Ind. Eng. Chem. Res. 2011, 50, 23–27
23
Sustainable and Scalable Synthesis of Piperylene Sulfone: A “Volatile” and Recyclable DMSO Substitute Gregory A. Marus, Eduardo Vyhmeister, Pamela Pollet, Megan E. Donaldson, Veronica Llopis-Mestre, Sean Faltermeier, Renee Roesel, Michael Tribo, Leslie Gelbaum, Charles L. Liotta, and Charles A. Eckert* School of Chemical & Biomolecular Engineering, School of Chemistry & Biochemistry, Georgia Institute of Technology, and Specialty Separations Center, Atlanta, Georgia 30332-0100
An essential feature of chemical research is advancing new laboratory findings to a form useful to industry. Dimethyl sulfoxide (DMSO) is an important dipolar, aprotic solvent for conducting chemical reactions. Unfortunately, separation of the products of reaction from the solvent is difficult and expensive. We have proposed the use of piperylene sulfone (PS) as a substitute for DMSO, and herein, we establish a roadmap to its sustainable and scalable synthesis. PS is a potentially important new dipolar aprotic solvent that has solvent properties similar to those of DMSO; in contrast to DMSO, PS is fully recyclable and undergoes a reversible retrocheletropic reaction at 110 °C, permitting facile solvent removal and recycle. PS is synthesized by the reaction of trans-piperylene with sulfur dioxide. Because PS is not commercially available, we synthesized laboratory quantities using a method not sustainable on a large scale because of expensive chemicals and considerable waste generation. To develop and optimize a scalable process, we (1) determined the kinetic parameters associated with the reaction by employing in situ proton NMR measurements and (2) studied the effects of radical inhibitors in reducing unwanted side reactions. In addition, we recovered PS from the reaction mixture through a sustainable CO2 separation method, which resulted in a substantial waste reduction. Our development of a more efficient, safe, and sustainable scaleup method for PS thus illustrates an important aspect of chemical research: the need to render the results usable and useful to industry. Introduction We present a paradigm for process scaleup using the example of the synthesis of piperylene sulfone (PS), a “volatile” and recyclable dimethyl sulfoxide (DMSO) substitute, providing opportunities for implementation in industry. Factors that were considered include reactor design and engineering, chemistry modification, and increased attention to safety. Polar aprotic solvents such as DMSO offer important advantages such as the single-phase dissolution of inorganic salts and organic substrates. However, solvent removal and product purification present a major challenge because of DMSO’s high boiling point (189 °C), often resulting in waste-intensive product isolation strategies. Further, the solvent removal step is often product-specific, and recycling of the polar, aprotic solvent is generally not economical. We have demonstrated the properties and uses of PS as the first recyclable dipolar, aprotic solvent.1 PS has solvatochromic properties that are very similar to those of DMSO, indicating similar solvent strength.1 However, unlike DMSO, PS has a thermal “switch,” activated at temperatures over 100 °C, where PS breaks into two gases, trans-piperylene and sulfur dioxide, for easy separation. Subsequently, trans-piperylene and sulfur dioxide can be recombined to reform PS, for recyclability. PS has now been used at the laboratory scale as a recyclable and volatile dipolar aprotic solvent for nucleophilic substitution, aerobic (Tempo) alcohol oxidations, and the in situ catalyzed hydrolysis of β-pinene.1-3 PS is volatile in the sense that the solvent is removed in the form of two gaseous species upon heating, which permits separation of products and solvent recycle. * To whom correspondence should be addressed. E-mail: cae@ gatech.edu.
PS has been synthesized by a method known in the literature.4 The procedure involves addition of piperylene, a radical inhibitor (typically either hydroquinone, p-t-butyl-catechol, or phenylβ-naphthylamine; the structure of the latter is shown in Figure 3A below),5-8 and an excess of liquid sulfur dioxide to a stainless steel bomb. Only the trans isomer of piperylene reacts to form PS; thus, yield is calculated based on the trans isomer content. Furthermore, decomposition of the PS yields solely trans-piperylene. This method can be used to purify the raw piperylene to the trans isomer.9 An excess of sulfur dioxide is commonly used in the literature4,6,9 to maximize the PS yield and minimize undesired polymerization of trans-piperylene, a reaction that is often observed with conjugated dienes.10 After addition of all reactants, the reactor is sealed and shaken for 2 days at room temperature (20 °C). The synthesis reaction is shown in Figure 1. After the inhibitor is removed, the yield is observed to range from 45% to 55%,3,6,11 and PS is then purified by distillation at reduced pressure. More recently, other researchers have adopted this procedure.3 Numerous physical properties of PS have been determined (including the density and refractive index,12 equilibrium constants and decomposition,5 rates of thermal dissociation,11 and heat of formation5), as well as a method of enriching the trans isomer9 and characterization studies including optical activity and purification.6 To make the separation of the inhibitor less waste-intensive for large-scale production, a more efficient separation strategy than used in the literature6 was devised. This new separation
Figure 1. Synthesis of PS from trans-1,3-pentadiene and SO2.
10.1021/ie100920y 2011 American Chemical Society Published on Web 07/14/2010
24
Ind. Eng. Chem. Res., Vol. 50, No. 1, 2011
strategy results in a significant reduction of waste generation. The literature method previously developed to separate the inhibitor from the PS involved an extraction into water saturated with sodium chloride, followed by a back-extraction into dichloromethane, with subsequent removal of the dichloromethane and filtration.6 The major disadvantage of this separation method is waste generation (over 100 mol of waste per mole of PS generated). In this work, we achieved a more fundamental understanding of the synthesis of PS to improve reaction performance and facilitate the implementation of largescale production. Results and Discussion We developed a sustainable protocol by improving reaction performances and developing a more efficient separation strategy. Synthesis of PS. Steps we have taken to make the synthesis of PS more feasible on a large scale include the following: (1) determination of the kinetics of PS formation from piperylene and sulfur dioxide, (2) minimization of undesirable side reactions by adjusting dilution and the structure of the inhibitors, (3) use of the kinetic information to develop a scaleup design, and (4) validation of the scaleup design on a laboratory scale using raw industrial material (mixtures of cis- and trans-piperylene). PS is synthesized by the cheletropic reaction of trans-1,3pentadiene (trans-piperylene) and sulfur dioxide, as illustrated in Figure 1. Although the reaction of conjugated dienes is known in the literature,5 critical scaleup data such as kinetics of the synthesis reaction, side product formation, and so on are scarce. We used in situ proton NMR measurements in high-pressure NMR tubes to monitor the real-time conversion to PS from 94% trans-piperylene and SO2 with respect to time (details provided in the Supporting Information). By monitoring in situ reaction progress, we determined that the time to reach completion occurred within 12 h at 295 K, as can be seen from the constant value of conversion in Figure 2 (below). Kinetics. We used in situ proton NMR measurements to obtain the kinetic parameters for PS formation (see the Supporting Information). In brief, specific NMR absorptions from piperylene and PS were used to calculate the conversion of the reaction, and gravimetric measurements were used to calculate the exact concentration of each component in the NMR tube and used for conversion calculations. The combined mole balance and rate law yields the kinetic expression for the formation of piperylene sulfone (PS) (eq 1). dCPS ) kCPiperylenenCSO2m dt
(1)
The reaction is first-order with respect to each reactant (n ) 1 and m ) 1), as determined by the integral method (R2 is greater than 98%). A sensitivity analysis confirms first-order behavior by comparison of the regression coefficients for other reaction orders. From the average of our experiments using the integral and differential methods, the kinetic rate constant is k ) 0.022 ( 0.003 L/(mol h) at 295 K. We have determined the rate constants at two temperatures (283 and 295 K) and estimated an approximate activation energy of 40 kJ/mol. Dilution. Only two significant products can be formed: (1) PS from the cheletropic reaction of trans-piperylene with SO2 and (2) undesired polymeric products from piperylene polymerization. Complete conversion of trans-piperylene was observed in all cases, as confirmed by NMR spectroscopy.
Figure 2. Effect of diluting piperylene with excess sulfur dioxide on the PS yield at four experimental concentrations, analyzed by in situ 1H NMR spectroscopy at 295 K. No inhibitor was present.
Figure 3. Structures of inhibitors: (A) neutral inhibitor and (B) ionic inhibitor.
Figure 4. Effect of neutral inhibitor (N-phenyl-2-naphthylamine) on PS yield at 8.8 mol % piperylene, as analyzed by in situ 1H NMR spectroscopy at 295 K.
Although dilution of the trans-piperylene with excess sulfur dioxide is known to increase the PS yield, we sought to achieve a deeper understanding. By in situ proton NMR measurements, we measured the formation of PS at a range of dilute piperylene concentrations, as shown in Figure 2. The relative concentration of trans-piperylene is varied from 26 to 2.73 mol %. At the highest concentration of piperylene tested (26 mol % piperylene), the PS yield was 48% (the remaining piperylene polymerized). At the lowest concentration (2.73 mol % piperylene), the PS yield increased to 76%. Figure 2 shows that, as the concentration of piperylene decreases (relative sulfur dioxide concentration increases), PS yield increases, but polymerization still occurs at all conditions tested. These results indicate that dilution alone is insufficient to achieve complete conversion to PS. Although dilution does have a substantial effect on reduction of polymer formation, the use of an inhibitor can further increase the yield and make the synthesis more practical for industrial use. Inhibitor Effect. Two radical inhibitors, N-phenyl-2-naphthylamine (neutral inhibitor) and 8-anilino-1-naphtalenesulfonic acid magnesium salt (ionic inhibitor), shown in Figure 3, were investigated over a range of piperylene concentrations, in all
Ind. Eng. Chem. Res., Vol. 50, No. 1, 2011
Figure 5. Effect of the ionic inhibitor (8-anilino-1-naphtalenesulfonic acid magnesium salt) on PS yield at 7.4 mol % piperylene, as analyzed by in situ 1H NMR spectroscopy at 295 K.
Figure 6. Cycle for the sustainable CO2 separation of PS and ionic inhibitor.
cases at an inhibitor mole fraction of 0.05. When using the neutral inhibitor, 100% PS yield was not obtained, yet an increase in dilution increased the PS yield and decreased polymerization (Figure 4).
Figure 7. 1H NMR spectrum of PS after sustainable CO2 separation.
25
The neutral inhibitor is more effective at preventing polymerization than dilution alone. A comparison between 4.5 mol % piperylene in Figure 2 and 8.8 mol % piperylene with neutral inhibitor in Figure 4, shows a difference from 68% PS yield to 83% PS yield, respectively. However, even at dilute piperylene concentrations (8.8 mol %) with the N-phenyl-2-naphthylamine inhibitor, complete conversion to PS was not achieved (only 83% conversion to PS, as illustrated in Figure 4). The ionic inhibitor gave complete yield of PS, and no polymer formation was detected by 1H NMR analysis. Thus, the ionic inhibitor is superior to the neutral inhibitor (constant mole fractions of inhibitor), as demonstrated by Figures 4 and 5. Several experiments have been run under similar conditions with the use of inhibitor and demonstrate the reproducibility of these results. Additionally, the ionic inhibitor allows enhanced separation from the product, PS, as discussed further in the Separation section. As the next step, we proceeded from PS synthesis on the milligram scale in NMR tubes to synthesis on the multigram scale (25 mL of piperylene added to at least 200 mL of liquid SO2 and 2.75 g of neutral inhibitor), a scale still suitable for a university laboratory. The use of expensive and highly concentrated 94% pure trans-piperylene is not feasible for large-scale industrial production; thus, we used the kinetic information found to synthesize larger amounts of PS from an industrial source of piperylene that is available in bulk quantities. The raw industrial piperylene that we used was composed of a mixture of cis and trans isomers (approximately 50% trans and
26
Ind. Eng. Chem. Res., Vol. 50, No. 1, 2011
Figure 8. 13C NMR spectrum of PS after sustainable CO2 separation.
40% cis), as well as other impurities. After 15 h, we attained a 75% yield of PS (based on the trans isomer content) and then separated the PS from the inhibitor using a literature method.6 Under similar conditions of temperature, dilution with excess SO2, and constant mole fraction of inhibitor, the yields between NMR reactions and scaleup were similar. The list of chemical shifts from the 1H NMR spectrum verifies nearly pure production (and is given in the Supporting Information). Separation. The literature method6 involves an extraction into water saturated with sodium chloride, followed by a backextraction into dichloromethane, subsequent removal of the dichloromethane, and filtration as the final step. This separation strategy is undesirable for three reasons: (1) It is waste-intensive because over 100 mols of waste are generated for each mole of PS produced (over 55 kg of waste per kilogram of PS, assuming equal volumes of aqueous and organic components and regarding the organic material as waste after the separation); (2) it is expensive because of the large amount of solvents involved in extractions and back-extractions; and (3) it uses a halogenated solvent, dichloromethane. For a sustainable industrial separation, we developed a liquid CO2-induced separation method that takes advantage of the low solubility of the ionic inhibitor in nonpolar liquid CO2 (Figure 6). Note that carbon dioxide is only being usedsnot generatedsin the separation process and will be recycled. Assuming recovery and recycle of CO2, minimal waste is generated using this separation method. The CO2 experiment was performed at a temperature of 294 K and pressure of about 59 bar. Because of the commercial unavailability of PS, we first demonstrated a proof-of-principal separation of DMSO from a soluble mixture of DMSO and ionic inhibitor. The sustainable separation results in a reduction from 3.70 ( 0.17 to 0.07 ( 0.04 wt %. Additionally, the ionic inhibitor has limited solubility in sulfur dioxide; it is over an order of magnitude less soluble in sulfur dioxide than the mole fraction (0.05) initially used with the neutral analogue. Therefore, the inhibitor mole fraction was reduced from 0.05 to 0.0007, and PS was synthesized using the same method on a 48 mL scale. Subsequently, we carried out the sustainable CO2 separation of PS from a mixture of PS and ionic inhibitor. The exact initial concentration of inhibitor in PS was found from ultraviolet-visible (UV-vis) spectroscopy. Then, liquid CO2 was added (PS is soluble in the nonpolar CO2 phase). As the amount of CO2 increases, it acts as an antisolvent, causing the ionic inhibitor to precipitate out of the CO2-expanded PS. The CO2 + PS phase was sampled and analyzed by UV-vis spectroscopy and indicated a reduction in the inhibitor concentration from 3.56 ( 0.05 to 0.07 ( 0.02 wt %. This sample of PS was also analyzed after the separation,
using 1H and 13C NMR spectroscopies, as shown in Figures 7 and 8, and was verified according to literature sources.13,14 The presence of ionic inhibitor was not detected in either NMR spectrum. Conclusions We have demonstrated a method for transitioning from a laboratory synthesis to a less waste-intensive, safer, and more sustainable method for the industrial production of a specialty chemical. The example shown is the scalable synthesis of the fully recyclable dipolar aprotic solvent, piperylene sulfone. For this reaction, we attained the kinetic parameters of PS formation, demonstrated suppression of undesirable polymerization, and developed a novel separation method to minimize waste. Thus, the PS production demonstrates the progression of innovative techniques and laboratory discoveries to a more useful form. This research is a critical first step toward the implementation of laboratory research into industry. Acknowledgment We express gratitude to Eli Lilly and AMPAC Fine Chemicals for their support of this work. Supporting Information Available: PDF describing the materials, NMR procedure, reactors, and scale-up procedure used in this work for the preparation of piperylene sulfone (PS). This information is available free of charge via the Internet at http://pubs.acs.org/. Literature Cited (1) Vinci, D.; Donaldson, M.; Hallett, J. P.; John, E. A.; Pollet, P.; Thomas, C. A.; Grilly, J. D.; Jessop, P. G.; Liotta, C. L.; Eckert, C. A. Piperylene sulfone: A labile and recyclable DMSO substitute. Chem. Commun. 2007, (14), 1427–1429. (2) Donaldson, M. E.; Mestre, V. L.; Vinci, D.; Liotta, C. L.; Eckert, C. A. Switchable Solvents for in-Situ Acid-Catalyzed Hydrolysis of β-Pinene. Ind. Eng. Chem. Res. 2009, 48, 2542–2547. (3) Jiang, N.; Vinci, D.; Liotta, C. L.; Eckert, C. A.; Ragauskas, A. J. Piperylene sulfone: A recyclable dimethyl sulfoxide substitute for coppercatalyzed aerobic alcohol oxidation. Ind. Eng. Chem. Res. 2008, 47, 627– 631. (4) Craig, D. The geometric isomers of piperylene. J. Am. Chem. Soc. 1943, 65, 1006–1013. (5) Drake, L. R.; Stowe, S. C.; Partansky, A. M. Kinetics of the Diene Sulfur Dioxide Reaction. J. Am. Chem. Soc. 1946, 68, 2521–2524. (6) Krug, R. C.; Rigney, J. A. Unsaturated Cyclic Sulfones. 4. Isomeric 2-Methyldihydrothiophene 1,1-Dioxides. J. Org. Chem. 1958, 23, 1697– 1699.
Ind. Eng. Chem. Res., Vol. 50, No. 1, 2011 (7) Staudinger, H.; Ritzenthaler, B. High polymers links, 104 AnnouncementsOn the adsorption of sulphur dioxide on ethylene derivatives. Ber. Dtsch. Chem. Ges. 1935, 68, 455–471. (8) Frey, F. E. Process of Separating Geometric Isomers of Piperylene from Each Other. U.S. Patent 2,430,395, Nov. 4, 1947. (9) Frank, R. L.; Emmick, R. D.; Johnson, R. S. cis- and transPiperylenes. J. Am. Chem. Soc. 1947, 69, 2313–2317. (10) Fried, J. R. Polymer Science and Technology; Prentice Hall: Englewood Cliffs, NJ, 1995. (11) Grummitt, O.; Ardis, A. E.; Fick, J. Thermal Dissociation of Methyldihydrothiophene-1-dioxides. J. Am. Chem. Soc. 1950, 72, 5167– 5170. (12) Morris, R. C.; Finch, H. D. E., V. Separation of Diolefins. U.S. Patent 2,373,329, April 10, 1945.
27
(13) Yamada, S.; Ohsawa, H.; Suzuki, T.; Takayama, H. Stereoselective Synthesis of (E)-Conjugated, (E,Z)-Conjugated, and (E,E)-Conjugated Dienes via Alkylation of 3-Sulfolenes as the Key Step. J. Org. Chem. 1986, 51, 4934–4940. (14) Chou, T. S. T.; Hua, H.; Chang, L. J. Study of the alkylation reactions of sulfol-3-enes. J. Chem. Soc., Perkin Trans. 1 1985, (3), 515– 519.
ReceiVed for reView April 19, 2010 ReVised manuscript receiVed June 18, 2010 Accepted June 21, 2010 IE100920Y
