A 1 in t3}!!! hhr .....u? I A fit it ‘ u: .u”. X ...V:: a. . 5.71.1. .. .1 . 31.. .. .1x 1:. gr. ‘ A 4 la ..qu .1? 1.5.2 ...-...qu l... hr :7 $5. .. .. purgié. ‘ z ....A . .331... . a a ... ‘ ,2 I; t A. I \ sir: (a. at... 3?... 5: 5.1.2.1.: (w .. é . (r 3!? :5 Al: I... ...... .. (1.11:: t 3:. El .... P . 2,: “33:, I S I ...O) .x 1..» :7!‘ 2x39.» .1 I! “va5 ...Q1 1": \ voter {.12 11.0 . wifiueufu. s4... 5 :flxxa $1.:- :\I¥.i;)s..t!. 3 ...-V2 Is. .132. .20: t)...£t.l: : 7o) xbsv sxviwu. -..Oxi59:. v r. xx . ,. ... t .t. ‘._.......¢..>E: ,bintrii-\rc Ewuvflgukkh (u a (1530...... I. in... ..h o. ...r L -..- 8,152}. I .a to»! ....) 4! (125.9411: 3. .5549»... rizud "Vi THEsm: 'f/ "I ‘/ C7 ,‘ Date weasm LIBRARIES ““ “H“ “ “ “““““3““3“““3“3““3 “““ “3“““ This is to certify that the thesis entitled A Preliminary Study of the Conversion of Glycerol To 1,3-Propanediol presented by Scot J. DeAthos has been accepted towards fulfillment of the requirements for M.S. Chemical Engineering degree in W/ Major professor 8/29/95 0-7 639 MS U is an Affirmative Action/Equal Opportunity Institution LIBRARY Michigan State University ' PLACE ll RETURN BOX to remove this checkout from your record. To AVOID FINES return on or More dde due. usu leAn Afflnnetlve Action/Ewe) om» mailman .1 A PRELIMINARY STUDY OF THE CONVERSION OF GLYCEROL TO 1,3-PROPANEDIOL BY Scot J DeAthos A THESIS Submitted to Michigan State University in partial fulfillment of the requirements for the degree of MASTER OF SCIENCE Department of Chemical Engineering 1 995 ABSTRACT A PRELIMINARY STUDY OF THE CONVERSION OF GLYCEROL TO 1,3-PROPANEDIOL By Scot J DeAthos A novel process has been developed to convert glycerol to 1,3- propanediol. The reaction scheme proceeds in four steps. First, the primary hydroxyl groups of glycerol are protected from dehydroxylation by forming 1,3- benzylidene glycerol. Next, the secondary hydroxyl group is replaced with tosylate, a better leaving group. Hydrolysis of the dioxane ring is the third step, yielding 2-tosyloxy-1,3-propanediol. Finally, the tosylate from the second carbon is removed to produce 1,3-propanediol. Experiments have been designed and completed to test the potential for the proposed reaction scheme. The current overall 1,3-propanediol yield is 27%. Significant increases in yield may be obtained with further study. The economic potential of this process was investigated. Success of this process is determined by reducing the selling price of 1,3-propanediol below the current market price of $12.00 per pound. At yields above 75%, the selling price of 1,3-propanediol is less than $10.00 per pound. ACKNOWLEDGMENTS I would like to thank Dr. Martin C. Hawley for his guidance in this investigation. Keyi Wang and Todd Furney are also recognized for their technical expertise and their willingness to share equipment and space during the development of this process. I would also like to thank The Amoco Foundation, Consortium for Plant Biotechnology Research, Michigan State University Crop and Food Bioprocessing Center and the Chemical Engineering Department at Michigan State University for their financial support. TABLE OF CONTENTS LIST OF TABLES ................................................................................................ vi LIST OF FIGURES ............................................................................................. vii KEY TO ABBREVIATIONS .................................................................................. ix INTRODUCTION .................................................................................................. 1 Literature Review .............................................................................................. 3 Objective ........................................................................................................... 6 EXPERIMENTAL RESULTS AND DISCUSSION ................................................. 8 Preliminary Study .............................................................................................. 8 Protection ........................................................................................................ 10 Tosylation ........................................................................................................ 13 Hydrolysis ....................................................................................................... 1 5 Hydrogenation ................................................................................................. 1 7 ECONOMIC STUDY ........................................................................................... 20 Introduction ..................................................................................................... 20 Summary ......................................................................................................... 20 Discussion ....................................................................................................... 22 Process Tour ............................................................................................... 22 Capital Investment ....................................................................................... 23 Manufacturing Cost ..................................................................................... 24 Raw Materials .......................................................................................... 24 Utility ........................................................................................................ 24 Fixed Manufacturing Cost ........................................................................ 25 Sensitivity Analysis ...................................................................................... 26 Trouble Areas .............................................................................................. 27 SUMMARY .......................................................................................................... 29 FUTURE INVESTIGATION ................................................................................. 31 Appendix A--Preliminary Procedures .............................................................. 35 Hydrolysis Procedure .................................................................................. 35 Hydrogenolysis Procedure .......................................................................... 35 Appendix B 1,3-Propanediol Procedures ....................................................... 37 Protection Procedure ................................................................................... 37 Tosylation Procedure ................................................................................... 40 iv Hydrolysis Procedure .................................................................................. 40 Hydrogenolysis Procedure .......................................................................... 41 Appendix C Design ..................... ' ................................................................... 43 Manufacturing Costs .................................................................................... 46 Raw Material ............................................................................................ 46 Utility .......... ‘ .............................................................................................. 4 6 Process Equipment ...................................................................................... 48 Sensitivity Analysis ...................................................................................... 51 List of Assumptions ...................................................................................... 59 LIST OF REFERENCES ..................................................................................... 60 LIST OF TABLES TABLE 1 HYDROGENOLYSIS RESULTS. YIELDS ARE CALCULATED FROM HYDROLYSIS STARTING MATERIAL. ...................................................................................... 19 TABLE 2 ECONOMIC SUMMARY TABLE. ................................................................... 21 TABLE 3 HEATS OF FORMATION FOR EQUILIBRIUM CALCULATIONS. ACTUAL VALUES WERE TAKEN FROM PERRY'S.21 ....................................................................... 43 TABLE 4 COMPLETE SUMMARY OF THE PROPOSED DESIGN. ...................................... 45 TABLE 5 RAw MATERIAL COST. .............................................................................. 46 TABLE 6 PROCESS EQUIPMENT FOR THE PROTECTION REACTION ............................. 48 TABLE 7 PROCESS EQUIPMENT FOR THE TOSYLATION REACTION. ............................ 49 TABLE 8 PROCESS EQUIPMENT FOR HYDROLYSIS REACTION. .................................. 50 TABLE 9 PROCESS EQUIPMENT FOR THE HYDROGENOLYSIS REACTION. ................... 50 vi LIST OF FIGURES FIGURE 1 DIRECT CONVERSION OF GLYCEROL TO 1 ,3-PROPANEDIOL. ......................... 3 FIGURE 2 WALBORSKY'S REACTION TO PRODUCE 1,2- AND 1,3-PROPANEDIOL. ........... 4 FIGURE 3 CONVERSION OF ACROLEIN TO 1 ,3-PROPANEDIOL ...................................... 4 FIGURE 4 SYNTHESIS SCHEME To CONVERT GLYCEROL TO PD. .................................. 7 FIGURE 5 MODEL PATHWAY TO PRODUCE 1,2-PROPANEDIOL ..................................... 8 FIGURE 6 1 ,2-PROPANEDIOL HYDROGENOLYSIS PRODUCT CHROMATOGRAM. ............ 10 FIGURE 7 PROTECTION REACTION TO FORM 1,3-BENZYLIDENE GLYCEROL ................. 1 1 FIGURE 8 GAS CHROMATOGRAM OF PROTECTION REACTION PRODUCT PRIOR TO CRYSTALLIZATION. .......................................................................................... 12 FIGURE 9 GAS CHROMATOGRAM OF PROTECTION REACTION PRODUCT AFTER CRYSTALLIZATION. .......................................................................................... 12 FIGURE 10 PROTON NMR OF THE PROTECTION REACTION PRODUCT - 1,3- BENZYLIDENE GLYCEROL ................................................................................. 13 FIGURE 11 CARBON 13 NMR OF PROTECTION REACTION PRODUCT - 1,3-BENZYLIDENE GLYCEROL. .................................................................................................... 13 FIGURE 12 TOSYLATIQN REACTION TO PROVIDE A BETTER LEAVING GROUP. .............. 14 FIGURE 13 PROTON NMR OF 2-TOSYLOXY-1,3-BENZYLIDENE GLYCEROL. ................ 15 FIGURE 14 HYDROLYSIS OF 2-TOSYLOXY-1 ,3-BENZYLIDENE GLYCEROL. ................... 15 FIGURE 15 HYDROGENATION REACTION TO ELIMINATE THE TOSYLOXY GROUP. .......... 17 FIGURE 16 STAINLESS-STEEL REACTOR FOR HYDROGENATION REACTION. ................ 36 FIGURE 17 DEAN-STARK APPARATUS USED FOR THE PROTECTION REACTION TO PRODUCE 'I ,3-BENZYLIDENE GLYCEROL TAKEN FROM ADVANCED PRACTICAL ORGANIC CHEMISTRY. .................................................................................... 38 FIGURE 18 CALIBRATION CURVE OF GC AREA AND PD CONCENTRATION. ................. 42 FIGURE 19 PROCESS FLOW DIAGRAM FOR THE CONVERSION OF GLYCEROL TO 1,3- PROPANEDIOL. ............................................................................................... 44 FIGURE 20 EFFECT OF 1,3-PROPANEDIOL PRODUCTION RATE ON THE SELLING PRICE. THERE IS LITTLE CHANGE IN SELLING PRICE AFTER A PRODUCTION RATE OF 1 TON PER DAY. ....................................................................................................... 51 vii FIGURE 21 EFFECT OF 1,3-PROPANEDIOL YIELD ON SELLING PRICE. CURRENT YIELD IS 27% . ........................................................................................................... 52 FIGURE 22 EFFECT OF P-TOLUENE SULFONYL CHLORIDE COST ON SELLING PRICE ...... 52 FIGURE 23 EFFECT OF RAW MATERIAL COST ON SELLING PRICE. ............................... 53 FIGURE 24 EFFECT OF VARIABLE MANUFACTURING COST ON SELLING PRICE. ............. 53 FIGURE 25 EFFECT OF CAPITAL INVESTMENT ON SELLING PRICE. .............................. 54 FIGURE 26 EFFECT OF THE SIZE OF THE HYDROLYSIS REACTOR ON SELLING PRICE. 54 FIGURE 27 EFFECT OF THE SIZE OF THE HYDROGENOLYSIS REACTOR ON SELLING PRICE. ........................................................................................................... 55 FIGURE 28 EFFECT OF SELLING PRICE ON R.O.I. THE BREAK-EVEN SELLING PRICE Is $8.19 PER POUND .......................................................................................... 55 FIGURE 29 EFFECT OF PYRIDINE RECOVERY ON 1,3-PROPANEDIOL SELLING PRICE. 56 FIGURE 30 EFFECT OF BENZENE AND LIGROIN REAGENT RECOVERY ON 1,3- PROPANEDIOL SELLING PRICE. ......................................................................... 56 FIGURE 31 EFFECT OF TOTAL MANUFACTURING COST ON ROI. ................................. 57 FIGURE 32 EFFECT OF FIXED CAPITAL COST ON ROI. ............................................... 57 FIGURE 33 EFFECT OF LESS EXPENSIVE RAW MATERIALS ON PD SELLING PRICE. ...... 58 VIII KEY TO ABBREVIATIONS _bbreviation Dficjption BG Benzylidene glycerol C13 Carbon 13 DMDT 2,2-dimethy-4yl methyl-1,3-dioxolan p-toluene sulfonate Ha/M Hydrogen and a metal catalyst HPA 3-Hydroxypropanal NAD Nicotinamide adenine dinucleotide PD 1 ,3-Propanediol PhCHO Benzaldehyde TSCI p-Toluene sulfonyl chloride TSOH p-Toluene sulfonic acid °C Degrees Celsius ea Each 9 Gmms h Hours Liters kW Kilowatt MJ Megajoules ’ mg Milligram ml Milliliter M Molar N Normal lb Pounds (weight) ix psia psig sqft AH ‘10 ROI U.S. Pounds per square inch (absolute) Pounds per square inch (gauge) Square feet Heat capacity Change in enthalpy Joules Molar flow rate Heat duty Temperature Return on investment United States of America INTRODUCTION A large portion of our chemical industry depends on fossil fuels such as petroleum and natural gas. This dependence is a severe handicap because of their limited amount. Until recently, fossil fuels were an acceptable fuel source because of the affordability of fossil mass. Now, however, the increase in cost of these materials has made it imperative to search elsewhere for fossil mass replacements. A replacement Should come from a renewable source. This would ensure that depletion and therefore high cost do not occur in the future. Recent emphasis has been placed on the development of methods to convert biomass to organic solvents, alternative fuels and polymer intermediates. Successful methods may replace the need for petroleum in the chemical industry. Biomass, unlike fossil fuel, is a renewable resource. The majority of biomass consists of forestry and agriculture residues, and municipal solid wastes. It has been estimated that the US. generates about 1 billion dry tons of these residues and wastes annually.1 Grohman et al. report that the annual world biomass production is as high as 1012 dry tons.2 There exists a large untapped renewable resource. Because of the lack of reasonable methods to economically convert biomass to useful chemicals, much of the biomass is discarded as waste. Biomass conversion to polymer intermediates is becoming more promising because of improvements in biotechnology and organic chemistry. Sugars are currently being chemically converted and bioprocessed to polymer intermediates and solvents. If these new processes are successful, the selling price may decrease. An example of an expensive polymer intermediate is 1,3- propanediol (PD), which costs over $12.00 per pound. PD also has possible uses as solvents and additives for lubricants. If the cost of PD could be reduced Significantly, demand would likely increase. PD is a long sought after polyester intermediate. Recently, Shell Chemical used PD to produce a new polyester fiber called Corterra.3 The common propanediol is 1,2-propanediol which is much less expensive than PD and easily produced. Common monomers used in polyester production are ethylene glycol, 1,4-butanediol and 1,6-hexanediol. When PD is used as a monomer instead, the polyesters are not as rigid as when ethylene glycol is used nor as flexible as when the larger diols are used. To decrease the cost of PD, an inexpensive starting material must be found. Currently, hydrogenolysis of sugars from biomass is being widely investigated. Some of the desired products of this process are glycerol, propylene glycol, and ethylene glycol. Glycerol is an example of an inexpensive starting material for PD production, at $1.00 per pound.4 Unfortunately, the selectivity for hydrogenolysis of sugars to glycerol is poor, but may be improved by using properly developed catalysts. If in the future selectivity for glycerol is increased, its cost will likely decrease. Therefore, glycerol is a good choice for a PD starting material. Literature Review 1,3-Propanediol is synthesized either by organic processes or, more recently, by bacterial fermentation. The proceSs of converting glycerol to PD involves the dehydroxylation of the secondary hydroxyl group (Figure 1). Unfortunately, the procedure of direct conversion produces 1,2-propanediol with negligible selectivity for 1,3-propanediol. OH A A -water H OH OH O 0“ Glycerol 1,3-Propanediol Figure 1 Direct conversion of glycerol to 1,3-propanediol. The reduction of ethyl glycidate with lithium aluminum hydride has been reported as a method of PD production.5 Figure 2 Shows this reaction. This method of synthesis only has a conversion of 56% and a selectivity of 8% for 1,3-propandiol. Therefore, it is not economically sound to utilize this process at a large scale. 4 OH o\ LiAlH4 A - I" + W E 560° OH OH OH 8% Figure 2 Walborsky's reaction to produce 1,2- and 1,3-propanediol. PD also is produced from the hydration of acrolein in aqueous solution over a fully hydrated alumina-bound zeolite. The product of this reaction is 3- hydroxypropanal (HPA) which is then hydrogenated to 1,3-propanediol in the presence of Raney nickel (Figure 3). Unruh et al. report results for acrolein to HPA for continuous and batch processes.6 The highest selectivity and conversion for the continuous process are 90% and 77%, respectively. Over 98% of unreacted acrolein is accounted for at the end of the reaction. The optimum selectivity and conversion for a batch process occurred at temperatures around 60°C and are 70.5% and 59.4%, respectively. No mention is made for the accountability of unreacted acrolein in this case. Acrolein 3-Hydroxypropanal 1 ,3-Propanediol Figure 3 Conversion of acrolein to 1,3—propanediol. Hydrogenation of the aqueous solution from acrolein hydration was performed next to produce PD. The optimum results occur when hydrogen is introduced at 250 psig at a temperature of 40°C. The selectively and conversion are 100% and 95%, respectively. The yield from a continuous process is calculated to be 83%, while a batch process yields 65%. Spectrum Chemical Manufacturing Corporation quotes the price of acrolein at $51.00 per pound for a 1 Kg package. Larger quantities cannot be purchased, according to sales representatives at Spectrum, because acrolein is hazardous to ship in large quantities. Therefore, organizations who use acrolein normally synthesize it on site. The production cost of acrolein was unavailable. Other methods of converting glycerol to 1,3-propanediol have been reported. One process produces 1,3-propanediol by reacting glycerol and synthesis gas in a basic organic solvent in the presence of tungsten and a group Vlll metal-containing catalyst. Che reports that the reaction yields 1,3- propanediol, 1,2-propanediol, and n-propanol after 24 hours.7 The 21% PD selectivity and long reaction time make this process undesirable. Recently, there has been in-depth research on producing 1,3-propanediol by bacterial fermentation of glycerol.“9 In most organisms, C02, acetate, butyrate, lactate, and ethanol are by-products of glycerol fermentation, and thus a cause of carbon loss. The theoretical yield is limited because of the by- products. The by-products of fermentation are necessary to reduce NAD needed in the reduction of glycerol to 1,3-propanediol. Another problem with fermentation is difficult product recovery. Generally, high product concentrations inhibit either cell growth or the fermentation process. This necessitates dilute product concentrations which implies high volume fermentors and costly post- reaction concentrating processes. In comparison to fermentation, catalytic hydrogenation conserves all carbon atoms, higher concentration of products may be achieved and reactions are relatively fast. Objective The purpose of this study is to develop an alternative process to produce PD from glycerol that competes with the current market price. To satisfy this objective, two goals were met. These goals are: 1. Develop and test a novel reaction scheme to convert glycerol to PD. Attempts at direct conversion of glycerol to 1,3-propanediol using dehydroxylation reactions were unsuccessful. The primary hydroxyl groups on glycerol were preferentially dehydrated to yield 100% 1,2-propanediol. Therefore, another method had to be developed. This method is a multistep process. Figure 4 shows the proposed scheme. The first step is to protect the primary hydroxyl groups from dehydroxylation. Next, the secondary hydroxyl group is replaced with tosylate to provide a good leaving group. The third step is to hydrolyze the carbon-oxygen bond in the dioxane ring. Finally, the carbon- oxygen bond in hydrogenized to replace the tosylate with hydrogen. Experiments are performed to verify that the proposed path yields PD. H s +PI'IC! Io +TsCl +HQO / H" +Hz / M X H+ (E03333 -HCI ”E PhCHO TSOH A03 OH OH 0333333 OH Figure 4 Synthesis scheme to convert glycerol to PD. 2. Investigate the economics by developing a preliminary design of a plant for this process. An economic study was completed to determine if this process is able to compete with current PD production processes. The parameter used for comparison to current processes is the selling price of PD. A plant was designed to calculate the selling price by determining the manufacturing and capital costs needed to support the proposed process. EXPERIMENTAL RESULTS AND DISCUSSION Preliminary Study Detailed procedures for the protection of the primary hydroxyl groups of glycerol and tosylation of this intermediate already existed.‘°'11 The last two steps had to be developed. Unfortunately, the intermediates produced during the reaction scheme could not be purchased commercially. For this reason, an analogous model pathway, whose intermediates could be purchased, was developed to determine if the last two steps in the reaction scheme (Figure 4) were possible. This model consists of hydrolysis then hydrogenolysis to produce 1,2- propanediol. Figure 5 shows the reaction pathway. The tosyl intermediate similar to the primary reaction tosyl intermediate is 2,2-dimethy-4yl methyl-1,3-dioxolan p-toluene sulfonate (DMDT). OH OH OTs H2804 '/|\3 H2/M 3/|\| ——> ——> -Acetone NaOH 0 O >< OH OTS OH Figure 5 Model pathway to produce 1,2-propanediol The first step of the model is hydrolysis of the tosylate. Attempts at hydrogenolysis prior to hydrolysis were unsuccessful. Showler presents a general procedure for acid hydrolysis.12 A detailed procedure may be reviewed in Appendix A. DMDT was placed in a glass vial equipped with a magnetic stirring bar. A ratio of 1 ml of 0.01 N p-toluene sulfonic acid for 60 mg of DMDT was added to the Vial. The reaction mixture was stirred at 80°C until DMDT was dissolved and a clear solution was obtained. After the reaction was complete, the odor of acetone was present. Acetone presence was confirmed by gas chromatography (GC). Acetone was removed by purging with air, then the reaction mixture was neutralized with 0.1 N sodium hydroxide. To hydrogenize the tosyloxy group, the neutralized solution was placed in a stainless steel reactor (Figure 16, Appendix A) with Raney nickel catalyst and a magnetic stir bar. Hydrogen gas was added to obtain 250 psig. The reactor was heated to 150°C. A GC standard of 1,2-propanediol was measured for comparison to the reaction mixture. Figure 6 shows that a peak at 1.89 minutes appeared during hydrogenolysis. The standard confirms that 1,2-propanediol elutes at 1.89 minutes. Quantitative measurements of 1,2—propanediol were not completed. However since 1,2-propanediol is formed, this method is feasible. Since the model pathway was successful in producing 1,2-propanediol, the first two steps of the primary reaction scheme were tested. These are the protection and tosylation steps. ‘IO figure 6 1,2—Propanediol hydrogenolysis product chromatogram. Protection Dehydroxylation of the secondary hydroxyl group proceeds with very low selectivity; the primary hydroxyl groups are much more reactive. For this reason, benzylidene glycerol (B6) is proposed as an intermediate for the conversion of glycerol to PD. BG is formed when glycerol is reacted with benzaldehyde in benzene and ligroin reagent under acidic conditions. The procedure may be reviewed in detail in Appendix B. This procedure is a modification of Baggett et al.10 The reaction protects the primary hydroxyl groups by forming a dioxane ring. This leaves only the secondary hydroxyl group exposed for elimination (Figure 7). Figure 7 Protection reaction to form 1,3-benzylidene glycerol Separation of 86 from 1,2-benzylidene glycerol, solvent, and starting material is done by crystalliZation and filtering.13 This procedure may also be reviewed in Appendix B. Evidence exists that 1,2—benzylidene glycerol and BG exist in equilibrium.'3'“"15 Therefore, interconversion is possible. Also, 1,2- benzylidene glycerol does not crystallize under these conditions.13 Thus, the filtrate is recycled to the reactor. The protection reaction yields both the cis and trans diasteriomer of BG. However, only the cis-isomer is isolated by fractional crystallization.16 A cis- trans-complex is also crystallized in a smaller amount; the trans-isomer only crystallizes as a complex with the cis-isomer. Identification of benzylidene glycerol is made by melting point, GC analysis and proton and C13 NMR. The melting point is 68-75°C. Baggett et al. reports the melting point of the cis-isomer to be 79.5-80.5°C.10 The cis-trans-isomer complex has a melting point of 63- 64°C.16 It appears that both isomers are present as crystals. This is acceptable Since both isomers can be converted to PD. GC analysis during the reaction Shows that three products are formed. These products are believed to be the cis and trans isomers of B6 and 1,2- 12 benzylidene glycerol mentioned earlier. Figures 8 and 9 Show the GC analysis of products before and after fractional crystallization, respectively. The peak at a retention time of 3.50 minutes is believed to be 1,2-benzylidene glycerol Since it does not crystallize (Figure 9). The peaks at 3.65 and 3.83 minutes are the isomers of BG. The GC method may be reviewed in Appendix B. Proton and C13 NMR provided further verification that BG was isolated (Figures 10 and 11). Juaristi at al. reports the proton and C13 NMR results for comparison.17 The experimental and Juaristi’s NMR results are in agreement. Figure 8 Gas Chromatogram of protection reaction product prior to crystallization. fir"- [.391 ’ 3.33. 8.??3 Figure 9 Gas Chromatogram of protection reaction product after crystallization. 13 L I LN I “LA T"' I""T'** I fi—o rrrr'; ' 'I ' '1 "TT'* *T' T 'I“ - a I) I! H I. 9 l 7 5 5 e .‘I : I 0 one Figure 10 Proton NMR of the protection reaction product - 1,3-benzylidene glycerol. . A JILA .A- A1 ... J "J AMA _4MA‘. - A“; V... _ 'v— fi' r—— v—v ' vv—vvv v w .Vv ,— ——- IIIIIII‘FFITITIUIII{IIIITIIIo00:!uloTIIITIIIIIIIITIIIII.I]ililngITIITIIIIITIIIIIoil’IIT].saololl. 180 160 140 120 100 80 60 4o 20 mm. figure 11 Carbon 13 NMR of protection reaction product - 1,3-benzylidene glycerol. The yield of cis-BG is 23.1%, which is comparable to literature values."3 However, because of recycling, emphasis was not placed on improving yield. Tosylation The second step is to replace the hydroxyl group with a better leaving group. A tosyloxy group, a very good leaving group, is proposed to replace the secondary hydroxyl group (Figure 12). The method of tosylation is a modification 14 of the Feiser et al. procedure.11 BG, and TSCI are allowed to react in pyridine at 10°C. Pyridine functions as a base and solvent for this reaction. The detailed procedure of this replacement and isolation of 2-tosyloxy-1,3-benzylidene glycerol may be reviewed in Appendix B. Figure 12 Tosylation reaction to provide a better leaving group. 2-Tosyloxy-1,3-benzylidene glycerol was confirmed as the reaction product by proton NMR (Figure 13). Peaks occurring in the product sample are in agreement with published NMR data.17 The product is isolated as a fine white crystal. Juaristi et al. reports yields as high as 98%.17 However, yields achieved in this lab were no higher than 80%. Pyridine is proposed to be separated from water and recycled to the reactor. Experiments for this separation have not been performed. This may prove to be difficult; an azeotropic mixture is formed between pyridine and water.19 In HySim computer simulations, distillations of pyridine and water under vacuum would not break the azeotrope. Figure 13 Proton NMR of 2-tosyloxy-1,3-benzylidene glycerol. Hydrolysis Hydrolysis of the dioxane ring is catalyzed by aqueous acid.12 Figure 14 shows the acid-hydrolysis reaction. Methods exist for dioxane hydrolysis by using sulfuric and hydrochloric acid.‘°'12 A detailed procedure is reviewed in Appendix B. OT T S O +H20 / Fr A —> E0 -PhCHO N OH OH Ph Figure 14 Hydrolysis of 2-tosyloxy-1,3-benzylidene glycerol. 16 The insolubility of the starting material in water enabled the reaction progress to be monitored. When the solid 2-tosyloxy-1,3-benzylidene glycerol was no longer seen in the reaction vessel, the reaction was determined to be complete. Verification of this is difficult by GC, as the starting material and product could not be detected with the current detector. Therefore, other methods will have to be used to determine yield. Benzaldehyde is formed during the hydrolysis of 2-tosyloxy-1,3- benzylidene glycerol. Since the reaction occurs in an aqueous medium, the immiscible benzaldehyde settles to the bottom of the reaction vessel when cooled. This allows for easy removal of benzaldehyde from the 2-phase mixture. Benzaldehyde was found in only trace amounts in the aqueous phase by GC. Methods of benzaldehyde extraction from water by various organic solvents are reported in the literature but were'found unnecessary.10 The easy and nearly complete separation of benzaldehyde from an aqueous medium allows for a simple recycle of benzaldehyde back to the protection reaction. The solubility of 2-tosyloxy-1,3-propanediol in benzaldehyde was not determined, but is believed to be low. Isolation of 2-tosyloxy-1,3-propanediol was attempted. An extraction method was tried. Hexanes and methanol were used to extract the hydrolysis product. 2-tosyloxy-1,3-propanediol was expected to have a greater solubility in the more polar methanol. Some solubility was also expected in hexanes because of the hydrophobic toluene group. The alcohol phase was removed and dried. A salt formed after drying; this is believed to be sodium tosylate. The extraction 17 and drying process was repeated six times. The precipitated salt was removed and the remaining oil was analyzed by GC and proton NMR. The NMR spectra contained peaks attributed to 2-tosyloxy-1,3-propanediol.18 However, other peaks were also present. The results from GC analysis also Show an impure product. Hydrogenation The final step in the series is hydrogenation of the carbon-oxygen bond to yield PD. Figure 15 Shows the reaction. This reaction is catalyzed by nickel. Since the hydrolysis product could not be purchased or isolated, the hydrolysis reaction mixture was used as the starting material for hydrogenolysis. The procedure for this reaction may be reviewed in Appendix B. Ts +H2/M a -TSOH O—l GI O~l G—l Figure 15 Hydrogenation reaction to eliminate the tosyloxy group. Table 1 shows a summary of the results of hydrogenation. Two different catalysts were studied to determine their effect on product yield. Nickel is suspected to be easily poisoned by sulfur. Since TSCI and TSOH contain sulfur, nickel may not be a good choice for a catalyst. Therefore a ruthenium catalyst, also known as a good hydrogenation catalyst, was used.20 The advantage of the 18 ruthenium catalyst is that a treatment of ruthenium with sulfur does not completely inhibit catalysis. The amount of nickel poisoning was not determined because of the inability to isolate the hydrolysis product. However, a study was performed by varying the amounts of catalyst with respect to the starting material. In doing so, it is expected that poisoning occurs if the maximum yield at any given time decreases as the ratio of catalyst to starting material decreases. The highest yield achieved is 27% with excess nickel. The yields of the hydrogenolysis reaction (Table 1) were calculated with respect to the starting material in the hydrolysis reaction; the hydrolysis product was not separated or purified prior to entering the hydrogenolysis reactor. Yields were calculated based on the maximum amount of PD produced at any given time. The yield increased as the ratio of nickel catalyst to hydrolysis starting material increased. PD yield when using ruthenium as a catalyst stayed constant at 1.0%. n-Propanol is detected during the reaction when either catalyst was used. This is produced when PD is hydrogenolyzed irreversibly. During hydrogenolysis, the maximum of PD is produced within two hours. After this time, the amount of PD decreases in the reactor and n-propanol increases. Other products that may appear under hydrogenolysis conditions are methane and ethane. Since these are gases at room temperature, they could not be detected by the current analytical technique. 19 Table 1 Hydrogenolysis Results. Yields are calculated from hydrolysis starting material. Catalyst mole ratio Temperature Yield cat:tosyl 60-62% Ni on kiesgguhr 0.7:1 40 °C 5.6% 0.13:1 85 °C 1.0% 07:1 85 °C 19% 9:1 85 °C 27% 5% Ru on Carbon 0.1 :1 85 °C 1.0% 0.3:1 85 °C 1.0% 1:1 85 °C 1.0% ECONOMIC STUDY Introduction A plant has been designed for a batch process to produce 1 ton of 1,3- propanediol per day from glycerol. This process has promise because an inexpensive starting material such as glycerol might lower the production cost of PD. The objective is to lower the selling price of PD significantly below the current market price and still maintain an ROI above 15%. The following is an investigation to determine if the proposed process is economically feasible. This study does not include specific equipment specifications or any optimization. However, preliminary design of the plant is completed to determine if further research hours and capital may be warranted. Summary The proposed plant consists of a multi-step synthesis of PD from glycerol. A process flow diagram may be reviewed in Appendix C. Equilibrium calculations were completed to determine if the last two steps may result significant yields; the reaction theoretically may go to 100% completion (Appendix C). Recycle streams are proposed where possible to minimize raw material cost. The total capital investment and the discounted cash flow rate are $4.5 million and 20%, 20 21 respectively. PD has a current market price of $12.00 per pound. Under the proposed batch process, the theoretical price calculated at 100% yield could be as low as $9.61 per pound. To achieve this price, four complete batch reactions are proposed per day with a daily production rate of 1 ton. Table 2 Shows a summary of various economic values calculated for the proposed batch process. Table 2 Economic Summary Table. Current MaIket Price $12.00 / lb Calculated SellingPrice (100% yield) $9.61 Raw Materials Cost $3,800,000 Manufacturing Cost $5,740,000 Total Capital Investment $4,520,000 Return On Investment (ROI) 23.5% Payout Period 3.9 years The basis for this design is Production rate of 1 ton/day. DCF of 20%. Solvent, acids and bases, and unreacted starting material are 100% recycled. Yields are 100% unless otherwise specified. PI“? P Process variables for future consideration are, but not limited to, the reaction temperatures and pressures, alternative raw materials, possible reduction of reactor volumes and feasibility of the assumption of complete recycle. 22 Discussion Process Tour Glycerol, benzaldehyde and p-toluene sulfonic acid in a mixture of ligroin and benzene as solvent are fed to the protection reactor (Figure 19, Appendix C). The protection reaction occurs at 85°C. After the reaction is complete, the reaction mixture is transferred to a crystallizer to precipitate the 1,3-benzylidene glycerol. After 1 hour at 0°C, the slurry is transferred to the first of two plate and frame filters, where BG is separated from the filtrate. Cold benzene and ligroin are added to wash the crystals. The filtrate is recycled back to the reactor to recover benzene, ligroin, and unreacted starting materials. The crystals are removed and sent to the tosylation reactor. BG crystals are mixed with TSCI in pyridine in one of four tosylation reactors and allowed to react at 5°C. After 12 hours, the reaction mixture is mixed with ice then sent to the second plate and frame filter. Pyridine is separated from the filtrate and recycled back to the tosylation reactor. The crystals are transferred to the hydrolysis reactor. Hydrolysis of the dioxane ring occurs in the next reactor. The crystals from the second filter are combined with water and 0.01M TSOH in the hydrolysis reactor to react at 120°C. When the reaction is complete, the 2-layer mixture iS transferred to a horizontal tank to remove benzaldehyde. Benzaldehyde is then recycled to back to the protection reactor. The aqueous layer is transferred to the hydrogenation reactor. The tosyloxy group is removed in the fourth reactor at 23 85°C by catalytic hydrogenation. Hydrogen is supplied at 250 psig to the reactor. Equipment needed for the separation of PD from reaction by-products and water has yet to be determined. Capital Investment This design proposes the use of 7 reactors, 1 crystallizer, 2 plate and frame filters and 1 horizontal tank. Equipment for separation of PD from by- product in the final reaction was not determined. Overestimating the cost by 15% is considered sufficient to cover the cost of the unknown equipment. A list of major equipment and some design specifications are found in Tables 6-9 in Appendix C for each step in the process. The majority of equipment cost is attributed by the reactors. The process equipment cost was calculated using capacity exponents. Perry’s Chemical Engineering Handbook provided the exponents needed for equipment cost calculations.21 Installed costs were considered to be 1.5 times the delivered cost. All equipment is constructed of stainless steel, type 410. The cost of stainless steel is determined by multiplying the delivered cost of carbon steel equipment by 2.21 Total capital investment calculated is based on a Lang factor of 4.1. This factor assumes that the process is built on real estate already in possession. The total capital investment including 10% of the fixed capital as working capital is $4,520,000. Ten percent of the fixed capital and a ten year project life were used to estimate straight-line depreciation. A $0 salvage value after ten years is assumed. 24 Manufacturing Cost Flaw Materials The total cost of raw materials per year is $3,750,000. This accounts for 65% of the total manufacturing cost. p-Toluene sulfonyl chloride is the largest contributor at $2,820,000 per year. Table 5 in Appendix C shows the raw material unit cost and total cost. The cost of the catalyst for hydrogenolysis was neglected for two reasons. First, no cost could be found for large quantities. Secondly, the cost of catalyst regeneration is assumed to be negligible. Raw material cost may be artificially low. This design assumes that 100% recycle of benzaldehyde, benzene, ligroin, and pyridine is possible. An investigation of the sensitivity of the selling price of PD on recycle streams is discussed in the Sensitivity Analysis Section. Optimization of the amount of solvents, acids, and bases has not been determined. The amounts used were sufficient to allow for a reaction. If these amounts were decreased, the resulting selling price of PD would decrease also (see Sensitivity Analysis Section). Utility The proposed process requires heating and cooling of reaction mixtures. Reactors for the protection, hydrolysis and hydrogenolysis steps require heating. The tosylation reactor and the crystallizer operate under refrigerated conditions. An estimate of how much the energy cost for heating and cooling was calculated 25 using energy and mass balances. Eighty percent efficiency was considered when calculating both heat and refrigeration requirements. Heating and cooling were accomplished by using 100 psig saturated steam and electricity, respectively. The cost of utility is $434,000 per year. The unit cost of steam and electricity are $2.65/1000 pounds and $0.15 per kWh, respectively.22 Peters and Timmerhaus report that utility cost for a typical chemical process is approximately 10-20% of total manufacturing cost.22 The estimated utility cost falls within this range. Calculations for utility cost may be reviewed in Appendix C. Fixed Manufacturing Cost The following costs estimates were taken from Perry’s Chemical Engineers’ Handbook.21 Table 4 shows the annual fixed manufacturing cost. Operating labor for the proposed process is estimated at four people per eight hour shift at $18.00 per hour. One person will be assigned to each step of the process. The supervisor’s salary is determined to be 10% of the operating labor. Plant maintenance is estimated at 5% of the fixed capital cost. The plant and payroll overhead are 75% and 10% of the operating labor, respectively. To account for any laboratory work needed, a rate of 9% of operating labor is used. Finally, taxes and insurance are estimated at 3% of the fixed capital. The total fixed manufacturing cost is $1,563,000 per year. 26 Sensitivity Analysis A sensitivity analysis was completed to determine how process design variables affected the selling price of 1,3-propanediol and return on investment (ROI). The Figures discussed in this section may be found in Appendix C. A production rate of one ton per day was chosen as a basis for study. This rate was found to be acceptable for industrial applications The effect of a change in production rate may be seen in Figure 20. As production rate decreases the selling price of PD increases exponentially. The price is stable at the chosen basis. In fact, little change in selling price occurs at production rates above the basis. Product yields play an important roll in the calculation of product selling price. Figure 21 shows the effect of yield on selling price. At the current yield of the process, the selling price of PD is over $40.00/lb. This value is significantly higher than the current market price. At yields above 70%, little change in selling price occurs. The majority of raw material cost arises from TSCI. Figures 22 and 23 Show the effect of TSCI and raw material cost on PD selling price, respectively. It is apparent that if a leaving group could be found that is 50% less expensive than TSCI, the selling price of PD will decrease by over $1.50/lb. Variable manufacturing cost affect PD selling price similarly (Figure 24). The effect of a change in the total manufacturing cost on the percent return on investment (ROI) is large when PD selling price is fixed at $9.61/lb. Figure 31 shows that ROI is effected linearly when the manufacturing cost is changed. 27 Capital investment and individual equipment cost has little effect on PD selling price (Figures 25-27). This justifies assumptions on equipment design. A change of 50% only changes PD selling price by $0.90 per pound. ROI is also affected by a change in fixed capital cost. As fixed capital cost increases, the ROI gradually decreases (Figure 32). The effect of PD selling price on ROl can be seen in Figure 28. The break-even selling price was found by setting the annual revenue equal to zero and maintaining a ROI of 23.5%. This value is $8.19 per pound. As the selling price of PD increases, the ROI increases linearly. Figures 29 and 30 were developed to determine how an error in the assumption of 100% recovery for recycle affects the selling price of PD. If 100% of pyridine is not recovered, the selling price of PD increases sharply and quickly falls short of being competitive. Ligroin and benzene recycle have similar effects on the selling price. Finally, Figure 33 shows how selling price is affected by pyridine recycle when pyridine and TSCI cost are decreased. TSCI is decreased by 80% to simulate using methane sulfonyl chloride in its place. The significant point to note here is that the minimum selling price is $5.60/lb as a result of using methane sulfonyl chloride. Trouble Areas Problems were identified during the economic investigation. These areas must be improved or verified before this process may be considered worthy of implementation. An obvious problem is the low yield of PD. Currently, the yield is 28 no higher than 27%. The overall yield achieved must be greater than 70% to compete with current processes. The recovery of all the pyridine in the tosylation step for recycle is necessary for a low selling price of PD. If only 50% of pyridine is recovered, the selling price of PD reaches $79.00 per pound. The cost of TSCI is 75% of the total raw material cost. To reduce this cost, an alternative leaving group must be found. One example is methane sulfonyl chloride. Attempts at finding the market price were unsuccessful. However, an estimate of the price difference between TSCI and methane sulfonyl chloride was determined by comparing prices of similar mole amounts given in The Aldrich Catalog.23 The prices compared are of 1 Kg quantities where methane sulfonyl chloride is found to be 80% less expensive than TSCI. If this comparison is valid for larger quantities, methane sulfonyl chloride is a good candidate to replace TSCI in the tosylation step as long as product yield are unaffected or improved. The last improvement necessary is the cost of the hydrogenation catalyst. Market prices for large quantities of nickel or ruthenium catalysts were not available. And since the extent of catalyst deactivation was not determined, the cost of the catalyst was neglected by assuming that the cost of catalyst regeneration was negligible. If this assumption is incorrect, the selling price of PD could increase. SUMMARY Two objectives were established for this investigation. 1. Develop and test a novel reaction scheme to convert glycerol to PD. 2. Investigate the economics by developing a preliminary design of a plant for this process. A novel process has been designed to produce 1,3-propanediol (PD) from glycerol. This new process proposes to convert glycerol to PD in four steps. The reaction scheme may be seen in Figure 2. The purpose of the first reaction step is to protect the primary hydroxyl groups of glycerol from dehydroxylation by forming a dioxane ring. The second reaction replaces the secondary hydroxyl group with tosylate, a better leaving group. In the third reaction, hydrolysis of the dioxane ring yields 2-tosyloxy-1,3-propanediol. The last reaction removes the tosylate from the number two carbon to give 1,3-propanediol. Experimental investigation was completed to test the potential of significant PD yield. Current PD production methods result in a PD selling price of $12.00 per pound. The proposed process has the potential of decreasing the selling price of 29 3O PD by over 20%. A preliminary design indicates a total capital investment of $4.5 million is required with a return on investment of 23.5%. The total manufacturing cost predicted iS $5.7 million a year. Nearly half of the manufacturing cost is attributed to the cost of TSCI which provides the tosylate for the second reaction. Process optimization was not completed. However, a sensitivity analysis was done to determine the effect of process variables on the selling price of PD. Results of this analysis provide direction for future investigation and may be reviewed in the following section. The highest overall yield to date is 27%. This yield results in a PD selling price of over $40.00 per pound. Improvements of experimental procedures must be accomplished before the implementation of this process. FUTURE INVESTIGATION As a result of the economic study, areas in need of further investigation were identified. These areas are: 1. Find a replacement for TsCI. Since TSCI is not able to be regenerated and it is highly expensive, a replacement is necessary to yield a low selling price of PD. A possible replacement is methane sulfonyl chloride. Preliminary investigation estimates that the cost of methane sulfonyl chloride is 80% less than TSCI on a mole basis. 2. Utilize a less expensive base than pyridine in the tosylation step. Pyridine forms an azeotrope with water. Therefore, separating the two may be difficult. A possible technique for pyridine separation may be to extract pyridine with a non-polar solvent such as benzene. Pyridine could be removed from benzene by distillation. Another method may be the absorption of water into an absorbent such as magnesium sulfate. Another possibility may be to add a third solvent that will break the azeotrope during distillation. If pyridine cannot be 100% recycled, then an alternative base must be found. 31 32 3. Optimize solvent amounts, temperatures and pressures. Experiments on the four different steps should be repeated with changes in experimental variables. Investigation into the minimum amount of solvent required was not completed. The amounts used were sufficient to produce the desired product. If the solvent amounts needed decrease Significantly, the cost of process equipment will decrease, resulting in a lower PD selling pfice. 4. Separate PD from final product stream. The isolation of PD from the final reaction mixture was not attempted. Therefore, the selling price of PD does not reflect the cost of separation. Purchasing and operating separation equipment is expensive. This equipment cost will have an affect on PD selling price and therefore must be determined to accurately predict the feasibility of the process. 5. Isolate the hydrolysis product and determine the yield. The hydrolysis product was not isolated because of the lack of analytical equipment needed to detect it. Extraction with hexanes and methanol was used in a failed attempt to isolate the hydrolysis product; proton NMR revealed impurities. Other extraction solvents Should be attempted. Analysis by HPLC should be used to detect the reaction products in the future. 6. Improve PD yield In the hydrolysis and hydrogenolysis steps. The best yield of PD obtained to date is only 27% based on the amount of hydrolysis starting material. If the assumption that the recycle of unreacted starting materials and by-product equilibrium in the first two steps is valid, 33 then the 27% yield is the overall reaction yield. This results in an estimated selling price of PD well above the current market price. Investigation into catalyst poisoning, reaction temperature and pressure, and other catalysts is necessary to increase PD yield. Also, quantification and minimization of n- propanol production must be investigated for the hydrogenolysis reaction. Another option to improve PD yield may be to conduct hydrogenolysis of 2-tosyloxy-1,3-benzylidene glycerol prior to hydrolysis. Once the tosyloxy group is bound to the dioxane ring, a choice had to be made to open the ring next or remove the tosyloxy group. The former was chosen based on initial investigation and existing methods found in the literature. However, future work should be completed to determine if this is the best route. . Determine the solubility of the hydrolysis product (2-tosyloxy-1,3- propanediol) in benzaldehyde. Benzaldehyde product in the hydrolysis reaction is recycled back to the first reactor. If the hydrolysis product is significantly soluble in benzaldehyde, the yield of PD will be reduced. If it is soluble, then extraction or distillation of benzaldehyde may be necessary prior to recycle. . Determine the effect of nonadiabatic reactors and crystallizers on utility cost. All process equipment utility costs were based on the assumption that they were adiabatic. For this reason, utility estimates may be low. Calorimetry studies of reaction mixtures will be necessary to determine the exothermic or endothermic nature of the reaction. Also, these studies will assist in 34 determining the heat capacities of the starting materials and products not found in the literature. . Calculate the utility requirement for reactor mixing. Power requirements for the mixing required in process equipment were not determined. Therefore, it is currently unknown if this cost will significantly affect the selling price of PD. Rheology studies of the reaction mixtures may be necessary to determine the Viscosity of the reaction mixtures in order to determine power requirements. APPENDICES APPENDIX A Appendix A-Preliminary Procedures 1,2-Propanediol was produced to determine if the hydrolysis and hydrogenolysis steps in the principle investigation were possible. Since detailed procedures were already developed for the protection and tosylation reactions of the principle pathway, only the analogous hydrolysis and hydrogenolysis steps were needed. Hydrolysis Procedure The procedure for hydrolysis was the first step developed. This step hydrolyzes the oxygen-carbon bond on the number two carbon of 2,2-dimethy- 4yl methyl 1,3-dioxalan p-toluene sulfonate. First, add to a glass vial with stir bar approximately 60mg of tosylate and 0.10N TSOH in 2 ml of water. Next, heat to 85°C in a water bath for 30 minutes. Once the tosylate is dissolved, the reaction is assumed to be complete. The result is a clear solution containing acetone which is an expected product. The reaction mixture is then neutralized with 0.10N NaOH and purged with air to remove acetone. Hydrogenolysis Procedure The next step is hydrogenolysis of the carbon-oxygen bond on the third carbon to eliminate the tosyloxy group. To do this, neutralize the product from the hydrolysis experiment with 0.1N NaOH. Place the neutralized mixture in a 35 36 stainless-steel reactor equipped with a stir bar. The reactor can be seen in Figure 16. Next, add nickel on kieselguhr as catalyst. Seal the reactor and heat to 150°C in an silicon oil bath. Purge the sealed reactor with nitrogen, then remove nitrogen by vacuum. Repeat the purging process several times. Once this is complete, add 250 psig hydrogen and allow to react for 1 hOur. At the end of one hour, 1,2-propanediol is present. Figure 16 Stainless-steel reactor for hydrogenation reaction. The reactor in Figure 16 has four important components. They are labeled A, B, C and D. The tubing labeled A is where hydrogen gas enters the reactor. Location B is where samples leave the reactor. The valve at point C is a two-way valve that allows for on-line sampling. The interior of the reactor has a magnetic stir bar labeled D. APPENDIX B Appendix B 1,3-Propanediol Procedures Protection Procedure Benzylidene glycerol (BG) is proposed as a precursor needed to convert glycerol to PD. To prepare BG, glycerol is combined with 2% excess benzaldehyde and 0.01N p-toluene sulfonic acid (T SOH) in a two-neck round bottom flask with a magnetic stir bar. Add a 40/60 ratio of benzene to ligroin reagent as solvent. The ideal volume of solvent has not been determined. About 3 times more solvent than benzaldehyde by volume is sufficient. In one neck, insert a thermometer with stopper and in the other attach a Dean-Stark trap. The Dean-Stark trap is used to remove water that is boiled off from the reaction vessel. Since the trap is also graduated, reaction progress may be monitored by quantifying the amount of water that has accumulated. A small condensation column is attached to the trap to avoid loosing any solvent or starting materials. Figure 17 Shows the apparatus. Next place the round bottom flask in a sand or oil bath and heat above 85°C while stirring with magnetic stir bar. Another method of determining when the reaction is complete is by observing the temperature. During the reaction, the temperature of the mixture in the flask will not rise above 85°C. Once glycerol is consumed, the temperature will rise to a new boiling point of 180°C. 37 38 1 condense: .— J J reaction flask 3:23 --— heating bath Figure 17 Dean-Stark apparatus used for the protection reaction to produce 1 ,3-benzylidene glycerol taken from Advanced Practical Organic Chemistry.“ During this temperature rise, benzene, ligroin, and benzaldehyde are being volatilized. Recovery of the solvent and unreacted benzaldehyde is essential for recycle. Therefore, prior to the temperature rise, remove all water from the Dean-Stark trap. Allow the temperature to rise and collect the solvent and benzaldehyde for future batch reactions. When the boiled off solvent and benzaldehyde were analyzed by GC, only trace amounts of BG were present. Since this was being recycled, quantification was not completed. Fractional crystallization is needed to separate 1,2-benzylidene glycerol from the isomers of 1,3-benzylidene glycerol. This procedure is presented by Hill et al.13 First, dissolve the mixed benzylidene glycerols from the reaction vessel in a mixture of ligroin and benzene. To 609 of mixed benzylidene glycerols add 150 ml of ligroin and 90 ml of benzene. Then freeze for approximately one hour. 39 Crystals may or may not be present at this time. Slowly add benzene with shaking until oil is dissolved. Crystals Should form if they have not done so already. Whether crystals have formed or not, place in freezer for an additional hour. Crystals will form during this time. Filter crystals in a Buchner funnel under a vacuum and wash with a cold 60/40 mixture of ligroin and benzene. Continue to wash crystals until the yellow color (benzaldehyde) is removed. The crystals are white. Juaristi et al. report a yield of mixed 1,3-benzylidene glycerols of 40.3%.” To increase the purity of BG, recrystallize by dissolving in the same ratio of ligroin to benzene and heat to 40°C to supersaturate. Slowly cool to room temperature then place in freezer for 1 hour. Next, repeat filtration and washing procedure above. Gas Chromatography Method Gas Chromatography analysis for the protection reaction was done using a Hewlett Packard model 5890 Series II gas chromatograph with a Scientific 30m X 0.320mm 1.80 micron DB-624 column. The injector and FID detector temperatures were set at 300°C. Initial temperature was set at 150°C for 3 minutes. Temperature was ramped at 40°C per minute until 320°C was reached, which held for 5 minutes. Chromatographs were recorded on a Hewlett Packard 3396A integrator. The solvent used for analysis of BG was benzene which has a retention time under these conditions of 1.39 minutes. 40 Tosylation Procedure The second step required for the conversion of glycerol to PD is replacing the hydroxyl group with a better leaving group on BG. A tosyloxy group provided by p-toluene sulfonyl chloride (TSCI) was chosen. The method to tosylate the fifth carbon of BG is a modification of Fieser et al.11 Combine BG with 5% excess TSCI in a glass vessel. Next add cold dry pyridine in a ratio of 20 ml to 1 g of BG. Then place in refrigerator at ca. 10°C. A white fine precipitate will form which is pyridine hydrochloride. After 12 hours or when no more pyridine hydrochloride is formed, remove from refrigerator and pour with stirring over 6. grams of ice for 1 ml of pyridine. The tosylate crystallizes immediately. Continue stirring until ice melts. Filter the crystals in a Buchner funnel under vacuum. Wash with cool water if necessary. Dry crystals in vacuo to remove water. Hydrolysis Procedure Add approximately 0.25 Molar 2-tosyloxy-1,3-benzylidene glycerol in water, 0.01 N p-toluene sulfonic acid and a magnetic stir-bar to a glass jar with a sealable top. Place the glass jar in a sand bath at 120°C for 4.0 hours. The reaction product will separate into two phases when cooled. The bottom yellow phase is benzaldehyde and the clear top phase is aqueous. 2- Tosyloxy-1,3-propanediol is believed to be in the aqueous phase. Extract benzaldehyde with a syringe and recycle to the first reaction. The aqueous phase is then neutralized with sodium hydroxide and undergoes a hydrogenolysis. 41 Gas Chromatography Method The progress of the reaction is difficult to monitor by GC. Benzaldehyde may be detected by gas chromatography. However, the two phase nature of the reaction mixture makes it difficult to extract a representative sample. Therefore, G0 was not used to monitor reaction progress. It was used to detect the purity of the starting material and by-products formation. The GC method is Similar to the one used in the protection step. A difference lies in the temperatures and holding time. The initial temperature is held at 50°C for 1 minute then ramps at 40°C per minute to 260°C. The starting material and product, 2-tosyloxy-1,3-propanediol cannot be observed by 60 under these conditions. Hydrogenolysis Procedure The last step in the series of reaction is hydrogenolysis of the carbon- oxygen bond (Figure 15) to remove the tosylate from 2-tosyloxy-1,3-propanediol. The same reactor found in Appendix A Figure 16 is used for this reaction. This reaction occurs in a basic medium with a metal catalyst. Two different catalysts were used. A comparison of these catalysts may be reviewed in Table 1. To perform the hydrogenolysis, first neutralize, with sodium hydroxide, the hydrolysis product after it is separated from benzaldehyde. Next, add a 1:1 mole ratio of catalyst to hydrolysis starting material to the reactor. Place hydrolysis product in reactor with a magnetic stir bar. Close the lid and place in oil bath at 85°C. Purge reactor with nitrogen then remove nitrogen by vacuum. Repeat 42 purging process three times then supply hydrogen at 250 psig. Allow system to react for four hours. A calibration curve of GC peak area and PD concentration was created to assist in quantifying PD yield. The two data points for each concentration represents a range with 95% confidence limits. Four measurements were taken for each data point with an injection sample of 1 III. 1 .8)E+07 1.60E+07 «- 1 1.40E+07 .. 1.20E+07 -* 1.00E+07 +- 8.(X)E+06 ‘- Q 8 m 9 GO Area (1uL hjecllon) 4.mE+O6 .. 2.00E+06 ~» 3 I .p ... 0.00E+00 1 : I f i I 0.00E+00 2.00E-03 4.00E—03 6005-03 8.00E-03 1005-02 1205-02 1.40E-02 Concentratlon (g/ml) Figure 18 Calibration curve of GC area and PD concentration. Gas Chromatograph Method The same 60 technique as earlier was used for hydrogenolysis analysis. The only difference was the initial and final temperatures were set at 100°C and 260°C, respectively. The peak for n-propanol and PD elute at 1.23 and 2.77 minutes under these conditions, respectively. APPENDIX C Appendix C Design This appendix is dedicated to presenting supporting tables, figures and calculations used in the economic study of the proposed process. Equilibrium Calculations Table 3 Heats of formation for equilibrium calculations. Actual values were taken from Perry's.21 Chemical Estimated AH. Actual AH, 1,3-propanediol -464.9 kJ/mole Hydrogen (9) -241.8 kJ/mole 2-Tosyloxy-PD 96.2 kJ/mole 2-Tosyloxy-BG 181 kJ/mole Water -68.3 kJ/mole Benzaldehyde -87.0 kJ/mole The equilibrium constant was estimated by Keq = exp(—AH%T] where R is the gas constant, T is the absolute temperature and the change in entropy is assumed to be zero. The equilibrium constant for the hydrolysis and hydrogenolysis reactions are 1018 and infinite, respectively. 43 O 32.5.. 9... 385.: £1.03}. 3:: o_.. _ U E====== U l; S: :0: 2 .06 ......s U AI .05.. z. .o o :2on no 5.339092: 8.33. .1330 tr (V “U 35.5.8.3; _ 292.853 1:. UV 9. 8... .3905... . O 8.33. o. GC.-use: c9213.. ._. 2:3: 25.: .0.» ...-...: E. AND—as: E 9%: 0.3:... 3.25.5 505$ EU o 322... an 8.8:: c3822.. ...-3 283993053 2 13:13. o .362. o H :o.» z 5.8 .H 83.. 126E i....=:§§z=s§%§.“/”\ .235 Figure 19 Process flow diagram for the conversion of glycerol to 1,3-propanediol. 45 Table 4 Complete summary of the proposed design. Production Rate 1.0 ton/day Variable Manufacturing Cost Raw Material Cost Unit Cost ($/yr) ($/yr) Glycerol $1 .00/lb. 850,000 Benzaldehyde $1 .05/Ib. 4,900 TSCI $1 .60/lb. 2,819,000 TSOH $5.86/lb. 0 Hydrogen $0.03/mole 69,300 Pyridine $3.57/lb. 0 Ligroin Reagent $0.96/lb. 0 Benzene $0.16/lb. 0 Water $0.80/1 0009al 2,700 Total Raw Materials 3,746,000 Utility 100psig steam $2.65/1000lb. 21,000 Electricity $0.15/kWh 412,900 Total Utility 433,900 Total Variable Cost 4,179,900 Fixed Manufacturing Cost Operating Labor $18.00/hr 605,000 Supervisory Labor 10% of Op lab. 60,500 Plant Maintenance 5% of Fixed Capital 206,000 Indirect Cost Plant Overhead 75% Op lab. 454,000 Payroll Overhead 10% Op lab. 60,500 Lab Work 9% Op lab. 54,400 Property Tax 2% of Fixed Capital 82,200 Insurance 1% of Fixed Capital 41,100 Total Fixed cost 1,563,000 Total Manufacturing Cost 5,742,900 Fixed Capital Cost 4,112,000 Working Capital 10% of Fixed Capital 411,000 Depreciation 10% of Fixed Capital 411,000 Total Capital 4,523,000 Federal Income Tax 35% 351,000 DCF 20% Cash Flow 1,063,000 Net Profit 651,900 Revenue 6,745,000 Selling Price ($/lb.) 9.61 Return on Investment 23.5% 46 Manufacturing Costs Raw Material Table 5 Shows a summary of the raw material unit cost and total annual cost used in the conversion of glycerol to PD. Total raw material cost is based on the assumption that 100% of solvents, acids and bases, and unreacted starting material is recycled. The effect of this assumption is reviewed in the Sensitivity Analysis section. Table 5 Raw material cost. Source Chemical Unit Price Total Cost (x1000) ($/yr) Chemical Marketing Glycerol $ 1.00/lb. 850 Reporter Chemical Marketing Benzaldehyde $ 1.05/lb. 56.3 Reporter Chemical Marketing Benzene $ 0.17/lb. 0 Reporter Chemical Marketing Pyridine $ 3.57/Ib. 0 Reporter Spectrum Chemical MFG Ligroin $ 0.96/lb. 0 Aceto Corp TSCI $ 1.55/Ib. 2,819 Spectrum Chemical MFG TSOH $ 5.87/lb. 0 AGA Gas Hydrogen gas $ 0.03/mole 69.3 Utility The reactors and crystallizers proposed are assumed to be adiabatic. Once the desired temperature is reached, it can be maintained at a zero energy requirement. Another assumption that was made is the heat capacity of the reaction mixtures is equal to the heat capacity of the solvent. Energy for heating is provided by 100 psig steam at $2.65 per 1000 lb. Electricity provides the 47 energy for cooling at $0.15 per kWh. The energy required to heat the protection reactor to 85°C is calculated by 358 AH = j open" 298 Cp = —33.92 + 0.4739 x T — 3.017E - 04 x T2 + 7.130E — 08 x T3 _ J AH _ 5,490 /mole Q = AH x m _ J l T 5’490 Knole x 35900 mo es batch _ BTU _ 197 M%atch :7 746,000 /day efficiency = 0.8 = Q = 933,000 BTU/day The heat capacity was assumed to be that of the solvent benzene”. The energy needed to cool the mixture from 85°C to 0°C in the crystallizer is 388,000 kWh/yr. Similarly, energy needed for cooling the tosylation reactor from 25 to 10°C was calculated and found to be 1,810,000 kWh/yr. This number is high because four refrigeration units are required. Water is the solvent in the hydrolysis and hydrogenolysis reactors. Therefore, the heat capacity of the reactor mixture was assumed to be that of water. The energy required to heat the hydrolysis reactor from 25 to 120°C is 7,840,000 BTU/day. The hydrogenolysis reactor is heated from 25 to 85°C. This requires 7,670,000 BTU/day. The total energy required for the proposed process costs $434,000 per year. The utility required to operate mixers, lighting and other appliances are neglected. However, a more detailed design must take these costs into consideration. 48 Process Equipment The equation used to calculate the cost of equipment is M & S(current) C t = as M & S (source) Size up x C t x [Size(reference)) 08 (reference) where M83 is the Marshall and Swift Index which is currently 1022.8 for a chemical process.26 Capacity exponents were found in Perry’s Chemical Engineers’ Handbook.21 The installed cost was determined to be 1.5 times the delivered cost. All equipment is proposed to be constructed of 410 stainless steel. The cost of this is two times that of carbon steel. Step one of the series of reactions is the protection reaction. Table 6 shows a summary of the equipment needed for this step. Table 6 Process Equipment for the Protection Reaction. Equipment Description Capacity Exponent Installed Cost Jacketed Reactor w/mixer 3,187 L 0.53 $85,500 Crystallizer 0.6 ton/batch 0.59 $20,420 Filter, Plate and Frame 600 sqft 0.55 $56,400 The first reactor is a jacketed vessel with a mechanical mixer. The size of the reactor was calculated based on the volume of benzene, ligroin reagent and benzaldehyde contained. The product is crystallized by a forced circulation crystallizer. The size of the crystallizer is based on the weight of product recovered. The crystals are filtered by a plate and frame filter. The size of the filter is an approximation based on small scale experiments. A first pass estimate 49 of the size of the filter is done using a ratio of the filter area of small scale filtration to the mass of solids and assuming that the filter area and mass of product are proportional.21 The second step in the series of reactions is tosylation. The equipment needed for this phase of the reaction is shown in Table 7. Since this reaction requires 12 to 24 hours to complete, four reactors are proposed. These reactors are equipped with mechanical refrigeration units to cool the reaction to 10°C. The size of the refrigeration units are calculated from the energy requirement.21 The energy requirement calculation may be reviewed in the Utility section of this Appendix. Table 7 Process Equipment for the Tosylation Reaction. Equipment Description Capacity Exponent Installed Cost 4 Vertical Tanks w/mixer 9,000 L ea. 0.5 $226,000 4 Mechanical RefLiqeration 400 kW ea. 0.73 $167,000 Filter, Plate and Frame 1,100 sqft 0.55 $80,000 The hydrolysis reaction occurs in a jacketed reactor with a mechanical mixer. The size of this reactor is based on the volume of pyridine used. In addition to the reactor, a tank is proposed to allow removal of benzaldehyde for the reaction mixture (Table 8). The final step requires a reactor that is able to withstand a pressure of 250 psig. For safety, a reactor is proposed for a rating of 500 psia. Table 8 Process Equipment for Hydrolysis Reaction. 50 Equipment Description Capacity Exponent Installed Cost Jacketed Reactor 11,900 L 0.53 $171,800 w/mixer Horizontal Tank 11,900 L 0.57 $26,800 The cost difference factor for the higher pressure rating was found in Perry’s.21 Table 9 shows a summary of the equipment for this step. The capacity of the reactor was calculated based on the volume in hydrolysis reactor minus the volume of benzaldehyde removed. An investigation of additional equipment needed for this step was not determined. It is anticipated that a distillation column is necessary to remove PD from water. Table 9 Process Equipment for the Hydrogenolysis Reaction. II Equipment Description Capacity Exponent installed Cost II Jacketed Reactor 7 w/mixer rated at 500 psia 11,600L 0.53 $476,000 II 51 Sensitivity Analysis The following are figures created to determine the effect of different parameters on process economics. Interpretation of these figures can be found in the Economic Study of the thesis. 35 I A 30 ‘r E. a 25 1* g 20 a. O. I E n a 10 "' . . I c» " 0 II E 3 5 4 0 1 1* I 4 i I t l . 0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 2 1,3-Propanedol Production Rate (T on/Day) Figure 20 Effect of 1,3-propanediol production rate on the selling price. There is little change in selling price after a production rate of 1 ton per day. 52 100 90 «- 70 4* 60 .0- Selllng Price of 1,3-Propanedol (slpound) 8 I 40 4* I 30 - I 20 T I . I 10 <~ . ' u 0 i i , 1 I t 4 t l 0% 1 0% 20% 30% 40% 50% 60% 70% 80% 90% 1 00% % Ybld Figure 21 Effect of 1,3-propanediol yield on selling price. Current yield is 27% . ‘5 CD ... 12 «~ 3 u i . a; 11 .. V I § . 10 «- I II S, . v: . 9 ->- ‘6 § I o. ' 8 ‘r 3 .. 8 7 .. -50% 40% -30% -20% -10% 0% 10% 20% 30% 40% 50% % Change In TsCI Cost Figure 22 Effect of potoluene sulfonyl chloride cost on selling price. 53 I! CD II A 12 "‘ I 2 § I E. 1‘ " . a . 10 1» g .. o. «5. ' 9 .. 3 I g - g . a ll 7 w -50% -40% -30% -20% -10% 0% 1 0% 20% 30% 40% 50% % Change In Raw Material Cost Figure 23 Effect of raw material cost on selling price. u g 12 ‘1’ . g I v 11 J'- I I 10 ~- II 3’3 ' 9 4 ‘6 - .fii . a g u 7 .. ll -50% -40% -30% -20% -10% 0% 10% 20% 30% 40% 50% % Change In Varlable Manufacturing Cost Figure 24 Effect of variable manufacturing cost on selling price. 54 A 12 ‘- E E; 11 -~ II I I 10 "F . I II I 3) I I v: II . 9 w "6 E § 22 7 .. -50% 40% -30% -20% -10% 0% 1 0% 20% 30% 40% 50% % Change In Capital Investment Figure 25 Effect of capital investment on selling price. 1‘ CD i 12 "" E 11 «- g 10 -~ II I I I I u I I ' ' " CL 5’3- 9 -. '5 E e~~ a .5 51 7 .. -50% -40% .3096 -20°/o 40% 0% 1 0% 20% 30% 40% 50% °/o Change In Reacbr 3 Volume Figure 26 Effect of the size of the hydrolysis reactor on selling price. 55 A 12 "" E 2 11 «~ E 10 ~ . I I Q n . ' . I IF I I '6 g .. C» .-‘-E 8 7 .. -50% -40% -30% -20% -10% 0% 1 0% 20% 30% 40% 50% % Change In Reactor 4 Volume Figure 27 Effect of the size of the hydrogenolysis reactor on selling price. 80% II 60% ~~ I 40% ~- § 20.. ._ E g 0% 1 I 1 g g 4 5 6 z 11 112 a: -20% I «10% ~~ I -60% Selling Price of 1.3-Propenedol (Slpound) Figure 28 Effect of selling price on R.O.l. The break-even selling price is $8.19 per pound. 56 so, 20 -- Selling Price of 1,3-Propanediol (WM) 10 1* 0 l l l 1 l 50.0% 1 1 1 1 I I i T l l I T I 55.0% 60.0% 65.0% 70.0% 75.0% 80.0% 85.0% 90.0% 95.0% % Pyridine Recovery Figure 29 Effect of pyridine recovery on 1,3-propanediol selling price. 15 1 00.0% d d d d d o d N 0) & L l l A l T I Y 1 Selling Price of 1,3-Propanediol (st/pound) (O i 8 1 I l l l l J 50.0% I l I l I l I l l I 7 55.0% 60.0% 65.0% 70.0% 75.0% 80.0% 85.0% 90.0% 95.0% % Recovery of Benzene and Ligroin 1 00.0% Figure 30 Effect of benzene and ligroin reagent recovery on 1,3-propanediol selling price. 57 N 5h :5 Retum on Investment (%) 65.0% + 55.0% -~ 45.0% ~~ 35.0% ~~ 25.0% 1; 15.0% + 0 5.0% '11- 60% -40% -30% 1 l 1 I -20% 40% -5.0%0‘% 45.0% *- 50% 25.0% °/o Change In Total Manufacturing Cost Figure 31 Effect of total manufacturing cost on ROI. Return on Investment 1 I I I 45.0% - 35.0% ~ I 0 25.0% - A 1' I 15.0% ~ 5.0% *- l l 1 l 1 1 -50l0°/o 410.0% I -30.0% I I I -20.0% -10.0%5_oo,90% 10.0% -15.0% *- 45.6% 20.0% I 30.0% I 40.0% 50. 0% % Change in Fixed Capital Cost Figure 32 Effect of fixed capital cost on ROI. 58 25 20 ,;- --------- Base 80% less emensive than ......... Mdim g I." _—0.0IN &0H .45; .5 .. g ........ ‘3 .......... E g... ._ a: ............ v”. 5 41- 0 I I ' I . I I . . 50.0% 55.0% 60.0% 65.0% 70.0% 75.0% 80.0% 85.0% 90.0% 95.0% 100.0% % Change in Base Recovery Figure 33 Effect of less expensive raw materials on PD selling price. 59 List of Assumptions 8. 9. . Heat capacities of reaction mixtures are assumed to be those of the solvents. Reaction vessels are assumed to be adiabatic. Equipment requiring utilities are assumed to be 80% efficient. Depreciation is 10% of Fixed Capital with a 10 year life and 0 salvage value. Federal taxes are 35%. Property for the plant is already owned. Working capital is 10% of Fixed Capital. Recycles are assumed to be 100% unless specified otherwise. Installed equipment cost is 1.5 times delivered cost. 10.A Lang factor of 4.1 is sufficient when calculating Total Capital Investment. 11.Four operators per shift is sufficient for plant operation. 12. Four batches can be produced in one day. 13. Hydrogenolysis catalyst is able to be regenerated at negligible cost. 14. Negligible effects of process scale up. 15.A production rate of 1 ton per day is marketable. LIST OF REFERENCES ‘ Barrier J. W. and Bulls M. M., Feedstock Availability of Biomass and Wastes In Emerging Technologies for Materials and Chemica_ls_from Biomass , (Edited by Rowell R. M. etal.), American Chemical Society, Washington DC 1992, pp 410-421 . 2 Grohmann, K; Wyman, C. E. and Himmel M. E., Potential for Fuels from Biomass and Wastes In Emerging Technologies for Materials and Chemicals from Biomass, (Edited by Rowell R. M. et al.), American Chemical Society, Washington DC 1992, pp 354-392. 3 Stinson, S. C., Fine and Intermediate Chemicals Makers Emphasize new Products and Prgcesses, C&E News 73(29):10. ‘ Chemical Marketing Reporter 16 January 1995. 5 Walborsky, H. M., Columbini, C., Reduction of Glycidic Esters with Lithium Aluminum Hydride, J. Org Chem, 27 2387 1962. 6 Unruh, J. D., Ryan, D. A., Nicolau, l., Method for the Manufacture of 1,S- Propanediol, US PATENT US 5093537 A, 1992. 7 Che, T. M., Satalytic Sonversion of Glycerol and Synthesis Sas to M, US PATENT US 4642394 A, 1987. 8 Heyndrickx, M., De Vos, P., Vancanney, M., De Ley, J. (1991) '_l'_I_1_e_ Fermentation of | ceroI b Iostridium bu n’cum LMG1212t and 121 t and Q. pasteurianum LMS 32S5, Appl Microbiol Biotechnol 34:637. ’ Dabrock, B., Bahl, H., Gottschalk, G. (1992) Parameters Affecting Solvent Production by Clostridium pasteun'anum, Appl Environ Microbiol 58:1233. 1° Baggett, N., Brimacombe, J. 8., Foster, A. 3., Stacey, M., Whiffen, D. H., Aspects of Stereochemistgy. Part IV.* Configuration and Some Reactions of the 1,S-O-Benylidengglygergls (S-Hydroxy-2-phenyl-1,S-diolxans), J. Chem. Soc. 1 9602574. ‘1 Fieser, L. F., Fieser, M. Reagents for Qrganic Synthesis, Vol. 1, Wiley, New York, 1967:1180. ‘2 Showler, A. L., Darley, P. A. Sondensation Products of Slycerol with Aldehydes and Ketoges. 2-Substituted m-Dioxan-S-gls and 1,S-Dioxolan§-4- Methangls, Chem. Rev., 67:427 1967. 60 61 ‘3 Hill, H. S., Whelen, M. S., Hibbert, H. Studies on the Reactions Relating to Sarbohydrates and Polysaccharides. XV. The lsomeric Benzylidene Slycerols, J. Amer. Chem. Soc. 1928, 50:2235. “ Trost, B. M. ed. Somprehensive Qrganic Synthesis, Vol 6, Pergamon Press. 1991, 659. ‘5 Piantadosi, C., Anderson, C. E., Brecht, E. A., Yarbro, C. L. The Preparation of Syclic Glycerol Apptals by Transacetalation, J. Org. Chem. 1963 28:2425. ‘6 Dobinson, B., Foster, A. B. As ects of tereochemist . Part VI.‘ confi uration of the S-Hydroxy-Z-p-nitrpphenyl-1,3-dioxans and Further Observations on trans-5-Hydroxy-2phenyI-1,S—dioxan, J. Chem. Soc. 1961, 2338. ‘7 Juaristi, E., Antunez, S. Sonformational Analysis of S-Substituted 1,S- Dioxanes. . tu of the Attractiv auche Effect in - -C-O e ments, Tetrahedron 1992 48:5941. 1° Chong, J. M., Sokoll, K. K., nthesis f2-O-Protected I cerol Derivatives Organic Prep Proc Int 25(6):639. ‘9 Budavari, s. The Mprck Indpx, 11111 ed. Merck & Co., Inc. 1989. 2° Dubeck, M, Knapp. G. (3., Sulfide-Modified Ruthenium Satalyst, us PATENT us 4430253 1934. 2‘ Perry, R. H. and Green, D. W. Per ’s hemical En inee ”Handbook, Sixth ed. McGraw-Hill 1984. 22 Peters, M. S., Timmerhaus, K. 0., Plant Design and Economics for Shemical Engineers, 4th ed. McGraw-Hill, Inc. New York 1991. 23 The Aldrich Chemical Company Satalog: Handbook of Fine Shemicals, 1992- 1993. 2‘ Casy,M., Leonard, J., Lygo, B., Procter, G. Advanced Praptica'l Qrganic Chemistm, Blackie Academic & Pofessional, Glascow 1993 pp134. 25 Reid, R. C., Prauznitz, J. M., Poling, B. The Properties of Sases and Liguids, 4th ed. McGraw-Hill New York 1987. 26 Zanetti, R. J., ed., Chemical Engineering, Vol 102:7, July 95 pp 192. MICHIGAN STATE UNIV. LIBRARIES 111111 III111111111111111111111111III 31293014107340