Recharge Kinetics of the Porous Lead Dioxide Electrode : I

. The Effect of Structural Changes

Per Ekdunge and Daniel Simonsson

J. Electrochem. Soc. 1985, Volume 132, Issue 11, Pages 2521-2529.

doi: 10.1149/1.2113616

Email alerting Receive free email alerts when new articles cite this article - sign up in
service the box at the top right corner of the article or click here

To subscribe to Journal of The Electrochemical Society go to:

http://jes.ecsdl.org/subscriptions

© 1985 ECS - The Electrochemical Society

JOURNAL OF THE ELECTROCHEMICAL SOCIETY

ELECTROCHEMICAL
--- AND TECHNOLOGY
SCIENCE
~ NOVEMBER
1985

Recharge Kinetics of the Porous Lead Dioxide Electrode

I. The Effect of Structural Changes
Per Ekdunge and Daniel Simonsson
Department of Chemical Technology, The Royal Institute of Technology, S-I O0 44 Stockholm, Sweden

ABSTRACT
The changes in structure and kinetics during charge of the porous PbO2 electrode have been studied by means of
electrochemical and structure characterizing methods. It was found that the potentiostatic charge current after an initial
period of growth, due to nucleation and growth of PbO2~ is proportional to the decreasing amount of remaining lead sul-
fate. The polarization curves have a Tafel slope of about 60 mV/decade at overvoltages lower than 0.1V and about 180
mV/decade at higher overvoltages. The form of these curves can be explained by a mechanism with two consecutive
single-electron transfer reactions. The results from these specific studies of the positive electrode and charge experi-
ments with complete lead-acid cells show that the fastest way to charge lead-acid batteries is to charge at a cor~stant cell
voltage of a value just below that which leads to unacceptable gassing rates.

In spite of its great age, the cheap and reliable lead-acid sic kinetics, structural changes, current distribution, and
battery is still the single most used battery in the world. concentration changes is still waiting to be explored, es-
Consequently, it has been and still is the target of inten- pecially with regard to the anodic process. The aim of the
sive research. Th e research area has been reviewed by present work was to determine the relative importance of
Burbank et al. in 1971 (1) and by Rtietschi in 1977 (2). various structural and kinetic parameters for the rate at
Vinal's classic t e x t b o o k (3) on the subject from a techno- which lead-acid batteries can be charged. This is a ques-
logical point of view has been succeeded by B o d e s tion of high practical importance, not least in the applica-
"Lead-Acid Batteries," published in 1977 (4). In compari- tion as a power source in electric vehicles. The interest
son to the discharge process, less scientific interest has was focused on the positive electrode since in most sys-
been paid to the charge process. This topic has more of- tems this is generally the limiting one, both at discharge
ten been approached with t h e technological aim of and charge.
finding the optimal m e t h o d of charging or procedures for
very rapid charging. One of the results of these studies of Experimental
black-box character is the classical rule of t h u m b which The investigations were performed with commerical
goes by the name of the ampere-hour law and states that a battery electrodes from Noack, Sweden, with a thickness
battery can accept as m u c h current (in amperes) as it of 2.0 mm. The lead grid was alloyed with 2A-2.6% anti-
needs in capacity (in ampere-hours) until it reaches full m o n y and lesser amounts of arsenic (0.09-0.11%), tin
charge. (0.03-0.04%), and selenium (0.02%). Electrodes for the in-
On the other hand, fundamental studies have been vestigations were cut into the size 60 x 60 m m ~-. The
made of specific p h e n o m e n a in the electrodes of the working electrode was placed centrally in a Plexiglas cell.
lead-acid battery. Studies have thus b e e n reported on the Two lead electrodes were used as counterelectrodes. T he
electrocrystallizafion of lead dioxide from lead sulfate electrolyte was 5.0M H~SO4 in large excess. The electrode
formed on smooth electrodes of electrodeposited lead di- potential was measured relative to a Hg/Hg~SO4 reference
oxide or lead (5-8). electrode. The potentiostatic experiments were per-
A lot of research has also been devoted to the electrode formed with a Wenking 68TS10 potentiostat. In order tc
kinetics of t h e lead dioxide electrode (e.g., 9-13), but it has obtain a uniform lead sulfate distribution in the elec-
been mainly with simplified systems with a plane elec- trodes, they were discharged with a low current density
trode and where the sulfuric acid has been substituted ( - 2 0 A/m~) in the standard case. The discharge was inter-
with an electrolyte which does not give a solid product rupted at a potential of 0.9V. The electrode was at rest for
such as lead sulfate. A real battery electrode deviates sig- 2h before the charge.
nificantly from the majority of systems used in these For comparisons, recharge experiments were also per-
studies. A main difference is its porous structure, which formed with complete lead-acid cells with electrodes of
introduces the additional complexities of current distri- the full size. Six cells were assembled; each with three
bution and concentration gradients in the pores. The po- positive and four negative plates. In five of these cells,
rous structure is in itself of decisive importance for the the distance between the electrodes was 5 m m and was
electrode kinetics. Specific surface area and pore size controlled by vertical separator tubes of the type used for
distribution are important structural parameters. These tubular electrodes. In one cell, the electrodes were closely
parameters are not constant in the porous lead dioxide packed with ordinary separators between them. The cells
electrode because of the structural transformations that were made of Plexiglas. They were closed and gastight;
occur between the two solid phases: lead dioxide and evolved gases were allowed to escape from the cell
lead sulfate. These structural changes have been studied through an outlet connected to an oxygen sensor (Servo-
in optical and scanning electron microscopes (14-19) and m ex OA250). The gas evolution rate was measured volu-
with structure characterizing methods (19-20). metrically with a graduated tube. The temperature was
The overall kinetic behavior of the porous le~d dioxide measured with a t h e r m o m e t e r inserted through the lid via
electrode as a result of the mutual dependences of intrin- a sealing rubber stopper. A thermostat bath was used to

2521
2522 J. Electrochem. Soc.: ELECTROCHEMICAL SCIENCE AND TECHNOLOGY November 1985

control the temperature in runs at nonambient tempera- I I 1 1

ture.
In all experiments, the charges were recorded by elec-
tronic coulometers simulataneously with the overvoltage {x3
or current.
The structural changes were followed by visual obser- >
vations with SEM (Jeol JSM2) and by measuring the
changes in specific surface area according to the BET 40.2
method (Micromeritics Accusorb 2100 E) and pore size I 1 3
distribution by means of mercury porosimetry (Aminco
5-7118). Samples taken for these observations were ~ 0.1
washed according to a special procedure in order to avoid
large concentration gradients which can cause redistribu-
tion of the lead sulfate in a bipolar process driven by the
concentration elements thus created. The samples were 0
-3OO -200 -100 0
placed in a beaker filled with a stirred 5M H~SO4 solution. State of Charge, Ah/m 2
This solution was then progressively diluted with dis-
tilled water, which was added continuously with a rate Fig. 2. Overvoltage-time curves at galvanostatic charge of fully dis-
charged (curve 1) and partially discharged (curve 3) PbO2 electrodes.
adjusted so that all the sulfuric acid had been washed out
Current density = 200 A/m 2. Curve 2: charge with 20 A/m ~ to 80%,
overnight. In this way, the concentration gradients within followed by charge with 200 A/rn '2.
the porous electrode were limited to small values.
The changes in effective c o n d u c t i v i t y of the pore elec-
trolyte were determined from the ohmic voltage drop re- charged electrode which has been recharged to the same
quired to pass a given current through the unloaded elec- state of charge. It can also be noted that the inflection
trode. An x-ray diffraction analysis showed that the active point of curve 3 in Fig. 2 is at an overvoltage of approxi-
mass of the fully charged positive electrode consisted al- mately the same value as the m i n i m u m in the recharge
most entirely of fl*PbO~. curve (curve 1) of the fully discharged electrode.
The potentiostatic charge curves (Fig. 3) can be inter-
Results preted in the corresponding way. The falling potential
Galvanostatic and potentiostatic charges.--Charging
after the initial peak value at the galvanostatic charge has
its correspondence in a rising current at potentiostatic
with a constant current gave the overvoltage-time curves
shown in Fig. 1 and 2. In these curves the Ri-free over- charge. After a m a x i m u m (corresponding to the mini-
m u m in the galvanostatic charge curve), the current de-
voltage has been plotted against a nominal state of charge
defined as cays almost linearly with the state of charge. Since the
current efficiency is somewhat less than 100%, especially
at higher overvoltage, the nominal state of charge (q) is
q = qd + i dt [1] somewhat higher than the true state of charge. The
curves therefore do not level off at the oxygen evolution
current until q assumes positive values. Figure 4 shows
where qd is the charge output in the previous discharge. that the current density at potentiostatic charge is not a
The general form of these curves can be interpreted by state function with respect to the PbSOJPbO~ composi-
formally distinguishing four different regions. When the tion. The maximum current was roughly the same (within
constant current is applied to the fully discharged elec- 20%) irrespective of whether the electrode had been dis-
trode, the potential immediately rises to a peak value. The charged to I00, 70, or only 20% (curves I-3 in Fig. 4). When
overvoltage then rapidly decreases to a flat m i n i m u m at the electrode had been discharged with a high current
about 10-20% charge. In a third region, the overvoltage in- density (-I000 A/m2), the potentiostatic charge current
creases monotonically until it finally reaches the plateau was somewhat higher.
corresponding to the oxygen evolution reaction on the The rising transients in Fig. 3 and 4 and the falling over-
fully charged electrode. voltage in Fig. 1 and 2 can be ascribed to nucleation and
The potential peak decreases with the decreasing dis- growth of lead dioxide. This process governs the kinetics
charge capacity at increasing discharge current. The min- only up to 10-20% charge. The amount of charge con-
imum, on the other hand, first decreases with increasing sumed in this region is independent of the current or the
discharge current density and then increases at very large potential. An initial galvanostatic charge with low current
discharge currents.
Figure 2 shows that the charge curve is flatter at lower
current densities. The recharge of a partially discharged I _3 I I I I
electrode requires a lower overvoltage than a fully dis- 60O

I I I I I _ 5OO

> 0.3 1 2 40C

d 3
~r 300
0.2

o : m

o.1 o

lOO

0
State of Charge, Ah/m a 0 -400 - 300 -200 -100 0
Fig. 1. Overvoltoge-time curves at galvanostotic charge with 200 State of charge, Ah/m 2
A/rn 2 of on electrode discharged with - 3 0 (curve 1), - 1 0 0 (curve 2), Fig. 3. Potentiostatic charge of o PbO2 electrode at different over-
- 2 0 0 (curve 3), - 5 0 0 (curve 4), - 1 0 0 0 A/m ~ (curve 5). voltages: 100 (curve 1), 200 (carve 2), 266 mY (curve 3).
Vol. 132, No. 11 POROUS LEAD DIOXIDE ELECTRODE 2523

I I I I I at higher overvoltage, where the back reaction can be ne-

glected. Piecewise fits of this expression are in good
4 agreement with the experimental points in Fig. 5. At
400 % F~.2 s 5 --
lower overvoltage, a n ~ 1 and S,joL ~ 3 A/m s. The Tafel
slope increases abruptly at an overvoltage of 0.1V. Above
<c 300 this value, a n = 0.32 and S~joL = 47 A/m 2. Calculations
with a rate equation derived from Eq. [2] showed that the
current density distribution in the porous electrode was
r 2oo / practically uniform at all current densities. When the
overvoltage approaches a value of 0.3V, the side reaction
of oxygen evolution becomes significant. The Tafel slope
o
loo of 95 mV/decade found for this electrode reaction is in
agreement with the results of Rfietschi et al. (21). Figure 5
o I I I also shows the cathodic overvoltage measured in the
-400 -300 -200 -100 0 same way at different states of charge during the recharge
State of Charge, Ah/m 2 process.
Fig. 4. Potentiostatic charge of an electrode discharged under dif- Charging of full-size cells.--Preliminary experiments
ferent conditions. Curves 1-3: discharged with - 2 0 A/m ~ to 100 with potentiostatic charge of lead dioxide and lead elec-
(curve 1), 70 (curve 2), and 2 0 % (curve 3). Curves 4, 5: discharged trodes, respectively, showed that the positive electrode
with - 1 0 0 0 A/m ~ to 100 (curve 4), and 4 0 % (curve 5). Curve 6: dis- was the limiting one both at discharge and at charge. The
charged with - 2 0 A/m 2 to 100% and recharged galvonostatically behavior of the lead-acid cell would then be controlled
with 20 A/m ~ to 10%. mainly by its positive electrode. Experiments with com-
plete cells with full-size electrodes were therefore under-
density (20 A/m ~) to 10% followed by a potentiostatic taken in order to compare the behavior observed for the
charge resulted in a monotonically decreasing current co- single lead dioxide electrode with that of the complete
inciding with the decreasing current of the complete po- cell. Comparisons were made between charging with con-
tentiostatic charge curve (Fig. 4). The initial deviation be- stant current (I), constant voltage (U), and constant cur-
tween the two curves can probably be explained by the rent followed by constant voltage (IU), respectively. The
differences in electrolyte concentration profiles. The ris- voltage-time curves at charging with constant current are
ing transients were also observed at the potentiostatic shown in Fig. 6 together with the total volume of evolved
charge of onlY partially discharged electrodes, although gases. The c o m m o n feature of the different charge curves
its relative importance decreases with decreasing is a continuously increasing voltage up to a state of
discharge. charge where the voltage rises abruptly to a much higher
The dependence of overvoltage on current density at level while simultaneously the gassing rate increases rap-
different states of charge is shown in Fig. 5. These curves idly. A cell voltage of, e.g., 2.6V is reached after 71, 79, and
were obtained by applying different constant currents of 90% charge at 150, 100, and 50 AJm2, respectively. The
short duration at three different states of charge fraction remaining to be charged into the battery is thus
(20,50, andS0%), while the electrode was otherwise approximately proportional to the current density. This is
charged with a low constant current density in order to in accordance with the first-order dependence on lead
avoid nonuniform current distribution and the establish- sulfate as expressed in Eq. [2]. The galvanostatic charge
ment of concentration gradients. The overvoltage was curves in Fig. 6 exhibit a steeper rise in overvoltage than
corrected for the contribution from the increasing acid those of a single PbO~ electrode in Fig. 1.
concentration gradient by extrapolation back to time Charging with constant current until the cell voltage
zero. The anticipated linear decrease in current density reaches 2.50V followed by charging at constant voltage
with the increasing state of charge indicated by the poten- gave the results shown in Fig. 7. Compared to charging
tiostatic oxidations was tested by a rate equation of the with constant current, this procedure has the advantage
form that excessive gassing is automatically avoided by not al-
[ anF ) lowing the voltage to increase above a prescribed value.
i = S~j,,L (qlqd) exp ~ - - ~ - - ~ [2]
According to Fig. 7, the charging current decays expo-
nentially during the period of constant voltage. The
I I I I lilt I I I I I tilt
1,1t{v) /o ~ I I I I

o3L / o / ~ (
2,9 .e

l .o ./..,/. 2,8

2,7 ,so r)/

>=
2,6
2 100
O 2,5

2,4- , .2';
...~,~ _...= /.,,," I ~
50
I "1 / - . ~ ./ 1/ ]
i I I I IIIII i I t I I III I .o/ / / A r sA
010 50 100 500 2,2i
t
l il(A/n~-)
Fig. 5. Polarization curves for two different PbO2 electrodes fully 2,1
discharged with - 2 0 A/m ~ and recharged with 20 A/m s to 20 (trian- t ......... "-I'" i i
gles), 50 (squares), 80 (filled circles), and 120% (open circles) (oxy- -300 -200 -100 0
gen evolution). Solid lines: anodic current. Dashed line: cathodic cur- State of charge (Ah/m 2)
rent. The anodic polarization curves drawn for the partial charges are Fig. 6. Changes in cell voltage and accumulated gas volume at gal-
fits to the relation i = SmjLq/qd wherej is given by Eq. [12] with c~ r = vanostatic charge at different current densities. Discharge current
0.32, Sm2jo,rL = 47 A/m s, Jo,rlj.,o = 25, and % = 0.5. density = - 120 A/m 2.
2524 J. Electrochem. Sac.: E L E C T R O C H E M I C A L SCIENCE AND TECHNOLOGY N o v e m b e r 1985

\o I I I I I I
1000 \
1,5,

N"~
E

>.
i-
~100
Z
LM
"
.

9\
\

t 1,C

~ 05-
~ -
1

10 O--

I I I I 1 I
0 1 2 3 3,0 3,5 103/T
TIME (h) Fig. 8. Arrhenius plot of the charge rate coefficient defined by Eq.
Fig. 7. Current density vs. time at a charge with constant cell volt- [4].
age = 2.50V (squares), 22~176 and IU charging with current limi-
tation to 22A at different temperatures: 3.5~ (nablas), 25~ (trian-
gles), and 52~ (circles). tural transformations: occasionally, we could observe
how lead sulfate crystals were transformed into lead diox-
ide with the macrostructure preserved (Fig. 13). This type
experimental curves deviate from the straight In i-t lines of transformation has earlier been observed by Simon
at lower current densities, since the electrochemical gas et al. (14, 16). Interesting observations were made
evolution becomes a progressively higher fraction of the concerning the structural transformations of the larger
total, decreasing current. The disadvantage of IU charg- lead sulfate crystals formed on the exterior surface at low
ing is obviously that it does not take full advantage of the discharge rates. The lead dioxide formed during charge
high charge acceptance in the beginning. Charging at was held within the dimensions of the original lead sul-
constant voltage utilizes this possibility of high initial cur- fate crystals. A dense microcrystal]ine wall of lead diox-
rents. The curve for charging at a constant cell voltage of ide was formed which eventually was ruptured at excess
2.50V in Fig. 7 is somewhat convex after the maximum. charge and revealed an inside growth of small lead diox-
This is due to a temperature rise from 22 ~ to 29~ In the ide crystals into the form of a compact kernel (see Fig. 10
linear portion of the curve, the temperature varies only and 14-17).
between 29 ~ and 28~ The changes in specific surface area are of great impor-
When Eq. [1] and [2] are combined, we obtain the varia- tance when attempting to explain the changes in elec-
tion of current density with time trode kinetics. Although the BET method does not distin-
guish between the surface areas of PbO~ and PbSO4,
i = ii exp [ - k (t - t~) ] (V ~ const) [3]
important conclusions can be drawn when combined
where index 1 denotes the time after which Eq. [2] is valid,
and

(~nF~ ~ ,, ,
k = SmJoL exp --R~j/~-qd) [4]

The coefficient k is a rate coefficient which can be deter-

mined from the slopes of the linear portions of the curves
in Fig. 7. The temperature dependence of k at constant
cell voltage follows an Arrhenius relation (see Fig. 8).
Structural changes.--The morphological changes dur-
ing charge were followed by taking out samples at differ-
ent states of charge for visual observations by means of a
scanning electron microscope. These observations
showed that larger lead sulfate crystals were formed at
lower rates of discharge (Fig. 9 and 10). The size of the
lead sulfate crystals could, furthermore, be related to the
volumetric restriction imposd by the porous structure. On
the exterior surface, where there are no such restrictions,
the lead sulfate can grow to large, closely packed crystals
(Fig. 9 and 10). This also holds for larger voids in the po-
rous electrode (Fig. 11), while otherwise the lead sulfate
crystals were smaller and randomly dispersed in the po-
, 10 /Lmj
rous structure (Fig. 12). In most cases, no firm conclu- Fig. 9. Lead sulfate crystals formed on the exterior surface of the
sions could be drawn as for the mechanism of the struc- electrode at a discharge with - 2 0 0 A/m '2 to - 3 2 0 Ah/m 2.
Vol. 132, No. 11 POROUS LEAD DIOXIDE E L E C T R O D E 2525

Fig. 10. Lead sulfate crystals formed on the exterior surface at a Fig. 13. Formation of microporous lead dioxide within the dimen-
discharge with - 3 0 A/m z to -410 Ah/m2. sions of the original lead sulfate crystals formed during discharge.
Electrode discharged with - 3 0 Mm ~ and recharged with 200 A/m ~ to
- 4 8 Ah/m2.

Fig. 11. Formationof dense agglomerates of lead sulfate crystals in Fig. 14. SEM micrograph showing the beginning of the transforma-
a larger void. Electrode discharged with - 3 0 A/m 2. tion of large lead sulfate crystals into microcrystalline lead dioxide in
the form of shell and kernel.

Fig. 15. SEM micrograph showing the transformation of large sul-

Fig. 12. Morphology of the interior of the electrode. Discharge with fate crystals into microcrystolline lead dioxide in the form of shell and
- 3 0 A/m2 to - 3 5 0 Ah/m2. kernel.
2526 J. Electrochem. Soc.: E L E C T R O C H E M I C A L SCIENCE AND TECHNOLOGY N o v e m b e r 1985

I I I J 1 t I I I

6
QJ / I
E 5
d 4

I,-' 3
L,l,,t
m 2

, I 1 I I , I , i
50 100
Charge, %
Fig. 18. Changes in specific surface area during charge. The sur-
face area is given in the unit of square meter per equivalent gram of
PbO2.

Fig. 16. SEM micrograph of large sulfate crystals transformed into face area during charge can thus be attributed to the un-
microcrystalline lead dioxide in the form of shell and kernel. covering of this encapsulated PbO~, while at the same
time a large inner surface area of P b S Q is created inside
with other pieces of information. The changes in specific the PbSO4 crystals as they dissolve. Although residual
surface area during charge are illustrated in Fig. 18. The PbO2 may act as nuclei for the lead dioxide formed dur-
fully discharged electrode has a specific surface area ing charge, new nuclei will probably be formed in parallel
which is almost three times smaller than that of the fully at the high initial overvoltage. These new PbO2 crystals
charged. The latter is 6.4 mS/g, which according to the ge- will give an additional contribution to the surface area.
ometric relationship for spherical particles After the initially rapid growth of surface area, the
continued increase is much slower because the surface
Sp = 6/dp [5] area of PbO~ grows slower than linearly with the a m o u n t
of charge, while the surface area of PbSO4 decreases.
corresponds to a mean particle diameter of 0.1 /zm. The The changes in pore size distribution are plotted in Fig.
mean particle diameter of the fully discharged electrode 19. The fully charged electrode had a total porosity of
can, in the same way, be calculated to be 0.4/zm. These
56%, including closed pores and pores smaller than 0.035
values can be comPared to those estimated from micro-
~m, and a mean pore diameter of but 0.15 ~m. Using a cy-
graphs, which are about 0.1 /zm for the lead dioxide and lindrical pore model, the porosimetry data can also be
1-10/zm for the lead sulfate. The specific surface area of used to estimate the specific surface area. The values ob-
the fully discharged electrode must, therefore, to a major tained in this way were 40-50% lower than the corre-
extent, be attributed to residual lead dioxide which has sponding values measured by the B E T method. Besides
not been covered with PbSO4 but is either electrically in- the inaccuracy of the cylindrical pore model, part of this
sulated or electrochemically nonactive. The dashed line discrepancy can be attributed to pores smaller than 0.035
in Fig. 18 represents the change in specific surface area ~m, which contribute little to the pore volume but appre-
of the lead dioxide only according to a structural model ciably to the specific surface area.
assuming sPherical particles shrinking uniformly during The changes in the porosimetry curves in Fig. 19 also
discharge. The measured surface area of the fully dis- reflect the fact that the molar volume of lead sulfate (48.9
charged electrode is lower than that predicted by this cm3/mol) is almost twice that of lead dioxide (25.5
model. This is also to be expected since microscopy stud-
cm:~/mol). The plugging effect of the lead sulfate in the
ies have revealed that much of the PbO~ remaining at the
pores affects the effective transport coefficients for diffu-
end of discharge have been encapsulated within PbSO4 sion and migration. The changes in effective conductivity
crystals (14, 16). The initially rapid growth of specific sur- are plotted in Fig. 20. As expected, the effective conduc-
tivity increases during charge, due to the opening up of
pores, but then reaches a flat m a x i m u m at about 80%
charge. The labyrinth factor ~ = ~f~-/K,r~ of the fully
charged electrode is 0.15 including the effect of the grid
and about 0.19 after correction for the grid, which
occupies about 20% of the total projected electrode area.

Discussion a n d C o n c l u s i o n s
The results should be interpreted with respect to both
the individual factors and their mutual dependence. It
can thus be concluded that the initial part of the charge
process is governed by the formation of electrochemically
active PbO~ surface. As already discussed, Fig. 18 indi-
cates that this occurs not only by nucleation and growth
of new PbO~ crystals, but also by the uncovering of en-
capsulated or covered residual lead dioxide. This hypoth-
esis is also supported by the observation that the duration
of this process is independent of current density or
overvoltage and that it does not affect the later period of
the recharge.
The rising potentiostatic transient reaches a m a x i m u m
after recharge to only 10-20%. While this first, shorter pe-
riod is controlled mainly by the available surface area of
Fig. i 7. Detail of Fig. 16 lead dioxide, the major portion of the charge process
Vol. 132, No. 11 POROUS LEAD DIOXIDE ELECTRODE 2527

I f where the latter reaction in turn is composed of several el-

ementary steps. Vetter's analysis can be extended by tak-
0,5 ing into consideration the interplay of diffusion and in-
trinsic electrode kinetics in a rate equation. Assuming
3 first-order, purely consecutive reactions, the kinetic
0,4- equation at high overvoltage for this scheme would be of
the form
2
~nF ]
Sj0 exp (
.~ 0,3 --
i=
\ --RF-,) [8]
S2jo ( 1 L ) [ omF ]
0,2 -- 1+~ ~ + DceqS1-----~.~ exp ~--R-~-~?)
o.
Equation [8] takes into account the dissolution rate of
PbSO4 and the diffusion of Pb 2~ ions over a characteristic
0,1 - -
distance L and across a mean cross section $1.~, which is
some weighted average of the surface areas of PbSO4 ($1)
I I and PbO~ ($2). When the dissolution and diffusion rates
oo,1 o,I I Io are high in comparison to the charge transfer rate, the
Pore diameter, ~ m second term in the denominator vanishes and the current
is proportional to the surface area of lead dioxide. It is
Fig. 19. Pore size distribution of o porous Pb02 electrode. 1: Fully
only at the limiting current density, when the reaction
discharged. 2: 50% charged. 3: Fully charged.
rate is controlled by the dissolution and/or diffusion, that
the current is proportional to the surface area of lead sul-
fate. Equation [8] thus predicts ~-i curves for different
seems to be controlled by the availability of lead sulfate.
states of charge which approximately coincide at low cur-
It is obvious that the anodic and cathodic reactions are
rent densities, diverge at higher current densities, and
different in their dependence on the structural changes
finally reach their different limiting values appearing as
and cannot be microreversible at the current densities
vertical lines in the ~-i diagram. The different curves in
employed. The cathodic polarization curves in Fig. 5 do
Fig. 5 obviously do not follow this pattern. The curves at
not differ significantly from each other, which indicates
different states of charge do not coincide until very close
that the working surface area of the active lead dioxide is
to the equilibrium potential, and no distinct limiting cur-
relatively constant after a recharge to 20%. This is in
rents can be observed at higher overvoltages.
agreement with the results from the surface area
An alternative kinetic model takes into account that dif-
measurements (Fig. 18) when recalling that the main frac-
fusion of Pb 2~ ions and their reaction occur in parallel as
tion of this total specific surface area can be attributed to
the lead dioxide. well as in consecutive steps. The physical picture is then
similar to that in a heterogeneous catalyst particle (23). An
Contrary to this, the anodic reaction is apparently first
order of magnitude calculation shows, however, that even
order with respect to the a m o u n t of remaining lead sul-
in this case, the diffusion rate of lead ions is sufficient to
fate. The charge reaction, therefore, seems to be confined
balance their consumption rate in the charge reaction.
to a zone very close to the surface of the lead sulfate crys-
The switch from a low to a high Tafel slope in Fig. 5 can
tals. An evident explanation would be that the reaction
be explained by a mechanism with two consecutive
proceeds through a solid-state mechanism. On the other
single-electron transfer reactions. This case has been
hand, it is more commonly believed that the transforma-
treated by Vetter (24) for the general scheme
tions between lead sulfate and lead dioxide occur via a
dissolution precipitation mechanism involving dissolved S~S~+ e- [9]
Pb 2~ ions. Vetter (22) has shown that such a mechanism is
possible in the lead dioxide electrode, and that the volu- Sm ~ So + e- [10]
metric diffusion rate of P b 2§ ions over the microscopic
distances is sufficiently high to allow an efficient utiliza- Sr ~- So + 2e- [1]]
tion of the whole PbO2 surface. According to this mecha-
nism, the oxidation reaction can be formally written as Writing down the rate equation for each step and
eliminating the concentration of the intermediate Sm be-
PbSO4 (s) --->P b ~ + SO4~ [6] tween the two equations give the following relation be-
tween current density and overvoltage
P b 2~ + 2H~O --. PbO2 (s) + 4H ~ + 2e [7]
j = 2j,,,, . e x p ~--~-~}
I I I I $

2F
1 - exp (
--h-~-~)
TE 10 - ..f"--' l+j,,,r exp(
\

l+~o-~r )
[12]
Jo.o RT F~?
where J0.r and j,.. are the exchange current densities and
a,. and O~oare the charge transfer coefficients of reactions
[9] and [10], respectively. At high anodic overvoltages, Eq.
O
[12] is simplified to
BO
.>
"6 j = 2j0.r exp ~ - - - ~ ) [13]

The geometric current density of the porous electrode at

uniform current distribution would then be
o I ~ I J
-400 -300 -200 -100
State of Charge, Ah/m 2 i = SmL(q/qd) 2j,),.9 exp ( --R-T
~rF n ) [14]
Fig. 20. Changes in effective conductivity during charge. Two dif-
ferent electrodes. which is similar to Eq. [12]. The values of the charge
2528 J. Electrochem. Soc.: E L E C T R O C H E M I C A L S C I E N C E A N D T E C H N O L O G Y N o v e m b e r 1985

transfer coefficient ar and the pre-exponential factor can odic reaction proceeds via a dissolution-precipitation
thus be explicitly determined from the slope and inter- mechanism. The Tafel slope even indicates that the pre-
cept of the Tafel line at overvoltages higher than 0.1V. cipitation of lead sulfate from Pb 2~ and SO4'-'- ions might
The obtained value of the charge transfer coefficient, a, be the rate-determining step at sufficiently high current
= 0.32, is rather low and indicates an unsymmetrical en- densities.
ergy barrier. Assuming that ao = 0.5 and using the quo-
tient JoflJo.o as a fitting parameter, the complete polariza- Acknowledgments
tion curve can be constructed from Eq. [12] multiplied by This work was financially supported by the National
SmL(q/qd). The fitted curves are drawn as solid lines in Swedish Board for Technical Development (Project no.
Fig. 5 with a fitted value ofj0.r/J0.o = 25. 83-3262). The SEM studies were performed by Mats
The reality corresponding to the formal Eq. [9] and [10] Hallberg.
is not possible to elucidate from current density overvolt- Manuscript submitted Dec. 5, 1984; revised manuscript
age curves only. Several possible mechanisms can be pro- received May 21, 1985. This was Paper 22 presented at the
posed, but the discrimination between these requires fur- New Orleans, Louisiana, Meeting of the Society, Oct. 7-12,
ther information, for example, about electrochemical 1984.
reaction orders.
The situation is more complicated when the cathodic The Royal Institute of Technology assisted in meeting
the publication costs of this article.
Tafel-line in Fig. 5 is taken into consideration. It is re-
markably straight in spite of the fact that the current dis-
tribution in this case with a low Tafel slope would be- LIST OF SYMBOLS
come successively more n o n u n i f o r m with an increasing
current density, thus leading to an increasing Tafel slope. ceq equilibrium concentration of Pb 2~ ions (mol]m :~)
It has a slope of about 30 mV/decade, which is consistent D diffusion coefficient of Pb ~ ions (m'-'/s)
dp particle diameter (m)
with results obtained on plane electrodes (25). This slope geometric current density (A/m 2)
is consistent with a rate-determining chemical step after i0 apparent geometric exchange current density
both electrons have been transferred. One thinks, on the (A/re'-')
first hand, of the precipitation of lead sulfate. The ob- j true current density (A/m2)
served value of the cathodic transfer coefficient is not 3o true exchange current density (Aim~)
k charge rate coefficient (h-')
consistent with the transfer coefficients obtained for the ko rate constant for the dissolution of PbSO4 (moYm~/s)
anodic reaction. The value of the parameter S~joL, ob- L thickness of one symmetric half of the porous elec-
tained from the intercept with the log i axis is about 3 trode (m)
A/m ~, which is in approximate agreement with the value n n u m b e r of electrons transferred in the electrode
ofj0.o obtained for the anodic reaction. On the other hand, reaction
q a m o u n t of charge remaining to be recharged
the exchange current density jo obtained on a flat elec- (Ah/m'~)
trode of electrodeposited fl-PbO2 is of the order of 0.1 q~ charge output in the previous discharge (Ah/m'-')
A/m ~ (24). With S~ = S~ ~ 107 m - ' and L = 10-~m, this S specific surface area (m~/kg)
value of Jo gives an estimated value of S~oL = 103 Aim2. S~ specific active surface area ( m - ' )
Even after a correction for the roughness factor of the S, specific surface area of PbSO4 ( m - ' )
plane electrode, this value is orders of magnitude higher $2 specific surface area of PbO2 (m-')
S,.2 mean specific cross-section area for the diffusion of
than that obtained for the porous electrode. Pb 2~ ions
The inconsistency between the anodic and cathodic
branches at higher overvoltages is difficult to explain in Greek Letters
view of the reversiblity of the lead dioxide electrode. a charge transfer coefficient
Probably, a change in mechanism occurs in the region of charge transfer overvoltage (V)
lower anodic overvoltages: for example, a change from a p density of electrode material (kg/m :~)
solid-state to a dissolution precipitation mechanism. This Subscripts
hypothesis is partly supported by the SEM micrographs,
from which the latter mechanism seems to be the more r first electron transfer step
probable at the cathodic reaction, while the first-order o second electron transfer step
dependence on lead sulfate at anodic overvoltages above REFERENCES
30 mV is in favor of a solid-state mechanism. 1. J. Burbank, A. C. Simon, and E. Willihnganz, in "Ad-
The SEM micrographs, unfortunately, do not give a n vances in Electrochemistry and Electrochemical
unequivocal discrimination between these two mecha- Engineering," Vol. 8, P. Delahay, Editor, p. 157
nisms at recharge. The formation of the core-shell struc- (197!).
ture in Fig. 16-17 evidently requires transfer of lead spe- 2. P. Rfietschi, J. Power Sources, 2, 3 (1977/1978).
cies via the solution, while the structure in Fig. 13 is more 3. G. W. Vinal, "Storage Batteries," 4th ed., J o h n Wiley
and Sons, New York, (1955).
consistent with a solid-state reaction. 4. H. Bode, "Lead-Acid Batteries," Wiley-Interscience,
I n conclusion, the experimental results can be com- New York (1977).
bined to give an understanding of the effect of the struc: 5. M. Fleischmann and H. R. Thirsk, Trans. Faraday
rural changes on the electrode kinetics of the porous lead Soc., 51, 71 (1955).
dioxide electrode. During discharge, the cathodic reac- 6. S. G. Canagaratna, P. Casson, N. A. Hampson, and K.
Peters, J. Electroanal. Chem., 79, 281 (1977).
tion is distributed (although not uniformly) over t h e en- 7. P. Casson, N. A. H a m p s o n , and K. Peters, ibid., 83, 87
tire PbO~ surface. As the discharge proceeds, the PbSO4 (1977). -n
crystals grow larger, overlap, and merge into larger but 8. N. A. H a m p s , , S. Kelley, and K. Peters, J. Appl.
fewer crystals. The total surface area of PbSO4 will then Electrochem., 10, 261 (1980).
grow only slowly. This explains why in Fig. 4 approxi- 9. D. Spahrbier, Dissertation, Technische Hochschule
mately the same m a x i m u m current is obtained at poten- Stuttgart, Stuttgart, Germany (1960).
10. N. A. Hampson, P. C. Jones, and R. F. Phillips, Can. J.
tiostatic charge of fully and partially discharged elec- Chem, 45, 2045 (1967).
trodes. 11. N. A. Hampson, P. C. Jones, and R. F. Phillips, ibid.,
At the same time, the amount of lead dioxide decreases 46, 1325 (1968).
and thus a successively larger amount of charge is needed 12. N. A. Hampson, P. C. Jones, and R. F. Phillips, ibid.,
for the formation of new lead dioxide i n the electro- 47, 2171 (1969).
13. H. Ogura and T. Sekine, Denki Kagaku, 41, 638 (1973).
crystallization process. The rising transients in Fig. 4, 14. A. C. Simon, C. P. Wales, and S. M. Caulder, This Jour-
therefore, have a longer duration after a deeper discharge. nal, 117, 987 (1970).
The morphology of the lead sulfate crystals formed dur- 15. A. C. Simon, S. M. Caulder, and J. T. Stemmle, ibid.,
ing discharge, especially the larger crystal size in larger 122, 461 (1975).
voids and on the exterior surface, indicates that the cath- 16. A. C. Simon and S. M. Caulder, in "Power Sources 5,"
Vot. 132, No. 11 POROUS LEAD DIOXIDE ELECTRODE 2529

D. H. Collins, Editor, p. 109, Academic Press, Lon- 20. W. Kappus and A. Winsel, ibid., 8, 159 (1982).
don (1975). 21. P. Rfietschi, J. B. Ockerman, and R. Amlie, This Jour-
17. S. Hattori, M. Yamaura, M. Kohno, Y. Ohtani, M. Ya- nal, 107, 325 (1960).
mane, and H. Nakashima, in " P o w e r Sources 5," 22. K. J. Vetter, Chem.-Ing.-Tech., 45, 213 (1973).
D.H. Collins, Editor, p. 139, Academic Press, Lon- 23. C. N. Satterfield, "Mass Transfer in Heterogeneous
don (1975). Catalysis," Chap. 3, M.I.T. Press, Cambridge, MA
18. P. Casson, N. A. Hampson, K. Peters, and P. Whyatt, J. K(1j97~e
Electroanal. Chem.. 93. 1 (1978). 24. . . tter, "Elektrochemische K i n e t ~ , " Springer
19. P. Reinhardt, M. Vogt, and K. Wiesener, J. Power Verlag, Berlin (1961).
Sources, I, 127 (1976/1977). 25. D. Simonsson, This Journal, 120, 151 (1973).

Recharge Kinetics of the Porous Lead Dioxide Electrode

II. The Effect of Sulfuric Acid Concentration
Per Ekdunge and Daniel Simonsson
Department of Chemical Technology, The Royal Institute of Technology, S-1 O0 44 Stockholm, Sweden

ABSTRACT
Anodic polarization curves for partially recharged porous lead dioxide electrodes have been recorded at different
sulfuric acid concentrations. These curves can be fitted by the rate equation for two consecutive single-electron transfer
reactions. The dependence of the two exchange current densities on the sulfuric acid concentration is strongly negative.
The apparent electrochemical reaction order for the first transfer step approaches - 3 as the sulfuric acid concentration
approaches 7M. The polarization curves also depend on the acid concentration during the previous discharge. At the
same current density, the overvoltage at recharge in 5M H2SO4 was much higher when the electrode had been dis-
charged in 1M H2SO4 compared to a discharge in 5M H2SO4. This effect can probably be attributed to the larger lead sul-
fate crystals formed at lower sulfuric acid concentrations.

In the previous paper (1), the recharge kinetics of the 1). This cell design with its two symmetrical Luggin capil-
porous lead dioxide electrode in 5.6M H~SO4 was studied laries was also suitable for the determination of the effec-
with regard to the effects of structural changes. The pres- tive conductivity and for diagnosing the current distribu-
ent paper deals with the effects of changes in sulfuric acid
concentration. Although of great importance, this param-
eter has not been studied often before. Many investiga-
tions of the electrode kinetics of the PbO~ electrode have
been made with electrolytes g i v i n g soluble Pb (II) spe-
cies, e.g., HC104 and NaC104 (2-5) or HNO:~ (6), since this is
the only way to vary the concentrations Of Pb 2~ and H ~ in-
dependently and at a constant ionic strength. The results
of these studies are not directly applicable to the positive
electrode in the lead-acid battery in which the electrolyte
is sulfuric acid with a concentration of generally 4.5-5M.
The concentration decreases during discharge and in-
creases during charge. Due to the concentration gradients
required for the diffusion process, the sulfuric acid con-
centration in the interior of the electrode may deviate (!)
significantly from that in the free electrolyte. The acid
concentration affects not only the electrode kinetics, but
also the m a x i m u m utilization of the active material of the
electrode (7-9).
The complexity of sulfuric acid as the electrolyte in the
porous lead dioxide electrode is due to not only its two
dissociation steps and the low solubility of lead sulfate, / /

but also to the fact that sulfate ions are specifically ad- t

sorbed on the lead dioxide surface (10, 11). r

The goal of the experimental studies reported here was

\
to investigate the effect of sulfuric acid concentration on
the recharge kinetics and as a specific result obtain a / /
practically useful rate expression, Which relates current
density to charge transfer overvoltage and includes pa-
rameters taking into account the effects of changes in z /

composition of both the solid electrode and the electro-

lyte. A kinetic expression of this form is of great use in / /

mathematical models for analyzing and predicting the \ I 9 " j|

/

electrochemical behavior of the porous lead dioxide

electrode.
x~- 2
Experimental Procedure \ / r

The electrochemical measurements were performed in Fig. 1. Experimental cell with the porous PbO2 electrode as a flow-
two different cells, one of which has been described through electrode. 1: Electrolyte flow from a peristaltic pump. 2: Porous
earlier (1). The other cell was designed with the porous lead dioxide electrode. 3: Lead electrode. 4: Luggin capillaries. 5: Grad-
lead dioxide electrode as a flow-through electrode (Fig. uated tube.