Codimerization Reactions

Lastly, in perfluorobutadiene s codimerization reaction with butadiene, a significant amount of Diels-Alder adduct is obtained, with the perfluorodiene acting as the diene component [125] (equation 105)... 

The codimerization reaction we will discuss in this chapter belongs to the last type. The reaction and the reactants involved will be the simplest of them all, i.e., a 1 1 codimerization of ethylene and butadiene to form C dienes. 

Catalysts for this codimerization reaction can be derived from prac-tially all the Group VIII transition metal compounds. Their catalytic properties, such as rate, efficiency, yield, selectivity, and stereoselectivity, vary from poor to amazingly good. Some better-known catalyst systems and their product distributions are listed in Table I. As one can see, the major codimerization product under the given condition is the linear 1 1 addition product, 1,4-hexadiene. The formation of this diene and its related C6 products will become the center of our discussions. The catalyst systems that have been investigated rather extensively are derived from Rh, Ni, Co, and Fe. We shall cover these systems in some detail. A lesser-known catalyst system based on Pd will also be briefly discussed. 

Before we continue I would like to mention briefly the practical reasons leading to the study of this codimerization reaction. 

In the codimerization reaction, both reactants are present in large excess compared to the catalyst concentration. The selectivity toward a 1 1 codimerization to form 1,4-hexadiene, instead of a random oligomerization, represents a rather unique reaction, especially in view of the fact that the same catalyst also dimerizes ethylene to butene (3) at about the same rate as the codimerization. The explanation forwarded by Cramer (4, 7) is based on the overwhelmingly favored stability of the tt-... 

In this codimerization reaction, the predominant complex is 3, which should lead to the ethylene-butadiene codimerization product. If the ethylene in complex 3 is displaced by butadiene to form 7 before the insertion reaction takes place, then a C8 or higher olefin could be formed... 

Hexadiene is the immediate product found in the codimerization reaction described above in a mixture of ethylene and butadiene. However, the reaction will not stop at this stage unless there is an overwhelming excess of butadiene and an adequate amount of ethylene present. As the conversion of butadiene increases, some catalyst begins to isomerize... 

Hexadiene which is formed by 1,4-addition of hydrogen and a vinyl group to butadiene, is the predominant product in the codimerization reaction. However, there is always a small amount (1-3%) of 3-methyl-... 

The essential steps in the nickel-catalyzed 1 1 codimerization reaction, which involve hydride addition to butadiene and ethylene coordination to the metal atom, were first proposed by Kealy, Miller, and Barney (35) and were later demonstrated by Tolman (40) using a model complex. Tolman prepared the complex H—Ni+L PFe [L = (EtO)3P] and showed that, after prior dissociation to form H—NiL3, it can react with butadiene to form a 7r-crotyl complex 19. 

A tetracoordinated complex (20)4 was actually isolated. Complex 20 in the presence of ethylene forms the coordinated complex 21, as can be seen from H NMR. Complex 21 is a model of the intermediate for the additional reaction to form C6 dienes. The model catalyst had been shown to be a codimerization catalyst under more severe conditions (high temperature), although the rate of reaction was very slow compared to the practical systems. These studies are extremely useful in demonstrating the basic steps of the codimerization reactions taking place on the Ni atom. The catalytic cycle based on these model complexes as visualized by Tolman is summarized in Scheme 7. A more complete scheme taking into consideration by-product formation can be found in Tolman (40). 

During the catalyst activation reactions in Kealy s catalyst system discussed above [Eqs. (6)—(7) and (8)—(10)], a monoalkylaluminum chloride is formed. Because the initial dialkylaluminum chloride R2A1C1 is present in excess, the effective cocatalyst could be the original R2A1C1 or the RA1C12 formed during the reaction. To clarify the role of the aluminum component in the actual codimerization reaction the following catalyst combinations have been tested ... 

The Influence of Phosphine Ligands on the Codimerization Reaction with Ni[COD]2 and i-BuA Cl2... 

Both the Co and the Fe systems have very similar chemistry for the 1 1 codimerization reaction. Although they are almost identical in catalytic selectivity, they do differ in other catalytic properties, especially the rate of reaction (66). In practice, the Co system is superior to the Fe system our discussion will therefore focus mainly on the former system. 

In the literature there are many reports of the formation of active catalyst for the 1 1 codimerization or synthesis of 1,4-hexadiene employing a large variety of Co or Fe salts, in conjunction with different kinds of ligands and organometallic cocatalysts. There must have been many structures, all of which are active for the codimerization reaction to one degree or another. The scope of the catalyst compositions claimed to be active as the codimerization catalysts is shown in Table XV (69-82). As with the nickel catalyst system discussed earlier, the preferred Co or Fe catalyst system requires the presence of phosphine ligands and an alkylaluminum cocatalyst. The catalytic property can be optimized by structural control of these two components. 

The codimerization reaction with unsymmetrically substituted silylalkynes is highly regioselective, specifically giving the isomer with the sily lated carbon attached to the metal. In an analogous manner, metallabicyclo[3.1.0]hexene derivatives of both titanium and... 

The mechanism of the codimerization reactions is depicted in Scheme 15367. It involves the initial coordination of the reactants (i) which is common to both proximal and distal reaction modes. For the proximal mode oxidative addition to the metal occurs next (ii) affording spirometallacyclopentane, which subsequently undergoes cyclopropylvinyl-homoallyl type rearrangement to metallacyclohexane (Hi), and finally demetalation (iv) via reductive elimination. On the other hand, the distal pathway involves ring-opening to TMM intermediate complex (v), followed by successive oxidative addition to a a,n-a y complex (vi) and demetalation (vii). 

Dimerization and codimerization reactions are widely used on an industrial scale either to provide chemicals of high added value or to upgrade by-product olefinic streams coming from various hydrocarbon cracking processes (steam or catalytic cracking) or hydrocarbon forming processes (Fischer-Tropsch synthesis or methanol condensation) (e. g., according to eq. (1)). 

The cyclodimerization of methylenecyclopropanes has already been discussed in Sect. 3.2. For the sake of completeness we just mention a some what misleading reaction, in which methylenecyclopropane apparently serves only as a source of butadiene (Eq. 79) 185) before we turn our attention to the more interesting codimerization reactions. 

These codimerization reactions are mainly limited by the degree of n-bond strength of the electron deficient alkenes to Pd(0). Strongly bonded ligands may prevent any interaction of the metal with the methylenecyclopropane. Typical examples of too strongly bonded alkenes are maleic anhydride, acrolein and acrylonitrile. On the other hand, too weak interactions may result in cyclodimerization of the methylenecyclopropane rather than codimerization. 

A large variety of codimerization reactions under the catalytic action of copper complexes is known. Usually, these reactions proceed via carbene intermediates and provide substituted ethenylcyclopropanes. Most of the catalysts for these reactions consist of copper(I) chloride and a phosphorus ligand, such as triphenylphosphane or triphenyl phosphite. Under the influence of these catalysts, carbenes are presumably formed from various substituted cyclopropenes at temperatures ranging from —40 to - -20°C, and these carbenes can be trapped by reaction with alkenes. ... 

The tendency for homo-cyclodimerization, which may be an important side reaction, is much smaller for 3-cyclopropyl-3-methylcyclopropene than for 3,3-dimethylcyclopropene. Therefore, the former is preferably used in codimerization reactions. 

When phosphane-free nickel complexes, such as bis(cycloocta-l,5-diene)nickel(0) or te-tracarbonylnickel, are employed in the codimerization reaction of acrylic esters, the codimer arising from [2-1-1] addition to the electron-deficient double bond is the main product. The exo-isomer is the only product in these cyclopropanation reactions. This is opposite to the carbene and carbenoid addition reactions to alkenes catalyzed by copper complexes (see previous section) where the thermodynamically less favored e Jo-isomers are formed. This finding indicates that the reaction proceeds via organonickel intermediates rather than carbenoids or carbenes. The introduction of alkyl substituents in the /I-position of the electron-deficient alkenes favors isomerization and/or homo-cyclodimerization of the cyclopropenes. Thus, with methyl crotonate and 3,3-diphenylcyclopropene only 16% of the corresponding ethenylcyc-lopropane was obtained. Methyl 3,3-dimethylacrylate does not react at all with 3,3-dimethyl-cyclopropene, so that the methylester of tra 5-chrysanthemic acid cannot be prepared in this way. This reactivity pattern can be rationalized in terms of a different tendency of the alkenes to coordinate to nickel(O). This tendency decreases in the order un-, mono- < di-< tri- < tet-... 

Additionally, regioisomeric differentiation can result at some stage of the actual codimerization reaction, e.g. the initial C —C bond-formation step of the distal opening (A) or in the course of the cyclopropylmethyl to but-3-enyl rearrangement of the proximal opening (B). The occurrence of formal [2-I-2]-cycloaddition products of MCP and alkene can also be rationalized from this mechanistic scheme. Thus, these products could arise from the spirometallaheptane by reductive elimination instead of the cyclopropylmethyl to but-3-enyl rearrangement. 

Di- and oligomerization reactions of methylenecyclopropanes are also the most important competing side reactions in many of the codimerization reactions employing methylenecyclo-propane (MCP), which are summarized in Section 2.2.2.3. They are specially favored if the cosubstrate, i.e. the alkene in a [3-I-2]-cycloaddition reaction, is only weakly bound to the metal center thus allowing it to be replaced by a second molecule of MCP. 

Most remarkably, no codimerization involving the unsaturated ester molecule is observed, even though nickel compounds such as the acrylonitrile complex are efficient catalysts for such codimerization reactions vide infra). An additional homodimer, l,3-bis(methylene)cy-clohexane (6), is formed as a minor product with phosphane-modified nickel catalysts. Only trace amounts of product 6 are obtained with maleic anhydride as cocatalyst. 

Even simple, nonstrained alkenes, such as ethene, are capable of undergoing a codimerization reaction with MCP in the presence of palladium(O) catalysts. However, besides the cycloadduct methylenecyclopentane (11 20.4% yield), one of the homodimers of MCP, 5-methylene-spiro[2.4]heptane (7), is formed predominantly (64% yield) even when a large excess of ethene is present. ... 

Asymmetric modifications of hydrovinylation are one of the earliest examples of successful asymmetric transition metal catalysis. After optimization of various dimerization and codimerization reactions using phosphane modified nickel catalysts, the first examples of asymmetric olefin codimerization were reported with n-allylnickel halides activated by organoaluminum chloride and modified by chiral phosphanes7. Thus, codimerization of 2-butene with propene using n-allylnickel chloride/A]X, (X = Cl, Br) in the presence of tris(myrtanyl)phosphane gives low yields of (—)-( )-4-methy 1-2-hexene (I) with 3% ee7,7 . 

Mechanisms illustrating the possible types of processes involved in the reactions are suggested for the codimerization reaction [Eq. (87)J and the dimerization reaction [Eq. (90)] (Schemes 6 and 7). 

---