INVESTIGATEGNS 0F POLYEALOGEN COMPLHES Thesis for the Degree of P21. D‘ MICHIQAN STATE UNIVERSITY TERRY SURLES 1970 LIBRARY Michigan State University This is to certify that the thesis entitled Investigations of Polyhalogen Complexes presented by Terry Surles has been accepted towards fulfillment of the requirements for __Eh_.D_._ degree in .Chemisiny. Major professor» Date June 26, 1970 0-169 INVESTIGATIONS OF POLYHALOGEN COMPLEXES By Terry Suries AN ABSTRACT OF A THESIS Submitted to Michigan State University in partial fulfiliment of the requirements for the degree of DOCTOR OF PHILOSOPHY Department of Chemistry 1970 ABSTRACT INVESTIGATIONS OF POLYHALOGEN COMPLEXES By Terry Surles Spectroscopic techniques have been used in this investigation to study the bromine chloride complexes with heterocyclic amines as well as in the equilibria in bromine trifluoride solutions. Conductometric techniques were also used in the study of bromine trifluoride solutionsg Bromine chloride complexes with pyridine, 4-picoline, 2,4~, 2,5-, 2,6-, 3.4-, and 3,5-lutidines, and 2,3,6-collidine have been prepared. The complexes, except those of 3,4-lutidine, 4-picoline, and pyridine, are quite stable at room temperature under anhydrous conditionso Formation constants of these complexes were determined in carbon tetrachloride solutions, Comparison of these values with the formation constants of iodine(I) chloride and iodine bromide shows that the order of Lewis acid strength is 101 > IBr > BrClq Far infrared spectra of the complexes indicate that the fundamental Br-Cl vibration band is 1 shifted by complexation with pyridine from 430 to 280 cm0' and that the Br-N stretching vibration occurs at 192 cm '1, The addition of strong acids and bases to liquid bromine tri- fluoride has been studied by Raman and infrared spectrosc0pic techniqueso The individual bands in the Raman spectrum of liquid Terry Surles bromine trifluoride have been resolved and vibrational assignments have been made. The fundamental vibration frequencies for the bromine trifluoride monomer in liquid are 236(a1), 269(b2), 34l(b]), 531(a1), l 6l3(b1), and 672(a1) cm.‘ The fundamental frequencies for the tetra- fluorobromate(III) anion in liquid bromine trifluoride are 249(b19), l 302(a2u), 455(b29)’ 528(a19), and 570(bu) cm. The Raman spectrum of potassium tetrafluorobromate(III), KBrF4, had fundamental vibrational 1 These data give frequencies at 530(alg)’ 455(b2g)’ and 242(b19) cm," strong support for a square planar (point group 04h) configuration for the BrF4' anion in both liquid bromine trifluoride and solid KBrF4° Data for the difluorobromonium(III), BrF2+, cation are not as clear. The Raman spectrum of solid BrFZSbF6 has bands attributable to BrF2+ at 734, 704 and 308 cm.“ The infrared spectrum of BerAsFG has bands 1 attributable to 8er+ at 732 and 705 cm.‘ These data support a bent cationic structure (point group c2v)' The cation in liquid bromine l l trifluoride has a band at 625 cm." in the Raman spectrum at 635 cm“- in the infrared spectrum. Experimental problems are encountered which make it difficult to observe other bands, if any, which might be due to the cation. Raman, infrared, and electrical conductance studies were carried out on liquid mixtures of bromine trifluoride with hydrogen fluoridee The experimental data reveal the existence of a fluoride ion transfer equilibrium HF + BrF3e==HF2- + BrF2+. The equilibrium constant for the above reaction was calculated to be ~lO'3a A new band was observed at 662 cm;-1 in the Raman spectrum for dilute solutions of + bromine trifluoride in hydrogen fluoride which may be due to the BrF2 cation solvated by HF molecules” Conductance measurements show a Terry Surles decrease in molar conductivity from pure bromine trifluoride to ~70 mole percent BrF3° From 70 to 5 mole percent BrF3, the molar conductivity increases. From 5 to 2.5 mole percent BrF3, the molar conductivity decreases and then increases from 2.5 to 0 mole percent BrF3. The peculiar behavior of the molar conductance in dilute solu- tions bromine trifluoride may be due to the formation of a new, less mobile Species as evidenced by the appearance of a new Raman band at 662 cm."1 Raman, infrared, and electrical conductance studies were carried out on liquid mixtures of bromine trifluoride with chlorine trifluorideo The experimental data reveal the existence of a fluoride ion transfer equilibrium ClF3 + BrF3=F=fl=BrF4- + ClF2+. The equilibrium constant for the above reaction was calculated to be ~l0‘4. It is seen that chlorine trifluoride is a weak base (fluoride ion donor) toward bromine trifluoride and that bromine trifluoride is only a slightly stronger base toward hydrogen fluoride: Thus, the base strength decreases in the order ClF3 > BrF3 > HFO INVESTIGATIONS OF POLYHALOGEN COMPLEXES By Terry Surles A THESIS Submitted to Michigan State University in partial fulfillment of the requirements for the degree of DOCTOR OF PHILOSOPHY Department of Chemistry 1970 ACKNOWLEDGMENTS I first wish to gratefully acknowledge the major professor of this paper, Dr. Alexander I. Popov. Without his patient guidance, encour- agement, and understanding through numerous trials and tribulations, this work could not have been completed. I also wish to acknowledge my most able mentor at Argonne National Laboratory, Dr. Herbert H. Hyman. His understanding and knowledge of the problems involved with thesis research were indispensible in this work. Special appreciation is extended to Mr. Lloyd A. Quarterman who personally taught me a great deal about the manipulation of halogen fluorides and kept my spirits high when the going got rough. Thanks is extended to Dr. Thomas Pinnavaia who served as the second reader for this thesis and made many valuable suggestions for this work. Thanks is entended to just about everyone whom I came into contact with at Argonne National Laboratory. Their helpfulness and efficiency helped to reduce my time as a graduate student. These people include Ed Klocek, Norm Nelson, and Ted Bezak and others in the machine shop, Bob Nelson, Dave Lind, and Glenn Mack in the instrument shop, James C. Sullivan, Irv Sheft, Dr. Howard Claassen, and Dr. Henry Hoekstra for their advice and assistance, and Elaine Longhini, Dorothy Maes, and Joanne Harmon of the typing pool for their typing efficiency. The author also gratefully acknowledges the support of an Associated 11' Midwest Universities-Argonne National Laboratory Predoctoral Fellowship. I would like to thank my parents for their interest and encouragement in my education. Finally, I gratefully acknowledge my wife, Jan, for her steady and emotional support in helping me through this last year. TABLE OF CONTENTS ACKNOWLEDGMENTS ............................ ii LIST OF TABLES ........................... vi LIST OF FIGURES. . . . . . . . . . . . . . . . . . . . . . . . . . . ix GENERAL INTRODUCTION .......................... 1 PART I. BROMINE CHLORIDE COMPLEXES .................. 4 HISTORICAL INTRODUCTION ..................... 4 EXPERIMENTAL PART ........................ 7 Chemicals .......................... 7 Preparation and Stability of Solutions ............ 8 Preparation of Bromine Chloride Complexes ........... 9 Spectral Measurements ..................... ll Analyses ........................... l4 Calculations ......................... l4 RESULTS AND DISCUSSION . . . . . . . . . . . . . . . . ..... 26 PART II. SPECTROSCOPIC STUDIES IN LIQUID BROMINE TRIFLUORIDE. . . . . . . . . . . . . ...... 31 HISTORICAL INTRODUCTION ..................... 31 TECHNIQUES FOR MANIPULATING HALOGEN FLUORIDES .......... 38 Materials of Construction ................... 38 Joints . . . . . . . . . . . . . . . . . . . . ........ 44 Valves . . . . . . . . . . . . . . . . . . . . . . . . . . . .46 Reaction Tubes ........................ 47 Conductivity Equipment . . . . . . . . . . . ......... 5l Density Measurements ..................... 51 Pressure Measurements ..................... 56 APPLICATION OF SPECTROSCOPIC TECHNIQUES TO THE STUDY OF BROMINE TRIFLUORIDE SOLUTIONS .............. 57 Spectral Cells for Visible and Ultraviolet Spectra ...... 57 Raman Cells for Liquid Samples ................ 57 iv Raman Cells for Solid Samples ................. 59 Infrared Spectral Cells .................... 6l Infrared Instrumentation ................... 66 Visible and Ultraviolet Instrumentation ............ 66 Raman Instrumentation ..................... 67 EXPERIMENTAL PART ........................ 68 Chemicals. . . . . . . . . . . . . . ............. 68 Preparation of Solutions ................... 7O Transferring of Solutions in the Vacuum Line ......... 7l Preparation of Solid Compounds ................ 72 Calculations ......................... 73 RESULTS AND DISCUSSION ..................... 98 Electrical Conductivity of Pure Bromine Trifluoride ...... 98 Spectroscopic Studies of Acid-Base Equilibria in Bromine Trifluoride . . ................. lOl Bromine Trifluoride-Hydrogen Fluoride System . ....... 119 Bromine Trifluoride-Chlorine Trifluoride System. . . . . . . 128 LIST OF REFERENCES ......................... 141 APPENDIX. SUGGESTIONS FOR FUTURE NORK ............... l45 LIST OF TABLES Table Page 1. Interhalogen Compounds ........... . ......... 2 2. Bromine Chloride Complexes with Pyridine and Pyridine Derivatives. . . . .......... . . . . . . . . . . . . .lO 3. Formation Constants of Amine-Bromine Chloride Complexes in Carbon Tetrachloride at 25° ................. 12 4. Spectrophotometric Data for the Pyridine-Bromine Chloride Complex in Carbon Tetrachloride. . . . . . . . . . . . . . . . .l9 5. Spectrophotometric Data for the 4-Picoline-Bromine Chloride Complex in Carbon Tetrachloride ............ 20 6. Spectrophotometric Data for the 2,4-Lutidine-Bromine Chloride Complex in Carbon Tetrachloride . . . . . . ...... 21 7. Spectrophotometric Data for the 2,5-Lutidine-Bromine Chloride Complex in Carbon Tetrachloride . . . . . ..... . .22 8. Spectrophotometric Data for the 2,6-Lutidine-Bromine Chloride Complex in Carbon Tetrachloride ............ 23 9. Spectrophotometric Data for the 3,4-Lutidine-Bromine Chloride Complex in Carbon Tetrachloride ............ 24 10. Spectrophotometric Data for the 3,5-Lutidine-Bromine Chloride Complex in Carbon Tetrachloride ............ 25 11. Comparison of the Formation Constants of Pyridine Complexes with ICl, IBr, and BrCl in Carbon Tetrachloride at 25° . . . . .28 12. Infrared Absorption Bands of Bromine Chloride-Pyridine and Substituted Pyridine Complexes in the 667-130 cm' Region. .29 13. Some Physical Properties of the Halogen Fluorides ........ 32 14. Properties of Teflon ...................... 41 15. Properties of Kel-F ....................... 43 16. Raman Data for Liquid Bromine Trifluoride ............ 76 vi T7. 18. 19. 20. 21. 22. 23. 24. 25. 26. 27. 28. 29. 30. 3]. 32. 33. 34. Raman Data for Potassium Fluoride Solutions in Bromine Trifluoride I ...................... 77 Raman Data for Potassium Fluoride Solutions in Bromine Trifluoride II ..... . . . . . . . . ........ 78 Raman Data for Cesium Fluoride and Rubidium Fluoride Solutions in Bromine Trifluoride . . . ...... . ...... 79 Raman Data for Antimony Pentafluoride Solutions in Bromine Trifluoride I ...................... 8O Raman Data for Antimony Pentafluoride Solutions in Bromine Trifluoride II ..................... 81 Raman Data for Arsenic Pentafluoride Solutions in Bromine Trifluoride ....................... 82 Raman Data for Mixtures of Hydrogen Fluoride and Bromine Trifluoride I ............ . ......... 83 Raman Data for Mixtures of Hydrogen Fluoride and Bromine Trifluoride II ...... . .............. 84 Raman Data for Mixtures of Hydrogen Fluoride and Bromine Trifluoride III ..................... 85 Raman Data for Mixtures of Hydrogen Fluoride and Bromine Trifluoride IV ............. . ....... 86 Raman Data for Mixtures of Hydrogen Fluoride and Bromine Trifluoride V ...................... 87 Raman Data for Mixtures of Hydrogen Fluoride and Bromine Trifluoride VI ..................... 88 Conductivity Data for Mixtures of Hydrogen Fluoride and Bromine Trifluoride I .................... 89 Conductivity Data for Mixtures of Hydrogen Fluoride and Bromine Trifluoride II ..................... 90 Conductivity Data for Mixtures of Hydrogen Fluoride and Bromine Trifluoride III ................... 91 Raman Data for Mixtures of Bromine Trifluoride and Chlorine Trifluoride I . . .> ................ 92 Raman Data for Mixtures Of Bromine Trifluoride and Chlorine Trifluoride II. . .' ................ 93 Raman Data for Mixtures of Bromine Trifluoride and Chlorine Trifluoride III .................. 94 vii 35. 36. 37. 38. 39. 40. 41. 42. 43. 44. 45. 46. 47. 48. 49. 50. 51. 52. Raman Data for Mixtures of Bromine Trifluoride and Chlorine Trifluoride IV ................... 95 Raman Data for Mixtures of Bromine Trifluoride and Chlorine Trifluoride V ..... . ............. 96 Conductivity Data for Mixtures of Bromine Trifluoride and Chlorine Trifluoride . . . . ................ 97 Temperature Dependence of the Specific Conductivity of Liquid Bromine Trifluoride ......... . ........ 99 Conductivity of Dilute Solutions of Volatile Impurities in Bromine Trifluoride at 25° ................. lOO Raman Spectrum of Liquid Bromine Trifluoride . . . . ..... 102 Vibrational Spectrum of Liquid Bromine Trifluoride . . . . . . 106 Vibrational Spectrum of the Tetrafluorobromate(III) Anion. . . 107 Vibrational Spectrum of Liquid Bromine Trifluoride Containing Acid Solutes ........ . ......... . . 113 Raman Spectra of Difluorobromonium(III) Cation . . . . . . . . 114 Vibrational Spectrum of Difluorobromonium(III) Cation. . . . . 118 Concentration of Species Present in Mixtures of Hydrogen Fluoride and Bromine Trifluoride . . . . . .......... 120 Values of the Equilibrium Constant for the Bromine Trifluoride-Hydrogen Fluoride System . . . ..... . . . . . 123 Conductivity of Bromine Trifluoride in Hydrogen Fluoride . . . 125 Vibrational Spectrum of Liquid Chlorine Trifluoride ...... 129 Concentration of Species Present in Mixtures of Bromine Trifluoride and Chlorine Trifluoride ........ . . . . . 137 Values of the Equilibrium Constant for the Bromine Trifluoride-Chlorine Trifluoride System. . . . . . . . . . . . 138 Conductivity of Bromine Trifluoride in Chlorine Trifluoride. . 140 viii LIST OF FIGURES Figure Page 1. Absorption spectra of the 3,4-1utidine-bromine chloride system in carbon tetrachloride solutions . . . . ........ 13 2. Far infrared spectra of pyridine and of the py-Brcl complex ............................. 15 3. Far infrared spectra of 4-picoline and of the 4-pic-BrC1 complex ....................... 16 4. Molar absorptivity of the 3,5-1utidine-bromine chloride system at 305 mp ...... . ................. 18 5. A Kel-F valve ......................... .48 6. A dissembled Kel-F valve .................... 49 7. Kel-F reaction tubes . . . . . . . ............... 50 8. Vacuum line used for halogen fluorides ............. 52 9. Conductivity cell for highly conductive solutions ........ 53 10. Conductivity cell for poorly conductive solutions ........ 54 11. Kel-F pycnometers used to measure densities of bromine trifluoride solutions ................ 55 12. The Raman cell disSembled .................... 58 13. The Raman cell ......................... 60 14. The diamond-window, infrared, short path length cell ...... 64 15. The dissembled diamond-window, infrared, short path length cell ........................... 65 16. Distillation still for bromine trifluoride ........... 69 17. Raman spectrum of liquid bromine trifluoride ......... 103 18. Resolved Raman spectrum of liquid bromine trifluoride. . . . . 104 19. Infrared spectrum of liquid bromine trifluoride ........ 108 ix Figure 20. Raman spectrum of 2.1 M potassium fluoride in bromine trifluoride ..................... 21. Raman spectrum of solid potassium tetrafluorobromate(III), KBrF4 ............................ 22. Raman spectrum of 2.9 M_antimony pentafluoride in bromine trifluoride ..................... 23. Raman Spectrum of solid difluorobromonium(III) hexafluoroarsenate(V), BrFZAsFG, and solid difluorobromonium(III) hexafluoroantimonate(V), BrFZSbF6 . . 24. Raman spectrum of a bromine trifluoride-hydrogen fluoride mixture (0.424 mole fraction bromine trifluoride). ..... 25. Specific conductivity for the bromine trifluoride- hydrogen fluoride system ........... . ...... 26. Raman spectrum of liquid chlorine trifluoride. ....... 27. Infrared spectrum of liquid chlorine trifluoride . . . . . . 28. Raman spectrum of 2.0 M_arsenic pentafluoride in chlorine trifluoride .................... 29. Molar conductivity of the bromine trifluoride-chlorine trifluoride system ..................... 30. Raman spectrum of a bromine trifluoride-chlorine trifluoride mixture (0.64 mole fraction bromine trifluoride) ............ . ........... Page . 109 . 110 . 115 . 116 . 121 . 126 . 130 . 131 . 132 . 134 . 135 GENERAL INTRODUCTION A characteristic feature of the members of the halogen family is the ease with which they react with each other to form a series of interhalogen compounds (Table 1). The number of halogen atoms in an interhalogen molecule is always even and only binary compounds are known. . The chemistry of interhalogen compounds has a long history and dates back to 1814 when Gay-Lussac discovered that iodine monochloride and iodine trichloride can be prepared by direct combination of the two elements.1 The ability of the halogen or interhalogen compounds to form a series of polyhalide ions by the interaction with a halide ion was demonstrated shortly thereafter by Pelletier and Caventou2 who prepared strychninium triiodide. The first halogen fluoride to be prepared was iodine pentafluoride reported by Kammerer in 1862.3 The chemical literature abounds with reports on the behavior and properties of interhalogen compounds. In general, three main avenues of research were followed: 1. study of the formation and physical properties of polyhalide ions; 2. study of donor-acceptor complexes formed by interhalogen compounds; 3. the application of liquid interhalogen compounds as nonaqueous solvents, especially from the point of view of preparative chemistry and of acid-base equilibria. The present research project was undertaken with two goals in mind. The first was to study preparation and properties of molecular TABLE l.--Interhalogen Compounds XY ClF BrF IF BrCl IC1 IBr XY3 ClF BrF w IF IC1 XY5 XY7 ClF5 IF7 BrF5 IF5 3 complexes of bromine chloride which hitherto have been unknown and the second, to apply spectroscopic techniques to the elucidation of acid-base reactions in liquid bromine trifluoride. I. COMPLEX COMPOUNDS OF BROMINE CHLORIDE HISTORICAL INTRODUCTION The fact that addition products may be formed from interactions of stable molecules has been recognized since the early days of chemistry. Much of this early work has been summarized in a review 4 Instances of additive combination between aromatic by Pfeiffer. hydrocarbons and other organic compounds as well as certain inorganic compounds were known at that time, but their existence was in apparent violation of the existing rules of chemical bonding. The stability of these additive compounds varied over a wide range. In some cases stable solid adducts exhibiting integral stoichiometries were formed, while, in other cases, the existence of a molecular complex could only be inferred from changes in color or in some other physical properties when the reactants were mixed together in solution. The whole subject of donor-acceptor compounds has been recently covered by Mulliken and Person5 and Foster.6 Donor-acceptor complexes formed by halogen and interhalogen molecules have received particularly close attention during the past decade. Complex compounds of iodine, iodine monochloride, iodine trichloride, and iodine bromide with a large number of organic electron donors have been either proposed or studied in solutions and their properties have been characterized. Addition compounds with halogen fluorides are practically unknown due to their high reactivity. Of the remaining interhalogens, the information concerning bromine chloride complexes is nearly completely lacking. Bromine chloride has a curious history. The discoverer of 5 bromine, Balard,7 noted in 1826 that when liquid bromine and chlorine are mixed, the color of the solution is quite different from the additive color and postulated that an addition compound was formed. Subsequent attempts by numerous investigators to isolate the resulting compound were unsuccessful and the existence of bromine chloride was disputed for over a century. It was not until 1929 that 8 Gillam and Morton unambiguously identified the formation of bromine chloride in bromine-chlorine mixtures in carbon tetrachloride solutions by Spectrophotometric techniques. Bromine chloride is highly dissociated into its component parts in the vapor phase, in the liquid phase, and in solutions. In fact, it appears to be impossible to isolate the compound in the pure form. 9-12 Many investigators have studied the dissociation of bromine chloride in the vapor phase by various techniques. The values obtained for K vary between 0.10 and 0.14 at room temperature. The first diss attempt to determine the degree of dissociation of bromine chloride in carbon tetrachloride solution by Spectrophotometric techniques was 13 who reported a dissociation constant of 14 made by Barratt and Stein, 0.28. However, it was pointed out by Vesper and Rollefson that Baratt and Stein made a wrong assumption that bromine chloride does not absorb at the wavelengths at which they measured absorption and, therefore, their results were erroneous. A later Spectrophotometric study showed that the equilibrium constant for the reaction 2 BrCl :== Br2 + C12 15 Thus, bromine in carbon tetrachloride solutions at 25° was 0.145. ichloride is found to be quite unstable either in the vapor phase or in solution. 6 Because of this instability, very little work has been attempted in which bromine chloride may act as a Lewis acid. In 1931, Williams briefly mentioned the preparation of a pyridine-bromine chloride complex,16 but no details are given on the properties of the resulting compound. More recently, the preparation of the 4-g;amy1pyridine- bromine chloride complex has been reported by Zingaro and Nitmer.]7 Scott has determined the electron-accepting ability of bromine chloride, relative to other halogens and interhalogens, from the free energy of the formation of the trihalide Br2C1' anion in aqueous ‘8 The following order of Lewis acid strengths for the solution. halogens and interhalogens was obtained: IC1 >> BrCl > IBr >> 12 > Brz >> C12. This work was undertaken in order to study possible formation of bromine chloride complexes with heterocyclic amines and to compare its complexing ability with that of the other halogens and inter- halogen compounds. EXPERIMENTAL PART Chemicals Pyridine: Fisher "Certified" pyridine was refluxed over granulated barium oxide for twelve hours and fractionally distilled through a 60 cm. Vigreaux column. The distillate was then stored in the dark to minimize photo-decomposition. One hundred ml. portions of the product were then refluxed over granulated barium oxide for two hours, fractionally distilled through a 20 cm. helices-packed column and stored over sodium hydroxide pellets until needed; bp760 = 115° 19 (lit. = 115.58°). bp760 All of the substituted pyridines used in the investigation were obtained from the Aldrich Chemical Company. These compounds were used without further purification aside from drying. Bromine: Liquid bromine was obtained from J. T. Baker Chemical Company and was used without further purification. Chlorine: Chlorine gas was obtained from Matheson Company and was used without further purification. Carbon tetrachloride: The purification of carbon tetrachloride 2° ‘9 = 76.75°). has been described previously; = 76.8° (lit. bp760 bp760 Dichloromethane: Fisher "Certified" dichloromethane was refluxed for 24 hours over granulated barium oxide and distilled through a one meter helices-packed column; bp760 = 39.9° (lit.19 bp760 = 39.95°). 8 1,1,2-trif1uorotrichloroethane: Matheson, Coleman and Bell “Spectroquality” 1,1,2-trif1uorotrichloroethane was refluxed for twelve hours over granulated barium oxide and fractionally distilled through a one meter, helices-packed column at a reflux ratio of 8:1; bp743 = 46.8° (lit.2] bp760 = 47.6°). 1,2-Dichloroethane: 1,2-dichloroethane was purified by the method used by Vogei;22 bp760 = 83.4° (1it.‘9 bp760 = 83.483°). Preparation and Stability of Solutions Stock solutions of the halogens were prepared by dissolving appropriate amounts of the halogens in the respective solvents. The exact concentration of the stock solutions was determined by iodometric titration. The solutions were then diluted to the desired concentra- tion. Stock solutions of the respective pyridines were made by weighing the compound into a volumetric flask and diluting with purified solvent. Bromine chloride solutions were prepared by mixing the carbon tetrachloride solutions of the two halogens. In the mole- ratio study, the concentration of the bromine chloride was calculated from the known value of the dissociation constant in carbon tetra- chloride solutions.15 The mixed solutions of bromine chloride and of the respective pyridine were prepared just before each measurement. Contact of solutions with the atmosphere was kept to a minimum, but no attempt was made to do all the transfers in a completely inert atmosphere. Preliminary results have indicated that brief contacts of the solution with the atmosphere did not alter the experimental results. Likewise, it was found that carbon tetrachloride solutions of bromine chloride and the pyridines were stable for at least two 9 hours at room temperature after mixing, while, in general, spectral measurements were completed within 5-10 minutes of the initial mixing. The concentration of the solutions was kept low (110'3 M) in order to avoid formation of the solid complex. Preparation of Bromine Chloride Complexes Equimolar solutions, 0.1 M, of bromine chloride and the respective pyridine were prepared in l,1,2-trifluorotrichloroethane. It was found that purer bromine chloride-pyridine products were obtained if the preparation was carried out at low temperatures. These solutions were cooled in a carbon tetrachloride-monochlorobenzene slush bath (ca. -30°C) and then rapidly mixed in a dry box in an anhydrous nitrogen atmosphere. The precipitate was filtered, washed with purified 1,1,2-trifluorotrichloroethane, and dried under vacuum. The list of the complexes prepared, their melting points, and the analytical data are given in Table 2. In all cases, the complexes were obtained as yellow or light .yellow microcrystalline powders. They slowly react with atmospheric moisture. When the complexes were kept in a dessicator, the pyridine, 4-picoline, and 3,4-lutidine complexes lost bromine chloride after several days. However, the 2,4-lutidine, 2,5-1utidine, 2,6-lutidine, 3,5-lutidine, and 2,3,6-collidine complexes remained unchanged for at least eight weeks. Attempts to prepare bromine chloride complexes with 2-f1uoro- pyridine, 4-hydroxypyridine, 2-chloropyridine, and 2-bromo-5-nitro- pyridine were unsuccessful. It appears that the inductive effect of the halogen and of the OH group decreased the donor strength of the pyridine to the point where the interaction with bromine chloride 10 o.mo~ o.F~F m.pp— m.N_— m.m~_ N.NPP m.~__ m.nm venom N.¢op o.mpp m.~_p m.P_p m.Pp_ m.PPP m.p__ N.mm .ee_eu pcmpm>wamu upcumeouoa eo.e Nm.m mm.e mN.e mm.e aN.e eN.e m~.e eN.e mm.e om.e m~.e eeeed .ee_eo .mmmmummm mm.¢ no.¢ mo.¢ mm.m mm.m mm.m mm.v mo.¢ mo.¢ mo.¢ mo.¢ mo.¢ eased .ee_eu :omoguxr mm.o¢ No.o¢ Fm.mm mm.mm mm.nm mm.mm om.~m mm.mm No.mm mm.~m m¢.mm mu.mm eased .eepeu concmo op .mom .omopumOF ._m> .pwhm .om-mm .eop-~op .wop-eo_ .em-mm .oop-mm .em-em .mm-ea e.eo_-eo_ .Qez Fugmuowauw —ugmu—Founo.m.~ Poem-eeh-m.m _uem-e=4-e.m _uem-ese-e.m Poem-esh-m.m _uem-p=4-e.m _uem-sa cczanou mw>wgm>wgmo mcwuwcxd use wcwunga new: mmxwpqeou mnmcopsu mewsogm--.~ m4m<~ 11 was too weak to produce a solid complex. Attempts to make the 3,4-lutidine complex in dichloromethane and 1,2-dichloroethane were also unsuccessful. Spectral Measurements Ultraviolet absorption spectra were obtained on a Cary Model 14 Spectrophotometer. Either 1 cm. or 5 cm. pathlength cells were used. All measurements were carried out at room temperature of ~25°. The donor/acceptor ratio was varied within such limits so as to give a good spread of experimental points. Spectra of the bromine chloride-Lewis base solutions containing a fixed amount of bromine chloride and a variable amount of the respective base were obtained in the 450-300 mp spectral range. A typical set of spectral curves is illustrated in Figure 1. All pyridine derivatives listed in Table 3 gave well defined isosbestic points. In general, the position of the absorption band of the complex did not change appreciably with the nature of the base. It varied from 304 mu for the 3,4-lutidine-bromine chloride complex to 310 mp for the 2,6—1utidine-bromine chloride complex. Studies of absorbance yer§u§_time were conducted using bromine chloride as the solute in 1,2-dichloroethane and in dichloromethane. It was found that the solutions were Quite unstable and the spectra changed rapidly with time. Apparently, the bromine chloride reacted with the solvent. Far infrared absorption spectra of Nujol mulls of the solid complexes were obtained with a Perkin-Elmer 301 Spectrophotometer. The frequency scale was carefully calibrated in the usual manner by using water vapor as the reference. Typical absorption curves are 12 TABLE 3.--Formation Constants of Amine-Bromine Chloride Complexes in Carbon Tetrachloride at 25° Complex Kf x 10'3 (lit. mole'])* Py-BrCl 1.2 i 0.1* 4-Pic-BrCl 2.1 a 0.2 2,4-Lut-BrCl 6.7 i 0.7 2,5-Lut-BrCl 5.2 i 0.4 2,6-Lut~BrCl 1.5 i 0.1 3,4-Lut-BrCl 4.2 1 0.4 3,5-Lut-BrC1 6.8 t 0.8 * Standard deviations for independent runs l3 l2 l.6"// /0 In 1 E) In 0) ‘4 QDSO F) (D I ABSORBANCE 0; F) 01 1 'fi S) .a ' 1 1V ‘N (12 \ l l l l 300 Figure l. 320 340 360 380 400 X(myl Absorption spectra of the 3, 4— lutidine- bromine chloride system in carbon tetrachlorideusolutions where C 6. 62 x 10 t x 10 M (1 o. oBr $3) 06 628 (3g 1. 24; (4V 2. 47; 5§3 71 6 4. 94; (7)6 ,( ) 7. 42; (9) 8. 65; (10) 12.6; 11) 18.5; (12) 618 14 shown in Figures 2 and 3. The mulls were placed between two polyethylene discs separated by a polyethylene spacer. In order to obtain the absorption bands of the respective pyridines, polyethylene liquid cells obtained from Barnes Engineering Company were used. The instrument was continually purged with nitrogen. The bands obtained for the uncomplexed pyridine agreed closely with those obtained by Frank and Rogers.23 Analyses Microanalyses for carbon, hydrogen, and nitrogen were made at Spang Microanalytical Laboratory, Ann Arbor, Michigan. Iodometric equivalent was determined by dissolving the complexes in aqueous acetic acid and titrating the resulting solutions by standard iodometric technique. Calculations Spectral data obtained on the bromine chloride—respective pyridine mixtures were used to calculate the formation constants of the complexes using the mole-ratio method described in a previous publication.24 The formation constants of the complexes were determined in carbon tetrachloride solutions. In a mixture containing bromine chloride and a base 8, the total absorbance at a given wavelength in a 1 cm. cell is expressed by +eC+eC (l) A BrCl BB cc = EBrClC where €BrC1’ EB, and EC are the molar absorbtivities of BrCl, B, and the complex, and CBrCl’ CB, and CC are the equilibrium concentrations l5 .Auunv xoamfioo Hohmnzm 63p %0 com a v ocfiwflsma mo mhpooam cosmthATgmm .m ohswflm .Ivco on. com 0mm one I _ _ 0 III-.." \\ z 44 I/ e I \ J . a \\l// \ a— _ 1.. ON , x z a _ , x / ~ _ . ._ a x e u n . _. , a . . 1 8 — _ , ~ _ . r \ . L _._ .. . 1 C ... I om .. L . \ ’l\s 11 00 oo. l6 .Anunv onQEoo HomeOHmue one mo use A V ocHHOOHana eo oppoomm nonmsmcflnewm .m onzwfim _.>:o on. com own One _ _ o .llll(\\11, \tlIInud"dudududllllllllllllllflu a, \\. I \ l om ’ \ a s x \ ’ \\|.I.I / ~s 1.. 0.? . . I; . . L z . C x L f.\ .l O@ 1 0m 00. 17 in moles/liter of the three species. If one assumes a 1:1 complex in solution, the absorbance equation becomes A=e (ct -C)+e(Ct-C)+eC (2) BrCl BrCl C B B C C C where CBrCl and C3 are the stoichiometric concentrations of BrCl and of the base, respectively. The molar absorptivity of the bromine chloride complex was determined from a limiting absorbance measurement25 (Figure 4), while 88 was determined from measurements on the solutions of the pyridines in carbon tetrachloride. Once these values were determined, the equilibrium concentrations of the complex, and hence the formation constant, could be calculated from eq. 2. Correction was applied for the dissociation of uncomplexed bromine chloride.15 It has been previously shown that the linearity of the Benesi- Hildebrand plot by itself is not a sufficient criterion for assuming a simple 1:1 interaction between the donor and the acceptor 26 Therefore, calculations were carried out over at least molecules. three different wavelengths over at least a 20 mp range. The results at different wavelengths were in good agreement. 18 .18 mom pm Eopmzm evapoaco oQHEopn 1 ocflvflpsaumem on» mo th>Hth0mno mmaoz .: ohsmfim meagre mzzzomm E mz_o_S.Tm.m do oEE who: 0v Om ON 0. O _ L _ _ 00v 000 CON. loom. l OOON l OOVN 00mm All/\lldEOSBV HV'TOW 19 m mo_ x P.o e N._ no. x F.o e N._ mo. x _.o . N.P x o.o m~.o me.F eeeeeesa. o._m ~.em o.m~ Poem. mmw wee owe 6. Noe.o mcm.c ~_a.o ommwoo.o Nee.o .Ne.o 4mm.o eo~_oo.o Fam.o ome.o mme.o mawpoo.o em~.o mem.o mmm.o ommwooo.o oe~.o F_m.o epm.o e_meooo.o ee~.o mw~.o Pm~.o eemmooo.o N-.o me~.o ee~.o ~__mooo.o mo~.o ~m~.o em~.c oeueooo.o mm_.o eo~.o eo~.o moemooo.o wep.o e~_.o pa_.o emmmooo.o ee_.o ee..o mmF.o eoa_ooo.o map.o eo_.o moo.o Nmmoooo.o eo_.o mmo.o emo.o emeoooo.o .ed oo.m u eemeep eeea -o_ x em.m u Hegemg .mm 1 asee : 3 n e mmma s opm< e mom< eeeeeesau mcwcopsoecum» concmo cw ongsou ouwcopgu mcvsocmumcwupgxa as» co mama oweumeopogaocpumgm--.e mom Py. The exception occurs in the case of the 2,6-1utidine complex, where the order is Pic > 2,6—Lut > Py. This inversion is not particularly surprising. It has been shown that bulky Lewis acids, such as iodine(I) chloride, form stronger complexes with pyridine than with 2,6-lutidine. This effect is due to the steric effect of the two methyl groups in the latter compound. Also, pyridine forms a stronger complex with boron trifluoride than 2,6-lutidine.27 Bromine chloride is not as large a moiety as iodine monochloride or boron trifluoride, therefore, in this case, the 2,6-lutidine complex is still stronger than the pyridine complex, but weaker than the picoline complex. On the other hand, the 2-methy1 group of the 2,4-lutidine and 2,5-lutidine does not interfere with the complex formation and, since these amines are stronger bases than either picoline or pyridine, they form 26 27 stronger complexes with bromine chloride. A comparison of the formation constants for the complexes of iodine (I) chloride, iodine bromide, and bromine chloride indicates ‘8 The that the order differs somewhat from that predicted by Scott. results, shown in Table 11, indicate that the order of Lewis acid strength is IC1 > IBr > BrCl. It is possible that the order may be dependent on the reaction medium. In any case, the difference in the acceptor strength of bromine chloride and of iodine bromide is not great. Far infrared spectra of the pyridine and the pyridine-bromine chloride complexes are shown in Table 12. The analysis of the spectra is complicated by the fact that the amines have numerous bands in this region. Tentative assignments are possible in the case of pyridine which does not have bands below 402 cm.']. 1 In this case, it seems reasonable to assume that the 280 cm." band is due to the Br-Cl stretch. When bromine chloride is complexed, the Br-Cl bond length would be expected to increase and the force constant of the bond would decrease. Thus, the frequency at which the Br-Cl stretching vibration occurs would be shifted to a lower frequency. It is 1 interesting to note that this shift from the 430 cm.’ band of the uncomplexed bromine chloride is much larger than the one observed for the I-Cl stretch which is at 375 cm.'] for the uncomplexed 1 28 molecule and 280 cm.‘ in the pyridine-iodine chloride complex. The shift seems to be even larger for the 2,4-lutidine-bromine chloride complex where the Br-Cl stretch appears to be at 238 cm.']. Analogous to the case of iodine chloride, the “BrCl band, upon complex- ation, becomes broader and more intense. 28 TABLE 11.--Comparison of the Formation Constants of Pyridine Complexes with IC1, IBr and BrCl in Carbon Tetrachloride at 25° Amine Ic1a KfIBra BrCl Pyridine 4.8 x 105 1.3 x 104 1.2 x 103 2-Picoline 8.9 x 105 2.4 x 104 2.1 x 103 2,6-Lutidine 8.9 x 104 3.7 x 103 1.5 x 103 aReference 24 29 em-z 6 Foucm n Npm .n .mmm_ .a .z .coumucecd ..oo ucegpmoz =6> .emanumpoz oweoueeo eo meoumam: .mgmnNLmz .mm 6eN_ e_N mmN.NeN mmm poem p34 m.m QON nomN NNN omm mNe mNm oceepuah m.m ONe Npm oem Now Poem1uzhue.m mmN N_e Pom eNm Nam mcpupuah e m uomp mPN nFmN mNe emm mom ~ucm-uahuo.N omN N_e emm mem mcpeppzh o N NFN .me Pom mem Pocmuuahum.N ONN NmN 0mm Poe mNe eNe New mcpupuah m. N oeNF am me mme oem _ucm-u:4-e.N em_ PNN PmN Noe mNe m_m ”mm mcpepuah e N 6_emN_ am omN Fwe New _ocm-uwa e Az>v__m Nee m_m eeepe6ea-e 6_HNmF ae omN NNe Fmo Fogmuzm Noe Noe 6=ee_csa some unecopcu mcweogm Apusuv mecca uczoaeou copmmm Eu om— New on» cw mmxmpaeou meeeeesa eeeeeeemesm eee meeeeesm -mupcopsu mcpeogm mo mecca coppagoma< umcmcmcH--. Np m4m0000 .3 .eemmeeeg6m .26 .Anompv mpmp .m ..Em;u .mgocH .FFmZUom .m .> can ..Lw .xo—Lovczw .u .u .066_e= .0 .0 .00605200 .2 .0 .0006600 .0 .0 .20000 .0 .0 .066360 .0 .0 .606: .3 .060>000__0 .0 a .Auompv ..> .z .xso> 36: ..6:H mmmgd owemumu< .0 .Fo> .exgumwemsu cmmopmze ..um .ccmEpau .> .emmuwocspa cmmopmz mo mm_pgmqoca Pmuwswcu use p06060500 .cempm .46 A.ee_ ..0_0 N.~ mp.m muop x e.m e.mN mp.N wucmpmazm mean mm 60000660 go: come.N m-o_ x _.m m.mN _m.F omom.N m-o_ x _o.m N.mN mP.P mucmpmnam mean we mpnmpomw poc mN.F A.0e_ ..m~-0 ~00._ ---------- 0.00 ---- emom._ mrop x m.e op.mN nmm.o N10. x m.— o.wN Pwm.o A—-.EU ragov A.=.mv Ammxnmuv AemN .vwaawP .ps\mv cow wucou “cmpmcou pcoeoe summcoo owewuwamm.copaogh mpoawo Add . m.oop om.pe mm.mNP o.e—u mn.__ —.oo_u 0e _e060 Na.e 00.0 A.6600 0N- m.oou mm.m mop- Nm.ouu w.mm—n as N00 000 6000 00 dem 020 020 0_0 0_0 uFQ muwcoapm :mmopm: ammuwcozpu cmmopm: 6:» mo mmwpgmaoca qupmxzd weomnu.m_ m4mHe> miaox < .m ohswam 00m 000m Dec—D.» Imbom coo: 09.25 octet 3:00 :25 $95 2660 “7.3. 050.com cote... 398m 3m 08.3w .:m .25 .n..z «0-0.x 5:00 52m .:.m ._5m Ill" IT _ 8:66 0:3 6665 32 056.com $95 i 595 6.1.3. 49 ‘— VALVE HANDLE (BAKELITE) .—PACKING NUT (BRASS) -—TEFLON PACKING a ._ KEL-F STEM TEFLON O-RING SPLIT COLLAR (BRASS—r Figure 6. A dissembled Kel-F valve. 50 .393. coapoeon muaom .N. 90de a C. . l 9 0.9.15: r . or a t . r p I. I a a 1. .h A .l ,1 1’ pl . 51 The above discussion has listed the basic equipment needed for work with bromine trifluoride and with halogen fluorides in general. Figure 8 Shows the vacuum line used in this work. Conductivity equipment Figures 9 and 10 show conductivity cells for highly conducting and poorly conducting solutions respectively. In the former case, a 1, while a representative cell shown in Figure 10 had a cell constant of 0.0418 cm.". typical cell constant was 2.76 cm.‘ In each case, the body of the cell was made of Kel-F, the bright platinum electrodes were made from platinum sheets, and the leads pass through Teflon which is compressed between metal plates to effect a vacuum-tight seal. All measurements were made in a closed system. When the measurements were being taken, the conductivity cells were placed in an ice bath or in a thermostat at the appropriate temperature. Since the thermal conductivity of the rather heavy plastic body is quite low, temperature equilibrium is established rather slowly. Conductivity measurements were made on a Wbyne Kerr Auto-Balance Precision Bridge Model B331. The precision of the instrument is a 0.1%. Actual precision of the measurements is probably : 1.0%. Density measurements A Kel-F pycnometer was constructed to make density measurements, (Figure 11). It was calibrated with carbon tetrachloride. The measure- ments were usually made on about Six m1. of solution. The precision of the measurements was i 0.6%. The error lay mainly in the fact that the pycnometer had to be handled excessively in putting it on and off the vacuum line to make the weighings. w n . no I“‘! " I: . . In. I"D‘H“II‘DOI¢I O I C ‘I Q‘s - .......x...c... 52 Vacuum line used for halogen fluorides. Figure 8. PLATINUM ELECTRODE“ 53 /—PLATmuu - 10% RHODIUM i FORCINC PLATE TAPERED TEFLON PLUG (.31 T 1/4' HOKE NICKEL omennasu VALVE FLUOROTHENE SINTERED DISC 13>: 3/4" NICKEL FLARE FITTING FLUOROTHENE SLEEVE FLUOROTHENE TUBE PLATINUM ELECTRODE FORCING PLATE PLATINUM LEAD Figure 9. Conductivity cell for highly conductive solutions. 54 Thermocouple fl— Electrode Leads Leod Forcing Plate bl Tapered Teflon Plug .__ 3/4" Nickel Flare Fitting <—- Fluorothene Tube Teflon Electrode : Aligning Block I I415. :3>‘ Electrodes Figure 10. Conductivity cell for poorly conductive solutions. .mcoapSHOm wuanosamanp mcasoyn mo mmfiufimcwo mgsmmmfi 0» com: mumpmsocozn minM .HH onsmfim ,. :35. 4...? . 56 Pressure Measurements A Monel Bourdon-type gauge (from Helicoid Gauge Co.) was ' incorporated into the vacuum system and was used for pressure meas u rements . THE APPLICATION OF SPECTROSCOPIC TECHNIQUES TO THE STUDY OF BROMINE TRIFLUORIDE SOLUTIONS Spectral Cells for Visible and Ultraviolet Spectra Aluminum oxide in the form of Linde synthetic sapphire is trans- parent in the ultraviolet and visible regions and can, therefore, serve as window material. Of equal importance is the fact that it is chemically inert toward the halogen fluorides. The body of the cell is made of Kel-F. A recess is machined into the metal backing pieces to hold the sapphire window which is pressed against the Kel-F body as the screws are tightened. With a carefully machined and cleaned Kel-F body, no gasket is needed to effect a vacuum-tight seal. A Teflon gasket may be used, but it is possible to effect a satisfactory seal without its aid. Sapphire windows are l/l6" thick. In assembling the cell, the sapphire window may crack when the backing screws are tightened only slightly more than is required for vacuum tightness. Therefore, care must be taken in the assembly of the cell. Raman Cells for Liquid Samples As in the case of the ultraviolet and visible spectral work, it was found that synthetic sapphire was the easiest material to use for the windows. Figure l2 shows the dissembled cell. The body of the cell was made of Kel-F. The sapphire windows were obtained from Perkin-Elmer Corporation coated with a reflecting material in order to 57 58 .anpsmmmflw HHmo adewm one .NH mesmfim 82.8 835834 N 3353 2235 Eozom .0 3353 2238 no... .0 flmxmow .3on .¢ 32m E35534 .m E t: .u no... 635524 ._ 59 increase the pathlength of the incident light beam through the cell, except for a portion of the top window which was left uncoated to allow for the passage of the incident light beam. Thus, the incident light beam would enter the cell through the part of the top window which had not been coated with the reflective material. This beam would then be reflected back and forth through the solution in the cell by the reflective coatings, thus increasing the accuracy of the measured spectrum. Since sapphire is brittle, Teflon gaskets were inserted between the Kel-F body and the windows and between the windows and the outer metal pieces. The assembled cell is shown in Figure l3. The cell was mounted on a platform which sat on three set screws. These set screws could be adjusted to vary the height and the tilt of the cell, thus putting the cell in the position in which the intensity of light going to the detector was maximized. Raman Cells for Solid Samples These cells were merely lengths of l/8" 0.0., l/32" thick Kel-F tubing with one end closed off and leveled so that the incident beam of light could enter through this end with a minimum of scattered light. The other end of the tube was flared and a Kel-F valve was attached to the tube through this flare. As in all the previous cases, this cell was vacuum-tight. In a later part of this work, when an Argon ion laser became available, regular sealed l/4" 0.0. Kel-F tubes were used as both solid and liquid sample cells. 60 #1- "I”. ‘- The Raman cell. Figure 13. 6l Infrared Spectral Cells This part of‘the work presented the biggest experimental problem. Only certain materials are transparent in the region that we wished to investigate andvof these common window materials, all were known to react rapidly with liquid bromine trifluoride. Numerous attempts were made before a successful cell was built. These attempts will be summarized below. Infrared Cell with Irtran-Z as Window Material: Irtran-Z, obtainable from Eastman Kodak, is a polycrystalline zinc sulfide which is supposed to be resistant to corrosive halogen fluorides. However, it is not transparent below 900 cm.'] and, therefore, it was only of very limited usefulness in this work. We found that Irtran-Z windows were indeed resistant to bromine trifluoride and were able to reproduce the reported spectrum of liquid bromine trifluoride in the overtone region.52 Since Irtran-Z is resistant to bromine trifluoride, it was felt that it could be used to detect the HF2' vibrations (lOOO-ZOOO cm.']) in the study of the hydrogen fluoride-bromine trifluoride system. It was found, however, that the.solutions of hydrogen fluoride in bromine trifluoride rapidly attacked the window material and, therefore, no spectra could be obtained. Infrared Cell with Irtran-6 as Window Material: Irtran-6, also obtained from Eastman Kodak, is a polycrystalline CdTe, which was claimed to be inert towards corrosive liquids, such as bromine l trifluoride and, in addition, be transparent down to 200 cm.‘ . When liquid bromine trifonride was put into a cell with Irtran-6 windows, 62 however, it immediately reacted with the windows. Therefore, we were unable to use Irtran-G as a window material. An Infrared Cell with a Reflectance Attachment: Since Claassen, 51 gt_gl, obtained an infrared spectrum of gaseous bromine trifluoride using silver chloride windows, it was assumed that a spectrum of liquid bromine trifluoride could also be obtained provided that only the vapor came into contact with the windows. Claassen, et_al,5] did notice that the windows were slowly attacked by the bromine trifluoride vapor so that they could be used only once. Therefore, it was thought that a cell could be assembled with which the infrared spectrum of liquid bromine trifluoride could be obtained before the windows were destroyed. For this end, a reflectance set-up was devised. A Model 134 Horizontal Reflectance Attachment was obtained from Barnes Engineering Corporation. A nickel body was made in which a thin layer of liquid bromine tri- fluoride could be placed. The incident beam could then be reflected into the cell through the silver chloride windows and reflected back off the bottom of the cell cavity. In practice, this method was unsuccessful, since the bromine trifluoride vapor rapidly attacked the silver chloride windows and the attack on the windows was much too rapid to make any spectral measurements. Infrared Cell with Teflon Sheetingiand Kel-F Grease: This technique was similar to the one employed by McDowell and Asprey73 in their work on bromine pentafluoride vapor. A very thin coating of Kel-F grease was applied to the silver chloride windows and a piece of l/4 mil. (0.00025") thick Teflon was placed over the windows. However, liquid bromine trifluoride leaked through the thin, porous Teflon sheet, dissolved the Kel-F grease and then rapidly attacked the 63 silver chloride windows. Thicker Teflon sheet could not be used because the material has only a limited transparency in the 300-700 cm.'] region. An Infrared cell with Diamond Windows: It was finally found that diamonds could be successfully used as the window material and with this cell, it was possible to obtain infrared spectra of corrosive liquids in the 200-900 cm."1 spectral region. The diamond windows were obtained from General Industrial Diamond Company.. One diamond was an irregular oval with diameters of the major and minor axes being 0.37" and 0.26", respectively. The other diamond was approximately 0.28" in diameter and more nearly circular. The cell body (Figures 14 and 15) was machined from Kel-F with 0.38" diameter central hole and a surrounding circle of six l/8" bolt holes. The diamonds and spacer could be set in the hole with very little extra space left over. Stepped Teflon plugs were constructed to fit snugly through the hole of the cell body. The lip resting outside the hole in the cell body was 0.l2" wide. The plug body was 0.38" in diameter, essentially identical to the diameter of the cell body hole. A hole of 0.20" in diameter was drilled through the Teflon plugs to allow for the passage of the light beam. A thin spacer, cut from Teflon sheet, Was placed between the diamond windows and the cell assembled between metal plates. As the bolts were tightened, the Teflon plugs were squeezed against the diamonds and the cell body, forming a vacuum-tight seal. In moSt of our research with bromine trifluoride, we have used 0.0005" in. thick Teflon for the spacer, which is thin enough to permit observation of all but the most intense bands. Thicker spacers, of course, would be easier to handle. 64 smeoH sung ho . .Haoo p am Umhmhmcfi .3ovcH3iucosmHu one .rfi mMSMHm 65 (5) Kel-F Body (6) Split Collar Nut (7) Kel-F Tubing (8) Assembled KeI-F Valve (I) Metal Plate (2) Teflon Plug (3) Diamond Window (4) Teflon Spacer (4) THE DIAMOND WINDOW, SHORT PATH LENGTH, INFRARED CELL Figure 15. The dissembled diamond-window, infrared, short path length cell. 66 The top of the cell body was constructed to form a l/4" male flare. A drilled hole 3/32" in diameter connected this opening and the space between the windows. A valve was attached to the flare by means of a small piece of Kel-F tubing. The assembled cell was attached to a vacuum line where it was filled and emptied. Pressure could be applied to the solution in the cell to insure that the narrow space between the diamonds was filled. After each run, the cell was taken apart and cleaned with water, acetone, and anhydrous methanol. A blank was run on the cell before each filling. A strong band was observed at 485 cm.'], which was due to the diamond windows. The cell was mounted on a stand which was movable in all three dimensions. By adjusting the stand's parameters, the light being transmitted through the cell could be maximized. All spectral runs were made at speeds of 8 or 3.2 cm 'l/min. in order to insure proper pen response. Although this cell has only been used in our bromine trifluoride work, it seems reasonable to assume that it can be used for other, more volatile, corrosive liquids. Infrared Instrumentation Infrared spectra were obtained in a Beckman Model l2 Infrared Recording Spectrophotometer which was fitted with a beam condenser attachment in order to increase the light intensity transmitted through the sample. Neutral density filters were used in the reference beam. 67 Visible and Ultraviolet Instrumentation Ultraviolet and visible spectra were obtained on a Cary Model 14 Recording Spectrophotometer. Raman Instrumentation Raman spectra were obtained for bromine trifluoride solutions and solid samples on a Cary Model 8l Raman Laser Recording SpectrOphotometer equipped with a He-Ne laser. Some measurements were also made on a Spex l40l Raman Laser Recording Spectrophotometer equipped with an Ar ion laser, when the instrument bacame available. EXPERIMENTAL PART Chemicals Bromine Trifluoride: Bromine trifluoride was obtained from the Matheson Company. It was purified by two fractional distillations. The first distillation was carried out in the Chemical Engineering Division of Argonne National Laboratory. The second distillation was carried out in the still shown in Figure 16. The nickel pot containing the impure bromine trifluoride was placed in a ceramic furnace wrapped with heating tape. This pot was kept at 80°C. The Kel-F column was wrapped with asbestos and heating tape and was kept at 50°C. The column was filled with Kel-F chips. This system was connected to the vacuum line. The condenser was kept at -l92°C. The first 20% of distillate was discarded as it probably contained large amounts of highly volatile impurities, such as hydrogen fluoride, bromine monofluoride, bromine, and bromine pentafluoride. The last l0% of distillate was also discarded as it may have contained some metal impurities. The middle fraction was saved and further purified by pumping at room temperature for two hours to remove any remaining volatile impurities. The purified product was pale yellow in color 74). It was stored in and had a melting point of 8 8°C (lit = 8.77°C Kel-F containers. Chlorine Trifluoride: Chlorine trifluoride was obtained from Matheson Campany. It was purified by two vacuum distillations at low 68 69 de. luori trif e bromin i n still for llat a Disti 16. Figure 70 temperature. The purified product was a very pale straw yellow liquid, with a specific conductivity at 25° of 4.8 x 10'9 ohm" cm." (lit. = 4.9 x 10'9 ohm" cm.'1 67). Hydrogen Fluoride. This reagent was purified by a method suggested 75 Commercial hydrogen fluoride obtained from by Runner, Ml. Matheson Company was absorbed on anhydrous sodium fluoride, and the resulting NaHF2 was heated in vacuo at l50°C to remove volatile impurities. Hydrogen fluoride was then regenerated by heating at 400°C and collecting the gas in a nickel vessel. The product was further purified by distillation from a still which was made entirely from Kel-F, except for a platinum cold finger. This procedure resulted in anhydrous hydrogen fluoride with a conductivity below l x l0'6 ohm-1 cm.‘1 at 0°C. Arsenic Pentafluoride: This reagent was obtained from Ozark- Mahoning Company. It was purified by vacuum distillation in a Kel-F vacuum system at low temperature. Antimony Pentafluoride: This reagent was obtained from Matheson Company. It was purified by treatment with anhydrous sodium fluoride in a Kel-F vessel. Following this treatment, the volatile impurities were removed by maintaining the product under vacuum for three hours. The compound was distilled twice under vacuum. Alkali Metal Fluorides: The anhydrous sodium, potassium, rubidium, and cesium fluorides were of reagent grade quality. They were dried for 24 hours at 140°C prior to use. Preparation of Soltuions Depending on the nature of the salute, the solutions were prepared in different ways. 7] The simplest case is that of a stable, non-volatile solid, such as an alkali metal fluoride, to be dissolved in bromine trifluoride. The dried sample was weighed into a Kel-F vessel, which was then connected to the line through a valve. The reaction vessel-valve system was weighed first empty and then with the added soluble material. The system was evacuated and bromine trifluoride was added directly into the reaction vessel. The vessel-valve system was removed from the line and weighed again to determine the bromine trifluoride content. The antimony and arsenic pentafluoride solutions were prepared by distilling the proper amount of solute into a tared reaction vessel. After reweighing the vessel, the required amount of bromine trifluoride was added to the vessel, whereupon the vessel was weighed once again. For spectral measurements, solutions for the hydrogen fluoride- bromine trifluoride system and the chlorine trifluoride-bromine trifluoride system were prepared in the same manner as the acid solute solutions. A slightly different method was followed for the solutions used for the conductance measurements. Since the slightest impurity will increase the specific conductance of hydrogen fluoride, or of chlorine trifluoride, the vacuum system had to be rigorously protected from water vapor in the atmosphere, even when vessels were taken off the line to be weighed. Therefore, a flow of helium was sent through the system whenever a vessel was taken off the line to be weighed. Transferring of Solutions in the Vacuum Line The following sequence of operations was employed in transferring liquidfrom the preparation tube to a cell where a measurement is to 72 be made. Three valves, one connected to the tube containing the solution, one to the cell to be filled, and one to the vacuum line, are connected to a Y joint through flexible Kel-F tubing. The valves to the vacuum line and the cell are both opened and that part of the system is evacuated. Both valves are then shut. The cell containing the solution is raised so that the attached valve points downward. When this valve is opened, the pressure above the halogen fluoride solution forces the liquid into the Y and associated tubing. Enough liquid is allowed to run into the tubing to fill the cell.‘ The valve is then shut. The cell is then made the lowest point of the system, the valve to the cell is opened, and the liquid drains into the cell. The valve on the cell is then shut, the tubing and Y are evacuated, and the cell is removed from the line. The error in concentration introduced by this procedure is limited to the loss of halogen fluoride or hydrogen fluoride vapor within the Y and tubing. With the vapor volume comparable to the liquid volume, this procedure would amount to a few tenths of a percent and, in practice, an error of one percent might result. Preparation of Solid Compounds Potassium tetrafluorobromate(III), KBrF4, was prepared by adding a slight excess of bromine trifluoride to anhydrous potassium fluoride in a Kel-F reaction vessel on a vacuum line. The reaction was allowed to proceed at room temperature. The excess bromine trifluoride was then removed under vacuum. The residue was in the form of a solid white,microcrystalline powder. The melting point was 332°C (lit, 330°C 49). 73 Difluorobromonium(III) hexafluoroantimonate(V), BrFZSbFG, was prepared in the same manner as KBrF4. A light yellow, microcrystalline solid was obtained with a melting point of 128°C (lit49 = 130°C). The yellow coloration was probably due to some bromine trifluoride which was absorbed on the solid and could not be pumped off. Difluorobromonium(III) hexafluoroarsenate(V), BrFZAsFG, was prepared on a vacuum line by slowly distilling arsenic pentafluoride into a Kel-F reaction vessel containing some bromine trifluoride. An excess of arsenic pentafluoride was added and the reaction allowed to go to completion. The excess arsenic pentafluoride was pumped off and a white solid remained. Found: Br 25.3%, As 26.2%, Calc. for BrFZAsF Br 25.2%, As 23.7%. 6: Due to the corrosive and reactive nature of BerAst, it was difficult to prepare for analytical analysis. This was compounded by the fact that fluorine interferes with the determination for arsenic. Thus, it was felt the above numbers were within experimental error. Calculations In order to determine quantitatively the equilibria existing in bromine trifluoride-hydrogen fluoride mixtures, it was necessary to obtain quantitative relationships between the integrated intensities of the bands due to BrFZ+ and BrF4' ions and their concentrations. Increments of potassium fluoride were added to liquid bromine tri- fluoride and the changes in the integrated areas of the 528 and 455 cm."1 Raman bands of the BrF4' ion were observed. The integrated areas were measured by multiplying the height by their half-width. It was found that the integrated areas were doubled when a spectrum of 0.91 M KF solution in bromine trifluoride was obtained. The following ratio 74 was used to calculate the concentration of BrF4' in basic solution: Integrated area of BrF4 band in basic solution Concentration of BrF4' in basic solution ’= Integrated area of BrF4 band in pure BrF3 Concentration of BrF4' in pure BrF3 From this equation it was found that pure bromine trifluoride is about 0.9 M_in BrF Therefore, the concentration of BrFZ+ should also be 4 0.9 [L Since the concentrations of ions in pure bromine trifluoride were now known, it was possible to determine the concentrations of the Ber+ and BrF4' ions in the bromine trifluoride-hydrogen fluoride and bromine trifluoride-chlorine trifluoride systems from the following relationship: Integrated area of BrF ' (or BrF2+) band in HF-BrF3 (or Cl 3-BrF3) system - + Concentration of BrF (or Ber ) in HF-BrF3 (or ClF3-BrF3) system Integrated area of BrF4' (or BrF2+) band in pure BrF3 Concentration of BrF4' (or BrF2+) in pure BrF3 75 In this manner the concentrations of BrFZ+ and BrF4' were found in the bromine trifluoride-hydrogen fluoride and bromine trifluoride-chlorine trifluoride mixtures. The concentration of HFZ' anion in the first system was found by subtracting the concentration of BrF4' ion from that of the BrFZ+ ion. The concentration of Cle+ cation in the second system was found by subtracting the concentration of BrFZ+ ion from that of the BrF4' anion. From the above data, it was possible to cal- culate the equilibrium constants for the fluoride ion transfer reactions in the two systems [BrF2+][HF2'] [C1F2+][BrF4'] K = and K = 1 IBrF3IIfiFI 2 IC|F3IIBrF3I since the initial concentration of bromine trifluoride, chlorine trifluoride, and hydrogen fluoride were known. The data and results are given in Tables 16, 23-28, 32-36, 46, 47, 50, and 51. 76 TABLE 16.--Raman Data for Liquid Bromine Trifluoride Bandcgosgtion ¥g¥;(BrF3) Relative Half-width Integsitid IntenSity (cm ) Ratio 673 1.00 78 22 625 0.09 5 28 581 0.65 29 40 531 0.12 15 17 528 0.41 17 50 490 1.14 28 70 455 0.14 6 40 428 0.50 15 70 341 0.23 11 34 265 0.08 4 30 249 0.02 2 10 236 0.14 5 30 77 cc 0 ep.o cu m ¢P.o emu om m mo.o on m mo.o mew co m pp.o es s mo.o mom mm m_ o~.o mm o_ om.o Fem em a —m.o mm o m~.o mme om op Fm.o on o— mm.o mme om w_ Pm.o om m_ om.o ome mm mm .wm.o so mm e..P mmm mp mm mp.o m_ op mp.o —mm oc mm w¢.o me om n¢.o _mm em mm oo._ om No oo._ mum 3.3. 0.53. xuwmcmucu sawmcmpcm Apisov A—45uv sawmcmpcm umumcumucm apisuv xuwmcmucH umumcmmucu cowppmoa gou_z-c_o= o>cnm_om Amlcmv_.\> gos.z-cpo= o>loa_om Amocmv_>\> scam ma; 5 e. .2. 3 no... 5 e. y. 3 .H .muwco=_cwch mcpsocm cm mcopuapom mvwcozpm Eswmmmpom co$ mama :msmmui.n_ m4m

wuopmm Ammcmvpax> guvwzimpmz m>wumpmm Ammcmvp>\> vcmm mlcm c. l¥.m mc.o mics a. L¥.m c_.l .HH .oc_co=_c_cc aspects c_ neocoslom occcoz_L Escmmoooa can some =o5mm--.mp msmcpo_om Amocmvpo\> cou_3-cpm= o>compoa Amocmv_>\> scam macs c. amu.m mm._ macm =_ lam.m m.~ muwcoapwwc» mewsocm c. mco_o=Aom ouccos_L assumes“ as. ouALospl 53.nou can some =oga¢--.mA u4mAoa_o¢ Amocmv_>\> goeAz-cpo= o>Aoo_om Amlcmvpp\> scam macs a. maam.m A.~ macm a. mlsm.m m.~ .H .oeccos_c_ce occeocm c. neocoapom newcoapcoocoa seoewocq cot some =agmx--.o~ mom<~ 81 mm o ep.o on o eA.o 0mm es s As.s as s as.s sss ss s.s ss.o cm s ss.s ass as __ s_.s cm s sA.s Ass es s_ s¢.s ss s ss.s sss ss ss ss.s sc s_ ss.s sse s_ s_ e_.s s. s sc.s Ass ss ss ss.s mm s_ ss.s Ass as s sc.s ss cc ss.s sss s_ s_ 4A.s cm s_ ss.s sss ss ss ss._ ss ss ss.A mas occss occss Auwmcmpca zu_m=maca A -Euv A isov AuAmcmucfi emuscmmucH A isov sawmzopcu couscmmucH cowpwmoa accs_;-c_ss o>ccscos Asacsv_>\> sssAz-aAs= occcc_os Asacsv_>\> sass sacs cc sass.m ss.s sacs cc sass.m e.c .ss .ssccosccccc occsocs cc scocc=_os osccosAcsccoa csoecc=< coc moss amass--.A~ asssa 82 TABLE 22.--Raman Data for Arsenic Pentafluoride Solutions in Bromine Trifluoride 3.0 MAst in BrF3 Band V/V (BrF ) Relative Half-w dth szgtipn Igtéggagéd Intensity (cm' ) Ratio 673 1.00 68 30 674 0.27 18 28 625 0.46 22 40 581 0.25 11 42 531 0.12 ' 11 17 490 0.25 7 60 428 0.28 7 65 374 0.18 10 30 341 0.18 9 34 265 0.07 3 32 236 0.14 4 45 83 em N—. FN.o em FA AN.o Fem es AA ss.s . ss ac ss.s sss as s __.s as s s..s sss ss ss es.s . ss ss ss.s sse ss A. ss.s ss ss ss.s sss sc NA _A.s s_ s_ __.s css ss ss ss.s ss . ms as.s css ss s.s sc.s ss s s_.s sss ss ss so._ as sA os.c sss occss ocoss sucscoccs sucscoccs A_-sov as... gas ass. as... mass“... sass sacs cocsoscc spas ssa.o sacs coccosca o_oe Ass.s .H .muscozAcPLP mcweocm was aupco=Aa cmmocuxr mo mmczuxaz com sumo assumiu.m~ m4m sacs coccoscc o_oe sss.s .HH as em NA me e_ ee m NA em on N_ AA o. oe mm em A em em isov xuwscmucH simps: m>spsAmm m mm.o mm.o e_.o om.o me.o m_.o mm.o m~.o oo.A osusm sawscmucH umpscmaqu Asacsv_>\> acs coccoscc sacs sss.s Fem mme mme ome mmm Amm Amm mmo msm say A - =o_ocsoa ecsm .mumcospcwcs mewsocm acs av_copa :mmocuxr no macapxwz com sumo :smeii.em m4msuspmm A~.o oe.o -.o mm.o ~m.o ~_.o mm.o em.o oo._ opusm AusmcmucA cmuscmmucm Asacsv r<> acs coccoscc vos ese.s A scsA on on oe mo om mp oe oe em isuv mp mm susscmpcA z-cAss occcsAos m mA.o Ae.o Am.o mm.o me.o NA.o mm.o mm.o oo.A opusm Ausscmpcm umuscmmucA AsacsVA>A> acs coccoscc vos sAs.s Fem mme mme om? mum pmm —mm mum mum AAiEov cocoasoa sass .mvscoapcwcs mcssocm vcs mcscoapa :mmocc»: co smcauxwz coc some :sEsm--.mm asses 86 on A Fm.o mm A om.o Fem mm A Ae.o mm A me.o mme oe e NN.o em e A~.o mme mo eA om.o oA MA om.o ome om m mm.o om m mm.o mum o_ A NA.o oA o NA.o Amm oe AN . mm.o oe mA Am.o Amm me m.o mm.o mm o mm.o mmm em me oo.A om we oc._ MAo osusm ospsm AuwmcmucA AuwmcmpcA A -Euv Apiwuv AuwscmpcA emascmwij . Awmmovs Awwmcwumm umumcwwwcw cowwwmoa cues: cps: m>wuspmm A acmv A sue. 4A I .p A m A a m A u m macs =o_puscm mpoe mem.o macs :owpuscm ons AAm.o .>H .mvsco:_cwcs mewsocm ucs muscoaAa cmmocuxr co macauxwz com sumo csssmii.mm mAmwuspmm A acmVA>A> guuwzicAs: m>ApsAmm AmacmVA>A> cams sacs coccoscc ocos AAA.s sacs coccosca voa sss.s .> .waLO:FL.T:. GCwEOAm flcm mUwLozpu— cmmOLbe— .._.o me3HXwZ Lou... duce :mEmmll.NN MAm<._. 88 oe m— Ne.o mum oe o mp.o mum om we oo._ on sacs coccosca ocos sss.s ss AA 4A.s as s ss.s sss ss s. ss.s es s sA.s sss ss ss ss.A sA ss ss.A sAs owusm osusm Ascscoccs Ascscoccs A -sov Awieuv Auwscmpcm uauscaoucu A isov AusscmucA umpscmmpcA cowpwsoa cos.z-c_s= oscsscos AsacsVA>A> scsAz-a_s= s>ccsAos AsacsVA>A> scss coccoscc vos ssA.o sacs coccoscc vos sAA.o .H> .muwcozAcAcA ocpsocm was descenpa camocnsr mo smcapxwz so» muse assumii.w~ u4mA> scsAz-a aAss o>AcsAss AsacsaA>A> scss sacs coAcoscc oAss ss.o sacs coccoscc voe ss.s .AA .osAcosAcAcA ocAcoAss scs osAcosAcAcA ocAsocs as socschz coa scss cs5ss--.ss asssA 94 em eA mm.o mm oA om.o wme oe A sA.o mm s AA.o sse Nm mm so.A ms es so.A ose me sA As.o Am oA os.o ssm sA NA mA.o sA s mA.o Ass oe AN ss.o sm mA Ae.o Ass om e mo.o om m mo.o mms em oA oo.A om om oo.A mAs om me Am.o mm mm As.o sAm om NA Np.o om- s Np.c smA ospss oApss AuAscmucA susscmucA A -suv A usuv AuAscmucA emuscmmpcA A -Eov suAscmucA umpscmmucs :oWAAsoa sss_3-cAss scAssAss AsacssA>A> css.;-cAss s>AssAss AsacsaA>A> sass macs :oAuoscA mAaE As.o macs :oAposcA one ms.o .AAA .ssAcssAcAcA scAcsAss scs ssccssAcAcA scAescs as sscssxcz csa sass csass--.ss asssA 95 on e me.o Ns s se.o sNe NN N AA.o sN m AA.o mme oe oA om.o oe mA ss.o ome mN m oA.o Nm s os.o st sA e eA.o eA s eA.o Amm oe s os.o oe NA NA.o Ass oN A so.o 0N N so.o st om m.eA so.A ON Nm so.A mAs sN sm mm.m Nm NA eo.s sAs NN sAmA ss.o ON sA ss.o smA oAuss oAuss summcmpcA AuAmcmpcA A isuv AAusov AuAmcmucA couscsmpcA A -Euv AuAmcmucs umuscmmucs :oWuAmoa sssA:-aAs= s>AcsAos AsacsVAsas assA3-cAss s>AssAss AsacssAsas scss sacs csAsosca oAse As.s sacs ssAssscc sAss ss.s .>H .mossoapmwsk acho—zu vcm mvwsoapmwch mchoss mo mmaszwz com mums cosmmii.mm m4m

AssAss AsacsOA>A> assAz-aAss s>AssAss AsacsOA>As sass sacs asAsosca sAsa sA.s sacs asAsoscc sAsa N~.s .> .vaLozAAALA chaoAsu was muAgozAAALA chanm so mmsauxAz Low mums Emzmmii.mm u4sH0mmm .mH asswsm .120 CON OOn 00¢ 80 00m 00A é . _ \ mm. mmm A 31m A mNm Sn mmc .mm omc ALISNBLNI _ _ a _ scs 105 its intensity increased with increasing temperature. It seems reason- able to assume, therefore, that this band may be due to a bromine tri- fluoride dimer or, in any case, a lower polymer than that represented 1 by the 428 and 581 cm.' bands. The infrared spectrum of bromine trifluoride is shown in Figure 19 and the vibrational assignments are listed in Table 41. Comparison is made with the Raman spectrum of the compound as well as the spectrum 76 of gaseous bromine trifluoride observed by Selig, et al. The spectrum obtained for the overtone region agrees well with the results of 52 There is a strong band at 490 cm.-1 Haendler, et_al. in the infrared spectrum of liquid bromine trifluoride. While its behavior has not been followed as a function of temperature, its intensity decreases upon the addition of an acid or a base. It is probably due to the same species which shows a Raman band at 490 cm."1 This evidence suggests the presence of a dimer without a center of symmetry and a fundamental 1 vibration at 490 cm." in both the infrared and Raman spectra. The Raman spectrum of a 2.1 M_potassium fluoride solution in bromine trifluoride is shown in Figure 20. The addition of potassium fluoride to bromine trifluoride resulted in a sharp increase in intensity of the 528 and 455 cm.“ 1 bands as well as in the intensity of the much weaker 249 cm." band. A comparison of the spectrum of the tetrafluorobromate(III), BrF4', anion with the spectrum of the quasi- isoelectronic xenon tetrafluoride78 is shown in Table 42. The Raman spectrum of solid potassium tetrafluorobromate(III), KBrF4, is shown in Figure 21. After the completion of this work, Shamir and Yaroslavsky79 published the Raman spectrum of cesium tetrafluorobromate(III). As shown in Table 42, their results clearly agree with ours. Three Raman-active fundamentals were observed for BrF4', two 106 acmsaOAss< OA moamcmamss Om>cmsno so: AEV NeN 3.s mmN sm>cmsso no: ssscssas sea Aas sss z sss Aaa sss ssscssss sea Ass sss.sss.sss a Ass Asa ass s.a Nss Ass Ass.Ass a.a Ass A3>a sss ass NAs Assa Ass.sAs.sss ssccssss sea Assv aAs s.a sAs Asa sss.sss s.a sAs Ass sAs :sEss OmcscAcA amass smcscacA ssacs sasssss sacs sassAs woAaosAwAaA wcAans OAOOAA we Ezapumam AccoAuanA>ii.Ae msscmsso so: OAm mme meN Nom mNm sAcss sacs s acs sAssAs aA acs sA sA s m>AuuscA s sA s sm>cmsaO econ: sA mocmcwamss pcmscsAms< :oA:< AAAAkuOEogaoaozAAOLumA Ac Eaguumnm AscoAuOAnA>ii.Ne usso .AAssAA NAss cm ..aoas .scsaA .scssassss .s .a sas asses .a .sa .AsssAA sAss «Mm ..asas .s .ass .ssssa .a sas .assaass .z .3 .csscss .s .as NA NA NA m NA m NA Om>cmsao mcmaz + x AzaA sss AzsA ssN AssNA so ass AsAsA so Nss Asssa A. Asa ass AsssA as AaA ass Assmv N> Om>cmsso no: ANAsA N> Om>cmsso Ho: x sss sss ox AsaA sss AsaA sss AsssA As sss Asssa s. Asa ass AsssA A» sss Asscc s. A.a Nsc Nacs asaascs aA -saas asAssAss sass ssasassAs aA -sasa asAssAss sass mossAos sAoa saAaAscass ssAcosAcAcA saAascs sAssAs as aacaooas Asascsscacs--.ss assaA 114 TABLE 44.-~Raman Spectra of Difluorobromonium(III) Cation BrF2A5F6 solid BrFZSbF6 solid BrF2+ 734 (6) 709 704 (100) x 694 677 (75) 598 637 (68) 573 550 (78) 520 520 (24) 381 490 (22) 363 33:) 308 (9) x 280 (19) 273 (13) 234 (7) .mOHAOSHmssp massosp cs messosamwfiamg hcoEspSs .Am 0N .Ho 5590QO anewm .NN asswsm 75 cos oos ooe com com OOA A A A A A 115 sacs a. sass was A A A A A ALISNBLNI 116 B'Fz As F6 . (la/LA 1 l l l I l 800 600 400 200 INTENSITY n Ber Sb F6 M w l l l l 1 700 600 500 400 300 200 CM" Figure 23. Raman spectra of solid difluoro- bromonium(III) hexafluoroarsenate(V), BrF AsF , and solid difluorobro- mongum éII) hexafluoroantimonate(V), BrF2Sb 6o 117 species with C2v symmetry, with a total of fifteen fundamental vibrations, all Raman active, might be expected. The observations summarized in Table 44 are similar to, but not 82 identical with, those of Christe and Schack. The only assignment that seems certain for the 8er+ cation is that for the symmetric 1 stretching vibration at 704 and 709 cm.’ in the antimony and arsenic complexes, respectively. Tentative assignments, however, are also 1 given in Table 45 for the asymmetrical stretch, 03, at 734 cm.’ and the bending mode, 02, at 308 cm.']. Thus, while the square planar BrF4' anion is found unequivocally in both solid and solution, the nature of the cation in acid bromine trifluoride solutions is not spectroscopically similar to the species found in solid complexes and the shape of the cation in solution cannot be readily inferred from the vibrational spectra. While for the solid, the x-ray evidence for a bent Ber+ cation is conSistent with the spectroscopic observations, in solutions, the species present is apparently somewhat different. It is probably due to the fact that the BrFZ+ cation is strongly solvated. With salvation, the Br-F bond distance would increase and the force constant would decrease. There- fore, the observed frequency for the Raman active stretching vibrations 1 would decrease from ~700.cm.- to ~625 cm.‘1 Preliminary work on the spectra of arsenic pentafluoride solutions in chlorine trifluoride has shown that the v] fundamental vibration for C1F2+ is observed at 782 cm.'], which is about 25 cm.’1 below its frequency in the solid 81 CleAsFG. cause the v3 and v] fundamental vibrations to be decreased by about 1 In bromine trifluoride, a stronger solvating effect might 80 cm.’ from where they occur in the solid.' 118 TABLE 45.-Vibrational Spectrum of Difluorobromonium(III) Cation + +. +. , +‘ BrF2 in BrF2 1n .ClF2 in CD?2 in a . BrF2SbF6 Liquid BrF3 ClFESbF6 Liquid ClF3 704 R b 625 R 805, 809 R 782 R 705 IR 635 IR 308 R 387 R 830R a - Reference C. b + . This work, BrF2 1n BrFeAsF6 119 Bromine Trifluoride-Hydrogen Fluoride System A typical Raman spectrum of a bromine trifluoride-hydrogen fluoride mixture is shown in Figure 24 and the observed bands are tabulated in Table 40. It should be noted that the v] symmetric stretching vibration of the HFZ- cation, which should occur at 600 cm.'],83 was not observed. It is quite possible that the band is so broad that it cannot be resolved. Woodward and Tyrell84 did not observe a1600 cm.'] band in aqueous solutions containing the HFZ' anion. We tried to find this band in a 2 fl_potassium fluoride solution in hydrogen fluoride without success. In essence, therefore, the addition of hydrogen fluoride to liquid bromine trifluoride resulted in changes in the intensities of some bands originally present, but not in the appearance of new bands. In order to determine quantitatively the equilibria existing in bromine trifluoride-hydrogen fluoride mixtures, it was necessary to obtain quantitative relations between the integrated intensities of the bands due to the Ber+ and BrF4' species and their concentrations. We first determined the concentration of these ions in pure bromine trifluoride by adding measured amounts of potassium fluoride in liquid bromine trifluoride and observing the change in the integrated areas of 1 the 528 and 455 cm.’ bands. For example, it was found that for a 0.91 M_KF solution, the intensities of the two bands were doubled. It was estimated that the concentration of BrF4' and the concentration of the BrFZ+ ions in pure bromine trifluoride was about 0.9 M, The concentration of BrF4' and Ber+ ions was, therefore, calcu- lated for each of the solutions prepared and, assuming electrical neutrality, the concentration of HF2' ion was calculated by difference. 120 om.o mm.o 00.0 mm.o mA.o Am.o Nm.o me.o ce.o eN.o mA.o AA.o oo.o a: sA.o ---- sN.s sA.s ss.s sA.s ss.s ss.s ss.s NA.A ss.s AN.A As.s sA.A As.s sA.A AA.s sA.A ss.s sA.A ss.s As.A ss.s ss.A ss.s ss.A As.s As.s -sacs +sacs AcssAAAssAsaA coAuscucmucou uAcoA N.Ns A.ss A.As s.Ns A.sN s.AN s.os s.AA s.sA N.sA s.A s.s s.A ss.s cscAAAssAoa as O.N A.e w.e o.A N.oA m.AA o.NA A.mA m.eA o.mA m.AA m.mA 0.0A m.ON cmAAAAsonE sss.s ssA.s sAA.s sA.s ss.s ss.s As.s ss.s Ns.s ss.s sA.s ss.s ss.s OO.A :oAuuscA onE sacs aoApscucmocos AsApAcA msAcoaAAAcA ocAEocs sas mOAcosAa :mOocOA: so smczuxAz :A pcmsmca sonwOm co :oApscpcmucosii.Oe sAssposncoo OHAAumam .mm osswsm sacs 20.55... 32,. 00.. omd 00.0 ovd 0N0 o . o _ A A A I N l . s F o . .. . . m 0 o O. O 1 . . o «W . oo o m o o r04 0 o I m .. o. A A _ A .-wo awwo (9.01 X) 04100 03:15 127 The molar conductance of the system also fell from the value for pure bromine trifluoride to a minimum at ~70 mole percent bromine 'trifluoride. From this point, it increased steadily to a maximum at ~5 mole percent bromine trifluoride, went through a minimum at .2.5 Inole percent bromine trifluoride and again increased with dilution (Table 48). It seems reasonable to postulate that the initial addition of hydrogen fluoride to bromine trifluoride disrupted the chain conducting inechanism operating in pure liquid bromine trifluoride. Therefore, although the total ionic concentration in the solution increased, both the molar and equivalent conductances decreased until ionic concentra- tions of 8er+ and HFZ' are sufficiently large to contribute significantly to the conductance. In going from 70 to 30 mole percent bromine trifluoride, the equilibrium favored increased formation of ionic species. The mixture appeared somewhat less viscous and an increase in the mobilities of the ions must be expected. In solutions containing less than 30 mole percent bromine trifluoride, the total concentration of ions decreased with increasing concentration of hydrogen fluoride and the conductivity also decreased. The peculiar behavior of the molar conductance in the 5 to 2.5 mole percent bromine trifluoride solutions might be due to the formation of a new and possibly less mobile species as indicated by the appearance of the 662 cm."1 Raman band. As we approach pure hydrogen fluoride, bromine trifluoride becomes a simple base and, as expected, its molar conductance increased with increased dilution. 128 Bromine Trifluoride-Chlorine Trifluoride System The Raman spectrum of liquid chlorine trifluoride is shown in Figure 26 and the bands are listed in Table 49. In general,.the results agreed with those of Jones, et a1.85 with the exception of a strong band at 502 cm."1 which we did not observe. The infrared spectrum of liquid chlorine trifluoride, which does not appear to have been reported previously, is shown in Figure 27. The Raman and infrared spectra are in essential agreement with the recently published results for gaseous chlorine trifluoride.76 No extraneous bands have been found in the liquid which could be attributed to ionic or polymeric species. The absence of ionic species in any appreciable concentrations is in agreement with the low specific conductance of the 1iquid.°7 Figure 28 shows the Raman spectrum of a 3.1 M_solution of arsenic pentafluoride in chlorine trifluoride. The elimination of bands attributable to the solvent and to the AsF6' anion left only one band at 782 cm." which cannot be attributed to the two above species. There is little doubt that this band is due to the v] symmetric stretching vibration of the difluorochlorinium(III), ClF2+, cation formed in the reaction + + AsF ' C1F 2 6 3 + AsFS—A= ClF It is interesting to note that the band is observed at a lower frequency in solution than in the solid complex ClFZAsF6 (806 cm.'1).81 The same behavior has been observed for the BrF2+ cation. The decrease in the frequency of the 0] band in solution of the two cations may well be due to their salvation. In such a case, it would be expected that the Cl-F and Br-F force constants would decrease and shift their bands 129 :3 A. :3 as :3 s. AAsA s5 AssA ss AAsA s» ucmEOOAmm< s ssA .ssA s .a NsA s> NOA .o .c a sss .Nss as .a sss : sas : Ass s st .o .c s sss z.a Ass.Ass OmcscwcA amass smaAO snowmss OA mocmcwwmss sm>cmsno so: u .o .c s ssA ANNA ssA s> mOA .o .a a sAs AssAA sAs a sAs AsA sss a ass .6 .a a st AsA ass OmcscAcA csEss saAs sAssAs ssacssAcAcA oaAcsAas sAsscA as ascsosas AsasccscsA>--.ss assaA 130 00m .ovssosamssp aCHAOHno Usswsa mo adhpoomm assmm .mN osswsm . I .2 0 00m 00¢ 000 000 00A. 000 A A A A A sas sas: AllSNBlNI 131 I m 1 1 I I T Liquid 01F, INTENSITY 1:? Jill L==l 800700600500'400300200 CM" Figure 27. Infrared spectrum of liquid chlorine trifluoride. 132 .mcssosHAssp massoano as mOHAOSHmspch 3:8me m o.N mo 85.50on 888m .wN onsmfim 00m 00¢ 000 .150 A A 000 00k 000 A A A saa aa sass mas _ A A ALISN31NI 133 to lower frequencies. Much of the comparative behavior of the two liquids can be described in terms of weaker interaction between chlorine trifluoride molecules and their derivative ions than between bromine trifluoride molecules and their derivative ions. A typical Raman spectrum of a bromine trifluoride-chlorine trifluoride mixture is shown in Figure 30. In essence, the addition of chlorine trifluoride to bromine trifluoride did not produce any substantial change in the spectrum. When the spectra were resolved, however, it became evident that the decrease in the intensity of the BrF4' bands was definitely smaller than that predicted on the basis of dilution factor. This observation is consistent with the postulated existence of the equilibrium + ClF ——‘«—ac1F+ + BrF- B"F3 3 2 4 It should be noted, however, that no band was observed at 782 cm.-1 which would correspond to the C1F2+ cation and, therefore, the equi- librium lies largely to the left. Chlorine trifluoride is probably a much weaker fluoride ion acceptor (Lewis acid) than bromine trifluoride. Many bases are known in bromine trifluoride while the C1F4' anion has been detected only 8 . . 6 There is no ev1dence as when stabilized in a few specific lattices. to their comparative strengths as fluoride ion donors, and even the evidence for their relative acid strength is qualitative. A very superficial approach based on their comparative bond strengths87 would suggest that chlorine trifluoride is a stronger base (weaker average bond energy, better donor) than bromine trifluoride, but the slight extent to which the equilibrium is demonstrated in solution indicates that the overall difference is small. 134 1000* 100 5 MOLAR cowoucnvmr (cmzohm-h x103 '8 0.1 o I l l I 1.00 0.80 0.60 0.40 0.20 0.00 MOLE FRACTION Br F3 Figure 29. Malar conductivity of the bromine trifluoride - chlorine trifluoride system. 135 .Aosseossascc panache. cospoahm macs :93 0.3538 ocsnosfimshu ocssosso 1 ovshogfishp 0550.3. 8 mo 85.50on 356m .1 5 cos 8.. com com 02. .Om ossssa 000 A A A A A sacs asccosca as: 5.0 A A A A ALISNBLNI 136 The concentration of the BrF4' and BrFZ+ ions in the liquid mixtures was calculated from the values of the integrated areas of the Raman bands. The concentration of the C1F2+ cation was calculated by subtracting the concentration of the BrFZ+ ion from that of the BrF4' (Table 50). Since the initial concentrations of chlorine trifluoride and bromine trifluoride were known, the value of the equilibrium constant (C1F2+)(BrF4') (C1F3)(BrF3) could be calculated. The results of these calculations are shown in Table 51. These values are only approximate since a small error in the determination of the integrated areas of the BrFZ+ and BrF4' bands is compounded in the calculations of the equilibrium constant. Neverthe- less, the values indicate the correct order of magnitude of the constant and show that chlorine trifluoride behaves as a very weak base in bromine trifluoride solutions. The calculations also show that the maximum concentration of the ClFZ+ cation in the mixtures is < 0.1 M_ 1 band due to and, therefore, too low to observe the band at 782 cm.- this species. The infrared results for the bromine trifluoride-chlorine tri- fluoride system agreed with the Raman data. The change in the bromine trifluoride spectrum relative to pure bromine trifluoride is due to the diluting effect of chlorine trifluoride. A plot of the molar conductance of bromine trifluoride ver§u§_the mole fraction of bromine trifluoride in the chlorine trifluoride-bromine thifluoride mixture is shown in Figure 29. In agreement with the 68 results of Toy and Cannon, the conductivity increased rapidly from 137 No.0 NN.o ON.o e.mA O.N AmA.o ss.s ss.s ss.s N.sA s.s sAs.s ss.s As.s ss.s s.AA A.s sAs.s ss.s ss.s ss.s s.AA a.s ssa.s As.s ss.s ss.s s.A A.sA sas.s ss.s Ns.s ss.s a.A s.sA sas.s As.s NA.s AA.s s.s A.sA sss.s As.s NA.s AA.s A.s s.AA sas.s +NaAO reacs +Nacs cmuAAAsvoe cmuAAAsons coApusca ons AcssAAAssAsaA saAs sacs coApscucmucoo uAcoA :oApsaucmucou AsApAcA ssAcssAaAcA saAcsAas sas ssAcssAcAcA chEocs so smcsprz chEocs :A ucmsmca smAuoam AO :oApscpcmuaosui.Os sAs