THIN FILMS FOR ADVANCED BATTERIES
This article also appears in the Oak Ridge National Laboratory
Review (Vol. 25, No. 2), a quarterly research and development
magazine. If you'd like more information about the research
discussed in the article or about the Review, or if you have any
helpful comments, drop us a line. Thanks for reading the Review.
Batteries are a familiar part of everyday life. Many of us depend
on heavy lead-acid batteries to start our cars; a variety of dry
cells to operate cameras, toys, and portable sound systems; small
lithium-iodine batteries to power cardiac pacemakers; and tiny
lithium-magnesium oxide coin cells to run watches.
Packing more power into batteries while reducing their size and
weight has become a goal for researchers because of growing
interest in smaller, lighter batteries for a variety of purposes.
Tiny batteries smaller than a button are needed to provide
electrical energy for computer memory chips, medical implant
devices, and radiofrequency transmitters in car key sets to prevent
car theft. Medium-size batteries are useful for consumer
electronics, such as laptop computers and cellular phones.
Large-scale, lightweight batteries are needed to power electric
A major issue in the development of new batteries is the materials
used. Many household batteries contain lead, cadmium, and
mercury--all toxic materials. New Jersey now has a law that
requires manufacturers to eliminate mercury from batteries by the
end of 1995 and mandates that collection programs be set up for
certain batteries containing mercury, lead, or cadmium.
The focus of research is to create batteries that contain fewer
toxic materials, are rechargeable, and are able to pack in more
energy per unit weight and volume. Batteries using lithium have
yielded promising results.
The size and weight of a battery are generally thought to be
related to the amount of energy it stores. Hence, the larger a
battery, the more electrical energy it can supply. Electrical
energy is expressed in watt-hours--the product of the current,
voltage, and discharge time when the battery is in use. However,
more important figures of merit for a battery are the energy
density, which is watt-hours divided by battery volume in liters,
and specific energy, which is watt-hours divided by battery weight
in kilograms. Although a lead-acid battery in a car can produce a
large amount of energy, its energy density and specific energy are
low. On the other hand, lithium-manganese oxide coin cells in
watches cannot supply the same amount of energy, but their energy
density and specific energy are much higher.
At ORNL a group led by John Bates of the Solid State Division has
developed a thin-film lithium microbattery for computer memory
chips that is much thinner than plastic wrap. "The purpose of these
batteries is to hold the memory until the power comes back," says
Bates. If such a battery could be scaled up to propel electric
vehicles, he calculates it would be an improvement over the
lead-acid battery by a factor of 8 in energy density and a factor
of 10 in specific energy.
The main goal of the ORNL work, however, is to find ways to develop
microbatteries for microelectronics. Integrated circuits on silicon
chips have made possible the reduction of computer size. As a
result, modern appliances and cars are now controlled by
microprocessors. However, partly because of differences in the
levels of research activity and money, reduction in the sizes of
batteries has not kept pace with the reduction in the sizes of
For example, consider nonvolatile computer memory chips. They have
nonrechargeable batteries as backup so that the information stored
as electronic charges is not lost in case of a power failure.
However, each lithium battery is many times larger than the chip
using it. ORNL's development of thin-film lithium batteries offers
the means to scale down the sizes of batteries to more closely
match the sizes of microelectronic components. The ORNL work may
result in the first practical rechargeable microbattery.
BATTERY ON A CHIP
Solid-state processes for depositing thin films on a substrate are
being used at ORNL and elsewhere to increase the amount of energy
that can be stored in a battery per unit weight and volume. ORNL's
Ceramic Thin Films Group in the Solid State Division has made a
number of thin-film lithium microbatteries that have remarkably
high specific energies and energy densities. The cells have open
circuit voltages at full charge of 3.7 to 3.8 volts, the highest
achieved to date in a thin-film battery. This success has attracted
the attention of the Eveready Battery Company, which has also been
developing thin-film batteries. In March 1992, ORNL and Eveready
signed a cooperative research and development agreement (CRADA) to
facilitate the commercialization of thin-film batteries.
According to Bates, the current goal of the ORNL group, which
includes Nancy Dudney, Greg Gruzalski, and Chris Luck, is to make
thin-film batteries that can be deposited directly either on the
reverse side of a computer memory chip or onto the chip's
protective ceramic package. "The idea," he says, "is to incorporate
the battery into the integrated circuits of computer memory chips
during their manufacture. This approach would eliminate the
manufacturing step of soldering large batteries onto circuit
Bates believes that these solid-state thin-film batteries will also
offer these advantages over competitive technologies:
- They are rechargeable.
- They have high power and energy densities.
- They can be fabricated to virtually any size onto a variety
of substrate materials, such as semiconductors, ceramics,
- They can be fabricated using standard deposition techniques
and mild deposition conditions.
- They can operate over a wide temperature range, even at
temperatures near the melting point of lithium; ORNL cells
have been operated between -15øC and 150øC.
- They contain no liquid components and produce no gases.
"Our thin-film battery," Bates says, "can be deposited onto any
substrate that can withstand a temperature of 50øC, including a
variety of polymers. In addition, the shape of the battery can
conform to the shape of the support. For example, it could be
fabricated on a cylinder. Our battery can be made as small as any
multilayer device or as large as the deposition equipment allows.
For example, a thin-film battery could be deposited onto an entire
sheet of window glass."
Bates and his colleagues have been searching for the right
combination of materials to make reliable backup power sources for
computer memory chips based on complementary metal oxide
semiconductor (CMOS) technology. Currently, standby power for
computer memory chips is supplied by nonrechargeable lithium-
manganese oxide or other lithium batteries, which are soldered to
circuit boards or incorporated into a self-contained package. These
batteries are larger than the chips for which they provide backup
energy. Someday they may be replaced by smaller thin-film lithium
batteries based on the ones developed at ORNL using a newly
discovered material for the electrolyte.
The ORNL group is also adapting this microbattery technology for
use in miniature radiofrequency transmitters expected to be
commercially available someday. The group hopes to contribute to
the development of scaled-up batteries for consumer electronics and
electric cars (see sidebar below).
WHY THIN-FILM LITHIUM CELLS?
Why did Bates and his colleagues choose to investigate thin-film
lithium cells? First, they were attracted to lithium's advantages.
Because lithium has a small atomic mass and the highest
electrochemical potential for a metal, it makes a good reactant for
a battery that must have high cell voltage and high specific power.
They chose thin films for batteries because they make an effective
cell which can be manufactured by the same processes used by the
electronics industry. Battery cell components can be prepared as
thin (~1µm) sheets built up as layers. The area and thickness of
the sheets determine battery capacity.
Deposition of thin films increases the contact area of the cell
components, resulting in a high fraction of reactants. Thin films
result in higher current densities and cell efficiencies because
the transport of ions is easier and faster through thin-film layers
than in bulk materials.
The major challenge to the development of the lithium cells was to
find an electrolyte that satisfactorily conducted ions and was
stable in contact with lithium. In the fall of 1991, groups at
Eveready and the University of Montpellier in France reported the
development of solid-state rechargeable thin-film cells with
The French group's cells had low current densities and short
lifetimes, but the Eveready cells had excellent current densities
and were charged and discharged hundreds of times. The Eveready
cells use a titanium sulfide cathode and an oxysulfide-based
electrolyte. In both cases, however, the cells required an extra
layer of lithium iodide to protect the electrolyte from attack by
lithium. At ORNL Bates' group has developed an even better cell
using an oxynitride-based electrolyte, which is stable with a
BUILDING A BETTER BATTERY
The ORNL research program began in November 1986 at a Materials
Research Society meeting when Jim Roberto, director of the Solid
State Division, heard a talk on microbatteries by Minko Balkanski
of the Universiti Pierre et Marie Curie in Paris, France. There
Roberto discussed with Bates the possibility of ORNL conducting
research in support of microbattery development. In April 1987,
Bates and Dudney submitted a seed money proposal entitled
"Micropower Sources" and obtained internal funding from ORNL. By
November 1987, they had their first vacuum chamber for film
deposition. In June 1988, Bates and Dudney submitted a proposal on
"Microionics: Materials and Devices" and received support for
fiscal year 1989 from the Director's R&D Fund at ORNL. In November
1988 Bates' group fabricated the first vanadium oxide cell ever
made at ORNL and possibly anywhere else. It marked the beginning of
a series of successes for the group in microbattery development.
"Our work started from scratch," Bates says. "We had no deposition
equipment and no experience in thin films. We began our electrolyte
studies using a lithium phosphosilicate system."
The electrodes of the ORNL thin-film battery are lithium (Li) and
noncrystalline vanadium oxide (V2O5). Vanadium oxide was selected
as the cathode because it is an intercalation compound that permits
a lithium ion to move into and out of a framework without causing
more than a small expansion or contraction of the structure.
Lithium ions move into the V2O5 structure during discharge of the
cell and are forced out of the structure during recharge. The
amorphous material is preferred over the crystalline form because
three times more lithium ions can be inserted into the amorphous
cathode, thus making a battery that has a higher capacity.
In a thin-film lithium cell (see the schematic diagram of its
operation on p. 53), lithium ions leave the anode, diffuse through
the electrolyte film, and reach the V2O5 cathode. At the same time,
electrons travel through the external circuit to power a device and
then to the cathode to "combine" with the Li+ ions so that the
compound retains a net neutral charge. The cell reaction can be
represented asxLi + V2O5 = Lix V2O5.
In their effort to develop an improved electrolyte, the ORNL
scientists took into account the fact that many inorganic compounds
are better ionic conductors in the amorphous state than in the
crystalline form. For example, the conductivity of amorphous
lithium phosphate having the composition 0.6Li2O:0.4P2O5 is 109
times as high as that of crystalline lithium orthophosphate
To avoid the need for a layer of lithium iodide to protect the
electrolyte from being attacked by lithium, Bates consulted the
literature for some clues. "I learned that adding nitrogen to
sodium metaphosphate glasses improves their durability in contact
with air and water vapor. So we decided to sputter the lithium
orthophosphate in nitrogen rather than the standard gas mixture of
argon and oxygen. The resulting film was an oxynitride that
contained about 3 at. % nitrogen. It seems that nature much prefers
oxygen to nitrogen in compounds such as these, but the presence of
this small amount of nitrogen greatly improves the performance of
Currently, the electrolyte used exclusively in ORNL's thin-film
lithium batteries is lithium phosphorus oxynitride (LiPON), the
first film of which was grown at ORNL in February 1991. The LiPON
electrolyte, which is deposited over the cathode, outperforms
competitive electrolytes, such as the oxysulfide electrolytes
employed by Eveready.
To analyze the composition of the electrolyte, the group has relied
on resonance ion backscattering performed by Ray Zuhr of the Solid
State Division, electron spectroscopy for chemical analysis and
Auger electron spectroscopy carried out by Ashok Choudhury of the
Metals and Ceramics Division, and proton-induced gamma emission
analysis conducted by their collaborator, Dave Robertson, of the
University of Kentucky at Lexington.
"Tests show that the nitrogen-containing electrolyte has 30 times
the lithium ion conductivity of a film of pure amorphous lithium
ortho-phosphate," Bates says. "This is important because every
battery has an internal resistance to ion transport. Our goal is to
overcome this resistance by increasing the ion conductivity of the
electrolyte. The result will be a better battery."
HOW FILM IS DEPOSITED
The thin-film lithium cell is fabricated at ORNL by depositing
successive layers of the cathode, electrolyte, and anode using
direct-current and radiofrequency magnetron sputtering and thermal
evaporation. The battery is only 6 microns thick, or one-third the
thickness of plastic wrap (a micron is a millionth of a meter), and
cells have been deposited on alumina or glass substrates.
Magnetron sputtering is done in a vacuum chamber at near room
temperature in pure argon, oxygen, and nitrogen gases. A thin
2.5-cm-diam disk of a target material is loaded into a commercial
magnetron sputter source. The targets are vanadium metal for the
cathode and Li3PO4 for the electrolyte. A high voltage is applied
to the target material. At low gas pressures, the voltage generates
a plasma discharge. The positively charged ions in the plasma are
accelerated to the target material because of its negative charge,
and upon bombardment, some of the target atoms are ejected, or
The sputtered atoms then condense on the battery substrate
positioned about 5 cm from the sputter target. Permanent magnets
positioned beneath the target enhance the sputtering efficiency by
confining the electrons in the plasma close to the target surface,
thereby increasing the ionization of the atoms in the sputtering
"Currently, the deposition rates are quite low because pushing for
a more rapid rate of film growth would sacrifice the film's
uniformity and quality," Bates says. "But this production problem
should be considerably reduced when larger sputter sources are
The final cell layer, the lithium anode, is deposited by thermal
evaporation under vacuum.
CRADA WITH EVEREADY
The ORNL group has apparently solved the problem of lithium attack
on the thin-film electrolyte. However, a remaining problem that
must be addressed is protecting the lithium from corrosion in air.
Currently, thin-film batteries must be stored in a protective argon
atmosphere. "We must find a way to seal up the battery and make it
self-contained," says Bates.
The researchers are now working on a project to package thin-film
batteries under a CRADA signed in March 1992 between Energy Systems
and the Eveready Battery Company. Eveready had approached ORNL
about a cooperative research effort because the company had not
found a packaging solution and because the smaller-scale deposition
equipment at ORNL allows researchers to try out ideas rapidly. The
deposition system at Eveready is much larger because it is designed
for small production runs, not research.
"We will work with Eveready on determining which material could
best seal up the battery without altering the properties of our
films," Bates says. "We are testing the results of depositing a
variety of materials onto the lithium as a sealant film."
Bates predicts the group will develop a self-contained thin-film
battery of more than 3.5 volts during 1992. Then he hopes the group
may become involved in research on scaling up thin-film batteries
for use as lightweight, energy-efficient sources of power for
electric cars (see sidebar "From Chips to Cars" below).
FUTURE OF THIN-FILM BATTERY RESEARCH
The use of LiPON in thin-film batteries will make possible higher
voltage cells based on lithium cobalt oxide (LiCoO2) or lithium
manganese oxide (Li2Mn2O4). Raising the voltage of these cells is
important because it increases their energy density.
Another goal of the ORNL research is to enhance the current density
and pulse capability of the thin-film cells. One problem in the
ORNL cell is that the current appears to be limited by lithium ion
transport in the vanadium oxide cathode.
The discovery of lithium phosphorus oxynitride and its excellent
properties as an electrolyte has opened up a new area of research
for thin-film materials, which will be explored in a Basic Energy
Sciences program of the Department of Energy. In addition, the new
field of microionics could become a major target of research at
"We hope to continue to focus our basic research on improving
battery technology," Bates says. "Our work in microbatteries and
our work with other groups on microelectronics could be extended to
development of, for example, miniature radiofrequency transmitters
and remote microsensors for environmental or biomedical
In conclusion, the program started by Bates, Dudney, and others
could expand to scale up microbatteries being developed for
computer chips into macrobatteries that could be useful in many
ways, including powering electric vehicles. Use of thin films for
batteries to power everything from chips to cars may someday be
seen as a classic example of getting more from less.
A battery is one of two kinds of electrochemical devices that
convert the energy released in a chemical reaction directly into
In a battery, the reactants are stored close together within the
battery itself, whereas in a fuel cell the reactants are stored
externally. This conversion of chemical energy to electrical energy
is potentially 100% efficient, whereas the conversion of chemical
energy to mechanical energy via a thermal conversion (e.g.,
internal combustion of gasoline in cars) always results in heat
transfer losses limiting the intrinsic efficiency.
The first electric battery may have been made in ancient Egypt, but
historians often credit Italian physicist Alessandro Volta, who in
1800 assembled a series of silver and zinc disks that sandwiched
cardboard disks soaked in saltwater. The disks served as
electrodes, and the saltwater was the electrolyte.
Familiar batteries of today are the flashlight (or dry cell)
battery, which uses manganese oxide and zinc for the electrodes and
a paste as an electrolyte, and the lead-acid car battery, which
uses lead and lead oxide for the electrodes and sulfuric acid for
the electrolyte. The lead-acid battery, which was invented in 1860,
is unsurpassed for vehicle uses because it can deliver a large
The electrodes are the positively charged pole (cathode) and the
negatively charged pole (anode) of a storage battery. The
electrolyte is a chemical compound that separates the electrodes
and conducts ions released during discharge. The electrolyte forces
the electrons to flow in the device's external circuits.
Chemical energy is converted into electrical energy by an oxidation
reaction in which electrons are released to an external circuit
through the anode. Simultaneously, a reduction reaction removes
electrons from the external circuit through the cathode. The
electrolyte functions to "control" the rate of reaction between the
anode and cathode by forcing electrons to move through external
circuits, producing energy. However, the electrodes must not touch
each other to avoid internal short circuits. (See the examples of
battery reactions shown in the schematic.)
Some cells operate as two independent half cells in which reactions
occur by ion exchange with the cell electrolyte. For others, a net
cell reaction results in the formation of a new compound such as
lithium iodide or sodium sulfide. Such a reaction requires the
transport of an ion through the electrolyte from one electrode to
Among the important characteristics of battery cells are
- Cell voltage--ideal or open circuit voltage, which is higher
than the actual cell voltage when the current flows through
the cell (cell efficiency is determined by the ratio of
actual voltage to the ideal voltage)
- Power output--the magnitude of the current that can be
delivered at a given voltage
- Power density--power output per weight or volume of the
battery system, which includes the cell materials,
reactants, and necessary packaging
- Capacity--the amount of chemical reactants that can be
stored and effectively used in the battery. Batteries rarely
can be used until fully discharged, so that one of its
reactants is totally consumed
- Shelf life--the amount of time a battery can retain its
original properties when not in use
- Rechargeability--the ability of the battery to be restored
to a useful condition.
Batteries are designed differently for various applications. "For
computer memory backup power," Bates says, "the battery should have
the appropriate cell voltage, a long shelf life, rechargeability,
and the ability to be integrated with the memory chips during
electronic fabrication. For vehicle applications, the battery
should offer high energy density, large current output, full
rechargeability, rapid recharge, safety, and a reasonable
From Chips to Cars
In the fall of 1991, the U.S. government gave a jump start to the
nation's battery research by combining its resources with those of
the Big Three automakers in Detroit.
Because of renewed concerns about air pollution and future oil
shortages, there has been a resurgence of interest in electric
cars. But to be acceptable to the public, such vehicles will need
batteries that are low in cost and able to keep the vehicle
operating for many miles before a recharge is needed. In addition,
the time to recharge the battery should be short.
DOE has established an Advanced Battery Consortium in which
national laboratories, universities, and other companies will
collaborate on battery research with Ford, Chrysler, and General
Motors. The latter company is already developing electric vehicles
that would use lead-acid batteries. One reason for this move by DOE
is to make the United States more competitive in battery research
and electric vehicle development; nations taking the lead in these
areas are Great Britain, Germany, Japan, and Canada.
ORNL is expected to be named a lead laboratory for DOE and the
consortium in at least one area of electric vehicle development:
lightweight body materials. ORNL may also play a role in the
development of electric vehicle battery materials.
The potential competitors for electric vehicle propulsion include
lead-acid (Pb/acid) batteries, sodium-sulfur (Na/S) batteries, and
lithium-iron sulfide (Li/FeS) batteries. The lithium-iron sulfide
battery has higher specific energy and energy density than the
other two. However, according to Bates, a scaled-up thin-film
lithium battery would be an improvement over the lithium-iron
sulfide battery by a factor of about 4 for specific energy and by
a factor of about 5 for energy density (see table).
"Improved power density is achieved by using a lithium battery,"
Bates says. "By making a thin-film battery, we hope to increase the
maximum discharge current and decrease the time required to
recharge the battery."
(keywords: thin films, batteries, microbatteries)
Please send us your comments.
Date Posted: 2/7/94 (ktb)