THE LNIVERSITY OF MICHIGAN INDUSTRY PROGRAM OF THE COLLEGE OF ENGINEERING KINETICS OF THE THERMAL DECOMPOSITION OF ETHANE TO ACETYLENE IN NONUNIFORM TEMPERATURE FIELDS Gordon Do Towell September 1960 IP-454

Doctoral Committee: Professor Joseph J. Martins Chairman Associate Professor Richard Bo Bernstein Professor Stuart W. Churchill Professor Guiseppe Parravano Professor Lawrence Ho Van Vlack

ACKNOWLEDGEMENTS The author wishes to express his appreciation to the members of his committee for their guidance during the course of this worko Special thanks are due to Professor Joseph Jo Martin who suggested the topic for this research and served as Chairman, and also to Professor Lawrence H. Van Vlack who served as Acting Chairman during the temporary absence of Professor J. Jo Martino The author is indebted to the Horace Ho Rackham School of Graduate Studies for financial aid through the award of University Fellowships for two yearso Finally, the author would like to thank his wife for her constant encouragement throughout and for typing the first draft of this manuscript. The preparation of the final form of this manuscript by the Industry Program of the College of Engineering is also very much appreciatedo -ii -

TABLE OF CONTENTS Page ACKNOWLEDGEMETSo e o o o o o o o oa o o. a o o o o o o a 6 o o o a D O a O a O O a O O * * ii ABSTRACT o,o,. o o. o. o o * o * * o0 o 9 0 o O O o o o o 111 LIST OF FIGURES.0 0 0 0 0 @ O 0 0 0 0 0 9 Oa O O O 0 0 0 * 0 0 a0,, O O 0 O O O O iX LIST OF APPENDICESo o.ooo o.. o.. o...... o ooooo oo o... o o 0 0 ii Io INTRODUCTION, * 99 o, o l Q 9 9 a o o o o o 0 o o o o.... o o o o0 0. o 9 00 o a 1 A. Scope of Research................................ o 1 B, Review of Literatureo o o o.,..o o o,,.,....,,,o o o o o o o,,,,,,,,,, o.. 3 Revew of Literatureo................................... 7* A. Introduction.o o o... o OOoO a o * 4 4 0 t> O o o O O O ~ 0 7 B. Reaction Equilibriao...................... oo.oo o... o 8 C, Temperature Distributiono ooooooooooooooo 9 ooo0oaa ooo 8 D, Determination of the Order of Reaction with an Arbitrary Temperature Distribution,..o.......o...... 11 E Derivation of the Kinetic Equations for an Arbitrary Temperature Distribution, o.....,,,o. oo...o, 14 F. Solution of the Equations for the Kinetic Parameters, 16 G, Calculation of Rate Constants and Smoothing of the Data. O 0 - - 6 0 a 9 O 9 a O O a O O O O O a O O a O O O O 0 0 a O O 0a 0 0 19 III. APPARATUS AND EXPERIMETAL TECENIQ UEo.... o o. o o.o o,., 21 Ao Description of Apparatus o.o..o 0... o...0............9. 21 B, Experimental Technique..oo.oooo...o...o.oooo*........ 27 C. Chemical Analysiso 9 9o,o.. Ooo.0 o 29 D. Experimental Program, o o o.... o o o ooO... 30 IV. EXPERIMENTAL RESULTS AND DISCUSSSION......o o o. o oo o o o. o. o 32 A, Thermal Decomposition of Ethane e........O..o...o..... 32 (i) Product Distribution,..,,,,oo.,.,....... 32 (ii) Order of Reaction...0b0*o000.obo900.o...o...... 34 (iii) Rate Constants, o o o 0 * O O O O O O.O o o o.. 34 (iv) Temperature Distribution..oooO............. 43 vi

TABLE OF CONTENTS CONT'D Page B. Thermal Decomposition of Ethylene..,....*............. 48 (i) Product Distribution,.......*o....,.....,,4** 48 (ii) Order of Reaction, o, ~......o,..e..oO. * 50 (iii) Rate Constants,..............<..**..***^*, o* 50 C, Thermal Decomposition of Acetylene...*... *.......,,,. 55 (i) Product Distribution,,, l.,..., o........ 55 (ii) Order of Reaction,,,.,.,.. *,.,. 58 (iii) Rate Constants,.......................... * 58 D< Inhibition of Thermal Decomposition..,......,,... 61 (i) Propylene Inhibition,,,. o.. *............,,. 61 (ii) Inhibition of Ethane Decomposition by Reaction Products65, *,........,,..oo o 65 V. REACTION MECHANIEMS.O ^,.,,.o....,, o..., o..,..., 69 A. Review of Literature on Reaction Mechanisms,.,,...... 69 B., Ethane Decomposition Reaction Mechanisms., o.....,,,* o, 71 C. Ethylene and Acetylene Decomposition Reaction Mechanisms o.. o. o *. O O....**. *...... o ~ ~ o*o.<~.. 78 VI. KINETIC CORRELATION FOR THE COMPLEETE SERIES OF REACTIONS IN THE FORMATION OF ACETYLENE FRC ETEYLE NE,..,,......... 80 A. Overall Kinetic Correlation..,....,oo,....,.....* * 80 B. Use of the Correlation to Predict Product Distributions Over a Wide Range of Reaction ConditionsO, * * o e *.*........ 0.* 0**.e* 00.. 0 OO***O 85; VII. CONCLUSIONS.o.. oo a O a o a 0,..... a * a a o0. 95 APPENDICES,,.o o,, o o o 0, o o0 o, o..... o o. o..o.,........ *... 98 BIBLIOGRAPHY,*.. 0 * a 6 a.,,*.a 0 0.......~ o.*. 0 0*. a.,... *, 0 *. o 123 NaMENCLATRE. a o,..... o. * *..... 126 vii

LIST OF TABLES Table Page I Comparison of Correlation with Experiment,..........., 84 II Raw Data: Product Distribution Experiments......*. o.* * 99 III Raw Data: Reaction Order Experiments............. 100 IV Raw Data: Rate Constant Experiments...... *.....,........ 101 V Raw Data: Inhibitor Experim ents,.........,........., *. 103 VI Results: Product Distribution Experiments....... * o.,,,*. 105 VII Results Reaction 0rder Experiments ~....... o,. o.*,,, 106 VIII Results: Rate Constant xperiments.........,........**.. 107 X Results: Inhibitor Experiments..................... 108 X Approximate Values of Bond Energies,.,,.....,, O.... 122 viii

LIST OF FIGURES Figure Page 1 EEquilibrium Constants,*.*................0............... 9 2 Schematic Temperature Profiles.................. 10 3 Solution for Kinetic Parameters,.,..0 0............ 18 4 Data Smoothing (Fictitious Points),....0 o o.0* o 0.... o..o. 20 5 Diagram of Apparatus...............oO*o*.~~ 22 6 Furnace Details. o....... o O *..... O... a o. * *. o 23 7 Constructional Details of the Packed Reactor (No. 1)..... 25 8 Constructional Details of the Annular Reactors (Nos, 2 and 3)...,.. *.... 0...0......* a... o'. a*..... o...o 26 9 Thermocouple Details,,.. o o. o o.................o.... 28 10 Product Distribution in Ethane Decomposition0............. 33 11 Ethane Decomposition Orders of Reaction...... o o o 0 0 0 0. 35 12 Typical Experimental Temperature Profile,..... o........o 37 13 First Order Rate Constants for the Disappearance of Ethane o....., 0* 0 * 000f *o o ~* * 0.o 6*000**.- *.. 0 0o obooo. 38 14 Solutions for All Possible Pairs from Four Points......0o 40 15 First Order Rate Constants for Formation of Ethylene and Methane from Ethane,..0 o........ o. * o. oo 0 41 16 First Order Rate Constants for Ethane Disappearance, a Comparison with Literature Data.......... o,.o..... 42 17 Radial Section Through Reactors Showing Location of Thermocouples*.., o * *.. o* * * O. 0 * * o 0 # 0 * * a# a 000 44 18 Effect of Radial Temperature Distribution,.. o...... 46 19 Product Distribution in Ethylene Decomposition........... 49 20 Ethylene Decomposition Order of Reaction, o................. 51 ix

LIST OF FIGURES CONT'D Figure Page 21 Ethylene Disappearance First Order Rate Constants......... 52 22 Acetylene Formation from Ethylene First Order Rate Constants.. o. o o o....... o...... o....Q....o o o O 53 23 Polymerization of Ethylene Second Order Rate Constants... 54 24 Ethylene Disappearance First Order Rate Constants, a Comparison with Literature Data*.,...o,.,...,......... 56 25 Product Distributions in Acetylene Decomposition.......... 57 26 Acetylene Decomposition Orders of Reaction..o.,. *.... 59 2.7 Acetylene Disappearance First Order Rate Constants....... 60 28 Carbon Formation from Acetylene First Order Rate ConStantS,00 *~< 0 0 0 00 0 0 000o 0 0 0 0 0 000 00000 ft 0 0o<o0 62 29 Acetylene Polymerization Second Order Rate Constants..... 63 30 Inhibition of Ethane Decomposition with Propylene......... 64 31 Inhibition of Ethylene Decomposition with Propylene....... 66 32 Inhibition of Ethane Decomposition with Butadiene........ 67 33 Chain Lengths for Ethane Decomposition with Propylene Inhibitiono.. *... o. o O.. o..o o O.... o o * * o o e o.. o. o. O o... 74 34 Chain Lengths for Ethane Decomposition,...........o...oo *. 75 35 Ethane Decomposition Activation Energy for Reaction Inhibited with Propylene*..*..o..,. a... o....... 77 36 Rate Constants for the Six Major Reactions.......o.....ooo 82 37 Effect of Ethane on Formation of Acetylene from Ethylene., 86 38 Calculated Product Distribution Throughout the Reactor for Run 129,,00.0,,0.,0o,00000o 000 O 87 39 Computed Product Distributions versus Residence Time (1 atm, no diluent, 1000~C and 1200~C)....,.oo000000o.oo0 89 x

LIST OF FIGURES CONT'D Figure Page 40 Computed Product Distributions versus Residence Time (1 atm, no diluent, 1400~C and 1600~C)....,..... oo 90 41 Computed Product Distributions versus Residence Time (1 atm, 75% diluent, 1000~C and 1200~C).....,... o 91 112 Computed Product Distributions versus Residence Time (1 atm, 75% diluent, 1400~C and 1600~C).....oo, 92 43 Computed Product Distributions versus Residence Time (0.25 atm, no diluent, 1000~C and 12000~C )... o.........* 93 44 Computed Product Distributions versus Residence Time (0.25 atm, no diluent, 1400oC and 1600oC)...,O..,,,,.O. 94 xi

LIST OF APPENDICES Appendix Page I Raw Exerimental Data~.............a.................. 98 II Results Calculated from Raw Data...,,........* o a....... 104 III Calculation of Correct Mean Temperature for a Linear Radial Temperature Distribution.... o....., o o.. a O o ao o 109 IV Sample Calculations.o.........o.,,............... 111 V Modification of Calculations to Handle Parallel Reactions. a a oa. a a o o a a0 0. a..... a.a a a. a a o o a. 114 VI Calculations to Show Effect of Reverse Reaction,......... 116 VII Flow Diagrams for Computer Programs,..................... 118 VIII Approximate Values of Bond Energies...a............... 122 xii

I. INTRODUCTION A. Scope of Research The main objective of this research is a kinetic study of the thermal decomposition of ethane to acetylene. The reaction conditions necessary are temperatures of about 10000C and higher, and residence times of the order of 1/100 of a second. The thermal decomposition of ethane consists mainly of a series of consecutive reactions proceeding through ethylene to acetylene and finally to carbon. A thermodynamic equilibrium limitation in the formation of acetylene from ethylene is the reason for the high temperature. The reactions are very fast at these elevated temperatures so that the residence time must be very short in order to prevent complete decomposition to carbon and hydrogen, A batch kinetic experiment is obviously out of the question for this rapid a reaction. The experimental program was carried out using a steady state flow system consisting essentially of an electrically heated tubular reactor through which the gas passed at high velocity The gas temperature distribution was measured with movable thermocouples. In a conventional kinetic study the experiments are carried out isothermallyo The reaction conditions necessary in this study made it impossible to even approach an isothermal experiment because of the heat transfer rate limitationo Therefore, the apparatus used in this work was designed to measure an accurate gas temperature distributionO A mathematical technique was developed for analyzing the non-isothermal kinetic data for the order of reaction and the rate constants. A digital computer was required to obtain solutions to the equations developed PL<-~

-2since the large amount of computation required would cause hand calculation to be prohibitively slow, This method of kinetic analysis of data obtained with an arbitrary temperature distribution is applicable to all kinetic studies and particularly to the case of fast high temperature thermal reactions. Previous workers in the literature have been faced with the problem of a non-uniform temperature distribution and the usual solution has been to estimate some average constant temperature for an arbitrary fraction of the reactor. This kind of approximation results in considerable scattering of the data0 Experiments carried out such that the whole series of consecutive reactions are occurring together are very difficult to analyze for the individual rate constants. Therefore, the thermal decompositions of pure ethane, ethylene and acetylene were investigated independently. In each of these independent studies the main reaction and the products of any significant side reactions were established and the order of reaction and the rate constants at various temperatures were determined. The results on the individual steps of the reaction were then combined and an overall kinetic correlation was developed for the complete set of reactions. The overall kinetic correlation can now be used to predict the product distlbuxtion for any reaction conditions. Some experiments were carried out under conditions where all of the steps occurred to significant amounts to check out the predictions of the correlation. The correlation was also used to show the variation of product distribution throughout the course of the overall reactiono Some calculations were also made to investigate the effect of various reaction, conditions, outside the experimental range of this work, on the product distributionso

Some effort was devoted to the consideration of possible mechanisms for the reactions, although this was not a primary objective of this work. Previous workers have shown that high temperature thermal decomposition of hydrocarbons occurs at least in part by free radical mechanisms. Therefore, a few experiments were carried out in which a free radical inhibitor was added in various amounts. These data together with the rest of the data were considered from the viewpoint of reaction mechanism and possible mechanisms are discussed in a speculative mannero Bo Review of Literature The pertinent literature can be divided into a number of sections, the first and largest of which is concerned with the kinetic data for the thermal decomposition of ethane at relatively low temperatures (5000C to 8000C). The next and much smaller section deals with the same reaction in the temperature range 800~C to 11000Co There has also been considerable work devoted to the determination of the mechanism of the thermal decomposition reaction of ethane from a free radical point of view. The last two sections are quite small and deal with the thermal reactions of ethylene and acetylene, mostly at lower temperatures (800~C) where polymerization predominates. These sections of the literature are reviewed briefly below and any literature data that can be compared directly with this work are discussed in greater detail in the section on experimental results. The kinetic data on the decomposition of ethane at lower temperatures (500~C to 800~C) are contained in the papers of Pease,(33) Frey and Smith,(20) Paul and Marek,(2) Sachsse(37) and Steacie and Shane.(43)

4 The work of these authors was reviewed by Steacie(44) and Brooks et al.(l0) The agreement between the various workers is good and the data are presented as first order rate constants for the disappearance of ethane. The reaction was found to be homogeneous and the main products were ethylene and hydrogen (no acetylene can be produced at these temperatures because of an equilibrium limitation) The kinetics of ethane decomposition in the temperature range 750~C to 1000~C have been investigated by Eastwood and Potas(l16) Hepp et alo(24) and Kinney and Crowley.(26) Schutt(38) reviewed these papers and compared the data with some commercial scale measurements. This higher temperature work agrees reasonably well with the lower temperature data but with much more scatter evident which is probably due to the uncertainty of the temperatureo The kinetic correlations were confined to first order rate constants for the disappearance of ethane. Some experiments were made by Tropsch and Egloff ( at temperatures up to 1400~C but only product distributions were obtainedo There is some literature concerning commercial and pilot plant processes for the high temperature pyrolysis of ethane and other saturated hydrocarbons to ethylene and acetylene. The significant authors in this field are Farnesworth et alo,(19) Bogart et alo,(7) Sittig,(40) Bixler and Coberly(6) Hasche(23) and Akin et alo(l) These papers do not contain any kinetic data but only process descriptions and product distributionso There is a considerable literature on the study of the mechanism of the thermal decomposition of ethaneo Rice and Dooley(36) detected free radicals by the lead mirror technique and Eltenton(l17l8) detected free radicals with a mass spectrometer0 Rice and Herzfeld(35) have suggested

,5a free radical mechanism and this work has been the basis for many further studies which are reviewed very well by Benson. (5) The literature on the mechanisms will be discussed in more detail in section VA of this work. It is pointed out that even after many years of study the free radical mechanism is still not well enough understood to make quantitative calculations of great accuracy. In a recent piece of work by Snow et al. (41) the most up to date free radical mechanisms and their appropriate rate constants were used to try and predict the existing literature experimental results on ethane pyrolysis to ethylene and hydrogen at low to moderate temperatures. In order to obtain agreement the radical mechanism rate constants had to be changed by a trial and error procedure. The kinetics of the thermal reactions of ethylene have mostly been studied at relatively low temperatures where polymerization predominates over decomposition. This work is contained in the papers of Burk et alo,(ll) Dahlgren and Douglas,(l1) and Molera and Stubbs.(31) They found the polymerization reaction to be homogeneous and second ordegr. Tropsch et al.(47) obtained some product distribution data at temperatures up to 1400~C. The thermal decomposition of acetylene at high temperatures is often an explosive reaction as shown in the early work of Bone and Coward. (910) Pease,(34) Zelinski(49) and Taylor and van Hook46) studied the kinetics of the polymerization of acetylene at moderate temperatures and found it to be homogeneous and second order. A small amount of kinetic data on the thermal decomposition of hydrocarbons has been reported rcently by workers using shock tubes as a means of obtaining high temperatures. Greene et al.(2l) and Miller(30)

-6have obtained some shock tube kinetic data on ethane and ethylene pyrolysis and Aten and Greene(4) on acetylene pyrolysis.

II. THEORY A. Introduction When ethane is subjected to thermal decomposition at elevated temperatures the major products are ethylene, acetylene, carbon and hydrogen. The work in the literature indicates that the reaction proceeds through a series of consecutive steps as outlined below. Ethane -* Ethylene -> Acetylene -> Carbon Hydrogen is formed at each step and there are also some small amounts of side reactions. In this kinetic study each step in the series was investigated separately as this approach will yield much more information than attempting to interpret data resulting from experiments in which all of the reactions were occurring. In fact, data resulting from all of the reactions occurring together would be almost impossible to interpret so as to obtain individual rate constants. A kinetic study was carried out on each of the steps and the conversions were kept low so that the primary reactions of any particular step could be studied. The product distributions were determined for each step at various amounts of conversion. These data were useful in determining which were the primary reactions and which were the secondary reactions. The determination of the order of a particular reaction and the method of obtaining the kinetic rate parameters from the data will be discussed later in this sectiono 7T

008. B. Reaction Equilibria C2H6 T C2H4 + H2 () C2H1. C2H2 + H2 (2) C2H2 - 2C + H2 (3) The equilibrium constants for the three reactions shown above are plotted against reciprocal temperatures in Figure 1. We can determine from this figure the temperatures at which the equilibrium yield of the products become appreciable. The equilibrium yield of ethylene becomes significant in reaction (1) at about 7850C (the equilibrium constant is 1 at this temperature)o In reaction (2) the equilibrium yield of acetylene reaches a significant value at about 11150C. The equilibrium position for reaction (5) is almost completely on the product side over the whole temperature range. The equilibrium yield of the products can be increased somewhat by a reduction in pressure (by inert dilution or by reducing the total pressure) since the forward reactions result in an increase in volume. Therefore, the products could be made at somewhat lower temperatures than previously indicated. However, generally speaking, it can be said that temperatures of 10000C and upwards will be necessary to produce high acetylene yields. Equilibrium considerations only place certain limits on the product distributions. The actual product distributions will, of course, depend upon the actual rates of the various reactions, the investigation of which is the purpose of this work. C. Temperature Distribution The thermodynamics of the system indicate that temperatures in the range 800~C to 1000C and higher are necessary to carry out these

' ( ( )' jaG oqea CSjus aGJ: )'*U1; uT s4tre'suno umnqtjqnTTba'* aJnSl-T >1o 1 01 x 01 6 8 L 9 G - O ~~~-6-~~r-^~ I ^01 3 5 I? ii O o -6

-10reactionso The rates of these reactions are quite high at these elevated temperatures so that the products will be mostly carbon and hydrogen unless the residence time is kept very small. (residence times are of the order of 1/100 of a second)o Obviously, a batch kinetic experiment is out of the question at this small a residence time, so steady state flow kinetic experiments were carried outo When a kinetic study is carried out it is extremely useful to have an isothermal experiment, ioe. temperature profile 1 in Figure 2, A set of isothermal experiments at different temperature levels allows relatively simple processing of the data to obtain the kinetic rate parameters [a log (rate constant) versus reciprocal temperature plot can be used]. The short residence time required for the reactions under study 1 2 Temp. Distance along Reactor Figure 20 Schematic Temperature Profiles.

allhere gives rise to a temperature profile like number 2 in Figure 2o The shape of curve 2 is due to the fact that the heat can only be transferred to the gas at a finite rate. The main point of this discussion is that a square profile like curve 1 cannot be obtained or even approached. This fact is true for all high temperature fast reactions. Many workers in the literature have been faced with this problem and the usual solution adopted is to estimate a mean temperature that is applied over an arbitrary fraction of the heated section of the apparatuso These approximations result in considerable scatter of the data especially since the shape of the temperature profile is dependent very closely on the flow rate (which determines the heat transfer rate) which is changed considerably from run to run during a kinetic studyo In this work no attempt was made to try and approach a square temperature profile (ioeo an isothermal experiment). The temperature profile was allowed to take whatever shape it wanted to and then the actual temperature distribution was measured carefullyo The use of this experimental data to determine rate constants is now quite difficult since the rate constants themselves are functions of temperatureo The equations and their method of solution to obtain these rate constants are developed in the next few sections. D. Determination of the Order of Reaction with an Arbitrary Temperature Distribution The data taken to study the order of reaction require a simple mathematical treatment to deal with the arbitrary temperature profileo In an isothermal kinetic study the order of reaction can be found from a set of data at constant temperature in which the rate is measured at

-12various concentrations of the reactant. The concentration of the reactant can be altered by changing the total pressure or by diluting with an inert gas. The amount of reaction is kept very low so that the concentration of the reactant is essentially unchanged and the rate measured can be considered as a differential rate. A plot of log (reactant concentration) versus log (rate) will result in a straight line the slope of which is the order of the reactiono The same technique can be used with an arbitrary temperature profile provided that the profile remains constant for the set of runs. This will be demonstrated by the following derivation, Suppose an irreversible reaction of the form A -> Products for which the rate equation in a flow reactor is - V R (C)C (1) where NA is the flow rate of reactant A in gm moles/sec., VR is the reactor volume in liters, n is the order of reaction, k is the rate constant in liter ni/gm, mole n-l sec., CA the concentration of reactant in gmo moles/litero The rate constant can be expressed by the Arrhenius equation as k A -JE/r (2) k =Aq, where A is the pre-exponential factor with the units of k, E is the activation energy in cal/gm. mole, T is the temperature in ~K and R is the gas constant in cal/gm. mole.

-13Substituting in Equation (1) -^ -= A - r (C)h (3) The concentration is related to the mole fraction by the following expression~ _ N,' MA P (4) C0 -NT RT RT where NT is the total flow rate in gm moles/seco, P is the pressure in atmospheres and MA is the mole fraction of Ao Also =VR (5) where a is the cross-sectional area of the reactor in cm2 and i is length along the reactor in cmo Substituting expression (4) and (5) in Equation (3) A - A - -E/RTr MA P (6) CLde LR J Assuming that the conversion is small so that MA stays essentially constant, that the pressure is constant and that there is a negligible radial temperature gradient, we can integrate Equation (6) along the length of the reactor to obtain the following expression: A, u. _. TCa) L / NA =. A/ -L -RT(e (7) where T( ) is the temperature distribution function with respect to reactor length, l NA is the small. decrease in moles/sec. of A, and L is the total reactor length in cmso

-14This reduces to L _E/R~ = Nc ] m 0 8@)mt vrtX(8) (8) The value of the expression inside the square brackets in Equation (8) will be a constant provided that the temperature distribution T(i) is constanto Taking logarithms of Equation (8), Equation (9) results. Lo6 ( N) = m Lo r(M ) +-L LOG(cCCOwTPT) (9) We can see that a plot of log (ANA) against log (MA) will result in a straight line the slope of which will be the order of reaction n, if the temperature distribution for the set of runs is kept constant. A similar derivation to this has been presented by Lee and Oliver. (28) Eo Derivation of the Kinetic Equations for an Arbitrary Temperature Distribution We will assume an irreversible homogeneous reaction of the type A -> B + C It also is assumed that plug flow is obtained and that there are negligible radial temperature variationso There is, of course, a large variation in longitudinal temperature distribution T(Y) the form of which is only known as a set of measured temperatures at known positionso The rate equation for this reaction is - aW =Aa (0o) A'q.

-15Let the feed rate of A be F gm. moles/sec, the fractional conversion of A be x, and the feed rate of inert gas ND gm. moles/sec. Then Wh _ V(\ w) (11) and N\ = -F X (12) Also - P F R RT P- FC\_') + ]R- (13) and dR = ad- 1 (14) Where the symbols are as defined in Section IID and the nomenclature. Substituting (12), (13), (14), (2) in (10) we obtain -E/ - - (15) Rearranging Equation (15) F _Ac______ = A (16) and integrating (16) along the length of the reactor from 0 to L, over the temperature distribution T(2), and from the inlet conversion xi to the outlet conversion xo we obtain _ L = T -..,. AyP) L- - /R'te) o. ~ ~~~~~~~~~~~T~e.)

-16The values of a,L are known dimensions of the reactoro The experimental data from a run will be values of F, ND, P, xo, xi and T(W)o The temperature distribution will only be known as a set of temperature measurements at known distances along the reactor axiso Fo Solution of the Equations for the Kinetic Parameters The rate constant k cannot be solved for directly from the data obtained in this work since it is temperature dependent and the data is obtained in a non-isothermal system. However, the more basic kinetic parameters, pre-exponential factor A and the activation energy E can be solved for. The mathematical problem is to solve Equation' (17) for A and E. The order of reaction n is determined experimentally from a separate set of data by the method developed in Section IID so that n will have a numerical value. Rearranging Equation (17) we obtain A = a P t x! i iL- X rL - (18) Jo TO"(Q Since the temperature distribution T(W) is only known graphically, the integral containing it will have to be evaluated numericallyo The numerical integrations are carried out using Simpson's ruleo The data obtained from one experiment are insufficient to solve Equation (18) since an infinite number of values of the two parameters A and E would satisfy the

-17equation. Unique values of A and E can only be obtained from Equation (18) by taking the data from two separate experiments (with different experimental conditions) and solving simultaneously. Let the data from the two different experiments be denoted by subscripts 1 and 2 respectively. Then substituting each of the sets of data into Equation (18) we obtain the following two equations a R. |S I/' l \ a+ Nfl/F Api/ -c L \-x D Tilie Jo TC - F' + A = a CcJx 2; L - - 1 / e-E/RT )z I (20) Jo T)2 We can now set about solving Equations (19) and (20) simultaneously to obtain values of E and A. The method of solution used is to assume values of E and calculate values of A from Equations (19) and (20) the integrals being evaluated numerically. The solution is obtained when for a chosen value of E the values of A computed from (19) and (20) are identical. These calculations represent considerable computational labor which would result in the method being prohibitively time consuming if carried out by hand. The calculations were however, readily handled with the aid of a

-18digital computer so they were programed for an IBM 704 digital computer. The flow diagram for the calculation upon which the program is based is contained in Appendix VII. The input for the computer program is the experimental data for a pair of runs and a set of values of E that brackets the solutiono The output from the computer is two sets of A values corresponding to the set of E values chosen. If these E and A values are plotted it would look like Figure 3 below: Data Set 2 Data Set 1 Log A E Figure 35 Solution for Kinetic Parameterso The intersection of the two lines represents the solution for E and A. It is found that the logarithm of A is practically a linear function of E so that linear interpolation from the four pointsbracketing the intersection is quite accurate.

-19 A sample calculation using actual experimental data is contained in Appendix IV. The technique developed here enables one to calculate E and A values from any pair of experiments. This method of calculation is quite sensitive to the accuracy of the data so that fairly small experimental errors product large variations in the E and A valueso The next section will develop a method of smoothing the data and will show why the method of calculation is so sensitive. Go Calculation of Rate Constants and Smoothing of the Data Now that we have computed the E and A values from the data from a pair of experiments, we can go back and calculate rate constants from the Arrhenius expression A e e (21) e The temperatures that are used in Equation (21) are the maximum temperatures of the temperature profiles for the pair of experiments, No error at all is introduced or assumption made here in selecting the maximum temperature to use in Equation (21) since the rate constant calculated is then plotted against this temperatureo If a number of experiments are carried out at different temperatures, we can obtain solutions for each pair of experiments. In this way we can build up the usual log (rate constant) versus reciprocal temperature plot (see Figure 4). It is important to note that this plot has been obtained without the requisite of an isothermal experimento The sensitivity of the calculation of E and A to experimental error is illustrated in Figure 4. The E values (the slopes) and the A values (the

-20intercepts) can be seen to differ quite markedly depending on which pair of points are selected for the solution. However, the dotted line is a Lnk Ln k A\ l/T Figure 4o Data Smoothing (Fictitious Points). fair representation of the data so that the in k versus l/T plot is seen to be a convenient way of smoothing the data.

.21IIIo APPARATUS AND EXPERIMENTAL TECHNIQUE Ao Description of Apparatus A diagram of the apparatus is shown in Figure 5o The gases used in this study were obtained in cylinders from the Matheson Company, Inc. The ethylene and acetylene were 99~5 mol % purity, the ethane 97 0 mol % purity and the nitrogen 99o996% purity (8 p pm oxygen)o More detailed analyses of the feed gases are contained in. Table IIo The gases were supplied from the cylinders at a steady pressure of 10 to 20 poSoio by Matheson No0 1 single stage regulators. The gas flowed throughout the system in 1/4 or 5/8 in. copper tubing with brass compression fittingso The flow rates of the various gases used were controlled manually with 1/8 in.. needle valves and were measured with various sizes of rotameters (Matheson Universal Flowmeters)o Th.e gases then passed into the ceramic reactor which was contained inside the furnace. Temperatures were measured with Chromel P-Alumel thermocouples and pressures were measured with mercury or water filled manometers. A sample of the gas stream was taken as it left the reactoro A vacuum pump was connected to the sample system to facilitate air removal from the sample bulbs. The gas then passed through a wet test meter which measured the volumetric rate and then out through a vent. The details of the furnace are shown in Figure 6. The alundum muffle was obtained from the Norton Company and was 1-1/2 in. bore with 1/4 i.n walls of type RA 139 materialo The platinum wire for the winding was obtained from the Baker Platinum. Division of Engelhard Industries, Inco and was 0.020 ino in diameter and 50 fto long. The furnace was

SCALE POTENTIOMETER MOVABLE- FURNACE THERMOCOUPLE ARIABLE VOLTAGE | 1 —-u- i | I=TRANSFORMER \ VOLTMETER ~ --- 220 V. SUPPLY /* C r~____ — SPL ^-THERMOMETER ROTAMETERSO M V 1/8"1 NEEDLE, TO VENT VALVES I r CYLINDER HYDROCARBONI BY-PASS - REGULATORS R~ * ^TOSA IR BLAST M t ( M COOLINGSAMPLE BULB WE SATURATOR METER NITROGEN -LMANOMETER VACUUM PUMP CYLINDER iHYDROCARBON CYLINDERS Figure 5. Diagram of Apparatus.

ALUNDUM MUFFLE WITH PLATINUM WINDING CEMENTED TO IT 3/64 SHEET STEEL 1 2.D., 2" O.D. 1/2" TRANSITE BOARDS BLOCKS POWER INPUT ---,,,ITERMINAL c9 SQUARE 3/64 SHEET CONTAINER< _ — LOOSE BUBBLE ALUMINA INSULATION \ \ ----------------- ~,."''__ _ _15" Figure 6. Furnace Details.

-24. designed to operate up to 1600~C and the watt density on the winding at 1.500 watts power input was 40 watts/sqocmo The winding was all in one piece and the turns were placed closer together at the ends to compensate for the large end heat losses. The actual spacings were (starting from one end) 1 ino at 8 turns/ino, 1 in, at 6 turns/ino, 4 ino at 4 turns/in,, 10 ino at 3 turns/in., 4 ino at 4 turns/in,, 1 ino at 6 turns/ino, 1 ino at 8 turns/in. Double lead in, wires for the power supply were used to prevent overheating in the passage through the insulation. The platinum winding was cemented to the muffle with Norton BA 1139 cement. Norton Bubble Type alumina was used as the high temperature insulation close to the muffle, and was poured in looselyo Johns Manville Superex (a silica type insulation) blocks were used for the lower temperature insulationo The whole furnace was contained in a light sheet steel rectangular box. Transite boards (1/2 ino thick) were bolted onto both ends of the sheet steel container to center and support the muffle and to contain the loose insulationo Three different reactors were used in the course of the experiments. The first of these reactors is illustrated in Figure 7 and was a 3/4- in. I.Do packed reactoro The other two reactors were concentric tube type reactors as shown in Figure 8. The packed reactor was the initial design and the purpose of the packing was to prevent radiation from the reactor wall affecting the center thermocouple reading. This ensures that the center thermocouple is reading the true gas temperature at that point. A second thermocouple sheath was placed between the reactor and the muffle However this design of reactor was found to be limited in its use since its pressure drop became too high as the

-25MULLITE CENTER THERMOCOUPLE SHEATH 1/8" I.D.,3/16" O.D. L I RUBBER - RING 1/4" BRASS UNION --- 3/8" BRASS TEE N /GAS - WATER COOLING COILS INLET _ /' RUBBER O-RING I \X 1#c 0 SQUARE BRASS ll^ES^ WFLANGES O FURNACE MUFFLE MULLITE REACTOR MULLITE WALL- / Ft 1 3/4" I.D. I" O.D. THERMOCOUPLE SHEATH oA T ALUMINA PACKING 1 14 MESH M W 1:^ -RUBBER O-RING I" VEECO BRASS SEAL SURFACE 2 3 U = 62 IN./IN. VOLUME GAS OUTLET Figure 7. Constructional Details of the Packed Reactor (No. 1).

-26MULLITE THERMOCOUPLE SHEATH NO:2 1/8"I.D., 3/16" O.D. NO:3 1/8" I.D., 7/32".D. RUBBER O-RING / —— RUBBER O-RING CONNECTOR 3/8 ALUMINUM TEE GAS INLET ---- --- (Nos. 3 N -— RUBBER O-RING I IULLITE REACTOR-B -I REA 1/4" 1.0., 5/16" 0.0. TOP OF.-IN FURNACE r. FURNACE MUFFLE SURFACE/VOLUME NO:2 64 INZ./IN NO:3 128 IN2 /IN3. BOTTOM OF REACTOR IDENTICAL TO TOP Figure 8. Constructional Details of the Annular Reactors (Nos. 2 and 3).

-27flow rate increased, so that the concentric tube type reactor was constructedo The great gas velocity through the narrow annulus in this reactor ensured that the center thermocouple read the true gas temperature (this is discussed further in Section IVA(iv))o The reactors and thermocouple sheaths were made from Vitreous Refractory Mullite (type MV 30) obtained from the McDanel Refractory Porcelain Companyo The seals were made on each end with rubber O-ringso The top seal on the packed reactor was water cooled and the bottom seals were cooled by an air blasto The cooling of the seals was to lessen the heat deterioration of the rubber O-rings. The thermocouples were made from 26 gauge Chromel P-Alumel wires and could be moved up and down inside the thermocouple sheath (see Figure 9)~ A scale was mounted vertically so that the position of the thermocouple in the sheath could be determined preciselyo B. Experimental Technique About 5 or 6 hours were required to heat the furnace up to operating temperature as heating rate was limited to 5 C per minute to prevent thermal shock damage to the alundum muffleo The feed rates of the various gases were observed with the rotameters and were controlled manually with the needle valveso The temperature was controlled manually with the variable voltage transformero A sample bulb was inserted in the exit line and the air pumped out of it with the vacuum pumpo When the system had reached steady state a gas sample was obtained by diverting the exit gas stream into the sample bulb by closing the bypass line. At that time a temperature profile was obtained as quickly as possible

-28THERMOCOUPLE WIRES 26 GA. CHROMEL- P ALUMEL MULLITE INSULATOR 1/16"O.D. 1/32"I.D. THERMOCOUPLE SHEATH BEAD 3I I- 16 DIA. Figure 9. Theocouple Details. Figure 9. Thermocouple Details.

-29along the length of the reactor~ The inlet and exit manometers were read and the volume of gas passing out through the wet test meter was recorded for a measured timeo The temperature and pressure in the wet test meter were noted and the atmospheric pressure was read and recorded. The experimental conditions were then changed to the next set of desired values and the system was allowed to come to steady state again (which would take from 1/2 hour to 1-1/2 hours). Air was passed through the hot reactor after each series of runs to burn off any carbon deposited during the reactiono Co Chemical Analysis The analyses of the gas samples were carried out on a mass spectrometer (Consolidated Engineering Corporation Type 21-103B)o The carbon analysis was obtained by material balanceo The mass spectrometer is a comparative instrument in that a quantitative analysis of a mixture can only be obtained by comparing the sample cracking pattern with those for the pure components in the mixture. The cracking pattern is the relative amounts of ion fragments at each mass to charge ratio resulting from electron bombardment of the sampleo If a pure sample of a compound in the mixture is not available, then a comparison can be made with published cracking patterns(3), although this is less accurate because patterns change slightly from machine to machineo In this work pure samples were available for most of the components (C2H6, C2H4, C2H2, CH4, H2, N2, C3H6, C3Hg, C6H6, 1-3 C4H6) of the mixtures so that reference was only made to the literature patterns in the case of some minor components (C5H4, C4H4, C4H2, C5Hg).

-30A difficulty arises when a component has a number of possible structures for the same chemical formula because the cracking patterns are quite similaro If considerable amounts of one of these components are present in a mixture, it can usually be identified but if it is only present in small quantities, identification of the structure is practically impossible. The difficulty is increased if the compound is also one for which a pure sample is not available. Most of the components of the mixtures analyzed in this work only have one possible structure The major product C4H6 has three possible structures 1-3, 1-2 butadiene and butadiyne but was present in large enough amounts to allow positive identification with the mass spectrometer as 1-3 butadiene. The other components with alternative structures were, C3H4E methylacetylene or allene, C4H4 vinylacetylene or butatriene, and C5H6 cyclopentadiene or a penten-yne. These possibilities are discussed further in the sections on product distributions. D. Experimental Program The overall reaction was broken down into three steps for the purposes of experimentation, namely the thermal decompositions of ethane, ethylene and acetyleneo Sets of runs were carried out for each of the compounds to determine the product distribution, the orders of the reactions and the kinetic rate constants. The product distributions were determined at various conversions with the temperature profile held constanto The purpose of these runs was to show the origin and order of appearance of the various products formed. The orders of reaction were studied by a series of runs in which

-31e concentration of the reacting component was varied by nitrogen dilution whilst keeping the temperature profile constanto The kinetic rate constants were investigated with a series of runs at spaced temperature intervals with the flow adjusted to give the desired conversiono In all of the runs the amount of reaction was kept reasonably low so that the initial reactions in each case would predominate over the secondary reactions of the products. The conversions were not usually much below 1.0% since the accuracy of the chemical analysis would be lessened at this conversione The conversions were as high as 60% - 70% in some of the higher temperature runs since the limitations of the equipment would not permit a high enough flow rate to reduce the residence time sufficiently. Some data were also obtained at high conversions using ethane feed and producing ethylene, acetylene and carbon as well as other side products. These runs were used to check out the overall correlation resulting from the combination of the data on the individual stepso Some data were also obtained on the effect of addition of propylene which acts as a free radical chain inhibitoro

IVo EXPERIMENTAL RESULTS AND DISCUSSION A Thermal Decomposition of Ethane (i) Product Distribution A set of runs (Nos 80, 81, 83, 84) was carried out at approximately the same temperature in which the product distributions were measured at various levels of conversiono The amount of conversion was varied by changing the flow rate which alters the residence time. The temperature need not be controlled or measured precisely for these runs since the product distribution is not affected very much by moderate variations in temperatureo Table II contains the raw data and Table VI the conversions and product distributions (expressed as moles formed per 100 moles of ethane reacted) computed from the raw dataa. The product distributions are plotted against the conversion in Figure 10 and this plot is studied to determine the primary stable molecular species that result from the thermal decompositiono The primary products are those products that appear in significant amounts at low conversions and the secondary products are those that appear only after considerable conversion has taken placeo It can be seen. from Figure 10 that ethylene and hydrogen together with lesser amounts of methane and a small quantity of propane are primary molecular products. The only secondary product'to occur is 1-3 butadiene (C4H6) which appears only i.n small. amounts at the highest conversion. The propylene shown in the analy'sis is not a reaction product but is an impurity present in the ethane feed. Previous workers(12 24)48) in the literature have found the same products as were detected in this worko -32

-33120 non: 816 ~l w 00 C 1 2H0 0 4 uX z 80 )~_ RUNS 80, 81,83,84 0 Tmax 816 IC 0 0 o CH4 Ir LL0 C3H8 _ 0 10 20 30 % CONVERSION Figure 10. Product Distribution in Ethane Decomposition.

-34(ii.) Order of Reaction This set of runs (Nos 49 - 53) was undertaken to determine the order of reaction for the formation of the primary molecular products with respect to the ethane concentrationo The experiments were conducted at constant temperature profile and the mole fraction of ethane was varied with nitrogen dilution. The total flow rate was kept constant for all of the runs so that the temperature profile would not be disturbed. The conversions were kept to a low value so that a differential rate measurement was obtained and also so that the mole fraction of ethane was not changed significantlyo The raw data are contained in, Table ILBand the calculated rates and mole fractions are to be found in Table VIIo The logarithm of the rate is plotted against the logarithm of the mole fraction (arithrmetic mean of inlet and outlet values) of ethane in Figure 11 the slope of which represents the order of reaction (the validity of this is shown in Section IID)o The disappearance of ethane and the formation of ethylene and methane are seen to be first order in ethane concentration This is in agreement with the previous work of Pease(33) and Frey and Smith(20) at 600~C and Calderbank and Hovnanian(12) at 800 C who showed that the disappearance of ethane was first ordero The last named authors also showed that the formation of methane was first order in etrhane concentrationo (iii) Rate Constants Now that we have established that the significant primary molecular products are ethylene and methane (together with the attendant hydrogen in quantities that fulfill the material, balance) and that the

0.5 —--— _____ ---- RUNS: 49-53 Tmax: 8740 C _ zId C2 H FORMATION-, SLOPE 1.00 w 0.2 z 0 I-Q~0 SLCH FORMATION SLOPE I JI LL. 0.1 0-j~~~ ________-C____ _ ______ _____ — C H DISAPPEARANCE SLOPE = 1.00 0.05 -- ^ ------- - - ---— ^ — - -- - - ------ -6 -5 -4 10 10 10 REACTION RATE,GM. MOLES /SEC. Figure 11. Ethane Decomposition Orders of Reaction.

-36reactions are first order in ethane concentration,'we can set about determining the rate constants. A number of experiments (Noso 18, 19, 21, 22, 41 - 46, 128, 129) were carried out each with a different temperature profile in order to obtain the rate constantso A typical temperature profile is shown in Figure 12o The conversions were kept fairly low so that the primary reactions would predominate and this was achieved by reducing the residence time (by increasing the flow rate) as the temperature was raisedo Nitrogen dilution was used to reduce the effect of reverse reaction (some calculations showing the reverse reaction to be negligible are in Appendix VI) and to lower the conversion, The raw data for these runs are presented in Table IV. The method of calculation of the rate constants from these data obtained with a nonuniform temperature distribution has been developed in Section II of this work. A sample calculation showing all of the numerical details is contained in Appendix IVo It is recalled that values of the activation energy and the preexponential. factor of the Arrheniu.s rate equation are computed from pairs of data points, These kinetic constants are then used with the maximum temperature for a run. to compute a value of the rate cons tant at that temperature. The assumptions used in the derivation were irreversible homogeneous reaction, negligible radial temperature gradient and plug flowo Tale eVII contains the first order rate constants for the disappearance of ethane and these are plotted versus reciprocal. temperature in Figure 153 The rate constants were determined from pairs of runs (indicated in Ta;b.LeTZI) which were adjacent in temperature. It is apparent that very many pairs of points could be selected from a set of runs (Noso

-371100 RUN NO. 45 1000 800 --- EQUIVALENT SQUARE 700 W 600 I-. 500, /5 GAS FLOW 400 300 200 100 4 8 12 16 20 24 28 DISTANCE FROM REACTOR INLET, IN, Figure 12. Typical Experimental Temperature Profile.

-383 10. _________ I. z 0 w RUNS REACTOR 1;:(0: — 18,19,21,22 #I 0 42-46,128,129 2 + 96-98 *3 3 -I 10.. 7.0 8.0 9.0 10.0 Xlo4 T ~K Figure 13. First Order Rate Constants for the Disappearance of Ethane.

-3918, 19, 21, 22) and the values of the rate constants comparedo The results are contained in Thble VII and are plotted in Figure 14 which shows that there is very close agreement between. the different solutionso Therefore, in the rest of this study only adjacent pairs of points are used to calculate the rate constantso The presentation of the rate data as first order rate constants for the disappearance of ethane is a simplification as we have seen that there are two parallel reactions forming ethylene and methane. The calculation method has to be modified a little to enable the rate constants for the two reactions to be evaluated. These modifications are explained in detail in Appendix V. The rate constants for the first order formation of ethylene and methane are contained inTableVIh and are plotted against reciprocal temperature in Figure 15o The majority of the literature data is presented as first order rate constants for the disappearance of ethaneo The results of a number of workers (including some shock tube work) are plotted in Figure 16 and are found to compare well with the results of this worko The data of this work show less scatter than the literature data especially at higher temperatures and this is attributed to the careful measurement of the temperature profile and its proper consideration in the calculationso A definite downward curvature of the rate constants plot is exhibited by the results of this worko The literature data do not disagree with this curvature although the curvature is not apparent from the literature data alone because of the large amount of scattero The curvature apparently indicates a decreasing activation energy but this is found not to be the explanationo. In a later section of this work the curvature is shown to

-4o10.0 6.0 2.0 -.I I \-1 0 w0.5 C) "* RUNS 18, 19,21,22 \l 0.05 l --- 0 5 10.0 z w0.5 0.05 9.0 I la 9.5 10.0 x- X 10 T~ K Figure 14. Solutions for All Possible Pairs from Four Points (First Order Rate Contents for Disappearance of Ethane).

-413 I0 JoC) L, 2 2 O 2( 10 zw i 0 w - I10 10 E RUNS: 18.19,21.22 42-46 128,129 -- - -- - 96-98 --------— N \ RATE BASED ON MOLS C - 10 70 8.0 9.0 10.0 Figure 15. First Order Rate Constants for Fomnation of Ethylene and Methane from Ethane.

-420 100 -io~~~ 0 -: 0 O - DAO F TH WOR 0_ 0 _ _ KN A ( _ _ 10 ~~~~~~~~io~~~~~~~~~~~~~~~~~~~~~~~x \0 MILLER (30) -2 + HAGUE AND WHEELER (22) X 10 HEPP ET AL (24) o 0 STOARCH AND GOLDEN (43) X D I EASTWOOD AND POTAS (16) 0 SCHUTT (38) D KINNEY AND CROWLEY(26) X MAREK AND Mc CLUER (29) A KUCHLER AND THIELE (27) -3 M STEACIE AND SHANE (43) 10 - - * SACHSSE (37)'3 1-5X (0 3 4 5 6 7 8 9 10 11 12 — 0( T K Figure 16. First Order Rate Constants for Ethane Disappearance, a Comparison with Literature Data.

-43be a result of inhibition of the rate by secondary products of the reaction which are formed at higher temperatureso One previous piece of work by Calderbank and Hovnanain(l2) contained some rate constant values for the first order formation of methane which agree well with this worko In the temperature range 730~C to 900~C for this work the first order rate constants for ethane disappearance (and ethylene formation) show an activation energy of 82~4 k cals/gm. moleo Above 900~C the reaction becomes inhibited by reaction products (the empirical representation of the data in this region is developed further in Section VIA). This value of the activation energy is a little higher than that found by the previous workers whose values lie between 66 and 75 k c'als/gm. mole. Over the whole temperature range studied in this work (730~C to 1160~C) the first order rate constants for methane formation show an activation energy of 6609 k cals/gmo mole which agrees quite well with the value of 64 4 k cals/gm, mole reported by Calderbank. (12) Various workers have shown that the reaction is homogeneous,33,43) however, this was checked with a few runs carried out in a reactor (No. 3) which had a different surface to volume ratio. These points are plotted on Figure 13 and are seen to be coincident with the data in Noo 2 reactor, which indicates that the reaction is homogeneouso (iv) Temperature Distribution The problems that will be considered here are firstly, do the thermocouples in the various reactors read the true gas temperature and secondly, can the radial temperature gradients be considered negligible. Figure 17 contains radial sections through the various reactors used

-44REACTOR NO: I CENTER THERMOCOUPLE SHEATH WALL THERMOCOUPLE SHEATH I/8"I.D., 3/16" O.D. REACTOR WALL / I DIA. HEATED MUFFLE REACTOR NO: 2AND3 /"2 DIA. REACTOR WALL --- THERMOCOUPLE 1/4" I.D., 5/16"O.D. SHEATH NO: 2 I/8"I.D.,3/16"O.D. NO: 3 1/8" I.D., 7/32"0.D. Figure 17. Radial Section Through Reactors Showing Location of Thermocouples.

_45showing the location of the thermocoupleso The first reactor used in the experimental work was No. 1 and it was designed to contain refractory packing which would shield the center thermocouple from radiation from the hot reactor wallo The gas velocity through the packing in reactor No. 1 is quite high so that the packing and the gas will be close to thermal equilibriumo The center thermocouple in reactor No 1 is considered then to measure the true gas temperature at the center of the reactoro A second thermocouple was placed alongside the reactor wall in the annulus between the reactor and the furnace muffle The purpose of this thermocouple was to measure the wall temperature of the reactor, however, it can be seen that this thermocouple will read a temperature greater than the desired wall temperature because of the radiation from the hot furnace muffle (there is no radiation shield for this thermocouple or gas flow to reduce the radiation error). Runs 18, 19, 21 and 22 were carried out in reactor No 1. and both the center temperature (Tc) and the wall temperature (Iw) profiles were measured (see Table IV)o Then the rate constants were calculated using in one case the center temperature and in the other case the radial distributiono A linear radial distribution is assumed between Tw and Tc and some calculations contained in Appendix III show that the correct mean temperature to use in the calculations is Tc + 075 (Tw - Tc). The values of the rate constants for the two cases are contained in Table VLI ard are plotted versus reciprocal temperature in Figure 18. It is seen that the difference in the rate constants is quite small and it is pointed out that this difference is greater than the maximum possible error since it is known that the wall thermocouple will be reading greater than the true wall

-4610.0 5.0 \\ - BASED ON Tc 2.0 - BASED ON Tc - 3/4 (Tw- Tc) o I \I \ c- RUNS 18,19,21,22 REACTOR #l 1 Tc CENTER TEMP. Tw WALL TEMR 0.2 01 0.05 8.0 9.0 10.0 x 104 T~K Figure L8. Effect of Radial Temperature Distribution

-47temperature, and the assumption of linear temperature distribution is an unfavorable oneo It is concluded then that the radial temperature distribution is negligible in reactor Noo 1 so that the center temperature can be used in the calculations of rate constantso Reactor NoO 1 was found to be unsuitable for the higher temperature experiments since a pressure drop limitation would not allow high enough flow rates to give the low residence times required~ Therefore, reactors Noo 2 and No. 3 were built and used for the higher temperature experiments. These two reactors are similar in design and are just simple annular spaces between two tubes with the center tube acting as the single thermocouple sheatho These annular reactors are of much smaller diameter than the packed reactor so that they have a higher gas velocity and a lower mass flow rateo These last two factors both will reduce the radial temperature gradients below that experienced in the packed reactor so that we can safely say that the radial temperature gradient is negligible with respect to its effect on the computed values of the rate constantso These annular reactors do not have any radiation shielding between the reactor wall and the center thermocouple sheath so that the question arises of whether the center thermocouple is reading the true gas temperature. The gas velocity through the annulus is very high ( of the order hundreds of feet per second) and this will reduce the radiation error. Some experiments were carried out in the annular reactor within the same temperature range as the previous experiments in the packed reactor and the values of the rate constants were coincident (see Figure 13). Since we have already established that in the packed

reactor the true gas temperature is measured, we can conclude that the center thermocouple measures the true gas temperature in the annular reactors o B, Thermal Decomposition of Ethylene (i) Product Distribution The product distribution runs (Noso 86 - 89, 126) were carried out in a similar fashion to those for ethane. The raw data is contained in Table II, the results in Table VI and the product distribution is plotted against conversion in Figure 19o Figure 19 shows the major pri~ mary molecular products to be acetylene, hydrogen and 1-3 butadiene and the major secondary products to be C4H4, benzene, methare and carbono Propane, propylene, ethane, C4H2, C3H4 and C 5H6 are also formed in small amounts. The C4H4 is thought to be vinylacetylene as a structure containing three double bonds seems unlikely, and the C4H2 can only be diacetyleneo The C^H4 could be either propadiene or methyl acetylene and the structure of the C:H6 is not known Previous workers have studied the. product distribu.tion in. this reaction although, mostly at lower temperatures where polymerization predominateso Burk et alo(11) found the primary products to be butene (C4H8) acetylene and hydrogen at 625 Co Dahlgren and Douglas(13) found propylene, butene, butadiene and ethane as primary products at 4800C to 580~Co The results of this work agree with the literature results except for the product butene which was not detected atll. lin this work. In the kinetic study carried out in this work the primary products are considered to be acetylene, hydrogen and 1-3 butadieneo

-49100 RUNS 86-89,126 90 -Tmox 996 C 80 70 60 w 50 Ia: w 40 ----- W 40 ~0 < 10 0 0+ 10 I I' 20 7- 0 0 0C I6 _____ _-__ \<~t ~ i I dH < 2 C5H6< 2 4 ____6 __ 2 0 0 5 10 15 20 25 30 35 40 45 50 % CONVERSION Figure 19. Product Distribution in Ethylene Decomposition.

-50(ii.) Order of Reaction The orders of reaction for the formation of the primary products from ethylene deco:rmpos:lltion were determined i.n the same way as for ethane. The raw data (Runs 54, 58,) are found in Table III, the computed results in Table VII and the rat es are plotted versus mole fraction in Figure 20. The polymerizati.on products included C4H6, C4H4, C5H6 and C6H6o The acetylene formation is seen to be first order and the polymerization rate second order -with respect to ethylene concentration The literature data are mostly at lower temperatures where polymerization predominates and a homogeneous second order -reaction is reported. (13 3^1) (iii) Rate Constants The raw data are contained in Table IV (Runs 24, 67 68, 47, 48, 99 - 101) and the results in Table'VTUL Figure 21 shows a plot of the first order rate constants for ethylene disappearance, however, this is a simplification of the reac-ti-on since there are actually two reactions occurring in parallel. Some data were obtained on reactor No. 3 which. has a different surface to volume ratio and these also are plotted on Figure 21.o The data from the two reactors are coincident so that this agrees with already reported factil3) that the reaction is homogeneous. Using a modification (see Appendix V) to the calculation method developed in Section II, the rate constants for the first order decomposition and the second order polymerization are calculated, listed in Table VIII, and plotted versus reciprocal temperature in Figures 22 and 235 There is very little literature ki.:ne.tic data in the temperature range of th.is work

0.5__VRUNS: 54-58 __ Tmax: 1082 0C _________ ___ H FORMATION _ w SLOPE =1.07 w cr' 02 _ IZ POLYMER FORMATION (C BASIS)- r 2 4 SLOPE SLOPE 1.78 0 0~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~00 0.05 _ u - ~ ~ ~~~~~ O. ----- -------— 0-0-0Z / I A P E R N E' u -------- - _ —_ _ —--------— ^ -- -^ -- - ^ - - -- 2 4 SLOPE t 1.38 - 005 4 10 10 10 REACTION RATE, GM. MOLES/SEC Figure 20. Ethylene Decomposition Order of Reaction.

-52RATE BASED ON MOLS C2 100 K\| E 64.5 KCAL/GM. MOL A 4.84 x 10 SEC C,) z 0 w ________ ________ RUNS REACTOR 0 24,1 0 47,48,67,68 #2 + 99-101 #3 7.0 8.0 9.0 I 4 x I10 T0 K Figure 21. Ethylene Disappearance First Order Rate Constants.

-53100 RATE BASED ON MOLS C2 E = 76.2 KCAL/GM.MOL 13 -I A= 1.76 x 10 SEC 10 0~\ _.... z z \ 0 \ w \ 0.I \ 7.0 8.0 9.0 I X 10 T~ K Figure 22. Acetylene Formation from Ethylene First Order Rate Constants.

-545 10. 1 RATE BASED ON MOLS C2 E = 60.3 KCAL/GM.MOL A= 2.63 x10 LITER/GM. MOL-SE ~0 ~~~~~\..J 0 J r — IJ \ 0 I 4 T x 104 T~K Figure 23. Polymerization of Ethylene Second Order Rate Constants.

-55with which to compare the data. Miller(30) and Greene(21) have some shock tube data which is based on a first order decomposition and their data are compared with this work in Figure 24 and the shock tube data is seen to have a lower activation energy although the rate constant values agree around 1160~Co In the temperature range 875~C to 1200~C the activation energy for first order acetylene formation from ethylene is found to be 76.2 kcal/gm. mole and for second order polymerization of ethylene 60o3 kcal/gm. moleo Molera and Stubbs(3l) around 700~C reported 75 kcal/gm. mole for the first order decomposition and 35 kcal/gmo mole for the polymerizationo Other work is reported in Steacie(44) at lower temperatures (600oo~C - 700~C) and the activation energy values for second order polymerization are around 40 kcal/gmo moleo C. Thermal Decomposition of Acetylene (i) Product Distribution The raw data is contained in Table II (Runs 115, 78, 127) the results in Table VI and the product distribution is plotted versus conversion in Figure 25~ The primary molecular products are seen to be carbon, hydrogen and C4H4 together with benzene as the only secondary product of any magnitudeo Small amounts of methane ethylene, diacetylene (C4H2) and C3H4 were also detected. The C4H4 is thought to be vinylacetylene and the C3H4 methylacetylene since the starting material was acetyleneo

-563 10 1 E =64.5 KCAL/GM MOL ~o~~~ O\ — bJ 2 z * GREENE ET AL (21) \ ~ 10 () ( SHOCK TUBE DATA 0 GREENE ET AL (21) COMPUTED BY MILLER (30) I0 O MILLER (30), E =21 KCAL/GMMOL \ 0 I0 4.0 5.0 6.0 7.0 8.0 9.0 I 104 T OK Figure 24. Ethylene Disappearance First Order Rate Constants, a Comparison with Literature Data.

MOLS OF PRODUCT PER 100 MOLS OF ACETYLENE REACTED o r\o ( A.b (A o -4 OD O OO~~~~~~ N~ (~ 01 o)O 0 0 N 00 ~~ ~~ ~~ ~~ ~~0 0 0 0 0 0 0 0 0 0 0 0 01 ~'x 3 C x (1) ~_ -— ~ I......T^ > ~+ 3 j 0+ _ _. o /I \ tI Id 0 ^l fTh ~~ ^ ~ AOo / /\ BOPi 4 N ~ ~ N _ ___________ - +_______________ — 0 OC+ 0 m) o.)P-. O +H0 H*.) Z Oc+ 0~~~~~~~~~~~ m - 0 _ 0 0~~~~~~~~~~~~C U)+. /o C) ) A A A 01

-58(ii) Order of Reaction These data (Runs 1,12 -115) are contained in, Table V, the results in. Table VII and the rates of reaction are plotted against acetylene concentration in Figure 260 The amount of carbon formed was calculated by material balance. The carbon formation is found to be first order and the polymerization second order (actually, slope is 1o5 but second order is assumed) with respect to the acetylene Considerable nitrogen dilution, was used to prevent detonation of the acetyleneo The polymerization products include C4H4 6 C6H6, and CLH20 Steacie(44) reported on the small amount of literature data on acetylene decomposition and the order is usually found to be two9 although the data, reported is at much lower temperatures than this work so that polymerization predominate S (iii) Rate Constants The raw data (Runs 110 - 112,.116 117 and 1.02 - 104) are contained in Table IV, the results in Tae\nEII, and the first order rate constants for the disappearance of acetylene are plotted in Figure 270 These data are seen. to scatter a little more and this is probably due to the fact that the carbon formed has to be determined indirectly by material balance. Data from reactors 2 and 3 (different s/v ratios) are shown in Figure 27 to be coincident (within the experimental error) wnhich shows the reaction to be homogeneous as already reported by Zelinski^(49) The rate constants for the formsration of carbon and polymer

RUNS: 112 -115 Tmax: 10430 C 0.2 POLYMER FORMATION (C BASIS) [u SLOPE = I.52 tX S L 1.52 S CARBON FORMATION I->^~ |~SSLOPE 1.00 / O0 w 0.1 2, 0 _________ R RATE____ ___ H G. - DISAPPEARANCE Figure 26. AcetySLOPE = 1.40 IA W. 0.05 -J 0 0.02,65 i-4 REACTION RATE, GM. MOLES / SEC Figure 26. Acetylene Decomposition Orders of Reaction.

-60RATE BASED ON MOLS C2 — 0 —-E-E= 52.0 KCAL/GM. MOL A= 9.30 x 10 SECi 0 w 10 U) RUNS REACTOR 0 110,-112,116,117 "2 + 102-104 - 3 7.0 8.0 9.0 x104 T~K Figure 27. Acetylene Disappearance First Order Rate Constants.

-61are contained in TabLe VII and are plotted in Figures 28 and 29~ In the temperature range 9000C to 11500C the activation energy for first order formation of carbon from acetylene is 61l9 kcal/gmo mol and for second order polymerization of acetylene is 4406 kcal/gmo mol The literature rate data(44) on the decomposition of acetylene are rather sparse, are imostly at lower temperatures and are correlated with a second order model for which activation energies between 50 and 4t0 kcal/gm. mol are reported. Aten and Greene(4) for their shock tube work reported a second order decomposition with an activation energy of 29 kcal/gm. molo Do Inhibition of Thermal Decomposition (i) Propylene Inhibition Previous workers(36,17y18) have shown that the thermal decomposition of ethane occurs at least, in part by a free radical process. Propylene and nitric oxide are compounds that react readily with free radical species(44) and have been used in various studies(4p225) to study the reaction mechanism. A set of experiments (Runs 70 - 74) was carried out at a constant temperature profile with varying additions of propyleneo The conversions were kept low so that a differential rate could be measured. The raw data are contained in Table V and the rates of formation of the products calculated from the raw data are presented in Table IX The rates are plotted against the fracti.on of propylene present in the feed in Figure 30. It is seen that the rate of ethylene formation is markedly reduced by small quantities of propylene which is indicative of a chain reactiono It is noted that the methane and propane rates are unaffected

-6250 40 30 RATE BASED ON MOLS C2 ~-~~20 ____\ _ E =61.9 K CALS/GM.MOL 10 -- A =9.7x10 SEC 9 8 7 \ 8 5 2 4 z 8 H ______II__l.7 3 z 0'- 2\ 1.0 9 --- RUNS: 110-112,116,117 ----.8.7.6.5.4.3.2 6.5 7.0 7.5 8.0 8.5 9.0 TOK 10 Figure 28. Carbon Formation from Acetylene First Order Rate Constants.

-63RATE BASED ON MOLS C2 10E =44.6 KCAL/GM MOL ______ A =3.19 x10 LITER/GMMOL-SEC 0 (I, (O 4 ol I \ 0 z \ 0 ~ O\ ~i i RUNS: 110-112,116,117 6.5 7.0 7.5 8.0 8.5 9.0 )I x104 Figure 29. Acetylene Polymerization Second Order Rate Constants. Rate Constants.

6 s0 5 ] ___________________RUNS: 70-74 Tmox 866 OC Cn 0 LJ (V) 4 4 C,) Ir -J 0 Z3 3 - C —---— HC4 FORMA-TIONW CH4 FORMATION z 2 0 0 2 4 6 8 10 12 14 16 18 20 22 MOLS C3H6 2 x 10 MOLS C. H Figure 30. Inhibition of Ethane Decomposition with Propylene.

-65by the propylene additiono A possible explanation can be advanced now as to why the rate constants for ethylene formation (Figure 15) and ethane disappearance (Figure 13) fall off at high temperatures and is that the reaction products at higher temperatures inhibit the reaction rate, The product distribution data for ethylene shows that a small amount of propylene is formed as the ethylene decomposes so that this will cause some inhibition of the rateo The reaction products from ethane obtained at the higher temperatures of this work were then tested to determine their inhibition effect on the rate (these experiments are described in Section IVD(ii))o The effect of propylene addition was also investigated for the thermal decomposition of ethylene, the results of which are contained in Tables V and IX (Runs 90 - 94). The rates of reaction are plotted against propylene fraction in Figure 31o There does not seem to be much inhibition effect but the data is somewhat inconclusive due to the fact that the propylene itself decomposes at the temperature of these experiments (see Run No 95). (ii) Inhibition of Ethane Decomposition by Reaction Products The various reaction products were added in turn to ethane to see if they could inhibit the reaction rate The data are contained in Tables V and IX (Runs 119 - 125), and they show that butadiene (C4H6) is quite a strong inhibitor whereas ethylene, acetylene, hydrogen do not inhibit the reaction at allo The effect of butadiene as an inhibitor is shown in Figure 32 and is noted that the methane and propane rates are not affected as was the case with propylene additiono It is concluded

12 RUNS: 90 - 94 10 I"0 1 Tmax 1093 ~C x o 0 w V0 z n 8C -------------— C 2 H F O R M A T IO N - LFJ _1 2 0 0 0 2 4 6 8 10 12 14 16 18 2FORMATION 0 2 4 6 8 I0 12 14 16 18 20 22 MOLS C3H6 x02 MOLS C2H4 Figure 31. Inhibition of Ethylene Decomposition with Propylene.

RUNS: 119-121,Do 4 Tmax: 880 ~C I A. +I A C FORMATIONHCH4 FORMATION 0 4 8 12 16 20 24 28 32 0: /- CH4 FORMATION CsH, FORMATION 0 4 8 12 16 20 24 28 32 MOLS C4H6 2 MOLS C2H6 Figure 32. Inhibition of Ethane Decomposition with Butadiene.

-68then that the curvatures of the plot's in Figures 13 and 15 are due to inhibition by the reaction products butadiene and propyleneo These inhbitiooni experiments are discussed further in the next section on reaction mechanism.

V. REACTION MECHANISMS Ao Review of Literature on Reaction Mechanisms There has been a tremendous amount of work carried out to determine the reaction mechanism of the thermal decomposition of saturated hydrocarbons, especially ethane, and this literature has been very well reviewed in four recent books by Steacie, 1954(44) Brooks et al., 1955}0) Semenov, 1959,(39) and Benson, 1960j(5) and the following discussion is based upon these works. The following brief review is almost exclusively limited to the thermal decomposition of ethane. A number of early workers were able to show the existence of free radicals in the decomposition of ethaneo Rice and Herzfeld(35) proposed a free radical chain mechanism in 1934 and this type of mechanism has been widely accepted and used as a basis for a large amount of later work. A simplified form of the Rice-Herzfeld mechanism for ethane decomposition will be presented and discussed now. The ethane molecule splits at the weakest link (the carbon bond) to form two CH3 readicals. These radicals then attack the ethane to form the C2H5 radicals. A free radical chain is set up which produces the bulk of the products and this chain reaction is terminated by some reaction which removes radicals. The following set of equations gives a qualitative picture of the mechanisms o C2H6 - 2 CH3 Radical initiation CH3 + C2H6 - C2H5 + CH4 -69

-70C2H5 - C2H4 + H Chain reaction C2H6 + H C2H5 + H2 CH5 + H --- C2H6 Chain term.ination CH5 + C2H5 -- CH8 Using the steady state assumption for the free radicals, the rate of disappearance of ethane can be solved for in terms of the rate constants for the individual. reactions and the concentration of ethane and is found to be approximately first order in ethane, which agrees with experiment. The activation energy for the overall reaction is shown to be less than. the value for the primary split into radicals wrhich is al.so in agreement with experimento The difference, however. is fairly small since the overall activation energy is dominated by the primiary radical. split because the reactions between a radical and a molecule have a very small activation energy and radical radical reactions have almost zero activation energy. The rate of reaction, hoowever, can be quite high if the chain length is long. (The chain length is defined as the rate of chain reaction divided by the rate of radical initiation) The very powerful inhibition. effects exhibited by nitric oxide and propylene are explained by removal of the chain. carryi.ng radicals whi ch reduces the chai.n length (i.e. the reaction rate)~ The actual mechanism of radical removal. by these two inhibitors is not very well understood. Addition of large amounts of inhibitor does not reduce the rate to zero and there has been much controversy over nwhether the residual mechanism is free radical or molecular in natureo Some recent work u.sl.ing isotopic mixing has established that the residual reactl.on is free radical i.n

-71natureo This type of mechanism can explain in a qualitative manner all of the observed experimental facts but the prediction of quantitative rates of reaction from data on the individual steps of the mechanism is not yet possible. A recent piece of work by Snow et alo (41) tried to predict the literature experimental data using a more complex form of the Rice Herzfeld mechanisms and the latest experimental data on the individual mechanism steps. Quantitative agreement could only be obtained by empirically changing some of the rate constants for the individual radical steps. In spite of the disappointing quantitative results from the Rice-Herzfeld type mechanisms, there is no doubt that these mechanisms are correct in principle and that the discrepancies probably arise because of a lack of good data on the individual radical. reactions and oversimplification of the reaction mechanism schemeo The nature of reaction mechanisms in the decomposition of ethylene and acetylene have not been studied very well and the little work that has been done is reviewed by Steacieo(44) The mechanisms proposed are quite speculative and there is considerable doubt as to whether the mechanism is free radical or molecular in natureo Bo Ethane Decomposition Reaction Mechanism The purpose of this discussion is to qualitatively show that the experimental facts in this work are consistent with the current free radical mechanisms proposed in the literature rather than to try and deduce any new facts concerning the mechanism on the basis of these experiments. This approach is taken here since it is not the objective

_72. of this study to investigate mechanisms nor is the experimental technique suitable to do soo Nevertheless, this discussion proves illuminating especially with regard to the inhibition experiments. The primary stable molecular products are found to be ethylene, hydrogen, methane and propane. The ethylene and hydrogen are, of course, the obvious products, but the methane and propane are not. A Rice-Herzfeld type mechanism, as outlined in the previous section, shows that methane can be formed by the attack c a methyl radical on an ethane molecule, and that propane can result from a chain termination reaction between two radicals. The activation energy for the decomposition 70 kcal/gm. mole; this is the true value discussed later an in this section and not the apparent value that is used in t:he kinetic correlation) is seen to be fairly close to but less than the bond dissociation energy (see Table X for approximate values) for the splitting of ethane into two methyl radical.s This fact is also consistent with the predictions of the Rice-Herzfeld mechanismso The Rice-Herzfeld mechanisms can be shown to lead to first order kinetics (an experimental fact) by making certain assumptions about the radical termination reactionso The experiments carried out with propylene inhibition are significant as the marked reduction in reaction rate by small additions of propylene is indicative of a chain reactiono The rates of formation of the products are plotted against propylene fraction in the feed in Figure 30. It is noted that the ethylene rate is very rapidly reduced whereas the methane rate is unaffectedo The ethylene rate is seen to approach a limiting rate as the propylene concentration increases rhich i.s almost the same as the constant methane rate. The propylene inhibition can be

-73 explained if the propylene reacts with the chain carrying radicals (H and C2H5)o The methane rate remains unaffected so that this means that the radical initiation steps (which produce the methane) are unaffected by the propyleneo The residual rate at full inhibition is then probably due to the radical initiation reactions. The chain length of the reaction is defined as the rate due to chain reaction divided by the rate of radical initiation and we can see that this is equivalent to the rate of ethylene formation divided by the rate of methane formationo The chain length is then computed on this basis and plotted against the propylene fraction in the ethane feed (Figure 33)~ The chain length is seen to vary from 24 down to 1.5 as the propylene amount is increased with the most rapid decrease occurring at low fractions of propyleneo The upper limit of 24 is due to the normal. chain termination reactions of the uninhibited reactiono The inhibition of ethane decomposition by nitric oxide has been investigated by Staveley(42) and Hinshelwood and Hobbs(25) and they found chain lengths ranging from 4 to 21o Dinstes et al. (1) showed that propylene inhibited ethane decomposition but no chain lengths were reported. A weakness of the Rice-Herzfeld mechanisms is that they predict a very much greater chain length than is found experimentally o The curvature of the rate constants plot for ethane disappearance and ethylene formation can. now be explained on the basis of varying chain lengtho The chain length is simply represented by the ratio of the ethylene formation rate constant to the methane formation rate constant and the values of the chain length are shown in. Figure 340 The chain lengths are never as high as 24 since there is a small impurity of

24 20 RUNS: 70-74,82,85 Tmax 830 OC -870 ~C (.9 16 z z <[ 12 I 0 LIMIT —........ 0 2 4 6 8 10 12 14 16 18 20 22 MOLS CH6 2 MOLS C2H6 Figure 33. Chain Lengths for Ethane Decomposition with Propylene Inhibition.

FIRST ORDER RATE CONSTANT, I/SEC 0I^) OL Co 0__ow ~~~~~~~~~~~~~o~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ r>/ $ -.- --- -r —. o'i"D.. — / x —~0a f o ~ ^ y O1 3 I i el' 0j t-* y > o~~~~~~~~~~~~;>^^ -- — 7 ^ - -R S - ^ — -- - - -- - ---- - - - m ----— ( —- -. -- --- -- --- -- ---

-76propylene in the ethane feed used which limits the maximum possible chain length to about 15 which.s the value at the lower temperatures in Figure 34. As the temperature is raised the chain length increases at first and then drops rapidly. The rapid drop is explained by the fact that at these higher temperatures considerable amounts of the secondary product butadiene are formed which was found to be an inhibitor (see Section IVD). The small increase in chain length at lower temperatures cannot be explained very well except for the suggestion that the amount of inhibition resulting from the propylene impurity in the feed may change with temperature. The curvature of the rate constant plot for ethylene formation has been explained on the basis of varying chain length rather than saying that the activation energy is decreasing. It will be of interest to determine whether the activation energy of the inhibited reaction does in fact remain almost the same, o hich it will if the effect of the inhibitor is on the chain length rather than on the primary split (the dominating reaction with regard to activation energy) into methyl radicals o Some experiments (Runs 82 and 85) were carried out with quite large additions of propylene inhibitor to determine the activation energyo The results are compared with the essentially uninhibited reaction results in Figure 55 and it is seen that the activation energies are about the same. This substantiates the conclusion that inhibition of ethane decomposition by reaction products is due to shortening of the chain length rather than decrease in the activation energy. Previous workers who studied nitric oxide inhibition found the activation energy for the inhibited reaction to be the same (Kuchler and Theile(27)) or a little higher (Steacie and Shanev *4^) than for the uninhibited rea.ction.o

-7730.0. 200 -- RUNS: 82,85 10.0 H. 5.0 DATA FROM h- ~'> -FIG. 13 0 w 2.0 C3 H6,C2 H6 =a 0.072 0 H) _RE =73 K CAL/GM. MOL / Q5 0.2 8.0 9.0 10.0 x 104 TOK Figure 35. Ethane Decomposition Activation Energy for Reaction Inhibited with Propylene.

-78C, Ethylene and Acetylene Decomposition Reaction Mechanisms The Rice-Herzfeld mechanisms only apply to the decomposition of saturated hydrocarbons so are of no use for ethylene and acetylene decomposition. There are no firm ideas in the literature about the mechanism of decompositiz.on of ethylene and acetylene and the little work that does exist is discussed by Steacie (44) who only concludes that the reactions contain in part at least some free radical processeso A consideration of the bond energies in ethylene and acetylene (see Table X) shows that the carbon hydrogen bonds are weaker than the double and triple carbon carbon bonds so that the inference is made that the decomposition occurs by hydrogen removalo The energy requirements for hydrogen removal are quite high (around 100 kcal/gmo mole) and th.i.s would indicate lower reaction rates for acetylene and ethylene decomposition than for ethane decomposition as is found experimentallyo The act ivation energies for acetylene and ethylene dehydrogenation are found expe:y rimentally to be 76 and 62 kcal./gmo mole respectively which. are inconsistent with the high bond energieso Some experiments were carri.ed out with propylene addition in ethylene decomposition. The results are plotted in Figure 31 but are rather inconclusive because it is:found that; the propylene is decomposing at the temperatures used to form ethyl.ene, acetylene and methane, If we consider the points at low propylene fraction (where the effects of propylene decomposition will be small), it is seen that the acetylene formation data scatter too much to a meaningfulo The only conclusion that can be drawn. from te data is that the polymerizati.on to butadiene i.s not inhil.bited by propyleneo

V79The product distribution experiments can be studied to obtain the sequence of formation of the various products which can be the basis of speculation into possible mechanismso In the case of ethylene decomposition (see Figure 19) the butadiene appears to be formed from polymerization of ethylene rather than the reaction of acetylene with ethylene. The carbon appears to result from the decomposition of acetylene since none is formed until the acetylene quantity has become appreciable, The vinylacetylene (CH4) can result either from dehydrogenation of butadiene or the polymerization of acetylene. The benzene can result from the reaction between vinylacetylene and acetylene. The experiments on acetylene decomposition (see Figure 25) show that the carbon can result by direct dehydrogenation of the acetylene and the vinylacetylene from the polymerization of two acetyleneso The benzene can result from the reaction of vinylacetylene with acetylene. It is emphasized that these last comments based on the product distributions are purely speculative since they pre-suppose a molecular process for which there is no evidence for or against in this work.

VIo KINETIC CORRELATION FOR THE COMPLETE SERIES OF REACTIONS IN THE FORMATION OF ACETYLENE FROM ETHANE Ao Overall Kinetic Correlation The rate data on the individual reaction steps are now combined to give an overall. kinetic correlation. Kinetic correlations have been developed for the individual thermal. decompositions of ethane, ethylene and acetylene in Section IV. These are now combined to give the following overall scheme: First order C^H6 = C2H4 + H2 (1) First order C2H - 2CH - H (2) First order CH4 =^ C2H2 + H2 (3) Second order C2H4 = 1/2(C4 polymer) + H2 (4) First order C2H2 = 2C + H2 (5) Second order C2H2 1/2(C4 polymer) (6) These reactions represent the best correlation of the rate data and are not intended to demonstrate the reaction mechanism. The negative hydrogen in reaction (2) is necessary to maintain material balance. The C4 polymer in reactions (4) and (6) includes all the C4, C5 and C6 products and also the small amounts of C3 products formed. The C4 polymer is assumed to have the approximate formula C4H4 for material balance purposes. The assumption is made that all of the methane formed results from the decomposition of ethane. It is pointed out that the experimental conditions in this work were such that the reverse reactions (1 and 3) were negligible (See Appendix VI). The rate constants for the six -80

-81reactions are summarized in Figure 36 and the values of the activation energies and pre-exponential factors are indicated on this figure. The ethylene decomposition data are represented empirically by three straight line sections with different apparent activation energies. This overall kinetic model can be used to calculate the product distribution at all points in any reactor under any reaction conditions. Some additional experiments (Runs 128 - 131) were carried out with ethane feed at reaction conditions such that all of the reactions were occurring (ioe. very high conversions) to provide a check on the correlation. The correlation was also checked out against some of the experiments that were carried out in the measurement of the rate constants for the individual reactions. The rate equations based upon the six reactions already discussed are: AVc, iCl K ] Wc"6 C CC.] (22) AVR r.,6 K A k c -- - (24) AVR L 2 1C 25 ci WCAC = 2 2 ^Ccz14 (25) VVR.

-82\ \ RATE CONSTANTS ON C2 BASIS \| \ xlOI 4 E A TOK KCAL / i10d - \ -- GM. MOL I /SEC \ \ I >8.45 k, 82.4 6.04x 10 \ \k' |>6.6,<8.45 k 35.4 1.25 x 08 \n / ^< 6.6 k, 66.9 3.20x ld 10 \ 1 k3 76.2 1.76x 10 6\. P (0 k 5 61.9 97010 Figure\\ 36. RateConstants, for theSixMajoLITER / \.GM.MOL-SEC \ \\ k \ 60.3 2.63 x10 \ \ Z),0 \\ \ (1) k2H6 6 CH +H 1sod(4 H =12Cpl44.6 3.19 x 20 cn \\ 10 10,, 10' 6.0 70 80 9.0 1.0 1.1 TK X 10 Figure 36. Rate Constants for the Six Major Reactions. (1) C2H6 C2H4 + H2 1st order (4) C2H = 1/2 (C4 poly.) + 2 2nd order (2) C2H6 = 2CH4 - H2 1st order (5) C2H2 2C + 12 1st order (3) C2H4 C2H + H 1st order (6) C2H2 = 1/2 (C4 poly.) 2nd order

-83NC~ = 2 k4LCc2i + j6L Cc2tJ (26) LNc = L c(27) a VR a w+htec-eNzis Cr2 ikn4 g (28) where N is the rate in gm. moles/secon. C is the concentration in gm. -moles/liter, and Kc the equilibrium constant in concentration unitso'The rate constants (k) are represented as temperature functions by the familiar Arrhenius equation I = A -E/RT (21) The equilibrium constants are also represented by an exponential temperature function of the same form as the Arrenhius equationo These equations were solved for the product distributions with the particular flow rates and temperature distributions by Euler'sfinite difference technique. oThis calculation is quite tedious so it was programmed for machine calculationo The flow diagram for the computer program is contained in Appendix VII. The computed product distributions are compared with the experimental distributions for a number of runs in Table I (Calco l)o Runs Nos. 4-3 - 46 and 128 - 131. are all. for ethane feed and good agreement is observed for the ethane, ethylene, methane and hydrogen quantitieso The predicted carbon quantity seems to be somewhat high although the run

-84TABLE 1. Comparison of Correlation with Experiment Run No. Mole % C2H6 C2H4 CzH2 CH4 C4 H2 C N2 Exp. - 0.086 12. 5 1.05 0.424 1.29 85.8 116 Calc. - 0.005 12.1 - 1.22 0.836 1.68 84.2 Calc. 2 - 0.005 12.1 - 1.22 0.836 1.68 84.2 Exp. 0.167 10.2 4.48 0.686 1. 17 6.73 0.895 76.6 48 Calc. 1 0.014 10.0 4.12 0.007 1.39 8. 18 1.30 75.0 Calc. 2 0.014 9.99 4. 13 0.007 1.39 8. 19 1.30 75.0 Exp. 7.83 8.76 0. 136 0.723 0.08 8.39 - 74.0 43 Calc. 1 7.91 9.25 0.016 0.435 0.01 8.74 73.6 Calc. 2 7.91 9.09 0.086 0.435 0.05 8.89 - 73.5 Exp. 5.23 10.45 0. 344 0.725 0. 151 10.6 - 72.4 44 Calc. 1 4.49 11.76 0.064 0. 628 0.029 11.3 - 71.7 Calc. 2 4.48 11.24 0.316 0.625 0.138 11.8 - 71.4 Exp. 5.83 7.33 0.246 0. 691 0.078 7.27 - 78.5 45 Calc. 1 5.89 7.80 0.036 0. 539 0.009 7.35 - 78.4 Calc. 2 5.88 7.48 0.236 0. 538 0.057 7.62 - 78.2 Exp. 2.95 8.96 0.799 1.17 0.118 10.2 - 75.7 46 Calc. 1 2.93 10. 10 0. 175 1.09 0.044 9.80 - 75.8 Calc. 2 2.92 9.07 0.813 1.08 0.185 10.6 - 75.2 Exp. 4.68 7.74 0.484 0.970 0.233 8.13 - 77.6 128 Calc. 1 4.28 9.07 0.093 0.917 0.023 8.60 - 77.0 Calc. 2 4.26 8.37 0.533 0.911 0.122 9. 19 - 76.6 Exp. 0.985 8.25 1.96 1.94 0.547 12.4 - 73.9 129 Calc. 1 0.547 10. 14 1.01 2.41 0.211 11.7 0.10 73.9 Calc. 2 0. 554 7.75 2.52 2.36 0.450 13.7 0.35 72.3 Exp. 0.128 5.65 4.22 2.47 0.732 15.4 - 73.2 130 Calc.1 0.018 6.02 3.44 3.08 0.652 15.9 1.46 69.4 Calc. 2 0.016 4.15 4.08 3.01 0.836 17.7 2.48 67.7 Exp. Reactor plugged during run 131 Calc. - 0.09 0.91 2.49 0.974 20.3 10.7 64.5 Calc. 2 - 0.07 0.78 2.49 0.997 20. 5 10.8 64.4

-85(No. 131) when the reactor plugged with carbon coincides with a marked increase in computed carbon quantity~ The quantity of acetylene and C4 polymer computed is considerably less than the experimental value at low concentrations of acetylene although the agreement is quite good when appreciable amounts of acetylene are formedo The ratios of the predicted to the experimental acetylene quantity are plotted against the ethane to ethylene ratio in Figure 37, and it is seen that there is a definite trendo This high initial rate of acetylene formation is attributed to the attack of methyl. radicals from the ethane on the ethylene to product acetylene and C4 polymerO The rate of acetylene formation and C4 polymer then is multiplied by a x ethane/ethylene ratio where a is an empirical constant (found to be 5~5 by selection of the best fit to the data). Runs 116 and 48 were experiments on acetylene and ethylene decomposition respectively and the calculated and experimental results are seen to agree fairly wello The product distributions were recalculated including the empirical correction discussed above and the results are compared with experiments in Table I (Calco 2) to show that the agreement is satisfactoryo The correlation allows the prediction of the product distribution throughout the reactor and this was done for Run No. 129, the results of which are plotted in Figure 38. Bo Use of the Correlation to Predict Product Distributions over a Wide Range of Reaction Conditions The correlation can be used to predict the product distributions for a wide range of reaction conditionso The correlation is based on data taken in the temperature range 730~C to 1330 ~C at a total pressure of 1 atmosphere with nitrogen dilution from 60% to 90%. Some example

-86cU Xn I < IIE NN5 0 04 <Mj~~J (OLS CH_ Figure 37. Effect of Etude on Formation of Ac-etylene from Eth /lene 2 0 0.2 0.4 0.6 0.8 1.0 MOLS C 2H6) MOLS C.H4 Figure 37. Effect of Ethane on Formation of Acetylene from Ethylene.

-871200 1100 __I 1000 1TEMPERATURE 900 PROFILE 800 700 600 ----- 14 i | I I\ 1 lXI I 1'1I/ | EXPTL POINTS __\ / | X /CH4-\ /I: 40 EXPTL POINTS_ ______ _____ lose C2H6_2H4. C2H2- - C4 -~v DISTANCE ALONG REACTOR IN INS. Figure 38. Calculated- Product Distribution Throughout the Reactor for Run 129.

-88calculations were carried out which represent an extrapolation of this data since the reaction conditions considered were total pressures of 025 atm. (with no diluent) and 1. atm. (with no diluent and 75% diluent) at temperatures of 1000~C, 1200~C, 1.400~C and 1600~Co These calculations were carried on a digital computer using Euler's method to solve the differential rate Equations (Nos. 22 to 28) The calculations by this finite difference technique are made very much more convenient if the temperature of the gas is allowed to rise at a finite rate at the entrance of the reactor rather than as a step function to the maximum temperature. Incidentally, this sort of temperature profile is much more realistic from a practical point of vieWo Therefore, the temperature in the first reactor increment was assumed to be 5000C and rose by 1000C in each successive increment until the desired maximum temperature was reached. In these example calculations no consideration was given to the heat transfer limitations that would have to be considered in a reactor design as detailed design such as this was not an objective of this worko The results of these calculations are shown graphically in Figures 39 to 44 in which the yields of products (expressed on the basis of 100 moles of ethane feed) are plotted against reactor increment. The residence time scale is shown beneath the reactor increment scale as this is a much more general parameter. The residence time scale is noted to be nonlinear with respect to equal reactor increments and this is due to changing temperature and extent of reaction at each point. The temperature distribution, pressure conditions and amount of diluent are shown on each graph. The acetylene yield increased with increase in temperature and decrease i.n pressure and decrease in The residence times were very small especially at the highest temperatureso

PRESSURE: 1.0 ATM. PRESSURE: 1.0 ATM. 0 DILUENT: NONE DILUENT: NONE ------- Tmx. 14000 ~C Tmox: 1200 ~C F? 1400 5 —---- 1400 < ~C(0 O- 15000 TEMP.2 TEMP. - 600 600 120. —-- " 120 -- 110 110FCiH6 9H CrH6 //_ 6 H2 o 100 - --- 100ad —w - 90 90 sou 8 80 80 0 70 70 0) 8 60 60 -J 0 50 50 o 40 40 > 30 30 RESIDENCE TIME, MILLISECS RESIDENCE, MILLISECS diluent, 100000 and 120000).

PRESSURE: I ATM. PRESSURE: I ATM. o DILUENT: NONE DILUENT: NONE Tmax: 1400 C Tmox 1600 ~C 1400 1400 TEMP. _ TEMP. L. 100c 0000.600 6001O 120 H1O ---------— H 110 C2Hr C Hs 100 0.22 0.5 05 0 02 100 - 90 / H2 9 80 80 RE-, - - N / ~- R TM 70 70 o 60 60~ J C2Ha 0 50 C-H 5u) d 40 40 300- 30 w C 2___ 4 __ __ 5: 20 2 I0lo J/^^^^ ^.POLY_ 10 1. 0 5 10DiSTANCE15 00 5 DISTANE15 20 o01 I I DISTANCE I o i I DIST I 0 0.25 0.5 0.75 1.0 0 0.025 005 0.075 0.10 0.125 RESIDENCE TIME, MILLISECS RESIDENCE TIME, MILLISECS Figure 40. Computed Product Distributions versus Residence Time (1 atm, no diluent, 1400oC and 1600Cc).

PRESSURE: I ATM PRESSURE: I ATM o DILUENT: 75% % DILUENT: 75%0 w Tmux: 1000 OC Tmox 1200 0C F 1400 1400 a 1000 --- ooo 1000_ TEMP. — _TEMP 600 600 120 120 110 -------- 110 C2H --- H2 C2H6 H2 o9 0 90 / I C2H4HI 80 8C —-- 0 1\~~ _ __ I~^2 PL _ __^ // ^ PL._ ^ o 70 70 5 8 D6 60 cn 50 2 50 C 1 17 Figure~~~~~~~~~~~~~~~~~ ~ 41 Coptd rdc itiuin essRsdneTm lam 5 o 40. din 40 a 20C w 5:: 30 11 1 I I)30 20 2C CH, C H POLY. OV IO _ _ _ _ _ 0 5 10 15 20 5 10 15 20 1 1, DISTANCE DISTANCE 0 I I I I I __I _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 0 25 50 75 100 125 150 0 5.0 75 10.0 12.5 15.0 175. 20.0 RESIDENCE TIME, MILLISECS RESIDENCE TIME, MILLISECS Figure 41. Computed Product Distributions versus Residence Time (i atm, 75% diluent, 10000C and 12000C).

PRESSURE: I ATM. PRESSURE: I ATM.. DILUENT: 75 % DILUENT: 75% W Tmox: 1400 ~C Tmox: 1600 ~C _____ _ TEMP. n 1000 -- ---- 1000L. - 600 600 120 120 110 110 C 2H C2,H 1 00 100 — C w H u- 90 90 H 80 / 80 / =Q' I ____ / / / 60 60 C-._) 0 50 C 50 4 __ _/ 40C —) 60 6 ~~~~20 20~~~ 20 — C2I 20 C,o --'"Q ^ ^ ^ Sh POLY. 10 00 5 K) 15 20 0 5 10 15 20 0 I TANCE I DISTANCE I IDISTANCE 0 0.5 1.0 1.5 2.0 0 0.025 0.050 0.075 RESIDENCE TIME, MILLISECS RESIDENCE TIME, MILLISECS Figure 42. Computed Product Distributions versus Residence Time (1 atm, 75% diluent, 1400oC and 1600~C).

PRESSURE: 0.25 ATM. PRESSURE: 0.25 ATM. o. DILUENT NONE DILUENT: NONE W Tmox 1000 ~C Tmox: 1200 ~C - 1400 1400 | 1000 1000TEMP. Li, TEMP. -- 600 600 120 i -" H120 110 H 110 —- H2 C2H ( o 100 100 W w u" 90 I / CH90 _170 0 70 0 0 60 C60 o 50 50 2 o 40 60 20 20o 4 0 C4 HC. CHC 20 o O10o. I10 CH4 CH o ~ POLY. 0 5 10 DISTANCE 15 20 0 5 10DISTANCE 15 20 0 1 I I I I I I I 1 1 i I 50 100 150 200 0 5 10 15 175 20 RESIDENCE TIME, MILLISECS RESIDENCE TIME, MILLISECS Figure 43. Computed Product Distributions versus Residence Time (0.25 atm, no diluent, 10000C and 1200~C).

o PRESSURE: 025ATM. PRESSURE: 0.25 ATM. o DILUENT: NONE DILUENT: NONE Tmox: 400 ~C Tmox 1600 ~C I- 1400 1400 zt 1 4 0 0. I I / I.TIEMR TEMP. I 1000 100o' 600 600 120 - 1 I —--- 120 110 110 C H 26 o 100 1 w H 2 - _ _ _ _ _ _ _ _ * 90 90. \/ C2H6 / 8 80 H2 0 707 -j__ __ _ 7 CH4 4 o \ / ^ TO —-- \ / __ 0 6060 2H2 0 50 50 0 40 40 — 30C -- i C H * 30 10 10 0 5 0 ISTANCE15 20 O01 DISTANCE 0 DISTANCE 0 0.5 1.0 1.5 2.0 2.25 2.5 0.025 0.05 0.075 0.10 0.125 RESIDENCE TIME, MILLISECS RESIDENCE TIME, MILLISECS Figure 44. Computed Product Distributions versus Residence Time (0.25 atm, no diluent, 1400~C and 1600~0).

VII. CONCLUSIONS A technique for carrying out a kinetic study in a non-isothermal field has been developed. The technique requires considerable mathematical labor (which is easily handled on a digital computer) but the need for an isothermal experiment is removed. The reactions studied in this work were fast, high temperature reactions for which the isothermal experiments can only be crudely approximated. Precise kinetic data were obtained and it is concluded that this technique will have wide general application in the field of fast high temperature kinetic studies. The reaction studied in this work was the thermal decomposition of ethane to acetylene. The reaction proceeded through a series of consecutive steps from ethane to ethylene to acetylene to carbon. The kinetics of thermal decomposition of ethane, ethylene and acetylene were studied separately because of the complexity of the overall reaction, The kinetic data for ethane decomposition was correlated with a model of two homogeneous parallel first order reactions to ethylene and methane. The rate constants for ethylene formation fell off considerably at higher temperatures and this was due to inhibition of the reaction by the secondary reaction products butadiene and propylene. The data of this work were found to be generally in agreement with previous workers The ethylene decomposition was correlated by a model of two homogeneous parallel reactions which were a first order formation of acetylene and a second order polymerization to butadiene (which subsequently reacted to vinylacetylene and benzene). -95

-96The acetylene decomposition was correlated by a model of two homogeneous parallel reactions which were a first order formation of carbon and a second order polymerization to vinylacetylene (with subsequent reaction to benzene)o The decomposition of ethane was found to be a chain reaction and it was concluded that the inhibition by the reaction products was due to shortening of the chain length. This is the only conclusion concerning mechanism that can be made as a result of this worko The decomposition of ethane was, however, shown to be consistent with the free radical chain mechanism ideas of Rice and Herzfeldo No conclusions could be made with regard to the mechanism of ethylene and acetylene decompositiono The kinetic data on the individual steps were combined and an overall correlation developed. Some experiments were carried out in which all of the reactions were occurring and these product distributions were compared with calculated values using the overall correlationo Reasonable agreement was obtained so it is concluded that this correlation can be used to predict the product distribution for any desired reaction conditionso A few calculations were carried out using this correlation over a wide range (part of the range represents an extrapolation) of the variables, temperature, pressure and residence time. The acetylene yield was found to increase with increase in temperature and decrease in pressureo The residence times were very small especially at the higher temperatures.

m97Future work of interest would be to investigate methane and propane as starting materials. It is thought that ethylene would be the intermediate preceding acetylene for these feed materials too so that a considerable part of this kinetic study would be applicable.

APPENDIX I RAW EXPERIMENTAL DATA -98

TABLE II. Raw Data: Product Distribution Experiments (All runs in reactor No. 2) Run No. 80 81 83 84 86 87 88 89 126 115 78 127 Temp. ~C (max) 816 816 816 816 996 996 996 996 996 1043 1052 1036 Feed Gases Hydrocarbon C2H6 C2H6 C2H6 C2H6 C2H4 C2H4 C2H4 C2H4 C2H4 C2H2 C2H2 C2H2 C2H6 C2H4 C2H2 Hydrocarbon Feed Rate 7.72 2.42 4.81 1.18 4.06 1.60 3.75 1.07 1.71 8.35 9.08 6.08 Gm.ml/sec. x10-5 x10-4 x10-4 x10-3 x10-5 x10-4 x10-4 x10-3 x10-3 x10-5 x10-5 x10-4 N2 Feed Rate 2.72 7.30 1.65 3.31 1.63 5.14 1.22 3.12 5.02 2.81 5.18 6.65 Gm. ml/sec. x10-4 x10-4 x10-3 x10-3 x10-4 x10-4 x10-3 x10-3 x10-3 x10-3 x10-3 x10-3 Pressure, Atm. 0.992 0.993 0.999 1.011 0.992 0.993 0.999 1.011 1.030 1.000 1.037 1.037 k0 Product Distribution Mole% C2H6 13.82 21.35 21.2 25.3 0.192 0.196 0.155 0.133 0.058 - - 0.029 96.8 - - C2H4 7.28 3.20 1.40 1.13 9.19 16.3 19.57 23.7 24.45 0.027 0.367 0.136 2.1 99.8 C2H2 - - - - 1.78 1.57 1.05 0.269 0.080 2.50 8.02 6.97 - - 97.9 CH4 0.495 0.20 0.075 0.048 2.01 0. 945 0.376 0.037 0.007 0.022 0.200 0.054 0.05 - H2 6.32 2.58 0.757 0.429 7.46 4.53 2.59 0.830 0.385 0.057 1.92 0.116 - - - C3Hg 0.086 0.041 0.020 0.028 0.049 0.046 0.087 0.055 0.053 - - 0.026 0.08 - - C3H6 0. 154 0.210 0.264 0.282 0.076 0.145 0.118 0.075 0.075 - - - 1.0 C3H4 - - - - 0.049 0.095 0.078 0.035 - 0.029 0.043 0.091 - - - C4H6 0.031 - - 0.253 0.583 0.631 0.428 0. 258 - - - C4H4 - - - - 0.194 0.296 0.232 0.057 0.018 0.064 0.554 0.286 - - - C4H2 - - - - 0.012 0.031 0.061 0.040. 0.003 0.009 0.024 0.033 - - - C5H6 - - - - 0.071 0.133 0.086 0.015 0.006 0.006 - 0.024 - - - C6H6 - - - - 0.703 0.495 0.173 0.028 0.011 0.011 0.723 0.042 - - - C - - - - 4.95 0.853 - - - 0.218 5.46 1.04 - N2 71.9 72.5 76.3 72.8 73.1 74.0 74.9 74.3 74.6 97.2 82.8 91.2 - 0.21 2.12 1. C4H8 2. Acetone

TABLE III. Raw Data: Reaction Order Experiments (All runs in reactor No. 2) Run No. 49 50 51 52 53 54 55 56 57 58 115 114 113 112 Temp. ~C (max) 874 874 874 874 875 1082 1082 1082 1082 1082 1043 1043 1043 1043 Hydrocarbon C2H6 C2H6 C2H6 C2H6 C2H6 C2H4 C2H4 C2H4 C2H4 C2H4 C2H2 C2H2 C2H2 C2H2 Hydrocarbon Feed Rate 2.49 4.46 6.32 1.02 1.68 2.28 4.00 6.10 9.92 1.59 8.32 1.42 2.06 3.89 Gm.ml/sec. x10-4 x10-4 x10-4 x10-3 x10-3 x10-4 x10-4 x10-4 x10-4 x10-3 x10-5 x10-4 x10-4 x10-4 N2 Feed Rate 4.52 4.36 4.11 3.62 3.02 4.55 4.41 4.15 3.72 3.16 2.80 2.77 2.68 2.56 Gm.ml/sec. x10-3 x10-3 x10-3 x10-3 x10-3 x10-3 x10-3 x13 x10-3 x10-3 x10-3 x10-3 x10-3 x0-3 Pressure, Atm. 0.998 0.998 0.998 0.998 0.998 1.001 1.001 1.001 1.001 1.001 1.000 1.000 1.000 1.000 H 0 Product Distribution Mole% C2H6 4.65 8.18 11.66 19.0 30.7 - - - - - - - - C2H4 0.494 0.938 1.33 2.21 3.45 4.14 7.00 10.36 16.02 24.55 0.027 0.032 0.030 0.068 C2H2 0.026 0.015 0.063 0.097 0.123 0.501 0.821 1.36 2.29 3.38 2.50 4.15 5.93 10.46 CH4 0.029 0.049 0.073 0.132 0.227 0.028 0.069 0.136 0.351 0.743 0.022 0.033 0.044 0.078 H2 0.384 0.701 1.08 1.80 2.89 0.509 0.985 1.71 3.25 5.46 0.057 0.093 0.147 0.292 C3H8 - 0.037 0.059 0.079 0.139 - - - - - - - - C3H6 0.038 0.077 0.111 0.182 0.304 - - -- - - - C3H4 - - - - - - - - - - 0. 029 0. 048 0.063 0.108 C4H6 - - - - - 0.051 0.122 0.255 0.480 0.852 - - - - C4H4 - - - - - - 0.072 0.175 0.363 0.633 0.064 0.158 0.278 0.633 C4H2 - - - - - - - - - - 0.009 0.015 0.031 0.064 C5H6 - - - - - - - - - 0.126 0.006 0.015 0.028 0.052 C6H6 - - - - - - 0.018 0.049 0.150 0.311 0.011 0.018 0.052 0.207 C - - - - - - - - - - 0.218 0.382 0.516 0.986 N2 94.3 90.0 85.6 76.2 62.15 94.8 90.9 86.0 77.0 64.0 97.2 95.5 93.3 88.0

TABLE IV. Raw Data: Rate Constant Experiments Run No. 18 19 21 22 42 43 44 45 46 118 96 97 98 128 129 130 131 Reactor No. 1 1 1 1 2 2 2 2 2 2 3 3 3 2 2 2 2 Temp. ~C (max) 743 765 816 843 841 896 971 996 1055 782 843 901 1010 1032 1138 1225 1338 Hydrocarbon C2H6 C2H6 C2H6 C2H6 C2H6 C2H6 C2H6 C2H6 C2H6 C2H6 C2H6 C2H6 C2H6 C2H6 C2H6 C2H6 C2H6 Hydrocarbon Feed Rate 3.80 9.20 8.35 8.74 1.16 1.85 4.24 6.10 6.31 1.41 4.72 1.024 2.22 8.50 8.65 8.32 4.84 Gm.ml/sec. x10-5 x10-5 x10-4 xlO-4 xl0-4 xl0-4 xl0-4 xl0-4 xl0-4 xl0-4 xl0-5 xl0-4 xl0-4 xl0-4 x10-4 xl0-4 xl0-4 N2 Feed Rate 1.90 4.86 3.88 3.81 4.35 8.01 1.86 3.50 3.54 3.75 2.46 5.13 1.12 4.80 4.88 4.52 3.34 Gm.ml/sec. xlO-4 x10l4 X10l3 xl003 xIO4 x04 x103 x l3 x10-3 xl0-4 xlO-4 xl0-4 xl0-4 x0-3 xl0-3 xl0-3 xl0-3 Pressure, Atm. 0.965 0.973 1.071 1.071 0.982 0.990 0.994 1.010 1.003 0.978 0.979 0.982 1.006 1.032 1.039 1.042 1.048 Product Distribution Mole % CZH6 15.22 14.50 15.72 13.87 13.83 7.83 5.23 5.83 2.95 26.05 11.8 7.19 1.44 4.685 0.985 0.128 (1) C2H4 1.08 1.15 1.75 3.85 5.71 8.76 10.45 7.33 8.96 0.850 3.51 7.64 10.32 7.74 8.25 5.65 C2H2 - - - - 0.040 0.136 0.344 0.246 0.799 - - 0.066 0.921 0.484 1.96 4.22 CH4 0.048 0.097 0.085 0.210 0.317 0.723 0.725 0.691 1.17 0.047 0.236 0.633 1.666 0.970 1.94 2.47 H2 0.517 0.836 1.092 3.10 4.91 8.39 10.58 7.27 10.17 0.373 2.94 7.14 12.42 8.13 12.4 15.4 C3H8 - - - - 0.075 0.061 - - - 0.080 0.081 0.047 0.027 0.058 0.028 - C3H6 0.128 0.125 0.130 0. 115 0.124 0.091 0. 107 0.098 0.093 0.276 0. 102 0.081 0.068 0.092 0.076 0.043 C3H4 - - - - - - - - - - - - 0.021 0.034 0.053 0.080 C4H6 - - - - 0.043 0.082 0.151 0.078 0.118 - - 0.069 0.175 0.114 0.147 0.073 C4H4 - - - - - - - - - - 0.101 0.065 0.183 0.310 C4H2 - - - - - - - - - - 0. 023 0. 009 0.035 0. 093 0 C5H6 - - - - - - - - - - - 0.032 0.029 C6H6 - - - - - - - - - - - - 0.060 0.030 0. 096 0.166 N2 82.91 83.3 81.3 78.9 75.1 74.0 72.4 78.5 75.7 72.1 81.4 77.2 72.7 77.6 73.9 73.9 Center Wall Center Wall Center Wall Center Wall Temp. ~C 538 538 538 538 538 711 538 727 404 560 640 431 457 480 564 525 457 603 487 552 682 at one 657 632 643 621 632 776 638 788 606 704 718 592 545 603 688 684 652 735 656 726 904 inch 733 727 738 732 721 807 721 827 738 772 796 664 633 727 812 843 846 868 827 900 1016 intervals 743 743 765 765 777 824 785 849 783 840 837 735 707 754 828 871 901 904 885 948 1082 727 739 761 766 805 830 822 860 829 857 877 781 782 782 843 899 956 942 943 995 1116 703 717 743 749 816 827 840 857 836 874 891 827 827 777 835 891 965 957 967 1018 1143 677 689 721 732 812 818 843 851 833 870 904 849 871 771 827 895 973 970 991 1039 1169 648 660 696 704 805 805 838 842 817 865 910 870 893 756 810 883 970 979 1010 1054 1188 623 635 673 682 793 793 830 832 802 858 915 883 918 740 794 872 966 988 1029 1071 1219 606 615 655 660 782 782 820 822 790 852 919 896 930 729 785 865 966 1000 1053 1102 1243 593 603 642 649 771 770 811 812 779 851 922 909 942 717 776 857 966 1013 1079 1132 1266 587 596 632 638 754 755 801 804 775 849 927 921 954 712 776 855 968 1021 1095 1155 1288 582 592 622 627 738 741 791 793 771 852 932 932 966 707 773 854 969 1030 1112 1180 1307 580 590 616 627 721 724 777 780 776 854 941 943 979 710 780 858 974 1030 1119 1191 1319 577 590 610 621 701 710 762 768 780 865 950 956 992 713 785 862 979 1031 1127 1203 1327 571 590 599 618 679 693 744 752 793 874 962 970 1006 720 799 872 989 1032 1132 1212 1335 562 590 588 610 657 680 721 735 806 885 969 985 1021 728 814 882 999 1032 1138 1220 1338 554 588 576 604 632 652 692 716 816 895 971 993 1038 731 819 885 1004 1017 1136 1222 1335 545 582 560 593 604 621 655 691 835 896 971 995 1049 733 824 888 1010 996 1134 1225 1321 - 566 - - - 593 616 659 841 893 964 996 1052 717 807 885 999 973 1123 1216 1304 - 538 - - - 560 577 630 840 883 939 992 1055 700 791 857 988 949 1112 1205 1282 593 830 854 888 971 1050 656 746 816 956 816 1093 1186 1232 --- 793 781 793 922 1030 614 703 775 924 682 1075 1167 1160 711 593 621 757 979 429 554 506 716 - 924 993 1057 449 - 538 631 782 - - - 395 - 774 819 827 (1) Reactor plugged with carbon during this run

TABLE IV. (Continued) Run No. 24 67 68 47 48 99 100 101 110 111 112 116 117 102 103 104 82() 85 Reactor No. 1 2 2 2 2 3 3 3 2 2 2 2 2 3 3 3 2 2 Temp. 0C (max) 899 1046 1096 1132 1197 1024 1079 1130 937 990 1043 1099 1118 1049 1013 949 832 872 Hydrocarbon C2H4 C2H4 C2H4 C2H4 C2H4 C2H4 C2H4 C2H4 C2H2 C2H2 CH2 C2H2 C2H2 C2H2 C2H2 C2H2 C2H6 C2H6 Hydrocarbon Feed Rate 1.25 5.65 1.73 5.34 8.53 2.77 7.95 2.74 3.94 1.44 3.77 8.75 1.32 2.06 1.04 8.17 2.32 5.12 Gm. ml/sec. xl0-4 xlO-5 xl0-4 xl0-4 x10-4 x10-5 xl0-5 xl0-4 x10-5 x10-4 x10-4 xl0-4 xl0-3 xl0-4 xl0-4 x10-5 xl0-4 xl0-4 N Feed Rate 4.05 3.81 1.14 2.39 3.63 2.52 6.59 1.55 3.84 1.07 2.48 4.80 6.98 1.55 6.48 5.43 7.11 2.30 Gm.ml/sec xl0-4 xl0-4 xl0-3 xl0-3 xl0-3 xl0-4 xl0-4 xl0-3 xl0-4 xl0-3 xl0-3 xl0-3 xl0-3 xl0-3 x10-4 xl0-4 xl0-4 xl0-3 Pressure, Atm. 0.973 0.975 0.985 1.009 1.020 0.986 0.995 1.024 0.980 0.985 1.000 1.025 1.067 1.019 0.995 0.992 0.993 1.008 Product Distribution Mole % C2HM - - - 0.147 0.165 0.030 - - 0.061 - - - - - - - 21.9 16.35 C H4 19.59 6. 12 6.56 11.2 10.1 5.09 5.69 8. 13 0. 119 0. 152 0.068 0.085 0. 118 0.057 0.083 0.086 1.86 1. 16 C2H4 0.562 2.65 3.20 3.15 4.45 1.64 2.44 3.27 6.61 8.70 10.4 12.3 12.7 9.31 10.0 10.0 - 0.035 CH4 0.64 0.916 0.660 0.593 0.680 0.835 0.525 0.616 0.065 0.079 0.077 0.104 0.115 0.061 0.095 0.070 0.302 0.229 H2 2.43 5.74 5.32 5.32 6.68 3.63 4.37 5.52 0.256 0.298 0.290 0.418 0.409 0.321 0.494 0.306 0.896 0.625 C3H8 - - - 0.027 - - - - - - - - - - 0.075 0.074 C3H6 - - - - - 0.028 - 0.047 - - - - - - - - 1.73 1.67 C3H4 - 0.066 0.070 0.106 0.113 0.021 0.035 0.064 0.064 0.081 0.107 0.149 0.173 0.085 0.079 0.109 - - C4H6 1.14 0.126 0.126 0.222 0.182 0.079 0.077 0.127 - - - - - - - - - - C4H4 - 0.273 0.282 0.401 0.405 0.146 0.220 0.347 0.496 0.624 0.627 0.669 0.612 0.538 0.742 0.726 - - 0 C4H2 - 0.047 0.070 0.078 0.101 0.041 0.048 0.077 0.031 0.052 0.063 0.075 0.072 0.046 0.048 0.036 - - (. C5H6 - - - - - - - 0.044 0.042 0.059 0.051 0.049 0.056 0.043 0.056 0.055 - - CGH6 0.17 0.400 0.349 0.232 0.260 0.173 0.178 0.247 0.273 0.200 0.205 0.153 0.151 0.146 0.382 0.323 - - C - 1.91 0.764 0.840 0.888 1.74 1.23 0.928 1.24 1.56 0.983 1.27 1.36 0.976 1.48 0.474 - - N2 75.4 82.0 82.7 77.8 76.0 88.1 85.2 80.6 90.6 88.6 87.2 84.7 84.0 88.5 86.5 87.7 73.4 79.8 Temp. ~C 638 546 577 511 554 420 375 453 423 549 450 521 477 424 543 512 438 471 at one 777 721 782 639 684 720 651 671 777 691 615 616 590 603 657 632 577 608 inch 849 846 907 769 778 838 816 890 842 832 710 734 701 780 829 754 718 664 intervals 880 973 1030 831 835 956 982 967 907 891 806 799 769 857 882 807 766 720 893 1002 1059 893 893 984 1021 1043 921 950 860 863 835 935 938 860 816 743 899 1032 1088 935 919 1013 1060 1071 937 970 916 898 864 971 963 888 829 767 893 1040 1092 977 944 1019 1069 1100 937 988 947 932 893 1007 988 916 832 780 881 1047 1096 999 970 1024 1077 1108 933 989 979 961 920 1021 997 927 822 793 867 1043 1095 1021 999 1022 1075 1117 931 990 998 990 946 1035 1007 939 812 804 850 1041 1094 1038 1027 1014 1073 1120 927 988 1017 1012 982 1042 1010 943 804 815 832 1038 1092 1054 1055 1010 1071 1123 923 985 1027 1035 1019 1047 1013 949 796 823 810 1035 1091 1069 1079 1007 1070 1124 921 984 1035 1054 1043 1048 1013 949 791 832 782 1032 1064 1082 1103 1007 1070 1127 919 982 104b 1074 1068 1049 1013 947 785 842 758 1030 1038 1093 1127 1008 1070 1127 918 980 1043 1081 1082 1047 1008 942 788 851 730 1027 1058 1104 1143 1010 1071 1127 917 978 1043 1089 1096 1043 1004 938 790 859 699 1024 1080 1116 1162 1011 1072 1128 916 975 1043 1093 1106 1042 996 928 799 868 663 1020 1072 1127 1177 1016 1075 1130 915 972 1041 1099 1118 1041 989 918 807 870 629 1015 1066 1132 1190 1021 1080 1125 910 959 1038 1088 1107 1029 969 891 812 872 593 999 1041 1132 1193 1012 1071 1121 904 946 1017 1077 1099 1019 949 866 818 862 552 983 1017 1124 1197 1004 1065 1098 873 910 995 1054 1080 982 904 812 806 851 - 943 972 1117 1190 977 1038 1074 842 871 960 1032 1061 946 860 760 795 830 - 903 928 1089 1173 947 1011 1030 788 818 926 990 1021 895 801 701 762 808 - 836 735 1038 1130 888 957 985 734 765 987 947 981 843 743 641 730 756 - 770 731 831 939 830 904 711 630 556 803 724 757 605 467 391 - 550 - 680 - - 778 466 560 438 526 349 549 - - 367 - - - (2) 1. 78 x 10 - gm. mol/sec. propylene added in feed (3) 5. 12 x 10-5 gm.mol/sec. propylene added in feed

TABLE V. Raw Data: Inhibitor Experiments Run No. 70 71 72 73 74 119 120 121 122 123 124 125 90 91 92 93 94 95 Temp. ~C (max) 863 863 863 863 863 880 880 880 880 880 880 880 1091 1091 1091 1091 1091 1127 Hydrocarbon CzH6 C2H6 C2H6 C2H6 C2H6 C2H6 CZH6 CZH6 C2H6 CzH6 C2H6 C2H6 C2H4 CGH4 CZH4 C2H4 CGH4 C3H6 Hydrocarbon Feed Rate 5.50 4.97 5.07 5.08 5. 16 2.24 2.23 2.33 2.26 2.26 2.25 2.24 1.07 1.07 1.07 1.07 1.07 3.21 Gm.ml/sec. x10-4 x10-4 x10-4 x10-4 x10-4 x10-4 x10-4 x10- 4 x 10- 4 x10- 4 x10-43 x10-3 x10-3 x10-3 x0-3 x0-4 Inhibitor C3H6 C3H6 C3H6 C3H6 C3H6 C4H6 C4H6 C2H4 C2H2 H2 -C3H6 C3H6 C3H6 C3H6 Inhibitor Feed Rate 5.21 4.59 1. 14 3.28 1.08 -- 7.02 7.22 1.87 9.62 1.87 -- -- 1.91 7. 18 1.12 2.25 Gm. ml/sec. x10-8 x10-6 x10-5 x10-5 x10-4 x10-6 x10-5 x10-5 x10-6 x10-5 x10-5 x10-5 x10-4 x10-4 N2 Feed Rate 4.17 4.22 4.21 4.20 4.15 2.17 2.16 2.18 2.18 2.18 2.16 2.18 3.29 3.26 3.30 3.29 3.32 3.20 o Gm. ml/sec. x10-3 x10-3 x10-3 x10-3 x10-3 x10-3 x10-3 x10-3 x10-3 x10-3 x0-3 x0-3 xlo-3 xl0-3 x0-3 xl0-3 xl3 x 3 Pressure, Atm. 1.004 1.004 1.004 1.004 1.004 0.990 0.990 0.990 0.990 0.990 0.990 0.990 1.016 1.016 1.01 6 1.016 1.013 Product Distribution Mole % C2H6 10.29 9.52 9.89 10.13 10.52 7.34 7.88 8.44 7.55 7.49 7.38 7.55 0.211 0.140 0.105 0.118 0.120 0.140 C2H4 1.19 0.886 0.722 0.496 0.213 1.94 1.47 1.05 2.26 1.80 1.83 1.76 18.92 19.8 20.2 20.35 20.55 3.12 C2H2 - - - - - - - - - 0.391 - - 1.64 1.95 1.34 1.34 1.33 1.59 CH4 0.047 0.057 0.065 0.059 0.066 0.121 0.118 0.130 0.119 0.117 0.117 0.116 0.398 0.475 0.793 1.02 1.62 3.62 H2 0.939 0.601 0.454 0.253 0.089 1.59 0.995 0.428 1.46 1.42 2. 51 1.42 3.34 3. 17 2.68 2. 55 2.80 3.09 C3/H8 0.039 0.041 0.038 0.026 0.055 0.036 0.035 0.041 0.037 0.038 0.040 0.031 0.034 0.050 0.047 0. 079 0.072 C3H6 0.001 0.096 0.240 0.691 2.26 0.071 0.091 0.132 0.074 0.075 0.073 0.069 0.094 0. 161 0.608 1.09 2.07 1.31 C3H4 - - - - - - - - - - - - 0.092 0.102 0.196 0.290 0.433 0.748 C4H6 - - - - - - 0.289 2.87 - - - - 0. 573 0. 567 0. 529 0. 539 0. 536 0. 164 C4H4 - - - - - - - - - - - - 0. 311 0.292 0.246 0.228 0.232 0.208 C4H2 - - - - - - - - - - - - 0.077 0. 071 0.057 0.058 0.052 0.040 C5H6 - - - - - - - - - - - - 0.058 0.071 0.090 0. 106 0. 151 0.116 C6H6 - - - - - - - - - - - - 0. 175 0. 152 0. 161 0. 179 0.208 0.535 C 87.5 88.8 88.6 88.4 86.9 88.9 89.1 87.0 88.5 88.7 88.0 89.2 74.0 73.1 73.0 72.0 N2 87.5 88.8 88.6 88.4 86.9 88.9 89.1 87.0 88.5 88.7 88.0 89.2 74.0 73.1 73.0 72.0 69.80 85.2

APPENDIX II RESULTS CALCULATED FROM RAW DATA -104

-105TABLE VI. Results: Product Distribution Experiments Hydrocarbon Feed C2H6 Run No. Temp. Conv. Hydrocarbon Product Distribution OC (max) % Moles/100 moles of hyd. feed reacted C2H4 CH4 H2 C3H8 80 816 33.2 96.5 7.06 92.0 1.25 81 816 11.3 96.3 7.35 94.9 1.51 82 816 4.35 96.8 6.70 78.7 2.08 83 816 2.29 97.0 6.05 72.3 4.73 Hydrocarbon Feed C2H4 C2H2 CH4 H2 C4H6 C4H4 86 995 49.2 19.9 22.5 83.7 2.84 2. 18 87 995 29.4 23. 1 13.8 66.7 8.53 4. 33 88 995 14.8 30.8 11.0 76.0 18.5 6.80 89 995 7.17 14.7 2.04 45.4 23.4 3. 10 126 995 3.86 8.20 0. 67 39.4 26.4 1.80 C6H6 C C2H6 C3H8 C3H6 86 7.90 55.5 2. 16 0.54 0.86 87 7.24 12.6 2.86 0.68 2. 12 88 5.07 - 4. 55 2.56 3.47 89 1. 55 - 7.27 3.02 4.09 126 1. 11 - 5.97 5.43 7.72 Hydrocarbon Feed C2H2 C4H4 C6H6 C H2 C2H4 78 1043 44. 6 8. 55 11.2 84.2 29. 7 5. 67 127 1043 19.6 16.8 2.42 61.0 6.8 8.0 115 1043 12.5 17.8 3.05 60. 5 15.8 7.4 CH4 C3H4 78 3. 10 1. 13 127 6.0 5.45 115 3.12 7.60

-1O6TABLE VIL Results: Order of Reaction Experiments Hydrocarbon Feed CzH6 Temp.~C Hydro- Hydrocarbon Reaction Rate RunNa (max.) carbon Gm. moles/sec. Mole.Fr.1 Disappearance Formation C2H6 C2H4 CH4 49 874 0.0486 2.04x 10-5 1.98 x 10-5 1.42 x 10-6 50 874 0.0857 3.82 x 10-5 3.69 x 10-5 2.40 x 10-6 51 874 0. 122 5.55 x 10-5 5.38 x 10-5 3.52 x 10-6 52 874 0.200 9.25 x 10-5 8.78 x 10-5 6.29 x 10-6 53 874 0.322 1.44 x 10-4 1. 38 x 10-4 1. 15 x 10-5 Hydrocarbon Feed C2H4 C2H4 C2H2 Polymer2 54 1082 0.0444 2.93 x 10-5 2.41 x 10-5 5.57 x 10-6 55 1082 0.0763 6.09 x 10-5 4.00 x 10-5 2.32 x 10-5 56 1082 0.115 1.09x10-4 6.61x10-5 5.20x10-5 57 1082 0. 183 2. 18 x 10-4 1. 14x 10-4 1. 12 x 10-4 58 1082 0.284 3.82 x 10-4 1. 68 x 10-4 2.21 x 10-4 Hydrocarbon Feed C2H2 C2H2 C Polymer2 112 1043 0. 119 8.26 x 10-5 2.88 x 10-5 6.82 x 10-5 113 1043 0.0656 3.50 x 10-5 1.48 x 10-5 2.76 x 10-5 114 1043 0.0452 2.08 x 10-5 1. 11 x 10-5 1.53 x 10-5 115 1043 0.0270 1.04x 10-5 6.30 x 10-6 7.26 x 10-6 1. Arithmetic mean of inlet and outlet 2. C4's, C5's, C6's, CH4 (Expressed on C2basis)

-107TABLE VIII. Results: Rate Constant1 Experiments Pair of Activation Pre-exp. Temp. l/Temp. Rate Pair of Activation Pre-exp. Temp. 1/Temp. Rate Runs Energy Factor (max) x 104 Const. Runs Energy Factor (max) x 104 Const. kcal/gm. mole sec-1 ~C oK- 1 sec-1 kcal/gm. mole sec-1 ~C ~K-1 sec-1 Disappearance of ethane3, first order 112 10 1043 7.60 19.9 116 57.78 7.66 1099 7.29 48.1 18 19 743 9.83 0.117 19 765 9.64 0.310 112 1043 7.60 21.2 117 6.847 4.87 xla 1 118 7.19 84.5 18 87.44 7. 17 743 9.83 0.115 21 816 9.17 2.02 102 3.7x 1049 7.56 17.6 103 42.38 1.78x10 1013 7.78 11.3 18 743 9,83 0.113 22 8497 2.04x1017 842 8.95 4.78 103 1013 7.78 8.59 104 17.01 6.61x0 3 949 8.18 6.07 19 9 765 9.64 0.289 21 84.70 1.95 x 1017 816 9.17 1.99 21 39.45816 9.17 1.99 Methane formation from ethane, first order 19 8 765 9.64 0.287 22 82.39 6.30 l016 842 8.95 4.70 18 7 743 9.83 0.00301 21 9154 1.41 017 816 9.17 0.0606 21 77.78 7.61 x 1015 816 9. 17 1.89 22 842 8.95 4.52 19 4 26 x 1014 765 9.64 0.0117 69.18 4.26x1014 22 842 8.95 0.123 118 760 9.68 0.209 42 70.54 3.78x1019 841 8.96 6.57 118 53.72 47 109 760 9.68 0193 42 841 8.96 0,132 42 62 841 8.96 5.20 43 896 8.55 21.4 43 47.95 6.28 x 108 896 8.55 0.703 44 971 8.03 2.44 43 896 8.55 19.3 44 5504 367971 8.03 79.5 45 536 7.5.996 7.88 5.35 46 5366 1055 7.52 13.7 45 39.45 6.45108 996 7.88 104 Polymerization of ethylene, second order2 128 36.42. 8 1032 7.65 143 129 1138 7.08 406 24 63.80 899 7.53 145 67 63.8 1014 10 47 7.57 2,560 82 4 832 9.05 1.21 8 65.94 1.33 X 1013 832 9.05 1.21 85 872 8.72 3.52 67 7 1047 7.57 2, 150 68 7250 6.366x a 0 1096 7. 30 6,000 96 70.54 3.22 a 1014 843 8.95 5. 17 97 901 8.51 23.8 42 1132 7.12 12,500 9 56.70 7.57 x 1011 901 8.51 21.2 98 1010 7.79 167 Acetylene formation from ethylene, first order 19 103.8 2.a18 x 1021 765 9.26 2,81 19 1 2.18x10-2765 9.64 0.312 4 899 8.53 0.0652 103.82148x121 765 9.64.312 2 81.467 l O 1047 7.57 3.87 2 74.32 1.14 x 1015 927 9. 13 1.72 22 956 8.89 4.23 67 952 8 34 1015 1047 7.57 4.13 68 1096 7.30 14.5 lDisappearance of ethylene, first order 47.1132 7.12 24.2 --------- 48 61.93 1.03 x 10 11 97 6.80 65.2 24 69 899 8.53 0.435 67 64194 5.42 1011 047 7.57 9.65 Carbon formation from acetylene, first order 67 13 047 7.57 8.55 -Tcalohrseonetr 68 76.28 3.74x10 1096 7.30 24.4 937 8.26 0.813 110 55.937 8.26 0.813 ~~~47 1132 7.12 45.6 ~~ ~ I 55.39 7.90 10 990 7.91 2.16 48 53.61 9.90 x 109 32 7.9 2 456 116 72.75 4.15 x 1012 1099 7.29 10.6 99 1024 7.71 7.48 100 60 1.03x 1079 7.40 20.9 112 1043 7.60 3.70 117 82.64 1.90 1014 18 7.19 19.50 100 1079 7.40 18.2 101 70.01 3.79 x 1014 1130 7.12 46.5 Acetylene polymerization, second order2 Disappearance of acetylene, first order 10 937 8.2 6 2,810 110 5.8 15a111 937 8.26 3.40 111 5908 1.56 1011 990 7.91 9.54 112 1043 7.60 11,970 116 35.64 9.83 0 1099 7.29 20,650 112 1043 7.60 13,300 117 7.35 1 1118 7.19 34,100 1. Rate constant defined on C2 basis 3. Ethylene formation rate constant is.95 times ethane 2. Rate constant units are liter gin. mol~1 -ec~1 4. Based on Tc + 3/4(Tw - Tc), all other based on center Tc 2. Rate constant units are liter gm. mo1-1 sec

-108TABLE IX. Results: Inhibitor Experiments Temp. Mole Ratio Run No. (max) Inhibitor Inhibitor Rate of formation ~C to Feed Gm. moles/sec. CzH4 CH4 C3H8 Ethane Feed x105 xl06 x107 70 863 C3H6 0. 0003 5.01 2.03 15.9 71 863 C3H6 0. 0092 3. 17 2. 54 15.9 72 863 C3H6 0. 0225 2. 38 2. 92 14.0 73 863 C3H6 0. 0677 1.29 2.66 8.25 74 863 C3H6 0. 209 0. 434 2.92 21.6 119 880 None 0 4.26 2.96 8.78 120 880 C4H6 0.0312 3.09 2.86 8.52 121 880 C4H6 0.310 2. 13 3.28 10.3 122 880 C2H4 0. 0600 3.69 2.92 9.10 123 880 C2H2 0. 0424 3.94 2.88 9.30 124 880 H2 0. 100 4.02 2.89 9.72 125 880 None 0 3.81 2.84 7.57 C2H2 C4H6 Ethylene Feed x105 x105 90 1091 None 0 7.30 2.55 91 1091 C3H6 0.0179 8.70 2.52 92 1091 C3H6 0. 0672 6.05 2.38 93 1091 C3H6 0.105 6. 15 2.64 94 1091 C3H6 0.211 6.35 2.48

APPENDIX III CALCULATION OF CORRECT MEAN TEMPERATURE FOR A LINEAR RADIAL TEMPERATURE DISTRIBUTION Rewriting Equation (18) A = ~ ( \RiJ[H I -ze - x (18) AL e7 E/t)() dependent is the integral / -/i- -r ) A (29) jo Ttr The objective now is to determine the correct value of temperature (T) to substitute into Equation (29) when a linear radial temperature gradient is assumed. This was accomplished by a numerical calculation with n taken to be unity. The reactor was divided into five equal area rings and the temperature at the mid point of each of the rings determined from the temperature gradient. The linear temperature gradient was thereby approximated by a series of steps. The values of e /R/T were then evaluated for each of the steps and the arithmetic mean of the five values was taken as the correct solution. The value of T that corresponds to this mean value of e E/R/T was then the correct mean temperature, The numerical work is summarized below for a linear temperature distribution 154o~R at the center (TC) and 1600~R at the wall (TW).'109

-110Fractions of radius that give five equal 0o446 o 632 Oo 775 0.895 1.0 area rings Mid points of rings 0.223 Oo 539 0.703 Oo835 Oo947 Temps. at mid points OR 1553.4 157204 1582.2 1590 1596.7 Values of -E/R/T x 1021 0.974 1.60 2.09 2.50 2o93 "E/RT / 921 Therefore, the mean value of e /T is 2.02 x 10 - which corresponds to a mean temperature of about- 07 of the difference between the wall (Tw) and center (TC) temperatures. In view of the approximate nature of the.calculation, a mean temperature of 3/4 of the difference is used, i.e. T mean = T( 3/4 (TTC (0) C ~ W

APPENDIX IV SAMPLE CALCULATIONS The pair of runs 47 and 48 (ethylene feed) are selected for this example solution. The kinetic model used for this calculation is a first order disappearance of ethylene. The experimental data for these two runs are: Run Number 47 48 Inlet conversion (xi) 0 0 Outlet conversion (xo) 0.360 o.434 Feed rate C2H4 gm.moles/sec(F) 5.54 x 10 8.55 x 10 -4 Feed rate N2 gm.moles/sec(ND) 2.39 x 10-3 3.63 x 10 Mean pressure atm. (P) 1o009 1.020 Reactor Number 2 2 Temperature Profile ~C(T() ) 511 1021 1127 554 999 1177 1-inch intervals 639 1038 1132 684 1027 1190 (read down columns) 769 1054 1132 778 1055 1193 831 1069 1124 J 835 1079 1197 893 1082 1117 893 1103 1190 935 1093 1089 919 1127 1173 979 1104 1038 944 1143 1130 999 1116 831 970 1162 939 778 Rewriting Equation (18) from section IIF F(}^\)rnQC1\ oo 4 ^F r t-E/RT(Q) = \ r/^oc; ^ \(18) - -111- E / RT -111

-112Since the kinetic model in this case is a first order reaction n is 1o Trial values of E are assumed and values of A are computed from (18) using the data of run 47 and then the data of run 48~ The calculations are carried out on a digital computer and the results are: Assumed E values Ln A(A in 1/hr) kcal/gmo mol. Run #47 Run #48 27,78 21o684 22.054 38.89 25.795 26 e016 44.44 27.841 27~981 50O10 29.882 29.938 55.55 31o918 31.888 61oll 33~952 33.834 66.67 35.982 35o776 72.22 38.010 37T715 77T78 40o035 39 651 83~33 42.058 41,584 88.89 44.079 43.515 1llo11 52.149 51o224 The solution is the E value for which the A values are equal and it can be seen from the results above that this lies between the E values of 50o10 and 55~55~ It was found that in A is almost linear with respect to E so that linear interpolation is quite accurate, which gives the solution E 53.61 kcals/gmo molo and in A - 31o205 ioe A = 3,56 x 1013 hr1 or 9.90 x 109 sec -1 The values of the valurate constant of the Arrhenius equation using the maximum temperature for each of the runs,

-113Run #47 Run #48 T max ~C 1132 1197 T max OK 1405 1470 -E/RT 1 53o61 53.61 A e sec 9.90 x 109 x e R x1405 9.90 x 109 x e R x 147 k sec 45o6 106.5 1/T~K 7.12 x 104 6.80 x 10 These two values of the rate constant can be found on Figure 21.

APPENDIX V MODIFICATION OF CALCULATIONS TO HANDLE PARALLEL REACTIONS The equations developed in section II are of the form A -> B + C It was found, however, that the decompositions proceeded by two parallel steps of the form, A -~ B + C A - D A, B and D are hydrocarbons and C represents hydrogen. The equations have to be modified so that the rate constants for the formation of B and D can be determinedo Rewriting Equation (18), F ( RA{ r F\- Nb F A X. \ CPJJ A \-. J! (18) L J o T@) it can be seen that the only part of (18) that has to be modified is the integral, rfi r \ F x t ND A ^~ xz (31) Let the total conversion of A be x, the conversion (XB) of A to B be Px and the conversion (XD) of A to D be 5x. p and 5 are constants and their sum is unity. -ll4

-115The (1 + x + ND/F) in Equation (31) represents the total number of moles divided by the feed rate F. so that for the two parallel reactions this becomes \- oC~- Z o- C + ~ 4 /w (32) which is \ + x(2 4I-i^ F -1+ND9(33) The (1 - x) in Equation (31) represents the mols of A left divided by the feed rate F and this remains unchanged. Therefore, the integral (31) for the case of the two parallel reactions above becomes, / + r + (Z+ - ) + N/F A (34) This Equation (34) can now be written twice in terms of each of the parallel reactions since xB = Px and xD = 6x, 2q3+/, I X (2 /9 A W^ (35) JOC^ BJ v L 1N) Equations (35) and (36) replace the upper integral in Equation (18). The new equations have the same form as the old ones, the only difference being the addition of some constant factors so that the method of solution proceeds as before.

APPENDIX VI CALCULATIONS TO SHOW EFFECT OF REVERSE REACTION Nitrogen dilution was used to reduce the reverse reaction effects of the reactions, C2H6 v C2H4 + H2 (1) C2H4 C2H2 + H (2) Some calculations were made and are summarized here to show that the effects of reverse reaction are negligible. The rate equations considering the reverse reactions will be Coand H6 disappearance rate k1 CC2G- V (37)< and K ^ C024 disappearance rate= k2 C C - C- C4 (38) where kl, k2 are rate constants, Kc 1, Kc2 are equilibrium constantso The second terms in Equations (37) and (38) represent the reverse reaction effects. The first and second terms are calculated for a number of experimental runs using exit concentrations and the second term expressed as a percentage of the first termo Partial pressures which are proportional to the concentrations are used in the calculations summarized below. Reaction C2H6 C2H4 + H2 Run NOo T max C2H6 press0 C2H4 x H2 Ratio % 0~C atm Kp -- atm 18 743 0.152 0.000113 0o074 118 782 Oo261 Oo0000317 0.012 21 816 0.157 Oo000119 0.076 44 971 0.0523 Oo00103 1.97 46 1055 0.0295 Oo000364 1.24 130 1225 0.0128 o000oooo668 0.522 -116

-117Reaction C2H4 v C2H2 + H2 Run No. T max C2H4 press. ~2H4 x H2 Ratio % ~C X atm. Kp,2 atm. 24 899 0.196 0.00286 1.46 67 1046 0.0623 0.0037 5.94 68 1096 0.0661 0.0222 3.36 47 1132 0.1125 0.00139 1.23 48 1197 O 102 0.00125 1.23 It is seen that the reverse reaction term can be considered. negligible, so that an irreversible reaction can be assumed.

APPEND IX VII FLOW DIAGRAMS FOR COMPUTER PROGRAMS (i) Solution for E and A values The equation that is to be solved is Equation (18) below f\0 ~ i - - (18) L., T(ce) The method of solution is to obtain a set of A values for an assumed set of E values for each experimental run. A pair of runs are considered together to obtain unique solutions for E and Ao The integrals in Equation (18) were evaluated numerically by Simpson's Ruleo The upper integral was divided into 20 increments for the integrationo The reactor length was divided. into 25 increments and the temperature distribution was approximated by a series of small steps so that the temperature in each increment could be assumed constanto The reaction order n is known from a separate set of experiments so that it will have a numerical value at the time of calculation. The first step of the program evaluates the conversion integral, the second step the temperature integral and the final step the A value corresponding to the E value assumed. The next E value is selected and the calculation repeated. The flow diagram for this calculation follows. -118

-119Evaluation of upper integral I^ E^ T~l/ I —--- I -X, l ---- 11 =1 -* --- ^ l) = X ~ l -l) ^-A oc — I1 1 1* _ D || i-Su y - 0 Evaluation of lower integral J 0- _2. _\ l\ ) - Calculation of A 1? y are indexes Y. iS G, Hu are intermnediate variables

-120(f Rate Eqa-uation- for Pii sectiont ui) ion of, ixr reactions tat ^earppeoThe rate equations for the VIB are: (22) A ______ - (23) A~~~fr. L 1 r T ^~~~~~~~~~~(4) (25) AA1^ + ~ ~~~~~~~~~(26) (27) iM-^ ~CC -WCZ\<~V~Cc ~~~~~~A2-~~~~~~~~~~~~2 c -1 I f? 1 (28) \rN [c% -L'VJ

-121The rate constants (k) are represented. by the Arrhenius equation l< = A e R (21) The reactor is divided into 25 increments and. the temperature profile is approximated by a number of small steps so that the temperature is constant in each increment. The rate dN/daVR is approximated by AN/AVR and the AN in each increment is computed using the concentrations at the end of the previous increment. In this way the calculation proceeds throughout the reactor. The flow diagram for this calculation follows. -N) ^) ^'AN k)AV. C(0) - c^) - ^ TLjjK are indexes representsEquations (22) to (28)

-122APPENDIX VIII TABLE X APPROXIMATE VALUES OF BOND ENERGIES Bond energies in kcal C2H5 - H CH3 H3 98 83 Steacie (44) CH2CH - H CHC - H 104 121 C- H C - C 92 77 Daniels 4) C = C C' C 12.2 200

BIBLIOGRAPHY 1. Akin, G. A., Reid, T. F. and Schrader, R. J. American Institute of Chemical Engineers Meeting, Chicago, Illinois, December 9-14, 1957. 2. A.PoIo Research Project #44, Selected Values of Physical and Thermodynamic Properties of Hydrocarbons and Related Compounds, Carnegie Press, 1953. 3. A.P.I. Research Project #44, Catalog of Mass Spectral Data, Carnegie Institute of Technology, 1959. 4. Aten, C. F. and Greene, E. F. Disc. Faraday Society, 22, 162, (1956). 5. Benson, S. Wo The Foundations of Chemical Kinetics, McGraw-Hill Book Company Inc., New York, 1960. 6. Bixler, G. H. and Coberly, C. W. Ind. Eng. Chem. 45, 2596, (1953). 7o Bogart, M,, Schiller, G. R. and Coberly, C. Wo Petroleum Processing, 8 377, (1953). 8. Bone, W. A. and Coward, H. F. J. Chem. Soc., 93, 1197, (1908). 9. Bone, W. A. and Coward, H. F. ProcO Chem. Soc. (London), 24, 167, (1908) 10. Brooks, B. T., Boord, C. E., Kurtz, S. S., Jr., and Schmerling, Lo The Chemistry of Petroleum Hydrocarbons, Vol. II, Reinhold Publishing Company, New York, 1955. 11. Burk, R. E., Baldwin, B, G. and Whitacre, CO Ho Indo Engo Chem., 29, 326, (1937). 12. Calderbank, P. H. and Hovnanian, V. P. Ind. Chemist, 33, 557 (1957). 13. Dahlgren, G. Jr., and Douglas, J. E. J. Amo Chemo Soc., 80, 5108, (1958). 14. Daniels, F. Ind. Eng. Chemo, 35, 5o4 (1943). 15. Dinstes, A. I., Kvyatkovskii, D. A., Stepukhovich, Ao D. and Frost, A. V. Zhur. Obsh. Khimo, 7, 1754, (1937). 16. Eastwood, S. C. and Potas, A. E. Petroleum Engro, 19, 43, (1948). 17. Eltenton, G. C. J. Chemo Phys., 10, 403, (1942)o 18. Eltenton, G. C.o J. Chem. Phys., 15, 455, 465, 474, (1947). -123

-12419o Farnesworth, J Fo., Manes, Mo, McGurl, G. V. and. Bretz, Go M. Ind. Engo Chem., 47, 1517, (1955). 20. Frey, Fo E. and Smith, Do Fo Ind. Engo Chemo, 23, 948, (1928). 21o Greene, Eo Fo, Taylor, Ro L. and Patterson, W. Lo Jro, Jo Physo Chem., 62 238 (Feb. 1958), 22, Hague, E. No and Wheeler, Ro. V. J Am. Chem. Soc., 51, 378, (1929)o 23. Hasche, Ro Lo Chem. and Met. Engro, 49, #7, 78, (1942). 24. Hepp, H. Jo, Spessard, Fo Po and Randall, J, Ho Ind. Engo Chem., 41, 2531, (1949)o 25. Hobbs, Jo E. and Hinshelwood., Co N. Proc. Royo Soco (London), A167, 439, (1938). 26. Kinney, Ro E. and Crowley, Do Jo Ind. Eng. Chemo, 46, 258, (1954), 27. Kuchler, L and Thiele, H. ZO Phys. Chemo, B.42, 359, (1939). 28. Lee, E. H. and Oliver, G. D. Indo Eng. Chemo, 51, 1351, (1959). 29. Marek, L. Fo and McCluer, W. R. Ind. Engo Chem., 23, 878, (1931)o 30o Miller, I. Fo University of Michigan Industry Program, Report IP-392 October 1959 (Ph.Do Thesis, University of Michigan)o 31. Molera, M. Jo and Stubbs, F. J. Jo Chem. Soco 381, (1952)o 32. Paul, Ro Eo and Marek, L. F. Ind. Engo Chemo, 26, 454, (1934). 33 o Pease, Ro No Jo Amer. Chemo Soco 50, 1779, (1928)o 34. Pease, Ro No J. Amero Chemo Soco 51, 3470, (1929)o 35. Rice, Fo 0Oo and Herzfeld, Ko Fo Jo Am. Chemo Soco, 56, 284, (1934). 36. Rice, Fo 0. and Dooley, M. Do Jo Amo Chieo Soco, 55, 4245, (1933). 37o Sachsse, Ho Z. Physik, Chemo, B31, 87, (1935)o 38. Schutt, H. Co Chemo Engr. Progo, 55, #1, 68, (1959)o 39o Semenov, No N. Some Problems in Chemical Kinetics and Reactivity, VolSo I and II, Princeton University Press, Princeton, 1959o 40o Sittig, Mo Petroleum Processing, July, 1011 (1955).

-12541, Snow, R. H., Peck, R. Eo and Von Fredersdorff, C. Go A.I.ChoE. Journal, 5, 304 (1959)o 42, Staveley, L. Ao K, Proc. Roy. Soc. (London) A162, 557 (1937). 43. Steacie, E. WO R. and Shane, Go Cano J. Research, B18, 351 (1940)o 44. Steacie, E, Wo R. Atomic and Free Radical Reactions, Vols I and II, Second Edition, Reinhold Publishing Company, New York, 1954. 45, Storch, H. H. and Golden, PO L. Indo Eng. Chemo, 26, 56, (1934)o 46. Taylor, H. A. and van Hook, A. Jo Phys, Chemo, 39, 811 (1935)o 47. Tropsch, H., Parrish, C. Go and Egloff, Go Ind. Engo Chem., 28, 581, (1936). 48. Tropsch, H. and Egloff, G. Indo Eng. Chemo, 27, 1063, (1935)o 49o Zelinski, N. Do Ber., 57, 264 (1924).

3 9015 032 40 NOMENCLATURE A Pre-exponential factor in Arrhenius equation (sec )* a Cross sectional reactor flow area (cm2) C Concentration (gm. mole/liter) E Activation energy in Arrhenius equation (kcal/gm. mole) F Feed rate of hydrocarbon (gmo moles/sec) Kp Equilibrium constant pressure units, concentration units k Rate constant (sec-1)' L Total length of reactor (cm) 1 Arbitrary length along reactor (cm) M Mole fraction N Flow rate of any component (gmi moles/sec) ND Flow rate of inert diluent (gm. moles/sec) NT Total flow rate (gm. moles/sec) P Pressure (atmospheres) R Gas constant T Temperature (~K) TC, W Reactor temperature at center, wall (~K) VR Reactor volume (liters) x. Conversion inlet, outlet 1, 0 a,, Constants These units are for first order reaction. Units for second order reaction are liters gmo mol-1 sec-1l ~126