Previous papers in this series (1, 2), pointed out that ruthenium indenylidene and vinylidene complexes, 1 and 2, where R is a phenyl (Ph) or a cyclohexyl (Cy) substituent, constitute a convenient alternative to the first generation Grubbs ruthenium alkylidene metathesis pre-catalysts, 3 and 4 (3, 4), used on a large scale in organic (5–9) and polymer syntheses (10–14).

View this figure

View this figure

Variations in the ligand sphere of the ruthenium atom have been made. These involve mainly the association of certain ancillary ligands, such as N-heterocyclic carbenes (imidazolin-2-ylidene and its saturated derivative), heterocyclic N-donating ligands (pyridine or different N-heterocycles), O- and O,N-chelated ligands (such as phenyl isopropoxy, Schiff-bases), 5 to 12, (R = Ph or Cy, R′ = Ph or Me) (15–31). However, the synthesis of novel metathesis initiators of improved accessibility and with a better application profile is still a challenge to organometallic and metathesis researchers.

View this figure

The present work is devoted to a vast and promising class of ruthenium metathesis pre-catalysts, derived essentially from neutral and cationic ruthenium allenylidene complexes, 13 and 14, respectively, (R = Ph or Cy, X = PF₆, BF₄, BPh₄, OTf), that gained wide applicability as metathesis initiators due to their easy accessibility and good to excellent catalytic properties.

View this figure

Types of Ruthenium Allenylidene Complexes

The family of neutral and cationic ruthenium allenylidene complexes is very large (32–37) and a considerable number of its members are now recognised as active metathesis catalysts (5, 6, 33).

Neutral Ruthenium Allenylidene Complexes

Three neutral, coordinatively unsaturated 16-electron ruthenium allenylidene complexes, of different types, namely the bisphosphane complex 15, the imidazolin-2-ylidene complex 16 and the bimetallic complex 17, have been prepared and evaluated for their catalytic efficacy in alkene metathesis reactions (38, 39).

View this figure

View this figure

View this figure

The bisphosphane complex, 15, is the allenylidene analogue of Grubbs catalyst 3, but with PCy₃ ligands (4). A complex similar to 15 but having PPh₃ groups instead of PCy₃, seems to be rather unstable under normal conditions and rearranges readily to indenylidene, as will be shown in a later section. More stable but less active, the imidazolin-2-ylidene complex, 16, which is the allenylidene analogue of complex 5, stems from complex 15 by simple ligand substitution. The binuclear allenylidene complex, 17, is a highly active metathesis ruthenium complex, and is related to the binuclear benzylidene complex [Ru₂Cl₄(=CHPh)(p-cymene)(PCy₃)] reported earlier by Grubbs and coworkers (40).

Cationic Ruthenium Allenylidene Complexes

Up to now, a vast library of cationic, coordinatively saturated 18-electron ruthenium allenylidene complexes, has been reported (41–48) and many of these have been successfully applied in various alkene metathesis (49) and ring-opening metathesis polymerisation (ROMP) reactions (50). Essentially, the main representatives of this class of allenylidene complexes, for example, complexes 18–21, contain η⁶-arene ligands associated with phosphane and chloride, in conjunction with a ‘non-coordinating’ counterion X⁻.

View this figure

By varying the substituents at the phosphine (R = Ph, Cy, i-Pr) and allenylidene moieties (R′ = Ph, p-chlorophenyl, p-methoxyphenyl, etc.), and the nature of the counterion X⁻ (X = PF₆, BPh₄, BF₄, OTf, etc.) a great number of this type of ruthenium complex could be conveniently prepared. Due to easy accessibility, their potential as metathesis pre-catalysts has been extensively evaluated. Some of them allowed metathesis reactions to be performed in ionic liquids of the type 1-butyl-3- methylimidazolium salts, as their cationic nature ensured a high solubility of the catalytic species in the reaction medium (51).

Incorporation of imidazolin-2-ylidene as the ancillary ligand in the arene ruthenium allenylidene complexes, such as 22, provided very active initiators in the ROMP of cycloolefins (52).

View this figure

Moreover, binding the arene ligand with the dihydroimidazolin-2-ylidene unit through a methylene tether, in a half-sandwich mode, led to the η¹:η⁶-arene-carbene ruthenium allenylidene complex, 23 which was highly active in ring-closing metathesis (RCM) of 1,6-dienes (53). Complex 23 allowed the metathesis of dienes to be selectively directed towards either RCM or cycloisomerisation products.

View this figure

A structurally different cationic ruthenium allenylidene complex 24, with two phosphines containing hemilabile methoxy groups, has also been reported (54). This bidentate ruthenium complex showed a lower activity in RCM of dienes and ROMP of cycloolefins as compared with the arene ruthenium complex of type 18, but the synthesis and structure of 24 are challenging issues.

View this figure

Numerous advantages are offered in many applications by the heterogenised version 25 of the arene ruthenium complex 18, obtained through immobilisation on polystyrene (55).

View this figure

This supported ruthenium pre-catalyst has been reused several times in RCM of dienes, but this was possible only after reactivation in a separate process.

Miscellaneous Ruthenium Allenylidene Complexes

The exchange of the triphenylphosphines in [RuCl₂(PPh₃)₃] with water-soluble phosphines, followed by reaction with 3,3-diphenylpropyn-3-ol (in methanol (MeOH)) gave the dimeric ruthenium allenylidene complex 26 which was able to initiate, under certain conditions (protic solvents), ROMP of cycloolefins (56). For instance, in ROMP of cyclopentene, complex 26 led to considerable polymer yields when, in order to activate the catalytic system, Brönsted acids were used as co-initiators.

View this figure

A large array of neutral and ionic ruthenium allenylidene complexes, bearing quite diverse ligands and substituents, has been prepared and their catalytic properties in various chemical transformations investigated (57–62). Some of these ligands or substituents might induce asymmetry in the metal complex, generating chiral ruthenium allenylidene catalysts of significance for enantio-selective metathesis catalysis.

Synthesis of Ruthenium Allenylidene Complexes

Neutral Ruthenium Allenylidene Complexes

The neutral, 16-electron ruthenium allenylidene complex 15 was synthesised almost simultaneously by Fürstner and Hill (38), and Nolan and coworkers (39) by reacting [RuCl₂(PPh₃)₃] or [RuCl₂(PPh₃)₄] with commercially-available 3,3-diphenylpropyn-3-ol, followed by PCy₃, see Equation (i).

View this figure

A similarly straightforward route for synthesis of the PPh₃ analogue (15a) has been reported by Hill and coworkers (63) but subsequent detailed investigation of its structure has proved that the compound is really the ruthenium 3-phenyl-1-indenylidene complex 11, supposedly formed by intramolecular rearrangement of an allenylidene structure, 15a, previously formulated for this complex (39b), see Equation (ii).

View this figure

Synthesis of complex 15 can also conveniently start from the commercially available ruthenium dimer 27, which reacts with 3,3-diphenylpropyn-3-ol and 2 equivalents of tricyclohexylphosphine. Further substitution of a phosphane group with 1,3-dimesitylimidazolin-2-ylidene readily affords the ruthenium imidazolin-2-ylidene allenylidene complex 16 (39) (Scheme I).

View this figure

Single-crystal X-ray studies on these two ruthenium allenylidene complexes, 15 and 16, provided significant information about bonding in the solid state and electronic environment at the metal centre (39). Unfortunately, the experimental catalytic activity of complexes 15 and 16 in some ring-closing metathesis of dienes was low, although both complexes are rather stable under the reaction conditions. However, a valuable representative of this group, the highly active bimetallic ruthenium allenylidene complex 17, has been prepared in appreciable yield from the same ruthenium dimer 27 and the ruthenium complex 15 (38), see Equation (iii).

View this figure

This latter complex, 17, has been successfully employed in an array of RCM and enyne metathesis reactions.

Cationic Ruthenium Allenylidene Complexes

It is noteworthy that a wide range of cationic 18-electron ruthenium allenylidene complexes, prepared by several research teams using the commercial 3,3-diphenylpropyn-3-ol as the allenylidene moiety source, came as a valuable bonus to state-of-the-art metathesis catalysis (40–47). Thus, a convenient method for synthesis of the ruthenium complex 18a (R = Ph, Cy or i-Pr) is based on the commercially available ruthenium dimer 27, which is first converted to the corresponding monophosphane complex 28 (R = Ph, Cy or i-Pr). The latter, 28, is then reacted with 3,3-diphenylpropyn-3-ol, in the presence of NaPF₆ (or NaBPh₄, etc.) in MeOH at room temperature to give high yields (92–97%) of allenylidene product (49), see Equation (iv).

View this figure

In the final complex, the sterically encumbered phosphane ligand R₃P prevents further attack of MeOH at the electrophilic α-C atom of the allenylidene group, which would lead to the formation of the Fischer-type carbene complex [(p-cymene)(R₃P)ClRu=CH(OMe)CH=CPh₂] (64). Several other ruthenium allenylidene complexes, such as 18b (R = Ph, p-chlorophenyl, p-methoxyphenyl, CR₂ = fluorenylidene, etc.) could be obtained by a similar method, in aprotic solvents. Thus, on treatment with AgX (X = PF₆, BF₄, OTf, etc.) in CH₂Cl₂, the monophosphane complex 28 forms the rather stable cationic 16-electron complex, 29. The latter compound reacts rapidly with the suitably substituted propyn-3-ol (in CH₂Cl₂ at room temperature) to give the corresponding ruthenium allenylidene complex 18b in appreciable yields (80–95%) (49, 65), see Equation (v).

View this figure

It is important to emphasise the ready transformation of 18-electron ruthenium allenylidene complexes into their ruthenium indenylidene counterparts, via the alkenylcarbyne Ru species, in the presence of strong acids, such as triflic acid (TfOH). In this respect, Dixneuf and coworkers (66) clearly evidenced, by ¹³C{¹H} and ¹H NMR, the generation of the corresponding alkenylcarbyne derivative from complex 18b (R = Ph, X = OTf), at –40°C. However, at –20°C the latter intermediate (an alkenyl carbyne derivative) was transformed into the 18-electron arene ruthenium indenylidene complex, 30 (Scheme II).

View this figure

Remarkably, this ionic, 18-electron arene ruthenium indenylidene complex, arising from the easily-made allenylidene precursor, showed exceedingly high catalytic activity in various metathesis reactions, such as RCM of N-containing dienes (e.g., of N,N-diallyltosyl amide), enyne metathesis (e.g., of N,N-allylpropargyltosyl amide), ADMET (e.g., of 1,8-decadiene) and ROMP of low-strain cycloolefins (e.g., cyclopentene and cyclooctene), even operating at 0°C.

A further variation on these ruthenium complexes consists of the replacement of the p-cymene ligand with other arenes, capable of binding coordinatively at the Ru centre, prior to introduction of the allenylidene ligand in the final step. Thus, cleavage of the commercial ruthenium dimer 27 with an aralkyl phosphine, in CH₂Cl₂ at room temperature, followed by an intramolecular substitution of the p-cymene ligand with the tethered phenyl ring (in chlorobenzene at 140°C) yields the ruthenium complex 31. This intermediate is finally converted, in substantial yields (83–91%), into the new chelated ruthenium allenylidene complex 32, via the above discussed protic or aprotic approach (49) (Scheme III).

View this figure

The single-crystal X-ray studies of complexes 18a (R = i-Pr), 31 and 32 gave, for the first time, interesting data about the structure of these types of complexes and convincingly unravelled the nature of the coordination bonds of the ligands around the ruthenium centre (49).

Catalytic Properties of Ruthenium Allenylidene Complexes

Except for a few cases, both families of ruthenium allenylidene complexes, that is: the neutral, coordinatively unsaturated 16-electron and the cationic, coordinatively saturated 18-electron complexes, evidenced good to excellent catalytic activity and selectivity in various metathesis reactions (RCM, ADMET, enyne metathesis, ROMP). By virtue of these catalytic properties and taking into account their easy accessibility from commercial reagents, the class of ruthenium allenylidene complexes holds promise as challenging candidates for productive applications in organic and polymer syntheses. For instance, the neutral, 16-electron ruthenium complex 17 proved to be a highly effective pre-catalyst for RCM of α,ω-dienes and dienynes, at ambient temperature (38). This complex is largely tolerant of functional groups including amide, sulfonamide, bromide, ester, ether, siloxane and fluorenylmethoxycarbonyl substituents. It allows facile and efficient formation of variously functionalised mono- and bicyclic ring systems. Among these compounds, we should mention the structurally complex precursors to exaltolide, a musk odorant, and epilachnene, an insect repellent alkaloid isolated from the pupae of a Mexican beetle.

Syntheses of cyclic compounds via RCM and enyne metathesis using the above binuclear ruthenium allenylidene pre-catalyst are illustrated in Equations (vi) and (vii).

View this figure

View this figure

The catalytic performances of the cationic, 18-electron ruthenium complexes of type 14 in RCM of a variety of dienes and in enyne metathesis have been extensively evaluated by Dixneuf and Fürstner (49). The most important finding was that the metathesis activity of these complexes depends essentially on the nature of the phosphane ligand, the remote substituent of the allenylidene moiety and the nature of the “non-coordinating” counterion. For instance, detailed studies on the catalytic activity of cationic complex 14, X = PF₆, in RCM of N,N-diallyltosyl amide, revealed that when varying the phosphane ligands the catalytic activity decreased in the following order: PCy₃ > Pi-Pr₃ >> PPh₃, in accordance with previous observations for the related ruthenium alkylidene pre-catalysts (67). With 14-PF₆ (R = PCy₃) (2.5 mol%) N,N-diallyltosyl amide was quantitatively converted to N-tosyl dihydropyrrole, after 4 h reaction time, in toluene, at 80°C (isolated yield 83%), see Equation (viii).

View this figure

The above ruthenium pre-catalyst has been applied in RCM of several representative diene substrates allowing synthesis of essentially all ring sizes greater than four, including mono and bicyclic compounds, in good to excellent yields. As expected, the formation of medium sized rings required particularly long reaction times (up to 100 h) and high dilution conditions, while decomposition of the catalytic species seemed to occur with a rate similar to that of the productive RCM. Despite these inherent inconveniences, a large number of heterocyclic compounds pertaining to different heterocyclic systems could also be obtained in good yield (77–88%) by this procedure, see Equations (ix) and (x)

View this figure

View this figure

Of great synthetic utility, the complex 14-PF₆ (R = PCy₃) allows efficient synthesis of a set of uncommon macrocyclic compounds; for instance, smooth cyclisations of dienes A and B to the 16- and 18-membered cycloalkenes I and II, respectively, have been successfully achieved by this procedure, see Equations (xi) and (xii).

View this figure

View this figure

As just mentioned, compound I is a precursor of the valuable macrocyclic musk, Exaltolide (a perfume ingredient). Furthermore, disaccharide III, obtained by cyclisation of the polyfunctional diene C, see Equation (xiii), constitutes an advanced intermediate for synthesis of tricolorin A, a carcinostatic resin glycoside.

View this figure

A characteristic feature of these ionic catalytic systems is the fact that the counterion exerts a particular influence on the reactivity of the ruthenium allenylidene complex as well as on the selectivity to metathesis products (48). Thus, whereas cation 14, associated with PF₆, BPh₄ or OTf, leads to excellent yields at 80°C, in toluene, its combination with OTf was found to be effective – even at room temperature. This behaviour was tentatively ascribed to the weakly coordinating propensity of the triflate group which may assist the decomplexation process of the p-cymene ligand from the 18-electron allenylidene pre-catalyst and may also potentially stabilise, in solution, the resulting 14-electron Ru(II) species (49), see Equation (xiv).

View this figure

Surprisingly, the allenylidene complex 14 having BPh₄⁻ or BF₄⁻ as counterions is less active and selective, giving rise to numerous side reactions. For instance, with BF₄⁻ as the counterion, the ruthenium complex 14 induced cycloisomerisation of N,N-diallyltosyl amide to methylenecyclopentane (42%), with concomitant cyclisation to N-tosyl dihydropyrrole (31%) and isomerisation to a new diene (18%) (48), see Equation (xv).

View this figure

Relevant studies on the effect that some additives have on the catalytic activity of the allenylidene complex 14 during N,N-diallyltosyl amide cyclisation pointed out that addition of Lewis or Brönsted acids (e.g., BF₃/Et₂O, HBF₄/Et₂O, F₃CSO₃H) increased selectively the RCM activity while almost completely suppressing the above cycloisomerisation. By contrast, ammonium salts (e.g., n-Bu₄NF) dramatically decreased the overall activity of the complex (49). These interesting observations constitute valuable information for assessing the reaction mechanisms promoted by cationic ruthenium allenylidene complexes.

It is worth noting that allenylidene ligands included in complexes 18 to 20 exert a pronounced influence on the activity of these catalytic systems, and essentially on the selectivity of the reaction products (49). For instance, in the reaction of N,N-diallyltosyl amide discussed above it was possible to alter the catalytic activity of the complex or to switch the reaction pathway from RCM to predominantly cycloisomerisation, simply by varying the distal para-substituents in the phenyl rings of the allenylidene units. This unprecedented result seems to indicate that the allenylidene moiety or a species derived thereof serves as a permanent ligand to the Ru-template throughout the entire catalytic cycle. This behaviour differs fundamentally from that of the parent Grubbs alkylidene complexes 3 and 4, in which the Ru=CHR fragment of the catalyst precursor intervenes only in the initiation process when CHR is cleaved off by reaction with the alkene substrate generating the Ru=CH₂ unit, further responsible for the propagation (68).

Very importantly, ruthenium complex 14 can be photochemically activated efficiently by irradiation with UV light (300 nm), as was found earlier for the immobilised phosphane complex 25 (R = Cy) (69). Thus, irradiation of a solution of this catalyst and N,N-diallyltosyl amide allows RCM of the unsaturated substrate to proceed in appreciable yield (81%), even at ambient temperature (49), see Equation (xvi).

View this figure

Such a photochemical activation technique has been successfully applied to a number of enyne substrates which could be conveniently converted into the substituted 3-vinyl-2,5-dihydrofuran derivatives, under these relatively mild conditions (70).

Enyne metathesis, a versatile method for the synthesis of unsaturated bicyclic and polycyclic compounds, has also been effectively performed using the cationic ruthenium allenylidene complex 14. A representative example is the synthesis, in good yield (86%), of 2,2′,5,5′-tetrahydro-3,3′-bifuran from the corresponding dienyne using 2.5 mol% complex 14-BF₄, in toluene, at 80°C (48), see Equation (xvii).

View this figure

Related with the cationic p-cymene ruthenium allenylidene complexes discussed above, the chelated compound 32 (with OTf⁻), containing the tethered phenyl ring instead of p-cymene as the ligand, showed good activity and stability in RCM reactions, allowing synthesis of macrocyclic compounds in appreciable yield, see Equation (xviii).

View this figure

One of the basic structural parameters for this type of chelated ruthenium complexes, namely the nature of the ‘non-coordinating’ counterion X⁻, was found to play an important part in determining their activity and stability in metathesis reactions.

Conclusions

The well defined, neutral and cationic ruthenium allenylidene complexes constitute a new, attractive class of highly effective pre-catalysts for various olefin metathesis reactions. They are easily accessible from commercial reagents, display a wide range of activity and selectivity in metathesis reactions, and show good tolerance towards a large array of functional groups.

This class of ruthenium complexes allows the synthesis of many hetero- and carbocyclic compounds by ring-closing metathesis (RCM) and enyne metathesis as well as production of specialty polymers from cycloolefins by ring-opening metathesis polymerisation (ROMP). Their activity and selectivity can be finely tuned by: suitable variations in the ligand sphere of the metal, appropriate substitution on the allenylidene moiety and a proper choice of the ‘non-coordinating’ counterion.

Of great significance for practical applications is the observation that the catalytic activity of the cationic, 18-electron ruthenium complexes can be further enhanced by addition of Lewis or Brönsted acids or by irradiation with UV light. Single-crystal X-ray studies unambiguously determined the structures of the neutral, 16-electron, and cationic, 18-electron ruthenium allenylidene complexes and also the defining features of the metal-ligand bonds.

The majority of the ruthenium allenylidene complexes are rather stable at room temperature or upon heating. Compelling proof exists that some ruthenium allenylidene complexes rearrange intramolecularly to the parent ruthenium indenylidene complexes, the latter also displaying high metathesis activity and selectivity. Elegant NMR studies clearly evidenced, for the first time, the transformation of cationic ruthenium allenylidene complexes into the indenylidene congeners via alkenylcarbyne ruthenium species.

Overall, the ready accessibility and good to excellent catalytic properties of the well-defined ruthenium allenylidene complexes turn out to be valuable bases for further development and refinement of this class of metathesis pre-catalysts.