Hydroformylation membranes

In the early 1970 s, Bayer et al. reported the first use of soluble polymers as supports for the homogeneous catalysts. [52] They used non-crosslinked linear polystyrene (Mw ca. 100 000), which was chloromethylated and converted by treatment with potassium diphenylphosphide into soluble polydiphenyl(styrylmethyl)phosphines. Soluble macromolecular metal complexes were prepared by addition of various metal precursors e.g. [Rh(PPh3)Cl] and [RhH(CO)(PPh3)3]. The first complex was used in the hydrogenation reaction of 1-pentene at 22°C and 1 atm. H2. After 24 h (50% conversion in 3 h) the reaction solution was filtered through a polyamide membrane [53] and the catalysts could be retained quantitatively in the membrane filtration cell. [54] The catalyst was recycled 5 times. Using the second complex, a hydroformylation reaction of 1-pentene was carried out. After 72 h the reaction mixture was filtered through a polyamide membrane and recycled twice. 

After a hydroformylation run, the reaction solution was subjected to ultrafiltration using an asymmetric polyethersulfone membrane (MWCO 50 kDa) supplied by Sartorius. A retention of 99.8% was found. When the catalyst solution was recycled, virtually the same catalytic activity was observed again (165 TO h 1). Repetitive recycling experiments resulted in 2-7% loss of rhodium, which was subscribed to partial oxidation of the phosphine ligand. 

Since none of the commercially available nano- or ultrafiltration membranes so far shows real long-term resistance against organic solvents under the reaction conditions needed for a commercially interesting hydroformylation process and since no prices are available for bulk quantities of membranes for larger scale applications, considerations about the feasibility of such processes are difficult and would be highly speculative. 

Figure 5.8. Membrane steps as a constituent part of aqueous-biphasic hydroformylation A+B->C+D...

Sequential hydroformylation/reductive amination of dendritic perallylated polyglycerols with various amines in a one-pot procedure to give dendritic polyamines in high yields (73-99%). Furthermore, the use of protected amines provides reactive core-shell-type architectures after deprotection. These soluble but membrane filterable multifunctional dendritic polyamines are of high interest as reagents in synthesis or as supports in homogeneous catalysis as well as nonviral vectors for DNA-transfection (Scheme 18) [65]. 

A few years ago, a new class of ligands namely the sulfonated phosphites (for examples see Table 7, 132, 133) was described.283 287 They show remarkable stabilities in water compared to conventional phosphites such as P(OPh)3 and rhodium catalysts modified with 132 exhibited much higher catalytic activities in the hydroformylation of 1-tetradecene than conventional Rh/P(OPh)3 or Ph/PPh3 catalysts even at lower reaction temperatures.285,286 Sulfonated phosphite ligands may play a role in the emerging field of biphasic catalysis in ionic liquids15 22 or in combination with membrane separation of the metal complexes of these bulky ligands. 

The same dendritic ligands, but used in combination with rhodium, were utilized in hydroformylation reactions [46]. Preliminary experiments with this catalytic system in a nanofiltration membrane reactor, however, showed that this membrane set-up was not compatible with the standard hydroformylation conditions because of its temperature and solvent restrictions. 

Kim and Datta [1991] tested the above concept with the homogeneously catalyzed ethylene hydroformylation by hydridocarbonyltris (triphenyl phosphine) Rh (I) catalyst dissolved in dioctyl phthalate solvent. They concluded that for effective separation of the product, the transport resistance of the catalyst layer should be less than that for the membranes by controlling the liquid loading of the catalyst layer. They also pointed out that the organic membranes used can not withstand the aggressive reaction conditions and suggested that ceramic membranes appear to hold promises for practical applications. 

Membrane technology is a recent development to separate (or concentrate) water-soluble catalysts (mainly hydroformylation catalysts) [147, 149], although a prior art is known [194, 195]. There are proposals for the use of immobilized or re-immobilized aqueous phases for large-scale processes (cf. Ref. [222] and Section 3.1.1.6). Carbon dioxide as a solvent for biphasic hydroformylations has been described by Rathke and Klinger [184], although the use of CO2 for hydroformylation purposes was described earlier [185]. For the use of supercritical CO2 cf. Section 3.1.13 with non-aqueous ionic liquids cf. Section 3.1.1.2.2. Investigations with supercritical water are in an early state (e. g., Ref. [223]). 

Tricyclodecane dialdehyde (TCD-dial) the starting product for TCD-diamine, is formed by hydroformylation of dicyclopentadiene, DCP. This hydroformylation was used as a test reaction for re-immobilized catalysts in membrane techniques (eq. (6)). 

In order to stabilize the ammonium salts, the presence of free amine in the system is recommended. With respect to this additional requirement, an ideal system must have the same good retention for amines as well as for phosphorus and rhodium. Therefore the choice of distearylamine could only be regarded as a good compromise of hydroformylation requirements and membrane separation properties. 

The separation of the Rh-distearylamine-TPPTS catalyst system by membranes was tested on pilot plant scale with crude aldehyde from the hydroformylation of DCP. Figure 2 shows the principle of the membrane separation step. Within the module, the mixture of crude oxoaldehyde, toluene, free ligands, and the Rh catalyst complex coming from the reactor is parallel- pumped to the surface of the membrane. Only aldehyde and higher-boiling products pass through the membrane. The concentrate of Rh complex and ligands is recycled back to the reactor. 

J. Feldman and M. Orchin, Membrane-supported rhodium hydroformylation catalysts. /. Mol. Catal, 63 (1990) 213. 

Reetz et al.59 have introduced polypropylenimine (PPI) dendrimers as the core for building phosphine-coated constructs that can complex with Rh(COD) BF4, where COD = 1,5-cyclooctadiene, to instill the desired catalytic character. Hydroformylation of 1-octene with these metallodendrimers was shown to have turnover numbers that were comparable to those of monomeric analogs. It was pointed out that these catalysts could be easily recovered by means of membrane separation technology.60 Gong et al. have used water-soluble, phosphonated dendritic... 

Some further special technical aspects should be mentioned. The intensive mixture of the two liquid phases is an important condition for obtaining high reaction rates. This mixing can be achieved in bubble columns, tray columns or in stirred-tank reactors. In the few publications on industrially realized two-phase reactions the stirred tank reactor is always cited, but without detailed information on the stirring device. One further possible way to increase the mass transfer between the two liquid phases is by the influence of sonification. Cornils et al. applied this technique in the hydroformylation of hexene or diisobutene and found a considerable increase in the turnover numbers [93]. Another possibility for increasing the mass transfer may be by the use of microemulsions and micellar systems [94], which can be reached by addition of certain surfactants. This aspect is discussed in Sections 3.2.4 and 4.5. The separation of catalyst compounds in two-phase systems in combination with membranes has been studied recently by Muller and Bahrmann [95],... 

Mutual Optimization of Hydroformylation and Membrane Filtration Step... 

Tab. 3 Hydroformylation of dicyclopentadiene and membrane separation of the catalyst system (with variation of amines). 

Membrane separation conditions feed, reaction product of hydroformylation of dicyclopentadiene solvent, about 50% toluene membrane, UF-PA-5/PET 100 from the former Hoechst AG overflow, se 200 Lfr1 separation temperature, 40°C pretreatment of membranes in water at 80°C for 10 min transmembrane pressure, 10 bar. 

Test runs with low P/Rh ratios and Rh concentrations while hydroformylating DCP showed excellent membrane separation results but decreasing activity data. This failure in the optimization approach of the hydroformylation and membrane separation step without regard to long-term stability again underlines a basic problem in catalyst development, the coincidental consideration of different contradictory circumstances. A second series with a high P/Rh ratio of 100 was performed with the same catalyst system and butyraldehyde from the hydroformylation of propene as feed. 

This replication series (for details cf. [38]) showed overall excellent results in the hydroformylation as well as in the membrane separation step. With a high P/Rh ratio no deactivation was observed. The activity of the catalyst remained sufficient and high amounts of permeate with stable flow rates of 10 to 12 Lnr2h in the... 

Hydroformylation conditions feed, propene, 270 bar temperature, 125 °C P/Rh ratio, 2 1 reaction time, 2 h Rh concn., 20 ppm membrane separation conditions feed, butyraldehyde membrane type, UF-PA-5/PET 100 from the former Hoechst AG pressure, 15 bar temperature, 40°C amount of permeate, 1st stage, 91-84% 2nd stage, 95-92%. 

Just as with hydrogenations, hydroformylations, etc., a major reason for performing catalytic oxidations in water is to provide for facile recovery of the catalysts, by simple phase separation, from the product which is in an organic phase. However, many examples of catalytic oxidations in water involve water-soluble substrates and/or products. In this case catalyst recovery can be facilitated by using polymeric water-soluble ligands (see below) in conjunction with separation with an ultrafiltration membrane [3] or by other measures. 

Further progress is expected from new developments and combinations of processes. Thus, it would be possible to make the disposal of the gaseous (and highly pure) waste gas streams (residual propane content of the propylene feed) cost-effective and a source of electric power by connection to novel, compact, membrane fuel cells. Potential synergisms would also occur in the operating temperature of the cells (medium-temperature cells at 120 °C using the residual exothermic heat of reaction from the oxo reaction), the membrane costs by means of combined developments (e.g., for membrane separations of the catalysts [22]), and also in the development of the zero-emission automobile by the automotive industry. The combination of hydroformylation with fuel cells would further reduce the E-factor - thus approaching a zero-emission chemistry. ... 

---