Cyclic Ruthenium-Alkylidene Catalysts for Ring-Expansion Metathesis Polymerization

Abstract

A series of cyclic Ru-alkylidene catalysts have been prepared and evaluated for their efficiency in ring-expansion metathesis polymerization (REMP). The catalyst structures feature chelating tethers extending from one N-atom of an imidazolylidine ligand to the Ru metal center. The catalyst design is modular in nature, which provided access to Ru-complexes having varying tether lengths, as well as electronically different NHC ligands. Structural impacts of the tether length were unveiled through ¹H NMR spectroscopy as well as single-crystal X-ray analyses. Catalyst activities were evaluated via polymerization of cyclooctene and key data are provided regarding propagation rates, intramolecular chain-transfer, and catalyst stabilities, three areas necessary for the efficient synthesis of cyclic poly(olefin)s via REMP. From these studies, it was determined that while increasing the tether length of the catalyst leads to enhanced rates of polymerization, shorter tethers were found to facilitate intramolecular chain-transfer and release of catalyst from the polymer. Electronic modification of the NHC via backbone saturation was found to enhance polymerization rates to a greater extent than did homologation of the tether. Overall, cyclic Ru-complexes bearing 5- or 6-carbon tethers and saturated NHC ligands were found to be readily synthesized, bench-stable, and highly active catalysts for REMP.

Introduction

The exploration of Ru-based metathesis catalysts has opened doorways to multiple areas of synthetic and polymer chemistry.^1,2 Advances in these areas have been made possible via development of new catalyst scaffolds based upon bis(phosphine) complex 1 (Figure 1), or those bearing N-heterocyclic carbene (NHC) ligands such as 2 and 3. The introduction of catalysts based on 1 – 3, but predisposed for specific tasks, has further expanded the potential of olefin metathesis. For example, areas such as solid-supported catalysis,³ asymmetric olefin metathesis,⁴ tandem catalysis,⁵ living polymerization,^1a,6 acyclic diene metathesis,⁷ and stereoselective cross-metathesis (CM) have each benefited from breakthroughs in catalyst design, development, and application.⁸

Figure 1.

Representative Ru-based metathesis catalysts.

Recently, a Ru-based catalyst design was reported that featured a chelating N-to-Ru tether (Figure 2).⁹ Whereas the catalytic activities of 4_cyc – 6 _cyc have not been explored, 7_cyc was found to mediate the synthesis of cyclic polymers from cyclic monomers (Scheme 1).^10,11 This ring-expansion metathesis polymerization (REMP) afforded the ability to produce cyclic polymers on large scale from diverse, readily-available cyclic monomers.^12,13 While the high catalytic activity of 7_cyc was desirable, the synthesis and storage of this compound were complicated by instability.

Figure 2.

Cyclic Ru-alkylidene metathesis catalysts.

Scheme 1.

Proposed REMP catalytic cycle.

To realize the potential in the area of cyclic polymer chemistry, catalysts should be readily synthesized in good yields, be easily purified to eliminate any acyclic contaminants, and have an appropriate balance of stability (e.g., during storage as well as polymerizations) and activity. To address issues of stability, we envisioned that catalysts with shorter tether lengths, such as 4_cyc, 5_cyc, and 6_cyc, which contain 4-, 5-, and 6-carbon tethers, respectively, may be advantageous. A potential drawback, however, is that this may be accompanied by decreased catalytic activities. Therefore, we designed catalysts to incorporate two key structural features, shortened tether lengths and saturated NHC backbones, expected to synergistically provide REMP catalysts of high stabilities and activities.¹⁴

We also hoped to further elucidate the mechanism by which REMP proceeds through judicious catalysts design. Initially, REMP was proposed to proceed via a ring-expansion initiation event from a cyclic Ru-alkylidene catalyst (Scheme 1) and propagate as cyclic monomers were incorporated into a growing cyclic polymer. Upon consumption of monomer, a final catalyst release step would provide the original catalyst and the desired cyclic polymer. The polymerization mechanism depicted in Scheme 1 has several intriguing features including: 1) opening of a chelated Ru-alkylidene catalyst, 2) propagation with the prospect of competing intramolecular chain-transfer events, and 3) a final release of the original catalyst via intramolecular CM.

Many scenarios are consistent with Scheme 1, depending on the relative rates of initiation, propagation, intramolecular chain-transfer, and catalyst release. Initial studies using catalyst 7_cyc demonstrated the ability to control polymer molecular-weight (MW) using the monomer/catalyst loading.¹¹ This corresponds to a regime in which nearly complete initiation occurs, and catalyst release does not take place prior to complete monomer consumption. Another key observation was that after complete conversion of monomer, the MW of the cyclic polymers progressively decreased in the presence of 7_cyc, indicating significant amounts of intramolecular chain-transfer (Scheme 1). Alternatively, if the rate of propagation is much greater than that of initiation, and the rates of intramolecular chain-transfer and catalyst release are negligible, then all monomer species may be incorporated into a number of macrocycles equal to the number of catalyst molecules that initiated. This latter scenario would yield cyclic polymers in which Ru is incorporated into the backbone. Finally, a catalyst that shows relatively rapid release from a propagating cyclic chain may be expected to initiate, propagate (e.g., oligomerize), and then release, ultimately providing multiple polymer rings from each catalyst complex. Therefore, understanding how the catalyst design influences the relative kinetics of these processes is central to controlling the nature and distribution of products obtained via REMP.

To better understand each of the mechanistic aspects of REMP, and provide guidance for REMP catalyst design, we sought to investigate a homologous series of cyclic catalysts of varying N-to-Ru tether lengths (Figure 2). The tether length may be central in controlling structural features of the catalyst such as: 1) inherent ring-strain in the cyclic Ru-complexes, 2) relative orientations of the NHC and PCy₃ ligands about the metal center, and 3) rotation about the Ru-alkylidene (i.e., Ru=C-R) bond. As will be discussed below, a combination of NMR spectroscopy and single-crystal X-ray analyses of cyclic catalysts ultimately revealed key connections between their structures and activities.

Considering each step in the REMP cycle, it was expected that the tether length ideal for polymerization activity might be unfavorable for catalyst release. Specifically, intramolecular metathesis to reform and release the initial catalyst from the polymer is expected to be most efficient for shorter tether lengths. In contrast, increased tether lengths may be beneficial for polymerization rates, considering longer tethers may increase ring-strain of the catalyst or provide necessary flexibility within the structure. Encouraged by the modular nature of the NHC ligand, and the possibility of controlling REMP catalyst activities via tether length, we prepared and analyzed a homologous series of cyclic REMP catalysts (4_cyc – 7_cyc, Figure 2), as well as analogues possessing saturated imidazolinylidene ligands. Herein, we report the study of their activities in various steps of the REMP cycle, as well as key structure-activity relationships.

Results and Discussion

Catalyst Syntheses

The syntheses of complexes 4_cyc and 5_cyc were previously described by Fürstner.⁹ To our knowledge, however, their catalytic activity has not been reported. Catalysts 6_cyc and 7_cyc were prepared analogously, as described in Scheme 2. The corresponding imidazolium salts (8) were first obtained by alkylation of 1-mesitylimidazole.¹⁵ Ligand exchange was then achieved via deprotonation of the imidazolium salt, followed by addition of bis(phosphine) complex 1 (8/1 molar ratio = 2:1) to give “open” complexes 4_acyc – 7_acyc.^{10, 16} In general, ligand exchange proceeded smoothly and the desired non-chelated complexes were isolated in good yields after chromatography on silica gel.^17,18 Intramolecular metathesis/cyclization was conducted in a PhH/pentane mixture (1:15 v/v) at 70 °C and 0.001 M to give the final “closed” complexes 4_cyc – 7 _cyc. Each of the catalysts could be purified by chromatography on silica gel, however, purification of 4_cyc – 6_cyc was more efficiently accomplished via recrystallization from Et₂O/pentane.¹⁵

Scheme 2.

Synthesis of cyclic REMP catalysts 4_cyc – 7_cyc.

We noted that the efficiency of cyclization of open complexes 4_acyc – 7_acyc to give cyclic catalysts 4_cyc – 7_cyc is highly dependent on the tether length (Table 1). Ostensibly, the ability of an open complex to undergo intra- versus intermolecular metathesis events gives some indication of the tendency for the proposed catalyst release step in Scheme 1. Table 1 summarizes the results of cyclization reactions for each catalyst at 0.01 and 0.001 M. In each case, yields were markedly improved at lower concentration (0.001 M) as determined by ¹H NMR analysis of the crude reaction mixtures. At 0.01 M, additional alkylidene peaks were observed via ¹H NMR spectroscopy that were upfield of signals characteristic of 1, 4_acyc – 7_acyc, or 4_cyc – 7 _cyc. These signals may be attributed to CM products such as those arising from CM involving styrene (formed as a product in the cyclization step), or the terminal olefin on an NHC ligand with another Ru-complex. As expected, these intermolecular metathesis events were significantly diminished at lower concentration. Prolonged reaction times did not result in increased yields of the desired cyclic species, rather decomposition was observed. It is worth noting that in the case of 7_acyc, no product was observed when cyclization was conducted at 0.01 M. Therefore, catalyst release during REMP may be slow in comparison with other chain-transfer events when 7_cyc is employed.¹⁹

Table 1.

Cyclization to give cyclic catalysts 4_cyc – 7_cyc.^a

cyclic catalyst	tether length	yield (%)b
		0.01 M	0.001 M
4cyc	4	62	81
5cyc	5	76	97
6cyc	6	81	97
7cyc	7	0	63

^aReactions conducted in dry C₆D₆ under N₂ atmosphere at 80 °C for 1 h.

^bDetermined by ¹H NMR spectroscopy.

Considering the enhanced activity observed from saturation of the NHC backbone (cf. 2 and 3),¹⁴ we were motivated to investigate cyclic catalysts incorporating this design feature. As depicted in Scheme 3, a PhCH₃ solution of N-mesitylethylenediamine (9)²⁰ was treated with HC(OEt)₃ in the presence of catalytic PTSA and stoichiometric bromo-olefin at 110 °C. This one-pot procedure effected cyclization and alkylation to provide the imidazolinium salts 10 in excellent yields.²¹ Unfortunately, attempts at direct deprotonation of 10 using KHMDS in the presence of bis(phosphine) complex 1 were complicated by NHC dimerization and provided low yields of the desired products.²² Alternatively, treatment of 10 with NaH in CHCl₃ cleanly provided neutral adducts 11.²³ Heating THF solutions of 11 (0.001 M) in the presence of 1 (11/1 molar ratio = 2:1) accomplished ligand exchange as well as cyclization to provide the desired cyclic catalysts 5_cyc·H₂ and 6_cyc·H₂ in 46 and 57% overall yields, respectively.²⁴ Although 5_cyc·H₂ and 6_cyc·H₂ were each isolable via chromatography on silica gel, both were found to be crystalline solids and were routinely recrystallized by slow addition of pentane into saturated PhH solutions of the complexes. Similar to 5_cyc and 6_cyc, the saturated catalysts 5_cyc·H₂ and 6_cyc·H₂ displayed good stability both in the solid-state and in solution.²⁵

Scheme 3.

Synthesis of “saturated” catalysts 5_cyc·H₂ and 6_cyc·H₂.

Structural Analyses

The structural impacts of changing the tether lengths of catalysts 4_cyc – 7 _cyc resulted in significant differences in catalyst activities (see polymerization studies below for more discussion). In addition to understanding the structure-activity relationships pertaining to REMP catalysts, a more general understanding of catalyst architecture may lead to breakthroughs in catalyst design as well as fundamental mechanistic insights of olefin metathesis. Cyclic catalysts 4_cyc – 6_cyc were found to show tether length-dependent trends in three key structural parameters summarized in Table 2:1) rotation about the Ru1-C2 bond, 2) the C1-Ru1-P1 bond angle, and 3) the Ru-C1 bond length (Figure 3).²⁶

Table 2.

Selected ¹H NMR and single-crystal X-ray data for 4_cyc – 6_cyc, 5_cyc·H₂ and 6_cyc·H₂.

Catalyst	4cycb	5cycb	6cyc	5cyc·H2	6cyc·H2
NMR data
δ H2 (ppm)a	19.70	20.50	19.71	20.39	19.61
3JH2,P1 (Hz)a	14.1	10.5	5.1	9.3	5.0
dihedral angles (°)
Cl2-Ru1-C2-C3	51.3	26.6	16.2	18.3	21.1
N1-C1-Ru1-C2	156.5	162.2	153.5	160.0	151.2
bond angles (°)
C1-Ru1-P1	171.0	166.0	163.3	165.3	168.8
N1-C1-Ru1	128.8	127.0	126.6	124.3	124.0
N2-C1-Ru1	127.8	128.9	129.4	128.1	129.5
N1-C1-N2	103.4	104.1	103.6	107.5	107.2
bond lengths (Å)
Ru1-C1	2.076	2.091	2.113	2.072	2.084
Ru1-C2	1.812	1.806	1.823	1.821	1.800
Ru1-P1	2.402	2.421	2.421	2.417	2.423

^aData taken in C₆D₆ at ambient temperature.

^bSee ref 9.

Figure 3.

(top): X-ray crystal structures of 5_cyc·H₂, 6_cyc, and 6_cyc·H₂. Solvent molecules and hydrogens have been removed for clarity. Ellipsoids are drawn at the 50% probability level. (bottom): ¹H NMR spectra (C₆D₆) of alkylidene proton of 5_cyc·H₂, 6_cyc, and 6_cyc·H₂.

Although many structural features of 4_cyc – 7_cyc are best observed via solid-state analysis, rotation about the Ru1-C2 bond is manifested in the coupling constants between the P1 and H2 (i.e., the alkylidene proton, bonded to C2) atoms in the ¹H NMR spectra (Table 2 and Figure 3). Complexes 4_cyc and 5_cyc, which were previously characterized in solution and solid-state,⁹ displayed coupling constants of ³J_H2,P1 = 14.1 and 10.5 Hz (solvent = C₆D₆), respectively. The smaller coupling constant observed from complex 5_cyc, in comparison with 4_cyc, indicated that the corresponding atoms in the former are closer to a perpendicular arrangement. Consistent with this trend, a smaller coupling constant was observed in the ¹H NMR spectrum of 6_cyc (i.e., ³J_H2,P1 = 5.1 Hz), indicating that the alkylidene proton (H2) was projected nearly perpendicular to the Ru1-P1 bond. The ¹H NMR spectrum of 7_cyc revealed a coupling constant of ³J_H2,P1 = 10.2 Hz, which may be ascribed to the increased ring size (cf. 6_cyc) inducing twist about the Ru1-C2 bond.²⁷

To further investigate the structures of the cyclic catalysts, we compared single-crystal X-ray data of 4_cyc – 6 _cyc, as well as saturated analogues 5_cyc·H₂ and 6_cyc·H₂.²⁶ The crystal structures of these complexes confirmed a variable degree of rotation about the Ru1-C2 bond, as determined from the Cl2-Ru-C2-C3 dihedral angles (Table 2). Overall, for 4_cyc – 6_cyc, decreased ³J_H2,P1 values corresponded to decreased dihedral angles suggesting that the solution and solid-state structures of the catalysts are similar. It should be noted that while the ³J_H2,P1 values observed from 5_cyc·H₂ and 6_cyc·H₂ were consistent with each complex’s respective unsaturated analogue, solid-sate analysis revealed that the Cl2-Ru-C2-C3 dihedral angles were not consistent with the trend observed from the unsaturated series.

Stepwise increase in the tether lengths was found to cause increasing nonlinearity in the C1-Ru1-P1 bond angles. Specifically, catalysts 4_cyc, 5_cyc, and 6_cyc have C1-Ru1-P1 bond angles of 171.0, 166.0, and 163.3°, respectively. One rationale for this trend may be that increasing the tether length caused the NHC ligand to tilt to accommodate the increased steric demand of the tether. A consequence of this tilt is that the Mes group is forced closer to the PCy₃ group which may facilitate dissociation of the phosphine during initiation (see below for a comparison of catalyst activities).^14a In addition, the tilted conformation may be difficult to restore which would hinder catalyst release and thus manifest in more rapid polymerization. Steric demands also resulted in discernable increases in the Ru1-P1 and Ru1-C1 bond lengths as the tether length was increased. For example, upon extension of the tether, the Ru1-C1 bond length increased from 2.076 Å for 4_cyc to 2.113 Å for 6_cyc. The saturated catalysts, 5_cyc·H₂ and 6 _cyc·H₂, showed changes in their Ru1-P1 and Ru1-C1 bond lengths that were consistent with those observed in the unsaturated series.

Ring-Expansion Metathesis Polymerization

To our knowledge, the catalytic activities of complexes 4_cyc – 6_cyc have not been explored in either small-molecule or polymer syntheses. In addition, electronic variants such as saturation of the NHC backbone have not been investigated. To compare the activities of 4_cyc – 7_cyc in REMP, we examined their relative efficiencies in the polymerization of cyclooctene (COE) to poly(cyclooctene) (PCOE). As can be seen from the data presented in Figure 4, the relative efficiencies of the catalysts showed a strong dependence on the length of the chelating tether.²⁸ In general, increased tether length was accompanied by an increase in catalytic activity. For example, comparison of the unsaturated catalysts revealed the time required to reach >95% conversion was nearly 24 h for 4_cyc (green line), approximately 8 h for 5_cyc (purple line) and 6_cyc (blue line), and less than1 h for 7_cyc (red line).

Figure 4.

REMP of COE using catalysts 4_cyc (green), 5_cyc (purple), 5_cyc·H₂ (orange), 6_cyc (blue), 6_cyc·H₂ (black), and 7_cyc (red). Conditions: CD₂Cl₂, 40 °C, [M/C]₀ = 1000:1, [M]₀ = 0.5 M. Conversion determined by ¹H NMR spectroscopy.

Saturation of the NHC backbone was found to dramatically increase catalyst activity. As expected, 5_cyc·H₂ and 6_cyc·H₂ each displayed faster polymerization rates than their unsaturated analogues 5_cyc and 6 _cyc, respectively (Figure 4). Surprisingly, the rate acceleration resulting from NHC saturation appeared to be greater than for homologation of the tether length. Specifically, 6_cyc·H₂ was found to achieve >95% conversion in shorter reaction times than did 7_cyc. Similarly, 5_cyc·H₂ was found to be a more active polymerization catalyst than 6_cyc. Overall, the data revealed that judicious combinations of shorter tether lengths (i.e., 5- or 6-carbon tethers) and NHC backbone saturation (e.g., 5_cyc·H₂ and 6_cyc·H₂) provided a desirable balance of catalyst stabilities, synthetic accessibility, and activities.

Catalyst stability takes on particular importance in REMP as decomposition of the catalyst before, during, or after polymerization could potentially lead to linear polymers, instead of the envisioned macrocycles. In addition, relative stabilities are important factors in the general development of new metathesis catalysts, especially considering that catalyst stability and activity are often inversely related.²⁹ To explore the stabilities of REMP catalysts 4_cyc – 7_cyc during polymerization reactions, we plotted the ln([COE]) versus time for REMP of COE (Figure 5). The logarithmic plots were found to be linear (R² values ranged from 0.969 – 0.997) between 20 and 80% conversion of COE to PCOE, indicating that catalyst decomposition was negligible in all cases during the time of the polymerization reactions. Closer examination of the plots revealed that the only discernable deviations from linearity (i.e., pseudo-first-order rate kinetics) involved apparent increases in the rate of monomer consumption. This observation can be rationalized by a relatively slow initiation period which would manifest in a gradual increase in the number of propagating polymer chains, and concomitant increase in the rate of conversion.

Figure 5.

Log plots for REMP of COE using catalysts 4_cyc (green), 5_cyc (purple), 5_cyc·H₂ (orange), 6_cyc (blue), 6_cyc·H₂ (black), and 7_cyc (red). Linear least-squares fitting gave R² values of: 4_cyc, 0.997; 5_cyc, 0.969; 5 _cyc·H₂, 0.991; 6_cyc, 0.998; 6_cyc·H₂, 0.990; and 7_cyc, 0.991. Conditions: CD₂Cl₂, 40 °C, [M/C]₀ = 1000:1, [M]₀ = 0.5 M. Data recorded using ¹H NMR spectroscopy.

Catalyst Release

A unique aspect of REMP, in comparison with ring-opening metathesis polymerization (ROMP), is the requirement for an intramolecular chain-transfer event with the olefin nearest to the NHC to release the initial cyclic catalyst and provide a cyclic polymer free of Ru (Scheme 1). While removal of Ru from linear polymers obtained via ROMP can be done efficiently using a terminating agent, such as ethyl vinyl ether, this method is incompatible with REMP as it may result in linear polymer formation.³⁰ Given the importance of catalyst release from the cyclic polymers, we investigated each catalyst’s propensity to undergo intramolecular cyclization during polymerization that would be indicative of the catalyst’s ability to be released from a polymer.

We envisioned that conducting polymerizations using “open” catalysts 5_acyc – 7_acyc would provide insight into each catalyst’s ability to perform intramolecular CM to release “closed” catalysts 5_acyc – 7_acyc.³¹ Propagation via growing Ru-alkylidene species A (Scheme 4) would inherently compete with catalyst cyclization (e.g., A → 5_cyc + B), and provide an indication of each catalyst’s propensity to be released from the polymer chain.

Scheme 4.

Proposed species observable upon ROMP of COE using open catalysts 5_acyc – 7_acyc.

To investigate, we conducted polymerizations of COE using open catalysts 5_acyc – 7_acyc in CD₂Cl₂ at 40 °C ([M/C]₀ = 250:1, [M]₀ = 0.5 M) and monitored the alkylidene region of the ¹H NMR spectrum as the reactions progressed. Each Ru-complex shown in Scheme 4 was identified by characteristic chemical shifts of the corresponding alkylidene protons. In CD₂Cl₂, complexes 5_acyc – 7 _acyc gave sharp benzylidene resonances as singlets at δ = 19.30 ppm, whereas propagating species (A) displayed broad multiplets at δ = 18.69 ppm. Cyclic catalysts 5_cyc, 6_cyc, and 7_cyc displayed signals at δ = 20.23, 19.35, and 19.67 ppm, respectively, with multiplicities matching those in Table 2.

We first examined open catalysts 6_acyc and 7_acyc as these were representative of the most active cyclic catalysts (6_cyc and 7_cyc, respectively) for this series. Catalysts 6_acyc and 7_acyc gave similar results, and polymerization of COE was found to reach completion faster than did cyclization of 6_acyc and 7_acyc to give 6_cyc and 7_cyc, respectively. Specifically, complete conversion of COE was achieved in less than 5 min for each catalyst.³² The mole fraction of cyclic catalyst (6_cyc/7_cyc) observed at this point, however, was only ca 10%, relative to 6_acyc/7_acyc (ca 30%) and A (ca 60%). Continued heating resulted in diminished amounts of 6_acyc/7_acyc and 6_cyc/7_cyc in each case, with concomitant increases in the relative amounts of A. As will be discussed in the next section, the continued progression to form A may have been due to incorporation of free cyclic catalyst into the polymer chains. After ca 1 h, only trace amounts of cyclic species 6_cyc/7_cyc could be observed. Overall, these results suggested that cyclization is not favored over polymerization for catalysts bearing 6- or 7-membered tethers, and that cyclization to release catalyst 6_cyc or 7_cyc after polymerization is not likely.

We next investigated the behavior of 5_acyc under the same conditions as described above. In contrast to the longer tethered analogues 6_acyc and 7_acyc, polymerization reactions using 5_acyc revealed much faster cyclization (to give 5_cyc) relative to polymerization. Figure 6 shows the mole fraction of each catalytic species (5_acyc, A, and 5_cyc) as well as the conversion of COE to PCOE over time. As can be seen, almost complete formation of cyclic catalyst 5_cyc was observed after ca 45 min, at which time the polymerization had reached only 48% conversion. Moreover, the amount of catalytic species within the polymer chains (A) quickly diminished to nearly undetectable amounts. The data presented in Figure 6 supports that cyclization to form 5_cyc is favored over propagation and that the background rate of cyclization (i.e., 5_acyc → 5_cyc) is also significant for this catalyst. In addition, the persistent amount of 5_cyc that is observed relative to propagating species (A) suggested that incorporation of 5_cyc into existing polymer chains is unlikely. The continued conversion of COE to PCOE, with persistent observation of only 5_cyc, suggested that this cyclic catalyst reached an equilibrium between propagating (e.g., A) and resting (i.e., 5_cyc) states that strongly favored the latter.

Figure 6.

Left axis: Conversion of COE to PCOE using 5_acyc (black). Right axis: Mole fraction of 5_cyc (red), 5_acyc (blue), and A (green). Conditions: CD₂Cl₂, 40 °C, [COE/5_acyc]₀ = 250:1, [COE]₀ = 0.5 M. Conversion determined by ¹H NMR spectroscopy.

Collectively, the experiments investigating the behavior of open catalysts 5_acyc – 7_acyc revealed that controlling the tether lengths of cyclic catalysts may dictate polymerization kinetics with regard to polymer MWs and polydispersities. For example, shorter tether lengths may facilitate catalyst release during polymerization (Scheme 1), ultimately leading to multiple macrocycles produced from each catalyst molecule. Alternatively, REMP catalysts displaying little tendency to be released from a cyclic polymer may provide access to cyclic block copolymers or other advanced macrocycles.

Interaction Between Free Catalyst and Polymer

As mentioned previously, it may be possible for a cyclic catalyst to equilibrate with poly(olefin)s and become incorporated (or reincorporated) into a polymer chain. This equilibrium, depicted in Scheme 5, may be tether length-dependent given that ring-opening of the catalyst may be a driving force toward incorporation into the polymer. With regard to REMP, the reversibility of intramolecular chain-transfer and catalyst release (Scheme 1) would result in an equilibrium amount of Ru species contained within the final cyclic polymers. Therefore, understanding each catalyst’s affinity toward polymer incorporation is important for understanding the potential purity of the cyclic polymers. To investigate, we prepared linear PCOE via ROMP using acyclic catalyst 3 in the presence of 3-hexene as a chain-transfer agent. This provided a hydrocarbon polymer (M_n = 150 kDa, PDI = 2.1) which closely resembled the PCOE obtained via REMP in composition.³³ The linear PCOE was then treated with each of the cyclic catalysts 5_cyc – 7_cyc (olefin/catalyst molar ratio = 100:1) in CD₂Cl₂ at 40 °C. The equilibration of catalyst and polymer was monitored via ¹H NMR spectroscopy using anthracene as an internal standard; key NMR signals of the cyclic catalysts and incorporated species were similar to those observed in the previous section (Scheme 4).

Scheme 5.

Equilibration of cyclic catalyst and linear PCOE.

As expected, incorporation of cyclic catalyst into the polymer chain was dependent on the tether length of the catalyst. Specifically, after 1 h ca 11% of catalyst 7_cyc had become incorporated into the polymer, whereas catalyst 6_cyc showed only 3% incorporation over the same time period. Catalyst 5_cyc, however, revealed no incorporation even after extended periods (up to 6 h). To compare, catalyst 5_cyc·H₂ was also studied and gave similar results as 5_cyc. Overall, although the amount of incorporated catalyst was small in each case, there appeared to be some equilibration of free catalyst into the poly(olefin) depending upon the length of the catalyst tether. Notably, equilibration appeared to be slow in comparison with polymerization data presented above. Therefore, avoiding prolonged polymerization reaction times may be sufficient for minimizing the amount of Ru in the resulting polymers.

Conclusions

In summary, we describe the synthesis and characterization of a series of cyclic Ru-alkylidene catalysts with particular focus on their ability to mediate ring-expansion metathesis polymerization. Both catalyst tether length as well as NHC electronics were found to significantly impact different aspects of the polymerization mechanism. Whereas shorter tether lengths were more efficient for catalyst release from the polymer, the caveat for these systems was found to be slower polymerization rates. Saturation of the NHC backbone, however, increased polymerization efficiency and effectively balanced activity loss due to shortening of the tether. Overall, catalyst stabilities were found to be good over the course of the polymerization experiments. The ability to control catalyst activity by a combination of tether length and ligand electronics may lead to new opportunities in olefin metathesis and catalyst design.

Experimental Section

Materials and Methods

¹H and ¹³C NMR spectra were recorded using a Varian Mercury 300 or Varian Inova 500 spectrometer and were routinely run using broadband decoupling. Chemical shifts (δ) are expressed in ppm downfield from tetramethylsilane using the residual protiated solvent as an internal standard (CDCl₃ ¹H: 7.26 ppm and ¹³C: 77.0 ppm; C₆D₆ ¹H: 7.20 ppm and ¹³C: 128.0 ppm). ³¹P NMR spectra were externally referenced to 85% H₃PO₄ (0.00 ppm). Coupling constants are expressed in hertz (Hz). THF, CH₂Cl₂, Et₂O, pentane, PhH, PhCH₃, and C₆D₆ were obtained from solvent purification columns.³⁴ CD₂Cl₂ used for NMR-scale experiments was distilled over CaH₂ under N₂ prior to use. CHCl₃ was distilled over P₂O₅ under N₂ prior to use. Ru-complex 1 was obtained from Materia, Inc. All other solvents and reagents were of reagent quality and used as obtained from commercial sources. Chromatography was performed with neutral silica gel (TSI Scientific, 230–400mesh, pH 6.5 – 7.0). Crystallographic data have been deposited at the CCDC, 12 Union Road, Cambridge CB2 1EZ, U.K., and copies can be obtained on request, free of charge, by quoting the publication citation and the deposition numbers 687290 (5_cyc·H₂), 683585 (6_cyc), and 687247 (6_cyc·H₂).

Cyclic complex 5_cyc·H₂

In a Schlenk tube, chloroform adduct 11a (200 mg, 0.51 mmol) was dissolved in dry THF (515 mL) under at atmosphere of dry N₂. To the solution was added Ru-complex 1 (210 mg, 0.26 mmol). The flask was sealed and the reaction mixture was stirred in an oil bath at 70 °C for 2 h. Afterward, the cooled reaction mixture was concentrated under vacuum, redissolved in a minimal amount of PhH, and treated dropwise with pentane until crystallization ensued (X-ray analysis was performed on crystals obtained in this manner). The solids were collected by vacuum filtration, rinsed with 5% Et₂O/pentane, and dried under vacuum to provide 147 mg (81% yield) of the desired complex as a tan solid. ¹H NMR (500 MHz, C₆D₆): δ 20.39 (dt, ³J_H,P = 9.3 Hz, J_H,H = 4.5 Hz, 1H), 6.93 (s, 2H), 3.17 (t, J = 10.0 Hz, 2H), 3.10-3.09 (m, 2H), 2.85 (t, J = 10.0 Hz, 2H), 2.75 (br, 2H), 3.71 (s, 6H), 2.60-2.53 (m, 3H), 2.32 (br 2H), 2.23 (s, 3H), 1.95-1.92 (m, 6H), 1.78 (br, 6H), 1.68 (br, 2H), 1.47-1.45 (m, 6H), 1.34-1.29 (m, 10H), 1.19-1.17 (m, 2H). ¹³C NMR (125 MHz, C₆D₆): δ 216.0 (J_CP = 85.9 Hz), 138.4, 137.6, 136.6, 129.9, 57.9 (J_CP = 4.9 Hz), 51.2 (J_CP = 3.5 Hz), 48.5 (J_CP = 2.7 Hz), 47.9, 32.1 (J_CP = 15.1 Hz), 29.8, 28.2 (J_CP = 10.4 Hz), 27.3, 26.8, 26.7, 21.1, 20.0. ³¹P NMR (121 MHz, C₆D₆): δ 27.0. HRMS m/z calcd for C₃₅H₅₇Cl₂N₂PRu [M⁺] 708.2680, found 708.2659.

[1-(6-Heptenyl)-3-mesitylimidazolylidene]RuCl₂(=CHPh) (PCy₃) (6_acyc)

Imidazolium bromide 8_n=5 (400 mg, 1.10 mmol), was suspended in dry PhCH₃ (7 mL) under dry N₂. To the solution was added NaOtBu (106 mg, 1.10 mmol) and the resulting mixture was stirred at RT for 12 h. Ru-complex 1 (453 mg, 0.55 mmol) was then added in a single portion and the resulting mixture was stirred for 1 h during which time a color change from purple to brown was observed. Upon completion, the mixture was filtered through a thin pad of TSI silica gel using Et₂O/pentane (1:4 v/v) as eluent. The filtrate was concentrated under vacuum without heating. Purification by column chromatography on TSI silica gel under N₂ pressure (10% Et₂O/pentane) provided 408 mg (90% yield) of the desired compound as a red-purple powder. ¹H NMR (major isomer) (300 MHz, C₆D₆): δ 19.85 (s, 1H), 7.05-6.94 (m, 2H) 6.56-6.55 (m, 1H), 6.31-6.15 (m, 3H), 5.90-5.76 (m, 1H), 5.15-5.03 (m, 2H), 4.70 (t, J = 7.7 Hz, 2H), 2.65-2.53 (m, 4H), 2.13 (s, 3H), 1.97-1.08 (m, 37H), 1.80 (s, 6H). ³¹P NMR (major isomer) (121 MHz, C₆D₆): δ 34.4. HRMS m/z calcd for C₄₄H₆₅Cl₂N₂PRu [M⁺] 824.3306, found 824.3298.

Cyclic complex 6_cyc

Ru-complex 6_acyc (400 mg, 0.48 mmol), was dissolved in dry PhH (30 mL) and pentane (450 mL) in a Schlenk tube under dry N₂. The mixture was then placed in a oil bath at 70 °C and stirred for 1 h. Upon completion, the solution was cooled to RT, transferred to a round-bottom flask, and concentrated under vacuum without heat. The crude material was triturated with 20% Et₂O/pentane (50 mL) for ca 20 min. The solids were then collected via vacuum filtration, rinsed with pentane, and dried under vacuum to provide 335 mg (96% yield) of the desired compound as a red-brown powder. X-ray quality crystals were obtained by slow addition of pentane to a PhH solution of the complex. ¹H NMR (300 MHz, C₆D₆): δ 19.71 (dt, ³J_H,P = 5.1 Hz, J_H,H = 5.5 Hz, 1H), 6.90 (s, 2H), 6.35 (d, J = 1.8 Hz, 1H), 6.14 (s, 1H), 2.74-2.51 (m, 3H), 2.51 (s, 6H), 2.23 (s, 3H), 1.99-1.31 (m, 40H). ¹³C NMR (125 MHz, C₆D₆): δ 186.0 (J_CP = 84.7 Hz), 138.2, 137.1, 129.4, 128.3, 123.6 (J_CP = 2.3 Hz), 120.4, 62.9, 47.0, 32.4 (J_CP = 16.9 Hz), 31.5, 29.8, 28.1 (J_CP = 10.0 Hz), 26.8, 22.9, 21.1, 19.7. ³¹P NMR (121 MHz, C₆D₆): δ 33.3. HRMS m/z calcd for C₃₆H₅₇Cl₂N₂PRu [M⁺] 720.2680, found 720.2671.

Cyclic complex 6_cyc·H₂

This compound was prepared analogously to 5_cyc·H₂ from chloroform adduct 11b (280 mg, 0.69 mmol) and Ru-complex 1 (285 mg, 0.35 mmol) in THF (650 mL) to provide 369 mg (74% yield) of the desired complex as a red-brown solid. X-ray quality crystals were obtained by slow diffusion of pentane into a Et₂O/PhH (20:1 v/v) solution of the complex. ¹H NMR (500 MHz, C₆D₆): δ 19.61 (dt, ³J_H,P = 5.0 Hz, J_H,H = 5.7 Hz, 1H), 6.93 (s, 2H), 3.33-2.83 (m, 4H), 2.70 (s, 6H), 2.63-2.56 (m, 3H), 2.22 (s, 3H), 1.93-1.30 (m, 40H). ¹³C NMR (125 MHz, C₆D₆): δ 215.3 (J_CP = 80.1 Hz) 137.5, 137.0, 129.9, 129.5, 62.8, 51.7 (J_CP = 3.3 Hz), 47.7, 46.9, 32.2 (J_CP = 16.5 Hz), 29.7, 28.1 (J_CP = 10.1 Hz), 27.9, 26.8, 25.6, 23.6, 21.1. ³¹P NMR (121 MHz, C₆D₆): δ 30.4. HRMS m/z calcd for C₃₆H₅₉Cl₂N₂PRu [M⁺] 722.2837, found 722.2808.

Cyclic complex 7_cyc

This compound was prepared analogously to 6_cyc from open complex 7_acyc (150 mg, 0.18 mmol) in PhH (10 mL) and pentane (170 mL). Upon completion, the cooled reaction mixture was concentrated under vacuum without heat, then triturated with 20% Et₂O/pentane (10 mL) for ca 20 min. The solids were collected via vacuum filtration and further purification via column chromatography on TSI silica gel under N₂ pressure (30% Et₂O/pentane) provided 56 mg (42% yield) of the desired compound as a light brown powder. ¹H NMR (500 MHz, C₆D₆): δ 19.37 (dt, ³J_H,P = 10.2 Hz, J_H,H = 5.5 Hz, 1H), 6.92 (s, 2H), 6.38 (d, J = 1.5 Hz, 1H), 6.19 (s, 1H), 3.66 (br, 2H), 2.62-2.57 (m, 3H), 2.54 (s, 6H), 2.24 (s, 3H), 2.08-1.61 (m, 24H), 1.50-1.42 (m, 2H), 1.34-1.28 (m, 12H), 1.22-1.17 (m, 2H). ¹³C NMR (125 MHz, C₆D₆): δ 184.7 (J_CP = 97.5 Hz), 138.1, 137.6, 129.2, 128.4, 128.3, 128.1, 127.9, 123.3 (J_CP = 3.3 Hz), 119.9, 60.4, 47.1, 32.7 (J_CP = 16.1 Hz), 29.9, 28.2 (J_CP = 9.5 Hz), 26.8, 26.5, 24.7, 21.1, 20.7, 19.7. ³¹P NMR (121 MHz, C₆D₆): δ 26.3. HRMS m/z calcd for C₃₇H₅₉Cl₂N₂PRu [M⁺] 734.2837, found 734.2814.

1-(6-Heptenyl)-3-mesitylimidazolium bromide (8_n=5)

This compound was prepared analogously to 8_n=3,4,6 from N-mesitylimidazole (1.00 g, 5.37 mmol) and 7-bromo-1-heptene (1.0 mL, 6.55 mmol) in PhCH₃ (20 mL).⁹ Upon completion, the reaction mixture was concentrated under vacuum and the crude material was suspended in Et₂O (100 mL) and vigorously stirred for 12 h to produce a fine white suspension. The solids were collected via vacuum filtration under a stream a N₂ to provide 1.81 g (93% yield) of the desired compound as an off-white powder.³⁵¹H NMR (300 MHz, CDCl₃): δ 10.31 (dd appearing as t, J = 1.5 Hz, 1H), 7.94 (dd apearing as t, J = 1.7 Hz, 1H), 7.21 (dd apearing as t, J = 1.7 Hz, 1H), 6.94 (s, 2H), 5.76-5.63 (m, 1H), 4.96-4.84 (m, 2H), 4.67 (t, J = 7.4 Hz, 2H), 2.29 (s, 3H), 2.01 (s, 6H), 2.01-1.91 (m, 4H), 1.43-1.30 (m, 4H). ¹³C NMR (125 MHz, CDCl₃): δ 141.1, 138.1, 137.8, 134.0, 130.6, 129.7, 123.2, 123.1, 114.7, 50.1, 33.2, 30.2, 28.0, 25.3, 21.0, 17.5. HRMS m/z calcd for C₁₉H₂₇N₂ [M⁺] 283.2174, found 283.2186.

1-(5-Hexenyl)-3-mesitylimidazolinium bromide (10a)

To a solution of HC(OEt)₃ (10 mL) and PhCH₃ (10 mL) in a 50 mL round-bottom flask was added PTSA·H₂O (39 mg, 0.20 mmol), N-mesitylethylenediamine (9) (729 mg, 4.09 mmol), and 6-bromo-1-hexene (0.66 mL, 4.91 mmol). The flask was fitted with a H₂O-jacketed condenser and the reaction mixture was stirred under N₂ in an oil bath at 110 °C for 10 h. Afterward, the cooled reation mixture was concentrated under vacuum. The crude product was treated with Et₂O (40 mL) and vigorously stirred for 2 h to produce an off-white slurry. The solids were collected via vacuum filtration, rinsed with Et₂O, and dried under vacuum to provide 1.31 g (91% yield) of the desired compound. ¹H NMR (300 MHz, CDCl₃): δ 9.50 (s, 1H), 6.87 (s, 2H), 5.80-5.67 (m, 1H), 5.03-4.92 (m, 2H), 4.30-4.11 (m, 4H), 3.94 (t, J = 7.2 Hz, 2H), 2.26 (s, 6H), 2.24 (s, 3H), 2.12-2.05 (m, 2H), 1.76-1.66 (m, 2H), 1.49-1.42 (m, 2H). ¹³C NMR (125 MHz, CDCl₃): δ 158.7, 139.9, 137.6, 135.0, 130.4, 129.7, 115.2, 50.9, 48.7, 48.1, 32.8, 26.4, 25.2, 20.8, 17.9. HRMS m/z calcd for C₁₈H₂₇N₂ [M⁺] 271.2174, found 271.2161.

1-(6-Heptenyl)-3-mesitylimidazolinium bromide (10b)

This compound was prepared analogously to 10a from HC(OEt)₃ (7.0 mL), PhCH₃ (7.0 mL), PTSA·H₂O (27 mg, 0.14 mmol), N-mesitylethylenediamine (9) (500 mg, 2.80 mmol), and 7-bromo-1-heptene (0.51 mL, 3.36 mmol) to provide 951 mg (93% yield) of the desired compound. ¹H NMR (500 MHz, CDCl₃): δ 9.31 (s, 1H), 6.77 (s, 2H), 5.70-5.61 (m, 1H), 4.90-4.81 (m, 2H), 4.20-4.16 (m, 2H), 4.12-4.08 (m, 2H), 3.78 (t, J = 7.3 Hz, 2H), 2.16 (s, 3H), 2.15 (s, 6H), 1.96-1.91 (m, 2H), 1.64-1.58 (m, 2H), 1.35-1.30 (m, 2H), 1.27-1.21 (m, 2H). ¹³C NMR (125 MHz, CDCl₃): δ 158.5, 139.7, 138.1, 134.9, 130.3, 129.5, 114.4, 50.8, 48.6, 48.0, 33.1, 27.9, 26.7, 25.3, 20.7, 17.8. HRMS m/z calcd for C₁₉H₂₉N₂ [M⁺] 285.2325, found 285.2310.

1-(5-Hexenyl)-3-mesityl-2-(trichloromethyl)imidazolidine (11a)

Under an atmosphere of dry N₂, imidazolinium bromide 10a (443 mg, 1.26 mmol) was dissolved in dry CHCl₃ (6 mL). NaH (95 wt%, 38 mg, 1.51 mmol) was then added portionwise under a stream of N₂. The resulting mixture was placed in an oil bath at 55 °C and stirred for 10 h. Afterward, the cooled reaction mixture was diluted with Et₂O (100 mL), filtered through a thin pad of silica gel, and concentrated to provide 305 mg (62% yield) of the desired product as a pale yellow oil. ¹H NMR (300 MHz, CDCl₃): δ 6.86 (s, 1H), 6.85 (s, 1H), 5.91-5.78 (m, 1H), 5.07-4.95 (m, 2H), 4.73 (s, 1H), 3.86-3.78 (m, 1H), 3.62-3.55 (m, 1H), 3.41-3.32 (m, 1H), 3.23-3.16 (m, 1H), 3.10-3.02 (m, 1H), 2.98-2.90 (m, 1H), 2.35 (s, 3H), 2.70 (s, 3H), 2.25 (s, 3H), 2.16-2.09 (m, 2H), 1.68-1.43 (m, 4H). ¹³C NMR (75 MHz, CDCl₃): δ 147.8, 147.7, 142.8, 138.8, 138.7, 134.9, 132.6, 129.9, 129.5, 114.5, 108.2, 94.2, 58.2, 52.9, 52.6, 33.7, 29.7, 26.1, 20.7, 19.8, 19.3. HRMS m/z calcd for C₁₉H₂₇Cl₃N₂ [M⁺] 388.1240, found 388.1225.

1-(6-Heptenyl)-3-mesityl-2-(trichloromethyl)imidazolidine (11b)

This compound was prepared analogously to 11a from imidazolinium bromide 10b (730 mg, 2.0 mmol), CHCl₃ (8 mL), and NaH (95 wt%, 101 mg, 4.00 mmol) to provide 670 mg (83% yield) of the desired product as a pale yellow oil. ¹H NMR (300 MHz, CDCl₃): δ 6.88 (s, 1H), 6.86 (s, 1H), 5.92-5.79 (m, 1H), 5.08-4.96 (m, 2H), 4.75 (2, 1H), 3.87-3.80 (m, 1H), 3.65-3.58 (m, 1H), 3.42-3.33 (m, 1H), 3.25-3.17 (m, 1H), 3.10-3.02 (m, 1H), 3.00-2.92 (m, 1H), 2.37 (s, 3H), 2.29 (s, 3H), 2.27 (s, 3H), 2.14-2.05 (m, 2H), 1.68-1.54 (m, 2H), 1.50-1.38 (m, 4H). ¹³C NMR (75 MHz, CDCl₃): δ 142.8, 139.0, 138.6, 134.8, 132.6, 129.9, 129.4, 114.3, 108.2, 94.2, 58.4, 52.9, 52.6, 33.8, 30.1, 28.8, 26.4, 20.7, 19.8, 19.3. HRMS m/z calcd for C₂₀H₂₉Cl₃N₂ [M⁺] 402.1396, found 402.1382.

Supplementary Material

1_si_001. Supporting Information Available.

Additional representations of Figure 4, NMR spectra for all new compounds, CIF files containing X-ray structural data, atomic coordinates, thermal parameters, bond distances, and bond angles for complexes 5_cyc·H₂, 6_cyc and 6_cyc·H₂. This material is available free of charge via the Internet at http://pubs.acs.org