seven days at room temperature, and gave a negative test for peroxide; however, when heated a t 170' for an hour and a half, the red solid decomposed to a mixture of white and yellow solids which gave the tests that would be expected of cesium superoxide. Both the potassium and cesium extracts decompose violently in water, with flashes of light accompanying the decomposition. Analyses of the extracted potassium product indicate that the red solid contains about 5,iC;; KaO and 45% available oxygen, i. e., oxygen liberated upon complete decomposition, as compared with the values 34.1 and 45.9% calculdted for the decomposition of KO3. This fact would al'pear tn substantiate to a large extent the findings of Kazarnovskii, h'ikolskii and Abletsova6 that the red solid is composed primarily of a substance having the formula Koa; hence, it should be termed potassium ozonide. In addition, thermal decomposition of the material a t 50' for one hour liberates approximately the quantity of oxygen calculated for the decomposition of KO3 to KO2 (found, 1 (LO; calcd., 13.4). lfeasurement of the oxygen and iodine liberated when acidified potassium iodide solution is used as a deconiposant indicates that the oxygeniodine ratio is about 2.5 : 1. The solution remaining after aqueous decomposition of the red solid retains 3 very weak oxidizing power. Inasmuch as the behavior of this red solid differs from that of the sodium ozonate in that the latter is incapable of oxidizing acidified potassium iodide, considerable doubt exists that the potassium estrdct is a single chemical species. Analyses of the extracted cesium products proved very inconclusive. The per cent. of Cs2O was much higher than that calculated on the assumption that the red cesium solid was CsOl. As in the case of the extracted potassium material, the results of oxygen-iodine determinations showed the oxygen-iodine ratio to be approximately 2.5: 1. However, in this instance the oxidizing power displayed by the solid was in no way affected by aque-
[CONTRIBUTION KO.
89
FROM THE INSTITUTE FOR
ous or acidic decomposition ; the resulting solutions displayed the same amount of oxidizing action toward iodide ion as did the solid material. In this respect the behavior of the extracted potassium and cesium materials differ considerably. These observations would seem to lend weight to the thesis that the extracted material is not a single chemical species of the formula CsOa, particularly in view of the fact that sodium ozonate does not oxidize iodide ion. Acknowledgment.-The authors are indebted i o Professor Arthur n'. Davidson for his critical reading of this manuscript.
Summary Ozone-treated sodium hydroxide decomposes in water with evolution of oxygen, but does not liberate iodine from acidified potassium iodide solution. Ozonated potassium and cesium hydroxides also evolve oxygen upon decomposition in water. hut liberate iodine from acidified iodide solution. -1nalyses of the gross products of ozonization of sodium and cesium hydroxides produce data which are most reasonably explained by assigning the forniulas NaOs and Cs03 to the colored, paramagnetic materials. 'The colored products of ozonization of potassium and cesium hydroxides are soluble in liquid ammonia; evaporation to dryness of the resulting solutions produces red solid substances which decornpose violently in water with evolution of oxygen, and liberation of energy in the form of light. Although analysis of the material extracted from ozone-treated potassium hydroxide appears to indicate that the red solid consists mainly of the cornpound KOS, chemical behavior would seem to indicate that the extract is not a single chemical spe cies. -1nalysis of the extracted cesium material even more strongly suggests the presence of another substance. T.A!TRESCE,
ATOMICRESEARCH AND COLLEGE ]
RECEIVED JUNE 7, 1950
KANSAS
THE DEPARTMENT OF CHEMISTRY, IOTV.4 STATE
Ionization and Hydration Equilibria of Periodic Acid' BY
c. E. CROUTHAMEL, A. h l . HAYES.4ND D . S . MARTIN
Introduction Kinetic studies of the reaction between periodic acid and ethylene glycol* and other oxidation reactions involving periodate have large variations in their rates with temperature and hydrogen ion activity. X more detailed knowledge of the species present in solution should be of assistance in explaining these variations. The apparent ionization constants of periodic acid have already been evaluated a t 25°.3 No differentiation between the degree of hydration of ionic and molecular (1) Work performed in the Ames Laboratory of the Atomic Energy Commission. (2) Frederick R. Duke, THIS JOURNAL, 69, 3054 (1947). (3) C.E.Croutharnel, d a l , ibid., 71, 3031 (1949).
species was made, thus the apparent ionization cotistnnts written in more general form would be ( n ~ - ) . % l / ~ a o=
Kl'
(uRt)Sn?/Sal = K2'
(1) ( 2)
where a H + is the hydrogen ion activity, and Baa, Zal and Za2 are, respectively, the sums of the activities of undissociated forms, singly charged and doubly charged ionic species. Since the ionization of three hydrogen ions was observed, the acid was designated as the hydrated species, H5106. In the first paper2 i t was established that the absorption maximum a t 2225 A. was caused by univalent ions. Further examination of this peak under various conditions was interpreted to in
I
Jan., 1931
IONIZATION AND HYDRATION EQUIIJDRIA O F PERIODIC
ACID
83
dicate that the single, dehydrated, univalent species, 104, is responsible for this absorption, and that solutions could be adjusted so that no interfering absorption was present. This made it possible to determine the molar extinction coefficient of this ion. In an analysis of the spectrophotometric absorption i t was therefore possible to determine the concentration of this ion, a feature which permitted a detailed analysis of dehydration phenomena. The behavior in acid solutions was adequately described by the three equilibria
+ + +
HSIOs ++ H f HIIOS-; K1 HIIOc- ++ 2Hr0 1 0 4 - i KD EIrIOti- ++ H f H~IOF,-;K?
(3) (4)
(5)
The constants for these reactions have been evaluated over a temperature range of @-ioo as have the apparent constants, K,' and K2'. Experimental Materials and Instruments.-In the spectrophotometric studies a Cary Recording Spectrophotometer (Model 12) was employed. This instrument recorded directly the optipn. cal density of the solution, defined in this paper as log I o / I , Fig. 1.-The optical density a t 2225 A. of potassium where 10and I represent the incident and transmitted light intensities, respectively. A Beckman glass electrode pH periodate solutions (2.37 X lo-' molar) vs. PH a t various meter (model G) calibrated with standard buffer solutions temperatures. was used for pH measurements. Solutions used in the ultraviolet absorption measurements were prepared from reagent grade samples of potassium metaperiodate manufactured by the G. F. Smith Chemical Co. and the General Chemical Co. No appreciable iodate was detectable by titration with standard thiosulfate in acid and in neutral buffered solutions. The oxidizing capacity of a weighed sample I l ea corresponded to a purity of better than 99.9%. Standard stock solutions of periodate were prepared by dissolving accurately weighed pure potassium metaperiodate in 500 ml. of solution. For the greatest accuracy standard solutions were prepared each day. Doubly distilled water from alkaline permanganate, specific conductance of 0.9 X 10mho cm.-l, was used throughout. Methanol when used was purified by refluxing in alkali and iodine and redistilling. Dilute periodate solutions (2 X lo-' M ) in methanol were stable for a day as measured by the spectrophotometer. 2 , Temperature Control.-Absorption measurements with 05 the Cary instrument a t 32.5" and below were made by suitI ably controlling the room temperature and allowing sufficient ~ 1 ' 1 1 -_J time for the solutions and instrument to attain thermal equi- 0 0O!I lb d0 j0 4 0 50 6 0 7 0 8% 90 100 librium. Copper tanks, constructed so that they would fit To mpwoturo;C into the sample chamber next to the silica cell, were filled Fig. 2.-The optical density a t 2225 A. of potassium periowith water a t the proper temperature to act as a thermal ballast. At higher temperature the cell containing the date solutions (2 37 X lo-' molar) vs. temperature for sample was immersed in a water-bath a t the correct tem- various acid concentrations. perature until thermal equilibrium was attained. The cell was then dried and placed in the sample chamber as quickly measurements a Beckman model 1190 glass electrode and a as possible. The time needed was from ten to fifteen seconds Beckman model 1170 calomel electrode, recommended for to transfer the cell, dry and make a measurement. Cooling use over this temperature range, were used with the model rates were found to be less than 0.5-lo/min. in the tem- "G" pH meter. perature range of the measurements. Temperature fluctuaCalibration was made a t the proper temperature with tions in these measurements made therefore a negligible standard buffer solutions after thermal equilibrium was attained. These curves are given in Fig. 3 . Additional contribution to the errors. 3 . Spectrophotometric Measurements.-Potassium potentiometric titrations of paraperiodic acid with potasmetaperiodate solutions a t various pH values were scanned sium hydroxide in various methanol-water mixtures a t 25' a t different temperatures with the Cary Instrument using (see Fig. 4) were performed. 1.000-cm. silica cells. The pH adjustments were made with perchloric acid or with carbonate free sodium hydroxide. Discussion and Results Figures 1 and 2 are plots of the optical densities of various I. Spectrophotometric Studies.-Consideraperiodate solutions ( c a . 2 X lO-'M) a t several PH values and temperatures. "hen necessary, corrections were made tion of the plots of optical density a t 2225 for volume changes by using density-temperature informa- (Figs. 1 and 2 ) led t o the following observation for pure water. The optical density of pure potassium metaperiodate solutions in methanol-water mixtures were tions: (1) The optical density of solutions a t a measured a t 0". These solutions were diluted to volume pH value of GI where essentially all the periodate a t this temperature to avoid differences in the coefficient of would be present as monovalent ions, varied conexpansion caused by the varying solvent composition. siderably with temperature and approached an 4. Potentiometric Measurements.-Potentiometric titrations of paraperiodic acid with 0.1795 NohTaOH were upper limiting value a t higher temperatures. 12) The optical density approached a lower limiting For these made at temperatures of 0.5, 25.0 and 45.0
b
A.
.
Vol. 73
C. E. CROUTHAMEL. -4. AI. HAYESAND D. S. MARTIN
84
salts suggested the forms HJ06- and 1 0 4 - for the monovalent ions and this was later substantiated by thc equilibrium behavior. At a pH of 10.5 there was very little change in the optical density a t 2225 A. with temperature from 0 to 45' and Beer's law was obeyed in this region. These facts seem to indicate that dehydration involving dimerization ( i e . , as in the formation of Iz09z species) was of no importance. However, a simple dehydration of the H ~ I O G ~ to form HI05- could occur and not cause apparent __ -1 discrepancies in Beer's law a t a given temperature. Also, if the extinction coefficient of the H3IO6= 0 *e 60 e. e* r! and HIOb- ions are not appreciably different a t Fig. 3.-Potentiometric titration of aqueous periodic 2225 A., temperature variations of optical density acid solutions with sodium hydroxide a t various tem- would be very small. A study of the hydration equilibria of periodic acid in the alkaline region a t peratures. high and low temperature is now in progress in value a t low pH values and low temperatures. this Laboratory and will be reported a t a later date. The equilibria necessary to describe the observed (3) The optical density of solutions in the PH (3) and (4). It was range 10.5-12 did not change appreciably with data have been given as temperature from 0 to 50' when the volume also necessary to determine which species absorbed changes of pure water were used to correct the a t 2225 A. I t was known3 that the undissociated concentrations. (4) Beer's law was obeyed a t a acid and the divalent ion did absorb a t this wave length, however, the absorption by these species given PH and temperature. could be kept to less than 1% by proper pH adjust__ ,303ment. Analysis of the absorption curves for a potassium inetaperiodate solution (at a pH of ,530) a t two temperatures (Fig. 5) showed that a t wave lengths of 2225 A. and greater only one species was absorbing within the accuracy of measurement. (Under these conditions a t least U 99.97, of the periodate present exists as some -w singly charged ion.) At wave lengths below this :609 . value two species were absorbing as shown by the 6c€ 1isobestic point a t 2050 A. This was demonstrable .CO - ri since the ratio, optical density 2So/optical density Oo, remained constant for wave lengths of 2225 A. and greater to within the accuracy possible, ca. -_ * ___.I __ _0 8 * 5:&. The absorption was greatest a t the higher " temperature, hence this was associated with a Fig. 4 --Potentiometric titration5 of perlodic dcld n ~ t h dehydration process and the metaperiodate ion, potassium hydroxide in various methanol-water mixtures IO4-, was assigned to this absorption. 1
59.
44
m,
72
56
O l 7 W H NOO"
(e),
f'
100
. _ i _
4
I)
0
2
rn
.i
#I
00
KO"
IO
10
12
1.
25': 0 , aq. s o h 2 5 " , 0 , 8570 CHyOH by 40% CH30H by V O ~ dt
VOI
, m,
To interpret these observed facts, i t was first necessary to assume that absorption of light a t a wave length of 2225 A. corresponds to an electron transition and should be characteristic of a given ionic or molecular species. The energy involved in the absorption of light of this wave length is so large that moderate temperature changes (i.e., 0-100') should have virtually no effect upon the population statistics of this process, i e . , the extinction coefficient of a species should not be a function of temperature in the range observed. I n such a case changes in optical density with temperature would be caused by a change o f species. Subsequent experiments confirmed the validity of this assumption. The original choice of species was governed by the known chemistry of periodic acid and its salts. The known polybasicity of the acid and the fact that vacuum drying a t 100' is necessary to prepare the meta acid led to the choice of HbIOe as the undissociated acid. The well known para- and metaperiodate
LI
-Z
o
I
I
I
o
T
k
I
x
7
1
~
0
0
Ilqlfron.
Fig. 5.-The optical density of a potassium periodate solution (PH 5.80) as a function of wave length.
By plotting the optical densities of potassium metaperiodate solutions uerm- temperature (at a pH 6) the extinction coefficient of the metaperiodate ion, IO4-, could be estimated With this
IONIZATION AND HYDRATION EQUILIBRIA OF
Jan., 1951
information together with the total periodate in solution and the assumption that the activity coefficients of the two forms of monovalent ions were equal, the dehydration constant KD =
(6)
(MIOi-/MH&I06-)a2H20
could be computed. Figure G is a plot of the optical density versus temperature of these solutions. The limiting value a t high temperatures
85
PERIODIC .4CID
mental error, the value obtained a t high temperatures in aqueous solutions. Plotting of these data yielded a straight line with a slope of -2.0 and the intercept value checked the value obtained for log K D a t this temperature. Values of the apparent first ionization constant, KI', were computed from the optical density-pH curves shown in Fig. 2 using the method already d e ~ c r i b e d . ~The log K,' was plotted versus the reciprocal of the absolute temperature in Fig. 8. log KD and log ( K D 1) were also plotted on this figure. If the constants K I ' , K D and Kl are taken as
+
2.08
K1' =
2.04
G +
'5 2.00 0
1.96
i t follows that
1.92
or
v
.e
1.88
= KD
+ 1)
log Kt' = log K ,
l!ItrDa
lA4 1.80
0 20 40 60 80 100 Temperature, "C. Fig, 6.-The optical density of potassium periodate (2.0 X loA4molar and PH 6) us. temperature.
was used for the calculation of the value 10,700 as the molar extinction coefficient of 1 0 4 - . The test for the validity of reaction (4) was afforded by measurements of the variation of optical density, and hence of the (MIO,-/MH,IO~-) ratio with changes of the aH,O in water-methanol solutions. For the reaction considered, this ratio should satisfy the relation (MIO,-/MH4IOa-)
aHrIO6-)/aHsIO6
KI' = KI(KD
0
log
+
(~H')(~HrIOe-)/~HsIOa
(a10d-/aHr106-)a2Hz0
I
2@
(UH+)(UIOi-
K1 =
= log (aIOr-/aHrIOa-)
+ log (KD
(8) (9) (10)
(11) 1)
(12)
hence the log Kl may be obtained by subtraction 1) from log K1'. The resulting of log ( K D values for log K1have been plotted in Fig. 8. The slopes of these curves made possible the evaluation of the heat of reaction for
+
+
HIIOB- C, 2H20 IO4-: AHo-roo = 10.9 kcal.(l3) HjIO6 -++2H2O H C Io4-: AH260 = 10.9 kcal. (14) HJO6 -e+ H + H4106- L\HzB0:= 0. kcal. (15)
+
+
+
11. Potentiometric Titrations.-The determination of a true second ionization constant for
=
This is an equation of a straight line with a slope of -2 and an intercept, log KD, when log (MIo,-/MH~Io~-) is plotted versus log U H ~ O . Methanol was employed to vary the activity of water since i t neither reacts with periodate nor absorbs appreciably a t this wave length, and $I3 measurements showed that in methanolwater mixtures appreciable amounts of the undissociated acid were not formed. The activity of water in aqueous methanol mixtures was computed from known vapor composition informa-log a H 2 0 . tion at 250*4 containing the Same Fig. 7.-The log ~l~Io,-/MH410,us. log uHio in various methamount of potassium metaperiodate in various anol-water mixtures a t 0 ', methanol-water mixtures a t 0' were measured on the spectrophotometer and the resulting in- HbIOa from spectrophotometric data would require formation is given in Fig. 7. The ratio (MIo,-/ further analyses of the spectra in the alkaline MH,IO~;)was calculated using the value for the region. However by a treatment similar to the extinction coefficient obtained by extrapolating above where the optical density curve versus mole fraction K? = ( U H + ) U Z / U H ~ I O ~ (16) of water t o zero. The extrapolation was necesand sary since a t mole fractions of methanol greater K2' (aH+)~Z/(aE4IOea104-) (17) than 0.8 it was apparent that interaction of the periodate with methanol became important. the following expression can be obtained which This extinction coefficient checked, within experi- corrects for the dehydration of the univalent ion
+
(4)
J. B. Ferguson and W. S. Funncll, J . Phys. Chcm., 88, 1
(1929).
Kn
6
&'(KD
+ 1)
(18)
acid, HI04, would be expected to be a very strong acid similar to perchloric. Theoretical treatmentKapbhave indicated that two properties appear to classify the inorganic oxygen acids with respect to ionization constants into distinct groups. First, the formal charge on the central atom and, second, the number of nonhydroxyl oxygen atoms in the molecule. Iodine in periodic acid has a formal change of, nc = 1 a n d the non-hydroxyl oxygen atoms are, 71 = 1 , 2, 3 for HBI06,HJOe- and H3106- species, respectively. Using the semi-empirical expression n (4.0) derived by Kicci,5 pK = S.0 - m (9.0) an improved correlation is obtained between calculated and observed values when the true ionization constants are used. In Table I1 a summary of the ionization constants is given.
+
I
TABLE I1 COMPARISON OF OBSERVEDAND CALCULATED VALUESOF THE IOKIZATION CONSTASTSOF PERIODIC ACID AT 25"
'\
hpparen t 3
True
Calcd.6
PK
p K (this paper)
9K
PK?
1.6 SA
3.3 6.7
pt;'::
14.3
...
3.0 7.0 11 . 0
Ph-1
I 1
29 3 0
1
31 - 3 2
33
f Fig. 8.-Iog
x
34
35
36
37 3 8
-
IO'
+ I),
Ku; log (Ru
log K ' and log
k'l VJ 1, T .
Values of Kz' were calculated from the series of potentiometric titrations with solutions thermostated a t 0.5,25.0 and 45.0' (Fig. 3 ) by the methods used p r e v i ~ u s l y . ~Values of K z and K,' from these experiments have been included in Table I. I t should be noted that the apparent constant Kz' a t 2.5' agreed closely with the constant obtained from the spectrophotometer data in the preceding paper. TABLE I (Values of KS and --log Kz'
7 40 8 40 8 63
T,
ac
0 3 35 0 45 0
KI')
log (Ku
i- 1 ) 0 95 1 61 3 13
-log E;#
6 45 6 69 6 50
The value of AH obtained from plotting log Kt' versus 1 /I' was approxitnately 11 kcal. a t room temperature. This corresponded to the heat effect of the dehydration reaction of the monovalent ion and indicates that a t near room temperatures the major portion of the divalent ion can be described as H310e". The values of K:! remain essentially constant over the temperature range within the accuracy of measurement, again, agreeing with the characteristics of other oxyacids. Figure 4 shows potentiometric titrations made in methanol-water mixtures. An unusual effect was observed here since the pH values in methanolwater mixtures were lower than in water. The dehydrating effect of methanol apparently overcame the usual effect of depression of the ionization of the acid when methanol is used. In this regard it is also well to noint out that metaneriodic
-
I t is interesting to note that the first apparent ionization constant of carbonic acid has a very small value because of a hydration equilibria between carbon dioxide and the undissociated acid as reported first by McBains and later investigator~.'*~ From the equilibrium constant K D and AHD for the dehydration of HdIOs-, the entropy change, ASD was computed to be 44 e.u. Contributions to this entropy change were indicated by resolving the reaction into three steps.
-+- H4106- (9) A&(H1106-) (19) H410e-(g) +2H20(1) -k 1OI-k) A S = 2S"[H20(1)] f SOIIOi-(g)] So[HJO~-(g)] (20) 104-(g) +10,- (as), A&(Io,-) (21) HJ06- (as)
-
AS, represents the entropy change for the hypothetical process of taking one mole of the ion in question from its standard state in aqueous solution to a standard state of ideal vapor a t one atmosphere. This quantity for various ions has been discussed by Frank and Evans.Q Using 9[H20(1)] = 16.75 e.u.l0 and calculating by the methods of statistical mechanics a value of - 12.5 e.u. for So[IOp-(g)] - So[HJOs-(g)],ll the differ(5) (a) A. Kossiakoff and D. Harker, THISJOURNAL, 60, 2047 (1938); (b) John E. Ricci, ibid., TO, 109 (1948). (6) McBain. J . Chem. Soc., 101,814 (1912). (7) G.A. Mills and H. C. Urey, T m s JOURNAL, 62, 1019 (1940). ( 8 ) A. R. Olson and P. V. Youle, ibid., 69, 1027 (1940). (9) H. S.Frank and M. W.Evans, J . Chcm. Phys., 18,507 (1945). (10) W. F. Ciauque and J . W. Stout, THIS JOURNAL, 68, 1144 (1936). (11) The 101-was assumed t o be tetrahedral with 1-0 distances of 1.79 A. From a set of estimated tetrahedral frequencies the vibrational contribution to the entropy w s s calculated t o be 11 e. u. The HdOI was assumed to have the oxygens arranged octahedrally with all 1-0 distances equal t o 1.83 h. The 0-H distances were taken as 0.97 A. and the I-0-H angles as 105O. The vibrational entropy was calculated t o be 16 e. u. from a set of estimated octahedral frequencies and a treatment of the hydrogens' motion as restricted rotation around the 1-0 bond with a very high two-madmum barrier. The symmetry of the 8. ion WPI) tnken to be D4h..u
-
-
Jan., 1951
IONIZATION AND HYDRATION EQUILIRRIA OF
ence in ASv for the two species would be
- 33.5 + 12.5 = 23 e. u.
SV(H4106) - &(IO4-) = 44
(22)
This difference would seem t o be surprisingly high. A part of the difference might be explained as due to loss of rotational entropy in the cage of the solvent. However, a fairly large fraction of i t would still have to be accounted for as due to a difference of the two ions in destroying the structure of the solvent, the A s s t discussed by Frank and Evans. 111. Solubility of Potassium Metaperiodate.Inconsistencies exist in the solubility values reported for KI04 in the literature. The early values of T. V. Barkerlzanh are apparently too high when compared with the later values of A. E, HillL3 and J. H. Jones.I4 From equilibria 9 and 17 it can be seen that the solubility of potassium metaperiodate should be increased by both excess acid and alkali. Thus it is not surprising to find that potassium hydroxide does not depress the solubility of potassium metaperiodate salt according to the common ion effect.I3 Also it is known that strong acids dissolve potassium metaperiodate quite readily. Using the data of and Jones" for the solubility of potassium metaperiodate the solubility product a t various temperatures has been calculated. The following equations should apply to solutions of potassium metaperiodate with no excess acid or alkali (pH 6-6.8). ( J ~ I o ~ - / ~ ~ ~ H ~=I KD o~-)
+
[KDIKD 11~vT= 1I'fIOd( u K + ) ( ~ I o , - )= K,,
+
Y r 2 ( ~ ~ ~ h t ) ( ~ ~ T ) [ ~ D11 / K=D Ksl,
6 1 0 4 ( s ) -E-+
K+
+ 104-
A H = 15.1 kcal.
(231 (24) (25) (26) (27)
\Vhere XT is the total periodate concentration, moles/liter, in solution and y + is the mean activity coefficient calculated by Jones." Using the solubility data of Jones for potassium metaperiodate a t 25' where yt = 0.866, XI
