–P2 = 164.32(5).

Placing a suspension of 2b in benzene under an atmosphere of carbon monoxide resulted in an immediate colour change from orange to yellow,79 accompanied by replacement of the 31P{1H} NMR resonance of 2b with a broad doublet at δP = 52.5 (1JRh–P = 130 Hz). The methylene protons appeared as a broad singlet peak in the 1H NMR spectrum at δH = 5.25. The broad peaks suggest a dynamic process in which the CO ligand is coordinating to the metal centre to form [RhCl(CO){C(NCH2PCy2)2C10H6}] (17b) and dissociating to reform 2b (Scheme 5). This exchange precluded the acquisition of informative 13C NMR spectral data. Furthermore, attempts to isolate the product resulted in either reversion to 2b, or apparent decomposition.80 Solution IR spectroscopy of the reaction mixture showed a broad, weak band in the carbonyl region at 1988 cm−1 with a shoulder at 2000 cm−1. The weakness of the bands and the fact that there were two and not one may suggest that the sample was decomposing under the nitrogen atmosphere in the infrared cell. Given the poor quality of these spectroscopic data, the proposed formation of 17b remains speculative. Another possible formulation is that CO had displaced the chloride anion to give a dicarbonyl cationic species, which may account for the second infrared CO stretching band, though such a species would be expected to exhibit very low solubility in the nonpolar benzene solvent.


image file: c8gc03094d-s5.tif
Scheme 5 Reversible reactions of complex 2b with CS2 and CO.

Studies of catalytic reaction routes

In addition to stoichiometric studies of 2b with CO2, and the analogous CS2 and CO, the reactions of 2b with the other substrates present during the formylation and methylation reactions were investigated. We undertook this to obtain an understanding of mechanism and product selectivity during the catalysed reactions.

A commonly held reaction mechanism for the N-formylation and N-methylation of amines with CO2 starts with metal catalyst promoted hydrosilylation of CO2 to silyl formate. There are then two possible routes leading to the final formation of the N-methylamines where the amine substrate reacts either with the silyl formate intermediate (Pathway 2) or the reduced silyl ether intermediate (Pathway 1, Fig. 4).32,81–83


image file: c8gc03094d-f4.tif
Fig. 4 Proposed reaction pathways for N-formylation and -methylation of amines with CO2 and phenylsilane. Pathway 1 (green): reduction of silyl formate to silyl ether, then reaction with amine substrate to form the final methylamine product; Pathway 2 (pink): formation of formamide by reacting amine substrate with silyl formate, then reduction of formamide to methylamine. In this work, the transition metal catalyst [M] is complex 2b.

On treating complex 2b with 2 molar equivalents of aniline in toluene-d8, no new products were formed – with no change observed in either the 1H or 31P{1H} NMR spectra at room temperature or after heating at 90 °C. The same observations were also noted when aniline was added under the CO2 atmosphere. On treating complex 2b with PhSiH3 in deuterated toluene a mixture of products was formed, including multiple rhodium species containing metal bound hydrides, as confirmed by 1H and 31P NMR spectroscopy (ESI).

The treatment of 2b with 2 molar equivalents of PhSiH3 under a CO2 atmosphere was then undertaken in order to further evaluate the plausibility of the general mechanism proposed in Fig. 4. A number of products formed, and based on 1H NMR spectra of the products at room temperature, the new species included silyl formate in small quantities and the silyl ether intermediate (Fig. 4 and Fig. S7). The formation of silyl formate was confirmed on observation of formic acid following hydrolysis of the products using 1H NMR spectroscopy. By 31P NMR spectroscopy, the formation of Rh bound reaction intermediates at room temperature was confirmed, however, these could not be isolated (Fig. S8). Upon raising the reaction temperature to 70 °C, all the new peaks that were seen in the 31P{1H} NMR spectrum at room temperature disappeared and the only peak present in the 31P{1H} NMR spectrum at that temperature was the doublet corresponding to the starting complex 2b. This suggested that the reaction of 2b with PhSiH3 and CO2 was rapid, and that the integrity of 2b was re-established after the reaction, indicating the potential for 2b as a recyclable catalyst.

Having established that 2b can readily promote the formation of silyl formate on reaction with CO2 and PhSiH3, and subsequently silyl ether, thus paving the way to the methylated product, we proceeded to explore whether the final N-methylamine product can form through Pathway 1 as part of the mechanism (Fig. 4). As a first step, PhSiH3 was added to complex 2b under a CO2 atmosphere, and the mixture was allowed to react for 16 h at 90 °C under analogous conditions to those used for the methylation catalysis. The 1H NMR spectrum of the reaction mixture showed the presence of silyl ethers, free PhSiH3 and no silyl formate (Fig. S9). The CO2 atmosphere was removed from the system with 3 freeze–pump–thaw cycles. Aniline (5a) was then added and allowed to react for another 16 h, affording N-methylaniline (13a) and N,N-dimethylaniline (15a), with some unreacted aniline present (Fig. S10–12). This suggests that the methyl group could be transferred from the intermediate silyl ethers onto the nitrogen atom of amine substrates as shown in proposed Pathway 1 (Fig. 4). No formylation products were observed in the reaction mixture, showing that Pathway 1 is not reversible.

In order to investigate the viability of Pathway 2 (Fig. 4) the N-formyl reaction product, formanilide (6a) was used as a substrate (Scheme 6). Under a nitrogen atmosphere, with no additional CO2 present, 3 molar equivalents of PhSiH3 were added to 6a in the presence of the catalyst 2b. This led to a clean reduction reaction that resulted in the corresponding N-methylaniline product (13a, Scheme 6), indicating that Pathway 2 of the mechanism is also plausible as a route for the methylation reaction.


image file: c8gc03094d-s6.tif
Scheme 6 Reduction of formanilide (6a) to N-methylaniline (13a) in the presence of PhSiH3 and complex 2b.
Time course studies. The mechanistic pathways for the methylation reaction (Scheme 4) were further investigated using a time course study of the reaction under catalytic conditions (toluene at 90 °C; 5 mol% of 2b, PhSiH3 (4 equiv.) and a CO2 atmosphere) (Fig. 5). In the initial stages of the reaction (up to 5 h), where the concentration of PhSiH3 will be highest, only the methylation products N-methylaniline (13a) and N,N-dimethylaniline (15a) were observed. Significant amounts of the formylation products 6a and 14a formed in the final part of the time course (5–7 h), where the PhSiH3 concentration was lower. The only product that decreased in concentration between 5 h and 7 h (22% vs. 3% respectively) was the methylation product 13a which was consumed as a precursor to 14a.
image file: c8gc03094d-f5.tif
Fig. 5 Reaction profile of N-methylation of aniline in toluene at 90 °C; ca. 5 mol% of 2b, PhSiH3 (4 equiv.) and a CO2 balloon was used. 1,3,5-trimethoxybenene (0.2 equiv.) was added as internal standard. The reaction was monitored by 1H NMR spectroscopy.

These data match well with the mechanistic proposal (Fig. 4), where both pathways require PhSiH3 for the formation of the N-methylation products. When the relative concentration of PhSiH3 drops, the reduction of the N-formyl products is slowed down, and build-up of the relative concentration of 6a and 14a occurs. This also links to the data seen when comparing the effect of different molar equivalences of PhSiH3 used for the methylation reaction (Table 1, entries 5–8). Lower relative ratios of PhSiH3 promoted the formylation products, while higher molar equivalences of PhSiH3 resulted in the methylation products as the mayor or exclusive species formed.

Observations gathered from the mechanistic investigations, scope studies and the screening of reaction conditions were in line with reaction pathways proposed herein and in the literature.32,81–84 Both proposed pathways in Fig. 4 were shown to be viable under the reaction conditions of catalysed methylation of amines; it is also clear that product conversions and ratios are highly dependent on the identity of the substrates but also the silane reagent and the relative quantity of it. Under the catalysis conditions, catalyst 2b remained intact (resting state) as a four-coordinate, chloride bound complex after catalysis as shown in spectroscopic studies.

Conclusions

In conclusion, the formation of the per-NHC rhodium complex 2b is described as well as an in-depth structural characterisation of 2b, following up on its’ initial report in a preliminary communication.54 The complex 2b was found to be an active catalyst for the N-formylation and -methylation of a wide range of amine substrates using CO2 as the carbon source. Selectivity between the formylation and methylation reactions was achieved by varying the solvent, temperatures and quantity of hydrosilane added, and has the potential to lead to control of product formation.

The reactivity of 2b was probed using a range of small molecules as well as reagents for methylation of aniline. Complex 2b was found to reversibly coordinate CS2, CO2 and CO on the basis of crystallographic and spectroscopic data; 2b can also activate C–Cl and Si–H bonds via oxidative addition, making it a potential catalyst for other reactions. In the context of N-methylation of aniline, 2b reacted with PhSiH3 and CO2, but not aniline. Rhodium-mediated hydrosilylation of CO2 was likely the first step of the N-methylation reaction catalysed by 2b and precedes attack by aniline (or other amine substrates). Plausible pathways for the 2b catalysed formylation and methylation of amines have been proposed by examining the results of stoichiometric studies and the reaction profile as well as results from reaction condition optimisation.

Conflicts of interest

There are no conflicts to declare.

Acknowledgements

We gratefully acknowledge the Australian Research Council (DP110101611) for funding. R. H. L. acknowledges Macquarie University Research Training Pathway Scholarship.

Notes and references

  1. Q. Liu, L. Wu, R. Jackstell and M. Beller, Nat. Commun., 2015, 6, 5933–5947 CrossRef PubMed.
  2. P. Braunstein, D. Matt and D. Nobel, Chem. Rev., 1988, 88, 747–764 CrossRef CAS.
  3. Y. Li, X. Cui, K. Dong, K. Junge and M. Beller, ACS Catal., 2017, 7, 1077–1086 CrossRef CAS.
  4. T. Sakakura and K. Kohno, Chem. Commun., 2009, 1312–1330 RSC.
  5. M. Aresta, A. Dibenedetto and A. Angelini, Chem. Rev., 2014, 114, 1709–1742 CrossRef CAS PubMed.
  6. H. Kolbe and E. Lautemann, Annalen, 1859, 113, 125–127 CrossRef.
  7. R. Schmitt and E. Burkard, Ber. Dtsch. Chem. Ges., 1877, 2699 Search PubMed.
  8. E. Blondiaux, J. Pouessel and T. Cantat, Angew. Chem., Int. Ed., 2014, 53, 12186–12190 CrossRef CAS PubMed.
  9. O. Jacquet, C. Das Neves Gomes, M. Ephritikhine and T. Cantat, ChemCatChem, 2013, 5, 117–120 CrossRef CAS.
  10. Y. Li, X. Fang, K. Junge and M. Beller, Angew. Chem., Int. Ed., 2013, 52, 9568–9571 CrossRef CAS PubMed.
  11. F. D. Bobbink, S. Das and P. J. Dyson, Nat. Protocols, 2017, 12, 417–428 CAS.
  12. H. G. Grant and L. A. Summers, Aust. J. Chem., 1980, 33, 613–617 CrossRef CAS.
  13. B. C. Chen, M. S. Bednarz, R. Zhao, J. E. Sundeen, P. Chen, Z. Shen, A. P. Skoumbourdis and J. C. Barrish, Tetrahedron Lett., 2000, 41, 5453–5456 CrossRef CAS.
  14. K. Kobayashi, S. Nagato, M. Kawakita, O. Morikawa and H. Konishi, Chem. Lett., 1995, 24, 575–576 CrossRef.
  15. C. Fang, C. Lu, M. Liu, Y. Zhu, Y. Fu and B.-L. Lin, ACS Catal., 2016, 6, 7876–7881 CrossRef CAS.
  16. B. Testa, P.-A. Carrupt, P. Gaillard, F. Billois and P. Weber, Pharm. Res., 1996, 13, 335–343 CrossRef CAS.
  17. E. S. Istvan and J. Deisenhofer, Science, 2001, 292, 1160–1164 CrossRef CAS PubMed.
  18. U. Fagerholm, Pharm. Res., 2008, 25, 625–638 CrossRef CAS PubMed.
  19. A. Malkia, L. Murtomaki, A. Urtti and K. Kontturi, Eur. J. Pharm. Sci., 2004, 23, 13–47 CrossRef CAS PubMed.
  20. E. J. Barreiro, A. E. Kummerle and C. A. M. Fraga, Chem. Rev., 2011, 111, 5215–5246 CrossRef CAS PubMed.
  21. H. Schoenherr and T. Cernak, Angew. Chem., Int. Ed., 2013, 52, 12256–12267 CrossRef CAS PubMed.
  22. E. Buchdunger, J. Zimmermann, H. Mett, T. Meyer, M. Mueller, U. Regenass and N. B. Lydon, Proc. Natl. Acad. Sci. U. S. A., 1995, 92, 2558–2562 CrossRef CAS.
  23. J. Zimmermann, E. Buchdunger, H. Mett, T. Meyer and N. B. Lydon, Bioorg. Med. Chem. Lett., 1997, 7, 187–192 CrossRef CAS.
  24. J. Zimmermann, E. Buchdunger, H. Mett, T. Meyer, N. B. Lydon and P. Traxler, Bioorg. Med. Chem. Lett., 1996, 6, 1221–1226 CrossRef CAS.
  25. J. Zimmermann, P. Furet and E. Buchdunger, ACS Symp. Ser., 2001, 796, 245–259,  DOI:10.1021/bk-2001-0796.ch015.
  26. F. Fache, L. Jacquot and M. Lemaire, Tetrahedron Lett., 1994, 35, 3313–3314 CrossRef CAS.
  27. R. A. da Silva, I. H. S. Estevam and L. W. Bieber, Tetrahedron Lett., 2007, 48, 7680–7682 CrossRef CAS.
  28. J. R. Harding, J. R. Jones, S.-Y. Lu and R. Wood, Tetrahedron Lett., 2002, 43, 9487–9488 CrossRef CAS.
  29. T. Maarten and R. Govind, EP375333A2, 1990 Search PubMed.
  30. O. Jacquet, C. Das Neves Gomes, M. Ephritikhine and T. Cantat, J. Am. Chem. Soc., 2012, 134, 2934–2937 CrossRef CAS PubMed.
  31. C. Das Neves Gomes, O. Jacquet, C. Villiers, P. Thuery, M. Ephritikhine and T. Cantat, Angew. Chem., Int. Ed., 2012, 51, 187–190 CrossRef CAS PubMed.
  32. O. Jacquet, X. Frogneux, C. Das Neves Gomes and T. Cantat, Chem. Sci., 2013, 4, 2127–2131 RSC.
  33. Y. Li, I. Sorribes, T. Yan, K. Junge and M. Beller, Angew. Chem., Int. Ed., 2013, 52, 12156–12160 CrossRef CAS PubMed.
  34. K. Beydoun, G. Ghattas, K. Thenert, J. Klankermayer and W. Leitner, Angew. Chem., Int. Ed., 2014, 53, 11010–11014 CrossRef CAS PubMed.
  35. K. Beydoun, T. vom Stein, J. Klankermayer and W. Leitner, Angew. Chem., Int. Ed., 2013, 52, 9554–9557 CrossRef CAS PubMed.
  36. L. Zhang, Z. Han, X. Zhao, Z. Wang and K. Ding, Angew. Chem., Int. Ed., 2015, 54, 6186–6189 CrossRef CAS PubMed.
  37. K. Beydoun, K. Thenert, E. S. Streng, S. Brosinski, W. Leitner and J. Klankermayer, ChemCatChem, 2016, 8, 135–138 CrossRef CAS.
  38. Z. Yang, B. Yu, H. Zhang, Y. Zhao, Y. Chen, Z. Ma, G. Ji, X. Gao, B. Han and Z. Liu, ACS Catal., 2016, 6, 1268–1273 CrossRef CAS.
  39. T. V. Q. Nguyen, W.-J. Yoo and S. Kobayashi, Adv. Synth. Catal., 2016, 358, 452–458 CrossRef CAS.
  40. T. V. Q. Nguyen, W.-J. Yoo and S. Kobayashi, Angew. Chem., Int. Ed., 2015, 54, 9209–9212 CrossRef CAS PubMed.
  41. X. Frogneux, O. Jacquet and T. Cantat, Catal. Sci. Technol., 2014, 4, 1529–1533 RSC.
  42. P. Daw, S. Chakraborty, G. Leitus, Y. Diskin-Posner, Y. Ben-David and D. Milstein, ACS Catal., 2017, 7, 2500–2504 CrossRef CAS.
  43. O. Santoro, F. Lazreg, Y. Minenkov, L. Cavallo and C. S. J. Cazin, Dalton Trans., 2015, 44, 18138–18144 RSC.
  44. H. Liu, Q. Mei, Q. Xu, J. Song, H. Liu and B. Han, Green Chem., 2017, 19, 196–201 RSC.
  45. S. Zhang, Q. Mei, H. Liu, H. Liu, Z. Zhang and B. Han, RSC Adv., 2016, 6, 32370–32373 RSC.
  46. K. Motokura, N. Takahashi, D. Kashiwame, S. Yamaguchi, A. Miyaji and T. Baba, Catal. Sci. Technol., 2013, 3, 2392–2396 RSC.
  47. L. Gonzalez-Sebastian, M. Flores-Alamo and J. J. Garcia, Organometallics, 2015, 34, 763–769 CrossRef CAS.
  48. Z.-Z. Yang, B. Yu, H. Zhang, Y. Zhao, G. Ji and Z. Liu, RSC Adv., 2015, 5, 19613–19619 RSC.
  49. I. Özdemir, B. Alici, N. Gurbuz, E. Cetinkaya and B. Cetinkaya, J. Mol. Catal. A: Chem., 2004, 217, 37–40 CrossRef.
  50. W. A. Herrmann, J. Schuetz, G. D. Frey and E. Herdtweck, Organometallics, 2006, 25, 2437–2448 CrossRef CAS.
  51. T. Tu, J. Malineni, X. Bao and K. H. Dötz, Adv. Synth. Catal., 2009, 351, 1029–1034 CrossRef CAS.
  52. H. Tsurugi, S. Fujita, G. Choi, T. Yamagata, S. Ito, H. Miyasaka and K. Mashima, Organometallics, 2010, 29, 4120–4129 CrossRef CAS.
  53. G. Choi, H. Tsurugi and K. Mashima, J. Am. Chem. Soc., 2013, 135, 13149–13161 CrossRef CAS PubMed.
  54. A. F. Hill and C. M. A. McQueen, Organometallics, 2012, 31, 8051–8054 CrossRef CAS.
  55. A. F. Hill and C. M. A. McQueen, Organometallics, 2014, 33, 1909–1912 CrossRef CAS.
  56. K. Verlinden and C. Ganter, J. Organomet. Chem., 2014, 750, 23–29 CrossRef CAS.
  57. W. P. Fehlhammer and W. Finck, J. Organomet. Chem., 1991, 414, 261–270 CrossRef CAS.
  58. P. Bazinet, G. P. A. Yap and D. S. Richeson, J. Am. Chem. Soc., 2003, 125, 13314–13315 CrossRef CAS PubMed.
  59. P. Bazinet, T.-G. Ong, J. S. O'Brien, N. Lavoie, E. Bell, G. P. A. Yap, I. Korobkov and D. S. Richeson, Organometallics, 2007, 26, 2885–2895 CrossRef CAS.
  60. A. F. Pozharskii and V. V. Dal'nikovskaya, Usp. Khim., 1981, 50, 1559–1560 CAS.
  61. For a regular pentagon and hexagon, the corresponding angles α are 72 and 60, respectively.
  62. V. M. Ho, L. A. Watson, J. C. Huffman and K. G. Caulton, New J. Chem., 2003, 27, 1446–1450 RSC.
  63. A. Prades, M. Poyatos, J. A. Mata and E. Peris, Angew. Chem., Int. Ed., 2011, 50, 7666–7669 CrossRef CAS PubMed.
  64. H. Valdes, M. Poyatos and E. Peris, Organometallics, 2013, 32, 6445–6451 CrossRef CAS.
  65. For a review of these methods see: E. Peris, Top. Organomet. Chem., 2007, 21, 83 CrossRef CAS.
  66. H. Werner, Angew. Chem., Int. Ed., 2010, 49, 4714–4728 CrossRef CAS PubMed.
  67. K. Q. Vuong, M. G. Timerbulatova, M. B. Peterson, M. Bhadbhade and B. A. Messerle, Dalton Trans., 2013, 42, 14298–14308 RSC.
  68. R. H. Lam, D. B. Walker, M. H. Tucker, M. R. D. Gatus, M. Bhadbhade and B. A. Messerle, Organometallics, 2015, 34, 4312–4317 CrossRef CAS.
  69. L. D. Field, B. A. Messerle, M. Rehr, L. P. Soler and T. W. Hambley, Organometallics, 2003, 22, 2387 CrossRef CAS.
  70. S. Burling, L. D. Field, B. A. Messerle and P. Turner, Organometallics, 2004, 23, 1714–1721 CrossRef CAS.
  71. L. P. Soler, PhD thesis, University of Sydney, 1999.
  72. X.-D. Li, S.-M. Xia, K.-H. Chen, X.-F. Liu, H.-R. Li and L.-N. He, Green Chem., 2018, 20, 4853–4868 RSC.
  73. J. Y. Zeng, M.-H. Hsieh and H. M. Lee, J. Organomet. Chem., 2005, 690, 5662–5671 CrossRef CAS.
  74. M. Doux, N. Mezailles, L. Ricard and P. Le Floch, Organometallics, 2003, 22, 4624–4626 CrossRef CAS.
  75. A. F. Hill, A. J. P. White, D. J. Williams and J. D. E. T. Wilton-Ely, Organometallics, 1998, 17, 3152–3154 CrossRef CAS.
  76. K. K. Pandey, Coord. Chem. Rev., 1995, 140, 37–114 CrossRef CAS.
  77. C. Bianchini, D. Masi, C. Mealli, A. Meli and M. Sabat, Organometallics, 1985, 4, 1014–1019 CrossRef CAS.
  78. E. Lindner, B. Keppeler, H. A. Mayer, K. Gierling, R. Fawzi and M. Steinmann, J. Organomet. Chem., 1996, 526, 175–183 CrossRef.
  79. Though by no means definitive, in many cases decolourization of d6 or d8 complexes often accompanies an increase in coordinative saturation.
  80. (16) Decomposition was suggested by peaks in the NMR spectra that matched those observed when 2b was exposed to air, despite all manipulations being carried out under nitrogen using standard Schlenk techniques. This decomposition product could not be well characterised, despite several attempts, but gave a distinctive 31P NMR doublet resonance at P = 30 (1JRh–P = 102 Hz) and most likely corresponds to a dioxygen adduct [RhCl(2-O2){C(NCH2PCy2)2C10H6}].
  81. W. Sattler and G. Parkin, J. Am. Chem. Soc., 2012, 134, 17462–17465 CrossRef CAS PubMed.
  82. R. Lalrempuia, M. Iglesias, V. Polo, P. J. Sanz Miguel, F. J. Fernandez-Alvarez, J. J. Perez-Torrente and L. A. Oro, Angew. Chem., Int. Ed., 2012, 51, 12824–12827 CrossRef CAS PubMed.
  83. K. Motokura, D. Kashiwame, A. Miyaji and T. Baba, Org. Lett., 2012, 14, 2642–2645 CrossRef CAS PubMed.
  84. M. Hulla, G. Laurenczy and P. J. Dyson, ACS Catal., 2018, 8, 10619–10630 CrossRef CAS.

Footnote

Electronic supplementary information (ESI) available. CCDC 898438. For ESI and crystallographic data in CIF or other electronic format see DOI: 10.1039/c8gc03094d

This journal is © The Royal Society of Chemistry 2019