Journal o f Solution Chemistly, Vol. 13, No.

9, 1984

Triiodide Ion Formation Equilibrium and Activity

Coefficients in Aqueous Solution
Donald A. Palmer, 'R. W. Ramette, 2and
R. E. Mesmer '
Received March 15, 1984

The equilibrium quotient .for the formation o f triiodide was studied as a .function
o f temperature, 3.8-209.0~ and ionic strength, 0.02-6.61. The best-fit value
for the molal equilibrium constant at 2 5 ~ is 698+--10 and the corresponding
partial molal enthatphy, entropy, and heat capacity o f formation are: A H ~ =
-17.0+0.6 kJ-mo1-1, A S ~ = -0.6+0.3 J - K l - m o l -|, and A C ~ = - 2 l + 8
J-Kl-molL. Activity coefficients o f iodine were determined as a ./'unction o f
ionic strength (NaClO 4) at 2 5 ~ and conclusions are drawn as to the car-
responding ionic strength dependence r the triiodide anion.

KEY WORDS: Iodine; triiodide; aqueous; thermodynamics; solubility.

1. INTRODUCTION
The equilibrium involving the formation of triiodide is
represented by

I2(aq) + I-(aq) = l~(aq) (1)

The data summarized in Table I clearly illustrate how intensively this
reaction has been studied in the past and that the equilibrium quotients
are remarkably consistent regardless of the technique used in their
determination. A closer inspection of these results reveals that some-
what higher QI (the subscript indicates the equation describing the
reactions involved) were obtained when higher concentrations of iodine
and/or iodide were employed. Such deviations can be attributed to the

1Chemistry Division, Oak Ridge National Laboratory, Oak Ridge, TN 37831.

2Carleton College, Northfield, MN 55057.

673
0095~9782/84/0900-0673503.50/0 9 1984 Plenum Publishing Corporation
674 Palmer, Ramette, and Mesmer

Table I. Summary of Literature Molar Thermodynamic Data for

Reaction (1) Q

[I-]total 104112] QI AH ACp, 1

Year M M 25 ~ kJ-mol l J-Kl-mo1-1 ~ Method Ref.

1896 0.06 b'c 716 dist. d 1

1906 0.10 b 726 dist. 2
t910 0.02 b 13.2 e 730 cond. 3
1928 0.02 < 2 705 dist. 4
< 0.1 < 4,4 720 dist. 4
1951 0.002 <0,01 708 -21.1 68 1-39 spect. 5
1952 < 1.0 0/ 768 -16.4 148 25-63 dist. 6
1955 <0.07 b 13.2 e 766 20-30 solub. 7
1960 <0.006 <0.5 721 -16.5 -254 15-35 spect. 8
1963 0.05 <6.7 727 -18.4 1 0-40 dist. 9
1965 0.01 13.3 e 748 -17.2 -35 15-35 solub. 10
1973 <0.0001 < 13.3 e 719 -15.5 10 0-56 spect. II
1973 770 -18.3 106 10-80 dist. 12

aQ, AH, and ACp, the thermodynamic parameters tbr Eq. (1) were calculated from a
least-squares fit ol the observed K 1 vs. temperature data using the generalized equation,
b c
pK I = A / T + B + C log T. Selected data taken from this reference. No acid added
to suppress hydrolysis, dDistribution of iodine between aqueous CCI 4 phases.
eSolubility limit of iodine taken from Ref. 10. IQ vs. [121 data were extrapolated to zero
free [12].

presence of other polyiodine species of the type I;, I42-, I62-, etc.
During the early stages of our investigation of the disproportiona-
tion reaction

312(aq) + 3H20 = IO;(aq) + 5I-(aq) + 6H+(aq) (2)

to temperatures in excess of 200~ the need for accurate ther-

modynamic data pertaining to Eq. (1) at these conditions became ap-
parent. The enthalpy change for the reaction in Eq. (1) AH1 given in
Table I are sufficiently varied so as to introduce considerable uncer-
tainties into the values of Q1 at the higher temperatures. Moreover, an
extension of the currently available temperature range should lead to a
more precise determination of the heat capacity change for Reaction
(1) ACp,I which would otherwise have to be set at zero in view of the
scatter shown in Table I. On the other hand, reactions in which the
ions of the same charge are on both sides of the equation are generally
Triiodide Ion Formation Equilibrium 675

believed to have small and constant absolute values for their heat
capacities of reaction. Empirically it is found that when hydrolysis reac-
tions of weak acids are written in their basic or anionic forms to obviate
the change in charge, the absolute magnitude of z~Cp.~ is significantly
reduced and is rendered nearly independent of temperature. ~3~4~Thus,
Reaction (1) presents an ideal opportunity to test this idea, in particular
because the ions involved appear more similar than in the case of
hydrolysis reactions where the behavior of hydroxide ions is equated to
that of quite different anions.

2. EXPERIMENTAL

2.1. Materials
Commercial, sublimed iodine was resublimed prior to use. Potas-
sium iodide was dried at 110 ~ under vacuo for about 24 hours, while
the water used to prepare the solutions was deionized, and then doubly
distilled.

2.2. Methods
(1) Spectrophotometric measurements were made using a one cm
cell with a Pye-Unicam PUS800 instrument at the wavelengths 370,
350, and 288 nm. The stock solutions were typically (a) 104 molal (m)
iodine in 0.02m KI and 10-4rrt HCIO4, (b) 0.02rn KI with 10gin HC104,
and (c) 0.02m NaC104, likewise 104m in HC104. In 100 g of solution
containing 5 g of stock (a), the respective weights of stock solutions
(b) and (c) were varied to give the required range of iodide concentra-
tions and a constant ionic strength of 0.02. The iodine concentration in
stock solution (a) was checked continuously during each series of
measurements by controlled potential coulometry (Princeton Applied
Research, Model 371 potentiostat and Model 379 coulometer). Using a
platinum gauze electrode, both the iodine and iodide concentrations
were determined to within +0.1%. (~5~ At least nine concentrations of
iodide were measured to establish a mean value of K~. The molar ab-
sorptivities of all the species are listed in Table II. The spectrum of the
triodide ion continued to change with increasing iodide concentration
indicating the presence of other polymeric forms of iodine as previously
observed by Chia. (f6~ Finally, above 45~ the absorbance decreased
with time, presumably due to the loss of iodine from the stoppered
cell, hence imposing an upper limit on the applicability of this method.
676 Palmer, Ramette, and Mesmer

Table II. Molar Absorptivity of Iodine Species at 25 0(2 o

Wavelength (nm)
Species 370 350 288 Ref.

12 31.7 16.1 83.5 This Work

18. 95. 5
l- 0,009 0.011 0.057 This Work
13 18980 25750 38790 This Work
26400 40000 5

a Units: l-moll-cm -1.

(2) Distribution measurements were made in which iodine vapor

was allowed to partition between two stirred solutions in separate
chambers of a closed system. One compartment contained an acidified
iodide solution and the other an acidic iodate solution of matching ionic
strength. After equilibrium was established, which required from
several days at room temperature to only hours at temperatures near
200~ both solutions were sampled. The total iodine (I2 + I3) and
iodide (I- + I3) concentrations in the former compartment were deter-
mined by coulometry. The I2 content of the latter solution was
measured spectrophotometrically at 460 nm where the molar ab-
sorptivity of iodine is 730.9 1-mol-Lcm ~. Note that iodate was present
merely to eliminate interference from possible iodide impurities, as dic-
tated by Eq. (2). Perchloric acid was added in order to minimize iodine
hydrolysis.
For temperatures from 0 to 100~ the outer vessel was a one
liter PFA Teflon flask (Savillex Corp.) which could be immersed under
argon pressure in a bath controlled to - 0 . 1 ~ and sampled through
minivalues. (17~ The inner compartment consisted of a 150 cm 3 Pyrex
beaker mounted on legs. At higher temperatures, a commercially pure
titanium pressure vessel was used. The vessel was seated in a Marshall
tube furnace with temperature control to within ---0.2 ~ In this case,
solutions were sampled through Teflon tipped titanium tubing sealed in
the cold zone by titanium valves. It is interesting to note that slow
reduction of iodate was observed at the high temperatures, as welt as
significant oxidation of iodide. The presence of perchloric acid was
later found to be responsible for the oxidation reaction. Although both
effects did not influence the results below 210 ~ at higher tempera-
tures these reactions became too fast to allow any meaningful measure-
ments to be made.
Triiodide Ion Formation Equilibrium 677

(3) One experiment was carried out at 25 ~ using the more con-
ventional distribution method which involved partitioning iodine
between acidic aqueous and carbon tetrachloride phases. This experi-
ment was conducted in sealed centrifuge tubes which were equilibrated
in a rotating thermostat controlled to --0.1 ~ The solutions were
analyzed by coulometry, including the carbon tetrachloride portion
which required longer times for complete reduction of its iodine
content, but the result was nevertheless equally reproducible.
Table III. Molal Equilibrium Constants for Reaction (1) Determined
by Spectrophotometry a

Kt
~ 370 n m 350 n m 288 n m

15.3 841 - 7 8 4 0 +- 7 843-+ 6

25.0 690-+ 1 687-+ 2 6 8 9 +- I
35.0 531-+7 527+7 541-+7
44.7 443-+ 4 4 4 0 +- 2 452-+ 3

alonic strength = 0.02.

3. RESULTS AND DISCUSSION

The values of the K t given in Table III were determined

spectrophotometrically as outlined above. The use of the term equi-
librium constant rather than equilibrium quotient Q, at these low ionic
strengths will be justified later in the text. Good consistency was ob-
tained among the K t values at the three measured wavelengths. Their
mean values at each temperature are included in Table IV together
with the results of the vapor transfer experiments. The combined data
are plotted in Fig. 1 which also shows the line resulting from a least-
squares fit (~8/according to Eq. (4),

logK~ = A / T + B + ClogT (4)

where T is the absolute temperature. The values of the parameters

determined in this manner are A = 555.0, B = 7.355, and C = -2.575.
The resulting change in the partial molal quantities for Reaction (1)
and their uncertainties as calculated by propagation of the errors in the
three parameters are as follows: AH, ~ -17.0--0.6 kJ-mol", AS1~
678 Palmer, Ramette, and Mesmer

Table IV. Molal Equilibrium Constants for Triiodide Formation

oC 104 [mj2lb mtlOtat t, K1 K 1 (calc) c

3.8 4,4 0.051 1160 1170

15,3 0.5 0.02 841 a 878
25,0 0.5 0.02 689 a 698
25,0 6, 0.02 701 698
35.0 0,5 0.02 533 a 558
35.1 6.9 0.050 550 556
44.7 05 0.02 445 a 454
50.2 4.0 0.010 409 405
50.3 7.1 0,050 432 405
50.3 8.7 0.050 423 405
75.2 4.7 0,010 255 252
75.5 11 0.050 244 251
98.8 13 0.050 174 167
100.1 8.5 0.049 159 165
100.3 5.1 :0.0093 166 165
126.3 10 0.048 116 Ill
152.6 l3 0.048 69.4 77,1
152,6 7.3 0.047 74.5 77.1
178.6 | 1 0.045 54.7 55,7
208.8 16 0,043 :37.2 39.5
209.9 12 0.044 45.1 39.4

a Spectrophotometric data the remainder being from distribution studies, bunits of

molality, m o l - k g . CCalculated from Eq. (4).

-0.6• J-Kt-mol -t, and ACo~ = -21• J-K-t-mo1-1 at 25~ The

partial molal enthalpy and heat capacity changes are in agreement with
those reported by Ramette and Sandford, (t~ despite the narrow tem-
perature range investigated.
More significantly, the small and constant value of A C~,~ is in
accord with the hypothesis stated earlier that with ions of the same sign
on both sides of an equation, the associated change in heat capacity is
small and nearly independent of temperature. For example, the AC~~
for the hydrolysis of CO2, "4'~9~when written in its basic or anionic form

CO2 + OH- = HCO3 (5)

is -110• J-Kl-mol -~ at 25 ~ which is considerably less, in absolute

terms, than the corresponding values for the acidic form
 Triiodide Ion Formation Equilibrium 679

, r ; , I ' ; ' ' I ' ' '' ' I '

3.0

2.0

1.5 , , , , I , , i , I J ~ J , 1

2.0 3.5 3.0 3.5

10:3/'17
Fig. 1. Plot of the log of the equilibrium constant for Reaction (1) vs. l / T
plot showing the best-fit line obtained from Eq. (4).

CO2 + H20 = H § + HCO~ (6)

for which ~C0 ~ = -340-+30 at 25~ -430-+40 at 200~ and -1300

J-KLmol 1 at 300 ~ Furthermore, it is significant that ~Cp ~ for Reac-
tion (1) is substantially less than that for Reaction (5) reflecting pos-
sibly the greater similarity between the solvation of anions I- and I3, at
very low ionic strengths (see discussion below), than that between the
OH- and HCO~ ions.
The effect of ionic strength on the equilibrium of Reaction (1) is
slight as exemplified by the Q, values in column 4 of Table V. They
were obtained using method (2), which describes the vapor transfer ex-
periments, except in this case the solutions were saturated with iodine.
Note that the solubility of iodine in 0.02m NaC104 (0.0005m [H+]) is
only slightly lower than the generally accepted values reported in pure
water at 25 ~ typically 1.326x 103M. u~ Since Q, is virtually independ-
ent of ionic strength it is essentially equivalent to K~ at ionic strengths
< 0.05. At first glance, one might infer that with increasing ionic
strength the activity coefficients of the anions remain similar, while
3' (I2) remains near unity. On the contrary, the activity coefficient of
680 Palmer, Ramette, and Mesmer

Table V. Molal Equilibrium Quotients and Activity Coefficients

for Reaction (1) in Sodium Perchlorate Media at 25 ~

'y (13 ) y m(Nal 3)

1~2 total 3 a
u m r- 10 m12 Q1 y(l?) 7(1-) y_+(Nal) y+_(Nal) y+_(Nal 3)

0,020 2.011 1.315: 744.7 1.000

0,514 2.063 1.177 723.5 1.117 1.150 0.698 1,072 0.75
1,05 2.081 1,023 719.3 1.286 1.331 0.684 1.154 0.79
2.19 2,206 0.7851 746.3 1.675 1.672 0.722 1.293 0.93
3.50 2,332 0.6149 762.4 2.139 2.089 0,810 1.445 1.17
4.98 2.500 0.4523 794.6 2.908 2.725 0.954 1,651 1.58
6.61 2.662 0.3393 762,2 3.876 3,787 1.167 1.946 2.27

a Molality of saturated iodine solution.

iodine, calculated from the solubility data in Table V and shown in

column 5, varies markedly with ionic strength. This is shown graphi-
cally in a form of the Setschenow plot in Fig. 2 where the dashed lines
represent the best fit for such neutral molecules as CO 2 in NaCI and
NaNO~ solutions, (2~ as well as I2 in NaNO~, ~23~CH 4 in NaC1, ~4~ and
SiO2 in NaC1. ~25) Although there are no comparable literature data for
NaC104, 12 appears to behave like CO2 in NaCI solutions. The impor-
tant conclusion is that the ratio of the individual ionic activity coef-
ficients, 3' (I;)/3' (I-) in the perchlorate media shown in column 6 of
Table V, varies dramatically with increasing ionic strength, thereby of-
fsetting the increase in y (I2). That these two effects almost exactly
cancel presumably indicates that the components of I-I-I- (namely I-!
and I-) behave as if they were independent of each other within the
triiodide anion. In order to assess the magnitude of 7 • (KI3) from this
ratio, the mean activity coefficient of KI in NaC104 must be known, i.e.

5' +_(KI3) = [K13` (I2) / Q1]'/~y_~(KI)

In the absence of these data, the ratio

3` _. (NaI 3)/3` _+(NaI) = 3` _+(KI 3)/3` + (KI)

at the trace concentrations of K +, I;, and I- employed in these experi-

ments, can be calculated instead (Table V, column 8). The values of
3' _+(NaI) shown in column 7 were calculated from Pitzer's ionic inter-
Triiodide Ion Formation Equilibrium 681

1.5 I I ] r I 1

/ /~f//"

1.0
.......-'...'"
#.
0.5
//'" ...'S"~ ..... "'-(e) /m"

0.0

-0.5 i I i I i I i I ~ I J 1 i
0 1 2 3 4 5 6 7
I
Fig, 2 The relationship between the trace mean activity coefficients for Nal(O) and
Nal 3 ( 0 ) with the ionic strength made up of NaCIO4. For comparison, the activity coef-
ficients for 12 in NaCIQ (ml), as well as : (a) CO 2 in NaCI, Refs. 20-22; (b) SiO2 in
NaCI, Ref. 25; (c) CO2.in NaNO 3, Refs. 20-22; (d) 12 in NaNO 3, Ref. 23; and (e) CH 4 in
NaCI, Ref. 24.

action model (26'27) using the mean activity coefficients of NaI and
NaC104 listed by Robinson and Stokes r and by assuming the mixing
parameters 0 and 9 are zero. The latter assumption is unlikely to lead
to significant errors, particularly as substituting the corresponding
parameters for the mixture NaC1 + NaNO3, (0.016 and -0.006,
respectively) led to a maximum deviation in 7_+ (NaI) of + 3% over the
range of ionic strengths, 0.5 to 7.0. Thus, the mean activity coefficient
of NaI3 shown in column 9 of Table V is significantly larger than for
most common 1-1 salts, such as NaI. These results are plotted in Fig.2
to better illustrate this point.
The QI presented in Table V at 0.02m ionic strength is marginally
682 Palmer, Ramette, and Mesmer

greater than the mean value of 698 (Table IV), obtained at a con-
siderably lower iodine concentration, 5 x 10-Sm. The difference is
presumably due to the presence of such higher polyiodine species as
I~. (1~ The results shown in Table I exhibit the same trend. The iodine
distribution experiment described in Method (3) resulted in a K~ value
of 728 at 25~ and at the intermediate iodine concentration of
5.536x 10~-41m. Note that the distribution cofficient for I2 between the
CC14 and H20 phases was 84.7.
It might be suggested that species such as I~ could bias the cal-
culated values o f 7+-(NaI3) presented in Table V. However, a direct
spectrophotometric m e a s u r e m e n t of Q~ at an ionic strength of 1.05
(NaCIO4) gave a value of 6 5 1 • at an iodine concentration low
enough that I~ can be neglected. T h e corresponding value for
~ / , ( N a I 3 / 3 , + ( N a I ) is then 1.173, very similar to 1.154 given in Table
V. In addition, the formation equilibrium quotient for I-~ is expected to
be unaffected by ionic strength as is Qv

ACKNOWLEDGMENT

Financial support for this work was provided by the Division of

Chemical Sciences, Office of Basic Energy Sciences, U.S. D e p a r t m e n t
of Energy, under contract W-7405-eng-26 with the Union Carbide Cor-
poration. R,W,R. acknowledges the financial support of the Oak Ridge
Science Semester of the G L C A / A C M consortia. The authors are also
indebted to H. Holmes for his invaluable discussions and assistance
with the ion interaction treatment of electrolytes.

REFERENCES
1. A. A. Jakowkin, Z. Physik. Chem. 20, 19 (1896).
2. W. Maitland, Z, Elekttvchem. 12, 263 (1906).
3. W. C. Bray and M. J. Mackay, J. Am. Chem. Soc. 32, 914 (910).
4. G. Jones and B. B. Kaplan, J. Am. Chem. Soc. SO, 2066 (1928).
5. A. D. Awtrey and R. E. Connick, J. Am. Chem. Soc. 73, 1842 (1951).
6. M. Davies and E. Gwynne, J. Am. Chem. Soc. 74, 2748 (1952).
7. L. 1. Katzin and E. Gebert, J. Am. Chem. Soc. 77, 5814 (1955).
8. G. Daniele, Gaz. Chim. Ital. 90, 1068 (1960).
9. E. N. Rengevich and E. A. Shilov, Ukr, Khim. Zh. 28, 1080 (1962).
10. R. W. Ramette and R. W. Sandford, Jr., J. Am. Chem. Soc. 87, 5001 (1965).
11. J. D. Burger and H. A. Liebhafsky, Anal. Chem. 45, 600 (1973).
12. W. Eguchi, M. Adachi, and M. Yoneda, J. Chem. En,~,,. Jpn. 6, 389 (1973).
13. C. F. Baes, Jr. and R. E. Mesmer, Am. J. Sci, 281, 935 (1981).
14. C. S. Patterson, G. H. Slocum, R. H. Busey, and R. E. Mesmer, Geochim.
Triiodide Ion Formation Equilibrium 683

Cosmochim, Acta 46, 1653 (1982).

15. R. W. Ramette and D. A. Palmer, J. Solution Chem. 13, 637 (1984).
16. Y.-t. Chia, Ph.D. Thesis, University of Calilbrnia, Berkeley (1958).
17. D. A. Palmer, R. W. Ramette, and R. E. Mesmer, J. Solutioi~ Chem. 13, 685
(1984).
18. W. R. Busing and H. A. Levy, ORGLS, A Geimral FORTRAN Least Squares
Program, ORNL-TM-271, Oak Ridge National Laboratory (1962).
19. R. H. Busey and R. E. Mesmer, J. Chem. Ei~g. Data 23, 175 (1978).
20. A. E. Markham and K. A. Kobe, J. Am. Chem. Soc. 63, 449 (1941).
2[. K. Onda, E. Sada, T. Kobayashi, S. Kito, and K, lto, J. Chem. E~Tg. Data 3,
18 (1970).
22. A. Yasunishi and F. Yoshida, J. Chem. EiTg. Data 24, 11 (I979).
23. A. V. Kiss and A. Urmanczy, Z. Aaorg. All~em. Chem. 2D2, 172 (1931).
24. R. K. Stoessell and P. A. Byrne, Geochhn. Cosmochim. Acta 46, 1327 (1982).
25. W. L. Marshall and J. M. Warakomski, Geochim. Cosmoch#n. Acta 44, 915
(1980).
26. K. S. Pitzer and G. Mayorga, J. Phys. Chem. 77, 2300 (1973).
27. K. S. Pitzer and J. J. Kim, J. Am. Chem. Soc. 96, 570l (1974).
28. R. A. Robinson and R. H. Stokes, Electrolyte Solutiom, 2rid edn.,
(Butterworths, London, 1959), p. 492.