OLEFIN METATHESIS: THE EARLY DAYS
Recognizing the role of metal carbenes was key in realizing the promise of olefin metathesis
Robert H. Grubbs of California Institute of Technology and Richard R. Schrock of Massachusetts Institute of Technology immediately come to mind at the mention of olefin metathesis. However, the advent of olefin metathesis featured many industrial researchers, who, while working at major U.S. petrochemical companies, discovered the reaction. It also involved several academic and French Petroleum Institute chemists, who helped establish the role of metal carbenes in the reaction.
Olefin metathesis was first observed in the 1950s by industrial chemists. In 1956, Herbert S. Eleuterio, at DuPont's petrochemicals department, in Wilmington, Del., obtained a propylene-ethylene copolymer from a propylene feed passed over a molybdenum-on-aluminum catalyst. Analysis showed that the output gas was a mixture of propylene, ethylene, and 1-butene. And when he tried the experiment with cyclopentene, "the polymer I got looked like somebody took a pair of scissors, opened up cyclopentene, and neatly sewed it up again," he tells C&EN.
Chemists at other petrochemical companies were getting similar baffling results. Edwin F. Peters and Bernard L. Evering recorded in U.S. Patent 2,963,447, assigned in 1960 to Standard Oil Co. of Indiana, that propylene combined with molybdenum oxide on alumina treated with triisobutyl aluminum yields ethylene and butenes. In 1964, Robert L. Banks and Grant C. Bailey, of Phillips Petroleum, Bartlesville, Okla., reported the disproportionation of propylene to ethylene and butenes using molybdenum hexacarbonyl supported on alumina [Ind. Eng. Chem. Prod. Res. Dev., 3, 170 (1964)].
Like magic, the chemistry was spewing products that could not be explained by the reactions of olefins known at the time. In 1967, Nissim Calderon and others at Goodyear Tire & Rubber, Akron, Ohio, figured out what was going on.
The unexpected products are due to cleavage and reformation of the olefins' double bonds. One carbon of the double bond of one olefin, along with everything attached to it, exchanges place with one carbon of the double bond of the other olefin, along with everything attached to it. The Goodyear researchers named the reaction "olefin metathesis" [Tetrahedron Lett., No. 34, 3327 (1967).]
The Goodyear team performed experiments with butene and deuterated 2-butene in the presence of a homogeneous catalyst [J. Am. Chem. Soc., 90, 4133 (1968)]. At about the same time, Johannes C. Mol and others at the University of Amsterdam, in the Netherlands, independently reached the same conclusion with propylene and carbon-14-labeled propylene in the presence of a heterogeneous catalyst [Chem. Commun., 1968, 633].
THE RACE TO EXPLAIN the reaction was on. According to a once-popular hypothesis, which was favored by Calderon and is sometimes referred to as the conventional mechanism, a cyclobutane intermediate complexed to the metal is formed [J. Am. Chem. Soc., 90, 4133 (1968)]. However, olefin metathesis forms no cyclobutanes, and putting cyclobutanes into olefin metathesis systems does not produce alkenes. Thus in 1971, Roland Pettit, then a chemistry professor at the University of Texas, Austin, proposed a tetramethylene complex, in which four methylene units are bonded to a central metal atom [J. Am. Chem. Soc., 93, 7087 (1971)].
In the late 1960s, Grubbs was a postdoc at Stanford University, where his interests in organometallic chemistry, catalysis, and reaction mechanisms converged. He first heard of olefin metathesis at a seminar there. "The transformation is unique, and we had no idea how it happens," he tells C&EN.
In an attempt to explain olefin metathesis, Grubbs--who by then had moved to Michigan State University, East Lansing--proposed that the redistribution of groups around the double bonds was due to a rearranging metallacyclopentane intermediate [J. Am. Chem. Soc., 94, 2538 (1972)]. Later, he suggested that one mode of rearrangement could lead to formation of a cyclobutane complexed to a metal carbene [Inorg. Chem., 12, 2166 (1973)].
These proposals were incorrect. Had people been aware of work by French chemists published in the early 1970s, some of these ideas might not have come up.
In 1971, two chemists at the French Petroleum Institute, Yves Chauvin and his student Jean-Louis Hérisson, suggested that olefin metathesis is initiated by a metal carbene. The metal carbene, they proposed, reacts with an olefin to form a metallacyclobutane intermediate that breaks apart to form a new olefin and a new metal carbene, which propagates the reaction [Makromol. Chem., 141, 161 (1971); because of a typographical error in the journal's running head, this paper is sometimes erroneously cited with a 1970 publication date].
This paper, everyone agrees, was the first to envision correctly a key role for metal carbenes in olefin metathesis and the events that lead to exchange of groups around carbon-carbon double bonds. Chauvin's work gave the field "a chance to move away from its state of alchemy, although it took several years before the mechanism was experimentally supported and widely accepted," says K. C. Nicolaou, a chemistry professor at Scripps Research Institute and the University of California, San Diego.
Chauvin, now retired, says three papers published in 1964 led him to the hypothesis. The first was from Ernst Otto Fischer at the University of Munich, Germany, about a new type of metal-carbon bond exemplified in the metal carbene (methylmethoxycarbene)pentacarbonyl tungsten--(CO)5W=C(CH3)(OCH3) [Angew. Chem. Int. Ed., 3, 580 (1964)]. The second paper was from Giulio Natta at the Industrial Chemistry Research Institute at Milan Polytechnic, Italy, describing the ring-opening polymerization of cyclopentene with triethylaluminum and hexachlorotungsten [Angew. Chem. Int. Ed., 3, 723 (1964)]. And the third was from Phillip Petroleum's Banks and Bailey on the disproportionation of propylene.
Like magic, the chemistry was spewing products that could not be explained by the reactions of olefins known at the time.
"APPARENTLY, these papers had nothing in common," Chauvin tells C&EN. "But for me, they were a revelation." The papers of Natta, Banks, and Bailey indicated that cyclopentene polymerization and propylene disproportionation are the same reaction, he explains. Therefore, they must involve the same type of intermediate species. "With the paper of Fischer, I felt that these species could be metal carbenes," Chauvin continues.
With Hérisson, Chauvin studied the coreactions of acyclic and cyclic olefins. They found that the main products of cyclopentene and 2-pentene were C9, C10, and C11 dienes in a ratio of 1:2:1. That three products are formed is key, because the conventional mechanism--the most popular hypothesis in the late 1960s--predicts only the C10 product. The results show complete exchange between the olefin starting materials, which is possible with a metal carbene intermediate but not with a cyclobutane intermediate.
The reaction of cyclooctene and 1-pentene, however, gave almost entirely the C13 product predicted by the conventional mechanism. The French authors stated that their scheme for olefin metathesis--involving a reaction between metal carbenes and olefins--could not explain the result.
For several years after the Chauvin paper was published, consensus regarding the
involvement of metal carbenes in olefin metathesis did not emerge. For example, Mol concluded in 1975 that the available data supporting the Chauvin hypothesis did not exclude other mechanisms. And in 1977, Calderon wrote, "The present state of knowledge does not permit a clear-cut selection of a single scheme [among four proposed, including Chauvin's] over the rest."
However, suggestions that metal carbenes are key were coming from various labs. For example, in 1972, Michael F. Lappert and coworkers at the University of Sussex, Brighton, England, showed that various rhodium complexes catalyze the metathesis of tetraaminoethylenes, which are highly electron rich olefins, and that rhodium carbenes are formed during the reaction [Chem. Commun., 1972, 927]. But the relevance of this reaction to ordinary alkenes was "an open question" at the time [Chem. Soc. Rev., 2, 99 (1973)].
Meanwhile, two years after proposing a role for metal carbenes, Chauvin showed that a small amount of propylene is formed from 2-butene in the presence of tungsten hexachloride and methyllithium or tetramethyltin [C. R. Acad. Sci. Paris, 276, 169 (1973)]. The product, Chauvin tells C&EN, could be explained by replacement of a ligand on tungsten with a methyl group and elimination of a methyl hydrogen to form a W=CH2 species, which reacts with 2-butene.
Perhaps the most significant clue came from the University of Wisconsin, Madison, in 1974. Charles P. Casey and Terry J. Burkhardt discovered that the metal carbene (diphenylcarbene)pentacarbonyltungsten reacts with isobutene to form as the major product a new olefin, 1,1-diphenylethene [J. Am. Chem. Soc., 96, 7808 (1974)].
In a separate experiment with 1-methoxy-1-phenylethylene, they determined that one fragment of this alkene goes into a new olefin while the other fragment is incorporated into a new metal carbene, (phenylmethoxycarbene)pentacarbonyl tungsten. It was the first demonstration that both fragments of a metal carbene exchange with both fragments of an olefin's double bond.
An exchange between a metal carbene and an olefin is now known to be a fundamental step in olefin metathesis. Casey and Burkhardt were the first to show that such a reaction occurs. However, their work did not address what happens between olefins.
"We did not know that we were doing olefin metathesis when we started," Casey tells C&EN. "At the time, we were more interested in using the metal carbene to make cyclopropanes. When we submitted the paper, we thought we were proposing a new mechanism for olefin metathesis that involved metal carbenes. One of the referees pointed out that such a mechanism had been proposed before but without supporting chemical models. What we had was the first model reaction."
From the publication of Chauvin's paper in 1971 until early 1975, only two papers cited it for its mechanistic value: one by the group of B. A. Dolgoplosk in Moscow [Dokl. Akad. Nauk. SSSR, 202, 871 (1972)] and the other by Casey and Burkhardt. The latter wrote that the "equilibrium between a metallocyclobutane and a metal complex containing both an alkene and a carbene ligand provides a sufficient mechanism for olefin metathesis. Such a scheme has been previously proposed by Chauvin."
That the results that perplexed Hérisson and Chauvin do not contradict a metal carbene mechanism became clear from a communication by Thomas J. Katz, a chemistry professor at Columbia University, and his graduate student James McGinnis [J. Am. Chem. Soc., 97, 1592 (1975)]. The paper analyzes the kinetics of olefin metathesis and addresses whether the reaction of a cyclic alkene with an acyclic alkene should give three products or just one.
PHOTO BY MAUREEN ROUHI
The answer is: It depends on the substituents of the acyclic olefin. "What Chauvin did not recognize is that, when a metal carbene reacts with an olefin, two metal carbenes can result and the more stable one will be formed in larger amount," Katz explains.
"Katz recognized that just because you have a highly active metal carbene does not mean that it's not governed by the regular laws of organic chemistry," notes Samuel J. Danishefsky, a chemistry professor at Columbia and the director of the Laboratory for Bioorganic Chemistry at Memorial Sloan-Kettering Cancer Center, New York City.
If the groups around the double bond of the acyclic olefin are sufficiently different, one metal carbene product will be favored over the other, Katz explains. Only one product will be formed, the same one predicted by the conventional mechanism. If the groups are not sufficiently different, olefin metathesis gives three products: that predicted by the conventional mechanism and two cross-products.
The 1975 paper was the first from Katz's lab on olefin metathesis. Katz says he had been following the literature and thought that none of the proposed mechanisms was right. "I could see how it worked, but trying to prove it was a nightmare. It took years. We started in the late 1960s," he says.
Like many of his contemporaries, Katz initially was unaware of Chauvin's work. He says he didn't find Chauvin's paper until he was ready to write the manuscript for his own paper on the mechanism. Katz and McGinnis cite Hérisson and Chauvin, crediting the French chemists for recognizing the significance of cross-products--three products instead of one--in olefin metathesis. And in a footnote, Katz and McGinnis point out that Hérisson and Chauvin "were the first to publish the carbene mechanism, but they could not easily reconcile" the results for cyclooctene and 1-pentene.
BY 1974, preparing and characterizing metal carbenes had become a major activity in organometallic chemistry. Noting the relative stability and ease of preparing metal carbenes, as well as the Casey-Burkhardt demonstration of the exchange between an alkene and a tungsten carbene, Grubbs also proposed a metal carbene mechanism [J. Am. Chem. Soc., 97, 3265 (1975)], published 11 weeks after Katz's. Instead of using differently substituted olefins as Chauvin and Katz did, Grubbs used isotopically labeled olefins to track the exchange of groups. The pattern of exchange, he and coworkers Patrick L. Burk and Dale D. Carr concluded, supports a metal carbene mechanism. A note added in proof cites Katz's paper.
The work of Katz in 1975 was the first to unambiguously substantiate the carbene mechanism for the olefin metathesis reaction, says Nicolaou, who once was a postdoc in the Katz lab. "In particular, Katz proposed that initiators for olefin metathesis reactions be sought among isolable metal-carbene complexes and used such discrete compounds as initiators of olefin metathesis."
The last sentence of Katz's paper reads, "The mechanism suggests that initiators for olefin metathesis be sought by synthesizing simple alkyl-substituted metal-carbene species and four-membered rings."
"This prediction and the work that followed from it represent a significant step forward in the early development of this important reaction," says James L. Leighton, an associate professor of chemistry at Columbia. Up to that point, researchers--including Chauvin, Grubbs, and Katz--had been initiating olefin metathesis reactions with ill-defined mixtures. "If the metal carbene mechanism is correct, you might start from a metal carbene species," Leighton says.
Katz's paper makes two other predictions: a cycloalkene that is unsymmetrically substituted at the double bond will yield regularly substituted polymers, and metal carbynes will initiate metathesis of acetylenes.
Within a year, Katz had proven the first two predictions. With McGinnis and undergraduate student Samuel Hurwitz, he reported the first use of an isolable metal-carbene complex--(diphenylcarbene)pentacarbonyltungsten--to initiate metathesis of unsymmetrically substituted ethylenes [J. Am. Chem. Soc., 98, 605 (1976)]. With Steven J. Lee and Nancy Acton, he showed that the metal carbene initiates polymerization of various cyclic olefins [Tetrahedron Lett., No. 47, 4247 (1976)]. And with Acton, he showed that another carbene, (methoxyphenylcarbene)pentacarbonyltungsten, also initiates polymerization of cyclobutene and norbornene [Tetrahedron Lett., No. 47, 4251 (1976)].
AS DANISHEFSKY puts it, Katz "actually put in a carbene--made by someone else to be sure, but it was a well-characterized carbene--and showed that it gave metathesis. No one can deny that. And then he proposed that this is the right way to do olefin metathesis."
In other publications, Katz and coworkers proved the second prediction. They showed that (diphenylcarbene)pentacarbonyltungsten forms perfectly alternating polymers from 1-methylcyclobutene and 1-methyl-trans-cyclooctene [J. Am. Chem. Soc., 98, 606 and 7818 (1976)].
"To my knowledge, these papers are the first examples of anyone using isolable metal carbenes to initiate metathesis," Leighton says.
Five years later, Schrock proved the third prediction. With Jeffrey H. Wengrovius and José Sancho, he prepared neopentylidyne complexes with tungsten and showed that these metal carbynes react with various acetylenes to give the expected metathesis products [J. Am. Chem. Soc., 103, 3932 (1981)].
Just as metal carbenes polymerize cyclic alkenes, so must they polymerize acetylene, which might be considered a cycloalkene of ring size two. Such a route to polyacetylenes was first proposed by Toshio Masuda and coworkers at Kyoto University, Japan, who showed that phenylacetylene is polymerized by tungsten hexachloride or molybdenum pentachloride [Macromolecules, 8, 717 (1975)]. Using Casey's carbene, Katz and Lee demonstrated that isolable metal carbenes initiate acetylene polymerizations, giving particularly clean polymers [J. Am. Chem. Soc., 102, 422 (1980)].
Metal-carbene-propagated acetylene polymerization has an interesting corollary. When a molecule that has both a carbon-carbon triple bond and a carbon-carbon double bond is exposed to a metathesis initiator, it should give a diene. With Timothy M. Sivavec, Katz first demonstrated this rearrangement, also called the ene-yne metathesis reaction [J. Am. Chem. Soc., 107, 737 (1985.)]. It is based on an earlier insight by Katz that polymerizing acetylenes should induce the metathesis of olefins [J. Am. Chem. Soc., 102, 7940 (1980)].
Katz, however, is not widely associated with olefin metathesis. "Clearly, the great advancement of the field and acceptance of the method in chemical synthesis can be attributed to the enormous body of cutting-edge work by Grubbs and Schrock," says Jay S. Siegel, a chemistry professor at the University of California, San Diego. "However, what Katz did was seminal. He was the guy who first pointed out how to get over the hill."
"What Katz did was seminal. He was the guy who first pointed out how to get over the hill."
KATZ WITHDREW from olefin metathesis research in the late 1980s. "I honestly don't know," he replies when asked why he left the field. He admits to being a research vagabond, working in an area for a while, solving some interesting problems, and then moving on. That there is no such thing as a "Katz catalyst" for olefin metathesis may be the reason that he is not linked to the reaction.
But several sources also have told C&EN that Katz's contributions are not consistently cited in related publications. One example that has been pointed out is a book chapter by Schrock, published in 1998. Titled "Olefin Metathesis by Well-Defined Complexes of Molybdenum and Tungsten," the chapter makes no mention of Katz's 1976 work with tungsten carbenes.
Experimental support for the metal carbene mechanism came from different groups, including those of Grubbs and Schrock. Schrock says he does not cite Katz's work extensively because "Katz was working with at least a gray box if not a black box. He never showed what type of compound was actually responsible for the metathesis reaction."
Regarding Katz's proposal to use metal carbenes as initiators of olefin metathesis, Schrock says, "My paper on an alkylcarbene complex of tantalum [J. Am. Chem. Soc., 96, 6796 (1974)] had already appeared in plenty of time for [Katz] to see, but he did not reference it. I remember seeing his 1975 publication and noting that he ignored my contribution and that what he proposed had already happened."
Schrock recalls becoming interested in olefin metathesis because the question of its mechanism was heating up. He joined the staff of DuPont's Central Research Department as a research chemist in 1972, soon after completing a postdoctoral stint at Cambridge University.
At DuPont, Schrock began exploring the organometallic chemistry of tantalum. Within two years, he had planted what he says was the seed that led to the flowering of olefin metathesis--the preparation of a tantalum carbene, the subject of the 1974 paper that Schrock says Katz ignored.
This tantalum carbene is different from the so-called Fischer metal carbenes, such as the tungsten carbenes Katz used. In Fischer metal carbenes, the metal is in a low oxidation state. In the tantalum carbene and in other so-called Schrock carbenes, the metal is in the highest possible oxidation state. Emphasizing this distinction, Schrock consistently refers to high-oxidation-state metal carbenes as metal alkylidenes, although he agrees that they are also correctly referred to as metal carbenes.
Schrock's 1974 paper details the preparation of the tantalum carbene, its characterization, and the mechanism of its formation. But it does not say that this metal carbene, or any metal carbene, is relevant to olefin metathesis.
By 1975, Schrock had moved to MIT, where he spent another five years developing the tantalum chemistry to the point that it could be applied to metals more classically associated with olefin metathesis. Thus, he proceeded to make high-oxidation-state tungsten carbenes using the principles gleaned from tantalum chemistry. The tungsten carbenes in turn led him to the highly active and now-renowned molybdenum carbenes.
During those five years, he showed that the tantalum compounds form metallacyclobutanes. The first compounds, however, were not productive in olefin metathesis; the metallacyclobutanes rearranged before they could form a new olefin. Schrock's lab prepared and characterized many new tantalum, as well as niobium, complexes and tested them in olefin metathesis.
FINALLY IN 1980, Schrock reported that his lab achieved metathesis of cis-2-pentene with complexes of the type [P(CH3)3](O-tert-C4H9)2(Cl)M5C(H)(tert-C4H9) [M = Nb or Ta]. The paper shows results for the niobium complex only [J. Mol. Catal., 8, 73 (1980)]. In later publications, Schrock says that this work represents "the first time that productive metathesis of a simple olefin starting with a well-characterized carbenoid complex had been observed" [for example, J. Organometal. Chem., 300, 249 (1986)].
Yet Katz had previously demonstrated olefin metathesis with (diphenylcarbene)-pentacarbonyltungsten, a Fischer-type carbene that Casey and Burkhardt prepared and characterized in 1973 [J. Am. Chem. Soc., 95, 5833 (1973)] and for which a full crystal structure was published four years later [J. Am. Chem. Soc., 99, 2127 (1977)].
Schrock tells C&EN that the metal in that carbene is not in a high oxidation state. Katz's work, he explains, "proved only what we knew already, that tungsten is an active metal for metathesis. It did not prove that the metal carbene is responsible for the reaction. It is possible that a very small amount of high-oxidation-state tungsten carbene species is formed under the reaction conditions described by Katz."
Grubbs agrees in part: "If you put a metal complex in, you heat it up, and a reaction happens, you have no idea what's going on. It's not what you put in. What's important is showing that what you put in actually did the chemistry."
Grubbs's own winning catalysts are not of the high-oxidation type. Grubbs wrote about his first ruthenium catalyst [J. Am. Chem. Soc., 114, 3974 (1992)]: "Typically, high oxidation state metallaolefins are called alkylidene complexes while low oxidation state analogues are referred to as carbene complexes. The new complex described here does not show all of the characteristics of either of these two classes of complexes."
The distinction between high- and low-oxidation-state carbenes is misleading, Grubbs says. "I've tried to contend that there is a continuum of reactivity. The Fischer systems tend to be electron poor at the carbon center while the Schrock systems tend to be electron poor at the metal center, and they give different chemistries. I don't know where the ruthenium catalysts fit in between, but they work. If we had stuck with those oxidation-state classifications, we would have been out of luck."
Grubbs and Schrock deserve accolades for olefin metathesis, but in the early days, others helped to unlock the key to the reaction. Chauvin proposed the involvement of metal carbenes. Katz suggested preparing metal carbenes and using them to initiate olefin metathesis and then showed that isolable metal carbenes initiate olefin metathesis. Grubbs and Schrock indisputably led the efforts to prepare the catalysts that have allowed the promise of olefin metathesis to be realized. The rest, as the saying goes, is history.