Hydroformylation catalyst decomposition

Conventional triorganophosphite ligands, such as triphenylphosphite, form highly active hydroformylation catalysts (95—99) however, they suffer from poor durabiUty because of decomposition. Diorganophosphite-modified rhodium catalysts (94,100,101), have overcome this stabiUty deficiency and provide a low pressure, rhodium catalyzed process for the hydroformylation of low reactivity olefins, thus making lower cost amyl alcohols from butenes readily accessible. The new diorganophosphite-modified rhodium catalysts increase hydroformylation rates by more than 100 times and provide selectivities not available with standard phosphine catalysts. For example, hydroformylation of 2-butene with l,l -biphenyl-2,2 -diyl... 

During a 33 h continuous hydroformylation run using this set-up, no catalyst decomposition was observed and Rh leaching into the scC02/product stream was less than 1 ppm. The selectivity for the linear nonanal was found to be stable over the reaction time with n/iso = 3.1. During the continuous reaction, alkene, CO, H2 and C02 were separately fed into the reactor containing the ionic liquid catalyst solution. Products and unconverted feedstock dissolved in SCCO2 were removed from the ionic liquid. After decompression the liquid product was collected and analysed. 

Thermal decomposition of RhH(CO)(PPh3)3, the well known hydroformylation catalyst, in the absence of H2 and CO leads to a stable cluster shown in Figure 2.36 containing p2-PPh2 fragments [31], Under hydroformylation conditions also other products are found such as benzaldehyde, benzene, and diphenylpropylphosphine. 

Cluster or bimetallic reactions have also been proposed in addition to monometallic oxidative addition reactions. The reactions do not basically change. Reactions involving breaking of C-H bonds have been proposed. For palladium catalysed decomposition of triarylphosphines this is not the case [32], Likewise, Rh, Co, and Ru hydroformylation catalysts give aryl derivatives not involving C-H activation [33], Several rhodium complexes catalyse the exchange of aryl substituents at triarylphosphines [34] ... 

The hydrolytic decomposition of a potential fluorophosphite ligand would generate free fluoride ions which would be expected to be detrimental to the activity of a hydroformylation catalyst. The patent literature contains abundant references to the detrimental effects of halogens (6) on hydroformylation catalysts, and based on the patent information, one could not reasonably expect a halophosphite to be a successful hydroformylation ligand. However, a second publication by Klender (7) shows that exposure of / and other fluorophosphites to moisture at temperatures of 250°C to 350°C does not generate fluoride, even at part per million levels. 

Catalyst decomposition is, overall, receiving little attention in academic work on homogeneous catalysis, and only in recent years has research on decomposition and stabilization of organometallic catalysts started to expand (116), with emphasis on reactions of significant commercial interest such as hydroformylation (117), metathesis 118), crosscoupling, and polymerization 119). Ligand decomposition seems to be a key issue for industrial application, because it affects the total number of turnovers, TON. Phosphine decomposition is an unavoidable side reaction in metal-phosphine complex-catalyzed reactions and the main barrier for commercial application of homogeneous catalysts. There are a few exceptions to this statement for example, the rhodium tppts-catalyzed hydroformylation of propene, a process developed by Ruhrchemie-Rhone Poulenc (now Celanese). 

For the different hydroformylation processes described above, the catalyst separation and recycling remain a constant preoccupation. This point is particularly crucial when a very expensive metal is used as catalyst (rhodium). The recycling is either operated by chemical transformation or by direct distillation, depending on the catalyst and its stability. In that way, the development of the aqueous biphasic process can be considered as an important breakthrough (4-6). The separation is operated by decantation, which simplifies the process scheme and limits the risks of catalyst decomposition during distillation. Even if the 0x0 Ruhrchemie/Rhone-Poulenc process presents undeniable advantages, this process remains limited to short-chain olefins (C2-C5) because of the low solubility of higher olefins in water which renders the reaction rates too low for viable processes [7]. 

HCo(CO), an exceptionaUy acidic transition metal "hydride", is formed by tfie reaction of Hj and COjfCOlg. HCo(CO)g is a trigonal bipyramidal d , 18-electron complex in which the hydride occupies an apical position. High temperature is required for industrially useful rates of hydroformylation catalyzed by HCo(CO)g. Moreover, high pressure of CO is required to prevent formation of higher cobalt clusters and of metallic cobalt. The rate of hydroformylation catalyzed by cobalt carbonyl depends on [H ] and Thus, increasing the pressure of a 1 1 mixture of CO Hj has little effect on tire rate but prevents catalyst decomposition. 

Several factors are needed to realize the tandem process. First, the catalyst should not be poisioned by the high concentration of amine. Tlie use of bidentate ligands helps to retard catalyst decomposition by this route. Second, the catalyst must be sufficiently electron rich to catalyze hydrogenation of the relatively electron-rich enamine, while not being too electron rich to prevent hydroformylation. These properties have been achieved with a number of systems, many of which have been outlined in a review on tandem processes initiated by hydroformylation. Some of these catalysts lack phosphine ligands, while others contain PPhj or more modern bisphosphine or bisphosphite ligands shown by the examples in the equations above. 

In the hydroformylation of lower alkenes using a modified cobalt catalyst complex separation is achieved by distillation. The ligands are high-boiling so that they remain with the heavy ends when these are removed from the alcohol product. Distillation is not possible when higher alcohols or aldehydes are produced, because of decomposition of the catalyst ligands at the higher temperatures required. Rhodium complexes can usually also be removed by distillation, since these complexes are relatively stable. 

The hypothesis that the cobalt carbonyl radicals are the carriers of catalytic activity was disproved by a high pressure photochemistry experiment /32/, in which the Co(CO), radical was prepared under hydroformylation conditions by photolysis of dicobalt octacarbonyl in hydrocarbon solvents. The catalytic reaction was not enhanced by the irradiation, as would be expected if the radicals were the active catalyst. On the contrary, the Co(C0)4 radicals were found to inhibit the hydroformylation. They initiate the decomposition of the real active catalyst, HCo(C0)4, in a radical chain process /32, 33/. 

In recent years the interest in hydroformylating higher alkenes with catalysts other than cobalt has increased. Platinum and palladium based catalysts have been studied and the results of the latter [10] seem very promising. Platinum has been known for many years to have a high preference for the formation of linear products, but ligand decomposition hampers applications [11]. 

Garland et al. have developed a powerful method for the reconstruction of individual pure component spectra from complex catalytic mixtures [20]. Using this band-target entropy minimization (BTEM) protocol, he was able to identify the mononuclear rhodium acyl intermediate in the hydroformylation reaction of 3,3-dimethylbut-l-ene starting from Rh4(a-CO)9(p-CO)3 as catalyst precursor [21]. In addition to the catalyst precursor and the more stable decomposition product... 

---