Isomerization and Dehydrocyclization of 1,3-Pentadiene

University of Central Florida STARS

Retrospective Theses and Dissertations

Summer 1979

Thomas E. Marcinkowski University of Central Florida

Part of the Chemistry Commons Find similar works at: https://stars.library.ucf.edu/rtd University of Central Florida Libraries http://library.ucf.edu

This Masters Thesis (Open Access) is brought to you for free and open access by STARS. It has been accepted for inclusion in Retrospective Theses and Dissertations by an authorized administrator of STARS. For more information, please contact [email protected].

STARS Citation Marcinkowski, Thomas E., "Isomerization and Dehydrocyclization of 1,3-Pentadiene" (1979). Retrospective Theses and Dissertations. 433. https://stars.library.ucf.edu/rtd/433 IS0~1ERIZATION AND DEHYDROCYCLIZATION OF 1,3-PENTADIENE

TH0~· 1AS E. MARCINKOWSKI B.S., St. Leo College, 1977

RESEARCH REPORT

Submitted in partial fulfillment of the requirements for the degree of Master of Science: Industrial Chemistry 1n the Graduate Studies Program of the College of Natural Sciences at the University of Central Florida; Orlando, Florida

Sununer Quarter 1979 ABSTRACT

Pipcrylene concentrate is ~ complex mixture of 5-carbon

unsatur.1tcd hydrocarbons obtained as a by-product when naphtha

or gas oils are cracked. The major component in this mixture is

1,3-pentadiene.

During the course of this study, a number of trials, utilizing

liquid phase reaction conditions, were made to investigate the

geometric isomeri:ation of 1,3-pentadiene and its separation from

the piperylene concentrate.

Isoiaeri:ation \\·as accomplished employing catalytic amounts of

iodine at temperatures ranging from 0°C to reflux. Using this

method. tl1e maximum amount of trans-1,3-pentadiene obtained was

70~ as compared to 51% in the piperylene concentrate. Recovery of

the proJuct \\·as 909o, \vi th the remainder being diiodo compounds and

polymer. Isomerizations employing catalytic amounts of potassium

tcrt-butoxide were also investigated. Using this anionic isomeriza

tion, the theoretical amount of trans-1, 3-pentadiene (84~6 @ 20°C)

\..;as obt~ined in the product. The greatest dra\-.rback with this

technique \vas the low recovery (50°a), Jue to the extensive polymer

format .ion.

Successful separation of 1,3-pentadiene from the mixture was

accomplishct.l through cuprous chloride complexing. Utili:ing this

technique, SL~~o of the 1, 3-pentadiene was recovered from the mixture, with the separated product being 99.9% pure 1,3-pentadiene. Separ~tion of trans-1,3-pcntadienc from the mixture was accomplished through a Diels-:\lJer reaction \oJith maleic anhydride. Since this dienophilc will react readily with trans-1,3-pentadiene but not ci.s-1,,:)-pent:ldiene:, this method offered an easy and efficient means of removing the former isomer from the mixture. In attempting to reverse this Diels-Alder, via pyrolysis, many products were obtained; incluJing those present in the original mixture.

The ·~~s phase dehydrocyclizati_on of l, 3-pentadiene \\'as investigated in a 316 stainless steel tubular flow reactor utilizing various heterogeneous and homogeneous catalysts. The selectivity to cyclopentadiene was greatest (60%) in the presence of a hydrogen sulfide promoter. For all other catalysts, the selectivity remained relatively constant L30~). This constant selectivity over a wide range of par~meters indicates that a significant amount of competing side reactions are prevailing within the preheater section of the apparatus. ACKNOIVLEDGE~·IENTS

TI1e author wishes to express his appreciation to Dr. Guy

~·Iattson for his personal quidance and profound patience which he exibited during the course of this project; Dr. Chris A. Clausen and Dr. John T. Gupton for their suggestions and encouragement leading to the completion of this work; and Dow Chemical USA,

Louisiana Division, for making this research possible through indirect support.

The author \vould also like to extend his thanks to his father for his support throughout the authors lmdergraduate and graciuate years.

iii CONTENTS

Introduction 1

Uses and Outlook for Various Components in the Piperylene Concentrate 4 1,3-Pentadiene 4 Cyclopentene 5 Cyclopentadiene and Dicyclopentadiene 6 Chemical Separation of 1,3-Pentadiene From Piperylene Concentrate Using Sulfur Dioxide 9 Separation of 1,3-Pentadiene From Pipery1ene Concen- trate By Cuprous Ammonium Chloride Cornplexing 11 Separation of Trans-1,3-Pentadiene From Piperylene Concentrate Via Dicls-Alder Reaction With ~·laleic Anhydride 13 Thermodynamics 16 Isomerization of 1,3-Pentadiene 17 Dehydrocycli:ation of 1,3-Pentadiene 19

Experimental 23

Analysis 23 Analytical Standards 25 Liquid Phase Reactions of 1,3-Pentadiene 26 Halogen Catalyzed Isornerizations 26 Base Cataly::ed Isomerization 26 Separation of 1,3-Pentadiene From The Piperylene i'·lixturc By Cuprous Ammonium Chloride Treatment 27

Synthesis of 3-~Iethyl-1,2,3,6-Tetrahydrophthalic Anhydride Via Die ls-Alder Reaction With ~laleic Anhydride 29 Pyrolysis of 3-Nethyl-1,2,3,6-Tetrahydrophthalic Anhydride 30 Pyrolysis Procedure 32 Vapor Phase Isomerization and Dehydrocyclization of 1,3-Pentadiene 32

lV CONTENTS (cont.)

Paaeb Catalyst Preparation 36 Operating Procedure 38

Results and Discussion 40 Separation of 1,3-Pentadiene From Piperylene Concen- trate Using Cuprous Ammonium Chloride 40 Iodine Isomerization of 1,3-Pentadiene 40 Potassium tert-Butoxide Isomerization of 1,3-Pentadiene 45 The Synthesis and Pyrolysis of 3-Methyl-1J2,3,6-Tetra- hydropht!1alic Anhydride Sl Thermal and Catalytic Dehydrocyclization of 1,3-Penta- diene 53

Conclusions 92

References 94

v LIST Or TABLES

Table Title __Pacreb_

l Typical Composition of Piperylene Concentrate 2

II Some Properties of Tne Common Sulfones 10

III Rate Constants For Various Diels-Alder React

lons of Several Dienes With Maleic A.nhvdride.; 15

IV nesu1ts of A Dehydrocyclization Experiment

Performed Bv Hutchincrs75b 21

Composition By Weight of Chemical Sample Co~pany Piperylene 27

VI Data For The Separation of 1,3-Pentadiene From Piperylene Concentrate Through Cuprous Chloride Complexing 41

VII Data For The Iodine Catalyzed Isomerization of 1,3-Pentadiene 44

VIII Data For The Potassium tert-Butoxide Isom erization of 1~3-Pentadiene 47

IX Products Obtained From The Pyrolysis of 3-~lethyl-1, 2, 3, 6-Tetrahydrophthalic Anhydride 52

X Constant Dehydrocyclization Reaction Parameters 55

XI Composition By Weight of The ';Pure'; Component Feeds 57

XII Dehydrocycli::J.tion Products Obtained From TI1e ''Pure" Component Feeds 58

XIII Retention Time And Identity of Products Found In Effluent Strean 60

XIV Results of 1be Thermally Initiated Dehydro cycli:ation of Piperylene In A Packed ru1d Unpacked Reactor 61

Vl LIST OF TABLCS (cont.)

Title __Pacre.:>_

\\' Cracking Data Obtained In A Clean And Coked Reactor 66

X\' I Results of The Hydrogen Sulfide Pro~oted Dehydrocyclization of Piperylene 69

XVII Results of The Dehydrocyclization of Piperylene Over A Sulfided Stainless Steel Bed In The Presence and Abscence of Sulfur Dioxide 80

X\'TII Results of The Dehydrocyclization of Piperylene Over Silica Gel And Alumina Catalyst Beds 85

vii LIST OF FIGURES

f igtlTC Title __Paaeo_

l The ~·1orc Comr:1on Chiaro Cor1pounds of Cyclopentadiene 8

Pyrolysis Apparatus 31

3 Vapor Phase Isomerization-Dehydrocyclization Apparatus 33

4 Proposed ~1echanism For The Iodine Catalyzed fsomeri:ation of 1,3-Pentadiene 43

5 Possible Side Reactions During Iodine Catalyzed Isomerization of 1,3-Pentadiene 46

Proposed ~·Iechnnism For The Potassium tert Butoxide Catalyzed Isomerization of 1,3-Pentadiene so - I Effect of Temperature On The Conversion of Piperylene And Selectivity To Cyclopentadiene In A Packed And Unpacked Reactor 63

8 Effect of Temperatur-e On The Conversion of Piperylcnc And Selectivity To Cyclopentadiene In The Presence of Nitrogen And Steam Dilutents 65

9 Effect of Contact Time On TI1e Conversion of Piperylene And Selectivity To Cyclopentaciiene Over A Stainless Steel Bed 67

10 Effect of Hydrogen Sulfide And ~itrogen On The Conversion of Piperylene And Selectivity To Cyclopentadiene At Various Temperatures 71

11 Effect of Hydrogen Sulfide And Steam On The Conversion of Piperylene And Selectivity To Cyclopentadiene At Various Temperatures 7'2

1.2 ~lcchanism For The Hydrogen Sulfide Promoted Dehydrocycli:ation of 1,3-Pentadiene 74

13 ~lcchanism For The Thermally Initiated Dehydrocycli:ation of 1,3-Pentadiene 75 viii LIST OF FIGURES (co~t.)

Title

Effect of Hydrogen Sulfide Concentration On TI1e Conversion of Piperylene ~nd Selectivity To Cyclopentadiene 76

15 Effect of Contact Tioe On The Conversion of Piperylene And Selectivity To Cyclopentadiene In The Presence of Hydrogen Sulfide 78

16 Effect of A Stainless Steel A~d-A Sulfided St~inless Steel Bed On The Conversion of Piperylene And Selectivity To Cyclopentadiene In Relation To Contact Time 82

17 Effect of Te~perature On The Conversion of Piperylene And The Selectivity To Cyclo pentadiene In The Presence of Sulfur Dioxide 83

18 Effect of Contact Time On The Conversion of Piperylene And The Selectivity To Cyclo pentaJiene Over Silica Gel 87

19 Effect of CoTJtact Time On Tne Conversion of Piperylene And TI1c Selectivity To Cyclo pentadiene Over Aluoina 88

20 Proposed Carbonium Ion Mechanism For The Cyclization of 1,3-Pentadiene Over Alumina so

ix INTRODUCTIO)J

Tremendous quantities of ethylene and propylene are produced eacn year .for use as starting materials for a Hide variety of petro- che:nicals. These were, and still are, produced by the thermal cracking of condensate from natural and refinery gases. But due to the decreasing production of natural gas and the increasing raw material restrictions ln the United States, there is a trend towards the usc of heavier naphthas and gas oils as cracking stock 1 .

Due to the large amounts of ethane present in the lighter gases, relatively small amounts of by-products are formed on the pyrolysis. But when the heavier naphthas and gas oils are cracked, large amounts of by-products are produced. Of particular interest to this project is a by-product stream consisting of a mixture of

5-carbon unsaturated hydrocarbons. The major component in this distillation fraction is 1,3-pentadiene, known commonly as pipeT)'- lene; hence, the name "piperylene concentrate" has been given to this mixture. A typical composition of piperylene concentrate is shown in Table I.

The amount of piperylene concentrate produced 1s dependent

? on the feed employed-. In general, the heavier the naphtha or gas oils, the greater the amount of piperylene formed. Also, the more severe the cracking conditions are, the greater the diene content.

The total production of the 5-carbon stream from cracking of naphtha and gas oils has been estimated to be approximately five 2

TABLE I

Trpi.cal Composition By \\'eight Of Piperylene Concentrate

Component Percent

2-mcthyl-2-butene 2.7

2-rnethyl-1-butene 5.2

isoprene 14.4

cyclopentcnc 24.9

trans-1,3-pentadiene 22.5

cis-1,3-pcntadicne 16.1

Remainder* 14.2

* consists mainly of c4 to c6 unsaturated hydrocarbons, benzene,

dic;:c1opcntadiene and other heavy materials in varying amounts. 3

I million tons per year-. With this amount being produced, it is incvit3blc that the piperylene concentrate fraction will become an imnort~nt tJ.ctor in the economics of the petrochemical industry.

Despite the large volume of piperylene concentrate available, relatively few commercial uses have been developed 3 . At present, the pipcrylene concentrate is used to produce resins which find end usc as tackifiers for various adhesives.

A co~~crcial epoxy hardener is obtained through a Diels-

Alder reaction with maleic anhydride. The acids and esters of this adduct can be used as plastisizers and softening agents for resins, 4 plastics, gums, and lacquer films . ~·Ialeic anhydride can also be copolymeri:ed \vi th piperylene to encorporate the very reactive anhyJride groups into the polymer5 . The excess amounts are being used ~s fuel for combustion furnaces and as a gasoline blend feed- stock. These uses are not economically desirable, since the fuel value of the piperylene c~ 6.~¢/#) is much less than its resin value c~ 9.1¢/~).

The main drawback of the piperylcne concentrate is the large amount of impurities it contains. These impurities have a depre- ciating affect on the quality of the hydrocarbon resin products.

It is therefore desirable that these be removed in order to nroduce superior products.

It has then been the main objective of this project to improve the quality of the piperylene concentrate. Emphasis was placed on removal of the alkenes and the isomerization of cis-1,3- 4

pentaJicnc to trans-1~3-pentadicne. Both of these changes will subst:antially upgrade the hydrocarbon resins produced from the conccntr:J.te. Another quality improvement sought was the cyclization anJ Jcllydrocycli:ation of tl1e 1~3-pentadiene to produce cyclopentene and cyclopentadiene respectively. Both of which have a hig!1er value than the piperylene concentrate.

Presently, few industrialized schemes are being utilized to produce ?urer 1,3-pentadicne from the piperylene concentrate. lioh·cvcr, it is widely known that the Japanese are using the GPB and GPI process to produce high quality 1,3-butadiene and isoprene6 .

~ippon Zcon Company is employing this technology to produce highly 2 pure cyclopentadiene and 1,3-pentadiene in high yields

Uses and Outlook for Various Components 1n the Piperylene Concentrate

The c!1cmistry and the applications of the major components in ? 6-26 the piperylene concentrate have been reviewed-' . These reports indicJ.te the importance of separating the components of the p1pery- lene concentrate.

1,3-Pcntadiene

Polymers of 1, 3-pentadiene \vere first prepared by J. TI1iele

?'J in 1901 using hydrogen chloride and heat._.:.. Currently, polymers and co-polymers with 1, 3-butadiene are prepared \vi th transition metal catalysts to produce elastomers. Two resins now produced .... from pure 1,3-pentadiene are Quintone and QuintolL.

Quintone is a resin used as a tackifier for pressure sensitive adhesives, hot melt adhesives, and in rubber compounding. This 5 material has been Jescribcd as "synthetic polyterpenerr since it has .:haractcristics sirnil::1r to polyterpenes with regard to color, odor, stJ.bility, creep strength!' etc. Quintal, which is a liquid po lyr::8 r o i 1, 3-pcnt ::1diene is \'ery similar to linseed oil. Due to its quick drying properties, it can be used in paints, coatings, sealants, and caulking COQpounds.

Cyclopcntene

Cyclopentenc has promising \:ormnercial potential for use in the rubber industry. This is due to the similar properties that the vulcani:ed polymer of cyclopentene has to cis-polybutadiene

,1 and cis-polyisoprene- •

These unsaturated polymers of cyclopentene are known as pol)~cntenamers and are produced using transition metal catalysts tnroug1, 1 a r1ng. open1ng . mec1an1sm 1 . 21' 27 . Originally, it was thought to proceed through the typical sigma bond rupture, but this was ?Q proven false by Calderon- . He and his co\vorkers showed that the polymers h·ere formed by a double bond interchange, or olefin 28 metathesis, typified by the follo\·Jing

,CII .~ .. ( CI L,) -- ~ W - n~CH CH-----HC---.__ ( CH ) :::::-- \\T CH ) ~ I / ( 2 2 n --...... 'CH-----HC '/' n ,;..--~--- -- 6

The most LOQmon catalysts presently used are tungsten and ~olybdenum

comt)lexes such as tungsten hexachloride with a trialky1 aluminum.

T!1es e c~t alysts arc stereospeci fie. Tnngsten yields the trans -1,5- polypen~cnamer and molybdenum, the cis-1,5-polypen~enamer.

Cyclopentadiene and Dicyclopentadiene

Dicyclopentadiene is the dimer obtained from cyclopentadiene on standing or heating. It deconposes to cyclopentadiene when heated to 1G0°C. In terms of general chemical reactions, cyclopentadicne IS a very versatile compotul.d. It owes its chemical reactivitv to three structural characteristics: the double bonds, their conjugation, and the activated methylene group. Although very versatile, cyclopentadiene still has few industrialized applications.

The chemistry of cyclopentadiene has been reviewed exhaus- . 7-16 17-21 t1 vely . Its uses have been surveyed , and lvill be reviewed

J.n the follo\ving paragraphs.

In general, it is used as a raw material for all types of res ins~ ~1eat resistant polymers, insecti cidcs, flame- retardants, anJ as a precursor to cyclopentcne which is used as an elastomer monomer.

Streams which contain 75% dicyclopentadiene can be thermally polymerized to produce a low grade resin used as a tackifier for . . 23-25 rubber and as a coat1ng varnlsh The resin made from higher purity dicyclopcntadicnc resembles natural rosin in terms of physical properties. Since some of these resins contain double 7 bonds, their properties can be improved through l7lodification.

These J1odi.fiablc res1ns find use as rubber tackifiers, pressure 2 sensitive tapes, :1ot-melt adhesives, coatings, .inks, and paints ' 25

~·n1cn Jrying oils, such as soybean, linseed, marine, and vegetable oils, are reacted with dicyclopentadiene they dry more 22 rapidly and have increased \vater and alkali resistance .

The halogen derivatives of cyclopentadiene are used as fire-

. - . . .. 2, 7,8,11,18-20 f 1 rc::.1ruJ.nts and lnscctlCll1es . s ome o tne more conunon chloro compounds of cyclopentadiene are illustrated in Figure 1.

All of the compounds shown are potent insecticides or pesticides.

Dcchlor311 (~lire:x) and Het acid are also used to impart fire-retard- 2 ancy properties into polyester resins and polyurethane foams .

Dicyclopcntadiene finds use as a vulcanizing agent for ? ethylene anJ propylene copolymers-. A superior vulcanizing agent,

\..rhi ch acts more quickly than dicyclopentadiene is ethylidene no rbornene. It is synthesized from cyclopentadiene and 1, 3-butadicne.

Propenyl norbornene, produced from cyclopentadiene and 1,3-pentadienc, also has good vulcanizing properties. It should also be chcapeT to produce than ethylidene norbornene since its synthesis

0 + Cl 0 \ 1101 S+O 1C12 0 0 - -~' Endosulfan Het Acid Cl Cl ,.,.. ~ Cl ., Cl / c cr c1 0 ~1I c Cl Cl le» J Jl \:2 Heptachlor c c( / J Cl · Cu/EtOH Aldrin Cl / I Cl Kepone H202 Cl / Cl Cl 1 Chlordane c~\ 1 ?;tl Cl -{:a' ~ ·Cl :~cr,o Pentac® Dieldrin

Figure 1. The More Common Chloro Compounds of Cyclopentadiene . 00 9

involves one less step .

.\s mentioned earlier, cyclopentadiene is a precursor for

eye lopentenc ~·;hich is used to produce the 1,5-polypentenaners.

Cyclopentadiene is also used to produce norbornene resins through 2 polymerization \vj th a vinyl rnonorner These find use in various molding applications. Norbornene rubbers are finding application as sound-shielding and soundproofing materials.

C!1emical Separation of 1, 3-Pentadiene From Pipei}',.lene Concentrate

Using Sulfur Dioxide

:.tany conjugated dienes form crystalline monomeric sulfones 27 \~hen heated with sulfur dioxide under pressure . One exception to this general statement is cyclopentadiene, which reacts with 7 sulfur dioxide to form t~e polysulfone resin shown below.

502 0 0 11-0-11s s 0 II II 0 0 n

The reaction of dienes with sulfur dioxide is reversible. At high temperatures, usually above 100°C, the sulfone dissociates.

Since most dienes form crystalline sulfones, which can be decomposed easily to yield the original dienes, they can be conven- iently separated from alkene hydrocarbons. ,.,1ne propert1es. 27 o f some conunon sulfones are illustrated in Table I I. These sulfones, especially that of 1,3-butadiene, have some commercial usefulness.

The hydrogenated sulfone of butadiene is sold under the generic name "Sulfolane"27 . It has a melting point of 28°C, boiling point 10

TABLE II ?7 Some Properties of the Common Sulfones-

Oicnc Sulfone Melting Point Decomposition oc Temperature °C

1, 3-butadicne 65 125

1,3-pcntadiene oil 100

isoprene: 63 125

cyclopcntaJiene 125

of 285°C, is chemically inert, non-corrosive, and 1s very miscible with \vater.

Due to its high selectivity toward aromatic compounds, it

1s used in extractive distillation and solvent extraction processes.

Two examples in which sulfolane has been successfully used are:

t • 29 1 . . . t 11c U d ex tcc.1n1que , w 1ere 1t 1s an al ternat1 ve sol vent to aqueous diethylcncglycol in the solvent extraction of aromatics; and in 27 t h e Su.. 1".r1nol process , \\flere1 1t . can b e com b.1ne d \~lt . h monoet h ano 1 amine t:o remove hydrogen sulfide, mercaptans, carbonyl sulfide, and carbon dioxide from hydrogen, natural or synthesis gas.

The sulfone formation technology has been applied by

Cra1g . 30-32 , to separate 1 , 3 -penta d.1ene f rom var1ous . unsaturat e d hydrocarbons and for the isomerization of cis-1,3-pentadiene to trans-1,3-pcntadiene. 11

Separation of 1,3-Pentadiene From Piperylene Concentrate By

Cunrous .\nunonium Chloride Complexing

fhc formation of unscn:urated hydrocarbon complexes with

.salts of heavy metals of group IB and IIB of the periodic 33 sys tern have been knoVv11 for some time. In 1898, Chavastelon reported the reaction between acetylene and cuprous chloride. 34-41 ~.]ore recent work , involves using the complexing ability of the ;;1ctal salts, usually cuprous chloride, to separate various alkcncs, cyclic alkenes, and dienes from one another.

The stoichiometry of these stable complexes formed involves 33 38 one group IB metal atom for each pi bond of the hydrocarbon '

The stability of the complex is dependent on the type of hydrocarbon involved. In general, the order of increasingly stable complex formation is: alkenes, acetylenes, cycloalkenes, dienes, . 33 37 and cyc1olllenes ' . The type of anion present in the salt is also important in complex formation, for CuCl and CuBr will complex

\vith unsaturated hydrocarbons, whereas CuO, CuCN, Cui, and Cu 2s \vi 11 not .).) .

One advantage to the pi complexes between the metal salt

~md the double bond is that it does not convert to a sigma complex.

The normal chemical reactivity of the cornplexed double bond is 33 effectively inhibited . Evidence of this is the lack of isomerization and polyrneri:ation of the tmsaturated hydrocarbon during complex formation and decomposition of the complex to recover the pur.i fied hydrocarbon. 12

Complex formation has been achieved in the gas phase, where 38 the ~ctal salt is Jispersed on a carrier or in a 50/50 mixture 33 \oJi th .::lass beads for use in a fluid bed lll1i t More generally

ho\·;evcl', ::he rc3.ction is ~arried out in the liquid phase, Hith the 34-38 metal salt in an aqueous solution or slurry The aqueous

solutions of the copper salts are prepared by·using various

so 1 ut1z1ng. . agents 34' 35' 39 , suc11 as anunon1um. c h 1 or1 . d e. As shown 3.+ 35 by Lur' c ' , the amount of soluti:ing agent in the solution has

a marked affect on the rate and extent of complex formation.

At the higher concentrations, t\venty five weight percent or more,

the complcxing is sJo\.; and incomplete. This seems likely, since

these agents and other alkali salts or bases, can be used to . 3-+ 35 decomuose the complexes to regenerate the purified hyarocarbon ' .

Decomposition of the complex can also be accomplished by heating . 33-35 or rcJuc1ng the pressure .

.\ny alkene present in the hydrocarbon mixture will also have 33 a marked aifect on the rate of complex formation . In the gas

phase, the al kenes wi 11 dis sol \·e large arnotmts of cuprous chloride by '.'t·eakly complcxing wi t.h it. This enables the dienes to react

more quickly throu~1 a liquid-solid reaction rather than a gas- solid reaction.

Separation of 1,3-pentadiene from other hydrocarbons, includ-

1ng cyclopcntene, has been accomplished through the use of cuprous 33 36 40 41 chloride complexes ' , ' It was first found by Ward and

Fa"f k. ·1n 36 , that the complex involved two moles of cuprous chloride 13 per r:-1olc of 1, 3-pentadicnc and had the following formula:

The separation of l, 3-pentadicne from aliphatic alkenes pro\·ccl to be a simple chore, since its complex \vas an insoluble solid which could be recovered from the solution and soluble complexes easily. The cyclopentene present in the mixture would also precipitate, but its complex formation was prevented by the

3.Jdition oi JliUTionium chloride and increasing the temperature to

35°C. The cyclopentene cuprous chloride complex is unstable at these conditions. This allows only the more stable 1,3-pentadiene cuprous chloride complex to form.

Using this method, 54 9o of the 1, 3-pentadiene was recovered from the pipcrylene concentrate in 99. 9go purity. The contaminant being a trace of cyclopentene.

Separation of Trans-1,3-Pentadiene From Piperylene Concentrate By

Dicls-.\ldcr Reaction With r.taleic Anhydride

The Diels-AlJer adduct of trans-1,3-pentadiene and maleic 3 anhydriJe is used cor.unercially as 311 epoxy hardener . Its acid and esters arc used as plastisi:ers and softening agents for . 4 resins, plastics, gums, an d lacquer f 1lms .

The Dicls-Alder reaction of trans-1,3-pentadiene and maleic anhydride to form 3-methyl-1,2,3,6-tetrahydrophthalic anhydride has been used as an cffecti ve means to separate trans-1, 3-penta- ' , d b 42-..J. 7 dicne from cis-1,3-pentadicne ::m d otner ny rocar ons . This can be accomplished since trans-1,3-pentadiene reacts readily with 14

maleic ~nh,:Jride, while the cis-isomer ~vill not react at an 4 3 ' -+ 4 ' -+ 6 ' ·+ 7 apprcci3.blc rate . Cis-1, 3-pentadiene t-1ill react slo~vly 43 \•:i. t~1 ;:1.1lcic anhydride, but only ~..1ndcr ',rigorous conditions

In order for a 1,3-diene to react with maleic anhydride,

. . . l . . d f . 44' 4 7 1t must ex1st 1n tae ClSOl con ormat1on . 1nis conformation is necessary in order for the 1,3-diene to exist as a coplanar system \vi th one of its sides completely exposed to the dienophile.

By vbs\Jrvin~ a scale model of cis-1,3-pentadiene, it can be seen that the cisoid conformation is sterically hindered by the protrudin~ methyl group, ho\..:e\·er, a scale model of its adduct shows no stcric hindrance. Since the planar cisoid conformation is required to expedite the transition state, the slow rate of this reaction must be due to the hindrance of the planar cisoid conformation and not a sterically hindered product.

The reaction rates of various 1,3-dienes with maleic anhydride 47 have been studied . It was found that trans-1, 3-pentadiene reactcJ much faster than either 1,3-butadiene or isoprene, but much slower than cyclopentadiene. This is depicted in the rate constants 7 listed in Table III-l . The increase in reaction rate as proceeding from 1,.3-butadicne to trans-1,3-pentadiene is due to the methyl groups ability to release electrons into the system. The rapid rate at which cyclopentadiene reacts is attributed to the low activation energy of 8.5 kcal as compared to 11.7 kcal for butadiene.

This lo\v activation energy \-.ras expected, since no change in conformation 1s necessary to form the required planar transition state. 15

TABLE III

Rate Constants for Various Diels-Alder Reactions of Several Dienes with Maleic Anhydride

Reaction Rate Constant 47 Diene Temperature oc 1/mol/h butadiene 25 0.19

isoprene 25 0.57

trans-1,3-pentadiene 25 0.92

cyclopentadiene 25 -200.00

cyclopentadiene -40 4.00

cyclopentadiene -60 0.72 16

Thcrr.:ouvnJ.mics

TI1crmodynarnic values for various hydrocarbons, including the 48 49 pent;1 dlcncs,. \vere report e d b y K1·1 patr1c. k~, et a l . and.. Dous 1 ln· . fhc values produced through these studies proved to be inaccurate.

The data implied that the stability of cis-1,3-pentadiene was greater than trans-1,3-pentadiene at all temperatures above 298°K,

\vhcreas the reverse has been shown to be true. Therefore, the thermod~·naf.lic llatet of the pentadienes and the other components of 50 the piperylenc concentrate reported by Messerly, et al. and

Stull, et a1. 51 have been used.

The equilibrium constants for the geometrical isomerization of 1,3-pentaJiene were calculated using the linear equation given 5? by Egger and Beason -. They studied the geometric isomerization of 1,3-pentadicne with nitric oxide over the temperature range of

400 to 670°1\. Their data was fitted by least squares using a rcS_;ular rc~rl!ssion program to yield the following linear equation,

(1037±28) 2.303Rlo0: K = -(.14±.05) + --- t/c T with R in cal/molc °K.

Thermodynamic data for cyclopcntadiene at low temperatures 53 was given by Kistiakowsky et al. . The free energy of cyclopentadicne at the higher temperatures \'Jas estimated by neglecting the effects of the conjugation and assuming the following to be true. 17

The equilibrium constant for the conversion of the piperylene into cyclopentadicne could then be calculated by:

6G = -RT ln K T) ..L with R in cal/mole °K.

Isomerization of 1,3-Pentadiene

The isomerization of simple alkenes with small amounts of a . . 53-57 h~lo0;en has been knohn for qu1te some t1me Benson, et al.ss,sg sho\.;ed that small catalytic amounts of iodine vapor at 200-300°C

\-:as capable of not only positional but geometrical isomerization of alkenes. Rate studies of these isomerizations have been made 60 and concluded the following general rate formula :

rate a (I ) ~ (alkene) 2

Interpretation of this half-order dependance on the iodine indicates that rart of the mechanism entails the addition of an iodine atom 60 to the carbon-carbon double bond . If this is true, the rate detcrr:1ining step must be either the internal rotation in the intermediate raJical or the addition or removal of the iodine atom. . 60 61 The general proposed mechan1srn ' for any atom or radical cataly:cd cis-trans isomeri:ation has been shown to be consistent with a consecutive step mechanism. This involves the formation of an intermediate radical followed by internal rotation in this radical.

If A 1s allowed to represent the radical catalyst, then this 18

mechanism :n~y be de pi cteJ 8.5 follo\vS:

R R" a R R_.. \

\ + A ~ b u. \ ·A

c' 1l c

R_.. b"" R"" A + pd a R RriA

The rate determining step for this general mechanism has been shown,

for some simple alkenes, to be the internal rotation in the inter- . d. 152,60,61 me dlate ra lea . But, if the Carbon-A bond strength and

the rcsonanl:~ energy introduced by the alkyl group(s) totals to

more than the strength of a carbon-carbon double bond (approximately

58 kc~l) then the rotation in the intermediate radical will occur

faster th:tn either the addition or removal of the radical catalyst

and h·i 11 not L:ontrol the rate of isomerization.

(someri:ation using radical catalysts, mainly nitric oxide and iodine, l1ave been applied to 1,3-pentadiene by several research-

4 2 , 5: , G1 - 6 3 Tl F k l 4 2 _ 4 ers . 1us. ~ran' et a . ref 1 uxed pure trans- or pure

cis-1,3-pentadiene with traces of iodine to obtain, after 18 hours,

86 9o trans- and 14(lo cis-1, 3-pentadiene in both cases. They also passed separately, pure trans- or pure cis-1,3-pentadiene through a vertical glass tube at 600°C to obtain 45~ and 40~ cis-1,3-penta- 52,61 diene in the liquid products respectively. Egger and Benson 19

isomeri.:2J l, 3-pcnt~dicn<.: l,.;ith nitric oxide over the temperature ranee Jt -l00°K to 598.S°K and obt~ined 73~ to 69% trans-1,3-penta-

Jicnc in the product. SimilJ.r results \\·ere obtained using an iodine 62 ~·atJ ...1 :·~t. ~\o• Ilr(;r J.n J ,S. 1nt~ . 1 t use d a mlcroreactor. to study the effects of alumina on 1,3-pentadiene at 760°K. They reported a proJuct \~hich contained 63.lgo trans-1,3-pentadiene. Wells and .. 1 63 \\ ' 1 son , using cabal t po\,'der and cabal t supported on alumina,

t~) nhtJ. in ab0ut ~5~ ) trans-l, 3-pentadiene at -+33°K using either of these c~talysts.

Dc hydro c ~· c I i : :1 t ion of 1 , 3- Pent ad i en e

The Jchydrocycli:~tion of 1,3-pentadiene has been attempted h\· various researchers. It should be noted, ho\,·ever, that most of these researchers used pure 1,3-pentadiene as their starting matcri:1l ~ \,·hcrc~s, in this study, the piperylene concentrate \\ras useJ.

The thermally induced transformations of 1,3-pentadiene has

. 1 1 ~ l . k . b--l • b 5 d l . . 1Jcen ;-;tu J ICl t)y ~1u1 ·1n an 11s coworKers. They passed pure

1,.)-pcntaJicnc through a quart: tube reactor at ~50-550°C and atmospheric pressure to obt:1in less than 1.09o cyclopentadiene at

RO~, pcntaJ icne conversion. \\1lcn the pressure h·as dropped to

9 20 mm llg, thP yield of cyclopcntadiene \vas less than 1. S o at 13°o

0 p c n tad i c n e con v e r s ion . At 15 atmospheres , 9 ..J. a of the pentad i en e was converted into r.S to c aromatics and high boiling polymeric 10 residual hYdrocarbons.

KenncJv ()b, \vas u b le to obtain 7. 3l~ eye lopentadiene by passing 20

1,3-pcntadienc through a stainless steel tubular reactor at 600°C and .)0 !'lm II~ ,..·ith a contact time of ~.4 seconds. Under identical conJi tions, he produced 3. 9 9a and 9. 2~.5 cyclopentadiene over silicon

.. J () 7' (J ~ d . k l . 6 7' 69 . ~arDt 0 an JaC c1a1n respcct1vely. 70 71 Timashev and Gregorovich produced 12.4 mole percent butadicne and 7 mole percent cyclopentadiene in a quartz tube reactor at 700°C and SO mm Hg. The contact time was 1.2 seconds. .., -,- 7 _,,.) 0 l3oJn3rvuk ct a 1 . produced 30~ cyclopentadiene and ~0% butadiene at a pcntadiene conversion of 25%. Their quartz reactor

\,•as heated to 700°C at 50 nun Hg. The space time was reported as -1 800 l.sec.mole . They concluded that the overall yields of cyclopcntadicne and butadiene were not only dependent on the pressure, but also ho~ the partial pressure of the diene was obtained. Thus, h'hcn water \•;as used as the dilutent, it promoted dimer and polymer formation, h·hcreas nitrogen did not. 74 75 Hutchings ct al. , used hydrogen sulfide as a promoter to produce cyclopcntene and cyclopentadiene from 1,3-pentadiene while reducing cracking. Table IV briefly illustrates their results.

Other promoters, such as hydrogen bromide, hydrogen fluoride, and carbon tetrachloride, have been found to promote the dehydro . 76 77 cycli:ation reaction while reducing crack1ng ' 78 Bencsi cracked n-pentane in the presence of hydrogen over a

Pt/Siu, catalyst at s:s °C and atmospheric pressure to obtain 2.1 ~o cyclopentaJicne, 2.S<~ cyclopentene and 6.S?o cyclopentane \vith a selectivity of 34%. Witl1out hydrogen present, he obtained 5.3%, 21

TABLE IV

Results of a Dehydrocycli:3.tion Experiment 75 P ertorme~ d 'oy Hutc~ h. 1ngs

Reactor Temperature 650 650

Mole Percent H S 0 100 2 Conversion 21.8 15.2

Components

methane 6.24 1.97

c'"' 4.62 1.01 c_ 1.96 1.40 .) butcnes 2.55 2.85

hutaJicne 28. 80 4.30

pentane 0. 4 7 1.18

1-pcntene 0.59 4.00

2-pentcne 2.04 10.20

cyclopcntcne 6.16 49. so

cyclopentadienc 26.70 20.40 22

2.lc~' ~d ~. -f~ of these products respectively with a selectivity of

Various cracking catalysts have also been utilized to produce 79 eye 1 cpcnta d1cnc. t- rom 1 ,~-penta~ d.1ene. Th us, S1u1·1n~1 . k. was aleb to obtain 18~ cyclopentadiene at 600°C and 20 mm Hg over a Al o -cr o - 2 3 2 3 K,O catalyst (42-7-1 mole % respectively) with a space velocity of -1 1.0 hour . Under identical conditions, over a 5% Pt/C catalyst

0 th~\· oht;Jj ned 1 :-' ) cyclopenta.diene. 80 Shuikin and Tulupov produced 9. 7?o cyclopentadiene over a

AL,O~-Cr_,O_ (1: 1 by \veight) at 600°C and 10 rnm Hg \vith a space - .) - .) -1 velocity of 0. 5 hour . TI1e conversion of 1, 3-pentadiene was 35.2%.

At atmospheric pressure the yield of cyclopentadiene dropped to 0. 4'1o at 71.9~ 1,3-pentadiene conversion. 67 Kennedy and Hct:e1 obtained up to 9% cyclopentadiene over various heterogeneous catalysts at 600°C and 20 nun Hg with a contact 81 time of 0.1 seconds. Thus, over fused alumina , they produced 8.69o cyclopcntaJicne ~t 69.1 o., 1, .3-pentadiene conversion. At 600°C and 82 200 nun ilg over activated alumina they obtained 4 . .3 96 cyc1opentadiene at 43(1o 1, 3-pentadiene conveTsion. At 600°C and atmospheric pressure 83 over a chromic oxide supported on alumina catalyst , they obtained

2.8~ cyclopentadicne at 70% diene conversion and 8.3% cyclopentadiene 84 cyclopcntadiene over silica ge1 at 62.7~ conversion.

85 0 Gi. tis 3.nd Rozengart produced 7°~ cyclopentadiene and ~. 0 5 cyclopcntcne with a selectivity of 35 and l0°o respectively at 615oC.

The Jienc conversion was 20~. EX PER I:.IENTAL

Pip::;ry lene is a cormnon name for 1_, 3-pentadiene. In this report, 1,3-pentaJlene will be referred to by IUPAC nomenclature.

"Piperylene Concentrate" is the name given to a mixture of unsaturated hydrocarbons obtained as a by-product distillation fraction from a naptha cracking process. The term piperylene concentrate will be used to designate this mixture. A typical composition of this material was presented in Table I.

Analysjs

The components of the starting materials and products were separated and analyzed on a Perkin-Elmer Sigma I Gas Chromatographic system equipped with flame ionization and thermal conductivity detectors. This system integrates the area under each peak, compares each peak area to the total peak area, and calculates the

\-:eight percent of each component using a response factor.

The response factors employed were taken from literature values 86 . The values given were all approximately 1.0. The two exceptions were benzene 1.1~, and toluene 1.07. The response factors for the hydrocarbons not listed in the literature were calculated using standard samples. All 1vere found to be approximately 1.0.

The colwnn consisted of a ten foot section of one-eighth inch thin \vall stainless steel, packed with 20 weight percent sebaconitrile

(Pfal tz and Bauer, Inc.) on 80-90 mesh acid washed Anakrom C22 24

(Ana labs Inc.) follo\ved by a 20 foot section of one-eighth inch thin Hall stainless steel, packed \

(2-r::ethoxyethyl) adip::tte (Supelco Inc.) on 60-80 mesh Chromsorb W l~1alabs Inc.) non-acid washed. With only the injector end connected, the column was conditioned with a helium flow of 25 ml per minute at 100°C for ten hours. The exit end was then connected in parallel to the flame ionization and thermal conductivity detectors via a one-eighth inch stainless steel tee which had been packed \vi th the Chromsorb W to decrease dead volume.

The starting materials and products were analyzed isothermally at 60°C with a helium flo\v of 25 ml per minute. The injector temperature \oJas set at 75°C and the flame ionization detector at

250°C. These conditions were held until all major peaks eluted, usually about 22 minutes. Heavies (residues after distillation of liquid product) \vere analy:ed on the sebaconitrile column. This proceeded isothermally at 100°C with a helium flo\~ of 30 ml per minute. The injector was set at 100°C and the flame ionization detector at 250°C. Under these conditions, all peaks eluted within

90 minutes. Hydrogen \vas analy:ed on a ten foot by one-eighth inch stainless steel column packed \vith 80-100 mesh Carbosieve B and operated at 70°C with a helium flow of 15 ml per minute. The injector temperature was set at 75°C and the thermal conductivity detector at 150°C.

Since the analyzer was not equipped with a backflush accessory, the column tended to load up with heavies after prolonged use. To 25 minimi~c this load up, the colunm was purged for one hour at 100°C with a helium flow of 45 ml per minute '.·:henever the retention times of the components Jecreased more than 2°o from their set points.

This analytical procedure has a precision \\jhich indicates a relative standard deviation less than or equal to 0.02% for all components except butadiene, cyclopentene, and isoprene which are less than or equal to 0. 03go, and cyclopentadiene and trans-1, 3- 87 pcntJ.diene \vhich are less than or equal to 0.08% . It is also indicated that the values obtained will not vary more than 2o from the averages, (i.e., trans-1,3-pentadiene will be ±0.16% relative at the 95~6 confidence level).

Analytical Standards

The pure samples of cyclopentadiene required for gas chroma tograph peak identification and response factor calculations, were obtained by depolymerizing dicyclopentadiene (technical grade) using 88 the r.1cthod of :.loffctt . Using this method, a product which \vas approximately 94'?6 pure cyclopentadiene was obtained. This was increased to 99.5+% by fractionating the product through a nine-inch glass column packed with Penn State extruded packing and collecting the fraction which distilled bet\veen 39 and 42\JC. This fraction \vas then used immediately for the response factor calculations.

Standard samples of trans-1,3-pentadiene (Aldrich) were used without further purification; while the samples of cyclopentene

(.J. T. Baker), cyclopentane (J. T. Baker), pentenes (Pfal tz and

Bauer, Inc.), 2-methyl-1-butene (Pfalt: and Bauer, Inc.), and 26

:-methy 1-:-butene (East.man) were all fractionated before using.

Liquid rh2SC Reac"tions of 1,3-Pentadiene

A number of trials, utilizing liquid phase reaction conditions,

·~-;ere m:tdc to investigate the geomet.ric isomerization of the penta- diencs and their separation from the mixture. 42 Halogen Catalyzed Isornerizations

A mixture of 100 grams of piperylene concentrate and 0.6 to

1. 3 6rar:1s iodine (or bromine) \oJas refllLxed with a trace of hydro quinone for :..t to 72 hours. The piperylene was then removed by distillation through a twelve inch Vigreux column and analyzed.

A similar mixture of piperylene concentrate and iodine was placed inside a stoppered flask and stored at 0°C. Small portions, approximately 25 rnl, of this mixture \-Jere removed periodically, distilled as before, and analyzed. 89 Base C~taly:cd Isornerization

These isomerizations were performed in a 100 ml three-necked round bottom flask equipped \oJith a thermometer, condenser, and stirrer.

To the apparatus \vas added 50 ml (0. 2 mole pentadiene) of pipcrylcnc concentrate and l gram (. 01 mole) of potassiu..rn tert butoxide. This was carried out both with and without a nitTogen blanket. This \vas repeated using 20 ml (0. 08 mole pentadiene) pipcrylenc concentrate, and 5.0 grams (.045 mole) potassiwn tert butoxidc i.n SO ml of dimethyl sulfoxide. This mixture was stirred for 24 hours under nitrogen, both at ambient temperature and reflux. 27

Tltis was repeated us1ng ~0 ml (0.08 mole pentadiene) piperylene conccntr~te and 10.0 grams (0.9 mole) potassium tert-butoxide in

50 :.11 of dimethyl sulfoxide. Similar mixtures Here stirred tu1der ni tTogcn for: :.~ hours at l5°C, 12 hours at 35°C, and 12 hours at reflux.

At the end of the reaction period, the mixture was quenched with 250 rnl of ice-water containing 25 rnl cyclohexane. The organic l:Jyer ,\.3.5 ,..;asncd once \'/ith 100 ml of ice-l-.rat:er, dried over molecular sle\·e (1~:\), distilled, and analyzed.

Separation of 1,3-Pentadiene from the Piperylene Mixture by

Cuprous Ammon1urn. C h lor1'd e Treatment 30

For all of the following work, piperylene obtained from the

Chemical Sample Company was used. Its composition is presented in

Table V.

TABLE V

Composition By Weight of Chemical Sample Company Piperyler.e

Component Percent

lights 1.0

2-methyl-2-butene 2.0

cyclopentene 17.7

trans-1,3-pentadiene 67.8

cis-1,3-pentadiene 11.5 28

Into ~ 500 ml three-necked flask equipped with a mechanical stirrer md thermometer, ~-.ras placed 64 grams (0.65 mole) cuprous

chloride, 32 :.;rams (0.6 mole) ar.llllonium chloride, 80 ml \vater, 4 ml

concentrated hydrochloric acid, and 2 grams copper turnings. The

flask was stoppered and stirring was started. On stirring, the temperature decreased rapidly to approximately l0°C with the formation of a deep bro\vn slurry. After one-half hour, stirring h·as discontinued and the mixture was allo\ved to set overnight.

The flask was then placed in a 25°C water bath and 40 ml

(0. 31 mole diene) of piperylene was added. The stirrer was started and the reaction proceeded \vi th a slight evolution of heat and the

formation of a greenish-yellow precipitate. After stirring for one hour, the \vater temperature \vas increased to 35°C and the stirring was continued for an additional five to six hours. The mixture was then fi 1 tcred and \vashed with warm water to remove any residual

~mmonium chloride. The prouuct h·as allowed to stand open, \vith occasional stirring, to allow any unreacted hydrocarbon to evaporate.

This procedure gave 60 grams, or 70% yield of complexed product based on the dicnc.

The complexcd product and 175 ml of water was then placed into a 250 ml ~lorton flask \vhich was equipped with a thermometer, mechanical stirrer, and an unpacked factionating distillation apparatus. The stirrer was started and the temperature of the pot

\-.ras increased very slowly (approximately l0°C per hour). The decomposition proceeded rapidly at a pot temperature of 63 to 72oC, 29

giving ll. 75 grams (54~)) of pure 1, 3-pentadiene.

S~rnthesis of 3-01ethyl-l, 2, 3,1}-Tetrahydrophthalic Anhydride Via 42 Diels-,\lder R0action h'ith 0Ialcic Anhydride

To a 250 ml round bottom flask equipped with a mechanical stirrer and condenser, was added 28 grams (0.3 mole) of maleic

anhydride, 110 ml piperylene (0. 7 mole trans-1,3-pentadiene), 100 ml benzene, and 0.1 gram of picric acid. The flask was then fitted

\•;ith :1 thermometer 311d stirring was started. Heat \~as applied to

st~rt the reaction which continued with the evolution of a consider-

able amount of heat. After stirring for 24 hours, the benzene and

unrcacted piperylene was removed by simple distillation tmder a

reduced pressure (about 380 Torr). The residue which was left, was

allo~ed to crystallize and then heated with a 20 volume percent

solution of benzene in pet ether. The solution was then treated with :\'orit-A, filtered, and cooled. The crystals formed were

\vashed with cold benzene-pet ether solution to remove unreacted

maleic anhydride, giving 21.5 grams, or 45~o yield (based on maleic 43 anhydride) of product with a melting point of 62°C. (lit. = 63°C)

l11e above procedure was repeated with refluxing the mixture

for 1-l hours. The yield dropped to 25°o, apparently due to the loss

of piperylcne through the condenser. With acetone as the solvent,

the reaction was carried out at l0°C for 48 hours. After workup,

and on cooling, 9. 5 grams of crystalline product \'las obtained. The

mother liquor from this crystallization was then cooled to approxi-

mately -45°C in an acetone-dry ice mixture to give an additional 30

7.0 grams of product. The total yield using this method was 35% of theory. 42 Pyrol\·sis of 3-C.lethyl-1,~,3,6-Tetrahydrophthalic Anhydride

The :1pparatus used for this pyrolysis is shown in Figure 2.

The pyrolysis tube consisted of a 12 inch long, one-half inch 316 stainless steel pipe filled with glass beads (4 mm diameter). This was heated to, and maintained at, 575 to 600°C by means of a

Lindbcr.; single :one tubular furnace. The temperature \vas monitored through thermocouples attached at various points around the tube and connected in parallel to give an average temperature reading.

The reactant hopper consisted of a 250 ml pressure equilized dropping funnel which was connected to the pyrolysis tube by a

24/40 to 10/30 reducing union which was inserted into a drilled out one-quarter inch stainless steel S\vagelok union. The 10/30 portion of the reducer was \vrapped with Teflon pipe tape before being inserted. Once joined, the entire joint \vas \vrapped tightly with glass cloth tape to obtain a gas tight seal.

The exit end of the pyrolysis tube \vas connected to a three necked round bottom blask containing an excess sodium carbonate solution \vhid1 \vas maintained at 60 to 70°C. The purpose of this solution was to dissolve the maleic anhydride, preventing recombina tion with the diene. The 1,3-pentadiene was then distilled off and collected in a receiving flask and a hydrocarbon trap, both of which were cooled by Jry icc. - Adduct Hopper

To Hood

Receiver with Na 2co 3 Solution

Figure 2. Pyrolysis Apparatus 32

Pvrolysis Procedure

One hundred sixty-six grams (1.0 mole) of 3-methyl-1,2,3,6- tetrJ.hyJrophthalic anhydride \vas placed into the dropping funnel and h~atcd to 100°C to insure complete melting of the adduct.

Once mel ted, and with a nitrogen flow of approximately 400 ml per minute, the adduct was slowly dropped (20 drops per minute) through the pyrolysis tube and into the sodium carbonate solution. The

\'Ol:Itilc portion \.;as distilled out of the flask and collected to give 26 grams or 38°o crude product. The crude product was then distilled and analyzed.

Vapor Phase I someri :ation and Oehydrocyclization of 1, 3-Pentadiene

The apparatus utili:.ed in this portion of this study is illustrated in Figure 3. In general, this system consists of a dilutcnt inlet, a two-stage steam generator, a hydrocarbon inlet and prcheater, a tubular reactor, a \vater scrubber, and water and hydrocarbon traps.

Stage I of the steam generator consisted of a 3/8-inch 316 stainless steel pipe tee \vhile Stage II was a twelve-inch length of 1/4-inch stainless steel tubing. The entire steam generator h1as wrapped \vith two layers of asbestos heating tape and insulated

\-lith one-half inch of asbestos cloth. Water was pumped, via two

30 ml disposable syringes and a Stage Instrwnent model 355 variable flow syringe pump, through a 1/ 16-inch stainless steel tube into the steam generator (Stage I). Flash vaporization effects \vere minimized by the use of this small inlet. The relatively large H 0 2 I Sample Ports To ,I I N /Air Hood _J 1 2 1 Piperylene - L 1 1 Steam Generator

[- Reactor !l_ 1~,1 I I~ Furnace I \Vaporizer I_ ~'_j I Water Trap Water l 1/ Scrubber l Hydrocarbon Figure 3. Vapor Phase Isomerization Trap Dehydrocyclization Apparatus 34 volume (10 ml) oi Stage I of the steam generator and a dilutent stream of nitrogen entering from the rear of the steam generator

(St.1.gc T), 3.lso helped to vaporize the water, minimizing flash

1.·apor1:ation. The exit end of the steam generator (Stage II) \~as attached to a 1/4-inch stainless steel Swagelok tee. One end of this tee \vas attached to a pressure gauge (Master gauge Type 100-f\1- moncl, ~1arsh Instrument Company) and the other to the hydrocarbon

\·a pori :cr.

The hydrocarbon vaporizer consisted of a five-inch length of

1/-l-inch stainless steel tubing \vi th a Swagelok stainless steel tee connection at its entrance, which served as the hydrocarbon inlet.

A 1/8-inch hole was drilled through one side of the vaporizer section. A 12-inch length of 1/8-inch stainless steel tubing was fitted into this hole and welded into place. This served as the inlet for gaseous promoters. TI1e entire vaporizer section was heated by means of t\vo layers of ~0 gauge Nichrome wire which had been insulated from the vaporizer and each other with asbestos paper. The vapori:er was then wrapped in glass cloth tape and insulated with one-inch of asbestos cloth.

Piperylene was pwnped, in the same manner as the water, through a one-inch hypodermic needle (gauge 25) into the vaporizer. TI1e liquid hydrocarbon feed thus mixed with the preheated dilutent

(usually steam and nitrogen) \vhich completely vapori:ed it before entering the prcheater.

The preheater consisted of a six-inch piece of 1/4-inch 316 35 stainless steel tubing i-Jhich connected the vaporizer to the re:1cror inlet. TI1e function of the preheater \vas to heat the g:1scous hydrocarbon-Jilutent mixture to a temperature Hhich was approximately l00°C below that of the reactor.

The reactor consisted of a 3/8-inch 304 stainless steel threaded pipe having a volume of 7.1 ml. It was placed in the center of a Lindberg Single-Zone tube furnace (~1odel 54031) equipped h·i th a solid state temperature controller (type 2200,

~lodel 59344). Its temperature was monitored with thermocouples positioned at various points around the reactor and connected in parallel in order to obtain an average temperature reading. The exit was attached to a 12-inch piece of 1/4-inch 316 stainless steel tubing \~hich '"as heated by a means similar to that of the inlet vapori:er. To the end of this exit tube was attached a

316 stainless steel Swagelok tee. A stream of diluting nitrogen flo\-.rco through this tee and met the effluent stream, cooling it and insuring its complete vapori:ation.

\Vater vapor \vas removed from the effluent by means of a \vater scrubber \vhich \vas made up from an eight-inch length of 3/4-inch pyrex tubing closed at one end and attached to a ~ 19/22 Claisen adaptor at the other. The water scrubber was wrapped \vith heating tape, insulated, heated to, and maintained at 45 to 50°C to insure none of the c hydrocarbons \vould condense. TI1e exit end of the 5 adaptor was fitted \vith a West-type condenser, through which heated

(50°C) \vater was pumped. The exit end of the condenser was 36 connected, via Tygon tubing, to the water trap \1/hich consisted of a 12S ill Erlenmeyer flask filled \vith 4 mm diameter glass beads.

Its purpose \vas to condense any remaining Nater vapor from the

·..:-fflucnt hefcrc entering the hydrocarbon trap.

The hydrocarbon trap consisted of two 25 ml test tubes which were filled \vi th 4 ml diameter glass beads and immersed in isopropyl alcohol cooled to -65°C.

S:.unplc ports were located at the reactor exit, after the h'ater scrubber, and after the hydrocarbon trap. The two former ports \'iere used in obtaining hydrocarbon samples, \-Jhi le the latter

\-:as used in obtaining hydrogen samples.

All temperature measurements were made with Chromel-Alumel thermocouples and read with an Omega model 200 digital thermometer, except the hydrocarbon trap temperature, which employed an iron constantan type J thermocouple and was read with a Fluke 2100A digital thermometer.

Catalyst Preparation

TI1e activated al urnina, 8-14 mesh (~1athes on, Co 1 ernan and Be 11) , and silica gel, 6-16 mesh, grade 03 (~Iatheson, Coleman and Bell),

\-Jere Jried overnight at 120°C prior to use. The stainless steel packing was the 316 stainless steel Penn State extruded packing.

Hydrogen sulfide (Air Products) and sulfur dioxide (Matheson) were supplied from a lecture bottle.

The sulfidcd 316 stainless steel surface was prepared by passing air through the reactor (500 ml/rnin) which was packed \vith 37

Penn St.:1tc extruded pack:.ng for approximately six hours at 600°C to oxidi:e the surface. The temperature was then dropped to 300°C with

~i.r rlO\ving. Once the reactor equilibrated at this temperature, the a1r \vas .r·lushcd ~·•ith nitrog~n. After a thorough flushing, a mixture of 10 to 25 volume percent hydrogen sulfide in nitrogen was passed through the reactor at a feed flow rate of about 100 rnl per minute.

·n1e sul fiding of the metal was very exothermic as indicated by an

1ncrease in reactor temncrature. This temperature was kept below

3S0°C by controlling the hydrogen sulfide flo\v rate. After the reaction \vas completed, as indicated by a stabilizing temperature, the reactor \vas flushed with nitrogen and brought to the desired conditions. The composition of sulfides on the metal was not determined. It is assumed that the bulk of the sulfide was iron sulfide \-:ith some chromium and nickel sulfides also being present.

~lolybdcnum does not oxidi:e or sulfide readily under the conditions ment1one. d90 .

The :11 urn ina s upporte<.l cat al ys t \vas prepared by the incipient

\vctncss ted1nique. In this technique, deionized water was added drop\-:ise to a measured amount of support material until the first sign of free water. The volwne of water added at this point was the water pore volume of the support. The desired amount of metal salt was then dissolved in that amolillt of water and added dropwise to the dry support as before. The \vater was then removed by gently heating the material leaving the metal salt behind. The heat was 38

applied s lo'.vly JS not to fracTure the pore structure and decrease its __;urface area. The salt can then be reduced '\'lith hydrogen, le:1vin~ the wetal deposited in the pores of the support. 'lliis method Insures that the metal was dispersed over the entire surface of the support and not deposited in the bulk.

The alumina supported platinum catalyst was prepared as follo\·.'S. A solution of 1.1266 grams of chloroplatinic acid in

39.5 rnl of deioni:cd \o;ater \vas slo\vly added to 51.13 grams of alumina \vith rapid stirring. This \.Jas then dried overnight at

120°C prior to calcination and reduction. This gave a catalyst h'hich contained 0. 82 weight percent platinum on alumina. Unfortu nately, time did not permit this catalyst to be fully evaluated.

Operating Procedure

l'.·ith nitrogen flo\ving through the system, all the components

\vere brought up to temperature. Once the desired temperatures

\~·ere obtained. water was pumped into the steam generator. The flow of nitrogen and steam was then adjusted to their desired rates.

The system was then allowed to equilibrate for one-half hour ~'V'ith minor adjustments JS required. Once equilibrated, the desired feed flow of piperylcne was introduced and the timer started. At a time of approximately two minutes into the reaction, a 0.5 ml sample of the effluent was taken and analyzed. Subsequent samples were taken every 20 minutes and analyzed until the reaction equilibrated and them for one hour longer. After the gaseous hydrocarbon was analyzed, a wet test meter was positioned after 39 the hyJrocarbon trap and 0. 5 ml samples of hydrogen were taken

(gencr~lly six) at five minute intervals and analyzed.

Run tiQes ran~ed from approximately one and one-half to seven hours, JepenJing on the catalyst employed and the conditions. After each run, the catalyst \vas cleaned \vi th a 1:1 weight ratio of steam and air for one hour to remove coke which was deposited during the reaction period. Scpara~ion of 1,3-Pentadicne from Piperylene Concentrate Using

Cunrous c:tl~.1 r i.dc ComnlexinB

Since the cuprous chloride complexes of simple aliphatic alkcnes are soluble in the salt solutions in which they are formed, they arc easily separated from the insoluble pentadiene and cyclo- 34-38 pentene complexes . Through the addition of ammonium chloride, the Jlh.cne anJ cyclopentene complexes will decompose leaving the more stable and insoluble pentadiene complex behlnd. 34 ' 35 ' 39 This

.:1J.Kcs :he :.Jcparation of 1, 3-pentadiene from the piperylene concen- tratc a relatively easy chore.

The d~t:1 for this separation lS shown in Table VI. The rcco\·cry \oJas above fifty percent, with the product being 99. 99o pure 1,3-pcntadiene. It was interesting to note the increase in the relative amount of trans to the cis lsomer. This can be cxplain~d, not by isomeri:ation, but by the stability of the

.).) corilplcxes. The cis isomer forms a more stable complex . Since the Jccomposi tion temperature \vas not taken much above the decorn- position temperature of the cis-1,3-pentadiene complex, some was left undccomposed in the pot, \'Jhile all of the trans-1, 3-pentadiene complex was completely decomposed.

Iodine isomerization of 1,3-Pentadiene

The isomeri::ation of 1,3-pentadicne by small amount.s of halogens, mainly iodine and bromine, is believed to proceed through 41

TAGLE 'JI

D:1t;1 For The Separation of 1,3-Pcntadiene From Piperylene L:unccntTate fhrou~h Cuprous Chloride Cornplexing

Reactor Temperature (°C) 25 Contact Time (hr) 7

~CC0\'Crv ( '',) 54

0 f-EED CO~iPOS IT I 0:\' ( wt o) trJns-1,.3-pentadiene 67.7 cis-1,3-pentadiene 11.8 cyclopcntenc 17.7

trans-1,.3-pentadiene 90.9 cis-1,.3-pent3diene 9.0 cyclopentene 0.1

DEC0~11'0S1TIO:-\ TE~!PERATURES (° C)

trans-1,3-pentadiene 66

"":''") cis-1.3-pentadiene I- Pot De composition Temperature 63-75 42

the frec-r~dic~l reaction path exem~lificd in Figure 460 , 61 . As

shohTI in the first step, the halogen may either abstract an allylic

hydrogen or ~dd to the double bond. The simultaneous addition of

the r1alo~en to form a dihalo compound should not occur to any

appreciable extent since the concentration of the halogen is low.

The low halogen concentration decreases the probability that the

adduct radical will react with molecular halogen. It should

instc~d, Jissociatc aga1n to yield the 1,3-pentadiene with or

1~ithout isomerization. This is not to say that no side reactions

\vill transpire, since during these isomerizations, two important

side reactions did indeed take place: polymerization and diiodide

formation.

These side reactions are explained through the energies

1nvo· 1 vc d . Eggar 5 ~' 60 , 61 , ver1·f· 1e d the rate d eternun1ng· · step 1n·

iodine isomeri:ation of simple alkenes is the internal rotation of

the intermediate radical. This is not the case with 1,3-pentadiene,

due to the additional resonance energy (-12.6 kcal) involved. TI1e

secondary carbon-iodine bond energy (-53 kcal) plus this resonance

energy, add up to more than the strength of a carbon-carbon double bond (58 kcal). Therefore, in the iodine isomerization of 1,3- pcntadiene, the addition or removal of the iodine is the rate

determining step rather than the internal rotation of the intermediate adduct radical.

The rcsul ts for the iodine isomeri zations of 1, 3-pentadiene are sho\vn in Table VI I. These results are not consistant \vith the 43

t I I i + HX ~

""'I I

r":/ X I

/\ i I I I v X v

~G~98 = -1.2 kcal ~\I

~H298 = -1.7 kcal RXN

Figure 4. Proposed i\lechanism for the Iodine Catalyzed Isomerization of 1,3-Pentadiene. TABLE VI I

Data for the IoJine Catalyzed Isomerization of 1,3-Pcntadicne

K ·A· ·k * (I 0 o. Temperature Contact ~1ole 'o ·o Trans Lsomcr 'o Trans Isomer Kt/c t/c 5, (oC) Iodiner. 111 lll Calculated .... Time (hr) Feed ** Product ** ------

42 20 O.lo 51.2 6~.7 2.30 4. o8

42 20 0.51 51.2 69.7 2.30 -~. B8 42 72 0.26 51.2 61.8 1.62 4.o8

0 22 0.40 51.2 60 .(> 1.54 6.29 0 72 0.40 51.2 60.6 1.54 6.29 0 144 0. 40 51.2 62.4 1.66 6.:.?9 0 648 0.40 51.2 62.9 1.70 6.29 0 7275 0.40 51.2 57.3 1.34 6. 29

* Based on the total piperylene feed. *·k Based on the amount of 1, 3-pentadiene in the piperylene feed. **·k Basis: 2.303R log Kt/c = -(0.14±0.05) + (1037± 28) T with R in cal/mole °K 45

2 exp~ctcu. resu l t3. ThJ c ca 1 cu 1 ate d equ1 . 1 1. b r1um· constant 5 shows that i:hc mixture should contain 83°u trans-1, 3-pentadiene relative to the tot~1l open chain 1, 3-pentadiene present, but after refluxing

.2or .20 l1ours, 69. -;--'~ tr~ns-1, 3-pcntadiene was detected and on reflu.xing for three days, only 60.996 trans-1, 3-pentadiene was detected.

The decrease in the trans-isomer on additional refluxing is

Jul: to the i_:1crcascd amount of polyrneri:ation and diiodide forma- ti.on. It has been sho\'.'11 that trans-1,3-pentadiene will dimerize 42 anJ polymeri:e much more rapidly than its cis-isomer This indicates that on longer refluxing, the amount of trans-1,3-pentadicne may increase, however, the polymerization reaction increases more rapidly. That is, the rate of polymerization and diiodide formation are faster than the addition or elimination of an iodine radi ca 1 from the adduct. Some of the possible but improbable side reactions are .sho\·Jn in Figure 5. As sho\vn by the thermodynamics, these siJe reactions should only occur to a small extent as compared to the cis- to trans-isomeri:ation of 1,3-pentadiene.

Potassium tert-8utoxidc fsomcri:ation of 1,3-Pentadiene

The tabuL1tcJ results for the potassium tert-butoxide isomer- i:ation of 1,3-pentadiene, both neat and in dimethylsulfoxide, are shown in Table VIII.

The effect of dimethylsulfoxide on the isomerization is appar- ent through a comparison of runs A and B. There \vas little, if any, isomcri:ation in run A, perhaps due to the insolubility of potassium 46

I I I I ~I I·1l

l1H298 = 10.7 l1G298 = 8.6

( )

J, 11 ~~ ~<~)~ •

1' RH -H:!gs = 13.0 .,.. G = 7.41 298 ~8H298 = 5.6 8G298 = 4. 8

F· gure 5. Possible Side Reactions During Iodine Catalyzed Isomerization of 1,3-Pentadiene. TABLE VIII

Data for the Potassium Tcrt-13utoxidc Catalyzed Isomerization of 1,3-PcntaJicne

52 + - 0 0 CalculateJ Contact D~lSO K t-BuO ·u Trans l$omcr o Temp. Trans Isomer * ·.1; ·)c 0 ·k K K Run (oC) Time (hr) Vol un1e ~D ~lol e ?o in Feed ·x* lll Product ** t/c tIc

A 25 18 neat 5.9 51.2 54.0 1 . l -; 5.37

B 25 24 so 5.9 51.2 79.8 3.95 5.37

c 20 24 70 44.0 57.6 83.6 5.10 5.63

0 20 24 90 36. 7 57.6 82.8 4.81 5.63

E 35 6 70 44.0 57.6 82.3 4.65 5.07

·x Based on 1, 3-pentadiene in feed. ** Based on 1,3-pcntadiene content. 2.303R log Kt/c = - ( 0. 14± 0. OS) + (1 037± 28) T with R in cal/mole °K 48 tert-butoxide in the piperylene. In contrast, in the presence of dimethylsulfoxide, the isomerization approaches the thermodynamic equilibrium value.

Dimethylsulfoxide enhances the rate of isomerization reaction 91 through its ability to: solvate the potassium ion , and activate 92 the weakly acidic allylic carbon-hydrogen bond . These abilities f ac1"1" 1tate t h e f ormat1on. o f t h e car b an1on . 1nterme . d" 1ate 92-95 . It has been postulated92 ' 94 , that the carbanion transition state, formed b ~~ the abstraction of an allylic hydrogen by the base, is the rate determining step. If this is true, then any factor which will decrease the energy of the transition state will increase the rate of its formation. The ability of dimethylsulfoxide to activate the acidity of the allylic hydrogen, must then increase the rate of intermediate formation enabling the thermodynamic equilibrium to be established sooner. This is supported by the data collected for runs C and Also seen, is the lack of substantial difference in the extent of isomerization in the presence of excess solvent or base.

Another factor which effects the ease of allylic hydrogen abstraction is the coplanarity of the alkene or diene system. As seen in the illustration below, a coplanar system must exist to " ~r~/ + B ~ j::U,,'"~ H .• • H- B 49

insure the max.1mum overlap bet1veen the TI bond and the rehybridizing p-bonJ oi the ~llylic carbon. This situation should not present

Jifficulties \·lith open-chain molecules such as 1,3-pentadiene, but has been sho\vn to play an important role in the isomerization of 96 ring systems

Through several base catalyzed isomerization studies of simple alkenes, the intramolecular cyclic transition state shown 97 bela~ ..· hJ.s been suggested . This suggestion is based on the strongly negative entropy of activation and observation that the isomeri:ation reaction takes place about sixteen times faster than the hydrogen-deuterium exchange reaction.

fhe proposed mechanism for the potassium tert-butoxide isomerization of 1,3-pentadiene is shown in Figure 6. This mechanism 1s a modi.fication of the olefin isomerization mechanisms given by 12 9 9 3 S c.nr1es, · I1c1m · ~ ' ..t., p r1ce· S , an d Ban~ kg , wh o postu 1 ate d th at th1 e intermediates \vere allylic carbanions. The proposed mechanism also shows ho\v small amounts of 1, 4-pentadiene and cyclopentene could form.

The major drawback found during this isomerization was the low so

9.. I KB + ~ ~ + KBH + I \ e . ~1 () ~ . ~ I\ ! 'V BKH ® r::::- HBK ~ ..; " 7 • / ~~ e· /1 , """ y I \ BKH (£) Ve ~e

BKH@ I

. ~

BKH ® ~ ~ 0 9

B = t-butoxide

Figure 6. Proposed Mechanism for the Potassium tert-Butoxide Catalyzed Isomerization of 1,3-Pentadiene. 51 recoveries of the pipcrylenc (

::'olymcri:ation.

The Svnt:1csis and !1vrol·.rsis of 3-i·.Iethyl-1,~,3,6-Tetr::lhydrophthalic

.\nhyJri Jc

The Diels-Alder reaction of trans-1,3-pentadiene and maleic anhydride to form 3-methyl-1, 2, 3, 6-tetrahydrophthalic anhydride was used to separate trans-1,3-pentadiene from cis-1,3-pentadiene and

:::12 other hydrocarbons present in the piperylene concentrate. The aJduct proJuced was then decomposed by pyrolysis. The first step

\~as easily accomplished since trans-1, 3-pentadiene reacts readily

\\'ith maleic anhydride \oJhile cis-1,3-pentadiene does not react at . ll 43,44,46,47 d . d" h 1 an\· apprec1a J c rate ue to 1ts protru 1ng met y group.

As indicated earlier (Experimental section), long contact times

(.:~l to 48 hours) were employed during this adduct formation. Tnese 47 long contact times arc not requ1re. d . Accor d"1ng to cra1g . , t h e rcJction \oJi 11 be compl cted in twelve hours. This was confirmed in l:.1tcr experiments in which similar yields were obtained 1n twelve hours -.)r less.

The adduct produced was then pyrolyzed in an attempt to produce pure tr~:ns-1,3-pentadicnc. TI1is pyrolysis was carried out in the apparatus depicted in Figure 1. On pyrolysis of the adduct, trans-

1,3-pcntadiene was obtained in fair yields. However, once freed from the adduct it was able to undergo· side reactions as it passed

Jo\._.n the heatcJ pyrolysis tube. The products obtained along with reaction conditions are sho\vn in Table IX. 52

TABLE IX

Products Obtained from the Pyrolysis of 3-~lcthyl-1, 2, 3, 6-Tet rahydrophthalic Anhydride

Reactor Temperature:

Contact Time: 5-6 sec

Recovery (as liquid): 38 wt %

ComT'u~i::ion of Liauid Products lTI Wei.ght Percent

Boiling Point oc 38-50 50-85 trans-1,3-pentadiene 25.3 30.8 cis-1,3-pcntadiene 14.8 20.8

1,3-butadiene 16.3 2.2 cyclopcntcne 1.~ 2.1 cyclopentadienc 34.6 38.9

Remainder * 7.6 5.2

.:.: C ~nd hea\·ier hydrocarbons 6 53

Th\..!rm~ l and C~ttalyt ic DehyJ.roc~.-cl i zat ion of 1, 3-Pentadiene

The ~~tuiument utili:ed in these runs \·.ras illustrated in

Fi:;urc .). The objective in constructing this apparatus was to

..:cvi_;..; .1 systcl!l \·:hich ~\·oulJ produce and reproduce consistant data. The system also had to be easy to operate and maintain.

The largest maintenance problem encountered '~as the clogging of the system due to polymer and coke formation. The two major fa~tors responsible for the excessive polymeri:ation Here found to be: 1) the presence of heavy material in the piperylene feed

.:1nd 2) the temperature of the hydrocarbon vaporizer. Distillation of the feed through a nine inch glass colunm packed with Penn

State Extruded Packing \~as sufficient to remove all of the heavy m3tcrials.

The extent of cracking and polymerization was also decreased by decreasing the temperature of the hydrocarbon vaporizer. It

,,.a~; oh:~crvr2d that if the dilutent gases were heated to a high

1 tenpcraturc b e f ore contact1ng· the· il~'droc,.,rbon! u feed, the amount of... crJ.cking wi.thin the reactor decreased. The extent of thermal cracking ~as also reduced considerably in the presence of steam.

The ste:1111 can bcnefi t the reaction in several ways:

1) Lowering the partial pressure of the piperylene allows the conditions to become more favorable toward the isomerization and cycli=ation reactions.

~) Removing carbon deposits by a "Water Gas r~eaction" during the

JchyJrocycli:ation, and Juring regeneration of the catalytic system. 54

c + IL>O l. ~ + (SJ ... g) co (g)

3) The: l "'.c~t ~apacity or· the super-he~ted steam can provide some of

the :1c-:.1t :1ccdcd for the Jehydrocycli::ation l'Caction.

The cracking, isomerization, and polymerization reactions of piperylene \.;ere found to be sensitive to the reactor surface. Due to this surface sensitivity, a small reactor was utilized. In

cat31Yst evaluation, this small reactor insured that the principle

reaction ·,,·as occurring on the catalyst surface rather than the walls of the reactor. To further decrease surface reactions, a greater

... L::o~nt vi .i.i1ert gas \'.'as used to reduce the diffusion of the piperylcnc molecules to the reactor walls.

The reaction conditions which were kept constant for all runs arc shoh·n in Table X. These conditions represent a compromise bcth·ccn optimum :1nd feasible conditions related to the equipment

limitations. For example, less cracking and higher selectivities we rc obscrvcu when the hydroc~rbon contributed S~o rather than 15?.5 to tile total vo 1 umetric feeJ. flO\'¥; ho'''ever, the latter was chosen clue to the inaccuracy of the syringe pump at the lo\ver settings

rcqui.rcJ for the former.

The data presented in this report are best understood if the fcllo\ving uefini tions are recogni:ed.

1) Key L:o1.1ponents - cyclopcntenc, trans-1, 3-pentadiene, and

cis-1,3-pentadienc.

~) Cant act Time - TI1e vo l wne of the reactor or unexpanded catalyst 55

TABLE \

Constant Dehydrocyclization Reaction Parameters

Steam Generator - Stage 1 125°C

Steam Generator - Stage 2 liyJroc:lrbon rapori:cr

Prchcater reactor temp. - 125°C

Sample Port

Hydrocarbon Trap

\"o 1 umc ~o Di 1utent 85% (75% H 0/10% N ) 2 2

Volume (~ Hydrocarbon 56

bed JiviJeJ b\r the volumetric feed flo~..,· rate at reaction

conditions.

3) Convcrs ion - The difference ln the ~.:eights of the key components

entering and leaving the reactor divided by the former, all

multiplied by 100.

4) Yield of Cyclopentn.diene - Weight of cyclopentadiene 1n the

effluent divided by the Neight of the key components in the

fccJ. all multiplied by 100.

5) Selectivity - Yield divided by the conversion, all multiplied

by 100.

To obtain an indication of the relative reactivities of the components in the piperylene concentrate, several runs utilizing

"pure" component feeds were made. The composition of these feeds arc shotvn in Table XI.

The relative reactivities based on conversion are: tr~lns-1,3-pcntadlenc > ~-methyl-1-butcne > cyclopentene >>

2-methyl-2-butcne >>> cis-1,3-pentadiene however. relative selectivities to cyclopcntadiene formation are: eye lupentcnc > cis-1, 3-pentadlene ~ trans-1, 3-pentadiene >>>>

2-methvl-1-butene > ~-methyl-2-butenc.

The Jat~ sho\vn in Table XI I was obtained by passing the above individual feeds through the packed (316 SS) reactor in a diluting stream of steam and nitrogen for one seconJ at 650°C.

·nH~ major reactions for l, 3-pentaJiene were isomerization and cyclization. TI1c isomerization of trans-1,3-pentadiene to the cis Ti\BLL XI

Colllposition by \\'eight of Lhe "Purc 11 Component Feeds

\Veigt1t Percent

trans-1,3- cis-1,3- 2-methy1- 2-mcthyl- cyc1opL;ntene isoprene Feed pentaJ.ienc pentadicne 1-butcnc 2-buteHc trans-1,3-pcntadienc 91.0 7.0 2.0 cyc1opentene 99.5

2-methyl-1-butene 99.6

2-methy1-2-butene 98.1 ] . 3 piperylene 69.5 12.6 16.1 1.8 58

TABLE XII

Dehydrocycli:ation Products Obtained from the ''Pure:' Component: Feeds

Feed ComE anent trans -1,3- cyclo- 2-methyl- 2-methyl- 0 \\'t ·v Composition pentadiene pentene 1-butene 2-butene

1,3-butadiene 10.7 0.4 0.2 0.6

2-methyl-1-butene 61.7 13.6

~- ;nethy l-:-butene 9.~ 66.:

3-methyl-1-butenc 9.7 4.3

isoprene 0.5 7.1 8.2 trans-1,3-pcntadiene 45.4 0.6 0.6 0.3

cis-1,3-pentadiene 26.5 0.4 0.1

cyclopentcne 1.8 65.5 <0.1

cyclopentaJicne 5.1 26.5 0.6 <0.1 59

. b , . I . 98 1somcr ecomes more preaomln~t as t 1e temperature 1s 1ncreased .

Th i.s J.ccotmts for the relatively lo\v conversion of the cis 1somer.

1t also indicates higher temperatures will be required to accomplish the c~·cli:~tion, since the reactive intermediate must. be in the cis configuration.

As sho\m, the dehydrogenation is the most prominant reaction for cyclopentene, \oJhi le the methylbutenes both isomerize (H-Shift)

;tnJ JchyJro~cnate readily. The products 3.re tabulated according to retention times and identity in Table XIII.

The results for the thermal conversions of piperylene within bot}1 packed and unpacked reactors are presented in Table XIV. The conversions are lower in the absence of any packing, however, the sclccti\'ities are greater. The maximum once through yields of cyclopentadiene were obtained at 700°C in both the packed and unpackcJ reactors. In the unpacked reactor the reaction was 38~o selective toward producing 18 weight percent cyclopentadiene, while only .:::s l~ selective to yield an equivalent amount of product in the prcs0nce of the packing.

The effect of surface on the convers1on of piperylene is illustratcJ in Figure 7. The conversions in the reactor packed with stainless steel are greater, indicating a surface reaction. TI1e maximum selectivity is less, and shifted to the lower temperatures for the pa.ckcd versus the tmpa.cked reactor. This indicates the optimum temperature for the reaction ls lower in the presence of the stainless steel packing, but a more efficient conversion takes 60

TABLE XIII

Retention Time and Identity of Products Found jn Effluent Stream

Retention Time* Peak ~umber (min) Identity

1 3.26 methane / 3.45 ethane/ethylene 3 3. 76 propane

-t' -L 01 propylene 5 4.55 1-butene 6 5.06

-I 5.46 trans-.2-butene 8 5.80 cis-2-butene 9 6.24 10 6.58 1,3-butadiene 11 7.29

1~ 7.85 ~-methvl-1-butene.; 13 8.35 14 8.89 2-methyl-2-butene

15 10.~2 16 11.53 isoprene 17 12.14 cyclopentene 18 13.73 trans-1,3-pentadiene 19 14.77 cis-1,3-pentadiene

~0 16.~2 cyclopentadiene

* Retention time decreases as column ages. TABLE XIV

Results of the TherJo~LJlly InitiatcJ DchyJrocycli:ation of Pipcrylcne j n a Packed and Unpacked I~cactor

Total K t/c Temp. Contact I Iydrocu rh on Selectivity to }\ Run jcC) Time (sec) Dilutent Packing Conversion Cyclopcntadiene t/c ---Calc.

Al 600 1.0 steam s.u 34.0 1. 7·~ 1.69

A2 650 l.U steam -- . 14.3 42.7 1.6b 1. 6·l

A3 700 1.0 steam 43. <-l 37.8 1. 63 1.59

A4 750 1.0 steam 88.2 1.9 1.42 1.55

AS 550 1.0 steam 316SS 2 • Jc 30.8 2.08 1.73

A6 600 1.0 steam 316SS 10.5 33.3 1.79 1.69

A7 650 1.0 steam 316SS 27.3 33.3 1 . 7.3 1.64

A8 700 1.0 steam 316SS 64.3 27.5 1.71 1.59

A9 550 1.0 nitrogen 316SS 2.0 41.1 2.28 1.73

AlO 600 1.0 nitro 5a en 316SS 9.8 28.7 1.74 1.69

All 650 1.0 nitrogen 316SS 73.4 20.3 1.67 1.64

0\ 1--l 1ABLE XIV (cont.)

Results of the Thermally Initiated Dchydrocyclizatlon of Piperylcne ]n a PackcJ and Unpacked Reactor

\\'eight Percent in Product Trans-1,3- Cis-1,3- Cyclo- 2-~.Icthy-f~ 1-tilltL·ne Run Cyclopentene Pentadicne Pentadiene Pentadicne 1,3-Butadicne and J- PcntC11e

Al 15.4 45.9 26.4 ] . 7 1.5 6.o

A2 13.2 41.4 25.0 G.1 4.8 5.0

A3 5.0 25.2 15.5 lb.4 15.9 3.7

A4 3.2 4. 7 3.3 1 . 7 2.1 1.7

AS 17.6 53.0 25.5 0.8 0.5 0.5

A6 15.6 46.1 25. 7 3.5 3.1 0.6

A7 11.2 37.6 21.7 9.1 9.0 0.7

AS 4.6 22.9 13.4 17.7 17.9 0.8

A9 17.1 54.2 23.8 0.7 0.3 0.6

AlO 16.0 45.7 26.2 2.8 3.0 0.7

All 3.1 14.9 8.9 15.3 20.1

Q\ N Figure '· Effect of Temperature on the Conversion of Piperylene and Selectivity to Cyclopentadiene in a Packed and Unpacked Reactor. 64 place i.n its absence. The stainless steel is apparently acting as a catalyst for the desired Jehydrocycli:ation reaction, however, it is :1lso .Lctin.s, to a greater extent, to adsorb and crac1: the pipcrylene.

The presence of steam increases the selectivity of the reaction by reducing cracking. This is shown in Figure 8. The ste3m may exert this effect by maintaining a clean metallic surface free r~ro;n carbon or met31 carbide. ~Ji trogen used alone as the dilutcnt does not have this effect. To support this assumption, the metal surfaces \"·ere coked by passing the gaseous feed through the reactor at a very slow rate in order to crack it severely.

Some condensed data obtained for the reaction over both a "clean"

(Runs 1:\ and~:\), and coked (Runs lB and 2B) surface are in Table XV.

As ii1JicateJ by the higher conversions and lower selectivities

(Runs lB and ~B), the coke effects the type of reaction occurring

\-Ji thin the reactor. Since the undesirable cracking reaction occurred more extensively in the presence of excess coke, steam was routinely employcJ as the Jilutent to help remove the carbon deposits during operation.

Figure 9 depicts the effect of contact time on the conversion and selectivity of the reaction. As the contact time increases, so docs the conversion. This is expected, since at the longer contact times the true thermodynamic equilibrium for the primary

But the lonoer contact times also reactions arc being approached. ~ enables the undesired secondary reactions of the pr1mary products 65

100

90 60

o\o 0 so (/) (l) 70 (1) s:: ...... (l) ro ·~ n "'0 • rt ("j :J ..... ~ 60 40 <: c:: ~· + + rt Co.. '< I t") n .. so "< ...... n 1-' c 30 0 0 ""0 •P'f 0 ~ .,. 0 ,_j ,..... rt (l) ~ 0.. >c ..... 0 ro ::l u 30 I 20 CD Q~ /, 20 ;· ' 10 '/ 10 ~ = ~/

550 600 650 700 Temperature (oC) Legend:

Stearn ..., ~ J ~ Nitrogen

F·gure 8. Effect of Temperature on the Conversion of Piperylene and Selectivity to Cyclopentadiene in the Presence of Nitrogen and Steam Dilutents. TABLE XV

Cracking Data Obtained in a Clean anu Coked l~vactor

Temperatur~ Contact Conversion of Selectivity to Run (oC) Tj lltC (sec) ~ ipcrylcnt! Cyclopcntadienc Cole lA 600 1.2 38.7 35.1 no

1!3 600 1.2 75.8 27.3 yes

2A 650 1.0 73.4 20.3 nu

213 650 1.0 76.2 1.8 yes

Q\ 0\ 67

100 -f I 90 j_ I 60

lI 80 •

o\o

so (f) n> 70 ro c: ~ Q) ro •M / (') -o rt / ...... _.;'"' 60 / < / 40 ...... ='.) rt '< I / t'i n ... 50 '< (') ~ ,...._. I I c 0 c 30 ~ ' , -- ---·---- (i) ·~ ~ •r .! () ""T'" ;... t ='r+ CJ p.l ~ +- ' -+ 0.. c 1-'. c (i) u 30 1J ::l 20 (i) ~":) t I 20

10 10 • ••

0.~ 0.5 1.0 1.5 2.0 Contact Time (sec I

Figure 9. Effect of Contact Time on the Conversion of Piperylene and Selectivity to Cyclopentadiene Over a Stainless Steel Bed. 68 to occur. The _;cl ecti \"i ty incrcn.ses tvi th increasing contact time, rc~chcd J. max1rnum at about 0.8 seconds, and then decreases. This inJi~atcs a cont~ct time of 0.8 seconds as the ontimum under the ..I. spec~ fi cd conJi tions. .-\bove 3.nd below this point, the secondary reactions are predominant.

Both the conversion of the piperylene and the selectivity toward cyclopentadiene are increased in the presence of hydrogen sulfiJ~}. :\ sum.rnarv or the results obtained using hydrogen sulfide as a promoter are shown in Table XVI. The increase 1n conversion and selectivity 1n the presence of hydrogen sulfide at various temperatures 1s compared to that 1n the presence of nitrogen as the dilutent in Figure 10, and to that in the presence of steam as the dilutcnt 1n Figure 11. In the presence of hydrogen sulfide, the con\'crsion increases linearly with increasing temperature,

\\"hcrcas in the presence of nitrogen, it increases slo\vly to ~10 9s

~t 600°C and then abruptly to 70~; at 650°C. In the presence of steam, this increase in conversion is less abrupt, being about 11 qo at o00°C and ~hS~, at 700°C. TI1e selectivity for the unpromoted rcJction tends to decline over the s1:.ated temperature range. llo,.,·cvcr, 111 the hydrogen sulfide promoted reaction_, the selectivity increases \vi th the temperature. This indicates that \vithout the promoter, the secondary reactions are occurring to a greater extent

\vi th increasing temperature.

The superior conversions and selectivities observed when hydrogen sulfide is used as a promoter are thought to be due to its TABLE XVI

Results of the llyJrogL:n Sulfide Promoted Dchydrocycliz.;ltjun of PipcrylcnL:

Total K Temp. Coil tact ~lo lc ~o or llydrocarbon Selectivity to t/..:. h (a C) T:imc ( S0C) Promoter P I\)motcr Conversion Cyclopcntadicne t/c Calc. Run --- --

B1 500 2.0 11 200 1 :~. 6 36.0 1.8~ 1.73 2s B2 600 2.0 II S 200 40.0 39.4 1.83 l.b9 2 B3 650 1.0 H S 50 43.0 41.0 1. 72 1.64 2 B4 650 1.0 II S 100 59.6 58.6 1.72 1.64 2 BS 650 1.0 ll s 200 50.6 53.8 1.6o 1.64 2 B6 650 2.0 H S 200 66.0 4 7. 9 1.77 1.64 2 B7 650 3.0 11 s 200 66.9 41.1 1.80 1.64 2 B8 700 2.0 li 200 91.8 53.2 2.07 1.59 2 s Results of thC:: Hydrogen S1tlfide PromotcJ DchyJro'"·yclizatiun of Pipcry1cnc

\~ejght Percent lTI ProthJct Trans-1,3- Cis-1,3- Cyclo- --~-_ ~I c thy 1 - l - But c n e Run Cyclopentenc PcntaJiene: Pcntadicnc ;1 ~uHI '11 CJl L' -----Pent J icnc 1,3-Butadiene ] - r\... t

B1 16.2 ~~ 3. 9 24.1 ·1. 9 0.4 3 . ()

B2 12.2 24.2 13.2 10.3 4.7 7.9

B3 9.9 28.9 16.8 17.6 8.0

B4 10.5 18. 1 10.5 3-'1. 9 6.8 3.3

BS 13.9 21.2 12.6 27.2 5.7

B6 10.3 14.6 8.2 31.6 8.5

B7 10.9 13.7 7.6 27.5 9.8

BB 3.4 3.1 1.5 48.8 7.5 71

100 +

90 90

80 /0 75 o\o (/) (1) 70 • ./ CD c I , ~ Q) CD •P"'4 n ~ rt ("j ...... +J 60 /cf < c: 60 ...... rt 0.. '< I t'} n ... so "< ~ n ~ c ~ 45 0 0 "'d ~ 0 + .,.I $-4 rt Q) ~~ Pl > ~ c !-'• 0 CD u 30 30 ('!)==' o·!=J

0 15 + 10 -----/ " 550 600 650 700 Temperature (oC) Legend:

J Nitrogen

._, J ~ Hydrogen Sulfide

Figure 10. Effect of Hydrogen Sulfide and Nitrogen on the Conversion of Piperylene and Selectivity to Cyclopentadiene at Various Temperatures. 72

! 1 ou ...

so 75 70

60 60

30 I ,. ~0 l I 15

I I 550 600 650 700 Tempera t ure ( OC)

Legend:

.) ~ Steam

IIydrogen Sulfide

Figure 11. Effect of Hydrogen Sulfide and Steam on the Conversion of Piperylene and Selectivity to Cyclopentadiene at Various Temperatures. 73

abilitY to act as a chain carrier. The proposed mechanism for the

' ...... ,. . . f I_j t h. 74 ~ . . promottJll reaCL.lon, :1 moa1rlcat1on o 1U c lngs mecnan1sm, 1s

dep.i.c:cd i.n Fi~urc 12. The mechanism for the thermal decomposition

of l, .3-pentadiene is sho\vn in Figure 13. In both mechanisms, the

initiating step is the homolysis of 1,3-pentadiene to form the

pentadienyl free-radical and a hydrogen atom. For the promoted

reaction, the hydrogen atom reacts with hydrogen sulfide to produce

r1olcculJr ~1ydro~en and a sulfhydryl free-radical (the chain carrier).

The chain carrier can then abstract an allylic hydrogen from

1,3-pent:ldiene to 11roduce the intennediate pentadienyl radical.

In order to produce cyclopentadiene, this radical must isomerize

to the cis configuration, transform to the boat conformation,

cycli:~ to the cyclopentadienyl radical which then can eliminate a

hyJro.:;cn :1tom to fonn cyclopentadiene. For the unpromoted reaction,

the chain carrier may be a methyl free-radical produced by the

crackin~ of 1,.3-pentadiene to butadiene.

rhc effect \vhich various amounts of hydrogen sulfide has on

the con\·crsion and selectivity is illustrated in Figure 14. As the

pcrccnta~e of hydrogen sulfide Ll00~~ = 1:1 mole ratio of piperylene .., c.onccntratc/ll...,S) is increased, both the conversion and selectivity

increase, pass through a maximum, and decrease in parallel. This

decrease at higher hydrogen sulfide concentrations might be explained

by the follow:i.n.g scheme which invol vcs the formation and consumption

of sulfhydryl radicals. 74

, I n i t i :it i on -/

rJropagation H· + ----..:; H + • SH 2

------.

~--

I I + H•

Tcrmin3tion ~ • SH + H· H2S

~ • SH + • SH ...:--- H2S2

H7 S,., ~ H") + s_.,

- '\ Ovcr:1ll I .\ +

Fi gurc 12. ~lechanism for the Hydrogen Sulfide Promoted Dchydrocycllzation of 1,3-Pentadiene. 75

Initiation

Propagation H· + )

> ~ + • CH 3

> ~. +CH4

~ ~ • ~ ~

~ ~ 111 + H· JV ~

Termination

Overall 2

Figure 13. Mechanism for the Thermally Initiated Dehydrocyclization of 1,3-Pentadiene. 76

100 (_ l I I ~0 + 90

75 o\o en

so 100 150 200

~ Hydrogen Sulfide

Fi.~;urc 14. Effect of Hydrogen Sulfide Concentration on the Conversion of Piperylene and Selectivity to Cyclopentadiene. 77

l I • + Ii.,S ----}- H2 + HS •

~' ' --../';/-...... ,'-...... / . + H S -----? ~''-~"-.. 2 + HS •

HS • + HS •

/ H.., +

The hydrogen sulfide can react \vith any free radical present, or

"'i th the re~ctor \-walls to form a sulfhydryl radical. With large

excesses oi hydrogen suliide, greater amounts of the sulfhyciryl

radical form. With increasing concentration of these radicals,

l:hc prooaoi 1 i ty th:1t they \\·ill react \"Vi th each other lS greater

than the prob:1bility of reacting with 1,3-pentadiene to form the pcntadicnyl radical. \\"hen t\\'O sulfhydryl radicals react, they can either dissociate back to the radicals or decompose to elemental hydrogen and sulfur. If this decomposition is significant:~ the

reaction uf the pentadiene will tend to follow the thermal reaction patiHvay rather than the promoted patln.;ay.

The effect \lihich contact time has on the convers1on and selectivity is illest:rated in Figure 15. As expected, the conversion of piperylene increases with increasing time spent inside the reactor. The relatively small slope of the conversion curve

illustrates the Jbility of the hydrogen sulfide to suppress the secondary reactions (c.f. Figure 9). The decrease 1n selectivity 78

100 l I I 90 + 90 I

75 o\.o (/) Q) 70 (1) c:: ~ C) ro • r-1 () ""'0 rt ,...... +-J 60 ~ < ::: / 60 ...... ".) rt '-": I .B--__ tr. n "' so • "< () --& - ~ r-..... 0 ,....._. ------45 '"0 ...... ro ..J. . ~ L) s.... I rt C) ~ > n. c ...... 0 (i) u 30 ::s 30 ro e,.':l f

~0 l I 15 10

tI 1. 0 1.5 2.0 2.5 3.0 Contact Time (sec)

Fl~urc 15. Effect of Contact Time on the Conversion of Piperylene and Selectivity to Cyclopentadiene in the Presence of Hydrogen Sulfide. 79

as the ~outact time 1s increased indicates that the secondary reactions ·.·ierc not ~:ompletcly suppressed. Also, over the range of contact t L1:1es observed, the selectivity \vas constantly decreasing

...1nd J.iJ not p3.Ss through a maximum as in the tmpromoted reaction

(c.f. Figure 9). This indicates the maximum selectivity might be at a contact time of less than one second at these reaction conditions.

The J~ta obtained for the reaction over the sulfided 316 stainless steel packing, both in a diluting stream of nitrogen and in sulfur dioxide, are shoh'Tl in Table XVII. The effect which the sulfided 316 stainless steel packing had on the conversion and selccti\·ity is compared to that of the metal itself in Figure 16.

'l11is comparison indicates the sulfided stainless steel causes less cracking or other side reactions than the unsulfided stainless st~el. The conversions of the piperylene are also more selective to cyclopcntadicne over the sulfided metal.

The conversion and selectivity with respect to temperature in the presence oi sulfur dioxide is sho\m in Figure 17. No firm conclusions can be made \vith regard to the pTesence of sulfur

JioxlJe Juc to the small nwnber of runs. Experimentation was terminated a short time after it began due to mechanical problems associated with it, i.e., rapid corrosion of brass needle valves.

In 11encral sulfur uioxic.le seems to increase the selectivity of .._"'0 ' the reaction ~t the lower temperatures but offers no assistance at the elevated temperatures employed. TAilLl: XV I I

Hesults of the Ochydrocyl:lization of Pii10rylcnc Over a Sulfide Stainless Steel BeJ in the Presence and Absence or ~;tllfur DjuAiJe

Promoter Total 0 K Temp. Contact or ~lole o llydrocarbon Selectivity to t/c ( OC) K Run Time (sec) Pack:ing Dilutent Promoter Convers:ion Cyclopent3Jicnc - t/~ ---Calc. Sul fidcd Cl 650 0.5 nitrogen 10.0 31.0 1.71 1.64 316SS C2 650 1.0 II nitrogen 29.4 35.7 1.70 1.64

C3 650 2.0 II nitrogen 45.6 28.9 1.73 1.64

C4 550 II 200 21.6 42.1 1.86 1.73 2.0 so 2 650 2.0 200 59.9 33.9 1. 85 1.64 cs " so 2 C6 700 2.0 so 200 85.3 24.4 1.45 1.59 " 2

00 0 TAbLE X\'1( (cont.)

Results of the Uchydrocyclization of Pipcrylcnc (J~,•cr a Sulfide Stainless Steel BL.:J in the Presence and Absence of ~~tllfur Dio.\.idc

\vci gh t Pcr~ent in 11 roduct Trans-1,3- Cis..:-1 3- Cyclo- 2-~Jcthyl-l- Butene ' Run Cyclopentene Pentadj enc: Pcntad:icne Pcntadicnc 1,3-Butadiene Jnd 1-Pcntcnc ------~ ------Cl 15.4 45.7 26.7 5.1 3.6 0.8

C2 11.6 36.1 21.2 10.5 8.4 1.7

C3 7.9 28.6 16.5 13. 2 13.5 2.2

C4 10.6 42.8 23.0 ~.1 1.3 4.3 cs 4.7 22.3 12.0 20.3 11.5 3.9

C6 1.3 7. 7 5.3 20.8 18.6 1. 2 82

100

90 90

75 o\o Ul Q) 70 C1) c:: ..._. Q) C1) ..... n "U rt ("j ...... +-J 60 < c 60 ..... ~ rt Q. / "< t") + n .. so "< ...... / ..._.(') c:: 0 0 p 45 "'0 ..... (1) VI 40 ::::3 r1' / /~ c ...... 0 (t) u 30 ::s 30 (i) oP /lo-r- / 20 rf' I + ~ 15 10 • I \ 0 b 0.1 0.5 1.0 1.5 2.0 Contact Time (sec)

Legend: , Sulfided 316 Stainless Steel , 316 Stainless Steel

Figure 16. Effect of a Stainless Steel and a Sulfided Stainless Steel Bed on the Conversion of Piperylene and Selectivity to Cyclopentadiene in Relation to Contact Time. 83

lUO l I 90 + 90 I '~ 80

I 75 o\o (f) C) 70 (1) c:: ...._. () ro •1""'1 I n .... I ,_..rt '~ ~ 60 ..I < :: 60 ,_.. :,; rt ·~ -I r'l / n so " '< ~ n..._. c I' 0 I 45 -:J 0 / 3t----___----- / () .; [) ~ :...... -- rt r:; --- PJ ;..... ---~- 0.. / ...... c= () L, 30 :::::1 +/ {,) / 30 '='::> ~ ~ '-.J 20 e 1- 15 10 I

550 600 650 700

Figure 17. Effect of Temperature on the Conversion of Piperylene and the Selectivity to Cyclopentadiene in the Presence of Sulfur Dioxide. 84

The Jat:.1 for the silica gel and alumina catalysts are shown in TJ.blc XVIII. The conversion and selectivity for the reaction un sili.c:.1 ~el \·w"ith respect to contact -rime are shown ln Figure 18.

The reaction on silica gel follows the same general trend as on the stainless steel packing (c.f. Figure 9), however, its surface is more inert in regard to cracking reactions. It does provide some cracking surface ho~ever, since the conversions are unselectively incr'~3scJ in its presence as compared to an empty reactor (c. f.

Figure 7). Due to the lack of acidity of the silica gel surface, the reaction should involve free-radicals and proceed by the mechJnism:-.f depicted in Figure 13. The alumina was found to be very acti \'C in converting the piperylene. This is shown in Figure 19.

The high Jctivity of the alumina is due to its acidity. The selecti\·ity of the reaction over the alumina follows an interesting path. At the short contact times, it is relatively high; however,

JS the contact time is increased, the selectivity decreases, passes throu~h a minimum and then slowly starts to increase. It may be, that J.s the cont.:1ct time increases from :ero to 1.5 seconds the amount of cracking \vill increase. At the longer contact times however, carbon is being deposited at a faster rate than it can be removed by the \vater gas reaction. TI1ese carbon deposits may block the acid sites of the alumina, thus decreasing cracking and increas-

1ng the selectivity of the primary reaction.

Olefin isomcri:ations, both double bond and skeletal, over alumina have been interpreted to involve carbonium ion mechanisms. TABLE \VI I I

l{csu1ts of the Dchydrocyc1ization o!" P.ipcry]t.;JlC Over Sil]ca Gel and Alu111ina Catalyst Beds

Total Temp. Contact Hydrocarbon Selectivity to Kt/c K Run ( OC) Time (sec) Catalyst Dilutent Conversion <:ycl opentadiene t/c Cal c..

Dl 600 0.5 SiO.J steam 5.9 16.9 2. 3l) 1.69 .... 02 600 1.0 5]02 steam 8.3 19.2 2.24 1.69 03 600 2.0 Si0 steam 20.2 1.69 2 11.4 2.17 04 650 0.5 Si0 steam 27.0 30.7 1.70 1.64 2 DS 650 1.0 Si0 steam 30.3 1.70 1.64 2 32.3 D6 650 2.0 S]02 steam 56.2 22.6 1. 71 1.64 07 650 3.0 Si0 steam 82.8 1.4:-1 1.64 2 16.1 08 650 5.0 Si0 steam 1. 75 1.64 2 97.1 0.5 09 550 1.0 A1 0 steam 10.5 10.0 2.01 1.73 2 3 D10 550 2.0 A1 steam 1.95 1.73 2o3 12.3 9.8 011 600 0.5 Al steam 1. 77 1.69 2o3 29.1 25.4 012 600 1.0 Al steam 1. 80 1.69 2o3 43.5 18.4

013 600 2.0 Al ')0'"? steam 54 .1 18.3 1. 79 1.69 .... .)

00 t.n T A~ .E XV 1 I I ( con t . )

Results of the Dehyd rocyc 1 i z at jon of Pi pcry L~.-·nc Over S i 1 i c a Gc 1 ~u1

\1/ej ght Percent ]_Jl 1>roduct Trans-1, 3- Cis-1,3- Cyclo- 2-~lethyl-1-Butene Run Cyc1opentenc Pentadicne Pentad i.cnc Pcntadicnc 1,3-Butadicnc :llld 1-Pcntcne ------...... ---- 01 17.6 51.8 21.9 1.0 0.3 1. 6

') ~· 02 17.5 49.7 22.2 1.6 0.5 .... ,.)

03 17.0 41.6 19.2 2.3 1.2 -~. 9 04 13.1 36.5 21.5 8.3 5.8 1.8 DS 11.6 34. 8 20.5 9.8 7.5 1.8 D6 7.3 22.2 13.0 12.7 9.3 3.0 07 2.9 8.0 5.6 13.3 12.1 6.6 08 0.6 1.4 0.8 0.5 0.3 0.2 D9 16.6 46.4 23.0 1.0 0.6 1.9 010 16.8 45.1 23.1 1.2 0.7 1. 6 011 11.7 36.4 20.6 7.4 1.6 8.0 012 9.1 29.4 16.3 8.0 2.0 10.8 013 6.7 24.3 13.5 10.0 2.6 11.9 87

I I .30 T 30 I

o\O

(/) ~ (t) s:: ...... u CD 20 + • 20 n --' r+ ·~ I p....J• ._J - ...... < c ------)o-1• () c r+ '< I "" ' t.r: " n " \, '< () " ~ ,- _,.. 0 c \ '"d +- ro Cf'. ;:::j s... / r+ / • () / PJ 10 / 10 > ~ ' c.. l !-"· =...... / I ~ c._; e· ;::::1 ~/ (D :lo ~------

0.5 1.0 1.5 2.0 Contact Time (sec)

Fi~ure lS. Effect of Contact Time on the Conversion of '' Piperylene and the Selectivity to Cyclopentadiene Over Silica Gel. 88

10() l l YO t l 80

75 o\O (f) Q) 70 (1) c ~ ~ ro •.-1 n v rt ~ 1-'· +-l 60 < c 60 1-'• ':.) rt '...-: - ~ -I I rr: n so ~_.,------'-< ~ / n ~ ,/"" :::: 45 .__.0 c .:t' -..J • .-1 , / (D ...,J •r. 40 ~ :... I rt (, ~ / ;..... /' 0.. c I-'• 0 (D u / ~ 30 30 (i) ::. ::. • t U,. ~() + 15

o.s 1. 0 1.5 2.0 Contact Time (sec)

Figure 19. Effect of Contact Time on the Conversion of Piperylene and the Selectivity to Cyclopentadiene Over Alumina. 89

The results obtained in this study, including cyclization reactions of pipcrylcnc, ~auld ln\·o l ve the carbonium ion mechanism proposed

1n Fi6ure 20. The presence of the 1-pentene and 2-methyl-1-butene found ln ~he effluent ~an be explained via this mechanism along

\vi th the isomeri: at ion ability of alumina. That is, the 3-methyl-

!-butene sho\.;n, can isomerize to 2-methyl-1-butene readily. The cycli:ation reaction should occur only to a limited extent, since it in\·ol vcs an .intermediate first degree carbonium ion. TI1e cyclopentadicne found, may be produced by the thermal dehydrogenation of the cyclopentene formed as sho\vTI by the mechanism below:

I 0 0 ' • ) 0 or Jircctly from 1,3-pentadiene adsorbed on the metal ('M).

Durj ng all runs, heavy material \.;as obtained. The heavies

\vhich ,,ere identified arc: JicyclopenTadiene, m-xylene, benzene, toluene, cyclohcxane, and cyclohexene. These products, except

Jicyclopentadiene, are formed through the dimerization of 1,3- pent aJicnc \1/hich has been sho\vn to proceed as follows, to form

l)l) 3-mcthyl-5-propenylcyclohexene-·. 90

"7' ~ R-H _, / //-.._ ',/ ...... , ...... , / ~- [ i· ' .,. R-H I I --.±. '! ~ \ "/

r. . 'I("' 1·1 guro - '. Proposed Carbonium Ion 0Iechanism for the Cycli:~tion of 1,3-Pentadiene Over Alwnina. 91

,r;---:.. \ I ' + I .I ', \..__

These dimerizations are then followed by dealkylation, dehydrogenation, and isomerization, to yield the above mentioned products. 92

CONCLUSIONS

Surveying the results obtained 1n this study lead to the iollo\.:ing conclusions:

1. The isomers of 1,3-pentadiene can be separated from the other components in piperylene concentrate through cuprous ammonium chloride complexing and subsequent decomposition of the complex.

2. Trans-1,3-pentadiene can be removed from piperylene concentrate via a Diels-Alder reaction with maleic anhydride. The reverse Diels-Alder (pyrolysis), to free the trans-1,3-pentadiene,

\\as found to produce many of the components of piperylene concentrate due to t!1e thermal cracking of the freed trans-1, 3-pentadiene.

~). The concentration of trans-1, 3-pentadiene in piperylene concentrate can be increased by refluxing with traces of iodine or catalytic a~ounts of potassiun tert-butoxide dissolved in dimethyl sulfoxide.

~- The isomers of 1,3-pentadiene can undergo dehydrocycli: ation tn a catalytic and non-catalytic reaction, for which a free radicJi mechanism is indicated.

5. The highest selectivity for the dehydrocyclization of

1,3-pentadienc was 60 %. This was achieved using 100 Tiole % hyJrogcn sulfide promotor, at 650°C and atmospheric pressure with a contact time of 1.0 seconds.

6. !'fatcT reduced cracking and prolonged catalyst life by reducing the carbonization of the catalytic surface through the 93 water ~as reaction.

7. The net!1od er.1ployed in vaporizing and preheating the

hvJrocarbon feed ~.:as critical in controlling the extent of cracking

·~,.ri t:J.in the reJ.ctOl'. It Has found, if the dilutent (most desirably

steam) \\as superheated before contacting the hydrocarbon feed the extent of cracking lvithin t!1e reactor was oinimized.

8. The relatively constant selectivities for the dehydro

cycli:a:i:Jn uf 1,3-pcntac.licne under thermal and heterogeneous catalytic conditions indicates that a substantial amount of competing

side reactions arc taking place wit!1in the preheater section of the

apparatus.

9. If this project is to be continued, it is suggested that

var 1ous horJogencous free-radical prol!lotors be enployed \~i th cmpl1asis on lower temperature operation. These may include ~Br,

(\H )S, various mercaptans, and the like. 4 10. It would also be desirable to purify the piperyleue

concentrate to ~ain information on the side reactions before

ernbarl,ing in ~ simular project. The most likely method seems to be an extractive distillation using various nitrogenous bases. REFERENCES

(1) Stinson, S. C. Chern. Eng. (fe,vs 1979, 57 (22), 32.

(2) 1 ~·oshiaki. ~\'. Chern. Econ. En?"._, •P\ev. ..!... 074...... ' 6 lr8)- ' oJ~6 •

(3) ~Jattson, G., University of Central Florida, Personal Communi cation, 1977.

(4) Soday, F. J. U.S. Patent 2 384 855, 1946.

(5) Gaylord, N. G. U.S. Patent 3 491 068, 1970.

( 6) :\i.e lly, T. ln "Encyclopedia of Chemical Processing and De si.sni'; ~·lcKctta, J. J., Ed.; ~larcell Dekker: New York, 1977; \'ol. 5, pp 110-157.

(7) \':irth, ~·1. ~1. Chern. Prod. 1956, 19, 352. , (8) \\'irth, ~I. ~1. Chern. Prod. 1956, 19, 400.

(9) Chandrasckaran, S.; ~.iark, H. F. In "Kirk-Othmer Encyclopedia of Chemical Technology", 2nd ed.; Standen, A., Ed.; John Wiley and Sons: ~ew York, 1964; Vol. 7, p 77.

(10) '\:ralo\·ec, R. B.; Louderback, H. B. In "Kirk-Othmer Encyclo pedia of Chemical Technology'!, 2nd ed.; Standen, A., Ed.; John Wiley and Sons: New York, 1964; Vol. 6, pp 688-703.

(11) ~·;ells, T. H.; :\'ilson, P. T. Chern. Rev. 1944,34, 1.

(12) "Chcmic3.l Technology: An i:ncyclopedic Treatment"; Harper and RoK: ~ew York, 1972; Vol. 5, pp 160-162.

(1.3) Kunt:, I. rn "Encyclopedia of Polymer Science and Technology", Bikalcs, 1\. ~1., Ed.; John ~'!ilcy and Sons: New York, 1966; Vol. ~, pp 563-567.

(14) Ferguson, L. N.; Paulson, D. R. '!Alicyclic Chemistry II"; FJ.anKlin Publishing Cornpa;1y: New Jersey, 1977, pp 47-85.

( 15) Lloyd~ 0. ''Alicyclic Compounds"; American Elsevier Publishing Company: New York, 1963; pp 91-93.

(16) Harkness, J. B.; et al. J. Chern. Phys. 1937, ~' 682. 95

(17) ~letcalf, R. L. In ''Kir~\:-Othmer Encyclopedia of Chemical Tech nology':, 2nd ed.; Standen, A. Ed.; John Wiley and Sons: New York, 1964; Vol. 11, pp 696-698.

( 1 S) P 1 i mmc r , .T • R . In ~'Pes tic ide Chemistry" ; T ah or i , A. S . Ed . ; Gordon Breach: ~Je\v York, 1972; Vol. I, pp 413-423.

(19) Brooks, G. T. "Chlorinated Insecticides~'; CRC Press: Cleveland, 1974; Vol. 1, Chapter 3.

(20) Morton, 01. "Science and Technology of Rubbern; Eirich, F. R. Ed.; Academic Press: New York, 1978; Chapter 2.

(21) ~1cBec, E. T.; Ungnard, H. E. Chern. Rev. 1958, 58, 249 .

. ...., ...., ) l-- Taketa, A. Chern. Econ. Eng. Rev. 1976, ~ (3), 26.

(23) ~cison, D. L. U.S. Patent 4 029 848, 1977.

(: ~) Lear~·. R. F. U.S. Patent 2 773 051, 1956.

(~5) Hamner, G. P.; Jones, T. G. U.S. Patent 2 750 359, 1956.

(26) Kudo, K.; et al. U.S. Patent 3 929 747, 1975.

(~7) Goldstein, R. F.; \Vaddams, A. L. "The Petroleum Chemical Indus try", 3rd cd.; E. & F.N. Span: London, 1967; Chapter 12.

(~8) Hughes, W. B. In "Homogeneous Catalysis - II"; Forster, D.; Roth, J. F., Eds.; American Chemical Society: Washington, D.C., 1974; Chapter 14.

(29) "Sulfolane Finds ;.Jew Uses In Extraction"; Chem. Eng. News 1964, 4~ r:o), so.

(30) Cr~ig, D. J. Am. Che~. Soc. 1946, 65, 1006. l31) Craig, D. U.S. Patent 7 347 667, 1944.

(3~) Craig, D. U.S. Patent 2 403 054, 1946.

(33) Long, R. B. AIChE Symp. Ser. 1972, 68 (120), 82.

(:>..f) Lur'c, ~I. A.; et al. Sintet. Kauchuk 1934, 3 (6), 19-29; Chen. Abstr. 1938, 32, 6099 .

(35) Lur'c:o ~I. A.; et al. Sintet. Kauchuk 193£~, .) (6), 13-18; Chern. Abstr. 1938, 32, 6099 . 96

r (36) \\'ard, ,\. L. ; nnkin, E. 1..... J. Am. Chern. Soc. 1947, 69, 657. ., (37) ;.torr is, L. c. U.S. Patent: .... 386 333, 1945 . (38) :.torr is, L. c. u.s. Patent -,.., 386 334, 1945.

(39) Scofield, R. C. U.S. Patent 2 963 517, 1960.

(40) Craig, D. U.S. Patent 2 322 281, 1943.

(41) B1ytas, G. C. Ger. Offen. 1 918 312, 1969.

(42) Frank, R. L.; et a1. J. Am. Chern. Soc. 1947, 69, 2313.

(·~3) C:rai~, ~- .J. :\m. Chem. Soc. 1950, 72, 1678.

(44) Robey, R. F.; Wiese, H. K. U.S. Patent 2 357 910, 1944.

( 45) Hitachi Cl1esica1 Company, Limited Fr. Patent 1 600 836, 1970.

(..l6) Robey, R. F.; et a1. J. Amer. Chern. Soc. 1941, 63, 627.

(47) Craig, D.; et al. J. ~~er. Chern. Soc. 1961, 83, 2885. l~SJ Kilpatrick, J. E.; et al. J. Research Nat1. Bur. Stand. 1949, 42, 225, Paper ~o. RP1964. l49) Douslin, D. R. Proc. Div. Refining, Aner. Petrol. Inst. 1970, 50, 189.

(50) ~Iesscrly, J. F.; et al. J. Chern. Eng. Data 1970, IS (2), 227. l 5 1) ~tu~ 11 , D . R, . ; et 2 1 . "Th• e ,,_.... h er!llCa . 1 T}. 1ernod ynam1cs. o f 0 rgan_tc. . Compounds"; John 1Vi1ey J.nd Sons: Ne\v York _, 1969; Chapter 9.

L52) Egger, K. W.; Benson, S.W. J. ~\m. Chern. Soc. 1965,87, 3311.

(S:;) Kistiakowsky, .J. R.; et al. J. Ar:1. Chern. Soc. 1936, 58, 146. t5Ln Laidler, 1\. J.: Loucks, L. F. In "Comprehensive Chenicnl Kin etics11; Bamford, C. H.; Tipper. C. F. H., Eds.; Elsevier Publishing Company: New York, 1972; Vol. 5, pp 28-36.

(SS) Cundoll, R. B. In nrrogrcss In Reaction Kinetics"; Porter, G., Ed.: Pergamon Press Limited: Oxford. 1964; Vol. 2, Chapter 4.

(56 ) l Icy , P . r L ; l\' ate r s , !V • ,\ • Chem . Rev . t 9 3 7 , : 1 ~ 16 9 .

(57) Golden, D. M.; et al. J. Am. Chern. Soc. 1964, 86, 5416. 97

(58) Benson, S. ~\r.; Bose, A. ~- J. Am. C!"lem. Soc. 1963, 85, 1385.

(59) Spe:iJ.lc, :\. J.; Bissilig, D. E. J. Am. Chern. Soc. 1963, 85, lSES.

(60) Benson,,'). 'v.; et al . .J. Am. CheEL Soc. 1965,87,468.

(61) Egger, K. \\'.;Benson, S. W. J. Am. Chern. Soc. 1965, 87, 3314.

(62) Rohrer, J. C.; Sinfelt, J. H. J. Phys. Chern. 1962, 66, 950.

(63) \':ells, P. B.; Wilson, G. R. Discuss. Faraday Soc. 1966, 41, 237.

(64) Shuikin. ::. I.~ et al. Bull. Acad. Sci. U.S.S.R., Div. Chen. Sci. 195-.l, 779-82 (Engl. Trans.) Izvest. Akad. Nauk S.S.S.R., otJe1. Khirrt. Nauk 1954, 898-903.

(65) Shuikin, ~- I.; Naryshkina, T. I. Bull. Acad. Sci. U.S.S.R., 0 1\·· ...... , ., ;1cr:1. Sc1.~ · 19-- ~b,, 1,..,._;/- 7 ~ s-.:> cr:..... ng.1 1ra:1s. J. r zvest ..L, ._·a. k d Nauk S.S.S.R., Otdel. Khirn. Nauk 1956, 1249-55.

(66) Kennedy, R. ~1. ; Hetzel, s. J. U.S. Patent I 438 398, 1948.

(6 7) Kenncdv, R. ~I. ; Het:el, s. J. Ind. Eng. Chern. 1950, 42, 547.

(68) Kenncdv, R. t.t . ; et al. U.S. Patent 2 438 401, 1948.

(69) Kennedy, R. ~1. ; et al. u.s. Patent 2 438 403, 1948.

(70) Tir.1ashcv, P. s. ; et al. Khim. Prom. 1968, -4t:- (2) , 106. ( 7 1) Grigorovich, B. A. ; et al. Russian Patent 186 446, 1966.

'T' (72) Bodnaryuk. 1 • et al. ~eftekhimiya 1969, 9 (S) ' 707-15. s. ; -

I (7 3) Bounaryuk, T. s. ; Grigorovich, B .. A. ~eftekhiniya 1967, I ( 2) ' 208-13.

(7-+) Hutc~1ings, D. A.; et al. :-\r.1er. Cher.1. Soc., Div. Petrol. Chern., Prepr. 1972, 17 (2), B37-B45.

(75) Frech, K. J.; et al. In "Ir!dustrial and Laboratory Pyrolysesn; Albright, L. F.; Crynes, B. L., Eds.; American Cher!lical Soci ety: Washington, D.C., 1976; Chapter 12.

(76) Frech, K. J.; ct al. Ger. Offen. 2 126 857, 1971. 98

(77) Bodnaryyuk, T. S.; et al. Neftekhimiya 1967, 7 (3), 389-91.

(78) 3enesi, !I. :\. Ger. Offen. ') 535 809, 1976.

(79) Shuikin, :.J. I.; ~aryshkina, T. I. World Petrol. Congr., Proc. 5th, :;e\v York, 1959, 4, 109-19 (Pub. 1960).

(80) Shuikin, N. I.; Tulupov, V. A. Bull. Acad. Sci. U.S.S.R., Div. Chern. Sci. (Engl. Trans.) 1955, 785-93; Invest. Akad. Nauk S.S. S.R., Otdel. Khi@. Nauk, 1955, 869-79.

(81) Kennedy, R. ~1.; et a1. U.S. Patent 2 438 402, 1948.

(8~) Hetzel, S. J.; et a1. U.S. Patent 2 438 404, 1948. lS3) !Iet:cl, S. J., et al. U.S. Patent 2 438 400, 1948.

(84) Kennedy, R. H.; et al. U.S. Patent 2 438 399, 1948.

(BS) (,ltl.s,~ · · ··~- 'l.'.; n... o:engart, '1,·. I. Izvest. Akad. Nauk S.S.S.R., S e r . Kh i rn . 19 7 5 , 7 , 16 8 5 .

(86) Diet:, \\'. A. J. Gas Chromatogr. 1967, 68, 6.

(87) Doh· Chemical Company A."1alytica1 ~!ethod #60719, Hay 18, 1977.

(88) :loffett, R. B. In norganic Synthsis Collective Volume IVn; Rab_io!1n, 0J. Ed.; John ~\'iley and Sons: New York, 1963; p 238.

(89) Bank, S.; et al. J. Amer. Chern. Soc. 1963, 85, 2115. l90) ~veis scr, 0. ; Landa, S. '!Sulphide Catalysts, Their Properties and .\pplications"; Pergamon Press: Oxford, 1973; Chapter 2.

(91) Ugclstad, J.; et al. Acta. C~ern. Scand. 1962, 16 (2), 323.

(92) Schriesheim, :\.; RO\-Je_, C.,\. Jr. J. Amer. Chern. Soc. 196~, 84, 3160.

(93) Bank., S.; ct al. J ..\mer. Chem. Soc. 1963_, 35, :115.

(94) Schriesheim, A.; et al. J. Amer. Chern. Soc. 1961, 83,

(95) Price, C. C.; Snyder, W. II. J. 1\rner. Chem. Soc. 1961,83,1773.

(96) Schricsheim, A; et al. J. Amer. Chern. Soc. 1962, 84, 3164. 99

(97) ~·1artin, D. ; Ilauthal, H. G. 1'Dimethyl Sulphoxiden, Halberstadt, E. S. Translater; John Wiley anti Sons: New York, 1975; Chapter I •

(98) Robey, R. F. Science 1942, 96, 470.

(99) f-r~nk, :t. L. ; ct al . Ind. Eng. Chern. 1947, _,39 893.