Modeling the Cost and Performance of

            Lithium-Ion Batteries for Electric-Drive

            Vehicles


            Final Report
&EPA
United States
Environmental Protection
Agency

-------
                  Modeling the Cost and Performance  of
                 Lithium-Ion Batteries  for Electric-Drive
                                        Vehicles

                                     Final Report
                                  Assessment and Standards Division
                                 Office of Transportation and Air Quality
                                 U.S. Environmental Protection Agency
                                       Prepared for EPA by
                                    Argonne National Laboratory
                               Chemical Sciences and Engineering Division
                                   Contract No. DE-AC02-06CH11357
                                      Submitted July 15, 2011
                   NOTICE

                   This technical report does not necessarily represent final EPA decisions or
                   positions. It is intended to present technical analysis of issues using data
                   that are currently available. The purpose in the release of such reports is to
                   facilitate the exchange of technical information and to inform the public of
                   technical developments.
&EPA
United States
Environmental Protection
Agency
EPA-420-R-12-023
August 2012

-------
                           TABLE OF CONTENTS







LIST OF FIGURES 	vii




LIST OF TABLES 	ix




ABBREVIATIONS 	x




LIST OF SYMBOLS 	xii




ACKNOWLEDGEMENTS 	xv




EXECUTIVE SUMMARY 	xvi




PREFACE TO THE FINAL REPORT	xvii




1. Introduction	1




2. Battery and Cell Design Format 	3




      2.1 Cell Design	4




      2.2 Module Design	5




      2.3 Battery Pack Design	6




3. Modeling of Battery Design and Performance 	9




      3.1 Criteria for Power, Energy, and Life 	9




      3.2 Voltage at Maximum Power	11




      3.3 Governing Equations 	17




      3.4 Calculation of the ASI 	18




            3.4.1 Current Collection Resistance	20




            3.4.2 Potential and Current Distribution	22




            3.4.3 Determination of Module Terminal Size	24




      3.5 Calculation of Battery Dimensions	25




            3.5.1 Cell Dimensions	25





                                       iii

-------
             3.5.2 Module Dimensions 	26




             3.5.3 Battery Pack Dimensions	26




       3.6 Additional Considerations 	26




             3.6.1 Maximum Electrode Thickness 	27




             3.6.2 Accounting for Parallel Cell Arrangements	33




             3.6.3 Accounting for Parallel Module Arrangements	33




4. Thermal Management	34




       4.1 Heat Generation Rates in the Battery Pack during Driving	34




       4.2 Heating under Adiabatic Conditions	35




       4.3 Active Cooling Systems	35




             4.3.1 Heat Transfer from Cell to Module Wall	36




             4.3.2 Heat Transfer from Module Wall to Flowing Coolant 	38




       4.4 Cooling and Heating Required to Maintain Pack Temperature	42




       4.5 Heat-up from Cold Ambient Conditions	43




5. Modeling of Battery Pack Manufacturing Cost	44




       5.1 Approach	44




       5.2 Materials Costs  and Purchased Items  	45




             5.2.1 Battery Specific Materials Cost 	45




             5.2.2 Purchased Items Cost	50




             5.2.3 Pack Integration Cost	50




       5.3 Baseline Manufacturing Plant	53




             5.3.1 Receiving and Shipping 	54




             5.3.2 Electrode Materials Preparation	56
                                          IV

-------
       5.3.3 Electrode Coating	56




       5.3.4 Calendering	57




       5.3.5 Inter-Process Materials Handling 	57




       5.3.6 Electrode Slitting 	58




       5.3.7 Final Electrode Drying	58




       5.3.8 Control Laboratory	58




       5.3.9 Cell Stacking	59




       5.3.10 Current Collector Welding	59




       5.3.11 Enclosing Cell in Container	59




       5.3.12 Electrolyte Filling and Cell Sealing	60




       5.3.13 Dry Room Management	60




       5.3.14 Formation Cycling, Final Cell Sealing, etc	60




       5.3.15 Module and Battery Assembly	61




       5.3.16 Rejected Cell and Scrap Recycle	62




       5.3.17 Baseline Plant Summary	63




5.4 Adjustment of Costs for Rates 	63




5.5 Plant Investment Costs	66




5.6 Unit Costs for Battery Pack 	66




       5.6.1 Variable Costs	66




       5.6.2 Fixed Expenses 	67




       5.6.3 Profits	68




       5.6.4 Battery Pack Warranty Costs 	68




5.7 Summary of Baseline Battery Cost	68

-------
6. Description of Spreadsheet Model and Instructions for Use 	71




       6.1 Background	71




       6.2 Instructions	71




             6.2.1 Enabling Calculation	71




             6.2.2 System Selection Worksheet	73




             6.2.3 Battery Design Worksheet	73




             6.2.4 Remaining Worksheets 	77




       6.3 Battery Design Format Requirements	80




       6.4 Troubleshooting and General Advice 	80




       6.5 Suggested Number of Cells, Modules, and Performance Inputs 	80




       6.6 Entering a New Material Couple	81




7. Illustrated Results	83




       7.1 Number of Cells in Series	83




       7.2 Cathode Materials	84




       7.3 Parallel-Connected Cell Groups and Electrode Thickness 	84




       7.4 Manufacturing Scale	86




8. Future Work	89




       8.1 Initial Power Designed at Differing Fractions of the Open-Circuit Voltage...89




       8.2 Optimum Battery Voltage for Minimum Drivetrain Cost 	89




       8.3 Multipurpose Battery Manufacturing Plants	90




       8.4 Stand-Alone Graphical User Interface for Model	90




References	91
                                           VI

-------
                                 LIST OF FIGURES

2.1    Prismatic cell and module design for battery packs	3

2.2    Cell sandwich inside of prismatic pouch cells	4

2.3    Coated current collector foil for prismatic electrodes	5

2.4    Hermetically-sealed module	6

2.5    Insulated battery jacket with enclosed modules that are cooled on their upper and lower
       surfaces by ethylene glycol-water solution	7

3.1.    Summary flow of the design model	10

3.2    a) Required change in [V/U] to maintain rated power with increases in internal
       resistance over the life of the battery, b) Increase in current due to lowered [V/U]	13

3.3    Change in heat rejection requirement from increases in resistance for batteries with
       different designed voltages at maximum power	14

3.4    Efficiencies for batteries designed to achieve maximum power at different fractions of
       their open-circuit voltage 	16

3.5    The change in current and potential within the positive and negative foils. The current
       collection design results in a uniform current distribution along the length of the foil ....23

3.6    Cell capacity simulated at the C/l and C/3 rate as a function of electrode thickness
       (loading) for NCA-Gr	29

3.7    Normalized electrolyte salt concentration at the end of discharge at the C/l and C/3
       discharge rates	29

3.8    Calculated ASI from a simulated 10-s, 5C discharge pulse for the NCA-Gr cell couple at
       60%SOC	30

3.9    The potential of the negative electrode versus a  hypothetical lithium reference electrode
       located in the center of separator during a 5C charge pulse for the NCA-Gr couple	31

4.1    Plot comparing the estimated resistance to heat transfer from the cell center to the
       cooled surface of the module to that calculated by the FlexPDE model 	38

4.2    Heat transfer from the module wall to the laminar flow heat transfer fluid. The
       temperature profile of the fluid is shown at different lengths down the path	39
                                           Vll

-------
4.3    Temperature profile in the heat transfer fluid for various fractions of the dimensionless
       path length	41

4.4    Correlation of model simulation results relating the Graetz number and mean Nusselt
       number for laminar flow between an insulated surface and the module casing	41

5.1    Metal ingot cost contribution to the current collector foils over a 20 year period	49

5.2    Baseline lithium-ion battery manufacturing plant schematic diagram 	54

5.3    Breakdown of installed capital equipment costs for the baseline plant	65

5.4    Breakdown of unit costs for baseline battery with total price to OEM of $2428 	70

6.1    Iteration must be enabled for the  spreadsheet model to function 	72

6.2    The specific cell chemistry for the battery design is selected on the System Selection
       worksheet	73

6.3    System Selection worksheet 	74

6.4    Top portion of Battery Design worksheet	75

6.5    Middle portion of Battery Design worksheet	76

6.6    Bottom portion of Battery Design worksheet	78

6.7    Summary of Results worksheet 	79

7.1    The effect of the number of cells for NMC441-Gr, 60-kW, PHEV25 packs with 10.7
       kWh total energy (70% useable)	83

7.2    Mass and volume of electric vehicle battery packs with lithium iron phosphate (LFP),
       lithium manganese-spinel (LMO) and lithium nickel-manganese-cobalt oxide (NMC441)
       positive electrodes versus graphite designed to deliver 150 kW of power at 360 V (25%
       SOC)	85

7.3    Battery pack price to OEM for LFP-Gr, LMO-Gr and NMC441-Gr battery packs for
       same designs as in Fig. 7.2. NMC441-Gr and LMO-Gr result in nearly the same price...85

7.4    Battery pack cost as a function of number of parallel cells and for different maximum
       electrode thicknesses	87

7.5    The effects of manufacturing rate on the price calculated by the model for battery packs
       of various cell chemistries, power capabilities and vehicle types 	88
                                          Vlll

-------
                                 LIST OF TABLES

3.1    Criteria for designing batteries for a specific end-use application	9

3.2    The effect of electrode loading on the price of a 17 kWh NCA-Gr PHEV40
       battery with  96 cells 	32

4.1    Sample calculations of composite thermal conductivities of cell structures
       across layer  and parallel to layers 	37

4.2    Range of parameter values for calculating heat transfer rates in FlexPDE model 	37

5.1    Details of stated costs for cathodes, anodes, electrolyte, and separator 	46

5.2    Cost equations for purchased items  	50

5.3    Costs to integrate battery pack into vehicle drivetrain	51

5.4    Summary table of the baseline plant	55

5.5    Materials yields during  electrode and cell fabrication	62

5.6    The effect of processing rate (R) on cost for various  scale factors	64

5.7    Battery pack manufacturing investment costs  	66

5.8    Unit cost of battery pack	67

5.9    Summary of results for  cost of baseline battery and that of similar batteries with
       double the power and double the capacity of the baseline battery	69

6.1    General suggestions for range of input parameters that change with battery type 	81
                                           IX

-------
                               ABBREVIATIONS




AST          area specific impedance




BOL         beginning of life




DMC        dimethyl carbonate




EC          ethylene carbonate




EMC        ethyl methyl carbonate




EOL         end of life




EV          electric vehicle




Gr           graphite




GSA         General, Sales, and Administration




HEV         hybrid electric vehicle




HEV-HP     high-power assist hybrid electric vehicle




LCO         lithium cobalt oxide




LFP          lithium iron phosphate




Li           lithium




Li-ion        lithium-ion




LMO        lithium manganese spinel




LMR        lithium and manganese rich




LTO         lithium titanate spinel




microHEV    micro or mild power assist hybrid electric vehicle




MW         molecular weight




NCA        lithium nickel cobalt aluminum oxide




NMC        lithium nickel manganese cobalt oxide

-------
NMP        N-Methyl-2-pyrrolidone




OCV        open-circuit voltage




OEM        original equipment manufacturer




PE           polyethylene




PET         polyethylene terephthalate




PHEV        plug-in hybrid electric vehicle




PP           polypropylene




R&D        research and development




SOC         state of charge




USGS        United States Geological Survey
                                         XI

-------
                                LIST OF SYMBOLS
Section 3, 4 and 6
a
ASI,
   energy
ASI,
   power
c
du


E


E


F


G


h


Hj


H/W
T  lor
him



In



'total



k
ratio of interfacial area to electrode volume, cm"1


area of the positive electrode, cm2


area of the terminal, cm2


area specific impedance for energy, ohm cm2


area specific impedance for power, ohm cm2


cell capacity, Ah.


parameter


specific heat capacity, J/g K


hydraulic radius, cm


total energy, Wh


energy usage rate, Wh/mile


Faraday constant, 96485.3  C/mol


mass flowrate, g/s


heat transfer coefficient, W/cm2 K


height of j, cm


aspect ratio of pouch cell


exchange current density related to the interfacial area, A/cm2


average current density, A/cm2


ionic limiting current density, A/cm2


local current density, A/cm2


total current density, A


thermal conductivity, W/cm K
                                           Xll

-------
Ij             length of j, cm




LJ            thickness of j, cm




m.j            mass of j,  g




n             parameter




Nj            number of j




[N/P]         negative to positive capacity ratio




P             battery power, W




Pbatt          maximum designed battery power, W




q             heating rate, W




Q            specific capacity of the electrode, mAh/g




rc            C-rate, h"1




rc,iim          limiting C-rate, h"1




TJ             radius of j, cm




R            universal gas constant, 8.3144 J/mol K




Rj            resistance of j, ohm




t             time, s




T             temperature, K




U0cv,p         open-circuit voltage at SOC for power, V




UOCV.E         open-circuit voltage at SOC for energy, V




v             square root of dimensionless exchange current




Vceii          cell voltage, V




[V/U]         fraction of the open-circuit voltage




Wj            width of j
                                            Xlll

-------
X




y
•^




a





P
Pj
Section 5.1

C
C0




Xi




MWi




MW




Section 5.4


C




C0




P




R





Ro
Cartesian coordinate, cm



compression factor




constant, ohm cm3




constant, ohm cm2




volume fraction of active material




first cycle efficiency




fluid viscosity, g/cm s




metal potential of foil k, V




density of j, g/cm3




conductivity of j, S/cm








final cost of lithiated oxide, $/kg




cost of raw material for component i, $/kg



baseline cost of lithiated oxide, $/kg



molar stoichiometry of component i



molecular weight of component i, g/mol



molecular weight of lithiated compound, g/mol






capital cost of an installed equipment for the designed battery, $



capital cost of an installed equipment for the baseline plant battery, $



power factor




designed battery processing rate for specific process step




baseline plant processing rate for specific process step
                                            XIV

-------
                            ACKNOWLEDGEMENTS

Support from the Vehicle Technologies Program, Hybrid and Electric Systems, initially under
Tien Duong and now David Howell, at  the U.S. Department of Energy,  Office of Energy
Efficiency and Renewable Energy, is gratefully acknowledged. The submitted manuscript has
been  created  by  UChicago Argonne,  LLC,   Operator  of Argonne  National  Laboratory
("Argonne"). Argonne, a U.S. Department of Energy Office of Science laboratory, is operated
under contract no.  DE-AC02-06CH11357. The U.S. Government retains for itself, and others
acting on its behalf, a paid-up nonexclusive, irrevocable worldwide license in said article to
reproduce, prepare  derivative works, distribute copies  to the public, and  perform publicly and
display publicly, by or on behalf of the Government. We especially thank Danilo Santini of
Argonne's Transportation R&D Center for his support and suggestions in carrying out this study.
Ralph Brodd reviewed our baseline  plant and  made  several suggestions which we  have
incorporated in the present design. Fritz Kalhammer and Haresh Kamath  of the Electric Power
Research Institute have reviewed our work over several years and made suggestions that resulted
in improvements. The work was done under the direction of Dennis Dees and Gary Henriksen of
Electrochemical Energy Storage who provided guidance in carrying out the work and preparing
this manuscript.
                                          XV

-------
                             EXECUTIVE SUMMARY

This report details the Battery Performance and Cost model  (BatPaC) developed at Argonne
National Laboratory for lithium-ion battery packs used in automotive transportation. The model
designs the battery for a specified power, energy, and type of vehicle battery. The cost of the
designed battery  is then calculated by accounting for every step in the lithium-ion battery
manufacturing process. The assumed annual production level directly affects each process  step.
The  total cost to the original equipment manufacturer calculated by the model includes the
materials, manufacturing, and warranty costs for a battery produced in the year 2020. At the time
this  report is written, this  calculation is the only publically available model that performs a
bottom-up lithium-ion battery design and cost calculation.

The  purpose of the report is to document the equations and assumptions from which the model
has been created.  A user of the model will be able to recreate the calculations and perhaps more
importantly, understand the driving forces for the results. Instructions for use and an illustration
of model results are also presented. Almost every variable in the calculation may be changed by
the user to represent a system different from the default values pre-entered into the program.

The  distinct advantage of using a bottom-up cost and design model is that the entire power-to-
energy space may be traversed to  examine the  correlation between performance and cost. The
BatPaC model accounts for the physical limitations of the electrochemical processes  within the
battery. Thus, unrealistic designs are penalized in energy density and cost, unlike cost models
based on linear extrapolations. Additionally, the consequences  on cost and energy density from
changes in cell capacity, parallel cell groups, and manufacturing capabilities are easily assessed
with the model. New proposed materials may also be examined  to translate bench-scale values to
the design of full-scale battery packs providing  realistic energy densities and prices to the
original equipment manufacturer.

The  model will be openly distributed to the public in  the year  2011. Currently, the calculations
are based in a Microsoft® Office Excel spreadsheet. Instructions are provided for use; however,
the format is admittedly not user-friendly. A parallel development effort has created an alternate
version based on a graphical user-interface that will be more intuitive to some users. The version
that is more user-friendly should allow for wider adoption of the model.
                                           XVI

-------
                      PREFACE TO THE FINAL REPORT

Changes were made to the draft report in direct response to the peer-review report generated by
ICE  International on  March 31, 2011  for the U.S. Environmental Protection Agency.  Other
changes were also made in response to additional private peer-reviews. Below is an incomplete
list of where changes have been made in response to the peer-reviews. Only the most important
changes are detailed below.

1.  Implemented liquid thermal management on all  battery packs. A new section was added to
explain the thermal  management strategy Section 4.  Costs  of the thermal  management are
addressed in 5.2.3.

2.  Changed the  default maximum single-sided electrode  thickness to  be 100 microns  while
continuing to allow the user to override with any value. A greatly expanded discussion  of the
transport issues, supported with validated physics-based models as well as a sensitivity analysis,
was  placed in Section 3.6.1. Additional instruction on how  to change the default value was
explicitly stated in Section 6.2.2.

3.  Changed allocation of overhead costs based on reviewers comments to more closely represent
a Tier 1 auto supplier. Exact values with discussion are in Section 5.6.

4.  Added cost estimates for automatic and manual disconnects necessary for safe operation of the
battery, Section 5.2.3.

5.  Added and  increased costs  of battery  management  system  to represent the complete
monitoring and control needs, Section 5.2.2 and 5.2.3.

6.  Increased cost of tabs to account for polymer sealant and use of isolation tape, Section 5.2.2

7.  Changed capital cost of the materials preparation  (lower) and coating (higher) in response to a
more detailed review of the cost inputs from an industrial partner, Sections.3.

8.  Reference and discussion on the validation of the model in Sections 1, 3.4, and 3.6.1.

9.  The definition of the designed fraction of the open-circuit voltage at maximum power and its
connection to existing vehicle battery systems is discussed in Section 3.2.

10. Breakdown  of the cathode material price into raw materials and baseline costs (processing,
profit,  utilities,  other materials)  for the oxides  based on  nickel, manganese, and cobalt is in
Section 5.2 with some sensitivity to cobalt prices.

11. Included  discussion of what the predicted model  price represents and  situations where it
might be conservative or optimistic, Section 1.

12. Increasing clarity on how to use the spreadsheet, Section 6 & 7.
                                          XVll

-------
13. Clarification of the production volume scaling methods, Section 5.4.

14. Discussion of where waste management costs are included in the estimation, Section 5.5

15. Explanation of approximate  method to account for modules, composed of the same cells,
connected in parallel, Section 3.6.3

16. Cell thickness changed, moving to thinner values, Section 3.1.
                                          XVlll

-------
                                   1. Introduction

The recent penetration of lithium-ion (Li-ion) batteries into the vehicle market has prompted
interest in projecting and understanding the costs  of this family of chemistries being used to
electrify the automotive powertrain. The model described here-in is a calculation method that
was developed at Argonne for estimating the manufacturing cost and performance of Li-ion
batteries for electric-drive vehicles including hybrid-electrics (HEV), plug-in hybrids (PHEV),
and pure electrics (EV). To date, a number of cost  models of various levels of detail have been
published in different forms.1"11 The cost of a battery will change depending upon the materials
chemistry, battery design, and manufacturing process.12"14 Therefore, it is necessary to account
for all three areas with a bottom-up cost model. Other bottom-up cost models exist but are not
available to the general public and have not been explicitly detailed in an open document. The
motivation for this work is based on a need for  a  battery cost model that meets the following
requirements:

       1. Open and available to the entire community
       2. Transparent in the assumptions made and method of calculation
       3. Capable of designing a battery specifically for the requirements of an application
       4. Accounts for the physical limitations that govern battery performance
       5. Based on a bottom-up calculation approach to account for every cost factor

The Battery Performance and Cost model (BatPaC)  described here-in is the product of long-term
research and development at Argonne National Laboratory. Over a period of years,  Argonne has
developed methods to design Li-ion batteries for electric-drive vehicles based on modeling with
Microsoft® Office Excel spreadsheets.12"20  These  design models provided all the data needed to
estimate the annual  materials requirements for manufacturing the batteries being designed. This
facilitated the next step, which was  to extend the effort to include modeling of the manufacturing
costs of the batteries. In the following sections of this document, a model is presented that meets
the above criteria and  may  be used to  analyze  the  effect of  battery design and  materials
properties on the cost of the final battery pack. Use  of BatPaC requires some basic knowledge of
battery packs;  however, a user does not  need to be an expert. For instance, the number of cells
and thus battery pack voltage must be specified by  the user.  However, default values are
available for more specific requirements such as experimentally measured values. In this way, a
person with reasonable knowledge of batteries may be able to conduct cost comparisons and
"what if studies.

The battery pack design and cost calculated in BatPaC represent projections of a 2020 production
year and a specified level of annual battery production, 20,000-500,000. As the goal is to predict
the future cost of manufacturing batteries,  a mature manufacturing process is assumed.  The
model designs  a manufacturing plant with the  sole purpose  of producing the battery being
modeled. The assumed battery design and manufacturing facility are based  on common practice
today but also assume some problems have been solved to result in a more efficient production
process and a  more energy dense battery.  Our proposed solutions do not have to be the  same
methods used  in the future by industry. We simply assume  the leading battery manufacturers,
those having successful operations in the year 2020, will reach these ends by some means.

-------
Establishing the validity of the model calculation is important to justify the conclusion drawn
from exercising the model. The design methodology used has been previously validated against
cylindrical wound cell  formats.15  The calculated materials quantities agreed with  the actual
values within  3 %. Moving to a prismatic format simplifies the current collection calculation
while leaving  the governing equations unchanged. The new approach developed for calculating
the cell impedance has been validated against experimental measurements from electrodes up to
100 (im in thickness.20 The module and battery jacket construction is of lighter construction
compared to contemporary designs. The battery pack energy density  calculated in  BatPaC is
higher than the battery packs used in the first versions of the Nissan Leaf and Chevrolet Volt for
equivalent cell designs  (calculated value  of 100 Wh/kg compared  to reported value near 84
Wh/kg for the Volt).21 Significant  engineering  advances  are necessary  to  minimize current
inactive material burden in the commercial pack designs thereby reducing the cost, mass  and
dimensions of future automotive battery packs. We  have  assumed a design that we believe  will
be representative of the engineering progress achieved by successful manufacturers in the year
2020.

Validation of the input material and capital costs are more difficult to achieve as few values are
publically available. We have relied, to a  large extent, on  private communications amongst
equipment manufacturers,  materials suppliers,   cell  manufacturers, and  original equipment
manufacturers. Variation does exist amongst the communicated values and we have maintained a
practical level of skepticism for their accuracy. Experts from all aspects of battery development
have reviewed the model both privately and as part of a formal peer-review process. While the
largest uncertainty in calculated values will exist in point cost estimates, the most instructive
information  may be  gained by examining ranges  in parameter values and relative  changes
between material properties (e.g. the advantage of moving to a manganese spinel cathode from a
layered-oxide material or from increases in cell capacity, etc).

The battery pack price to the OEM calculated  by the model inherently assumes the existence of
mature, high-volume manufacturing of Li-ion batteries for transportation applications. Therefore
the increased costs that current manufacturers face due to low scale of production, higher than
expected cell  failures in the  field,  and product launch issues are not accounted for in the
calculation.  The  model results for  year  2020   could  be considered very optimistic if the
transportation  Li-ion  market fails  to develop  as  a result of insufficient investment in product
research  and  development, reduced  motivation for lowering  petroleum consumption  and
greenhouse gas emissions, and/or a series of high-profile safety incidents.

-------
                      2. Cell and Battery Pack Design Format

Various cell and battery design concepts are under development at battery manufacturers. Based
upon experience gained from  extensive previous work, we have found the exact design of the
battery does not have an important effect on the cost  for a set cell chemistry; the amounts of
electrode materials and the number, capacity and electrode area of the cells, are the determining
cost factors. The most  common cell designs for batteries nearing large-scale production are
cylindrical wound cells, flat wound cells, and prismatic cells with flat plates. Cylindrical cells
probably have a slight advantage for the assembly of the electrode-separator unit because of the
ease of making a cylindrical winding. For the different cell designs, there are small differences in
the weights of the terminal extensions and the procedures for connecting these extensions to the
current collector sheets, with a small advantage for flat  plate cells. The flat-wound and flat-plate
cells form a more compact module and have better heat rejection capabilities than the cylindrical
cells. These small differences would have minor  effects on the cost of batteries produced in high
volume in  a mature, automated production plant and all of the cell designs can be adequately
cooled for  most applications. We conclude that the BatPaC cost calculations would be relevant
for batteries differing considerably from the selected design approach.

To provide a specific design for the calculations, a prismatic cell in a stiff-pouch container was
selected (Fig. 2.1). For this design, calculations of the current collector and terminal resistances
are easily done with a one-dimensional model, because the terminals are almost the same width
as the electrodes.
              Polymer Seal of
               Cell Container
                to Terminal
                   Ultrasonic Welds
                   of Terminal to
                   Collector Foils
     Terminal Seal
           Aluminum
          Conduction
             Channel
                                                   _
                                                            Cell Cross-Sections
\l
   After Edge Sealing
                                                            After Edge Shaping and
                                                             Addition of Aluminum
                                                             Conduction Channel
                                                   Cell with Stiff, Multi-
                                                     Layer Container
Figure 2.1 Prismatic cell in stiff pouch container with aluminum conduction channel added for
heat rejection

-------
The cells are hermetically sealed in module containers, which are then enclosed in an insulated
jacket. The module enclosure provides added protection of the cell seals from the diffusion of
moisture  from the external environment into the interior of the cells or alternatively loss of
electrolyte solvent from the cells. The exterior  surfaces of the modules are cooled by ethylene
glycol-water solution.  The calculated electrical  performance of a battery of this construction is
near optimum and the configuration is compact and light-weight. We have not likely selected the
most viable design in this short study;  there may be serious flaws in some details. However, the
calculated overall performance and low cost for the selected design will be challenging to match
in actual production and will only be met by the most successful manufacturers, those that will
dominate the market.

2.1 Cell Design

The prismatic cell of this design embodies individual positive and negative electrodes consisting
of current collector foils coated with electrode materials on both sides. The current collectors are
usually solid copper and aluminum foils for the positive and the negative. An illustration of a
segment of the cell is detailed in Figure 2.1.  Each electrode is  made up of active material
particles  held  together by  a polymeric  binder. A conductive  additive,  carbon black  and/or
graphite,  is added to the positive electrode and sometimes to the negative  electrode. The
electrodes and separator each  have porosity that is filled  with the electrolyte solution. During
discharge, the Li-ions move from the electrode particles into the electrolyte, across the separator,
and then insert into the particles composing the opposite electrode. The electrons simultaneously
leave the cell through the current collection system and then enter through the opposite side after
doing external work.  The materials currently  used in Li-ion cells are based on an intercalation
process. In this process, the Li-ion is inserted into or removed from  the crystal structure of the
active  material. The oxidation state of the active material, or host, is concurrently changed by
gaining  or  losing an  electron. Other  electrode  materials based  on conversion  reactions or
electrodeposition could be implemented into the model if the user desired.
                                                                        positive
                                                                        electrode
                                                          separator
                   Figure 2.2 Cell sandwich inside of prismatic pouch cells.

-------
The electrodes are easily and efficiently prepared by coating wide sheets of foil (up to 2-meters
in width) with uncoated strips running the lengths of the foil being coated. The  individual
electrodes can be cut from these sheets with little waste of electrode coating material or foil (Fig.
2.3). The separator for these cells can be handled as a single sheet that is folded back and forth as
the electrodes are  inserted. The electrodes are inserted so that all of the positive tabs extend
beyond the separator sheet in one direction and the negative tabs extend in the opposite direction.
The design model selects the number of electrodes to meet a set cell thickness determined by the
type of cell: HEV, 6 mm; PHEV, 8 mm; EV,  12 mm. These cell thicknesses are default values
and may be changed to suit the designer.   The cell terminals  are formed from flat stock to  be
almost as wide as the entire cell. They are  bent to the shape shown in Fig. 2.1 and ultrasonically
welded to the current collector tabs.  The cell stack is then sealed between the two halves of the
cell container. The cell housing material is  a tri-layer consisting of an outer layer of polyethylene
terephthalate (PEP)  for strength,  a middle  layer of 0.1-mm  aluminum for stiffness and
impermeability to moisture and electrolyte solvent vapors and an inner layer of polypropylene
(PP) for sealing  by heating.22'23 The two halves of the cell container are pre-shaped to facilitate
assembly. The aluminum foil in  the cell  container material provides  stiffness and  it may  be
increased in thickness  to assist in conducting heat to the module container. After sealing the
edges of the cell, the edges are flattened along the sides of the cell to  form a compact shape and
an aluminum conduction channel is added to assist in heat rejection at the sides of the cells.
._-,_	
 I  .'.   .
   :-  •

                                                             Electrode
                                                             Coating
                                                         ~  Uncoated Area
                                                            for CC Tabs

                                                          Lines Showing
                                                          Places to Cut for
                                                          Electrodes
 Figure 2.3 Illustration of coated current collector foil showing four rows of prismatic electrodes
                     before slitting or stamping into individual electrodes

2.2 Module Design

The module format is based on a casing  of 0.5-mm thick aluminum that is sealed by double
seaming, a process that is well established and inexpensive because it is  automated, rapid, and
uses low-cost equipment that is common  in the container industry. The sealing of the module
provides an additional barrier to the loss of electrolyte solvent from the cells and the entrance of
water vapor. These deleterious transfers through the seals of pouch cells may shorten their lives
to less than the desired fifteen years.22

-------
    Heat Transfer Surfaces
     on Top and Bottom of
       Container in  Contact
      with Cell Conductors
   Cell Terminal
   Connections
       Module
     Terminal
                                                            Double-Seamed
                                                            Module Closure
Figure 2.4 Hermetically-sealed module
The cells are placed on their sides in the module and the terminals of adjacent cells are connected
either mechanically with small bolts and flat springs to maintain contact or by laser welding.
Space is provided within the module casing on the left side, as sketched in  Fig. 2.4, for an
electronics package that includes cell monitoring for malfunctions (temperature and voltage) and
for state-of-charge (SOC) control. The SOC control is activated when ever the  battery is at rest
and it diverts charge from the cells at highest voltage to those at lowest voltage.

2.3 Battery Pack Design

The model designs the battery pack (Fig 2.5) in sufficient detail to provide a good estimate of the
total weight and volume of the pack and the dimensions of the battery jacket so that its cost can
be estimated. The modules  are arranged within the battery jacket either in a single row, with the
terminals facing the same side of the pack, or in an even number of rows with the terminals in
one row facing the terminal of an adjacent row. For a pack with a single row of modules, a
busbar must be provided to carry the current to the front of the battery pack. This feature results
in an additional cost for the busbar. For batteries with more than one row of modules (Fig. 2.5),
the terminals are laid  out on the module so as not to interfere with those on the opposite row of
modules, thus  conserving space in the battery pack. The modules in a row are interconnected,
negative to positive  terminals, by  copper connectors.  The  modules casings  are compressed
together by two steel sheets bound with steel straps at the front and back of the battery pack. The
compression is necessary to ensure intimate contact between the active layers that make up the
pouch cells that are  tightly fit into the modules.  The compressive force also serves to add
structural support to the module casings.

-------
        Coolant
         Outlet
                                                Compression
                                                Plate and Straps
Polymer Foam
to Block Flow
Pack Cover
Attached to
Module Tray
             Section A-A
            Coolant
              Inlet
                                                      Section B-B
                       Pack with Two Rows of Modules
       Section A-A
                            Section B-B
                        Pack with One Row of Modules
Figure 2.5 Insulated battery jacket with enclosed modules that are cooled on their upper and
lower surfaces by ethylene glycol-water solution

The modules are supported by a tray that provides  space for the heat transfer fluid (ethylene
glycol-water solution) to flow against the top and bottom of each module. All connections to the
pack terminals that lead to the exterior of the pack and signal wire feedthroughs can be made

-------
before inserting the attached modules into the jacket and making the final closure. The bolts
depicted in the diagram (Fig 2.5) for making this closure are for illustrative purposes only.

The battery jacket consists of a sheet of aluminum on each side of a 10-mm thick layer of ridged,
light-weight high-efficiency insulation. The thickness of each of the aluminum layers is selected
by the modeling program to be 1- to 2-mm thick, depending on the total volume of the modules.
The insulation slows the interaction of the battery with the external environment that cools the
battery in the winter and heats it in the hot summer weather.16

Although the main purpose of the battery pack design for the model is to provide a plausible list
of materials  to  estimate  the manufacturing cost of the battery, the overall design approach
permits the battery to be  shaped by the designer to fit dimensional objectives. If there is a height
restriction for the battery pack, a high ratio of height-to-width for the positive  electrode will
result in a battery of low height (the cells are placed on their side in the pack).

-------
                 3. Modeling of Battery Design and Performance

The design portion  of the model calculates the physical properties of a battery based on user-
defined performance requirements and minimal experimental data. An illustration of the model is
shown in Figure 3.1. The user is asked to enter a number of design parameters such as the battery
power, number of cells and modules, and target voltage at maximum power, etc. In addition, the
user must enter one  of the following three measures of energy: battery pack energy, cell capacity,
or vehicle electric range. Defining one of these values will determine the value of the other two.
An iterative procedure  then solves for the user defined energy parameter (energy, capacity, or
range) and remaining battery properties by varying the cell capacity and electrode thickness. The
result is the dimensions, mass, volume,  and materials requirements  for the cells, modules, and
battery pack.

The model has been designed to allow the user to enter as many customized values as desired. In
this way, the model allows flexibility in the battery chemistries studies and some of the cell,
module,  and battery design aspects. Hence, the  focus  of this  report is on  the method  of
calculation and not the exact values chosen for a specific capacity or cell thickness. However, the
default cell  design  parameters  as well  as  experimental data measured at Argonne  National
Laboratory,  for a number  of different battery chemistries  both commercial and developed at
Argonne, are available  for use within the model. There are five governing equations for battery
performance that  calculate  the  current density,  battery  energy, electrode  area,  electrode
thickness, and resistance. The voltage at maximum power and the area specific impedance  (ASI)
are two important parameters in the design model for calculating the battery performance. Most
of the discussion will be spent on these two properties.
3.1 Criteria for Power, Energy, and Life

In order to fully specify a battery design, the user of BatPaC must supply criteria for power,
energy, and life. These criteria will depend on the application for which the battery will be used.
While the user may change some of the settings as they prefer, we list our suggestions in Table
3.1. The battery type is defined by the end-use application. Hybrid electric vehicles (HEVs),
plug-in hybrid electric vehicles (PHEVs), and electric vehicles (EVs) have increasing levels of
electrical energy storage for use by the vehicle drivetrain. The model will use Table 3.1 or the
user's explicit inputs to size the battery correctly for the chosen application.

Table 3.1 Criteria for designing batteries for a specific end-use application
Battery Type
SOC for Rated Power, %
Power Duration, sec
SOC Range for Useable Energy, %
Cell Thickness, mm
microHEV
50
2
40-65
6
HEV-HP
50
10
40-65
6
PHEV
25
10
25-95
8
EV
25
10
15-95
12

-------
                    Battery Design Model
        Pack Requirements
        •   power
        •   energy or range
        •   number of cells
Key Constraints
•  max electrode thickness
•  target cell potential, V, at
   peak power
•  cell/module format

                       Iterative Spreadsheet
                       Solves for cell capacity
                       and designs battery pack
                       by varying:
                           1.  Cell area
                           2.  Electrode thickness
                           3.  Internal resistance
      Cell Chemistry
      Measured Properties
      •   pulse power ASI
      •   discharge ASI
      •   mAh/g, g/cm3
      •   electrode porosity
      •   SOC window
      •   physical properties
      ASI = area specific impedance
Figure 3.1 Summary flow of the design model

The microHEV is a micro or mild-hybrid that provides a moderate power level, -25 kW, for two
seconds. This  design is best suited for cell chemistries capable of very high power-to-energy
(P/E) ratios. The HEV-HP is a power-assist hybrid that provides the rated power for a full 10
second pulse. The power for both HEV applications is rated at 50 % state-of-charge (SOC). The
energy available for discharge and charging is 25 % of the total energy to ensure long cycle life.
As the capacity of the HEV cells is typically small, a cell thickness of 6 mm is used. The PHEV
utilizes a much larger  portion of the total energy, 70 %.  At the end of discharge, the PHEV
battery is operated  in a charge sustaining  mode. Therefore, the power rating for the battery is
   Calculated Battery
   Pack Properties
   •  dimensions
   •  volume & mass
   •  specific energy,
      power
   •  materials required
                                     10

-------
determined at 25 % SOC. PHEV cells should be much larger than HEV cells and thus a cell
thickness of 8 mm is  assumed. Finally, EV batteries use 80 % of their total energy with their
power rated near the end of discharge. Cell thicknesses are set to 12 mm to accommodate a high
capacity  design. As  noted later  in the  report,  selecting a  parallel arrangement  of cells
automatically  assumes  a cell thickness  of 6  mm  regardless  of the  end-use  application.
Additionally, the use of negative electrodes operating at potentials high above the lithium metal
potential may extend the upper end of the available SOC range from 95 to 100 %. The lithium
titanate spinel, Li4Ti5Oi2 (LTO), negative electrode is an example of intercalation electrode with
almost no risk of plating lithium metal during a charge pulse. On this basis, the available energy
for LTO-based Li-ion chemistries is suggested to be 75 % for PHEVs and 85 % for EVs.

In an established factory, the fixed design parameter is most likely the electrode area for a single
layer rather than a set cell thickness. To make higher capacity cells,  more  layers  of the
predetermined footprint are stacked, thus increasing the cell thickness. In our model, the plant is
constructed for the sole purpose of building the battery being designed. Flexibility to produce
other products is not taken into  account.  Most importantly, the model calculates very little
difference in cost whether the cell has large-area layers that are few in number as compared to a
high number of layers smaller in area. For ease of calculation, we use a  set cell  thickness  to
determine the number of layers. The area of each layer is set by the cell capacity requirement.

Accounting for capacity and power fade in the battery requires the user to design the battery with
the appropriate excess energy and power at the beginning-of-life (BOL). Defining the voltage at
which maximum power is achieved at BOL is one way to set the allowable power fade over the
life of the battery. This is discussed in detail in the following section. Capacity or energy fade
must be accounted for by  over-sizing the battery at BOL.  If  the  user has  a certain fade
requirement, then the BOL energy may be  increased to meet the end-of-life  (EOL) target.  The
design model does not attempt to predict fade rates or even suggest an allowable fade for a
specific  application. It is our view that many aspects of materials chemistry, cell design, and
battery use directly affect the  decay  of the battery pack. Hence, we allow the user  to make
accommodations for decay as he or she believes is necessary.

3.2 Voltage at Maximum Power

The  voltage at which a cell reaches the designed maximum power is one of the most important
factors in the design of a battery. However, this specification is one of the least discussed aspects
of battery design. The voltage  at maximum power, Vceii, is a measure of the largest polarization
the cell will undergo during operation at the BOL. This initial value has a direct effect on round-
trip battery efficiency, heat removal requirements, cold-cranking  power, and allowable power
fade. A basic calculation demonstrates the maximum achievable power for a battery at BOL is at
50 % of the  open-circuit voltage  (OCV). Operating at these conditions would  result in an
inefficient battery and require a significant cooling system to reject heat. More importantly, the
battery will never be able to reach this power level after any increase in impedance occurs. With
all certainty, the impedance of a battery will rise with time and the power rating of battery will
no longer be accurate.
                                           11

-------
We design the battery to achieve EOL power capabilities at a specified fraction of the open-
circuit voltage, [V/U], at BOL. This approach is  unique when compared to current design
practice of OEMs and cell manufacturers. However,  a characteristic value of [V/U] exists for all
batteries regardless of the battery design process. One may determine this value for an existing
system in a straightforward manner. The potential at maximum power is measured at the end of a
10 s pulse at the EOL power rating and the SOC used for the power rating of a specific battery
type (HEV,  PHEV, EV). The designed [V/U] value is the measured potential at  the end of the
pulse divided by the open-circuit potential reached long after the pulse.  This design point then
captures the degree to which the battery has been oversized to enable long-life, cold-start, and
efficient operation.  The remainder of this  section presents  a discussion for setting the BOL
voltage at maximum power at no less  than 80 % of the open-circuit  voltage, [V/U]  =  0.8.
Defining the voltage as a fraction of the OCV, allows  for direct calculation of all the necessary
battery properties (see for example Eq. 3.6 or 3.8 in the section 3.3).

The  allowable  increase in battery resistance over  the life of the battery is a function  of the
designed voltage  for maximum power. In general, designing the battery to achieve maximum
power at a higher [V/U] allows for larger resistance  or impedance increases over  the lifetime  of
the battery.  Figure 3.2 created from Eq 3.1  displays how  the voltage at maximum power  will
change  to meet the designed power as the internal resistance of the battery increases. Clearly,
achieving  BOL power at a high  fraction of the OCV allows for  greater  degradation within the
usable lifetime of a  battery. RI is  the initial resistance of the battery at BOL while R2 is the
resistance as the  battery ages. If the minimum voltage is 55 % of the OCV,  the allowable
increase in resistance for batteries designed for BOL max power at 70, 80, and 90 % OCV is 18,
55, and 175 %. The consequence of achieving the  power at lower and lower fractions  of the
open-circuit voltage is that  both electric  current and heat  generation  will increase over the
lifetime of the battery, Figure 3.2b and Figure 3.3. The proper design of a battery will account
for the changes over the entire lifetime and not just desired behavior at BOL.
                          V_
                          U



-y-
	
U _
(
U-
ll
"y"

[/
A

J
(3.1)
The  level of heat production is significantly different at BOL for batteries designed to meet
maximum  power at differing  fractions  of the open-circuit voltage.  We may compare the
differences in designed [V/U]  by assuming the resistive heating  (joule heating) is the most
significant factor in determining the heat generation,  Eq.  3.2. This  assumption is true for
moderate  to  high  rate  applications. We also  reasonably  assume the ASI  will not change
significantly  in  the range of  current densities  and electrode thicknesses  we  vary  in the
comparisons. From  this point, we can analyze the difference in heat generation from different
designed [V/U] values, Eq 3.3.
                                           12

-------
     I
     o
    O.
    re
    >
    O
    O
    u
    re
               Designed
              BOL [V/U] at
              max Power
        50

       80

       70

       60

       50

       40
 Designed
BOL [V/U] at
 max Power
          0     25    50    75   100   125   150   175   200
                      Increase in Resistance, %

Figure 3.2 a) Required change in [V/U] to maintain rated power with increases in internal
   resistance over the life of the battery, b) Increase in current due to lowered [V/U].
                                 13

-------
                                                  U2 1-
                  qj=(Aposl)2RJ=ItotalU\l-
V_
~u
                                                          U
(3.2)
                                            y
                                            U
                                                                              (3.3)
ouu
A Kn -
4OU
^9
0^
•* /inn -
C T'UU
o
t; i^n -
co oou
3_
0
c onn -
0 ouu
O
*- ocn _
CO ^OU
0
onn -
_ ^uu
0) 1c-n _
(0 1OU
CO
0
*- mn -
O 1 vlvl
_c
en _
ou
n _
Designed
BOL [V/U] at
max Power /
	 90%
	 80% /
---- 70% /
/
/
/
I /

J^
              0     25     50     75    100   125   150   175   200
                            Increase in Resistance, %

 Figure 3.3 Change in heat rejection requirement from increases in resistance for batteries with
                     different designed voltages at maximum power.

The ratio of resistances may be found  by equating the power for the two cases. Then the
resistances, and areas if the ASIs are equivalent, are determined solely by the fraction of the
open-circuit voltage at which they achieve maximum power, Eq 3.4. Then substitution will give
                                        14

-------
the ratio of heat production at maximum power for the two cases, Eq. 3.5. A battery that achieves
maximum power at 80 % of OCV will have a  heat production at maximum power that is 2.3
times higher than one designed at [V/U] = 90%. A battery producing power at 70 % of the OCV
will have 3.9 time higher heat generation than at  [V/U] = 90 %.
                                    owerlA
                                   powerl  pos2
"y"
u _
"y"
u _
(
,( ~
(
2\ ~
"y"
u _
"y"
u _

J
1
J
                                                                                    (3.4)
"y"
u _
"y"
_c/_
(>-
i-
"y"
"y"
j
j
                                                                                    (3.5)
Two straightforward design changes will enable operating a battery at 90 % of OCV compared to
80  %  while  maintaining  the  same  power output. First,  a second  identical  battery may be
connected in parallel to the original battery. This will lower the resistance of the battery pack by
one half but will also double the energy and cost of the battery. A more realistic approach is to
reduce the electrode thickness by coating a larger separator area. The capacity of the  cell is
maintained while minimizing increases in  cost from  a  larger separator, current collector and
packaging area. This approach is feasible as long as the reduced electrode thickness is above the
increase in AST, > 20 microns as discussed in detail below.

The efficiency of a battery defines the heat  rejection requirements  and may  be measured or
calculated. Measurement of round-trip  efficiency of  a  battery is  best performed  by using a
calorimeter to measure  the heat given off during the cycling of the battery.  The calorimeter
removes the requirement of knowing the exact SOC of  a battery during the entire drive cycle.
Calculation of the round-trip efficiency of a battery requires a detailed transient battery model
within a vehicle simulation program to exercise the  battery over the many acceleration and
deceleration periods that occur during a drive cycle. The interesting result is that the same battery
will have  different power ratings depending on what  level of round-trip efficiency  the user is
willing to  accept.

Figure 3.4 shows the efficiency of a battery as a function of the designed potential at which the
battery reaches maximum power. The figure is created using Equations  3.1  and 3.4 above. Each
line may be considered a different drive cycle, or  duty load, for a battery with  the same energy
but different impedance (changing separator area). The straight, solid black line represents the
efficiency of the battery operated only at maximum power,  P/Pmax = 1. In example, a battery
designed at [V/U]  = 0.8 will have  80 % efficiency for a single discharge pulse at maximum
power. Likewise, a battery designed at 0.9 will be 90 %  efficient at maximum  power. Batteries
are normally  operated in the area above the line  of the  maximum power.  Therefore, the other
curves represent the efficiency of discharging a battery at power levels below maximum power
                                           15

-------
(typical driving conditions). Consider two batteries each designed for a maximum power of 100
kW although one achieves this power at a [V/U] = 0.9 and the other at 0.7. If the two batteries
are discharged at 45 kW, P/Pnw* = 0.45, the battery designed at [V/U] = 0.9 will be 6.4 % more
efficient.  This is significantly less than the 20 % efficiency improvement realized when operated
at maximum power. The efficiency penalty is reduced as the battery operates less and less near
the designed maximum power.
             100%
              50%
                   0.5
0.6
0.7
0.8
0.9
                                Designed [V/U] at Max Power
Figure 3.4 Efficiencies for batteries designed to achieve maximum power at different fractions
of their open-circuit voltage. Comparative efficiency lines are shown for equivalent power
demands over a period of battery operation.
                                          16

-------
3.3 Governing Equations

The five coupled, algebraic equations that govern the battery design are presented in this section.
While these equations are perhaps the most important, many other equations  are used to fully
define the battery mass and volume. These other equations will be specified where necessary in
the following subsections.

The user of the model specifies the required maximum power, Pbatt, of the battery. This power is
translated to  a current density, /, in Eq 3.6 using the area of the positive electrode, Apos,  the
number of cells, NCeii, the open-circuit voltage at the SOC for power,  C/0cv,p, and the fraction of
the open-circuit voltage at which maximum power is achieved, [V/U].
                                  /=-
                                              batt
                                                     v_
                                                     u
(3.6)
The relationship between capacity and battery energy is described by Equation 3.7. Formally, the
energy of a battery is the product of the capacity and the average voltage at which the energy is
obtained.  The average cell voltage is approximated in Eq. 3.7 by subtracting the polarization
from discharging the battery at a C/3 rate from the  open-circuit voltage at the SOC for energy,
f/ocv,E- The energy for all batteries designed by the design model is calculated at a C/3 rate and
the average open-circuit voltage at 50 % SOC. The remaining necessary values are the capacity
of the cell, C, ASI for energy, ASIenergy, number of cells, and area of positive electrode. Either the
battery energy or capacity may be specified. The energy may alternatively be determined from a
stated range, fraction of total energy available, and energy usage rate for the vehicle (Wh/mile).
                                                 r ASI
                                                       en
                                                        ergy
                                                 3   A
                                                                                      (3.7)
                                                       pos   j
The area of the positive electrode in Eq. 3.8 is determined largely by the area specific impedance
for power, ASIpower, and resulting voltage drop. The voltage of cell at max power, Vceii,p, is found
from the product [V/U] £/0cv,p, where Uocv,p is the open circuit voltage at the SOC for max power.
In general, the area of the electrodes will increase if the ASI for power increases.  The areas of
the negative electrode  and separator are determined from the area of the positive electrode. The
negative electrode is taken to be 1 mm larger than the positive electrode in both height and width
to alleviate concerns of lithium plating during charge pulses. The separator area is slightly larger
than the negative electrode to prevent the electrical shorting of the two electrodes.
                                                   P
                                               power rbatt
                                                         u
                                                                                      (3.8)
The positive electrode thickness, Lpos, in Eq. 3.9 is determined from the capacity of the cell, C,
specific capacity of the electrode, Q, volume fraction of active material, sact, bulk density of the
                                            17

-------
active material, p, and the positive electrode area. The negative electrode thickness is determined
by its specific  reversible-capacity and the designed excess-capacity to prevent lithium plating
during charging. We have chosen a ratio of 1.25 negative to positive reversible-capacity (N/P
ratio) for the default value for the cells with graphite negative electrodes. LTO negative electrode
based cells are designed  at a  1.1  N/P ratio  because of the previously mentioned  minimal
possibility of lithium deposition. The maximum allowable electrode thickness is a user defined
value. The calculation for the electrode area changes when the designed thickness is greater than
the maximum allowed (Section 3.6.1).


                                                  .                                    (3.9)
Finally, the AST for power (and for energy to a lesser extent) is calculated using an expression
that  is  based on  the  electrode thicknesses,  the  current  density, and  the  C-rate. The exact
expression will be  discussed in the next  session. The AST in  Eq  3.10 shows  the basic
dependencies with a and P being constant valued parameters.
                                              pos
3.4 Calculation of the ASI

In most battery design scenarios, the ASI directly determines the electrode thickness to meet a
specified power-to-energy (P/E) ratio. From this electrode thickness, the area of the electrode is
set to meet the capacity requirements. Clearly, the ASI plays a significant role in the design of a
battery  and particularly in the case of the P/E ratios required by  automotive  applications.
However, the  ASI is not an inherent constant of a specific battery chemistry or cell design. The
measured value of the ASI is a complex combination of resistances within the battery resulting
from  the  physical processes occurring at different  length  and  time  scales. Consequently, the
measured value is a function of many factors  (state  of charge, pulse length, current density, C-
rate, particle size, transport and kinetic parameters, etc). The calculation used for the ASI in this
battery design model has been discussed in detail and validated against experiments elsewhere.20
The physical meaning of the equation will be discussed but those interested in the derivation are
directed to the separate publication. We note that the  ASI described here is slightly different than
the one addressed in the paper. The thermodynamic  component is removed that originated from
the change in  open-circuit potential with concentration for the intercalation materials. Equation
3.11  contains  the definition of the ASI used in this document. 7ti is a positive valued current
density for a discharge pulse. I& is equal to zero as it is during the relaxation  period after the
pulse. Time 0, tO, is the time just before a current pulse begins, time 1, tl, is the time just before
the current pulse ends, and time 2, ?2, is the time long  after the current pulse when the cell is at
open-circuit and the concentration gradients have relaxed. Therefore, this ASI  measurement is
not troubled by accounting for a change in  open-circuit voltage with the passage of current. In
general, the ASI measured  with this definition is similar,  although smaller,  in value  to  those
produced using the more standard definition  used elsewhere.
                                            18

-------
                                                                                     (3.11)
The AST for the electrochemical charge and discharge process is referred here-in as ASIechem. Our
calculation  approach  for  both the  AST  for  power and  for energy  involves adding  three
components together to reach the ASIechem, Eq. 3.12. The first two factors include impedance that
arises from the interfacial charge transfer and transport. The third factor is a lumped parameter
used to capture the remaining impedance.

                            A S7     — A <\JP°S -L A 
-------
                                      I = rcQpeaaLpos                                (3.15)

The cell AST for energy, ASIenergy, and power, ASIpower, are determined by adding the ASIechem to
that of the current collectors, ASICC, as discussed in the next subsection.  The difference between
ASIenergy and ASIpower is that the limiting currents are not important during the C/3 discharge for
energy  and the ASIconst is a different value for two cases. ASIenergy will always be higher than
ASIpoWer if a battery is operated far from the limiting current. The higher  impedance is due to the
formation of significant concentration polarizations during the longer time  scale of the  energy
discharge.  A reasonable rule-of-thumb is that  the ASIconst for energy is  2.2  times the value for
power in layered oxide materials such as LiNio.soCoo.isAlo.osOi.

3.4.1 Current Collection Resistance

The resistance from the conductors used to collect the current must be accounted for as they can
contribute  significant ohmic drop to the battery. The AST  used to calculate  the required cell
separator area, ASIpower, is  larger than the AST for the electrochemical charge and discharge
processes, ASIechem,p, as shown in Equation 3.16. The ASIechem value is typically measured from
experiments and must be added to the external resistances that arise from the  materials used to
conduct the electric current. These resistances  come from current collection in the cell and also
those on the module and battery pack level.


                       ASIpower = ASIechemf +ASICC  +ASl£ +^^L                (3.16)
                                                              Ncells

The current collector foil impedance, ASICC, is determined  from an analytical expression, Eq.
3.17, which accounts for the coated and uncoated region of the foil, labeled act for active and tab
respectively. The resistance factor, R/,  and the resistance of the current collector foils, Rcc, are
also shown for clarity in Eq 3.18 and 3.19. The factor of 2 in the R/term is due to assuming half
of the foil thickness carries the current produced on one side of the foil.  While all of the current
passes through the tab region, the magnitude of the current varies along  the height of the coated
foil as the reaction area continually contributes current to the foil. An equivalent length  for the
resistance calculation may be determined so that multiplication by the total current for a cell will
give the correct ohmic drop.  This  equivalent  length is H/3 if the current density  is relatively
constant over the  entire area. The derivation  of this  equivalent length as well as  an in-depth
discussion of the voltage and  current distribution in the foils may be found  in subsection 3.4.2.
Also in the later subsection, the assumption of constant current density is verified with numerical
modeling.
                        ASIcc=HaaWaaRcc=Rf
(3.17)
                            Rf=\	7	+	7	1                      (3-18)
                                 V   foil ,neg  foil ,neg     foil, pos  foil, pos
                                            20

-------
                                                                                     (3.19)
The cell terminals are ultrasonically welded to the ends of the current collector foil tabs. While
the welding removes this contact resistance, the AST of the terminal must be included in the total
cell resistance. The AST  of the cell terminals, ASI^,  is the summation of the positive and
negative cell terminals as shown in Eq 3.20. The dimensions for these terminals are set by the
calculated width of the cell and the user defined terminal thickness and height.

                                      1.11   #_                          (3_2Q)
                                   (7       (7      \W   J
                                     term,neg     term,pos J   term  term
                                                                pos
The AST for connection losses is the last term in the AST summation stated in Eq. 3.16. This AST
value is calculated by multiplying the ratio of cell positive electrode area to number of cells by
the  summation  of  the  resistances,  ^cnct,  for  cell  terminals,  module  terminals, module
interconnects, and batteries terminals. In this way, each cell shares in the burden of overcoming
the system losses from carrying the electric  current. The calculation of ^Cnct is detailed  in Eq.
3.21 with the individual sources of connection losses shown.  The voltage drop resulting from
cell-to-cell contact resistance, Rc^ct ,  is taken  to be 10~4£/OCv,E in Eq. 3.22, a small fraction of the
open-circuit voltage. A battery manufacturer would only tolerate a minimal voltage drop from
cell-to-cell contact. One connection method is to physically press the two cell terminals together.
This resistance could be lowered by increasing the physical pressure and contact area, or by laser
welding the terminals together. Regardless, the value used in the model is left to the choice of the
user to leave as is or to change to a different value.

                                        modR^ + (Nmod - Ifel + R%L                (3-21)


                                        = 10"4 Nc*U°">'E                              (3.22)
The module terminal resistance, R™^ , calculation in Eq. 3.23 is shown as an example of how the
terminal and interconnect resistances are calculated for the module and battery pack. The size of
the terminals  and thus  their resistance are determined from  a calculation based on a  pre-
determined allowable rate of temperature rise for the conductor. This approach is explained in
more detail in subsection 3.4.3.
                                   mod       " term     AT mod
                                   term =	7	Nterm
                                            21

-------
3.4.2 Potential and Current Distribution in the Current Collection Foils

The designed current collection system was evaluated using a numerical simulation package.
Equations 3.24-3.26 were solved for a steady state, isothermal, and 1-D simulation. Here, the
conductivity, 03, is the effective conductivity of l/2 of the foil (the other half carries the current
from the opposite side).  The bulk conductivity value, 03°, is multiplied by the thickness of the
conductor, L/2, to lower the dimension of transport.


                              In=—	H^	^^^                        (3-24)
                                                    -In                             (3-25)

                                                    ln                              (3-26)

The boundary conditions were set for both ends of each foil. The tab ends of the foils were set to
a specified voltage and the opposite ends of the foils were restricted to a no flux condition. The
simulation was performed using the foils defined in our battery design: 12 micron thick copper
foil and 20 micron thick aluminum. The cell length was 20 cm, the ASIechem was 30 ohm cm2, and
the Uocv and Vceu were set to 3.72 and 3.57  V respectively. Figure 3.5  shows the current and
potential distribution in the foils and in the cell resulting from the  simulation. The cell potential
along the  length of the foil varies  only by 1.5 mV from maximum to minimum difference. The
0.4 % variation in voltage results in a 0.9 % variation in current density. This verifies the current
density is  uniform along the length of the foil. This is also obvious from the linear relationship of
current with foil height in Fig. 3.5. The assumption of constant current density was tested in cell
heights up to 100 cm and found to be satisfactory. The assumption should be reasonable as long
as the ASIechem is at least twice the value of ASICC.  The simulated resistance of the foils  is found
to raise the ASIechem by 0.7 for an ASIpower of 30.7 ohm cm2. Additionally, the numerical result
verified that H/3 is the correct equivalent length to represent the ASICC for the cell. This may also
be found analytically, Eq. 3.27-3.29, if you assume an even current distribution. We have shown
that to be a reasonable assumption.


                                                                                    (3.27)
                                             21 gs -<&lne
               ASIechem + ASICC = — I   :   -   —
                                H  0      ln
                                            22
                                                                                    (3.28)
1     l   '         (3.29)

-------
               3.5750
            „  3.5740
          .2
          "^
           c
           0)
          +•>
           o
           Q.

          "55
           o
           c
           (0
3.5730
               3.5720
           O   3.5710
           0)

           "^
           '(75
           o
           Q.
           o
           o>
           Q.


           0)
           i_

           O
3.5700




3.5690

  0.12




  0.10




  0.08




  0.06




  0.04




  0.02




  0.00
                                                     0.0000
•Cell

•Positive foil

•Negative foil
        Negative foil

        Positive foil

       •current density
                             0.2
                        0.4
                 0.6
0.8
                                           x/H
                                                              a>

                                                     -0.0040  '•*=

                                                              O)
                                                              0)
            -0.0050




            -0.0060

            0.00492




            0.00491
                                                                             O
                                                                    0.00486
Figure 3.5 The change in current and potential within the positive and negative foils. The current

collection design results in a uniform current distribution along the length of the foil.
                                           23

-------
An analogous problem has been solved  by Euler and Nonnemacher and then communicated
repeatedly by Newman et al.24' 25 The analytical solution they presented may be used after a
slight alteration to dimensionalize the current density to the geometry of our concern, Eq. 3.30
and 3.31. This solution was reached assuming linear polarization behavior and is valid for cases
where the current density varies along the height of the current collector foil. Thus, this approach
is a more general solution than the one we use in the design model.
                                                 H2
                                                                a
                                                                  •"
                                                                     + -
                1 + -
                                                                   v sinh v
                                                                                   (3.30)
                                      H2
                                        echem V   pos    neg
                                                                                   (3.31)
3.4.3 Determination of Module Terminal, Battery Terminal, and Module Inter-connect Size

An important factor for setting the resistances of a module terminal, battery terminal, or module
interconnect is the allowable rate of temperature rise in the conductor at full power. We set the
acceptable rate of temperature rise, dT/dt,  at 0.2 °C/sec or a 2 °C rise for a 10-sec power burst
under adiabatic conditions. The heating rate, q, is then used  to determine the mass,  m,  of the
terminal  required for the designed battery  in Eq. 3.32. Since the heating  rate may also be
determined by Eq. 3.33, we may determine the cylindrical terminal radius and mass by assuming
a length,  //term- In this way, the size of the module terminal is redesigned during each simulation
to meet the specified power requirements and allowable temperature rise, Eq. 3.34. The mass of
the conductor is found to be inversely proportional to the allowable temperature rise.
   r  dT
 mCp~T
    p dt
                                                                                   (3.32)
                                        term  term
                                                                                   (3.33)
                             A    = 7A
                              term    pos
           dT_
r'term " term ^ p  ,
                                                        -1/2
(3.34)
A copper busbar must also be sized for batteries using a single row of modules.  We have
somewhat arbitrarily assumed a AV&, = 30  mV drop  across the busbar to be  allowable  at
maximum current. This value maybe  easily changed by the user. Equation 3.35 is used  to
calculate the mass of the busbar, m^,. The complicated expression for the volume of the busbar is
                                           24

-------
derived from  the  voltage drop, conductivity,  busbar width, Wbb,  and required busbar cross-
sectional area.


                                             ltOt\Wbb)
                                    nlbb ~ rbb
3.5 Calculation of Battery Dimensions

The goal of the model is to quantify how the various components of a specific battery design
sum to make the mass and volume of the battery pack. In this way, a true energy and power
density  can be calculated as well as  the exact materials requirement  to meet  this design.
Summing the mass of the components is relatively straight forward. Determining  the  total
volume that contains the components and required free volume is not as  obvious. The exact
calculations used in the design model are detailed below for the cell, module, and battery pack.

3.5.1 Cell Dimensions

The number of layers in each cell is approximated in Eq. 3.36 by accounting  for the compression
factor, .Xcomp, and the individual thicknesses of the current collector foils, Lf0a, electrodes, Lpos
and Lneg, separator, Lsep, and container, Lcont. Xcomp is usually taken to be 0.97. The Li-ion battery
chemistries this model was designed  for are assumed to undergo negligible volume change on
the cell level. No effort  was made to address possible changes in electrode volume upon cell
discharge or charge.

                                           T   _2/~   +T"eg
                       AT    _ y            cell   ^^cont ^ ^foil                        ,~ ,-, -^
                         layers ~   comp j mg   j pos ,J,T   ~j   ~j   7                 ^ '   >
                                    ^foil ^ ^foil ^ L\^sep ^ ^neg ^ ^ pos >

The Niayers approximation is  necessary  as the cell thickness is a  user  defined parameter. The
aspect ratio of the cell is also user defined; therefore, solving  for the width  also determines the
height of the cell as seen in Eq. 3.37.  The width is calculated from the number of layers and the
aspect ratio, H/W. The factor of 2 enters the denominator as both sides of the foil are assumed to
be coated.
                                                                                    (3-37)
Having determined the width and height of the  electrode, the rest of the cell dimensions are
relatively straightforward, Eq. 3.38 and 3.39. The width of the cell, Wceii, is 8 mm wider than the
positive electrode to allow for the larger separator area and pouch seals. The pouch  seals are
folded up, pressing along the inside wall of the module casing. The height of the cell, Hceu, is the
height of the positive electrode in addition to the distance for the terminals and connections to
the foil tab, Lterm>cnt. Our assumed design requires 15 mm for this distance. The volume of the cell
is the product of the three dimensions.
                                            25

-------
                                     WaU=Wpa+*                                (3.38)

                                  Hcdl = Hpos + 2Lmt                            (3.39)

3.5.2 Module Dimensions

The module dimensions are defined by Eq. 3.40-3.42. The height and length of the module are
both just 2 mm wider than the cell dimension. The width of the module is related to the total
thickness from all of the cells with allowance for a SOC controller at one end.

                                     Hmod=Wcdl+2                               (3.40)

                                     Lmod=Hcell+2                               (3.41)

                                Wmod=Lcell(Ncell/mod+l) + l                          (3.42)

3.5.3 Battery Pack Dimensions

The battery pack volume includes all of the modules, spacing for connections between modules,
channel for the cooling air to flow, Hair, thickness of the module compression plates, Lcomp, and
the battery pack jacket, Ljack (Eq.  3.43-3.45). Ljack includes  a 10 mm thick insulation layer
sandwiched between two aluminum walls for the container. The thickness of the aluminum wall
increases from 1 to 1.5 to 2 mm as the battery volume increases from < 20 L to < 40 L to larger
dimensions.  The layout  of the modules, number per row, Nmod/row, and number of rows, Nrow, is
also included. The final volume of the battery is the product of the three dimensions.  The space
left for connections between modules, Lgap, is a function of the number of rows of modules. Lgap
is equal to 8, 10, or 20 depending if there is  one, two, or four rows of modules. Three rows of
modules are not allowed  as the positive  and negative terminal for the battery would be on
opposite ends and thus not very practical. A number greater than four rows of modules is deemed
unnecessary.

                               Hbatt = Hmod + 2Halr + 2Ljack                          (3.43)


                         ^batt = N mod/ row "mod + "air + ^^comp + ^^ jack                   (3-44)

                               Wbatt=NrowLmod+Lgap+2Ljack                         (3.45)
3.6 Additional Considerations

A few situations may arise that require a change in the calculation method. These situations are
addressed in the subsections below. The inclusions of these calculations into the model allow for
a more realistic depiction of limitations often encountered by cell manufactures.
                                           26

-------
3.6.1 Maximum Electrode Thickness

A practical limitation exists  for the maximum achievable electrode thickness.  This  limitation
may be set by manufacturing capabilities, ionic and electronic current transport within the porous
electrode, susceptibility to plating  lithium on the negative electrode, or aging characteristics
related to adhesion to the current collector. Some of these challenges are discussed in great detail
in the following subsection. When the maximum electrode thickness, LmaX, has been reached on
either the positive or negative electrode, the electrode area equation is modified as shown in Eq
3.46.  The electrode thickness, Ltgt,  is the largest  electrode thickness,  negative  or positive,
calculated at the targeted fraction of the OCV [V/U].

                                     ALpr=Y^Apos                               (3.46)
                                              max

The area of the electrode is now determined by the cell capacity requirement to meet the battery
energy demands and not the  target voltage at maximum power. As a consequence, the battery
pack will operate at a higher [V/U] than originally  selected by the battery designer. The  new
[V/U] may then be calculated from Eq. 3.47 which is the solution to the quadratic found in Eq.
3.48.
                             2C7
                                cell
                        v
                        '        -*-    •»- T     \l-r-r   \2,   A  UU.ll     I}UWP.t~                   /,-* A l—l\
                                                                                    (3.47)
                        -r j     ^* -r -T      CKtt  •» I V  UKll '       -» T    i                       \     '
                                                     power
                                             V  N   A
                                              cell  cells pos
                                                                                    (3.48)
The maximum electrode thickness may have a large impact on the energy density and cost of
cells designed for high  energy and range. Nelson et  al. demonstrated this  concept in 2009
assuming a 100 micron  maximum electrode thickness.12'19  In 2010, Santini et al.  relaxed this
assumption to 300 micons; although, the thickest electrode discussed in the  paper was a 225
micron  graphite electrode  in the LMO-Gr EV with 100 mile  range.13  In conversations with
manufactures, 100 microns appears to be the general electrode thickness  used  for EV type cells
at the present time. However, Santini et al. has shown substantial increases in energy density and
decreases in cost if larger  electrode thicknesses may be  utilized. The challenges to achieving
thick electrodes, in addition to  those already mentioned, relate to fast charging while avoiding
lithium metal deposition, utilizing all of the materials reversible capacity, removing gases formed
during formation cycling,  wetting  the full porosity of  the electrode,  achieving defect  free
coatings, and  drying the thick  electrode at high rates.  Our  opinion is that the  successful cell
manufacturers will engineer ways to overcome these challenges to increase energy density and
lower cost.

3.6.1.1 Challenges of Large Electrode Thicknesses

Dependent  upon the battery chemistry  and designed P/E ratio,  the  maximum achievable
electrode thickness (loading) may have a significant effect on the end cost and  energy density of

                                            27

-------
a battery pack. For  batteries  designed at low  P/E ratios  or for cell  chemistries with  low
volumetric capacities, the designed  electrode thickness based on the target efficiency is often
larger than what is feasible during operation in  a transportation environment. This  subsection
explores some  of  the challenges  that  arise  in  the  electrochemistry  when larger electrode
thicknesses are utilized.

Argonne gained a wealth of experience in the NCA-Gr in 1.2 M LiPFe 3:7 EC:EMC  cell during
the Advanced Technology Development program sponsored  by the US Department of Energy
(DOE). Dees  and coworkers developed a world-leading  parameter set  for a numerical model
through  exhaustive electrochemical measurements, ex-situ  characterization techniques,  and
multi-scale modeling activities.26"35 The resulting phenomenological cell model founded on the
methodology  originating  from  John Newman (UC Berkeley) will  be used to evaluate the
electrochemical behavior of cells  using  thick  electrodes.36 The coupled,  non-linear partial
differential equations are solved with the finite element method using FlexPDE.

Simulated discharge capacity for the C/l  and C/3 discharge rate is shown in Figure  3.6  as a
function of electrode thickness. For reference, the target  positive electrode thicknesses for this
cell operating at a 5C-rate and a  [V/U] =  0.8  is  142 microns.  The line of 100 %  capacity
utilization is  also  shown  as a means to judge  the deviation from theoretical capacity. As
expected, the C/l  rate deviates more strongly  than the C/3  rate  with increasing electrode
thickness. The loss in capacity is a result of the cell hitting the discharge voltage cut-off, 3.3 V,
before all of the lithium has been transported from the negative to the positive electrode.

Figure 3.7 displays the normalized concentration  profile of the electrolyte salt, LiPFe, at the end
of a C/l and C/3 discharge for an electrode thickness of 245  microns. The C/l discharge results
in a positive electrode starved  of electrolyte salt. This  transport limitation results  in the cell
prematurely reaching the voltage cutoff. In order to overcome this limitation, the electrode would
need to be engineered with significantly reduced  tortuosity37  or utilize an electrolyte  with better
mass transfer characteristics. This behavior is exacerbated by lower temperatures, such as those
experienced during winter driving conditions. The fraction  of theoretical  discharge  capacity
begins to lower significantly at  thicknesses greater than  100 microns, 3.4 mAh/cm , at the C/l
rate and 175  microns, 6.4 mAh/cm  , at the C/3 rate. The electronic  transport properties of the
cathode material also play an important role in determining the current distribution  within the
electrode. While the NCA material has a reasonably high conductivity, other cathode materials
have  lower  valued  electronic conductivities  and, depending  on  the  conductive  additive
properties, may have different current distributions and limitations within the electrode.
                                            28

-------
                10
             •    6
             o
             re
             a.
             re


             
-------
The  simulated AST for a 5C, 10-s discharge pulse at 60 % SOC is shown in Figure 3.8 as a
function of electrode  thickness.  The initial decrease in the  AST is a mathematical result of
diminishing significance of the interfacial  impedance  as more current is  passed  in the same
geometric area. The AST then remains constant from 75 microns to nearly 400 microns. The
constant AST results from ohmic losses that  behavior linearly with applied current. The dramatic
increase in the AST at the largest electrode thicknesses results from limitations in electrolyte
transport within the porosity of the positive  electrode. This is similar to  what is displayed in Fig
3.7 above during the constant discharge at the C/l rate for an electrode thickness of 245 microns.

The most significant issue for pulse power operation with thick electrodes occurs on the negative
electrode during a charge or regen pulse, Figure 3.9. The potential of the negative electrode may
drop below that of a hypothetical lithium reference electrode during  a charge pulse, inferring an
undesirable  side  reaction of  lithium plating on  graphite.38 This behavior is exacerbated  by
increasing  electrode thickness.  Operation  at higher SOC  and  lower temperatures will also
increase the probability of lithium plating. The lithium  reference electrode is taken to be in the
center of the separator  layer. The two times  shown in the graph, 1-s and 10-s, represent different
polarization measurements  for  the electrode.  The 1-s value includes  all of the  interfacial
impedance and minor  contributions from  concentration polarization.  The longer time value
includes additional changes in potential due to the concentration gradient in the electrolyte. The
1-s time is the more accurate valuation of the tendency of the electrode to plate lithium.
                   50
                   40 •
                   so •
E
o
E
£
O

J  20
                   10
                         •Full Cell
                          Positive Electrode
                          Negative Electrode
                                                                  .1"
                               100        200        300        400

                                 Electrode Thickness, microns
                                                                       500
   Figure 3.8 Calculated AST from a simulated 10-s, 5C discharge pulse for the NCA-Gr cell
 couple at 60% SOC. Positive and negative electrode thicknesses are similar in value for this cell
 design. Transport within the electrolyte is not limiting until the electrode thickness approaches
                        450 microns for these simulation conditions.
                                            30

-------
    0.1
    0.0
 E
 = -0.1
>
>
   -0.2
   -0.3
   -0.4
              1s Pulse Time
              10s Pulse Time
               50
300
350
                                   100      150     200     250
                                 Electrode Thickness, microns
    Figure 3.9 The potential of the negative electrode versus a hypothetical lithium reference
  electrode located in the center of separator during a 5C charge pulse for the NCA-Gr couple.

The tendency for lithium plating appears to be significant at electrode loadings greater than 3.4
mAh/cm2 under these specific conditions. Lower continuous and pulse charge rates are most
likely necessary to ensure the safe operation and  long life of this battery. Fast charging of PHEV
and EV batteries is  an oft  discussed value-added  characteristic  necessary  to increase  the
attractiveness of electric vehicles to the consumer. However, this fast charge requirement may
require a lower loading design to prevent lithium plating from occurring. This would not be true
for batteries based on LTO negative electrodes, or possibly even non-graphitic carbon electrodes
(e.g. hard  carbon). The consequence  of  a  lower loading design is that higher quantities  of
inactive materials are used resulting in a more expensive and less energy dense battery.

A limit of 100 microns has been chosen for the default maximum electrode thickness. This
thickness represents a graphite electrode balanced to a positive loading of 3.5 mAh/cm2 and is
the largest thickness that AST measurements have been validated at Argonne. However, a low
volumetric capacity electrode, such as  LMO, will result in a lower area-specific capacity as the
limit will be determined by the positive electrode thickness. One domestic OEM has suggested
that at the time of this publication, state of the art electrode loadings for PHEV applications are
less than 2 mAh/cm . This low loading level was selected based on cold-start performance, life
testing,  and rate capability studies. We note that different electrolytes will likely have the most
significant effect on the transport limitations and result in different optimum electrode loadings.
Gel-based  electrolytes or  standard carbonate electrolytes mixed with  low molecular weight
polymers are often utilized in pouch cell based  batteries.  While the  ionic conductivity of these
systems may approach standard carbonate electrolytes used, diffusion of the salt is restricted.
                                31

-------
This decrease in mass transport will result in large increases  in  impedance  during  longer
discharges  (constant  speed highway driving). Hence,  companies  using an  electrolyte  with
sluggish  mass transport will require thinner electrodes  than what the model  would normally
calculate based on pulse power applications. Greater electrode thicknesses may be achievable in
the future  as  manufacturers  expend significant engineering efforts  to minimize the inactive
material  that lowers energy  density and raises cost. The concerns  over lithium plating  may
remain unless new,  less susceptible negative electrodes are developed  that still enable  high
energy density.

Table 3.2 displays a calculated sensitivity analysis for designing  the battery at different target
electrode loadings to achieve 17 kWh of total energy for the NCA-Gr cell couple for two power
levels, 110 kW and 60 kW of power. The consequence of varying electrode loading at constant
power and  energy is to change the fraction  of open circuit voltage at  which maximum power is
achieved at BOL. This quantitative comparison of electrode thickness with [V/U] is only valid
for the specific battery and cell chemistry designed in  the  table, while the qualitative  results
would hold for all systems. An alternate but equivalent way to create Table 3.2 is to maintain a
constant  value of [V/U], but vary the designed power level. As mentioned previously, private
communications  suggest that current electrode thicknesses used in PHEV  applications are near
50 microns or 2 mAh/cm2; however, the exact details of that cell configuration are unknown and
thus direct  comparison should be conducted with caution. For the cell chemistry shown in the
table, a loading  of  2 mAh/cm  with 110 kW of power correlates to achieving the maximum
power at BOL at [V/U] = 0.84. Our default suggestion of using 80% of open-circuit voltage
results in a moderately less expensive battery, albeit similar in value. If a lower P/E ratio battery
was designed, then the electrode thickness (loading) would be much larger for the same designed
fraction of open-circuit voltage. The sensitivity to [V/U] decreases with lower [V/U]  values
owing to the inverse  proportionality  of electrode loading to the P/E  ratio.  The calculated
designed electrode loadings for lower  P/E  ratios  will differ the most significantly from those
used in industrial practice today. This is in  part the  motivation to limit electrode thicknesses to
100 microns as well as the transport limitations discussed earlier. While the standard practices of
today are important,  the goal of the calculations is to  evaluate the potential  cost of Li-ion
batteries  in the future years after improvements have been made resulting from the competitive
marketplace.  Therefore, calculating a range of values will be the most instructive approach to
determining where future battery costs may fall.

Table 3.2 The effect of electrode loading on the price of a  17 kWh  NCA-Gr PHEV40 battery
with 96 cells
         17kWh             110kW                         60 kW
[V/U]
68
72
76
80
84
88
92
mAh/cm 2
3.74
3.45
3.09
2.67
2.19
1.65
1.04
microns
96
88
79
68
56
42
27
Price, $
4305
4358
4442
4572
4780
5163
6045
mAh/cm2
7.38
6.80
6.11
5.30
4.38
3.36
2.22
microns
188
174
156
135
112
86
57
Price, $
3954
3977
4014
4091
4188
4364
4749
                                           32

-------
3.6.2 Accounting for Parallel Cell Arrangements

The user of the design model may wish to use a parallel arrangement of cells within the larger
series arrangement of the battery pack. Several motivations exist for a parallel cell arrangement.
For example, a battery supplier may wish to only produce cells  of a specific capacity. The
manufacturer may only have the equipment to produce a certain size cell or they may encounter
engineering design problems for very large cells ( > 60 Ah). Thus, a cell group composed of
parallel connected cells may be necessary to meet the energy requirements while staying within
battery pack voltage and current requirements.

When the user chooses to have cells connected in parallel, the design model calculation includes
the appropriate factors necessary to account for changes in the resistance, volume, and mass of
the battery. The end result is a lower energy density for the battery from including the cell group
interconnects and additional inactive material  for each cell. Furthermore, the thickness  of the
cells is reduced to 6 mm to better suit the smaller cell format.

3.6.3 Accounting for Parallel Module Arrangements

Designing a battery composed of modules in parallel cannot be explicitly simulated in the model.
However, one may approximate this scenario by simply doubling the number of modules, and
thus battery pack voltage. The simulated battery pack  is nearly  the same as if constructed of
parallel packs.  A  simple  wiring  change would lower the pack voltage with a  proportional
increase in capacity.  The overall energy and power will be the same. Subtle errors will exist in
the size of the module terminals and conductors as the calculated total current will be less in the
series arrangement than that in a parallel configuration.
                                           33

-------
                              4. Thermal Management

The  power  and life  of lithium-ion batteries are  more drastically affected  by the  battery
temperature  than they are for most other batteries including those based on lead-acid,  nickel-
cadmium, and nickel-metal hydride systems. It is important that the temperature of a lithium-ion
battery be controlled  at all times, even when the battery is  at rest. Developing  schemes for
effectively controlling the pack temperature at minimum cost will certainly be important in the
success of this technology. The most difficult problem is the removal of heat generated within
the battery,  principally by ohmic heating. Avoiding excessive temperature rise during idle
periods in hot ambient conditions is also a problem. Either of these conditions might raise the
temperature  to well above 40°C, which enhances degradation reactions and shortens the  battery
life.  Thus, maintaining the battery temperature as near the minimum temperature for adequate
power will prolong battery life. Because the battery has poor power at low temperatures, heating
the battery from a very cold condition is necessary and especially difficult for large EV  battery
packs for which no assistance  is  available from the engine.  For electric-drive vehicles to  be
competitive  in the  market with conventional vehicles, these thermal control problems must  be
solved at moderate costs and by  means that do  not compromise the  safety of the vehicle  or
battery system.

The  BatPaC model has  a  separate worksheet  for estimating the thermal  parameters and
requirements for lithium batteries and for designing the thermal management system. The results
are transferred to the Battery Design worksheet to calculate  the mass,  volume and materials
requirements for the battery pack.

4.1 Heat Generation Rates in the Battery Pack during Driving

During driving, the heat generation rate depends  on the driving cycle  and the power of the
battery relative to the demands  of the cycle for the vehicle being driven. As discussed below, the
heat capacity of the battery pack smooths  out fluctuations in the heat generation rate. The rate
that the cooling system must handle is the average rate for the most difficult driving conditions to
which the battery pack will be subjected.

The  best way to determine the maximum cooling rate  requirement for the battery pack is by
vehicle simulation studies. These  studies require  a battery impedance  algorithm that makes
possible accurate estimates of internal heating and the use of vigorous driving cycles and high-
speed driving patterns. The results of vehicle simulation studies of battery heating can be entered
on line 19 in the Thermal worksheet to override the estimated default values. However, pertinent
results are  not always accessible  and therefore  we have provided  some  initial,  although
dramatically simplified, estimations for heat generation.

For  microHEV and HEV-HP  battery  packs, heat generation  is intermittent  and substantial
periods of little or no heat generation  exist in the load  profile. The model estimates the heat
generation rate  for 25-kW micro-HEV batteries at 100W. For the HEV-HP battery packs, the
maximum average  battery power is derived from  the energy requirement, default  of E = 300
Wh/mile, entered on the Battery Design worksheet (line 157). The energy use  requirement for
HEV-HP batteries is estimated to take place at an average driving  speed of 40 mph and involve

                                           34

-------
the battery 50% of the time.  These assumptions calculate an average estimated battery power
level of 6.0 kW. For a 50-kW battery pack, the heat generation is estimated as 5% of the average
battery  power.  The heat generation for  a HEV-HP battery  of different maximum power is
estimated to be inversely proportional to the battery pack rating as shown in Equation 4.1.


                             ?HEV-HP = £*40*0.5*0.05* —                        (4.1)
                                                        "batt

For PHEV  battery packs,  the maximum heating condition  is deemed  to occur during  the
declining charge period when essentially  all of the vehicle power is supplied by the battery.
Thus, the factors controlling heating during operation of PHEVs and EVs are similar and may be
treated in the same way. The highest rate of continuous  heat generation will occur on a very
vigorous driving cycle or when driving at sustained high speed. Continuous  discharge of the
battery  at constant power results in increasing impedance and increasing heat generation rate
because of solid-state diffusion overvoltage. The area-specific impedance (AST) will reach  2-3
times the 10-second AST in  about five minutes accompanied by a similar increase in the internal
heat generation rate. Whether or not steady high speed driving will result in greater internal heat
generation than a vigorous  driving cycle will depend on many factors including vehicle speed,
battery  power,  and electrode thickness. The energy usage requirement, default of E = 300
Wh/mile, is assumed to occur at a steady discharge of 60  mph in PHEVs and EVs. The vehicle
speed and energy usage assumptions provide an estimate  of the maximum  long-term discharge
rate. The heat generation is  calculated from this power requirement and AST for energy. Finally,
the maximum  heat  generation rate is assumed to be 1.3  times the  value at 60 mph. The heat
generated in the battery pack for these conditions is inversely related  to the pack power.

4.2 Heating under Adiabatic Conditions

A factor to be considered in thermal management is the substantial  heat capacity of the battery
pack.  This  smoothes out temperature fluctuations resulting from power bursts  so that  the heat
dissipation system need  only handle the  average heat generation  rate for the most  extreme
driving  profiles the battery is likely to encounter. For large PHEV and EV batteries  the heat
capacity of the battery will limit the temperature rise of the centerline of the cells by distributing
the heat throughout the battery until steady state is reached. For a large EV battery with power of
120 kW and energy of 50  kWh, the temperature rise under adiabatic  conditions may be only
15°C or less for a complete discharge and certainly less with a cooling  system even if it is only
moderately effective.  For HEV batteries, which  have high power-to-energy  ratios, the main
effect of the heat capacity will be to smooth out temperature fluctuations.

4.3 Active Cooling Systems

There are  several choices  of coolant that have been considered for cooling battery  packs
including air from the cabin, which may be heated or cooled, water-ethylene glycol solutions and
dielectric liquids such as transformer coolants. Air is the least expensive, but it is less effective
than the liquids because of its poor conductivity, the need for large flow passages and high
pumping power. Dielectric  liquids are expensive, but have the advantage of being compatible
                                           35

-------
with terminals and other parts at electrical potential. Water-50% glycol solution is inexpensive
and has good conductivity; we have selected it as the coolant for this study.

We selected  a general cell and  battery design that can be adapted to all of the electric-drive
batteries from  micro-HEVs packs to EV  packs  (section 2).   This  design incorporates a
hermetically sealed module closure. Unfortunately, the enclosure does not have sufficient surface
area to be cooled effectively by air. This design can be effectively cooled by liquids and requires
that only the module terminals  and connectors be protected from contact with a conducting
coolant, thus accommodating water-glycol coolant. In contrast, cooling individual cells would
require  flow  passages  between  the cells, which would add  to the pack volume. That  design
feature  would permit air cooling for  microHEV and HEV-HP batteries.  For larger batteries
requiring liquid  cooling, flow passages between the cells would probably require the use of a
dielectric  coolant  because  of  the difficulty of  protecting  so many  cell terminals  from a
conducting coolant. Alternatively, elaborate flow channels with engineered seals may be used to
contain an aqueous based heat transfer fluid.

4.3.1 Heat Transfer from Cell to Module Wall

As described in section 2, the cells transfer  heat  to the cooled walls of a hermetically sealed
module with  the aid of an aluminum heat conduction channel. Some of the heat is transferred
through the sides of the cell to the channel and from there to the module wall. The remainder is
transferred directly through the seal edge of the cell to the conduction channel flange which is in
contact with the module wall. Calculation of heat transfer in this two-dimensional array through
several  materials is  complex requiring a numerical model. The spreadsheet iterates  several
hundred times in reaching a solution, each resulting in a slightly different cell design. Thus, it
would be impractical  to imbed a numerical  model directly, which may increase the total
calculation time to many minutes. Instead,  a software program  based  on the finite element
method, FlexPDE 6.15 by PDE Solutions  Inc., was employed to calculate heat transfer rates  for
70 cell configurations.  The resulting  simulations were  empirically correlated so that  simple
equations occupying a few cells  in the spreadsheet could rapidly calculate the heat transfer rate
with only a small error.

An important requirement for calculating heat transfer rates within the  cell is to estimate  the
composite conductivities of the cell layers both parallel to the layers and across the layers. The
resulting conductivities vary considerably  with the relative thicknesses of the layers as shown in
Table  4.1, for  which  the  results are consistent with the  literature.39"43 These values  for
conductivities and a range of cell dimensional parameters (Table 4.2) were employed in selected
arrangements for calculating heat transfer rates  with the FlexPDE model for 70  representative
cells that covered a broader range of variables than is needed for practical cells. For each of these
cells,  the FlexPDE model calculated the temperature difference between the cell center  and  the
module housing per unit of heat generation, AT/q (°C/W), and the  fraction of the  total heat that
was transferred through the edge of the cell, qe/q. The balance was transferred through the side of
the cell to the  aluminum conductor, qs.  The division of  the heat transfer into  two routes is
represented by the equation:

                                   q/AT = qe/AT + q/AT                               (4.2)
                                            36

-------
Estimated values for g/Ar and g/Ar were determined by empirical correlation of the results
obtained for the calculation of the 70 cells by the FlexPDE model with the result shown in Fig.
4.1. Empirical values of these estimated values resulted in the equations:
                                           °-58
                                                  9r  L2nr°-75
                                                   Lceii  W
                            / A-T   7/cior 0.55, 0.58T   -0.21-,T,r0.7T  0.72
                          qs/AT = 1628ky   kx  Lceii   W  LAi
(4.3)
(4.4)
Table 4.1 Sample calculations of composite thermal conductivities of cell structures across layer
and parallel to layers

Layer Thicknesses, microns
Positive foil
Negative foil
Positive coating
Negative coating
Separator
Total bicell structure
Thermal Conductivities, W/cm-K
Aluminum
Copper
Positive coating
Negative coating
Separator
Across layers, kx
Parallel to layers, ky
Celll

20
12
30
40
20
212

2.0
3.8
0.013
0.013
0.0020
0.00689
0.4127
Cell 2

20
12
75
100
20
422

0.2.0
3.8
0.13
0.13
0.0020
0.00689
0.4127
Cell 3

20
12
150
200
20
772

0.2.0
3.8
0.13
0.13
0.0020
0.01045
0.1228
Cell 4

20
12
220
300
20
1112

0.2.0
3.8
0.13
0.13
0.0020
0.0.01112
0.0892
Table 4.2 Range of parameter values for calculating heat transfer rates in FlexPDE model


Conductivities, W/cm-K
Across layers, kx ~~l (a)
Parallel to layers, ky J
Cell edge, ke
Cell Dimensions, cm
Cell thickness, Lceii
Cell width, W
Cell edge thickness, Le
Aluminum conductor thickness, LAI^
Parameter Levels Evaluated
1

0.00689
0.4127
0.10

0.6
8
0.1
0.03
2

0.00689
0.4127


1.0
12

0.06
3

0.01045
0.1228


1.4
18

0.10
4

0.0.01112
0.0892






      kx and ky values are calculated as in Table 4.1 and were, thus, paired together in the Flex
PDE model calculations.
^The total conductor thickness consists of the conductor thickness itself plus twice the thickness
of the aluminum layer within the pouch material.
                                            37

-------
        2,0
        1.5  -
   a
   u
   C
   IA
   OJ
   a*
   (IJ
i.o  -
        0.5  -
        0.0
           0.0
                     0,5               1.0
                    Calculated Resistance, AT/q (oC/W)
1.5
2.0
Figure 4.1 Plot comparing the estimated resistance to heat transfer from the cell center to the
cooled surface of the module to that calculated by the FlexPDE model.

The average error in the estimated AT/q compared to the values calculated by the FlexPDE
model for the 70 cases studied was 6.0% and the maximum error was 13.0%. This accuracy was
deemed to be satisfactory in that for all practical battery designs, the error will be only a fraction
of a degree Celsius.

4.3.2 Heat Transfer from Module Wall to Flowing Coolant

Heat may be transferred  to and from the modules by flowing fluid directly across  the module
casing. For ease of communication, the focus of this discussion will be on cooling of the module
rather than heating for cold-climate operation. In theory, both liquid and air may be used as the
heat transfer fluid. However, the stacked cell design we have selected has minimal exposed area
relative to the overall volume of the cells. Our calculations have shown that liquid cooling is the
only feasible option for  this particular module design.  The superior heat  transfer of  liquids
(density and heat capacity) allows for implementation of this compact design without exposure
of the individual cells to the heat transfer fluid. A design of this kind should result in a lower cost
and higher energy density battery than a different design that cools individual cells.
                                            38

-------
The  model directly calculates the temperature drop  between the module wall and the heat
transfer fluid  for  a  set  pressure drop, fluid (coolant) temperature rise,  and fluid physical
properties. A 50/50 ethylene glycol, deionized water (EG/HiO) mixture was selected based on
the low cost and contemporary use in coolant systems. The default pressure drop was taken to be
10 millibar, but may be changed by the user if desired. The gap in which fluid flows is sized to
maintain the  target pressure drop  without going below a  minimum gap height of 3 mm. A
coolant temperature rise  of 1 °C was selected to establish a mass flow rate, but also may be
changed by the user.

Calculation  of the heat transfer coefficient  allowed for determination of the temperature
difference between the module and average  coolant temperature. A  schematic of the flow
passageway and change in temperature profile with distance is  shown below in Figure 4.2. The
outer wall of the flow passage is assumed to be perfectly insulated. The inner wall (module
casing) is assumed to have a constant heat flux perpendicular to the wall. Laminar flow was
assumed to simplify the calculation of the velocity profile (parabolic).
                           t
            t
t
Figure 4.2 Heat  transfer from the module wall to the laminar flow heat transfer fluid.  The
temperature profile of the fluid is shown at different lengths down the path.

Frequent use of dimensionless numbers  was necessary to adequately correlate the numerical
results into a generally useable  form. We define the Reynolds, Prandlt, Graetzl, and Nusselt
numbers  here for completeness.44 The Reynolds number, Re, is the ratio of inertial to viscous
forces. The Reynolds numbers were always less than 1000 confirming laminar flow. The Prandlt
number, Pr, is the ratio of the momentum diffusivity to thermal diffusivity. The Prandlt number
for the EG/H2O mixture is approximately 38. The Graetz number, Gz, is directly proportional to
the product of  the  Reynolds and  the Prandlt numbers. Moreover, the  Gz  value  is inversely
proportional to the distance down the fluid flow path, /, resulting in higher values near the  start
of the flow path.  Finally, the Nusselt number, Nu, is the ratio of the convective  to conductive
heat transfer. Here uave is the average fluid velocity, dn is the hydraulic  radius (twice the flow
gap), and // is the viscosity.  The heat  capacity, cp, thermal conductivity, k, and  heat transfer
coefficient, h, are the critical heat transfer values. The mass flow rate, G, and the width of the
channel, W, are the remaining parameters.
Re =
  Pr =
                                                                                    (4.5)

                                                                                    (4.6)
                                           39

-------
                               Gz = 2-L = 2RePr                           (4.7)
                                       kWl       I

                                       Nu=-^                                   (4.8)
                                              k
Coupled momentum and heat transfer has been solved previously by determining a number of
the eigenvalues for a series solution of a vast number  of various geometrical  configurations
related  to pipe, duct,  and parallel plate flow.45'46 We have chosen to reach the solution
numerically and then fit a correlation between the Graetz  number and the mean Nusselt number.
The empirical form provided by Nickolay and Martin provides an accurate means of correlating
the results over many orders of magnitude.47 The correlation, shown in Equation 4.9, relates the
Graetz number  and the limiting solution, Nuoo =  5.385, to the mean Nusselt number. Then the
mean heat transfer coefficient may then be directly calculated from Nu. Here n and C; are fitting
parameters.

                               Nu = [(Nu J* + (CjGz173 )" f                           (4.9)

The numerical model was solved with the finite  element method using FlexPDE software. We
note that the bulk or "cup mixing" fluid temperature in Equation 4.10, the average  temperature of
the fluid normalized by the fluid velocity profile, was necessary to reach the proper values.

                                                                                  (4.10)
The following important assumptions were used to reach a solution.

   1.  Flow of incompressible heat transfer fluid is laminar
   2.  Thermal diffusion is allowed up and down stream of the heat transfer (for convergence)
   3.  Boundary conditions: dT/dy = 0 at insulation; q = constant at module casing
   4.  Negligible radiative energy transfer
   5.  Steady state conditions reached

Figure 4.3 displays the temperature profile between  the module casing and the insulated wall for
various distances along the flow channel. The average temperature of the fluid has risen 1 °C at
the end of the flow path even though the maximum and minimum temperature is separated by
nearly 5 °C. The simulated change in average temperature down the  length of the flow channel
allows the calculation of the average heat transfer coefficient and thus Nusselt number. The
correlation, Eq 4.9,  determined from various  simulations conditions is shown in Figure 4.4. An
excellent fit is obtained  allowing for implementation of the correlation into the design and cost
model. This  correlation now enables  efficient and accurate calculations of the heat transfer
coefficient to be made in the spreadsheet informing the user of the effectiveness  of the thermal
management in the design.
                                           40

-------
                        Dimensionless distance
                            down fluid path
                      -1          -0.5          0          0.5           1

                        Dimensionless distance from centerline

      Figure 4.3 Temperature profile in the heat transfer fluid for various fractions of the

                               dimensionless path length.
                  100
                0)
                .0

                E
                3
                C
                =  10
                0)
                C
                «
                Q)
                          • Calculations

                         — Correlation
                       Nu« = 5.385

                       n = 3.592

                       G! = 2.255
                     0.1
1.0         10.0        100.0
   Graetz number, Gz
1000.0
Figure 4.4 Correlation of model simulation results relating the Graetz number and mean Nusselt

        number for laminar flow between an insulated surface and the module casing.
                                          41

-------
In general, the heat transfer from the module is improved by increasing the contact area and
increasing the fluid flow rate. The contact area may be increased by using cells with a higher
aspect ratio.  This  also results in a smaller temperature gradient within the cell as discussed
previously in section 4.3.1. Increasing  the fluid flow rate is accomplished by using a  lower
temperature rise and/or a larger target pressure drop. The gap height may prevent a change in a
single parameter from having a significant affect on the temperature drop. Physical limitations of
implementing a cooling  system should be considered when  moving to higher flow rates and
pressure drops. The user  should note that raising both of these parameters will increase the cost
of the battery design in  ways that the  model does not consider (e.g. more expensive pump,
increasing structural integrity, etc).

For high speed driving or very aggressive driving cycles, the temperature difference between the
surface of the cooled module  surface and the bulk of the  coolant may become fairly large
(>10°C). This coupled with the temperature rise within the cells could result in too high cell
centerline temperatures. This result can  be avoided  by controlling the inlet coolant temperature
as a variable  that is adjusted in a classic cascade automatic control system to control the module
wall temperature at the desired value. Thus, the temperature rise at the center of the cells will be
essentially held to that  resulting from  conduction within the cell and will  not be greatly
influenced by the temperature rise in the coolant.

4.4 Cooling and Heating Required to Maintain Pack Temperature

When parked in the sun for several hours, the internal vehicle temperature and, thus, that  of the
battery may become so hot that the life of the battery is reduced. To avoid this, the vehicle  air
conditioning  system  may  be actuated intermittently to  cool the  battery.  By allowing  the
temperature of the battery  to fluctuate by several degrees, it is only necessary to actuate the
cooling system about once per hour for a few minutes.  For a set of  target temperatures with a
difference  of 25°C and  with the default insulation thickness (10  mm) and default  thermal
conductivity  (0.00027 W/cm-K), BatPaC calculates the average cooling requirement to be about
60 W for PHEV-40 batteries. The performance coefficient of the vehicle air-conditioning system
might reduce the actual energy draw to less than half that, but heating  of the system outside of
the battery during the hour-long downtime periods would be a counter-acting factor. The BatPaC
model calculates the energy required for cooling of all types  of electric-drive vehicle batteries.
However, most HEVs may not have  electrically driven air-conditioning units and some other
method might be needed  to avoid very high battery  temperatures during parking such as thicker
insulation and fan cooling.

If the battery is to  deliver full power at startup, it must be at a temperature of at least 5°C. This
minimum temperature can be  maintained by heaters and circulation of the glycol  solution.
BatPaC calculates  the amount of power required to maintain the battery temperature for any  set
of battery and  ambient temperatures. PHEV-40 batteries would require about 50 W of heat to
maintain the battery temperature at 20°C above that of the ambient under steady-state conditions.
During recharging  this should be easily done for 20 hours at a  cost of $0.10 for an energy cost of
$0.10 per kWh. If  the vehicle is not at a source of power for recharging, limited energy (say 1-2
kWh) can be  drawn from the battery (if not blocked by a switch actuated by the driver) and then
                                           42

-------
automatically shut off after maintaining the battery temperature for one to three days, depending
on the ambient temperature.

4.5 Heat-up from Cold Ambient Conditions

All of the batteries for the various types of electric-drive vehicles will occasionally be exposed to
very cold temperatures, which will require  special heat-up procedures. All but the EV batteries
can be heated with the aid of the engine. This can be done with electric heaters operated from
power taken from the generator or from glycol solution from the engine cooling system. If the
latter, it might be prudent to isolate the engine coolant from the battery coolant by means of a
plastic heat exchanger.

Another method of heating the battery is by means of the electric heaters that should be available
for maintaining  the  battery temperature (section 4.4). BatPaC calculates the amount of heat
needed and the  time required  with  suitable heater  power.  For PHEV-40 batteries, about  15
minutes is  required with 2-kW heaters. This method of heating  will be slower than with the
engine coolant and even the latter would result in some delay before the battery is capable of full
power.

To avoid delay in starting vehicles from a cold startup, the driver could initiate heating by means
of a  remote device, which in the future may be a telephone. By this means, heating could  be
initiated either from heat drawn from the engine or electric heaters. Remote initiation of heat-up
would be especially important for an EV  away from  a charging station in  that no engine is
available to assist heating and the large size of the battery would result in a long heating period
with electric heaters. The BatPaC model estimates the time under these conditions.
                                            43

-------
                5.  Modeling of Battery Pack Manufacturing Cost

5.1 Approach

The manufactured cost of a battery pack is  calculated with input from the design information
generated in modeling the cell and battery pack performance. The design modeling determines
the annual materials and purchased items requirements. The manufacturing cost is then added to
these materials costs, along with a warranty cost, to reach the unit cost of a  single battery pack.
The manufacturing costs for the designed battery are scaled from a baseline plant. The baseline
plant was designed for a battery of intermediate size  and production scale  so as to establish a
center-point for other designs. The baseline plant accounts for the size, speed, number of units,
direct labor,  and depreciation of the capital cost for each processing step. These costs  are
adjusted  to meet the requirements for a plant producing the battery under study. The process
expenses are summed with the additional costs of operating the manufacturing facility. These
costs include launch costs,  working capital, variable overhead, general, sales, administration
(GSA),  research  and development,  depreciation,  and profit.  Additionally, the costs  for  the
thermal management, battery management  system,  and disconnects have been  estimated to
provide the total cost to the OEM for the integrated battery pack.

In this analysis, all costs  are evaluated for 2020 when large battery manufacturing plants  are
built. All dollar values are brought back to 2010 with allowance for inflation. In other words, all
costs and prices are in 2010 dollars.  Some materials and battery manufacturing  costs are  lower
than recent values, where we judged that processing improvements for high volume production
of materials would reduce costs.

The baseline manufacturing  plant was calculated for an  annual production rate of 100,000
batteries. The cost model accounts for different scales of manufacture by recalculating the costs
of each individual step in the manufacturing process. The changes in capital and operating costs
will change the calculated unit cost of the  battery pack. The parameters were determined to
provide reasonable estimates for manufacturing rates of 20-500 % of the baseline rate. Thus, for
a plant that is far different in size from the baseline plant,  for instance a pilot plant having an
annual production of only 5,000 battery packs per  year, the estimate from this study would be
expected to be less accurate than if determined in a study dedicated to that purpose.

To simplify the cost calculations, it was assumed that all hardware  items for the cells, modules
and battery will be purchased from a  vendor specializing in similar products. The costs for these
items were estimated to be a fixed value plus an additional value proportional to the weight of
the item, which is calculated during the battery design. In mature manufacturing plants in  2020,
toward which this study is directed, some items which are assumed to be purchased in this study
might actually be internally manufactured from raw materials. This would increase the number
of processing steps  needed  in  our  manufacturing simulation  and  thus complicate  the cost
calculations. Assuming that some parts would be purchased if they would actually be produced
from raw materials  would  tend to  underestimate capital  and labor costs and  overestimate
purchased items expenses. However, the net effect  would be a very small change to the overall
unit cost of the battery pack.
                                           44

-------
5.2 Materials Costs, Purchased Items, and Pack Integration

The  end battery pack  cost depends significantly on the cost of both the active and inactive
materials that compose the design. In this  subsection, the assumed material  costs and the
rationale behind them  are presented. We provide for  means to scale the materials cost with
production volume using the same method used for processing rates as discussed in section 5.4.
In general, the materials costs will be largely insensitive to production volume  since we have
assumed a  high volume market  already exists. Only the negative and positive electrode  active
materials are assumed  to have a minor benefit for larger scales of production. While we state
suggested materials costs and sensitivity to production  scale, the users of the cost model may
enter any value that he desires.

5.2.1 Battery Specific Materials Cost

The  largest contributions to the materials cost of the battery are from the following components:
positive and negative electrode active material, separator, electrolyte, and current collector foils.
The  choice of the materials often defines the size and performance of the battery as well  as the
cost. Many different variations  of materials are possible  in the Li-ion family  of chemistries.
However, we have chosen to focus on the different available positive electrode  materials with
less  attention on the negative  electrode. This reflects  the current research and  manufacturing
activities. The separator and the electrolyte are also both active areas of development. However,
the following battery  designs are  based  on a single  electrolyte and separator combination.
Including the cost and effect of  additives  and enhanced separators is  beyond the scope of this
work. The user is always able to modify the dimensions, cost, and ASI that may be required to
account for changes in these materials.

The  price of specific battery materials is of some debate. The values presented in Table 5.1
compare our suggested costs to those reported recently in the open literature. Our  values, as well
as the others in the table,  are  derived from conversations with material, cell, and original
equipment manufacturers. The sources are commonly anonymous and the accuracy of the values
is generally unknown. We present the comparison of published values so that the user of the cost
model may appreciate the accepted range of values for commonly used materials.

5.2.1.2 Positive Electrode Active Materials

The cost of positive electrode materials is driven to a large extent by the cost of the raw materials
from which it is made. The archetype Li-ion positive  electrode material,  lithium cobalt  oxide
(LCO),  was the original material commercialized in Li-ion  batteries for consumer electronics.
LCO has many excellent characteristics but is not considered a viable choice for use in Li-ion
batteries for automotive applications. One of the largest drawbacks of LCO, other than  safety
concerns, is the high and volatile cost of the cobalt. While tolerable in the consumer electronics
market, the cost is too high for use in an automobile  battery. Many other materials are in a
commercially viable state of development and are currently utilized in Li-ion batteries produced
today (Table 5.1) such  as lithium manganese spinel oxide (LMO) and lithium nickel manganese
cobalt oxide (NMC).3'6 The relative advantages and disadvantages of each material will not be
discussed here.
                                           45

-------
S ^ 2
Q ^ OO
* 


in
(N
(N
(N
in
in
i
^




o

ON



5
o
(D
ON (D .-., S
(D  
toOtoOtoOtoOtoOtoOtoOtoOtoOtoOtoO ^  ^g  ^-H
 o
in
•°,
•o"
_o
•_
a
(U
CO
£H
'o
-J— »
8
o
CO
D
T3
O
-4— *
cd
O
£

-4— »
CO
O
O
jO
CO

O
CO
-4— »
(U
Q
2
CS
H
Abbreviation

hemistry
U








Material
0
s
'si'
o
X
~S
a"
j


CD
o

cd
1
CO
CD
CO
CD
03
toO
03
s
LNMO
0
wo
1
wo
0
Z
,_)






*
CD
O
03
O
13
CO
in
!

2
CD
,_)




CD
O
-s
03
O
CD
Phospholivi
U
Z
(N
0
0
CD
^
CD
O
U
0
OO
CD
Z
,_)



*
CD
1
O
CD
O
-a
§


U
S
z
(N
0
wo
O
6*
U
CO
a
S
s
q
,_)



*
CD
"S
.g
s
o
CD
O
1
i— H
^^
^^
U
S
z
(N
0
wo
O
U^
si
q
,_)



*
CD
"S
4^
s
o
CU
O
1
03
U
S
z
1
(N
O
N
U
N
j?
CO
O
1
7<

*
CD
O
03
O
"S
(H
CD
03
^
1
O

(N
0
o
y
3




CD
1
13
o
CD
X
O
1
03
O


u"







1
Graphite an
O

(N
<5
wo
E^
3





•S
o
03
'co
Titanate spi
U
U
a
£*
°""^ Cu
H^ ^^
5 ^
^ SH
(N Q"
~ £








Electrolyte
Separator


CD
&
&
O
U





r^
o
o
8
u


luminum
<;





r^
o
o
8
u
o
U
1
o
o
'o
o
"O
-Q
O
U
o
CC
1
2
'ecipitation o
OH
6
C*
Ik
.3
o5
CC
"O
o
( \
* The cost of (



_u
3
'c/3
&
o

r
O
G°
'G
u
** Cost repres

-------
The amount of cobalt and nickel, as well as ease of manufacture, controls the end price for a
positive electrode material. For example, the NMC-441 is less expensive than the NMC-333 as
the cobalt quantity is significantly reduced. The market price for cobalt and nickel metal varies
dramatically  from year  to  year. Reducing  the quantities  of these  materials in the positive
electrode will reduce the total price and price volatility. Researchers at TIAX LLC have treated
this variation and shown the significant effect  on  end  battery cost.   The average traded metal
prices for the last 20 years is 48 $/kg and  15  $/kg for cobalt  and  nickel  respectively. These
numbers are based on historical prices for the metals as  collected by the United States Geological
Survey (USGS).48 The metal  prices are indicators for how the  intercalation material cost will
relate when compared to one another. The fact these materials are not earth abundant means they
will not benefit as much as other materials from increased scales of production.

We employ the relationship in Equation 5.1 to systematically calculate the cost of the transition
metal based spinel and layered compounds. The final cost, C, of the lithiated oxide depends on
the baseline cost, Co, and the contributions  of the lithium and transition metal raw materials, Q.
The molar stoichiometry, Xi, is transformed to a mass basis with the molecular weight of the raw
material, MWi, and the final product, MW. The baseline  cost is the sum of the cost for processing,
additional raw materials, and profit margin associated with the manufacture of the materials. We
assume a baseline cost of $7/kg for single metal containing oxides (LMO and LCO) and $16/kg
for the co-precipitated metal oxides such as NMC-333  and NMC-441. NCA is known to have a
slightly lower yield and requires additional raw materials resulting in an assumed Co = $20/kg.
The costs  for Li, Ni, Mn, and Co are taken to be 0.22, 0.85, 0.15, and 5.00 $/mol respectively.
Aluminum is assumed to be similar in cost to manganese for these calculations. One may directly
translate these numbers to raw materials costs resulting in $6/kg  for I^COs, $5.5/kg for NiSC>4,
$32/kg for CoSCM,  and $l/kg for MnSO/t. Calculations are also shown in Table 5.1 using
$2.5/mol for cobalt as a simple demonstration of the effect of cobalt on the end material cost.
                                                   x{C{MW{                          (5.1)
In general, earth abundant elements  should be the dominate  transition metals used if a low
positive  electrode cost is desired. Both  iron  and manganese are abundant and inexpensive
transition metals  for  intercalation materials.  Comparison  of the iron phosphate,  LFP,  to
manganese spinel, LMO, reveals how processing costs contribute to the end price of a material.
LMO is relatively easy to manufacture. In contrast, LFP requires a reducing atmosphere and a
carbon coating step to reach the  end product.  The increased complexity in the manufacturing
process is realized in  the price.  However, one could argue that the  manufacturing cost will
decrease with increased knowledge from larger scales of production.

5.2.1.2 Negative Electrode Active Materials

While several negative electrode materials exist for Li-ion batteries, carbon materials in the form
of graphite and/or hard carbon are still used in the vast majority of commercial cells. Graphite
offers the greatest energy density while hard carbon is said to  enable high  rate capability with
decreased risk of lithium plating (an undesired  side reaction)  during high charge rates. We have
chosen synthetic  graphite as  a generic carbon electrode in our model. The  price of  graphite is

                                            47

-------
much better understood  than  that  of most  of the positive  electrode  materials.  However,
significant differences in cost and performance will exist between synthetic, natural, and coated-
natural graphite. The method of production and necessary heat-treatment will control the end
cost. Graphite, although in different purity grades or microstructure forms, is  used in many
industries. This is in stark contrast to the positive electrode materials.

The lithium titanate electrode, LTO, offers an interesting  option compared to graphite.  Unlike
graphite, LTO operates within the stability window of the electrolyte. The higher  electrode
potential,  1.5 V vs Li,  dramatically reduces or eliminates the formation of the solid electrolyte
interphase (SEI). As a result, nanoparticle-based LTO may be implemented without concerns of
increased  side  reactions with the electrolyte.  The reduced dimensions increase the  available
surface area for reaction while simultaneously shortening  the diffusion length.  Both of these
factors combined with the lack of SEI dramatically reduce the impedance of the electrode.

5.2.1.3 Electrolyte and Separator

The  electrolyte used in this model is based on  a  lithium hexafluorophosphate salt,  LiPFe,
dissolved  in a carbonate  based solvent system. The carbonate solvent system is  a blend of
ethylene carbonate, EC, and a linear carbonate such as ethyl methyl carbonate, EMC, or dimethyl
carbonate, DMC. Other chemical additives may be used to lower the capacity and power fade of
the battery over time.  Polymers  may be added to the electrolyte as either a minor or major
component. This is not discussed  in any further detail in this work. The price of  18 $/kg, about
22 $/L, is only for the base electrolyte (i.e. no additional additives).

The  separator is  typically a porous membrane based on polypropylene  (PP) and sometimes
includes a polyethylene (PE) middle layer. PP and PE are very inexpensive raw materials and
thus the suggested cost of $2/m2  is in large part due to the manufacturing process required to
form the porous network in the membrane. As competition and scale of manufacture increase,
the prices of the separator may fall closer to $l/m  . However, the cost of improved technology
                                                                             ij
may offset some of this cost reduction, so we have retained our cost estimate of $2/m .

As safety is a  major concern for Li-ion batteries, the separator plays a key role in isolating the
oxidant from the fuel. If the two charged electrodes contact each other (short), then a run-away
reaction is possible. Separators have been designed to "shut-down" or melt at key temperatures.
The middle PE layer is the shut-down feature in our proposed separator. Ceramic coatings have
also  been used to ensure structural  integrity.  Many other approaches are being developed to
increase the safety of Li-ion batteries.  The user of the cost  model  should account for the
increased technology in the price and dimensions of the separator as needed.

5.2.1.3 Current Collector Foils

The current collector foils are based on copper metal for the negative electrode and aluminum for
the positive electrode. However, the LTO anode material, because of its high voltage relative to
lithium, enables the use of aluminum as the negative electrode current collector. The price of
these foils is based on raw materials and manufacturing costs. The aluminum foil is produced by
rolling of thicker stock foils into thinner and thinner sheets. On the other hand, copper foil is
                                           48

-------
more likely to be produced through an electrodeposition process. The foils are 12 microns and 20
microns thick for the copper and aluminum current collectors respectively. The foils used in
batteries have additional requirements beyond the cheapest product available. Surface treatments
are often necessary to promote adhesion of the electrode to the foil surface. In addition, alloying
of the foil may be necessary to achieve the required material properties for long life.

The raw material contributions  to the foil price will vary with the volatility of the market price
for the metals.  Figure 5.1 displays the metal ingot price contribution on a  $/m basis. These
numbers are based on historical prices for the metals as collected by the USGS.
48
The  values  for  both aluminum and  copper tend to vary  significantly over the time period
examined. The price for copper is more volatile and always more expensive than aluminum.
Analysis of Figure 5.1 reminds the user of the cost model that cost quotes are only valid for a
short period. As the market price  for raw materials  changes,  so will the price for the finished
product.
                                                                                     r\
Conversations with manufacturers  and suppliers lead us to take a price of 1.80 and 0.80 $/m  for
battery grade copper and aluminum foil respectively. We point out that the current metal  ingot
price is only a small contribution  to the end foil price being about 16 % of the aluminum foil
price and 23  % of  the copper foil price.  Thus,  a  doubling of the ingot prices would only
moderately increase the foil prices.
u.au •
Don _
.OU
*E
t/3- n vn -
O
2 n en -
0>J U.DU
_c
•** n £n -
^\>CrO^c,^^JJ-^

85 1990 1995 2000 2005 20
Year











10
Figure 5.1 Metal ingot cost contribution to the current collector foils over a 20 year period. The
average for that period is also shown. All costs are in 2010 US$.
                                            49

-------
5.2.1.4 Additional Electrode Components
The binder and conductive additive in the positive and negative electrodes add a small but real
cost to the battery. The conductive additive, more common for the positive electrode, was priced
at 6.80 $/kg  for a high purity and moderate surface area carbon  black material.  The binder,
perhaps PVDF or CMC based, is assumed to be  10  $/kg. The W-Methyl-2-pyrrolidone (NMP)
solvent for the PVDF binder is estimated to be 3.20  $/kg. Most  of the NMP is recovered after
evaporation and recycled as discussed in section 5.3.3. Only the small amount lost in processing
need be replaced. No cost is assumed for water used to in the electrode slurry processing.

5.2.2  Purchased Items Cost

Table 5.2 lists  the purchased items for the cell module and  battery jacket.  The cost of a SOC
controller for each cell, or group of parallel cells, is $2.50 plus a small factor for the cell capacity
(Ah), which allows for higher cell balancing currents  for larger  cells. The other components cost
a fixed amount plus an additional factor, which is proportional to their  mass, mi.  The cell
negative terminal and parallel cell group connection are both  made from nickel plated copper
sheet and thus have the same cost equation. The costs  shown for the terminals  include an
allotment for  isolation tape that is necessary to protect the electrical connection. The bus bar is a
fixed cost and is  only charged if a single  row of modules  is  used.  A single  row of modules
requires a bus bar in order to locate the positive and negative  terminals at the  same end of the
battery.

Table 5.2 Cost equations for purchased items
Component, /
SOC controller
Cell positive terminal
Cell negative terminal
Cell container
Aluminum heat conductor
Parallel cell group connection
Module terminals
Balance of module (casing)
Module interconnect
Battery terminals
Bus bar for one module row
Battery jacket
Cost Equation, $/unit
2.50 + 0.01C
0.25 + 4m/
0.25 + 6nn
0.20 + 3mf
0.10 + 4mr-
0.25 + 6mt
0.75 + 5nn
1.00 + 3mr-
1.00 + 5mr-
15.00 + 0.02/toto/
20.00
30.00 + Imi
Cost per unit
cell or parallel cell group
cell
cell
cell
cell
parallel cell group
module
module
module
battery pack
battery pack
battery pack
5.2.3 Pack Integration Cost

Various additional components and thus  cost are necessary to integrate the battery into  the
electric drive system, which adds cost. While it is not clear what should and should not constitute
the cost of the "battery pack," we present these additional items  in Table 5.3 in an attempt to be
complete. The model treats these values as a cost to the OEM for integrating the battery into the
vehicle. After all, the price of the entire system is of interest to the final consumer of the product.
                                           50

-------
The general conclusion is that the pack integration costs have the largest consequence for the
smallest batteries. The worst case is perhaps that of a small PHEV10 battery. The integration
costs of a PHEV10 battery carry the burden of charging from the grid, but provide only a modest
electric drive benefit. The fixed cost of pack integration may amount to 25 % of the battery pack
total even without  considering the  costs  of additional  powertrain  components.  Clearly,
understanding the entire cost of the electric drive system is  of importance to evaluating the true
value of the electrified vehicle to a consumer.

Table 5.3  Costs to integrate battery pack into vehicle drivetrain. $/kW numbers reflect maximum
kW of cooling or heating required.
          Battery Management System     MicroHEV    HEV-HP   PHEV & EV
              Current and voltage sensing, $       40          70          100
              Module controls, $/module          10          10          20
          Disconnect Units
              Auto, disconnect, $                50          70          200
              Manual disconnect, $               15          15          15
          Thermal Management System
              Baseline thermal system, $          30          80          120
              Additions to AC system, $/kW        40          40          40
              Heating system, $/kW              20          20          20

5.2.3.1 Battery Management System

The battery management  system (BMS), in  our assumed battery design, integrates the modules
and battery into the overall electric drive system. The BMS includes measurement and control
features such as the following:

   •   Measurement of battery pack current and voltage
   •   Balancing of the module  voltages (cell balancing done within module)
   •   Estimation of battery pack state-of-charge (SOC) and state-of-health (SOH)
   •   Estimation of module SOC and SOH
   •   Monitoring and signaling of battery thermal management

The cost of the BMS will scale with magnitude of battery current  and with the need to charge
from the electrical grid. Therefore the  PHEV and EV batteries will have  a higher burden from
the BMS.  The  micro-HEV is assumed  to have less complicated management and thus less cost
than the HEV-HP.

5.2.3.2 Manual and Automatic Disconnects

The manual and automatic disconnects integrate a high-level of safety and electrical management
into the electric drive system. The manual disconnect breaks the current flow pathway from the
high-voltage terminals to the outer system allowing for the safe service of the vehicle and battery
pack. This disconnect is designed to be operated when the electrical system is de-energized. The
automatic  disconnect is much more complex. This unit contains the connections for the high-
voltage  system to the rest of  the vehicle's electrical system: drivetrain,  grid  charging  (if
applicable) and accessories (high and  low voltage). Fuses are present as a hard-wired safety

                                           51

-------
device to prevent unusually large current  spikes  from damaging  the  battery or drivetrain.
Multiple contactors are used to appropriately channel electrical current depending upon normal
operation or grid-charging. Engaging the contactors requires that multiple safety interlocks are
established including isolation of the high voltage bus from the vehicle  chassis and an inertia
based sensor (crash protection). Finally, a small circuit is provided to  prevent arcing of the
current across the high-voltage contactor when the high-voltage circuit is closed.

The relative cost of the automatic disconnect amongst the various battery designs is driven by the
pack voltage, maximum battery current, and the need for charging from the grid. The voltage of
the pack has a significant effect if a  42 V  micro-HEV pack is considered. For this system,
electrical safety regulations allow a less complicated system to be used. Requiring higher battery
currents generally increases the cost of electronics and conductors. The additional complications
arising from grid-charging adds a significant additional cost to the PHEV and EV systems. It is
unclear  to the authors at this time what other factors may enable a lower burden  of external
safety  controls.   These  additional  costs  in  the automatic  disconnect  unit  have the most
pronounced effect on the cost of smaller batteries, as the burden amounts to a significant fraction
of the total cost.

5.2.3.3 Balance of Thermal Management System

The thermal management of the battery is crucial to meeting the life and safety  requirements of
transportation applications. The complexity of this system must be minimized  to reduce the cost
and size burden on the vehicle.  Our assumed design format uses liquid thermal management for
all vehicle  battery types. In practice, current microHEVs and  HEV-HPs  are more likely to be
cooled by blowing air. Air-cooling is  generally  less expensive than  liquid cooling, but is less
effective,  requires larger  system volume,  and  may result in substantial background noise.
Furthermore, directly cooling the pouch cells with air increases the  magnitude of oxygen and
water permeation through  the seals resulting in deleterious effects to  the fifteen year life of the
battery.  Air cooling is not  feasible with our current assumed battery format (stacked cells sealed
within module). The cells would need to be separated to allow air to flow past at least one side to
achieve sufficient heat transfer as earlier versions of this model had used.1  Future versions of the
model may include an option to select either air or liquid cooled. However, the main  goal of this
model is to explore the effect of battery performance  and materials chemistry on the price of the
battery to the OEM.

A single  refrigerant  compressor  is  used for  both the  cabin  air   and the battery cooling
applications. Likewise, the same radiator and fan as the cabin cooling will also be used for the
battery cooling refrigerant. Most OEMs appear to use an electric  compressor for all full-HEVs
(HEV-HP)  and  PHEVs/EVs  that are liquid cooled. The incremental cost  for  an electric
compressor at high volume in year 2020 will  likely be $200-300 more than the commonly used
$100  belt  driven compressor.  We do not include  this incremental  cost in our thermal
management system cost; however, we state it here for completeness. The additional cost to the
compressor for the battery cooling capacity is insignificant compared to burden of transitioning
to the electric compressor.  Experts in the field have informed us that the electric motor and high-
voltage invertor are the largest contribution to  the incremental cost of the electric compressor.
                                           52

-------
An expansion valve on the refrigerant line and a heat exchanger (chiller) transfers the thermal
energy from  the heat transfer fluid to the refrigerant loop. A 50/50 Dl-water/ethylene-glycol
solution is  selected as the heat transfer fluid. The assumed battery design has the heat transfer
fluid pumped over the module casing  to transfer heat from and to the cells. The battery may be
heated by either a positive thermal coefficient (PTC) or flexible mat heater. The PTC heater
would directly raise the temperature of the heat transfer fluid in a reservoir while the matt heater
would be placed under the battery jacket insulation.

The PHEV and EV batteries will likely have both active and passive thermal management modes
requiring some additional monitoring and an electrically  actuated  valve. We have assumed
decreasing  cooling costs  for the HEV-HP and microHEV systems without explicitly dictating
where the  savings originate.  In general, one would expect  smaller batteries to  have a less
complicated control system, lower flow rates and possibly even direct cooling of the cells with
the evaporator.
5.3 Baseline Manufacturing Plant

The baseline plant is designed to produce 100,000 NCA-Gr baseline battery packs per year. The
baseline battery pack produced by the plant has sixty, 40-Ah capacity cells, providing a total
pack power of 50 kW and total energy of 8.7 kWh. The battery will power 20 miles of vehicle
travel at 70% of the pack energy and 300 Wh/mile. The schematic diagram of the plant (Fig. 5.2)
is designed to illustrate the flow of materials through the plant and the relative floor areas for the
processing steps rather than representing a realistic plant layout. The overall manufacturing rate
of 100,000 battery packs per year is achieved by operating for three shifts at the equivalent of
300 days per year of fully effective production. There will be more than 300 days of operation,
but some days will have less than  100% effectiveness. The exceptions to three-shift operation are
the Receiving and Shipping sections, which are active for only two shifts per day. The cost
factors for the individual manufacturing steps in the baseline plant  are summarized in Table 5.4
and discussed in detail in the sections that follow. Most of the operations are carried out with
normal factory  atmosphere,  but the cell assembly process steps are completed in a dry room
atmosphere.
                                           53

-------
     Receiving
                    Electrode Materials
                       Preparation
                   >

                      Electrode
                       Coating

                      Solvent r.v.--1.-"-- Solvent
                    Evaporation
                                i

Recovery
                                                                  Calendering
Electrode
 Slitting
                                                              Vacuum
                                                               Drying
   Shipping
   Cell and
    Scrap
   Recycling
              Battery Pack
               Assembly
              and Testing
            Control
           Laboratory
            Module
           Assembly
 Charge-
Retention
 Testing
                                          Formation
                                            Cycling
 Final
  Cell
Sealing
                                 Air
                                Locks
                                                        Materials
                                                        Handling
  Cell Stacking
                                   Enclosing
                                    Cell in
                                   Container
           Current
          Collector
           Welding
Electrolyte Riling
and Cell Closing
                Assembly Route
                      Dry Room
                                    Outdoor dry-room air
                                    processing equipment
    The areas in this diagram for each processing step are approximately proportional to
    the estimated plant areas in the baseline plant.
Figure 5.2 Baseline lithium-ion battery manufacturing plant schematic diagram. Manufacturing
rate: 100,000 NCA-Gr battery packs per year, 50-kW pack power, 40-Ah capacity, 60 cells per
battery. Operating year: 300 days with three 8-h shifts (two shifts for receiving and shipping)
5.3.1 Receiving and Shipping

These operations require the moving equipment and  storage facilities common to  any such
factory facilities. The Receiving section handles slightly less than 6,000,000 kg of materials per
year and also has facilities  to handle and  store some of the electrode  materials  in a dry
atmosphere. The Shipping  section is  required to enclose  the battery packs in crates,  which
requires some automated equipment and more labor than is required for Receiving. Shipping also
handles  about 400,000 kg of scrap each year, which is broken down and prepared for shipping in
the Rejected Cell and Scrap Recycle section.  The estimated resources needed for the Receiving
and Shipping sections are shown in the table below.

Receiving
Off-loading
Moving
Storage
Shipping
Rate Factor
870,000 kWh/y
870,000 kWh/y
Direct Labor
3 per shift
6 per shift
Capital Equip.*
3.60 mil$ total
0.60
1.20
1.80
5.0 mil$ total
Plant Area, m2
900
900
''Total cost including installation
                                           54

-------









£>
cp
03
CD

CD
CL
J/3
"CD
O
0

•e?
CO
Q.
cO
O
<(
i
i_"
CD

O
CL

B

CQ
>
"^
i
O
in
CO
/IN










'u>
c
'Q.
Q.
Ic
CO
T3
CO

;|
CD
£>
£
M—
fr
T3
CD
«
lc
CO
o
i
CO
jtz
~«
i_
o
I
oo
Q)
Q)
£



CQ
0)
^
c
CQ
Q.


^j
s
CZ

Q
'3
w
d
O





k.
o



0)
k.
Q










0)
CQ
O
CO

c
^
Si
CQ
m

CD
75




0)
=
CQ
m

CD
CQ
O
CO

0)
"cu

CQ
CO
7s
CD
J5
P
^




0.
^
O LO CDCD OOOOCD CDCDCDCDCDCD OOOOCDCD^OOCDCDCDCDCDCD
tt O OO OOO OOOOOO OOOOOOOOOOOO
CQ
"•
CM O OO OOLO LOLOOOOO OOOOOOOOOOOO
co oo LO LO eg eg eg o o o o OOOOOOLOOOOOO
„
CO
CD
0. « {
c\T c\T 5-~ 5-~ 5-^
O co r- r- do'co r~- r~- r~- r~- r~- r~- "oor-r-co"r-r-r-c'?i--i--
tt d do "-'"-'d dddddd S-ddddS-dddSdd
tl 00 00 on on ~
l£ °' °' odd

£8 88 888 888888 8888888K8888
:= c\j co
I


„
O^" LOLO LOLO1^ LOLOh-LOLOLO h-h-LOLO^I^LO^LOLOh-LO
•JJO OO OOO OOOOOO OOOOOOOOOOOO

^ o o o o o o o o o o o o o o o o o o o o o o o o
^» o o o o o o o o o o o o o o o o o o o o o o o o
*- •*" •*" •*" oo" oo" •*" •*" i-~" oo" oo" •*" oo" CD" CD" ^- CD" •*" i-~" •*" ••- co" co" CD" oo"

O
X

Ji
|CCO CNJCNJ •^••^•CNJ C\JT— •^••^•CNJ'^- LOLOCOLOCNJOOCNJCOCDCDLOCD
C/)


"^n *t
DC 03 03 (Tj
CD B B ™ w






LO
eg
-
1







3
OJ







o
CD





8









£*~

Q
8^
CJ5
CO
eg"
o
w
CC^
„
"ro

i
| .
$
O
15
C
§
.c
"CD
|
.c
°
o

Q
§•
CD
r-
c
-^
CD
(0
o
• w
1 0°





J2
1
"5.
"5
E
W
0
O
(-

CO
QI
w
3
03
O
03
a p
CD ^
•^ cc
1 I
" 00
£ 2.
— SE

O) ^
•i O°
^
£ "
"to
.^ o
B
ra =5
[^
Q
CD
i 1



^_,
£
si
2

"0
O
^
O^
+-
Q

0)
.^
"0
w
"S
.c
"o

0
.0
«
o
<
o
<^-

^
w
_c
"Q
g1
g

p
£
,2
rn










CO
P


1^-
03
"
p
c|

^
o
II
to
3
w
[I
^

5
Q_
d
CD
£
•2
"CD
_w
co
_^-
0


_1_I
C
0
"u
03
.±±
£•
_Q
CO
0
4-I

§
CD
.g
"CD
_ra
CD
.C

'o
§
o
.2
^
"§
E

o
£
g
£^
£
£
03
Q
**—
.B
'S




CO
2
^
^
^
of
g^
o
O


o
O
CD
a
O"l
d
^
"^
T3
CD
•|
2.
"§

~-^
CD
-g
O
^
2
_Q
E

c
CD
1
                                                Illll      sssssss
 CU

•i-H

13
 05


X)


^



 o


r^H
3
^: "o

O o
CO O

Q_ CQ
    CO

    §.
       CO

      CQ
      O  CD

      °o
                    81
                    CO
                   m CD
a a     o  o z

Q) Q)   °i- °i- Q)
.^ ^     E  E ^


•^ ^    (3) (3) t—
eg r-    co co eg
LO co    ooo
                            5! 8
                            Is- O
 O  O  O  O  O -C


"E "E "E "E "E i


O) O) O) O) O) CD
CO CO CO CO CO T—
OOOOOO
                                                                   cncncncnocncncn
                                                                   ooooooooooooooo
                   cncD
                   oo-^
                                    cg

                                    oo"
                    O)O>O>O)O>O>    IDl^l^LOOJLOLOLOOOLOO)
                    O  O  O  O  O CD    i— i— i—  i—    i— i—  i— O O  i— CD
                    OJOJOJOJOJOO    COCOCOCO    CO CO  CO O T—  CO 00

                    oo"  oo"  oo"  oo"  oo"       CD" CD" CD"  CD"    CD" CD"  CD" CD"    CD"
CD ^
w o
	 o
                                                                      =i  « ° «
-21 s  1  a  .^ -e i
                           ;5-2o
                              ~ "-
                                        CD
                                              CD  ±=
                                                                    E  & | c
                                                                    Q. O Q. E

                                                                    ^_ CD "5 w
                                                                                                                         8
                                                                                                If &~! f  *f J
                                                                                                11=1^ISS
C/5
—
2
 ca
H
Receiving (two-shift o
Materials preparation
Positive electrode
Negative electrodf
Electrode coating
Positive electrode
Negative electrodf
Solvent recovery
Calendering
Positive electrode
Negative electrodf
Materials handling*
Electrode slitting
Vacuum drying
Control laboratory
Cell Assembly in Dry
Cell stacking
Current collector wi
Enclosing cell in cc
en
c
l^
B
>,
Q
'o
-CD
(5
2
"c
0
o
1
£•
Formation cycling
Final cell sealing
Charge retention test
Module assembly
_l_i 1—
P O
^ w
CD
W T3
W C
03 03
v ^z
0 "K
<5 «
Q-T3
£-B
B 8
!_L.
t±
Ic
w
O)
c
"o. _
Q. 03
Q.
W
CD
.C
'o
•E
!c
CD
*The baseline cost fo
the positive electrode
CD
2
$
o
w
"ra
^
**The baseline costs
capacity, the cost is
***The baseline cost
number of modules (
                                                                55

-------
5.3.2 Electrode Materials Preparation and Delivery to Coating

The electrode materials,  which consist of active material, carbon  (if necessary), binder and
binder solvent, are well mixed in small  batches in portable tanks. At the design production rate
in the baseline plant, each shift requires three tanks each holding about 1000 liters of positive
electrode material mix and three tanks each holding  about 900 liters of negative electrode
material mix. The section must be capable of exceeding  this design rate of production by at least
25% to catch  up in case of unscheduled downtime in Materials Preparation or in  some of the
section immediately following  that section.  The tanks  of prepared materials are moved to the
Coating section  and pressurized to push the coating paste  into the coating mechanism.  The
estimated resources needed are the following:
Materials Prep.
Positive
Materials
Storage tanks
Mixing tanks
Moving equip.
Negative
Materials
Rate Factor
1,71 0,000 kg/y
active material
1,2 10,000 kg/y
active material
Direct Labor
2 per shift
2 per shift
Capital Equip.*
2.0 mil$ total
1.00mil$
0.50
0.50
2.0 mil$ total
Plant Area, m
600
600
*Total cost including installation

5.3.3 Electrode Coating on Current-Collector Foil

The positive and negative electrode structures are formed by coating both sides of the current
collector foil. In the baseline plant, the coating lines are 1.5 meter wide continuous roll-to-roll
coating processes carried out at a line speed of 10 m/min.  The first set of coating and drying
stations coats one side of the current collector foil, drives off the solvent in a heated oven, and
turns the foil over while transferring it to a second set of stations. The second set of coating and
drying stations applies and  dries the remaining coating before the coated foil is wound into a
large roll at the end of the line. An advanced alternative would be to run the foil directly into the
calendering process. The negative and positive coating lines are very similar. However, some of
the negative material  is coated only on one side to provide the electrodes at the end of the cell
stacks. For the baseline plant, a total of 8,170,000 m2/y of coating (annual cell area) is required
for the positive electrode (slightly more for the negative electrode), which allows for the 5% loss
of cells expected to fail testing and inspection. A larger area of foil than the coated area must be
fed to  the coaters to  allow  for the  part of the foil that is not coated so as  to provide tabs for
welding to the terminals (about 10%) and to allow for trimming losses during electrode slitting
(8%). Also,  about  30% excess  coating capacity must be provided to allow  for unscheduled
downtime. Only one coating line is needed for each electrode type to meet these needs. If one
coating line breaks  down, the other coating line may change over temporarily to coat the other
electrode material.

The oven sections of the coating line are designed to dry coatings about 100 microns thick at the
coating speed of 10 m/min.  A thicker coating will require longer ovens at additional capital cost
which is provided in the adjustment of costs discussed in section 5.4. For the  same annual area
                                            56

-------
throughput, a coating line that coats both sides with a 300-micron coating would cost $9,500,000
rather than the $8,000,000 cost for the 100-micron coater. The binder solvent for the positive
electrode in the baseline plant is NMP, which must be recovered by condensation and recycled.
About 0.5% of the binder solvent is combusted with a thermal oxidizer and must be replaced. For
the negative electrode the binder is water, which need not be recovered. The estimated resources
to meet these needs are the following:
Electrode Coating
Positive Electrode
Uncoated area
Width of coater
Coating speed
Number of coaters
Maximum rate
Excess capacity
Negative Electrode
Solvent Recovery &
Oxidation
Rate Factor
8,170,000
m2/y cell area
8,170,000
m2/y cell area
1,527,000kg
NMP/y
Direct Labor
4 per shift
4 per shift
2 per shift
Capital Equip.*
8.0 mil$ total
18%
1.5m
10 m/min
One
13,000,000 m2/y
30%
8.0 mil$ total
3.0 mil$ total
Plant Area, m2
750
750
225
*Total cost including installation

5.3.4 Calendering

The materials leaving the coating lines may be stored on large rolls (see next section).  However,
typically the materials leaving the coaters would go directly to the calendering process in which
the coatings are compressed by rolling to meet the  specified void volume fraction, which will
later be filled with electrolyte. The calendering equipment must match the output of the coating
equipment producing 8,170,000 m2/y of cell area with a maximum rate of 13,000,000 m2 of foil
per year to meet  contingencies as in coating.  We estimate  three workers are necessary  to
collectively operate the two pieces of equipment. The estimated resources to meet these needs
are the following:
Calendering
Positive Electrode
Negative Electrode
Rate Factor
8,170,000
r\
m /y cell area
8,170,000
m2/y cell area
Direct Labor
2 per shift
1 per shift
Capital Equip.*
1.0mil$ total
1.0mil$ total
Plant Area, m2
225
225
*Total cost including installation

5.3.5 Inter-Process Materials Handling

For all processes (Fig. 5.2), work in progress must be transported and occasionally stored to
permit nearly-continuous operation of the equipment. Storage areas must be provided both inside
and outside of the dry room. Raw materials must also be moved to the processing sites, which for
those in the dry room means through a separate air lock for materials transfer.  One-third of the
                                           57

-------
total space for Inter-Process Materials Handling is within the dry-room for the baseline plant and
also for all other plants. The estimated resources to meet these needs are the following:
Materials Handling

Rate Factor
8,170,000
ij
m /y cell area
Direct Labor
4 per shift
Capital Equip.*
1.5mil$total
Plant Area, m2
900
*Total cost including installation

5.3.6 Electrode Slitting

The coated electrode foils are slit into strips between the coated sections and then into individual
electrodes as shown in Fig. 2.3. The estimated scrap loss of foil for this process is about 8%. The
estimated resources to meet these needs are the following:
Electrode Slitting

Rate Factor
8,170,000
r\
m /y cell area
Direct Labor
4 per shift
Capital Equip.*
2.0 mil$ total
Plant Area, m2
300
*Total cost including installation

5.3.7 Final Electrode Drying

In the absence of electrolyte, no harm is done by exposing the electrodes to normal factory air;
however,  the electrodes  must be dried by  heating under  vacuum prior to cell assembly.
Maintaining  extremely low moisture  conditions during cell  assembly is believed to be very
important in  achieving long battery life. The final drying step  coupled with dry room conditions
ensures a  minimal quantity of moisture will exist in the final  product. The pertinent processing
rate  in determining the resources  necessary for drying is the total amount of active  materials
processed per year (other electrode materials  are  approximately proportional), which for the
baseline plant is 2,950,000 kg/y  or 3,275 kg/shift. The individual electrodes exiting from the
electrode slitting process are separated into stacks by polarity,  loaded into vacuum drying ovens,
dried for several hours, and unloaded directly into the dry room. The estimated resources to meet
these needs are the following:
Electrode Drying
Dryer capacity
Number of dryers
Maximum rate
Rate Factor
2,950,000 kg/y
active material
Direct Labor
2 per shift
Capital Equip.*
1.6mil$total
600 kg/shift
8
4,320,000 kg/y
Plant Area, m2
300
*Total cost including installation

5.3.8 Control Laboratory

The purpose of the control laboratory is to ensure that the raw materials and the electrodes being
fabricated meet specifications.  Laboratory personnel collect or supervise collection of samples
and carry out analyses. The estimated resources to meet these needs are the following:
                                            58

-------

Control Lab
Rate Factor
869,000 kWh/y
Direct Labor
4 per shift
Capital Equip.*
1.5 mil$ total
Plant Area, m
300
*Total cost including installation

5.3.9 Cell Stacking

The cells are assembled in four steps, which are carried out in a dry room. The first of these steps
is cell  stacking. The primary rate factor that determines the cost for all steps in cell assembly is
the number of cells assembled per year. For cell stacking an additional cost factor is the capacity
of the  cells; large cells  usually require more electrodes of larger area and thus a more capable,
faster cell stacking machine. The method used to determine the extra costs of stacking equipment
is detailed in Table 5.4. The capacity of the cells is deemed  to have only a minor effect on the
other steps in cell assembly and is not taken into account  for  those  steps. The electrodes are
inserted in a folded separator sheet, the positive electrodes  tabs protrude on one side and the
negative electrodes tabs on the other. As in other parts of the plant, excess capacity is provided to
allow catching up after  unscheduled downtime. The estimated resources to meet these  needs for
the baseline plant are the following:
Cell Stacking
Stacking rate
Number of units
Maximum rate
Rate Factor
6,320,000 cells/y
Direct Labor
5 per shift
Capital Equip.*
4.0 mil$ total
5 cells/min
4
8,640,000 cells/y
Plant Area, m2
600
*Total cost including installation

5.3.10 Current Collector Welding

The current collector tabs for the negative and positive electrodes are welded to their respective
terminals by ultrasonic welding.  This procedure achieves  a connection of near-zero resistance
and avoids overheating the electrodes during the welding process. The estimated resources to
meet these needs are the following:
Tab Welding
Cell rate
Number of units
Maximum rate
Rate Factor
6,320,000 cells/y
Direct Labor
5 per shift
Capital Equip.*
4.0 mil$ total
5 cells/min
4
8,640,000 cells/y
Plant Area, m2
600
*Total cost including installation

5.3.11 Enclosing Cell in Container

The aluminum foil in the pouch container is sufficiently thick (100 microns default thickness) to
permit the use of stiff, pre-shaped pouch halves. The pouches are  assumed to be purchased as
finished parts. Each cell is enclosed in these containers, which are then partially sealed prior to
injecting electrolyte. The estimated resources to meet these needs are the following:
                                            59

-------
Enclosing cells
Cell rate
Number of units
Maximum rate
Rate Factor
6,320,000 cells/y
Direct Labor
3 per shift
Capital Equip.*
3.0 mil$ total
5 cells/min
4
8,640,000 cells/y
Plant Area, m2
600
*Total cost including installation

5.3.12 Electrolyte Filling and Cell Sealing

At this  station, the  cells  are evacuated, filled with electrolyte and temporarily sealed. The
estimated resources to meet these needs are the following:
Filling & 1st Seal
Cell rate
Number of units
Maximum rate
Rate Factor
6,320,000 cells/y
Direct Labor
5 per shift
Capital Equip.*
5.0mil$total
5 cells/min
4
8,640,000 cells/y
Plant Area, m2
900
*Total cost including installation

5.3.13 Dry Room Management

Excellent dry-room atmosphere is required for lithium-ion cell assembly. A maximum dew point
temperature of -40 °C is maintained in the room. The load on the dry-room drying apparatus is
determined by diffusion of water vapor through the walls, entry of air through the air locks, the
number of workers in the room, and the  need to admit some  fresh air to limit the build up of
contaminants such as electrolyte solvent vapor. These load factors are approximately a function
of the room area. Because  of the importance of the proper functioning of the dry room, two
workers are on duty at all times to monitor its performance. The equipment for circulation and
purification of the dry air will be located outside of the plant building, adjacent to the dry room.
The estimated resources to meet these needs are the following:

Dry Room
Operating
Area
3,000 m2
Direct Labor
2 per shift
Capital Equip.*
20.0 mil$ total
Air Locks, m
100
*Total cost including installation

5.3.14 Formation Cycling, Final Cell Sealing and Charge Retention Testing

Formation  cycling is expensive because  it takes considerable time  and each cell must  be
monitored separately. For plants to be  operated in 2020, we expect some improvements from
present day operations because of the urgency to improve and thus save cost. We project that the
entire formation cycling and testing can be done in two shifts.  These operations  consist of
charging the cell, discharging to full depth to measure capacity and impedance, followed by fully
recharging the cells. These tests will be carried out in large temperature controlled cycling units
                                           60

-------
that test 500 cells simultaneously, monitor each cell and automatically identify failed cells. The
capital cost of the cycling equipment is primarily a function of the annual number of cells to be
tested, but to a lesser extent on the capacity of the cells.

The short-term testing described above does  not detect cells that have self-discharge rates that
are slightly above normal, which could lead to catastrophic failures later. To detect such defects,
the cell charge is topped off and the cells are  stored for two weeks and then checked for loss of
charge. Most of the test period is spent in large racks in compact arrays, without electronic
monitoring. Incidentally,  the two-week long self-discharge testing requires less floor space than
for formation cycling, which lasts only two shifts.

The final cell sealing occurs between the formation cycling and charge-retention storage test.
Gas generated during formation cycling may  accumulate in the reservoir space that was created
during the temporary sealing  step.  This  gas  is removed by creating the  final seal below the
reservoir and trimming off the unwanted portion.
The estimated resources to meet these needs are the following:

Formation Cycling
Cell capacity
Number of cyclers
Cells per cycler
Length of test
Testing capacity
Final Cell Sealing
Charge Retention
Testing rack capacity
Racks per stack
Number of racks
Length of test
Testing capacity
Rate Factor
6,320,000 cells/y
6,320,000 cells/y
6,320,000 cells/y
Direct
Labor
8 per shift
2 per shift
3 per shift
Capital Equip.*
30.0 mil$ total
40 Ah
35
500
2 shifts
7,875,000 cells/y
2.0 mil$ total
4.75 mil$ total
500 cells
5
750
14 days
8,040,000
Plant Area,
m2
2200
450
900
*Total cost including installation

5.3.15 Module and Battery Assembly

Approximately 5%  of the cells are expected to fail the formation cycling and charge-retention
tests and these are sent to the Rejected Cell and  Scrap  Recycle section. The  accepted  cells
(6,000,000 finished cells per year)  are assembled into  modules by attaching the terminals
through laser welding or mechanical joining with spring loaded devices. Electronic circuit packs
are attached that occupy about the same volume as a cell. An aluminum heat conductor is placed
around every cell. These operations are carried out at four automated stations each capable of
handling about 280  cells per hour. For the module design being cost estimated in this model, the
                                           61

-------
module is enclosed in an air-tight aluminum container by double seaming. The processing rate
that determines the cost of module assembly is the number of finished cells that must be handled
per year.

The finished modules are assembled into battery packs with the  aid of automated stations. The
total cost of these stations is dependant mainly on the number of battery packs to be assembled
per year (100,000 for the baseline plant), but to a lesser extent on the number of modules per
pack. After assembly, the packs are moved to testing stations where they are discharged as a final
check  of impedance and  to  lower the  state of charge to a level  suitable  for shipping. The
estimated resources to meet these needs are the following:

Module Assembly
Number of stations
Cells/h/station
Capacity
Battery Pack Assembly
Modules/pack
Number of stations
Packs/h/station
Capacity
Battery Pack Testing
Rate Factor
6,000,000 cells/y
100,000 packs/y
100,000 packs/y
Direct
Labor
6 per shift
3 per shift
3 per shift
Capital Equip.*
6.0 mil$ total
4
280
8,060,000 cells/y
3.0 mil$ total
4
3
6
130,000 packs
3.0 mil$ total
Plant Area,
m2
600
450
450
5.3.16 Rejected Cell and Scrap Recycle

Scrap is generated in preparing the electrodes and by the rejection of 5%  of the cells that go
through formation cycling and  charge-retention tests. This scrap is gathered  and packaged for
shipment for recycling of the materials having value. No credit is taken for the value of the scrap
in this model except that the costs of gathering, sorting, packaging and shipping are understated
by about that value.  The main factor in determining  the cost of scrap recycle is the number of
cells rejected, which have to be  disassembled to recover the scrap, a labor intensive process. The
yields of materials in the various processing steps are shown in Table 5.5.

Table 5.5 Materials yields during electrode and cell fabrication
Material
Positive Electrode
Negative Electrode
Positive Current Coll.
Negative Current Coll.
Separator
Electrolyte
Material
Mixing
99
99




Coating
95
95
99
99


Electrode
Slitting
99
99
92
92


Cell
Stacking
99
99
99
99
98

Electrolyte
Filling





94
Total
92.2
92.2
90.2
90.2
98.0
94.0
The estimated resources needed for scrap recycle are the following:
                                           62

-------

Scrap Recycle
Scrap rate
Rate Factor
6,320,000 cells/y
Direct
Labor
5 per shift
Capital Equip.*
2.5 mil$ total
441 kg/shift
Plant Area, m2
600
5.3.17 Baseline Plant Summary

The processing rates and the primary cost factors for the baseline plant are summarized in Table
5.4. The main cost-determining rate of processing for each step is shown in the second column.
The requirements for direct labor, capital equipment and plant area, which are shown in detail in
the subsections above, are summarized in the table. It is seen that the plant requires a total of 90
workers per shift, $127,450,000 worth of capital equipment, and 15,425  square meters of plant
area to manufacture the baseline battery at a rate of 100,000 battery packs per year.
5.4 Adjustment of Costs for Varying Production Volumes

Production volume may affect the end price of the battery in two distinct ways. First, the user of
the model may change the annual production volume and every processing step will be affected.
Somewhat differently,  as  the  performance requirement and thus design is  changed, the
production of individual steps will change in non-uniform ways. As noted in Table 5.4, there are
many processing rates that must be considered in addition to the overall number of battery packs
manufactured per year. Each of these rates affects the costs of one or more steps in the process
and may have no effect upon the costs of other steps in the process. For instance, when the user
of the model increases the power of the battery packs without increasing the number of cells or
their capacity, the model increases  the area of the cells and decreases the electrode coatings
thicknesses. Such changes would result in an increase in the cost of the coating equipment, the
floor area  occupied by the equipment, and in the direct labor for that step in the process. It would
have no effect on the cost  of mixing the materials to be coated because the amounts  of  these
materials per battery back are unchanged under the assumed conditions.

The general approach to cost estimation of multiplying a known cost by the ratio of processing
rates  raised to a power has also  been applied to  the capital cost of  individual  items  of
          49
equipment.
                                      C = C0(R/Rof
(5.2)
Here, C0 is the capital cost of an installed equipment item designed for the baseline processing
rate, ^0. The power factor, p, relates the capital investment cost and the processing rate for the
manufacturing step.

If the value of p were 1.0, it would imply that the cost of the equipment item, or the equipment
items if  there are several in parallel,  would be directly proportional  to the processing rate.
However, the value of p  for the cost  of equipment is  frequently about 0.6 to 0.7 for many
manufacturing process steps  because the equipment is larger for the higher processing rates and
                                           63

-------
its cost is less than if it were directly proportional to the processing rate. For process steps
requiring the addition of many identical pieces of equipment for scale up, such as may be true for
formation cycling of battery cells, the value ofp may be as high as 0.9. The value ofp is unlikely
to reach 1.0 because the equipment cost includes installation,  for which  there is some savings
even in installing multiple units of the same processing capacity. The relationship between  cost
and processing rate for two-fold and three-fold rate changes is illustrated in Table 5.6.

          Table 5.6  The effect of processing rate (R) on cost for various scale factors
CICo = (R/R0f

Scale Factor, p
0.25
0.3
0.4
0.5
0.6
0.7
0.8
0.95
1.0
Cost Ratio, C/C0
R/R0 = 2
1.19
1.23
1.32
1.41
1.52
1.62
1.74
1.93
2.00
R/RO = 3
1.32
1.39
1.55
1.73
1.93
2.16
2.41
2.84
3.00
Similar equations have been applied for determining the effect of processing rate on the annual
hours of labor and the plant area required for a manufacturing step. In general, the value of p is
low for the labor equation, usually only 0.4 to 0.5, because only a relatively small addition to the
labor crew permits operation of larger equipment or of operating  several more units of the same
processing capacity.24 The value of p for the plant area required for a processing step is slightly
less than that for equipment. The floor area required for larger equipment or for more equipment
items of the same size is proportionately less than the increase in the processing rate because of
the more efficient use of the space occupied by the equipment and the savings in aisle area.

The value of the scale factors (i.e. p factors) for labor, capital equipment,  and floor area were
estimated  for each of the processing steps (Table 5.4).  The scale factors selected for the direct
labor requirement are usually only 0.4 to 0.5, which indicates considerable unit cost reduction for
increasing the plant throughput.

For most processing  steps, increasing the processing rate beyond that in the baseline plant would
result in a decision to increase automation or use faster equipment to mitigate the costs of higher
levels of throughput. Decreasing the processing rate would have the opposite effect. Some steps
in the process such as cell stacking, welding of current collectors, and formation cycling do not
appear to be easily automated beyond the level intended in the baseline plant and, thus require a
higher value for the scale factor of 0.8. This higher  scale factor results in achieving fewer
reductions in the cost per battery pack with increasing production  volume. Additionally, a higher
                                            64

-------
p factor results in a less severe penalty for lower production scale for an individual step in the
process.

There are five steps for which the cost of the capital equipment is affected by other factors than
the main processing rate for the process step. These are discussed in the footnotes at the bottom
of Table 5.4. For these steps, the costs that have been adjusted  for the changes in the processing
rate from the baseline rate are further adjusted to take into account the other cost  factors. The
cost of the coating equipment is adjusted for the amount of  solvents to be driven off of the
positive and negative electrodes; thicker coatings need longer, more expensive ovens to drive off
the additional binder solvent or the coater most be operated at lower speeds. The cost of the cell
stacking equipment and that of the formation cycling equipment, for which the main cost factor
in both cases is the number of cells to be fabricated annually, are also adjusted for the capacity of
the cells; larger cells require more expensive equipment. The cost of the capital equipment for
battery assembly is primarily a function  of the number of cells in the battery, but it  is also a
function of the number of modules that must be interconnected. This  dependence is accounted
for in the model with an additional multiplying factor.

A breakdown of the baseline plant capital equipment costs listed in Table 5.4 is illustrated in Fig.
5.3. The largest costs for capital equipment are for formation cycling and testing, cell assembly
in the dry room and electrode coating. These capital  costs are likely to be dominant in any
lithium-ion battery plant in the near future.
                            2%    7%
                                                15%
        29%
                                     28%
D Receiving and shipping

• Materials preparation

D Electrode coating

D Calendering

• Materials handling

D Electrode slitting

• Vacuum drying

n Control laboratory

• Cell assembly in dry room

• Formation cycling and testing

n Module and pack assembly

n Rejected cell and scrap recycle
Figure 5.3 Breakdown of installed capital equipment costs for the baseline plant
                                           65

-------
5.5 Plant Investment Costs

In this model, the calculated investment costs are defined as those directly related with building
and operating the plant (Table 5.6). Other costs that may require investment, such as research
and development, are added separately to the unit cost of the battery. The largest investment cost
is for the installed capital equipment. Each cost item for the battery under design is adjusted from
the estimate of the baseline plant. The plant cost is done in a similar way with a cost of $3,000
per square meter ($280/sq. ft) including land and utilities. The high  cost for land and utilities
accounts for both the area of the manufacturing facility as well as other land requirements such
as office buildings and waste water treatment requirements. Launch costs include plant start-up,
employee training and materials that are lost or recycled in early stages of production, beyond
the normal amounts. Launch costs are estimated to be 5 % of annual materials costs plus  10 % of
annual direct labor and variable overhead (Section 5.6). Working capital is needed to cover the
costs of payroll, receivables, and the inventories of raw materials, work in progress and finished
product.  These working capital costs are partially offset by bills that are payable. We estimate
the working capital to be 15 % of the annual variable costs.

Table 5.6 Battery pack manufacturing investment costs
Investment Costs
Capital Equipment
Plant Floor Space
Launch Costs
Working Capital
Description
Equipment costs including
installation
Space includes aisles and space
for unfinished processing
inventory plus land and utility
costs.
Plant start-up, training, out-of-
spec product.
Cash to meet payroll,
receivables, inventories of raw
materials and of unfinished and
finished product, minus
payables.
Method of Calculation
Estimates of costs for each
processing step at baseline rates
adjusted for actual rates.
Estimates of costs for each
processing step at baseline rates
adjusted for actual rates.
5% of annual materials cost,
10% of direct labor plus
variable overhead.
15% of annual variable costs.
5.6 Unit Costs for Battery Pack

The unit costs of the battery pack are calculated as summarized in Table 5.7.

5.6.1 Variable Costs

The costs of the materials and purchased items are based on the costs discussed in section 5.2,
and the annual amounts of materials are adjusted for the yields of materials (section 5.3) and
yield of cells. The direct labor is the sum of the labor cost for each step in the process, which are
                                            66

-------
each calculated for the baseline plant and adjusted for the rate associated with the battery under
study.  Variable overhead is  the cost of  indirect materials  and labor, utilities,  and  plant
maintenance. It is estimated to cost 40 % of direct labor costs and 20 % of depreciation.

5.6.2 Fixed Expenses

Fixed expenses include General,  Sales, and Administration (GSA), research and development,
and depreciation. The cost of GSA includes  the plant office, taxes on income and property, cost
of sales and insurance. It is estimated by the model as 25 % of direct overhead and depreciation.
Research and development (R&D) must be  carried out to ensure that the battery packs that are
produced in the plant and the means of production continue to be  competitive in the world
market with respect  to performance and price. The greater the investment in the plant and its
equipment, the greater is the need to be successful in the R&D effort.  Thus, the expenditure has
been set at 40 % of  the depreciation expense. Depreciation expense provides funding available
for future  investment in this plant  or another  venture to replace deteriorating  plant and
equipment. The equipment and plant are depreciated at straight-line rates for 6-year life (16.7 %
per year) and 20-year life (5 % per year).
Table 5.7 Unit cost of battery pack
Variable Costs
Materials and Purchased
Items
Direct Labor
Variable Overhead
Fixed Expenses
General, Sales, and
Administration (GSA)
Research and Development
Depreciation
Profit
Warranty
Description
All materials and purchased
items in finished product and
lost in processing.
Labor costs for operations and
immediate supervision.
Indirect materials, labor,
utilities, plant maintenance

Plant office, taxes on income
and property, cost of sales and
insurance expenses.
On-going research needed to
upgrade product and maintain
competitive position.
Provides funds for new
investments to replace those in
current equipment and plant.
Return on invested capital after
taxes.
Funds set aside for reimbursing
customers for battery pack
failures.
Method of Calculation
Based on prices of materials,
cost equations for purchased
items and yields.
Estimates of costs for each
processing step at baseline rates
adjusted for actual rates.
40% of direct labor cost plus
20% of depreciation

25% of direct labor and variable
overhead plus 25% of
depreciation.
40% of depreciation
16.7% of capital equipment cost
plus 5% of plant floor space
cost.
5% of total investment costs.
5.6% added to price based on
present worth of projected
payments.
                                            67

-------
5.6.3 Profits

The  profit goal for this type of venture varies  with the financial structure of the company,
especially regarding long-term debt. For the model, the profit is set to provide a 5 % return on
the total investment, which  is an  approximate average  for mature manufacturing as vehicle
battery production  is expected to  be  in 2020. In general, the chosen cost structure  and the
resulting margin are similar to a Tier 1  supplier in the automotive industry.

5.6.4 Battery Pack Warranty Costs

If a battery module  or an entire pack fails, the replacement will cost much more than the original
price paid by the OEM. It is  important that such events are rare, but provision must be  made to
reimburse the vehicle owner, especially in the early years  of the projected battery life. The extra
costs of replacing  the  battery will result from labor for testing  and replacing the  battery,
inventory costs for stocking replacement batteries, and servicing the battery controller if the new
battery is slightly different than the old battery. It is likely that the battery manufacturer will be
responsible for the  cost of the new battery, which we assume will be  equal to the cost  of the
original battery. The other costs of replacing the battery,  to the extent that they are covered by
the warranty, are assumed here to  be  covered by the automobile manufacturer and the dealer.
The  goal for  average battery life is  15 years  and  a warranted life of 10 years, with  full
replacement in the first five  years and shared cost  of replacement for the last five years  seems
appropriate. The vehicle owner would pay an increasing share of the cost from between 0 % at 5
years to 100 % at 10 or more years. With these assumptions, the cost to the battery manufacturer
will be equal to the present worth of the future costs of the new battery or modules as provided in
the warranty. The rate of battery failure will vary over  the life of the battery with a slightly
higher rate early in life, then a low failure rate followed  by a gradually increasing failure rate.
For purposes of calculation we assume a failure rate of 1.0 % per year throughout the warranty
period. With an internal rate of return of 8 % and calculated on a monthly basis, the present value
of the future costs would be  about  5.6 % of the price of the battery before adding the warranty
cost.
5.7 Summary of Baseline Battery Cost

The  spreadsheet version of the model, which is discussed in more detail in sections 6 and 7,
provides a summary sheet which is illustrated in Table 5.8 for the cost of the baseline battery and
that  of two  others. This breakdown of the battery costs, with  a brief summary of the design
values, illustrates the  effects of the cost factors. The second battery has twice the power of the
baseline  battery and the third battery has the same power as the baseline battery, but twice the
capacity. The number of cells is the same for each battery. The energy storage is slightly higher
for the  battery with  double power  because the voltage would be slightly higher during the
discharge to determine capacity. The battery with double the capacity has fewer electrodes which
are longer and  wider, because the cell thickness  is maintained, resulting in higher resistance in
the current-collector structure. The higher impedance lowers the voltage  during the discharge
capacity measurements and results in slightly less than twice the energy storage of the baseline
battery.
                                            68

-------
Table 5.8. Summary of results for cost of
double the power and double the capacity

Calculated Battery Parameters
Battery energy storage, kWh
Battery power at target % OCV, kW
Required battery power, kW
Capacity, Ah
Number of cells
Battery weight, kg
Battery volume, L
Weight and Volume of Components Exterior to Battery
Weight, kg
Volume, L
Cooling system power requirement, W
Vehicle electric range, miles
Investment Costs
Capital equipment cost including installation, mil$
Building, Land and Utilities
Area, m2
Cost, $/m2
Building investment, mil$
Launch Costs
baseline battery and that of
of the baseline
Baseline

8.7
50.0
50.0
40
60
59.1
31.2

11.0
5.6
1,760
20.3

127

15,374
3,000
46.1

battery
Double Power

8.8
100.0
100.0
40
60
72.4
37.1

9.0
4.0
649
20.5

149

18,165
3,000
54.5

similar batteries

with

Double Capacity Double Modules

17.3
86.9
50.0
80
60
108.2
55.2

11.0
5.6
857
40.5

161

19,204
3,000
57.6


17.4
100.0
100.0
40
120
116.0
59.7

9.0
4.0
724
40.6

212

24,284
3,000
72.9

Rate: 5% of direct annual materials + 10% of other annual costs
Total, millions
Working capital (30% of annual variable costs), mil$
Total investment, mil$
Unit Cost of Battery Pack, $
Variable Cost
Materials and Purchased Items
Cell materials
Cell purchased Items
Module
Battery pack
Total
Direct Labor
Electrode processing
Cell assembly
Formation cycling, testing and sealing
Module and battery assembly
Cell and materials rejection and recycling
Receiving and shipping
Control laboratory
Total
Variable Overhead
Total Variable Cost
Fixed Expenses
General, Sales, Administration
Research and Development
Depreciation
Total Fixed Expenses
Profits after taxes
Total unit cost per battery not including warranty, $
Summary of Unit Costs, $
Materials
Purchased Items not including cooling system
Direct Labor
Variable Overhead
General, Sales, Administration
Research and Development
Depreciation
Profit
Warranty (includes battery pack only)
Price to OEM for battery pack, $
Pack integration (BMS & Disconnects), $
Estimated cost to OEM for thermal management, $
Total cost to OEM for complete battery system, $
10.62
28.78
212.53



1,302
63
202
148
1,714

35
26
17
16
6
8
5
113
92
1,919

110
94
235
438
106
2,463

1,302
412
113
92
110
94
235
106
138
2,601
395
360
3,356
13.11
35.80
252.58



1,739
66
205
141
2,151

51
26
17
16
6
8
5
129
107
2,387

128
110
276
514
126
3,027

1,739
413
129
107
128
110
276
126
170
3,196
395
240
3,831
17.05
47.53
283.32



2,452
74
237
164
2,927

47
26
17
16
6
11
7
130
112
3,169

135
119
297
551
142
3,861

2,452
475
130
112
135
119
297
142
216
4,078
395
280
4,753
19.30
53.28
357.74



2,547
110
399
187
3,243

50
39
26
22
11
11
7
165
144
3,552

175
156
390
721
179
4,452

2,547
696
165
144
175
156
390
179
249
4,701
475
240
5,416
69

-------
Doubling the power does not add as much cost to the materials and purchased parts as doubling
the cell capacity. Most of the labor costs for the three batteries are similar with the major
difference being for the labor cost for electrode processing. The double power battery requires
greater labor costs  principally for coating  the larger electrode area. Capital equipment and
depreciation costs are higher for both  the high power and high capacity battery packs.  The
increases in capital equipment cost for the  high-power battery are for  coating, calendering,
materials handling  and vacuum  drying equipment. For the high-capacity battery,  the  main
additional capital equipment costs  are for the materials mixing, binder solvent  recovery, cell
stacking and formation cycling steps in the process.

Overall, doubling the power  of the battery  increases  the price by only 23  %.  Doubling the
capacity of the cells increases  the cost by 57 %, considerably more than for doubling the power.
Alternatively, doubling the number of baseline cells and modules within a larger battery jacket
(two rows of modules instead of one, twice the voltage, energy, and power) would increase the
cost by 81 %.

The summary of unit costs for the baseline battery pack, which is shown at the bottom of Table
5.8, is illustrated in Fig. 5.4.  The materials  and purchased items are the  largest costs for the
battery. For larger levels of production, these costs are even more dominant because the  scale
factors for these items are close to one.
                                       50%
               16%
n Materials

• Purchased Items not including
  cooling system
D Direct Labor

D Variable Overhead

• General, Sales, Administration

D Research and Development

• Depreciation

D Profit

• Warranty (includes battery pack
  only)
Figure 5.4 Breakdown of unit costs for baseline battery with total price to OEM of $2600. The
total cost to the OEM, including pack integration components, is $3,360.
                                           70

-------
       6  Description of the Spreadsheet Model and Instructions for Use

6.1 Background

Historically, the model has been based on Microsoft® Office Excel spreadsheets. The flexibility
afforded  by a spreadsheet approach has  been extremely useful to  the  development of  the
calculations. Until now, the model had been in a constant state of development. Changes to
parameters and equations  were made rapidly and frequently. The publication  of this report
represents the  first time a version of the model will be "frozen" for open distribution to  the
public. Advances  will continue to be made with the model, such as those discussed in the last
section of this report. However, distributions  of the revised model will be made in an orderly
fashion rather than the continuous improvement approach taken over the last number of years.

6.2 Instructions

The following  subsections are a brief explanation of how one may operate the spreadsheet based
model. The user  is  advised  to  save the original document separately  as a back-up copy.
Corruption of the  calculation is possible and will likely occur during use by someone unfamiliar
with the model.

6.2.1 Enabling Calculation

This Microsoft® Office Excel workbook requires the  use of iteration.  To enable this feature in
Office 2003, go to the "Tools" drop-down menu and select "Options."  On the calculation tab,
check the box next to "Iteration" and change the maximum number of iterations to 1000 (Figure
6.1). Perhaps most importantly, ensure the calculation is set to automatic and not manual. If the
iteration  is  not turned on,  the  software will present an error complaining about circular
references. If the model is opened while a different Excel spreadsheet is in use, the software will
also warn of an error. Simply close all Excel windows except for the  model; alternatively, one
could  re-enable the iterative function as discussed above. In newer versions of Excel such as the
Office 2010 edition,  the iterative function  may be enabled by going to File > Options >
Formulas.
                                           71

-------
 ] 9= Edt S»S" fserc
             look Eafa »ld™ t
             * -a a - ,/ 1 o •
                          -si *i
                                                  u
                                                         > •/, . -as iff _ -
                                                                         i * ' !>• s> i -3 <; I ./. » I J 4 «! £ & J

1
2
3
4
5
6
7
ToT
11
12
A



B
C D E
Cell


Positive Electrode
Active material capac
Weiaht %
Active material
Carbon
Binder
Binder sol\
Density, a
13 I Active mat
~U\ Carbon
15 I Binder
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
. <
Negative
N/P capac
Active mat
Weiaht %
Active mat
Carbon
Binder
Binder sol\


ent
_%
cm3
erial

Electrode
ty ratio aft
erial capac
erial


ent
Void. Vol% %
Density, q/cm3
Active material
Carbon
Binder


Positive Foil
Material
Thickness

urn
Negative Foil
••iftsssi'sd.

Selected System: LMO-G
ty, mAh/g: 100
F
G
Chemistry
NCA-G NMC-G
1160

89 89
6
5
NMP
32


1.825
1.77

5r formation 1.25
ty, mAh/g: 290

95
6
5
NMP
32
4.78
1.825
11.77

1.25
290
95
0 0
5 5
Water
34

2.24
1.95
1.1
Water
34
2.24
1.95
1.10

Aluminum
20
Aluminum
20

175

89
6
5
NMP
32
4.65
1.825
1.77

1.25
290

95
0
5
Water
34
2.24
1.95
1.10
H
l
J
Default Values
LFP-G

155

89
6
5
NMP
50
3.45
1.825
1.77

1.25
290

95
0
5
Water
34
224
1.95
1.10

Aluminum
20



Aluminum
20

£sante
LMO-LTO

inn
K
L
M
N


LMO-G Other


Override
Values




•t nn 	 	
cda | Intematwnal | save | Errachedang | SrjelSng | EecurRy
f? startup
^Nona
window opt*
r Page b,

Task Pane P Eprmulaha P 3atusbar P Widows In Tasltar
*„,»,*,»„* r ««„„,,„

aks F Row & cdlenn headers ^ HorEontd scral bar
p'Se.ovakjes F sheet tate
| OK | Cancel

	






~89 yt> \
6
5
Water
40
3.40
1.95
1.10
0
5
Water
34
2.24
1.95
1.10

Aluminum
20
Aluminum
20








































0


























P I






=|


















	 '-m
nn ' M.,4,l= ' unwind 4EMHMH&8F ly












Excel Options
General
| Formulas
Ft oof ing
Save
Language
Advanced
Cuitomtje Ribbon
Quick Autss Toolbar
Trust CtnW r
^-^j Change options related to formula calculation, performance, and error handlmg.
CakuiaLon options
Workbook Calculation . J Enable iterative calculation
® Autonwtk MasimumHtritioni; 1,000 -s-
Automatic except lor jjata tables
Waiimum Change1 0.001
.1 Manual
0 ftecakulatewoifcboot Defer* saving
WoiUngwtthlannulH
P4CI itltttntt Jtyte •
0 formula AutoCompiete
J: U?*labl< names in formulas
IE Use GetP>
•J Incomiitent calculated columniormula in tables J Unlocked cells containing formulas
151 Mumoers formatted as text or preceded by an apostrophe -f Data entered in a table is invalid

QK "] | Caind
Ssna Find a
»r" Filter- S*l*n-
Editing


Q










R









S










T *z



























Ready Calculate | | l-3f 3 EJ 100% -








+
Figure 6.1 Automatic iteration must be enabled for the spreadsheet model to function. The top
screen shot was taken from Microsoft
the 20 10 edition.
                                     ®
                                       Office Excel 2003 edition and the lower image is from
                                              72

-------
6.2.2 System Selection Worksheet

The cell chemistry is selected by copying the system designated at the top of a column, for
instance NCA-G in cell F3, pasting it into cell E3 (Figure 6.2). Any of the values in row E can be
overridden by entering the desired value in  column L. For example,  the maximum electrode
thickness  may be overridden by placing a new value in cell L53. The  selection of the  cell
chemistry also includes the associated prices at the bottom of the page.  These prices can also be
overridden by entering the desired values in column L. A full screen shot of the system selection
worksheet is in Figure 6.3. An alternative cell couple, NMC333-G, is pasted into column O as an
example of another commercially relevant battery chemistry.
                                         i [Compatibility Mode] - Microsoft E«e!
:±1 _jj ffl a a
|«™«i^ "S^"- =««" ^


1
2
3"
4
5
e
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
"27
28
29
E3 t
A B
t3»«. BF—I», ' ^ j ii^l
h/1 Gndtmei B HMdlnQi Zoo"* IM% loom to
selection
NCA-G
c

'D 1 E
S B
N(W Arrang
Window All

FW s spin
~Htd«
RWIH- a unM>*

F G
Cell Chemistry
Selected System:| NCA-G
Positive Electrode
Active material capacity, mAh/g: 160
Weight %
Active material
Carbon
Binder
Binder solvent
Void. Vol% %
Density, a/cm3
Active material
Carbon
Binder
Negative Electrode
N/P capacity ratio aft
Active material capac
Weight %
Active material
Carbon
Binder
Void. Vol% %
Active material
Carbon
Binder
Positive Foil
" system Selection Battei





=r formation
ty, mAh/g:







r
-------
                          Ji | NCA-G

                                 D
                    Selected System:[
Positive Electrode
Active material capacity  mAh/g:
Weight %
Active material
Carbon
Binder

           g/cm:
 13 Active material
 14 Carbon
 15 Binder
 16 Negative Electrode
 17 N/P capacity ratio after formation
 18 Active material capacity  mAh/g
 19 Weight %
 20 Active material
 21 Carbon
 22 Binder
 23
 24
 2S
 26 Active material
 27 Carbon
 28 Binder
 29 Positive Foil
 30 Material
 31 Thickness. u.m
 32 Negative Foil
 33 Material
 34 Thickness, ^m
 35 Separator
 36 Thickness, urn
 37 Void, Vol% %
 38 Density, g/cnr
 39 Electrolyte density, g/cm3
 40 Cell Voltage and Resistance Parameters
 41 OCV at full power (at SOC Tor power rating). V
 42 Open circuit voltage at 50% SOC V
 43 Solid state diffusion limiting C-rate (10-s). A/Af
 44 Negative electrode cm2/cm3
 45 Positive electrode cmVcm
 46 Electrode system ASI for power, ohm-cm'
 47    Selected ASI value
_48 i   At 50% SOC. 2-sec burst
 49~]   At 50% SOC. 10-sec burst
 50 1   At bottom  of SOC range. 10-sec burst
_51^ ASI correction factor for limiting current. %
 52 Electrode system ASI tor energy, ohm-crrf
 53 Maximum electrode coating thickness. u.m
 54 Available battery energy, % of total
 55    Selected % energy
 56l   microHEV  and HEV-HP
 57    PHEV
 58    EV
 59 Cell Materials Costs
 60 Positive Electrode. $/Kg
 61    Active material
 62^    Carbon
~6§1   Binder PVDF
 64    Binder Solvent (NMP)
 65 Negative electrode material S-'k
 66 |   Active Material
J»7_    Carbon Black
 68    Binder PVDF
 69    Binder Solvent
 70 Positive current collector foil.  S/i
 71 Negative current collector foil. S
 72 Separators. S/nf
 73 Electrolyte. S/L
 74
 75
                                         1.25
                                         330
                                           5
                                         Water
                                         3.551
                                         3.68
                                          27
                                         74000
                                         890

                                          30
                                          18
                                         23.6
                                          30
                                           3
                                         51.92
                                          100
Cell Chemistry
Default Values
-G j NCA-G
D 160
:','
5
NMP
32
4.78
1.825
7 1 77
5 1 25
D 330
95
0
5
er Water
34
4 224
5 1 95
• 1.10
mm Aluminum
20
>er Copper
12
20
50
6 046
> 1 20
51 3551
S 3 680
27
DO 74000
0 BQOO
18
6 236
30
3
)2 51 9
3 100

25
70
80
NCA-G
&;:;;! 3700
0:;;;:; 680
fl:!;!:! 1000

-------
                                'aC Model Beta [Compatibility Mode] - Microsoft



                                                     L^ New Window  3 Spirt

                                                     jjg Arrange All   "1 Hide

                                                        Freeze Panes - 3 Unhid
                                                                                                           Save    Switch
                                                                                                    ij  workspace Wind
                               f* \ PHEV
                            ram for Calculating Performance and Materials Requirements
                                        LiNi0.80Co0.15AI0.0502-Graphite
 4  System Chemistry Input
 5  Finished Cell Materials
 G  Positive Electrode g
 7  Active material
 8  Carbon
 9  Binder
                                      Battery 1   Battery 2  Battery 3   Battery 4  Battery 5
    Void
    Total
                   Vol %
 12 Negative Electrode, g
 13 Active material
 14 ^Carbon
 15 Binder
 17 Total
 18 Baance of Ceil
 19 Positive foil, m;
 20 Negative foil, m2
 21 Separator, m2
 22 Electrolyte . L
 23 Positive terminal assembly, g
 24 Negative terminal assembly, g
 25 Thickness of cell container aluminum layer, um
 26 Thickness of cell container (PET-AI-PP). urn
 27 Density of cell container. g/crn3
 28 Cell container (PET-AI-PP). g
 29 Cell mass, g
 30 Length-to-width ratio for positive electrode
 31 Cell thickness, mm
 32 Thickness of cell edge from positive electrode to outside of fold, mm

 34 Top of positive electrode to top of terminal, mm
 35 Cell Capacity Parameters	
 36 Positive active material capacity, mAh/g;
 37 Positive electrode capacity. Ah/cm3
 38 Negative active material capacity, mAh/g:
 39 Negative electrode capacity. Ah/cm3
 40 Negative-to-positive capacity ratio after formation
 41 Cell Voltage and Resistance Parameters
 42 OCV at full power, V
 43 Open circuit voltage average for discharge. V
 44 Electrode system ASI for energy, ohm-cm*
 45 Excess negative area, %
 46 Maximum allowable electrode coating thickness,  ^m
 47 Cell terminal contact voltage loss, % of cell OCV
 48 Rate of terminal temperature rise at full  power. °C/sec
 49 Target % OCV at full power
 50 % OCV at full power adjusted for thickness limit
 51 Battery Input Parameters
J2 Vehicle type (microHEV, HEV-HP, PHEV, EV)
 53 Duration of power burst (10 or 2). s
 54 Battery power, kW
 55 Number of cells per module
 56 Number of cells in parallel
 57 Number of modules in row
 58 Number of rows of modules
 59 Number of modules per battery
 60 Cells per battery
 61 Voltage drop for bus bar for packs with one row of modules. V
 62 Battery pack insulation thickness, mm
 63 Battery jacket total thickness, mm
 64 Number of batteries manufactured per year
 65 Cell Chemistry  Input
 66 Battery Performance and Design Input
 67
 63
 69
 70 Calculated Cell Parameters
 71 Capacity, Ah
 72 Cell group capacity
 73 Cell capacity
 74 ASI Calculation
 75 Limiting current density, mA/cnr1
 75 Limiting C-rate. A'Ah
 77 Electrode system ASI for power, ohm-cm'
 78 Current collector resistance parameter,  ohms
 79 Current collector ASI. ohms-cm2
Weight %
(89
I
100
Weight %
35
!
100
Thick., tim
20
12
20
r ym
jm



Density
478
1 825
1.77
2749
Density
224
1 95
1 10
1406
Density
270
892
046
120






192.95
13.01
1084
21680
11986
631
12617
0476
0517
0929
00693
9.5
31.2
100
150
2.2
23.0
579
192.39
12.97
10.81
21617
11947
629
125.75
0.571
0615
1 116
0.0711
96
31 7
100
150
22
236
599
191.80
12.93
1073
215.51
119.03
626
12530
0.726
0.774
1.419
0.0742
9.8
32.5
100
150
22
246
632
191 40
12.90
1075
215-05
11871
625
124.96
0893
0946
1 747
0.0775
101
33.4
100
150
22
257
668

191.10
1288
10.74

214.72

118.46

623

124.70

1.077
1 135
2 108
0.0813
104
34.3
100
150
22
26.9
708
























                                                                          1 30
                                                                          80
                                                                          1.0
                                                                          1 00
                                                                           15

                                                                        160
                                                                       0392
                                                                        330
                                                                       0 441
                                                                        125

                                                                       3551
                                                                       3680
                                                                        51.9
                                                                        234
                                                                        100
                                                                        001
   1 30
    8.0
    1 0
   1 00
     15
 160
0392
 330
0 441
 125

3.551
3680
 51 9
 228
 100
 001

Program for Calculating Performance and Materials Requirements
                 LiNI0.80Co0.15AI0.0502-Graphite
                                     Battery 1   Battery 2  Battery 3
005
80
80.0
PHEV
10
95
16
1
3
2
6
96
003
10
13
0 100,000
snts
0.05
SO
800

PHEV
10
110
16
1
3
2
6
96
003
10
13
100,000

























meters




Wcm'

















309
30.9

85
27


30.8
308

85
27


307
30.7

85
27


30.6
306

85
27


306
306

85
27
                         Figure 6.4  Top portion  of Battery  Design worksheet.
                                                                       75

-------
 Page
Layout IUJ Full Screef
  Workbook Views
                                [7] Rulei    [7] Formula Ba

                                G/j Gridlines J7| Headings

                                         Show
                                f* j  PHEV
                   B
                                          D
                                                                                      H
J2_ Vehicle type (microHEV, HEV-HP, PHEV, f
 53 Duration of power burst (10 or 2), s
 54 Battery power. kW
 55 Number of cells per module
 56 Number of cells in parallel
 57 Number of modules in row
 58 Number of rows of modules
 59 Number of modules per battery
 60 Cells per battery
                                                              PHEV
 61 Voltage drop for bus bar for packs with one row of modules. V
 62 Battery pack insulation thickness, mm
 63 Battery jacket total thickness, mm
 64 Number of batteries manufactured per year	
 65 Cell Chemistry Input
 65 Batteiy Performance and Design Input	
 67
 66
 69
                                                         10
                                                         50
                                                         16
                                                         1
                                                         3
                                                         2
                                                         6
                                                         96
                                                        003
                                                         10
                                                         13
                                                        100.000
PHEV
  10
  66
  16
  1
  3
  2
  6
  96
 003
  10
  13
100.000
PHEV
  10
  80
  16
  1
  3
  2
  6
  96
 0 03
  10
  13
100.000
PHEV
  10
  95
  16
  1
  3
  2
  6
  96
 003
  10
  13
 100.000
PHEV
  10
 110
  16
  1
  3
  2
  6
  96
 0 03
  10
  13
 100,000
                Program for Calculating Performance and Materials Requirements
                                  LiNi0.80Co0.15AI0.0502-Graphite
                                                                 Battery 2
meters



iA/cm2














30.9
30.9

35




27


 70 Calculated Celt Parameters
 71 Capacity, Ah
 72 Cell group capacity
_73 Cell capacity
 74 ASI Calculation
 75 Limiting current density. mA/cm*1
 76 Limiting C-rate. AWi
 77 Electrode system AS1 for power, ohm-cm2
 73 Current collector resistance parameter, ohms
 79 Current collector ASi. ohms-cm2
 80 Total cell terminal AS! ohms-cm'
 81 Cell and battery terminal connections, ohms
 82 Total cell hardware and battery resistance, ohm-cnr1
 83 Total cell ASI for power, ohm-cm4
 84 Total cell ASI for energy fC/3 rate), ohm-cm2
 85 Electrode Coating Thickness Calculation
 86 Positive electrode thickness parameter, urn
 87 Negative  electrode thickness parameter, jim
 88 Positive electrode thickness at adjusted % OCV,
 89 Negative  electrode thickness at adjusted % OCV.
 90 Cell Area Calculation
 91 Area determined at target % OCV
 92 Area limited by max  allowed electrode thickn
 93 Cell area based on total ASI for power, cm2
 94 Cell Dimensions
 95 Number of bicel! layers (97% packing density)
 96 Width of positive electrode,
 97 Length of positive electrode,
 98 Length of current collector tabs, mm
 99 Width of terminals, mm
 100 Length of terminal materia
 101 Width of cell, mm
 102 Length of cell, mm
 103 Volume of ceil, cm3
 104 Module Parameters
 105 Weight of each cell group interconnect (copper), g
 106 Module state-of-charge regulator assembly, g
 107 Terminal  heating factor. W/g
 108 Terminal  resistance factor A-ohms/cm
 109 Module terminals, if more than one module (each 2.0-c
 110 Module terminal resistance both terminals, oh
 111 Module wall thickness (aluminum), mm
 "112. Length of aluminum conductor, rnm
 113 Thickness of aluminum conductors, mrr
 114 Total weight of aluminum conductors, g
 115 Balance of module materials, g
 116 Module length, mm
 117 Module width, mm
 118 Module height, mm
 119 Module volume. L
 120 Module weight, kg
 121 Calculated Battery Parameters
 122 Total battery energy storage. kWh
 123 Useable battery energy storage, kWn
 124 OCV at full power. V
 125 Nominal battery voltage (OCV at 50% SOC)
 126 Battery power at target % OCV and SOC. kW
 127 Maximum current at full power. A
 128 Maximum current density at full power. mA/cm'
 129 C-rate at  full power, A/Ah
 130 Coolant space above and below modules and at end of jacket, mm
 131 Thickness of module compression  plates (steel), mm
 132 Battery pack le
 H « > M »HV4J-J.L
                                                                                  Battery 3
                                                        298
                                                      0 005409
                                                        0781
                                                        0.073
                                                       0.000372
                                                         17
                                                         143
                                                         185
                                                         16

I. mm













135
26
1*5
215
249

hm-cm2





3CV, |im
OCV, urn


ness.


O.BS7 0.918
306 31.4
55.2 55.7

99.7 747
1107 830
90.0 747
100.0 830

7911 10524
8 758 8 732
8.758 10524

0.975
324
56.6

58 5
650
58.5
65.0

13.401
8.705
13.401

1.036 1 102
33.7 35 1
57.5 58.5

473 39.1
52.6 43.5
473 391
52 6 43.5

16523 19961
8.687 8.674
16523 19961



























 Ready  Calculate
                           Battery Design    Summary of Results     Manufacturing Cost Calculations

                        Figure  6.5 Middle  portion of Battery Design worksheet
                                                                        76

-------
The  Battery Design worksheet automatically receives  input from the  System  Selection
worksheet. These values are shown in purple (Figures 6.4 and 6.6) and must not be altered on the
Battery Design worksheet. As explained above, cell chemistry values may be adjusted  on the
System Selection worksheet. The operator provides battery design input in the aqua colored cells
(Figures 6.4 and 6.6). The battery input parameters on lines 54 to 58 (Figure 6.4) and lines 162
to 164 (Figure 6.6) are the only input values that the operator is required to provide to study a
group of batteries.  The type of vehicle battery (microHEV, HEV-HP, PHEV, or EV) on line 52
in Figure 6.4, is another important variable to be specified. One performs the selection by typing
the name  of the vehicle  battery type in cell  F52.  While  the  correct spelling is important,
capitalization is  not. This selection  automatically determines the state of charge at which full
power is designated (thus, the open-circuit voltage and AST for full power) and the length of the
power burst (2 seconds for microHEV  and 10 seconds for  all others). It is expected that the
majority of the remaining  default values should serve well for most batteries; however, the user
may also change to their exact specifications.

The  cell capacity (lines 162  to  164  in Figure 6.6) can be set in any of three ways: (1) directly
specifying the capacity (Ah) on line  162, (2) specifying the total battery energy on line 163 or (3)
specifying the electric range of the vehicle (miles). Only one of the three lines should be filled in
and the others should be blank. The model will follow the directions of the top-most line with
non-zero values.

The number of batteries manufactured per year is selected on line 64 in Figure 6.4. Changing this
value from the default value of 100,000, which is the manufacturing rate for the baseline plant,
will change the manufacturing cost.

If it is desired to study more than five batteries in the same workbook it is only necessary to add
additional columns by copying the battery 5 column to the right as many times as desired. Care
should be taken  that the appropriate values are  maintained when the cells are copied over. The
aqua colored cells  are typically the  source of any problems. The same column additions most
also be done for all other worksheets containing calculations.

6.2.4 Remaining Worksheets

The  cost calculations are done on  the Manufacturing Cost  worksheet and the results for the
model are shown on the Summary of Results worksheet (Figure 6.7). No parameters need to be
entered on these worksheets by the operator; all of the input for these worksheets is from the
Battery Design and the Cost Input worksheets. Tables for presentations or for preparing  graphs
of the data can  be  assembled  at the bottom of either the Battery  Design or the Summary  of
Results worksheet.  These tables  can  be transferred to a blank worksheet for more complex
studies. For instance, results for different cell  chemistries  can be copied and pasted (special
paste, values and numbers formats) to a blank worksheet.  On the  last worksheets,  the  cell,
module, and battery design, as well as the baseline plant are sketched.
                                           77

-------
                                 : Model Beta [Compatibility Mode] - Microsoft E.
*V Home Insert Page Layout
Formulas Data Rev ew vie^
& Q = rgl y
! |7J1JJ Pag. Break Previ™ f ^^ ^ O PS Fg] ^ N™ Window E3 Spirt JJ |~ 1 CD
^ 13 Custom Wews « «™ W 9 arrange All — \ Hide j' "-•
Normal Page @ Gridlines El Headinss Z°°"i K»* Zoom to Sa.e Switch Macros
Layout U-l Full screen Selection gll Freeze panes * :, workspace Windows '
Workbook Views
Show Zoom Window Macros
F90 W- *| |*
ABC
D E F
90 |Cell Area Calculation
91 Area determined at target % OCV
92 Area limited by max. allowed electrode thickness.
93 Ce area based on total ASI for power, cm
94 Cell Dimensions
95 Number of bicell layers (97% packing density)
96 Width of positive electrode, mm
97 Length of positive electrode, mm
98 Length of current collector tabs, mm
99 Width ofterminals. mm
100 Length of terminal material, mm
101 Width of cell, mm
102 Length of cell, mm
103 Voumeofcell. cm*
104 Module Parameters









105 We ght of each cell group interconnect (copper), g

7911
8.768
8,758

17
143
185
16
135
26
145
215
249

0.0
106 Module state-of-charge regulator assembly, g 128
107 Terminal heating factor W/g 0.019
108 Terminal resistance factor. A-ohms/cm 0.00054
109 Module terminals, if more than one module (each 2.0-cm long), g 24
110 Module terminal resistance both terminals, ohms 0.0000120
1 1 1 Module wall thickness (aluminum), mm 0.5
112 Length of aluminum conductor, mm 185
113 Thickness of aluminum conductors, mm 0 40
114 Total weight of aluminum conductors, g
115 Ba ance of module materials, g
116 Module length, mm
117 Module width, mm
118 Module height, mm
119 Module volume. L






120 Module weight, kg
121 Calculated Battery Parameters
122 Total battery energy storage. kWh
123 Useable battery energy storage kWh
124 OCV at full power. V
125 Nominal battery voltage (OCV at 50% SOC)
126 Battery power at target % OCV and SOC. kW
127 Maximum current at full power, A
547
227
217
144
146
458
10.20
1071
750
340.9
3533
55.4
178
128 Maximum current density at full power. mA/cm2 20.31
129 C-rate at full power. A/Ah 5.8
130 Coolant space above and below modules and at end of jacket, mm 0.6
131 Thickness of module compression plates (steel), mm 1.5
132 Battery pack length, mm 461
133 Battery pack width, mm 471
134 Battery pack height mm

174
135 Battery volume, L 37.7
136 Weight of each module inter-connect (5-cm long), g 30
137 Weight of both battery terminals (each 5 0-cm long), g 74
138 Weight of module compression plates and steel straps, g 1537
139 Weight of bus bar for packs with one row of modules, g 0
140 Resistance of module interconnects if more than one module, ohms 0.0000903
141 Resistance of battery terminals
142 Power of battery heaters kW
143 Weight of battery pack heaters (0.1 kg'kW), kW
144 Battery jacket weight parameter, g/cm2
145 Battery coolant weight within jacket, kg
14G Battery jacket weight, kg
147 Battery weight kg
148 Pack integration (BMS & disconnects), kg
149 Pack integration (BMS & disconnects). L
150 Battery Cooling System


0.0000181
20
02
0.84
276
10.5
71.7
4.0
40

151 Heat generation rate for pack, W 3871
152 Cooling System Refrig 4
153 Weight and Volume of Cooling System Exterior to Battery
154 Weight kg 70
155 Volume. L 2.8
156 Vehicle Electric Range
157 Energy requirement. Wh/mile
158 Available battery energy. % of total



159 Vehicle range, miles
160 Cell Capacity Calculation
161 Select capacity, battery energy, or vehicle rang*
162 Capacity {Ah}
163 Battery energy jkWh)
164 Vehicle range (miles)
165 Capacity at C/3 Ah
166 Capacity holding



167 Positive electrode thickness
168 Positive electrode thickness holding
169 Convergence parameter
1701
1711


300
• 70
25.00
. but only one.
| 25.0
30.872
30.872
" 900
90.0
0.3
Restart (0/1) 1
"*t N 'BffiSsJH*' Battery i>esign bummary ot nesufcs rianura
G
10.524
8.732
10.524
19
145
188
16
137
26
147
218
256

00
123
0.019
000054
32
0.0000090
0.5
188
040
562
232
220
144
149
470
10.54
1071
750
3409
353.3
650
238
22.65
77
0.5
15
461
476
176
386
40
99
1577
00000674
00000135
2.0
0.2
084
258
10.3
735
4.0
40

3062
Refrig 4
70
2.8
300
70
25.00
25.0
30.782
30782
' 74.7
747
0.3
1
turing Cost C
H 1 1
13.401 16523
8.705 8 687
13.401 16.523
23 28
148 152
193 198
16 16
140 144
26 26
150 154
223 228
268 280
00 00
128 128
0019 0.019
0.00064 0.00054
39 46
00000073 0.0000061
05 05
193 198
0 40 0 40
588 616
239 246
225 230
144 144
152 156
4.91 5.14
11 10 11 72
1071 1071
7.50 7.50
340.9 340.9
3533 353.3
80 0 95 0
293 348
2189 21.08
96 114
0.5 04
15 1.5
461 461
485 495
179 183
40 0 41.7
49 58
122 145
1645 1719
0 0
00000547 00000461
00000109 00000092
2.0 2.0
02 02
0 84 0.84
239 226
103 10.5
77.0 80.8
40 40
40 40
2303 1829
Refrig 4 Refrig 3
70 50
2.8 2.0
300 300
70 70
25 00 25.00
25.0 25.0
30 688 30 624
30.688 30.624
' 58.5 ' 473
58.5 47.3
0.3 0.3
1 1
ilcuiations Cost Input
J K

19.961
8.674
19961

32
156
203
16
148
26
158
233
294

00
128
0.019
000054
54
0.0000053
05
203
0.40
648
255
235
144
160
640
12.41
10.71
750
340.9
353.3
110.0
403
20.21
132
0.4
15
461
506
187
435
67
168
1803
0
00000398
0 0000080
20
02
0.84
216
107
851
40
40

1501
Refrig 3

50
2.0

300
70
2500


25.0
30.576
30.576
1 39.1
39.1
0.3
L #f































B

































1
Price ot Modules AiHuHFAHillslSBli] < >_Q
Ready Calculate
             Figure 6.6 Bottom portion of Battery Design worksheet
                                          78

-------
             lodelBeta [Compatibility Mode] - Microsoft Excel
"^1 JJ Page Break Preview
l3 Custom Views
Normal Page g] rj
Workbook View!
idline! 0 Heading! Z°°"i ™* Z°°
Sele
Show Zoom
^ q£ New Window
~A „_
•* 3 Arrange All
nto
ton P Freeze Pane! -
23 split _u n nn
Save Swrtch
Ij Workspace Windows
Window
Macros
Macros
F5 - £ ='Battery Design'!F122

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
1B
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
31
32
A B • ~C D
E F G
H
Summary of Results
LiNi0.80Co0.15AI0.0502-Graphite
Battery 1 Battery 2 Battery 3
Calculated Battery Parameters
Battery energy storage kWh I 10 71 107
Battery power at target % OCV. kW 55 4 " 65.0
Required battery power, kW
Capacity. Ah
Number of cells
Battery weight, kg
600 650
31 31
96 96
71 7 73 5
Battery volume. L 377 38 S
We ght and Volume of Components Exterior to Battery
Weight kg 10 11 0
Volume L 5.6 5 6
Cooling system power requirement. W 1.551 1226
Vehicle electric range, miles 25.0 25 0
Investment Costs
Capital equipment cost including installation, mils 162 166
Bu Iding. Land and Ut iities
Area, m2 18.948 19433
Cost. S.'m! 3.000 3 000
Building investment. mil$ 56.8 58 3
Launch Costs
Rate" 5% of direct annual materials + 10% of other annual costs
Total, million! 1291 1332
Working capital (30% of annua variable costs), mill 3497 3613
Total investment mils 26647 273 36
Unit Cost of Battery Pack, $
Variable Cost
Materials and Purchased Items
Cell materials
Cell purchased Items
Modue

1.561 1.619
87 87
306 306
Battery pack 138 140
Total 2.081 2 153
Direct Labor
Electrode processing 37 40
Cell assembly 34 34
Format on cycling, testing and sealing 22 22
Modu e and battery assembly 20 20
Cell and materia s rejection and recycling 9 9
Receiving and sh pping 9 9
Control laboratory 6 6
Total 136 139
Variable Overhead 114 117
Total Variable Cost 2.332 2 409
General. Sales. Administration 137 140
Research and Development
119 122
Depreciaton 298 305
Total Fixed Expenses 554 567
Profits after taxes 133 137
Total unit cost per battery not ncluding warranty. S 3.019 3113
Summary of Unit Costs, S
Materials 1.551 1.619
Purchased Items not including cooling system 530 534
Direct Labor 136 139
Variable Overhead 114 117
General Sales. Administration 137 140
Research and Development
Depreciation
Profit
Warranty I ncludes battery pack only)
119 122
298 305
133 137
169 174
Price to OEM for battery pack. $ 3188 3287
Pack integration (BMS & Disconnects). S 435 435
Estimated cost to OEM for thermal management- S 320 320
Total cost to OEM for complete battery system, S 3943 4.042
Price to OEM for modules for one pack. $ 3014 3111
Chart Values






Vc



Ready Calculate
Range 25.00 25.00
Thickness 100 83
Weight 71.7 735
Volume 37.7 38.6
Cost 3014 3111
Cell Area 09 11
Itage (%OCV) 82 80

107
800
800
31
96
770
40-0
10
56
923
250
172
20177
3.000
60.5
1399
38.00
284 13
1.732
88
307
142
2.269
44
34
22
20
9
9
6
143
121
2.533
145
127
316
588
142
3.263
1 732
537
143
121
145
127
316
142
183
3.446
435
280
4.161
3.267
2500
65
77.0
40.0
3267
1.3
80

1
Battery 4
107
95.0
950
31
96
808
41.7
9.0
40
733
250
178
20.934
3000
628
14.72
40.06
295.40
1856
89
308
147
2398
48
34
22
20
9
9
6
148
125
2.671
150
131
328
609
148
3.427
1.855
544
148
125
150
131
328
148
192
3.619
435
240
4.294
3.436
2500
S3
808
41 7
3436
1.7
80




J
Battery 5
10.7
1100
110.0
31
96
85 1
43.5
90
4.0
601
250
134
21.722
3.000
652

15.52
4233
30738
1.991
90
309
152
2541
53
34
22
20
9
9
6
152
129
2822
155
136
340
631
154
3,607
1.991
551
152
129
155
136
340
154
202
3.809
435
240
4.484
3.620
2500
43
85.1
43.5
3620
20
80



K L M























































































"
N












































































.i.u...iaF^.i.









































mrannfTU tITi
',n3D 100% - ~H±)
Figure 6.7 Summary of Results worksheet
                     79

-------
6.3 Battery Design Format Requirements

As the battery design is based off an assumed format (Section 2), certain design requirements are
necessary to ensure the modeled battery is physically realistic. The dimensions of the calculated
battery pack should be examined.  Some final designs may benefit from changing the cell aspect
ratio, H/W, to fit the end-use application. One example would be, to change the height of the
battery pack. Also, for a set number of cells in the pack, changing the number of modules, thus
cells per module, allows for adjustment of the pack dimensions.

6.4 Troubleshooting and General Advice

The spreadsheet iterates to find  the solution and this sometimes causes error messages to appear
after an entry is changed. These errors can usually be removed by first correcting  any erroneous
entries (non-numeric, two decimal points, etc.). Then the cells may be reset to default values by
entering a "0" (i.e. zero) in the  restart cell, F170 in Figure 6.6. Finally, entering  a "1" in F170
restarts the iteration process leading to a successfully converged answer.

At some point, a user will ask the  model to design a battery that is outside the bounds of what is
allowable for the selected cell chemistry. The most common error is when too large of a P/E ratio
is requested. Two different physical limitations are approached with increasing P/E ratio. First,
the electrode thickness is  shrinking. At some point,  the value will become unrealistic and
eventually approach  0 crashing the  calculation. At the same time, the  C-rate for the active
material is approaching the limiting  C-rate defined in  the Cell Chemistry Worksheet.  As this
value is approached, the AST will increase to larger and larger values, which  thus demands
smaller and smaller electrode thicknesses. Eventually, the calculation will crash.

Common sense approaches to resolve these issues are  to use lower designed power  or higher
designed energy. The C-rate and electrode thickness are  easily viewed  in the model output.
These are found on the Battery  Design worksheet in row 129 for the C-rate and rows 88 and 89
for the electrode  thickness. Therefore, the  user may try designs  of increasing P/E ratios and
watch to see how the electrode  thickness and C-rate is changing. Different cell chemistries will
have  different  sensitivities to  the P/E ratio  depending on  the  defined limiting C-rate and
calculated AST for power. What is possible with the LMO-G system will not always be possible
with the NCA-G system. P/E ratios that satisfy the expression in Eq. 6.1 generally result  in
successful battery designs. Higher P/E ratios are allowable in some situations. Note that selecting
the microHEV design doubles the allowable C-rate since only two second pulses are used. The
limiting C-rate, rc,um, may be found in cell E43 on the System Selection worksheet and is carried
over to row 76 in the Battery Design worksheet.

                                                                                    (6.1)
                                        E   1.35

6.5 Suggested Number of Cells, Modules, and Performance Inputs

Table 6.1 presents some suggestions for the required inputs into the design model that might
change  depending on the type of vehicle battery being designed. These values are only


                                           80

-------
suggestions, but tend to be similar to practices used today or projected to be used in the future by
industry. If the calculated cell capacity is higher than 60 Ah,  the user should consider the
inclusion of parallel cells as an additional parameter to examine. The default energy usage rate in
BatPaC is 300 Wh/mi. This rate may be used to size the energy requirement based on a desired
electric range for the vehicle by specifying the distance in  row 164 on the Battery Design
worksheet.

Table 6.1 General suggestions for range of input parameters that change with  battery type
  battery type     modules/battery  OCV  @ 50% SOC  Power (kW)   Energy (kWh)
HEV-25
HEV-HP
PHEV
EV
1-4
2-4
4-6
4-6
40 - 200
160 - 260
290 - 360
290 - 360
25
25-80
40 - 160
80 - 160
0.6-1.5
1-2
4-30
20 - 200
6.6 Entering a New Material Couple

The user of the model may wish to examine an electrochemical couple that is not included as one
of the options available in the model. We list below a brief explanation on how to properly enter
new materials into BatPaC. Various properties may be calculated, found in literature or measured
in the laboratory. The self-consistency of the data used is very important.

Experimentally measured values required:
   1.  Half cell formation cycling data from positive and negative electrode
   2.  Half cell cycling data from positive and negative electrode at C/3 rate
   3.  Full cell open-circuit voltage measurement at 50 % and 25 % SOC
   4.  Full cell ASI measurement for 5C pulse at 50 % and 25 % SOC
   5.  Full cell ASI at 50% SOC during a C/3 discharge
   6.  Electrode void fractions, active material densities, electrode component weight percent
   7.  Estimated interfacial area from surface area (preferred) or particle size measurements

The ASI calculation includes some additional parameters that become important as the designed
P/E ratio increases above  10 h"1 or the electrode thicknesses decrease below 30 microns. The user
is referred to section  3.4 and the supporting manuscript from Gallagher et al.20 for the parameter
estimation  process.  An exchange current of 0.15 mA/cm2,  normalized  to the surface area
calculated using the BET method from nitrogen  absorption experiments, is used in the model in
row 77 in the Battery Design worksheet. While the exact value of the exchange current will vary
from  the material  to material,  the  general behavior of the ASI will  be  preserved with  this
assumption. If lower P/E  ratio designs are desired,  the exact valuation of the exchange current,
interfacial  area, and limiting  C-rate are  less  important. However,  the experimental ASI
measurement should  then come from a cell with similar P/E ratio (electrode loading). Electrode
thicknesses 40 microns or larger should be used  to minimize the  contribution of interfacial
impedance to the ASI measurement. Otherwise  the "ASI correction factor" may not accurately
remove the interfacial component. A reasonable approach for a first approximation of the ASI
parameters may  be to  select the same values for a similar material. For example, if the new
material is based on  nano-sized primary particles, then the parameters for LFP or LTO may be

                                           81

-------
close enough. The AST does depend on a large number of factors and a full determination of the
parameters is important to capture all of the physical behavior of the cell couple. If the desire of
user is to reproduce or to reverse engineer an existing cell, then many of the parameters may be
estimated from the electrochemical characterization of the full cell data. These electrochemical
results typical require a teardown  of the cell to measure the area  of the electrodes and their
loadings/thicknesses, assuming no legal contract has been established prohibiting the analysis of
the cell.

The  available lithium for cycling in a full cell configuration may be calculated from half cell
measurements. The calculation method may be found in the Capacity Calculator worksheet in the
spreadsheet model and is detailed below. Alternatively, the reversible capacity of the full cell in
the experiment may be normalized to the mass  of the positive and negative  electrodes while
carefully accounting for the negative to  positive capacity ratio.  The first cycle efficiency of a
positive or negative electrode based half-cell may be defined as the ratio of the first discharge
capacity divided by the first charge capacity, Equation 6.2. We have assumed the first discharge
capacity is equivalent to the reversible capacity when measured against lithium  foil (half-cell
arrangement).
                                          Qrev
                                     *=                                            (6'2)
The quantity of lithium consumed from the positive electrode in the negative electrode SEI, gs
may be calculated from Equation 6.3.  Here,  [N/P] is the negative to positive capacity ratio.
                       Qsel =
After one full cycle, the remaining lithium in the positive electrode available for cycling, Qpaci, in
a full-cell configuration (positive  electrode versus  non-prelithiated negative electrode) may be
calculated by choosing  the minimum value determined in Equation 6.4 below. Here we see the
possibility  that the positive electrode is unable to accept the full amount of lithium  released
during the  first charge cycle. This  so called "irreversible capacity" of the positive electrode
results in lithium residing in the negative electrode. While this excess  lithium may require
additional negative electrode  capacity, it  also provides some beneficial aspects to  cycle and
calendar life. ° We have chosen to set the  [N/P] = 1.25 for layered oxides positive  electrodes
(NCA, NMC441, NMC333) due to their tendency to have  a lower first cycle efficiency ~ 88%.
The lithium manganese spinel and lithium  iron phosphate cells have a high first cycle efficency
and thus we selected a [N/P] = 1.20. For positive electrodes with a high first cycle efficiency, the
reversible capacity of the cell  is reduced by the lithium consumed in the graphite electrode SEI
during the  formation cycle. Conversely, the lithium titanate spinel negative electrode does not
form an SEI and  is significantly  safer than the graphite electrode as  discussed in Chapter 5.
Thefore the [N/P] ratio is set to 1.1 for the cells based on lithium titanate spinel.
                           {
               Qf = MIN 27  i +  —-  - N/P  —-   27                    (6.4)
                                            82

-------
                                7. Illustrated Results

The BatPaC model may be used to study the effects of battery parameters on the performance
and the manufactured cost of the designed battery packs. A few examples are given below for the
effects of various parameters on battery pack volume, weight and cost.

7.1 Number of Cells in Series

For a set battery pack power, the number of cells in the pack has substantial effects on the price
of the pack, the pack voltage and the maximum current. These effects are illustrated (Figure 7.1)
for NMC441-Gr PHEV25  batteries (providing 25-mile electric range) with 60-kW power at a
[V/U] = 0.8. The price of the pack increases by 17% in changing the number of series-connected
cells in  the pack from 32  to 96 and the entire pack integrated cost increases by 15.7%. The
integrated cost includes  additions to the vehicle air-conditioning system to provide for battery
cooling  and the battery management system with  disconnects. The change in the  maximum
current,  resulting from differing pack voltages, would also  affect the cost of the motor and the
electronic  converter and controller, but in the opposite direction. As a result of these offsetting
effects on the total cost of the electric drivetrain, a  study is required to determine  the optimum
current at  maximum power as a function of the total battery pack power and other parameters
(see the  Future Work section).
            _ 4000
-g 3500  -
s.
-O
v
M
ta
              3000 -
              2500
              2000 -
            o
   1500  -
 O 1000  -
 
- 200   Q)

- 100  £

  0    >
U
E
3
E
1

•o
E
re
                           20      40      60      80     100
                           Number of Series Connected Cells
                                                      120
Figure 7.1 The effect of the number of series-connected cells for NMC441-Gr, 60-kW, PHEV25
packs with 10.7 kWh total energy (70% useable).

Current PHEV battery technology uses battery packs containing 80-96 series connected cells.
However, these series  connections are often composed of parallel cell groups. For instance, the
                                          83

-------
battery used in the first production model of the Chevrolet Volt is in a 3P-96S configuration.21
Three low capacity cells are connected in parallel forming a parallel cell group. Then 96 parallel
cell groups are connected in series. The cost savings from moving to larger format cells with
only series connections is discussed later in this section.

7.2 Cathode Materials

Lithium-ion batteries for PHEVs and EVs do not require a high P/E ratio or low AST to meet
their goals. The most important material properties for performance are high  specific capacity
(mAh/g), high cell voltage, and high electrode density.  Since the graphite electrode  is almost
universal in commercial  cells  (although not all graphite is the same), changes in the cathode
result in dramatically different calculated batteries. To compare the  performance of EV battery
packs made from various Li-ion chemistries, we designed the packs to provide 150 kW at 360 V
(25% SOC)  for a [V/U]  = 0.8. Each pack consisted of six modules containing 16, 16, and 18
cells for the cell chemistries LMO-Gr, NMC441-Gr, and LFP-Gr, respectively.  This calculation
assumes that large capacity cells may be reliably produced. Moving to a parallel connection of
smaller capacity cells would result in higher cost as discussed later in the section.

The NMC441-Gr system  has excellent energy density and low cost (Fig. 7.2 and 7.3). The LMO-
Gr system is less energy dense  than the NMC441-Gr couple, but equivalent in calculated price to
the OEM. The LFP-Gr system results in a battery that is larger and more expensive than the other
two chemistries. The mass-specific cathode raw material prices are 29, 20,  and  10  $/kg  for
NMC441, LFP and LMO respectively. The  differences in initial cost do not directly translate to
the end cost of the battery. The performance (exhibited by specific  capacity and voltage) affect
the quantity of both active and inactive material required. The NMC441 material achieves 175
mAh/g at a good  cell  voltage  and is representative  of  an  advanced, although  close to
commercialization, layered oxide cathode.51 The combination in voltage and capacity results in a
superior energy density  compared to the  other cathodes. The LMO  cathode  has similar cell
voltage to NMC441 and  low raw material cost but also  a low specific capacity of 100 mAh/g.
This low capacity results in a positive electrode loading limited by the  maximum achievable
electrode thickness -100 microns. The LFP electrode has moderate capacity, 150 mAh/g, and
raw  material  cost, but  exhibits a  lower cell voltage and  electrode  density. These poor
performance characteristics result in  a low energy density battery with a high price.

7.3 Parallel-Connected Cell Groups and Electrode Thickness Limits

BatPaC also  allows  the  user to create parallel cell groups and to set a maximum  electrode
thickness. The effect these two unique design factors have on battery price are illustrated below
in Figure 7.4 for the LMO-Gr and NMC441-Gr systems. In  this illustration, the PHEV battery
pack design parameters are 100 kW of power at a [V/U]  = 0.8 and 17 kWh of total energy. The
nominal battery pack voltage (OCV at 50% SOC) is around 360 V from 96 cell groups connected
in series. The number of cells  in the parallel cell group is varied from a single cell (no parallel
connections) to four. Two maximum electrode thicknesses of  100 and 200 microns are shown for
the LMO-Gr chemistry. In contrast to the NMC441-Gr, the LMO-Gr chemistry benefits form
allowing  larger electrode thicknesses. The thickness of the positive  electrode is limiting  the
LMO-Gr chemistry while the thickness of the negative electrode limits the NMC441-Gr
                                           84

-------
   700
 5? 600 -
   500 -
 _
 ,2,400
 01
 E 300
 _3
 o
 ^.200

 s
 ra 100
 DO
-Volume NMC441
-MassNMC441
-Volume LMO
•MassLMO
-Volume LFP
-MassLFP
       50
      100
200
250
                                            150
                                    Vehicle range (mi)
Figure 7.2 Mass and volume of electric vehicle battery packs with lithium iron phosphate (LFP),
lithium manganese-spinel (LMO) and lithium nickel-manganese-cobalt oxide (NMC441)
positive electrodes versus graphite designed to deliver 150 kW of power at 360 V (25% SOC).
             18000
•— • 16000 -
•l/V
   14000 -|
u
O
O 12000 -

 re
 re
EQ
   10000 H
    sooo H
    6000 -I
    4000 -
    2000 -
-NMC441
-LMO
-LFP
         50
        100
 200
250
                                             150
                                     Vehicle range (mi)
Figure 7.3 Battery pack price to OEM for LFP-Gr, LMO-Gr and NMC441-Gr battery packs for
same designs as in Fig. 7.2. NMC441-Gr and LMO-Gr result in nearly the same price.
                               85

-------
chemistry. The calculated value for the NMC441-Gr system never exceeds 100 microns for this
P/E ratio. The LMO-Gr is the least expensive in all cases. However, the difference between the
two chemistries lessens with smaller limiting electrode thickness. The costs will become even
closer for lower designed P/E ratios. In general, thicker electrodes reduce the cost of the battery
pack by lessening  the  amount  of inactive materials  used (separator,  current collector, etc).
Moving to 300 microns allows for greater savings in the LMO-Gr design but not the NMC441-
Gr design. However, a lower P/E ratio design for NMC441-Gr would take advantage of electrode
thicknesses greater than 200 microns.
           &*-
The cell capacity is  shown for the NMC441-Gr case limited to 100 microns. While the exact
values will change with cell chemistry, they will all be similar. The cell capacity is reduced by
one half as  a single  cell is added in  parallel.  This  approach  is commonly  used by  cell
manufacturers and OEMs that cannot reliably produce  or successfully operate cells of high
capacity for transportation applications. However,  this approach also increases the price of the
battery pack. In this example, the price is increased by ~ $500 when an additional string of cells
is incorporated in a parallel arrangement.

The model calculations show that the lowest cost  battery pack  will utilize thick electrodes and
large  capacity cells. In  current practice, these two  approaches have yet  to  be successfully
implemented within the entire community. In the challenge of lowering costs, it is useful to point
out the largest gains come from the initial advances (e.g. moving from 100 to 200 micron limit).
After  that point, the benefits are diminishing.

7.4 Manufacturing Scale

The effects of manufacturing scale come into the cost calculation even if the annual number of
packs produced is unchanged, but the design is altered  (e.g. power is increased). For a fixed
design, the effect of changing the scale of operations depends on the fraction of the total price
that is made up of materials costs. Unit materials costs change little with scale whereas the costs
per pack for labor, capital and plant area may decline substantially with increasing production
rates,  especially at low  production rates, Fig. 7.5.

The lines in the graphs  are for the best-fit power relationships through  the  data with power
factors of -0.076, -0.077, -0.147, and -0.211 from the top curve to that at the bottom. The least
negative power factor  is for the battery pack with the highest  fraction of materials cost in the
total  pack  cost. The more  negative power factors  result from a decreasing contribution of
materials cost as a fraction of the total pack cost. These power factors for equations of the cost of
a single unit can be converted to factors relating the total annual cost of manufacturing similar to
Eq. 5.2 by adding  1.0  to each power factor. Thus  the factors become 0.924, 0.923, 0.853, and
0.789. These large factors show only a small to moderate effect  of scale. When the power curves
are compared to the points in each of the graphs of Fig. 7.5, it  is apparent that the scale factors
approach one as the scale increases. This is because the model assigns a value of 0.95 for the
active materials and  1.0 for the balance of the materials. As the production  level increases and
the materials costs become a larger fraction of the total price of the battery, the  scaling power
approaches 1.0 and the  effect of scale become very  small. Likewise, the  effect of scale on battery
price  is much larger for HEV batteries than for EVs because materials costs constitute a smaller
                                            86

-------
portion of the total cost for HEV batteries. Increasing the production rate for HEV batteries will
result in a more dramatic reduction in cost than increasing the production rate for EV batteries.
               6000
               5500


          _   5000  -
          •u>

          2   4500  -
          LJJ
          O
          O   4000
          01
          r:   3500
          D.

               3000

               2500


               2000
NMC441-Gr
                           1234
                         Number of cells in parallel group
                        65.0
55.0  ~
                     h  45.0  -f
                              TO
                              CL
                              TO
                     \-  35.0  ^
                              
-------
             LLI
  14000


  13000 -


  12000 -


  11000 -


  10000 -
              gjj  9000
                 8000 -


                 7000 -


                 6000
                        y=15609x
                                  -0.076
                        y=13935x-°-077
                   EV150NCA-Gr
                   EV150LMO-Gr
                     10
                           100
1000
                        Manufacturing rate (1000's/year)
                 3500
                 3000 -
vv

uu
O
2
OJ
                 2500 -
                 2000 -
                 1500 H
                 1000 -
                  500
                             PHEV20LMO-Gr
                             HEV-25 LMO-Gr
                          = 4931.1x
                                   -°-147
                        y=2462.3x
                                  -°-211
                     10                   100                 1000

                        Manufacturing rate (1000's /year)

Figure 7.5 The effects of manufacturing rate on the price calculated by the model for battery
packs of various  cell chemistries, power capabilities and vehicle  types. The EV and PHEV
batteries are composed of 96 cells and the HEV-25 is composed of 48 cells.

-------
                                   8. Future Work

8.1 Initial Power Designed at Differing Fractions of the Open-Circuit Voltage

The objective of this study is to evaluate the various benefits and costs associated with over- or
under- sizing the initial power of the battery utilizing Argonne National Laboratory's BatPaC
(battery performance and cost model)  and Autonomie (vehicle model). The fraction of the open-
circuit voltage at which initial designed battery power is achieved has a direct impact on heat
generation during operation, performance under cold conditions, and acceptable power fade. The
vehicle focus will be on PHEVs for a mid-size vehicle glider and also a small SUV. Coupling the
vehicle and battery models will allow  for a complete study evaluating the benefits and tradeoffs
associated with this  critical design  parameter. This work will be carried out in a  cooperative
effort with Argonne's Transportation Technology R&D  Center, which will  perform  vehicle
simulation tests to determine the rate of heat generation in the battery pack.

The end result of this study will be  quantitative justification for designing a battery at a set
fraction of the open-circuit voltage. The justification will likely depend on battery chemistry and
powertrain type as differing architectures lead to differing benefits. These benefits may lead to
changes in the system design to maximize net present value. A secondary, but just as important,
benefit will be extrapolation of thermal management requirements from heat generation and cold
temperature performance calculated from various levels of sizing the battery.


8.2 Optimum Battery Voltage for Minimum Drivetrain Cost

For a set cell chemistry and set battery pack power, the cost of the pack increases as the number
of cells and the pack voltage are  increased (Fig. 7.1). The additional cost results primarily from
the cost  of additional state-of-charge equalization circuits and the additional  number of cells
needing formation cycling and testing. The increase in battery pack cost is almost linear with the
increase  in the number  of  cells. As the  number of cells is decreased, the  pack current at
maximum power becomes very high and the cost of the balance of the drivetrain increases at an
accelerating  rate. Thus,  there must be a number of  cells and an associated pack current at
maximum power for which the total cost of the drivetrain is at a minimum. This minimum would
be for the current at full power for which the slope  of the cost curve  for the balance of the
drivetrain versus current was equal to the negative of slope of the cost of the pack versus current.
This optimum current will increase with the pack power because the slope of the cost-versus-
current  curve increases with increasing power and,  therefore, the optimum current will also
increase.

To represent these phenomena in illustrating the model in Section 7, we used an equation (Eq.
7.1) for selecting the current at full power as a function of the battery pack power. This equation
is  just an estimate and it does not  provide for differences in the optimum current that would
result from differences in cell chemistries, which are known to affect the battery cost versus
power function.
                                           89

-------
A study is needed to determine the cost of electric motors and the electronic control equipment
required for the vehicle and battery pack as a function of power and maximum current capability.
Once appropriate cost curves are established, we will determine equations relating the optimum
battery current to the desired battery power taking into account the battery chemistry. We intend
to do this study with the cooperation of Argonne's Transportation Technology R&D Center.

8.3 Multipurpose Battery Manufacturing Plants

Our cost modeling is based on the concept that a manufacturing plant is constructed to produce a
single type of battery pack at a predetermined level of production. In practice, manufacturers will
have to produce several types of batteries within the same manufacturing plant and the levels of
production may fluctuate. We intend to investigate how this increased flexibility requirement
will  affect the various manufacturing costs. For example, the cell filling and sealing equipment
may be required to handle cells of different dimensions. This would most likely increase the
capital cost for this equipment. Alternatively, additional packaging and sealing lines might be
needed. We intend to evaluate combinations of vehicle battery  packs that are easily integrated
into the same plant. The manufacturing cost will most likely increase after these considerations
are built into  the model.

8.4 Stand-Alone Graphical User Interface for Model

The  spreadsheet  program described here-in  allows  versatility  in designing the battery  and in
calculating the  costs, but like all complex  spreadsheet programs it is not  user-friendly to those
unfamiliar with the details of calculation. In addition, the model is easily corrupted by  a poor
choice of input parameters. As a result, the final spreadsheet  program will be converted to a
stand-alone user-friendly application, primarily with the efforts  of Ira Bloom. Visual Basic for
Applications  will be used to hard code  in the model calculations and to also create the graphical
interface.  The  new user  interface  should allow for a  wide distribution  of the model  while
maintaining the ability  to change  the vast majority of  input parameters.  The retention  of  this
flexibility should make the model a valuable tool for those  interested in batteries regardless of
the specific material property or manufacturing cost structure the user seeks to analyze.
                                            90

-------
                                  REFERENCES

1. M. Anderman, F. Kalhammer, and D. MacArthur. "Advanced Batteries for Electric Vehicles:
An Assessment of Performance, Cost, and Availability." California Air Resources Board, June
(2000). Available from http://www.arb.ca.gov/msprog/zevprog/2000review/btap report.doc
(accessed on December 17, 2010).

2. B. Barnett, D. Ofer, C. McCoy, Y. Yang, T. Rhodes, B. Oh, M. Hastbacka, J. Rempel, and S.
Sririramulu "PHEV Battery Cost Assessment," 2009 DOE Merit Review, May (2009). Available
fromhttp://wwwl.eere.energy.gov/vehiclesandfuels/pdfs/merit_review_2009/energy_storage/
es_02_barnett.pdf (accessed on December 11, 2010).

3. B. Barnett, J. Rempel, D. Ofer, B. Oh, S. Sriramulu, J. Sinha, M. Hastbacka, and C. McCoy,
"PHEV Battery Cost Assessment," 2010 DOE Merit Review, June (2010). Available from
http://wwwl.eere.energy.gov/vehiclesandfuels/pdfs/merit_review_2010/electrochemical
_storage/esOO l_barnett_2010_o.pdf (accessed on December 11, 2010).

4. A. Dinger, R. Martin, X. Mosquet, M. Rabl, D. Rizoulis, M. Russo, and G. Sticher, "Batteries
for Electric Vehicles: Challenges, Opportunities, and the Outlook to 2020." The Boston
Consulting Group, Available from http://www.bcg.com/documents/file36615.pdf (accessed on
December 11,2010).

5. L. Gaines and R. Cuenca. "Costs of Lithium-Ion Batteries for Vehicles," Center for
Transportation Research, Energy Systems Division, Argonne National Laboratory, ANL/ESD-
42, Argonne, IL May (2000).

6. F.R. Kalhammer, B.M.  Kopf, D.H. Swan, V.P. Roan, M.P. Walsh, "Status and Prospects for
Zero Emissions Vehicle Technology: Report of the ARE Independent Expert Panel 2007,"
California Air Resources Board, April (2007). Available from http://www.arb.ca.gov/
msprog/zevprog/zevreview/zev_panel_report.pdf (accessed on December 17, 2010).

7. M. A. Kromer and J. B. Heywood, "Electric Powertrains: Opportunities and Challenges in the
U.S. Light-Duty Vehicle Fleet," Sloan Automotive Laboratory, Laboratory for Energy and the
Environment, Massachusetts Institute of Technology, LFEE 2007-03 RP, Cambridge, MA
(2007).

8. P. Mock, "Assessment of Future Li-Ion Battery Production Costs," presented at Plug-in 2009,
Long Beach, CA, (2009).

9. National Research Council of the National Academies, "Transitions to Alternative
Transportation Technologies - Plug-in Hybrid Electric Vehicles," The National Academies
Press, Washington, D.C. (2010). Available from http://www.nap.edu/catalog.php?record_id=
12826  (accessed on December 17, 2010).

10.  TIAX LLC, "Cost Assessment for Plug-In Hybrid Vehicles (SOW-4656)," Report to US
DOE Office of Transportation Technology,  October (2007).
                                          91

-------
11. M. Anderman "The Plug-In Hybrid and Electric Vehicle Opportunity Report: A critical
assessment of the emerging market and its key underlying technology: Li-Ion batteries,"
Advanced Automotive Batteries, May (2010).

12. P.A. Nelson, DJ. Santini, J. Barnes, "Factors Determining the Manufacturing Costs of
Lithium-Ion Batteries for PHEVs." International Electric Vehicles Symposium EVS-24,
Stavanger, Norway, (2009).

13. DJ. Santini, K.G. Gallagher, P.A. Nelson, "Modeling the Manufacturing Costs of Lithium-
Ion Batteries for HEVs, PHEVs, and EVs," International Electric Vehicles Symposium EVS-25,
Shenzhen, China, (2010).

14. K.G. Gallagher, P.A. Nelson, D.W. Dees, "PHEV battery cost assessment" 2011 DOE Merit
Review Presentation Washington, DC May 9-12 (2011). Available from http://wwwl.eere.
energy.gov/vehiclesandfuels/pdfs/merit_review_201 l/electrochemical_storage/esl 1 l_gallagher_
201 l_o.pdf (Accessed on July 12, 2011).

15. P. Nelson, I. Bloom, K. Amine, G. Henriksen, "Design modeling of lithium-ion battery
performance," Journal of Power Sources, 110, 437 (2002).

16. P. Nelson, D. Dees, K. Amine, G. Henriksen, "Modeling thermal management of lithium-ion
PNGV batteries," Journal of Power Sources, 110, 349 (2002).

17. G.L. Henriksen, K. Amine, J. Liu, and P.A. Nelson, "Materials Cost Evaluation Report for
High-Power Li-Ion HEV Batteries," Electrochemical Technology Program, Chemical
Technology Division, Argonne National Laboratory, ANL-03/05, Argonne, IL (2002).

18. P. Nelson, K. Amine,  A. Rousseau, H. Yomoto, "Advanced Lithium-Ion Batteries for Plug-in
Hybrid-Electric  Vehicles", International Electric Vehicles Symposium, EVS-23, Anaheim, Ca
(2007).

19. P. Nelson, "Modeling the Manufacturing Costs of Lithium-Ion Batteries for PHEVs,"
presented at Plug-in 2009, Long Beach, CA, (2009).

20. K.G. Gallagher, P.A. Nelson, and D.W. Dees, "Simplified Calculation of the Area Specific
Impedance for Battery Design," Journal of Power Sources, 196 2289 (2011).

21. "Chevrolet Volt: Development Story of the Pioneering Electrified Vehicle" ed. Lindsay
Brooke,  SAE International, Warrendale, PA pgs 215 (2011)

22. A.N. Jansen, K. Amine, and G. L. Henriksen, "Low-Cost Flexible Packaging for High-Power
Li-Ion HEV Batteries," Electrochemical Technology Program, Chemical Technology Division,
Argonne National Laboratory, ANL-04/09, Argonne, IL (2004).
                                          92

-------
23. P.A Nelson and A.N. Jansen, "Comparative Costs of Flexible Package Cells and Rigid Cells
for Lithium-Ion Hybrid Electric Vehicle Batteries," Electrochemical Technology Program,
Chemical Technology Division, Argonne National Laboratory, ANL-06/43, Argonne, IL (2006).

24. J. Euler and W. Nonnenmacher, "Stromverteilun in Porosen Elektroden," Electrochim. Acta,
2268(1960).

25. J.S. Newman and C.W. Tobias, "Theoretical Analysis of Current Distributions in Porous
Electrodes," J. Electrochem. Soc., 109 1183 (1962).
Electrode thickness limitations

26. "Diagnostic Examination of Generation 2 Lithium-Ion Cells and Assessment of Performance
Degradation Mechanisms," ANL Report, ANL-05/21, July 2005.

27. D. Dees, E. Gunen, D. Abraham, A. Jansen, and J. Prakash, "Electrochemical Modeling of
Lithium-Ion Positive Electrodes during Hybrid Pulse Power Characterization Tests" J.
Electrochem. Soc., 155 (8) A603 (2008).

28. D. Dees, E. Gunen, D. Abraham, A. Jansen, and J. Prakash, "Alternating current impedance
electrochemical modeling of lithium-ion positive electrodes," J. Electrochem. Soc., 152
(7),A1409 (2005).

29. D. Abraham, S. Kawauchi, and D. Dees, "Modeling the impedance versus voltage
characteristics of LiNio.gCoo.isAlo.osCV Electrochim. Acta, 53 2121-2129 (2008).
30. D. P. Abraham, S. D. Poppen, A. N. Jansen, J. Liu, and D. W. Dees, "Application of a
lithium-tin reference electrode to determine electrode contributions to impedance rise in high-
power lithium-ion cells," Electrochim. Acta, 49 4763 (2004).

31. D. P. Abraham, E. M. Reynolds, E. Sammann, A. N. Jansen, and D. W. Dees, "Aging
characteristics of high-power lithium-ion cells with LiNiO.8CoO.15A10.05O2 and Li4/3Ti5/3O4
electrodes," Electrochim. Acta, 51, 502 (2005).

32. D.P. Abraham, J. Liu, C.H. Chen, Y.E. Hyung, M. Stoll, N. Elsen, S. MacLaren, R. Twesten,
R. Haasch, E. Sammann, I. Petrov, K. Amine, G. Henriksen, J. Power Sources, 119-121 511
(2003).

33. D. P. Abraham, R. D. Twesten, M. Balasubramanian, J. Kropf, D. Fischer, J. McBreen, I.
Petrov, and K. Amine, "Microscopy and spectroscopy of lithium nickel oxide-based particles
used in high power lithium-ion cells," J. Electrochem. Soc., 150 (11) A1450 (2003).

34. A.M. Andersson, D.P. Abraham, R. Haasch, S. MacLaren,  J. Liu, K. Amine, "Surface
characterization of electrodes from high power lithium-ion batteries," J. Electrochem. Soc., 149
A1358 (2002).
                                          93

-------
35. D. Abraham, J. Knuth, D. Dees, I. Bloom, and J. Christophersen, "Performance degradation
of high-power lithium-ion cells - Electrochemistry of harvested electrodes," J. Power Sources,
170 465 (2007).

36. J. Newman and K. E. Thomas-Alyea, "Electrochemical Systems," 3rd ed., Wiley
Interscience, New York (2004).

37. Wei Lai, Can K. Erdonmez, Thomas F. Marinis, Caroline K. Bjune, Nancy J. Dudney, Fan
Xu, Ryan Wartena, and Yet-Ming Chiang, "Ultrahigh-Energy-Density Microbatteries Enabled
by New Electrode Architecture and Micropackaging Design" Advanced Materials, 22, E139-
E144 (2010).

38. Pankaj Arora, Marc Doyle, and Ralph E. White, "Mathematical Modeling of the Lithium
Deoposition Overcharge Reaction in Lithium-Ion Batteries Using Carbon-Based Negative
Electrodes," J. Electrochem. Soc. 146 (10) 3542-3553 (1999).

39. Y. Chen and J.W. Evans, "Thermal Analysis of Lithium-Ion Batteries" J. Electrochem. Soc.
143(9)2708-2712(1996).

40. S.C. Chen, C.C. Wan, and Y.Y. Wan, "Thermal Analysis of lithium-ion batteries" J. Power
Sources 140 111-124(2005).

41. H. Maleki, S.A. Hallaj, J.R. Selman, R.B. Dinwiddie, and H. Wang,  "Thermal Properties of
Lithium-Ion Battery and Components" J. Electrochem. Soc. 146 (3) 947-954  (1999).

42. M.  Sievers,  U.  Sievers,  and  S.S. Mao "Thermal  modelling of new Li-ion cell design
modifications" Forsch Ingenieurwes 74 215-231 (2010).

43. T.M. Bandhauer, S.  Garimella, and T.F. Fuller  "A Critical Review of Thermal Issues in
Lithium-Ion Batteries" J.  Electrochem.  Soc. 158 (3) R1-R25 (2011).

44. R. Byron Bird, Warren E. Stewart,  and Edwin N. Lightfoot, "Transport Phenomena," 2nd Ed.
John Wiley & Sons, Inc, New York, NY (2002).

45. R.K. Shah and A.L. London, "Laminar Flow Forced Convection in Ducts" Advances in Heat
Transfer Supplement, Academies Press, New York, NY (1978).

46. R. Siegel, E.M. Sparrow, and T. M. Hallman, "Steady Laminar Flow Heat Transfer in a
Circular Tube with Prescribed Wall Heat Flux," Appl. sci. Res. Section A, 7 386-392 (1958).

47. M. Nickolay and H. Martin, "Improved approximation for the Nusselt number for
hydrodynamically developed laminar flow between parallel plates," Int. J. Heat Mass Tran. 45
3263-3266 (2002).

48. United States Geological Survery, Commodity Statistics and Information, Available from
http://minerals.usgs.gov/minerals/pubs/commodity/ (accessed on December 17, 2010).
                                          94

-------
49. R.H. Perry, D.W. Green, and J.O. Maloney, Perry's Chemical Engineers' Handbook, Sixth
Edition, pp: 25-68 (1984).

50. M. Broussely, Ph. Biensan, F. Bonhomme, Ph. Blanchard, S. Herreyre, K. Nechev, and RJ.
Staniewicz, "Main aging mechanisms in Li ion batteries," J. Power Sources 146 90-96 (2005).

51. Sun-Ho Kang, Wenquan Lu, Kevin G. Gallagher, Sang-Ho Park, and Vilas G. Pol, "Study of
Lil+x(Mn4/9Col/9Ni4/9)l-xO2  Cathode  Materials for Vehicle  Battery  Applications" J.
Electrochem. Soc. 158 (8) A936-A941 (2011).
                                         95

-------