Functional Binders at the Interface of Negative

and Positive Electrodes in Lithium Batteries

by Fabian Jeschull

Licenciate Thesis

Abstract

In this thesis, electrode binders as vital components in the fabrication

of composite electrodes for lithium-ion (LIB) and lithium-sulfur bat-

teries (LiSB) have been investigated.

Poly(vinylidene difluoride) (PVdF) was studied as binder for sulfur-

carbon positive electrodes by a combination of galvanostatic cycling

and nitrogen absorption. Poor binder swelling in the electrolyte and

pore blocking in the porous carbon were identified as origins of low

discharge capacity, rendering PVdF-based binders an unsuitable

choice for LiSBs. More promising candidates are blends of

poly(ethylene oxide) (PEO) and poly(N-vinylpyrrolidone) (PVP). It

was found that these polymers interact with soluble lithium polysul-

fide intermediates generated during the cell reaction. They can in-

crease the discharge capacity, while simultaneously improving the

capacity retention and reducing the self-discharge of the LiSB. In con-

clusion, these binders improve the local electrolyte environment at the

electrode interface.

Graphite electrodes for LIBs are rendered considerably more stable in

‘aggressive’ electrolytes (a propylene carbonate rich formulation and

an ether-based electrolyte) with the poorly swellable binders

poly(sodium acrylate) (PAA-Na) and carboxymethyl cellulose sodium

salt (CMC-Na). The higher interfacial impedance seen for the conven-

tional PVdF binder suggests a protective polymer layer on the parti-

cles. By reducing the binder content, it was found that PAA-Na has a

stronger affinity towards electrode components with high surface are-

as, which is attributed to a flexible polymer backbone and a higher

density of functional groups.

Lastly, a graphite electrode was combined with a sulfur electrode to

yield a balanced graphite-sulfur cell. Due to a more stable electrode-

electrolyte interface the self-discharge of this cell could be reduced

and the cycle life was extended significantly. This example demon-

strates the possible benefits of replacing the lithium metal negative

electrode with an alternative electrode material.

To Karlheinz Jeschull (1936-2014)

“Wer aufhört besser zu werden,

hat längst aufgehört gut zu sein”

Acknowledgements

First of all I would like to thank my main supervisor Daniel Brandell for his support and enthusiasm, for the valuable discussions in our Friday meetings and for giving me the independence to pursue my own ideas within this project. I highly appreci-ate the many opportunities to go on conference travels and summer schools. I thank Kristina Edström for giving me the opportunity to work and study in the Ång-ström Battery group. I would also like to gratefully acknowledge Matthew J. Lacey for the successful and productive collaboration beyond the time of my degree project, for his time and interest in my new projects and for his help and advice in electrochemistry-related problems. Special thanks to Leif Nyholm for proof reading this thesis and to Gerrit Boschloo for taking on the role as opponent in this licenciate thesis. I thank the PUB-group (Tim Bowden, Daniel Brandell, Jonas Mindemark and Andre-as Bergfelt and Ronnie Mogenson) for the nice atmosphere during our regular Thursday meetings I wish to acknowledge the general help in the lab of Henrik Eriksson and Reza Younesi, Haptom D. Asfaw, Bing Sun, Fredrik Lindgren, Julia Maibach, Chao Xu, Bertrand Philippe, Jonas Mindemark and Andreas Bergfelt for the fruitful discus-sions and assistance with analytical tools, data interpretation or polymer synthesis. Thanks to my summer and project workers who contributed to my work: Andoria Kotronia (Summer 2014), Valentin Kopf (Summer 2015), Tayfun Koçak (Summer 2014) and Aqib Hayat (Autumn 2015). Not less important to mention is the good atmosphere in this group. Special thanks to all the people participating in our fika breaks, after works and weekend activi-ties: Mario, Bertrand, Sara, Matt, Solveig, Stéven and Marco, as well as the ‘Inor-ganic AW Group’ (Kristina, Linus, David, Tim, Johan (2x) and Rickard) Tim Nordh – ‘slapping bros’ forever ;) Thanks to my office buddy Linus von Fieant and to Rickard Eriksson for passing on his office spot to me - ‘Das Kontor’ was recently inaugurated Special thanks to Christoph Howe for the countless hours at klättercentret. Finally I would like to express my gratitude to my parents, Dirk and Claudia:

Ich danke euch für eure stetes Interesse an meiner Ausbildung und eure

Unterstützung.

List of Publications

This thesis is based on the following peer-review articles, which are referred

to in the text by their roman numerals.

I M.J. Lacey, F. Jeschull, K. Edström, and D. Brandell (2014)

Functional, water-soluble binders for improved capacity and

stability of lithium-sulfur batteries. J. Power Sources 264, 8-14

II M.J. Lacey, F. Jeschull, K. Edström, and D. Brandell (2014)

Porosity Blocking in Highly Porous Carbon Black by PVdF

Binder and Its Implications for the Li-S System. J. Phys. Chem.

C, 118, 25890-25898

III F. Jeschull, M.J. Lacey and D. Brandell (2015) Functional

binders as graphite exfoliation suppressants in aggressive elec-

trolytes for lithium-ion batteries. Electrochim. Acta, 175, 141-

150.

IV F. Jeschull, D. Brandell, K. Edström and M.J. Lacey (2015) A

Stable Graphite Negative Electrode for the Lithium-Sulfur Bat-

tery, Chem. Commun., 51, 17100-17103

Reprints were made with permission from the respective publishers.

Publications not included in this thesis

M. J. Lacey, F. Jeschull, K. Edström, D. Brandell (2013) Why PEO

as a binder or polymer coating increases capacity in the Li-S-system,

Chem. Commun., 49, 8531

L.T. Costa, B. Sun, F. Jeschull, and D. Brandell (2015) Polymer-

ionic liquid ternary systems for Li-battery electrolytes: Molecular

dynamics studies of LiTFSI in a EMIm-TFSI and PEO blend, J.

Chem. Phys. 143, 024904

Comments on my contributions to the publications

I I was part in the development of the electrode formulations,

investigated the polymer-polysulfide interactions and I was in-

volved in all discussions of the results

II I planned and prepared all samples for the physisorption ex-

periments, performed the measurements and analyzed the re-

sults. I took part in the writing process.

III I planned and performed all experiments and wrote the manu-

script (main author of the manuscript).

IV I planned and performed all experiments, analyzed the data

and took part in the writing process (main author of the manu-

script).

Contents

Introduction ................................................................................................... 12

Lithium Batteries .......................................................................................... 13 Lithium-Ion Batteries ............................................................................... 15

Commonly Used Electrode Materials .................................................. 16 Electrolyte components and the Solid-Electrolyte Interface (SEI) ...... 19

Lithium-Sulfur Batteries .......................................................................... 21 Sulfur-Carbon Composite Electrodes .................................................. 21 Cell Chemistry and Analysis of Lithium-Sulfur Batteries ................... 22 Self-discharge and the Redox Shuttle Mechanism .............................. 23 Lithium Metal Negative Electrodes ..................................................... 24

Functional Polymers as Electrode Binders ................................................... 25 Electrode Binders ..................................................................................... 26

The “Gold Standard” – Poly(vinylidene difluoride) ............................ 26 Water-based Binders for Negative Electrodes ..................................... 27 Binders for Novel Positive Electrodes ................................................. 28

Scope of this Thesis ...................................................................................... 29

Results and Discussion ................................................................................. 30 Binders for Sulfur-Carbon-Composite Electrodes ................................... 30

Pore-Blocking Effect of PVdF-based Binders in Composite

Electrodes ............................................................................................ 30 Water-Soluble Functional Binders for Sulfur Electrodes .................... 33

Binders for the Graphite Electrode ........................................................... 36 Binders as Surfactants and Exfoliation Suppressants .......................... 36 Optimization of Conventional Graphite Electrodes ............................. 43

A Stable Graphite Electrode for the Lithium-Sulfur System ................... 44

Conclusions ................................................................................................... 48

Sammanfattning på svenska .......................................................................... 50

Bibliography ................................................................................................. 52

Abbreviations

General

LIB lithium-ion battery

LiSB lithium-sulfur battery

EV electrical vehicle

HEV hybrid electrical vehicle

Chemicals

CMC-Na carboxymethyl cellulose sodium salt

DMC dimethyl carbonate

DME 1,2-dimethoxyethane

DOL 1,3-dioxolane

EC ethylene carbonate

LCO lithium cobalt oxide, LiCoO2

LFP lithium iron phosphate, LiFePO4

LNMC lithium nickel-manganese-cobalt oxide, LiNi1/3Mn1/3Co1/3O2

LP40 common non-aqueous electrolyte formulation, comprising

1 M LiPF6 in DMC and EC (v/v = 1:1)

LTO lithium titanate, Li4Ti5O12

NMP N-methylpyrrolidone

PAA poly(acrylic acid)

PAA-Na poly(sodium acrylate)

PC poly propylene

PEG poly(ethylene glycol) (typically low molecular weight)

PEO poly(ethylene oxide) (high molecular variety of PEG)

PS (soluble) lithium polysulfides, Li2Sx

PVdF poly(vinylidene difluoride)

PVdF-HFP poly(vinylidene difluoride-hexafluoropropane)

PVP poly(N-vinylpyrrolidone)

SBR styrene-butadiene rubber

TEGDME tetraglyme or tetra(ethylene glycol) dimethyl ether

TFSI bis(trifluoromethane sulfonimide)

Electrochemistry

C.E. Coulombic efficiency, Qdischarge/Qcharge

CV cyclic voltammetry

DoD depth of discharge

I current

Q charge per unit weight (mAh g-1

), 𝑄 = ∫ 𝑑𝑞𝑄

0

SEI solid-electrolyte interphase

SHE standard hydrogen electrode

V volt (SI unit), [V] = J/C

E electrochemical potential

Analysis

BET Brunauer-Emmett-Teller (surface area analysis)

BJH Barrett-Joyner-Halenda (pore size and volume analysis)

EDX energy-dispersive X-ray spectroscopy

EIS electrochemical impedance spectroscopy

SEM secondary electron microscopy

XPS x-ray photoelectron spectroscopy

12

Introduction

This year, for the first time in modern human history, the atmospheric con-

centration of CO2 passed the 400 ppm milestone (it was last reached 3 mil-

lion years ago)1. Anthropogenic climate change caused by consumption of

fossil fuels and emission of greenhouse gases has become a central scientific

and political challenge of our generation and the generations to come2. In the

light of continuous economic growth and a growing global population3 ener-

gy consumption and hence the concentration of greenhouse gases is expected

to rise further in the coming decades4. For instance, the present global annual

energy consumption of 14 TW (corresponding to 1010

metric tons of oil) is

expected to double until the middle of this century4,5

.

In order to mitigate CO2 emissions the current power generation needs to

be put on a sustainable and renewable basis6. The intermittent nature of re-

newable energy generators, e.g. wind or solar power, requires highly effi-

cient, large-scale energy storage technologies in order to balance the electri-

cal grid, i.e. to level the peak load7,8

. Storage devices for such facilities are

required where renewable energy accounts for > 10 % of production9. Today

only 1 % of the global energy production is put into interim storage5.

Pumped hydroelectric systems and to a much lesser extent compressed air

storage account for 98% of these storage systems10

. In the future they are

likely to be complemented by electrical energy storage systems for load lev-

eling to accommodate daily peak loads in the grid8. These systems allow a

more efficient use of the grid capabilities7. Here, electrochemical storage

systems, e.g. batteries, are attractive solutions due to their scalability and

pollution-free operation, high efficiencies and long cycle life10–12

.

The transportation sector currently accounts for about 20 % of all emis-

sions, a figure which needs to be greatly reduced in the future by use of (hy-

brid) electrical vehicles (HEVs and EVs) or fuel cell powered vehicles13

.

Either way, light, flexible and high energy density batteries will play a key

role as energy storage devices in these vehicles.

In this context, the state-of-the-art, rechargeable storage system with the

highest energy density is the lithium-ion battery.

13

Lithium Batteries

Lithium-ion batteries (LIBs) rapidly conquered significant parts of the ener-

gy storage market after their commercialization in 199111,14,15

. Lithium ex-

hibits the lowest electrochemical potential of all elements (-3.04 V vs. SHE),

thus yielding high cell voltages16

. The small and light lithium ion is stored

readily as a guest in host structures allowing for superior specific capaci-

ties10

. So far the energy density of LIBs (the product of cell voltage and ca-

pacity by normalized weight or volume), is larger than that for any rivalling

rechargeable battery system, e.g. lead-acid or nickel-metal hydride.

Most applications, including portable user electronics, stationary energy

storage (for the electrical grid) and electrical vehicles, require high volumet-

ric energy densities so as to maximize the capacity in a confined space9,17

.

The amount of stored energy in a state-of-the-art LIB has more than doubled

since its introduction9,16

, while the costs per kWh has decreased substantial-

ly, i.e. 10-fold to about 300-350 $/kWh16,18

. However, this goal was primari-

ly achieved by cell engineering (e.g. cell packaging)19

. From a material per-

spective LIBs appear to approach an intrinsic limit owing to capacity re-

strictions of the electrochemical active components currently used16

. Current

EV driving ranges between 150-200 km per charge† (90-100 Wh/kg and 450

kg battery weight)20

, are not competitive with combustion engines or fuel

cells in neither range nor price5. Furthermore, substantial amounts of energy

are required (~400 kWh) for material extraction and synthesis and battery

manufacturing, accounting for CO2 emissions of 75 kg/kWh of LIB5.

‘Post-lithium-ion batteries’ set out to overcome this limitation by means

of radically new cell concepts involving light non-metal positive electrode

materials, namely sulfur and oxygen. The abundance of these materials

could potentially reduce the costs of lithium batteries16

, help to produce bat-

teries more energy efficient5 and raise current energy densities multifold

21,22.

These new cell concepts provide new opportunities to meet the envisaged

storage capabilities, for instance, for the electrification of transport22

. Table 1

compares theoretical capacities and energy densities based on the pure elec-

trode materials for some established battery systems and lithium batteries as

a first figure of merit. A more detailed analysis can be found elsewhere16,17

.

† Exceptions being the models by Tesla Motors, due to lightweight design and the Belloré Bluecar, which employs a lithium-solid polymer battery

14

Table 1. Cell characteristics of selected battery systems.

Battery Specific capacities Energy densities Vcell

Anode[1]

Cathode[1]

gravimetric

(Wh/kg)

volumetric

(Wh/L)

Lead-acid PbSO4*H2O

(167/1051)

PbSO4*H2O

(167/1051)

167

(40-45)[2]

1051 2.0

Nickel-Metal

Hydride

LaNi5H6[3]

(306/1531)

Ni(OH)2

(289/1156)

178

(76)[4]

790

(275)[4] 1.2

Li-Ion Graphite

(372/818)

LiCoO2

(140/700)

LiFePO4

(170/612)

386.5

(120)[5]

408.4

1433.7

(285)[5]

1225.5

3.8

3.5

Li-S Li

(3884/2074)

Li2S

(1167/1937)

1973.6 2203.3 2.2

Li-O2

(non-aqueous

& aqueous)

Li

(3884/2074)

Li2O2

(1798/3620)

LiOH*H2O

(1276/1952)

3565.3

3266.1

3824.4

3419.7

2.9

3.4

[1] values given in brackets: theoretical gravimetric capacity & volumetric capacity, respectively

[2] practical achievable energy density, ref23 [3] Zuettel, A. (2003) Materials for hydrogen storage, Materials today, 6, 24-33

[4] Twicell HR-4/3A (3500 mAh) of SANYO

[5] US17670 4/3A-Battery (1200 mAh) of SONY Energytec Inc.

The gravimetric and volumetric energy densities refer to the discharged state. Note that Li-S and

Li-O2 cells are usually assembled in their charged states. The energy densities were calculated on the basis of E = (Cc*Ca)*Vcell/(Cc+Ca); Cc/a: specific capacities of the cathode and anode

respectively, Vcell: average cell potential based on thermodynamic data at 298 K21.Practical

energy densities are about a quarter to one half lower than their theoretical values due to addi-tional weight and volume from cell packaging, current collectors, electrolyte, etc.

15

Lithium-Ion Batteries

The first commercial lithium-ion battery comprised a graphite negative elec-

trode, a lithium cobalt oxide (LCO), LiCoO2, positive electrode and a non-

aqueous, carbonate-based electrolyte‡.

The system is commonly referred to as the “rocking chair” battery, since

lithium ions are moving back and forth between the two electrodes (Figure

1). On the graphite side ions intercalate between graphite layers while on the

cathode side ions are inserted into vacant sites in the crystal structure under

change of the oxidation state of the transition metal. Furthermore, by apply-

ing non-aqueous electrolytes it is possible to overcome the cell potential

limitations encountered in water-based battery systems and hence greatly

enhance the energy density. LIB electrolytes employ alkylcarbonates with

high anodic stabilities as solvents and a lithium salt with a weakly or non-

coordinating anion. The general redox reaction for a lithium-ion battery sys-

tem is given by:

LixCn ⇆ x Li+ + n C + x e

-

Li1-xCoO2 + x Li+ + x e

- ⇆ LiCoO2

Figure 1. Principle of a “rocking chair battery”. Upon discharge Li, which is interca-lated between the individual graphite layers is oxidized (LiC6 C6 + Li

+ + e

-) and

leaves the electrode. At the positive electrode Li-ions are inserted into the host struc-ture of the inorganic material due to change of the oxidation state of the metal ion (e.g. Co

4+ + e

- Co

3+). Electrons move through an external circuit to perform work,

while ions travel through the electrolyte in order to preserve charge balance.

‡ The terms anode and cathode are frequently used as synonyms for the negative and positive electrode, respectively. However, they coincide with the anode/cathode terminology during discharge of the battery only. According to the thermodynamic definition the anode drives the oxidation and the cathode the reduction. Therefore the definition of the electrode changes during a charge-discharge cycle. A negative electrode is always defined by its redox potential relative to its counter electrode (the positive electrode).

16

The combination graphite-LCO is predominantly found in today’s mobile

devices. Increasing demands on battery lifetime, energy density and price,

for instance for electrical vehicles (EV), however, challenge LCO as the

positive electrode material, since cobalt raises safety concerns and is a pricy

element24

. For EV applications LCO is therefore frequently displaced by the

cheaper, more benign lithium iron phosphate (LFP), LiFePO4. The capacity

of transition metal oxide cathodes is yet a major limitation to higher energy

densities (vide infra).

Lithium-ion batteries are assembled in the discharged state, meaning that

the positive electrode contains the entire lithium inventory. On the first

charge the graphite electrode is lithiated. The processes occurring at the in-

terface between the negative electrode and electrolyte during the first lithia-

tion are decisive for the stability of this system: at potentials below 1 V vs.

Li+/Li the formation of a kinetically stable interlayer, the so-called solid-

electrolyte interphase (SEI), protects the anode from irreversible reactions on

consecutive cycles. Unfortunately, upon repeated cycling of the battery

cracking of this layer leads to capacity fading due to continued side reac-

tions. In this respect, the presence of the SEI is both a blessing and a curse to

the stability of the lithium-ion system.

The development of new strategies to prolong the negative electrode life

and to maintain its high capacity by improving the properties of this inter-

phase therefore remains a topic of high interest to the battery community,

even after nearly 25 years since the introduction of the first LIB.

Commonly Used Electrode Materials

This chapter provides a brief overview over the most commonly studied

negative and positive electrode materials used in LIBs.

Negative Electrode Materials

Graphite. It is the most commonly used negative electrode because of its

high durability in combination with carbonate-based electrolytes. The mate-

rial exhibits a theoretical specific capacity of 372 Ah/kg and charge density

of 837 Ah/L. Besides titanate-based anodes this is the lowest storage capaci-

ty among the common negative electrode materials. The average delithiation

potential lies at 0.125 V vs. Li+/Li and is close to that of neat lithium, which

is why graphite is attractive for high energy density batteries.

The insertion of lithium into this material, the so-called intercalation (lat.

intercalare – insert (day or month) in the calendar), is a phenomena ob-

served frequently for layered structures, particularly graphite and transition

metal dichalcogenides25

(e.g. TiS2). Atoms, ions or small molecules squeeze

into the interplanar space of two neighboring graphene layers. The reversible

electrochemical intercalation of lithium into graphite was first reported by

Yazami and Touzain in 198326

. The intercalation occurs in a periodically

17

repeating stacking pattern of occupied and unoccupied layers. This process is

hence related to topotactical reactions in a given (crystalline)

superstructure27

. The range of known graphite intercalation compounds is

vast, Brönsted acids, such as H2SO4, HNO3 or HClO4 and Lewis acids, such

as FeCl3 or AlCl3 intercalate as donor-acceptor complexes, [Cx]+[X]

-. The

most prominent intercalants are the alkali metals Li, K, Rb and Cs. Metallic

intercalants are commonly referred to as donor compounds as they transfer

electrons to the graphite imparting the characteristic metallic and optical

properties (e.g. KC8 is bronze-colored). The rarely achieved limiting stoichi-

ometry is MC6 (M= Li, Sr, Ca, Ba, Eu, Yb) and typically requires elevated

temperatures for most compounds28

. Lithium ions on the other hand interca-

lates reversibly at ambient temperatures. Hence, graphite represents an ideal

storage material29

. As shown recently, intercalation of Na becomes thermo-

dynamically favorable by cointercalation with ethers30

.

Lithium intercalation proceeds primarily at the prismatic surfaces (edge-

planes)31

,via phase transitions between the fixed lithium-to-carbon ratios

LixC6 (x = 0.22 (III), 0.33 (II L), 0.5 (II), 1 (I)), which are commonly re-

ferred to as stages (stage indices indicated in roman numerals)32

. Each stage

is uniquely characterized by its stoichiometry, crystal structure and electro-

chemical potential29

. Under galvanostatic control the different stages of lithi-

um intercalation appear as flat plateaux in the chronopotentiogram (Fig. 2),

due to the presence of coexisting phases during a phase transition33

.

Figure 2. Structure of lithiated graphite (left) and electrochemical-related changes at the electrode surface during lithiation and delithiation of a graphite electrode.

Silicon (Si). By virtue of its high theoretical gravimetric capacity of 3589

Ah/kg and volumetric capacity of 2194 Ah/L (Li15Si4) lithium-silicon alloys

received much attention lately17,34

. The average delithiation voltage of silicon

is 0.4 V vs. Li+/Li, which is higher than for graphite

17. However, in terms of

energy density the gain in capacity outweighs the smaller voltage window17

.

The major difficulty with alloying materials, such as silicon, is their rapid

capacity fading as a result of their tremendous volume expansion during

lithiation. In case of Li-Si the volume expansion is about 280 % (Li15Si4).

18

This leads to rapid electrode disintegration as a result of the particle redistri-

butions during expansion and contraction35–37

. In course of the particle rear-

rangements the electrode wiring in the electrode is compromised38

and parti-

cles are rendered inactive on consecutive cycles39

. Even more troublesome is

the formation of a stable SEI. Even at moderate degrees of lithiation the in-

terphase cracks repeatedly under the induced stress, thus promoting contin-

ued electrolyte decomposition (causing low Coulombic efficiencies). Since

the SEI keeps growing the resistance at the electrode-electrolyte interface

increases, impeding lithium ion diffusion to the surface and hence the elec-

tron-transfer process38,39

. In summary both the electronic and ionic wiring in

a composite electrode decrease significantly due to volume expansion, which

ultimately decreases the number of particles that take part in the electro-

chemical reaction40

.

Strategies to mitigate the effects of volume expansion include buffering

the expansion by reducing the total amount of active material in the compo-

site electrode35

, limiting the total amount of lithium in the alloy (capacity

limitation)41

, the use of nanosized materials, particularly with tailored di-

mensions42

, electrode binders43–46

and electrolyte additives47

for improved

SEI formation. The combination of these approaches now allows extended

cycle life for silicon electrodes over more than 1000 cycles in half cell setups

with ‘unlimited’ lithium supply.

Lithium titanate (Li4Ti5O12, LTO). As for most transition metal oxides

(see below) its capacity is comparatively low (179 Ah/g & 615 Ah/L). Ow-

ing to the high operation voltage of 1.55 V vs. Li+/Li the LTO electrodes are

inherently safer and operate outside the range of electrolyte decomposition.

Unfortunately the energy density is thereby dramatically reduced19,48

. The

absence of a SEI in the ‘classical’ sense49

and its spinel-type structure

(Li[Li1/3Ti5/3]O4), which allows rapid solid state Li+-diffusion, lead to supe-

rior rate-capability19,50

. LTO-LFP full cells cycling reversibly for thousands

of cycles have been demonstrated51

. In primary lithium batteries, LTO is

even used as a positive electrode versus lithium metal. One way to restore a

high energy density involves the use of so-called high-voltage cathodes, i.e.

materials with operation voltages close to or higher than 5 V. This approach,

however, poses challenges to other cell components such as the electrolyte.

Positive Electrode Materials

In general, positive electrode materials are devided in three categories: 1)

inorganic materials, ii) small organic molecules52

and iii) polymers53,54

. The

most commonly used types of materials in LIBs so far are transition metal

oxides and chalcogenides of the general form LixMaXb (M=transition metal,

X=O,S,Se)32,55

. These materials provide a host structure into which guest

ions can be inserted/extracted reversibly (topotactic reaction). Typically, one

or more moles of alkaline ions per mole of MaXb can be stored at ambient

temperatures, yielding specific capacities of 100-200 mAh/g for most struc-

19

tures56

. The crystallographic structure changes to only limited degrees during

this process.

Three major families of structures (Figure 3) can be distinguished based

on their crystallographic characteristics that control the electronic conductiv-

ity (density of states and positions of the valence bands)56

and ionic transport

(ion diffusion framework host (1D, 2D or 3D) or a skeleton structure

(Goodenough et al.57

)48

: 1) layered oxides56,58

, 2) olivine compounds59–61

and

3) spinels62–64

. References19,51,56,65

provide detailed discussions on this topic.

Figure 3. Crystal structures of the three most commonly used metal oxide-based positive electrode materials for LIBs. Lithium ions are indicated as orange spheres.

Electrolyte components and the Solid-Electrolyte Interface (SEI)

Unlike water-based electrolytes, non-aqueous electrolytes act mainly as

Lewis bases. Hence, they provide high solubility for cations but only poor

affinity to anions, resulting in ion pairs or aggregates66

. Xu67

and Goode-

nough et al.56

summarized the properties of an ideal electrolyte as follows:

A large dielectric constant or strong donor-acceptor interactions yield a

high solubility of the lithium salt (1). The electrolyte should be electronically

insulating (2) and exhibit ionic conductivities of > 10-4

S/cm (3). (Electro-

)Chemical inertness (wide stability window) is required (4). The physical

properties should be preserved within a wide temperature range (e.g. viscosi-

ty, solvation ability, etc.) (5). For safety the flash point should be high and

the materials non-toxic (6). The ability to form a stable SEI is favorable (7).

An effective SEI is thin, exhibits a high ionic conductivity (rapid lithium

ion diffusion across the interface) and is elastic in order to avoid cracking

upon stress. In graphite electrodes the volume expansion due to intercalation

is still small (> 10%) compared to lithium alloys for instance17

. In fact crack-

ing of the passivation layer leads to new irreversible reactions and its refor-

mation can consume a considerable amount of the lithium inventory on the

long term. SEI growth is a kinetically controlled process that depends on the

reactivity of the electrolyte solvents employed (i.e. the rate of reduction)68,69

.

Hence, SEI properties, such as composition, may vary with the rate at which

it is formed. The general spatial composition is thoroughly described in liter-

20

ature32,69,70

and changes gradually from mainly inorganic compounds, such

as Li2CO368

and LiF71,72

, in inner SEI regions (close to the electrode surface)

to mostly organic lithium salts, e.g ROCO2Li73

, in outer regions74,75

.

Figure 4. Suggested electrolyte additives for conventional carbonate-based LIB electrolyte formulations.

The primary SEI forming component in LIBs when using a graphite elec-

trode is ethylene carbonate (EC), due to its high reactivity and the formation

of insoluble reduction products69,73

. However, due to its high melting point

(Tm = 35º C) it has to be blended with other, liquid carbonates with lower

viscosities, e.g. dimethyl- or diethylcarbonate. Another component prone to

decomposition, especially in presence of trace amounts of water, is lithium

hexafluorophosphate, LiPF6 - the most commonly used electrolyte salt.

Compared to other lithium salts, LiPF6 is relatively reactive, as the salt un-

dergoes hydrolysis for instance to form toxic HF76–78

, which raises safety

concerns. Moreover, EC and LiPF6 restrict the temperature range at which

LIBs are able to operate79–81

.

Small amounts of sacrificial electrolyte additives can be effective in form-

ing thinner layers with improved properties. The general principle of most

electrolyte additive is an early decomposition of the compound, prior to any

breakdown of major electrolyte components (“reaction-type additive”82

).

Figure 4 gives an overview of different classes of additives. The

carbonates83–87

, sulfoxides88–92

and sulfones93

are the most prominent exam-

ples. Alternatively, the SEI properties can be improved by surface modifica-

tions94–96

or so-called artificial SEI-layers (generated prior to operation)97–99

.

21

Lithium-Sulfur Batteries

In conventional positive electrode materials the transition metal oxide redox-

centers are “spatially diluted” as immobile positions in the crystal lattice.

The amount of charge stored in these materials is typically limited to a single

electron transfer reaction per formula unit21

. ‘Post-LIB’ systems, such as the

lithium-sulfur battery (LiSB), in which lithium polysulfides (PS) constitute

the redox-active species, are in this context appealing, since they are based

on multielectron transfer reactions with no heavy transition metals involved.

The cell concept of this battery system is completely different to that of

the established Li-ion technology. The overall cell reaction of a LiSBs is:

16 Li ⇆ 16 Li+ + 16 e

-

S8 + 16 Li+ + 16 e

- ⇆ 8 Li2S

The average cell voltage for this reaction is 2.2 V vs. Li+/Li when lithium

metal is employed as the negative electrode. The use of the pure metal is

favored, since alternative electrodes exhibit higher average (de)lithiation

potentials, which would further reduce the overall cell voltage.

The most distinct difference compared to LIBs is the chemistry that takes

place in the electrolyte, or more precisely at the electrode-electrolyte inter-

face. PS react at the electrode surface as soluble species. Only the charge and

discharge products, S8 and Li2S respectively, are insoluble in the electrolyte.

PS are thermodynamically unstable, unlike their sodium congeners100

. As

dissolved species they react in a complicated cascade of disproportionation

and (electrode confinded) electron-transfer reactions. The high solubility

also causes diffusion of PS out of the electrode into the bulk electrolyte and

eventually to the counter electrode. Side reactions at the counter electrode

can thus trigger an internal short-circuit known as the polysulfide redox shut-

tle. This adds a high degree of complexity to the understanding of this sys-

tem and compromises its reversibility (vide infra).

Sulfur-Carbon Composite Electrodes

Sulfur positive electrodes generally comprise a blend of sulfur, carbon and

binder. Because of the insulating nature of S8 and Li2S, between 20 to 30

wt.% of carbon is needed in order to make the composite sufficiently elec-

tronically conductive. The cell reaction occurs at the carbon-electrolyte in-

terface and high surface areas are therefore beneficial. Especially at the end

of a (dis)charge cycle, when insoluble products precipitate thin product lay-

ers may inhibit further electron-transfer reactions. Sulfur can be infiltrated

into highly porous carbons by solution or melt infiltration or freeze-

drying101

prior to electrode preparation,. Li2S electrodes102

, used for instance

as cathodes in lithium-ion sulfur batteries, can be obtained either by fabrica-

tion of Li2S/carbon composites in a moisture free environment103,104

or con-

22

verting sulfur chemically into Li2S prior to cell assembly105,106

. Li2S/C com-

posites can also be prepared directly from Li2SO4, Na2CO3 and an organic

precursor by a Leblanc-type process107

.

In 2009 the Nazar group reported for the first time confinement of PS in a

porous carbon matrix during battery operation108

, which can curtail the redox

shuttle substantially. Other examples quickly followed109–112

, triggering the

development of a rich variety of tailor-made meso- and microporous carbon

host architechtures113

. Hierarchical pore structures facilitate ion transport,

increase the electrode surface area and enable PS confinement in the

pores112,114,115

. More recently it was found that commercial carbons can com-

pete with these advanced nanomaterials101,116,117

. The adsorptivity of PS

within materials is generally a function of surface area, but is drastically

influenced by the surface polarity of the material. Hydrophobic carbons for

instance show little adsorption compared to polar (nano-sized) metal oxides

(e.g. TiO2 or MnO2)118

. Another strategy is to tailor the binder properties, so

that the polymer is located directly at the electrode interface where the cell

reaction occurs, which generates a specific, local electrolyte environment for

PS intermediates and their insoluble (dis)charge products119,120

.

Cell Chemistry and Analysis of Lithium-Sulfur Batteries

The reaction pathway of LiSB is generally very sensitive to the electrolyte

properties. PS react with carbonate solvents commonly used in LIBs121

.

Among the aprotic alternatives the glyme family (e.g. 1,2-dimethoxyethane

(DME) or tetraglyme (TEGDME)) and 1,3-dioxolane (DOL) are today the

most commonly used electrolyte solvents and these are sufficiently inert in

the given voltage range122

. Efforts in understanding the cell reactions date

back to the late 70s and 80s. Behind the simplistic two-plateau voltage pro-

file of a typical LiSB stands a complex cascade of various electron-transfer

and disproportionation reactions, as is illustrated in Figure 5. In a standard

ether-based electrolyte, i.e. DME:DOL containing lithium bis(trifluoro-

methane sulfonimide) (LiTFSI), the main intermediates are likely S62-

and

S42-

. The latter disproportionates into the final discharge product Li2S123

. On

top of that, different intermediate species can be favored depending on the

donor number of the solvent124–128

(Gutmann129

). A good example is the S3*-

-

radical124,130–132

.

A vast variety of different analytical techniques has been applied in order

to investigate the complex cell reaction, including ex-situ and in-situ X-ray

diffraction (XRD)116

, X-ray absorption near-edge spectroscopy

(XAS/XANES)133

, electron paramagnetic spin resonance (ESR/EPR)134,135

,

light absorption spectroscopy134,136,137

and Raman spectroscopy138

. Electro-

chemical methods, such as cyclic voltammetry123,139

and rotating ring disc

experiments140

, have been carried out in-situ and allow investigations on a

broad range of time scales.

23

Figure 5. Reaction scheme (left) of sulfur and polysulfide intermediates at the posi-tive electrode-electrolyte interface and the resulting voltage profile (right).

Self-discharge and the Redox Shuttle Mechanism

The diffusion of active material into the electrolyte is responsible for rapid

capacity fading and self-discharge in LiSBs. Naturally a soluble compound

will dissolve until saturation is reached. In case of the ether-based electrolyte

system the saturation limit for PS is relatively high (several mol/L). Redox

active species will move too far away from the positive electrode and be-

come inaccessible for the ongoing cell reaction. At the interface of the nega-

tive electrode, PS are reduced to short-chain PS or precipitate as insoluble

lithium sulfides, Li2Sx=1,2. Soluble reduction products can diffuse back to the

positive electrode and react with sulfur or long-chain PS. This phenomenon

is referred to as the ‘redox shuttle mechanism’ and constitutes the origin of

rapid self-discharge seen in this system. The lithium corrosion rate is highest

for the high-order PS, indicating a faster diffusion for these species126

. A

detailed mathematical description of the shuttle parameters has been given

by Mikhaylik et al.141

. Self-discharge tests have shown a rapid transition in

open-circuit potential (OCV) from the upper to the lower discharge plateau

within a couple of days142

. Consequently, various strategies have been re-

ported to either prevent PS dissolution by careful selection of the

electrolyte115,143–145

or inhibition of PS diffusion by the use of

interlayers146,147

and ion-selective separators148

.

Figure 6. Schematic polysulfide redox shuttle process in a LiSB.

24

Lithium Metal Negative Electrodes

Lithium metal is the negative electrode of choice in order to maximize the

comparatively low cell potential and thus energy density16

. So far lithium

metal electrodes have only been commercialized in lithium polymer batter-

ies, employing a solid polymer electrolyte12

. The electrode experiences se-

vere morphological changes. Because of the high reactivity of Li towards

almost any liquid electrolyte and the related, severe SEI formation, the metal

is (re)depositing unevenly, which may cause growth of lithium dendrites that

raise serious safety concerns (as in LIBs)12,145

. Exposure of pristine lithium

deposits on each cycle result in continued irreversible reactions and loss of

the lithium inventory. Hence, the roundtrip efficiency (Coulombic efficien-

cy, C.E.) is low and the cycle life short. Additional complexity comes from

dissolved PS and the redox shuttle in the case of Li-S batteries22

.

It has been shown that lithium dendrite growth is reduced when “modest

pressures” are applied149

. Recently, caesium and rubidium ions as electrolyte

additives have been reported to promote homogenous plating and mitigate

dendrite growth150

. Dendrite suppression could potentially be achieved by

the use of inorganic electrolytes as well145

, for instance lithium polysulfideo-

phosphates, Li3PS4+n151

, and Li2S-P2S5 glass-ceramics152–154

. Monte Carlo

simulations predicted a profound impact on the depth of discharge (DoD) on

the surface morphology. Cycling only small fractions of the lithium elec-

trode (high Li excess; low DoD) or cycling lithium at 100 % DoD (requires a

current collector) could improve the cycle life. Unfortunately, both options

come at the expense of dramatic losses in energy density149

. Lastly, Sion

Power introduced the electrolyte additive LiNO3, which reduces the rate of

self-discharge in Li-S cells and the internal redox shuttle, i.e. parasitic reac-

tions, tremendously149,155

. Nitrates likely react with PS to form a passivation

layer on the Li metal156,157

. As a result, the additive is consumed over time117

.

For these reasons replacement of lithium metal is discussed158

. Unfortu-

nately, alternative negative electrodes require lithiation prior to use or sulfur

has to be replaced by Li2S in the positive electrode. Owing to their high spe-

cific capacity silicon106,159,160

and tin161–163

are possible candidates. A less

viable solution is the use of hard carbons159

or graphite164–166

.

25

Functional Polymers as Electrode Binders

Polymers are a versatile class of materials that penetrate our society in eve-

ryday life. Like no other class of materials polymers represent functionality

and flexibility because of easy processing in nearly any form and shape de-

sired and the ability to tailor properties by choosing from a vast variety of

readily available building blocks. Most polymers are of organic nature which

limits their applications to temperature ranges between -100 and +300° C,

which is the relevant range for applications in commodity products.

Classic polymers, such as polyethylene (PE) or poly(ethylene terephtha-

lat) (PET), are well-known from plastic bags and bottles. The majority of

polymers, however, remain shapeless and unrecognized entities, yet they

take an active role in providing the physical or chemical properties desired.

This subgroup is commonly referred to as functional polymers. They are

recognized by their effect on a system or their specific role in a system rather

than their appearance as an (physical) object – frequently invisible and in-

conspicuous167

. The physical, chemical, biological and pharmacological

properties of a functional polymer depend on the (chemical) nature of their

constituents – the monomers168

. They are commonly classified as 1) anionic,

2) cationic, 3) neutral, hydrophilic and 4) neutral hydrophobic167

, depending

on which moieties the monomers carry. An overview of monomers for func-

tional polymers used in this thesis is provided in Figure 7.

Figure 7. Overview of different categories of monomers for functional polymers.

26

Electrode Binders

LIB electrodes are typically cast onto a (metal) current collector from slur-

ries that contain the blended electrode components. The resulting composites

are scaled up easily, enabling electrode fabrication at low costs. To improve

adhesion and mechanical strength of the electrode coatings on the current

collector a binder (or a blend of several binders) is added to the slurry.

Binders are high-molecular weight polymers with molecular weights of

several hundred thousand to several million grams per mole and with good

solubilities in the dispersion medium of the slurry. Moreover, binders are

surfactants that should assist in homogeneous mixing of the electrode com-

ponents. Since binders add dead-weight to a battery, their amount should be

reduced as much as possible. Until recently binders have been mostly recog-

nized for imparting composites better mechanical properties. With the ad-

vent of increased alloy anode research their beneficial impact on the surface

chemistry at the electrode-electrolyte interface has been acknowledged46

.

This chapter briefly reviews the development of functional binders for posi-

tive and negative electrodes.

The “Gold Standard” – Poly(vinylidene difluoride)

The traditional binder for LIB electrodes is poly(vinylidene difluoride)

(PVdF) and copolymers of the vinylidene difluoride (VdF) and hexafluoro-

propene (HFP) monomers, so-called PVdF-HFP (Kynar or KynarFlex), with

refined mechanical properties.

It is argued that its chemical inertness, adhesion strength to particles and

the current collector are the main advantages of PVdF169

. However, a closer

look at the polymer properties shows that the polymer swells extensively in

the presence of the electrolyte, thus loosing much of its mechanical

benefits170

. In fact, due to its good swellability PVdF(-HFP) polymers can be

employed as gel polymer electrolytes in LIBs171,172

. As a consequence the

adhesion properties are affected adversely as well173

, unless the polymer is

further modified174

. Electrode fabrication generally requires dissolution in

toxic and high-boiling solvents, like N-methyl-2-pyrrolidone (NMP) causing

additional expenses and safety issues during the fabrication of electrode

coatings from slurries and waste disposal175,176

. Furthermore, the stability of

the fluorinated polymer towards reducing agents177

leads to further safety

concerns in the event of a thermal runaway as a result of a formation of LiF

and the possible release of HF178,179

. Lastly, PVdF-based polymers are rela-

tively costly (15-18 EUR/kg)176

.

27

Water-based Binders for Negative Electrodes

There are three established binder alternatives in negative LIB electrodes:

cellulose-based polymers, particularly the sodium salt of carboxymethyl

cellulose (CMC-Na), poly(acrylic acid) (PAA) and its sodium salt

poly(sodium acrylate) (PAA-Na) as well as the latex styrene-butadiene rub-

ber (SBR). The chemical structures of these binders are depicted in Figure 8.

These binders are water-processable and thus environmentally more benign

than PVdF(-HFP) and, as they are industrially relevant chemicals180

, relative-

ly inexpensive (e.g. CMC-Na: 1-2 EUR/kg176

).

Figure 8. Chemical structures of the four most commonly used binders for negative electrodes.

CMC-Na was introduced by Drofenik et al. in 2003 as a part of a broader

study on surface modified graphite particles involving a range of cellulose-

based binders and gelatin94,95,181,182

. Their finding showed that small amounts

of gelatin or CMC-Na binder (a few wt.%) can suffice in order to obtain

superior cycling stabilities with reduced irreversible capacity losses. The

authors concluded that electrode fabrication with water-soluble binders leads

to a binder coating on the graphite particles95

and favorable SEI growth dur-

ing cell cycling94,182

. Lee et al. used CMC-Na in small amounts, as a thick-

ener, in order to adjust the viscosity of slurries containing SBR and natural

graphite183

. This function was ascribed to the pH-dependent swellability of

the binder, which is, incidentally, highest in deionized water. Hence, CMC-

Na provides a networking effect that promotes homogenous particle disper-

sion and the formation of a dense electrode microstructure. A later study by

some of the authors also showed that the adsorption of CMC-Na on the

graphite surface is out-competed in presence of PAA in the same slurry184

.

Unlike CMC-Na, PAA has little effect on the viscosity of the slurry. Binder

mixtures of PAA and CMC-Na showed better electrochemical performance.

A series of related studies by Komaba and coworkers stressed the superior

stability of graphite electrodes processed with PAA-based binders185–187

.

Deprotonation with alkalimetal hydroxides yields the corresponding poly-

28

acrylate. It was shown that the presence of alkali ions at the electrode inter-

face can be beneficial96,188,189

.

Even though PAA and CMC-Na are macroscopically very rigid and brit-

tle (high Young’s modulus)170

, a single polymer strand appears to be much

more resilient. Especially when the polymer is covalently tethered to the

particle surface40,190,191

. Hence, these polymers are more capable of accom-

modating large volume changes in alloy composite electrodes. Moreover,

favorable SEI growth is expected with these materials170,187,192

. This devel-

opment has initiated increased interest in exploring novel electrode

binders44,45,193,194

and the synthesis of tailor-made binder materials195–197

.

Binders for Novel Positive Electrodes

Many of the inorganic transition metal oxides are prone to undergo reactions

when processed in water. Therefore the purely water-based binders CMC-Na

and SBR can be difficult to implement. Luckily, the LFP electrode material

does not show signs of degradation during electrode fabrication using water-

based slurries176

. PAA is soluble in a variety of alcohols and exhibits a

strong attraction for positive electrodes as well.

As shown above the cell chemistry of novel positive electrodes for Li-S

and Li-oxygen batteries can be considerably more complex than those used

in conventional LIBs. This opens new possibilities for functional polymers

to intervene in the charge and discharge reaction respectively. One example

is the use of poly(ethylene oxide) (PEO) as binder for sulfur-carbon compo-

site electrodes. In such electrodes the polyglycol increases the solubility of

the reaction intermediates, which leads to an increased discharge capacity

and improved reaction kinetics198,199

. In other words the polymer alters the

local electrolyte environment in the electrode119

. Studies involving tethering

glycols onto carbon hosts108

or casting polymer coatings on top of elec-

trodes200

have shown similar effects.

Another study on Li-oxygen batteries shows the profound impact of the

donor number of the electrolyte solvent on the discharge reaction201

. Similar

observations were reported very recently in a mechanistic study on Li-S

cells202

. These solvents could potentially help to master the cell chemistry in

these systems. However, their application leads to an increased reactivity on

the negative lithium electrode. An electrode binder on the other hand re-

mains locally confined and established electrolyte mixtures can hence be

applied while exploiting the beneficial effects of these compounds.

29

Scope of this Thesis

In comparison to the classical picture of the binder being a largely inactive

component in composite electrodes, these polymers can in fact actively in-

fluence or even control the interfacial processes at the electrode-electrolyte

interface. This work investigates the role of some functional binders at the

electrode interface of negative electrodes of lithium-ion batteries (LIBs) and

positive electrodes of lithium-sulfur batteries (LiSBs).

Pore-blocking of PVdF-based binders (Paper II)

The complexity of lithium-sulfur batteries lies in the cascade reactions of

soluble sulfur species at the positive electrode. The disadvantages of the

established poly(vinylidene difluoride)-based binders were investigated by

galvanostatic cycling and nitrogen adsorption of PVdF(-HFP):carbon:sulfur

composites.

Water-based functional binders for the lithium-sulfur system (Paper I)

Alternative, functional polymers were explored that interact favorably with

intermediate species. New electrode binder formulations were examined

mainly by a combination of galvanostatic cycling and self-discharge tests.

Binders as exfoliation suppressants in graphite electrodes (Paper III)

In negative electrodes binders should provide good mechanical properties

and ideally assist in the formation of a stable solid-electrolyte interphase.

Poly(vinylidene difluoride), carboxymethyl cellulose sodium salt, styrene-

butadiene rubber and poly(sodium acrylate) were studied as binders for

graphite electrodes. The characterization of these electrodes was carried out

mainly in a half cell setup, using galvanostatic cycling and electrochemical

impedance spectroscopy. Morphological changes were investigated by sec-

ondary electron microscopy and energy-dispersive X-ray spectroscopy.

A stable graphite-sulfur battery (Paper IV)

In an attempt to replace the problematic lithium anode in LiSBs, a stable

graphite electrode was examined in a lithium-ion-sulfur setup, where the

graphite electrode was lithiated prior to its cycling against the sulfur positive

electrode.

30

Results and Discussion

Binders for Sulfur-Carbon-Composite Electrodes

The following chapter discusses investigations related to sulfur composite

electrodes for lithium-sulfur batteries and describes the replacement of the

most commonly applied binder material, poly(vinylidene difluoride), by a

blend of poly(ethylene oxide) and poly(vinylpyrrolidone) which provides

superior properties in terms of self-discharge and capacity retention.

Pore-Blocking Effect of PVdF-based Binders in Composite

Electrodes

Poly(vinylidene difluoride) (PVdF) and poly(vinylidene difluoride-

hexafluoropropane) (PVdF-HFP) are chemically and from a structural point

of view very similar polymers. Therefore little distinction is made in the

literature between these two materials. Together, they make up the largest

share of commonly used electrode binders. The presence of HFP in the co-

polymer makes the material less crystalline and somewhat easier to dissolve

in organic solvents172

. In Paper II, sulfur composite electrodes were pre-

pared by ball-milling a slurry comprising of either PVdF or PVdF-HFP,

sulfur and a porous carbon. A comparatively large amount of binder (15

wt.%) was used in order to exaggerate the differences in the physical and

electrochemical properties of these composites.

A simple solubility test was conducted in which the binders were dis-

solved in a standard lithium-sulfur electrolyte comprising DME:DOL (v/v =

1:1), 1 M LiTFSI and 0.25 M LiNO3. As shown in Figure 10, PVdF-HFP

yielded a homogenous PVdF-HFP:electrolyte gel, whereas a turbid gel was

obtained using PVdF, presumably due to the higher crystallinity of the pol-

ymer171,203,204

.

Figure 9. Swelling test of PVdF and PVdF-HFP in 1 M LiTFSI, 0.25 M LiNO3 and DME:DOL (v/v = 1:1) after gentle heating at 50° C.

31

A series of physisorption experiments (see Figure 11) was carried out with

carbon:binder (25:15) and carbon:sulfur:binder composites. In the latter case

sulfur was melt-infiltrated into the pores of the carbon host prior to the prep-

aration of the composite.

Figure 10. Differential (a, c) and cumulative (b, d) pore volumes obtained from physisorption on the pristine carbon, binder-infiltrated carbon and melt-infiltrated carbon. The latter was treated with a solution of the binder after sulfur infiltration.

The BET surface area of the mesopores in carbon:binder composites was

significant decreased after the carbon was immersed into the binder solu-

tions. The pore volume was determined by a single point measurement close

to ambient pressures (p/p0 ~ 0.975; corresponding to pore widths between 67

and 86 nm). The micropore surface area and pore volume were obtained by

the t-plot method at relative pressures below 0.2 p/p0 and disappeared com-

pletely in all cases after infiltration with both sulfur and binder. The differ-

ence of 56 m2/g between PVdF and PVdF-HFP composites was compara-

tively small considering the overall surface area of the carbon (< 5%) and

may originate from different surface wetting or penetration into the pores as

a result of different radii of gyration in NMP. Furthermore, the binder itself

contributed to the overall surface area in the infiltrated composites. It is pos-

sible that the polymers exhibit a different porosity when they crystallize in

the pores upon drying. Different strength of adsorption between nitrogen and

the polymer may cause small deviations between the two materials as well.

When using melt-infiltrated sulfur-carbon composites nearly of the pore

32

volume is occupied by sulfur and binder. In this context it should be noted

that it was reported recently that sulfur redistributes in porous carbons at low

relative pressures205

. However, based on the previous physisorption results

for carbon:binder composites, it is clear that any sulfur infiltration will fur-

ther reduce the porosity of the carbon host. Since sulfur infiltration is con-

ducted prior to the composite preparation (binder infiltration), only the re-

maining pore volume can be occupied by the binder.

Figure 11. Voltage profiles of melt-infiltrated sulfur-carbon composites with PVdF (top), PVdF-HFP (middle) and no binder (bottom) for the first three discharge cycles (cycle numbers indicated in the plots). The electrodes were cycled galvanostatically at a rate of C/20.

The voltage profiles of the corresponding composite electrodes cycled at a

cycling rate of C/20 are shown in Figure 12. Generally, PVdF-based elec-

trodes displayed lower discharge capacities than PVdF-HFP-containing elec-

trodes. The general trend was the same for both polymers though, the dis-

charge capacity dropped on the second cycle and started to increase again on

the third cycle. The peak plateaux shift to lower potentials between the first

and the third cycle, indicated the build-up of an overpotential. Hence, even

electrode binders with very similar chemical structure may induce different

electrochemical behavior. In order to further highlight this finding, an elec-

trode containing no binder at all (infiltrated carbon only) was also studied. In

33

contrast to the PVdF-based electrodes, the binder-free electrode exhibited

constant plateau potentials and higher capacities. The results suggest a pore

clogging effect of the electrodes fabricated with PVdF-based binders in a

standard ether electrolyte. This interpretation is in accordance with earlier

findings by Younesi et al., where PVdF was found to reduce the pore vol-

ume of porous carbons significantly when applied as electrode binder in

lithium-oxygen batteries, with detrimental effects on the capacity206

. In the

present study it was found that poorly swellable binders might trap PS inside

the pores of the carbon host and compromise the discharge capacity. Over

time the sulfur utilization improves again, since the binder is rendered more

permeable to electrolyte due to swelling. This could result in higher local

viscosities and the built-up of the observed overpotential.

Water-Soluble Functional Binders for Sulfur Electrodes

The previous results indicate that a binder-free electrode would allow facile

communication between components confined in the pores and in the bulk

electrolyte. In terms of fabrication, binder-free composite electrodes are

hardly a solution, since the mechanical properties and adhesion to the current

collector of such electrodes are poor. A binder that swells readily or at least

does not impede the cell chemistry is a compromise. One example is a water

processable binder mixture of carboxymethyl cellulose sodium salt (CMC-

Na) and styrene-butadiene rubber (SBR), which was used as a reference

system in Paper I. If binders at the same time fulfill a certain function in the

composite, they become useful additives in a composite electrode and repre-

sent a straightforward approach to improved battery performance. For in-

stance, earlier works indicated that PEO binder119

or glycols as electrolyte

additives increase the initial discharge capacity108,198

. Further, Seh et al. re-

ported in a recent study using DFT calculation that the interaction strength

between binder and PS intermediates depend on the nature of the functional

group. The binding energy of PS to amide groups was found to be about

twice the energy than that of PS to fluoride in fluoroalkanes. In order to

highlight this result the authors showed superior dispersion properties of

carbon black:Li2S:binder mixtures using poly(N-vinylpyrrolidone) (PVP)103

.

Three cathode binders (10 wt.% PEO, 10 wt.% PVP and PEO:PVP, w/w

= 4:1) were examined by galvanostatic cycling in comparison to a CMC-

Na:SBR-based electrode (Figure 13). The discharge profile of the PEO sam-

ple (10 wt.%) showed increased capacities initially but then faded rapidly the

next 50 cycles, in accordance with previous observations119

. The capacity of

PVP electrodes dropped more rapidly at the beginning but displayed better

capacity retention during prolonged cycling. An electrode containing a blend

of PEO:PVP yielded superior behavior and both in terms of capacity reten-

tion and discharge capacities.

34

Figure 12. Galvanostatic cycling of sulfur positive electrodes containing 10 wt.% binder (PEO, PVP, PEO:PVP or CMC:SBR), 50 wt.% sulfur and 40 wt.% carbon. A mixture of CMC-Na and SBR in a ratio of w/w = 2:3, was used in a reference cell.

The function of PVP can be visualized in electrolyte:binder solutions (both

PEO and PVP are soluble in the electrolyte): when a PS-electrolyte solution

is added, precipitation of a red solid is observed in presence of PVP and the

solution loses its color (Figure 14). The solid was not stable in dry state.

Exposure to air led to rapid decomposition into a yellow solid, indicating

disproportionation of PS to sulfur and lithium sulfide. PVP is known for its

complex formation with polyanions, e.g. with triiodite (I3-), and can be used

as an antiseptic207

. The insoluble product observed might be a result of ionic

cross-linking208,209

.

The use of a PEO:PVP blend leads to a symbiotic effect in which the

overall capacity is increased and the capacity fade decreased in comparison

to electrodes containing either only one of the two materials. In fact, a 4:1

mixture of this blend has been investigated previously at Hydro-Québec210

,

but the authors did not comment on the choice of their binder system.

To further show the retention effect of the PEO:PVP mixture the cell was

stopped after 5 discharge-charge cycles and the open-circuit potential (OCV)

was recorded as a function of resting time. Pronounced self-discharge, indi-

cated by a voltage drop at about 2.3 V to 2.15 V (corresponding to the upper

and lower discharge plateau in a conventional discharge cycle, respectively),

was found after about 2 days for the CMC-Na:SBR reference cell. Ryu et al.

ascribed the voltage drop to increased dissolution of PS as a result of a self-

accelerating process in which elemental sulfur dissolves to a small degree in

35

the electrolyte142

. A spontaneous ring-opening reaction generates high-order

PS, which in turn dissolves more sulfur and initiates the redox shuttle be-

tween the electrodes. The time it takes for a notable voltage drop to be seen

is therefore significantly prolonged in PEO:PVP electrodes (6-7 days).

Moreover, the cell equilibrates at an OCV of 2.4 V instead of 2.3V as in case

of CMC-Na:SBR and PEO electrodes.

Figure 13. Formation of a red insoluble precipitate (right) upon the addition of a polysulfide-electrolyte solution to a solution of PVP and electrolyte.

Figure 14. Self-discharge test of three sulfur electrodes containing a different binder (CMC:SBR, PEO or PEO:PVP). The cells were cycled for 5 galvanostatic cycles at C/5 and then allowed to rest at open-circuit potential for 7 days.

36

Binders for the Graphite Electrode

In the second part of this thesis, functional binders are employed for graphite

electrodes in LIBs and LiSBs. The stability of electrodes in ‘aggressive’

solvents, i.e. propylene carbonate and ethers, in dependence of the polymer

properties is investigated

Binders as Surfactants and Exfoliation Suppressants

Four binders CMC-Na, SBR, PVdF-HFP and poly(sodium acrylate) (PAA-

Na) (Paper III) were examined in graphite electrodes, containing 10 wt.%

of binder, using either a standard carbonate electrolyte mixture (ethylene

carbonate (EC):diethyl carbonate (DEC) (v/v = 1:1) and 1 M LiPF6 referred

to as LP40) or a propylene carbonate (PC) enriched version (PC:EC:DEC,

(v/v/v = 4:1:1) and 1 M LiPF6). Discharge profiles and Coulmbic efficien-

cies during 20 (LP40) and 50 (PC-rich formulation) cycles, respectively, are

given in Figure 16. As can be seen, all binders performed very stably using

LP40. However, high concentrations of PC led in the case of PVdF-HFP§ to

rapid electrode degradation as a result of co-intercalation and evolution of

gaseous reduction products causing particle cracking69,211,212

. In contrast,

PAA-Na and CMC-Na anodes maintained a stable performance.

Figure 15. Galvanostatic cycling (top row) and corresponding Coulombic Efficien-cies (C.E.) (bottom row) of graphite electrodes cycled in LP40 electrolyte (left col-umn) and a PC-rich formulation (right column). Note that PVdF-HFP containing electrodes were unable to cycle in presence of PC. A PVdF-HFP sample cycled in LP40 was added to plots in the right column for comparison.

§ Note that a discharge profile of a PVdF-HFP sample, cycled in LP40, was added in plots for the PC-rich formulation

37

Figure 16. Secondary-electron image of delithiated electrodes containing PAA-Na (a) and SBR (b), respectively, after 50 galvanostatic cycles.

A cycled PAA-Na and SBR sample were analyzed by SEM (Figure 17).

Graphite particles in the SBR electrode degraded to fan-like objects (about

twice the original size), which can be attributed to severe exfoliation due to

the internal pressures of gaseous reduction products from co-intercalated

solvent molecules. In contrast in the PAA-Na electrode graphite particles

remained unchanged but were covered by a smooth, insulating layer. The

charging effects can be attributed to the insulating salt deposits originating

from SEI formation. The stability of PAA-Na containing electrodes against

PC was previously investigated by the Komaba group187,213

. Cycling in pure

PC, however, was not possible in the present work. As mentioned by the

latter authors, graphite particles below a critical size of 3 µm could contrib-

ute to the stability in this solvent186,214

.

In the following the binder properties were correlated to the electrode per-

formance. Firstly, the degree of swelling was determined for the four poly-

mers (Figure 18). The swelling of CMC-Na might be somewhat overestimat-

ed, given the poor solubility of the cellulose-based polymer in almost any

solvent but water170,215

. Similar trends in swelling have been observed by

Magasinksi et al.170

and Wang et al.216

. Moreover, binder swelling in LP40

has recently been investigated in-operando by electrochemical quartzcrystal

microbalance with dissipation monitoring (EQCM-D). The results suggest

that CMC-Na behaves as a rigid polymer, whereas PVdF shows viscoelastic

behavior215

. Poor swellability in the electrolyte could present a mean to pro-

tect electrolyte components from direct contact with the electrode surface by

preventing their penetration at the interface.

Furthermore, the SEI growth and composition could benefit from the

presence of sodium ions at the interface188,189

, although a contact with the

electrolyte probably leads to rapid ion-exchange187

. The actual amount of

sodium incorporated in the surface layer and the solubility of sodium salts

(in comparison to corresponding lithium salts) remain largely unclear so far.

Therefore XPS analysis was applied to cycled cells (50 galvanostatic cycles

at C/10), indicating the presence of sodium salt deposits in the SEI of the

38

PAA-Na sample (Figure 19; not included in Article III). In contrast the

CMC-Na sample did not show any measureable amount of sodium in its SEI.

This was very likely an effect of sodium concentration in the polymers,

which accounts for about 25 wt.% of the molecular weight of PAA-Na com-

pared to 9 wt.% in CMC-Na (assuming a degree of substitution ~0.8).

Figure 17. Swelling test of various binders in the battery electrolyte LP40. The pol-ymers were cast in a teflon mold and immersed in the solutions for several days. The term degree of swelling refers to the gain in mass relative to the dry state mass.

Figure 18. XPS specta of the alkali metal ions in the SEI for a cycled PAA-Na and CMC-Na graphite electrode samples (Electrolyte: 1 M LiPF6 in EC:DEC:PC, v/v/v = 1:1:4), respectively. Note that the peaks in this plot were normalized to the highest peak intensity and not further calibrated. In the lower panel (Li 1s) the CMC-Na spectrum was aligned manually by -1.2 eV so as to coincide roughly with the Li 1s-peak of the PAA-Na sample.

39

Complementary, impedance spectroscopy experiments were carried out after

10 galvanostatic charge-discharge cycles in order to compare the interfacial

resistances of the electrodes after SEI formation (Figure 20). The impedance

after 6h of charge and discharge, respectively, increased in the order PVdF-

HFP < SBR < PAA-Na < CMC-Na. As expected for a polymer gel the re-

sistances of PVdF-HFP and SBR samples are comparatively low in contrast

to the poorly soluble PAA-Na and CMC-Na. CMC-Na gave rise to the high-

est interfacial resistance in this type of measurement. This is also in agree-

ment with the cycling data, which indicated lower capacities during the first

20 cycles. Despite the possibility that some of the active material was not

accessible at the start of the experiment (leading to a reduced capacity), a

high resistance at the interface could stop the intercalation reaction since the

lower cut-off limit of 5 mV would be reached prematurely. PAA-Na on the

other hand appeared to provide a balance between sufficient protection (in-

dicated by a relatively high interfacial resistance), and lithium ion transport

across the interface. Better transport properties might also originate from a

higher density of functional groups. In this context, Komaba et al. proposed

an ion exchange at the electrode interface, in which Li+ strips off its solva-

tion shell and becomes solvated by functional groups of the polymer, thereby

preventing a direct contact of the solvent with the electrode surface187,192

.

Figure 19. Impedance spectra in the Bode representation. The cells were paused every 2 h on the 11

th cycle (cycling rate C/10) to record an impedance spectrum in

the frequency range of 10 mHz – 200 kHz. The upper row shows the real (Z’, left) and imaginary (Z’’, right) parts of the phase vector against the frequency during lithiation (charge). In the bottom row the corresponding Bode plots during delithia-tion (discharge) are shown.

40

A higher resistance should also be reflected in the rate-capability of the elec-

trodes. In a rate-test with SBR, PAA-Na and CMC-Na containing electrodes,

cycled in LP40 electrolyte (Figure 21) the cells were subjected to 20 cycles

at a rate of C/10 before the rate was increased gradually every 10 cycles to

C/5, C/2, C, 2C and back again. As in the previous experiments, the CMC-

Na electrode showed a low initial capacity, which increased on the following

cycles. At a rate of C/2 all cells encountered difficulties to keep up with the

cycling rate. The capacity profile of the PAA-Na sample displayed fluctuat-

ing, yet high capacities, while SBR and CMC-Na electrodes showed a dis-

tinct drop in capacity at this rate.

Figure 20. Rate-capability tests of graphite electrodes containing different anode binders (CMC-Na (top), PAA-Na (middle) and SBR (bottom)) cycled in LP40.

The reversibility of PAA-Na electrodes on the reversed rate-steps is note-

worthy. Despite the complexity of a rate-capability test the results clearly

confirm PAA-Na as superior electrode binder in terms of reversibility and

cycling stability (the capacity remains constant at a given rate). The data for

the CMC-Na electrode further complement those of the impedance study

showing that the high interfacial resistance found for this electrode was also

reflected in a poorer rate-capability.

Lastly, the effects of lower binder contents were studied. This topic is

rarely addressed in academic studies, where contents of 10 wt.% are typical-

ly used, although such formulations are of little practical use.

41

Samples with various amounts of binder were prepared. The lowest

amount depended on the onset of an “infinite” irreversible reduction plateau

in the voltage profile. Interestingly, all electrodes were capable of cycling in

the PC-rich electrolyte formulation when the binder content was high

enough. This can be nicely shown with a PVdF-HFP sample containing 15

wt.% of binder. For SBR, PAA-Na and CMC-Na distinct differences were

observed in the cyclability of the electrode at reduced binder contents (see

Figure 7 in Paper III). The voltage profiles in Figure 22 show that CMC-Na

still provided sufficient protection even at contents as low as 5 wt.%. In con-

trast, the well-performing PAA-Na binder lost its protective effect at this

concentration.

Figure 21. Voltage profiles of PAA-Na and CMC-Na samples for various binder contents obtained for the first 2800 h of galvanostatic cycling at C/10 (top row). Pristine electrodes containing 5 wt.% and 10 wt.%, respectively, were investigated by EDX mapping (middle rows; sodium (green) and oxygen (blue) emission lines) and SEM (bottom row; SEM superimposed with EDX maps).

This behavior was studied using EDX mapping of CMC-Na and PAA-Na

samples with 5 and 10 wt.% binder contents, respectively, using the K-α

spectral line of sodium and oxygen in order to locate the binder on the elec-

42

trode surface (Figure 22). This method was successfully applied by Chong et

al217

in a previous study. It was found that PAA-Na appeared to accumulate

preferentially on the conductive carbon additive instead of depositing on the

graphite particles, where protection is required. In contrast, in CMC-Na

samples the polymer was distributed more evenly, indicated by the absence

of “hot spots” in the EDX map. Furthermore, spots with increased X-ray

emission and locations of agglomerated conductive additive coincide less

frequently. This indicates that binders with functional groups may act as

surfactants that stabilize slurries during electrode fabrication due to their

adhesion strength to the graphite surface183,184

and high surface areas prefer-

ably**. Battaglia and coworkers observed a similar adsorption effects be-

tween carbon black and PVdF composites218–221

. Moreover, the flexibility of

the polymer needs to be taken into account when the surface coverage of the

active material is discussed. CMC-Na by nature is a “stiff” polymer as a

result of its carbohydrate backbone, whereas PAA(-Na) consists of consider-

ably smaller monomer units with a flexible hydrocarbon backbone. This

might allow PAA(-Na) to attach effectively to a particular surface.

The influence of functional binders on graphite electrodes can thus be

summarized as follows:

Protection of the electrode surface by a resistive, yet ionically con-

ductive ‘artificial’ surface layer. The interfacial resistance can origi-

nate from poor solubility of the polymer in the respective solvent (in

other words, gelling leads to a lower resistance but less protection).

PAA-Na appears to provide a good balance between protection (re-

sistivity) and ion conduction across the interface.

Shielding the electrode from parasitic reactions or processes is a

sensitive function of binder content, i.e. surface coverage. CMC-Na

is more efficient in covering the active materials’ surface and can be

applied in considerably lower amounts than PAA-Na.

Functional binders can be viewed as surfactants that will cover high

surface areas preferably. The surfactant strength strongly depends on

the constitution of the polymer, for instance the density of functional

groups and the flexibility of the polymer chain.

This study aims to build a baseline for the comparison of different elec-

trode binders and provided insights into the protective mechanisms intro-

duced by the binder. As will be shown below, these findings have implica-

tions for other negative electrode materials or cell chemistries as well.

** BET surface areas: SuperP carbon black: 60 m2/g and coarse graphite powder < 10 m2/g

43

Optimization of Conventional Graphite Electrodes

The following example highlights how mechanical properties and electro-

chemical performance can be further tuned by using a mixed binder system.

In the previous study coarse graphite powders with particle sizes of

around 10 µm or more were employed as the main active material. In this

example the size distribution of the graphite was between 5.8 and 7.5 µm

in order to allow more efficient packing of the particles††, yielding denser

coatings. The binder mixture consisted of 2 wt.% CMC-Na and 6 wt.%

PAA-Na to yield an overall binder content of 8 wt.%. Upon drying, CMC-

Na coatings typically led to stress and strain on the metal current collector

(even for the 22 µm thick Cu foil used in this study) and bending of the coat-

ed foil was observed. Mixing CMC-Na with another binder (typically

SBR183,222

) and reducing the CMC-Na content to 2 wt.% can greatly reduce

the internal stress and prevent cracking of the electrode coatings.

Figure 22. Electrochemical performance of an optimized electrode formulation with a mixed binder system (left). The SEM images on the right side show a comparison between the electrode morphologies of the new and the “classic” formulation. The inset below the SEM image illustrates the flexibility of the new electrode.

The flexibility and resilience of the resulting coating is shown in Figure 23.

Apart from the choice of binder(s), a denser coating may help to obtain

higher mechanical stability. Fig. 23 (right) shows SEM images of electrodes

fabricated with coarse and fine graphite powder, respectively). If focus is

placed on the mechanical properties of the coating less binder is presumably

sufficient. However, as shown above, electrodes with higher binder contents

exhibit better electrochemical performance. The discharge profile and corre-

sponding Coulombic efficiency are also given in Fig. 23 (left). The opti-

mized electrode exhibited efficiencies of more than 99.5 % after 20 cycles

and achieved an average C.E. of 99.91 % during the first 100 cycles.

†† Data provided by www.matweb.com (search word: TIMCAL KS6 graphite). Size distribu-tion refers to the ‘d90’ value reported for this product, i.e. 90 % of all particles have a size in the stated range of particle sizes.

44

A Stable Graphite Electrode for the Lithium-Sulfur System

In the last part of this thesis (also described in Paper IV) a graphite elec-

trode with 10 wt.% of PAA-Na binder was tested in an ether-based electro-

lyte, which is commonly used in lithium-sulfur batteries.

For graphite electrodes to operate stably in an ether electrolyte a very ef-

fective SEI is needed. Typically the use of ether solvents leads to rapid exfo-

liation of graphite as a result of co-intercalation211,223,224

. A non-swellable

binder at the electrode surface, however, might be able to prevent solvent

molecules to co-intercalate into the graphite host.

Figure 23. Voltage profiles of graphite-lithium halfcells using the three different electrode binders CMC-Na (top row), PAA-Na (middle row), PVdF-HFP (bottom row) in the ether-based electrolytes DME:DOL (v/v = 1:1) + 0.25 M LiNO3 (middle column) + 1 M LiTFSI and DME:DOL (v/v = 1:1) + 1 M LiTFSI (right column), respectively. The rate of electrode cycling was C/10.

In Figure 24 the voltage profiles of graphite half cells with different anode

binders (CMC-Na, PAA-Na and PVdF-HFP, respectively) are depicted dur-

ing the first 40 h. The electrodes were cycled in a solvent mixture of 1,2-

dimethoxyethane (DME) and 1,3-dioxolane (DOL) (v/v = 1:1) with and

without the electrolyte additive LiNO3, respectively‡‡. It should be noted that

neither the electrode nor the electrolyte was further optimized in this setup.

Solely the PAA-Na sample operated stably upon repeated cycling and only

in the presence of LiNO3. Using this cell chemistry, an average Coulombic

efficiency of 99.4 % was achieved over 50 consecutive galvanostatic cycles

(1st and 2

nd cycle C.E.: 70 % and 93.5 %). PVdF-HFP electrode failed rapid-

ly in the chosen electrolyte, while the CMC-Na sample did not show any

significant lithium intercalation in these two electrolytes. LiNO3 appeared to

play an important role for the stability of the PAA-Na electrode in this par-

‡‡ Bearing in mind that the swellablity of polymers changes in the new electrolyte solvent

45

ticular electrolyte formulation. Aurbach and coworkers reported on the role

of nitroxide intermediates originating from the reduction of nitrate at the

lithium anode in conventional LiSBs156

. In secondary reactions mostly with

PS a more stable interphase is formed225

. A partial reaction of the formed

nitroxide intermediates with functional groups of the binder hence seems

likely. This could also explain why CMC-Na, with a lower density of func-

tional groups, did not exhibit the same protective effect in this electrolyte.

Figure 24. LiSB vs. graphite-sulfur cell. The lithium-sulfur reference cells are shown in the left panel. Thick and thin lithium foil refers to 125 µm and 30 µm thick foils, respectively. The thick foil used to charge the graphite electrode was also combined with a sulfur positive electrode and examined. In the right panel two graphite cells are shown against the thick Li reference. The prelithiation of cell #1 and cell #2 involved 9 and 6 galvanostatic cycles at C/10, respectively.

Lithium-ion-sulfur cells (or graphite-sulfur cells) were prepared by lithiating

the graphite electrodes in a half cell during 6 and 9 cycles, respectively, prior

to the assembly of the final cell. The sulfur positive electrode consisted of a

simple mixture of commercial carbons and a water-based binder mixture of

CMC-Na and SBR (w/w = 2:3). Sulfur was melt-infiltrated into the carbon

host prior to fabrication of the electrode slurry. Reference cells were assem-

bled with lithium metal as the negative electrode using two differently thick

metal foils (‘thin Li’: 30 µm and ‘thick Li’:125 µm). Moreover one reference

cell contained the thick Li-foil used to charge the graphite electrode in the

half cell as the negative electrode. The cycling data is presented in Figure 25.

Both thin and thick Li reference cells as well as the two graphite-sulfur

cells performed similarly in terms of discharge capacity over the first 60

cycles. The graphite electrode charged during 9 cycles exhibited some insta-

bilities during extended cycling. However, the random capacity fluctuations

did not seem to affect the overall cell behavior. In contrast, the reference cell

46

containing the cycled Li-foil from the charging process experiences very

rapid fading. It was reported recently that the morphology and surface com-

position of the SEI changes in LiNO3-containing ether electrolytes in pres-

ence of polysulfides225

. Since the first 6 cycles are performed in a graphite-

lithium half cell setup, no polysulfides are initially present. Moreover, the

porosity of the foil increased upon repeated cycling as a result of inhomoge-

neous Li stripping and plating. When the foil eventually is put into a lithium-

sulfur cell, the SEI on the foil is not sufficient and apparently no stable layer

is able to grow onto it during the following cycles.

Figure 25. Extended galvanostatic cycling of lithium-sulfur reference cells and a lithium-ion-sulfur cell with a graphite negative electrode that was lithiated during 6 cycles prior to use. With exception of the pre-cycled lithium foil (blue) all samples were cycled until no significant capacity remained.

Differences in cycle life and cycling stability become obvious after exceed-

ing 100 galvanostatic cycles (Figure 26). Clearly, the capacity fade of all

cells was significant. The thin lithium foil faded rapidly and failed complete-

ly before reaching 200 cycles, while accelerated degradation of the thick foil

started after ~250 cycles and culminated in cell failure after little more than

300 cycles. The graphite electrode, however, continued to cycle until

stopped at the 800th cycle. In contrast to LIBs, the graphite-sulfur cells repre-

sent balanced systems with a sulfur excess of 20-30 %. Therefore the impact

of irreversible losses on capacity fading is much stronger. For comparison,

Berg et al.16

and Rosenman et al.226

estimated the minimum Li excess of any

lithium metal-based cell design to at least 200-300 %, especially in cells

where several mAh cm-2

are exchanged per cycle226

. Although Li-S cells are

operated with a significant Li excess of > 500 % (thin) and > 2000 % (thick),

respectively, they failed much earlier than the cells with a graphite anode.

Part of the rapid failure of lithium-sulfur batteries is normally attributed to

the small electrolyte volumes of about 6 µL/mgsulfur also used in this

work101,117,227

. The constant morphological changes on the lithium surface

47

will lead to repeated exposure of pristine Li and thus continues electrolyte

decomposition. This also causes the formation of other parasitic and soluble

species149

. This inefficiency of the available Li inventory is reflected in poor

C.E. values of less than 95 % (Fig. 25). Moreover, a pronounced increase in

the internal resistance of the cell as a function of foil thickness was noted

(see the Supporting Information in Paper IV). Anyhow, in order for lithium-

sulfur batteries to exceed the practical energy densities of LIBs, the electro-

lyte volume would still need to be further reduced228

.

A self-discharge test (Figure 27) was carried out in order to examine the

stability of the SEI formed on graphite in the full cell. Previous experiments

in the half cell setup indicated that the PAA-Na electrode performed well in

absence of PS (Figure 1 in Paper IV). Electrodes with poor SEI forming

properties readily react with PS, thus triggering the redox shuttle and hence

self-discharge of the cell. This process can be observed by stopping the cy-

cling and following the evolution of the OCV with time. In conventional

LiSBs, a sudden drop after less than 3 days from a high voltage plateau at

about 2.3 V to a lower plateau at 2.15 V is usually observed, resembling the

voltage profile of the discharge reaction142

. Using a graphite electrode, how-

ever, the self-discharge was slowed down considerably with an extended

upper voltage plateau. The voltage commenced to drop slowly after 10-12

days. Based on these results the electrode interface appears to be inherently

more stable against parasitic reactions upon storage.

Figure 26. Self-discharge test for a lithium-sulfur cell (thick Li foil) and a lithium-ion-sulfur cell with a graphite electrode (lithiated during 6 cycles). The cell cycling was interrupted every three cycles by a resting sequence and the OCV was recorded. The resting period was gradually increased from 0.5 to 14 days.

The example of a lithium-ion-sulfur cell employing a graphite negative elec-

trode highlights the weaknesses accompanied with conventional LiSBs – the

SEI growth on the lithium. This is indicated by the significantly higher C.E.

as well as the reduced self-discharge rate when an alternative anode is em-

ployed and despite the use of small volumes of electrolyte.

48

Conclusions

When Sony introduced the first lithium-ion battery (LIB) containing a graph-

ite negative electrode it was the formation of a stable electrode-electrolyte

interphase imparted by the decomposition products of the carbonate-based

electrolyte solvent that allowed the battery to operate stably69

. It is the inter-

face that provides a long cycle life, safe operation and a stable capacity over

hundreds of cycles. Carbonate solvents contribute to this interface to a great

extent. Later, various carbonate- and sulfuroxide-based electrolyte additives

were shown to improve the SEI growth82

. Until the discovery that the brittle

polymers carboxymethyl cellulose sodium salt (CMC-Na)43,222,229

and

poly(acrylic acid) (PAA)45,170

could extend the cycle life of silicon electrodes

considerably, binders were mostly considered as inactive electrode compo-

nents and their potential as useful additives, located directly at the electrode-

electrolyte interface was largely ignored. In this thesis it is shown that bind-

ers should be seen as immobilized local components of the electrolyte, af-

fecting the ongoing surface reactions, particularly the cell reaction itself and

the formation of a solid-electrolyte interphase (SEI).

The commonly used poly(vinylidene difluoride) (PVdF) binder and related

copolymers exhibit poor properties when used as electrode binders in com-

parison to alternative polymers:

PVdF-based binders require toxic organic solvents during the fabrication

of electrode coatings. In porous materials, applied in so-called ‘post-LIB’

systems, the poorly swellable PVdF binder (ether solvent) can cause pore

clogging, thus leading to reduced capacities as a result of restricted access to

the electrode surface. The large degree of swelling in (LIB) carbonate sol-

vents on the other hand, results in poor mechanical properties and surface

adhesion, especially when binder contents of less than 10 wt.% are applied.

In contrast, the water-processable binders CMC-Na, styrene-butadiene rub-

ber (SBR) and poly(sodium acrylate) (PAA-Na) exhibit smaller degrees of

swelling. CMC-Na and PAA-Na yielded improved adhesion to the current

collector and better mechanical properties of the composite electrode. Due to

their poor swellability in the electrolyte, they gave rise to a higher interfacial

resistanc on cycled graphite electrodes, which greatly improved the stability

in the presence of parasitic electrolyte solvents such as propylene carbonate

(PC). Reducing the amount of ethylene carbonate (EC) in the overall solvent

49

mixture is favorable with respect to low temperature applications. Because

of a dense population of functional (carboxyl) groups these two polymers

can also be viewed as surfactants whose degree of surface coverage is a

function of the surface area. When the binder content is reduced, PAA-Na

shows a preference of covering the conductive additive and thereby loses its

protective properties. In samples containing CMC-Na binder, a much stiffer

polymer due to its carbohydrate backbone, the protection of graphite parti-

cles against PC is maintained for binder contents of at least 5 wt.%.

A graphite electrode with a PAA-Na binder was found to operate stably in an

ether-based electrolyte when LiNO3 is used as electrolyte additive. The elec-

trode served as the negative electrode in a lithium-ion sulfur battery. Higher

Coulombic efficiencies, significantly longer cycle life at similar discharge

capacities and a reduced self-discharge rate indicate the presence of superior

SEI properties on graphite electrodes. Unlike lithium negative electrodes, the

graphite electrode tolerated the use small electrolyte volumes (closer to prac-

tically relevant volumes) much better. Furthermore a balanced lithium-to-

sulfur inventory was achieved, with an excess of 30 % sulfur. In contrast,

lithium metal is commonly used in excess of more than 2000 %, thereby

increasing the cycle life and Coulombic efficiency of the system artificially.

As a representative of a versatile class of alternative electrodes, the example

of a graphite electrode, highlights the disadvantages encountered with the

use of lithium negative electrodes in LiSB.

The positive sulfur electrode in Li-S cells was improved with respect to ca-

pacity retention (i.e. sulfur loss in course of the polysulfide redox shuttle

mechanism of the soluble lithium polysulfide (PS) reaction intermediates) by

employing a binder mixture of poly(ethylene oxide) (PEO) and poly(N-

vinylpyrrolidone) (PVP) with a synergistic effect. While PEO increased the

solubility and reaction kinetics of PS locally in the porous electrode, PVP

prevented the soluble intermediates from escaping the positive electrode by

formation of an insoluble polymer-PS complex. The combination of the two

polymers yields higher discharge capacities and better capacity retention as

compared to a reference electrode containing a “neutral” binder mixture of

CMC-Na and SBR or the current standard PVdF(-HFP).

Studies on the relationship(s) between the chemical properties and constitu-

tion of polymers used as binders in LIBs and LiSBs can help in the devel-

opment and design of improved materials with tailor-made properties for the

corresponding cell chemistries.

50

Sammanfattning på svenska

När Sony lanserade det första litiumjon-batteriet med en negativ grafite-

lektrod var detta möjligt genom bildandet av en stabil fas mellan elektroden

och elektrolyten, känt som SEI-lagret (Solid Electrolyte Interphase). Denna

bestod av nedbrytningsprodukter från den karbonatbaserade elektrolytens

lösningsmedel. Det är detta SEI-lager som fortfarande avgör batteriets livs-

längd, säkerheten under drift och att en stabil kapacitet kan levereras under

hundratals cykler. Under senare års utveckling har olika karbonat- och sva-

veloxid-baserade tillsatser i elektrolyten förbättrat skiktets sammansättning.

Men ytterligare förbättringar kan göras. Fram till upptäckten att polymerer

som karboximetylcellulosa-natriumsalt (CMC-Na) och polyakrylsyra (PAA)

kan förlänga cykellivslängden av kiselelektroder avsevärt, hade dessa bin-

demedel för elektroderna mestadels betraktas som inaktiva komponenter i

batteriet. Det har i stort sett ignorerats att dessa komponenter har stor pot-

ential för förbättringar genom att de är placerade direkt vid elektrodens yta

mot elektrolyten. I denna avhandling visas att bindemedlen kan ses som en

lokal delen av elektrolyten som påverkar ytreaktionerna, inte minst den

elektrokemiska cellreaktionen själv, och bildandet av en fast fas mellan elek-

trod och elektrolyt.

Det vanligen använda bindemedelet polyvinylidendifluorid (PVdF) eller

besläktade sampolymerer med detta uppvisar flera mindre goda egenskaper i

jämförelse med alternativa polymerer: PVdF-baserade bindemedel kräver till

exempel toxiska organiska lösningsmedel under tillverkningen av elektroder.

I mer porösa elektroder, som används i vissa batterisystem under utveckling,

kan bindemedlet leda till att porerna sätts igen då polymeren inte kan svälla i

elektrolyten, vilket i sin tur leder till lägre kapaciteten till följd av begränsad

tillgång till elektrodytan. Den stora graden av svällning i karbonatlösnings-

medel resulterar dock i undermåliga mekaniska egenskaper. Under ett inne-

håll på 10 vikts-% bindemedel (vanligen använt i akademisk forskning) blir

vidhäftningen till strömtilldelaren dålig, och detta blir än värre vid vätning.

Närvaron av PVdF vid elektrodgränssnittet riskerar vidare att hämma tillväx-

ten av ett skyddande SEI-lager.

I motsats till PVdF uppvisar de vattenlösliga bindemedel CMC-Na, styren-

butadien-gummi (SBR) och polynatriumakrylat (PAA-Na) en mindre grad

av svällning, och ger sålunda både förbättrad vidhäftning till strömtilldelaren

51

och goda mekaniska egenskaper hos kompositelektroderna. I avhandlingen

uppvisas att de dåligt svällbara bindemedlen CMC-Na och PAA-Na har en

högytresistans på cyklade grafitelektroder, vilket kraftigt förbättrar stabilite-

ten i närvaro av aggressiva elektrolytlösningsmedel, så som propylenkarbo-

nat (PC). Att minska mängden etylenkarbonat (EC) i den totala lösningsme-

delsblandningen är särskilt gynnsamt för lågtemperaturtillämpningar. Den

täta förekomsten av funktionella karboxyl-grupper i dessa två polymerer kan

också ses som ytaktiva ämnen som ökar graden av täckning på det aktiva

materialet i elektroden. När halten bindemedel minskar, visar sig dock PAA-

Na täcker främst den elektronledande komponenten i elektroden, och därmed

förlorar den sina skyddande egenskaper. I prover innehållande CMC-Na-

bindemedel, vilket är en mer rigid polymer på grund av att den är kolhydrat-

baserad, kan grafitpartiklarna i elektroden skyddas mot PC ända ner till hal-

ter av 5 vikts-% bindemedel.

En grafitelektrod med bindemedlet PAA-Na befanns vara (kinetiskt) stabil i

en eterbaserad elektrolyt när LiNO3 användes som additiv till elektrolyten.

Elektroden fungerade här som negativ elektrod i ett ’litium-jon-svavel’-

batteri. En högre coulombisk effektivitet, en betydligt längre livslängd vid

liknande urladdningskapacitet och en reducerad självurladdning på grund av

migrerande polysulfider indikerade ett SEI-lager på grafit med goda egen-

skaper. Till skillnad från negativa elektroder av litium-metall tolererade gra-

fitelektroden små volymer elektrolyt (närmare praktiskt relevanta värden)

mycket bättre. Detta åstadkoms med celler innehållandes ett överskott av 30

% svavel, i uppenbar kontrast till de ’konventionella’ Li-S-batterier som

innehåller litiummetall i överskott av mer än 2000 %, och vars långa livs-

längd och höga coulombiska effektivitet därigenom väsentligen är artificiella

egenskaper.

Den positiva svavelelektroden i Li-S-celler förbättrades också med avseende

på kapaciteten som kan hållas kvar (dvs, begränsningen av förlusten av sva-

vel genom en skyttelmekanism i elektrolyten av lösliga litiumpolysulfider

(PS) som bildas intermediärt i elektrodreaktionerna) genom att använda ett

bindemedel bestående av en blandning av polyetylenoxid (PEO) och poly-N-

vinylpyrrolidon (PVP), vilka gav en synergistisk effekt. Medan PEO ökar

kinetiken för löslighet och reaktioner av PS lokalt i den porösa elektroden,

förhindrar PVP de lösliga mellanprodukterna från att diffundera från den

positiva elektroden genom bildning av olösliga polymer-PS-komplex. Kom-

binationen av de två polymererna ger högre urladdningskapacitet under

längre tid jämfört med en referenselektrod som innehåller en ’neutral’

blandning av bindemedel (CMC-Na och SBR) eller den nuvarande standar-

den PVdF(-HFP).

52

Bibliography

1Chem. World 5 (2015). 2 N. Armaroli and V. Balzani, Angew. Chem. Int. Ed. 46, 52 (2007). 3 W. Lutz, W. Sanderson, and S. Scherbov, Nature 412, 543 (2001). 4 N. Armaroli and V. Balzani, Energy Environ. Sci. 4, 3193 (2011). 5 D. Larcher and J.-M. Tarascon, Nat. Chem. 7, 19 (2014). 6 M.Z. Jacobson, Energy Environ. Sci. 2, 148 (2009). 7 Z. Yang, J. Zhang, M.C.W. Kintner-Meyer, X. Lu, D. Choi, and J.P. Lemmon, Chem. Rev. 111, 3577

(2011). 8 M.A. Pellow, C.J.M. Emmott, C.J. Barnhart, and S.M. Benson, Energy Environ. Sci. 8, 1938 (2015). 9 M.S. Whittingham, Proc. IEEE 100, 1518 (2012). 10 B. Dunn, H. Kamath, and J.M. Tarascon, Science (80-. ). 334, 928 (2011). 11 J.-M. Tarascon and M. Armand, Nature 414, 359 (2001). 12 M. Armand and J.-M. Tarascon, Nature 451, 652 (2008). 13 C. Samaras and K. Meisterling, Environ. Sci. Technol. 42, 3170 (2008). 14 Sony Corporation, (1990). 15 A. Yoshino, Angew. Chem. 124, 5898 (2012). 16 E.J. Berg, C. Villevieille, D. Streich, S. Trabesinger, and P. Novák, J. Electrochem Soc. 162, A2468

(2015). 17 M.N. Obrovac and V.L. Chevrier, Chem. Rev. 114, 11444 (2014). 18 G. Patry, A. Romagny, S. Martinet, and D. Froelich, Energy Sci. Eng. 3, 71 (2015). 19 J.B. Goodenough and K.S. Park, J. Am. Chem. Soc. 135, 1167 (2013). 20 V. Etacheri, R. Marom, R. Elazari, G. Salitra, and D. Aurbach, Energy Environ. Sci. 6, 2581 (2011). 21 P. Adelhelm, P. Hartmann, C.L. Bender, M. Busche, C. Eufinger, and J. Janek, Beilstein J.

Nanotechnol. 6, 1016 (2015). 22 P.G. Bruce, S.A. Freunberger, L.J. Hardwick, and J.M. Tarascon, Nat. Mater. 11, 19 (2012). 23 C.H. Hamann and W. Vielstich, Elektrochemie, 4. Edition (Wiley-VCH, Weinheim, 2005). 24 K. Zaghib, M. Dontigny, A. Guerfi, J. Trottier, J. Hamel-Pâquet, V. Gariepy, K. Galoutov, P.

Hovington, A. Mauger, H. Groult, and C.M. Julien, J. Power Sources 216, 192 (2012). 25 M.S. Whittingham, Science (80-. ). 192, 1126 (1976). 26 R. Yazami and P. Touzain, J. Power Sources 9, 365 (1983). 27 H. Zabel, Phys. Bl. 45, 123 (1989). 28 M.S. Dresselhaus and G. Dresselhaus, Adv. Phys. 139 (1981). 29 J.R. Dahn, Phys. Rev. B. 44, 9170 (1991). 30 B. Jache and P. Adelhelm, Angew. Chem. Int. Ed. 53, 10169 (2014). 31 T. Tran and K. Kinoshita, J. Electroanal. Chem. 386, 221 (1995). 32 M. Winter, J.O. Besenhard, M.E. Spahr, and P. Novák, Adv. Mater. 10, 725 (1998). 33 T. Ohzuku, J. Electrochem. Soc. 140, 2490 (1993). 34 M.N. Obrovac and L. Christensen, Electrochem. Solid State Lett. 7, A93 (2004). 35 S.D. Beattie, D. Larcher, M. Morcrette, B. Simon, and J.-M. Tarascon, J. Electrochem. Soc. 155, A158

(2008). 36 B.P.N. Nguyen, J. Gaubicher, and B. Lestriez, Electrochim. Acta 120, 319 (2014). 37 S. Park, T. Kim, and S.M. Oh, Electrochem. Solid State Lett. 10, A142 (2007).

53

38 J.H. Ryu, J.W. Kim, Y.-E. Sung, and S.M. Oh, Electrochem. Solid State Lett. 7, A306 (2004). 39 Y. Oumellal, N. Delpuech, D. Mazouzi, N. Dupré, J. Gaubicher, P. Moreau, P. Soudan, B. Lestriez, and

D. Guyomard, J. Mater. Chem. 21, 6201 (2011). 40 N. Delpuech, D. Mazouzi, N. Dupre, P. Moreau, M. Cerbelaud, J.S. Bridel, E. De Vito, D. Guyomard,

B. Lestriez, and B. Humbert, J. Phys. Chem. C 118, 17318 (2014). 41 M.N. Obrovac and L.J. Krause, J. Electrochem. Soc. 154, A103 (2007). 42 M.T. McDowell, S.W. Lee, W.D. Nix, and Y. Cui, Adv. Mater. 4966 (2013). 43 B. Lestriez, S. Bahri, I. Sandu, L. Roue, and D. Guyomard, Electrochem. Commun. 9, 2801 (2007). 44 I. Kovalenko, B. Zdyrko, A. Magasinski, B. Hertzberg, Z. Milicev, R. Burtovyy, I. Luzinov, and G.

Yushin, Science (80-. ). 334, 75 (2011). 45 C. Erk, T. Brezesinski, H. Sommer, R. Schneider, and J. Janek, ACS Appl. Mater. Interfaces 5, 7299

(2013). 46 D. Mazouzi, Z. Karkar, C.R. Hernandez, P.J. Manero, D. Guyomard, L. Roue, and B. Lestriez, J. Power

Sources 280, 533 (2015). 47 I.A. Profatilova, C. Stock, A. Schmitz, S. Passerini, and M. Winter, J. Power Sources 222, 140 (2013). 48 K. Zaghib, A. Mauger, H. Groult, J.B. Goodenough, and C.M. Julien, Materials (Basel). 6, 1028

(2013). 49 T. Nordh, R. Younesi, D. Brandell, and K. Edström, J. Power Sources 294, 173 (2015). 50 A. Guerfi, S. Sévigny, M. Lagacé, P. Hovington, K. Kinoshita, and K. Zaghib, J. Power Sources 119-

121, 88 (2003). 51 K. Zaghib, M. Dontigny, A. Guerfi, P. Charest, I. Rodrigues, A. Mauger, and C.M. Julien, J. Power

Sources 196, 3949 (2011). 52 B. Häupler, A. Wild, and U.S. Schubert, Adv. Energy Mater. 5, 1402034 (2015). 53 Z. Song and H. Zhou, Energy Environ. Sci. 6, 2280 (2013). 54 T. Janoschka, M.D. Hager, and U.S. Schubert, Adv. Mater. 24, 6397 (2012). 55 J. Molenda, Solid State Ion. 175, 203 (2004). 56 J.B. Goodenough and Y. Kim, Chem. Mater. 22, 587 (2010). 57 J.B. Goodenough, H.Y.-P. Hong, and J.A. Kafalas, Mat. Res. Bull. 11, 203 (1976). 58 T. Ohzuku and Y. Makimura, Chem. Lett. 30, 642 (2001). 59 A.K. Padhi, Nanjundaswamy, and J.B. Goodenough, J. Electrochem. Soc. 144, 1188 (1997). 60 N. Ravet, Y. Chouinard, J.F. Magnan, S. Besner, M. Gauthier, and M. Armand, J. Power Sources 97-

98, 503 (2001). 61 Y. Huang and J.B. Goodenough, Chem. Mater. 20, 7237 (2008). 62 Q. Zhong, A. Bonakclarpour, M. Zhang, and J.R. Dahn, J. Electrochem. Soc. 144, 205 (1997). 63 K. Amine, H. Tukamoto, H. Yasuda, and Y. Fujita, J. Power Sources 504, 604 (1997). 64 K. Edström, T. Gustafsson, and J.O. Thomas, Electrochim. Acta 50, 397 (2004). 65 J.B. Goodenough and Y. Kim, J. Power Sources 196, 6688 (2011). 66 J. McBreen, H.S. Lee, X.Q. Yang, and X. Sun, J. Power Sources 89, 163 (2000). 67 K. Xu, Chem. Rev. 104, 4303 (2004). 68 Y. Ein-Eli, B. Markovsky, D. Aurbach, Y. Carmeli, H. Yamin, and S. Luski, Electrochim. Acta 39,

2559 (1994). 69 P. Verma, P. Maire, and P. Novák, Electrochim. Acta 55, 6332 (2010). 70 S. Malmgren, K. Ciosek, M. Hahlin, T. Gustafsson, M. Gorgoi, H. Rensmo, and K. Edström,

Electrochim. Acta 97, 23 (2013). 71 D. Aurbach, Y. Ein-Eli, B. Markovsky, A. Zaban, S. Luski, Y. Carmeli, and H. Yamin, J. Electrochem.

Soc. 142, 2882 (1995). 72 A.M. Andersson and K. Edström, J. Electrochem. Soc. 148, A1100 (2001). 73 D. Aurbach, B. Markovsky, A. Shechter, Y. Ein-Eli, and H. Cohen, J. Electrochem. Soc. 143, 3809

(1996). 74 A.M. Andersson, A. Henningson, H. Siegbahn, U. Jansson, and K. Edström, J. Power Sources 119-121,

522 (2003). 75 K. Tasaki, A. Goldberg, J.-J. Lian, M. Walker, A. Timmons, and S.J. Harris, J. Electrochem. Soc. 156,

54

A1019 (2009). 76 R. Younesi, G.M. Veith, P. Johansson, K. Edström, and T. Vegge, Energy Environ. Sci. (2015). 77 A. V. Plakhotnyk, L. Ernst, and R. Schmutzler, J. Fluor. Chem. 126, 27 (2005). 78 C.L. Campion, W.T. Li, and B.L. Lucht, J. Electrochem. Soc. 152, A2327 (2005). 79 O. Yariv, D. Hirshberg, E. Zinigrad, A. Meitav, D. Aurbach, M. Jiang, and B.R. Powell, J.

Electrochem. Soc. 161, A1422 (2014). 80 D. Yaakov, Y. Gofer, D. Aurbach, and I.C. Halalay, J. Electrochem. Soc. 157, A1383 (2010). 81 E. Zinigrad, L. Larush-Asraf, J.S. Gnanaraj, M. Sprecher, and D. Aurbach, Thermochim. Acta 438, 184

(2005). 82 S.S. Zhang, J. Power Sources 162, 1379 (2006). 83 D. Aurbach, K. Gamolsky, B. Markovsky, Y. Gofer, M. Schmidt, and U. Heider, Electrochim. Acta 47,

1423 (2002). 84 Z.X. Shu, R. McMillan, J.J. Murray, and I.J. Davidson, J. Electrochem. Soc. 143, 2230 (1996). 85 R. McMillan, H. Slegr, Z.. Shu, and W. Wang, J. Power Sources 81-82, 20 (1999). 86 G. Chen, G. V. Zhuang, T.J. Richardson, G. Liu, and P.N. Ross, Electrochem. Solid State Lett. 8, A344

(2005). 87 M. Xu, L. Zhou, L. Xing, W. Li, and B.L. Lucht, Electrochim. Acta 55, 6743 (2010). 88 G.H. Wrodding, T.M. Wrodnigg, J.O. Besenhard, and M. Winter, Electrochem. Commun. 1, 148

(1999). 89 G.H. Wrodding, J.O. Besenhard, and M. Winter, J. Electrochem. Soc. 146, 470 (1999). 90 W. Yao, Z. Zhang, J. Gao, J. Li, J. Xu, Z. Wang, and Y. Yang, Energy Environ. Sci. 2, 1102 (2009). 91 K. Abe, T. Hattori, K. Kawabe, Y. Ushigoe, and H. Yoshitake, J. Electrochem. Soc. 154, A810 (2007). 92 K. Ciosek Högström, S. Malmgren, M. Hahlin, H. Rensmo, F. Thébault, P. Johansson, and K. Edström,

J. Phys. Chem. C 117, 23476 (2013). 93 R. Wagner, S. Brox, J. Kasnatscheew, D.R. Gallus, M. Amereller, I. Cekic-Laskovic, and M. Winter,

Electrochem. Commun. 40, 80 (2014). 94 M. Gaberšček, M. Bele, J. Drofenik, R. Dominko, and S. Pejovnik, Electrochem. Solid State Lett. 3,

171 (2000). 95 J. Drofenik, M. Gaberšček, R. Dominko, F.W. Poulsen, M. Mogensen, S. Pejovnik, and J. Jamnik,

Electrochim. Acta 48, 883 (2003). 96 S. Komaba, M. Watanabe, H. Groult, N. Kumagai, and K. Okahara, Electrochem. Solid State Lett. 9,

A130 (2006). 97 P. Verma, T. Sasaki, and P. Novák, Electrochim. Acta 82, 233 (2012). 98 P. Verma and P. Novák, Carbon N. Y. 50, 2599 (2012). 99 P. Verma, P. Maire, and P. Novák, Electrochim. Acta 56, 3555 (2011). 100 S. Wenzel, H. Metelmann, C. Raiß, A.K. Dürr, J. Janek, and P. Adelhelm, J. Power Sources 243, 758

(2013). 101 S. Urbonaite, T. Poux, and P. Novák, Adv. Energy Mater. 5, 1500118 (2015). 102 Y. Son, J.-S. Lee, Y. Son, J.-H. Jang, and J. Cho, Adv. Energy Mater. 5, 1500110 (2015). 103 Z.W. Seh, Q. Zhang, W. Li, G. Zheng, H. Yao, and Y. Cui, Chem. Sci. 4, 3673 (2013). 104 F. Wu, J.T. Lee, A. Magasinski, H. Kim, and G. Yushin, Part. Part. Syst. Charact. 31, 639 (2014). 105 Y. Fu, C. Zu, and A. Manthiram, J. Am. Chem. Soc. 135, 18044 (2013). 106 Y. Yang, M.T. McDowell, A. Jackson, J.J. Cha, S.S. Hong, and Y. Cui, Nano Lett. 10, 1486 (2010). 107 Z. Yang, J. Guo, S.K. Das, Y. Yu, Z. Zhou, H.D. Abruna, and L. a Archer, J. Mater. Chem. A 1, 1433

(2013). 108 X.L. Ji, K.T. Lee, and L.F. Nazar, Nat. Mater. 8, 500 (2009). 109 N. Jayaprakash, J. Shen, S.S. Moganty, A. Corona, and L. a. Archer, Angew. Chem. Int. Ed. 50, 5904

(2011). 110 Y.L. Cao, X.L. Li, I.A. Aksay, J.P. Lemmon, Z.M. Nie, Z.G. Yang, and J. Liu, Phys. Chem. Chem.

Phys. 13, 7660 (2011). 111 X.L. Li, Y.L. Cao, W. Qi, L. V Saraf, J. Xiao, Z.M. Nie, J. Mietek, J.G. Zhang, B. Schwenzer, and J.

Liu, J. Mater. Chem. 21, 16603 (2011).

55

112 C.D. Liang, N.J. Dudney, and J.Y. Howe, Chem. Mater. 21, 4724 (2009). 113 M.A. Pope and I.A. Aksay, Adv. Energy Mater. 5, 1500124 (2015). 114 M. Oschatz, L. Borchardt, K. Pinkert, S. Thieme, M.R. Lohe, C. Hoffmann, M. Benusch, F.M. Wisser,

C. Ziegler, L. Giebeler, M.H. Rümmeli, J. Eckert, A. Eychmüller, and S. Kaskel, Adv. Energy Mater. 4,

1300645 (2014). 115 J.T. Lee, Y. Zhao, S. Thieme, H. Kim, M. Oschatz, L. Borchardt, A. Magasinski, W. Il Cho, S. Kaskel,

and G. Yushin, Adv. Mater. 25, 4573 (2013). 116 J. Kulisch, H. Sommer, T. Brezesinski, and J. Janek, Phys. Chem. Chem. Phys. 16, 18765 (2014). 117 A. Jozwiuk, H. Sommer, J. Janek, and T. Brezesinski, J. Power Sources 296, 454 (2015). 118 C.J. Hart, M. Cuisinier, X. Liang, D. Kundu, A. Garsuch, and L.F. Nazar, Chem. Commun. 51, 2308

(2015). 119 M.J. Lacey, F. Jeschull, K. Edström, and D. Brandell, Chem. Commun. 49, 8531 (2013). 120 M.J. Lacey, F. Jeschull, K. Edström, and D. Brandell, J. Power Sources 264, 8 (2014). 121 J. Gao, M.A. Lowe, Y. Kiya, and H.D. Abruna, J. Phys. Chem. C 115, 25132 (2011). 122 S. Tobishima, H. Morimoto, M. Aoki, Y. Saito, T. Inose, T. Fukumoto, and T. Kuryu, Electrochim.

Acta 49, 979 (2004). 123 M.J. Lacey, K. Edström, and D. Brandell, Electrochem. Commun. 46, 91 (2014). 124 R.D. Rauh, F.S. Shuker, J.M. Marston, and S.B. Brummer, J. Inorg. Nucl. Chem. 39, 1761 (1977). 125 R.D. Rauh, K.M. Abraham, G.F. Pearson, J.K. Surprenant, and S.B. Brummer, J. Electrochem. Soc.

126, 523 (1979). 126 H. Yamin and E. Peled, J. Power Sources 9, 281 (1983). 127 H. Yamin, A. Gorenshtein, J. Penciner, Y. Sternberg, and E. Peled, J. Electrochem. Soc. 135, 1045

(1988). 128 E. Peled, Y. Sternberg, A. Gorenshtein, and Y. Lavi, J. Electrochem. Soc. 136, 1621 (1989). 129 V. Gutmann, Electrochim. Acta 21, 661 (1976). 130 D.-H. Han, B.-S. Kim, S.-J. Choi, Y. Jung, J. Kwak, and S.-M. Park, J. Electrochem. Soc. 151, E283

(2004). 131 B.S. Kim and S.M. Park, J. Electrochem. Soc. 140, 115 (1993). 132 W. Giggenbach, J. Inorg. Nucl. Chem. 30, 3189 (1968). 133 K.H. Wujcik, T.A. Pascal, C.D. Pemmaraju, D. Devaux, W.C. Stolte, N.P. Balsara, and D.

Prendergast, Adv. Energy Mater. 5, 1500285 (2015). 134 C. Barchasz, F. Molton, C. Duboc, J.-C. Leprêtre, S. Patoux, and F. Alloin, Anal. Chem. 84, 3973

(2012). 135 Q. Wang, J. Zheng, E. Walter, H. Pan, D. Lv, P. Zuo, H. Chen, Z.D. Deng, B.Y. Liaw, X. Yu, X.

Yang, J.-G. Zhang, J. Liu, and J. Xiao, J. Electrochem. Soc. 162, A474 (2015). 136 M.U.M. Patel, R. Demir-Cakan, M. Morcrette, J.-M. Tarascon, M. Gaberšček, and R. Dominko,

ChemSusChem 1177 (2013). 137 S.I. Tobishima, H. Yamamoto, and M. Matsuda, Electrochim. Acta 42, 1019 (1997). 138 M. Hagen, P. Schiffels, M. Hammer, S. Dorfler, J. Tubke, M.J. Hoffmann, H. Althues, and S. Kaskel,

J. Electrochem. Soc. 160, A1205 (2013). 139 R. Dominko, R. Demir-Cakan, M. Morcrette, and J.M. Tarascon, Electrochem. Commun. 13, 117

(2011). 140 Y.-C. Lu, Q. He, and H.A. Gasteiger, J. Phys. Chem. C 118, 5733 (2014). 141 Y. V. Mikhaylik and J.R. Akridge, J. Electrochem. Soc. 151, A1969 (2004). 142 H.-S. Ryu, H.J. Ahn, K. Kim, J.-H. Ahn, J. Lee, and E. Cairns, J. Power Sources 140, 365 (2005). 143 S. Zhang, K. Ueno, K. Dokko, and M. Watanabe, Adv. Energy Mater. 5, 1500117 (2015). 144 E.S. Shin, K. Kim, S.H. Oh, and W. Il Cho, Chem. Commun. 49, 2004 (2013). 145 L.M. Suo, Y.S. Hu, H. Li, M. Armand, and L.Q. Chen, Nat. Commun. 4, (2013). 146 Y.S. Su and A. Manthiram, Nat. Commun. 3, (2012). 147 X. Wang, Z. Wang, and L. Chen, J. Power Sources 242, 65 (2013). 148 I. Bauer, S. Thieme, J. Brückner, H. Althues, and S. Kaskel, J. Power Sources 251, 417 (2014). 149 Y. V. Mikhaylik, I. Kovalev, R. Schock, K. Kumaresan, J. Xu, and J. Affinito, in ECS Trans. (The

56

Electrochemical Society, 2010), pp. 23–34. 150 F. Ding, W. Xu, G.L. Graff, J. Zhang, M. Sushko, X. Chen, Y. Shao, M.H. Engelhard, Z. Nie, J. Xiao,

X. Liu, P. V Sushko, J. Liu, and J.-G. Zhang, J. Am. Chem. Soc. (2013). 151 Z. Lin, Z. Liu, W. Fu, N.J. Dudney, and C. Liang, Angew. Chem. Int. Ed. 52, 7460 (2013). 152 M. Agostini, Y. Aihara, T. Yamada, B. Scrosati, and J. Hassoun, Solid State Ion. 244, 48 (2013). 153 A. Hayashi, T. Ohtomo, F. Mizuno, K. Tadanaga, and M. Tatsumisago, Electrochem. Commun. 5, 701

(2003). 154 M. Nagao, A. Hayashi, and M. Tatsumisago, Electrochim. Acta 56, 6055 (2011). 155 Y. V. Mikhaylik and (Sion Power Corporation), (2008). 156 D. Aurbach, E. Pollak, R. Elazari, G. Salitra, C.S. Kelley, and J. Affinito, J. Electrochem. Soc. 156,

A694 (2009). 157 S.S. Zhang, Electrochim. Acta 70, 344 (2012). 158 F. Wu, H. Kim, A. Magasinski, J.T. Lee, H.T. Lin, and G. Yushin, Adv. Energy Mater. 4, 1 (2014). 159 J. Brückner, S. Thieme, F. Böttger-Hiller, I. Bauer, H.T. Grossmann, P. Strubel, H. Althues, S. Spange,

and S. Kaskel, Adv. Func. Mater. 24, 1284 (2014). 160 N. Liu, L. Hu, M.T. Mcdowell, A. Jackson, and Y. Cui, ACS Nano 6487 (2011). 161 J. Hassoun and B. Scrosati, Angew. Chem. Int. Ed. 49, 2371 (2010). 162 J. Kim, D.J. Lee, H.G. Jung, Y.K. Sun, J. Hassoun, and B. Scrosati, Adv. Func. Mater. 23, 1076

(2013). 163 J. Hassoun, Y.K. Sun, and B. Scrosati, J. Power Sources 196, 343 (2011). 164 X. He, J. Ren, L. Wang, W. Pu, C. Wan, and C. Jiang, ECS Trans. 2, 47 (2007). 165 S. Zheng, Y. Chen, Y. Xu, F. Yi, Y. Zhu, Y. Liu, and J. Yang, ACS Nano 7, 10995 (2013). 166 M. Agostini, B. Scrosati, and J. Hassoun, Adv. Energy Mater. 5, 1500481 (2015). 167 A. Göthlich, S. Koltzenburg, and G. Schornick, Chemie Unserer Zeit 39, 262 (2005). 168 A.D. McNaught and A. Wilkinson, IUPAC. Compendium of Chemical Terminology, 2nd Ed. (the

“Gold Book”) (Blackwell Scientific Publications, Oxford, 1997). 169 S.-L. Chou, Y. Pan, J.-Z. Wang, H.-K. Liu, and S.-X. Dou, Phys. Chem. Chem. Phys. 16, 20347

(2014). 170 A. Magasinski, B. Zdyrko, I. Kovalenko, B. Hertzberg, R. Burtovyy, C.F. Huebner, T.F. Fuller, I.

Luzinov, and G. Yushin, ACS Appl. Mater. Interfaces 2, 3004 (2010). 171 A.M. Stephan and Y. Saito, Solid State Ion. 148, 475 (2002). 172 J.Y. Song, Y.Y. Wang, and C.C. Wan, J. Power Sources 77, 183 (1999). 173 M. Yoo, C.W. Frank, S. Mori, and S. Yamaguchi, Polymer (Guildf). 44, 4197 (2003). 174 Y. Takahashi, F. Suzuki, and T. Iwasaki, US005415958A (1995). 175 J. Li, R. Klöpsch, S. Nowak, M. Kunze, M. Winter, and S. Passerini, J. Power Sources 196, 7687

(2011). 176 S.F. Lux, F. Schappacher, A. Balducci, S. Passerini, and M. Winter, J. Electrochem. Soc. 157, A320

(2010). 177 H. Maleki, G. Deng, and I. Kerzhner-Haller, J. Electrochem. Soc. 147, 4470 (2000). 178 P. Biensan, B. Simon, and J.P. Peres, J. Power Sources 81-82, 906 (1999). 179 L. El Ouatani, R. Dedryvère, J.-B. Ledeuil, C. Siret, P. Biensan, J. Desbrières, and D. Gonbeau, J.

Power Sources 189, 72 (2009). 180 K. Takamura and D. Urban, Polymer Dispersions and Their Industrial Applications (Wiley-VCH,

Weinheim, 2002). 181 R. Dominko, J. Drofenik, and M. Gaberšček, J. Power Sources 94, 97 (2001). 182 R. Dominko, M. Gaberšček, M. Bele, J. Drofenik, E.M. Skou, A. Wursig, P. Novák, and J. Jamnik, J.

Electrochem. Soc. 151, A1058 (2004). 183 J.-H. Lee, S. Lee, U. Paik, and Y.-M. Choi, J. Power Sources 147, 249 (2005). 184 J.-H. Lee, U. Paik, V.A. Hackley, and Y.-M. Choi, J. Power Sources 161, 612 (2006). 185 K. Ui, S. Kikuchi, F. Mikami, Y. Kadoma, and N. Kumagai, J. Power Sources 173, 518 (2007). 186 S. Komaba, K. Okushi, T. Ozeki, H. Yui, Y. Katayama, T. Miura, T. Saito, and H. Groult,

Electrochem. Solid State Lett. 12, A107 (2009).

57

187 S. Komaba, N. Yabuuchi, T. Ozeki, K. Okushi, H. Yui, K. Konno, Y. Katayama, and T. Miura, J.

Power Sources 195, 6069 (2010). 188 S. Komaba, T. Itabashi, T. Kimura, H. Groult, and N. Kumagai, J. Power Sources 146, 166 (2005). 189 S. Komaba, T. Itabashi, M. Watanabe, H. Groult, and N. Kumagai, J. Electrochem. Soc. 154, A322

(2007). 190 U. Maver, A. Žnidaršič, and M. Gaberšček, J. Mater. Chem. 21, 4071 (2011). 191 D. Mazouzi, B. Lestriez, L. Roue, and D. Guyomard, Electrochem. Solid State Lett. 12, A215 (2009). 192 N. Yabuuchi, K. Shimomura, Y. Shimbe, T. Ozeki, J.Y. Son, H. Oji, Y. Katayama, T. Miura, and S.

Komaba, Adv. Energy Mater. 1, 759 (2011). 193 F. Courtel and Y. Abu-Lebdeh, WO2011130837 A1 (2011). 194 B. Koo, H. Kim, Y. Cho, K.T. Lee, N.-S. Choi, and J. Cho, Angew. Chem. Int. Ed. 51, 8762 (2012). 195 C. Wang, H. Wu, Z. Chen, M.T. McDowell, Y. Cui, and Z. Bao, Nat. Chem. 5, 1042 (2013). 196 M.H. Ryou, J. Kim, I. Lee, S. Kim, Y.K. Jeong, S. Hong, J.H. Ryu, T.S. Kim, J.-K. Park, H. Lee, and

J.W. Choi, Adv. Mater. 25, 1571 (2013). 197 M. Wu, X. Xiao, N. Vukmirovic, S. Xun, P.K. Das, X. Song, P. Olalde-Velasco, D. Wang, A.Z.

Weber, L.-W. Wang, V.S. Battaglia, W. Yang, and G. Liu, J. Am. Chem. Soc. 135, 12048 (2013). 198 C. Barchasz, J.-C. Leprêtre, S. Patoux, and F. Alloin, Electrochim. Acta 89, 737 (2013). 199 C. Barchasz, J.-C. Leprêtre, S. Patoux, and F. Alloin, J. Electrochem. Soc. 160, A430 (2013). 200 J.-Q. Huang, Q. Zhang, S.-M. Zhang, X.-F. Liu, W. Zhu, W.-Z. Qian, and F. Wei, Carbon N. Y. 58, 99

(2013). 201 L. Johnson, C. Li, Z. Liu, Y. Chen, S.A. Freunberger, P.C. Ashok, B.B. Praveen, K. Dholakia, J.

Tarascon, and P.G. Bruce, Nat. Chem. 6, 1091 (2014). 202 M. Cuisinier, C. Hart, M. Balasubramanian, A. Garsuch, and L.F. Nazar, Adv. Energy Mater. 5,

1401801 (2015). 203 R.M. Michell, I. Blaszczyk-Lezak, C. Mijangos, and A.J. Müller, J. Polym. Sci. Part B Polym. Phys.

52, 1179 (2014). 204 M. Steinhart, P. Göring, H. Dernaika, M. Prabhukaran, U. Gösele, E. Hempel, and T. Thurn-Albrecht,

Phys. Rev. Lett. 97, 1 (2006). 205 C. Raiß, K. Peppler, J. Janek, and P. Adelhelm, Carbon N. Y. 79, 245 (2014). 206 R. Younesi, N. Singh, S. Urbonaite, and K. Edström, ECS Trans. 25, 121 (2010). 207 F. Haaf, A. Sanner, and F. Straub, Polym. J. 17, 143 (1985). 208 I. Iliopoulos and R. Audebert, Macromolecules 24, 2566 (1991). 209 B.S. Ghosh and J. Rasmusson, 1997 (1998). 210 C.-S. Kim, A. Guerfi, P. Hovington, J. Trottier, C. Gagnon, F. Barray, A. Vijh, M. Armand, and K.

Zaghib, J. Power Sources 241, 554 (2013). 211 D. Aurbach, B. Markovsky, I. Weissman, E. Levi, and Y. Ein-Eli, Electrochim. Acta 45, 67 (1999). 212 J.O. Besenhard, M. Winter, J. Yang, and W. Biberacher, J. Power Sources 54, 228 (1995). 213 S. Komaba, T. Ozeki, and K. Okushi, J. Power Sources 189, 197 (2009). 214 H. Buqa, a. Würsig, D. Goers, L.J. Hardwick, M. Holzapfel, P. Novák, F. Krumeich, and M.E. Spahr,

J. Power Sources 146, 134 (2005). 215 N. Shpigel, M.D. Levi, S. Sigalov, O. Girshevitz, D. Aurbach, L. Daikhin, N. Jäckel, and V. Presser,

Angew. Chemie Int. Ed. 54, 12353 (2015). 216 L. Wang, Y. Fu, V.S. Battaglia, and G. Liu, RSC Adv. 3, 15022 (2013). 217 J. Chong, S. Xun, H. Zheng, X. Song, G. Liu, P. Ridgway, J.Q. Wang, and V.S. Battaglia, J. Power

Sources 196, 7707 (2011). 218 G. Liu, H. Zheng, S. Kim, Y. Deng, a. M. Minor, X. Song, and V.S. Battaglia, J. Electrochem. Soc.

155, A887 (2008). 219 G. Liu, H. Zheng, A.S. Simens, A.M. Minor, X. Song, and V.S. Battaglia, J. Electrochem. Soc. 154,

A1129 (2007). 220 G. Liu, H. Zheng, X. Song, and V.S. Battaglia, J. Electrochem. Soc. 159, A214 (2012). 221 H. Zheng, R. Yang, G. Liu, X. Song, and V.S. Battaglia, J. Phys. Chem. C 116, 4875 (2012). 222 H. Buqa, M. Holzapfel, F. Krumeich, C. Veit, and P. Novák, J. Power Sources 161, 617 (2006).

58

223 D. Aurbach, J. Power Sources 119-121, 497 (2003). 224 D. Aurbach, A. Zaban, Y. Ein-Eli, I. Weissman, O. Chusid, B. Markovsky, M. Levi, E. Levi, A.

Schechter, and E. Granot, J. Power Sources 68, 91 (1997). 225 W. Li, H. Yao, K. Yan, G. Zheng, Z. Liang, Y.-M. Chiang, and Y. Cui, Nat. Commun. 6, 7436 (2015). 226 A. Rosenman, E. Markevich, G. Salitra, D. Aurbach, A. Garsuch, and F.F. Chesneau, Adv. Energy

Mater. 5, 1500212 (2015). 227 S. Urbonaite and P. Novák, J. Power Sources 249, 497 (2014). 228 M. Hagen, D. Hanselmann, K. Ahlbrecht, R. Maça, D. Gerber, and J. Tübke, Adv. Energy Mater. 5,

1401986 (2015). 229 J. Li, R.B. Lewis, and J.R. Dahn, Electrochem. Solid State Lett. 10, A17 (2007).