Journal of Power Sources, 47 (1994) 119-127 119

Development of an aluminiumhea water battery for sub-

sea applications

PK. Shen and A.C.C. Tseung

Chemical Energy Research Centre, Department of Chemist?y and Biological Chemisty, University
of Essex, Wivenhoe Park, Colchester CO4 3SQ (UK)

C. Kuo
Marine Technology Centre, lJniversi@ of Strathclyde, Glasgow G4 OLZ (UK)

(Received April 7, 1993; in revised form July 3, 1993; accepted July 5, 1993)

Abstract

There are two modes of operation for an aluminiumbater battery: use of hydrogen-evolving
or oxygen-reduction cathodes. In the first case, the current density is -10 mA cm-’ at
0.5 V. This leads to significant increase of pH and deposition of Ca/Mg(OH)z at the
cathode, leading to sharp decrease in battery performance. On the other hand, in the
oxygen-reduction mode, the current density is -0.1 mA cm-* at 1.4 V, because of the
low concentration of O2 in sea-water. Long-term tests of an Al/water battery using Teflon-
bonded Co,O,/C oxygen reduction cathodes in Brightlingsea Harbour, Essex, over a test
period of 70 days have shown that the performance of this battery is stable and that
it is possible to increase its performance by means of multi-cathodes on either side of
the anode. A conceptual design shows that such a battery has an energy density of
1008 Wh kg-’ over an operating period of one year, significantly higher than all the
conventional primary batteries considered for use in sub-sea environments.

Introduction

The growing interest in sub-sea production and exploration of oil and gas wells
has created the need to develop long-life and high energy density batteries. Primary
lithium batteries have the highest energy density (up to 500 Wh kg-‘) [l], but if these
batteries are used for sub-sea applications, there is a need to put them inside pressure
vessels, reducing their energy density and increasing the operating lost. On the other
hand, primary batteries based on the use of metal alloy anodes and oxygen-reduction
or hydrogen-evolution cathode should be able to satisfy the above requirements [2-53.
Sea-water is used as the electrolyte and dissolved oxygen or water is used as the
oxidant. Therefore, the only consumable item is the metal anode since dissolved oxygen
and water are supplied by sea-water. Moreover, since sea-water is used as the electrolyte,
the battery does not require vessel containment.

Some theoretical and practical considerations

Electrode reaction processes

Metal/water batteries can be operated under two modes:
(i) Mode 1 (using a hydrogen-evolving cathode)

0378-7753/94/$07.00 0 1994 Elsevier Sequoia. All rights reserved

120

anodic reaction: Al + 30H- - Al(OH)3 + 3e- (I)

cathodic reaction: 2H,O + 2e- - Hz + 20H- (2)
The operating voltage of such batteries is about 0.5 V at a current density of
-5 mA cmp2.
(ii) Mode 2 (using an oxygen-reduction cathode)
anodic reaction: Al + 30H- --+ Al(OH), + 3e- (I)’
cathodic reaction: O2 + 2H20 + 4e- - 40H- (3)
The operating voltage of such batteries is about 1.4 V at a current density of
-0.1 mA cme2. The low current from such batteries is due to the limited solubility
of oxygen in sea-water.
Natural sea-water contains dissolved salts, dissolved gases, organic substances,
colloidal matter, and particulates, both living and detrital. Most of the compounds of
sea-water are unreactive electrochemically [6]. Easily reducible species at the cathode
are dissolved oxygen, water and hydrogen carbonate. The formation or precipitation
of insoluble salts of calcium and magnesium through alkalisation of the cathode surface
by reactions (2) or (3) is detrimental, which will greatly reduce the output voltage of
the battery.
Oxygen is an extremely electroactive constituent of sea-water. The theoretical
half-cell potential for the reduction of oxygen in pH 8.2 sea-water is 0.504 V versus
saturated calomel electrode (SCE), compared with -0.726 V for hydrogen evolution.
The corresponding high potential implies that the battery voltage will be significantly
higher than that of the batteries which rely on hydrogen evolution as the cathodic
reaction. Though both the oxygen reduction and hydrogen evolution produce OH-
which alkalises the cathode surface, the degree of alkalisation of the surface by oxygen
reduction is far less than in the case of hydrogen-evolution cathode operating at
10 mA cme2, since the limiting current for oxygen reduction is likely to be -0.5 mA
cmm2. This should reduce the rate of precipitate build-up on the surface and part of
it may be dissolved more readily since the movement of sea-water will moderate the
pH of the solution near the surface of the electrodes.

Aluminium alloy anodes

Aluminium as an anode is attractive because of its very negative potential and
electrochemical equivalence and availability. However, it is subject to a high overpotential
in use which is thought to be due to the formation of passive deposits on the surface
of the electrode [7J. Aluminium alloys with superactive discharge characteristics in
alkaline batteries have been developed and they exhibit a potential that is approximately
700 mV more negative than structural alloys [8]. Alcan International Laboratories of
Banbury, UK, supplied their newly developed aluminium alloy, ABSOV, for this study.
This alloy contains 0.63% Mg, 0.14% Sn, 0.04% Ga and 0.0018% Fe. Figure 1 shows
the steady-state .polarisation curve of this alloy in 3% NaCl solution at room temperature.
The open-circuit potential of this alloy is about - 1.45 Vversus SCE which is significantly
better than the normal Al alloys (- - 1.1 V for AlZ2).

Improved cathode catalysts

Early work on corrosion batteries for military and biomedical applications have
relied on the use of Pt black cathodes, since cost is of secondary importance. For
sub-sea applications, cost must be minimised. The traditional way of reducing the cost
 121

-1.6 -1.56 -1.5 -1.45 -1 A -1.35 -1.3

-I
-1.25
Potential, V vs SCE

Fig. 1. Steady-state polarisation curve of an ABSOV Al alloy in 3% NaCI solution at room

temperature.

of precious metal cathode is to use Pt/C catalysts. By dispersing ultrafine Pt particles

(- 100 A) on high surface area C supports, the utilisation of available Pt is significantly
improved. However, since the current density of an Al battery operating with hydrogen-
evolving cathode is at the most - 10 mA cm-*, even at a loading of 1 mg Pt cm-‘,
the cost is still too high. Previous work has shown that it is possible to reduce the
Pt content in the catalyst by means of W03 as the active support instead of C. The
evolution of hydrogen can be extended to the WO, surface via the formation of the
conducting HXW03. Very high activity could be achieved at very low Pt loadings in
acidic media [P, lo]. It should be noted that the Pt/W03 electrode used in previous
work was based on the impregnation of chloroplatinic acid on WOs, followed by freeze-
drying and vacuum decomposition at 300 “C for 1 h. This preparation method would
only disperse Pt particles on W03 particles and the spillover reaction is limited to
sites where there is Pt/WO, particle contact and the utilisation of Pt is not high.
Recently, we prepared Pt/W03 electrodes using coelectrodeposition from a solution
containing chloroplatinic acid and dissolved tungsten and showed that Pt is very well
dispersed in the WO, matrix [ll]. Tests of a 0.2 mg Pt/W03 cm-’ hydrogen-evolution
cathodes in 3% NaCl solution showed that the performance is comparable with a
Pt/C cathode containing 1 mg Pt cm -‘. However, on testing the Pt/WO, hydrogen-
evolution cathode at 10 mA cm-’ in artificial sea-water, the performance of the battery
deteriorated after 10 h due to the accumulation of hydroxide deposits on the surface
of the electrodes. The relatively high current density will lead to a rapid increase in
the pH of the solution near the electrodes, leading to the precipitation of hydroxide
sludge.
Much higher battery voltages can be achieved if dissolved oxygen is used as the
oxidant. Moreover, the changes in pH of the solution will be minimal because of the
low operating current density. In this case, inexpensive, non-noble metal catalysts for
oxygen reduction have to be used. In this work, we concentrated our studies on the
Teflon-bonded cO,O,/C cathodes, since previous work by Tseung and coworkers [12]
has shown such electrodes possess very good performance in alkaline solutions.
Figure 2 shows the performance of a Teflon-bonded C&0.,/C electrode operating
at 0.1 mA cm-’ in artificial sea-water and the performance is similar to that of a
Teflon-bonded Pt/C (Pt loading: 0.4 mg cm-‘) oxygen-reduction cathode.
 i= O.lmA/sq.cm

0 5 10 15 20 25 30 5
Time, hours

Fig. 2. The performance of a Teflon-bonded Co,O,/C and a Pt/C (Pt =0.4 mg cm-*) cathode
operating at-O.1 rnA cm-' in artificial sea-water.

2-

1.75-

OS-
i= O.lm&q.cm

0.25-

OO
-I 0.2 0.4 0.6 0.6 1 1.2 1.4 6
Distance, cm

Fig. 3. The effect of anode/cathode gap on the battery voltage in 3% NaCl solution at 14 “C.

Battery development
For long-term operations in sub-sea environments, Al/water batteries using oxygen-
reduction cathode should overcome the following problems:
(i) Prevention of marine growth and the formation of hydroxide or calcareous
layers at the electrodes.
(ii) Minimum increase in IR losses due to the dissolution of the anode.
The effect of anode/cathode gap on battery performance operating with oxygen-
reduction cathode (mode 2) is given in Fig. 3. There is no significant change in
performance by varying the anode/cathode gap from 0.4 to 1.2 cm. In addition, when
the design is based on an anode with a cathode on either side, the results suggest
that even 20 cm thick anode could be used without any worries about the increase
in IR losses with time. Since the resistivity of sea-water is typically 25 a cm [13], the
IR loss is only 25 mV for a 10 cm gap under the operating current density of
0.1 n~4 cm-‘.
 123

(iii) Minimising of the self-discharge of anode. All active anode materials will
react with water thus reducing faradaic efficiency. Therefore, it is important to reduce
the amount of anode surface exposed to the sea-water to a minimum.

Evaluation of long-term performance in sub-sea environment

Installation of the underwater test facility

Three batteries were installed side by side. The surface areas of both anode and
cathode for each battery are 1 cm ‘. A cube of ABSOV Al alloy was used as anode
material and the cathode was a Teflon-bonded Pt/C (Pt loading: 0.4 mg cm-‘) electrode.
The anode/cathode gap was fixed at 0.5 cm. The batteries were bolted to a Dexion
box and the electrical leads and resistive loads were housed in a junction box bolted
at the top of a mooring pole in Brightlingsea Harbour, Essex. The Dexion frame was
bolted to the bottom of the pole and the clearance between the bottom of the sea
bed and the frame was about 60 cm. This is to ensure that the batteries will still be
immersed in sea-water at the low tide. After installation, the first measurement showed
that the performance was almost the same as laboratory test in 3% NaCl solution.
However, there was significant decrease in performance with time. The whole test
assembly was lifted out of the water after 80 days and examined. The anode/cathode
gap was gummed up with mud and whitish deposit (hydroxides). There was, however,
no evidence of any biological or marine growth on either the anode or the cathode.
One possible explanation is that the aluminium anode surface is continuously changing
because of dissolution, and no growth is possible. In the case of the Teflon-bonded
cathode, since Teflon is non wetting, the likelihood of marine growth is greatly reduced.
In addition, the increase in pH at the cathode is not conducive to the growth of
marine organisms.

Installation of the second test rig in a floating pontoon

In order to avoid the clogging of the anode/cathode gap by mud, the second test
rig was installed in a floating pontoon further away from the shore. The test assembly
was assembled about 30 cm below the water surface and the assembly was about
120 cm from the bottom at low tide.
Figure 4 shows all three batteries maintained stable output - 1.4 V (0.1 rn4
cm-‘) over a test period of 21 days, significantly better than the results obtained in
the first trial.

Sea trials of the third test rig using Teflon-bonded Co304/C cathodes
The third test rig was commissioned and two sets of batteries were tested. These
batteries were made with Teflon-bonded Co,O,/C cathodes. The results shown in
Fig. 5 are obtained from a set of batteries with the anode/cathode gap of 0.5 cm.
Figure 6 shows the results of another set of batteries which had a gap of 2 cm between
the anode and the cathode. There was hardly any difference in output between the
two sets of batteries up to 50 days, again confirming that at lower current densities,
the effect of IR losses due to the distance between two electrodes are low enough
to be of insignificant consequence. Postmortem examination of the first set of batteries
after 100 days revealed that the electrodes were covered with the whitish brown
deposits, indicating the inclusion of mud particles with the hydroxide precipitates. The
high coverage of the electrode surface by the deposits results in the decrease of the
battery performance. On the other hand, though there was some precipitate on the
 0.5 -
-
AV
-
0.4 BV
-
cv
-M-
Ai
-
Bi
-
Ci

0 2 4 6 6 10 12 14 16 16 20
Time, days

Fig. 4. The changes in battery voltage and current with time. Test site: floating pontoon, Teflon-
bonded Pt/C cathode. AV, BV and CV are the voltages of batteries A, B and C, respectively,
and Ai, Bi and Ci are the currents.

0.5 -
AV
... . .
:%/ 0.4 BV
-In-
1.4
> cv
gb 1.2’ 0.3 $ -
9 *x. ..- z Ai
li. -*. __*..
2 1’ ._.._‘._. ..?q I
E
E Bi
r: 0.8*---- Q “.#G . 0.2 6
-*-
m”
I. --*
0.6.
..*
-_._.__ ‘.r< _A -Ci

0.4. 0.1
__)
0.2

04 7
0 20 40 60 80 IO 0”
Time, days

Fig. 5. The changes in battery voltage and current with time. Test site: floating pontoon, Teflon-
bonded Co304/C cathode, anode/cathode gap: 0.5 cm. AV, BV and CV are the voltages of
batteries A, B and C, respectively, and Ai, Bi and Ci are the currents.

surface of the electrodes in the second set of batteries after a test period of 70 days,
it was far less serious than the first set of batteries, suggesting that the greater
separation between the anode and the cathode enables the hydroxide precipitates to
flow away from the batteries more readily.

Discussion and conclusions

The reason for choosing special Al alloy from Alcan International Ltd. for most
of this work is because of their lower self-discharge rate, compared with the Mg alloys
1141. However, it may be better to consider the use of Mg alloys as anodes in the
light of the results of the sea trials. The fact that we are getting Al(OH),/mud deposits
 125

0.6-
4
0.4- 0.1

0.2.

-r
OO
-I 20 40 60 60 108
Time, days

Fig. 6. The changes in battery voltage and current with time. Test site: floating pontoon, Teflon-
bonded Co,04/C cathode, anode/cathode gap: 2 cm. AV, BV and CV are the voltages of batteries
A, B and C, respectively, and Ai, Bi and Ci are the currents.

on the electrodes in sea trials is not surprising since the solubility product, Ksp, of
Al( is 1.3 X lo- 33. However, the solubility product of Mg(OH), is only 1.8 X lO_“.
In other words, the anodic reaction product of Mg hs significantly more soluble that
that of Al(OH)3 and will therefore be less likely to form precipitates on the electrode
surface.
Teflon-bonded Co,O,/C oxygen-reduction cathodes have proved to be almost as
active as Pt/C catalysts (Fig. 2) at low current densities. Moreover, test up to 120
days in sea trials have shown that they are not poisoned and that there is no growth
of marine life on the surface. This could be due to two reasons: (i) the presence of
polytetrafluoroethylene (PTFE) on the electrode surface, making it difficult for the
marine life to attach onto the surface, and (ii) increase in pH of the solution at the
surface of the electrode due to oxygen reduction. High pH solution is not a suitable
media for the growth of marine life.
The energy density for such a corrosion battery is higher than that of best primary
lithium battery. One method of improving the current output from such batteries is
using many parallel cathodes on each side of the anode. It is known that the current
output is limited by the oxygen reduction on the cathode. However, the current can
be increased by increasing the cathode area. Since the cathodes are porous, there is
adequate electrolyte pathway between each cathode layer. The use of such multi-
cathode battery will significantly increase the current output but without causing
significant changes in the pH at the cathode surface, thus avoiding the precipitation
of Ca/Mg(OH)z. Moreover, the higher power output from each battery will greatly
decrease the parasitic loss from the self-discharge of the anode, since the total exposed
area of the anode remains the same. Figure 7 shows the performance of a 1 cm2 Al
alloy anode/5cathode battery. The battery output was 0.5 mA cm-* for anode and
the performance is the same as the one with the current density at 0.1 mA cm-*.
In the conceptual design of a 25 W multi-cathode Al/water battery, it is assumed
that the operating current density is 0.2 mA cm-’ for each cathode and the operating
voltage is 1.3 V. By using - 1 mz Al anode and 5 X 1 mz cathode on either side of
the anode, the amount of Al alloy used in 1 year is:
W = (ItN)IF (4)
126

2-

new solution
1.6-
t
IJHI*( “)*_ q
>
& 1.2
d
?
z6 0.6 _
e:

0.4.

01, IO 20 30 40 60 60 70 0
Time, hours

Fig. 7. Voltage vs. time curve of an Al/5 cathode battery in 3% NaCl solution (0.5 mA cm-‘).

where I is operating current, f operating time, N the electrochemical equivalent of

Al, and the F is Faraday constant. The calculated weight of Al, including 30% self-
discharge, is -,74 kg. The weight of 10 cathodes is estimated to be 50 kg and the
weight of current collectors, reinforcing structures, etc., is estimated to be 50 kg. The
total dry weight of battery is 174 kg. This results in an energy density of 1260 Wh
kg-‘. It is important to note that the energy density of best primary lithium battery
is 500 Wh kg-’ (11. On the other hand, the battery voltage can be readily increased
to 15 V by a d.c. to d.c. converter [15]. This device can start at 0.7 V and the output
limit with -80% efficiency. The use of such a converter will only reduce the energy
density of the Al/multi-cathode battery to 1008 Wh kg-‘.
Further sea trials on offshore facilities over longer periods are necessary before
these batteries would be commercialised.

Acknowledgements

Financial support provided by the Offshore Supplies Office of the DTI and the
Petroleum Science and Technology Institute is gratefully acknowledged. The authors
would also like to thank the Harbour Master, Brightlingsea Harbour for providing the
facilities for the sea trials.

References

D. Linden, in D. Linden (ed.), Handbook of Batteries and Fuel CelLF, McGraw-Hill, New
York, 1984, p. 43.
8. Hasvold, in T. Keily and B.W. Baxter (eds.), Power Sources 13, International Power
Sources Symposium Committee, Leatherhead, UK, 1991, p. 307.
B.M.L. Rao, J. Zakrzewski and R.P. Hamlen, 7’he Electrochemical Society Spring Meet., Los
Angeles, CA, USA, May 7-12, 1989, Abstr. No. 10.
E. Budevski, I. Iliev, A. Kaisheva, A. Despic and K. Krsmanovic, J. Appl. Electrochem., 19
(1989) 914.
G. Scamans, J. Stannard, N. Fitzpatrick, B. Rao and R. Hamlen, Froc. Conj Defence
Oceanography, Brighton, UK, I990.
 127

6 M. Whitefield and D.R. Turner, in M. Whitefield and D. Jagner (eds.), Marine Electrochemistry,
Wiley-Interscience, 1981, p. 40.
7 P.G. Anderson and O.F. Devereux, J. Electrochem. Sot., 122 (1975) 267.
8 G.M. Scamans, J.A. Hunter and N.J.H. Holroyd, Proc. 8th Int. Light Metals Congr., Vienna,
Austria, 1987, pp. 699-709.
9 B.S. Hobbs and A.C.C. Tseung, Nature, 222 (1969) 556.
10 S.A. Abbarao, A.C.C. Tseung and D.B. Hibbert, J. Electrochem. Sot., 127 (1980) 1106.
11 A.C.C. Tseung, P.K. Shen and J. Syed-Bokhari, UK Patent Applic. No. 9104 377.8 (Mar. 1,
1991); PTC Patent Applic. No. PCTIGB 92100364 (Feb. 28, 1992).
12 S.P. Jiang, Z.G. Lin and A.C.C. Tseung, 1. Electmchem. Sot., 137 (1991) 764.
13 DegrCmout S.A., Water Treatment Handbook, Lavoisier Publishing, Paris, 1991, p. 490.
14 A.C.C. Tseung, P.K. Shen and C. Kuo, Development of a durable AI/H20 battery for sub-
sea applications, Final Rep., 1992.
15 E. Soumalainen, US Patent No. 4 970451 (1990).