You are on page 1of 6

Russian Journal of Electrochemistry, Vol. 37, No. 8, 2001, pp. 786791. Translated from Elektrokhimiya, Vol. 37, No.

8, 2001, pp. 916921.

Original Russian Text Copyright 2001 by Evdokimov.

Kinetics of the Sodium Chlorate Formation in Electrolysis

of Chloride Solutions with Use of Dimensionally Stable Anodes
S. V. Evdokimov
Russian Federation Scientific Center Karpov Research Institute of Physical Chemistry,
ul. Vorontsovo Pole 10, Moscow, 103064 Russia
Received November 15, 1999; in final form, October 23, 2000

AbstractThe kinetics of the formation of sodium chlorate during electrolysis of chloride solutions with use
of dimensionally stable anodes is studied. An analysis and calculations suggest that the experimental results are
well described by the proposed set of differential equations under the assumption that the homogeneous chem-
ical reaction of chlorate formation has second, rather than third, order by active chlorine.

INTRODUCTION Scheme (II) with an irreversible chemical stage for-

mally corresponds to a fast electrochemical process
This work continues the research we initiated into
accompanied by a subsequent slow homogeneous
the electrochemical and corrosion processes that occur
chemical reaction of order p (the process was consid-
during chlorate electrolysis with use made of dimen-
ered in [610])
sionally stable anodes DSA [1, 2].
In the conditions of chlorate electrolysis on porous Red = M M + ne, (III.1)
DSA, there occurs a fast primary process of the forma-
tion of chlorine molecules, which then undergo hydrol- pM = Ox, (III.2) III
ysis, with the subsequent slow homogeneous chemical provided the rate of the reverse chemical stage of con-
reaction of the NaClO3 formation [35] version of Ox to M can be ignored in scheme (III),

where M is a stoichiometric coefficient by component
2Cl = Cl 2 + 2e, (I.1) M. As shown in [610], sufficiently far from equilib-
rium, the Tafel slope of a polarization curve (PC) for
Cl 2 + H 2 O = HClO + Cl + H , (I.2) such an irreversible process in scheme III is
+ 2 M RT
HClO = H + ClO , (I.3) b = 2.3 -----------
- -------, (3)
p + 1 nF
2HClO + ClO = ClO 3 + 2H + 2Cl . (I.4) I while the reactions exchange current is
According to Ferster, the rate of a homogeneous nF 2
chemical reaction similar to stage (I.4) is described by i 0 = ------- ------------ v 0 c 0 D, (4)
M p + 1
the kinetic equation
where D is the diffusion coefficient for substance M in
v = k 0 [ HClO ] [ ClO ].
(1) the bulk solution, v0 is the chemical reaction rate under
open-circuit conditions, and c0 is the concentration of
As follows from (1), the homogeneous reaction of the component M at infinite distance from the electrode.
NaClO3 formation has third order by active chlorine
(AC) [3]. The AC concentration [Cl*] is In view of the above, for conditions of chlorate elec-
trolysis, the Tafel slope must equal 2.3RT/4F, because
[ Cl* ] = [ HClO ] + [ ClO ].

(2) n = 2, HClO = 1, and p = 3.
Formally, the NaClO3 formation can be represented
as involving two successive stages. Species of AC form THEORETICAL MODEL
in the first, fast and reversible, electrochemical stage. In When studying the NaClO3 formation kinetics, one
the second, slow, homogeneous reaction, these irrevers- can employ not only the method of steady-state PC, but
ibly convert to NaClO3 via a third order reaction: a non-steady-state method as well. The latter involves
analysis of relevant kinetic curves, specifically, time
Cl = Cl* + 2e, (II.1) dependences of concentrations of AC and NaClO3.
Suppose that a pure chloride solution is used in
3Cl* ClO 3 . (II.2) II experiments. By such a solution we mean an electrolyte

1023-1935/01/3708-0786$25.00 2001 I Nauka /Interperiodica


where initial concentrations of AC and NaClO3 are E, V

zero. 1.38
Once a polarization is imposed, in accordance with
scheme (II), AC starts to form in the solution at the rate
W0 = I/2FV, at the expense of electrolysis. Simulta-
neously, AC will be spent at the rate Wp = 3k[Cl*]p, due 1.34
to the NaClO3 formation at the rate W = k[Cl*]p. Quan-
tities Wp and W differ by a factor of three, because three
species of AC participate in the formation of one chlo-
rate ion. In these formulas, p is the reaction order by 1.30
AC, which can be other than three; = CE/100, where
CE is the overall current efficiency by AC and NaClO3,
which can vary with time; I is the current; V is the over- 3 2 1
all electrolyte volume; and k is an effective rate con- logi [A cm2]
stant, which can depend on pH. We assume that the
chloride concentration and the solution volume Fig. 1. Anodic PC for DSA (5 g m2 Ru, 2 cm2), recorded
undergo no substantial changes during the time period in 300 g l1 NaCl + 0.2 g l1 K2Cr2O7 solution of pH 6 to 7
it takes [Cl*] to reach a steady-state value. The time at a steady-state AC concentration and 90.
dependences of volume concentrations of AC and
NaClO3 satisfy the set of differential equations The sought-for parameters in (9) are varied until the
d[ Cl* ] I discrepancy between the theoretical and experimen-
---------------- = ----------- 3k [ Cl* ] ,
(5) tal dependences of v on [Cl*] is reduced to a mini-
dt 2FV
d [ ClO 3 ]

- = k [ Cl* ] .
------------------- (6)
Using the time dependence of CE and concentrations of The electrochemical behavior of DSA in a chlorate
AC and NaClO3, one can determine k and p by search- electrolysis was studied in a setup with common anodic
ing for optimum parameters in (5) and (6). To this end, and cathodic compartments [11]. The cathode was
one has to have an experimental dependence v = made of titanium, on which a protective porous film of
v([Cl*]) in an explicit form, where v is the rate of vari- chromates had been formed, the film substantially
ations in [Cl*] with time. Such a dependence can be diminishing the AC reduction current.
found by differentiating an experimental time depen- The experiments, in which current efficiencies for
dence of [Cl*], provided one succeeded in deriving AC and NaClO3 were determined, were carried out in
approximating functions for the experimental curves. the following manner. Upon pouring 600 ml of a work-
To the first approximation, one can describe the ing solution (5 M NaCl + 0.2 g l1 K2Cr2O7) into the
obtained experimental time dependences by polynomi- setup, we imposed polarization on the electrolyzing cell
als (I = 4.2 A) and started supplying an HCl solution
diluted in a 1 : 1 ratio. In the course of experiment, solu-
[ Cl* ] = a t , [ ClO 3 ] = b t .
r r
r r (7) tion samples were collected at certain time intervals. To
(r) (r) prevent chemical conversion of AC into NaClO3, the
samples were rapidly cooled with ice to room tempera-
According to linear programming (linear regression) ture. Then the overall content of [HClO] + [NaClO] +
calculations, the time dependences of concentrations of [NaClO3], was determined in these solutions by iodo-
AC and NaClO3 satisfy polynomials of fifth and third metric titration with concentrated hydrochloric acid.
order, respectively. The NaClO3 concentration in similar portions of solu-
From (7) and the polynomial coefficients, we com- tions was determined by separate iodometric titration
pute v(t) with dilute sulfuric acid.
d [ Cl* ]
a rt
v = ---------------- = . (8)
dt r
Figure 1 shows a typical anodic PC recorded on
Then we utilize the available time dependences of [Cl*] standard DSA in a 5 M NaCl + 0.2 g l1 K2Cr2O7 solu-
and v to construct an experimental dependence of v tion of pH 6 to 7 at 90 in conditions of chlorate for-
on [Cl*]. This dependence is described by the equation mation at a steady-state concentration of AC in solu-
v = ----------- 3k [ Cl* ] .
p 1 The experimental part of the work was carried out with the partic-
2FV ipation of V.V. Gorodetskii.



c, M CE, %
0.20 150

0.15 100

0 50 100 150 200 250
t, min
0 50 100 150 200 250
t, min Fig. 3. Time dependence of overall CE for AC and NaClO3
for 300 g l1 NaCl + 0.2 g l1 K2Cr2O7; pH 8 to 9, 90.

Fig. 2. Time dependences of concentrations of (1) AC and

(2) NaClO3 for 300 g l1 NaCl + 0.2 g l1 K2Cr2O7; pH 8 to the background of a higher AC concentration. Hence,
9, 90.
when searching for parameters k and p via (9), we will
assume 1.
tion. The electrodes open-circuit potential was 1.29 V. Using these data, equation (9), and nonlinear pro-
The PC has a long linear portion with a slope of gramming, we can find the rate constant for the NaClO3
32.7 mV, which is close to 2.3RT/2F, which corre- formation and the reaction order by AC. When search-
sponds to the preceding electrochemical stage (II.1) ing for optimum parameters, the initial value of p was
with an exchange current of about 1 mA cm2. Due to taken to be three. The calculation yielded optimum
the effect of the reverse process (chlorine ionization) order p = 2.2. If the initial approximation was p = 1, the
near the open-circuit potential, it proved impossible to result was the same, i.e. p = 2.2. The optimum rate con-
observe the lower Tafel portion with a smaller slope stant was the same in either case, specifically, 3k =
(2.3RT/4F), which corresponds to the subsequent slow 0.087 (with time expressed in minutes and volume, in
chemical stage (II.2). Thus the method of steady-state liters). The overall curve, computed with (9) for opti-
PC allows one to investigate only the kinetics of the mum parameters k and p is represented in Fig. 4.
preceding electrochemical stage. Information about
kinetic parameters of the homogeneous chemical reac-
tion of the NaClO3 formation can apparently be A MORE EXACT DEFINITION
obtained only with a non-steady-state method. OF THE THEORETICAL MODEL
We used such a non-steady-state method for study-
ing the kinetics of the NaClO3 formation at higher pH As the order of the NaClO3 formation reaction by
values of 8 to 9 (this was dictated by procedural consid- AC proved to be closer to two than to three, for a theo-
erations). Figure 2 shows time dependences of concen- retical model, instead of equations (5) and (6), with
trations of AC and NaClO3, which are typical for these allowance made for the approximation 1, we should
conditions. The symbols in Fig. 2 represent experimen- write the equations
tal points, and the lines were calculated with equations d [ Cl* ] I
---------------- = ----------- 3k [ Cl* ] ,
(7). With time, [Cl*] increases and tends to a steady- (11)
state value [Cl*]1, which must correspond to a zero dt 2FV
right-hand part in (5)
d [ ClO 3 ]

- = k [ Cl* ] .
I 1/ p ------------------- (12)
[ Cl* ] 1 = -------------- . (10) dt
From (12), we obtain the equation for the time
The NaClO3 concentration steadily increases with time dependence of the NaClO3 concentration:
in accordance with (6).

[ ClO 3 ] = k [ Cl* ] dt.
Figure 3 shows an experimental time dependence of (13)
overall CE (by AC and NaClO3). As seen, throughout
the entire experiment, overall CE of about 8090% Let us introduce the designation
remains nearly constant. The marked deviation of the
first two points from 100% is due to the inaccuracy of I
a = ----------- . (14)
determination of a low NaClO3 concentration against 2FV



Solution of (11) reduces to the standard integral v, mol l1 min1

a exp ( 2 3akt ) 1
[ Cl* ] = ------ ------------------------------------------ . (15)
3k exp ( 2 3akt ) + 1
Equation (15) explicitly defines the sought time depen- 0.002
dence of the AC concentration. As follows from (15), at
infinite time, [Cl*] tends to its steady-state value [Cl*]1
[ Cl* ] 1 = ------ . (16)
Equation (16) offers a simple means for evaluating the
rate constant through experimentally measured value of 0 0.1 0.2 c, M
the steady-state AC concentration
Fig. 4. Rate of change in [Cl*] vs. current concentration of
AC in 300 g l1 NaCl + 0.2 g l1 K2Cr2O7; pH 8 to 9, 90.
k = ------------------. (17)
3 [ Cl* ] 1
Thus, the combination of the above equations allows
Substituting (15) into (14) and integrating, we obtain one to calculate basic characteristics of a chlorate elec-
the sought time dependence of the NaClO3 concentra- trolysis.
As the reaction order by AC is equal to two, and not
[ ClO 3 ]

to three, reaction (I.4) is not an elementary one. Rather,
it represents, most likely, a sequence of bimolecular
= ------------------------------------------------------- + ( a/3k )t + const.
2 a/3k steps. With this assumption in mind and taking into
3k' ( exp ( 2 3akt ) + 1 ) account that no chlorate forms in pure HClO solutions
[4], we can divide the scheme of reaction (I.4) into two
Given the initial condition [ ClO 3 ] = 0 at t = 0, we find
the integration constant
HClO + ClO = ClO 2 + H + Cl ,
const = ( 1/3 ) a/3k. (19) +
HClO + ClO 2 = ClO 3 + H + Cl , (I.4b)
With the time dependences of concentrations of AC and
NaClO3 available, we can calculate the rates of their and assume that the process rate is limited by stage
increase with the equations (I.4a), while stage (I.4b) is fast. Then, the kinetic equa-
tion for the process rate takes on the form
v ( Cl* ) = a 3k [ Cl* ] ,
v = k 0 [ HClO ] [ ClO ].

v ( ClO 3 ) = k [ Cl* ] .
We can express concentration [ClO] in (27)
Then we can find currents corresponding to the forma- through [HClO] and [H+] from equilibrium condition
tion of AC and NaClO3 (I.3):
I ( Cl* ) = 2F v ( Cl* )V , (22) K D [ HClO ]
[ ClO ] = --------------------------,

[H ]
I ( ClO 3 ) ( ClO 3 )V ,

= 6F v (23)
where KD is the dissociation constant for hypochlorous
where V is the solution volume where these reactions acid. Equation (27) must be expressed through [Cl*], as
occur. We can also compute relevant CE for AC and one can measure experimentally [Cl*], rather than indi-

NaClO3, CE(Cl*) and CE( ClO 3 ), respectively: vidual concentrations of hypochlorous acid and
hypochlorite. From (2) and (28), we find [3]
I ( Cl* )
CE ( Cl* ) = ---------------- 100 %, (24)
[ H ] [ Cl* ]
[ HClO ] = ------------------------
-. (29)
[H ] + KD
I ( ClO 3 )

CE ( ClO 3 ) = ------------------- 100 %.

(25) Substituting (28) and (29) into (27), then
k 0 K D [ H ] [ Cl* ]
+ 2
The overall CE for AC and NaClO3 is
v = ---------------------------------------
-. (30)
CE = CE ( Cl* ) + CE ( ClO 3 ). ([H ] + KD)



CE(Cl*), % logc1 [M]



5 7 9 pH
0 50 100 150 200 250
t, min
Fig. 6. The pH dependence of steady-state AC concentra-
tion (based on data given in table) in 300 g l1 NaCl +
Fig. 5. Time dependence of CE for AC for 300 g l1 NaCl + 0.2 g l1 K2Cr2O7 at 90: (symbols) experiment and (solid
0.2 g l1 K2Cr2O7; pH 8 to 9, 90. line) calculation.

This implies that the effective rate constant of the pH. This conforms to relation (31), from which it fol-
NaClO3 formation is pH-dependent lows that the effective rate constant increases with [H+]
by first order, provided [H+ ] KD. At [H+ ] KD , a fur-
k0 K D[H ]
ther decrease in solution pH will make the rate constant
k = -------------------------------
-2 . (31) of the NaClO3 formation decrease (our estimate yielded
([H ] + KD)
KD = 6.6 108 at 90).
As follows from (16) and (31), at [H+ ] KD, the pH
COMPARING THE CORRECTED MODEL dependence of the logarithm of the steady-state AC
WITH EXPERIMENT concentration is linear, with a slope of 0.5. To verify
this conclusion, the relevant experimental dependence
When calculating dependences of concentrations of (Fig. 6) was approximated with the equation
AC and NaClO3 with (15) and (18), we employed the
optimum constant 3k = 0.055 mol l min1, which had log [ Cl* ] 1 = a + bpH. (32)
been found using a nonlinear programming method.
The curves thus calculated coincided with relevant The optimum coefficients, found with a linear regres-
experimental and calculated curves (Fig. 2). Figure 5 sion method, are a = 3.726 and b = 0.36. When deter-
shows time dependences of CE for AC: experimental mining these parameters, we ignored the first point,
(symbols) and calculated with (20), (22) and (24) (solid because [Cl*] is pH-independent at pH 66.5. The
line) and demonstrates good agreement between the slope value 0.36 we obtained is in a reasonable agree-
model and experiment. Similar kinetic measurements ment with its expected theoretical value of 0.5.
were taken at lower pH values of 7.47.8 (table). Similar non-steady-state measurements were car-
The calculations we performed showed that, in a ried out at pH 6 to 7. The conclusion was that a steady-
chlorate electrolysis, the rate constant of the NaClO3 state AC concentration occurs in this case much sooner,
formation increases to 3k = 1 with decreasing solution which makes it difficult to study the NaClO3 formation
kinetics. That is why we performed all the measure-
ments in this work at higher pH and tested basic proce-
Time dependences of concentrations and CE of AC and dural questions connected with the obtaining of exper-
NaClO3, measured in 300 g l1 NaCl + 0.2 g l1 K2Cr2O7; imental results and their subsequent mathematical
pH 7.47.8, 90C treatment.
CE(Cl*), CE(NaClO3),
t, min [Cl*], M [NaClO3], M
10 1.75 102 1.83 103 83.75 26.3 The author thanks E.V. Kasatkin for useful critical
30 4.75 102 6.8 103 69.70 30.1 advice aired when discussing this work.
60 4 102 2.8 102 28.10 58.6
120 4 102 6.75 102 13.40 67.8 REFERENCES
180 4.5 102 1.23 101 9.57 78.4 1. Evdokimov, S.V., Elektrokhimiya, 1999, vol. 35, p. 366.
240 4.5 102 1.54 101 7.18 73.8 2. Evdokimov, S.V., Elektrokhimiya, 1999, vol. 35, p. 202.



3. Kokoulina, D.V. and Krishtalik, L.I., Elektrokhimiya, An Introduction), Moscow: Vysshaya Shkola, 1975,
1971, vol. 7, p. 346. p. 319.
4. Kheifets, L.I., Goldberg, A.B., and Mazanko, A.F., Itogi 8. Bagotzky, V.S., Osnovy elektrokhimii (Fundamentals of
Nauki Tekh., Ser.: Elektrokhimiya, 1983, p. 244. Electrochemistry), Moscow: Khimiya, 1988, p. 272.
5. Yakimenko, L.M. and Seryshev, G.A., Elek-
9. Antropov, L.I., Teoreticheskaya electrokhimiya (Theo-
trokhimicheskii sintez neorganicheskikh soedinenii (The
retical Electrochemistry), Moscow: Vysshaya Shkola,
Inorganic Electrosynthesis), Moscow: Khimiya, 1984,
1984, p. 324.
p. 35.
6. Vetter, K., Elektrochemische Kinetik, Berlin: Springer, 10. Rotinyan, A.L., Tikhonov, K.I., and Shoshina, I.A., Teo-
1961. reticheskaya elektrokhimiya (Theoretical Electrochem-
7. Damaskin, B.B. and Petrii, O.A., Vvedenie v elek- istry), Leningrad: Khimiya, 1981, pp. 268, 302.
trokhimicheskuyu kinetiku (Electrochemical Kinetics: 11. Evdokimov, S.V., Elektrokhimiya, 2001, vol. 37, p. 420.