Digital Comprehensive Summaries of Uppsala Dissertations from the Faculty of Science and Technology 2075

Operando Characterisation of Lithium–Sulfur Batteries

YU-CHUAN CHIEN

ACTA UNIVERSITATIS

UPSALIENSIS UPPSALA

2021

ISSN 1651-6214 ISBN 978-91-513-1297-2 URN urn:nbn:se:uu:diva-453442

Dissertation presented at Uppsala University to be publicly examined in Polhemsalen,Ångströmlaboratoriet, Lägerhyddsvägen 1, Uppsala, Friday, 5 November 2021 at 09:15 for the degree of Doctor of Philosophy. The examination will be conducted in English. Faculty examiner: Professor Paul Shearing (University College London).

Abstract Chien, Y.-C. 2021. Operando Characterisation of Lithium–Sulfur Batteries. Digital Comprehensive Summaries of Uppsala Dissertations from the Faculty of Science and Technology 2075. 63 pp. Uppsala: Acta Universitatis Upsaliensis. ISBN 978-91-513-1297-2.

Lithium–sulfur (Li–S) batteries have been under the spotlight of research on electrochemical energy storage systems, primarily owing to their high theoretical specific energy (2552 Wh kg-1). So far, Li–S cells on the market have presented a specific energy of 400 Wh kg-1, which is superior to many commercial alternatives, but far below the theoretical value. At the same time, Li–S batteries encounter other problems that are generally not associated with the standard Li-ion batteries, such as low utilisation rate of active materials and short cycle life. These often originate from the unique catholyte nature and/or the low reversibility of the metallic Li electrode.

The dissolution and precipitation of elemental sulfur and lithium sulfide in the positive electrode are here investigated by operando X-ray diffraction (XRD) and small-angle neutron/X-ray scattering (SANS/SAXS) coupled with the Intermittent Current Interruption (ICI) method. The real-time internal and diffusion resistances are correlated to the kinetics of the precipitation of the crystalline species by operando XRD. Through operando SANS and SAXS, the formation of crystalline and amorphous solid-state discharge products and the compositional variation of catholyte inside the mesopores are linked to features in the resistance profiles. These studies indicate that the ionic transport limitation inside the positive electrode is the cause for the low sulfur utilisation during battery discharge.

To examine the impact of the repetitive precipitation on the functionality of the positive sulfur electrode, a method based on electrochemical impedance spectroscopy (EIS) was developed to track the electrochemically active surface area of the carbon matrix in-situ over extensive cycling. The investigation found no progressive passivation on the positive electrode despite the rapid decrease in specific discharge capacity. Additionally, a novel three-electrode setup for Li–S cells reveals a faster growth of the resistance on the metallic Li electrode along cycling. These findings suggest that primarily the negative electrode limits the cycle life.

Through providing the mechanistic insights of operational Li–S cells, this thesis demonstrates the value of simultaneous electrochemical and materials characterisations for understanding the complex Li–S system.

Keywords: lithium–sulfur batteries, operando characterisation, intermittent current interruption, X-ray diffraction, small-angle scattering, electrochemical impedance spectroscopy

Yu-Chuan Chien, Department of Chemistry - Ångström, Structural Chemistry, Box 538, Uppsala University, SE-751 21 Uppsala, Sweden.

© Yu-Chuan Chien 2021

ISSN 1651-6214 ISBN 978-91-513-1297-2 URN urn:nbn:se:uu:diva-453442 (http://urn.kb.se/resolve?urn=urn:nbn:se:uu:diva-453442)

To Ama and Agong

List of Papers

This thesis is based on the following papers, which are referred to in the text

by their Roman numerals.

I. Chien, Y.-C., Menon, A. S., Brant, W. R., Brandell, D., Lacey,

M. J. (2020) Simultaneous Monitoring of Crystalline Active Ma-

terials and Resistance Evolution in Lithium–Sulfur Batteries.

Journal of the American Chemical Society, 142(3):1449–1456

II. Chien, Y.-C., Menon, A. S., Brant, W. R., Brandell, D., Lacey,

M. J. (2021) Understanding the Impact of Precipitation Kinetics

on the Electrochemical Performance of Lithium–Sulfur Batteries

by Operando X-ray Diffraction. In manuscript.

III. Chien, Y.-C., Lacey, M. J., Steinke, N.-J., Brandell, D., Rennie,

A. R. (2021) Correlations between Precipitation Reactions and

Electrochemical Performance of Lithium–Sulfur Batteries. Sub-

mitted manuscript.

IV. Li, H., Lampkin, J., Chien, Y.-C., Furness, L., Brandell, D.,

Lacey, M.J., Garcia-Araez, N. (2021) Operando Characterization

of Active Surface Area and Passivation Effects on Sulfur-Carbon

Composites for Lithium-Sulfur Batteries. Submitted Manuscript.

V. Chien, Y.-C., Brandell, D., Lacey, M. J. (2021) Towards Reliable

Three-Electrode Cells for Lithium–Sulfur batteries. Submitted

manuscript.

Reprints were made with permission from the respective publishers.

Disclaimer: Parts of this thesis are based on my licentiate thesis titled “Inves-

tigations of the performance-limiting factors in lithium-sulfur batteries” (Upp-

sala University, 2019).

My Contributions to the Papers

I. I derived the theoretical foundation of the electrochemical method

with my supervisors’ input, conceptualised, executed and analysed

the experiments with co-authors’ assistance in the refinement of the

diffraction data. I wrote the manuscript with input from co-authors.

II. I conceptualised, executed and analysed the experiments with co-

authors’ assistance in the refinement of the diffraction data. I wrote

the manuscript with input from co-authors.

III. I conceptualised, executed and analysed the experiments with co-

authors’ assistance in the execution and interpretation of small-an-

gle scattering data. I wrote the manuscript with input from co-au-

thors.

IV. I fabricated the optimised electrodes, conducted the operando

measurements and participated in the data analysis. I participated

in writing the manuscript.

V. I conceptualised, executed and analysed the experiments. I wrote

the manuscript with input from co-authors.

Papers not Included in the Thesis

I. Chien, Y.-C., Pan, R., Lee, M.-T., Nyholm, L., Brandell, D., Lacey,

M. J. (2019) Cellulose Separators With Integrated Carbon Nano-

tube Interlayers for Lithium-Sulfur Batteries: An Investigation into

the Complex Interplay between Cell Components. Journal of the

Electrochemical Society, 166(14):A3235–A3241

II. Hernández, G., Naylor, A. J., Chien, Y.-C., Brandell, D., Min-

demark, J., Edström, K. (2020) ACS Sustainable Chemistry & En-

gineering, 8(27):10041–10052

III. Chien, Y.-C., Jang, H., Brandell, D. and Lacey, M. J. (2021)

Poly(Ethylene Glycol-block-2-Ethyl-2-Oxazoline) as Cathode

Binder in Lithium-Sulfur Batteries. ChemistryOpen, in press.

Contents

1 Introduction ......................................................................................... 13

1.1 Electrochemical Energy Storage ..................................................... 13

1.1.1 Rechargeable Batteries .......................................................... 14

1.1.2 Lithium-ion Batteries ............................................................. 14

1.2 Lithium–Sulfur Batteries ................................................................. 15

1.2.1 Cathode and Catholyte ........................................................... 15

1.2.2 Anode and Catholyte ............................................................. 18

2 Scope of the Thesis .............................................................................. 20

3 Experimental Methods ......................................................................... 22

3.1 Sample Preparation ......................................................................... 22

3.1.1 Preparation of Sulfur/Carbon Composite Electrodes ............. 22

3.1.2 Two- and Three-electrode Electrochemical Cells .................. 22

3.1.3 Electrochemical Cell Designs for Operando Measurements . 23

3.2 Characterisation Techniques ........................................................... 24

3.2.1 Intermittent Current Interruption Method .............................. 24

3.2.2 Electrochemical Impedance Spectroscopy ............................ 26

3.2.3 X-ray Diffraction/Wide-angle X-ray Scattering .................... 27

3.2.4 Small-angle X-ray and Neutron Scattering ............................ 28

4 Results and Discussions ....................................................................... 29

4.1 Illustration of a Discharging Positive Electrode ............................. 29

4.1.1 Precipitation and Energy Loss Processes ............................... 29

4.1.2 From Cell Parameters to Electrochemical Performance ........ 33

4.1.3 Dissolved, Crystalline and Amorphous Sulfur Species ......... 39

4.2 Investigations into the Cycle Life Limitations ................................ 45

4.2.1 Electrochemically Active Surface Area over Cycling ........... 45

4.2.2 Resistance Distribution and Its Evolution .............................. 49

5 Conclusions ......................................................................................... 53

6 Sammanfattning på svenska ................................................................ 55

7 Acknowledgements.............................................................................. 57

8 References ........................................................................................... 58

Abbreviations

BET Brunauer–Emmett–Teller theory/method

CCC Carbon nanotube-coated cellulose

CE Counter electrode

CNF Carbon nanofiber

CNT Carbon nanotube

CPE Constant phase element

DME 1,2-dimethyloxyethane

DOL 1,3-dioxolane

ECM Equivalent circuit model

EIS Electrochemical impedance spectroscopy

E/S Electrolyte-to-sulfur (ratio)

FWHM Full-width at half-maximum

gS Mass of sulfur in grams

ICI Intermittent current interruption (method)

KB Ketjenblack®

LiFSI Lithium bis(fluorosulfonyl)imide

LiTFSI Lithium bis(trifluoromethanesulfonyl)imide

OCV Open-circuit voltage

PEO Poly(ethylene oxide)

PVP Poly(vinylpyrrolidone)

RE Reference electrode

SAS Small-angle scattering

SANS Small-angle neutron scattering

SAXS Small-angle X-ray scattering

SEI Solid electrolyte interphase

SEM Scanning electron microscopy

SLD Scattering length density

SoC State of charge

SoD State of discharge

SSA Specific surface area

WAXS Wide-angle X-ray scattering

WE Working electrode

XRD X-ray diffraction

13

1 Introduction

1.1 Electrochemical Energy Storage1,2

Electrochemical energy storage systems operate by guiding the electron in-

volved in a pair of reduction and oxidation reactions through an electric cir-

cuit, where the electric current can do work when the redox reaction is spon-

taneous. This process of harnessing the free energy of redox reactions is

termed discharge. By the sign convention, the electrode where reduction takes

place, and thus the electrons flow into, is the positive electrode. The other

electrode is the negative electrode by the same rationale. The substances going

through these redox reactions in the electrodes are the active materials. The

redox reaction is spontaneous because the active material at the positive elec-

trode has a higher potential to reduce than that on the negative electrode. The

difference in the reduction potentials gives rise to the electromotive force (E),

which can be related to the Gibbs free energy (∆G) of the redox reaction by

∆𝐺 = 𝑛𝐹𝐸 (1.1)

where n is the number of moles of electrons transferred in the reaction and F

is the amount of charge carried by one mole of electrons, i.e. the Faraday con-

stant. In an electrochemical energy source, the electromotive force is meas-

ured by the open-circuit voltage (OCV), which is the potential difference be-

tween the positive and negative electrodes when no current is present.

To sustain the electric current, the charge difference caused by the move-

ment of electrons has to be balanced out. This is achieved by the ions in the

electrolyte. The electrolyte is an electron-insulating but ion-conducting me-

dium between the two electrodes, which is usually a liquid with high ionic

conductivity. As an electron moves out of the negative electrode, a positive

charge is left behind in the oxidized active material, whether it stays in the

electrode or dissolves in the electrolyte. This excess positive charge has to be

compensated either by an anion moving towards this electrode or a cation

leaving it. Due to the size difference, the ionic conductivity of the electrolyte

is usually much lower than the electronic conductivity of the electric circuit.

Depending on where the active materials are stored, electrochemical energy

storage devices can be categorized as batteries or fuel cells. In a fuel cell, the

active materials are constantly replenished, used and exhausted, as its name

14

suggests, whereas in a battery, the active materials are embedded and sealed

inside the cell.

1.1.1 Rechargeable Batteries2,3

Batteries discharge by spontaneous reduction at the positive electrode and ox-

idation at the negative electrode. The positive and negative electrodes are

therefore also named as the cathode and anode, respectively. For a rechargea-

ble battery, the process can be reversed by supplying electric current in the

opposite direction, which results in oxidation at the positive electrode and re-

duction at the negative electrode, also known as charging. The electrical en-

ergy is then stored chemically in a rechargeable battery.

Besides the two electrodes and the electrolyte, a modern battery usually

consists of a separator to minimize the required distance between the elec-

trodes. The separator is a porous insulating membrane, whose pores are filled

with liquid electrolyte in a battery. The two electrodes with this combination

in between can be stacked together densely without being short-circuited. This

configuration not only enhances the energy density of the battery system but

also minimizes the energy dissipated by the solution resistance of the electro-

lyte.

Battery systems are usually named after their components. For example:

lead-acid batteries, nickel metal hydride batteries, lithium-ion batteries. The

following subchapter will introduce the working principles of lithium-ion bat-

teries as the current dominant technology.

1.1.2 Lithium-ion Batteries3

Lithium-ion batteries are a set of battery systems that utilize Li ions as the

charge carriers but does not apply metallic Li as the negative electrode. This

distinction from Li batteries (or Li-metal batteries) was deliberately made in

the 1980’s due to the safety issue related to the Li metal anode. Metallic Li

has several advantages as an anode material, e.g. low reduction potential and

high specific capacity. However, it also has one problem that made it imprac-

tical in these early days of Li-battery development, which is still not really

solved to date: inhomogeneous Li deposition. The most dangerous impact of

dendritic growth of Li is that it can short-circuit the cell by penetrating the

separator. Therefore, researchers circumvented the use of metallic Li by pair-

ing graphite, as the negative electrode, with layered lithium metal oxides as

the positive electrode. Both active materials are categorized as intercalation

materials, which means that Li ions are inserted into the layered atomic struc-

ture of the material during the redox reaction while the layered structure re-

mains topologically the same. This setup sacrifices the specific energy but still

benefits from the high charge-to-mass ratio of Li ions and enhances safety

substantially. Thus, Li-ion batteries were successfully commercialized in

15

1990’s and has until now been the rechargeable battery system with the high-

est energy density on the market.

1.2 Lithium–Sulfur Batteries

The quest for more efficient battery technology did not stop after the commer-

cialization of Li-ion batteries. Despite the safety concerns discussed above,

the metallic Li anode is still regarded as one of the promising solutions to push

the specific energy density of commercial batteries beyond its current values.

To match the high specific capacity of Li, 3860 mAh g-1,3 sulfur is considered

a candidate for the positive electrode material due to its high specific capacity

among cathode materials, 1672 mAh g-1.4 Additionally, sulfur is non-toxic and

abundant in the earth’s crust.5,6 In an ideally balanced cell with complete uti-

lisation of both active materials, a Li–S battery has a theoretical specific en-

ergy of 2552 Wh kg-1 on the materials level based on the following reaction.4

2Li + S → Li2S (1.2)

The overall reaction above is made up of the following half-reactions taking

place at the positive and negative electrodes, respectively.

S + 2Li+ + 2e- → Li2S (1.3)

Li → Li+ + e- (1.4)

On the cell level, a specific energy of 400 Wh kg-1 has been demonstrated

commercially.7–9 This performance is superior to that of the current Li-ion bat-

teries on the market but still falls behind the theoretical value of the Li–S sys-

tem by a factor of six, which is high if we consider the practical specific energy

of Li-ion batteries only lags its theoretical value by a factor of three. Moreo-

ver, the short cycle life of the Li–S cells also hinders their wider applica-

tion.10,11 These issues can be attributed to several challenges that the develop-

ment of the Li–S technology is facing, which will be discussed below with the

introduction to the corresponding cell components.

1.2.1 Cathode and Catholyte

Due to the low electronic conductivity of sulfur, 5∙10-30 S cm-1,5 a conductive

matrix is required to facilitate the electrochemical reactions. A wide variety

of conductive materials have been investigated in literature,6,12,13 among which

carbonaceous materials constitute the majority due to their high conductivity,

low density and versatility. To form a mechanically robust electrode with the

16

above-mentioned powder materials, one or more polymers are used as bind-

ers.14–16 Various fractions of sulfur can be found in the published works with

the majority ranging from 50 % to 70 % by weight.17 Due to the relatively high

composition of the carbonaceous materials and their important role, the posi-

tive electrodes are often referred to as sulfur/carbon (S/C) composite elec-

trodes.

The working mechanism of a cathode in Li–S batteries is substantially differ-

ent from that in Li-ion batteries. The difference originates from the interme-

diate states between the charged state, elemental sulfur, and the discharged

state, lithium sulfide (Li2S).4 As sulfur is discharged in commonly used ether-

based electrolyte, lithium polysulfides (Li2Sx, x = 2–8) are formed first, before

they are eventually reduced to Li2S.18,19 These polysulfides are soluble to var-

ious extents in liquid electrolyte while sulfur and Li2S are not.20–22 Thus, in-

stead of staying in the solid state like a common cathode material of Li-ion

batteries, sulfur species go through dissolution and precipitation during the

course of discharge or charge. In other words, it is a catholyte system at most

states of charge (SoC) levels.

A typical potential profile of an S/C composite electrode consists of two plat-

eaus, which are more distinct upon discharge and closer to each other upon

charging, as shown in Figure 1.1. The assignment of redox reactions to the

potential plateaus are still being discussed since the exact reaction mechanism

is still under investigation. Nonetheless, it is widely accepted to associate the

upper plateau (~2.35 V vs Li/Li+) with the conversion of elemental sulfur to

the long-chained polysulfides (Li2Sx, x = 6–8) and the lower one (~2.1 V vs

Li/Li+) with that of the short-chained polysulfides (Li2Sx, x = 2–4) to Li2S.23,24

Figure 1.1. Schematic illustration of a Li–S cell (left) and a typical potential profile of an S/C composite electrode (right), where E is the potential and Q is the specific ca-pacity based on the mass of sulfur in grams (gS

-1).

17

Given the solubility of the reaction intermediates in the electrolyte, it is not

difficult to imagine that the active material of the positive electrode can diffuse

to the other side of the separator and react with the negative electrode without

contributing any current to the external circuit. Moreover, the reduced poly-

sulfides may diffuse back to the positive electrode and be oxidized if the cell

is charging.25 This loop of redox reactions is commonly referred to as “poly-

sulfide shuttle. It can be inferred from the above description of the polysulfide

shuttle that its consequences are self-discharge and low Coulombic effi-

ciency.25 Due to the faster kinetics of the reduction of the long-chained poly-

sulfides, the redox shuttle occurs to a larger extent at the upper potential plat-

eau.26,27 Thus, it can often be identified by a prolonged upper charging plateau

on the potential profile.25

Another challenge facing the positive electrode is the explanation for the

low sulfur utilisation. Despite the high theoretical specific capacity of sulfur,

only around 1000 mAh g-1 is reported in the majority of the current litera-

ture,6,28 which means only around 60% utilisation of the active material. This

discrepancy is a roadblock to the development of the Li–S technology since

there is already a considerable fraction of conductive carbon materials in the

positive electrode, which do not contribute to the capacity of the electrode.

The final step of the discharge process, which involves the precipitation of

Li2S, has been extensively investigated to identify the reason for low sulfur

utilisation.29–37 Early studies proposed that a dense two-dimensional film of

Li2S is formed on the surface of the carbon matrix, which eventually passiv-

ates the electrode, and prevents further reduction of sulfur.37,38 However, re-

cent reports have shown that porous Li2S particles can also precipitate under

certain conditions, which do not cover the entire surface of the electrode.32,39

Thus, instead of electrode passivation, the diffusion of polysulfides has been

suggested to be the reason for low sulfur utiliation.32 These conflicting claims

demonstrate that the complex discharge process in the positive electrode is yet

fully understood. A series of investigations of the discharge processes inside

the positive electrode will be presented in chapter 4.1.

Following the hypothesis of the Li2S passivation at the end of discharge, a

frequently proposed reason for the capacity fade of Li–S cells is the drop in

the electrochemically active surface area due to the irreversible coverage of

an insulating layer on the carbon matrix.40,41 Longer cycle life was reported

for cells with a cycling protocol that avoids the formation of Li2S. Similar to

the disagreement about the end of discharge, contradictory reports can also be

found in the literature, where cells achieved longer cycle life by avoiding the

upper plateaus.42 Nevertheless, the missing information in these works is the

quantification of the surface area of the carbon matrix. In chapter 4.2.1, a

method based on electrochemical impedance spectroscopy (EIS) will be intro-

duced to track the surface area along with the cycle number.

18

1.2.2 Anode and Catholyte

As mentioned in previously, metallic Li is the ideal material for the negative

electrode in Li-ion-based batteries for the reason of its high specific capacity

and low reduction potential. However, the safety risk brought by the dendritic

growth of Li is a roadblock yet to be removed, at least not in liquid electro-

lyte.43 In addition to the uneven deposition, inhomogeneous dissolution is also

a problem with the metallic Li electrode since it results in Li domains electri-

cally disconnected from the bulk. This so-called “dead Li” becomes inactive

and does not participate in the redox reaction anymore, which means a perma-

nent loss of capacity.44

Metallic Li electrodes face further challenge in Li–S batteries due to the

catholyte nature of the sulfur electrode. As discussed above, the soluble poly-

sulfides can diffuse between the positive and negative electrodes and go

through repeated redox reactions, which waste the stored charge in the battery.

For Li-ion batteries, an ideal SEI is expected to be both electronically insulat-

ing to stop further electrolyte decomposition and Li-ion conductive to main-

tain the redox reactions at the anode.2 For Li–S cells, the SEI has one more

mission—shielding Li from the polysulfide shuttle.43

The addition of lithium nitrate (LiNO3) as a co-salt in the electrolyte was

found to facilitate the formation of a protective SEI on Li to reduce the poly-

sulfide shuttle.45,46 It was concluded from surface analyses of Li that LiNO3 is

reduced by Li to LixNOy, which in turn oxidizes the sulfur species on Li.45,47

The oxidized sulfur species are effective to the passivation of Li because they

are stable components of the SEI, contrary to the reduced sulfur species.26,48

As introduced above, most of the reduced sulfur species, the lithium polysul-

fides, are soluble in the electrolyte. The insoluble Li2S can easily be oxidized

into polysulfides by disproportionation with the polysulfides in the catho-

lyte.45 This means the SEI on the metallic Li in a Li–S cell is much less robust

than that in a cell where the Li metal negative electrode is paired with a “stand-

ard” positive electrode, e.g. a Li+ insertion material.

Nevertheless, it has been demonstrated that LiNO3 is continuously con-

sumed in a Li–S cell,46 which means that the above-mentioned benefits are

temporary. As the passivation on the surface of Li deteriorates, more electro-

lyte is gradually decomposed.45 This is corroborated by the strong dependence

of a Li–S cell’s cycle life on the electrolyte volume.9,49,50 It is thus reasonable

to infer that metallic Li electrode may not always act as a non-limiting counter

electrode (CE) even when used in excessive dimensions. However, it is a com-

mon practice in the literature to investigate S/C composite electrodes in two-

electrode “half-cells” with a Li CE. The cell performance of such two-elec-

trode cells is often interpreted as the performance of the working electrode

(WE), which is the S/C composite electrode in this case. Hence, it is critical

19

to quantify the impact of the Li CE on the overall cell properties. To distin-

guish the electrochemical responses from WE and CE, a reference electrode

(RE) needs to be introduced. In chapter 4.2.2, a simple three-electrode setup

for the Li–S system will be proposed. With the Intermittent Current Interrup-

tion (ICI) method, the resistance contributions from both the positive and neg-

ative electrodes will be tracked at all SoC over the entire cycle life.

20

2 Scope of the Thesis

This thesis studied Li–S batteries during their operation in order to elucidate

the complex and changeable system. The ‘operando’ characterisations did not

only involve simultaneous materials characterisations, such as X-ray and neu-

tron scattering, but also real-time electrochemical characterisations, such as

resistance and specific surface area determinations, which were performed

while a Li–S cell was in operation. Two research questions were raised: What

mechanisms limits the discharge capacity of these cells? And what limits their

cycle life?

To investigate the reason for the discrepancy between the practical and the-

oretical capacities of the positive electrode, an experimental framework com-

bining operando X-ray diffraction (XRD) and the Intermittent Current Inter-

ruption (ICI) method was developed in Paper I, which renders the internal

and diffusion resistances. The evolution of the diffraction signal and the

transport properties were correlated directly to study the precipitation and dis-

solution processes of Li2S. The advantage of the combination was further

demonstrated in Paper II, where a series of operando XRD measurements

were carried out to evaluate the effects of several cell parameters on the pre-

cipitation of Li2S. This investigation aimed at providing the mechanistic ex-

planations for the strong dependence of the electrochemical properties of Li–

S cells on these parameters. Subsequently, the methodology was further de-

veloped for small-angle neutron and X-ray scattering, which captured the evo-

lution of the catholyte and amorphous discharge products, in addition to the

crystalline Li2S, in Paper III. The results indicated that a decrease in Li+ con-

centration in the catholyte inside the carbon matrix terminated the discharge

process.

In search of the limiting factor of the cycle life, the previously proposed

hypothesis of irreversible passivation of the carbon matrix in the positive elec-

trode was tested in Paper IV. An EIS-based technique was developed for in-

situ determination of the electrochemically active surface area, which indi-

cated that the decrease in discharge capacity is not correlated with the decrease

of electrode surface area. Lastly in Paper V, long-term cycling data of three-

electrode Li–S cells were reported. The resistance contributions from the pos-

itive and negative electrodes of an operational cell could thereby be measured

21

separately. The faster growth rate of the resistance of the metallic lithium elec-

trode suggested that electrolyte decomposition and/or the irreversibility of

lithium metal were the limiting factors for the cycle life.

22

3 Experimental Methods

3.1 Sample Preparation

The specifications of the materials used are listed in the corresponding articles.

3.1.1 Preparation of Sulfur/Carbon Composite Electrodes

The positive electrode was produced by following a previously optimized rec-

ipe for S/C composite electrode,51 where the sulfur content is 65% and the

porous carbon matrix is Ketjenblack (KB). First, sulfur was mixed with KB in

a mortar and heated to 155 °C for 20 minutes. Together with the other com-

ponents, the S-impregnated KB was ball-milled with ethanol or isopropanol

solution in deionized water. The slurry was cast onto a carbon-coated alumin-

ium foil, and then cut into 13-mm-diameter discs.

3.1.2 Two- and Three-electrode Electrochemical Cells

Unless specified otherwise, the two-electrode electrochemical cells in this the-

sis were made with R2025-format coin cells. Coin cells were chosen to avoid

the uncertainty of electrolyte amount resulting from the vacuum sealing pro-

cess for pouch cells. In contrast to Li-ion batteries, the electrolyte-to-sulfur

(E/S) ratio strongly affects the performance of a Li–S cell.49,52 Moreover, due

to the high specific capacity of both sulfur and lithium, the weight of electro-

lyte can easily dominate the overall specific energy.17,53 Another important

parameter is the sulfur-loading,51,54 which is defined as the mass of sulfur per

electrode area here. Due to the constant sulfur content in this work (65%), the

sulfur-loading is proportional to the thickness of the electrode.

Several three-electrode cells were employed in this thesis. In Paper I, a

three-electrode cell with a 16-mm-diameter Li metal disc as the reference elec-

trode was built with a homemade Swagelok® T-cell, as displayed schemati-

cally in Figure 3.2. In Paper V, a more robust three-electrode design was de-

veloped for long-term measurements. By inserting a polyimide-wrapped gold

microwire, minimal adjustment and thus minimal disturbance is introduced to

the commercial coin cell, as sketched in Figure 3.3.

23

Figure 3.1. Scheme of the components of a conventional coin cell

Figure 3.2. Scheme of the homemade three-electrode Swagelok® T-cell

Figure 3.3. Scheme of the homemade three-electrode coin cell

3.1.3 Electrochemical Cell Designs for Operando

Measurements

A scheme of the modified coin cell can be found in Figure 3.4. 7-mm holes

were drilled on both the stainless steel cap and can to avoid its high X-ray

absorption. The holes were sealed by polyimide tape on both sides in Paper

I, which was replaced by aluminium-coated polyimide in Paper II Similar

cell designs were also used in the Paper III with only adjustments on the

24

sealing material. Beryllium discs were used in place of the spacer owing to its

low X-ray absorption, rigidity to maintain a homogeneous stack pressure and

high electronic conductivity to establish a uniform current distribution. The

influence of the stack pressure and current distribution on the reactivity of the

electrode has been demonstrated computationally and experimentally.55 The

limitation of this cell design is the oxidation of beryllium at 4.2 V vs. Li/Li+,56

which is far above the maximum potential in the electrochemical tests carried

out in this work (2.6 V vs Li/Li+).

Figure 3.4. Scheme of the modified coin cell for operando X-ray-based measurements.

The modified cell for the neutron-based experiments share a design that is

similar to the above one. However, the windows are not made to increase the

transmission but to avoid the irregular scattering from the fine grains of stain-

less steel, which can vary substantially from sample to sample. In addition,

the cells have to be transported to the beamline at Institut Laue-Langevin in

France. Therefore, the sealing of the cells was made thicker.

3.2 Characterisation Techniques

3.2.1 Intermittent Current Interruption Method

The electrochemical cells in this work were mostly cycled at a constant current

but with periodic transient current pauses; a one-second pause every five

minutes. From the potential response to a current pause, resistance information

is derived. This technique of resistance measurement is termed the Intermit-

tent Current Interruption (ICI) method.49,57,58 The applied current is often la-

belled as the C-rate, where 1C equals 1672 mA gS-1.

Based on the porous electrode model proposed by de Levie,59 it is derived

in the supporting information of Paper I that the potential change (∆E) during

a current interruption can be expressed as the following function of step time

(t, defined as the time passed after the current is switched off). The mathemat-

ical relationship is shown in (3.1), where I is the current before the interrup-

tion, R´ and C are the solution resistance and double-layer capacitance per unit

length inside the pore, respectively.

25

∆𝐸(𝑡) = ∆𝐸(0) − 𝐼√4𝑅′

𝜋𝐶√𝑡 (3.1)

During the data analysis, a linear regression of the potential against the square

root of the step time is performed for each current interruption, as illustrated

in Figure 3.4. The intercept is the time-independent part of the potential

change (∆E(0)), which can be divided by the current to obtain the internal

resistance (R).

𝑅 = −∆𝐸(0)

𝐼(3.2)

The diffusion resistance coefficient (k) is acquired by dividing slope of the

regression line by the current.

𝑘 = −1

𝐼 ∙

𝑑∆𝐸

𝑑𝑡0.5(3.3)

By comparing (3.1), (3.3) and the mathematical expression of a Warburg ele-

ment used in the analysis of impedance spectra, as demonstrated in the sup-

porting information of Paper I, k can be derived to be proportional to the

Warburg coefficient (σ), which will be introduced in the next chapter.

𝑘 = √4𝑅′

𝜋𝐶= √

8

𝜋 𝜎 (3.4)

Figure 3.4. Example of the ICI analysis of a current interruption experiment. Data points and the regression line are plotted in red and blue, respectively. ∆E and t are the potential change and the step time as the current is switched off.

26

3.2.2 Electrochemical Impedance Spectroscopy1

The electrochemical impedance spectroscopy (EIS) is an alternating current

(AC) technique, which resolves the energy loss processes with different time

constants in an electrochemical system. In this work, the EIS is carried out by

enforcing sinusoidal potential E perturbation with a small amplitude |E| on the

cell.

𝐸 = |𝐸|𝑒𝑗𝜔𝑡 (3.5)

where j is the imaginary number, ω is the angular frequency and t is time. The

current response follows the general expression:

𝐼 = |𝐼|𝑒𝑗(𝜔𝑡+𝜃) (3.6)

where θ is the phase shift between the current and the potential. By applying

Ohm’s law, the impedance Z, which is analogous to the resistance under direct

current (DC), can be obtained.

𝑍 =𝐸

𝐼= |𝑍|𝑒−𝑗𝜃 (3.7)

𝑤ℎ𝑒𝑟𝑒 |𝑍| =|𝐸|

|𝐼|

Z can also be expressed in its real and imaginary parts.

𝑍 = 𝑍′ − 𝑗𝑍′′ (3.8)

𝑤ℎ𝑒𝑟𝑒 𝑍′ = |𝑍| cos 𝜃 𝑎𝑛𝑑 𝑍′′ = |𝑍| sin 𝜃

This procedure is repeated at various frequencies of interest, so a spectrum of

impedance is obtained and usually presented on a Nyquist plot, where -Z´´ is

plotted against Z´. Equivalent circuit models (ECM) are used to analyse the

impedance spectra. Two of the basic elements are a resistor with a resistance

of R and a capacitor with a capacitance of C, whose impedance ZR and ZC

contains only real and imaginary parts, respectively.

𝑍𝑅 = 𝑅 (3.9)

𝑍𝐶 =1

𝑗𝜔𝐶(3.10)

27

It is intuitive to visualise that ZR is a point on the Z´-axis and ZC is a line

overlapping the Z´´-axis. To fit the spectrum of an electrochemical system,

many or even infinite combinations of resistors and capacitors with different

resistance and capacitance, respectively, are used in series or in parallel. Com-

monly adopted complicated combinations can be found as individual elements

in the literature, such as the infinite Warburg element, which is used to de-

scribe diffusion processes and the retarded electrochemical double-layer ca-

pacitance in a porous electrode.59 Its impedance ZW can be expressed as:

𝑍𝑊 = 𝜎𝜔−0.5 − 𝑗𝜎𝜔−0.5 (3.11)

where σ is the coefficient of the Warburg element. Details on how it describes

a porous electrode can be found in the Supporting Information of Paper I.

In Paper IV, the analysis of impedance spectra was focused on the capacitor

at low frequencies, which is mostly contributed by the electrochemical dou-

ble-layer on the carbon matrix. With the specific capacitance of carbon, the

electrochemically active surface area can be calculated from the capacitance

derived from the impedance measurement.

3.2.3 X-ray Diffraction/Wide-angle X-ray Scattering

X-ray diffraction (XRD), also known as wide-angle X-ray scattering

(WAXS), is a materials characterisation technique that analyses the elastically

scattered photons with larger momentum transfer, which are attributed to the

diffraction phenomenon of crystals. As an array of atoms is illuminated by

photons with a wavelength close to their spacing, each atom behaves like a

point light source and the light from the atoms in the array can intensify or

cancel out each other in specific directions. The collective outcome of the

point light sources is termed diffraction. The relationship between the lattice

plane spacing (d) and the scattering angle (θ) is described by Bragg’s law:

𝑛𝜆 = 2𝑑 sin 𝜃 (3.12)

where n is a positive integer and λ is the wavelength of the incident X-ray. The

intensity of each peak is determined by the elements sitting on the planes in

the array since each element scatters X-rays differently. Generally, atoms with

more electrons interacts more with the electromagnetic wave, but this rela-

tionship is also wavelength-dependent. In summary, each crystalline material

has its own unique diffraction pattern, which can be used to identify it.60

In chapter 4.1, XRD is applied to an operating battery. The coin cell has to

be modified to avoid the high X-ray absorption of the stainless steel casing, as

explained previously in chapter 3.1.3. Through operando XRD, the evolution

of the crystalline active materials can be followed as the cell goes through

different SoC levels. For Li–S batteries, as mentioned above in chapter 1.2.1,

28

the active material of the positive electrode dissolves in the electrolyte as dis-

charging/charging starts, and precipitates at the end of discharge/charge. Since

both elemental sulfur and Li2S are crystalline, operando XRD is applied here

as a tool to track the process of precipitation and dissolution.

3.2.4 Small-angle X-ray and Neutron Scattering

Small-angle scattering (SAS) techniques also analyse the results from elastic

scattering, just like its wide-angle counterpart, but with small momentum

transfer, which are contributed by scatterers larger than atoms. Depending on

the specifications of the equipment, the probing scale can range from nano- to

micrometers.61 The momentum transfer (q) can be expressed in terms of scat-

tering angle (θ) and the wavelength (λ) of the incident beam:62

𝑞 =4𝜋 sin 𝜃

𝜆(3.13)

The scattering intensity (I) follows a general form:63

𝐼(𝑞) = 𝑁𝑉2∆𝜌2𝑃(𝑞)𝑆(𝑞) (3.14)

where N is the number of scatterers, V is the volume of a scatterer, ∆ρ is the

contrast of the scattering length densities (SLD) of the scatterers and their sur-

rounding medium, P is the form factor and S is the structure factor. In chapter

4.1.3, SAS experiments were conducted with both X-ray and neutrons. More-

over, cells with hydrogenous and deuterated electrolytes were probed by the

small-angle neutron scattering (SANS). These combinations rendered three

SLD contrasts and thus sufficient information for a definite interpretation of

the system. The simultaneously measured WAXS signal enabled the compar-

ison between crystalline and amorphous solid clusters in the system.

29

4 Results and Discussions

4.1 Illustration of a Discharging Positive Electrode

This chapter dives into the details of the discharge processes inside an operat-

ing S/C composite electrode and aims to provide a holistic picture of how

these processes are steered by the electrochemical cell parameters and how

they, in turn, impact the electrochemical properties of the cell.

4.1.1 Precipitation and Energy Loss Processes

In Paper I, an experimental framework of concurrent detection of the crystal-

line species and resistance measurements is established by combining oper-

ando XRD and the ICI method. The influence of the precipitation of the insu-

lating solid species on the functionality of the porous carbon matrix will be

exhibited.

4.1.1.1 Electrochemical equivalence of the modified coin cell

To justify the resistance measurement, the electrochemical properties of the

modified coin cell was first benchmarked against those of conventional coin

cells. The potential (E), internal resistance (R) and diffusion resistance coef-

ficient (k) of the modified and conventional cells are plotted in Figure 4.1. It

can be noticed that the electrochemical properties of the modified coin cell,

which has 10 µL gs-1 electrolyte, laid between those of the two conventional

coin cells with 6 and 10 µL gs-1 electrolyte. This suggests that the effective

loss of electrolyte was smaller than 4 µL gs-1. Therefore, it can be concluded

that the modified coin cell presented in this work had electrochemical proper-

ties that are comparable to a standard Li-S coin cell with a high S-loading

(>4 mgS cm-2), a moderate reversible specific capacity (~800 mAh gS-1) and a

relatively lean electrolyte volume (6–8 µL mgS-1). The sealing of the modified

cells were improved in Paper II and III, which enabled operando XRD meas-

urements with 6 µL mgS-1. Similar comparisons between the modified and

conventional cells are presented in the Supporting Information of the corre-

sponding manuscript.

30

Figure 4.1. Potential (E), internal resistance (R) and diffusion resistance coefficient (k) of the operando XRD cell (green) and two other unmodified coin cells with 6 (blue) and 10 (red) µL mgS

-1 electrolyte, respectively. The electrochemical parameters from the selected cycle (marked on the top of the panels) are plotted against the state of charge (SoC) represented by specific charge (Q). A complete plot with data from every cycle can be found in the supporting information of Paper I.

4.1.1.2 Operando XRD with concurrent resistance measurement

Figure 4.2 displays an overview of the operando XRD results with the elec-

trochemical properties plotted below it. The colour of the segments of the

curves of the potential, internal resistance and diffusion resistance coefficient

corresponds to the sulfur species present in the XRD pattern at the same time.

The initial orthorhombic α-S disappeared as the upper discharge plateau ended

in the first C/50 discharge. The internal resistance, which is the sum of solu-

tion and charge transfer resistance,49 decreased by a factor of one third as the

dissolution of α-S freed up the conductive surface in the porous carbon matrix

of the positive electrode. The internal resistance subsequently rose to a peak

situated at the SoC level between the two plateaus, which can be attributed to

the increased solution resistance caused by the high polysulfide concentration

at this SoC level,49,64,65 where no crystalline sulfur species were detected by

XRD. After one third of the lower discharge plateau, broad peaks of Li2S

31

started to appear in the XRD patterns. These can be ascribed to the very small

grain size and/or low crystallinity of Li2S according to previous work.66–68

Figure 4.2. Operando XRD data plotted against time. The XRD patterns are displayed as a heat map at the top, followed by the potential (E), internal resistance (R) and diffusion resistance coefficient (k). Error bars for R and k are estimated by the stand-ard deviation from the linear regression.

As soon as the cell was charged at C/25 in the first cycle, both the internal

resistance and diffusion resistance coefficient dropped immediately, although

32

there was still considerable intensity of the diffraction peaks of Li2S. This be-

haviour can be found in every cycle and corresponding data has also been

reported in EIS studies of Li-S cells.49,69,70 At the end of charging, elemental

sulfur crystallized in the monoclinic structure (β-S) instead of the initial or-

thorhombic structure (α-S). The formation of the thermodynamically less fa-

vourable allotrope have been frequently observed in previous operando XRD

works67,68,71 and has then been argued to be a consequence of precipitation

kinetics68.

The C-rate for both discharge and charge was then switched to C/10 for the

remainder of the cycles. After the second cycle, the internal resistance de-

creased and stabilized around the fourth cycle. The diffusion resistance coef-

ficient stayed roughly the same, but the error bars were much smaller after the

second cycle due to the larger current.

A trend in the diffusion resistance can be observed in Figure 4.2 after the

fourth cycle, where the profile of the internal resistance also had stabilized.

As Li2S appeared in the XRD patterns, the coefficient of diffusion resistance

jumped from below 5 Ω s-0.5 cm2 to a plateau around 12 Ω s-0.5 cm2, and there-

after increased rapidly again at the end of discharge. Upon charging, it

dropped immediately below 5 Ω s-0.5 cm2; the same as its value when there

was no Li2S detected. Since the diffusion resistance reflects the transport prop-

erties in the positive electrode, it is not surprising that it is influenced by the

evolution of Li2S. Nevertheless, it is interesting that the effect is asymmetric

upon discharging and charging.

4.1.1.3 Impact of Li2S precipitation on the transport properties inside

the porous carbon matrix

To further analyse the correlation between the diffusion resistance and the

amount of Li2S, the diffusion resistance coefficient is plotted against the inte-

grated peak area of the 111 reflection of Li2S in Figure 4.3. As shown in Equa-

tion (3.4), the diffusion resistance coefficient increases when the solution re-

sistance (R´) inside the porous carbon increases and/or the capacitance (C) of

the carbon surface decreases. The morphology of Li2S growth on open struc-

ture carbon matrices, e.g. carbon nanotube scaffolds, has previously been stud-

ied by SEM.35,37 These studies suggested that the precipitation proceeds as a

two-dimensional growth covering the surface of the conductive carbon after

nucleation. Therefore, an explanation for the slow increase of the diffusion

resistance coefficient when Li2S started to appear in the XRD could be the

drop in capacitance, stemming from the progressive coverage of the surface

of the pores in the carbon matrix.

As the Li2S phase continued to grow, more and more volume fraction of

the pores was occupied by insulating particles, instead of ion-conducting elec-

trolyte. This resulted in the increase in the solution resistance inside the pores

(R´) and consequently the diffusion resistance coefficient (k). On prolonged

33

discharge, the precipitates of Li2S may block the pores and the channels be-

tween them. Therefore, the diffusion resistance increased sharply, leading to

the termination of discharge. This reasoning is supported by several operando

spectroscopic experiments, which have demonstrated the presence of polysul-

fides in the positive electrode at the end of discharge.68,72,73

Figure 4.3. Diffusion resistance coefficient (k) plotted against the integrated area of the 111 reflection of Li2S. Data are from the fourth to the eighth cycles.

In summary, Paper I demonstrated the value of coupling a materials charac-

terisation technique with simultaneous resistance measurements. More de-

tailed discussion about the combined observations are elaborated in Paper II.

4.1.2 From Cell Parameters to Electrochemical Performance

On the basis of the experimental framework developed in Paper I, the work

in Paper II applied the methodology to a series of Li–S cells to understand

how the Li2S precipitation is altered by different parameters of the cells and

how the precipitation in turn impacts their electrochemical properties. The

sealing of the modified coin cells was improved by using aluminium-coated

polyimide films, which rendered comparable electrochemical performance of

the operando cells with the same E/S ratio as the commercial coin cells, as

demonstrated in the Supporting Information of Paper II. The integrated in-

tensity from the uniform polyimide films also served as a normalising stand-

ard, which allowed a comparison between multiple cells. Moreover, the higher

34

X-ray transmission of the new sealing enabled the collection of an XRD pat-

tern every 15 minutes, which is one fourth of the time required previously.

4.1.2.1 Effect of the sulfur-loading and E/S ratio

Four cells with two sulfur-loadings and two E/S ratios were fabricated with

the otherwise identical components to cross-examine the influence of both pa-

rameters. In contrast to Li-ion batteries, the electrode thickness and the elec-

trolyte amount have been well-recognised to be important factors in the spe-

cific capacity, which is proportional to the utilisation of active material, of the

Li–S system, even at low C-rates.20,51,54,74,75 However, their influence on the

electrochemical properties has not been fully mechanistically explained.

Two major sources of resistances appearing when a Li–S cell discharges

are the solution resistance which increases with the polysulfide concentration,

and the charge-transfer and/or diffusion resistances raised by the precipitation-

related processes. The viscosity of the electrolyte increases with the concen-

tration of polysulfides in the catholyte, which has its maximum between the

two discharge plateaus (see Figure 1.1).76 The effect of the precipitation is

more complicated. Initially, it actually reduces the solution resistance since it

removes polysulfides from the electrolyte, which can be observed in the de-

crease of internal resistance right after the voltage entered the lower plateau;

see Figure 4.4. However, the subsequent formation of Li2S deteriorates the

carbon matrix, as discussed in the previous chapter.

At the same sulfur-loading, the increase in E/S ratio decreases the concen-

tration of the polysulfides in the catholyte, which can be observed by the lower

peak in the internal resistance between the two discharge plateaus. The lower

polysulfide concentration also leads to a lower rate and amount of precipita-

tion. With less insulating precipitates, the ionic transport inside the porous

electrode remains undisturbed, as indicated by the low diffusion resistance

coefficient until the sudden increase at the end of discharge. However, the low

polysulfide concentration in the electrolyte could cause a more difficult re-

trieval of active material, resulting in a rapid increase in diffusion resistance.32

The mass transport overpotential may also have terminated the discharge pro-

cess earlier in comparison to the cell at a lower E/S ratio, and thus leading to

lower specific capacity.

35

Figure 4.4. Potential (E), internal resistance (R), diffusion resistance coefficient (k), normalised intensity of the 111 peak of Li2S (I) and its rate of changing with respect to the total charge passing through the cell (dI/dq) for cycle 5 plotted against the spe-cific capacity (Q). Data on discharge and charge are in blue and red, respectively. See Paper II for a detailed description.

The effect of sulfur-loading was more obvious at E/S = 6. Comparing the first

and third columns in Figure 4.4, the major difference can be identified as the

internal resistance. Since the precipitation rate, reflected by dI/dq, is similar

in both cells, the difference in the internal resistance, especially on the lower

discharge plateau, could be caused by the non-uniform reaction progress in

the axial direction of the electrode with higher sulfur-loading. The shorter up-

per discharge plateau of the cell with higher loading may also suggest an in-

complete dissolution process as a result of less homogeneous reactions across

the electrode.

36

4.1.2.2 Effect of the specific current

The impact of the specific current on the formation of Li2S can be found in

Figure 4.5.

Figure 4.5. Potential (E), internal resistance (R), diffusion resistance coefficient (k), normalised intensity of the 111 peak of Li2S (I) and its rate of changing with respect to the total charge passing through the cell (dI/dq) of the cell ‘S-L=2.5 E/S=6’ cycled at C/10, C/5, C/2 and C/20 are plotted against the specific capacity (Q). Data on dis-charge and charge are in blue and red, respectively. See Paper II for a detailed de-scription.

The first cell in Figure 4.4, ‘S-L=2.5 E/S=6’, was cycled at different C-rates.

An obvious feature in the results at C/2 was the missing lower plateaus. The

surge of the internal resistance could result from a more extreme case of non-

uniform reaction progression. It could be speculated that the Li+ transport in-

side the positive electrode cannot catch up with the high electronic current, so

37

the polysulfides on the separator have to be reduced further into shorter-

chained polysulfides, which causes the solution resistance to rise substantially.

A more mild case could happen at C/5, where the internal resistance between

the two discharge plateaus is substantially higher than that at C/10. The as-

sumption of the transport limitation of Li+ is supported by the peak of the dif-

fusion resistance coefficient at the same SoC.

At C/20, an increasing trend in dI/dq could be observed, which suggested

that the precipitation is accelerated towards the end of discharge. This indi-

cated a more homogeneous reaction progress in the positive electrode, where

the polysulfides had a narrow distribution of oxidation states. Since the for-

mation of Li2S should be the final step of the reduction, the precipitation was

concentrated at the end of discharge.

4.1.2.3 Effect of the electrolyte salt with a high donor number

In the last part of Paper II, the influence of the electrolyte salt was studied by

replacing 50% of LiTFSI with LiBr. It has been previously reported that LiBr

promotes the disproportionation reactions between polysulfides, which en-

hances sulfur utilisation in a non-encapsulating carbon matrix,33 as discussed

in detail in the manuscript. Here, this effect was tested in an S/C electrode

optimised for the specific surface area. Despite the higher precipitation rate,

the discharge capacity was not higher than for the cell with standard electro-

lyte. As demonstrated by the larger I/qd,l values of the cell with LiBr in Table

4.1, more Li2S was formed for the same amount of charge passed during the

lower discharge plateau. This could explain the high overpotentials observed

from the cell.

38

Table 4.1. Comparison of the specific discharge capacity (Qd,l) and total discharge capacity (qd,l) of the lower plateau, integrated intensity of the 111 reflection of Li2S (I), full-width half-maximum (FWHM) of the reflection from the second last meas-urement during discharge (to avoid current switching during the measurements) and ratio of the intensity to the total discharge capacity of the lower plateau (dI/dqd,l) of the cells. Standard deviations are in brackets. See Paper II for the full table.

Cell C-rate Qd,l /

mAh gS-1

qd,l /

mAh

I / a.u. FWHM / ° I/qd,l /

mAh-1

S-L=2.5

E/S=6

C/10 725 2.434 11.5(3) 1.22(2) 4.7

C/5 613 2.060 11.1(3) 1.36(3) 5.4

C/2 0 0 0 n/a n/a

C/20 761 2.556 13.3(3) 1.19(2) 5.2

S-L=2.5

E/S=10 C/10 627 2.091 8.2(2) 1.10(3) 3.9

S-L=3.3

E/S=6

C/10 710 3.140 14.4(3) 1.16(1) 4.6

C/10 724 3.203 14.8(3) 1.22(1) 4.6

S-L=3.4

E/S=10 C/10 529 2.378 5.4(2) 1.11(4) 2.3

S-L=3.3

E/S=6

TFSI/Br

C/10 515 2.285 15.4(3) 1.31(2) 6.7

C/20 613 2.718 19.0(3) 1.32(1) 7.0

C/13 470 2.085 14.1(3) 1.38(2) 6.8

39

4.1.3 Dissolved, Crystalline and Amorphous Sulfur Species

In Paper III, the experimental framework used in the previous two papers

was expanded by incorporating small-angle X-ray and neutron scattering to

characterise the discharge process at both the micro- and mesoscales.

4.1.3.1 Small-angle scattering data in three contrasts

As mentioned in chapter 3.2.4, the intensity of the scattering depends on the

contrast of SLD between the scatterer and its surrounding. In addition to the

intrinsic different contrasts with X-ray and neutron, the SANS experiments

were carried out with deuterated and hydrogenous electrolytes. These cells

were referred to as “SANS-D” and “SANS-H” cells, respectively, while the

SAXS measurements was conducted on the “SAXS” cell. The scattering data

from the three cells are displayed in Figure 4.6.

Figure 4.6. Azimuthally averaged SAS data from the first discharge of the cells a) SANS-D, c) SANS-H and d) SAXS and b) the first charge of the SANS-D cell. The intensity (I) was plotted against the momentum transfer (q). The measurements of the SANS-H and SAXS cells during the first charge can be found in the Supporting In-formation of Paper III.

40

In spite of the obvious variations in the scattering curves from the SANS-

D cell at different SoC, the scattering results from the SANS-H cell were prac-

tically constant throughout the SoC. Based on the SLD diagrams in Figure 4.7,

the lack of change from the SANS-H cell indicated that the structure of the

carbon matrix is not altered by the precipitation. Although the variations in

the scattering curves of the SAXS cell were not as apparent as those of the

SANS-D cell, the SAXS measurements were normalised to the transmission.

Thus, the changes could still be analysed when the background was sub-

tracted, as shown in Figure 4.8.

Figure 4.7. Schematic diagram of scattering length density of the main components of the positive electrode with a) neutron and b) X-ray (λ = 1.54 Å) at various states of discharge or charge. See Paper III for the full description.

Figure 4.8. a) The intensity difference (∆I) between the data from selected states of discharge in the first cycle and the first measurement in the SANS-D cell plotted against the momentum transfer, q. b) The square of the intensity difference (∆I2) be-tween the data from selected states of discharge in the first cycle and the first meas-urement in the SAXS cell plotted against q. ∆I2 was plotted because ∆I is mostly neg-ative above 0.1 Å-1. See Paper III for the ∆I-q plot.

41

It can be observed in Figure 4.8 that the scattering results from both SANS-

D and SAXS cells show two intensity contributions, one above 0.1 Å-1 and the

other between 0.1 and 0.3 Å-1. Since the contribution at lower q-range was

positive in SANS results but negative in SAXS measurements, it was not the

same as that at higher q-range, which contributed positively in both SANS and

SAXS. Therefore, a model with two independent spherical objects was used

to fit the data from both SANS-D and SAXS cells. The detailed description of

the model could be found in the experimental method of Paper III.

4.1.3.2 Interpretation of the two-sphere model

The parameters from the fitting results of the data from SANS-D and SAXS

cells are demonstrated in Figures 4.9 and 4.10.

Figure 4.9. The scale factor, scale; radius, r; potential, E; internal resistance, R; and diffusion resistance coefficient, k, of the positive electrode of the cell SANS-D during discharge and charge cycles. A two-sphere model was applied to fit the intensity dif-ference, ∆I, between the SANS data for each time interval and the first measurement. The parameters of the two sphere distributions are plotted in different colours and symbols. The scale factor here is the product of a term that would correspond to the number density of the objects that give rise to scattering and the square of the differ-ence in scattering length density, ∆ρ2. The first discharge and charge were at C/20 and C/10, respectively. The second and third cycles are at C/10 and C/5, respectively.

42

Figure 4.10 The scale factor, scale; radius, r; the integrated intensity from the 111 diffraction peak of Li2S, IWAXS; the potential, E; internal resistance, R; and diffusion resistance coefficient, k, of the positive electrode of the cell SAXS. A two-sphere model was applied to fit intensity differences, ∆I, between the SAXS data in each interval and the first measurement. The respective parameters for each distribution of spheres are plotted in a specific colour and shape. The scale factor here is the product of a number concentration of scattering objects and the square of the difference in scattering length density, ∆ρ2. The cell was discharged at C/50 and charged at C/25 in the first cycle, and subsequently cycled at C/10.

By comparing the second cycle from the SANS-D cell and the third to fifth

cycle of the SAXS cell, which were all cycled at C/10, the radius of the smaller

sphere, which are plotted in green squares, is consistent with both fitted re-

sults. The scale factor also follows the same trend, but the values are larger in

43

the fitting of the SANS results due to the different contrasts. The trend in the

scale factor of the smaller sphere matches that in the intensity from WAXS.

Thus, it could be concluded that the small sphere in the model accurately de-

scribed the Li2S precipitates.

The larger sphere, whose parameters are plotted in orange triangles,

showed a drop in scale factor but increase in radius at the end of second in

Figure 4.9. The same trend could also be observed in the results from SAXS

although the magnitudes were smaller. The larger sphere could thereby be as-

signed to the mesopores of the carbon matrix filled with catholyte. The evolu-

tion of the larger sphere near the end of discharge could then be attributed to

a drop in Li+ concentration, which would result in the dashed lines in the

scheme of SLD in Figure 4.7. Such shifts in SLD, both for X-ray and neutron,

would be consistent with the evolution of the larger spheres in Figures 4.9 and

4.10.

4.1.3.3 What stops the discharge process?

With the two spheres in the model assigned, further analysis could be carried

out to determine which parameter is correlated to the diffusion resistance of

the positive electrode. In Figure 4.11, the diffusion resistance coefficient was

plotted against the scale factors of the smaller and the larger spheres. In cycles

3–5, which showed reproducible features at C/10, the correlation between the

diffusion resistance coefficient and the scale factor of the smaller sphere re-

sembled the correlation between the diffusion resistance coefficient and the

diffraction/WAXS intensity from Li2S, as shown in Figure 4.3. Both indicated

that there was a sudden increase in the diffusion resistance when the amount

of Li2S reached a certain level. In the first manuscript, this correlation was

interpreted as the parabolic increase in k when the volume fraction of Li2S in

the pores grows linearly, as elaborated in Paper I.

Figure 4.11. The diffusion resistance coefficient (k) plotted against a) the scale factor for the smaller sphere, scaleB, and b) the scale factor for the larger spheres, scaleA, from the fitting results for the SAXS data shown in Figure 4.10.

44

Here, with the additional information on the catholyte in the mesopores, Fig-

ure 4.11b showed a more direct correlation between the drop in Li+ concen-

tration in the mesopores and the increase in the diffusion resistance. This

strongly suggested that a lack of Li+ inside the carbon matrix terminated the

discharge process at a medium C-rate, e.g. C/10 and C/20 in this study. How-

ever, this does not rule out the deterioration of the transport properties inside

the positive electrode caused by the formation of insulating Li2S particles. In

fact, the worsened transport properties could also contribute to the Li+ defi-

ciency inside the mesopores at the end of discharge.

In the first discharge, where the current was lower at C/50, of the SAXS

cell, the decreased slope in the diffraction intensity from Li2S and the linear

increase of the scale factor of the smaller sphere near the end of discharge

indicate the formation of amorphous solid discharge solid products, e.g. Li2S2,

which was proposed in previous operando XRD studies.36 The formation of

discharge products which require less Li per unit of S than Li2S, supported the

above observation of Li+-deficiency near the end of discharge.

Figure 4.12. Scheme summarising the discharge and charge processes during the lower voltage plateau, as described above. The approximated states of discharge (SoD) and charge (SoC) are shown at the bottom. The colour of the electrolyte inside the pores of the carbon indicates the Li+ concentration, which starts to drop inside the small pores. At a medium C-rate, C/10, the discharge process is stopped by the defi-ciency of Li+ inside the mesopores, whereas, at a low C-rate, C/50, less Li+-demanding LixSy forms and more discharge capacity can be delivered.

45

4.2 Investigations into the Cycle Life Limitations

The work in this chapter developed techniques to probe an operating Li–S

along its cycle life. Although the electrochemical methods are inherently “in-

situ” for a battery, attention needs to be paid to minimise the disturbance of

the measurements to the operation of the system, so that the methods can ap-

proach the full definition of “operando”.

4.2.1 Electrochemically Active Surface Area over Cycling

4.2.1.1 Method development in symmetric cells

As mentioned in the experimental methods, in this subchapter, a method for

quantifying the electrochemically active surface area of the positive electrode

was developed by characterising the capacitor element in the low frequency

range of the impedance spectra. Since the capacitance per unit area of carbon

is typically 10 µF cm-2,77 the capacitance derived from the impedance meas-

urement can be converted to the electrochemically active surface area.

The method was first verified in symmetric cells with electrodes made with

carbon nanotubes (CNT), which have well-defined pore structure. The sym-

metric cells avoided the interference from the redox reactions and thus sim-

plified the equivalent circuit. The porous electrodes were fitted with the de

Levie pore finite element (Ls) in the ZView software, whose impedance ZLs

can be written as:

𝑍𝐿𝑠 =𝑅𝐿𝑠

Λ0.5cot(Λ0.5) (4.1)

𝑤ℎ𝑒𝑟𝑒 Λ =1

𝐴𝐿𝑠+ 𝐵𝐿𝑠(𝑗𝜔)𝜙𝐿𝑠

In addition to the symbols introduced in chapter 3, RLs is the solution resistance

inside the pores, BLs is a parameter proportional to the electrode effective ca-

pacitance, ALs is a parameter proportional to the charge-transfer resistance and

φLs is the constant phase exponent, which accounts for the nonideality of the

capacitor. Since there should not be charge-transfer reactions in the symmetric

cells, ALs was set to 1020 in this work, so the element is equivalent to the finite

space Warburg element. The impedance spectra from the symmetric cells

could then be fitted with the ECM in Figure 4.13a, which also includes other

elements to account for the impedance from the electrolyte and the metallic

current collectors. The derivation of the capacitance from the above parame-

ters are elaborated in Paper IV.

46

a) b)

Figure 4.13. Equivalent circuits used to fit impedance data, and their corresponding impedance response. a) Resistor coupled, in series, to a combination of a constant-phase element (CPE) and a resistor in parallel, and then coupled, in series, to a de Levie pore finite length element with infinite charge transfer resistance. b) Resistor coupled, in series, to a combination of a CPE and a resistor in parallel. For simplicity, the impedance plots illustrate the ideal behaviour with unity values of the exponent of the CPE and de Levie elements.

Although fitting the spectra with the ECM in Figure 4.13a was theoretically

thorough, it involved several parameters that were not really required for ex-

tracting the capacitance of the electrode. Alternatively, the fitting could be

restricted to the low frequency range, where the capacitive behaviour was

dominant. The other impedance contributions at higher frequencies could then

be reduced to a resistor while the capacitive tail was fitted as a constant phase

element (CPE), whose impedance is:

𝑍𝐶𝑃𝐸 =1

𝑄𝐶𝑃𝐸(𝑗𝜔)𝑃𝐶𝑃𝐸(4.2)

where QCPE is the capacitance parameter in the unit of F s-(1-PCPE) and PCPE is

the exponent. It can be observed that if PCPE equals one, Equations 4.2 will be

analogous to Equation 3.10, which shows the impedance of a capacitor. A

resistor (Rct) was added in parallel to the CPE to account for the very slow

redox reactions that may be present in the Li–S cells, which are discussed be-

low. The resulting ECM is displayed in Figure 4.13b. The double-layer capac-

itance Cdl of the CPE in the ECM could be calculated by:78,79

𝐶𝑑𝑙 = 𝑄𝐶𝑃𝐸

1𝑃𝐶𝑃𝐸 (

𝑅𝑡𝑜𝑡𝑅𝑐𝑡

𝑅𝑡𝑜𝑡 + 𝑅𝑐𝑡)

1−𝑃𝐶𝑃𝐸𝑃𝐶𝑃𝐸

(4.3)

In the symmetric cells, since two identical electrodes were in serial, the capac-

itance of each electrode was twice the value of Cdl calculated above. A series

of CNT electrodes with and without sulfur infiltration were coated on several

different metallic current collectors and characterised by this method. It can

47

be observed in Figure 4.14 that all the electrodes showed a specific surface

area (SSA) close to the value obtained from a S/CNT electrode by nitrogen

adsorption-based method (based on Brunauer–Emmett–Teller (BET) theory).

This indicated that some pores, most likely the micropores (<2 nm), which are

accessible to the small nitrogen molecules, are not approachable by the larger

ions in the electrolyte.

Figure 4.14. Results of the specific surface area (SSA) of carbon in CNT and S/CNT electrodes, coated on different substrates, as indicated. The results of the BET specific surface areas are also shown to facilitate comparison. The specific surface area values are normalised by the mass of carbon, and the carbon loading values are normalised by the geometrical area of the electrode. See Paper IV for the complete description.

4.2.1.2 Mismatching trends of electrochemically active surface area

and the discharge capacity

The execution of the above method in a Li–S cell is much simpler if the im-

pedance of the redox reactions is excluded. Therefore, a test protocol was de-

signed to cycle the cell at a constant current between 1.8 and 2.6 V but raise

the potential to 2.8 V every five cycles when the cell reached 2.6 V. The in-

crease in potential should minimise the residual polysulfides in the electrolyte

and thus the redox reactions. Moreover, the protocol ensured that the imped-

ance measurements were conducted at the same SoC and that the deviation

from a standard cycling program is minimal. The electrochemically active sur-

face area of the positive electrode was tracked over 160 cycles, as demon-

strated in Figure 4.15. The SSA decreased in the initial cycles but stabilized

above 200 m2 gC-1 from the 50th cycle to the end of the cycling. The specific

capacity of the cell followed the same trend in the beginning but showed a

continuous decrease since the 80th cycle. The stable SSA during the capacity

fade indicated that there was little correlation between them. This observation

demonstrated that the positive electrode was not irreversibly passivated by the

48

insulating sulfur species, e.g. Li2S, as the discharge and charge processes re-

peated. Therefore, the decrease in capacity, which seems to occur without any

extensive passivation, should be caused by other factors.

Figure 4.15. a) Specific surface area (SSA) of carbon in the optimised S/C electrode formulation used in a Li-S cell (with carbon loading of 1.79 mgC cm-2). b) Cycle per-formance of the same Li-S cell.

49

4.2.2 Resistance Distribution and Its Evolution

As mentioned in chapter 1.2.2, the metallic Li is not an ideal counter electrode,

even when it is in excess amount. The previous chapter also revealed that the

positive electrode was not passivated irreversibly. In Paper V, a three-elec-

trode cell setup for the Li–S system was developed to resolve the resistance

contributions from the positive and negative electrodes, which could locate

where the deterioration of the system originates.

4.2.2.1 Three-electrode setup and the initial resistance distribution

The three-electrode cell is made by a simple modification of the conventional

coin cell, as sketched in Figure 3.3. A polyimide-wrapped gold microwire

(ø50 µm) was placed between two separators in the middle of the positive and

negative electrodes and then fit in the groove of the polypropylene gasket of

the coin cell before the cell was closed. The first method to employ the gold

wire as the reference electrode (RE) was to lithiate it to form LixAu (x = 0–1),

which had a stable potential of 0.31 V vs Li/Li+ and could thus be used as a

true RE.80,81 Alternatively, it was found that the gold wire could also serve as

a pseudo-RE in a Li–S cell without additional treatments thanks to the depo-

larising nature of the polysulfides in the electrolyte. Although pseudo-REs

may cause artefacts in the high frequency range of impedance measurements,

the resistance measured here by the ICI method is not affected. In addition,

the pseudo-RE showed better long-term stability since it avoided the reactions

with polysulfides that LiAu encountered at 0.31 V vs Li/Li+.

It may be worth mentioning that several other three-electrode cell designs

were tried, but there were difficulties in either controlling the electrolyte

amount at cell assembly or preserving the electrolyte during cycling. A com-

parison of the electrochemical properties of several cell setups can be found

in the Supporting Information of Paper V. The above issues might explain

why there has not been reports on long-term cycling of a three-electrode Li–S

cell in the literature.

The resistance distribution between the positive and negative electrodes in

the 10th cycle of four cells with different experimental parameters is exhibited

in Figure 4.16. The positive and negative electrodes were referred to as work-

ing electrode (WE) and counter electrode (CE) in the manuscript since the cell

is strictly speaking a half-cell, of which the CE is larger than the WE. The

cells were labelled with the E/S ratio and the type of the RE. From the two

columns on the left, it could be observed that the profiles of internal resistance

and diffusion resistance coefficient measured by the LiAu RE and the Au

pseudo-RE were in good agreement, despite the difference in the relative val-

ues of the potential. At a higher E/S ratio, e.g. 10, the internal resistance was

evenly distributed between the two electrodes while the diffusion resistance,

mostly at low SoC, was predominantly contributed by the S/C composite elec-

trode. The same trend in diffusion resistance was also shown by the cells with

50

an E/S ratio of 6, although the diffusion resistance coefficient of the WE

started to increase earlier during discharge. The reasons for this phenomenon

were explored previously in chapter 4.1.2. The internal resistance was also

dominated by the contributions from the S/C electrode during the lower dis-

charge plateau at E/S = 6.

Figure 4.16. Potential (E), internal resistance (R) and diffusion resistance coefficient (k) of the cell, working (WE) and counter (CE) electrodes in the 10th cycle. The E/S ratio, RE of each cell are marked. The sulfur-loading of the cell marked with ‘HL’ is 3.2 mgS cm-2 while that of the other three cells is 2.5 mgS cm-2. R and k of the WE, CE and the cell in all cells are higher during discharging than during charging, except for R at the end of charging of the two cells with E/S = 6. Note that the CE potential is approximately 0 V – ERE, and the cell voltage is identical for all cell.

51

4.2.2.2 Resolving resistance growth from the positive and negative

electrodes along cycle number

The evolution of the resistance contributions from both WE and CE in the first

cell in Figure 4.16 is demonstrated in Figure 4.17. From the 10th cycle in Fig-

ure 4.16 to the 60th cycle here, the resistance profiles did not vary noticeably.

However, the internal resistance of CE started to rise in the 70th cycle while

that of the WE remained at the same level. A substantial increase in the diffu-

sion resistance of the CE could also be identified near the end of discharge.

Both of these resistance growths on the CE continued in the 80th cycle, but the

RE did not function completely properly, as indicated by the negative diffu-

sion resistance coefficients.

Figure 4.17. Potential (E), internal resistance (R) and diffusion resistance coefficient (k) of the cell, working (WE) and counter (CE) electrodes in the 50th, 60th, 70th and 80th cycle of the cell with E/S = 10 and a Au pseudo-RE, i.e. ‘3-e E/S = 10 Au’ in Figure 4.16.

52

An overview of the resistance evolutions in all four cells is displayed in

Figure 4.18. The median of the internal resistance of the CE increased before

that of the WE, although the latter surpassed the former in the cell with the

higher loading. In Figure 4.18a, in the two cells with E/S = 10, the internal

resistance was dominated by the CE in the later cycles. This indicated that the

metallic Li electrode limited the cyclability of these Li–S cells even though it

had much excess capacity. This implication supported the conclusion from the

previous chapter, which stated that the passivation of the S/C electrode did not

affect the cycle life.

Figure 4.18. The median (solid line), maximum and minimum (ribbon) of the internal resistance (R) in each cycle of cells with a) E/S = 10 and b) E/S = 6. Note that the scales are different in parts a and b. The maximum of the y-axis is set to 60 (a) and 350 (b) for the clarity. The data from first 8 cycles of the cell ‘3-e E/S = 10 LiAu’ have large noise in R due to the insufficient resolution of the potentiostat, which was changed after the 9th cycle.

53

5 Conclusions

This thesis can be concluded by answering the two research questions that

were raised in chapter 2: What mechanisms limits the discharge capacity of

these cells? And what limits their cycle life?

To answer the first question, an experimental framework for the investiga-

tion of an operating positive electrode was established in Paper I by the de-

velopment of both an electrochemically equivalent modified cell for X-ray

based techniques and an electrochemical method that enabled simultaneous

measurements of the internal and diffusion resistances. The correlation be-

tween the amount of the insulating precipitates and the transport properties

inside the carbon matrix was revealed by the concurrent crystallographic and

electrochemical characterisations. Based on this framework, Paper II built

the missing mechanistic link between the parameters of Li–S cells and their

electrochemical performance by comparing the real-time variation of Li2S in

a series of consistent operando XRD cells. The experimental framework was

subsequently enhanced by incorporating small-angle X-ray and neutron scat-

tering techniques in Paper III, which revealed information about not only the

crystalline and amorphous solids but also the catholyte in the mesopores of

the carbon matrix. With the synchronous resistance measurements, a correla-

tion between the drop in Li+ concentration and the increase in diffusion re-

sistance was identified at medium C-rates. The formation of less Li-demand-

ing amorphous precipitates at a low C-rate corroborated this finding.

To summarise the first three papers and answer the first question, the ionic

transport can be identified as the main limiting factor for the discharge capac-

ity of a highly porous S/C composite electrode. At high C-rates, the discharge

process was terminated by the surge in solution resistance caused by a non-

uniform current distribution, as exhibited in Paper II. At medium C-rates, the

discharge process was stopped by the rapid increase of diffusion resistance,

which is correlated to a decrease in Li+ concentration inside the carbon matrix,

as demonstrated in Paper III. At low C-rates, higher sulfur utilisation rate was

achieved by forming amorphous discharge product in the last stage of dis-

charge, as observed in Paper III. However, if the amorphous product is Li2S2,

as the literature suggests,36 the discharge capacity will be limited by the solid-

state kinetics of the following reduction to Li2S. These observations implicate

that sustaining the flow of Li+ into the carbon matrix may be the key to en-

hance the sulfur utilisation in an encapsulating cathode.

54

To answer the second question, an investigation was first conducted on a

widely-presumed cause for the short cycle life of Li–S cells: cathode pas-

sivation. To precisely characterise the degree of passivation of a cycling S/C

electrode, an EIS-based technique was developed and verified in symmetric

cells. A test protocol was subsequently designed to measure the electrochem-

ically active surface area while minimising the disturbance to the cycling cell.

The observed stable surface area while the discharge capacity dropped sub-

stantially ruled out cathode passivation as a major cause of the short cycle life.

Finally, the attention was turned to the negative electrode. With the develop-

ment of a reliable three-electrode setup, the resistance contributions from the

positive and negative electrodes was resolved. The long-term measurements

revealed a faster increase in internal resistance on the negative electrode. This

not only indicated that metallic Li limited the cycle life, but also implicated

that the common practice of testing S/C electrodes in a “half-cell” with a me-

tallic Li counter electrode should be reconsidered.

In conclusion, this thesis work has not provided a solution to the problems

of the Li–S system, but it was never the true intention. Instead, it may in fact

have revealed more problems by attempting to push the boundaries of charac-

terising such a complex electrochemical system. The two main research ques-

tions raised here have been discussed frequently in the field. Until now, how-

ever, there is still limited fundamental understanding of the discharge process

and the limitation of the cycle life. Hopefully, the observations here can serve

as references for the future development and the methodology originated from

the works can inspire more sophisticated characterisation techniques, thereby

contributing to the advancement of the Li–S batteries.

55

6 Sammanfattning på svenska

Litium–svavelbatterier har betraktats som en av de mest lovande kandidaterna

för nästa generation av batterier, främst på grund av deras höga teoretiska

energiinnehåll per enhet massa (2552 Wh kg-1) och den stora mängden svavel

som kommer som en biprodukt från petroleumindustrin. Hittills har kommer-

siella litium-svavelbatterier uppvisat en energitäthet på 400 Wh kg-1, vilket

förvisso är överlägset många andra kommersiella batterier, men långt ifrån det

teoretiska värdet. Samtidigt innefattar litium-svavelbatterier andra problem

som i allmänhet inte är associerade med vanliga litium-jonbatterier. Exempel

på detta är en begränsad användning av det aktiva materialet och en kort livs-

längd för batterierna. Dessa problem härrör från de unika reaktionsmekan-

ismerna i svavel/kolkompositelektroden, som utgör batteriets positiva elek-

trod, och den låga reversibiliteten för den metalliska litiumelektroden, som är

den negativa elektroden.

Den komplicerade reaktionsmekanismen i den positiva elektroden är en av

anledningarna till det låga utnyttjandet av aktiva materialet. Vid urladdning

genomgår elementärt svavel en serie intermediära reaktioner och reduceras så

småningom till litiumsulfid (Li2S). De intermediära produkterna (Li2Sx, x =

2–8) är lösliga i olika omfattning i de oftast använda eterbaserade elektroly-

terna. Därmed kommer det aktiva materialet i den positiva elektroden att lösas

upp och fällas ut vid varje urladdning eller uppladdning. Tillväxten av isolerat

svavel och litiumsulfid kan då störa jontransporten inuti den positiva elektro-

den, vilket resulterar i en ofullständig batterireaktion och minskad kapacitet.

Därför är det avgörande att undersöka upplösningen och utfällningen av den

isolerande fasta fasen i den positiva elektroden.

I denna avhandling undersöks litium-svavelbatterier under det att de an-

vänds – operando – med tekniker såsom röntgendiffraktion, neutronspridning

(’small-angle neutron scattering’, SANS) och röntgenspridning (’small-angle

X-ray scattering’, SAXS), i kombination med den elektrokemiska metoden

Intermittent Current Interruption (ICI). Detta har alltså utförts i realtid, när

batterierna upp- och urladdas. Hur den inre resistansen samt diffusionsresi-

stansen hos batteriet korrelerar med kinetiken för utfällningen av de kristallina

substanserna har bestämts med hjälp av operando-röntgendiffraktion. Genom

operando-SANS och -SAXS har bildandet av kristallina och amorfa fasta ur-

laddningsprodukter samt variationen av elektrolyt inuti elektrodens mesoporer

56

kunnat studeras och kopplas till funktionerna i resistansprofilerna. Dessa stu-

dier har visat att transportbegränsningar av joner inuti den positiva elektroden

är den huvudsakliga orsaken till det låga utnyttjandet av svavel under batteri-

ernas urladdning.

För att analysera problemet med kort livslängd undersöker det fjärde delar-

betet i avhandlingen hur den repetitiva utfällningen av litiumsulfider påverkar

funktionaliteten hos den positiva elektroden. En metod baserad på elektroke-

misk impendansspektroskopi har utvecklats för att spåra den elektrokemiskt

aktiva ytarean hos kolmatrisen in situ under omfattande battericykling. Under-

sökningen fann dock ingen tilltagande passivering av den positiva elektroden,

trots den snabba minskningen av specifik urladdningskapacitet. Därför kunde

slutsatsen dras att passiveringen av den positiva elektroden inte är en begrän-

sande faktor för batteriets livslängd, vilket tidigare har diskuterats. Vidare, för

att studera den negativa elektrodens inverkan på livslängden hos batteriet ut-

vecklades en tillförlitlig treelektrodsuppställning för litium-svavelceller.

Längre resistansmätningar med denna treelektrodcell avslöjade en snabbare

tillväxt av resistans hos den metalliska Li-elektroden vid upprepad upp- och

urladdning. Resultaten från dessa studier tyder därmed på att främst den ne-

gativa elektroden begränsar livslängden hos litium-svavelbatterier.

Sammanfattningsvis har denna avhandling uppvisat nya insikter i hur kemin

fungerar i litium-svavelceller under det att batterierna används. Dessutom har

värdet av parallella elektrokemiska och materialkaraktäriseringsmetoder på-

visats, och hur kombinationen av dessa är viktiga för att förstå det mycket

komplexa litium-svavelsystemet.

57

7 Acknowledgements

Again, I am rushing to submit my thesis at the last moment, just like two years

ago. I will try to do my best without wine but please forgive me if I miss any

name.

Thank you, Daniel, for your support to my work that sometimes gets a bit

expensive and crazy. I also appreciate that you went along with all my side

projects, which may be the reason why I am still writing now at the last mi-

nute. Matt, thank you for coming up with the idea of constantly stopping the

batteries and hiring me after listening to a presentation on DFT and CAL-

PHAD. I learned a lot during my PhD in through our sometimes cynical con-

versations/messages about science/ions.

I would also like to express my gratitude to my collaborators. Ashok, thank

you for answering to my spontaneous call at STOE and show me the way to

TOPAS. Keep calm and carry on doing ICI. Will, thank you for the urgent

manuscript revising, which just happened this week. Thank you, Adrian, for

all your help and guidance during my adventure in the small q-range. Thank

you, Nuria, Henry and John for all the great ideas and work that led to the

inspiring manuscript.

Of course, I am only still sane today thanks to the ålcoholics, Wolfgang,

Titania, Gustavito, Pachecito, Sarmad alltså, Marki, Money, Yoni, Laura,

Zacharia, DJ Cha Cha, Malaka, Tusa, Katalin, crazy rich German, uptight,

stinky, Robin, Rassi, and….aaaaaaaahhhhh. The deadline is chasing me……

58

8 References

(1) Bard, A. J.; Faulkner, L. R. Electrochemical Methods: Fundamentals and Ap-plications, 2nd ed.; Wiley, 2000.

(2) Winter, M.; Brodd, R. J. What Are Batteries, Fuel Cells, and Supercapacitors? Chemical Reviews 2004, 104 (10), 4245–4270.

(3) Winter, M.; Barnett, B.; Xu, K. Before Li Ion Batteries. Chemical Reviews 2018, 118 (23), 11433–11456.

(4) Wujcik, K. H.; Wang, D. R.; Teran, A. A.; Nasybulin, E.; Pascal, T. A.; Pren-dergast, D.; Balsara, N. P. Determination of Redox Reaction Mechanisms in Lithium–Sulfur Batteries. In Electrochemical Engineering; John Wiley & Sons, Ltd, 2018; pp 41–74.

(5) Manthiram, A.; Chung, S.-H.; Zu, C. Lithium-Sulfur Batteries: Progress and Prospects. Advanced Materials 2015, 27 (12), 1980–2006.

(6) Pope, M. A.; Aksay, I. A. Structural Design of Cathodes for Li-S Batteries. Advanced Energy Materials 2015, 5 (16), 1–22.

(7) Shi, L.; Bak, S.-M.; Shadike, Z.; Wang, C.; Niu, C.; Northrup, P.; Lee, H.; Baranovskiy, A. Y.; Anderson, C. S.; Qin, J.; Feng, S.; Ren, X.; Liu, D.; Yang, X.-Q.; Gao, F.; Lu, D.; Xiao, J.; Liu, J. Reaction Heterogeneity in Practical High-Energy Lithium–Sulfur Pouch Cells. Energy & Environmental Science 2020, 13 (10), 3620–3632.

(8) Shateri, N.; Auger, D. J.; Fotouhi, A.; Brighton, J. An Experimental Study on Prototype Lithium-Sulfur Cells for Ageing Analysis and State-of-Health Es-timation. IEEE Transactions on Transportation Electrification 2021, 1–1.

(9) Ye, G.; Zhao, M.; Hou, L.-P.; Chen, W.-J.; Zhang, X.-Q.; Li, B.-Q.; Huang, J.-Q. Evaluation on a 400 Wh Kg−1 Lithium–Sulfur Pouch Cell. Journal of Energy Chemistry 2022, 66, 24–29.

(10) Zhu, K.; Wang, C.; Chi, Z.; Ke, F.; Yang, Y.; Wang, A.; Wang, W.; Miao, L. How Far Away Are Lithium-Sulfur Batteries From Commercialization? Frontiers in Energy Research 2019, 7 (November), 1–12.

(11) Ue, M.; Sakaushi, K.; Uosaki, K. Basic Knowledge in Battery Research Bridging the Gap between Academia and Industry. Materials Horizons 2020, 7 (8), 1937–1954.

(12) Li, Z.; Huang, Y.; Yuan, L. Status and Prospects in Sulfur – Carbon Compo-sites as Cathode Materials for Rechargeable Lithium – Sulfur Batteries. Car-bon 2015, 92, 41–63.

(13) Maria Joseph, H.; Fichtner, M.; Munnangi, A. R. Perspective on Ultrami-croporous Carbon as Sulphur Host for Li–S Batteries. Journal of Energy Chemistry 2021, 59, 242–256.

(14) Yuan, H.; Huang, J.-Q.; Peng, H.-J.; Titirici, M.-M.; Xiang, R.; Chen, R.; Liu, Q.; Zhang, Q. A Review of Functional Binders in Lithium-Sulfur Batteries. Advanced Energy Materials 2018, 8 (31), 1802107.

(15) Jin, B.; Li, Y.; Qian, J.; Zhan, X.; Zhang, Q. Environmentally Friendly Bind-ers for Lithium-Sulfur Batteries. ChemElectroChem 2020, 1–20.

59

(16) Dirlam, P. T.; Glass, R. S.; Char, K.; Pyun, J. The Use of Polymers in Li-S Batteries: A Review. Journal of Polymer Science Part A: Polymer Chemistry 2017, 55 (10), 1635–1668.

(17) Cao, Y.; Li, M.; Lu, J.; Liu, J.; Amine, K. Bridging the Academic and Indus-trial Metrics for Next-Generation Practical Batteries. Nature Nanotechnology 2019, 14 (3), 200–207.

(18) He, Q.; Freiberg, A. T. S.; Patel, M. U. M.; Qian, S.; Gasteiger, H. A. Oper-ando Identification of Liquid Intermediates in Lithium–Sulfur Batteries via Transmission UV–Vis Spectroscopy. Journal of The Electrochemical Society 2020, 167 (8), 080508.

(19) Pascal, T. A.; Wujcik, K. H.; Velasco-Velez, J.; Wu, C.; Teran, A. A.; Ka-pilashrami, M.; Cabana, J.; Guo, J.; Salmeron, M.; Balsara, N.; Prendergast, D. X-Ray Absorption Spectra of Dissolved Polysulfides in Lithium–Sulfur Batteries from First-Principles. The Journal of Physical Chemistry Letters 2014, 5 (9), 1547–1551.

(20) Dibden, J. W.; Smith, J. W.; Zhou, N.; Garcia-Araez, N.; Owen, J. R. Predict-ing the Composition and Formation of Solid Products in Lithium–Sulfur Bat-teries by Using an Experimental Phase Diagram. Chem. Commun. 2016, 52 (87), 12885–12888.

(21) Hannauer, J.; Scheers, J.; Fullenwarth, J.; Fraisse, B.; Stievano, L.; Johansson, P. The Quest for Polysulfides in Lithium-Sulfur Battery Electrolytes: An Op-erando Confocal Raman Spectroscopy Study. ChemPhysChem 2015, 16 (13), 2755–2759.

(22) Bouchal, R.; Boulaoued, A.; Johansson, P. Monitoring Polysulfide Solubility and Diffusion in Fluorinated Ether‐ Based Electrolytes by Operando Raman Spectroscopy. Batteries & Supercaps 2020, 3 (5), 397–401.

(23) Wild, M.; O’Neill, L.; Zhang, T.; Purkayastha, R.; Minton, G.; Marinescu, M.; Offer, G. J. Lithium Sulfur Batteries, a Mechanistic Review. Energy & Environmental Science 2015, 8 (12), 3477–3494.

(24) He, Q.; Gorlin, Y.; Patel, M. U. M.; Gasteiger, H. A.; Lu, Y.-C. Unraveling the Correlation between Solvent Properties and Sulfur Redox Behavior in Lithium-Sulfur Batteries. Journal of The Electrochemical Society 2018, 165 (16), A4027–A4033.

(25) Mikhaylik, Y. v.; Akridge, J. R. Polysulfide Shuttle Study in the Li/S Battery System. Journal of The Electrochemical Society 2004, 151 (11), A1969.

(26) Lacey, M. J.; Yalamanchili, A.; Maibach, J.; Tengstedt, C.; Edström, K.; Brandell, D. The Li–S Battery: An Investigation of Redox Shuttle and Self-Discharge Behaviour with LiNO3-Containing Electrolytes. RSC Adv. 2016, 6 (5), 3632–3641.

(27) Lacey, M. J.; Edström, K.; Brandell, D. Analysis of Soluble Intermediates in the Lithium-Sulfur Battery by a Simple in Situ Electrochemical Probe. Elec-trochemistry Communications 2014, 46, 91–93.

(28) Berg, E. J.; Trabesinger, S. Viability of Polysulfide-Retaining Barriers in Li–S Battery. Journal of The Electrochemical Society 2018, 165 (1), A5001–A5005.

(29) Lang, S. Y.; Xiao, R. J.; Gu, L.; Guo, Y. G.; Wen, R.; Wan, L. J. Interfacial Mechanism in Lithium-Sulfur Batteries: How Salts Mediate the Structure Evolution and Dynamics. Journal of the American Chemical Society 2018, 140 (26), 8147–8155.

(30) Cuisinier, M.; Hart, C.; Balasubramanian, M.; Garsuch, A.; Nazar, L. F. Rad-ical or Not Radical: Revisiting Lithium-Sulfur Electrochemistry in Nonaque-ous Electrolytes. Advanced Energy Materials 2015, 5 (16), 1401801.

60

(31) Xu, Z.-L.; Kim, S. J.; Chang, D.; Park, K.-Y.; Dae, K. S.; Dao, K. P.; Yuk, J. M.; Kang, K. Visualization of Regulated Nucleation and Growth of Lithium Sulfides for High Energy Lithium Sulfur Batteries. Energy & Environmental Science 2019, 3144–3155.

(32) Drvarič Talian, S.; Kapun, G.; Moškon, J.; Vizintin, A.; Randon-Vitanova, A.; Dominko, R.; Gaberšček, M. Which Process Limits the Operation of a Li–S System? Chemistry of Materials 2019, 31 (21), 9012–9023.

(33) Chu, H.; Noh, H.; Kim, Y.-J.; Yuk, S.; Lee, J.-H.; Lee, J.; Kwack, H.; Kim, Y.; Yang, D.-K.; Kim, H.-T. Achieving Three-Dimensional Lithium Sulfide Growth in Lithium-Sulfur Batteries Using High-Donor-Number Anions. Na-ture Communications 2019, 10 (1), 188.

(34) Yang, B.; Jiang, H.; Zhou, Y.; Liang, Z.; Zhao, T.; Lu, Y.-C. Critical Role of Anion Donicity in Li 2 S Deposition and Sulfur Utilization in Li–S Batteries. ACS Applied Materials & Interfaces 2019, 11 (29), 25940–25948.

(35) Kong, L.; Chen, J.-X.; Peng, H.-J.; Huang, J.-Q.; Zhu, W.; Jin, Q.; Li, B.-Q.; Zhang, X.-T.; Zhang, Q. Current-Density Dependence of Li 2 S/Li 2 S 2 Growth in Lithium–Sulfur Batteries. Energy & Environmental Science 2019, 12 (10), 2976–2982.

(36) Waluś, S.; Barchasz, C.; Bouchet, R.; Leprêtre, J.-C.; Colin, J.-F.; Martin, J.-F.; Elkaïm, E.; Baehtz, C.; Alloin, F. Lithium/Sulfur Batteries Upon Cycling: Structural Modifications and Species Quantification by In Situ and Operando X-Ray Diffraction Spectroscopy. Advanced Energy Materials 2015, 5 (16), 1500165.

(37) Fan, F. Y.; Carter, W. C.; Chiang, Y.-M. Mechanism and Kinetics of Li 2 S Precipitation in Lithium-Sulfur Batteries. Advanced Materials 2015, 27 (35), 5203–5209.

(38) Barchasz, C.; Leprêtre, J. C.; Alloin, F.; Patoux, S. New Insights into the Lim-iting Parameters of the Li/S Rechargeable Cell. Journal of Power Sources 2012, 199, 322–330.

(39) Pan, H.; Chen, J.; Cao, R.; Murugesan, V.; Rajput, N. N.; Han, K. S.; Persson, K.; Estevez, L.; Engelhard, M. H.; Zhang, J.-G.; Mueller, K. T.; Cui, Y.; Shao, Y.; Liu, J. Non-Encapsulation Approach for High-Performance Li–S Batteries through Controlled Nucleation and Growth. Nature Energy 2017, 2 (10), 813–820.

(40) Zheng, J.; Gu, M.; Wang, C.; Zuo, P.; Koech, P. K.; Zhang, J.-G.; Liu, J.; Xiao, J. Controlled Nucleation and Growth Process of Li 2 S 2 /Li 2 S in Lithium-Sulfur Batteries. Journal of The Electrochemical Society 2013, 160 (11), A1992–A1996.

(41) Lin, Y.; Chen, D.; Wang, S.; Han, D.; Xiao, M.; Meng, Y. Addressing Pas-sivation of a Sulfur Electrode in Li–S Pouch Cells for Dramatically Improving Their Cyclic Stability. ACS Applied Materials & Interfaces 2020, 12 (26), 29296−29301.

(42) Su, Y.-S.; Fu, Y.; Cochell, T.; Manthiram, A. A Strategic Approach to Re-charging Lithium-Sulphur Batteries for Long Cycle Life. Nature Communi-cations 2013, 4 (1), 2985.

(43) Cheng, X.-B.; Huang, J.-Q.; Zhang, Q. Review—Li Metal Anode in Working Lithium-Sulfur Batteries. Journal of The Electrochemical Society 2018, 165 (1), A6058–A6072.

(44) Fang, C.; Li, J.; Zhang, M.; Zhang, Y.; Yang, F.; Lee, J. Z.; Lee, M.-H.; Al-varado, J.; Schroeder, M. A.; Yang, Y. et al. Quantifying Inactive Lithium in Lithium Metal Batteries. Nature 2019, 572 (7770), 511–515.

61

(45) Aurbach, D.; Pollak, E.; Elazari, R.; Salitra, G.; Kelley, C. S.; Affinito, J. On the Surface Chemical Aspects of Very High Energy Density, Rechargeable Li-Sulfur Batteries. Journal of Electrochemical Society 2009, 156 (8), A694–A702.

(46) Zhang, S. S. Role of LiNO 3 in Rechargeable Lithium/Sulfur Battery. Elec-trochimica Acta 2012, 70, 344–348.

(47) Chien, Y.-C.; Pan, R.; Lee, M.-T.; Nyholm, L.; Brandell, D.; Lacey, M. J. Cellulose Separators With Integrated Carbon Nanotube Interlayers for Lith-ium-Sulfur Batteries: An Investigation into the Complex Interplay between Cell Components. Journal of The Electrochemical Society 2019, 166 (14), A3235–A3241.

(48) Zhang, L.; Ling, M.; Feng, J.; Mai, L.; Liu, G.; Guo, J. The Synergetic Inter-action between LiNO3 and Lithium Polysulfides for Suppressing Shuttle Ef-fect of Lithium-Sulfur Batteries. Energy Storage Materials 2018, 11, 24–29.

(49) Lacey, M. J. Influence of the Electrolyte on the Internal Resistance of Lith-ium−Sulfur Batteries Studied with an Intermittent Current Interruption Method. ChemElectroChem 2017, 4 (8), 1997–2004.

(50) Cheng, X. B.; Yan, C.; Huang, J. Q.; Li, P.; Zhu, L.; Zhao, L.; Zhang, Y.; Zhu, W.; Yang, S. T.; Zhang, Q. The Gap between Long Lifespan Li-S Coin and Pouch Cells: The Importance of Lithium Metal Anode Protection. Energy Storage Materials 2017, 6, 18–25.

(51) Lacey, M. J.; Österlund, V.; Bergfelt, A.; Jeschull, F.; Bowden, T.; Brandell, D. A Robust, Water-Based, Functional Binder Framework for High-Energy Lithium-Sulfur Batteries. ChemSusChem 2017, 10 (13), 2758–2766.

(52) Urbonaite, S.; Novák, P. Importance of “unimportant” Experimental Parame-ters in Li-S Battery Development. Journal of Power Sources 2014, 249, 497–502.

(53) Eroglu, D.; Zavadil, K. R.; Gallagher, K. G. Critical Link between Materials Chemistry and Cell-Level Design for High Energy Density and Low Cost Lithium-Sulfur Transportation Battery. Journal of the Electrochemical Soci-ety 2015, 162 (6), A982–A990.

(54) Sun, K.; Cama, C. A.; Huang, J.; Zhang, Q.; Hwang, S.; Su, D.; Marschilok, A. C.; Takeuchi, K. J.; Takeuchi, E. S.; Gan, H. Effect of Carbon and Binder on High Sulfur Loading Electrode for Li-S Battery Technology. Electro-chimica Acta 2017, 235, 399–408.

(55) Borkiewicz, O. J.; Wiaderek, K. M.; Chupas, P. J.; Chapman, K. W. Best Practices for Operando Battery Experiments: Influences of X-Ray Experi-ment Design on Observed Electrochemical Reactivity. The Journal of Physi-cal Chemistry Letters 2015, 6 (11), 2081–2085.

(56) Amatucci, G. G. CoO2, The End Member of the LixCoO2 Solid Solution. Journal of The Electrochemical Society 1996, 143 (3), 1114.

(57) Lacey, M. J.; Edström, K.; Brandell, D. Visualising the Problems with Bal-ancing Lithium–Sulfur Batteries by “Mapping” Internal Resistance. Chemical Communications 2015, 51 (92), 16502–16505.

(58) Chien, Y.-C.; Menon, A. S.; Brant, W. R.; Brandell, D.; Lacey, M. J. Simul-taneous Monitoring of Crystalline Active Materials and Resistance Evolution in Lithium–Sulfur Batteries. Journal of the American Chemical Society 2020, 142 (3), 1449–1456.

(59) de Levie, R. On Porous Electrodes in Electrolyte Solutions. Electrochimica Acta 1963, 8 (10), 751–780.

62

(60) Shriver, D. F.; Weller, M. T.; Overton, T.; Rourke, J.; Armstrong, F. A. Shriver & Aktins’ Inorganic Chemistry, 5th ed.; Oxford University Press, 2010.

(61) Pauw, B. R. Everything SAXS: Small-Angle Scattering Pattern Collection and Correction. Journal of Physics: Condensed Matter 2013, 25 (38), 383201.

(62) Sivia, D. S. The Basics of X-Ray and Neutron Scattering. In Elementary Scat-tering Theory; Oxford University Press, 2011; pp 63–92.

(63) Brand, H. E. A.; Scarlett, N. V. Y.; Grey, I. E.; Knott, R. B.; Kirby, N. In Situ SAXS Studies of the Formation of Sodium Jarosite. Journal of Synchrotron Radiation 2013, 20 (4), 626–634.

(64) Kolosnitsyn, V. S.; Kuz’mina, E. v.; Karaseva, E. v.; Mochalov, S. E. Imped-ance Spectroscopy Studies of Changes in the Properties of Lithium-Sulfur Cells in the Course of Cycling. Russian Journal of Electrochemistry 2011, 47 (7), 793–798.

(65) Zhang, T.; Marinescu, M.; O’Neill, L.; Wild, M.; Offer, G. Modeling the Volt-age Loss Mechanisms in Lithium-Sulfur Cells: The Importance of Electrolyte Resistance and Precipitation Kinetics. Physical Chemistry Chemical Physics 2015, 17 (35), 22581–22586.

(66) Cañas, N. A.; Wolf, S.; Wagner, N.; Friedrich, K. A. In-Situ X-Ray Diffrac-tion Studies of Lithium-Sulfur Batteries. Journal of Power Sources 2013, 226, 313–319.

(67) Waluś, S.; Barchasz, C.; Colin, J.-F.; Martin, J.-F.; Elkaïm, E.; Leprêtre, J.-C.; Alloin, F. New Insight into the Working Mechanism of Lithium–Sulfur Batteries: In Situ and Operando X-Ray Diffraction Characterization. Chemi-cal Communications 2013, 49 (72), 7899.

(68) Zhu, W.; Paolella, A.; Kim, C.-S.; Liu, D.; Feng, Z.; Gagnon, C.; Trottier, J.; Vijh, A.; Guerfi, A.; Mauger, A.; Julien, C. M.; Armand, M.; Zaghib, K. In-vestigation of the Reaction Mechanism of Lithium Sulfur Batteries in Differ-ent Electrolyte Systems by in Situ Raman Spectroscopy and in Situ X-Ray Diffraction. Sustainable Energy Fuels 2017, 1 (4), 737–747.

(69) Cañas, N. A.; Hirose, K.; Pascucci, B.; Wagner, N.; Friedrich, K. A.; Hiesgen, R. Investigations of Lithium-Sulfur Batteries Using Electrochemical Imped-ance Spectroscopy. Electrochimica Acta 2013, 97, 42–51.

(70) Risse, S.; Cañas, N. A.; Wagner, N.; Härk, E.; Ballauff, M.; Friedrich, K. A. Correlation of Capacity Fading Processes and Electrochemical Impedance Spectra in Lithium/Sulfur Cells. Journal of Power Sources 2016, 323, 107–114.

(71) Villevieille, C.; Novák, P. A Metastable β-Sulfur Phase Stabilized at Room Temperature during Cycling of High Efficiency Carbon Fibre-Sulfur Compo-sites for Li-S Batteries. Journal of Materials Chemistry A 2013, 1 (42), 13089–13092.

(72) Saqib, N.; Ohlhausen, G. M.; Porter, J. M. In Operando Infrared Spectroscopy of Lithium Polysulfides Using a Novel Spectro-Electrochemical Cell. Journal of Power Sources 2017, 364, 266–271.

(73) Dominko, R.; Vizintin, A.; Aquilanti, G.; Stievano, L.; Helen, M. J.; Mun-nangi, A. R.; Fichtner, M.; Arcon, I. Polysulfides Formation in Different Elec-trolytes from the Perspective of X-Ray Absorption Spectroscopy. Journal of The Electrochemical Society 2018, 165 (1), A5014–A5019.

(74) Chaa, E.; Patela, M. D.; Choi, T.-Y.; Choi, W. Lithium-Sulfur Batteries: The Effect of High Sulfur Loading on the Electrochemical Performance. ECS Transactions 2018, 85 (13), 295–302.

63

(75) Kang, S. H.; Zhao, X.; Manuel, J.; Ahn, H. J.; Kim, K. W.; Cho, K. K.; Ahn, J. H. Effect of Sulfur Loading on Energy Density of Lithium Sulfur Batteries. Physica Status Solidi (A) Applications and Materials Science 2014, 211 (8), 1895–1899.

(76) Ding, N.; Li, X.; Chien, S.; Liu, Z.; Zong, Y. In-Situ Monitoring Viscosity Change of Electrolyte in a Li-S Battery. Chem. Commun. 2017, 53, 10152–10155.

(77) Huang, J.; Sumpter, B. G.; Meunier, V. Theoretical Model for Nanoporous Carbon Supercapacitors. Angewandte Chemie International Edition 2008, 47 (3), 520–524.

(78) Brug, G. J.; van den Eeden, A. L. G.; Sluyters-Rehbach, M.; Sluyters, J. H. The Analysis of Electrode Impedances Complicated by the Presence of a Con-stant Phase Element. Journal of Electroanalytical Chemistry and Interfacial Electrochemistry 1984, 176 (1–2), 275–295.

(79) Lasia, A. Impedance of Porous Electrodes. Journal of Electroanalytical Chemistry 1995, 397 (1–2), 27–33.

(80) Bach, P.; Valencia-Jaime, I.; Rütt, U.; Gutowski, O.; Romero, A. H.; Renner, F. U. Electrochemical Lithiation Cycles of Gold Anodes Observed by in Situ High-Energy X-Ray Diffraction. Chemistry of Materials 2016, 28 (9), 2941–2948.

(81) Solchenbach, S.; Pritzl, D.; Kong, E. J. Y.; Landesfeind, J.; Gasteiger, H. A. A Gold Micro-Reference Electrode for Impedance and Potential Measure-ments in Lithium Ion Batteries. Journal of The Electrochemical Society 2016, 163 (10), A2265–A2272.

Acta Universitatis Upsaliensis Digital Comprehensive Summaries of Uppsala Dissertations from the Faculty of Science and Technology 2075

Editor: The Dean of the Faculty of Science and Technology

A doctoral dissertation from the Faculty of Science and Technology, Uppsala University, is usually a summary of a number of papers. A few copies of the complete dissertation are kept at major Swedish research libraries, while the summary alone is distributed internationally through the series Digital Comprehensive Summaries of Uppsala Dissertations from the Faculty of Science and Technology. (Prior to January, 2005, the series was published under the title “Comprehensive Summaries of Uppsala Dissertations from the Faculty of Science and Technology”.)

Distribution: publications.uu.se urn:nbn:se:uu:diva-453442

ACTA UNIVERSITATIS

UPSALIENSIS UPPSALA

2021