Technical documentation:

Anja Streit
Fabian Biebl
Barbara Planas

BATTERYDICT
User Guide
GeoDict release 2020
October 16, 2019

Fabian Biebl | Ilona Glatt | Mathias Fingerle

BATTERYDICT: CHARGING & DISCHARGING SIMULATIONS FOR LI-
ION BATTERIES 1
THEORETICAL BASIS 3
STATE OF CHARGE (SOC) 3
BATTERYDICT SIMULATION OF CHARGING AND DISCHARGING PROCESS 5
SIMULATION PARAMETERS 7
Tested parameter space 8

BATTERYDICT COMPUTATIONS 9
DESIGN BATTERY 10
Geometry 10
Anode IDs and Cathode IDs 12
CHOOSING A SOLVER 14
CHARGE BATTERY (BESTMICRO) 16
Geometry 16
Anode, Cathode, and Electrolyte/Separator 17
Solver Options 19
Equations 21
CHARGE BATTERY (BESTMICROFFT) 22
BATTERYDICT COMPUTATION RESULTS 24
RESULTS OF DESIGN BATTERY 24
RESULTS OF CHARGE BATTERY (BESTMICROFFT AND BESTMICRO) 30
RESULT FOLDERS 43
BatteryDesigner: the design battery folder 43
BESTmicroFFT and BESTmicro: the charge battery folders 43
B ATTERY D ICT : C HARGING & D ISCHARGING
SIMULATIONS FOR L I - ION BATTERIES

BatteryDict simulates the performance of secondary lithium ion batteries during

charging and discharging on the micro scale. With the Battery Designer, virtual cells
can be created, based on 3D structures of electrode materials with a representative
elementary volume (REV). The numerical calculation of the electrochemical processes
governing ionic transport in the different material phases and across their interfaces
are based on local charge and mass neutrality together with a Butler-Volmer reaction
model.
A battery in BatteryDict is composed of four different components: cathode, anode,
separator and the current collectors. The last two are represented as standardized
layers, whereas the cathode and anode are individually modeled structures, consisting
of multiple active materials, binder and electrolyte. Up to two different active
materials can be handled for each electrode from GeoDict 2019 on. From GeoDict
2020 on, it is also possible to perform a half-cell simulation, by modelling cathode or
anode as a lithium reservoir.
The input data for the electrode models can be either scanned data from µCT or
FIB/SEM images of real cathode and anode materials, or realistic structures created
in GeoDict. BatteryDict is capable of handling voxel-based structures, with a volume
large enough (REV) to simulate the behavior of a real battery electrode. For each
active material, the Open-Circuit Potential (OCV) curve can be specified and the
characteristic values for ionic and electronic transport in the active materials,
electrolyte and binder and carbon black, respectively, can be set. The separator and
current collectors are represented as homogeneous materials with a specified
thickness.
The model used for simulation and the parameters to be defined are explained in
more detail in the section Theoretical Basis. In the section BatteryDict Computations,
the use of the BatteryDict GUI and the different simulation options are outlined. The
results of the simulations and their interpretation are discussed in the section
BatteryDict Computation Results.
The characteristic application for BatteryDict is to simulate and optimize the charging
performance of a battery cell, based on realistic electrode microstructures. For
analyzing or modeling electrode microstructures, also the use of other GeoDict
modules is emphasized. Many properties of the microstructure, like porosity, pore size
distribution, surface area, tortuosity, thermal conductivity, thermal flux, electric
conductivity, electric flux, thermal expansion, permeability, diffusivity, etc. are vital
for the performance of a battery. The GeoDict modules FlowDict, GrainGeo, GrainFind,
PoroDict, DiffuDict and ConductoDict can be used to analyze these properties and to
obtain detailed information about the microstructure of the electrodes.
With ImportGeo-Vol 3D image data can be imported and segmented and be further
processed with the modules mentioned above. GrainFind and GrainGeo are powerful
tools to create Digital Twins based on image data or realistic microstructures from
scratch. Please refer to the other sections of the GeoDict User Guide for more
information.

GeoDict 2020 User Guide 1

BatteryDict simulations of charging Li ion batteries

2 GeoDict 2020 User Guide

 Theoretical basis

T HEORETICAL B ASIS
Upon charging a typical secondary Li-ion battery, the Li+ ions initially stored in the
active material of the cathode are transported to the anode side. During this process,
the cathode active material is oxidized and Li+ ions de-intercalate and diffuse through
the electrolyte, towards the anode. This electrochemical process is induced by the
applied charging potential, that drives the transport of free electrons through the
cathode active material and through the binder with carbon black to the current
collector. On the anode side, electrons flow from the current collector through the
carbon black and active material. The active material of the anode is reduced by the
Li+ intercalation.
During discharging, Li+ ions are transported in a corresponding process to the cathode
side, while the electron flow powers the consumer device. The amount of lithium an
electrode material can store per volume is defined as the volumetric energy density.

S TATE OF C HARGE (SOC)

There are 3 different states of charge in a battery:

Anode state of charge: It measures the Li-ion

content of the anode vs. the maximal Li-ion content
of the anode. The Li-ion content is the volume
integral of the Li-ion concentration.

For an active material voxel, the SOC is the voxel’s

Li-ion concentration divided by the voxel’s maximum
Li ion concentration.

The anode SOC is the mean value of the voxel SOCs

of all active-material voxels within the anode.

Cathode state of charge: It measures the Li-ion

content of the cathode vs. the maximal Li-ion
content of the cathode. The Li-ion content is the
volume integral of the Li-ion concentration.

The cathode SOC is the mean value of the voxel

SOCs of all active-material voxels in the cathode

GeoDict 2020 User Guide 3

BatteryDict simulations of charging Li ion batteries

Cell state of charge: It describes the overall state of charge of the battery cell. It is
0% if the anode is empty OR the cathode is full. It is 100% if the anode is full OR the
cathode is empty.

For a half-cell simulation the lithium reservoir has always the same capacity as
the other electrode.

4 GeoDict 2020 User Guide

 Theoretical basis

B ATTERY D ICT SIMULATION OF CHARGING AND DISCHARGING PROCESS

The main physical principles, describing the charging or discharging simulation in

BatteryDict, are conservation of mass for lithium atoms and conservation of charge.
The length scale considered in BatteryDict is above the scale where positive and
negative charges can be resolved. Therefore, no free charges need to be considered.
In the following, we shortly explain, how the principles of mass and charge
conservation apply in the electrolyte and in the active material:
1. In the electrolyte: The movement of Li ions is simulated in the whole electrolyte.
Due to the mass conservation for Li ions, the change of Li-ion concentration in
time is equivalent to the divergence of the flux of the Li ions. The movement of Li
ions, and therefore the flux, is driven by diffusion due to differences in Li-ion
concentration and by migration, due to the applied current. In the case of constant
Li transference number t+, the divergence of the migration is zero.
The current density of the electrolyte is given by gradients in the electric potential
and in the Li ion concentration. Due to the charge conservation, the sum of both
is zero.
2. In the active material: The movement of Li atoms is not coupled to the electric
field here. The reason is that Li ions entering the active material, are combined
directly with an electron. The resulting Li atoms do not contribute to the electric
current. Mass and charge conservation equations are therefore simpler compared
to the ones in the electrolyte. Again, changes in time of the Li concentration are
equivalent to the divergence of the flux of the Li concentration. This flux consists
here only of one term, the diffusion of Li atoms.
The current density is defined by gradients in the electric potential alone. This is
Ohm’s Law for the electrons.
3. The interface of electrolyte and active material is described by the Butler-
Volmer interface current density. This is the flux of Lithium from the active material
into the electrolyte and vice versa. This flux depends on the Li concentrations in
both the electrolyte and the active material, and on the difference in chemical
potentials. ise >0 means deintercalation, i.e. movement of Lithium from the active
material to the electrolyte, ise <0 intercalation.
Furthermore, during the intercalation of Li ions from the electrolyte into the active
material, these Li ions attain one electron per ion and become neutral in charge.
During the de-intercalation of Li from the active material into the electrolyte, every
Li atom loses one electron and becomes a positively charged Li-ion.
Hence, in the solid, the charge transport is performed solely by electrons, whereas
in the electrolyte, the charge transport is performed solely by ions.

Boundary condition at the current collectors: In BatteryDict, a constant current

is applied to the current collectors of the battery cell. This current can also be specified
via a charge rate, which has the advantage of easily defining in which time the battery
should charge. In the simulation result, it can easily be seen how far the cell potential
will rise for this charge rate, and also see if the charge rate diverges at some point,
which means that the battery may have been charged too fast.
This process is simulated at constant current, not with constant potential. Simulating
at constant current has the advantage of seeing when the battery cannot be charged
at a constant charge rate anymore due to the cell potential becoming too high.

GeoDict 2020 User Guide 5

BatteryDict simulations of charging Li ion batteries

6 GeoDict 2020 User Guide

 Theoretical basis

S IMULATION P ARAMETERS

The parameters occurring in the equations are explained in the following:

▪ Maximum Lithium Concentration: Maximum Lithium concentration that can be

stored to or taken out of the active material.

▪ Electronic Conductivity: Measures a materials ability to conduct electrons.

▪ Ionic Conductivity: Measures the electrolyte’s ability to conduct ions (such as Li

ions).

▪ Ionic Diffusion Constant: Measures a material’s ability to equilibrate differences

in concentration.

▪ Butler-Volmer Rate Constant: This is not a material parameter of one material

alone but depends on the active material as well as on the electrolyte. It is
necessary to define the interface condition between voxels of active material and
those of electrolyte. It characterizes “how easy” ions can be exchanged between
active material and electrolyte.

▪ Open-Circuit Potential (OCV) Function: The open-circuit potential is directly

related to the chemical potential of lithium within the active material.
Experimentally it is determined by measuring the potential difference between the
respective electrode and a lithium counter electrode in equilibrium. It depends on
the charging state of the electrode and is therefore a function of the state of charge
(SOC) of the electrode soc=c/cmax.

▪ Equilibrium Lithium Concentration: Concentration of lithium ions in the

electrolyte solution. This is the lithium concentration of the electrolyte as long as
it is in thermodynamic equilibrium. As soon as the battery starts charging or
discharging, local concentration gradients will build up in the electrolyte.

▪ Transference Number: In the absence of concentration gradients, the

transference number measures the fraction of ionic current carried by lithium ions.

References
A. Latz and J. Zausch. Thermodynamic consistent transport theory of Li-ion batteries.
Journal of Power Sources 196(6):3296-3302, 2011
J. Newman, K. Thomas-Alyea, Electrochemical systems, 2004, Wiley, ISBN
3175723993
A. Latz and J. Zausch. Thermodynamic derivation of a Butler–Volmer model for
intercalation in Li - ion batteries. Electrochim. Act., 110:358–362, 2013
T. Hofmann, H. Andrä, R. Müller, J. Zausch, Numerical simulation of phase separation
in cathode materials of lithium ion batteries, Int. J. of Solids & Struc., 100-101:456-
469, 2016

T. Hofmann, H. Andrä, R. Müller. A fast immersed interface method for the Cahn -
Hilliard equation with arbitrary boundary conditions in complex domains. Comp. Mat.
Sci, 140C:22-31, 2017.

GeoDict 2020 User Guide 7

BatteryDict simulations of charging Li ion batteries

T ESTED PARAMETER SPACE

For any given set of material parameters (like conductivities), material structure
parameters (morphological), and simulation parameters (like the applied current),
the solver in BatteryDict strives to perform a physically accurate simulation. This is
done by calculating the time evolution of the battery simulation step-by-step.
First, the solver works to establish the initial state or equilibrium state for every
simulation in BatteryDict. Then, the solver undertakes performing time steps. In every
time step, the solver first attempts a time step with the “maximum time step” given
by the user. If this time step does not result in a physically meaningful state of the
battery cell, the solver reduces the time step and tries again. This time-step reduction
is repeated until the solver finds a physically meaningful state. If it cannot find one,
the solver produces a warning message indicating that “Time step has not converged”.
In some cases, the cell-potential might rise well above any reasonable end-of-
charging voltage, like for example if a large charge rate is applied. The solver might
reach 10 V or even 100 V, but at some point, it will stop. In these kinds of situations,
the cell-potential diverges, but the solver produces a meaningful and helpful result.
However, there are other situations in which the cell-potential does not diverge, but
the solver stops regardless of that and, as mentioned above, the solver has not
converged.

In BatteryDict 2020, the possible reasons why the solver does not converge are
revealed in warnings, shown before the solver starts and written into the report (see
page 14ff.). The warnings help figuring out why the solver stopped and include:

▪ not having enough RAM,

▪ having very large particles with very low ionic diffusion at a very fast charge rate,
▪ having chosen untested material parameters.
For the third, different tests have been carried out during the development of
BatteryDict.
One of the tests is a variation of the parameters for a realistic example structure,
starting from the default values of the BatteryDict options dialog. From the default
values, parameters were varied one at a time, and it was checked, whether the solver
still produced meaningful results. Limits for this variation are defined by the regime,
in which the parameter does not make sense physically anymore, or the solver does
not produce meaningful results. This leads to a n-dimensional box. For a parameter
combination in this box, it is very probable that the simulation will work, but it is not
guaranteed since parameters are varied only one-by-one and the convergence
behavior of the solver depends also on the geometry.
If the parameter combination chosen in the BatteryDict options dialog is outside the
tested parameter range, a warning appears. This does not mean that the simulation
will not work, but the user should check carefully if parameters are chosen in a
physically meaningful range. If you have chosen physically meaningful parameters,
but the solver does not produce meaningful results, please let us know by sending
email to our support (support(at)math2market.de).
If two active materials are used in an electrode, an additional check controls that the
value ranges of the OCV curves of both materials overlap. Otherwise there is no well-
defined equilibrium state of the battery cell, which is the starting point for the
simulation. This equilibrium state is defined by the SOC range given in the BatteryDict
options dialog.

8 GeoDict 2020 User Guide

 BatteryDict computations

B ATTERY D ICT C OMPUTATIONS

Start BatteryDict by selecting Predict → BatteryDict in the menu bar.

In the BatteryDict section, the pull-down menu includes the possibility to Design a
Battery and to simulate the charging of a battery with BESTmicroFFT (Charge
Battery (BESTmicroFFT)) and BESTmicro (Charge Battery (BESTmicro)).

GeoDict 2020 User Guide 9

BatteryDict simulations of charging Li ion batteries

D ESIGN B ATTERY
To design a battery, choose Design Battery from the pull-down menu and click Edit
to choose the options.
The Design Battery dialog opens, containing three tabs where to define the
morphology of the battery (Geometry tab) and the structure for anode and cathode.

G EOMETRY
For both electrodes, either the current structure loaded in the GeoDict GUI can be
used, a previously saved structure can be loaded, or the electrode can be modelled
as a lithium reservoir.
Later on, we show how to model the electrode as a lithium reservoir to perform a
half-cell simulation with only one electrode modelled as a microstructure.

Choose Load Structure and use the Browse… buttons to select an input file for
anode and cathode, in *.gad or *.gdt GeoDict formats.
After opening these files, a 2D representation of the anode and of the cathode appear
in the diagram. For the example battery shown here, the anode material is a structure
created in GeoDict using the GrainGeo module and the cathode material is part of a
FIB/SEM image of a real cathode.

▪ A red dot is visible to the right of the Anode and Cathode buttons, if no files are
loaded at all.

▪ A yellow dot appears if a file is loaded but the dimensions or voxel length of the
anode and the cathode do not match.

▪ A green dot appears if the anode and cathode files are loaded successfully, the
dimensions of the two match each other, and the voxel length is the same for both
microstructures.

10 GeoDict 2020 User Guide

 BatteryDict computations

The Size of anode and cathode are shown as the number of voxels in each direction
below the diagram of the battery, and the Voxel Length is also displayed. After
choosing anode and cathode, the total number of voxels and total length of the
complete battery structure are displayed for each direction above and to the left of
the diagram.

The Rotate around X, Y and Z buttons are used to rotate the structure 90° in X, Y
or Z direction. Click Reset to go back to the original orientation of anode and cathode
in the battery structure.

GeoDict 2020 User Guide 11

BatteryDict simulations of charging Li ion batteries

The number of voxels of the separator can be defined in the GUI. The default is 3
voxels. The length of anode and cathode collector is fixed to be three voxels and
cannot be changed.
Finally, the Boundary Conditions in Tangential Direction are chosen in this dialog
box. Periodic or Symmetric boundary conditions can be selected. For periodic
boundary conditions the structure is assumed to repeat periodically in y- and z-
direction and for symmetric boundary conditions, the Li+ ions and electrons are
reflected at the boundary in y- and z-direction.
Note that the selection of boundary conditions is not stored in the *.gdt file containing
the created structure file.
Boundary conditions are applied while using the battery designer to identify and show
connected and unconnected material, since this depends on the continuation of the
material in y- and z-direction. For a charging simulation, it can be chosen independent
of the value selected here. Connected and unconnected material is identified again
during charging simulation dependent on the boundary conditions selected then.

A NODE ID S AND C ATHODE ID S

The next step is to assign electrolyte, active materials, as well as binder and carbon
black from the materials available in the structures of anode and cathode.

12 GeoDict 2020 User Guide

 BatteryDict computations

Under the Anode IDs and Cathode IDs tabs one or several Material IDs of the
structures can be assigned to electrolyte, binder & carbon black, and two different
active materials. In the example shown here, the anode consists of Active Material 1,
whereas the cathode contains Active Material 1 and Active Material 2. Other
possibilities are greyed out.

Click OK in the Design Battery dialog, to keep the parameters selected in memory,
or Cancel to discard the parameters.
Click Run in the BatteryDict section of the GUI to create the battery.
If the battery is created successfully, a result file with the default name
BatteryDesigner.gdr is created in the selected project folder, as well as a structure
file (*.gdt) in the subfolder BatteryDesigner.
The Result Viewer of the result file opens automatically after the generation, showing
the information of BatteryDesigner.gdr (see below Results of Design Battery).

GeoDict 2020 User Guide 13

BatteryDict simulations of charging Li ion batteries

C HOOSING A S OLVER
Choosing a solver is the first issue to address before running a simulation. In essence,
the particular advantages of the solvers BESTmicro and BESTmicroFFT are:

▪ BESTmicro is more robust and has been tested over a decade. The 2020
version is quite faster and its RAM consumption has been reduced.
▪ BESTmicroFFT is a little faster and needs less memory. However, Fraunhofer
ITWM has stopped developing BESTmicroFFT any further. They have fully
concentrated their efforts on improving BESTmicro. So, we strongly recommend
using BESTmicro instead of BESTmicroFFT.
In detail, the advantages of the two solvers are listed in the following table.

BESTmicro BESTmicroFFT

Experienced and robust solver. It Relatively new solver. No more

possesses many years of testing enhancements or bug fixes
within the CoRheoS1 framework. possible due to ceased
Robustness development at Fraunhofer
ITWM. Please use BESTmicro, if
you experience malfunctions
occurring with BESTmicroFFT.

Needs a considerable amount of Needs substantially less RAM. On

RAM (about 2.5 kB per Voxel). a normal workstation, the RAM is
Example: With 25 GB free RAM, not the limiting factor. There,
Memory one can simulate structures the solver should be able to
containing up to 10 million voxels. simulate using a battery cell
E.g., in this case, a structure of structure of size 500x500x500.
size 250x200x200 is possible).

• Since BatteryDict 2020, half-cell

simulations can be performed.
• Since BatteryDict 2020, an
estimation of the charge curve
out of equilibrium is available.
Additional • The separator can optionally be
treated as if the electrolyte
would be contained in a micro-
porous structure, in which the
pores are not resolved by the
voxel length.

Uses algebraic multigrid methods. Uses Fast Fourier Transform.

The solver needs at least 5 voxels It works best for periodic
in flow direction and at least 3 boundary conditions.
voxels in the other directions.
Method The solver is most efficient, if the
Otherwise it may not produce an
structure has 2n-1 voxels per
output. In case that your battery
edge, i.e. 2n-1 ∈ {1, 3, 7, 15, 31,
is too small, use ProcessGeo‘s
63, 127, 255, 511, …}. The
“Rescale” method to increase the
number n may be different for
size of your battery.
every edge and is an integer

14 GeoDict 2020 User Guide

 BatteryDict computations

number, i.e. n ∈ {1, 2, 3, 4, …}

and so on.
In case of periodic boundary
conditions, one can also choose
2n voxels per edge, i.e. 2n ∈ {2,
4, 8, 16, 32, 64, 128, 256, 512,
…}.
Also, numbers that are a little
below these optimal number
work nicely. However, one voxel
more than the optimal number of
voxels may cause a noticeable
loss of performance.

May produce noisy charge

curves above about 4 Volts of
Known
cell potential. In the noisy
issues
regions, the result fields may not
be very reliable.
1 see https://www.itwm.fraunhofer.de/de/abteilungen/sms/produkte-und-leistungen/best-battery-and-
electrochemistry-simulation-tool.html

GeoDict 2020 User Guide 15

BatteryDict simulations of charging Li ion batteries

C HARGE B ATTERY (BEST MICRO )

To simulate the charging or discharging of a battery with BESTmicro, choose Charge

Battery (BESTmicro) from the pull-down menu in the BatteryDict section.
Then, click Edit to adjust the solver options.

The BatteryDict (BESTmicro) dialog opens, containing six tabs.

G EOMETRY
Several options are available as Battery Input Mode to select the battery itself

▪ Design Battery opens the Battery Designer (see Design Battery above) to design
a new battery structure.

▪ Load Battery File allows to select a battery structure saved as *.gdt file

▪ Use Current Structure allows to use the battery structure already in memory.

16 GeoDict 2020 User Guide

 BatteryDict computations

A NODE , C ATHODE , AND E LECTROLYTE /S EPARATOR

Next, the material properties of the active materials, as well as of binder and
electrolyte may be changed under the Anode, Cathode and Electrolyte/Separator
tabs.

GeoDict 2020 User Guide 17

BatteryDict simulations of charging Li ion batteries

For both active materials, the simulation parameters Maximum Lithium

Concentration, Electronic Conductivity, Ionic Diffusion Constant, Butler-
Volmer Rate Constant, and Open-Circuit Potential Function can be chosen (see
Simulation Parameters above).
Values defining the Open-Circuit Potential Function can be modified directly in the
table. Additionally, the number of value points can be changed by deleting or inserting
new rows. Another possibility to input the Open-Circuit Potential Function is to
import it from a text file with two columns. The first column contains the state of
charge in percent, and the second column shows the related potential.
The chosen function is displayed as Potential over State of Charge in the plot on
the right of the table.
The Electronic Conductivity is defined under the Binder & Carbon Black subtab.

The following electrolyte parameters are defined under the Electrolyte/Separator

tab:

18 GeoDict 2020 User Guide

 BatteryDict computations

Equilibrium Lithium Concentration, Ionic Conductivity, Ionic Diffusion

Constant, and Transference Number (see Simulation Parameters above).
Additionally, after checking Modify Effective Separator Properties, Effective
Ionic Conductivity and Effective Ionic Diffusion Constant which are properties
of the separator can be defined. This accounts for the possibility that the electrolyte
properties can be different inside the separator compared to the electrodes due to the
microstructure of the separator.
If Modify Effective Separator Properties is left unchecked, the electrolyte in the
separator is assumed to have the same properties than the electrolyte in the
electrodes.

S OLVER O PTIONS
The options for the charging simulation are defined under the Solver Options tab.
From the pull-down menu Charge or Discharge Battery?, select to simulate the
charging or discharging of the battery. Here, a charging simulation is chosen.

The Applied Current can be specified by Charge Rate, Charge Density or by defining
the Current itself in Ampère (A).

GeoDict 2020 User Guide 19

BatteryDict simulations of charging Li ion batteries

With an entered Charge Rate, the applied current is related to the cells capacity and
is a measure of the duration of the battery charging. E.g. a charge rate of 1 for a
battery with capacity of 1Ah means that the battery is loaded completely in 1 hour
with a constant current of 1 Ampère (A). This time is unrelated to the simulation time
but instead to the time that the charging process itself requires. A charge rate of 2
means that charging the same battery with a current of 2 Ampère takes ½ hour.
Defining the Current or the Current Density instead means that the applied current
(in A) or the current density (in A/cm²) is defined directly.
Since BatteryDict2020, the state of charge (SOC, see State of Charge above) for the
start and the end of the simulation can be defined in two ways:

▪ as percentage of the SOC of the whole cell, or

▪ for the anode SOC and the cathode SOC, individually.

In the Output Options panel:

▪ Calculate Estimation Only can be checked, and then, the full simulation is not
performed, but only the cell potential is estimated. This estimation is much faster
than the full simulation. It considers the cell potential in equilibrium as well as the
charge rate and part of the structure information, like the solid volume
percentages of the different materials of the battery. From this information Ohm
losses and overpotentials at the surfaces are considered in the estimation.
Overpotentials due to diffusion are not considered in the estimation so far.

▪ The Cell-SOC Interval of Result Output defines the interval for the export of
intermediate result files in percent of the cell state of charge.

▪ Checking Final Simulated Time, the total time that the battery will be charged
can be defined. The simulation ends if the charging process exceeds the Final
Simulated Time even if the final state of charge for the simulation is not reached.
Note that the Final Simulated Time is not the time that the solver is given for
its computation, but it is the time, the simulated battery is given for charging or
discharging.
Example: Let us consider a battery that is charged from 20% state-of-charge to
70% state-of-charge at a charge rate of 1C. The battery takes 30 min to charge.
If the Final Simulated Time is set to 600 seconds (10 min), then in this example,
the battery will only be charged from 20% to 36.67% state-of-charge.
The boundary condition chosen in the Boundary Conditions in Tangential
Directions panel is linked to the Battery Input Mode (if Design Battery is chosen)
under the Geometry tab. Thus, changes in the boundary conditions under Solver
Options, changes them in the Geometry tab. However, if Load Battery File is chosen
for the simulation, the boundary conditions can be chosen independent of the
previously designed battery. The value for the boundary conditions is not stored in
the gdt file containing the battery structure. The connected and unconnected parts of
the active materials and of the electrolyte are dependent on the boundary conditions,
and are, therefore, identified when the simulation is started.
Finally, in the Additional Options panel the following options for the simulation may
be selected:

▪ Temperature defines the temperature during simulation in Kelvin (˚K).

The material properties of active materials and electrolyte can strongly depend
on the temperature. Because this temperature-dependence of the material
parameters is not considered in the equations solved during charging or

20 GeoDict 2020 User Guide

 BatteryDict computations

discharging, the temperature defined here needs to be consistent with the

material properties defined under the Anode, Cathode, and Electrolyte tabs.

▪ Maximum Simulated Time Steps defines the maximum time step in seconds
used in the simulation. For each time step in the simulation, the solver tries to
find an equilibrium solution fulfilling the given system of differential equations. If
the optimization does not find a valid solution fitting to the state of charge, this
time step is reduced, and the solver starts again. This procedure is repeated until
an appropriate time step is found. This time step is used for the next steps in the
simulation as well. After several successful steps with a certain time step, the
solver tries to increase this time step again, to reduce overall computation time.
The value defined in Maximum Simulated Time Steps gives an upper value for
this time step. If the size of the time step is oscillating a lot, the recommendation
is to reduce this maximum value.
If your simulation does not produce enough points in the plots of the result file,
then you might want to reduce the Maximum Simulated Time Step. This is
especially the case if you choose to simulate at a charge rate larger than 10C.

▪ Choose the number of processes used for the computation in Parallelization.

E QUATIONS
The differential equations solved in the charging simulation are listed under the
Equations tab and are described above in Theoretical Basis.

GeoDict 2020 User Guide 21

BatteryDict simulations of charging Li ion batteries

C HARGE B ATTERY (BEST MICRO FFT)

To simulate the charging or discharging of a battery with BESTmicroFFT, choose
Charge Battery (BESTmicroFFT) in the pull-down menu of the BatteryDict section
and click Edit to choose the solver options.
The BatteryDict (BESTmicroFFT) dialog opens, containing six tabs that are almost
identical to those for Charging Battery (BESTmicro).

The Electrolyte tab differs from the Electrolyte/Separator tab of the simulation
with BESTmicro. Here, the Electrolyte tab contains only the panel for the electrolyte.
For a simulation with BESTmicroFFT, the electrolyte in the separator always has the
same properties as the electrolyte in the electrodes.

Also, the Solver Options tab differs a bit from the one of a BESTmicro simulation.

22 GeoDict 2020 User Guide

 BatteryDict computations

On the panel Charging Options, it can be selected whether the battery is charged
or discharged. In contrast to a BESTmicro simulation the applied current is always
defined by the charge rate and the state of charge for start and end of the simulation
is always defined as a range of the cell state of charge.
On the panel Output Options, is possible to choose to Calculate the Equilibrium
Charge Curve Only, instead of performing the whole simulation. It is not possible to
estimate the charge curve out of equilibrium like in a BESTmicro simulation.

GeoDict 2020 User Guide 23

BatteryDict simulations of charging Li ion batteries

B ATTERY D ICT C OMPUTATION R ESULTS

R ESULTS OF D ESIGN B ATTERY

After running Design Battery successfully, a GeoDict result file (.gdr) is written and
the Result Viewer of this result file automatically opens. The created battery structure
is available in memory and shown in the Visualization area of the GeoDict GUI.

The .gdr file can be reopened at any time by choosing File → Open Results
(*.gdr)… in the menu bar. The Result Viewer opens at the Results tab.

In the Result Viewer, the list of all open result files is shown in a box in the upper
part. The name of the result file currently displayed is highlighted in blue. This box in
the upper part of the Result Viewer can be collapsed (and expanded) by pulling up
(or down) the dotted line under the box.

The green dot in the upper right part of the result viewer, indicates that the structure
from which these results were obtained is loaded in the Visualization area. This is
always the case directly after creating the battery structure. If a result file (*.gdr) is
reloaded later through File → Open Results (*.gdr)…, the dot is red. The structure
can be loaded into GeoDict memory and the visualization area by clicking the Load
Structure button in the result viewer.
At the top of the result viewer, clicking the button Excel (single table) exports the
information of this result file to Microsoft Excel®. A single worksheet is created that
contains the information under the Results tab from the result file.
With the Combine Results button, the results of several selected .gdr files can be
shown side by side in the Result Viewer.

24 GeoDict 2020 User Guide

 BatteryDict computation results

Collapse and expand

The Input Map tab contains all the input parameters of the battery creation.
The Log Map tab shows the runtime and the information about the system in which
the battery was created.

The Post Map tab shows parameter and graph data used for the plots on the Results
tab.

GeoDict 2020 User Guide 25

BatteryDict simulations of charging Li ion batteries

The Results - Report subtab lists the information on the volume fractions of the
different materials for the whole battery, as well as for anode and cathode separately.
It is differentiated between connected and unconnected parts of the material. The
unconnected parts do not contribute to the performance of the battery, a small part
of unconnected material is therefore desired.

Under the Results - Plots subtab, several geometric properties of the designed
battery are available as standard plots. Porosity, unconnected active material, and
unconnected electrolyte are shown as averaged values over each slice in y- and z-
direction.
At the separator and both current collectors, no values are shown for porosity, since
these structures are not resolved into a microstructure.
For the separator no value for unconnected active material is shown, since it does not
contain active material at all.

26 GeoDict 2020 User Guide

 BatteryDict computation results

In the same manner, no value for unconnected electrolyte is shown for the anode and
cathode current collector, since they do not contain electrolyte.

GeoDict 2020 User Guide 27

BatteryDict simulations of charging Li ion batteries

The Results - Map subtab displays information on the volume fractions of the
different materials of the whole battery, as well as for anode and cathode alone. This
are the values also shown under the Results – Report subtab.

28 GeoDict 2020 User Guide

 BatteryDict computation results

The Metadata tab displays a description of the .gdr file. The text can be modified
here, and the modification can be saved to the .gdr file.

As usual for all GeoDict result files, several buttons for further processing are available
at the bottom of the Result Viewer.

▪ Clean-Up/Pack: Zip the information contained in the result folder or clean-up the
result folder.
▪ Load Input Map: Reload solver options and material parameters used for the
computation of the results. After loading, these parameters are available in the
BatteryDict dialog box.
▪ Export: Export the results for postprocessing with another tool

▪ Excel (generic): Export the information of the section ResultMap of the

GeoDict result file to Microsoft Excel®. Optionally, the information contained
in the InputMap can be also exported. You can analyze computation results in
Microsoft Excel® using GeoDexcel provided with GeoDict. See the GeoDexcel
handbook for more information.

▪ Matlab: Open Matlab® if an installation and license is available, change to the

project folder, and load the result file with GeoLab. See the GeoLab handbook
for more information.

▪ Store As Html: Export the information shown in the Results tab to a html
file.

▪ Save Plots: Export all plots to the GeoDict result folder.

GeoDict 2020 User Guide 29

BatteryDict simulations of charging Li ion batteries

R ESULTS OF C HARGE B ATTERY (BEST MICRO FFT AND BEST MICRO )

After a charging simulation has finished, a GeoDict result file (.gdr) is written and the
Result Viewer of this result file automatically opens. Information, available in this file
is shown and can be visualized or further processed.
These results are accessible at any time by loading the *.gdr file from the menu bar
(File → Open Results (*.gdr) …). The Load Structure button and the buttons
available at the bottom of the Result Viewer are the same as for other result files
and are explained above.
In the following, results of the simulation with BESTmicro are shown.
The Input Map tab displays the input data for the simulation. All parameters can be
displayed here, e.g. the material parameters defined for the first active material of
the anode.

On the tab Log Map, the runtime of the simulation as well as information about the
machine used, are shown.

30 GeoDict 2020 User Guide

 BatteryDict computation results

The Post Map tab lists information about plots shown in the Results tab. Parameters
chosen for the plots as well as data for the graphs are shown here.

The Results tab shows the results of the computation in several forms.

GeoDict 2020 User Guide 31

BatteryDict simulations of charging Li ion batteries

The Results - Map subtab lists all values available in the Report tab of the GeoDict
result file.
In the Results - Report subtab, part of the information of the results map is
displayed, with explanations about the computation results. If there were warnings
during the simulation, then they appear here in the report (see Tested Parameter
Space above). The information displayed in the Report subtab can be exported to a
html file by clicking Export → Store As Html at the bottom of the Result Viewer.

Under the Results - Plots subtab, several standard plots for the visualization of the
BatteryDict simulation results are available: Cell Potential, Cell State of Charge,
and Transferred Charge plotted over Time. Additionally, the Cell Potential and
the Electrode State of Charge can be plotted over the Cell State of Charge.

Clicking the right

mouse button, allows
to change properties
of each plot shown,
or to save the image.

32 GeoDict 2020 User Guide

 BatteryDict computation results

In the Cell Potential over Cell State of Charge plot (seen above), the cell potential
while charging or discharging is compared to the cell potential in equilibrium, i.e. while
loading in infinite time. While charging or discharging the battery infinitely slow, the
battery is always in equilibrium state, leading to optimum values for the cell potential.
In contrast to this, charging or discharging the battery in finite time, the battery is
not in equilibrium state anymore. Charging curves have therefore always higher
values than the equilibrium curve, i.e. charging the battery requires a higher potential
than the optimum one. Discharging curves are always below the equilibrium curve.
Additionally, for the results of Charge Battery (BESTmicro), the estimation of the
cell potential is shown in this plot. Please note that BESTmicroFFT does not provide
an estimation of the cell potential. If Calculate Estimation Only was chosen as option,
then only the estimation of the cell potential and the equilibrium cell potential is
shown.

Due to simulation at constant current (not constant voltage), the state of charge
increases linearly with the simulated time. The trend of the cell potential over time is
therefore the same as for the Cell Potential over Cell State of Charge and the Cell
State of Charge over Time is a straight line.

GeoDict 2020 User Guide 33

BatteryDict simulations of charging Li ion batteries

Due to the simulation of charging with constant current (not constant voltage), the
curve Transferred Charge over Time is also a straight line.

Also, the state of charge of anode and cathode changes linearly over the cell state of
charge. The cathode of the example shown here has a larger capacity compared to
the anode. The cathode state of charge is therefore not zero, if the anodes state of
charge has reached 100%.

The Ion Concentration tab shows the mean concentration of lithium ions for each
slice in y-z-direction for a constant x value. It can be displayed as Ion Concentration
of Electrolyte, Ion Concentration of Active Material (for both active materials
together) or of each active material separately (Ion Concentration of Active
Material 1, Ion Concentration of Active Material 2).

34 GeoDict 2020 User Guide

 BatteryDict computation results

A separate curve is shown for each state of charge, with an intermediate result
available. As can be seen in the following plot, the concentration of lithium ions in the
cathode is decreasing during charging of the battery, while the concentration in the
anode is increasing.

The ion concentration of the cathode changes more uniformly if only the ion
concentration of the first active materials is selected to be displayed.
Note, that for the X-Position between 48 and 50 µm no Active Material 1 is contained
in the cathode. Therefore, a jump is visible there. For the anode, only one active
material was used in the battery cell. Therefore, the ion concentration shown for
Active Material 2 is zero.

GeoDict 2020 User Guide 35

BatteryDict simulations of charging Li ion batteries

In the beginning of the simulation, the Li-ion concentration of the electrochemically

connected part of the electrolyte is constant throughout the battery. Its value is the
equilibrium Lithium concentration specified in the settings.
During charging, the cathode releases Li-ions into the electrolyte. Therefore, the Li-
ion concentration grows in the electrolyte on the cathode side. At the same time, the
Li-ions intercalate into the anode.
Since, the diffusion of the Li-ions takes some time, the gradient of the Li-ion
concentration in the electrolyte is visible in the example

36 GeoDict 2020 User Guide

 BatteryDict computation results

The Potential tab shows the mean potential for each slice in y- and z-direction for
the solid part of the battery, as well as for the electrolyte. Again, as a default, for
each state of charge with an intermediate result available, a separate curve is shown
in the plot. Click the right mouse button in the plot to change view settings.
In the example, for every given time step the potential in the solid looks constant
within one electrode. This indicates that the conductivity is high enough. If the
conductivity would be much smaller, the potential would exhibit a drop within one
electrode.
The solvers calculate the battery charging by applying a constant electric current. The
necessary potential difference for keeping up the current is determined from the
electric current in the solids, the ionic current in the electrolyte, and the overpotentials
on the boundary of electrolyte and active material.

The current is only dependent on the potential difference and not on the total potential
offset. The latter can be chosen freely. The solvers choose the potential to be zero at
the anode current collector. Therefore, the potential in the anode always starts at
zero for all times. As explained before, the potential is as good as constant in the
anode. Hence, in this example, the potential in the anode solid is very close to zero
everywhere for all times.
Over the time, the intercalation of Li-ions into the anode becomes more costly. At the
same time, the potential gain from the deintercalation of Li-ion out of the cathode
decreases. Therefore, the potential difference between anode and cathode grows,
which can be seen in the example.
For the potential of the electrolyte, a gradient is visible between the anode current
collector and the cathode current collector. This gradient is increasing for higher
charge rates. Disregarding the noise, the potential grows from left to right. This comes
from the charge current that is driven by the ions resolved in the electrolyte. The

GeoDict 2020 User Guide 37

BatteryDict simulations of charging Li ion batteries

charge current leads to a growing potential in flow direction because the ionic
conductivity is not zero. The lower the ionic conductivity, the stronger the growth of
the electrolyte potential in flow direction. However, since the conductivity is
sufficiently large here, we can see that the voltage drop is only about 0.1 Volt. This
is negligible compared to the 4 Volt drop caused by the overpotentials at the boundary
of electrolyte and active material (see plot “Potential of Solids”). The potential
gradient is larger on the right, because there, the electrolyte is not as good connected
as on the left side.

The noise visible in the electrolyte potential stems from electrolyte voxels which are
not connected to the separator via other electrolyte voxels. In these ‘unconnected’
electrolyte voxels, the potential is rather extreme (either quite high or quite low).
The potential of the electrolyte voxels is shown in the figure above. On the right side
(cathode), unconnected voxels are red (high electrolyte potential) or blue (low
electrolyte potential).
The blue voxels are not connected to active material. This connection would be the
only way for electrolyte to change the potential (via the Butler-Volmer interface
current).
The red voxels are connected to active material and roughly have the potential of the
cathode solids minus the overpotential from the interface of active material and
electrolyte. They have a higher potential than the electrolyte connected to the
separator, because they do not have currents that would lead to a potential drop.
On the anode side, all voxels with electrolyte have a low potential even in the case of
them being connected to active material only, because they either have kept their
potential from the beginning (0 Volt for BESTmicro) or have the potential of the anode
solids plus the overpotential on the interface of active material and electrolyte.
Unconnected electrolyte voxels have a lower potential than the electrolyte connected
to the separator, because they do not have current that would cause a potential
difference.

38 GeoDict 2020 User Guide

 BatteryDict computation results

The Battery Geometry subtab can show the mean of porosity, unconnected active
material or unconnected electrolyte for each slice in y- and z-direction.

Solution files can be accessed via the Data Visualization tab. The number of
available solution files depends on the length of the interval selected for output of
intermediate results in Solver Options (see above).

For instance, let us select the solution

file at a state of charge of 40%. For the
example shown here, we just need to select
step 3. By clicking Load Results, the
available volume fields (Select Volume Fields)
of the file can be selected. Additionally, the
user may have a look at the file header (View
File Header).

GeoDict 2020 User Guide 39

BatteryDict simulations of charging Li ion batteries

Let’s say that we would like to load only the fields “Concentration” and “Potential”. In
this case we first click Uncheck all. Then, we select the choices
“Concentration:Concentration” and “Potential:Potential”. After clicking OK, both
selected fields are loaded into the GeoDict GUI.
The concentration field of lithium ions for the example computed above is visualized
as shown here.
More information on visualization of results is available in the Visualization handbook.

Check Allow clipping by distance to create a

distance map with respect to the solid material.
Its value is zero on the surface of the solid
material, inside the material the values are
negative and outside positive. This allows,
clipping any result field by its distance to the
surface of the solid material.
For example, if you are only interested in the
Li-ion concentration inside of the electrolyte,
check Allow clipping by distance when you
load the volume field. In this way, in the
Visualization, you can choose to clip the ionic
concentration by positive values of the created
distance map.

Check Keep existing Volume Fields if you

repeat loading result fields for another time
step or have other result fields in memory and
want to keep previously loaded volume fields.
If this box is not checked, previously loaded
result fields are discarded.
Note that for the solver BESTmicroFFT, only the
fields “Concentration” and “Potential” are
available.

40 GeoDict 2020 User Guide

 BatteryDict computation results

Through the Create Videos tab, a video showing the ion concentration while charging
or discharging of the battery can be created. The first part of the video shows the
cropping of the battery structure together with the ion concentration at the start of
the simulation.

A screen shot of such a video is shown here. The left side of the video shows a graph
of the cell potential over state of charge. The actual state of charge value is displayed
there. In the middle, the ion concentration in the battery structure is shown. On the
right, the actual state of charge of the battery is schematically shown.

GeoDict 2020 User Guide 41

BatteryDict simulations of charging Li ion batteries

Click Open Macro File, under the Create Videos tab, to open the GeoDict Python
macro used to create the video, if you are interested in modifying it with a text editor.
Click Edit Parameters… to modify the default settings of the video creation.

Check Keep Images to save not only the

video, but also the created images for the
ion concentration in a separate folder.
The number defined in Crop Frames
defines the time for the cropping of the
structure at the start of the video.
Crossfade Frames defines the time for the
visualization of the ion field. The shorter the
time defined here, the faster the charging
is shown in the video.
Waiting Frames: A higher number here
means longer pause before and after
charging.

Check Use Predefined View Settings to use the predefined orientation of the
structure for the video creation as shown in the example above. Uncheck this box to
use the orientation of the structure as shown in the GeoDict GUI.
Check Add Potential Plot to keep the graph with the potential curve at the left of
the video.
Check Show Color Bar to add an additional color bar for the ion concentration below
the structure.
The font size for the potential curve graph can be set with Fontsize for Plots.

42 GeoDict 2020 User Guide

 BatteryDict computation results

R ESULT F OLDERS

After designing a battery or running a charging simulation with BESTmicro or

BESTmicroFFT, result files and the corresponding result folder are found in the project
folder.

B ATTERY D ESIGNER : THE DESIGN BATTERY FOLDER

After the generation of a battery structure, a GeoDict result file (.gdr) is created in
the selected project folder. A subfolder with the same name as the result file contains
the generated battery structure.

The result file (.gdr) can be loaded at any time by selecting File → Open Results
(*.gdr) … in the menu bar.

BEST MICRO FFT AND BEST MICRO : THE CHARGE BATTERY FOLDERS
After running a charging simulation, a GeoDict result file (.gdr) is created in the
selected project folder. A subfolder with the same name as the result file contains two
subfolders called SolverConfig and SolverOutput.

GeoDict 2020 User Guide 43

BatteryDict simulations of charging Li ion batteries

SolverConfig contains files with the input parameters for the charging simulation.
The SolverOutput folder contains some log-files from the solver, as well as a
subfolder output with *.cap files containing volume fields. Such a *.cap file is written
for the initial state, for the final state, and for each time step with intermediate output.
These volume fields can be loaded from the Load Results button under the Result
Visualization tab of the Result Viewer or directly from the GeoDict GUI by selecting
File → Load Volume Field … in the menu bar.
The BESTmicro result folder contains some additional files as well as the two additional
folders configuration and geometry in SolverOutput. They contain some internal
solver information and cannot be loaded in the GeoDict GUI.

44 GeoDict 2020 User Guide

Math2Market GmbH
Richard-Wagner-Str. 1
67655 Kaiserslautern

Web: www.geodict.com

© Fraunhofer Institut Techno- und Wirtschaftsmathematik ITWM, 2003-2011.

© Math2Market GmbH, 2011-2020. All rights reserved.