Journal of Power Sources 112 (2002) 162–173

Design and analysis of aluminum/air battery system

for electric vehicles
Shaohua Yang, Harold Knickle*
Department of Chemical Engineering, University of Rhode Island, Kingston, RI 02881, USA
Received 6 April 2002; accepted 14 June 2002

Abstract

Aluminum (Al)/air batteries have the potential to be used to produce power to operate cars and other vehicles. These batteries might be
important on a long-term interim basis as the world passes through the transition from gasoline cars to hydrogen fuel cell cars. The Al/air
battery system can generate enough energy and power for driving ranges and acceleration similar to gasoline powered cars.
From our design analysis, it can be seen that the cost of aluminum as an anode can be as low as US$ 1.1/kg as long as the reaction product is
recycled. The total fuel efficiency during the cycle process in Al/air electric vehicles (EVs) can be 15% (present stage) or 20% (projected)
comparable to that of internal combustion engine vehicles (ICEs) (13%). The design battery energy density is 1300 Wh/kg (present) or
2000 Wh/kg (projected). The cost of battery system chosen to evaluate is US$ 30/kW (present) or US$ 29/kW (projected).
Al/air EVs life-cycle analysis was conducted and compared to lead/acid and nickel metal hydride (NiMH) EVs. Only the Al/air EVs can be
projected to have a travel range comparable to ICEs. From this analysis, Al/air EVs are the most promising candidates compared to ICEs in
terms of travel range, purchase price, fuel cost, and life-cycle cost.
# 2002 Elsevier Science B.V. All rights reserved.

Keywords: Aluminum/air batteries; Life-cycle analysis; Electric vehicles; Cost analysis

1. Introduction increases the material cost. To demonstrate the ability of this

system for vehicle applications, the range and acceleration
The Al/air battery system has a high theoretical voltage capability similar to internal combustion engine vehicles
(2.7 V), high theoretical energy density (8.1 kWh/kg-Al), (ICEs) on an economically sound basis must be provided.
low cost, an environmentally benign and recyclable product. Thus, the analysis and estimation of the performance, cost
It turns out to be a promising system for EVs. The Al/air and efficiency of the fuel (aluminum anode), the batteries,
yields energy densities that exceed many other couples and the vehicles powered by this system are all important
(Table 1) [1,2]. parts of this study.
Al/air batteries have the potential to be used to produce
power to operate cars and other vehicles. These batteries
might be important on an interim basis as the world passes 2. Fuel cost and efficiency
through the transition from gasoline cars to hydrogen fuel
cells cars. The Al/air battery system could generate enough 2.1. Fuel cost
energy and power for driving ranges and acceleration similar
to gasoline powered cars. The major problem in using this The presence of certain impurities in aluminum can
system is the low coulombic efficiency of aluminum in markedly affect the electrochemical behavior. For example,
strong alkaline media resulting from its high corrosion rate the corrosion rate is particularly sensitive to the concentra-
(hydrogen evolution reaction) and high level of polarization tion of iron [5]. Typically the anode uses aluminum of high
during discharge. Use of high grade (99.99 or 99.999%) purity 99.995 and 99.999% with small amount of other
aluminum doped with other minor elements such as Ga, In, elements, usually in combinations as ternary or quaternary
Sn, Mg, Pb, Hg, Mn, Tl, etc. [3–7] can reduce corrosion but alloys to achieve activation and inhibition of corrosion. The
production of aluminum, the cost of aluminum required by
*
Corresponding author. Tel.: þ1-401-874-5984; fax: þ1-401-782-1066. the Al/air battery system are reviewed and estimated in this
E-mail address: [email protected] (H. Knickle). section.

0378-7753/02/$ – see front matter # 2002 Elsevier Science B.V. All rights reserved.
PII: S 0 3 7 8 - 7 7 5 3 ( 0 2 ) 0 0 3 7 0 - 1
 S. Yang, H. Knickle / Journal of Power Sources 112 (2002) 162–173 163

Table 1 advantage of the solubility equilibrium of the alumina

Comparison of electrochemical couples [1,2] hydrates in caustic soda solution, in accordance with one
Couple Theoretical open Gravimetric or the other of the following equations:
circuit voltage (V)
Theoretical energy Actual energy AlOOH þ NaOH ! NaAlO2
density (Wh/kg) density (Wh/kg)

Al/air 2.7 8140 300–500

AlðOHÞ3 þ NaOH ! NaAlO2 þ H2 O
LiSOCl2 3.6 1462 374–440
AgZn 1.86 526 140–200 The Bayer process consists of four stages [10]:
AgCd 1.4 318 70–100
 Digestion—in which the finely ground bauxite is fed into
PbAC 2.15 252 30–45
NiCd 1.35 244 40–51 a steam-heated unit called a digester, where it is mixed,
NiMH 1.35 206 50 under pressure, with a hot solution of caustic soda. The
Li-ion 3.6 631 130 aluminum oxide of the bauxite and some of the silica react
Zn/air 1.4 1050 200–300 with the caustic soda forming a solution of sodium
aluminate or green liquor and a precipitate of sodium
aluminum silicate.
2.1.1. Aluminum production process
 Clarification—in which the green liquor or alumina-
In industry, extraction of aluminum metal takes places in
bearing solution is separated from the waste (the undis-
three main stages—bauxite mining, alumina production by
solved iron oxides and silica which were part of the
Bayer process, and alumina electrolysis (the Hall–Herout
original bauxite and now make up the sand and the red
process). To get high purity aluminum (>99.99%) electric
mud waste). This stage involves three steps: firstly, the
refining is needed [8].
coarse sand-sized waste is removed and washed to recover
caustic soda; secondly, the red mud is separated out; and
2.1.1.1. Bauxite miming [9]. Bauxite is the principal ore of
thirdly, the remaining green liquor is pumped through
aluminum. Bauxite is mined by surface methods (open cast
filters to remove any remaining impurities. The sand and
mining) in which the topsoil and overburden are removed by
mud are together pumped to residue lakes and the liquor is
bulldozers and scrapers. The underlying bauxite, broken by
pumped to heat exchangers where it is cooled from 100 to
explosives if necessary, is mined by front-endloaders, power
50–60 8C.
shovels or hydraulic excavators. Sometimes the bauxite is
 Precipitation—in this stage, the alumina is precipitated
crushed and washed to remove some of the clay and sand
from the liquor as crystals of alumina hydrate. To do this,
waste and then dried in rotary kilns. Other bauxites may just
the green liquor solution is mixed in the tall precipitator
be crushed or dried. The ore is then loaded into trucks,
vessels with small amounts of fine crystalline alumina,
railway cars, or onto conveyor belts, and transported to ships
which stimulates the precipitation of solid alumina
or refineries.
hydrate, as the solution cools. As the reaction is slow
Bauxite consists of aluminum oxide, Al2O3, more or less
and limited, it takes 50–80 h to obtain the desired degree
hydrated and containing various impurities, such as iron
of precipitation. When completed, the solid alumina
oxide, aluminum silicate, titanium dioxide, quartz, and
hydrate is passed onto the next stage and the remaining
compounds of phosphorus and vanadium. The composition
liquor, which contains caustic soda and some alumina,
of bauxites used in the aluminum manufacture varies a great
goes back to the digesters.
deal, but the variations generally fall within the limits shown
 Calcination—in the final stage, the alumina hydrate is
in Table 2.
washed to remove any remaining liquor and dried. Finally,
it is heated to about 1000 8C to drive off the water of
2.1.1.2. Alumina production. In nearly all commercial
crystallization, leaving the alumina, which is a dry, pure,
operations, alumina is extracted from the bauxite by the
sandy material.
Bayer process. The Bayer process, discovered by Karl Bayer
in 1888, is carried out entirely in the aqueous phase, taking The physical properties of this alumina, characterized by
its particle size and degree of calcination, are important to
the user. One such property is its content of residual com-
Table 2 bined water and absorbed water. Furthermore, the alumina
Typical composition of bauxites for aluminum manufacture [8]
should be as pure as possible. Alumina of good quality has
Composition Percentage the characteristics shown in Table 3.
H2O, combined 12–30
A2O3, total 40–60 2.1.1.3. Electrolysis. All commercial production of
SiO2, free and combined 1–15 aluminum is based on the Hall–Heroult process in which
Fe2O3 7–30 the aluminum and oxygen in the alumina are separated by
TiO2 3–4
electrolysis. This consists of passing an electric current
F, P2O5, V2O5, and others 0.05–0.2
through a molten solution of alumina and natural or
164 S. Yang, H. Knickle / Journal of Power Sources 112 (2002) 162–173

Table 3 Table 5
Composition of a good alumina [8] Comparable cost of aluminum production in 2001 [11]

Composition Percentage Company Cost of production (US$/kg)

H2O, combined (loss on calcination) 0.05–0.15 Kaiser (USA) 1.32

H2O, adsorbed (loss at 110 8C) 0.20–0.50 Alcoa (USA) 1.25
SiO2 0.005–0.015 Reynolds (USA) 1.19
Fe2O3 0.005–0.020 Alumax (USA) 1.18
TiO2 0.004–0.005 Alcan (Canada) 1.11
P2O5 <0.002 VAW (Germany) 1.30
V2 O5 <0.001 Nalco (India) 0.90
ZnO <0.010
World average 1.25
Na2O 0.40–0.80

synthetic cryolite (sodium aluminum fluoride). The molten 2.1.1.4. Electrolytic refining [8]. The Hall–Herout process
solution is contained in reduction cells or pots which are cannot ensure purity higher than about 99.9%; other
lined at the bottom with carbon (the cathode) and are techniques were therefore required when an extremely
connected in an electric series called a potline. Inserted high purity was desired.
into the top of each pot are carbon anodes, the bottoms of The principle of electrolytic refining, as described by
which are immersed in the molten solution. Betts in 1905, is still in use today. It is based on the use
The passage of an electric current causes the oxygen from of a bath containing three layers. The bottom of the cell,
the alumina to combine with the carbon dioxide gas. The which is a carbon anode, as well as its sides which are
remaining molten metallic aluminum collects at the cathode nonconductors of electricity, are covered with a dense layer
at the bottom of the pot. Periodically, it is siphoned off and of aluminum–copper alloy. Upon this layer rests another
transferred to large holding furnaces. Impurities are layer of electrolyte, which contains aluminum cations. The
removed, alloying elements added and the molten aluminum density of this layer is slightly below that of the aluminum–
is cast into ingots. The material and energy consummations copper alloy but above that of aluminum itself. Finally,
are listed in Table 4 [8]. covering these is a third layer, which is pure refined serving
A good portion of the metallic impurities present in the as the cathode. The metal may attain a purity of 99.995%.
raw materials, especially Fe, Si, Ti, V, and Mn, may also be The material and energy consummations are shown in
found in the aluminum. That is why the raw materials should Table 6 [8].
be prepared in as pure a state as possible. As a rule, the metal The cost of 1 kg of refined aluminum can be estimated as
is 99.6–99.8% pure. following:
The cost of aluminum (99.9% grade) can be estimated as refined aluminum cost
following:
¼ aluminum cost ð99:9% gradeÞ
aluminum ð99:9% gradeÞ cost þ electricity cost þ other material costs:
¼ alumina cost þ electricity cost According to Table 6, aluminum (99.9% grade) cost is
þ anode material cost þ cryolite cost: US$ 1.25/kg-Al and the electricity cost to refine 1 kg of
aluminum is US$ (18  0:06) ¼ US$ 1.08.
From Table 4, the electricity cost is US$ 0.06/kWh in the
Total cost of refined aluminum ð99:995% gradeÞ
USA and the production cost of alumina/kg-metal is about
US$ 0.30 [11]. ¼ US$ ð1:25 þ 1:08Þ ¼ US$ 2:33=kg-metal:
Total cost of aluminum/kg is about US$ (ð0:06  15Þþ
0:30) ¼ US$ 1.20 (excluding the cost of anode material and 2.1.2. Recycled fuel cost estimation
cryolite); this value approximately equals the cost listed in In an Al/air battery system, the anode used is of high
Table 5 (world wide average US$ 1.25/kg-metal). purity (99.995%) with a small amount of alloy elements that

Table 4 Table 6
Material and energy consumption of production for 1 kg of aluminum Material and energy consumption for production of 1 kg of refined
(99.9%) [8] aluminum (99.99%) [8]

Material and energy Material and energy

Alumina (kg) 1.9 Aluminum (99.9%) (kg) 1

Anode material (kg) 0.45 Bath material (kg) 0.06
Cryolite (kg) 0.07 Graphite (kg) 0.02
Energy to produce 99.9% Al (kWh/kg) 15 Energy to produce 99.99% Al (kWh/kg) 18
 S. Yang, H. Knickle / Journal of Power Sources 112 (2002) 162–173 165

Table 8
Thermodynamic properties of various chemicals [13]

Element or compound Formation energy, DHf8 (kJ/mol)

Al(s) 0
Al(OH)3 1272.8a
Al2O3 1669.8
H2O 241.8
O2 0
Fig. 1. Refined aluminum production process. a
Source: [14].

NaAlO2 which is precipitated in the form of hydrargillite,

Al2O3 3H2O or Al(OH)3. This is heated and decomposed
into Al2O3. The Al2O3 is regenerated to aluminum metal by
electrolysis.
Fig. 2. Aluminum recycling process in an Al/air EVs for computation of In our calculation of fuel energy cycle efficiency, our basis
theoretical fuel energy cycle efficiency. is 1 mol of aluminum. Assuming complete battery energy
discharge is 80% (20% unused aluminum). The heats of
formation used in our calculations are listed in Table 8. The
have positive effects on the performance of the anode, i.e. EVs aluminum fuel cycle is shown in Fig. 3.
high open circuit potential and low corrosion. Use of the
battery produces Al(OH)3. Al(OH)3 is then calcined into 2.2.1. EVs aluminum fuel cycle energy balance
Al2O3 followed by electrolysis to Al (99.995%). We assume components
the impurities such as Fe, Si and other harmful elements are There are four components of the fuel cycle energy
not added during this stage. Energy and material consump- balance of the EVs. These include the usable energy con-
tion is considerably reduced through this recycle process. tained in the aluminum anodes, the heat of reaction from
For comparison, the refining of aluminum from bauxite is Al(OH)3 to Al2O3, the heat of reaction from Al2O3 to Al, and
shown in Fig. 1 in a very simplified form. The aluminum unused aluminum reformed to anode plates. These are
recycle process from the chemical product of a battery depicted in Fig. 3.
system is shown in three steps in Fig. 2. The cost for
1. Energy contained in used aluminum: an optimistic value
recycling aluminum consists of the cost of electrolysis
for the useable energy density of aluminum is estimated as
and the cost of calcination. If the energy consumption is
564 kJ/mol (72% of theoretical value of 783 kJ/mol [12]).
still 15 kWh/kg-Al for electrolysis of alumina, the cost of
Thus, 0.8 mol aluminum has useable energy of 451 kJ.
electrolysis is approximately US$ 0.90/kg-Al (excluding
2. The heat of reaction from Al(OH)3 to Al2O3: the
other material cost, labor cost etc.). The cost of calcinations
reaction is
in the recycle process is about 22% (assuming that this is
equal to the energy requirement of calcination divided by AlðOHÞ3 ðsÞ ! 12 Al2 O3 ðsÞ þ 32 H2 OðgÞ þ DH

that of electrolysis; see Section 2.2 for energy requirements

The heat of reaction from the formation energies for
of each process) that of calcination or US$ 0.20. So the total
the above reaction is
cost of the anode with recycle is about US$ 1.10/kg-Al. The
cost of the anode with and without recycling is listed in DH
¼ 12 ð1169:8Þ þ 32 ð241:8Þ  ð1272:8Þ
Table 7. ¼ 325:2 kJ=mol:

2.2. Fuel efficiency (calculation basis: 1 mol Al)

When placed in the battery, aluminum reacts with oxygen

from the air and with water in the electrolyte to provide
power for the EVs. The product of the discharge reaction is

Table 7
Estimated cost of anode materials from electrolytic refining and recycled
from Al/air EVs

Anode material Cost (US$/kg)

Refined aluminum (from mine) 2.33

Recycled aluminum 1.10 Fig. 3. Fuel cycle of Al/air EVs: cycle used for computation of fuel energy
cycle efficiency.
166 S. Yang, H. Knickle / Journal of Power Sources 112 (2002) 162–173

Decomposition of 0.8 mol Al(OH)3 needs 260.2 kJ total cycle efficiency including recycle for the Al/air EVs in
energy to decompose into 0.4 mol of Al2O3. the steady state of (0:315  0:64) 20.2%.
3. The heat of reaction from Al2O3 to Al: From these efficiency calculations we construct Fig. 4. It
must be noted that, here the anode energy density of 564 kJ/
Al2 O3 ! 2Al þ 32 O2
mol (5.8 kWh/kg-Al, projected techniques; see Section 3) is
The heat of above reaction is 1669.8 kJ/mol. Electro- assumed. If the anode energy density of 418 kJ/mol
lyzing 0.4 mol Al2O3 needs 667.92 kJ in theory to (4.3 kWh/kg-Al, present techniques; see Section 3) is used
produce 0.8 mol of aluminum. Actually [8] industry uses in the calculation, the fuel energy efficiency in this cycle is
15 kWh to produce 1 kg of aluminum. Therefore, about 15%.
industry would use 1166 kJ to produce 0.8 mol alumi-
num.
4. Unused aluminum to anode plate: aluminum requires 3. Battery system analysis
only 26 kJ/mol to melt. So the energy requirement of
melting unused 0.2 mol Al is 5.2 kJ. 3.1. Cell performance model

The mathematical model of the Al/air cell provides the

2.2.2. Fuel energy efficiency estimation with recycle means to simulate the electrical characteristics of the Al/air
battery during changing operating conditions. Cell charac-
2.2.2.1. Theoretical efficiency. The theoretical energy teristics are also a key determinant of the physical char-
density of Al is 783 kJ/mol-Al. For a theoretical cycle for acteristics of the Al/air battery and its associated vehicle.
1 mol aluminum consumed, the total theoretical energy The power and energy characteristics of the cell principally
required is made up from the energy used in calcinations determine total battery mass (BM), which greatly influence
(325.2 kJ) and the energy used in electrolysis (834.9 kJ). total vehicle mass (VM) and cost [12].
The total theoretical energy require is 1160.1 kJ. Thus, the Voltage and current data from Behrin et al. [12] were
energy efficiency of the recycle process is (783/ modeled with a cubic polynomial using
ð1160:1  100Þ) 67.5% (see Fig. 2). This is the maximum
possible efficiency of the recycle process. V ¼ a þ bðirÞ þ cðirÞ2 þ dðirÞ3 (1)

2.2.2.2. Actual efficiency. From Fig. 3, the actual total where a ¼ V0 ; V0 is open circuit potential of the cell; r is the
energy required for recycling is made up of the three resistivity of electrolyte.
components the previous section (260:2 þ 1166 þ 5:2). The parasitic current is modeled as follows:
The total energy used is 1431.4 kJ. The energy efficiency
ip ¼ ip0 þ mi (2)
for regeneration and recycle is the energy produced from the
battery (451 kJ) divided by the energy required for recycle Eq. (1) can be rewritten as
(1431 kJ) or 31.5%.    
V b c d
A battery efficiency of 0.8, and motor/transmission effi- 2
¼1þ ðirÞ þ ðirÞ þ ðirÞ3 (3)
ciency of 0.8, results in the EVs operating efficiency of V0 a a a
(0:8  0:8  100) 64%. These factors result in an actual
The power density is
Pd ¼ Vi (4)

The energy density is

q0 P d
E¼ (5)
ði þ ip Þ
where E is the energy density (kWh/kg-Al); q0 ¼ F/
Weq ¼ 2:98 (kAh/kgeq); F ¼ 96487 (kC/kgeq); Weq ¼ 9
(kg-Al/kgeq).
Combining Eqs. (4) and (5) gives
q0 Vi
Ed ¼ (6)
ði þ ip Þ
or,
  
V i
Ed ¼ q0 V0 (6)0
Fig. 4. Fuel energy cycle efficiency in an Al/air EVs. V0 i þ ip
 S. Yang, H. Knickle / Journal of Power Sources 112 (2002) 162–173 167

Table 9
Parameters for calculating present and projected Al/air battery character-
istics at 1.0 M aluminate, 4N NaOH and 60 8C [12] (see Eqs. (1) and (2))

Parameters Present Projected

a (V) 1.8 2.2

b (mm) 13.95 13.95
c (mm2/V) 152.6 152.6
d (mm3/V2) 767.4 767.4
r (O m) 0.1866 0.1866
m 0.015 0.015
ip0 (kA/m2) 0.100 0.024

Fig. 5. Effects of corrosion current density on the energy density of anode

material.
Let Ed0 ¼ q0 V0 , then Eq. (6)0 becomes
  
Ed V i
¼ (6)00
Ed0 V0 ði þ ip Þ

3.1.1. Current and projected Al/air battery characteristics

The key present stage characteristics are parasitic current,
present stage power density, energy density, and voltage.
The parasitic current (corrosion current) has a great effect on
the energy density of the anode as shown in Fig. 5. The
effects of the present stage power density, energy density,
and voltage are shown in Figs. 6–8 (Table 9) [12]. The
present stage battery characteristics are based on Behrin et al.
[12] so-called ‘‘near-term’’ characteristics, with slight mod-
ification. This modification is based on our experimental
results and other authors’ work [15–17]. From Fig. 8 the
energy density is evaluated at a current density 2 kA/m2 for
normal operation.
The projected cell characteristics are similar to those
given by Behrin et al. [12], the improved characteristics
Fig. 6. Present voltage and energy density characteristics for an Al/air cell
could be achieved by improving air cathodes, anodes and
with an inter-electrode gap of 2 mm, and 1.0 M aluminate concentration at electrolytes. The projected characteristics of the cell are
60 8C. shown Figs. 9–11.

Fig. 7. Present voltage and power density characteristics for an Al/air cell with inter-electrode gap of 2 mm, and 1.0 M aluminate concentration at 60 8C.
168 S. Yang, H. Knickle / Journal of Power Sources 112 (2002) 162–173

Fig. 8. Energy density for the present stage Al/air cell at 60 8C. Fig. 11. Energy density for the projected Al/air cell characteristics at
60 8C.

The resulted energy density, peak power (PP) density for

present and projected characteristics are listed in Table 10.

3.2. Mass and volume of battery system analysis

3.2.1. Effect of power density and energy on mass and

volume of battery system
If the peak power of the battery is 34.8 kW, anode mass is
82.9 kg and total number of cells 136 (two module with 68
cells each), we calculated the dependence of the mass and
volume of the battery system on the peak power density and
energy density of anode. From Fig. 12, it can be concluded
that, an increase in both power density and energy density
can considerably decrease both mass and volume of the
battery system.

Fig. 9. Projected voltage and energy density characteristics for an Al/air 3.2.2. Mass distribution and volume of Al/air battery
cell with an inter-electrode gap of 2 mm, and 1.0 M aluminate
system
concentration at 60 8C.
The mass distribution and volume of the Al/air battery
system are shown in Figs. 13 and 14. It can be seen from

Fig. 10. Projected voltage and power density characteristics for an Al/air cell with inter-electrode gap of 2 mm, and 1.0 M aluminate concentration at 60 8C.
 S. Yang, H. Knickle / Journal of Power Sources 112 (2002) 162–173 169

Table 10
Present and projected battery characteristics

Parameters Present Projected

Ed (kWh/kg-Al) 4.3 5.8

Pd (kW/m2) 6.0 7.6

Fig. 14. Mass distribution of a fully fueled Al/air battery system (projected).
Peak power ¼ 34:8 kW, total mass ¼ 240 kg, total volume ¼ 329 l.

Fig. 12. Effect of power density on mass and volume of the battery
system.

Fig. 15. Cost distribution of a fully fueled Al/air battery system (2001
US$) (present characteristics). Peak power ¼ 34:8 kW, total cost ¼ US$
1120.

34.8 kW predicted are US$ 1019 or 32/kW (present) and

US$ 1120 or 29/kW (projected), respectively. The cost can
decrease 9% due to the improved characteristics. It can
also be seen from both Figs. 15 and 16 that, the major cost
items are the air cathode modules, which take up for
approximately 50% of the total cost. In this calculation,
Fig. 13. Mass distribution of a fully fueled Al/air battery system (present). the air cathode cost is US$ 100/m2 is assumed (the cost of
Peak power ¼ 34:8 kW, total mass ¼ 267 kg, total volume ¼ 371 l. some of the current commercial air cathode is higher than

Figs. 13 and 14 that, the mass of battery decreases 10% and

volume decreases 11% due to the improved battery char-
acteristics. And also, most of the battery system mass comes
from the reactants. For the present characteristics, the
reactants take up 67% of the total mass and for projected
characteristics, the reactants take up 70% of the total mass.
From Figs. 13 and 14, the energy densities calculated are
1300 and 2000 Wh/kg for present and projected techniques
respectively, these values are far larger than the values for
Al/air system listed in Table 1.

3.2.3. Cost distribution of the battery system (in 2001 US$)

Cost and cost distribution of battery system are shown in
Figs. 15 and 16 for both present and projected character- Fig. 16. Cost distribution of a fully fueled Al/air battery (2001 US$)
istics. The costs of the battery system with peak power (projected). Peak power ¼ 34:8 kW, total cost ¼ US$ 1019.
170 S. Yang, H. Knickle / Journal of Power Sources 112 (2002) 162–173

this value and are not used in our analysis). The cost of Al/air Table 11
battery is much less than the cost of the present fuel cell Manufacturing cost allocation by group and subgroup for ICEs [22]
stack (US$ 1219/kW) [18]. Vehicle group and subgroup Share of vehicle cost (%)

Group Subgroup

4. Life-cycle cost analysis of electric vehicles Body group 34.5

Engine group 16.00
Transmission group 4.75
Electric vehicle (EVs) have low emissions and are capable Chassis group 22.00
of using regeneratable energy. They are an alternative to Fuel storage 0.45
conventional ICEs. Many battery technologies have been Others 21.55
studied and appraised for EVs including lead/acid, zinc/air, Vehicle assembly 23.00
nickel metal hydride (NiMH), nickel cadmium batteries Total 100.00
[19–21]. The Al/air battery offers the potential of a very
high energy density system [16]. There are few papers in the
literature about the use of the Al/air battery system as a
power source for EVs. Life-cycle energy and cost analysis parts. Detailed cost allocation and calculation of ICEs
calculations have been made using a cost program developed (Table 11) were given by Vyas et al. [22]. Once the compo-
for estimating the life-cycle and operating costs of Al/air nents unique to the EVs have been defined and their man-
battery powered sub-compact cars. ufacturing cost assigned, it is possible to calculate the cost of
the entire vehicle.
4.1. Cost basis
4.2. Cost analysis model
An electrically powered passenger vehicle shares many
common components with a conventional vehicle. The The Al/air battery data and cost analysis were based on a
differences are mainly in the power train (engine and modified battery physical model and cost program originally
transmission group) and the fuel energy storage system, developed by Behrin et al. [12]. All the material cost is
i.e. Al/air battery system. Both types have similar bodies and updated to 2001 using the chemical process cost index [23].
chassis components. Combining these similar components The peak power requirement of the traction motor was com-
account for 78–80% of the total vehicle cost. Therefore, we puted to meet an acceleration requirement of 0–96 km/h in
base the EVs cost estimates mostly from available ICEs data about 20 s. The normal power (NP) used is calculated
and then make necessary adjustments for the differences in according to a mix of SAEJ1798 [24] and 96 km/h high
the hardware and system between the two for the remaining way cruising. Once the peak power, normal power, range and

Fig. 17. Schematic of cost analysis procedure.

 S. Yang, H. Knickle / Journal of Power Sources 112 (2002) 162–173 171

Table 12
Basic specifications of EVs analyzed

Vehicle characteristic Al/air Lead/acida NiMHa

Present Projected Initially After 20 years Initially After 20 years

b
PWR (W/kg) 33 33 33 33 33
Battery capacity (kWh) 204 204 19 12 32 23
Traction motor power (kW) 48 40 56 44 55 44
Battery mass (kg) 280 221 481 245 442 261
Vehicles mass (kg) 1280 1088 1571 1201 1525 1219
Estimated range (km) 400c 400c 122 98 212 192
a
Data predicted for 2000 [22].
b
PWR: traction motor peak power/(vehicle mass þ 136 kg of payload).
c
Battery has enough aluminum anode to have a range 1600 km requiring four stops for addition of water.

speed are assigned, the mass, volume and cost of the battery previous paragraph and some are set as default parameters,
system can be calculated (assuming the battery system to the such as energy density of aluminum anode, peak power/m2 of
traction motor efficiency of 0.85). the battery, concentration of electrolyte, and various material
The main choices in the power train design are dependent costs, etc. once the required parameters are set, then we can
upon the type of motor and controller to use. We analyzed perform battery calculation and cost analysis using the battery
the production of electric traction motor and controls under a physical model and cost program. The battery mass to vehicle
high volume production scenario. We assumed an induction mass ratio is confined to 0.2–0.3, which is generally accepted
motor in our cost analysis presented in this paper. The motor value. The output parameters are battery mass, vehicle mass,
costs US$ 17/kW and the controller US$ 45.50/kW [22]. battery manufacture cost (BMC) and vehicle manufacture
For the life-cycle cost analysis, the ICE car purchase price cost (VMC) and various cost related items.
is set at US$ 15,500. The fuel economy for ICEs is about
30 mpg [25]. Both ICEs and EVs are assumed to have a life 4.4. Specifications and cost allocations of vehicles
of 12 years and a range of 17,600 km per year. The total life- analyzed
cycle cost for the EVs or ICEs are calculated by adding up
the annualized initial user’s cost fuel cost and non fueling Table 12 indicates the basic specifications of vehicles
cost such as maintenance cost, tire replacement cost, insur- analyzed. The peak power to vehicle mass including 136 kg
ance cost, cathode replacement cost, and other cost. These of payload is kept at about 33 W/kg. Compared with lead/
are converted to annual cost/km. acid and NiMH EVs. Al/air EVs have less vehicles mass,
much higher battery capacity (at least 10 times that of lead/
4.3. Cost analysis procedures acid and 6 times that of NiMH). The projected improvement
of Al/air battery characteristics is due to the improvements
The schematic flow diagram of Fig. 17 describes the main of aluminum anode energy density from 4.3 to 5.8 kWh/kg
program for the EVs powered by the Al/air battery. The main and battery peak power from 6.0 to 7.6 kW/m2, thus the
input parameters are peak power of traction motor to vehicle resulting vehicles mass is lowered without sacrifice of the
mass (PWR), normal power, and by assuming the peak power battery capacity and vehicle range. The initial mass of the
of traction motor, we can calculate the vehicle mass. Other lead/acid and NiMH vehicles is quite heavy due to the initial
input parameters associated with vehicle and battery mass and high mass of the batteries of both these vehicles. The pro-
cost calculation are not shown here, some of them are stated in jected lead/acid and NiMH have battery mass comparable

Table 13
Cost allocation for EVs (%)

Vehicle cost component Al/air Lead/acida NiMHa

Present Projected Initially After 20 years Initially After 20 years

Common components and assembly 55.3 58.6 57.5 79.4 33.6 62.8
EV power train 28.1 25.6 23.0 7.7 13.1 6.1
Battery pack 14.5 13.6 17.3 10.6 52.0 29.3
Subsystem premiums 2.0 2.2 2.2 2.3 1.3 1.8
Total 100.00 100.00 100.00 100.00 100.00 100.00
a
Data predicted for 2000 [22].
172 S. Yang, H. Knickle / Journal of Power Sources 112 (2002) 162–173

Table 14 5. Summary
Cost ratios of Al/air, lead/acid, and NiMH EVs in comparison to ICEs

Items Al/air Lead/acida NiMHa The cost of aluminum as an anode can be as low as US$
1.1/kg as long as the reaction product is recycled. The total
Present Projected
fuel efficiency during the cycle process in an Al/air EVs can
Price ratio 1.3 1.25 1.25 1.9 be 15% (present stage) or 20% (projected), comparable to
Fuel cost ratio 1.35 1.01
that of ICEs. The battery energy density is 1300 Wh/kg
Life-cycle cost 1.22 1.14 1.2 1.68
per km ratio (present) or 2000 Wh (projected), far better than other
Estimated range 400b 400b 130 180 electrochemical couples shown in Table 1. The cost of
a
the battery system chosen is US$ 30/kW (present) or US$
Data predicted for 2000 [22].
b 29/kW (projected), far less than the present hydrogen fuel
Battery has enough aluminum anodes to have a range 1600 km
requiring four stops for addition of water. cell stack (US$ 1219/kW).
We have conducted Al/air EVs life-cycle analysis and
(slightly less than) with aluminum/air, but sacrifice some of compared the results with that of lead/acid and NiMH EVs.
the battery capacity reducing the range of these vehicles. Only the Al/air vehicles can be projected to have a travel
Also, only the projected Al/air battery EVs have a vehicle range comparable to ICEs. The purchase price of aluminum/
mass (1088 kg) similar to the average ICEs mass and a range air EVs would cost 25–30% more than the present ICEs. With
(400 km) comparable with that of the ICEs. the decrease in cost of the power train, the purchase price will
Table 13 indicates the costs allocations of vehicle com- likely drop farther. The fuel cost of the Al/air EVs at present is
ponents for each type of EVs. Compared with ICEs, the Al/ 35% higher than that of ICEs. But with improvement in
air EVs common components and assembly costs are battery characteristics, the fuel cost would fall to a level
assumed to be 55.3%. This is much less than that of ICEs almost equal to that of ICEs. The life-cycle cost of Al/air EVs
at 78–80% due to the higher cost of the EV power train and is 14% (projected) and 22% (present) higher than that of
battery pack. Compared with the lead/acid and NiMH EVs, ICEs. Only the lead/acid EVs life-cycle cost is comparable to
the Al/air EVs power train cost decreases from 28.1 to Al/air EVs life-cycle cost while the NiMH EVs is the highest.
25.6%, a difference of only 2.5%, while the latter two types From this analysis Al/air EVs are more promising candi-
of EVs power train cost decreases from 23.0 to 7.7% and dates, than lead/acid or NiMH, for replacement of ICEs
from 13.1 to 6.1%. In our calculation, we keep the cost of the considering travel range, purchase price, fuel cost, and life-
power train of EVs constant, and in the lead/acid and NiMH cycle cost.
EVs calculation (Vyas et al. [22]) the power train and battery
pack cost are both projected to decrease.
Acknowledgements
4.5. Purchase price and life-cycle costs
This project is partially funded through the RI State
Table 14 compares various cost ratios for the three EVs of Energy Office.
Al/air, lead/acid and NiMH to those of ICEs. The calculated
results for the actual purchase price and life-cycle cost are
not shown in Table 14 but are summarized here. The References
purchase price for ICEs is US$ 15,500 and for that of the
Al/air EVs is US$ 20,150 (present techniques) or US$ [1] D. Linden, Handbook of Batteries and Fuel Cells, NAVSEA Battery
19,300 (projected techniques). Fuel cost for ICEs is Document (NAVSEA-AH-300), 1st ed., McGraw-Hill, New York,
3.1 cents/km and for Al/air EVs is 4.2 cents/km (present) 1993.
or 3.2 cents/km (projected). Life-cycle cost per kilometer for [2] T.A. Dougherty, A.P. Karpinski, J.H. Stannard, W. Halliop, S.
Warner, Aluminum–air: status of technology and applications, in:
ICEs is about 22.0 cents/km and for Al/air EVs is about Proceedings of the Conference on Intersociety Energy Conversion
26.8 cents/km (present) or 25 cents/km (projected). Engineering, IEEE V2 (1996) 1176–1180.
From the above results and Table 14, we can see that, the [3] A.R. Despic, D.M. Drazic, P.M. Milovan, Electrochemically active
purchase price of Al/air EVs is 25–30% higher than the ICEs. aluminum alloy, the method of its preparation and use, US Patent
This value is almost equal to that of lead/acid EVs, and is lower 4,098,606 (1978).
[4] B.B. Jovanovic, A.R. Despic, D.M. Drazic, Electrochemically active
than that of NiMH EVs. At the present stage of development, aluminum alloy and composite, US Patent 4,288,500 (1981).
fuel cost of Al/air EVs would be 35% higher than that of ICEs. [5] P.W. Jeffrey, W. Halliop, F.N. Smith, Aluminum anode alloy, US
But with the improvement of battery characteristics (projected Patent 451,086 (1988).
battery aluminum anode characteristics), the fuel cost would [6] J.A. Hunter, G.M. Scamans, W.B. O’Callaghan, Aluminium
decrease to a value comparable to that of ICEs. Life-cycle cost batteries, Alcan International Limited, Montreal, CA, US Patent
4,942,100 (1990).
of Al/air EVs would be 14 and 22% higher than that of ICEs. [7] Z. Solomon, C.C. Norman, R.M. Mazgaj, Aluminum-consuming
These life-cycle costs are similar to lead/acid EVs and much fluidized-bed anodes, J. Electrochem. Soc. 137 (6) (1990) 1851–
lower than that of NiMH EVs. 1856.
 S. Yang, H. Knickle / Journal of Power Sources 112 (2002) 162–173 173

[8] F.M. Herman, J.M. John Jr., F.O. Donald, S. Anthony (Eds.), [18] K.S. Jeong, B.S. Oh, Fuel economy and life-cycle cost analysis
Aluminum and Aluminum Alloys, Encyclopedia of Chemical of a fuel hybrid vehicle, J. Power Sources 5 (1) (2002) 58–
Technology, 2nd ed., vol. 1, Wiley, New York, 1963. 65.
[9] http://www.eaa.net/pages/material/production.html, 2001. [19] A.K. Shukla, A.S. Arico, V. Antonucci, An appraisal of electric
[10] http://www.minerals.org.au/pages/page3_34.asp, 2001. automobile power sources, Renewable Sustainable Energy Rev. 5 (2)
[11] http://www.indianinfoline.com/sect/alum/ch07.html, 2001. (2001) 137–155.
[12] E. Behrin, R.L. Wood, J.D. Salisbury, D.J. Whisler, C.L. Hudson, [20] I. Rade, B.A. Andersson, Requirement for metals of electric vehicle
Design Analysis of an Aluminum–Air Battery for Vehicle Operations, batteries, J. Power Sources 93 (1/2) (2001) 55–71.
Transport system Research, Lawrence Livermore Laboratory, 1983. [21] R.F. Nelson, Power requirement for batteries in hybrid electric
[13] M.B. Gordon, Physical Chemistry, 3rd ed., McGraw-Hill, New York, vehicles, J. Power Sources 91 (1) (2000) 2–26.
1973, pp. 779–784. [22] A. Vyas, R. Cuenca, L. Gaines, An assessment of electric vehicle
[14] http://www.eiu.edu/eiuchem/GenChem/downloads/tutorial8.pdf, life-cycle costs to consumers, in: Proceedings of the Conference and
2001. Exposition on Total Life-Cycle, Argonne National Laboratory,
[15] D.W. Gibbons, E.J. Rudd, Development of aluminum/air batteries for Graz, Austria, 1–3 December 1998, SAE Technical Paper Series
propulsion applications, in: Proceedings of the 28th Conference on 982182.
Intersociety Energy Conversion Engineering, vol. 1, Atlanta, 8–13 [23] Economic Indicators, M&S Equipment Cost Index, Chemical
August 1993. Engineering, July 1981 to May 1982 and May 2001.
[16] R.P. Hamlen, W.H. Hoge, J.A. Hunter, W.B. O’Callaghan, Applica- [24] Recommended practice for performance rating of electric vehicle
tions of aluminum–air batteries, IEEE Aerosp. Electron. Syst. Mag. 6 battery modules, SAE J1798, January 1997.
(10) (1991) 11–14. [25] Effectiveness and Impact of Corporate Average Fuel Economy
[17] M.L. Doche, F. Novel-Cattin, R. Durand, J.J. Rameau, Characteriza- (CAFE) Standards, Committee on the Effectiveness and impact of
tion of different grades of aluminum anodes for aluminum/air Corporate Average Fuel Economy (CAFE) Standards, National
batteries, J. Power Sources 65 (1/2) (1997) 197–205. Academy Press, Washington, DC, 2001.