Separation of ethylbenzene from para-xylene by azeotropic distillation by Richard Lee Nelson A thesis submitted to the Graduate Faculty in partial Fulfillment of the requirements for the degree of MASTER OF SCIENCE in Chemical Engineering Montana State University © Copyright by Richard Lee Nelson (1962) Abstract: This investigation was performed to study the separation of ethylbenzene from para-xylene by azeotropic distillation. The apparatus consisted mainly of a fractionating column with a Corad condensing head, a refractometer, and a gas chromatograph. The entraining agents were chemical compounds that were available at a reasonable price, that boiled from about 40°C below to 20°C above the boiling point of ethylbenzene, and that contained active oxygen, nitrogen, or halogen atoms since compounds having these atoms are likely to azeotrope with ethylbenzene. For each entraining agent a mixture of the entraining agent and ethylbenzene was charged to the column and distilled. The overhead product, which should be the ethylbenzene-entrainer azeotrope, since all azeotropes formed were minimum boiling azeotropes, was analyzed to determine the azeotropic-composition, The usual method of analysis was reading the refractive index of the sample and then determining the composition from a previously prepared plot of refractive index versus composition. The charge to the column was then adjusted so that there was enough entrainer to azeotrope with all of the ethylbenzene present./Also, about as much para-xylene was added to the charge as there was ethylbenzene. This charge was distilled at total reflux for at least three hours. Then samples of the overhead and bottoms were analyzed for percent ethylbenzene and para-xylene on the gas chromatograph so that relative volatilities could be calculated using the Fenske equation. If the entrainers did not azeotrope with ethylbenzene, or if the azeotropes were less than 10% ethylbenzene* the entrainers were not used for a relative volatility determination. This method for determining azeotrope compositions seems to be reasonably good. The azeotropic composition between ethylbenzene and cellosolve found by this method at 640 mm. of mercury was 59 weight percent ethylbenzene. The literature(8) gives this azeotropic composition at 735 mm. of mercury as 56.7 weight percent ethylbenzene. The relative volatility between ethylbenzene and para-xylene using no entrainer was 1.037. The relative volatilities obtained from the best four entrainers were: 2-methyl butanol, 1.079; methyl isobutyl carbinol, 1.075; n-hexyl amine, 1.073; and methyl amyl alcohol, 1.073. The ratios of the number of plates needed to obtain 95% separation in the overhead and bottoms products using an entrainer to using no entrainer were .48, .50, .51, and .51, respectively for these entrainers.  O I SEPARATION OF ETHYLBENZENE FROM PARA-XYLENE BY AZEOTROPIC DISTILLATION by RICHARD L, NELSON A thesis submitted to. the Graduate Faculty in partial fulfillment of the requirements for' the degree of .MASTER OF SCIENCE .in Chemical Engineering dor Departments Chairman? Examining C'oimnf,Wee Graddafte Division MONTANA STATE COLLEGE Bozeman, Montana- December? 1,96.2 '■v//,;.■W iii ACKNOWLEDGEMENT The author wishes to acknowledge the staff of the.chemical engineering ,department of Montana State Collegey and in particular the research advisor Dr. Lloyd Berg, for their encouragementy sug­ gestions;, and criticisms which have been helpful to the completion of this research.project. The author also wishes to thank the National Science. Foundation fo,r their financial assistance. I ' ? ( iv TABLE OF CONTENTS Abstract...... ........... ...... ........ .............. . vi Introduction and Theory ............................................ .I Research Objectives............. ................ ............ 5 Equipment....................................................... 6 Chemical Compounds.................. .................... ...........9 '' Operating Procedure.......................12 Discussion of Results ......... ............... ............... .15 Conclusions.................... .19 Sample Calculations ............... ^20 Appendix........ ..... ...... ...... ...... ................22 ..37Literature Cited.. o o o e e e o *00»0«»0»« V LIST OF TABLES AND FIGURES Fî JU-IsS I e DiS-^IaSIll -Of App S Ia 3-f IlS eee6ooeeo»6oe»»»oooooooeeooooeo, oeo23 .Figure 2. Chromstogrsph Sepsrstion Using 7^.8-benzog[uinoline column,... .......... .. ..............24 Figure 3..Chromstogrsph Separation Using l^chloronaphthalene c o l u m n - . »25 Table I. Structures and Boiling Points of Entrainers............26 Table II. Azeotropic Composition of Ethylbenzene with. Entrainers ............... .......... . 30 Table III. Relative Volatilities for Entrainers Aiding Separation...................*o..*.......,.....35 Table IV. Relative Volatilities for Entrainers nbt Aiding Separation..................................35 Table V. Ratio of Number of Plates Nebded, Using Entpainer to Using no Entrainer for 95 # Separation.............. 36 Ti • ABSTRACT This investigation was performed to study the separation of ethyl,- benzene from, para-xylene by azeotropic distillation. The apparatus consisted mainly of a fractionating column with a Corad condensing Jaeadjt a refractometer, and a gas chromatograph,. The entraining ,agents were chemical .compomds that. were available at a .reasonable price* that boiled Trom about 40 dC below to. .20 0C above the boiling-point of ethylbenzene* and, that contained active oxygenj, .nitrogen* or halogen atoms since compounds having these atoms are likely to azeotrope with ethylbenzene. For each, entraining agent a mixture .of the entraining agent and' ethylbenzene was ,charged to the column and distilled. The overhead product^ which should be the ethylbenzene^ent.rainer. azeotrope* since all azeotropes formed.-were min­ imum boiling, azeotropes Sl was analyzed to determine the azeotropic composition. The usual method, of analysis was reading the refractive index of the sample and then determining the composition from.a previously prepared plot of refractive index versus composition. .The charge to the column was then adjusted so that there was enough entraine.r to. azeotrope with all of the ethylbenzene present ./Also* about as much para-xylene was added to the charge as there was .ethyl-' benzene. This .charge was distilled, at total reflux, for at least three hours. Then samples ,of the overhead and bottoms were analyzed for percent.ethylbenzene and para-xylene on the ,gas chromatograph so that relative.Volatilities could, be calculated, using the Fenske .equation. If the entrainers ,did not azeotrope with ethylbenzene* op if the. azeotropes were less than IO^ ethylbenzene* the entrainers were not used for a .relative volatility determination., . This method for deter­ mining.azeofrope.compositions seems to be reasonably good* .Theazeo- , tropic ,composition, between ethylbenzene and cellosolve found by this method at 64-0 mm..,of mercury was 59 weight.percent ethylbenzene. The literature(§.). gives this azebtropic composition at. 735 mm, ..of mercury as 56.7 weight percent' ethylbenzene. .The .relative volatility between, ethylbenzene end para-xylene using no entralner was 1.037 • The relative volatilities obtained fpom. the best fopr entrainers were5 Bumethyl butanol-j I „079.3 methyl i isobutyl carbinol* I,.075} n-hexyl. amine* I „073.} and methyl amyl aleOhpl* I.O73. ..The ratios.of the number of plates needed.to obtain 95^ separation in the,overhead-and bottoms products using an. entralner to using no entralner ,were .4-8* „50* .51* and. .51,. respectively for these entrainers. INTRODUCTION AND THEORY The purpose of this investigation ,was to study the ease ,of Z Ethylbenzene and.pararxylene are two. important.commercial chemicals Ethylbenzene is used to make styrene^ and, para-xylene is used, to make terephthalic acid, and plastics,. The present method, of separation of para-xylene and. ethylbenzene from an ethylbenzene-xylene mixture is quite .costly,. Pararxylene is crystallized from a solution at about -63 with, a yield,.of only about 65 percent. Ethylbenzene is then separated by distillation from, the-remaining .xylenesy. which includes s-ojiie para- xylene,. The distillation columns used for this separation reqtiipe a ■ large number' of theoretical plates^ This study was undertaken to if the separation.could be accomplished using,a smaller number of plates by employing.azeotropic distillation. Distillation is a process .of separation based on the difference in composition between a liquid, and the vapor formed from it. ,Since the difference in composition between a. liquid and the yapoh formed is Somewhat dependent on the difference in boiling points of the.,constit­ uents^, compounds with close boiling points such .as ethylbenzene and ' para-xylene are.difficult to separate by straight -distillation. .For this reasonji' azeotropic distillation was selected as one method, to separate these two compounds;., The method, selected, to measure the separation between ethylbenzene 2 and para^xylejae .was their- relative volatility. Th;e volatility ,df .a^ ■compound is a measure of its tendency to enter the vapor phase-* and, is equal to the mole fraction of thp component in the vapor phase \ divided, ,by the mole fraction of the component in the liquid phase." / The ease .of separation of a binary mixture by ,fractional, distillation \ depends ,on the relative volatility.of one.component to the other | component.jThe purpose of azeotropic distillation is.to change the relative, volatilities,of the components present. . Sinde all.systems for this study were homogeneous the discussion that follows .will be for homogeneous systems .only. An azeotrope.is a constant boiling mixture and. is usually the result of deviations from ideality caused by hydrogen bonds or internal pressure. A minimum boiling, binary azeotrope is a mixture. of the two. pure compounds that has a.boiling point which is less than the boiling.point of any other mixture containing the pure components. -In the sense that an azeotrope .will distill, .with no. change in, c OmposItionj,'it acts like a .pure .compound,. .However^ the composition of azeotropes .rarely corresponds to- any simple formula.^ and the azeotropic composition changes if. the pressure of distillation changes i. A maximum.boiling azeotrope is a mixture Of the two. pure compounds that has .,a. boiling, .point which. i$ greater than the boiling point .of any Other,mixture containing the two pure components. .No maximum boiling azeotropes were encountered in this study ̂ The reason .there were none is discussed, In the section CjHEMICAL COMPOUNDS, 3 Azeotropic distillation has been used quite successfully, to separate different classes of compounds „ Howeyer5,. the separation, of isomers or homologues generally presents a ,more difficult problem. One example- of this type of separation is between 2,6-lutidlne, J^picolines..-a-nd 4 -picOline(6) using acetic acid which forms a maximum boiling.azeotrope .vfith. each, of these .compounds „ The equation used to calculate the relative volatility was the I Eenske equation ,which can be written as follows: n+l % x e where o< = relative volatility =i number of theoretical plates in column = percent ethylbenzene in.overhead = percent para^xylene in overhead =s percent para-xylene- in bottoms =S percent ethylbenzene in bottoms This formula .results from applying the.definition of relative volatility to each, theoretical plate, in the column and. to the still-pot Which, acts as one theoretical, plate. 1 .> The Jpethod of analyzing samples to. obtain quantitative values for the overhead aqd bottoms ,.-compositiotis was a gas ,.chromatographs Gas chromatography uses a gas as the mobile phase. The Solute travels 4 through, the.column, as a plug.,of gas which Is partially dissolved. In^ df adsorbed on, the stationary phase. A sample of the. mixture to be analyzed is flash-evaporated,, at, one end of the column., %he time required for each component to emerge from the column 16 known as its retention time. This retention time is characteristic of the compound. .The chromatograph used had a thermal conductivity,cell that detects each component as it emerges from the column. A continuous recorder records each component as. a peak ..on the recording.paper. If two compounds have equal .thermal conductivities but different retention times ̂ the areas under, the peaks for each of the compounds should be proportional to the volume percent of each compound in the sample. . This method was used in this study since the thermal .conductivities of. ethylbenzene and para-xylene should be nearly equal. The accuracy.of this method is.evaluated, in the DISCUSSION OF RESULTS section. RESEARCH OBJECTIVES The main, objective of this project, was -to effect a separation between -ethylbenzene and para-xylene .using.azeotropic distillation, that would require a . smaller number .Of plates than would be, reeded to.effect the same separation using straight distillation. . The sOpar- ation. obtained.was to be measured by relative volatility between, ethyl­ benzene and para-xylene for each entrainer used. A-secondary objective was to determine azeotropic compositions between ethylbenzene and. various entraining agents. EQTJIBMEflTT The equipment used ..in this study -consisted .of .a fractionating ,.column* Corad condensing head*, still-pot* electrical heater j ref rap- t.ometer* Westphal .balance, Ohaus balance* Christian Bdeker chainomatic balance*, and .gas . chromatograph .with, a Minneapolis-Honeypell recorder,. The fractionating ,column, .was made. from, three concentric, glass, cylinders .each 43 inches high. The diameters of the three cylinders -were 1.0* I,.75* and 2.5 inches* respectively. The inside cylinder contained.4 5 inches of randomly packed*.. Oneeeighth inch stainless steel helice packing (Fenske rings). - supported .on a .cone-shaped -wire support. The middle column was wound with a Nicrome heating,coil so that the column could be operated, nearly adiabatically. The heat sup­ plied was controlled-with, a .Variac. connected to the heating ,.6011.. A mercury thermometer was attached to the outside of the inner column. The. inner .column had.ground, glass joints at top and bottom, so that the Coradcondensing head,.and the still-pot Could be. connected..,easily but tight enough to prevent any vapor from.leaking.from the column to the air.. The still-pot, used, was a one-liter flask with a. side arm for sampling .and. a thermometer wQH-- ■ ,This still-pot .Pas. heated .with a small .electric heater connected to a Variac. -.The W r l a c was used to control the heat input to the column. . A Corad,.condensing head was used to condense the vapors rising from. 7 A the fractionating column. Using this head,* reflux .ratios .of 2„5 to Ix 5 to I* 10 to lx. 20 to I, and. .30 to I could, be obtained. This head Was used.at total..ref lux .except when overhead samples were taken. Duririg this time*.a 30 to I .reflux ratio was used. A diagram -Of this .equipment is shown ih Figure I. An.OhaUs laboratory balance scale was used to. weigh the amounts of the various components charged to the column-. .In ,determining .azeotropic compositions*.a.Valentine.refractometer and a Westphal balance were used.. A. constant temperature water, ,bath was used to keep the temperature of the samples in.the refractometer at 20 degrees Centigrade. An Aerograph, gas .chromatograph made by the Wilkens Instrument ■ and Research Company was used to analyze the overhead, and botterns prod,-- ucts for percent ethylbenzene and. para-xylene. Two packed..chromatograph columns were. used. . One was a 1/4-inch copper, tube* 12 feet longX; packed With 60-9-0 mesh, Chromosorb P acid-washed packing*. containing 20 percent of 7>8-benzoquinbline .substrate. .This column was operated at 100 0C with a. helium, flow rate of about 9.0. milliliters per minute.. .The separ­ ation.obtained.between ethylbenzene and. para-xylene fop this column is shown in. Figure 2. . The. other column was a l/4-inch. copper - tube* 8 feet long* packed with 30-6.0, mesh Chromos orb F 15. percent of l^chloronaphthalene substrate. . This .column ,was .operated at acid-washed, packing* ,.containing i X- , about 50 0C with a helium flow rate of about 175 milliliters jser minute. The separation obtained between ethylbenzene and para-xylene for this column is shown in. Figure 3. A Minneapolis-Honeywell recorder was used with the chromatograph. A Christian Becker precision chalnOmatic balance, was used, to weigh the pieces of paper representing the. areas unler the: peaks from, the chromatograph. CHEMiCAL COMPOUNDS Chemicals, used, for this investigation were the variousentraining . agents* ethylbenzeneY and para-xylene., The ethylbenzene and para-xylene were obtained, from Phillips Petroleum ,Company. , These- compounds were, not less than. 99 mole percent pure^. and. were not redistilled. All liquids can be divided into the. following fiye classes .of compounds:(2) 1. .^Liquids capable of fqpming three-dimensional,.networks..of strong hydrogen, bonds. . - 2. . Other liquids composed of molecules ,containing both active hydrogen - atoms and. donor atoms. 3. Liquids composed.of molecules ,containing.donor atoms but no active-hydrogen atoms. 4. Liquids composed of molecules containing.active hydrogen atoms but no donor atoms. 5. All .other liquids." Ethylbenzene and pararxylene are class 5 liquids. Any liquid that, azeotropes with a .class 9 compound, will.always give positive deviations from Raoult-'s law- which causes minimum boiling ,azeotropes . . The entrainers were selected for. the following reasons t- I. be available at a reasonable price,. 2.. boil close to the boiling ppint of ethylbenzene.* 3. contain- oxygen^ nitrogen^, or halogen atoms , or. groups .containing these atoms. ■ The boiling points of the entrainers were, usually from 40. 0C below 1,0 to 15 0C . above the boiling point of ethylbenzene. Hdweyefj, ..a few entrainefs wene tried that boiled below this range to check the effect this would have,on the relative volatility. Hntrainers that were water soluble and that did not azeotrope with water would be desirable fop separation of the entrainer from the ethylbenzene^ but the 'number of compounds having these properties in this boiling, range is -Small. The following .list shows the entrainers that were used fop this ■ investigation* 2-methyI butanol 2/6^dimethyl moppholiue methyl is 0butyl. c arbinol isobutyl c-arb-inol n-hexyl amine acetic acid X n^butanol ethyl, ohlopo acetate amyl, acetate N-methyl piperazine 5.-h ex en~2» on e dimethyl, .ethanolamine methyl propyl. curbInol 2 j4-pentanedione 1-n.itro propane lj}-diamino propane methyl isoamyl ketone n-propanol 2-ethyl butanol n-amyl alcohol 2 -methyl pentanol 2 ohlopo ethanol 2-nitpo propone nltpoethane diethyl carbinol N-ethyl morpholine methyl .cplofo acetate methyl amyl.acetate dimethylainino propyl amine -dimethyl isoprOpanolamine 11 is0butanol I^ardioxane pyrrole ethyl butyl ketone cellosolye- ethyl butyrate methyl cellosolye p-fluoro toluene propylene.diamine isobutyl acetate ,Zrpicoline methanol ethylene diayiine methyl ethyl ketone morpholine methyl isobutyl ketone allyl propionate Mrmethyl morpholine amyl fQrmate 2-me thy I piperidine ' secondary butyl alcohol piperazine tertiary butyl alcohol nrpropyl propionate chlorOrE-pp'opanone pyrazine I^J-jIimethyl butyl amine tetrachloro ethylene Table I gives the structures . and boiling ,'points of these • entfainers at 760 .mm.. of ,mercury. . Most of the entrainers were .obtained fropu Union Carbide Chemicals Company^ Eastman Kodak Company^;. and. Fisher Scientific Company. Theae-. entrainers were redistilled to .remove high and low boiling, impurities 7 .. OPERATING PROCEDURE During this investigation^, .nearly-tii'e same .experimental, procedure was used for each entrainer. Therefore^ this section .will present the general procedure used* with detailed..descriptions of ,any variations in this procedure. . The. first step in .this, .Investigation was to determine azeotropic -compositions Petwden the entrainer and ethylbenzene. A rough, estimate of the azeotropic composition was made, from, the difference in boiling points ,of the. entrainer and, the ethylbenzene. A-mixture .of this . compos. ' sition. Was then charged, to the .column• Distillation at a 39 .to I reflux .ratio was .used until the distillate temperature became constant. Since the azeotrope Is the lowest boiling material* it should be the. first material to come -Off as ,.distillate. A .distillate sample, at this reflux, ratio was then taken. The distillate sample was then analysed, to determine azeotropic composition,., for nearly all of the ent r a i n e r s the refractive index of the entrainer was much lower than the refractive index.of the ethyls benzene. If so, the refractive, index of. the overhead from.the column (the azeotrope) was found using the refrac t Ometery and. the . compos it ion of the azeotrope w,aS .determined, from a. .previously prepared plot of refractive index versus.composition. In. some cases, the refractive index of the entrainer was too close to the refractive index of the ethylbenzene. Two methods.Wppe used, to • 1> determine the azeotropic composition of these samples. .Some .of the azeotropes were -.run through the .chromatograph and analyzed by measuring the- areas, under the peaks, as described in more detail la.t.er in. this section for ethylbenzene and para-xylene. .However., since the thermal ■ conductivities,Of the entrainer end the ethylbenzene might, be quite ,different^ - samples .of known, .weight per bent ethylbenzene and entrainer • Were run on the ,chroma to graph.. .Then a pint of actual-weight percent versus ,chromatograph area percent was. made. From the area percent for the azeotrope, the - azeotropic composition was then. • determined, from -this plot. This method was used.for the following .entrainersf 2tpIcoline^. p-'fluoro toluene, pypazine and pyrrole,. . The .other method used when the refrac.tiye indices were too close pas to,measure differences in densities with the W^stphal balance. .This was used.on only one. entrainer. .This,.entrainer was tetrachloro ethylene which, did. ,not azeotrope above 10 percent ethylbenzene, -After the azeotropic composition had. been, determined^ enough, ■additional, .entrainer or ethylbenzene W,as added to the column so that all.of the-.entrainer would azeotrope with all of the ethylbenzene. Enpugh para-xylene pas added, to the .column so. that the amount of para-, xylene was about equal to the amount of ethylbenzene in the.column. However, the exact amount of papa-xylene added is not important since the overhead and bottoms samples were analyzed on the chromatograph. DiStillation, at a reflux ratio, -of 30. to I was used .until the. temperature of the oyerhead. became constant at the temperature of the azeotrope, .The Corad.condensing, head, ̂ as then turned to infinite reflux, (all .distillate returned to the ,o-oluim).. . The .column was then .operated at total reflux for at least three hours „ At this timê .-. a.-.one milliliter sample of the oyerhead•and bottoms were taken. The reflux ratio used. Was 30 to I to keep the column as close to total ..reflux as possible. The Samples of the overhead and. bottoms products were then injected into the gas chromatograph with the column containing the 7v.8̂ benzo*. ■ quinoline substrate for analysis. If the entrainer did. not come out at^the same jtime as the ethylbenzene...or para-xylene^ the areas under the, peaks ,for ethylbenzene and para-xylene were analyzed by tracing the peaks on a piece of paper. . These tracings .were then cut from the paper arid weighed on ■ the Christian Becker Chaindmatie balance.. . The areas under the respective peaks were taken as a measure of the relative amounts of ethylbenzene and para-xylene present. If the entrainer did come out at the same time as the ethylbenzene or para-xylene using, this .column, the column containing the !-.Chloronaphthalene substrate was used. . Howevery, the separation on this column was not quite as good as the column containing the 7*8-benzoquinoline (see Figures 2 and 3). .Since the chromatograph columns did not completely separate ethyl­ benzene from.para-xylene^ the lines.were extrapolated as though, only : 'one .component had been present ds Shown In Figures.2 arid 3» DISCIISSIQII OF RESULTS The method used. In this inyestlgation to determine azeotropic, composition is probably not the most accurate method since the column used cannot completely separate ,a homogeneous azeotrope from a pure component. Howeyery If the boiling point of the.ethylbenzene-entrainer azeotrope is more than a few. degrees below the boiling point of the ethylbenzene^ nearly complete separation should be obtained.. The literature(8) giyeS the azeotropic composition between ethylbenzene and cellosolye as £>6.7 weight percent ethylbenzene at 735 mm.... of mercury A value determined by the method used for this investigation gave a value of ;59 weight percent ethylbenzene at 640 mm.. of mercury. Considering, the., difference in pressure^ : this • agreement se'ems (juite good. When the azeotropic composition was .determined to be less than 10 percent ethylbenzene^- the entrainer' was not used for a .relative vol­ atility determination. The main reason these entrainers were not used is that the possibility of one.of these.entrainers making such a. separ­ ation economical .is small since sQ much ■ entrainer would have to be. used for a small amount of ethylbenzene. AlBey If the azeotropic composition appears to be only thrd'e .or four percent ethylbenzene^ it is possible that the separation, in the column was not good, enough to separate completely the two ■ pure c o m p o n e n t s T h i s means that the compounds , might not have even azeotroped... The methods of analysis- fop azeotropic compositions were reasonably accurate. . The temperature of the refractometer was ,.controlled 16 ' , • - at 2.010.5 C. ,The .refractive. Index method Is probably the most accurate method of analysis used. For the chromatograph* five samples of known concentrations of ethylbenzene^para-xylene mixtures, were .analyzed, on the. .chromatograph,. The average error for these determinations was ,one percent. '.However* since the ethylbenzene and, entraincrs probably had, .-different thermal conductivities* plots, of actual weight percent entrainer to chromatograph, area percent entrainer were made. .Since this introduces another source of error* the error in these analyses •was probably between one and two percent. Although using, the Westphal balance to determine densities should.give reasonably accurate azeotropic compositions* the accuracy of this method is not important in. this report since the only system .that ,was. analyzed, this way did. not seem, to indicate an azeotrope between, the entrainer and the ethylbenzene,. ' . The entrainer for this ,system, (tetrachloro ethylene) has a.-density of. 1.623 gm/ml while the ethylbenzene has a density of 0 .867 grr/ml. .The relative volatilities found, in this investigation should, be .close to# put not precisely equal to*; actual relatiye volatilities since, the exact number of theoretical plates,was not known. .The. value used.in. this.report* 23- theoretical plates^ should be close to the ' actual value. However,.for comparing the various entralners* the calculated ralatiye,volatilities does giye a valid rating,of the entrainers except for errors In measurements as discussed in the next paragraph. .Table IH lists the relative volatilities for the entrainers. aiding separation, while Table IVilists the relative volatilities for Xl the entrainers that 4id not enhance the separation. I l s ^ the ratio, of percent of plates needed using an en,trainer to using no- entrainer (Table V) should, be correct even though .the number of theoretical plates ,may not be exact. The following ,discussion of error will deal only with error assuming the column, was found .experimentally to- have 23 theoretical plates. The average - error in area percents from .the chromatography as determined from.known samples^ was one percent, in error of one percent in overhead and bottoms samples could have caused an error in the relative volatility of 0.004- units. Alsoy several factors . such as boil-up rate in the column can change slightly the number of theoretical plates in. the column.- A -Change .of one theoretical, plate (:as from 23 to 24- theoretical plates) would change the relative volatility by Q.003 units. As an example of the reproducibility of results# two separate trials using no entrainer were performed. .The relative volatilities for these runs were 1.035 and. 1.039. - For ,calculations the average of 1.037 was used. Two correlations of relative volatility pith some property of the entrainers were examined. The first# boiling point# did not seem to have any correlation with relative Volatility. This can be seen, from the boiling points listed in. Table X which has . entrainers -arrange in decreasing order..of relative.volatility. The second correlation examined was the type of compound. .Several of the best entrainers' X 1.8 : were alcohols. Ho^ewr, several..of the other alcohols either did not azeotrope or gaye much smaller .relative volatilities. Correlation of alcohols by boiling points also, did not seem, to "show any meaningful trend. CONCLUSIONS It is possible to effect a .given separation, between.ethylbenzene and para-xylene with about one-half as .many theoretical plates using certain entraining, agents as by straight distillation. / SAMPLE CALCULATIONS Calculation of relative volatility Penske. eqiiat Ion vjhere o< = relative volatility A.~ number' .of theoretical -plates in coluim. y e = .percent ethylbenzene in overhead y-p = percent para^xylene in. overhead . Xp = percent, para-xylene in bottoms Xe = percent ethylbenzene in bottoms Sample calculation, using Zrmethyl butanol run:- ye .= -.9 . 7 8 5 mole, fraction ethylbenzene - yp =Q. 2 1 5 mole fraction para-xylene. Xp = 0.627 mole, fraction para-xylene X 6 = 0.575 mole fraction ethylbenzene = 1.079 Calculation, of number Of plates needed using.entrainer divided by number .of plates needed by straight distillation .to obtain 95^ . separation at total reflux. ĉ e - relative volatility using 2 -methyl ,butanol (I.O79) 2,1 ° Fractionating Column Still-Pot----> Heater Figure I. Diagram of Apparatus 24 ethylbenzene peak------ para-xylene ----peak line time Figure 2. Chromatograph Separation Using 7 ,8-benzoqulnoline Column 25 ethylbenzene para-xylene base line time Figure 5. Chromatograph Separation Using 1-chloronaphthalene Column 26 STRUCTURES M D BOILING POINTS OF ENTRAINERS TABLE I .E NTR AINER STRUCTURE BOILING POINT 0C 2 -methyl butanol CH^Ch 2CH(C1H5)CH2OH 12& methyl isobutyl ,carbinol (Ce 5)2Ch c r 2Ch o h c r 5 131.4 n-hexyl amine CE5 (GH2 )5NH2 132.7 n-butanol CH5CH2CH2CE2OH 117.7 amyl, acetate CH5COQ(CH2)4CH5 148 5-hexen-2-one CH2=CHCH2CH2COCH5 129,5 methyl propyl.carbinol CH5CE2CH2CHOHCH5 119.3- 1-nitro propane CH5CH2CE2NO2 132 methyl isoamyl ketone CH5COOH2CE2OH(CH5)2 144 2-ethyl butanol (C2H5 )2CHCH2OH 149.5 2-rmethyl pentanol CH3(CH2)2CH(CH5) CH2OH 148 2-?nitro propane CH5CH(NO2)CH5 120. diethyl .carbinol CH5CE2CH(OE)CE2CH5 115.6 methyl chioro acetate Ch2 CICOOCH5 131.5 27 TABLE I (conH) BOILING .E NTRHMER ,STRlfiGTURE PQINT °C dimethylamlno propylamine (CH3)2NCH2CH2Ch 2NH2 134.9 2^6-dimethyl morpholine NHCH2CH(CH5 )OCH(CH5 ) CH2 146.6 isobutyl carblnol (CH5)2CHCH2CH2Ofl. .1 3 9 .5 acetic acid CH5COOH.. 118.1 ethyl. Chlop1O acetate CH2ClCOOCfl2CH5 144.2 N^methyl piperazine CH5NCH2CH2MHCH2CH2 1 3 8 .0 dimethyl ethanolamine (CH5)2NCH2CH2OH 134.6 2 -p ent ane dione CH5COGH2COCR5 139 H 3^diamino.. propane NH2OH2GH2OH2MH2 1 3 5 .5 n-propanol CH5CH2CH2OH. 9 7 .2 n--amyl alcohol CH5CH2CH2OR2OH2OH 138 2-iChloro ethanol ClCH2CH2OH 128.8 nitpoethane CH5CH2NO2 • 114.8 N^ethyl morpholine O2H5NCH2CH2OOH2CH2 1 3 8.3, methyl amyl acetate CH5COOC.H.(OH5 )■ CH2CH,(OH5) £ ■ 146.1 28, TABLE I (.Con1J t). ENTRAINER I BOILING POINT 0G dimethyl Isopropanolamlne (CH=XgNCHgCH(Og)CH, !HS.? isobutanol (CH5)2CHCH2OH . 108.4 pyrrole NHCH=CHCH=CH 131 . cellos olye- CgHcOCHgCHgOH 135.1 methyl cellosol.ye CH5OCH2CH2OH 124.3 propylene diamine CH3CH(NH2)CH2NH2 119 .2-picollne N=C(CH3)CH=OHCH=GH 128 ethylene diamine NH2CH2CH2Nfl2 116.1 morpholine OcH2CH2NHCH2CH2 126^30 ally! propionate C2H5COOCH2CH=CH2 — . amyl formate HCOO(CH2)4CH3 130,4 secondary Isutyl alcohol CH5CH2CHOHCH3 99,5 tertiary butyl.alcohol. (CH3)5COH 82.8 chloro^ “Propan one CH3COCH2Cl 11,9 1^3-dlmethyl. ,butylamlne. CH3CH(CH3)CH2CH(Cfl3)NH2 .Crf-C= 29 TABLE. I (con.it). BOILING EN.TR ADJER STRUCTURE POINT 0C 1.,4-d.ioxane Oc h2CR^OCH2CH2I _ :_ • I 101.5 ethyl butyl ketone CgE^COfCHgi^CR^ 148-. 5 ethyl butyratp CH3CH2Ch 2COOC2H 5 121,3 p-fluoro .toluene C%CH=CHCH=CH(F), CH=CH 117 lsobutyl acetate GH3COOCH2CH(CH3)2 116.5 ,methanol CH3OR 64.6 ,methyl ethyl ketone CH3COC2H5 79,6 methyl lsobutyl ketone CH3COCH2CH(CH3)2 ' 119 N^methyl morpholine CR3NCH2GH2OCH2CH2 115.6 2-methyl piperidine .Ch 3CHNHCR2CH2 CRg CH2 119. piperazine NHCE_CH_NECH_GH_I C 2 ' 2 1 S 145 n-propyl proplonate CR3CR2COOC3H7 123.4 pyrazlne N=CRGHlNCH=OH I I 118- tetrachloro ethylene CCl2=CCl2 121.0. C 30. AZEOTROPIC COMPOSITION QE ETHYLBENZENE WITH ENTRAINERS ENTRAINER AZEOTROPIC COMPOSITION TJfflLE II 2-mefchyl ,butanol methyl is.obutyl carbinol n-hexyl amine n-butanol amyl acetate 5-hexen-2-one .methyl • propyl, carbinol 1- nitro propane methyl .isoamyl .ketone 2- ethyl butanol 2-methyl pentanol 2-nitro propane ,diethyl carbinpl methyl, ,chi,oro- acetate dimethylaminO propylamine 2^6,-dimethyl .morpholine isobutyl ,carbinol acetic acid ethyl .chloro.acetate N -methyI piperazine dimethyl ethanplamine (weight % ethylbenzene) 4.0 48 1.6. 27 9Q 35 .22 .44 M . 82 So 13 14 34 ■43 78. 5P 30 75 50. . 54 31 TABLE XE (AonH) .EJJTR AIKER 2 Apfentanfedione l>.3.-diamlnp. propane n-propanol n.-amyl alcohol 2-^chloro ethanol nitroethane K-ethyl morpholine methyl amyl acetate - d imethyl Is oprop anolamine isobatanpl pyArPle .cellos ol ye methyl cellosOlye propylene diamine 2-plcollne ethylene diamine morpholine • ally! propionate* amyl formate* secondary butyl alcohol* tertiary butyl alcohol* chloro«2—prop anone* AZEOTROPIC COMFOSITIOK (weight. % ethylbenzene). 63 70. .11 47 13 55 91 I2 ■ 1.6. 51. . . 59 47 40. 23 22 30 32 TABLE II (conlt) ENTRAIMER 1*3 rdimfe^hyl Mtylaminfe* lj4-dioxane* ■ ethyl butyl ketone* ,ethyl butyrate* p-fluoro toluene* lsobutyl ..acetate* methanol* methyl .ethyl ketone* methyl lsobutyl ketone* N-methyl mprpholine* 2-methyl piperidine*. p Iperazlne* n -propyl propionate* pyrazlne* tetrachlhro ethylene* AZEOTROPIC COMPOSITION (weight %. ethylbenzene.) *azeotrope% if formed.^ is less than IO^ ,ethylbenzene. TABLE. Ill b e l a t i v e v o l a t i l i t i e s FOR EBTRAINERS AIDING SEPARATION ENTR AINER RELATIVE VOLATILITY R^jnettiyl Hutanol . 1.079 .methyl isobutyl carbinol 1.07% n-hexyl .amine 1.973 n-~butandl 1.969 amyl acetate 1.067 5-hexen-RL-one I..Q$6 methyl propyl ,c arbinol 1.965 1-nitro propane 1.065 methyl .isoamyl .ketone 1 .9:60 2-ethyl butanol 1,059 2-methyl pentarjpl 1.459 2-nitro propane I.059 •diethyl .carbinol I.057 methyl .chloro" acetate I.057 dimethylamino propylamine 1.056 2;f;6^dimethyl morpholine 1.05N- isobutyl .carbinpl 1.953 acetic acid. 1.453 ethyl chloro acetate 1.052 N-methyl piperazine I.05Q dimethyl ethanolamine 1.049 TABLE .III (-c-onlt) .EBTR AINER'-I t '< ‘ RELATIVE, VQLATILITy 2 >4 -p entane diene 1,.046. 1,3-dlamlno propane 1.943 n-propanol 1 ,043 n^amyl .alcohol 1 .041 B-uPhloro ethanol i,o4Q nltroethane 1.039. no entraineh 1,4354 1,9)9 35 TABLE IV RELATIVE VOLATILITIES FOR ERTRAIRERS ROT AIDIRG SERARATIQR ERTRAIRER RELATIVE VOLATILITY no enlralnb'r 1.935, 1-939 R-ethyl morpholine I.P37 methyl ^jnyl acetate 1,937 dimethyl Is oprop Anolamlne 1.032 isobutanol - 1.Q30 pyrrole 1,028 -cellos olye 1.024 methyl .cellosolye 1.024 propylene-diamine 1.023 2-pi-coline 1.017 ethylene liamlne 1.016 morpholine 1.012 36 TJffiLE V RATIO OF .NUMBER OF PLATES NEEDED USING.ENTRAINER TO TJSING NO ENTRAINER F O R ^ SEPARATION E NTR AINER RATIO. R̂ jne.tJiyl .butanol „48 methyl Isobutyl -carb-lnol .50 n^hexyl amine .51 n-butanol 54 amyl,. acet ate „56 5-hexen-.2-one . 57 methyl propyl .carblnol ' " „58 1- nitrp propane .59 ■ methyl igo.amyl ketone „6.2 2- ethyl butanol „63 2-!methyl pentanol .63 2-nitre propane „ 6.3 .diethyl carblnol „65 methyl chloro. acetate. „6.5 dlmethylamlno.. propylamine , 67 2y6-dlmethyl morpholine „6.9 * 37 LITERATURE ,CITED 1 ., Berg^. L .. and Harrisonj, J,x. '^Evaluation of Azeotropic Entrainersylr -Ehemic al ,,Engineering Ryogres s > Vol., 43, Sept., 1947, ppy .487-496. 2. Berg, L.^ ,Harrison, J.y and, Ewell, R.,. "Azeotropic Distillation," .̂ -Industrial .and. Engineering ,.Chemistry, Vol. '3,6, 1944, pp .,8711-75. .3. Burchfield, H,., and Storrs, E,.,& Biochemical Applic.ations cf, Gas Chromatography, Academic. Press, Hew York, 1962. 4 . . Carlson,. Carl S »> and others. Distillation, Interscience Publishers, Inc. Hew York, 1951. 5. Carney, Thomas P ., Laboratory Fractional Distillation, The Mapmillan-' Company, Hew. York, 1949. 6. Coulson, E. A., and Herington, E. F. G.% Laboratory Distillation- Practicey Interscience Publishersy Inc., Hew York, 1958. ' 7. . Handbook. of ,Chemlsiry and P h y s i c s 4lst edition. Chemical Rubber Bublishing. Company, Cleveland, Ohio, 1959-60. .8. .Kieffer, ¥. ,F.y and, Grabiel, C. E.y "Azeotropes of 2 .-Ethoxyethanol With Alkyl Benzenes -industrial .and. Engineering Chemistry, Vol. 4"3, April, 1951, PP. 973-5. . 9. Hoebels>- Henry J., Wall, R. F., and Brenner> H.x Gas ChromatQgrsphy, Academic Press,. Hew-Yorky 19.61. 10.. Ehyslc;al,Properties ,of Synthetic Organic Compoundsy 1959 edition. Union Carbide Chemicals Chemicals Company. MONTANA STATE 762 1001 5079 4 N378 N337 cop .2 Nelson, R. L. Separation of ethylbenzene from para-xylene by azeotropic distillation I N A M E A N D A D D R E S S , 2 , — = 2 — I u / ' / ( " V i . r 5 ^ ^ / / f - j ' ■ " < / 7 23 /2- 37 - L S I