A. Photochemical Reactions Involving Tri-Iodine Ion. B. The Photochemical Reaction between Iodine Monochloride and Hydrogen Citation Ravitz, Sol Frederick (1931) A. Photochemical Reactions Involving Tri-Iodine Ion. B. The Photochemical Reaction between Iodine Monochloride and Hydrogen. Dissertation (Ph.D.), California Institute of Technology. doi:10.7907/xxeg-5247. https://resolver.caltech.edu/CaltechTHESIS:09092024-210528692 Abstract No abstract. Item Type: Thesis (Dissertation (Ph.D.)) Subject Keywords: (Chemistry) Degree Grantor: California Institute of Technology Division: Chemistry and Chemical Engineering Major Option: Chemistry Thesis Availability: Public (worldwide access) Research Advisor(s): Dickinson, Roscoe Gilkey Thesis Committee: Unknown, Unknown Defense Date: 1 January 1931 Record Number: CaltechTHESIS:09092024-210528692 Persistent URL: https://resolver.caltech.edu/CaltechTHESIS:09092024-210528692 DOI: 10.7907/xxeg-5247 Default Usage Policy: No commercial reproduction, distribution, display or performance rights in this work are provided. ID Code: 16706 Collection: CaltechTHESIS Deposited By: Benjamin Perez Deposited On: 10 Sep 2024 16:41 Last Modified: 10 Sep 2024 16:42 Thesis Files PDF - Final Version See Usage Policy. 19MB Repository Staff Only: item control page

A. PHOTOCHEMICAL REACTIONS INVOLVING TRI-IODIDE ION B. THE PHOTOCHEMICAL REACTION BETWEEN IODINE MONOCHLORIDE AND HYDROGEN Thesis by s. Frederick Ravitz In partial fulfillment of the requirements for the degree of Doctor of Philosophy California Institute of Technology Pasadena, California 1931 ACKNOWLEDGEi\[ENT The author wishes to acknowledge his gratitude and indebtedness to Dr. R.G.Dickinson, under whose guidance the following work was done. TABLE OF CONTENTS Page A. Photochemical Reactions Involving Tri-iodide Ion. I The Effect of Light on the FerrocyanideFerricyanide Iodine-Iodide Equilibrium. Introduction Apparatus Analytical Method The Nature of the Photochemical Reaction The Quantum Yield Discussion Summary II 9 The Photochemical Reduction of Ferric Iron in Tri-:i.odide Solution. Introduction Experimental Part Discussion Summary References B. l 2 2 3 5 8 10 10 14 16 17 The Photochemical Reaction between Iodine Monochloride and Hydrogen. Introduction Preliminary Experiments Preparation of Iodine Monochloride Method of Titration Preliminary Quantitative Experiments Final Photochemical Experiments Reaction Sys t em Optical System Experimental Procedure Experimental Results Quantum Yield Summary References 18 19 20 24 2'7 31 32 32 34 36 41 44 45 A. PHOTOCHEMICAL REACTIONS INVOLVING l'RI-IODIDE ION I. The Effect of Light on the Ferro cyanideFerricyanide Iodine-Iodide Equilibrium. [Reprint from the Journal of the American Chemical Society, 52, 4770 (1930).] (CONTRIBUTION FROM GATES CHEMICAL LABORATORY, CALIFORNIA INSTITUTE OF TECHNOLOGY, No. 269] THE EFFECT OF LIGHT ON THE FERROCYANIDEFERRICYANIDE IODINE- IODIDE EQUILIBRIUM BY ROSCOE G. DICKINSON AND S. FREDERICK RAVITZ RBCRIVBD AUGUST 20, 1930 PUBLISHED DRCBMBBR 18, 1930 It is well known that ferric salts oxidize iodides ih aqueous solution at a moderate rate and that ultimately an equilibrium, 2Fe+++ + 21-:::;:::'.: 2Fe + + + I2, is reached. Sasaki 1 investigated the effect of light on systems containing the substances involved in this equilibrium. Using sunlight and light from an incandescent lamp, he found that the illuminated systems always contained less titratable iodine than similar unilluminated systems. He further adduced evidence that the effect of the light was a displacement of the equilibrium rather than some other mode of disappearance of iodine. Further photochemical experiments were made by Rideal and Williams 2 who concluded, among other things, that the region of visible photoactive radiation was within the range 5500-6500 A., and that one quantum of absorbed radiant energy caused one molecule of iodine to react. However, 1 2 N. Sasaki, Z. anorg. Chem., 122, 61 (1922). E. K. Rideal and E. G. Williams, J. Chem. Soc., 127, 258 (1925). z. Dec., 1930 THE EFFECT OF LIGHT ON AN EQUILIBRIUM 4771 Kistiakowsky 3 extended these experiments and concluded that radiations of the wave lengths 5460, 4360 and 3660 A. were all effective; determinations of the quantum yield showed a disappearance of approximately one moiecule of iodine per quantum absorbed at each of these three wave lengths. In systems containing the ions Fe(CN) 6"' and Fe(CN) 6- - instead of the ions Fe+++ and Fe++, the thermal rates of reaction are again moderate, and directly measurable equilibria are again attained. The reaction may be expressed by the equation 2Fe(CN)s"' + 21- ~ 2Fe(CN)&-- + h Both the kinetics and the equilibria in these systems have been the subject of considerable study. 4 We have found that such systems are sensitive to light not absorbed by the complex cyanides. However, preliminary experiments showed that when light was directed into an equilibrium mixture containing the complex cyanides, free iodine was produced rather than consumed. The present paper describes experiments in which the light sensitivity of solutions containing the complex ions is compared with that of solutions containing the simple ions. Apparatus The source of illumination was a 500-watt lamp whose filament covered an area about 1.3 cm. square. The lamp was controlled with the aid of a variable resistance and an ammeter in series with it; the lamp current was maintained constant at 4.3 amperes. Light from this lamp passed through a 6% copper sulfate solution in a flat cell, 1 cm. thick; this filter absorbed the infra-red. This cell was fastened to the front of the thermostat, which was a large flat-sided glass jar, so as to leave a pocket 3 cm. in thickness between the cell and the jar. In this pocket was placed a second filter mounted on a sheet of brass with a rectangular opening; this filter served to isolate the desired spectral region. The pocket was filled with water to minimize reflection losses. The reaction cell which contained the solution to be illuminated was a flatsided bottle, the face of which was 4.5 X 10 cm., and the thickness 2.86 cm. This reaction cell was silvered externally except on the face toward the lamp. The thermostat with its attached filters was contained in a box provided with a shutter and thermal insulation. Arrangement was made so that the lamp and cell could, from one experiment to another, be brought into the same positions. The experiments were carried out at 0°; to permit this, the thermostat was kept well packed with clean ice except in front of the reaction cell, which was close to the thermostat wall. Water was circulated from bottom to top of the thermostat by an air lift. Ice was ·also kept in the upper part of the filter pocket. Analytical Method The solutions to be analyzed, whether ferrous-ferric iodine-iodide mixtures or ferrocyanide-ferricyanide iodine-iodide mixtures, were ordinarily not greatly removed 3 G. B. Kistiakowsky, Tms JOURNAL, 49, 976 (1927). F. G. Donnan and R. Le Rossignol, J., Chem. Soc., 83, 703 (1903); G. Just, Z. physik. Chem., 63,513 (1908); C. Wagner, ibid., 113,261 (1924); V. K. La Mer and K. Sandved, Tms JOURNAL, 50, 2656 (1928); V. K. La Mer and H. B. Friedman, ibid., 52, 876 (1930). • 4 3. 4772 ROSCOE G. DICKINSON AND S. FREDERICK RAVITZ Vol. 52 from equilibrium. They were analyzed for free iodine by the following method: from a 0.1 N thiosulfate solution, a solution of convenient strength, usually 0.004 N, was freshly prepared by dilution. Five cc. of this solution was pipetted into 50 cc. of water and starch was added. A sample of the solution to be analyzed was withdrawn from the reaction cell into a 10-cc. graduated pipet. The solution was then immediately run from this pipet into the thiosulfate solution to the appearance of a blue color, permanent on stirring. The elapsed time between the withdrawal of the sample and the attainment of the end-point was less than two minutes, and most of the solution analyzed was discharged into the thiosulfate in considerably less time. The suitability of this method for the titration of solutions containing ferrocyanide and ferricyanide has been discussed by La Mer and Friedman and by La Mer and Sandved.6 The Nature of the Photochemical Reaction As mentioned above, when light is directed into a ferrocyanide-ferricyanide iodide-iodide equilibrium mixture, the concentration of titratable iodine increases. We have supposed that this increase in iodine concentration arises from a displacement of the equilibrium, i. e., from the oxidation of iodide by ferricyanide. If this be the case then a solution which has been removed from equilibrium by shining in light should, when placed in the dark, return to equilibrium at the same rate as a solution of the same composition which has been equally removed from equilibrium by some other method, for example, by the dilution of a more concentrated equilibrium solution. We have tested this by the following experiment, which was carried out in an atmosphere of nitrogen in order to eliminate possible oxidation by dissolved air. Tank nitrogen which had been bubbled through two bottles of strong alkaline pyrogallol solution and one bottle of water was used. All flasks used were first filled with nitrogen by filling them with water and displacing the water with nitrogen. Solutions were made up of distilled water which had been boiled for several minutes and allowed to cool with a fairly strong stream of nitrogen passing through it; the solutions were transferred from one flask to another by the use of nitrogen pressure. Sixteen grams of c. P. potassium ferricyanide was dissolved in about 100 cc. of water, filtered into a 500-cc. volumetric flask, and a vigorous stream of nitrogen was passed through the solution for several minutes; 20.00 g. of c. P. potassium iodide was likewise dissolved in about 100 cc. of water and nitrogen was passed through this solution for several minutes. It was then transferred to the volumetric flask and water was added to the 500-cc. mark. The resulting mixture was transferred to an Erlenmeyer flask and nitrogen was passed through it for about ten minutes, during which time very little iodine was formed. The mixture was then carefully packed with ice and allowed to stand for several days at 0 0 until equilibrium was practically attained. To 250 cc. of the resulting equilibrium solution at 0° an equal volume of oxygenfree water (previously cooled to 0 0 ) was added. The flask containing the resulting solution was packed in ice water and allowed to stand in the dark. Samples of this diluted solution were withdrawn at intervals by forcing the solution into the 10-cc. graduated pipet with nitrogen pressure, and were titrated by the method described above. The results of these titrations are plotted as circles in Fig. 1. 6 V. K. La Mer and H.B. Friedman, Tms JOURNAL, 52,877 (1930); V. K. La Mer and K. Sandved, ibid., SO, 2659 (1928). 4. Dec., 1930 THE EFFECT OF LIGHT ON AN EQUILIBRIUM 4773 After about twenty-four hours, this diluted solution had nearly reached equilibrium; 160 cc. of it was then transferred to the reaction vessel in the ice thermostat and illuminated for three hours through the copper sulfate filter and a 1-cm. layer of 16% potassium chromate solution. Immediately upon ceasing the illumination the solution was placed in the dark and samples were withdrawn at intervals and titrated as before. The results of this series of titrations are shown as crosses in Fig. 1; the time origin for the plot relating to the illuminated solution was so chosen that the first point (highest iodine concentration) fell on the curve relating to the unilluminated solution. If the effect of light was simply the displacement of equilibrium, then the concentrations of all materials should be the same in the two solutions at this point. It may be seen from Fig. 1 o.oosr-----~---"T----...,...-----,------, 0.0071----,\-----+----+------+------+------l 200 400 600 800 1000 Time, minutes. Fig. !.-Comparison of change of iodine concentration with the time for two solutions of the same composition, one (circles) having been removed from equilibrium by dilution, the other (crosses) having been illuminated. The curve is drawn through the circles. that these two solutions returned toward equilibrium at the same rate at all times. Moreover, the final iodine concentration arrived at was the same in each case; twenty-nine hours after the beginning of each experiment that for the illuminated solution had become 0.00396 and that for the unilluminated solution, 0.00398 N. When similar experiments were carried out without the precaution of excluding air, the coincidence of the two curves was not so complete and was especially incomplete if the illumination was a prolonged one; a solution which had been given twenty hours' illumination finally came in the dark to an equilibrium concentration of iodine about 12% lower than that of a corresponding unilluminated solution. 4774 ROSCOE G. DICKINSON AND S. FREDERICK RAVITZ Vol. 52 In the experiments of Fig. 1, the concentration of the iodine before illumination was 0.00398 N; the equivalent concentration ·of the ferrocyanide (since the solution was made up from ferricyanide and iodide) was the same. During illumination, the iodine concentration.rose to 0.00636 N, an increase of 60%. If this increase were not due to a shift in the equilibrium, the illuminated solution would then not be expected to contain the same amount of ferrocyanide as the unilluminated solution of equal iodine concentration prepared by dilution, and an equality in the rate of return to equilibrium would not be expected, nor should the same equilibrium concentration of iodine be reached in the two cases. We take the observed equality as evidence that the photochemical reaction with the radiation used is a reduction of ferricyanide to ferrocyanide by iodide. Further ·evidence that this is the case is furnished by the following experiment in which the formation of ferrocyanide on illumination is shown directly. Two solutions, the first 0.008 Nin iodine and 0.016 formal in potassium iodide, and the other 0.010 M in potassium ferricyanide, were made up with no effort to exclude air, and brought to 0°. Equal volumes of the two solutions were then mixed, and half the resulting solution was immediately transferred to a reaction vessel placed in the ice thermostat. Illumination with the 500-watt lamp (through the copper sulfate and potassium chromate filters as above) was started at once. At intervals 10-cc. samples of the illuminated and of the unilluminated solutions were withdrawn simultaneously and pipetted into large test-tubes, each containing 50 cc. of water at 0 O to which had been added 2 cc. of a 2% solution of ferric ammonium sulfate in 0.012 N sulfuric acid. After :five minutes' illumination the test with the illuminated solution gave a decided blue color of ferric ferrocyanide, whereas the test with the unilluminated solution showed no visible trace of blue. After ten minutes the results were similar, the color obtained with the illuminated solution being somewhat stronger than before. After twenty minutes the illuminated solution gave a still stronger color, while the unilluminated solution now gave a faint blue color resulting from the formation of ferrocyanide in the comparatively slow thermal reaction. The Quantum Yield A series of experiments was made in which the photochemical rate of production of iodine in ferrocyanide-ferricyanide mixtures was compared with the rate of disappearance of iodine in ferrous-ferric mixtures. In carrying this out, a pair of solutions, one of each kind, was made up and the solutions allowed to come substantially to equilibrium at 0°; the concentrations were so chosen that the titratable iodine content was about the same in each. The ferrocyanide-ferricyanide solutions were all made up as described in the previous section, by starting with 16.00 g. of potassium ferricyanide and 20.00 g. of potassium iodide; in different experiments, the extent of dilution was varied. The ferrous-ferric solutions were made up by dissolving 4.82 g. of ferrous ammonium sulfate and 0.66 g. of ammonium sulfate in water containing 4.2 cc. of 6 N sulfuric acid; to this was added a solution of 3.74 g. of potassium iodide and the resulting solution diluted so as to give the desired free iodine concentration. The solutions were successively exposed in the same reaction cell under the same conditions. During each exposure samples with withdrawn from time to time and titrated. Dec., 1930 THE EFFECT OF LIGHT ON AN EQUILIBRIUM 4775 The results of two typical pairs of such experiments are shown plotted in Fig. 2. The abscissas are the time of exposure in minutes and the ordinates are the concentration of titratable iodine in millimoles per liter. Only at the beginning of each exposure can the rate of change of iodine concentration be considered to be due to light alone; when sufficient departure from thermal equilibrium has been attained, the thermal reverse reaction should lower the rate of change of iodine concentration. We have sought to decrease this difficulty by working at 0°. With exposures of one hour, in our apparatus, no appreciable change in rate was observable; consequently, as in Fig. 2, a straight line has been taken as the best representation of the data. The results of all pairs of experiments are summarized in Table I. The concentrations, given throughout in millimoles per liter, apply to the beginning of exposure. In Col. 2 is given the concentration of titratable iodine; this is equal (since the solutions were made up from iodide and ferricyanide or ferric iron) to one-half the millimolal concentration of ferrocyanide or ferrous iron as the case may 1.0.___ _ _..___ _ _~_ ____, be. The total iodide, given in Col. 3, 0 20 40 60 is the iodide taken minus the milliTime, minutes. equivalents of titratable iodine. Fig. 2.-Change of iodine concentration The concentrations of ferricyanide with the time in ferrocyanide-ferricyanide or ferric iron in Cols. 4 and 5 are solutions (circles) and in ferrous-ferric solutions (crosses) under the same illumination. similarly obtained from the titratable iodine and the amounts of ferricyanide or ferric iron taken. In Col. 6 is given the fraction of the titratable iodine existing as I 3 -. This was obtained using the equilibrium constant (I-) X (I 2)/(I3-) = 0.000686, which is the value obtained by Jones and Hartmann6 in 0.1 N potassium iodide. Although the solutions involved here are by no means perfect, the calculations suffice to show that most of the titratable iodine was in the form of I,-. The potassium chromate filter used in addition to the copper sulfate e G. Jones and M. L. Hartmann, Tms JOURNAL, 37, 241 (1915), 7. 4776 ROSCOE G. DICKINSON AND S. FREDERICK RAVITZ Vol. 52 TABLE I RELATIVE RATES OF PRODUCTION OF IODINE IN FERROUS-FERRIC AND IN FERROCYANIDEFERRICYANIDE SOLUTIONS No. Concentrations in millimoles per liter Titratable iodine = '/,(Fe(CN)o•·) Total (Fc(CN),•) (Fe+++) or '/,(Fe++) iodide +0.0108 1.03 .0105 44 .9 + .0111 1.03 1.51 .0108 33.7 + .0085 1.08 2.26 .0079 38.1 4a + .0086 1.07 4b 1.99 .0080 82.2 5a + .0152 1.00 4.04 .0151 5b 82 . 1 6a + .0152 1.07 4. 13 .0142 6b 45 .0 7a + .0008 0.16 7ba 1.08 .0050 a Air was not excluded. Filters: K2CrO, and CuSO, in Expts. 1-6 inclusive; methyl violet, Cu(NO,h and CuSO, in Expt. 7. la0 lb 2a 2b 3a 3b 0 0 0 0 1. 71 1. 71 1.87 1.78 1.34 1.46 1.53 1.59 7.53 7.98 7.59 7.93 1.86 1.96 117.2 7.10 116.8 7.64 87.6 8.52 98 .9 8.25 225.9 29 .04 225.8 29 . 13 116 .9 5.58 45.3 Rate of production of titratable (I,-) iodine, milliRatio of (I,)+ (I,•J moles/liter-min. rates 1.25 0.995 .889 .995 .899 .993 .912 .994 .906 .997 .969 .997 .968 .995 .893 filter in Expts. 1-6 inclusive was a 1-cm. layer of 16% solution. Examination with a spectroscope showed that this filter removed substantially all radiation below 5000 A. None of the materials involved in the experiments except 13 - and 12 gives appreciable absorption of visual radiation of wave lengths longer than 5000 A. when present in the amounts used. Consequently all of the light absorption is attributed to Ia - and 12. The total light absorption in any ferrocyanide- ferricyanide solution must have been nearly the same as in the companion ferrous-ferric solution, for the proportion of titratable iodine in the form 12 was not large even in these latter cases. Moreover, unpublished measurements made by Mr. C. F . Carlson in this Laboratory have shown that the absorption coefficient of I~ is the same as that of Ia- at about 5500 A., and that the absorption coefficients do not change rapidly in this neighborhood. In Col. 8 of Table I are given the ratios of the rates of appearance of iodine in the ferricyanide solutions to the rates of disappearance in the companion ferric solutions. For Expts. 1-6 inclusive, where the light absorption was due simply to Ia- (and I2), the ratios are within a few per cent. of unity in all cases, i.e., the quantum yield of the two reactions is substantially the same. If we may assume that the quantum yield in the ferrous- ferric solutions is the same at 0° as in Kistiakowsky's quantum yield determinations, we are led to conclude that in the ferricyanide reaction, one molecule of iodine is formed per quantum absorbed by 13 - • Dec., 1930 THE EFFECT OF LIGHT ON AN EQUILIBRIUM 4777 Experiment 7 was made using filters of methyl violet and copper nitrate solutions; the combination absorbed strongly above 5000 A. and transmitted freely below 4700 A. It accordingly transmitted radiation which was absorbable by ferricyanide as well as by la-. In this case the rate in the ferricyanide solution fell considerably below that in the ferric solution. Apparently, then, absorption by ferricyanide leads to little, if any, production of iodine. Discussion In the present photochemical experiments resulting in the reduction of ferricyanide, the absorbing species was almost exclusively Ia-. The formation of a theory of this reaction is hampered by lack of knowledge of the elementary act accompanying the light absorption. In reviewing the photochemical ferrous-ion oxidation, Kistiakowsky 7 considered two reaction mechanisms to be possible (I) (II) 1,- + hv - + 1- + 21 l + Fe++ - + 1- + Fe+++ Ia - + hv - + Ia - (activated) 11 - (activated) +Fe++-+ 21- +Fe++++ I I +Fe++-+ Fe+++ + 1- Neither elementary act here provides very naturally for the reduction of 2Fe(CN)6"", but either can be made to do so by supplementing with the reaction I+ 1--+ I2- and supposing that 12- (as well as Ia- (activated) in II) can act as a reducing agent. The hypothetical ion 12- has been employed by Wagner 8 to account for the inhibiting effect of the products in the thermal reactions 2Fe+++ + 21--+ 2Fe++ + 11 and 2Fe(CN)6• + 21--+ 2Fe(CN)s-- + Ia and for this purpose has been attributed both oxidizing and reducing properties. If it be introduced here, the following mechanisms of ferricyanide reduction result (I) (II) Ia - + hv - + 1- + 21 I+ 1---►. I2I2 - + Fe(CN)a• - + Is + Fe(CN)a-11 - + hv-+ 11 - (activated) 1,- (activated) + Fe(CN)a•-+ 12 + I + Fe(CN)s-1 + 1---►. 1212- + Fe(CN)a•-+ 12 + Fe(CN)s-- To these may be added a mechanism based on a third elementary act (III) 1,- + h11-+ 12 - + I I+ 1--+ 1112- + Fe(CN)••-+ 12 + Fe(CN)s-- 7 G. B. Kistiakowsky, "Photochemical Processes," Chemical Catalog Co., Inc., New York, 1928, p. 65. • C. Wagner, Z. physik. Chem., 113, 261 (1924). 9. 4778 ROSCOE G. DICKINSON AND S . FREDERICK RAVITZ Vol. 52 Any of these mechanisms is readily reconciled with our present knowledge of the photochemical ferrous-ion oxidation. In an investigation of the photochemical reaction between potassium oxalate and iodine in iodide solution, Berthoud and Bellenot9 developed a mechanism of that reaction which led to their empirical kinetic expression; this mechanism supposed that light absorbed by 12 (producing iodine atoms) was effective while light absorbed by Ia- was ineffective. This mechanism would then rule out any elementary act in which iodine atoms are produced from Ia-; this would eliminate I and III above. However, we are at present disinclined to accept this conclusion; for it is possible to set up a mechanism leading to the kinetic expression of Berthoud and Bellenot without assuming light absorbed by Ia - to be ineffective. 10 In conclusion we may say that whatever the elementary act accompanying light absorption by Ia- may be, it is clear that in it or in the subsequent reactions both oxidizing and reducing configurations are produced which are not to a great extent present in the unilluminated solution. The tentative assumption of the ion 12- is a simple hypothesis which satisfies these requirements. Summary It is found that when a ferrocyanide-ferricyanide iodine-iodide solution is illuminated with light absorbable by Ia- but not by the complex cyanides, reduction of ferricyanide and oxidation of iodide occurs. The quantum yield is found to be substantially the same as that in the oxidation of ferrous ion in iodine-iodide solutions. Some possible mechanisms of the re~ action are discussed. PASADENA, CALIFORNIA 9 A. Berthoud and H. Bellenot, J. chim. phys., 21, 308 (1924); Helv. Chim. Acta, 7, 307 (1924). 10 This is accomplished, for example, by the following mechanism which is similar to that of Berthoud and Bellenot but differs from it in two important respects: (1) light absorbed by Is- is effective; (2) the inhibiting effect of r- is attained through the formation of Ia- from r- and I. The reactions are: hv + Ia~ I2- + I (or r- + 21) (1) I+ 40,- ~ C2O,- + r(2) r,- + C2O,- ~ 2co2 + r2- + r(3) or I2 + C2O,- ~ 2COi + I2I2- + Ia- ~ r,- + r(4) r+r~r,(5) If the equilibrium (5) is assumed to be rapidly established, the usual treatment leads to the expression _ d(40r) = k v ✓~ 1 (I -) (~o,-) dt 2''-1 k, 0 I er-) for the case of low absorption, which agrees with that of Berthoud and Bellenot. A. PHOTOCHEM ICAL REACTIONS INVOLVING TRI-IODIDE ION II. The Photochemical Reduction of Ferric Iron in Tri-iodide Solution. 10. INTRODUCTION. Investigations of the effect of light o~ the equilibrii.u/ 1 ) :l- 'F~" ....... -+ Z -:r.- ~ .l Fe.._."" -+ rt.- have shown tha.t visible or near ultra violet displaces the equilibriwn to the left and does so at such a rate that two equivalents of' titra.table iodine a.re consumed per einstein absorbed. In a recent paper( 2 ) the authors described the effect of light on ferrocyanideferricyanide iodine-iodide equilibrium solutions in which the equilibrium is so displaced that t.vo equivalents of iodine are fonned per einstein absorbed. This result showed that absorption of light by tri-iodide ion leads to configurations having both oxidizing and reducing properties, and suggested that under suitable chosen conditions photochemical reduction of ferric ion might be observed. The present paper describes experiments in which this result was obtained. EXPERIMENTAL PART. The choice of experimental conditions was governed by the following considerations. The concentrations of all materials involved should be kept so small as convenient in order to reduce the rates of the therm.al reactions; the thennal ree.ctions would be further suppressed by working at 0°. The concentration of ferrous iron should be small compared with that of ferric; otherwise, the usual observation of a net photochemical oxidation of ferrous iron would be expected. The concentration of titratable iodine should, however, not be too SI:1al1 since this material furnishes the lightabsorbing constituent. adopted: The following procedure vras accordingly A dilute solution of ferric ru~~onium alum~ acidified with HzS0 4 was mixed with a dilute solution of iodine in potassium iodide, both solutions having been previously cooled to 0 °. The resulting I I l\. solution wu.s immediately separated into two parts., one of which was placed in the dark and the other illuminated in an ice thermostat as previously described( 2 ). The light used was that from a 500- w-att projection lamp filtered through 1 cm. of 16% ~Cr04 and 1 cm. of 6% Cu.S04 • The change in iodine concentration in each solution was follmved by withdrawing., from time to time, portions which were titrated against freshly prepared 0.004 n. thiosulfate. In s~file experiments the solutions were made up and handled under an atmosphere of nitrogen; this., hcnvever., did not alter the results. In Fig. 1 are shown the results of a typical experiment in which oxygen was excluded. The concentrations of iodine in the illumina.ted(l) and unilluminated(2) solutions are plotted against the tL~e. It may be seen that the iodine concentration increased more rapidly at first in the illuminated than in the unilluminated solution, but subsequently did so less rapidly when suf'ficient ferrous ion had been formed to make its rate of photochemical oxidation appreciable. All experiments of this type yielded sL~ilar results., and we have attributed them to a photochemical oxidation of iodide by ferric iron., r; being the light absorber. However., in view of the fact that the differences between the illuminated and unilluminated solutions were not large, we have sought oonfirmation of this oonclusion in an attempt to detect more rapid development of ferrous ion in the illuminated solution. For this purpose a solution was made up and divided as described above. At intervals., 10 oc. sa.~ples of the illuminated and unillumL"'lated solutions were simultaneously withdrawn and pipetted into large test tubes# each containing 40 cc. of water at 0° to whioh had been added 2 cc. of 2r; potassium. ferricye.nide. Throughout \3 . the experiment (50 minutes) the illuminated solution produced a decidedly stronger blue color of ferrous ferricyanide than did the unilluminated. This experiment, shmving the photochemical formation of ferrous iron, is then in agreement with those in which the iodine was followed and seems hardly to be open to the possibility that the result was produced by traces of oxygen. That the more rapid reaction in the illuminated solution was not due simply to a higher temperature resulting from the illumination was shown by an experiment in which the unilluminated solution was purposely kept at a higher temperature (5°) than the illuminated (a thermometer in the illuminated solution read below o.5° throughout). The results of this experiment were sL~ilar to those of the preceding. Finally an experiment was performed to determine whether ferrous iron and iodine were found in equivalent amounts by the illumination. A solution was made up as i n the preceding experiments and the iodine concentrations in both the illuminated and unilluminated portions followed by wi thdr~1.wing samples for titration. Alternating with the thiosulfate ti trations, 10cc. samples were simultaneously withdra~m from the illuminated and dark portions and pipetted into test tubes containing potassium ferricyanide as described above. From a burette, 0.001 n. ferrous ammonium sulfate solution wa s :i.rri.media.tely titrated into the test tube to which had been added the unilluminated sarn.ple until the color matched that produced by the illuminated sample. From this, the differencec in ferrous iron concentration in the t,.~o solutions were calculated; by plotting the results of the thiosulfate ti trations, the dif'ferences in iodine concentrations at the same t:i.rnes were found. The results are shown in Table I. Table I. Time of illuru•ination •(min. ) Dii'. in iod.ine cone. Dil'. in ferrous cone. (milli-equiv. /liter) (milli-equiv./liter) 14 0.075 o.os 23 .12 .18 38 .15 .24 53 • 17 It may be noted that (1) the difi'erences betvreen the iodine titrations were small (about 0.2 cc.)., (2) a color match was depended on for the ferrous detenn.inations., and (3) the .conditions under which the collodial ferrous ferricyanide was fonned were somewhat different in the two solutions matched. In vievv- of these facts., the agreement is a.s good as could be expected and shows that., at least as to order of magnitude, the iodine and ferrous iron a.re produced on illumination in equivalent a.mounts. We have also attempted to show a photochemical reduction of iodine by ferrocyenide by illuminating a very dilute solution of potassium iodide., iodine and potassium ferrocyanide. attempts have not been successful. However, the The thermal rates in these solutions were so fast as to mask any effects of the order of magnitude of those described above. DISCUSSION. In general, when a solution containing ferrous e.nd ferric iron., iodine and iodide is under illumination., four reactions may be presurned to be occurring, n.runely., a. photochemical and a thermal oxidation of ferrous iron and a photochemical and a thermal reduction of ferric iron. The two thennal reactions are well knovm (l) , a s is also the photochemical oxidation of ferrous iron (I} . The photochemical reduction of ferric iron, demonstrated in the present paper, has not previously been found. The existence of this reaction is, however, not in disagreement with previous experimental work., for the specific rate of the reaction is apparently much lower than that of the photochemice.l oxidation of ferrous iron. Sasaki illuminated solutions containing appreciable amounts of ferrous iron at the start and consequently always found a net disappearance of iodine. S:L'Tlilarly, the present results probably do not create any difficulty with Kistiakowski's quantum yield determinations; for in these the ratio of ferro~s to ferric iron present was sufficiently large so that the photochemical reduction could hardly have run.oUJ,ted to more than a fm~ the oxide.tion. per ceny of 16, SUMM.ARY. It has been found that when solutions containing iodine., ferric iron, and iodide ion are illuminated immediately af'ter being prepared, iodine and ferrous iron in approximately equivalent amounts, are at first fonned more rapidly than in similar unilluminated solutions., indicating that the effect of light is to increase the rate of reduction of ferric iron by iodide, After a short time, however., the illuminated solutions reach a steady state, owing to the fact that the photochemical oxidation by iodine of the ferrous iron formed becomes e.pprecia.ble. Attempts to observe a. greater rate of oxidation of ferrocyanide by iodine in illuminated solutions containing ferrocyanide ion, iodide ion., e.nd iodine., than in corresponding unilluminated solutions., were unsuccessful. The thermal rates, hovrever, were so rapid that they would have masked any effects of the order of magnitude of those observed in the photochemical reduction of ferric iron by iodide. REFERENCES 1. (a.) Sa.ski., Zeit.anorg.Chem • ., 122., 61 (1922); (b) Rideal and Williams., J.Chem.Soc . ., 127., 258 (1925); (c) Kistiakowski., J . .f\m.Chem.Soc . ., ~• 976 (1927). 2. Dickinson and Ravitz., J • .Am.Chem.Soc . ., 52., 4270 (1930). (Pa.rt A - I of this Thesis) 1' " f or examp 1 e, ',vagner., Zei• .,_"• p h ys. Ch em. , 11'.(., 261 (1Cl24). v _ 3 • ,;;,ee., B. THE PHOTOCHEMICAL REACTION BETVV°EEN IODINE MONOCHLORIDE AND HYDROGEN IN'.lRODUCTION The photochemical reactions between hydrogen and the halogens have been the object of considerable study. At the time this work was started no investigation had been reported of the behavior of hydrogen with a mixed halogen, such as iodine monochloride, under the influence of light, so it was thought of interest to study that behavior. Also it was hoped that this might give some information regarding the mechanism of the hydrogen-chlorine reaction. Rollefson(l)gave reasons for believ- ing that only excited chlorine atoms can start the latter reaction, the normal atoms being without effect. Unfortunately, it is not yet definitely knovm whether iodine monochloride dissociates optically into two normal atoms or into a normal iodine atom and an excited chlorine atom. Gibson and Ramsperger,( 2 ) who investi- gated the absorption spectrum. of ICl, at first interpreted their results as indicating dissociation into norm.al atoms, but later Gibson( 3 ) decided that dissociation into a normal iodine atom and an excited chlorine atom was more probably the case. Shortly after the present work had been started and a definite photochemical reaction found, Rollefson and Lindquist< 4 ) reported that "it is not possible to bring about the fonna.tion of hydrogen chloride by the opticnl dissociation of iodine monochloride in hydrogen~ either in the presence or absence of chlorine molecules". However., Mellor e.nd Iredale., ( 5 ) and also Ashley and V/est, ( 6 ) 19. shortly afteri.vards stated that they observed a fairly rapid photo-chemical reaction between ICl and hydrogen. To date, though, no quantitative infonna.tion concerning this reaction has appeared in the literature. PRELIMINARY EXPERIMENTS. In order to determine whether iodine monochloride and hydrogen would react under the influence of light some simple qualitative experiments were made vrith some ICl furnished by Dr. D. M. Yost. A small glass bulb was filled with I Cl in the manner in which Victor Meyer bulbs are usually filled, and placed, along with a heavy piece of glass rod, in a large cylindrical glass tube. The latter was filled with hydrogen at atmospheric pressure and then sealed off at both ends; by shaking the ttlbe the glass rod was made to break the bulb of ICl. The tube, which was not quite saturated with I Cl, was placed in a water thermostat and illwninated at room temperature with a 500-watt projection lamp. Within five minutes small crystals of iodine c.o uld be seen separating out on the walls of the flasks. Upon breaking the tube open after the illumination, white fumes appeared, indicating the presence of hydrogen chloride or hydrogen iodide. Several such experiments were made with different amounts of ICl and at different temperatures~ but all gave substantially the same results. Also mixtures of ICl and hydrogen kept in the dark showed no signs of' reaction even after a week. z,o. Two possible reactions between ICl and hydrogen suggest themselves: I. J½ + 2 ICl = 2 HCl + I 2 2. Hz+ ICl = HCl + HI . l However, experiments made by Dr. H. C. Ramsperger of this ~aboratory shovired that HI and ICl react to .form hydrogen chloride and iodine so rapidly that the rate cannot be followed. Hence, if any HI is formed it is used up immediately, so equation l would represent the over-all reaction. The possibility of photochemical reaction between hydrogen and the iodine formed is eliminated by the fa.ct that Coehn and Stuckardt( 7 ) and Bowen(B) found that even at temperatures as high as 270° illumination of mixtures o.f hydrogen and iodine in glass vessels produced no reaction. PREPARATION OF IODIHE MONOCHLORIDE Iodine monochloride is usually prepared by passing chlorine over iodine until the solid has completely liquified, and then fractionally distilling the product. This method, however, is open to serious objection since it is knovm that the vapors of ICl are richer in chlorine than is the solid, giving rise to a product containing more than the stoichiometric proportion of chlorine.( 9 ) To overcome this difficulty, the ICl used in the experiments to follow was prepared by directly combining weighed quantities of chlorine and iodine. Purification of Iodine. The best availabl e commercial iodine, the manui'acturer 1 s analysis of which gave the iodine content as 99. 97%, 1,rns used. This was purified by mixing intimately 50 grams of it with 15 grams of c.p. potassium iodide and 15 grams of c.p. calcium oxide, and subliming the iodine from the mixture at a temperature never over 75°C. The first and last portions of the sublimate were discarded, and the remainder 1 about 40 grams, was resublimed in vacuo at temperatures below 70° through a series of U-tubes. Purification of Chlorine. Tank chlorine was purified by passing it through the follovn.ng all-glass chain: (1) a wash bottle, 16 cm. deep, filled with glass beads and containing a satu1·ated solution of potassium permanganate; (2) a similar wash bottle containing distilled water; (3) another similar wash bottle containing concentrated sulruric acid; (4) au-tube filled with glass-1vool; and (5) a long tube filled with dessichlora. Preparation of ~ Iodine Monochloride. A weighed glass bulb was filled with about 3.5 cc. of chlorine as follows (see fig. 1): Chlorine from the purifying chain passed through the small glass tube Finto the weighed bulb G, where it was liquified with liquid air. Care vras taken to keep the end of tube F above the surface of the liquid chlorine, and to keep the chlorine flowing fast enough so that it issued out through the calcium chloride tube at all times. The latter tube served to ""e r oAbb trC:.o c.ti ons. tight- f itti"j a., b,c - ➔ Clz. -- - p ,b Ca.Cl, tube r:--w--a~ F ,3 . i. <,; , -~Ill F I (. I I I I I , Le I ( I l ' 'I I . I I I I I I I A ' 6 - ., ; 1 ; I e: F,-9. Z. I I J ' I ,_ , , I I I I I I I I I I I I I I I I I I I ·, '1 I\ ~ I I t . ~ ;\ ~ ,,:, I I 'I I I D ~ I ( ~~~-,Flu tll ?-l prevent the condensation of atmospheric moisture in G. When the desired amount of chlorine had been liquified, the end of tube F was dravm up to a point above P, at which point the neck of bulb G was sealed off to a fairly long capillary. The bulb of liquid chlorine was allowed to come to room temperature, and was then weighed, together with the sealed off neck. The difference between this weight and the original weight of the bulb gave the weight of the chlorine. This was 4.8434 grams. The weight of iodine stoichiometrically equivalent to 4.8434 grams of chlorine is 17.3357 grams. To allow for evapora- tion, 10 mg. more than this weight of the purified iodine was weighed out into a glass tube closed at one end. For the combination of the iodine and the chlorine, the reaction chamber shown in figure 2 was made. The bulb of chlorine " \ G.c.t.Cl was,..in arm A, the tube of iodine in arm B, and the tops of the two arms were sealed off. The system was then pumped down as rapidly as possible and the reaction chamber sealed off from the pumps. The arm containing the chlorine was immersed in a carbon dioxide-ether freezing mixture until the chlorine was frozen, and then by moderately shaking the apparatus, the heavy glass hammer C was ma.de to break the tip of the chlorine bulb. The chlorine was allowed to evaporate slowly, and when it had all evaporated, the mixture of iodine and chlorine was repeatedly distilled from one portion of the apparatus to another to insure complete reaction. Finally every trace of the mixture was condensed in the container D, which was then sealed off from the rest of the system. The ICl container was so designed that small quantities of the liquid could be transfered without contamination. done in the following manner: This was The ends of the capillary tips E and F were broken off and the section G (which was ground to fit at H) was attached. By blowing (through a calcium chloride tube) at E, a few drops of ICl were forced into G, through which the liquid passed into a weighed length of 1.5 mm. tubing, the closed end of which was dra.vm to an easily breakable capillary. 111/hen filled with the desired amount of ICl, the open end of the small tube was sealed off and the tube of ICl was weighed together with its sealed-off piece, After the desired number of tubes had been filled, the tips E and F of the container were again sealed off. METHOD 01'' TITRATION. Since the photochemical reaction is 2161 +Biz= 2HC1 + I 2 , it can be followed by determination of total free halogen present (i.e . ., ICl + 12 ) and of acid. This i,vas done by (a) dissolving the sample in potassium iodide solution and titrating to an endpoint with thiosulfate solu-f:;ion, then (b) adding potassium iodate and titrating the iodine thus liberated with more thiosulfate. The reactions involved are: {a) ICl +KI= I 2 + KCl {b) 6 H+ + I0 3 - + 5 r- = 3~ + ~ H,,aO , hence ea.ch equiw,lent of ICl and ea.ch equivalent of acid gives one equivalent of titratable iodine. From the equation for the photochemical reaction, it is evident that for every two equivalents of ICl used up, one of acid and one of iodine are fonned, so that the total number of equivalents of thiosulfate used for both the free halogen and the acid should be equal to the number of equivalents of ICl originally present. If 11 a 11 represents the number of cos. of thiosulfate solution used for the titration of free halogen., and 11 b 11 the number used for the acid, then 2b is the '"'e.-n fraction of ICl which has reacted. This method of analysis was tested as follovrs: a number of tubes were filled with ICl as described in the preceding section., but were made with brealcable tips at each end. A file scratch vras made across the center of' each tube., and it was then placed in a flask containing 15 cc. of 30% KI solution. The tube was broken at the file scratch by placing it across a glass rod and striking it sharply with another glass rod; the tips were also broken with a glass rod. 0.1039 n. thiosulfate solution was run in until an endpoint was reached with starch solution as an indicator, then 5cc. of 3% potassium iodate solution were added and more thiosulfate solution was run in until an endpoint was reached again. In some experiments definite amounts of standard HCl solution were added before titrating. Blanks made on the potassiurn iodate showed that neither contained any acid or any free halogen. The results are shovm in table I. It will be noted that in experiments 2., 3., and 4 some acid was found., although none was added, and that the percentages of ICl are somewhat lovr, In these experiments the broken off. tips retained a little of the ICl so that it could not react with the KI, and due to the deep color of the solution, this was not noticed until several ccs. of thiosulf~te solution had been added. The Table I. Expt. Wt. ICl No. used ICl found at_ wt. /0 Vol. acid added 1 0.2421 o. 2417 99.83 none none 2 .1821 .1803 99.00 ti ti 0.1a 3 .2411 .2362 98.00 II tt • 70 4 .2021 .2012 99.55 " " .01 5 .1799 .1801 100.07 tt tt none 6 .1865 .1868 100.14 ti " II 7 .2054 .2050 99.84 10.00 9.23 9.23 100.00 8 .2007 .1999 99.61 15.00 13.85 13.83 99.87 Vol. thio for a.oid. Calo. Found % none small capillary tips were then crushed with a glass rod and the titration continued. Since it is lmown that chlorine oxidizes tetrathionate to sul:f'ate(lO), these results are probably due to a s~ilar oxidation by ICl of the tetrathiona.te produced by the reaction between iodine and thiosulfate, according t o the equation Evidently for every ti.venty-eight equivalents of ICl reacting in this manner, only fourteen will appear in the free halogen titration; the other fourteen vrill appear as tt,renty equivalents of adid. Hence, if xis the number of equivalents of thiosulfate used for the first endpoint, and y the number used for the acid, then if no acid were initially present, the actual number of equivalents of ICl is X -r, 14 y :m . Making this correction gave the Z7. follovrlng results*: Table Ia. Expt. No. wt. ICl used. ICl found (uncorr.) wt. % 2. o.1a21 o.1a03 99.00 0.1813 99.60 3. . 2411 .2362 98.00 .2404 99. 72 4. .2021 .2012 99.55 .2013 99. 59 ICl found (corrected) Wt. % In all subsequent experiments., special care was taken that all the ICl was dissolved by the KI solution before any thiosulfate was added. Tables I and Ia indicate that the ICl is ver-:,' pure., the mean of the eight values(using the corrected values for 2., 3 and 4) being 99.80%., with an average deviation o:f 0.17:,.{. Ex- periments 7 and 8 show that the method outlined for deterrnining acid is quite satisfactory. PRELilv!INARY QUA...liTTI TATIVE EXPERIMENTS. A number of preliminary photochemical experiments were made before it was found possible to make the analyses after illumination agree with the amounts of iodine monochloride used. Only after the apparatus shovm in :figure 3 vras adopted., were satis:factory results obtained. * More striking a.re the results obtained (with a different sa:m.ple of ICl) in another series of titrations for whioh it was necessary to make a simiiar correction: %ICl. 94.58 98.63 98.70 101.0 101.0 101.2 Correoted: 100.6 100.7 101.0 101.0 101.0 101.2 Apparent: s A Flowme-te y- To pumps . K OH and manometer R (, 6 ) H D t 7a,,,k.., C Hz · FIGURI= J. Tank hydrogen was passed through the purifying chain (consisting of tubes of copper turnings, potassium hydroxide sticks, and phosphorus pentoxide respectively., as shO'V'm) into the reservoir A, and out through a flowmeter. This purifying chain served to remove oxygen and moisture from the hydrogen. The hydrogen was allowed to fl0v1 through the reservoir for several hours to S'l'reep out all the air, then for at least two hours with the tube of copper turnings strongly heated, before being used. The ICl., contained in a small tube with a breakable tip, was introduced by means of the side-arm E. This side-ann also con- tained a small piece of iron rod sealed in a glass tube, which, by means of a small solenoid., could be made to break the tip of the ICl tube whenever desired. The 1·eaction vessel R was eonstructed from a one liter round-bottomed flask, and had a diameter of about 12 em. Fi, F2, G and H constituted an automatic circulating device. G was a plunger of thin walled glass tubing, 16 cm. long, in which vras sealed a 14. 5 cm. length of steel rod about 4 mm. in diameter., and which fitted very closely, yet freely., within tube H; the latter was about 42 cm. long. F1 and F were solemoids, 2 each 3 inches long and 1 inch in diameter., and each wound with 3500 turns of no. 24 enamel-covered copper wire. By mechanically al·t;ernating the current from one solenoid to the other, the plunger G could be ma.de to oscillate back and forth through almost the full length of H about once a second., and quite effective circulation of the gas was thus obtained. The system was filled with the reaction mixture in the following manner: The apparatus was completely evacuated belovr stopcock 1. Liquid air was placed at U-trap J to condense out mercury vapor from the system., and then at u-traps C and D. The ICl in the bulb in arm E was completely solidii'ied with liquid air., then the tip of the ICl bulb was broken. When all the ICl had distilled into C and D. the arm E ,va.s sealed off. Any air introduced by the ICl bulb vms pumped off. and the hydrogen was let into the reaction flask from the reservoir A by means of the 11 doser 11 B. which was the space between stopcocks 1., 2 and 3 0 When the desired amount of hydrogen had been introduced, the reaction vessel was sealed off at the constriction K. The volume of the reservoir A had been directly determined before assembling the apparatus by measuring the volurne of water which it could hold., and found to be 255 cc. By expanding air from A into B., and into the volume below stopcock 2., before and after sealing off at K, and measuring the resulting pressures on the manometer., the following volumes had been calculated: B., 37.0 cc.; total volume below stopcock 2., about 1155 cc.; volume oi' reaction system (below constriction K). about 1150 cc. From these volumes and the weight of ICl used., the partial pressures of ICl and of hydrogen could be determined. After the hydrogen had been introduced and the reaction system sealed off, the liquid air was removed from the u-traps and the ICl allov,ed to evaporate into the flask. The latter was then illuminated with a 500-watt projection lamp whose filament was 17 cm. from the surface of the flask. During the illumination, water at 15° c. was kept at the u-trap D, and the circulating pump was operated continuously; considerable iodine always condensed out at D. At the end of the illumination, liquid air was placed at D and the pump operated rapidly for at least an hour to condense out all the ICl, I 2 and HCl. The trap D was then sealed off and the condensed materials titrated. Two runs were made in this manner under exactly the same conditions with approximately the same amounts of ICl, but greatly different hydrogen pressures. hours in each case. was shown. The time of illumination was four A marked dependence on the hydrogen pressure The results are given in table II. Table II. Wt. ICl used. Initial pressures. ICl ~ Vol. 0.1039 n. thio (cc.) For free For acid Total. halogen. Found Cale. % ICl Consumed % 1. o.1304 13. o 60 14.00 1.32 15.32 15.45 99.2 17.3 2. .1457 14. 5 330 11.55 5.65 17.20 17.28 99.6 65.7 FINAL PHOTOCHEMICAL EXPERIMENTS. The method described in the preceding section, although it could undoubtedly be applied, is hardly practical for an extended series of experiments. The principal objection is that only one single measurement can be made with a particular mixture of ICl and hydrogen. The fact that fairly long illuminations are required, even with umnonoohromatized light from a 500-watt lamp, would make the process exceedingly slow and tedious, and this would be much more the case if monochromatic light were used. Accordingly, an entirely different analytical procedure was adopted, and it was decided to follow the rate of the reaction by measuring, by means of a photocell, the change in the absorption of light by the ICl as the reaction proceeded. ~action System. The reaction system used was Sllnilar to that shown in figure 3, except that a better manometer than that used previously was attached at the 11 doser 11 B., and a liquid air trap WB:S inserted between stopcock 2 and the constriction K. The purpose of this trap was to catch mercury vapor., and any condensable materials which might be carried into the reaction vessel by the hydrogen. The volume of the system below the constriction was determined ver~ carefully by expansion of air from A into B, and into the part of the system below stopcock 2 before and after sealing off at the constriction; the resulting pressures were measured on the new manometer. The volume was found to be 1163 cc. The hydrogen and the ICl were introduced into the systern in exactly the same manner as before. However., the pressure of' the hydrogen, instead of being determined from the volume relations, was read directly on the manometer bei'ore sealing the reaction flask off at the constriction. Optical Syste~. The light source consisted of a quartz mercury vapor arc lamp which was always operated at a current of 4.0 amperes, and with a potential drop of 33 volts across the electrodes. The lamp was housed in a wooden box 33 cm. high, 20 cm. vnde, and 25 em. long. The box was open at the top and back, and had a rectangular hole 6.5 x 14 cm. in the front., in front of' which was mounted a shutter. Across the hole, inside the box., was fastened a brass filter compartment, 3.5 cm. thick, with glass windows 6 x 9 cm. Water was f'lovred through this filter compartment continuously while the lamp was running. By placing the lmrer pa.rt of the lamp in a horizontal v-shaped slot cut in a piece of heavy brass which was fastened permanently to the box, the lamp could always be replaced in exactly the same position, very close to the filter compartment. 0 To isolate the mercury blue line, 4358 A, a filter consisting of a 2 nun. thick piece of Noviol A glass and a 3 mm. thick piece of Cornung G-585 glass, with a 1 cm. layer of s~r CuS04 solution betv,een the:m, was used. This filter was mounted over a square hole, 4.75 cm. on a side, in a strip of brass which slid smoothly into the filter compartment. The photocell circuit was as follows: ➔ 1)~1 3 8 0 q "' /_J 5 Fig. 4 The Burt sodium photocell C was mounted on a Bakelite socket in a closed box 20 cm. high, 12 cm, wide, and 12 cm. long, which had a 1.9 cm. diameter hole in the front. The socket., and the two binding posts which passed through the side of the box., were cast in sulfur plugs to insulate them completely from the box. The photocell had a grounded metal guard ring, A, fastened tightly around its neck. A high sensitivity galvan01neter, G, w-as used. rt had a resistance of 595 ohms and a sensitivity of 11.,300 megohms. It was protected by the 200,000 ohm resistance,~ and wns almost critically damped by the 24.,000 ohm resistance, Rz. The source of potential, B, 11vas tvm 45-volt B batteries; these were mounted on The swi toh, S, consisted of tvro small pools of vtire mercury in a block of paraffin, with a removable ~bridge. vi'h~rever paraffin blocks. it was necessary to suspend the connecting wires, they were suspended frmn paraffin blocks. very necessary. All the above insulating precautions were Before they were taken, numerous leaks manifested themselves in the electrical circuit, but these disappeared one by one as the various precautions were taken, and were finally entirely eliminated. The boxes containing the photocell and the mercury lamp were firmly fastened to a rigid base, facing each other at a distance of 14 cm. apart. The base was mounted on brass runners so that the entire assemblage could be slid forward and back. Two stops were arranged so that in one position there was a free path betvreen the t\vo boxes, while in the other, the reaction flask crune directly betllreen them. The entire reaction system, except for the front and back of the reaction flask itself, was painted with optical black to prevent the entrance of stray light. Experimental ~~~• During all :measurcments of light transmissions with the photocell, a brass diaphragm, 2.5 cm. in diameter, was allm.ys inserted in the filter compartment. This was done to reduce the galvanometer deflections to the desired amount, and also to insure that the only light reaching the photocell was that passing through the center of the spherical flask where the windows could be considered as essentially plane and parallel. n Before filling the reaction flask with iYdrogen and iodine monochloride, a determination was made of the ratio of the light intensities with and without the flask in the path of the light bewn. This was done in the following manner: first. with a free path betvveen the two boxes, the shutter was opened. At the end of thirty seconds, the galvanometer was read and the shutter closed. At the end of another thirty second period, the galvanometer was again read and the shutter opened; then, a third reading was made after thirty seconds with the shutter open. The difference betvreen the average of the two galvanometer readings with the shutter open and that vii th the shutter closed gave the light intensity in e.rbi trary units. Next., the optical assemblage was pushed into place with the flask in the light path, and the readings repeated. This was contil1ued until three measurements had been made with a free path, and tvro with the flask in the path of the light. The mean of the two latter measurements, divided by the mean of the three former, gave the transmission of the empty flask, which was 0.876. All transmission determinations were made in the above manner, except that usually only tvro measurements were made with a free light path, and one with the flask in the path of the light. If I represents the intensity measured through the flask, and Ir the intensity with a free light path, then o. 876 Ir would be the intensity I 0 measured through the flask if empty, and I actual transmission. '1a° (= I ) the o. 876 Ir The reaction system was filled with ICl and hydrogen as previously described, and a transmission measurement made before starting an illumination in order to determine an absorption coefficient for the ICl. The 2.5 om. diaphragm was then withdrawn and the flask illuminated. During the illwnination, the automatic circulating pump i.ms operated continuously., and u-trap D was kept innnersed in a Dewar flask full of water a.ta definite temperature, somewhat below room temperature, in order to keep the iodine formed during the photochemical reaction at a definite and rather low pressure. Tra..."1.smission moa.surements were made at intervals of a.bout 15 minutes. Five series of measurements were made with the reaction flask filled with 0.1144 grams of !Cl, and 596 nun. of hydrogen. The weight of ICl used corresponded to an initial ICl pressure of 14.03 nun. Since the hydrogen is so largely in excess, it can be regarded as constant throughout the run. The initial value of the transmission was 0.2029. law to be obeyed., Assuraing Beer's log.!_= - kp. In this case., log.!.... is -0.6927 and pis 14.03 mm., Io Io so k = O. 04937. This value was used in calculating the pressures of iodine monochloride throughout the run. experiments are shown in table III. The results of the The temperature of the reaction flask was 23° throughout. Plotting the partial pressures of I Cl against ;[, I 0 A t gave the curves shown in figure 6. The values used f'or I 0 wero the means of the values obtained at the beginning and end of each time interval. Since the points are somewhat irregular, a straight line was chosen as the best curve in each case. It was f'ound that on dividing the slope of each curve by the average value of the fraction of' light absorbed during the particular run, the resulting 31. TABLE III l\-\z_ Time (min) I/I 0 Io = 59G l'l"YY\-- Prc1 mm (1-I/! 0 ) I0 t Temp. of I 2 Trap Pr2 nun A. 0 14 29 44 59 0.3288 .3358 .3413 .3549 .3655 9.31 9.14 8.99 9.32 8.95 9.79 9.60 9.46 9.14 8.85 0.6712 .6642 .6587 . 6451 .6345 0 129 265 402 540 17°0 1. 54 B. 0 15 32 45 61 75 o.4054 .4180 .4244 .4370 .4465 .4572 9.05 9.31 9.24 9.25 9.27 8.97 7.94 7. 67 7. 54 7.28 7.09 6.88 o.5946 • 5820 . 5756 • 5630 • 5535 .5428 0 138 295 415 563 691 17° 1.54 c. 0 15 30 45 60 75 o. 6079 .6193 .6239 .6378 .6491 .6624 8. 83 8. 70 8.90 8.69 8.97 8.98 4.38 4.22 4.15 3.96 3.80 3.62 0.3921 .3807 .3761 .3622 .3509 .3376 0 132 264 396 528 663 15° 1.28 D. 0 16 30 45 60 75 o.7509 • 7558 . 7675 . 7809 • 7903 . 7954 9.61 9. 42 9.34 9.36 9.24 9.81 2. 52 2.46 2.33 2.18 2.07 2.01 o.2491 .2442 .2325 .2191 .2097 .2046 0 152 283 423 562 705 10° o.oaoa E. 0 15 30 45 60 75 o.s121 . 8224 • 8285 . 8337 .8464 • 8582 8. 76 9. 03 9.67 9.69 9.57 9.58 1.83 1. 72 1.66 1.60 1.47 1.35 0.1879 .1776 .1715 .1663 .1536 .1418 0 133 273 418 563 707 10° 0.0808 4-0 . values were fairly constant over the entire range., as shown in the following table: Table IV - dr, -------d(~ T At) 0 l - I/I 0 Ratio A 0,60186 0.6547 6.00284 13 • 00160 • 5686 .00282 a .00112 .3833 .00292 D .00078 .2265 .00344 E .00058 .1663 .00324 Run The ra-f:;io is ~ - de_ d.(I"r9 ~t) Since -dp/dt is the rate of the • 1-T./:ro reaction., while I 0 (1 - I/I 0 ) is proportional to the rate of absorption of light., it is evident that the constancy of this ratio indicates that the rate of the reaction is proportioal to the rate of absorption 0£ light. Ap,pe.rently., the rate of photochemical reaction., under the conditions studied., is independent of the concentrat:tons of ICl ( except insofai~ as it af:feets; the absorption of light) and of the products of the reaction. Also.,- since the rate of reaction is proportional to the number of molecules of ICl decomposed per unit time, and the rate of absorption of light is proportional to the number of quanta absorbed per unit time., it follows that the number of molecules of ICl reacting per quantum of light absorbed., i.e., the quantum yield., is constant under the conditions of the experiments. In the above consideration, the fraction of the total absorption which was due to iodine was taken as negligible. by the follmving calculation: This is justified When t?10 absorbing species, both of which obey Beer's la,.•;, are present., the fraction x 1 of the absorption which is due to one of them is given by the expression x1 = ___A~1~cr1___., A1c1+A2 c2 41. in which A1 and A2 are the respective absorption coefficients., and c 1 a.nd _c 2 the respective concentrations. Applying this_ expression to run E., in which iodine would have the greatest effect., it was found that about 3.0 % of the light absorption was due to iodine. The pressure taken for ICl was 1.35 mm . ., and that for iodine 0.0808 nnn. (its vapor pressure at 10°). The absorption coefficient for ICl was calculated from the initial transmission measurement., and the initial ICl pressure, e.nd found to be 75. 7. That for iodine at the same wave-length had been determined by Wellman and Ravitz in this laboratory and found to be approximately 40. (Log 10 I/I 0 = -Acl; c i n mols ·.per liter, 1 in centimeters.) The Quantum Yield. In order to obtain an idea. of the order of magnitude of the quantum yield under the conditions studied., a Moll thermopilegalvanometer system wa.s calibrated uga.inst a standard lamp, the directions furnished by the Bureau of Standards being carefully followed. The quartz window· of the thermopile was removed during the calibration. It was found that each centimeter galvanometer deflec- tion corresponded to 1.30 ergsjmm?--/sec. Next., the deflections pro- duced by a steady light source., with and 1.vithout a 7-ohm shunt across the galvanometer, were measured, and it was found that the ratio of the deflection ,vi thout the shunt to that with the shunt vras 2. 39. With conditions exactly the same as during the illuminations of the rel-hydrogen mixture, but with the reaction flask out of the light path., the thermopile (with the quartz window· in place) was placed at a position corresponding to the center of the exposed part of the flask; an1': average deflection of 21.3 cm. was observed., with 42. the 7-ohm shunt across the galvanometer. At a position 2.5 cm. to the right of the center the deflection vms 12.1 cm., and at a position 2.5 cm. to the right of, and 2.5 cm.. above the center it was 7.5 cm. Dnmedi a tely following these measurements, the brass diaphragm was inserted in the filter compartment and a reading made with the photocell in the usual manner, an average deflection of 9.42 cm. being obtained. Assurn.ing,for the purpose of calculation, an average of 20 cm. deflection with the thermopile for an area 2 cm. in radius abou t the center of the portion of the flask exposed to the light ( the diameter of which was 8. 0 cm.) and 9 cm. for the remaining area, a very approximate value of the quantum yield can be deter!nined. A deflection of 20 cm. with the shunt corresponds to one of 47.8 without-the shunt, or 62.2 ergs/mni2/sec., while an area 2 cm. in diameter is 1257 mm2 • Similarly, a 9 cm. deflection corresponds to 28.0 ergs/mm2 /sec., and the area receiving this is 3770 nun2 • Taking into account the transmission of the thermopi le window, which is 90J~, it is found that the amount of' li ght reaching the surface of the flask was 2 :x: 10 5 ergs/sec. , or 1. 2 x 10 7 ergs/min. Since the deflection with the photocell was 9.42, al cm. galvanometer deflection with the photocell corresponds to 1.27 x 106 ergs/min. at the surface of the flask. In run A the average value of I 0 ·was 9.14 cm., corresponding to a. va lue of 10.45 cm. with t.he flask out of the light beam. Hence during that run, 1.27 x 106 x 10.45 or 1. 33 x 10 7 ergs reached the surface of the flask. If it be assumed that 10% of the light is re- flected at the surface, and that t he average absorption, 65.5%, measured during run A (see Table IV) represents the average absorption throughout the cell during the run, then 7. 83 x 10 6 ergs/min. are absorbed. 4 '3. From Table IV., -dpr a t) for run A was o. 00186. d(~ 0 •r 1 t ip · 1yi n ing this by the average value., 9.14 cm . ., of I 0 during the run gives -dp/dt as 0.0170 nmymin . ., from which., assuming the perfect gas equation it is found that 1.07 x 10- 6 mols of ICl were consumed per minute. One einstein of light of wave length 4358 A is 2. 73 :x: 1012 ergs. Hence., in one minute., 1.07 x 10- 6 mols of ICl reacted for every 7.83 x 10 6 ergs., or 2.87 x 10- 6 einsteins of light absorbed. This cor- res ponds to a quantum yield of 0,37 molecules of ICl reacting per quan- tum absorbed. 'l.'his value for the quantum yield is undoubtedly a minimum., and the actual value may be several times l arger. In the first place, the light intensities at the surface were probably less than the value taken. In the second place, since the flask was spherical with only a six inch radius, the reflection at the surface was certainly considerably greater than the icy; assumed. And finally, the absorption measured was through the longest possible path., whereas a. fairly large amount of the light must have traversed a considerably shorter path. 4-4-. SUMMARY The effect of light on 1nixtures of iodine monochloride a.nd hydrogen ha.s been investigated, vri th the following results: 1. Iodine monochl o ride and hydrogen react under the influence of light to give hydrogen chloride and iodine. 2. The reaction is much faster at high hydrogen pressures than at low. 3. tvidence has been found tha.t, at a constant hydrogen pressure of a.bout 600 mm., the rate 9f the reaction is dependent only on the amount of light absorbed. a.t tt-e. ~li."'-e, k.ljJri,I\~ .p,-.c.ss ur-~ 4. An approximate determination of t he quantum yield has A been ma.de , yielding a minimum value of 0.37 molecules of iodine monochloride decomposed per quantum of li ght absorbed. REFE..~ENCES 1. Rollefson, J.l1m.Chem.Soc., 51, 770 (1929). 2. Gibson and Ramsperger, Phys.Rev., 30, 598 (1927). 3. Gibson, Zeit. f. Physik, 22_, 692 ( 1928). 4. Rollefson and Lindquist, J .Am. Chem. Soc., ~ , 2793 (1930). 5. Mellor and Iredale, Nature, 127, 93 (Jan. 17, 1931). 6. Ashley and West, Nature, gz_, 308 (Feb. 28, 1831). 7. Coehn and Stuckardt, Zeit.phys.Chem., ~ , 722 (1916). 8. Bowen., J.Ghem,Soc., ~ , 1233 (1924). 9. Mellor, 10. Mellor, ibid., vol. X, p. 615. 11 Gomprehensive Treatise on Inorganic Chemistry ," Vol. II, p. 117; fig. 24, p. 115.