Bis(Cyclic Alkyl Amino Carbene) Ruthenium Complexes: A Versatile, Highly Efficient Tool for Olefin Metathesis.

The state-of-the-art in olefin metathesis is application of N-heterocyclic carbene (NHC)-containing ruthenium alkylidenes for the formation of internal C=C bonds and of cyclic alkyl amino carbene (CAAC)-containing ruthenium benzylidenes in the production of terminal olefins. A straightforward synthesis of bis(CAAC)Ru indenylidene complexes, which are highly effective in the formation of both terminal and internal C=C bonds at loadings as low as 1 ppm, is now reported.

Significant efforts have been made in the last two decades to develop efficient olefin metathesis (OM) processes.1 Those efforts focused on the proper selection of the substrate structure,2 purification of the starting material,3 and most importantly, on the modification of (pre)catalysts.4 The efficiency of OM catalysts still remains too low for many transformations, prohibiting wider implementation of this technology in the industry. This is particularly true for the processes in which commodity and specialty chemicals are formed, for which the turnover numbers (TONs) of at least 50 000 and 35 000, respectively, should be obtained.5 Notably, however, since the development of N‐heterocyclic carbene (NHC)‐ligated Grubbs (1),6 Hoveyda–Grubbs (2),7 and indenylidene (3)8 catalysts, the vast majority of new ruthenium‐based complexes for OM have been synthesized by modifications of these parent structures (Figure 1).4

Figure 1

NHC‐ and bis(NHC)‐ligated complexes.

Ligation of two identical [bis(NHC)Ru]9 or different [(NHC)(NHC′)Ru]10 NHCs to the ruthenium center has been significantly less explored. Some of those complexes revealed interesting features such as good efficiency in ring‐opening metathesis polymerization (ROMP) (4),11 effectiveness in the formation of tetrasubstituted double bonds (5),12 or mechanochemical activation (6).13 In general, the complexes containing two identical NHCs exhibit low activity, as exemplified by catalyst 6, whereas the ligation of two different NHCs is somewhat tedious.

The presence of NHC ligand is a common feature for basically all modern ruthenium catalysts designed for general applications. Those complexes show, in the majority of cases, a low level of degenerate (unproductive) metathesis and low stability of ruthenium methylidenes. This characteristic makes them practically useless in the industrially important cross‐metathesis of unsaturated fatty acid derivatives such as methyl oleate (MO, 7) with ethylene, commonly referred to as ethenolysis.14 Linear α‐olefins (LAOs, for example 8 and 9, Scheme 1), obtained as a result of ethenolysis, can be easily transformed to higher added‐value products.15 Benzylidene ruthenium complexes containing the CAAC ligand16 (for example, 12) are the only catalysts that can reach the required level of TON in the ethenolysis of MO.17 Of note, even at a catalyst loading as low as 3 ppm, the ethenolysis of 1 million metric tons of oils would require a few metric tons of catalyst. Therefore, a safe and economically viable method of catalyst synthesis is crucial for industrialization.

Scheme 1

A) Ethenolysis of methyl oleate; B) state‐of‐the‐art method for the synthesis of (CAAC)Ru benzylidenes.

It is important to note that until now, the (CAAC)Ru benzylidene complexes have only been obtained from Hoveyda first‐generation catalyst 11,18 the preparation of which is burdened with significant safety and/or processing issues. The reported methods of synthesis of 11 require 1) application of diazocompound at −78 °C,19 2) conducting the Wittig or Stille coupling reaction to synthesize 2‐isopropoxystyrene,20 or 3) using as much as 9 equiv of 2‐isopropoxypropenyl benzene (13).21

No reports on the practical use of (CAAC)Ru benzylidenes in the formation of internal C=C bonds can be found.22 This might be due to the difficult synthesis of these catalysts and lack of their commercial availability. The second fact that can discourage chemists from their general use is a relatively high degree of degenerate metathesis exhibited by the CAAC complexes which can potentially reduce productive TON.23 We hypothesize, however, that the high stability of the CAAC‐ruthenium methylidenes can prevail over degenerate metathesis and result in a very effective formation of internal C=C.

Thus, we aimed to develop one of the most versatile ruthenium based catalysts for OM, which could deliver both internal and terminal olefins at very low loadings and with high selectivity. For the above‐mentioned reasons we decided to examine the possibility of synthesizing the CAAC catalysts directly from the readily accessible and patent free first‐generation complexes bearing two phosphine ligands.

To start with, indenylidene complex 14 a 24 was reacted with CAAC generated in situ from 2 equivalents of salt 10 a 22a with the use of lithium hexamethyldisilazane (LiHMDS) (Scheme 2).

Scheme 2

Synthesis of bis(CAAC)Ru indenylidene complex.

A short time for the deprotonation step and application of LiHMDS proved to be critical for a good yield. Commonly utilized KHMDS or longer times of deprotonation in some cases completely prohibited the formation of products.25 The product was isolated as a red solid in 70 % yield. Surprisingly, the signal from the phosphorus atom was not present in 31P NMR spectra, which suggests substitution of both phosphines by CAACs. The 1H NMR spectra of product was complicated owing to the presence of rotational isomers,26 and could not provide reliable information about its structure. Nevertheless, the mass spectra, elemental analysis, and single‐crystal X‐ray diffraction analysis confirmed the formation of bis(CAAC)Ru complex 15 a. The yield of 15 a was improved to 86 % by increasing the excess of 10 a to 3 equivalents. Interestingly, even with 1.25 equiv of 10 a, complex 16 a was not observed on a TLC plate even though complex 14 a was not fully consumed. This fact suggests a strong steric repulsion between CAAC and tricyclohexylphosphine in 16 a, which facilitates dissociation of the latter and complexation of the second CAAC.27 From practical and economical point of view, complex 14 b is the most preferred Ru source.24 Conveniently, 15 a was obtained from 14 b as a sole product with 59 % yield. Analogously, the treatment of CAACs, generated by deprotonation of 3 equiv of salts 10 b–h with 14 b, provided exclusively bis(CAAC) complexes 15 b–h (Table 1). In the case of complex 15 h, the two isomers were separated by crystallization.

Table 1

Synthesis of bis(CAAC)Ru indenylidene complexes.

R	Ar	CAAC precursor	15 (yield [%])
Me		10 b	15 b (37)
Ph		10 c	15 c (37)
Me		10 d	15 d (35)
Ph		10 e	15 e (28)
Ph		10 f	15 f (60)
Me		10 g	15 g (41)
Me		10 h	15 h (39)

Crystals suitable for single‐crystal X‐ray diffraction were obtained for complexes 15 a–c, 15 e, and 15 f (selected bond lengths and angles within coordination sphere are given in the Supporting Information, Table S2). These complexes show a distorted square pyramidal geometry with the carbon atom of the indenylidene ligand in the apical position (Figure 2). A similar geometry of the coordination sphere can be found in CAAC‐benzylidenes (for example, 12 a)17a, 22a and (NHC)(NHC′)indenylidene complexes (5).12 The CAACs exhibit stronger σ‐donation abilities than the NHCs. Consequently, the Ru−Ccarbene bond in 12 a (1.9482(14) Å) is noticeably shorter than Ru−Ccarbene in complex 2 (1.981(5) Å). Surprisingly, the Ru−Ccarbene bonds in complexes 15 were found to be significanly longer than Ru−Ccarbene in 12 a and very similar to the Ru−Ccarbene distances in 5 (2.105(4) and 2.091(4) Å). For example, the distances in 15 f are Ru−C7carbene 2.093(11) and Ru−C46carbene 2.103(11) Å.

Figure 2

X‐ray crystal structures of 15 a and 15 f. Ellipsoids are set at 30 % probability; hydrogen atoms are omitted for clarity.37

To examine the possible initiation pathway of 15, we reacted 15 a and 15 f with doubly chelating olefin 13. Clean conversion of 15 f to complex 12 f was observed within 1 h in the presence of 1.2 equiv of 13 (toluene, 60 °C), and the product was isolated in 86 % yield. Under the same conditions no conversion of 15 a was observed within two days. However, in the presence of 2 equiv of CuCl, complex 12 a was formed within 30 minutes.28 This results strongly support the classical mechanism in which active, 14‐electron species are formed via neutral ligand release and also prove that for 15 a dissociation of CAAC is the rate‐limiting step. NMR kinetic studies were undertaken to gain more information about the possible mechanism of initiation of 15 f. The rate of the reaction between 15 f and olefin 13 ([D8]toluene, 60 °C) proved to be independent of the concentration of 13 (see the Supporting Information) and was in good agreement with the stability of 15 f in solution (which appears to be limited by the high lability and low stability of CAAC).25 Experimental results suggest that a dissociative mechanism operates for 15 f (Scheme 3). This, however, cannot be extrapolated to all complexes 15, and further theoretical and kinetic studies are needed to shed more light on the initiation mechanism of these (pre)catalysts.29

Scheme 3

Proposed mechanism for initiation of 15 f.

Complexes 15 a–h and highly efficient in ethenolysis reaction complex 12 g were tested in the benchmark ring closing metathesis (RCM) of diethyl diallylmalonate (DEDAM, 17, Figure 3), to compare their activity.

Figure 3

Reaction profiles for RCM of DEDAM with catalysts 12 g, 15 a–h.

The structure of the CAAC ligand proved to have a dramatic influence on the initiation rate. Complexes 15 a–d exhibited a very low (or lack of) activity, which could be expected in the case of (pre)catalysts bearing two strongly binding CAACs. These complexes can be activated by CuCl (Supporting Information). Unexpectedly, complexes 15 e–h showed a promising, moderate‐to‐high activity under these mild conditions.

For the complexes with a symmetric substitution of the N‐aryl ring (15 a–c, 15 f), the ratio of isomers in the solution can be conveniently determined by 1H NMR based on the chemical shift of the characteristic proton of the indenylidene ligand (see the Supporting Information for details).30 Crystal structures and 1H NMR spectra of 15 f and 15 c suggest that introduction of a phenyl ring into the quaternary carbon atom favors the formation of a rotamer having N‐aryl groups on the opposite side with respect to the coordination pyramid base. These complexes are more active than their close analogues which contain methyl instead of a phenyl substituent (15 a, 15 b) and which exist mainly in the rotamer having N‐aryl groups on the same side with respect to the coordination pyramid base. However, specific rotamer does not ensure high activity, as exemplified by the poor conversion of 17 obtained with 15 c. The electronic properties of CAACs in 15 a and 15 g must be similar since these ligands differ only in the position of one methyl group in the N‐aryl substituent. At the same time, a striking difference in activity of 15 a and 15 g was observed. Therefore, the most plausible explanation for the very broad range of activities covered by catalysts 15 is that the degree of steric repulsion between two CAACs determines the initiation rate.31

Next, ethenolysis of methyl oleate 7 was performed. Also in this case, 12 g was used for the comparative reasons.

Initially, reactions were run using 10 ppm catalyst loading in neat MO at 150 psi of ethylene (99.99 % purity).32 The catalysts were compared at the conditions at which they were most efficient. Catalyst 12 g provided maximum TON at 40 °C in just 2 h. The bis(CAAC)Ru indenylidene complexes 15 required a temperature between 50 and 60 °C and 4 h to reach maximum TON. Even at this increased temperature, the slow initiators, namely 15 a,b, and 15 d, did not provide significant amounts of ethenolysis products. On the other hand, the highly active complexes 15 e–g delivered products with over 50 000 TON. Under these conditions, catalyst 15 g was the most efficient (TON 60 000), slightly outperforming 12 g (TON 56 000). The most efficient initiators, 12 g and 15 f,g, were tested at 5 ppm loading (Table 2, entry 2, 9, and 11). Upon reduction of catalyst loading 12 g performed the best (TON 94 000) followed by 15 g (TON 86 000) and 15 f (TON 74 000). It was not possible to further effectively reduce the catalyst loading (drop of TON was observed for each catalyst at 3 ppm), which is most probably due to the insufficient purity of MO or lack of a glovebox.33 Because of the easy synthesis, 15 g and especially 15 f are (from an economic point of view) an interesting alternative to 12 g. For tests in the formation of internal alkenes (Scheme 4), catalyst 15 f was selected owing to 1) its good activity at low temperatures, 2) high efficiency, and 3) the inexpensive starting material (2,6‐diethylaniline) used in the synthesis of the CAAC precursor 10 f. First, the efficiencies of 15 f and state‐of‐the‐art complex 1 in RCM of highly polar proline derivative 19 (Table 3) were compared. The reaction was run in toluene at 0.25 m concentration, with the use of 60 ppm of initiators. We were pleased to see that 15 f (92 % isolated yield) significantly outperformed 1 (50 % GC yield). It was reported that productivity of active species generated from 1 is reduced by PCy3 reuptake and additionally by attack of the free PCy3 on methylidene carbon.34 In respect to that, the low stability of CAAC liberated from 15 f can potentially contribute to high efficiency of this complex. The RCM of 21 promoted by 50 ppm of 15 f delivered an azepine derivative 22, a useful building block in the synthesis of Cathepsin K inhibitors, with over 99 % yield.35 Excellent result was achieved for a highly challenging macrocyclization of 23 run at 5 mm concentration (91 % yield of 24, 250 ppm of 15 f). Compatibility of 15 f (1000 ppm catalyst loading) with sterically crowded substrates was proved by RCM of 25 which provided tetra‐substituted 26 in 79 % isolated yield.

Table 2

Ethenolysis of methyl oleate with catalysts 12 g, 15 a–h.[a]

Entry	[Ru], ppm	T [°C]	Conv. (Select)[b] [%]	Yield [%]	TON
1	12 g, 10	40	61 (91)	56	56 000
2	12 g, 5	40	51 (92)	47	94 000
3	15 a, 10	60	<1	–
4	15 b, 10	60	3	–
5	15 c, 10	60	21 (92)	19	19 000
6	15 d, 10	60	1	–
7	15 e, 10	60	57 (89)	51	51 000
8	15 f, 10	50	60 (88)	53	53 000
9	15 f, 5	50	41 (91)	37	74 000
10	15 g, 10	55	69 (87)	60	60 000
11	15 g, 5	60	48 (90)	43	86 000
12	15 h, 10	60	37 (91)	34	34 000

[a] Reactions in neat MO, 150 psi of ethylene; entries 1,2: 2 h, entries 3–12: 4 h. [b] For details regarding calculation of selectivity, yield, and TON, see the Supporting Information.

Scheme 4

Metathetic transformations accomplished with catalyst 15 f.

Table 3

Results of OM reactions.[a]

Substr.	Prod.	[Ru], ppm	T [°C]	C [m]	Conv. (select.) [%]	GC Yield [%][b]	TON
19	20	15 f, 60	60	0.25	96 (98)	(92)	15300
19	20	1, 60	60	0.25	51 (98)	50	8300
21	22	15 f, 50	55	0.25	>99 (>99)	>99 (91)	20 000
23	24	15 f, 250[c]	70	0.005	95 (96)	91[d]	3640
25	26	15 f, 1000[e]	70	0.25	90 (>99)	(79)	790
8	27	15 f, 1	60	neat	64 (98)	63 (55)	315000
8	27	1, 2	60	neat	33 (95)	31	77 500
8	27	2, 2	60	neat	16 (88)	14	35 000
9	28	15 f, 2	60	neat	69 (98)	68 (62)	170000
7	27+28	15 f, 5[c]	55	neat	45 (>99)	45	90 000
9+29	30	15 f, 200[f]	60	0.5	99 (98)	97[g] (95)	4850
31	32	15 f, 250	60	0.25	94 (>99)	(79)	3160
33 [h]	34	15 f, 1000	27	0.1	–	(87)	870

[a] In toluene, reaction time 2 h. [b] Isolated yield provided in brackets. [c] Catalyst added in 5 portions. [d] E/Z=65:35. [e] Catalyst added in 10 portions. [f] Catalyst added in 4 portions. [g] E/Z=87:13. [h] reaction in DCM with 1.5 mol % of CuCl, reaction time 10 min, PDI 1.81.

Self‐cross‐metathesis (SM) of simple, terminal alkenes is an example of a transformation which requires very high TONs to be economically viable. In our hands, SM of 1‐decene 8 proceeded up to 63 % of GC yield when run with 1 ppm of 15 f, which relates to a TON27 of 315 000.36 Moreover, a very high selectivity of 98 % was observed without any additive typically applied to inhibit the C=C bond migration. Under the same conditions, 2 ppm of catalyst 1 delivered dimer 27 with only 31 % yield and 95 % selectivity (TON27 77 500), whereas 2 ppm of initiator 2 gave 14 % of product with a poor selectivity of 88 % (TON27 35 000). Similarly, 2 ppm of 15 f delivered diester 28 with excellent TON28 of 170 000. Self‐metathesis of methyl oleate 7 proceeded up to 45 % of conversion in the presence of 5 ppm of 15 f. Almost quantitative yield in challenging cross‐metathesis of ester 9 with electron‐deficient methyl acrylate 29 was achieved with 200 ppm of 15 f. Additionally, alkene–alkyne (ene‐yne) RCM of 31 and ROMP of norbornene 33 were successfully accomplished with the use of 250 and 1000 ppm of 15 f, respectively. The obtained polynorbornene 34 was characterized by a number‐average molecular weight of 61 kg mol−1, weight‐average molecular weight of 110 kg mol−1, and polydispersity (PDI) of 1.81.

In summary, we have synthesized for the first time the bis(CAAC) ruthenium complexes. Activity of complexes 15 depends on the steric repulsion between CAAC ligands. Importantly, synthesis of 15 requires only 6 steps starting from commercially available raw materials and is not burdened with any safety or processing issues. The level of TON required for production of commodities (50 000) was significantly exceeded with complexes 15 f and 15 g in ethenolysis of methyl oleate. Testing our hypothesis we confirmed that 15 f can efficiently and selectively promote formation of internal olefins in RCM, CM, SM, ene‐yne RCM, and ROMP. A TON exceeding 300 000 was achieved in self‐metathesis of 1‐decene promoted by 15 f.

As a service to our authors and readers, this journal provides supporting information supplied by the authors. Such materials are peer reviewed and may be re‐organized for online delivery, but are not copy‐edited or typeset. Technical support issues arising from supporting information (other than missing files) should be addressed to the authors.

Supplementary

Click here for additional data file.