PLACE ll RETURN BOX to remove this checkout from your record. TO AVOID FINES return on or before due due. DATE DUE DATE DUE DATE DUE , .W.’ ’ 1' 3 a?! ' {11; n 12' MSU Is An Affirmative AotionlEqual Opportunity Institution CATALYTIC CONVERSION OF D-GLUCITOL AND ISOSORBIDE ' USING ZEOLITE CATALYSTS by Gary J. Dozeman A Thesis Submitted to Michigan State Universlty In partial fulfillment of the requirements tor the degree at MASTER OF SCIENCE Department of Chemical Engineering 1989 QOO\959 ABSTRACT CATALYTIC CONVERSION OF D-GLUCITOL AND ISOSORBIDE USING ZEOLITE CATALYSTS by Gary J. Dozeman The catalytic conversion of D-glucitoi and isosorbide (1,4:3,6-Dianhy- dro-D-Glucitol) to volatile organic products has been carried out using Y and ZSM-5 zeolite catalysts. D-glucitol was dehydrated in the liquid phase in a batch reactor with decahydronaphthalene or deionized water as the reac- tion medium. lsosorbide, a principal product of D-giucitol dehydration, was reacted in the vapor phase in a differential flow reactor capable of operating at 600°C and 1000 psig. The primary catalytic activity observed in the liquid phase was dehydration with both water and organic soluble products formed. Conver- sions of up to 98% were achieved in under 4 hours of reaction time at 260°C. Major reaction products including anhydro-alditols and furfurals were identified by GC/MS. Vapor phase catalytic activity included dehydra- tion and cracking with conversions up to 100% to gaseous products without detectable tarring. The high activity and regenerative properties of the zeo- lites could lead to a viable biomass conversion process. Dedicated to Amy E. Relsterer for her love. support. and undying faith in my abilities. Acknowlegements i would like to thank Dr. Dennis J. Miller for his support, patience, and guidance throughout my graduate studies; the Michigan State University Mass Spectrometry Facility for assistance with the mass spectral data; and the REED Bioprocessing Fund for the financial support on this project. TABLE OF CONTENTS List of Tables ......................................................................................... ix List of Figures ........................................................................................ x Chapter 1: introduction and Background ................................................... 1 1.1 Introduction ................................................................................. 1 1.2 Background ................................................................................. 3 1.2.1 Sugar Conversion Via Homogeneous Catalysis ........................ 3 1.2.2 Applications of Zeolite Catalysts ............................................ 3 1.2.3 Biomass Conversion Using Zeolites ........................................ 4 1.3 Proposed Research ...................................................................... 8 1.3.1 Proposed Reaction Pathway ................................................. 8 1.3.2 Choice of Catalysts and Conditions ...................................... 10 1.3.3 Analytical Methods .............................................................. 11 1.4 Specific Research Objectives ........................................................ 15 Chapter 2: Liquid Phase Apparatus ......................................................... 17 2.1 Catalyst Preparation .................................................................... 17 2.2 Catalyst Calcination/Reduction and Regeneration ........................... 18 2.3 Batch Reaction Apparatus ............................................................ 18 2.4 Liquid Sampling Apparatus ........................................................... 22 2.5 Reaction Product Analysis Equipment ...................................... . ..... 24 2.5.1 Product Drying and Separation ........................................ . ..... 24 2.5.2 Gas Chromatography ........................................................... 25 2.5.3 High Performance Liquid Chromatography .............................. 25 2.5.4 Gas Chromatography/Mass Spectrometry ................................ 25 Chapter 3: Vapor Phase Reactor Design and Apparatus ............................ 26 3.1 Conceptual Design of Vapor Phase Reactor ................................... 26 3.2 Vapor Phase Reaction System Design ........................................... 27 3.2.1 Pressure Vessel ............................................................ . ...... 27 3.2.2 Flange and Seal .................................................................. 28 3.2.3 inner Reactor Tube .............................................................. 30 3.2.4 Reactor Furnace ................................................................... 32 3.3.5 Temperature Measurement and Control ................................... 32 3.2.6 Gas Flow System .................................................................. 35 3.2.7 Reactant Pumping System .................................................... .- 37 3.2.8 Product Recovery System ...................................................... 38 Chapter 4: Experimental Conditions and Procedures .................................. 40 4.1 Catalyst Preparation .............................................................. . ..... 40 4.1.1 lon-Exchange/Metal Loading .................................................. 41 4.1.2 Calcination, Reduction, and Regeneration ............................... 43 4.2 Liquid Phase Reactions ................................................................ 45 4.2.1 Loading and Operation of Reactor .......................................... 45 4.2.2 Kinetic Runs ........................................................................ 47 4.2.3 Reaction Product Removal .................................................... 47 4.3 Vapor Phase Reactions ................................................................ 48 4.3.1 Loading and Operation of Reactor .......................................... 48 4.3.2 Reaction Product Removal ................................................... 50 4.4 Reaction Product Analysis ........................................................... 51 4.4.1 Water-Soluble Products ....................................................... 51 4.4.2 Organic-Soluble Products ..................................................... 52 4.4.3 Reaction Product Yield Calculations ...................................... 53 vi Chapter 5: Results and Discussion of Liquid Phase Experiments ................ 54 5.1 Preliminary Experiments ................................................................ 54 5.2 Results of Liquid Phase Experiments .............................................. 57 5.2.1 Summary of Experiments ....................................................... 57 5.5.2 Identification and Yield Calculation of Water Soluble Products .............................................................................. 57 5.2.2.1 GC/MS Applications ....................................................... 59 5.2.2.2 Peak 255 ...................................................................... 59 5.2.2.3 Peak 329 ...................................................................... 63 5.2.2.4 Peak 338 ...................................................................... 65 5.2.2.5 Peak 347 ...................................................................... 65 5.2.2.6 Peak 358 ...................................................................... 67 5.2.2.7 Peak 369 ...................................................................... 67 5.2.2.8 Peak 416 ............................................................... . ...... 67 5.2.2.9 Peak 420 ............................................................... . ...... 68 5.2.2.10 HPLC Applications ...................................................... 68 5.2.3 Identification and Yield Calculation of Organic Soluble Products ............................................................................. 68 5.2.3.1 Peak 77 ....................................................................... 70 5.2.3.2 Peaks 86 & 98 .............................................................. 70 5.2.3.3 Peak 111 ...................................................................... 72 5.2.3.4 Peaks 118 & 139 ........................................................... 72 5.2.3.5 Peak 165 ...................................................................... 73 5.2.3.6 Peaks 205 & 237 ........................................................... 73 5.2.3.7 Peak 0 ........................................................................ 74 5.2.3.8 Peak T ......................................................................... 74 5.2.3.9 Peaks 277, 283, 299 ...................................................... 74 5.2.3.10 Peaks 307, 312. 323. & 332 .......................................... 75 5.2.3.11 Review of Organic Products .......................................... 75 5.2.4 Quantitative Analysis ............................................................. 75 vii 5.3 Discussion of Results .................................................................... 82 5.3.1 Effects of Reaction Conditions ............................................... 82 5.3.2 Observed Reaction Pathways .................................................. 85 Chapter 6: Results and Discussion of Vapor Phase Experiments .................. 89 6.1 Preliminary Experiments ................................................................ 89 6.1.1 Reaction Conditions. .............................................................. 90 6.1.2 Blank Runs ........................................................................... 91 6.2 Results of Vapor Phase Experiments ............................................... 93 6.2.1 Summary of Experiments ........................................................ 93 6.2.2 Summary of Recovered Reaction Products ............................... 93 6.2.3 Quantitative Analysis ............................................................. 99 6.3 Discussion of Results ................................................................... 101 Chapter 7: Conclusion and Recommendations .......................................... 104 7.1 Liquid Phase Experiments ............................................................. 104 7.2 Vapor Phase Experiments ...................................................... . ....... 106 7.3 Process Considerations ................................................................. 107 Appendix A ............................................................................................ 108 Appendix B .......................................................................................... 113 List of References .................................................................................. 144 viii @QNOJUI-hQNd .3 0 LIST OF TABLES Catalysts Prepared for Liquid and Vapor Phase Experiments ................ 42 Summary of Blank Runs. Liquid Phase Experiments ............................ 55 Summary of Liquid Phase Experiments ............................................... 58 Summary of Water Soluble Products .................................................. 69 Summary of Organic Soluble Products ............................................... 77 Summary of Quantitative Results for Decalin Solvent Reactions ............ 79 Summary of Quantitative Results for Water Solvent ............................ 80 Summary of Vapor Phase Experiments ............................................... 91 Summary of Catalyst Weights for Vapor Phase Experiments ................. 100 Summary of Quantitative Results for Vapor Phase Experiments ............ 102 «5013M DQNQU‘I 11 12 13 14 15 16 17 18 19 20 21 22 23 LIST OF FIGURES Proposed Reaction Sequence ............................................................ 9 Structure and Pore Size of Zeolites ..................................................... 12 General Fragmentation Scheme for TMS Carbohydrate Ethers .............. 14 General Fragmentation Scheme for Carbohydrate Degradation Products ....................................................................................... 16 Sketch of Calcination/Reduction Reactor ..................................... . ..... 19 Sketch of Caicination/Reduction Process ............................................ 20 Sketch of Batch Reaction System ...................................................... 21 Sketch of Kinetic Run Collection Tube on Batch Reactor ...................... 23 Reactor Tube Assembly .................................................................... 29 Reactor Seal Assembly ..................................................................... 31 Furnace .......................................................................................... 33 Temperature Measurement and Control in V.P.R. ................................. 34 Vapor Phase Reaction System .......................................................... 36 Product Trapping System ................................................................... 39 HPLC Tracings of Water Soluble Products ........................................... 60 GO Tracings of Water Soluble Products ............................................. 61 Total ion Count Chromatogram (K1, sample #5) .................................. 62 Structure of Reaction Products ......................................................... 64 Non-Cyclic Reaction Product Formation ............................................. 66 GO Tracing of Organic Products ........................................................ 71 Spirocyciic Lactone Structure and Fragmentation ................................. 76 Kinetic Run Reaction Products Summary ............................................ 81 Observed Reaction Pathways ........................................................... 86 24 25 26 27 8-1 8-2 8-3 8-5 8-6 8-7 8-9 8-10 8-11 8-12 8-13 8-14 8-15 B-16 B-17 8-18 8-19 8-20 8-20 8-21 8-23 8-24 8-25 60 Tracing of Blank2 and Y62,500 Experiments ................................ 95 GC Tracing of MC-3.500 Experiment ................................................. 96 GO Tracing of Gas Products ............................................................. 97 GC Tracing of MC-3.800 Experiments ................................................ 98 Mass Spectrum of Peak 255 ............................................................ 113 Mass Spectrum. of Peak 329 ............................................................ 114 Mass Spectrum of Peak 338 ............................................................ 115 Mass Spectrum of Peak 347 ............................................................ 116 Mass Spectrum of 2,5-Anhydro-D-Mannltol ..................................... 117 Mass Spectrum of Peak 358 ............................................................ 118 Mass Spectrum of Peak 369 ............................................................ 119 Mass Spectrum of Peak 416 ............................................................ 120 Mass Spectrum of Peak 420 ............................................................ 121 Mass Spectrum of D-Giucitol ....................................................... 122 Mass Spectrum of Peak 77 ............................................................ 123 Mass Spectrum of B-Angeiica Lactone ........................................... 124 Mass Spectrum of Peak 86 ............................................................ 125 Mass Spectrum of Peak 98 ............................................................ 126 Mass Spectrum of Peak 111 .......................................................... 127 Mass Spectrum of Peak 118 .......................................................... 128 Mass Spectrum of Peak 139 .......................................................... 129 Mass Spectrum of 2-Furyl Methyl Ketone and 5-Methyi Furfurai ....... 130 Mass Spectrum of Peak 165 .......................................................... 131 Mass Spectrum of Peak 205 .......................................................... 132 Mass Spectrum of Peak 237 ......................................... 133 Mass Spectrum of 4-i-iydroxy-6-Methyl-Pyran-z-one and lsomaltoi. 134 Mass Spectrum of Decalin ............................................................. 135 Mass Spectrum of Tetralin ............................................................ 136 Mass Spectrum of Peak 277 .......................................................... 137 xi B-26 B-27 B-28 8-29 8-30 8-31 Mass Spectrum of Peak 283 ..................................................... 138 Mass Spectrum of Peak 299 ..................................................... 139 Mass Spectrum 01 Peak 307 .................................................... 140 Mass Spectrum of Peak 312 ..................................................... 141 Mass Spectrum of Peak 323 ..................................................... 142 Mass Spectrum of Peak 332 .................................................... 143 xii CHAPTER 1 INTRODUCTION AND BACKGROUND 1.1 Introduction Biomass is a potentially inexhaustible alternate raw material for the production of traditionally petroleum-based chemicals and fuels. At current prices and avaiiibility, however, products from biomass conversion technolo- gies such as fermentation, pyrolysis, hydrogenative liquefaction. and gasification/synthesis gas conversion are unattractive compared to those from crude oil processing. Biomass conversion processes require a large capital expenditure and high energy input which makes them uncompetitive with the established petroleum industry. The shortcomings of existing biomass conversion systems dictate that continuing research should focus on developing a system which is more cost competitive. increasing the energy efficiency of the conversion process is possible by replicating the reactions occuring in the natural conversion of 2 biomass by microorganisms. The use of synthetic, inorganic catalysts may fa- cilitate such replication, as inorganic catalysts are applicable in many reac- tion systems and in general are easier to prepare and handle than enzymes or microorganisms. From a process standpoint, therefore, inorganic cata- lysts are preferable to organic catalysts for biomass conversion; the key to a successful biomass conversion process lies in selective replication of the energy-efficient reaction pathways. Substantial research has already been done on the use of inorganic catalysts for conversion of biomass. Of the catalysts studied, the most ef- fective and widely used are the zeolite catalysts. Zeolites are crystalline, mi- croporous aluminosiiicates which function as solid acids. The pores are of molecular dimensions and thus may provide shape selective properties. The chemical structure is thermally stable and can be easily altered by cation ex- change to exhibit a desired catalytic activity. Results from previous studies in biomass conversion using zeolites are very encouraging. The biomass conversion reactions observed over these catalysts include dehydration, dehydrogenation, hydrogenation, and decarboxylation. With this controllable diverse reaction capability, zeolites have potential to provide an energy-efficient solution to the conversion problem. Continued research in this area may lead to a breakthrough in biomass processing. This research focuses on.developing a novel process for the conver- sion of D-glucitol, a biomass derived product, using zeolite catalysts. A reaction pathway has been proposed which utilizes the zeolite-catalyzed dehydration and hydrogenation reactions. identification of reaction prod- ucts, investigation of conditional effects, and determination of the most de- sirable zeolite properties for the reaction are the primary objectives. The remainder of this chapter gives some background in the conver- sion of biomass and biomass-derived products. it also provides an outline of the proposed project and reviews specific research objectives. 3 1.2 Background 1,3,1 5293! anvgrgign Via. Hgmgggngggg Qgtglxglg The combination of acid-catalyzed reaction and thermal degradation found in the food processing industry has been the focus of a considerable amount of carbohydrate research. Many of these natural reactions are classi- fied as Maillard reactions and are poorly understood. The resultant research into this field has produced insight into sugar reaction and degradation products which may have applications in this study. The acid-catalyzed dehydration of hexitols (monomer sugar alcohols) to form primarily tetrahydrofuran products has been accomplished using min- eral acids [1, 2]. Barker [3] gives initial results of the dehydration of D- giucitol using 2N hydrochloric acid with the discovery of the formation of anhydrohexitois by derivatization and GC analysis. it was found that the first step in D-giucitoi dehydration produced 1,4-anhydro-D-glucitol, 3,6-an- hydro-D-glucitol, and small amounts of 2,5-anhydro-L-iditoi. Barker goes on to explain formation mechanisms and discusses the effect of configura- tion of various hexitols on anhydride products. Beck et al. [4] continued this research with a thorough study of alditol dehydration products using NMR. They heated solutions of D-giucitoi in 3M sulfuric acid for up to 172 hours in order to determine the dehydration pathway. initial dehydration products were identified as 1,4-anhydro- and 3,6-anhydro-D-glucitoi, 2,5-an- hydro-L-iditol, and 2,5-anhydro-D-mannitol. With further reaction, they ob- served the formation of 1,4:3,6-dianhydro-D-glucitol from the 1,4 and 3,6- anhydro-D-giucitol compounds. Kinetic studies yielded rate constants for the formation and subsequent reaction of each product. They also reported a slow decomposition of the dianhydride compound with prolonged heating in the acid solution. More recent studies on the formation of isosorbide (1,4:3,6-dianhydro-D-glucitoi) from D-giucitol using various organic acids 4 and reaction conditions have been conducted by Fleche and Huchette [5]. Results in the literature indicate the formation of anhydrohexitol compunds as the initial products in D-glucitol dehydration. initial research into sugar degradation revealed a singular result for al- most aii types of sugars. Sugar solutions were degraded in high temperature acid to produce primarily 5-(hydroxymethyi)-2-furfural (HMF). This com- pound was found to be the major product in the reaction of all hexoses [6] including glucose [7], fructose [8], and sucrose [9]. Gardner and Jones [10] have patented a process for the commercial production of HMF from hydro- lyzed cellulose. Shaw et al. [11, 12] identified many other compounds from the acid- cataiyzed degradation of various hexoses. They found significant amounts of furfural compounds such as 5-methyi-2-furfural, 2-(2-hydroxyacetai)-2- furfural, and the HMF degradation products levulinic acid and formic acid. in a more recent study, Popoff et al. [13] identified a myriad of cyclic ke- tones, substituted benzenes, and very complex iactone compounds. Feather and Harris [14] give an excellent review of the reaction pathways leading to the above compounds and discuss the effects of reaction conditions and reactant configuration on the resultant products. Results of these studies suggest that further reaction of the initial an- hydro-sugar compounds may produce products similar to those found in sug- ar degradation. Various furfurais, pyrones, and lactones all are likely prod- ucts to be found in this research. 2. A ii in of Zeoll l i The dehydration of alcohols over zeolites is well documented and has been previously reviewed [15, 16, 17, 18, 19]. The aluminosiiicate pore structure, shape selectivity, and thermal stability properties make zeolites excellent dehydration catalysts. The best known zeolite-catalyzed reaction is the methanol to gasoline process (MTG) developed by Mobil [20]. This process has resulted in several commercial installations where the zeolite 5 ZSM-5 is used to effectively dehydrate methanol and ethanol in route to gas- oline. Kudiac [15] demonstrated that the Ca-Y zeolite effectively dehydrat- ed hexanol and hexanedioi in a batch process. This information, combined with the capability of zeolites to accomodate large molecules, makes the zeo- lite catalyst an excellent candidate for sugar alcohol dehydration. The dispersion of hydrogenating agents such as Group VIII metals in zeolites to form dual functional catalysts is also well documented in the literature [21, 22]. Metal-loaded zeolites are employed in petrochemical pro- cesses for hydrocracking, seiectoforming. and hydroisomerization [23, 24]. Hydrocracking uses the acidic sites of the zeolite to crack and isomerize the reactant while the metallic sites hydrogenate the products. There are sever- al interesting literature reviews on the reaction sequences and pathways of hydrocracking using bifunctional zeolite catalysts [25, 26]. 1,2,; Blgmggs Convmlon Using Zeglitgs Catalytic conversion of biomass using zeolites has been accomplished by both direct and indirect methods. Direct conversion involves contacting high molecular weight biomass material with the catalyst with minimum prior processing. indirect conversion is the result of reacting a biomass-derived product with the catalyst. Typical processes involved in product extraction include fermentation, hydrolysis, and pyrolysis. Direct conversion of cellulose and starch to hydrocarbon products was accomplished by Chen and Koenig [27] using the synthetic Mobil zeolite ZSM-5. Results showed relatively low yields of usable hydrocarbon prod- ucts and significant catalyst coking problems. Weisz et al. [28] also used ZSM-S to convert natural oils to hydrocarbons. The conversion of rubber la- tex, corn oil, and peanut oil produced an array of olefins and aromatic com- pounds. The major reaction observed was dehydration which leads to the ar- omatic structures. More importantly, this research demonstrated the capability of zeolites to accomodate and effectively convert very large mole- cuies. 6 The extension of the Mobil MTG process to ethanol [16, 29] has re- sulted in a major biomass application [30]. Ethanol and similar components obtained from biomass fermentation can be directly converted to gasoline fraction compounds in high yields. The reaction can also be run in somewhat dilute conditions to reduce the energy-intensive distillation step required to recover ethanol from the fermentation broth. The ZSM-5 zeolite used in this system effectively dehydrates and then shape-selectively transforms the alcohols into hydrocarbons. This is an example where the integration of an established process like fermentation with a catalytic process reduces the en- ergy required to effectively process the biomass. A considerable amount of research has gone into integrating catalysis with pyrolysis. Pyrolysis involves heating biomass, usually wood, in the ab- sence of oxygen to produce a pyroiytic oil. The oil, which contains a large variety of oxygenated compounds, is then reacted with a zeolite catalyst to produce usable hydrocarbons. Chantal et al. [31] and Frankiewicz [32] used ZSM-S and Y zeolites, respectively, to demonstrate that the pyroiytic oil could be effectively upgraded; however, char formation was a major prob- lem. Model oil components such as cyclic ketones [33], phenols [34], fur- furals [35], and heavy alcohols [32. 33] were all reacted over various zeolites to study reaction products and pathways. Diebold and Scahiil [36] bypassed the charring problem by reacting the vapors produced by the pyrolysis of wood directly with ZSM-5 catalysts. it was found that the vapors were much more reactive, and much higher conversions of the oxygenated pyroly- sis products resulted. These studies showed that zeolites could effectively deoxygenate a number of compounds which were oxygenated to various de- grees. The major reaction observed was dehydration with decarboxylation al- so occurring. There are many examples of zeolite-catalyzed reactions of biomass hy- drolysis products. The hydrolysis process involves the reaction of biomass polymers such as cellulose with acid solutions to produce monomers. Sugars 7 such as D-glucose. D-fructose, and sucrose are the major products in biomass hydrolysis. These resultant sugars have been converted directly to hydrocarbon products using zeolites with results similar to those seen previ- ously in the cellulose reactions [27, 37]. The major reactions observed are dehydration to anhydro-sugar compounds. further degradation to a variety of furfurais, and condensation to polymeric compounds. A major drawback of sugar conversion with zeolites is the rapid formation of nonvolatile tars which render the catalyst ineffective in a very short time. As mentioned earlier, most hydrocracking reactions are carried out on petroleum byproducts, but several papers have been published on com- pounds which can be obtained from biomass. Arena [38] has patented a pro- cess to hydrocrack polyois such as D-glucitol. The patent describes the use of a Group VIII metal deposited on an alkali earth metal-exchanged solid sup- port. This catalyst hydrocracked aqueous solutions of D-glucitol to yield eth- ylene glycol, glycerol, and 1,2-propanediol at conversions of up to 58%. Weitkamp, Jacobs, and Ernst [39] effectively hydrocracked substituted cyclo- hexanes using Pt-exchanged ZSM-5. The compounds studied are similar to expected products from D-giucitoi dehydration and thus suggest that zeo- lites will react with rather large cyclic compounds. As outlined, the use of zeolite catalysis has been integrated into all ar- eas of biomass conversion. Reactions have been catalyzed using unproc- essed and processed biomass materials which include many types of com- pounds of various molecular weights. The goal of this research is to inte- grate some of these zeolite-catalyzed reactions into a novel biomass processing pathway to facilitate a high-yield route to hydrocarbon com- pounds. 8 1 .3 Proposed Research Pr R l n P hill! The proposed reaction sequence for the conversion of biomass to hy- drocarbons using zeolite catalysts is similar to that presented in earlier re- search. [15] The conservation of the sugar carbon chain by oxygen removal through dehydration and hydrogenation may lead to higher carbon conver- sions. Also, the use of alditols may reduce the charring reactions seen in the previous sugar/zeolite reactions. The reaction sequence integrates exist- ing processes with proven zeolite-catalyzed reactions already shown to be feasible for similar compounds: (1) Sugar Formation Biomass (cellulose) --------- hydrolysis -------- >Monomer Sugars (2) Alditol Formation Sugars (D-glucose) ----- hydrogenation ----- > Alditols (D-glucitoi) (3) Dehydration Alditol ------------ acid catalyst --------- > Polyhydroxyiated Olefin (4) Hydrogenation Polyhydroxyiated Olefin --------- hydrogenation---->Lower Poiyoi The compounds represented by poiyhydroxylated olefin and lower poly- ol can be understood by referral to the structural scheme in Figure 1. Repetition of reactions (3) and (4), along with possible product cracking. to- gether can be summarized by the following: (5) Alditols dehydration, hydrogenation--->Oxygenated compounds cracking and Hydrocarbons The hydrolysis of biomass to recover sugars, Reaction (1), is a proven process often used to prepare feedstock for fermentation. Requirements 9 PROPOSED REACTION SEQUENCE hydrolysis Biomass CH CH 2 O OH H OH D-Glucose D-Glucitol Polyhydroxyiated Olefin Lower Poiyoi > Sugars ( D-Glucose) HOH HOH hydrogenation HO CH ‘ 7 metal catalyst H OH H OH HEOH D-Glucitoi HCH ll dehydration OH HO H H OH H OH HEOH Polyhéd roxylated leiin acid catalyst hydrogenation 7 Lower Poiyoi metal catalyst dehydration Oxygenated Compounds and Hydrocarbons V hydrogenation cracking Figure 1. Proposed Reaction Sequence 10 for effective hydrolysis depend upon the material being processed. Common sugar extraction is done using strong acid solutions at moderate tempera- tures and pressures [40] to give high monomer yields. For the reaction path- way proposed, the recovered sugar is D-giucose. The hydrogenation (reduction) of sugars to form sugar alcohols or polyois (Reaction (2)) is also well understood. Arena has patented several processes using Group Viii metal-impregnated pyropolymers [41], ruthenium dispersed on a-alumina [42], and nickel [43] and ruthenium [44] deposited on titanium dioxide to effectively hydrogenate D-glucose to D-giucitoi. Aque- ous solutions of the sugars are reacted in both batch and fluidized bed reac- tors to yield up to 98% conversion of D-giucose with 95% selectivity to D- giucitoi. Reaction (5), as a composite of Reactions (3), (4) and cracking, repre- sents the focus of this research and involves the integration of the zeolite catalysts into the reaction sequence. The proven dehydration capabilities of zeolites will be combined with the proven hydrogenation properties of Group VIII metals in a dual-functional catalyst. This results in a reaction to deoxygenate the sugar alcohol to hydrocarbon products. The cracking reac- tion is added to accomodate isosorbide which is a major product of D-gluci- toi dehydration. The volatility of this compound and the applications of dual functional zeolites in hydrocracking may make vapor phase cracking to small- er hydrocarbons feasible. 1.2hi f isanCnlin All of the literature reviewed to this point lends support to the use of zeolite catalysts to accomplish the proposed reaction scheme. Zeolites pro- vide bifunctional capabilities, selectivity to minimize the number of products, and thermal stability to withstand reaction conditions. Many of the papers reviewed involved reactions of large, substituted compounds which are simi- lar in structure to the sugar alcohol D-glucitol. A strongly dehydrating 11 zeolite with hydrogenation activity could prove to be an effective biomass conversion catalyst. A Ca-exchanged Y zeolite and Mobii’s HZSM-5 zeolite were chosen as the acidic dehydration catalysts. Hydrogenation capability is provided by the addition of both Pt and Ni metals. The Y zeolite has the large pore faujasite structure while ZSM-S has a channeled medium pore structure of a synthetic zeolite (see Figure 2). The differing structure and pore size could lead to different products because of shape selectivity. Kerr [45] gives a good review of the synthesis, structure, and reactivity of these two catalytically important zeolites. Previous work on this project with Ca-Y zeolite [15] used decahydronaphthalene (Decalin) as the reaction solvent in a batch process. This solvent was chosen for its suitability in hydrogenation processes and al- so for its large size which may inhibit entrance into the zeolite pores. Deca- lin will be used in this continuing research but the focus will be on using wa- ter as the reaction solvent; deionized water is a more practical solvent for biomass conversion. The H-ZSMS catalyst is chosen primarily for its capability for dehydration in aqueous reaction solutions. Primary reactions for catalyst effectiveness studies and product identification will take place in a batch reactor. The batch system provides high conversions of reactant and thus high concentrations of products, mak- ing their identification easier. Volatile products obtained from D-glucitol de- hydration will be further reacted in a designed vapor-phase differential flow reactor. 1 . Anal i a M h d This section provides background for the analysis of expected reac- tion products using Gas Chromatography/Mass Spectrometry (GC/MS). Mass spectrometry is the primary tool used for unknown compound identification in all areas of research. When combined with gas chromatography, a series of unknowns can be separated and analyzed with 12 Y ZBOIHB . C . ”5.6A . 10 mm . HZSM-5 Zeolite Figure 2. Structure and Pore Size of Zeolites 13 ease. The spectrum of fragment masses provides a distinct fingerprint for each type of unknown compound. The molecular weight and the masses and amounts of the breakdown fragments are then used in the determination of the structure of the unknown. The ability to analyze carbohydrates using GC/MS has resulted from breakthroughs in derivatization of non-volatile compounds. Hydroxyl groups found on all carbohydrates can be reacted with a derivatizlng agent which replaces the hydrogen with a volatilizlng agent such as an acetate or trimethylsiiyl group. This type of reaction yields volatile products which can be separated by gas chromatography. The most effective derivatizlng technique for alditols and similar com— pounds is trimethylsiiyl derivatization [46]. it has distinct advantages over methyl ether and acetate derivatives in ease of preparation and simplicity of fragmentation pattern. Kudiac [15] gives a review of the structure-revealing fragmentation patterns encountered in TMS derivatives (see Figure 3) and discusses the problems of rearrangement and isomerization. The major ions obtained in mass spectra of TMS-derivatized compounds result only from the cleavage of 0-0 bonds in the carbohydrate chain or from the loss of TMS groups. These major ions can be traced backward to reveal the structure of the parent compound. This technique will be employed to determine the structure of all water soluble, non-volatile products like the anhydrohexitol compounds discussed earlier. The expected organic degradation products will also be analyzed using GC/MS. Since compounds of this type are volatile, no derivatization proce- dure is necessary. Previous MS studies of furfural-type compounds have re- vealed severai characteristic fragmentation patterns. Furfurals always exhib- it an abundant parent ion (m/z 110, m/z 126) and usually have a strong loss of the hydrogen associated with the aldehyde substituent (M-1 ion). Anoth- er characteristic is the consecutive loss of two CO molecules [47]. The MS characteristics of possible hydroxylated pyrones (126) include the immediate 14 «35w 29.22.0an 92... Lo. oEosom cozanEaaE .2050 .n 952". 955 .385»: 53:83 «286:. H 2:010. 955 .385»; >555 3:365 H mEPONIo- 965 m2... So: 32 H «:0. :2 E93 .0 mmmE H s. 82.8.:33.2..3 mum—.1: cart: omumwie‘ mom's— : = > > 2254.10 - 1055 - :92: - 225:0 - ow .. E m-. u s- nap . E > > > :92: . mzo . 2:090 - 15 loss of a CO molecule followed by the loss of the prominent substituent on the ring [48]. Finally, Iactone products (m/z 144, m/z 154) can often be rec- ognized by characteristic base peaks produced from the cleavage of the alkyl side chain from the ring. Abundant ions of mass 85 and 99 usually signal a iactone structure. Molecular weight determination of lactones is often diffi- cult due to the loss of one or possibly two molecules of water from the par- ent compound [49]. These guidelines are reviewed in Figure 4 and are used in the identification of all volatile reaction products. 1.4 Specific Research Objectives The objectives of this research center around determining the feasibili- ty of the proposed biomass conversion proposal. 1) The first objective is the identification of all reaction products obtained from the D-giucitol reac- tions using the Ca-Y zeolite with decalin as a reaction solvent. New analytical procedures have been developed to isolate the organic products from the decalin phase. 2) The second objective is to test the feasibility of reacting D-giucitol in the presence of HZSM-S in aqueous solutions. The deionized water reaction solvent makes analytical procedures easier and more accurate. 3) Finally, the major D-giucitol dehydration product, isosor- bide, will be reacted further in a newly designed and constructed vapor- phase reactor. These studies will investigate the reactivity of the initial dehydration products and will aid in determining the feasibility of the overall reaction pathway. 16 furfural; + _' ‘—'—’ - CO. CO I | ____. _. \ R 0 CH0 R 0 0-0 R + R-Hm/z96 R-Hm/295 R=Hmlz39 R - CH3 m/z 110 R - CH3 m/z 109 R . CH3 m/z 53 n O O C 0 \cszo m/z 111 0H+ + O=C=CH=CO l I m/z 69 ngggnes JV \ CH3 \ \ CH3 H30 0 0 H30 0 0 £\ :L [\ :L H30 0 O O O + + m/z 99 m/z 85 Figure 4. General Fragmentation Schemes for Carbohydrate Degradation Products CHAPTER 2 LIQUID PHASE APPARATUS This chapter describes the equipment used in the liquid phase experi- ments for the dehydration of D-glucitol. Sections 2.1 and 2.2 describe the catalyst handling and preparation equipment, Sections 2.3 and 2.4 describe the batch reaction apparatus, and Section 2.5 describes the equipment used in reaction product analysis. 2.1 Catalyst Preparation Standard laboratory equipment was used in catalyst preparation. The ion-exchanging and metal loading operations required a magnetic stir- rer/heater and several Erlenmeyer flasks with condensers for reflux. Recov« ery of the catalyst from the exchange solution utilized a vacuum filtration flask, Buchner funnel and #42 filter paper. The recovered catalyst was dried in a Cenco Constant Temperature Laboratory Oven before caicina- tion/reduction. 17 18 2.2 Catalyst Caicination/Reduction and Regeneration The catalyst caicination/reduction and regeneration procedures were carried out in a reactor designed by Treptau and Miller. [50] A representa- tion of the reactor is given in Figure 5. it consists of an internally-insulated stainless steel pressure vessel with a removable cover. The gas inlet and outlet as well as the power and thermocouple leads are accomodated by two Conax fittings located in the cover. Catalyst is placed in the alumina reac- tor tube which is heated by a coil heater. Temperature is controlled by an Omega Model 2011 Programmable Temperature Controller via two thermocou- ple probes in the catalyst. Figure 6 gives a schematic of the overall calcination/reduction pro- cess. Gas flow to the reactor is controlled by a flow control valve down- stream from the regulated reactant gas cylinders. Gas flows into the reac- tor and down the annular space between the outside of the reactor tube and the insulation. It then enters the alumina tube, flows upward through the packed catalyst. and exits the reactor. The exit gas is routed to either a bubble flowmeter or a fume hood by way of a switching valve. All 1/8 " tub- ing and valves are stainless steel and all 1/16" tubing is copper. Several precautions were taken to insure safe operation of the caicina- tion/reduction reactor. The reactor and all gas lines were thoroughly purged with nitrogen whenever switching reactant gases was necessary. The reac- tor was also situated in a firebrick enclosure to prevent a fire hazard while operating at high temperatures. 2.3 Batch Reaction Apparatus All liquid phase D-glucitol dehydration reactions were carried out in the batch reactor system represented in Figure 7. The reactor is a stainless steel Parr 4561 High Presssure Mini Reactor with a 300 ml capacity. it is rated for operation up to 3000 psig at 25°C ........................................................................................................................................... 2am»... amt? ... .....u.h~.w...um... ..........n.U.n....wu.w.u.n.u.u....u......... ..m. ”22.?.3.x.u,3+".wwwxnfix.“ “...”.nmmaflm....u.a.wwieeska...”.u.u.”,...n.n.w,u.u.u. ...... ........... ................... ... ... .....2 ....x.....z.. ......n.. ................................ 12...}...3 sw.m.V......... .....w...m.u.u ........................................................................................................................... e and Power Leads Gas Outlet and Thermocou ie Probe 20 Vent to Hood (i Power and Thermocouple Leads Flow Control Valve Switch Valve i iii E Catalyst Powder Reactor Clllllllllll Bubble Flowmeter Air N2 H2 Figure 6. Sketch of Caicination/Reductlon Process 21 Magnetic Stirrer Drive Cooling Water Lines Thermocouple Probe .1 4 5 l ] 7 3 Vent g 1 $ 2 n O_O OI: 6 Heater H N 2 2 Gas Collection Vacuum System System Parr Bomb Reactor Figure 7. Sketch of Batch Reaction System 22 and has a maximum temperature of 350°C. it features a removable glass lin— er and a water-cooled magnetic agitator drive. The reactor agitator speed. heating rate, and temperature are controlled by an external control module . The reactor is a pressure vessel with a removable cover which con- tains all the necessary fittings. Fittings located in the cover include gas in- let and outlet valves, a 2000 psig rupture disc. pressure gauge, thermocouple probe, and a dip tube for liquid sampling. The magnetic stirrer drive is also located in the cover and extends upward to connect to the drive motor when situated in the reactor stand. The pressure vessel is heat- ed by a quartz mantle heater; temperature is controlled by the internal thermocouple probe. The reactor is charged from regulated gas cylinders by way of a switching valve and a flexible gas service line. Post-reaction gas is routed to a vent or to the gas sample collection system which consists of a small cylinder with a pressure gauge and a septum sampling fitting attached. The gas outlet is also connected to a Matheson Model 1400 vacuum pump for rapid evacuation of the reactor system. All 1/8 in. 0.0. tubing, valves. and Swagelok fittings in the system are stainless steel. The only exceptions are the high pressure nylon gas service line and the vacuum pump connection which consists of 1/4" OD. copper tubing fitted to the heavy rubber vacu- um tubing. The reactor system is situated in a fume hood with Plexiglas doors to provide protection in case of reactor failure. Other safety precautions in- clude purging and evacuation of the system when switching reactant gases as well as operation of the fume hood while a reaction is taking place. 2.4 Liquid Sampling Apparatus A liquid sampling system was designed for the batch reactor described above for collection of time-dependent reaction data. Figure 8 illustrates Thermocouple Probe Gas inlet U 1 Gas Outlet 4! § i if 5 : I .1 4f Dip Tube Collection_, Tube f Stirrer 9 a? CDC) .. v Vacuum System Test Tube Figure 8. Sketch of Kinetic Run Collection Tube on Batch Reactor 24 the collection tube and the connection to the batch reactor. The collection tube consists of a 12 in. long, 1/4 in. OD. copper tube with the appropriate fittings for connection to the existing dip tube valve on the reactor cover. The tube volume allows for collection of a 9 ml sample which corresponds to a standard test tube volume. The outlet of the collection tube is fitted with a valve for isolation of the sample and a fitting to allow connection to the existing vacuum system for evacuation between samples. 2.5 Reaction Product Analysis Equipment Pr Dr in n ar i n The reaction products present in the aqueous phase are predominately non-volatile and thus required derivatization for gas chromatography (GO) or gas chromatography/mass spectrometry (GC/MS) analysis. The trimethylsiiyl derivatization process requires very dry samples so a drying ap- paratus using nitrogen has been constructed. Nitrogen gas from a regulated cylinder is routed through a fitting which allows attachment of 1/4 in. Tygon tubing. The tubing leads to a glass tee which splits the flow into two streams for drying two samples at a time. Regular syringe-type needles are attached to the ends of the tubing and the tubing is clamped in place over the surface of a stirrer/heater. The needles extend into 25 ml Erlenmeyer flasks which hold the aqueous samples. The combination of very low heat and low nitrogen flow rates effectively dries the samples overnight. The organic products contained in the decalin solvent require separa- tion for analysis. This is accomplished by methanol extraction and the use of a 250 ml separation flask to separate the respective organic layers. The methanol is then removed by using a Wheaton Micro Rotary Evaporator. Successive extractions and evaporations produce a concentrated fluid containing the organic reaction products which is analyzed by GC. 25 r m r h The gas chromatograph used is a Varian Model 3700 with a Supelco SPB-1 0.53 mm l.D., 30 meter fused silica capillary column. A Flame ioniza- tion Detector (FID) on the GC is used to generate an output signal which is collected using a Hewlett-Packard 3390A Integrator. The carrier gas (helium) flow rate of 30 cc/min is split into 2.5 cc/min through the capillary column and 27.5 cc/min through the make-up gas line. Hydrogen and air flow rates into the system are 30 cc/min and 300 cc/min respectively. An in- jection size of 1 pi and a temperature program of 55-225°C at a rate of 5°C/min with an initial 3 minute hold is used for most GC analyses. All gas chromatography results were obtained using this system. thPrfrmn Lil hrm rh The HPLC used is a Waters Model 600 Multisoivent Delivery System Chromatograph with a Bio-Rad HPX-87H ion Exclusion Organic Acid Analy- sis Column. The solvent used in the system is 0.01 N sulfuric acid at a flow rate of 0.5 ml/min which provides a back pressure of 426 psig. An injection size of 10 pl was used along with a column temperature of 65°C. The detec- tor used is a Waters Model 410 Differential Refractometer. Data are collect- ed using the WlRC Data Aquisition Package by Waters on an iBM XT Person- al Computer. All HPLC results were obtained using this system. WW All GC/MS data were generated by a LKB Bromma 2091 Mass Spectrometer located in the Michigan State University Mass Spectrometry Facility. The spectrometer is attached to a Shimadzu Model GC-9A Gas Chromatograph equipped with a DB-1 20 meter capillary column. A typical temperature program of 55-225°C at a rate of 2°C/min with an initial 3 minute hold is used for the GC/MS analyses. CHAPTER 3 VAPOR PHASE REACTOR DESIGN AND APPARATUS This chapter describes the conceptual design and apparatus of the va- por phase reaction system used for isosorbide hydrocracking. Section 3.1 reviews the desired capabilities of the system and Section 3.2 gives the de- sign criteria for each component. The equipment used in reaction product analysis is the same as that used in the liquid phase D-giucitol dehydration experiments. 3.1 Conceptual Design of Vapor Phase Reactor it is desired to design an apparatus to carry out heterogeneous catalytic reactions with vapor phase reactants and products. A fixed bed, flow reactor design was chosen because of excellent results in similar reac- tion studies. Since the apparatus is to be used in exploratory research, the following criteria are included in the design to ensure satisfactory performance. 26 27 A) The system must be capable of carrying out reactions simultaneous- ly at 600°C and 1000 psig. B) The reactor must be constructed of a material which is economical and allows a wide range of reactant and product properties. C) The system must provide leak-free performance under extreme con- ditions. D) The reactor design must allow easy charge, caicination, and recov- ery of catalysts. E) The system design must include accurate and automatic tempera- ture control of catalyst bed. F) The flow system must be capable of producing accurate and stable gas flow rates. G) The reactor must be able to accomodate the introduction of liquid phase reactants and reactant gases simultaneously. H) The trapping system must allow complete recovery of reaction products for analysis. in the following sections, a detailed apparatus description will illus- trate how each of the above design criteria were achieved. 3.2 Vapor Phase Reaction System Design 3.2.1 Pressurs Vesssl in order to achieve design criteria A) and B) described above, an 18 in. long, 1.25 in. 0.0. x 0.5 in. ID. seamless tube of 304 stainless steel was chosen as the pressure vessel for the reactor. The resultant wall thickness of 0.375 in. provides ample temperature and pressure capabilities; calcula- tions of maximum allowable pressure versus temperature are given in Appen- dix A. The 304 stainless steel was chosen because of its machinabiiity, resistance to hydrogen embrittiement, and availablity in a wide variety of siz- es. The OD. and LO. of the pressure vessel were not specifically designed, 28 but rather were chosen to fit heating element dimensions and desired cata- lyst volumes within the limits of commercially available tube sizes. A schematic of the outer reactor tube is seen in Figure 9. At the reac- tor inlet 0.5 in. NPT threads were machined to accomodate a dual 1/16 in. probe Conax fitting used to introduce tubing for the liquid and gas reac- tants. The opposite end the outer tube was welded to a bolted flange assem- bly which accomodates the reactor seal (Section 3.2.2). A copper sleeve with and 0.0. corresponding to the heating element diameter was fit over the outer tube. This sleeve is included to minimize thermal gradients in the catalyst bed. .2.2 Fl n n al The primary reactor seal is designed to facilitate frequent catalyst re- moval and replacement. A bolted flange assembly was chosen because of its simplicity. economy, effectiveness in providing a leak-free seal, and availabili- ty. The flanges used were 0.5 in. thick, 2.75 in. diameter 304 stainless steel blanks obtained from Huntington Laboratories, inc. The blanks allowed custom boring and seal groove milling to fit this design. The side of the flange with vacuum seal grooving was welded to the reactor tube leaving a flat side for the seal system. Figure 9 gives shows how the flanges were welded to the reactor tubes and how they fit together to provide seal com- pression. Appendix A includes the flange calculations which confirm ade- quate pressure and temperature capabilities. . . The choice of a seal for this design was dictated by need for reusabili- ty and simultaneous high temperature and high pressure requirements. Eiastomeric O-rings provide adequate pressure capabilities but are very expensive and are not reusable in high temperature apparatus. Metal seals provide both high pressure and temperature capabilities but are not reus- able. A seal which provides both high performance and reuse is the resilient metal seal. Thermoweli Fitting S Flange creens \ l i l l [2:] "if: till—'1 Int—‘1 llll Threaded Sections L Inner Tube Disassembied Ream” 0""9‘ Conax Copper Heatsink Fitting Flange \ ............................ Outer Tube Reactor Tubes Assembled Figure 9. Reactor Tube Assembly 30 The seal chosen for this reactor is the Type H15015-03 resilient metal seal from Helicoflex. Figure 10 provides the dimensions and the compo- nents which make up the seal. The Helicoflex seal has a maximum operating temperature of 600°C and a maximum pressure capability of 11,000 psig. The groove dimensions and the bolt loads required for compression of the metal seal are given in Appendix A. Since the seal is resilient, reuse is possi- ble via greater compression with each successive reseaiing. Figure 10 shows the entire seal assembly. The respective flanges are brought together by sliding the inner reactor tube into the outer tube. The seal is compressed in the groove of the outer tube flange by tightening the six bolts between the flanges. Effective sealing was achieved at a moderate bolt torque. With each reuse. the seal requires greater compression which is limited by the groove dimensions. The dimensions are such that the seal can- not be overtightened. The Helicoflex seals have been reused up to five times at high pressures without leaking. MW To provide easy catalyst handling the inner reactor tube was designed in sections as illustrated in Figure 9. The main body of the tube consists of a seamless section of 0.5 in. 0.0., 0.409 in. l.D 316L stainless steel. The 316L alloy was chosen for its excellent welding properties in thin-walled tub- ing. The 0.5 in. 0.0. tubing was machined into threaded sections of varying lengths which allow a catalyst bed of any desired volume. Stainless steel screening is inserted between the sections and holds the catalyst in place when the sections are screwed together. The last 5 in. section of tubing is welded to the inside of the inner tube flange. This insures a flat surface for seal compression. A 2 in. section of 1.25 in. 0.0. tubing was welded to the outside of the flange to provide for reactor outlet and thermoweli fittings. The result is an all-welded inner reactor tube with a variable length catalyst chamber which slides into the outer tube and seals with the joining 31 Outer Lining: Aluminum inner Lining: 316L S. S. lnconel Spring Seal Dimensions Flange Welded to Inner Tube \ Seal A . \_ ///E A \\ inner Tube __ Outer Tube Flange Welded to uter Tube //// V : Groove Dimensions Figure 10. Reactor Seal Assembly 32 of the bolted flanges. The inner tube also accomodates the axial thermoweli which will be described in a following section. When combined with the reusable seal, the inner tube design provides fast access to the catalyst. The bottom of Figure 9 illustrates the reactor fully assembled and shows that the assembled inner tube extends just past the heated zone represented by the copper heatsink. W The heating system for the vapor phase reactor consists of a furnace constructed of 10 standard firebricks and two Melien 2 and 3/8 in. 0.0. x 6 in. long, Al203 nichrome wire semicircular heating elements. The heating ele- ments have a maximum temperature of 1100°C and are wired in series to make use of a 240 volt power input. As seen in Figure 11, the firebricks are shaped to accomodate the heating elements and wiring system in a horizon- tal position. The bricks were cut to fit the dimensions of the reactor to in- sure proper insulation at high reaction temperatures. Two cast iron reactor support stands complete the furnace apparatus. The outer reactor tube fits into the rectangular stands at flats ground into the tube wall 2 in. from ei- ther end. The stands are adjustable in height and provide a rigid, non-rotat- able base for reactor service. 3.2.5 Temperature Messureme_nt_and Control Accurate temperature control of the catalyst bed is accomplished by an axial thermoweli which runs the length of the inner reactor tube. The top of Figure 12 shows the thermoweli apparatus. The 1/8 in. diameter, 18 in. long 316 stainless steel well is fixed in the center of the inner tubing by a bored tube fitting which is threaded (1/8 in. NPT) into the end of the reac- tor outiet. This thermoweli accomodates an 18 in. long Omega chromel- aiumel thermocouple probe equipped with an integral handle. The thermocou- ple can be moved to monitor the temperature at any point in the reaction zone. The probe is connected to an Omega Model 4000 Programmable 8 in. 0.25 in. Rerod- Flrebrick POWBI‘ Leads— Closed Furnace 11.25 in One Half, Top View a) ...— Outer Tube IIII Gr0und Fiat Adjustable Reactor Stand Figure 1 1 . Furnace 1/8x12 in. 8.8. Thermoweli Thermocouple s\\\\\ To Controller Temperature Control for Catalyst Bed § 1/2 in. Heat Tape :\\ .;. ;.;.|| .;. ;.;.-.:,: Q Surface Thermocouple TSOHn??§.Ep° Temperature Control for Reactor Outlet Figure 12. Temperature Measurement and Control in V.P.R. 35 Temperature Controller which controls power output to the reactor furnace. The programmable controller not only gives control over temperature and heating rate, but also allows ramp and soak programming for automatic tem- perature control in calcination cycles. Because of the high boiling point of the reactants and products used in the experiments, a heating system was designed for the reactor outlet. The bottom of Figure 12 shows the use of a 1/2 in. wide, 24 in. long heating tape to prevent the voiatiiized products from condensing in the reactor out- let. A surface thermocouple was placed between the reactor tubing and the heating tape to provide input to another Omega temperature controller. The heat tape had a maximum temperature of 220°C and was set at 200°C for the majority of the experiments. The use of the heat tape is only dependent on the boiling point of the reactants and products being used in the sys- tem. The conduction of heat down the reactor tube from the furnace is quite extensive and is adequate for reactants or products which boil below 150°C. 3.2.6 Gas Flow System Figure 13 shows a schematic of the entire vapor phase reaction sys- tem including the gas flow apparatus. Reactor gas pressure is controlled by the cylinder regulators while flow rate is controlled by a Linde FM4575 Mass Flowmeter/Flow Controller operating with a Model 70 0-200 SCCM H2 mod- ule. The flow controller can be computer interfaced and has 4 channels for blending of gases. Gas flows from the regulated cylinders (hydrogen, nitrogen, com- pressed air), through a 3-way switching valve. a 1000 psig rupture disc, the flow control module, and into the reactor inlet. The reactor inlet consists of a Conax fitting with 2-1/16 in. sealed probe entrances. The gas tubing ex- tends into the inner reactor tube entrance 5.5 in. to insure direct flow through the catalyst bed. Reactor pressure is monitored by a 0-1000 psig 36 3.0503000 ] E236 00:003. owe—E 33> .nw 230E FLTTIITJL E o > L 3:03:00 000:.au 0. 5:03:00 003. .00: 0. Gauamv _Omm0> E0~m>m hOufl-SDOC Obaflacbl CO_~OO__OO mflo 53305) On w~cu~0m¢¢ D_SG_J u. ¢ .2 a: «z m. C O 0. or 3 .. N F 02... 2... 00¢ 00. 9.3.. 0E: S kw”... v .2031 o 0 0300.2 20:05.00 30.... i ...If. A i V 000:. 203000.505. 20?. 3:5 2030008205. 00250 ........ r l. ....... t l t 0005:“. ‘ 3:02:00 30E 2 37 Ashcroft gauge downstream from the reactor outlet and two product collec- tion traps. The gas pressure is stepped down from reactor to atmospheric pressure by a Linde Model 863810 Line Regulator. Exit gas is routed to a bubble flowmeter or directly to the vent. The system is also connected to a vacuum source for reactor evacuation when changing reactant gases. All tubing, Swagelok fittings, and valves in the gas flow system are stainless steel and rated for high pressure operation. The Fike rupture disc was installed upstream from the reactor to protect the system from cylinder regulator failure or operator error. The entire high pressure apparatus is contained in a fume hood with Plexiglas doors to insure safety from gas leaks or reactor failure. .27Re tan Pum in m The introduction of liquid reactants into the vapor phase reactor was achieved by direct injection of liquid into the heated reaction zone. Figure 13 outlines the apparatus involved in this process. Because direct injection requires a pumping system capable of pumping against the inlet gas pres- sure. an Eidex Model A-30-S High Pressure Metering Pump was chosen for liquid injection. it features a variable flow rate from 0.36 to 90 ml/hr and is capable of pressures up to 5000 psi. This pump provides very accurate flow rates even at very low settings. Liquid reactant enters the pumping system from 250 ml Erlenmeyer flasks through HPLC solvent filters. A switching valve allows choice be- tween two possible liquids. The liquid enters the pump through 1/8 in. Te- flon tubing and exits into 1/16 in. 0.0. stainless steel tubing. A coil of ap- proximately 15 ft. of 1/16 in. 0.0. tubing was used as a pulse dampener since the pump operates with a single piston action. The liquid reactant then travels through another valve before injection directly into the catalyst bed by way of the sealed probe entrance in the Conax fitting described earli- er. This system allows easy control of liquid composition and flow rate and 38 can be easily disconnected from the reactor for thorough flushing of the pump and tubing. P R v r m The product recovery system for the vapor phase reactor is designed to allow easy recovery of both liquid and gas reaction products. Figure 13 ii- lustrates the layout of the collection process. Reaction products travel down the heated reactor, through the heated reactor outlet, and downward into the first trap which is immersed in an ice bath. High-boiling point prod- ucts condense and are collected while lighter products continue into the liq- uid nitrogen trap. The condensed products from the ice trap are recovered by following a run by draining the trap after the system is brought down to atmospheric pressure. Valves are situated such that the liquid nitrogen trap can be isolated to contain the gaseous products when the trap is warmed to room temperature. The gaseous products can then be transferred to a down- stream gas sampling apparatus. ln each case, the traps are situated vertical- ly with the cylinders extending downward into the vessels containing the cooling fluid. Figure 14 shows the design of both the ice bath and liquid nitrogen traps. Both traps are made up of 1/8 in. stainless steel tubing, stainless steel Swagelok fittings, and Whitey 40 ml stainless steel gas collection cylin- ders with a pressure rating of 1400 psi. Vapor products travel down the dip tube, into the trap, and the apprOpriate species condense and collect in the bottom of the trap: Uncondensed products travel back up 26 the top. flow through the annular space surrounding the dip tube, and exit through the side port of the trap tee. This system provides simple leak-free collec- tion and can be easily disassembled for cleaning or replacement. This summarizes all the components of the designed vapor phase reac- tion system. A detailed operating procedure will be outlined in the following chapter. 39 Eo.a>m 9:002... «0300:. .3 0.52“. l! 0>_a> 000505 sum «2 2.6... A... Sam 5.25 8. $53 90 L F 1 2205.5 .2 oe||v r. . _ All 5:02.00 as Tl «33> .250 .2000... Egan—Om. CHAPTER 4 EXPERIMENTAL CONDITIONS AND PROCEDURES This chapter discusses the experimental procedures used in both the liquid and vapor phase experiments. Section 4.1 discusses catalyst prepara- tion and calcination/reduction procedures. Sections 4.2 and 4.3 discuss pro- cedures used in operating the liquid phase and vapor phase apparatus respec- tively. Finally, Section 4.4 outlines the procedures used in reaction product analysis. 4.1 Catalyst Preparation The catalysts obtained for this study were the Y zeolite of form Y-62 and the synthetic zeolite ZSM-S. The Y zeolite was obtained from the Union Carbide Corporation and the ZSM-5 zeolite from Mobil Corporation. The catalysts were delivered in both powder and 1/16 in. AIZOS-bonded extru- date forms for use in the liquid and vapor phase experiments respectively. 40 41 Both catalysts came in ammonium exchanged form which required the ion-ex- change and calcination procedures described in the following sections. Ta- ble 1 lists the catalysts prepared for use in these experiments 4.1.1 lgn-ExshangslMefal Lgadlng The ammonium form of the zeolites was modified to activate the de- sired catalytic properties. This was done by calcining the zeolite in air to give the active hydrogen form, or by ion-exchange with a desired ion like cai- clum. The Ca-Y zeolites used in this study were formed by contacting the ze- olite with a solution of calcium nitrate. Successive contacts with fresh solu- tlon insured complete ammonium/calcium ion-exchange. The loading of noble metals on the zeolites was also accomplished by contacting the catalyst with a solution of the metal salt. The Pt form of the Ca-Y zeolite and the Ni form of the ZSM-5 were both made using this proce- dure. Quantitative loadings were possible by using the amount of metal salt required to give desired loading of metal on the catalyst. Complete deposi- tion was achieved in only one contact. The following procedure was used for both ion-exchange and metal loading. Since several catalysts in various amounts were made this way, the weight of the salts used is given in a percentage and the amount of catalyst to be prepared is 10 grams: A 500% excess of the desired ion salt was dissolved into 500 mi of deionized water. Ten grams of catalyst was then placed in an Erlenmeyer flask and 100 mi of deionized water was added. The flask was then situated on a heater/stirrer, a water condenser was attached to the top, and the mix- ture was heated under reflux for 1 hour to degas the catalyst. Approximate- ly 100 ml of the salt solution was added over a period of 2 hours followed by refluxing for 2 more hours. The mixture was filtered, washed with "deion- ized water, and dried in a 120°C oven for approximately 2 hours. This en- tire procedure was repeated four times to insure complete ion-exchange. 42 Table 1 Catalysts Prepared for Liquid and Vapor Phase Experiments Caicinationl Caicinationl Catalyst % Metal Form Reduction Temp. Reduction Time Ca Y-62 0 % powder 500/ NA°C 5 hrs./ NA [22] Ca Y-62 5 % Pt powder 300/ 500°C 4 hrs./ 4hrs. H-ZSMS 0 °/. powder 550/ NA°C 5 hrs/ NA [27) H-ZSMSI 5 °/. Ni powder 550/ NA°C 5 hrs.l 4 hrs. H Y-62 0 % extrudate 500/ NA°C 5 hrs./ NA H-MC3 0 % extrudate 500/ NA°C 5 hrs./ NA H Y-62 0.5 % Pt extrudate 400/ 400°C 5 hrs./ 4 hrs. H MC-3 0.5 % Pt extrudate 400/ 400°C 5 hrs./ 4 hrs. 43 in loading the metal, the catalyst was placed in twice its weight of deionized water. The metal salt solution was prepared by dissolving the ap- propriate amount of salt in 350 times its weight of deionized water. Using the the same procedure described above, the metal salt solution was added to the catalyst mixture and refluxed for 2 hours. The catalyst was then fil- tered, washed. and dried to prepare for calcination/reduction. l in l R d l n R n r l n The final step in preparing the active catalyst was caicina- tion/reduction which improved stability and activity at elevated reaction tem- peratures. The calcination step was carried out at temperatures greater than 300°C with compressed air to drive off ammonia, reducing the NH4 ex- changed catalysts to an active hydrogen form. The calcination step also re- moved adsorbed water from the active sites of the zeolites. The reduction step was run at temperatures greater than 500°C with hydrogen to reduce the deposited metals to their zero-vaient state. The calcination/reduction procedures for catalyst used in the liquid phase experiments (zeolites in the powder form) were performed in the reac- tor apparatus designed by Treptau and Miller and described in Chapter 2. The calcination of catalysts for the vapor phase experiments (zeolites in the 1/16 in. extrudate form) was carried out in the reactor for vapor phase reac- tions. The liquid phase zeolite catalyst calcination procedure began with the loading of the powder catalyst into the ceramic reaction tube. The tube was loaded by placing a plug of glass wool in the bottom and filling to a depth of about three inches which corresponds to the height of the heating coil. The tube was then screwed into the cover of the reactor, which pushes the ther- mocouple probe down into the catalyst. The cover was then lowered onto the pressure vessel and the reactor was sealed. All gas tubing, thermocou- ple, and power leads were then connected to the appropriate outlets. The 44 controller was then turned on, set to the correct temperature program, and started. While the controller warmed up, a hydrogen flow of 30-40 cc/min was started by opening the flow control valve upstream of the reactor and monitoring the gas flow downstream with a bubble flowmeter. Once set, the system operated automatically for the entire calcination or reduction cycle. When the reactor cooled after a cycle, the catalyst was removed and stored in sealed bottles in a dessicator. Special care was taken when changing gas flows to insure adequate nitrogen purging to avoid an explosion. The regeneration of liquid phase catalysts was also carried out using the same reactor. Spent catalyst was filtered from the reaction solution, washed in deionized water, rinsed in acetone, and dried overnight in a 120° C oven. Once enough dried spent catalyst was obtained, it was loaded into the reactor and calcined as described earlier. The result was a regenerated catalyst with the same appearance and virtually the same reactivity. The calcination/reduction procedures for the vapor phase experiments were carried out in essentially the same manner in the differential flow reac- tor. The inner reaction tube was assembled fully except for for the last two sections. A stainless steel screen with a center hole to accommodate the thermoweli was slid down and seated on the threads of the final section as- sembly. The section to contain the catalyst was then screwed into place and catalyst was poured into the annular region between the thermoweli and the tube walls until the section was full. Another screen was placed on top of the full section and secured by the final inner tube section. The reactor was then sealed as described earlier, the thermocouple probe was inserted into the thermoweli, and the temperature controller was activated and set to the appropriate temperature program. The route of the gas flow into the reactor depends upon the type of procedure being run. The compressed air required for calcination is routed past the flow controller and enters the reactor through the liquid reactant 45 inlet. (see Figure 13) The flow was routed through the bottom of the first trap by closing the valve immediately downstream (valve 8 in Figure 13). This allows the connection of a bubble flowmeter directly to the fitting on the bottom of the first trap and eliminates the use of the rest of the gas flow system. Reduction with hydrogen is carried out with the existing hydrogen flow controller and the same first trap exit. The entire procedure was fully automated and required the turn of a valve and a change of temperature program to begin. The regeneration of the extrudate catalysts in-situ was accomplished simply by purging the system of the reactant gas, changing the route of the gas flow through the first trap, and changing the temperature program on the controller. The regenerated catalyst had the same clean white appear- ance and reactivity as the freshly prepared catalyst. 4.2 Liquid Phase Reactions All liquid phase reactions were carried out in the batch apparatus de- scribed in Chapter 2. The entire apparatus was contained within an operat- ing fume hood behind Plexiglas doors for every experiment. Figure 7 aids in illustrating the following procedure, 4 1 L adln n eration fRe r Two grams of reactant and one gram of catalyst were placed in the glass liner of the Parr reactor along with 125 ml of reaction solvent (Decalin or deionized water). The liner was placed in the pressure vessel, the reac- tor was sealed, and all gas, cooling water, power, and thermocouple connec- tions were made. The reactor vessel was then charged with 20 psig of nitro- gen purge gas by slowly opening valve 4 until reactor pressure reached the regulated pressure. A gas flow was then started by opening valve 5 and slowly opening valve 7 until bubbling could be heard within the reactor 46 vessel. The reactor was purged for approximately five minutes. All nitrogen in the system was then bled out and all valves were again closed. The reaction system was evacuated by turning on the vacuum pump and opening valve 6. By slowly opening valve 5, the reactor was gradually pumped down. When the pressure gauge reached full negative deflection ( usually within five minutes), valves 5 and 6 were closed and the vacuum pump was shut down. Reactant gas (H2 or N2) was introduced into the system in the same manner as the purge gas. The reactor was first charged to a low regulated pressure and a gas flow was started as before. After five minutes of flow, valves 7 and 5 were closed and the reactor was allowed to reach the desired reaction pressure as set by the regulator pressure. Once stabilized, all valves were closed for operation. After raising the heating mantle around the pressure vessel, the con- troller was turned on to start the heating cycle. The stirrer was also set to the desired second level speed (40% 0f maximum). Cooling water was start- ed to protect the magnetic drive from the high reaction temperatures. The heating rate was set on high until approximately 60 degrees below the de- sired reaction temperature at which time it was turned off. Once the reactor reached the desired temperature, the heating rate was set on low to main- tain the set point. By turning off the power, the chances of overshoot were minimized so an accurate reaction temperature could be attained. Reac- tion times were measured from the moment the reactor reached the desired temperature. At the completion of a run, the power to the heater was shut down and the heating mantle was lowered from the pressure vessel. The stirrer continued to operate until the reactor had cooled to room temperature. Once cooled, the reactor was depressurized by opening valves 5 and 7 and venting the reaction gas. 47 4.2.2 Klnstls flgns Studies in product formation versus time were performed using the ki- netic apparatus described in Figure 8 of Chapter 2 . Deionized water was used as the reaction solvent to facilitate collection and N2 was used as the gas for all kinetic runs to avoid possible explosion. The reaction was started in the same manner as previously described. The collection tube was fitted into place on valve 8 which is at the end of the reactor dip tube. With valve 8 closed, the collection tube was pumped down by opening valve 9 to the connected vacuum line. Valve 9 was then closed and the vacuum line was disconnected. After the desired reaction time had passed. valve 8 was opened to allow the pressurized reactor efflu- ent to fill the evacuated collection tube. Valve 8 was then closed, trapping the hot sample in the copper tubing. The outlet of the collection tube was then placed into the opening of a clamped test tube. Valve 9 was slowly opened to allow the collected fluid (9 ml) to empty into the test tube. The sample was then sealed with Parafilm and allowed to cool. Before collection of another sample, the collection tube was disconnected, rinsed with deion- ized water, and reassembled. 4.2. R a i n Pr (1 tRemoval After the reactor cooled to room temperature and the reaction gas was vented, the reaction products were removed. This was accomplished by carefully opening the reactor, emptying the glass liner, and rinsing all parts in contact with the reaction mixture with deionized water. The following procedure was used for accurate recovery of all reaction products. After a run was completed, the reactor was opened and the head was placed in a ring stand which situates the internal apparatus over a 1000 mi beaker. The glass liner was removed from the pressure vessel and its con- tents were emptied into the beaker. A wash bottle filled with 250 ml of deionized water was then used to rinse the glass liner, pressure vessel, and 43 all reactor apparatus of all catalyst and residue. All rinse water was collect- ed in the beaker. The beaker containing the reaction mixture and the rinse water was then placed on a stirrer/heater and stirred with low heat to maximize the re- moval of all products from the catalyst and insure thorough contact of all solvent phases. After stirring for 30 minutes, the mixture was filtered using a vacuum filtration flask. This separated out the catalyst which was washed and filtered repeatedly with the remaining wash water. The resultant filtered effluent was allowed to settle and separate into the respective phases if the reaction solvent was organic. Once settled, pi- pets were used to take samples of each phase for analysis. The recovered catalyst was washed in acetone, dried overnight in a 120°C oven, and stored until regeneration. 4.3 Vapor Phase Reactions All vapor phase reactions were carried out in the apparatus described in Chapter 3. The entire apparatus is contained in an operating fume hood with Plexiglas shields to insure safe operation. Figure 13 aids in understand- ing the following procedure. 4.3.1 Lgadlng and Operation of Re_actor The operation of the vapor phase reactor began with the loading of the catalyst, sealing of reactor, and the calcination cycle all described in earlier sections. The following procedure begins after the calcination cycle has been completed. After the catalyst has been properly calcined/reduced, the gas flow was shut down and the system was rerouted for a reaction run. This was done by disconnecting the bubble flowmeter from valve 10, closing valve 10, and opening valve 8 to allow flow through the downstream apparatus. 49 The line regulator was then set to zero and the reactor was charged with 50 psig of nitrogen by turning on the flow controller (set at 200 cc/min) and opening valves 1. 4. and 5 in that order. Once the downstream pressure gauge reached 50 psig. the line regulator was adjusted to 15 psig to estab- lish a flow through the entire system. The reactor was purged for 10 min- utes to remove all residual gases. After purging, valve 1 was closed and the pressure. in the reactor was allowed to reach zero. The reactor system was then evacuated by turning on the vacuum pump. opening valve 13. and switching valve 11 to the vacuum line. The system was pumped for about 10 minutes to insure proper evacua- tion before the valves were returned to their former state and the vacuum pump was shut down. The reactant gas flow (H2) was started in the same manner as the purge gas. Valves 2, 4, and 5 were opened and the line regulator was closed to allow the reactor to pressurize to the desired reaction pressure. Once the pressure gauge read the desired value. the line regulator was opened to 15 psig to start the gas flow. The flow controller was then set to the de- sired rate and allowed to stabilize. The actual gas flow rate was then checked with the bubble flowmeter to insure accuracy. Adjustment of con- trol valve 14 was often necessary to establish a stable flow readout on the flow controller. When the flow rate had stabilized, the furnace and heating tape power supplies were turned on and the controllers were set to the appropriate temperature programs. While the reactor heated up, the liquid reactant pumping system was readied for operation. The HPLC pump was checked for proper priming and was set to 4 ml/min for flushing. With the outlet line disconnected from the closed valve 6. valve 7 was switched to the deionized water flask. The pump was turned on and approximately 50 ml of water was pumped into a beaker to insure that the system was properly flushed and operating correctly. 50 After flushing, valve 7 was switched to the reactant flask and approximately 10 ml of reactant was pumped to prime the pump for a run. The pump was then turned off, set to the desired flow rate. and the outlet tubing was con- nected to valve 6. When the reactor and reactor outlet temperatures had stabilized. the traps were immersed in their respective cooling baths and the run was start- ed by turning on the HPLC pump and opening valve 6. The system operated automatically until the run was ended by closing valve 6 and turning off the pump. The outlet tubing from the pump was immediately disconnected and the system was flushed with deionized water as described earlier. Valve 2 was then closed and the reactor pressure was allowed to fall slowly to atmo- spheric pressure over a period of 1 hour. Once the pressure was released. the temperature controllers were shut down and the reactor was allowed to cool before the removal of the reaction products. 4.3.2 Reaction Product Removal Both liquid and gas phase products were collected in the vapor phase experiments. The procedures used in each case will be described in this sec- tion. After the reactor cooled to less than 50°C, the ice bath was removed from the trap for recovery of liquid reaction products. The exterior of the ice trap was dried with a towel and the outlet of valve 10 was thoroughly cleaned. A vial was then raised into place under the trap outlet using a screw stand and valve 10 was opened to empty the trap. Afterithe con- tents were collected, the vial was closed and refrigerated. The reactor was then disassembled and all apparatus in contact with the reaction mixture was rinsed with a quantitative amount of deionized water. This procedure collect- ed all starting material and reaction products which remained in the reactor or in the catalyst. The rinse water was sampled and analyzed along with the trapped products to complete a mass balance. 51 After the reactor reached atmospheric pressure. the liquid nitrogen trap was isolated by closing valves 8 and 9. The trap was then slowly warmed to room temperature by allowing the remaining liquid nitrogen to boil off. Once warm. the entire downstream gas sampling system was pumped down by activating the vacuum pump, setting the line regulator to zero. and opening valves 11, 12. and 13. When the pressure gauge regis- tered full zero deflection, the valves were closed in the same order. Valve 11 was then switched to the pressure gauge line and valve 9 was slowly opened to allow collected gas pressure to be measured. A gas sample was collected by switching valve 11 to the gas collection line. opening valve 12 to allow gas to fill the collection cylinder which is monitored by a small pres- sure gauge, and then closing valve 12 trapping the gas. The gas was then sampled for GO analysis with a gas-tight syringe through the septum port. 4.4 Reaction Product Analysis The reaction products from these experiments can be categorized into water-soluble and organic-soluble species. Procedures for each type of product are described separately. Since many different concentrations of products were encountered, the amounts of agents used is expressed as a percentage of product concentration for simplicity. 4.4.1 Water-Soluble Products The products contained in the aqueous phase were predominately non- volatile and required derivatization of GC and GC/MS analysis. The follow- ing trimethylsiiyl derivatization procedure proved effective in volatilizlng these products [51]. Approximately 2.5 ml of the aqueous phase was placed in a 25 ml Erlenmeyer flask which was situated in the sample drying apparatus de- scribed ln Chapter 2. A nitrogen flow rate was then established so that the 52 gas flow from the needle suspended above the surface of the fluid created a small depression. The heater/stirrer was turned on at its lowest setting and the sample was dried to a residue which was usually accomplished in 6-8 hours. When dry. the sample was removed from the apparatus and 1 ml of py- ridine was added to the flask. The pyridine dissolves the residue and acts as a solvent for the derivatizing agent. The mixture was then transferred to a small vial where a 500% excess of BSTFA (N.O-bls (Trimethylsilyl) trifluoroacetamide) was added. The excess was calculated on the number of theoretical hydroxyl groups present in the sample versus the number of trimethylsiiyl groups available for donation. The sample was then thoroughly mixed. warmed for 30 minutes in the 55°C 60 oven, and stored overnight in the refrigerator for complete reaction. Due to the hindered configuration of the reaction products. the derivatization reaction often took several days to reach completion. Once complete, the samples were ready for CC and (SC/MS analysis. 4.4.2 r anl -Soluble Pr duct The organic-soluble products contained in the decalin reaction solvent required extraction for CC and GC/MS analysis. The only other organic sol- vent found which was somewhat immiscible with decalin was methanol. The following procedure describes the extraction of the organic phase using methanol and the apparatus outlined in Chapter 2. The recovered‘organic phase was poured into a 250‘ml separation flask along with a like amount of methanol. The two phases were then thor« oughly mixed until the products giving the golden color of the decalin phase were transferred into the methanol phase. The decalin phase was drained in- to a storage vessel, and the methanol phase was emptied into a boiling flask. The boiling flask was connected to the rotary evaporator apparatus where the methanol was evaporated. The concentrated fluid was extracted 53 and evaporated repeatedly until all traces of the decalin phase were re- moved. The result was a mixture of organic phase reaction products which could be analyzed by 60 without the interference of the decalin reaction sol- vent 4 R l P Yl l l I h After the water soluble and organic soluble reaction products had been identified by GC/MS, quantitative procedures were developed in order to calculate product yields. Samples of the recovered water soluble products (Sections 4.2.3 and 4.3.2) were analyzed using HPLC. Mole percent yields were then calculated from the HPLC peak areas using response factors from standard runs. A mass balance was then performed by comparison of the total amount of product recovered to the amount of starting material. The amount of start- ing material not recovered in the aqueous phase was assumed to be present as organic soluble products. Justification of this assumption will be dis- cussed in the next chapter. The relative yields of the recovered organic soluble products were calculated by subtracting the total molar percentage of water soluble prod- ucts from 100. and multiplying by the GC area percentage of each organic soluble component. This results in a yield percentage for each organic prod- uct which totals the amount of starting material not recovered in the aque- ous phase. Attempts at resolving the actual concentrations of the organic soluble products were unsuccessful due to the difficulty in separating the products from the decalin reaction solvent. CHAPTER 5 RESULTS AND DISCUSSION OF LIQUID PHASE EXPERIMENTS This chapter presents the results of the D-glucitol and isosorbide reac- tions performed in the liquid phase. Section 5.1 reviews preliminary experi- ments, Section 5.2 outlines the experiments and the analytical procedures used in reaction product identification, and Section 5.3 discusses the overall liquid phase results. 5.1 Preliminary Experiments A series of preliminary experiments were run to investigate the performance of the batch reaction system described in Chapter 2. These blank runs were conducted not only to study the response of the system. but also to test the viability of the analytical procedures to be used in fur- ther studies. A summary of conditions for the blank runs can be found in Table 2. The first blank run was performed to check compatibility of the deca- lin solvent (125 ml) with D-glucitol (2 grams) under typical reaction 54 Run One Two Three Four 55 Table 2 Summary of Blank Runs, Liquid Phase Experiments Materials Decafin D-GIUCIIOI Water D-Glucitol Decafin Ca-Y catalyst DecaHn H-ZSM-5 catalyst M2.” I I n 3.1.9.19. 260°C 0% 160 psig H2 1 hr. 260°C 0.9% 60 psig N2 1 hr. 260°C no 160 psig H2 effect 1 hr. 260°C no 160 psig H2 effect 1 hr. 56 conditions. This run also tested the quantitative accuracy of the product re- covery procedures outlined in Chapter 4. Results showed a complete recov- ery of the starting material. Spectra from CC and HPLC showed no reac- tion products formed from the thermal degradation of D-glucitol in the deca- lin solvent. The second blank run was performed in order to observe the pressure response of the reactor when loaded with water(125 ml). The reactor was also loaded with D-glucitol (2 grams) to again test the analytical product re- covery procedures. The pressure of the system was monitored as the reac- tor was slowly heated to the typical reaction temperature of 260°C. A linear pressure increase was observed from the charge pressure of 60 psig to a maximum of 780 psig at 260°C. Quantitative results showed a 0.9% loss of starting material although no reaction products were found in the GC and HPLC analysis. This small loss probably is a result of the use of water as both the reaction solvent and the rinsing fluid. Not all water can be recov- ered leading to a small loss of the water-soluble reactant. Runs Three and Four were performed to determine whether either cata- lyst would have an effect on decalin, leading to unwanted organic products. The reactor was loaded with 1 gram of catalyst and 125 ml of decalin, and heated to 260°C for 1 hr. The GC analysis of the reaction solvent showed no apparent breakdown of the decalin with either the Ca-Y or l-l-ZSM-5 catalyst. The results of these preliminary experiments show that no products are formed from the starting material in the absence of a catalyst at typical reaction temperatures. The quantitative procedures for product recovery from both the decalin and water reaction solvents were also proven accu- rate. 57 5.2 Results of Liquid Phase Experiments .2. mm r fEx rlm nt Table 3 lists all experiments run in the liquid phase reactor apparatus. Catalyst type, reaction conditions, reaction time, and the conversion of starting material are provided for each experiment. Previous work [15] showed encouraging results from the Ca-Y cata- lyst in decalin so further experiments were run under similar conditions to complete the identification of all reaction products. Reactions were run at both 240° C and 260° C at various reaction times to yield information on the reactivity of the starting material as well as the product formation path- ways. Experiments were also run to catalytically convert D-glucitol in aque- ous solutions using HZSM-S zeolite. Experiments were first run in decalin for comparison to the Ca-Y catalyst in reactivity and product formation. Subsequent experiments were then run in water solutions. The use of water as a reaction solvent allowed the collection of samples over time using the ki- netic apparatus described in Chapter 2. This resulted in a much faster analy- sis of the reactivity of the HZSM-S catalyst. Several experiments were also run using 1.4:3,6-dianhydro-D-glucitol (isosorbide), a major intermediate reaction product of D-glucitol dehydra- tion. The objective of these experiments was to better understand the over- all reaction pathway by examining both the reactivity of and products formed from the intermediate. 2 l nlflalnnYll lulatlanaer lblePrd The water soluble reaction products from the liquid-phase experiments were analyzed using high performance liquid chromatography (HPLC) to de- termine product yields and gas chromatography/mass spectrometry (GC/MS) of trimethylsiiyl derivatives to identify individual reaction products. 58 Table 3 Summary of Liquid Phase Experiments Reactant: D-Glucitoi Solvent: Decalin .5191. 422mm CAYO.5,240 Ca Y-62,0% Pt CAY1.0.240 Ca Y-62.0°/o PI CAY2.0,240 Ca Y-62.0°/o Pt CAYO.5 Ca Y-62.0% Pt CAY1.0 Ca Y-62,0°/o Pt CAY2.0 Ca Y-62.0°/o Pt CAY4.0 Ca Y-62,0°/o Pt CAYP11.0 Ca Y-62,5°/o Pt CAYPt2.0 Ca Y-62,5% Pt 22.0 HZSM-5,0°/o Ni ZNi2.0 HZSM-5,5°/o Ni Reactant: D-Glucltol Solvent: Deionized Water .mE 9.13m ZW2.0.180 HZSM-5,0% Ni ZW2.0 HZSM-5,0% Ni ZWNi2.0 HZSM-5,5°/o Ni K1 HZSM-5,0°/o Ni Reactant: Isosorblde Solvent: Decalin £19.12 M DEG-1 Ca Y-62,0% Pt DEGZ Ca Y-62.0% Pt bees Ca Y-62,0% Pt DEG4 HZSM-5,0% Ni Reactant: lsosorblde Solvent: Deionized Water gxpt. Qatalyat K2 HZSM-5,0% Ni Mn. 240° 0 240° C 240° C 260° C 260° C 260° C 260° C 260° C 260° C 180° C 180° C Tamp. 180° C 260° C 260° C 260° C M2; 260° C 260° C 260° C 180° 0 260° C Praaagra Lima anvaratan 160 psig H2 0.5 hr. 92% 160 psig Hz 1.0 hr. 94.6% 160 psig H2 2.0 hr. 96.5% 160 psig Hz 0.5 hr. 96.4% 160 psig H2 1.0 hr. 96.9% 160 psig H2 2.0 hr. 97.4% 160 psig H2 4.0 hr. 98.2% 160 psig H2 1.0 hr. 84% 160 psig H2 2.0 hr. 88% 160 psig H2 2.0 hr. 97.5% 160 psig H2 2.0 hr. 92.7% Preaaara 1mg anverslan 60 psig N2 2.0 hr. 8.5% 60 psig N2 2.0 hr. 89% 400 psig l-i2 2.0 hr. 95% 60 psig N2 080 hr. 0-92.5% (9 samples) Praaaara 3111a anveralan 160 psig H2 1.0 hr. 22.5% 160 psig H2 2.0 hr. 29.2% 160 psig H2 3.0 hr. 35.6% 160 psig H2 2.0 hr. 26.9% Praasura 111a Qanvaralan 60 psig N2 0-4.0 hr. 0% (10 samples) 59 A major portion of this research focuses on the comparison of prod- ucts formed using Ca-Y zeolite in the organic solvent with those formed us- ing HZSM-5 in water. Figure 15 provides high pressure liquid chromato- grams of water soluble reaction products for a typical Ca-Y decalin run (CAYO.5) and a typical HZSM-S water run (K1, sample #7). Qualitative com- parison of the two spectra show that similar products are formed albeit in different amounts. Figure 16 provides a comparison of the gas chromato- grams of derivatized water soluble products from a Ca-Y decalin (CAY2.0). HZSM-S decalin (22.0). and HZSM-S water (K1. sample #7) runs. Again. qualitative comparison shows the same reaction products are present in each sample. .2216 IM A Ilcatl n Successful derivatization and separation of the water soluble products using 60 prompted the use of GC/MS to identify individual reaction prod- ucts (K1, sample #5). The total ion count chromatogram of the derivatized sample with the eight major peaks labeled according to scan number is shown in Figure 17 (K1, sample #5). The mass spectrum of each peak is ana- lyzed in the following sections using the methods outlined in Chapter 1 to identify each component. The mass spectrum for each peak is found in Ap- pendix B. 5.23.2 Peak 25: The mass spectrum of peak 255 is given in Figure B-1. The mass spec- trum shows a molecular weight of 290 which was determined by the parent ion at m/z 290 and the characteristic fragmentation at m/z 275 (M15). The absence of a prominent peak at m/z 187 (M-103) indicates that there are no primary hydroxyl groups present on the molecule. This indicates a ringed structure such as isosorbide which exhibits no primary hydroxyl fragments and is an expected product of this reaction. A derivatized isosorbide stan- dard matched the retention time and mass spectrum of the product represented by peak 255 exactly. Peak 255 is therefore identified as 311000001 010.: “am MI! (II-flu t lupin £010.: kulrut M41140 um film: m beret-rt “Ml an '8 “(S .Cluslt count In!» Kim kptrm 03413-8! 11131 b.0001 WI‘ 00m .8 10"2 volt. x 10-3 volts 60 i E S. “l i 6.00 -t n J < Ca-Y, Decalin Solvent a.“- ‘.00- E 3 3.00 « 1 “an“; __ _.. _ / ° N M . Y I r _l _ ' '_- _1 0.00 0.50 1.00 1.50 2. x 101 ninutes 7: 0.00- HZSM-s, Water Solvent 6.00- i F. g .- s ‘3 4.00 ~ 5! T T ' I ' I 0.00 0.50 1.00 1.50 x 101 minutes Figure 15. HPLC Tracings of Water Soluble Products 27.66 27'69 2‘88? 61 Ca-Y, Decalin Solvent HZSM-s. Decalin Solvent HZSM-s. Water Solvent Figure 16. SC Tracings of Water Soluble Products STOP STOP STOP 62 A: oEEaa .5: ESuoEEoEo 8:00 :0. .23. .5 952". om¢ 59¢ smm com smm saw emu as“ am a .J. .~_m¢mmm uuaa_ own 2.3 8;; «on :8 93 m? «an mm-_=a-mm mzcmcuu: Tm: zubm swim—1mm: + 2.: .5m Exam mzuzm :u mmzdmmmmmbs— 63 1.4:3.6-dlanhydro-D-glucitol (isosorbide) with a derivatized molecular formu- la of 06H602(OTMS)2. A representation of this molecule is given in Figure 18. 53,24 Paak 323 The mass spectrum of peak 329 (Figure B-2) exhibits the presence of two compounds of molecular weights 362 and 380 as evidenced by parent ions and characteristic fragments. The compound of mass 362 is represented by the parent ion at mlz 362. The presence of mlz 259 (M-103) indicates a primary hydroxyl group. The peak at mlz 319 (M-43) is indicative of a characteristic pinacol rearrangement discovered in previous studies [15]. Prominent peaks at mlz 157 and mlz 147 are also present which are indications of adjacent primary and secondary hydroxyl groups. Evidence of both a pinacol rearrangement and a primary hydroxyl group suggest a non-ringed structure. Feather and Harris [14] describe common enediol formation from D-glucose and related compounds by acid catalysis. The structure of this compound follows their hypothesis and a tentative assignment is made as 3,5,6-trihydroxy-8-hexene- 2-one. The structure is given in Figure 18. The derivatized molecular formu- la is CSH-,O(OTMS)3 which corresponds to MW of 362. The presence of the mass 380 compound can be seen from the charac- teristic fragments at m/z 365 (M-15) and m/z 290 (M90). The peak at mlz 277 indicates the presence of a primary hydroxyl group. A mass of 380 indicates a compound which is not fully derivatized; the incomplete derivatization is most likely the result of the hindered configuration of the anhydro-alditol compounds in accomodating the rather large TMS groups. A fully-derivatized single dehydration product of D-glucitol has a MW of 452 (06H80(OTMS)4 . 452). the same compound with a hydroxyl group underivatized yields a mass of 380 (06H802(OTMS)3 . 380). From the frag- OH OH 1 ,4:3,6-Dlanhydro-D-Glucltol o -o —o -I I I-o—o—o o o I 3,5,6-Trlhydroxy-3-hexene-2-one, Peak 329 Figure 18. Structure of Reaction Products 65 mentation patterns of the dehydration products analyzed. this compound is identified as 1,4-anhydro-D-glucitol with one hydroxyl group underivatized. 5,2.2.4 PQQK 555 The mass spectrum of peak 338 also represents a mixture of two com- pounds. The mass spectrum is given in Figure B-3. The compound of mass 380 is evident from the parent ion at mlz 380 and from peaks at m/z 365 (M-15) and mlz 290 (M-90). Mass 380 is again a marker for incomplete derivatization. The fragmentation pattern for this this compound is essentially the same as that seen in peak 329. suggesting another dehydration product with one primary hydroxyl group. Because of its longer retention time, this compound is identified as 3.6-anhydr07D-glucl- tol with one hydroxyl group underivatized. Evidence for a compound of mass 452 can be seen from the frag- ments at m/z 362 (M-90), mlz 349 (M-103) via loss of a primary hydroxyl group. and m/z 259 (M-90-103). The loss of a secondary hydroxyl group is seen by the fragments at m/z 247 (M-103-102) and m/z 157 (M-90-103- 102). Evidence of the loss of another secondary hydroxyl group is seen at mlz 145 (M-103-102-102). which also gives a peak at m/z 306. This com- pound also shows the distinct loss of a pinacol ion at m/z 319 (M-90-43). A possible structure of this compound is 4,5,6.-trihydroxy-hexane-Z-one. which is seen in Figure 19. Previous studies [14] site the formation of an in- termediate before 3,5.6-trihydroxy-3-hexene-2-one (peak 329) is formed. Figure 19 also provides a possible reaction pathway for the formation of both of these products. This compound is tentatively identified as the inter- mediate of molecular formula 06H80(OTMS)4. 5,2,2.5 Paak 547 The mass spectrum of peak 347 (Figure B-4) represents a compound of mass 452. This is deduced from the the parent ion at mlz 452 and charac- teristic fragments at mlz 437 (M-15) and m/z 362 (M-90). A very strong H H H O O O : H ICI CICHHCI CI CIH H O H H H H #0 2 u." H H H HH HHH OOHOOO H lClClCIClClCIH H H O H H H Mass 452, Peak 338 Mass 362, Peak 329 D-Glucltol Figure 19. Non-Cyclic Reaction Product Formation 67 peak at m/z 103 indicates that more than one primary hydroxyl group may be present. Evidence for the loss of a secondary hydroxyl group is not found. An expected reaction product which exhibits a mass of 452 and has two pri- mary hydroxyl groups are the 2.5-anhydro compounds. GC and HPLC stan- dards using 2,5-anhydro-D-mannitol gave retention times very close to this compound. The mass spectrum of the 2,5-anhydro-D-mannitol standard is given in Figure B-5 and matches this compound exactly. This peak is identi- fied as a mixture of 2,5-anhydro-D-iditol and 2,5-anhydro-D-mannitol which have a molecular formula of C6H80(OTMS)4. 5,2.2,§ Paak 555 The mass spectrum of peak 358 (Figure B-6) also represents a mass 452 compound. This compound exhibits characteristic M-15 and M-90 peaks and gives evidence of a primary hydroxyl group at mlz 349 (M-103) and mlz 259 (M-90-108). The peak at m/z 157 (M-90-103-102) also gives evidence of a secondary hydroxyl group. The only expected reaction product to exhibit this fragmentation are single dehydration products. 1,4- or 3,6-anhy- dro-D-glucitol. Since the former exhibits a shorter retention time, this peak is identified as 1,4-anhydro-D-glucitol with a molecular formula of 06H80(OTMS)4. 5,33,? Peak 565 This compound of mass 452 is evident from the characteristic M-15 and M-90 ions. The mass spectrum of peak 369 found in Figure B-7 shows an almost an identical fragmentation pattern as that seen in peak 358. With the longer retention time. this peak is identified as 3,6-anhydro-D-glucitol. 5.2.2.5 Paak 415 The mass spectrum of peak 416 in Figure B-8 represents a compound of mass 542. This is seen from the fragment ions at mlz 527 (M-15) and mlz 452 (NI-90). Gas chromatography standards show TMS derivatization problems with the starting material D-glucitol, resulting in peaks at the same 68 retention time. Fully derivatized D-glucitol exhibits a derivatized molecular weight of 614 (06H8(OTMS)6 - 614); D-glucitol with one hydroxyl underivatized exhibits a molecular weight of 542 (CBH1OO(OTMS)5 - 542). Therefore. this peak is identified as D-glucitol with one hydroxyl underiva- tized. 53.2.5 Paatt, 439 Peak 420, seen in Figure B-9, represents a compound of mass 614. Gas chromatography and HPLC standards of D-glucitol match this com- pound exactly and Petersson's [46] MS standard of TMS derivatized D-gluci- tol (Figure B—9) also matches this compound. Peak 420 is identified as the starting material D-giucitol. .2.2.1 HPL A lie I it Because the derivatization step required for CC analysis proved time consuming and was inaccurate due to incomplete derivatization of certain products, HPLC was used to determine yields of reaction products. injection of standard samples and comparison with gas chromato- grams were used to identify reaction products and calculate product yields from the HPLC results. The different concentrations of reaction products in various experiments provided a link between GC and HPLC spectra. The use of HPLC provides direct injection of the aqueous phase without the dry- ing and derivatization steps which add error in determining product yields. A complete listing of the GC and corresponding HPLC retention times for each compound identified can be seen in Table 4. 5,2.5 ldantlfiaatlgn and Ylald Calgalatlon at Qrganla §glubla Produata Reaction products were evident in the organic phase by the deep gold- en color of the post-reaction organic solvent. Direct injection of the post- reaction solvent into the GC yielded only very large peaks for decalin which masked all product peaks. This warranted the separation of the reaction products by methanol extraction as outlined in Chapter 4. After extraction. 69 Table 4 Summary of Water Soluble Products at; B, llmg LIPLQ B, Tlmg EJ291151 W OH lsosorbide O 28.9 min. 17.2 min. 0 OH MW 8 290 1,4-Anhydro-D-Glucitol 35.1 min. 13.8 min. 38:1 a OH OH MW - 452 3,6-Anhydro-D-Glucitol 36.1 min. 16.4 min. 0 Q0” OH OH OH MW = 452 2,5-Anhydro-D-Mannitoi 34.6 min. 11.4 min. O HO W... OH MW =- 452 2,5-Anhydro-L-lditol O 34.6 min. 11.3 min. OH HO OH MW 3 452 Mass 362 Enone Figure 18 33.5 min. 13.9 min. Mass 452 Ketone Figure 19 34.3 min. unknown D-Glucitol Figure 19 39.1 min. 13.1 min. MW - 614 70 the resultant dark brown fluid was analyzed by GC without interference from decafin. Gas chromatography of the extracted organic products revealed a large number of volatile compounds with retention times from 3 to 24 min- utes (Figure 20). The CAY2.0 tracing represents a typical array of lighter or- ganlc reaction products which result from reaction at 260°C. The CAY1.0.240 tracing is part of a series of experiments performed at 240°C to isolate intermediate organic soluble products found at higher retention times. The same products are present in each case but different tempera- tures and reaction times provided higher concentrations'of the respective products which makes analysis by mass spectrometry much easier. The following sections describe the identification of major organic reaction prod- ucts. Mass spectra of all the labeled peaks in Figure 20 are found in Appen- dix B. 5.2.5.1 Paak 77 The mass spectrum of peak 77, found in Figure B-11. represents a compound of mass 98. The simple fragmentation pattern of m/z 98, mlz 83. m/z 82, m/z 69, and lower ions is typical of a small ringed compound with a furan-type structure [52]. Shaw et al [11] recovered several mass 98 com- pounds in the acid-catalyzed degradation of sugars. One particular com- pound. B-angelica Iactone. has the same fragmentation pattern as peak 77 (Figure B-12) [53]. Peak 77 is identified as 5-methyl-2(5H)-furanone (B-an- gelica iactone) with the molecular formula 05H802. Leonard [54] postulated the formation of this iactone from levulinic acid, which is a common degrada- tion product of 5-hydroxymethyl furfural. W The mass spectra of peaks 86 and 98 have essentially the same fragmentation patterns (Figures B-13 and B-14). The molecular weight of 110 is apparent from the abundunt parent ion at m/z 110. The breakdown in- 6.02 2“ I$.8i I7. I4 71 237 'r x, o a oat ' in: 205 1?. 2’3. 5: n2 1“ 139 3 :2 3 ‘3 l 165 j i a l 86 9." 3 a? 1308 ’i a? i °' “- “i . r0 0'0 a: -> . :2 i Q'SI \LJ : V on _ --__; Experiment CAY2.0 T 253 283 299307323 332 ‘3‘“ 3. 9:8? as: 59-35 2773 as“ (are " fir-.312 [i EL» .05 Experiment CAY1.0,240 Figure 20. CC Tracing of Organic Products 72 to the fragments found at m/z 81 and mlz 53 is typical of a substituted furan reviewed in Chapter 1 [47]. The peak at m/z 81 (M-29) represents the loss of the substituent on the ring which is most likely a carbonyl group (COH). The mass 53 fragment is the typical rearrangement structure re- viewed earlier. A furan type compound with an similar expected fragmenta- tion pattern is 2-furanacetaldehyde [55]. Since no standard samples were obtainable, peaks 86 and 98 are tentatively identified as compounds of molecular formula CSHGOZ with structures similar to 2-furanacetaldehyde. 5,2,5,5 Eaak 111 Peak 111 was found to be a major organic product in all experiments. The mass spectrum can be found in Figure B-15. This compound also exhib- its an abundant parent ion at m/z 110 and has a simple breakdown characteristic of a substituted furan. The peak at m/z 95 (M-15) represents the loss of a methyl group substituent. Investigation of mass 110 furan type compounds revealed a standard match (Figure B-18) [56]. Peak 111 is identified as 2-furyl methyl ketone with a molecular formula of 06H602' 5.13.4 Paak 118513; Peaks 118 and 139 are mass 110 compounds which have similar break- down patterns (Figures B-16 and B-17). Both peaks follow a typical furan breakdown but ions at both m/z 95 and mlz 81 give evidence of different structure. The peak at mlz 109 represents the loss of a hydrogen from a car- bonyl group which is evident from the m/z 81 (M29) ion. The m/z 95 (M- .. 15) ion represents the loss of a substituted methyl group. The presence of both ions yields evidence of separate substitution. A mass 110 compound which explains this result is 5-methyl furfural. This compound was found to be a major sugar degradation product [11] and is a major product in these ex- periments. Gas chromatography of a 5-methyl furfural standard revealed a retention time which matched peak 139 but did not explain peak 118. Fur- ther investigation of this compound revealed a standard match (Figure B- 73 18) [57]. Peaks 118 and 139 are identified as 5-methyl furfural with a molecular formula of CSHSOZ' W The mass spectrum of peak 165 found in Figure B-19 exhibits the breakdown of a mixture of mass 110 and mass 112 compounds. The mass 110 compound follows the fragmentation of 5-methyl furfural discussed earli- er. The mass 112 compound could be 5-methyl furfuryl alcohol which is the only furan compound of the appropriate molecular weight. Typical break- down patterns for 5-methyl furfuryl alcohol would be m/z 95 (M-17, -OH), mlz 83-82-82 (M-17-CH2). and the rearrangement structure at mlz 53 which are all evident for this compound. The gas chromatogram of a standard sample revealed a shorter (12.0 min) retention time. The identification of this mass 112 compound as 5-methyl furfuryl alcohol (C6H802) is tentative. W The mass spectra of peaks 205 and 237, represented in Figures 8-20 and B-21. indicate compounds of structures differing from substituted furans. Mass 126 compounds found in previous studies include 5- hydroxymethyl furan, 2,(2-hydroxyacetyl) furan. isomaltol (2-acetyl-3- hydroxyfuran), and various pyrones in very small quantities [11, 12, 13]. Peaks 205 and 237 exhibit similar fragmentation patterns with the exception that the parent ion at .m/z 126 is not seen in the mass spectrum of peak 205. The breakdown pattern is typical of a pyrone compound reviewed in Chapter 1 [48]. Pyrones usually exhibit the loss of CO (M28) and the loss of the primary substituent like a methyl group (M-15) before rearranging in- to a structure which gives an ion at mlz 69. This type of fragmentation fits the compounds in these peaks perfectly. A matching mass spectrum was found for 4-hydroxy-6-methyl-2-pyrone which can be seen in Figure B-22. Subsequent research with an actual standard of this compound revealed a longer (14.5 min) GC retention time. Although the mass spectra match 74 very well. the reaction product can not be positively identified. lsomaltoi. al- though of furan structure, was found in much higher concentrations in litera- tures studies and exhibits a similar mass spectrum (Figure B-22). The break- down of this unknown follows the characteristics of a pyrone of molecular formula C6H603 but a definite identification can not be made at this time. W All peaks labeled as peak D, in both 60 tracings in Figure 20, repre- sent the reaction solvent decalin. The mass spectrum for decalin can be seen in Figure B-23. The two peaks evident in the GC tracings represent the cis and trans isomers of the mass 138 (C10H18) reaction solvent. The fragmentation of decalin is typical for a ringed hydrocarbon in that many ions are present from a long series of rearrangements. 5.2.5.5 Peak T All peaks labeled with T have been identified as the major decalin impurity tetrahydronaphthalene (C1OH12). The mass spectrum of tetralin can be found in Figure B-24. Tetralin is the result of incomplete hydrogena- tion of naphthalene in route to decalin. Therefore tetralin exhibits a molecu- lar weight of 132 along with a typical ringed hydrocarbon breakdown. 2. . P ak 277 2 8. 2 Peaks 277, 283 and 299 all exhibit the same fragmentation pattern (Figures B-25, B-26, B-27). These compounds may represent intermediates in the formation of the furfural compounds identified earlier since they were isolated from the experiments run at 240°C. Peak 277 may contain another decalin impurity along with a compound which fragments like those in peaks 283 and 299. The mass spectra provide an abundant parent ion at mlz 154. Also seen is the apparent loss of a water molecule at m/z 136 (M-18). Fur- ther fragmentation follows a classic iactone breakdown outlined in Chapter 1 [49]. Previous studies in sugar degradation have not reported the isola- tion of similar high mass compounds. The degradation of decalin can be 75 ruled out because blank runs showed no breakdown of the solvent with the catalyst. Investigation of mass 154 compounds revealed a series of lactones classified as spirocyclics which gave similar breakdown patterns [58]. The ambiguity of the spirocyclic structure (see Figure 21) and the molecular formula (C8H1003) do not fit into the sequence of reactions known to oc- cur in sugar degradation. Therefore the identification of the compounds rep- resented in peaks 277, 283, and 299 will be classified as mass 154 lactones. .2 .1 P k 7 12 2 The mass spectra for peaks 307, 312, 323, and 332 shown in Figures B-28 through B-31. All of the compounds represented by these peaks exhib— it the same fragmentation pattern. All have an apparent molecular weight of 154 and have nearly the same breakdown pattern as encountered in peaks 277, 283, and 299. This series of compounds exhibits a small mlz 154 par- ent ion and a large m/z 136 (M-18) ion while the previous mass 154 com- pounds exhibited a large parent ion and a large m/z 111 ion. This fragmenta- tion is also explained by a similar series of spirocyclic lactones [59]. Again the structure (see Figure 21) and molecular formula (09H1402) of these compounds does not fit the reaction scheme. Further investigation did not reveal any definite identification of these compounds. These peaks will also be identified as mass 154 lactones. .2..11vaw f r ni Pr A summary of all isolated organic soluble products is given in Table 5. Structure, molecular weight, and GC retention times have also been provid- ed for each compound. Since the organic soluble products are volatile with- out derivatization, quantitative studies were performed using GC peak area. 4 n I Iv An i I After the identification of the major reaction products was completed, quantitative studies were performed to reveal reaction pathways and deter- mine conditional effects. The procedures outlined in Chapter 4 were used to 76 0 II L L /[/\ +_+ mm... mlz 1 54 (Cal-l1 003) ,9 I? O = o / H2C mlz 154 (C9H1402) m/z 111 (100%) i mlz 98 Figure 21. Spirocyciic Lactone Structure and Fragmentation Summary of Organic Soluble Products 940.0221 5-MethyI-2(5H)-Furanone 2-Furanacetaldehyde 2-FuryI-Methyl Ketone 5-Methyl Furfurai 5-Methyi Furfuryl Alcohol 4-Hydroxy-6-Methyl-2-Pyrone Mass 154 Lactones (Group 1) Mass 154 Lactones (Group 2) §olvant Decahydronaphthalene Tetrahydronaphthalene 77 Tables 5311191010 H30 H O \\ MW-98 l l i? O MW-110 l l 9| 0/ MW-110 l I? O MW=110 -|——T ‘I'\ o/‘I‘OH MW-112 OH H3 0 *0 MW .126 Unknown Unknown Mil—'9. 03 MW-‘I38 ()3 MW .132 9.9m 6.0 min. 9.0 - 9.7 min 10.5 min. 11.0 - 12.1 min. 12.9 min. 13.7 min. 20.6 - 21.4 min. 21.8 - 23.3 min. GQ R. Time 15.4 - 17.9 min. 18.7 min. 78 calculate product yields with the assumption that the amount of starting ma~ terial not recovered in the aqueous phase is present in the organic soluble products. To justify this step, every possible route of reactant/product loss was investigated. The blank runs showed that no starting material was lost due to ther- mal degradation or reaction with the solvent. Several experiments were per- formed to test recovery of standard amounts of starting material from cata- lyst-containing solutions. Results showed quantitative recovery was achieved using the developed methods indicating that no product was ab- sorbed into the porous catalyst following reaction. -A reactant/product loss could also occur via the charring of reactant on the catalyst surface during reaction. Several studies were done using ESCA to compare the carbon con- tent of fresh and spent catalyst. Results showed no significant buildup of carbon on the spent catalyst surface. The remaining possible pathway for the loss of reactant is conversion to products which are not soluble in the aqueous phase, or detectable with the analytical techniques employed. However, the compounds recovered in other studies involving sugars and sugar products were all detectable using derivatization and gas chromatography, so the formation of undetectable products is unlikely. Finally, no gaseous products were formed, as evidenced by no significant pressure rise from the starting gas charge after reaction, and the absence of peaks from gas chromatography of collected samples. Table 6 provides the results of quantitative analysis on the decalin sol- vent experiments. The percentage yields for each product are listed along with the total percent conversion. Table 7 provides the yield percentages of water soluble products from Kinetic Run 1 (K1). The percentage of each product from each timed sample is listed along with a total conversion at each interval. 79 Table 6 Summary of Quantitative Results for Decalin Solvent Reactions Reactant: D-Glucitol Catalyst: Ca-Y62,0% Pt Time (hrs.): 0.5 1.0 2.0 0.5 1.0 2.0 4.0 Temp (°C): 240 240 240 260 260 260 260 Mum D-Glucitol 8.2 5.4 3.5 3.6 3.1 2.6 1.9 1,4-Anhydro 28.5 17.9 6.2 12.4 8.9 4.2 1.7 3,6-Anhydro 3.8 2.8 1.9 2.9 2.1 1.3 0.8 2,5-Anhydro 1.9 1.9 0.8 2.1 0.4 0.2 0.1 Enols, Enones 1.0 2.0 3.9 1.0 1.5 1.9 1.9 isosorbide 23.2 33.0 36.8 25.4 29.2 36.6 35.5 B-Ang. Lactone 0.0 0.0 0.0 0.1 0.0 0.0 0.0 2-Furanacetal. 1.2 5.2 8.4 11.2 7.7 1.3 11.7 2-F. M. Ketone 7.7 10.4 16.1 16.5 16.2 19.1 17.1 5-M. Furfurai 1.4 2.5 8.3 8.7 9.7 8.4 12.0 5-M. Furfuryl Alc. 0.0 0.1 0.9 0.1 0.0 0.0 0.0 Pyrones 3.8 6.7 10.3 11.9 13.1 17.8 11.0 Magma 19.3 12.3 3.7 4.3 8.2 6.0 6.1 % Sorbitol Conv. 91.8 94.6 96.5 96.4 96.9 97.4 98.1 Reactant: D-Glucitoi lsosorbide Catalyst: Ca-Y62 HZSM-S HZSM-S Ca-Y62 HZSM-5 5% Pt 0% Ni 5% Ni 0% Pt 0% Ni Time (hrs.): 1.0 2.0 2.0 2.0 2.0 Temp (°C): 260 180 180 260 180 Progugta D-Giucitol 9.4 2.6 7.0 0.0 0.0 1,4-Anhydro 26.0 2.4 7.7 0.0 0.0 3,6-Anhydro 8.6 0.4 1.3 0.0 0.0 2,5-Anhydro 0.0 0.0 0.8 0.0 0.0 Enols, Enones 5.8 3.3 0.2 0.0 0.0 lsosorbide 12.2 49.5 44.2 75.6 73.1 B-Ang. Lactone 1.7 0.0 0.4 0.3 1.7 2-Furanacetal. 4.3 0.6 1.2 1.0 6.7 2-F. M. Ketone 14.0 9.4 7.0 1.1 1.3 5-M. Furfurai 0.4 7.0 3.6 1.3 0.4 5-M. Furfuryl Aic. 0.0 0.0 0.0 0.3 0.0 Pyrones 1.5 9.1 7.2 15.5 8.9 mt; 154 Lagtgnaa 15.6 45.6 19.4 4.9 7.9 % Conv. 90.6 97.4 93.0 24.4 26.9 80 Table 7 Summary of Quantitative Results for Water Solvent Kinetic Run 1 (Exp. K1) Reactant: D-Glucitol Catalyst: HZSM-S, 0% Ni Temp (°C): 260 Tlme(mln.) 0 15 30 45 60 90 120 150 180 mm: 1,4-Anhydro (1) 4.6 10.4 12.5 25.6 29.0 28.0 23.6 17.5 15.0 3,6-Anhydro (2) 1.6 4.9 6.4 5.6 4.8 3.1 2.5 2.3 1.3 2,5-Anhydro (3) 3.5 4.5 5.3 5.9 5.6 5.7 6.5 7.0 7.0 Enols, Enones (4) 0.0 1.7 3.1 6.4 7.2 9.2 9.7 10.4 10.5 Isosorbide(5) 0.0 3.4 11.0 18.1 256 37.9 46.7 53.7 54.9 - l i 4 1 9.0 8.2 % Conv. 9.7 24.4 35.9 61.9 72.8 83.9 89.1 91.0 91.8 81 EaEEam 8:02.... cozoaom :3: 0:25. .3 0.52“. 2.3 m2: OO- . 8— ON— Om xi IX iXIl OH - JDI it? . 'I. it Li 33030-0 029800.. mococm .m_o:m $352-94.. 925266 8355-: DXDNOO .00 won o0— CI'IBIA Z 310W 82 5.3 Discussion of Results .1 if IR l n 11 II 11 Close inspection of the conversion and product distribution results presented in Tables 3, 6, and 7 reveal several characteristics which effect the catalytic conversion in the liquid phase experiments. Previous studies [15] revealed that temperature has the strongest ef- fect on conversion in the decalin Ca-Y experiments, with reaction tempera- tures of at least 240°C required for significant conversion. The Ca-Y experi- ments performed in this research confirmed this effect with conversions above 90% at temperatures at or above 240°C. The HZSM-S experiments in decalin were found to require much lower temperatures (180°C) for conver- sion of greater than 90 %. This is because HZSM-5 is a much stronger acid dehydration catalyst. Comparison of the 240 °C and 260°C Ca-Y experiments in Table 6 al- so show a significant temperature effect. Although the total conversions of D-glucitol are similar, the relative amounts of the reaction products are different. The lower temperature runs show a greater recovery of water soluble products and larger amounts of intermediate products like 1,4-anhy- dro-D-giucitol and the m/z 154 lactones. The effect of pressure is insignificant when compared to temperature. Previous studies [15] showed that an increase in pressure of 800 psig only increased the conversion by 9%, whereas an increase in temperature of 20°C (220-240°C) increased conversion by 53%. These results warranted no fur- ther investigation of pressure effects in this research. Tables 6 and 7 also provide results for the comparison of reaction sol- vents. By comparing typical one hour runs at identical temperatures for each solvent, it is evident that the reaction solvent affects reaction rate and product formation. The decalin experimental results show a much higher 83 conversion and greater concentrations of organic soluble products. The wa- ter experimental results show a lower conversion and virtually all starting material is accounted for in the water soluble products; small amounts of or- ganic-soluble products were detected in the gas chromatograms of the water reaction solutions but constituted less than 5% of the starting material. The decalin solvent promotes the further conversion or degradation of water soluble products into organic soluble products. Literature results show that anhydrosugar products degrade into furfural type compounds much faster in anhydrous conditions than in water [14]. Experiments were performed using the reaction product isosorbide to reveal the mode of formation of the organic products. Blank runs using isos- orbide in decalin without catalyst resulted in a negligible (2%) loss of start- ing material at a reaction temperature of 260°C, showing that isosorbide is thermally stable at reaction conditions. The DEG1 experiment reviewed in Table 3 shows a 22.5% conversion of isosorbide with Ca-Y zeolite present under identical conditions. The K2 experiment (Table 3) tested the catalytic effect of HZSM-S on isosorbide in a water reaction solvent. Results for this experiment show no loss of starting material after 4 hours at 260°C. These results prove that the formation of organic products from isosorbide is a result of a combination of both an organic solvent and the presence of a catalyst, and that the organic products detected in the water solutions must result from the single dehydration products of D-glucitol. The decalin solvent in contrast provides the anhydrous conditions required for the cata- lytic conversion of isosorbide and other initial water-soluble dehydration products. These experiments also show that isosorbide is a relatively unreac- tive product of the liquid phase reactions. The effects of reaction time are evident from the concentrations of products listed in Tables 6 and 7. inspection of the yield distributions for the 260°C Ca-Y experiments show the rapid formation of intermediate 84 products which slowly react further with time. A better example of the ef- fect of reaction time can be seen in Table 7 and the corresponding Figure 22. The results from this kinetic experiment with a water reaction solvent show the formation and subsequent reaction of all identified water soluble products. Results from both solvents show very fast reaction rates and the formation of a greater amount of organic soluble products with increasing re- action time, although the total quantity of organic soluble products formed in water is again small. Several experiments were performed to reveal the effects of different types of catalysts on the conversion of D-glucitol. The comparison of 2 hour Ca-Y and HZSM-5 runs in decalin show almost identical overall conver- sions. The difference between the runs lies in the temperature required for high conversions. The stronger, more acidic HZSM-5 catalyst required a 180°C reaction temperature to attain the same conversion as the Ca-Y cata- lyst at 260°C. An interesting result, although expected because of the strong temperature effect, is the greater amount of organic products formed with Y zeolite at the higher reaction temperature. Other experiments were run to test the effect of deposited Group VIII metals on the reactivity of both catalysts. The comparison of Ca-Y and Ca- Y,5% Pt experiments under identical conditions show differences in overall conversion and resultant product concentrations. The exchange of Pt onto the catalyst removes Ca ions which in turn reduces the dehydration activity of the zeolite. This results in lower D-glucitol conversion. The lower activi- ty also results in a greater concentration of intermediate products like 1.4- anhydro-D-glucitol compounds. Another result is the greater concentration of the enoi or enone compounds with the Pt exchanged catalyst. Platinum may promote the dehydrogenation required for the formation of these type of products. Similar results were seen for the HZSM-S and HZSM-5,5% Ni catalysts. identical runs with each catalyst resulted in a lower conversion 85 and a lower dehydration activity evident from the higher concentrations of intermediate products. The desired hydrogenation activity of the Group VIII metals was not observed in these experiments. v R l nP hw The identification of the reaction products and the calculation of their relative amounts at different reaction time intervals led to the formulation of a reaction pathway outlined in Figure 23. The dehydration of D-glucitol to form the water soluble compounds is an expected result from the literature reviewed in Chapter 1. An exception to the expected results was the observed formation of the enoi and enone compounds. The literature sites the formation of similar compounds as vari- ous intermediates, but never comments on their isolation from acid-catalyzed sugar reactions [14]. The formation of 1,4:3,6-dianhydro-D-glucitoi by the second dehydration of the 1,4- and 3,6-anhydro-D-glucitol compounds is al- so expected from previous results in alditol dehydration. The data from Kinetic Run 1 presented in Figure 22 provide insight in- to the preferred reaction pathways. it is evident that 1,4-anhydro-D-giuci- tol is initially formed in the greatest amount. When the amount of starting product has been depleted, further dehydration of 1,4-anhydro-D-glucitol compound to isosorbide is observed. The small amount of 3,6-anhydro-D- glucitol initially formed rapidly reacts to form isosorbide. This result sug- gests 3,6-anhydro-D-glucitol as the desired route to isosorbide while the preferred initial dehydration product is 1,4-anhydro-D-glucitol. The forma- tion of the 2,5-anhydro-alditol and enoi/enone compounds is slow and there is no further reaction observed in the water solvent experiments. Long reaction times would result in the similar organic product formation seen in the decalin experiments. The pathways for the formation of the organic soluble products were revealed by several methods. Time interval methods like those used in the water experiments did not provide consistent or conclusive results. The con- 86 0:352. .3103: 002030 .2 2:2“. .225“. 15.3.... .... “.3: 2.0.3. .5503 .>.:m.« eu»..eo.3eo-:e.:u.« ‘\\\/4 ./‘K OIO, 3:0 «865-0 «0 mac—0655. . m: — — III PliPllli —Il AV pm”. my canmvflmu I _ Il— JI— .a l .u n a u c. w H... .a «8:20 .aaxm £63! '2'” mi 00 .vu 0.52m '71.! '91 (0'82: 8 9V6 88 ( 02" 99 l 62" EI'Q .9. m... le¢ 6"? [8'6 [0' I2 6912 use ta'6l [8'9l St'tl tz'll 86'tl W” J . mu C 2' fit It tl'S 621 fi 99 c a" j . ‘£ Ilsa? 38'? 95'? Figure 25. GO Tracing of MC-3,500 Experiment 97 dOIS 20.52... one .o 9:09.... 00 .3 2:2“. £92 98 8co§3axm 3990: .o 89:02... 00 ...a 2:2“. IL. ”} p F 28'tl l8 “'9! ll 6"1 {9'2 3(3 99 The addition of platinum may have increased the formation of gas products (higher hydrocracking activity) or may have lowered the activity of the cata- lyst enough to exhibit the conversion of thermal degradation products to gas products seen in the Y-62 experiment. n n l l The quantitative analysis of vapor phase experiments was based on the amount of isosorbide. liquid products. and gas products recovered since the reaction products were not identified. The procedures outlined in Chap- ter 4 were used to calculate yields of each type of product. The assumption that the amount of starting material not recovered was converted into detectable products was justified by the investigation of reactant loss routes. Since the reaction system is capable of collecting liquid and gas prod- ucts and the quantitative recovery methods have been proven, the only apparent route for reactant loss would be from far formation on the catalyst or sides of the reactor tubing. The blank runs showed no tar formation on the glass beads or reactor tubing although thermal degradation was taking place. To test for far formation on the catalysts, several experiments were run to measure catalyst weight after calcination/reduction, rinsing, and dry- ing. The results from these experiments were then used as a comparison to the actual catalyst runs in which the catalyst was weighed before calcination and after the recovery procedures were completed. All results (see Table 9) showed very little weight gain indicating that reactant loss due to far forma- tion on the surface of the catalyst was negligible. The appearance of the re- actor tubing also indicated that a negligible amount of isosorbide was lost by tar deposition. The formation of gas and liquid products presented problems in calculating yields. The molar amount of gas products formed was estimated using the ideal gas law from the expansion of the trap contents into the pressure gauge system. By assuming an average molecular weight of 40 100 Table 9 Summary of Catalyst Weights for Vapor Phase Experiments Wt. before Wt. after Elm. 9.2191111. MI "I | n 8111291893.?! Calcination MC-3 1.4208 9. 1.4343 g. Blank2 Glass Beads 4.2785 g. 4.2620 9. 1403.500 MC-3 1.4257 g. 1.3743 g. MC3.8OO MC-3 1.4278 9. 1.3731 g. 101 g/gmol. the amount of gas products was calculated (see Appendix A for calculations). The amount of liquid products was estimated by multiplying the relative peaks areas of the GC chromatograms by a response factor calculated from relative peak areas of standard concentrations of 2-furfural which is a similar organic compound. Table 10 provides the results of quantitative analysis on the vapor phase experiments. The percentage of each type of product recovered is listed along with a total isosorbide conversion. 6.3 Discussion of Results Inspection of the conversion and product distribution presented in Ta- ble 9 and in Figures 25, 26, and 27 reveal characteristics of the catalytic conversion of isosorbide in the vapor phase experiments. Temperature has a strong effect in that little or no conversion of isos- orbide was detected below 500°C. The blank run results showed that isosor- bide thermally degraded into volatile compounds at the same temperature. Nearly complete conversions were achieved at 500°C, so reactions were not . run at higher temperatures. The effect of hydrogen pressure is evident in the comparison of the M03500 and MC3.800 experiments. The higher hydrogen pressure results in the recovery of more gas products suggesting higher hydrocracking activi- ty or an enhancement of thermal degradation products which are in turn con- verted to gas products. The effects of different types of catalysts on the lsosorbide conver- sion are given in Figures 24 and 25. Comparison of the Y-62 and MC-3 product chromatograms show formation of a much different array of organic products. The Y-62 experiment shows the formation of compounds similar to those seen via thermal degradation, while the MC-3 experiment exhibited products similar to those seen in the liquid phase experiments. The 102 Table 10 Summary of Quantitative Results for Vapor Phase Experlments £393; ngalyst 1mm Prsssgrs P[%sng'qstus|d firsgagassts % anv. Blank1 - 300°C 500 psig. 0% 0% 2.6% Blank2 - 500°C 500 psig. 12.7% 51.1% 85.8% Y62.500 HY62 500°C 500 psig. 4.9% 66.7% 84.3% MC3,500 HMO-3 500°C 500 psig. 12.1% 51.1% 97.1% MC3,800 HMO-3 500°C 800 psig. 1.5% 66.7% 100% MCSPt HMO-3.05% Pl 500°C 500 psig. 2.9% 66.7% 97.0% 103 stronger. more acidic MC-3 catalyst led to more dehydration activity while the Y-62 catalyst seemed to merely aid thermal degradation to gas prod- ucts. The 800 psig experiment with the MC-3 catalyst exhibits a product distribution similar to that for the Y-62 catalyst. suggesting that the catalyt- ic activity of the MC—3 catalyst can be controlled by varying the hydrogen pressure. Higher pressures may lead to high yields of cracked gas products while lower pressures may produce high yields of volatile liquid products. The primary catalytic activity observed in the vapor phase experi- ments was hydrocracking. lsosorbide was effectively cracked by both Y and MC-3 (ZSM-5 extrudate) catalysts to produce llght liquid products and gas products. An encouraging result is that little or no tar formation was detect- ed in these experiments. Previous sugar conversion studies using zeolites reported the formation of tars which quickly deactivated the catalysts. The major problem in these experiments is the inability to attain an ac- curate material balance. The production of both liquid and gas products makes quantitative recovery difficult. The loss of product gases is the most probable cause of the inaccurate balance. The existing gas collection appara- tus is effective in collecting a representative gas sample but could be allow- ing some of the total amount of gas produced escape due to an insufficient volume. A new design or renovation of the existing apparatus for quantita- tive recovery of gases is needed for further research. CHAPTER 7 CONCLUSIONS AND RECOMMENDATIONS 7.1 Liquid Phase Experiments The results of the liquid phase experiments presented in Chapter 5 give evidence that the catalytic dehydration of D-glucitol can be effectively accomplished using zeolite catalysts. The primary catalytic activity observed for the conversion of D-gluci- tol was dehydration. Both the Ca-Y and ZSM-S catalysts exhibited this dehydration activity with the initial formation of anhydro-alditol compounds such as 1,4-anhydro-D-glucitol. This single dehydration product was then further dehydrated to 1,4:3,6-dianhydro-D-glucitol (isosorbide) which was the major reaction product. Small amounts of organic soluble products such as 2-furyl methyl ketone and S-methyl furfural were also isolated from the D- glucitol reactions. Experimental results also showed that D-glucitol could be converted in water reaction solvent as well as in an organic solvent like 104 105 decahydronaphthalene (decalin). The stronger. more acidic ZSM-5 zeolite proved to be an effective dehydration catalyst in water. Comparison of products from a Ca-Y run in decalin to a ZSM-5 run in water showed the for- mation of identical products with the anhydrous conditions of the decalin promoting higher conversions of organic soluble products. The presence of Group VIII metals such as platinum and nickel on the zeolite catalysts produced no apparent hydrogenation activity. Comparison of non-metal to metal catalysts in both reaction solvents showed only a de- crease in catalytic activity due to the replacement of acidic dehydration sites with the metal ions. Quantitative studies of the D-glucitol reactions resulted in conditional effects and reaction pathways. Temperature was found to have the greatest effect with no activity at 220°C and conversion up to 98% at 260°C in un- der four hours of reaction time in both reaction solvents. The comparison of the decalin and water solvents showed identical product formation with the decalin promoting the formation of organic products. Similar D-glucitol conversions in the decalin solvent were seen at 260°C for the Y catalyst and 180°C for the ZSM-5 catalyst indicating that ZSM-5 is a much stronger dehydration catalyst. The observed reaction pathways suggested that the organic soluble products resulted primarily from the degradation of the sin- gle dehydration products. lsosorbide was found to be relatively unreactive in liquid phase experiments. Recommendations for further. research in the liquid phase conversion of D-glucitol involve enhancing the activity of the zeolite catalysts toward the production of one type of product. The single dehydration products were found to be reactive and easily converted to organic products by ther- mal degradation. Selectivity towards a single dehydration compound could be achieved by lowering the activity of the catalyst by changing reaction conditions or by ion exchange. lsosorbide, although unreactive in the liquid 106 phase. is relatively volatile and could be further converted in a downstream gas-phase conversion system. The selective production of isosorbide is best accomplished with the ZSM-5 catalyst in the water solvent; yields of up to 75% are possible with the current reaction conditions. The use of water as a reaction solvent is most practical since it is used in all of the processes contained in the proposed reaction scheme. 7.2 Vapor Phase Experiments The results of the vapor phase experiments presented in Chapter 6 suggest that isosorbide can be converted into liquid and gas products. Pre- liminary experiments show that high conversions of isosorbide have been ac- complished at 500°C and 500-800 psig of hydrogen using zeolite catalysts in the fixed-bed. flow reactor described in Chapter 3. The observed catalytic activities include dehydration and cracking. Blank runs using glass beads showed that isosorbide degrades into volatile liquid compounds at 500°C. The use of Y-62 zeolite at the same conditions produced smaller amounts of the same compounds as well as gas products in- dicating cracking activity. The use of MC-3 (ZSM-S extrudate) at the same reaction conditons produced dehydration products similar to those seen in the liquid phase, organic soluble products. By increasing the hydrogen pres- sure to 800 psig, the MC-3 catalyst produced cracked gas products identical to those seen in the Y-62 experiment. This suggests a direct connection be- tween hydrogen pressure and cracking activity. . A major result from these experiments is that little or no tar forma- tion was detected. Previous zeolite-catalyzed sugar conversion processes had serious problems with the reactant forming tar on the catalyst surface, deactivating the catalyst. lsosorbide may be the answer to the tarring prob- lem and provide a link between biomass sugars and hydrocarbon products. 107 Further investigation into isosorbide conversion should begin with the identification of all reaction products. The volatility of the observed products should make gas chromatography/mass spectrometry a practical identification tool. Once the products are identified, reaction conditions can be studied for the optimum conversion to desirable products. The final step is to design a product collection system which is capable of producing accurate quantitative results by collecting all gas products. 7.3 Process Considerations It has been demonstrated that the reaction pathway proposed in Chapter 1 provides a feasible route from biomass-derived sugars to organic chemicals. The idea of a continuous process beginning with the aqueous- phase hydrolysis of biomass to produce sugars. subsequent aqueous phase sugar reduction to alditols, dehydration of D-glucitol to isosorbide in aque- ous solutions, and finally hydrocracking vaporized solutions of isosorbide to produce deoxygenated organic chemicals is very possible. The use of water as a solvent throughout the process would require only simple separations between process steps. Recommendations for further research in developing this process in- volve integrating the separate processes. The zeolite-catalyzed reactions would have to be optimized for highest conversion to desirable products at a variety of reactant concentrations. The efficiency of the system would al- so have to be studied extensively to determine whether the entire conver- sion process can compete economically with existing processes. APPENDIX A 108 APPENDIX A DESIGN CALCULATIONS FOR VAPOR PHASE REACTOR P ure Ve I I la I n The design pressure of the vapor phase reactor is 1000 psig at a tem- perature of 600 C. Since the reactor wall thickness was not designed but rather chosen for convenience. the following design equation was solved for maximum pressure at a given wall thickness of 0.375 in. t - P 1le [63] SE - 0.6 P P - Internal pressure tm - wall thickness (0.375 in.) SE - allowable metal stress (7.700 @ 600 C) [64] RI - inside radius of tube (0.5 in.) Solving for internal pressure P .gives 6,079 psi. This proves that the ' pressure vessel of the reactor is vel=y capable of performance at the design pressures and temperatures. Wu}. The following equations were taken from the Helicoflex manual for cal- culation of metal seal groove dimensions ' —II_—_ A = 0.732 in. ',.- is “9:22... 2: W22 h . 22.92:? A23; x. . B a O .99 2 ln . .9392“? I ‘33 . 2:132 - .- 591‘s ll C = 0.130 In. V \V a C ::Dm Mes/3(9)) h - C - 0.027 - 0.103 in. Gext - B + 0.022 - 1.014 in. For internal pressures >290 psig. the seal requires radial support on the 0.0. of the groove. The groove can be IOpen on the ID. of the groove. 109 Seal Calcglatlons: LoadI §trsss. and Forge Qalsulatlons The following constants and equations from the Helicoflex manual were used to calculate the loads. stresses. and forces associated with the metal seal. These calculations were used in the flange design found in the following section. Seal Type: H15015-03 Applications: Max. Temp - 600 C Seal l.D.: 0.732 in. Max. Pres. - 11,000 psi Seal O.D.: 0.992 in. Lining: 316L Stainless Steel Section Diameter: 0.130 in. Design Constants: Y2 - 2855' lblin Y1 - 857 lb/in Pu - 7250 psig Dj : seal mean reaction diameter D] - Seal ID + Seal OD - 0.862 in. 2 Fi : Total load required to compress seal to operating point FI a 1: DI Y2 - 7731.5 lbs Ff : Total hydrostaticforce l=f .. 1: P (022/4 = 583.6 lbs Fm : Total load to be maintained on seal In service Fm . the greater of 1c DI Y1 or 1: DI Y2 P/Pu - 2320.8 lbs : Total load to be applied during assembly Fs - Ff + Fm - 2904.4 lbs : Total load applied by bolts during assembly Fb . Fi .- 7731.5 lbs Sb : Resulting stress on bolts F b Sb — = 30.9755 psi N (Ab) N=#ofbolts=6 _ Ab = section area of bolt (1/4 in.) = 0.0416 in2 110 The total load to be applied by bolts during assembly (Fb) was used to calculate the total bolt area needed by the flange. The proof strength of 1/4 In. bolts was found to be 100 kpsi [65] however 50,000 psi was used as a safety factor. Total bolt area needed - Fb I 50,000 = 0.15463 inz Tensile strength area of 1/4 in. bolts - 0.0416 in2 Using 6 bolts. Area 'éyailable . 0.2496 in2 The result is that a flange using six 1/4 in. diameter bolts provides plenty bolt area for safe performance at the design conditions. Flan alcul lon The following equation was used to calculate the required thickness of a flange for operation in the design conditions. The actual thickness of the flange used was 0.5 in. so this calculation serves as a check. t2 splzis + 1)(A' - mm + (s - 1)(A"/U2) - (s + 3)] {661 41sec K12an1 1+3s A' =- __ 4- K12 - 1 2(1.+ s) 2(1 + s)K12 an1 A.. 8 is -1lPbL—thL—Plb.D—PPDP—lhtb— _D I_Eh —. —D h — PPD—PI’IP ‘11 I4“—.— 1 IIIJul—qdd 41 I ‘1—————4d I I._——— OdIdinuII-q b-— qulI Iuttdd—mdndnjn 41- — .d— uddI «I—u thdtq 1 mm“ bmu 9; mm rfi .9. . r m.. ms . .3seez .eep..m "use. mmumm ue-.a “ammmsue use.2 so. u_a mmm a seam mm-.=slmm macacumz .-ma zmam cuts. mm: + 2:. Jam z.=so mauzm =0 Lm=.~mmmmmsa_ 114 «.3 3.3.. .o .528on «an: .«.m 952". ssv 9mm son arm can own new smm 9mm sum ssm com com arm sum smm sem smm smm ._IPPDI—DDPP—PPip—Dbt_bPlP—Plbbp—PDLP—FDPr—bFir—IPFPP—PIPP—PIbID_trtlbD—DDPD ‘ I ‘1 III I._ I. I1— II‘d—1-IIII IId-_Id II._1- III II‘I “IdII II‘ T m.m am ppm mmm TrT I mmm h x 9mm sum cam sad sad as" so— am" s¢u am" emu sad and am am or so am av b Frrt— D D P Pt—I t.’ L D Pt. P 1‘ I II..- ——— GIII IIl—= -‘_11II 0 pm" hum mm hm" T I mm" "a" 1 mam l we. mp Avsou: .¢¢huqm "wood vmubm no-.h .m«s¢vmr use—m Lou u_h mmm a snow mal—afilmm uzomcuu: «in: tuba omlau.mu= + a:— 42m z—zao msuzm In hmz.mmmmm~pc_ an” 800m .0 5330on was: .flum 0.2-2n— aav aam aam at.” aum amm avm amm saw an aam amm aam arm aam aww avm a amm Til—IITIZT ...—...r_ti.r_._..._.=IFFLIFT.......LisLI:_...._.:. _ — - a Iq—l «AAA—i --__._ 41...- 4-2—..- 1. mum ma amm mhm «mm rvm. t mmm l can an aam am“ am" a: a3 am— av— am« am" a: aau am am ah am am av 115 am" r: ma— Avaouz .vvh—«m nuaau aauam non—h “ammmm—mu nae—m sou u—h amm a zoom amt—:almm maaacuu: ulaa Imam amlau mu: + fit- 43m z_:am mauzm.:u mmz.mmaammha_ 116 3a soon. 3 83.8on «was. in 2:2... aov amv. avv amv .amv auv aav amm aam arm amm amm avm amm amm I—II’D-I’II-I ILILIPII_Ir’PbP’II.-IIIIl-P’PI_IIIPP-’I-I—ID’I-IIIIbI’IIIbr'II «Le __ 2. NW 1. . p 3 pm x II I rI amm aum aam amm aam arm aem amm avm amm amm a—m aam am“ am -‘DDP.-rPDD-DD.bI—.£DI—bb’ v-I -bpbb-PDtrb—D.b’bbr.bl_h+bLF-DFIDID-.PleIt-IPPDr-IPDIPL ... I.-— 1 __ A I: 4M... .._ q Au: l 1&4 11 T me . m m m m . w r mam amm ta 1... a3 a: sea . am“ av“ and am—. a: aa— am or am am av $.30...“ ... {if x. . . . .33. . ... 0 Tu If .2..-. K; ._r. . .Iflxfiaar... . . ml. «-.ir...rr$r. . . . . . L = _ _ __ mm _ m 2. P a r a . s :— i am" I P: mad. ma 1 .sseez .va—m unaa— mmuam "on: “2.3.me $.33 .3..— u: pvm a zoom mal :1... am umaacuuc “la: zmam amtau mu: + 2: Jam z-zam mauzu :0 mm... mmaammpa. 117 .2355 .o..==a:.o.o.u>:=<.m.« .o 83.8on new: .mé 2:9“. amv amv avv amv aav a—v aav aam aam arm amm amm avm amm amm PII’I-P’PI- ’I-PIII_IIbI-FDIII-IPP ‘ P P _ P P P I’ — P I, t’.IP — P P Iq I I-Prr’ - PI erIPlI’F _ IPI’ ._ _ ems I aux was I I I. Id ..II I14 I av I a amm aum aam amm aam arm amm amm avm amm amm aum aam am" am— I D D —It Fl .PLPPLI I.Ibpbb.IlP—.IPPD—pbbllb— FILIPPr D-DD ‘ I l _ p D ’ .qi .- — Id I- — IIq — I. a. J- I— — = r q u—. II. n so _ _ _ _ = _ __ _ _ _ _ - T m m l mv « .. mm. i . mam mmm . B. a... 8. am. as. am. ...... a: ...: so 8 E. 8 a... as rain...” 3.2.3.1 "Tr.“ . .2:— . ..m m _ ...-.-L: _J_:_-.-a_. ”—4..-. .-__._...x_..£ “Jeni. ._m T.._. .4. . . . . l mm m p . m w as. 23. mm. ms. t ..eee: .vvhzm nuaau vvnm— no.2. .mmmhmmmu "as... ...... 0:. be" a 5.0m malasclmu zExm amulmm maaacumz «a: gab—22¢: ama>=zcm.m uzm :mzu mmzmmaamuaa— 118 a3 .33. .o 538on owns. .06 832.... due. chm sum 9mm owe one one So cam smm sow arm 5mm 9mm new 5mm smm bPP?}.>P>>ppub>bP>LPP>>rrbPrrpbr.>>—>>P>->>>>p>-Lr-?5».hbkr»->’v>~#tP>P?thb&hru . . .. . .u ‘ . .0 . ‘ . . . w. . - ‘ ‘- s n s; .0. .0. ; ‘ .. ; ~ u.. s . ‘ . s . s.~ .0 s. . ; ITUII'I NJ ‘ 1 . mmm S m mom nfmWIamdIJ331IaHdIAa3WIamdI433WIamdIAfl37IadflIAz:ngzflllumllanIIQQFIAHWl pL$$rbyt>pbbbltp>rb?-++rrb-b‘tbbPhp>-P>P>-P.L?p>?>rppbb’bhb}bnbpu>-L>>’-’»»Pr»b> 1. _ J “—4 #111: 111: :1 .= -.-—1 : .—_ -. ‘ :: :44: z.— ._44 1 0.5.. i. 1 :‘ .._:: J‘._.1_‘1 m a p _ E m _ p: m... m. llj'IT‘ Avaoux .vvb—«m unss— smumm uo-mh .samammmu nae—m no“ u—h own a :num man—:Hnmm nuance“: «In: zmau 59.350: + 3: 42m z-nao mauzm :u mmzdmmmmmhe 119 «on :2. .o 8.38on «on: .né 2:9“. .aq¢ awe sec emu. sue .auv 99¢ amm smm sum com smm gem smm smm _-’-pr-[-—--r-—L>>>.>P>LhL>L’—--->P>[’_}»»’_E-r_srb>_;£r»P>»--—--»L _ . :_ ‘._J J ‘ mom w mc w 2.3 cox Tr own can can amm emu chm com 8mm scm smm 9mm sum cam 32 as _.. p . .x. .-0._»_ . .xb »;_..,. . »‘_ . .-“‘_‘_ .\b .u__ a.. . .‘_ n_”.._. .yb I.m. . . . a." p ”m. p _ 0 . _.‘_ p p‘¢_w “ p‘» v r mum _ ¢u~ _ ; _ _ 1 a m w "a" r mmm hum mom a son a: 03 am— a: emu smq a: as am am 2. am am 3‘ h.pb-PrDr-bPPLFP’Pb-Pp’ub—‘Drblhl- LPIP-IDPI-DP’LPDIbFD_PLPPPLler_ 41.10: W a 4 2 41 .= J-_1. : ._ —-. J ——44 J: .1 4 Wm . r’-bP’L-’DD’-’>F’-”"-)F’r_>b”-’II’I—i”’—””_L*”bLDPi-’b‘I r’bP-D‘t-F'}’_’F>+_ ‘ J— I. -1: — «q _ e m_ x I mam arm can amm scm smm sum sum mam 9mm sum arm mom amm gum 8mm smm sum 7.3—m on . ._ ... . _ . In T {Ln—9v.“ F vi.“ . . r__. :4 C.” rxi . pvm _m~ ppm ; w p m m a N I ham mmm U j [I .I bcm 3m com 03 do" a: so" am" a: emu am“ a: as" so so 2. so am 2.. m3 pm" 2: mm b: m3 Aaaoaz .v:.««m "use“ Nuumm "01:. .vvuummmq ”yo—m ham Q; @SV a :mom moi—:Humm mzamcumz Tm: zmbm swusufiu: + 3: .5m 2330 msuzm :u mmzdmmmmmg— 121 owe ago... .o 53.3%.. was: .96 2:2“. smm 3m amm omm 3m sum saw 93 a? 92. smv a: amv smc a; a3 smm smm -’£}--?”.l’llx’—'b’u’-i "! pPr-F”’—r"”—r”pl’_b"lrbx-rr”-”IIL-”"-P-”—"F’b”’--’t’ d . a q ‘ql 11 4 mLm ; d.” w ¢ - wmm mm x .. amm ohm mom smm gum smm 9mm sum mam smm amm arm sum smm 2% 9mm 0mm 3m PI‘?D-IPPb-PbbP-D-‘D|P-Dbb Lib Ibbb— D‘b+-bibbbpbh‘IP.IbPLL-LDbI—#bPPPbbDP—bbb Pbtbb—‘b b+ ‘ d 4 d u -— d. — u- q: 1- d- 1m 44 q q <-—. Al I mmm d m _ .LN pp m m d w ram rum I am new 2.; 2:. a: S. a: a: 2: ea c: 3: so 8 2. 8 am 3 mp :— 2; mg mp Aasonz .evpudm "use" mmumm u.._a .mmv¢mp- ”#0“; so“ u_a sm¢ a caom mm-_=n-mm mzamcuuz «an: zmam smug—.mu: + 2: .5m ...—zoo mauzw :u mmzdmmmmmro— 122 3:03.06 3 828on «was. 67m 2:2“. 03 09 of can 02 ofi 02 ’ + ’ bL ’ b P b ’ ’ b ’ t b ’ - ’ ’ h b [byl’ ’ p h r ’ ’ ‘ L- ’ ’ A d 4 a A .4 1 a T q. a . .. m __ . a _ u l . m “ g u . 8. a . m s . u _ a .9. 8. .9 8.. . ....“ . c a 9:” 52w x 93 5:8. "98 u w . _ . . . _ . _ h . . _ u t . .. J r . «lo , x x . o: u 1 .3 3. ..R 02 ..9 09 3a 8.. can a on a + b } } bL D > sh b b b «r L< b4 ? b i b n b h 0 b L b b b p b ’ b h L r) b O 9898 1.) 123 ,2. soon. .0 528on «no: .Ié 952". sa_ mm mm mm so. mp my me so mm am mw a¢ -b ppL-pp—ppp- pupp—Pnpp whipp—-n-—~h-. -.—-p-——-——b—-bp _ _ _ __ _ ___ . mm . m¢ mm 1 T mm Avsquu .mmmmm "use“ asum nonma .m~m_m "ya.“ go“ u_a pp a seam mm-=oa-md E: 2.: .5m hung—mm zozcacmuun zczNoa uzw zuzu mmzémmmng 124 2.203 «2.3.24. 3 £38on «was. .36 952“. P D 02 o: 00' our 0: 8' on on as on on 3 on on o— r f r p h b r P P|t P . 14 . ‘_ a: _. =—_ :4 .. , o o: \ ..Efiofium .ococfismagflm :33 556 mm 125 mm u—flGl .0 53.3001” ”no: .Mrum 9.30.".— 92 we: so: mm so we so mp E. we so mm am m¢ av n+b r— r#- b pbrb—pp-p—pnp——bLIP —pb|P—B- —nb-F-|Pbp—npb|P—bbpp— -— _b— 4‘ ‘ i‘ \ N ‘ ——‘_ _ ___ -4 do mm 8 r 3.3L.“ .ommom— "was: film "on; .3..me nae—m Lou 0:. mm a coom wanes—Fm: «a: 3: .5m .8255: zozcncmumn zczNon uzm :mzu mmzémmmng 126 om 800n— uO Eahuo¢am mam: .Q—um Guam-m a: we as: mm mm mm mm mp 2. we so mm mm mv aw P mp-Lp—u"-m_mnb“ --— bmlm J—p.- P-pb.w--1~m-—--T“V—r mg—m_nn bin 1 r mm I B r Avaomu .svmmm "use: mmum uo-mh .musmmm nan—m so“ u_h mm a :mom mmucowum: «me at: 42m huaaoxm zo:hc=¢¢umn zczNon uzm :wzu mm=.¢mmmm—«s_ 127 p: .30.. .0 53826 «no: 67m 2:2". ufl m: a: ma 2: mm 8 mm 8 pp-p—pt- P-pPFE—rrpbrp ‘ ‘ ‘ ppP——uppp—Frpn—mIPPn—pppp—m mp 9p we so m cm mv s¢ _ _ _ 9: mm 1. n—F-pn——-pP—r-PP—Ppp-—pbpp d. 411 uq—-—— ‘ .¢¢p:«m "was: mmup "on; .mmcpms— Awnamu nae—m sou u_p «a: a zoom mmucuasm: «m: 3: .5m .3292; zozcncmumn zccha uzm zuzu mmzémmmng 128 Q: saun— uO Eahaomnm an“: .070 9533—“— smd m: a: ma: as: mm so we so mp sp me so mm mm m¢ s¢ pppP pp. Pbb :P- pup. rub: p... PP ppnp—DLPPFPP-hrnbrr—Cup—P bLbh-p ppr _ q _ _q__ ‘ __ _: I we «m mm mm I . . 33¢ mdpmo "use“ mm.~ no._h .emmsmm ”vo_m no“ u_~ m__ a zoom mm-=~a-m_ «m: 3: .5m puzaozm 22.235.qu zctha uzm zmzu mmzémmmmze 129 on. .23.. .0 533on mass. .570 2:2". mm: m: a: we: as: mm mm mm em mp Sp we mm mm mm mv EV h...p ppb—ppbPFpp-rrthr_ .bbP—PF 1“ 1 ‘ a: PbF—Pb+-_--— P—brnL pp-P ..i 1. i 4 “m" «m 33mm .mmmmvm "use: puum ”on: .msmppsu "we; .3: u: mm: a zoom amuse—(m— am: 3: .5m pug—SE 22.—.25:qu zczNoa uzw zmzu mmzémmmmza 130 .325“. 35226 2.3 2.23. 3522 3.5“; .o 838on «was. .mpé 952“. 3‘19; ofi o a P-P Nwo ”ON Nov now now 00— :.omn.<<. ~o m: mo mumoo.o~: 4<¢3um:u.~.4pzpmz.m 3.9;. B“. oorbmzo 8“ . ON .ov ow om . 400: n.m¢n.<< No m1 mo mumno.odu wZOpwx 4>Ipuz 4pm:m.~ 131 m3 scan. .0 523on mass. .36 2:2“. mm: m: a: ma: so: mm mm mm mm mp sp mm mm mm mm m¢ aw —p-Ph—Pp - ...—pbPLF-p r-be—brp- .pb—ppnp—p-p—— ppb—b-pb— p—ppL_ pr-F-pup . __ _ ___ ____ _ _____fi mm , S .. a: m ex r mm . «tam: .as¢mm "ass" .a"«« u.._s .mmmo_m u«o_m so“ u_a we" a cuom mm-=.a-m_ Sn 3: .5m pug—SE zapcncmuua zczNoa uzm guru mmzémmmmze 132 new .32. .0 53.026 322 .cué 2:2“. mm: m: a: mm: as: mm mm mm mm mp op mm mm mm mm mv ac _-Pb_-b- p-pn—pP-pr— up bkbn—h- p—_Pw+—phbn_bpph—r pler—pp—p—P-hn—m-pprhppb—rppp _____ _— . .———_ . pm 1 , T pm 1 ._ mmx I mm I 33% .msmsom "use: «Sm: “on: .m—muvm "3o: .3..— u: mam a zoom manna—(m «m: 2: .5m .8259: zo_.—.c=cmum= zczNoa uzu :uxu mmzémmmng « 133 Eu 33.. .0 5283.0. 322 .pué 9:3.“— am— mm: am: m: a: ma: aa: ma am ma aa mp ap mm am mm am mv av FE pPPthP—pph pPhl—thb— pn-pnPn—Fn P-p inb—pPPrF :- —pPPhrL j? berwchp—nFPh I ‘ 1m._33:6.axoaazé .o Enhoonm «ma: .«Né 2:2“. bbbbD>D$DDLPDPDDDPDD bib-DDDDDDD. om" .ov" o d b . FL D'D _ ooa om ov. ON munn-o< no m: mu o~_no.m~_ nzxomo>1 .. n . Jppwo< .. Nu 40:23.53 . 134 3' 0: 3' one oprp on»: on on Oh on on 9. on a 3 h b r b r # . P h —- u «—- 14 88838" muoflampm o _ .2 / ox ..vmfioaumlmxohvzsnv decimaqfimmnmm «Ciao . on: 135 :20000 no E=h~00nm was: .ON.m 0.36-u. an; .2: 2.: am. a: 2: so so 2. 8 8 8 am— .vvpuum nuaau vauva "camp .vamaamm nae—m Lou u_p mp: a coom aas>ozuau maaacumz : an at: mam puaaaaa uaa azau zatNaa atm zmzu mm:.vmaaau—«_ 136 5.9.3... .0 5:33am «was. .vué 2:2... av: mm: and mm: am: m: a: ma: aau ma am ma aa mp ap mm am mm am mv av mm: va: «a .vvam unaa— vvuam ”on: .mpampmm "33m .3.— u: mpm a ...5m aan>oznau uaaaaum: : aa 3: .SN puaaaaa "Ea azau zazaaa uzm guru mmzvmaaazz 137 apd am: am:. ppm .39. .o 8382....“ own—z .mu.m 959“. a 3 am: a: aa.— aa aa ap am am av . ampmm unaau Nana: non: .amammm "we; .3“ 0:. ppm a zoom maugozsm uaaacmu: «a: 2: Jam zctNaa avaua.;cu mzm aura mmzmmaaaama: 138 no" xaom .o Enzomnm awn: .oué 2:9“. am: am: 5 am." am— : aa: aa aa ap am am av am 3: mm .vvam unaau mmnau "on: .maaammm nae—m ...; ...:p mam .— zoom mango—Tm uaaacuuz «a: 3: .5m zczaaa avmua.;cu uzu :mzu mmzmmaaaama: 139 am: 99“ ann— no 5.9—coma“ ”no: .pNum G530:— am: av: am: am— 3: aa: aa :— .vvp:m unaau aauam "on: .avmmamm waaaaumz 3: 2: Jam zczaaa a aa ap am am av am mm 2.3.: .3“ 0:. mam a scam mango—Tm vmaa.;¢u uzm nmzu amp-.mmmaaama— 140 am: pan zucn— «O Enhaomnm ”an: .QNum 0530-"— am: av: and am: a: aau am aa ap am am av am mm: pm .vvpzm nuaa— mmuam "on: .vvmvupp n+3.— .Bu a: pan 3 zoom mango—Tm maaacum: «a: 3: .5m zazNaa avmua.;cu uzm zuau mmzdmmaaama: 141 am: "PM ¥G¢m «O Sabaooam mam: .mflum 0530—“— am: av: am: am: a:: aa: aa aa ap am am av am mm: _ pm mm .vvp::m unaa: mmuam "our... .ammammp nab—n: .3: a: m:m a cuum mannozna maaaauuz :aa 2: .5m zazaaa avmua.:>cu atm zmzu mmzmmaaaama: 142 man 32. 3 $3325 amas— .on...m 8.3.“. am: am: av: am: am: a:: aa: aa aa ap am am av am mm: .vvp::m unaa: mafia "on: .aaammav nae—m .3: a: mam : 53m mango—Tm maaacum: :aa 3: :am zazNaa avmna.:>au mzu :mzu mmzmmmaaama— 143 ap: «an 39.. 3 52325 «3.2 .53 2.3.“. am: am: av: am: am: a:: aa: aa aa ap am am av am mm: :a: pm mm .vvp::m nuaa: m:umm uol:p .vapamma nae—m Lou u—p mam a caum massazum maaaaum: :aa at: :am zczNaa avmua.:>cu uzm zmzu mm:.mmmamama: LIST OF REFERENCES [1] [2] [3] [41 [5] [6] [7] [8] [9] [10] [11] [12] [13] [14] [15] [16] [171 [13] [19] 144 LIST OF REFERENCES Hockwett, FLC. et. al.. J. Am. Chem. Soc.. 68. 930. (1946). Wiggins, L.F., Adv. Carbohydr. Chem.. 5, 191, (1950). Barker, R.. J. Org. Chem.. 35, 461. (1970). Beck. K., Pederson, 0., Thorgersen. H.. Acta Chemica Scandinavica. B 35. 441, (1981). Fleche. G. and Huchette. L.. Starch. 38, 26. (1986). Anet, E.F.L.J., Chemistry and Industry, 262, (1962). Mednick, M.L.. J. Org. Chem, 27, 398. (1962). Feather. M.S. and Harris. J.F.. Carbohydr. Res.. 15. 304, (1970). Turch, G.R. and Tatani, J.M., Carbohydr. Res., 15, 233. (1970. Gardner, JD. and Jones. R.F., United States Patent 3,483.338(1980). Shaw, P.E.. Tatum J.H.. Berry, R.E., Carbohydr. Res., 5, 266, (1967). Ibid., Carbohydr. Res., 16. 207, (1971). Popotf. T. and Theander, 0., Acta Chemica Scandinavica, B 30, 347, (1976). Feather, M.S. and Harris, J.F., Adv. Carbohydr. Chem., 28. 165, (1980). Kudlac. G.A.. M.S. Dissertation. Michigan State University, East Lan- sing, MI (1988). Chang. 0.0. and Silvestri, A.J., J. Catal.. 47. 249. (1977). Rudham. R. et. al., eds., Catalysis by Zeolites. Elsevier Scientific Pub- lishing Company, 113. (1980). Minachev, K.M.. Garanin, V.I., and Isakov. Y.l., Izv. Akad. Nauk. SSSR Ser. Khim., 1635. (1964). Jacobs. P.A., Carboniogenic Activity of Zeolites, Elsevier Scientific Publishing Company. 101, (1977). [20] [21] [22] [23] [24] [25] [25] [27] [23] [29] [30] [31] [32] [33] [34] [35] [35] [37] [38] [39] [401 145 Meisel, S.L., McCullough, J.P., Lechthaler, C.H., Weisz, P.B., CHEMTECH, February, 86, (1976). Rabo. J.A., Schomaker, V., and Pickert, P.E., Proc. 3rd Int. Congress on Catalysis 1964, North Holland Publishing Company, 2, 1264, (1965) Kubo, T., Arai, H., Tominaga, H., and Kunugi, T., Bull. Chem. Soc. of Japan, 45, 607, (1972). Rabo. J.A., ed., ACS Monograph #171, American Chemical Society, Washington 0.0., 714, (1976). Ibid., 522, (1976). l Chen, N.Y. and Degnan, T.F.. Chem. Eng. Prog., February, 32,(1988). lmelik. B. et. al., eds., Catalysis by Zeolites, Elsevier Publishing Com- 1 pany, Amsterdam, 279, (1980). f Chen, N.Y. and Koenig, L.R., United States Patent 4,503,278 (1985). ”F. .. Weisz, P.B., Haag, W.O., Rodewald, P.G.. Science, 206, 57, (1979). Oudejans, J.C., van den Oosterkamp, P,F., van Bekkum, H., Appl. Catal., 9. 317, (1983). Chen, N.Y., CHEMTECH, August, 488, (1983). Chantal, P.D., Kaliguine, S., Grandmaison, J.L., Appl. Catal., 10, 317, (1984) Frankiewicz, J.C., Occidental Petroleum Company, Irvine, CA., (1984). Soltes, E.J., Milne, T.A., eds., ACS Symposium Series #376, American Chemical Society, Washington 0.0., 328, (1988). Chantal, P.D., Kaliguine, S., Grandmaison, J.L., Appl. Catal., 18, 133, (1985). Soltes, E.J.,Milne, T.A., eds., ACS Symposium Series #376, American Chemical Society, Washington 0.0., 311, (1988). Ibid., 264, (1988). Haniff, M.l., Dao, L.H., Appl. Catal., 39, 83, (1988). Arena, B.J., United States Patent 4,496,780 (1983). Jacobs, P.A., et. al., eds., Structure and Reactivity of Modified Zeo- lites. Elsevier Scientific Publishing Company, Amsterdam, 279, (1984). Boyles, D.T., Bio-Energy Technology, Thermodynamics, and Costs, Ellis Howard Limited, 46, (1984). [41] [42] [43] [44] [45] [461 [47] [43] [49] [50] [51] [52] [53] [54] [55) [56] [57] [58] [59] [60] [61] 146 Arena, B.J., United States Patent 4,380,679 (1983). Arena, B.J., United States Patent 4,380,680 (1983). Arena, B.J., United States Patent 4,382,150 (1983). Arena, B.J., United States Patent 4,413,152 (1983). Kerr, G.T., Catal. Rev. Sci. Eng., 23(182), 281, (1981). Petersson, G., Tetrahedron, 25, 4437, (1969). Porter, O.N., Mass Spectrometry of Heterocyclic Compounds, John Wiley and Sons, New York, 153, (1985). Ibid., 203. (1985). Ibid., 262, (1985). Treptau, M.H. and Miller, D.J., Ind. Eng, Chem. Res., 26, No. 10, 2007, (1987). Pierce Chemical Company Catalog, 186, (1987). Porter, Q.N., Mass Spectrometry of Heterocyclic Compounds, John Wiley and Sons, New York, 146, (1985). EPA/NIH Mass Spectral Data Base, Vol. 1, US. Department of Com- merce, 69, (1978). Leonard, R.H., Ind. Eng. Chem., 48, 1331, (1956). Porter, O.N., Mass Spectrometry of Heterocyclic Compounds, John Wiley and Sons, New York, 153, (1985). Registry of Mass Spectral Data, Vol. 1, John Wiley and Sons, New York, 128, (1974). Ibid., 128, (1974). Porter, Q.N., Mass Spectrometry of Heterocyclic Compounds, John Wiley and Sons, New York, 267, (1985). Ibid., 266, (1985). Krishna, H., Kalluny, M.H., Ambidge, C. Tidwell, T.T., Carbohy. Res., 158,253,(1986L lmelik. B. et.a|., eds., Catalysis by Zeolites, Elsevier Scientific Publish- ing Company, Amsterdam, 93, (1980). u '_l ”HF." -.'l"_— PL. [62] [63] [64] [65] [66] [67] 147 Occelli, M.L., eds., ACS Symposium Series #375, American Chemical Society, Washington 0.0., 34, (1988). Perry, Fl.l-l., Green, D.W., Maloney, J.O., Perry's Chemical Engineering Handbook Sixth Edition, McGraw-Hill, 6-73, (1984). Ibid., 6-80. (1984). Shigley. J.E., Mechanical Engineering Design, 3rd Edition, McGraw-Hill New York, Chapter 6, (1971). Waters, EC. and Taylor, J.H., Mech. Eng., 49, 531, (1927). Perry, R.H., Green, D.W., Maloney, J.O., Perry's Chemical Engineering Handbook Sixth Edition, McGraw-Hill, 23-44, (1984).