. A . ‘ u , . , , , u... 3 .
1 v1.2. ‘ , ‘ . . , , . ‘ ‘ . . . $53.»
L: . LL. . . . . . ‘ . . . . . L... ~ Sty“?

4:4:

. 7‘ l g
4....» duty} 1
. .. .13. A. A

 

.z

N.
M.

 

 

 

 

 

THES!
Z
’?Fol

This is to certify that the

dissertation entitled

PEO-BASED POLYMER ELECTROLYTES FOR
SECONDARY LITHIUM BATTERIES

presented by

Micah K. Stowe

has been accepted towards fulfillment
of the requirements for

—_P_b.rD...— degree in _Chemjs_tmL_

.cg/Jzigfiwx

Major professor

Date April 9. 2001

MSU i: an Affirmative Action/Equal Opportunity Institution 0-12771

{LIBRARY 7
‘ Michigan State .
University f

——

PLACE IN RETURN BOX to remove this checkout from your record.
TO AVOID FINES return on or before date due.
MAY BE RECALLED with earlier due date if requested.

 

DATE DUE DATE DUE DATE DUE

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

8/01 cJCIRC/DateOuepss-p. 15

 

PEO-BASED POLYMER ELECTROLYTES FOR SECONDARY
LITHIUM BATTERIES

By

Micah K. Stowe

A DISSERTATION
Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of

DOCTOR OF PHILOSOPHY

Department of Chemistry

2001

PE(

Poiyeth
mirgeable it
witty and I
mi’yfeihylene (.
Edited £13514

“5‘ SIL'face f,

The fill

CTS R;

 

 

 

ABSTRACT

PEO-BASED POLYMER ELECTROLYTES FOR SECONDARY
LITHIUM BATTERIES

By

Micah K. Stowe

Polyethers mixed with lithium salts are excellent candidates for electrolytes in
rechargeable lithium batteries. Polyether systems with low crystallinity result in fast ion
mobility and therefore high conductivities. In this work the properties of several
poly(ethylene oxide) based electrolytes are examined with an emphasis on systems with
reduced crystallinity including, composite polymer electrolytes, oligomeric polyethers,

and (AB)“ microblock copolymers.

Highly conductive and processable composite polymer electrolytes were made
using surface functionalized fumed silica fillers and PEGDME-SOO (LiClO4, O/Li = 20).
The fillers were both hydrophobic and cross-linkable and formed an open three-
dimensional network in the electrolytes due to van der Waals forces. The open network
allowed for high ionic mobility and provided for the mechanical stability of the
composite. Methacrylate monomers of differing hydrophobicity were added to cross-link
the silica network and impart permanent mechanical stability. The optical, conductive,
thermal, mechanical, and kinetic properties of the composites are examined as a

functionof monomer hydrophobicity and filler surface chemistry. It was found that

hyIi'Op-hobic r:
phase separat:
the electroliic
85—95% com:
Machine .2
wide for h-
E‘flhyl metht.

tszszzmce.

A bias.
Extremes I}
I'anabie tempt-
We compare;
Cysh‘liniiy at :
haddiuon. st.
Ifcrease CQ'Rla
“3935an Ih.

Mei an

 

hydrophobic monomers such as butyl methacrylate and octyl methacrylate preferentially
phase separate onto the filler surface while hydrophilic methyl methacrylate is soluble in
the electrolyte phase. The composites were both photochemically and thermally cured to
85-95% conversion of monomer to polymer. Hydrophilic monomers such as methyl
methacrylate are more compatible with the electrolyte after polymerization and therefore
provide for better mechanical properties in the composite. However, unpolymerized
methyl methacrylate can react at the electrodes resulting in increased interfacial

resistance.

A branched oligomeric polyether, star(12)PEO, was prepared and characterized.
Electrolytes formed from star(12)PEO and LiClO4 were characterized by DSC and
variable temperature impedance spectroscopy. The properties of the branched system
were compared to linear PEGDME-SOO. It was found that branching completely inhibits
crystallinity and improves the low temperature properties of polyether-based electrolytes.
In addition, star(12)PEO added to high molecular weight poly(ethylene oxide) help to
decrease crystallinity. Finally, electrolytes were made from LiClO4 and (AB)n microblock
copolymers that contain a repeating pattern of exact length segments of poly(ethylene
oxide) and poly(ethylene). The thermal and conductive properties were studied as a
function of the polyether block length. The conductivity of the electrolytes increases with
the mole fraction of the polyether block, but electrolytes with >10 ether segments

crystallized, leading to decreases in conductivity.

To My Family

iv

I \I
this work.
wmmmn
personm‘ly
him for It.
Mmha an.
“FPO“ Ila.
Brett. Gm

IOT “Chill

My
0f fail. sho.
Wheel Kelli
Efiatne“.1
I... SM
YOur IOVC 4

JESSICA Bu

-1”,
3‘ ”e4 .

M“ {his

 

 

ACKNOWLEDGMENTS

I would like to express my thanks first to God for giving me the opportunity to do
this work. I have been blessed with a wonderful advisor, Professor Greg Baker, who is an
outstanding scientist and mentor. He has allowed me to grow both professionally and .
personally during my time at MSU and I would like to express my deepest gratitude to
him for this, and for his continuing friendship. I would also like to thank my parents
Marsha and Bob Stowe for encouraging me to follow my dreams. Your constant love and
support have made all the difference. To the rest of my family, Penny, Steve, Tyler,
Brett, Grandmas, and Grandpas. ...I am so grateful to have you all in my life. Thank you

for everything.

My time in Michigan has been full of new experiences and fun times...the colors
of fall, shoveling snow, freezing rain, learning that you have to kick the snow out of your
wheel wells if you want your tires to turn, humid summers, big ten sports (Go Spartansl),
great new friends and much more. A special thank you goes to Michael Miller who I
have shared much of this with, including the trials and tribulations of graduate school.
Your love and support has meant so much. 1am also so very grateful for the friendship of

Jessica Barker. Graduate school would not have been the same without you to confide in.

I would also like to thank all past and present Baker group members whom I have
shared this experience with, who have shared their knowledge, and have made this a fun

time... Chun (Honey), Gao, Kirk, Erin, Fadi, J .B., Tianqi, Sue, Tara, Cory, Yiyan, Hou,

Mao and

an unexpe

C c
assistance
Merlin Br
thanks to I

and Jerem:

 

 

Mao and Chris (honorary member). Much thanks also to Lisa. Our friendship has been

an unexpected and wonderful gift.

Completion of this dissertation would not have been possible without the
assistance of my guidance committee; Dr. Rob Maleczka, Dr. Gary Blanchard, and Dr.
Merlin Bruening. Thank you all for your time and assistance over the years. Finally,
thanks to our collaborators at NC. State; Dr. Saad Khan, Dr. Peter Fedkiw, Jeff Yerian,

and Jeremy Walls for your contributions to this work.

vi

DuofAr’

LISI Of Tat I

LutofScE

Unoffg;

 

Chapter
l

Chap“

TABLE OF CONTENTS

Page
List of Abbreviations ...................................................................................................... x
List of Schemes ............................................................................................................... xi
List of Tables ................................................................................................................... xii
List of Figures .................................................................................................................. xiii
Chapter 1. Introduction ........................................................................................ 1
I. Battery Systems ....................................................................................... l
1. General ................................................................................... 1
II. Ionic Conductivity ................................................................................... 9
1. General ................................................................................... 9
2. Impedance Spectroscopy .................................................. . ..... lO
3. Electrolytes for Li Batteries .................................................... 13
III. Mechanical Properties ............................................................................. 38
1. Dynamic Rheology ................................................................. 38
IV. Previous Work ......................................................................................... 40
1. Composite Polymer Electrolytes ............................................ 40
V. Monomer Behavior ................................................................................. 44
1. Suspension Polymerization .................................................... 44
2. Phase Separation ..................................................................... 48
3. Radical Polymerization .......................................................... 49
Chapter 2. Composite Polymer Electrolytes
Results and Discussion ................................................................... 54

vii

I.

H.

1]].

IV.

V.

Chapter 3.

I.

II.

111.

Chapter 4.

I.

Hydrophobic Cross-linkable Fumed Silicas. ........................................... 56

Composite Polymer Electrolytes ............................................................. 59

1. Cross-linked Composite Polymer Electrolytes ....................... 59

Composite Properties Based on Monomer Hydrophobicity .................... 64

1. Optical Properties .................................................................. 64

2. Conductive Properties ........................................................... 68

3. Thermal Properties ................................................................ 76

4. Mechanical Properties ........................................................... 88

5. Kinetics of Polymerization .................................................... 91
Composite Properties Based on Filler Surface Characteristics ............... 108
1. Mechanical Properties ............................................................ 108
Conclusions .......................................................... 116

Limiting Crystallinity in PEO

Results and Discussion ................................................................... 120
Preparation and Characteristics of star(12)PEO Electrolytes .................. 121
Plasticization of PEO .............................................................................. 130
l. Oligomeric PEO’s .................................................................. 130
2. Fillers ...................................................................................... 138
Alternating Microblock Copolymers ....................................................... 148
Experimental .................................................................................... 166
Materials .......................................... A ........................................................ 166

viii

Bibliograpi

 

 

11. Analytical Methods ................................................................................. 167

General Procedure for the Impedance Analyzer .......................... 171

III. Hydrophobic-Polymerizable Silicas ........................................................ 175

IV. Preparation of Composites ...................................................................... 176

Octyl methacrylate .......................................................................... 176

V. Preparation of Star(12)PEO .................................................... . ................ 177
Bibliography ............................................................................................................... 178

ix

AIBN
BIIA
CPE
DEA
D.\IAE.\IA
DNIPA
DSC

E0

is

FIIR
HINS
I11

WA
51,

MW
.\.\IR
0m
ODMCS
ODTCs
OTCS
OTMS
PDI
PEGDME
PEO
PMMA
q

R.

S

SE1
SPE

 

AIBN
BMA

CPE

DEA
DMAEMA
DMPA
DSC

LIST OF ABBREVIATIONS

2,2’-Azobisiso(butylronitrile)

Butyl methacrylate

Composite polymer electrolyte
Diethyl amine
Dimethylaminoethylmethacrylate
2,2’-Dimethoxy-2-phenyl-acetophenone
Differential scanning calorimetry
Ethylene oxide

Fumed silica

Fourier transform infrared spectroscopy
High resolution mass spectroscoppy
Multiplet

Methyl methacrylate

Number average molecular weight
Molecular Weight

Nuclear Magnetic Resonance

Octyl methacrylate
Octyldimethylchlorosilane
Octadecyltrichlorosilane
Octyltrichlorosilane
Octyltrimethoxysilane
Polydispersity index

Poly(ethylene glycol) dimethyl ether
Poly(ethylene oxide)

Poly(methyl methacrylate)

Quartet

Rate of polymerization

Singlet

Solvent electrode interface

Solid polymer electrolyte
Transmission electron microscopy
Crystallization temperature

Glass transition temperature
Thermogravimetric analyses
Melting Temperature

Trimethylsilyl propyl methacrylate
Triplet

Ultra violet

Scheme

Scheme 1.
Scheme 3.
Scheme 3.

Scheme 4.

Scheme 5.

Schemc 6.
Scheme 7_
Scheme 8.
Scheme 9'

Scheme 10.

 

LIST OF SCHEMES

Scheme Page
Scheme 1. Preparation of methoxy—linked PEO (PEMO) ......................................... 23
Scheme 2. Synthesis of a highly conductive network polymer ................................. 28
Scheme 3. Synthesis of FS(T OM) silica from A200 ................................................ 43

Scheme 4. Mechanism for radical formation of some common
radical initiators ....................................................................................... 50

Scheme 5. Schematic of the photo cross-linking set-up for UV

curing of composites ............................................................................... 60
Scheme 6. Exploded view of the thermal curing set-up ........................................... 62
Scheme 7. Diagram of the experimental set—up for measuring turbidity .................. 66
Scheme 8. Preparation of difunctional monomers for future polymerization ........... 118
Scheme 9. Preparation of star(12)PEO ..................................................................... 122
Scheme 10. Synthesis of (AB)n microblock copolymers ............................................ 149

xi

Table

Table I.
Table 2.
Table 3.

Table 4.

Table 5.
Table 6.
Table 7,

Table 8.

 

Table

Table 1.

Table 2.

Table 3.

Table 4.

Table 5.

Table 6.

Table 7.

Table 8.

Table 9.

Table 10.

Table 1 1.

Table 12.
Table 13.

Table 14.

LIST OF TABLES

Page

Examples of primary batteries ............................................................... 6
Examples of secondary batteries................... 7
Some common small molecule hosts for electrolytes

and their properties. ................................................................................. 16

Properties of composites using 10 wt% and 40 wt%

methacrylate monomers ........................................................................... 78
Physical properties of methacrylate monomers ...................................... 81
Solubility of AIBN in different solvents .............................................. 105

Molecular weights of THF soluble polymer extracted

from thermally polymerized composites ................................................. 106
The variation of Tg with salt concentration for

star(12)PEO and PEGDME-SOO. T c for PEGDME-SOO

mixtures is also given .............................................................................. 127
Melting transitions and heats of fusion for mixtures of

PEG-3400 with star(12)PEO and PEGDME-SOO ................................... 133
Crystalline transitions and heats of fusion for mixtures of

PEG-3400 with star(12)PEO and PEGDME-SOO ............................ 134
Properties of FED-3400 composites containing 10 and 20 wt%

of filler ..................................................................................................... 144
Physical properties of [C41I:C4EO,,]n polymers 152
Thermal properties of [C41tC4EOy]n (y = 6-14) electrolytes obtained

from DSC scans at 10 °C per minute 156

Products of mixed silylation reactions .................................................... 175

xii

Rpm

Emmi
Emmi
Emmi

Emme4.

Emmi

Egnel4.

Figure 15.

Esme I6

 

Figure

Figure 1.

Figure 2.
Figure 3.
Figure 4.
Figure 5.

Figure 6.

Figure 7

Figure 8.
Figure 9.

Figure 10.
Figure 11.

Figure 12.

Figure 13.
Figure 14.

Figure 15.
Figure 16.

LIST OF FIGURES

Page
Schematic of a genric Zn/Clz battery and the direction of
crrent flow under and applied load .......................................................... 1
Representation of ideal and actual discharge curves over time ............... 3
Discharge behavior of high capacity and high rate cells ......................... 4

Common cell configurations................. 5
Impedance analyzer set-up with variable temperature
capacrty 12
Example of a Nyquist plot obtained from variable frequency

Impedance measurements ........................................................................ 13
Structures of potential host polymers for

Side and top view of the helical structure of PEO .................................. 20
Temperature dependent conductivity of PEO/LiClO4 (O/Li = 6) ........... 22
Polymer structures designed to limit crystallinity........... 24

Common low Tg polymers used to increase the flexibility

of PEO ..................................................................................................... 25
Structures of polyether functionalized polystyrene ................................ 26
Polyether macromolecule used to form gel electrolytes .......................... 30

Effect of alumina filler on the resistance of a PEO(8)LiClO4

SPE .......................................................................................................... 32
Effect of filler particle size in a PEO(10)Nal SPE .................................. 33
TEM image of5 wt% A200 in mineral 01136

xiii

Figure 17.

Figure 18.

Figure 19.

Figure 20.

HE‘dre 38.

F?
1

We 29

r...

c we 30

Figure 31

 

Figure 17.

Figure 18.

Figure 19.
Figure 20.

Figure 21.

Figure 22.

Figure 23.

Figure 24.

Figure 25.

Figure 26.

Figure 27.

Figure 28.

Figure 29.

Figure 30.

Figure 31.

Schematic showing the thixotropic behavior of fumed silica

Systems .................................................................................................... 37
Standard parallel plate configuration for dynamic

rheology experiments .............................................................................. 39
Variable strain and stress experiments............. 40

Conductivity and modulus fo fumed silica composites

as a function of wt% filler ....................................................................... 42
Representation of a suspension polymerization. ..................................... 45
Three scenarios for monomer location in CPE’s ..................................... 47
Functionalized fumed silicas ................................................................... 55

1H NMR of trimethoxy silane residue from FS-TOM after

surface functionalization ......................................................................... 57
The mol% of each monomer added to the surface

of A200 based on the initial feed ratio .................................................... 58
The effect of pyrex versus quarts cover slide on the

w% BMA bound to the silica surface after UV curing ............................ 61

EUR spectra of a composite containing 10 wt% BMA before

andafterthermalcuring.............. ........................................................ 63
UV and thermal curing of composite electrolytes using BMA

as the monomer ................................................................................... 65
Results of turbidity measurements on composites

containing MMA, BMA, and OMA ........................................................ 67
Conductivity of composites made using 13 wt% TOM(4: 1) and

10 wt% methacrylates before UV polymerization ................................... 69
Conductivity of composites made using 13 wt% TOM(421) and

10 wt % methacrylates after UV polymerization .................................... 7O

xiv

,
figm3,

Hmm37

1figure 38.

Eam39

fiflm40

E®m4t

Emu:

Figure 32.

Figure 33.

Figure 34.

Figure 35.

Figure 36.

Figure 37.

Figure 38.

Figure 39.
Figure 40.

Figure 41.

Figure 42.

Figure 43.

Figure 44.

Figure 45.

Conductivity of cured and uncured CPE’s as a function of

OMA concentration ................................................................................. 72
An idealized view of OMA composites before and

after polymerization ................................................................................. 73
Conductivity of cured and uncured CPE’s as a function

of MMA concentration. . . . . .................................................................... 74
An idealized view of MMA composites before and

after polymerization......................' ........................................................... 75
Conductivity of cured and uncured CPE’s as a function

of BMA concentration ............................................................................. 77
DSC heating scans of uncured composites containing 10 wt%

and 40 wt% methacrylate monomers ...................................................... 79
DSC cooling scans of uncured composites containing 10 wt%

and 40 wt% methacrylate monomers ...................................................... 8O
DSC heating and cooling scans of OMA ................................................ 82
DSC heating scans of composite containing 40 wt%

methacrylates before and after polymerization ........................................ 84
DSC cooling scans of composite containing 40 wt%

methacrylates before and after polymerization ........................................ 85
TGA of composite filler materials after crosslinking with 10%
methacrylates and solvent extraction ....................................................... 87
Dynamic frequency sweep of composites containing 10 wt%

OMA and 15 wt% FS-TOM(80:20) filler ............................................... 89
Dynamic frequency sweep of composites containing 10 wt%

MMA and BMA using 15 wt% FS-TOM(80:20) filler ........................... 90
Kinetic curves for the thermal cross-linking of composites

containing 10 wt% MMA, BMA and OMA ............................................ 92

XV

figm46

Hgm47

Emm48

Epm49

FQMSO

FleureSI.

Figure 46.

Figure 47.

Figure 48.

Figure 49.

Figure 50.

Figure 51.

Figure 52.

Figure 53.

Figure 54.

Figure 55.

Figure 56.

Thermal cross-linking of 10 wt% MMA composites done

at 80 °C with and without filler .........................................................

Thermal cross-linking of 10 wt% MMA composites done

at 80 °C with and without LiClO4 .....................................................

Thermal cross-linking of 10 wt% BMA composites done

at 80 °C with and without LiClO4..................

Thermal cross-linking of 10 wt% OMA composites done

at 80 °C with and without LiClO4......................

Kinetic curves for the curing of 10 and 40 wt% BMA
containing composites at 80 °C. The effect of initiator

concentration was studied using 1% to 0.25% AIBN in 40 wt%

BMA composites ...............................................................................

The linear portion of the kinetic curves for the cross-linking
of BMA containing composites at 80 °C. The slope of each

curve, calculated from the least squares fit to the data,

is shown next to each data set ...........................................................

Kinetic curves for the curing of 10 and 40 wt% OMA

containing composites at 80 °C .........................................................

The linear portion of the kinetic curves for the cross-linking
of OMA containing composites at 80 °C. The slope of each

curve, calculated from the least squares fit to the data,

is shown next to each data set

Kinetic curves for the curing of MMA containing

composites at 80 °C ...........................................................................

FTIR spectra of PEGDME—SOO samples showing the effect

Of AIBN and 02 on the chemical stability of PEGDME-SOO ...........

Dynamic frequency sweep of a composite containing 5 wt% R805 .......

xvi

...... 94

...... 95

96

97

99

100

101

102

103

107
108

Figure 57.
Figure 58.

Figure 59.

Figure 60.

figure 61.
Figure 62.
Figure 63.
Figure 65.
Figure 66

F1336 68.

 

Figure 57.

Figure 58.

Figure 59.

Figure 60.

Figure 61.

Figure 62.

Figure 63.

Figure 64.

Figure 65.

Figure 66.

Figure 67.

Figure 68.

Figure 69.

Dynamic rheology of composites with 15 wt% FS-TOM fillers.

The degree of n-octyl substitution on the filler surface is varied ............ 110
Dynamic rheology of composites with 10 and 15 wt%

FS-TOM(45:55) filler .............................................................................. 111
Dynamic rheology of composites with 15 and 20 wt%

FS-TOM(80:20) fillers. Data for 15 wt% FS-TOM(66:33)

is shown for comparison .......................................................................... 112
Effect of FS-TOM filler surface on the conductivity of

composites with BMA before and after UV curing and

without added monomer .......................................................................... 114

Effect of FS-TOM filler surface composition on the fraction

of BMA attached to the surface ............................................................... 115
13c NMR of star(12)PEO ........................................................................ 123
DSC heating scans of star(12)PEO at several salt concentrations .......... 125

The variation of T3 with the molar salt fraction for star(12)PEO

and PEGDME-SOO .................................................................................. 126
Conductivity of star(12)PEO electrolytes as a function of

the O/Li ratio ........................................................................................... 128
Arhenius plots of the conductivity of several star(12)PEO

Electrolytes .............................................................................................. 129
DSC heating and cooling scans for different compositions of

star(12)PEO and PEG (Mn=3400) ........................................................... 131
DSC heating and cooling scans for different compositions of
PEGDME-SOO and PEG (Mn=3400) ....................................................... 132
Spherulitic morphology of PEG-3400 viewed at 50x

Magnification through cross-polarizers ................................................... 135

xvii

 

Emm70

hmm71

fiwm7l
fiamIi
Hamii
Humii
Figure 76.
RWW.

Figure 78.
Figure 79.
fiau80
Hire 81.

E2

ere 82.
IE“

wrest

fiflm&i

 

Figure 70.

Figure 71.

Figure 72.
Figure 73.

Figure 74.
Figure 75.

Figure 76.
Figure 77.
Figure 78.

Figure 79.

Figure 80.

Figure 81.

Figure 82.

Figure 83.

Figure 84.

Morphology of A. PEG-3400, 20 wt% star(12)PEO and B. PEO-

3400 50 wt% star(12)PEO viewed 50x magnification through
cross-polarizers ........................................................................................ 136
Morphology of A. FED-3400, 20 wt% PEGDME-SOO and B. PEO-

3400 50 wt% PEGDME-SOO viewed 50x magnification through

cross-polarizers ......... . .............................................................................. 137
Functionalized fumed silica fillers .......................................................... 139
DSC heating and cooling scans of PEG-3400 with A200-dry ................ 140
DSC heating and cooling scans of PEO-3400 with AME-8b .................. 141
DSC heating and cooling scans of PEO-3400 with R974 ....................... 142
DSC heating and cooling scans of PEG-3400 with R805 ....................... 143

DSC heating and cooling scans of PEG-3400 with 20 wt% fillers ......... 145
Illustration of how PEO chains may interact with R805, A200

and FS-AME8b fillers based on their surface properties ...................... 147
Ideal packing structure of [C41t4EOy],,l polymer chains

here y = 7, and the geometry of the double bonds is trans ...................... 151
Melting and glass transitions for [C41tC4EOy]n polymers with and

without salt .............................................................................................. 154
Evolution of the thermal properties of the polymer (C41tC4EOx),..

with increasing salt content. The numbers refer to the O:Li ratio .......... 155
DSC traces for polymer salt complexes prepared from

[C41cC4E014]n and LiClO4 (O/Li=24, 32, & 48) ...................................... 157
DSC traces for polymer salt complexes prepared from

[C41cC4EOm]n and LiClO4 (O/Li=24, 32, & 48) ...................................... 158
DSC traces for polymer salt complexes prepared from

[C41tC4EOg]n and LiClO4 (O/Li=24, 32, & 48) ....................................... 159

xviii

Figure 85.

Figure 86.

Figure 87.

Figure 88.

Figure 89.

DSC traces for polymer salt complexes prepared from

[C4,1'tC4EO7]n and LiClO4 (O/Li=24, 32, & 48) ....................................... 160
DSC traces for polymer salt complexes prepared from

[C41tC4E06]n and LiClO4 (O/Li=24, 32, & 48) ....................................... 161
DSC traces for polymer salt complexes prepared from

[C41tC4EOy]n and LiClO4 (O/Li=32) ....................................................... 163
Temperature dependent conductivity for [C41tC4EOy]n / LiClO4

Complexes with O/Li = 32 ...................................................................... 164

Impedance apparatus set-up .................................................................... 172

xix

l. Genera
Butt
electrical ei
portable ele
energy dire.
reactions. Ti
”Ode (the s:
electrode ”CT
Tells) in the
A“ electrolii
béiu’een the a

load COmI-‘ICIC

 

Chapter 1. Introduction

I. Battery Systems

1. General

Batteries harness chemical energy and convert it to a usable and storable form of
electrical energy. This electrical energy can then be called on to power cars, toys and
portable electronic devices. The active materials contained in a battery convert chemical
energy directly to electrical energy through electrochemical oxidation and reduction
reactions. The major components of an electrochemical cell consist of two electrodes: an
anode (the site of oxidation) and a cathode (the site of reduction), an electrolyte, and an
electrode separator. The anode and cathode materials define the potential difference (in
Volts) in the cell and therefore its theoretical capacity (in Ampere hours per gram, Ah/g).
An electrolyte sandwiched between the electrodes allows for the transport of ions
between the anode and cathode during discharge. Electrons flowing through an external

load completes the circuit.

   

e' e'
anode cathode
Zn C12

T

electrolyte

Figure 1. Schematic of a generic Zn/Clz battery and the direction of current

flow under an applied load

Figur
Zinc and the

from the ano

potentials are
anode

Cdlhm

0‘ 6rd

in prZiL

“TIA-
”Ll-(41 EX

 

Per:

Figure 1 shows a simple representation of a battery where the anode material is
Zinc and the cathode is C12. Upon applying a load, i‘.e. during discharge, electron flow is
from the anode to the cathode. The half cell reactions and the corresponding reduction

potentials are shown in equations 1 and 2 at 25°C.

anode: Zn ———> Zn2+ + 2e' v = -0.76 (1)
cathode: C12 + 2e- ——> 2Cl‘ V = 1.36 (2)
Overall: Zn + C1; —> ZnClz V = 2.12 (3)

The capacity of a cell is expressed as the total quantity of charge in ampere-hours
involved in the electrochemical reaction per gram of electrode materials. For the Zn/Clz
battery the theoretical capacity by weight is 0.394 Ah/g which is the energy that can be
withdrawn from a fully charged cell under specific discharge conditions per gram of
active electrode material. Note that the capacity can also be calculated on a volume
basis. The theoretical energy density (ED) of a battery, takes into account both voltage
and theoretical capacity and is commonly expressed in watthours/gram or liter (Wh/g or
Wh/L). The calculation of energy density in terms of equivalent weight is shown in

equation 4.

Watt.hour/gram capacity = 2.12 V X .394 Ah/g = 0.838 Wh/g (4)

In practice, the theoretical energy density is never realized. Figure 2 illustrates a

typical experimental discharge curve.1 In the idealized case, discharge of the battery

proceeds at
capacity is f:

similar to cur

Hgllre 2. RC[

In bott
resistive elem
CUTVC I CUT
ditch

" mg? tart

1038le “116'

'al
““1 35 POlLtng

 

0t rationS at t‘
on
[he materlii

[tan
‘ “Primed

 

In gen.»

‘1

I

proceeds at the theoretical voltage until the active materials are consumed and the
capacity is fully utilized. The voltage then drops to zero. Actual discharge curves are

similar to curves 1 and 2 in Figure 2.

 

 

 

 

 

T Ideal Curve
Curve 1

>

6 cm

0)

:3

T:

>
Time —>

Figure 2. Representation of ideal and actual (Curves 1 & 2) discharge curves over time.

In both cases an immediate IR drop (V=IR) is observed due to the presence of
resistive elements in the electrode materials as well as the electrolyte. Compared to
Curve 1, Curve 2 represents a cell with either a higher internal resistance, higher
discharge rate, or both, and shows a characteristic increase in its sloping profile.
Increased internal resistance can be caused by irreversible reactions at the electrodes as
well as polarization of active materials (for instance, the build-up of a high concentration
of cations at the cathode which can hinder lithium ion diffusion). Therefore, depending
on the materials and conditions of discharge, the result is a cell capacity 30-75% lower

than expected.

In general, there are two types of batteries: primary and secondary. In a primary

cell, the chemical components are consumed irreversibly during the discharge process

and the batter
cells are up
low/moderate
measure of hi
need for grea
current drain.
between high

Ofa Primal} C

Fig
(C;

Primdr.

 

n.
.13 La} design.

C r. ;
0.5.Mdered [O

0..

‘4
l

.14")- CCU-S -

 

"l- .
“3th
adIOrn

" l

and the battery is disposed of after discharge. Secondary cells are rechargeable. Primary
cells are typically used when a high capacity (high energy density) is needed at
low/moderate discharge rates, i.e., when power requirements are low. Power is a
measure of both voltage and current (P = V * I or I2 * R). High power devices imply a
need for greater efficiency in delivering power even at higher discharge rates (higher
current drain). However, capacity is usually lower. Figure 3 illustrates the difference
between high capacity and high-rate (high power) cells. Curve 1 represents the behavior

of a primary cell and curve 2 a secondary cell.

Curve 1 — High Capacity

s Curve 2 — High Rate

Capacity —p

 

 

Discharge current —>

Figure 3. Discharge behavior of high capacity (Curve 1) and high rate
(Curve 2) cells.

Primary cells are usually designed to contain the maximum of active material
possible for a given cell configuration. The bobbin and button cell constructions are
typical designs for primary cells (Figure 4). They are usually light in weight and are
considered to be very convenient. The Zinc-MnOz battery is one of the most popular
primary cells and is used in many portable consumer products such as tape recorders,

calculators, portable radios, and toys. Other batteries in this category such as Mercad and

LilSOClg are

respectively a

 

 

 

 

 

Li/SOClz are used for more specialized high and low temperature applications

respectively and are limited by high cost (Table 1).

 
   
  
   
   
 

   
  
  
 
 

Cathode

Cathode Current collector

Cathode

Separator
with S
electrolyte eparator
wrth

electrolyte

Anode Anode
Anode

Current colle 0
Ct r Current collector

 

Jelly-roll construction Bobbin construction

Anode

  
 
 

- Separator
Cathode

Button cell construction

Figure 4. Common cell configurations

 

Battery
System

Aqueous
Mercad
Zinc-MM

Silver oxi

Aprotic O
LiMnO;

non-Aeuer

19800;

\

Reef.
high PW'er
efferent-ems.
conflSuration

am electrod‘

Table 2..1 B

   
  
  

flit‘lifil batten)
thal {he CHCI

.1 ‘ .

'h.
M are added

 

 

 

 

 

Battery Anode Cathode Voltage, Capacity, Energy Density, use temp,
System V Ah/kg Wh/kg Wh/L °C
Aqueous Alkaline Electrolyte Systems

Mercad Cd HgO 0.9 163 55 230 15 to 70
Zinc-MnOz Zn MnOz 1.5 224 125 330 -20 to 45
Silver oxide Zn Ag20 1.6 180 120 500 -20 to 70
Aprotic Organic Electrolyte System

Li/MnOz Li MnOz 3.0 286 230 550 -20 to 70
non-Aqueous

Li/SOClz Li 8002 3.6 379 320 700 ~100 to 70

 

Table 1. Examples of Primary Batteries

Rechargeable, or secondary batteries, are especially attractive because of their
high power output (constant performance as the discharge rate increases) and cost
effectiveness. Typical secondary battery designs such as the spirally wound “jelly roll”
configuration (Figure 4) aim to minimize internal resistance by emphasizing high surface
area electrode materials. Some examples of common secondary batteries are listed in
Table 2.1 By far the most common is the lead-acid battery used in automobiles. The
nickel batteries as well as lithium batteries are often used in consumer electronics. Note
that the energy density of primary cells is typically higher than secondary cells due to

their higher capacities. Capacities of secondary cells are always lower due to features

that are added so it can be recharged.

f

Battery Systc
lead-Acid

Agueous Ali
Nickel-Cad
Nickel-metal
hydride
Nickle-iron

M
Li ion-organ.
Li metalcrg.
Li metal-puts

 

 

 

 

 

Battery System Anode Cathode Voltage, Capacity, Energy Density, use temp,
V Ah/kg Wh/kg Wh/L °C

 

 

Lead-Acid Pb PbOz 2.0 120 35 70 ~40 to 60
Aqueous Alkaline Electrolyte Systems

Nickel-Cad Cd Ni oxide 1.2 181 35 80 -20 to 45
Nickel-metal (MH) Ni oxide 1.2 206 50 175 -20 to 45
hydride

Nickle-iron Fe Ni oxide 1.4 224 30 55 -20 to 70
Organic Electrolfie Systems

Li ion-organic C LixCoOZ 4.0 90 200 -20 to 70
Li metal-organic Li MnOz 3.0 286 120 265 -20 to70

Li metal-polym Li V6013 3.0 200 350

 

Table 2. Examples of Secondary Batteries

In both types of cells, the lithium batteries are far superior in energy density. A
high ED is expected for lithium cells due to the high standard potential of lithium
compared to its low equivalent weight. This makes lithium batteries very attractive for
use in portable, light-weight, high-powered consumer electronics. However, lithium
anodes can be problematic due to their sensitivity to water and reactivity with organic
electrolytes. Upon discharge, electron transfer reactions with electrolyte materials forms
a thin porous passivating layer on the electrode surface known as the solid electrolyte
interface (SE1). This layer acts as a protective barrier to further reaction with the
electrolyte while allowing access of lithium ions to the electrode. The SEI should be
porous and relatively thin, 50-250 nm compared to an electrolyte thickness of 50—150 um.
Unfortunately, cycling often results in replated lithium that is more porous, and with
higher surface area deposits. This further increases reactivity of the anode, even in

otherwise stable solvents. In most commercial lithium batteries, the lithium metal anode

 

has been repl.
ion cell the e.
significant ar.
The chemical

A variety or

 

appropriate e;

As di~
their behavit
Performtmce.
elemOlyte. t

 

“ample if
PtrformmCe
EMOum li‘.
electroI“e 2n

elecuglflc is

remand

appllCaUQn\

has been replaced by a lithium intercalated graphite material (Li:C = 1:6). In a lithium
ion cell the carbon anode (as well as cathode materials) can reversibly accept and donate
significant amounts of lithium without affecting its mechanical and electrical properties.
The chemical potential is almost identical to lithium metal (0.0-0.5 Volts versus lithium).
A variety of highly conductive organic electrolytes form stable SEIs, and given an

appropriate cycling rate, no redeposition of lithium as metal occurs.

As discussed, electrode materials define the overall voltage and ED of a cell and
their behavior during charge and discharge are significant factors that limit battery
performance. However, the overall performance of the cell also depends on the
electrolyte. If a poor electrolyte is used, even the best electrode materials won’t provide
acceptable cell performance. The materials used for the electrolyte are critical. For
example, if the electrolyte is not able to form a thin stable SE1, the cell will fail.
Performance issues such as chemical, electrochemical, and thermal stability are of
paramount importance. Once those issues have been satisfied, the conductivity of the
electrolyte and its behavior during cycling becomes the focus. The conductivity of an
electrolyte is highly variable depending on the salt used, salt concentration, and solvent.
The remainder of this work will concentrate on electrolyte properties for lithium battery

applications.

1. General
Ionic
conductivity

and can be re

Here qi is th.
R‘pmsented -,
aFilled field
“3310 see It
mobility. h.

Ck‘iiertdent.

 

Elmrolyte I.”
electrolytes. .
fillets 1'0“ 11‘.
elcctmlm c.

FUlchEr ("TE

 

II. Ionic Conductivity

1. General

Ionic conductivity results from the transport of ionic charge. The total ionic
conductivity in a sample, 0', is a sum of contributions from free ions, cations and anions,

and can be represented by equation 5.
0 = Z Qinibi (5)

Here qi is the carrier charge, ni is the number of carriers per unit volume, and bi (also
represented as ui) is the mobility of each ion relative to its average velocity due to an
applied field of unit strength. Since velocity is also a function of size and weight, it is
easy to see that small, light, electropositive metals like Li+ would be desirable for fast ion
mobility. In addition to the charge/size ratio, ion mobility is also highly temperature
dependent. High ionic conductivity is a result of ions being able to diffuse through an
electrolyte medium. Since it is difficult for ions to move freely in a crystal lattice,
electrolytes are rarely used below their melting, Tm, or glass transition, Tg, temperatures
where ion mobility is hindered. The relationship of the conductivity of a homogeneous
electrolyte can be described by an Arrhenius type equation called the Vogel-Tammann-

Fulcher (VTF) equation (6).

cm = A exp (-E./(T-T.» (6)

HCTC A 15 [hi
is the appart
conductivity

difference b.

2. Measurit

Ollll‘i

the resistant

COItductivitx

mClPIOCa] OE

and

where A (In

{GROWS then

 

 

Here A is the pre-exponential factor and is related to the number of charge carriers and E.
is the apparent activation energy for ion transport. T0 is the temperature at which the
conductivity is zero and is usually taken to be 50 K lower than Tm or T3. The larger the

difference between T0 and the use temperature, T, the higher the conductivity.

2. Measuring Ionic Conductivity - Impedance spectroscopy

Ohms law (7) describes the current, I as a function of the applied voltage, V and

the resistance, R of an electrolyte.

R: VII (7)

Conductivity, 0', is directly related to Ohms law in that 0' (in S/cm) is defined as the

reciprocal of resistivity, p (in Qcm)(equation 8)

o = Up (8)
and

P = R (M) (9)
where A (in cmz) is the area and l (in cm) is the length of the electrolyte sample. It

follows then that

O'=l/RA (10)

10

In or
necessary tol
practice. ele
stainless stee
the intercala

hence. only I

Rests
Si‘ictroscopy
containing c
inversely pr
dcP'Endent. ;
mitum. The

“hen Xbx

'
4 I

Mathemdlic -.

Where 2‘ IS .

51“ 9* andi -
{Epresenled .
I

. r m
q it No:

 

In order to characterize the conductive properties of the electrolyte alone, it is
necessary to minimize the resistive components of the electrode materials and the SEI. In
practice, electrodes of a highly electronically conductive and inert material, such as
stainless steel or platinum, are used in place of actual working electrode materials. Here,
the intercalation of ions and reactions with the electrode surface are minimized and

hence, only the resistive components of the electrolyte are measured.

Resistivity and therefore conductivity is measured best by impedance
spectroscopy (IS). Here an alternating voltage signal is applied to an electrolyte
containing cell (Figure 5), and the resulting ac. current measured. This current is
inversely proportional to the impedance, Z, of the electrolyte, and is frequency
dependent. Z contains both a capacitive, Xc and inductive, XL element and is complex in
nature. The real component of Z is the resistance and can be extracted from equation 11
when XL-Xc = 0.

2: [(XL-Xc)2+R211/2 (11)

Mathematically, the impedance is commonly expressed as.
Z=Z’-iZ” (12)

Where 2’ is the real component (2’ = Z cos 9), Z” is the imaginary component (2’ = Z
sin 0), and j is the imaginary symbol. Z is a vector quantity in the complex plane and is

represented at each frequency by a point on a complex impedance diagram termed a

Nyquist plot (Figure 6). At a specific frequency, where XL-XC= 0 and the applied

11

voltage and
difference. 6
intercept is

calculated.

 

voltage and resulting current come into phase with each other so that their phase angle
difference, 0, equals zero. At this frequency a linear spike intercepts the real axis. This
intercept is the real bulk impedance, Rb, from which p of the electrolyte can be

calculated.

oven~p

  

 

Impedance
Analyzer

    

 

polymer stainless steel electrode
electrolyte

Figure 5. Impedance analyzer set—up with variable temperature capability

12

1 0000

 

 

 

i
8000 f
E : o=(1/Rb)(L/A)
I: 6000 t
O a
f" E x
I}! 4000 : a:
X
F W‘Xxxx x x f
2000 i x b
if iii“)?
i-
0 L 111L111 1 11 iii

 

0 2000 4000 6000 8000 10000
2’ (Ohm)

Figure 6: Example of a typical Nyquist plot obtained from variable frequency

impedance measurements

3. Electrolytes for Lithium Batteries

Organic electrolytes for lithium batteries are made by dissolving a lithium salt
into an organic medium. Some common salts used are LiPFg, LiN(CF3802)2, LICF3SOz,
and LiClOa.2 The salt is chosen for its thermal, chemical, and electrochemical stability.
It must also have a low enough lattice energy that dissociation into ions in solvents is
relatively easy. Similarly, the solvent must have a high enough dielectric constant to
overcome the lattice energy of the salt and insure the dissolution of the ions. A good

solvent will dissolve a variety of salts and be chemically, thermally, and

 

electrochemic
VOlIS). Alllli‘

and therefore

 

ability of the
different sali~
used. usually
solvent vvher
LIVICFSO;
to have the h.

detail later).

 

oxide) (PEO‘

Lithit
molecules, y
“Penman“:

this hopping

a0

(I
'3

‘ fighteg

'1
LL16 [O the ‘

0100111“ 9‘

requnxib}

qurlbum

Each lUn .

electrochemically stable over the voltage window typical for lithium batteries (0 to 4.5
Volts). Although this medium can be a solid or liquid, the conductivity of the electrolyte
and therefore the performance of the battery at a given temperature, depends on the
ability of the ions to move freely from electrode to electrode. The conductivities of
different salts can vary several orders of magnitude depending on the salt and solvent
used, usually due to different degrees of ion aggregation or in the affinity of ions toward
solvent where the size and charge dispersion in the anion differ. In the case of
LiN(CF3802)2, also known as lithium imide, the large sulfonamide anion has been shown
to have the beneficial effect of decreasing crystallization in polyethers (described in more
detail later). As a result, the room temperature conductivity at O/Li = 27 in poly(ethylene

oxide) (PEO, Mn = 2000) is 1.8 X 10’5 S/cm versus 6.6 x 10'6 S/cm when LiC104 is used.3

Lithium ions move through interaction with lone pair electrons present on solvent
molecules, hopping from one site to the next. Several molecular simulations4 and
experimental FTIR, Raman, and impedance studies3 have been undertaken to understand
this hopping mechanism, and the interactions of lithium cations, anions, and their ionic
aggregates with several solvent hosts. The degree of “free” ions has been shown to
decrease at low temperatures and high salt concentrations (lower than O/Li = 20) in PEO
due to the formation of immobile aggregate species and is also highly dependent on the
mobility of the solvent.5 It has been well established that anions and triple aggregates are
responsible for a large fraction of the electric current in most electrolytes with very little
contribution from “free” lithium ion species. The fraction of the total current carried by

each ion is called the ion transference number. Anion and cation transference numbers,

14

t+ and t-. C
. _6
techniques.
measuremen:
the range of '.
lrnprovemen'.
chain or part;

positive effc.

a. Small "(7

Curre-
Obtain high I
lithium bait».
mmF‘eIItture,

Sinai] Organt
e”Wed. a1
SOlVemS are
Cihilene Car

‘ l‘.

 

dimethyl cad

0. a .
FTL‘RHIES '41

l

1
IO ' .

l“?
u°h COIt'duct
L

1+ and t-, can be measured by a number of electrochemical and diffusion-based
techniques.‘73 Pulsed field gradient NMR is nucleus specific and often allows the
measurement of both simultaneously. Transference numbers for lithium are typically in
the range of t+ = 0.2-0.4 depending on the temperature, solvent, salt, and concentration.‘5
Improvements in transference numbers can be made by tethering the anion to a polymer

chain or particle. This also reduces polarization at electrode surfaces which has an overall

positive effect on battery properties.

a. Small Molecule Electrolytes

Currently, most ambient temperature lithium batteries utilize liquid electrolytes to
obtain high ionic conductivities and good cycling characteristics. Current commercial
lithium batteries use small molecule organic electrolytes that are liquids at room
temperature. Some common solvents and their physical properties are shown in Table 3.
Small organic molecules from the carbonate and ether families are most commonly
employed, although rarely used alone.7'9 Most often, binary or ternary mixtures of these
solvents are used to obtain good conductivity and minimize crystallinity. For instance,
ethylene carbonate (EC) has a desirable dielectric constant for solvating salts, but is a
solid at room temperature and hence is a poor conductor of ions. Combining EC with
dimethyl carbonate (DMC) in a 50:50 ratio results in a liquid with good dielectric
properties at room temperature. DMC alone would not be desirable alone due to its
lower dielectric constant and relatively low viscosity. Although low viscosity leads to

high conductivity, it also contributes to the potential for electrolyte leaking from cells. In

15

addition. D

possible.

 

Stru

 

O’
ethylene
C .

of

g

pr Opylcnc

Hgao.

dimeth v i

Q
‘i-butyr. I

I
\
./

K

 

Tab

 

addition, DMC’s boiling point is relatively low and over-pressurization in cells is

 

 

possible.
Structure Mp Bp Dielectric const. Viscosity Conductivity
(°C) (°C) s (20°C) /cP(25°C) o(S/cm), 1M LiAsF6
it
o 0 36.4 248 89.78 1.93 6.97x10'3
\__/
ethylene carbonate
0
o/Iko -48.8 240 66.14 2.53 5.28x10'3
‘CH3
propylene carbonate
0
-2
macho/ma 4.6 91 3.12 0.59 1.1x 10
dimethyl carbonate
0
o: y 43 203 39.0 1.75 1.01 x 10'2
'y-butyrolactone
O
6 7 -109 66 7.75 0.48 1.29x10'2
tetrahydrofuran

 

 

 

 

 

 

 

 

Table 3. Some common small molecule hosts for electrolytes and their properties

Carbonates and other small molecule electrolytes result in excellent room
temperature conductivities that range from 10'3t010'2 S/cm.1o This is close to the
conductivity of aqueous/alkaline electrolytes (10'2t010'l S/cm). Unfortunately, aqueous
electrolytes are unstable toward lithium electrodes and cannot be used. Desirable
features of the carbonate and ether families are their electrochemical stability (up to 5.1 V

versus Li for carbonates and 4.0 V for ethers) and their ability to quickly form a thin

16

 

stable passii
unifomiity t

and am"

The
to their liqt; i
561332111011 0
501m recogr
due to volai

0rflame mo

 

stable passivating layer, the SEI (solvent electrode interface).11 The thickness and
uniformity of the SEI for several electrolytes has been characterized by FTIR‘Z, STM13,

and AFM14 and are thought to be composed of LiCO3 and alkoxides.

The advantages of small molecule electrolytes aren’t without complications. Due
to their liquid-like nature, separators (usually polypropylene) need to be used to assure
separation of the electrodes. This adds extra cost and decreases ED in cells. In addition,
some recognized safety hazards are leaking in improperly sealed cells, pressure build up
due to volatile components, and flammability of organic materials. In addition small
organic molecules can jointly intercalate with lithium into the electrodes, causing
electrode volume changes and cracking. This is especially problematic in propylene
carbonate (PC) containing electrolytes.15 All of these things can eventually lead to

failure of batteries.

b. Polymer electrolytes

In the late 1970’s Armand et al. showed that polymers could successfully replace
common liquids as electrolytes.16 Over the past 20 years, researchers have looked to
polymeric systems as the ideal electrolytes to improve battery technology. Solid polymer
electrolytes (SPE’s) have several advantages over their liquid counterparts. The
entanglements of polymer chains afford elastomeric strength and flexible mechanical
properties, and for this reason, SPE’s have often been termed “plastic” electrolytes. This
“plastic—like” nature, also allows for excellent adhesive properties, which in turn provides

for good electrode contact and also helps to stabilize the SEI. Polymers are inherently

l7

 

It

less volatile
pressum be

additional c

The
into electro.
battery. A:
SPE with 1..
high Curren
CD'Stalline
“Orphous
retlions are

1631016 and

Poly
Comm“ so.

Wigner hO'x.

 

less volatile and have safety advantages over small molecules. Corrosion due to leakage,
pressure build-up due to evolved gases, and the need for an electrode separator as an

additional component are eliminated.

The high molecular weight of polymer electrolytes prevents joint intercalation
into electrodes and can improve the lifetime of the electrode materials and therefore the
battery. Another advantage stems from the ease in creating very thin (50 um) films of
SPE with large surface-to-thickness ratios.2 This is an important feature since it enables
high current densities in batteries using polymer electrolytes, which are usually senti-
crystalline in nature. It is well known that ions only conduct in SPE’s through
amorphous regions where local segmental motion is possible. Therefore, crystalline
regions are highly undesirable. The large surface to thickness ratio helps to counter this

feature and allows for the potential use of semi-crystalline SPE’s.

Polymer electrolytes are usually made by dissolving a polymer and salt in a
common solvent and then removing the solvent to obtain a homogeneous SPE. Typical
polymer hosts have the same qualities as small molecule electrolytes in terms of electron
donor ability and polarity, thermal stability, and electrochemical stability. Oxygen,
nitrogen, and sulfur containing polymers such as polyethers, poly(ethylene imine),
poly(N-propylaziridine), and poly(alkylene sulfides) have been studied" (Figure 7).
Little success has been seen with nitrogen and sulfer contaning polymers. Crystallinity

and hydrogen bonding in poly(ethylene imine) and low salt solubility in poly(N-

 

propylaZinu

applications

Odee (FED

MW) ran 4..

 

At higher 31‘

propylaziridine) result in low conductivities of 10'6 S/cm, too low for practical

applications. Sulfur containing polymers have similar problems.

Poly(methylene oxide) ‘6‘OCH2%
Poly(ethylene oxide) AQOCHZCH2—>
Poly(trimethylene oxide) fOCHzC H2CH2}
Poly(propylene oxide) ‘60CHQOH‘>—

CH3
Poly(ethylene imine) . fCHchz—NHfi'
Poly(N-propylaziridine) fCHgCHg—IINI-a'

CHzCHgCHa

Poly(alkylene sulfides) -<-(C Han—8%— n = 2 - 6

Figure 7. Structures of potential host polymers for SPE’s

c. Poly(ethylene oxide)

By far the most successful host to date is the ether based polymer polyethylene
oxide (PEO). Polyethylene oxide is made by the cationic or anionic ring opening
polymerization of ethylene oxide and can result in polymers ranging in molecular weights

(MW) ranging from 1000 to 5 x 106.2’18’19 Below MW = 600, PEO is a viscous liquid.

At higher MW ’3, PEG is a waxy solid with a glass transition temperature (Tg) near —65°C

19

 

and a melti

history. Pr

W.

ETI'ST' It.

and a melting temperature (Tm) anywhere from 60-70 °C, depending on MW and thermal

history. Pristine PEO readily crystallizes in a 72 helical structure (Figure 8),

 

 

 

 

-a--

Figure 8. Side and top view of the helical structure of PEO.

which contains seven ethylene oxide repeat units with two turns in a fiber period of 19.3
A. Under tension PEO can also crystallize in a planar zigzag conformation?“21 The
structural features of PEO under various conditions have been confirmed through X-ray

analysis and extensive spectroscopic studies?"29

When PEG and other polymers are cooled from their melts, they produce semi-

crystalline morphologiess'30 Layer-like crystallites are separated by disordered regions

20

 

which lead
crystallizab'.
crystalline p
specific 1028
is hard to ac.
are often It
primarily by
mobility of 1
with the mm
In Chitin fold
CD'Staliizati.
SPhleiteg_
“Tim exzrm

 

0X\ a
'g~n Un‘]

which leads to a two phase system of crystalline and amorphous regions. Non-
crystallizable portions of polymer chains accumulate in the amorphous parts of a partially
crystalline polymer and include entanglements, endgroups, short chains branches, and
specific local conformations that oppose transformation into a uniform chain. Because it
is hard to achieve ordering over a very long range, the polymer chains in polymer crystals
are often folded. Unlike small molecules, crystallization of polymers is governed
primarily by kinetic criteria rather than equilibrium thermodynamics. Due to the limited
mobility of the polymer chain, the structure that develops at a given temperature is that
with the maximum growth rate rather then the lowest free energy. Therefore, an increase
in chain folding and growth rate is observed with increased super cooling.31 The result of
crystallization in layers is the formation of spherical crystals of polymer called
spherulites. Spherulites are formed by nucleation followed by even radial growth.32
When examined using a polarizing optical microscope, spherulites exhibit a characteristic

Maltese cross pattern of light extinction.33

The electrostatic interactions (8 = approx 5), electron pair donating ability
(approx 22), and most importantly, the optimal spatial distribution of the solvating
oxygen units, make PEO a superior SPE candidate. Polyether hosts such as
poly(methylene oxide), (CHzO)n, poly(trimethylene oxide), (CH2CH2CH20),., and
poly(propylene oxide), (CH2CH(CH3)O),,, are thought to be unable to adopt the
conformations necessary for multiple inter- and intramolecular cation-polymer
coordination. As a result, less salt is dissolved and there are less charge carriers to

conduct.34

21

 

 

Eve
conductivity
PEO at OIL _
its perform.
mechanical
hence low

indicate he

 

Crystalline :
On heating
l‘héfic result

and allow tl

Ft

Even though PEO can dissolve a variety of salts well, its room temperature
conductivity is still quite low. Figure 9 illustrates the temperature dependent behavior of
PEO at O/Li = 6.35 It is clearly seen that although PRO is a good conductor above 60 °C,
its performance drops off rapidly as the electrolyte crystallizes. It seems that the good
mechanical properties at room temperature come at the expense of immobile ions, and
hence low conductivity. Phase diagrams of PEO with several salts are available and
indicate both amorphous and crystalline regions of PEO-salt complexes. In the
crystalline state, lithium is known to preferentially coordinate with 3-4 oxygen atoms.6
On heating above 60 °C, the ions become more mobile, but the loss of the crystalline

phase results in poor mechanical properties, which can result in creep of the electrolyte

and allow the electrodes to come into contact and short-circuit the cell.

 

1 .E-02
1 .E-03
/ Tm

1.E-04

1 .E-O5

o (S/cm)

1 .E-06

1 .E-O7

 

 

1.E-08~ e . - .
2.4 2.6 2.8 3.0 3.2 3.4

1000/T (1/K)

Figure 9. Temperature dependent conductivity of PEO/LiClO4 (O/Li = 6)

22

 

 

while main
ways to di
amorph00s

the strategy

 

technique is
chments (3

methylene

11.3537 Th

HFOC

l*‘ere 310d
5x 105 1
Some Oll‘r
phlfie

it

it.

may .

$611.6

c. Advances in PEO based SPE ’s - Amorphous Polymer Architechturcs

In an attempt to improve the conductive properties of PEO based electrolytes
while maintaining good mechanical properties, researchers have come up with several
ways to disrupt PEO crystallinity and increase the fraction of the highly conductive
amorphous phase. Redesigning polymer structures to create amorphous materials follows
the strategy of creating less regular structures by disruption of crystalline segments. One
technique is to create defects in the regular pattern of linear chains as in methoxy-linked
segments of PEO (Mn=400 for PEO segments), or PEMO. PEMO, a copolymer of
methylene and PEO, was first synthesized and characterized by Booth et al. (Scheme

1136'” The electrolyte properties of PEMO (M., = 50,000, O/Li = 25 using LiCF3so3)

KOH
H(—OCH2CH2>OH ——> CHZfOCH2CH2>rCH2-<OCHZCH27\r—n—}
m CHZC|2 m n

m = 4 - 14 PEMO

Scheme 1. Preparation of methoxy-linked PEO (PEMO).

were studied using impedance spectroscopy and was found to have a conductivity of O' =

5 x 10'5 S/cm at 25 °C.:38 Similar results were obtained recently by Kerr and Buriez.39

Some other commonly used techniques to increase the volume fraction of the amorphous
phase include synthesis of random or blocky copolymers, comb polymers, and
crosslinked networks (Figure 10). Typically the polymer is made from PEO segments

that serve as the conducting medium, linked by a highly flexible polymer segment or

23

 

backbone.

than PEG '4'

'llt’

P01}.
P013dim 61h
Copolylnert
Combinatjo,
morn [9mm

W‘
J] pOOr n

 

backbone. The highly flexible polymers are characterized by a low Tg (usually lower

than PEO at —60 °C), and must be thermally, chemically, and electrochemically stable.

/——/
6‘blocky,9 _

“random” M

l 7 L
°branched copolymers
) '7

~copolymers

°network polymers

 

Figure 10. Polymer structures designed to limit crystallinity

Polyethylene (Tg = -128 °C), polyphosphazene (Tg = -70 °C) and
polydimethylsiloxane (Tg = -123 °C) have been used as the backbone in many comb-type
copolymers that incorporate PEO segments (Figure 11). A common result of such
combinations is seen in poly(methoxyethoxy ethoxyphosphazene) (MEEP), which has
room temperture conductivities of 10‘1 S/cm, 3 orders of magnitude larger than PEO but

with poor mechanical properties.40

24

 

p01

Fig

50!]

Ch‘
standing n
crysu‘tlhml
polyethylfi
stability. I

of losing 0

In 1
satisfy botl
designed a

lFigure 12

Method for

Base 2012mm incorporation of PEO

R
polyphosphazene ‘6N=ll3% For MEEP...
II; R = (QC HchzlzocHa
(EH3 C|H3
Polydimethylsiloxane fSI‘J—O‘jr f§i—O
CH3 (C H20H20)nc H3

polyethylene fCHZC Hg‘j' ffiDHC H2—>

(C HZCH20)nCH3

Figure 11. Common low 'I‘g polymers used to increase the flexibility of PEG and

some examples of how PEO is incorporated into the structure

Chemical cross-linking“ and blending42 MEEP with PEG resulted in stable free
standing materials but with no gain and even slight losses in conductivity. Limiting

43,44 and

crystallinity by copolymerization of PEO with polymethyl disiloxanes
polyethylenes45 resulted in conductivities of 5 x 10'4 S/cm, but at a loss of mechanical

stability. In addition, polysiloxane bonds are susceptible to hydrolysis. Overall, the trend

of losing one property for the other has been difficult to rectify.

In recent years, more complicated polymer architectures have been designed to
satisfy both the conductive and physical properties requirements of electrolytes. Janasch
designed a comb-type polymer based on a polyether functionalized polystyrene backbone

(Figure 12).46 PEG and a random compolymer of PEO (80%) and polypropylene oxide,

25

 

PPO. (20‘??-

than polyl‘.

 

 

from 22 CC I

I
further depl
hydloplmb;

Mechanics"

electrolues

 

[0J5 Sl'cm
respective}
FPO P01}‘n

0C due to <

PPO, (20%) were grafted (Mn = 2000) onto the para postion of a styrene monomer and
then polymerized (Mn = 32,000). Addition of PPO decreased the melting temperature,
from 22 °C (100% PEO) to —16 °C. End capping with hexadecanoyl chloride resulted in
further depression of Tm to -21 °C. It was concluded from DSC experiments that the
hydrophobic end-caps phase separated and therefore acted as physical cross-links.
Mechanical stability was mainly attributed to the polystyrene backbone. At 20 °C
electrolytes using LiCF3SO3 (O/Li = 40) gave conductivities of 1 x 10's, 8 x 106, and 4 x
10" S/cm for polymers with end-capped grafts, non-capped grafts, and PEG grafts
respectively. Smooth conductivity curves were observed from —20 to 100 °C for PEO-
PPO polymers while those using PEO alone showed a steep decline in conductivity at 30

°C due to crystallization.

CH3

0%CH20H20'2TH QaEéCH2CH20>féHCHZOfiH

O I 0.8 0.2

A) PEO only B) non-capped PEO-PPO

2% °
egeeesoieeeeesfiiccw CH
3
08 7?.

0.2

O C) end-capped PEO-PPO

Figure 12. Structures of polyether functionalized polystryrene

26

 

segment n:

P0l}'€ICCll' -

(Ol'Ll = 43
have more
has been I?
branched .3
Hence. it l

ngldity of 1

“ti
1‘ lmmoblt
Shon E0 t“
the Shear m
were ”Cree;
a Slmllar.

df‘ieltlpcd

Willem

C(Wf

”WmEhs

SPE

ilit]
l,n 1,0 I“
Ate SM

 

Several research groups highlighted the importance of ethylene oxide (E0)
segment mobility in both main-chain, branched, and network copolymer systems.3'39'47
Polyelectrolyte networks formed using tri-functional urethane crosslinkers and PEG of
M, = 600 and 2000 gave conductivities of 2.0 x 10‘8 and 4.5 x 10'5 S/cm respectively
(O/Li = 43, Li imide, 20°C).3 The longer PEO chains in the network were determined to
have more segmental motion. A key observation in main chain and network polymers
has been that E0 segments in the main-chain are limited by extended chain mobility, but
branched systems with short E0 segments have fast molecular motion in the side-arms.48

Hence, it may be possible to tune mechanical and conductive properties by tuning the

rigidity of the backbone and the length of the side-arms.

Watanabe showed that methylation of short E0 chain ends leaves them free to act
as immoble internal disrupters of crystalline regions.49 As the degree of methylation of
short E0 chains in a branched network polymer electrolyte increased from 0-65%, both
the shear modulus and Tg decreased, while conductivity increased. Although these SPE’s
were “creep-free”, conductivity never exceeded 10’5 S/cm at room temperature. Recently
a similar, but improved class of network PEO electrolytes (Scheme 2) has been
developed with hyperbranced side arms with both good mechanical properties and an
ambient conductivity of 1 X 10“t S/cm.50 A small amount of crosslinking is observed due

to some diacylated monomer and possible photocrosslinking.

SPE’s of this type could be potentially used in rechargeable batteries, particularly

for large scale industrial use, electric vehicles and uninterrupted power supplies operating

27

/O\ O
CH30C H20 H200 H2CH200 HZCH—CHZ + CH: 3C H2

CHaOC H20 HQOC HQCHZOH
KOH

 

V

CH30CH2CH20CHQCHZO+E€CH2THOt<CH2CHZOfi~OH
n

CHgOC H2C H200 H20 H20

CH2=CHCOOH
PTS/HQ

 

V

CH30CH2OH20CH2CH20€€CHZTHOfiCHQCHZOfiCOCH=CH2
n

CH30CHQC HzOCHzCHgO

UV radiation
DMPA
8r Metal-Salt

Network Polymer
Electrolyte

Scheme 2. Synthesis of a hi ghly conductive network polymer. PTS =
p—toluenesulfuric acid, HQ = hydroquinone, DMPA = 2,2-dimethoxy-2-
phenylacetophenone, DMPA = dimethoxyphenylacetate

28

 

at modern:

often leqn:

 

ti Gel Elt‘
Re
Abraham t

SPE incorij

 

the “gel e
molecule

netwom l.
large C0”
WWII] lt‘
SL”Tommi
Slime ‘11 1
E5] is f

UUOUEh

Synthe

at moderately elevated temperatures. Unfortunately, complicated multistep sythneses are

often required, making production of these systems costly and time consuming.

d. Gel Electrolytes

Research on polymer gel electrolytes has rapidly grown since 1990 when
Abraham and Alamgir first reported liquid-like conductivity of up to 4 X 10‘3 S/cm for a
SPE incorporating polyacrylonitrile and a carbonate based electrolyte.51 The idea behind
the “gel electrolyte” is that a high molecular weight polymer is swollen with a small
molecule electrolyte, usually termed a plasticizer. A physically cross-linked polymer
network (formed through chain entanglements or chemical crosslinking) is filled with a
large continuous phase of electrolyte. When sufficiently large, these pockets conduct
lithium ions as an unhindered liquid. Meanwhile, mechanical stability is imparted by the
surrounding network. Ease of preparation makes gel electrolytes particularly attractive
since it requires simple mixing of components. In some cases a chemically cross-linked
gel is formed by including a functionalized polymer or monomer and cross-linking

through UV irradiation, electron beam irradiation, or thermal treatment.

A random PEO-PPO trifunctional acrylate monomer (Figure 13) was recently
synthesized and polymerized (photopolymerization using DMPA photoinitiator) in the
presence of several small molecule electrolytes.52 The wt% of the macromonomer was
varied from 25 to 75 wt%. The room temperature conductivity of gel electrolytes

containing 35 wt% and 10 wt% macromonomer (1 M LiBF4 in y-butyrolactone) was 2 x

10’3 and 5 x 10'3 S/cm respectively. Films of the gel electrolytes are claimed to have an

29

 

elastic-like

versus thel

 

Fig

Va;
fluoride) =l
hexflllUQn.
demonic.
0&3“ Used
Allhough :
and meeh.
examPle, ;
Stabll

it." p:

it,

 

10!"
p "mef St

 

 

elastic-like nature, even for >80 wt% electrolyte, and show improved cycling capacities

versus their liquid counterparts.

9H3 0
CHg—O—ECHQCHZO CHCHZOHJK:
9()—,‘g 0.2 n
9H3 °
CH—O—gCHZCHZO CHCHZ =
‘20—86 0.2 n
9H3 0

 

Figure 13. Polyether macromonmer used to form gel electrolytes.48

Various polymer hosts such as polyacrylonitrile (PAN),53'55

poly(vinylidine
fluoride) (1>le=),5“'58 poly(methyl methacrylate) (PMMA),59'6° poly(vinylidine fluoride-
hexafluoropropylene) (PVdF-HFP),61'62 and PEG”66 have been proposed to form gel
electrolytes. As might be expected, the carbonate family of small molecule plasticizers is
often used as the conducting phase, as well as small molecule and oligomeric ethers.
Although the above mentioned gel electrolytes have excellent conductivities (10'3 S/cm)
and mechanical properties at ambient temperatures, there are some drawbacks. For

example, large volume ratios of liquid electrolyte (up to 85%) can cause mechanical

stability problems, particularly at elevated temperatures.5

Mechanical stability is not a concern for PVdF-HFP containing systems as the

polymer structure is highly insoluble and rigid. Unlike other gel-type systems, electrolyte

3O

 

is diffused int
Z-phase electl

being comme

Wesli‘
PEO~LiClOr
over 100 CC
ceramic dlSpt

Since then. 5

9- “110.1,“ (
1' diSperspd

Inen ;
Sl'SlemS’ and
ShOWn to in.
demOnStrata
increase in C
that the Exw

l -
he "mdller I...

 

is diffused into a porous PVdF-HFP film and no real mixing occurs. For this reason these
2—phase electrolytes are often classified as porous SPE’s.67 This SPE system is currently

being commercialized for secondary lithium batteries.

Weston and Steele showed that adding inert inorganic fillers such as a—Aleg. to
PEOeLiCloa improved the mechanical properties of amorphous polymers at temperatures
over 100 °C (in the liquid regime).68 This improvement was explained assuming the
ceramic dispersoid provided a supporting matrix for the amorphous polymer complex.
Since then, several research groups have added inert inorganic filler materials to their
gelled systems to reinforce and improve mechanical propertiesag’n Still, the volatility of
small molecule electrolytes in these systems is a detriment to the safety and lifetime of a

battery, a problem that some have remedied by using oligomeric PEO’s.72

e. Filler-based electrolytes
1. dispersed fillers

Inert inorganic fillers improve the mechanical properties of amorphous electrolyte
systems, and finely dispersed inorganic fillers such as $02, A1203, and TiOz, have been
shown to increase the amorphous phase of semi-crystalline SPE’s. This was initially
demonstrated by Liquan, who observed both improved mechanical properties and an
increase in conductivity by adding 'y-Ale3 to a PEO-NaSCN complex.73 He suggested
that the extent of the latter phenomenon depends greatly on the particle size of the filler,

the smaller the particle size, the better ability to disperse evenly in the polymer matrix.

31

 

cryst
anal;
lDS(
nunl

AlgC

When highly dispersed in a polymer matrix, ceramic fillers affect the
crystallization rate of the polymer.”75 Several analytical methods such as impedence
analysis, Raman spectroscopy, X—ray, NMR, and differential scanning calorimetry
(DSC), have been used to probe the crystallization charateristics of PEO containing a
number of fillers. Figure 14 illustrates the results of impedance analysis of 10% wt 7-

A1203 in (PEO)3-LIC104.76 Without added filler the electrolyte resistance, which is

 

 

 

 

 

 

 

 

1.00E+06
AA
A
8.00E+05 ~ ‘ ‘
g A
5 6.00E+05 r
d) . .
g o with alumina tiller
(6
E; 400505 A no tiller
(0
0)
tr
2.00E+05 ~
A o
4: 0 ‘” ‘” . .
0.00300 ‘ ‘ 4
0.0 5.0 10.0 15.0 20.0

Time (103 min)

Figure 14. Effect of alumina filler on the resistance of a PEO(8)LiC104 SPE at
30°C. Adapted from Croce et al.72

32

 

int

13

SUI

do

it

inversely related to conductivity, increases due to formation of a crystalline phase. The
importance of particle size is shown in Figure 15, where 10 wt% of 2-7 pm 0-A1203
powders were dispersed in PEO.77 Composites made with smaller particles (higher
surface/ area ratio) show superior, more consistent bulk conductivities. Larger particles
do not sufficiently inhibit crystallization as seen by the sharp inflection point at 1000 /

T(K) = 2.9, the approximate Tm of the PEO salt complex.

 

 

 

 

-3 i 6
It
I 0 ,
g /v' 3
a -5 4w" 0 .
b | I °
3 -6 l- . I
. I
-7 _
him/V.
O
'8 r .1 r r I r
2.5 3 3.5
1000/T (K)

Figure 15. Effect of filler particle size (10 wt% A1203) in a PEO(10)NaI SPE.
Adapted from Croce et al.7

33

It is proposed that the mechanism by which inorganic particles disrupt
crystallinity is related mainly to surface chemistry. Polymer chains, such as PEO, can
interact with polar groups, such as surface silanols (~OH) on SiOz particle surfaces. The
dipolar and hydrogen bonding interactions in these groups can act as cross-linking centers
for PEO segments, thus lowering the reorganizational tendancies of polymer chains and
inhibiting the recrystalization kinetics.78 The higher the surface area of the particle, the
more interaction with polymer chains, and hence, the greater the stability of the
amorphous phase. An added benefit is that these inorganic powders have the ability to
trap remaining trace liquid impurities such as water.79’80 Therefore, it has been shown
that inorganic fillers actually improve lithium interfacial properties. It has also been
suggested that structural modifications of the polymer chains at the filler surface may
induce highly conducting Li+ pathways at the surface. This has led to research on
ferroelectric materials such as BaTiO3 as ceramic fillers,81 since ferroelectric materials
have a permanent dipole moment that may expand the highly conductive surface

electrolyte layer.

Further evidence of surface effects comes from recent findings that up to 20 wt%
of inorganic fillers increases the conductivity in amorphous hyperbranched PEO
electrolytes. Since these electrolytes cannot crystallize, it was suggested that surface

interactions account for a plasticizing effect in the polymer.82

34

 

CC

sh

ell

iii

2. aggregated fillers

While composite polymer electroytes (CPE’s) made from high molecular weight
PEO have excellent mechanical and conductive properties at elevated temperatures, the
conductivities at ambient temperatures are still too low (10'8 to 10'5 S/cm). It has been
shown that at ambient temperatures, highly dispersed silicas (HDS’s) form mechanically
stable structures in small molecule liquid electrolytes through hydrogen bonding. At
loadings as low as 4% by weight, HDS’s form loose gel-type structures in mixtures of
EC, PC, and DME carbonates, with conductivities >10'3 S/cm.83 If hydrogen bonding
improves the mechanical properties in a polar medium, the effects should be even greater
in a non-polar system. A transmission electron microscopy (TEM) image of 5 wt%
hydrophilic silica (SiOH functionalized, Degussa A200) dispersed in mineral oil, shows
a hydrogen-bonded network with up to 200 nm pockets of continuous mineral oil phase
(Figure 16).84 Because of the large difference in polarity between silanol functionalized
A200 and the non-polar oil, a stronger network should form compared to the previous

85 For

example. Similar network structures are formed in reverse phase systems.
example, a hydrophobic filler, such as R805 (50% n-octyl modified A200) was dispersed
in a polar oligomeric PEO (Mn = 500) salt mixture. In both cases the driving force for
network formation is the difference in polarities of the filler surface and the continuous

phase. In the first case, hydrogen bonding affords mechanical stability, in the second

case van der Waals forces.

35

 

Figure 16. TEM image of 5 wt% A200 in mineral oil

 

 

"
all:

Both composite systems are considered to be thixotropic in nature (Figure 17) in
that the networks aren’t permanently/chemically linked and can be disrupted by shear

with a resulting decrease in viscosity. Upon resting, the network is reformed and the

0 0 0 0 0

ago go°oo°o 3“ ? Q

0080 000000 ——> a

960 °g%°o°o° 3 a g

no 000 0 o e of 589
primary aggregates
particles (low viscosity)

ii

.0". ‘4’
(

-'. . ‘o‘o

1"‘6 6 .- . O
{'0 I '
.- i' . .
‘ I ' O

O D. .- ’4 (’ .
0 u 'Oce'o .0

r v' .3
‘ 3) ‘3‘"
t, o 4., oc

a
‘
’ .0 t

 
   
      
    
    
    

  

shear

M
I . 0.
<——-—
s_e ——>
o r
:2 rest
O D

   
    

.0
0

3-D aggregates agglomerated
(low viscosity) network
(high viscosity)

Figure 17. Schematic showing the thixotropic behavior of fumed silica networks

thickening of the composite is re-established. Thixotropic materials are particularly

attractive for manufacturing purposes as they flow easily into various shaped molds and

regain mechanical stability upon resting.

The degree of association between various particles in the electrolyte, and
therefore the mechanical stability of these composites is an important parameter to assess.

While TEM gives a visual picture of the degree of network formation, one of the most

37

 

iii“.

m5

obi

Sll’

sensitive techniques used to probe the mechanical stability of networks is dynamic

rheology.

111. Mechanical Properties

1. Dynamic Rheology

Dynamic rheological experiments probe the strength of network formation by
applying a low—amplitude sinusoidal deformation (7) to a sample at a frequency ((0) and
maximum strain amplitude (yo) (see equation 13).86 Since thixotropic networks aren’t
permanent and may be disrupted at very high strain amplitudes, it is important that the
maximum 70 be within the linear viscoelestic regime (ie, the network must stay intact
during dynamic frequency sweeps at a given strain value). A suitable strain amplitude
can be determined by holding wconstant and varying the strain until a yield stress is
obtained. This yield stress is indicative of the onset of structural disruption at a given
strain value.

Y: Yo Siam!) (13)

The sinusoidal stress reponse (I) of the sample may be written as in-phase and out-of-

phase component as shown in equation 14.

I = G'yo sin(0)t) + (3"70 cos(0)t) (14)

38

. n '1'“ ‘

 

(II

CL)

The in-phase component is attributed to the energy stored in the sample and thus
defines the elastic modulus, G’, a measure of the solid-like character of a material. The
out of phase component is attributed to energy dissipated and thus defines the viscous
modulus G", a measure of a material’s liquid—like character. In short, for liquid samples
G" dominates, and for solid samples 6' dominates. The values of these moduli as a

function of frequency are used to characterize materials.

Figure 18 shows a standard parallel plate configuration where a composite is
sandwiched between two stainless steel plates at a given thickness. A schematic showing
yield stress determination and the follow up dynamic strain and shear experiments
appears in Figure 19. First, a variable strain experiment determines the strain level
where the elastic network breaks down (the yield stress). The shear experiment is then
run at a low enough strain level to preserve the network while varying the frequency. For
solid-like behavior, a constant, frequency independent modulus is observed with G’ > G".
A rigid physical gel with significant network formation should have an elastic modulus of

approximately 103 Pa.87

" f \ Parallel Plates

Sample

 

Figure 18. Standard parallel plate configuration for dynamic rheology experiments

39

 

ll__

Yield Stress (1,. = (3’?)
A A

 

 

 

 

 

 

 

Ga
/ G,
<51 G” 5”:
'0 (D
g) 3 G"
.‘I/
> >
% Strain (987) Frequency, (log 0))

Figure 19. Variable strain and stress experiments

IV. Previous Work

1. Composite Polymer Electrolytes

Previously, Hou studied CPE’s using low molecular weight polyethylene glycol
dimethylethers (PEGDME’s) of Mn=350-600, LiClOa, and several types of fumed
silicas.84 Salt concentration, molecular weight of the PEGDME’s, and wt% filler were all
optimized to give the best combination of mechanical properties and conductivity.
Several fumed silicas provided by Degussa were used and many were custom
synthesised. A200 was selected for chemical functionalization due to its high surface
area, abundance of surface hydroxyls, and the availability of small primary particles with
diameters of approximately 12 nm. The salt, LiClO4, is readily available, dissolves easily

in PEG, and is chemically and electrochemically stable. An O/Li ratio of 20 — 40 gave

40

Opt

COl

t'is

 

Nix

I
n.1,

p
l

Ni. l_

0r ll

optimal conductivity. It is well known that at low salt concentrations (O/Li > 40),
conductivities are low due to lack of charge carriers. At high salt concentrations (O/Li <
20), ions and ion aggregates act as transient cross-links or as agents to increase the local

viscosity of polymer chains. This results in a more rigid system and lowers conductivity.

Studies of the mechanical properties of CPEs determined that at least 5-15 wt%
filler was needed to form a strong network and impart good mechanical pr0perties. More
hydr0phobic fillers such as R805 required only 5 wt% in PEGDME-SOO electrolyte to
form rigid gels with a frequency independent 6’ of 103 Pa. Composites made using 5
wt% of a less hydrophobic, methyl functionalized silica, showed liquid-like (frequency

dependent) behavior and 6’ values of onlleo- 101 Pa due to poor network formation.87

The conductive and mechanical properties of fumed silica CPEs are essentially
decoupled (Figure 20).?”88 Negligible losses in conductivity are observed as filler is
added, while the mechanical properties significantly improved. These results point to an
important characteristic of the composites, that the mechanical and conductive properties

can be optimized independently.

Functionalizing the silica surface with both hydrophobic octyl groups and
crosslinkable propyl methacrylate groups, (Scheme 3) gave strong thixotropic networks
through van der Waals interactions, which could be permanently stabilized through UV
or thermal crosslinking. Since very few interparticle connections were made using filler

alone, small amounts (10 wt%) of a methacrylate monomer (usually butyl methacrylate)

4l

 

“in

in:

p01

all

 

 

 

 

 

106g ’ ° ~ o . “2w » a.
A 5 "' o —t
M -
e, i '2 103 3
(D 105 r ‘ 5?
m“ : ‘ 3+
2 I d :3.
8 : ~ E
g 9
4 _ ~
.9 10
i3 : . 7er
_<_u _<———————
LU _ 333
103 _ PEG-0M, Li lmide ‘w'
: i l l r r r l A l r r r i u r l 1 l r l r 1 10-4
0 5 10 15 20 25

Fumed Silica (R805) weight °/o

Figure 20. Conductivity and modulus of fumed silica composites as a function of wt%
filler (PEGDM, Mn = 250, CID = 20).

were added to improve connectivity. As a result, plastic-like CPE’s were formed
with a two order of magnitude increase in modulus with little cost to conductivity.89 The
increase in modulus is thought to result from significant phase separation and
polymerization of butyl methacrylate (BMA) monomer onto the filler surface. Therefore,

a more rigid and permanent network was created.

Another benefit of filler-based CPE’s and other solid-like systems is the improved
stability of the SEI. Essentially, the SEI is mechanically stabilized by solid-like
electrolytes, which prevents exfoliation of the SE] from the electrode surface and

subsequent reaction with more electrolyte. In addition, the filler may act as an impurity

42

 

HO OH
H. g ..
HO OH
OH
A200
0
WSKOCHals
CH3
v HNtCH20Hala

 

FS(TOM)

Scheme 3. Synthesis of FS(TOM) silica from A200

43

scavenger. Recent cycling studies at N .C. State have shown that added butyl
methacrylate monomer increases the interfacial resistance in lithium button cells from 6 x
102 Gem2 to l x 103 Qcmz, almost half an order of magnitude. Since some monomer is
left unpolymerized after crosslinking, unreacted monomer can diffuse to and react at the
electrode surface, causing a thickening of the SEI and therefore increasing the resistance

at the electrode surface.

In continuation of previous work by Hou, further studies of filler surface
characteristics, the hydrophobicity of methacrylate monomers, and the polymerization
environment have been undertaken to further understand and optimize CPE’s made using
FS-TOM fumed silicas. The results of kinetic studies are presented in later sections that
elucidate the fate of the monomer and provide a better understanding of the distribution
of monomer in the composite and how it evolves during polymerization. Before these
results are presented, a brief discussion of suspension and some basic concepts of radical

polymerization will be presented.

V. Monomer Behavior

1. Suspension Polymerization

In atypical suspension polymerization, one or more water-immiscible monomers
containing the polymerization initiator are dispersed into droplets in an aqueous medium
by strong stirring or mechanical agitation. A suspending/dispersing agent is also added to

hinder coalescence of droplets by decreasing the surface tension between monomer and

the aqueous phase. Suspending agents are most effective when present in the surface
layers between the water and monomer droplets and can be water soluble polymers such
as polyvinyl alcohol, polymethacrylic acid salts, and PEG, or insoluble powders, such as
hydroxyapatite, TiOz, talc, and Al hydroxide.90 Once formed, spherical droplets, usually
0.1 to 3 mm in diameter, are polymerized. Because the product of suspension
polymerizations are polymer beads, this process is often termed “bead” polymerization.

A schematic representation of suspension polymerization is shown in Figure 21.

 

 

 

 

 

 

 

 

 

 

 

 

 

Figure 21. Representation of a suspension polymerization where D = dispersant

and I' = initiator.

45

Because each monomer droplet is in essence a mini-reactor, suspension
polymerizations are often thought of as water-cooled bulk polymerizations. Early kinetic
studies on bulk versus suspension polymerization of MMA and styrene have shown good
agreement in time-conversion curves, heats of polymerization, and dependence on
initiator concentration. Rates of polymerization were not much influenced by bead size

or type of suspending agent.91

Although cross-linking of CPE’s does not result in beads of pristine polymer, the
suspension polymerization model should have many parallels. An idealized picture of
the CPE network is shown in Figure 22. A hydrophobic methacrylate monomer such as
BMA is dispersed over the hydrophobic silica surface (Figure 22A). Since BMA would
normally phase separate from PEGDME-SOO, the filler acts as a dispersing agent in the
polar PEGDME-SOO electrolyte phase. Note that AIBN is an oil soluble initiator, and is
dissolved in the monomer prior to dispersion. Polymerization then proceeds through
either photochemical or thermal cross-linking following bulk-like polymerization
kinetics. This is an idealized picture. The adsorbtion of monomer to the filler surface
depends greatly on the monomer hydrophobicity compared to that of the solvent. If the
monomer is too hydrophilic the monomer will dissolve in the electrolyte phase leading to
a solution-type polymerization (Figure 22C). An intermediate case is illustrated in
Figure 22B. The kinetic differences between suspension (bulk-like polymerization) and

solution polymerization will be presented in an upcoming section.

46

 

 
    
        
 

monomer &
initiator

silica
PEGDME-500
LiClOa

 

 

 

 

 

 

 

Figure 22. Three scenarios for monomer location in CPE’s. A) Completely phase separated
and absorbed onto filler surface. B) Intermediate case between A & C, plus some phase
separation in the electrolyte phase. C) Completely dissolved in the electrolyte phase.

47

2. Phase Separation
Mixing behavior of two solvents is governed by the free energy of mixing

(equation 15).

ACimix = AHrrrix — TASmix (15)

where AG“" is the Gibbs free energy change upon mixing, AHmix is the enthalpy change
on mixing, T is the absolute temperature, and A8,,“" is the entropy change on mixing. A
negative value for AGm,x means that the mixing process will be spontaneous while a
positive value leads to two or more phases from the mixing process. Polymers have a
very low entropy of mixing due to their high molecular weight. Therefore, AHmix is

usually the deciding factor in determining the sign of AGmix.

AHmix is related to the polarity of two substances to be mixed, and can be

expressed as shown in equation 16 in terms of their solubility parameters 61 and 82.

AHmix/V = (51 - 52)2 (Pl<P2 (16)

Here V is the volume of the mixture and <01 and (p2 are the volume fractions of each
component. The solubility parameter, 8, of each component is a measure of the cohesive
energy density (CED), AE", of each (equation 17). The CED is the energy of
vaporization per cm3. Polar molecules or those with H-bonding usually have large values

of AE".

48

5: (AEV/ V)“2 (17)

In equation 16, the solubility parameter difference, A5 = 61 — 52, must be small for
misciblity. A guideline for the miscibility of polymer-polymer systems has been
developed.92 For polymers with no hydrogen bonding interactions, those relavent to this

work, A8 can be no larger than 0.2-1.0 for miscibility.

3. Radical Polymerization

A free radical is an organic fragment possessing an odd or unshared electron.
Although several vinyl monomers such as styrene, and MMA are capable of self
initiation, these processes are slow and an initiator is usually needed to start the
polymerization process. Initiators can be photochemically, thermally, or redox activated
and come in a range of solubilities. Here we limit our discussion to photo and thermal
initiators. Some commonly used initiators in this category are peroxides (dialkyl, diacyl,
peresters, and hydroperoxides) and azo compounds. Free radicals are obtained by
thermal or photoinduced. homolysis of the weak CO or C-N bonds. Scheme 4 shows the
mechanism for radical formation of three common initiators: dibenzoyl peroxide, di-t-
butyl peroxide, and 2,2’~azo-bis-isobutronitrile (AIBN). Azobisnitriles are especially

attractive for kinetic work since decomposition rates are almost independent of the

solvent.”92

49

Three general processes comprise the radical polymerization mechanism;
initiation, propagation, and termination. There are two initiation steps, radical formation,

and the addition of the initiator radical to monomer. Propagation involves growth of a

hv or A
“W O —“*
di-benzoyl peroxide 10
CH CH
I 3 I 3 hv or A (EH3
H3C—Cl;—O—O—(|3—CH3 -———-—> 2 H3C—C—'Ou
I
CH3 CH3 CH3
di-t-butyl peroxide 1
0
II
CH /C\
2 - 3 + CH3/ CH3
(EH3 (EH3 hv or A (EH3
No—c—NEN—c—CN ————> 2 NC—?- + N2
CH3 CH3 CH3
AIBN

Scheme 4. Mechanism for radical formation of some common radical initiators

50

polymer chain by rapid sequential addition of monomer to the active chain end. Finally,
termination of a growing chain can occur by combination of two active centers,
disproportionation through abstraction of a hydrogen, or chain transfer. The first two
processes eliminate the active centers involved in polymerization while chain transfer
terminates the chain, but forms a new radical that can initiate a new chain. Inter- or intra-
molecular chain transfer is a common side reaction in bulk polymerizations where the
concentration of monomer is high. Fewer side reactions are observed in dilute solution
polymerizations which result in more ideal polymerization kinetics. Ideal kinetics
assumes that for every radical center, one monomer will react and that these reactions are

irreversible. Ideal radical polymerizations follow the rate expression (equation 18).

Rp = -d[M]/dt = kp [M'][M] ( 18)

Here Rp is the rate of polymerization, -d[M]/dt is the change in monomer concentration
[M] over time, [M'] is the total concentration of chain radicals, and kp is the
polymerization rate constant. It is assumed that the net rate of change of [M‘] is zero
during propagation and the reaction is said to be under “steady state conditions”. This
implies that during propagation, the rate of radical initiation is the same as termination,
hence [M’] is a constant during propagation and is only dependent on the initial rate of

initiation, R; and termination constant, k. of radicals. Hence...

[M] = (Rt/2h)” (19)

51

 

'6. o. c I: 4.. .1 . . g . . w .e . .. .. {r 34 ml 0.. .
i ; ,

The factor of 2 is used since two growing chains end in each termination event.

Bulk polymerizations are not only faster due to their concentrated nature, but are
non-ideal. Formation of high molecular weight polymer increases viscosity, therefore
decreasing the mobility of growing chains. In turn, this decreases the likelihood that two
polymer radicals can terminate by coupling or disproportionation. Monomer diffuses to
the active chain end to sustain propagation. Bulk polymerizations typically follow the

rate expression (equation 20)...

R..=k.tMl°‘tIl“3 (20)

where or at 1 and often [3 = V2.

For both the solution and bulk process, the rate of thermal initiation is related to
the initiator efficiency, f, the rate coefficient for initiator dissociation, a factor of 2 to

account for two radicals formed from each initiator, and initiator concentration, [I].

Ri=2fkdlll (21)

f is _<_ 1, since not all radicals successfully initiate chains. Note that kd follows Arrhenius
behavior kd = Ad exp(-Ea/RT). Therefore, increasing the temperature increases Ri.

Substitution gives the rate of thermally initiated polymerization.

R. = k. (fkt / kt)”2 iMl"‘tIl“2 <22)

52

CUE

The rate of photoinitiation is given by a similar expression, but molar absortivity
of the initiator, e, the quantum yield, (it, and the intensity of the incident light, 10, must be

considered (equation 23).
Ri = 24> 8 loll] (23)
giving

R. = k. (it) e I./ tom iMl"‘tIl“2 (24)

Hence it is typical that the rate of polymerization for both processes to have a square root

dependence on initiator, but differ orders in monomer concentration.

53

Chapter 2. Composite Polymer Electrolytes

Results and Discussion

This section focuses on the preparation and characterization of CPE’s
incorporating oligomeric PEO (PEGDME-SOO), LiClO4 salt (O/Li = 20), functionalized
fumed silicas, and methacrylate monomers. The methacrylates used were methyl
methacrylate (MMA), butyl methacrylate (BMA), and octyl methacrylate (OMA). Both
MMA and BMA are available commercially, while OMA was synthesized in-house.
These monomers vary in hydrophobicity due to the length of the alkyl side-chain, MMA
being the most hydrophilic, and OMA the most hydrophobic. The effect of monomer
properties and filler functionalization are examined as a function of polymerization rate,

conductivity, and the mechanical properties of the CPE’s.

The terminology used to describe the modified silicas is as follows: for fillers
synthesized in-house, FS is used to represent Fumed Silica and is followed by an
acronym representing the surface modification. These include FS-PMA (Propyl
MethAcrylate modified surface) and FS-TOM where TOM describes the starting
materials Trimethoxypropyl Methacrylate and Octyltrimethoxy. The exceptions to this
naming scheme are those fumed silicas used as received from Degussa, such as A200 (a
hydrophilic, silanol functionalized fumed silica) and R805 (a hydrophobic, n-octyl
modified fumed silica) (Figure 23). For FS-TOM silicas (see Scheme 3) the relative
amount of each group on the silica surface is denoted in parenthesis after the acronym.

The first number denotes the percent of propyl methacrylate (PMA), the second refers to

54

the percent coverage by n-octyl groups. For instance FS-TOM(4:1) (or FS-TOM(80:20))

would denote a filler surface with a surface coverage of 80% PMA and 20% n-octyl

groups.

A-200

R-805

FS-PMA

FS-TOM

3:02

3:02

Sio2

3'0 H

‘0 H

' O
ZOBSW

O
O
Sioz -O-S NOV
‘0
\
-\ O
\ o\\

Figure 23. Functionalized fumed silicas

55

hydrophilic, disperses
inPEO

hydrophobic, forms strong
networks in PEO

hydrophobic, forms strong
networks in PEO,
cross-linkable

hydrophobic, forms strong
networks in PEO,
cross-linkable

I. Hydrophobic - Crosslinkable Fumed Silicas

Silylation conditions for a maximum surface coverage of 50% were optimized
previously by Hou.84 Surface coverage was monitored by FTIR, looking for the
disappearance of the isolated silanol peak (3744 cm") and the appearance of the new C—H
absorption bands (2800—3000 cm’l), as well as C=C (1705 cm") and C=C (1640 cm!)
bands from the cross-linkable groups. Thermogravimetric analysis (T GA) and lithium
aluminum hydride titration techniques supplemented FTIR in some cases. With surface
coverage techniques firmly established, FS-TOM silicas were prepared as before,
monitoring the surface attachment of the n-octyl versus propyl methacrylate groups by
1H-NMR and TGA (thermogravimetric analysis). Known amounts of ocytltrimethoxy
silane (OTMS) and trimethoxysilyl propylmethacrylate (TMSPM) were weighed and
mixed to give the desired mole fraction of each prior to mixing with an A200 / toluene /
diethyl amine (DEA) slurry. After reaction, the relative and total degree of attachment of
each silane was determined from the mass and 1H-NMR spectra of the residue. As can be
seen in Figure 24, the methyl group adjacent to the double bond of TMSPM (1.95 ppm)
and the terminal methyl group on OTMS (0.89 ppm) can be easily distinguished and
integrated. After some simple calculations using the residual weight, the relative amount
of each group attached to the surface can be extrapolated from that not attached to the
surface. The weight loss expected for each FS-TOM was calculated on this basis and
compared to the actual weight lost by TGA. In TGA experiments it was assumed that all

methoxy groups convert to siloxane linkages during silylation and that upon heating, all

56

silicon-carbon bonds break and along with residual silanol groups, condense to give

siloxane linkages.
0

{CH3}

CH3 in OTMS

J

 

CH3 in TOM

34

 

 

 

Figure 24. 1H NMR of the trimethoxy silane residue extracted from FS-TOM
after surface functionalization is complete.

Hou showed that compared to OTMS, TMSPM preferentially attaches to the silica
surface. This behavior was reproduced and is seen quite clearly in Figure 25 for a 1:1
molar mixture of OTMS:TMSPM; 80% of the surface attached groups are from TMSPM
and only 20% are from OTMS. This observation can be explained by the structural
differences in the two silanes. A200 is hydrophilic and selective adsorption of the more

hydrophilic TMSPM silane is expected due to its ester functionality. Because this

57

preferential adsorption can be predicted, FS-TOM silicas have been prepared with
varying fractions of each group on the surface by use of the appropriate feed ratio of each

alkoxy silane.

 

 

 

 

100
I
OTMSPM l
a) 80 * IOTMS O
8 I
‘t
:3 I
‘0 60 .
.9.
'00, o I
'D
B 40 -
°\o o
‘5 o
E 20 I
o
o
o 1 l I l
0 20 4o 60 80 100

mol% used in rxn

Figure 25. The mol % of each monomer added to the surface of A200 based
on the initial feed ratio.

58

11. Composite Polymer Electrolytes

Once FS-TOM’s were prepared they and other fillers were mixed with a
methacrylate monomer, initiator, and electrolyte components to form thixotropic CPE’s.
The amount of filler required to form free standing gels is dependent on its surface
properties and will be discussed in more detail later. In kinetic and conductivity studies
the CPEs contain13 wt% of FS-TOM(4:1), 10 wt% methacrylate monomer, and 1 wt%

AIBN based on monomer, unless otherwise specified.

1. Cross-linked Composite Polymer Electrolytes

Our desire was to create dimensionally stable cross-linked CPE’s that were easily
processable by either a photochemical or thermal cross-linking process. Versatility in
this step is important for future processing applications. Photochemical cross-linking is
particularly desirable for samples that are to be further studied by impedance
spectroscopy. A cross-linking cell was developed that has minimal leakage under photo-
processing conditions and can be converted to an impedance measurement cell. The cell
and the geometry for photo-cross-linking are shown in Scheme 5. The uncross-linked
CPE was transferred to a cylindrical Teflon cell with a stainless steel end-cap. The cell
had a Teflon spacer (1.2 cm in diameter, 2 mm thick) that defined the volume of the cell.
A glass cover slip was carefully placed over the composite and the cell was exposed to a
medium pressure UV lamp under an N2 atmosphere. The UV lamp was water-cooled to
minimize IR heating of the sample. Once cross-linked, impedance analysis was

performed by carefully removing the glass cover slip and replacing it with another

59

stainless steel end-cap. Both end-caps then acted as electrodes for impedance

measurements, which are discussed later.

‘—

ili ill

Teflon

 

SS electrode conductivity cell cross-
section

Scheme 5. Schematic of the photo cross-linking set-up for UV curing
of composites

UV initiated polymerizations typically result in fast polymerization kinetics. This
is desirable for fast processing, however 200-300 nm light has also been shown to
photodegrade polymethacrylates.93 Homolytic B—cleavage at the carbonyl eventually
leads to generation of carbon monoxide, carbon dioxide, methyl, and methoxy radicals.
Pressure build-up associated with these gases is highly undesirable in sealed batteries.
However, it was of interest to determine if faster polymerization kinetics could be
obtained in our composites. Medium and high pressure mercury lamps generate broad
energy distributions with peaks at 254, 280, 297, 303, 313, 334, and 366 nm.94 Quartz

transmits light above 200 nm but pyrex filters out the lower wavelengths, transmitting

60

light only above 325 nm. Pyrex and quartz cover slips were compared to determine
whether the wavelength of the source had a significant effect on polymerization rate.
The CPEs contained 10 wt% BMA. The polymerization rate was monitored by
extracting each time-point sample with acetone to remove electrolyte and unreacted
BMA. The amount of BMA attached to the silica surface was then quantified by TGA.
The shapes of these curves are identical to those of Hou, who used FI‘ IR to monitor the
loss of the C=C acrylate bond and the shift of the C=O from 1705 cm"1 to 1724 cm'l.
Figure 26 demonstrates that the polymerization rate for our AIBN initiated system is
independent of the type of cover slip used. Since no significant rate increase was
observed using quartz, pyrex cover slips were used for all other UV photo-

polymerizations due to their convenience and their ability to minimize photodegration.

 

 

 

 

30
2 25 - I I I
E 20 '
8 8 O.
z m
E": o
3.3 l
031:) 10 .. l
E .0 oQuartz
o\°
g 5 .- IPyrex
I
o I I I I I
0 5 10 15 20 25 30
UV(min)

Figure 26. The effect of pyrex versus quartz cover slide on the weight % BMA
bound to the silica surface after UV curing.

61

Thermal cross-linking has the advantage of being applicable to a wide range of
sample geometries, and not needing optical access for curing. We thermally cured CPEs
using a variable temperature IR cell. Composites containing 10 wt% BMA were
sandwiched between two 32 mm diameter NaCl plates, with a 0.015 mm thick copper
spacer and put into the IR cell (Scheme 6). Once placed in the path of the IR beam, the
composite was heated to 80 °C. This temperature affords reasonable polymerization
times since the first order rate constant, kd, of AIBN is 1.5 x 104 s", which corresponds
to a half-life (rm) of approximately 60 minutes.95 Kinetic data were obtained by
monitoring the disappearance of the C=C double bond (1640 cm") during
polymerization. This disappearance and that of the C=C-H vinyl proton (3105 cm") as

well as the C=O shift are apparent in Figure 27 .

Copper spacer
with methacrylate
containing composite

 

Infrared
Detector .4“; ‘ Beam

 

 

Thermocouple
Power Supply

Scheme 6. Exploded view of the thermal curing set-up

62

relative absorbance

 

 

 

 

 

 

1.2 - .
C=O shiftto
higher frequency
1 - V
i
0.3 - g
0.6 r 5
0-4 ' _ "1 50min
0.2 F
0min A
O I I J I I

 

 

3000 2700 2400 2100 1800 1500
frequency (cm'1)

 

_ C=C-H at 3105 cm" I I C=C at 1640 0111’1 J

 

 

 

Figure 27. FTIR spectra of a composite containing 10 wt% BMA,

1% AIBN before polymerization (0 min) and after thermal polymerization
at 80 °C (50 min). The 50 min spectra has been shifted up 0.3 absorbance
units for clarity.

63

Figure 28 compares the photo and thermally initiated processes. Both sets of data
are reproducible and show that the cross-linking reaction is complete on the timescale of
minutes. It was expected and observed that the photo-initiated process is faster since the
rate of radical formation is faster in photochemical processes. The molecular weights of
BMA extracted from each composite after polymerization are consistent with a higher
rate of radical formation. The number average molecular weight, M“, from the photo-
initiated process is 33,400 with a polydispersity index (PDI) of 1.8 while for the thermal
process Mn=83,000 with a PDI of 2.5. Both PDIs are representative of a chain growth
process, while the fewer initiating radicals in the thermal process allows for longer chain
growth and higher molecular weight. Note that the thermal process involves a 2 minute
warm-up period where radicals are slowly initiated. This should have only a small effect

on polymerization time but may be responsible for a slightly higher PDI.

III. Composite Properties Based on Monomer Hydropobicity

1. Optical Properties

Because of differences in polarity, each methacrylate monomer should have a
different solubility in PEGDME-SOO. The difference is solubility parameter, A8, for each
monomer relative to PEGDME-SOO is as follows; A8 MMA = 0.8, A8 BMA = 1.2,
A5 OMA = 1.5. These numbers should increase as salt is added and the PEGDME phase

becomes more polar. Therefore, phase separation is expected for BMA and OMA.

Methacrylate containing composites show differences in optical clarity, which is believed

to be the result of phase separation. Upon polymerization, the solubility of the resulting

polymers is expected to decrease further due to a decrease in system entropy.

 

 

 

 

100 . . t
I l
I
. I
O
80 - l
O
I
S
.E 60 II .
a)
E u
o
0 40 .O OUV
58 lThermalC=C
I
20 P
I
0 l 4 I
0 20 40 60 80
Time (min)

Figure 28. Kinetics of cross-linking composite electrolytes using butyl
methacrylate as the monomer. The thermal polymerization was carried

out at 80 °C

Turbidity experiments were done to investigate phase separation. Phase
separation typically forms domains that are large compared to visible light. Thus, a probe

beam is effectively scattered in a phase separated sample. Measurement of the attenuation

65

of the beam gives the turbidity, a semi-quantitative measure of the degree of phase
separation. The probe beam was light from a 632 nm HeNe laser light source with a wide

area silicon photodiode as the detector (Scheme 7 ). The relative turbidity of uncured

/

thz ‘ Light

HeNe laser fl cell w. ——> transmitted

-—-—-\

composite electrolyte
before cross-linking

 

 

 

 

Scheme 7. Schematic diagram of the experimental setup for measuring
turbidity

samples of electrolyte were measured as a function of wt% monomer in the composites.
An optically clear sample containing no added monomer as a zero turbidity reference.
The data show distinct differences among the methacrylates, the results of which are
shown in Figure 29. At 10 wt%, the most hydrophobic monomer, OMA, results in a
slightly opaque composite with 92% transmittance. These composites become stronger
scatterers with added OMA, eventually reaching a minimum transmittance of 65% at 40
Wt% OMA. Based on solubility parameter considerations, the most hydrophilic

monomer, MMA, should be most soluble in the electrolyte phase. MMA gives optically

66

clear composites up to approximately 30 wt%, but show definite signs of phase
separation at 40 wt%. Composites incorporating BMA show intermediate behavior. 10
wt% BMA composites are optically clear, but become opaque at 20 wt% with dramatic

decreases in transmittance between 20-30 wt % BMA.

 

 

 

 

 

 

 

100

90 -
a:
0
C
m
E
E 80 _ + BMA
2 —I— MMA
g + OMA
o\°

70 -

60 41 l I I

 

0 10 20 30 40 50
Weight % monomer

Figure 29. Results of turbidity measurements on composites
containing MMA, BMA, and OMA

The turbidity phenomena can be explained by considering the FS-TOM(4:1)

fumed silica to be the site for nucleation of monomer droplets. In the case of OMA and

67

BMA, the hydrOphobic surface layer of the fumed silica matches the monomer solubility
parameter better than the electrolyte, and thus one can envision adsorption of monomer
onto the surface layer, and growth of the layer to form droplets of sufficient size to scatter
light. This effect can also be probed by studying conductivity data, thermal properties,
mass balance by TGA, and finally by following the kinetics of the curing reaction for

each monomer.

2. Conductive Properties

Conductivity data for composites with 10 wt% of each monomer before and after
polymerization are shown in Figures 30 & 31 respectively. Data for the PEGDME-SOO
(O/Li = 20) electrolyte without added filler or monomer is shown for comparison. Recall
that 13 wt% filler has little if any effect on conductivity. All electrolytes exhibit
temperature dependent Arhenius behavior, with the conductivity before cross—linking
dependent on the monomer structure. Adding MMA increases the conductivity relative
to the base electrolyte, BMA causes little change, and OMA induces a decrease in
conductivity. These data can be understood by considering the two limiting ways that the
monomer can affect the electrolyte (recall Figures 22A & 22C). First, the monomer
could simply act as a low molecular weight diluent and dissolve in the electrolyte phase.
The expected result would be a slight decrease in viscosity, which should be manifested
as an increase in conductivity. This effect is seen with MMA. If the added monomer is
insoluble in the electrolyte and adsorbs to the filler surface, then the effects should be
limited to simply decreasing the volume fraction of the conductive phase. OMA typifies

that behavior. These results are consistent with the turbidity data.

68

Conductivity (S/cm)

 

1 .0E-02

 

 

 

 

_'_ + MMA before UV
_ —O— BMA before UV
_ +OMA before UV
- --PEO-500
1.0E-03 ;
1.0E-04 ‘ ' l
2.75 2.95 3.15 3.35

1000rr (K)

Figure 30. Conductivity of composites made using 13 wt% TOM(411) and
10 wt% methacrylate before UV polymerization. Data for PEGDME-SOO
(O/Li=20) are shown for comparison.

69

 

 

 

 

1.0E-02 b
i
: + MMA after UV
—I— BMA after UV
i +OMA after UV
’5? .
Q
93,
.3
-2 1.0E-03 :
‘6' .
3 b
.0
c I
o I
O
1.0E-04 ' ‘ '
2.75 2.95 3.15 3.35
1000/T (K)

Figure 31. Conductivity of composites made using 13 wt% TOM(4: 1) and
10 wt% methacrylate after UV polymerization. Data for PEGDME-SOO
(O/Li=20) are shown for comparison.

70

After cross-linking, all three composites had the same conductivity. Note that the
conductivity of the composite made with OMA was unchanged from its previous value
before polymerization. Thus, both OMA monomer and polyOMA (pOMA) are
immiscible the electrolyte phase and it is likely that OMA polymerization occurs

predominantly on the filler surface. Upon polymerizing, MMA and BMA are expected to
become less soluble in the electrolyte phase and may adsorb more to the silica surface.
However, this does not imply that polymerization occurs at the surface and that the
polymer chains are covalently attached to the surface. The drop in conductivities seen for
the MMA and BMA containing composites may be due to a combination of an increase
in the viscosity of the electrolyte phase due to dissolved polymer and a decrease in the

volume fraction of the conductive phase as is seen with OMA.

The location of the monomer in our CPE’s can be further elucidated by tracking
the conductivity with increases in the amount of monomer added. The data show clear
differences in the conductive behavior of composites made with 10 wt% to 40 wt%
methacrylate. In the case of OMA, adding more monomer simply decreases the volume
fraction of electrolyte, and causes a corresponding decrease in conductivity Figure 32.
An idealized illustration of the CPE network before and after polymerization for the case

of coIrlplete phase separation is shown in Figure 33A & 33B respectively.
The data for MMA show markedly different behavior. As seen in Figure 34,

mcreasing the wt% of MMA before polymerization continuously increases the

conductivity, presumably due to a decrease in viscosity. Once polymerized, the

71

Conductivity (S/cm)

1 .0E-03

 

: +0MA before uv

_ —l—— OMA after UV
1.0E-04 :-

i
105-05 - . - - . - .

 

 

 

5 10 15 20 25 30 35 40 45

Weight % Methacrylate

Figure 32. Conductivity of cured and uncured CPE’s as a function of
OMA concentration.

72

A. 10 wt% OMA B. 40 wt% OMA

Before and after polymerization Before and after polymerization

 

 

 

 

Figure 33. An idealized view of OMA composites before and after polymerization;
A. 10 wt% OMA, B. 40 wt% OMA

Characteristic drop in conductivity is seen, but unlike the OMA composites, the
Conductivity is the same at all monomer concentrations. Poly(methyl methacrylate)
(pMMA) should be more miscible in the electrolyte phase than either pBMA or pOMA,
Since A5 = 0.8 for the MMA system, within the acceptable range for miscibility of non—
hydrogen bonded polymers. Miscible pMMA chains should be flexible and cause a
viSCOsity increase similar to that seen in plasticized pMMA electrolytes, with little to no
effect on conductivity. As a control, 0.025g of pMMA (Mn = 12,000) was dissolved in
0‘2 g of PEGDME-SOO electrolyte With heating. This simulates a composite made from
10 wt% MMA monomer after polymerization. The pMMA was fully miscible and no
clystallization or phase separation was observed after 1 week. In addition, the

Conductivity measured for this sample was 2.38 x 10‘4 S/cm, compared to 2.41 x 10“

SIC-m for the composite. Therefore, the conductivity drop for all concentrations can be

73

Conductivity (S/cm)

1 .0E-02

 

 

 

' +MMA before UV
—I— MMA after UV
#9
1015-03 : //
L \I‘ —I——/-
1.0E-04 :-
1 .OE-OS I I I I I I I

 

 

 

5 10

1 5 20 25 30 35

Weight % Methacrylate

40

45

Figure 34. Conductivity of cured and uncured CPE’s as a function of

MMA concentration.

74

explained by an increase in viscosity due to soluble pMMA, as opposed to a volume
effect as in the case of OMA containing composites. An idealized illustration of the
MMA composite before and after polymerization is shown in Figure 35. Figure 35A

represents a CPE with 10 wt% MMA. Before polymerization, MMA is evenly dispersed

throughout the electrolyte phase. After polymerization, some pMMA is covalently

attached or loosely adsorbed to the filler surface, while the remainder is dissolved in the

A. 10 wt% MMA B. 40 wt% MMA

Before polymerization Before polymerization

   

UV or thermal UV or thennal
curing curing
After polymerization

After polymerization

 

 

 

 

 

Figure 35. An idealized view of MMA containing composites before and after
Polymerization.

75

electrolyte phase. Figure 358 represents the 40 wt% MMA case. Turbidity
measurements indicate some phase separation at high wt% MMA. After polymerization,

more pMMA may be dissolved in the electrolyte phase, but not enough to significantly

hinder the mobility of ions.

Finally, consistent with the turbidity experiments, the conductivity data for BMA
composites show intermediate behavior (Figure 36). From 10-20 wt% BMA, the
conductivity mirrors that seen with MMA. At concentrations higher than 20 wt% BMA,
the conductivity drops, as was seen in the OMA composites. This indicates that BMA
and pBMA is partially miscible in the electrolyte, but that phase separation becomes

more pronounced at higher concentrations.

3- Thermal Properties

Differential Scanning Calorimetry is useful for studying the effects of monomer on
composite properties. It was shown previously that filler has little to no effect on the
therIlla] properties of CPE’s.84 Results for liquid electrolyte with and without filler (no
added monomer) are presented here for comparison. Table‘4A shows several trends in
T8, Tm, and Tc (crystallization temperature) data for composites made with 10 and 40
Wt% monomer. The corresponding DSC heating and cooling scans can be seen in

Figures 37 & 38 respectively. Thermal data for pure monomers are also shown in Table
5.

76

Conductivity (S/cm)

1 .0E-03

 

 

 

 

 

I ‘— \
I \
1.0E-04 :
I +berore UV \5
. +after UV
1'0E_05 I I I I I I I
5 10 15 20 25 30 35 40 45
Weight % Methacrylate

Figure 36. Conductivity of cured and uncured CPE’s as a function of
BMA concentration.

77

5038381835 be Us o3 £88 menace Ea messes Omn—

Eob 3:830 083 an. £2883an :oeefiwambo e5 .eh .mEBEoQEB mes—£2 .003 on :oufitofibom 3505 Bee
moon—beefing: axe; ow 53> 328980 «c 33.—25$ .m demtafioo 8m :32? as him 5223 new 53, 8383 out
8:208:68 8m 333m 588288 cam—E852: ace; ow can 55:» S mama: anemone—co mo 33.—38m .< .9 855a.

 

 

 

 

 

 

 

 

 

 

 

5.3. m9- 3.- 5.9.- 9% es 3 8-2:- TR- 2059.
EN- e5- 08. :5- Re 3 25: see: tee- Emma?
i L. 1 8. 22.5...

films g 6252.? $5 :4 Gasesese 35 :4 Gamay a5 5 Sarah 6.?— 295m

 

 

gaging—om “new 3:89:00 .m

$559. 5e 93. u :3 58 u. 86. u wees: 8 oh Nee:

 

 

 

 

 

 

 

 

 

 

 

 

3m- 3.x- SN- Nd- 25: sees an an- 8e: «Sosa.
25: 22. oem- 2- 55 we «:2 m- 35. 355.8
:8. NS. 5.2- 3:- on m- wem 99.- 88.55 sizes?
om- Na- 3m- 2 5.? es 3 5.9.- 68.55 <20s2
25: 25: 2m- 2- ea 2 at. 3 3m- Assamese
25: see: «.3. 3. 25a 25a 2:. 3. 2. sizes
25: see: new- on 25: 25: 2e .2 x. 22 SEE

SNuQQ
OGO: DGOG ~.O01 fl .N Oficfi UGO—.— m .00 ~.© hm: SWImU—Zn—OWHH
3.5 m4 Ghee m5 =< 6.1a m5 m< Cease as :4 6.2.; 6.? 23%

 

 

 

epsecega 2&3 garages see; 2... 5.; 85388 .<

78

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

PEGDME-500(O/Li=20) A
Composite - No Methacrylate A
10% MMA A,“
8 40% MMA /\w —_ 5
c: :
w i
40% BMA _ J4“—
400/0 OMA fl_ - -fi~+"‘-u—
Tm of PEG phase
-1 00 -80 -60 -40 -20 0 20
Temp (°C)

Figure 37. DSC heating scans of uncured composites containing 10 wt% and

40 wt% methacrylate monomers.

79

Endo —>

 

PEGDME-500(O/Li=20)

 

Composite - No Methacrylate

10% MMA
40% MMA

”Y

10% BMA

 

1M

------- L

40% BMA

i

10% OMA

V

 

40% OMA

W

I j l I I I

-100 -80 -60 -40 -20 0 20
Temp (°C)

L

 

 

 

Tc of PEG phase

 

 

 

 

Figure 38. DSC cooling scans of uncured composite containing 10 wt% and
40 wt% methacrylate monomers.

80

 

 

 

Monomer Molecular Weight-(g/rml) Melting Point °C Boiling Point °C
MMA 100.1 -48.0 100
BMA 142.2 -75.4* 163
OMA 198.3 -46.5 (-58, -63) 70 @ 0.4 torr

 

 

 

 

 

 

*at - 142°C, BMA undergoes a crystalline to vitreous transition

Table 5. Physical properties of methacrylate monomers

In the case of 10 wt% monomer composites, two distinct melting transitions were
observed using OMA, a TIn at — 49.9 °C resulting from a pure OMA phase, and a
transition at 6.1 °C, close to the Tm (8.4 °C) found in monomer-free composites.
Conversely, BMA and MMA composites show only one melting transition, confirming
that these monomers are miscible at 10 wt%. In addition, the T1; of MMA composites is
almost 20 °C lower than of that made with BMA. This is consistent with dissolved
MMA decreasing the viscosity of the composite, as was implied by the conductivity
results. Composites at 40 wt% monomer show distinct signs of phase separation with
multiple melting and crystallization transitions, even in the case of MMA. This was
anticipated from the high turbidity seen for all composites at 40 wt% monomer. The low
temperature melting transition seen in the heating scan of 40 wt% OMA is especially
Complicated. The DSC heating scan of pure OMA is equally complex, with multiple
melting and crystallizations transitions (Figure 39). A similar complexity in the DSC

Scan of the composite would be expected.

81

 

 

 

Endo —>

 

 

 

 

kheating—P
‘— cooling
-100 ~30 -60 -40 ‘ -20 0
Temp (°C)

Figure 39. DSC heating (top) and cooling (bottom) scans of
octyl methacrylate (OMA)

It is interesting to note that even at 40 wt%, DSC scans of BMA composites show
no Sign of phase separation. This is likely due to the lack of crystallinity in BMA, as it
has been shown to be an amorphous (glassy) solid in the region scanned. One clear trend
at 40 wt% monomer is that the more miscible the monomer, the more the thermal

t - o ' 6‘ 9’ ' ‘ ‘
Iiallsmons of the electrolyte devrate from that of the monomer-free composrte. This is

82

most apparent in the crystallization behavior of composites (cooling scans). The purer

the electrolyte phase, the less the monomer interferes with crystallization.

The thermal properties of composites with 40 wt% monomer after thermal
polymerization are reported in Table 4B and the DSC heating and cooling scans are
shown in Figure 40 & 41 respectively. An appreciable degree of phase separated
polymer should show a unique thermal signature for the homopolymer.92 For the pMMA
containing composite, there is a single Tg at —60 °C which is characteristic of the
electrolyte. This confirms that pMMA produced during cross-linking is fully soluble in
the electrolyte and as an added benefit, hinders crystallization of the electrolyte phase.

pBMA and pOMA phase separate, but unique Tg’s for these polymers were not observed
However, unlike the pMMA case, a Tm for the PEGDME electrolyte phase is easily seen.

An interesting qualitative observation is that after polymerization, the pBMA and pOMA

Composites appear to seep electrolyte while pMMA composites remain homogeneous in

nature. It is likely that as the polymer encapsulated composites cool from

Polymerization, volume changes occur in the network due to decreasing polymer

SOILIbility and densification.

A clear trend is observed in the Tg’s for the series, decreasing as the polymer

hydl‘Ophobicity increases. This is most likely due to a decrease in dissolved polymer.

However, the fact that these Tg’s are much lower (~60 °C pMMA, -69 °C pBMA, and —74
Q pOMA) than the Tg reported for composite with no added monomer (-54 °C) must be

addressed. The Tg’s approach the value reported by Hou (-71 °C) for samples that were

83

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

PEGDME-500 (LiClO4, O/Li = 20) A
f 40% MMA J/\ _ x
o _ _ A__
.0
LE
40% BMA __.,_._..,--._»-—~~..__...,_,___
--_, “h—._
40% OMA w _,._._~—\.\___
__ _____._ —“\._._.
I I I I I I

 

-1 00 -80 -60 -40 -20 0 20

Figure 40. DSC heating scans of composites containing 40 wt% methacrylates
before (thin line) and after (thick line) thermal curing at 80 °C.

84

 

PEGDME-500 (LiClO4, O/Li = 20)

 

 

12

 

 

 

 

 

 

 

T 40% MMA -
W V’—
O
"U 1
C
LU
40% BMA J
w
40% OMA

 

 

 

 

 

 

-100 -80 -60 -40 -20 0 20

Figure 41. DSC cooling scans of composites containing 40 wt% methacrylates
before (thin line) and after (thick line) thermal curing at 80 °C.

85

flash cooled from the melt. Flash cooling provides an amorphous solid as a starting
point, and Tg is expected to be lower than for semi-crystalline samples formed by slow
cooling. The similarity of these Tg’s to the flash cooled electrolyte is consistent with
dissolved polymethacrylate interfering with the crystallization process. In fact, compared

to the 40 wt% composites before polymerization, Tc for. the PEGDME phase is broader

and has a slightly smaller AH after polymerization

Hydrophobic monomers that do not dissolve in the electrolyte should tend to
adsorb to the hydrophobic filler surface. Specific information on the degree of phase
separation and therefore the polymerization that occurs on the FS-TOM(4:1) surface can
be probed by measuring the amount of polymer that is covalently attached after curing.
Since the silica portion of the filler is thermally stable to >900 °C, organic fragments
attached to the surface burned off during a TGA experiment and the loss in weight can be
measured. The results from these mass-balance experiments are shown in Figure 42.
'TGA curves for A200 and FS-TOM(4:1) result in relative weight losses of approximately
0.5% and 4.5% respectively. The weight loss for A200 is from water absorbed on the
Surface and from condensation of silanols. The mass lost from FS-TOM(4: 1) is from the
organic groups attached to the surface, plus a small contribution from residual surface
hydroxyls. Significant weight losses are seen for cured composites, confirming the
polymer is bound to the surface during cross—linking. Of the three composites tested, the
Composite with 10 wt% OMA lost the most weight upon heating. It follows that OMA
f0I‘ms the most polymer at the silica surface due to its high degree of phase separation

and adsorbtion compared to BMA and MMA. Composites made with the most

86

°/e weight loss

 

 

 

 

 

 

 

 

 

1 00
A200
FS-TOM
90
80
70
MMA
60 PMA
OMA
50 I I I I I I

 

100 200 300 400 500 600 700 800

Temperature (°C)

Figure 42. TGA of composite filler materials after cross-linking with
10 wt% methacrylates and solvent extraction. Data also are shown for
A200 and FS-TOM(4:1).

87

hydrophilic and soluble monomer, MMA, had the least weight loss and the least amount

of polymer covalently attached to silica.

4. Mechanical Properties

In terms of mechanical properties, dynamic rheological measurements done at
room temperature before cross-linking show that composites containing 10 wt% OMA
have slightly better properties than composites using BMA, MMA, or no monomer.
Figures 43 & 44 give the results for composites prepared from each of the three
monomers and 15 wt% FS-TOM(4:1). Two plots are shown for clarity. The OMA
composite (Figure 43) has the highest modulus with G’ dominating over the entire
frequency sweep. Frequency independent (solid-like) behavior is maintained from .01 s'1
to 1 s". It is presumed that phase separation of the OMA strengthens the as is it deposits
on the silica network. The non-zero slope above 1 5'1 indicates flow and therefore
disruption of the network structure. In the absence of monomer, this deviation occurs at
lower frequencies and G” begins to dominate (crossover point), indicating poorer

mechanical properties than the OMA containing composites

Results for the MMA and BMA containing composites are shown in Figure 44. In
addition to non-zero slopes for G’ and G” curves, G” crosses over G’ indicating liquid-like
behavior. G” dominates the entire frequency sweep for the MMA containing composite
and therefore the MMA containing composite is the most liquid-like. Note that the

MMA composite should have the lowest modulus, however, due to the volatile nature of

88

G’, G” (Pa)

 

1 .0E+03

' I G’-OMA
- 1:1 G”-OMA
-----G'-15°/. FS-TOM(80:20)
-- - -G”-15%FS-TOM80:20)

 

 

 

 

 

1.0E+02 ‘ '
0.01 0.1 1 10

Figure 43. Dynamic frequency sweep of composites containing 10 wt%
OMA and 15 wt% FS-TOM(80:20) filler. G’ and G” for a composite with
no added monomer are shown as a solid and dashed line respectively.

89

 

 

 

 

 

 

1.0E+03
. A G’-BMA
A Gn'BMA O <>
- o G’-MMA 0 A
o G”-MMA A
- “021% FS-TOM(80:20) 8
__ - —G”-15%FS-TOM(80:20) 8
’3 r
9; /
ED .
i9
1.0E+02 ' ‘
0.01 0.1 1 ‘0
freq (3")

Figure 44. Dynamic frequency sweep of a composites containing 10 wt%
MMA and BMA using 15 wt% FS-TOM(80:20) filler. G’ and G” for a
composite with no added monomer are shown as a solid and dashed line
respectively

90

MMA it is difficult to obtain an accurate measurement of G’. Therefore, the MMA result

is shown for qualitative purposes.

5. Kinetics of Polymerization

Phase separation of the monomer and electrolyte has a significant effect on optical,
conductive, thermal, and mechanical properties of our CPE’s. Signatures of phase
separation can also be seen in the kinetics of the curing reaction. The samples were cured
at 80 °C and the progress in the curing reaction was monitored by IR spectroscopy. As
discussed previously, an ideal free radical polymerization typically follows first order
kinetics and can be described by Rp = d[M] /dt = kp[R*][M], where Rp is the rate of
polymerization, d[M]/dt is the change in monomer concentration over time, kp is the rate
constant for propagation, [R*] is the concentration of active growing chain ends (a
constant at steady state), and [M] is the concentration of monomer. Recall that bulk
polymerizations are often non-ideal due to their concentrated nature and Rp = p[M]°‘[I]B
where or at 1 and often [3 = V2.96 The kinetic data should reflect the solubility of the
monomer in the electroyte. Insolubility should yield bulk-like behavior (high [M]), while
soluble monomers should follow solution polymerization kinetics (low [M]). The kinetic
data for composites made with 10 wt% MMA, BMA, and OMA indicate a difference in
RI) (Figure 45). For MMA, the monomer is soluble and therefore the effective monomer
concentration is relatively low; hence the slower polymerization kinetics and different
shaped curve. The MMA polymerization follows first order kinetics at 10 wt%

monomer. Composites made with BMA and OMA have kinetic curves that resemble

91

 

 

 

 

100
I ' o l
I
I ‘ o
80 b A ..
I A"
. v
c d
E, 60 - X
g? I
S
0 40 - "
o\°
I
" eMMA
20 - IBMA
I AOMA
0. I I I I
0 20 40 60 80 100
Time(min)

Figure 45. Kinetic curves for thermal cross-linking of composites containing
10 wt% MMA, BMA, and OMA. The cross-linking was carried out at 80 °C.

92

bulk-like polymerizations where the local concentration of monomer is high and the data
do not fit first order kinetics. Typical limiting monomer conversions in bulk
polymerizations of methacrylates range from 80-95%.96 This is consistent with

conversions seen for these polymerizations.

Polymerization of MMA-containing electrolytes were run with and without filler.
As shown in Figure 46, the kinetic data for the two experiments are different, and thus
the fumed silica must play a role in the curing reaction. The simplest explanation that is
consistent with the turbidity and TGA data is that the monomer concentrates at the silica
surface. In the absence of the hydrophobic filler, the effective concentration of monomer
is low because the monomer completely disperses in the electrolyte. With added filler,
the monomer absorbs to the filler surface, increasing the effective monomer
concentration at the surface and the rate of polymerization. Note that BMA and OMA
phase separate from the electrolyte without filler present, therefore, this experiment can

only be done using MMA.

The lithium salt also plays a role in the distribution of monomer. Shown in Figures
4749 are the results for three sets of polymerizations, one for each monomer with and
without salt. The addition of salt increases the polarity of the PEGDME phase and
therefore helps drive phase separation of hydrophobic monomers. In the two extreme
cases of the hydrophilic MMA and hydrophobic OMA, no significant change in rate is
seen. However, for the intermediate case of BMA, added salt increases the

polymerization rate so that the shape of the kinetic data matches that of the fully phase

93

 

100

 

 

 

O
o ' A
O
80 " ’ ‘
. A
O
O A
c O
o _ AA
0)
0
CE) A‘
0 40 -° ‘
O
o\ A
. A A Average - No Filler
20 ,A OAverage w/Fillef'
OIIILIIIIIIJLIiJIIJII
O 50 100 150 200
Time (min)

Figure 46. Thermal cross-linking of 10 wt% MMA composites done at
80°C with and without filler

94

 

100

 

 

 

80 - o 0
.9
' ’0’ <>
0
f; 60 - as <><><>
i2 00
8 40 - 0°
o\° ’ o 10%MMA(salt)
<> 0 10%MMA-no salt
20 r.
0
O k l J I I
0 20 40 60 80 100

Time (min)

Figure 47. Thermal cross-linking of 10 wt% MMA composites at 80
°C

95

% Conversion

 

 

 

 

100
II E]
I
80 - l' D
[:1
E]
El
60 -
. a?
[:3
El
40 - Cl I10%BMA(salt)
E El 10%BMA-no salt
20 PB
E
O I l l L
0 50 100 150
Time(min)

200

Figure 48. Thermal cross-linking of 10 wt% BMA composites at 80 °C

with and without LiCl04 (O/Li = 20).

96

 

100 A

 

 

 

A A
a A A
80 AAA
I “A
A AA
60 - A
g A
E .
d}
g 40 - AA
8 A A 10°/oOMA(salt)
.\° 2:3 A 10%OMA-no salt
20 g)
A
0 k ' '
0 50 100 150
Time (min)

Figure 49. Thermal cross-linking of 10 wt% OMA composites at 80 °C
with and without LiClO4 (O/Li = 20).

97

separated composite using OMA. Without salt, the kinetics resemble data for composites

containing the miscible MMA

The effect of monomer concentration and the concentration of AIBN were also
investigated for the case of monomers that phase separate (BMA, Figures 50 & 51, and
OMA, Figures 52 & 53), the polymerization kinetics at all concentrations behave as a
suspension (bulk-like) polymerization. Rp doubles when the initiator concentration is

increased four-fold (i.e...Rp 0t [1]”2), and the monomer concentration, Rp decreases by

half when the BMA monomer concentration is quadrupled (Rp 0t [Mil/2). At higher
monomer concentrations (40 wt %), the kinetic data for both cases show an increase in
slope at conversions over 30% (see the difference in slope for OMA in Figure 53). This
effect, known as autoacceleration, is due to the characteristic increase in the viscosity of
the bulk system. The polymerization rate accelerates as the bimolecular termination
reactions are inhibited by reduced diffusion of growing chain ends. Monomer is smaller

and can diffuse faster, hence there is no change in propagation rate, but an increase in RP.

We have seen in previous results that MMA containing composites do not phase
separate to the same degree as those prepared from BMA and OMA. As a result, the
kinetic curves for MA polymerization indicate solution-type polymerizations where a
solvent is added to monomer as a diluent. Here the PEGDME phase acts as the solvent.
A clear shift from solution-type behavior to a bulk-like polymerization is seen in Figure

54 as MMA concentration and phase separation increases. Note that Rp increases with

98

100

 

 

 

 

IIIAA
O ”f A
O A
80 b . I A
O
I
c . l A
A
O
3 ’ I A
c:
8 I ‘
40 - I A O10°/oBMA 1%AIBN
o\° A
‘ I 40%BMA 1°/oA|BN
9 I
A 40%BMA .25%AIBN
50 100 150 200

 

Time (min)

Figure 50. Kinetic curves for the curing of 10 and 40 wt% BMA containing
composites at 80 °C. The effect of initiator concentration was studied using
1% to 0.25 % AIBN in 40 wt% BMA composites

99

100

 

 

 

 

 

 

 

 

 

 

 

 

  

 

 

 

y = 1.79x - 7.12
80 ' y=3.86x-7.18 Y=1-00X+2-73||
A
S 60 -
1‘72
0)
>
C
O
0 40 b
e\°
O10°/eBMA 1%AIBN
2° ' I 40%BMA 1%AIBN
A 40%BMA .25%AIBN
O I I I
0 20 40 60 80

Time (min)

Figure 51. The linear portion of the kinetic curves for the cross—linking of BMA
containing composites at 80 °C. The slope of each curve, calculated from the least
squares fit to the data, is shown next to each data set.

100

%Conversion

 

 

 

 

100
o o 9
o
I o
80 - I 9
I
I o
I o
60 .
o
. O
40 _ I . 940/.
I10%
0
e.
20 - .
I 9.
o
o
O I I I I I
0 20 40 60 80 100
Time (min)

Figure 52. Kinetic curves for the cross-linking of 10 and 40 wt% OMA
containing composites at 80 °C.

101

 

100

 
   

 

 

 

 

 

 

 

, a . Jv"= 5.719;"526-01
80 J y = 2.46x 4. 2.58-
c
.9 6O -
Q
0)
>
c
8
°\° 40 -
y =1.65x - 12.25
20 -
I 10%0MA
O 40%OMA (low conversion)
A 40%OMA (high conversion)
0 I l I l
O 20 40 60 80 100
Time (min)

Figure 53. The linear portion of the kinetic curves for the cross-linking of OMA
containing composites at 80 °C. The slope of each curve is calculated from a least
squares fit to the data and shown next to each data set.

102

100

 

 

 

 

I 9’ ‘
c w
.9 0
9 60 - -
<1)
>
5 5 A10%
0 40
\o '9 920%
O
I I40%
20 -
0 I I I l
0 50 100 150 200
Time (min)

Figure 54. Kinetic curves for the polymerization of MMA containing
composites at 80 °C.

103

monomer concentration due to the increase in local monomer concentration. The higher
effective monomer concentration is also seen in the resulting molecular weight of

(pMMA) extracted from each composite after polymerization. Values of Mn and PDI for
these MMA-based composites at 80 °C are: 10 wt%, Mn = 51,900 (PDI=1.7), 20 wt%,

Mn = 56,000 (PDI=1.8), 40 wt%, Mn = 91,700 (PDI=1.9).

The composites, particularly the phase-separated cases, are thought to cure via a
suspension type polymerization. The bulk characteristics of these polymerizations are
clear from their kinetic profiles and the behavior seen when monomer concentration is
varied. A potential competing polymerization mechanism is emulsion polymerization
where the initiator resides in the solvent (usually water in the classical examples) and the
monomer is phase separated into small micelles and larger monomer droplets. The first
step in initiation occurs when initiator reacts with small amounts of monomer in the
solvent phase to create activated oligomers. Polymerization then proceeds by diffusion
of initiated oligomers in and out of micelles. Over time, monomer from droplets diffuses
into the polymerizing micelles until beads of polymer are eventually formed. These

polymerizations are fast and normally result in polymers of very high molecular weight.

To determine if our polymerizations were emulsion or suspension-like in nature, it
was necessary to determine the location of the “active” initiator in our composites. The
solubility of AIBN in PEGDME-500, PEGDME-500 (O/Li=20), MMA, BMA, and OMA
was measured by dissolving the maximum amount of AIBN in each solvent at room

temperature. The 1H NMR shift of the methyl group on AIBN is 51.69, which is easily

104

 

resolvable from the methyl groups at the end of the PEGDME-500 chain, and from the
protons on the double bond of the methacrylate groups at 53.35 and 51.91 respectively.
Comparisons of the the NMR integration data yielded the solubility limit in terms of mg
of AIBN per gram solvent. These results are shown in Table 6. AIBN has the highest

solubility in the electrolyte phase, 7 times more soluble than in OMA, and 3.5 times more

 

 

 

soluble than in MMA.
Solvent AIBN Solubility
(mg/iSolvent)
PEGDME— 500 (O/Li=20) 30.0
PEGDME-500 27.0
MMA 8.5
BMA 5.1
OMA 4.5

 

 

 

 

Table 6. Solubility of AIBN in different solvents

If the majority of AIBN decomposes in the electrolyte phase, it may be possible
that activated monomer diffuses from the electrolyte phase to the phase separated
monomer to initiate polymerization. If so, the kinetics of polymerization would follow
those of an emulsion rather than suspension polymerization. Our current kinetic data
indicate a bulk-like, suspension polymerization where “active” initiator is contained in
the monomer. Further clarification of the mechanism was obtained by varying the
temperature of the polymerization. In suspension polymerizations the molecular weight
decreases with increasing temperature, while the molecular weights of emulsion

polymerizations typically increase with temperature due to increased monomer diffusion.

105

Table 7 gives the results for 10 and 40 wt% MMA, BMA, and OMA composites. In all
cases a decrease in molecular weight is observed, consistent with an increase in the
concentration of AIBN radicals in the vicinity of the monomer. Therefore, even though
the solubility of AIBN is higher in the electrolyte phase, the majority of “active” AIBN is

predissolved in the monomer.

 

 

 

 

 

 

 

Monomer: MMA BMA OMA
Wt% Temp °C Mn PDI Mn PDI Mn PDI
10% 80°C 51,900 1.7 83,000 2.5
85°C 26,600 2.0 63,700 1.9 40,500 2.1
95°C 18,900 1.8 29,300 1.5 18,600 1.9
105°C 17,900 1.7 18,000 1.5 18,300 1.6
40% 30°C 91,700 1.9 75,400 3.8
85°C 52,500 2.3 40,500 2.8 51,600 3.9
95°C 22,600 2.2 14,700 2.4 12,800 4.0
105°C 21,900 1.8 19,800 2.0 4,600 3.8

 

 

 

 

 

 

 

Table 7. Molecular weights of THF soluble polymer extracted from thermally
polymerized composites. All polymerizations were done in the presence of
PEGDME-500 (O/Li=20), 13 wt% FS-TOM(4:1), and a 1 wt% AIBN, based on

monomer.

The high solubility of AIBN in the electrolyte phase is a point of concern
however. The formation of a high concentration of radicals in the polyether medium
could result in the breakdown of the oligomeric PEGDME chains, especially in the
presence of 02. This is unlikely due to the relatively low stability of the AIBN C-N=N
bonds in comparison to the OH, CC, or C-O bonds in a polyether. As a control, 1 mg of
AIBN was dissolved in 0.77 grams of PEGDME-500 electrolyte (the amount in a typical

1 gram composite) and the mixture was heated in the presence and absence of oxygen.

106

The FI‘IR data in Figure 55 show no evidence for an AIBN / polyether reaction when the
reaction is carried out in a helium filled drybox. However, when run in an open-air
environment, FI‘IR shows development of a small carbonyl peak at 1721 cm". These
results highlight the importance of maintaining an oxygen free environment when

working with these composite polymer electrolytes.

 

 

 

 
 
    

 

 

 

 

 

 

 

 

  
 
   

LiCIo4
’2 PEGDME 500 -1
a) - 1721 cm (C=C)
C + LiClO4, AIBN
CD .
E '" 02 W
m _
O
E, PEGDME-500
e + LiCIO4, AIBN
8
.0
(U
PEGDME-500

+ LiCIO4

 

 

PEGDME-500

 

 

l I l l l I

3600 3100 2600 2100 1600 1100 600

frequency (cm'1)

Figure 55. FI‘IR spectra of PEGDME-500 samples showing the effect of
AIBN and 02 on the chemical stability of PEGDME-500. Samples were
heated at 100 °C overnight.

107

IV. Composite Properties based on Filler Surface Characteristics

1. Mechanical Properties
Different types of silica surface functionalization change the mechanical properties

87 Dynamic rheology was used to elucidate the mechanical properties

of the composite.
of composites, this time varying the silica surface and wt% filler. To validate the
instrumental technique, a composite with 5 wt% R805 (n-octyl surface) was prepared and
the dynamic behavior measured on a Rheometric Mechanical Spectrometer (RMS-800).

Figure 56 shows the same frequency independent behavior and high elastic modulus over

the entire sweep range that had been previously reported.87

 

1.0E+04

O G’
o G”

U'I'Ur

. +

i

G’, G” (Pa)

N'U'I

A

ow

,00 0000
0000000000000

1.

 

 

 

1.0E+02 u .
0.01 0.1 1 10

freq (3")

Figure 56. Dynamic frequency sweep of a composite containing 5 wt%
R805.

108

Figure 57 illustrates the changes in rheological properties of 15 wt% filled
composites as the fraction of n-octyl chains on the filler surface increases and PMA
decreases. We would expect that the more polar PMA chains would have increased
interaction with the solvent and therefore decreased flocculation and lead to weaker
network formation. Indeed this trend is quite apparent. The data show that for
composites more hydrophilic than R805 (FS-TOM(80:20) and FS-TOM(67:33)), G’
dominates the composite behavior at low frequencies (long relaxation times), and there is
a crossover of G” to more liquid-like behavior at higher frequencies. In the FS-
TOM(45:55) system where the surface coverage is >50% octyl chains, the highest
modulus is observed and there is no crossover of G”. In other words, the dynamic
frequency behavior begins to parallel that of R805. These results point to strong van der
Waals attractive forces in this most hydrophobic FS-TOM that leads to a more elastic

network.

The amount of filler needed to form an elastic network is called the critical overlap
volume, (pv, and is a measure of interaction between the surface groups on filler particles.
Increased attractive forces should result in a decrease in (pv. Low filler loads are desirable
for high energy densities as the amount of active materials in a battery can be maximized.
Composites made using 15 wt% FS-TOM(45:55) have a modulus of 104 Pa. When the
filler is decreased to 10 wt%, G’ decreases to 2 x 102 Pa but the composite remains
mechanically stable with G’ > G” and there is only a slight increase in the slope is

observed at the end of the sweep (Figure 58). It is likely that the critical overlap volume

109

G’, G” (Pa)

 

1 .0E+05

U'I'U"

 

 

 

 

 

 

 

1 .0E+04
1 .0E+03
1.0E+02 e '
0.01 0.1 1 1o
freq (3")

Figure 57. Dynamic rheology of composites with 15 wt% FS-TOM fillers. The
degree of n-octyl substitution on the filler surface is varied. G’ = closed symbols,
G” = open symbols

110

 

1 .0E+05

 

         

G’, G” (Pa)

 

 

 

;
1.0E+04 AAA" —
; AAAAAfiAAAA
1.05103 1;
C .5! E a
_!.
b I ~ 5 E E
: III-IIISDEB'
BBC] {:1 BEDS D
11.0E+02 * ‘ ‘ ‘ "'l“ I l I 111111 1 n 1;....
0.01 0.1 1 1o

freq (3")

Figure 58. Dynamic rheology of composites with 10' and 15 wt% FS-TOM(45:55)
filler. G’ = closed symbols, G” = open symbols

lll

 

1 .0E+O4

 
 
 
  
   

—I— G’-15%FS(80:20)
-a— G”-15%FS(80:20)
+ G’-15%FS(66:33)
+ G”-15%FS(66:33)
+ G’-20%FS(80:20)

  

 

 

 

 

_ —e— G”-20%FS(80:20)
Tc?
3:
ED 1.0E+03 t
E :
1-
F
1.0E+02 n l 111.11; I n lllllll J n llllll
0.01 0.1 1 10

freq (5“)

Figure 59. Dynamic rheology of composites with 15 and 20 wt% FS-TOM(80:20)
fillers. Data for 15 wt% FS-TOM(66:33) is shown for comparison.

112

for FS-TOM(45:55) is close to 10 wt%. A lower concentration should result in the same
behavior seen in FS-TOM(80:20) and FS-TOM(66:33) composites, which require >10
wt% to give non-flowing gels upon standing. Dynamic frequency sweeps show that at
higher loadings of 20 wt% FS-TOM(80:20) composites still have significant liquid-like
characteristics with G” dominating the entire sweep (Figure 59). The modulus is
comparable to composites using 15 wt% FS-TOM(66:33), which is more hydrophobic
and shows an increased G’ presence. The best mechanical properties and lowest filler

loadings (5 wt%) are obtained using R805.

It should be mentioned that although no rheological experiments were on cross-
linked composites, these silica systems have previously shown an increase of at least 4

decades in elastic modulus after polymerization.97

2. Conductivity

While the mechanical behavior of these composites shows a major dependence on
the number of surface octyl chains, the conductivity is completely independent.
Composites were prepared using fillers such as R805, FS-TOM’s of varying surface
ratios, and FS-PMA (propyl methacrylate functionalized A200). The data show that the
conductivity is approximately the same for all composites studied before and after
polymerization with BMA (Figure 60). Recall that the independence of conductive
properties from mechanical properties has also been shown when varying weight percent

filler in an R805 based composite (Figure 20).

113

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

1.0E-03
' 9 before UV curing
. I after UV curing
- A no BMA
O O
A ’ t ”
E ° .
£9
59,
3‘ M A
E ‘ ' ‘
g I
2 . * ' L“
o l
0 L /
100% n-octyl 100% pMA
1.0E-O4 v ' ' '
0 20 4o 60 80 1 00

percent crosslinkable groups

on silica surface

Figure 60. Effect of FS-TOM filler surface on the conductivity of composites
With BMA before and after UV curing and without added monomer.

114

The attachment of BMA to each type of filler surface was studied by TGA after
polymerized composites were extracted. The consistency in the conductivity data using
different FS-TOM’s indicate that phase separation onto the filler surface should be
independent of filler hydrophobicity. The TGA data in Figure 61 agree and show that
the surface attachment of BMA is independent of how many cross-linkable groups are
present. In the case of fully octyl functionalized silica no significant polymer attachment

is expected or observed due to the lack of a surface anchor.

 

 

 

 

 

 

 

 

 

 

 

 

100
80 _ ° 0 7
o o
.. o
'0 ° /
S o 100%PMA
0
e 8 60 ' .
a)
E g
0 3
g 0)
E .3 40 ~
“5 '23
C
.9
0
g 20 -
100%n- l
K‘ Ody
O I I I L
o 20 4o 60 80 100

percent crosslinkable groups on
Silica Surface

Figure 61. Effect of FS-TOM filler surface composition on the fraction
of BMA that is bound to the surface during UV curing

115

V. Conclusions

Composite polymer electrolytes made from functionalized fumed silica fillers are
excellent candidates for use in ambient temperature lithium batteries. They are simple to
prepare and their thixotropic nature allows for ease in processing. Tuning the mechanical
properties is simply a matter of varying the hydrophobicity of the filler. Fillers are easy
to prepare and can be made cross-linkable. In addition, since the uncross-linked
composite is a physical gel, it can be prepared ahead of time and will conform to a variety
of shapes that may be realized in future battery systems. These shapes can be made
permanent by addition of a monomer/initiator and effected through photo or thermal

cross-linking.

Monomer hydrophobicity and its incompatibility with the electrolyte has a
significant effect on the optical properties and polymerization behavior of the composite.
Small effects are seen for conductivity and mechanical properties. Regardless of physical
properties, all monomers polymerize to the expected 80-95% conversion. However,
strong phase separation for more hydrophobic monomers such as OMA may help prevent
later reaction of uncross-linked monomer at electrode surfaces due to trapping in the
polymer phase (recall that an increase in resistance at the lithium anode was observed
when using 10 wt% BMA). Regardless of the electrochemical benefits, physical data
from DSC experiments and qualitative observations indicate that strongly phase separated

monomers lead to inhomogeneity in cross-linked samples due to macroscopic phase

116

separation of the electrolyte. This could eventually lead to failure of a battery cell

through loss of electrolyte by leaking.

Conversely, the data for composites containing MMA show little to no phase
separation up to 20 wt% MMA. Even at 40 wt%, polymerized composites showed good
mechanical and conductive properties with no visible leakage of the low molecular
weight electrolyte. It is concluded that MMA’s solubility in the electrolyte phase makes
it more compatible for use as a polymerized composite. After cross-linking, significant
amounts of pMMA are covalently attached to the filler surface, which increases the
mechanical strength. However, the problem of incomplete polymerization may be worse
for MMA because of its high misciblity. Compared to OMA, it is more likely that free
MMA and BMA will be found in the electrolyte phase. MMA is the most volatile of the
three monomers studied. This could lead to pressurization in cells containing
unpolymerized monomer and may complicate large-scale thermal polymerizations using

MMA.

A possible solution to these problems may be the use of a difunctional monomer
of compatible polarity to the electrolyte and having a low volatility. Difunctionality
would help ensure at least one monomer unit is part of the cross-linking reaction and
decrease the chance that free monomer reacts with the electrodes. A higher molecular
weight would decrease volatility and pressure build up, and a polarity similar to the
electrolyte would afford a homogeneous, seep-free electrolyte. Such a monomer could be

easily prepared in (Scheme 8) from an oligomeric PEO diol or a short alkane diol such as

117

1,4-butanediol. Reaction with methacryloyl chloride would give the dimethacrylate. The
use of a cross-linkable filler would still afford thixotropic composite materials. Recall
that these systems are highly desirable during room temperature processing and can be
made more durable after cross-linking. Ease of processing coupled with a high modulus

and the possibility of a variety of sizes and shapes is not possible for network and gel

electrolytes.
HOCHZCHZCHZC H20H

OR

H0<CHZCH20—}nH

n =4-6

H20 Cl
CH3

 

O

V
0

CH3 H30
OR
0 O

HzcngC’iC'lzCHZWCHz

CH3 H30

Scheme 8. Preparation of difunctional monomers for future polymerizations

118

The nature and the amount of filler has been shown to have little effect on the
conductivity of our composites whether polymerized or not. However significant effects
on mechanical properties have been shown. Customizing the surface of fillers is a
potentially costly step, and the need for a customized cross-linkable surface should be
examined in future studies. R805 may be used alone if the covalent surface attachment of
monomer is not a factor in reinforcing network properties. TGA results have shown that
hydrophilic monomers such as MMA significantly adsorb onto hydrophobic filler
surfaces. Therefore covalent attachment may not be necessary to increase the strength of
the filler network. If R805 is determined to be unsuitable alone, a small amount of a
commercial cross—linkable filler in combination with R805 may give acceptable results.
Recall that BMA attachment was independent of the percentage of cross-linkable groups

on the FS-TOM filler surface.

119

Chapter 3. Limiting Crystallinity in PEO

Results and Discussion

We have shown that composites based on PEGDME-500, methacrylate
monomers, and fumed silica fillers are excellent candidates for ambient temperature
secondary lithium batteries. The liquid nature of the electrolyte allows for conductivities
in the appropriate range (>104 S/cm) for ambient temperature use. At the same time,
mechanical stability is imparted to the system through hydrophobic and cross-linkable
fillers that form elastic networks with little to no degradation of conductivity. A
challenge for battery technology is the reduction of crystallinity in PEG-based systems,
both at ambient and low temperatures (T = -40 °C). Batteries for low temperature use are
needed in cold climates and for space applications. Oligomeric PEO’s, binary and
ternary mixtures of carbonates and other small molecule electrolytes, and branched PEO
polymers designed to have limited crystallinity are being evaluated for low temperature
applications. Unfortunately, branched as well as linear oligomeric PEO’s crystallize at
low temperatures. Hence, the PEGDME-500 electrolyte currently used for composites
may be unsuitable for low temperature use. Conductivities as high as 10'3 S/cm at —20°C
were reported for ternary mixtures of DMC, EC, and PC which form a low viscosity

electrolyte with a depressed tendency for crystallization.98

In this chapter, the synthesis and characteristics of a new branched oligomeric
PEO called star(12)PEO is described. The star architecture is designed to reduce

crystallization and improve the low temperature properties of electrolytes. Plasticization

120

of high molecular weight PEO systems by both star(12)PEO and PEGDME-500 was
investigated to gain a better understanding of how the shape of oligomeric plasticizers
affect PEO crystallinity. Similarly, fillers ranging in hydrophobicity were incorporated
into a solid PEG and tested for their ability to inhibit crystallization. Finally, the
conductivity and physical properties of a series of (AB)n polymer electrolytes
incorporating short PEG and poly(ethylene) segments have been studied. The effect of

the E0 segment length on crystallization and conductive properties will be presented.

1. Preparation and Characterization of Star(12)PEO Electrolytes

Branching PEO has been an effective strategy for disrupting crystallinity and
increasing the conductivity of PEO based electrolytes. However, the synthetic pathways
to these high molecular weight PEO’s can be complicated and costly. We developed a
simple oligomeric version of a branched PEO in hopes of improving the low temperature
properties of PEO electrolytes. This star-shaped analog of PEGDME-500 is called
star(12)PEO, where the number 12 represents the number of oxygen atoms in the
structure. In terms of composition, the number of E0 units in star(12)PEO was designed
to match those in PEGDME-500. The synthesis of star(12)PEO is shown in Scheme 9.
The starting materials are inexpensive and a pure product can be obtained through
vacuum distillation in relatively good yield. Star(12)PEO was characterized through 1H
NMR, 13C NMR, and GCMS. Since the 'H NMR signals for star(12)PEO overlap,
confirmation of the structure was obtained by 13C NMR. Seven distinct carbon

resonances can be seen in the 13C NMR spectrum (Figure 62). The identification of the

121

quartemary carbon and methyl carbon were confirmed by Distortionless Enhancment by

Polarization Transfer (DEPT) NMR.

 

K metal / THF
CH34<OCH2CH2}20H > CHa-(OCHZCHgtO'W
\ Br

0 Br

8 Br
0 B

O\/\O/
0N

1.
O?

O

2

Scheme 9. Preparation of star(12)PEO.

122

ngcvwvommmficmwmoomcvo0032.thchwhow

 

:85 a.

 

 

 

853333 e E22 o: .ae 2:3..—

E3

 

 

98% £0

 

 

 

 

Ammoommommoommuamoofovu \

 

83on «EU

 

 

 

 

a

 

 

123

Electrolytes with different O/Li ratios were prepared by dissolving LiClO4 in 1
gram quantities of star(l2)PEO. The thermal and conductive properties of the
electrolytes were studied by DSC and Impedance Spectroscopy. DSC heating scans of
the electrolytes, taken after flash quenching from the melt, are shown in Figure 63. Only
glass transitions were observed, with the Tg increasing as the salt concentration increases
(O/Li decreases). This effect is commonly seen in PEO electrolytes, and usually is
ascribed to interactions between lithium cations and oxygen on the PEO chains. The
cations act as transient cross-linkers, decreasing the chain mobility and increasing Tg.
Figure 64 shows that the relationship between Tg and mole fraction of salt in the
electrolyte is linear. For comparison, the data for PEGDME-500 (the linear oligomer) are
included in Figure 64 and are nearly identical to the star(12)PEO data over the range
shown. The shape of the oligomer should have a minimal affect on its interactions with

salt, therefore similar Tg behavior is expected.

The difference in oligomer shape does have a significant effect on crystallization
behavior. As seen in the DSC scans for star(12)PEO, there is no evidence for
crystallization or melting at any of the salt concentrations or for the pure oligomer.
Previously reported DSC scans of PEGDME-500 at varyious O/Li ratios show markedly
different behavior with obvious crystalization and melting transitions throughout the
series (DSC scans not shown).84 Table 8 lists the physical properties of both oligomers
for reference. The linear structure of PEGDME-500 allows for more uninterupted van
der Walls and dipolar interactions between chains, but the central sp3 hybridized carbon

in star(12)PEO acts as an inherent defect, shortening the effective PEO chain length that

124

can participate in crystallization. Because of geometrical restraints, star(12)PEO cannot

crystallize in the 72 helix seen for crystalline PEO.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

no salt
139 _
_““““—' J A
__ J 39
J‘ ' v 35
_ —_/-
29
~————o——'/‘
T 26
I‘— '—-/-
8 —/- 1 22
C .»— 18
LL] W/x MW '
15f
r‘ J‘
12
fl - J .- “M " 8
M- M 4
z ”5.3",”
-150 -100 -50 ° 50 100

Temperature (°C)

Figure 63. DSC heating scans of star(l2)PEO at several salt concentrations
(O/Li ratios). The scans were run at 10 °C per minute after slow cooling from
the melt.

125

Temperature (°C)

 

-10 1
-2o - °

-30 u

-70 . I. O star-PEG
W I PEGDME-500

-80 - l
-90 l:
-100

r I I I

o 0.1 0.2 0.3 0.4
Xsan (mole fraction)

 

 

Figure 64. The variation of Tg with molar salt fraction for star(12)PEO
and PEGDME-500

126

 

 

 

 

 

O/Liratio: 3 4 8 12 15 18 22 26 29 35 39 60 139 nosaltl
Tg star(12)PEO ~18 -49 -61 -67 -73 -75 -78 ~80 -81 -82 -85 -91 -93
T8 PEGDME-500 -50 -58 -67 -69 -73 -74
"‘Tc PEGDME-500 -14 -40 -53 -61 -64

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Table 8. The variation of Tg with salt concentration (O/Li ratio) for star(12)PEO and
PEGDME-500. Tc for PEGDME-500 mixtures is also given. (Note: There is no Tc
for star(12)PEO mixtures). All PEGDME data were obtained by first flash quenching
to —100°C and then heating at a rate of 10°C per minute. The star( 12)PEO data were

independent of whether the samples were initially slow cooled or fast quenched to
—100 °C.

The lack of crystallinity at low temperatures in the star(12)PEO electrolytes
should result in better low temperature conductivities compared to linear PEO systems
which crystallize. Impedance measurements were run between —40 °C and 80 °C. Figure
65 shows that at 20 °C the maximum conductivity of star(l2)PEO electrolytes (1 x 10"
S/cm) falls between O/Li = 18-29. At 80 °C the maximum (1 x 10'3 /S/cm) shifts to O/Li
= 15 due to increased solvation of charge carriers. At lower temperatures, salt
aggregates form more easily lowering conductivity. At 10 °C the conductivity maximum
(2 x 10'5 S/cm) shifts to lower salt concentrations O/Li = 29-35. Arhenius plots for
several salt concentrations are shown in Figure 66. Each data set shows curvature over
the full range of temperatures, —40 °C to 80 °C, indicating a dependence of conductivity
on free volume (VTF behavior). This densification of PEO chains at low temperatures

makes ion mobility more dependent on chain relaxations.

127

 

 

 

 

1.0E-02 :
' aooc
E "'-..
Q 1.0E-03 :
£0, El
:3 : 20°C
'2". . ”s
3
'2 o
8 1.0E-04
. -10°c 9
. ‘AA‘
. A
A
1.0E-05 ' '
o 50 100 150
O/Li ratio

Figure 65. Conductivity of star(12)PEO electrolytes as a function of the
O/Li ratio.

128

 

 

 

 

1.0E-02
A O/Li=15
A O/Li=18
1.05.03 - l l . ’ O’L'=22
. o O/Li=29
’E‘ ' . .oxLi=35
Q I
Q 1.0E-04 - e g
.é‘ . ‘
> A U
z:- A A
8 . '
‘2 1.0E-05 ' 2'
O A I
0 A .
9
1.0E-06 P ‘ .
2 8 .
1.0E-07 ' ‘ ' l-—
2.5 3 3.5 4 4.5
1000/T (K)

Figure 66. Arhenius plots of the conductivity of several star(12)PEO
electrolytes.

129

II. Plasticization of PEG

1. Oligomeric PEO’s

In gel electrolytes, the conductivity of high molecular weight PEG and other
polymers can be improved by the addition of small molecules called plasticizers that
hinder crystallization and increase the mobility of polymer chains. Here we study the
effects of addition of star(12)PEO and PEGDME-500 on the thermal properties of a
higher molecular weight PEO, Mn=3400 (PEG-3400) which readily crystallizes. Both
oligomers were added to PEG-3400 at 5 wt%, 20 wt%, and 50 wt%, and the resulting

melting and crystallization behavior were studied via DSC and optical microscopy.

The DSC heating and cooling scans for star(12)PEO are shown in Figure 67
while those for PEGDME-500 are shown in Figure 68. Tables 9 and 10 list values for
melting and crystallization transitions, including those for pure PEG-3400 and
PEGDME-500 for reference. All melting temperatures reported as the peak of each
transition. Pure PEG-3400 shows 2 closely spaced melting transitions likely due to
different sized lamella that form upon crystallization. Thinner lamella are less stable
(higher surface to volume ratio) and have lower melting points. The addition of either
oligomer decreases the crystallinity and melting temperature of PEO-3400 markedly.
Qualitatively, pristine PEG-3400 is a brittle solid at room temperature, and incorporating

50 wt% of either oligomer results in mixtures that are soft and waxy.

130

 

 

Heating Scans

 

PEG-3400

 

5%

 

20%

—>

50%

 

 

 

 

[Cooling Scans
PEG-3400

it
it
v
V._,

Endo

 

20%

50%

 

 

 

 

 

-100 -50 0 50
Temp (°C)

Figure 67. DSC heating and cooling scans for different compositions of
Star(12)PEO and PEG M..=3400.

131

 

 

 

 

Heating Scans

 

PEG-3400

 

5%

 

20%
50%

 

 

 

 

—>
|

 

 

 

 

 

 

 

 

'8 Cooling Scans

C -

LIJ PEG 3400 - __
5%
v“
V

-1 00 -50 O 50
Temp (°C)

Figure 68. DSC heating and cooling scans for different compositions of

PEGDME-500 and PEG Mn=3400.

132

 

 

normalized for wt of total sample normalized for wt of PEO 'm sample
wt% star(12)PEO Tm1(°C) Tm2(°C) AH,(J/g) AH2(J/g) AH total AHrU/g) AH2(J/g) AH total

 

 

5 ~ 59 177 177 177 177
20 49 58 37 115 152 37 137 173
50 47 54 33 38 71 62 72 134
100 - -

 

wt% PEGDME-500

 

 

 

 

 

 

 

 

 

 

 

 

 

5 ~ 60 138 138 138 138
10 ~ 58 158 158 188 188
50 9 52 67 122 189 127 232 359
100 16 ~ 88

 

Table 9. Melting transitions and heats of fusion for mixtures of PEG-3400 with
star(12)PEO and PEGDME-500. Data for 100% PEGDME-500 and PEG-3400 are
shown for comparison. The data are from second DSC heating scans at a rate of 5°C per
minute. AH values were calculated based on the total weight of the sample and for the
amount of PEG-3400 in the sample

 

normalized for wt oftotal sample normalized for wt of PEO in sample
wt% star(12)PEO Tc1(°C) Tcz(°C) AH,(J/g) AH2(J/g) AH total Alina/g) AH2(J/g) AH total

 

 

5 ~ 41 ~16l ~161 ~161 ~l61
20 - 37 ~146 ~146 ~173 ~l73
50 ~ 32 ~73 ~73 ~139 ~139
100 ~ ~

 

wt%PEGDME-SOO

 

 

 

 

 

 

 

 

 

 

 

 

5 ~ 41 ~135 ~l35 ~135 ~135
10 17 38 ~29 ~152 ~181 ~34 ~181 ~215
50 8 28 ~69 ~l l9 ~187 ~13l ~226 ~356
100 9 ~ ~90

 

Table 10. Crystalline transitions and heats of fusion for mixtures of FED-3400 with
star(12)PEO and PEGDME-500. Data for 100% PEGDME-500 and PEG-3400 are
shown for comparison. The data are from second DSC heating scans at a rate of 5°C per
minute. AH values were calculated based on the total weight of the sample and for the
amount of PEG-3400 in the sample

133

 

The ability of star(12)PEO and PEGDME-500 to inhibit crystallization are
different. The DSC data for 50 wt% PEGDME-500 show an additional crystallization at
7.7 °C and melting at 8.7 °C which nearly matches the crystallization and melting
behavior of pure PEGDME-500. Therefore, the PEGDME-500 system is a 2 phase
system at room temperature consisting of an amorphous liquid primarily composed of
PEGDME-500 and a crystalline solid of mostly PEG-3400. Since star(12)PEO is not
crystallizable, no crystalline/melting transitions should be seen for this branched system
even if it is phase separated. Since the phase separation behavior is unclear in this case
and at low weight percent oligomer for both cases, the thermal data in Tables 9 & 10
have been normalized for both miscible (total sample weight) and phase separated
(weight of PEG-3400 in sample) possiblities. In both cases star(12)PEO seems to be a
better inhibitor of crystallization. In the phase separated case, AH(fus) of pristine PEO-
3400 drops from 251.0 J/g to 134.1 J/g with 50 wt% star(12)PEO and only to 232.2 J/g

using PEGDME-500.

Optical microscopy clearly shows phase separation at 50 wt% for both oligomers
at room temperature. Images were taken under cross-polarizers at 50x magnification
after cooling samples at a rate of 5 °C/min (the same rate as used for DSC scans). A
dendritic morphology (Figures 70B and 71B) was observed instead of the typical
Spherulites seen for PEO~3400 (Figure 69). Dendritic crystallites have also been
observed for the case of PEO mixed with poly(ethylene oxide)~bIock-poly(dicyclohexyl
itacon'ate),99 where it was determined that the primary nucleation of the PEO chains was

disturbed by the presence of the block copolymer. During the growth of dendritic

134

Spherulites, the concentration of locally available and crystallizable PEO segments is
decreased, and crystallization becomes dependent on a secondary crystallization process,
the diffusion of crystallizable chains to the crystallization site. Slow diffusion results in

the dendritic morphology. The dendritic morphology was also observed at 20 wt %

 

Figure 69. Sperulitic morphology of PEO-3400 viewed at 50x magnification
through cross-polarizers. The sample was prepared by cooling from the melt
at a rate of 5 °C/min.

star(12)PEO, though to a lesser degree (Figure 70A). At 20 wt% PEGDME-500 there is
no observable phase separation by DSC and a simple increase in spherulite nucleation

sites is observed due to the disruptive effect of linear oligomer (Figure 71A).

135

 

Figure 70. Morphology of A. PEG-3400, 20 wt% star(12)PEO and B. PEO-3400
50 wt% star(12)PEO viewed 50x magnification through cross-polarizers. Mixtures
were cooled from the melt at a rate of 5 °C/min.

136

 

Figure 71. Morphology of A. PEO-3400, 20 wt% PEGDME-500 and B. PEG-3400
50 wt% PEGDME-500 viewed 50x magnification through cross-polarizers. Mixtures
were cooled from the melt at a rate of 5 °C/min.

137

2. Fillers

The use of high surface area fillers to decrease crystallinity in solid polymer
systems has been well established, but the effect of the surface chemistry of the filler on
crystallization has not been studied. The fillers used in this study range from
hydrophobic to hydrophilic, and were included in the polymer at 10-20 wt%. The silicas
include: R805 (n-octyl), R974 (dimethyl), FS-AME8b (ethylene oxide, average E0
length = 8), A200-dry (minimum Si-OH), and A200-wet (hydrated surface). Composites
were prepared by incorporating the filler into PRO-3400 above the PEO melting point.
The structures of all silicas including R974 and lFS-AMESb are shown in Figure 72.
Thermal data for each sample were obtained from DSC scans which are shown in

Figures 73-76.

The DSC scans in Figures 73-76 show that the addition of the fillers to PEG-3400
had no significant effect on Tm, but all fillers decreased the onset of Tc. The AH data in
Table 11 show that, with the exception of A200-wet, all fillers at 10 wt% loading
decreased the crystallinity of pure PEO by approximately 36%. The DSC scans at 20
wt% (Figure 77) show similar trends for R805, R974, and A200-dry with the
crystallinity reduced to 55% of the value seen for pure PEG-3400. A200-wet silica had
the largest effect, reducing crystallinity to 40% compared to pristine PEG-3400. In
addition, both Tm and Tc were lowered by approximately 10 °C. These results are most
likely due to the increase in surface hydroxyl groups and therefore are increased in
dipolar interactions with the PEO chains, subsequently decreasing their ability to

crystallize. It is also possible that surface adsorbed water may solvate the PEO phase and

138

.mm

A~200

R-974

R-805

FS-AMEBb

TOM

R-711

“3,0 H
3102 §~o H
f0 H

.........

fi' , . “1:; /O\
sro2 .;:—o;sv- ....O

139

hydrophilic, disperses
in PEO

hydrophobic, forms
moderately strong networks
in PEO

hydrophobic, forms strong
networks in PEO

hydrophilic, compatible with
PEG, disperses in PEO

hydrophobic, forms strong
networks in PEO,
crosslinkable

hydrophobic, commercial
analog of TOM, exact
structure uncertain

Figure 72. Functionalized fumed silica fillers

 

 

Heating

 

 

 

PEG-3400

 

——>

A

 

 

 

 

 

 

 

 

 

 

 

 

 

 

20%
g Cooling
LU PEG-3400
10% _' _
20%
V
0 20 40 6‘0

Temp(°C)

Figure 73. DSC heating (top) and cooling (bottom) scans of PEG-3400
with A200-dry. The scans were run at 5 °C per minute.

140

 

 

 

 

Heating

 

PEG-3400

10% A
e/\__

20%

 

 

——>

 

 

 

Cooling

 

Endo

PEG-3400

_10% V

20%

 

 

 

 

 

 

 

 

Temp(°C)

Figure 74. DSC heating (top) and cooling (bottom) scans of FED—3400
with AME8b. The scans were run at 5 °C per minute.

141

 

 

 

 

Heating

 

PEG-3400

 

10%

—->

20%

#%

 

 

 

 

Cooling

 

Endo

PRO-3400

 

10% e

W

I I I

o 20 4o 60
Temp(°C)

20%

 

 

 

 

Figure 75. DSC heating (top) and cooling (bottom) scans of PEG-3400
with R974. The scans were run at 5 °C per minute.

142

 

 

 

 

 

 

 

 

 

 

 

PEG-3400
10% A
T 20% _,_/\
'8 Cooling
2
LIJ
PEG-3400 ._

 

 

10%

V
V
V

 

 

 

 

 

0 20 4o 60
Temp (°C)

Figure 76. DSC heating (top) and cooling (bottom) scans of PEG-3400
with R805. The scans were run at 5 °C per minute.

143

 

 

Sample Tm(°C) AH (J/g) Tc(°C) AH (J/g)
PEG-3400 60.8 251 41.4 -268
Most hydrophobic 10% R805 59.1 167 36.2 - 169
filler
20% R805 59.6 137 37.5 -137
10% R974 61.1 162 39.7 ~166
20% R974 59.4 143 38.4 - 139
10% AME8b 61.9 165 41.2 -160
20% AME8b 59.7 104 39.8 - 133
10% A200 59.4 160 38.5 -180
20% A200 60.1 135 38.0 - 139
i 10% A200(wet) 62.6 123 40.2 - 125
Most hydrophilic
filler 20% A200(wet) 50.7 101 30.5 -1 l3

 

 

 

 

 

 

 

 

 

 

 

Table 11. PrOperties of PEG-3400 composites containing 10 and 20 wt%
of filler. The data for PEG-3400 is provided for comparison. All data
were taken from DSC scans at 10 °C per minute.

144

 

Endo

 

 

r Heating

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

PEG-3400 J—
R805 _,_./\___
R974 A
AME8b A
A200 _’_/L_
Cooling
PEG-3400
R805 \/
R974 V
AME8b V
.200 \/
A200 thdrated) V

\/f

20 40 60
Temp(°C)

Figure 77. DSC heating (top) and cooling (bottom) scans of PEG-3400 with
20 wt% fillers. The scans were run at 5 °C per minute.

decrease crystallinity. TGA weight loss experiments show that A200-dry loses 0.5 wt%

from the loss of surface hydroxyls and adsorbed water, while A200-wet loses 1.5 wt%.

One surprising observation was the effectiveness that E0 functionalized FS-

AME8b has on limiting crystallinity at 20 wt% filler. The filler is dry (T GA results show

only a 0.5 % weight loss after 60 minutes at 120 °C) and thus surface water cannot

explain the observed effect. It is possible that FS-AMESb particles with their randomly

placed and relatively short PEO chains, have the same effect as star(12)PEO. In other

words, the short tethered PEO chains disperse into the PEO medium and inhibit
crystallization. If true, there should be a large difference in the degree of solvation, or

effective particle radius, between FS-AME8b and R805 and A200—dry. As shown for
PEGDME-500 electrolytes, the hydrophobic n-octyl chains on R805 result in phase
separation of the particles from the polar PEO medium. The more hydrophilic A200-dry

should disperse and have good interaction with PEG at the particle surface where the
silanol groups are located. This would result in a smaller effective radius for both R805

and A200-dry compared to FS-AMESb. Figure 78 illustrates how these polymer
surfaces might interact with PEG. The dashed circle around each particle represents the
effective particle radius, or the extent to which each particle can reach into the PEO
medium. Note that the surface interactions on R805 may not account for the mechanism
crystallinity disruption in PEO. It is more likely that network aggregates of R805 in the
melt stage hinder PEO crystallization kinetics as forming lamella must overcome the

attractive van der Walls forces between R805 particles if they are to crystallize fully.

I46

 

 

 

Figure 78. Illustration of how PEO chains may interact with R805, A200, and
FSOAMESb fillers based on their surface properties. Dashed circles represent
the effective particle size based on the solubility of the surface modifier in the
PEO medium

147

 

We conclude that compatible short—chain plasticizers are effective at limiting the

crystallinity of solid PEO systems. This finding is similar to that found for branched

PEO polymers.49 PEO samples using filler alone are brittle, while samples using

star(12)PEO may creep at higher oligomer ratios. A composite made with both FS-
AME8b and star(12)PEO in a PEG-3400 may result in minimal crystallinity while
maintaining good mechanical properties. For low temperature applications, star(12)PEO

electrolytes with a hydrophobic filler such as R805 may be the best choice as these

electrolytes are amorphous at all temperatures.

Finally, the effect of short PEO segments in a microblock copolymer were

examined. The low Tg polyethylene (PE) block should help to perturb crystallinity in the

PEO segments and improve the conductivity of the polymer.
III. Alternating Microblock Compolymers

The mobility of cations in polyether-based solid polymer electrolytes is coupled
to the segmental motion of the polymers, and thus the ionic conductivities of polymer
electrolytes correlate with the glass transition temperature of the polymer host.2’17'34'100
Well-designed polymer electrolytes have room temperature conductivities that range
from 104 to 10'5 S/cm, values that are almost independent of the details of the polymer
structure, and characteristic of polyethers in general. Recently we investigated a series of
(AB)n microblock copolymers that contain alternating polyethylene oxide and alkenyl

blocks (refered to as polyethylene, PE, blocks for simplicity).4"-”101'103 The preparation

148

 

 

and characterization of these polymers is reported elsewhere.104 For reference, the

synthetic route, generic structure and the abbreviations for the polymers are shown in
Scheme 10, where 1: represents the two carbon atoms that form the double bond in the
polymer. The structures of the (AB)n polymers are related to the many “linked”

polyethers that have been studied as host polymers for solid polymer electrolytes. The

H2C=CH-(CH2)4-O' Na” + Tso(CH20Hzo))—(Ts
H2C=CH—(CH2)4-O-(CHZCHZO);(CH2)4-CH=CH2
metathesis catalyst

—{(CH2)4—CH=CH-(CH2)4-O-(CHZCH20)-:— + H2C=CH2

 

[C41tC4EOx]n

Scheme 10. Synthesis of (AB),. microblock copolymers

syntheses of the linked polyethers were motivated by the need to minimize the poorly

105

conducting crystalline phase of PEG-salt complexes. A variety of functional groups

were used to link oligomeric segments of PEO, including methylenes,”’1°6'1°7

109,110

siloxanesfa'108 aluminates, and other groups. The use of oligomeric PEO with a

149

 

 

distribution in chain lengths was preferred since the mixture of lengths tends to inhibit

crystallization.‘ 11

The (AB)n system here differs in that the segments of the polymer are intentionally
designed to have exact lengths. This allows us to study in detail the evolution of the
physical and conductive properties of the polymers as a function of the segment length.
In addition, the regularity of the segment length leads to self-assembly of the polymers
into 2-dimensional layered structures that contain alternating layers of ion conducting
polyethers separated by insulating alkenyl segments. An idealized layered structure of
the (AB)n system with all double bonds in trans conformation is shown in Figure 79.
Recall that PEO readily crystallizes in a 72 helix, while PE crystallizes in a zig-zag

conformation.

The use of alkenyl (PE) segments to link polyethers is an attractive approach
because they are highly flexible, and provide a site for further elaboration of the polymer
to give branched or cross-linked materials. The ethylene oxide segments provide the lone
pairs of electrons needed to dissolve lithium salts, and thus the addition of LiClO4 should
lead to good room temperature ionic conductivities. Previously we reported the
properties of polymer salt complexes for [C41tC4EOx],,, x = 36.45 The conductivities of
the polymers increased as the length of the E0 segment increased. One reason for the
increase was the different coordinating behavior of the polymers. For x = 3, two E03

segments coordinate with each lithium ion, while for x = 4,5, only a single E0x segment

150

 

 

 

 

Figure 79. Ideal packing structure of [C.,,{]C4EO,]n polymer chains
where y = 7, and the geometry of the double bonds is trans.

151

Y“.-

was needed for coordination to Li+. Obviously, the conductivity of the polymer
electrolyte cannot continue to increase without limit, and thus we synthesized new (AB)n
polymers to identify the limiting conductivity for this polymer system. In this work, we
extend our earlier study to polymers with longer E0 blocks (x=6, 7, 8, 10 and 14) and use
DSC and AC impedance measurements to compare these results with those obtained for
the shorter blocks. The thermal characterization data for the polymers are shown in
Table 12. The T35 of the polymers increase with increasing E0 length, approaching that

of high molecular weight PEO. All of the polymers are crystalline, and the melting

 

JE(CH2),,—cH=CH—(CH2),,—-0(CH20H20)y—}n

 

 

 

 

 

 

 

 

 

polymer x,y Mn PDI T g Tm
[C 41tC 4E03]n 4,3 32,200 2.12 -79 4
[C 41tC 4E0 4]n 4,4 28,100 2.16 -77 -5
[C 41t:C 4E05]n 4,5 31,500 2.01 -74 7
[C 41tC 4E06]n 4,6 48,800 1.84 -72 16
[C 41tC 4E07]n 4,7 55,500 1.79 -67 8
[C41CC413081n 4,8 19,400 1.61 -71 18
[C 41tC 4E010]n 4,10 98,500 1.38 -65 12
[C41cC4EOM]n 4,14 41,700 1.52 -55 24

 

 

 

 

 

 

 

 

a. Temperatures reported in °C

Table 12. Physical properties of [C41ttC4EOy]n polymers.

152

 

points show an interesting odd-even effect (Figure 80, squares). Studies of small
molecule analogs show that the odd-even effect is related to a planar zig-zag
conformation for the E0 segment instead of its usual helical conformation.1°1'102 This is
likely a result of PE segment influence on crystallization behavior. Both Tg and Tm
increase with increasing E0 segment length. The details of the DSC scans are
complicated, with the scans typically showing multiple melting and crystallization
transitions. Only the highest melting transitions are included in Table 12. It is important
to note that for all polymers with E0 block lengths <14, the highest melting transition is

below room temperature and thus the polymers are amorphous.

Polymer/salt complexes were prepared from LiClO4. As shown in Figure 81, the
addition of the salt has a simple effect. The crystallinity due to the pure polymer phase
decreases monotonically and eventually is lost. At the same time, the Tg increases with
salt content (Figure 80, diamonds). Plotting the Tg as a function of the mole fraction of
salt in the complex (data not shown) yields a linear relationship similar to the data shown

for star(12)PEO.

Table 13 outlines the thermal properties of the series of polymers with longer E0
units at O/Li ratios = 24, 32, and 48. The DSC heating scans for each polymer are also
shown in Figures 82-86. We characterized the entire series of [C41cC4EOx].,/LiClO4
complexes and found that the conductivity maximum occurred at an O/Li ratio of 32. At
this ratio, DSC scans after flash quenching from the melt show that the polymers

typically retain some remnant fraction of the pure homopolymer phase. As shown in

153

 

 

Temperature (°C)

 

40

 

E
20-»
I U I
- IE].- I
07- I
-2o-E
40%
t 000’ ’
-80": ..
-100. l l I l = l 4 4 l ;

DI

 

 

I Tm

9 T9

[:1 Tm(salt)
<> Tg(salt)

 

 

8

 

 

10
Number of EO units

0
01

Figure 80. Melting and glass transitions for [C4EJC4EOy]n polymers with

and without salt.

154

15

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

4
____..e—a-—~#r- ._
8
_— f—
12
——-——'—"'/-_
16
PM
24
-.—i—'-‘-'/_
o 1. ., 32
U 9 ‘. “,3 v. .. ”w“ fl _H_,,.-..,...._..~.,--~-.-..~-m~—~——~—
C
tu
48
64
pure
-100 -50 0 50 100

 

Tern perature (°C)

Figure 81. Evolution of the thermal properties of the polymer (C4DC4EOx),._
with increasing salt content. The numbers refer to the O:Li ratio. Adapted

from Qiao.45

155

 

 

 

 

 

 

 

 

EOunits O:Liratio Tg(°C) Tc(°C) Tm1(°C) Tmz(°C)

24 -63 -41 -12 ~9

6 32 -64 -48 -1o ~9
48 -68 -57 -12 ~9
24 -6O -39 -5

7 32 -64 -48 -4 ~15
48 -67 -55 -6 ~15
24 -59 -38 -2

8 32 -63 -47 -1
48 -66 -57 1 ~9
24 -57 -33 8

10 32 -61 -43 9
48 -64 -53 12
24 -57 -34. 20

14 32 -60 -43 18
48 -63 -53 22

 

 

 

 

 

 

 

 

 

Table 13. Thermal properties of [C41tC4EOy]n (y = 6-14) electrolytes obtained from
DSC scans at 10 °C per minute

156

 

48

 

—>
I
|

32

 

Endo

24

 

 

 

l l I L l l l l l

-100 -80 -60 -40 -20 0 20 40 60 80 100

 

Temperature (°C)

Figure 82. DSC traces for polymer salt complexes prepared from
[C4EJC4E014]n and LiClO4 (O/Li=24, 32, & 48).

157

1r

 

 

48

 

—>

32

 

Endo

 

 

I I I j I I I I I

-100 -80 -60 -40 -20 0 20 40 60 80 100

 

 

Temperature (°C)

Figure 83. DSC traces for polymer salt complexes prepared from
[C4DC4E010],. and LiClO4 (O/Li=24, 32, & 48).

158

 

48

 

—>

32

 

Endo

24

 

 

 

I I I I I I I I I

-100 -80 -60 -40 -20 0 20 40 60 80 100

 

Temperature (°C)

Figure 84. DSC traces for polymer salt complexes prepared from
[C4DC4E03],. and LiClO4 (O/Li=24, 32, & 48).

159

 

 

48

—>

 

32

Endo

 

24

 

 

 

 

I I I I r I I I I

-100 -80 -60 -40 -20 0 20 4O 60 80 100

Temperature (°C)

Figure 85. DSC traces for polymer salt complexes prepared from
[C4DC4EO7]n and LiClO4 (O/Li=24, 32, & 48).

160

_ -_‘"“H.fl-.-

 

 

 

'48

 

 

—»

32

Endo

24

 

 

 

I I I I I I I I I

-10 -80 -60 -40 -20 0 20 40 60 80 100
0

 

Temperature (°C)

Figure 86. DSC traces for polymer salt complexes prepared from
[C4DC4E06],, and LiClO4 (O/Li=24, 32, & 48).

161

 

Figure 87 for an O:Li ratio of 32, the polymers with different E0 lengths have similar
glass transitions and crystallization temperatures, and differ primarily in the details of the

melting transitions.

Temperature dependent conductivities were measured for these materials and the
results are shown in Figure 88. The data from our earlier study45 (open symbols) are
plotted for comparison. The conductivities increase with increases in the "PEG-content"
of the polymers, presumably due to a simple increase in the volume fraction of the
conductive phase in the materials. Room temperature conductivities reach 5 x 10'5 S/cm,
a value typical of amorphous PEO materials. For salt complexes with O/Li+ = 48 (data
not shown), the data show a large drop in conductivity at room temperature. Examination
of the DSC scans confirms the formation of a pure crystalline homopolymer phase,
identical to that of the salt-free homopolymers. Crystallization of the polymer host
apparently decreases the mobility of the polymer chain segments in the amorphous phase,
and leads to the observed drop in conductivity. We note that PEO/salt complexes show
the opposite effect. For these materials, the well-know decline in conductivity at room
temperature stems from the formation of a crystalline PEG/salt phase that immobilizes

the lithium salt.

In summary, despite their modest PEO volume fractions, these high molecular
weight polymers give room conductivities as high as 5x10'5 S/cm. Crystallinity begins to

limit conductivity for samples with E0 blocks >10. The crystalline phase formed,

162

 

 

 

14

 

 

 

Endo

 

 

I I I I I I I I I

-100 -80 -60 -40 -20 0 20 40 60 80 100

 

 

 

Temperature (°C)

Figure 87. DSC traces for polymer salt complexes prepared from [C4DC4EOy]n
and LiClO4 (O/Li=32). The number next to each trace is y, the number of
ethylene oxide units in the PEO block of the polymer.

163

 

 

o" (S/cm)

 

 

 

 

 

 

1.E-03
I.
. a
s a
1
1.5414 :" a 8 O
’ Q a.
: El 0
. O D e
' El
' @EO14 O O
AEO10
: eEoe
; @EO? O
. OEOS
_ 1:11:04
<>EO3
1.E-06 ‘ ‘ '
2.8 3 3.2 3.4
1000/T (K)

Figure 88. Temperature dependent conductivity for [C4DC4EOy]n/
L100;

164

however, is a pure homopolymer phase, and no samples studied showed evidence for
formation of a crystalline polymer/salt phase. Crystallization in the (AB)n system could
be futher frustrated by introducing branches at the unsaturated sites along the chain via
hydrosilation chemistry. Short polyether-based branches would also increase the salt-
host fraction in the polymer and may improve conductivity. In addition, the double
bonds of the polymers also provide sites for cross-linking reactions which could increase

mechanical stability if needed.

165

 

 

 

Chapter 4. Experimental

1. Materials

Unless otherwise specified, ACS reagent grade starting materials and solvents
were used as received from commercial suppliers without further purification. Aerosil
200 (A200) was a gift from Degussa A.G., Frankfurt, Germany. This silica has a surface
silanol content of 1 mmol/g, and was pre-treated by hydrating over a half-saturated
NH4N03 solution in a dessicator for at least one week. Commercially available methyl
methacrylate (MMA) and butyl methacrylate (BMA) were distilled over KOH to remove
di-tert-butyl hydroxytouene (BHT) radical inhibitor. Radical inhibitor, di-t-butyl
hydroxy toluene (BHT) was removed from polyethyleneoxide dimethylether, Mn = 500
(PEGDME-500) by passing it through a hydroquinone-based inhibitor removal column

packing from Aldrich.

166

 

 

 

 

II. Analytical Methods

1H and 13‘C nuclear resonance analyses were carried out at room temperature in
deuterated chloroform (CDCl3) on a Varian Gemini-300 spectrometer with the solvent

proton and carbon signals being used as chemical shift standards.

A Nicolet IR/42 spectrometer purged with dry nitrogen was used to obtain
infrared spectra. Solid samples used were 1 cm2 pressed pellets prepared from ca. 10 mg
of functionalized silica for a 2% silica to KBr ratio by weight. Thermal polymerization
spectra of composites for kinetic data were obtained by sandwiching the composite
between two 32 x 3 mm NaCl salt plates with a 0.015 mm copper spacer and heating to
the desired temperature in a Spectrotech variable temperature IR apparatus. Spacers of
from 0.015 to 0.5 mm thicknesses were used, all giving the same experimental result, but
with the 0.015 mm spacer giving the best signal to noise ratio. Hence, this was the
thickness of choice. All spectra reported were acquired by signal averaging 32 scans at a

resolution of 4 cm".

Kinetic polymerization data was extrapolated from IR spectra by hand
normalizing and subtracting spectra on Excell. Spectra at several time points were
compared to initial spectra of un-polyrnerized composite for each sample. Due to
sensitivity in the fingerprint region (400-800 cm"), not all spectra were to scale and

normalization was accomplished by comparing the heights of the carbon-hydrogen

167

'1va iv'..1".'-$E 1 HT

 

 

 

stretching region (2800 3000 cm'l), which were at a constant and high concentration due
to the unchanging nature of alkyl segments in PEGDME-500. After normalization, the
disappearance of the acrylate C=C stretch, and hence % monomer consumed, was

measured by subtracting the signal at time, t to t = 0 min (initial monomer concentration).

Unless otherwise specified, differential scanning calorimetry (DSC) were
performed under a helium atmosphere at a heating rate of 10 °C/min on a Perkin Elmer

DSC 7. The DSC temperature was calibrated with an indium standard.

 

Thermogravimetric analyses (TGA) were performed in air atmosphere at a heating rate of
10 °C/min on a Perkin Elmer TGA 7 instrument. Samples for TGA measurements were

first dried in vacuum at 100°C overnight.

The rheological properties of composites were measured in dynamic shear using a
Rheometrics Mechanical Spectrometer (RMS 800) with 50mm parallel plate geometry set

at 1 mm thickness. A steady pre-shear of 0.5 rad/s was applied to composites for 60

 

seconds followed by a 120 second rest period to erase any sample loading history. In all

experiments a low strain amplitude of yo = 1.3% was maintained to ensure measurements
were done in the linear viscoelastic regime. The oscillatory frequency, (1), was varied

from .01 to 100 st.

The light scattering properties of composites were probed by passing a medium

pressure He laser beam through 1 cm2 quartz cells filled with composite. The degree of

168

~—

 

phase separation and therefore scatter was qualitatively measured as intensity of light to
pass through the cell and be received by a silicon photodiode hooked up to a multimeter.
Before measurement composites were carefully evacuated to remove air bubbles. All
samples were compared to a standard composite of 13% TOM (4:1) and 87% PEGDME-
500 (O/Li=20) which is optically clear. An average of four readings was taken at several

different vertical positions on the cell to insure no macro phase separation was occurring.

The electrical properties of the samples were measured by AC Impedance
Spectroscopy over the frequency range 6 Hz - 11 MHz using an HP 4192A LF
Impedance Analyzer. An applied voltage of 0.1 V was used to collect the data.
Conductivity measurements were taken at room temperature (approx 20°C), 30°C, 45°C,
55°C, 65°C, and 80°C, with the samples held at each temperature for at least 15 minutes
under a nitrogen gas purge before being measured. Low temperature measurements were
done in a glove bag purged with nitrogen for at least 2 hours. A dewer filled with liquid
nitrogen was place in the glove bag as a water trap. To effect cooling, a metal coil was
place in series with the nitrogen line connected to the conductivity cell. This coil was
submerged in liquid nitrogen and the flow rate maximized. The length of tubing from the
coil to the conductivity cell was minimized and insulated to decrease wamring of the
cooled nitrogen by the atmosphere. The lowest temperature achieved was —50°C. Once
cooled sufficiently, the nitrogen flow rate was turned down and impedance measurements
were taken at from —40°C to 10°C at 5 degree intervals while warming up. All
electrolytes were prepared in a helium filled dry box by dissolving LiClO4 in the PEO-

based medium of choice. Uncrosslinked composite samples containing methacrylate

169

1"; JJiL'wT-ur. 0'

 

 

 

 

monomer were void of AIBN initiator to insure that little or no cross-linking occurred

during measurements.

Cross-linked samples were prepared by sandwiching uncross-linked electrolyte
between a glass slide and a stainless steel disk, with the desired thickness of 0.2 mm
maintained using a Teflon spacer. The composites were polymerized under a nitrogen
atmosphere by placing each sample 2:1 cm from a water cooled 450 W medium pressure
mercury lamp. Cross-linking was initially performed using both pyrex and quartz cover-
slips. It was found that there was no difference in polymerization rate and pyrex was
used thereafter. After polymerization, the glass slide was peeled off gently and replaced
with another stainless steel disc for impedance analysis. A general operating, set-up, and
data retrieval procedure for the HP 4192A LF Impedance Analyzer in Dr. James Dyes

lab, room 10 Chemistry follows.

High resolution mass spectra (HRMS) data were obtained at the Michigan State
University Mass Spectrometry Facility which is supported, in part, by a grant (DRR-
00480) from the Biotechnology Research Technology Program, National Center for
Research Resources, National Institutes of Health. A direct inlet was used, and samples

were held at 30 °C for 2 minutes followed by a 16 °C/min warm up to 300 °C.

170

 

 

 

General procedure for the HP 4192A LF Impedance Analyzer

1. Set-up

Figure 87 gives a general idea of what the Impedance set-up looks like. A
specially made impedance measurement apparatus has been created to do 2 probe,
variable temperature (room temperature to 200°C) A.C. impedance. For this purpose
there are two BNC adapters at the base of the apparatus for attaching leads to and from
the 2-probe port of the impedance analyzer as well as a J-type thermocouple. For
temperature control, the thermocouple on the apparatus should be plugged into a J-type
temperature control box while the power supply cord is plugged into a variac, which is in
turn plugged into the temperature control box. The temperature control box is then
plugged into a basic power supply. There is also removable glass bell covering the
apparatus and an inlet and outlet at the base for N2 or some other inert gas. This is to

allow impedance measurements on air sensitive compounds if needed.

Once set up, turn on the impedance analyzer and the temperature control box (this
will give a room temperature reading even if not doing variable temperature runs). Place
the sample cell between the two stainless steel electrode contacts. One is under the
stainless steel lid, which lifts off, the other is on the metal finger. Turn on the variac and
the set control box to the desired temperature if a run at elevated temperatures is desired.
For solid samples, let equilibriate 20-30 minutes at each temperature, liquids,

approximately 15 minutes.

171

 

 

 

 

I Frequency display

 

 

 

Impedance Analyzer
2 probe port

 

 

 

 

 

 

 

   

 

Temp Control Box

 

to power supply

V‘
a

20°C

 

 

 

 

 

 

 

 

 

 

 

 

 

———

 

N2 out N2 in

 

 

 

 

 

 

Figure 89. Impedance apparatus set-up

172

 

2. Taking Measurements

Once set up, make sure that the two leads in the apparatus attached to each
electrode contact are securely in place in the base pins. Open the ACIMPD software in
the CONDUCTIVITY program. The voltage frequency range should already be set from
13 Hz to 1 1 MegaHz and the measurement period at 1 second. The voltage amplitude
will default to 0.005 and must be changed to 0.01 Volts. Click on RUN, enter a FILE
NANIE, and then click SAVE. This will begin your run, each of which takes

approximately 1-2 minutes.

3. Working with Data

To work with the data once the run is done, open up the desired file in the Excel]
program (it must be delimited to convert data to columns). The file can be found in the
C: Drive in the CONDUCTIVITY Folder (you may want to create your own folder here

as it makes files easier to find). Once open in Excel, there will be three data
colurnns...Column A = Z (complex impedance), Column B = phase, Column 3 =
frequency. The data in Columns A&B is used to make the Nyquist plot but is in spherical
coordinates and must be converted to rectangular coordinates first. The calculations are
as follows...the X-axis = column A X cos(Column B X 3.14159/ 180), the Y-axis = -
Column A x sin(Column B x 3.14159/ 180). Note: The X-axis represents the Real
Impedance (Re(Z)) and the Y-axis is the Imaginary Impedance (Im(Z)). Once plotted,
the graph that results should be similar to the one shown in Figure 6, where the bulk

resistance is the intercept on the X-axis of the low frequency spike. If there is no true

173

 

‘M' 1—
p

 

 

intercept, you must extrapolate back by estimating where the semicircle would touch
down on the X-axis. The value for conductivity is obtained through equation 11 (0' = 1/

RA).

Also note, when saving your Excel file you must remember to save it as an Excel
Workbook. When finished with the experiment you can close the ACIMPD Labview
window by clicking EXIT, then closing the window, and finally hitting QUIT for

Labview.

174

f A‘-‘ a: [ar’w fi

 

 

 

III. Hydrophobic-Polymerizable Silicas

FS-TOM. The experimental procedure described in this example was used to make
all FS-TOM silica fillers. To 15 g of hydrated A200 in a 500 mL round bottomed flask
was added 450 mL of toluene containing 1.5 mL (15 mmol) of diethylamine. The flask
was attached to a mechanical shaker, and a mixture of octyltrimethoxysilane (OTMS) and
trimethoxysilylpropyl methacrylate (TPM) in 30 ml of toluene were added. The ratio of
OTMS to TPM was dependent on the degree of hydrophobicity desired. The reaction
was allowed to proceed at room temperature overnight. The product was separated by
filtration, and washed with 3 x toluene and 3 x diethyl ether. The residual toluene/diethyl
ether was evaporated and the solid was crushed via mortar and pestal and transferred into
a Schlenk flask and dried under vacuum at 120°C overnight. The degree of surface

functionalization in each silane depended on the starting ratio (Table 14).

 

 

 

 

 

 

 

 

Sample Name A200 TPM TPM OTMS OTMS
(TPMIOTMS) (g) (8) (mo!) (8) (mmol)
FS-TMSPM 15 3.0 12.4 o 0
FS-TOM(4:1) 15 1.5 6.2 1.5 6.2
FS-TOM(4:3) 15 1.2 4.8 1.7 7.2
FS-TOM(1:1) 15 0.75 3.0 2.25 9.6
FS-TOM(1:4) 15 0.33 1.33 2.7 11.5
FS-TOM(1:9) 5 .05 0.2 .95 4.1
FS-OTMS 15 O 0 3.0 12.4

 

 

 

 

 

 

 

 

Table 14. Products of mixed silylation reactions.

175

IV. Preparation of Composites

Composites in 1-5 gram quantities were prepared in a helium filled dry box. In
each instance, AIBN initiator was predissolved in a methacrylate (MA) monomer of
choice, then a fumed silica was added so the maximum of MA/AIBN mixture would
absorb on the surface. In most cases 13 wt % of fumed silica was used as it was
determined to be the minimum amount of filler needed to form a non-flowing gel.
Monomers used were methyl methacrylate (MMA), butyl methacrylate (BMA), and octyl
methacrylate (OMA) and were incorporated at desired ratios (usually 10 wt%). Next,
polyethylene glycol dimethyl ether, Mn = 500 (PEGDME-500) with LiClO4 salt
predissolved to a ratio of O/Li=20 was added. Finally, mechanically blending of
composites with a small-scale blender paddle, adapted from a drill bit and attached to a
drill, was employed for approximately 2-3 minutes. Composite samples were used

immediately after preparation

Octyl methacrylate. To a 1000 mL round bottomed flask were added 500 mL
CC14, 58.8 g (451.5 mmol) of l-octanol, and 15 g of crushed 3A molecular sieves. The
mixture was heated to reflux and 57.6 ml (587.1 mmol) methacryloyl, chloride in 80 mL
CC14 was added over a period of 30 min under argon. Heating was continued overnight.
The molecular sieves were removed by filtration, and after removing the solvent, the
product was purified using vacuum distillation at 60°C, 10 mtorr; yield = 95%. IH NMR:
56.07 (1 H, m), 55.52 (1 H, m), 54.11 (2 H t), 51.92 (3 H, 8), 51.63 (2 H, m), 51.25 (10 H

m), 50.86 (3 H, t). Procedure adapted from Hou.84

176

 

 

V. Preparation of Star-PEO

Star(12)PEO: To a 2000 ml round bottom equipped with a magnetic stir bar was
added 85.7 g (.71 mmol) diethylene glycol monomethyl ether, 800 ml dry THF
(Na/benzophenone), and 27.9 g (.7lmmol) potassium metal. The mixture was refluxed
overnight, approximately 12 hours until all potassium was reacted. A solution of 45.9 g
(.ernmol) pentaerythritol tetrabromide in 200 ml THF was added slowly via syringe.
The solution immediately turned yellow and eventually cloudy and was let stir at 100°C
for 2 days. The reaction mixture was cooled and quenched with 2 N HCl and extracted
with 3 x 200 ml diethylether. Resulting ethereal solvent was combined, dried over
MgSO4, filtered, and condensed. The resulting oil was left to dry over vacuum at 100°C
overnight, washed with activated carbon, and run through a short silica plug. Finally, the
product was obtained through vacuum distillation using a jacketed distillation apparatus
and heating tape. Star(12)PEO distilled at 225-235°C, 43 mtorr with final yield of 26.2
g, yield = 40.8%. 1H NMR: 53.64-3.42 (8 H, m), 5342-338 (2 H, broad 8), 5338-324
(3 H, broad s). '3 C NMR: 571.84 (CH2), 570.87 (CH2), 570.35 (CH2), 570.26 (CH2),
569.87 (CH2), 558.88 (CH3), 545.31 (C). HRMS calc. for C25H52Ol2 544.3459, found

344.3444.

177

 

10.

11.

12.

13.

14.

15.

Bibliography

Linden, D. Handbook of Batteries; second ed.; McGraw-Hill Inc.: New York,
1995.

MacCallum, J. R.; Vincent, C. A. Polymer Electrolyte Reviews - 2; Elsevier
Science Publishers Ltd.: New York, 1989.

Carvalho, L. M.; Guegan, P.; Cheradame, H.; Gomes, A. S. Eur. Polym. J. 2000,
36, 401-409.

Borodin, O.; Smith, G. D. Macromolecules 1998, 31, 8396-8406.

Gray, F. M. Solid Polymer Electrolytes--Fundamentals and Technological
Applications; VCH publishers, Inc.: New York, 1991.

Vincent, C. A. Prog. Solid State Chem. 1987, I 7, 145-261.

Chen, H. P.; Fergus, J. W.; Jang, B. Z. J. Electrochem. Soc. 2000, 147, 399-406.
Klassen, B.; Aroca, R.; Nazri, M.; Nazri, G. A. J. Chem. Phys. B. 1998, 102,
4795-4801.

Ding, M. S.; Xu, K.; Jow, T. R. J. Electrochem. Soc. 2000, 147, 1688-1694.
Scrosati, B. Electrochim. Acta 2000, 45, 2461-2466.

Peled, E.; Golodnitsky, D.; Menachem, C.; Bar-Tow, D. J. Electrochem. Soc.
1998, 145, 3482-3486.

Aurbach, D.; Bin-Ely, Y. J. Electrochem. Soc. 1995, 142, 1746.

Ogumi, Z.; Inaba, M. Bull. Chem. Soc. Jpn. 1998, 71, 721.

Chu, A.; Josefowicz, G.; Farrington, J. J. Electrochem. Soc. 1997, 144, 4161.

Blomgren, G. E. J. Power Sources 1999, 81 , 112-118.

178

 

 

 

l6.

17.

18.

19.

20.

21.

22.

23.

24.

25.

26.

27.

28.

29.

30.

31.

Armand, M. B.; Chabagno, J. M.; Duclot, M. Second International Meeting on
Solid Electrolytes St. Andrews, Scotland, 1978.

MacCallum, J. R.; Vincent, C. A. Polymer Electrolyte Reviews-1; Elsevier
Applied Science: New York, 1987; Vol. 1.

Staudinger, H.; Lohmann, H. Ann. Chim. 1933, 505, 41.

Hill, F. N.; Bailey, F. E.; Fitzpatrick, J. T. Ind. Eng. Chem. 1958, 50, 5.
Takahashi, Y.; Sumita, I.; Tadokoro, H. J. of Polymer Sci. .' Polymer Phys. Ed
1973, 11, 2113-2122.

Tashiro, K.; Tadokoro, H. Reports Prog. Poly. Phys. Jpn. 1978, 21, 417-420.
Tadokoro, H.; Chatani, Y.; Yoshihara, T.; Tahara, S.; Murahashi, S. Makromol.
Chem. 1964, 73, 109-127.

Yoshihara, T.; Tadokoro, H.; Murahashi, S. J. Chem. Phys. 1964, II, 2902-2904.
Koenig, J. L.; Angood, A. C. J. of Polymer Sci: Part A-2 1970, 3, 1787-1796.
Liu, K.-J.; Parsons, J. L. Macromolecules 1969, 2, 529-533.

Machida, K.; Miyazawa, T. Spectr. Acta 1964, 20, 1865-1873.

Matsuura, H.; Miyazawa, T. Spectr. Acta 1967, 23A, 2433-2447.

Matsuura, H.; Fukuhara, K. J. Mol. Struct. 1985, 126, 251-260.

Matsuura, H.; Fukuhara, K. J. Polym. Sci. Pt. B-Polym. Phys. 1986, 24, 1383-
1400.

Kanig, B. Progr. Colloid Polym. Sci. 1975, 57, 176.

Strobl, G. R. The Physics of Polymers: Concepts for Understanding Their
Structures and Behavior; 2nd ed.; Springer-Verlag Berlin Heidelberg: New York,

1997.

179

 

 

32.

33.

34.

35.

36.

37.

38.

39.

40.

41.

42.

43.

45.

46.

47.

e1 Maaty, M. I. A.; Hosier, I. L.; Bassett, D. C. Macromolecules 1998, 31 , 153-
157.

Vaughan, A. S.; Bassett, D. C. Comprehensive Polymer Science; Pergamon Press:
New York, 1989; Vol. 2.

Gray, F. M. Solid Polymer Electrolytes: Fundamentals and Technological
Applications; VCH Publishers, Inc.: New York, 1991.

Chabagno, J. M.;; Institute NationalPolytechnique, Grenoble, 1980.

Craven, J. R.; Mobbs, R. H.; Booth, C.; Giles, J. R. M. J. Makromol. Chem,
Rapid Commun. 1986, 7, 81.

Craven, J. R.; Nicholas, C. V.; Webster, R.; Wilson, D. J .; Mobbs, R. H.; Morris,
G. A.; Heatley, F.; Booth, C.; Giles, J. R. M. 1987, 19, 509-516.

Nicholas, C. V.; Wilson, D. J .; Booth, C. 1988, 20, 289.

Buriez, 0.; Han, Y. B.; Hou, J .; Kerr, J. B.; Qiao, J.; Sloop, S. E.; Tian, M. M.;
Wang, S. G. J. Power Sources 2000, 89, 149-155.

Blonsky, P. M.; Shriver, D. F. J. Am. Chem. Soc. 1984, 106, 6854-6855.

Shriver, D. F. J. Electrochem. Soc. 1987, 134, 270.

Abraham, K. M.; Alamgir, M. Chem. Mot. 1991, 3, 339-348.

Nagaoka; Watanabe, M. J. Polym. Sci. Polym. Lett. 1984, 22, 659.

Spindler, R.; Shriver, D. F. J. Am. Chem. Soc. 1988, 110, 3036-3043.

Qiao, J .; Chen, Y. Y.; Baker, G. L. Chem. Mat. 1999, 11, 2542-2547.

Jannasch, P. Macromolecules 2000, 33, 8604-8610.

Mertens, I. J. A.; Wubbenhorst, M.; Oosterbaan, W. D.; Jenneskens, L. W.; van

Tumhout, J. Macromolecules 1999, 32, 3314-3324.

180

 

 

48.

49.

50.

51.

52.

53.

54.

55.

56.

57.

58.

59.

61.

62.

Cho, K. Y.; Lee, K. H.; Park, J. K. Polym. J. 2000, 32, 537-542.

Kono, M.; Hayashi, E.; Watanabe, M. J. Electrochem. Soc. 1998, 145, 1521-1527.
Nishimoto, A.; Agehara, K.; Furuya, N.; Watanabe, T.; Watanabe, M.
Macromolecules 1999, 32, 1541-1548.

Abraham, K. M.; Alamgir, M. J. Electrochem. Soc. 1990, 13 7, 1657-1657.
Kono, M.; Hayashi, E.; Nishiura, M.; Watanabe, M. J. Electrochem. Soc. 2000,
147, 2517-2524.

Lee, K. H.; Park, J. K.; Kim, W. J. J. Polym. Sci. Pt. B-Polym. Phys. 1999, 37,
247-252.

Choe, H. S.; Carroll, B. G.; Pasquariello, D. M.; Abraham, K M. Chem. Mat.
1997, 9, 369-379.

Croce, F.; Brown, S. D.; Greenbaum, S. G.; Slane, S. M.; Salomon, M. Chem.
Mat. 1993, 5, 1268-1272.

Williamson, M. J .; Hubbard, H. V. 8.; Ward, 1. M. Polymer 1999, 40, 7177-7185.
Mustarelli, P.; Quartarone, E.; Capiglia, C.; Tomasi, C.; Ferloni, P.; Magistris, A.
J. Chem. Phys. 1999, 111, 3761-3768.

Wang, H. P.; Huang, H. T.; Wunder, S. L. J. Electrochem. Soc. 2000, 147, 2853-
2861.

Hou, x. P.; Siow, K. s. Polymer 2000,41, 8689-8696. I

Rajendran, 8.; Uma, T. Bull. Mat. Sci. 2000, 23, 27-29.

Fuller, J .; Breda, A. C.; Carlin, R. T. J. Electrochem. Soc. 1997, 144, L67-L70.
Saito, Y.; Capiglia, C.; Yamamoto, H.; Mustarelli, P. J. Electrochem. Soc. 2000,

147, 1645-1650.

181

 

 

63.

65.

66.

67.

68.

69.

70.

71.

72.

73.

74.

75.

76.

Reiche, A.; Sandner, R.; Weinkauf, A.; Sandner, B.; Fleischer, G.; Rittig, F.
Polymer 2000, 41, 3821-3836.

Xu, W.; Siow, K. S.; Gao, Z. Q.; Lee, S. Y. J. Electrochem. Soc. 1999, 146, 4410-
4418.

Nishiura, M.; Kono, M.; Narnegaya, M.; Matsuda, Y. Electrochem. Sol. St. Lett.
1998, 1, 246-248.

Kono, M.; Hayashi, E.; Watanabe, M. J. Electrochem. Soc. 1999, 146, 1626-1632.
Murata, K.; Izuchi, S.; Yoshihisa, Y. Electrochim. Acta 2000, 45, 1501-1508.
Weston, J. E.; Steele, B. C. H. 1982, 7, 75.

Park, H. C.; Chun, J. H.; Kim, S. H.; Ko, J. M.; Jo, S. I.; Chung, J. S.; Sohn, H. J.
J. Power Sources 2001, 92, 272-276.

Appetecchi, G. B.; Dautzenberg, G.; Scrosati, B. J. Electrochem. Soc. 1996, 143,
6-12.

Abraham, K. M.; Koch, V. R; Blakley, T. J. J. Electrochem. Soc. 2000, 147,
1251-1256.

Reiche, A.; Dlubek, G.; Weinkauf, A.; Sandner, B.; Fretwell, H. M.; Alam, A. A.;
Fleischer, G.; Rittig, F.; Karger, J .; Meyer, W. J. Phys. Chem. B 2000, 104, 6397-
6407.

Liquan, C. Chem. J. Chinese Univ. 2 1986.

Borghini, M. C.; Mastragostino, M.; Passerini, S.; Scrosati, B. J. Electrochem.
Soc. 1995, 142, 2118.

Capuano, F.; Croce, F.; Scrosati, B. J. Electrochem. Soc. 1991, 138, 1918-1922.

Croce, F.; Scrosati, B.; Mariotto, G. Chem. Mot. 1992, 4, 1134-1136.

182

77.

78.

79.

80.

81.

82.

83.

84.

85.

86.

87.

88.

Wieczorek, W.; Such, K.; Wycislik, H.; Plocharski, J. Solid State Ion. 1989, 36,
255-257.

Scrosati, B.; Croce, F.; Persi, L. J. Electrochem. Soc. 2000, 147, 1718-1721.
Appetecchi, G. B.; Croce, F.; Dautzenberg, G.; Mastragostino, M.; Ronci, F.;
Scrosati, B.; Soavi, F.; Zanelli, A.; Alessandrini, F.; Prosini, P. P. J. Electrochem.
Soc. 1998, 145, 4126-4132.

Appetecchi, G. B.; Croce, F.; Mastragostino, M.; Scrosati, B.; Soavi, F.; Zanelli,
A. J. Electrochem. Soc. 1998, 145, 4133-4135.

Sun, H. Y.; Takeda, Y.; Irnanishi, N.; Yamamoto, O.; Sohn, H. J. J. Electrochem.
Soc. 2000, 14 7, 2462-2467.

Wen, Z. Y.; Itoh, T.; Ikeda, M.; Hirata, N .; Kubo, M.; Yamamoto, O. J. Power
Sources 2000, 90, 20-26.

Eichinger, G.; Fabian, M. J. Power Sources 1997, 68, 387-391.

Hou, J. Composite Polymer Electrolytes using F unctionalized

F umed Silica and Low Molecular Weight PEO:

Synthesis and Characterizatoin; Michigan State University, 1997.

Joseph, R.; Zhang, 8.; Ford, W. T. 1996, 29, 1305.

Ferry, J. D. Viscoelestic Properties of Polymers; John Wiley and Sons, Inc.: New
York, 1961.

Raghavan, S. R.; Riley, M. W.; Fedkiw, P. S.; Khan, 8. A. Chem. Mat. 1998, 10,
244-251.

Walls, H.; Zhou, J.; Yerian, J.; Fedkiw, P. S.; Khan, S. A.; Stowe, M.; Baker, G.

J. Power Sources 2000, 89, 156-162.

183

 

 

89.

90.

91.

92.

93.

94.

95.

96.

97.

98.

99.

100.

101.

102.

Fan, J .; Raghavan, S. R.; Yu, X. Y.; Khan, S. A.; Fedkiw, P. 8.; Hon, J .; Baker, G.
L. Solid State Ion. 1998, 111, 117-123.

Polymerization Processes; John Wiley and Sons, Inc.: New York, 1977.

Warson, H. The Application of Synthetic Resin Emulsions; Benn: London, 1972.
Polymer Handbook; fourth ed.; John Wiley and Sons, Inc.: New York, 1999.
Electronic and Photonic Applications of Polymers; American Chemical Society:

Washington DC, 1986.

fr I ”Junta”, a-W‘]?
1 .

Hoyle, C. E.; In Core-active Photoinitiators for Surface Polymerization; Hoyle,
C. E. and Kinstle, J. F., Ed.4l7;American Chemical Society: Washington, DC,
1989.

Fossey, J .; Lefort, D.; Sorba, J. Free Radicals in Organic Chemistry; John Wiley
and Sons: New York, 1995.

Elias, H. G. An Introduction to Polymer Science; VCH Publishers, Inc.: New
York, 1997.

Khan, 8. A., Collaborator, NC. State.

Smart, M.; Ratnakumar, B.; Surampudi, S. J. Electrochem. Soc. 1999, 146, 486-
492.

Buzarovska, A.; Koseva, S.; Cvetkovska, M.; Nedkov, E. European Polymer
Journal 2001, 37, 141-149.

Meyer, W. H. Adv. Mater. 1998, 10, 439-+.

Chen, Y.; Baker, G. L. J. Org. Chem. 1999, 64, 6870-6873.

Chen, Y.; Baker, G. L.; Ding, Y. Q.; Rabolt, J. F. J. Am. Chem. Soc. 1999, 121,

6962-6963.

184

103.

104.

105.

106.

107.

108.

109.

110.

111.

Qiao, J .; Chen, Y.; Baker, G. L.,, (submitted to Macromolecules).

Chen, Y. Synthesis, Modeling and Characterization of Ethylene Oxide-Segmented
Microblock Copolymers; Michigan State University: East Lansing, 1999.
Berthier, C.; Gorecki, W.; Minier, M.; Armand, M. B. Solid State Ionics 1983, 11,
91.

Craven, J. R.; Mobbs, R. H.; Booth, C.; Giles, J. R. M. J. Macromol. Chem. Rapid
Commun. 1986, 7, 81.

Booth, C.; Nicholas, C. V.; Wilson, D. J .; In Polymer Electrolyte Reviews - 2;
MacCallum, J. R. and Vincent, C. A., Ed.2;Elsevier: Essex, England, 1989.

Sun, J. 2.; MacFarlane, D. R.; Forsyth, M. J. Polym. Sci. Pol. Chem. 1996, 34,
3465-3470.

Onishi, K.; Matsumoto, M.; N akacho, Y.; Shigehara, K. Chem. Mot. 1996, 8, 469-
472.

Giles, J. R. M. Solid State Ionics 1987, 24, 155.

Wagener, K. B.; Boncella, J. M.; Nel, J. G. Macromolecules 1991, 24, 2649-2657.

185