8480
Ind. Eng. Chem. Res. 2007, 46, 8480-8489
Environmentally Benign Catalytic Hydroformylation-Oxidation Route for Naproxen Synthesis Kalpendra B. Rajurkar, Sunil S. Tonde, Mahesh R. Didgikar, Sunil S. Joshi, and Raghunath V. Chaudhari* Homogeneous Catalysis DiVision, National Chemical Laboratory, Pune 411 008, India
Hydroformylation of 6-methoxy-2-vinylnaphthalene (MVN), using homogeneous Rh(CO)2(acac) as a catalyst and a chelating bidentate ligand (1,2-bis-(diphenylphosphino) ethane, dppe), followed by oxidation of the product (2-(6-methoxynaphthyl) propanal, 2-MNP) has been studied as an alternative route for the synthesis of D,L-naproxen. The feasibility of the MVN hydroformylation route has been demonstrated, and a detailed study has been reported on the key hydroformylation step. The roles of the catalyst, ligands, and solvents, as well as the effect of reaction conditions on the reaction rate and regioselectivity of the product 2-MNP, have been investigated. With Rh(CO)2(acac) as a catalyst and dppe as a ligand, >98% selectivity to 2-MNP (an important precursor to D,L-naproxen) has been achieved. A possible mechanism to explain the variation in regioselectivity with Rh(CO)2(acac) as a catalyst and dppe as a ligand has been discussed. The kinetics of the hydroformylation step has been investigated and a rate equation has been proposed. The second step in the proposed route for naproxensthe oxidation of 2-MNP to 2-(6-methoxynaphthyl) propanoic acid (2-MNPA, or naproxen)shas been studied using Na2WO4 as a catalyst and tetrabutyl ammonium hydrogen sulfate (TBAHS) as the phase-transfer catalyst with H2O2 as the oxidant for the first time. Screening of the catalysts that consisted of the early transition metals, such as salts of tungsten, vanadium, and molybdenum showed that Na2WO4 gives the best performance for the oxidation step with >80% selectivity to 2-(6-methoxynaphthyl)propanoic acid (2-MNPA/naproxen). This study would be valuable in developing a new environmentally benign route for naproxen synthesis. Introduction Profens, which are 2-arylpropionic acids, have significant importance, because of their application as nonsteroidal antiinflammatory drugs (NSAIDs).1 Among the various derivatives of this class, ibuprofen and naproxen represent the most important examples. The commercial processes for ibuprofen recently have undergone significant changes to replace the stoichiometric synthetic organic routes by environmentally benign catalytic routes.2 The recently developed three-step catalytic route for ibuprofen, using catalytic acylation, hydrogenation, and carbonylation, represents one of the best examples of the use of catalysis for cleaner processes in pharmaceuticals.3 Naproxen is another important drug in this category, which is currently manufactured by multistep stoichiometric synthetic routes: (a) the Syntex process, starting with β-naphthol and involving stoichiometric bromination, methylation, and alkylmetal coupling reactions to yield naproxen; (b) the Zambon process, involving acylation of nerolin (2-methoxynaphthalene), ketalization, bromination, hydrolysis, and reductive cleavage as the key steps; and (c) asymmetric hydrogenation of 6-methoxy naphthacrylic acid, using the Ru-(S)-BINAP catalyst.1 Attempts toward the direct synthesis of chiral naproxen via a chiral pool using (S)-lactate, and asymmetric hydroformylation followed by oxidation, were also made. These routes suffer from drawbacks such as the use of hazardous reagents and the generation of undesired waste that consists of inorganic salts. Therefore, it is most desirable to develop an environmentally benign catalytic route for the synthesis of naproxen. In this paper, we report a detailed study on the two-step hydroformylation-oxidation route for the synthesis of D,L-naproxen (2* To whom correspondence should be addressed. Tel.: +91-202590-2770. Fax: +91-20-2590-2621. E-mail address: rv.chaudhari@ ncl.res.in.
MNPA) (see Scheme 1). As a first step, the hydroformylation of 6-methoxy-2-vinylnaphthalene (MVN) was investigated using a Rh(CO)2(acac) catalyst to understand the role of ligands, reaction conditions, and kinetics. In the second step, the oxidation of 2-MNP with H2O2 as the oxidant has been investigated using Na2WO4, Na2MoO4, Na2VO4, PTA, and PMA as catalysts. Experimental Section 1. Materials. RhCl3·3H2O and the phosphine ligands were procured from Aldrich (St. Louis, MO) and used as such without further purification. Rh(CO)2(acac) was prepared according to the literature procedure.4 Na2WO4, tetrabutyl ammonium hydrogen sulfate (TBAHS), oxalic acid, and all the solvents were procured from SD Fine Chemicals (Mumbai, India). CO with 99.9% purity (Matheson, USA) and H2 with 99% purity (Industrial Oxygen Company, Mumbai, India) were used as received, and syngas mixture in the required CO/H2 ratio was prepared in a reservoir. 2. Hydroformylation Experiments. The reactions were performed in a 50-mL high-pressure stainless-steel reactor (Parr Instrument Co., USA) that was equipped with a pressure transducer, automatic temperature control, and magnetic drive with variable speed. A reservoir filled with syngas (CO:H2 ) 1:1) was connected to the reactor via a constant pressure regulator. This enabled continuous feeding of the syngas from the reservoir, as per consumption in the reactor, while maintaining a constant pressure in the reactor. In a typical experiment, the reactants and catalyst were charged into the reactor and the reactor was flushed with nitrogen and syngas. Following this, the reactor contents were heated to the desired temperature under low stirring (200 rpm). Once the temperature was attained, the stirring was stopped and the syngas was pressurized as required
10.1021/ie0700866 CCC: $37.00 © 2007 American Chemical Society Published on Web 05/10/2007
Ind. Eng. Chem. Res., Vol. 46, No. 25, 2007 8481 Scheme 1. Synthesis of D,L-Naproxen via a Hydroformylation-Oxidation Route
Scheme 2. Plausible Mechanism for Rhodium-Catalyzed MVN Hydroformylation to 2-MNP and 3-MNP
into the reactor. The reaction was started by increasing the agitation speed to 1000 rpm. During the course of the reaction, samples were withdrawn periodically and analyzed by gas chromatography (GC) for reactant and products. For the hydroformylation reactions, a syngas (CO:H2 ) 1:1) reservoir was used as a source, which was connected to the reactor through a constant-pressure regulator. In this mode, we maintained a constant pressure inside the reactor and the pressure drop in the reservoir was measured as a function of time. The synthesis gas consumption-time profiles were evaluated from these observations. The gas absorption data were collected by reading the pressure drop in the reservoir as a function of time. For kinetic measurements, the reactions were conducted for a fixed time duration, whereas, for the screening studies, the reactions were conducted to high levels of conversion of MVN. At the end of the reaction, the autoclave was cooled and final samples were analyzed for reactants and products.
3. Oxidation Experiments. Oxidation of 2-(6-methoxy-2naphthyl) propanal (2-MNP) was conducted in a 15-mL twonecked jacketed reactor using methyl ethyl ketone (MEK) as a solvent. In a typical experiment, desired amounts of 2-MNP, MEK, Na2WO4 (as a catalyst), and TBAHS (as a phase-transfer catalyst) and the oxidant (30% H2O2) were mixed and stirred at constant temperature, using a magnetic stirrer. 2-MNP was analyzed by GC, and the products were analyzed by highperformance liquid chromatography (HPLC). The experiments were conducted for a duration of 5 h with different sets of initial conditions. 4. Analytical Methods. The products of the reactions were identified using gas chromatography-mass spectroscopy (GCMS) analysis on an Agilent Series 6890N gas chromatograph that was equipped with a model 5973N mass-selective detector. Liquid samples were analyzed for the quantification of 2-MNP on a Hewlett-Packard 6890 Series gas chromatograph that was
8482
Ind. Eng. Chem. Res., Vol. 46, No. 25, 2007
Figure 1. Effect of ligand concentration on rate of hydroformylation of MVN and n/iso ratio. Conditions: MVN, 10.7 × 10-2 kmol/m3; Rh(CO)2(acac), 7.7 × 10-4 kmol/m3; pressure (CO:H2 ) 1:1), 5.52 MPa; time, 120 min; solvent, NMP; and total volume, up to 2.5 × 10-5 m3.
Figure 3. Effect of catalyst concentration on rate of hydroformylation of MVN and n/iso ratio. Conditions: MVN, 0.1070 kmol/m3; (dppe) L:Rh ) 4:1; pressure (CO:H2 ) 1:1), 5.52 MPa; time, 120 min; solvent, NMP; and total volume, up to 2.5 × 10-5 m3.
of polymethylsiloxane). The oven temperature was programmed in the range of 373-573 K. The quantification of 2-MNP was done based on a calibration curve. The quantitative analysis of the products, D,L-naproxen and 2-acetyl-6-methoxynaphthalene, was performed using a Symmetry Shield RP-8 column (5 µm, 4.6 × 250 mm) on a series 1100 Agilent HPLC instrument that was equipped with a DAD, at 210 nm. A solution of 40% acetonitrile and 0.1% glacial acetic acid filtered through a 0.45 µm Vericel membrane filter was used as the mobile phase. The flow rate of the mobile phase was maintained at a value of 1 mL/min (isocratic). Results and Discussion Figure 2. Effect of temperature on hydroformylation of MVN. Conditions: Rh(CO)2(acac), 7.7 × 10-4 kmol/m3; (dppe) L:Rh ) 4:1; MVN, 0.108 kmol/m3; pressure (CO:H2 ) 1:1), 5.52 MPa; time, 180 min; solvent, NMP; and total volume, up to 2.5 × 10-5 m3.
controlled by the HP-Chemstation software and equipped with an autosampler unit, using an HP-1 capillary column (0.32 mm ID, 30 m length, 0.25 µm film thickness with a stationary phase
The proposed hydroformylation-oxidation route to naproxen involves the following reaction steps: (i) the hydroformylation of MVN to the regioisomers 2-MNP and 3-MNP, and (ii) the oxidation of 2-MNP to D,L-naproxen (2-MNPA), using H2O2 as the oxidant. The stoichiometric reactions are shown in Scheme 1. Because of the fact that only 2-MNP is useful for naproxen synthesis, the regioselectivity is the most important issue in the catalytic hydroformylation step. Therefore, the
Ind. Eng. Chem. Res., Vol. 46, No. 25, 2007 8483 Table 1. MVN Hydroformylation: Screening of Catalysts and Ligands Regioselectivity (%) run
1 2 3 4
5 6 7 8 9 10
catalyst
ligand
run time (min)
MVN conversion (%)
iso-
n-
turnover number, TON
RhCl3 HRhCO(PPh3)3 Rh(COD)Cl Rh(CO)2(acac)
Conditions: Catalyst, 8.7 × 10-4 kmol/m3; PPh3:Rh, 6:1; MVN, 0.108 kmol/m3; Pressure (CO:H2 ) 1:1), 8.2 MPa; Temperature, 373 K; Solvent, NMP (up to 2.5 × 10-5) PPh3 60 79.2 51.4 48.5 105.7 PPh3 60 80.5 64.3 64.0 107.5 PPh3 60 90.3 72.0 27.8 120.6 PPh3 60 95.7 76.3 23.0 127.8
Rh(CO)2(acac) Rh(CO)2(acac) Rh(CO)2(acac) Rh(CO)2(acac) Rh(CO)2(acac) Rh(CO)2(acac)
Conditions: Catalyst, Rh(CO)2(acac), 7.7 × 10-4 kmol/m3; L:Rh ) 4:1; MVN, 0.108 kmol/m3; Temperature, 373 K; Pressure (CO:H2 ) 1:1), 5.51 MPa PPh3 120 100.0 76.0 24.0 127.8 dpph 120 70.0 84.3 15.5 93.5 dppp 120 50.0 86.1 14.0 66.8 dppe 120 35.0 95.0 5.0 46.7 dppe 300 79.3 94.3 5.4 dppe 600 95.0 95.2 5.5
objective of this work was to investigate the role of different ligands, solvents, and reaction conditions in regioselectivity of the hydroformylation step. Because of the fact that there is no prior literature on the oxidation of 2-MNP, it was also the objective of this work to study the oxidation step using different catalysts and oxidants. The results are discussed below. 1. Hydroformylation of MVN. (A) Catalyst/Ligand Screening. Initial hydroformylation experiments were performed to study the effect of the catalyst precursors and ligands. The results for different catalyst precursors [RhCl3, Rh(COD)Cl2, HRhCO(PPh3)3, and Rh(CO)2(acac)] and ligands are presented in Table 1. It was observed that Rh(CO)2(acac), with PPh3 as a ligand, gave the highest activity (with a turnover frequency (TOF) of 127.8 h-1) with a regioselectivity of 76% to 2-MNP. The bidentate ligands improved the regioselectivity to 2-MNP significantly, being highest for dppe (95%) for a conversion of 35% and 100% for reaction durations of 2 and 8 h, respectively. It was observed that the high regioselectivity was independent of the conversion levels. The rate of MVN hydroformylation was higher with TPP as a ligand, but the n/iso ratio was also higher (i.e., 0.316), which is not desirable. It was observed that the regioselectivity to iso-aldehyde is enhanced (n/iso ) 0.053) with dppe as a ligand, even though the rate of MVN hydroformylation was lower, compared to that using TPP. The effect of ligand concentration on the rate of hydroformylation of MVN and the n/iso ratio were also studied at 373 K, using dppe as a ligand. It was observed that the rate of hydroformylation increases as the dppe concentration increases, passes through a maximum at an L:Rh ratio of 2:1, and then decreases, as shown in Figure 1. Beyond the typical L:Rh ratio, the rate starts to decrease, because of the formation of a catalytically inactive dimeric species.5 The n/iso ratio also increases as the ligand concentration increases until an L:Rh ratio of 4:1 is attained, after which it is independent of the concentration of the ligand. In the case of bidentate ligands, the increasing ligand concentration may cause a reversible formation of the catalytically inactive dinuclear rhodium complex species,5 thereby causing a decrease in the rate of reaction. This is indeed observed in the ligand concentration effect studied using dppe as the ligand with Rh(CO)2(acac) as the catalyst (see Figure 1). The olefin insertion step is reversible at higher temperatures and at low partial pressures of CO or H2, in the case of isoaldehyde formation, whereas these parameters do not affect the olefin insertion in the formation of the normal product.6 Therefore, it is expected that the regioselectivity to the isoaldehyde is adversely affected under these conditions, as observed in this study. A plausible mechanism for rhodium-
turnover frequency, TOF (h-1)
105.7 107.5 120.6 127.8
63.9 46.7 33.4 23.4
Table 2. MVN Hydroformylation: Solvent Screeninga Regioselectivity (%) solvents
MVN conversion (%)
iso-
n-
methyl ethyl ketone, MEK N-methyl pyrrolidone, NMP chlorobenzene ethyl acetate tetrahydrofuran, THF
42 38 29 22 18
94 94 91 91 92
6 6 9 9 8
a Conditions: catalyst, Rh(CO) (acac), 7.7 × 10-4 kmol/m3; (dppe):Rh 2 ) 4:1; MVN, 0.108 kmol/m3; temperature, 373 K; pressure (CO:H2 ) 1:1), 5.51 MPa; time, 180 min; solvent, up to 2.5 × 10-5 m3 total volume.
catalyzed MVN hydroformylation to 2-MNP and 3-MNP is given (see Scheme 2). (B) Effect of Solvent. The results of solvent screening, using solvents with varying polarities, are shown in Table 2. The low solubility of MVN in different solvents was a major constraint; therefore, polar solvents were not used for the screening study. It was observed that MVN conversions of 42% and 40% with >99% selectivity to the aldehyde products (2-MNP and 3-MNP) were obtained using MEK and N-methyl pyrrolidone (NMP) solvents, respectively. Moderate conversions (in the range of 18%-29% were obtained in hydroformylation experiments, using other solvents, such as tetrahydrofuran (THF), ethyl acetate, and chlorobenzene. It was observed that the regioselectivity varied only marginally (n/iso ) 0.0106-0.0109), although considerable variations in conversions were observed with the solvents screened. Although MVN has moderate solubility in ethanol, the latter as a solvent in the hydroformylation of styrenes is known to give acetals as undesirable byproducts in considerable quantities7 and, therefore, was not used in the present study. Generally, the polar solvents showed a remarkable effect on the activity and regioselectivity in the hydroformylation of MVN. (C) Effect of Temperature. The effect of temperature on the conversion of MVN and the n/iso ratio is shown in Figure 2 for 353-383 K. It was observed that the conversion increases from 13% to 48% as the temperature increases from 353 K to 383 K. The formation of n-aldehyde increases (1.5%-13%) as the temperature increases. (D) Effect of Catalyst Concentration. The effect of catalyst concentration on the rate of hydroformylation was studied at an MVN concentration of 0.107 kmol/m3 and CO and H2 partial pressures of 2.068 MPa each. The range of catalyst concentration for the study was 3.86 × 10-4-1.23 × 10-3 kmol/m3. The L:Rh ratio was maintained constant during this study. The rate varied with a linear dependence on catalyst concentration. There is a
8484
Ind. Eng. Chem. Res., Vol. 46, No. 25, 2007
Figure 5. Effect of CO concentration on rate of hydroformylation of MVN and n/iso ratio. Conditions: MVN, 0.1070 kmol/m3; Rh(CO)2acac, 7.7 × 10-4 kmol/m3; (dppe) L:Rh ) 4:1; PH2, 2.758 MPa; time, 120 min; solvent, NMP; and total volume, 25 mL.
Figure 4. Effect of MVN concentration on the rate of hydroformylation of MVN and n/iso ratio. Conditions: Rh(CO)2(acac), 7.7 × 10-4 kmol/m3; (dppe) L:Rh ) 4:1; pressure (CO:H2 ) 1:1), 5.52 MPa; time, 120 min; solvent, NMP; and total volume, up to 2.5 × 10-5 m3.
marginal decrease in the n/iso ratio with increasing catalyst concentration, as shown in Figure 3. (E) Effect of MVN Concentration. The effect of MVN concentration on the rate of hydroformylation and regioselectivity of the aldehyde products (2-MNP and 3-MNP) was studied over a range of MVN concentration of 0.053-0.0214 kmol/ m3, at CO and H2 partial pressures (PCO and PH2, respectively) of 2.76 MPa each, a catalyst loading of 7.72 × 10-4 kmol/m3 in NMP as a solvent, in a temperature range of 353-373 K. The results are presented in Figure 4. The rate increases linearly as the MVN concentrations increase in the range studied. The first-order kinetics with MVN is expected in hydroformylation reactions, as the enhanced olefin concentration will increase the formation of active alkyl rhodium species concentration. The n/iso ratio of the aldehyde products decreased marginally as the MVN concentration increased, as presented in Figure 4. (F) Effect of Partial Pressure of Carbon Monoxide (PCO). The effect of PCO on the rate of hydroformylation and the n/iso aldehyde ratio was studied at PH2 ) 2.76 MPa and a catalyst concentration of 7.72 × 10-4 kmol/m3 in NMP in a temperature range of 353-373 K and an MVN concentration of 0.107 kmol/
m3. The rate initially increased as PCO increased, and then it decreased with further increases in PCO. The inhibition in rate with enhanced CO pressure, because of the formation of the inactive dicarbonyl and tricarbonyl species of rhodium, is a wellknown phenomenon in hydroformylation chemistry.8 The n/iso ratio was also observed to be strongly dependent on PCO. The n/iso ratio decreased as PCO increased. The PCO and PH2 values are known to affect the regioselectivity at higher temperatures, because of the different behavior of the alkyl rhodium intermediate toward the β-hydride elimination. Lazzaroni et al. have shown, through deuteroformylation experiments at higher temperatures, that the β-hydride elimination from the branched alkylmetal occurs to a larger extent at lower CO and H2 partial pressures, unlike the β-hydride elimination from the linear one, which occurs only to a small extent even at low pressure.6 The results obtained in the present study are given in Figure 5. (G) Effect of the Partial Pressure of Hydrogen (PH2). The effect of PH2 on the rate of hydroformylation and the n/iso aldehyde ratio was studied at a PCO value of 2.76 MPa, a catalyst loading of 7.72 × 10-4 kmol/m3 in NMP solvent in a temperature range of 353-373 K, and at an MVN concentration of 0.107 kmol/m3. The results are presented in Figure 6. The rate was determined to have a first-order dependence on PH2. Because the oxidative addition of hydrogen to the acyl carbonyl rhodium phosphine species is the rate-determining step, an increase in PH2 would lead to an increased rate of hydroformylation. The n/iso ratio was observed to decrease with increasing hydrogen pressure at all the temperatures studied, as observed
Ind. Eng. Chem. Res., Vol. 46, No. 25, 2007 8485
Figure 7. Effect of total pressure on the hydroformylation of MVN. Conditions: Rh(CO)2(acac), 8.1 × 10-4 kmol/m3; (dppe) L:Rh ) 4:1; MVN, 10.8 × 10-2 kmol/m3; temperature, 373 K; time, 240 min; solvent, NMP; and total volume, up to 2.5 × 10-2 m3.
Figure 8. MVN hydroformylation: experimental versus predicted rates.
Figure 6. Effect of H2 concentration on the rate of hydroformylation of MVN and n/iso ratio. Conditions: MVN, 10.7 × 10-2 kmol/m3; Rh(CO)2(acac), 7.7 × 10-4 kmol/m3; (dppe) L:Rh ) 4:1; PCO, 2.758 MPa; time, 120 min; solvent, NMP; and total volume, up to 2.5 × 10-5 m3.
in Figure 6. This result is concurrent with the explanation given previously. The effect of the total pressure on the conversion of MVN and the n/iso ratio is also shown in Figure 7. It was observed that the conversion increases from 63% to 85% as the pressure of the synthesis gas increased from 2 MPa to 8 MPa without affecting the n/iso ratio. (H) Kinetic Study. An important objective of this work was to investigate the kinetics of hydroformylation of MVN to 2-MNP and 3-MNP, using Rh(CO)2(acac) as a catalyst and dppe as a ligand. The kinetics of hydroformylation of MVN was studied under the following range of conditions (i) PCO, 0.345.51 MPa, (ii) PH2, 0.69-3.45 MPa, and (iii) MVN, 0.25 to 1.0 kmol/m3 in a temperature range of 353-373 K. For the kinetic studies, only the initial rate data, wherein the concentration of the olefin changed marginally (∼10%-15%), was considered. Preliminary experiments were conducted to ensure the material balance and reproducibility of the experiments. In these experiments, the amounts of MVN consumed, products (2-MNP and 3-MNP) formed, and syngas consumed were recorded. A typical
Figure 9. MVN hydroformylation: Arrhenius plot.
concentration-time profile is shown later in this paper as Figure 10. Generally, it was observed that the material balance of the syngas (CO + H2), and the MVN consumed, were consistent with the amount of total aldehyde products (2-MNP and 3-MNP) formed. The formed products were confirmed by GC-MS. The observation that agitation speed has no effect on the rate indicated the absence of mass-transfer resistance. The observed dependency of the rate on different parameters indicates that the hydroformylation of MVN is first order, with respect to the catalyst and dissolved hydrogen concentrations, and a negative order, with respect to PCO. The effect of MVN concentration on the initial rate showed first-order kinetics. To fit the rate data, several rate equations were examined using a nonlinear regression analysis. The results on parameters estimated for different models are presented in Table
8486
Ind. Eng. Chem. Res., Vol. 46, No. 25, 2007
Table 3. Rate Models Examine to Fit the Data on MVN Hydroformylation
riso ) rn )
φmin
k1A*B*CD
353
Rate Model 1 2.53 × 102
7.64 × 100
4.99 × 100
1.31 × 10-14
(1 + K2B*)2
363
3.45 × 102
1.45 × 101
1.80 × 101
8.33 × 10-17
373
6.96 × 102
2.47 × 101
5.26 × 101
4.52 × 10-14
k1A*B*CD
353
Rate Model 2 4.02 × 101
5.32 × 103
1.42 × 10-5
2.83 × 10-11
(1 + K2B*)
363
4.55 × 101
7.25 × 104
1.05 × 10-4
1.30 × 10-10
373
6.12 × 102
1.15 × 102
2.07 × 10-2
3.99 × 10-11
k1(A*)1.5B*(C)1.5D
353
Rate Model 3 3.29 × 101
1.96 × 10-4
3.01 × 10-3
2.75 × 10-11
(1 + K2B*)2
363
4.55 × 101
1.21 × 10-4
1.50 × 10-3
1.30 × 10-10
373
7.43 × 102
1.19 × 10-4
3.05 × 10-3
6.61 × 10-10
k1A*B*CD
353
Rate Model 4 3.29 × 101
1.96 × 10-4
3.01 × 10-3
2.75 × 10-11
(1 + K2B*)3
363
4.55 × 101
1.21 × 10-4
1.50 × 10-3
1.30 × 10-10
373
7.43 × 100
1.19 × 10-4
3.05 × 10-3
6.61 × 10-10
353
Rate Model 5 3.29 × 101
2.50 × 10-6
5.13 × 10-4
2.75 × 10-11
363
4.55 ×
10-5
10-3
1.30 × 10-10
373
6.40 × 101
5.20 × 10-3
4.19 × 10-10
k3A*B*CD
k3(A*)1.5B*(C)1.5D (1 + K2B*)2
k3A*B*CD (1 + K2B*)2
riso ) rn )
k3
(1 + K2B*)
riso ) rn )
k2
(1 + K2B*)2
riso ) rn )
temperature, T (K)
k3A*B*CD
riso ) rn )
rate model
k1A*B*CD (1 + K2(B*)2) k3A*B*CD
(1 + K2(B*)2)
k1
3; for this purpose, an optimization program based on Marquardt’s method was used. The objective function was chosen as follows: n
F)
[RAi - Rai]2 ∑ i)1
(1)
where F is the objective function to be minimized (φmin), representing the sums of squares of the difference between the observed and predicted rates, n is the number of experimental data; RAi and Rai represent predicted and experimental rates, respectively. The minimum absolute squared-error objective function (Φmin) was selected as the basis for the discrimination of the kinetic models. The values of rate parameters and Φmin are presented in Table 3. Models 2, 3, and 4 are not consistent with the observed rate dependence. Also, the values of Φmin, for these models, were greater than that for model 1. Φmin for model 5 is higher than that of model 1. Therefore, model 1 was considered the best model for representing the kinetics of hydroformylation of MVN using Rh(CO)2(acac)as the catalyst and with dppe as a ligand. The corresponding rate equations are
riso ) rn )
k1A*B*CD (1 + k2B*)
2
k3A*B*CD (1 + k2B*)2
(2)
101
5.06 ×
2.92 × 10-4
1.63 ×
where k is the intrinsic reaction rate constant (given in units of m3/kmo13); A* and B* are the concentrations of H2, and CO in NMP at the gas/liquid interface (given in units of kmol/m3), respectively; and C and D represent the concentrations of the catalyst and MVN (kmol/m3), respectively. The rate parameters in eqs 2 and 3 are presented in Table 3. The experimental rates were compared with those predicted by the rate model (eqs 2 and 3) in Figure 8, which shows a reasonably good fit of the data. The average deviation in the predicted and observed rates was determined to be in the range of 5%. The Arrhenius plots showing the effect of temperature on the rate parameters are shown in Figure 9, from which the activation energy was evaluated as 13.24 and 31.02 kcal/mol. To verify the applicability of the kinetic model under integral conditions, experimental data on the liquid-phase concentrations of MVN, 2-MNP, and 3-MNP as a function of time were obtained. The variation of the concentration of MVN, 2-MNP, and 3-MNP can be represented by the following mass-balance equations, for conditions of constant synthesis gas pressure in the reactor. For 2-methoxy-6-vinylnaphthalene:
k3A*B*CD k1A*B*CD dD + ) 2 dt (1 + K2B*) (1 + K2B*)2
-
(3) For 2-methoxynaphthylpropanal:
(4)
Ind. Eng. Chem. Res., Vol. 46, No. 25, 2007 8487
k1A*B*CD dE ) dt (1 + K B*)2 2
(5)
For 3-methoxynaphthylpropanal:
k3A*B*CD dG ) dt (1 + K2B*)2
(6)
with initial conditions of
t)0 D ) D0 E)0 and
G)0
(7)
For the data at a constant synthesis gas pressure, the equations were solved numerically, using the Runge-Kutta method to obtain the concentration of MVN, 2-MNP, and 3-MNP, as a function of time. For this purpose, intrinsic rates were used. The comparison of the experimental and the predicted results for 373 K are presented in Figure 10, which show excellent agreement. These results indicate that the rate model proposed on the basis of the initial rate data is also applicable over a wide range of conditions and can be used for design and scaleup purposes. To show the difference in fitting between model 1 and the other models, a typical substrate consumption-time profile is shown in Figure 11, which clearly indicates the suitability of model 1. (I) Oxidation of 2-MNP. It is well-known from previous work that the oxidation of organic substrates by tungstate catalysts proceeds via a peroxo-complex intermediate.9 Such types of complexes of transition metals such as molybdenum, vanadium, and especially tungsten, etc., particularly heterogenized on supports such as hydrotalcites or resins, in their highest oxidation states are known to be excellent oxidation catalysts for the oxidation of benzylic and secondary alcohols,10 epoxidation of olefins,11 oxidative bromination of phenols,12 etc. Therefore, experiments on the oxidation of 2-MNP were performed using molybdenum, vanadium, and tungsten compounds, with H2O2 as the oxidant and TBAHS as a phasetransfer catalyst. The results are presented in Table 4, which show that the Na2WO4·2H2O catalyst with TBAHS gives the best results (19.3% conversion of 2-MNP and 82.3% selectivity for naproxen), compared to other catalysts. Hence, Na2WO4· 2H2O was used as the catalyst for further experiments. The effect of Na2WO4·2H2O concentration on activity and selectivity for the oxidation of 2-MNP was studied, and the results are presented in Table 5, which showed that conversion of 2-MNP increased as the catalyst concentration increased (from 5.5% to 23.1%), whereas the selectivity to naproxen decreased marginally (from 81.2% to 74.7%). The solubility of the aldehyde, 2-MNP, in the organic solvents is generally very low, and this was the constraint involved in studying the solvent effect. The 2-MNP was soluble in the desired concentrations only in ethyl acetate and MEK. We found that the reaction was very slow in ethyl acetate, because hydrogen peroxide forms a separate phase in the reaction mixture. Therefore, MEK was used as the solvent. The results on the effect of other reaction parameters, such as TBAHS, H2O2, and 2-MNP concentrations, as well as the temperature, on the conversion of 2-MNP and the selectivity to naproxen are presented in Tables 6-9. With
Figure 10. MVN hydroformylation: experimental versus predicted concentration-time data. Conditions: Rh(CO)2(acac), 7.7 × 10-4 kmol/ m3; (dppe) L:Rh ) 4:1; MVN, 12.00 × 10-2 kmol/m3; pressure (CO:H2 ) 1:1), 5.52 MPa; time, 600 min; solvent, NMP; and total volume, up to 2.5 × 10-5 m3.
Figure 11. MVN hydroformylation: experimental versus predicted concentration-time data for MVN conversion with different models. Conditions: Rh(CO)2(acac), 7.7 × 10-4 kmol/m3; (dppe) L:Rh ) 4:1; MVN, 12.00 × 10-2 kmol/m3; pressure (CO:H2 ) 1:1), 5.52 MPa; time, 600 min; solvent, NMP; and total volume, up to 2.5 × 10-5 m3. Table 4. Oxidation of 2-MNP: Effect of Screening Catalystsa Selectivity (%) catalyst
2-MNP conversion (%)
naproxen
ketone
Na2WO4‚2H2O Na2MoO4‚2H2O NaVO3 H3PO4‚12WO3‚xH2O H3PO4‚12MoO3‚xH2O
19.3 6.7 9.3 4.2 2.6
82.3 52.3 56.5 71.2 66.7
12.2 45.2 40.2 24.7 28.2
a Reaction conditions: MNP, 10.7 × 10-2 kmol/m3; catalyst, 0.010 kmol/ m3; TBAHS, 0.010 kmol/m3; H2O2, 0.250 kmol/m3; MEK, up to 5 × 10-6 m3 total volume; temperature, 298 K; agitation, 16.6 Hz; and time, 6 h.
increases in the TBAHS concentration, the conversion of 2-MNP, as well as the selectivity to naproxen, was observed to increase (see Table 6). The effect of H2O2 concentration (Table 7), showed a gradual increase in conversion of 2-MNP with a loss in selectivity to naproxen. The parameter variation studies showed that the optimum results were obtained at 298 K, where the conversion of 2-MNP was moderate, but excellent selectivity
8488
Ind. Eng. Chem. Res., Vol. 46, No. 25, 2007
Table 5. Oxidation of 2-MNP: Effect of Catalyst Concentrationa Selectivity (%) Na2WO4‚2H2O (kmol/m3)
2-MNP conversion (%)
naproxen
ketone
0.005 0.01 0.02
5.5 19.3 23.1
81.2 82.3 74.7
18.6 12.2 23.1
a Reaction conditions: MNP, 10.7 × 10-2 kmol/m3; TBAHS, 0.010 kmol/ m3; H2O2, 0.250 kmol/m3; MEK, up to 5 × 10-6 m3 total volume; temperature, 298 K; agitation, 16.6 Hz; and time, 6 h.
Table 6. Oxidation of 2-MNP: Effect of TBAHS Concentrationa Selectivity (%) TBAHS (kmol/m3)
2-MNP conversion (%)
naproxen
ketone
0.005 0.010 0.021
15.2 19.3 21.6
60.1 82.3 85.3
36.2 12.2 14.0
Reaction conditions: MNP, 10.7 × 10-2 kmol/m3; Na2WO4·2H2O, 0.010 kmol/m3; H2O2, 0.250 kmol/m3; MEK, up to 5 × 10-6 m3 total volume; temperature, 298 K; agitation, 16.6 Hz; and time, 6 h. a
Table 7. Effect of H2O2 Concentrationa Selectivity (%) H2O2 (kmol/m3)
2-MNP conversion (%)
naproxen
ketone
0.12 0.25 0.50
10.8 19.3 24.1
78.3 82.3 65.5
14.2 12.2 33.9
a Reaction conditions: MNP, 10.7 × 10-2 kmol/m3; Na WO ·2H O, 2 4 2 0.010 kmol/m3; TBAHS, 0.010 kmol/m3; MEK, up to 5 × 10-6 m3 total volume; temperature, 298 K; agitation, 16.6 Hz; and time, 6 h.
Table 8. Oxidation of 2-MNP: Effect of Substrate Concentrationa Selectivity (%) MNP (kmol/m3)
2-MNP conversion (%)
naproxen
ketone
0.05 0.107 0.22
28.3 19.3 15.8
70.6 82.3 84.3
26.5 13.6 15.1
a Reaction conditions: Na WO ·2H O, 0.010 kmol/m3; TBAHS, 0.010 2 4 2 kmol/m3; H2O2, 0.250 kmol/m3; MEK, up to 5 × 10-6 m3 total volume; temperature, 298 K; agitation, 16.6 Hz; and time, 6 h.
Table 9. Oxidation of 2-MNP: Effect of Temperaturea Selectivity (%) temperature (K)
2-MNP conversion (%)
naproxen
ketone
273 298 308
6.0 19.3 15.8
85.6 82.3 60.8
14.5 13.6 36.7
a Reaction conditions: MNP, 10.7 × 10-2 kmol/m3; Na WO ·2H O, 2 4 2 0.010 kmol/m3; TBAHS, 0.010 kmol/m3; H2O2, 0.250 kmol/m3; agitation, 16.6 Hz; and time, 6 h.
(up to 82.3%) to the carboxylic acid (D,L-naproxen) could be achieved. It was observed that the major byproducts of the reaction were ketone and 2-acetyl-6-methoxynaphthalene, which may be due to oxidation at the tertiary carbon. In the parametric effect studies, the best results were obtained at a substrate concentration of 0.107 kmol/m3, a catalyst concentration of 0.01 kmol/m3, and the H2O2-to-substrate ratio of 1.5.
Conclusion The hydroformylation of 6-methoxy-2-vinylnaphthalene (MVN), using homogeneous Rh complex catalysts with dppe as the ligand, has been studied, with the objective of exploring it as the key step in the synthesis of naproxen. It was observed that a Rh catalyst, in the presence of an optimized concentration of the bidentate phosphine ligand-dppe under the reaction conditions in the present study, gave a high regioselectivity to the branched isomer 2-MNP. The 2-MNP thus obtained could be oxidized under mild conditions in a facile manner using the sodium tungstate-tetrabutyl ammonium hydrogen sulfate (TBAHS) catalyst system, and hydrogen peroxide, to yield the racemic naproxen with high selectivity. The present study elucidates the potential application of the hydroformylation-oxidation route for the production of naproxen. Nomenclature A* ) concentration of H2 in NMP in equilibrium with the gas phase (kmol/m3) B* ) concentration of CO in NMP in equilibrium with the gas phase (kmol/m3) C ) concentration of the catalyst (kmol/m3) D ) concentration of 6-MVN (kmol/m3) E ) concentration of 2-MNP (kmol/m3) G ) concentration of 3-MNP (kmol/m3) k1, k3 ) reaction rate constants (m9/kmol3/s) K2 ) constant in eq 2 (m3/kmol) N ) agitation speed (rpm) PH2 ) partial pressure of hydrogen (MPa) PCO ) partial pressure of carbon monoxide (MPa) R ) rate of hydroformylation (kmol/m3/s) t ) reaction time (min) Acknowledgment Authors K.B.R., S.S.T., and M.R.D. wish to thank the Council of Scientific and Industrial Research (CSIR), Government of India, and Task Force (under CMM0005) for financial support as Senior Research Fellows. Literature Cited (1) Harrington, P. J.; Lodewijk E. Twenty years of naproxen technology. Org. Process Res. DeV. 1997, 1, 72. (2) Akbarali, P. M.; Vijaya, R. K. K.; Gani, R. S.; Krishna, S.; Venkatrraman, S.; Mahalinga, M. Process for producing Ibuprofen sodium dihydrate, U.S. Patent No. 7,084,299, August 1, 2006. (3) Elango, V.; Murphy, M. A.; Smith, B. L.; Davenport, K. G.; Mott, G. N.; Moss, G. L., Method for producing Ibuprofen, U.S. Patent No. 4,981,995, January 1, 1991. (4) Varshavskii, Y. S.; Cherkasova, T. G. A simple method for preparing acetylacetonate dicarbonyl rhodium(I). Russ. J. Inorg. Chem. 1967, 12, 899. (5) Castellanos-Pa´ez, A.; Castillo´n, S.; Claver, C.; van Leeuwen, P. W. N. M.; de Lange W. G. J. Diphosphine and dithiolate rhodium complexes: characterization of the species under hydroformylation conditions. Organometallics 1998, 17, 2543. (6) Lazzaroni, R.; Raffaelli, A.; Settambolo, R.; Bertozzi, S.; Vitulli, G. Regioselectivity in the Rhodium-catalyzed hydroformylation of styrene as a function of reaction, temperature and gas pressure. J. Mol. Catal. 1989, 60, 1. (7) Nair, V. S. Hydroformylation of Olefins Using Homogeneous and Biphasic Catalysis, Ph.D. Thesis, University of Pune, India, Chapter II, 1999, p 136. (8) Deshpande, R. M.; Chaudhari, R. V. Kinetics of hydroformylation of 1-hexene using homogeneous HRh(CO)(PPh3)3 complex catalyst. Ind. Eng. Chem. Res. 1988, 27, 1996.
Ind. Eng. Chem. Res., Vol. 46, No. 25, 2007 8489 (9) Sheldon, R. A.; Arends, I. W. C. E.; Dijksman, A. New developments in catalytic alcohol oxidations for fine chemicals synthesis Catal. Today 2000, 57, 157. (10) Sato, K.; Aoki, M.; Takagi, J.; Noyori, R. Organic solvent- and halide-free oxidation of alcohols with aqueous hydrogen peroxide. J. Am. Chem. Soc. 1997, 119, 12386. (11) Hoegaerts, D.; Sels, B. F.; de Vos, D. E.; Verpoort, F.; Jacobs, P. A. Heterogeneous tungsten-based catalysts for the epoxidation of bulky olefins. Catal. Today 2000, 60, 209.
(12) Sels, B.; De Vos, D.; Buntinx, M.; Pierard, F.; Kirsch-De Mesmaeker, A.; Jacobs, P. Layered double hydroxides exchanged with tungstate as biomimetic catalysts for mild oxidative bromination. Nature 1999, 400, 855.
ReceiVed for reView January 14, 2007 ReVised manuscript receiVed April 3, 2007 Accepted April 5, 2007 IE0700866
