Study of ionic and electronic transport

properties influencing the performance of

Lithium-ion batteries

Internship Report/Master Thesis

by

Julius Akinribido
Surface, Electro-, Radiation, and Photo-Chemistry (SERP+)

§ January – July 2022.

Supervised by: Olivier Dubrunfaut

Loïc Assaud
Jean-Claude Badot
Table of Contents

Acknowledgments ......................................................................................................................... 3
Abstract.......................................................................................................................................... 4
Abbreviations ................................................................................................................................ 5
1 Introduction ........................................................................................................................... 6
1.1 Background ................................................................................................................. 6
1.2 Lithium-ion Batteries .................................................................................................. 6
1.3 Challenges of Lithium-ion Batteries ........................................................................... 8
1.4 Electronic and Ionic Limitations in Lithium-ion Batteries ......................................... 9
1.5 Broadband Dielectric Spectroscopy (BDS) .............................................................. 10
1.6 Objectives of the study.............................................................................................. 12
2 Materials and Methods ....................................................................................................... 13
2.1 Materials ................................................................................................................... 13
2.2 Material Characterizations ........................................................................................ 13
2.2.1 XRD Analysis ................................................................................................... 13
2.2.2 SEM – EDS Analysis ........................................................................................ 13
2.3 Sample Preparation ................................................................................................... 13
2.4 BDS Measurement .................................................................................................... 14
3 Results and Discussion ........................................................................................................ 15
3.1 Physico-Chemical Characterizations ........................................................................ 15
3.1.1 X-ray Diffraction (XRD) .................................................................................. 15
3.1.2 Scanning Electron Microscopy (SEM) ............................................................. 15
3.1.3 Energy Dispersive Spectroscopy (EDS) ........................................................... 17
3.2 Broadband Dielectric Spectroscopy (BDS) .............................................................. 17
3.2.1 Ex-situ Measurements (Dry Electrode) ............................................................ 17
3.2.2 Wet Electrode Measurements ........................................................................... 22
4 Conclusions and Perspectives ............................................................................................ 26
References .................................................................................................................................... 27

2
Acknowledgments

I am grateful for the opportunity to join the Erasmus Mundus SERP+ Program. The experiences
of the two-year master have set a landmark for my academic and professional careers.
My heartfelt thanks go to my thesis supervisor and co-supervisors, Dr. Olivier Dubrunfaut, Dr.
Loïc Assaud, and Dr. Jean-Claude Badot, for their continuous guidance and support during the
internship, and Prof. Sylvain Franger for his insightful comments.
I thank the Institut de l’Énergie Soutenable (IES Paris-Saclay) for the financial support of my
project, as well as collaborating institutes, Institut de Chimie Moléculaire et des Matériaux d'Orsay
(ICMMO) and Laboratoire de Génie électrique et électronique de Paris (GeePs), for the wonderful
environment created to attempt high-quality research. Also, thanks to Serge Albacha, Anshuman
Agrawal, Benjamin Rondeau, Yutzil Segura, Florian Guenard, Hadeer Khalifa, and other ERIEE
team members, for their company and scientific discussions.
Special thanks to the director of the SERP+ program, Prof. Sandrine Lacombe, and the local
coordinator from the University of Porto, Prof. Eduardo Marques, and also to the excellent staff
team that efficiently delivered the master's program. I also appreciate the program manager, Eva
Renouf, and secretary, Kali Sidibé, for providing prompt information and guidance throughout the
master's program.
My appreciation goes to my home country professors, Prof. Folahan Adekola and Prof. Alafara
Baba, for their incredible support and for providing the necessary foundation and inspiration to
help their students achieve academic and research excellence.
Many thanks to my SERP+ colleagues for their encouragement and support, and to my friends,
Ademola Adeniji, Raj Shekhar, Nitik Bhatia, Merin Kennedy, and Kiran Hidayat, for the great
moments we shared. To my mother, Feyisayo, and siblings – Ademola, Victoria, and Itunu, for
always supporting me.

3
Abstract
A major factor in optimizing lithium-ion batteries for electric vehicle applications and
decarbonizing road transport is the ionic and electronic transport in the composite positive
electrode. This study presents the use of broadband dielectric spectroscopy (BDS) with frequencies
ranging from a few Hz to several GHz to provide a multiscale description of the electronic and
ionic properties of two EV-grade battery active materials: LiNi0.5Mn0.3Co0.2O2 (NMC532) and
LiNi0.8Mn0.1Co0.1O2 (NMC811), by measuring the permittivity and conductivity as well as the
different dielectric relaxations resulting for interfacial space-charges and ionic motions. Dry
measurements were obtained using an electrode mix with varying amounts of active material,
binder (PVdF-HFP, poly-vinylidene fluoride-co-hexafluoropropylene, and additive conductive
agent (carbon black). 1M LiPF6 solution in an ethylene carbonate/diethyl carbonate mixture
(EC/DEC 1:1 v/v) was used as electrolyte for wet electrode measurements and ionic observations.
The obtained results showed higher electronic conductivity in NMC811, lower electronic
percolation threshold of conductive additive, and lower ionic and electronic activation barrier,
projecting a better rate performance and higher energy density for the nickel-rich electrode.

Keywords: Lithium-ion battery, permittivity, dielectric, electronic conductivity, ionic

conductivity.

Résumé
L’un des défis majeurs dans l'optimisation des batteries lithium-ion pour application aux véhicules
électriques et la décarbonation du transport routier est une meilleure compréhension des
phénomènes conduisant à une limitation du transport ionique et électronique au sein de l'électrode
positive. Cette étude présente la mise en œuvre de la spectroscopie diélectrique large bande (BDS)
utilisant des fréquences allant de quelques Hz à plusieurs GHz pour fournir une description multi-
échelle des propriétés électroniques et ioniques de deux matériaux actifs de batteries utilisées dans
les véhicules électriques : LiNi0.5Mn0.3Co0.2O2 (NMC532) et LiNi0.8Mn0.1Co0.1O2 (NMC811), en
correspondantes, résultant des charges d'espace interfaciales et des mouvements ioniques. Les
mesures à sec ont été obtenues en utilisant différentes électrodes présentant des quantités variables
de matière active, de liant (PVdF, poly-vinylidene fluoride-co-hexafluoropropylene), et d'additif
conducteur (noir de carbone). Une solution de LiPF6 1M dans un mélange de carbonate
d'éthylène/carbonate de diéthyle (EC/DEC 1 :1 v/v) a été utilisée comme électrolyte pour les
mesures d'électrodes imprégnées, permettant de caractériser les mouvements ioniques. Les
résultats obtenus ont montré une conductivité électronique plus élevée dans le cas du NMC811,
un seuil de percolation électronique plus bas du noir de carbone, et une barrière d'activation ionique
et électronique plus basse, conduisant à de meilleures performances et une densité d'énergie plus
élevée pour l'électrode riche en nickel.

Mots-clés : Batterie lithium-ion, permittivité, diélectrique, conductivité électronique, conductivité

ionique.

4
Abbreviations

AM – active material
BDS – broadband dielectric spectroscopy
CB – carbon black
EDS – energy-dispersive x-ray spectroscopy
EIS – electrochemical impedance spectroscopy
EV – electric vehicle
e- – electron
LCO – LiCoO2
LFP – LiFePO4
LIB – lithium-ion battery
LP40 – 1M LiPF6 solution in an ethylene carbonate/diethyl
carbonate mixture (EC/DEC 1:1 v/v)
m – Archie’s exponent
NMC – LiNixMnyCo1-x-yO2
NMC532 – LiNi0.5Mn0.3Co0.2O2
NMC811 – LiNi0.8Mn0.1Co0.1O2
PVdF-HFP – poly-vinylidene fluoride-co-hexafluoropropylene
SEM – scanning electron microscopy
TM – transition metal
XRD – x-ray diffraction
ε – permittivity
Δε – dielectric strength
ρn – electronic resistivity
σe – electronic conductivity
σi – ionic conductivity
𝝓 – porosity fraction

5
1 Introduction
1.1 Background
In an electrochemical cell, energy can be stored as chemical energy and converted into electrical
energy to power various devices and equipment when needed1. A battery comprises several
electrochemical cells connected in series and/or parallel to give the required voltage and capacity2.
In most applications, the ability to recharge the cells (secondary cells) is preferred to the non-
rechargeable cells (primary cells). Hence, the secondary cell design requires energy-delivering
chemical reactions, which can be reversed by supplying electrical current in the opposite
direction1.

The transportation sector accounts for about 16.2% of global greenhouse gas emissions, and
decarbonizing road transport is projected to cut emissions by about 12%3. This can be achieved by
developing and optimizing advanced energy storage devices to eliminate tailpipe emissions and
overcome the intermittency challenges of wind and solar sources4. When compared to other battery
systems, lithium-ion batteries (LIBs) are currently the most versatile and efficient option for eco-
friendly electric vehicle applications4,5. This is due to their performance characteristics, which
include higher energy and power density, long cycle life, rapid charge capabilities, and low self-
discharge rate6. As a result, LIBs have dominated consumer electronics and mobile device
applications6,7.

1.2 Lithium-ion Batteries

Lithium-ion batteries depend on the ability of layered and structured transition metal compounds
to undergo reversible intercalation of lithium ions in an electron-transfer process8. A lithium-ion
cell consists of a positive electrode, a negative electrode, electrolyte, separator, and current
collectors (Figure 1). During charging, the electrolyte facilitates the transport of lithium ions from
the positive electrode to the negative electrode, where they are stored by intercalation, and
electrons leave the positive electrode to the negative one through the current collectors and external
circuit. The reverse mechanism occurs during discharge, the positive electrode is re-lithiated, and
electrons flow to it, powering an external device. The separator allows only the passage of lithium
ions, thereby preventing a short-circuit6. The electrochemical reactions are shown in
Table 1.

Figure 1. Schematic of electrochemical reactions in a lithium-ion cell.9

6
 Table 1. Electrochemical reactions in lithium-ion batteries*.1
𝐶ℎ𝑎𝑟𝑔𝑒
Positive electrode: LiCoO2 ⇌ Li1−𝑥 CoO2 + 𝑥Li+ + 𝑥𝑒 −
𝐷𝑖𝑠𝑐ℎ𝑎𝑟𝑔𝑒
𝐶ℎ𝑎𝑟𝑔𝑒
Negative electrode: C + 𝑥Li+ + 𝑥𝑒 − ⇌ Li𝑥 C
𝐷𝑖𝑠𝑐ℎ𝑎𝑟𝑔𝑒
𝐶ℎ𝑎𝑟𝑔𝑒
Overall reaction: LiCoO2 + C ⇌ Li1−𝑥 CoO2 + Li𝑥 C
𝐷𝑖𝑠𝑐ℎ𝑎𝑟𝑔𝑒
* for a Co-based positive and C-based negative electrodes.

The positive electrode in a commercial lithium-ion cell consists of aggregates of

electrochemically-active material, a conductive additive (mostly carbon black), and a polymeric
binder to ensure composite durability, homogeneously mixed and coated onto a current collector
(Figure 2). The negative electrode used is mostly highly-intercalating graphite; hence, the positive
electrode material and composition variation are more relevant for power performance and cost
optimizations of the battery.

The composite nature of the positive electrode gives rise to several interfaces, including a
macroscopic interface between the electrode material and the current collector; mesostructural
interfaces (micrometric) between clusters (aggregates of particles) of active materials; and
nanometric level interfaces between active material particles and carbon black additive10. The
liquid electrolyte occupies the defined porosity, creating additional solid-electrolyte interfaces.
These interfaces create a compounding limitation to charge depletion and influence the conduction
of lithium ions and electrons through the electrode. These interfacial processes and overall cell
performance can be influenced by changing parameters such as the nature and amount of active
and non-active materials, the size distribution, porosity, electrode thickness, nature of the
electrolyte, etc.

Figure 2. Composite architecture of the positive electrode.10

Lithium metal oxides or lithium metal phosphates are used as positive electrode active materials
in commercial lithium-ion batteries6. LiCoO2 (LCO) has a capacity of 155mAh/g, a nominal
voltage of 3.9 V, and good electrochemical performance. However, its high-cost cobalt demands
call for other active material configurations for applications in electric vehicles 6. The high
phosphate stability in LiFePO4 (LFP) batteries ensures better safety and a reduced risk of thermal
runaway. LFP batteries are also less expensive, with a specific capacity of 160 mAh/g and a voltage
of 3.40 V6. The most commercialized electrode types are LiNixMnyCo1-x-yO2 (NMC) with varying

7
amounts of transition metals nickel, manganese, and cobalt and are capable of initial capacity
greater than 200 mAh/g6. NMCs are made by partly substituting Co in LiCoO2 with Ni and Mn to
achieve improved electrochemical performance while reducing the material cost. High discharge
capacity is demonstrated in Ni-rich compositions; Mn-rich compositions maintain better cycle life
and thermal safety; Co-rich compositions provide excellent rate capability, allowing composition
modifications to achieve the desired behavior5. The multivalence states Ni2+/Ni3+ and Co2+/Co3+,
and their occupation fractions determine the performance properties of the NMC composition.
Figure 3 summarizes the effect of these composition variations on thermal stability, discharge
capacity, and capacity retention11.

There is increasing interest in Ni-rich electrodes for EV applications due to increased specific
capacity, electronic conductivity, and low cost12. With increasing nickel content, Wang et al.
(2018) reported a broadening of the Ni3+/4+ eg* band, lowering the Fermi level, and decreasing
electronic activation energy13. However, due to Ni mixing with Li sites, nickel-rich cathodes suffer
structural degradation during cycles5,12.

Figure 3. Relationship between thermal stability, discharge capacity, and capacity retention of different NMC
compositions.11

1.3 Challenges of Lithium-ion Batteries

The major challenges in developing lithium-based rechargeable batteries for electric vehicles
include improving safety, increasing energy density by increasing the capacity of reversible
lithiation, improving the power density, allowing fast charge/discharge cycles while optimizing
for low-cost, environmentally friendly materials, and good chemical compatibility to reduce
battery ageing and promote long service life14. Unlike the thermodynamic properties, like specific
capacity and energy density, the kinetic behavior of NMCs, specifically the power/rate capability,
which is primarily determined by the electronic and ionic transport processes, is still poorly
understood and crucial for designing high-performance lithium-ion batteries13.

The target metrics for efficient lithium-ion battery technology in EV applications (Figure 4) reveal
the areas where improvements are needed and the minimum goals to be attained15. NMC-based
lithium-ion batteries currently cover the largest area in the metric chart in terms of specific energy
and power, and recent developments have reduced irreversible capacity loss and enhanced the rate
capabilities6.

8
 Figure 4. Important target metrics for lithium-ion batteries in electric vehicles.15

The cost of lithium-ion batteries has dropped significantly in recent years, with a market average
of $137 per kWh, and is expected to go below $100 per kWh by 202316. According to Baker
McKenzie's 2022 report, the global manufacturing capacity of lithium-ion batteries has expanded
from 14 GWh in 2010 to 457 GWh in 2020, and it is predicted to exceed 2,000 GWh in 2028 due
to increasing legislation favoring the production of battery-powered electric vehicles16.

1.4 Electronic and Ionic Limitations in Lithium-ion Batteries

To close the gap between the theoretical and actual rate performance capabilities of lithium-ion
batteries, improvements in the materials, electronic and ionic conductivities are required17.
However, there is limited literature on the interplay between electrode composition, transport
limitations, and electrochemical performance of the battery18. Increasing the percentage of
conductive additive and reducing the porosity results in higher electronic conductivity. However,
electrodes with less porosity have intrinsically slower ion transport, which poses a limitation17.
The convolution of electronic and ionic conduction processes is typically hard to distinguish13.

Several studies combine electrochemical impedance spectroscopy (EIS) and direct-current (DC)
polarization methods to evaluate ionic and electronic conduction in battery cells. These methods
allow for studying charge transfer at the macroscopic interface, ionic diffusion during cycling, and
the overall conductivity17,19. In a study by Wang et al. (2018), a combination of EIS with DC
polarization methods was used to separately measure ionic (σi) and electronic (σe) conductivities
of LiCoO2 and NMCs13. With increasing Ni content, they observed a significant increase in both
σi and σe, accompanied by considerable reductions in the activation barriers, accounting for the
improved rate capability of the nickel-rich NMCs13. Noh et al. (2013) also established a strong
correlation between high rate capability, electronic conductivity, and chemical Li diffusivity in
several NMCs11. A different study combined experiments and modeling to explore the relationship
between rate capability and microstructure in a LIB porous electrode18. They demonstrated that
the porosity of the carbon-binder domain was a vital parameter in obtaining the ionic and electronic
percolation network and obtained a threshold value for NMC loading (92 wt.%) over which the
electronic transport limitations dominate the rate performance of the electrode18.

9
However, a more fundamental understanding of interfacial and rate mechanisms is required10. The
proposed approach, Broadband dielectric spectroscopy (BDS), involves identifying limiting
phenomena by multiscale decomposition studies of the hierarchical electrodes and correlation with
the electrochemical performance of the cells20.

1.5 Broadband Dielectric Spectroscopy (BDS)

Broadband dielectric spectroscopy (BDS) classically studies the movement of ions and dipoles in
the presence of an electric field21. This technique replaces the light source with an oscillating
voltage, and matter interaction is determined by the perturbation of the field21. BDS investigates
the relaxation and rotation of dipoles at frequencies below 1012 Hz (Figure 5), and the frequency
at which a particular dipole responds is related to the "relaxation time" of that system (Figure
6a).10,21

Figure 5. Electromagnetic spectrum with wavelength, frequency, and energy with common spectroscopic
techniques listed.21

Broadband Dielectric Spectroscopy (BDS) may also be used to describe the electronic transport in
electrode materials for lithium batteries by subjecting it to a time-dependent electric field, which
can be described by the time-dependent polarization(s) and relaxation(s)22. In electronic
conductors, several polarization mechanisms occur at different scales (from interatomic to
macroscopic sizes) with distinct characteristic frequencies. The smaller scales respond at higher
frequencies, and the macroscopic scale has its response at lower frequencies due to the time scale
of its polarization (Figure 6b)10. The response is expressed by frequency-dependent complex
conductivity σ(ω), resistivity ρ(ω), and relative permittivity ε(ω), given by equation (1) where ε0
= 8.84 × 10−12 F·m−1 is the vacuum permittivity23.

𝜎(𝜔) = 𝜌(𝜔)−1 = 𝑖𝜔𝜀0 𝜀(𝜔) (1)

10
 Figure 6. (a) Schematic representation of the real (ε') and imaginary (ε") parts of the permittivity vs. frequency for
a solid. (b) Schematic description of a hierarchical architecture at different scales of a powdered material:
different sources of polarization vs. frequency and size.10

A decomposition procedure of the Nyquist plots of complex permittivity (ε′′ vs. ε′), resistivity (ρ′′
vs. ρ′), and conductivity (σ′′ vs. σ′) is used to evidence the different relaxations. The frequency-
dependent permittivity of a conducting material is generally given by equation (2)23.

𝜀𝑚𝐿 − 𝜀𝑚𝐻 𝑠−1

𝜎𝑑𝑐
𝜀(𝜔) = 𝜀∞ + [∑ ] + 𝐴(𝑖𝜔) + (2)
(1 + (𝑖𝜔𝜏𝑚 )1−𝛼𝑚 )𝛽𝑚 𝑖𝜔𝜀0
𝑚

Where ε∞ is the residual permittivity, ω = 2πν is the angular frequency in rads −1 (ν being the
frequency expressed in Hz). For each relaxation, εmL and εmH are the low and high-frequency limits
of the permittivity (Δεm is the dielectric strength), τm is the mean relaxation time; A, αm, and βm are
fitting parameters10. The term in brackets corresponds to the sum of the m relaxation functions
described by the empirical Havriliak–Negami (HN) functions24 because electrical polarizations at
different scales are additive due to their vector character. The relaxation functions that fit most
experimental spectra are Debye (αm = 0, βm = 1) and Cole−Cole (αm ≠ 0, βm = 1) functions whose
Nyquist representations display arcs of a circle centered on the real axis for the first one and below
this axis for the second one23. The term A(iω)s−1 (0 ≤ s ≤ 1), whose contribution occurs in the low-
frequency range, is due to a more or less disorderly conductive network (at a macroscopic scale).
It can also be due to distributions of contact resistances and capacitance between the metallization
and the sample. The Nyquist plot of such behavior is a straight line merged with the sample dc
conductivity contribution σdc/iωε023. The decomposition of the dielectric spectra into several
contributions (DC conductivity and relaxations) is carried out by successive subtractions of the
contributions starting from the lowest frequencies to the highest ones 23.

11
In the investigation of the electrical conductivity of a composite electrode, several types of
polarizations involving dielectric relaxations appear from low to high frequencies. Polarization of
space-charge due to the sample/current collector interface; polarization of aggregates (or clusters)
and particles due to resistive junctions; and electron transfers at nanometric or interatomic scale10.
BDS is a unique technique for these interfacial phenomena, allowing complex permittivity and
conductivity measurements over a wide frequency range from 10 Hz to 18 GHz and with
temperature variations23. The first detailed study of the technique in NMC (LiNi1/3Mn1/3Co1/3O2)
observed that the main limiting factor of NMC conductivity lies at the junctions between the
particles within the various clusters. The charge carriers are highly obstructed at this level,
resulting in a conductivity loss of about 3 to 4 orders of magnitude22.

1.6 Objectives of the study

This study aims to better describe the charge transport in commercial lithium-ion battery
electrodes. This will be achieved by studying several electrodes and changing parameters such as
the type and composition of active and non-active materials and linking the BDS measurements:
ex-situ (dry and wet electrodes), in-situ (electrode in a full cell), and operando (cycling electrode)
to the electrochemical performance of the cell. In-situ and operando experiments are not reported
in this study but are essential to the overall goal and will be carried out in upcoming experiments.
Identifying and discriminating the multiscale electronic and ionic contributions will help optimize
battery materials for high capacity and power applications such as electric vehicles.

12
2 Materials and Methods
2.1 Materials
The NMC powders used in this study were industrially supplied EV battery-grade active material
powders; LiNi0.5Mn0.3Co0.2O2 (NMC532) and LiNi0.8Mn0.1Co0.1O2 (NMC811). Other materials used
include poly-vinylidene fluoride-co-hexafluoropropylene (PVdF-HFP) (as the polymeric binder),
carbon black (TIMCAL SUPER C65) (as conductive additive), acetone (99.8%) (as slurry
solvent), 1M LiPF6 solution in an ethylene carbonate/diethyl carbonate mixture (EC/DEC 1:1 v/v)
(LP40 as electrolyte), a 3-component mould (7 mm), Hydraulic press, Micrometer screw gauge (±
0.1 µm), Beakers, and a Magnetic stirrer.

2.2 Material Characterizations

2.2.1 XRD Analysis
The x-ray diffraction patterns of the NMC powders were measured with a PANalytical XPert Pro
x-ray diffractometer equipped with a Cu Kα1 source (λ = 1.5408 Å) operating at 40 kV and 30 mA.
The patterns were acquired in a θ–2θ configuration with a 2θ range of 10 − 80° and a step size of
0.033º.

2.2.2 SEM – EDS Analysis

The surface morphology and cluster size analysis of the electrode materials were obtained by SEM
analysis performed using a Zeiss Supra 55 variable-pressure scanning electron microscope with a
field emission gun, equipped with an energy-dissipative X-ray detector for chemical
microanalysis.

2.3 Sample Preparation

The required amount of PVdF-HFP (binder) was mixed with 20 ml of acetone (99.8%) and stirred
for 3 hours in a magnetic stirrer at room temperature. Calculated quantities of NMC and CB to
give the needed volumetric percentages were added to the solution and stirred for 15 hours to
ensure component homogeneity. The solvent was then evaporated at 50 ºC, and the mixture was
ground and vacuum-dried for 2 hours at 90 ºC. The dry electrode material is subsequently stored
in a desiccator to avoid interaction with moisture.
Cylindrical pellets were prepared by compacting specific amounts of electrode powder in a 3-
component mould (diameter = 7 mm) at 250 MPa. The thickness of the resulting pellet was
measured with a micrometer screw gauge, and the pellet was transferred to a sample holder (to
prevent an air gap with the instrument). Silver paste was applied to both faces of the cylindrical
pellet, as shown in Figure 7, to ensure good metallic contact with the instrument's waveguide (3
mm) and short-circuit (7 mm). The weight and volume percentage of the prepared electrode are
shown in Table 2. Sample porosity is determined from each component's density, volume fraction,
and the electrode's volume.
Table 2. Weight and Volume percentage of prepared electrodes.
Samples Weight % Volume %
NMC AM PVdF-HFP CB NMC AM PVdF-HFP CB Porosity
A 97.5 2.5 0 67.5 4.5 0 28
B 95.5 2.0 2.5 64.5 3.5 4.0 28
C 93.5 3.0 3.5 61.5 5.2 5.3 28
D 91.5 4.0 4.5 58.2 6.8 7.0 28

13
2.4 BDS Measurement
The BDS setup (Figure 7) consists of a cylindrical electrode sample of known composition and
dimension with silver paint on both faces placed in a disk-shaped sample holder, at the
discontinuity of the coaxial waveguide in an ex-situ cell, connected to a Keysight 4990 low-
frequency impedance analyzer and an Agilent E8364B high-frequency network analyzer. A tight
sealing and a counter electrode are included for in-situ and operando measurements (with
electrolyte). In this study, however, due to technical constraints, electrodes wetted with electrolyte
were used in an ex-situ cell to observe the ionic contributions.

Figure 7. BDS Setup. Left: Cylindrical electrode sample dimension and silver metallization. Center:
Measurement devices (a) Ex-situ cell (b) In-situ/Operando cell; A: sealing (tight glass window) B: sample and
sample holder C: short-circuit (aluminium grid) D: porous glass paper separator E: counter-electrode (Li).20,25.
Right: Impedance analyzer.

The electromagnetic wave admittance of the sample was measured over multiple frequency bands,
with a frequency overlap ranging between 50 – 120 MHz in the two analyzers, confirming the
agreement of the measurement23. The frequency limitation of the setup is the theoretical cut-off
frequency of 19.6 GHz for the standard "APC7" used in the "SuperMit cell "26. An excitation port
at the input of the coaxial line enables the generation of an incident transverse electromagnetic
(TEM) field and, at the same time, absorbs the reflected mode. The complex admittance Y is
related to the reflection coefficient at the discontinuity by equation (3). Where Y0 = 0.02 S (50 Ω)
is the characteristic admittance of the coaxial line27.

1−Γ
𝑌 = 𝑌0 = 𝐺 + 𝑗𝐵 (3)
1+Γ

A calibration procedure is used to obtain a reference plane to minimize systematic errors generated
by the analyzers, waveguides, and dc-block in measurements. The calibration kit includes a 50 Ω
impedance load with the reflection coefficient Γ(ω) = 0, an open-circuit with capacitance C = 89
femtofarad (impedance Z = 1/jCω), and a short-circuit with the reflection coefficient Γ(ω) = –1
(and impedance Z = 0)20.
Following a successful calibration and complex admittance measurement, the permittivity ε(ω) =
ε′(ω) − iε′′(ω) and conductivity σ(ω) = σ′(ω) − iσ′′(ω) of the sample are calculated. The dielectric
spectra comprise 402 measuring points, and temperature-varied measurements were obtained
using dry nitrogen flux at temperatures between 200 and 300 K.

14
3 Results and Discussion
3.1 Physico-Chemical Characterizations
3.1.1 X-ray Diffraction (XRD)
The XRD diffraction patterns of active material powders NMC532 and NMC811 are shown in
Figure 8a. The two materials show sharp and well-defined diffraction peaks, indicating high
crystallinity.
(a) (b)

Figure 8. (a) X-ray diffraction patterns of NMC532 and NMC811 active materials, (b) NMC lattice with
the R3m space group.5

The peaks can be indexed to a hexagonal layered structure with space group R3m (α-NaFeO2
structure)13, with lithium (Li), oxygen (O), and transition metals (TM) occupying alternate atomic
layers, and the TM located at the center of the oxygen octahedron with same atomic site occupation
probability (Figure 8b)5. The R3m phase can be categorized as hexagonal according to the crystal
family, trigonal according to the crystal system, and rhombohedral according to the lattice
system23. The evident separation of peaks (006) and (102), as well as (108) and (110), indicates a
well-defined layered structure13,23.

3.1.2 Scanning Electron Microscopy (SEM)

The SEM images reveal the morphology and distribution of the active material powder and
electrode materials. Both materials are composed of spherical-shaped aggregates (clusters) with a
large size distribution. In NMC532 (Figure 9), the average cluster size is 13 ± 4 μm, while in
NMC811 (Figure 10), the average size is 10 ± 4 μm. With the addition of binder (PVdF-HFP) and
carbon black, there is a nearly homogenous distribution of the CB/binder network throughout the
electrode mix. Also, cluster fractures due to electrode compression were observed in the
calendered electrodes (Figure 9d).

15
(a) (b)

(c) (d)

Figure 9. SEM images of NMC532 clusters (a,b), and electrode mix with 5% CB (c,d)

(a) (b)

(c) (d)

Figure 10. SEM images of NMC811 clusters and particles. (a,b), 5%CB (c,d)

16
3.1.3 Energy Dispersive Spectroscopy (EDS)
A complementary chemical microanalysis of the active material powders by EDS was used to
obtain the chemical stoichiometry of the materials (Figure 11). The obtained sample composition
confirms a good agreement with the expected stoichiometry for each active material – with
NMC811 having ca. 81% Ni and NMC 532 having ca. 50% (Table 3).

Table 3. Percent composition of NMC532 and

NMC811 by EDS.

TM NMC532 NMC811
Ni (%) 49.57 81.23
Mn (%) 30.70 8.79
Co (%) 19.72 9.99

Figure 11. EDS Spectra of NMC532 and NMC811.

3.2 Broadband Dielectric Spectroscopy (BDS)

The scale-dependent relaxation frequency in a geometrical approach gives smaller-size systems a
relaxation response at higher frequencies and the larger scales in the hierarchy at lower
frequencies10. Additionally, in a kinetic approach, a shift to higher frequency is also observed with
increasing density and the mobility of the charge carrier (e.g., ionic contributions are generally
observed at lower frequencies than electrons)10.

3.2.1 Ex-situ Measurements (Dry Electrode)

The evolution of the real parts of permittivity and conductivity with frequency for NMC532 and
NMC811 dry electrodes are shown in Figure 12 and Figure 13 respectively. The permittivity
decreases as a function of frequency due to scale-lowering and decreasing charge density. The
conductivity figures show increased conductivity with frequency due to increasing charge
depletion and less contact resistance as it approaches the core active material interfaces. With the
introduction of the conductive additive, carbon black (CB), there is a significant increase in the
dc-conductivity at a lower frequency (< 107 Hz) due to the occupation of CB within the cluster
networks and a lowering of contact and space-charge resistances. The dc/sample conductivity
saturation is observed at 7% CB for NMC532 and 5% CB for NMC811.

17
 (a) (b)

Figure 12. Real parts of (a) permittivity and (b) conductivity as a function of frequency for NMC532 dry electrode with
increasing CB vol.%.

(a) (b)

Figure 13. Real parts of (a) permittivity and (b) conductivity as a function of frequency for NMC811 dry electrode with
increasing CB vol.%.

The decomposition of the dielectric spectra into several contributions (dc conductivity and
relaxations) is carried out by successive subtractions of the contributions starting from the lower
frequencies. The low-frequency part of the spectrum is well fitted by a quasi-vertical straight line
(P1 contribution) corresponding to the sample dc conductivity (Figure 14a). After subtracting the
P1 contribution, the subsequent dielectric relaxations are fitted to the Cole-Cole function (Equation
(3). The relaxations P2b, P2, P3, and P4, are thus space-charge relaxations associated with the
binder/NMC assembly, sample polarization, active material cluster, and particle polarizations,
respectively23. From the resistivity Nyquist (Figure 14b), the conductivity values of the sample
and cluster are also obtained.

18
 (a)

1.0×105 Hz

(b)

Figure 14. Nyquist plots of (a) complex permittivity, ε′′(ω) vs. ε′(ω) (b) complex resistivity, ρ′′(ω) vs. ρ′(ω) of
NMC532 dry electrode, 0% CB at RT (293 K).

The evolution of the dielectric spectra with temperature provides additional information on the
thermal activation of the polarizations and conduction mechanisms in the dry electrode. With
NMC532 in the absence of CB (Figure 15), the sample has a significant activation barrier at low
and high frequencies, evidenced by the lowering of the permittivity and conductivity with
decreasing temperature. This thermal activation becomes less significant with increasing CB
vol.%, and the near-zero conductivity activation with 7% CB (Figure 16b) represents the behavior
of a percolated electrode as the quasi-metallic CB/binder composite becomes the principal
electronic pathway20.

19
 (a) (b)

Figure 15. Real parts of (a) permittivity and (b) conductivity as a function of frequency for NMC532 dry
electrode 0% CB with temperature varied.

(a) (b)

Figure 16. Real parts of (a) permittivity and (b) conductivity as a function of frequency for NMC532 dry
electrode 7% CB with temperature varied.

The activation of the relaxation frequencies and conductivity values associated with the various
phenomena follows the Arrhenius law (Equation(4)23, where ν0n and σ0n are the pre-factors; En is
the activation energy; T is the temperature, and k = 1.38 × 10−23 JK−1 is the Boltzmann constant.
The resulting Arrhenius plots for NMC532 dry electrode with 0% CB are presented in Figure 17.

𝐸𝑛 𝐸𝑛
𝜈𝑛 = 𝜈0𝑛 exp (− ) 𝜎𝑛 = 𝜎0𝑛 exp (− ) (4)
𝑘𝑇 𝑘𝑇

20
 (a) (b)

Figure 17. Arrhenius plots of (a) dielectric relaxation frequency (b) real conductivity in the sample/binder,
sample for NMC532 dry electrode 0% CB.

A comparison of the obtained parameters and activation energies in NMC532 and NMC811 is
presented in Table 4. With scale increase, i.e., from particle to cluster to sample, the activation
barrier increases, and the relaxation frequency and conductivity values reduce exponentially. The
cluster, sample, and dc conductivity of the nickel-rich NMC811 dry electrode is one order of
magnitude higher than NMC532 (p-type semiconductor) and with a lower activation barrier, in
agreement with increased electron density in nickel-rich electrodes and NMC811 as an n-type
semiconductor; with its Fermi-level closer to the conduction band5.

The threshold of electronic percolation was also observed at a lower CB vol.% for NMC811 – 4%
CB, while it is 5% CB for NMC532, suggesting a possible increase in specific capacity due to less
required non-active material.

Table 4. Obtained dielectric parameters for NMC532 and NMC811 dry electrode 0% CB at RT (293 K).
NMC532 NMC811
Sample/binder Conductivity, σsb (S/m) 1.19 × 10–4 1.27 × 10–3
Sample Conductivity, σs (S/m) 4.57 × 10–3 1.81 × 10–2
Cluster Conductivity, σcl (S/m) 1.35 1.80 × 101

Sample/binder Polarization, νsb (Hz) 1.09 × 104 1.26 × 105

Esb (eV) 0.25 0.17
Sample Polarization, νs (Hz) 1.26 × 106 4.08 × 106
Es (eV) 0.22 0.19
Cluster Polarization, νcl (Hz) 1.55 × 108 3.17 × 109
Ecl (eV) 0.19 0.16
Particle Polarization, νp (Hz) 1.26 × 109
-
Ep (eV) 0.16

21
3.2.2 Wet Electrode Measurements
To observe the ionic motion and the influence of adsorbed ions on the electrode conductivity and
NMC cluster polarization, liquid electrolyte LP40 (1M LiPF6 solution in an ethylene
carbonate/diethyl carbonate mixture 1:1 v/v) was introduced to the compressed electrode to fill in
the porous layer. A comparison between the dielectric spectra of the dry and wet electrodes is
presented in Figure 18 and Figure 19 for NMC532 and NMC811, respectively.
(a) (b)

Figure 18. Real parts of (a) permittivity and (b) conductivity as a function of frequency for NMC532 dry and wet
electrodes at 0% and 5% CB.

(a) (b)

Figure 19. Real parts of (a) permittivity and (b) conductivity as a function of frequency for NMC811 dry and
wet electrodes at 0% and 5% CB.

With the introduction of ions, There is a significant increase in the permittivity and conductivity
values, especially at lower frequencies, characteristic of the ionic contributions. The influence of
CB addition on the conductivity of the wet electrode is also less significant when compared to the
dry electrode; this may be due to the lowering of CB conductivity in the presence of the
electrolyte28. The Nyquist decomposition (Figure 20) reveals five dielectric relaxations with low
relaxation frequency W1 associated with the polarization of the ionic double-layer at the sample-
electrolyte interface. The W2 relaxation is due to ionic motion and is thermally activated up to the
freezing temperature of the electrolyte between 248 and 253 K28. The other relaxations, W3, W4,
and W5, are associated with the sample, cluster, and particle polarizations, respectively. These

22
polarizations are associated with a significant increase in dielectric strength Δε, corresponding to
the space-charge or double-layer capacity. This increase is a result of the presence of adsorbed
ions in the space-charge regions, which greatly decreases the thickness of the space-charge ω (as
Δε α ω–1)20,23. The conductivity Nyquists (Figure 20b) is more suitable to observe the effective
conductivity contributions as ion and electron flows are carried out in parallel23.
(a)

(b)

Figure 20. Nyquist plots of (a) complex permittivity, ε′′(ω) vs. ε′(ω) (b) complex conductivity, σ′′(ω) vs. σ′(ω)
of NMC532 wet electrode, 0% CB at RT (293 K).

23
The evolution of the dielectric spectra with temperature (Figure 21) reveals thermal activation
across all frequencies. At low frequency (< 107 Hz), a significant drop in conductivity and
increased activation energy is observed below the freezing point of the electrolyte (< 253 K). The
dielectric parameters of the wet electrode also follow the Arrhenius law (4). The effect of
electrolyte freezing on the relaxation frequency of ionic motion (Figure 22a) and ionic conductivity
(Figure 22b) can be observed as the values drop sharply below freezing temperature.
(a) (b)

Figure 21. Real parts of (a) permittivity and (b) conductivity as a function of frequency for NMC532 wet
electrode 0% CB with temperature varied.

(a) (b)

(c) (d)

Figure 22. Arrhenius plots of (a) relaxation frequency of the ionic motion ν2 in NMC532
(b) bulk ionic conductivity σsi. Comparison of NMC532 and NMC811 (c) cluster relaxation
frequencies ν4 (d) cluster conductivity σcl in dry and wet electrodes at RT.

24
The lower values of cluster relaxation frequency observed in the wet electrodes (Figure 22c) can
be linked to changes in the space-charge depletion layer. The relatively high charge density on the
cluster surface increases the time scale of electrical polarization and lowers the relaxation
frequency. The conductivity, however, is increased in the wet electrodes (Figure 22c) due to the
elimination of cluster surface resistance. Hence, in the wet electrodes, we observe the core
conductivity of the cluster23. A summary of the obtained dielectric parameters in NMC532 and
NMC811 wet electrodes is presented in Table 5.

Table 5. Obtained dielectric parameters for NMC532 and NMC811 wet electrode0% CB at RT (293 K) and the
activation energy above freezing temperature.
NMC532 NMC811
Bulk ionic Conductivity, σsi (S/m) 2.43 × 10–1 3.38 × 10–2
Local ionic Conductivity, σi (S/m) 8.64 × 10–1 7.23 × 10–2
Cluster Conductivity, σcl (S/m) 1.72 2.30 × 101

Double-layer capacitance, ν1 (Hz) 1.47 × 102 6.85 × 100

E1 (eV) 0.27 -
Ionic Motion, ν2 (Hz) 1.02 × 105 3.75 × 104
E2 (eV) 0.23 0.26
Sample Polarization, ν3 (Hz) 1.14 × 106 7.12 × 106
E3 (eV) 0.21 0.19
Cluster Polarization, ν4 (Hz) 8.87 × 106 6.67 × 107
E4 (eV) 0.21 0.14
Particle Polarization, ν5 (Hz) 2.45 × 109 1.36 × 109
E5 (eV) 0.18 -

A set of forward−backward ionic movements describes the long-range conductivity σsi (or bulk
ionic conductivity), and the short-range conductivity σi (or local ionic conductivity) is linked to
hopping between adsorption sites20,23. Using Archie's law (5) to relate the liquid electrolyte
conductivity σel and the ionic conductivity σsi, we can estimate the electrode tortuosity, defined as
the ratio of the actual distance traveled by the ions per unit length of the medium. The tortuosity
is a measure of the lengthening of the transport channel by the twisted structures of the pores in
the presence of the solid electrode materials7,23.

𝜎si = 𝜎el . 𝜙 𝑚 (5)

Where 𝜙 is the porosity volume fraction and the exponent m is Archie's exponent, correlated to
the tortuosity factor. With σel = 1 S/m (LP40)28, and 𝜙 = 0.28, the calculated m value is 1.1 for
NMC532, and 2.7 for NMC811. The m value in the NMC811 electrode agrees with that obtained
for similar electrodes23, while its low value in NMC532 may be due to microporosity (additional
ionic pathways within the AM clusters).

25
4 Conclusions and Perspectives
This study presents the multiscale description of ionic and electronic transport in two electric-
vehicle grade lithium battery materials; NMC532 and NMC811, using broadband dielectric
spectroscopy to provide fundamental and intrinsic details of the ion and electron transport and to
gain additional information on how the composite variations influence its properties and ultimately
the performance. Figure 23 summarizes the relaxation frequency and energetic barriers (activation
energies). NMC811 is generally associated with higher frequency and lower activation energy for
each contribution. However, the lower barrier to ionic motion in NMC532 may be due to its
microporosity. In addition, the wet NMC811 cluster has a lower relaxation frequency barrier
relative to the dry cluster, which implies improved electrical polarization in the presence of the
electrolyte. The nickel-rich electrode (NMC811) is therefore projected to have better rate
performance.

Figure 23. Activation energies vs. relaxation frequencies at RT attributed to ionic motions in the pore network
filled with LP40 and electronic motions in the samples and clusters of NMC532 and NMC811 with 0% CB.

As a perspective, it will be essential to correlate these results with the full-cell rate performance
by multiple galvanostatic cycling measurements. Further BDS measurements, including full-cell
in-situ and operando measurements described in the report, will also be important to characterize
the electrode behavior. Other complementary measurements include electrochemical impedance
spectroscopy (EIS), to describe the correlation of the low-frequency ionic contributions and
macroscopic electronic contributions in the complete cell, and focused ion beam–scanning electron
microscopy (FIB-SEM), to provide an improved morphological description from the 3D geometry
of the electrode. The results obtained from this study will serve as additional preliminary results
in supporting theoretical studies, electrode modeling, and the overall study of transport limitations
in composite battery electrodes and material optimizations, especially for electric vehicle
applications.

26
References

1. Kazda, T. & Vanýsek, P. The chalkboard: Lithium batteries as electrochemical sources of

energy. Electrochemical Society Interface 25, 47–49 (2016).
2. Tarascon, J.-M. & Armand, M. Issues and challenges facing rechargeable lithium batteries.
www.nature.com (2001).
3. Emissions by sector - Our World in Data. https://ourworldindata.org/emissions-by-sector.
4. Kittner, N., Lill, F. & Kammen, D. M. Energy storage deployment and innovation for the
clean energy transition. Nature Energy 2, (2017).
5. Sun, H. & Zhao, K. Electronic Structure and Comparative Properties of LiNi xMnyCozO2
Cathode Materials. The Journal of Physical Chemistry C 121, 6002–6010 (2017).
6. Chen, X., Shen, W., Vo, T. T., Cao, Z. & Kapoor, A. An overview of lithium-ion batteries
for electric vehicles. in 10th International Power and Energy Conference, IPEC 2012 230–
235 (2012). doi:10.1109/ASSCC.2012.6523269.
7. Besnard, N. et al. Multiscale Morphological and Electrical Characterization of Charge
Transport Limitations to the Power Performance of Positive Electrode Blends for Lithium-
Ion Batteries. Advanced Energy Materials 7, (2017).
8. Molenda, J. Li-ion batteries for electric vehicles. Annales UMCS, Chemistry 66, (2011).
9. The Internal Structure of Lithium Polymer Battery. https://www.ludabattery.com/the-
internal-structure-of-lithium-polymer-battery/.
10. Badot, J. C., Lestriez, B. & Dubrunfaut, O. Interest in broadband dielectric spectroscopy to
study the electronic transport in materials for lithium batteries. Materials Science and
Engineering B: Solid-State Materials for Advanced Technology 213, 190–198 (2016).
11. Noh, H. J., Youn, S., Yoon, C. S. & Sun, Y. K. Comparison of the structural and
electrochemical properties of layered Li[NixCoyMnz]O2 (x = 1/3, 0.5, 0.6, 0.7, 0.8 and 0.85)
cathode material for lithium-ion batteries. Journal of Power Sources 233, 121–130 (2013).
12. Lipson, A. L. et al. Stabilizing NMC 811 Li-Ion Battery Cathode through a Rapid
Coprecipitation Process. (2021) doi:10.1021/acsaem.0c03112.
13. Wang, S., Yan, M., Li, Y., Vinado, C. & Yang, J. Separating electronic and ionic
conductivity in mix-conducting layered lithium transition-metal oxides. Journal of Power
Sources 393, 75–82 (2018).
14. Goodenough, J. B. & Kim, Y. Challenges for rechargeable Li batteries. Chemistry of
Materials vol. 22 587–603 Preprint at https://doi.org/10.1021/cm901452z (2010).
15. Tamirat, A. G., Guan, X., Liu, J., Luo, J. & Xia, Y. Redox mediators as charge agents for
changing electrochemical reactions. Chemical Society Reviews vol. 49 7454–7478 Preprint
at https://doi.org/10.1039/d0cs00489h (2020).
16. Baker McKenzie. Battery Storage-a global enabler of the Energy Transition 2022. (2022).
17. Chen, Y. H., Wang, C. W., Zhang, X. & Sastry, A. M. Porous cathode optimization for
lithium cells: Ionic and electronic conductivity, capacity, and selection of materials. Journal
of Power Sources 195, 2851–2862 (2010).
18. Hamed, H. et al. Demystifying Charge Transport Limitations in the Porous Electrodes of
Lithium-Ion Batteries. Advanced Energy Materials 10, (2020).
19. Dedryvère, R. et al. Electrode/electrolyte interface reactivity in high-voltage spinel
LiMn1.6Ni0.4O4/Li4Ti5O12 lithium-ion battery. Journal of Physical Chemistry C 114, 10999–
11008 (2010).

27
20. Seid, K. A. et al. An in situ multiscale study of ion and electron motion in a lithium-ion
battery composite electrode. Advanced Energy Materials 5, (2015).
21. Woodward, W. H. H. Broadband Dielectric Spectroscopy - A Practical Guide. in ACS
Symposium Series vol. 1375 3–59 (American Chemical Society, 2021).
22. Seid, K. A. et al. Multiscale electronic transport in Li1+xNi1/3-uCo1/3-vMn1/3-wO2: A
broadband dielectric study from 40 Hz to 10 GHz. Physical Chemistry Chemical Physics
15, 19790–19798 (2013).
23. Agrawal, A. et al. Influence of a Liquid Electrolyte on Electronic and Ionic Transfers in a
LiNi0.5Mn0.3Co0.2O2/Poly(vinylidene Fluoride- co-hexafluoropropylene)-Based Composite
Material. Journal of Physical Chemistry C 125, 17629–17646 (2021).
24. Böttcher, C. J. F., Bordewijk, P. & Rip, A. Theory of electric polarization. (Elsevier
Scientific Pub. Co, 1973).
25. Belhadj-Tahar, N.-E. & Fourrier-Lamer, A. Broad-Band Analysis of a Coaxial
Discontinuity Used for Dielectric Measurements. 346 IEEE Transactions On Microwave
Theory And Techniques vol. 34 (1986).
26. Sadou, H. et al. An approach based on ANFIS and input selection procedure for microwave
characterization of dielectric materials. COMPEL - The International Journal for
Computation and Mathematics in Electrical and Electronic Engineering 37, 799–813
(2018).
27. Belhadj-Tahar, N.-E., Fourrier-Lamer, A. & de Chanterac, H. Broad-Band Simultaneous
Measurement of Complex Permittivity and Permeability Using a Coaxial Discontinuity.
IEEE Transactions On Microwave Theory And Techniques vol. 38 (1990).
28. Panabière, E., Badot, J. C., Dubrunfaut, O., Etiemble, A. & Lestriez, B. Electronic and Ionic
Dynamics Coupled at Solid-Liquid Electrolyte Interfaces in Porous Nanocomposites of
Carbon Black, Poly(vinylidene fluoride), and γ-Alumina. Journal of Physical Chemistry C
121, 8364–8377 (2017).

28