258 RIDEAL AND WILLIAMS THE ACTION OF LIGHT ON XLII1.-The Action of Light on the Ferrous Ferric Iodine Iodide Equilibrium. By ERIC KEIGHTLEY RIDEAL and EDWARD GARDNER WILLIAMS. THE exceptions to Einstein’s laws of pliotochemical equivalence may be divided into two classes according as the number of quanta absorbed is (1) greater than the number of molecules decomposed, the excess quanta being emitted as fluorescence or dissipated in heating the reacting system or (2) less than the number of molecules decomposed. In class (Z) chain reactions occur of the atom type as is exemplified in the reactions C1 + H = HC1+ H and H + CH = HCl + C1 (Nernst 2. Elektrochem. 1918 24 335), the product of the reaction with the original energy of excitation together with the energy liberated by the reaction may excite other molecules on impact (Christiansen and Kramers 2.phy8ikaZ. Chem., 1924,104,452) as e.g. in the chorina~on of toluene a t low temper-atures without a chlorine carrier (Book and Eggert 2. Elektrochem., 1923 34.0 521). It is to be anticipated that if a photosensitive system could be formed in which chain reactions of either of these ssort could be eliminated the Einstein law of photochemical equi THE FERROUS FERRIC IODINE IODIDE EQUILIBRIUM. 259 valence would be obeyed. Conditions for such evidently obtain in systems in which the energy may be rapidly dissipated no atom chains being formed and in which the energy of excitation is relatively smail. It has long been known that the interaction between ferric salts and iodides to produce ferrous ions and free iodine is revemible and that the reaction expressed by the equation dark light 2Fe"' + 21' T+ 2Fe" + I, is photpensitive.Sasaki (Mem. CoZZ. Sci. Kyoto 1922 5 5) has shown more recently that under conditions of uniform illumination with light from an electric lamp it photochemical equilibrium different from the dark equilibrium is attained. It was considered that this reaction might reasonably be expected to obey the Einstein law of photoequivalence. An investigation of the photodynamics of the system revealed the fact that this law was rigidly fulfilled the mechanism of the reaction being expressible in the form I' If\ /I' \I' It/ \It I' \Fe" 3- Fee*/ -t hv = If-Jj'e*'* + Jj'e***-I' + I' : I-1'-I I f / and that the minimum value of a photochemically active quantum corresponded to a wave-length X = 5790 A.or a potential of 2.14 volts. Further investigation indicated that the photosensitive consti-tuent was the iodine; both the iron salts and the iodine ion being inactive over the region of spectrum employed. The value of the quantum determined in this way X = 5790 8. E = 49,200 cals. per gm.-mol. or V = 2-14 volts agrees very closely with the resonance potential of the iodine molecule 2.34 & 0.2 volts as determined by Foote and Mohler (" The Origin of Spectra," p. 78), thus supporting the evidence obtained from this investigation that the photochemical action results from the optical excitation of the iodine and that the level of energy corresponding to the resonance potential of the iodine molecule is also a chemically active state of excitation.E x P E R I M E N T A L. (a) Isolation of the Photochemically Active Light of Minimum Quantum Value. Equal volumes of two solutions of the compositions (A) Ferric ammonium afurt 0.02 molar ammonium sulphate 0.02 molar, E 260 RIDEAL AND WILLIAMS THE ACTION OF LIGHT ON made up to N/10 by the addition of sulphuric acid and (B) potassium iodide 0.09 molar were mixed. The following reaction occurs: Fe,( SO,), (NH,),S04,24H,0 + ( NH,),S04 + 4*5KI + 5H,S04 yt 2FeS0 + 2(NH,),S04 + K,SO f- I + 2.5KI + 5H,SO,. The liberation of iodine proceeds rapidly a t the start in the dark, and equilibrium is almost attained a t the end of 24 hours but a slow reaction goes on for several weeks.As source of light a 1000 c. p. gas-filled lamp (110 volts 10 amps.) was employed in conjunct'ion with a series of " M " light filters (Wratten type) for isolating comparatively narrow regions of the spectrum. The cuts of these filters as published by the makers were confirmed by means of a direct-vision spectroscope and found to be accurate under the conditions of the experiment. Both the gelatin light filter and the flat-sided glass-stoppered reaction vessel were maintained a t a uniform temperature of 25" in glass-sided troughs by means of a water-circulating system from a controlled thermostat. I n order to obtain approximate equality of intensity in the transmitted light the distance of the lamp from the filter was changed for each combination of filters according to their transmission as published by the makers.To follow the change in equilibrium the amount of iodine in solution was estimated by titration of 5 C.C. of solution run into 100 C.C. of water a t 0" with approximately N/200-sodium thio-sulphate solution using starch as indicator. The usual precautions concerning the quantity of starch and standardisation of the weak thiosulphate solution were taken. The following results were obtained :-Wave-Iength in A. of light transmitted. 6500-7000 6000-7000 5000-6300 4000-5 100 5100-5500 4300-4700 4870-5040 Exposure in hours. 2 2 2 2 2 2 2 1 Titration value for the unilluminated solution in C.C. less that for illuminated. 0 0.60 0.80 0.15 0 0 0 8.75 The region of the spectrum responsible for the photochemical change appears from these data to lie between h = 5500 and 6500 A.The quartz mercury vapour lamp emits a group of strong yellow lines a t about X = 5790 8. which can be isolated conveniently by means of a colour filter prepared by the Kodak Co. Replacing the 1000 c. p. lamp by a quartz mercury vapour lamp the following data were obtained the time of exposure being in all cases 2 hours THE FERROUS FERRIC IODINE IODIDE EQUILIBRIUM. 261 Titration value in C.C. Filter. Light transmitted in A. NI2OO-thio. Nil Complete spectrum 6.6 E 5790 5769 5679 (and red lines) 1.05 c 4916-4078 4047 0 H 5461 4916 4359 4348 4339 0 Mercury groen 5461 0.20 These results confirm those obtained with the 1000 c.p. electric lamp and show that in addition to the ultra-violet light the lines 5700 5769 5679 are photochemically active (the red lines although absorbed by the solution being very faint contributed but little if any energy to the system). The line X = 5790 A. corresponds to an energy of excitation of 49,200 calories or a potential of V = 2.14 volts a value as we have seen almost identical with the resonance potential of the iodine molecule. 274 yellow (b) The Photochemically Actire Constituent. As has already been indicated it would appear from the above experiments that the iodine was the photochemically active constituent. Since we are dealing with a reversible system any alteration in the equilibrium constant would entail a relative alteration in the velocity coefficients of the two reactions : (1) I + 2Pe" -+ 2Fe"' + 21' (2) 21' + 2Fe"'+ 2Fe" -i- 1, k , ka It was accordingly important to find out whether this apparent relative alteration in the velocity coefficients under illumination was produced by the actual alteration in k1 and that k remained unaltered under illurnination.In order to establish this point a modification of the Harcourt and Esson experiment as used by Donnan and Le Rossignol (J., 1903 83 703) to determine the order of reaction between the ferricyanide and iodide of potassium was employed. Fifty C.C. of the iodide solution B to which 2 C.C. of a standard thiosulphate solution and two drops of starch solution had been added were placed in a vessel provided with a stirrer and maintained at 25" in a thermostat.An equal volume of the ferric iron solution A at 25" was added and the time of appearance of the blue colour noted. Immediately on its appearance a fresh quantity of thiosulphate solution was added and the time of reappearance of the starch iodine blue observed. With the precautions as to strength of solution and use of the delivery pipette given by Donnan and Le Rossignol the reaction was carried out at 25" in the dark and in the light of the 1000 c. p. electric lamp. The rates of reaction both in the dark and under illumination were found to be identical 262 RIDEAL AND WILLIAMS THE ACTION OF LIGHT ON In the above reaction the iodide-ion concentration was maintained constant during the conversion of the ferric into ferrous ion and the system is insensitive to illumination.Thus the alteration in the equilibrium of the system under illumination is due to an increase in the velocity of the reaction :-I + 2Fe" --+ 2Fe"' + 21'. If the increase in the velocity coefficient of this reaction nnder illumination were due 60 a photochemical increase in the chemical activity or thermodynamic concentration of the iodine molecules, should exhibit a definite elec-tromotive force. Such cells constructed with platinum electrodes a n d v a r i o u s strengths of iodine solutions in potassium iodide showed after removal of all traces of dissolved oxygen no electromotive force on illumination of one half of %he cell. It therefore seems probable that the production of excited iodine molecules by the photochemical action is small compareJ with the rate a t which the reaction 2~ + I p- 21' occurs a t the electrode.(c) The Application of the Einstein Law of Photoepicalenee. E'or the determination of the amount of radiation absorbed and the concomitant chemical action produced by the radiation the following apparatus was employed (Fig. 1) : A. Mercury vapour lamp. B. Screen. C. Cooling trough. D. Blackened metal plate carrying gelatin light filter E. 3'. Cooling trough containing glass-stoppered reaction vessel G. H. Moll thermopile. I. compensating leads and galvanometer. The Moll thermopile H was first calibrated by mmns of a Leslie cube precautions being taken to ensure " black body " radiation as nearly ideal as possible in the circumstances by screening the system of Leslie cube and thermopile with non-conducting material faced with reflecting surfaces in which accurate openings were cut.The radiant energy received by the thermopile from the Leslie cube was calculated by the formula of Lummer and Pringsheim (Ann. PhysiE 1897 63 395) and found to be 1-23 x lo5 ergs per sec. I THE FERROUS FERRIC IODINE IODIDE EQUILIBRIUM. 263 The readings of the galvanometer connected t o the thermopile during the radiation fluctuated slightly between 35.4 and 36.4 cm., yielding a mean value of 35-95 cm. Hence the energy per cm. deflection is 3.423 x lo3 ergs per second. Allowing for the fact that lamp-black surfaces emit and absorb but 00% of the true black-body radiation the thermopile constant is 2-77 x 103.The volume of the solution in G covered by the funnel of the thermopile was found both by measurement and by calculation to he 31 C.C. The mercury vapour lamp was started and allowed to attain a steady state the current being checked by means of an ammeter connected in the circuit. The thermopile suitably protected from draughts was allowed to come into equilibrium with its surroundings, precautions being taken to keep the room temperature as steady as possible. The bottle was first filled with water and a number of readings were taken of the deflection caused on the galvanometer scale when the screen B was raised. The bottle was then filled up to the mark with the dark equilibrium solution and a number of readings taken as before. Since even with the utmost precaution mercury vapour lamps are apt to vary in intensity of illumination a number of independent experiments were performed of which the two appended represent the extremes of variation.The curve of the rate of disappearance of iodine under the same conditions was also obtained by means of a large number of experiments which agreed fairly closely. Water. Deflection on galvanometer scale in cm. Initial. Final. in cm. / \ TotaI deflection 7.3 27.8 7.0 27.5 7.0 27.2 5.4 27.8 5.6 28.2 5.6 28.2 (I) 7.2 27.5 (11) Dark equilibrium mixture. 6.9 15.1 (I) 7.0 15.0 6.9 14-8 7.0 15.3 5.4 13.7 (I1) 5.6 14-2 5.4 14.0 Energy absorbed by solution in (I) = Energy absorbed by solution in (11) = Mean = 20.2 22.6 8.3 12.3 x 2.77 x los ergs per second.14.03 x 2.77 x 10s ergs per second. 13.2 x 2.77 x 108 ergs per second 264 RIDEAL AND WILLIAMS THE ACTION OF LIGHT ON The rate of disappearance of iodine under the same conditions was next investigated. At various times 5.15 C.C. were withdrawn run into 100 C.C. of cold water and titrated with approx. N/lOO-thiosulphate the following data being obtained from which the velocity ccefficient [k = l/t.log a/(a - x)] has been calculated a = 7-5 x 0.0012696 g. of 12. T N/lOO-Thio. (mins.). (c.c.). t . X. k. 0 7.50 15 0.0006345 0.00463 15 7.14 30 0.001 142 0.00430 52 6.20 60 0.002030 0*00404 120 5.40 90 0.003173 0.00455 192 4-97 422 4.91 Mean = 0.00438 The reaction appears to follow the unimolecular law as indicated by the values of Ic hence the initial rate of disappearance of iodine may be calculat8ed.dx/dt = kc = 0.00438 x 0.000005 x 7.5 x 6.06 x molecules I2 per minute for 5-15 C.C. = 0.996 x 1016 molecules I2 per see. for 31 C.C. We can now proceed t o calculate the number of quanta absorbed Wave-length = 5790-66 A. Magnitude of quantum = h v = 6.55 x per see. by the solution. Frequency v = 517G x 1014. x 5.176 x 1014 = 3.39 x erg. Energy absorbed by solution = 13.2 x 2-77 x 103 ergs per see. Hence number of quanta absorbed by solution = 1.078 x 1016 per sec. - _____ -There are thus 0.996 x 1016 molecules of iodine being decomposed per second by 1.078 x 1016 quanta. The reaction thus obeys the Einstein law of photochemical equivalence under the condit,ions of the experiment.(d) The Temperature Coefiicient of the Reactions. The influence of temperature on the dark equilibrium system 2Fe" + I r j 2Fe"' + 21' is negligibly small but the rate a t which equilibrium is attained is very dependent on the temperature; this rate was investigated at 25" and 33-25" and the equilibrium point was reached in each case though more rapidly in the warmer solutions. By estimation of the amount of iodine liberated after mixing the solutions A and THE YERROlTS FERRIC IODISE IODIDE EQUILIBRIURI. 265 a t the two temperatures the temperature coefficient was calculated from the velocity coefficients to be for a rise in temperature of lo", k35*/k250 = 2.713. The iiifucnce of temperature on the rate of attainment of the light equilibrium was investigated by a potentio-metric method.A photovoltaic cell coiisisting of bright platinum electrocks inserted in a constricted U-tube containing the dark equilibrium mixture was eniploycd. On illumination of one limb of the U-tube with i~ioiioc~lromat'ic yellow radiation at 25" the equilibrium concentration in the exposetl limb is altered arid the rate of alteration can 1;e dctcrniined froiii thc electromotive force of the ccll. On cxtting off the radiation. the return of the photo-chciuical equilibriuni niihture to tlie dark equilibrium can readily he followed from the electromotive force of the cell. As typical of the readings obtaiticrt 1::- this method the follon-ing values may be cited. Time iii I n j 11s. 0 20 40 GO 90 113 9 ,..- 43 liilli-volt,s.10.30 10.30 10.20 !b70 8.85 6.45 6.25 5.90 Time i n rnins. 0 10 20 3 0 40 i) Teniperut,nre 35". Jlilli- lime in volts. inins. 0.0 50 1.40 60 2.80 70 4.85 80 6.40 90 7.63 100 Nilli-volts. S.60 9.25 9.65 10.00 10.17 10-17 From tlie velocity coefficiciits of the rate of attainment of photo-cheinictL1 equilibriuni the t emperature coefficient k,,./k,,. (light) was fouiid to be 1-17. It is probable that the velocity of the true piic;tocheiiiicaI reaction is not iiiflueiiced by temperature and that the value 01 1.1T obtai!ied is due to the effect of the increase i n velocity of tlicl dark reaction 11 ith thc tenipcrature huperiinposed o i l the vcblocitjr of Piic light roac'tio!i.I t has already bwil iiotctl that altliough the liberation of iodine from the reactiirg system co~i~ineiicw rilpidly J e i the reaction comes 1 ) u C slon-1) to eqi~ilihriuni. At the ciid of 24 l~oim thc average airlount of iodine 1ik)crdtctl as tfcieriniiied 1);)~ titration with thio-sulphatc v-as fouiid cqcivalc~nt to 7-50 C.C. of S/lOO-thiosulphate €or 5.15 C.C. of the equilibrium mixture. Although the reaction is apparently completed n e ~ ertheless a ~ O T V reaction is still pro-ccecliig towards a true ecluilibriurn. This ec1ui;ibriurn was estab-l i s l i c d at the eiicl of 3 i~ic~ntiis and found cqual to 8-37 c.c. of S/lOCI-tliiosiili~~i~ttc. Sincc the equilibrium attained is the result of the attaininelit of cqual velocity for two reactions 2Fe *+ I 4 K 266 RIDEAL AND WILLIAMS THE ACTION OF LIGHT ON 2Fe"'+ 21' and 2E"e"'+ 21'- 2Fe"+ I, it is clear that if the active masses of equilibrium concentrations of the reactants could be raised in equal proportion the reaction velocities of each reaction would proceed more rapidly and equilibrium would be more rapidly attained.The active masses or thermodynamic concentrations of the active ions whether simple as indicated above or complex as imagined by Sasaki (2. angew. Chem. 1024,137,181) can readily be raised by the addition of a neutral salt such as potassium chloride (Bronsted 2. physikal. Chem. 1922 103 307). I n the following curves (Fig. 2) the rate of attainment of equilibrium in thc dark for three mixtures to which varying amoiints of potassinm chloricte have been added are given.Fra. 2. Time minutes. It is clear from the curves that the augmentation in thermo-dynamic concentration of the reacting ions by the addition of the potassium chloride has as anticipated augmented the rate a t which true equilibrium is attained. That the separate reaction velocities are in reality augmented by the addition of potassium chloride was confirmed by repeating the modified Harcoixrt and Esson experiment as already described in presence of 1437-potassium chloride and in the ordinary solution without the salt addition. The marked increase in the velocity of the forward reaction on the addition of potassium chloride is evident from the following data : Time in minutes for appearance of starch-iodide colour. Thioadded(c.c.) 1 2 3 4 5 G 7 8 9 No KCl added 0.5 0.8 1.5 2.8 4-7 8.15 13.7 25-0 In 1.6N-KCl ...0.4 0.6 1.0 1.6 2-4 3.6 5.6 8.9 16. THE FERROUS FERRTC IODINE IODIDE EQUILIBRITTM. 267 At equilibrium in the dark the rates of the forward photosensitive reactions acre given in absence and presence of potassium chloride by the equations dzldt = Ic .f(I,)f (Fe**)2 - a and dx'ldt = II:.f'(I,)f'(Fe")2 = b where f andf' are the activity coefficients in absence and presence of the neutral salt. On illumination of the solutions tlhe respective velocities become dx/dtfil. = II .f(Iz)f(Fe")2 + ](I,) = a + x, dz'ldtill. = II .f'(12)f(Fe")2 + I(&) = b + X, where I is the intensity of the photoactive radiation. wiOhout and in presence of salt will evidently be, The ratio of the velocities of the two dark and light reactions Without KCl present Vill./Vdark = (a + x)/a = 1 + z/a.With ICC1 present Viu,/Vdark = (b + x)/b = 1 + xjb. Thus the greater the absolute velocity of the forward reaction a t equilibrium the smaller will be the apparent effect of the light on the equilibrium. We should thus anticipate that the shift in the equilibrium obtained with the solution containing potassium chloride on exposure to light will be far less than the shift in the equilibrium of the solutlion in the absence of potassium chloride on exposure to identical radiation since the addition of salt has as has been indicated by the previous experiments raised the value off' considerably above f or b > a. These expectations were fully realised by determining the rate of change of the dark equilibrium mixture on illumination both by the potentiometric method and by titration.The following data were obtained for a dark equilibrium mixture containing 1 -5N-potassium chloride exposed to mono-chromatic radiation a t 25". Time (mins.). 0 1 s 10 26 30 Milli- N/lOO-Thio. Time Milli- N/lOO-Thio. volts. (C.C.). (mins.). volts. (c.c.). 0 7.86 41 2.40 -0.1 - 52 2.60 I 0.8 - 60 2.65 7.30 1-05 - 70 2.70 -1.90 - 170 2.70 7.30 2.15 7.75 The final value of 2.70 mv. as compared with a value of 10.3 mv. for the solution in the absence of potassium chloride (see Table, p. 265) and a shift in the equilibrium concentration of iodine equivalent to only 0.56 C.C. of N/lOO-thiosulphate solution as com-pared with the value 2.50 C.C.obtained without the potassium chloride indicated clearly the effect of augmenting the velocity of the two reactions proceeding in either direction at equilibrium by elevation of the thermodynamic concentrations of the reactants. K" 268 THE FERROUS FERRIC IODINE IODIDE EQUILIBRITJM. The value of the equilibrium constant for the reaction K = (Fe")(12)$/(Fe"')(I') has been determined by Maitland (2. Elektrochem. 1906 112 263) and by Bronsted and Pedersen (2. physikal. Chem. 1922,103 307). Maitland obtained the value K = lO2'005 whilst the data of Bronsted and Pedersen yielded a value of K 2 y = 212.1 ~t~ilising a value of L = (I')(12)/(I'J = 0.0061. = 0.00611 for the equilibrium constant The data obtained in the above experiments permit us to calculate the value of K .The original concentration of ferric ions = 0.02 gm.-atom per litre. The titre of the dark equilibrium solution is 8.37 C.C. of N/100-Y Y Y Y Y 9 of iodide ions = 0.045 gm.-atom per litre. thiosulphate for 5.15 C.C. Hence a t equilibrium we obtain : (Fe"') = (2.0 - 8-37/5-15)10-2 gm.-atom per litre. (Fee*) = 8.37/5.15.10-2 gm.-atom per litre. (1') = 8.37/5.15.10-2 gm.-atom per litre. we obtain Also if the original iodide concentration be x and the final .z - a, I,' = 3a I = x - 2a, = 2.875 x I' = x - a and (x - 2a)(x - a)/% = 0.00611 . . . . . (1) * where x = (4.5 - 8.37/5.15)10-2 gm.-atom per litre gm.-atom per litre. Inserting this value of x in (I) and solving for a we obtain a = 0.58 x gm.-atom per litre.Hence (1') = (2.575 - 0.58j x 10-2 = 2.30 x 10-2 gm.-atom per litre. (1.625 x 10-2)(1*625 x 10-2)a ~ 23.6, (0.380 x 10-2)(2-30 x Hence K25" = value in close agreement with that of Bronsted and Pedersen. The slight discrepancy between 23.6 and 21.1 for the values of K in the two investigations is probably accounted for by the fact that Bronsted and Pedersen utilised only chlorides and iodides of iron and potassium whilst in the present case ammonium sulphate was present in addition. Xummary. The reaction 2Fe"'+ I ~2 2Fe"+ 21' is photosensitlive to both ultra-violet and visible light. The region of visible photoactive radiation is within the range 5500-6500A. with an apparen SHORT THE CONDENS-4TION OF PHENYLETHPLAMINE ETC. 269 maximum a t 58008.The tri-iodide ion is the photoactive constituent and the energy of excitation corresponding to a wave-length 5800 A. is equivalent to 2.14 volts a value almost identical with the resonance potential of the iodine molecule. It is found that one quantum of absorbed radiant energy causes 1 mol. of iodine to react; the mechanism of the reaction can be expressed in the 1-for111 2Fe1 + I,' -/- hv 1 2FeT3 + 1'. The dark eyuilil:)r.ium coils1 ant of the reaction is K = 23-6 = (Fe"')(T2}~/(Fe"')(I') a t 23". The temperature coefficient of the liberation of iodine in the dark is k,; ,ikZ5 - 2.723 whilst that of the photochemical reaction is ,'Idz5 = 1.17. The addition of potassium chloride raises the theimodynaiuic concentrations of all the reactants and although the equilibrium point is unchanged the rate of attainment is very considerably increased.Proof is given that equilibrium is the result of attainment of equal velocities of the forward and the back reaction since the effect of radiation of constant intensity on the equilibrium attained is less when the thermodynamic concentrat ions of the reactants (and thus the reaction velocities of the two reactions} are raised by the addition of potassium chloride. Our thanks are due to the Department of Scientific and Industrial Research for a grant out of which a part of the cost of the apparatus was provided. IA A u OR -i TO RT o 17 P HITS I ( .A 1 C H F JI r s-rR\-, G'AMI3RTI)C~E. [ R P c P ~ u P ~ ' Ncrc~t-r~~bcr 1&h 1024. 258 RIDEAL AND WILLIAMS THE ACTION OF LIGHT ON XLII1.-The Action of Light on the Ferrous Ferric Iodine Iodide Equilibrium.By ERIC KEIGHTLEY RIDEAL and EDWARD GARDNER WILLIAMS. THE exceptions to Einstein’s laws of pliotochemical equivalence may be divided into two classes according as the number of quanta absorbed is (1) greater than the number of molecules decomposed, the excess quanta being emitted as fluorescence or dissipated in heating the reacting system or (2) less than the number of molecules decomposed. In class (Z) chain reactions occur of the atom type as is exemplified in the reactions C1 + H = HC1+ H and H + CH = HCl + C1 (Nernst 2. Elektrochem. 1918 24 335), the product of the reaction with the original energy of excitation together with the energy liberated by the reaction may excite other molecules on impact (Christiansen and Kramers 2.phy8ikaZ. Chem., 1924,104,452) as e.g. in the chorina~on of toluene a t low temper-atures without a chlorine carrier (Book and Eggert 2. Elektrochem., 1923 34.0 521). It is to be anticipated that if a photosensitive system could be formed in which chain reactions of either of these ssort could be eliminated the Einstein law of photochemical equi THE FERROUS FERRIC IODINE IODIDE EQUILIBRIUM. 259 valence would be obeyed. Conditions for such evidently obtain in systems in which the energy may be rapidly dissipated no atom chains being formed and in which the energy of excitation is relatively smail. It has long been known that the interaction between ferric salts and iodides to produce ferrous ions and free iodine is revemible and that the reaction expressed by the equation dark light 2Fe"' + 21' T+ 2Fe" + I, is photpensitive.Sasaki (Mem. CoZZ. Sci. Kyoto 1922 5 5) has shown more recently that under conditions of uniform illumination with light from an electric lamp it photochemical equilibrium different from the dark equilibrium is attained. It was considered that this reaction might reasonably be expected to obey the Einstein law of photoequivalence. An investigation of the photodynamics of the system revealed the fact that this law was rigidly fulfilled the mechanism of the reaction being expressible in the form I' If\ /I' \I' It/ \It I' \Fe" 3- Fee*/ -t hv = If-Jj'e*'* + Jj'e***-I' + I' : I-1'-I I f / and that the minimum value of a photochemically active quantum corresponded to a wave-length X = 5790 A.or a potential of 2.14 volts. Further investigation indicated that the photosensitive consti-tuent was the iodine; both the iron salts and the iodine ion being inactive over the region of spectrum employed. The value of the quantum determined in this way X = 5790 8. E = 49,200 cals. per gm.-mol. or V = 2-14 volts agrees very closely with the resonance potential of the iodine molecule 2.34 & 0.2 volts as determined by Foote and Mohler (" The Origin of Spectra," p. 78), thus supporting the evidence obtained from this investigation that the photochemical action results from the optical excitation of the iodine and that the level of energy corresponding to the resonance potential of the iodine molecule is also a chemically active state of excitation.E x P E R I M E N T A L. (a) Isolation of the Photochemically Active Light of Minimum Quantum Value. Equal volumes of two solutions of the compositions (A) Ferric ammonium afurt 0.02 molar ammonium sulphate 0.02 molar, E 260 RIDEAL AND WILLIAMS THE ACTION OF LIGHT ON made up to N/10 by the addition of sulphuric acid and (B) potassium iodide 0.09 molar were mixed. The following reaction occurs: Fe,( SO,), (NH,),S04,24H,0 + ( NH,),S04 + 4*5KI + 5H,S04 yt 2FeS0 + 2(NH,),S04 + K,SO f- I + 2.5KI + 5H,SO,. The liberation of iodine proceeds rapidly a t the start in the dark, and equilibrium is almost attained a t the end of 24 hours but a slow reaction goes on for several weeks. As source of light a 1000 c.p. gas-filled lamp (110 volts 10 amps.) was employed in conjunct'ion with a series of " M " light filters (Wratten type) for isolating comparatively narrow regions of the spectrum. The cuts of these filters as published by the makers were confirmed by means of a direct-vision spectroscope and found to be accurate under the conditions of the experiment. Both the gelatin light filter and the flat-sided glass-stoppered reaction vessel were maintained a t a uniform temperature of 25" in glass-sided troughs by means of a water-circulating system from a controlled thermostat. I n order to obtain approximate equality of intensity in the transmitted light the distance of the lamp from the filter was changed for each combination of filters according to their transmission as published by the makers.To follow the change in equilibrium the amount of iodine in solution was estimated by titration of 5 C.C. of solution run into 100 C.C. of water a t 0" with approximately N/200-sodium thio-sulphate solution using starch as indicator. The usual precautions concerning the quantity of starch and standardisation of the weak thiosulphate solution were taken. The following results were obtained :-Wave-Iength in A. of light transmitted. 6500-7000 6000-7000 5000-6300 4000-5 100 5100-5500 4300-4700 4870-5040 Exposure in hours. 2 2 2 2 2 2 2 1 Titration value for the unilluminated solution in C.C. less that for illuminated. 0 0.60 0.80 0.15 0 0 0 8.75 The region of the spectrum responsible for the photochemical change appears from these data to lie between h = 5500 and 6500 A.The quartz mercury vapour lamp emits a group of strong yellow lines a t about X = 5790 8. which can be isolated conveniently by means of a colour filter prepared by the Kodak Co. Replacing the 1000 c. p. lamp by a quartz mercury vapour lamp the following data were obtained the time of exposure being in all cases 2 hours THE FERROUS FERRIC IODINE IODIDE EQUILIBRIUM. 261 Titration value in C.C. Filter. Light transmitted in A. NI2OO-thio. Nil Complete spectrum 6.6 E 5790 5769 5679 (and red lines) 1.05 c 4916-4078 4047 0 H 5461 4916 4359 4348 4339 0 Mercury groen 5461 0.20 These results confirm those obtained with the 1000 c. p. electric lamp and show that in addition to the ultra-violet light the lines 5700 5769 5679 are photochemically active (the red lines although absorbed by the solution being very faint contributed but little if any energy to the system).The line X = 5790 A. corresponds to an energy of excitation of 49,200 calories or a potential of V = 2.14 volts a value as we have seen almost identical with the resonance potential of the iodine molecule. 274 yellow (b) The Photochemically Actire Constituent. As has already been indicated it would appear from the above experiments that the iodine was the photochemically active constituent. Since we are dealing with a reversible system any alteration in the equilibrium constant would entail a relative alteration in the velocity coefficients of the two reactions : (1) I + 2Pe" -+ 2Fe"' + 21' (2) 21' + 2Fe"'+ 2Fe" -i- 1, k , ka It was accordingly important to find out whether this apparent relative alteration in the velocity coefficients under illumination was produced by the actual alteration in k1 and that k remained unaltered under illurnination.In order to establish this point a modification of the Harcourt and Esson experiment as used by Donnan and Le Rossignol (J., 1903 83 703) to determine the order of reaction between the ferricyanide and iodide of potassium was employed. Fifty C.C. of the iodide solution B to which 2 C.C. of a standard thiosulphate solution and two drops of starch solution had been added were placed in a vessel provided with a stirrer and maintained at 25" in a thermostat.An equal volume of the ferric iron solution A at 25" was added and the time of appearance of the blue colour noted. Immediately on its appearance a fresh quantity of thiosulphate solution was added and the time of reappearance of the starch iodine blue observed. With the precautions as to strength of solution and use of the delivery pipette given by Donnan and Le Rossignol the reaction was carried out at 25" in the dark and in the light of the 1000 c. p. electric lamp. The rates of reaction both in the dark and under illumination were found to be identical 262 RIDEAL AND WILLIAMS THE ACTION OF LIGHT ON In the above reaction the iodide-ion concentration was maintained constant during the conversion of the ferric into ferrous ion and the system is insensitive to illumination.Thus the alteration in the equilibrium of the system under illumination is due to an increase in the velocity of the reaction :-I + 2Fe" --+ 2Fe"' + 21'. If the increase in the velocity coefficient of this reaction nnder illumination were due 60 a photochemical increase in the chemical activity or thermodynamic concentration of the iodine molecules, should exhibit a definite elec-tromotive force. Such cells constructed with platinum electrodes a n d v a r i o u s strengths of iodine solutions in potassium iodide showed after removal of all traces of dissolved oxygen no electromotive force on illumination of one half of %he cell. It therefore seems probable that the production of excited iodine molecules by the photochemical action is small compareJ with the rate a t which the reaction 2~ + I p- 21' occurs a t the electrode.(c) The Application of the Einstein Law of Photoepicalenee. E'or the determination of the amount of radiation absorbed and the concomitant chemical action produced by the radiation the following apparatus was employed (Fig. 1) : A. Mercury vapour lamp. B. Screen. C. Cooling trough. D. Blackened metal plate carrying gelatin light filter E. 3'. Cooling trough containing glass-stoppered reaction vessel G. H. Moll thermopile. I. compensating leads and galvanometer. The Moll thermopile H was first calibrated by mmns of a Leslie cube precautions being taken to ensure " black body " radiation as nearly ideal as possible in the circumstances by screening the system of Leslie cube and thermopile with non-conducting material faced with reflecting surfaces in which accurate openings were cut.The radiant energy received by the thermopile from the Leslie cube was calculated by the formula of Lummer and Pringsheim (Ann. PhysiE 1897 63 395) and found to be 1-23 x lo5 ergs per sec. I THE FERROUS FERRIC IODINE IODIDE EQUILIBRIUM. 263 The readings of the galvanometer connected t o the thermopile during the radiation fluctuated slightly between 35.4 and 36.4 cm., yielding a mean value of 35-95 cm. Hence the energy per cm. deflection is 3.423 x lo3 ergs per second. Allowing for the fact that lamp-black surfaces emit and absorb but 00% of the true black-body radiation the thermopile constant is 2-77 x 103. The volume of the solution in G covered by the funnel of the thermopile was found both by measurement and by calculation to he 31 C.C.The mercury vapour lamp was started and allowed to attain a steady state the current being checked by means of an ammeter connected in the circuit. The thermopile suitably protected from draughts was allowed to come into equilibrium with its surroundings, precautions being taken to keep the room temperature as steady as possible. The bottle was first filled with water and a number of readings were taken of the deflection caused on the galvanometer scale when the screen B was raised. The bottle was then filled up to the mark with the dark equilibrium solution and a number of readings taken as before. Since even with the utmost precaution mercury vapour lamps are apt to vary in intensity of illumination a number of independent experiments were performed of which the two appended represent the extremes of variation.The curve of the rate of disappearance of iodine under the same conditions was also obtained by means of a large number of experiments which agreed fairly closely. Water. Deflection on galvanometer scale in cm. Initial. Final. in cm. / \ TotaI deflection 7.3 27.8 7.0 27.5 7.0 27.2 5.4 27.8 5.6 28.2 5.6 28.2 (I) 7.2 27.5 (11) Dark equilibrium mixture. 6.9 15.1 (I) 7.0 15.0 6.9 14-8 7.0 15.3 5.4 13.7 (I1) 5.6 14-2 5.4 14.0 Energy absorbed by solution in (I) = Energy absorbed by solution in (11) = Mean = 20.2 22.6 8.3 12.3 x 2.77 x los ergs per second.14.03 x 2.77 x 10s ergs per second. 13.2 x 2.77 x 108 ergs per second 264 RIDEAL AND WILLIAMS THE ACTION OF LIGHT ON The rate of disappearance of iodine under the same conditions was next investigated. At various times 5.15 C.C. were withdrawn run into 100 C.C. of cold water and titrated with approx. N/lOO-thiosulphate the following data being obtained from which the velocity ccefficient [k = l/t.log a/(a - x)] has been calculated a = 7-5 x 0.0012696 g. of 12. T N/lOO-Thio. (mins.). (c.c.). t . X. k. 0 7.50 15 0.0006345 0.00463 15 7.14 30 0.001 142 0.00430 52 6.20 60 0.002030 0*00404 120 5.40 90 0.003173 0.00455 192 4-97 422 4.91 Mean = 0.00438 The reaction appears to follow the unimolecular law as indicated by the values of Ic hence the initial rate of disappearance of iodine may be calculat8ed.dx/dt = kc = 0.00438 x 0.000005 x 7.5 x 6.06 x molecules I2 per minute for 5-15 C.C. = 0.996 x 1016 molecules I2 per see. for 31 C.C. We can now proceed t o calculate the number of quanta absorbed Wave-length = 5790-66 A. Magnitude of quantum = h v = 6.55 x per see. by the solution. Frequency v = 517G x 1014. x 5.176 x 1014 = 3.39 x erg. Energy absorbed by solution = 13.2 x 2-77 x 103 ergs per see. Hence number of quanta absorbed by solution = 1.078 x 1016 per sec. - _____ -There are thus 0.996 x 1016 molecules of iodine being decomposed per second by 1.078 x 1016 quanta. The reaction thus obeys the Einstein law of photochemical equivalence under the condit,ions of the experiment. (d) The Temperature Coefiicient of the Reactions.The influence of temperature on the dark equilibrium system 2Fe" + I r j 2Fe"' + 21' is negligibly small but the rate a t which equilibrium is attained is very dependent on the temperature; this rate was investigated at 25" and 33-25" and the equilibrium point was reached in each case though more rapidly in the warmer solutions. By estimation of the amount of iodine liberated after mixing the solutions A and THE YERROlTS FERRIC IODISE IODIDE EQUILIBRIURI. 265 a t the two temperatures the temperature coefficient was calculated from the velocity coefficients to be for a rise in temperature of lo", k35*/k250 = 2.713. The iiifucnce of temperature on the rate of attainment of the light equilibrium was investigated by a potentio-metric method.A photovoltaic cell coiisisting of bright platinum electrocks inserted in a constricted U-tube containing the dark equilibrium mixture was eniploycd. On illumination of one limb of the U-tube with i~ioiioc~lromat'ic yellow radiation at 25" the equilibrium concentration in the exposetl limb is altered arid the rate of alteration can 1;e dctcrniined froiii thc electromotive force of the ccll. On cxtting off the radiation. the return of the photo-chciuical equilibriuni niihture to tlie dark equilibrium can readily he followed from the electromotive force of the cell. As typical of the readings obtaiticrt 1::- this method the follon-ing values may be cited. Time iii I n j 11s. 0 20 40 GO 90 113 9 ,..- 43 liilli-volt,s.10.30 10.30 10.20 !b70 8.85 6.45 6.25 5.90 Time i n rnins. 0 10 20 3 0 40 i) Teniperut,nre 35". Jlilli- lime in volts. inins. 0.0 50 1.40 60 2.80 70 4.85 80 6.40 90 7.63 100 Nilli-volts. S.60 9.25 9.65 10.00 10.17 10-17 From tlie velocity coefficiciits of the rate of attainment of photo-cheinictL1 equilibriuni the t emperature coefficient k,,./k,,. (light) was fouiid to be 1-17. It is probable that the velocity of the true piic;tocheiiiicaI reaction is not iiiflueiiced by temperature and that the value 01 1.1T obtai!ied is due to the effect of the increase i n velocity of tlicl dark reaction 11 ith thc tenipcrature huperiinposed o i l the vcblocitjr of Piic light roac'tio!i. I t has already bwil iiotctl that altliough the liberation of iodine from the reactiirg system co~i~ineiicw rilpidly J e i the reaction comes 1 ) u C slon-1) to eqi~ilihriuni.At the ciid of 24 l~oim thc average airlount of iodine 1ik)crdtctl as tfcieriniiied 1);)~ titration with thio-sulphatc v-as fouiid cqcivalc~nt to 7-50 C.C. of S/lOO-thiosulphate €or 5.15 C.C. of the equilibrium mixture. Although the reaction is apparently completed n e ~ ertheless a ~ O T V reaction is still pro-ccecliig towards a true ecluilibriurn. This ec1ui;ibriurn was estab-l i s l i c d at the eiicl of 3 i~ic~ntiis and found cqual to 8-37 c.c. of S/lOCI-tliiosiili~~i~ttc. Sincc the equilibrium attained is the result of the attaininelit of cqual velocity for two reactions 2Fe *+ I 4 K 266 RIDEAL AND WILLIAMS THE ACTION OF LIGHT ON 2Fe"'+ 21' and 2E"e"'+ 21'- 2Fe"+ I, it is clear that if the active masses of equilibrium concentrations of the reactants could be raised in equal proportion the reaction velocities of each reaction would proceed more rapidly and equilibrium would be more rapidly attained.The active masses or thermodynamic concentrations of the active ions whether simple as indicated above or complex as imagined by Sasaki (2. angew. Chem. 1024,137,181) can readily be raised by the addition of a neutral salt such as potassium chloride (Bronsted 2. physikal. Chem. 1922 103 307). I n the following curves (Fig. 2) the rate of attainment of equilibrium in thc dark for three mixtures to which varying amoiints of potassinm chloricte have been added are given.Fra. 2. Time minutes. It is clear from the curves that the augmentation in thermo-dynamic concentration of the reacting ions by the addition of the potassium chloride has as anticipated augmented the rate a t which true equilibrium is attained. That the separate reaction velocities are in reality augmented by the addition of potassium chloride was confirmed by repeating the modified Harcoixrt and Esson experiment as already described in presence of 1437-potassium chloride and in the ordinary solution without the salt addition. The marked increase in the velocity of the forward reaction on the addition of potassium chloride is evident from the following data : Time in minutes for appearance of starch-iodide colour. Thioadded(c.c.) 1 2 3 4 5 G 7 8 9 No KCl added 0.5 0.8 1.5 2.8 4-7 8.15 13.7 25-0 In 1.6N-KCl ...0.4 0.6 1.0 1.6 2-4 3.6 5.6 8.9 16. THE FERROUS FERRTC IODINE IODIDE EQUILIBRITTM. 267 At equilibrium in the dark the rates of the forward photosensitive reactions acre given in absence and presence of potassium chloride by the equations dzldt = Ic .f(I,)f (Fe**)2 - a and dx'ldt = II:.f'(I,)f'(Fe")2 = b where f andf' are the activity coefficients in absence and presence of the neutral salt. On illumination of the solutions tlhe respective velocities become dx/dtfil. = II .f(Iz)f(Fe")2 + ](I,) = a + x, dz'ldtill. = II .f'(12)f(Fe")2 + I(&) = b + X, where I is the intensity of the photoactive radiation. wiOhout and in presence of salt will evidently be, The ratio of the velocities of the two dark and light reactions Without KCl present Vill./Vdark = (a + x)/a = 1 + z/a.With ICC1 present Viu,/Vdark = (b + x)/b = 1 + xjb. Thus the greater the absolute velocity of the forward reaction a t equilibrium the smaller will be the apparent effect of the light on the equilibrium. We should thus anticipate that the shift in the equilibrium obtained with the solution containing potassium chloride on exposure to light will be far less than the shift in the equilibrium of the solutlion in the absence of potassium chloride on exposure to identical radiation since the addition of salt has as has been indicated by the previous experiments raised the value off' considerably above f or b > a. These expectations were fully realised by determining the rate of change of the dark equilibrium mixture on illumination both by the potentiometric method and by titration.The following data were obtained for a dark equilibrium mixture containing 1 -5N-potassium chloride exposed to mono-chromatic radiation a t 25". Time (mins.). 0 1 s 10 26 30 Milli- N/lOO-Thio. Time Milli- N/lOO-Thio. volts. (C.C.). (mins.). volts. (c.c.). 0 7.86 41 2.40 -0.1 - 52 2.60 I 0.8 - 60 2.65 7.30 1-05 - 70 2.70 -1.90 - 170 2.70 7.30 2.15 7.75 The final value of 2.70 mv. as compared with a value of 10.3 mv. for the solution in the absence of potassium chloride (see Table, p. 265) and a shift in the equilibrium concentration of iodine equivalent to only 0.56 C.C. of N/lOO-thiosulphate solution as com-pared with the value 2.50 C.C.obtained without the potassium chloride indicated clearly the effect of augmenting the velocity of the two reactions proceeding in either direction at equilibrium by elevation of the thermodynamic concentrations of the reactants. K" 268 THE FERROUS FERRIC IODINE IODIDE EQUILIBRITJM. The value of the equilibrium constant for the reaction K = (Fe")(12)$/(Fe"')(I') has been determined by Maitland (2. Elektrochem. 1906 112 263) and by Bronsted and Pedersen (2. physikal. Chem. 1922,103 307). Maitland obtained the value K = lO2'005 whilst the data of Bronsted and Pedersen yielded a value of K 2 y = 212.1 ~t~ilising a value of L = (I')(12)/(I'J = 0.0061. = 0.00611 for the equilibrium constant The data obtained in the above experiments permit us to calculate the value of K .The original concentration of ferric ions = 0.02 gm.-atom per litre. The titre of the dark equilibrium solution is 8.37 C.C. of N/100-Y Y Y Y Y 9 of iodide ions = 0.045 gm.-atom per litre. thiosulphate for 5.15 C.C. Hence a t equilibrium we obtain : (Fe"') = (2.0 - 8-37/5-15)10-2 gm.-atom per litre. (Fee*) = 8.37/5.15.10-2 gm.-atom per litre. (1') = 8.37/5.15.10-2 gm.-atom per litre. we obtain Also if the original iodide concentration be x and the final .z - a, I,' = 3a I = x - 2a, = 2.875 x I' = x - a and (x - 2a)(x - a)/% = 0.00611 . . . . . (1) * where x = (4.5 - 8.37/5.15)10-2 gm.-atom per litre gm.-atom per litre. Inserting this value of x in (I) and solving for a we obtain a = 0.58 x gm.-atom per litre. Hence (1') = (2.575 - 0.58j x 10-2 = 2.30 x 10-2 gm.-atom per litre.(1.625 x 10-2)(1*625 x 10-2)a ~ 23.6, (0.380 x 10-2)(2-30 x Hence K25" = value in close agreement with that of Bronsted and Pedersen. The slight discrepancy between 23.6 and 21.1 for the values of K in the two investigations is probably accounted for by the fact that Bronsted and Pedersen utilised only chlorides and iodides of iron and potassium whilst in the present case ammonium sulphate was present in addition. Xummary. The reaction 2Fe"'+ I ~2 2Fe"+ 21' is photosensitlive to both ultra-violet and visible light. The region of visible photoactive radiation is within the range 5500-6500A. with an apparen SHORT THE CONDENS-4TION OF PHENYLETHPLAMINE ETC. 269 maximum a t 58008. The tri-iodide ion is the photoactive constituent and the energy of excitation corresponding to a wave-length 5800 A. is equivalent to 2.14 volts a value almost identical with the resonance potential of the iodine molecule. It is found that one quantum of absorbed radiant energy causes 1 mol. of iodine to react; the mechanism of the reaction can be expressed in the 1-for111 2Fe1 + I,' -/- hv 1 2FeT3 + 1'. The dark eyuilil:)r.ium coils1 ant of the reaction is K = 23-6 = (Fe"')(T2}~/(Fe"')(I') a t 23". The temperature coefficient of the liberation of iodine in the dark is k,; ,ikZ5 - 2.723 whilst that of the photochemical reaction is ,'Idz5 = 1.17. The addition of potassium chloride raises the theimodynaiuic concentrations of all the reactants and although the equilibrium point is unchanged the rate of attainment is very considerably increased. Proof is given that equilibrium is the result of attainment of equal velocities of the forward and the back reaction since the effect of radiation of constant intensity on the equilibrium attained is less when the thermodynamic concentrat ions of the reactants (and thus the reaction velocities of the two reactions} are raised by the addition of potassium chloride. Our thanks are due to the Department of Scientific and Industrial Research for a grant out of which a part of the cost of the apparatus was provided. IA A u OR -i TO RT o 17 P HITS I ( .A 1 C H F JI r s-rR\-, G'AMI3RTI)C~E. [ R P c P ~ u P ~ ' Ncrc~t-r~~bcr 1&h 1024.