MANSEL DAVIESAND EMRI'SGWYNNB
2i48
aqueous solutions, and thus only half as many equivalents of CaHd would be available for reduction as a result of the proposed disproportionation, It is seen (Fig. 2) that a t higher acidity p-chlorovinylphenyliodonium chloride is not completely reduced. Instead, the solvent decomposition (hydrogen wave) merely precedes and prevents formation of waves I11 and IV. Quantitative Analysis of Iodonium Salts.Diffusion currents of all the reduction waves were found to be proportional to concentration when tested between 5 X 1M and 1 X M in 50% alcohol as a solvent. The solubility of the diphenyliodonium salts limits the analytical determinations to concentrations in the region of 0.001 AI. Acidities in the region pH 6-7 are recommended since higher acidities do not allow full development of all waves, and basic solutions decompose iodonium salts. Maxima in any of the waves are not particularly large or troublesome, see Fig. 3. These niaxima are completely absent a t iodonium salt concentrations around 0.001
[CONTRIBCTIOS F R O M
1.01. i 4
M . At higher concentrations, the maxima are easily eliminated by the addition of small quantities of gelatin or long chain quaternary ammonium salts as suppressors. Gelatin is recommended since concentrations slightly in excess of the "Maximum Suppression Point" (M.S.P.)9t21 do n o t alter the polarographic results. The use of, for example, lauryl trimethylammonium bromide introduces complications. The minimum concentration of quaternary ammonium salt necessary to eliminate the maximum in wave 111 (its M.S.P. valuez1) shifts the El/, value of.wave IV to an erroneous smaller negative voltage. Apparently, the KHg* free radicals, postulated in the reduction, result in another example of substantivity causing an anomalous"-21maxima suppression behavior. Acknowledgment.-The support of this investigation by the Research Corporation is acknowledged with appreciation. (21) E. L Colichrnan, Tws
JOI-RNAI.,
73, 4036 (1050).
PORTLAKD, OREGON
T H E EDWARD DAVIESLABORATORIES, THE GXIVERSITY
C(OLLEC7E O F \VALES, ABERYSTWYTH]
The Iodine-Iodide Interaction BY MANSEL DAVIESAND EMXYS GWYNNE RECEIVED JUXE 8, 1951 The well-known equilibria in dilute aqueous iodide-iodine solutions have been re-examined systematically a t a number of temperatures by the partition technique. The increase in heat content (-8) associated with the formation of the triiodide ion is found to be markedly temperature dependent. The next detectable step in polyiodide formation is the occurrence of 21,- e I,-. The structure and binding in the triiodide ion are discussed on the basis of an electrostatic model.
+
The equilibrium IIz$ IChas been one of the Experimental most frequently studied of ion-molecule interacDistribution measurements between aqueous and carbon tions. The well-known distribution method was tetrachloride phases a t 25.00 f 0.02,38.38 It 0.03, 49.65 f systematically applied to its study a t 23' by 0.05, 63.05 f 0.05' form the basis of the results presented. . R. chemicals, appropriately dried when used as volumetJakowkin'; his results and those of many later A ric standards, were used throughout the work, together with workers serve to show that the process is not con- calibrated weights and carefully standardized volumetric fined to the triiodide stage. Most of the numerous apparatus. In view of the numerous previous studies of this determinations of the equilibrium constant (K3) for system, between which marked discrepancies are often found (see later), the experimental errors in our methods the triiodide formation have been made at 25' and were evaluated a t each stage. Equilibrium was established few explicit evaluations of the corresponding in- in an all-glass cell of the form sketched by Brown and Burys crease in heat content (m3)have been reported; and samples of the layers were displaced by air pressure with in fact Dawson's?"and Jones and Kaplan'sZbpapers precautions against loss of volatile iodine. Such samples for analysis were weighed and concentrations converted to are exceptional in quoting Kg values a t two tem- voliime units largely on the basis of recorded densities6 and, perature~.~ Although these authors did not calcu- for aqueous K I + I p solutions a t temperatures other than late AH3, this has been done by M~elwyn-Hughes.~25", from oiir owii determinations of a few points over the On the practical side, the purpose of the present ranges involved. Frcquently standardized thiosulfate was for the 1 2 cstimations. The aqueous phase was throughwork was to make a systematic tleterininatioii of used out made 0.0010 'iA in M?S04;this concentration o f acid 11113 and also an examination of what further equi- sufficing t o ensure that the hydrolysis of I2 to 1310 is neglibria play a significant role in dilute solutions. For ligible.7 It is worth mentioning that the presence of the led to significantly better reproducibility in the discomparison with the former, a calculation of the acid for iodine itself between the two solvents. tribution ratio (KI) expected net interaction energy arising principally In each run with the potassium iodide solutions dilution was continued until the total titratable iodine reached the from an ion-induced dipole term has been made. (1) A. A. Jakowkin, 2. ghysik. Chcm., 20, 19 (1896). (2) (a) H. M. Dawson. J . Chem. Sor , 238 (1901); (b) C, Jones and R . B. Kaplan, THISJ O U R N A L , 60, 1845 (1928). (3) See, honever, A . D Awtrey and R. E. Connick, ibitd , '73, 1842 (3961). i?) E. A. Moelwyn-Hughes, "The Kinetics of Reactions 1 1 1 5iolutiou," 2 n d Edition, Oxford University Press, 1917, p. 191.
lower____ limit for accurate determination: accepting a maxi-
F 3 Iiroivn nnrl C. R. Rary, Trans. Chcm. Soc., 123, 2430 ( 1923 J (6) "I,an~olt-IlBrnstein," 5th Edition, Vol. I, p. 430; "I.C.T.," Vu1 111. / I 1:+2 ( 7 ) \V C Bray and G . hI. J . hIacKay, TIZISJ O U R N A L , 33, 932 (1010), 33, 1185 (1911).
June 5, 1952
IODINE-IODIDE INTERACTION
degree of agreement in the Ka values as [I2]tends to zero Over forty-fold concentration range of iodide. To evaluate AHl, ;.e., the heat content change
mum uncertainty of 0.15%, this concentration was about 1.5 X molar. This procedure led to a minimum extrapolation in the plots of Kr against free iodine concentrations to zero concentration.
than f 0 . 5 unit. Regular variations of K1 with concentration of the form shown, and of the same order, appear a t other temperatures. Of previous tainly amongst the most carefully determined, although he did not suppress the hydrolysis. At the lowest concentrations measured by him the hydrolysis of the aqueous I 2 exceeds 1%: with 0.0010 M HzS04 this is reduced to less than 0.10%. Whilst Jakowkin’s figures show a concentration dependence of the same type as ours, they are in the 88.‘ Bray and MacKay9 chose a region 85 value K1 = 90, although estimates as low as 82.6 have been reported.1° The ratio of the solubilities in the pure solvents a t 25’ is quoted as 89.6.11 The small but real variation of K1 with concentration arises, of course, from the non-ideality of the iodine solutions. The general features of iodine solutions have formed a particular topic in a number of Hildebrand’s studies, l2 and the K1 variation, whose form could be “explained” by an association of the iodine in the carbon tetrachloride (cf. Hildebrand and Jenks), has been discussed explicitly by Hildebrand and Scott.l 1 The solubility relations of iodine in the individual solvents lead to a confident prediction of an increase in K1with increasing concentration. In calculating the equilibrium constant Ka, we have used the K1 value appropriate to the carbon tetrachloride concentration of iodine. As a check on the influence of ionic strength upon the activity of aqueous iodine, K1 was determined a t 25’ in presence of potassium nitrate (a neutral salt unlikely to interact specifically with the halogen) up t o the maximum ionic strength subsequently encountered. For C, = 4.8.10-2 M , and (KNOs] = 1.00 M , a value K1 = 90.5 == I 0.3 was found, showing no real deviation from the value with respect to water (90.6) a t the same concentration. This absence of a general salt effect on the iodine activity agrees with numerous previous indications for the present and similar systems provided the aqueous halogen concentration is below 0.1 M.’ It is given further considerable support by the
-
(8) C , . A Linhart. i b i d . , 40, 158 (1918). (9) !A.’ C Bray and G. M J. MacKay, i b i d . , 32, 1207 (1810). (10) J. N. Pearce and W. G. Eversole, J. Phyr. Chcm., 28, 245 f 1924)
(11) J. H. Hildebrand and R. L. Scott, “The Solubility of Nonlllectrolytes,” 3rd Edition, Reinhold Publishing Corp., New York, N. Y., 1950, p. 208. (12) (a) J. H. Hildebrand and C. A. Jenks, THISJOURNAL, 42, 2180 (1920); (b) J. H. Hildebrand, ibid., 51, 66 (1929): ( c ) 0.R. Negisbi, L. H. Donnally and J. H. Hildebrand, ibid., 5& 4793 (1933).
2749
8
92
0
’
’
’
n ~~
2 4 6 8 [I21 ccl4 in units of 10-2 g. mole. per liter. Fig. 1.-A, present values; El, Linhart’s values.
10
in the heats of solution of iodine in the two solvents a t (ideally) zero concentration. From solubility datal3 the limiting values of the heats of solution a t saturation (for the same mean temperature range as above) were evaluated as I 2 (solid) I 2 (CClr soln.), A H = $6920 f 380 cal. per mole
12 (solid)
12
(aq. soln.), A H = $6150 f 100 cal. per mole
The difference of these figures, i.e., +770 f 480 cal. per mole provides adequate agreement with the directly determined AH1. The Triiodide Equilibrium.-The constant for this is taken in the form
Although Lewis and Randall14 were of the opinion that the activity coefficient f1,- is less than fI-, Jones and cite sufficient reasons for assuming them to be equal. A s R has been shown not to vary with ionic strength over the range in which we are interested, i t is taken as unity in the following calculations. At 25” the distribution of iodine between CCL and approximately 1.0, 0.10 and 0.025 M K I solutions was measured a t a series of iodine concentrations. For the other temperatures only two K I concentrations were used, 0.025 M IC1 being omitted a t 38 and 49”) and 1.0 M K I a t 68’. Table I summarizes the data for 25.00 f 0.02O, the terms being calculated on the assumption that 1 3 - formation is the only interaction occurring ; the units are moles (or g.-ions) per liter. The data a t other temperatures are shown in Fig. 2. It is seen that there are marked systematic trends in Kat but that as [Iz] tends to zero the various series a t any one temperature converge t o a common value. This means that, as antici(13) A. Seidell, “Solubilities of Inorganic and Metal Organic Compounds,” 3rd Edition, D. Van Nostrand C o . , Inc.. New York, N. Y . ,
1941.
(14) G. N. Lewis and M. Randall, “Thermodynamics.” McGrawHill Book Co., Inc., New York, N. Y.,1823, p. 636.
"7n0
[I
1
+ [I1
1.069 1.059 1.059 1.039 1.032 1.014 1 ,008 1.004 0 . 1010 .lo02 ,1001 . lON3 ,1000 .loo0 0.02482 ,02482 .02482 ,02482 .02482 .(I2482
800
,.ABL LE I 1
[rri x
Pi 1
K?
A 766
0 4917
12x7
4 167 3 964 2 520 1 991 1 212 0 594 0 230 8 120 1 807 0 920 471 20,5 1Od 7 794 7 682 5 202 2 073 0 891 0 4.74
3300 3129 2016 1577 0960 0456 0178 0 03951
1086 1058 954.1 906.1 862.8 798.1 783.6 791.1
in$
0125J
.00660 00348 00 153 00078 U 009271 009233 007068 003338 .(I01595 00078
