Cyclic Alkyl Amino Carbene (CAAC) Ruthenium Complexes as Remarkably Active Catalysts for Ethenolysis

Abstract

An expanded family of ruthenium metathesis catalysts bearing cyclic alkyl amino carbene (CAAC) ligands is reported. These catalysts exhibited exceptional activity in the ethenolysis of the seed oil derivative methyl oleate. In many cases, TONs >100,000 were achieved, at only 3 ppm catalyst loading. Remarkably, the most active catalyst system was able to achieve a TON of 340,000, at only 1 ppm catalyst loading. This is the first time a series of metathesis catalysts has exhibited such high performance in cross metathesis reactions employing ethylene gas, with activities sufficient to render ethenolysis applicable towards the industrial scale production of linear alpha-olefins (LAOs) and other terminal olefin products.

Graphical Abstract

A new series of cyclic alkyl amino carbene (CAAC)-containing olefin metathesis cataysts was synthesized, and were found to exhibit unprecedented activity in the ethenolysis of methyl oleate. This work advances the state-of-the-art of the ethenolysis reaction, and is expected to find particular use in large scale industrial applications.

The transformation of small molecule chemical feedstocks to high-value chemicals has been a long-standing challenge that has received a significant resurgence of interest in the chemical sciences. This is a result of recently introduced programs promoting the use of greener chemistry practices, as well as the rising costs associated with the production of fine chemicals from petrochemicals. Consequently, the ability to access high-demand products from renewable sources such as oleochemicals presents a cost-effective and environmentally friendly alternative.^[1]

Olefin metathesis reactions, such as cross-metathesis (CM), ring-closing metathesis (RCM), and ring-opening metathesis polymerization (ROMP), all of which generate a new internal olefin, have enjoyed widespread popularity in both academic and industrial settings as a result of their general applicability, ease of use, and Non-prohibitive costs.^[2] Ruthenium-based metathesis catalysts are ideal for such transformations as a result of their generally robust nature, which enables handling in air, and imparts good tolerance to a variety of functional groups and trace impurities. All of these are necessary prerequisites when subjected to raw materials or biomass.

Many renewable or bio-based materials, such as fatty acids originating from seed oils and their derivatives, contain at least one unit of unsaturation, providing a synthetic handle for derivatization by olefin metathesis catalysts. The CM reaction with ethylene (2), commonly referred to as ethenolysis, has significant potential as a clean, scalable, and sustainable solution for the production of linear alpha olefins (LAOs) (eg. 3 and 4) from the natural oils found in oleochemicals such as methyl oleate (MO, 1) (Scheme 1). LAOs are direct precursors to a variety of commodity chemicals with applications as fuels, surfactants, lubricants, waxes, perfumes, antimicrobial agents, and thermoplastics. In addition, LAOs can be rapidly elaborated to more expensive products such as agrochemicals, insect pheromones, and pharmaceuticals.^[3]

Scheme 1.

Ethenolysis of the seed oil derivative methyl oleate (1).

The production of terminal olefins from the ethenolysis of seed oil derivatives using metathesis catalysts has been previously demonstrated. However, the high catalyst loadings required (10 – 100 ppm) to achieve an acceptable yield of terminal olefins render these reported procedures cost prohibitive on an industrial scale.^[4],[5],[6] In general, catalyst turnover numbers (TONs) of at least 35,000 and 50,000 are recommended in the manufacturing of specialty and commodity chemicals, respectively.^[4] In the ethenolysis of the benchmark substrate MO (1), standard ruthenium-based metathesis catalysts such as 5 – 8 afforded TONs of only 2,000 – 5,000. This stands in contrast to the extremely high activity normally exhibited by these catalysts in CM with terminal or internal olefins. For example, TONs as high as 470,000 have been achieved with 8 in the CM of MO and 2-butene.^[7] The most active catalyst for ethenolysis in the literature to date is cyclic alkyl amino carbene (CAAC) complex 10 (Scheme 2), which has been previously reported to generate a TON of 35,000 in the CM of MO with ethylene.^[4],[5] As a result of the lack of a catalyst sufficiently active to produce teminal olefins using ethylene gas, industrial scale ethenolysis is currently accomplished using higher olefins as ethylene surrogates.^[6b] Catalyst 7, for example, is able to achieve the in situ ethenolysis of MO with a TON as high as 192,900 with propylene gas.^[4a] However, there is a need to develop catalysts capable of achieving high activity in ethenolysis reactions when ethylene gas is utilized directly. Whereas CM with higher olefins necessarily results in a substantial amount of undesired internal olefins being produced as byproducts, the only products derived from CM with ethylene are terminal olefins. This intrinsic advantage promotes both increased yield and ease of purification of the desired terminal olefin products, and is a particularly important consideration for bio-refinery feedstocks in which multiple downstream products are produced.^[3]

Scheme 2.

Synthesis of CAAC complexes under study (isolated yield in brackets).

Herein, we report the discovery of the most active ethenolysis catalysts to date. In many cases, TONs surpassing 100,000 were achieved for the ethenolysis of MO, using ethylene gas. Remarkably, certain catalyst systems even exhibited TONs approaching 200,000, with the highest TON achieved being 330,000. This represents the first time that reaction conditions have been developed in order for ethenolysis to proceed efficiently on an industrial scale.

Despite the promising results previously exhibited by 10 in the ethenolysis of MO,^[4],[5] CAAC ligated ruthenium complexes have yet to be investigated in detail. In particular, it was envisioned that more in-depth structure/activity relationship (SAR) studies would facilitate the development of new, more efficient catalysts. Thus, a variety of new catalysts were prepared through modifications of exisiting literature procedures (Scheme 2).^{[5],[ 8 ]} Known CAAC catalysts (10, 13, 18, 25) were screened alongside the new catalysts, in order to ensure accurate SAR comparisons within the series.

Initially, derivatives of the previous benchmark catalyst 10 were targeted, in which only the ortho substituents of the N-aryl ring were varied (9 – 13). It is worthy to note that the syntheses of 9 and 10 are low yielding (ca. 20%), and purification is cumbersome. Furthermore, complexes bearing even smaller ortho substitution such as N-mesityl or N-2-isopropylphenyl, were unable to be accessed in even small amounts. In contrast, 11 – 13 can be produced in high yield (78 – 86%) and isolated without difficulty. This is hypothesized to be related to the stability of the free carbene intermediate that is generated in situ, which might otherwise be expected to decompose rapidly in the absence of steric protection. It was envisioned that this decomposition pathway could be circumvented through the installment of larger substituents at either R⁴ or R⁵. Indeed, this strategy proved to be successful, and we were able to readily access a variety of new complexes in moderate to high yield (29 – 82%). Several of the new backbone containing catalysts included N-aryl substitution that was previously inaccessible (14, 15, 17, 21), meanwhile others (16, 18 – 20, 22 – 25) were synthesized in order to provide a more thorough SAR study. Single-crystal X-ray diffraction of 11 and 24 revealed distorted square-pyramidal geometries, and structural parameters, including bond lengths and angles, were consistent with those found previously for 10 and 13 (Figure 2). Moreover, catalyst 24 exhibits a CAAC ligand featuring a chirogenic center as well as two different ortho N-aryl substituents. Accordingly, the single crystal of 24 revealed both N-aryl rotamers (24a and 24b), in a ratio of 64:36 respectively.⁹

Figure 2.

Solid-state structures of 11 and 24 (24 crystallized as 24a and 24b, in a ratio of 64:36), with thermal ellipsoids drawn at 50% probability.¹⁴ For clarity, hydrogen atoms have been omitted. Selected bond lengths (Å) for 11: C1-Ru 1.928(6), C19-Ru 1.824(6), O1-Ru 2.301(4), Cl1-Ru 2.3374(16), Cl2-Ru 2.3165(17); for 24a: C1-Ru 1.940(7), C24-Ru 1.836(9), O1-Ru 2.332(8), Cl1-Ru 2.3356(18), Cl2-Ru 2.3271(13); for 24b: C1-Ru 1.931(12), C24-Ru 1.828(18), O1-Ru 2.325(15), Cl1- 2.335(4), Cl2-Ru 2.307(3).

Once in hand, catalysts 9 – 25 were examined in the ethenolysis of 1 using ethylene (Table 1). Reaction conditions were adapted from those that were previously reported in the literature, and were initially re-optimized using benchmark catalyst 10 (neat MO, 40 °C, 150 psi ethylene).^[4] The only deviation from the published procedures was the use of higher purity ethylene (99.95%) than previously reported (99.9%). We were pleased to find that this simple modification appeared to already result in a substantial increase in activity: 10 ppm loading of catalyst 10 resulted in a TON of 67,000, whereas the benchmark value for 10 published in the literature is a TON 35,000.^{[4b],[10].11[12],[13]} We were delighted to find that the TON of catalyst 10 further increased to 120,000 upon reduction of the catalyst loading to 3 ppm. Thus, all subsequent reactions were run at 3 ppm catalyst loading, which was also expected to provide greater differentiation in activity between promising catalysts than at 10 ppm.

Table 1.

Ethenolysis of methyl oleate (1) using catalysts 9 – 25.^a

catalyst	conversion (%)b,c	selectivity (%)b,d,e	yield (%)f	TONg
9	37	86	32	110,000
10	42	88	37	120,000
11	59	92	54	180,000
12	18	94	17	57,000
13	19	97	18	60,000
14	22	63	14	47,000
15	26	86	22	73,000
16	42	92	39	130,000
17	19	78	14	47,000
18	13	97	13	43,000
19	16	97	15	50,000
20	<5%	---	---	---
21	41	83	34	110,000
22	46	85	39	130,000
23	48	88	43	140,000
24	57	94	54	180,000
25	47	98	46	150,000

^aReaction conditions: catalyst (3 ppm), C₂H₄ (150 psi, 99.95% purity), 40 °C, 3 h.

^bDetermined via GC, using dodecane as an internal standard.

^cConversion = 100 – [(final moles 1) X 100/[initial moles 1)]

^dSelectivity for ethenolysis products (3 and 4) over self-metathesis products (3a and 4a).

^eSelectivity = 100 X (moles 3 + 4)/[(moles 3 + 4) + (2 X moles 3a + 4a)].

^fYield = Conversion X Selectivity/100.

^gTON = Yield X (initial moles 1/moles catalyst)/100.

Remarkably, at 3 ppm catalyst loading, most catalysts surpassed a TON of 100,000! Specifically, catalysts 11 and 24 emerged as the most efficient, with TONs of 180,000. Catalyst activity correlates with N-aryl substitution, with larger substituents at R¹ and R² generally resulting in higher TONs, although this can also have a deleterious effect (as in 12 and 13, compared to 9 – 11). The ideal combination thus far appears to be when R¹ is small and R² is large, as in catalysts 11 and 24 (R¹ = Me, R² = iPr). Interestingly, substitution at R⁵ by a phenyl ring resulted in an overall improvement of activity, especially for N-2,6-diisopropylphenyl catalyst 25. Replacement of R⁴ and/or R⁵ by ethyl, propyl, or cyclohexyl did not result in a significant change in the TON (as in 15 – 19), although an adamantyl substituent (20) resulted in complete loss of activity. Interestingly, while consumption of MO appears to be the most important determining factor in the overall yield of ethenolysis products, increased N-aryl substitution on the catalyst appears to strongly favour selectivity for terminal olefins. These trends in selectivity and TON are most evident in the series of catalysts bearing a phenyl ring on the backbone at R⁵ (21 – 25).

A plausible explanation for the high activity exhibited by CAAC catalysts in ethenolysis transformations might be a result of increased stabilization of the ruthenium methylidene intermediate generated in the presence of ethylene gas. Ruthenium methylidenes are known to decompose rapidly via insertion of the N-aryl substituent into the methylidene carbene, which subsequently generates various ruthenium hydrides that are inactive in metathesis transformations.^[15] CAAC ligands are known to be more electron donating than their N-heterocyclic carbene (NHC) counterparts.^[16] Thus, when used in place of NHCs, it is expected that the increased electron density at ruthenium might somewhat stabilize the otherwise highly reactive and electron deficient methylidene intermediate.^{[15b,c], [17]} Substitution of the ortho N-aryl substituents with a larger sterically encumbered group would also be expected to significantly decrease the rate of termination by insertion into the [Ru]=CH₂ bond. However, increased substitution can also hinder coordination of olefins to the ruthenium metal center. Diminished reactivity with increasing N-aryl substitution was indeed noted when initiation rates of selected catalysts (9 – 14) were measured following exposure to n-butylvinylether.^[18] When initiation rates are compared to TONs, it is clear that both the slowest initiating catalysts (12, 13) and the fastest initiating catalyst (14) exhibit the lowest TONs (Figure 3). This is likely a result of diminished catalytic rate for the former group and an increased susceptibility to decomposition for the latter. This study illustrates the importance of this delicate balance, as reflected in the superior TONs exhibited by catalysts 11 and 24, possessing asymmetric N-aryl substituents. In these systems, the smaller substituent (R¹ = Me) faciltates rapid coordination of the incoming olefin substrate, whereas the larger substituent (R² = iPr) prevents decomposition of the methylidene intermediate. This asymmetry might be expected to exhibit a greater effect in CAAC ligated catalysts, as the steric interaction of the ortho substituents on the N-aryl ring with the two adjacent geminal methyl substituents would be expected to influence the conformation of the N-aryl ring. If the larger substituent resides closer to the ruthenium metal center, the methyl group which is inherently more susceptible to CH insertion would be directed away from the reactive ruthenium methylidene.

Figure 3.

Comparison of initiation rates and TON for catalysts 9 – 14.

It has been postulated that high activity in ethenolysis might be correlated to the tendency of a catalyst to undergo degenerative metathesis events, through the preferential formation of 2,4-metallacycles rather then 2,3-metallacycles, which would result in increased selectivity for terminal olefins in the product distribution.^[19] This is a powerful design principle in the context of achieving high kinetic selectivity for terminal olefins, when employing higher olefins such as propene and 1-butene gas as ethylene surrogates. However, when ethylene gas is employed, it is more likely that the lower selectivities for terminal olefins exhibited by previous generations of catalysts (5 – 8) is primarily a result of rapid catalyst death in the presence of ethylene. This would translate to products reflecting the kinetic distribution of rapid unselective cross metathesis reactions of 5 – 8 with both ethylene and terminal olefins.^[20] A lthough CAAC-ligated ruthenium-based catalysts have been demonstrated to engage in metathesis reactions more slowly than phosphine or NHC-ligated catalysts,^[5a],[18] they also appear to persist for a much longer time in the presence of ethylene.^[21] This would allow for the ethenolysis reaction to proceed closer to completion in order to achieve the equilibrium ratio of terminal olefin products, and provides a feasible explanation for the notable increase in activity exhibited by this family of catalysts.^[22]

Finally, given the notable dependance on the TON with respect to the purity of the ethylene employed, as seen earlier for catalyst 10, we briefly explored the effect of utilizing an even higher purity ethylene source (99.995% vs. 99.95%) at different loadings of catalyst 11 (Table 2).²³ A dramatic increase in TON was noted at 1 ppm catalyst loading. To the best of our knowledge, this represents the highest value available in the literature for any ethenolysis catalyst to date (TON 340,000).

Table 2.

Effect of C₂H₄ (2) source on activity of catalyst 11.^a

Loading	purity of 2	yield (%)	TON
3	99.95%	54	180,000
3	99.995%	53	180,000
2	99.95%	48	240,000
2	99.995%	49	245,000
1	99.95%	13	130,000
1	99.995%	34	340,000

^aReaction conditions, and calculations, are as listed in Table 1.

In summary, a new series of ruthenium metathesis catalysts, bearing CAAC ligands, is presented that displays exceptional activity in ethenolysis reactions. In the cross metathesis reaction of the seed oil derivative methyl oleate (1) and ethylene gas (2), TONs >100,000 are generated in many cases, which surpasses the minimum value of 50,000 required to be considered economically sustainable on an industrial scale. Furthermore, even higher TONs (180,000 – 340,000) were obtained in some cases. These are the highest values recorded in the literature to date for an ethenolysis reaction, and the only reported TONs >50,000 using ethylene gas specifically. As a result, it is envisioned that this work will find substantial application in the continued development of new methodologies and processes directed towards the economically and environmentally sustainable production of LAOs, as well as other valuable terminal olefins, especially through the transformation of seed oils and their derivatives.

Figure 1.

Selected ruthenium metathesis catalysts previously studied for ethenolysis.