VAPOR PRESSURES OF ACETQNITRILE SOLUTIONS Thesis Ivar m Degree of Ph. D. MICHIGAN STATE UNIVERSITY Wilbert Hui-ton, Jr. 1959 THESIS v0.2. LIBRARY Michigan Stats Univcnity II’HCHIGQN STATE UNIVERSITY (Hip EAST LANSING, MICHIGAN VAPOR PRESSURES 0F ACETONITRILE SOLUTIONS By Wilbert gutton, Jr . AflIESJB Submitted to the School for‘Advanced Graduate Studies of Michigan State University of Agriculture and Applied Science :Ln'partial mlfillment of the requirements for the degree of DOCTOR 0F PHIHBOPHY Department of Chemistry 1959 Gnaflbl l?"9"bo ACKNOWLEDGMENT It is impossible for the author to cite all the members of the faculty and staff and the many graduate students who have given freely of their advice and assistance in matters both academic and personal during his tenure at Michigan State University. To that large number of friends and associates he expresses his sincere thanks. He is especially grateful to Dr. James L. Hall whose help, encouragement and advice have made this manuscript possible. He acknowledges the assistance and suggestions given in matters dealing with the construction of the apparatus by Mr. F. E. Hood and Mr. F. A. Bette. The cooperation, faith and understanding given him by his wife and family have been an inspiration through- out his work. He extends his appreciation for the assistance and experience afforded him by the E. I. du Pont de Neumours and Company, Incorporated, in granted him the du Pont Teaching Fellowship for the 1957-1958 academic year. The last six months of this research was supported by a grant from the National Science Foundation, NSF- 07386. WET-WW ii VITA Wilbert (Bill) Hutton, Jr., was born July 30, 1927 in Denver, Colorado. After his graduation from West Denver High School he enrolled in the University of Denver where he received the Bachelor of Science in Chemistry degree in 1950. He entered the School for Advanced Graduate Studies at Michigan State University that same year. For some five and one-half years he was a.Graduate Teaching Assistant in the Department of Chemistry. His studies were interrupted in 1951; when he was inducted into the armed forces where he served two years with the United States Army Medical Corps. He returned to Michigan State University in 1956. He was honored by being granted the DuPontheaching Assistant- ship for the academic year 1957 to 1958. From March until.September of 1959 he has held a Special Graduate Research Assistantship under a grant from the National Science Foundation. He is a member of Sigma Xi, Phi Lambda Upsilon and the Amer ican Chemical Society . iii VAPOR PRESSURES OF ACETONITRIIE SOLUTIOIIB By Wilbert Hutton, Jr. AN ABSTRACT Submitted to the School for‘Advanced Graduate Studies of Michigan State University of Agriculture and Applied Science in partial fulfillment of the requirements for the degree of _ DOCTOR OF PHILOSOPHY Department of Chemistry Year 1959 spree. {9,34% X. 7% ABSTRACT An attempt was made to measure the vapor pressures of solutions of 1-1 electrolytes in acetonitrile. The activity of the solvent in these solutions was to have been calculated using the Berthelot equation as modified by Lambert and his co-workers for dimeric vapors. From the determined solvent activities, mean activity coefficients for the solute ions were to be calculated using the method of Randall and Longtin. The purpose of this research was threefold: to determine vapor pressure-concentration data for salt .solutions suitable as standards for isopiestic measurements, to determine the limits to which the Debye-.-.~ Hfickel theory could be applied to these solutions, and to gain some insight into the ion association of 1-1 electrolytes in acetonitrile. A mill-type manometer employing the expansion of a brass bellows as the pressure-sensing device was constructed. The instrument utilized an optical lever to magnify the bellows expansion in order to increase its sensitivity. The results of the measurements were ambiguous. The difficulty lay in the construction of the bellows manometer. Measurements were not reproducible nor could any significance be attached to their values. It is believed that more refined methods of machining than are- available in the average machine shop are required to fashion an instru~ ment of the necessary sensitivity. Suggestions are given as to improve- ments that may be made to rectify the inconsistencies in the apparatus. TABLE OF CONTENTS Page I. STATEMENT OFPROBLEM..............‘........'......'..'............ II. INTRODUCTION.................................................. H Source and Purification of Reagents........................ Description of Distillation Apparatus and Its Operation. . . . Preparation of Solutions................................... Vapor Pressure Apparatus................................... Bellows Hanometer and Constant Temperature Bath......... Mercury Manometer....................................... General_Information...........H....._....................... weration of the Vapor Pressure Apparatus.................. IV. DISCUSSION.................................................... v. summmmammmm ammonium WH0.00000000000000.0000...000000000000ooeoeooooooooooooooooeeo vi 1 2 25 28 31 - 3h 3h 38 39 h0 I h2 55 57 60 TABLE II m IV VII VIII LIST OF TABLES Page Vapor Pressure of Acetonitrile at 25°C........................ Vapor Pressures of Acetonitrile at 25°C....................... VaporoPressures of 0.1th Molal Potassium Thiocyanate Solution at 25 COOOOOOOOOOOOOOOOOOOOOOOOOO000.0000...COOOOOOOOOOOCO‘OOO. VaporoPressures of 0.113 Halal Potassium Thiocyanate Solution' at as COOKOOOOOOOOOOO0..0......0..OOOOOOOCOOOOOOOOOOOOOO.00...: VaporOPressures off-0.0175 Molal'Potassium ‘Thiocyanate Solution at 25 00.0,...OOOOOOOOOOOOOOOOOOOOOOOOOOOOO'OOOOOO0.0.00.0.0... J VaporoPressures of 0.0239 101.91 Potassiu- miocyenate Solution at 25 0......OOOO'OOOOOOOOOOS000.00...0.COOPOOOOOOOOOOCOOOOOOOOO Vapor Pressures of 0.111; Molal Potassium 'Thiocyanate Solution' at 25 OOOOOOOOOOOOOO....0..00..O.0.00.00ODOOOOOOOOOOOOOOOOOIOOO Vapor oPressures of 0.1014 Holal Potassium Thiocyanate Solution at 25 COO-OOOOCOOOOOOOO0.0.00000000000000000000000000000'OOOOOfO Vapor Pressures 'of 0.0570 Holal Pptassium 'Thiocyanate Solutign at 25 GOD‘S-"0.00.0.0.0..0000000.000.000.000...OOOOOOOOOOOCOO0.. Vapor‘Pressures of ‘0 .0147? 'Molal Potassium 'Iodide Solution‘at ' 25 GO.OOONOOOOODOOOOOOOOOOOOOOOOOO0.0.00.0...OOOOOOOOOOOOOOO... Vapor Pressures 'of 0.0937 'Molal Potassium 'Iodide Solution 'at ' 25 OOOOOOOOOOQOOOOOOOOOOOOO0.00000000000000000000.00.00.00.00. vii 10 1:5 1&5 ho 1:6 h? ’47 h8 1:8 I49 19 FICHJRE 1 2 Fifitéli'sxom-soxmz—u 20 21 LIST OF FIGURES Distillation Apparatus........................................ Complete Vapor Pressure Apparatus.............................. Vapor Pressure Flask Cross-Sectional View of the Bellows Manometer................. Bellows nanometer Assembly-External View..................... Bellows Manometer Housing..................................... External Mirror Support....................................... hternal Mirror Support Ring.................................. Bellows Housing Base Ring..........................'........... Bellows Housing Retaining Ring................................ Manometer Base...............................6................. Upper Retaining Screw.......................................'.. Bellows Support............................................... Lens Mount.................................................... Mirror Platformand Groove.................................... Mirror Pivot Base and Needle Guide............................ Mirror Comartment...................,................_......... Details of Mirror Chamber Assembly............................ External. Mirror Assenbly...................................... External Mirror Bracket....................................... Detail of hternal mrror Adjustment Mechanism................- .. viii Page 27 32 33 35 37 6h 65 65 66 66 67 68 70 70 7O 73 73 76 77 77 STATEMENT OF PROBLEM The demands for new materials being made by our rapidly advancing technology has accelerated interest in nonaqueous solvents as media for inorganic reactions. Of the many substances finding use as nonaqueous solvents few have been investigated extensively so that the full range of their potential usefulness cannot be evaluated. me present knowl- edge of solutions of electrolytes in such media is quite limited and there is a need for much more fundamental data for these systems before any successful theory can be developed. 2 It was the purpose of this research to investigate the vapor pressures of solutions of electrolytes in acetonitrils. It was hoped that the results of such measurements would aid in future studies of reactions in this solvent by contributing toward the following: the determination of concentration limits to which the theories for aqueous solutions of electrolytes may be applied in this solvent; the develop- ment of a satisfactory series of vapor pressure standards suitable for further vapor pressure measurements by an isopiestic method; the applic- ability of vapor pressure measurements as an indication of the extent and nature of ion association in this solvent. INTRODUCTION A study of the nature of solutions of a solute in any solvent revolves around the determination of the changes in chemical potential (Gibb's partial molal free energy) with concentration. Experimentally this is usually done through the measurement of the activity of the solvent and the calculation of the activity of the solute using the Gibbs-Duhem equation or the reverse, i.e., measurement of the activity of the solute and calculation of the activity of the solvent. Several textbooks (1,2,3) have excellent discussions of the various methods of determining activity coefficients. These discussions are chiefly con- cerned with aqueous solutions but the same principles can be applied to nonaqueous solutions with certain modifications . 0f the many methods of measuring activity coefficients of electro- lytes, those employing vapor pressure measurements seem attractive to a study of acetonitrile solutions. Vapor pressure data should furnish an excellent means of investigating the nature of concentrated solutions since they can be taken over a wide range of concentration and temperature. Methods of measuring vapor pressures have been classified intothree general categories (1): the static method, the dynamic method, andsthe isopiestic method. A static method consists of allowing the vapor of the solutionsto pass into some suitable manometric device where its pressure is measured. As the particular method used in this research is of this category, the application and limitations of such a procedure shall be discussed in detail later. A dynamic method operates on the principle that when a solvent- free, inert gas is passed successively through a solvent, a solvent absorber, an aqueous solution, and a second absorber, then, under the proper emerimental conditions, the amount of vapor absorbed in the - first absorber is prOportional to the vapor pressure of the solvent; the amount of vapor absorbed in the second absorber is proportional to the vapor pressure of the solution. Bechtold and Newton (h) and Vashburn and Heuse (5) have used such a method in studies of aqueous system with excellent results . The isopiestic method is perhaps the most simple and has been the one most extensively used. It depends upon the principle that two solu- tions of nonvolatile solutes will distil from one to another until the vapor pressures of both solutions become equal. This is a comparative method in that the vapor pressure-concentration curve of some reference electrolyte must be known for the determination of the vapor pressure of the unknown solution. This method suffers from the fact that it gives values for activity coefficients to an accuracy determined bythe accuracy of the reference solution. However, very careful measurements of aqueous solutions of sodium chloride, potassium chloride and sulfuric acid have been made by static methods and these results closely evaluated by... Robinson and Harned (6) am! others. As a result, standards are very well established for gqgems solutions and cover a wide concentration range. Another objection to this method is that the more dilute solutions require a considerable time to reach equilibrium--in many instances several weeks are required (7). However, the results of studies of Baldes and Johnson (8), Brady, 9_t_ a1. (9) and Higuchi and his co-Iworkers (10) appear to offer a method of minimizing this difficulty. These workers report devices for measuring the vapor pressures of solutions using a technique employing matched thermistors. The measurement is based upon the cooling effect as solvent evaporates from small samples of the solution as compared to a reference. The small samples required and the application of a recording device to indicate when equilibrium is reached have shortened the time necessary to make a determination. The measurements of vapor pressures of solutions of inorganic salts in nonaqueous solvents have been few in number. Tower and Germann (ll) measured the vapor pressures of salts in several organic solvents and calculated apparent formula weights of solutes at various concentrations. Lithium chloride and potassium chloride, gave values less thanthose calculated from the simple formula in methanol and ethanol but gavewalues greater than the true formula weights in other organic solvents.-. They report the formula weights remain relatively constant with changes in dilution and interpret this as indicating that the degree of dissociation remains about the same in dilute and in concentrated solutions. bullioecopic measurements of Jones (12) on potassium iodide solutions in methanol and ethanol confirm the above results. Larsen am Hunt (13) measured vapor pressures of several amoniun salts in liquid ammonia at 25°C. They determined the relative activity of the solutes and found the deviations from Raoult's law to be quite pronounced in the dilute as well as in the concentrated range. It was suggested these deviations were due to the solutes' being associated and .the resulting ions highly solvated. Ritchey and Hunt (114) extended this work and determined the activity coefficients of ammonium chloride in liquid ammonia at 25°C. These workers measured the vapor pressure dif- ference between pure liquid ammonia and the solutions with a mercury manometer equipped with a traveling microscope for recording the positions of the menisci. Pressure differences were measured to 0.01 cm. The activity coefficients were calculated using a modification of ..the method of Randall and White (15). Application of the .Debye-Hfickel theory to .these solutions indicated that the theory held satisfactorilyto concen- trations of 0.(X)l molal and below. Abe (16) measured the vapor pressure of saturated ammonium chloride solutions; in liquid ammonia at temperatures from -20.00° to 7.1306. uni calculated the heat of solution for this salt. Shatenshtein and Believe (1?) determined the osmotic coefficients of liquid ammonia solutions of ammonium acetate, ammonium chloride, amonium bromide, amonium iodide, amenium nitrate, sodium chloride and sodium nitrate. The low values they obtained were attributed as being the result of the low dielectric constant of the solvent. The osmotic coefficients increased as didnthe dielectric constant and the conductivity coefficients with the lowering of the temperature. The vapor pressure measurements of Linhard (18) in liquid ammonia solutions of the alkali halides and alkaline earth halides indicated a high degree of ionic association and solvation of the ions. Solvation was noted to increase in the order: Cl'< Br- < I" ancl Cs‘ < Rb. < K... < Na... Solubility increased in the same order. Strontium nitrate and barium nitrate were observed to give abnormally high vapor pressure depressions. This was attributed as being due primarily to a high lattice energy rather than excessive solvation. Vapor pressure studies of solutions of aluminum chloride in liquid phosgene have been made by Germann and McIntyre (19) at temperatures of 0° and 25°C. Oosaka (20) utilized a dynamic air-saturation.method to study the vapor pressures of sodium iodide solutions in acetone. The temperature range of 10° to 50°C. was investigated. The compouni Hal-3(cH3),co. separated from the saturated solution at 25.500. Above this temperature it was observed to change to NaI. Deviations from Raoult's law were quite apparent. I - An.interesting study of solutions of lithium perchlorate in.diethyl ether was reported by Ekelin.and Sillén (21). Calculations of activity coefficients‘were made and plots of these values against the corresponding values of the mole fraction of solute gave a curve showing a definite maximum. .Attempts to apply the theory of’Hdgfelt (22) for characteriz- ing the nature of the specie present at the maximum were not too satis- factory. However, the results seemed to indicate the formation of lithium perchlorate-~ether complexes. Eversole and Hanson (23) investigated solutions of potassium iodide in liquid sulfur dioxide. Vapor pressure data were used to calculate the activity of the solute using the method of Randall and Longtin (2h). The limiting values for the function hi/g at infinite dilution were calculated from the ‘Debye limiting law for activity coefficients using the equation given by Eversole and Hart (25). Three distinct regions were observed in the plot of log EN (N m mol fraction of solute) against Hg. The observance of a minimum at E - 0.20 was attributed to the same phenomenon observed in aqueous solutions, i.e., solvation of increasing amounts of solute which tends to immobilize considerable solvent and as a result, increases the activity of the solute. Reports of previous workers on the isolation of the sulfur dioxide addition product with potassium iodide (KIthOz) seem to add credence to this reasoning. An attenmt was made to investigate the probability of ion association occurring in this solvent of dielectric constant ranging from lh.53 to 11.7 over the temperature range studied. Calculations of Elf/.1; values were made assuming association and subsequent dissociation of the ion- pair (KI)2 :4, x’ e KI; or 1:21“ o 1'. These values did not approach the limiting values as well as those calculated assuming dissociation of a simple binary salt. Selection of the Method for Determinigg Vapor Pressures of Aceto- nitrile Solutions. The absence of amr reliable concentration-vapor . pressure data for solutions of acetonitrile makes it impossible to employ the isopiestic technique. A dynamic method requires an efficient means of absorbing the solvent vapors from a gas stream. The success of this method in aqueous systems is largely due to the very excellent chemical desicfints such as magnesium perchlorate. A satisfactory absorbent for acetonitrile is not available and the only apparent recourse would be to remove the actonitrile vapors from the gas stream by condensation. It was felt that the difficulties in the construction and the operation of equipment to perform these operations to the high degree of accuracy desired would be considerable. A static method for determining the vapor pressures was selected as it seemed likely to present fewer problems in the construction of apparatus while the very excellent measurements obtained for aqueous solutions using this method seemed to recommend it for basic studies of the type proposed. Excellent early measurements of the lowering of the vapor pressure of water by dissolved salts were made by Frazer and Lovelace and their co-workers (26,27). Their apparatus was carefully constructed, and using highly refined techniques they measured the vapor pressures of. solutions of the nonelectrolyte, mannite, and of potassium chloride in aqueous solutions. Vapor pressure lowerings to 0.001 mm. of mercury were measured using a Rayleigh manometer (28) and a temperature control of i 0.001°0. was maintained. Gibson and Adams (29) employed an apparatus of simple construction designed to allow the measurement of not only the difference in vapor pressures between solvent and solute ( A p), but also the absolute vapor pressures of the pure solvent (20) and of the solution (p). these data enable calculations of the activity of water (p/po) to be made in three ways: 13/12. 3 (pa-Apro 3 p/(p *Ap)» They obtained agreement between the solvent activity calculated from each of the above relations to three parts in fifty-seven hundred. The accuracy of their apparatus was enhanced by using nabutyl pthalate (density 2: 1.0h18 g./cc. at 25°C.) as the manometer fluid instead of mercury. This liquid not only has a vapor pressure less than mercury, its low density results in a displacement in the manometer arm much larger than would be given by mercury. They reported solvent activities for aqueous solutions of lithium chloride, potassium thiocyanate, ammonium chloride and sodium chloride which have been shown to be in good agreement with other measurements . Shankman and Gordon (30) measured the vapor pressures of aqueous solutions of sulfuric acid using a similar apparatus but enploying Cenco- Hyvac pump oil (density at 25°C. is 0.895 g./cc.) as the manometer fluid. These results are among the most accurate to date. The high order of success using this instrument made it the first choice as an apparatus to be considered for use in measurments of acetonitrile solutions. The outstanding advantages of the apparatus of interest here are the increased accuracy and sensitivity to be gained from a manometer fluid less dense than mercury, the relative simplicity of the apparatus, and its versatility in affording means for measurement of the absolute values of the pressure of the solvent and the solution. One of the main problems in the use of the above apparatus was. to find a suitable manometer liquid. This becomes a bit more difficult when dealing with acetonitrile than it is when measuring aqueous vapor pressures- Many of the oils showing no solubility in water show some 10 solubility in acetonitrile and are not applicable. Another limiting factor, perhaps the most serious, is that although the vapor pressure of acetonitrile is somewhat uncertain it can be said to be in the neighborhood of nine centimeters of mercury, a value much higher than the vapor pressure of water at 25°C. (2.3756 cm. of Hg). (See Table I for a partial list of vapor pressures reported for acetonitrile.) The result of -this is that in order to utilize a manometer of a convenient size (less than one meter) the density of the manometer fluid must be at least 1.22..g./cc., using 13.5 g./cc. as the density of mercury at 25°C. A density of about two would permit a manometer of a convenient size for immersion in a thermostated bath. Many promising liquids including fluorocarbons and silicone oils were tried but none proved satisfactory. um I VAPOR ramsmus or mums u 25°C. Vapor Pressure, Method of Deriving cm. / Hg . Value Reference 8.887 Calculations from Heim (31) emirical equation fitting emerimental data. 9 .225 Antoine equation Dreisbach and Martin (32) 9.231; Direct measurement Dreisbach (33) 8.833 Antoine equation Brown and Smith (3h) 11 In lieu of the above difficulties, other methods for measuring vapor pressures of solutions were investigated. It was decided, since there is such a discrepancy in the value for the vapor pressure of acetonitrile itself, and because of the lack of dependable vapor pressure-concentration data for is0piestic standards, that it would be of primary importance to measure the absolute pressure of the solution and solvent rather than the pressure differences. Most instruments for measuring large pressures accurately depend upon the mechanical displacement of a suitable membrane or diaphragm and the magnification of this movement to a high degree of accuracy. In some instances movement of the pressure-sensing material is constant enough to be calibrated to allow pressure to be read directly, or, as is usually the case, the instrument is used as a null-reading. device. The glass-spoon gauge or sickel gauge has found many applications especially in the measurement of vapor pressures of corrosive materials (35). It is a rather fragile device and limited in range. A null—reading instrument employing a rubber diaphragm has been constructed by Amsel ani Uittwer (36) which is claimed to measure vapor pressures .in the range 2 x 10" mm. to 760 mmof mercury. The magnification of the move- ment of these pressure~sensing devices has been achieved by optical methods (36) or by utilizing electromechanical transducers (37). East and Kuhn (38) reported construction of instruments employing standard brass bellows as pressure-sensing membranes and optical systems for magnification. 'lheir devices measured pressures in ranges from 0 to 100 mm. down to 0 to 2.5 mm. of mercury. It was claimed that sensitivities to 1/2000 mm. of mercury were obtained in some devices with high magnification. Allen, gt £1. (39) applied one of the instru- ments constructed by East and Kuhn (hO) to the study of vapor pressures of benzene and benzene solutions of diphenyl, dipherwlmethane and dibenzyl. ' The bellows manometer should be quite applicable to measurements of acetonitrile solutions. Used as a null-reading device, it allows the vapor pressures to be read in liquid nanometers without the vapors and the manometer fluid coming in contact. The instrument and vessel contain- ing the solution to be measured could both be immersed in the same thermo- stated bath, and thus any condensation of the vapors between the solution and pressure-measuring device would be eliminated. The liquid manometer would not of necessity be in a thermostated bath so most of the size limitations would be eliminated and added accuracy through the use of manometer fluids of low density could be obtained. The success with the use of brass bellows as pressure-sensing devices encouraged the attempt at constructing an instrument in the machineshop of the Kedzie Chemical Laboratory. If a successful instrument working on this principle couldwbe constructed without need of precision machinery and tools uncommon to the average machine shop it would be of possible value for fundamental pressure measurements of marv materials other. than acetonitrile solutions. With such a device it should be possible to measure vapor pressures of fluorine compounds,many of which, while extremely reactive towards glass and rubber, are somewhat passive toward copper. The standard brass bellows itself is relatively inexpensive and easily 13 available. The various bellows manufacturing conpanies can also supply bellows made of such metals as copper, Monel metal and Inconel on special order. Because of its sturdy construction, an instrument of this type could serve as a standard piece of laboratory equipment for vapor pressure measurements of a wide variety of materials. An attempt was made to construct such an instrument without any special tools or materials other than a toolroom lathe and a vertical milling machine. Magnification of the bellows displacement was tobe attained by a simple optical lever system. Pressures were to be measured using a cathetometer capable of reading to 0.01 mm. A detailed description of the bellows manometer construction is given in the Appendix complete with diagrams. The calculations of activity coefficients from vapor pressure measure— ments of solutions of salts in nonaqueous solvents have been made. in liquid ammonia (13,110 and in liquid sulfur dioxide (23). The method of calculation in this research is to follow the same general procedure as used in the above works. For the calculations of the meanaactivity coefficients of the l-l electrolytes a simple binary dissociation process is assumed. in outline of the method of calculation is presented below. Muation of State for Gaseous Acetonitrile. Acetonitrile vapors at 25°C. deviate considerably from ideality and it is necessary to use fugacities instead of pressures in the thermodynamic treatment. -Fugacities of the vapor may be determined if an equation of state for the vapor is known. Lewis and Randall (hl) illustrate examples of a possible calcu- lation. The only information on an equation of state for acetonitrile recorded in the literature is that of Lambert arxi his co~workers (J42). They studied the behavior of the second virial coefficient, g, of the Berthelot equation of state. Before proceding with this discussion, the results of their work should be mentioned in detail. The Berthelot equation of state can be written in the form PV a n(RT «0 BP) (1) where i i M 61‘ a. __C .. _.Q ? a. r. <1 >~ m A _'I_'c and 2c are the critical temperature (degrees Kelvin) and critical pressure respectively. Lambert determined the values for _B_ from plots of experimentally determined :11 products against 3: for several different substances at various values of the absolute temperature using a nBoyle's law apparatus.” Comparison of the experimentally determined values for- g with those calculated from the critical data and equation 2 showed that substances fall into two classes. Class I included gases where the experimental value of g did not depart significantly from the value calculated from equation 2. In this class were ethane, ethylene, cyclo- hexane, benzene, diethyl ether and ethyl chloride. Class II included gases where the measured values of g were consistently very much higher (over 100 percent at the lower temperatures) than the calculated values. Acetone, methanol and acetaldehyde and acetonitrile fell into this classification. They found that the equation a g - pfz satisfied the data in the -2 range of temperatures considered (5 << b_'1_‘ ). Values of the constants 15 _a and _b_ were calculated from the critical data and were found to give a good empirical fit with the measured values of _B_ for the nonpolar sub- stances in Class I. As a result of their investigations it was concluded that values of _B_ from equation 2 apply satisfactorily to the dipole interactions or ”generalized" intermolecular attraction (i.e.., inter- actions not of sufficient energy for dimerization) in all molecules. In the case of the polar molecules of Class II where dimerization occurs the second virial coefficient can be expressed by the effects of dimeri- zation superimposed on :3. Application of the effect of dimerization on the second virial co- efficient as treated by Schafer and F02 Gazulla (143) led to an expression for an equation of state for dimerized gases of Class II rv .. n(RT + BOP). (3) _B_o is the second virial coefficient which is obtained experimentally. Considering the dimerization equilibrium ——-9 ‘2 «___ 2A and its dissociation constant, _Igp, where x 3/ P " DA pAz they derived the expression for go in terms of 11 BO 9 [B - R'r/Kp] - (h) 5? can be calculated from measured values for go and values of _B_ calculated using equation 2 and the critical constants. Plots of values for 1gp calculated from the measurements on acetonitrile against _l/T fall on a straight line. Calculation of the Fugacity of Acetonitrile Vapors. The volume of one mole of an associated, nonideal vapor can be expressed from equation 3 as vzé-m-fi Bo (5) where _v is the molal volume. It has been shown (hl) that the fugacity, _f, of an imperfect gas is related to its pressureby the equation Rl‘dlnfnvdp. (6) Substituting the value for z from equation 5 gives Rlenfa(-§r-¢ Bo)dp. (7) Integrating between the limits 2 e- _g to 2 up gives P , 0 Then substituting for _B_.o the expression given in equation )4 where _s is expressed in terms (of '_1j by the empirical equation 2 Beaobr" (9) results in P‘ amt-"2 1 Infalnp§SO(-Er——fi—)dp. (10) Integration gives .2. .2- P. -_P_. 2.303.10gp p(RT RT3 K). (11) 17 Upon substitution of the values for acetonitrile as given by Lambert _e_t ale (he): a s 66.3, p e 11.93 x 107 and gp .. h-.87 at 25°C. results in log(f/p) - (Coll—17» (12) Calculation of the Activity of the Acetonitrile in Salt Solutions. The activity of the solvent, 31, is by definition 3.1: f/-fo (13) where _f_'o is the fugacity of the pure solvent and f is the fugacity of the solution at some concentration. These can be calculated from equation 12. _ , Calculation of the Jigtivity of the_§:olute. In this discussion the following symbols are used: an is the molal concentration, i.e., the number of gram formula weights of salt per 1000 g. of solvent K - lab/Ml, where H; is the molecular weight of the solvent. Ve and v- are the rumber of positive and negative ions formed by the dissociation of a salt formula unit. v- v++v_ 31 . activity of the solvent as determined by vapor pressure measurements. N1 - K/(K o'Dm) - 1/(ler) - mole fraction of solvent. r - . N/Nl 5- (vm)/K up mole ratio of solute ions to solvent molecules. N - film/(K e Vm) - mole fraction of solute ions. H is the formula weights of salt per 1000 cc. of solution - (1000p r)/(Mi1’ + Her). 18 Q is the density of the solution. Po is the density of the pure solvent. M2 is the formula weight of the salt. ac s. ill/amt. av" is the geometric mean activity of the * e... ions'in solution, the subscript 2 indicating the concen- tration units. Randall and White (15) modified the divergence function, h, of Lewis and Randall (hl) to apply to solutions of electrolytes. Eicpressed in terms of _1; the function becomes h e ln al/r e 1. (lb) Ritchey and Hunt (1)4) used this function to calculate solute . activity coefficients of axmnonium‘chloride solutions in liquid ammonia. Randall and Longtin (21;) have defined a similar function, _h_' , based on mole fractions to be h' . (1n a./N1)/r. (15) Eversole and Hanson (23) applied this function to the calculation of activity coefficients of potassium iodide in liquid sulfur dioxide. The I'special devices“ necessary for application of these two functions to the calculation of activity coefficients of strong electrolytes have been discussed in a paper by Eversole and Hart (25). This paper also includes derived equations relating the solute activities and activity coefficients for different concentration scales. Because the _h' function appears to lend itself to a more accurate extrapolation to 19 infinite dilution than does the _h function it was chosen for the work presented in this thesis. 4 . It can be easily shown by methods given by Randall and Longtin (21;) that H a - ~ ln f+ a -h' - 2 I (he/rimr‘} (16) O I, - - where _fi - gN/fl. In order to evaluate the terms necessary for this. calculation, the values of _ht/ri (i.e., 324354111. )‘ can be plotted _ 1" .. against 5%. Using the complete curve of hf/ri against ri, _h' ”can be evaluated for a chosen molality by calculating the value of 3% at the desired molality and then multiplying, it by the corresponding value of y/gi. The integral term can be determined by graphical integration between 3% - _Q and 3% . r71“. The areas under the curve can be measured with a planimeter. - The extrapolation of this curve to infinite dilution can be carried out using the .Debye..- limiting law for activity coefficients. Bversole and Hart (25) have also formulated the equations for this purpose. , 'According 'to the .Debyo-Huckel theory (3), in a dilute solution of a single salt, . an (as) - 1 Ali (17) 1 where " 11,22 i' V'.‘ 23 a; A - (9.90 x 102) v1 [——*5.-T-—-1 (18) - . - 1 _ V1 - volume of l g. f. w. of solvent in cc., ,1); . the dielectric constant ofthe solvent, T is the absolute temperature 2 and Z are the valences of the positive and negative ions of e - the salt. 20 It can be shown that A, “Li...“ (19) 11 I'll 4} r142 ' 3 Substituting 19 into 1'] and dividing by 3/2 3 1n al N a, A [ l.___£._000 ] /2 a 11.1. (20) r ‘ r From equation 20 3/ 2 Lim n: . “Egfifia—J (21) . 1 r' ->O 1.1/2 ‘ 2 Evaluating 21 for lwl electrolytes in acetonitrile using the values V1 m 52.80 cm:3 121 an 36.7 at 25°C. (hl) (00 a 0.777 g. per .53 at 25°C. '- 1 one obtains 3.77 as the limiting value at infinite dilution. Differentiating equation 20 it becomes _1 _ M1 ,3 . 1000 e A 2 _d_9/dr _ M _, % dri 3000 A FMJV + rMe ] g Mlv- §_r--M_2 LZ—(v M1 + rM2)2]r (22) o '5 and .1 Lim 1‘11;le 3" 0 if dP/drf-‘m, (23) r2->»0 dr Thus the limiting lepe of the h'/§% curve is zero. A similar calculation of the h/:% curve results in a limiting slope of'g. As the slopes of both of these curves are negative in regions of moderate concen- tration one can assume the slope of the h! function will approach zero at ‘ .1 1This is also the limiting value of h/ge. 21 infinite dilution while the curve using the h function must pass through a maximum in the dilute range. From the values of";i it is possible to calculate the activity co- efficients in other concentration units if desired (2h). The extent of the conformance of the acetonitrile solutions to Debyewflfickle theory may be estimated by comparing the linearity of the curve obtained by plotting l‘nflf2 against}!2 in the dilute region. Accuracy of the extrapolation can.be confirmed by comparing the SJOpe of the curve in this dilute region with that calculated from the Debye limiting law (23). Ion Association. In a study of solutions of electrolytes at various concentrations the concept of ion association must be taken into con- sideration. In very dilute solutions where the pOpulation of solvent molecules is much greater than that of solute ions any individual ion is probably surrounded by an atmosphere of solvent molecules extending out from the ion a distance of several solvent molecules. As the concen- tration of solute ions increases and the ratio of ions to solvent mole- cules increases the extent of this solvation atmosphere or solvation sheath lessens and the ions become closer together. The concept of ion association provides a relatively simple and self-consistent method of dealing with the situation that arises when the ions of opposite charge come close enough that their mutual electrical attraction'becomes greater than their thermal.energy. The result is that they form virtually a new entity in the.solution. In the case of symmetrical electrolytes ion~pairs will have no net charge though they should have a dipole moment. 22 They will not contribute to the electrical conductivity of the solution while their thermodynamic effects on the solution will.be the removal of a certain number of ions from the solution and replacing them with half the number of dipolar molecules. As more ions become involved, higher ionvassociation products can result. Bjerrum (hS), assuming the ions to'be rigid nonpolarizable spheres contained in a medium of fixed dielectric constant, developed a theory of ion association by considering the factors which determine the form- ation of ionwpairs under the influence of coulombic forces. He considered any two ions at a critical distance, 3, to be associated where '2 Z I e2 qs + - (25) thT and'g is the charge on the electron, §* and gr are the charges on the ions,‘g is the Boltzmann constant and,§ is the dielectric constant of the medium. Solvents of lower dielectric constant should favor ionrpair formation to a marked degree. Bjerrum has shown that at the critical distance,lg, the potential energy of the ionrpair is £53 and the energy necessary to separate the pair is of the same order of magnitude as their thermal motion. From the expression for.g, it is apparent that the smaller the dielectric constant the larger this critical distance, and ion pairing can occur even at distances greater than the ordinary ionic distances. 7 .. Fuoss and Kraus have extended the treatment to higher ionrassoci- ation.products such as triple-ions (ho) and quadripoles(h7). 23 The theory has been applied to electrical conductance. with some success in dilute solutions in water and solvents of very low dielectric constants V(h8). The nature of ion-«association products have been in- vestigated in aqueous systems by other means such as spectrophotometric, solubility, electromotive force and to a lesser degree, vapor pressure measurements . Mercier and Kraus (149) tabulated the vapor pressure lowering of _ (CH3)3NHBr in liquid bromine at 25°C. and found vapor pressure lowering much less than that predicted by Raoult’s law in the very dilute solu- tions. it the higher concentrations the vapor pressure lowering increased steadily. This phenomenon was interpreted as showing a high degree of association from ion-pairs up to quadripoles or more complex polyionic structures atgthese low concentrations. Debyémi‘ckel theory is not applied to solutions where ion associ- ation is appreoiable. Moderately low activity coefficients are believed to be an indication of Bjerrum ion-pair formation. Therefore, it is believed data from an accurate study of vapor pressure behavior of . acetonitrile solutions may show some indication as to the existence of ion-pairs and higher ion-association products in this solvent. Such- knowledge will be of importance in understanding and interpreting results of mrther studies of solutions of such salts. A very excellent paper by Kraus (hB) on the ions-association concept reviews the results of much of the work done to date in studying this phenomenon both in aqueous and nonaqueous systems . There is apparently 214 a reversal of the mass law in highly concentrated solutions, and Kraus 'writesz For every salt in any solvent of dielectric constant such that association occurs in less concentrated solutions, there is a concentration at which the degree of association is a maximum. Toward lower concentrations, the degree of association de-‘ creases with descending concentration, toward higher concene tration, it decreases with increasing concentration. Although there is no theory available at this time to account for such behavior, it is mentioned here as a phenomenon that should be present in the more concentrated acetonitrile solutions of ionic salts. 25 WWAL Source and Purification of Reagents All chemicals in this study, except those listed below, were ordinary grades and were used without further purification. Potassium Thioqyanate. Baker and Adamson reagent grade potassium thiocyanate was purified and dehydrated after the procedure of Kolthoff and Lingane (50). The salt was recrystallized from demineralized water then twice from absolute ethanol. After preliminary drying over sulfuric acid the salt was transferred to a vacuum oven at 100°C. The last traces of'water were removed by drying for two hours at 150°C. then for ten minutes at 200°C. The white crystalline product was stored in a tightlyb capped weighing‘bottle over anhydrous magnesium perchlorate. Potassium Iodide. Mallinckrodt analytical reagent grade potassium iodide was recrystallized from demineralized water then dried in a vacuum oven at 100°C. until a constant weight was attained. The white salt was stored in a darkened desiccator over magnesium perchlorate.1 ‘égetonitrile. Matheson Coleman and Bell practical grade acetonitrile (b.p. 80~82°C.) was purified and dehydrated following a procedure similar to that of‘uard (51). The impure reagent was allowed to stand over potassium hydroxide for several days with intermittent shaking to remove any acetic acid formed by the hydrolysis of acetonitrile. The liquid tAll solutions of potassium iodide in acetonitrile were protected from the light and no indication of decomposition was observed. 26 was then predried with anhydrous calcium chloride, decanted into a glass- stOppered flask containing phosphorus pentoxide and refluxed for several days in order to remove the hydrolysis product, acetamide, and most of the water. The refluxed acetonitrile was then distilled into another flask containing fresh phosphorus pentoxide. Only the fractions boiling in the 81-7-8200. range were retained. This distillation process was repeated four times from fresh portions of phosphorus pentoxide.:l After the fifth distillation, the flask containing the acetonitrile over phosphorus pentoxide was transferred to the closed fractional.- distil- lation apparatus shown in Figure 1. Here it was distilled three times, in a manner described later, in an inert, water-free atmosphere. The first and second distillations were made from phosphorus pentoxide and the third from anhydrous potassium carbonate which had been dried at 170°C.2 The initial and final ten-milliliter portions from each distil- lation were discarded. The purified solvent had physicd. properties in agreement with the literature values . 1Walden and Birr (52) report a red-"orange product formed in the refluxing and distillation procedure and state that after six such dis- tillations the product did not yield a color. Smith and Witten (53) found that after ten distillations from fresh phosphorus pentoxide the color continued to form. However, these authors noted no change in physical properties of the distillate after the third distillation. In this work the observations of the latter authors were confirmed; indeed, the red- orange residue formed even after fifteen distillations. It is believed that the red-orange material is a nonvolatile product of a reaction between acetonitrile and phosphorus pentoxide. III‘alden and Birr (52) reported that some of the phosphorus pentoxide was carried over into the distillate with the acetonitrile. The potassium carbonate treatment was to prevent this from taking place. j) = TO M9(Clo4} DRYING TUBE 2 W U FIGURE I ' DISTILLATION APPARATUS 28 Qer imental A .' . I_..i_terature Refractive index: 1130 - 1.31042 , DD?" -.l.3hh11 (32) Boiling point: 81.3%. at 7M; mm. Hg 80.0.00. at 760 m.(32) 81.5 c. at 751 m.(51) ‘1 _s -1 incl!) ohms (51) ..e .. Specific conductance: 2.2 x 10 ohm No indication of water was observed when the copper acetylide test of Weaver (5b.) was made. Description of Distillation Apparatus and Its gperation The apparatus shown in Figure 1 was designed to perform the following functions under a dry, inertatmospherex (1) distillation at known pressures; (2) collection or rejection of fractions as desired! (3‘) trans- ference of pure distillate to a storage container from which sanples could be removed as needed. me fractional distillatim apparatus consisted of a 26 in. electric- ally heated column 1; packed with glass helicies and fitted with a still head. A glass dripper :3 contained a small magnet and was fastened to the end of the graduated receiving cylinder 9 so that it could swing freely over the outlet tubes which led to the collection flask g, the waste flask g, and the storage flask Q. in" external magnet was used to position the dripper so that distillate collected in the graduated receiv- ing cylinder could be directed to the desired vessel. Prepurified nitrogen: could be introduced through the 'gas bubbler g or through the two-stay stopcock a. The second position of! stopcock 3 opened to a vacuum line A; 1 "-r — a- ' a'lletheeen, Coleman and Bell. purity 99.9 per cent. 29 through which the entire system could be evacuated. A one~meter length of capillary tube 3 was immersed just below the surface of a pool of mercury and, served as a safety valve and a rough manometer. The magnesium perchlorate drying tube referred to in the figure was kept closed from the system by a stopcock and only opened when it was desired to flush the apparatus with nitrOgen at a flow rate faster than the capillary would permit. The storage vessel g was wrapped in black paper and aluminum foil to protect the solvent from prolonged eXposure to light to avoid any photochemical decomposition. Before each distillation the entire system was evacuated and flushed with nitrogen several times. During the distillations a constant flow of nitrOgen through the system was maintained until the distillation temperature was reached. The distillate was collected in the graduated receiver 9. The most volatile fraction was discarded to the waste container 3!. The middle fractions were collected in the same way and transferred to the receiving flask :3, if there was to be a subsequent distillation. Flask l_3_ contained phosphorus pentoxide or anhydrous potassium carbonate depending on the stage of the purification procedure. After the last portion of the distillate had been discarded, the system was allowed to cool while nitrogen was admitted at a rapid rate through stapcock g. Flask _B_ was removed and rapidly stoppered. If further distillatiom were to be made, another flask containing the required drying agent was put in its place or else a cap was fitted to the taper. The waste flask H was emptied and returned to its position then the 3C distilling flask g nc removed ard the fracti-mnmg onlumn capped. The process of evacuation and flushing with nitrogen was repeated to remove any air which had been introduced. Flask g was then attached to the column under a stream of nitrogen and the distillation resumed. Tube 11: sewed as a waste vessel. to prevent any uniesired liquid from getting into the. tubing cozmected to the storage flask. Befcre the. purified acetonitrile was delivered to the storage flask, a few millic liters were allm-zed to flow into. tube 2 for rinsing purposes. The stop-c cooks above tube :1: “were closed, stopcocks E3 and 12 were Opened, and the acetonitrile was then delivered to the storage flask _Q. Stapcock i3 served as a pressure release valve while the storage flask was filled. A stepcock with a similar function was attached to tube 2. All stopcock-s adjacent to flask Q were Opened only under an inert, anhydrous atmosphere. Samples were collected from the storage container into a solvent flask constructed of a lO/hO female standard taper and a stopcock fastened to a 200 ml. Pyrex glass bulb. This solvent flask was attached to the storage vessel at the taper if, stopcock _B_; was then closed, stopcock Elli opened and the solvent allowed to flow into the flask. In most cases the vapor pressure of the acetonitrile in the storage flask was sufficient to deliver the desired amount of solvent. If more was needed it could be forced out by introducing nitrOgen through :1 and either §fl or g3. It should be mentioned here that acetonitrile is an excellent solvent for petroleum products and also attacks rubber. Dow Corning silicone highuvacuum grease and polyethylene tubing proved satisfactory and were 31 used when needed on all areas exposed to the solvent either as a vapor or in the liquid state. Prepara: ion c.f_S_ol.ut i on _s‘ Prior to making any solution. the solvent was freed of an}r dissolved gases that might introduce errors in the vapor pressure measurements. To do this the female adapter 13 was removed from the vacuum system (see Figure 2.) and the solvent flask attached. StOpcocks gig, g}; and 1523 were Opened and the system evacuated. After closing gig the stop» cock on the solvent flask was opened and the solvent distilled into ampule £1, which was immersed in an isOprOpanol~--Dry Ice bath. In order to accelerate the transfer, the tubing between the solvent flask and ampule was wound with nichrome wire and warmed electrically. A shielded light bulb was used to warm the solvent flask. After sufficient solvent was frozen in ampule fl, stopcock kg was closed and Egg Opened to allow the frozen solvent to be degassed while it was warmed slightly. Stopcock kg was then closed, the cold bath transferred to ampule A13 and the solvent distilled into the second ampule where it was again degassed as described above. This procedure was repeated several times to insure complete degassing. The solvent flask was removed and the female adapter _'1_‘;_2_ was replaced. 1 dried salt was transferred in a dry box to the vapor pressure flask, Figure 3, which contained a Teflon-covered magnetic stirrer bar. Care must be taken in this operation to insure a quantitative transfer due to the static electricity present in the dry atmosphere of the box. The amount of salt added was determined by weight difference. m3H wkquEOU N memuE 2 D m i I. -q '- 0am mzsa zoaaua «< .< r x . o m: oh . m m h F l «W _ mm mznm zsao<> omwwumv O» 0.4 9. 2m 2% . . c_ m 10-40 STANDARD MERCURY TA PER RESERVOIR \ VACUUM STOPCOCK (3mm BORE} 19-38 3 TA NDARD TA PER DRAWING AC TUA L SIZE T EF LON COVERED MA GNE TIC S TIRRE R BA R FIGURE 3 VAPOR PRESSURE F LASK 3h The vapor pressure flask was. then attached to the vacuum system at the femal adapter 24g, Figure 2. No lubricant was used because a vacuum“ tight fit was insured with a mercury seal. The flask was evacuated, then innuersed in a Dry IOG‘LdSOpI‘OpanOl bath and the desired amount of degassed acetonitrile distilled onto the solute. The stopcock on the vapor pressure flask was closed; the flask was removed from the vacuum system, washed with acetone and allowed to come to room temperature in a desiccator. The flask with solution was weighed and the amount of solvent determined by difference. Vapor Pressure. Apparatus A schematic diagram of the apparatus is shown in Figure 2. It con- sists essentially of three parts: (1) a bellows manometer and vapor pressure cell immersed in a constant temperature bath; (2) a mercury manometer; (3) the solvent degassing and collection section which was described previously. Parts (.1) and (2) are described in detail below. Bellows nanometerl and Constant Temperature Beth. A cross~sectional view of the bellows manometer is shown in Figure )4. The vapor to be measured was admitted to the inside of the bellows. Any pressure dif~ ference between the inside and outside of the bellows resulted in an upward or downward movement. This movement was transmitted by a needle resting on the top of the bellows to a small. face-«surfaced mirror arranged on needle supports such that it tilted with the motion of the 1The construction of the bellows manometer is described in the Appendix. [NEOPRENE ”O -”R1NG LENS NEOPRENE GASKET NEEDLES 4 NEOPRENE ”O—RING” MIRROR STAINLESS STEEL SINKS ‘ BRASS BELLOWS SUPPORT SHAFT SCALE: 1 in.=1in. I.'A'A‘. AAAAAAAAAAAAAAAAAAAAA " AAAAAAAAAAAAAAAAAAAAAAA CROSS-SECTIONAL VIEW OF THE BELLLOWS MANOMETER 36 bellows. The movement of the mirror, which is prOportional to the pressure difference, was magnified by the Optical lever principle. The image of a hairline from a galvanometer light source was reflected downward onto the surface of the small mirror from a large adjustable mirror outside the instrument (see Figure 5). The light beam was re- flected from the small mirror out through a lens and focused on a verti~ cal scale four meters from the instrument. A very small displacement of the bellows resulted in a pronounced movement of the hairline image on the vertical scale. The bellows manometer was used as a mill—reading instrument. The inside of the bellows was connected to the vacuum system and to taper 1;]; through stopcock §_-_2_ (Figure 2). The vapor pressure flask was connected to this 10/140 taper end sealed by surrounding the joint with mercury. The outer side of the bellows was connected to the vacuum system and a mercury manometer through stopcock _S_-_§. The bellows manometer and the vapor pressure flask were immersed in a constant temperamre bath and the entire assenbly connected to the vacuum system by means of a flexible glass spiral. It was found that the system reached equilibrium more rapidly when the solutions were stirred so provisions were made to stir the samples in the vapor pressure flask while in the constant temperature bath. This was achieved by sealing a magnetic stirrer in a watertight copper can. A little anhydrous calcium sulfate was added to remove any condensible water vapor. The cord from the motor was encased in Tygon tubing and led to a small rheostat outside the bath for control of the speed of stirring . P LIGHT BEAM \@ ‘ M I R R OR + AyUSTMENT L J, ' l * ,, . l . . RUBBER SPR’NGS <§ i i /BUSHING . IE . 1234 l P l J SCALE ’INCHES FIGURE 5 BELLOWS MANOMETER ASSEMBLY EXTERNAL VIEW 38 A copper tank 2 immersed to about one inch below the surface of the bath surrounded the bellows manometer and the vapor pressure flask. This modification was necessary to minimize rapid temperature fluctuations. The water in the thermostat was circulated around the copper tank by a pump. A micro-set mercury thermoregulator placed in the bath controlled two infrared heating lamps connected through an electronic relay with a mercury switch. Tap water circulating through a copper coil was used for cooling. This arrangement enabled a temperature control of i- 0.00300. at room temperature in the vicinity of the vapor pressure flask. The temperature was observed with a Beckmann thermometer which had been calibrated in the bath with a standardized platinum resistance thermometer. A Hoke needle valve 11 connected the manometer to a second vacuum pump and a supply of nitrogen. This allowed the system to be brought to atmospheric pressure before removing the vapor pressure flask and also made it possible to equalize the pressure on the outside of the bellows with the vapor being measured on the inside. In order to minimize vibrations the bellows manometer was mounted on a spring-suspended base that was supported in the bath by aluminum rods mounted in rubber bushings . Herc-um Manometer. The mercury manometer served to measure the pressure of the nitrogen necessary to restore the bellows to the null- point position. This pressure is equal to the vapor pressure of the solution unier investigation. The mercury column was measured with a cathetometer graduated to 0.01 mm. In order to be able to make all measurements on a rising miniscus the manometer was modified as suggested 39 by Allen, gt _al. (39). A mercury reservoir 3 was attached to the manometer and a small glass pointer built into the leftwhand column. The measurements were made by introducing mercury from the reservoir until the meniscus in the left-hand coluxnn was just touching the pointer. Once this position was established all further measurements could be made by reading the height of mercury in the rightehand column only. The left-hand mercury column was positioned by using a magnifying lens to observe when the tip of the pointer and its image were coincident on the bright mercury surface. This position could be established within the precision of the cathetometer. Because the manometer was constructed of 10 mm. I.D. Pyrex tubing which was not uniform a meniscus correction was applied. A thermometer immersed in a tube of mercury suspended between the arms of the manometer was used to determine the temperature. A microscope substage lamp which gives a soft but bright light was used to illuminate the menisci. General Information. The vacuum line passed through a mercury dif~ fusion pump backed by a Cenco~Hyvac mechanical pump. A liquid air trap preceded the diffusion pump and a conventional trap containing Dry Ice . and isopropanol protected the mechanical pump. The pressure in the system was measured with a calibrated McLeod gauge. A pressure in the order of l x 10-15mm. of mercury could be obtained throughout the apparatus. Apiezon-N vacuum grease and Apiezon~H vacuum wax were used on stop- cocks ani tapers not in contact with acetonitrile vapors. Those stopcocks exposed to acetonitrile, Sml, Sw2, SmlO, _S__-_2_L_‘_L_ and §__1_g_ (Figure 2) and the taper on the vapor pressure flask were greased with Dow Corning silicone I40 high-vacuum grease. The metal to glass taper leading from the outside of the bellows was sealed with Apiezonww. I The metal to glass taper connecting the inside of the bellows to the system required special treatment. The best solution found was to coat the forward end of the taper with silicone grease and the rear three—quarter portion with Apiezonww. The entire external surface of the Joint was then painted with Glyptal enamel.1 This gave a rigid, vacuum- tight seal that showed no attack by the solvent after months of use. Operation of the Vapor Pressure Apparatus The vapor pressure flask containing the solution to be studied was connected to 1.3;; (Figure 2) with the mercury seal. The bellows manometer was covered with an electrical heating tape and warmed to 50 or 60°C. with all stopcocks open except §_«_~_l., §;§ and §:_§ while the system was evacuated. The tenperature was maintained with the aid of a bimetallic thermoregulator between the tape and the manometer wall. The system was pumped down to about 0.1 micron then the tape removed. With further pumping pressures of the order of 0.01 micron or less were attained. It was found that the heating procedure was necessary to allow the evacuation of the manometer to such low pressures probably because of adsorption of vapors in the pores of the brass bellows. The mill point of the bellows was then set by focusing the hairline on the chart using the external mirror. ”Trade name for an enamel with low vapor pressure manufactured by General Electric Conpany. 141 Stopccocks if. and §;2 were closed and the solvent vapors introduced to” the inside of the bellows by slowly Opening stopcock §;_l. Nitrogen was then introduced through stopcock §-_-_§ and control valve 3 until the hairline image was nearly to the zero mark on the scale. Stopcocks fl and _S_-_-_1 were then closed and _S_-_-_§_ opened. More nitrogen was added to the mercury reservoir until the left-hand mercury column made contact with the pointer. These last two steps were repeated until the left-hand mercury column and the pointer were just in contact and the bellows was at the null-point setting. The system was allowed to stand until equilibrium was attained. This was determined when no further movement of the hairline was observed and when measurements of the height of the mercury column taken at one- hour intervals remained constant. This value was taken as the uncorrected vapor pressure of the solution. In some measurements the left-hand mercury column was not brougxt to the pointer, but rather stopcock E21 was left open and measurements of both the left and right columns were made. The precision was the same with either method . At the end of the run stopcock §;l_. was closed, the pressure in the system was brought to that of the atmosphere by introducing nitrogen, the water was pumped from the bath, and the next sample fitted to 2:15. h2 DISCUSSION In order to evaluate better the results of the vapor pressure measurements the limits of accuracy of the various parts of the apparatus is presented. Temperature was a variable directly affecting the measurements at three positions on the apparatus; the thermostat which maintained the sample and bellows manometer at constant temperature, the liquid manometer from which values of pressure were measured, and the glass tubing connect- ing the liquid manometer to the outside portion of the bellows. It was observed that the movement of the bellows was extremely sensitive to temperature changes. During early attempts at making measure- ments, considerable difficulty was experienced in maintaining a constant setting of the hairline on the scale. This was believed to be due to sudden temperature fluctuations at various parts of the apparatus in the thermostat when suddenly contacted by portions of the rapidly circulating water not at the equilibrium bath temperature. To remedy this effect a copper tank was placed in the thermostat surrounding the bellows manometer and vapor pressure flask on four sides and the bottom. The tank was sub- merged about one inch below the surface of the bath. The Beckmann thermometer was immerseddn the tank next to the vapor pressure cell. The water outside the tank was circulated by the pump and its temperature controlled as described previously (see page 38). The net effect was to have the bellows and flask assenbly in a relatively quiet bath of water 143 surrounded by a water atmosphere at 25.00 i 0.01%. The temperature in the vicinity of the cell controlled well to 25.000 i 0.003%. at which all measurements were made. This seemed to limit any extreme and sudden changes in temperature ani the hairline position was stabilized con- siderably. Mercury was used for the experiments designed to test the performance of the apparatus. The temperature as measured by a thermometer immersed in a tube of mercury beside the manometer did not alter more than one degree during any measurement. The factor for converting the manometer readings to zero degrees centigrade was not altered significantly by these fluctuations so one can assume a negligible error introduced here. There was a relatively large volume in the portion of the apparatus between the liquid manometer and the bellows due to the spiral-tube connection. Sudden changes in room temperature were observed to be sufficient to introduce troublesome changes in the position of the mercury column during a measurement. For this reason, the spiralwas encased in a Styrafoam boat and the attendant tubing covered with glass wool insulation. After this precaution, the mercury columns showed no noticable fluctuations once set for a measurement . It is believed, after the above modifications, that errors due to temperature changes either in the bath or in the room were negligible. The hairline position on the chart could be read to i: 0.3 m. A movement of 0.1; mm. on the chart corresponds to a pressure change of 0.01 mm. of mercury, around the null point. Since the mercury column could be read to but i 0.02 mm, the hairline position could be established M with a better accuracy than one could read the cathetometer. Any error due to reading the hairline position was, therefore, negligible in the over-all measurements . The cathetometer was equipped with a telescope focused at one meter from the mercury column. Readings could be «reproduced to but i- 0.02 m. V due chiefly to vibrations caused by machinery in the laboratory. The positioning of the pointer in the left-hand mercury column could be done to i 0.02 mm. also. This results in a maximum error of i- 0.014 mm. in the measurement of the height of mercury. ‘ The greatest source of error in concentration measurements was in the weighing. Assuming no loss of solute during the transfer, its weight could be determined to i 0.02 mg. The weight of the vapor pressure flask, magnetic stirrer bar and solvent sample averaged about 85 grams am could be weighed accurately to within 1.0 mg. Depending on the sample, the weight of the solvent could be determined to an accuracy of about 0.02%. On the basis of the above considerations it is believed the error of any one measurement should be no more than i 0.0).; mm. of mercury in pressure and the uncertainty in concentration should not be geater than i 0.0006 molal units. Tables II through XI show the values obtained for various measure- ments of acetonitrile vapors after being in contact with the inner bellows compartment for considerable lengths of time. Equilibrium should be attained when the pressure, after steadily increasing, reaches a constant value with time. As one can see from the experimental data, the pressures gradually increased over several hours and then began to fluctuatebetween 145 TABLE II VAPOR pressures OF ACETONITRIIE .AT 25°C. Equilibrium Observed Average Vapor Mean.Deviation Time‘ Pressure Pressure from Average (hours) (cm. of Hg) (cm. of Hg) ' (cm. of Hg) 73 .0 9.390 ' 7u.l 9.1.07 86.3 9.hh6* 89 .0 9 .hzaw 91.75 9 405* 9 “1‘31 °°°°5 95 .8 9 .h26* *These four values were used to compute the average vapor pressure. TABLE III VLPOR mssvms 0F 0.1m. MOLAJ. pm... 1er THIOCTANATE SOLUTION u 25°C. Equilibrium Observed Average Vapor Mean.Deviation Time" Pressure Pressure from.£verage (hours) (cm. of Hg) (cm. of Hg) (cm. of Hg) 25.h 9.168 28.1 9.918 59 oh 9 .169* 73.1 9.172% 9.170 0.002 82.0 9.173* fThese three values were used to compute the average vapor pressure. h6 TABLE IV VAPOR.PRESSURES OF 0.113 MOLAL POTASSIUM THIOCYANATE SOLUTION AT 2500. Equilibrium Observed Average Vapor Mean Deviation Time Pressure Pressure from Average (hours) (cm. of Hg) (cm. of Hg) (cm. of Hg) 11.5 8.883 20.0 8.955% 23.6 8.95h% 8.952 0.003 2h.5 8.9h6* * These three values were used to compute the average vapor pressure. TABLE V VAPOR.PRESSURES OF 0.0175 MOLA1.POTASSIUM THIOCXANATE SOLUTION AT 25°C. ~ Equilibrium Observed Average Vapor Mean Deviation Time Pressure Pressure from Average (hours) (cm. of Hg) (cm. of Hg) (cm. of Hg) 17.0 8.871 18.25 8.899% 20.5 8.906% 22.5 8.897% 8.900 0.002 27.0 8.897* *These four.va1ues were used to compute the average vapor pressure. h? TABLE VI VAPOR PRESSURE 0F 0.0239 PDLAL POTASSIUM THIOCIANATE SOLUTION AT 25°C. fi... v' — Equilibrium Observed Average Vapor Mean Deviation Time Pressure Pressure from Average (hours) (cm. of Hg) (cm. of Hg) (cm. of Hg) 25 '5 8 e90h* 27.5 8.906% 28 075 809% 31 .5 8 .870 ' 8 .903 O .002 32.75 8.903* 33 .0 g 8 .901... f *These five values were used to compute the average vapor'pressure. TABLE VII VAPOR mum 0F 0.11. now. POTASSIUM THIOCYANATE SOLUTION AT 25°C. A _lL Equilibrium ’ Observed Average of Mean Deviation Time" Pressure Pressure from Average (hours) (cm. of Hg) (cm. of Hg) (cm. of Hg) 13 05 90176 18.0 9.188%- 21 .9 g 9 .185-:- 9 .185 O .002 23 .5 9 .183-16 “These three values were used ,to Compute the average vapor pressure. 148 TABLE VIII VAPOR PRESSURE OF 0.1014 MOLAL POTASSIUM THIOCYANATE SOLUTION AT 25°C. Equilibrium Observed Average Of Mean Deviation Time Pressure Pressure from Average (hours) (cm. of Hg) (cm. of Hg) (cm. Of Hg) 20 .5 9 .365);- 9 .365 O .002 22 .5 9.362;. *These three values were used to compute the average vapor pressure. TABLE II VAPOR PRBSURES OF 0.0570 Mom. POTASSIUM THIOCIANATE SOIUTION AT 25°C. Equilibrium Observed Average of Mean Deviation Time“ Pressure Pressure from Average (hours) (cm. of Hg) (cm. of Hg) 7. (cm. of Hg) 12.1 9.231% 12 .9 9 .2364 9 .235 O .002 1.1401 9 ‘239* *These four values were used to compute the average vapor pressure. h9 TABLEJX VAPOR PRESSURE 0F 0.01m mm. POTASSIUM IODIDE SOLUTION AT 25°C. Equilibrium Observed Average Of Mean Deviation Time“ Pressure Pressure from Average (hours) (cm. of Hg) (cm. of Hg) (cm. Of Hg) 114.9 10.1I21 15.9 10.132 19.65 10.1;147 20.85 10.158 21.7 10.150 22.95 10.1463 33.6 10.h90* 35.0 10.14912- 3? .1 10. 163* 10 .h91 0 .ml 38 .6 10 .h89e {These four values were used to compute the average vapor pressure. TABLE II VAPOR PRESSURE OF 0.0937 TDLLL POTASSIUM IODDE SOLUTION AT 25°C. Equilibrium Observed Average of Mean Deviation Time" Pressure Pressure from Average (hours) (cm. Of Hg) (cm. Of Hg) (cm. Of Hg) 116.5 8.85011- 11.9014 8.879% 8.859 0.01 226 .6 8 .852A *These three values were used to compute the average vapor pressure. 50 high and low values. Equilibrium was assumed to have been reached when these fluctuations became evident and the equilibrium vapor pressure was recorded as the average of the pressure measurements over this period. All pressures in the Tables have been reduced to 0°C. and standard gravity and have been corrected for mercury capillary depression (55). Table II shows the vapor pressure data when acetonitrile was the only substance present in the vapor pressure flask. NO constant value was attained after 96 hours; however, the pressure readings began to fluctuate after about eighty-eight hours of equilibration time. Assuming equilibrium was reached by this time then by averaging subsequent pressure values, a value of 9.191 cm. of mercury was obtained for the vapor pressure of acetonitrile at 25.0000 i 0.0030 C. This value is considerably higher than anygof those reported in the literature (see Table I, page 10). The fact that fluctuations of such large magnitude resulted at the end of the measurements when a constant value was expected seems to indicate erratic behavior in the instrument. Further measurements have confirmed this. The values for the vapor pressures of potassium thiocyanate solu- tions, given in Tables III through II, were taken by averaging the last of a series of measurements after equilibrium was believed to have been attained. In no case was a constant vapor pressure reached, but as was the case with the pure solvent, the pressure increased steadily with time for several hours and then began to fluctuate between high and low values. The measurements used in determining the equilibrium vapor pressures are 51 listed in these tables together with the average deviation of these values from the average for any one solution. One Observes a fair degree of precision in most of the measurements. The vapor pressures Of the potassium thiocyanate solutions increase with increasing concentration. One cannot account for such behavior on . the basis Of present theories. Plots of vapor pressure against concen- tration and vapor pressure lowering against concentration do not give any indication Of a smooth curve and the determination of the activity Of the solvent is not meaningful. Several measurements (see Tables IV, VII and VIII) were made with solutions of about the same concentration. The extreme differences in the values Obtained indicate the measurements are irreproducible and erratic. On this basis any-explanation or attempt to use the data in evaluating activities or interpreting behavior Of the solutions is not Justified. Confirmation Of the lack Of reproducibility Of theapparatus was further confirmed by a measurement on a potassium iodide solution. As seen in Table I, a 0.01377 93 solution gave a vapor pressure value Of 10.191 cm. of mercury. mviously no validity can be put in this nuber if one is to recognize the 9.1431 cm. value obtainedfor the solvent alone or in any one Of the literature values which are still lower in magnitude.- ‘From the results Of these measurements it is obvious n0 conclusions can be drawn other than that the present apparatus fails to make true reproducible measurements . 52 Since the limitations of all other parts Of the apparatus have been fairly well defined and cannot be designated as the source of errors of such magnitude, it must be the bellows manometer itself which was at fault. The fact that there was good precision in the measurements made at any one null-point setting while the mercury manometer and bellows were in communication indicates there is no leakage between the two bellows compartments. A temperature differential between.both sides Of the bellows cannot account for the trouble since ample time was allowed for the entire apparatus to come to temperature equilibrium. It appears to the author that the troubles Observed can be due, in the main, to faults in the construction of but two parts Of the instru- ment. The bellows itself may not be returning to its correct position when the hairline indicates the null point. This WOUld be the case if the bellows was displaced in random lateral directions during expansion or if there was an extremely high degree Of hysteresis in its movement. Of these two possibilities, the former seems to this author to be the most likely‘because at the low pressures involved the forces on the bellows should most certainly lie well within the region of low hysteresis. NO lateral displacement on expansion was Observed prior to assembly of the instrument. However, as only crude methods Of detecting this property ‘were used, it is possible that lateral movement did occur sufficient to account for the observed discrepancies but small enough to escape detection. The other likely source of the difficulty is in the mounting Of the mirror. This seemed to be a most likely source of error and every S3 effort was taken to minimize this in constructing the instrument. Due to the very small lever arm on the mirror and the high magnification afforded by the optical system, any slight change in location of the mirror on its pivot points could result in quite large movements of the hairline on the chart. To prevent this the needle points must be rounded on the tips so as to permit rocking and lessen their penetration into the surface of the metal comprising the sinks and groove. The sides of the sinks and groove must be very smooth and free of tool marks or ridges so the needles will not get caught on the sides but will rest on the bottom at all times. Obtaining these conditions with the equipment of the average shop was found to be extremely difficult and.it may not be possible to do a sufficiently precise job without special tools. There is, of course, the possibility that the inconsistencies observed were due to alterations in the solvent during the degassing procedure and while equilibrating in the bath. Care was exercised to keep the solvent and the solutions shielded from direct exposure to strong light to reduce the possibilities of any photochemical decomposition. The samples were at no time subjected to excessive heating; the ampules were merely warmed with a light bulb during distillations. In spite of these precautions such a possibility does exist and should not be ignored; however, if any alteration in the acetonitrile does occur due to chemical reaction, it is felt the results should not'be so great as to account for all the inconsistencies in the measurements. The possibility of leakage of water from the thermostat into the samples was investigated and no such evidence could be found. The fact Sh that the vapor pressures do level off somewhat after a prolonged time would indicate that water is not continually leaking into the system. It is concluded that the bellows manometer as constructed does not give the reproducibility in the measurement of acetonitrile vapor pressures necessary to permit any degree of confidence to be placed in the values obtained, to calculate activity coefficients, or to reach any conclusions as to the behavior and constitution of acetonitrile-salt solutions. SS SUGGESTIONS FOR FURTHER WORK The problem as outlined in the introduction to this thesis seems sufficiently sound in principle to merit further investigations along this line. Effort should be directed toward correcting the difficulties connected with construction of the bellows manometer so its potentialities can be realized. Suggestions are given below as to what might be done in future work toward this end as well as to Obtain additional information pertinent to a study of this kind. The method of sealing the instrument as described in the Appendix is not too satisfactory. It is suggested that a larger bellows housing be constructed so that a single 0~ring resting directly on the manometer housing would suffice for the top seal instead of the rather inefficient small O-ring and flat gasket combination used in the described apparatus. Care should be given to obtain a bellows which is free of any random lateral movement upon expansion. The improvement of the construction of the sinks and groove presents a more formidable problem. It is suggested that they be constructed from a stainless steel alloy which can be hardened. Unless a watchmaker's lathe can be Obtained, this author has no suggestion for a more efficient method for polishing the sinks than the one used. In lieu of this and the probable significance of this portion of the apparatus, it may be advisable to resort to some other method of measuring the bellows expansion than the one attempted here. Connecting the bellows by a lever to an 56 electromechanical pressure-sensing transducer or to a mirror mounted to pivot in bearings of low friction, such as are found in watches, might still maintain the manometer at a sufficiently high sensitivity to be satisfactory. In the event of the construction of a successful instrument, other lines of investigation should be explored to extend the significance Of the proposed problem. The value for go used in the modified Berthelot equation (see page 15) for calculating the fugacity of the acetonitrile vapors is an extra- polated value. Lambert _e_t ;a__l.= (142) obtained ,the'data at‘temperaturesr above 120°C. The excellent linearity of the plot somewhat Justifies the long extrapolation to 25°C. but this should certainly be confirmed by experimental measurement of pressure-volume-temperature data for aceto- nitrile vapors at lower temperatures. . Since the major difficulty with fluorocarbons as manometer fluids lay in their densities and high viscosities, it may be possible to com- pound a mixture of various such liquids to yield a fluid possessing the desired properties. If such be the .case, the more simplified techniques of Gibson and Adams (29) could be applied to the vapor pressure measure- ments. . The advantages of this method would encourage investigations along this line . S7 BIBLIOGRAPHY (1) R. A. Robinson and R. H. Stokes, "Electrolyte Solutions,“ Butterworth Scientific Publications, London, England, 1955. (2) H. S. Harned and B. B. Owen, "The Physical Chemistry of Electrolytic Solutions,“ 2d. EdL, Reinhold.Publishing Corp., New York, New York, 1950. (3) H. Falkenhagen, "Electrolytes," Translated by R. P. Bell, The Clarendon Press,.0xford, England, 1931;. (14) M. F. Bechtold and R. F. Newton, J. Am. Chem. Soc., fig, 1390 (19140). (5) E. W. Washburn and E. 0. Heuse, _i_._b;i_q., 21, 309 (1915). (6) R. A. Robinson and H. s. Harned, Chem. Rev., _2_§, 1.19 (191.1). (7) R. A. Wynveen, Ph. D. Thesis, Michigan State University, 1959. (8) E. J. Baldes and A. F. Johnson, Biodynamica, No. Q], (1939); cf., E. J. Baldes, ibid., No. to, (1939). (9) A. 1. Brady, H. Huff and J. w. McBain, J. Phys. Chem., 5, 301; (1951). (10) w. I. Higuchi, H. A. Schwartz, E. G. Rippie and T. Higuchi, 3.1.33.9.” .62, 996 (1959)- (11) o. F. Tower and A. F. o. Germann, J. Am. Chem. Soc., 1Q: 2M9 (191k). (12) H. c. Jones, .2. physik. Chem., 2;, 111. (1899). ' (13) H. Hunt and w. E. Larsen, J. Phys. Chem., L8, 801 (19310. (11.) H. w. Ritchey and H. Hunt, 2232-: y}, be? (1939). (15) M. Randall and A. H. white, J. Am. Chem. Soc., kg, 2511. (1926). (16) S. Abe, K. Watanabe and R. Here, J. Soc. Chem. Ind., Japan, 3.3.. Suppl. binding 61.2 (1935); C- A» 2.9.. 5332 (1935). (17) A. I. Shatenshtein and L. S. Uskova, Acta Physicochim. U.R.S.S., 2.. 337 (1935) (in German); 0. 1., 22. 61.89 (1935). (18) M. Linhard, Z. physik. Chem., A115, 168 (1936). 58 (19) A. F. O. Germann and G. H. McIntyre, J. Phys. Chem., 22, 102 (1925). (20) H. Oosaka, Bull. Inst Phys. -Chem. Research (Tokyo), 19,1.66 (Abstracts 118 149) (in English) published with Sci. CPapers Inst. Phys. -Chem. Research (Tokyo), _5, No. 301. (l931)3C . A.25, 5610 (1931) (21) K. Ekelin and L. o. Sille’n, Acta Chem. Scand., 1, 987 (1953). (22) E. He'gfelt, Aota Chem. Scand., 5, 11.00 (1951). (23) V. G. Eversole and A. L. Hanson, J. Phys. Chem., 1:11: 1 (191.3). (21.) M. Randall and B. Longtin, ibid., kg, 306 (191.0). (25) W. G. Eversole and T. F. Hart, ibid., 26, 555 (191.2). (26) J. C. W. Frazer and B. F. Lovelace, J. Am. Chem. Soc., 56, 21.39 (1911;); cf., B. F. Lovelace, J.C ..w Frazer and B. Miller, ibid., 8. 515 (1916) (27) 21.1210)“. Frazer, B. F. Lovelace and '1‘. H. Rogers, ibid., L, 1793 19 (28) Lord Rayleigh, z. physik. Chem., 31, 713 (1901). (29) R. E. Gibson and L. H. Adams, J. Am. Chem. Soc., 52, 2679 (1933). (30) s. Shankman and A. R. Gordon,.gggg.,.§l, 2370 (1939). (31) 0. Heim, Bull. Soc. Chim. Belg.,.y2, h67 (1933). (32) R. R. Dreisbach and R. A. martin, Ind. Eng. Chem.,‘gg, 2875 (19h9). (33) R. R. Dreisbach, “Physical Properties of Chemical Substances, Dow Chemical Company, Midland, Michigan, 1953. (314) I. Brown and F. Smith, Aust. J. Chem., 1, 269 (1951.). (35) R. E. Dodd and P. L. Robinson, “Experimental Inorganic Chemistry,“ Eleevier Publishing 00., New York, New York, 1951., p. 129. (36) O. Amsel and G. Uittwer, Z. angew. Phys., 5, 20 (1956). (37) P. A. Gigulre and I. D. Liu, Can. J. Technology, 111.: 1.73 (1957). (38) H. G. East and H. Kuhn, J. Sci. Instru., 22, 185 (191.6). 59 (39) P. H. Allen, D. H. Ehrerett and M. F. Penny, Proc. Roy. Soc. (London), 2232, 11.9 (1952); cf.,D. H. Eyerett and M. F. Penney, ibid., A212, 161. (1952). '“"" ""‘"— (1.0) M. F. P. Wagley, _I_._o’_$_e_ M. F. Penney, Private communication. (1.1) G. N. Lewis and M. Randall, l‘Cl‘hermodynamics and the Free Energy of Chemical Substances ," lst Ed. , McGraw-Hill Book Co., Inc. , New York, New York, 1923. (1.2) J. D. Lambert, G. A. H. Roberts, J. S. Rowlinson and V. J. Wilkinson, Proc. Roy. Soc. (London), A126, 113 (191.9). (1.3) K. 'Schafer and o. R. Foz Gazulla, z. physik. Chem., 1352, 299 (191.2). (1.1.) P. Debye and H. Stack, "Tables Anuelles Internationales de Constantes et Données Numériques," Vol. XI, McGraw~Hill Book Co., Inc., New York, blew York, 1937, ppo 22-270 (h5) N. Bjerrum, “Selected Papers," Einer Munksgaard, Copenhagen, Denmark, l9h9, p. 108. . (1.6) R. M. Fuoss and C. A. Kraus, J. Am. Chem. Soc., 55, 2387 (1933). (1.7) R. M. Fuoss and c. A. Kraus, ibid., 51, l (1935). (1.8) C. A. Kraus, J. Phys. Chem., Q, 129 (1956). (1.9) P. L. Mercier and C. A. Kraus, Proc. Natl. Acad. 'Sci., 2.2, 1.87 (1956). (50) I. M. Kolthoff and J. J. Lingane, J. Am. Chem. Soc., 51," 2126 (1935); cf., 1. M. Kolthoff and V. A. Stenger, "Volumetric Analysis," Vol. II, Interscience Publishers, New York, New York, 191.7. - (51) W. Ward, Ph. D. Thesis, State University of Iowa, 1958. (52) P. Walden and. E. J. Birr, Z. physik. Chem., 1.1.1., 269 (1929). (53) J. W. Smith and L. B. Witten, Trans. Faraday Soc., 31, 1301. (1951). (51.) E. R. Weaver, J. Am. Chem. Soc., 56, 21.62 (1911.); cf., K. G. Stone, |'Determination of Organic Compounds ," McGraw-Hill Book Company, Inc. , New York, New York, 1956, p. 1.. . (55) A. Weissburger, ”Physical Methods of Organic Chemistry," Part I, Interscience Publishers, Inc. , New York, New York, 191.9, p. 11.6. APPENDIX APPENDIX Construction of the Bellows Manometer The essential features of the manometer are shown in Figure h. A description of the Operational principles of the instrument precedes the details of its construction in this Appendix. 2 The movement of the bellows caused a small platform to tilt. A plane, face-surfaced mirror fixed to a platform transmits the tilt to an optical lever similar to that used in galvanometers. Another face-surfaced mirror located outside the instrument reflects the beam in a horizontal direction (see also Figure 5). A beam of light focused a hairline through a tele- scope onto the tilting mirror. This image was reflected through the lens onto the large mirror and projected onto a screen four meters distant. The manometer was made airtight with neoprene O-rings and gaskets. Two tubes led from the manometer, each terminating in a lO/hO male standard taper which was machined from Monel metal. One tube was connected to the inside of the bellows, the other to the outside. The instrument could thus be used as a differential manometer over a wide range of pressures. The platform rested on three needles: two were mounted in a rigid base, one behind the other; a third, which had double points, rested on top of the bellows and supported the platform. The double-pointed needle and one of the others rested in stainless steel polished sinks; the third needle rested in a stainless steel groove bolted to the platform. This groove could be moved back and forth and locked in place to position 61 all three needles. This arrangement had the advantage that small lateral displacements, which most bellows show on extension, do not affect the length of the lever which was determined by the distance between the points of contact of the needles on the platform. This distance was small-«3 mm. As a result, very small displacements of the bellows were possible and hysteresis and deviations from Hooke's Law were minimized. Arw sudden bursts of pressure or rough handling could cause the mirror and needles to be disarranged. In order to minimize this danger, hooks were fashioned to prevent the platform from being jerked off the needles and a guide around the double-pointed needle served to keep it from falling into the bellows compartment. The Bellows. Several standard brass bellows of varying sizes were ~ obtained from the stock of a bellows manufacturing firm.1 Standard bellows come in two basic types-~regular and extrafluiblr—differingin the depth of the convolutions relative to the outside diameter. Sanples of both typeswere obtained. Selection was made firstly on the basis of the size limitations dictated by available materials and secmdly on the basis of the greatest flexibility (i.e., linear expansion with pressure) with the least lateral displacement during expansion. These effects were measured qualitatively. An extraflexible bellows with an inside diameter of 1 15/61. in. and an outside diameter of 2 in. was chosen. Calculations based on the manu- facturer's data for this bellows showed a sensitivity of 0.1 m. of linear expansion per centimeter 0f mercury pressure. As the pressures to be measured were not to exceed 10, cm., the total expansion of the bellows lclifford Manufacturing Company, Haltham 51., Massachusetts. should not have exceeded 1.0 mm. With this small maximum expansion during a measurement, no deformation of the bellows or hysteresis should be observed. The Manomete‘r Housing. The dimensions for the housing are given in Figure 6. The material used was brass. The housing was divided into two chambers: the bellows chamber, 2 1/8 in. in diameter and 3 1/1. in. high; an upper chamber, 1 5/8 in. in diameter and 1 3/1. in. high. The center of the upper chamber was offset 3 mm. from the center of the bellows chamber to allow for the mirror lever arm. The two chambers were connected by a threaded 3/1. in. hole for the needle guide. "me upper chandler was threaded for a retaining screw. The outside of the housing was turned down one inch to a diameter of 2 5/16 in. around the center of the upper chamber to form a shoulder for the external mirror support ring, Figure 8. The machining and threading was done on a toolroom lathe with a four-jawed independent chuck. The Bellows Housi_ng Base _R_.:1_ng. This simple brass ring, Figure 9, was bored to fit around the base of the bellows housing. It was silver soldered flush with the base to form a flange for sealing the housing to the base of the manometer. After soldering, it was placed in the lathe and the under surface smoothed and polished to insure airtight contact with the O-ring seal. The Manometer Base. A diagram of the brass circular manometer base is shown in Figure 11. A groove 3/8 in. wide and having an outside diameter of 2 7/8 in. was bored to a depth of 5/16 in. As this groove 63 was to accommodate the O-ring gasket, care was exercised to make the floor as smooth as possible. A hole 1 3/8 in. in diameter was milled in the center to a depth of l/h in. to accommodate the brass support, Figure 13. This support served to decrease the free volume inside the bellows and to support the bellows in case it was suddenly compressed due to a break in the vacuum system or similar accident. The support was silver soldered in the hole taking care to make the seal airtight. T510 3/16 in. holes were bored from opposite sides of the base to be intersected by similar holes at right angles. One of these latter holes opened between the inside of the groove and the bellows support, the other extended through the center of the bellows support and out the tOp. Five-sixteenths 0.D. copper tubing 1 1/2 in. long was silver soldered into the holes on the sides of the base, and the opposite ends silver soldered to lO/hO male standard tapers machined from Monel metal. These provided entrances to the inside and outside of the bellows. Six holes around the base were threaded to accommodate 12-214 screws. After the bellows was in place (see page 69) a retaining ring, Figure 10, was machined to slip over the housing and rest on the lower flange made by the brass ring soldered to the housing base. One-fourth inch holes in this retaining ring were centered above the screw holes in the base am served to fasten the housing and base together and to compress the ' O-ring laying in the groove of the base thus making the necessary air- tight seal. The Sinks and Groove. One-eighth inch stainless steel rod was used for the three sinks. The rod was mounted in a lathe and drilled to a , DRILL -I 5’." T READ 1 perinch DRILL-1V4" I l I y ' “ ~ .— -'| M J‘ 24 pcrmc / . l V 4L / L / / / Z <{l 3'7”” fi / V‘ / / I / / / / | j C I / / é A SCALE: / l j 1in.=1in. / | / E u C ___JL L————“—_§£~'_—*———4 FIGURE 6 BELLOWS MANOMETER HOUSING / . V4.28 THREADS PER INCH. (see FIGURE 21 for details of adjust- ment knob.) 25’4' l ill T WI3+$ WI— Id/H . I I‘—'9Is—’I"1Ie SCALE: . 1 in.= 1in. FIGURE 7 I—— m——4 EXTERNAL MIRROR SUPPORT 372.. fi] ll TAP FOR 10-24 SCREW” ' k— L—fi —_—'—fir ~A ! i __ _ _,---.;,-_. . l ' I V; I _ m N _ ”.11- _ I TAP FOR I 1/4'-23$1.=r- - — _L‘m‘I‘M SCREW II. ' ' SCALE (in): ULIJ FIGURE 8 EXTERNAL MIRROR SUPPORT RING SCALE — 1;— 1in.=1in. . ———-—--‘—7 r — “E “E— "“‘-2 ' l_ : —__.____§IZ __Z -I m- ‘Q FIGURE 9 BELLOWS HOUSING BASE RING 6 HOLES. 7/32' D’AM- ([3 1 2 3 l l l SCALE Va": 1 inch FIGURE IO HOUSING RETAINING RING 2.79m ES SCALE: 1 Ida‘l in. FIGURE I I MANOMETER BASE it ——_fln: : x 0.: -o Og— O c-~ '— EL}; ‘eo‘n‘fiabq .- —— >h—gkmn —- ¢ w-vn-"_'q'g _ I fie U 12 -<-I-- _. 2 h: SCALE: 1in.=1in FIGURE I2 UPPER RETAINING SCREW FIGURE :3 SUPPORT BELLOWS 69 depth of 3/32 in. A60O countersink was used as an endmill to sharpen the bottom of the hole. The rod was then cut off 1/8 in. from the end. An attempt was made to bring the bottoms of the sinks to a fine pointed recess and to polish them free from ridges and burrs. This was done ‘with a piece of drillrod that had been turned and filed to a very sharp point and then hardened. The 1/8 in.sinks were clamped in.a.horizontal position on a vertical milling machine and polished with the hardened rod and emery powder. The groove (see Figure 15) was made from a piece of stainless steel milled to a block 3/8 x 2 x 3/32 in. The block was stood on edge and milled along the long edge at a hSo angle. The slots for the set screws were bored and filed, then the two 1/1; in. pieces cut from the block to make the groove. This procedure gave the two halves a common edge insuring a close fit. The edges were polished on wet emery paper (No. 320), so they matched exactly and no light could be seen between them when.butted tOgether. A large magnifying glass was used in the operation. The groove was mounted in the slot in the mirror platform as shown in Figure 15. Rbunting the Bellows. The bellows was clamped gently in the.head- stock of a lathe, the center of the top determined, and the area tinned with lead solder. The bottom of one of the steel sinks was also tinned and held in place in the center of the bellows using a 60° lathe center. The solder was then melted with a torch and the sink fused to the bellows. Care was taken so that the sink remained upright and at right angles to the top surface of the bellows . SCALE: 1 In. =Iin. _ 452%; FIGURE l4 ZI- LENS MOUNT ”I“? _ w my _ __ Sly/’55.!- grill rod, Iago/and!”— _ _ _ a 5m 0 5 am ' ee , I L231 TI polish an hfirelen. 37. P 22 _ ®__ _ T— _ _ ___ i ,\ Push m in sink holes. _I._ GROOVE: Stalnless steeI. I. . - L_. __. *l' #5.- - __ 4L... Eton aled holes to allow for 4 I '\ :1: J13“. posigoéving In snug-fitting Iraub ”t Mam, £11,, T7}; 1 slot. pIeces needed. ‘ I ‘ escape/I '- I"'.'-.:'“1 I use, 45" \ __ _ .L was... is???“ 54” h—u.’-§‘°"""*‘ *5 ' GROOVE *— y.~—.. IE. ’SCALE:1In.=1I‘n. FIGUREIS {figs-a MIRROR PLATFORM AND GROOVE .. _ l” T ' I ‘i’IO'P'I ' Holes (Or phonograph needles. Needles i] mgr—57.51" cut 5/s m. long. . " hi.-..+.._,J-Ij . MIBBQ_REL¥.O_T ELSE FIGURE l6 MIRROR PIVOT BASE AND NEEDLE GUIDE 37. mos. yum. SCALE: 1in.= 1 in. 71 The inside surface of the bottom of the bellows was coated with a smooth layer of solder prior to the installation of the sink. The base of the bellows support was also tinned and the bellows fitted over the support and the base warmed until the solder in the joint flowed freely. To prevent leakage a small bead of solder was laid around the outside of the bellows where it butted flush against the base. It should be empha- sized that it was of utmost importance to solder the bellows in a vertical position. To insure this the bellows was checked when it was in position over the support with its lower rim flush against the base. If it was not perpendicular, the bottom of the rim was filed until it was. The soldering operation was done with the bellows fit in the same position. The bellows sealed to the base assembly was submerged in water and compressed air was slog introduced with care not to expand it to excess. No evidence of leaks was noted. Similar results were obtained using Freon gas and a General Electric leak detector. Mror Platform. The diagram is shown in Figure 15. The plat- form was constructed of brass. A slot was milled along one short edge to accommodate the two halves of the stainless steel grove. Two screws were threaded into the center of the slot. These served to lock the stainless steel groove in place when it was set so the apex of the groove ' was directly across from its corresponding sink. Holes were drilled for the two sinks and the sinks pressed into place. Two pieces of 1/16 in. brass rod were soldered to the midpoint of the short edges of the plat- form. These were for the hooks which kept the platform from falling off 72 its needle supports. The mirror could be glued to the platform surface or clamped into place just before installation. It was necessary to know the depth of the sinks and the groove in order to determine the lengths of the supporting needles so that the platform would rest parallel to the base. This was determined in the following manner. The smooth top of the platform was covered with plastic insulating tape and placed face down on the table of a vertical milling machine. A sharp needle was inserted in the chuck of the machine and centered over the apex of the sink. A three volt light bulb was connected in series with the chuck, the platform, and a dry cell.battery so that the bulb would light when the needle touched the platform and completed the circuit. The depth was determined by measuring the dif- ference in the reading on the micrometer dial on the spindle of the milling machine (graduated to 0.001 in.) when the needle touched the platform surface and when it touched the bottom of the sink. A similar procedure gave the groove depth. The platform was then filed or solder was added until the best balance over its center was attained. It was the objective in the above Operations tosgain a platform with a low center of gravity and nearly frictionless movement on its supports. The Mirror Compartment. This part of the apparatus, Figure 17, con- sisted of a cylindrical brass chamber which was grooved at the tOp end to accommodate an 0-ring and machined flat at the bottom for a tight fit against a neoprene gasket. The lower half of the cylinder was threaded on the inside. The mirror pivot base upon which the mirror was mounted could be raised or lowered on these threads to adjust the distance between the mirror and the bellows. 32 THREADS A e—r-. - —— ——.—— —T—. O O-RING . 353 K J CHANNEL "—- “11:: j: ' . - "a. I I «IIIr——%‘——I'T I'——7%"—"I :_~ _ _ __ CLOSE-UP OF GROOVE TO " «1. TAKE tar/:0” 1y; 0. 0.. V; w 3.;— , _ 45— NEOPRENE O-RIIVG". 4.1; om SCALE: 1I'n.=1in. I'M I , FIGURE I7 MIRROR COMPARTMENT BRASS WASHER ’NEOPRENE "O -RING" HOOKS / Drawing not to scale.) 'NE OPRE NE -. ”FLAT” GASKET FIGURE l8 DETAILS OF MIRROR CHAMBER ASSEMBLY 7h The Mirror Pivot Base. The mirror pivot base drawn in Figure 16 was a small brass threaded disk 3/16 in. thick. It contained the two stationary needles for supporting the mirror and the hooks that prevented the mirror from being jarred off these supports. .1 3/16 in. hole was bored with its center 1.5 m. from' the center of the disk. has double- pointed needle extending through this hole was fitted into a sink on the mirror, and rested in the sink on the bellows. The holes for the two stationary needles were drilled 1.5 m. from the center of the disk directly back from the small hole and to a depth such that they would support the mirror parallel to the face of the disk. This depth was calculated knowing the depth of the groove and the sink (see preceding section). The holes were bored exactly the same size as the needles or 0.001 in. undersize and the needles sweat fit. rwo’hooke were soldered in holes drilled on a line through the center of the disk and parallel to a line joining the two needles. the hooks were fashioned to bend around the small rods extending from opposite sides of the mirror platform. They did not touch or interfer with the free movement of the platform in any way, but yet were close enough to the rods so that the platform could not be taken cospletely off the needles with the hooks in place. The Needles. The two stationary needles were nickel-plated phono— graph needles . rho double-pointed needle was fashioned from a nickel- plated dueling needle. he points of the needles were rounied in a lathe using a carborundum oil stone until they fit the curvature of a circle 0.01 in. in diameter viewed through a hand microscope. They were polished with wet 320-grade emery paper. 75 A needle guide, Figure 16, was screwed into the threaded hole connecting the mirror and bellows chambers. It consisted of a disk with a 3/16 in. hole drilled in its center. The double-pointed needle passed through this hole without touching the sides of the disk. The guide's function was to prevent the needle from falling into the lower compart- ment. Mm and Lens Mount. The lens used for focusing the projected hairline onto the graduated screen was a simple plano-convex lens ground from a spectacle lens blank. Its focal length was approximately four meters and the diameter of the lens, 1 3/8 in. It was mounted in a recessed aluminum washer (Figure 114) with an epolqr resin. The bottom of the washer was polished to serve as the upper surface of an O—ring seal. flipper Retaining Screw. An aluminum ring with an inside diameter 1 1/8 in. (see Figure 12) was threaded to match the top of the bellows housing. Four slots arranged at 90° intervals across the top served as lock points for a bar used to tighten the screw firmly in place. A brass washer 1 5/8 in. 0.D. and 1 1/8 in. I. D. was placed between the retaining nut and the lens mount when the manometer was assembled. The Bacternal Mirror Assembly. The details of the various parts of the mechanism are shown in Figures 7, 19, 20, and 21. The mirror used was a 3 x 2 3/14 in. face-surfaced plane mirror mounted on a brass plate bolted to a bracket so its angular position with respect to the bracket could be altered. The bracket pivoted in an adjustment mechanism that was in turn fastened to the aluminum ring which fit around the top of the >mm2mmm< mOmmS m