THE UNIVERSITY OF MICHIGAN INDUSTRY PROGRAM OF THE COLLEGE OF ENGINEERING HYDRATION OF PROPYLENE WITH A CATION EXCHANGE RESIN CATALYST John Harold Hiestand A dissertation submitted in partial fulfillment of the requirements for the degree of Doctor of Philosophy in The University of Michigan Department of Chemical Engineering 1961 May, 1961 IP-515

Doctoral Committee: Associate Professor Kenneth F. Gordon, Chairman Associate Professor Wilbur Co Bigelow Professor Giuseppe Parravano Assistant Professor Dale F. Rudd Associate Professor Milton Tamres

ACKNOWLEDGMENTS The author wishes to express his appreciation to the members of his committee for their discussion and comments on the various aspects of this work, and particularly to Professor K. F. Gordon for his guidance during the completion of the experimental work and to Professor R. R. White for his guidance and encouragement throughout the beginning of this investigation; to the shop and office staff who gave freely of their time and aid; and to the many graduate students who contributed much through their stimulating discussions of the study. Materials supplied by the Phillips Petroleum Company and Dow Chemical Company are gratefully acknowledged. Financial assistance in the form of fellowships and grants from the Socony-Mobil Oil Company, Gulf Research and Development Company, E. Io du Pont Company, and Dow Chemical Company is much appreciated. Finally, the author wishes to thank his family: his wife, Carol, and his parents for their inspiration and assistance. ii

TABLE OF CONTENTS Page ACKNOWLEDGEMENTS.......................... ii ABSTRACT................v..................................... iv LIST OF TABLES................................O........... o vi LIST OF FIGURES............................................ ii INTRODUCTION.................................................. 1 THEORY......... *......................................... 5 EQUIPMENT.................................................... 10 MATERIALS................................................... 17 DATA AND RESULTS.......................................... 20 CONCLUSIONS........................................ 52 APPENDICES A Derivation of Rate Equation........................... 53 B Derivation of Modified Rate Equation.............o.. 54 C Sample Calculations.................................. 61 D Experimental Procedure............................o 65 E Chemical Analysis of Samples......................... 67 F Estimate of Error..**o*@o***.Osee 70 F Estimate of Error.................................. 70 G Heat of Reaction..................................... 73 H Correlation of Data................................... 75 NOMENCLATURE..................................... 81 BIBLIOGRAPHY............................................ 85 iii

ABSTRACT The hydration of propylene to isopropyl alcohol in the presence of an acid cation exchange resin catalyst was studied in a flow reactor under steady-state conditions. The experimental variables investigated were: Temperature: 100~ to 160~C Pressure: 450 psig and 1440 psig Resin size: 32-42 mesh and 100-150 mesh Hydrogen ion concentration: 0.462-1.678 meq./ml Feed composition: 0-60% (wt. % isopropyl alcohol) Incremental reaction rates were measured as a function of temperature, pressure, resin size, hydrogen-ion concentration, and feed composition. The reaction rate, correlated as a function of temperature and concentration in the liquid on the propylene-free basis, is: r = exp(-21,600/RT){(-37.94 + 0.668t - 0.0026t2) 109 + (2.059 - 0.06517t) lolx + (1.952 + 0.00937t) 1 o1x2} gm.-moles/eq,-min. @ 450 psig and r = exp(-21,600/RT){C.68 - 0.005t) 100 - (4.32 - 0.023t) 101x + (20.98 - 0.122t)- 1011x} gm.-moles/eq.-min. @ 1440 psig where R = 1.987 cal/gm.-mole-~K, T = ~K, t = ~C, and the mole fraction of isopropyl alcohol in the liquid (propylene-free), x = 0 to 0.12. This yields an apparent activation energy of 21,600 cal/gm.-mole for the 8% crosslinked, sulfonated polystyrene resin catalyst used, Dowex 50WX8. Resin size had no apparent effect on the reaction rate, indicating that diffusion is not important in this system under the conditions investigated, iv

The rate per unit of catalyst volume is- not'a.-Alnear function of hydrogen-ion concentration. Initial rates are correlated by: r. = {12.952[H+] + 4.002[H+]2 x 10-3 gm-moles/liter-min. where [H+] = meqo/mlo of saturated resin. Because the formation of di-isopropyl ether becomes increasingly noticeable above 140~C, only a few exploratory runs were made in this area, and no attempt was made to correlate the datao The resin half-life is 155 days at 1300C, found by extrapolating data at 200~C and 170~C on a log half-life versus 1/T plot. v

LIST OF TABLES Table Page I Hydration Catalysts....... o....................... e 3 II Ion Exchanger Catalyzed Reactions.................. 4 III Heater Design Data............................ 14 IV Incremental Reactor Data, Group I.,......... 2... 23 J Incremental Reactor Data, Group II o....o........o, 25 VI Incremental Reactor Data, Group III................ 27 VII Incremental Reactor Data, Group IV............... 29 VIII Incremental Reactor Data, Group V................. 31 IX Incremental Reactor Data, Group VI............... 33 X Incremental Reactor Data, Group VII............ 35 XI Summary of Correlations................ o.. o.. 37 XII Integral Reactor Data, Group VIII.............o.. 40 XIII Integral Reactor Data, Group IX................ 40 XIV Integral Reactor Data, Group X.................. 41 XV Integral Reactor Data, Group XI,............,...... 41 XVI Resin Half-Life Data.,.......................... 48 XVII Hydrogen-Ion Data....,.........................o 48 vi

LIST OF FIGURES Figure Page 1 Apparatus......................................... 11 2 Heater Details.................................... 15 3 Wiring Diagram.....................................o 16 4 Group I Data...................................... 24 5 Group II Data................................. 26 6 Group III Data................................... 28 7 Group IV Data...................................... 30 8 Group V Data.............................. 32 9 Group VI Data...................................... 34 10 Group VII Data..................................... 36 11 Summary of Correlations.......................... 38 12 Group VIII Data................................. 42 13 Group IX Data.................................. 43 14 Group X Data.............................. 44 15 Group XI Data..................................... 45 16 Log ri vs. 1/T O................................ 46 17 Resin Half-Life................................... 49 18 Volumetric Rate vs. [H]............................. 51 19 Ternary Liquid Phase Diagram @ 25~C................ 56 20 Estimated Binary Phase Diagram @ 500 psi........... 57 21 Estimated Ternary Liquid-Vapor Phase Diagram @ 14o~C, 500 psi............................. 58 @ l4o0c, 500 psi..58 22 Graphical Differentiation....................... 64 23 Example of Sample Analysis........................ 69 24 Correlating Constants, 450 psig................. 78 25 Correlating Constants, 1440 psig.................. 79 26 rexp vs rcalc..-........9.................... 80 vii

INTRODUCTION Since the first commercial production of isopropyl alcohol, about 1920, it has been widely used as a solvent, a dehydrating agent, a disinfectant, and a raw material for the manufacture of acetone and other compounds. Production in 1960 was 1,250,000,000 pounds with a (15) value of $88,360,000. ( Petroleum refining companies are currently interested in a relatively simple and inexpensive method of converting by-product propylene to isopropyl alcohol for use as an anti-icing additive for gasoline. The most obvious route for making isopropyl alcohol is the catalytic hydration of an olefin in the vapor phase or the use of a mineral acid or the acid form of a cation exchange resin as the catalyst in the liquid phase. General Background of Hydrolysis and Hydration The hydrolysis of olefins to alcohols has been carried out with a number of catalysts. The original commercial process(l) is a two-step operation. The olefin is first contacted with concentrated sulfuric acid in a counter-current absorber to form the alkyl sulfuric acid. This intermediate product is passed-through a hydrolyzer and reacted with water forming the alcohol and sulfuric acid which are separated by distillation. Recently many alternate processes have been investigated. Some make use of direct hydration on a solid catalyst while others involve direct hydration with strong mineral acidAs or their salts. -1

-2Most reactions in the solid catalyst group take place at high temperature and high pressure in the vapor phase. The mineral acid or salt reactions occur in the liquid phase. The use of the acid form of cation exchangers as hydration catalysts has also become of interest. Table I lists some of the hydrolysis and hydration catalysts investigated with their references. McDonald and Hamner(27 28,29) discuss the latest developments in the field. General Background of Ion Exchanger Catalysis Like mineral acids, the acid form of cation exchanger catalysts are highly active at relatively low temperatures and. pressures because of the high hydrogen ion concentrations. However, the advantage of direct reaction on a solid catalyst, namely the easy separation of catalyst from product, is enjoyed. Ion exchangers have been used as catalysts for many types of organic reactions. Table II lists reactions which are catalyzed by either cation or anion exchangers or in some cases both, presumably through different mechanisms. The reaction under investigation is catalyzed by the acid form of a cation exchanger, Dowex 50, a sulfonated polystyrene crosslinked with divinylbenzene. This resin, described by Bauman and Eichhorn(5) has been used as a catalyst in numerous investigations of organic reactions!3554'18,1125) The newer form of the resin, Dowex 50W, sulfonated. under controlled conditions, was used in this study.

-3TABLE I HYDRATION CATALYSTS Catalyst Author (14) (26) Mineral Acids Ipatieff & Mo e ), Majewski & M ek Runge, et a1 4 Stanley, et al4 Sherwood.36). Ion-exchangers Chambers(8) Douglas(11) Keith(16), Kreps & Nachod.(19, Langer(21), Young(44) Inorganic Oxides Levy & Greenhalgh(23), Lukasiewicz, et al(24) (alumina, silica, zir- Muller & Waterman(30), Reynolds & Pittwell(325 conia, thoria, titania, Robinson(33), Runge. et al(34) tungstic, and ferric) Activated Copper Cottle & Young(9) Permanganates, Alumi- Teter, Gring, and Hettinger(42) nates and Silicates Silicaphosphoric Acid Wegner(43) Chrysocolla Smith(39) Halogenated Poly- Friedman & Morritz (2) carboxylic Acid Organic Nitrogen Bases Bent & Wik(6)

-4TABLE II ION-EXCHANGER CATALYZED REACTIONS Reaction Type Author Iydaration-dehydration Hamilton & get.ner(13), Tessman(20) Reed, et al(31 ) Sussman41 ) Alcoholysis Sussman(41 Esterification Kressman(20), Sussman(41) Sucrose Inversion Kressman(20), Sussman(41) Acetal Synthesis Kressman() Sussman( 41) Alkylation Kelly(17) Condensation Astle & Pinns(2), Kressman(20)

-5THEORY Hydration Reaction The hydration of propylene to form isopropyl alcohol was investigated using a cation exchange resin as catalyst. cation exchange OH H2C -C- CH3 + H20 < H2C - CCE3 H " resin H The reactants form a vapor-liquid mixture since the temperature required to obtain useful rates is above the critical of propylene. Vapor-phase chromatography was used to determine the products'by comparison of chromatograms with those of standard samples of expected products, isopropyl alcohol, di-isopropyl ether, n-propyl alcohol, and acetone. The only compounds found in the product were isopropyl alcohol and di-isopropyl ether. The rate of formation of di-isopropyl ether is much slower than that of isopropyl alcohol; therefore, this reaction does not consume an appreciable amount of reactants below 160~C. cation exchange H H 212C C - CH5 + H20 <3c 3 c o - C - CH H resin 3 3 In runs at atmospheric pressure and 150~ C where the reaction mixture is entirely vapor phase, no reaction was apparent with the cation exchange resin catalyst, so the vapor-phase rate must be negligible in comparison with the catalyzed liquid.phase rate. Marberry(25) found that the rate of cumene hydroperoxide decomposition catalyzed by Dowex 50 resin was dependent upon the water concentration in his reaction mixture.

-6Perhaps liquid water must be present for a reaction to take place in the resin phase. Liquid water causes an expansion of the resin allowing molecules to diffuse into the resin more freely, and it causes the functional groups to become more highly ionized favoring the carbonium ion reaction. Since the vapor-phase rate is negligible, the reaction may be considered to take place between the propylene dissolved in the water and the water itself. Inside the resin phase, this reaction probably takes place much the same as the carboniuma ion reaction in aqueous acid. The resin-liquid-vapor reaction system is much too complex to determine the mechanism from overall rate data alone. However, the carbonium ion mechanism(22) is chosen as a possible mechanism and a rate equation is developed which justifies the correlation of the rate data.. k HC C - C + H (C - C- ) (1) H k H k OH (HICC +H C C )+ + H02~ -H3C H cC.. C. C + (2) Tk4he equation r The equation for the rate of formation of isopropyl alcohol is derived in Appendix A. d[i-C3H70H] klk3[H [[i-C H70B] d.t k 2k [H {[C3] [H 20] 3 — 5}161 (3) dt k k2 5[H 01 2 K The reaction is exothermnic. The heat of reaction, calculated in Appendix GI is -10.74 kcal/gm.-mole corresponding to a heat generation of 2.69 - 107 cal/min. or 0.187 - 7.43 watts for rates measured in the

-7range 100~ - 140~Co A typical power input to the reactor heaters is 130 watts. It is apparent that the heat of reaction is negligible compared to the heat required to maintain the reactor at operating temperature, Nature of the Catalyst Ion exchange materials have functional groups, capable of ionizing, attached to a non-ionizing skeleton which is insoluble in the exchange medium, Some occur naturally, such as silicates, micas, and feldspars, but the synthetic resins are more specific in their action since they contain only one functional group. Cation exchange resins may contain carboxylic, phenolic, sulphonic, or phosphonic groups. Resins containing sulphonic or phosphonic groups are the most useful for acid catalyzed reactions since both are strong acid groupso Dowex 50 is a polystyrene, cross-linked with divinylbenzene and sulfonated with H2S04. A newer type, Dowex 50W, is sulfonated under controlled conditions yielding a product which has a lighter color and a greater mechanical strength due to less charring of the organic material and smaller internal stresses, Dowex 50W was chosen as the catalyst for this reaction because other investigators had difficulty with Dowex 50 disintegrating at lower temperatures than those required for this reaction. In the Dowex 50 resin, the sulphonic acid groups are attached to the benzene rings with the gel structure shown on the following page. The amount of cross-linking determines the absorptive and swelling capacities of the resin. Highly cross-linked resins have low absorptive and swelling capacitieso As a result, the rate of absorption

8- CH CH - CK2 - CH - CH2 - CH - C 2 - CH.SO3H SO3H 1SO3H - CH, - CH - CH2 - CH - CH2 - CH- CHIL2 CH - 1^1 rli"r 2 bVL>SO H SO H SORH 5 5 5 An of large molecules is limiteda by a highly cross-linked. resin because the resin structure offers an obstruction to the migration of these molecules. Converselyg small molecules are little affected by cross-linking. An 8% cross-linked resin was chosen because it has the optimum cross-linking for most systems; it neither swells excessively nor offers too much diffusional resistance. For the sake of simplicity, most authors assume that the gel phase has a continuous "liquid" structure. This phase is in equilibrium with the external liquid phase; the concentrations,of the gel phase are determined. by the external concentrations and. the phase distribution coefficients of each of the components, Ion exchange resins selectively absorb materials from a multicomponent mixture. This can be described by the phase distribution coefficients which are the ratios of internal to external concentration of each component. In the treatment of ion exchange catalysis data phase distribution coefficients receive various degrees of attention, Some authors ignore them; others use them but make simplifying assumptions such as their independence of temperature and concentration. It is

-9now recognized that distribution coefficients are important and are functions of concentration and temperature. Correlating Equation In Appendix B. the rate equation is developed from the theory, justifying the form of the equation used for correlating the rate data, Although the data are correlated by the equation, this does not prove that the assumed mechanism is correct. Since the reaction system is complex and the phase diagram for the gel-liquid-vapor system of the three-component mixture is not available, it is necessary to simplify the derivation of the rate expression to arrive at an equation which may be written as a function of the liquid-phase mole fraction of isopropyl alcohol on the propylenefree basis. d[i-C5HOR] kik [H] l t3 70H] k3 {A+Bx+Cx2+Dx3+Ex4+Fx5} (4) dt G The constants, A through G, are defined in Appendix Bo Although G is a function of x, it is assumed to be constant over the range of x investigated. A more exact rate equation would be a polynomial of higher degree, substituting a power series of x for G hb.wenrer. rate data would scarcely warrant being fitted by an equation containing more than two or three constants while Equation (4) has eight constants after combining k1 and k3. The correlations are made with three-constant equations.

EQUIPMENT Flow Description Much of the equipment used was designed for the dehydration (37) of n-butyl alcohol. 7) Details of the original equipment may be found in the dissertations of Sliepcevich(38), Dale(lO), and Douglas(ll). Several modifications were made to adapt the equipment to the present investigation. As shown in Figure l1 distilled water was pumped from its gaging cylinder (A), through a cation exchanger bed for metal ion removal (B), to the reactor (C). Oil was pumped from its gaging cylinder (D) to the first cylinder of the mercury displacement feed system (E), displacing mercury into the second cylinder (F) and in turn propylene to the reactor (C). To provide positive pressure to the pump intakes, assuring uniform pumping rates, a nitrogen pressure of 50 psi was applied to each gaging cylinder. The two feed streams entered the reactor inlet tee from opposite sides, mixed in the center, and passed into the reactor through a feed preheater section where the reaction mixture was heated to reaction temperature. A stainless steel cone, with a 0o.00o6 diameter hole at the apex, was placed in each of the feed lines at the entrance of the inlet tee. The feed streams were thus jetted into the tee producing good mixing of the two phases. After leaving the preheater, the two-phase mixture passed through the resin catalyst which surrounded a thermowell containing a traversing thermocouple (G). A gas loaded back-pressure regulator valve maintained the reactor pressure. The -10

-11- U V)II. 1 W > U m0 I- I 0I — () W S a.~ W0 Wa I / I- >I aeeW g^^ ^ -)1 - -f - ao0a - O 4W n. uo P, I ol < 2 < &W UW U>. HW ( —IX-z "30NIIAz z''"~~~~~~~ 13 ~ - ~<:F - WV~w r Z~~~~~~ eu

-12~ prod.uct passed to a separator from which the propylene flowed to a water bubbler and a wet test meter vented- to the atmosphere. Apparatus Description The gaging cylinders (A) and (D) were constructed of 5" D. x 36" brass pipe with pipe caps threaded and soldered on each end.. Each cylinder was equipped with a pair of Penberthy gage valves and a 3/4-" O,D, heavy wall pyrex gage glass for the measurement of pump input rates. The pump was a Hills-McCanna type HAJD-5/8", dual unit rated at 10,000 psi maximum pressure with adjustable rates of 0.08 to 0.8 gph. The check valves of this pump operated erratically and were replaced by external twin seal check valves, Autoclave Engineers type 50K-4400. This valve has a spring loaded check which offers a metal to metal taper seal plus a neoprene "0" ring seal. The modified pump operated well at all pressures and flow rates used, i.e., 0.5 to 15cc/min. The vessel used to contain the resin bed (B) was an Amiijco type 406-38A reactor vessel, rated at 15,000 psi, with inside dimensions of 1" D. x 8" L, The mercury displacement vessels were obtained from equipment storage. Vessel (E) was an Amninco type 21-4750 reaction vessel of manganese steel, rated at 15,000 psi, with an approxirmate volume of 1750 ml. Vessel (F) was an Aminco type 41-4675 reaction vessel of manganese steel, rated at 15,000 psi, with an approximate volume of 1150 ml. The tube which dipped into the mercury in each of these vessels was 1/4" stainless high pressure tubing threaded to the cap.

-13The reactor vessel was constructed of 19-9 W-Mo forged product of Universal Cyclops Company. It was approximately 30" long with a 2 1/2" O.D. and 3/4" I.D. The volume of the reactor with the 1/4" OoD. thermowell in place was approximately 150 ml. The reactor was lined with a 30" x 3/4" O.D, x 0.020" wall thickness tantalum liner, and the thermowell jacketed by a 34 1/16" x 1/4" I.D. x 0.010" wall thickness tantalum sleeve. To prevent bypassing of the reactants, the liner was rolled. into place with a heat exchanger tube roller and the thermowell jacket was welded around the top. The resin was supported by 2" of pyrex wool in the bottom of the reactor, and. approximately 2" of pyrex wool was placed on top of the bed to prevent the resin from being blown back into the inlet lines in the event of an accidental pressure release. The reactor pressure was controlled with a Grove Model 90W back pressure regulator fitted with a teflon diaphragm, loaded with gas pressure from a nitrogen cylinder. This valve could be used to regulate the pressure from 100 to 2000 psi. The temperature was controlled by a 300 watt heating wire wound over the reactor length. To obtain a uniform temperature distribution, the reactor was divided into 5 longitudinal heater sections: two end sections, 3 1/2" long, adjoining 7" sections with a 10" center section. On top of the 300 watt wirey each section was wound with a 100 watt heating wire controlled by a separate variable transformer, A thermocouple (between the reactor inlet flanges) actuated the temperature controller which regulated the feed preheater, a 275 watt heating tape wrapped around the reactor inlet tube.

-14As seen in Figure 2, the reactor was covered by a double layer of glass cloth held in place by Scotch No. 27 High Temperature Electrical Tape. Heater TOTT, 300 watts, was wound on this at 1/8" spacing along the entire reactor length. This was covered by a layer of No. 27 tape and two more layers of glass cloth over which were wound the five sectional heaters, "X", "T" "M" "B"T and "E", These were also covered with tape and glass cloth for insulation. All of the heaters were wound on a lathe to obtain uniform wire spacing and tension. Table III shows the pertinent design data and measured ratings of the completed heaters. Figure 3 shows the heater wiring diagram. TABLE III'IEATER DESIGN DATA Measured Heater Wire Size Length Resistance Actual Rating rtt0 22 ga. 31" 105 ohms 349w @ 240v, 1 45A r"X" 30 gaO 3" 114 ohms 126w @ 20v, 1. 05A'"T" 28 ga. 7" 150 ohms 96w @ 120v, 0 80A t"M" 26 ga. 10" 140 ohms 103w @ 120v, 0. 86A r"B"r 28 ga. 7" 148 ohms 97w @ 120v, 0 81A "E"T 30 ga. 3" 140 ohms 103w @ 120v, 0. 81A NOTE: All heaters were wound from Chromel A wire at 1/8" spacing.

-15HEATER WIRESOTCH NO.27 TAPE CLOTH _\^ ^^ [ ^^^^1 ~~~~~DETAIL A REACTOR SECTION " I I-GLASS CLOTH (2 LAYERS) 2-HEATER WIRE 3-TAPE DETAIL A Figure 2, Eeater. Details.

-16-. t~ t.~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~. z o 0 P~~~~~~~~~~~~~~~~~~~~~~~it Z~~~~~~~~~~~~~~~~~~~~~~~~~O3 o c O hp CO~5 0~~~ ScI3NuI~iSNVW. 319VI~1VA ea% sc (> co~~~ w~~~~~~~~r 3 j~ 0 - 4~~

MATERIALS Chemicals The propylene, supplied by the Phillips Petroleum Company, was Technical Grade Propylene, 95 mole % minimum. Our mass spectrometer yielded the following analysis: Mole % Propylene 96.53 Propane 3.47 The isopropyl alcohol, supplied by the J. T, Baker Chemical Company, was reagent grade containing a maximum water impurity of 0,04%. The water was taken from the East Engineering Building distilled water supply. This water is single distilled and contains some metal ions introduced in the storage and distribution system, Cation Exchange Resin The Dowex 50WX8 resins supplied by the Dow Chemical Company. is a monosulfonated polystyrene with 8% divinylbenzene cross linking. One sample was screened from a batch of 50-100 mesh resin and another sample from a batch of l00-200 mesh resin in the following manner: 1. The resin was allowed to stand submerged in distilled water for at least 10 minutes. 2. A small amount of resin was poured on a series of Taylor screens, 32, 42, and 100 mesh. 3. The resin on the 32 mesh screen was washed repeatedly until most of it had passed through. -17

-184. The resin remaining on the 42 mesh screen was washed several more times. 5. The fractions were separated and the process repeated for another small amount of resin. When approximately 600 ml. of the 35242 mesh resin had been collected. it was put through the above sequence two more times. The 100-200 mesh resin was screened as above using 100L 150, and 200 mesh screens. The resin obtained from screening was washed several times with 6N HC1 and then with distilled water until the effluent was approximately neutral as determined by pH paper. The resin was analyzed for capacity and water content by the following technique: 1. 5 mL completely saturated samples were measured in a 5 ml. graduate, 2. The samples were transferred to Buchner funnels and centrifuged 4 minutes to remove excess water. 3* The samples were weighed and transferred to Erlenmeyer flasks. 4. Approximately 20 ml, of 1 M NaCl solution and 2 drops of phenolphthalein solution were added and the samples were titrated with 1 N NaOHE solution, 5. Another small amount of the resin was dried in a vacutm oven for approximately 6 hours at 100~ C 6. Approximately 2 gram samples were weighed. out into Erlenmyer flasks.

-19% 7. Approximately 20 ml, of 1 M NaC1 solution and 2 drops of phenolphthalein solution were added and the samples were titrated with 1 N NaOH solution. From the data obtained by the aborve the following calculations were made: 1) meq../ml. (wet), 2) meq./gm. (wet), 3) meqo/gm. (dry), and 4) % water.

DATA AND RESULTS Rate data were obtained for the continuous hydration of propylene to isopropyl alcohol using an acid cation exchange resins Dowex 50WX8, as catalyst. It is hoped that these results will be useful for extrapolating rates for this and similar reactions. The range of variables was.: Temperature 100~ - 160~C Pressure 450 and 1440 psig Resin Size 52-42 and 100-150 mesh Hydrogen-Ion Concentration 0.462 - 1. 678 meq./ml, Feed Composition'(liquid) 0 - 60o isopropyl alcohol The reactor was operated as a) an incremental reactor and b) an integral reactor. In the incremental reactor propylene and mixtures of isopropyl alcohol and water were used as feed with relatively high flow rates and low conversions. In the integral reactor propylene and water were. fed and the feedL rate:.was varied. In both, sufficient propylene was fed to maintain a vapor phase as well as the liquid and resin phases over the reactor length. Two runs were made to obtain resin life. A series of runs was also made at 130 C and 1440 psig to determine the effect of hydrogen-ion concentration in the resin on initial reaction rate. Incremental Data and Results Data at varying feed compositions, taken by the incremental method. were obtained for seven conditions. They were plotted as rate -20

versus composition on a propylene-free basis in Figures 4 - 10 and in Tables IV - X. A detailed experimental procedure is outlined in Appendix D and the product analysis procedure is found in Appendix E. The feed rate was adjusted so that the change in concentration of isopropyl alcohol was less than 5 w%. %. The incremental rate was then calculated by the formula, Rate = FAx/N. This was assumed to be the rate at the average of inlet and outlet composition. This rate was then corrected for isopropyl alcohol lost in the vapor product and for the deviation from the correlating temperature. Sample calculations are given in Appendix C. The plotted data show that a parabolic curve fits the data. Data were fitted by the least-squares method, and the mean deviation, _ y was calculated. Points falling outside the range of 2.5c (98% confid-ence interval) were discarded and the data were recorrelated. A summary of the correlations is given in Table XI and Figure 11. An estimate of the experimental error is given in Appendix F. The data, taken to determine the effect of temperatureS pressure and resin size on the reaction rates, were divided into eleven groups in Tables IV - X and Tables XII - XV. A comparison of Groups III and IV shows the effect of resin size. Both were run under the same conditions, 150~C and 450 psig, but the resin for Group III was 32 - 42 mesh and that for Group IV was 100 150 mesh. Although the correlations are not identical, the deviation is less than 2. 5cr therefore, the effect of resin size on the reaction rate is assumed to be negligible. * StatisticaLly speaking, a is the standard deviation of an infinite sample, The a used here is s, the mean deviation of a finite sample. -21

-22A comparison of Groups IV and V shows the effect of pressure on reaction rate. Both were run with temperature and resin size the same, but pressure was raised from 450 to 1440 psig. The rates increased, presumably due to an increased solubility of propylene in the liquid phase. Equation (4), the theoretical equation, predicts that one equation could correlate all of the data. The constants, A through G, are made up of the fundamental system constants, i.e., solubility constants, distribution coefficients, equilibrium constant, and molal volume, as shown in the derivation in Appendix B. These system constants are functions of system properties, concentration, temperature, and pressures The constants, A through C, obtained for the correlations of data were correlated as simple functions of temperature, with the knowledge that they are complex and unknown functions of temperature. Substituting these functions of temperature for the constants in the rate equation, a single rate correlation was obtained at each of the pressures which correlated all of the data with a fair degree of accuracyo The calculations are shown in Appendix H. r exp(-.21,600/RT){(37. 94 + 0. 668t- 0. 0026t2) 09 + (2.059 - 0.6517t) 1010x + (L952 + 0.00937t) 1011x2} gn. -moles/eq. -min at 450 psigo and r " exp(-216600 /RT){(. 68 - 0.005t) 10 - (4. 32 - 0. 02t) - lOlx + (20.98- 0.122t) lOlx2} gm.-moles/eq. -in (6) at 1440 psig

-230 xl rd )0 "t 4 0- \ 0\ —,' \D 01: \O \ 0 - -.- 0 n 0.c CO 4, o O 0 d O O O O ONJ O H C'O CK nCQ o H 0 0 0 0 0 \ O -~0 O U% O'O O -t O,-=t k o ~ O 0 o,O 0,-.-i o H o\ L4 — t 0 nm o C t) t m O 0 4' 0 CY s ON rl r r-i rr 0Hc: O 0 t 0N0000000000 o S o O4 Cr r \ o'r r- \ -' O 0 O O 0a a a) l ~ O\~ ONL 0 —{ ON 0 N 0 — O p F 39OQ O r0\ ri- Cl o rl O r O( O0 O O 0 0 0 0 0 00 0 0 0 0 0 0 0 P - o c Lroo 00 10r- o — o o- o \ o — L cN H N. D rJc\j — t C\iH CMl r{ H r-10 0 IH ~ ~ 0OJC 0- H OO 44 O H t3 0 o 00000000000000,'H 0 fiO < (MOJCj~o-C0 rLO O {> ^ c) o o oo o o o. o ooOHOJH P; d... 4 * * * ~ * ~ * ~ * ~ ~. o * H s, -rl o Yo' CY o O o cH0 T0 - o O \ k \ - c- ) Q 0 00 00 0 0 -* O * * (.0 "\ C V OOOOOOOHOOO X t Cxl Ctl Ql st 4 1 (\JHC'l rJ H 4', d.. 00.-. o. 0".. E.O O 0000 L0r- 0 4,- (r)01\ 0co H -010-I L-(l CY) 0\ (. CC 00 - 00^o~~~~~~~~~, -i-I,- d,'0 H \ O S S O O O rwl C l ( 1 ) O O s O rd C r ^ 0 ** q0 ~ H 4C)~ LC\ 0 Hbr- LO r 4 O O O O O O O O O OO rOH r N1 Pi o * oN 0 -q r\croorCH0H0 — 1'- 0oN'H O4X a Ha a! ooH * m o i * H 0!4'H0 OOOOOOOOOOOOHHHr- 0 H -p ar Co (L s Lr \ _:I \O O O H \ c Lr\ 0 01 p ONM C\1 cLc-^O r( C O \O HL O N Cl r 040 O\ H -C' \r\oE-OJHLCN Pi\ A r p4 Lr\- t- t - t Lr\ Lp \c n Lr\i Lr Lr\ \ Pi *H.d 1 r rl.. H'Q 0 40 S So N mc0-n'-t —-os -0 0 I < 0 00 \0O \H n 00\ 0- O -- o- \ c 0 t* * * * * *.... 0. r C) 4'- 0 *rl -P~ 0 <Uzc 1 n a L5 CXc *) ~ 4 O r~l Co XD ~ 4 LA X O z M z a wO L o o o t- O 0 o C) c 0 O O O \0 \ - Lr-0 <0 CCOHl- r -I -00 -l O CC) J 0 Q U 0 CO O 4 c CymHHC H H-r- 0 (- NC-)t -— CLr\ Ho r- CO@~~~ *H* * * * * * * * * * * * * a 0 r-l P ir\Lr\ L r- t OJO H C cl O rC Cl H H S rI cP 4-, C) 0U 0 H 0H O Q'r- 01 -t"- - Cr\ 0x VO O O O O O O r u -7 r rdl rd r I0 J 0101

-24I.OxIO,.i CONDITIONS: IO0~C, 450 PSIG, 0 32 -42 MESH RESIN I I I I I _I X, -MO- LEFRACTION O ISOPROP ALCOHOL IN PROPLENE FREE LIQUID Fgur 4 Gro I D0at 0 0.02 0.04 0.06 0.08 0.10 0.12 X, MOLE FRACTION OF ISOPROPYL ALCOHOL IN PROPYLENE FREE LIQUID Figure 4. Group I Data,

-25~J N::0 L0 H "-' S G -:) \N o\ H O\ H O\-4t-OH N CM^ -t m0 -\ 01 CY,-)\cY) Gj O'I-rN -c( L OC\0H 0 C:, H 9 CLr rH M( LnH N- GD ON GD'O O OO GD - n ONG'GDD\ o O On Qa) 4- tQ - El \EJ Q\co9 \CC co CC) \ C \ C) LN n- O c\ 9- H 0) C k0 LN-t K ( ll -C \J Ci r-I ol 0l K l ~- m O - 0! H H K K CT- C-, H- *'-I - I -I 0, H oIj r 4-o ) m _:t - \D 0\ t C\ o Cy.) O \ Co OC\ L — cO O \ O O\ CYICY) rq cO CYI \ 0 rl q O Ql I Q HHHH -HHHrl 0\ r- L C \ Cm rqC)oo 0co L \o~oCC) oo^-o -too.O \ It- 0 \i- 0 OO COG \ e I cOnO \O I O O D c c)l C)OCC) rr rH\O HH HH\ o H HH 4' o O+kO-=tk-. CCM L\-4 Ho tN -oCOM CY-) ONcr cc1 ONOkOtf t O G - cO H0H H H d d 0 o H H o 0 0 0 d O 0 0 00 - H Ho H ON -ONC 000N0- I H -0 HD 0 0 00 H'OONGDGDO NONC G 0 ON F-I 0 c^ r-i r0q- r -q rH rC oO ci OG rG CM O -0 0 0 0OH 0 0 CMCM0 0 0 0 0 0 0 0 H I HCMH0 0 c) 0 c H0 0 0H OGD 0 H MO~r-IHCf)O- OJOOOOOOCr i o H 0I t-~GDn 0 CMi-GOO O CM~CMCMCtc< ON.!\ 00I HHY) 0000 0000 -- O 0H0 M O0 CM 00000000000000000 0 0000000000000.H04.r. F) - - Or H < Hc G0 O\NC\0 ojt oj\ M \ i OH C O(OnH- 0CY -G n- O c MH\ L-n 0cot ON -O-NOOc\ O p d 0\Gco-:I-t —— tON\0 m I C Y'-, I \f 6 Lf\ co O co 0 ON GD ON~ () C cO H\ H-i t-M I C In H- H \D \0 O tU- — NON0 0-\ Lf\kO 0 ON () a~c; M ~r~H H 0- ~rl d ) U 0 0 ta c o * G0 cLr' CM cM c\o \c Hla\r cot cO - -C O- 0 \ —t r\ H' *o ( c O f\M -- 0 c -n t-C O\ i n-Or\\GCY)0 —-= CY ON Lr -- C \O COYI 0 ON -qHLGcoOO\CY) 0 1J O fl) 0\() iO cU lroLr\ rq 0\ N m C MO U H_ f \< \>0 \(co 0 O\ V\ \O O \ (I O 0 * —MM.Jo0G0 GD H 0 N 0 00HHCMC0 n')-\0kG -— GD H 0HH('L H 0\ o 0 0 0 0 r q 01\ \0 t-O N1\ ON C\ 0:I- H:I- ct M CY\ co t -t G\ H - t V \ CY,) \0 -:t G \ C a-t -\ C CM c g H **I C\!>0 - 0n\ 6 0O\ \ 0 o co a- H Y,) cj 0 tcO 0 C M CC i\ B O O O ON O ON O ON H L- O - Ot f ik\ LONONO ON LON t-O O\O \.0 t-'0 GD ).0 -: (nQ CY) Pi 0 + 00 oo~ ~ ~ ~ ~ \ 0\C LcoC t-c O c c CY R- n 0 \ C) t \ - o C) CMOHO CMn CMn CY,) r CM H H H H H H CY)n Cf -Z n-zi-n -OOH H \O q.4 CH 0 A...0 0 4 a U) HO N CY')0HONHCCY,)0C-\O C\C\] COO Y i c -J0)>- 0 O JO )l\?U) c\ jt 0\\ it COjr:frLr C)i\H I\C, t — c? CYO,)r\Lr\L\ r\ L C p^O L- 0 co \O 0 0l \ \O o\O-4 Cft _ CU CtO \C\-CM CMON MOO \G H HVNc -\ 0O \ NHi0 a,\GcoDCM 0 -H CM 0H r HO) P4 ^ 1: a) 0 OH OO DCOsNL-: \ 6 U t) CM\JGGD GD ONON -i ti rl t ON C\MC\C* G i H r ac H.P 4 C) V 0 4'-P 0 H C croc~oH t — HO NCIzi -coco GONON — Ho cO o o 0CC) ON O H' H U) 0Q 0 OG C-p GNC o-'. O N N0 O pC; O O O O O O O 2 O r( O ON ON ON ON*'.0 G-* t- G GD GD O OOOH <U 0 O OOOOOO OOOO OOOOOO HH OO HH

-266.0 xl03 —... CONDITIONS: 120~C 450 PSIG., 32- 42 MESH RESIN 5.0x10i3 __ _ _ _ ~<~) ~ —- LEAST- SQUARES CORRELATION.... G.~ —- nGENERAL CORRELATION 4.00 0 O X DATA DISCARDED BY 2.5 C TEST 0 w -' + DATA DISCARDED BY LOW ~C~o 301 r PROPYLENE MATERIAL BALANCE 0 2.0x10l3 4 ---- 0 30 - -~ Figwe 5. GraL'p II Data. i.Oxi(dX 0 2xlO 4xl02 6xlO'2 8xl'O2 Ix 0"l 1.2x I"' 1.4x10l' X, MOLE FRACTION OF ISOPROPYL ALCOHOL IN PROPYLENE - FREE LIQUID Figure 5. Group II Data.

-270 rlQ r U HUr I O \ cO MN H LN O O0\ D -r On L Lr\D U) *Q 0C Mt N H cO Lrc t>- ONO ON 0 C\ LrU C O -- -OD o \O L ) t- r-H O CO O O O \\0\ CU' L N \OI- \ Ir O C\ J O ct a pO 0\ p- C* -4 \O \<D L~\ Lr^ -}* m C C H O' 0 O 1-I N 0 r; b H CM 0\ -, 0 \ O ~ \0 co-:CO-OO CO CO co o0 4 0 C \) O O ts Ef U od o~co~ o0 o; C oC~ 0 t C ol 0 000 CO" (0; C; 0; 0a 0; 0 C a o c C\J C'l CU C n C CC CI CY CLI Cj CL I CL N CU E1 r - r — r -~ r - r- r H r - - r r-H - rH-I H r-' i H 0\ 0C 0 0\ O 0 0 1' L \ - - C \O 10 O\'\ 0\ L — ) n Lrn -- r' O-I 0 00 -- O O -I LrU\ r- H CM cO - 0\ 0 0 O \,\ O I \ a o(\ a,\ OO0 CO \C C 0 (\,W r- rH r- l rl CMOJ cr o o O LO o r- lO n on uno o c ~} O 00000-00 0 0| 0 0\0 000000 <0\c \ H H O\ o _: E N H H- \ un S c0 CC - 0 o 0 \ N. CvJ n N 8CYCY) -t \8 L1\ (Y) (\j o r- \10 )(n C\ 1HH HH H H ~. oooooooooooooooooooo Cr 4.......... t- H C O H 0 - dI) o 0 0O < O o ON-O - t> - \o o n O oN \ _ 0 H.. - O H fc> E o-,o-. "l,-I r (lrx 0 H. 0' 0- ~C - -': C.- C C 0 O C ON.. H <r;, - -IH 0 ('LIH l>- - o C 0 0r H ux c W^ ~~~ C~ ~ -P- CT*x r IR o00 0o o o o N0COoO o oOO o U) ^ t>*-cO\NN Lr* (n( (^*o tsot0 t C000 000 000 000 00 ^cn rI 1: 1 \ 0 \ 1N\ H 0y a 0NOL N 01 O N 0 r-O PO O H OH {=! 0 h*H O (OOO OOH LH A CU; t I C C 0 *l *..) N........n. r t (r) 0 rlO rl N I d rH, O0o~ H 0 *r0 0 0 L l L* ort. O O O HO Lr ( H ) o E,-I k < 4, oo - H.H -: o:.O- -o,-, oIn. o h r r-Cr Hn 0Ho1t'-O J0Lrlo 0''O- 0 o H H HH O O CU4 LO 0 0 * I l L\ 00 \ O\ r0 r- c - O\ le 0 0 NHH r O r( C\ O O \ ( CO' cO HH H O 1 H\ 0 rU Q ) 0000 0000 0 HHO OCOO^CML-O^- - OOHO C)> Oo H) 0 c oo *, *c- * o o coO** * c rl ri ~l rl O rl U) H r-I rd H C H d HH r- \ \ O r-'l r,'-I r ( O RH t * - ^t ^- r( ~c Lr- ~ b — -. C O O 0 L O\ O r( d - o D C'd 6 ~ M()- Pt -- Ln L! N IO J r l V 1o O t s C tI 1 LO O\D - a 0 ooooooooooH- OH OOH r O C0 U)o 0.U) L' Q. "X L O k.O..O D - b-. - - o r O t a.O.) p 000 -4 O C)\,On.9 I r- CON COA H H - L L \D r- 0 C C,\ C-

i-0 3-8 IOxlO1 0~\ | l CONDITIONS 130~C,450 PSIG., 32-42 MESH RESIN \ \ -- LEAST - SQUARES CORRELATION 9.0x10 \ I O. -- GENERAL CORRELATION \-+I X DATA DISCARDED BY 2.50' TEST \+ DATA DISCARDED BY LOW PROPYLENE 8.0xlO3 _- MATERIAL BALANCE 7.0 x I__ I I I 1 \-3 X 4. xIO 3.0x10 I —-- + J \ L\ X, MOLE FRACTION OF ISOPROPYL ALCOHOL Figixe 6. Goup III Data 2.0x10 - 1.0x10"' ------ 2.0 xl( — 0 2xl02. 4xl()2 6xl0'2 8x10z2 IxlO1' 1.2x1lO' 1.4x1lO' X. MOLE FRACTION OF ISOPROPYL ALCOHOL IN PROPYLENE- FREE LIQUID Figure 6. Group III Data

-290 e Vr rl \ ON \r o 0 \-D N-\ KIN CC C) H rq H H H H-I HO ( 5 -t ON 0 0OHO\Lr0 00oC' O * Q Ct c O O 1> LD Lr0 0- CI N- C\ U) # O CO On O NO NO kn-O z Oi - 0) o co m 0 0 0 C00 0 0 \C0 0 cC CC' CC @ @ C0 C- 1 C CU 40 o, ooooooC c o o\oc\o\ d o\ N H o o o n\ o;I o mo H 0 HE1 rH H r HH - -r H0t t-I m \@~ t-! o r 4 CO 0 () CX \~ H\ c O 0 \ CH 0 H 0 \ C\ H 0 -0 0 H -o O cO Lr\ 0 -o c- rol o-H \ H.- co \O H:o ON Lr O - I- O O N LC 0 O XD 0 C H N H oo _ o Lr \ cU 0000000 0000 k ~ Oe Ol O~ O O OO O\ O\ O OO 00000000000000 0 cl C\ O kD0 H t — u \ o r o~a oo;ojmooa H nmaoL 0 o < \ o o - \- o m Cyo-) 0 o o" ) U'0 OO OOO OO O0x 0 c L -LO -"' kO @- 4 r-H - -I H -- r-I Hr - H V E. ~....... ~0 0,1 COk b — - I.cU1- O G-. COO _ E3 O -LL O *HO * HH 2 *H 0- 0 H OO - \ CC)- -ItC),\ O ON' 0\ ON- o o ca CC', —co 1\ -I,- -I CO *rl - H ^~~~o0 I r; dc; H O 0 CUU H o @c4 uOO \ m N t o r u n — O H U PO CO o \0 1 O O C Y O \Oc Cj c *r H0 H..HH.H HH_ I 4 H I ~ U r a) q O U P Oi CO O 0 0 O O i U (III\ \ r- H HH 4r- d a:) c cu o ~ OC CxO OJ f OL C ( k Q O Oi nXO Cn O C) cO OU O n3 a o O *H O M o o H m t- O a 0 (U CDrf FC a; L r\Cj L H a r CO -CI- C\ tL- ) t- - O rO oA ) C\JN oCo \D @ cCO-CjLr-\Ct- - +O t- OHd Lw -:0- Lnr\o\ cO c 0 coO GO N C), OR aO ts L O c C'CC''H o4 a o oH C@ LC'- z ONLnOHr O C —Hrtl- 0 U) 0 OOOOOOOOOOHr Q - C) cO7crO\0 H0 OO O - 4r) C I -I I-I I-I I- I- I I-I a) COONONONONOOOOHHH~~~~~r HHHHHHH~~~~~~~~~

-30IIrldo CONDITIONS' 130 ~C, 450 PSIG., 100-150 MESH RESIN -3' 9.X0 I -- LEAST - SQUARES CORRELATION 8.0n... - - - GENERAL CORRELATION 8.0x103 ------ -3________. -3. 0 x 10 U) \ ILJ 0 25xO 4xOO 6xO 8xlO WxC0 1.2xlO 1.4xl IN PROPYLENE - FREE LIQUID Fiure 7. Group IV Data"' 0 2.0 x IO — 0 2xl(2 4xld0 6x10"^ 8xl0'2 IxIO" 1.2x10l 1.4xl0( X, MOLE FRACTION OF ISOPROPYL ALCOHOL IN PROPYLENE - FREE LIQUID Figure 7. Group TV Data

-310 x r ( L rl\ P > - Ln LI O oo O Lr\ -:o) o a\ ON r- WH < \. 00 o Lr\ 4. C** ~\ 0 n C\ H Ho nI CO L * \ _t C\ ~~~~~~O m; 6~ ~ - *H000 )-O\- O * CO O\ \ — - L —:I OOO\(O OHO O\C\ )'P CY') C\J r C r) CUj NO C CM CU CQ ~El ~ ~ ~r-t ~-q,-t r- rl, -. rl r- r-I r-I +5 r4 0o t- o C" -0 H' C- ON C'I C O a oo...ooooo o@'a. 0' U 0X a,\O 0. 0 " 0 C\ o H iH H r - H r-I H H- ~r'l 0 0,\ 0 - Lr\CO co-.tkO \ CoY-, r)-.:- \0 + 0o -ot0CO: c;O:,-; -HO r ooo s o O \c 0 O C\ 0 o\ o o F i P HHH H - H, 1 HHH -tC\ CO 0 C\C H H r-lQ 1-UtL - t cor o \tCM- c v -O co crncO PI \ cO \o \o - O\ cO 0 o- Lr\ Lr\ d re) \H r co LH\ o CC H - L\ \O f \Q H \HCO L H N N L 4- c m4 MC) 0000000 0'0 0 0 U) 44 OH 0 O 0 0 0 00 0 0 0 0 0 0 0: r -o- OP CC cmCMJ cmJcJC JCM H r r-H o O o O O O O oO O O O O h *... *.C. I 0 E I a H r U. C H CO O 0 O Lt CO n ~L H H~ hD~~~ o \\O p' o,, co 0 c l. —. —. _.. co -,- O - 0 H H Y) 0 d H H H R N 0od O - n m C —-C CC r —I Od C) HC'HHHH HHaHH H 1 E-1 h' *0 >0 c0 0 0 0 0 0.0 -I 0r 0; r-?4 r asC nL o 0 ~H PL..-) 0 Ik tr kO 00 t O D O 1 ~ \-ILf^ \ 0 0 O^O\ C\r-4OJt *H 0 PC)~ ~~a0o ri 0 oL u O\ r-c mc Lr\ cO Lr\ -- 0 < 0 * 0 - "' * * * *'- * * *\' * ^ \0 0 P-I d; 0 e A X CO C O \ co ~ L ro c) 0r t) r\- cy H r'- OOH I..C\J k. \ <)O b-CO 0, CU 0 C0 0 0 - iO' H V H O CH P1 O\ c- o r4 L \ cn cl cr\ \n \ (> r4 t H 0' 0' O" X O' OXHO 0,- 0 - H r r-0I )rlJ r - I HHHH r- O -P03 Pi OP n Lr\C —O HO m H n UH- O A) O O L —- -C -o tC 0 -c _ _ — I - H c OC) X n r Lo C0 lr~O r- Lo HC C H H HoH H H H H 4 a) U) X o 1Y CJ 1 H r l m X 1Y 0 o i C) OO O O O cO OO X LA c H C) li { e-l~ o o; o oo o od od oI; d ( o r pC) HHHHHH~~~U

-3220x10l -3 \ CONDITIONS 1300C, 1440 PSIG., 100-150 MESH 18X10 xIO- --.LEAST-SQUARES AND GENERAL CORRELATION 16x10(5 14 x IO) - 14x10 -- - — 6- - -- - - -- - -1. x cn 12x10 d'" j-. 0x10N -- - - - - -- F- - - - 0 6.0x I0$ 4.0x10 3~ 0 2x102 4xlO12 6xl-2 8xl162 I x10' 1.2x10' I.4xl-0 X, MOLE FRACTION OF ISOPROPYL ALCOHOL IN PROPYLENE - FREE LIQUID Figure 8. Group V Data,

-33Q<-43 N -:l- \OJ t ^OO H-t- I O H r-|OJ -:ICO CO O 0 r\ O L H OJ 0.,, H q 0,o o\ rl., t- t cO, \,- -\ o.,, "- o or oA -- o.-, o,-, SO P \ 0 00co -=* t"- \d -=t Nc S 0 O Lrd rc c 0 1 d Lr^ N^ K^ r^' cKc' -t:: r\ 4 o i rHH r HI r- CM r-I rd H r'' -P P c o - - O,, oH o, 0 -,t,-0,oH o,0 C,,C-O'y \O O~0-~ C-.-' 0 H- - 0C —' L'n ON 0 ON' o H \O" CO cO —\t CM \O O\O CO H\ a) d COC'1 o- 00 0 0\ I i.-P',r\M O'\ O o C- C O\ CO\C-Lr'-0, COOO C- Oin C- Oo \O (U ro r 1.1 A * * i * * * * **\i * *1 NP cOO'cOOO 6 C- H \CI- 4 -I- I MC, —',\ 1 C\,-C\ 0.\ -,1 I,\ 0oC-CO H. HH H H OOONO O,- r- I 0-. I, O-I, - N O-O-I r-I C\j CM\j t — c ON\r- t LcnCC- -- HCOr- CO) c, CLr\ --- c —- t O CO \O -0 P.C- O (C oC ) CM) CO 0 co,-t 0 O CO' - Ok 0 H\ COc o - \CO CO oC-1-CO CM ONclMOCOCO H H LN-J- L C —ONLINCO-1- LNCOOO H\O C O-: ~ O 0000 H 000 00ONH 0 00 r-r-000 0000000000000000000 0000000 C I 1H 0 nr co COC)COIcn L \1OOON\ CU k \ OO H\ \1 CM -4 CO rH-O m CM N -tONCYM C\ -H E-H CN cO H A o f t cO \1 \Q CM L - L o CO -t \ C iO\r)CO LnCLr\ H Q-i- \lcOCO H -- CMUJ -o LLr\ Cv - CYC) Ml — O \OLC —- \.O \'0C -C\ -jt -0 C-CO P4 I- H- HHi-I i-1 r rH HH H HHHHHHH-HHHl a) ^C C \ CM 0 LI\ LI- H NDON — LO OCMl \>0 L\ 0\ \O t- (OYO\ C00 CO N, NE-l C V S\ M C\ -l c [- \' t — L\ X C —LCN0-\CO o CM C Lrr-l-0 t —-'\ H 0 a) I CY) r-i \ c4 I L C-CO H \ \. C C t — co a r4 c;f;4 co rl CY) i +; r 0 0 H-+ \8 * 1 P 0O O 0O CO C H C r CM \ H- OC, OHCMJ \1 0CO -C C —-n CY,)H tO\ C0 C- *q -H 0 oR\ C -:I(4HLL C\ o0ofHHCMCr- O-\-t -3Lf C OC\-oo HO CMHYCM) COC O a H o H 0 0 I H'O'-OL c; OCO<oocM.CMCOOO LHCMjcoCML(\NOrNONOLNCOrn U'0'OO LC\ Nn LN in CrO'O- 4 L L( C O N-\I- 0 \- -- C- C — C —t C —cO C- L(N * o v r\tr,\r\roop, o -.,-Lr,\Lp-,,.,-,o L- - -taLr\ ooCt — or\ N c coo NC c C>- o0-tCM jn i HLiNO Qo NC NC.p CMLInO\ OCM 10 OCM\OCD-N I 0\OCt —\O — CMOJOeCMOLCN cN Nn ) * *............s H.......).. o HH H HHHHHHH ~k OH 0. I. I...-I. I. r- I ~ r4 0 0A ~rl C) &'~ vn Lr,\ N \O O'\O7\C H \ CM C\|t- uLrn Lr OJt \O t — O tN -- O rAl COH -ONOj CJ - o mI \\, I. coE —-C CcO I \C-,- c o M -o-,_ I o'\ CP ".I, I ~ O.I,- O" O~,I r'-I I I ~-,'-'- rO" OI I: -: C- C — H C O t - -- Pc 0 HH CM CM CM CMO Crc ziH- LN t O\ C-C- CO 00 CO

-3418 x1 3 --- I 1CONDITIONS: 140~C, 450 PSIG., 32-42 MESH RESIN O _ I I I I I I -3 0 a I o _ -- LEAST- SQUARES CORRELATION 16xl 10 -- GENERAL CORRELATION ~ I — --- -- X DATA DISCARDED BY 2.50' TEST -x13 + DATA DISCARDED BY LOW PROPYLENE 14 X10 - - \ x I | L MATERIAL BALANCE z -3 2O xlO6 10x10C6.0xlOI cc \' 4.0 x-IO Figuxe1+' 9 \ -3. x + 2.0 x 10 0 2xl0- 4xl0'2 6xl0'2 8x10z2 IxIl0' 1.2x10l' 1.4x10' X, MOLE FRACTION OF ISOPROPYL ALCOHOL IN PROPYLENE- FREE LIQUID Figure 9. Group VI Data.

-35~e,.J oi..ooo b^ Q * O O OO 8 )t *0 S 9. 0. -. -. 0 *) H K r) \0 C 0 K ) 4 q d ( # D a a C OHr-f:O O r\ U0 +4 o 000 O \QQQ Eo r-HHHHHHHH H U 0 00\0 \O\0 0 O O\ O0\ O0 O (3N O ON ON H H H *, e) cnJ rt a\ I CC) ~ 0 C 1H 0 H3N o J ( e 0 0' A pq H H C0 ) f\ro-..= -. O' ~... O..-. O. o. oY- O O' cH ri P) n o X -^~4 X -c t co H <Hj C' Ncon O - ~\l — t \ r (oo C 1O t. 4 0 O l l Eoo o *. I C*Cr-i rl - 0 0 0 H U i) Ot O(.\H cJ LC \ HO'- H p) r O cO \)OO t- - OO Id 1~d i~l cf~ o~ od c-r- k0 OH r-P C ^ A0 W X H \ 0' *\ O *c \ b) oYi OH-'\L\ C C H1J m L O,o0 H ~ ~ ~ ~Co O 1 i O 0no\ C l3 C CC( CJ C\ CM cn () )OC) U)re O; O O O rf r( aUc OC U) 0( ONC H H g m o U) U) N Lr \ r CO O \COn Lr\ U) \O O O HO Y ~; * H H H r ~rf ^-O H PL O -+ <3P- )

-36T 0 z z C 0 2 o i' -,, co >_f1 ~ 1 0 ~ r I~l cc IL -- I - -h - - - - - o 0 o I 0 0! _..K " I I. N I O " - -. i - 2 0, C0 0 H -— ) W 5, - a. 2 w h - 0 0; _,_e_ 0'o o < ) Q M 0 LL ------ - - - 0?o 1o o0 -o z' 0 <'0'c'0' 0 ^j o rO N N N -- ~' 0110 to lo I 0 lo I0 lo to I_ lo 0 NIHI -'03 / S10W - Wflt~b

-37TABLE XI SUMMARY OF CORRELATIONS Group I r- 103 = 0.6348 - l0llx + 66.59x2, 100~C, 450 psig Group II r. 103 = 4.668 - 49.65x + 249.2x2, 120~C, 450 psig Group III r- 103 - 10,10 - 121.5x + 603.1x2, 130~C, 450 psig Group IV r* 103 - 10.12 - 146.7x + 745.4x2, 130 C, 450 psig Group V r< 103 = 19.89 - 213.6x + 992.3x2, 130~C, 1440 psig Group VI r' 103 = 17.23 - 248.4x + 1147x2, 1400C, 450 psig Group VII r. 103 - 36,72 - 324.9x + 1454x2, 1400C, 1440 psig General Correlations: t = l00~ - 140~C, P = 450 psig r - exp (-21,600/RT){(-37.94 + 0.668t - 0.0026t2). l09 + (2.059 - 0.06517t) - 1010x + (1.952 + 0.00937t) * 1011x2} gm-moles/eq. -min. t = 130~ - 140~C, P = 144o psig r = exp(-21,600/RT){(l.68 - 0.005t) - 1010 - (4.32 - 0.023t) 011x + (20.98 - 0,122t) 1011x2} gm-moles/eq.min. R = 1.987 cal/gm-mole~.K T:', K, t = ~ C.

-3840x10-'" 36.x 10 32 x 10328x103. 24X1I g 20x103 - 140 ~C,1440 PSIG LaJ -J 16 XO 12 x10l _xI _ ________ ~_____ — 1300C,1440PSIG -3 _ __ _ __ _ _ I$__1300C 450 PSIG 4 x I03 1 —---- 1400C,450 PSIG 1200C,450 PSIG 0 I00~C,450PSIG 0 0.02 0.04 0.06 0.08 0.10 0.12 X, MOLE FRACTION OF ISOPROPYL ALCOHOL IN PROPYLENE FREE LIQUID Figure 11. Summary of Correlations.

-39where R = 1.987 cal/gn.-mole-~K, T = ~K, t = ~C, and x 0 to x s= 0.12, the liquid-phase mole fraction of isopropyl alcohol on the propylenefree basis. These correlations are shown as dotted curves in Figures 4-10. On figures which show only one curve, the general correlation is identical with the individual correlation. Integral Data and Results Data from integral reactor operation are plotted in Figures 12-15o The differentiated data are given in Tables XII-XV and the calculations in Appendix C. These yield initial rates at higher temperatures as shown in Figure 16, log initial reaction rate versus 1/T. The initial rates at 150~C and 160~C deviate from the expected straight line., This could be due to the resin diffusion becoming a controlling factor or to a combination of a number of factors such as the changing of distribution coefficients in addition to the increased diffusional resistance. Further investigation in this region was not undertaken because of the increased rate of ether formation above 140~Co The apparent reaction activation energy was found to be 21,600 cal/gm- mole from. Figure 16. The initial rates are given by: ri = 4.87x109 exp(-21,600/RT) gm-moles/eq.-min. at 450 psig,,O0-140~ C.(7) r. = 10.05x109 exp(-21,600/RT) gm-moles/eq.-min. at 1440 psig, l30l400C (8) 130o140o C where T = oK R = 1.987 cal/gm-mole-e K Chemical analysis of the reaction products shows that increased temperature, pressure, and alcohol concentration favor the formation of di-isopropyl ether. Data were not taken above 160~C where the competing reaction of ether formation becomes important.

~40<,.4oTABLE XII INTEGRAL REACTOR DATA - GROUP VIII Conditions: 140~C, 1440 psig, 100-150 mesh resin Runs: 117-121 103 x N/F Ax AN/F x. x10O3 gm-mol. AN/F eq. -min. 0 0 40, 6.0241.6654 36.17.0241.6654 29.5.0156.5902 26.45.0397 1.2556 24.9.0517 2. 332 22. 15.0913 3. 5876 21 7.0554 2 593 21 37,1468 6.1806 21.6.1518 6.893 22.02,2986 13.074 TABLE XIII INEGRAL REACTOR DATA - GROUP IX Conditions: 150~C, 1440 psig, 100-150 mesh resin Runs 1.22-126 x 103 x N/F Ax A/F xL x103 gm-mol. Au/ZF eq.,gmin. 0 0 41.5.0129 35127 41.09.0129.3127 - 40.8.0045 2797 16.09.0174.5924 40.2.0355 8904 39, 92.0529 1.4828 39.2,o44O 1.1215 35924.0969 2. 6043 40.3.0941 2.0727 45.42 1910 4.6313

-41TABLE XIV INTEGRAL REACTOR DATA - GROUP X Conditions: 160~C, 1440 psig, 100-150 mesh resin Runs: 127-130 r x 103 x N/F Ax,^AN/F -A x1053 gm-a1 AN/F eq. -min. 0 0 78.8.0282.4621 60 92.0282.4621 37.6.0057.2913 19 46,0338.7534 25.0.0159.7844 20 28.0497 1.5378 20.6.0417 L 1878 35.07 0,914 2.7256 TABLE XV INTEGRAL REACTOR DATA GROUP XI Conditions: 140~C0 1440 psig, 32-42 mesh resin Runs: 1531 133-135 r x 103 x N/F Ax AN/F x103 gm9mol, —.N/F eq. -min. O O 35.9.0305 9046 33.73.0305.9046 30 7.0457 1.6618 27.41.0762 2. 5664 22.4.0332 1.2187 27.21.1093 3.7851 18.0.0409 3.2889 12. 42.1502 7.0740

-4244 xo103 40xlCOLa3 ______ I I I I_ - CONDITIONS: 140~C, 1440 PSIG, 100-150 MESH RESIN 36x103 32xl 10 ~ 28x10_ -J o -. ~ 24 x1012 xl0 0 0.02 0.04 0.06 0.08 0.10 0.12 0.14 0.16 X, MOLE FRACTION IPA Figure 12. Group VIII Data

-4348X 10I 4 03 l CONDITIONS: 150~C, 1440 PSIG., 100-150 MESH RESIN 44x10 - _ 24 0 I ] o 40x10 4X1 61. w -J 0 36x103 - 32x 10 - -- - -- - -- -- - -- - -- -... 24x10...... 0 2xl0'2 4xl102 6xlIO2 8xl102 IxlO'- 1.2x10I' X, MOLE FRACTION IPA Figure 13. Group IX Data.

-4488x10 80 xl -3 72x103 CONDITIONS: 160~C, 1440 PSIG., 100-150 MESH RESIN z 64x103 0 %.: 56x10 IO3 -j -3xl(_ 1., 0 2410 48x 6X0l0 8X10 1XI 1.2X16 X, MOLE FRACTION IPA Figure 14. Group X Data.'- 40xl63 32x10324x10 -- 16xl0' m m.. 0 2xl062 4xl0'2 6xl0-z 8xlO' 1x10' 1.2xl0'1 X, MOLE FRACTION IPA Figure 14. Group X Data,

-4540 x I O 4Ox..... 36 x0 I _____ CONDITIONS: 1400C, 1440 PSIG 32-42 MESH RESIN 32x IIO —-- 24xl" 0oxi3 -3w 16x1004x 12 x IO 4 i 8x IO F"iure 15. Group XI Data

-46t,~c.3 160 140 120 100 100x1 | i- I' i'l - "1 ~~~50 ~xl0- l l |~ + POINTS TAKEN AT 1440 PSIG | l | \ lDISCARDED FOR HIGH ETH RATE 0 -.o 51x103 -- X X16 23xl0 2.4x- 1 2.5xIo - 2.6 x 10 27 x 10' I~3 l' / \ 2 1 /T OK16. Log vs. 1/T Figure 16. Log ri vs. 1/T

-47Resin Life Data and Results For the resin life runs, the reactor was operated at a high temperature for three days with analyzed new resin. The resin was analyzed for ion exchange capacity and water content. The first run was made at 200~C and the second at 170~C. The deactivation of the resin is considered to take place due to the desulfonation reaction. The half-life of the resin was calculated for each of these temperatures assuming the desulfonation reaction to lbe irreversible and first order in sulfonic group concentration (ion exchange capacity). Table XVI shows the results of the resin analyses. Figure 17 is a plot of halflife versus 1/T. The half-life at 130~C, found by extrapolating the high temperature data, is 155 days. After use at high temperature, the resin became lighter in color and the water content of the saturated resin decreased. This is presumably due to an increased cross-linking of the gel. The high temperature and pressure caused no apparent cracking or charring of the resin beads. Hydrogen-Ion Data and Results The initial rates were obtained for resin which had been neutralized with NaOH to approximately 35/4 1/2, and 1/4 the original hydrogen-ion concentration. Incremental runs were made with pure water feed and the rates were corrected to zero alcohol concentration using the correlation of the Group V data. The data are shown in Table XVII and Figure 18. Rate in gm-moles/liter-m.in versus hydrogen-ion concentration does not yield a linear relationship as might be expected. however, the presence of the sodium ion in the resin could change the

-48TABLE XVI RESIN HALF-LIFE DATA Run 148 Run 156 t = 200~C t = 170~C P = 1440 psig P = 1440 psig 50WX8, 32-42 mesh 50WX8, 100-150 mesh New Resin Used Resin New Resin Used Resin Activity 4.930 meq/gm 0.517 meg/gm 4.861 meq/gm 3.854 meq/gm % Water 52.83% 6.19% 55.22% 51.40% Run duration = 76 hrs. Run duration = 73 hrs. tl/2 = 23.3 hrs. = 0.971 days tl/2 = 218.3 hrs. = 9.10 days Ave. measured temp. = 201.6~ C Ave. measured temp. = 166.8~C absorptive properties of the resin, changing the distribution coefficients, and yield the above result. The data are correlated by: ri = {12.952 [H+] + 4.002 [H+]2}x 10-3 gm.-moles/liter-min. (9) where [H ] = meq./ml. of saturated resin. TABLE XVII HYDROGEN-ION DATA [H+] meq/ml ri[H+] gm.-moles/liter ri, gm-moles/eq.-min 1.678 33.554 x 10-3 19,890 x 10-3 1.343 24.380 18.131 0.978 15.917 16.275 Q.462 7.100 15.368

-49I /T ~K-' 2.1 xl13 2.2 x tK 2.3x103 2.4 X103 2.5X10"3 2.6 10" 900.0 800.0 - 600.0 400.0 -- / 200.0 / / 100.0 80.0 / 10.0 / 20.0 / ~/ 6.0 200 190 180 170 160 150 140 130 120 110 t, / Figure 17. Resin Half-Life 4 ~ ~ ~ Fiue1.0 /ei afLf

-503 LCONDITIONS: 140 ~C, 1440 PSIG. 40x 10 0 INCREMENTAL DATA, 32-42 MESH RESIN A INTEGRAL DATA, 100-150 MESH RESIN 0 INTEGRAL DATA,32-42 MESH RESIN -363 -O LEAST-SQUARES CORRELATION OF INCREMENTAL DATA 36 x lo B 32x10 W E I. 28x10' - 2- 24x103 - 0~~~~~ ~0 20x10 24x1~$ 16xdl.. 0 0.04 0.08 0.12 0.16 XMOLE FRACTION OF ISOPROPYL ALCOHOL IN PROPYLENE- FREE LIQUID Comparison of Integral and Incremental Data The above plot compares the data taken at 140~C and 1440 psig by both the integral and incremental methods (Groups VII, VIII, and XI). All points meet the 2.5a test with the least-squares curve obtained from the incremental data correlation. This comparison is sufficient evidence that either method of reactor operation yields data of the same reliability and that integration of incremental data is valid for integral reactor design. Since resin size has an insignificant effect on rate, both resin sizes were used for this plot. Incremental runs were not made at 150~ or 160~C; therefore no comparison of integral data at those temperatures was possible.

-5124x10) Z 20xI i -3 F 12x1l. 16xl{3.... 0 J-3 [ +] HYDROGEN ION CONCENTRATION, EQ/LITER Figure 18. Volumetric Rate vs. [~H-+]

CONCLUS IONS 1. The hydration of propylene with the cation exchange resin catalyst, Dowex 50WX8, is favorable in the range 100~-160~C and 450-1440 psig, The initial reaction rates for these conditions are: ri = 4.87x109 exp(-21^600/RT) g. -moles/eq.-min. at 450 psig, 100-140~C ri = 10. 05x109 exp(-21, 600/RT) gm -moles/eq. -min. at 1440 psig, 130-140~C -where T ='K., R = 1.987 cal/mole - 0K. 2. The rates become constant at a certain product composition presumably due to increasing solubility of the propylene in the liquid phase. 3. By assuming a mechanism, a rate equation may be derived which satisfactorily correlates the data for a mole fraction of isopropyl alcohol up to 0.12 and a temperature range 100~ ^140~C. (It is given on page as Equation (4)). 4. The competing reaction to form di-isopropyl ether is favored by increased temperature, pressure, and alcohol concentration becoming Important above 160~C where it corresponds to approximately 10% of the propylene. 5. The half-life of the resin, found by extrapolating data at 200~C and 170~C, is 62, 155, and 370 days for 140~0 130~, and 120~C respectively. -52

APPEND IX A DERIVATION OF RATE EQUATION The rate equation for the carbonium ion mechanism may be derived as follows: k H2C = C - CH + H+ zL (HC - C - CH5)4 (i) H 2 H k OH (H3C - C CH3 + H20 HC - CH3 + H (2) H kiH k4 The rate of formation of isopropyl alcohol may be written: Ii-c7HOH] = k3[c3H7 ][H20,] - k4[i-C3H70H1][H] (10) dt The concentration of the carbonium ion, [C^H7 ], is indeterminant, but may be found by assuming it to be small and its net rate of formation to be zero. d[CH ] + + + t 0 [C 6 ] -k2[Cc ]s + k[i-C5H7OH][H k[C3H ]s.s.[ H] (11) Solving Equation (1l) for [CH7+] s. the steady-state concentration of [H7 1.5s carbonium ion, and substituting the result in Equation (10) yields the rate equation for isopropyl alcohol in terms of product and reactant concentrations. [ kl[C3H6H[+]e + k4[i-C3H70H][H I [CR H1 (12) 7 s. k2 + k3[H20] d[i-C3'7OH] klk3[H [iC iH (] dt k 2+kH2 { [C03H[H 2o] 2.... where K = k31/k2k4, the equilibrium constant. -53

APPENDIX B DERIVATION OF MODIFID RATE EQUATION Beginning with the rate equation derived in Appendix A. d[i-C H7OH] kk5[EH] [i-C3H7OH] -^3 7 klk3 -- f{[C5H6[HH0] K —- -- dt k2+k [H20] 6 2 K which is the rate equation for the carbonium ion reaction in a homogeneous liquid-phase reaction, the rate equation for the reaction in the resin phase may be written by adding the distribution coefficients. i+tI01 i >[C 3H61 sc[H20,sW];A~ } (13) at kk+k [H201 [[ K where [ ] is the concentration of a component in the external liquids phase at the surface of the resin and a i-s the resin phase distribution coefficient. Simplifying assumptions may be made to reduce the number of unknown coefficients. 1, Concentration of each of the components is equal to the mole fraction of the component divided by the molal volume of the liquid. 2. The mole fraction of propylene is small and is a function of the mole fraction of isopropyl alcohol. 3. The distribution coefficients are not functions of composition. One condition necessary to validate all of the above assumptions is that the range of composition must be very small~ This condition is -54

-55fulfilled since the range of the mole fraction of isopropyl alcohol used in the correlations is 0 to 0.14. A change of 0,14 in the mole fraction of isopropyl alcohol will yield a change of approximately 50% in the molal volume, if the solution is ideal. However, assumption 1 will stand since the purpose of this derivation is to show the general form. of the correlating equation. Solubility of propylene in water-siopropyl alcohol mixtures is not known, Azarnoosh and McKetta(3) have published the solubility of propylene in water in the range 200-500 psig and 100-220~F. This data may be extrapolated. to reactor conditions with some degree of reliability with the least confidence being placed in the high temperature, high pressure valueso The range of the mole fraction of propylene is 0, 00091 - 0.00322, An estimated ternary phase diagram may be prepared from estimated binary diagrams and a ternary low temperature liquid.-phase diagram. Figure 19 is the ternary liquid-phase diagram., Figure 20 shows the estimated binary diagrams, and Figure 21 is the estimated ternary diagram at reactor conditions. From this diagram it can be assumed that a parabolic curve of the form Xp = a + bx + cx2 will approximate the liquidus line. It can be shown that xA and xW, the mole fractions of alcohol and water, may be written as functions of x, the mole fraction of alcohol on the propylene-free basos, amd Xp, the mole fraction of proplyleneo XA x(l-xp) (14) W (l -x) (l-x) (15) Letting xp = a+ bx + cx2 (16)

-560 CYhLo ^^^ ^^^^^ P^^r^ ^^^i I,. ZAn ^^^\^\^\^\^^^\^^\^\ p o > oId ^'^l^' ^> ^>' > ^> > > w -H I1 0

-57o. --- " ------ - -- --- d z -r ow N — n 0 0 o U N N - - _H Pi 0 < 0 od z U) C$ Lz e -rN~0 ~ i 0 - zo rd a 0 I 0 0 0 HC du.,-_ _ —-0 0 o o o o o o In O In o0 N4 N - - q' 3UnJiVu3dM31L

-5800 P4 CS, > ~0 C.00 ~~~~~0 OD 0 0~~~~0 CL ~ ~ ~ I~ co~acr dl~~~~~~~~~~~~~~~~~~~~~l~~~~~F 0 p.: IL L i~ P-1 O~~~~~~~~~~~~~~~~ W.& ~~~~~~~~~~~~~~~0'd ra I0 =,~~~~~~~~~~~~~~~~~~~~4 Q>~~~~~~ a 0 Ne I

-59XA = x(l-a-bx-cx ) = (l-a)x-bx -cx3 (17) xw = (l-x)(l-a-bx-cx2) = (l-a)+ (a-b-l)x+ (b-c)x2+cx (18) Now the rate equation may be expressed as a function of x, d[i-C3H70H] klk5[H+] r 1 _ + dt V kV+k3[ (l-a) + (a-b-l)x + (b-c)x2 + cx3]aW J {(a-a2 )par + (a2-a+b-2ab)oxpOx + (2ab-b2-b-2ac+c)lpCWx2 + (b2+2ac-2bc-c)Qpcx3 + (2bc-c2)PaWx4X+c2aaWx5 -_ [(l-a)x-bx2-cx ]lA} (19) where V = molal volume of liquid. For an ideal solution, V = xAV + wVw + p (20) For small values of x, it is assumed that V and -1 are constant. k2V+k [(1-a) + (a-b-l)x + (b-c)x2+cx ]3c The rate equation may be written, d[i-CHB70H] klk jH ] 2 4 di - -COH k3H —-= {A + Bx + Cx + Dx + Ex4 + Fx5} (4) dt G where, A (a-a )clap B = (b-a-2ab+a2)cpaW -VA- (l-a) K,D 2(2ac-c 2bca 2, 2 cx VA D = (2ac-c-2bc+b-)b + c.A. + K

-6oE = (2bc-c2)Ca F - -c2opW G = {k2V+k3[(1-a) + (a-b-l)x + (b-c)x2 + cx3]aw} V

APPENDIX C SAMPLE CALCUTATIONS Raw Data, Run 51 Product Samples Analysis (wt. %) Weight Collected @ 10 minute intervals H 0 IPA G 86.26 13 74 26.35 gm. H 86.o6 13. 94 30 09 gm. I 85.98 14 02 26,47 gmo J 85.98 14 02 28.29 gm, K 85.90 14. 1o Average 86. 01 13.99 27,80 gn. Feed Sample Analysis 89 90 10.10 Reactor Bed: 0,2745 eq. Dowex 50WX8, 32-42 mesh, Average Temperature: 130. 9~ C Gas Produact Rate:'0.00248 ft. /min. Gas Product Temperature: 76, 20 F Liquid Product Temperature: 24. 3 C Calculations The average rate is calculated by the approximations F (xf xi) N x 10.10/60. 09. 03259, mole fr. IPA in feed. 10o.o0 10/60.09 + 89 9/18. 02 x 1399/60.09 -.. 04651, mole fr. IPA in producto 135 99/60. 09 + 86.01/18,02 -61

-6227.80 (13 99/60.09 + 86.01/18.02) = 13916 g-moles/ 10 100 minute, molal feed rate. (o. 13916 )(o. o4651 # 0. 03259) 5 r (= l1)(- 04651 03 =2 7.057 x 10 3 gmwmoles/eq.-min. 0 2745 This rarte is the uncorrected reaction rate based upon the li-quid-phase measurements. A small amont of the product is carried away by the propylene flashed off in the liquid separator and a correction is made in the rate for this lost product. The composition of the vapor is assumed to be propylene with water and isopropyl alcohol in equilibrium with the liquid product. Vapor pressure of IPA @ 24.3~ C = 41 4mm Hg. Vapor pressure of H20 @ 76.20F F 23.1mm Hg. Volume rate of prod. gas @ STP = 2.48 x 103 x 492 x 74 536.2 760 = 2.22 x 10- ft. 3/min Volume rate of dry prod. gas @ STP = 2.22 x 103 x (1- 1 ) = 2.15 x 103 ft. /min. 454 Molal rate of dry prod. gas = 2,15 x 10 x -- 359 = 2.72 x 10,3 gn-moles/min, Alcohol rate in prod. gas = 2. 72 x 10o3 x 41-. x 0 04651 720 7.28 x 10(6 gm-moles/min. Alcohol rate/eq. in:prod. gas 7.28 x 10"6 x 1. 0,2745 2. 65 x 10-5 gm-mole/eq. wmin. Rate = (7.057 + 0.027) x 10-3 = 7.084 x 10'3 gm-mole/eq..min.

-63The rate is also corrected for temperature difference. This data point was used for a correlation at 1300C, but the average measured temperature was 130.9~ C. The measured rate was multiplied by a correction factor f = exp[(-E/R)(l/To - 1/T)] where E is the experimental apparent activation energy. For this run, f = exp[(-21,600/1.987)(1/403.1 - 1/404.0)] = 0.942, Rate = 7. 084 x 10-3 x 0.942 = 6.673 x 10-3 gm-moles/eq-min. The usual technique of graphical differentiation was used to differentiate the integral data to obtain point rates. The data were differenced as shown in Tables XII - XV on pages 40 and 41, and the differences plotted as shown in Figure 22. The smoothed curves were drawn and point rates obtained.

-6440x0-3 - 1..... 50x1 - GROUP VIII, 140 C,1440PSIG, GROUP IX, 1500C 100-150 MESH 1440PSIG 100-150 MESH 36x1I0 I3 46xI03 32x10 3 42 x103 —.1 N/F, EO-MIN/MOLE N/F, EQ-MIN/ MOLE 28 xl 3 0 38x10 x — i z L GROUP X, 160C,1440 PSIG, GROUP X,140C'lZ 1 \ 1 1 100-150 MESH61 24x10 34x10 I Uj 0 3_-_,3 -3 2 OxI0. ---- 30xO10 ----- - 0 0 4 6 8 10 12 14 0 I 2 3 4 5 I| ~ N/F, EO-MIN/MOLE N/F, EQ-MIN/ MOLE 8 o0xl0'i\ — I I - Figure 42x2hical Dif iai ^Z U \ GROUP X, 1600C,1440 PSIG, GROUP XI,140~C <, o o,100-150 MESH 1440 PSIG -3 \.o.oc.o s. -3 32-42 MESH 64x10 36x10 48x10' 28xle.- -, 32x10,'- 20xlo? \ 16x0 0.4 0.8 1.2 1.6 2.0 2.4 2.8 0 2 4 6 8 10 N/F, EQ-MIN/MOLE N/F, EQ-MIN/MOLE Figure 22. Graphical Differentiation

APPENDIX D EXPERMENTAL PROCEDURE An alcohol-water mixture of approximate composition was prepared from Reagent Grade Isopropyl Alcohol and distilled water from the East Engineering Building distilled water supply. Before each run this mixture was charged into the water gaging cylinder (A), shown in Figure 1l page T and flushed through the pump. This mixture was pumped through the reactor system for approximately an hour at a rate of approximately 4 -cc/minute. Propylene was charged into its storage cylinder (F) by allowing liquid propylene from an inverted cylinder to displace the mercury back into the oil cylinder (E) and oil to a collecting vessel at atmospheric pressure. Propylene pressure was increased from vapor pressure to reactor pressure by pumping oil into the displacement system with the valve in the propylene feed line closed. The preheater controller was set to the desired temperature. All heaters were turned on and full voltage was applied to heater "TO" for quick heat-up. Reactor temperature was measured periodically, and the voltage to heater "0" was reduced as the desired temperature was approached. Heater currents had been predetermined and only minor adjustments were necessary. After the alcohol-water mixture had flowed for sufficient taime the pump was adjusted to a feed rate calculated to yield a 2-5% change in alcohol concentration. After the reactor reached operating temperature and the propylene reached operating pressure, the propylene feed -65

-66valve was opened and its feed rate was adjusted. The reactor was allowed to come to equilibrium for a period in which at least one free reactor volume of liquid passed through it. During this time, checks and adjustments were made to keep the temperature stable. At the end of the equilibration period the run was begzu, The temperatures of the feed systems1 product streams, and the reactor were measured and recorded. The liquid product receiver was emptied, the wet test meter reading was recorded, and an electric timer was started. Independently, the levels in the oil and water gaging cylinders were recorded with the starting of a stopwatch. Four liquid samples were collected at equal time intervals after the start of the run (eig. at 5, 10 15, and 20 minutes after time 0). The time interval was chosen to yield 20-40ce liquid samples. At the time of collection of the last sample, the wet test meter read. ing was again recorded. Reactor temperature was read at 10 minute intervals during the run. This was done by means of a traversing copper-constantan thermocouple. The readings were taken at 2 1/2" intervals throughout the reactor length. The reactor temperature was then calculated as a time-length average for the ruw The liquid samples were weighed and analyzed by a refractive index technique described in Appendix E.

APPENDIX E CHEMICAL AIALYSIS OF SAMPLES Gas chromatograms of preliminary run samples showed the presence of only isopropyl alcohol, di-isopropyl ether, and waters An attempt to devise a quantitative analytical technique for this system using gas chromatography was not successful. Analysis of this system was devised by Brey. (7) The technique used two of three physical properties (refractive index, density, and viscosity) of the mixture, depending upon the range of composition encountered. Brey measured all of these properties accurately for standard mixtures throughout the entire spectrum of compositiono Since the product mixtures obtained from this study were limited to 0 to 50% isopropyl alcohol and low concentrations of ether1 a modified refractive index technique was used with Brey's data. This technique involves the measurement of the refractive index of the sample, titration with di-isopropyl ether to saturation1 and ameasurement of the refractive index of the saturated sample. From the phase diagram on which the refractive index contours have been plotted. the composition of the sample may be determined. The isopropyl alcohol concentration was determined with high accuracy on three standard samples with this technique, and it was adopted for the chemical analysis of the samples. In the range where the isopropyl alcohol concentration approaches'40-50o the accuracy of this technique is reduced because the rate of change of refractive index with ether addition is small; -67T

-68i. e, the refractive index of the saturated sample is not much di~fferent from that of the sample itself although a large amoamnt of ether has been adeiedo Only a small number of samples fell into this range. Figxre 23 gives an example of the use of this technique.

-69i/\/ / \ /. /e\, 0,\o o i o 0 0.l ILI 0 wo0 0 L~0 0 ~o ~, ~~~ I — PI 0, 0 c ~I U) o~~ a, O u,~~~~~~~~~~~~~ O,~ 0 O e~~~

APPENDIX F ESTIMATE OF ERROR Calibrations The traversing thermocouple was calibrated against a standard thermometer in the range of temperatures used. Pressure gauges used in the system were calibrated with an Aminco Dead Weight Pressure Tester. All temperatures and pressures used herein are the corrected values, Measurements The sources of error in the experimental results are due to two groups of measurements: a,) the measurements used for the calculation of the rates, and b.) the measurements used for the correction of the rates to the desired conditions. An estimate of the error of each of these measurements yields an estimate of error of the corrected rates. Measurement Maximum Error %Error a. Product rate + 0.02 gm + 0.10% Composition + 0.20 wt,% + 10% Resin capacity + 0.01 meq/gm.(wet) + 0.43% b. Gas rate + 0.01 ft.3 + 0,03% Product temperatures + 0.50F + 005% Barometric pressure + 20mm Hg + 0. 09% Reactor temperature + 0.1~ C + 1. O0o0 The best method would be to obtain an error estimate for each run; however, one run was chosen which gives a maximum estimate. The items of group (a) were used to calculate the rate. Preliminary calculation showed that maximum values of F and Ax and a minimum value of N yielded a maximum error estimate. Choosing Run 79, where F = 0.322 gm-moles/min. + 0,000322 gm-moles/min. Ax = 0.0135 + 0.00135 -70

-71N - 0.2745 eq. + 0. 00118 eq. Rate - - F 15.83 x 10 gm.-moles/eq.-mino N / N V (Fb) + (-Axf) P = probable error = N Where, + n error of N + 5 = error of x + f = eror of F + 15 83xlO3 L18xlO0 3)2 (.322xl.35x10O 3)+( L35xlO02x322x10 4)2 P 0.2745 p =+\/l893 x 10'8 + 4 o55 x 10 + 158 x g mes/e mio 0.2745 0.2745 Group (b) measurements have an effect only on the correction of the rate to correlation conditions, The correction for the alcohol lost in the vapor may be written as follows: C = G - X A' B D - x where, G = gas flow rate - 1.65 x 103ft.3/min. + lo65 x 10"5 X = conversion term =4. 49gm-moles/ft.3 (includes pressure correction) A = water correction term o1 - Oo 9635 + 0.001 B = temperature cojrrection term = 0 909 (correction to STP) p D A approximation of Henry' s Law Constant *- A o 0.624 + 070016 x mole fraction of alcohol in liquid product 00135 + 0.00135 P probable error a XB. (ADxg)2 + (GDxa (GAdx) (GAD)2

-72P = + 4.49x.909 (1.339x10-8)2 + (1.39x10-9)2 + (3.43xlO-8)2 + (1.339x10l7)2 P = + 4.49x. 909g 1. 926x1 014 + 4.49x. 909x1.9x lO-7 = + 5.67x10-7 gm-mol/ eq. -min. An estimate of the maximum error due to temperature correction is made by substituting the expected temperature difference into the correction factor calculation. The difference in correction factors multiplied by the rate yields the probable error. t = 140.9~ C f = exp[(-21,600/1.987)(15. - 1)] = 0.944 1 1 f' exp[(-21,600/1.987)(41 1 - il)] = 0.950 P = + 15.83 x 10-3 x (0.950 - 0.944) = 0.095 x 10-3 The probable error due to all of the experimental readings is due to the combination of the errors of the calculated rate plus the errors of the corrections. P = total probable error = + (1.585)2 + (0.0006)2 + (0.095)2 x 10-3 P = + 1. 586 x 10-3 gm-moles/eq. -min.

APPENDIX G HEAT OF REACTION The heat of reaction is calculated. from the heats of forma. tion of the products and reactants at 25~ C H20 (1) -- 2 + 1/2 02 + 68.3174 kcal/mole C3H6(g) ---- 3C + 35 - 4.879 kcal/mole 3C + 4H2 + 1/2 02 —-C 3 70H (1) - 74.32 kcal/mole C3H6(g) + HO0 (l)- i-C3H70H (1) 10.88 kcal/mole The heat of reaction at 1000C is calculated by adding the heat required to raise the prodicts and reactants from 25~C to 100~C, A_ = AC dT The C data are not available for all of the compounds; therefore, AJH for each compound will be evaluated by the best method available and the sum of the AH's calculated.. 1. C3H6(g)(250C) -- C5H6(g)(10ooC).2 120 (l)(250c) - o (H)(loo~c) 2< H^2O (1)(25~C) - 20 BO(1)(100~ C) 3. i-C3H70OH (1)(25~C) — i-C3i 70H (1)(100 C) 1. AH ( T )1 T1T ( ~)2 T (frcm generalized charts) AS = (0.3) (537) - (2.4)(672) 161-1612 = -1451 Btu/lb-mole A..- 1451/1.8. -806 cal/gm-mole 2. Al = H.-H2 (from steam tables) AH = 45.02 - 180.07 =:135.05 Btu/lb AE -135. 05 x 18402/1.8 = -1352 cal/gm-mole -73

.743. AH C (ave) AT 0.654x-75 x 49 cal/gm = -49x60o09 p A -2942 cal/gm-mole H M= -Ap r = 2942 - (1352-1451) = 139 cal/gm-mole Hp at 100~C = -10.88 + 0. 14 = -10.74 kcal/gmimole At 100~C, Rate = 1 x 10`3 gm.-moles/eq.-min. The reactor contains = 0.25 eq. of catalyst Heat evolved by reaction 1 x 10'3x10.74 x 1053x0,25 = 2. 69 cal/min. 24 69x60/860 57 f 0.187 watts

-75APPEMIX H CORREIATION OF DATA The corrected data were grouped according to the reactor conditions and a three^coxstant equation was fitted to each group of data by the least-squares technique. The mean d.eviation of the data from the correlations was calculated. and. the data which deviated. by more than 2.5a were discarded, and the correlation process repeated. From. Figure 16, the apparent activation energy is 21,600 cal/gm-mole. Each of the correlating constants, A through C, was divided by exp (-21,600/RT) yielding a new set of constants, A' through C', which correspond to A through C of Equation (4), For the data at 450 psig, these constants were then correlated as functions of temperature in Figure 24, and the temperature functions substituted into the rate equation r = exp(-21600/RT)t{A + +B + C'tx } (21)

-76-,dI r, i r- rl, 0 0 0 0 0 0 o o - r- o o\ 9.X M. S O O O r- O r - 0 0 0 0 H 0 H Hl H H H H H4 Hi S o 0 0 0 0 0 0 b H H H H H r-I H H n 0 C n r-' H \D CO 0 -4. LC\ \-O - LH I C R I IB I I I -4. r1 o3 c5o 0 0 o o oN oN oN N r-. aO r 0 0 00 0 0 0 X C r r- ll 4 C r0 — t n i - t 0 0 co 0 0 0 0 0 0 0 Nr-i 0 jrX ^,(\ XC\,x Cx 04 t0 c 0 H 0 1 l B I I -. 1 1 O ~~ C O O Q O O r H O H H 8 8 8 8 8 8 8 H H.- C\1 I- ON i IO o O bP1 Ct1 H tCC'V --- -. 4-P \-C) \IO 0 n 0 0 0 0 4 X XD~~~0 0 0 o + 0 ~ 0 o a, - c-.3 ON - L L o 1 -n ON 1 L\ C-O n co C\1 H * K * O X o n o&lt —4.- i04 H rA- 04 0o \D.- O H -N H H C - J u 0t 04 9 0' 0 o H \9 Ln 4 — a n > C\ (Y'd oC 0. n \ 0o O LO C\ 4. 0 0 ON \- - H H H H- H HK O H1 H4 H r H H ^H 0q,- -~- - C

-77AL, BI, and C were obtained by fitting the best least squares curve'with two or three constants through the points shown in Figure 24. A2, B2p and. C2 were obtained assuming a linear relationship to temperature and placing a straight line through the points as shown in Figure 25. The resultant rate equations are: at 450 psig r " exp(.021,600/RT) {(-57.94 + 0.666t - 0.0026t2) 109 + (2.059 - o0o6517t) - l10 + (1.952 + 0.00957t) 10 x2} gm-moles/eq. -min. at 1440 psig r = exp(-21,600/RT) {(l.68- 0.005t) 1010 - (4.32 - 0.025t) 10 l1x + (20.98 - 0.122t) 1011x2} gm-moles/eq. -min. Figure 26 shows the experimental rate vs. the rate calculated by both general correlations. All the 98 runs used for obtaining the correlations are plotted. There is no apparent trend away from the diagonal indicating the validity of the rate expression.

-786x10 2 x 10 ________ <4xlO9....... --- 100 110 120 130 140 t,~C I0 4 x 10 100 110 120 130 140 t, ~C 4 x o10' 02x 100 110 120 130 140 t,~C Figure 24. Correlating Constants, 450 psig.

-79-.9 IxlO~ 0.9x10'~ - 130 140 t,~C 1.2x 10" - 1.1 xlO" i1.0 xIO 0.9x10" 0.8x10ld 130 140 t,o 6.0xl0 5.0x1 i 4.OxIO - 130 140 tC Figure 25. Correlating Constants, 1440 psig.

-80x x %%%X I n J 0 0 Ix~x \ o -o x I 0 X X\X\ do x x\ -, X \ x w ~ " S o I IIdX3j I 0 0 0 0 N X X Xx \ \:~ ^ 0 0 5~ 0 x0xx\ x NxWJX- S -.X' x x< XX xx ^ mX X d d d 0'NIVA -'03/S3q0I - AIMVI9''dX3j

NOMENCLATURE A through G constants Cp heat capacity, cal/gm-~C E activation energy, cal/gm-mole F molal feed rate, moles/min. H enthalpy, cal/gm-mole K chemical equilibrium constant N equivalents of resin in reactor P atmospheric pressure, mm Hg R gas constant T absolute temperature, ~K v molal volume, liters/mole a, b c constant s f error of F k reaction rate constant n error of N p partial pressure, mrriHg r reaction rate, gmo-moles/eq.-min. t temperature, OC; time x mole fraction a phase distribution coefficient 6 error of Ax Subscripts A alcohol P propylene -81

-82W water f f inal i initial p products r reactants s surface s s. steady state v volumetric

BIBLIOGRAPHY 1. Anonymous. Chem. and Met. Eng., 23, (1920) 1230. 2. Astle, Melvin J., and Pinns, Murray L. J. Org. Chem., 24 (1959) 56-60. 3. Azarnoosh, Azarpoor, and McKetta, John J. J. Chem. Eng. Data, 4, (1959) 211. 4. Barker, G. E., and White, R. R. Chem. Eng. Prog. Symp. Series, No. 4, 48, (1952) 75. 5. Bauman, W. C., and Eichhorn, J. J. Am.. Chem, Soc., 69, (1947) 2830. 6. Bent, Franklin A., and Wik, Simon N. U.S. 2,036,317 (April, 1936). 7. Brey, W. J. Anal. Chem., 26, (1954) 838-42. 8. Chambers, Robert R. U.S. 2,803,667 (August, 1957). 9. Cottle, Delmer L., and Young, David W. U.S. 2,810,759 (October, 1957). 10. Dale, C. B. Ph.D. Thesis, The University of Michigan (1955). 11. Douglas, W. J. M. Ph.D. Thesis, The University of Michigan (1958). 12. Friedman, Bernard S., and Morritz, Fred L. U.S. 2,830,091 (April, 1958). 13. Hamilton, G. E., and Metzner, A. B. Ind. Eng. Chem., 49, (1957) 838. 14. Ipatieff, V. N., and Monroe, G. S. J. Am. Chem, Soc., 66, (1944) 1627. 15. Katzen, R. Pet. Ref., 40, (1961) 161. 16. Keith, Carl D. U.S. 2,805,261 (September, 1957). 17. Kelly, J. T. U.S. 2,843,642 (July, 1958). 18. Klein, F. G., and Banchero, J. T. Ind. Eng. Chem., 48, (1956) 1278. 19. Krep s Saul I., and Nachod., Frederick C. U.S. 2,477,380 (July, 1949). 20. Kressman, T. R. E. Ind. Chemist, 36, (1960) 3-8. 21. Langer, Arthur W., Jr. U.S. 2, 861, 045 (November, 1958)o -83

-8422. Leffler, J. E. The Reactive Intermediates of Organic Chemistry, p. 137, Interscience Publishers, New York, 195b., 23. Levy, Norman, and Greenhalgh. Robert K. U.S. 21683,753 (July, 1954). 24. Lukasiewicz, Sigmund, J., Denton, W. I., and Plank, Charles J. U.S. 2,658,924 (November, 1953). 25. Marberry, J. E. Ph. D. Thesis, The University of Michigan (1960). 26. Majewski, F. M., and Marek, L. F. Ind. Eng. Chem., 30, (1938) 203. 27. McDOnald, D. W., and Hamnerg W. F. Ind. Eng. Chem., 48, (1956) 1621. 28. McDonald, D. W., and Hamner, W. F, Ind. Eng. Chemo 49, (1957)o 1518. 29.. McDonald, D. W. and Hamner, W. F. Ind. Eng. Chem., 52, (1960) 962. 30. Muller, J., and Waterman, H. J. Brennstoff - Chem., 38, (1957) 357. 31. Reed, L. M. Wenzel, L. A., and O'Hara, J. B. Ind. Eng. Chem., 48 (1956) 205. 32. Reynolds, Peter W. and Pittwell, Laurence R. U.S. 2,7555309 ( July, 1956). 33. Robinson, Ralph L. U.S. 2,769,847 (November, 1956). 34. Runge, Von Franz, Bankowski, Otto, and Hoffmann, Gerhart Brennstoff - Chem., 34, (1955) 330. 35. Saletan, D. I., and White, R. R. Chem. Eng. Prog. Symp. Series No. 4, 48 (1952) 59. 36. Sherwood, Peter W. Pet. Engr., 28, (1956), C33& 37. Sliepcevich, Co M., and Brown, G. G, Chem, Eng. Progress, 46, (1950) 556. 38. Sliepcevich, C. M. Ph.D. Thesis, The University of Michigan (1947), 39. Smith, Carl D. U.S. 2,797,247 (June, 1957). 40. Stanley, H. M., Youell, J. E. and Dymock, J. B. J. Soc. Chem. Ind. 23, (1934) 205. 41. Sussmann, Sidney, Ind. Eng. Chem., 38, (1946) 1228.

-8542?. Teter, John W., Gring, John L., and. Hettinger, W. P., Jr. U.S. 2,830,090 (April, 1958). 43. Wegner, C. U S. 2,876,266 (March, 1959). 44. Young, David W. U.S. 2,813,908 (November, 1957)