DECOMPOSITION OF SODIUM HYPOCHLORITE: THE UNCATALYZED REACTION1

ABSTRACT 'The decomposition of sodiu~n hypochlorite has been re-examined. The results show that Foerster and Dolch's mechanism of the decomposition to chlorate and chloride is correct; they postulated a slow bimolecular reaction to chlorite and chloride, followed by a faster reaction of chlorite with more hypochlorite. Values of the rate constants of both steps are reported; they make the activation energies 24.8 l;cal./g~n-molec~~le the first step and 20.8 kcal./gm-molecule for for t h e second. The rates are such t h a t a t 40" C. a solution of sodium hypochlorite will contain about lyo as much chlorite a s hypochlorite. The rate is strongly affected by changing ionic strength; a t low ionic strengths it is nearly constant or falls slightly; above about 0.8, the rate rises and a t high ionic strengths the rise is quite rapid. No signs of specific catalytic effects of sodium chloride, hydroxide, or carbonate could be observed, and it seems probable t h a t earlier reports of this were due t o variations in ionic strength. T h e decomposition t o chloride and oxygen has been measured and is a unimolecular reaction, which is possibly, but not certainly, uncatalyzed. Values of its rate constant are reported; they also are much altered by changing the ionic strength.

Although this reaction was first investigated a considerable number of years ago, there are several matters connected with it that are still not entirely clear. Briefly, the position seems to be this. T h e best early work on the subject was that of Foerster and his co-worlters, particularly Foerster and Dolch (2). They found it to be a second order reaction, and consequently deduced t h a t the mechanism was:
2NaOC1 + NaClOz NaCl NaOCl NaCIO? + NaC103

+ NaC1 .

T h e first step is the slower. They showed by an independent experiment t h a t the second step was indeed faster, and their rate constants gave activation energies of 224 and 20%kcal./gm-molecule for the first and second step respectively. In addition to the reactions given above it has long been known that decomposition t o sodium chloride and oxygen also occurs. Foerster and Dolch's rates are the total reaction rates, though they knew that the reaction to chlorate was very much the more important. They also found that the rate increased as they added sodium hydroxide. Later Giordani (3) investigated the reaction, and concluded that it was a combinatiori of a termolecular reaction giving chlorate, and a bimolecular reaction giving oxygen. He also found a marked effect of sodium hydroxide, b u t in the opposite direction from that fou~lci Foerster; and, a t least from about 0.5 to 1.0 M sodium hydroxide, by obtained rate constants proportional to (Na0H)-?. H e therefore assumed that hypochlorous acid was essential t o the reaction, and that it went in one step, as follows:
OCI-

+ 2HOC1+

C103-

+ 2HC1

He proposed that the reaction to oxygen was

20C1- - 2C1,

+ 0,.

'Manuscript received November 14, 1955. Conlribzrlion from the Departn~entof Chevcistry, University of Toro~zto,Toronto, Ontario.

465

466

CANADIAN JOURNAL O F

CHEMISTRY. VOL.

34. 1956

Skrabal and Skrabal ( G ) examined the dependence of rate on p H , and agreed with Foerster (and Giordani) that the rate was second order a t high pH. Pierron (5) examined the relative stability of lithium, sodium, and potassium hypochlorites. He seemed t o favor Giordani's mechanism, b u t supposed t h a t alkali metal peroxides were formed in the alkaline solutions, and that these promoted the reaction. Barredo (I) examined the reaction and found it to be second order; but he also found the rate t o be altered by addition of chloride, and over a certain range the rate was proportional t o the chloride concentration. He believed that this explained Giordani's results. Such a dependence on chloride concentration would of course make the reaction autocatalytic. The present author (4) has examined the rate of decomposition of neutral or slightly acid hypochlorite solutions, and found t h a t under these conditions the main decomposition was of hypochlorous acid molecules, by a second order reaction. I t can be calculated from these results (combined with those in the present paper) that once quite a small excess of sodium hydroxide is present, the concentration of hypochlorous acid is so slow as t o make only a negligible contribution to the total decomposition. I t seemed possible that these reported effects of sodium hydroxide and chloride might really be due to changes in the ionic strength. In earlier work the ionic strength was not always kept constant from run to run; and it is not always possible in re-examining the data to deduce what the ionic strength reallj. was. Accordingly the effect of ionic strength on the rate of reaction was examined in the present work. Foerster and Dolch's results seemed to establish t h a t the reaction was second order; however, in view of contradictory later results, this point was checked. Sodium hypochlorite decomposes both to chlorate and chloride, and to oxjygen and chloride. The relation between these reactions was by no means clear, and the oxygen evolution had been relatively little investigated. Hence in the present work the two reactions were measured, so that their kinetics could be determined separately. Since the reaction to oxygen is strongly catalyzed by certain metallic oxides, notably copper or nickel oxides, there has always been some doubt that there really is ally appreciable uncatalyzed evolution of oxygen. Hence i l l the present work, an attempt was made to be sure that the oxygen evolution was not catalyzed; and to find the order of this reaction, and the effect on the rate of varying temperatures or ionic strength.
MATERIALS AND EXPERIMENTAL METHODS

S o d i u m Hypochlorite This was produced in solution by the usual method of passing chlorine through cold sodium hydroxide solution. The product always contained a small excess of sodium hydroxide to avoid hydrolysis to hypochlorous acid. The chief uncertainty with these solutions was that they might contain traces of metallic oxide which could act as a catalyst. Such impurities could easily be present in any commercial sample of sodium hydroxide to a n extent that would be undesirable, and it would be very difficult to measure how much

LISTER: UNCATALYZED REACTION

467

was present accurately enough to allow for its catalytic effect. I t was found however that if nickel or copper salts were added t o the hypochlorite solution, they could be removed by the following procedure. A sinall amount of calcium chloride was added, and then enough aqueous sodium carbonate to precipitate all the calcium. The calcium carbonate was filtered off by a sintered glass disk. I t was found that if this procedure was repeated four or five times, the rate of gas evolution fell rapidly t o a constant low value; further precipitations did not reduce the rate any more. Presumably this procedure removed the copper or nickel salts by coprecipitation with the calcium carbonate. The fact that a final slow rate was reached is some evidence that this was the rate of the uncatalyzed reaction. I t was checked that calcium carbonate did not affect the rate; nor did pyrex glass, since the same rate was obtained in a polyethylene flask.

Sodium Chlorite This was the best available commercial material, manufactured by the Mathieson Alkali Works, recrystallized from water. I t contained about 0.04% of chlorate, and a trace of chloride. Sodium Carbonate T h e reagent grade of sodium carbonate, manufactured by Merck Chemical Co., was recrystallized from water. Sodium Chloride The reagent grade of sodium chloride, manufactured by the General Chemical Co., was recrystallized from water. Apparatus The solutioils were contained in a 1-liter flask fitted with (i) a mercury sealed stirrer, (ii) a side arm, normally closed, through which samples could be taken, and (iii) a glass capillary tube leading to a water-jacketed gas burette. The flask was immersed in a water thermostat of conveiltional design, which was found to maintain the temperature over long periods of time to f0.1" C., though the swing during any on-off cycle was only about 0.02" C. The temperature was read on a thermometer graduated t o 0.1" C. T h e oxygen evolved was measured in a gas burette graduated in 0.1 ml. In practice the volume could be easily read to 0.05 ml. In reducing the observed volumes to N.T.P. it was always assumed t h a t the oxygen was saturated with water vapor a t the temperature of the burette. T h e confining liquid was mercury, but during a run a trace of water usually condensed out of the oxygen, especially in the runs a t higher temperature. T h e pressure of the oxygen was adjusted t o atmospheric pressure, as shown by a small mineral oil manometer; the atmospheric pressure was noted t o the nearest mm.

Analytical Methods T h e solutions were analyzed a t various times for sodium chloride, hypochlorite, chlorite, chlorate, hydroxide, and carbonate. The methods were the same as those outlined by the author in a previous paper (4).

468

CANADI.4N

JOURNAL O F

CHEMISTRY. VOL.

34, 195G

EXPERIMENTAL PROCEDURE

The general procedure in almost all runs was as follows. A quantity of the stock hypochlorite solution was poured into the reaction flask, the amount added being weighed to the nearest 0.1 gm. Sodium chloride, sodium carbonate, or water (as required) were also weighed out and added. The final volume was always close to 900 ml. When appreciable additions were made to the stock hypochlorite solution, a sample was pipetted out and its density determined; from this the volume of solution could be found, so that the oxygen evolution per liter of solution could be calculated. The flask was then brought to the right temperature in the thermostat, and a t intervals samples were taken for analysis. Between samples the gas evolution was followed by means of the gas burette. I t was found that the gas evolution was usually slower a t first but built up to a practically constant rate. I t is believed that this is due to the building up of a certain degree of supersaturation in the solution, and that eventually the steady rate of gas evolution equals the true reaction rate; it is always the steady rate that is reported in the following data. The only exceptions to this procedure were in two runs intended to discover whether the pyrex flask catalyzed the reaction. In one of these the flask was filled with pyrex beads, and in the other a polyethylene flask was used. These runs are not reported in detail, but it was found that they gave the same rates as in the ordinary pyrex flask. I t was also checked that calcium carbonate, which might be present in traces from the purification process, did not affect the rate. In a few runs sodium chlorite solutions were used, b u t the general procedure was the same.
EXPERIMENTAL RESULTS

The results for the decompositioil of the stock hypochlorite solution a t various temperatures are given in Table I. Table I1 gives various rates of gas evolution. In this table are given data on some runs which were done with filtered hypochlorite solution, but which had not had the complete calcium carbonate treatment. These gas rates are high relative to later runs, so their rate constants for oxygen evolution have little significance; however the data are needed for calculations on the rates of reaction leading t o chlorate. I t is believed that the rate constants of the reaction to chlorate obtained from these runs are reliable, since (i) it was found that the purification process did not affect the rate of chlorate production, but only that of oxygen, and (ii) it was found (to anticipate results which will be reported later) that catalysts such as copper oxide did not affect the rate of chlorate formation. These runs are also of interest because they show that the rate of gas formation is proportional to the hypochlorite concentration, a t least in these solutions; and they throw some light on the activation energy of the (presumably) uncatalyzed reaction to oxygen. Table I11 shows the effect of ionic strength on the rate constant t o chlorate, and Table IV shows the effect of ionic strength on oxygen evolution. Table V gives results on the effect of sodium carbonate, and sodium hydroxide. Table VI gives data on the chlorite-hypochlorite reaction.

LISTER: UNCATALYZED REACTION

TABLE I
-,

Run 1

I emp., " C. 40

50

60

40

60

'rime, min. 0 138 289 581 1244 0 171 279 521 0 112 250 380 0 193 409 1386 0 358

(CIO-), M 1.556 1.548 1.545 1.532 1.506 1.497 1.476 1.457 1.433 1.356 1.316 1.278 1.237 1.201 1.192 1.187 1.164 1.134 1.045

(ClOa-), M 0.007 0.008 0.0075 0.008 0.0075 0.0095 0.010 0.013 0.012 0.011 0.015 0.010 0.011 0.006 0.008 0.009 0.007 0.007 0.009

(CIOa-), M 0.044

0.047 0.051 0.059 0.061 0.067 0.070 0.079 0.105 0.112 0.128 0.138 0.152 0.154 0.154 0.162 0.174 0.199

(NaCI), M 1.80

1.83 1.84 1.89 1.94 2.02 2.05 2.08 2.09

Remarks Ionic strength 3.79 (NaOH) = 0.32 d l (NaaCOa) = 0.02 Ad As in run 1

As in r ~ 1 n ~

As in run 1

As in run 1

50 50

0 347 658 0

1.481 1.434 1.397 1.320

0.009 0.007 0.010

0.178 0.088 0.103 -

1.84 1.90 1.96

As in run 1 As in run I

0 82 210 302

0.973 0.961 0.942 0.929

0.007

0.015

2.48

Ionic strength 3.63 (xaOH) = 0.042 M (NazCOa) = 0.036 M

TABLE I1 Run Temp., " C. Mean (CIO-), crm-niol./l. Gas rate, ml. N.T.P./min. per 1. of soln.

KO,n1in.-'

470

CANADIAN JOURNAL O F CHEMISTRY. VOL. 34, 1956

TABLE I11
-

Temp., Ionic Run O C. strength

Time, mln.

(CIO-1, M

(ClOz-), Af

(Clan-),

(NaCI), M

Iie~l~arks

As

ill

run I )

(NaOH) = 0.28 (NapC03) = 0.02 M (NaOH) (Na2COI)

= =

0.29 iM 0.05 M

(NaOH) = 0.32 M (NazC03) = 0.02 M (NaOH) = 0.29 M (Na2COa) = 0.02 M (NaOH) = 0.18 M (NarCOn) = 0.03 iM

(NaOH) = 0.23 11f (NalC03) = 0.01 M

(NaOH) = 0.28 M (NarC03) = 0.03 M

(NaOH) = 0.19 M (Na?C03) = 0.02 M (NaOH) = 0.06 M (NazCOn) = 0.02 M

(NaOH) = 0.11 M (Na?C03) = 0.01 M (NaOH) = 0.05 iM (Na?COa) = 0.00 M

(NaOH) = 0.28 iM (Na?COn)= 0.03 M

LISTER: UXCATALYZED REACTIOX

TABLE I\!

DATAON
R L ~Typ., C.

RATES OF EVOLUTION OF OXYGEN

Ionic strength

Mean (C10-), gm-mol./l.

Gas rate, ml. N.T.P./min. per I. of soln.

TABLE \'

RLIII

Temp., " C.

Time, min.

(CIO-),

(C102--),
llf

i l I

(CIOs-), ilf 0.070

Remarks Ionic strength 4.57 (NaOI-I) = 0.28 A4 ( N a L 0 3 ) = 0.51 A l Ionic strength 4.25 (NaOH) = 0.07 A[ (Na?COa) = 0.265 d l Ionic strength 3.70 M (NaOH) = 0.023 i (Sa?CO,?)= 0.019 M Ionic strength 3.73 (NaOH) = 0.275 d l (Ka?COa)= 0.04 M

0.086 0.133 0.118

0.161

Ionic strength 1.75 (NaOH) = 0.06 M (Xa?CO,) = 0.02 M

As in r u n 22

C 4 K A D I 4 N JOURNAL O F CHEMISTRY.

VL O.

34. 195G

TABLE VI

DATA ON

CHLORITE-HYPOCHLORITE

REACTION

Ionic strength 3.79 Run Temp., " C.

Time, mln.

(C10-),

(ClOs-),

(C102-),
J M

DISCUSSION O F RESULTS

I t is apparent from these results that the main reaction in the decompositioii of sodium hypochlorite is the oiie to chlorate, and only a small part goes to oxygen. Consequeiitly the over-all kinetics are those of the reaction to chlorate, and the results of Table I fit the equation of a second order reaction: this can be seen from the long run, run 7, or by comparing runs 1 and 4, or 3 and 5 . The possibility of catalysis by chloride, or effects of sodium hydroxide, will be considered later when the rate constants have been evaluated. Table I1 shows that for these runs the reaction to oxygen is first order, though owing to the relatively small amounts of gas evolved the rates are not very accurate. Any comparisons of this sort must be made between runs a t the same ionic strength; fortunately the decomposition of hypochlorite does not change the ionic strength during a run. The mechanism of the reaction is evidently that proposed by Foerster, with

LISTER: UNCATALYZED REACTION

473

the addition of the reaction to oxygen. Let the rate constants of the various reactions be kl, k2, and ko as follows:
NaOCI 2NaOCI + NaCI + NaC102 + NaC102 + NaCl + NaCI03 NaOCI + NaCl + $02
kI k 2 ko
=

and let the concentratioils be: (C10-) = x, and (ClOz-) equations are (assuming no effect of chloride) :

y. Then the rate

These equations are difficult t o solve rigorously, but various simplifying assumptions can be made appropriate to the conditions of the reaction. I n particular the chlorite concentration is low and changes very little during a run; so approximately dy/dt = 0. Equation [i] then becomes:

whose solution is most useful in the form: [iii] where x

=
- -

l x

l -+-(e"b xo k o

1-

1)

x~ a t t

0 ; and b

+kl+ko/xo.

T o evaluate kl it seemed simplest to plot l / x against t, and to determine the slope of the best line through the points. Calling this slope s, we get:
S =

l/x-l/xo t

b -(ekO'-1) kot

b(1+$kot+&02t2+

. . .).

As kot is always small (never above 0.04), t o a good approximation:

s = b(l+$kot)

and

ko was determined as explained in the next paragraph, and kl was then fouild from equation [iv]. I t will be seen t h a t kl could also be obtained from the chlorate concentrations, and this was occasionally done, but the hypochlorite ailalyses are probably the more accurate. For the determination of ko, let v be the quantity of oxygen evolved a t time t, in units of gm-atoms of oxygen per liter of solution. Then:

Putting in the value of x from equation [iii], integrating, and since v t = 0, we get:

0 at

474

CANADIAN J O U R N A L O F

CHEMISTRY. VOL.

34. 1956

T h i s is awkward t o apply, b u t suppose we take a s a n approximate value of K O :

i.e. the average rate of gas evolution divided b y the mean hypochlorite concentration during the measuring period; then from equations [iii] a n d [v], expanding in powers of t u p to t 2 :

T h e term in t2 gives the error introduced b y this approximation for k o . Using typical figures for 60" C., this term is a b o u t 2.5X10-*t2. T h e gas collection period never ran above 400 min., when the error is about 0.4%. Hence this approximation is sufficiently accurate. T h e values of k1 a n d k o given below are from equations [iv] a n d [vi]. T h e evaluation of k2 will b e discussed in a later section. T h e results in Table I1 give values of ko which are appreciably higher t h a n those obtained for the most carefully purified solutions. Nevertheless these results are of interest a s showing t h a t the gas r a t e is proportional to the hypochlorite concentration ( a t least for these solutions); a n d the d a t a a r e needed in evaluations of k l . T h e mean k o values for each run are:
Run Temp. ko

1 40 2.28

4 40 2.48

7 50 6.95

2 50 6.81

3 60 19.45

5 60 19.4

" C.
XlO-%in.-l

T h e average k o a t each temperature was used in obtaining k l for these runs. I t is interesting t o note t h a t although a trace of impurity must be present, these runs give a n apparent activation energy of a b o u t 21 kcal./gm-mol., which is considerably higher t h a n for a n y of the catalyzed reactions t h a t were examined. T h e effect of ionic strength can be obtained from Tables I11 a n d IV. Froill the gas rates in Table IV, we get:
Iiun 9 Iorlic strength 5.83 ko (60 C.) 25.1

10 4.84 10.2

14 4.71 17.6

15 3.69 9.4

16 2.76 5.7

21 1.75 4.2

17 1.74 4.7

18 0.59 5.0

X1O-i min.-I

k o varies considerably with ionic strength. As will be seen in the next section, a parallel change is also found for k l . T h e results in runs 21, 22, and 23 enable k o t o be found a t different temperatures b u t t h e same ionic strength. T h e results are:
Temp. ko 60 4.2 70 0.2 75 16

" C.
XlO-%ir~.-l

These make t h e activation energy about 20 Itcal./gm-mol. R a t h e r surprisingly this is lower than from runs 1 t o 7, b u t possibly the difference is due t o experimental error, for the actual rates are very small. However the value seems t o be about 20 to 21 lical. I t is also surprising t h a t the activation energy seems t o be lower t h a n for the reaction t o chlorate. T h i s malies it doubtful that we have a n uiicatalyzed reaction; the catalyst is not pyrex glass, nor cobalt,

LISTER: UNCXThLYZED REACTION

475

nicltel, or copper (which give lower activation energies). However it is impossible t o be sure t h a t traces of some unltilown catalyst are not present, and there must remain this element of doubt in the interpretation of the results. T h e runs in Table I give the kl values in Table VII. I t will be seen, for instance by comparison of the rate coilstants of runs 3, 5, and 11, t h a t the
TABLE VII
(NaCI), i\l Iiu11 Temp.,

" C.

kr, min.-'(gm-mol./l.)-l

Initial

Final

Ionic strength

5n

3.62

purification did not affect kl; to anticipate results reported in another paper, addition of catalysts such as copper also did not affect kl. Table VII also gives the sodium chloride concentrations during the runs. All these runs except run 8 were a t a n ionic strength of 3.79; in run 8 it was 3.63. Hence it is possible to discover whether sodium chloride has a specific catalytic effect, apart from its contribution to the ionic strength. Although the range of concentrations might perhaps have been profitably extended, these runs definitely show that sodium chloride has no effect on k1, provided the ionic strength remains unchanged. In particular, run 8 gives a value of k1 perhaps slightly low for this ionic strength, although the sodium chloride concentration is some 25% higher than in run 3. T h e long runs 5 and 7 show no sign of autocatalysis, although the sodium chloride increased 10-1570 during them. Further evidence in support of this contention is provided by the runs with sodium carbonate present. At an ionic strength of 3.79 the mean rate constants are:
Temp.

kI

40 1.02

50 3.61

60 11.4

" C.
XlOP min.-'(grn-rnol./l.)-I

These give a good linear plot of log k1 against 1/T, and make the activation energy 24.8 (to the nearest 0.2) kcal./gm-mol. Foerster's constants a t 25' and 90' C. give a value close to 26 kcal., but his result a t 50' C. does not fall on a straight line with these values in a plot of log kl against 1/T. Turning now to the effect of ionic strength on kl, we get from Table I11 for k1 a t 60' C.:
Run Ionic strength

k1
Run Ionic strength
k1

9 5.83 17.4 20 2.61 7.6

10 4.84 16.75 17 1.74 5.5

14

4.71 15.65 21 1.75 5.4

19 4.16 12.1 18 0.59 3.85

11 3.79 11.4 13 0.515 4.3

15 3.69 11.2

8 3.63 10.5

16 2.76 7.7

12 2.73 7.8

X10-6 min.-'(grn-mol./l.)-I

476

CANADIAN JOURNAL O F CHEMISTRY.

VL O.

34, 1956

These values give somewhat the same type of trend that might be expected for the activity coefficients of sodium hypochlorite over this range of ionic strengths. T h e trend in ko is somewhat, though not very, similar. Run 24 makes k l = 1.98X10-5 mill.-'(gm-mol./l.)-I a t 50' C. and an ionic strength of 2.61. This is proportionally a somewhat larger drop than a t GO0 C., but not very much. Runs 25 and 26 (Table V) give the effect of adding sodiunl carbonate; they make kl a t 50' C.:
Run 25 Ionic strength 4.57
26 4.25 3.95 1.99

kI
(NaCI) initial

4.25 1.58

X10-j n1in.-'(gm-mol./l.)-I

Relative to the more carefully measured rate a t an ionic strength of 3.79 these runs show an increase ill k 1 which is nearly what might be expected purely from the increase in ionic strength. The increase is perhaps a little less than in the runs when sodium chloride was added : of course specific effects are bound to occur, and one would not expect two different salts to give exactly the same change in kl for the same change in ionic strength. This is especially so when the two salts, as here, are not of the same ionic type. Nevertheless the main changes in k l do seem t o be attributable to the changes in ionic strength. Runs 21, 22, and 23 enable us to evaluate kl a t three temperatures a t another ionic strength (1.75) :
Run Temp. kl
21 60 5.4 22 70 16.0 23 75 24.0

" C.
X10-5 min.-'(gm-rnol./1.)-~

T h e figures a t the higher temperatures are rather rough; they give a moderately linear plot of log k l against 1/T, with a slope corresponding to an activation energy of about 24 kcal. The experimental error is too large to say whether there is any real difference between the activation energies a t ionic strengths of 3.79 and 1.75. Table V gives the data a t low sodium hydroxide concentrations. The bulk of the runs in this paper were done in the presence of 0.32 M sodium hydroxide. Some of the runs a t low sodium hydroxide concentrations have already beell considered, and found to fall illto line with the other runs (e.g. runs 8 and 21). T h e k l values for various runs which provide evidence on the effect of sodium hydroxide are as follows:
Run (mean) Temp. 60 Ionic strength 3.70 (NaOH) 0.32
27 60 3.70 0.023 10.9 8 60 3.63 0.042 10.5 21 60 1.75 0.057 5.42 17 60 1.74 0.19 5.48

(mean)
50 3.79 0.32 3.60

kI

11.4

28 50 3.73 0.275 3.48

" C.
ill

These results show that sodium hydroxide has a negligible effect apart from its contribution to the ionic strength. A t very low sodium hydroxide concentrations hydrolysis t o hypochlorous acid and its decomposition become important; but the rate of this reaction has been measured (4), and it call be calculated t h a t the sodium hydroxide would have to be close to 0.001 M before this reaction made much difference to the total rate.

LISTER: UNCATALYZISD KPSACTION

477

Finally the results in Table VI give information on the hypochlorite-chlorite reaction. Firstly runs 29 and 30 show t h a t the decomposition of chlorite by itself is entirely negligible. From the remaining runs k:! was evaluated a s fol1 0 ~ ~As one molecule of chlorite removes one of hypochlorite, and since this s. reactio~lis much faster than the decomposition of hypochlorite (kl), then approximately : cEx/dt = dy/dt = -k?xy where, as before, x is hypochlorite and y is chlorite. The solutioil of this equation is:

where c = xO-yo. 111 reactioils of this sort k2 is ilorinally obtained froill this equation, but in the present case a small correction has to be applied for the clecomposition of the hypochlorite, which is slow but not negligible in coinparison. I t therefore seemed simpler to take an approximate value of kz defined by k2(approx.) = corrected rate/(meau ?c) (mean y) . By "correctecl rate" is meant dx/dt between any two successive readings minus the part of the rate due to decompositioil of hypochlorite, which call be calculated from kl. If the values of x and y from equation [vii] are substituted in this expression, and the exponentials expanded in powers of t, we get: [viii] k2(approx.)
=

k ? [ l - (.vi'+xoyo+y$) (kz%2/12)

+ ...] .

This gives the error in ks in this approximation: in the runs reported it rose to about 1%, but the error could then be allowed for. T o examine the validity of our methocl of allowing for kl ant1 ko, we call subtract equations [i] and [ii] to get: dy/dt -d.t-/dt = Qkl.xz+kox . Substituting the value of .2: froin [vii], integrating, and expanding in powers of t , it is found t h a t the fractioilal error in k2 so introduced is: Fractional error
=

(Gkly0+3k1xo- ko) . xoyokat. 3klxo+2ko

T h e size of this error was checked for runs 31 to 34, and ill no case was above 0.3y0. Consequer~tlyit was assumed to be adequate to obtain k:! simply b y taking the observed slope of x (or y) over each interval between analyses, correctiilg for the decompositioil of hypochlorite if the slope of x is used, and dividing the slope by the mean x times the mean y over the interval. A small correction was applied in accordance with equatioil [viii]. T h e mean values of k:! for each run were:
Run Temp.

k2

31 50 2.75

32 50 2.77

33 45 1.64

34 40 0.97

" C.
X10-3~~~in.-l(gm-mol./l.)-1

478

CANADIAN JOURNAL O F CHEMISTRY.

VOL. 34.

1956

These results give an activation energy of 20.8 kcal. In all these runs the ionic strength was 3.79, and the sodiurn hydroxide co~lcentrationwas 0.32 M. Foerster's results made the activation energy 203 lccal., in good agreement. but the difference may be However his constant a t 50 C. is about 8.3 X due to a different ionic strength. These values of kz mean t h a t a decomposing sodium hypochlorite solution will contain about 1/100 as many chlorite ions as hypochlorite ions a t 40' C.; this was roughly found. If the constants k l a n d k p are written as AecEIRT, A factors are fairly similar: the

T o surllmarize these conclusions, it is believed that these results show t h a t (i) the mechanism of Foerster et al. via chlorite is well established; (ii) the first stage to chlorite has an activation energy of 24.8 kcal., and is slower than the secoild stage, which has an activation energy of 20.8 kcal.; (iii) the' rates are strongly influenced by the ionic strength; (iv) added sodium hydroxide, carbonate, or chloride changes the rate by changing the ionic strength, b u t apart from this they exert no specific catalytic effect; and (v) there is a si~nultaneous unimolecular decomposition to oxygen, which is possibly, but not certainly, uncatalyzed. REFERENCES
1. BARREDO, M. G. Anales fis. y quim. (Madrid), 37: 220. 1941. J. F. and DOLCH, Z. Elektrochem. 23: 137. 1917. P. 2. FOERSTER, 3. GIORDANI, Gazz. chim. ital. 54: 844. 1924. F. M. 4. LISTER, W. Can. J . Chem. 30: 879. 1952. P. 5. PIERRON, Bull. soc. chim. France, 10: 445. 1943. 6. SKRABAL, 11. and SKRABAL, Monatsh. 71: 251. 1940. R.