The Rhodium Process

The high selectivity and mild conditions make the rhodium process more attractive than the cobalt one for the manufacture of n-butyraldehyde. The high cost of rhodium makes near-complete catalyst recovery a must for the commercial viability of the process. As we shall see, this has been achieved by developing an elegant separation method based on water-soluble phosphines. 

The catalytic cycle and the catalytic intermediates for the rhodium-plus-phosphine-based process are shown in Fig. 5.1. It is important to note that hydroformylation with rhodium can also be effected in the absence of phosphine. In such a situation CO acts as the main ligand (i.e., in Fig. 5.1, L = CO). The mechanistic implications of this is discussed later (Section 5.2.4). 

The following points are to be noted. First of all, complexes 5.1, 5.3, 5.5, and 5.7 are 18-electron complexes, while the rest are 16-electron ones. Second, conversions of 5.3 to 5.4 and 5.5 to 5.6 are the two insertion steps. The selectivity towards n-butyraldehyde is determined in the conversion of 5.3 to 5.4. It is possible that a rhodium-isopropyl rather than rhodium-propyl complex is formed. In such a situation on completion of the catalytic cycle isobutyral-dehyde will be the product. In practice both the n-propyl and the /-propyl complexes of rhodium are formed, and a mixture of n-butyraldehyde and /-butyraldehyde is obtained. This aspect is discussed in greater detail in the following section. Third, the catalyst precursor 5.1 undergoes ligand dissocia- 

TABLE 5.1 Process Parameters for Several Hydroformylation Processes 

Process parameters Cobalt Cobalt + phosphine Rhodium + phosphine 

---

In contrast to the rhodium process the most abundant iridium species, the catalyst resting state, in the BP process is not the lr(l) iodide, but the product of the oxidative addition of Mel to this complex. 

The rhodium process is highly selective and operates under mild reaction pressure, 400 to 1000 psig. However, because of the high price of rhodium, an efficient recovery of the catalyst is essential. An expensive rhodium recovery section is an integral part of any new acetic acid plant (15-16). This can be a substantial financial burden, especially in a smaller plant. 

The reaction effluent of the rhodium process is subjected to flash distillation under mild conditions in order to maintain the activity of the sensitive catalyst and to contain the costly metal in a limited section of the plant. The effluent of the nickel process is subjected directly to distillation of the products with considerable savings. 

Several economic evaluations of the nickel process as compared with the rhodium process, made by Halcon and by independent consultants, agreed that there is savings of about Iq/lb of acetic acid in favor of the nickel process. The difference is derived essentially from utility savings due to the higher reaction rate, simplified separation and lower water concentration. The nickel process consumes less than half the energy needed for the rhodium-catalyzed process. The catalyst inventory, and the equipment needed for its recovery contribute to the higher cost of production in the rhodium case. 

The rhodium process has been in commercial use since 1976. In the interim the increasing importance ot environmental, health and safety considerations in industrial processes make the use of less-extreme reaction conditions, catalyst longevity and the minimization of side-product formation of crucial importance. 

The costs of the rhodium process are, however, higher owing to the required work up, catalyst recycling, and corrosion problems. Therefore, intensive research is being carried out to develop heterogeneous rhodium catalysts. However, this has so far been thwarted by the low stabihty of die catalysts. 

Since Shell s report on the use of phosphines in this process [3], many industries started applying phosphine ligands in the rhodium process as well [52]. While alkylphosphines are the ligands of choice for cobalt, they lead to slow catalysis when applied in rhodium catalysis. In the mid-sixties the work of Wilkinson showed that arylphosphines should be used for rhodium and that even at very mild conditions very active catalysts can be obtained [9]. 

While reaction rates for cobalt catalysts depend on the methanol concentration and partial pressure of carbon monoxide, the rates with rhodium catalysts are independent of both reactant and product concentrations. The reaction mechanism for cobalt catalyst in methanol carbonylation is similar to hydroformyla-tion but is different for rhodium catalysts. The catalyst intermediate in the rhodium process is Rh(CO)2l2 and has been identified at 100°C and 6 atm by infrared spectroscopy. Reaction with methyl iodide in the rate-determining step then forms a methyl rhodium complex that rapidly gives an acyl complex (CH3CORh(CO)l3)2 from reaction with the high carbon monoxide content in the reactor. The complex decomposes after reaction with water to produce acetic acid, and with simultaneous regeneration of the catalyst ... 

Thus, as expected the reaction in Scheme 9.10 (66 67) is much faster for iridium. In and of itself this does not mean that the iridium catalyst is therefore faster than the rhodium catalyst As we have learned before, the reductive elimination may be slower for iridium. Apparently, this is not the case. Migration is now the slowest step (67 69) [126]. In contrast to the rhodium process, the most abundant iridium species, the catalyst resting state, in the BP process is not 66, but the product of the oxidative addition of Mel to this complex, 67. 

Since 1960, the Hquid-phase oxidation of ethylene has been the process of choice for the manufacture of acetaldehyde. There is, however, stiU some commercial production by the partial oxidation of ethyl alcohol and hydration of acetylene. The economics of the various processes are strongly dependent on the prices of the feedstocks. Acetaldehyde is also formed as a coproduct in the high temperature oxidation of butane. A more recently developed rhodium catalyzed process produces acetaldehyde from synthesis gas as a coproduct with ethyl alcohol and acetic acid (83—94). 

Catalyst recovery is a major operational problem because rhodium is a cosdy noble metal and every trace must be recovered for an economic process. Several methods have been patented (44—46). The catalyst is often reactivated by heating in the presence of an alcohol. In another technique, water is added to the homogeneous catalyst solution so that the rhodium compounds precipitate. Another way to separate rhodium involves a two-phase Hquid such as the immiscible mixture of octane or cyclohexane and aliphatic alcohols having 4—8 carbon atoms. In a typical instance, the carbonylation reactor is operated so the desired products and other low boiling materials are flash-distilled. The reacting mixture itself may be boiled, or a sidestream can be distilled, returning the heavy ends to the reactor. In either case, the heavier materials tend to accumulate. A part of these materials is separated, then concentrated to leave only the heaviest residues, and treated with the immiscible Hquid pair. The rhodium precipitates and is taken up in anhydride for recycling. 

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. 

This reaction is rapidly replacing the former ethylene-based acetaldehyde oxidation route to acetic acid. The Monsanto process employs rhodium and methyl iodide, but soluble cobalt and iridium catalysts also have been found to be effective in the presence of iodide promoters. 

Liquid-phase oxidation of lower hydrocarbons has for many years been an important route to acetic acid [64-19-7]. In the United States, butane has been the preferred feedstock, whereas ia Europe naphtha has been used. Formic acid is a coproduct of such processes. Between 0.05 and 0.25 tons of formic acid are produced for every ton of acetic acid. The reaction product is a highly complex mixture, and a number of distillation steps are required to isolate the products and to recycle the iatermediates. The purification of the formic acid requires the use of a2eotropiag agents (24). Siace the early 1980s hydrocarbon oxidation routes to acetic acid have decliaed somewhat ia importance owiag to the development of the rhodium-cataly2ed route from CO and methanol (see Acetic acid). 

This process is one of the three commercially practiced processes for the production of acetic anhydride. The other two are the oxidation of acetaldehyde [75-07-0] and the carbonylation of methyl acetate [79-20-9] in the presence of a rhodium catalyst (coal gasification technology, Halcon process) (77). The latter process was put into operation by Tennessee Eastman in 1983. In the United States the total acetic anhydride production has been reported to be in the order of 1000 metric tons. 

Ligand-Modified Rhodium Process. The triphenylphosphine-modified rhodium oxo process, termed the LP Oxo process, is the industry standard for the hydroformylation of ethylene and propylene as of this writing (ca 1995). It employs a triphenylphosphine [603-35-0] (TPP) (1) modified rhodium catalyst. The process operates at low (0.7—3 MPa (100—450 psi)) pressures and low (80—120°C) temperatures. Suitable sources of rhodium are the alkanoate, 2,4-pentanedionate, or nitrate. A low (60—80 kPa (8.7—11.6 psi)) CO partial pressure and high (10—12%) TPP concentration are critical to obtaining a high (eg, 10 1) normal-to-branched aldehyde ratio. The process, first commercialized in 1976 by Union Carbide Corporation in Ponce, Puerto Rico, has been ficensed worldwide by Union Carbide Corporation and Davy Process Technology. 

The switch from the conventional cobalt complex catalyst to a new rhodium-based catalyst represents a technical advance for producing aldehydes by olefin hydroformylation with CO, ie, by the oxo process (qv) (82). A 200 t/yr CSTR pilot plant provided scale-up data for the first industrial,... 

Alternatively, butadiene can be oxidized in the presence of acetic acid to produce butenediol diacetate, a precursor to butanediol. The latter process has been commercialized (102—104). This reaction is performed in the Hquid phase at 80°C with a Pd—Te—C catalyst. A different catalyst system based on PdCl2(NCCgH )2 has been reported (105). Copper- (106) and rhodium- (107) based catalysts have also been studied. 

The processiag costs associated with separation and corrosion are stiU significant ia the low pressure process for the process to be economical, the efficiency of recovery and recycle of the rhodium must be very high. Consequently, researchers have continued to seek new ways to faciUtate the separation and confine the corrosion. Extensive research was done with rhodium phosphine complexes bonded to soHd supports, but the resulting catalysts were not sufficiently stable, as rhodium was leached iato the product solution (27,28). A mote successful solution to the engineering problem resulted from the apphcation of a two-phase Hquid-Hquid process (29). The catalyst is synthesized with polar -SO Na groups on the phenyl rings of the triphenylphosphine. 

Hydroformylation catalyzed by rhodium triphenylphospine results in only the 9 and 10 isomers in approximately equal amounts (79). A study of recycling the rhodium catalyst and a cost estimate for a batch process have been made (81). 

Precious Meta.1 Ca.ta.lysts, Precious metals are deposited throughout the TWC-activated coating layer. Rhodium plays an important role ia the reduction of NO, and is combiaed with platinum and/or palladium for the oxidation of HC and CO. Only a small amount of these expensive materials is used (31) (see Platinum-GROUP metals). The metals are dispersed on the high surface area particles as precious metal solutions, and then reduced to small metal crystals by various techniques. Catalytic reactions occur on the precious metal surfaces. Whereas metal within the crystal caimot directly participate ia the catalytic process, it can play a role when surface metal oxides are influenced through strong metal to support reactions (SMSI) (32,33). Some exhaust gas reactions, for instance the oxidation of alkanes, require larger Pt crystals than other reactions, such as the oxidation of CO (34). 

The direct combination of selenium and acetylene provides the most convenient source of selenophene (76JHC1319). Lesser amounts of many other compounds are formed concurrently and include 2- and 3-alkylselenophenes, benzo[6]selenophene and isomeric selenoloselenophenes (76CS(10)159). The commercial availability of thiophene makes comparable reactions of little interest for the obtention of the parent heterocycle in the laboratory. However, the reaction of substituted acetylenes with morpholinyl disulfide is of some synthetic value. The process, which appears to entail the initial formation of thionitroxyl radicals, converts phenylacetylene into a 3 1 mixture of 2,4- and 2,5-diphenylthiophene, methyl propiolate into dimethyl thiophene-2,5-dicarboxylate, and ethyl phenylpropiolate into diethyl 3,4-diphenylthiophene-2,5-dicarboxylate (Scheme 83a) (77TL3413). Dimethyl thiophene-2,4-dicarboxylate is obtained from methyl propiolate by treatment with dimethyl sulfoxide and thionyl chloride (Scheme 83b) (66CB1558). The rhodium carbonyl catalyzed carbonylation of alkynes in alcohols provides 5-alkoxy-2(5//)-furanones (Scheme 83c) (81CL993). The inclusion of ethylene provides 5-ethyl-2(5//)-furanones instead (82NKK242). The nickel acetate catalyzed addition of r-butyl isocyanide to alkynes provides access to 2-aminopyrroles (Scheme 83d) (70S593). 

The 0X0 process for higher alcohols CO -1- H9 -1- C3H6 /1-butanal further processing. Catalyst is rhodium triphenylphos-phine coordination compound, 100°C (212°F), 30 atm (441 psi). 

Acetic acid from methanol by the Monsanto process, CH3OH -1-CO CH3COOH, rhodium iodide catalyst, 3 atm (44 psi), 150°C (302°F), 99 percent selectivity of methanol. 

---