Citation
The development of a high energy density lithium/oxygen battery

Material Information

Title:
The development of a high energy density lithium/oxygen battery
Creator:
Cordova, Stephen Gary
Publication Date:
Language:
English
Physical Description:
ix, 34 leaves : ; 28 cm

Thesis/Dissertation Information

Degree:
Master's ( Master of science)
Degree Grantor:
University of Colorado Denver
Degree Divisions:
Department of Chemistry, CU Denver
Degree Disciplines:
Chemistry

Subjects

Subjects / Keywords:
Lithium cells ( lcsh )
Oxygen ( lcsh )
Electric batteries -- Design and construction ( lcsh )
Electric batteries -- Design and construction ( fast )
Lithium cells ( fast )
Oxygen ( fast )
Genre:
bibliography ( marcgt )
theses ( marcgt )
non-fiction ( marcgt )

Notes

Bibliography:
Includes bibliographical references (leaf 34).
Thesis:
Department of Chemistry
Statement of Responsibility:
by Stephen Gary Cordova.

Record Information

Source Institution:
|University of Colorado Denver
Holding Location:
Auraria Library
Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
263685099 ( OCLC )
ocn263685099
Classification:
LD1193.L46 2008m C67 ( lcc )

Full Text
THE DEVELOPMENT OF A HIGH ENERGY DENSITY LITHIUM/OXYGEN BATTERY
by
Stephen Gary Cordova
B.S. Chemistry, B.S. Biology, University of Colorado Denver, 2001
A thesis submitted to the
University of Colorado at Denver/Health Sciences Center
in partial fulfillment
of the requirements for the degree of
Master of Science
Chemistry
2008


2008 by Stephen Gary Cordova
All rights reserved.


This thesis for the Master of Science
degree by
Stephen Gary Cordova
has been approved
Date


Cordova, Stephen, G.( Master of Science, Chemistry)
The Development of a High Energy Density Lithium/Oxygen Battery
Thesis directed by Dr. Mark Anderson
ABSTRACT
A Li/oxygen cell which gives a specific energy of more than 700 Wh/kg was developed. The
cell consists of a carbon cathode which scavenges oxygen from the atmosphere, a polyolefin
separator, a metallic lithium anode, and a lithium salt in an organic electrolyte. It is contained
within a plastic packaging material, which doubles as a semi-permeable membrane to
oxygen. The effect of cell components such as carbon type and electrolyte type was
evaluated. Each carbon type prefers a particular electrolyte. The most important factor was
found to be the composition of the electrolyte. Emphasis was put on lowering cell weight,
primarily by lowering electrolyte amounts (over 50% of the cell weight). The rate capability of
the cell was enhanced by the use of catalysts, while membranes allowed for the selective
permeation of oxygen while preventing the entry of moisture. It is expected that a careful
optimization of cell parameters would lead to specific energies in excess of 1000 Wh/kg.
This abstract accurately represents the content of the candidates thesis. I recommend its
publication.
Signed
Mark Anderson


ACKNOWLEDGMENT
This dissertation could not have been written without Dr. Mohamed Alamgir, who not only
served as my supervisor and instructor, but also encouraged and challenged me throughout
my academic program and employment opportunities. The artistic talent of Mr. Tim Eckert,
and assistance of Mr. James Spence and Mr. Rob Banaszak allowed me to complete my
research. Mrs. Trade Thomas and Dr. Daniel Rivers, also supported and assisted in several
aspects of this project.
The assistance and direction of Dr. Donald Zapien and Dr. Mark Anderson was an invaluable
commodity throughout the preparation and writing of this dissertation. The encouragement
and cooperation of Dr. John Lanning and Dr. Larry Anderson is also appreciated. I would also
like to thank Dr. Susan Schelble for her support and assistance in scheduling.
Finally, I would like to thank the United States Air Force for their funding and support of this
project, specifically Dr. Joe Feldner and Dr. Dave Pickett.


TABLE OF CONTENTS
Figures......................................................viii
Tables.......................................................ix
Chapter
1. INTRODUCTION..............................................1
1.1 METAL-AIR CELLS...........................................2
1.2 LI-AIR CELLS..............................................2
1.2.1 LI/02 CELL REACTIONS....................................2
1.2.2 ADVANTAGES OF LI/AIR CELLS..............................3
1.2.3 PREVIOUS STUDIES........................................4
1.2.4 PRESENT LIMITATIONS.....................................5
2. EXPERIMENTAL..............................................6
2.1 CATHODE MAKING PROCESS....................................6
2.1.1 BELLCORE METHOD.........................................6
2.1.2 DRY PASTE.............................................. 7
2.1.3 WET PASTE...............................................7
2.1.4 CURING..................................................8
2.2 CATALYST..................................................8
2.2.1 COBALT PHYTHALOCYANINE CATLAYST.........................8
2.2.2 PLATINUM METAL CATALYST.................................9
2.3 ELECTROLYTE FORMULATION..................................10
2.4 ANODE PREPARATION........................................10
2.5 ALUMINUM MESH PREPARATION................................12
2.6 SEPARATOR................................................12
2.7 PACKAGING/MEMBRANE MATERIAL..............................13
2.8 CELL FORMULATON/ACTIVATION...............................13
2.9 CELL TESTING.............................................14
3. RESULTS..................................................17
3.1 ALUMINUM GRIDS...........................................17
3.2 ELECTRODES.............................................. 17
3.2.1 CURING.................................................19
3.3 CATALYSTS................................................19
3.4 SEPARATOR................................................21
3.5 RATES....................................................21
3.6 ELECTROLYTE STUDIES..................................... 23
3.6.1 SALT EFFECT............................................26
3.7 PACKAGING/MEMBRANE.......................................27
4. DISCUSSION...............................................29
4.1 FUTURE WORK........................................... 31
4.1.1 TEMPERATURE............................................32
4.1.2 SAFETY.................................................32
4.1.3 CALENDAR LIFE......................................... 32
4.1.4 MODULAR DESIGN.........................................33
vi


Bibliography


LIST OF FIGURES
Figure
1.1 A TYPICAL METAL-AIR CELL.................................2
1.2 SCHEMATIC OF A TYPICAL LI/02 CELL........................3
2.1 THE STRUCTURE OF COPC....................................9
2.2 THE STRUCTURE OF COTMPP..................................9
2.3 A LISTING OF SOLVENTS AND SALTS TESTED FOR LITHIUM/AIR CELLS .... 11
2.4 A TYPICAL LI/02 CELL....................................14
2.5 THE HARDWARE USED TO SECURE AND APPLY MILD PRESSURE
TO THE CELL............................................15
2.6 FINAL ASSEMBLY OF THE LI/02 CELL........................15
3.1 SURFACE AREA OF SEVERAL CARBON BLACKS...................18
3.2 THE DISCHARGE CAPACITY OF A NUMBER OF CARBON BLACKS IN A
1 M LIPF6 EC/PC SOLVENT.................................19
3.3 THE EFFECT OF THE PORYPHIN CATALYST (COPC) ON THE KETJEN
BLACK SYSTEM...........................................20
3.4 COMPARISON BETWEEN A SUPER-P PT CATALYZED ELECTRODE AND
A BASELINE ELECTRODE...................................21
3.5 SPECIFIC CAPACITY OF KETJEN BLACK ELECTRODES............22
3.6 CARBON UTILIZATION OF SUPER-P USING THREE DIFFERENT
SOLVENTS.............................................. 24
3.7 DISCHARGE OF THREE DIFFERET CARBON TYPES IN A 1 M LIPF6
EC/PC SOLUTION........................................ 25
3.8 COMPARISON BETWEEN DIFFERENT SOLVENT SYSTEMS............26
viii


LIST OF TABLES
Table
1.1 THE REACTIONS OF 02 WITH LITHIUM AND THE THEORETICAL SPECIFIC
ENERGIES................................................4
2.1 VARIOUS BAGS USED FROM THE CRYOVAC COMPANY ALONG WITH
THE OXYGEN AND WATER VAPOR TRANSMISSION RATES..........13
3.1 COMPARISON OF THE RATE CAPABILITY OF THE TWO PRIMARY
CARBONS EVALUATED .....................................23
4.1 THE CALCULATED ENERGY DENSITY OF A TYPICAL SUPER-P CELL
USING TEG OR NMP AS THE ELECTROLYTE SOLVENT............30
IX


CHAPTER 1
INTRODUCTION
The world is drawing down its oil reserves at an unprecedented rate, with supplies likely to be
constrained by global production capacity by 2010As non-renewable resources like fossil fuels are
depleted around the world, it is necessary to develop new energy conversion devices. Lightweight, low
cost, safe and reliable power sources are needed to provide power for several areas of industry.
Specific energies of 700 Wh/kg have been achieved with Li/02 cells, while over 1000 Wh/kg is viable.
The potential for high energy Li/air cells is significant when considering other types of energy sources.
The majority of modem day batteries are inadequate for many high energy applications. Currently,
both primary and secondary batteries give 100 200 Wh/kg, while fuel cells approach 400 Wh/kg2.
Metal-air cells have already proven to be a viable energy producing device capable of much higher
energy densities.
Initially efforts are focused on implementing Li/air cells in man-portable battery packs for the
military. The technology must be adapted to meet the high energy challenge of unmanned-aerial
vehicles (UAVs)3. At present, power sources available to the military provide only marginally
adequate operating times for electrically-powered UAVs. Li/air cells will significantly extend the
flying times of UAVs.
The cells could provide low-rate (~0.1 mA/cm2), high energy power for several fields,
including; transportation, military, medicinal, electronic industries, and miscellaneous recreational
applications. The military relies on lightweight materials, which will allow the foot soldier to be agile
in his movement while not hindering the proliferation of electronic gear he carries. There are several
possible areas where Li/air cells could be applied in a military device, including: direct power for
tactical radio, forward field charging of rechargeable batteries, SATCOM radios, nightscope power,
guidance systems, surveillance, and sensors. Li/air batteries could provide a longer duration of steady
state power with fewer batteries than conventional systems employed in the military, such as Ni/Cd or
Ni/MH4. Li/air could serve as a backup power supply or as hearing aids for the medical field. With the
consistent development of power hungry electronic devices, there has been an ever-growing need for
1


higher energy density power sources such as Li/air. A few of the possible applications in the
electronics industry include: cellular phones, portable chargers for cellular phones, and personal digital
assistants (PDAs), camcorders, and notebook computers5.
Metal-Air Cells
A metal-air cell has a metal as the anode and 02 from the air as the cathode. Since the 02 from
the air is outside of the cell, its specific energy is high. During discharge of the cell the metal anode is
consumed during an oxidation process releasing free electrons that flow through an external circuit,
supplying a load, as Figure 1.1 shows. Once the metal anode is consumed the cell may be, in principle,
refueled by removing the consumed metal and electrolyte and replacing it with fresh electrolyte and
metal. The oxide products can be reprocessed into the original metal as well. The entire process is
therefore a closed cycle. Multiple cells may be combined to create the desired load for an application.
Although this technology has been proven to produce electricity in an environmentally sound
way, it has not yet been developed sufficiently to be a viable energy source for the general
public. Zn/air cells have been used widely in recent years as hearing aide batteries. There are also plans
for using Zn/air cells for vehicles used in mass transportation.
rctnlyle
Figure 1.1. A typical metal-air cell. During discharge the metal anode is oxidized and supplies
electrons through an external load. The air cathode absorbs oxygen were it is reduced by
metal ions dissolved in the electrolyte6.
2


Li/Air Cells
Li/O^ Cell Reactions
Figure 1.2 shows the typical schematic of a Li/02 cell. The cathode is a carbon/binder mix
pasted on a coated aluminum mesh. Oxygen from the atmosphere permeates the carbon cathode where
it is reduced. The cathode is separated from the anode by a polyolefin separator, which allows Li+ to
pass. The anode metallic lithium pressed to a copper or nickel mesh. The cell is wetted with
electrolyte, where the majority of the electrolyte is absorbed by the cathode. An oxygen permeable
plastic packaging material encloses the entire cell.
Oiygei PerneaMe Packagiig Membraie
Separator/Electrolyte
Ni current collector
2 (from air)
Carbon Electrode
Lithium
Figurel. 2 Schematic of a typical Li/02 cell. The upper light blue region represents the oxygen
permeable packaging membrane. The porous cathode is represented by the dark grey
region which serves as an absorbent for oxygen. The yellow area depicts the
polypropylene separator. The light grey region represents the lithium anode which is
pressed to either a nickel or copper mesh. The cell thickness on average is 5 mm.
Advantages of Li/02 Cells
Metal/oxygen batteries are unique in that one of the reactants, oxygen, is not stored within the
cell. Atmospheric oxygen reacts with Li+ transported by the electrolyte to form either Li20 or Li202 at
the cathode. Atmospheric oxygen diffuses through the cathode membrane by dissolving in the
electrolyte and, subsequently, gets reduced on the carbon electrode. Specifically, the reaction is
3


thought to occur on the surface of the carbon.
Metal/oxygen batteries have been developed based on Fe, Zn, Al, Mg, Ca, and Li. The open
circuit voltage, however, rarely approaches that of the Li/02 system (~3V). Most systems developed so
far have been based on alkaline electrolytes. Whereas most metals employed are relatively heavy with
high standard reduction potentials, lithium affords the highest energy density primarily due to its light
weight and low standard reduction potential.
The Li/02 aqueous electrolyte battery has been studied but suffers from corrosion of the
lithium electrode by water.7 Abraham and Jiang8,9 were the first to report a Li/02 organic electrolyte
battery. The battery was shown to have an open-circuit voltage close to 3 V, and an operating voltage
of 1.5 to 2.8 V. With cobalt phthalocyanine as a catalyst for the air electrode, there was good
coulombic efficiency upon recharge over several cycles. The Li20 and Li202 products formed are
insoluble in the organic electrolyte and cathode (Table 1.1).
Table 1.1 The reactions of 02 with lithium and the theoretical specific energies associated with the
reactions. The more efficient reaction involves the reaction of four moles of metallic
lithium with one mole of 02 to form two moles of lithium oxide. The less efficient reaction
entails two moles of metallic lithium reacting with one mole of oxygen to form one mole of
lithium peroxide.
Reaction Capacity, Ah Voltage, V Specific Energy, Wh/kg
Without 02 With 02
4 Li + 02 2 Li20 4 26.8 2.5-2.8 11,000 5,000
2 Li + 02 Li202 2 26.8 2.5-2.8 11,000 3,300
The reactions occur at a load voltage of 2.5 2.8V. The lithium oxide product is favored over lithium
peroxide because it is the more efficient product of the two. The precipitates plug the pores of the
carbon electrodes, which limits cell life. The discharge product at the surface of the air electrode limits
discharge capacity and rate capability by preventing access to the interior of the electrode10.
Previous Studies
The effects of air cathode formulation and electrolyte composition on discharge capacity, rate
capability, and rechargeability were previously studied.10 The air electrodes physical properties, its
4


thickness, porosity, and volume fraction of the carbon black are all significant factors in determining
cell performance. It was found that electrolyte formulation had the most dramatic influence on both
cell performance and on the nature of deposit formed.11 Particular electrolytes, however, preferred
specific carbons. The ability of an electrolyte to wet a particular carbon is an important factor in
determining cell capacity. It was shown that oxygen solubility in the electrolyte had a drastic effect on
cell performance10. Low oxygen solubility, in the form of the Bunsen coefficient (a), and the diffusion
constant calculated from the viscosity (q), have a direct impact on discharge capacity and rate
capability. Discharge capacity could be increased by increasing oxygen concentration in the
electrolyte, either by electrolyte reformulation or by increasing oxygen partial pressure. To increase
oxygen partial pressure, an oxygen atmosphere was provided in the form of a pressurized oxygen bag.
The battery was therefore designated a Li/02 battery as opposed to a Li/air battery.
Present Limitations
Although the potential exists for a high specific energy battery, lithium/air cells have many
barriers which must be overcome.
1) Lithium is intrinsically very reactive. If not protected, it could potentially react
with nitrogen or water. This requires that the cell be protected, if not isolated,
from all potential reactants besides oxygen.
2) Additionally, provisions must be made to allow for a high activity of oxygen at
the cathode surface.
3) In order to utilize the potential of lithium to its fullest, the weight of cell
components must be minimized.
4) Lithium is a safety hazard. It has the potential to explode or ignite.
5) Provisions must be made to allow for scaling-up of the processes.
6) Finally, the electrochemical characteristics of the cell must be optimized to
afford the highest energy density possible at an acceptable rate.
The purpose of this thesis was to develop a practical Li/02 cell by understanding the issues which limit
its capacity and rate.
5


CHAPTER 2
EXPERIMENTAL
The general experimental procedures, materials treatment, and cell construction were as
follows. Air sensitive experiments were carried out in an MBraun Corporation argon-filled dry box
maintained at <5 ppm H20.
Ketjenblack manufactured by Akzo Nobel Polymers, Super-P carbon black manufactured by
M.M.M. Carbon Belgium, Chevron Shawiningan Black acetylene black manufactured by Chevron
Phillips Chemical Company (SAB), Black Pearls 2000 manufactured by Cabot Corporation, ASI, or
Norit manufactured by Norit Technologies were evaluated as possible carbon materials. If the carbon
purity was in question, it was washed by stirring in an excess of acetone. Following the wash, the
carbons were filtered and dried at 80 C under vacuum overnight.
Cathode Making Process
Bellcore Method
A modified Bellcore method12 was used at the outset in order to construct porous electrodes.
The Bellcore process involves making a carbon slurry in a solution of PVdF (Kynar 2801, Atofina) in
acetone. Dibutyl phythalate (DBP) (purchased from Aldrich and used as received) was added as a
plasticizer. The paste was blended at low speed for 10 minutes. The paste was cast on a glass substrate
at a desired thickness (typically 2 mil 200 mil) and allowed to dry slowly through evaporation to
prevent film warping. The film was cast at half the thickness of the final cathode, since two halves
were used for a final cathode.
A treated aluminum grid was used as a current collector for the cathode. The current collector
was sandwiched between two cathode films. The films, held between a nonstick mylar sheet, were
pressed to the grid by way of a double-roll, three stage heated laboratory laminator. Lamination was
6


typically accomplished at 130 C using light pressure.
To make the final porous electrode, it was necessary to extract the DBP. A number of
solvents can be used (diethyl ether, hexane, ethanol, methanol, petroleum ether, and cyclohexane) to
accomplish the extraction. Diethyl ether was preferred due to its fast, complete extraction and its rapid
evaporation rate. Depending on the laminate/solvent ratio, laminate thickness, and extraction
temperature, two or three 5-15 minute extraction steps were sufficient to remove the DBP from the
laminate. The porosity of the final electrode can be controlled, to some extent, by adjusting the DBP
percentage. Heat and pressure of the laminator is also crucial in the final electrode porosity.
Dry Paste
Dry paste cathodes were prepared following the procedure of Read10. Carbon air cathodes
were prepared by mixing a certain wt %, typically 20 %, polytetrafluoroethylene, PTFE (Dupont,
Teflon PTFE 30), emulsion (0.05 to 0.5 pm PTFE particles in water) with a selected carbon. The
carbon was wet with a 1:2 ethanol/water (v/v) mixture and mixed with a spatula. After the slurry was
homogenous, it was placed under vacuum at 80C to dry. The dried paste was ground in a blender to
form a fine powder. It was then dry pasted onto preweighed treated A1 or Cu grids. Cathodes were cold
pressed (between 2000 8000 psi) for 2 minutes. In addition, other cathodes were also hot pressed
(250 300C) at similar pressures and times to improve electrode integrity.
Wet Paste
Wet paste cathodes were prepared following the procedure of Read11. Carbon air cathodes
were prepared by mixing a certain wt %, typically 20 %, polytetrafluoroethylene, PTFE, emulsion
(0.05 to 0.5 pm PTFE particles in water) with either Super-P, Chevron, Cabot, or Norit. The carbons
were wet with a 1:2 ethanol/water (v/v) mixture before the PTFE emulsion was added and the paste
was mixed. Two 6 mil metal shims with 100 cm2 (there were also two shims with 10 cm2 for
preliminary cells) openings were used to hold the aluminum grid down for pasting. A stainless steel
spatula was used to pull the paste across the grid using the shim as a thickness gauge. The assembly
was flipped over and more paste was added to the opposing side. A constant thickness was therefore
achieved on both sides of the aluminum grid. The shims were removed and the excess solvent removed
7


with blotting paper and mild pressure. The pasted electrodes were then dried at 80C in a vacuum oven
to completely remove the solvent. Drying the electrodes with minimal cracking was accomplished by
keeping the electrodes between blotting paper with mild pressure while in the vacuum oven. Cathodes
varied in thickness according to the final weight desired for the cathode.
Curing
Curing of the cathode was done to improve electrode integrity. The cathode before curing was
very porous and lacked particle to particle cohesion. By curing the cathode, a cohesive cathode could
be obtained while maintaining the electrode porosity.
Curing of the cathode was achieved by heating the polymeric binder (PTFE) near its softening
temperature (300 C) while applying pressure. This was achieved by placing the cathode between two
stainless steel plates and bringing the apparatus to temperature in a furnace. The plates were pressed at
a moderate pressure (-4000 psi) for 2 minutes. The cathode was placed between aluminum shims prior
to adding heat and pressure to allow the cathode to be released after curing. Alternatively, cells were
pressed without heat (cold pressed).
Catalyst
Cobalt phthalocyanine (CoPC, Figure 2.1) and cobalt (tetramethoxy phenyl, Figure 2.2)
porphyrin (CoTMPP) were purchased from Alfa Aesar or Aldrich and used as received. Hydrogen-
hexachloroplatinate was purchased from Alfa Aesar and used as received.
Cobalt Phthalocyanine Catalyst
Carbon containing a cobalt porphyrin catalyst was prepared as follows: 0.25 g of cobalt
phthalocyanine or cobalt (tetramethoxy phenyl) porphyrin (CoTMPP) (0.1 % 5% relative to carbon)
was dissolved in 30 ml of concentrated [30 weight percent (w/o)] sulfuric acid. The viscous liquid was
poured onto 10 g of carbon forming a wet paste. The paste was added to distilled water, allowing the
CoPC or CoTMPP to precipitate and deposit on the carbon matrix. The suspension was allowed to stir
for 30 minutes to allow for deposition. Distilled water was added and decanted repeatedly until the
8


carbon slurry was brought to a neutral pH. After the slurry was deemed neutral, the carbon slurry was
filtered and dried under vacuum at 80 C overnight.
Figure 2.1 The structure of CoPc. CoPC is
composed of a porphyrin ring surrounding a
cobalt center.
Platinum Metal Catalyst
There are several methods for the preparation of a dispersed Pt catalyst. In the most common
colloidal method12'13'14'15,1617, the reagent containing the metal, namely Pt in the form of hydrogen-
hexachloroplatinate, is reduced in a way that produces platinum particles in the colloidal form which
deposit on a carbon support. This adsorption occurs alongside that of by-products which must be
removed by heat treatments. In another preparation, the support is impregnated with the compound
containing the platinum, filtered, dried, and then the hydrogen-hexachloroplatinate is reduced to Pt
metal using a stream of hydrogen at high temperatures18,920,21.
In addition to these two methods, another adsorption method was used. Hydrogen-
hexachloroplatinate was reduced with formic acid according to the procedure set forth by Gonzalez22.
The supported Pt catalysts (Pt/C, 5 wt.%) were prepared by precipitating the Pt particles onto the
carbon surface through the reduction of hydrogen-hexachloroplatinate with formic acid. An
appropriated mass of the carbon powder substrate was suspended in a 0.1 M formic acid solution. This
suspension was heated to 80 C and a solution of hydrogen-hexachloroplatinate is added in three
9


stages. The new addition was made only after there was a negative test with potassium iodide for the
presence of hydrogen-hexachloroplatinate in a side reaction vessel. The suspension was left to cool to
room temperature, and then the solid filtered and dried in an oven at 80 C for 1 hr.
Electrolyte Formulation
Propylene carbonate (PC), ethylene carbonate (EC), 1,2-dimethoxyethane (DME), diethyl
carbonate (DEC), dimethyl carbonate (DMC), ethyl methyl carbonate (EMC), y-butyrolactone (gBL),
N-methyl pyrrolidinone (NMP), N-dodecyl pyrrolidinone (DP), N-octyl pyrrolidinone (OP),
tetra(ethylene glycol) dimethyl ether(TEG) were purchased from either the Ferro corporation or
Aldrich and used as received. Lithium hexafluorophosphate purchased from Ferro Chemicals, was
used to prepare the majority of electrolyte solutions Figure 2.3.
Anode Preparation
The anode for the prototype cells was made by sandwiching a small piece (~ 2 cm2) of copper
mesh between two layers of 7 mil 210 cm2 metallic lithium received from the Ferro corporation (Ferro
Chemicals, Cleveland, OH). A copper tab was attached to the copper mesh by an ultrasonic weld prior
to the addition of the lithium so as to allow an attachment for the current lead in the final cell. The two
pieces of lithium were mechanically rolled together around the copper mesh using a polyethylene
roller.
10


Solvent Tested Structure Boiling Point, C
DEC Diethyl Carbonate 4a 127
EMC Ethyl Methyl Carbonate 109
* DME 1,2-Dimethoxy Ethane 0 r1- 85
DMC Dimethyl Carbonate 90
EC Ethylene Carbonate 0 248
PC Propylene Carbonate A H 241
y-BL y-Butyro Lactone 0 202
NMP N-Methyl-2-Pyrrolidone 1 Cr 202
TEG Tetra (ethylene glycol) dimethyl ether 275
OP N-Octyl Pyrrolidone 307
DP N-Dodecyt Pyrrolidone <4 Salts Tested LiPF6, UCF3SO3, LiBF4 145 (0.3 mm Hg)
Figure 2.3 A listing of solvents and salts tested for Li/air cells. In addition, the boiling points of the
various solvents are listed.
11


The copper mesh and current collector were left as small as possible and only added to allow for an
external lead; the metallic lithium served the dual purpose of carrying the cell current and acting as the
anode.
Aluminum Mesh Preparation
To ensure good electrical contact between the electrode material and current collector, the
metal grid must be treated.
Annealed A1 grids (All Foils) must be etched to remove the layer of oxide on the surface.
Following etching, the metal must be coated with an electrically conducting coating. Grid treatment
steps are as follows:
1. Etching: The A1 grid is etched in a 1 M NaOH solution in water, for a short time at room
temperature.
2. Rinse: The A1 grid is rinsed thoroughly in deionized water and acetone.
3. Coating: The A1 grid is coated with a carbon black emulsion (ADCOTE 50C12), which
allows for adhesion and high electrical conduction between the electrode materials and
metal grid.
ADCOTE 50C12 is an alcohol-based suspension of a copolymer of polyethylene with acrylic acid
routinely used as a surface bonding treatment for many classes of dissimilar materials. The stock
ADCOTE 50C12 along with conductive carbon were ball milled for 1 hour at room temperature in a
laboratory ball mill. The viscous suspension was diluted 1:1 with ethanol and well stirred. The material
was sprayed on both sides of the A1 grid using an air gun. After drying, a smooth, black, conductive
layer was formed on the metal surface.
Separator
A single-ply or multi-ply separator either coated or uncoated with a PVdF copolymer
manufactured by Celgard Microporous membranes (Charlotte, North Carolina) was used as separator
for the prototype Li/02 cells. The type of separator was chosen based on its ability to be wetted by the
12


electrolyte. For the EC/PC electrolyte, a Celgard 3401 or 3501 separator was employed since these
were the only two separators wetted by the electrolyte.
For preliminary studies, a glass microfiber separator manufactured by the Whatman
Corporation was used to optimize the cells discharge characteristics. A glass microfiber allows the
cell to hold an excess of electrolyte which allows the cell to be moist during discharge.
Packaging/Membrane Material
The packaging/membrane material was provided by the Cryovac Sealed Air Corporation
(Duncan, SC). The bags were chosen due to their light weight, mechanical strength, quick heat sealing
capability, and selective permeability.
Table 2.1 shows the various bags used from the Cryovac corporation and their respective oxygen and
water transmission rates. The bag was chosen based on their oxygen transmission rate. Ideally, one
uses a membrane which allows desirable quantities of 02 to pass without the passage of any moisture
at all. A high oxygen transmission rate, allowed for a greater number of auxiliary reactants to enter the
cell, such as water vapor and nitrogen. Many times a higher oxygen transmission rate bag was
necessary due to the low rate capability of the cell.
Table 2.1 Various bags used from the Cryovac company along with the oxygen and water vapor
transmission rate.
Oxygen Transmission Rate (cc @ 73F, 1 atm., m2,24 hrs) Water Vapor Transmission (gm @ 100UF, 100% RH, 100 in.2, 24 hrs)
E 2300 5000 Not Given
Super L 3000 0.65
SES 340 13,800 2.6
HP 2700 >10,000 Not Given
Cell Formulation/Activation
A prototype cell was constructed by stacking electrodes into a bicell configuration. An anode
was flanked on both sides by a cathode. To begin, electrodes, separator, packaging material and
electrolyte were prepared as described above and transferred into the glove box. Electrodes and
13


separator were stacked following the configuration in Figure 1.2. Cathodes were wet with electrolyte
using a laboratory dropper until there was no further uptake. The separator was positioned on top the
moist cathode and allowed to draw up any surface electrolyte atop the cathode. Caution was taken
throughout the process not to allow the cathode and anode to come into contact shorting the cell. The
packaging material enclosed the cell and was sealed on all sides using a thermal impulse sealer. The
tabs extending out of the cell were reinforced with a thicker packaging material.
Cell Testing
Each Li/02 cell (Figure 2.4) was composed of a single bicell and compressed using a cell
holder and fasteners.
Figure 2.4 A typical Li/02 cell. The cell is composed of two cathodes surrounding one anode to form
a bicell. One aluminum and one copper tab extend outside the packaged cell to carry the
current. The blue packaging material surrounding the tabs is used for strength and
structural support. The translucent material surrounding the electrodes serves as both a
packaging material and a semi-permeable membrane.
The carbon Kevlar and epoxy resins were purchased from Ashland Chemical (Pueblo, CO) and used as
received. The carbon Kevlar was stacked in layers and formed into plates using the resin. After drying
of the resin, plates were formed that were both strong and rigid. The cell was compressed by the
carbon Kevlar cell holders fastened by nylon nuts and bolts (Figure 2.5). The material was fabricated
14


to expose the majority of the cathode surface while still applying mild pressure. The entire cell
assembly was lightweight, rigid and functional (Figure 2.6).
An oxygen atmosphere was achieved by using a modified, large sealable plastic bag. One
comer of the bag was cut perpendicular to the comer.
Figure 2.5 The hardware used to secure and apply mild pressure to the cell. The framework of the
holder was made of resin-bound carbon Kevlar which is high in tensile strength and light
in weight. The interior section of the holder was milled out so as to allow for oxygen
permeation. The framework was held together by nylon nuts and bolts.
15


Figure 2.6 Final assembly of the Li/02 cell. The cell is contained within the framework while still
having the majority of the cell exposed to the atmosphere.
Two pieces of thin (1 mil) copper foil (4 x 'A in.) were used as leads to span from the interior of the
bag to the exterior. A small strip of Surlyn (DuPont corporation, Wilmington, DE) was applied by a
thermoseal to the center of each of the leads for a mounting point to the large plastic bag. The two
copper leads were sealed to the comer of the bag using a thermoseal process. Finally, two wires with
alligator clips on each end were used as a means of connecting the cell to the bags interior leads.
The cell was placed inside the bag and connected to the leads, noting polarity. The bag was
inflated with industrial grade oxygen and sealed. Read10 showed the oxygen consumption at low rates
(0.05 mAh/cm2) to be near 3 mAh/mL 02. Bags capable of providing at least 3 times the necessary
oxygen were used. The exterior leads were connected to the Arbin cycler (Arbin Instruments, College
Station, TX) and discharged following the appropriate schedule. Cells were typically discharged at a
constant current of 0.05 0.20 mA/cm2 to 1,5 V. The cycler allows multiple cells to be discharged at
various currents following a predetermined schedule.
16


CHAPTER 3
RESULTS
As a matter of convention, the capacity of the lithium/oxygen discharge reaction for any
particular cell has been normalized to the weight of the carbon in the air cathode.
Aluminum Grids
Treated grids exhibit improved adhesion to electrode materials. Electrodes therefore exhibit a
lower resistance; in turn, the overall cell resistance is lowered as well. No deleterious effects of the
treatment on battery performance was realized, such as, capacity fading. Treatment of the aluminum
grid was deemed successful/unsuccessful by determining the AC impedance across a standard length.
Copper mesh was used as a reference since it has a very low internal resistance. Coating of the A1 grid
was determined successful as long as the AC impedance across a certain length of material was no
more than double that of the copper grid.
Electrodes
It was determined that with the Bellcore method it was very difficult to make electrodes with
high carbon loading or thick electrodes using this process. Excess PVdF is necessary to bind the high
surface area carbons to the electrodes, which lowers the specific energy of the electrodes. Despite this
obvious disadvantage, we have used this process to screen out potential carbon electrodes.
The PTFE wet paste method, which is a variation of the dry paste method, is the most widely
used and conventional process to fabricate porous carbon electrodes. The technique is mainly used for
oxyhalide batteries such as Li/SOCl2, Li/S02Cl2 and Zn/air batteries. Although it is a straightforward
method, controlling porosity and fabrication of a crack-free electrode is an art and varies strongly with
the surface area of the respective carbons. The greatest advantage of the process, however, is the ability
17


to prepare electrodes with high carbon content and any thickness. Since our goal is to develop very
high specific energy batteries, our preferred process was a wet paste, PTFE-based process or a
variation thereof.
In preliminary trials, a 1 M solution of LiPF6 in EC/PC was used as the electrolyte to test for
the most formidable cathode material. The EC/PC solvent system was chosen due to its relatively low
vapor pressure and high boiling point compared to other organic electrolytes. LiPF6 was chosen as the
salt because of its stability and high ionic conductivity. Carbons with high surface areas (Figure 3.1)
were chosen since it is believed that oxygen reduction occurs on the surface of the carbon active
material. The reduction process is more efficient with a high surface area carbon.
Tvce of Carbon Black Surface Area. m2/a
SuperP 60
Chevron 40
Ketjenblack 1200
Norit 1800
ASI
Figure 3.1 Surface area of several carbon blacks. ASI was not listed as it is unknown. Ketjenblack
and Norit have the have an extremely high surface area compared to Super-P and Chevron.
It was found that in the chosen electrolyte (1M LiPF6 : EC/PC) Ketjenblack allowed for the
highest capacity (Figure 3.2). It is thought, however, that capacity is more dependent upon other
factors. Oxygen must dissolve in electrolyte absorbed by the cathode. The ability of electrolyte to wet
the carbon is thought to be an important factor in carbon utilization. The porosity of the cathode is also
thought to be important. Electrolyte/carbon matching is therefore crucial in obtaining the highest
possible carbon utilization.
18


Figure 3.2 The discharge capacity of a number of carbon blacks in a 1 M LiPF6 EC/PC solution.
Ketjenblack gives the highest discharge capacity, while the Super-P and Chevron carbons
give inferior performance in this solvent.
It was determined that, although a Ketjenblack cathode allowed for a very high capacity (up
to 2000 mAh/g) at an adequate rate (0.2 mA/cm2), it was not desirable for this cell. The electrolyte
uptake for this material was excessive and caused the weight of the final cell to exceed acceptable
parameters. The specific energy of the final cell was therefore inferior to prior expectations. For this
reason, alternate electrodes received the main focus thereafter. In particular, Super-P was chosen since
its electrolyte uptake was less than half of Ketjenblack.
Curing
There was little difference between hot and cold pressed electrodes. Although electrodes gave
a lower resistance when hot pressed, a cold pressed electrode could achieve a similar resistance by
increasing the pressure slightly.
19


Catalysts
There were two catalysts which showed promise but only with Ketjenblack as the carbon.
Cobalt phthalocyanine (CoPC) and cobalt (tetramethoxy phenyl) porphyrin (CoTmPP) proved to
enhance the rate capability of the Ketjenblack system while raising the voltage. As Figure 3.3 shows,
the voltage was raised from 2.6 V to 2. 8 V while the capacity was increased (~ 10%).
Figure 3.3 The effect of the poryphin catalyst (CoPC) on the Ketjenblack system. The catalyzed
system allows for both an increased load voltage as well as overall discharge capacity.
The metal catalyst (Pt) was found inactive with every type of carbon. This was a perplexing
result since these catalysts are routinely employed for use in fuel cells using similar cathodes. Figure
3.4 compares a catalyzed and uncatalyzed Super-P electrode.
20


0)
9)
£
- Carbon = Super-P
- Electrolyte = TEG LiPFs
- Rate = 0.2 mA/cm2
R Catalyst
Uncatalyzed
0 200 400 000 800 1000
Discharge Capacity, mAh/g
Figure 3.4 Comparison between a Super-P Pt catalyzed electrode and a baseline electrode.
Separator
For testing purposes, a glass microfiber (Whatman) was used as a separator between the
cathode and anode. Even though the Whatman separator allowed for the maximum capacity to be
achieved, it could not be used in a practical cell due to it being extremely absorbent and therefore
heavy. The cell would have excess weight due to excess electrolyte and an overweight separator. A
glass microfiber separator, however, does offer the advantage of providing a reservoir of electrolyte.
The cell can therefore give its maximum capacity at its maximum rate. This separator was therefore
used in the initial stages of testing and also for preliminary testing of cells.
For practical cells, a polypropylene separator (Celgard) was used. This separator, although not
allowing for a reservoir of electrolyte, provided a lightweight medium which absorbed only as much
electrolyte as necessary.
Rates
Figure 3.5 shows the discharge behavior of an air cathode in 1 M LiPF6 EC/PC electrolyte at
21


three different rates. Voltage is plotted vs. specific capacity relative to the Ketjenblack carbon weights.
Capacity, mAh/g
Figure 3.5 Specific capacity of Ketjenblack electrodes. Rates were examined with Bellcore style
electrodes. A 1 M LiPF6 50/50 EC/PC electrolyte was used for all studies.
The rate of a given cell was crucial in optimizing its capacity. As with any cell, too high a
current will not allow for the highest capacity. Conversely, having an insufficient current will prevent
the cell from giving its maximum capacity because oxygen permeation will exceed the rate and react
with Li, and electrolyte will evaporate. The cell will therefore fail prematurely (not cathode limiting)
due to electrolyte loss over time since this is a semi-closed system. The rate was chosen to optimize the
capacity according to each cell type. Further, rate was adjusted according to the mass of active cathode
material, excluding binder and current collector weights. For the 0.6 mg/cm2 loaded cathodes shown in
Figure 3.5, the current is 0.12 mA/cm2-0.48 mA/cm2. This corresponds to 0.2 mA/mg-0.8mA/mg of
Ketjenblack carbon, a very low rate compared to standard Li-ion cells.
A heavier loaded cathode increases the overall thickness of the electrode. A thick electrode
increases Li-ion polarization upon charge/discharge, which decreases rate capability. The 0.6 mg/cm2
Ketjenblack electrode is considered to be a lightly loaded cathode when compared to conventional Li-
22


ion cells, but it is still not capable of high rates. From Figure 3.5, the cell is able to discharge at
continuous rates of 0.2 mA/mg, and pulse rates of 0.4 mA/mg-0.8 mA/mg.
The packaging material/semi-permeable membrane exhibited large influence on rate
capability. A cell could range from a closed system to an open system. A cell would be deemed a
closed system if the membrane lets in no oxygen, whereas an open system is a cell which is exposed to
the external atmosphere without a barrier. Membranes which were semi-permeable to oxygen were
utilized in the majority of the systems employed. In some instances, however, an open system would
be employed. This system would be used when trying to determine the maximum rate capability or
when a semi-permeable membrane was insufficient in sustaining oxygen levels within the cell.
In addition to the membrane, there were several other factors which contributed to rate
capability. Carbon type was crucial in determining rate capability. Certain carbon types would sustain
high rates (Ketjenblack, 0.2 mA/cm2) while others (Super-P, 0.1 mA/cm2) would not be as capable. In
addition, catalyzed electrodes would often, but not always, sustain a higher rate than non-catalyzed
electrodes. In many instances, therefore, a non-catalyzed electrode was chosen over a catalyzed one.
Further, complicating the situation was the electrolyte. The electrolyte had to be matched to a given
cathode in order to attain the full potential of a given cell (Table 3.1).
Electrolyte Studies
A number of solvents and salts in a variety of concentrations were tested for their ability to
provide the highest capacity at a reasonable rate. Figure 2.3 shows solvents and salts tested for use
with Li/02 cells.
Table 3.1 Comparison of the rate capability of the two primary carbons evaluated. The high rate
capability of the catalyzed Ketjenblack allowed for a membrane semi-permeable to 02 to be
used. The Super-P material had to be open to the oxygen atmosphere due to the lesser rate
capability. The electrolyte system was chosen based on the carbon used.___________
Carbon Catalyst Electrolyte Membrane / Oxygen Permeability (cc/m2/24 hrs., 1 atm) Rate Capability (mAh/cm2)*
Ketjenblack CoPC EC/PC 1M LiPF6 E 2300 5000 0.2
Super-P None TEG or NMP 0.5M LiPF6 None 0.1
The rate capability was determined by the maximum current allowable for obtaining the optimum
capacity.
23


Electrolyte plays an important part in both rate capability and capacity. Since oxygen is an
external reactant, it must be capable of entering the cell by dissolving in the electrolyte. The electrolyte
proved to have the greatest effect on both capacity and the formation of precipitates which often times
led to cell failure. Different solvents exhibited different effects on various cathode materials. A 0.5 M
solution of lithium hexachlorophosphate (LiPF6) in tetraethylene glycol (TEG) showed the greatest
capacity with Super-P (SP) and Chevron, while a 1 M solution of LiPF6 in a 1:1 ethylene carbonate
(EC) / propylene carbonate (PC) electrolyte solution gave the highest capacity with Ketjenblack.
Figure 3.6 compares three different solvents used in electrolytes, TEG, NMP, and EC/PC, with their
respective carbon systems, while Figure 3.7 shows the discharge curves of three different carbons in an
EC/PC system.
Each system was unique in its own mode of failure. For low boiling solvents, DEC, DMC,
and to some extent NMP, the drying out of electrodes was the primary cause of cell failure.
Figure 3.6 Carbon utilization of Super-P using three different solvents. The three systems were
discharged at 0.1 mA/cm2 to 1.5 V. The EC/PC system exhibits very poor capacity with
Super-P as the cathode. The TEG system gives nearly 2000 mAh/g of carbon. In addition,
the discharge curve is smooth throughout. NMP gives the highest discharge capacity. The
discharge curve, however, is rough throughout most likely due to solvent loss and
precipitate formation.
24


Figure 3.7 Discharge of three different carbon types in a 1 M LiPF6 EC/PC solution. An EC/PC based
system showed its greatest performance using Ketjenblack as the cathode. Ketjenblack,
not only, exhibited the greatest capacity, but also, had an extremely high voltage at an
unsurpassed rate capability.
Alternatively, EC/PC based systems failed due to precipitate formation throughout the cell,
primarily on the cathode. The generation of precipitates choked the cell, not allowing for sufficient
oxygen to enter the cell. NMP exhibited both solvent loss in the cell and precipitate formation. In
addition, the metallic lithium on the anode side appeared tarnished, most likely from solvent attack.
TEG based electrolytes showed neither solvent loss nor precipitate formation, and cell failure was
thought to be from electrode limitations. A Super-P cell using NMP as the electrolyte gave the highest
capacity of any cell tested (-2400 mAh/g,Figure 3.6). The rate capability for these cells was not high
(0.1 mA/cm2), however, and the discharge extended over a six day period.
Due to the long discharge, the cell lost solvent over time and was dry in the end. In an effort to combat
solvent evaporation, N-octyl pyrrolidinone (OP) and N-dodecyl pyrrolidinone (DP) were used in the
place of NMP. It was thought that by keeping the functionality of the pyrrolidinone ring and extending
the aliphatic chain on the nitrogen group, one could retain capacity while preventing solvent loss due
to a lower vapor pressure of the solvent. As Figure 3.8 shows, however, both DP and OP were inferior
to an NMP system in terms of capacity.
Both DP and OP cells remained wet throughout their discharge. The cells were far inferior in
25


terms of capacity. The electrolyte viscosity was increased by using DP and OP, which probably did not
allow the cell to sustain the discharge rate due to a decrease in electrolyte conductivity.
Salt Effect
LiPF6 was chosen as the primary salt of choice due to its high ionic conductivity and stability
in organic solvents. Lithium, because of its high reactivity, has a narrow range of compatible salts and
solvents.
Figure 3.8 Comparison between different solvent systems. N-methyl pyrrolidinone was used as a
reference to N-octyl-pyrrolidinone and N-dodecyl-pyrrolidinone solvents. All cells used
Super-P electrodes and were discharged at a rate of 0.1 mA/cm2.
LiPF6 is stable over a wide electrochemical window, soluble in a number of organic solvents, highly
ionizable leading to a high conductivity in various nonaqueous solvent systems, and electrochemically
stable on both the anode and cathode.
The concentration of LiPF6 is usually important in sustaining rate capability in many lithium
cells. An insufficient concentration of LiPF6 often hampers rate due to a low concentration of Li+. Salt
concentration, however, did not greatly influence the rate of Li/02 cells. Presumably, the concentration
26


range used (0.5M 1M) was above the threshold level where salt concentration would start to
influence the rate of these cells. Further, the rate of oxygen permeation was thought to be the rate-
determining step far outweighing any influence salt concentration may have on the cell.
Furthermore, excess amounts of LiPF6 add weight to the final cell decreasing the overall energy
density. Ultimately, a lower concentration (0.5M) was employed over the initial concentration (1M)
used in preliminary cells.
Packaging/Membrane
The ideal situation for this cell would allow for the transfer of 02 to cross the cathode dictated
by the discharge rate. In addition, there should be a barrier which prevents excess molecules from
entering the cell. An excess migration of 02 into the cell will force undesirable reactions of 02 with Li
to occur. Because of this, the rate of each cell must be matched to the oxygen permeability so as to
maximize capacity and rate capability.
Each system sustained a certain rate. Ketjenblack showed its optimum performance in
conjunction with a 1 M solution of LiPF6 in a 1:1 mixture ofEC/PC. Super-P exhibited its optimum
performance in a 0.5 M solution of LiPF6 in TEG or N-Methyl pyrrolidone (NMP). The Ketjenblack
system showed the greatest rate capability. Rates of 0.2 mAh/cm2 were sustained even with thick
cathodes (50 mg of Ketjen/cm2). The Super-P system was not capable of sustaining such a high rate.
Therefore, rates of 0.1 mAh/cm2 were standard for thick cathodes (50 mg of Super-P/cm2).
Due to the lower rate capability of the Super-P system, the packaging material had to be
punctured to allow for an enhanced oxygen permeability, as compared to the Ketjenblack system. This
enhancement was required for both TEG and NMP based electrolytes. This system, being directly
exposed to oxygen, posed an additional problem to sustaining the cells performance. It proved difficult
to keep solvent within the cell due to extended discharge period of the cell (extending anywhere from 5
- 10 days). NMP boils at 202 C and has a vapor pressure of 0.24 mm Hg/20 C. NMP evaporated
from the cell due to the long discharge periods, and the cell was thought to fail due to lack of solvent
within the cell rather than precipitates formed on the cathode. TEG, on the other hand, boils at 275 C
and has a vapor pressure less than 0.01 mm Hg/20 C, and this system was capable of keeping the cell
moist throughout its discharge.
The Ketjenblack system, especially the catalyzed electrode, showed higher rate capability
27


than that of the Super-P system. For this reason, a lower oxygen activity within the cell was acceptable.
A barrier to a direct oxygen atmosphere was therefore employed successfully. The packaging material
served this purpose while simultaneously serving as a structural support. A packaging material used in
the food processing industry was found to be optimum for this system. The plastic bag chosen was a
model E 2300 packaging material from the Cryovac corporation. This material was found to be
lightweight and sturdy while allowing oxygen permeation at a controlled rate (5000 cc/m2/24 hrs., 1
atm). The E 2300 packaging material controlled the rate of oxygen permeation and prevented the loss
of solvent. With the packaging material, however, the cell was capable of delivering its full capacity
and was thought to fail because of precipitations within the cell rather than loss of electrolyte.
28


CHAPTER 4
DISCUSSION
A cell delivering a capacity of 2,000 mAh/g of carbon, which has been achieved, gives a
specific energy of 700 Wh/kg. By assuming 2,500 mAh/g, which should be attainable, a cell exceeding
900 Wh/kg should be realized. There are, however, several details to consider and problems to remedy.
Advancements in electrode formulation, including a suitable carbon supported catalyst, are needed for
an optimum cathode. In conjunction with an optimized cathode, an electrolyte is needed which is
capable of working over a wide temperature range. Increases in energy density can be realized by
understanding the weight ratios of cell components and adjusting them accordingly.
Table 4.1 shows the weight percentage of each component in the cell. It was quickly realized
that electrolyte uptake was of primary concern due to it being more than half the weight of the final
cell.
The majority of cell weight is due to the electrolyte contained in the porous cathode. Since
electrolyte is necessary for cell function, the only way to lower this weight percentage is to find a
carbon material with a low electrolyte uptake. Super-P was found to have an electrolyte/carbon weight
ratio of 3.5 as opposed to that of Ketjenblack which had a ratio near 8. Super-P was therefore chosen
over Ketjenblack, even though it was inferior in rate capability. If Super-P is to be used in the future,
the rate capability must be improved. Alternatively, another high capacity carbon should be found
which will allow for a high rate capability.
29


Table 4.1 The calculated energy density of a typical Super-P cell using TEG or NMP as the
electrolyte solvent. An 80/20 ratio of carbon to binder was used as the cathode
composition. The amount of lithium was chosen so as to allow the cell a 10% excess
relative to the overall cell capacity. The membrane weight was not applicable in this
system since the packaging material served as both a packaging material and a semi-
permeable membrane.
- Assume 2000 mAh/g utilization per g of Carbon
W eight, g % \Y eigh t
Carbon 2.50 15
Binder 0.40 2
AI Grid 0.46 3
Lithium 1.42 8
Cn arid 0.70 4
Carboa l tilizatioagAh/#!..
Electrolvte/Carbon (wt ) 3.5
Electrolyte Cathode 8.75 51
Electrolyte Separator 0.70 4
Cclgard 0.22 1
M ea bran e 0.00 0
Packaging 2.00 12
T abs 0.00 0
Total 17.15

Cell Capacity. Ah 5
Voltage, V 2.6
Specific Eitrir.Wi/kt 758
Assumptions
80% carbon
20% binder
10% extra
- 2500 mAh/g utilization leads to > 900 Wh/kg!
The carbon/electrolyte interaction should be studied extensively to find the maximum specific
capacity. An unexpectedly high utilization was obtained from an NMP or TEG based electrolyte with
Super-P as the cathode. A TEG electrolyte gives around 2000 mAh/g and appears to be limited by the
cathode. NMP, on the other hand, appears to fail through solvent loss by evaporation. Similar, lower
vapor pressure solvents (OP and DP) were used for evaluation to extend cell capacity. Both OP and DP
have longer alkyl side chains than NMP. Neither solvent increased cell capacity, however. Rather, cell
capacity was reduced, probably due to an increase in electrolyte viscosity leading to a decrease in
conductivity.
Super-P has a relatively low surface area. It is thought that surface area contributes to the
overall carbon utilization, but that other factors are of greater importance; such as, porosity and
30


carbon/electrolyte interactions. Electrolyte is thought to fill the porosity of the carbon electrode. In
turn, capacity is thought to be dependent upon the porosity of the electrode. There is therefore an
interplay between cell weight (through electrolyte uptake), and cell capacity (carbon porosity).
Carbon/electrolyte interactions are thought to be crucial in improving cell capacity and rate capability.
Cells can have a two to threefold increase in capacity by simply changing the electrolyte. Modest
surface area carbons with minimal electrolyte uptake in combination with TEG or NMP can lead to
high specific energies.
The packaging/membrane material should be capable of preventing N2 and moisture to enter
the cell. As of now, the oxygen permeable membrane is a major barrier preventing the advancement of
this research. There is a limited availability of membranes, and they must be taken from other fields
and prior technology. The matching of rate capability with the membrane permeability is therefore
difficult.
The densification of the cathode was critical in achieving optimal performance. Applying
little or no pressure did not allow for sufficient contact between adjacent cathode particles. Exerting
excess pressure on the electrode caused the final cell to not exhibit any significant capacity, probably
because the electrode became brittle leading to insufficient contact between cathode particles. The
resistance of the electrode was therefore very high across the length of the electrode (~20 G/cm) for
insufficiently pressed electrodes. The resistance of a properly pressed electrode was 0.5 fl/cm. The
electrode resistance, in turn, typically increased the resistance of the final cell by a factor of three.
Future Work
A number of issues related to the development of a Li/Air battery have been identified,
including: temperature, safety, state of charge (SOC) determination, rechargeability and calendar life.
Temperature
The principal concern is performance of the cell at temperatures down to -30 C, but high
temperature performance (e.g. desert conditions) are also of interest. Future work will focus on making
additional prototype cells that can be used to demonstrate the performance of Li/Air cells at various
temperatures. A comparison of the discharge profiles to temperature will be taken into account. Tests
31


will be performed at -30 C, 0 C, room temperature (~25 C), and 50 C. Since electrolyte
conductivity is the single most dominant factor in determining the discharge behavior especially at low
temperatures, different electrolytes at will be used at different temperatures. For example, current cells
are constructed using a lithium hexafluorophosphate in TEG or NMP, which is especially suitable for
use at room temperature and above. The composition of the electrolyte at subambient temperatures,
however, will most likely need a less viscous solvent such as diethyl carbonate (DEC) in order to
maintain sufficient conductivity.
Safety
Lithium/oxygen cells can be thought of as unsafe due to the presence of metallic lithium.
Safety is particularly important when high energy batteries are employed. A broad range of hazards
should be studied to fully assess the safety aspects of the cell. Tests will be performed to characterize
the response of the cell to expected and worst-case scenarios:
1. Mechanical Abuse Tests includes mechanical shock, drop, penetration, rollover,
immersion, and crush tests. The outcome of these tests may dictate the type of packaging
and preferred orientation of the cells or modules.
2. Vibration Tests includes cyclical tests of varying magnitudes, which simulate the
exposure of cell components. These tests characterize the effect of long term vibration and
shock on the performance and service life of battery technologies.
Information gained from testing will be used to quality their safe construction and operation
and to identify design deficiencies. Upon initial testing, further research will be conducted to determine
means of improving cell safety.
Rechargeabilitv
The cycling ability of the cell will be tested. It may not be possible for the cell to charge
sufficiently. If the cell is capable of charging, the number of satisfactory cycles will be determined. In
addition, the electrolyte formulation will be adjusted to allow for cell cycling if possible.
32


Calendar Life
Prototype cells will be built to be used in demonstrating the effect of extended storage of the
Li/Air cells at room temperature on capacity. Tests will be performed to measure capacity loss when
the battery is not used for an extended period of time.
Upon completing the initial tests, further research will be conducted to determine how and
where improvements can be made, such as sealing materials used to activate and deactivate the cell.
Finally, cells will be constructed that can serve as a baseline to further research efforts.
Modular Design
Future work will focus on the development of a modular system. The single-cell system will
have to be extended to a multi-cell system. The modular system will have to allow for the stacking of
single cells to attain the desired final voltage and capacity. The stacking of cells will require the cells
be separated to allow for oxygen permeation between the layers. A successful final design will allow
for oxygen to reach all areas of every cathode with sufficient activity. Also, the design will have to
accommodate the interior areas of the cell. It will require the separation of bicell layers to allow for the
passage of oxygen. Separation could be achieved by placing a lightweight plastic mesh between the
bicells. The mesh will have to be of sufficient thickness to allow for proper separation.
The final cathode will also have to be catalyzed to increase the activity of oxygen. By
catalyzing the cathode, not only will the rate of the cell be accelerated, but the requirement for oxygen
permeation throughout the final cell will also be lessened.
A successful cell would allow the battery to be sealed with a pealable tape, resembling that of
a Polaroid picture. Prior to use, the tape would be removed activating the battery to let oxygen into the
carbon electrode. Ideally, the battery could be resealed after each use or it could be recharged and used.
A successful final design will also require all materials to be lightweight and resistant to the
semi-corrosive electrolyte. In addition, allowances should be made for voltage sense at the modular
level.
33


REFERENCES
1 Douglas-Westwood Ltd., an energy industry consulting firm based in Canterbury, England.
2 Fuel Cell Handbook (fifth edition), EG&G Services, Parsons Inc., October 2000.
3 Putt, R., Electric Fuel Corporation, US CECOM RDEC
4 COTS journal, pp. 19, May 2003.
5 Koretz, B., 3300 mAh Zinc-Air Batteries for portable consumer products, Electric Fuel Unlimited,
6 Figure modified from the following website:
http://www.vu.union.edu/~sabourii/background info.htm
7 E.L., Littauer and K. C. Tsai, J. Electrochem. Soc., 124, 850 (1977).
8 K.M. Abraham and Z. Jiang, J. Electrochem. Soc., 143, 1 (1996).
9 K.M. Abraham and Z. Jiang, U.S. Pat. 5,510,209 (1996).
10 J. Read, J. Electrochem, Soc., 149 (9), A1190-A1195 (2002).
11 J. Read, J. Electrochem. Soc., 150 (10), A1351-A1356 (2003).
12 Carey, J., Some Characteristics of Kynar PowerFlex-LBG, a New High-Performance Binder for
Bellcore PLiON Batteries, December 1997.
13 H. G. Petrow, US Patent 3,992,331 (1976).
14 K. Aika, J. Inst. Catal., Hokkaido Univ., 24, 54 (1976).
15 V.M. Jalan, US Patent 4, 136, 059 (1979).
16 H. Bonnemann, Angew. Chem., 103, 1344 (1991).
17 N. Giordano, Electrochim. Acta, 38, 913 (1993).
18 P.L. Antonucci, J. of Applied Electrochem., 24, 58 (1994).
19 D. Richard, Preparation of Highly Dispersed, Carbon Supported, Platinum Catalysts, Elsevier
Science Publishers B.V., Amsterdam, 1987, 71-81.
20 T. Frelink, J. Electroanal. Chem., 382, 65 (1995).
21 K. Amine, J. Chem. Soc., Faraday Trans., 91, 4451 (1995).
22 S. Ye, J. Electroanal. Chem., 415, 115 (1996).
23 Gonzalez, E., J. of New Materials for Electrochem. Systems, 6, 1-8 (2003).
34