Hydrocarbonylation olefin

Rhodium Ca.ta.lysts. Rhodium carbonyl catalysts for olefin hydroformylation are more active than cobalt carbonyls and can be appHed at lower temperatures and pressures (14). Rhodium hydrocarbonyl [75506-18-2] HRh(CO)4, results in lower -butyraldehyde [123-72-8] to isobutyraldehyde [78-84-2] ratios from propylene [115-07-17, C H, than does cobalt hydrocarbonyl, ie, 50/50 vs 80/20. Ligand-modified rhodium catalysts, HRh(CO)2L2 or HRh(CO)L2, afford /iso-ratios as high as 92/8 the ligand is generally a tertiary phosphine. The rhodium catalyst process was developed joindy by Union Carbide Chemicals, Johnson-Matthey, and Davy Powergas and has been Hcensed to several companies. It is particulady suited to propylene conversion to -butyraldehyde for 2-ethylhexanol production in that by-product isobutyraldehyde is minimized. 

Often the aldehyde is hydrogenated to the corresponding alcohol. In general, addition of carbon monoxide to a substrate is referred to as carbonylation, but when the substrate is an olefin it is also known as hydroformylation. The eady work on the 0x0 synthesis was done with cobalt hydrocarbonyl complexes, but in 1976 a low pressure rhodium-cataly2ed process was commerciali2ed that gave greater selectivity to linear aldehydes and fewer coproducts. 

Hydroformylation, or the 0X0 process, is the reaction of olefins with CO and H9 to make aldehydes, which may subsequently be converted to higher alcohols. The catalyst base is cobalt naph-thenate, which transforms to cobalt hydrocarbonyl in place. A rhodium complex that is more stable and mnctions at a lower temperature is also used. 

Olefin hydrocarbonylation can be used in conjunction with oxidative addition to prepare indanones and cyclopentenones, but the reaction is limited to terminal alkenes.243... 

The Allylic Exchange of RCHaCH with HCo(CO)4 A second scheme for the isomerization of olefins by HCo(CO)4 consists of an exchange of an allylic hydrogen of the olefin with the hydrogen of the hydrocarbonyl through a six-membered transition state 18) ... 

Several related reations are summarized in Scheme 4.13. When a nucleophilic OH or NH2 group exists at the proper position of the olefinic substrate, cyclohy-drocarbonylation provides a lactone or a lactam. Allyl alcohols" " and homoallyl amines are converted into y-lactones and 8-lactams in high enantiomeric excesses. Thiol is employable as a nucleophile for hydrocarbonylation of a C—C double bond. A successful catalytic carbonylation is reported for hydrothiocarbonylation. ... 

One process that capitalizes on butadiene, synthesis gas, and methanol as raw materials is BASF s two-step hydrocarbonylation route to adipic acid(3-7). The butadiene in the C4 cut from an olefin plant steam cracker is transformed by a two-stage carbonylation with carbon monoxide and methanol into adipic acid dimethyl ester. Hydrolysis converts the diester into adipic acid. BASF is now engineering a 130 million pound per year commercial plant based on this technology(8,9). Technology drawbacks include a requirement for severe pressure (>4500 psig) in the first cobalt catalyzed carbonylation step and dimethyl adipate separation from branched diester isomers formed in the second carbonylation step. 

This review deals with the recent developments in the transition metal-catalyzed carbonylation reaction, especially hydroformylation, hydrocarbonylation, and oxidative hydrocarbonylation reactions of olefins, referring to literature since 1994. Because of the importance of carbonyl functionality in organic chemistry and the ideal atom efficiency of... 

The addition of cobalt hydrocarbonyl to olefins has been investigated and information on the detailed mechanism of the reaction obtained. The reaction of 1-pentene with cobalt hydrocarbonyl to produce a mixture of 1- and 2-pentylcobalt tetracarbonyls was shown to be inhibited by carbon monoxide (46). The inhibition very likely arises because the reactive species is cobalt hydrotricarbonyl rather than the tetracarbonyl. The carbon monoxide, by a mass action effect, reduces the concentration of the reactive species. 

Cobalt hydrocarbonyl is a very reactive compound. It reacts extremely rapidly with triphenylphosphine, probably by a first-order dissociation mechanism, producing cobalt hydrotricarbonyl triphenylphosphine (44). This demonstrates the very ready replacement of one ligand by another. Cobalt hydrocarbonyl also catalyzes the isomerization of olefins. Under conditions of the hydroformylation reaction, olefin isomerization is observed. But there is controversy as to whether or not rearranged aldehydes (aldehydes which cannot be produced by simple addition to the starting olefin) are produced mainly by rearrangement of an intermediate in the reaction (28, 75, 55) or by reaction of isomerized olefins (55). 

Manganese hydrocarbonyl, though much less reactive than cobalt hydrocarbonyl, does add to some activated olefins. Tetrafluoroethylene for example reacted to give tetrafluoroethylmanganese pentacarbonyl (95). 

Metal Hydrides. Metal hydrides generally react readily with acetylenes, often by an insertion mechanism. Cobalt hydrocarbonyl gives complicated mixtures of compounds with acetylenes. The only products which have been identified so far are dicobalt hexacarbonyl acetylene complexes (34). Greenfield reports that, under conditions of the hydroformy lation reaction, acetylenes give only small yields of saturated monoaldehydes (30), probably formed by first hydrogenating the acetylene and then reacting with the olefin. Other workers have identified a variety of products from acetylene, carbon monoxide, and an alcohol with a cobalt catalyst, probably cobalt hydrocarbonyl. The major products observed were succinate esters (74,19) and succinate half ester acetals (19). 

Figure D shows some olefin insertion reactions. Hydride additions to olefins have been known for a long while. Among these many examples, manganese hydrocarbonyl, and cobalt hydrocarbonyl, magnesium hydride, diborane, alkylalu-minum hydrides, germanium and tin hydrides all add quite readily to olefins. These last two cases are questionable because the mechanism is not clear. Some of these additions occur without a catalyst some are speeded up by ultraviolet light some are catalyzed by Group VIII metals. So it is not clear whether all these reactions are the same or whether there are several different mechanisms.

Figure F shows some acetylene insertion reactions. These, too, are similar to the olefin insertion reactions. The manganese and cobalt hydrocarbonyls again add. Chloronickelcarbonyl hydride, which I believe is an intermediate in many of the nickel carbonyl-catalyzed reactions, adds to olefins. Diborane and the aluminum hydrides also add.

The mechanism most consistent with all the data is an ionic acid opening of the epoxide —apparently where the hydrocarbonyl is used as an acid to attack the epoxide— which is more sensitive to steric effects than to electronic factors. This conclusion may at first appear to be inconsistent with our previous finding that isobutylene reacted with cobalt hydrocarbonyl to give exclusively addition of the cobalt to the tertiary position. The inhibitory effect of carbon monoxide on that reaction, however, indicated that it was probably cobalt hydrotricarbonyl that was actually adding to the olefin and steric effects would be expected to be much less important with the tricarbonyl than with the tetracarbonyl (7) Apparently he feels now that the former reactions really involve the tricarbonyl, loss of CO being important to get the reaction running whereas epoxide attack perhaps involves a tetracarbonyl, steric factors are more important here. 

Despite very extensive studies on this reaction, there is still considerable uncertainty about its mechanism. The reaction occurs at about the same rate in a wide variety of organic solvents, including benzene, heptane, and alcohol, suggesting that polar intermediates are not involved (Wender et al., 41). Reaction of the olefin with preformed cobalt hydrocarbonyl also gives the aldehyde product (Wender et al., 4 ). This, together with the observation that cobalt hydrocarbonyl is formed under hydroformylation conditions in the absence of olefin, but cannot be detected in the presence... 

It should be noted that the Natta-Martin mechanism, while satisfactory from the kinetic standpoint, does not assign any role in the reaction to cobalt hydrocarbonyl. Hence, it is not readily reconciled with the evidence for the formation of the latter under hydroformylation conditions and its known reactions with olefins (Kirch and Orchin, 43). 

We first established that hydrocarbonylation reactions occur with cis-stereochemistry (29, 16) and that asymmetric induction occurs before or during the formation of the metal alkyl intermediate (5, 6). This means that is either during the 7r-olefin complex formation between catalyst and substrate or during the insertion of the 7r-complexed olefin into the M-H bond. Therefore, the model should focus on the interactions between the substrate double bond and the catalytically active metal atom of the catalyst. 

Figure 2. Models for the transition states controlling asymmetric induction in the hydrocarbonylation of olefins...

Asymmetric hydrocarbonylation is a promising method for synthesizing optically active oxygenated compounds from prochiral olefins. Despite the reaction conditions, which include high carbon... 

The fact that a model for the transition state controlling asymmetric induction based on steric interactions allows us to correctly predict the type of prevailing regio- and stereoisomer for about 85% of the asymmetric hydrocarbonylation experiments (including hydroformylation and hydrocarbalkoxylation) is an indication that asymmetric induction in these catalytic reactions is based mainly on steric interactions. The data obtained so far do not allow us to establish whether the more stable or the less stable 7r-olefin complex intermediate is the one that reacts preferentially. However, the regularities that we observed indicate that the kinetic features are the same, at least in most of the experiments. 

In this review the synthetic aspects of asymmetric hydroformylation will be discussed first the experimental data relevant to attempt a rationalization of the results will then be considered. The closely related synthesis of optically active aldehydes by hydroformylation of optically active olefinic substrates in the presence of achiral catalysts7,8 and the different asymmetric hydrocarbonylation reactions, such as the synthesis of esters from olefins, carbon monoxide and alcohols in the presence of optically active catalysts9 , are beyond the scope of this review and will not be discussed here. 

The noncatalytic reaction of cobalt hydrocarbonyl with olefins to produce aldehydes... 

When a large excess of olefin was used, the cobalt hydrocarbonyl was completely used up in Eq. (2), and Eq. (3) did not occur. In this case the products have been recovered as the triphenylphosphine derivatives, or as the esters by reaction of the acylcobalt carbonyls with iodine and an alcohol. 

The conditions under which cobalt hydrocarbonyl was reacted with olefin were also found to affect the distribution of products and the extent of isomerization of excess olefin (62, 73, 147). At low temperatures (0° C) under carbon monoxide (1 atm) very little isomerization of excess 1-pentene occurred and the main product was the terminal aldehyde. Under nitrogen or under carbon monoxide at 25° C, extensive olefin isomerization occurred and the branched aldehyde was mainly produced. The olefin isomerization is most satisfactorily accounted for by an equilibrium between alkylcobalt and olefin-hydride cobalt complexes [Eqs. (9) and (10)]. The carbon monoxide inhibition is most easily explained if the isomerization proceeds via the tricarbonyls rather than tetracarbonyls. This also explains why ethylcobalt tetracarbonyl is not in equilibrium with hydrocarbonyl and ethylene under conditions where the isomerization is rapid (62, 73). 

Prior to Takegami s studies, the effect of isomerization of acylcobalt carbonyls on the products of the reaction between cobalt hydrocarbonyl and olefins had received little attention. Terminal olefins had been found to give a mixture of linear and branched products at low temperatures under carbon monoxide, and this was taken as reflecting the mode of addition of cobalt hydrocarbonyl (62, 73, 147). In view of the slow rate of isomerization of acylcobalt carbonyls this seems justified. However, it is worth noting that branched products predominated in the reaction of 1-pentene with hydrocarbonyl under nitrogen even when the olefin had isomerized only to the extent of 50% (73). Both isobutylene and alkyl acrylates had been found to produce branched products. It was suggested that isobutylene, with an... 

Takegami et al. (147) reexamined the reaction of olefins with cobalt hydrocarbonyl to determine the effect of reaction variables and the possibility of isomerization on the structure of the acylcobalt carbonyls formed. Products were recovered as their ethyl esters as described previously. Additionally, the uptake of carbon monoxide was measured. This is important since the presence of an acylcobalt tricarbonyl can be inferred when the amount of ester produced exceeds the amount of carbon monoxide absorbed. 

In view of the fact that no free hydrocarbonyl is normally detected in the presence of olefin under Oxo conditions (92, 107), Eq. (22) seems more likely than Eq. (21). 

The stoichiometric hydroformylation of olefins with cobalt hydrocarbonyl is also inhibited by an atmosphere of carbon monoxide (62, 73) (Section II, A) and this has been shown to involve a CO inhibition of alkylcobalt carbonyl formation (Eq. (18)). 

Piacenti et al. suggested that the different results at low and high carbon monoxide pressure were due to different catalytic intermediates (A and B) under the two sets of conditions. Thus at low pressures A caused a rapid olefin isomerization and the formation of similar product distributions of aldehydes from 1- and 2-pentene. At high pressures little olefin isomerization occurred and 1-olefin yielded significantly more straight-chain aldehyde than 2-olefin. This would seem consistent with Heck and Breslow s mechanism (62) if A were an acylcobalt tricarbonyl in equilibrium with isomeric olefin-cobalt hydrocarbonyl complexes and B were an acylcobalt tetracarbonyl. 

The present authors feel this point needs further investigation in view of the results of Takegami et al. They found that the isomerization of the acylcobalt tetracarbonyl was very solvent-dependent, and it could well be that conditions in the hydroformylation of olefins and orthoformates were sufficiently different to cause faster isomerization in the former case. Thus, for example, the presence of olefins in the former case may contribute to a faster isomerization, or perhaps orthoformates, like tetrahydrofuran, inhibit the isomerization. A further factor to be considered is the presence of cobalt hydrocarbonyl, which must be present in larger amount in the case of olefin hydroformylation. Takegami et al. (143) have shown that cobalt hydrocarbonyl strongly promotes the isomerization of phenylacetylcobalt car-... 

Some light has been thrown on this unusual reaction by a study of the reaction of cobalt hydrocarbonyl with olefins under nitrogen (14). It has also be discussed recently by Heck (59). 

A useful study has just been completed by Roos and Orchin (125), who have examined the effect of ligands such as benzonitrile on the stoichiometric hydroformylation of olefins. A variety of such reagents (acetonitrile, anisole) were found to act in a similar manner to carbon monoxide by suppressing the formation of branched products and the isomerization of excess olefin. The yield of aldehyde was also increased by increasing ligand concentration up to 2 moles per mole of cobalt hydrocarbonyl. Benzonitrile was not found to affect the rate of the reaction of cobalt hydrocarbonyl with acylcobalt tetracarbonyl, so the ligand must have affected an earlier step in the reaction sequence. It seems most likely that cobalt hydrocarbonyl reacts with olefin in the presence of benzonitrile to form an acylcobalt tricarbonyl-benzonitrile complex which is reduced more rapidly than the acylcobalt tetracarbonyl. 

As with the hydroformylation of olefins, aldehydes are expected by a reduction of acylcobalt carbonyls by cobalt hydrocarbonyl. They are formed in small amounts for a number of epoxides (145). 

In view of the many differences noted above between the hydroformylation of olefins and epoxides, it is not surprising to find that changes in structure result in a different order of reactivity in each case. Thus for epoxides the reactivity to cobalt hydrocarbonyl is cyclohexene oxide > propylene oxide, whereas with olefins the order is terminal olefins > internal olefins > cyclic olefins (145). 

This isomerization to ketones also occurs under the milder conditions under which cobalt hydrocarbonyl is reacted with epoxides, however, and it seems likely that cobalt hydrocarbonyl was also present under the conditions of Eisenmann s experiment. Heck has therefore suggested that the mechanism could involve the formation of a hydroxyalkylcobalt carbonyl followed by elimination to produce the enol form of the ketone in the same way that alkylcobalt carbonyls can give olefins. 

Originally, Piacenti et al. explained the formation of isomeric products in terms of an equilibrium of alkylcobalt carbonyls with olefin-hydrocarbonyl complexes as in the Oxo reaction. More recently, however, they have noted that the conditions under which n-propyl orthoformate gave no isomeric products (below 150° C, carbon monoxide pressure 10 atm) are conditions under which isomerization occurs readily in the hydroformylation of olefins (115). Since alkylcobalt carbonyls were formed in both reactions they dismissed the possibility that this isomerization was due to alkyl- or acylcobalt carbonyls. The fact that Takegami et al. have found that branched-chain acylcobalt tetracarbonyls isomerize more readily than straight-chain acylcobalt tetracarbonyls would seem to fit in quite well with the results of Piacenti et al., however, and suggests that the two findings may not be so irreconcilable as might at first appear (see Section II, B,2). 

---