\ - _ SPECTROPHOTOMETRIC STUDIES OF BROMINE COMPLEXES Thesis for TIM Degreo OI pk. D. MICHIGAN STATE UNIVERSITY Henry Barber 1958 THESIS 1'? (‘3 \J I CK.) This is to certify that the thesis entitled SpectrOphotometric Studies of Bromine Complexes presented by Henry Barber has been accepted towards fulfillment of the requirements for Ph.D. degree in Chemistry 3 Major professor E5 I Date 15 July 1958 0—169 PLACE IN RETURN BOX to remove this checkout from your record. To AVOID FINES return on or before date due. MAY BE RECALLED with earlier due date if requested. DATE DUE DATE DUE DATE DUE 00"03 6/01 c:/C|RC/DateDue.p65‘p. 15 “wright by Henry Berber 1960 SPECTROPHOTOMETRTC STUDIES OF BROMINE COMPLEXES By Henry Barber A THESIS Submitted to the School for.Advanced Graduate Studies of Michigan State University of.Agricu1ture and .Applied Science in partial fulfillment of the requirements for the degree of DOCTOR OF PHILOSOPHY Department of Chemistry 1958 ACKNOWLEDGMENTS The author wishes to express his sincere appreciation to Professor James C. Sternberg for his guidance and assistance throughout the course of this investigation. Appreciation is also extended to the Research Corporation for a Frederick Gardner Cottrell grant which provided financial assist- ance during the academic year 1957-1958. ii —-'_-‘._-.‘ _.__M‘—~ .v .3... “~ 5.8.4.: in . no. ' .‘V:. v ‘| r VITA Henry Barber candidate for the degree of Ioctor of Philosophy Dissertation: "Spectrophotometric Studies of Bromine Complexes." Outline of Studies: Major subject - Physical Chemistry Minor subjects -- Physics, Mathematics Biographical Items: Born, May 16, 1923, Mechanicville, New York Undergraduate Studies, B. 3., University of Kentucky, 19h8-l952 Graduate Studies, Michigan State University, 1952-1958 Experience: Graduate.Assistant, Michigan State University, 1952-1957 Special Graduate Research Assistant, Michigan State University, 1957-1958 Member of American Chemical Society, Society of the Sigma Xi iii To Liz SPECTROPHOTOMETRIC STUDIES OF BROMINE COMPLEXES Henry Barber ANAABSTRACT Submitted to the School for Advanced Graduate Studies - of Michigan State University of Agriculture and Applied Science in Lartial fulfillment of the requirements for the degree of DOCTOR OF PHILOSOPHY Department of Chemistry Year 19 58 Approved qqflm, ‘4 C . WW3) I'- tn 4 v. vufl "'0 :‘1? or v v v‘r. I >“Ap-‘r3 ' V. ‘54 a...» p . O 37"» I; "“C so. . .. fa n.‘v v ,.‘_ ..._-. ABSTRACT A spectrophotometric study was undertaken to investigate the nature of complexes involving molecular bromine in solution. Initial studies of the complexing of bromine in aqueous solution were made with the system, watermbromine, which shows, in addition to the visible bromine absorption, a short wavelength band in the approximate wave- length region, 2550-2650R. These early studies led to the major. portion of the present work in which the formation equilibria of the tribromide and dibromochloride ions in aqueous solution at 250 were investigated. Bromine complexing with acetonitrile and ethylene dichloride was also investigated spectrophotometrically. With the water-bromine system, the interpretation was necessarily qualitative because of the instability, as indicated by the time- dependent Spectral changes, of the very dilute bromine solutions. Assuming that the low wavelength band is due principally to the presence of the tribromide ion, which is formed through bromine hydrolysis, a reaction scheme was then postulated to account for the observed changes. The system, water-bromine-sodium bromide, was ultimately chosen f or quantitative study of the tribromide ion equilibrium; formation of the tribromide ion results from the reaction, Br2 + Br- = Br; Ultraviolet. absorption by the trihalide ion is characterized by a very intense band with a maximum at 2660K. A method was developed for vi treatment of the data, which led to values of 3.h6 x 104 and 11.3 liter mole.1L for the molar extinction coefficient at the band maximum and the equilibrium constant, respectively. The value for the: equilibrium constant is in agreement with those obtained by other investigators using non-spectrophotometric methods based upon distribution experiments. The dibromochloride ion equilibrium study was first attempted with the system, water-bromine-sodium chloride. However, a complex reaction sequence seemed to prevail, as evidenced by the non-concordant results. It was decided to‘modify the Spectrophotometric approach to the dibromo- chloride equilibrium by study of the system, water bromine-sodium chloride-sodium bromide; the undesirable effects are suppressed by addition of the bromide. Through use of the data already obtained for the tribromide study, this system did lend itself to quantitative treatment, deepite the presence of both the tribromide and dibromochloride ions. The molar extinction coefficient at the band maximum (2b3of) and the equilibrium constant were calculated to be 2.19 x 104 and 1.39 liter mole-1', respectively. The value for the equilibrium constant is in excellent agreement with those resulting from non-spectrophotometric methods. The system, acetonitrile-bromine, is characterized by an intense Ultraviolet absorption with a peak at 2690.8. As with the water-bromine Study, a quantitative treatment proved unfeasible due to time-dependent Spectral changes. Nevertheless, some significance could be ascribed to the observed changes. The absorption is believed to be due to the ion vii pair, CH3CNBr‘Br3‘. The postulated equilibria for the system, including the formation equilibrium for the ion pair, can lead to an explanation for the spectral changes. The ethylene dichloride-bromine system also exhibits an intense absorption in the ultraviolet region of the spectrum. The maximum of this band could not be observed because of the limited transparency of the solvent itself. It is probable that the 1:1 complex, CICHBCH201Br2, is the absorbing species. The work with ethylene dichloride also led to a novel means for purification of this solvent; the method should prove feasible as a means of obtaining Spectroscopic-grade ethylene dichloride. viii TABLE OF CONTENTS :mnnumrmnn.u.u.u.n. ...... ”.n. ...... n.”.n.u ...... .n PART I Tribromide and Dibromochloride Ion Formation Equilibria in.Aqueous Solution HmmMMLammununuunnnnunnnnnuunq ...... .H The water-Bromine System.................... ....... ........... The'Water-Bromine-Bromide System.............................. The‘WaternBromine-Chloride System............................. Emalnation.of Activity Coefficients........................... Introduction...........,................................... Bromine in Aqueous Salt Solutions................... ...... . Hydrobromic Acid in.Aqueous Sodium.Bromide.... ...... ....... Hydrobromic Acid in.Aqueous Sodium Chloride................ Conversion.of.Activity Coefficients........................ THE PRESENT INVESTIGATION-RESULTS AND DISCUSSION................ The water-Browne systemOOODJOOOO'O‘OO0.0000000‘OOOOOOOOOOOI‘OI The'Water-Bromine-Sodium Bromide System....................... Determination of e2 - Hydrolysis Neglected (Method 1)... Determination of oz -Hydrolysis Considered (Method 2).. Determination of K3 (Method 3)................... The'Water-Bromine-Sodium Chlorége System...................... Determination of K4 and 82380” Hydrolysis Neglected' (Method l).............................. ............ .. Determination of K4 and 82380- Hydro ysis Considered (Method 2).............................................. The'WaterfiBromine-Bodium Chloride-Sodium.Bromide System........ mtrOductionOOOIOOOOO. IDIOIOOOOOOIOCO. 000000 0.0.0.0009 ..... Determination of e 00 for Tribromide Ion........ Determination of tfiéa Ratios, 2% 2380/82660 and ezsoo/ezobo for Dibromochloride Ion................................. Determination of e , e g , and e for Dibromo— Chloride Ion'.O??WCCO?OWOCOODOC?@IO00.0.09...0...... Detemimtion 0f K4... ..... 00.00.000.000.000000000000000... Determination.of Concentrations, Absorbances, and Extinction Coefficients of the Individual Trihalide Ions TABLE OF CONTENTS - Continued Page mpmmM-ICOOIOOOOIOOOOO ..... ..OOOOOOOOOOOOOOOOOOOOO00.0.0.0... 80 Source and Purification of Reagents . . . . . . . . ................... 80 water 0 O O O I O O O O O I I O O O O O C O O C 0 O O O O O O O O O O O O O O O O O O ......... O 0 O O O 80 Bromine C O O O O O O . O O O O O O O D O O O O O O O D O 9 0 O O O O O O O O O I O O O O O O O I O O C O O O O 80 Mtg D O O I O O 5 O O O O O O O O O I O O O O O O I O O O O O O O O I O O O O O O O O O O O O O O O O O D D I O 81 Apparatus..................................................... 81 Spectrophotometers, Absorption Cells and Cell Holder. . . . . . . 81 Apparatus and Method for Preparation of Solutions ............. 87 Apparatus and Method for Filling the Absorption Cell. . . . . . . . . . 89 Determination of Volume of Solution and Volume of Solvent. . . . . 92 Analysis of Solutions for Bromine Content. . ....... . . . . . . . . . . . . 9); Conversion of Activity Coefficients . . . . . . . . ............... . . . . 96 SWY.‘....I......OO'OOOOOOOOOOIOOOO- OOOOOOOOOOOOOOOOOOOOOOOOOOOO 100 PART II Bromine Complexing in Several Organic Solvents HISTORICAL sumsr ................. 10h Pertinent Past Studies.................. ................. ...... 10h um PRESENT TNVESTIGATIONuRE‘SULTS AND DISCUSSION. . .. ........... . 107 The Cyclohexane-Bromine System ...... . 107 The Acetonitrile-Bromine System.... ...... .. 10? The Ethylene Dichlorid e-Brontine System. ...... . ......... . ...... 111 The fluorocarbon-Bromine Systems.-. ..... 113 31(me Source and Purification of Reagents... ...... .. ..... . 115 Cyclohex'ane....................... ............... ..... ..115 Acetonitrile..........................'..................... 115 Ethylene Dichloride... ........ . ........... ....... ....... 116 Apparatus.................... ..... ........... .....117 SMARY.....,.. ............ . .................. 138 LITERANRECITED OOOOOOOO 00.00.00... 000000 .....OOOOCCOOOOOOOOOOOQCI121 -.-v-§ n .1 dd J“ v—« -s.. -.. TABLE I. II. III. VI. VII. VIII. IX. X. IXI. LIST OF TABLES Determination of the Molar Extinction Coefficient, 82660’ for Tribromide Ion. Hydrolysis Neglected.......... ....... Determination of the Molar Extinction Coefficient, ezooo’ for Tribromide Ion. Hydrolysis Considered................ Determination of the Formation Equilibrium Constant, K3, and the Molar Extinction Coefficient, e , for Tri- _ 2660 bromide Lanoootoooooooooooooooooooooaooooooooooo‘ococo-coo Determination of the Formation Equilibrium Constant, K4, and the Molar Extinction Coefficient, 82380’ for Dibromo- _chloride Ion. Hydrolysis Neglected. ....... ............... Calculated Values Used for Determination of the Formation Equilibrium Constant, K4, and the Molar Extinction Coefficient, e , for Dibromochloride Ion. Hydrolysis 2380 . GonSideredOOOOOIOOOOOOOOCIOOOO'OOOOO_OO.....OOOOOOODOOOOOOO Calculated Values Used for Determination of the Formation Equilibrium Constant, K4, and the Molar Extinction Coefficient, e , for Dibromochloride Ion. Hydrolysis 2380 conSideredoooooonoooo-on.coo ooooo ooooococooétoo-0.0000000. Data Used. for Determination of the Molar Extinction Coefficients, e2500 and 82380’ for Tribromide [on......... Calculated Values Used for Determination of the Ratio of the Molar Extinction Coefficients, e /e for Dibromochloride Ion.................ngg..?§§9i........... Calculated Values Used for Determination/of the Ratio of the Molar Extinction Coefficients, e e , for Dibromochloride Ion.................ggqg..?§§9.......... Pertinent Data for SYStem: Hzo-BTZFN8C1-N8Br............. Calculated Values.Used for Determination of the Molar Extinction Coefficients, 82380 and 82660’ for Dibromo- Chloride [onOOIOOCOII.0O.......OIOOOOOOOOOOOOO0.00.0...0’. xi Page 32 35 h3 M7 N8 62 65 LIST OF EKBLES - Continued IABLE XII. XIII. XIV. XV. XVI. XVII. Page Calculated Values Used for Determination of the Molar Extinction Coefficients, e and e , for Dibromo- Chloride IronOOOOOOOO0.000.?5W00600?§§9000000..OOOOOO;OOOO 6'] Concentrations of the Various Species Present in System: HZO-Brg-NRC1-NaBI'...“...................o......o..e....... {2 Calculated Absorbances of Tribromide and Dibromochloride Ions in Studies of System: HZO-Bré-NaCl-NaBr..... ...... .. 73 Molar Extinction Coefficients‘of the Tribromide and Dibromochloride Ions at Various'Wavelengths......... ...... 7d Absorbance Corrections for the One Millimeter Quartz Cell and Solvent (water) at Various Wavelengths................ 8h Results of Check on Optical Bath Length of the One Milli- meter Quartz Cell by Use of a Reference Solution.......... 85 XVIII. Absorption Data for Evaluation of the Relative Optical XIX. XX. XXI. XXII. Path Lengths of the One Millimeter and One Centimeter WtzGel-1300.......IIOOO'OOCOODIOOO'OOIOO0.0.0.0.. ..... O 86 Calibration of Volumetric Solution Flask............ ...... 93 Data for'Computing Salt Concentrations of the Solutions... 95 Analysis of Solutions for Bromine Content................. 97 Summary ofoctivity Coefficient Values and Related Data... 98 xii FIGURE 1. LIST OF FIGURES Visible and ultraviolet absorption spectrum of a dilute aqueous solution of bromine............................... Ultraviolet absorption spectrum of hypobromous acid in aq‘leous SOlutionOOOOOO......OOO'..DOOOOO'OO00.0.0000!0.00.0 Effect of light on the ultraviolet absorption of a dilute aweous SOlution Of bromj—neOOOO000..........OOOOOOOOOOOOIO Ultraviolet spectral changes exhibited by a dilute aqueous solution of bromine when light is excluded from the system Ultraviolet Spectral changes exhibited by a dilute aqueous solution of bromine when system is first studied in the "light" and then in the "dark."........................... Ultra‘iolet absorption Spectrum of the HZO-BrZ-NaBr system at 25 COOOOOIOOOOOOOOOO..O'OOQOOOOOCOO....OOOOOO'OOOOOOO’. Test of the linear relationShip used for determination of . the formation equilibrium constant, K3, and the molar 10. extinction coefficient, 62660’ for tribromide ion......... Ultraviolet absorption spectrum of the H20—Br2-NaCl system at 25 C..........'............'......'I...’.IO........'O.'. Test of the linear relationship used for determination of the formation equilibrium constant, K4, and the molar extinction coefficient, 82380’ for dibromochloride ion. Hydrolysis neglected.........,...................... ...... Test of the linear relationship used for determination of the formation equilibrium constant, K4, and the molar ' extinction coefficient, 82380’ for dibromochloride ion. 11. 12. mdrOlij-s conSideredOOOOOI.0.0.0.0000.00.000.00.000...DI.0 Test of the linear relationship used forvdetermination of the ratio of the molar extinction coefficients, e2380/ 8266O"f°r dibromochloride ion............................ Test of the linear relationship used for determination of the ratio of the molar extinction coefficients, 82500/ 82660’ for dibromochloride ion....... ..... .. ..... ......... xiii Page 2h 26 28 h2 h9 59 LIST OF FIGURES - Continued FIGURE 13. lb. 15. 16. 17. 18. 19. 20. 21. 22. Page Test of the linear relationship used for determination of the molar extinction coefficients, e and e , for dibromochloride ion.................2889......€é§9........ 66 Test of the linear relationship used for determination of the molar extinction coefficients, e and e , for dibromochloride ion.................?599......géég........ 68 Analysis of ultraviolet absorption spectrum of the H20- Br2*NaC1-N&BI‘ SYS‘bem at 25 Coooutcast-00000000000000 oooooo 77 Molar extinction coefficientawavelength relationship for tribromide and dibromochloride ions....................... 78 Modified cell holder for the one millimeter quartz cell... 83 Apparatus for preparation of solutions.................... 88 Manipulation of bromine in.vacuum system. ....... .... ...... 90 .Apparatus for filling the absorption cell...... .......... . 91 Ultraviolet absorption spectrum of a dilute solution of bromine in acetonitrile................................... 108 Ultraviolet absorption spectra of ethylene dichloride (Spectrum,n) and of a dilute solution of bromine in ethylene dichloride {Spectrum B)...a............. ..... .... 112 xiv INTRODUCTION Theoretical and spectroscopic considerations suggest an ionic character for certain excited electronic states of aromatic molecules. .A study was undertaken to demonstrate, by chemical behaviour, the ionic character of these excited molecules. It was sought to promote a photo- chemical ring bromination of an aromatic compound containing an aliphatic side chain, through irradiation with wavelengths of ultra- violet light selected to correspond to the appropriate absorption band of the aromatic compound; normal photobromination, in which the actinic radiation is absorbed by the bromine, leads to side-chain bromination of such substituted aromatic compounds through a free-radical mechanism. A first requirement for the study was an inert solvent, one trans- parent throughout the ultraviolet region; the high reactivity of bromine seriously complicates the problem of finding such a solvent. The solvents investigated included cyclohexane, ethylene dichloride, and acetonitrile, the latter two being studied the more extensively. While considerable effort was expended, and a novel purification procedure was devised for one of the solvents, ethylene dichloride, the aim of Obtaining a suitable one was not achieved. Several fluorocarbon solvents were tried but those readily available were also found to be unsatisfactory. The use of acetonitrile led to a particularly complicated inter- amlion with bromine, involving initial complexing followed by ion pair formation and dissociation; the system seemed to defy quantitative study. In an attempt to obtain a simpler prototype for the acetonitrile— bromine system, the trihalide ions in aqueous solution were selected for study. The trihalide study itself has become the major portion of this investigation, and accordingly is presented first in this thesis. The other phases of the study are presented in later sections. A spectrophotometric study was made of the formation equilibria of the tribromide and dibromochloride ions in aqueous solution. This study has led to a determination of the formation equilibrium constants and molar extinction coefficients for these two trihalide ions. The numerous earlier studies of these trihalide equilibria have, with two exceptions, been based upon distribution experiments rather than Spectrophotometric methods, and detailed data on the light absorption characteristics of these ions have been largely lacking; such data are essential to photochemical studies, to theoretical interpretations of the light absorption process (including the nature of the ground and excited states), and to possible analytical applications of spectro- V photometric methods to these systems. > . The review of past work, and the various phases of this study, Will be subdivided according to the apparent complexity of the system. We will first consider the binary system, water-bromine. PART I Tribromide and Dibromochloride Ion Formation Equilibria in Aqueous Solution HISTORICAL SUP/MARY The Water-Bromine System The ultraviolet absorption.spectrum of the system, water-bromine, does not appear to have been interpreted satisfactorily. The following remarks of Bovis (1) seem to be typical of those present throughout the literature: Bromine in water exhibits two absor tion maxima, one of which lies at 1:1ch and the other at 2600.8. Since only the former is exhibited by gaseous and by liquid bromine, the band at 2600 must be characteristic of bromine in solution. Figure l, a spectrum taken from the paper of Katzin (2), shows the visible and ultraviolet absorption of this system. Figure 2, a spectrum also reproduced from the literature (3), represents the absorption of hypobromous acid in water. The latter has been included because of the important role of this acid in the system. The lack of complete inter- pretation of the spectrum reflects the actual complexity of the system. The more important of the reactions occurring in.an aqueous solu- tion of bromine will now be considered. Bromine hydrolysis proceeds as follows: Bra t H20 == H+ 4 Br- 'r HOBr (l) In.the presence of light, the hypobromous acid undergoes the decompo- Sition, HOBr . hv = H+ 4- Br- 4- $02 (2) Besides the photochemical step, the formation of HOBr leads to ABSORBANCE 10 25 I L0 (150 0.25 (LIO- (105- 0.25 r' 0.0! 1 1 2400' 2800' 3200* 3600' 4000* 4400- 4dddr WAVELENGTH (K) Figure l. Visible and ultraviolet absorption spectrum of a dilute aqueous solution of bromine (Katzin (2)). 200L 100' MOLAR EXTINCTION COEFFICIENT l l 2500 3000 3500 WAVELENGTH (3) Figure 2. Ultraviolet absorption spectrum of hypobromous acid in aqueous solution (Anbar and Dostrovsky (3)). additional side reactions. It appears that its decomposition proceeds by two simultaneous, independent paths, SHOBr = H’ . Bro; + 2H20 + 2hr2 (3) and. LLHOBr =2 02 + 2H20 + 2Br2 ()4) In the dark, reaction (3) predominates. Hypobromous acid is also a very weak acid, its dissociation occurring as follows: HOBr = H‘ » Bro‘ (5) Another reaction which may be significant is the oxidation of hydro- bromic acid by ' ‘bromic acid, i.e., 6H, + sBr" : Bro; = 3Br2 4 3H20 (6) The literature concerning reaction (1), the extensively studied bromine hydrolysis, has been reviewed by Liebhafsky (h); a value of 5.8 x 10“9 for the hydrolysis constant seems to be the most acceptable one. Reaction (2) , the photochemical one, has been studied by Pagel and Carlson (5). Reactions (3) and ()4) were investigated by Pollak and Doktor (6). Shilov (7) .has reported a dissociation constant of K = 2.06 x 10"9 for reaction (5); this value is in good agreement with that of other investigators. Reaction (6) is discussed by Bette and MacKenzie (8) . Standard oxidation potentials for reactions ( 3), (u), and (6) are 0.07, 0-36, and O.h3 volts respectively. These values have been calculated from tabulated half-cell potentials (9). The above oxidation, potentials lead to values of 2.59 x 10, 6.62 x lo5 and 3.08 x 108 for the re"Spective equilibrium constants. The addition of bromide ion to the water-bromine system effects considerable simplification, since repression of the hydrolysis reaction virtually eliminates the complex side reactions of hypobromous acid. The system, water-bromine-bromide, is considered next. The water~Bromine—Bromide System The reaction of molecular bromine with bromide ions in aqueous solution yields the anionic donor-acceptor complex, Bra“. The formation reaction is as follows: Br2 + Br- e Brs- ‘With excess bromine, an additional complex Br5-, may be formed (10) according to the overall equation, 2Br2 + Br- = Br5- The equilibrium.constants for these reactions are: f .. K a 8 Br ‘) ' Br3_ 3 _ 1"L323L1Qf 3‘Br 2 8'Br Bra Br fBr 2 fBr' an; (Brat far; 2 _ n- _L‘ e"‘ ' 2 f - aBrg 8Br (Br?) (Br ) fBr2 fBr where the quantities in parentheses represent concentrations of the particular species, and the symbols a and f the activities and activity coefficients, respectively. The first of these reactions is the more important, with the extent of the second being relatiVely negligible under most conditions. In addition, in some studies, the bromine hydrolysis reaction, Br2 + H20 = H“ .. Br” + HOBr must also be considered, although it is appreciably repressed in sys- tems containing added bromide. Most of the earlier studies of the waterbbromine-bromide system were based on a partition or distribution method. By these means, the concentration of free bromine in a solution containing both free bromine and bromide ions is accurately determined after permitting the volatile bromine to reach a state of equilibrium, via the vapor phase, with a similar solution which is, however, free of bromide ions. Since the activity of bromine at equilibrium is the same in both solutions, the concentration of free bromine in the aqueous bromide solution can easily be determined if the activity coefficients are assumed to be the same in both solutions. In the earlier work, activity coefficients were generally neglected, although in several more recent studies they have been included. The most recent work on the tribromide equilibrium is that of Scaife and Tyrell (ll). Here again, a distribution method was applied. It was found advantageous in their work to use a solvent medium of constant acidity and of high and almost constant ionic strength. The slight acidity helped to suppress the hydrolysis reaction and the constant ionic strength to ensure that activity coefficients are independent of changes in concentration of the complex species. These [investigators studied the system at three temperatures; 5, 25, and 350 C. iMean.values of K3 for these temperatures were found to be 19.85, 15.53, and l5.28 kg mole-l, respectively (all concentrations were based on the molality scale). .-.A selection of the best accepted values is shown in the following table, which has been taken from the paper of Scaife and Tyrell. Salt ‘ Temp. (00) K3(liter mole‘l) m KBr 0 ' 19.6 (12) ICBr (I = 0.5)* 16.5 18.2 (13) NaBr (I = 0.5?" 16.5 17.11 (13) KBr (1 = 0.5?“ 21.5 17.b (13) NaBr (1 ... 0.5)““ 21.5 16.6 (13) LiBr (I = 0.5)* 21.5 16.2 (13) KBr 25.0 16.1 (it) KBr 32.5 15.11 » (15) 8 I = ionic strength The only previous spectroscopic studies of the tribromide ion in (aqueous solution are those carried out by Gilbert, gt_§l, (l6) and :Rob (17). A number of other trihalide ions, including the dibromo- Na2,O4 for nitrous oxide in sodium bromide, =-. t for bromine in sodium bromide, (10g ' )/‘1 = (K1)H20 and BI.2 (l/r)N3.Br By substitution of the quantities on the right into the expression, 1 (log t )/n Br, in Na,so, {—02 U V» N20 in N83,. v—vfi— 1 (“'03 “ V" N,0 in Na,so, we obtain, by cancellation of like terms, the reduced expression, “(1)1120 and Br, (l/r)NaBr The latter is identical to the expression desired, i.e., the ratio (log 1 )/n, for bromine in sodium bromide. Griffith, e3 31. (13), using the above expressions, calculated the value of ( logz)/}1 for bromine in sodium bromide. The data shown in the Preceding table were first supplemented with the corresponding value for bromine in aqueous sodium sulfate. By partition experiments, the par- til—tion coefficient for bromine between carbon tetrachloride and an a(Bleeous solution of sodium sulfate (ionic strength a 0.5) was found to be 31 -3 . Since the partition coefficient of bromine between water and Ca1‘an tetrachloride is 27.5, it follows that the activity coefficient 17 of bromine in the sodium sulfate solution is 31.3/27.5 = 1.131;. Hence, (log t )/n = (log 1.13h)/0.5 a 0.109. It is then inferred, from considerations discussed in the preceding paragraph, that for bromine in sodium bromide, (log I )/n = 0.109(0.807) = 0.088. Since the expression, (log i’)fiu, is a constant for a given system, the activity coefficient for bromine in any given solution of sodium bromide can then be easily calculated. For the system, water—bromine-sodium chloride, the ratio was determined in.an analogous manner; i.e., (log t )/n = 0.109(0.93h) = 0.102. For the mixed salt system studies, the activity of bromine was estimated as follows. From the preceding discussion, it was seen that for bromine in aqueous sodium bromide, the value of (log 3')Au = 0.088. Similarly, for bromine in aqueous sodium chloride, (log X’)Au = 0.102. For any mixture of the two salts, it would appear that the (log a )/,u ‘value should be closely estimated by the expression, (log 8 )Au = 0.088 + m‘ (0.01h) Elle term m' represents the ratio of the molality of sodium chloride to idle sum of the molalities of both salts in a given run. The value, 0.o1u, is the difference in the (log 7 )/u values for bromine in each of the aqueous salt solutions, i.e., 0.102 - 0.088 = 0.011;. Mirobromip Acid in Acnxeous Sodium Bromide The activity coefficient 0f hydrobromic acid in aqueous sodium bIVDDIide was calculated from data available in the literature. It was 18 necessary to make two plots of the literature data. For the first, values of the mean ionic activity coefficient of the acid in sodium bromide for a constant acid concentration of 0.01m.were plotted versus the concentration of the salt (25). From data of the mean ionic activity coefficient of hydrobromic acid in water (26), a second plot of the mean activity coefficient of the acid versus the concentration of the acid was prepared. By proper interpolation from these two plots, ‘with respect to the individual runs, a third graph was obtained for subsequent use in the calculations. This third graph was drawn for each run. The significance of this latter plot may perhaps be best understood by reference to a particular run. For run 1h5, the following data are pertinent: molality of solution = 0.195 1+ of HBr in H,0 = 0.783 — (HBr conc. = 0.195m) 8+ of I-IBrinNaBr = 0.769 ‘ (0.01m HBr, 0.185m NaBr) In the plot, the log It for hydrobromic acid was plotted versus the 1n01ality of the acid at constant total molality.. Extrapolation to zero mOlality yields a value of X i for the acid which should be quite accurate. It will be recalled that the concentration of the acid, fOI'med from the hydrolysis of bromine in aqueous sodium bromide is reBlatively'very'low, even in dilute solutions of the salt. This method ibis been widely used by other investigators. For example, Hawkins (2?) has shown that for hydrochloric acid-uniunivalent halide mixtures, this f_——————_—_————_wfiT-uv* 1? linearity persists at constant total molalities as high as 6 m. Other similar systems have also been studied in this manner. Eydrobrondc Acid_in Aqueous Sodium Chloride The activity coefficient of hydrobromic acid in sodium chloride was calculated from the expression, XiHBr in Nam : ( xiii—CE; in NaCl) ( Ya)~1\IaBr in H20) ViiIaCl in H20 For each calculation, the quantities on the right side of the equation were associated with‘salt concentrations equal to those for the run under consideration. From the following discussion, it would appear that the above expression should approximate the desired quantity fairly well. Using the definition of mean ionic activity coefficient, the equation can be expressed as follows: ”if ( 2’01“).5 (xNa+)% ”hr“? xiHBr in NaCl = C1ium chloride solution alone is considered. In both instances, the erfirironment of the chloride ion is essentially the same. If one now cc>nsiders the environment of the sodium ion for the systems, sodium bromide in water and sodium chloride in water, it is clear that a 2O bromide and chloride environment respectively are involved. Assuming that the effect on the sodium ion by the two different species is not appreciably different, we can cancel these terms and arrive at the following: _ . i- _ i XiHBr in NaCl " (111*) (xBr) The expression on the right side is, by definition, the mean ionic activity coefficient of the hydrobromic acid. For a given run, the value of xiH'Gl in N aCl was calculated in the same manner as was the coefficient for hydrobromic acid in sodium bromide (preceding section). The values for the mean ionic activity coefficients of the sodium bromide and sodium chloride in water were taken from the literature (28). Conversion of ActivitLCoefficients With all subsequent calculations, it will be noted that concen- trations are expressed in terms of molarities. Since the literature data used for activity coefficient calculations are based on the molality scale (as discussed in the preceding pages), it was necessary to convert all activity coefficients. The conversion was made by use of the formula (29), fi = (l + 0.001nMB)(do/d) Xi" where the quantities are defined as follows: f = mean molar ionic activity coefficient + m ---= molality of solution WB = molecular weight of the solute do = density of the solvent d = density of the solution +o< ll mean molal ionic activity coefficient 21 The analogous formula, f = (l r 0.001mflB)(do/d)x','was employed for conversions involving the non-electrolyte, bromine. For the case of a mixed electrolyte, the extended formula, fi = (1 + 0.001 XmWBXdO/d)? i, was used for converting from the molal to the molar scale (29). 22 THE PRESENT .TJWESTIGATION-«RESULTS AND DISCUSSION The various systems studied in this investigation are the follow- ingt water-bromine, water-bromine-bromide, water-bromine—chloride, and water-bromine-chloride-bromide. Each of these will now be considered in some detail. The Water~Bron1ine System The water-bromine system was initially studied in an attempt to elucidate the origin of the short wavelength band in the ultraviolet region of the spectrum. our studies clearly demonstrate that this band can be ascribed to the presence of the tribromide ion, which results simply from the reaction of molecular bromine with the bromide ions formed in the hydrolysis step, i.e., H20 + Br2 e HOBr + H" + Br- Bra 4- Br‘ =- Br,“ In an attempt to verify this interpretation quantitatively, a study was planned in which both reactions, the hydrolysis and the tribromide formation equilibria, would be accounted for in the mathematical treat- ment of the data. Unfortunately, however, quantitative treatment was not possible because of the spectral changes which occurred with time. When no precautions were taken to exclude light, the intensity of. the absorption minmm in the 2550 to . 26503 region always showed a definite increase with an accompanying shift to longer wavelengths. A typical increase would amount to approximately one-tenth of an 23 absorbance units the shift would be of the order of five angstroms. A period of several hours was usually necessary for these changes to occur. This instability made quantitative treatment impossible. Figure 3, a typical spectrum for the system, illustrates the observed trends. Two 200~watt tungsten lamps, placed several feet from the absorption cell during the exposure periods, were used as light sources. The changes mentioned above could be accounted for by the subse- quent behaviour of the hypobromous acid which is formed in the hydrolysis. Hypobromous acid readily undergoes decomposition in one of two ways. In the presence of light, the decomposition proceeds largely in.accord- ance with the following reaction [reaction (2), early discussion]: 2HOBr e 2H+ + 2Br--+ 02 The bromide ion formed enters into the tribromide equilibrium, thus increasing the tribromide ion concentration, Therefore, because of the intense absorption of light by the tribromide ion, even a very small change in its concentration would result in an appreciable change in the absorption spectrum. The shift may be explained by the depletion of the hypobromous acid together with the accompanying increase in the tribromide ion. .Anbar and Dostrovsky (3) and other independent workers have studied the ultraviolet absorption of hypobromous acid in aqueous 8Clution (Figure 2). The acid has two absorption maxima, one at 2600 auni the other at 3200K. The short and long bands have molar extinction VEChues of approximately 95 and 35 liter mole".1 cm"1 respectively. Since 'flhe tribromide absorption maximum occurs at 26608, the observed wave— length shift could thus be explained. 2t sofipoppsoosoo o5. cofiwmadmmhm poems 333855 cognac mg 4. geomam .308 msOH N m bampgunpagd ma msfisoan mo 38de mason 5mm poopoooh mg m époomm meowpfiuom 93 .Ho .ofiusonn .Ho noampsaom msoosvd 3.ch d Mo soapmnomnd poaOfiboApHs one no ”Em: mo 98me .m madman a thozu4u><3 c. .7 c. c. c. c. 3 Z Z Z Z Z Z Z 0 .7 o o a n m a me n n n n h w .e. O O O O O c. O 9 8 9 8 8 6 c. 9 a 4 A . Jfi 4 i q _ 1 i e 1 n _.0 1N6 1nd 4 1¢.O Lmd m 0.0 Ind and lmd go; BONVBUOSBV 25 When precautions were taken to exclude light, it was observed that again a true-dependent shift to longer wavelengths occurred in the 2550 to 26653 region. Under these conditions, however, the absorption maximum decreased in intensity. An accompanying slight increase in the bromine band intensity at 39003 was also noted. The nagnitudes of the changes for the short wavelength band were approximately the same as for the case discussed in the preceding paragraph. Typical spectra for this system are shown in Figure )4. Since even the "dark" system displayed instability, no reproducible quantitative results were obtained, and any explanation of the behaviour must be Speculative. The following combination of effects seems con- sistent with the earlier discussion and is capable of explaining the observed trends. If the tribromide ion is indeed the main contributor to the short wavelength hand, then since this band decreases in intensity, ‘while the bromine band intensity at 39008 undergoes an increase, it follows that the bromide ion concentration must decrease. The shifting of the short wavelength band to longer wavelengths also indicates a lowering of the hypobromous acid concentration, so that the hydrogen ion concentration mst increase to maintain the hydrolysis equilibrium. From a consideration of the important reactions, one might arrive at the following explanation for the observed spectral changes. Hypobromous acid would, of course, be produced in the hydrolysis step. Its decompo- sition in the dark could then proceed as shown by reaction (3), i.e., SHoBr = H“ + Bro,“ + zsr2 + 2H20 (3) 26 .Hsaos myoH N m hampcsfixohamc we osflsonn mo soapsnpcoosoo one .haosflpooamon hopcfl mason cosmm new meow nonpoooa ones 0 new m depoomm muonpsHom one mo soapsndnoaa scams hfiopdflpmssfi posfimpno mm: <.Esepomam .smpmmw one Eopm popsaoxo me pnmflfi_sons osfisonn mo soapsaom mucosvc opsaflp m an pmpfinflnxw momsmno Haywoomm endofibwppfib .: caswfim a. 5053i; '1 8’9? -49|93 9993 ooos - oow - 0669 - 02.92 - oer; — 992: 402.09 -4 9962 -( a u 92 e euz i 9392 J one 6—“ niwru BONVBHOSBV ._ v . {I . _. O . .. [as .1 v I. Ii . Vii... . '5" 'V' . H .r takes!!!» . .. . .. 1".rr 27 Subsequent oxidation of hydrobromic acid would then occur according to reaction (6) . 6H+ + SBr- 4- Bro; == BBr2 + 3H20 (6) If all of the bromic acid produced by (3) were not consumed in (6), the conditions stipulated above would be satisfied . The hydrogen ion concentration would increase, that of the bromide decrease, that of the hypobromous acid decrease, and that of the free bromine increase. Reactions (2) and (1;) need not be considered since their occurrence requires the presence of light. The ionization of hypobromous acid can alas be overlooked; the dissociation constant is so smll that the concentrations of the ionization products would be insignificant. From the known chemistry of bromine and its acids, this explanation seems to be a reasonable one. One of our runs was initially studied in the "light" and then followed in the "dark." The interesting results are shown in Figure 5. It is seen that the changes are precisely what one might predict from the "light" and "dark" reactions discussed separately above . The short Wavelength band first increased in intensity with an accompanying shift to longer wavelengths. When light was later excluded from the system, this same band decreased in intensity with the shift toward longer Wavelengths again occurring. In addition, the bromine band was observed to Show a slight increase during the "dark" stage of the study. 28 IF 8 15233; .asaos muoa x H zHopwstopmqm ma ocfisopn mo sowpmnpsoosoo one .oasmoaxo enwaa mo mason moanp mo poommm one msonm o scapoomm .ampma mason mead popaooou n can soapsaom map mo soapsnsmoaa hoped afloeceoosfi coast he. a Essen 2.3.... on... :35 oofioosoo one s es. 4 eneooom . is? one. 5. son... as case? one. 3.“ 8...ch the he seems dos. seasons mo soapsaom mucosuc unease s an popfinfinxm mowsono asnpooam poHOstngD .m wasMHm c. .v c. c. c. c. c. .6 .6 .6 .6 .6 .6 .6 .6 n. .V .b a. .v .6 nu .b n. I. q. .3 b. .7 .2 Au nu .o .L .a .a .L .z I. I. .6 .9 .L .I .3 n. n. n6 nu nu a: lbw a. a. m. aw. n. .u .3 a. _IIIIIAl A q . .4. a _ a e . . a _. .\MA 4 L ..o ///IIIIIII‘ InNAu 0.0 O . < mag Iaxo Ififio ..mAu Jame lmwo J.G.O n... BONVBHOSBV 5.... VFW . 1,. H. . . .... #01! [Viki qy¢§ v.3]..a.‘. Q tw’ofl... : Q. . Oh . . .. . F). l‘ o . O muffle Algio‘li‘kb‘ét.‘o v. s1“)... 1 .. ... 29 The Water--Bromine—Sodium Bromide System The instability in the above system was eliminated in the next phase of our study by the addition of sodium bromide, which suppresses the hydrolysis. Runs 138—110 and lh2-152 were concerned with this portion of our work. The spectra obtained verified the anticipated stability; absorption readings for any one run at a given wavelength were virtually constant with time in all cases. The observed absorption maxima occurred at 26602 in these runs, indicating no contribution from hypobromous acid. A typical Spectrum is shown in Figure 6. Deterndnation of 9266c - gxdroysis neglected (Method 1) Various methods of treating the data were employed. The expressions used in the first method will now be derived- Consider the following equilibrium: ‘ Br2 + Br" . Br,“ Let a be the initial concentration of bromine, b the initial concen- 131‘ation of bromide, y the equilibrium concentration of the tribromide ion, fBr the activity coefficient of bromine, and f 4 and f - the 2 - Br3 Br activity coefficialts of the tribromide and bromide ions respectively. 1:118:11, , '1‘ - y ( Br3 ) K3 “ (£3,371. 4m: _~,,) ' (r33 Neglecting the y term relative to b, and assuming the ratio, (fBr3-)/(£Br-), to be essentially unity, we get 30 000? Aqda edmv .AdHos mica a mmm.m ma moHEOhn.s:Hnom can mo noflpduenmucoo mnB .nmaos rioa x m©.m ma seasons one mo soapdnpumosoo one .oomm pd Empmhm Ameumhmuomm one no sshpoomm dowpmnomnd poacabwnpab .0 onewfim Anvzhozm4u><3 C. C. c. c. c. .6 .6 7g 8 9 .7 z 0 8 9 .7 O 0 0 0 O O O O O O O 0 O O O 0 0033 VAV ed . adv O; BONVBUOSGV 31 Y Y (£31.37; : flb (ng271.11. in) K3 = Solving for y gives Kaab(f ” ) -2 y . 1 + K3b<£Bré7_ The justification for neglecting the y term relative to b can be seen from a consideration of the relative concentration.va1ues of a and b (Table 1). In most of the runs, b is more than a hundredfold greater than a. Since y'is necessarily less than a, the difference between b and y'is even greater than.between.b and a. Using a value of K3 = 17 liter molefl, the tribromide ion concentration was calculated for each run from the data. The value, K3 8 17 liter mole'l, is representative of those in the literature (11). The molar extinction coefficient, 62660, was then calculated from the.expression, e = A/yl, where.A is the observed absorbance and l is the path length of the quartz cell used in the studies. Table 1 shows the results of these calculations. Detenrnnation of eoéw- Hydrolysis Considered (Method 2) A second and somewhat lengthier method of treating the data in- ciluded the effect of hydrolysis. Let a, b, and y represent the same (plantities as before. The two equilibria, H20 t Br2 = HOBr + H+ + Br- Br2 + Br- = Bra- 32 . 3H, assesses? 5. trees to mefipflsseessm t mum . A m a t * fifism I 6”.me mm.m 4w.m mm.H mmo.m ~.wH: ow.m mma om.m mm.m mm.a mwo.a 5.:Hm :m.m HmH mo.m :m.m m4.H mam.a :.©oa HH.: omH om.m :m.m Hm.H NNH.H ,mm.mo Ho.: mqa mm.m $.m .84 034 meg. £8 9.? om.m mo.m mm.a HHN.H . om.mm sm.m 52H em.m es.m NN.H mso.H Hm.oe . mm.m . see fizm mm.m moé 30.4. 3.? 38 mi? H .m 0m.m mm~.o mHO.H mam.m m©.m 44H m m om.H mmo.o HHo.H qwm.z :m.: mad m m mo.m Hmo.o moo.H 0mm.: Hb.4 «za 4 m Hm.m w:~.o woo.H Hmo.m :w.o 04H oo.m mmw.o mom.o moo.H Hem.a mm.m. mmfl om.m ma:.o m:a.o mootfi . mmba.0 :m.m . mmfi rig 5 $2 5 Ease sew: tattoo $2 3 $3 5 $0.52 A.so\mmmW\pmpHHv Amopwm\mmflosv .ss Hv th>H904_ Anmpfia\moaoev Anmpfifl\moaosv gem A mv.mmoo A v.0soo 000m as «am Anv.onou Adv.uqou .ss seas: ..ssm date 3 efissomss smsz Essa. ssm EPEH l'... I ‘7'"-. I'.’ I.--’ I’i.-i ’ "- 1", ..-- . 1".--”i’- ."".“- iAmsoepsHseaso 3.. Be: 3 .. me to asset 88 afiomsmz mHmSomBm on meezommefi mos . e .EEBES _zoEoEE $.52 as so 5:ng ngmfi. 33 are now considered. If x is the concentration of the bromide ion pro- duced from the hydrolysis, the equilibrium expressions become K __ xegb + x -1) (fH"’)(fHOBr)(fBr') h a - x - v ' iv, f ( y) C 31.27 and (again.assuming fBr3--= fBr-), Y' K': = —' v .- ’ , _ ‘ (fBrlea - .. was . 1 fl Then assuming further that f f , since these are both neutral HOBr ” Br2 species, the ratio of the equilibrium constants gives __., x2113 + x LYN" . ,_ ,' Kh/‘Ks y- W . (er)(fBI' )\fBr2) Xffib_:_x fQY)2 2, y (ffHBr) karz) sari /K /K )%= 3((p F x;yL (f )"f Y; \ h 3 y; mr \ Bra. Itearrangement of terms gives (Kb/x3)?t x2+x(b-y)= ( ' a; - (1) ~fsmrxfsrz)” .Aggain.assuming f , = f , the product of the equilibrium constants Brz HOBr ginres K K n :zy (fH‘)(fBr‘) a - ;~ 2 h (a x 3'7 Bra v 2 “2y (fmr) (a - i - fl" (731.27 3‘4 V % Ky Writer) “9%) = (a .... - y) {77; Br2° Solving for x, _ X :3 (53:3)1fga'v’ y)... ‘ <2) iii-3%. 4. (KhK3)i. (rm) Equations (1) and (2) were solved for x by choosing arbitrary values for y. When x is plotted versus y for each equation, the point of intersection of the two curves obtained gives the values of x and y consistent with the two equilibria for the system. This procedure was followed for each run. The results of these calculations are shown in Table II. Because of the very dilute salt solutions used for runs 138 and 139, the usual method for determining the mean ionic activity coefficient 01‘ the hydrobromic acid was not possible; i.e. , the literature data did not attend into this very dilute range. Calculations for the two runs Were made with the assumption that the activity coefficient was equal to unity. Runs above 1146 were not analyzed according to this method. With these later runs, hydrolysis effects were insignificant, and the me‘thod ishreduced to the one used in the preceding section. The very Small discrepancy between the results obtained by the first and second m'i‘thods, as seen from a comparison of Tables I and II, indicates that the hydrolysis equilibrium can be neglected in all runs with added bromide . .liflr a. .- lllflw .s. S 3 . ASN.3 maggoflua dun mmonp mo mbfipwfinmmmfimmm .f, I m same a J sea emu n .H * e4.m n use: . swa H8 . mien mm; mono mmoé we: mm.m oi m:.m Ho.m Ho.m mo.H H-.o mzo.a o:.mfi, om.m mqa o:.m mm.m mH.m mm>.o zom.o mHo.H mam.m mo.m 34H em.m mm.o mm.a mmo.o mmm.o HHo.H zem.z :m.4 med mm.m 8;. SN $0.0 48.0 mnoé same, 2.: «fl mm.m m.oa mm.m mam.o mmm.o woo.a Hmo.m p :~.w 04H He.m :.HH H:w.0 mom.o H). moo.H Ham.H ww.m and mo.m H.NH 5mm.o mza.o H). Noo.H mmom.o :m.m mmH $13. 5 $3” 5 $3 cc Essa new: *3 1mm 33m $3 xv $2 cc 62 Tamwmmoiinmflud AamfiQmoHosv AnmpaHBmHosv .5: S . ..Hmoo Aumfluakmaosv Anmfludmoaosv HE A 33300 Axvmfimhog T3650 M88 pd hinged 3.» .980 63.250 . Ba. 30: B pests .....sm deem 3 sessssomfl. , ..Hmez Beam «em Essa A aahmccfipcasoaso ca comb mica N m.m n x and we a mm_mo mmsas>v QmmmmHmzco mHmMAdmnwm 28 afizomfima mom .SQNs ..EEBES zoflofissr mics Be so 2825ng HH mqmda 36 geterndnation_of £3, and e2 6 Method _ A third method of calculation was tried in which both K 3 and e2 660 were calculated. The fact that K 3 was evaluated by this method probably makes it the most desirable of the three. In this case the hydrolysis reaction was neglected, as was shown to be justified by the previous calculations, because of the direct suppression of the hydrolysis by the added bromide. Developnent of the expression used, in the calculations proceeds as follows: Considering only the tribromide equilibrium, Br2 + Br“ = Br; and 3’ K3 "” TfBrawa. The - r) where a, b, and y are the concentrations of bromine, bromide, and tri- bromide, respectively, as in the previous methods. Since 13 is much greater than y in the runs where bromide has been added, the equation becomes 3" ‘ K3 2 (£51.93 - 3’7? By rearrangement and substitution of the value A/el for y, the concen- tration of Bra’, abl(fBr2)/A = l/K3e + b(fBr2)/e Tnererore, a plot of abl(fBr2)/A versus has],r ) should yield a straight 2 line with intercept equal to l/Kae and slope equal to l/e. Figure 7 Shows this plot. A least squares analysis of the data resulted in 37 THE canoe 5 nope: copped» 3:35 .sow mofisoanfinp no.“ .8omm .pcmfiofimmmoo soapofifia 8305 one. 93 «mm .pgmdoo Sadness. soggom one .3 cogéopoc .3.“ com: aEmnogfluoa ndocfln one mo pmme é madmwa X m; 0.. 0.0 . a q q — q q d u — d J 1 J .....H . A mass ax N a) Sevens n a (,0: m values of 17.3 liter mole“1 and 3.146 x 104 for K3 and 82660’ respectively. Table III includes the values used for the least squares treatment. The excellent agreement between these spectrophotometric results and the results of earlier distribution studies verifies the identity of the hand at 26608 as due to the tribromide ion and affords, for this A ion, a value for the molar extinction coefficient which can be utilized :A as an analytical tool in any further studies 3“" The WaterwBromine—Sodimn Chloride System «as-"- an . Retgnuination _of K4i~§awr_1d_e2380_- Eydrogysis Neglected_0‘lethgd ll The first series of runs in regard to the dibromochloride ion were studied in the system, water-bromine—sodium chloride. Although hydrolysis in this system is not directly suppressed, as was the case With the system, water-bromine-bromide, it should nevertheless be diminished appreciably because of the reduction in bromine concentra- tion through the dibromochloride equilibrium. The first set of calcu-r lations were therefore made with the assumption that the hydrolysis effects could be neglected. It was also necessary to assume (analogous to the assumption, fBr - - fBr'” mde for the tribromide ion study) 3 that fB '- - f -. Preliminary studies had indicated that the rZCl Cl h ‘- — ——.—¢ *Using the triiodide absorption maximum at 35202, Custer and Natelson (31) have made spectrophotometric determinations of microquantities of iodine. A 5% KI solvent was utilized for complexing of the iodine. As little as 0.2 micro grams of iodine per milliliter could be determined. 39 ImHoE noon: 95“ n AccommxohmonopfiH Mgmds 14b. our discrepancy, and it was decided to attempt a hydrolysis correction in our calculations. The derivation of. the expressions used in the cachulations is as follows. Consider first the trihalide equilibrium, Br2 ; c1.“ = Brgol‘ The equilibrium constant is __ 2 K4 - (fBrzja - x - 276 (1) Here, a is the initial bromine concentration, 2 is the equilibrium trihalide concentration, x is the extent of the hydrolysis reaction (as in Method 2 for the tribromide system), c is the initial concen- tration of chloride, and fBr is the activity coefficient of bromine in 2 the salt solution. The assmmption has again been made that the activity coefficients of the trihalide and halide ions are essentially equal. The accompanying hydrolysis reaction is, as before, H20 1!» Bra a HOBr + Br- + H? The equilibrium constant for this reaction is K = 7* X3 T . EH+)(fBr*v)fi(£HOBr)— h (a ‘ - E) fiBrz) x3 2 TEE-2 - a (fax...) <2) It has again been assumed that the activity of the bromine is equal to that of the hypobromous acid; thus cancellation of these quantities gives equation (2) . Substituting the expression, A/el, for z and mlving (l) for 3: gives LS K4c(fBr2) + l 40 Br 2 x== a. -A/el (3) Solving (2) for the expression, (a - x - z) , with subsequent substitu- tion of this quantity into (1) gives, upon solution for z, n z =- Eff“ (fiHBr)2(P r2) 3: A/el (14) I {_1‘ - substimtion of (3) into (LL) and rearranging gives 11‘ AKh12o2(rB )2 i- - La r2 . o , -— a - _fi—w . (K4e)'5= acl(fB )(K4e) ~—Ac(fB )(K4) ‘ A (5) (f+tmr)2 r2 . r2 For Simplification, let (5') be represented. as follows: Here, 1:1 as (K4e)'§, X2 = K4e, and X3 = K4. The significance of the (Identities, t, u, and v, is apparent from a comparison of equations (5) and ('6). Division of (6) by t gives A ' x1 = (u/t)X2 - (v/t)X3 - A/t or X1 = u‘Xz - v‘X3 - w’l . (7) Where up . u/t, v' = v/t, and w? . A/t. For any two individual runs, we would have: Run #1 X1! ul'x2 - vl'X3 ~‘ w’l? Run #2 X1 = ugxg - vaixg- w2' ng X1 from the last two equations and rearranging gives to (“1' " V2,) . _(vvz' "" V13) X3 - X2 (8) (112‘ " 111') (L12’ " 111') Since X2 and X3 are both constants, then a plot of the quantity on the left side of equation (_8) versus the coefficient of X3 should give a straight line with slope and intercept equal to X3 and ~X2 respectively. Ten of the runs applicable to this treatment were considered in the calculations. Table V shows the values of u', v' , and w' as calculated from the data for each of the ten runs. Table VI gives the calculated values of the two quantities plotted for each of the pair of runs considered. Figure 10 shows the resulting plot. Examination of the graph clearly reveals that the hydrolysis corrections alone do not alter the results markedly. An approximate calculation, directly from the graph, leads to approximate values of 0.75 liter and mole”1 and 3.3 x 104 for K, and 82380’ respectively. A least squares analysis did not seem to be warranted because of the scatter of the results of these calculations. It appears that some effects considerably more pronounced than hydrolysis are necessary to explain the discrepancy in the results. A Possible reaction sequence is the following: Bra Bra Bra ., Cl- = BrZCl” == BrCl + Br- 4. 01" - u .. ClBrCl h? TABLEV . CALCUIATED VALUES USED FOR DETERMINATION OF THE FORMATION EQHlLIBBIHM CONSTANT, K4, AND THE MDLAR EXTINCTION COEFFICIENT, e2380, FOR DIBROMOCHIORIDE ION HYDROLYSIS CONSIDERED % (Value of Kh = 5.8 x 10"9 used in Calculations)“ . ' ‘_~_* A— # _-_—_— fiv _v— ——f—w-—v—7v—V fifi fi '7 V—V _vwr— ——— u! vI -3 w’ _3 Run No. (x 10) (x 10 ) (x 10 ) 156 1.1LN 1.630 1.332 157 1.015 0.8618 1.513 158 0.9657 0.6092 1.621 159 0.9170 0.3102 1.678 160 1.093 1.253 1.822 161 ' 1.756 1.598 2.683 165 2.098 3.178 2.229 167 1.751 1.698 2.81.7 168 1 . 887 2 . 808 2 .hh2 169 2.139 3.506 2.313 A (f A v! :3 Ci Biggfi w! :3 _l .—‘—-—_.~ AfglzcnfBii): g— [Mh12c2(rBrPf]-§r (fiHBr)2 (fiHBr)2 _— _ — w a v v fi—vfi v—‘ v—f “— w — fl * iterature Value (h) 88 MBLE VI CALCULATED VALUES USED FOR . DETEEH‘GINATION OF THE FORMATION EQUILIBRIUM CCNSTANT, K4, AND THE MOLAR EXTINCTION COFTFICIEINT, 82380’ FOR DIBRDMOCI-LORIDE ION HYDRDLYSIS CON S IPERED (Wl' '9 We') (Vz' " Y1?) H 112“ - ul' Tug - u T) Runs No. (x '10‘4) (x 10"“) 156 and 161 -2.182 -0.058L7 156 and 165 ~O.9LLM4 1.630 .f 156 and 167 ~2.h98 0.1118 _; 156 and 168 -1-h9h 1.585 t 156 and.169 -0.9860 1.885 157 and 161 «1.528 0.9881 157 and 165 -0.6632 2.187 157 and 167 -1.813 1.136 157 and 168 -1.019 2.232 157 and 169 -0.711h 2.352 158 and 161 .1.294 1.286 158 and 165 - -0.5391 2.277 158 and 167 -1.563 1.387 158 and 168 . -0.8916 2.387 158 and 169 . ~0.5901 2.h67 159 and 161 ~1.150 1.530 159.and 165 -0.h678 . 2.h37 159 and 167 ~ -1.802 1.668 159 and 168 -0.7873 2.575 159 and 169 -0.519h 2.615 160 and 161 -1.880 0.5185 160 and 165 ~0.8057 1.923 160 and 167 -2.166 0.6763 160 and 168 ~ -1.28h 1.958 160 and 169 - I.-0.8513 ‘ g 2.153 —~—-f v—V—vfi —v w—v fi w _r—‘ 1:9 TE, @389 5. pmpmfimwmnm poppoaa moan: .pmhmpfimdoo manhoaphm .noH mpfihodzoosounflp .Hom . mm Jamwoflwmoo soapofifiB H39: was. was 3a .pgpmdoo EBuBHHHHGm nofipgom 0%. mo coflpgfimpmp pom pom: mflnmaowpdama .38qu one. .Ho paws .OH madman 1 q u — d u u d — I u q 1 fl - moNl _ ..N f: .3 ..>.. .3 x x O.Nu X X X \I \I X X 7m. IIM! 0.7. . . x x m m. x u. X X 0.... X X X X X X X X X x ufid: X 50 The complete mathematical treatment of this sequence of reactions cannot be carried out with the available data. However, since these additional effects, as well as hydrolysis, would be retarded by added bromide, it was decided to modify the spectrophotometric approach to the dibromochloride equilibrium by study of the system, water-bromine- sodium chloride-sodium bromide. 51 The Water-Bromine—Sodium Chloride-Sodium Bromide System Intraduction The study of this system was undertaken in an attempt to minimize the possible side equilibria which complicated the interpretation of the water-bromine-chloride study. a The mathemtical treatment of this system must consider the r‘ sinultaneous equilib ria, Bra ~. 01" = Br201‘ It“. .111 3:. u u—— -— -ln 1'. .. -.- Br2 + Br' = Bra- The interpretation is somewhat complicated by the fact that the ultra- violet absorption bands of the tw0 trihalide ions overlap slightly so that the tribromide ion concentration cannot be determined from its absorbence at 026602 alone. In addition, the tribromide extinction coefficient must be known at whatever other wavelengths are used for estimating the concentration of the dibromochloride ion. It was therefore necessary to determine values of the tribromide ion extinction coefficient at wavelengths other than its absorption Wm, 2660.8. The two wavelengths chosen were 2500 and 23803. The latter was chosen because the apparent maximum of the dibromochloride ion appeared to be located at this wavelength. The 25008 value was decided upon because it represents a wavelength intermediate between the naxima for the two separate ions. 52 Determination of 32380 and 82500 for Tribromide Ion The new extinction values required for the tribromide ion were readily obtained from scans of the tribromide band from the earlier study on that ion. The relationships used were simply 82500 g 32660(A2500/A2660) and 82380 "‘ €2660(A2380/A266O) Not nl of the runs made in the tribromide study could be used for the present calculations. In some cases, due to the fact that these present evaluations were not anticipated at the time, insufficient data Were available-that is, absorbence readings for the required wave- lengths were not recorded in all. instances. However, a sufficient number of runs were studied over the necessary wavelength span to permit the desired calculations. For example, the evaluation of e2500 was made from the data obtained in seven of the fourteen runs studied. Similarly, five runs were involved for the evaluation of 32380 . The results of these calculations are shown in Table VII. It is apparent that the agreement of the extinction values is rather good, a fact which probably strengthens any interpretation of these first studies. Subsequent calculations for the mixed salt system, water-bromine- sodium bromide-sodium chloride, yield results which seem to add further Validity to this belief. 53 TABLE VII DATA USED EOE DETERMINATION OF THE MDLAR EXTINCTION COEFFICIENTS, ' e2500 AND 82380, FOR TEIBROMIDE ION *— W vw—urv— — _u‘ h. _— #7 W WW —. _ w w—v ._—_ V _— ___'—— ‘— —.7 v—V—v fl __.. ~ ‘ Molar Ext. Absorbance (A) Absorbance (A) Coef. (e Run. at 25008 .at 26603 (liter/molg9gm.) Number (1 mm. light path) (1 mm. light path) (3: 10*) 138 0.071 0.1h5 1.69 139 0.150 0.265 1.63 1RD 0.370 0.7h8 1.71 lid; 0.365 0.752 1.68 1115 0.5014 1.03 1.69 1h7 0.619 1.28 1.67 11.8 0.580 1.21 1.66 Mean = 1.68 . Molar Ext. Absorbance (A) Absorbance (A) Coef. (e at 23808 at 26608 (liter/ /moi398n.) Number (1 mm. light path) (1 mm. light path) (x 10‘3) tun; .1._1_ V'Ydiidi fifififififi «ofiifi ,1, —' u. 65 .1. 1175 0 .135 l .03 h . 5h 1L6 0.170 1.2? b.82 11;? 0.175 1.28 b.73 1L8 0.170 1.21 h.86 Mean =-- 14.72 Sh . | , .‘ ~ Deteri'rgnatipn of the Ratios, 3238C5826SC and 82500122558 for D_bromo chloride Ion The mixed salt system requires a. different mthemtical treatment from those used in the previous calculations. The development of one of the methods will now be shown. Consider two wavelengths at which the molar extinction coefficients fl ,4 of the tribromide ion are Imam. Thai, Ann = eflm + ezlm <1) and 5:1 A'TZ ‘3 ye(.Y) '*‘ ez2(2) (2) The absorbsnces per unit length of cell at the two wavelengths, A 1 and A2, are represented by mm and “1‘? respectively. (In the previous sections, the absorbance has been represented by A, i.e., the absorbence associated with the actual light path (1 mm.) of the cell used in the studies. However, in several of the subsequent derivations, it will be convenient to use the absorbsnce per unit length of cell; therefore, both notations, A and A', will be utilized). II'he concen- tration of the tribromide ion is denoted by y and that of the dibromo- chloride ion by z, the designation being the same as in the preceding Sevisions. Molar extinction coefficients are again represented by e; 33.1 would thus be that of the tribromide ion at the first wavelength, at? . The unknown extinction coefficients of the dibromochloride ion are e z 1 and e72. By simultaneous solution of (1) and (2), y and 2 can be expressed in terms of absorbances and extinction coefficients, i.e., 55 z = e ZCA'Tl) ' e 1(A'T2) (3) ey2 ezl - eyl 822 and I .. t e22(A T1) 821” T2) 3' ” $09.37*: 5462;) “0 The two equilibria considered are Brz + Br- a Br3- (5) and. Br2 + 01‘ = Brzcl‘ (6) Using the definitions indicated for y and z , the equilibrium constants of (5) and (6) can be expressed as follows: K3 = m ‘7) r2 WW “8) r2 Where a is the initial concentration of bromine, and b and c are the bromide and chloride concmtrations respectively. Because of the- I‘elaiively high values of b and c relative to a, the exact expressions, (my) and (c-z), representing the equilibrium concentrations of the two h‘3:L"Lde ions, have been simplified to b and c. As in the preceding Stnldties, the activity coefficient of bromine is designated by fBrg’ and a"31-‘dh'ity coefficients of the ions can be considered to cancel from the e2Knivessions . 56 From (7) and (8), 'K Cy f = g; (9) Substitution of the expressions for y and 2, as given in (3) and (14), into (9) gives .15.: a 5321mm) "Lezleznc K fl a ~ —— ‘ (10) 4 Fawn) “eyzq'nm’ Rearrangement then results in the following: [e (A' )- e (A' )JKab 1 T2 ~ Tl . , . JIL‘“"e'TAT§T"‘””' g K4822) " K4<3 Gopm comm OOwN OO¢N CONN T H t q l— u — q . ll _ m IJIV - . I N . mm .. m w «IJIV L n «IV D Lualomaaoo' NOIlomlxa anon 79 dibromochloride ion band has its absorption maxilrmm at approximately 21608.. Since the absorbence at this wavelength can be closely approxi— mated from spectrum C, it was possible to calculate the extinction coefficient for the band peak. The computation is as follows: > - a — 0'370' a 4 e2.21.30 “ “ta-211306“l " '1‘39‘3: lo'r’ICO'l' 7 2'19 x 10 It may be concluded that the results obtained strongly support the assignment of the indicated absorption band to the dibromochloride ion- 8O EX PEIRD’ITNTAL Source and Purification of Reagents Water Purification of the water was effected by two distillations. The first of these was distillation of the original tap distilled water from a dilute alkaline permanganate solution through a three foot column packed with glass beads; the column contained approximately six theoretical plates. A middle fraction from this first distillation E 5?, then underwent a second distillation through a column similar to the first one. In the latter case, however, the distillation was carried out in a nitrogen atmosphere. Again, only a middle fraction was c34>:Lll_ected. Purity of the product was determined conductimetrically. Coniuctance values were of the order of 0.5 x 10-6 reciprocal ohms in Virtually every case . abomine Bromine purification also utilized a distillation. The bromine was first refluxed over phosphorous pentoxide and potassium bromide for a Period of one hour. This was followed by distillation with only a mid cfle fraction retained. In a typical plrification, approximately one hundred milliliters would be used initially; about fifty milliliters "D1130 be retained after the distillation. Mallinckrodt (N. F. V) bromine was the starting material. The phosphorous pentoxide and Potassium bromide were both 'Baker Analyzed? Reagent products. 81 The purified bromine was subsequently stored in vacuo; the container was kept immersed in a dry ice~¢hlorofom-carbon tetrachloride cooling bath. This purification is a slight modification of that used by Bauer and Daniels (31) , who used barium oxide rather than phosphorous pentoxide. w. Mallinckrodt "Analytical Reagent" Grade sodium bromide and chloride were used without further purification. Apparatus fliflmowmiESLiAPEOQREQ-ic813—3111 Cell Holge}: The early qualitative studies concerned with the water-bromine system were made with a Beckmen DK-2 recording Spectrophotometer. A pair of 'one centimeter quartz cells were used in this phase of the work . With the exception of these first studies, the remainder were made With a Beekman Model DU spectrophotometer. A one millimeter cylindrical quartz cell was used for these quantitative runs. All of the runs made with the Beckmm DU were carried out at 25°C. The cell compartment was equipped with a pair of Beekman thermospacers, t111-‘r‘iough which water from a thermostated bath was pumped . Temperatures Were controlled, by these means, to approximately 25 i 0.500. The cylindrical shape of the special one millimeter cell required a Special cell holder, since its diameter was too large for insertion into one of the rectangular compartments of the standard Beckman cell holder. A satisfactory holder was constructed by modification of one 82 of the ordinary Beckman one «centimeter cell holders. A sketch of the modified piece is shown in Figure 17. The inner sections of the original- holder were removed, and a U—shaped aluminum assembly was fitted into the resulting rectangular skeleton. In use, the cylindrical cell would be placed into the U-cavity, which was lined with felt to hold the cell in a secure position. The aluminum assembly was so positioned that the cell was centered with respect to the second of the four original cell positions in the holder. To eliminate light scatter- ing and reflection, the diameter of the light beam was restricted to a 3/8 inch opening, which permitted light to pass through the cell without striking the side walls. With this arrangement, a reference cell could not be used. However, because of the more critical necessity for reproducing cell position with the small cylindrical cells, it was felt that the usual lateral motion of the entire cell holder to inter- Change sample and reference cells would not have been satisfactory. A Somewhat less convenient but more reproducible technique was used, employing the one cell and making the necessary corrections. It was found that with the cell holder in a stationary position, the position of the cell was quite reproducible; for any one of the solutions studied, alternate removal and replacement of the cell invariably gave the same abacrbance reading. Therefore, for any solution, the correction necessary was subtraction of the absorbence reading of the cell filled With the solvent, water. Since neither of the salts, sodium bromide or Chloride, show" any absorption in the range of wavelengths studied, this Procedure proved to be a satisfactory one. In practice, while taking 83 FELT-LINED AREAS ALUMINUM INSERT FOR SUPPORT OF I MM. QUARTZ CELL 3/8" SLIT ALUMINUM msem ~ AS SHOWN ABOVE INSERT FITTED INTO MODIFIED BECKMAN CELL HOLDER Figure 17. Modified Cell Holder for the one millimeter quartz cell. 8h readings from wavelength to wavelength, the cell was first removed, and the instrument then balanced with no obstruction whatsoever in the optical path. This served as the reference balance, analogous, of course, to the usual balancing of the instrument with reference cell and solvent. In the calibration of the cell with solvent, the same referalce was used; thus, the overall correction would simply be the abSorption due to the cell and solvent. Since the latter are constant for any given wavelength, they would therefore represent the correction necessary in all of the runs. These corrections are listed below (Table XVI) for the wavelengths under consideration. TABIEXVI ABSORBANCE CORRECTIONS FOR THE} ONE MELD’UETER QUART? CELL AND SOLVENT (WATER) AT VARIOUS WAVMNGTHS __‘.__‘ _ _— A— _— *v Viv w “Va—v ‘ —— w v—v—r #— —— _- _ H _ *_ _J r‘ w Wavelength (R) . Ab Sorbance Correction 23 80 -O .0 85 2.1.00 -0 .082 2500 —0 .0 7 7 2600 -0 .062 2660 -0 .050 2700 -o.050 2800 -O .050 2900 -O .050 3000 -0 .050 ‘l _ _ _n. _ __ _— * ——v——w‘—' —— V—w fiv—vw W W _r The indicated 1.000 millimeter optical- path length of the cell was checked by meamring the absorption of a reference solution. An a‘lkaline solution of potassium chromate for this plrpose has been described by Hanpt (32). Included in Haupt's paper is a table of .85 absorbence readings for the entire ultraviolet portion of the spectrum. The standard solution used was one containing 0.01400 g. .of potassium chromate per liter in 0.05N potassium hydroxide. The results obtained in the present work as well as those reported by Haupt are shown in Table XVII . TABLE XVII RESULTS OF CHECK 0N OPTICAL BATH LENGTH OF THE ONE yILLIMETEE QUARTZ CELL BY'USE OF A REFERENCE SOLUTION” ‘— ' Ll __ I -. ‘ 1 17—v W“ _v—w W—fi —-_ v—v—v ‘_ _— .~__ —w ‘— I Ab sorbance i 2' (from Haupt) (Corrected) (1 cm. light path) Absorbance Absorbance Absorbance Wavelength(B.) Correction (Total) fififi —— j—v: fi_—v _— 2200 -0.078 0.118 0.0LI0 0.hh6 2250 ~O .073 O .093 0 .020 0 . 221 2300 -0 .07).; 0 .090 O .016 0 .l71 2350 -O .078 0 .098 ‘ O .020 0 .210 2h00 ~0.082 0.098 0.029 0.295 ,2h50 -0.080 0.119 0.039 0.396 *HanptIS studies with the reference solution are discussed above. AS indicated in the preceding discussion, the correction values repre- 3exit the absorbence of the cell plus solvent versus air; the total. abB-orbance is that of the solution plus cell versus air. The corrected Values obtained are within experimental error smaller than Haupt*s by the anticipated factor of ten. The results of another interesting check are summarized below (Table XVIII). In this study, the one—millimeter cell was compared with one marked 1.000 centimeters. Here too, the cells were used 86 TABLE XVIII ABSORPTION DATA 1“OR EVALUATION OF THE RELATIVE OPTICAL PATH LENGTHS OF THE ONE MIILBETER AND ONE CENTDIETER QUARTZ CELLS/r fim Absorbance Cgrrection Absorbance (Total) Absorbance (Corrected) Wave- ”Tmmnf 1.0007}: 1.000 mm. 1.000 cm. 1.000 mm. “1.000 cm: island) cell . cell cell cell cell cell 2200 -0.078 -0.096 0.193 1.211 0.115 1.111 2250 —o.073 -0.089 0.1112 0.800 0.069 0.711 2300 «0.0711 -0.09h 0.136 0.725 0.062 0.631 2350 ‘-0.078 -0.106 0.157 0.918 0.079 0.812 2380 -0.085 —0.112 0.18h 1.13 0.099 1.02 21:00 -0.082 -0.113 0.199 1.29 0.117 1.18 2150 -0 .080 -0 .107 0 .2142 l .68 0 .162 l .57 _ wk — fl _. _fi —fi _v—— #— AA A; — ‘_ ~__ __ _ _A *The solution (absorbing medium) used in this study is discussed below. individually to obtain readings, i.e., no reference cell was used. A Solution considerably more concentrated in potassium chromte than that used for the preceding study served as the absorbing medium. Since attention-here was directed primarily to the relative absorbances Obtained with the two cells, no effort was made to prepare the solution quantitatively. As indicated by the data, the results appear quite . good. The acceptance of the value, 1.000 mm., seems to be a valid one. It will be noted that the above studies were limited to the short 1wavelength region of the ultraviolet. This range only was simdied betcaulse, at the time, major concern was with the 23808 position. 87 However, there seems to be no reason to doubt that comparable results would have been obtained at longer wavelengths. Apparatus and Method for Preparation of Solutions Special techniques were developed. for premring and transferring the solutions. The discussion which follows illustrates the procedure used in a typical run . The second water distillation, as mentioned previously, was per- formed in a nitrogen atmosphere. A sketch of the apparatus is shown 1 3. f .1’ ... :Ln Figlre 18. (All joints, stoppers, and caps were of the ground-glass type: stopcocks A and B were of Teflon). The desired weighed quantities of sodium bromide and/or chloride were introduced into the empty volu- metric flask before beginning the distillation. Stirring, provided by a. magnetic stirrer (encased in glass), was necessary for solution of the salt and subsequent attainment of a homogeneous system. At the c{Lose of a distillation, the flask was removed from the distillation at-PZ£ZJa.‘1‘a.tus and the neck opening C immediately stoppered; nitrogen flow was maintained during this operation. The bromine was then introduced with the aid of a hypodermic syringe and micro-pipette. To accomplish this, the elbow-like piece D, leading upward from the neck of the Volumetric flask, was partially lifted. from the flask, the outlet E capped, and the bromine released into the flask. The piece D was then lifted clear of the flask and the neck opening F stoppered. With this Procedure, the bromine was introduced against a continuous flow of IL'I-‘l‘lrogc-in. The purpose for taking the above precautions was to prevent 88 NITROGEN FLOW MAGNETIC STIRRER T FLOW D DISTILLATE FROM / DI STILLATION APPARATUS B I F__s—> NITROGEN / POROUS OUTLET FOR NITROGEN Figure 18. Apparatus for preparation of solutions. 89 entry of any carbon dioxide into the system. In the early studies, where the added salt content was very low, the increased hydrogen ion concentration occurring from the presence of any carbon dioxide might very well have affected the equilibrium concentration of bromide ion, and, consequently, the tribromide ion concentration. This procedure ms used throughout the work. The flask itself had previously been completely covered with black photographic cloth, thus minimizing any effects produced by light. Bromine was transferred from its storage vessel in the vacuum line by a. vastnun distillation into a cold finger, from which it was removed with the aid of the hypodermic syringe and micro-pipette. The removal was nude in a stream of nitrogen._ Figure 19 shows the section of the vacuum system used for the distillation. The nitrogen flow was used, of course, only after the distillation, while removing the bromine. It ms found that Hooker Fluorolube (Grade MG) served satisfactorily as a. lubricant for stopcocks and Joints in the vacuum line. As would be expected, the more common greases, including the Dow—Corning silicone type, were subject to bromine attack, this effect being observed by the liberation of hydrogen bromide and darkening of the grease. Apparatus and Method for Filling the Absorption Cell A Vacuum technique was utilized to fill the cell. Figure 20 is help-fir]. in describing the procedure used. The diagram shows the elbow- like Piece D refitted to the solution flask. The stem of the absorption cell was attached as shown to an intermediate section containing the .Empmcnm assomS CH oqwsoan .Ho soapmaafindz .3 magmas 9O uziomm oumzuozoo 1:3 mung—.... 01.00 memm> mo<¢0hm . 304a zmeomtz ell J. a2...“— 2330<> O... 91 ABSORPTION __S—>. cELL R TTO VACUUM LINE FITTED TO SOLUTION FLASK Figure 20. Apparatus for filling the absorption cell. A 92 Teflon stopcock G. By proper manipulation of the stopcocks, the cell was first evacuated and then filled. The cell, together with the small section H containing the stopcock G (closed), was then removed as a unit and absorption readings taken. All regions of the apparatus, including the cell itself, were protected from light by careful- masking with black cloth. Measurements with the spectrophotometer were made in a darkened room. Determination of Volume of Solution and Volume of Solvent The vollune of solution used in each of the runs was determined as follows. (During the distillation, the volumetric solution flask was fIL‘Lled approximately to the 1000 cc. line etched on the neck of the container. However, upon removal. of the tube leading nitrogen into the System, the level would be several centimeters lower. Hence, the final vOZLume of solution, for each run, would be slightly less than 1000 cc.) For each run, the distance from the 1000 cc. marking to the meniscus formed by the salt solution was measured; the exact volume was then determined by raking the necessary correction. This correction was determined directly from the data of a previously performed calibration of - the volumetric flask . Table XIX shows the data obtained in the Calibration, which was carried out with ordinary distilled water. The. flask was first filled to the level for which the meniscus—to-line distance was equal to h.5 cm. A burette was then used. to add the increments of water corresponding to 0.1 cm. changes in meniscus height. It Should be mentioned that the volume data for the runs were obtained 93 (m1.) Correction to 1000 cc. Mark (cm.) DistanceuMeniscus TABLE XIX ml.) Correction CALIBRATION OF VOLUMRI'RIC SOLUTION FLASK (cm.) to 1000 cc. Mark Distance-«Meniscus 92hx-81h703703603692592 S/m/m66777888999000011 2. 111111 h567890123h567890123h5 2222223333333333h14141414h 2570250I92579136803691h6 0001111122223333hhhh555 123145678 90123.“..5678 90123 00000000011111111112222 9b. prior to the addition of bromine. The volume increase resulting from the bromine addition could be neglected since, in every case, less than 0.0S cc. of the liquid bromine was used. However, the volume of the glass-encased stirrer was considered and the necessary correction made. Table XX contains the detailed data for this phase of the work. The use of the molality scale in our calculations required knowl- edge of the exact quantities of the solvent, mter, used in each run. These data were obtained by preparing a second solution for each of those studied. In the preparation of these duplicate solutions, ordinary distilled water was added to the same weight of salt as used for the "mm" solution, and dilution effected to the same volume as for the latter. In this instance, the volume of water was measured to obtain the desired data. Table XX also contains these data. Analysis of Solutions for Bromine Content Analysis of the salt solution for bromine content was carried out by titration with standard thiosulfate solution. Because the solutions were qlite dilute in bromine (3 x 10-4 to 10"3 molar), it was necessary to use thiosulfate solution of about 0.01N. A rather large aliquot of the Salt solution had to be used for each titration so as to require sufficiently large volumes of the titrating solution. For eacn run, JGl'l-I‘ee separate one hundred milliliter aliquots of the salt solution were titrated . The usual analytical method was employed—the aliquot was added to an acidic potassium iodide solution and the iodine thus liberated titrated using starch solution as an indicator. It was found DATA FOR COMPUTING SALT CONCENTRATIONS OF THE SOLUTIONS __... ~ TABLEXX m _A w—w—v 95 Volume Distance Correction Volume Volume Volume (Meniscus to (Preceding Correction of of Run 1000 cc. Mark) Column) (Stirrer) Solution Water No. (cm.) (cc.) (cc.) (cc.) (00.) 138 1.8 h.3 2.8 992.9 992.6 139 1.8 8.3 2.8 992.9 992.5 11.0 2.0 11.9 2.8 992.3 991.3 1.112 2.8 6.8 3.0 990.2 989.0 11.3 3 .h ' 8.7 3 .0 988.3 987.1 11.1. 3.11 8.7 3.0 988.3 986. 11.5 3.8 10.0 3.0 987.0 982.7 1116 3.7 9.6 3.0 987.17 979.1 1177 2.0 11.9 3.0 992.1 971.7 1118 3.7 9.6 3.0 987.11 959.8 1119 3.7 9.6 3.0 987.1: 971.9 150 1. .5 12 .2 3 .0 9811.8 960 .1 151 3.7 9.6 3.0 987.1; 931. 152 17.3 11.5 3.0 985.5 923.2 155 1.8 h.3 2.2 993.5 980.7 156 1.8 17.3 2.2 993.5 976.3 157 1.2 2.9 2.2 991.9 987.1 158 0.0 0.0 2.2 997.8 992.8 159 1.3 3.1 2.2 9911.7 990.8 160 3.5 9.0 2.2 988.8 978.1 161 1.14 3.3 2.2 991.5 985.3 162 1.9 11.6 2.2 993.2 9811. 163 1.6 3.8 2.2 9917.0 985.0 161., 1.7 17.0 2.2 993.8 981.1 165 1.7 8.0 2.2 993.8 Who 166 2.17 5.9 2.2 991.9 985.0 167 2.8 6.8 2.2 991.0 9811.0 168 2.0 11.9 2.7 992.1. 976.1 169 2.5 6.2 2.7 991.1 972.9 171 17.0 10.6 2.7 986.7 977.1 172 3.1 7.7 2.7 989.6 980.9 173 17.0 10.6 0.9 988.5 970.7 1714 2.2 5.11 0.9 993.7 973.7 175 17.2 11.2 0.9 987.9 970.8 176 3.8 10.0 0.9 971-0 4 1 J l I . 989.1 ‘“_ 96 that the volumes of titrating solution regiired for each of the three titrations were very nearly equal, the differences usually being of the order of a fraction of one per cent. Removal of the one hundred milliliter aliquots ,Was made as rapidly as possible with a pipette. At all other times, the flask was tightly stoppered' to prevent any loss of the volatile bromine. Of course, it would be expected that little loss would occur under any circumstances, since most of the bromine was present, under the existing equilibrium conditions, as the nonvolatile tribromide ion. The titration data for all of the runs considered in the preceding sections are shown in Table XXI. Conversion of Activity Coefficients As indicated previously (page 20), t was necessary to convert all activity coefficient values from the molality to the molarity scale. The conversion formulae were also briefly discussed in that section. Table XXII shows the results of all of the conversion calculations. For each run, the molality and molarity of the solution (with respect to salt concentration) is listed as well as the molar and molal activity coefficients for bromine and. hydrobromic acid . The density of each solution is also included. Several aSpects of the calculation procedure should be pointed out. The simplest of the conversion formlae is, as shown previously, fi = (1 7 0.001mmB)(d°/d) “'1 A value of 0.9970, the density of water at 25°C was used for do (33). T TABLEXXI ANALYSIS OF SOLUTIONS TOR BROMINE CONTENT h—h.‘ ww— ~— *_ — a; w A— ‘— ~- w —_ ...—‘ w _— 97 Ml. of Na23203 Required for Bra Conc. Run Titration of 100 cc. of Sol'n. Normality of' (moles/liter) Number Tit. 1 Tit. 2_‘iit.3f Mean— 11.1159203 (x 104) 138 5.50 5.87 5.85 5.87 0.01076 2.98 139 7.20 7.20 7.23 7.21 0.01076 3.88 180 12.55 12.50 12.52 12.52 0.01076 6.78 182 8.80 8.78 8.70 8.75 0.01076 8.71 183 7.86 7.87 7.88 7.87 0.01076 8.28 188 6.96 6.76 6.80 6.88 0.01076 3.68 185 7.20 7.12 7.18 7.17 0.01076 3.86 186 7.15 7.15 7.18 7.16 0.01076 3.85 187 7.28 7.15 7.18 7.19 0.01076 3.87 188 7.12 7.12 7.18 7.13 0.01076 3.88 189 7.88 7-86 7-85 7.85 0.01076 8.01 150 7.66 7.60 7.63 7.63 0.01076 8.11 151 7.32 7.35 7.29 7.32 0.01076 3.98 152 7.18 7.22 7.17 7.18 0.01076 3.87 155 7.30 7.80 7.80 7.37 0.01060 3.91 156 7.58 7.83 7.55 7.51 0.01060 3.98 157 7.58 7.55 7.66 7.58 0.01060 8.02 158 7.88 7.82 7.50 7.87 0.01060 3.96 159 7.08 7.00 7.02 7.03 0.01060 3.73 160 7.71 7.71 7.78 7.72 0.01060 8.09 161 17.32 17.36 17.80 17.36 0.01060 9.20 162 16.50 16.53 16.53 16.52 0.01060 8.75 163 17.76 17.67 17.75 17.73 0.01060 9.80 168 16.63 16.52 16.65 16.60 0.01060 8.80 165 17.25 17.19 17.22 17.22 0.01060 9.12 166 15.88 15.82 15.88 15.86. 0.01060 7.93 167 18.00 18.03 17.97 18.00 0.01060 9.58 168 16.92 16.88 16.98 16.93 0.01060 8.97 169 17.88 17.82 17.72 17.79 0.01060 9.83 171 3.73 3.85 3.80 3.79 0.01055 2.00 172 7.62 7.59 7.65 7.62 0.01055 8.15 173 7.68 7.75 7-75 7-73 0.01055 h-08 178 7.68 7.59 7.65 7.63 0.01055 8.03 175 7-50 7.58 7.50 7-53 0.01055. 3.97 176 17.78 17.78 17.70 17.75 0.01055 9.37 A; I 1 9 8 TABLE XXII SUMMARY OF ACTIVITY COEFFICIENT VALUES AND RELATED DATA W Dmsity Molar M0131 of Molarity Molality Activity Activity Run Solution of of Coef; -. Coef . No. (grams/cc.) Solution Solution Br2(f) Emmi)" BrXfl HBrCK'Qi- 138 0.998 0.009623 0.009626 1.002 ~21. 1.002 m1 139 0.998 0.01371 0.01372 1.003 «21 1.003 «.1 11.0 0.999 0.03051 0.03058 1.007 0.857 1.006 0.855 1112 1.001 0.08386 0.08391 1.009 0.838 1.009 0.838 11.3 1 .001 0 .08578 0.08580 1.011 0 .832 1.009 0 .831 11.14 1 .008 0 .08598 0 .08613 1. .019 0 . 808 1 .018 0 .803 185 1.013 0.1980 0.1989 1.085 0.771 1.080 0.768 11.6 1.030 0.8021 0.8055 1.095 0.762 1.086 0.756 1117 1.061 0.8250 0.8825 1.211 -- 1.186 -- 188 1.083 1.108 1.180 1.296 —- 1.260 -- 1119 1 .053 0 .6972 0 .7082 1.172 _— 1.158 -- 150 1.082 1.068 1.091 1.279 1.288 .- 151 1.162 2 .187 2 .275 1.679 -- 1 .586 -- 152 1.365 8.187 8.870 2.638 _- 2.87h -- 155 1.027 0.7267 0.7361 1.208 -- 1.189 -— 156 1.036 0.9571 0.9738 1.279 0.799 1.257 0.785 157 1.019 0.5017 0.5057 1.135 0.789 1.126 0.783 158 1.012 0 .3889 0 .3866 1 .090 0 .788 1.085 0 . 780 159 1.008 0.1766 0.1773 1.087 0.768 1.083 0.765 160 1.029 0.7328 0.7808 1.203 0.767 1.190 0.759 161 1.019 0 .5278 0 .5323 1 .188 0 . 788 1 .133 0 . 781 162 1.012 0 .8183 0 .8221 1.118 .- 1 .108 _— 163 1.103 0.8193 - 0.8231 1.115 -- 1.105 _- 168 1.025 0.6901 0.6990 1.198 0.761 1.178 0.752 165 1.081 1.079 1.101 1.322 0.825 1.296 0.801 166 1 .021 0 .5215 0 .5252 1.139 ~- 1 .131 - 167 1.021 0.5232 0.5269 1.180 0.786 1.132 0.781 168 1.038 0.9102 0.9258 1.263 0.791 1.283 0.779 169 1.085 1.138 1.156 1.336 0.831 1.312 0.816 171 1.027 0.5912 0.5969 1.158 .— 1.186 -- 172 1.022 0.5373 0.5820 1.188 - 1.138 _- 173 1.082 1.028 1.087 1.283 «- 1.260 -- 178 1.081 1.089 1.111 1.328 .- 1.298 - 175 1.038 0.9218 0.9381 1.268 -- 1.286 _— 176 1.037 0.9962 1.015 1.292 .- - 1.269 - _fi ¥ {1 '1 J _—v‘— *The activity coefficient of hydrobromic acid was not required in all of the runs. 99 Knowing the volume and density of water, the weight of the salt(s), and the volume of the solution, the density of the solution, d, was then calculated from the expression, d= _fi lOO SUMMARY An ultraviolet spectrophotometric investigation has been made of the formation equilibria of the tribromide and dibromochloride ions in aqueous solution at 25°C. The formation equilibrium constants were determined and compared with values obtained by other workers—in the past work, non-spectrophotometric methods were used almost exclusively. Molar extinction coefficients for the W0 ions have been determined at a number of wavelengths. The system, water—bromine-sodinm bromide, was chosen for the tri— bromide study, formation of the tribromide ion resulting from the reaction, Br2 ~i- Br- =1 Br; For this reaction, the equilibrium constant is K a'Br‘ .. (3133-) £13333- : :_ wfifiifi = W 3 831.2 3B1.” (Br2)(.Br ) fBrZ Br” where the quantities in parentheses represent concentrations of the mrticular species, and a and f the activities and activity coefficients, respectively. Absorption of the trihalide ion is characterized by a Very intense band with a peak at 26608. A graphical means was employed for treatment of the Spectral and equilibrium data. The equilibrium Constant, K3, and the molar extinction coefficient at the band mxirmim, 101 ..l 826 50, were found to be 17.3 liter mole and 3 .86 x 104 respectively. Other workers, using distribution methods, have reported values for K3 of approximately this value (ll). Job, the only other investigator who has made a Spectral determination, has reported a value 28.6 liter _1 mole. (17). To ascertain the effects of bromine hydrblysis on the tribromide ion formation, several methods of Calculation were tried . With solu- tions dilute in Sodium bromide, the'hydrolysis reaction, H20 + Bra .. HOBr + H!" Br“ could conceivably, because of bromide formation and bromine consumption, affect the tribromide eq;ilibrium. However, even in solutions as dilute as 0.01 molar in added bromide (bromine concentrations varied from 10-4 to 10"3 molar), hydrolysisswas shown to be virtually completely alppressed. The dibromochloride ion study proved to be much more complicated. For this study, the system, water-bromine—sodium chloride, was chosen. Ultraviolet absorption for this system is characterized by an intense band with a peak at 23808. The formation reaction and equilibrium constant expression are analogous to those for the tribromide ion. The results obtained, were, however, inconsistent with the rather con- cordant, results in the literature. The equilibrium constant for this System was found to be 0.567 in contrast with a literature value of 1«39 liter mole"I (20) . Complicating side-reactions are believed to 13183 a role in this system. A possible reaction scheme may be the 102 following: Br3 N Br; + - Brgcl = BrCl -:~ Br 4. 01‘ ll _, ClBrCl Br, + 01‘ Since the addition of bromide would retard these side effects, it was decided to study the system, water-bromine-sodium chloride-sodimn bromide. The analysis was thereby complicated since the sirmltaneous equilibria , Br2 + 81": Br201‘ Br2 - Br” = Br; had to be considered. In addition, the spectra were more complex since the ultraviolet absorption bands of the two ions overlap. Nevertheless, by considering the absorption of both ions at several Wavelengths, a method of Solution was developed. A value of 1.39 liter mole"1 , in very good agreement with the reported literature values, was calculated for the equilibrium constant. The method also 13K1‘t0 a characterization of the dibromochloride ion.absorption.band for the spectral region studied. At the wavelength of maximzm absorp— tion, 2830K, the molar extinction coefficient was estimated to be 2.1.9 X 1040 103 A qualitative study was made of the water-bromine system. This study was undertaken to elucidate the short wavel ength band exhibited by this system in the approximate range, 2550 to 2650K. Numerous investigtors have reported the presence of the band (2) but no Satisfactory explanation has heel presented as to what may be the absorbing species. The study was necessarily qualitative because of the instability, as indicated by the Spectral changes, of the very dilute bromine solutions. Nevertheless, two distinct types of changes, one occurring in the dark and the other in the light, were recognized. In both cases, the changes consisted of a shifting of the band toward longer wavelengths; in the dark, however, the shift was accompanied by a. decrease in band intensity whereas an increase in intensity occurred in the "light" study. To explain the dual behavior, it was assumed that the tribromide ion is present in the system, and that the absorp— tion band under consideration is produced principally by the trihalide absorption. The tribromide ion would be formed by the reaction of bromine and bromide ion, the latter. produced by the hydrolysis of bromine. Hypobromous acid, also produced in the hydrolysis reaction, also absorbs in this region of the spectrum. The absorption of the tribromide ion and of the acid, together with the changes produced by the "light" and "dark" decomposition of the acid, can lead to the Plausible explanation presented and thus account for the observed Spactrum and its time-d ependence . ‘ 1 Q- . 9%“)...- 10h As mentioned at the begitming of the thesis, attempts to obtain a suitable solvent for the initially prOposed photochemical study were not successful. However, several interesting spectral phenomena were observed in these early studies, and these will be discussed in this section. . a order of presentation will be similar to that followed in Part I. Pertinent background material from the literature will first be reviewed. A discussion of our findings will then follow. Finally, experimental aSpects of the studies will be considered. HISTORICAL SUMMARY Pertinent Past Studies Popov and Swenson (18) have studied the systems, acetonitrile-- tetrabutylammonium tribromide and ethylene dichloride-tetrabutylammonium tribromide. These investigators obtained quantitative data on the absorption spectra of the tribromide ion in each of the solvents. In acetonitrile, the absorption mimum occurs at 2690.8; the molar extinction coefficient was estimated to be 55,000. In ethylene di- chloride, the absorption maximum is located at 27308 and the extinction coefficient was calculated to be 514,000. The tribromide ion is formed directly from the complete dissoci- ation of the tribromide salt. Because polyhalide ions themselves dissociate in acetonitrile, it was found necessary to add an excess of tetrabutylamnonium bromide to the solutions to suppress the further diBE’KDGZLation. Since the simple halides show no appreciable absorption at the wavelength investigated, no error was thereby introduced. PART II Bromine Complexing in Several Organic Solvents 105 Another series of studies that seem to be related to those in the present investigation are those made by'Keefer and Andrews (31;) . These workers, in addition to others, have clearly shown that bromine forms a. 1:1 complex with a large number of compounds. For example, they have studied such systems as bromine-methyl- iodide in carbon tetrachloride, bromine-n butyl bromide in the same solvent, iodine-t butyl chloride in heptane, and others. Using the formation equilibrium expression, X 2 + RY == RYX2 with equilibrium constant, RYX K é‘xfiz IEQRLTY together with the spectral data, it was possible to calculate the equilibrium constant for each system. The concordant results obtained by making the. calculations at different wavelengths seemed to leave little doubt as to the nature of the phenomena involved. Keefer and Andrews, in addition to other independent workers, have stated that in suchsystems, the halogen molecule functions as an electron acceptor. A structure for the complex based on a resonance hybrid of 3:35.323}; and related electronic forms seems reasonable. This belief is further supported by noting that the equilibrium constants for the formation of the alkyl iodide-iodine complexes in heptane solu- tion decrease with change in the structure of the alkyl group in the order, t-butyl > i-propyl > ethyl > methyl. This stability series is the reverse of that for the relative inductive effects of the alkyl 106 groups, and also the reverse of the effect which would result from steric factors. The Spatial geometry of the complexes may be similar to that proposed by Pauling (35) for trihalide ions. The trihalide ion has been shown, by X-ray methods, to be linear in crystals of alkali compounds. It is probable that the configuration is that of a trigonal ’7] bin with halogen atoms at the two pyramidal apices and the three unshared electron pairs in the equatorial plane. Still another series of studies pertinent to the present one is a very recent one by Popov and Deskin (36) in which various iodine halides, in the solvent, acetonitrile, were considered. These workers found that iodine monochloride, iodine bromide, and iodine each form 1:1 complexes with acetonitrile. Both their conductance and spectro- photometric data indicate that a definite change occurs with time for each of the systems. The slowly increasing conductance and accompany- ing absorption changes were interpreted as being associated with a slow transition of the "outer complex," CH3CND(, to the "inner complex," (CHachjf Xe". One other very interesting aspect of the system, acetonitrile-iodine, is that the triiodide ion is formed. The presence of this ion is indicated by its ultraviolet absorption, characterized by two intense absorption peaks, one at 3600.8 and the other at 2910.8. 107 THE PRESENT DIVFSTIC‘ATION--RESULTS AND DISCUSSION The Cyclohexane-Bromine System Despite the ideal ultraviolet transparency of the cyclohexane, the system formed with bromine was not found sufficiently stable to warrant _ its use. The instability was observed by studying the intensity of the [11,! bromine absorption'with time at the wavelength of maximum bromine . absorption, hZOOX. Studies were made in the dark as well as under normal conditions of room lighting. Since this cycloparaffin should be I relatively inert to bromine attack, it must be concluded that all traces of reactive impurities were not removed in the purification. As was stated previously, attempts to pirify this solvent were not extensive since it was believed, at the time, that another solvent, suitable for our purposes, would easily be found. The Acetonitrile-Bromine System As indicated previously, acetonitrile did not prove to be useful as a solvent primarily because of the intense ultraviolet absorption exhibited by this system. The absorption maximum of this band, which is shown'in Figure 21, is at approximately 27003. Popov's spectral investigation (18) 'of tetrabutylarmnonium tribromide in acetonitrile Sllggested that the tribromide ion may be the absorbing Species; both the intensity (qualitatively) and location of the bands in the two. systems are in agreement. In addition, it will be recalled that in the System, acetonitrile-iodine, Popov reported the presence of the 108 0009 .oflfipunsopood 5H ocfisosn mo oOHPSHom 333.6 m .Ho Estevan oofiohpomno peacebwhpflb .Hm madman a. 53392: [.92 SW 9923 -1 00V? -‘ 0662 it 4 9222 d OLOQ ‘4 9E6? .. SILZ -* 8292 — 81793 A 61.92 ““92 1. to l 0.0 I. ad BONVGUOSBV 109 triiodide ion (36). It would appear that in the analogous system, acetonitrile-bromine, the tribromide ion might well be present. The tribromide ion in the present system could be formed as follows: the solvent would first interact with the bromine to form a 1:1 complex, CH3CN + Bra ‘3 CHSCNBI'E Dissociation of the complex in the fairly highly-ionizing acetonitrile (dielectric constant = 37.5 at 20°C) might then proceed .as follows: CHBCNBrz = (CH3CN'Br+Br-) = CH3CNBr+ + Br“ {2) The tribromide ion could then be formed in the usual mnner, i.e., Br2 + Br- = Br; (3) When studied as a function of time, the spectra indicated that the SYStem underwent a gradual change. The nature of these changes seemed ’50 be dependent on the manner in which the solution was prepared. When no dilution was involved (bromine added directly to the solvent), the intense ultraviolet band exhibited a gradual increase in intensity. In all cases, this increase was observed on the band shoulder, since the absorption at the maximum was beyond the absorbance range of the inStI'ument. However, when such a solution was diluted so as to observe the absorption at the peak (27008) , the intensity showed a slow decrease With time. 110 Attempts to explain the dual behaviour of the system on the basis of the proposed equilibria (:1), (2), and (3) failed. An alternative explanation which seems in better agreement with the observed behaviour involves the two principal equilibria: CH3CNBr2 CH3CN + Br2 a ‘ H + _ CH3CNBr Br and CH3CN'Br2 4L _ 2 ". ‘ _ = CH3CNBr Br3 + CH3CN CHacNBr Br or CH3CNBr2 L _ H + _ + Bra = CHBCNBr Br3 CH3CNBr Br (1) (ha) (hb) The bracketed equilibria indicate that no distinction has been made between the "outer complex" and the "inner complex" ion pair in the reactions considered here. Equilibrium (1) is postulated to be rapid, With (kg) or (hb) slower equilibria. These may be accompanied, to a lesser extent, by the slow ionizations, CHacNBr'" e Br" CH30NBr*Br' CH30NBr*’Br3' = CH3CNBr+ +- Br; . Br; = Br;2 + Br The observed tribromide absorption is attributed principally to the ion pair, CH30NBr+Br3-', in accord with the ion-pairing tendencies in many TE 111 non-aqueous solvents, and in accord with the tendency of ions involved in ion pairs to show absorption very similar to that of the free ions. The slow attainment of equilibrium according to (ha) or (Lib) would explain the gradual changes in absorption intensity. In a freshly prepared solution (ha) or (hb) would slowly form the tribromide ion pair, with increasing absorbance as a function of time. Dilution would drive (ha) or (hb) towards the left so that the tribromide ion pair concentration immediately after dilution would be excessive, and its absorbance would decrease as the new equilibrium is approached. The Ethylene Dichloride-Bromine System Our studies of the system, ethylene dichloride-bromine, clearly indicate that some type of interaction results. It was observed that 8- Very intense new ultraviolet absorption occurs upon mixing the reactants; the long wavelength edge of the band is in the vicinity of ZSOOX. Figure 22 shows the absorption of the solvent alone and also Of a very dilute solution of bromine in the solvent. Unfortunately, the limited transparency of the solvent would not permit the whole of the band to be observed. It seems probable that complexing may be involved. As with the preceding examples of complex formation, the electrophilic bromine could form a loose "donor acceptor" or charge- transfer" type bond with the nucleophilic chlorine of the solvent molecule. A literature search revealed that no thorough studies have been previously made with this system. Buckles and Mills (3?) concluded that complexing does not occur and justified their conclusion on the Am sEteooomV opiodtae oqoatfipo ca @9293 mo .cofipsaom opsdmp a mo mad 2 Eapoonv opfipoggp oqodfipo mo wppomam cospmpomnw poH0fi>thHD .mm onsmfim 112 6 5333: 0009 —4 co» .. 0669 -+ 01.99: acts seas 0L0: - 9962 -w an 92 a 9| [.3 _-1 61.92 -1 9 n 92 99E? «.0 n6 v.0 0.0 0.0 Nd m6 m6 BONVBUOSBV all.» tn»? 113 basis of the relatively non—polar nature of the ethylene dichloride molecule. However, it appears that their studies were not extended to sufficiently short wavelengths; the absorption would therefore not have been noted. Another factor which may have caused these investigators to fail to observe the intense absorption might be concerned with solvent r purity. We have found that with extensive purification it was possible 1 to extend the ultraviolet cut-off of the pure solvent down to approxi- mately 22808. The purified ethylene dichloride used by Popov (18) for ' his tetrabutylammonium halide studies was transparent down to about :0 23508. It may very well be that the solvent used by Buckles had a cut- - off too near the region of abSOrption to have detected the latter. The possibility of the reaction of bromine with the solvent was considered but does not seem very probable. Study of the visible absorption band of bromine with time indicated no depletion of the halogen. In addition, any reaction would very likely have resulted in the formation of hydrogen bromide. The ultraviolet absorption of gaseous hydrogen bromide consists of a primary band with absorption Insacilmim, at 28703 and a secondary band with its maximum at BhBOX (38). The Fluorocarbon-Bromine Systems A few fluorocarbons were also tried as possible solvents. Those examined were Fluorolube (Hooker - F. 5.), perfluorokerosene (Du Font), and 2, 2, 3 trichlorheptafluorobutane (Hooker). The ultraviolet cut-off Values for these compounds were found to be 2850, 3280, and 25003, respectively. The rather high cut-offs precluded their use with bromine mi - languid.“ 11).; unless purification improved their ultraviolet transparency. However, because of the advantages of using solvents more readily available, purification was not attempted. 115 EXPERIMENTAL Source and Purification of Reagents gvyglohexane The cyclohexane (Eastman) displayed a weak absorption at approxi— mately 2900K with a somewhat stronger band at about 26008. Purification was attempted by passing the liquid through a l—meter, hO-mm. i.d. column packed with silica gel (39). This procedure resulted in a productrwith a cut—off of approximately 2100X as compared with one of approximately 22003 for the unpurified liquid. The absorption at 29003 diaplayed by the latter was virtually eliminated by this procedure. Acetonitrile Several plrification schemes were tried for the acetonitrile (Matheson, Coleman, and Bell). The unpurified solvent, with a cut—off at 2180.2, showed an absorption in the 2600-29003 region. A purification method used by Wawzonek and Runner (ho) proved to the most efficient of those tried. For a typical purification run, the procedure was as follows. Five hundred milliliters of the solvent was fractionally distilled. A middle fraction of M40 cc. was then added to the same VOlume of saturated potassium hydroxide solution and the mixture agitated for a period of six hours. After separation from the aqueous phase, an aIllple supply of sodium carbonate was added to the organic phase and this mfiqcture stirred for two hours. The separated organic phase was then 116 distilled from phosphorous pentoxide. A middle fraction of 300 cc. showed excellent ultraviolet transparency, the cut—off being in the nrange 1990—20708. Istgzlene Dichloride Purification of the ethylene dichloride was also, attempted by various means. Eastman solvent (White Label) displayed appreciable absorption in the 2700-29008 region with a cut-off value of approxi— mately 2h708. Traces of benzene also appeared to be present as indi- cated by slight absorption at 2610 and 25508, these being superimposed on the main spectrum. Of the various schemes attempted for purification, the following appeared to be most effective. Quantities of the reagents used are included, although they are probably not critical. Four hundred milli- liters of the solvent, )3, cc. of bromine, and 5 grams of iron filings were Mixed in a liter flask, and the mixture was then refluxed over- Iright. The solution was then functionally distilled and the spectrum of each fraction obtained. Cut-off values for all fractions in the first 320 cc. of distillate were in the range 2380—21:208. Ninety nulli- liters from the middle fractions was then mixed with 1 cc. of bromine and this mixture was refluxed overnight. Fractional distillation of the latter followed by spectral studies of the fractions showed that the initial 60 cc. of distillate had cut-off values in the range 2270- 22902. The addition of the iron catalyst was necessary to remove the benzene present, the benzene being bromimted with the catalyst presents- 117 Separation was subsequently readily effected in the distillation Step. The boiling points of ethylene dichloride and benzene are 83.5 and 80.1%, respectively; separation by fractional distillation would have been extremely difficult. 0n the other hand, the high boiling point of the brominated product, bromobenzene, effects easy separation. In runs with bromine present, but in the absence of the iron catalyst, the benzene band was observed in its entirety with the fractions obtained in the subsequent distillation. Refluxing of the solutions was carried out in a lighted hood, the light being furnished by two 200-watt tlmgsten lamps. It was noted that best results were obtained with a straight reflux condenser-~for some inexplicable reason, use of a spiral reflux condenser repeatedly resulted in a less transment solvent. Since the purification method described results in a solvent with excellent ultraviolet tranSInrency, it should serve satisfactorily as a general method for obtaining spectroscopic~ grade ethylene dichloride. The 'plrification method is based upon the following crude, but rather fruitful, analogy: groups displaying chromophoric behaviour in the near ultraviolet region of the Spectrum tend also to be reactive with bromine. The high boiling points of the resultant brominated products facilitate their removal by distillation. Amatus All of the Spectral studies considered in this section of the thesis Were made with a Beckman DK-2 recording spectrophotometer. A pair of One-centimeter quartz cells was used throughout. The systems were Studied at room temperature . 1’18 SUMMARY The studies considered in this section of the thesis were a direct conseqtence of attempts to obtain a solvent which is both inert to bromine and tranSparent throughout the ultraviolet region of the Spectrum. A solvent with these characteristics was desired for another investi- gation. Although a satisfactory solvent was not found, the studies revealed several interesting Spectral phenomena. These findings are presented. The solvents investigated are cyclohexane, acetonitrile, and ethylene dichloride, the latter two being studied rather extensively. Several fluorocarbon compounds were 3130 tried as possible solvents. Despite the ideal ultraviolet transparency of the cyclohexano, the system formed with bromine was not found to be sufficiently stable. All traces of reactive impurities were apparently not removed in the pirification. Acetonitrile did not prove to be a satisfactory solvent, primaril y because of the intense ultraviolet absorption exhibited by the system, acetonitrile-bromine. The absorption maximlm of this band is at approximately 27003. Popov’Sspectral investigation of tetrabutyl- ammonium tribromide in acetonitrile (18), a study in which the spectral Characteristics of the tribromide ion in this solvent were defined, suggested the presence of this ion in the system, acetonitrile-bromine. Both the intensity Qqualitatively) and location of the bands in the two systems are in agreement. Other evidence also indicate that the 119 tribromide ion is the absorbing. species. For example, Popov has found that the triiodide ion is produced in the system, acetonitrile-iodine ( 36) - The intense band formed in the acetonitrile—bromine system was found to undergo very slow changes with time. Two distinct trends were recognized; when the solution was prepared directly, i.e., bromine added to the acetonitrile, the 27003 band underwent a slow increase in intensity--when the solution was prepared by dilution, the band in- tensity slowly decreased. Plausible explanations have been presented for the mechanism of tribromide ion formation as well as for the changes described above. The. following two equilibria have been postulated to explain the observed results: CH3CNBr2 CHacN + Bra :8 H + _‘ (1) ' CH3CNBr Br and CHBCNBrz _ 2 H_ + _ = CHscNBr'Bra -+- CH3CN (2) CHBCNBr Br or CHBCNBrz J __ . H + _y +- Br2 = CH3C1\IBr"Br:3 (2a) CH3CNBr Br ' The observed tribromide ion absorption is attributed to the ion pair, CHacNBr'le-a‘. Considering (1) ass rapid and (2) or (2a) as slower equilibria, the time-dependent spectral changes can be explained. 120 Ethylene dichloride, after extensive purification, was found to be inert to bromine. In addition, the purified solvent alone is fairly transparent in the ultraviolet regien, the cut-off value being approxi~ mately 2280X. However, a.solution.of bromine in ethylene dichloride exhibits an intense absorption.in the region below 25002. Since only the long wavelength edge of the band could be observed, it must be lassumed that the band peak is located at a shorter wavelength than that at which the solvent itself begins to absorb. It is probable that the 1:1 complex, ClCH20H2ClBr2, is formed in this system; this species could account for the absorption, .A purification method for ethylene dichloride was devised which Should serve satisfactorily as a general means for obtaining a spectro- scopic grade of the solvent. The purified ethylene dichloride is more transparent than the product used by other workers. The method involved two treatments of the solvent with bromine, followed, in.each case, by a.flactional distillation, 'With the first treatment, an.iron.catalyst was used to promote bromination of trace impurities; in the second, a light bromination.was utilized. The relatively higher boiling points of the brominated products facilitated separation by the subsequent dis- tillatidns. Several fluorocarbons'were alSo tried as possible solvents. waever, the rather poor transparenqy of those investigated precluded their use with bromine. Purification was not attempted because of the advantages of using solvents more readily available. (1) P. (2) L. (3) M. (h) H. (S) H. (6) F. (7) s. (8) R. ..(9) N; N. (10) G. (11) D. (12) G. 121 LITERATURE CITED Bovis, Compt. rend., 1.82, 57 (1927); C. A., 11;, 3158 (1927). I. Kat-zin, J. Chem. Phys., g9, 1165 (1952). Anbar and I. Dostrovsky, J. Chem. Soc., 1105 (1951;). A. Liebhafsky, J. Am. Chem. Soc., 6, 1500 (19314). A. Pagel and W. W. Carlson, J. Phys. Chem., Ag, 613 \1936). Pollak and E. Doktor, Z. anorg. allgem. Chem., 12.0, 89 (1931). A. Shilov, J. Am. Chem. Soc., 9‘9, 190 (1938). H. Bette and A. N. Ma‘cKenzie, Can. J. Chem., g9, 655 (1951.). M. Latimer, "Oxidation Potentials," Prentice-Hall, Inc., New York, 1., 1952, Jones and S. Baeckstrom, J. Aim-Chem. Soc., 28, 1517 (1931;). B. Scaife and H. J. v. Tyrrell, J. Chem. Soc., 386 (1958).. 'Jones and M. Li. Hartmann, Trans. Am. Electrochem. Soc., 22., 295 (1916‘) . (13) R. O. Griffith, A. McKeown, and A. G. Winn, Trans. Faraday Soc., gg, 101 (1932).. (lb) G- (15) F. (16) F. (17) P. (18) A. N. Lewis and. H. Storch, J. Am. Chem. Soc., 92, 25th (1917). P. Worley, J. Chem. Soc., 531, 1107 (1905). I L. Gilbert, R. R. Goldstein, and T. M. Lowry, ib_i__d_., 1092 (1931). Job, Ann. Chim., _9_, 135 (1927). I. Popov and R. F. Swensen, J. Am. Chem. Soc., 11, 3721;, (1955). (19) A. A. Jakowkin, Z. physik. Chem., 29., 19 (1896). (20) E. A. Dancaster, J. Chem. Soc.,. 12g, 2038 (192m. (21) P. Ray and P. v. Sarkar, J. Chem. Soc., 121, 11.119 (1922). 122 ('22) M. Randall and C. F. Failey, Chem. Rev., 1,, 271 (1927). (23) H. S. Sherrill and E. F. Izard, J. Am. Chem. Soc., 53., 1667 (1931). (21;) P. Debye and J. McAulay, Physik. 2., 26, 22 (1925). (25) H. s. Harned and w. J. Hamer, J. Am. Chem. Soc., 55, M96 (1933). (26) H. S. Harned, A. S. Keaton, and J. G. Donelson, ibid., 58, 989 (1936)- (27) J. a. Hawkins, ibid., 51,, tho (1932). (28) R. A. Robinson and H. S. Harned, Chem. Rev., _2_8., h19 (19141). t 5 {as L! O“ . (29) R. A. Robinson'and R. H. Stokes, "Electrolyte Solutions," Butterworths Scientific Publications, London, 1955, p. 31. (30) J. J. Custer and S. Natelscn, Anal. Chem., 21., 1005 (19119). 355" (31) W. H. Bauer and F. Daniels, J. Am. Chem. Soc., 5_6_, 1005 (1931;). (32) G. w. has, J. Pot. Soc. Am., 12, Am (1952). (33) "Handbook of Chemistry and PhySics," 3lst ed., Chemical Rubber Publishing Co., Cleveland, Ohio, 19119, p. 1720. (311) R. M. Keefer and L. J. Andrews, J. Am. Chem. Soc., 11L, 1891 (1952). (35) L. Pauling, "The Nature of the Chemical Bond," 2nd ed., Cornell University Press, Ithaca, N. 1., l9h8, p.111. (36) Alexander I. Popov and William A. Deskin, J. Am. Chem. Soc., Q0, 2976 (1958). (37) R. E. Buckles and J. F. Mills, ibid., 1.5.: 552 (1953). (38) J. Ronand, Ann. phys., g, 527 (1931.). (39) B. J. Mair and A. F. Forziati, J. Research Nat. Bur. Standards, 22, 151 (19hh1v ' (140) S. Wawzonek and M. E. Runner, J. Electrochem. Soc., 22, 1457 (1952). win! WY 1I1 16111111111111[(1111 :91 111 31293 01763 03