UCRL-JC-l20308

PREPRINT

A Refuelable ZindAir Battery for

Fleet Electric Vehicle Propulsion

J. F. Cooper
D, Fleming
D. Hargrove
R Koopman
K. Peterman

This paper was prepared for submittal to the

1995 SAE Future Transportation Technology
Conference and Exposition
Costa Mesa,CA
August 7-10,1995

April 20,1995
 DISCWMER
Thisdoclrment wasprepared as an accaunt of work sponsored by an agency of
the United states Government. Neither the United states Government nor the
University of W o m i a nor any of their employees, makes any warranty, express
or implied, or assumes any legal liability or responsibility for the accuracy,
completeness, or uduhess of any infonnatian, apparatus, product, or process
disdosed, or represents that its use would not infringe privately owned rights.
Referenae herein to any spedfic commercial product,process, or service by trade
name, tradema& manufacturer, or otherwise, does not necessarily constitute or
imply its endorsement, recommendation, or favoring by the United States
Government or the UNversity of W o m i a . The views and opinions of authors
expressed herein do not neessady state or reflect those of the United States
Government or the University of California, and shall not be used for advertising
orproductendorsementpurposes.
 DISCLAIMER

Portions of this document may be illegible

in electronic image products. Images are
produced from the best available original
document.
 Paper Submitted to the 1995 SAE Future Transportation Technology Conference and Exposition
Costa Mesa CA; August 7-10.1995
Manuscript 95-FIT-32 Session FTT-IS

A Refuelable Zinc/Air Battery for Fleet Electric Vehicle Propulsion

John F. Cooper, Dennis Fleming, Douglas Hargrove, Ronald Koopman and Keith Peterman
Lawrence Livermure National Laboratory, L369 LivermoIe CA 94550
ABSTRACT
We report the development and on-vehicle testing of an engineering prototype zinc/air
battery. The battery is refueled by periodic exchange of spent electrolyte for zinc particles
entrained in fresh electrolyte. The technology is intended to provide a capability for nearly
continuous vehicle operation, using the fleet's home base for 10 minute refuelings and zinc
recycling instead of commercial inhstructure. In the battery, the zinc fuel particles are stored in
hoppers, from which they are gravity fed into individual cells and completely consumed during
discharge. A six-celled (7V)engineering prototype battery was combined with a 6 V leadacid
battery to form a parallel hybrid unit, which was tested in series with the 216 V battery of an
electric shuttle bus over a 75 mile circuit. The battery has an energy density of 140 Wh/kg and a
mass density of 1.5 Isg/L. Cost, energy efficiency, and alternative hybrid configurations are
discussed.

INTRODUCTION
The use of electric storage batteries to power commercial and industrial fleet vehicles has
a number of attractive features. The fleet's home base facilities and personnel constitute a pre-
existing infrastructure for vehicle maintenance, battery recharging and technical support.
Missions are ofien repetitive (or at least predictable), and daily usage can be high enough to
allow rapid write-offof the generally high initial investment in electric propulsion. Examples of
fleet electric vehicles include electric shuttle and transit busses, commercial delivery and
passenger vans, enclosure-operated vehicles such as fork-lifs trucks and mall shuttles, and
specialty applications ranging from aircraft service vehicles to silent military equipment. To
undertake long-duration missions (e.g., lasting a substantial fraction of the day), the batteries
might be periodically exchanged, "opportunity'' charged, or extended with heat engines.
However, these approaches increase cost and servicing complexity while restricting mission
flexibility.
Alternatively, the batteries may be mechanically recharged by a rapid exchange of fresh
reactants for reaction products--while the products are externally recycled In recent years, a
number of efforts have been made to develop high energy density batteries which are refuelable
or reconstructable [l-51,as a means of combining long range with a capability for nearly
continuous use through periodic refbeling.
The objectives of the battery development work reported here are: (1) develop an
engineering prototype based upon our self-feeding cell invention [11; (2)determine energy and
power characteristics; (3) demonstrate rapid refueling; and (4) show that the self-feeding cell
operates under repmentative road vibrations and accelerations.

-1-
TECHNICAL APPROACH
Svstem, Our a p p a c h is shown schematically in Figure 1. The battery consumes zinc
metal and atmospheric oxygen, and produces a reaction product consisting of a liquid suspension
of zinc oxides and potassium zincate in an alkaline electrolyte:
Zn + 1/2 02 (air)+ KOH =KznooHaq (1)
Refueling is accomplished by replacing the spent electrolyte with fresh electrolyte and zinc
pellets. The spent electrolyte is reprocessed in equipment at the fleet's home base to recover
zinc-depleted elecmlyte and zinc fmes, which are pressed into uniform 1mm pellets. The cost
of the recovery unit is small compared with the cost of the battery because it operates
continuously and hence at a lower power level than the battery. In addition, it lacks the
expensive air electrode of the zinc/air battery, which accounts for half the estimated battery cost
of $50/kW.

electrolwis
cells

spent electrolyte
I
Pellet
press
)r)-
wztzgw q
-
e s h electrolyte
and entrained pellets
storage tank
for pellets and discharge
electrolyte cellsand
hoppers erecrroiyre
storape
ZINCIATR BATTERY
Figure 1. The battery consists of a stack of bipolar cells which are gravity fed with zinc particles
from overlying hoppers, and an electrolyte storage tank. Refueling exchanges spent electrolyte
for fresh electrolyte with entrained zinc particles. Electrolysis equipment recycles zinc and fresh
electrolyte from the spent electrolyte.
Self-feeding cell. During discharge, zinc pellets fall continuously from the overlying
hopper into a n m w (4 mm) opening at the top of each cell (Figure 2). Because the cell gap is
only a few times larger than the particles, the particles do not close pack--even when subjected to
strong vibrations. Rather, particle bridging and voids develop, which gives the bed an artificially
low packing density of about 40%. Electrolyte flows upward through the cell and hopper to
remove heat and reaction products. The expanded bed offers little resistance to electrolyte flow.
The wedge-shape of the cell allows dissolving zinc particles to fall into an increasingly narrow ,
gap to maintain the gap/particle size ratio (c 5 ) needed for bridging. The losses from air and
electrolyte pumping and shunt currents sum to about 1%of gross power [7].
Refueling. The 1 mm particles are too heavy to be easily levitated into a uniform slurry 1
by flowing electrolyte in the battery. Rather, the particles are pushed along the base of a
horizontal fill tube common to a multicell stack by the impulse of flowing electrolyte. The

-2-
particles fall through slots (6 x 16 mm) leading into the hoppers. Because the anodes may be
shorted through particles in the fill tube, it is necessary to shut off the air supply to the cells
during filling and to continue electrolyte flow after filling to remove excess zinc and electrolyte
from the fill tubes.
7inc recovery and recV& Using a combined electrochemical and mechanical process,
zinc pellets were produced from recycled battery reaction product, Zinc was recovered from
spent battery reaction products by deposition onto a nickel substrate under conditions of current
density and flow resulting in a porous deposit called "mossy zinc." This deposit is approximately
15-20 % metal. It does not adhere to the electrode but detaches under its own weight to collect at
the bottom of the electrolysis cell. The deposit is pressed into 1 mm x 1 mm cylindrical pellets of
60%density using a simple pin and die press. Details of the process will be reported in a later
publication. Experimental work reported here uses commercial 0.5- or 1 mm cut Wire as fuel.

with entrained
particles

b- Anode chambers

Electrolyte in
Figure 2. Zinc pellets or particles are stored in hoppers above individual cells, and are gravity fed
into the cells. Packing and clogging in the bed are prevented by the formation of bridges and
voids, allowing electrolyte to circulate freely and with low hydraulic resistance.
TECHNICAL,RESULTS
Battery components. Our electrochemical cell is comprised of two components: (1) A
plastic frame contains the hopper, channels for air and electrolyte circulation, seals and a porous
separator 191. (2) A bipolar plate supports both anode and cathode 181 and is provided with feed-
throughs to electrically connect the adjacent cells. Up to 12 cells can be combined with the
electrolyte storage tank (Figure 3). The number of cells determines power, the size of the storage
tank determines energy content.
The plastic fmme is shown in Figure 4. The electrode area (250 cm2) and hopper volume
(150 cm3) are chosen to yield a peak power of about 125 W and a nominal energy of up to 525
Wh (at 1.2 kW/m2) -- a combination anticipated for the shuttle bus application requiring a 5-10
hour rate and modest power. (Electrode area, hopper volume and storage tank volume can be
varied independently to match vehicle power and energy requkments.) The long electrolyte

-3-
channels between cells reduce shunt losses for a 12 cell stack to below 0.5% of gross power.
Electrolyte flows upwards at a phase v e l d of 1cm/s, while air flows downwads at 4 times
the rate requiredby stoichiometry for 6 Wm!f-the maximum cell current density.
The bipolar plate supports a metal basket attached to the anode suxface and a ribbed frame
on the cathode d a c e to hold the air electrode. The plate was fabricated from electronic circuit
board (62 mil;both sides copperclad; glass-filledepoxy resin).
Refuelinq. Refueling of a 12-cell stack of 250 cm* celis was developed before discharge
tests. The particles wem fed into the assembled stack (Figure 3) through two internal 2.5 cm
diameter tubes, driven by combine flow of 4.5 Umin with a total presswe drop of 2 kPa across
the stack. As shown in figure 5, the cells filled sequentially at a rate of 4 cells per minute. Three
c

parallel sets of thm 12cell stacks c o n n d in hydraulic series can thus be =fueled in under 10
minutes and will contain 56.7 kwh of deliverable energy [a.

Figure 3. The twelve-cell stack and electrolytetank is shown here, assembled for refueling tests.
Cooling fms (Aavid Engineering, Inc. NH) were added to the storage tank prior to discharge
tests on 3- and 6-celled units.

-4-
 -s-

t - 000

- os0
- 001

- 0083
-
-
3

jI-
OS&

ZT T4- OOP

.............
.. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. ..
....................
and wedge angle of this cell were reproduced in the 250-cm2cell used in laboratory refueling
and in vehicle tests. After refueling tests on the twelvecell stack, this stack was disassembled
for tests in multiples of 1,3, and 6 cells together with the same storage tank (sized for 12 cells)
and auxiliaries.
The three cell stack was discharged for five hours then held on standby for 12 days,
while the cells remained iilled with cold, supersaturated electrolyte and the air flow was shut off.
No deterioration in the polarization curves was noted following mctivation. This indicates that
the strongly reducing potential of the oxygen-depleted zinc-filled cells does not cause
deteriorationof the separatoror the air elecmde catalyst.
The six cell stack (the largest discharged) was tested on board the electric bus, with the
storage tank half fded with electrolyte. The power and voltage of the stack (averaged per cell)
are plotted against current density in Figure 7. Superimposedare the voltage and power curves
for the 80 cm2 cells (Figure 6). The difference in voltage and power density is attributed to the
use of conductive silver-loaded epoxy resin cement to connect air electrodes to their substrates
rather than the solder connection used in the smaller cells. The 10 OC lower temperature of
operation also lowered the power density because of the lower elecmlyte conductivity.

18.
5
1.6

1A
4
N >
z
E 12
9
-e
Y
3 1 =

n 2
OS 8
0.6

0.4
1
02

0 0
0 1 2 3 4 5 6 7 8 9

Current density, W m 2

Figure 6. Power densities and voltages are given for discharge of the 80 cm2 cell at a
temperature of 65 OC. Curves are for three levels of electrolyte ampere-hour loading: (a) 36
Amter; (b) 54 Ah/liteq and (c) 126 Ah/liter. The cell configuration and wedge angle were
replicated in the 250 cm2 cells.
Vehicle Test Configuration and Results. On-vehicle tests were conducted in order to
demonstrate the operation of the self-feeding cells under typical driving vibrations and
accelerations, and to demonstrate a multicell level of system integration. The Santa Barbara
Metropolitan Transit District (SBMTD) provided us with a 22 foot electric shuttle bus (Figure 9) c
for purposes of conducting this test. One 6-cell (7 V) battery was cabled in electrical parallel
with a leaWacid battery consisting of three cells (6V;Chloride Motive Power, Ltd., Model

-6-
S32Y11) to form the basic unit of a parallel-configureddual-battery hybrid power source. This
unit was placed in series with 108 l d a c i d cells of the same model (216 V) which normally
provide all power. Both batteries in the hybrid unit were protected against reverse polarization
with 300 A diodes. Current and voltage were monitored for each zinc/& cell, for the zinc/&
stack, and for the drive train. A second on-board computer allowed measurement of power and
total energy use with a 4-second time constant for the battery train. The bus was driven on a
paved street around a 2 mile loop for 5 hours (including a 1hour rest) until the lead/acid battery
train was 80% discharged. Test conditions and salient results are reported in Table 1.

1.4 3 , 1.4

1.2

r:
0
0.4 J
v)

0.2

0 0.2 0.4 0.6 0.8 1 1.2

Current density, W m 2
Figure 7. Cell voltage and power density (averaged per cell) are plotted against current density
for the 6 cell stack at 55fl "C. The power density and voltage curves from Figure 6 are
superimposed. The differencereflects the lower temperature of operation and the use of a poorly
conductive silver epoxy cement to bond the air electrode to its supparting conductor.
Table 2 tabulates the weights of the components and reactants used in the test. Energy
yield is determined by the amount of consumable zinc with the electrolyte as the limiting reactant.
The weights of the test cells are reported. One-half of the weight of the 12-celltank was counted
for this h e l l test. Air pumps, scrubbers and motors will serve the entire vehicle battery array;
accordingly, the value reported is based on available equipment pro-rated from 55 kwh scale
down to the &cell battery. A specific energy of 127 WMrg is indicated for the test equipment.
The listing for production units (141 W g ) is based on the same chemistry and configuration,
but with 10%additional zinc and electrolyte. The production battery figure also assumes use of
lightened cell frames constructed of filled polymers and transfer plates of 40 mil board with
removal of unused cladding. We can fill the production cell hoppers to capacity and increase
electrolyte and tank size accordingly, which leads to an energy density of 150 Wh/kg. These
differences will be seen to be insignificant for a bus battery, as this spread (127-150 Wh/lcg)
conresponds to about a 1 % spread in resultant vehicle weights with the same range.

-7-
Table 1. Test Conditions and Results for on-vehicle test of zinc/air in dual-

25 I v v

f 160
3
$20
Q)
C
140 -
L.

3
Q)

= 120 0
Q),

C 15 Q Energy from Zn
0
L
1 0 0 cn
3
0 0 Total energy
80 Q)
10 C
% 60
(XI
c -Zn Power
C
F (XI

-cn
c
Q)
40
5 5 C

c 20
N
0 0
0 100 200 300 400
Time, s

Figure 8. Instantaneous power, integrated power from Zinc/air and time-averaged total energy
from parallel unit is plotted against drive time for a 330 s, 3.22 lan (2 mile) cycle. (Cycle No.
32). This part of the on-vehicle test was conducted at lower-than-optimum temperam of 42 OC.
While the peak rated power of the cells (750 W) and weight (26kg) might suggest a low ,
"specific power" of only 30 Wkg, such a conclusion would be misleading. The weight and
volume of this kind of battery is dominated by the weight and volume of reactants and not by the
weight and volume of the cells. Tripling the electrode area for the same tank and amount of
reactants would triple the power to 2.25 kW.However this increase would add < 3 kg (or 10%)
to weight to the battery -- an amount representing the additional electrodes and support (lower

-8-
part of the cell, Figuxe 4). Unlike secondary and reconstructable zinc/air batteries, the energy
and power of this battery 8fle indepndent parameten andcan be adjusted tomeet the needs of the
vehicle. Specific power has no unique meaning here, and cannot be used as a scale factor.
Rather, weight and volume can be estimated from the separate coefficients for peak
power and nominal energy. The coefficients 2.8 kg/kW-peak and 6.5 kg/kWh-nominal can be
derived from the "production" column in Table 2. The comsponding volume coefficients are 3
ukw-peak and 4 WWh-nominal. Thus the weight of a 27 kW-peak battery delivering 46.8
kwh of energy at 80% DOD is 456 kg and occupies about 315 L (11 ft3). Because of the zinc
reserye in each cell, the battery can be discharged to 100%DOD without ham.

Table 2. Battery component weights (6-cell stack and down-sized tank); Peak
rated power, 750 W.
Component Test equipment Production units Projected units
kg kg kg

aBased'mconsumable zinc fuek excludes undischarged zinc mewe of 8-1096 in cells.

bAuxitiarieS (air pumps, motors and scrubbers) sewe entire battery arrays; quoted values are pro-rated to &ells.

Table 3. Comparison of lead/acid and zindair (specific energy, 141 Wh/kg) in

shuttle bus on SBMTD urban dutv " cvcle.
"
Battery Test Battery Battery Required Efficiency, Energy Range
weight weigh@ volume, energyb % use km (mi)
ksr(Ib\ L(ft3) kwh kwh

%%c/airweight is weight to deliver required energy at 80% exhaustion of consumable reactants from a battery of
specific energy, 141 Whikg.
bVehicleenergy-usescale factor is 256 kJ/tonne-km (104 Wh/ton-mile) for the urban driving cycle.

In Table 3, we have calculated the weight and volume of the zinc/& battery that might be
used on the SBMTD shuttle bus in its current duty cycle. Keeping either delivered energy
constant (at 80% DOD)or the range constant, we have calculated reduced vehicle mass and total
AC energy consumption using a vehicular specific energy consumption rate of 104Wton-mile
(SBMTD data, from the Santa Barbara Waterfront Route). Keeping range constant, the vehicle
gross weight drops by 1600 kg--equivalent to 20 passengers. Despite the lower energy
efficiency of the zinc/& battery, total energy use is 20% lower than that of the ledacid-powered
bus because of the lower vehicle mass, all else being equal. The scaling of the zinc/air battery is

-9-
based on the mass and volume coefficients of power and energy, with energy computed at 80%
exhaustion of available energy. Tire and brake life and handling are also improved by weight
reduction. The number of cells is fixed at 216 (1.23 V nominal) and the peak rated power is 27
kW. The power can be altered by altering the cell size or the number of cells.
%

ECONOMY, ENERGY USE AND HYBRID CONFIGURATIONS

Any detailed cost estimate is beyond the scope of this work. However, the 20% c

reduction in weight and energy use for fixed range and mission (passenger-miles) indicates a
significant reduction of energy cost and tire and brake ware. The cost of zinc/air battery is
dominated (>50%) by the cost of the air electrode and its supporting structure, which we
estimate to be $SO/kW-peak + $2&Wh-capacity &om a large-scale cathode production cost of
$120/m2 (using continuous production techniques and cobalt macrocyclic catalysts).
The zinc battery described in Table 3 has sufficient power for the bus, giving the vehicle
a powedweight ratio of 6.8 W/kg-gross (vs. the 5.2 W/kg-gross available with leadacid).
Nonetheless, consideration of a power leveled, parallel hybrid is encouraged for three reasons:
(1) the powerheight ratio can be i n d greatly, giving the vehicle greater route flexibility and
highway-safe acceleration; (2) regenerative braking becomes possible, increasing efficiency by
ca. -10 96;and (3) the life of the air electrode is maximized when current peaks are avoided. The
electrodes used in this test operated without failure or serious deterioration for over 12,000 hours
in a 7.5 M KOH + 1M aluminate electrolyte at 60 "C and 2 Wm2[8].
CONCLUSIONS
The essential objectives of this on-vehicle test were met. The self-feeding cells
functioned correctly over the 5 hour test, without becoming clogged or starved. The refueling
tests (on the 12 cell stack) showed fding rates of 4 cells/minute, which allows any number of
parallel sets of 36 cells (19 kwh per set) to be filled in parallel in 10 minutes. The battery
functioned in a hybrid with the 6 V leadacid pack, which was subjected to a rigid (high
impedance) c m n t source and a pulse-width modulated controller. The indicated specific energy
(based on the electrolyte reactant limited to 250 Am)was 141 Wh/kg, for laboratory test cells
reduced to commercial production techniques. A disappointing feature of this work was the
lower power obtained from the stack--the consequence of the use of conductive silver-epoxy
rather than low-melting solder, cold bonding, or leaf metal bonding techniques used elsewhere
for connecting air electrode screen substrates to conductors. Subsequent tests of the battery will
correct this deficiency.
Acknowkkmm
Work pelfonned trnder the auspices of the U.S.Department of Energy by the Luwrence Livermore National
Laboratory under contract number W-740.5-ENG-48. We gratefully acknowledge the support of the Lawrence
Livennore National Laboratory Institutional Research and Development Program in the initial phases of this
work. We gratefully acknowledge the support for this project by the Department of Transportation (Federal
Transit Administration) through its contractor, CALSTART, INC. (Burbank,CA). Finally we express our
g r a t i t h to the Santa Barbara Metropolitan Transit District for allowing us to use the Electric Shuttle Bus in this
remarch.

REFERENCES
1. John F. Cooper, Dennis Fleming, Ronald Koopman, Douglas Hargrove, Arturo Maimoni
and Keith Peterman, "Testing of a refuelable zinc/air bus battery," Roc. of the
Electrochemical Society, May 21,1995; LLNL Report UCRL-JC-119686 December 1994.

- 10-
2. Jonathan R. Goldstein and Binyamin Koretz, "Tests of a full-sized mechanically
nxhargeable zinc-air battery in an electric vehicle," Proc. Intersociety Energy Conversion
Engineering Conference, Paper No. 93410, Atlanta Georgia, August 1993.
3. Evans, J. W.and Savaskan, "A zinc-air cell employing a packed bed anode," J. Appl.
Electrochem., 21,105-110 (1991)
4. Sierra Alcazar H. B., Nguyen P. D. and Pinoli, A. A., Technology Base Research on
ZinclAir Battery Systems, LBL-Report No. 23960, September 1987.
5. Appleby, J. and M.Jaquier The C. G.E. Circulating 25ncIl-k Battery, J. Power Sources
VOI. 1, pp. 17-34 (1976fl7)
6. John F, Cooper, Dennis Fleming, Larry Keene, Arturo Maimoni, Keith Petennan and
Ronald Koopman, "preliminary Results in the Refueling of a MultimIl Module," (Paper No.
-
AIAA 94- 3841; hoc. 29th Intersociety Energy Conversion Engineering Conference;
Monterey, CA August 7-12,199+ LLNL R F NO.UCRL 3C 117946, August 1994.
7. John F. Cooper, Larry Keene, Jon Noring, Arturo Maimoni and Keith Peterman,
"Regenerative Zinc/Air and Zinc/Femcyani& Batteries for Stationary Power Applications,"
Roc. 36 Power Sources Conference, Cherry Hill, NJ, June 6-9 1994; LLNL Report
UQur117248-Rev. 1, M a y 1994.
8. The separator was a @orated polymer membrane manufactured by Pall RAI Manufacturing
Company, Inc.; Hauppauge, NY.
9. Air electrodes were manufactured by Eltech Research Company, Painesville OH; Model AE-
100 catalyzed with cobalt etramethoxy-phenyl-porphyrin (Co TMPP). Life data from
private communication from Dr.Eric Rudd, Eltech Research Company, May 1993.

Figure 9. Shown is the electric shuttle bus used in testing the 6-cell stack (loanedto us by the
Santa Barbara Metropolitan Transit District, a CALSTART,Inc. affiliate).

-11-