A STUDY OF THE HYDROLYSIS REACTIONS OF I>IISIcI3 AND MOZSI DI2 IN A LAMINAR FLOW TUBULAR REACTOR Dissertation for IheDegree of Ph. D. MICHIGAN STATE UNIVERSITY - IOHN HOWARD CAMERON 1975 IIIIIIII III IIIIIIIIIIII I II 3219301087 6146 This is to certify that the thesis entitled A STUDY OF THE HYDROLYSIS REACTIONS OF PhST'C'I3 AND MeZST'CT2 IN A LAMINAR FLOW TUBULAR REACTOR presented by John Howard Cameron has been accepted towards fulfillment of the requirements for Eh_._D_.____deg1-eein Chemical Engineering “(M/U fh— £7; ' {c «t;- ”((7 Major professor 0/; [/2 14"75' Datefl- / I ,/ ABSTRACT A STUDY OF THE HYDROLYSIS REACTIONS OF PhSiCl AND MezSiCl IN A LAMINAR LON TUBULAR EACTOR By John Howard Cameron This work contains a study of the hydrolysis reactions of PhSiCl3 and MeZSiCl2 in a laminar flow tubular reactor. The aims of this work were to model these reactions, investigate the mechanisms involved, determine the effect of HCl on these reac- tions, and estimate the effects of the plug flow and isothermal assumptions for the reactor. The hydrolysis reactions of chlorosilanes consist of the replacement of the chlorines by hydroxyl groups. Little work has been done on these reactions due to their fast rate and complexity. There are three reasons for the fast rate of these reactions rela- tive to similar carbon reactions. These are the ionic nature of the silicon-halogen bond, the larger diameter of silicon compared to carbon, and silicon's unfilled 3-d orbitals. The complexity of these reactions is due to the effect of HCl on the system and a competing condensation reaction. HCl has the ability to either catalyze or suppress the hydrolysis reactions of chlorosilanes and is also known to catalyze the condensation reaction. John Howard Cameron In this study, the individual hydrolysis reactions of the chlorosilanes were followed using an infrared spectrophotometer. The absorption spectrum of the chlorosilanes and their hydrolysis products were obtained. Using these spectra, infrared band assignments were made for the reactants, unstable intermediates and products. A variable length reactor surrounded by a heat exchanger and coupled with an infrared spectrophotometer was used to monitor the reactions. The silane solvent mixture entered the reactor through a long needle, similar to a hypodermic needle. By varying the position of this needle in the reactor, the length of the reaction zone was varied. Using this technique, data on concentra- tions versus residence time was obtained. The hydrolysis reactions of PhSiCl3 can be written as a series of three irreversible reactions. K PhSiCl3 + H20————l—+ PhSiCl2(OH) + HCl K PhSiC12(0H) + H20-———§—+ PhSiCl(0H)2 + HCl K + H O-———§-+ PhSi(0H)3 + HCl PhSiCl(0H)2 2 To determine the effect of HCl on the reactions, the ini- tial concentration of HCl in the system was varied. The mechanism of this effect is complex, with the reaction being catalyzed through the stabilization of the transition state intermediate by John Howard Cameron the chloride ion and being suppressed by the reaction of water with the proton to form the hydronium ion. However, the principal effect of HCl was to suppress the rate of these hydrolysis reac- tions. The effect of HCl was incorporated into the set of differ- ential equations describing the reactions by using the term [HCl]". The rate constants and n were found to have the following values at 0°C: K1 = 220, K2 = l6.5, K3 = 40.6 (liters/mole)1+"/seconds and n = -0.6l2. The second system studied was the hydrolysis reactions of MeZSiClz. The following series of reactions were found to describe these hydrolysis reactions. K . l . MeZSTCl2 + H20-———————+ MeZSTCl(0H) + HCl K MeZSiCHOH) + H20 ._—-_:2.__--* MeZSi(OH)2 + HCl The rate constants found to describe these reactions at 0°C are K1 = 38.3, K2 = ll.l, K; = 20.2 (mole/liter-second). By study- ing these reactions at different temperatures, the activation energies were determined: AE1 = 3.6 kcal/mole, AE2 = 10.5 kcal/ mole, AEé = l4.0 kcal/mole. These reactions were modeled assuming that plug flow existed and that the system was isothermal. A principal aspect of this study was to determine the magnitude of the error introduced by these assumptions. The study of these assumptions begins with John Howard Cameron a material and energy balance on a laminar flow tubular reactor. The model is based on a ring-shaped element in the reactor with conduction and convection of energy in the axial direction and conduction in the radial direction. Partial differential equations describing the temperature and concentration profiles in the reac- tor are developed. Using a finite difference technique and physical constants similar to those found in the hydrolysis reac- tions of MeZSiClz, a computer program is employed to integrate these equations. Second order rate constants needed to give the same conversion in a laminar flow reactor with heat transfer are compared to those in an isothermal plug flow reactor. It was determined that the rate constants determined assuming isothermal plug flow are approximately 10 percent lower than those determined ,assuming laminar flow with heat transfer. This produces a bound of lo percent between these two models. A STUDY OF THE HYDROLYSIS REACTIONS OF PhSiC'l3 AND MeZSiCl IN A LAMINAR 2 FLON TUBULAR REACTOR By John Howard Cameron A DISSERTATION Submitted to Michigan State University in partial fulfillment of the requirements for the degree of DOCTOR OF PHILOSOPHY Department of Chemical Engineering 1975 ACKNOWLEDGMENTS The author wishes to express his sincere appreciation to Dr. Martin C. Hawley for his guidance and advice throughout the course of this work. Appreciation is also extended to the members of my Thesis Guidance Committee for their advice and suggestions. The author is indebted to the Division of Engineering Research for their support of this project and to the Dow Corning Corporation for furnishing the necessary reagents. The advice and effort of Don Childs in the fabrication of the experimental appa- ratus is appreciated. Special thanks is given to my wife, Pam, for her under- standing and encouragement. ii TABLE OF CONTENTS LIST OF TABLES . LIST OF FIGURES Chapter I. II. III. IV. INTRODUCTION . A. General Chemistry of Silicon B. Scope of Past Work C. Scope of This work . D. Organization of This Work EXPERIMENTAL . A. Apparatus . B. Temperature Control C. Reagents . 0. Experimental Technique THE HYDROLYSIS REACTIONS OF PHENYLTRICHLOROSILANE A. Interpretation of the Infrared Spectra Obtained During the Hydrolysis Reactions of Phenyltrichlorosilane . B. Preparation of PhSi(OH) C. Condensation Reactions . D. Calculation of Concentrations . . E. Conditions for the Hydrolysis Reactions of PhSiCl . . . . . . F. Result; and Discussions . G. Solvent Effect . H. Concluding Remarks THE HYDROLYSIS REACTIONS OF DIMETHYLDICHLOROSILANE A. Interpretations of Infrared Spectra Obtained During the Hydrolysis Reactions of Dimethyldichlorosilane Condensation Reaction Page vii 61 66 Chapter Page C. Calculation of Concentration . . . . . 68 D. Conditions for the Hydrolysis Reactidns. of Me SiCl2 . . . . . . . . . . . . . 68 E. HCl Effect . . . . . . . . . . . . 69 F. Results and Discussion . . . . . . . . . 7l G. Temperature Effect . . . . 78 H. Conclusions on the Hydrolysis of MezSiCl2 . . . 85 V. ANALYSIS OF A LAMINAR FLOW REACTOR . . . . . . 88 A. Sc0pe of Past Work . . . . . . . . . . 88 8. Scope of This Study . . . . . . 89 C. Development of the Laminar Flow Model . . . . 90 D. Conclusions . . . . . . . . . . . . . 99 VI. CONCLUSIONS . . . . . . . . . . . . . . l05 BIBLIOGRAPHY . . . . . . . . . . . . . . . . 109 NOMENCLATURE . . . . . . . . . . . . . . . . llZ APPENDICES . . . . . . . . . . . . . . . . . ll4 A. Data Used to Model the Hydrolysis Reactions of PhSiCl3 and MeZSiCl2 . . . . . . . . ll5 8. Numerical Method and FORTRAN Program Used to Analyze a Laminar Flow Reactor With Heat Transfer . . . . . . . . . . . . . . . 127 iv Table 10. 11. A-1. A—2. A-3. LIST OF TABLES Infrared Assignments for PhSiCl3 and Its Hydrolysis Products Conditions for Condensation Reaction of PhSi(OH)3 Extinction Coefficients for PhSiCl3 and Its Hydrolysis Products . . . . Parameters Used to Model the Hydrolysis Reac- tions of PhSiCl3 . . . . Physical Constants of Acetonitrile and l,2-Dimethoxyethane Parameters Used to Describe the Hydrolysis Reac- tions of PhSiCl3 in Acetonitrile . . Infrared Assignments for Me SiCl and Its 2 2 Hydrolysis Products . Extinction Coefficients for Me SiCl and Its 2 2 Hydrolysis Products . Parameters Used to Model the Hydrolysis Reactions of Me SiCl . . . . . . . . 2 2 Rate Constants for the Hydrolysis Reaction of MeZSiCl2 at Different Temperatures Activation Energies for the Hydrolysis Reactions of MeZSiCl2 . . . . . . Summary of Initial Run Conditions for the Hydrolysis Reactions of PhSiCl3 . . . . . . . Absorption and Concentration Values at Various Residence Times for the Hydrolysis Experiments of PhSiCl3 Summary of Initial Run Conditions for the Hydrolysis Reactions of PhSiCl3 in Acetonitrile . . Page 23 30 33 4O 47 54 61 68 72 81 85 116 117 120 Table A-4. A-S. A-6. A-7. Absorption and Concentration Values at Various Residence Times for the Hydrolysis Reactions of PhSiCl3 in Acetonitrile . . Summary of Initial Run Conditions for the Hydrolysis Reactions of MeZSiClz . Absorption and Concentration Values at Various Residence Times for the Hydrolysis Experi- ments of MeZSiCl2 . . . . Absorbance Data Obtained During the Condensation Reaction of PhSi(OH)3 vi Page 121 122 123 125 LIST OF FIGURES Figure l. Reactor Apparatus . 2. Reactor and Surrounding Heat Exchanger . 3. Constant Flow Infrared Cell and Cell Holder 4. Infrared Spectrum of PhSiCl3, Hydrolysis Reaction Products and (PhSiC12)O . . . . . . 5. Infrared Spectrum of PhSiCl3 6. Infrared Spectrum of PhSi(OH)3 7. Absorbance of PhSi(OH)3 Versus Residence Time 8. Beer's Law Application: the 620 cm.1 and 517 cm-1 Bands of PhSiCl3 . . . . 9. Spectra Obtained During the Hydrolysis Reactions of PhSiC13, Run 21. [HCl]o = 1.46 Moles/Liter 10. Spectra Obtained During the Hydrolysis Reactions of PhSiC13, Run 24. [HCl]o = 0.0 Moles/Liter . 11. Concentration of Silanes Versus Mean Reaction Time for the Hydrolysis of PhSiCl3, Run 20 12. Concentration of Silanes Versus Mean Reaction Time for the Hydrolysis of PhSiCl3, Run 21 13. Concentration of Silanes Versus Mean Reaction Time fbr the Hydrolysis of PhSiC13, Run 22 14. Concentration of Silanes Versus Mean Reaction Time for the Hydrolysis of PhSiC13, Run 24 15. Concentration of Silanes Versus Mean Reaction Time for the Hydrolysis of PhSiCl3, Run 26 16. Infrared Spectrum of 1,2-Dimethoxyethane vii Page 14 l6 17 24 25 28 32 34 36 37 41 42 43 44 45 49 Figure Page 17. Infrared Spectrum of Acetonitrile . . . . . . . 50 18. Infrared Spectrum Obtained During the Hydrolysis of PhSiCl3 in Acetonitrile . . . , . 51 19. Concentration of Silanes Versus Mean Reaction Time for the Hydrolysis of PhSiCl3, Run 17 . . . . . 55 20. Concentration of Silanes Versus Mean Reaction Time for the Hydrolysis of PhSiCl3, Run 18 . . . . . 56 21. Infrared Spectrum of MeZSiCl2 . . . . . . . . . 63 22. Infrared Spectrum of Me ZSiCl Its Hydrolysis and Common Condensation Producés . . . 65 23. Spectra Obtained During Run 44 . . . . . . . . 70 24. Concentration of Silanes Versus Mean Reaction Time for the Hydrolysis of MeZSiClz, Run 44 . . . . . 73 25. Concentration of Silanes Versus Mean Reaction Time for the Hydrolysis of MeZSiClZ, Run 45 . . . . . 74 26. Concentration of Silanes Versus Mean Reaction Time for the Hydrolysis of MeZSiClz, Run 46 . . . . . 75 27. Concentration of Silanes Versus Mean Reaction Time for the Hydrolysis of MeZSiClz, Run 37 . . . . . 76 28. Concentration of Silanes Versus Mean Reaction Time for the Hydrolysis of MeZSiClz, Run 38 . . . . . 77 29. Concentration of Silanes Versus Mean Reaction Time for the Hydrolysis of MeZSiC12, Run 42 . . . . . 82 30. Concentration of Silanes Versus Mean Reaction Time for the Hydrolysis of MeZSiClz, Run 49 . . . . . 83 31. 1/Temperature Versus Loge (Rate Constant) for the Hydrolysis Reactions ofMeZSiCl2 . . . . . . 86 32. A Cylindrical Element in a Laminar Flow Reactor . . . 93 33. Extent of Reaction for Laminar Flow With Heat Transfer Versus Extent of Reaction for Isothermal Plug . . . . . . . . . . . . . 100 viii Figure Page 34. Ratio of Rate Constants Versus Extent of Reaction for Zero, First, and Second Order Reactions . . . 102 35. Temperature Rise in a Laminar Flow Reactor With Heat Transfer to the Wall . . . . . . . . . 103 ix CHAPTER I INTRODUCTION A. General Chemistry of Silicon The hydrolysis reactions of halosilanes are faster than those of similar carbon compounds. Frequently, these reactions are too fast to be studied by conventional techniques and techniques designed for fast reactions must be employed to study them. Their complexity and fast rate are the primary reasons for the lack of study on the hydrolysis reactions of halosilanes. Silicon compounds are generally tetravalent. From the study of these compounds and orbital energy considerations, it is concluded that silicon like carbon makes use of sp3-o bonding (l: p. 3). However, due to the unfilled 3-d orbitals, silicon has the ability to expand its valence shell to include pentacovalent and hexacovalent silicon. Unlike carbon, the double oxygen bond in silicon is unstable (2: p. 1). Instead, silicon forms single bonds with oxygen and silicon-oxygen chains known as polyorgano- siloxanes. The silicon in polyorganosiloxanes may have one, two or three organic groups attached to it. The possible forms of this silicon unit in the polyorganosiloxanes are: 0 0 R R I I I I O - Si - 0 0 - Si - II 0 - Si - 0 0 - Si - R I I I O O R R Polyorganosiloxanes are normally producedeIMTmonomeric organosilicon compounds containing both organic and silicon func- tional groups. These silicon functional groups are hydrolyzable groups attached to the silicon atom. Polyorganosiloxanes are formed by the reaction of these monomeric organosilicon compounds with water. During this hydrolysis reaction, the functional groups are replaced by hydroxyl groups. These silanols further react by condensation accompanied by loss of water to form poly- organosiloxanes. The hydrolysis and condensation reactions of monomeric organosilicon compounds containing 1, 2, 3, or 4 func- tional groups are: Silane Unstable Intermediate Siloxane Unit R3SiX + R3Si(OH) + R3Si01/2 RZSiX2 + RZSi(0H)2 +»R28i0 RSiX3 + RSi(0H)3 +RSiO3/2 SiX4 + Si(OH)4 + SiO2 The silicon-halogen bond is extremely reactive and reac- tions involving this bond are generally fast. There are three principal reasons for the reactivity of the silicon-halogen bond. First, the silicon-halogen bond is strongly polar. Second, the radius of a silicon atom is 1.17 A compared to 0.77 A for a carbon atOm. Therefore, an attacking nucleophile would have easier access to the silicon atom in a silicon compound than to a carbon atom in a similar carbon compound. Third, since the 3-d orbitals of the silicon atom are unfilled, these orbitals may take part in the transition intermediate. In cases where a lower energy state for the intermediate occurs due to the incorporation of the 3-d orbitals, the silicon species would be more reactive. Industrially, the most important monomeric organosilicon compounds are those containing chlorine atoms as the functional groups. Through a variety of methods and conditions, these organochlorosilanes are hydrolyzed with water to form organosilanols which condense to form the various polyorganosiloxanes. Sommer (l: p. 78) studied the hydrolysis reactions of * optically active compounds R3Si C1. 0 E3593—+ (+) R Si*OH + HCl 3 'k (+) R351 C1 + H2 + 6.4° + 20.5° Such reactions were found to be 90 percent sterospecific and to proceed with inversion of configuration; for example, cis-chlorosilane gives a trans-silanol. From such reactions, it has been proposed that the hydrolysis reactions involving organo- chlorosilanes take place through backside attack by the water molecule. The transition intermediate for such reactions would involve a separation of charges. A typical intermediate for these reactions is: s[6+ H20 - - - s: - - - CT5'JS /' \ This intermediate would be stabilized by solvent molecules (S). ‘Since such intermediates contain a separation of charges, the more polar the solvent molecules, the greater their ability to sta- bilize the intermediates. B. Scope of Past Work The hydrolysis reactions of the halosilanes that have been studied are principally those of the monohalosilanes. These reac- tions are generally slower and less complex than those of the di- or trihalosilanes. Recent studies of those reactions include those of Milishkevick et a1. 1971 (3), Allen and Modena 1957 (4), and Chipperfield and Prince 1963 (5). Shaffer and Flanigen 1957 (6) investigated the combined hydrolysis-condensation reactions of alkyl and aryl chlorosilanes in batch experiments. The studies of Shaffer and Flanigen (6) and Chipperfield and Prince (5) both used electrolytic conductance to follow the hydrolysis of the chlorosilanes. Shaffer and Flanigen (6) slowed the reactions by saturating the solution with HCl and cooling the reactants. This allowed the reactions to be followed by conducto- metric titration. Chipperfield and Prince (5) designed a flow reactor and monitored the electrolytic conductivity as a function of reactor length. Both of these techniques follow the HCl con- centration, which allows observation of only the rate limiting hydrolysis or condensation step. Shaffer and Flanigen (6) expressed the rate of reaction with respect to the change in water concentration as: d(H20) dt = k (H20)m (ESiCl)n (IICI)p (2.1) The hydrolysis reactions were found to be first order with respect to water. Hence, the value of m in equation (2.1) is l. The value of n for the reactions of PhSiCl3 was found to be 1.9. How- ever, for the reactions of(CH3)ZSiC12, n was found to be -l/3. This decrease in initial rate of reaction with an increase in dimethyldichlorosilane was not understood. No corrections were made for any equilibriums in the system. This was mentioned as a possible source of error in the work of Shaffer and Flanigen. No value of p was given in the paper. However, it was observed that an increase in HCl concentration decreased the rate of the hydrolysis reactions. For the hydrolysis of PhSiCl3, a 10 percent increase in HCl concentration can double the half-time of the reaction. 1,2-dimethoxyethane saturated with HCl was used as the solvent for the majority of the hydrolysis experiments. However, the hydroly— sis reactionscHiCH3SiCl3, CGHSSiC13, and (CH3)ZSiC12 were also measured in a dioxane-HCl solution. The rates of reaction of CH3SiCl3 and (CH3)ZSiCl2 were found to be three times as fast in dioxane-HCl than in l,2-dimethoxyethane HCl. However, the hydrolysis of CGHSSiC13 was observed to be slower in dioxane-HCl. When titrating PhSiCl3 in an initially neutral solution at 0°C, Shaffer and Flanigen determined that only completely condensed polysiloxanes (RSiO3/2)x were formed. While at -78°C, (RSiC12)0 was found to be an intermediate product. Only in the case of (C6H5)ZSiCl2 were indications of stable groups such as =Si-OH Cl found. From their research, Shaffer and Flanigen reached the following generalizations. First, the rate of hydrolysis reac- tions of a series of silanes SiCl4 - R3SiCl are related in the following manner. SiCl4 > RSiCl3 > RZSiC12 > R3SiC1 Second, for a silane containing more than one chlorine, the rate of hydrolysis of the first chlorine atom is much faster than that of the remaining chlorines. Kleinhenz and Hawley 1970 (7) studied the hydrolysis reactions of PhSiCl3 at 0°C in an experimental flow system similar to the one used in this study. The reactions were followed using an infrared spectrophotometer. The hydrolysis reactions of PhSiCl3 were modeled as three irreversible reactions in series, first order with respect to water and silane, and second order overall. The rate equations used to describe these reactions are: d[PhSiCl3] . dt = -k] [PhSTC13] [H20] (2.2) d[PhSiC12(OH)] . dt = R1 [PhSTC13] [H20] - k2 [PhSiC12(OH)] [H20] (2.3) d[PhSiCl(OH)2] dt = k2 [PhSiC12(OH)] [H20] - k3 [PhSiCl(0H)2] [H20] (2.4) [PhSi(OH)3] = [PhSiCl3]o + [PhSiC12(OH)]0 + [PhSiCl(OH)2]o + [PhSi(OH)3]O - [PhSiC13] - [PhSiC12(0H)] - [PhSiC1(OH)2] (2 5) Equation (2.5) is a material balance on the system with [PhSiC13]O, [PhSiC12(OH)]O, and [PhSiCl(0H)2]0representing the initial concentrations of PhSiCl3, PhSiC12(0H), and PhSiC1(0H)2, respectively. To fit the rate constants, the reactions were modeled for both plug and laminar flow. In laminar flow the velocity distri- . bution is: U = 2 Ub [1 - (r/R)2) (2.5) The residence time for each element in the cylindrical cross section is: t = L/U (2.7) The velocity distribution was used as a weighing factor to obtain the bulk concentrations by integrating the point concentrations over the reactor cross section. R 2n I C(L,r/R,Ub) U(r/R,Ub)r dr 0 c = (208) b A Ub R2 Using this analysis, rate constants, which give the same conversions for either plug or laminar flow, were determined for different order reactions. For second order irreversible reactions, the average ratio of the rate constant determined assuming laminar flow to that determined assuming plug flow is 1.176. The rate constants for equations (2.2) through (2.4) were determined by bracketing the data points between a model for laminar flow and a model for plug flow. Using this method, the rate constants were determined to be k1 = 1500.0, k2 = 77.5, k3 = 1000.0 liters/mole- seconds. All hydrolysis experiments were run with no initial HCl concentration and no attempt was made to determine the effect of HCl on the system. C. Scope of This Work The literature available on the hydrolysis reactions of halosilanes deals almost exclusively with the monohalosilanes. Though the reactions of the monohalosilanes are of little indus- trial importance, they have been studied to a considerably greater extent than those of the di- or trihalosilanes. This occurred since monohalosilanes reactions tend to be slower and less complex than those of the di- or trihalosilanes. Their hydrolysis reactions can be followed through either an increase in the HCl concentration or a decrease in water concentration. In order to adequately determine the kinetics of such reactions, it is necessary to determine the individual reactants and product concentrations during the reactions. A technique has been devel- oped in this research which allows the determination of the con- centrations of the initial chlorosilanes, unstable intermediates, and products of the hydrolysis reactions to be followed during these reactions. When Kleinhenz and Hawley (7) studied the hydrolysis reac- tions of PhSiC13, no effort was made to determine the effect of HCl on these reactions. Shaffer and Flanigen (6) observed a marked decrease in the rate of reaction with an increase in the HCl con- centration. They gave the decrease of the rate of the combined hydrolysis and condensation reactions of PhSiCl3 as an example of this effect. Here, a definite decrease in the rate of reaction was observed with an increase in the HCl concentration. To mini- mize this effect, Shaffer and Flanigen studied the reactions in a saturated solution of HCl. One of the objects of this present study was to determine the effect of HCl on the hydrolysis reac- tions. The hydrolysis reactions of PhSiCl3 and MeZSiC12 were studied during this research. These compounds and their hydroly- sis reactions are industrially important in the manufacture of polyorganosiloxanes. A laminar flow tubular reactor coupled with an infrared spectrophotometer was utilized to follow the hydroly- sis reactions. Infrared spectra of the reactants, unstable intermediates, and products were obtained during the experiments. Peak assignments were made for the various compounds which are 10 present during the reactions and extinction coefficients determined for these compounds. Using these spectra, concentration data versus time in the reactor were obtained. The experimental apparatus was modified to handle corrosive solutions of HCl. This enabled the initial concentration of HCl to be varied and the effect of HCl on the hydrolysis reactions to be determined. In this study models for the hydrolysis reactions of PhSiCl3 and MeZSiCl2 are presented. These models are based on an SnZ-Si type reaction and describe the individual hydrolysis reactions. Sn2-Si reactions are nucleophilic substitution reactions of silicon having a transition state intermediate containing a separation of charges. Rate constants are determined from experimental data for the indi- vidual hydrolysis reactions. With the exception of the worinleein- henz and Hawley (7), previous work on these reactions has only described the rateSnythe combined hydrolysis-condensation reactions. The temperature effect on the hydrolysis reactions of MeZSiClz is described. The effect of HCl on the hydrolysis reactions is clarified by describing the conditions under which HCl catalyzes or suppresses the hydrolysis reactions of the chlorosilanes. By describing the individual rates of the hydrolysis reactions, the temperature effect, and the effect of HCl, the knowledge of these hydrolysis reactions is significantly increased. Finally, a model of a laminar flow reactor is presented. This model may be used to determine under what conditions the plug flow and isothermal assump- tions are justified when investigating reaction rates in a flow reactor. 11 0. Organization of This Work The body of this work is contained in five chapters. Chapter II describes the experimental apparatus and techniques employed to study the hydrolysis reactions. Chapter III describes the hydrolysis reactions of PhSiCl3. This chapter begins with the interpretation of the infrared spec- tra obtained during the hydrolysis experiments. The competing condensation reaction is then described. The conditions for the hydrolysis reactions of PhSiCl3 are discussed and a model for these reactions presented. The solvent effects are then investi- gated. The chapter ends with some concluding remarks on these hydrolysis reactions. Chapter IV describes the hydrolysis reactions of MeZSiCTZ. This chapter begins with an interpretation of the infrared spectra obtained during the hydrolysis reactions of MeZSiClz. The com- peting condensation reactions for this system are then discussed along with any possible effect of the condensation reactions on this study of the hydrolysis reactions. The experimental condi- tions for these hydrolysis reactions are described and a model developed. The temperature effect on these reactions is then dis- cussed. This chapter ends with concluding remarks on the hydrolysis reactions of MeZSiClZ. Chapter V contains a development of a laminar flow reactor model with heat transfer. This chapter begins with a discussion of past work that has been done on this type of model. The basis 12 for the model is introduced and the model developed. The chapter ends with a discussion of conclusions reached through this analysis. Chapter VI contains a general discussion of conclusions reached during this study. Appendix A contains the data from the laminar flow reactor used to model the reactions. Appendix 8 contains the development of the computer program used to solve the partial differential equations describing the temperature and concentration profiles in the reactor. CHAPTER II EXPERIMENTAL A. Apparatus Figure 1 illustrates the apparatus used for the hydrolysis experiments. The reactants-solvent mixtures were contained in two polyethylene tanks. The driving force for the system was a tank of dry nitrogen. The line pressure from the nitrogen tank was maintained at approximately 10 p.s.i.g. during the runs. This pressure forced the solutions from the tanks through the rotame- ters. The rotameters were calibrated and these calibrations checked after each run. After the reactants-solvent mixtures left the rotameters, they flowed through the temperature bath and then into the reactor. Polyethylene tubing (1/4 inch 0.0.) was used for the flow lines to the rotameters, from the rotameters to the constant temperature bath, and from the constant temperature bath to the reactor. The polyethylene tanks and tubing, being resistive to the corrosive nature of chlorosilanes and hydrogen chloride, were found to be ideal for handling solutions of these reagents. The heat exchanger consisted of a series of coiled stainless steel tubing, which insured good heat transfer. Since the contact time between the reactants and stainless steel was very short, only a minimum of corrosive reaction occurred here. 13 :oU .m._ 2. .9095 o... 14 .mzumgmaa< LouummmII.F mczmwd 50m 9.293th .ccficou ®I Q83; {{(rxglj 53001 2205202 .— 53 .95 .203 "0.2.. T _ 2 05365 $5052 15 The polyethylene tubing proved a good insulating material. Besides being constructed from polyethylene, the tubes from the heat exchanger to the reactor were wrapped in 1/2 inch of insula- tion, which kept the temperature rise during this portion of the reactants' journey to a minimum. The reactor is shown in Figure 2. It consisted of a stainless steel tube, 0.216 cm in diameter, surrounded by a heat exchanger. Water was pumped from the constant temperature bath through the heat exchanger. This aided in maintaining a constant temperature during the reaction. The silane-solvent mixture entered the reactor at a side port near its base, and the water-solvent mixture entered the reactor through a needle. This needle is similar toga hypodermic W needle and has three small holes at 120° from each other at its) __.__—_._—- end. As the water-solvent mixture was forced through these holes, -_.-._.__.....7 N“ A.“ "u—-— -4.— it mixed with silane-solvent_mixture. Thiswpointw 3 defined as W”...- H”. “An—.— m w“- the beginfljflg 0f EPP,Y939PIQDPZQPG- Based on the stability of the infrared spectra,it waswfpund_thatfithe holes in the hypodermic needle are small enoughfito insure good mixinggjn the react9r_and _._.._’-_____.__——- 7 prevent the development of streamline regions of high concentra- tion, which flow through the reactor without mixing. av .—— u..-“ From the reactor, the reactants flowed into the infrared cell, Figure 3. Considerable turbulence occurred where the reactor makes a 60° turn when entering the infrared cell and again at a 90° turn where the reactants flowed up between the salt plates. It was necessary to place a gasket between the reactor and salt plates 16 ._._mU a.— .mecmcuxu pom: mcwucaogcam new Louumwmu-.~ mesmwd 3.0)) mczooU F .9; Eu SN. _ "IL .23 ".22.. T IutU W0 17 E00m .m._ .cmu—o: —pmo new ppmu 60.8.»:H zopm acmumcouuu.m mgamwm . . . Lez. 0.7.00 co=0h ‘I. .0003 100.. 33. :3 so: :8 m Immx n - may. 00:0.Em of .2... O I * \l \l 0__u002 95.031? I. .. .2000. a .95 ".05... "”913' "H 18 to prevent any reactants from leaking at this junction. The salt plates were made of thallium bromide-iodine (KRS-S). These plates allowed observation of the low infrared region characteristic of the silane-chlorine vibrations (650 cm-1 to 400 cm'l). The reactants entered the infrared cell at its base and left at the top. From here, they flowed into a con- tainer for disposal. The spectrophotometer employed for this study was a dual beam, Perkin-Elmer 337 model. The dual beam arrangement com- pensated for the solvent absorption. 8. Temperature Control The reactants were cooled to 0°C using a constant temperature bath containing distilled water and ice. To main- tain temperatures above 0°C, an on-off type heater and cooling coil were employed. Tap water at approximately 16°C was constantly circulated through the coil. A centrifugal pump was employed to maintain constant circulation of water in the bath. To maintain temperatures below 0°C, a methanol-water bath was employed. Here, methanol was added to the water to lower the freezing point to the desired temperature. Dry ice was then added to the mixture until a liquid-solid equi- librium occurred. Using this method it was possible to main- tain the temperature to within 0.5°C of the desired temperature during each experiment. 19 C. Reagents The solvent used in the majority of the experiments was l,2-dimethoxyethane (Ansul Ether 121*). This solvent has the necessary characteristics of being a solvent for water, halo- silanes, and silanols, and of not absorbing strongly in the lower infrared region. The ether was dried over CaSO4 and distilled from KOH in a nitrogen atmosphere. Only the middle fraction of the distillate, that of a constant boiling point, was used. The dis- tillation process removed inhibitors found in the ether. A procedure was developed to recycle the ether after it had been used. The used ether was first titrated with aniline to remove the hydrogen chloride. It was then filtered to remove the aniline hydrochloride. Next it was distilled from KOH and then stored with CaSO4, to further dry the ether. The ether was dis- tilled a second time before use. An infrared spectrum of the ether showed that no silanes remained in the recycled ether. Matheson gaseous HCl was used to prepare the HCl ether solutions for the runs requiring initial concentrations of HCl. The solutions were prepared by passing the gaseous HCl through the ether. The concentration of HCl was then determined by electro- lytically titrating the HCl ether solutions with a standard solution of NaOH. To prepare the silane solutions, the silane was distilled in a nitrogen atmosphere directly into the ether. This procedure avoided any reaction between the silane and atmospheric moisture. *Ansul Company, Marinette, Wisconsin. 20 The silane solvent mixture was then weighed to calculate the con- centration of silane. 0. Experimental Technique Before each experiment, the experimental apparatus was flushed with acetone and then dried with nitrogen. By first evacuating the tank and then drawing the reagents in under this vacuum, it was possible to avoid any contact with atmospheric moisture. I At the beginning of each run, an initial spectrum of the silane was obtained before any water had been added to the system. This spectrum allowed the initial concentrations of the reactants to be calculated. During the runs, the flow rates of both the water-solvent and silane-solvent mixtures were monitored using rotameters. The initial concentrations of the reactants were corrected for the dilution due to the mixing of the streams. To obtain concentration data versus residence time in the reactor, the hypodermic needle was pushed the maximum distance 1 to 400 cm") scanned. into the reactor and the spectrum (650 cm- The hypodermic needle was calibrated, by centimeters, such that its distance in the reactor could be read at the needle input to the reactor. Using the flow rates from the rotameters, the reactor area, and reaction zone, the residence time could be calculated. The needle was then moved out to a desired distance increasing the residence time and again the spectrum was recorded. Using this technique, a series of concentrations versus residence times was obtained. CHAPTER III THE HYDROLYSIS REACTIONS OF PHENYLTRICHLOROSILANE The experimental techniques described in Chapter II were used to study the hydrolysis reactions of phenyltrichlorosilane. The object of this study was to obtain experimental data on the hydrolysis reactions, model the reactions at 0°C, determine the rate constants for this model, and determine the effects of hydro- gen chloride on these reactions. To obtain experimental data on these reactions, the posi- tion of the hypodermic needle through which the water-solvent mixture entered the reactor was varied. A steady state of con- centration of reactants occurred along the length of the reactor. Therefore, by varying the length of the reaction zone, the resi- dence of the reaction was also varied. The hydrolysis reactions of phenyltrichlorosilane were studied in three parts. First, these reactions were studied at 0°C with varying concentrations of HCl. Here, the initial con- centration of HCl was varied from 0.330 to 1.46 moles/liter. The concentration of silanes present was varied for 0.073 to 0.095 moles/liter with enough water initially present (0.215 to 0.291 moles/liter) to ensure complete reaction. Next the hydroly- sis reactions were studied with no HCl initially present at 0°C. 21 22 The amount of water initially present varied from 0.259 to 0.0807 moles/liter. At an initial water concentration of 0.259 moles/ liter, the chlorosilanes were completely converted to silanols. However, with an initial water concentration of 0.0807 moles/ liter, the reaction stopped with mainly unstable intermediates present. This allowed the reaction to be studied when complete hydrolysis did not occur. A total of five runs with approximately nine data points per run at times from 0.066 to 2.0 seconds were used to model the hydrolysis reactions of phenyltrichlorosilane in l,2-dimethoxyethane. ‘To determine the effect of solvent on the rate of the hydrolysis reactions, two hydrolysis runs were made using acetoni- trile as the solvent. Approximately 0.07 moles/liter of silane and an excess of water were used for these runs. Approximately 18 data points were obtained, which allowed the rate of these reac- tions in acetonitrile to be compared with the rate in 1,2- dimethoxyethane. The data collected during the experiments consisted of the infrared absorption of several peaks versus residence time in the reactor. In order to utilize this data, it was necessary to assign these peaks to the various reactants, unstable intermediates and products of the hydrolysis reactions. 23 A. Interpretation of the Infrared Spectra Obtained'During the Hydrolysis Reactions of Phenyltrichlorosilane The assignments which were made for the absorption peaks of PhSiCl3 and its hydrolysis products are listed in Table 1. TABLE l.--Infrared Assignments for PhSiCl3 and Its Hydrolysis Products. PhSiCl3 . . . . . . . . . 590 cm" and 518 cm“ PhSiC12(OH) . . . . . . . . 572 cm“ and 527 cm“ PhSiCl(OH)2 . . . . . . . . 542 cm“ PhSi(OH)3 . . . . . . . . 485 cm" and 465 cm" The infrared spectra characteristics of PhSiCl3, the hydroly- sis reaction products of PhSiCl3 and (PhSiC12)20 are shown in Figure 4. These spectra along with spectra of the hydrolysis experiments were used to deduce the species present during the hydrolysis experiments. Figure 5 shows the infrared spectrum of PhSiCl3 obtained during the hydrolysis experiments. H. Kriegsmann and K. H. Schowlka (8) conducted an extensive study of the various si- lanes and assigned the 590 cm'1 peak of the PhSiCl3 spectrum to the asymmetric vibration of SiCl3 and the 518 cm'I peak to its symmet- ric vibration. The 620 cm'1 peak was assigned to the Ph-Si vibra- tion. Smith (9) also made these assignments for PSiCl3 and noted 1 peak has a shoulder at approximately 583 cm']. that the 590 cm' Preliminary hydrolysis experiments with no initial HCl present and with excess water produced spectra similar to that of 24 2.555: m I new muuavogm :o_uummm mwmapoguaz . Fuwmgm mo E:.uumam um.m.wcuuu.e 0.3mm; .720. 5230.... can _ 900 I can . m_ e.._w_ 2_ w_ w_ W_ ”0.2.. m a .s m m n$93.... .s_ m_m_ m_ s._ m_ m_ 0.392... m m .03.: .83 be, 2.. :0 xom: 3 Mum”... M .2 gal _ «few... 25 (z) aoueaatwsueu) .. 81 cow ow 1 co 1 om 1 com .mpuwmga we E=.pumam um.m.$:~II.m 0.:mwd AFIEUV aucwzcwgu com com com a 1 q 26 1 PhSi(OH)3 over the range of from 400 cm" to 740 cm"1 at long reactor lengths. At shorter reactor lengths, intermediate bands 1 1 appeared at 572 cm” , 542 cm", and 527 cm' . The 572 cm" and 527 cm'1 peaks appeared immediately after the water was added to the system, were well correlated, and decreased as the reactor 1 peaks were length was increased. The 572 cm"1 and 527 cm' present and similar in appearance during all runs. For PhZSiClz, the symmetric and asymmetric peaks of =SiCl2 occur at 540 cm'1 and 1 572 cm' , respectively. Therefore, the 572 cm'1 peak was assigned to the asymmetric vibration of SiCl2 in PhSiC12(OH) and the 527 cm.1 peak was assigned to the symmetric vibration. At all times the 542 cm"1 peak was quite small and somewhat obscure due to the absorption of a solvent peak at 547 cm']. Therefore, the quanti- tative interpretation of the peak was difficult. C1 C1 I I The absorption spectra characteristic of Ph-Si-O-Si-Ph C1 C1 was not seen during the experiment. Also, a separate experiment showed that the condensation reaction was considerably slower than the hydrolysis reaction at low concentrations of HCl. B. Preparation of PhSi(OH)3 The infrared spectrum of phenylsilanetriol from 4000 cm"1 to 650 cm'1 has been published (10). However, the spectrum 1 was not available in the literature. between 650 cm.1 and 400 cm" Therefore, to confirm that phenylsilanetriol was the final product from the hydrolysis reaction of phenyltrichlorosilane and to study 27 its condensation reaction, it was necessary to prepare phenyl- silanetriol and obtain its spectrum. I Phenylsilanetriol has been prepared by several authors. Takiguchi (11) prepared phenylsilanetriol by direct hydrolysis of phenyltrichlorosilane using aniline as an hydrogen chloride accepter. This procedure was somewhat complex and involved the evaporation of a considerable amount of solvent. Tyler (12) prepared phenylsilanetriol through the hydroly- sis of phenyltrimethmoxysilane at 10°C using acetic acid as a catalyst. On standing, condensation gradually occurred in some of Tyler's samples, while others were kept for one year with no signs of decomposition. Phenylsilanetriol was prepared using a method similar to that of Tyler's. In a three-neck, one-liter flask, 198 g (1.0 mole) of phenyltrimethmoxysilane were added to 108 g (6.0 moles) of water. No acetic acid was added in orderto minimize the amount of condensation occurring. After constant agitation at 16°C for four hours, shiny, white, flat crystals precipitated. After six hours of constant agitation, the crystals were filtered and dried overnight. The product weighed 75 g (0.49 moles), which gives approximately a 50 percent yield. Figure 6 shows the spectrum’ 1 l to 400 cm' . The phenyl- 1 of phenylsilanetriol from 800 cm' silanetriol spectrum has two peaks at 745 cm'1 and 705 cm' which compare with those of the published spectrum plus peaks at 485 cm'1 and 465 cm']. 28 (z) aouenilwsueul .mfixov.me. .0 e=.eueem ee.e..e.--.o 5.3m.. apieuv Xucmzcm.d o ooe . omm . owe . 3w. . owe ON H“ e“ em om fl . .rx. e cop 29 C. Condensation Reactions To determine the effect of the condensation reaction, the reaction between PhSiCl3 and PhSi(OH)3 was studied. Kleinhenz and Hawley (7) reported that no reaction occurred when PhSiCl3 and PhSi(OH)3 were mixed in the reactor at 0°C. Here a 14 weight percent mixture of PhSi(OH)3 in ether and a 4.95 weight percent mixture of PhSiCl3 in ether were mixed in the reactor. No signifi- cant changes in infrared absorption of either PhSi(OH)3 or PhSiCl3 were observed with changes in the reaction zone. This indicated that with no hydrogen chloride present the condensation reaction is very slow. 7 During the hydrolysis reaction, HCl is produced. It has also been observed that HCl catalyzes the hydrolysis reaction through the following mechanism (13): First, the hydrogen ion joins the silanol oxygen. + (1) H+ + Si-OH e———-—+ Si-OII2 The second reaction consists of a nucleophilic attack by the oxygen on one silanol on the silicon of the silanol molecule in the oxonium form. + 2-—> Si-O-Si + II+ + H (2) SiOH + Si-OH O 2 The first reaction is quite rapid, while the second occurs more slowly and is the rate controlling step. To determine the effect of HCl on the condensation reac- tion, PhSiCl3 and PhSi(OH)3 were mixed in the reactor with an 30 initial HCl concentration of 0.390 moles/liter. The conditions for this reaction are given in Table 2. TABLE 2.--Conditions for Condensation Reaction of PhSi(OH)3. ————_ _———— Temperature . . . . . . . . . 0°C Flow Rates . . . . . . . . . 33.9 ml/min Initial Concentrations . . - . . Moles/Liter PhSi(OH)3 . . . . . . . . . 0.073 PhSiCl3 . . . . . . . . . . 0.114 PhSiC12(OH) . . . . . . . . 0.056 PhSiCl(OH)2 . . . . . . . . 0.0 Total Silanes . . . . . . . . 0.243 HCl . . . . . . . . . . . 0.390 In the hydrolysis reactions with no initial concentration of HCl, the concentration of HCl reaches approximately 0.3 moles/ liter at the end of the reaction. At this time, the trichlorosilane and dichlorosilane have reacted leaving principally phenylsilane- triol, approximately 0.1 moles/liter. Hence, the conditions of the condensation reaction study with HCl initially present were considerably harsher than those encountered during the hydrolysis experiments with no initial concentration of HCl. A reaction was detected during this condensation study by a decrease in the PhSiC13 and PhSi(OH)3 and an increase in the PhSiC12(OH) peak. Some of the possible reactions occurring here are: 31 (1) Ph-Si-DH + Ph-Si-OH'--» Ph-Si-O-Si-Ph + H20 (2) Ph-Si-OH + Ph-Si-C1-- » Ph-Si-O-Si-Ph + HCl (3) PHSiCl3 + H20 -—-——+-PhSiC12(OH) + HCl The decrease in the PhSiCl3 concentration and increase in PhSiC12(OH) may be explained by [1] reaction (1) followed by reac- tions (2) and (3), or [2] an unknown amount of H20 in the PhSi(OH)3 stream. The decrease in the PhSi(OH)3 peaks can only be explained through a condensation reaction. Either reactions (1) or (2) are ' occurring, or both reactions are occurring simultaneously. Hence, the rate of the condensation reaction may best be determined through the decrease in PhSi(OH)3. The concentrations of the various silanes versus residence times are given in Appendix A-7. Figure 7 is a plot of absorption of PhSi(OH)3 versus time during the condensation reaction. The mean time on the figures refers to the residence time in the reactor. The half life, the time required for one-half of the initial concentration to react, of PhSi(OH)3 is approximately 0.5 sec. In run 24 the half life of PhSiCl3 is approximately 0.0015 sec. and that of PhSiC12(OH) is 0.1 sec. Therefore, for the hydrolysis reactions with an initial HCl concentration greater than 1.0 moles/liter, the condensation effect can be seen at long reaction times where the amount of PhSi(OH)3 present falls below that predicted by the model. 2.0 1.8 1.6 1.4 Absorbance :3 ie on 32 Legend 0 PhSi(OH)3 .2. Initial Concentrations 0.390 Initial Moles MHCl = Liter of Solution _ Initial Moles MPhSiCl3 ' 0'114 Liter of Solution M . _ I Initial Moles Ph$1(OMB ' 0°073 Liter of Solution O l l 3.4 .'5 .15 .‘7 .85'9 1:0 1 {,2 1,4 t = Mean time; sec. Figure 7.--Absorbance of PhSi(OH)3 Versus Residence Time. 33 0. Calculation of Concentrations Kleinhenz and Hawley'(7)experimentally determined, using infrared spectrophotometry, that the difference of the absorption of PhSiCl3, A, and the corresponding base-line absorption, A0, was proportional to the product of the concentration,C,ofPhSiCl3 in the ether solution and the infrared cell path length, d (i.e. , the Beer's law relationship is applica- 1 ble). Figure 8 illustrates Beer's Law application to the 620 cm' and 517 cm"1 peaks of PhSiCl3 for concentrations of PhSiCl3 below 0.1 moles/ liter. A0 "' A = EdC (3.1) It was assumed that the same relations were valid for each reaction species being monitored, Here, 5 is the extinction coefficient or pro- portionality constant. The base-line technique was used to measure the absorption of 1 peak of each peak. The extinction coefficient, 6, for the 465 cm- PhSi(OH)3 was found by reacting PhSiCl3 with an excess of water at long reaction times. The 518 cm'1 peak was used to determine the concentra- tion of PhSiCl3 and the 527 cm" peak was used for PhSiC12(0H). The extinction coefficients for these peaks were determined using a least squares fit for absorption data obtained at different reaction times when reacting 1 mole of PhSiCl3 with 1/2 mole of water. The extinction coefficients obtained are: TABLE 3.--Extinction Coefficients for PhSiCl3andIts Hydrolysis Products. 446 (chm)"1 390 (chm)'1 237 (chm)'1 8517 e527 e455 Absorbance 34 .9- / JB‘ Beer's Lab I .7" / / / .6- ’ I -l ,5- 517 cm I 620 C111;l O .4" o .3-1 I .2“ 1" 0° '0 l I l ' l I l .0 .1 .2 .3 .4 .5 .6 Concentration of PhSiCl3 (Moles/Liter) Figure 8.--Beer's Law Application: the 620.cm'1 and 517 cm- Bands of PhSiC13. / 35 E. Conditions for the Hydrolysis Reactions of PhSiCl3 The reagents were prepared and drawn into the holding tanks as described earlier in Chapter II. The initial concentrations of the silanes were determined from the reference spectrum and are shown in Table A-1 along with the initial water and hydrogen chloride concentrations. Since a major object of this investi- gation was to study the HCl effect, the initial HCl concentration was varied from 0.0 to 1.46 moles per liter for different runs. During each run approximately 0.3 moles per liter of HCl were produced. The total silane present was varied from 0.072 to 0.095 moles per liter and the initial H20 concentration from 0.0807 to 0.291 moles per liter for the various runs. At the higher water concentrations, complete hydrolysis of PhSiCl3 occurred. This allowed the final hydrolysis products of PhSiCl3 to be studied. At lower water concentrations, unstable intermedi- ate products occurred and reached a steady state. This allowed the assignments for these products to be made. Some of the spectra obtained during two different hydroly- sis experiments, runs 21 and 24, are shown in Figures 9 and 10. These two experiments illustrate the hydrolysis reactions in media where no HCl is initially present and where a high concentration of HCl is initially present. Run 24 contained no initial concen- tration of HCl, while run 21 contained 1.46 moles per liter of HCl initially. At the end of run 24, the spectrum closely approximates that obtained for PhSi(OH)3. 36 aouenthsueul ) (2 o ..ee.4\me_ez me.. I HFQIH ._N eem .m_u_me. 0o mcowpummm mmmx_o.cxz mg» m:..=a nmcwmpoo m.uumamII..wm.:mwu AFIEuv.aucm=cm.m 004 some see ome .ooe omm “7.44.3 ..d.J 1‘4.J o~ I I oe I I ‘oe cm. 1' I .eem .o I e I .umm mme.o I . oo.r mm. I I O q-WuJ ON I oe I r. on em I I .55.. eeeee.memv I m .eem mm..o I . r .eem oeo.c I e . muzmmu.um I . I .N :31 cop 37 (z) aoueqatmsueal .Lmungmmpozo mo mcoepummm mpmxpogvm: ems» ooe ome O quqJ cm I oe I om I o» I.eem mom.. I . oo_r O . .1-J cm I oe I cm I cm I I.eem mm..o oopI AFIEUV aucmzcmcm. cow omo -d_I_ .owm mmv.m u h Auswh m a may .umm mmo. o I h em_u=H .em cam .m_u.me. 5.59 3538 95831.2 0.53... wees mm_.o I .. UTT—r muzmm..mm em cam 38 To determine the cell length, d, the empty sample cell transmittance was measured and a number of interference bands obtained. Smith (14) reported that the cell length, d, is related to these bands by: d = n/(2(u] - 112)) (3.2) Here n is the number of maxima between the wave numbers n} and “2' Due to the high concentration of HCl, the rate of the hydrolysis reaction in run 21 is considerably slower than that of 1 and 465 cm"1 are smaller and run 24. Also, the peaks at 485 cm' less distinct than those in run 24. Some condensation does take place in run 21 and this effect can be seen in Figure 12 (on page 42), where the amount of silanes present (points) falls below that predicted by the model (solid lines). F. Results and Discussions The following reaction steps were used to model the hydrolysis of PhSiCl3. K (1) PIIsI'CT3 + H20 —'--> PhSiC12(OH) + HCl K (2) PhSiC12(OH) + Hzo-——2—+ PhSiC1(OH)2 + HCl K (3) I>IIsI'CT(OII)2 + H20—-—3-+ PhSi(OH)3 + HCl were found to be essentially irreversible. 39 Under the conditions used in this study, these reactions It was assumed that the reactions were first order with respect to H20 and the silanes. The effect of HCl was incorporated into the rate expressions using the term [HCl]". The rate expressions used to describe these reac- tions are: d [PhSiC13] dt d [PhSiC12(OH)] = -I<1 [PhSiCl3] [H20] [HCl]n dt d [PhSiC1(0H)2] = R1 [PhSiCl3] [H20] [HCT]" - K2 [PhSiC12(OH)] [H20] [HC1]" Eff d [PhSi(OH)3] K2 [PhSiC12(0H)] [H20] [HClJn - K3 [PhSiCl(0H)] [H20] [HCI]n dt = K3 [PhSiCl(OH)2] [H20] [HCT]" [H20] = [H20]O - [PhSiC12(OH)] - 2[PhSiCl(0H)2] - 3[PhSi(OH)3] [HCl] = [HOT]o + [PhSiC12(OH)] + 2[PhSiCl(OH)2] + 3 [PhSi(OH)3] (3.3) (3.4) (3.5) (3.6) (3.7) (3.8) 40 Here, [HCl]o and [H20]o are the initial concentrations of HCl and H20, respectively. Increasing the initial HCl concentration had the effect of decreasing the rate of the hydrolysis reaction. Using an unweighted squares type curve fitting program (14), the constants listed in Table 4 were determined to fit the various data sets. Figures 11 through 15 are plots of the data points and model (solid lines) ver- sus mean residence time for the various runs. The solid lines through the data points represent twice the estimated standard deviation of the data. The kinetic parameters were determined assuming that the reactor was isothermal and that plug flow existed in the reactor. The effect of the plug flow assumption has been studied by Klein- henz and Hawley (7) for reaction orders between 0 and 3. Their analysis employed an integration of the concentration-velocity product for each reaction over the flow cross-sectional area. They concluded that for first and second order reactions, the rate TABLE 4.--Parameters Used to Model the Hydrolysis Reactions of PhSiC13. Linear Estimate of Parameter Standard Deviation K - 220 (l' 'T" 121 1 — . Tters/mole) /sec. . K - 155(1' 1”" 105 2 . iters/mole) /sec. . _ . 1+n K3 - 40.6 (liters/mole) /sec. 4.90 -0.612 0.0809 3 ll 41 Initial Concentrations Legend R". 3?: I Silane 0 PhSiC'fiOHI M11 5 W o PhSi(OH), Mug, 1.21 Wis-9M .09' .08 e I =1 ill/L $.06 ii 3.05 2 294 PhSi(OH), E T ,g1x3* a‘IlI' I 8 e QI- cur Figure ll.--Concentration of Silanes Versus Mean Reaction Time for the Hydrolysis of PhSiCl3, Run 20. 42 Initial Concentrations Figure 12.--Concentration of Silanes Versus Mean Reaction Time for the Hydrolysis of PhSiCl3, Run 21. M R012! Moles I s. o Phsacuom “TE .9992 WM 0 PhSi(OH); 3.1463W .08- .2 ‘1 gm- ° goo ° 0 § ms: each) ‘5-05 ° ° PhSi(OH .§ 0 2'“ o g s u.03- .02- RIISICI, o o Phsacuou, .01- 0 I 2 .3 A .s 3 '7 5 5; -Moon Time sec 43 Initial Concentrations Legend Run 22 0 Ph s:cu,(ou_) M91323: M ' 0 Ph Si(OH), M1533 W 09" .08+ .3 _, o \ $.07» 306. PhSiCl,(OH) O o I ‘3 3.05: a PhSi(OH), .5 ' 1304 E ° ° g /"‘SiC'3 O u.03 ° 0 o O .02 ' o Phsac1(ou), tn- 0 .l .2 3 A» 5 .6. .} .6 7- Mean Time; see Figure l£L--Concentration of Silanes Versus Mean Reaction Time for the Hydrolysis of PhSiCl3, Run 22. 44 Initial Concentrations 7 Legend 0 PhSiCl2(OH) 0 Ph Si (on)3 .09r E08“ 3'07“ 0 PhSi on), I 0 306 :7. . 3.05- a .g gm- o § 593- . O .02 - PhSiCl,(Ol-l) .ov ° PhSiCl (OH 2 - PhSiCla ° 0 1 J l 1 1 1 I it: 0 J .2 ._ .3 A .5 .6» .7' .8 t-Afloan Thus; on: Figure l4.--Concentration of Silanes Versus Mean Reaction Time for the Hydrolysis of PhSiCl3, Run 24. 45 Initial Concentrations Legend 3323802 m Silane o PhSiCl2(OH) M20000 W o PhSi(OH); Mflcfi’ .000 W 09~ as. C) .1: ° 0 '4 o §.07- 0 ° , . g o Phsucugou) _ £06" 2 i3 395 .g 5404 5 $03 .02 PhSiCI(OH)2 9" a a PhSi(OH): #— PhSiCl: D 1 1 r l l 1 1 _1 0 .1 .2 3 A .5 .0 .7 .0 . 7-4Moan Than; see Figure l5.--Concentration of Silanes Versus Mean Reaction Time for the Hydrolysis of PhSiCl3, Run 26. 46 constants determined assuming a plug flow reactor are approxi- mately 16 percent lower than the actual rate constants in a laminar flow reactor. This produces a l6 percent bound between two extremes, that of laminar flow and that of plug flow. Chapter V of this study analyzes both the plug flow and isothermal assumptions. These two assumptions tend to counteract each other. The plug flow assumption produces a lower rate Con- stant, while the isothermal assumption produces a higher rate con- stant. For a second order reaction with thermodynamic and rate constants similar to those found in the hydrolysis reactions, the rate constants determined using both assumptions are approximately l0 percent lower than the rate constants in a laminar flow reactor with heat transfer. In the actual system neither of the models completely describes the situation. Although the Reynold's number is well within the laminar flow region, there is considerable mixing in the reactor where the reactants are combined and again where the reactor effluent enters the infrared cell. Also, there is a cer- tain amount of mixing in the system due to radial diffusion. Therefore. the kinetic parameters determined with the plug flow isothermal assumptions are within this l0 percent bound. The rate of hydrolysis of the first chlorine atom in PhSiCl3 is extremely fast and near the limit of the system to determine. Therefore, this first rate constant K1 should only be considered an estimate of its actual value. The parameters 47 K2, K3 and n are well defined based on the fit of the data and the linear estimates of the standard deviations. G. Solvent Effect To determine the solvent effect on the hydrolysis reactions of PhSiCl3, the reactions were studied in acetonitrile. Acetoni- trile has the necessary characteristics of not absorbing in the 600-400 cm'1 region and of being a solvent for both water and the silane. The physical constants of acetonitrile and l,2-dimethoxy- ethane are shown in Table 5. The dipole moment of acetonitrile is 37.5 Debyes. while that of l,2-dimethoxyethane is 7.2. The higher the dipole moment the greater the ionization ability of the sol- vent. For an SnZ-Si reaction mechanism, the rate controlling transition intermediate would be of the form: \/ 6_ 0 - Si - Cl I TABLE 5.--Physical Constants of Acetonitrile and l,2-Dimeth0xyethane. 5+ H:2 -‘ Acetonitrile l,2-Dimethoxyethane Molecular weight 41.053 90.123 Boiling point (°C) 8l.6 93.0 Density at 20°C 0.78 0.8665 Refractive index l.344 l.3796 Dipole moment (Debyes) 37.5 at 20°C 7.20 at 25°C 48 This transition state involves the separation of charges. Any solvent which stabilizes this intermediate would increase the reaction rate. Hence, if this mechanism is correct, the rate of reaction would be faster in acetonitrile than in l,2-dimethoxy- ethane. The infrared spectra of l,2-dimeth0xyethane and acetoni- trile are shown in Figures 16 and 17, respectively. Acetonitrile 1 region. Some of has no absorption bands in the 400 to 600 cm' the spectra obtained during run I? are shown in Figure 18. The spectra are quite similar to those obtained using l,2-dimethoxy- ethane as a solvent. The spectra obtained during the hydrolysis reaction at long reaction times are similar to those of PhSi(OH)3. Reagent grade acetonitrile was prepared by drying over calcium sulfate (Drierite) followed by distillation from phos- phorous pentoxide. Only the middle fraction, approximately 80 percent, was used for the hydrolysis study. To prevent poly- merization of acetonitrile, the phosphorous pentoxide was limited from 5 to 10 grams per liter of solution. The amount of water remaining in the acetonitrile can be calculated from the initial spectrum of PhSiCl3 by comparing the amount of PhSiCl3 and PhSiCl2(0H) initially present. Using this method, the amount of water present was determined to be less than 0.5 moles per liter of solution. Since the effect of hydrogen chloride was not studied in acetonitrile, the reactions were modeled as three irreversible SnZFSi reactions in series. 49 (%) aoueaitwsueut one com .mcmsumzxosumewo4~.F to Ezcuumam cmcmcmcHuu.mp mgzmwm A_-Euv xucmzcmeu com com com om r u u 50 (z) aouenngwsueul .mpwcuwcoumu< to Eacuumam nmcmcmcH--.n_ mcamwu Apisuv xocmscmgu ooc ccm can con cow a q a a d - - om u cc 1 on 1 g; 51 (z) aouenzgwsueul 0 0 00¢ 0mm 0 d a q q u oe F 1 on ow I 1 .00m mmm.o u h [— cop I om ow I on ow COP .Umm OGP.O M P .m_wcpccoumo< c. «Fuwmga A—nsov xucwacmgm ooe 0mm a...J I Wow"... mood u H mwmxpocuxz mgu mcwcao cmcmmuao Easpumam umcmgecmlu.wp mcamwm 0mm 4 . a 1.0mm mm~.o u h muzmmmmmm mp cam 52 K1 (1) PhSiCl3 K2 (2) PhSiClz(0H) + H20----——->-PhS'iC1(0H)2 + HCl K3 (3) PhSiC1(OH)2 + H20--———+ PhSi(OH)3 + HC1 + H20 ---+ PhSiC12(OH) + HC1 It was observed that the third reaction (3) was faster than the second and that a steady state approximation was applicable to PhSiCl(0H)2. The rate expressions for these reactions may then be written as: d [PhSiC1]3 . dt = -K.I [PhSiC13] [H20] d [PhSiC12(OH)] dt = K1 [PhSiC13] [H20] - -K2[PhSiC12(OH)] [H20] d [PhSiC1(OH)]2 dt = K2 [PhSiC12(OH)] [H20] - K3 [PhSiC1(OH)2] [H20] d [PhSi(OH)3] dt = K3 [PhSiCl(0H)2] [H20] Assuming a steady state on PhSiCl(0H)2. (3.9) (3.10) (3.11) (3.12) 53 K2 [PhSiC12(0H)] [H20] - K3 [PhSiCl(0H)2] [H20] = o (3.13) K2 [PhSiCl(0H)2] = KS-[PhSiC12(0H)] Equation 3.12 may now be written as: d [PhSi(OH)3] dt = K2 [PhSiC12(0H)] [H20] (3.14) Using this steady state approximation, the hydrolysis series reac- tions for PhSiCl3 can be written with two rate constants K1 and K2. The following reasoning was used to estimate the extinction coefficients of the reactants in acetonitrile. The intensity of an absorption band in solvents of refractive indexes n1 and n2 has been expressed by Brown (16) as: 2 l + c - C/ng 2 1 1'1 l (3.15) 'jPhé? 3 2 1 + C - C/n Here, A is the absorption in the two solvents, and C is a constant which depends on the geometry of the solute molecule. The refrac- tive index of acetonitrile is 1.34 and that of l,2-dimethoxyethane is 1.38. Since these refractive indexes are essentially the same, one would expect the same band intensity in both solvents. Not enough data was obtained in acetonitrile to calculate all band intensities used in the hydrolysis study of PhSiCl3. However, the 465 cm'] band for PhSi(OH)3 was determined to be 255 liters/mole- centimeter. This compares with 237 for the same band in 54 l,2-dimethoxyethane. Therefore, the extinction for acetonitrile were assumed to be the same as those for l,2-dimethy0xyethane. To determine the rate constants K1 and K2, a least squares program was employed. The values and linear estimate of standard deviation are shown in Table 6. Due to the fast rate of reaction, the first rate constant K1 is near the limit of the system to determine. This is evident from the large value of the linear estimate of standard deviation. Therefore, K1 should only be con- sidered an estimate of its actual value. Figures 19 and 20 show the concentrations of the silanes versus the residence time for runs 17 and 18. The initial condi- tions for these experiments are given in Table A-3 in Appendix A. The significance of this study of the reactions in acetoni- trile is the marked increase in the rate of reaction rather than a model for the system in acetonitrile. This increase in rate is especially evident from the value of K2. Kleinhenz and Hawley (7) determined a value of 77.5 liters/mole-second for this rate constant in l,2-dimethoxyethane using a similar irreversible SnZ-Si mechanism. This compares with a value of 122 liters/ TABLE 6.--Parameters Used to Describe the Hydrolysis Reactions of PhSiCl3 in Acetonitrile. .fié— Linear Estimate of Rate Constant Standard Deviation 180. liter/mole-sec. 147. K _.a II 122. liter/mole-sec. 37.9 Concentration of silanes (moles/liter) 55 Initial Concentrations _ Legend Bflfl_ll A PhSlC13 . . . ' = Moles of Total Silane o SE?1%0§)3H) HSi 0'0701 Liter of Solution . M = 0 218 Moles of Water Added ’HZO ’ Liter of Solution M = 0 0 Moles of HCl Added HCl ' Liter of Solution .0 9 “ 0 8 ” .0 7 ” ___ -0 6 F G 0 .0 5 “ O .0 4 .0 3 - .0 2 O .o 1 ' ° 1 l 4 1 I 1 I .l .2 .3 .4 .5 .6 .7 .8 -.9 t - Mean time; sec. Figure l9.--Concentration of Silanes Versus Mean Reaction Time for the Hydrolysis of PhSiCl3, Run 17. Concentration of silanes, moles/liter 56 Initial Concentrations " Legend Run 18 A PhSiC13 . ° PhSiClz(0H) M = 0 069 Moles of Total Silane F 0 PhSi(OH)3 Si ' Liter 0f Solution M = 0 22 Moles of Water Added 1 0” H20 ' Liter of Solution M = 0 0 Moles of HCl Added .0 9- H01 ' Liter of Solution .0 8- 0 7- 0 a 0 o 0 6” 0 .0 5- o .0 4 0 3- .0 2 0 ‘° 0 O O .0 l, A O 'l 2 3 4 .15 L 16 1 i 1 8 9 E - Mean time; sec. Figure 20.--Concentration of Silanes Versus Mean Reaction Time for the Hydrolysis of PhSiC13, Run 18. 57 mole-second for K2 in acetonitrile. This increase in rate can be accounted for by the higher dipole moment of acetonitrile compared with l,2-dimethoxyethane. This increase in dipole moment enables the solvent to better stabilize the transition state intermediate. H. Concluding Remarks It was found that the hydrolysis of the first chlorine atom of PhSiCl3 is much faster than that of the remaining two. This is similar to the observation of Shaffer and Flanigen (6), that the first hydrolysis reaction in a series is faster than the following hydrolysis reactions. The second and third hydrolysis reactions were much closer in their rates. However, the hydroly- sis of the third chlorine was slightly faster than that of the second. Hence, a small amount of PhSiCl(OH)2 was present during the reaction. A major observation was the effect of HCl, which decreased the rate of hydrolysis. This observation was supported by the study of Shaffer and Flanigen (6), who observed the decrease in the rate of hydrolysis of PhSiCl3 by saturing l,2-dimethoxyethane solution with HCl. The effect of HCl was incorporated into the kinetic equations in the form of [HCl]n and the value of n deter- tnined using a least squares fit. The effect of HCl can be accounted for by considering the equilibrium of H+, C1' in the solvent with water. H” + 01' + H20 *2. H30" + 01' 58 The hydronium ion H30+ should be considerably less reactive than water. As the amount of HCl increases, more water becomes tied up as hydronium ion. Therefore, an increase in the HCl concentration would suppress the hydrolysis reaction. However, the effect of HCl is complicated by changes in the medium with changes in the amount of HCl present. The chloride ion has the ability to stabilize the transition state intermediate and increase the rate of reaction. However, for the hydrolySis of PhSiCl3 with a relatively small amount of water present, the prominent effect of HCl is to decrease the rate of reaction through formation of the hydronium ion. HCl is also known to catalyze the condensation reaction. During the residence times used for these studies, no indication of condensation was observed in the experiments where there was no initial concentration of HCl. However, in the experiments with an initial HCl concentration the condensation effect can be seen at long residence times where the amount of silanes present falls below that predicted by the model. Therefore, the dual effect of HCl of suppressing hydrolysis and catalyzing:condensation causes condensation to occur before hydrolysis is complete at high HCl concentrations. In a saturated solution, it might be reasonable Cl Cl 1 l to expect such groups as Ph-Si-O-Si-Ph, which were indicated by C1 C1 Shaffer and Flanigan's (6) study, to exist. Hence HCl has a major role in determining the products obtained during the hydrolysis of halosilanes. 59 The hydrolysis reactions of PhSiCl3 were modeled assuming that they were first order with respect to both water and silane. Such a model is consistent with the SnZ-Si mechanism that has been proposed to describe these reactions (1: p. 93). These studies were conducted in a polar solvent. In a nonpolar or slightly polar solvent, the rate of reaction with respect to water or silane should be greater than one. The addition of water or silane to a slightly polar solvent would increase the ability of the medium to stabilize the transition state intermediate. Hence, the rate of reaction with respect to water or silane would be greater than one. CHAPTER IV THE HYDROLYSIS REACTIONS OF DIMETHYLDICHLOROSILANE The hydrolysis reactions of dimethyldichlorosilane were studied using the experimental techniques discussed in Chapter II. The objects of this study were to obtain experimental data on these reactions, model these reactions, determine the rate con- stants for this model, investigate the effects of hydrogen chloride on these reactions and determine the temperature effect. The data was collected by varying the length of the laminar flow reactor. Since a steady state occurred in the reactor, it was possible to study the concentrations at one residence time. By varying the length of the reactor, data on concentrations versus residence time were obtained. The hydrolysis reactions of Me25i012 were studied in three parts. First, these reactions were studied with initial concen- trations of 0.57 and 0.48 moles/liter of HCl and silane concentra- .ti0ns of approximately 0.1 moles/liter at 0°C. Since at most 0.2 moles/liter of HCl are formed during these reactions, more HCl is initially present in these runs than is formed during the hydrolysis reactions. Approximately ten data points at times from 0.064 to 3.62 seconds were obtained during each of these runs. Next, three runs at 0°C with no initial concentration of HCl were made. The concentration of silane varied from 0.10 to 60 61 0.12 moles/liter with excess water initially present. Again ten data points were obtained during each run at times ranging from 0.18 to 2.13 seconds. To determine the effect of temperature on these reactions, the reactions were also studied at 267.5 and 252.0°K. Using the data from these runs, the runs at 273.0°K, and measurements on the equilibrium shift with temperature, the activation energies of the various reactions were determined. The data obtained consisted of a series of absorption peaks versus residence time. In order to utilize this data, it was necessary to assign these peaks to the various reactants, unstable intermediates, and products of the hydrolysis reactions of MeZSiClZ. A. Interpretation of Infrared Spectra Obtained During the Hydrolysis Reac- tions of Dimethyldichlorosilane The following spectra assignments were made for the absorp- tion bands of MeZSiCl2 and its hydrolysis products: TABLE 7.--Infrared Assignments for MeZSiCl2 and Its Hydrolysis Products. Me251012 . . . . . . . . . 532 cm'1 and 470 cm"1 MeZSiCl(OH) . . . . . . . . 572 cm“ and 542 cm‘1 1 MEZSi(OH)2 . . . . . . . . 510 cm 62 The infrared spectrum of Me25i012 is shown in Figure 21. This spectrum was obtained using a 0.1 M solution of dimethyl- dichlorosilane in l,2-dimethoxyethane. The asymmetric and symmetric 1 1 vibrations of SiCl2 occurred at 532 cm' and 470 cm' , respectively. Smith (17) published an infrared spectrum of gaseous MeZSiClz with the asymmetric and symmetric vibrations of SiCl2 occurring at 1 1 553 cm' and 473 cm" , respectively. The shift in frequencies of these vibrations may be accounted for by the difference between the gaseous and liquid states (18: p. 41). Similar differences were noted by Smith (17) for SiC14. After water was added to the system, the bands at 532 cm-1 and 470 cm"1 gradually decreased and new peaks appeared at 542 cm-1 1 and 572 cm' . The 572 cm'] peak was assigned to the Si-O vibration 1 and the 542 cm" peak to the Si-Cl vibration in Me25i01(0H). Since MeZSiCl(OH) is an unstable hydrolysis intermediate, its infrared spectrum has not been published. However, the 542 cm"1 assignment 1 for Si-Cl falls in the range (465 cm' to 560 cm'1) given by Smith (9) for the Si-Cl vibrations. As the residence time was increased, the peaks at 572 cm-1 and 542 cm'1 gradually decreased and a new peak appeared at 510 cm']. This peak was assigned to the Si-O vibration in Si(0H)2. No published infrared spectrum for MeZSi(0H)2 includes the infrared region below 900 cm']. The reason for this lack of information on the spectrum of Me25i(0H)2 is the highly unstable nature of MeZSi(0H)2. Since several authors have reported preparing MeZSi(0H)2 using a variety of techniques, MeZSi(0H)2 does not appear to be 63 (z) aoueqthsueul .mpuwmmmz mo Ezguumqm cmemgmcmii._~ weaved “Fusuv aocmacmgu 0 see com com com com . d d) d u q d d cm 1 cc i om ow oer 64 significantly less stable than other silandiols. However, the con- densation of MeZSi(0H)2 is base catalyzed to such an extent that the hydroxyl groups in glass cause condensation to occur. To avoid this problem,most researchers recommend the use of quartz materials in the preparation of this compound. To confirm the assignment of the 510 cm'] to MeZSi(0H)2, the silandiol was prepared using a technique similar to that of Takiguchi (ll). Takiguchi prepared Me25i(0H)2 in an ether solution using aniline as a hydrogen chloride acceptor. To prepare MeZSi(0H)2, 0.164 moles of Me2Si012 were slowly added to a cooled l,2-dimethoxyethane solution (0°C) containing 0.33 moles of aniline and 0.33 moles of water. Infrared spectrum of samples taken while adding MeZSiClz showed an infrared peak at approximately 510 cm']. This peak increased in strength as addi— tional MeZSiCl2 was added to the solution. However, condensation was a problem and due to heating of the cell, may have occurred while scanning the infrared spectrum. Figure 22 shows the bands of MeZSiClZ, its hydrolysis Me Me and common condensation products, such as HD-Si-O-Si-OH. It is in Me interesting to note that the spectra of the condensation products do not correspond to any of the spectra obtained during the hydrolysis experiments. This supports the conclusion that little condensation occurs during the hydrolysis experiments. 65 Legend M Medium VW Very Weak MeZS‘iCl2 M M MeZSiC1(OH) IN N MeZSi(OH)2 M HO(Me)ZSiOSi(Me)2(OH) M VW (Me SiO) 2 3 M (MeSiO) 4 M f l T 1 l l 700 600 500 400 Frequency (cm']) Figure 22.--Infrared Spectrum of MeZSiClz, Its Hydrolysis and Common Condensation Products. 1 66 B. Condensation Reaction The condensation reaction of Me25i(0H)2 has been studied by Chrzczonowicz and Chojnowski (13). The reaction was found to involve Sn2 substitution and to be acid catalyzed. The rate expres- sion used to describe this reaction is: :g—tLS—‘O—HL Km [HCl] [31011]2 (4.1) The reaction is first order with respect to hydrogen chloride and second order with respect to the silandiol. At 25°C and using dioxane as a solvent, KIII was determined to be 0.33 moles.2 liter2 sec.']. At the concentrations of HCl present during the hydrolysis study, the rate of this condensation reaction at 25°C is 100 times slower than the hydrolysis reaction at 0°C in l,2-dimethoxyethane. One may expect an increase in the rate of condensation in 1,2- dimethoxyethane. However, Shaffer and Flanigen (6) found mixed results when comparing the combined hydrolysis-condensation reac- tions of MeZSiClz and PhSiCl3 in these solvents. The rates of the reactions of MezSiCl2 was three times as fast in dioxane as in l,2-dimethoxyethane. However, the rates of the reactions of PhSiCl3 were slower in l,2-dimethoxyethane than in dioxane. Yet, even an increase of three times in the condensation reaction would have little effect on this comparison of the hydrolysis and conden- sation reactions of MeZSiC12. Therefore, one can conclude that the condensation reaction of MeZSi(0H)2 does not occur to any significant extent under the conditions present during the hydrolysis reactions experiments. 67 Several authors state that in comparing the hydrolysis and condensation reactions of halosilanes, hydrolysis occurs first, leading to silanols, -diols, or -triols. These then undergo con- densation with loss of water. Various examples are given here as further evidence that condensation does not occur during the time of the hydrolysis reactions studied in this work. Noll (2: p. 109) gives a general mechanism for the hydrolysis-condensation reactions of halosilanes. Here, complete hydrolysis occurs followed by condensation. Petrov et a1 (19: p. 1) give the following scheme for the hydrolysis of MeZSiClz. 2 H 0 . 2 ‘ . . Roberts and Caserio (20: p. 1193) state that the hydrolysis of MeSiCl3 gives MeSi(0H)3 which is unstable and rapidly undergoes condensation with the loss of water. Condensation can also occur between organohalosilanes and organosilanols as shown below: Si-X + HO-Si ---————+ Si-O-Si + HX Noll (2: p. 204) states that this reaction occurs on heating organohalosilanes and organosilanols. Also, these reactions may also occur during the hydrolysis if the hydrolysis reaction is slow. The fast rate of the hydrolysis reactions of MeZSiClz makes this type of reaction extremely unlikely. 68 C. Calculation of Concentration Since Beer's law held for PhSiC13, it was assumed to hold for MeZSiCl2 and its hydrolysis products. The concentrations of MeZSiClz, MeZSiCl(0H), and Me25i(0H)2 were determined from the 1 1 and 510 cm'], respectively. The absorption for 532 cm" , 572 cm" the various peaks was determined using the base-line technique. A material balance for the total silane present during several runs was used to determine the extinction coefficients. Using a least square fit, the following extinction coefficients were obtained: TABLE 8.—-Extinction Coefficients for MeZSiCl2 and Its Hydrolysis Products. Linear Estimate of Extinction Coefficient Standard Deviation 8532 = 122.3 (chm)'1 4.9 8572 = 119.8 (chm)'1 7.2 8510 = 102.2 (chm)'1 15.0 0. Conditions for the Hydrolysis Reactions of MezSiClz The reagents were prepared for the hydrolysis experiments using the methods described in Chapter II. They were drawn into the tanks and an initial spectrum of the silane obtained. The initial concentrations of the various silanes were determined from this reference spectrum. These concentrations were corrected for 69 dilution due to the addition of the water-solvent mixture. The initial conditions for the experiments used to describe the hydroly- sis reactions of MeZSiCl2 are given in Table A-5 in Appendix A. Complete hydrolysis of MeZSiClZ was observed in the experi- ments. However, MeZSiCl(0H) and Me23i(0H)2 followed by the 542 1 peaks, respectively, reached an equilibrium cm'1 and 510 cm' state. A higher concentration of water shifted the equilibrium toward MeZSi(0H)2. Spectra obtained during run 44 are shown in Figure 23. At 1 1 the end of the run, the peaks at 542 cm' and 510 cm" are still present. E. HCl Effect To determine the effect of HCl on the reaction, several runs were made with an initial concentration of HCl. The initial HCl concentration of approximately 0.5 moles/liter is considerably stronger than the approximately 0.2 moles/liter of HCl formed dur- ing the runs. The data obtained during runs 37 and 38 are given in Table A-6 in Appendix A. No difference between the rate of reactions was observed for runs with an initial HCl concentration and those with no HCl initially present. Therefore, the reactions with HCl initially present can be described by the same model as those with no HCl initially present. 7O (%) aoueqatmsueul .ve cam mc_czo cmcwmuno meuomam11.m~ acumen A—lsuv zucmscmcu ooe omo ooe omm .-__ .___J cm 1 cc 1 oo 1 ow I .eem eN._ n e r ooP ' cm I 1 ow 1 oo 1 cm i u.umm o—N.o n H oopr .umm mo¢.o n H F I I fi.eem Nee.o n e W l ITj r j .umm 5mm.o u H muzmmmumm cc cam 71 F. Results and Discussion The following reaction steps were used to model the hydrolysis reactions of MeZSiC12. K1 (1) MeZSiC1 + H20 --—-+ MeZSiC1(0H) + HC1 2 K (2) MeZSiCl(0H) + H20-———§—+Me25i(0H)2 + HCl K (3) Me25i(0H)2 + HC1-———3—+ MeZSiCl(OH) + H20 At the concentrations used in this study, HC1 had little effect on the reaction rates. The reactions were assumed to be first order with respect to the water and silane and second order overall. The rate expressions used to describe the hydrolysis reactions of MeZSiCl2 are: d[Me SiCl ] dt 2 2 = -K] [MeSiClz] [H20] (4.2) d M 5'01 0H di e2 1 ( )1: K1 [MeZSiClZ] [H20] -K2 [MeZSiCl(OH)] [H20] + K; [MeZSi(0H)2] [HCl] (4.3) d[Me Si(0H) ] dt 2 2 = K2 [MeZSiC1(0H)] [H20] - K; [Me25i(0H)2] [HCl] (4.4) 72 [H20] = [H2010 - [MeZSiC1(OH)] - 2[MeZSi(0H)2] (4.5) [HC1] = [HC1]o + [MeZSiC1(OH)] + 2[MeZSi(OH)2] (4.6) Here, [H20]o and [HC1]0 are the initial concentrations of H20 and HCl, respectively. Using the data obtained during runs 44, 45, and 46, the con- stants listed in Table 9 were determined with the aid of a least squares type curve fitting program (15). The data points were weighted based on estimates of their variances. Figures 24, 25 and 26 are plots of the data points and model (solid lines) versus residence time for runs 44, 45, and 46, respec- tively. Figures 27 and 28 are plots of the data points and model versus residence time for runs 37 and 38 and illustrate the effect of HC1 on the reactions. The kinetic parameters were determined assuming that the system was isothermal and that plug flow existed. The effect of these assumptions is discussed in Chapter V and results in TABLE 9.--Parameters Used to Model the Hydrolysis Reactions of MeZSiClZ. Parameter 32332455321353; K1 = 38.3 liters/mole-sec. 4.81 K2 = 11.1 liters/mole-sec. 1.94 K. = 20.2 liters/mole-sec. 4.77 Concentration of silanes, moles/liter 73 Initial Concentrations Legend Run 44 F M S'Cl : M235}(0§) M5. = O 106 Moles of Total Silane D M9251c1(05) 1 ° Liter of SOTution = Moles of H20 Added MH20 0'24] Liter of sqution‘ M = 0 0 Holes of HCl Added .08" HC1 ' Liter of Solution .07L 0 (:1 D D D D 06” ° .05 ~ .04 A A A A .03" A .02 “ A .01 F . ' O 1 1 1o 10 1 1 1‘ 1 1 1 1 .1 .2 .3 .4 .5 .6 .7 .8 .9 1.0 1.2 t - Mean time; sec. Figure 24.--Concentration of Silanes Versus Mean Reaction Time for the Hydrolysis of MeZSiClz, Run 44. 1 .4 Concentration of silanes, moles/liter ,- 74 Initial Concentrations F Legend Run 45 oMeZSiCl M g 0 103 Moles of Total Silane AMeZSiCl 0H) Si ' Liter of Solution t3MeZSi(OH)2 M _ .M9195 of H20 Added H20 ’ 0°236 Liter of Solution M = 0 0 Moles of HC1 Added .08- HC1 ' Liter of Solution D .071- C1 .06 0 Cl ' C1 .05 F .04 ___ A A A .03 - A A O A A .02 - O o o .01 “ O O 1 L l i l ' I l l l l .2 3 .6 .7 .8 .9 1.0 1.2 1.4 t - Mean time; sec. Figure 25.--Concentration of Silanes Versus Mean Reaction Time for the Hydrolysis of MeZSiClZ. Run 45- Concentration of silanes, moles/liter 75 Initial Concentrations Run 45 M _ = 0 1,. Moles of Total Silane $1 ' ”“ Liter of Solution _ Moles of H20 Added M - 0.195 . r F Le end H20 L1ter of‘Solut1on 0MeZSiCl M = 0 0 Moles of HCl Added 0 MezSiC110H) HC1 ' Liter of Solution 0 D 0.8 ” 0 0 D C) 0.7 ‘ 0.6 ' 0.5 ‘ 0.4 ~ A 0.3 " 0 . A . A o.2~ ' _ O O 0.1 ' o .1 .2 .3 .4 .5 .6 .7 .8 .9 1.0, 1.2 t - Mean time; sec. Figure 26.--Concentration of Silanes Versus Mean Reaction Time for the Hydrolysis of MeZSiClz, Run 45, Concentration of silanes, moles/liter .12 .11 .10 ~ .09 .08 .07 .06 .05 .04 .03 .02 .01 76 . Initial Concentrations Run 37 . = Moles of Total Silane ASi 0'1283 Liter of Solution . = Moles of 520 Added [Eggend MH20 0°286 Liter of Solution 0 M8231C1 W 1 f A 4 S'Cl H .. = “9 es 0 HC1 AddEd r o M:§51(0§?2) ~IHC1 0'569 L1ter of Solution _ A A A A A D D D 00 C1 0 111101 V11 1111-1 L1L4 .l .2 .3 .4 .5 .6 .7 .8 .9 1.0 1.2 1.4 1.6 1.8 t - Mean time; sec. Figure 27.--Concentration of Silanes Vensus Mean Reaction Time for the Hydrolysis of MeZSiClz. Run 37. Concentration of Silanes, moles/liter 77 Initial Concentrations Legend Run 38 .121 0Me SiCl - Moles of Total Silane C’Me 25i(0H) M - = 0.109 ° ‘ J AM6251C1(08) Sl L1ter of Solution .11 ' Moles of H20 Added M = 0, H20 418 Liter of Solution 10 - ' _ Moles of Hcl Added TMHCl ' 0'4815 Liter of S01ution A A .09 - A A .08“' .07 ‘ .06 - .05 ~ .04 D D .03“ D .02 - .01 - O 1 L 1 1 L 1 1 1 1 1 1 g 1 1 1 1 l 1 .l .2 .3 .4 .5 .6 .7 .8 .9 1.0 .1.2 1.4 1.6 1.8 t - Mean time; sec. Figure 28.--Concentration of Silanes Versus Mean Reaction Time for the Hydrolysis of MeZSiClz, Run 38. 78 approximately a 10 percent bound between the parameters determined assuming isothermal plug flow and those determined assuming laminar flow with heat transfer. 0. Temperature Effect Only limited information is available on the temperature effect on the hydrolysis reactions of chlorosilanes. Generally, these reactions are highly exothermic due in a large part to the heat of dissolving hydrogen chloride in the water solvent mixture. However, in a sturated solution of hydrogen chloride, there may be an actual cooling effect of the reaction due to the vaporization of the hydrogen chloride. The only apparent study to calculate the activation energy of the hydrolysis reaction of chlorosilanes was conducted by Shaffer and Flanigen (6). Their technique was described earlier and measures only the rate limiting step in the combined hydroly- sis and condensation reactions of chlorosilanes in a sturated solu- tion of hydrogen chloride. For the hydrolysis of MeZSiClz, Shaffer and Flanigen determined that their indicated hydrolysis end product was a chlorine end blocked linear polysiloxane. The rate of reac- tion was measured at 0°C and 25.7°C. Using the rates calculated at these temperatures, the activation energy was determined to be approximately 25 kcal/mole for the combined hydrolysis condensation reactions. One objective of this research was to determine the activation energies for the individual hydrolysis reactions of 79 MeZSiClz. To accomplish this, the hydrolysis reactions of MeZSiCl2 were studied at temperatures ranging from 16°C to -21°C. The higher temperature (16°C) was maintained through a combination of an on-off heater, a water circulating pump, and a cooling coil circulating water at 15°C. To obtain temperatures below 0°C, a methanol water solution was used. Here methanol was added to reduce the freezing point of the solution to the desired tempera- ture. A liquid-solid equilibrium was obtained by adding dry ice to the solution. Through the use of these techniques, the temperature was maintained to within 0.5°C of the desired temperature. The solution in the constant temperature bath was circula- ted through the heat exchanger surrounding the reactor. At 252°K, a sludge formed in the constant temperature bath, which caused some problems in circulating the solution through the heat exchanger. A marked difference in the intensities of the absorptions of the infrared bands occurred with the changes in temperature. As the temperature was increased a decrease in band intensity was noted and a similar increase in intensity occurred as the tempera- ture was decreased. This observation has been made by other inves- tigators in the infrared region and has been attributed to one of two effects. Slowinski (21) observed this decrease in intensity of absorption with an increase in temperature. It was observed that a band at 100°C was in some cases only 70 percent as intense as the same band at 25°C. This effect was attributed to the presence of rotational isomers. Brown (22) attributed this effect to colli- sions of the solute molecules with the walls of the solvent cage. 80 To correct for this effect, the extinction coefficients obtained at 0°C were multiplied by an appropriate constant based on a material balance on the system. This approximation is based on the fact that the extinction coefficients are all of similar magnitude and are due to similar vibrations. Therefore, one would expect the rate of change of the extinction coefficients with respect to tem- perature to be similar. It was necessary to use this method of 4 determining the extinction coefficients at different temperatures since not enough data was available for an independent evaluation of the extinction coefficients at each temperature. At 16°C, the rate of reaction was extremely fast. This plus the loss of accuracy in determination of the concentrations caused by the decrease band intensities made the calculation to the hydrolysis rate at 16°C impossible with the desired degree of accu- racy. Therefore, only temperatures at or below 0°C were used to determine the activation energy. The hydrolysis of MeZSiC12 is described by one irreversible and one reversible reaction. The activation energies were deter- mined by measuring the reaction rates at 0, -5.5 and -21°C. The rates obtained for these reactions at the three temperatures are shown in Table 10. Figures 29 and 30 (on pages 82 and 83) are plots of concentrations versus residence time for the hydrolysis reactions at 267.5 and 252.0°K, respectively. Due to the slower rates of reaction at 252.0°K, the rate constants were determined with more accuracy at this temperature. This resulted in a smaller linear estimate of standard deviation 81 TABLE 10.--Rate Constants for the Hydrolysis Reaction of MeZSiC12 at Different Temperatures. 0 Rate Constant Linear Estimate of Temperature ( K) (liter/mole-sec.) Standard Deviation K1 273.0 38.3 4.81 267.5 34.0 9.91 , 252.0 23.7 5.71 K2 273.0 11.1 1.94 267.5 7.37 3.94 252.0 3.58 1.3 K2 273.0 20.2 4.77 267.5 19.2 15.4 252.0 14.6 6.47 for these rate constants compared to those determined at 267.5°K. Considerably more data was available at 273.0°K, which accounts for the smaller standard deviations in the rate constants at this tem- perature. The standard deviations for the parameters at 267.5°K are high since only one data set was used to determine these parameters. In determining the activation energies for these. reactions, the rate constants at the various temperatures were Concentration of silanes, moles/liter 82 Initial Concentrations Run 42 Legend "- ~101' . _ Moles of Total Silane SMEZSICHOH) MSi ' 0‘12“ Liter of Solution oMEzS1C12 Me Si(0H) . 2 2 = Moles of Water Added .09 MHZO 0.323 L of S M = 0.0 Moles of HCl Added L of S 1 1 1 1 #1 0 .1 .2 .3 .4 .5 .6 .7 .8 .9 1.01.1 1.2 1.3 1.4 t - Mean time; sec. Figure 29.--Concentration of Silanes Versus Mean Reaction Time for the Hydrolysis of MeZSiClz, Run 42. 7*S.2<”zer< Concentration of silanes, moles/liter 83 ‘ Initial Concentrations —9-—L° °"° Run 49 .12r ° "825102 . 2 MeZSi(OH)2 M . = 0 07.7 Moles of Total Silane MeZSiCl(0H) ~ 51 ' Liter of Solution .llr _ Moles of Water Added _ MH 0 - 0'180 L of S .10 2 = Moles of HCl Added -09L MHCl 0'0 L of s .08~ .07- .05” A A A /O’“D\A A .05“ .04" .03" D o l 1 l o .2 .4 .6 .8 1.0 1.2 TIA 1.6 1.8 E - Mean time; sec. Figure 30.--Concentration of Silanes Versus Mean Reaction Time for the Hydrolysis of MeZSiClz, Run 49. 7fi2‘2gulfikf 84 given the appropriate weights based on their linear estimate of standard deviation. The activation energies and frequency factors were calcu- lated for the two forward reactions using the reaction rate con- stants determined at the different temperatures. The rate con- stants K; were not known accurately enough to allow determination of the activation energy of the reverse reaction. To determine the activation energy of the reverse reaction, the equilibrium constant K2/Ké was measured at 273.0, 262.0, 267.5 and 252.0°K. Using these equilibrium constants, the difference between the activation energies of the forward and reverse reactions was calculated using the derivation shown here: K2 = K20 e (4.7) n 1 -AE2/RT K20 e (4.8) 7‘ ll . . (~AE2 + AEé)/RT Kz/K2 = KZO/KZO e (4.9) The activation energies and frequency factors for the two forward reactions and for the reverse reaction are given in Table 11. Figure 31 (on page 86) isia plot of 1/temperature versus loge (rate constant) for reactions 1 and 2. This illustrates a graphical technique which may be used to calculate the activation energies. The solid lines are drawn using the values of the activa- tion energies and frequency factors given in Table 11. 85 TABLE ll.--Activation Energies for the Hydrolysis Reactions of MeZSiClz. Linear Estimate of Reaction Standard Deviation Forward Reaction 1 Activation energy = 3.6 x 103 cal/mole 5.70 x 101 Frequency factor = 3.05 x 104 liter/mole-sec. 3.34 x 103 Forward Reaction 2 Activation energy = 1.05 x 104 cal/mole 6.4 x 101 Frequency factor = 3.02 x 109 liter/mole-sec. 3.4 x 108 Reverse Reaction 2 AEé - A52 = 4.37 x 103 cal/mole 1.2 x 103 4 Activation energy 1.4 x 10 cal/mole H. Conclusions on the Hydrolysis of MeZSiC12 This study demonstrated that the hydrolysis reactions of MeZSiC12 can be described by rate equations that are first order with respect to water and silane and second order overall. Such expressions are consistent with the Sn2-Si mechanism which is believed to describe the hydrolysis reactions of halosilanes. The hydrolysis of the first chlorine atom in Me25i012 is faster than that of the second. This is similar to the observa- tions made in the case of the hydrolysis reactions of PhSiCl3 and is consistent with the observation of Shaffer and Flanigen (6). Both the first and second hydrolysis reactions of MeZSiCl2 appear 86 5T 4 _ Reaction 1 3:? IS 3 ' in c O U .8 .2 0 01 £3 2 - Reaction 2 'l— 0 1 L 1 .4 .0036 .0037 .0038 .0039 .0040 l/Temperature (°K'1) Figure 31.--l/Temperature Versus Loge (Rate Constant) for the Hydrolysis Reactions of MeZSiC12. 87 to be reversible. However, under the conditions of this study, the equilibrium of the first reaction is shifted to the right to such a degree that it can be described as an irreversible reaction. The presence of HC1 had little effect on the rate of reac- tion. This indicates that the hydronium ion H30+ is more reactive in the hydrolysis of MeZSiC12 than in the hydrolysis of PhSiC13. Since the condensation reaction is catalyzed by HC1, the presence of HCl still plays an important part in determining the final products of the combined hydrolysis-condensation reactions. CHAPTER V ANALYSIS OF A LAMINAR FLON REACTOR The analysis of the rate data collected during these studies assumed that the reactor system was isothermal and that plug flow existed. These assumptions are commonly made in studying reactions in flow systems. In this chapter, the significance of these assump- tions is investigated. A. Scope of Past Work Several authors have studied these assumptions independently by either studying the plug flow assumption in an isothermal reactor or the isothermal assumption in a plug flow reactor. The assump- tion of plug flow in an isothermal system has been investigated by Kleinhenz and Hawley (7) and by Johnson (23). Johnson calculated the fraction of reagent reacted for both plug and laminar flow for first and second order reactions. From the comparison of the fraction reacted, one can determine the significance of the plug flow assumption in an isothermal reaction system. Kleinhenz and Hawley (7) plotted the ratio of conversion of laminar to plug flow for reaction orders from O to 3. They extended this analysis by calculating the reaction rate constants needed to achieve a certain conversion for plus or laminar flow. From the ratios of such rate constants, one can determine the significance of the 88 89 plug flow assumption on the determination of the reaction rate constants. Kleinhenz and Hawley concluded that the reaction rate constants determined assuming plug flow are a reasonably good estimate of the actual rate constants. Huang and Barduhn (24) have analyzed the isothermal assumption in a plug flow reactor. Considering an exothermic reac- tion, temperature profiles occur along the length of the reactor, since the heat generated by the reaction exceeds the heat trans- ferred through the reactor walls to the surrounding medium. At a certain length in the reactor, the temperature reaches a maximum and a hot spot occurs. If the reactor is long enough the tempera- ture in the reactor eventually reaches that of the surrounding medium. A factor, which may be used to correct the rate constants obtained assuming an isothermal reaction, is presented in the article. Merrill and Hamrin (25) investigated the laminar flow assumption with radial molecular diffusion for a three-halves order reaction. They studied the diffusion effect for an isothermal reactor, a reactor with heat transfer through the walls of the reactor, and for an adiabatic reactor. From this study, they con- cluded that radial diffusion affected the concentration profiles near the wall but had little effect on the overall conversion. 8. Scope of This Study The analysis presented here investigates the combined temperature and laminar flow effects on the concentrations and 90 temperatures in a flow reactor. Differential equations are devel- oped which describe the laminar flow system with radial heat con- duction. These equations are then written in dimensionless form with the appropriate dimensionless constants. Using numerical integration techniques and values for the dimensionless constants from the study of the hydrolysis reactions of MeZSiClz, the equa- tions are integrated to obtain concentration and temperature profiles in the reactor. From these profiles, the extent of reaction for laminar flow is obtained. The extent of reaction for plug flow ver- sus that of laminar flow is plotted for zero, first, and second, order reactions. From such plots, the combined effects of the iso- thermal and plug flow assumptions can be determined. This data can then be used to determine the effect of these assumptions on the determination of the rate constants. C. Development of the Laminar Flow Model The model developed here is based on laminar flow in a circular tube. The aim of this model is to determine the concen- trations and temperature profiles in a laminar flow reactor. Bird, Steward and Lightfoot (26: p. 46) have expressed the change of axial velocity with respect to radial direction for laminar flow in circular tubes as: dV 2 = AP r (5.1) HF“' 2p L 91 Here, AP is the pressure drop along the length of the tube, r is any radius in the tube, u is the viscosity of the fluid and L is the length of the tube. Integrating and applying the boundary conditions that V2 is zero at r = R, we have: 2 _ AP R r 2 V2 _ 4DT—E'11 - (R0 ] (5.2) Since at r = 0, the maximum velocity is obtained, equation (5.2) may be written as: _ :2 vz — vmax [1 - (R) 1 (5.3) where 2 _APR Vmax ' 4U L (5'4) The average velocity in the tube may be calculated by integrating VZ over the area of the tube and dividing by the area. 2 If n r VZ dr d8 = AP R2 Sn L = Z 11 n r2 dr d6 (5.5) Hence, the average velocity in the tube is equal to one-half the maximum velocity. (5-6) (V > = Z 92 Figure 32 illustrates a cylindrical element in reactor with laminar flow. An energy balance over the cylindrical element gives the following terms: Energy in by qur - 2n r Az conduction at r Energy out by q I - 2n (r + Ar) Az conduction at r + Ar r r+Ar Energy in by qzlZ - 2n r Ar conduction at 2 Energy out by q | . Zn r Ar conduction at z + A2 2 z+Ar Energy in with p C V(T - TO)|z - Zn r Ar flowing fluid at z p Energy out with p C V(T - T )l - 2h r Ar flowing fluid at z + Az p O Z+AZ Energy produced in 2n r Az Ar (-re) (-AH) ring-shaped element For an nth order reaction, we have: dC _ A _ -AE/RT N N re - T - -KO e CO (1 - X) (5.7) Equating the energy outputs to the energy inputs plus the energy produced h1the element and dividing by 2n Ar Az gives: (rqr)|r+Ar ' (rqr)lr + r qzlz+Az - qzlz Ar AZ + r p Cp VZ A2 2 + r re (-AH) = 0 (5.8) 93 .copomma zopm .5553 m E 29:um 30:31.8 <1... mm 9:5: 30,; yo cowpumcwo _ . . :o_pu:ucou .. cologcou A. cowuozucou 94 Taking the limit as A2, Ar + 0: 8 3 q .2;.= - %-5F-(rqr) - 525..» (.AH)re (5.9) Q) Introducing the velocity distribution for laminar flow and Fourier's law for heat conduction: - 21.. _ El q.--K..: q.--K.. ‘ (540) We now have the partial differential equation: 2 A r 2 3T _ l_§_. gl_ 3 T p Cp VM [1 ' (R9 ] 52'- K[r 3r [r 8r] ' [322]] - (-AH)re (5.11) The heat conduction in the z direction isnormallyrsmall in comparison with the heat convection term. Therefore, the heat con- duction term (azT/azz) may be dropped from equation (5.11) and we obtain the following equation: A r 2 81 _ 5_§__ gl_ 9 Cp VM[]'(R( ] —2" r r [r 3r] - AHKOe'AE/RT CE (1 - X)N (5.12) This is a partial differential equation describing the temperature distribution in a laminar flow reactor. The following boundary conditions may be applied to equation (5.12). 95 . - .31: B.C. l. at r — 0, 3r 0 B.C. 2: at r = R, T = Twall B.C. 3: at z = 0, T = T0 (for all r) In this analysis, we will consider only reactors where the initial temperature, To, is equal to the temperature at the wall, Twall' To simplify the manipulations involved, equation (5.12) is written in dimensionless form through the introduction of the fol- lowing dimensionless variables. 0 C T 0 = (:35703' (Dimens1onless temperature) n = z/R (Dimensionless length) E = r/R (Dimensionless radius) Equation (5.12) is now written as: Ev(]_gz)flw9§=i L m 2?. p p M o cp R dn rR at R o Cp he - AHK e'AE/RTc’gu-X)N (5.13) 0 Through the use of the following dimensionless constants, equation (5.13) is: 96 2 . 2 do a so a o -y/e N (1 - E ) --= - -- 'E + Be (1 - X) (5.14) dn 6 BE SE7- AE p C Y = RCTLAH500 _ K a- A p Cp VM R N-1 -RKOCO 8’ v M Equation (5.14) represents the final form of the partial differential equation describing the temperature distribution in a laminar flow reactor. This equation is a function of the extent of reaction, X. Therefore, the extent of reaction must be deter- mined for the laminar flow system. Merrill and Hamrin (25) have shown that the radial diffu- sion has a small effect on the overall conversion in a laminar flow system. Therefore, the radial diffusion term can be neglected in considering the conversion in laminar flow. Since the flow in the tube is quite high, we can also neglect the axial diffusion in the reactor. Writing a material balance on the ring-shaped element in Figure 31, we have: Material in at z VZ 2n r Ar C with flowing fluid Material out at z +~Az VZ 2n r Ar C with flowing fluid Material reacted in 2h r Ar Az (-re) ring-shaped element 97 Equating the inputs and outputs for this system, we have: Vz 2n r Ar ClZ - Vz 2n r Ar Cl - 2n r Ar Az (-re) = 0 (5.15) Z+AZ Dividing by A2 and taking the limit as Az goes to zero: (5.16) < 0.20) Nlo ll 1 When substituting the laminar velocity profiles for V2, the reaction equation for re and the extent of reaction for concentration, equation (5.16) becomes: vM [1 - (r/R)2] 332(- = Koe'AE/RCT (1 - X)NCON" (5.17) Equation (5.17) written in dimensionless form along with equation (5.14) describe the temperature and concentration profiles in the laminar flow reactor. dx _ B -Y/6 N —— - ——————e (1 - X) (5.18) d” (1 - £2) 2 (1 - 52) ga- - 3(1- x)N e'Y/e =%1§-§—+ 53%;] (5'19) 0 for all 5 .e—W’ lWV4/7g B.C. l for equation (5.18) at q = 0, X o C T B.C. 1 for equation (5.19) at n = o, e = 17-3515! for all t: O o C T B.C. 2 for equation (5.19) at E = l, 0 = 1:3fijtfl'for all n o 98 36 ,E=0foralln B.C. 3 for equation (5.19) at E = 0 The integration of equations (5.18) and (5.19) gives the radial temperature and concentration profiles along the length of the reactor. Equations (5.20) and (5.21) are used to find the average concentration and mixing cup temperature in the reactor. This is the concentration and temperature that would be measured if at one point along the length of the reactor, the reaction was stopped and the solution allowed to flow into a cup. The tempera- ture in this cup is the mixing cup temperature and the concentration is the average concentration. 24 6R X(€.n,6)V(€) X = (5.20) ave n V R Tmixing cup = N V R2 (5'21) In equations (5.20) and (5.21), the radial concentration and temperature profiles are multiplied by the velocity at their radial position and integrated over the area of the reactor. These are then divided by the average velocity multiplied by the area of the reactor. The numerical integration technique, numerical values for the constants used in the integration, and a computer listing of the program are given in Appendix B. 99 0. Conclusions This analysis of the isothermal and plug flow assumptions in a laminar flow reactor enables the significance of these assump-‘ tions to be studied. In the case of the hydrolysis reactions of MeZSiClz, this analysis shows that the isothermal and plug flow assumptions have little effect on the extent of reaction. This is seen in Figure 33, where the extent of reaction for laminar flow with heat transfer follows quite closely the extent of reaction for isothermal plug flow, especially for a second order reaction. SMEminaLflgw 95995195199-result:-injheeeesurenfnt ofma ——‘————4—~ " 4 «W1...— ... p.-h——._‘_...— lower rate constant, the slight increase in temperature actually W's-“r“ "“‘ 'TT' reduces thefldifference between laminar and plugmflow. The tempera- ture rise is not great enough to significantly affect the concen- trations in the reactor. The extent of reaction predicted considering laminar flow, heat of reaction, and radial heat transfer are still below those of an isothermal plug flow reactor. This analysis demonstrates that it is possible to study reactions in laminar flow and predict the temperature and laminar flow effects. Kleinhenz and Hawley (7) demonstrated the magnitude of the laminar flow assumption on the rate constants. The maximum error introduced by this assumption is approximately 16 percent. When the temperature effect is introduced, the maximum error is reduced to approximately 10 percent. Due to the mixing in the reactor, this error will be further reduced. Therefore the values of the rate constants determined in this study are well within this 10 percent bound. 100 1.0 Laminar f1 flow '6” Plug 0w -‘-zero order first order ~5” econd order XL (Conversion Laminar Flow) l .— 1— O .. .. . ... .. .1 .2 .3 .4 .5 .5 .7 .8 .9 X0 (Conversion plug flow) Figure 33.--Extent of Reaction for Laminar Flow With Heat Transfer Versus Extent of Reaction for Isothermal Plug. 101 Using rate and thermodynamic constants similar to those found in the hydrolysis reactions of MeZSiCl2 with an initial temperature of 0°C, the extent of reaction for laminar flow with heat transfer to the wall was obtained. Figure 34 illustrates the extent of reaction for an isothermal plug flow reactor versus that for the laminar flow reactor with zero, first, and second order reactions with heat transfer. To indicate the effect of laminar flow with heat transfer to the wall of the reactor on the rate constants, the ratio of the rate constants for an isothermal plug flow versus those for laminar flow with heat transfer are plotted in Figure 35. The temperature effect for a second order reaction with constants similar to those for MeZSiCl2 hydrolysis reactions is shown in Figure 34. In the adiabatic case the temperature reaches a maximum of 2.6°C above the initial temperature at the completion of the reaction. With heat transfer to the sides of the reactor, the temperature rise follows that of the adiabatic reactor until the extent of reaction reaches 0.1; from here the temperature falls below that for the adiabatic case. At an extent of reaction of 0.9, the temperature for the laminar flow reactor is approximately l.2°C below that of the adiabatic reactor. As the reaction reaches completion, the rate of reaction falls and heat transfer is more significant. If the reactor is long enough, the tempera- ture in the reactor will eventually reach that of the surrounding medium in the heat exchanger. As Huang and Barduhn (24) have shown in their analysis, the temperature in a plug flow reactor KP/KL 102 1.0 -‘—— Zero order .9 ‘ Second order First order —" \ .8 ~ .7 — .6 - .5 — 4 1 1 1 1 1 1 1 1 1 O .l .2 .3 .4 .5 .6 .7 .8 .9 XL (Conversion laminar flow) Figure 34.--Ratio of Rate Constants Versus Extent of Reaction for Zero, First, and Second Order Reactions. 103 -. ._Faz 8;“ oe eoemeaee “mo: new: setummm 30pm emcwEmA m cw mmwm monumewasmhul.mm oegmwm Aeoweoeom co eeoexwv 4x e. a. . m. a. m. N. _. a . . _ _ . d omem oeoeneee< 30 8513 aunqeuadmal 104 also reaches a maximum. However, since heat transfer in the plug flow reactor is gnEter than the heat transfer in the laminar flow reactor, the maximum temperature will be lower in the plug flow reactor. This analysis can be applied to flow reactors to determine the accuracy of the values of the rate constants determined from data on such reactors. It would be possible to predict at what initial concentrations a large error is introduced into the values of the rate constants by the laminar flow and isothermal assump- tions. This should aid investigators collecting data on existing processes containing flow reactors. CHAPTER VI CONCLUSIONS This study demonstrates the use of a laminar flow reactor to investigate fast reactions. The unique design of the flow system enables the residence time of the reactants to be varied without changing flow rates or location of the detector. Chapter of this study shows the effect of the plug flow and isothermal assumptions on the rate constants determined using these assump- tions. In both the hydrolysis reactions of PhSiCl3 and MeZSiClZ, the chlorines were replaced by hydroyl groups in series. In these studies, the hydrolysis of the first chlorine atom was con- siderably faster than that of the remaining chlorines. This is similar to the observation of Shaffer and Flanigen (6) in their study of various hydrolysis reactions of chlorosilanes. It is proposed that the hydrolysis reactions of chloro- silanes proceed through the following mechanism: \/ s -- —-—-> ...+ --- -... (1) H20 +/51C1 ..__. M H20 S|1C1 M - 2’ (2) M +H20-Si-Cl M — HO-Si—+ HC1 | \\ Reaction (1) represents the fast formation of the transi- tion state intermediate and is followed by the slower breakdown 105 106 of the transition intermediate, reaction (2). M represents sta- bilization of the intermediate by the medium. In both the hydrolysis reactions of PhSiCl3 and MeZSiClz, rate expressions first order with respect to water were used to describe the reaction. These rate equations were developed to model the reactions in a polar solvent. In a nonpolar solvent, the order of reaction with respect to water should be greater than one due to changes in the medium. Water has the ability to sta- bilize the transition state intermediate in the hydrolysis reac- tions. Hence, the addition of water to a nonpolar medium would noticeably increase the ability of the medium to stabilize the intermediate and thus increase the rate of reaction. One of the principal aims of this study was to investigate the effects of HCl on the hydrolysis reactions of the chlorosilanes. The addition of HCl to the initial reagents markedly suppressed the rate of the hydrolysis reaction of PhSiC13. This effect can be explained by the reaction of the ion H+ with with water to form the hydronium ion H3O+. Due to its positive charge, the hydronium ion is considerably less reactive toward the silane than the water molecule. No effect on the rate of the hydrolysis reactions of MeZSiCl2 was observed with changes in the initial HC1 concentra- tion. This difference in the effect of HC1 compared to its effect on PhSiCl3 can be explained by the nature of the substitutent groups attached to the silicone. In the case of PhSiC13, the phenyl group is electron withdrawing which decreases the electron 107 density on the silicon. In MeZSiC12, the methyl groups do not exhibit this electron withdrawing effect. Therefore, the hydronium ion is more reactive toward the silicon in MeZSiClZ. These differ- ent polar effects demonstrate the importance of the polar nature of the substituents on the rate of reaction. A similar effect of the substituents has been observed in the case of base or acid catalyzed condensation reactions of silandiols (2: p. 213). In acid catalyzed condensation, electron negative groups decrease the electron density of the oxygen on the silanol, thus making attack by H+ more difficult. In the case of base catalyzed condensation, the electron withdrawing groups have the opposite effect and facilitate the nucle0philic attack of OH' on the silicon. Under certain conditions, it may be reasonable to expect that HCl may actually catalyze the hydrolysis reaction of organo- chlorosilanes. One example of this may be a pseudo-first order reaction where H20 is present in excess. Here, the amount of water in the hydronium form would not be sufficient and the increase in HC1 would increase the polarity of the medium. Thus, an actual increaSe in the rate of reaction would result. The significant aspects of this study are the demonstration of the laminar flow technique to study fast reactions, the descrip- tion of the reaction models and mechanism, the effect of HC1 on the systems, and the investigation of the plug flow and isothermal assumptions. 108 The reactions studied were modeled assuming that plug flow existed and that the system was isothermal. A principal aspect of this study was to determine the magnitude of the error intro— duced by these assumptions. The study of these assumptions begins with a concentration-temperature model of a laminar flow tubular reactor with heat transfer to the wall of the reactor. The model is based on a ring-shaped element in the reactor with conduction , and convection of energy in the axial direction and with conduction in the radial direction. Plots of concentration for a laminar flow reactor with heat transfer versus an isothermal plug flow reactor are presented. From this analysis, the effect of these assumptions on the rate constants is determined. BIBLIOGRAPHY 109 11. 12. 13. 14. 15. 16. BIBLIOGRAPHY Sommer, L. H., Stereochemistry, Mechanism, and Silicon, McGraw Hill, New York (1965). Noll, W., Chemistry and Technology of Silicones, Academic Press, New York (196812 ' Mileshkevick, V. P.; Nikolaev, G. A.; Evdokimov, V.F.; and Karlin, A.V., Zh. Ohsch. K1im., 41 (3), 643 (1971). Allen, A. D., and Modena, G., J. Chem. Soc., 3567 (1963). Chipperfield, J. R., and Prince, R. H., J. Chem. Soc., 3657 (1963). Shaffer, L. H., and Flanigen, E. M., Phys. Chem., 61, 1591 (1957). Kleinhenz, T. A., and Hawley, M. C., M.S. Thesis, Mich. State Univ. (1970). Kriengsmann, H., and Schowtka, K. H., Z. Physik. Che., 209, 261 (1958). Smith, A. L., Spectrochim. Acta., 23A, 1075 (1967). Sadtle Standard Spectra, Sadtler Research Laboratories, Phila- delphia, Penn. 12137 (1972). Takiguchi, T., J. Am. Chem. Soc., 2359 (1959). Tyler, L. J., J. Amer. Chem. Soc., 77, 770 (1955). Chrzczonowicz, S., and Chojnowski, J., Rochzinki Chemii. Ann. Soc. Chim. Polororum, 36, 1293 (1962). Smith, D. C., and Miller, E. C., J. Opt. Soc. Am., 34: 130 (1944). . Dye, J. L., and Nicely, V. A., J. Chem. Ed. 48: 443 (1971). Brown, T. L., J. Chem. Phys., 24, 1281 (1956). 110 17. 18. 19. 20. 21. 22. 23. 24. 25. 26. 111 Smith, A. L., J. Chem. Phys., 21, 1997 (1953). Kendall, D. N., Applied Infrared Spectroscopy, Reinhold Pub- lishing Corp., New York (1966). Petrov, A. D.; Mironov, B. F.; Ponomarenko, V.A.; and Cherngshev, E. A., Synthesis of Organisolicon Monomers, New York, Consultants Bureau (1964). Roberts, J. D., and Caserio, M. C. Modern Organic Chemistry, W. A. Benjamin, New York (1967). Slowinski, E. J., and Claver, G. C., J. Opt. Soc. Am., 45, 1281 (1956). Brown, T. L., J. Chem. Phys., 24, 1281 (1956). Johnson, M. M., Ind. Eng. Chem. Fundam., 9, 681 (1970). Huang, J. S., and Barduhn, A. J., A. I. Ch. E. J., 20,6,1228 Merrill, L. W., and Hamrin, C. E., A. I. Ch. E. J., 16, 194 (1970). Bird, R. 8.; Stewart, W. E.; and Lightfoot, E. N., Transport Phenomena, John Wiley & Sons, Inc., New York (1960). NOMENCLATURE Absorption Base line absorption Constant in equation (3.15) Concentration (moles/liter) Average concentration in reactor (moles/liter) Initial concentration (moles/liter) Heat capacity (cal/g-°C) Infrared cell diameter (cm) Activation energy (cal/mole) Heat of reaction (cal/mole) Thermal conductivity (cal/sec-cm-°C) Rate Constant for laminar Flow Rate constant for plug flow Frequency factor Rate constants+defined by equations (3.3)-(3.6) (liters/mole) n/sec Rate constants defined by equations (4.2)-(4.6) (liters/mole-sec) Rate constants defined by equations (3.9)-(3.14) (liters/mole-sec) Length of reactor (cm) Moles/liter Power to which concentrations are raised Number of maximum equation (3.2) 112 "1’ n2 70.0 .D N '5 113 Refractive index of solvent Pressure drop in reactor (dyne/cmz) Heat flux (cal/cmz-sec) Heat flux (cal/cmz-sec) Reactor radius (cm) Gas constant Reactor radius (cm) Rate of reaction Temperature (°C) Reference temperature (°C) Temperature at wall of reactor (°C) Velocity in reactor (cm/sec) Maximum velocity in reactor (cm/sec) Extent of reaction Axial direction in reactor Extinction coefficient (liters/cm-mole) Dimensionless length Dimensionless temperature Viscosity (poise) Wave numbers (cm-1) Dimensionless radius Density (g.cm3) APPENDICES 114 APPENDIX A DATA USED TO MODEL THE HYDROLYSIS REACTIONS OF PhSiCl3 AND MeZSiC12 115 APPENDIX A DATA USED TO MODEL THE HYDROLYSIS REACTIONS OF PhSiC13 AND MeZSiCTZ TABLE A-l.--Summary of Initial Run Conditions for the Hydrolysis Reactions of PhSiCl3. Concentration* Ru" (M°‘°S/L‘t°”) 20 21 22 24 25 PhSiCl3 0.0379 0.0672 0.0849 0.0329 0.061 PhSiCl(OH)2 0.0351 0.0278 0.00712 0.0447 0.0242 Total Silane 0.073 0.095 0.0918 0.0776 0.0852 Water 0.215 0.291 0.242 0.259 0.0807 HC1 1.21 1.463 0.338 0.0 0.0 Temperature 273.0°K 273.0°K 273.0°K 273.0°K 273.0°K *Concentrations are based on mixed reactants. 116 117 TABLE A-2.--Absorption and Concentration Values at Various Residence Times for the Hydrolysis Experiments of PhSiC13. Absorption Concentration (Moles/Liter) t(Sec.) 518 cm" 527 cm- 465 cm- PhSiCl3 PhSiC12(0H) PhSi(OH)3 Run 20 initial concentrations (moles/liter): = 0.215, MHCl = 1.21 Cell spac1ng = 0.116 cm. 0.066 trace 0.272 0.024 trace 0.05997 0.008693 0.165 0.0 0.223 0.028 0.0 0.0492 0.01014 0.231 0.0 0.199 0.035 0.0 0.0439 0.0127 0.297 0.0 0.191 0.045 0.0 0.0421 0.0163 0.429 0.0 0.142 0.090 0.0 0.0313 0.326 0.561 0.0 0.118 0.095 0.0 0.0260 0.0344 0.752 0.0 0.084 0.095 0.0 0.0186 0.0344 Run 21 initial concentrations (moles/liter): = 0.291, MHC1:=1'463 Cell spac1ng = 0.0113 cm. 0.066 trace 0.322 0.013 trace 0.07275 0.004825 0.159 0.0 0.276 0.046 0.0 0.0624 0.0171 0.224 0.0 0.273 0.049 0.0 0.0617 0.0182 0.291 0.0 0.225 0.053 0.0 0.0508 0.0197 0.357 0.0 0.226 0.075 0.0 0.0511 0.02784 0.423 0.0 0.174 0.085 0.0 0.0393 0.03155 0.489 0.0 0.155 0.090 0.0 0.0350 0.0334 0.753 0.0 0.105 0.122 0.0 0.0237 0.0453 118 TABLE A-2.--Continued. é Absorption Concentration (Moles/Liter) E (Sec.) 1 1 518 cm" 527 cm' 455 cm” PhSiC13 PhSiClZ(OH) PhSi(OH)3 Run 22 initial concentrations (moles/liter): MH 0==0.242, MHC1:=0'338 Cell spacing = 0.0154 cm. 2 0.066 0.0 0.456 0.033 0.0 0.0757 0.009 0.159 0.0 0.351 0.082 0.0 0.0583 0.02237 0.198 0.0 0.286 0.096 0.0 0.0465 0.0262 0.291 0.0 0.192 0.135 0.0 0.0319 0.0368 0.423 0.0 0.222 0.161 0.0 0.03687 0.0439 0.489 0.0 0.207 0.184 0.0 0.0344 0.0502 0.621 0.0 0.144 0.210 0.0 0.0239 0.0573 0.753 0.0 0.119 0.211 0.0 0.01976 0.0576 0.951 0.0 0.126 0.220 0.0 0.0209 0.0600 1.083 0.0 0.054 0.218 0.0 0.00897 0.0595 Run 24 initial concentrations (moles/liter): H 0 = 0.259, MHCl = 0.0 Cell spac1ng = 0.0119 cm. 2 0.066 0.0 0.189 0.076 0.0 0.0406 0.0268 0.1387 0.0 0.129 0.135 0.0 0.0277 0.0477 0.217 0.0 0.056 0.156 0.0 0.0120 0.0551 0.299 0.0 0.036 0.176 0.0 0.0077 0.0621 0.386 0.0 0.020 0.183 0.0 0.00429 0.0646 0.482 0.0 0.015 0.206 0.0 0.00322 0.0727 0.614 0.0 0.014 0.209 0.0 0.0301 0.0737 1.926 0.0 0.010 0.218 0.0 0.00215 0.077 119 TABLE A-2.--Continued. Absorption Concentration (Moles/Liter) 5 (Sec.) 1 1 518 cm" 527 cm‘ 455 cm‘ PhSiCl PhSiC12(OH) PhSi(OH)3 3 Run 26 initial concentrations (moles/liter): MH 0 =0.0807, MHC1:=0‘0 Cell spacing = .0121 cm 2 0.066 0.0 0.383 0.0010 0.0 0.0812 0.000349 0.1023 0.0 0.373 0.0028 0.0 0.079 0.000573 0.168 0.0 0.36 0.00563 0.0 0.0763 0.00196 0.234 0.0 0.3096 0.0105 0.0 0.0656 0.00366 0.300 0.0 0.322 0.0164 0.0 0.0682 0.00572 0.366 0.0 0.336 0.0136 0.0 0.0712 0.00474 0.433 0.0 0.342 0.0206 0.0 0.0725 0.00718 120 TABLE A-3.--Summary of Initial Run Conditions for the Hydrolysis Reactions of PhSiCl3 in Acetonitrile. Concentration* Run (Moles/Liter). 16 17 PhSiCl3 0.0669 0.064 PhSiC12(OH) 0.0032 0.0046 Total Silane 0.0701 0.069 H20 0.218 0.22 Temperature 273.0°K 273.0°K *Concentrations are based on mixed reactants. 121 TABLE A-4.--Absorption and Concentration Values at Various Residence Times for the Hydrolysis Reactions of PhSiCl3 in Acetoni- trile. Absorption Concentration (Moles/Liter) E (Sec.) 1 1 1 518 cm- 527 cm- 465 cm- PhSiCl PhSiClz(0H) PhSi(OH)3 3 Run 16 initial concentrations (moles/liter):' MH Cell spacing = 0.0121 cm. 0.0735 0.0735 0.147 0.220 0.294 0.441 0.6613 1.0286 2.79 Run 17 initial concentrations 0.076 0.088 0.028 0.026 0.014 0.001 0.0 0.0 0.0 0.049 0.068 0.073 0.051 0.055 0.033 0.01 0.0 0.0 Cell spacing = 0.0123 cm. 0.095 0.142 0.190 0.237 0.248 0.379 0.474 0.569 0.663 0.853 1.396 O .059 .022 .038 .055 .055 .035 .030 .024 .030 .001 0000000000 0 0.112 0.085 .097 .098 .104 .057 .043 .037 .001 000000000 0 .065 . 0.136 0.133 0.146 0.171 0.158 0.181 0.195 0.207 0.216 0.141 0.163 0.005 0.005 0.002 trace 0.0 0.0 0.0 (moles/liter): M 0.144 0.223 0.210 0.188 0.190 0.214 0.220 0.232 0.227 0.237 0.242 .0107 .0040 .0069 .0100 .0100 .0064 .0055 .0044 .0055 trace 0.0 OOOOOOOOO = 218, M 11 000000000 OOOOOOOOO .0103 .0144 .0154 .0110 .0116 .0070 .0021 .0 .0 0.22, M .023 .0176 .0201 .0202 .0216 .0118 .0135 .0089 .0114 trace 0. 0 HC1 HC1 0.0 0.0474 0.0464 0.0509 0.0596 0.0551 0.0631 0.0680 0.0722 0.0753 = 0.0 0.0494 0.0765 0.0720 0.0645 0.0650 0.0734 0.0755 0.0796 0.0779 0.0813 0.0831 122 TABLE A-5.--Summary of Initial Run Conditions for the Hydrolysis Reactions of MeZSiC12. Concentration* Run ("°‘°5/L‘t°') 37 38 42 44 45 45 49 MeZSiClz 0.1108 0.0937 0.0575 0.0545 0.07 0.0595 0.017 Me25101(0H) 0.0175 0.0148 0.0503 0.0394 0.0279 0.048 0.05125 Me25i(0H)2 0.0 0.0 0.0163 0.011 0.0051 0.0155 0.00884 Total Silane 0.1283 0.109 0.1241 0.106 0.103 0.123 0.077 HC1 0.569 0.4815 0.0 0.0 0.0 0.0 0.0 H20 0.286 0.418 0.323 0.241 0.236 0.195 0.180 TeTpegature 273.0 273.0 267.5 273.0 273.0 273.0 252.0 °K *Concentrations are based on mixed reactants. 123 TABLE A-6.--Absorption and Concentration Values at Various Residence Times for the Hydrolysis Experiments of MeZSiClz. Absorption Concentration (Moles/Liter) ESec.) 1 1 532 cm" 572 cm‘ 510 cm‘ MeZSiClz MeZSiCl(OH) Me25i(0H)2 Run 37 initial concentrations: MH 0 = 0.320, MHCl = 0.569; temperature = 273.0°K; cell spacing = 0.0115 cm.2 0.0537 0.038 0.141 0.021 0.025 0.0179 0.1587 0.018 0.149 0.029 0.0123 0.108 0.0247 0.2951 0.0085 0.130 0.325 0.0058 0.094 0.0277 0.4234 0.001 0 123 0.040 0.0007 0.089 0.0341 0.5781 0.0 0.128 0 035 0.0 0.0925 0.03059 1.187 0.0. 0.125 0.038 0.0 0.0904 0.0324 1.559' 0.0 0.131 0.039 0.0 0.0947 0.0333 1.888 0.0 0.130 0.040 0.0 0.0940 0.0341 2.524 0.0 0.129 0.039 0.0 0 0933 0.0333 Run 38 initial concentrations: M 273.0°K; cell spacing = 0.0115 cm. 0. 0. .001 0.0913 0.333 0.6985 1.410 1.794 2.251 2.707 3.62 0 0 0. 0 0 0 000°C) 01 003 0.127 0.121 0.121 0.128 0.128 0.127 0.132 0.131 0.017 0.026 0.039 0.040 0.039 0.041 0.035 0.037 0.0068 0.0020 0.0007 0.0 0.0 0.0 0.0 0.0 0.102 0.0922 0.0878 0.0878 0.0929 0.0929 0.0922 0.0958 0.0951 H20 = 0.459, MHCl = 0.4815; temperature = 0.0145 0.0222 0.0333 0.0342 0.0333 0.0350 0.0299 0.0316 124 TABLE A-6.--Continued. Absorption Concentration (Moles/Liter) {(Sec.) 1 532 cm‘1 572 cm' 510 cm“ MeZSiClz MeZSiCl(OH) Me25i(0H)2 Run 42 initial concentrations: MH20==0.180, MHc1==0.0; temperature = 267.5°K; cell spacing = 0.010 cm. 0.121 0.121 0.161 0.3565 0.484 0.611 0.930 2.39 2.62 Run 44 initial concentrations: 273.0°K; cell spacing = 0.00961 cm. 0.0156 0.202 0.265 0.327 0.389 0.452 0.826 1.20 1.823 2.1342 0 0 0 0 0 0 0 0 0 0 .01127 .01127 .0081 .009 .0 .0 .0 .0 .0 .0137 .00888 .00478 .001 .0032 .0 0000 0.0918 0.0918 0.0931 0.0988 0.09503 0.0962 0.0956 0.0954 0.0954 0.0746 0.0729 0.0709 0.0747 0.0726 0.0749 0.0726 0.0742 0.0746 0.0765 0.032 0.0322 0.0300 0.0472 0.0565 0.057 0.0583 0.0589 0.0585 MH20 0.0172 0.0282 0.0287 0.0344 0.0308 0.0386 0.0450 0.0455 0.0450 0.0448 0.00606 0.00529 0.00388 0.00274 0.00139 0.00689 0.00767 0.0 0.0 = .241, MHC] = 0.0112 0.0073 0.0039 0.0008 0.0026 0.0 0.0 0.0 0.0 0.0 0.0744 0.0744 0.0754 0.0801 0.0771 0.0781 0.0776 0.0774 0.0774 0.0304 0.0306 0.0285 0.0450 0.0538 0.0543 0.0555 0.0561 0.0553 0.0; temperature = 0.0648 0.0633 0.0616 (0.0549 0.063 0.065 0.063 0.0644 0.0647 0.0663 0.0175 0.0287 0.0292 0.0350 0.0313 0.0392 0.0457 0.0462 0.0457 0.0455 125 TABLE A-6.--Continued. Q—b Absorption Concentration (Moles/Liter) t (Sec.) 1 1 532 cm'1 572 cm" 510 cm' Me25i012 MeZSiCl(0H) MeZSi(OH)2 Run 45 initial concentrations: MH20==.236 M, MHCl = 0.0; temperature = 273.0°K; cell spacing = 0.0098 cm. _ 0.0921 0.0921 0.143 0.207 0.334 0.416 0.588 0.715 1.224 1.86 0.0348 0.0274 0.0162 0.0172 0.0140 0.0155 0.013 0.0052 0.013 0.006 0.065 0.066 0.0687 0.065 0.070 0.075 0.074 0.0701 0.0672 0.0669 0.025 0.028 0.0315 0.0303 0.0313 0.0300 0.0318 0.034 0.046 0.0517 .0279 0.0554 0.0562 0.0585 0.0554 0.0597 0.0640 0.0631 0.0598 0.0573 0.0570 0.0250 0.0280 0.0315 0.0303 0.0313 0.0300 0.0318 0.0340 0.0460 0.0517 Run 46 initial concentrations: MH20==0.195 M, MHC1:=0'03 temperature = 273.0°K; cell spacing = 0.0105 cm. 0.0792 0.0337 0.105 0.02365 0.0792 0.0356 0.0993 0.0241 0.142 0.03128 0.1063 0.02859 0.206 0.2592 0.1091 0.03192 0.3328 0.02141 0.0975 0.02556 0.4596 0.02158 0.09832 0.0275 0.5864 0.0158 0.0967 0.0430 0.0252 0.0266 0.0234 0.0194 0.0160 0.0161 0.0122 0.0835 0.0790 0.0846 0.0868 0.0776 0.0783 0.0770 0.0220 0.0224 0.0266 0.0297 0.0238 0.0232 0.0363 126 TABLE A-6.--Continued. Absorption Concentration (Moles/Liter) 1: (Sec.) 1 1 1 532 cm- 572 cm- 510 cm- MeZSiClz MeZSiC1(0H) MezSi(0H)2 Run 49 initial concentrations: MH20==0.180, MHCl = 0.0; temperature = 262.0°K; cell spacing = .0104 cm. 0.158 0.0242 0.211 0.0453 0.00643 0.0596 0.015 0.281 0.0124 0.203 0.0515 0.00329 0.0573 0.017 0.3456 0.0112 0.194 0.0548 0.00297 0.0548 0.0160 0.410 0.0127 0.206 0.0527 0.00337 0.0582 0.0154 0.5394 0.0124 0.203 0.0584 0.00329 0.0574 0.0171 0.8721 0.0131 0.207 0.0605 0.00348 0.0585 0.0177 1.315 0.00796 0.202 0.0623 0.00211 0.0571 0.0182 1.638 0.00351 0.203 0.0577 0.00093 0.0574 0.0169 1.961 0.001 0.206 0.07328 0.00026 0.0582 0.0215 2.284 0.001 0.204 0.0707 0.00026 0.0576 0.0207 3.059 0.0 0.205 0.07945 0.0 0.0579 0.0233 4.803 0.0 0.190 0.09938 0.0 0.0549 0.0291 TABLE A—7.--Absorbance Data Obtained During the Condensation Reaction of PhSi(OH)3. Time (sec.) 0.0 0.195 0.30 0.49 0.525 0.945 1.29 Absorbance of 1.6 PhSi(OH)3 0.64 .608 1.31 1.12 0.735 1.01 APPENDIX B NUMERICAL METHOD AND FORTRAN PROGRAM USED TO ANALYZE A LAMINAR FLOW REACTOR WITH HEAT TRANSFER 127 APPENDIX B NUMERICAL METHOD AND FORTRAN PROGRAM USED TO ANALYZE A LAMINAR FLOW REACTOR WITH HEAT TRANSFER As developed in Chapter V, the differential equations describing the temperature and concentration profiles in laminar flow are: 2%: -—£——2—e'Y/6 (1 - X)N (3.1) (1 - E) (1 - 521-99 -[(1- x1” e‘Y/°] =9- 99+ 5 32° (8 2) an E SE 3E2 ° with the following boundary conditions: B.C. l for equation (8.1) at n = O, X = O for all E . p C Tw B.C. l for equat1on (8.2) at n = 0, 0 = 1:3fi763-f0r all E o C T B.C. 2 for equation (8.2) at E II _I o _ W O-TTA-Eyqfor 311 TI 30_ 0, 5E—- 0 for all n B.C. 3 for equation (8.2) at E 128 . 129 Here, an explicit method is developed to integrate these partial differential equations. The partial derivatives may be approximated by the following finite-difference formulas: a_x_ . 3+1.i 3.1 n M (8.3) 8 . - 8 £2 = J+Isl .151 8. - 8 . §9_= J.i+l 3.1 320 = ej,i+1 ' 291.1 + 63.1-1 (3 5) 85 AEZ Substituting these expressions into equations (8.1) and (8.2), we have: 0. . - 0. . -y/0. . _ 2 j+1,l 3,1] = _ J,1 (l E )1 An 8(1 Xj,i) e r + g 9.1.81 ' 9.1.11 E 45 J 0. . - 20. - + + €[t191'1'1 .3912 331-11 (3.7) 05 J x. . - x. -Y/9- . 3+1.;n3.1 _ ___£L?_.e 3,1 (1 - x)N (8.8) 130 Solving for ej+l,i and xj+l,i: -y/8. . 8‘1””. = [An/(l -€2)] * [B * (l - xj,i)2] *e J" + (a/E) * [(8j,i+] - 8.,1)/A§] + a * (83.,m - 2 * 8. . J 2 + ej’1_])/A€ ) + ej,‘i -y/8. . Xj+1,1- = [(An * 8w - €2)]*e 3" * (1 - x“)N ” xjn' 3,1 (8.9) (B.l0) The following constants were used to model the reactions in the laminar flow reactor. radius Constant Symbol Value Units Derivation Heat capacity CD .438 cal/g.°C l,2-dimethoxyethane Density p .8683 g/ml l,2-dimethoxyethane Initial 0 temperature To 273.0 K system H§2§c3§on AH -lOOO0.0 cal/g-mole bond energy differ- ences = -7000 call gmole for 1 Cl replaced by 1 0H Initial con- . centration Co .1 g-mole/liter system 2 Thermal con- cal/sec cm ductivity K .00033 °C/cm ethyl ether Maximum velocity Vm 34. cm/sec system Reactor R .108 cm system l3] Constant Symbol Value Units Derivation half life of .5 sec. Fgggggncy Ko l.4)<108 mole"]sec'1 for second order reaction Activation range of experi- energy AE 10000'0 cal/g-mole mental values Gas constant RC 1.987 cal/g-mole/°C constant Dimensionless parameters for equations (8.9) and (B.l0) used for second order reaction. Constant Derivation Value Symbol hiProgram mum w 103 826 Const (1) temperature -AH *CO ‘ K -4 Alpha -:-—-- 2.363 x 10 ALPHA * 'k * p Cp Vm R R*Ko 5 Beta —-——7;jr 6.44 x 10 BETA Vm Co AE*p*C 3 Gamma '———£—— 1.914 x l0 GAMMA RC*AH*CO Temperature THETA (J,I) Extent of reaction X (J’I) Program LAM begins with the input and output of the data necessary for this integration. The sum of the point velocities multiplied by the area is then calculated. This value is used to calculate the mixing cup temperature and the extent of reaction. Next the concentration and temperature profiles are calculated 132 using equations (5.20) and (5.2l). The inner D0 loop calculates the temperature and extent of reaction across the radial direction and the outer DO loop advances along the axial direction of the reactor. The mixing cup temperature and extent of reaction are then calculated and printed. These values are compared to the isothermal plug flow reactor and their ratio to that of the iso- thermal plug flow case printed. m.u.x< .ou .moM.co. oaua .m m.»< mzp mpafiamuao u op can: zozh mza nuumcoxa umurp .aopuama 304a a<..~m.~m.x<..com.-.x..com.-.<»mxp..m.hmzou:mm~wzur_o l- - Abaapnau ~ouaap. paapao. haaz~u coua axx .pazmoc ~c~ wzuh.\m$+0¢¢¢-¢wwflwt*h - (hm004hwuh—UOJM> xqam th t \\. (:13; Cum m2>m AcaNoAO. Ibnmx !: a- -.-Il- : El:l|.!mt>lblmt> Mt$mlbmliillxl ax 9 x0 ax mxo t —x t “atom t > n In!) .Ntouona . o I) u > Iflll! s. -.asflzcm cm Co-.: 1-xllll oN\~xouHx ocnuz> cooumz>m a >h~UUJO> uo< no zo~hta<~odofivx< n N o.N\.A~I<~o~ofi.K o “(moaofivxwu .MWmaofimM . " |.§}l . ll illlmt>m\$KD*~K#ul# wmtnmttmxIaavtt>ta<~wmofi vbw>< u o.N\..~I<—o~ofi. XSINNH< Zo~u¥fi an O l I l- I ll l I - ..... '|-i l Tl:.l,ullll| l OwIOHD o.~\onuH o.~u~ OoOHN ll . I I II I|x | owonmpm NDZ~FZO w13h<1w12mh wo< QZ< ZO—pwaz no hzwbxm no ZOHhzou Lo o~h\m3~a P 1m 2 #2 0 Q 0 D L u o L 0 u up 2 Z O I—IQJuJ III. K IF h H fix 4 : bu a m Mu w I II -I h pp h .‘leSVMu'cflIw- Sample Output HIGAN STRTE UNI V. @IIIII IIIII IIII IIIIII IIII II IIIIIII IIII IIIIIII I III