Exploring Bis-Cyclometalated Ru(II) Complexes as Active Catalyst Precursors: Room-Temperature Alkene-Alkyne Coupling for 1,3-Diene Synthesis

Late transition metal-mediated C–H bond activation has become a popular method to generate metal-carbon σ bonds in metallacycle synthesis. These cyclometalation reactions are usually facilitated by a heteroatom (X)-based functional group nearby the target C–H bond.^[1] Such directed C–H bond activation strategy has been widely used to generate cyclometalated late transition catalysts and catalyst precursors in various homogeneous catalytic processes.^[2] Among reported metallacycles as catalyst precursors, a dominant majority feature one 5- or 6-member chelating ring with a η²-[C,X] ligand as both anionic carbon-donor and neutral heteroatom-donor. By contrast, bis-cyclometalated late transition metal complexes with two η²-[C,X] ligands ^[3] are rarely exploited as organometallic catalysts.^[4-6] We herein report the synthesis, structural characterization, and catalytic applications of several bis-cyclometalated ruthenium(II) complexes with η²-[C,N] and η²-[C,O] ligands. The potential of these ruthenacycles as catalyst precursors is demonstrated by a room-temperature catalytic alkene-alkyne coupling to synthesize α,β,γ,δ-unsaturated esters and amides.

We have recently reported a bis-cyclometalated octahedral Ru(II) complex, {Ru(η⁴-cod)[η²-HN=C(C₆H₅)C₆H₄]₂} (1a), as a catalyst precursor in Ru(II)-catalyzed [3+2] carbocyclization between N–H ketimines and alkynes using a N-heterocyclic carbene ligand IPr (Scheme 1).^[7] The proposed mechanism involves carbon-carbon bond formation by alkyne insertion into the Ru–C σ bond of a ruthenacycle intermediate (A→B), presumably facilitated by the IPr ligand that has replaced cod ligand on Ru center. Thus, the substrate-derived η²-[C,N] imine ligands appear to play the dual role of actor ligand and spectator ligand, eventually incorporated into the cyclization product and replaced by incoming ketimine substrates via cyclometalation. It is noteworthy that 1a did not react with alkyne substrates without added IPr ligand, suggesting significant ligand effect on its stability and reactivity. We envision that bis-cyclometalated 1a and its structural analogs can be further explored as Ru(II) catalyst precursors with η²-[C,X] ligands solely as spectator ligands that occupy four of the six coordination sites and affect reactions occurring at the other two cis-coordination sites (Scheme 2, C). In particular, ancillary ligands (L) can be replaced by alkene/alkyne substrates via π-complexation (C→D), setting the stage for C–C bond formation by oxidative cyclization (D→E).^[8] The latter transformation is a key step in a number of Ru-catalyzed C–C coupling reactions such as alkene-alkyne (enyne) couplings for diene synthesis,^[9] [2+2] or [2+2+2] cycloadditions, ^[10] and [2+2+1] cycloadditions such as the Pauson-Khand reaction.^[11] With easily accessible η²-[C,X] ligands via C–H activation, bis-cyclometalated Ru(II) complexes may serve as an attractive alternative to existing Ru catalysts,^[12] allowing modular catalyst design and tunable ligand effects on catalyst efficiency and selectivity.

Scheme 1.

Proposed C-C bond formation by alkyne insertion into Ru-C bond of ruthenacycle intermediates for Ru(II)-catalyzed [3+2] imine/alkyne carbocyclization.^[7]

Scheme 2.

Envisioned C-C bond formation by oxidative cyclization with bis-cyclometalated Ru(II) complexes having two η²-[C,X] ligands

As reported previously,^[7] complex 1a was synthesized by room-temperature cyclometalation of benzophenone imine with a commercially available Ru(II) π-allyl complex [(cod)Ru(η³-methallyl)₂] in a 2:1 ratio (Scheme 3). A bis-cyclometalated ketone analog {Ru(η⁴-cod)[η²-OC(C₆H₅)C₆H₄]₂} (1b) was synthesized using benzophenone in a similar fashion, albeit with lower reactivity and requiring heating at 80 °C for complete conversion. The chelating cod ligand in both 1a and 1b could be replaced by two pyridine ligands via heating in neat pyridine at 60 °C to form bis(pyridine)-ligated 2a and 2b respectively. The solid-state structures of complexes 1a,^[7] 1b, 2a and 2b were determined by single crystal X-ray diffraction (see Supporting Information for details). All four of these bis-cyclometalated Ru(II) complexes displayed a near-octahedral Ru(II) center with two cis η²-[C,X] ligands, where the two N or O atoms were trans to each other and the two Ru–C σ bonds were cis to each other.^[7][13]

Scheme 3.

Synthesis of bis-cyclometalated Ru(II) complexes with η²-[C,N] benzophenone imine and η²-[C,O] benzophenone ligands.

The catalytic activity of bis-cyclometalated Ru(II) complexes was evaluated by intermolecular alkene-alkyne coupling between diphenylacetylene (3a) and methyl acrylate (4a) to form a (2E,4Z)-1,3-diene product 5a (Table 1).^[9][14-16] Using 5 mol% [(cod)Ru(η³-methallyl)₂] (6) as catalyst precursor and no added ligands led to only 12% conversion after heating at 80 °C for 24 hours in toluene (entry 1). By contrast, in situ generated ruthenacycle 1a by pre-activation of 6 with 2 equivalents of benzophenone imine effectively promoted formation of 5a in quantitative yield (entry 2). Compared to benzophenone imine ligand, much lower reactivity was observed when catalyst pre-activation was carried out using other aromatic compounds that are capable of generating bis-cyclometalated Ru(II) complexes.^[13] For example, using 2-phenylpyridine and benzophenone ligands both result in lower than 10% conversion (entries 3,4; see Supporting Information for complete results).

Table 1.

Development of the catalytic reactions.^[a]

Entry	Ru Catalyst	Ligand	Solvent	Method[b]	Temp.	Yield (%)[c]
1	[Ru(cod)(C4H7)2] (6)	none	toluene	A	80 °C	12
2	6	Ph2C=NH	toluene	B	80 °C	>98
3	6	2-phenylpyridine	toluene	B	80 °C	8
4	6	Ph2C=O	toluene	B	80 °C	<5
5	6	Ph2C=NH	toluene	B	r.t.	>98
6	6	Ph2C=NH	toluene	A	r.t.	15
7	1a	none	toluene	A	r.t.	>98
8	1a	none	THF	A	r.t.	96
9	1a	none	DME	A	r.t.	66
10	1a	none	DMF	A	r.t.	93
11	1a	none	hexane	A	r.t.	82
12	1b	none	toluene	A	r.t.	<2
13	2a	none	toluene	A	r.t.	>98
14	2b	none	toluene	A	r.t.	<2

^aConditions: 3a (0.20 mmol, 1 equiv), 4a (2.0 equiv), Ru catalyst (0.05 equiv), ligand (0.10 equiv), solvent (0.5 mL), room temperature (20-22 °C) or 80 °C, 24 h.

^bMixing methods: (A) All components were mixed without pre-activation; (B) Ru precursor 6 and the ligand were added to toluene and stirred at 80 °C for 30 min; the mixture was then cooled down to room temperature and added with substrates 3a and 4a.

^cGC yields with n-dodecane as internal standard.

Gratifyingly, the in situ formed 1a was sufficiently active at room temperature, promoting quantitative formation of 5a after 24 hours with 5 mol% catalyst loading (entry 5). To our best knowledge, this is the first example of room-temperature catalytic acrylate-alkyne coupling to form α,β,γ,δ-unsaturated esters and complements other catalyst systems for mild enyne couplings.^[15][16] Notably, skipping the 80 °C pre-activation led to much lower catalyst reactivity (entry 6). Thus, isolated 1a was used as catalyst precursor to further evaluate the solvent effect, with highest reactivity observed in toluene and THF solvents (entries 7-11). Under optimized conditions, room-temperature coupling between 3a (1.0 equiv) and 4a (2.0 equiv)^[17] proceeded smoothly in toluene solvent with 5.0 mol% 1a, giving 5a in quantitative yield by GC analysis (entry 7). The pyridine-ligated bis(imine) complex 2a was less stable than 1a in solution phase but displayed comparable catalytic activity (entry 13). By contrast, the bis(ketone) analogs 1b and 2b were virtually unreactive as catalyst precursors (entries 12 and 14). Such reactivity distinction is consistent with the mechanistic hypothesis that cod or pyridine ligands will be replaced by alkene/alkyne substrates (Scheme 2, C→D), thus having little effects on catalytic activity beyond the initial stage of catalyst pre-activation. On the other hand, the η²-[C,X] imine or ketone ligands are expected to stay on the Ru center throughout catalytic cycles and play a dominant role on catalyst activity.

With the standard reaction conditions established, various internal alkynes and acrylic esters or amides were studied for Ru-catalyzed room-temperature alkene-alkyne coupling (Table 2). Coupling between 3a and unsubstituted alkyl acrylates proceeded smoothly to form 1,3-diene products 5a-d and 5f in over 90% yields and with exclusive stereoselectivity for the (2E,4Z)-isomers. For phenyl acrylate coupling product 5e, the yield was improved from 58% to 87% by replacing 1a with bis(pyridine) ruthenacycle 2a as the catalyst precursor. Such reactivity enhancement is likely due to facile catalyst activation by substrate replacement of more labile pyridine ligands compared to the chelating diene ligand. When coupling between 3a and 4a was scaled up from 0.2 mmol to 20 mmol, the loading of 1a could be reduced to 1.0 mol% to acquire 5a in 90% isolated yield (4.8 gram purified product) over 48 hours. Coupling between 3a and N,N-dimethyl acrylamide gave product 5g in 72% yield, while heating was needed to improve the yield of N,N-diethyl product 5h to 85% at 60 °C. Compared to less reactive N,N-dialkylacrylamides, N-isopropyl- and N-t-butylacrylamide reacted with 3a in good reactivity to form products 5i and 5j, although the latter required 2a as catalyst precursor for satisfactory yield. The scope of alkyne substrates was studied by coupling reactions with methyl acrylate (4a) to give products 5k-v. High reactivity and regioselectivity was observed for phenylacetylene derivatives with alkyl substituents (products 5k-s), favoring the formation of 4-alkyl-5-aryl regioisomer in >10:1 selectivity. The mild reaction conditions allow good compatibility with functional groups such as acyl, formyl, and Br substituents (products 5p, 5r and 5s), providing synthetic handles for further functional group transformations. Aliphatic internal alkynes such as 3-hexyne and 4-octyne displayed lower reactivity than aromatic alkynes, and a 2:1 alkyne/acrylate stoichiometry was used to get coupling products 5t and 5u in moderate yields. Coupling between 4a and terminal alkynes generally suffered from low reactivity and gave a complex mixture of products.^[18] Nevertheless, coupling between 4a and phenylacetylene was effectively catalyzed by 2a to form product 5v with (E,E)-stereoselectivity in 65% isolated yield.^[14b]

Table 2.

Scope of the catalytic alkene-alkyne coupling.^[a]

^aConditions: 3 (0.20 mmol, 1.0 equiv), 4 (2.0 equiv), 1a (0.050 equiv), toluene (0.5 mL), 20-22 °C, 24h; averaged isolated yields of two runs.

^bIsolated yield under scale-up conditions: 3 (20 mmol, 1.0 equiv), 4 (2.0 equiv), 1a (0.010 equiv), toluene (6.0 mL), 22 °C, 48h.

^cUsing 2a as catalyst precursor.

^dReactions at 60 °C

^eCombined yield of two regioisomers (ratios determined by NMR analysis); structure of major isomer is shown.

^fUsing 0.20 mmol methyl acrylate as limiting reagent and 2.0 equiv of alkyne.

^gIsolated yield for the major stereoisomer in 5:1 selectivity; minor isomer was not purified.

Three types of reaction mechanism have been proposed for Ru-catalyzed alkene-alkyne couplings to form 1,3-dienes:^[14][19] (1) C–C bond formation by alkene-alkyne oxidative cyclization (Scheme 2), followed by β-H elimination and C–H reductive elimination (Path 1); (2) alkyne insertion into a Ru hydride, followed by alkene insertion into the resulting Ru-alkenyl linkage and subsequent β-H elimination (Path 2); (3) sp² C–H bond activation of alkene, followed by alkyne insertion into Ru alkenyl, and C–H bond formation by either reductive elimination or protonation of Ru–C bond (Path 3). Although the latter two pathways cannot be completely ruled out, the oxidative cyclization mechanism (Path 1) is most consistent with the observed regio- and stereochemistry in coupling products. In particular, high regioselectivity with non-symmetric alkyne substrates (products 5k-s) supports C–C bond formation by oxidative cyclization (Path 1) or alkyne insertion into Ru-alkenyl linkage (Path 3), not by alkyne insertion into Ru–H linkage (Path 2).^[19] The complete lack of (2Z)-stereoisomers as coupling products also argues against the proposed alkene C–H activation stereochemistry in Path 3, which should favor (2Z)-isomers by ester- or amide-directed C–H activation/cyclometalation.

The proposed oxidative cyclization pathway has prompted us to extend our study to other mechanistically related C–C couplings using current catalyst system. Thus, 1a was found to catalyze room-temperature dimerization of methyl acrylate with high efficiency and exclusive tail-to-tail regioselectivity (Eq. 1).^[17] In addition, a [2+2] norbornene/alkyne cycloaddition was effectively catalyzed by 1a at 120 °C (Eq. 2), which further supports the proposed Ru(II)/Ru(IV) catalytic cycle involving oxidative cyclization.^[19][20]

In summary, we have developed a new class of bis-cyclometalated Ru(II) catalyst precursors with readily available η²-[C,X] ligands based on aromatic N–H ketimines and ketones. The catalytic activity of bis(imine) complex 1a was demonstrated with several catalytic C– C coupling reactions by proposed Ru(II)/Ru(IV) catalytic cycles involving oxidative cyclization. A room-temperature alkene-alkyne coupling was promoted to form α,β,γ,δ-unsaturated esters and amides with high regio- and stereoselectivities, good functional group tolerance, and very high catalyst efficiency in a representative gram-scale synthesis. The major limitation of current catalyst system is the limited scope of alkene substrates,^[21][22] which we aim to improve by a more systematic future study on structure-reactivity correlations in bis-cyclometalated Ru(II) complexes with various η²-[C,X] ligands.

(1)

(2)