Cyclometallated Ir(III), Rh(III) and Ru(II) complexes as catalysts for the cyclotrimerisation of 1,6-diynes with monoynes

A new series of Rh, Ir and Ru precatalysts for the [2+2+2] cyclotrimerisation of 1,6-diynes with monoynes is reported. The precatalysts are reduced in situ to the active catalysts by reduction with alcohols. The precatalysts activity is in the order Ru>Rh>Ir which reflects the ease of this reduction. The Rh and Ir precatalysts require temperature in excess of 140 o C allowing their preparation in 2-methoxymethanol at 125 o C. The mechanism of this process is discussed


Introduction
The exploitation of cyclometallated complexes in catalysis has recently evolved as a broad new strategy.A variety of palladacycles incorporating cyclometallated phosphines, 1 phosphites, 2 carbenes, 3 imines, 4 heterocycles, 5 thioethers, 6 and oximes 7 have been reported to catalyse carbon-carbon (Heck, Suzuki) and carbon-nitrogen bond forming processes with high turnover numbers. 8,9Additionally, chiral palladacycles have been shown to catalyse carbon-carbon bond forming processes such as the aldol reaction, 10 Michael addition 11 and cyclopropanation reactions 12 with high enantiomeric excesses.Studies on the synthesis and catalytic behaviour of orthometallated complexes of Rh(II) 13 , Rh(III) 14 , Ir(I) 15 , and Ru(II) 16 have revealed active catalysts of high efficiency.For example, Nishiyama et al. reported that the chiral orthometallated rhodium (III) complex 1 effects the catalytic enantioselective allylation of aldehydes. 17Other useful applications of the cyclometallated transition metal complexes include electroluminescent/photoluminescent devices 18 and antibacterial agents. 19 1 Rn Grigg was Chairman of the RSC Heterocyclic Group during the period 1983-1985.

Results and Discussion
In this paper we report the synthesis of cyclometallated Rh(III), Ir(III) and Ru(II) complexes and their application as precatalysts for cyclotrimerisation of 1,6-diynes with monoynes.These were postulated as potential precatalysts that upon in situ reduction would liberate coordinatively unsaturated non-phosphine ligated M(I)/M(0) catalysts.Initally we explored the synthesis of ortho-metallated complexes of Rh(III) and Ir(III) with N-phenylpyrazole as the ligand (for the first generation catalysts) employing the protocol used by Nonoyama for their synthesis. 20Thus the corresponding transition metal chloride and N-phenylpyrazole in 2-methoxyethanol under reflux afforded the dimeric complexes 2a,b in 55-65% yield.Attempts to prepare cycloruthenated complexes by the same method failed, yielding a black powdery substance believed to Ru black.Hence, the ortho-metallated ruthenium(II) complex 4 was prepared in 61% yield by transmetallation of the corresponding mercury(II) complex 3 with [Ru(p-cymene)Cl 2 ] 2 in acetonitrile at room temperature (Scheme1) 21 or stirring [Ru(p-cymene)Cl 2 ] 2 and N-phenyl pyrazole in the presence of sodium acetate in DCM at room temperature which afford 4 in 73% yield Confirmation of the structure of 4 was provided by X-ray crystallography (Figure 1).Next we varied the electronic/steric properties of the ligands employing 2phenylbenzothiazole, N-(m-methoxyphenyl)pyrazole, and 3,5-dimethyl-2-phenylpyrazole as ligands in the preparation of cyclometallated complexes 5 and 6 as second generation precatalysts.Dimeric cyclometallated complexes 5a,b were obtained in 77-86% yield.Complex 6 was obtained in 45% yield and its X-ray crystal structure established that it was a monomer (Figure 2).Steric hindrance of the methyl groups on the pyrazole ring of the ligand prevent formation of the dimeric Rh(III) complex.The metal catalysed [2+2+2]-cycloaddition reactions of alkynes is one of the most direct methods to generate polysubstituted benzenes, pyridines and annelated benzene derivatives.Early studies involved the use of a stoichiometric amount of the cobalt(0) complex, C p Co(CO) 2 , to promote the cyclotrimerisation. 22A wide variety of functional groups such as alkyl, aryl, vinyl, CH 2 OH, CO 2 R, NR 2 , SR, SiR 3 , B(OR) 2 , Br and I are tolerated in the process. 23,24We reported that a catalytic amount of Wilkinson's catalyst, RhCl(PPh 3 ) 3 effects [2+2+2]cycloadditions analogous to those promoted by stoichiometric amounts of the C p Co(CO) 2 complex. 25Since then various other transition metal catalysts based on Ni, Co, Pd, Cr, Rh, Fe, Zr, Nb, and Ir have been developed for alkyne trimerisation reactions. 26Cyclometallated complexes of Rh(III), Ir(III) and Ru(II) have not, to our knowledge, been reported as precatalysts for the [2+2+2]-cycloaddition of alkynes.This encouraged us to evaluate complexes 2-6 for such reactions.
Initally, the 1,6-diynes 7a-d [25][26][27] , were reacted individually with propargyl alcohol (5 mol eq) in t BuOH at appropriate temperatures in the presence of the first generation pre catalysts 2a (2 mol%), 2b (2 mol%), and 4 (4 mol%) affording the corresponding benzene derivatives 8a-d (Scheme 2, Table 1).Excess of propargyl alcohol was employed to effect reduction of the precatalysts to the active catalysts.C) precatalysts, function at lower temperatures.This temperature variation reflects the relative ease, and rate, of reduction of the precatalysts to their respective lower valent active catalysts.Another interesting feature was that no product was observed (nmr) for 1-2 h indicating an induction period.Although the reaction showed high chemoselectivity for cycloadduct 8, the dimer 9, arising from the corresponding diyne, was also observed in some cases (Table 1).The diynes 7b and 7d afforded the corresponding cycloadducts exclusively in the presence of precatalysts 2a and 2b, while, 6-13% dimerisation was observed with precatalyst 4. Diyne 7d reacts much more slowly than diyne 7b.Among the precatalysts evaluated for the reaction, the rate of the reaction decreased in the order Ru > Rh > Ir whilst chemoselectivity was, in general, in the reverse order.
Next the second generation ortho-metallated Rh(III) complexes 5a, and 6 were evaluated in the [2+2+2]-cycloaddition of 7b with propargyl alcohol and 2-butyn-1,4-diol (Table 3).Results from the evaluation of precatalyst 2a are also included in Table 3 for comparison.When the sterically more hindered Rh(III) complex 6 was utilised in the [2+2+2]-cycloaddition of the 1,6diyne 7b with propargyl alcohol and 2-butyn-1,4-diol, a decrease in the reaction temperature required to effect 100% conversion and an increase in the rate of the reaction were observed compared to the reactions effected by 2a ( Table 3, entries 4 and 7).Note that since 6 is a monomeric complex 4% of this precatalyst was added.A possible explanation for the increased activity of the precatalyst 6 could be more facile conversion of the precatalyst 6 to the active Rh(I) species (discussed in the following mechanism section).Rh(III) complex 5a was found to be the best precatalyst among the cyclometallated complexes studied for the [2+2+2]cycloaddition of 1,6-diynes with propargyl alcohol and 2-butyn-1,4-diol (Table 3, entries 2 and 6).No diyne dimerisation product was observed in these processes except for entry 5.
Initial coordination of the rhodium(I) catalyst to the 1,6-diyne is followed by oxidative addition-cyclisation to produce the rhodacyclopentadiene complex 11.Coordination of the monoyne then gives complex 12 which undergoes either monoyne insertion to give 13 or a Diels-Alder type reaction to give 15.Finally, reductive elimination 13 → 14 / 15 → 16 gives the products and regenerates the Rh(I) catalytic species.In recent years the number of characterised metallocyclopentadienes 28,29 and metallocycloheptatrienes 30 has continued to expand.The unresolved problem of whether the Diels-Alder like pathway 13 → 16 (Scheme 5) or the insertion pathway 12 → 13, or both, play a role in the [2+2+2]-cyclotrimerisation has seen some clarification.have provided data and mechanistic arguments that both pathways may be involved, at least when R 3 P ligands are present, depending on steric effects and ligand dissociation rates.Their interpretation of factors favouring the Diels-Alder like sequence involves the incoming alkyne accessing a vacant apical Ir coordination site 18 with its regiochemical approach trajectory minimising steric interaction between the R and R 1 groups in the reacting partners leading to the meta-substituted product 16 (Scheme 4).Accessing the metallocycloheptatriene pathway is believed to arise from steric hindrance in the pentacoordinate complex 20 and / or to hemilability of L favouring the equilibration 20 21 in which steric interaction between R / R 1 is minimised.The regiochemistry of the insertion step 21 → 22 (Scheme 5) which leads to the ortho-substituted product 14, would then suggest that Ir-C bond formation is in advance of C-C bond formation and that steric effects at Ir are the key determinant.The ferrocenyl phosphine DPPF 19 was found to be the best, of those ligands surveyed, for ortho-selectivity

Sheme 5
The cyclometallated complexes utilised in our studies differ from those normally employed in that they are M(III) (M = Ir, Rh) complexes and lack phosphine ligands.The former require an initial reductive step, to generate the catalytically active M(I) species from the respective M(III) complexes.The Ru(II) complex 4 is assumed to undergo facile reduction to ruthenium(0) nanoparticles 32 and this accounts for the lower temperature and shorter reaction times in this case.An induction period 1-2 h was observed in the [2+2+2]-cycloaddition reactions when the first generation catalysts were used which points to a slow generation of the active catalysts.A plausible pathway for the generation of M(I) species, is shown in Scheme 6.A possible explanation for the increased activity of the second generation Rh precatalysts could be the rate of conversation of the Rh(III) complex to the active Rh(I) species.In the case of precatalyst 6 this conversion may be acclerated by relief of steric strain, occasioned by the bulkiness of the ligand.Thus the active catalysts 25 in our case differ substantially from those employed previously.They should be accessible by addition of appropriate alcohol primers and their regioselectivity remains to be explored.In conclusion, we report the synthesis of a variety of novel cyclometallated Ir(III), Rh(III) and Ru(II) complexes and their catalytic activity in the [2+2+2]-cycloaddition reaction between 1,6-diynes and alkynes.The Ir(III) precatalysts complex 2b proved highly chemoselective.stated.Compounds 8a-d and 2a,b have been previously reported. 20,25-27X-Ray data for 4 (CCDC 624678 and 6 (CCDC 628187) has been deposited in the Cambridge Crystallography Database.

General procedure for cyclometallated complexes of rhodium(III) and iridium(III)
Either rhodium(III) chloride hydrate or iridium(III) chloride hydrate (2.5 mmol) was added to the appropriate ligand (6 mmol) in 2-methoxyethanol (25 m1) and the mixture boiled under reflux for the appropriate time.The resulting suspension was filtered to afford the product, which was further purified by crystallisation if necessary.
General procedure for cyclisation of 1,6-diynes with monoynes 1,6-Diyne, monoyne and the catalyst were mixed in tert-butanol in a Schlenk tube and stirred and heated at the appropriate temperature for the appropriate time.The solvent was then removed under reduced pressure and the residue was purified by flash column chromatography to afford the product.

R
Very recently Takeuchi et al.

Table 2 .
The a Obtained from 1 H-NMR; Values given in brackets are the isolated yield of the major cycloadduct 11 (based on conversion)