urdflfihnfi .w. V Ii: . . , . .
13.. .... u. ; ,. . If,
NM £19. . A Wfik‘?‘ {JWHQMWJYI .

., rhuvwvv. V: .lt.l

       

 

    

 

 

 

T}; t ;
7.. 1x45”

.. i. an aim“.

a. 3.1.2.

  

 

txrfinU

. ,v‘ ”Uh i

, . ; .3,” ”In

.5 4"! l. .

. a? 3hr“? Sow; n

“tfin; Le numwkmfi. 5
1.....‘ ..

. , .1}. kmmrimm

x .. i... «3.1.8;ch

.. firiririf _

gflwuafifl70? . XL”

 

.( vs .

u . ¥ .Axkt II

:4. fig! GL‘ 7.
I:

.mfinwéfifiu

2.3 I .. )fluwfyfi’.

v...‘ 1.3.11
. ; .6........r&om\~rn. 3:1
g.» . 3.... . : 5&3? s.
R§.nnr..:¢fiu8. sing.» 9%.: .m
$3.... 3.. 3...! I9‘. . . fl
mt! 354%,.3
{ammufaa ..

  

c,

 

             

 

( .. \l.

|.n I 5 film: . , i 4 10 A;
1'. . 1.!!- 1 it I ~ I 3 {nun-18

an"... x 1;:

      

 

.
is, v?
N‘rkth... liq.

 

.5-‘-‘."-‘.; ‘7
““"“ “'7'

 

7”

 

M)!
.P ‘3’!

a"

is:
I. ‘31:!!!
(I: 6.; .( {in-n1:

Viilé c$ stiltlnxil. 4
7r . 2.....1OXII .
.25: 113V37. I 1.. v A

;

 

      

\._

         

.Lr

 

. . ... ..§... .... w . tie," ‘ 4v. ..... ;
r _ F» ; gratui‘a My; 1 . ..
‘ , 363‘s. ( ‘1 l ‘

  

 

7"" F‘ LIBRARY
1 l Michigan 5}“
,. l ;,l “fluently

This is to certify that the
dissertation entitled

PROTON CONDUCTING MATERIALS: SYNTHESIS,
CHARACTERIZATION, AND APPLICATION IN FUEL CELLS

presented by
QIN YUAN

has been accepted towards fulfillment
of the requirements for the

DOCTORAL degree in CHEMISTRY

7 Major1 Professor’s Signature

’5 Decemée/ 250 7

Date

 

 

 

MSU is an Affirmative Action/Equal Opportunity Employer

 

 

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
MAR O 8 2013

1097 12
TV"!

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

5/08 K:lProj/Acc&Pres/ClRC/Dateom.indd

7, V,r V_ 7 ~————————

PROTON CONDUCTING MATERIALS:
SYNTHESIS, CHARACTERIZATION, AND APPLICATION IN FUEL CELLS

By

Qin Yuan

A DISSERTATION
Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of
DOCTOR OF PHILOSOPHY
Chemistry

2009

ABSTRACT

PROTON CONDUCTING MATERIALS:
SYNTHESIS, CHARACTERIZATION, AND APPLICATION IN FUEL CELLS

By

Qin Yuan

Recently, there has been considerable interest in the development of fuel
cells, especially hydrogen powered fuel cells, which cleanly convert the chemical
energy of fuels directly into electric energy and heat with high efficiency. A key
component of hydrogen based fuel cells is the proton exchange membrane
(PEM), which transports protons from the anode to the cathode and separates
fuel and oxidant. Nafion, the prototype PEM, has a high proton conductivity and
chemical and mechanical stability at moderate temperatures in the harsh fuel cell
environment. Despite the high performance of Nafion in fuel cells, its high
manufacturing cost, drastically reduced proton conductivity above 80 °C,
environmental incompatibility, and severe methanol permeability limit its use in
fuel cell applications.

We synthesized poly(sodium 4-styrenesulfonate)-coated silica
nanoparticles via surface initiated atom transfer radical polymerization (ATRP).
After acidification, we used a solution-casting process to embed the proton
conducting nanoparticles in a poly(vinylidene fluoride) (PVDF) matrix. In this
blend PEM system, the silica nanoparticles were the proton transport medium,
while PVDF gave the membrane favorable mechanical properties: toughness,

resilience, and minimal swelling. The proton conductivities were comparable to

Nafion 117, and depended on the volume fraction of the nanoparticles in the
membarne.

We also synthesized uniform magnetic nanoparticles with magnetite cores,
a silica buffer layer, and poly(4-styrenesulfonic acid) grown from the particle
surfaces. Applying an external magnetic field during membrane casting aligned
the particles in the through-plane direction, increasing the conductivity for a given
volume fraction of particles. These “magnetic” PEMs had higher conductivity and
better dimensional stability than Nafion 117.

Sulfonated polyimides (SPls) have proven to be one of the best
alternatives to Nafion due to their excellent thermal, chemical, and mechanical
stability, high proton conductivity, good film-fonning ability, and low gas
permeability. Using a polycondensation reaction, we synthesized a SPI and
prepared PEMs from the SPI/PVDF blends. Transmission electron microscopy
showed that the blends have a phase-separated morphology that evolved into a
bicontinuous morphology when the SPI weight fraction exceeded 30 %. The
conductivities were comparable to those of Nafion 117. These blends are good

candidates for high temperature fuel cell applications.

To my family

iv

ACKNOWLEDGMENTS

I would like to sincerely thank my advisor, Dr. Gregory Baker for his
guidance, assistance, and patience. My five-year experience in the Baker group
is one of the most important treasures in my life. I would like to thank my
committee members, Dr. William Wulff, Dr. Chi-Kwong Chang, and Dr. Merlin
Bruening for their guidance and helpful suggestions.

My thanks go to both previous and current Baker group members: quei,
Ping, Ying, Leslie, Feng, Bao, DJ, Erin, Sampa, Torn, Hui, Gina, Quanxuan, Heyi,
Wen, and Yiding and all my friends at Michigan State University. Thanks to Dan
Holmes, Kathy Severin, Riu Huang, and Alicia Pastor for their help on various
instruments.

Finally, I would like to thank my family for always believing in me and my
husband, Jun, for his love, support, and encouragement. I have benefited a lot

from his polymer background.

TABLE OF CONTENTS

LIST OF TABLES ............................................................................................... viii
LIST OF FIGURES ............................................................................................... ix
LIST OF SCHEMES ........................................................................................... xiv
LIST OF ABBREVIATIONS ................................................................................. xv
Chapter 1 Introduction ...................................................................................... 1
Nafion conducting mechanism .......................................................................... 2
lnorganic-Nafion composite membranes ........................................................... 6
Silicon oxide nanoparticles. .......................................................................... 6
Metal oxide nanoparticles ............................................................................... 7
Metal nanoparticles ......................................................................................... 8
Self-humidifying materials ............................................................................... 9
Block copolymers as an approach to alternatives to Nafion membranes ........ 12
Sulfonated block co-polyimides PEM ............................................................ 12
Sulfonated block co-po|y(arylene ether sulfone) PEM .................................. 14
Sulfonated poly(arylene ether sulfone)-b-polyimide PEM ............................. 16
Sulfonated polysulfone-b-poly(vinylidene fluoride) PEM ............................... 19
Sulfonated poly(styrenesuIfonate)-b-poly(methylbutylene) PEM .................. 20
Polymer blends as an approach to alternatives to Nafion membranes ............ 21
Van der Waals/dipole-dipole interaction ....................................................... 23
Hydrogen bonding interaction ....................................................................... 25
Electrostatic interaction (ionic cross-linking) ................................................. 26
Covalent cross-linking ................................................................................... 27
Controlling the ionic channels of PEMs ........................................................ 29
Electric field aligned blend membranes ........................................................ 30
Magnetic field aligned blend membranes ..................................................... 32

Chapter2 Direct Polymerization of Sodium 4-Styrenesulfonate from Silica
Nanoparticle Surface: Kinetics, Characterization, and Application in Fuel

Cells .............................................................................................. 37
Introduction ...................................................................................................... 37
Results and discussion .................................................................................... 40
Surface initiated ATRP of sodium 4-styrenesulfonate ................................... 4O
Fabrication and characterization of composite PEMs from the blending of
silica-PSSH nanoparticles and PVDF ........................................................... 52
Conclusion ....................................................................................................... 63
Experimental section ....................................................................................... 65

vi

Chapter3 Magnetic Nanoparticles: Synthesis, Characterization, and

Application in Fuel Cells ................................................................... , ............... 75
Introduction ...................................................................................................... 75
Results and discussion .................................................................................... 79
Synthesis and characterization of magnetic nanoparticles ........................... 79
Synthesis of poly(4-styrenesulfonic acid)—coated magnetic hybrid
nanoparticles (MNT-PSSH) by a “post-sulfonatlon”method .......................... 87
Direct polymerization of sodium 4-styrenesulfonate ..................................... 94
Fabrication and characterization of "magnetic" composite PEMs ................. 97
Conclusion ..................................................................................................... 104
Experimental section ..................................................................................... 105

Chapter4 Sulfonated Polyimide and PVDF based Proton Exchange

Membranes for Application in Fuel Cells ....................................................... 112
Introduction .................................................................................................... 1 12
Results and discussion .................................................................................. 114
Synthesis and characterization of sulfonated polyimide (SPI) .................... 114
Fabrication and characterization of PEMs from the blending of SPI (triethyl
ammonium salt) and PVDF ......................................................................... 120
Conclusion ..................................................................................................... 135
Experimental section ..................................................................................... 1 36

Chapter5 Synthesis and characterization of poly(4-styrenephosphonate)

for high temperature fuel cell application ...................................................... 141
Introduction .................................................................................................... 141
Results and discussion .................................................................................. 143
Conclusion ..................................................................................................... 150
Experimental section ..................................................................................... 150
References ........................................................................................................ 1 53

vii

LIST OF TABLES

Table 1. Characteristics of particles isolated at different polymerization time ..... 48
Table 2. Summary of silica-PSSH/PVDF membrane characteristics ................... 54

Table 3. Summary of SPI/PVDF membrane characteristics .............................. 121

viii

LIST OF FIGURES

Figure 1. Chemical structure of Nafion .................................................................. 1

Figure 2. Gierke’s cluster-network model for the morphology of hydrated Nafion..
................................................................................................................... 2

Figure 3. Polymer-bundle model. (a) Schematic view of polymer aggregate
domains inside a bundle. (b) Schematic view of nanoscopic swelling
between three polymer aggregates inside a bundle with different water
contents. (from almost dry membrane (left) up to diluted solution (right)). .3

Figure 4. Parallel water-channel model of Nafion. (a) Two views of an inverted-
micelle cylinder, with ionic side chains filling the water channel and
hydrophobic polymer backbone on the outside. (b) Schematic view of the
approximately hexagonal packing of inverted-micelle cylinders. (c) Cross-
sections through water channels (white) and crystallites (black) in the non-
crystalline Nafion matrix. ............................................................................ 4

Figure 5. Preparation procedure of Nafion-silica hybrid membranes .................... 7

Figure 6. Schematic illustration of multi-layer self-assembly of Pd nanoparticles
and Nafion ionomers onto Nafion membranes ........................................... 9

Figure 7. Schematic diagram of preparing exfoliated Pt-clay/Nafion composite
membrane ................................................................................................ 11

Figure 8. Structure of segmented bisulfonated poly(arylene ether sulfone)-b-
polyimide (BPSH-b—Pl) copolymer ............................................................ 17

Figure 9. Tapping mode AFM phase images of BPSH-b-Pl copolymers with
different block lengths. (both block lengths increase from top to bottom).
................................................................................................................. 18

Figure 10. Specific interactions between blending materials ............................... 22

Figure 11. TEM images of blend membranes with different degrees of sulfonation
and blend ratios: (a) 824F22; (b) $24F41; and (c) S39F22. (SxxFyy,
where xx represents the sulfonation level and yy stands for the content of
PVDF in the blend membrane) ................................................................. 24

Figure 12. (a) Schematic view of the polymer blend. (b) A schematic diagram of
the formation of a H-bonding complex between the sulfonic acid group and
PEO units ................................................................................................. 26

ix

Figure 13. Structure of polysulfone bearing 4-nitrobenzimidazole side groups ...27

Figure 14. Schematic view of the proton transfer mechanism with electrostatic
interaction between SPEEK/PSf-NBlm blending materials ...................... 27

Figure 15. Illustration of possible SPEEK cross-linking reaction ......................... 28

Figure 16. Schematic illustration of precursor particles with hydrophobic cores
(grey) and ionic surface (dark) and a film fabricated by self-assembly of
the particles .............................................................................................. 30

Figure 17. Light microscopy photographs of (a) non-ordered; and (b) ordered ion-
exchange particles in the polymeric component of the rubber. 1OOX,
electric field: 1330V/cm, 50 Hz ................................................................. 31

Figure 18. Optical micrographs of the alignment of sulfonated crosslinked
polystyrene ion exchange resin in a poly(dimethylsiloxane) matrix under
AC electric field of different magnitudes at 5 Hz frequency ...................... 32

Figure 19. Schematic diagram of the proposed mechanism for the formation of
magnetic particles .................................................................................... 34

Figure 20. Images of SPEEK membranes with magnetic nanoparticles (a)
randomly dispersed; and (b) aligned in magnetic field ............................. 35

Figure 21. FTIR spectra of (a) Snowtex-XS nanoparticles, CTAB salt; (b)
nanoparticles with surface anchored initiators; and (c) silica/PSSNa-3
nanoparticles. Samples were made into pellets with KBr ......................... 43

Figure 22. Thermogravimetric analysis of (a) Snowtex-XS nanoparticles, CTAB
salt; (b) silica-initiator nanoparticles; (c) silica-PSSNa-1 nanoparticles; (d)
silica-PSSNa-2 nanoparticles; (e) silica-PSSNa-3 nanoparticles; and (f)
linear PSSNa. See Table 1 for details of silica-PSSNa-(1-3). All samples
were run in air at a heating rate of 10°/min ............................................... 44

Figure 23. 500 MHz 1H NMR spectra of a representative polymerization sample
for the kinetic study .................................................................................. 45

Figure 24. Surface polymerization kinetics of sodium 4-styrenesulfonate from
silica nanoparticles. The polymerization was run at room temperature in
water/methanol (v/v = 3/1) and an initial monomer concentration ([M]o) of
sodium 4-styrenesulfonate is 0.5 mM. The ratios of
[M]o:[standard]:[surface initiator]: [CuBr]:[CuBr2]:[bpy] were approximately
70:10:1:2:O.2:6 ......................................................................................... 46

Figure 25. TEM images of (a) dried Snowtex-XS nanoparticles, CTAB salt; and
(b) silica-PSSNa-3 nanoparticles ............................................................. 49

Figure 26. FTIR spectra of (a) silica-PSSNa nanoparticles; and (b) silica-PSSH
nanoparticles ............................................................................................ 50

Figure 27. Thermogravimetric analysis of (a) Snowtex-XS nanoparticles, CTAB
salt; (b) silica-initiator nanoparticles; (c) silica-PSSNa nanoparticles; and
(d) silica-PSSH nanoparticles ................................................................... 52

Figure 28. (a) Schematic illustration of composite proton exchange membrane
fabrication. (b) Photo of a 4x4 cm2 composite membrane with 40 wt%
silica-PSSH particles ................................................................................ 53

Figure 29. Water uptakes of composite membrane as a function of particle
contents. The membrane water uptake was measured gravimetrically at
room temperature ..................................................................................... 56

Figure 30. Water uptakes as a function of membrane IEC. The membrane water
uptake was measured gravimetrically at room temperature ..................... 57

Figure 31. Membrane proton conductivities as a function of particle content for
composite PEMs at 100% RH and room temperature (0) and 60 °C (A).
Data for Nafion 117 at room temperature (I) and at 60 °C (I) was shown
for comparison .......................................................................................... 59

Figure 32. Membrane proton conductivity as a function of lEC for composite
PEMs at 100% RH and room temperature (O) and 60 °C (A). Data for
Nafion 117 at room temperature (I) and at 60 °C (I) was shown for
comparison ............................................................................................... 60

Figure 33. Membrane proton conductivity as a function of water uptake for
composite PEMs at 100% RH and room temperature (O) and at 60 °C (A).
Data for Nafion 117 at room temperature (I) and at 60 °C (I) was shown
for comparison .......................................................................................... 61

Figure 34. Cross-sectional TEM images of membranes with (a) 40 wt% particles;
and (b) 50 wt% particles ........................................................................... 62

Figure 35. TEM image of MNT nanoparticles ...................................................... 81
Figure 36. X-ray diffraction pattern of powder MNT nanoparticles. The vertical
bars correspond to the reflections and their relative intensity as recorded
in ASTM data cards for cubic Fe304 ......................................................... 82
Figure 37. TEM image of MNT-silica nanoparticles ............................................. 83

Figure 38. (a) TEM image of MNT-silica. (b) EFTEM iron map. (c) EFTEM silicon
map .......................................................................................................... 84

xi

Figure 39. Photos of MNT-silica nanoparticles dispersed in water (a) without an
external magnetic field (a fine dispersion of particles in water); and (b) with
an external magnetic field (particles were localized toward the magnet)...
................................................................................................................. 85

Figure 40. TEM images of chain-like structures formed by MNT-silica in response
to an external magnetic field (~1 Tesla) ................................................... 86

Figure 41. FT-IR spectra of (a) MNT-silica nanoparticles; (b) MNT-initiator
nanoparticles; (c) MNT-PSt nanoparticles; and (d) MNT-PSSH
nanoparticles. Samples were made into pellets with KBr ......................... 89

Figure 42. TGA analysis of (a) MNT-silica; (b) MNT-initiator; (c) MNT-PSSNa; (d)
MNT-PSt; and (e) MNT-PSSH .................................................................. 91

Figure 43. TEM image of (a) isolated MNT-PSt nanoparticles; and (b) “coffee ring
stain”-like aggregated MNT-PSt nanoparticles ......................................... 93

Figure 44. TEM image of MNT-PSSNa nanoparticles. Black spots are the Fe304
cores, the gray middle layer is the silica coating, and the light gray outer
shell is composed of PSSNa .................................................................... 95

Figure 45. FT-IR spectrum of MNT-PSSNa nanoparticles. The sample was made
into a pellet with KBr ................................................................................. 96

Figure 46. Room temperature hysteresis curves of magnetic nanoparticles. The
inset demonstrated the data around zero field from -200 Oe to 200 Oe.
................................................................................................................. 98

Figure 47. (a) Schematic illustration of corn osite proton exchange membrane
fabrication. (b) Photo of a 2.5x2.5 cm composite membrane with 30 wt%
MNT-PSSH particles cast in presence of an external magnetic field ........ 99

Figure 48. Polarization curves of composite membranes with different magnetic
nanoparticle contents and Nafion 117 at room temperature ................... 100

Figure 49. Cross-sectional TEM image of a 30 wt% MNT-PSSH aligned
membrane .............................................................................................. 101

Figure 50. (a) Cross-sectional TEM image of aligned composite membrane with
30 wt% MNT-PSSH content. (b) EDX line-scan along the line displayed in
(a) ........................................................................................................... 103

Figure 51. 500 MHz 1H NMR spectra of the triethylammonium salt form of SPI
synthesized with (a) dried starting materials, reagents, and solvents; and
(b) dried starting materials, but reagents and solvents used as received
............................................................................................................... 117

xii

Figure 52. FTIR spectrum of SPI (triethylammonium salt). The sample was made
into a pellet with KBr ............................................................................... 119

Figure 53. Thermogravimetric analysis of SPI (triethylammonium salt), run at a
heating rate of 10 °C/min in air ............................................................... 120

Figure 54. Water uptakes of blend membrane as a function of the SPI content.
The membrane water uptake was measured gravimetrically at room
temperature ............................................................................................ 123

Figure 55. Water uptakes as a function of membrane IEC. The membrane water
uptake was measured gravimetrically at room temperature Water uptake
as a function of membrane IEC. ............................................................. 124

Figure 56. Blend membrane proton conductivity as a function of SPI content for
composite PEMs at 100 % RH and room temperature (O) and 60 °C (A).
Data for Nafion 117 at room temperature (I) and at 60 °C (I) was shown
for comparison ........................................................................................ 126

Figure 57. Membrane proton conductivity (room temperature, 100 % RH) as a
function of lEC ........................................................................................ 127

Figure 58. Blend membrane proton conductivity (room temperature, 100 %
humidity) as a function of wateruptake .................................................. 128

Figure 59. Cross-sectional TEM images of blend membranes with different SPI
content: (a) 10 wt%; (b) 20 wt%; (c) 30 wt%; (d) 40 wt%; and (e) 50 wt%

............................................................................................................... 131
Figure 60. Surface SEM images of blend membranes with different SPI content:
(a) 10 wt%; (b) 20 wt%; (c) 30 wt%; (d) 40 wt%; and (e) 50 wt% ........... 134
Figure 61. 1H NMR spectra of poly(4-bromostyrene) and poly(4-
styrenephosphonate) .............................................................................. 145
Figure 62. 31P NMR spectrum of poly(4-styrenephosphonate) .......................... 146

Figure 63. DSC analyses of poly(4-bromostyrene) and poly(4-
styrenephosphonate). DSC samples were run in a nitrogen atmosphere at
a heating rate of 10 °C/min ..................................................................... 147

Figure 64. FTIR spectra of poly(4-bromostyrene) and poly(4-styrene
phosphonate). Samples were made into pellets with KBr ...................... 148

Figure 65. Thermogravimetric analysis of (a) poly(4-bromostyrene); (b) poly(4-

styrenephosphonate). All samples were run in an air atmosphere at a
heating rate of 10 °C/min ........................................................................ 149

xiii

 

LIST OF SCHEMES

Scheme 1. Synthesis of sulfonated block co-polyimides ..................................... 14

Scheme 2. Synthesis of segmented sulfonated multiblock co-poly(arylene ether
sulfone) with different linkage composition ............................................... 16

Scheme 3. Synthesis of Bisphenol A PSf-b-PVDF copolymers .......................... 20

Scheme 4. Synthetic route to hybrid poly(4-styrenesulfonic acid)/silica
nanoparticles ............................................................................................ 42

Scheme 5. Schematic diagram illustrating the structure of a composite
membrane ................................................................................................ 54

Scheme 6. Schematic illustration of two strategies for the synthesis of poly(4-
styrene sulfonic acid)-coated magnetic nanoparticles .............................. 80

Scheme 7. Schematic illustration of (a) synthesis of MNT nanopaticles; and (b)
structure of PAA-capped MNT nanoparticles ........................................... 80

Scheme 8. Synthesis of poly(4-styrenesulfonic acid)-coated magnetic hybrid
nanoparticles ............................................................................................ 88

Scheme 9. Synthesis route to poly(sodium 4-styrenesulfonate)-coated magnetic

hybrid nanoparticles ................................................................................. 94
Scheme 10. Synthesis of sulfonated polyimide (SPI) ........................................ 115
Scheme 11. Synthesis route to poly(4-styrene phosphonate) ........................... 144

Images in this dissertation are presented in color.

xiv

AC
AFM
BPSH
CTAB
Da
DMF
DMFC
DEG
DMSO
DSC
EDX
EFTEM
emu
Et3N (TEA)
GPC

h

Hz

lEC

mA

LIST OF ABBREVIATIONS

conductivity

ampere

alternating current

atomic force microscopy
bisulfonated poly(arylene ether sulfone)
cetyltrimethylammonium bromide
Dalton

dimethylforrnamide

direct methanol fuel cell
diethylene glycol

dimethyl sulfoxide

differential scanning calorimetry
dispersive X-ray

energy-filtered transmission electron microscopy
electromagnetic unit

triethyl amine

gel permeation chromatography
hour

hertz

ion exchange capacity

infrared spectroscopy

miliampere

XV

 

[Mic

meq
MHz
MNT

Mw
NaSS
NBlm
NMR
Oe

pmn
PAA
PDDA
PDI
PEl
PEM
PEO
PES
Pl
PMB
PMMA

meta

mole per liter

initial monomer concentration
milliequivalent

megahertz

magnetite

number average molecular weight
weight average molecular weight
sodium 4-styrenesulfonate
4-nitrobenzimidazole

nuclear magnetic resonance
oersted

para

parts per million

poly(acrylic acid)
poly(diallyldimethylammonium chloride)
polydispersity index

poly(ether imide)

proton exchange membrane
poly(ethylene oxide)

poly(ether sulfone)

polyimide

poly(methylbutylene)

poly(methyl mathacrylate)

xvi

PSf
PSS
PSSA
PSSH
PSSNa
PSt
PVDF

RH

SEM

SPEEK

SPEKK
SPI

TEOS

TEM

TGA
THF

XRD

polysulfone
poly(styrenesulfonate)

sulfonated polystyrene
poly(4-styrenesulfonic acid)
poly(sodium 4-styrenesulfonate)
polystyrene

poly(vinylidene fluoride)

room temperature

relative humidity

Siemens

scanning electron microscopy
sulfonated poly(ether ether ketone)
sulfonated poly(ether ketone ketone)
sulfonated polyimide

reaction time

temperature

tetraethyl orthosilicate
transmission electron microscopy
glass transition temperature
thermal gravimetric analysis
tetrahyd rofu ran

wflw

X-ray diffraction

xvii

Chapter 1 Introduction

Recently, there has been considerable interest in the development of fuel
cells, especially hydrogen powered fuel cells, which cleanly convert the chemical
energy of fuels directly into electric energy and heat with high efficiency. The
proton exchange membrane (PEM), which transports protons from the anode to
the cathode and acts as a separator to prevent mixing of reactant gases, is one
of the most important components of hydrogen fuel cells.1 Currently, the state-of-
the-art PEM is Nafion which was commercialized by DuPont in the late 19605. As
shown in Figure 1, Nafion has a perfluorinated backbone with pendent
fluorosulfonic acids and has demonstrated excellent performance in the harsh
fuel cell environment.2 In PEM fuel cells, Nafion can continuously work for up to
60,000 h with high conductivity (80 mS/cm) at 80 °C.3 Nafion’s outstanding
performance as a PEM has spurred researchers to study the morphology of
hydrated Nafion, and by inference, the mechanism responsible for its proton

conductivity.

fiCFZ—CFZHIF—Canb—I—

CFz-i::O(CF2)2—SO3H

Figure 1. Chemical stmcture of Nafion.

Nafion conducting mechanism

  
   
  

K— 5 nm —>-
so;

 

so; so;

   

so; so;
M so;

   
 

so; 505

       

  

so; 50’ 1'so;

  

Figure 2. Gierke’s cluster-network model for the morphology of hydrated Nation.4

A common starting point for understanding hydrated Nafion is Gierke’s

cluster model?"7

Gierke et al. proposed that water-swollen Nafion is comprised of
approximately spherical ionic clusters with an inverted micelle structure,
interconnected by small channels to form a continuous conductive pathway as
shown in Figure 2. As the water content increases, the ionic clusters grow and
combine into larger but fewer clusters in fully hydrated Nafion materials. Though
the cluster-network model has been widely used to explain Nafion’s morphology-
performance relationships, recent experiments have inspired alternative models
for Nafion’s morphology. Rubata and co-workers introduced a polymer-bundle
model, where elongated polymeric aggregates are surrounded with ionic charges
(Figure 3).‘Ho The local orientation for these aggregates had a correlation length
of a few hundreds of nanometers at both high and low hydration levels. The main

difference from the Gierke model is that water molecules are located in an

intrinsic fibrillar structure rather than spherical cavities.

 

 

Figure 3. Polymer-bundle model. (a) Schematic view of polymer aggregate
domains inside a bundle. (b) Schematic view of nanoscopic swelling between
three polymer aggregates inside a bundle with different water contents. (from
almost dry membrane (left) up to diluted solution (right)). (Reprinted with
permission from Macromolecules 2004, 37, 7772-7783. Copyright 2004
American Chemical Society)

The recent work of Schmidt-Rour supports a parallel water-channel model
for Nafion.11 They proposed that hydrophilic side groups align to form water
channels with an average diameter of 2.4 nm (Figure 4b). The channels were
stabilized by stiff hydrophobic backbone segments surrounding the channels,
resulting in a cylindrical inverted micelle structure (Figure 4a). Consistent with
other models,6'12 the Schmidt-Rour model gave a 10-20 nm repeat length

between crystallites. The crystallites are an important component in Nafion’s

structure as they function as physical cross-links responsible for Nafion’s
mechanical stability, as shown in Figure 4c. The water-channel model explains
many properties of Nafion, especially its high proton conductivity and water

permeability.13

    

 

 

hannel Crystallite

0

H2

Figure 4. Parallel water-channel model of Nafion. (a) Two views of an inverted-
micelle cylinder, with ionic side chains filling the water channel and hydrophobic
polymer backbone on the outside. (b) Schematic view of the approximately
hexagonal packing of inverted-micelle cylinders. (c) Cross-sections through
water channels (white) and crystallites (black) in the non-crystalline Nafion matrix.
(Reprinted with permission from Nat. Mater. 2008, 7, 75-83. Copyright 2007
Nature Publishing Group)

Though the actual morphology of hydrated Nafion is still under

investigation, it is widely agreed that Nafion’s phase separation into two domains

accounts for its high proton conductivity.""1“'15 The two domains function
separately: the hydrophilic domain affords proton transport while the hydrophobic
domain provides mechanical and chemical stability. This has been confirmed by
the work of Xu and coworkers, who found that when Nafion was dissolved and
recast into membranes, its proton. conductivity was four orders of magnitude
lower due to the break-down of Nafion’s phase-separated structure.16 Despite
having favorable PEM characteristics for fuel cell applications, Nafion has
significant drawbacks. The most critical limitation is its high price, which is a
significant cost of a fuel cell device and has limited the market penetration.17 Also,
its poor performance in direct methanol fuel cells (DMFCs), and a drastic
decrease in proton conductivity at elevated temperatures and/or low humidity
limit Nafion to applications to temperatures <90 °C."2

For applications in automotive, stationary, and portable power areas,
PEMs must satisfy several critical requirements including (1) high proton
conductivity (>10 mS/cm at 25 °C), (2) low electronic conductivity, (3) good
chemical, mechanical and thermal stability, (4) low permeability to fuel and
oxidant, and (5) low cost (<$200/m2).18 For fuel cells to replace combustion
engines, the US. Department of Energy has set a target for PEM performance: a
proton conductivity of 100 mS/cm at operating conditions of 120 °C and 50 %
relative humidity.

Extensive research has focused on modifying Nafion to overcome the
drawbacks of Nafion and make it more suitable for fuel cell applications, as well

as developing alternatives to Nafion.

lnorganic-Nafion composite membranes
Silicon oxide nanoparticles

SiOz has been extensively explored as an additive to enhance Nafion’s
physical properties.“”25 Nafion/silica composite membranes were prepared by
the addition of silica nanoparticles or sol-gel polymerization of silica precursors
catalyzed by the sulfonic acid groups of Nafion. The as-synthesized membranes
were reported to have outstanding performance at high temperatures. For
example, organic-inorganic hybrid Nafion membranes were synthesized via an in
situ sol-gel process, which formed mesoporous silica with sulfonic acid groups
(Figure 5).19 The mesoporous silica resulted from the co-condensation of
tetraethoxysilane and 2-(4-chlorosulfonylphenyl)ethyltrimethoxysilane during the
recasting of Nafion membranes, using an organic surfactant as the template. The
homogeneous composite membranes had improved dimensional stability and
water management compared to conventional Nafion membranes. More
importantly, including silica in the membrane increased the proton conductivity at
95 °C and over the entire range of relative humidity, compared to recast Nafion
and Nafion 112 membranes, which was ascribed to the presence of the

hygroscopic inorganic silica phase within the Nafion.

 

 

I Organosilane solution I I Nafion solution I

 

 

Drying under vacuum
Re-dissolving in ethanol

20 wt% Nafion
ethanol solution

Mixing
Hydrolyzing at rt for 20 h

Dissolving in ethanol

 

 

I Organosilica sol I

 

 

 

 

 

I Hybrid Nafion-organosilica solution I

 

Casting onto a glass support
Drying at ambient conditions
Annealing at 120°C for 10 h

 

I Hybrid Nafion-silica membrane I

 

Figure 5. Preparation procedure of Nafion-silica hybrid membranes.19

Metal oxide nanoparticles

TiOz has a high surface zeta potential and proton-induced charge density
over an extended temperature range (25-200 °C) and efficiently binds water
molecules through hydrogen bonding.”27 The electrochemical performance of
Nafion-TiOz composite membranes was investigated in a high temperature H2/02
fuel cell. The (composite membrane had a two-layer structure and the TiOz
powder was mainly deposited on one side of the membrane. The performance of
a composite membrane with 20 wt% Ti02 content, tested at 50 % RH, was
comparable to a conventional Nafion membrane at 100 % RH. The improved
water retention of the composite membrane suggests the possibility of fabricating

composite PEMs for operation at 120 °C and 50 % RH conditions without loss of

performance. Due to their excellent properties, a wide variety of metal oxides

have been exploited as additives to improve Nafion’s performance.”29

Metal nanoparticles

Metal nanoparticles can enhance Nafion’s performance in fuel cells.
Positively charged Pd nanoparticles were produced by the alcohol-based
reduction of metallic ions with poly(diallyldimethylammonium chloride) (PDDA) as
the surfactant.30 Using the electrostatic interactions between the positively
charged Pd nanoparticles and the negatively charged sulfonic acid groups of
Nafion ionomers, multi-layers of Pd nanoparticles and Nafion ionomers were self-
assembledonto Nafion membranes, generating an effective methanol-blocking
proton exchange membrane (Figure 6). Self-assembly of 3 double-layers of Pd
nanoparticles and Nafion ionomers provided a composite membrane with 99 %
less methanol crossover and a slight decrease in proton conductivity compared

to the original Nafion membrane.

   
 

 

 

 

 

 

 

Pd particle

_¢,, —8 /
E w . Keep in Pd E W - Keep in Nafion E
3 solutions 3 ionomersolutions 3 Nafion
:3 -—-cn 3 —U’ m 3 ionomer
a 5 O 3
w w , s w I g
a a v
_‘ __m -I —'U) 'I
N S o 3
a t?- o 9’ I m

_m --C0

0 _ ,2 _

I
E

Figure 6. Schematic illustration of multi-layer self-assembly of Pd nanoparticles
and Nafion ionomers onto Nafion membranes.30

Au nanoparticles can also improve Nafion's performance in direct
methanol fuel cells. The self-assembly of ~4 nm charged Au nanoparticles onto
Nafion 212 membrane surfaces formed a monolayer structure and drastically
reduced the methanol crossover from 168 to 18 mA/cm2 under the conditions of
2 mol/L methanol and 60 °C.31 The self-assembly of Pt and other metal
nanoparticles onto the surface of Nafion also has been reported,32 however,

these methods usually increase cell resistance and reduce proton conductivity.

Self-humidifying materials

The poor proton conductivity of Nafion at elevated temperatures is
attributed to vaporization of water leading to the lack of proton transport vehicles.
A class of inorganic-organic composite membranes, “self-humidifying" proton
exchange membranes, was developed by Uchida et al. to improve the high
temperature performance of Nafion.33'34 Generally, “self-humidifying" membranes

contain highly dispersed Pt nanoparticles, which catalyze recombination of

diffused H2 and O2 inside the membranes, generating water molecules that “wet"
the membranes. A hygroscopic metal oxide is included to adsorb the water
produced at the Pt particles and at the cathode. The resulting composite
membranes exhibited lower gas and methanol crossover as well as improved
water uptake compared to Nafion membranes. In addition, the ohmic resistance
of membranes decreased when operated with dry H2 and 02 at 80 °C and
ambient pressure.

Recently, Gao and coworkers developed a self-humidifying membrane
from an exfoliated Pt-clay/Nafion nanocomposite.35 Instead of using spherical
hygroscopic particles, which may have poor compatibility with the membrane
material and reduce membrane toughness, they used clay platelets to balance
the water content and enhance the mechanical properties of the composite
membrane. Monolayers of Pt nanoparticles (2-3 nm) were chemically anchored
to the exfoliated clay surfaces through chemical vapor deposition to stabilize Pt
nanoparticles (Figure 7). The composite membranes were prepared by simple
recasting a Nafion solution and Pt-clay. With dry H2 and 02, the power density of
the Pt-clay/Nafion composite membrane was 170 % higher than commercial

Nafion 112 of a similar thickness tested under the same conditions.

10

Metal-organic precursor

     

 

 

 

\ ’ Pt nanoparticles a
l
Vaporized c C a a a 0
organic modifier . ‘ I a
. . s’ I , ) cvo .,
. I; . U Vacuum ib—
Clay with organic modifier
Exfoliated clay
Calcinatioaneduction D 0
W 2—I
5 wt% Nafion ' e
— a I
. Electro e
Pt-clay casting Anode
Dry H2 8- —>

Figure 7. Schematic diagram of preparing exfoliated Pt-clay/Nafion composite
membrane. (Reprinted with permission from Langmuir 2008, 24, 2663-2670.
Copyright 2008 American Chemical Society)

Incorporation of inorganic materials in Nafion is a feasible and effective
method for improving membrane stability as well as reducing water swelling and
methanol crossover. However, improving these properties usually comes at the
cost of decreased proton conductivity. More importantly, the cost of Nafion
remains a barrier to wider application of Nafion-containing composite membranes.
Nafion’s cost and its poor performance at high temperature has stimulated
research into alternatives to Nafion.

The most widely studied. alternatives to Nafion are hydrocarbon

15,36-57

polyelectrolytes, such as sulfonated polystyrene, sulfonated polyimides,5M4

14.85-121 122-142

sulfonated poly(ether ether ketone)s, and sulfonated polysulfone.
The proton conductivity of PEMs is a function of both the chemical structure of

the polymeric materials and the membrane morphology. Therefore, most

11

research is aimed at generating inexpensive, chemically and mechanically stable
proton conducting membranes from existing polymer materials instead of

synthesizing new polyelectrolytes.

Block copolymers as an approach to alternatives to Nafion membranes
Nafion’s proton conducting mechanism suggests that developing ionic
channels inside membranes by microphase or nanophase separation of
hydrophilic and hydrophobic segments would be the key to PEMs with high
proton conductivity. Designing well-defined block copolymers with incompatible
blocks is a useful approach to achieve microphase-separated structures, such as
cylinders, lamellar, and spheres.”3 Extensive research effort has been devoted
to PEMs fabricated from block copolymer systems, which have proton conducting
hydrophilic blocks and mechanically stable hydrophobic blocks. These materials
provide an insight into the relationship between polymer structure, morphology
and proton conductivity?“""8'134'144’156 ln block copolymer-based PEMs, phase-
separated morphologies are common; the ionic phase is composed of hydrophilic
segments that support proton conductivity while the insulating hydrophobic phase

affords dimensional stability.

Sulfonated block co-polyimides PEM
Polyimides are often described as high-performance materials due to their
excellent thermal, chemical, and mechanical stability, good film-forming ability,

and low gas permeability. Sulfonated block co-polyimides are widely studied as

12

1,82,83,114,149,157,158 Usually, the synthesis of

alternatives to Nafion membranes.
sulfonated block co-polyimides involves a two-pot procedure as shown in
Scheme 1.83 Fluorinated block co-polyimides were synthesized with different
block lengths and diamine compositions, and were found to have proton
conductivities that depended on the block lengths. The conductivities of
sulfonated block copolymers increased with the block length, which was
explained as the influence of block lengths on the distribution of ionic channels.
Furthermore, the conductivities of sulfonated block co-polyimides were higher
than the corresponding random copolymer membranes, which was ascribed to

ionic channels formed by phase separation of hydrophilic and hydrophobic

sequences in block co-polyimides.

13

Scheme 1. Synthesis of sulfonated block co-polyimides.83

H03 0 6 O F 0 6 o
”sznmma) WWW..-”
I I
0 H033 .
. Q H00H HOO% Q . +H F3 Hoo<|3_l Q moo F3 H
2W 9 .0

 

 

 

Sulfonated block copoly(amic acid)

IChemical imidization

I‘m

Sulfonated block copolyimide

Sulfonated block co-poly(arylene ether sulfone) PEM

Another extensively explored block copolymer system is sulfonated block
co-poly(arylene ether sulfone)s. These polymers are one of the most promising
low-cost alternatives to Nafion membranes. Like polyimides, poly(arylene ether
sulfone)s have excellent thermal, chemical, and mechanical stability, superior
hydrolytic stability, and good processability. PEMs solvent-cast from hydrophilic-
hydrophobic poly(arylene ether sulfone) block copolymers (Scheme 2) showed
increases in proton conductivity and water uptake as the block length is

increased.1"""'156 Using shorter blocks resulted in reduced water uptake, and

14

lower proton and water transport, comparable to random copolymers. However,
membranes produced with long block lengths exhibited higher proton
conductivities than random copolymers with similar ion exchange capacity (lEC)
values under partially hydrated conditions, which was ascribed to the unique
microstructure and phase-separated morphology of hydrophilic-hydrophobic
block copolymers. Water analysis supported different microstructures and
morphologies for the block copolymer and random copolymer: phase-separated
morphologies in block copolymers but closed and isolated morphological
structures in random copolymers, despite having similar chemical structures and
chemical compositions. This implies that proton transport is controlled by not only

the acidity, but also the morphology of the membrane system.

15

Scheme 2. Synthesis of segmented sulfonated multiblock co-poly(arylene ether
sulfone) with different linkage composition. 56

 

 

SO3Na 0
0.3-3.8 30.; + HOMOH Cl-Q-fi-Q—CHHMOH
NaO I DMAc I DMAc
Toluene Toluene
K2003 K2C03
03
K409330§0omox K4+M§04e+x
F DMAc
X Cyclohexane
F F o F
F 0- 000,011-00- 0x
DMAc V
Cyclohexane
K2003

050E300 om§0lI

Sulfonated poly(arylene ether sulfone)-b-polyimide PEM

In addition to block co-polyimides or block co-poly(arylene ether sulfone)s,
PEMs were also fabricated from more complicated block copolymer architectures.
As shown in Figure 8, PEMs based on a bisulfonated poly(arylene ether
sulfone)—b—polyimide (BPSH-b—Pl) copolymer exhibit well-defined hydrophilic and
hydrophobic domains, and the block copolymer morphologies correlate to the

block lengths (Figure 9).149 Tapping mode AFM phase images show that the

16

dark ionic regime and the bright nonionic regime increased when the block
lengths increased, and eventually forming a long-range bicontinuous lamellar
microstructure. In addition to the well-defined phase-separated morphology, the
hydrolytic stability of the block copolymer in 80 °C water was greater than

random polyimide copolymers.

II gewmloleewmli
inter = 391%“

Figure 8. Structure of segmented bisulfonated poly(arylene ether sulfone)-b-
polyimide (BPSH-b—Pl) copolymer.

17

 

Figure 9. Tapping mode AFM phase images of BPSH-b—Pl copolymers with
different block lengths. (both block lengths increase from top to bottom)
(Reprinted with permission from J. Polym. Sci. Pol. Chem. 2007, 45, 4879-4890.
Copyright 2007 Wiley InterScience)

Sulfonated polysulfone-b—poly(vinylidene fluoride) PEM

Poly(vinylidene fluoride) (PVDF) is chemically and mechanically stable
and is commonly used as the hydrophobic component of block copolymer-based
PEMs. For example, Holdcroft et al. synthesized sulfonated polysulfone-b-
poly(vinylidene fluoride) (PSf-b-PVDF) copolymers from the polycondensation of
an q,uI-dihydroxybisphenol A and d,w-dibromo poly(vinylidene fluoride) as
illustrated in Scheme 3.123 After sulfonation, the copolymers were cast into PEMs
to test their performance. Despite their relatively low ion exchange capacity, the
sulfonated PSf-b-PVDF copolymers exhibited higher proton conductivities than
sulfonated polysulfones, which was attributed to greater phase separation in the
block copolymer system, as has been proposed for other block copolymer

systems.

19

Scheme 3. Synthesis of Bisphenol A PSf-b-PVDF copolymers.123

H O
(Willow—(jail + MTHCM
H3 0

1) DMAc/Toluene, 150-160 °C
K2003, Refluxing, 15 h

2) Excess bisphenol A, 150-160 °C
K2CO3, Refluxing, 5 h

HOIQEQOQEQOIHQIbm

H3
+
Br-CF2-CF2-I-CH2-CF2'I1-Br
I,NaH rt, 6h

H3

-CF -CF H -cr=
00 040334 2 2+cz 2 ,
H3 n

m=4, 11,21

Sulfonated poly(styrenesulfonate)-b-poly(methylbutylene) PEM

In addition to high performance polyelectrolyte systems, similar phase-
separated structures have been developed in block copolymer systems
containing polystyrene. For example, PEMs obtained from
poly(styrenesulfonate)-bIock-poly(methylbutylene) (PSS-b—PMB) copolymers
spontaneously formed hydrophilic PSS and hydrophobic PMB domains, which
resulted in microphase separation.159 The water uptake of the PEMs, which had
a hydrophilic domain width ranging from 2 to 5 nm, increased as the surrounding

air temperature increased from room temperature to 90 °C at a constant relative

20

humidity, whereas Nafion 117 became drier under the same conditions.
Consistently, the proton conductivity of these PEMs increased with increasing
temperatures. This unexpected behavior might be due to capillary condensation,
which suppressed solvent evaporation at elevated temperatures. This interesting
discovery might enable high temperature PEM fuel cells.

Block copolymers also show better performance at low hydration levels
compared to homopolymers and random copolymers, which results from ionic
channels formed upon microphase separation of the hydrophilic and the
hydrophobic blocks. However, obtaining high proton conductivity and mechanical
stability in block copolymers usually correlates with difficulty in optimizing the
physiochemical properties of individual blocks and increased membrane
synthesis and production costs. In addition, some block copolymers suffer from

severe swelling and poor mechanical strength problems.160

Polymer blends as an approach to alternatives to Nafion membranes
Polymer blending is a useful route to Nafion alternative membranes.
Instead of integrating the hydrophilic segment (proton conductivity) and the
hydrophobic component (mechanical strength) in a single polymer structure,
blending a polyelectrolyte with a non-conductive hydrophobic material has
become a popular route to PEMs.‘5°'61'1°“““51"70 Blending is usually based on
commercially available and inexpensive polymer precursors and thus it is a
simple and cost-effective route to PEMs. Compared to block copolymer-based

PEMs, the blending concept eliminates the need to optimize the mechanical,

21

chemical and physical properties in a single polymer chain and it has the
potential to develop a wide variety of PEMs. A range of membrane properties,
such as proton conductivity, water uptake, and mechanical strength, can be
easily tailored by (1) adjusting the relative ratio of two blending materials, (2)
varying the ion exchange capacity of the hydrophilic component, or (3) modifying
the backbones of the blend components.

Polymer blends have been extensively studied and have had some
success as Nafion alternative membranes. In order to form stable membranes,
research attention needs to be directed to the different kinds of interactions

between the blending materials, as shown in Figure 10.170

van der Waals interaction I—CH3 ---- H30-I

dipole-dipole interaction It?

0:?

electrostatic interaction I-88 H fi_I
(ionic cross-linking) 3 3

H
hydrogen bonding I—O,‘ [b—I
H
H 4
covalent cross-linking -N
I
H V

Figure 10. Specific interactions between blending materials.170

mfiuans puoq Bugseeioul

 

 

22

Van der Waals/dipole-dipole interaction

PVDF has been widely studied as the hydrophobic material in blend
membranes.“9'1°8'171'173 The large dipole moment associated with the CF2 moiety
favors mixing with polar polymers such as poly(methyl methacrylate) (PMMA). A
clever blending example is the mixing of sulfonated poly(styrene-b-methyl
methacrylate) (PSSA-b—PMMA) and PVDF; PMMA is miscible with PVDF but
PSSA is immiscible. Since the blend membranes were stained with mm, the
dark regimes represent the hydrophilic component (PSSA) while the brighter
areas correspond to the hydrophobic PMMA and PVDF (Figure 11).49 The
membrane proton conductivity increased with the degree of sulfonatiOn in PSSA-
b-PMMA and also was enhanced by a well-ordered phase-separated morphology.
The membranes in Figure 11a and Figure 11b had the same degree of
sulfonation but the membrane shown in Figure 11a had smaller domains due to
a lower PVDF content. Not surprisingly, it also had a higher proton conductivity.
Interestingly, a comparison between the two membranes in Figure 11a and
Figure 11c, which had the same PVDF content but different sulfonation levels,
shows that the hydrophilic domain increased as the sulfonation level increased
due to the aggregation of more sulfonic acid groups. The membrane in Figure
11c formed the best-ordered phase-separated microstructure and exhibited the
highest proton conductivity. The morphological structure of the blend membrane

was easily controlled by varying the blend ratio of the two blending materials.

23

 

Figure 11. TEM images of blend membranes with different degrees of
sulfonation and blend ratios: (a) 824F22; (b) $24F41; and (c) 839F22. (SxxFyy,
where xx represents the sulfonation level and yy stands for the content of PVDF
in the blend membrane) (Reprinted with permission from J. Power Sources 2009,

188, 127-131. Copyright 2009 Elsevier)

24

Hydrogen bonding interaction

Blend membranes from sulfonated poly(ether ketone ketone) (SPEKK),
the proton-conducting component, and poly(ether sulfone) (PES) resolved the
swelling problem of SPEKK to some extent despite the immiscibility of SPEKK
with I=>Es.169 Blending PES with the highly sulfonated SPEKK provided robust
membranes. In this example, the membrane morphology can be controlled by
balancing the polymer composition, mixing ratios, solution casting solvent and
temperature. At certain sulfonation levels of SPEKK, the blend membrane phase
separated into co-continuous morphologies. Similar improved properties were
seen in blends of sulfonated polyimide (SPI) and PES.60 The addition of PES
reduced the excessive swelling of SPI membranes and greatly increased both
the stability and tensile strength of the SPIs membrane.

Hong and coworkers proposed a convenient way to fabricate PEMs from
blends of polystyrene (PS), poly(ethylene oxide) (PEG), and a fine powder of
poly(styrene sulfonic acid) (PSSA) powder.174 In the blend membrane, PSSA
acted as a proton source and as a compatibilizer, a polymer that improves the
miscibility between the two polymers, in this case, hydrophilic PEG and
hydrophobic PS (Figure 12a). PS provided the mechanical strength of the blend
membrane while PEO conveyed protons to the cathode under an electric field as
illustrated in Figure 12b. When the PEG and PS mix well, the hydrophobic PS
component limits water evaporation at elevated temperatures which may be

relevant to designs of PEMs for application at elevated temperatures.

25

 

PSSA

Figure 12. (a) Schematic view of the polymer blend. (b) A schematic diagram of
the formation of a H-bonding complex between the sulfonic acid group and PEG
units.174

Electrostatic interaction (ionic cross-linking)

Ionic cross-linking in blend membranes utilizes acid-base interactions
such as those between sulfonic acid groups and the nitrogen atoms of imidazole
groups. These materials have been explored in direct methanol fuel
cells.85'"°'"9'175 Blending sulfonated poly(ether ether ketone) (SPEEK) with
polysulfone-4-nitrobenzimidazole (PSf-NBlm, Figure 13) formed membranes
with lower methanol crossover than SPEEK and Nafion 115 membranes in direct
methanol fuel cells.85 The size of the ionic clusters formed from the sulfonic acid
groups of SPEEK increased in the blend membrane, owing to the addition of the
polar nitrobenzimidazole side groups to ionic clusters. These membranes had
reduced uptake of water and methanol/water solutions. Membranes with 2.5 and
5.0 wt% PSf-NBlm contents had higher proton conductivities than SPEEK, and
the imidazole groups were thought to facilitate proton conduction from one
sulfonate to another (Figure 14). Imidazoles can act as a proton conducting
solvent and replace the acid/water complex in conventional PEMs, making high

proton conductivity under anhydrous conditions possible.

26

No2

Figure 13. Structure of polysulfone bearing 4-nitrobenzimidazole side groups.85

@ Polysulfone backbone
'10 4313?...211521... 6-1%131‘31

N02 N02 N02

Figure 14. Schematic view of the proton transfer mechanism with electrostatic
interaction between SPEEK/PSf-NBIm blending materials.85
Covalent cross-linking

One disadvantage of ionically cross-linked blend membranes from PSf
and polybases is that the electrostatic interaction breaks down in aqueous
environment at elevated temperatures (T>70-90 °C), leading to brittle
membranes.17° Cross-linking via covalent bonds is more robust and has played
an important role in improving membrane performance."6"76'"7 For example, the
thermally activated condensation of SPEEK and polyols bridges neighboring
chains (Figure 15). Cross-linking improved the stability and mechanical strength
of SPEEK membranes and suppressed excessive water uptake. Though cross-

linking reduced the number of available sulfonic acid groups for proton

27

conduction, the blend membrane showed only a slight reduction in proton

conductivity compared to a pure SPEEK membrane.

+00 0 =53“

803H
H0. H2
+ (H0— H)x (no-81),, + 2H2O
HO" H2 802 _0- H2

i033” got a ‘61...

Figure 15. Illustration of possible SPEEK cross-linking reaction.177

Both polymer blends and block copolymers have had some success as
alternatives to Nafion membranes. In block copolymers, the hydrophilic and
hydrophobic components are part of a single polymer structure, which lessens
concerns of miscibility and the interactions of different chemical blocks. Blend
membranes distribute the functions of proton transport and mechanical stability
into different materials, which cannot be realized in homopolymer based
membranes. Despite significant improvement in fuel cell membrane performance,
controlling membrane microstructure has been mostly ovenooked. It is known
that the phase-separated morphology of Nafion is critical for its success as a fuel
cell membrane, and controlling the microstructure in alternative membranes is

desirable. For example, with no control over the alignment of proton conductive

28

polymers, blend membranes are generally associated with improved mechanical

properties but with a sacrifice of proton conductivity.

Controlling the ionic channels of PEMs

Recently, Gao et al. utilized self-assembly of surface-charged latex
particles to form ionic channels for proton conduction.178 In their work, the latex
particles were synthesized by aqueous emulsion polymerization of two
hydrophobic monomers, butyl acrylate and methyl methacrylate, a cross-linker,
N,N’-methylene-bis-(acrylamide), and a charged monomer, sodium styrene
sulfonate, resulting in particles with charged surfaces. Membranes were
developed by solvent casting an aqueous dispersion of cross-linked particles,
followed by acidification, which provided particles with hydrophobic cores and
proton-conducting surfaces. Continuous ionic channels naturally developed upon
annealing, resulting in no or a very low percolation threshold (Figure 16). In
these membranes, the charged particles were not physically packed as in
amorphous membranes, but instead they adhered to each other by either
intersphere cross-linking and/or entangling chains. Therefore, the resulting
membranes had significantly higher proton conductivities than amorphous
membranes with the same charge content. However, the membranes swelled
excessively reducing their mechanical stability. A better approach is needed to

solve the swelling problem while minimizing the percolation threshold.

29

 

Figure 16. Schematic illustration of precursor particles with hydrophobic cores
(grey) and ionic surface (dark) and a film fabricated by self-assembly of the
particles. (Reprinted with permission from Macromolecules 2005, 38, 5854-5856.
Copyright 2005 American Chemical Society)
Electric field aligned blend membranes

Gao’s model suggests that aligning proton conductive particles in a
suitable hydrophobic matrix might lower the percolation threshold and solve the
swelling problem. Particle alignment can be accomplished by applying an
external driving force while curing membranes, such as an electric or magnetic
field. Oren et al. first utilized this principle to align ion exchange particles in a two-
component RTV silicon rubber matrix by the application of an AC electric field
during the fabrication of ion-exchange membranes.”9 As shown in Figure 17a,
particles dispersed randomly in the absence of an electrical field. With the
application of the electric field, long chain-like aggregates aligned along the field
(Figure 17b). The alignment shifted the percolation threshold for ionic conduction,

from 40 % particle content for membranes with randomly distributed particles to

10 % for membranes with ordered particles. For a given ionic conductivity,

3O

membranes with aligned particles required fewer particles than conventional
membranes with randomly dispersed particles. The benefits of a low particle

content include reduced water uptake and increased mechanical stability.

 

Figure 17. Light microscopy photographs of (a) non-ordered; and (b) ordered
ion-exchange particles in the polymeric component of the rubber. 100X, electric
field: 1330V/cm, 50 Hz. (Reprinted with permission from J. Membr. Sci. 2004,
239, 17-26. Copyright 2004 Elsevier)

Oren et al. successfully reduced the percolation threshold, but the
conductivity mechanisms for ions and protons may be different. Shaw et al.
extended Oren’s work to proton conductivity, and investigated the effects of the
electric field amplitude and frequency on the alignment of sulfonated crosslinked
polystyrene ion exchange resin (acid form) in a poly(dimethylsiloxane) matrix
perpendicularly to the membrane plane.180 As illustrated in Figure 18, a
membrane with 5 wt % particles showed no alignment in AC fields <150 V/mm at
5 Hz; alignment only occurred when the electric field exceeded a certain
threshold. Membranes with aligned particles had higher proton conductivities and

less water uptake than membranes with unaligned particles.

31

 

Figure 18. Optical micrographs of the alignment of sulfonated crosslinked
polystyrene ion exchange resin in a poly(dimethylsiloxane) matrix under AC
electric field of different magnitudes at 5 Hz frequency. (Reprinted with
permission from ANTEC 2005, 1731-1735. Copyright 2005 Society of Plastics
Engineers)

Later, the same group extended the electric field alignment of
polyelectrolytes to a blend of SPEKK and a poly(ether imide) (PEI), which

resulted in a thousand-fold Increase in conductivity at low humidity due to the

formation of aligned proton conducting channels in SPEKK.101

Magnetic field aligned blend membranes
Although the application of electric fields has successfully aligned

polyelectrolytes while casting blend membranes, there are limitations to this

32

technique. There must be sufficient dielectric contrast between polyelectrolyte
and the matrix, and the matrix must sustain a high field and have sufficient
mobility. These restrictions have hindered the broad application of this
processing technique due to a limited choice of matrix materials. An alternative
technique, using magnetic fields as the external driving force, is more suitable for
polyelectrolyte alignment in blend membranes, as most polymers are anti-
ferromagetic and hence are good matrix materials. Unlike electric fields,
magnetic fields are not restricted in terms of the applied field strength, but only
require that the polyelectrolytes have sufficient magnetization to respond to a
magnetic field.

Recently, the Shaw group synthesized proton conducting magnetic
nanoparticles via the emulsion polymerization of styrene, divinylbenzene, and
sodium 4-styrenesulfonate (NaSS) in the presence of a y-Fe203 dispersion.
Subsequent acidification provided the sulfonic acid, as demonstrated in Figure
19.181 During the membrane casting process, an external magnetic field was
used to align proton conductive magnetic nanoparticles in SPEKK. Successful
alignment of the magnetic nanoparticles can be seen in Figure 20b, while
processing the membrane without a magnetic field resulted in randomly
dispersed particles, as illustrated in Figure 20a. Not surprisingly, the conductivity
of membranes with aligned particles was twice that of unaligned controls.
However, the proton conductivity of the blend membrane was much lower than a
standard SPEKK membrane, which was ascribed to the decreased lEC of the

blend membrane.

33

O = monomer droplet/polymer
O = Myristic acid-treated v-F0203

 

 

 

 

 

 

303'Ila Increasing time of reaction
/sosua
Moo 8
3 x 8038.3
Mao; —> —>
k303Na
$03M SO3Na
Water + NaSS + Initiator A,

Time = 20 min

6h

Figure 19. Schematic diagram of the proposed mechanism for the formation of
magnetic particles.

 

, ‘ «I» .
Iv 5.“: . . 4.. - I? 49:. V

    

 

    

v

.A’ ,.‘, ,

4 a“
I

1 .1 ’.I I
v \f‘" . ' ‘ , w
A . I - ‘ . ‘F
c». . 1 1 mm 2. .- ~ '1. 4.444
s. "-o . '1'." ;-..-_ .. --- . p .. .
.‘- ‘ ... m '. '. .~ <A Kanqfir. ‘42 v
.. . . .- .. “W. a ......» 42.
~ six-”q. -< “‘3 1‘ "fir w). . :. MR. __ \ '
.5". ."' .-..\ “*u.. " . "‘ '

Figure 20. Images of SPEEK membranes with magnetic nanoparticles (a)
randomly dispersed; and (b) aligned in magnetic field. (Reprinted with permission
from J. Membr. Sci. 2007, 303, 64-71. Copyright 2007 Elsevier)

While Shaw et al. showed that magnetic fields can be used to align
magnetic nanoparticles in a SPEEK matrix, there are limitations to Shaw’s

approach. The y-Fe203 particles were confined to the surfaces of composite

35

particles instead of being encapsulated by the polymer, making y-Fe203 particles
vulnerable to the strong acidic environment of PEMs (y-Fe203 particles dissociate
in acid). Secondly, the magnetic particles were aligned in the plane of the
membrane, while real PEMs require through-plane proton conducting paths.
Finally, both the magnetic particles and the matrix (SPEEK) are hydrophilic,
which would likely result in excessive swelling and low mechanical strength
membranes. Nevertheless, using magnetic fields to align conductive paths may

be a useful technique for improving blend membrane performance.

36

Chapter 2 Direct Polymerization of Sodium 4-Styrenesulfonate from Silica
Nanoparticle Surface:

Kinetics, Characterization, and Application in Fuel Cells

Introduction

Ionic transport in polymers primarily depends on the number of charge
carriers and their mobility. Practical polymer electrolytes are required to provide
high ionic conductivity and chemical and thermal stability, while maintaining
sufficient mechanical properties to serve as an electrode separator. In the case
of lithium ion systems, conductivity is correlated with the segmental mobility of
the polymer host, and therefore polymers with low glass transition temperatures
(T9) are ideal hosts for electrolytes."’2'183 However, low Tg materials have poor
mechanical properties and must be used in conjunction with an external
electrode separator. Similar challenges are encountered with membranes for fuel
cell applications. For example, highly sulfonated polymers have good proton
conductivities, but tend to swell excessively.2

One solution to the challenge of synthesizing polymers that
simultaneously have high conductivity and good mechanical properties is to
design electrolytes from polymers that spontaneously separate into ion-
conducting and hydrophobic phases. To ensure high conductivity, the phases
need to be bicontinuous while maintaining good mechanical properties. There

151,184

are many variants of this approach, including block copolymers, polymer

blends,108 and various nanoparticles dispersed in polymer hosts.”31

37

The phase-separated polymers approach is particularly useful in the
design of proton exchange membranes (PEMs) for fuel cells. Structural studies
of hydrated Nafion, the prototype PEM, show a continuous network of proton
conducting channels stabilized by the perfluorocarbon backbone of the
polymer.“'11 Recent alternative membranes, designed to overcome Nafion’s
limited temperature range (<90 °C), mimic the Nafion structure by incorporating
both hydrophobic segments and highly sulfonated blocks to effect phase
separation on the nano—scale.134'148'185

A simpler approach to bicontinuous electrolytes is. to blend a
polyelectrolyte material with a hydrophobic non-conductive material.6°'61'1°1'119‘6"
‘69 The blending concept eliminates the need to optimize the proton transport
and physical properties in a single polymer chain. Blending is usually based on
commercially available and inexpensive precursor polymers, and therefore
blending can be a simple and cost-effective route to PEMs. A popular
hydrophobic material in blend membranes is poly(vinylidene fluoride) (PVDF),
which is easy to process into membranes, commercially available, and
chemically and mechanically stable.49'1°8'171‘173 A widely studied polyelectrolyte
material for PEMs is poly(4-styrenesulfonic acid) (PSSH) because of its high
sulfonation level, low cost, and the simplicity of synthesis.”41

Blends also have limitations. They must be available with a stable
bicontinuous morphology, and the polymers that comprise the conductive phase
cannot leach from the PEM. Leaching can be inhibited by binding polymers to

nanoparticles, in effect, increasing their polymer molecular weight and reducing

38

their mobility. Core-shell particles coated with proton-conducting polyelectrolytes

186,187 and

can take advantage of both the inorganic and organic functional groups
are good candidates for the hydrophilic component in blend membranes.

There are many methods that generate polymer shells from inorganic

"188,189 ”190,191

nanoparticle cores, such as the “grafting from and “grafting to
techniques, and electrostatic self-assembly.192 Among these, atom transfer
radical polymerization (ATRP) has been the primary methodology for growing
polymers of precise molar mass, architecture, and functionality from inorganic
nanoparticles.186'183'193'200 ATRP has been applied to a rich variety of monomers
to prepare polyelectrolyte end chain-immobilized nanoparticles.

In this report, we test the utility of blends comprised of functionalized
nanoparticles in a hydrophobic matrix (PVDF). Silica particles were modified with
a thick layer of PSSH via ATRP. These particles can be synthesized by two
methods: grow polystyrene from silica particles and then sulfonate the
polystyrene, or directly polymerize sodium 4-styrenesulfonate onto silica
particles199 followed by acidification. The “post-sulfonation” approach usually
results in low degrees of sulfonation, cross-linking, and polymer chain
degradation,2 while direct polymerization of sodium 4-styrenesulfonate places
one sulfonic acid on each monomer, occurs at ambient temperature, and has no
significant side reactions.199

Herein, we report the synthesis and characterization of hybrid

nanoparticles (silica-PSSNa) comprised of ~7 nm diameter silica cores and a

poly(sodium 4-styrenesulfonate) (PSSNa) shell. After acidification, the hybrid

39

nanoparticles were blended with PVDF, and composite membranes were
prepared using a solution-casting method. The as-synthesized membranes had
proton conductivities comparable to, or higher than Nafion 117 membranes.
Compared to conventional PSSH blend membranes, the advantage of this
strategy is 100 °/o degree of sulfonation owing to the presence of one sulfonate
group per polymer repeating unit. In addition, despite the high degree of
sulfonation, PSSH-coated silica nanoparticles do not leach from the membrane

when hydrated.

Results and discussion

Surface-initiated ATRP of sodium 4-styrenesulfonate

The synthetic route to hybrid poly(4-styrenesulfonic acid)/silica particles is
shown in Scheme 4. Colloidal silica (Snowtex-XS, Nissan Chemical Corporation)
was received in the form of an aqueous dispersion in pH 9-10 water. Ion
exchange with cetyltrimethylammonium bromide (CTAB) rendered the particles
hydrophobic, causing them to aggregate.201 Figure 21 shows IR spectra for the
Snowtex-XS particles and their derivatives. Bands at 3000-2800 cm'1 confirmed
the presence of the CTAB ion on the particle surface. Treating the particles with
(11-(2-bromo-2-methyl)propionyloxy) undecyltrichlorosilane immobilized an a-
bromodimethyl ester, an ATRP initiator, to the silica nanoparticle surface. The
appearance of a band at 1730 cm‘1 (C=O), a broad C-O stretching band at 1200-
1000 cm“, and further development of the C-H bands at 3000-2800 cm'1 are

consistent with the successful anchoring of the initiator to the nanoparticle

4O

surface. Finally, PSSNa was grown from the nanoparticle surface at room
temperature using ATRP (water/methanol (v/v = 3/1), [M]o:[surface
initiator]:[CuBr]:[CuBr2]:[bpy] = 70:1:2:0.2:6) to yield hybrid poly(sodium 4-
styrenesulfonate)/silica (silica-PSSNa) nanoparticles. The particles were
collected by centrifugation followed by further purification and no PSSNa polymer
was found in the reaction solution implying that the polymerization was restricted
to the particle surfaces. The IR spectrum of silica-PSSNa-3 (Table 1) is
consistent with a substantial shell of PSSNa grown from the particles: new bands
appeared for sp2 C-H stretching (3200-3000 cm"), a sharp in plane bending peak
at 1010 cm‘1 and a pattern of combination and overtones (2000-1800 cm")
consistent with a 1,4-substituted benzene ring, and most prominently, va and vs
bands for the sulfonate group (S=O, 1200 and 1040 cm"). Acidification with 1M

HCI converted the sodium salt to the corresponding sulfonic acid.

41

Scheme 4. Synthetic route to hybrid poly(4-styrenesulfonic acid)/silica
nanoparticles.

CI
l_ O
CI-ir(CH2)1To-$‘_
I Br
———>

Toluene, 65 °C

//—©—803Na

CuBr, CuBr2, bpy
H20, CH3OH, rt

 

   

silica silica-initiator

\ . 0
- = :st—(CH2);1-o-$(;I
Br
II

n n
1MHm
= l ,.
03H O3Na

silica-PSSH silica-PSSNa

 

 

 

 

 

   

42

 

Absorbance
O’

 

 

 

4000 3500 3000 25.00 2000 1 500 1000 500
Wavenumber (cm'1)

Figure 21. FTIR spectra of (a) Snowtex-XS nanoparticles, CTAB salt;

(b) nanoparticles with surface anchored initiators; and (c) silica/PSSNa-3
nanoparticles. Samples were made into pellets with KBr.

Thermal gravimetric analysis (TGA) results are consistent with the IR data.

As shown in Figure 22, CTAB modified Snowtex-XS nanoparticles have an onset

for weight loss at ~230 °C, and a final weight loss of ~13 % at 900 °C, which

corresponds to decomposition of CTAB, and loss of water due to condensation of

surface silanols. The silica-initiator particles also showed an onset for weight loss

at ~230 °C, Their 27 % weight loss at 900 °C is consistent with thermal

degradation of surface initiator moiety.

43

100

 

 

 

 

 

87.6%
80-

?: 73.2%

C

.9

#

cc)

5 60 -

L

g) 54.5%

g 47.3%
40 _ 39.4%

37.0%

20 .

I ' I I j I 1 l
200 400 600 800 1000
Temperature (°C)

Figure 22. Thermogravimetric analysis of (a) Snowtex-XS nanoparticles, CTAB
salt; (b) silica—initiator nanoparticles; (c) silica-PSSNa-1 nanoparticles; (d) silica-
PSSNa-2 nanoparticles; (e) silica-PSSNa-3 nanoparticles; and (f) linear PSSNa.
See Table 1 for details of silica-PSSNa-(1-3). All samples were run in air at a
heating rate of 10°lmin.

To synthesize particles with predictable surface polymer molecular
weights, we characterized the kinetics for surface-initiated ATRP of sodium 4-
styrenesulfonate. Since tracking the course of the polymerization was
complicated by the poor solubility of silica-PSSNa particles in the polymerization

solution (water and methanol), we chose to follow monomer consumption by 1H

44

NMR, using sodium p-toluenesulfonate as an internal standard because of its
similar structure and solubility, and its distinct proton chemical shifts. Monomer
consumption was calculated from the integration of the 1H NMR resonances for
the methyl group of sodium p-toluenesulfonate (2.27 ppm) and the vinyl proton at

5.3 ppm in sodium 4-styrenesulfonate (Figure 23).

a
e H at
H

H \ c H3
b d bi
d e c'

3N8 3N8
a
b a'

 

 

 

 

 

 

 

 

 

If] I I I I l I r I I I I I I I I I I I I' I I I I T l I I T I I I

7 6 5 4 3 2
ppm

Figure 23. 500 MHz 1H NMR spectra of a representative polymerization sample
for the kinetic study.

The plot of ln([M]o/[M]) as a function of polymerization time is nearly linear
up to 69 % conversion (240 min) consistent with a surface polymerization first-
order in monomer concentration (Figure 24). At higher conversions, curvature in
the polymerization data indicates a decreasing polymerization rate, likely due to

fewer growing chains as a result of termination (Equilibrium effects might also

45

affect the polymerization, but [M]eq is not known for sodium 4-styrenesulfonate).
We were pleased that the silica nanoparticles did not have a negative effect on
the surface polymerization. Samples taken at t = 20 min (conversion = 10 %,
silica-PSSNa-1), t= 1 h (26 %, silica-PSSNa-2), and t= 4 h (69 %, silica-PSSNa-

3) were isolated by centrifugation and purified for further characterization.

 

2.0 -

1.6- I

1.2-I

0.8 -

|n([M]o/[M])

0.4 -

 

00 -Il ' I ' r ' I ' I '
0 100 200 300 400 500
Time (min)

 

Figure 24. Surface polymerization kinetics of sodium 4-styrenesulfonate from
silica nanoparticles. The polymerization was run at room temperature in
water/methanol (v/v = 3/1) and an initial monomer concentration ([M]o) of sodium
4-styrenesulfonate is 0.5 mM. The ratios of [M]o:[standard]:[surface initiator]:
[CuBr]:[CuBr2]:[bpy] were approximately 70:10:1:2:O.2:6.

Three samples: silica-PSSNa-l, silica-PSSNa-Z, and silica-PSSNa-3,
which were prepared by polymerizing sodium 4-styrene sulfonate from silica-

initiator surfaces for increasing reaction times, were characterized by TGA

(Figure 22). Their weight loss at 900 °C increased from 45 % (silica-PSSNa-1) to

46

53 % (silica-PSSNa-Z) to 61 % (silica-PSSNa-3), indicating that the PSSNa
molecular weight increases with the polymerization time. The TGA trace of linear
PSSNa exhibited three mass loss events. While definite assignments are not
possible without corroboration by other techniques, the weight losses likely
correspond to elimination of residual water, desulfonation, and finally degradation
of the polymer backbone commencing at ~450 °C. The onset for degradation of
silica-PSSNa-1 was ~280 °C, much lower than silica-PSSNa-2 and silica-PSSNa-
3. The small mass loss, the shorter polymerization time, and a degradation onset
comparable to the silica-initiator particles all argue for silica-PSSNa-1 being
comprised of low molecular weight PSSNa chains. Interestingly, linear PSSNa
had a nonvolatile residue at 900 °C even in an oxidative atmosphere?02
Consistently with the NMR data, the TGA results indicate increasing
molecular weights with polymerization time. Using the residual mass of both the
silica-initiator nanoparticles and linear PSSNa at 900 oC, and the monomer
conversions determined by 1H NMR, the expected TGA weight loss at 900 °C
was calculated for silica-PSSNa-1, silica-PSSNa-2, and silica-PSSNa-3 (Table 1).
The data are in reasonable agreement with the actual TGA losses, ~4 % less
than expected. The result suggests that the amount of surface-grafted PSSNa
increased linearly with monomer conversion, polymerization in solution was
insignificant, the number of growing chains on particle surface was constant, and

the surface ATRP was a living/controlled reaction from I: 20 min to t= 4 h.

47

Table 1. Characteristics of particles isolated at different polymerization time.

 

 

 

 

Particles ATRP time Monomer Calculated Actual TGA
(min)a conversion TGA weight weight
(°/.)" retention at retention at
900 °c (%)° 900 °c (%)"
silica-PSSNa-1 20 10.4 58.1 54.5
silica-PSSNa-2 60 25.7 50.1 47.3
silica-PSSNa-3 240 69.4 43.3 39.5

 

 

 

 

 

 

 

a The polymerization was run at room temperature in water/methanol (v/v = 3/1)
and an initial monomer concentration ([M]o) of sodium 4-styrenesulfonate is 0.5
mM. The ratios of [M]o:[standard]:[surface initiator]: [CuBr]:[CuBr2]:[bpy] were
approximately 70:10:1:2:O.2:6.

determined by 1H NMR, using sodium p-toluenesulfonate as an internal
standard

° run in air at a heating rate of 10 °C/min

The CTAB-coated silica particles have a diameter of around 7 nm,
consistent with the data provided by the supplier. As shown in Figure 25a, the
dried particles agglomerate, but after surface polymerization, the cores of the.
silica-PSSNa nanoparticles are dispersed, consistent with polymerization of
PSSNa from the particle surfaces (Figure 25b). Imaging the polyelectrolyte

phase required staining with metal ions.

48

 

Figure 25. TEM images of (a) dried Snowtex-XS nanoparticles, CTAB salt; and
(b) silica-PSSNa-3 nanoparticles.

49

The hybrid nanoparticles which were blended with PVDF to make
composite membranes were named silica-PSSH nanoparticles and their
precursors were silica-PSSNa hybrid nanoparticles. As displayed in Figure 26,
the FT-IR spectrum of silica-PSSNa hybrid nanoparticles was dominated by the
characteristic absorption peaks of PSSNa with the strong O—H stretching bond at
3435 cm’1 from water absorbed by the hygroscopic surface poly(sodium 4-
styrenesulfonate). The absorption bands of silica-PSSH nanoparticles were
similar to the silica-PSSNa nanoparticles, except for the stronger and broader O-
H stretching peak, which is due to the strong hydrogen bonding between sulfonic

acid groups and absorbed water.

 

Absorbance

 

 

 

 

4000 3500 3000 2500 2000 1500 1000 500
Wavenumer (cm'I)

Figure 26. FT IR spectra of (a) silica-PSSNa nanoparticles; and (b) silica-PSSH
nanoparticles.

50

As shown in Figure 27, the weight loss of silica-PSSNa particles at 900 °C
was about ~58 %, which indicated a thick polymer layer on the surface of the
hybrid nanoparticles. A comparison of the TGA traces of silica-PSSNa and silica-
PSSH particles suggested that the free acid was less thermally stable than the
sodium salt in an oxidative atmosphere. The initial weight loss for silica-PSSNa
nanoparticles occurred at ~400 °C, while silica-PSSH degraded earlier and at
900 °C left a lower non-volatile residue than its sodium salt. The silica-PSSH
nanoparticles degraded in two steps: the first step from 200 to 400 °C should
correspond to desulfonation, while the second step was more likely the

degradation of the polystyrene backbone.

51

 

 

 

 

 

 

 

 

a
a 87.6%
80 -
g3 bfi. 73.2%
C
.9
E 60 -
9
9
’5, c
g 40 A 41.8%
‘ d
fi24.0%
20 -

 

 

 

I ' I I ' T
400 600 800 1 000

Temperature (°C)

I
200

Figure 27. Thermogravimetric analysis of (a) Snowtex-XS nanoparticles, CTAB
salt; (b) silica-initiator nanoparticles; (c) silica-PSSNa nanoparticles; and (d)
silica-PSSH nanoparticles.

Fabrication and characterization of composite PEMs from the blending of
silica-PSSH nanoparticles and PVDF

In conventional polystyrene based PEMs, the sulfonic acid groups are
introduced after polymerization by sulfonation reagents, such as chlorosulfonic
acid, concentrated sulfuric acid, or fuming sulfuric acid. However, most “post-
polymerization” sulfonation reactions are associated with low degrees of

sulfonation, cross-linking, and polymer degradation? Since it is widely accepted

52

that PEM conductivity is strongly related to the number of charge carriers (H‘),
the preference is the highest degree of sulfonation without sacrificing the PEMs’
mechanical properties. The direct polymerization of sulfonated styrene-like
monomers provides one sodium sulfonate group per repeating unit, 100 %
sulfonation.

Composite membranes were obtained by solution-casting a DMF mixture
of silica-PSSH hybrid particles and PVDF onto glass slides at 50 °C followed by
further drying under vacuum overnight (Figure 28). The dried membranes (15-35
pm thick, ~3x0.5-1 cm?) were peeled from the glass slides once immersed in
water. In composite membranes, silica-PSSH hybrid particles are the proton
donors and PVDF provides mechanical stability (Scheme 5). Table 2
summarizes the composition and experimental results for the composite

membranes.

 

(a)

 

silica-PSSH PVDF
in DMF in DMF

I |

Spread onto glass slide

 

 

Dry
Cut to size

Figure 28. (a) Schematic illustration of composite proton exchange membrane
fabrication. (b) Photo of a 4x4 cm2 composite membrane with 40 wt% silica-
PSSH particles.

4—4—4—

 

53

Scheme 5. Schematic diagram illustrating the structure of a composite
membrane.

Red: sulfonated polystyrene
functionalized nanoparticles

03H

Blue: PVDF

 

Table 2. Summary of silica-PSSH/PVDF membrane characteristics.

 

 

 

 

 

 

 

 

PEMs IEC Conductivity Conductivity Swelling Water
(particle (meqlg) at rt (mSIcm) at 60 °C ratio (%) uptake
wt%) (mSIcm) (wt%)

0 0 0 0 <3 0

10 0.47 2 4 <3 8

20 0.93 22 38 <3 15

30 1.40 27 61 <3 25

40 1.86 60 11 <3 43

50 2.33 70 12 <3 95

Nafion 1178 0.915 62 11 1815 21

 

 

 

 

 

 

 

a Nafion 117 was included for comparison.

Proper “water management" is critical for PEMs. While water contributes
to the dissociation of SO3H groups and acts as “vehicles” that transport protons

from the anode to the cathode, membranes with high water uptake are usually

54

 

associated with severe swelling and poor mechanical strength. A proper water
uptake is essential for facilitating high proton conductivity as well as maintaining
reasonable membrane stability. As shown in Figure 29, the water uptake
increased linearly with the silica-PSSH particle content up to ~37 wt%. Above
37 wt%, the water uptake increased significantly, indicating a change in
membrane phase morphology?03'205 Generally, when the volume fraction of the
conductive phase in a composite membrane is low, sulfonic acid groups are
isolated in ionic clusters in the matrix, and only a fraction of the clusters are
accessible to water. However, when the particle content reaches the percolation
threshold, the volume fraction at which the conductive phase becomes
continuous, the water uptake increases rapidly. PVDF is hydrophobic and thus

pure PVDF membrane (with 0 wt% particle) has a zero water uptake.

55

 

100

7

80‘ '0’
A ’0'
o\° I
E 60- .t
v I
(D ;
t5 :
“" i
Q. _ o
:3 40 ',
L— ,'
tn ,.
II-l 'v
(0 ,7"
E 20-

o”"
,1"
O-v'"

 

 

 

o 1'0'20'30 4o 50
Particle content in membranes (wt%)

Figure 29. Water uptakes of composite membrane as a function of particle
contents. The membrane water uptake was measured gravimetrically at room
temperature. (The dotted line is a guide to the eye)

The ion exchange capacity (lEC), expressed as mmol acid/g sample,
defines the density of charge carriers in a membrane and is directly related to a
membrane’s ability to absorb and retain water. Water uptake is directly related to
membrane lEC values and a sharp increase in lEC value is usually interpreted as
exceeding threshold for forming continuous hydrophilic domains.5°'2°3'2°6 As
shown in Figure 30, the IEC tracks the water uptake vs particle content, as
expected, since there is a 1:1 correspondence between the silica-PSSH particle

content and the IEC.

56

Wtaer uptake (wt%)

 

100

80-

60-

4o-

20-

 

 

 

0.5

r.
J"
110 115
lEC(meq/g)

2.0

2.5

Figure 30. Water uptakes as a function of membrane IEC. The membrane water
uptake was measured gravimetrically at room temperature. (Dotted lines were a
guide to the eye)

Proton conductivity is one of the most important parameters for evaluating

the performance of PEMs as well as the practicability of fuel cells. Since the

proton conductivity is strongly related to the number of charge carriers (H‘), a

high silica-PSSH particle content is preferred without the sacrifice of PEMs’

mechanical properties. As seen in Figure 31, the pure PVDF membrane (0 wt%

particle) has no measurable proton conductivity, consistent with its lack of charge

carriers. However, membranes with silica-PSSH particles were conductive, and

the conductivity increased linearly as a function of particle content.

57

Nafion 117 was used as a benchmark and as a control for the accuracy of
conductivity measurements. At room temperature, the proton conductivity of
Nafion 117 was 62 mS/cm, which is comparable to the reported data for Nafion
117 at room temperature (60 mS/cm)?07 Membranes with 40 and 50 wt%
particles had conductivities of 60 and 70 mS/cm at room temperature,
respectively, comparable to or higher than Nafion 117 run under the same
conditions. The high conductivity of the PEMs is consistent with a continuous
hydrophilic phase comprised of hybrid particles and water, similar to the
continuous proton network of channels of Nafion.11

The proton conductivity was relatively low at room temperature but
increased rapidly as the temperature was raised to 60 °C. Proton transport is
thermally activated, so higher temperatures lead to improved proton

conductivities.

58

140

 

 

 

 

120-
A I
E
2 100-
U)
.5.
Q 80-
.2
8 60- e
‘O
c:
O
o 40-
c
.9
O
n- 20"
O-Q

 

I I I

I7 1 r
0 1O 20 30 4O 50 60

Particle content in membranes (wt%)
Figure 31. Membrane proton conductivities as a function of particle content
for composite PEMs at 100% RH and room temperature (O) and 60 °C (A). Data

for Nafion 117 at room temperature (0) and at 60 °C (I) was shown for
comparison. (Lines were a guide to the eye)

Proton conductivity is strongly correlated to IE0, since the IEC defines the
carrier density in membranes. Figure 32 shows the proton conductivity as a
function of membrane IEC. As displayed, the conductivity increased linearly with

the increasing IEC.

59

 

 

 

 

 

120-
E
.2 100-
(I)
.5.
g‘ 80'
.2
g 60-
.0
8
0 40-
c
.9
9 20-
D.

o-e

0.0 0.5 1.0 1.5 2.0 2.5

IEC (meqlg)
Figure 32. Membrane proton conductivity as a function of lEC for composite
PEMs at 100% RH and room temperature (O) and 60 °C (A). Data for Nafion
117 at room temperature (0) and at 60 °C (I) was shown for comparison. (Lines
were a guide to the eye)

It was of great interest to investigate the proton conductivity vs water
uptake, because water molecules were needed as the mobile phase to facilitate
proton transport. As demonstrated in Figure 33, there was an almost linear
relationship between proton conductivity and water uptake up to around 40 wt%,
which was followed by a far slower increasing rate. This result suggested a

threshold close to 40 wt% water uptake, which was corresponded with 40 wt%

particle content in composite membranes.

60

 

.5

O

O
I

(I)
O
I

Proton conductivity (S/cm)
b 0)
<3 C:

N
O
I

 

 

 

20 40 60 80 1 00

I ' I

0"?
0

Water uptake (wt%)

Figure 33. Membrane proton conductivity as a function of water uptake for
composite PEMs at 100% RH and room temperature (O) and at 60 °C (A). Data
for Nafion 117 at room temperature (C) and at 60 °C (I) was shown for
comparison. (Lines were a guide to the eye)

The nature of the acid groups, chemical composition, and microstructure
of membranes all contribute to the desirable properties of PEMs. In order to
study membrane microstructure, the cross-sectional slices of composite
membranes with 40 and 50 wt% particles were analyzed with TEM (Figure 34).
Stained with Ag" ions, the hybrid nanoparticles showed up as dark regions in a

bright PVDF matrix. A phase-separated morphology was observed in both TEM

images. The hybrid nanoparticles aggregated to form continuous domains

61

immiscible with PVDF matrix, which is consistent with the plot of water uptake vs

lEC values.

 

Figure 34. Cross-sectional TEM images of membranes with (a) 40 wt% particles;
and (b) 50 wt% particles.

62

It has been a significant challenge to design PEMs with high proton
conductivity and minimal swelling. Dimensional stability, which is the change in
the measured membrane size for the dry and water saturated states, is a very
important parameter for practical fuel cell applications. If membranes swell too
much, electrodes can detach from the membrane surfaces as most electrodes
don’t swell, resulting in mechanical stress?08 As listed in Table 2, Nafion 117
swells significantly, 18 % from the dry to wet state. The severe expansion of
Nafion 117 explains its poor performance in direct methanol fuel cells, as the
permeability of Nafion to methanol is due to the high swelling of Nafion in water.
However, in both dry and hydrated states, all composite membranes with
different particle contents did not show any measurable size change (<3%). The
minimal water swelling is likely due to the crystalline structure of the hydrophobic
PVDF, which suppressed the water swelling of silica-PSSH particles in hydrated

membranes.

Conclusion

Direct polymerization of sodium 4-styrenesulfonate was carried out on
~7 nm silica particles by ATRP at room temperature in aqueous system. This
approach generated polyelectrolyte-coated silica nanoparticles with 100 %
polymer sulfonation degree. Kinetic study demonstrated that the surface
polymerization was a well controlled polymerization with first order kinetics up to
six hours followed by slightly slower reaction rate, which might be due to polymer

chain termination. This methodology provides an easy approach to generate

63

polyelectrolyte functionalized hybrid nanoparticles with a high surface-to-volume
ratio. TEM images proved that the anionic polyelectrolyte assisted the dispersion
of hybrid particles in water and minimized particle aggregation. IR spectroscopy
confirmed the presence of surface functional groups on silica particles and the
analysis of the TGA data confirmed the kinetic study results.

Solution-casting the mixture of the acidified hybrid particles and PVDF
formed composite proton exchange membranes. In addition to traditional
advantages of blend PEMs, such as easy design, synthesis, and processing, the
composite membrane system does not have the side effects related to “post-
sulfonation” reactions. Both the membrane water uptake and proton conductivity
increased linearly with the increase of particles in composite membranes and a
drastic increase showed up above a percolation threshold. The composite
membranes with 40 and 50 wt% particle contents revealed proton conductivities
comparable to or even higher than Nafion 117 under the same testing conditions.
The excellent conductivity might be due to either the high lEC value or the
phase-separated morphology of composite membranes. No water swelling was
observed with the composite membranes as the crystalline structure of the
hydrophobic PVDF matrix constrained the expansion of hybrid nanoparticles.
Moreover, this composite system is more cost effective compared to the

perfluorinated PEMs due to much easier synthesis and processing.

64

Experimental Section
Materials

Sodium 4-styrenesulfonate, CuBr (99.999%), CuBr2, (99.999%),
cetyltrimethylammonium bromide (CTAB), PVDF (MW = 534,000 Da), 2, 2’-
dipyridyl (bpy, 99+ %), and Nafion 117 were used as received from Sigma-
Aldrich except for bpy, which was sublimed prior to use. Snowtex-XS colloidal
silica, a gift from the Nissan Chemical Corporation, was received as an aqueous
dispersion of pH 9-10. Toluene was distilled over sodium using benzophenone as
an indicator and dimethylformamide (DMF) was dried over activated 4A
molecular sieves. The disodium salt of N,N,N’,N’-ethylenediaminetetraacetic acid
(EDTA-2Na) was purchased from Spectrum. All other chemicals and solvents
were ACS reagent grade and used as received from commercial suppliers

without further purification unless otherwise specified.

Characterization.

1H NMR analyses were performed at room temperature in deuterated
water on a Varian UnityPlus 500 spectrometer at 500 MHz with the residual
solvent proton signals as chemical shift standards. FT-IR spectra were acquired
from a Mattson Galaxy 300 spectrometer purged with dry nitrogen, with the
signal averaging 128 scans at a resolution of 4 cm". All IR samples were dried,
mixed with KBr, ground, and then pressed into pellets.

Thermogravimetric analyses (TGA) were carried out in air on Perkin-Elmer

TGA 7 instruments at a heating rate of 10 °C/min. TGA samples were dried under

65

vacuum at 80 °C overnight prior to use. Dried samples were held at 120 °C in the
TGA apparatus for 30 min prior to initiating the run.

The microstructure of particles and composite membranes were imaged
using either a JEOL-1OOCX or a JEOL-2200FS transmittance electron
microscope. To prepare the TEM sample for dried Snowtex-XS, particles were
dispersed in toluene by ultrasonication, and then a drop of the suspension was
drop-cast onto a carbon-coated copper mesh grid followed by solvent
evaporation in air. The TEM sample of silica-PSSNa particles was prepared in a
similar way from their water suspension. To prepare membrane cross-sections
for TEM analysis, membranes were immersed in 0.5 M AgN03 solution overnight
to stain the ionic domains by ion exchange of sulfonic acid groups for silver.
Samples were rinsed with water, and then dried under vacuum at 80 °C overnight.
The stained membranes were embedded in epoxy resin, dried at 50 °C for 24 h,
and then sectioned with an ultramicrotome to generate ~100 nm thick slices. The

slices were picked up with copper grids for TEM analysis.

Isolation of silica nanoparticles from Snowtex-XS

Cetyltrimethylammonium bromide (3 g) was added to 100 mL of Snovvtex-
X8 and the mixture was stirred at 50 °C for 30 min. The silica particles were then
collected by centrifugation and re-dispersed in 100 mL of Milli-Q water to remove
excess salt from the particle surfaces. After repeating the washing process four

times, the silica particles were dried under vacuum at 80 °C overnight.

66

Immobilization of ATRP initiators onto silica particles

Dry silica particles (1 g) were dispersed in 100 mL of dry toluene by
ultrasonication. Using a syringe, 0.5 mL of freshly synthesized initiator, (11-(2-
bromo-2-methyl)propionyloxy)undecyltrichlorosilane,196 was added to the
particles and the reaction mixture was stirred under a N2 atmosphere at 65 °C
overnight. The crude initiator-coated particles (silica-initiator) were isolated by
centrifugation, re-dispersed in 20 mL of dry toluene, precipitated from 100 mL of
pentane, and then collected .by centrifugation. The dispersion-precipitation-
centrifugation washing process was repeated six times to remove physisorbed
initiators from particle surfaces. Silica-initiator particles were then dried under

vacuum at 80 °C overnight.

Surface ATRP of sodium 4-styrenesulfonate

The surface-initiated ATRP of sodium 4-styrenesulfonate was carried out
at room temperature under N2 with CuBr/CuBr2 and bpy as the catalytic system.
Sodium p-toluenesulfonate was used as an internal standard for determining
monomer conversion. The ratios of [Monomer]o:[standard]o:[surface
initiator]o:[CuBr]o:[CuBr2]o:[bpy]o were approximately 70:10:1:2:O.2:6, with the
initial concentration of surface initiator estimated from TGA analyses of silica-
initiator particles. A typical synthesis is described. Sodium 4-styrenesulfonate
(0.43 g, 2.09 mmol), sodium p-toluenesulfonate (60 mg, 0.31 mmol), CuBr2 (1.4
mg, 0.006 mmol), bpy (28 mg, 0.18 mmol), 3 mL of Milli-Q water, and 1 mL of

methanol were charged into a 10 mL Schlenk flask equipped with a magnetic

67

stirrer. After complete dissolution of the solids, silica-initiator particles (62 mg,
0.03 mmol of surface initiator (calculated from TGA)) were added and then the
mixture was sonicated for 10 min to fully disperse the particles. After four freeze-
pump-thaw cycles, the Schlenk flask was filled with nitrogen at room temperature
and sonicated for 15 min to afford a fine dispersion of silica-initiator particles. An
initial sample (~0.2 mL) was removed via a syringe and diluted with 1 mL of Milli-
Q water and designated as sample I = 0. Finally, CuBr (8.6 mg, 0.06 mmol) was
charged into the Schlenk flask under a flow of nitrogen and with a rapid stirring.
At various reaction times, 0.2 mL samples were transferred from the
reaction mixture to vials, and after diluting with 1 mL of Milli-Q water, the vials
were shakened to terminate the surface polymerization. The samples were left
undisturbed overnight, allowing the particles to settle at the bottom of the vial as
a gel. The upper water layer was removed from the vial and dried under vacuum,
and the remaining solids were analyzed by 1H NMR for conversion. An aqueous
solution of EDTA-2Na was added to the particle suspension to remove residual
Cu2+ until the faint blue color associated with the particles had disappeared. The
hybrid particles were then isolated by centrifugation and washed extensively with
water to remove excess monomer, standard, and EDTA-2Na. Finally, acetone
was added to precipitate the particles, which were then collected and dried under

vacuum at 80 °C overnight.

68

General procedure for acidifying silica-PSSNa particles

A 1 M HCI solution was added to an aqueous suspension of silica-PSSNa
to exchange Na+ with H“. After 48 h, the silica-PSSH nanoparticles were
centrifuged, washed extensively with water, precipitated from acetone, and dried

under vacuum at 80 °C overnight.

Solution polymerization of sodium 4-styrensulfonate - linear PSSNa

A round-bottom flask was charged with sodium 4-styrenesulfonate (1.0 g,
4.9 mmol) and 5 mL of ethylene glycol, and the mixture was stirred until
homogeneous. The reaction system was purged with N2 for 30 min, and then the
flask was transferred to an oil bath maintained at 60 °C. The catalyst system was
made by dissolving Na2S205 (0.035 g, 0.18 mmol) and K2S2Oa (0.065 g, 0.24
mmol) in 1 mL and 2 mL of Milli-Q water, respectively. The two solutions were
injected into the reaction flask, separately, and then the reaction mixture was
stirred under a N2 atmosphere at 60 °C for 5 h. The crude poly(sodium 4-
styrensulfonate) (PSSNa) was precipitated from acetone, washed with acetone,

and then dried under vacuum at 80 °C overnight.

Preparation of composite membranes

Composite membranes were fabricated by the solvent-casting method.
This procedure describes the preparation of 10 mg membranes. Predetermined
amounts of PVDF and silica-PSSH nanoparticles were dissolved in DMF in

separate vials. Each solution was fixed at 0.05 g/mL, and the total mass of the

69

particles and PVDF was 10 mg. The two solutions were combined and stirred
until the system was homogeneous. Membranes were solution-cast by pouring
the mixture onto a clean glass plate at 50 °C, typically covering 4x1 cm2 area.
After 1 hour, the membrane had turned translucent, and the glass plate was
transferred to a vacuum at 50 °C for further drying overnight. The dried
membranes (15-35 pm thick) were cut to size (~3XO.5-1 cm?) and then removed
from the substrate by immersing the plate in water. The as-synthesized
membranes were soaked in 1 M HCI for 48 h to ensure that all sulfonate groups
in the membranes were in their acidic form. The membranes were then washed
extensively with water, and prior to conductivity tests, PEMs were stored in Milli-

Q water.

Conductivity measurements

The proton conductivities of membranes were measured with an
alternating current (AC) impedance analyzer HP 4192A over the frequency range
from 5 Hz to 13 MHz. Membranes were placed between two Pt electrodes in a
home-made Teflon cell. The entire setup was kept in Milli-Q water at controlled
temperatures for at least 10 min to saturate membranes and then impedance
was measured in in-plane direction. From the Nyquist plot, the resistance of the
membrane was estimated, and then the conductivity of the membrane was
calculated using the electrodes distance and membrane cross-sectional area, as

shown in equation 1:

L
“R7 ‘1’

70

Where ais the conductivity of the membrane, L represents the distance between
two electrodes, A stands for the membrane cross-sectional area, and R is the

bulk membrane resistance estimated from the Nyquist plot.

Water uptake

Water uptake of all membranes was measured gravimetrically. After
equilibration in Milli-Q water for two days at room temperature, membranes were
removed from water, and weighed after the surface water was quickly removed
by a Kim-wipe. Membranes were then dried under vacuum at 50 °C overnight
and the weight of dried membrane was measured. The membrane water uptake

was calculated according to equation 2:

 

W — W
Water uptake = wet dry x100% (2)
Wary

where Wwet stands for the mass of wet membranes and Wdry represents the

weight of dried membranes.

Dimensional stability

Membranes were saturated in Milli-Q water for two days at room
temperature prior to measuring the dimensions of swollen membranes. After
drying under vacuum at 50 °C overnight, the dimensions of the dried membrane
were measured. The swelling ratio was calculated from the length (longest

dimension) of the membrane, as shown in equation 3:

71

l -l
Swelling ratio = M x100% (3)
law

where [wet and law represent the lengths of wet and dried membranes,

respectively. The uncertainty in the measurement is <3%.

Theoretic ion exchange capacity (lEC)

The ion exchange capacity of membranes was estimated from the TGA
weight retention data of particles and particle weight fractions in membranes. The
lEC was calculated according to equation 4:

m +
IEC = —H—- (4)
dey

where mH+ expresses the quantity of sulfonic acid groups.

Calculation of surface initiator in silica-initiator particles from TGA data
We assume that mbs (900 °C) = mi (900 °C), where mbs and msi denote

the masses of a single bare silica particle and a single silica-initiator particle at
900 °C in TGA analysis, respectively. Using the residue mass of the silica (or
silica-initiator) as reference, the mass increase of silica-initiator to silica particles

is calculated from equation 5:
[(1--WR,,)/W-itsi —(1—WRbs)/WRbS:Ix100%=22.5% (5)
Where WRbS and WRS, are the weight retentions of bare silica particle (87.6 %)

and silica-initiator particle (73.2 %) at 900 °C in TGA, respectively. Therefore, the

mass of surface initiator on 62 mg silica-initiator particles is shown in equation 6:

72

22.5%x(62x WRSi) = 10.21 mg (6)

The thermally decomposable moiety of the surface initiator (the hydrocarbon

portion) has a molecular weight (MW) of 320 g/mol. Thus, the calculated amount

of surface initiator in 62 mg silica-initiator particles is 10.21/MW = 0.03 mmol.

o
ESi-(CH2)fiOd$<_
Br
k—W—J
Mw = 320

Calculation of TGA weight retentions of particles isolated at different
polymerization time

The calculation of the TGA weight retention of silica-PSSNa-1 particle
isolated at 10.4% monomer conversion is illustrated. The weight retentions of

silica-PSSNa-2 and silica-PSSNa-3 particles were calculated by the same

method and the results are summarized in Table 2. In addition to mbs (900 °C) =
ms; (900 °C), we also assume that the weight retention of silica-PSSNa particles
correspond to the residue of silica-initiator and the surface PSSNa. So mSID (900

°C) = ms; (900 °C) + mp5; (900 °C), where m;p and mpss represent the masses of

a single silica-PSSNa particle and PSSNa on a single silica-PSSNa particle at
900 °C in TGA analysis, respectively. Figure 22 showed that the weight retention
of linear PSSNa was 37.1 % at 900 °C, which would be used as a reference for
surface PSSNa, and silica-initiator particles retained 73.2 % of its mass at 900 °C
in TGA. So the TGA weight retention of silica-PSSNa-1 could be calculated from

equation 7:

73

Wsi x 73.2% + Wmo x Conv x 37.1%
Wsi 1' Wmo x Conv

 

x100% (7)

Where W3; and Wmo represent the initial weights of silica-initiator particles and

monomer charged in the surface polymerization, respectively, and Com, stands
for the monomer conversion. For the silica-PSSNa-1 particle, the experimental
data are W3; = 62 mg, Wmo = 430 mg, and Com, = 10.4 %. Therefore, the

calculated TGA weight retention is 58.1 %, which is reasonably close to the

actual data obtained from the TGA curve.

74

Chapter 3 Magnetic hybrid nanoparticles:
Synthesis, Characterization, and Application in Fuel Cells

Introduction

Recently, there has been considerable interest in the development of fuel
cells, especially hydrogen powered fuel cells, which cleanly convert the chemical
energy of fuels directly into electric energy and heat with high efficiency. A key
component of hydrogen based fuel cells is the proton exchange membrane
(PEM), which transports protons from the anode to the cathode and separates
fuel and oxidant. Nafion, the prototype PEM, has a high proton conductivity and
chemical and mechanical stability at moderate temperatures in the harsh fuel cell
environment. Although the actual morphology of hydrated Nafion is still under
investigation, it’s widely agreed that microphase-separation of Nafion’s
perfluorocarbon backbone and pendent fluorosulfonic acid groups accounts for
its high proton conductivity.“"“'15 The two domains of Nafion provide two critical
properties: the hydrophilic domain supports efficient proton transport while the
hydrophobic domain provides mechanical and chemical stability. Despite the high
performance of Nafion in fuel cells, its high manufacturing cost, drastically
reduced proton conductivity above 80 °C, environmental incompatibility, and
severe methanol permeability limit its use in fuel cell applications.

Developing alternatives to Nafion is a major challenge in current fuel cell
research. The usual design strategy for such membranes has been to synthesize
sulfonated analogues of thermally stable polymers, such as sulfonated

poly(arylene ether ether ketone),14 sulfonated polyimides,” sulfonated

75

3 and sulfonated polystyrene.160 Membranes derived from these

polysulfone,12
polymers have advantages over Nafion in terms of cost and reduced methanol
cross-over, but in their hydrated state, they generally suffer from poor
dimensional, chemical, and mechanical stability?“'1‘51'16‘3'168 The problems
associated with sulfonated polymers can be overcome by the use of polymer
blends, which are comprised of a hydrophilic polyelectrolyte phase embedded in
a non-conductive material for reinforcement.

Composite membranes made from the blending of a hydrophilic proton
conductor and PVDF have been investigated in our group. The hydrophilic proton
conductors were poly(4-styrenesulfonic acid)-coated silica nanoparticles, which
were synthesized by the surface-initiated ATRP of sodium 4-styrenesulfonate
from silica nanoparticles followed by acidification. The resulting composite
membranes have comparable or better performance than Nafion 117 at the same
testing conditions. However, the composite membrane system relied on the self-
assembly of the hydrophilic nanoparticles in the hydrophobic PVDF matrix and
there was no control over the alignment of particles in matrix material.

The distribution of proton conductors in membranes has a dramatic effect
on the performance of membranes. Aligning polyelectrolytes along the proton
transport direction could create conductive paths, which reduce the bulk
membrane resistance and improve proton conductivity. Electric fields have been
ued to align proton conducting polyelectrolytes while casting blend
membranes.1°1'18°'208 The polyelectrolytes were aligned into long linear chains

extending across the membranes. Membranes with aligned polyelectrolytes had

76

higher proton conductivities and less water uptake than membranes with
unaligned polyelectrolytes

Although applied electric fields have successfully aligned polyelectrolytes
while casting blend membranes, there are limitations to this technique. There
must be sufficient dielectric contrast between polyelectrolyte and the matrix, and
the matrix must sustain a high field and have sufficient mobility. These
restrictions have hindered the broad application of this processing technique due
to a limited choice of matrix materials. An alternative technique, using magnetic
fields as the external driving force, is more suitable for polyelectrolyte alignment
in blend membranes, as most polymers are anti-ferromagetic and hence are
good matrix materials. Unlike electric fields, magnetic fields are not restricted in
terms of the applied field strength, but only require that the polyelectrolytes have
sufficient magnetization to respond to a magnetic field.

Recently, Shaw et al. ued a magnetic field to align proton conductive
magnetic nanoparticles in sulfonated poly(ether ketone ketone) (SPEKK) during

the membrane casting process.181

Successful alignment of magnetic
nanoparticles in SPEEK can be seen in aligned membranes, while randomly
dispersed particles showed up in membranes without magnetic field processing.
Not surprisingly, the conductivity of membranes with aligned particles was twice
that of unaligned controls. However, the proton conductivity of the blend

membrane was much lower than a standard SPEKK membrane, which was

ascribed to the decreased lEC of the blend membrane.

77

While Shaw et al. showed that magnetic fields can be used to align
magnetic nanoparticles in a SPEEK matrix, there are limitations to Shaw’s
approach. The y-Fean particles were confined to the surfaces of composite
particles instead of being encapsulated by the polymer, making y-Fe203 particles
vulnerable to the strong acidic environment of PEMs (y—Fe203 particles dissolve
in acid). Secondly, the magnetic particles were aligned in the plane of the
membrane, while real PEMs require through-plane proton conducting paths.
Finally, both the magnetic particles and the matrix (SPEEK) are hydrophilic,
which would likely result in excessive swelling and low mechanical strength
membranes.

Herein, we systematically investigated the synthesis and characterization
of proton conducting magnetic nanoparticles, which were composed of magnetite
as the core, silica as the middle layer, and poly(4-styrenesulfonic acid) as the
polymer coating. The magnetite core enables magnetic alignment, the silica
shields magnetite from the acidic environment of PEMs, and the poly(4-
styrenesulfonic acid) coating supports proton conduction. “Magnetic” composite
membranes were then fabricated from a blend of proton conducting magnetic
nanoparticles and PVDF. During the solution casting of membranes, an external
magnetic field was applied perpendicularly to the membrane surface to induce
the alignment of nanoparticles in PVDF matrix across the membrane. The
“magnetic” composite membranes approach may be more cost effective than the

perfluorinated PEMs due to the much easier synthesis and simpler processing.

78

Results and discussion

Synthesis and characterization of magnetic nanoparticles

As displayed in Scheme 6, the desired magnetic nanoparticles were
synthesized by two different strategies. After coating silica onto magnetite (MNT)
particles to form MNT-silica nanoparticles, the particle surface was chemically
functionalized with ATRP initiators, providing a macroinitiator (MNT-initiator). In
the “post-sulfonation” approach, polystyrene was grown from the surface of the
nanoparticles followed by sulfonation to obtain proton conducting magnetic
nanoparticles (MNT-PSSH). The second strategy utilized the direct ATRP of
sodium 4-styrenesulfonate from MNT-initiator nanoparticles to generate the salt
form of proton conducting magnetic nanoparticles (MNT-PSSNa). The magnetic
composite nanoparticles, MNT-PSSH, have the dual properties of magnetic
susceptibility from the magnetic inorganic core and high proton conductivity from

the poly(4-styrenesulfonic acid), anchored to the surface.

79

Scheme 6. Schematic illustration of two strategies for the synthesis of poly(4-
styrene sulfonic acid)-coated magnetic nanoparticles.

n ClSOgH - n
03H I ‘

\
Ion exchange _
Post-sulfonation

SO3Na Initiator Sl02
(2 C: (2.: .
H

7 Direct polymerization
of sulfonated monomers

  

The MNT nanoparticles were synthesized according to a literature
procedure.209 FeCl3 was hydrolyzed by water produced in situ from the reaction
between sodium hydroxide and poly(acrylic acid) (PAA) generating Fe(OH)3
(Scheme 7). Fe(OH)3 was then reduced to Fe(OH)2 in diethylene glycol (DEG) at
220 °C, and converted to Fe3O4 by dehydration at high temperature. The MNT
were stabilized by electrostatic interactions between the iron cations and the
carboxylate anions of the polyacrylate surfactant, resulting in water soluble MNT

nanoparticles.

Scheme 7. Schematic illustration of (a) synthesis of MNT nanopaticles; and (b)
structure of PAA-capped MNT nanoparticles.

(a) . (b)
diethylene glycol
NaOH + FeCla + PAA ——-> i=e3o4 PM
220 °C

80

 

Figure 35. TEM image of MNT nanoparticles.

The high temperature hydrolysis reaction yielded monodisperse magnetite
nanocrystals with an average diameter of 12 nm (Figure 35). In addition, the high
temperature annealing during the reaction provided highly crystalline magnetite
nanoparticles, which was confirmed by X-ray diffraction (XRD). The XRD powder
pattern in Figure 36 shows the characteristic 220, 311, 400, 422, 511, and 440

reflections for cubic Fe304.2°9

81

 

311

220

  

111

Intensity (a.u.)

Powder MNT nanoparticles

  

 

 

 

 

 

20 I 40 ' 60 80 I 100
2-Theta (deg)

Figure 36. X-ray diffraction pattern of powder MNT nanoparticles. The vertical
bars correspond to the reflections and their relative intensity as recorded in
ASTM data cards for cubic Fe304.

The MNT nanoparticles were coated with silica to screen particle-particle
dipolar attractions, prevent their aggregation, and shield the MNT nanoparticles
from the acidic environment of PEMs. The surface coating was carried out in a
basic ethanol/water mixture with the MNTs acting as the nucleation seeds.21o
This strategy provided easy control over the resulting particle (MNT-silica) size
since the thickness of silica layer depends on the ratio of base:alcohol:water and
the reaction time. The as—synthesized MNT-silica nanoparticles were water
soluble due to surface silanol groups.

The size and morphology of MNT-silica nanoparticles were characterized

by TEM (Figure 37). The monodisperse MNT-silica nanoparticles are spherical

82

and have an average diameter of 50 nm with very little aggregation. The MNT
cores (dark color) are encapsulated by a ~20 nm thick silica layer. Some MNT-
silica nanoparticles had a single MNT nanoparticle as its core, but others had
multiple MNT cores, which might have resulted from the rapid growth of silica
layer onto MNT nanoparticles, or that some particles are ferromagnetic rather

than superparamagnetic.

 

Figure 37. TEM image of MNT-silica nanoparticles.

Energy-filtered transmission electron microscopy (EFTEM) is an efficient
analytical tool for obtaining the chemical composition of materials on the
nanometer and even subnanometer length scale. in EFTEM, the elemental

distributions are displayed as false color images. The images show that iron

83

(Figure 38b) only appears in the core while silicon (Figure 38c) is limited to the
shell, which was in good agreement with the actual morphology shown in Figure

38a.

 

Figure 38. (a) TEM image of MNT-silica. (b) EFTEM iron map. (c) EFTEM silicon
map.

84

The MNT-silica nanoparticles need to be superparamagnetic for their
application in fuel cell membranes. Superparamagnetic materials have no net
magnetism, but can be aligned in an external magnetic field. MNT-silica
nanoparticles were water soluble and did not exhibit any evidence for
ferromagnetic properties (Figure 39a). However, in presence of an external
magnet, the well-dispersed MNT-silica nanoparticles were localized by the
magnet demonstrating magnetic properties (Figure 39b). The response of MNT— ‘
silica nanoparticles to an external magnetic field is very useful for aligning

particles.

 

Figure 39. Photos of MNT-silica nanoparticles dispersed in water (a) without an
external magnetic field (a fine dispersion of particles in water); and (b) with an
external magnetic field (particles were localized toward the magnet).

The induced alignment of MNT-silica nanoparticles can be examined
microscopically. A TEM sample was prepared by adding a drop of the dilute

MNT-silica nanoparticle water dispersion to a TEM sample grid which was placed

in an external magnetic field. As shown in Figure 40, the four images captured

85

from different positions on the TEM grid show particles aligned into chain-like
structures along the magnetic field direction. The aggregation of MNT-silica
nanoparticles is consistent with the attractive interaction between magnetic
moments generated in response to the external magnetic field. This result
suggests the possibility of using an external magnetic field to align proton-

conducting magnetic particles while casting membranes.

 

 

) ~\ 5.
.39.: “33"
"‘2: .
a ‘1
I” * -
“s we
"L‘sh‘Q- .‘-..aa.r. 2"“
, .
~. \ v ,
”S“ _
F ‘ .
" \ «r
. NH
2W“

Figure 40. TEM images of chain-like structures formed by MNT-silica in
response to an external magnetic field (~1 Tesla).

86

Synthesis of poly(4-styrenesulfonic acld)-coated magnetic hybrid
nanoparticles (MNT-PSSH) by a “post-sulfonation” method

The silica shell of MNT-silica nanoparticles not only improves the chemical
stability of MNT nanoparticles in acidic conditions, but also facilitates the
introduction of surface functional groups through the reaction between surface
silanols and various coupling agents. As illustrated in Scheme 8, organosilane
initiators were chemically immobilized onto MNT-silica nanoparticle surfaces to
. generate sites for ATRP. Polystyrene (PSt) was then grown from the nanoparticle
surface to yield hybrid nanoparticles (MNT-PSt) with thick polymer shells. In this
polymerization, we also added a free initiator, ethyl 2-bromoisobutyrate, to initiate
polymerization in solution. The polymer formed in solution is often used as a
proxy for the polymer grown from the surface, and characterizing the polystyrene
formed in solution would provide some information about the polystyrene grown
from the surface. GPC analysis of polystyrene formed in solution gave a number
average molecular weight (Mn) of 56,000 Da and a polydispersity index (PDI) of
1.08. The low polydispersity is comparable to ATRP reactions in the absence of
nanoparticles.211 Finally, sulfonation provided poly(4-styrenesulfonic acid)—coated

magnetic hybrid nanoparticles (MNT-PSSH) for the proton conducting material in

“magnetic” composite membranes.

87

Scheme 8. Synthesis of poly(4-styrenesulfonic acid)—coated magnetic hybrid
nanoparticles.

Cl

0
- Cl-Si- CH —0
F6304 SIOZ a ( 2)‘1 ‘Sr
TEOS Br
0 ———* :

Toluene, 65 °C

 

 

 

 

 

MNT MNT-silica MNT-initiator
\
_ > ._ _ 0
M — /s. (CH2)11O-‘$:‘
Br
CuCl,anbpy

xylene, 130 °C, 6h

" 44”,!” 3!] CISO3H, 90 °c f n
:> (“'9‘ H . J CICH2CH2C| \H' . _ C:

‘Vuga'

MNT-PSSH MNT-PSt

FT-IR was used to confirm each surface reaction (Figure 41). The MNT-
silica nanoparticles have an absorption band at 1650 cm'1, which was ascribed to
the C=O bond from the surfactant (PAA) of Fe304 core. New bands at 3000 crn'1
to 2900 cm'1 (aliphatic C-H stretching) and 1750 cm'1 (C=O) in the spectrum of
MNT-initiator confirmed the successful anchoring of the initiator to MNT-silica
nanoparticles. After surface polymerization of styrene, the IR spectrum of the
MNT-PSt nanoparticles was dominated by the lcharacteristic bands of
polystyrene: aromatic C—H stretching at 3100 cm'1 and 3000 cm", aliphatic C—H
stretching from the polymer backbone at 3000 cm'1 to 2900 cm“, and the

overtones and combinations consistent with a mono-substituted benzene ring at

88

2000-1800 cm'1. After sulfonation, the IR spectrum of MNT—PSSH shows strong
O—H stretching modes from 3500 cm'1 to 2500 cm'1 due to water hydrogen-
bonded to surface sulfonic acid groups. Bands for the sulfonic acid groups were
partially obscured by a strong absorption band at 1050 to 1300 cm", from $102,

but the S=O symmetric stretching mode at 1040 cm'1 is partially visible.

 

 

 

 

 

d
8
C
(U
E
8 c
.Q
<
b
A a W
4000 3000 2000 1 000

-1
Wavenumber (cm )
Figure 41. FT-lR spectra of (a) MNT-silica nanoparticles; (b) MNT-initiator
nanoparticles; (c) MNT-PSt nanoparticles; and (d) MNT-PSSH nanoparticles.
Samples were made into pellets with KBr.
Thermal gravimetric analysis (T GA) of the magnetic nanoparticles was
used to establish the thermal stability of the polymer layer, and to estimate the

surface functional group coverage for the nanoparticles (Figure 42). The

89

cumulative weight loss at 800 °C for MNT-silica particles was ~6 %, which is
likely due to the loss of adsorbed water and condensation of surface silanol
groups. The TGA trace for MNT-initiator particles shows an additional weight loss
of ~10 %, from the degradation of surface initiators. After surface-initiated ATRP
of styrene from MNT-initiator nanoparticles, the MNT-PSt particles showed a
sharp weight loss at ~300 °C, characteristic of the depolymerization of
polystyrene chains. The 86 % weight loss at 800 °C indicates a thick surface
coating of PSt. The thermal degradation of MNT-PSSH nanoparticle shows two
steps: the first, from 200 to 400 °C, is assigned to desulfonation, while the

second, from 400 to 640 °C, is degradation of residual polystyrene.

90

 

 

100

 

 

 

a
94 %
b
84 %
80 -
a",
8 60 -
E
9 c
9 45 %
E 40 -
.9
é’
20 -
d 14 %
13 %
0 I fi'i T ' I ' l
200 400 600 800

Temperature (°C)
Figure 42. TGA analysis of (a) MNT-silica; (b) MNT-initiator; (c) MNT-PSSNa; (d)
MNT-PSt; and (e) MNT-PSSH.

TEM analysis shows that surface-initiated ATRP yielded MNT-PSt
nanoparticles with an average diameter of 150 nm (Figure 43). The
nanoparticles were well-defined with a black core of Fe304, a grey middle layer
due to silica, and a thick outer shell composed of polystyrene. The polymer shells
have an average thickness of 50 nm and were uniform for all nanoparticles,

which is consistent with controlled surface growth of polymer chains by ATRP.211

91

The morphology of the Fe304 and silica layer in MNT-PSt was unchanged from
the image shown in Figure 37 for MNT-silica nanoparticles.

Interestingly, the isolated MNT-PSt nanoparticles shown in Figure 43a
also formed “coffee ring stain-like” aggregates (Figure 43b) due to the way the
TEM sample was prepared. A drop of the toluene dispersion of MNT-PSt
nanoparticles was dropped to the water surface, and a TEM grid submerged in
water was slowly raised to capture nanoparticles. During solvent evaporation,
toluene might have carried nanoparticles to solvent frontier, where the particles

aggregated.

92

 

500 nm

 

(b)

 

Figure 43. TEM image of (a) isolated MNT-PSt nanoparticles; and (b) “coffee
ring stain"-like aggregated MNT-PSt nanoparticles.

93

Direct polymerization of sodium 4-styrenesulfonate

In contrast to the “post-sulfonation” approach, which introduces the proton
conducting sulfonic acid groups after polymerization, the polymerization of
sodium 4-styrenesulfonate from MNT-initiator surfaces directly generates the
sodium salt of MNT-PSt. As illustrated in Scheme 9, surface-initiated ATRP of
sodium 4-styrenesulfonate from MNT-initiator nanoparticles in aqueous solution
at room temperature yields a dense shell of poly(sodium 4-styrenesulfonate)
(PSSNa). Acidification of the hybrid MNT-PSSNa nanoparticles generates poly(4-
styrenesulfonic acid) coated magnetic nanoparticles for the proton conducting

material in PEMs.

Scheme 9. Synthesis route to poly(sodium 4-styrenesulfonate)-coated magnetic
hybrid nanoparticles.

 

 

Cl" f—(CH2)"
{I 5%; r0803Na

Toluene, 65 °C CuBr, CuBr2, bpy
HZO/CH3OH, rt, 8 h

MNT-silica MNT-initiator MNT-PSSNa

Ll .
’1
> ,
‘l
\

 

 

O

= 5910””?J3g
r

The surface polymerization of sodium 4-styrenesulfonate from MNT-

 

 

 

initiator formed MNT-PSSNa nanoparticles with an average diameter of 100 nm
(Figure 44). The morphology of Fe304 and silica layer in MNT-PSSNa was
similar to the MNT-silica nanoparticles displayed in Figure 37. The hybrid silica-

PSSNa nanoparticles have uniform polymer shells with an average thickness of

94

25 nm, which indicated that surface polymer chains were grown at comparable

rates.

 

Figure 44. TEM image of MNT-PSSNa nanoparticles. Black spots are the Fe304
cores, the gray middle layer is the silica coating, and the light gray outer shell is
composed of PSSNa.

The FT-IR spectrum of MNT-PSSNa nanoparticles was dominated by the
characteristic peaks of PSSNa (Figure 45). The strong O—H stretching bands at
3435 cm‘1 was due to water absorbed by the hygroscopic polyelectrolyte.
Absorption peaks between 3100 cm“1 and 3000 cm'1 were assigned to aromatic
C-H stretching, while bands from 3000 cm'1 to 2900 cm" resulted from aliphatic

C—H stretching. In addition to the bands at high wavenumbers, the presence of

95

PSSNa was confirmed by the 8:0 asymmetric and symmetric stretching bands

at 1200 and 1040 cm", respectively.

 

Absorbance

 

 

 

 

4000 ' 30'00 ' 2600 I 1600
Wavenumber (cm'1)

Figure 45. FT-IR spectrum of MNT-PSSNa nanoparticles. The sample was made

into a pellet with KBr.

The TGA curve of MNT-PSSNa, as presented in Figure 42c, exhibited
two major mass loss events, which we interpreted as the loss of sulfonate groups
at approximately 400 °C, and degradation of the polymer backbones
commencing at around 450 °C. It is noteworthy that PSSNa had a nonvolatile
residue (likely NaZO) at 800 °C in an oxidative atmosphere.202 Thus, 45 % weight
retention of MNT-PSSNa particles at 800 °C underestimates the amount of

polymer on the hybrid nanoparticles. A comparison between the TGA traces of

96

MNT-PSSNa and MNT-PSSH particles suggests that the free acid was thermally
less stable than the sodium salt in an oxidative atmosphere. The MNT-PSSNa
nanoparticle had its first dramatic weight loss commencing at 400 °C while MNT-

PSSH degraded much earlier at ~200 °C.

Fabrication and characterization of “magnetic” composite PEMs

The magnetic properties of different nanoparticles were tested with a
vibrating sample magnetometer at room temperature. As shown in Figure 46, the
nanoparticles had a strong magnetic response to a varying external magnetic
field. The field-dependent magnetization plots (inset) indicated that the magnetic
nanoparticles were slightly ferromagnetic with very small remanence (the
magnetization that persists after an external magnetic field is removed) and
coercivity (the intensity of an external magnetic field required to reduce the
magnetization of that material to zero after the magnetization of the sample has
been driven to saturation), which might be due to some ferromagnetic impurities.
The saturation magnetization value of MNT-silica nanoparticles was 4.9 emu/g
but decreased in polymer coated magnetic nanoparticles, primarily due to the
nonmagnetic polymer shell since the units for magnetization are expressed as

per gram of sample.212

97

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

6
—MNT—silica
—MNT-PSt
4' —MNT-PSSNa
Z")
3 2-
E
i”,
g o ’
'«g #11 1o
' I
.N
*5 _2_ 0.5- /
8)
m . o.o 4,
E —4- '0-5'l /
-1.o - 1‘ .
-6 -230 .100 o 100 200
-10000 -5000 0 5000 10000

Applied magnetic field (0e)
Figure 46. Room temperature hysteresis curves of magnetic nanoparticles. The
inset demonstrated the data around zero field from -200 Oe to 200 Oe.

Generally, “magnetic” composite membranes were obtained by solution-
casting a DMF mixture of MNT-PSSH hybrid particles and PVDF (Figure 47).
Membranes with aligned MNT-PSSH nanoparticles were cast in the presence of
an external magnet aligned perpendicularly to the membrane plane. The ~1
Tesla field was maintained until all solvent had evaporated to prevent
randomization of nanoparticles upon removal of the external magnetic field.
Membranes were light brown color, possibly due to the dark color of iron oxide

(Figure 47b).

98

 

(a) MNT-PSSH (bl

in DMF

 

I
Spread onto glass slide perpendicular \ "
to external magnetic field

1..

Cut to size . ‘-

 

Figure 47. (a) Schematic illustration of composite proton exchange membrane
fabrication. (b) Photo of a 2.5x2.5 cm2 composite membrane with 30 wt% MNT-
PSSH particles cast in presence of an external magnetic field.

The performance of membranes can be evaluated by their polarization
curves. Generally, there are three divisions in a polarization curve: activation
polarization, ohmic polarization, and concentration polarization. When the current
is zero, the open circuit voltage is the cell voltage. When the current starts to flow,
the voltage decreases owing to the activation polarization, after which the
polarization curve becomes linear due to the ohmic polarization. This region is
associated with the ohmic losses and its slope is correlated to the resistance of
the membrane-electrode assembly. The polarization behavior of several
composite membranes in fuel cells is displayed in Figure 48. The experiments
were run under fully humidified conditions, with H2 and 02 as the fuel and oxidant,
respectively. Nafion 117 and pure PVDF membranes were used as controls.
Pure PVDF membranes (0 wt% MNT-PSSH nanoparticles) had little or no proton
conductivity, which is not surprising since there was no conductive ionomers. The

resistance of a membrane with 20 wt% particles was slightly higher than Nafion

99

117, but membranes with 30 wt% and 40 wt% particles performed better than
Nafion 117 in the ohmic region, and the performance of these “magnetic"

membranes increased with the MNT-PSSH content.

 

 

 

 

 

 

 

1 1 +Nafion 117
k +0 wt% particle (pure PVDF)
20 wt% particle
0'8 _ + 30 wt% particle
+40 wt% particle
2. 0.6 1'—
o
O}
3 II
T; a
> 0.4
II
0.2 ~
II
II
0 g l l l l
0 1 2 3 4 5 6

Current Density (mAlcmZ)
Figure 48. Polarization curves of composite membranes with different magnetic
nanoparticle contents and Nafion 117 at room temperature.

Ideally, PEMs should have high proton conductivity and minor membrane
swelling. Membrane dimensional stability was obtained by measuring the
dimensional changes between the dry and water saturated states. Membranes
that swell excessively can detach from the electrode surfaces as most electrodes
do not swell, resulting in mechanical stress.208 Nafion 117, equilibrated in water
for 48 h, increased 18 % in length compared to the dry film.15 The severe

expansion of Nafion 117 explains its poor performances in direct methanol fuel

100

cells, as the permeability of Nafion to methanol is related to the high swelling of
Nation in water. However, all of the “magnetic" membranes swelled <3 %. The
low water swelling behavior may be due to the rigid matrix of crystalline PVDF,
which suppressed the water swelling of MNT-PSSH particles in hydrated
membranes.

Cross-sections of a 30 wt% MNT-PSSH aligned membrane were studied
by TEM (Figure 49). The dark area represents nanoparticles and the light
regions are PVDF. The magnetic nanoparticles aligned into elongated domains
within the PVDF matrix, presumably due to the application of an external

magnetic field during the film casting process.

 

Figure 49. Cross-sectional TEM image of a 30 wt% MNT-PSSH aligned
membrane.

101

Dispersive X-ray (EDX) line scanning was used to determine the
elemental composition of the phases seen in the TEM images (Figure 50a).
Fluorine, which is associated with the PVDF matrix, was found evenly along the
scan line (Figure 50b). Sulfur, carbon, and oxygen, associated with the proton
conducting MNT-PSSH nanoparticles, were concentrated in the dark area, which
confirmed the phase-separated morphology of composite membranes. Nearly all
of the dark regions are connected into a network of conducting channels, which
likely contributed to the outstanding performance of the aligned composite

membrane with 30 wt% MNT-PSSH particles shown in Figure 48.

102

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

o ' 1 3 71? o 1 2 pm
Sulfur Ka1 Carbon Ka1_2
2 )- i 3
5 2
1 -
1
o ' ' I l 0"
o 1 2 pm 0 1 2 pm
Oxygen Ka1 Fluorine Ka1-2

Figure 50. (a) Cross-sectional TEM image of aligned composite membrane with
30 wt% MNT—PSSH content. (b) EDX line-scan along the line displayed in (a).

103

Conclusion

Proton conducting magnetic nanoparticles with a magnetite core, silica
buffer layer, and a polyelectrolyte outer shell were synthesized by two different
strategies, both of which utilized surface ATRP from the same macroinitiator. All
synthesized magnetic nanoparticles had narrow size distributions. The uniform
thickness of surface polymer shell in MNT-PSt and MNT-PSSNa nanoparticles
implied the controlled “living” polymerization characteristics of surface ATRP.
Applying an external magnetic field to MNT-silica nanoparticles aligned these
nanoparticles in magnetic field direction.

Solution-casting the mixture of MNT-PSSH particles and PVDF formed
“magnetic” composite proton exchange membranes, which were associated with
advantages including low cost, and easy design, synthesis, and processing
compared to perfluorinated PEMs. In the polarization experiments, “magnetic”
composite membranes with 30 and 40 wt% particles revealed better performance
than Nafion 117 under the same testing conditions. The magnetic nanoparticles
in aligned composite membranes were noted to form chain-like aggregates in
PVDF matrix leading to a phase-separated morphology.

In addition to proton exchange membrane, the polyelectrolyte coated
magnetic hybrid nanoparticles might find applications in other areas, like ion
exchange resin, water purification, and biology due to the versatile properties of

magnetite and surface polyelectrolytes.

104

Experimental Section
Materials

Styrene, sodium 4-styrenesulfonate, CuBr (99.999%), CuBr2 (99.999%),
PVDF (MW = 534,000 Da), 2, 2’-dipyridyl (bpy, 99+ %), 4,4’-dinonyl-2,2'-dipyridyl
(anbpy), and Nafion 117 were obtained from Sigma-Aldrich. Poly(acrylic acid)
(PAA, MW = 2,000 Da) was purchased from Monomer-Polymer & Dajac Labs.

Styrene was purified by passing through activated basic alumina column followed
by distillation and bpy was sublimed prior to use. Toluene and tetrahydrofuran
(THF) were distilled over sodium and benzophenone. Dimethylforrnamide (DMF)
was dried over activated 4A molecular sieves. The disodium salt of N,N,N’,N'-
ethylenediaminetetraacetic acid (EDTA-2Na) and diethylene glycol (DEG) were
purchased from Spectrum and Fisher Scientific, respectively. All other chemicals
and solvents were ACS reagent grade and used as received from commercial

suppliers without further purification unless otherwise specified.

Characterization

IR spectra were acquired from a Mattson Galaxy 300 spectrometer purged
with dry nitrogen, with the signal averaging 128 scans at a resolution of 4 cm".
All lR samples were dried, mixed with KBr, ground, and then pressed into pellets.
Thermogravimetric analyses (TGA) were carried out in air on Perkin-Elmer TGA
7 instruments at a heating rate of 10 °C/min. TGA samples were dried under
vacuum at 80 °C overnight prior to use. Dried samples were held at 120 °C in the

TGA apparatus for 30 min prior to initiating the run.

105

The crystallographic study of the synthesized magnetite powder was
performed on a Rigaku X-ray diffractometer using Cu Kc radiation at room
temperature. The magnetic properties of different magnetic nanoparticles were
studied at room temperature with a Quantum Design magnetic property
measurement system (MPMS), which utilized a superconducting quantum
interference device (SQUID) magnetometer.

The microstructure of particles and composite membranes were imaged
using either a JEOL-1OOCX or a JEOL-2200FS transmittance electron
microscope. To prepare TEM samples for MNT and MNT-PSSNa nanoparticles,
their dilute water solutions were drop-casted onto carbon-coated copper mesh
grids followed by solvent evaporation in air. For MNT-PSSH, a toluene solution of
particles was dropped onto Milli-Q water, and as the toluene spread, it carried
nanoparticles on water surface. A TEM grid submersed in water was slowly
moved up to water surface to “fish" the nanoparticles. To prepare TEM sample of
membrane cross-sections, membranes were processed as follows: dried under
vacuum at 80 °C overnight, embedded in epoxy resin, dried at 50 °C for 24 h,
and finally sectioned with an ultramicrotome to generate ~100 nm thick slices.
The slices were picked up with copper grids for TEM analysis.

Polymer molecular weights were acquired with gel permeation
chromatography (GPC). This characterization was done at 35 °C using two PLgel
10mixed-B columns in series with THF as the eluting solvent at a flow rate of 1
mL/min and a Waters 2410 differential refractometer as the detector. Mono-

disperse polystyrene standards were used to calibrate the molecular weights.

106

The absolute polymer molecular weights could be calculated using an Optilab
rEX (Wyatt Technology Co.) and a DAWN E08 18 angle light scattering detector

(Wyatt Technology Co.) with a laser wavelength of 684 nm.

Synthesis of magnetite (MNT, Fe304) nanoparticles

The magnetite nanoparticles were synthesized according to the procedure
of Yin’s work with some moolifications.209 A solution of NaOH (2 g, 50 mmol) and
20 mL of DEG was heated at 120 °C for 1 h under a nitrogen purge, and then
was cooled to 70 °C with stirring until the solution turned to dark red so the
dissolution of NaOH in DEG was complete. To a 50 mL 3-neck round bottom
flask, were added poly(acrylic acid) (0.432 g, 6 mmol acrylic acid repeating units),
FeCla (0.488 g, 3 mmol), and 22.5 mL of DEG. The mixture was heated at 220 °C
under nitrogen for 30 min with mechanic stirring, and then 6 mL of NaOH/DEG
stock solution was quickly added. The mixture rapidly turned black, and was kept
at 220 °C for another 20 min before slowly cooling to room temperature. The
particles were collected with the help of a permanent magnet and after adding
nitrogen-purged ethanol, they were re-dispersed in 0.5 mL of nitrogen-purged
Milli-Q water. This washing process was repeated for three times to remove
excess reactant and especially physisorbed poly(acrylic acid). The Milli-Q water
and ethanol must be purged by nitrogen for at least 30 min prior to use and the
washing process must be finished very quickly to prevent further air oxidation of
the magnetic particles. Finally, the magnetite nanoparticles (MNT) were made

into 0.5 wt% dispersion by dispersing in 45 mL of Milli-Q water.

107

Coating MNT nanoparticles with silica to synthesize MNT-silica

Using the method described by Deng et al. with some modification, silica-
coated magnetic nanoparticles were prepared in basic alcohol/water mixture at
room temperature with MNT as nucleation seeds.21o To 11 mL water suspension
of MNT, were added 110 mL of Milli-Q water, 480 mL of ethanol, and 10 mL of
ammonia solution. After ultrasonication for 15 min, the mixture was vigorously
stirred. Then, 2 mL of tetraethyl orthosilicate (TEOS) was added to the dispersion
and the mixture was stirred at room temperature for 2 h forming silica coating on
the surface of MNT nanoparticles. A solution of 2 M HCI was then added
dropwise to flocculate the fine dispersion. The resulting particles (MNT-silica)
were then collected with a permanent magnet and washed with ethanol and
water six times until the supernatant was neutral. Finally, the MNT-silica

nanoparticles were dried under vacuum at 80 °C overnight.

Immobilization of organosilane initiator onto MNT-silica to make MNT-
intiator

Dry MNT-silica nanoparticles (0.2 g) were dispersed in 60 mL of dry
toluene by ultrasonication. Using a syringe, 0.1 mL of freshly synthesized initiator,
(11-(2-bromo-2-methyl)propionyloxy)undecyltrichlorosilane,196 was added to the
particles, and the reaction mixture was stirred under a N2 atmosphere at 65 °C
overnight. The resulting initiator immobilized magnetic nanoparticles (MNT-
initiator) were collected with a permanent magnet, washed with dry THF six times,

and dried under vacuum at 80 °C overnight.

108

Surface ATRP of styrene from MNT-initiator to make MNT-PSt

A 10 mL Schlenk flask was charged with styrene (2.7 _mL, 23.3 mmol),
MNT-initiator (0.05 g), ethyl 2-bromoisobutyrate (3.4 ,uL, 0.023 mmol), 4,4’-
dinonyl-2,2’-dipyridyl (anbpy) (0.29 g, 0.714 mmol), and 1.1 mL of xylenes.
After four freeze-pump-thaw cycles, the flask was backfilled with nitrogen and
CuCl (23.3 mg, 0.238 mmol) was added. The mixture was treated with another
three freeze-pump-thaw cycles and then stirred at 130 °C for 6 h in a nitrogen
atmosphere. After quickly cooling down to room temperature in air, the
polymerization system was diluted with 20 mL of THF and the magnetic particles
were collected with a magnet. The supernatant containing linear polystyrene was
condensed and precipitated from methanol to yield the polystyrene homopolymer.
The isolated magnetic particles were washed by redispersing in THF and then
being collected with a magnet. This washing process was repeated three times to
fully remove linear polystyrene formed in solution. Finally, polystyrene and the
magnetic nanoparticles (MNT-PSt) were dried under vacuum at 80 °C overnight,

respectively.

Sulfonation of MNT-PSt to make MNT-PSSH nanoparticles

The sulfonation reaction was carried out by adding 0.05 mL of
chlorosulfonic acid to a refluxing dispersion of MNT-PS nanoparticles (0.1 g) in
50 mL of 1,2-dichloroethane. Dark brown precipitate immediately formed from the
homogeneous dispersion. The precipitate was washed with THF, being careful to

remove physisorbed acid, and finally dried under vacuum at 80 °C overnight.

109

Surface ATRP of sodium 4-styrenesulfonate from MNT-initiator surface to
synthesize MNT-PSSNa

In this surface-initiated ATRP reaction of sodium 4-styrenesulfonate,
CuBr/CuBr2 and bpy were applied as the catalytic system. Sodium 4-
styrenesulfonate (0.43 g, 2.09 mmol), MNT-initiator nanoparticles (62 mg), CuBr2
(1.4 mg, 0.006 mmol), bpy (28 mg, 0.18mmol), Milli-Q water (3 mL), and
methanol (1 mL) were added to a 10 mL Schlenk flask equipped with a magnetic
stirrer. Particles were dispersed in the reaction mixture by ultrasonication for 10
min. After four freeze-pump-thaw cycles, the Schlenk flask was backfilled with
nitrogen and CuBr (8.6 mg, 0.06 mmol) was added. The polymerization was
carried out in a nitrogen atmosphere at room temperature for 8 h. The as-
synthesized magnetic nanoparticles were collected with a permanent magnet
and then washed with Milli-Q water for further purification. Finally, the magnetic

nanoparticles (MNT-PSSH) were dried under vacuum at 80 °C overnight.

Fabrication of “magnetic” composite membranes

Composite membranes were fabricated by the solvent-casting method.
This procedure describes the preparation of 40 mg membranes. Predetermined
amounts of PVDF and MNT-PSSH nanoparticles were dissolved in DMF in
separate vials. Each solution was fixed at 0.05 g/mL, and the total mass of the
particles and PVDF was 40 mg. The two solutions were combined and stirred
until the system was homogeneous. Membranes were solution-cast by pouring
the mixture onto a clean glass plate at 50 °C, typically covering 4x4 cm2 area,

with an external magnetic field applied perpendicularly to the glass slide surface

110

until all solvent evaporated. After 1 hour, the membrane had turned translucent,
and the glass plate was transferred to a vacuum at 50 °C for further drying
overnight. The dried membranes (15-35 pm thick) were then removed from the
substrate by immersing the plate in water. The as-synthesized membranes were

stored in Milli-Q water prior to tests.

Dimensional stability

Membranes were saturated in Milli-Q water for two days at room
temperature prior to measuring the dimensions of swollen membranes. After
drying under vacuum at 50 °C overnight, the dimensions of the dried membrane
were measured. The swelling ratio was calculated from the length (longest
dimension) of the membrane, as shown in equation 1:

l —l
Swelling ratio = —w—et——dl x100% (1 )
’dry

where [wet and [dry represent the lengths of wet and dried membranes,

respectively. The uncertainty in the measurement is <3%.

111

Chapter 4 Sulfonated Polyimide and PVDF based Blend Proton
Exchange Membranes for Application in Fuel Cells

Introduction

Recently, there has been considerable interest in the development of fuel
cells, especially hydrogen powered fuel cells, which cleanly convert the chemical
energy of fuels directly into electric energy and heat with high efficiency. A key
component of hydrogen based fuel cells is the proton exchange membrane
(PEM), which transports protons from the anode to the cathode and separates
fuel and oxidant. Nafion, the prototype PEM, has a high proton conductivity and
chemical and mechanical stability at moderate temperatures in the harsh fuel cell
environment. Although the actual morphology of hydrated Nafion is still under
investigation, it’s widely agreed that microphase-separation of Nafion’s
perfluorocarbon backbone and pendent perfluorosulfonic acid groups accounts
for its high proton conductivity.“'1“'15 The two domains of Nafion provide separate
properties: the hydrophilic domain supports efficient proton transport while the
hydrophobic domain provides mechanical and chemical stability. Despite the high
performance of Nafion in fuel cells, its high manufacturing cost, drastically
reduced proton conductivity above 80 °C, environmental incompatibility, and
severe methanol permeability limit its use in fuel cell applications.

Developing alternatives to Nafion, and especially those that can operate at
high temperatures, is a major challenge in current fuel cell research. The usual
design strategy for such membranes has been to synthesize sulfonated

analogues of thermally stable polymers, such as sulfonated poly(arylene ether

112

ether ketone),14 sulfonated polyimides,83 sulfonated polysulfone,123 and
sulfonated polystyrene‘so. Membranes derived from these polymers have
advantages over Nafion in terms of cost and reduced methanol cross-over, but in
their hydrated state, they generally suffer from poor dimensional, chemical, and
mechanical stability.2'61'"51'166'168 The problems associated with sulfonated
polymers can be overcome by the use of polymer blends, which are comprised of
a hydrophilic polyelectrolyte phase embedded in a non-conductive material for
reinforcement.

The two components of a blend membrane provide complementary
functions, with the polyelectrolyte providing proton conductivity while a second
polymer affords mechanical integrity. In addition, blending usually involves
commercially available polymers or inexpensive polymer precursors and thus
blending is a simple and cost-effective route to produce PEMs. More importantly,
the blending approach eliminates the need to optimize the mechanical, chemical
and physical properties in a single polymer chain and it has the potential to
develop a wide variety of PEMs.

The blend morphology is crucial since the conducting phase must be
continuous for meaningful proton conductivity. Nafion’s high proton conductivity
has been ascribed to the development of a continuous channel morphology when
hydrated, and analogous bicontinuous blends should also exhibit high proton
conductivities. The selection of appropriate hydrophilic and hydrophobic blending

materials must consider the interactions between the two blending

113

17° in order to form continuous conducting channels and high proton

components,
conductivities analogous to Nafion.157

The naphthalenic sulfonated polyimides (SPls) are a widely studied
polyelectrolyte system due to their excellent thermal, chemical, and mechanical
stability, high proton conductivity, good film-forming ability, and low gas
permeability.‘"32's?““'149'157'158 However, SPls also tend to swell extensively.
Poly(vinylidene fluoride) (PVDF) is an excellent candidate for the hydrophobic
host in blend membranes; PVDF is commercially available, chemically and
mechanically stable, and can be easily processed into membranes.49'1°8"7"173 In
addition, PVDF based membranes have reduced methanol permeability and
swelling ratios, and exhibit good performance in direct methanol fuel cells
(DMFC).1°8'213

Herein, we describe blend membranes derived from SPI and PVDF. The
SPI was synthesized by direct polymerization of nathphalenic dianhydride and
sulfonated diamine monomers. With its high degree of sulfonation, SPI has
excellent proton conductivity, while the rigid PVDF matrix, incompatible with SPI,
provides excellent mechanical and dimensional integrity. In addition, micro-phase

separation of the two incompatible materials generates proton conducting

channels comprised of SPI aggregates, and enhances proton conductivity.

Results and discussion
Synthesis and characterization of sulfonated polyimide (SPI)
The sulfonated polyimide homopolymer was synthesized by step growth

polymerization as illustrated in Scheme 10.214 4,4'-Diaminobiphenyl 2,2:-

114

disulfonic acid (BDSA) was firstly converted to its triethylammonium salt, and
then benzoic acid was introduced as the catalyst for the polycondensation of
1,4,5,8-naphthalene tetracarboxylic dianhydride (NTDA) and BDSA in m-cresol at
180 °C. The high reaction temperature helps remove water from the system and
drive the reaction toward polymer formation, which ensured the synthesis of high
molecular weight SPI. Furthermore, under the reaction conditions, polymer
growth (amide formation) and imidization occurred spontaneously. The as-
synthesized SPI in its triethylammonium salt form was a hard and tough solid and
was difficult to grind, implying a high polymer molecular weight. Using the
sulfonated diamine monomer avoided the drawbacks associated with “post-
sulfonation” method, which usually results in low sulfonation degrees, cross-

linking, and polymer chain degradation.2

Scheme 10. Synthesis of sulfonated polyimide (SPI).

 

-+
$03H SO3NH(Et)3
‘ , N(Et)3
HZN NH2 > H N NH
0 Q m-cresol, rt 2 Q Q 2
H03S (Et)3Ht§93

 

 

O O
, O,
N N
Benzoic acid, m-cresol, 80 °C, 4h; Q

180 °c, 14 h 0 0

 

115

The chemical structure of the SPI triethylammonium salt was confirmed by
1H NMR spectroscopy (Figure 51 a). Proton peaks from the ethyl groups at ~1.1
and 3.0 ppm confirm the presence of the triethylammonium moiety in the polymer.
Genies and others‘58'214 assigned peaks at 8.7-8.8 ppm to the naphthalenic and
ammonium protons, while the peaks at 7.35, 7.8, and 8.0 ppm were attributed to
phenylene protons in the BDSA polymer segments. While our spectra showed
comparable data, we found that the peaks were broadened, and the peak areas
for the resonances at 8.9 ppm, and 8.0 ppm were significantly larger than the
peaks at 7.35 and 7.8 ppm. The broadening of the NMR resonances is consistent
with a rigid, high molecular weight polymer.

The diminution of the peaks at 7.35 and 7.8 ppm suggests that they are
related to the polymer end groups, and that our material may have a higher
molecular weight. Usually, the starting materials for polycondensation reactions
are dried, but reagents and solvents are often used as received?”214 In our SPI
syntheses, we dried or distilled all starting materials (BDSA and NTDA), reagents
(benzoic acid and triethyl amine), and solvent (m—cresol) prior to the
polycondensation reaction in order to obtain high molecular weight. We surmised
that trace amounts of water in reagents and/or solvents limit the molecular weight,
and to confirm this assumption, we carried out the polycondensation of dried
BDSA and NTDA, but used reagents and solvents as received. As shown in
Figure 51 b, the 1H NMR spectrum of the as-synthesized SPI is consistent with

that reported in the literature.158'214

116

“:20 DMSO 9

 

 

 

 

 

 

 

 

 

 

 

 

 

ppm
('3)
a b xzm-cresol
e x
......,....,....,....,....,
9.0 8.5 8.0 7.5 7.0 6.5
ppm

Figure 51. 500 MHz 1H NMR spectra of the triethylammonium salt form of SPI
synthesized with (a) dried starting materials, reagents, and solvents; and (b)
dried starting materials, but reagents and solvents used as received.

117

FT-IR spectroscopy confirms the triethylammonium salt form of SPI
(Figure 52). The spectrum shows strong, broad, O—H stretching bands at 3435
cm'1 due to water absorbed by the hygroscopic polymer. Bands between 3200
cm'1 and 3000 cm‘1 were assigned to aromatic C—H stretching while bands from
3000 cm'1 to 2900 cm'1 resulted from aliphatic C—H stretching of the
triethylammonium moiety. The characteristic C=O symmetric and antisymmetric
stretching bands at 1718 cm'1 and 1680 cm", and the C—N—C bending mode at
1345 cm'1 confirmed the presence of the naphthalenic imide. In addition to the
characteristic peaks at high wavenumbers, the SPI structure was supported by
the S=O antisymmetric and S=O symmetric stretching bands at 1185 cm’1 and
1035 cm", respectively. The absence of the characteristic peak at 1780 cm'1 for a

polyamic acid indicates a high degree of imidization.

118

 

 

1680
1718 1345 1185
030-
1035

8
c 0.28 -
N
.0
L
O
3
< 0.26 —

024-

 

 

 

4000 3500 3000 2500 2000 1500 1000 500
Wavenumber (cm'1)

Figure 52. FTIR spectrum of SPI in its triethylammonium salt form. The sample
was made into a pellet with KBr.

Thermogravimetric analysis (TGA) was performed to investigate the
thermal stability of SPI in its triethylammonium salt form. As presented in Figure
53, there are two major weight loss events. The first weight loss (~30 %)
commenced at ~200 °C and continued to 400 °C, which is likely due to
desulfonation.214 The weight loss observed above 400 °C corresponds to the
oxidative degradation of the aromatic backbone of the polymer. The onset for
weight loss is at ~200 °C, which demonstrates that the thermal stability of the SPI

polymer is suitable for high temperature PEM applications.

119

100

 

 

 

 

 

80-
3‘
g 60b
”5
£3
9
E40-
.9?
a)
2
20-—
0 l l 1 I
0 200 400 600 800

Temperature (°C)

Figure 53. Thermogravimetric analysis of SPI (triethylammonium salt), run at a
heating rate of 10 °C/min in air.

Fabrication and characterization of PEMs from the blending of SPI
(triethylammonium salt) and PVDF

The triethylammonium salt of SPI was soluble in aprotic solvents (NMP
and DMSO) and m-cresol. Blend membranes were obtained by solution-casting a
DMSO mixture of the SPI (triethylammonium salt) and PVDF onto flat glass
surfaces. Once soaked in water, the dried membrane samples were easily
peeled from the glass substrate, and after acidification, the membranes were
tough and yellow. Table 3 summarizes the composition and experimental results

from the blend membranes.

120

Table 3. Summary of SPI/PVDF membrane characteristics.

 

 

 

 

 

 

 

 

 

PEMs lEC Conductivity Swelling Water uptake
(SPI wt%) (meqlg) at rt (mSIcm) ratio (%) (wt%)
0 0 a 0 <3 0
10 0.26 a 0 <3 2
20 0.51 a 0 <3 6
30 0.77 a 7 <3 1 1
40 1 .03 a 41 <3 25
50 1 .29 a 75 <3 47
Nafion 117" 0.91 ‘5 62 1815 21

 

 

 

 

 

 

a IEC values were calculated from the SPI repeating unit (Mn = 778 g/mol) and
the composition of blend membranes.
b Nafion 117 was included for comparison.

A key parameter of PEMs is water absorption. Membranes require water
to dissociate $03H groups and act as “vehicles” to transport protons from the
anode to the cathode; however, too much water usually results in severe swelling
and poor mechanical strength. Ideally, the water uptake facilitates high proton
conductivity without sacrificing a membrane’s mechanical properties. As shown
in Figure 54, water uptake by SPI blends increased linearly up to about 28 wt%
SPI content. However, beyond 30 wt% SPI, water uptake increased drastically,
indicating a change in the membrane morphology.215 Since PVDF is hydrophobic,
the SPI phase is primarily responsible for water absorption. When the SPI

content in blend membranes is low, most sulfonic acid groups are isolated in

121

ionic clusters distributed throughout the continuous PVDF matrix, and water may
be accessible to only few clusters. As the SPI content increases, accessible ionic
clusters either increase in size or link to adjacent clusters, increasing the
measured water uptake. When the SPI content in blend membranes exceeded
~30 wt%, the membrane morphology of isolated ionic clusters in PVDF
transformed to a bicontinuous network of water swollen channels in a PVDF
matrix. Above the apparent percolation threshold, water uptake increased linearly,
but at a higher rate. Not surprisingly, a pure PVDF membrane (0 wt% SPI) had

zero water absorption.

122

50

 

 

 

f

40-
$3
1% 30-
0)
ft; ’A
Q l.
3 20- ,
3
(U
3 10- ,4"

""" ".‘.-"'
. vvvv ‘ '.'.
o-ir'

 

SPIs content in membranes (wt%)

Figure 54. Water uptakes of blend membrane as a function of the SPI content.
The membrane water uptake was measured gravimetrically at room temperature.
(The dotted line is a guide to the eye)

The ion exchange capacity (lEC), expressed as mmol acid/g of sample,
defines the number of charge carriers in a membrane and is directly correlated to
a membrane’s ability to absorb and retain water. As demonstrated in Figure 55,
the water uptake is linearly correlated with the lEC values of the membranes up
to about 0.78 meq/g (~28 wt% SPI). However, beyond this value, water uptake

increased more rapidly, indicating a change in the membrane morphology. This

sharp increase in water uptake is consistent with a percolation phenomenon

123

where isolated hydrophilic domains transform to a continuous network in the

PVDF matrix.5°'2°6'215

 

 

 

 

50
.A'
404
:\o‘ f.
E 30«
g A
E .
g 20-
31:»
(U
3 10- ‘
.... "'."".
.... “"'
0~A""
l ' l ' l ‘ T
0.0 0.4 0.8 1.2
lEC (meqlg)

Figure 55. Water uptakes as a function of membrane IEC. The membrane water
uptake was measured gravimetrically at room temperature. (The dotted lines are
a guide to the eye)

Proton conductivity is one of the most important parameters for evaluating
the performance of PEMs as well as the practicability of fuel cells. It is widely
accepted that the proton conductivity of PEMs is strongly related to the number
of charge carriers (H‘), so an ideal membrane would have a high sulfonic acid

content without sacrificing the PEMs’ mechanical properties. The conductivity

tests were conducted with the cell submersed in Milli-Q water (100% RH), and

124

proton conductivities were measured for the series of blend membranes as a
function of the weight fraction of SPI (Figure 56). Pure PVDF membranes (0 wt%
SPI) and blend membranes with <30 wt% SPI fraction had no measurable proton
conductivity, which is consistent with isolated ionic clusters. However, beginning
at 30 wt% SPI, the presumed percolation threshold, proton conductivity
increased linearly from 7 to 75 mS/cm at 50 wt% SPI.

The proton conductivity of Nafion 117 was measured as a control for
conductivity experiments. At room temperature, the conductivity of Nafion 117
was 62 mS/cm, which is reasonably close to the reported conductivity data (60
mS/cm) for Nafion 117 at room temperature.207 The proton conductivity of a 50
wt% SPI blend PEM was higher than Nafion 117 at both room temperature and
60 °C. The high conductivity of the blend membranes is consistent with a
continuous proton conducting SPI phase separated from the hydrophobic PVDF
analogous to Nafion 117.11

The proton conductivity of blend membranes and Naflon 117 were
relatively low at room temperature but increased rapidly as the temperature was
raised to 60 °C. The elevation of temperature might have stimulated the
dynamics of proton transport and structural reorganization, which favors fast
proton conduction.216 Moreover, temperature increase might activate both the
proton diffusion and molecular diffusion leading to improved proton conductivities

at high temperatures.216

125

160

 

E 120-

a, I
.5,

.é‘

.5 80‘

0

3

'2 O
O

0

f), 404

2

D.

 

 

 

0 t ? ? l l l l
0 10 20 30 40 50 60
SPIs content in membranes (wt%)

Figure 56. Blend membrane proton conductivity as a function of SPI content for
composite PEMs at 100 % RH and room temperature (O) and 60 °C (A). Data
for Nafion 117 at room temperature (0) and at 60 °C (I) was shown for
comparison. (The lines are a guide to the eye)

Proton conductivity is strongly correlated to the IE0 and generally the
proton conductivity of PEMs increase with the IEC. Figure 57 shows the proton
conductivity as a function of membrane IEC. As with the water uptake data, there

is an apparent percolation threshold at ~ 0.7 meq/g, followed by a linear increase

in conductivity with increasing IEC.

126

 

80

O)
O
l

404

Proton conductivity (mS/cm)
N
O

 

 

 

i

A A
0.4 0.8 1.2
lEC (meqlg)

.0
0

Figure 57. Membrane proton conductivity (room temperature, 100 °/o RH) as a
function of IEC. (The line is a guide to the eye)

It was of great interest to investigate the proton conductivity vs water
uptake, as water molecules were needed as the mobile phase to facilitate proton
transport. As demonstrated in Figure 58, there was an almost linear increasing
relationship between proton conductivity and water uptake above ~10 wt%, which

was assumed to be the percolation threshold.

127

 

80

 

 

 

A
E 60-
(I)
E
.2
.15. 40- A
‘0'
3
'0 1
C
8
c 20"
2
2
0- A
01AA A
0 10 20 30 40 50
Water uptake (wt%)

Figure 58. Blend membrane proton conductivity (room temperature, 100 °/o
humidity) as a function of water uptake. (The lines are a guide to the eye)

The nature of acid groups, chemical composition, and membrane
microstructure all contribute to the properties of PEMs. To obtain direct
morphologic information, ~100-nm-thick cross-sectional slices of blend
membranes were analyzed by TEM. Figure 59 displays images of SPI/PVDF
blend membranes with different blending ratios. After staining with Ag”, the SPI
appears as dark regions, while the matrix material, PVDF, is bright. A phase-
separated morphology was seen for all compositions. Blend membranes with 10
- 30 wt% SPI show randomly distributed ionic domains; the domain size seems
constant, but the areal density of domains increases. TEM images of membranes

with >30 wt% SPI content show ionic domains interconnected into a continuous

128

 

network. The images in Figures 59c and Figures 59d provide direct evidence
for a percolation phenomenon near 30 wt% SPI. The TEM image of a membrane
with 40 wt% SPI content shows an apparent bicontinuous morphology with
conducting paths for proton transport, which is consistent with the conductivity
and water uptake studies. The ionic domains have sizes of 30-100 nm. A finer
phase-separated morphology was found in the 50 wt% SPI blend membrane,
and the domain size dropped to 20-40 nm. Scattered black dots seen in most
images are likely silver oxide clusters generated through the photolysis of silver
nitrate, which was applied for staining blend membranes. While the cross-
sectional TEM images provide direct evidence for a phase-separated structure in
dried membranes, water saturated membranes might have different ionic domain

sizes or domain connections.

129

 

 

 

Figure 59. Cross-sectional TEM images of blend membranes with different SPI
content: (a) 10 wt%; (b) 20 wt%; (c) 30 wt%; (d) 40 wt%; and (e) 50 wt%.

As PEMs are used as not only a proton conductor but also a separator to
prevent the mixing of fuels with the oxidant, PEMs cannot have pin holes, which
will allow permeation of fuels. We used SEM to obtain the surface microsthcture
of blend membranes that had been kept in water for over a week, and then dried

(Figure 60). No holes were seen for any compositions. The SEM image of the

131

membrane with 10 wt% SPI content shows a homogeneous surface structure.
The roughness of the blend membrane surfaces increases with the SPI content.
The image in Figure 60d shows a spherical morphology in a 50 wt% SPI
membrane. It is noteworthy that the SEM images were taken from the blend
membranes which had been kept in water for over a week, implying that blend

membranes are hydrolytically stable.

132

 

 

133

 

Figure 60. Surface SEM images of blend membranes with different SPI content:
(a) 10 wt%; (b) 20 wt%; (c) 30 wt%; (d) 40 wt%; and (e) 50 wt%.

Ideally, PEMs should have high proton conductivity and minimal
membrane swelling. Membrane dimensional stability was characterized by
measuring the dimensional changes between the dry and water saturated states.
Membranes that swell excessively can detach from the electrode surfaces as
most electrodes do not swell, resulting in mechanical stress.208 As illustrated in

Table 3, Nafion 117 equilibrated in water increased 18 % in length compared to

134

the dry film. The severe expansion of Nafion 117 explains its poor performances
in direct methanol fuel cells, as the permeability of Nafion to methanol is related
to the high swelling of Nafion in water. However, all of the blend membranes
swelled <3 %. The very low water swelling behavior may be due to the rigid
matrix of crystalline PVDF, which suppressed the water swelling of SPI in

hydrated membranes.

Conclusion

A SPI with a 100 % degree of sulfonation was prepared by the direct
polycondensation reaction of a naphthalenic dianhydride monomer and a
sulfonated diamine monomer, which avoided the disadvantages associated with
“post-sulfonation” method. The chemical structure of the SPI was verified by 1H
NMR and FT-IR spectra. A series of tough and smooth proton conducting
membranes were prepared using SPI to blend with PVDF followed by
acidification, which revealed comparable or even higher proton conductivity than
Nafion 117 at the same testing conditions. The proton conductivity and water
uptake of blend membranes were proportional to the SPI content above the
percolation threshold, which was confirmed by TEM cross-sectional images.
Membranes with higher SPI content also showed a coarser surface morphology.
The blend membranes showed a high dimensional stability which might be due to
the rigid hydrophobic PVDF matrix. This blend membrane system is more cost

effective compared to Nafion due to much easier synthesis and processing.

135

Experimental Section
Materials

Unless otherwise specified, all chemicals and solvents were ACS reagent
grade and used as received from commercial suppliers without further
purification. Dimethyl sulfoxide (DMSO) was dried over activated 4 A molecular
sieves, and triethylamine (TEA) and m-cresol were distilled before use. 4,4”-
Diamino-biphenyl 2,2’-disulfonic acid (BDSA), received from TCI, was purple.
BDSA was dissolved in an aqueous TEA and then precipitated upon addition of
1M HCI BDSA. The process was repeated until BDSA was obtained as a white

solid, and then it was dried under vacuum at 70 °C overnight prior to use. Nafion
117, PVDF (MW = 534,000 Da), and 1,4,5,8-naphthalene tetracarboxylic

dianhydride (NTDA) were purchased from Sigma-Aldrich. NTDA and benzoic
acid were dried under vacuum at 170 °C and 60 °C overnight, respectively, prior

to the polycondensation reaction.

Synthesis of sulfonated polyimide (SPI)

Under N2, BDSA (0.69 g, 2mmol), TEA (0.7 mL, 5 mmol), and m-cresol (6
mL) were added to a 100 mL three-neck round bottom flask and were vigorously
stirred with a mechanical stirrer. After the BDSA was completely dissolved,
NTDA (0.539, 2 mmol) and benzoic acid (0.34 g, 2.8 mmol) were added. The
reaction system was stirred at 80 °C for 4 h, and then additional m-cresol (20 mL)
was added to dilute the viscous mixture. The reaction temperature was increased

to 180 °C. After 14 h, the heating was discontinued, 100 mL of DMSO was added,

136

 

and the system was allowed to cool. The mixture was poured into 500 mL of
ethyl acetate to precipitate the polymer. The crude polymer was isolated by
filtration and was re-precipitated twice by dissolving the polymer in 100 mL of
DMSO and precipitating into 500 mL of ethyl acetate. Finally, the sulfonated

polyimide (SPI) was dried under vacuum at 80 °C overnight (1.1 9).

Film casting and membrane acidification

Blend membranes were fabricated by the solvent-casting method. This
procedure describes the preparation of 10 mg membranes. Predetermined
amounts of PVDF and SPI (triethylammonium salt) were dissolved in DMSO in
separate vials. Each solution was fixed at 0.02 g/mL, and the total mass of the
particles and PVDF was 10 mg. The two solutions were combined and stirred
until the system was homogeneous. Membranes were solution-cast by pouring
the mixture onto a clean glass plate, typically covering ~4x1 cm2 area, and
carefully dried with infrared heat at ~60 °C. After 3 hours, the membrane had
turned translucent, and the glass plate was transferred to a vacuum at 80 °C for
further drying overnight. The dried membranes (5-15 pm thick) were cut to size
(~3XO.5—1 cm2) and then removed from the substrate by immersing the plate in
water. The as-synthesized membranes were soaked in 1 M HCI for 48 h to
ensure that all sulfonate groups in the membranes were in their acidic form. The
membranes were then washed extensively with water, and prior to conductivity

tests, PEMs were stored in Milli-Q water.

137

 

Characterization.

1H NMR analyses were performed at room temperature in deuterated
dimethyl sulfoxide (DMSO) on a Varian UnityPlus 500 spectrometer at 500 MHz
with the residual solvent proton signals as chemical shift standards. FT -IR
spectra were acquired from a Mattson Galaxy 300 spectrometer purged with dry
nitrogen, with the signal averaging 128 scans at a resolution of 4 cm". All IR
samples were dried, mixed with KBr, ground, and then pressed into pellets.

Thermogravimetric analyses (TGA) were carried out in air on Perkin-Elmer
TGA 7 instruments at a heating rate of 10 °C/min. TGA samples were dried under
vacuum at 80 °C overnight prior to use. Dried samples were held at 120 °C in the
TGA apparatus for 30 min prior to initiating the run.

The microstructure of blend membranes was imaged using a JEOL-
100CX transmittance electron microscope. To prepare membrane cross-sections
for TEM analysis, membranes were immersed in 0.5 M AgNOa solution overnight
to stain the ionic domains by ion exchange of sulfonic acid groups for silver.
Samples were rinsed with water, and then dried under vacuum at 80 °C overnight.
The stained membranes were embedded in epoxy resin, dried at 50 °C for 24 h,
and then sectioned with an ultramicrotome to generate ~100 nm thick slices. The

slices were picked up with copper grids for TEM analysis.
Conductivity measurements

The proton conductivities of membranes were measured with an

alternating current (AC) impedance analyzer HP 4192A over the frequency range

138

from 5 Hz to 13 MHz. Membranes were placed between two Pt electrodes in a
home-made Teflon cell. The entire setup was kept in Milli-Q water at controlled
temperatures for at least 10 min to saturate membranes and then impedance
was measured in in-plane direction. From the Nyquist plot, the resistance of the
membrane was estimated, and then the conductivity of the membrane was
calculated using the electrodes distance and membrane cross-sectional area, as
shown in equation 1:

L

T. ‘1’

0'

Where ais the conductivity of the membrane, L represents the distance between
two electrodes, A stands for the membrane cross-sectional area, and R is the

bulk membrane resistance estimated from the Nyquist plot.

Water uptake

Water uptake of all membranes was measured gravimetrically. After
equilibration in Milli-Q water for two days at room temperature, membranes were
removed from water, and weighed after the surface water was quickly removed
by a Kim-wipe. Membranes were then dried under vacuum at 50 °C overnight
and the weight of dried membrane was measured. The membrane water uptake

was calculated according to equation 2:

W .
Water uptake = W“ dry x100% (2)
Wary

 

where Wwet stands for the mass of wet membranes and Wdry represents the

weight of dried membranes.

139

Dimensional stability

Membranes were saturated in Milli-Q water for two days at room
temperature prior to measuring the dimensions of swollen membranes. After
drying under vacuum at 50 °C overnight, the dimensions of the dried membrane
were measured. The swelling ratio was calculated from the length (longest

dimension) of the membrane, as shown in equation 3:

l -I
Swelling ratio = —w—et-—dry- x100% (3)
[dry

where Iwet and [dry represent the lengths of wet and dried membranes,

respectively. The uncertainty in the measurement is <3%.

Theoretic ion exchange capacity (lEC)

The ion exchange capacity (lEC) of membranes was calculated from the
molecular weight of SPI repeating unit and SPI weight content in each membrane.
The lEC was calculated as shown in equation 4:

m +

lEC = -—H— (4)
Wary

where mH+ expresses the quantity of sulfonic acid groups.

140

Chapter 5 Synthesis and characterization of poly(4-styrenephosphonate)
for high temperature fuel cell application

Introduction

The proton exchange membrane (PEM) is one of the most important
components of a fuel cell. It must have high proton conductivity, dimensional
stability, controlled swelling in water, thermal and chemical stability, and low cost.
These requirements place conflicting demands on materials, such as a high
proton conductivity (favored by liquid-like materials) with dimensional stability,
making it difficult to realize all desired features in a single material. A significant
limitation of current PEMs is poor performance at elevated temperatures >100 °C
and under low humidity conditions.

The primary challenge is the design of polymer morphologies that provide
dynamic hydrogen bonding networks and high conductivities at low relative
humidities, with retention of the physical characteristics required of practical
membranes. Kreuer et al. have proposed criteria for proton conducting

membranes that operate under anhydrous conditions: 217

(1) a high concentration
of charge carriers, preferentially via self ionization (self dissociation), (2) a high
dielectric constant, (3) and a dynamic hydrogen bonding network that facilitates
proton transport without externally supplied water. A significant number of

d,2'218’ 223 most following the design

phosphonated membranes have been reporte
strategies previously used for sulfonic acids. There also have been several

attempts to construct phosphonic acid-based membranes that satisfy the criteria

141

elucidated by Kreuer but the conductivities of the resulting membranes were
modest?“225

Compared to sulfonic acid-based polymers, the synthesis of
phosphonated polymers is limited by the available phosphonation reagents and
more complex synthetic procedures. In addition, high proton conductivities
require a high degree of phosphonation because of the lower acidity of
phosphonic acid than sulfonic acid. There are several reports on proton
conductive polymers, in which phosphonic acid groups were connected to
aromatic rings through alkylene linkers"?226 However, the benzylic alkylene
spacer was very vulnerable to the oxidizing environment of fuel cells. Directly
tethering phosphonic acids from aromatic ring would provide materials with
greater thermal and chemical stability.

One route to phosphonated polymer materials is the direct polymerization
of phosphonic acid-functionalized monomers. Nevertheless, because of side
reactions, most attempts to polymerize phosphonic acid-functionalized
monomers have resulted with oligomeric products.227 Thus, post-polymerization
functionalization of suitable aromatic polymers is a more practical route. A few
procedures have been studied for introducing phosphonic functional groups into
aromatic polymers. The Friedel-Crafts reaction with phosphorus trichloride as the
starting material is one of the most common, efficient, and cost-effective
procedures for the direct phosphonation of aromatic rings, but the reaction has
poor selectivity and its application to polystyrene system led to cross-linked

products?27 There has been some success in creating direct aryl-phosphonate

142

functional groups in various polymers, such as poly(arylene ether),
poly(phenylene oxide), and poly(phenylsulfone) by using the Michaelis-Arbuzov

220,221,227 However

reaction and Pd(0)—catalyzed P—C coupling reaction.
purification of the polymer product and the expense of the Pd catalyst limits this
approach.

Recently Allcock and coworkers reported the successful incorporation of
phosphonic moieties into the aromatic rings of polyphosphazene by bromination,
lithiation, and finally reaction with diethyl chlorophosphate.228 In the present study,

we use a similar strategy to prepare poly(4-styrenephosphonate) from

brominated polystyrene and study its properties.

Results and discussion

The synthetic route to poly(4-styrene phosphonate) is shown in Scheme
11, The Cu(l) catalyze ATRP of 4-bromostyrene at 110 °C in xylenes was well-
controlled. GPC analysis of the poly(4-bromostyrene) product after precipitation
into methanol provided an Mn of 16,500 Da and a PDI of 1.01. To avoid inter— and
intramolecular coupling of the polymer,228 n-BuLi in THF at -78 °C was added
dropwise to the THF solution containing poly(4—bromostyrene) at -78 °C. The
addition of diethyl chlorophosphate provided the white phosphonated product
after precipitation from THF into water, and cyclohexane, respectively. GPC

analysis of the polymer showed a narrow molecular weight distribution (PDl=1.1)

and a molecular weight (Mn = ~16,200 Da) comparable to the parent poly(4-

143

bromostyrene), consistent with negligible backbone cleavage and side reaction in

the lithiation-phosphonation reaction.

Scheme 11. Synthesis route to poly(4-styrene phosphonate).

 

 

 

IP\
\ Etc-g— £8.13 1. EtO élOEt n
\CH 0 n-Butyllithium THF, -78 °c, 3 h
3
p > >
CuCl, xylene THF, -78 °C, 1h 2. rt, overnight

anbpy, 110 °C, 5 h

EtO’ "‘OEt
Br 0 '

The 1H NMR spectra of poly(4-bromostyrene) and poly(4-
styrenephosphonate) are shown in Figure 61. The broad peaks in poly(4-
bromostyrene) at ~6.3 and 7.2 ppm, associated with a para-substituted aromatic
ring, shifted to 6.4 and 7.4 ppm after the phosphonation reaction. The peaks at
1.3 and 4.1 ppm corresponding to the phosphate-ester groups in poly(4-
styrenephosphonate) were assigned to the methylene and methyl protons,
separately. Integrations of peaks at 4.1 and 6.4 ppm indicate that ~80 % of the
bromo groups have been phosphonated. Figure 62 shows the 31P NMR
spectrum of poly(4-styrenephosphonate). The single resonance at ~19 ppm rules

out intramolecular phosphonation

144

n
b
a

:30,

P C d
H3C HZCO’ "\OC H2CH 3
O c

 

 

 

 

 

 

 

 

 

0
PIN“

Figure 61. 1H NMR spectra of poly(4-bromostyrene) and poly(4-
styrenephosphonate). _

145

 

 

 

a
EtO’fi‘OEt
a O
30 20 10 0 -10 -20 ppm-30

Figure 62. 31P NMR spectrum of poly(4-styrenephosphonate).

The glass transition temperature (T9) is related to the polymer free volume,
and often correlates with transport of small molecules or ions. The DSC scan of
poly(4-bromostyrene) showed a glass transition temperature (T9) of 125 °C
(Figure 63). The phosphonated product, poly(4-styrenephosphonate), revealed a
T9 of 95 °C, which was ~30 °C lower than that of the parent polymer. The
decrease in T9 in the phosphonated product might be due to the increase of free

volume as the diethyl phosphonate group was bulkier than the bromo group.

146

 

 

 

 

 

1 Poly(4-bromostyrene)
B
O
E
‘5
o
:l:
Poly(4-styrenephosphonate)
50 1 00 1 50 ' 200

Temperature (°C)

Figure 63. DSC analyses of poly(4-bromostyrene) and poly(4-
styrenephosphonate). DSC samples were run in a nitrogen atmosphere at a
heating rate of 10 °C/min.

Figure 64 shows FT-IR spectra that confirm the phosphonation reaction.
Both poly(4-bromostyrene) and poly(4-styrenephosphonate) showed the
expected absorption bands at 3200-3000 cm'1 and 3000-2800 cm'1 from aromatic
and aliphatic C—H stretching vibrations, respectively. The two sharp absorption
peaks of C—Br at 1010 and 1073 cm'1 are characteristic peaks of poly(4-

bromostyrene). New absorption peaks due to the diethyl phosphonate groups are

found in the spectrum of poly(4-styrenephosphonate). Absorption peaks at 1025

147

and 1052 cm'1 were both attributed to P—0 and the peak at 1246 cm'1 to P=O,

which confirmed the happening of the phosphonation reaction to poly(4-

 

  
 

 

bromostyrene).
1 025
n
0
U
E H3CHZCO’fi‘OCHZCH3
.9 O
I-
O
In
.0
<
n
m

 

 

 

3000 2500 2000 1500 1000 500
Wavenumber (cm'1)

Figure 64. FTIR spectra of poly(4-bromostyrene) and poly(4-styrene
phosphonate). Samples were made into pellets with KBr.

Figure 65 compares the thermal stability of the parent polymer and the
phosphonated product. The TGA trace of poly(4-bromostyrene) indicated a
dramatic mass loss at around 250 °C, which is likely due to depolymerization,

and a minor weight loss occurred at 420 C, probably related to some crosslinking

148

during degradation. The TGA curve of poly(4-styrenephosphonate) exhibited
three mass loss events. The initial mass loss started at about 230 °C and
continued to about 270 °C, which might be attributed to the dissociation of diethyl
phosphonate groups into free acid and ethylene. The second stage occurred
between 270 and 295 °C due to the cross-linking of phosphonic acids in terms of
P-O—P linkage forming anhydrides.220 The last mass loss beginning at around

420 °C might result from the degradation of polymer backbone.

100

 

 

-h 0') 00
O O O
r r r

Weight retention (%)

N
O
1

 

 

 

O L l 4 l n I

200 400 600 800
Temperature (°C)

 

Figure 65. Thermogravimetric analysis of (a) poly(4-bromostyrene); and (b)
poly(4-styrenephosphonate). All samples were run in an air atmosphere at a
heating rate of 10 °C/min.

149

I _

Conclusion

Poly(4-styrenephosphonate) was successfully prepared from poly(4-
bromostyrene) via lithiation and subsequent phosphonation. The as-synthesized
polymer was soluble in common organic solvents, such as THF. This approach
generated phosphonated polymers with a high degree of phosphonation and no
apparent side reactions. The phosphonated polymer might find its application in

fuel cell applications at low humidity conditions.

Experimental Section
Materials

Unless othenlvise specified, all chemicals and solvents were ACS reagent
grade and used as received from commercial suppliers without further
purification. N-Butyllithium (2.5 M in hexanes), 4,4’-dinonyI-2,2’-dipyridyl
(anbpy), and 4-bromostyrene were obtained from Sigma—Aldrich. 4-
Bromostyrene was passed through an activated basic alumina column followed
by distillation prior to use. Tetrahydrofuran (THF) was distilled over sodium and

benzophenone.

Characterization.

1H NMR (500 MHz) spectroscopy was performed at room temperature in
CDCI3 solvent on a Varian UnityPlus 500 spectrometer with the residual solvent
proton signals as chemical shift standards. 31P NMR (121 MHz) spectra were

obtained on a Varian UnityPlus 300 spectrometer. IR spectra were acquired from

150

a Mattson Galaxy 300 spectrometer purged with dry nitrogen, with the signal
averaging 128 scans at a resolution of 4 cm". All IR samples were dried, ground
and mixed with KBr and then pressed into pellets.

Thermogravimetric analyses (TGA) were carried out in air on Perkin-Elmer
TGA 7 instruments at a heating rate of 10 °C lmin. TGA samples were dried
under vacuum at 80 °C overnight prior to test. Dried samples were held at 120 °C
in the TGA apparatus for 30 min before the heating proceeded. Differential
scanning calorimetry (DSC) measurements were performed using a TA DSC
0100 with a heating/cooling rate of 10 °C /min under nitrogen.

Polymer molecular weights were measured by gel permeation
chromatography (GPC) at 35 °C using two PLgel 10p mixed-B columns in series
with THF as the eluting solvent at a flow rate of 1 mL/min. A Waters 2410
differential refractometer was used as the detector. Mono—disperse polystyrene

standards were used to calibrate the molecular weights.

Synthesis of poly(4-bromostyrene) by ATRP

To a 10 mL Schlenk flask equipped with a magnetic stirrer, were added 4-
bromostyrene (4.5 mL, 34.4 mmol), ethyl 2-bromoisobutyrate (50 pL, 0.34 mmol),
anbpy (0.42 g, 1.02 mmol), and xylenes (1.5 mL). After four freeze-pump-thaw
cycles, the flask was backfilled with nitrogen and CuCI (34 mg, 0.34 mmol) was
added. The mixture was degassed by three additional freeze-pump—thaw cycles,
and then stirred at 110 °C for 5 h in a nitrogen atmosphere. The polymerization

system was quickly cooled to room temperature, exposed to air, and diluted with

151

 

5 mL of xylenes. Precipitation from methanol gave poly(4-bromostyrene) as a
white solid. The polymer was redissolved in THF and precipitated by the addition
of methanol to remove impurities. Finally, the poly(4-bromostyrene) was dried

under vacuum at 80 °C overnight.

Synthesis of poly(4-styrenephosphonate)

Under a nitrogen purge, n-butyllithium (5 mL, 12.5 mmol) was added
dropwise to 250 mL of THF in a 1 L three-neck round bottom flask equipped with
a magnetic stirrer and equilibrated at -78 °C. A THF solution (150 mL) of poly(4-
bromostyrene) (0.32 g, 1.75 mmol repeating unit) was cooled to -78 °C and then
transferred to the n-butyllithium solution via cannula. The mixture was held at
-78 °C for 1 h and then diethyl chlorophosphate (3.9 mL, 27 mmol) was injected
dropwise. After an additional 3 h at -78 °C, the reaction mixture was allowed to
warm to room temperature overnight. The crude polymer was collected by
filtering the reaction mass, condensing the THF solution, and then precipitating
into water. The polymer was purified by precipitating from THF into water and
cyclohexane, respectively. Finally, the solid polymer was dried under vacuum at

50 °C overnight.

152

(1)

(2)

(3)
(4)
(5)
(6)

(7)
(8)

(9)
(10)

(11)
(12)

(13)

(14)

(15)

(16)
(17)
(18)

References

Einsla, B. R.; Kim, Y. 8.; Hickner, M. A.; Hong, Y. T.; Hill, M. L.; Pivovar,
B. S.; McGrath, J. E. J. Membr. Sci. 2005, 255, 141-148.

Hickner, M. A.; Ghassemi, H.; Kim, Y. S.; Einsla, B. R.; McGrath, J. E.
Chem. Rev. 2004, 104, 4587-4611.

Kerres, J. A. J. Membr. Sci. 2001, 185, 3-27.
’Mauritz, K. A.; Moore, R. B. Chem. Rev. 2004, 104, 4535-4585. _
Hsu, W. Y.; Gierke, T. D. Macromolecules 1982, 15, 101-105.

Gierke, T. D.; Munn, G. E.; Wilson, F. C. J. Polym. Sci.’ Pt. B-Polym.
Phys. 1981, 19, 1687-1704.

Hsu, W. Y.; Gierke, T. D. J. Membr. Sci. 1983, 13, 307-326.

Rubatat, L.; Rollet, A. L.; Gebel, G.; Diat, O. Macromolecules 2002, 35,
4050-4055.

Rubatat, L.; Gebel, G.; Diat, O. Macromolecules 2004, 37, 7772-7783.
Rollet, A. L.; Diat, 0.; Gebel, G. J. Phys. Chem. B 2002, 106, 3033-3036.
Schmidt-Rohr, K.; Chen, Q. Nat. Mater. 2008, 7, 75-83.

Roche, E. J.; Pineri, M.; Duplessix, R.; Levelut, A. M. J. Polym. Sci. Pt. 3-
Polym. Phys. 1981, 19, 1-11.

Choi, P.; Jalani, N. H.; Datta, R. J. Electrochem. Soc. 2005, 152, E123-
E130.

Zhao, C. J.; Li, X. P.; Wang, Z.; Dou, Z. Y.; Zhong, S. L.; Na, H. J. Membr.
Sci. 2006, 280, 643-650.

Chuy, C.; Ding, J. F.; Swanson, E.; Holdcroft, S.; Horsfall, J.; Lovell, K. V.
J. Electrochem. Soc. 2003, 150, E271-E279.

Zaluski, C.; Xu, G. Macromolecules 1994, 27, 6750-6754.
Jagur—Grodzinski, J. Polym. Adv. Technol. 2007, 18, 785-799.
Flint, S. D.; Slade, R. C. T. Solid State Ion. 1997, 97, 299-307.

153

(19)

(20)
(21)

(22)

(23)

(24)

(25)

(25)

(27)

(28)

(29)
(30)

(31)

(32)

(33)

Pereira, F.; Valle, K.; Belleville, P.; Morin, A.; Lambert, S.; Sanchez, C.
Chem. Mat. 2008, 20, 1710-1718.

Tang, H. L.; Pan, M. J. Phys. Chem. C 2008, 112, 11556-11568.

Yuan, J. J.; Pu, H. T.; Yang, 2. L. J. Polym. Sci. Pol. Chem. 2009, 47,
2647-2655.

Jin, Y. G.; Qiao, S. 2.; Zhang, L.; Xu, Z. P.; Smart, 8.; da Costa, J. C. 0.;
Lu, G. Q. J. Power Sources 2008, 185, 664-669.

Lin, Y. F.; Yen, C. Y.; Ma, C. C. M.; Liao, S. H.; Lee, C. H.; Hsiao, Y. H.;
Lin, H. P. J. Power Sources 2007, 171, 388-395.

Yen, C. Y.; Lee, C. H.; Lin, Y. F.; Lin, H. L.; Hsiao, Y. H.; Liao, S. H.;
Chuang, C. Y.; Ma, C. C. M. J. Power Sources 2007, 173, 36-44.

Kannan, A. G.; Choudhury, N. R.; Dutta, N. K. J. Membr. Sci. 2009, 333,
50-58.

Chalkova, E.; Pague, M. B.; Fedkin, M. V.; Wesolowski, D. J.; Lvov, S. N.
J. Electrochem. Soc. 2005, 152, A1035-A1040.

Chalkova, E.; Fedkin, M. V.; Wesolowski, D. J.; Lvov, S. N. J.
Electrochem. Soc. 2005, 152, A1742-A1747.

Lepiller, C.; Gauthier, V.; Gaudet, J.; Pereira, A.; Lefevre, M.; Guay, D.;
Hitchcock, A. J. Electrochem. Soc. 2008, 155, 870-378.

Jalani, N. H.; Dunn, K.; Datta, R. Electrochim. Acta 2005, 51, 553-560.

Tang, H. L.; Pan, M.; Jiang, S. P.; Wan, Z. H.; Yuan, R. Z. Colloid Surf.
A-Physicochem. Eng. Asp. 2005, 262, 65-70.

Mu, S. C.; Tang, H. L.; Wan, Z. H.; Pan, M.; Yuan, R. Z. Electrochem.
Commun. 2005, 7, 1143-1147.

Pan, M.; Tang, H. L.; Jiang, S. P.; Liu, Z. C. J. Electrochem. Soc. 2005,
152, A1081-A1088.

Uchida, H.; Mizuno, Y.; Watanabe, M. J. Electrochem. Soc. 2002, 149,
A682-A687.

Uchida, H.; Ueno, Y.; Hagihara, H.; Watanabe, M. J. Electrochem. Soc.
2003, 150, A57-A62.

Zhang, W. J.; Li, M. K. S.; Yue, P. L.; Gao, P. Langmuir 2008, 24, 2663-
2670.

154

(36)
(37)

(38)
(39)
(40)

(41)

(42)
(43)
(44)
(45)
(45)

(47)

(48)

(49)
(50)

(51)

(52)

(53)

Sangeetha, D. Eur. Polym. J. 2005, 41, 2644-2652.

Chuy, C.; Basura, V. l.; Simon, E.; Holdcroft, S.; Horsfall, J.; Lovell, K. V.
J. Electrochem. Soc. 2000, 147, 4453-4458.

Lee, W.; Kim, H.; Lee, H. J. Membr. Sci. 2008, 320, 78-85.
Brijmohan, S. 3.; Shaw, M. T. Polymer 2006, 47, 2856-2864.

Gubler, L.; Prost, N.; Gursel, S. A.; Scherer, G. G. Solid State Ion. 2005,
176, 2849-2860.

Gasa, J. V.; Boob, 8.; Weiss, R. A.; Shaw, M. T. J. Membr. Sci. 2006,
269, 177-186.

Ding, J. F.; Chuy, C.; Holdcroft, S. Macromolecules 2002, 35, 1348-1355.
Ding, J. F.; Chuy, C.; Holdcroft, 8. Adv. Funct. Mater. 2002, 12, 389-394.
Ding, J. F.; Chuy, C.; Holdcroft, S. Chem. Mat. 2001, 13, 2231-2233.

Shi, Z. Q.; Holdcroft, S. Macromolecules 2005, 38, 4193-4201.

Sherazi, T. A.; Ahmad, 8.; Kashmiri, M. A.; Kim, D. S.; Guiver, M. D. J.
Membr. Sci. 2009, 333, 59-67.

Li, D. R.; Chen, J. H.; Zhai, M. L.; Asano, M.; Maekawa, Y.; Oku, H.;
Yoshida, M. Nucl. Instrum. Methods Phys. Res. Sect. B-Beam Interact.
Mater. Atoms 2009, 267, 103-107.

Huang, Y. F.; Hwang, B. J.; Lin, C. W. J. Appl. Polym. Sci. 2009, 113,
342—350.

Choi, W. H.; Jo, W. H. J. Power Sources 2009, 188, 127-131.

Zhang, Z. C.; Chalkova, E.; Fedkin, M.; Wang, C. M.; Lvov, S. N.;
Komarneni, S.; Chung, T. C. M. Macromolecules 2008, 41 , 9130-9139.

Silva, A. L. A.; Takase, l.; Pereira, R. P.; Rocco, A. M. Eur. Polym. J.
2008, 44, 1462-1474.

Sahu, A. K.; Selvarani, G.; Bhat, S. D.; Pitchumani, S.; Sridhar, P.;
Shukla, A. K.; Narayanan, N.; Banerjee, A.; Chandrakumar, N. J. Membr.
Sci. 2008, 319, 298-305.

Rao, S. S.; tha, P. K.; Rath, S. K.; Palla, K. B.; Patri, M. J. Appl. Polym.
Sci. 2008, 110, 2041-2046.

155

(54)

(55)

(55)

(57)

(58)

(59)

(50)

(51)

(52)

(53)

(54)

(55)

(55)

(57)
(58)

(59)

(70)

Qiu, J. Y.; Zhao, L.; Zhai, M. L.; Ni, J. F.; Zhou, H. H.; Peng, J.; Li, J. 0.;
Wei, G. S. J. Power Sources 2008, 177, 617-623.

Fu, R. Q.; Woo, J. J.; Seo, S. J.; Lee, J. 8.; Moon, S. H. J. Membr. Sci.
2008, 309, 156-164.

Lin, C. W.; Huang, Y. F.; Kannan, A. M. J. Power Sources 2007, 164,
449-456.

Elabd, Y. A.; Napadensky, E.; Walker, C. W.; Winey, K. I.
Macromolecules 2006, 39, 399-407.

Miyatake, K.; Zhou, H.; Matsuo, T.; Uchida, H.; Watanabe, M.
Macromolecules 2004, 37, 4961-4966.

Zheng, G. W.; Gong, C. L.; Wen, 8.; Xie, X. L. Acta Phys.-Chim. Sin.
2008, 24. 855-860.

Wang, L.; Yi, B. L.; Zhang, H. M.; Xing, D. M. J. Phys. Chem. B 2008,
112, 4270-4275.

Shu, Y. C.; Chuang, F. 8.; Tsen, W. C.; Chow, J. D.; Gong, C.; Wen, S. J.
Appl. Polym. Sci. 2008, 108, 1783-1791.

Zhang, G. M.; Guo, X. X.; Fang, J. H.; Chen, K. C.; Okamoto, K. I. J.
Membr. Sci. 2009, 326, 708-713.

Sutou, Y.; Yin, Y.; Hu, 2.; Chen, 8.; Kita, H.; Okamoto, K.; Wang, H.;
Kawasato, H. J. Polym. Sci. Pol. Chem. 2009, 47, 1463-1477.

Chen, X. B.; Chen, P.; Okamoto, K. I. J. Appl. Polym. Sci. 2009, 112,
3560-3568.

Chen, K. C.; Chen, X. B.; Yaguchi, K.; Endo, N.; Higa, M.; Okamoto, K. I.
Polymer 2009, 50, 510-518.

Zhu, X. L.; Pan, H. Y.; Liang, Y. F.; Jian, X. G. Eur. Polym. J. 2008, 44,
3782-3789.

Zhao, L.; Huang, Y. 0.; Song, Y. J.; Piao, Y. M. e-Polymers 2008.

Loredo, D. E. 8.; Paredes, M. L.; Sena, M. E. Mater. Lett. 2008, 62,
3319-3321.

Li, W. M.; Cui, Z. M.; Zhou, X. C.; Zhang, 8. B.; Dai, L.; Xing, W. J.
Membr. Sci. 2008, 315, 172-179.

Li, N. W.; Cui, Z. M.; Zhang, S. 8.; Li, S. H. J. Polym. Sci. Pol. Chem.
2008, 46, 2820-2832.

156

(71)
(72)

(73)

(74)

(75)

(75)

(77)

(78)

(79)

(80)

(81)

(82)

(83)

(84)

(85)

(85)
(87)

Lee, C. H.; Park, C. H.; Lee, Y. M. J. Membr. Sci. 2008, 313, 199-206.

Deligoz, H.; Vatansever, 8.; Oksuzomer, F.; Koc, S. N.; Ozgumus, 8.;
Gurkaynak, M. A. Polym. Adv. Technol. 2008, 19, 1792-1802.

Deligoz, H.; Vantansever, 8.; K00, S. N.; Oksuzomer, F.; Oezgumus, 8.;
Gurkaynak, M. A. J. Appl. Polym. Sci. 2008, 110, 1216-1224.

Chen, X. B.; Yin, Y.; Chen, P.; Kita, H.; Okamoto, K. I. J. Membr. Sci.
2008. 313, 106-119.

Wang, L.; Meng, Y. Z.; Wang, S. J.; Zhu, G. M. Chem. J. Chin. Univ.-
Chin. 2007, 28, 1408-1410.

Pinto, B. P.; Maria, L. C. D.; Sena, M. E. Mater. Lett. 2007, 61, 2540—
2543. '

Li, Y. H.; Jin, R. Z.; Wang, Z.; Cui, Z. M.; Xing, W.; Gao, L. X. J. Polym.
Sci. Pol. Chem. 2007, 45, 222-231.

Li, Y. H.; Jin, R. Z.; Cui, Z. M.; Wang, Z.; Xing, W.; Qiu, X. P.; Ji, X. L.;
Gao, L. X. Polymer 2007, 48, 2280-2287.

Li, N. W.; Cui, Z. M.; Zhang, S. B.; Xing, W. Polymer 2007, 48, 7255-
7263.

Lee, 8.; Jang, W.; Choi, 8.; Tharanikkarasu, K.; Shul, Y.; Han, H. S. J.
Appl. Polym. Sci. 2007, 104, 2965-2972.

Hu, Z. X.; Yin, Y.; Chen, S. W.; Yamada, 0.; Tanaka, K.; Kita, H.;
Okamoto, K. J. Polym. Sci. Pol. Chem. 2006, 44, 2862-2872.

Asano, N.; Miyatake, K.; Watanabe, M. J. Polym. Sci. Pol. Chem. 2006,
44, 2744-2748.

Nakano, T.; Nagaoka, 8.; Kawakami, H. Polym. Adv. Technol. 2005, 16,
753-757.

Yin, Y.; Fang, J. H.; Watari, T.; Tanaka, K.; Kita, H.; Okamoto, K. J.
Mater. Chem. 2004, 14, 1062-1070.

Li, W.; Fu, Y. Z.; Manthiram, A.; Guiver, M. D. J. Electrochem. Soc. 2009,
156, B258-8263.

Zhong, S. L.; Liu, C. G.; Na, H. J. Membr. Sci. 2009, 326, 400-407.

Zhang, G.; Fu, T. Z.; Shao, K.; Li, X. F.; Zhao, C. J.; Na, H.; Zhang, H. J.
Power Sources 2009, 189, 875-881.

157

(88)

(89)
(90)

(91)

(92)

(93)

(94)

(95)

(95)

(97)

(98)

(99)
(100)

(101)

(102)

(103)

(104)

Norddin, M.; Ismail, A. F.; Rana, D.; Matsuura, T.; Tabe, S. J. Membr. Sci.
2009, 328, 148-155.

Lin, C. K.; Kuo, J. F.; Chen, C. Y. J. Power Sources 2009, 187, 341-347.

Fu, T. Z.; Zhong, S. L.; Cui, Z. M.; Zhao, C. J.; Shi, Y. H.; Yu, W. 2.; Na,
H.; Xing, W. J. Appl. Polym. Sci. 2009, 111, 1335-1343.

Colicchio, I.; Wen, F.; Keul, H.; Simon, U.; Moeller, M. J. Membr. Sci.
2009, 326, 45-57.

Zhong, S. L.; Cui, X. J.; Fu, T. 2.; Na, H. J. Power Sources 2008, 180,
23-28.

Zhao, C.; Li, X.; Lin, H.; Shao, K.; Na, H. J. Appl. Polym. Sci. 2008, 108,
671-680.

Wang, J. L.; Song, Y. L.; Zhang, C.; Ye, Z. 8.; Liu, H.; Lee, M. H.; Wang,
D. H.; Ji, J. B. Macromol. Chem. Phys. 2008, 209, 1495-1502.

Seo, D. W.; Park, H. S.; Choi, S. W.; Jeong, Y. 6.; Hong, T. W.; Kim, W.
G. Polym. J. 2008, 40, 979-985.

Ramani, V.; Swier, 8.; Shaw, M. T.; Weiss, R. A.; Kunz, H. R.; Fenton, J.
M. J. Electrochem. Soc. 2008, 155, B532-8537.

Park, H. 8.; Sec, D. W.; Choi, S. W.; Jeong, Y. 6.; Lee, J. H.; II Kim, 0.;
Kim, W. G. J. Polym. Sci. Pol. Chem. 2008, 46, 1792-1799.

Li, X. F.; De Feyter, 8.; Vankelecom, I. F. J. J. Membr. Sci. 2008, 324,
67-75.

Kim, D. S.; Guiver, M. D. J. Polym. Sci. Pol. Chem. 2008, 46, 989-1002.

Hasani-Sadrabadi, M. M.; Emami, S. H.; Ghaffarian, R.; Moaddel, H.
Energy Fuels 2008, 22, 2539-2542.

Gasa, J. V.; Weiss, R. A.; Shaw, M. T. J. Membr. Sci. 2008, 320, 215-
223.

Fu, T. Z.; Wang, J.; Ni, J.; Cui, Z. M.; Zhong, S. L.; Zhao, C. J.; Na, H.;
Xing, W. Solid State Ion. 2008, 179, 2265-2273.

Fu, T. Z.; Cui, Z. M.; Zhong, S. L.; Shi, Y. H.; Zhao, C. J.; Zhang, G.;
Shao, K.; Na, H.; Xing, W. J. Power Sources 2008, 185, 32-39.

Do, K. N. T.; Kim, D. J. Appl. Polym. Sci. 2008, 110, 1763-1770.

158

(105)

(106)

(107)

(108)
(109)

(110)

(111)

(112)

(113)

(114)

(115)

(116)

(117)

(118)

(119)

(120)

Zhong, S. L.; Cui, X. J.; Cai, H. L.; Fu, T. Z.; Zhao, 0.; Na, H. J. Power
Sources 2007, 164, 65-72.

Zhao, C. J.; Wang, 2.; Bi, D. W.; Lin, H. D.; Shao, K.; Fu, T. Z.; Zhong, S.
L.; Na, H. Polymer 2007, 48, 3090-3097.

Li, X. F.; Zhang, G.; Xu, 0.; Zhao, C. J.; Na, H. J. Power Sources 2007,
165, 701 -707.

Jung, H. Y.; Park, J. K. Electrochim. Acta 2007, 52, 7464-7468.

Huang, M. Y.; Chen, H. Y.; Guo, J. 2.; Wang, Z. T.; Xu, L.; Wang, Y. X.
Acta Phys.-Chim. Sin. 2007, 23, 44-49.

Fu, Y. Z.; Manthiram, A.; Guiver, M. D. Electrochem. Solid State Lett.
2007, 10, B70-B73.

Cai, H. L.; Shao, K.; Zhong, S. L.; Zhao, C. J.; Zhang, G.; Li, X. F.; Na, H.
J. Membr. Sci. 2007, 297, 162-173.

Zhong, S. L.; Liu, C. G.; Dou, Z. Y.; Li, X. F.; Zhao, C. J.; Fu, T. 2.; Na, H.
J. Membr. Sci. 2006, 285, 404-411.

Zhao, C. J.; Lin, H. D.; Shao, K.; Li, X. F.; Ni, H. 2.; Wang, 2.; Na, H. J.
Power Sources 2006, 162, 1003-1009.

Wu, H. L.; Ma, C. C. M.; Li, C. H.; Lee, T. M.; Chen, C. Y.; Chiang, C. L.;
Wu, C. J. Membr. Sci. 2006, 280, 501-508.

Wang, Z.; Li, X. F.; Zhao, C. J.; Ni, H. 2.; Na, H. J. Power Sources 2006,
160, 969-976.

Mikhailenko, S. D.; Robertson, G. P.; Guiver, M. D.; Kaliaguine, S. J.
Membr. Sci. 2006, 285, 306-316.

Mecheri, B.; D'Epifanio, A.; Di Vona, M. L.; Traversa, E.; Licoccia, 8.;
Miyayama, M. J. Electrochem. Soc. 2006, 153, A463-A467.

Liu, B. J.; Robertson, G. P.; Guiver, M. D.; Sun, Y. M.; Liu, Y. L.; Lai, J.
Y.; Mikhailenko, 8.; Kaliaguine, S. J. Polym. Sci. Pt. B-Polym. Phys. 2006,
44, 2299-2310.

Fu, Y. Z.; Manthiram, A.; Guiver, M. D. Electrochem. Commun. 2006, 8,
1386-1390.

Xing, D. M.; Yi, B. L.; Liu, F. Q.; Fu, Y. Z.; Zhang, H. M. Fuel Cells 2005,
5, 406-411.

159

(121)

(122)
(123)

(124)
(125)

(126)

(127)

(128)

(129)

(130)

(131)
(132)

(133)

(134)

(135)
(136)

(137)

(138)

Li, X. F.; Liu, C. P.; Lu, H.; Zhao, C. J.; Wang, Z.; Xing, W.; Na, H. J.
Membr. Sci. 2005, 255, 149-155.

Karlsson, L. E.; Jannasch, P. J. Membr. Sci. 2004, 230, 61-70.

Yang, Y. 8.; Shi, Z. Q.; Holdcroft, S. Macromolecules 2004, 37, 1678-
1681.

Shahi, V. K. Solid State Ion. 2007, 177, 3395-3404.

Zheng, G. W.; Gong, C. L.; Wen, 8.; Zhou, H. B.; Xie, X. L. Acta Phys.-
Chim. Sin. 2009, 25, 575-582.

Ni, H. 2.; Wang, Z.; Chu, M. L.; Zhang, M. Y.; Na, H. Acta Polym. Sin.
2007, 868-874.

Wang, 2.; Ni, H. 2.; Zhao, C. J.; Zhang, M. Y.; Na, H. J. Appl. Polym. Sci.
2009, 112, 858-866.

Takamuku, S.; Akizuki, K.; Abe, M.; Kanesaka, H. J. Polym. Sci. Pol.
Chem. 2009, 47, 700-712.

Qi, Y. H.; Gao, Y.; Tian, S. H.; Hlil, A. R.; Gaudet, J.; Guay, D.; Hay, A. S.
J. Polym. Sci. Pol. Chem. 2009, 47, 1920-1929.

Matsumoto, K.; Higashihara, T.; Ueda, M. Macromolecules 2009, 42,
1161-1166.

Lee, J. K.; Li, W.; Manthiram, A. J. Membr. Sci. 2009, 330, 73-79.

Feng, 8. G.; Shang, Y. M.; Xie, X. F.; Wang, Y. Z.; Xu, J. M. J. Membr.
Sci. 2009, 335, 13-20.

Bai, Z.; Shumaker, J. A.; Houtz, M. D.; Mirau, P. A.; Dang, T. D. Polymer
2009, 50, 1463-1469.

Badami, A. 8.; Roy, A.; Lee, H. 8.; Li, Y. X.; McGrath, J. E. J. Membr. Sci.
2009, 328, 156-164.

Yang, J.; Li, Y. X.; Roy, A.; McGrath, J. E. Polymer 2008, 49, 5300-5306.

Paul, M.; Park, H. B.; Freeman, B. D.; Roy, A.; McGrath, J. E.; Riffle, J. S.
Polymer 2008, 49, 2243-2252.

Schuster, M.; Kreuer, K. 0.; Andersen, H. T.; Maier, J. Macromolecules
2007, 40, 598-607.

Chang, B. J.; Kim, D. J.; Kim, J. H.; Lee, S. B.; Joo, H. J. J. Membr. Sci.
2008, 325, 989-996.

160

(139)
(140)

(141)
(142)

(143)
(144)

(145)

(146)

(147)

(148)

(149)

(150)

(151)

(152)
(153)

(154)

(155)

(156)

Bae, B.; Miyatake, K.; Watanabe, M. J. Membr. Sci. 2008, 310, 110-118.

Zhang, X. P.; Liu, S. 2.; Yin, J. J. Polym. Sci. Pt. B-Polym. Phys. 2006,
44, 665-672.

Zhang, X. P.; Liu, S. 2.; Yin, J. J. Membr. Sci. 2006, 275, 119-126.

Liu, 0.; Kyriakides, 8.; Case, 8. W.; Lesko, J. J.; Li, Y. X.; McGrath, J. E.
J. Polym. Sci. Pt. B-Polym. Phys. 2006, 44, 1453-1465.

Olson, D. A.; Chen, L.; Hillmyer, M. A. Chem. Mat. 2008, 20, 869-890.

Yu, X.; Roy, A.; Dunn, 8.; Badami, A. 8.; Yang, J.; Good, A. 8.; McGrath,
J. E. J. Polym. Sci. Pol. Chem. 2009, 47, 1038-1051.

Ghassemi, H.; McGrath, J. E.; Zawodzinski, T. A. Polymer 2006, 47,
4132—4139.

Roy, A.; Hickner, M. A.; Yu, X.; Li, Y. X.; Glass, T. E.; McGrath, J. E. J.
Polym. Sci. Pt. B-Polym. Phys. 2006, 44, 2226-2239.

Wang, H.; Badami, A. 8.; Roy, A.; McGrath, J. E. J. Polym. Sci. Pol.
Chem. 2007, 45, 284-294.

Li, Y. X.; Roy, A.; Badami, A. 8.; Hill, M.; Yang, J.; Dunn, S.; McGrath, J.
E. J. Power Sources 2007, 172, 30-38.

Lee, H. 8.; Badami, A. S.; Roy, A.; McGrath, J. E. J. Polym. Sci. Pol.
Chem. 2007, 45, 4879-4890.

Roy, A.; Yu, X.; Dunn, 8.; McGrath, J. E. J. Membr. Sci. 2009, 327, 118-
124.

Badami, A. 8.; Lane, 0.; Lee, H. 8.; Roy, A.; McGrath, J. E. J. Membr.
Sci. 2009. 333. 1-11.

Roy, A.; Lee, H. 8.; McGrath, J. E. Polymer 2008, 49, 5037-5044.

Nakabayashi, K.; Matsumoto, K.; Ueda, M. J. Polym. Sci. Pol. Chem.
2008, 46, 3947-3957.

Nakabayashi, K.; Matsumoto, K.; Higashihara, T.; Ueda, M. J. Polym. Sci.

Pol. Chem. 2008, 46, 7332-7341.

Matsumoto, K.; Ozaki, F.; Matsuoka, H. J. Polym. Sci. Pol. Chem. 2008,
46, 4479-4485.

Lee, H. 8.; Roy, A.; Lane, 0.; Dunn, 8.; McGrath, J. E. Polymer 2008, 49,

71 5-723.

161

(157)

(158)
(159)

(160)

(161)

(162)

(163)

(164)

(165)

(166)

(167)
(168)

(169)

(170)
(171)
(172)

(173)
(174)

Okazaki, Y.; Nagaoka, 8.; Kawakami, H. J. Polym. Sci. Pt. B-Polym.
Phys. 2007, 45, 1325-1332.

Zou, L.; Anthamatten, M. J. Polym. Sci. Pol. Chem. 2007, 45, 3747-3758.

Park, M. J.; Downing, K. H.; Jackson, A.; Gomez, E. D.; Minor, A. M.;
Cookson, D.; Weber, A. Z.; Balsara, N. P. Nano Lett. 2007, 7, 3547-3552.

Tsang, E. M. W.; Zhang, Z.; Shi, Z.; Soboleva, T.; Holdcroft, 8. J. Am.
Chem. Soc. 2007, 129, 15106.

Di Vona, M. L.; Sgreccia, E.; Licoccia, S.; Khadhraoui, M.; Denoyel, R.;
Knauth, P. Chem. Mat. 2008, 20, 4327-4334.

Cho, C. G.; Kim, 8. H.; Park, Y. C.; Kim, H.; Park, J. W. J. Membr. Sci.
2008. 308, 96-106.

Robeson, L. M.; ku, H. H.; McGrath, J. E. J. Membr. Sci. 2007, 302,
70-77.

Kim, D. H.; Choi, J.; Hong, Y. T.; Kim, S. C. J. Membr. Sci. 2007, 299,
19-27.

Compan, V.; Femandez-Carretero, F. J.; Riande, E.; Linares, A.; Acosta,
J. L. J. Electrochem. Soc. 2007, 154, B159-B164.

Wu, H. L.; Ma, C. C. M.; Liu, F. Y.; Chen, C. Y.; Lee, S. J.; Chiang, C. L.
Eur. Polym. J. 2006, 42, 1688-1695.

Swier, 8.; Shaw, M. T.; Weiss, R. A. J. Membr. Sci. 2006, 270, 22-31.

Gasa, J. V.; Weiss, R. A.; Shaw, M. T. J. Polym. Sci. Pt. B-Polym. Phys.
2006, 44, 2253-2266.

Swier, S.; Ramani, V.; Fenton, J. M.; Kunz, H. R.; Shaw, M. T.; Weiss, R.
A. J. Membr. Sci. 2005, 256, 122-133.

Kerres, J. A. Fuel Cells 2005, 5, 230-247.
Mokrini, A.; Huneault, M. A. J. Power Sources 2006, 154, 51 -58.

Piboonsatsanasakul, P.; Wootthikanokkhan, J.; Thanawan, 8. J. Appl.
Polym. Sci. 2008, 107, 1325-1336.

Mokrini, A.; Huneault, M. A.; Gerard, P. J. Membr. Sci. 2006, 283, 74-83.

Hong, L.; Chen, N. P. J. Polym. Sci. Pt. B-Polym. Phys. 2000, 38, 1530-
1538.

162

(175)

(176)

(177)

(178)

(179)
(180)
(181)
(182)
(183)
(184)

(185)
(186)
(187)

(188)
(189)

(190)
(191)

(192)

(193)

Fu, Y. Z.; Manthiram, A.; Guiver, M. D. Electrochem. Commun. 2007, 9,
905-910.

Zhong, S. L.; Cui, X. J.; Cai, H. L.; Fu, T. Z.; Shao, K.; Na, H. J. Power
Sources 2007, 168, 154-161.

Mikhailenko, S. U. D.; Wang, K. P.; Kaliaguine, 8.; Xing, P. X.; Robertson,
G. P.; Guiver, M. D. J. Membr. Sci. 2004, 233, 93-99.

Gao, J.; Lee, D.; Yang, Y. 8.; Holdcroft, S.; B. J.

Macromolecules 2005, 38, 5854-5856.

Fnsken,

Oren, Y.; Freger, V.; Linder, C. J. Membr. Sci. 2004, 239, 17-26.
Boob, S. B.; Shaw, M. T. ANTEC 2005, 1731-1735.

Brijmohan, S. 8.; Shaw, M. T. J. Membr. Sci. 2007, 303, 64-71.
Hou, J.; Baker, G. L. Chem. Mat. 1998, 10, 3311-3318.

Qiao, J.; Chen, Y. Y.; Baker, G. L. Chem. Mat. 1999, 11, 2542-2547.

Harrison, W. L.; Hickner, M. A.; Kim, Y. 8.; McGrath, J. E. Fuel Cells
2005, 5, 201-212.

Miyatake, K.; Watanabe, M. J. Mater. Chem. 2006, 16, 4465-4467.
Pyun, J.; Matyjaszewski, K. Chem. Mat. 2001, 13, 3436-3448.

Liu, T. Q.; Jia, 8. J.; Kowalewski, T.; Matyjaszewski, K.; Casado-Portilla,
R.; Belmont, J. Macromolecules 2006, 39, 548-556.

Li, D. J.; Sheng, X.; Zhao, B. J. Am. Chem. Soc. 2005, 127, 6248-6256.

Unsal, E.; Elmas, B.; Caglayan, B.; Tuncel, M.; Patir, 8.; Tuncel, A. Anal.
Chem. 2006, 78, 5868-5875.

Bridger, K.; Vincent, B. Eur. Polym. J. 1980, 16, 1017-1021.

Bridger, K.; Fairhurst, 0.; Vincent, B. J. Colloid Interface Sci. 1979, 68,
190-195.

Dorris, A.; Rucareanu, 8.; Reven, L.; Barrett, C. J.; Lennox, R. B.
Langmuir 2008, 24, 2532-2538.

Radhakrishnan, B.; Ranjan, R.; Brittain, W. J. Soft Matter 2006, 2, 386-
396.

163

(194)

(195)

(196)

(197)

(198)
(199)

(200)

(201)
(202)

(203)

(204)

(205)

Zhang, X. H.; Hong, L.; Liu, Z. L.; Lee, J. Y. Colloid Polym. Sci. 2008,
286, 1351-1360.

Wang, Y. P.; Pei, X. W.; He, X. Y.; Yuan, K. Eur. Polym. J. 2005, 41,
1326-1332.

Matyjaszewski, K.; Miller, P. J.; Shukla, N.; Immaraporn, B.; Gelman, A.;
Luokala, B. B.; Siclovan, T. M.; Kickelbick, G.; Vallant, T.; Hoffmann, H.;
Pakula, T. Macromolecules 1999, 32, 871 6-8724.

Pyun, J.; Jia, 8. J.; Kowalewski, T.; Patterson, G. D.; Matyjaszewski, K.
Macromolecules 2003, 36, 5094-5104.

Zhao, H. Y.; Kang, X. L.; Liu, L. Macromolecules 2005, 38, 10619-10622.

Chen, X. Y.; Randall, D. P.; Perruchot, 0.; Watts, J. F.; Patten, T. E.; von
Weme, T.; Armes, S. P. J. Colloid Interface Sci. 2003, 257, 56-64.

Vo, C. D.; Schmid, A.; Armes, S. P.; Sakai, K.; Biggs, S. Langmuir 2007,
23, 408-413.

Han, S. J.; Hyeon, T. Chem. Commun. 1999, 1955-1956.

Jiang, D. D.; Yao, Q.; McKinney, M. A.; Wilkie, C. A. Polym. Degrad.
Stabil. 1999, 63, 423-434.

Summer, M. J.; Harrison, W. L.; Weyers, R. M.; Kim, Y. 8.; McGrath, J.
E.; Riffle, J. 8.; Brink, A.; Brink, M. H. J. Membr. Sci. 2004, 239, 199-211.

Wiles, K. B.; Wang, F.; McGrath, l. E. J. Polym. Sci. Pol. Chem. 2005, 43,

2964-2976.

Kim, Y. 8.; Hickner, M. A.; Dong, L. M.; Pivovar, B. 8.; McGrath, J. E. J.
Membr. Sci. 2004, 243, 317-326.

(206) Sankir, M.; Kim, Y. 8.; Pivovar, B. 8.; McGrath, J. E. J. Membr. Sci. 2007,

(207)

(208)
(209)

(210)

299, 8-18.

Fontanella, J.. J.; McLin, M. G.; Wintersgill, M. C.; Calame, J. P.;
Greenbaum, S. G. Solid State Ion. 1993, 66, 1-4.

Liu, D.; Yates, M. Z. J. Membr. Sci. 2009, 326, 539-548.

Ge, J. P.; Hu, Y. X.; Biasini, M.; Dong, C. L.; Guo, J. H.; Beyerrnann, W.
P.; Yin, Y. D. Chem-Eur. J. 2007, 13, 7153-7161.

Deng, Y. H.; Wang, C. C.; Hu, J. H.; Yang, W. L.; Fu, 8. K. Colloid Surf.
A-Physicochem. Eng. Asp. 2005, 262, 87-93.

164

(211)

(212)
(213)
(214)

(215)

(216)

(217)

(218)

(219)

(220)

(221)

(222)
(223)
(224)

(225)

(226)

(227)

(228)

Wang, Y.; Teng, X. W.; Wang, J. 8.; Yang, H. Nano Lett. 2003, 3, 789-
793.

Vestal, C. R.; Zhang, Z. J. J. Am. Chem. Soc. 2002, 124, 14312-14313.
Xue, S. N.; Yin, G. P. Polymer 2006, 47, 5044-5049.

Genies, C.; Mercier, R.; Sillion, B.; Cornet, N.; Gebel, G.; Pineri, M.
Polymer 2001, 42, 359-373.

Sumner, M. J.; Harrison, W. L.; Weyers, R. M.; Kim, Y. 8.; McGrath, J. E.;
Riffle, J. 8.; Brink, A.; Brink, M. H. J. Membr. Sci. 2004, 239, 199-211.

Jang, W.; Sundar, 8.; Choi, 8.; Shut, Y. 6.; Han, H. J. Membr. Sci. 2006,
280, 321 -329.

Schuster, M.; Rager, T.; Noda, A.; Kreuer, K. D.; Maier, J. Fuel Cells
2005, 5, 355-365.

Lafitte, B.; Jannasch, P. In Advances in Fuel Cells; Kreuer, K.-D.,
Nguyen, T., Zhao, T., Eds.; Elsevier: 2007; Vol. 1, p 119-185.

Bock, T.; Mulhaupt, R.; Mohwald, H. Macromol. Rapid Commun. 2006,
27, 2065-2071.

Subianto, 8.; Choudhury, N. R.; Dutta, N. K. J. Polym. Sci. Pol. Chem.
2008. 46, 5431-5441.

Bock, T.; Mohwald, H.; Mulhaupt, R. Macromol. Chem. Phys. 2007, 208,
1324-1340.

Kreuer, K. D. Chem. Mat. 1996, 8, 610-641.
Souzy, R.; Ameduri, B. Prog. Polym. Sci. 2005, 30, 644-687.

Rager, T.; Schuster, M.; Steininger, H.; Kreuer, K. D. Adv. Mater. 2007,
19,3317. .

Steininger, H.; Schuster, M.; Kreuer, K. D.; Maier, J. Solid State Ion.
2006, 177, 2457-2462.

ltoh, T.; Hirai, K.; Tamura, M.; Uno, T.; Kubo, M.; Aihara, Y. J. Power
Sources 2008, 178, 627-633.

Jakoby, K.; Peinemann, K. V.; Nunes, 8. P. Macromol. Chem. Phys.
2003, 204, 61-67.

Allcock, H. R.; Hofmann, M. A.; Wood, R. M. Macromolecules 2001, 34,
6915-6921.

165