Asymmetric Hydroformylation of Styrene in Supercritical Carbon

The laboratory and supported catalysts were prepared by Nick Kingsley, ..... P. G.; Macomber, D. W.; Willging, S. M. Isolation and Characterization of...
1 downloads 0 Views 167KB Size
Download PDF
1324

Ind. Eng. Chem. Res. 2006, 45, 1324-1330

Asymmetric Hydroformylation of Styrene in Supercritical Carbon Dioxide Angela M. Kleman and Martin A. Abraham* Department of Chemical and EnVironmental Engineering, The UniVersity of Toledo, Toledo, Ohio 43606

Hydroformylation reactions in supercritical carbon dioxide can provide an environmentally conscious method of producing aldehydes for fine-chemical and pharmaceutical products. Asymmetric ligands, such as (R)BINAP, can be used to provide an enantioselective product. When these reactions are performed in environmentally benign solvents, such as supercritical carbon dioxide (scCO2), additional environmental benefits are derived, such as ease of recycling of the solvent and unconverted reactants and elimination of the need for organic solvents. Rhodium-based catalysts have been prepared through the solid-phase reaction of catalyst precursors in supercritical carbon dioxide and evaluated for the hydroformylation of styrene to produce 2-phenylpropionaldehyde. Triphenylphosphine and (R)-BINAP were tested as ligands, and their effects on the reaction were examined. The catalyst, formed in situ in supercritical carbon dioxide (scCO2), was found to promote the hydroformylation of styrene with enantiomeric selectivity when (R)-BINAP ligands were used. Introduction Hydroformylation reactions, in which an olefin reacts with carbon monoxide and hydrogen to produce an aldehyde in the presence of a catalyst, were first discovered by Otto Roelen in 1938. In the 1960s, Wilkinson began investigating phosphorusmodified rhodium catalysts;1 these catalysts were found to improve the regioselectivity of the hydroformylation reaction toward the formation of the linear product. Wilkinson’s research showed that rhodium provided chemo- and regioselective hydroformylation.1 The regioselectivity of the reaction, which is the ratio of linear to branched products, depends on the substrate being used. When styrene is used, the branched aldehyde should be produced in abundance according to the known electronic preference for internal hydroformylation.1 For a system in which unmodified rhodium-based precursors were used between 20 and 100 °C, van Leeuwen states that the formation of the linear aldehyde during styrene hydroformylation increased as temperature increased. van Leeuwen did not note selectivity changes for a phosphorus-modified rhodium-based system. Reinius and Krause observed regiocontrol during the hydroformylation of styrene with an in situ formed rhodium-phosphorus complex;2 they also stated that the factors that most greatly affected regioselectivity were the natures of both the substrate and the modifying ligand. Phosphite ligands were examined for hydroformylation in the 1960s, but triphenylphosphine (PPh3) continued to be the ligand of choice; the interest in phosphite ligands arose again in the 1980s because workers at Union Carbide made bulky monophosphites with better stability, and these materials were found to give very high rates (with high regioselectivity to linear aldehydes).3 Eastman’s BISBI is an example of a diphosphine ligand that became very popular for rhodium hydroformylation because of the high regioselectivity obtained and the ability to recover the ligand following reaction.4 Although phosphorusbased ligands often provide desired selectivity, they also decrease the rate of reaction.5 Rhodium-BINAP complexes are often studied for their enantioselective hydrogenation abilities,6 although other ligands * To whom correspondence should be addressed. E-mail: [email protected].

have been used successfully to perform asymmetric hydroformylation reactions. For example, a Rh-BINAPHOS complex supported on a highly cross-linked polystyrene was found to provide nearly 100% regioselectivity and 82% enantiomeric excess (ee) when styrene was used as the substrate in the absence of organic solvents.7 Other olefins, including vinyl acetate, 1-alkenes, and fluorinated alkenes, were also successfully converted into the corresponding isoaldehydes with high enantiomeric excess. In recent years, diphosphite and phosphinephosphite ligands have been developed to provide greater enantioselectivity and activity for selective hydroformylation.8 Supercritical carbon dioxide has previously been evaluated in efforts to replace organic solvents that have traditionally been used in the hydroformylation reaction.9 Carbon dioxide is naturally abundant; it is nontoxic, readily available in large quantities, and relatively inexpensive. Because carbon dioxide is not consumed by the reaction, it can be recycled and reused for future applications. Supercritical carbon dioxide exhibits favorable properties of both a liquid and a gas; it has high miscibility with reactant gases, no liquid/gas phase boundary, very low viscosity, very high diffusivity, and high compressibility.10 These solvent properties provide a suitable environment in which to conduct the hydroformylation reaction, which minimizes their environmental implications and allows for a more sustainable process to produce aldehyde products. Additionally, aldehyde formation is often slower in traditional solvents, such as toluene, than in scCO2.11 Early work of Lin and Akgerman demonstrated that the rhodium catalyst [(COD)Rh(Et-DuPHOS)][BARF] {BARF ) tetrakis[3,5-bis(trifluoromethyl)phenyl]borate} could be used to promote the regioselective hydroformylation of styrene in scCO2 and that pressure could be used to influence the selectivity of the reaction.12 In later work, rhodium catalysts bound to a fluoroacrylate copolymer backbone through phosphine ligands were synthesized, shown to be soluble in scCO2, and demonstrated to be active for the regioselective hydroformylation of styrene at two different temperatures (323 and 348 K) and three different pressures (172, 207, and 241 bar).11 A catalyst prepared from [(OC)2Rh(acac)] and (R,S)-BINAPHOS in CH2Cl2 was used to promote the asymmetric hydroformylation of styrene in the presence of compressed CO2 to give appreciable asymmetric induction [ee ) 66% (R)] under conditions close to the

10.1021/ie0509701 CCC: $33.50 © 2006 American Chemical Society Published on Web 01/19/2006

Ind. Eng. Chem. Res., Vol. 45, No. 4, 2006 1325

critical properties of pure CO2, but very low ee’s were obtained at high CO2 densities.13 Additional research by Matsuda et al. indicates that enantioselectivity can be increased by decreasing the pressure while maintaining a constant temperature.14 There is a growing market for specialty chemicals produced via asymmetric hydroformylation of alkenes with applications of the resulting aldehydes in the production of pharmaceuticals, fragrances, and pesticides.15,16 Chiral amino acids and peptides are available via this route, and hydroformylation of vinylarenes is a convenient entry to substituted 2-arylpropanoic acids used as precursors to analgesics such as ibuprofen, ketoprofen, and naproxen. In particular, the branched aldehydes obtained from the hydroformylation of styrene can be converted to various nonsteroidal antiinflammatory agents.17 The high value of these specialty chemicals, compared to bulk aldehydes such as butanal, suggests that an environmentally benign process developed using scCO2 as the reaction medium might have economic incentives not available in the bulk-chemicals industry. In fact, finechemicals production, which often has poor atom economy and results in the generation of a greater mass of waste than mass of product, would benefit from new environmentally benign syntheses. We have previously studied homogeneously and heterogeneously catalyzed 1-hexene hydroformylation in supercritical carbon dioxide (scCO2) and have shown that the reaction rate and the regioselectivity can be varied by changes in pressure at constant temperature.18 The current study examines the enantiomeric hydroformylation of styrene to produce 2-phenylpropionaldehyde using a catalyst that is prepared in situ within the reactor, or bound to a solid support, without any added solvent. By performing the reaction using supercritical carbon dioxide and a nonsoluble or bound enantioselective catalyst, we demonstrate a hydroformylation reaction with a reusable catalyst that minimizes environmental impacts and improves economic viability. Experimental Section Hydroformylation of styrene in supercritical carbon dioxide was used to evaluate various methods of in situ catalyst preparation with variations in rhodium-to-phosphorus ratios. The 300-mL stirred batch reactor, described previously9,18 was manufactured by Autoclave Engineers. It is operable to 370 bar at 344 °C. The catalysts used in all experiments were solids in the form of granules or powders. A 5-µm nylon mesh envelope (Spectrum Labs) was created by cutting two small rectangles of mesh; three of the rectangle’s edges were fused together using a variablewattage soldering iron. A measured amount of catalyst or catalyst precursors was added to the pouch through the open end. The remaining edge was then soldered to secure the catalyst inside the pouch as a preventative measure to reduce catalyst disbursement throughout the reactor and to enable catalyst recovery. The catalyst envelope was clamped between two heavy-gauge screens mounted on disks located on the stirring shaft. The screens and disks were configured so that two catalyst envelopes could be used during reactions; however, only one catalyst envelope was used during any reaction. After the reactor had been sealed with the catalyst envelope attached to the stir column as previously described, the system was powered on so that the system pressure and temperature could be observed. The heating jacket was set to provide a temperature slightly greater than room temperature, so as to minimize temperature-fluctuation effects during pressurization. The reaction vessel was flushed using CO2 with a maximum

pressure no greater than 34 bar. The CO2 was purged from the reaction vessel via the sample loop and discharged into the hood; the purging allowed the pressure inside the reactor to drop to atmospheric pressure. An equimolar mixture of carbon monoxide and hydrogen was added to a sample cylinder (Whitey) from a gas cylinder (AirGas); the sample cylinder was pressurized to 95 bar. The sample cylinder was then weighed, and its contents were brought into equilibrium with the sealed reaction vessel. After equilibration, the sample cylinder was reweighed to determine the amount of carbon monoxide and hydrogen that had been added to the reactor. The temperature of the heating jacket was then incrementally increased until the desired temperature was obtained; this process was carefully observed so as to minimize overshoot, with a maximum overshoot no greater than 10 °C above the desired temperature. After the vessel was heated to the desired temperature, carbon dioxide could be added to adjust the pressure in the reaction vessel. At this point, the system underwent a 24-h period in which in situ mixing and system stabilization occurred. This pretreatment period was also designed to allow time for a solid-phase reaction between the catalyst precursors to yield the active catalyst; the variation of pretreatment conditions describes their effect on catalyst performance. Most of the catalysts used were produced in the reaction vessel. Catalysts were generally produced in situ from catalyst precursors, [RhCl(C8H12)]2 and the phosphorus component, either (R)-BINAP or triphenylphosphine, that were placed in the mesh envelope that was sealed and affixed to the stir bar. The reactor was sealed and then brought to the specified temperature and pressure with a desired composition of CO and H2 or CO, H2, and CO2. To allow for in situ catalyst preparation, the contents of the reactor were maintained at the given temperature and pressure with mixing for 24 h. [RhCl(C8H12)]2 was prepared in the Department of Chemistry at The University of Toledo using the procedure of Giordano and Crabtree.19 (R)-BINAP was purchased from Strem Chemicals with a stated purity of 98% and used as received. Triphenylphosphine was purchased from Aldrich Chemical Co., Inc., at a purity of 99% and used as received. Several other catalysts (described as “prepared in the laboratory”) were prepared in the Department of Chemistry at The University of Toledo and characterized by 31P solid-state NMR spectroscopy. These catalysts included Wilkinson’s catalyst,20 a solid catalyst with a Rh/P ratio of 1:3,21 and a solid rhodium-BINAP complex.22 At the conclusion of the 24-h period, the reactor was pressurized using liquid CO2 to obtain a pressure approximately 35 bar lower than the desired reaction pressure. The reaction system was again allowed to stabilize, normally requiring no greater than 30 min to achieve stability. Known amounts of styrene and n-heptane (used as an internal standard) were added to a bypass section located within the feed line. After stability had been achieved, the styrene and n-heptane present in the bypass were pushed into the reactor using liquid CO2; this process continued until the desired reaction pressure was achieved, usually requiring no more than 3 min to obtain the desired pressure. Styrene and n-heptane were purchased from Sigma Aldrich Chemical Co. with purities of 99% and greater than 99%, respectively, and used as received. The start of the experiment was marked from the time the contents of the bypass line were flushed into the reactor. Sampling began at 15 min; additional samples were taken generally at 30, 45, 60, 75, 90, and 120 min, with hourly

1326

Ind. Eng. Chem. Res., Vol. 45, No. 4, 2006

Table 1. Reaction Specifications pretreatment conditions rxn code

component(s)

1F 2G 3H 4J 5K 6L 7M 8N 9P 10Q 11R 12S 13T 14U 15V

[RhCl(C8H12)]2 and PPh3 [RhCl(C8H12)]2 and BINAP [RhCl(C8H12)]2 Wilkinson’s catalyst C96H80Cl2P4Rh2 (PPh3) [RhCl(C8H12)]2 and PPh3 [RhCl(C8H12)]2 and PPh3 [RhCl(C8H12)]2 and PPh3 [RhCl(C8H12)]2 and PPh3 [RhCl(C8H12)]2 and PPh3 [RhCl(C8H12)]2 and PPh3 [RhCl(C8H12)]2 and BINAP C96H80Cl2P4Rh2 (BINAP) [RhCl(C8H12)]2 and PPh3 [RhCl(C8H12)]2 and BINAP

reaction conditions

Rh/P temp pressure temp pressure Ratio (°C) (bar) (°C) (bar) 1:8 1:4 N/A 1:3 1:3 1:34 1:8 1:34 1:8 1:8 1:2 1:4 1:2 1:8 1:8

90 90 90 90 90 50 50 50 50 90 90 90 90 90 50

45 45 45 45 45 137 137 137 137 137 45 45 45 137 137

90 90 90 90 90 75 90 90 75 90 90 90 90 75 75

171 171 171 171 171 205 171 171 205 171 171 171 171 205 171

samplings continuing for the next 4 h of the experiment. If additional samples were needed, they were taken at the 9-h mark and then every 6 or 12 h thereafter. The reaction’s progress was monitored in terms of styrene conversion and product formation by gas chromatography. Products were identified using a Hewlett-Packard 6980 gas chromatograph and a 5973 mass-selective detector. A chiral Supelco Beta Dex 225 column (30 m × 0.25 mm × 0.25 µm film, column number 16651-018) was used to achieve separation of the styrene reactant and the enantiomeric products. A temperature ramp was used, with an initial temperature of 40 °C for 5 min that was then increased at a rate of 10 °C/min until a temperature of 85 °C was reached; the total time needed to complete each sample analysis was 39.5 min. Samples were collected from the reactor by expansion through 1 mL of ethyl acetate in a 5-mL screw-cap vial. The sample was bubbled through the solvent. The organic components were trapped in the liquid solvent, whereas the CO, H2, and CO2 escaped; therefore, these components were not quantified. The known samples used in producing the calibrations were prepared using purchased materials of high purity. The enantiomeric 2-phenylpropionaldehyde products were purchased as a racemic mixture. Both enantiomeric products were obtained from the analysis; however, no effort was made to identify the individual enantiomers because the separate enantiomers were not found to be commercially available. Results Table 1 summarizes the experiments that were conducted. The components used include the following: [RhCl(C8H12)]2, triphenylphosphine (PPh3), and (R)-(+)-2,2′-bis(diphenylphosphino)-1,1′-binaphthyl (BINAP). The atomic ratio of rhodium to phosphorus is noted in the table, as it was varied for experiments to determine its effects on selectivity. The yields of all products were calculated using the calibration obtained from direct injection of standards into the GC. The styrene yield is the amount of reactant remaining in the reactor relative to the initial amount of styrene injected into the reactor; as a reactant, the yield of styrene decreases with time. Styrene and 2-phenylpropionaldehyde were the only products observed on a regular basis, with small amounts of ethylbenzene observed in some experiments. Although the sum of the yields should have been 1 in all cases, this was not actually achieved, which was ultimately attributed to the sampling process in relation to the calibration method. GC calibrations were obtained by mixing

Figure 1. Styrene conversion and yield of phenylpropionaldehyde product from experiments 2G and 12S using [RhCl(C8H12)]2 and BINAP and reaction at 90 °C and 171 bar, demonstrating reproducibility of experiments.

known quantities in a vial and running these known samples through the GC/MS. However, reaction products were obtained by expanding samples through an ethyl acetate solvent. The difference between the calibration and the sample was attributed to differences in the dissolution of n-heptane (used as an internal standard), styrene, and products from the gas phase into the ethyl acetate solvent as these species were expanded from the reactor. To provide consistent data for comparison purposes, a normalized yield was calculated according to the equation

yn,i )

yi ys + yp1 + yp2

(1)

where yn,i is the normalized yield of component i, yi is the yield of component i, ys is the yield of styrene, and yp1 and yp2 are the yields of the 2-phenylpropionaldehyde enantiomers. Replicate experiments (2G and 12S) were performed to examine the reproducibility of the experiments. The experiments were conducted with the following conditions: Rh/P ratio of 1:4, pretreatment conditions of 90 °C and 45 bar, and reaction conditions of 90 °C and 173 bar. Figure 1 shows the disappearance of styrene and the formation of one of the phenylpropionaldehyde products and illustrates that the results were reproducible within experimental limits. The styrene yield decreased steadily, indicating complete conversion around 40 h, and the product peak increased to a yield of 50%. Figure 2 compares the 31P MAS NMR spectra of pure triphenylphosphine and the catalyst remaining in the envelope after reaction 1F, in which [RhCl(C8H12)]2 and PPh3 were used to make a catalyst in situ. Because the peaks of PPh3 and the catalyst formed in situ were observed at different locations, it is believed that the catalyst was formed from triphenylphosphine and [RhCl(C8H12)]2 during the in situ pretreatment conditions. Additionally, the observed doublet peak centered at 3.4 ppm with JRhP ) 107.4 Hz is within the range reported for rhodium phosphine complexes. Therefore, it can be concluded that the catalysts using PPh3 for ligands were generated during in situ pretreatment. The effect of reaction temperature and pressure on the conversion of styrene is shown in Figure 3. Experiments 7M (90 °C and 171 bar) and 9P (75 °C and 205 bar) used [RhCl(C8H12)]2 and PPh3, a Rh/P ratio of 1:8, and pretreatment conditions of 50°C and 137 bar, whereas experiments 6L (90 °C and 171 bar) and 8N (75 °C and 205 bar) used [RhCl(C8H12)]2 and PPh3, a Rh/P ratio of 1:34, and pretreatment conditions of 50°C and 137 bar. Regardless of the pretreatment

Ind. Eng. Chem. Res., Vol. 45, No. 4, 2006 1327

Figure 2. Solid-state 31P MAS NMR spectra of catalyst residues from experiment 1F with peaks at 34.0 and 32.7 ppm, compared to that of pure triphenylphosphine showing a peak at -7.1 ppm.

Figure 3. Effect of reaction temperature and pressure on styrene conversion. Catalysts prepared from in situ reaction of [RhCl(C8H12)]2 and PPh3, with Rh/P ) 1:8 or 1:34, at pretreatment conditions of 90 °C and 131 bar.

conditions, the reactions at 90 °C go to completion much faster than the reactions at the lower temperature. The reactions at 90 °C achieved nearly complete conversion in less than 10 h, whereas the reactions at the lower temperature required nearly 60 h to obtain a similar yield. At 75°C and the higher phosphorus loading of reaction 6L, the reaction proceeded very slowly, achieving only about 20% conversion after 80 h. The presence of phosphines during hydroformylation reactions is believed to decrease the rate at which products are produced;23 however, regioselectivity increases as the amount of phosphorus per rhodium atom increases. To understand these effects during reaction in supercritical CO2, a series of experiments were performed in which the Rh/P ratio was varied under otherwise identical conditions. The effect of phosphine loading is explored in Figure 4, in which experiments 1F (1:8), 11R (1:2), and 3H ([RhCl(C8H12)]2 only) compare the styrene conversion obtained as the Rh/P ratio was varied among catalysts generated using [RhCl(C8H12)]2 and PPh3, with in situ preparation conditions of 90 °C and 45 bar and reaction conditions of 90 °C and 171 bar. Additional comparisons using experiments 6L (1:34) and 9P (1:8), conducted with in situ preparation conditions of 50 °C and 137 bar and reaction

Figure 4. Effect of phosphorus loading on styrene conversion. [RhCl(C8H12)]2 and PPh3 with in situ pretreatment conditions of 90 °C and 45 bar and reaction conditions of 90 °C and 171 bar or in situ preparation conditions of 50 °C and 137 bar and reaction conditions of 75 °C and 205 bar.

conditions of 75 °C and 205 bar (reproduced from Figure 3), further illustrate the effects of Rh/P ratio variances. At both sets of conditions, the reaction with the greatest phosphine loading had the lowest reaction rate. It is interesting that the experiment using only [RhCl(C8H12)]2 showed a significant induction time before the catalyst became active, indicating the importance of the phosphorus ligand in activating the catalyst for hydroformylation. It can also be seen from the data of Figure 4 that the catalyst prepared at 50 °C and 137 bar had a greater styrene conversion than the catalyst prepared at 90 °C and 45 bar (without added CO2), even though the latter catalyst was tested at a higher reaction temperature. This phenomenon is further explored in Figure 5, which depicts the effects of pretreatment conditions; experiments 1F (90 °C and 45 bar), 7M (50 °C and 137 bar), and 10Q (90 °C and 137 bar) were conducted using [RhCl(C8H12)]2 and PPh3, a Rh/P ratio of 1:8, reaction conditions of 90 °C and 171 bar, and variable in situ preparation conditions. These graphs show that the experiment using a higher pretreatment pressure provided a higher rate of reaction among

1328

Ind. Eng. Chem. Res., Vol. 45, No. 4, 2006

Figure 5. Comparison of the performance of catalysts prepared with varying in situ pretreatment conditions. All experiments were performed with [RhCl(C8H12)]2 and PPh3, a Rh/P ratio of 1:8, and reaction conditions of 90 °C and 171 bar.

Figure 6. Styrene conversion obtained from catalyst prepared in situ with supported catalysts. All experiments had identical pretreatment conditions of 90 °C and 45 bar and reaction conditions of 90 °C and 171 bar.

experiments with a 90 °C pretreatment temperature. Additionally, among the reactions with the pretreatments at higher pressures, the reaction with the catalyst prepared at the lower temperature provided a higher rate of reaction. As the catalyst pretreatment conditions moved away from the critical conditions of CO2, a successively poorer catalyst was formed. This is consistent with the previous observations of Leitner, who reported improved performance from the catalyst when used for reaction at conditions close to the critical point of CO2.15 The presence of CO2 during pretreatment might have provided a solvent environment in which the catalyst precursors effectively interacted, thereby creating a good catalyst. However, in the absence of CO2, the catalyst formation reaction is slow, and a poor-quality catalyst was produced. One of the goals of the present work was to compare the performance of in situ prepared hydroformylation catalysts with a rhodium catalyst grafted onto a phosphinated silica support. Figure 6 compares the conversion obtained from catalysts with the lowest atomic ratios, pretreatment conditions of 90 °C and 45 bar, and reaction conditions of 90 °C and 171 bar; these graphs show results from experiments 4J (Wilkinson’s catalyst), 5K (Wilkinson’s catalyst immobilized on a silica support), and 11R ([RhCl(C8H12)]2 and PPh3). Wilkinson’s catalysts and the immobilized catalyst have Rh/P ratios of 1:3; the experiment in which the catalyst was made in situ has a Rh/P ratio of 1:2. The catalyst made in situ performed somewhat better than the two catalysts produced outside of the reactor, both of which performed similarly. It is possible that the catalyst prepared during the in situ pretreatment was more highly soluble in the

Figure 7. Performance of catalysts prepared in situ using the BINAP ligand (2G) and a Rh/P loading of 1:4 with a Rh-BINAP catalyst produced in the laboratory (13T) and with Rh(PPh3) catalysts of similar phosphorus loadings prepared in situ (Rh/P ) 1:8, 1F; Rh/P ) 1:2, 11R). All reactions have in situ pretreatment conditions of 90 °C and 45 bar and reaction conditions of 90 °C and 171 bar.

supercritical fluid, leading to the greater performance of the catalyst, although the lower phosphorus loading also might have led to enhanced catalyst performance. Figure 7 compares the performance of catalysts prepared in situ using the BINAP ligand (2G) and a Rh/P loading of 1:4 with a Rh-BINAP catalyst produced in the laboratory (13T), and it compares these results to two catalysts prepared in situ from [RhCl(C8H12)]2 and PPh3 with phosphorus loadings Rh/P ) 1:8 (1F) and Rh/P ) 1:2 (11R). All reactions had in situ pretreatment conditions of 90 °C and 45 bar and reaction conditions of 90 °C and 171 bar. Both catalysts prepared with BINAP ligands performed similarly to the catalyst prepared with the PPh3 ligand at the higher loading. The Rh/P sample with low PPh3 loading had the highest rate of reaction; as previously discussed, the presence of phosphorus decreases the reaction rate. The catalyst prepared with BINAP in the laboratory appeared to have a short induction period, but otherwise performed roughly equivalently to that prepared in situ. Thus, it appears that the BINAP ligand performs comparably to the PPh3 ligand (in terms of activity) and that in situ production is a viable route for the preparation of the active catalyst from [RhCl(C8H12)]2 and BINAP. The use of BINAP as a ligand would be expected to impart enantioselectivity to the styrene hydroformylation reaction. To determine whether enantioselectivity could be obtained in these experiments, the ratio of GC integration from the first enantiomeric product to the second product was calculated. Figure 8 compares this ratio for two experiments using BINAP ligand (one conducted at 90 °C and 171 bar using pretreatment conditions of 90 °C and 45 bar and an optimized experiment in which the pretreatment was conducted at 50 °C and 171 bar and the reaction was performed at 75 °C) with the results from two baseline experiments using PPh3 as the ligand (conducted at 90 °C and 171 bar using pretreatment conditions of 90 °C and 45 bar). Reaction 2G, performed using BINAP ligand at 90 °C, showed some tendency to preferentially form the second product at short reaction times. Although there is substantial scatter in the data at short reaction times (corresponding to low conversion), the product peak ratio from reaction 1F, performed with PPh3 ligand, is consistently closer to 1 than that obtained from reaction 2G, performed with the BINAP ligand. Because the PPh3 ligand should provide no enantioselectivity, we take this result as a preliminary suggestion that the BINAP ligand did yield the expected enantioselectivity. The loss of enanti-

Ind. Eng. Chem. Res., Vol. 45, No. 4, 2006 1329

effective for stereoselective hydroformylation as required in pharmaceuticals production. Acknowledgment The laboratory and supported catalysts were prepared by Nick Kingsley, a graduate student in the laboratory of Prof. Mark Mason of the Department of Chemistry at The University of Toledo. Characterization of these materials was also completed with the assistance of Nick Kingsley and The University of Toledo Instrumentation Center. Financial support for this project was provided by the University of Toledo Department of Chemical and Environmental Engineering. Literature Cited Figure 8. Comparison of the product peak ratio for two experiments using BINAP ligand (one conducted at 90 °C and 171 bar using pretreatment conditions of 90 °C and 45 bar and an optimized experiment in which the pretreatment was conducted at 50 °C and 171 bar and the reaction was performed at 75 °C) with the results from two baseline experiments using PPh3 as the ligand (conducted at 90 °C and 171 bar using pretreatment conditions of 90 °C and 45 bar).

oselectivity at longer reaction times suggests racemization of the product or deactivation of the selective catalyst, similar to that previously observed by Shibahara et al.8 For reaction 15V, performed at a lower reaction temperature previously shown to favor enantioselectivity,28 the ratio of the integration of peak 1 to that of peak 2 remained near 0.8 throughout the course of the experiment; this ratio corresponds to an enantiomeric excess of approximately 10-15%. Judging from previous work on asymmetric hydroformylation,15 the reactions in which BINAP ligands were used should have provided enantioselective products, whereas the reactions using triphenylphosphine ligands should not have produced enantioselectivity. Despite the appearance of only mild enantioselectivity with the use of BINAP, these results were seen to be similar to previous results reported in the literature. Hegedus et al.24 reported an enantiomeric excess of 13% for conditions most similar to those used in the current work; this is compatible with the 10-15% enantiomeric excess observed in the current work. Hegedus et al. noted a high enantioselectivity at fairly low reaction temperature, much lower than that used in the current work, and decreasing enantioselectivity as the temperature increased. Thus, it is concluded that the mild enantioselectivity observed herein is consistent with previous asymmetric hydroformylation results. Conclusions Hydroformylation catalysts were successfully produced in scCO2 in situ using [RhCl(C8H12)]2 and triphenylphosphine or BINAP through reaction in supercritical CO2, as determined both by NMR analysis and by the formation of reaction products. The variation in catalyst performance brought on by varying catalyst preparation conditions further confirms that active catalyst was prepared through the in situ reaction. The most active catalyst was formed by pretreatment at conditions close to the critical point of CO2. The catalysts prepared in situ performed comparably to those prepared in the laboratory and to catalysts that were immobilized on solid supports. Catalysts prepared from BINAP ligands demonstrated mild enantioselectivity during styrene hydroformylation, with enantiomeric excesses similar to those found in the literature, demonstrating that the formation of a solid catalyst in supercritical CO2 is

(1) van Leeuwen, P. W. N. M.; Claver, C. Rhodium Catalyzed Hydroformylation; Kluwer Academic Publishers: Boston, 2000. (2) Reinius, H. K.; Krause, A. O. I. Hydroformylation of Functional Alkenes with Heterodonor Phosphine Rhodium Catalysts: Substrate or Ligand Directed Regioselectivity? Catal. Lett. 2000, 70, 149-154. (3) Abatjoglou, A. G.; Bryant, D. R. Ionic phosphites and their use in homogeneous transition metal catalyzed processes. U.S. Patent 5,059,710, 1991. (4) Buhling, A.; Elgersma, J. W.; Nkrumah, S.; Kamer, P. C. J.; van Leeuwen, P. W. N. M. Novel amphiphilic diphosphines: Synthesis, rhodium complexes, use in hydroformylation and rhodium recycling. J. Chem Soc., Dalton Trans. 1996, 10, 2143-2154. (5) van der Slot, S. C.; Duran, J.; Luten, L.; Kamer, P. C. J.; van Leeuwen, P. W. N. M. Rhodium-Catalyzed Hydroformylation and Deuterioformylation with Pyrrolyl-Based Phosphorus Amidite Ligands: Influence of Electronic Ligand Properties. Organometallics 2002, 21, 3873-3883. (6) Bunten, K. A.; Farrar, D. H.; Poe, A. J.; Lough, A. Stoichiometric and Catalytic Oxidation of BINAP by Dioxygen in Rhodium(I) Complex. Organometallics 2002, 21, 3344-3350. (7) Shibahara, F.; Nozaki, K.; Hiyama, T. Solvent-free asymmetric olefin hydroformylation catalyzed by highly cross-linked polystyrene-supported (R,S)-BINAPHOS-Rh(I) complex. J. Am. Chem. Soc. 2003, 125 (28), 8555-8560. (8) Dieguez, M.; Pamies, O.; Claver, C. Recent advances in Rh-catalyzed asymmetric hydroformylation using phosphite ligands. Tetrahedron: Asymmetry 2004, 15 (14), 2113-2122. (9) Tack, T. T. Performance of Rhodium and Platinum Supported Catalysts for Hydroformylation in Supercritical Carbon Dioxide. M.S. Thesis, The University of Toledo, Toledo, OH, 2003. (10) Koch, D.; Leitner, W. Rhodium-Catalyzed Hydroformylation in Supercritical Carbon Dioxide. J. Am. Chem. Soc. 1998, 120, 13398-13404. (11) Kani, I.; Flores, R.; Fackler, J. P.; Akgerman, A. Hydroformylation of styrene in supercritical carbon dioxide with fluoroacrylate polymer supported rhodium catalysts. J. Supercrit. Fluids 2004, 31 (3), 287-294. (12) Lin, B.; Akgerman, A. Styrene hydroformylation in supercritical carbon dioxide: Rate and selectivity control. Ind. Eng. Chem. Res. 2001, 40 (4), 1113-1118. (13) Kainz, S.; Leitner, W. Catalytic asymmetric hydroformylation in the presence of compressed carbon dioxide. Catal. Lett. 1998, 55 (3-4), 223-225. (14) Matsuda, T.; Kanamaru, R.; Watanabe, K.; Harada, T.; Nakamura, K. Control on Enantioselectivity with Pressure for Lipase-Catalyzed Esterification in Supercritical Carbon Dioxide. Tetrahedron Lett. 2001, 42, 8319-8321. (15) Jessop, P. G.; Ikariya, T.; Noyori, R. Homogeneous Catalysis in Supercritical Fluids. Chem. ReV. 1999, 99, 475. (16) Simmons, M. S. The Role of Catalysts in Environmentally Benign Synthesis of Chemicals. In Green Chemistry: Designing Chemistry for the EnVironment; Anastas, P. T., Williamson, T. C., Eds.; ACS Symposium Series 626; American Chemical Society: Washington, DC, 1996; Chapter 10. (17) Ojima, I. Catalytic Asymmetric Synthesis; Wiley-VCH: New York, 2000. (18) Tadd, A. R.; Marteel, A.; Mason, M. R.; Davies, J. A.; Abraham, M. A. Hydroformylation of 1-hexene in supercritical carbon dioxide using a heterogeneous rhodium catalyst. 1. Effect of process parameters. J. Supercrit. Fluids 2003, 25 (2), 183-196. (19) Giordano, G.; Crabtree, R. H. Di-µ-chloro-bis(η-4-1,5-cyclooctadiene) dirhodium(I). Inorg. Synth. 1990, 28, 88-90.

1330

Ind. Eng. Chem. Res., Vol. 45, No. 4, 2006

(20) Gassman, P. G.; Macomber, D. W.; Willging, S. M. Isolation and Characterization of Reactive Intermediates and Active Catalysts in Homogeneous Catalysis. J. Am. Chem. Soc. 1985, 107, 2380-2388. (21) Marteel, A. E. Development of Heterogeneous Catalysts for Hydroformylation of 1-Hexene in Supercritical Carbon Dioxide. Ph.D. Dissertation, The University of Toledo, Toledo, OH, 2003. (22) Bunten, K. A.; Farrar, D. H.; Poe, A. J.; Lough, A. Stoichiometric and Catalytic Oxidation of BINAP by Dioxygen in Rhodium(I) Complex. Organometallics 2002, 21, 3344-3350. (23) Bensaid, L. F. Rhodium-Catalyzed Hydroformylation of 1-Hexene Using Acidic Phosphines and Synthesis of a New Phosphine Ligand for

Use in Supercritical Carbon Dioxide. M.S. Thesis, The University of Toledo, Toledo, OH, 2002. (24) Hegedus, C.; Madarasz, J.; Gergely, I.; Szollosy, A.; Monsees, A.; Riermeier, T. Bakos, J. Electronic and Steric Effects of Ligands as Control Elements for Rhodium-Catalyzed Asymmetric Hydroformylation. Part 3: Highly Active Hydroformylation of Styrene. Tetrahedron: Asymmetry 2004, 15, 2507-2513.

ReceiVed for reView August 25, 2005 ReVised manuscript receiVed December 14, 2005 Accepted December 15, 2005 IE0509701