.--

+ ,4...” “a.“

.fi uufl-Eu'

Ev

nib-3:1-

huh-mm.

W

.15. - rl:":~':<'

 

 

is ;‘
.:. , ,~ f‘
m: 'i-fizggfi
£535,133:

.«1.
. :2;-:;:5« ,.
. x mm... ’
h- ». V)
33-33»; "‘
"ht--

 

m LIBRARY

:30 o “l Michigan State
5 4,7 6,; 5 o ‘7 University

This is to certify that the
dissertation entitled

Characterization and Manipulation of Vesicles Formed from
1,2-Butylene Oxide/ Ethylene Oxide Diblock Copolymers

presented by

John Keith Harris

has been accepted towards fulfillment
of the requirements for the

PhD. degree in Degrtment of Chemistry

 

 

MMQJM

Major Professor’s SWture

December 5, 2003

 

Date

MSU is an Affirmative Action/Equal Opportunity Institution

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

 

 

 

 

 

 

 

 

 

 

 

 

 

 

6/01 c/CIRC/Datemest-pts

 

CHARACTERIZATION AND MANIPULATION OF VESICLES
FORMED FROM 1,2-BUTYLENE OXIDE/ ETHYLENE OXIDE
DIBLOCK COPOLYMERS

By

John Keith Harris

A Dissertation

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

DOCTOR OF PHILOSOPHY
Department of Chemistry

2003

ABSTRACT

CHARACTERIZATION AND MANIPULATION OF VESICLES FORMED
FROM 1,2-BUTYLENE OXIDE! ETHYLENE OXIDE DIBLOCK COPOLYMERS

By
John Keith Harris

Vesicular structures are promising materials for encapsulation applications, and
vesicles composed from block copolymers are especially attractive in this regard because
of their cost, purity, tunability and stability. However, vesicle formation generally requires
specialized procedures. This dissertation demonstrates that 1,2-butylene oxide/ ethylene
oxide di—block copolymers spontaneously form um—sized, multi-lamellar vesicles
(“onions”) over a broad range of copolymer concentrations upon simple mixing with
water. Slow copolymer hydration without agitation produces giant vesicles with
diameters greater than 0.1 mm.

This behavior is consistent with a geometric argument that asserts vesicle formation
is preferred only when the proper ratio of hydrophobe volume to the product of head
group area and hydrophobe length is achieved. The relatively large volume to length ratio
of the butylene oxide hydrophobe is critical to vesicle formation, but formation also
depends on the ethylene oxide head group length. For BO block lengths 15 units or less,

BO/EO copolymers with E0 blocks between 5 and 15 units long spontaneously form large

MLVs upon simple mixing with water. EO lengths greater than 15 result in small vesicles

or no vesicles at all. We also found that EO block length polydispersity strongly affects
structure formation. Substitution of an 803' group for the E0 block produced small

vesicles, further evidence that the BO geometry is key. For BO/EO copolymers with BO
block lengths of 15 to 25 units, ultra-sonication is required for vesicle formation, and the
resulting structures have diameters of about 70 nm.

BO/EO MLVs formed by hand-shaking have shelf lives of over 2 years and survive
thermal cycling, mild sonication and moderate shear. In contrast, extrusion of the same
dispersion through a porous membrane reduces vesicle sizes, although final structure
diameter is independent of shear rate and the pore size used for extrusion. Vesicle
formation is reversed by thermal cycling. Ultra-sonication also reduces structure diameter.

Copolymer composition was analyzed using nuclear magnetic resonance (NMR),
gel permeation chromatography (GPC) and mass spectrometry (MS). Vesicle capacity for
encapsulating both hydrophobic and hydrophilic materials was explored. Phase diagrams
were prepared of the copolymer, a model hydrophobe and water. The release rate of an
encapsulated hydrophile was also studied. Structures were characterized using plane-
polarized light microscopy, dynamic light scattering and cryogenic scanning electron

microscopy.

To Sharon, my “pearl of great price ".

All the wonderful events of my life have happened with you at my side.

iv

ACKNOWLEDGEMENTS

The Apostle Paul wrote “No man lives unto himself and no man dies unto himself”
(Rom. 7:14). Likewise a project of this scope is the result of the efforts and contributions
of numerous individuals, first and foremost of whom was my wife, Sharon. I could not
have begun this project without her support and would not have been able to continue
without her encouragement. Through numerous challenges, she was my stabilizing force,
my strength and my comfort. Without her, I would certainly have given up on this long
ago.

This research was the product of a coordinated interaction between Michigan
State University and The Dow Chemical Company that required a great deal of work on
the parts of Gene Rose and Merlin Bruening. Gene acted as the liaison between Dow and
MSU, but more importantly was my technical guide, my mentor and quite frequently my
sounding board. As my advisor, Professor Merlin Bruening undertook the challenging
task of working with an older student, who had spent a decade in industrial research and
was commuting hours each day. He was honest, meticulous and demanding and I am
unquestionably a better scientist because of his efforts. This achievement would not have
been possible without him and l doubt that few of his colleagues would have even
considered such a role. I am extremely grateful that both of these men thought it
worthwhile to invest so much in helping me reach this goal.

While these gentlemen played key roles in this project, my colleagues also

contributed significantly to my success. Individuals like Tom Kalantar, Mladen Ladika,
Steve Mork, Greg Rickle, Gary Strandburg, Sonke Svenson, and Chris Tucker provided

invaluable technical expertise at a moments notice. Mark Nace, Bob Whitmarsh and

Mark Warny supplied synthesis “know-how” and materials on a regular basis and I
appreciate their efforts. Bruce Bell, Deborah Rothe and David Williams provided critical
analytical data on the copolymers and structures. 1 had numerous productive
conversations with Professors Paschalis Alexandridis (SUNY at Buffalo), Frank Bates
(Univ. of Minnesota), Eric Kaler (Univ. of Delaware), and Robert Prud’homme
(Princeton) all of whom provided me with valuable insights and continuous
encouragement.

As with most research, quality lab assistance was critical. My heart-felt thanks go
out to Stacie Young whose patience and commitment were above and beyond
expectation. She was there day-in and day-out and I grew to depend on her more than I
realized. Thank-you for all of your hard work, Stacie. Likewise Doug Ebaugh, Jack
Hoffman, Dubravka Ladika, and Andrew Schirmer all contributed hours doing repetitive
tasks so that I would be free to do my research. Thanks to you all.

I also must say thanks to David Forgach and Paul Glugla. When I entered the
research arena these men took me under their wings, encouraged me and helped me
develop fundamental research skills. They were my mentors and my friends and much of
what I demonstrated as a researcher that earned me this opportunity, I owe to their
patience and kindness. Thank-you both.

Finally, I am deeply indebted to the leadership at Dow that allowed me to
undertake this project. Individuals like Pat Andreozzi, Dave Beechuk, Bob Cipriano,
Tim Fawcett, Pat Green, Erin O’Driscoll, and Dane Wright saw this as an opportunity
rather than a burden and were willing to endure some inconvenience so that I could

pursue this goal. My thanks to each of you.

vi

TABLE OF CONTENTS

List of Schemes ........................................................................................................ xi
List of Tables ............................................................................................................ xii
List of Figures .......................................................................................................... xiii

Chapter 1. Introduction to Vesicle Formation

1.1 Molecular Self-assembly and Overview of the Dissertation ........................... l
1.2 Solvent-solute Interactions and Pair Potentials .............................................. 4
1.3 Amphiphiles and the Critical Micelle Concentration ...................................... 6
1.4 The Thermodynamics of Micellization .......................................................... 9
1.5 Aggregation Numbers ................................................................................... 12
1.6 The Critical Packing Factor Model and Aggregate Shape .............................. 16
1.7 Vesicles ......................................................................................................... 24
1.8 Block Copolymers ......................................................................................... 28
1.9 1,2-Butylene Oxide/ Ethylene Oxide Diblock Copolymers ............................ 29
1.10 References .................................................................................................... 31

Chapter 2. Synthesis and Characterization of 1,2-Butylene Oxide!

2.1

2.2

2.3

Ethylene Oxide Diblock Copolymers

Overview ...................................................................................................... 38
Parr Reactor Synthesis ................................................................................... 4O
PolyGlycols Mini-Plant Synthesis ................................................................. 44

vii

2.4 Copolymer Purification ................................................................................. 45
2.5 Copolymer Characterization .......................................................................... 49
2.6 Analysis by Nuclear Magnetic Resonance Spectroscopy ............................... 49
2.7 Gel Permeation Chromatography ................................................................... 53
2.8 Copolymer Analysis by Mass Spectrometry .................................................. 56
2.9 Mass Spectrometry Results ........................................................................... 59
2.10 Conclusions ................................................................................................... 67
2.11 References .................................................................................................... 67

Chapter 3. Effect of Butylene Oxide and Ethylene Oxide

Block Length on Structure Formation

3. 1 Introduction ................................................................................................... 69
3.2 Screening 'Tests for Liquid Crystalline Phases ............................................... 72
3.3 Spherical Lamellar Structures ........................................................................ 80
3.4 Vesicle Structures ......................................................................................... 85
3.5 Cryo-SEM ..................................................................................................... 87
3.6 Effect of EO Length on Vesicle Formation .................................................... 89
3.7 Effect of Concentration on Vesicle Formation ............................................... 93
3.8 Effects of Polydispersity on Vesicle Formation ............................................. 96
3.9 Giant Vesicles ............................................................................................... 98
3.10 Conclusions ................................................................................................... 106
3.11 References .................................................................................................... 107

viii

Chapter 4. Vesicle Characterization and Stability

4.1 Introduction to Vesicle Stability .................................................................... 110
4.2 Yield of Structures ........................................................................................ 1 l 1
4.3 Thermal Stability of Liquid Crystalline Structures ......................................... 1 12
4.4 Thermal Stability of Vesicle Dispersions ....................................................... l 13
4.5 Structural Stability ........................................................................................ 1 14
4.6 Extrusion of Vesicles .................................................................................... 1 16
4.7 Ultra-sonication of Vesicles .......................................................................... 1 18
4.8 Examination of the Effect of Added Hydrophobe on Structure Formation ..... 120
4.9 BO/EO Copolymer Phase Diagrams .............................................................. 122
4.10 Hydrophile Loading - Experimental Procedure .............................................. 128
4.11 Hydrophilic Loading and Release .................................................................. 132
4.12 Conclusions ................................................................................................... 137
4.13 References .................................................................................................... 138

Chapter 5. Spontaneous Vesicle Formation from

5.1

5.2

5.3

5.4

5.5

Poly(l,2-Butylene Oxide) Sulfate Oligomers

Introduction ................................................................................................... 141
Oligomer Synthesis and Characterization ...................................................... 142
Vesicle Formation and Stability ..................................................................... 146
Conclusions ................................................................................................... 150
References .................................................................................................... 150

ix

Chapter 6. Summary and Future Work

6. 1 Background .............................................................................................. 152
6.2 Summary .................................................................................................. 1 53
6.3 Future Research ....................................................................................... 155
6.4 References ............................................................................................... 157

Appendix A Geometric Limitations Imposed by Surfactant

Packing Shape
A1 The Critical Packing Factor Model ........................................................... 159
A2 Aggregation Number versus Head Group Size .......................................... 159
Appendix B Processing the Mass Spectrometry Data ....................... 161

List of Schemes

Scheme Page
Scheme 2.1 Synthesis of 1,2-butylene oxide/ ethylene oxide copolymers;

(1) BO oligomer formation, (H) BO oligomer with active EO site,

(HI) acid quenching of reaction. ............................................................ 38
Scheme 5.1 Poly(l,2-butylene oxide) oligomer structure before ( 1a)

and after (lb) sulfonation. ..................................................................... 143

xi

Table

Table 2.1
Table 2.2
Table 2.3
Table 2.4

Table 3.1

Table 4.1

Table 8.1

Table B.2

Table B.3

Table B.4

List of Tables

Page
BO/EO copolymers screened for evidence of micellar formation. ......... 43
Copolymers prepared by the Mini-plant Process. .................................. 48
Copolymer compositions derived from mass spectrometry data. ............ 65
Comparison of compositions based on analytical method. .................... 66

Composition, molecular weight, density and concentration range for
PLM-detectable vesicle formation for BO/EO diblock copolymers. ..... 90
BO oligomer concentrations in BO/EO copolymers. ............................. 126
A portion of the mass spectrometry data for copolymer

L1150—51A(B011E012), arranged by increasing mass, after

being processed to remove multiple-charge signals. .............................. 165
Calculated masses of BO/EO diblock copolymers for compositions

BO3EOO through B014E024. .............................................................. 167

A portion of the copolymer L1150-51A mass spectrometry data,

arranged by increasing intensity, and including a

“range” designation. ............................................................................. 168
Example of copolymer compositional distribution spread sheet

showing total intensity of each composition. ......................................... 169

xii

Figure

Figure 1.1

Figure 1.2

Figure 1.3

Figure 1.4

Figure 1.5

Figure 1.6

Figure 1.7

Figure 1.8

Figure 2.1

Figure 2.2

Figure 2.3

Figure 2.4

Figure 2.5

List of Figures

Page
The balance of forces acting on a surfactant molecules to produce
micelles. ............................................................................................... 8
Various states of surfactant aggregation within the bulk solution. ......... 13
Optimum head group size as a balance of attractive and repulsive
forces. .................................................................................................. 15
The four fundamental micellar self-assembled units. ............................. 17
Two types of aggregate bilayer structures; flexible and rigid. ................ 18
Eight basic packing configurations for various surfactant molecules
in a rnicelle. .......................................................................................... 20
Two types of closed bilayer structures; uni-lamellar and
multi-lamellar. ...................................................................................... 24
Two regions of vesicles structure are potential encapsulating
areas. .................................................................................................... 26
A representative BO/EO molecule showing amphiphile shape. ............. 39

NMR spectrum of BO/EO diblock copolymer M1038A, (BOl 1E012). . 51

GPC chromatographs of the two copolymers whose synthesis

was described in section 2.2. ................................................................. 55
Compositional profile of copolymer L1150-51A (angled view)

determined from ESI Mass spectrometry. ............................................ 61
Compositional profile of copolymer L1 150—5 1A (overhead view)

as determined from ESI Mass spectrometry. . ........................................ 62

xiii

Figure 2.6

Figure 2.7

Figure 3.1

Figure 3.2
Figure 3.3
Figure 3.4
Figure 3.5

Figure 3.6

Figure 3.7

Figure 3.8

Figure 3.9

Figure 3.10

Figure 3.11

Figure 3.12

Compositional profile of copolymer L1150-51B (angled view)

as determined by ESI-mass spectrometry. ............................................ 63
Compositional profile of copolymer L1150-51B (overhead view)

as determined by mass spectrometry. .................................................... 64
Micellar aggregate structures and the resulting meso-phases. ................ 71

Tetragonal meso-phase produced by B01 1E012 in water (70 wt %). 75

Hexagonal meso-phases produced by B01 lEOlg. ................................. 77
Planar lamellar formed by BO] 113012 in water (30 wt %). .................... 78
Example of screening test results of B0] IEOIZ copolymer. .................. 79
Maltese Cross formation occurs in plane-polarized light because of

the uni-axial liquid crystalline structure of spherical lamellae. .............. 81

Vesicles of BO] 1E0” (20 wt %) viewed through white light (above)

and plane-polarized light (below). ........................................................ 82

Butylene oxide oligomer, BO 12, (2 wt %) dispersion viewed under

partially plane-polarized light microscopy. ............................................ 83

Vesicle of B011E012 “budding” from myelin tip formed at

copolymer/ water interface. . ................................................................. 84
Plane-polarized light micrograph of B01 1E0” (1 wt %)

without slide cover. .............................................................................. 85
Plane-polarized light micrograph of vesicles from BO] IEOIZ

dispersion (1 wt %). ............................................................................. 86

Cryo-SEM micrograph of B0] IEOIZ vesicle (1.0 wt %) taken

10 weeks after formation. ..................................................................... 87

xiv

Figure 3.13

Figure 3.14

Figure 3.15

Figure 3.16

Figure 3.17

Figure 3.18

Figure 3.19

Figure 3.20

Figure 3.21

Figure 4.1

Figure 4.2

Cryo-SEM of vesicle from B01 15012 ( 1 wt %) showing numerous

layers of MLV structure. ...................................................................... 88
Plane-polarized light micrographs of an B011E019 aqueous

dispersion (1 wt %) before and after heating. No vesicles are visible

prior to heating (left), but numerous structures are visible after

heating at 85° C for 3 hours (right). ....................................................... 92
Mean structure diameters as a function of EO block length for

B01 1E0x 00polymers. .......................................................................... 93
301 1E0” dispersion (15 wt %) showing vesicles with planar

lamellae. ................................................................................................ 95

Plane-polarized light micrograph of B01 1E08 (BL50—1500A) shows

helical myelin at water/ surfactant interface upon hydration. ................. 98
Experimental set-up for photographing giant vesicle formation. ............ 101
Giant vesicles from copolymer B012E095 in glass cell. ....................... 102
Giant vesicles dispersed after slight agitation. ....................................... 103

Plane-polarized light micrograph of giant vesicle isolated from
copolymer of second experiment. ......................................................... 104
Particle size distribution of MLVs in a 5.0 wt % aqueous dispersion

of B01 1E012 before and after various post-formation treatments. ........ 1 15

Particle size distribution before and after extruding a 5.0 wt %

dispersion of B01 1E0” through a 5.0 pm or 0.45 pm filter. ................ 1 18

XV

Figure 4.3 Phase behavior of BO6EO4 with dodecane and water at

room temperature. ................................................................................ 123
Figure 4.4 Phase behavior of B0, 15012 with dodecane and water at

room temperature. ................................................................................ 124
Figure 4.5 Phase behavior of 801 1E019 with dodecane and water at

room temperature. ................................................................................ 126

Figure 4.6 Phase behavior of BOIZEOIO (mass spectral composition) with

dodecane and water at room temperature. ............................................. 127
Figure 4.7 The fraction of fluorescein dye as a function of time retained by

30205012 uni-lamellar vesicles. .......................................................... 135

Figure 4.8 Sum of retained and released fluorescein dye for each batch
of vesicles. ............................................................................................ 137

Figure 5.1 FT-IR spectra of the B017 oligomer before and after sulfonation. ......... 144

Figure 5.2 Cryo-SEM image of spontaneously-formed vesicles in a 1 wt%

EtB011S04Na dispersion. .................................................................... 147
Figure 5.3 Diameters of vesicles formed from ultra-sonicated EtBOl IEOIZ
(O) and EtBOl 1SO4Na (I) (1 wt %) as a function of temperature. ......... 148

Figure B.1 Mass spectra of three BO Oligomers along with calculated average

molecular weights and polydispersities. ................................................ 166

xvi

Figure B.2 Compositional profile of copolymer L120-15A (angled view).

Figure B.3

The mean composition of the above peaks is BOIOEO6, while NMR
indicated a composition of BO6EO4. The expected composition
based on synthesis formulation was EOlOBOIO. ................................. 170

The compositional distribution of the BO oligomer from
Batch M 1038. Note that the majority of the material (peaks

with highest intensity) lies on integer values of B0 length. ................... 171

xvii

Chapter 1

Introduction to Vesicle Formation

1.1 Molecular Self-assembly and Overview of the Dissertation

Micrometer- and nanometer-sized structures are employed in applications ranging
from material modification to encapsulation. Nevertheless, synthesis of micro- or nano-
structures such as exfoliated clays, inorganic nanoparticles, carbon nanotubes, hollow
latexes, and hollow polyelectrolyte spheres is generally labor or energy intensive."4
Development of systems that self-assemble at the molecular level to yield such structures
should significantly reduce the effort required to produce large quantities of micro- and
nano-materials.5 Such self-assembly systems could be driven either kinetically or
thermodynamically.

A thermodynamic mechanism for molecular self-assembly already exists for
selected molecules and offers several advantages over conventional processes for micro-
and nano-structure generation. The process is an equilibrium condition, however, and
requires careful control over molecular architecture and environmental conditions to
function correctly. Because the self-assembled structure is the preferred molecular state,
these compounds generally have high yields and reduced levels of waste material. Self-
assembly also provides fine control over material properties because structure formation
is governed at the molecular level, where small changes can dramatically alter structure
energies and geometrical"7 This allows tailoring of structure properties (i.e., size, shape
or durability) using only slight changes in environment or molecular composition.
Additionally, because the desired structure is the preferred molecular arrangement,

minimal defects occur leading to improved material properties.8

While development of a self-assembling system requires careful control of
molecular architecture and environmental conditions, numerous examples of such
systems exist. In nature, phospholipids self-assemble into vesicles, the protective
membranes that house biological cells and protect them from a hostile environment.
Phospholipids form an encapsulating molecular bilayer that is robust, flexible and self-
repairing. The bilayer can also be perm-selective, allowing for controlled transmembrane
transport, and modification of vesicle surfaces permits partitioning of structures into
selected environments such as specific biological domains or organs. These properties
make vesicles ideal candidates for a host of encapsulation applications including
improved drug solubility, increased “release—rate” control, enhanced biological imaging
techniques, and artificial cells for oxygen transport or artificial photosynthesis.9'13
Vesicular structures have also been used as additives for modification of polymer
material properties. '4

Unfortunately, phospholipids have inherent operational limitations as well as
processing constraints. Synthesis and purification of these materials is costly, and vesicle
formation often requires additives and development of highly specific formulations.
Structure generation can also be laborious and time consuming. Finally, phospholipid
structures frequently have limited stability during exposure to thermal variations, salinity
changes, shear or organic solvents.

In an effort to overcome the limitations of phospholipids, this work explores the
use of selected diblock copolymers as vesicle-forming materials. Synthesis and

purification of copolymers is often straightforward and can yield larger quantities of

materials more quickly and at relatively lower cost than phospholipids. Rapid synthesis

also means that the effects of changes in monomer or block lengths on structure stability
can be evaluated faster and more thoroughly and, once understood, these compositional
changes should allow for tailoring of vesicle properties for specific applications.
Specifically, this dissertation examines the formation of vesicles by a series of 1,2-
butylene oxide/ ethylene oxide diblock copolymers. Copolymer synthesis is rapid and
versatile, and allows for accurate control of block lengths. Moreover, 1,2-butylene oxide/
ethylene oxide (BO/E0) copolymers are innocuous, bio-compatible and inexpensive.
This dissertation is divided into six chapters. The first provides background
information on the nature of surfactants, the driving force behind self-assembly, and the
model that I use to explain the impact of the BO/EO copolymer geometry on structure
formation. Chapter 2 then focuses on copolymer synthesis and characterization, while
Chapter 3 examines the effect of block lengths and formation processes on structure
formation. In Chapter 4, I discuss structure stability with respect to thermal and
mechanical stresses, methods of altering the mean particle size and vesicle release rates.
To better understand vesicle formation, I also prepared sulfonated butylene oxide
Oligomers, and Chapter 5 presents the vesicle-forming properties of these materials.
Chapter 6 examines the conclusions of this work and proposes future research directions.
I chose 1,2-butylene oxide/ ethylene oxide diblock copolymers based on a
fundamental understanding of the driving forces behind self-assembly and structure
formation. To help explain this project and its goals, I present below a brief description
of these forces and their effects on self-assembly. Section 1.2 discusses the fundamental
interactions involved in solvation. In Sections 1.3 through 1.5, I address the driving

forces behind surfactancy, self-assembly and structure formation. I describe the impact

of molecular geometry on structure size and shape in Section 1.6, and Section 1.7
discusses the nature of vesicles; their properties, uses and limitations. I then show how
others have used di- and tri-block copolymers to form self-assembled structures in
Section 1.8. Finally, in light of these topics, Section 1.9 presents my reasoning for
selecting 1,2—butylene oxide/ ethylene oxide copolymers for vesicle formation and the

experimental course I pursued.

1.2 Solvent-solute Interactions and Pair Potentials

A heuristic principle of solvent-solute interactions is that “likes dissolve likes”.
From a thermodynamic perspective, this translates into a comparison of solute-solute and
solvent-solvent interactions with solvent-solute interactions. The net force between any
two molecules is called the pair potential and reflects the sum of the attractive and
repulsive forces that exist between them. The similarity of the solute-solute and solvent—
solvent pair potentials is the overriding factor determining solubility.

Israelachvili discusses pair potentials and the two dominant types of
intermolecular forces that produce them.” The first forces are the electrostatic
interactions that arise from fixed charges or permanent dipoles in the molecules. These
forces act over short ranges and adhere to typical coulombic laws governing the
interactions of charged species. The second forces result from induced dipoles and also
extend over only a few A.

If the solvent pair potential is comparable to the solute pair potential, solvation
can occur spontaneously because the solute-solvent pair potential will likely be

comparable to both the solute and solvent pair potentials, and entropy will drive

dissolution. Solute dissolution can affect the system free energy however, by changing
the solute-solute or solvent-solvent pair potentials. For example, the solute molecule may
be of sufficient size that it allows the solvent molecules to pack in a more efficient
manner than occurs in pure solvent. The solution is then at a lower energy state than
when either component was pure and solvation is generally observed as an exothermic
reaction such as dissolving MgSO4 in water. Systems such as this are extremely stable
because there is no incentive for solute aggregation due to the high free energy costs.

Conversely, if the solute pair potential is lower than the solvent pair potential,
solvation can occur only while additional energy, e.g., stirring, is provided. For instance,
if the solvent molecules interact in a specific manner, such as a dipole-dipole orientation,
and the solute molecule disrupts the “order” of the solvent system, then distributing the
solute throughout the solvent would require an increase in the internal energy of the
solution, an undesirable outcome. Therefore the lowest energy state occurs when the
solute molecules are interacting with the fewest solvent molecules. Solutions produced
from these solute-solvent systems would be precarious at best, because any loss of energy
could initiate solute coalescence and lead to phase separation. Therefore, these systems
generally require constant agitation or the solution transitions to a dispersion.

Finally, the solvent may induce some change in the solute dipole (or vice versa)
that alters the net force between similar molecules. For example, ionization of the solute
would convert the pair potential, which may have been mildly attractive in the uncharged
state, to repulsive. Likewise molecules that had not been particularly polar when pure,
may become polarized in the presence of selected solvents. Polyethers, for instance,

become much more polar when dissolved in water than in their neat form. In this

manner, the solvent can create a physical change in the solute that alters the solute-solute
pair potential.

To summarize, molecules vary chiefly in their polarity, or polarizability, which
affects their tendency to interact. Aliphatic groups, like alkanes, have low polarizability
and low interaction potentials, while water, which is highly polar, has a strongly
attractive pair potential. In general, the more polar or polarizable a solute, the more polar
the corresponding solvent must be for solvation to occur. As differences in polarity
between solvent and solute increase, the solute solubility decreases.

It should be noted that the interaction potentials do not require homogeneity of the
solute molecules. A sufficiently large solute could contain a polar, non-polar and
polarizable group within the same molecule. The solvent molecules would interact with
each group differently, generating a range of intermolecular forces that vary with the
group polarity. In fact, molecules with different “groups” possess interesting and

desirable properties.

1.3 Amphiphiles and the Critical Micelle Concentration

If the solute molecule is not homogenous with respect to polarity, that is, there are
portions of the molecule that differ significantly in polarity from other portions of the
same molecule, then solvent selection becomes more challenging. Molecules such as this
are referred to as amphiphilic indicating that they are composed of at least two dissimilar
fractions that prefer different solvents. These fractions are respectively labeled lyophilic
(“solvent-loving”) and lyophobic (“solvent-fearing”) depending on the solvent used.

With respect to water, the hydrophile is the polar group, typically ionic, but any

functional group that provides hydrogen-bonding sites will suffice. The hydrophobe is
frequently an alkane, but any poorly polarizable group can serve this function. The
hydrophile and hydrophobe are commonly referred to as the “head” and “tail”,
respectively, of the amphiphile.

Most amphiphilic molecules are not readily soluble in water.16 Instead, upon
mixing with water the amphiphile migrates to a surface, such as the air-water interface,
where the hydrophobe can be displaced, at least partially, from the aqueous medium.
Because of this tendency to move to a surface, amphiphiles are also called “surfactants”.
Once the interface is saturated with surfactant, however, any additional molecules must
remain in the aqueous phase.

Amphiphile solubility is a function of its size and polarity. If the hydrophobe is
sufficiently large, or if the concentration of amphiphiles is sufficiently high, it becomes
thermodynamically unfavorable for the molecules to be dissolved in the aqueous phase.
In this case, to minimize contact between the two incompatible components, i.e. the
hydrophobe and water, the amphiphiles begin to assemble into structures. Once structure
formation begins, the level of “free” surfactant, or monomer, in the system remains
constant and all additional surfactant is incorporated into structures, or micelles. The
concentration at which structures begin to form is called the critical micelle
concentration, or CMC.

The CMC is determined by tracking specific physical phenomena of the system
such as turbidity or surface tension as a function of surfactant concentration. However,
regardless of the technique employed, CMC determinations have an inherent level of

uncertainty and are somewhat method-dependent.l7 Examination of more than one

system property may be required to obtain precise CMC values. Nonetheless an
approximate CMC value is usually sufficient for self-assembly studies.

In aqueous systems, a balance of three physical properties determine the CMC
(Figure 1.1). The hydrophobe insolubility is balanced against the hydrophile solubility,
which is dependent on factors such as the hydrophile-water pair potential, the number and
accessibility of hydrogen-bonding sites, and hydrophile-hydrophile interactions.‘8
Additionally, repulsive interactions between head groups will increase surfactant
solubility, while attractive interactions reduce solubilization and facilitate longer range
interactions such as gel formation or other rheological changes. In general, surfactants
with ionic head groups have higher CMCs than those with nonionic head groups for a
given hydrophobe.18 Increasing the magnitude of the amphiphile’s hydrophilic portion
will increase CMCs slightly while increasing the hydrophobic portion will significantly
decrease the CMC. Micellization, however, is not due to an attraction between the

Head Group Interaction
<==> <2: <==> <==> <==>

 

<: Hydrophilic .. . Hydrophilic :>
O O
O O
O O
O O
O . O

 

 

:> Hydrophobic c

Figure 1.1. The balance of forces acting on surfactant molecules to produce micelles.

(Shown as hemi-micelle).

hydrophobic groups. Tanford has shown that hydrophobes have about the same pair
potential with water as they do with each other,19 which is much lower than the water pair

potential. The next section explains the driving force behind micellization.

1.4 The Thermodynamics of Micellization

Micellization is a thermodynamically driven process'8 and both phase-separation
and mass-action models can describe this reaction.20 Rosen used work by Molyneux and
Birdi to present an explanation of micellization based on the phase-separation model, and
I follow that treatment here.“22 Equations 1.1 to 1.3 show the micellization reaction and

the chemical potentials of the surfactant in the aqueous, um)“, and micellar, “mic, phases.

Hydrophobemon A Hydrophobemic eq. 1.1
“mon = “omon + RT In amon e‘l- 1'2
“mic = u°mic + RT In a mic eq. 1.3

The chemical potential gives the total molar free energy for the molecules including the
pair potential and contributions from its thermal energy. The difference in molar free

energy between the states (um-c - umon) is the free energy change on micellization,
AG°mic, and is given by equation 1.4.
AG°mic = (u°mic + RT 111 amic) - (u°mon + RT 1“ amon) eq- 1-4
At equilibrium, AG°mic = 0, and substitution of K = am! am)“, produces equation 1.5,
AG°mic = -RT In K eq. 1.5
where AG°mic = u°mic - Mme“. Birdi and Molyneux also assumed that the amphiphiles

behave as a pure phase in the micelle (activity of 1) and as if at infinite dilution (activity

coefficient of 1) in the aqueous phase. Thus, the standard-state free energy change on
micellization as a function of the surfactant concentration is given by equation 1.6, where

XCMC is the mole fraction of surfactant molecules dissolved in water at the CMC.

(Below the CMC, micellization cannot be an equilibrium process as AG will be greater

than zero, and once the CMC is reached, XCMC is constant because all added surfactant
forms micelles.)

AG°mic = -RT ln XCMC eq. 1.6
Of course, AG°miC can be expressed as a function of enthalpy, AH°mic, and entropy,
AS°mic, changes (eq 1.7).

AG°mic = AH°miC -TAS°mic eq. 1.7
If the micelles do not interact to facilitate or interfere with additional micelle formation,

then AH°mic and AS°mic should be relatively constant over the temperature range
examined. Thus, the derivative of AG°mic with respect to temperature is given by
equation 1.8.

d(AG°mic)/dT = -AS°mic eq. 1.8

By monitoring the change in CMC with respect to temperature, the free energy

can be used to determine the system entropy change on micellization. Likewise, if

AS°mic is held constant over the same temperature range, the change in enthalpy can be

described by equation 1.9. Using these equations, the changes in free energy, enthalpy

and entropy of micellization can be determined as a function of the surfactant

T2d(AG°miC fI‘)/dT = -AH°mic eq. 1.9

10

concentration in the system The data from numerous studies on many types of
surfactants produce a consistent pattern: negative free energy changes on micellization

typically result from large positive changes in AS°miC. Frequently the change in
enthalpy, AH°mic, is positive, but even when negative, the value of TAS°mic dominates

the free energy change. Thus, the process of micellization is driven principally by
entropy gains. How then, does molecular self-assembly lead to increased entropy in the
system?

There are two explanations for this entropy increase. Aranow and Witten23
suggest that water molecules surrounding the hydrophobe restrict its rotation about its C-
C bonds. This reduces the hydrophobe’s configurational entropy, forcing it to maintain
an elongated and rigid shape. In a spherical micelle however, the hydrophobe is
partitioned into the center of a sphere filled with other hydrophobic moieties. This
should yield a configurational entropy similar to that possessed by the hydrophobe in its
pure, liquid state.

The second explanation focuses on the organization of water molecules
surrounding the hydrophobe.24 Shinoda performed an extensive study of enthalpy
changes that occur during micellization and concluded that at low temperatures, the
presence of the hydrophobe reduces the number of hydrogen-bonding sites accessible to
the water molecules surrounding it.25 Fewer sites mean that the hydrogen-bonds that do
form tend to be less flexible. This produces a rigid network of water molecules, a
“pseudo-ice” state, which decreases the system entropy. Minimizing the number of water
molecules in direct contact with the hydrophobe minimizes the formation of this rigid

state and increases the system entropy. Thus, hydrophobe aggregation to reduce contact

11

with water is entropically highly favorable. Geometrically the structure with the smallest
surface area to volume ratio is the sphere, therefore surfactants assemble in spherical
structures whenever possible to minimize contact with water.

Further work by Evans26 and Beesley et al.27 showed that micellization requires
that solvent molecules contain multiple hydrogen-bonding sites. Strong hydrogen bonds
provide the driving force for cooperative interactions among solvent molecules at the
hydrophobe surface. Hence, aqueous surfactant assembly is not due to attraction between
hydrophobes, but rather occurs as an expression of hydrophobe expulsion by water. The
hydrophobes are forced together by the strong attraction of water molecules to one

another and the need to maintain the mobility of these molecules.

1.5 Aggregation Numbers

Section 1.4 described the entropic driving force behind micellization but said
nothing of the size, shape or lifetimes of the micelles themselves. To gain a better
understanding of micellar structures and any intermediate aggregates that the monomers
may form, I now examine the thermodynamics of the surfactant system as a function of
aggregation number.15 '19 Clint presents an in-depth discussion of the multiple
equilibrium model using statistical thermodynamics.19 To begin, I assume that the
surfactant is in one of two states; a solvated monomer unit, M, or in an aggregate
structure composed of N molecules, MN, as shown in equation 1.10. In this model,

NM A MN eq. 1.10

4———-)

I do not consider the aggregate as a separate phase. Equating the chemical potentials for

the aggregation reaction gives equation 1.11 where I substitute “k” for “R” to indicate

12

that the free energies are per molecule. Note that XN is the mole fraction of surfactant

N u°M + NkT 1n XM = u°N + kT 1n (XN/N) eq. 1.11
molecules in structures with aggregation number N. Thus XN/N is the micellar
concentration of structures having aggregation number, N. Dividing equation 1.11 by N
yields equation 1.12, where 11°}; is the standard state chemical potential of a surfactant
molecule in the aggregate, rather than the aggregate itself.

11°... + kT ln xM = 11°}. + kT/N ln (XN/N) eq. 1.12
To completely define the system, I need only remember that the total solute concentration
is given by equation 1.13 (See Figure 1.2). Solving this equation for XN gives

00

C=X1+X2+X3+...=£XN eq.1.13
N=l

 

Trirner X1
O’\\.—O
O& 1101 S

 

 

 

Figure 1.2. Various states of surfactant aggregation within the bulk solution. Adapted

from Israelachvili, Intermolecular and Surface F orces.28

13

equation 1.14 which allows us to directly compare the surfactant population in each of
the aggregate states as a function of surfactant concentration in the monomeric state.
XN = le1 exP[(11°1 ' “O’Nyle lN eq. 1-14
This equation shows that, as expected, when no] - uO'N = 0, then X1 = XN
because if there is no chemical potential difference, then there is no incentive for
molecular aggregation based on a decrease in entropy. Note also that increasing values
of N lead to exponentially decreasing values of XN. That is, the larger the aggregation

number, the fewer such structures form. For sufficiently low monomer concentrations,

X]. we find that X1 >> X2 >> X3 >>. .. XN. Thus, at low concentrations, most of the
molecules will be independent monomers, such that X1 z C. Additionally, if uO’N is

relatively constant with aggregation number, entropy losses greatly favor small
aggregates. Equation 1.14 also suggests that there is a threshold concentration limit at

which this equation is no longer true. Since XN cannot be > 1, as X1 approaches the value
represented by exp[-(u°1 - uO’N)/kT], it can increase no further. This limit in monomer

concentration occurs at the CMC as shown in equation 1.15. Thus, the greater the free
energy change between monomeric and micellar surfactant molecules, the lower the
CMC. The CMC is dependent almost entirely on the change in chemical potential that
occurs as the hydrophobe moves from the water phase to the micelle interior. The value

of N at this concentration is Nmin’ the minimum number of molecules necessary to

stabilize the aggregate and produce a micelle.29

CMC z exp[-(u°1 - uO’N)/kT] eq. 1.15

14

One final aspect of this model is the relationship of enthalpy gains to micelle size.

Initially, 11°, N decreases substantially with increasing N due to entropic gains associated

with excluding the hydrophobe from water. However as the structure size increases, the
effective surface area per molecule decreases. If the head group is ionic, a repulsion
develops as head groups are forced into closer contact. As shown in Figure 1.3, the
balance of attractive and repulsive head-group interactions yields a minimum free energy

state at some optimum head group area, a0. Larger spherical aggregates have head-group

areas that are energetically unfavorable, and the size distribution of micelle structures is

centered about structures having the optimal head-group area.

 

Total
Energy

 
     
 

 

Minimum energy
a = a0

l

 

   

(——— Attractive energy

Interaction free energy, 1.10,,

K Repulsrve energy

 

 

 

Surface area per molecule, a

Figure 1.3. Optimum head group size as balance of attractive and repulsive forces.

Adapted from Israelachvili, Intermolecular and Surface Forces.28

15

Therefore from a statistical thermodynamic approach we can see that while
intermediate molecular aggregates are possible, they are not prevalent. Instead the
dispersion is composed primarily of monomer until the concentration reaches a threshold
level when a specific aggregate, the micelle, is formed. The model also shows the direct
relationship between the chemical potential change per molecule and the CMC. Finally,

this model suggests that a major factor limiting structure size is head group area.

1.6 The Critical Packing Factor Model and Aggregate Shape

In section 1.3, I described how amphiphilic molecules differ from typical solutes
by possessing both a “solvent-loving” and “solvent-hating” portion. Section 1.4 showed
that above a certain concentration these molecules form micellar structures. Structure
formation is a thermodynamic process driven by the entropy gains of water as the
hydrophobes are forced into a separate phase. The previous section (1.5) discussed the
relationship between the aggregate size and its probability of formation. In this section I
will describe how the size and shape of the hydrophobe and hydrophile influence the type
(shape) of micellar structure formed. First, I will briefly describe the possible types of
structures that can form, and then I will present a model explaining the factors that
determine which structures form from a given surfactant.

The most commonly observed aggregate structure is the spherical micelle, though
numerous other aggregate forms have been identified including elliptical, cylindrical, and
Worm-like micelles, discs, bilayers (continuous and discrete), and even reverse micelles
(Figure 1.4).30 Spherical micelles are dynamic structures that exist as unique entities for

fractions of a second before disintegrating to form new structures milliseconds later. '9'31

l6

Their short lifetimes make them more reflections of a free-energy minimum than an
actual organized unit. Increasing the size of the hydrophobe will decrease the CMC, and
can increase the stability of structures, however, the same surfactant geometric properties
must be maintained to produce spherical micelles. Typical spherical micelles have

diameters slightly less than twice the length of the fully elongated hydrophobe.

~3$§Sr§3fl9®$u$50~6

5'
IIIIIIIIIIIIII

   

Spherical Lamellar

   

Cylindrical Reverse

Figure 1.4. The four fundamental micellar self-assembled units.

Cylindrical micelles also have the hydrophobe partitioned to the structure center,
but, because of a smaller head group, they are geometrically unable to form spheres.
Instead, they produce a variety of aggregate shapes including rods, “threads” or “worms”
depending on micellar length (longest dimension) and the frequency of bends in the
structure. These micelles have a considerably larger aggregation number than spherical
micelles and therefore tend to have longer average lifetimes.”33 Rod-shaped micelles

must lose far more surfactant molecules before N drops below Nmin and spontaneous

disintegration of the structure takes place. Thus, while the rate of surfactant exchange
between solution and cylindrical micelles is similar to that with spherical micelles, the
size of the rod-shaped structures makes them exist as entities for longer periods.

Bilayer structures may be planar or flexible depending on the hydrophobe fluidity
and the size and flexibility of the hydrophile (Figure 1.5). Molecules where the head

group and hydrophobe diameters are nearly identical generally yield flat, planar

 

Rigid Flexible
Bilayer Bilayer

Figure 1.5. Two types of aggregate bilayer structures; flexible and rigid.

18

structures such as micellar discs. When the head group is slightly larger than the
hydrophobe, more flexible bilayers may create an extensive network, as in a
bicontinuous, or sponge-phase, or may form discrete spherical structures called

vesicles.For structure formation to occur, the surfactant standard state chemical potential

in an aggregate must be less than that of the monomeric surfactant, 1.1%; < [1%, (equation

1.15) for some value N. The value of N, while indicative of the number of monomers
incorporated into each structure, does not identify the structure shape or explain why the
various types of micelles form. Since self-assembly is entropically driven, one might
expect that any shape that removes a similar number of hydrophobes from the aqueous
phase would have a similar free energy change and therefore be equally likely to develop.
This is not observed however. Instead, dispersions of “pure” surfactant (low variability

in molecular composition) are typically populated by two surfactant states: the monomer

and a distinct micellar structure suggesting that uON is not equal for all types of

structures.

One model that describes why a single structure predominates, and defines the
variables that influence the type of aggregate structures a given amphiphile will produce,
is based on surfactant geometry and packing. As shown in Figure 1.1, both the
hydrophobe insolubility and the hydrophile solubility significantly impact the surfactant
CMC. Likewise, the shape of both the hydrophobe and the hydrophile impact the
geometry of the micelle formed because of the packing constraints presented by each
group. Israelachvili et al. described the constraints on micellar shape in terms of a critical
packing factor (CPF), ¢,6 which is related to the hydrophobe volume, V, the critical chain

length of the hydrophobe, IC, and the optimum cross-sectional head group area, do, as

19

shown in equation 1.16. The value of d) is indicative of the type of micellar structure the
amphiphile will form.
¢=Vllcao eq.1.16
Figure 1.6 shows how values of V, 16 and a0 limit the possible shapes that the

surfactant can assume, and thereby limit the types of aggregate structures the surfactant
can form For example, the combination of a large head group and a flexible tail of
moderate length will generally produce the conical shape. If the head group shrinks,
because of a change in salinity or temperature, the surfactant may form a truncated cone.
Similarly, if the tail is replaced with a more bulky one, due to unsaturation or branching,

its volume may increase and, again, the surfactant shape moves from conical to truncated

  

Cone

'0.

  

‘1.
i
l
3
p

   

-
~-
~" ‘ n-

Cone

Figure 1.6. Eight basic packing configurations for various surfactant molecules in a

micelle. Adapted from Israelachvili, Intermolecular and Surface Forces. 28

20

cone. Some surfactants, like phospholipids, have multiple tails, giving the surfactant a
cylindrical shape. Finally, amphiphiles with either small head groups or long or semi-
rigid tails produce inverted cones. To be more quantitative, if the CPF lies between 0 and

1I3, Israelachvili’s model predicts that optimal packing will occur when the surfactant

assumes a conical shape within a spherical micelle (Figure 1.6). (See Appendix A for
more details on the packing factor model.) This occurs because the head group, no, is
large and the tail group is narrow and flexible enough to allow the surfactant to form a
tapered shape that readily packs into spheres. This head/ tail combination is common
with mono-valent ionic surfactants having an aliphatic hydrophobe, although nonionic
head groups of comparable diameter will also produce spherical and oblate micelles.
If the head group size is reduced, by the addition of salt or by blending of ionic
and nonionic surfactants, or if a hydrophobe with a larger volume-to-length ratio is

selected, then 4) rises to between 1/3 and 1/2, and the surfactant adopts a truncated cone

shape within a cylindrical micelle. These structures have a circular cross-section, but an
elongated Z-axis.

When the head group diameter is similar to the hydrophobe diameter, then the
CPF increases to between 1/2 and 1, and the amphiphile molecules assume a cylindrical
shape. These surfactants assemble into bilayered, or lamellar, aggregates. Lamellar
structures consist of two surfactant monolayers paired such that the ends of the

hydrophobes from the first layer contact the ends of the hydrophobes from the second. A
CPF value closer to 1/2 indicates a more flexible bilayer while a value near 1 indicates a

planar structure (see Figure 1. 5). If the CPF value exceeds 1, reverse micelles form.

21

The critical hydrophobe chain length, IC, is generally taken to be about 80 % of

the tail’s fully-elongated length.34 The 20% reduction in length accounts for the normal,
random movement of the tail in a liquid environment. Other variations in chain length
and flexibility occur due to the twists and kinks associated with cis and trans
conformations as well as the extent of saturation.

While the CPF model uses the optimum head group area, ac, to predict structure
type, the model remains valid when (1 ¢ a0, however these are metastable structures and

do not represent the lowest energy state aggregates. In either case, the head group size
depends on two factors: the hydrated hydrophile diameter and the extent of interaction
between adjacent head groups. Ionic head groups are generally a single atom or a
polyatomic complex and are treated as a point charge. They repel one another and
present a slight resistance to micellization. Values of a for ionic surfactants depend on
solvent ionic strength, the presence of organics, and temperature. Ionic head groups may
also be influenced by the valency of the ions in solution.

Nonionic head groups, such as ethylene oxide, can interact and, under certain
conditions, can form well-ordered systems. They usually contain multiple hydrophilic
units rather than a single group and oppose micelle formation, but to a lesser extent than
the ionic groups. (This is evidenced by the lower CMC values for surfactants with the
same hydrophobe, but nonionic rather than ionic head groups”) Solution salinity
influences nonionic head group size, but to a lower extent than ionic groups, although
nonionic groups are more temperature sensitive.

The head group area, a, depends on the model’s location for the hydrophile/

hydrophobe interface. Once selected, the head group area then becomes the total

22

interfacial area divided by the aggregation number. While an absolute value of a is
critical to some research, it is much less important to the work I will present. Instead I
am primarily interested in how changes in head group size affect micellar structure. It is
also important to mention that while the head group is generally modeled as a sphere, this
approach is not applicable if the head “group” consists of multiple hydrophilic entities.
For nonionic hydrophiles, which may be 20 times longer than a typical ionic group and
consist of dozens of atoms, the head group “shape” is not obvious, and determination of
head group area is complicated.

The most common head groups used in this work are ethylene oxide (E0)
Oligomers. The relationship between E0 block length and head group area in micellar
structures is not first order as it is at the air-water interface.35 Instead, as the E0 block is
hydrated, it forms an expanding coil whose diameter is dependent on E0 length.36 The
most pronounced changes in coil size, and hence head group area, occur with increases to
shorter E0 blocks. Alexandridis et al. showed that the head group area for E02 to E010
varies with the square root of the number of EO units.37 Thus, using an E0 block as the

hydrophile, one can generate various values of a0 simply by changing the block length

during synthesis. The hydrophilicity of EO head groups decrease with increasing
temperature due to changes to the water structure around the E0 groups, changes in the
hydrogen bonding to the E0 groups and changes in the conformation of the E0 group.‘9
All contribute to reduce the E0 head group area at higher temperatures. Note that (I) is

inversely proportional to no, so decreasing the head group area is expected to produce

larger structures. (See appendix A)

23

1.7 Vesicles

Vesicles form when selected surfactants self-assemble into a bilayer that curves in
on itself to form a hollow, spherical nano-structure. The bilayers form closed structures
to reduce contact between the water and the hydrophobic edge of the bilayer. While all
bilayer structures suffer the free-energy costs of the “edge effect”, only flexible lamellar
structures can negotiate the bending necessary to form a closed structure and eliminate
bilayer edges.

Vesicles may consist of a single or multiple bilayers (Figure 1.7). Single-bilayer
vesicles are divided into three groups; small unilamellar vesicles (SUVs) with diameters
of less than 100 nm, large unilamellar vesicles (LUVs) with diameters from 100 nm to 10
um and giant unilamellar vesicles (GUVs) with diameters reported of up to
approximately 5 mm.”39 While more time-consuming to prepare, LUVs offer an
efficiency advantage over SUVs in terms of contained volume per mole of

phospholipid.40 Multi-lamellar vesicles (MLV) are composed of up to hundreds of

 

Figure 1.7. Two types of closed bilayer structures; uni-lamellar (left) and multi-lamellar

(right).

24

concentric bilayer spheres. These structures are rarely less than 250 nm in diameter, but
generally have the same upper size limits as LUVs. The separation between bilayers in
an MLV is not fixed, and the number of layers contained in two MLVs of the same
diameter could vary substantially.

Vesicles are of particular interest because of their encapsulation properties. They
are found extensively in biological systems as cell membranes and display substantial
barrier properties. The bilayer is robust, resilient and self-repairing, making these
structures well suited for isolating their contents from the dispersing medium. The
structure’s aqueous core contains hydrophilic materials, while the bilayer interior
sequesters hydrophobic while hydrophobic molecules (Figure 1.8). The bilayer can be
modified for selected permeability, controlled release and system compatibility.

Although the question of whether or not vesicles are thermodynamically stable

“'42 they can achieve a steady-state equilibrium with the bulk

structures is debatable,
dispersion, and substantial shelf lives have been observed that range from weeks to years.
While individual surfactant molecules may move laterally within a layer, the exchange
rate between the two layers is generally low and the exchange rate with aqueous phase
monomers is lower yet.43 For these reasons and others, vesicles are considered ideal
candidates for encapsulation research.

Laboratory preparation of vesicles was first reported by Bangham, and
encapsulation applications quickly followed.44 The use of vesicles as drug delivery
agents was soon postulated and has recently received significant attention.‘“”46 Fendler

demonstrated the encapsulation of several amino acids and studied the effects of

electrolyte and various encapsulated materials on vesicle stability.47 Researchers have

25

used giant vesicles, which due to their low radius of curvature place less strain on the
bilayer, as tools for studying selective transport across cell membranes.48 Vesicles laden
with contrast agents such as radiolabeled monoclonal antibodies, ferrites or paramagnetic
materials, have successfully increased imaging quality in physiological studies.49
Research has also been conducted using structures as modified solar collectors by loading
the vesicles with photo-reactive materials.9 Others have explored using vesicles as a
blood substitute by packing the structures with hemoglobin.50 Work by Meier has
recently shown that modified vesicles can function as enzymatic nano-reactors.5 ' Natural
lipids, particularly phospholipids and lipid/ cholesterol mixtures, are the most commonly

used materials for vesicle formation. A phospholipid consists of a phosphate hydrophile

Hydrophilic

materials

3."; .
, ._ -i§
\ .

 

Figure 1.8. Two regions of vesicle structure are potential encapsulation areas.

26

and generally a diglyceride containing two similar-length aliphatic hydrophobes.
Variations in hydrophobe length and saturation, as well as variations in head group
substituents, produce a wide range of phospholipids. As a result, numerous vesicular
structures of varying sizes and stabilities can be produced from these materials.
Vesicle generation processes are often time-consuming or elaborate.“40
Typically the phospholipid is mixed with other lipids or cholesterol to modify the
physical characteristics of the bilayer and make vesicle formation easier. In a common
preparation procedure, the mixture is dissolved in an organic solvent, filtered, cast as a
film and dried (frequently freeze-dried) before phase separation of the phospholipid from
the additives can occur. The dried film is then slowly hydrated with a solution of salts,
buffers, and other additives and/ or stabilizers to produce structures. After initial
swelling the mixture can be hand-shaken or ultra-sonicated to accelerate structure
formation. More elaborate processes, which involve aggressive agitation, extensive
dialysis, solvent exchange, solvent injection and even oscillating electric fields, are
common. Finally, the structures are homogenized using extrusion, freeze/ thaw cycles or
sonication to produce a narrow size distribution. Once homogenized, the vesicles can
also be readily separated from residual materials in the aqueous phase by mechanical or
gel filtration. Final structures can be characterized using aqueous size exclusion
chromatography, light scattering techniques, optical microscopy, cryo-scanning electron
microscopy and cryo-transmission electron microscopy.

Although formation of vesicles from phospholipids is common, lipids possess
inherent chemical and mechanical limitations restricting their applicability as

encapsulants.52 Besides the preparation difficulties, the structures tend to be temperature,

27

shear and saline sensitive making them vulnerable in certain environments. The
phosphate head group, like most ionic species, is strongly influenced by the presence of
salts in the aqueous phase,53 and certain solvents can selectively remove the additives,
destabilizing the structures. The hydrophobes frequently have a phase transition

1.31 Attempts have been made over

temperature that limits their thermal stability as wel
the past four decades to improve the stability of phospholipid-based vesicles by chemical
modification of the lipid and in situ polymerization of the structure.“55 However, these

have met with limited success and have, in turn, introduced new complications to

structural stability.

1.8 Block Copolymers

Some limitations of phospholipid vesicles can be overcome by using amphiphilic
block copolymers as surfactants. Block copolymer synthesis can be a simple, well-
understood process that involves a single step per block, and typical syntheses have high
yields and do not require extensive product purification. Thus, in most cases, copolymers
are much less expensive than phospholipids. Block copolymers can also be synthesized
with specific physical properties, such as morphology or melting point, by selecting the
appropriate hydrophobic or hydrophilic monomer. The asset of copolymers which is of
most interest to this work, however, is the ability to produce surfactants with varying
geometric shapes by appropriate selection of monomer type and block length.

Block copolymer surfactants can form a variety of aggregate structures. Using a
solvent exchange method, Laibin and Eisenberg showed that polystyrene! poly(acrylic

acid) diblock compositions form a myriad of structures and complex shapes in an

28

aqueous system.“5 Alexandridis studied ethylene oxide/ propylene oxide/ ethylene oxide
(EO/PO/EO) mixtures in xylene/ water systems and found numerous aggregate structures
such as spherical and rod-shaped micelles, lamellae and reverse micelles.57 Ding and Liu
produced vesicles in solvent mixtures of THF and hexane using polyisoprene/ poly(2-
cinnamoylethyl methacrylate),58 and Meier prepared vesicles in water with several siloxyl
containing copolymers.59 Using poly(amino acids), Tirrell has produced materials that
form vesicles, as well as other structures, in water.60 In spite of these successes,
preparation of a copolymer specifically for vesicle formation is not trivial.

There remains an unmet need for a material which is easy to synthesize, bio-
compatible and which forms structures without the need for solvents. The ideal
compound could be tailored to produce range of structure sizes for various applications.
A copolymer that would allow for preparation of elevated vesicle concentrations would
advantageous as well. Additionally, improved stability, high loading capacities and

increased release rate control would be beneficial in any encapsulation application.

1.9 1,2-Butylene oxide/ Ethylene oxide Diblock Copolymers

The previous examples show that both ionic and nonionic surfactants can form
vesicles, however nonionic surfactants offer several advantages over ionic materials.
Nonionic copolymers demonstrate lower pH and saline sensitivity than ionic surfactants,
and vesicles produced from nonionic copolymers would be expected to show this same
reduced sensitivity.61 Nonionic amphiphiles generally have lower CMCs, so vesicles can
be prepared at lower concentrations with a higher fraction of surfactant in the structures.

Ethoxylated copolymers have inherent biocompatibility, making them appealing as drug

29

delivery agents.62 Finally, the head group size of an ethoxylated material can be easily
altered by adjusting the ethoxylate block length during synthesis.‘51

Several groups examined structure formation with ethoxylated copolymers.
Maskos and Harris demonstrated vesicle formation from 1,2-butadiene/ ethylene oxide
copolymers and successfully cross-linked these structures after formation.63 Discher et
al. demonstrated large vesicle formation from ethyl ethylene/ ethylene oxide diblock
copolymers and showed that these structures display mechanical stabilities 10 times
greater than phospholipid-based structures.64 Alexandridis studied butylene oxide/
ethylene oxide (BO/BO) diblock copolymers and found that they formed more than six
different structural phases in a ternary copolymer/ water/ xylene system, although no
vesicles were observed,65 while Booth focused on using BO/EO copolymers as gel
formers for possible application as drug delivery agents.66'67 He extensively explored
BO/EO micellar behavior, however be consistently worked with materials that have a
larger head group than the CPF model predicts will form vesicles.

The 1.2-butylene oxide (BO) group offers several advantages as a hydrophobe. A
BO unit displays about the same free energy change on micellization as a methylene
group making it a substantial hydrophobe.‘58 Hydrophobe length can be controlled using
anionic polymerization and the BO block remains a liquid at molecular weights of more
than 3000 amu. Additionally, the B0 monomer is less prone to rearrangement than
propylene oxide, so control of the hydrophobe shape is superior.‘59 However, the BO
volume-to-length ratio is the biggest factor in choosing it as a hydrophobe.

The CPF model indicates that phospholipids form vesicles because the large

volume-to-length ratio of their two-tailed hydrophobe produces a cyclindrical surfactant

30

geometry that promotes bilayer forrmtion. I believe that the large volume-to-length ratio
of the BO hydrophobe can produce the same results. However, the hydrophobe must be
matched with the proper size head group for vesicle formation to occur. Therefore I
surmised that butylene oxide/ ethylene oxide diblock copolymers would be potential
vesicle formers when prepared with the correct block lengths.

This work addresses the synthesis and characterization of low molecular weight
butylene oxide/ ethylene oxide diblock copolymers and examines their ability to form
vesicles in water. Remarkably, micron-sized vesicles form upon simple mixing of certain
copolymers with water. Vesicle formation is, however, a function of block length,
particularly the E0 block length, and the CPF model correlates well the vesicle-forming
properties of different copolymers. BO/EO vesicles are stable over months and even
years, and release rates from these materials are very slow. One area that I have only
begun addressing is the development of polymerizable BO/EO surfactants. Given the
simplicity of vesicle formation form BO/EO copolymers, I anticipate that these materials

may have applications in a variety of encapsulation applications.

1.10 References

(1) Jacoby, M. "Nanoscale Electronics"; Chemical and Engineering News, September
30, 2002, PP 38-43

(2) Borman, S. "Chemistry Highlights 2001 "; Chemical and Engineering News,
December 10, 2001, pp 45-55

(3) Borman, S. "Chemistry Highlights 2002"; Chemical and Engineering News,
December 16, 2002, pp 42-47

31

(4)

(5)

(6)

(7)

(8)

(9)

(10)

(11)

(12)

(13)

(14)

(15)

Sauer, M.; Streich, D.; Meier, W. Advanced Materials (Weinheim, Germany) 2001,
13, 1649-1651.

Tanev, P. T.; Pinnavaia, T. J. Science (Washington, D. C.) 1996, 271, 1267-1269.
Israelachvilli, J. N.; Mitchell, D. J.; Ninham, B. W. J. Chem. Soc., Faraday Trans. 2
1976, 72, 1525-1568.

Fairley, P. "Nano Technology: The Start of Something Big"; Chemical Week,

December 12, 2001, pg 23-26

Mueller, P.; Rudin, D. 0.; Tien, H. T.; Wescott, W. C. Formation and Properties of
Bimolecular Lipid Membranes. In Recent Progress in Surface Science; Danelli, J.
F., Pankhurst, K. G. A., Riddiford, A. C., Ed.; Academic Press: New York, NY,
1964; Vol. 1, pp 379-393.

Sisido, M. Photoenergy harvesting on two-dimensional vesicular assemblies. In
Vesicles, First ed.; Rosoff, M., Ed.; Marcel Dekker: New York, NY, 1996; Vol. 62,
pp 563-592.

Tilcock, C. Liposome Technol. (2nd Ed.) 1993, 65-87.

Chang, T. M. S. Journal of Internal Medicine 2003, 253, 527-535.

Gerasimov, O. V.; Rui, Y.; Thompson, D. H. Triggered Release from Liposomes
Mediated by Physically and Chemically Induced Phase Transistions. In Vesicles,
First ed.; Rosoff, M., Ed.; Marcel Dekker: New York, 1996; Vol. 62, pp 679-746.

Lasic, D. D. Liposomes in Drug Delivery. In Vesicles, First ed.; Rosoff, M., Ed.;
Marcel Dekker: New York, 1996; Vol. 62, pg 448.

Dean, J. M.; Lipic, P. M.; Grubbs, R. B.; Cook, R. F.; Bates, F. S. Journal of
Polymer Science, Part B: Polymer Physics 2001, 39, 2996-3010.

Israelachvili, J. N. Intermolecular and Surface Forces, 2nd ed.; Academic Press:
San Diego, CA, 1992; pp 16-21.

32

(16)

(17)

(18)

(19)

(20)

(21)

(22)

(23)

(24)

(25)

(26)

Clint, J. H. Surfactant Aggregation; Chapman & Hall: New York, NY, 1992; pp 1,
2.

Myers, D. Surface Science and Technology; VCH Publishers: New York, NY,
1988; pg 106.

Rosen, M. J. Surfactants and Interfacial Phenomena, Second ed.; Wiley-
Interscience: New York, NY, 1989; pp 120-133.

Clint, J. H. Surfactant Aggregation; Chapman & Hall: New York, NY, 1992; pp
100-115.

Myers, D. Surface Science and Technology; VCH Publishers: New York, NY,
1988; pg 96.

Molyneux, P.; Rhodes, C. T.; Swarbrick, J. Transactions of the Faraday Society
1965, 61, 1043-1052.

Birdi, K. S. ACS Symposium Series 1975, 9, 233-238.

Aranow, R. H.; Witten, L. Journal of Physical Chemistry 1960, 64, 1643-1648.

Franks, H. S.; Evans, M. W. Journal of Physical Chemistry 1945, I3, 507.

Shinoda, K. Journal of Physical Chemistry 1977, 81, 1300-1302.

Evans, D. F. Langmuir 1988, 4, 3-12.

(27) Beesley, A. H.; Evans, D. F.; Laughlin, R. G. Journal of Physical Chemistry 1988,

92, 791-793.

(28) Israelachvili, J. N. Intermolecular and Surface Forces, 2nd ed.; Academic Press:

San Diego, CA, 1992; pp 345-372.

33

(29)

(30)

(31)

(32)

(33)

(34)

(35)

(36)

(37)

(38)

(39)

(40)

Wall, S. N .; Aniansson, G. E. A. Journal of Physical Chemistry 1980, 84, 727-736.

Rosen, M. J. Surfactants and Interfacial Phenomena, Second ed.; Wiley-
Interscience: New York, NY, 1989; pp 112-114.

Fendler, J. H. Membrane Mimetic Chemistry: Characterizations and Applications
of Micelles, Microemulsions, Monolayers, Bilayers, Vesicles, Host-Guest Systems,
and Polyions; John Wiley and Sons: New York, NY, 1982; pp 25, 26.

Shioi, A.; Harada, M.; Adachi, M. Recent Research Developments in Physical
Chemistry 1999, 3, 149-182.

Aniansson, G. E. A. Progress in Colloid & Polymer Science 1985, 70, 2-5.

Tanford, C. The Hydrophobic Efi‘ect: Formation of Micelles and Biological
Membranes, Second ed.; WIley-Intersciences: New York, NY, 1980; pg 52.

Schick, M. J. J. Colloid Sci. 1962, 17, 801-813.

Schick, M. J. J. Am. Oil Chemists' Soc. 1963, 40, 680-687.

Alexandridis, P.; Athanassiou, V.; Fukuda, S.; Hatton, T. A. Langmuir 1994, 10,
2604-2612.

Betageri, G. V.; Jenkins, S. A.; Parsons, D. L. Liposome Drug Delivery Systems;
Technomic Publishing Company Inc: Lancaster, PA, 1993; pp 1,2.

Tien, H. T. Journal of Physical Chemistry 1967, 71, 3395-3401.

Betageri, G. V.; Jenkins, S. A.; Parsons, D. L. Liposome Drug Delivery Systems;
Technomic Publishing Company Inc: Lancaster, PA, 1993; pp 7- 17.

34

(41) MacDonald, R. C. The Relationship and Interactions Between Lipid Bilayer
Vesicles and Lipid Monolayers at the Air/Water Interface. In Vesicles, First ed.;
Rosoff, M., Ed; Marcel Dekker: New York, NY, 1996; Vol. 62, pp 3-48.

(42) Gradzielski, M. Journal of Physics: Condensed Matter 2003, 15, R655-R697.

(43) Fendler, J. H. Membrane Mimetic Chemistry: Characterizations and Applications
of Micelles, Microemulsions, Monolayers, Bilayers, Vesicles, Host-Guest Systems,
and Polyions; John Wiley and Sons: New York, NY, 1982; pp 145-153.

(44) Bangham, A. D.; Standish, M. M.; Watkins, J. C. Journal of Molecular Biology
1965, 13, 238-252.

(45) Gregoriadis, G. Fate of injected liposomes: Observations on entrapped solute
retention, vesicle clearance and tissue distribution in vivo. In Liposomes as Drug
Carriers. Recent Trends and Progress; Gregoriadis, G., Ed.; Wiley & Sons: New
York, NY, 1988; pp 176-188.

(46) Lasic, D. D. Liposomes in Drug Delivery. In Vesicles, First ed.; Rosoff, M., Ed.;
Marcel Dekker: New York, NY, 1996; Vol. 62, pp 447-476.

(47) Tran, C. D.; Klahn, P. L.; Romero, A.; Fendler, J. H. Journal of the American
Chemical Society 1978, 100, 1622-1624.

(48) Tien, H. T. Bilayer Lipid Membranes (BLM), Theory and Practice, 1974; pg 672

(49) Tilcock, C.; Utkhede, D.; Meng, G. Vesicles as imaging agents. In Vesicles, First
ed.; Rosoff, M., Ed.; Marcel Dekker: New York, NY, 1996; Vol. 62, pp 593-648.

(50) Takeoka, S.; Ohgushi, T.; Sakai, H.; Kose, T.; Nishide, H.; Tsuchida, E. Artificial
Cells, Blood Substitutes, and Immobilization Biotechnology 1997, 25 , 31-41.

(51) Graff, A.; Winterhalter, M.; Meier, W. Langmuir 2001, 17, 919-923.

35

(52)

(53)

(54)

(55)

(56)

(57)

(58)

(59)

(60)

(61)

(62)

(63)

Gerasimov, O. V.; Rui, Y.; Thompson, D. H. Triggered Release from Liposomes
Mediated by Physically and Chemically Induced Phase Transitions. In Vesicles,
First ed.; Rosoff, M., Ed.; Marcel Dekker: New York, 1996; Vol. 62, pp 679-746.

Hianik, T.; Ottova-Leitmannova, A.; Tien, H. T. Surfactant Science Series 1996,
62, 49-77.

Sisson, T. M.; Srisiri, W.; O'Brien, D. F. Journal of the American Chemical Society
1998, 120, 2322-2329.

Sells, T. D.; O'Brien, D. F. Macromolecules 1994, 27, 226-233.

Laibin, L.; Eisenberg, A. Journal of the American Chemical Society 2001, 123,
1012-1013.

Alexandridis, P.; Olsson, U.; Lindman, B. Langmuir 1998, 14, 2627-2638.

Liu, G.; Ding, J. Journal of Physical Chemistry B 1998, 102, 6107-61 13.

Nardin, C.; Hirt, T.; Leukel, J .; Meier, W. Langmuir 2000, 16, 1035-1041.

Haverstick, K.; Pakalns, T.; Yu, Y.-C.; McCarthy, J. B.; Fields, G. B.; Tirrell, M.
Polymeric Materials Science and Engineering 1997, 77, 584,585.

Rosen, M. J. Surfactants and Interfacial Phenomena, Second ed.; Wiley-
Interscience: New York, NY, 1989; pp 23-31.

Watson, K. J .; Anderson, D. R.; Nguyen, S. T. Macromolecules 2001, 34, 3507-
3509.

Maskos, M.; Harris, J. R. Macromolecular Rapid Communications 2001, 22, 271-
273.

36

(64) Discher, B. M.; Won, Y.-Y.; Ege, D. S.; Lee, J. C.-M.; Bates, F. S.; Discher, D. E.;
Hammer, D. A. Science 1999, 284.

(65) Alexandridis, P.; Olsson, U.; Lindman, B. Langmuir 1997, 13, 23-34.

(66) Kelarakis, A.; Havredaki, V.; Yu, G.-E.; Derici, L.; Booth, C. Macromolecules
1998, 31, 944-946.

(67) Pople, J. A.; Hamley, I. W.; Fairclough, J. P. A.; Ryan, A. J.; Komanschek, B. U.;
Gleeson, A. J .; Yu, G. E.; Booth, C. Macromolecules 1997, 30, 5721-5728.

(68) Derici, L.; Ledger, 8.; Mai, S.-M.; Booth, C.; Hamley, I. W.; Pedersen, J. S. Phys.
Chem. Chem. Phys. 1999, I, 2773-2785.

(69) Yu, G.-E.; Masters, A. J.; Heatley, F.; Booth, C.; Blease, T. G. Macromol. Chem.
Phys. 1994, 195, 1517-1538.

37

Chapter 2

Synthesis and Characterization of 1,2-Butylene
Oxide! Ethylene Oxide Diblock Copolymers

2.1 Overview of Synthesis

The 1,2-butylene oxide! ethylene oxide (BO/E0) diblock copolymers used in this
work were prepared by sequential anionic polymerization, which exhibits pseudo-living
characteristics for low molecular weight blocks.l Synthesis began with dissolution of the
alkoxide initiator in toluene, addition of butylene oxide monomer and polymerization of

the B0 block (Scheme 2.1). The E0 monomer was then added and polymerized from the

 

O
CH3CH2CHCH2 I
CH3CH20+K T l D CH3CH20[CH2CH(CH2CH3)O-]n +K
0 “CDC

0
1V
H2CH2C
II

CH3CH20[CH2CH(CH2CH3)O]“[CHZCHZO']m +K

11V
111

CH3CHZO[CHZCH(CH2CH3)O]"[CHZCH20]m-H + KOAc l

Scheme 2.1. Synthesis of 1,2-butylene oxide! ethylene oxide copolymers; (1) BO
oligomer formation, (H) BO oligomer with active E0 site, (III) acid quenching of

reaction.

38

BO oligomer, and the reaction was finally quenched by the addition of a strong acid.
This synthesis affords a diblock copolymer with an aliphatic-initiated hydrophobe and a
hydroxyl-terminated hydrophile (Figure 2.1). The terminal hydroxyl functionality

strongly influences the hydrophilic nature of the head group.2
0 0x 0% O O O O
/\ \K \/\ N V\ /V V\ /V \
O O O O O O O H
Figure 2.1. Representative BO/EO molecule showing amphiphile shape.

Copolymer composition dramatically affects surfactant performance and is always
a concern.3 While this synthesis produces the desired end groups, inefficient coupling of
EO groups to the BO oligomer can lead to uncapped, residual BO oligomer.
Additionally, as in all copolymer syntheses, the product is not homogenous, but rather a
mixture of copolymers having EO and B0 blocks of varying lengths. Since EO length
strongly affects structure formation,4 characterizing the copolymer is critical to
understanding and predicting aggregation. Finally, byproducts or contaminants must be
identified to ensure the desired copolymer performance.

This chapter will discuss the synthesis of BO/EO Oligomers and the subsequent
copolymer purification and characterization. I will address specific characterization
procedures, highlighting the advantages and drawbacks of each. Finally, I will present
profiles of the copolymer compositional distribution and explain why the methodology

used to determine the “average composition” was chosen.

39

2.2 Parr Reactor Synthesis

This work began by examining a range of BO/EO copolymers previously
prepared by The Dow Chemical Company using a lab-scale reactor. Lab-scale batches
are denoted with the prefix “L“ and were synthesized using either a 450 or a 600 mL Parr
Reactor (Parr Instruments, Moline, IL). The reactors consist of two components: a head
unit, which contains the stirrer, thermocouple well, cooling line and assorted ports, and a
base, which is the actual reaction vessel. Both reactors use the A1120HC head unit, but
the 450 mL reactor uses base unit 452HC2, while the 600 mL reactor uses base unit
452HC3. The polymerizations were performed in batch mode, with all of the monomer
added and mixed before heating began. Typical copolymer yields were between 50 and
200 mL per batch or about 90 % of added monomer.

The Parr Reactor is equipped with a temperature controller that continuously
adjusts the heating rate, and the temperature is monitored using redundant thermocouples
inserted into a well filled with silicone oil. The controller also monitors the reactor
pressure and controls a magnetically-coupled, shaft-driven stir rod. A cooling line
circulates a refrigerated solution through a coil in the reaction solution for more precise
temperature control and acts as a fail-safe device in case of an uncontrolled, exothermic
reaction. The reactor also has a sampling port to extract product, a blanketing port for
pressurizing the headspace above the solution, and an injection port for introducing
materials into the reactor. A 750-psi frangible disk on the head unit prevents over
pressurization, and a 0—300 psi pressure gauge serves as back-up to the pressure

transducer.

40

Ethylene oxide (EO, liquefied gas, 99.5+%), 1,2-butylene oxide (BO, liquid,
99+%), toluene (HPLC grade), and potassium ethoxide (powder, 95%), were purchased
from Sigma-Aldrich. Dry molecular sieves were added to the toluene upon receipt and to
the BO approximately 24 hour before use. Otherwise, materials were used as received.
The quality of the E0 supply was erratic, and oligomerization in the cylinder was
common. To test monomer quality, ~ 0.5 g of the material was injected into a glass bottle
and allowed to evaporate. The bottle was inspected for a residue that would indicate
oligomerization. Several synthesis runs had to be aborted due to oligomerized E0.

Variations in alkoxyl block lengths, which are unavoidable in this type of
polymerization, can be reduced by using low monomer concentrations and only alkoxide
salts as chain initiators. Such conditions were used in all of the lab-scale syntheses. (The
conjugate alcohol for a given alkoxide is frequently added to the reactor to provide an
inexpensive source of additional initiation sites, but these sites are active for only a
portion of the total synthesis time and thus, contribute to E0 capping inefficiencies.)

An example of the lab-scale synthesis procedure is as follows: the potassium
ethoxide initiator (12.63 g), dried toluene (250 g) and 1,2-butylene oxide (110 g) were
added to an oven-dried 600 mL Parr reactor in a dry box blanketed with dry N2, and the
reactor was immediately sealed. The reactor was then purged with N2 for 5 min,
blanketed with 30 psi of N2 and heated to 110 °C with constant stirring. Within 2 to 5
minutes of reaching 110 °C, a controlled, exothermic reaction reaching ~ 115 °C was
observed. The reactor was adjusted to maintain 115 °C for about 4 hours, then cooled to
5 °C (6 °C below the boiling point of E0), and approximately 5 mL of the B0 oligomer

solution was removed. (The N2 pressure inside the reactor forces the oligomer out the

41

sampling port so opening the reactor is not necessary to remove this aliquot.) Next, the

N 2 pressure was dropped to ~ 2 psi and liquid E0 (75 g) was transferred, using the

internal pressure of the E0 cylinder, from the supply cylinder to the reactor via 1/8”
stainless steel tubing. Caution: E0 is a suspected teratagen and should be handled with
care.5 The reactor was re-heated to 90 °C, and a second controlled exothermic reaction
reaching about 120 °C was observed and allowed to run to completion, ~ 2 hours, before
cooling the reactor back to room temperature. At this point, the reaction was either
complete, or a portion of the product could be removed via the sampling port and a
second addition of E0 to the remaining product could be conducted.

In this example, a second E0 addition was performed after half (~ 240 g) of the
initial product was removed. The reaction of the removed material was terminated by
addition of 2 mL of hydrochloric acid (12 N), and the copolymer was set aside. A second
aliquot of E0 (18 g) was added to the remaining copolymer, and the reactor was re-
pressurized with N2 to about 30 psi and re-heated to 90 °C. Another exothermic reaction
reaching 120 °C was observed, and the reaction ran to completion in ~ 2 hours. The
reactor was cooled to room temperature, purged with N2 and the reaction terminated by
injection of hydrochloric acid. The product was transferred to a glass container to stand
overnight.

The copolymer was extracted using a toluene! 3 M NaCl aqueous dispersion and
decantation of the organic, upper phase. The solvent was removed by evaporation. If salt
was found in the product, the copolymer was mixed 1:1 with DI water and the resulting
emulsion heated until phase separation occurred (~ 80 °C). The organic phase was

decanted and the residual water removed by evaporation. Synthesis yield was calculated

42

by dividing the final copolymer mass by the total mass of monomers added and
multiplying by 100. This synthesis produced 96 g from the first E0 addition (Batch#
L1150-51A) and 82 g from the second addition (Batch# L1150-51B) for a combined
yield of greater than 85 %.

Aqueous samples of various concentrations were prepared from each copolymer
listed in Table 2.1 and allowed to equilibrate for ~ 20 days at room temperature. These
samples were then examined using white-light and laser-light scattering and plane-
polarized light microscopy (see section 3.3) to ascertain which materials demonstrated
micellar structure formation properties. (The strong hydrophobic nature of some
copolymers lead to dispersions and emulsion formation rather than micellar structures.)

Those that formed structures were examined further.

Table 2.1. B0/EO copolymers screened for evidence of micellar formation.

 

 

 

 

 

 

 

 

 

 

 

 

 

Theor. Actual Theor. E0 Actual
Sample BO block BO block block lengthII E0 block Wt % PEO

Name lergth' length" length“
L1573-9a 6.3 6.2 5.1 5.1 30
L120-15a 9.8 10 10 10.2 35
L1150-51a 9.8 10 11 11.1 40
L1150-51b 9.8 10 18 18.2 51
L1 150-55 13.2 14 23 22.7 50
L1573-65b 13.2 14 35 35.0 60
L120-53a 16.7 17 15 14.8 35
L741-45b 16.7 17 44 44.4 61
L120-29b 20.2 21 50 50.0 60
L1150-21a 23.6 24 23 17.0 30
L1 150-21b 23.6 24 40 39.7 50
L1 150-21c 23.6 24 93 92.6 70
L741-68b 34.0 34 101 101 64

 

 

 

 

 

 

 

 

3 Calculated from ratio of monomer to initiator

b Average value determined from NMR spectra. (see Section 2.4 for details)

9 Based on GPC calculations of total MW using NMR data for B0 MW.
(see Section 2.6 for details)

43

2.3 PolyGlycols Mini-Plant Synthesis

To prepare larger quantities of copolymer, and to evaluate the feasibility of large-
scale production, several B0/E0 batches were prepared at the PolyGlycols Mini-plant
Facility of The Dow Chemical Company in Freeport, Texas. The copolymer was
prepared using conditions as close as possible to those that could be used in a commercial
operation. The mini-plant employed a continuous addition, solvent-free synthesis in
contrast to the batch, solution process used in the lab-scale runs. The mini-plant
procedure also held the copolymer at high temperatures for longer periods that are
frequently more than twice as long as reaction times in the lab-scale synthesis.
Polymerizations were performed in an 8-L reactor and typically produced between 2.5
and 3.5 L of copolymer per batch. Using the mini-plant process, multiple additions of E0
were conducted to produce copolymers with sequentially longer E0 blocks from the
same B0 Oligomers. This resulted in a series of copolymers that differed only in their
E0 block lengths.

Batches prepared by the mini-plant are denoted by the prefix “M”. The recipe
and general synthesis procedure were provided to the reactor operator. In a
representative procedure, 227 g of 24 wt % potassium ethoxide in ethanol (Sigma-
Aldrich, ~ 4.94 mol of initiator sites) was used as the initiator. The reactor was loaded
with the initiator solution and 20 g of the B0 monomer and heated to 120 °C with
constant agitation. Over the next 6 hours, additional monomer was added at a rate of 10
g/ min for a total of 3710 g (51.5 mol) of B0 monomer. The material was held at 120 °C
for another 3 hours, and the reactor was allowed to cool and stand at room temperature

overnight while under N2. The following day, 2435 g (55.3 mol) of E0 monomer was

added over a 4-hour period while the reactor was held at 120 °C. (The rate of addition
automatically varied with the reactor internal pressure, which decreased as monomer was
consumed by the reaction.) The material was held at 120 °C for 4 more hours, and the
reactor was allowed to cool to room temperature. Continuous stirring continued

overnight under flowing N2. The following morning the product was removed, tested,

labeled and shipped.

Generally, fundamental research was conducted using materials produced from
the Parr reactors, while studies of dispersion properties, such as stability and
encapsulation capacity, were conducted on materials from the mini-plant. Any variation
from this pattern in the following discussions will be noted. In either case, the copolymer
composition was determined using the same characterization processes. Specific

purification procedures are described below.

2.4 Copolymer Purification

Copolymer purification was required to remove residual solvent, unreacted
monomer, and contaminants from the synthesis mixture. Due to the slight possibility of
residual ethylene oxide monomer, care was taken to minimize exposure to the copolymer
prior to complete purification. Low E0 capping efficiency is the primary source of
contamination in this work, as it yields B0 oligomers with no E0 head group.6 Butylene
oxide polymerizes much more slowly than does ethylene oxide, primarily because of
steric hindrance. However, once the B0 oligomer reacts with an ethylene oxide unit,
steric hindrance is minimal, and EO-capped oligomers tend to grow at the expense of

uncapped B0 oligomers. As a result, the final product contains some fraction of non-

45

ethoxylated B0 oligomers, while capped oligomers have E0 blocks that are longer than
would be expected from the reactant ratio. Thus the slow initiation rate and low capping
efficiency lead to polydispersity in E0 block lengths.

Monomer isomerization and product oxidation can also affect the final copolymer
composition, but these unwanted reactions can be reduced by minimizing both synthesis
temperature and duration. E0 does not undergo rearrangement, but B0 can form crotyl
alcohol when the anionic active site on a growing chain extracts a proton from a B0
monomer instead of opening the epoxide ring. This process terminates the growing chain
and transfers the anionic site to the monomer, which becomes unstable and rearranges to
produce CH3CH=CHCH20-, ionized crotyl alcohol. The anionic crotyl alcohol acts as an
initiator and starts a new oligomer chain. Thus, in principle, crotyl alcohol formation can
significantly reduce the average B0 block length by premature block termination and an
increase in the number of oligomeric chains. Nevertheless, rearrangement of B0 is a
relatively rare process. Crotyl alcohol can be detected by NMR, but was not found in any
of the copolymers studied.

Copolymer oxidation cleaves the ether bond and produces peroxides, aldehydes
and ketones.6 This process takes place in the presence of strong base and oxygen at
elevated temperatures, so minimizing exposure to these conditions is important. Hence,

the reactor was blanketed with N2 to limit exposure to 02 and H20. As with
isomerization, oxidation products can be detected by NMR if they are present at levels
high enough to affect surfactant performance. We did not observe any oxidation products

using proton NMR, but 13C NMR showed trace levels of aldehyde formation, particularly

in the mini-plant materials.

46

The previously described extraction process used in the lab-scale synthesis
(Section 2.2) was not used with the mini-plant batches because of its cost. Instead, mini-
plant batches stood undisturbed for at least 48 hours before purification. During this time
most copolymers separated into a clear, upper phase and a much smaller Opaque, lower
phase, although in some cases formation of multiple phases was observed. Phases were
isolated by decanting and labeled with a suffix to denote position, “A” being uppermost,
i.e. “M 1038A”. The copolymer was then vacuum filtered to remove any residual salts
such as potassium acetate. (Viscous copolymer batches were thinned with hexane to
reduce viscosity and make filtration easier.) After paper filtration, the surfactant was
vacuum filtered through 40-60 mesh sintered glass filters. Any residual volatiles were
removed by rotary evaporation at a minimum of 75 °C for 2 hours. (If hexane had been
added to thin the copolymer, it was removed in this step.) NMR was used to confirm
solvent removal and estimate product composition.

Samples that displayed discoloration or did not appear homogeneous after solvent
removal, were purified by centrifugation. These mixtures were centrifuged (IEC, Centra
GP8R, Needham Heights, MA) at 3400 rpm for between 1 and 24 hours to isolate the
contaminant or induce phase separation. Salt contamination was apparent as “dust” on
the centrifuge tube bottom, while other contaminants were visible as phase-separated
components. The clarified material was decanted, and the contaminant was discarded.
These same purification procedures were used on copolymers that I later synthesized
using the Parr reactor.

The following is a list of the copolymers prepared at the PolyGlycols mini-plant.

The number of phases identifies the extent of separation that occurred in the batch over

47

time. Copolymer phase separation is generally attributed to residual salt which can
complex with the E0 chain, or to crystallization of the longer (> 25 units) E0 blocks.
Batches M1190, M2013 and M2075 were a series of copolymers made with sequentially
increasing E0 block lengths. Many of the compositions with shorter E0 blocks in these
runs were single phase while those having E0 blocks greater than 16 tended to form 2
phases. The average B0 block length differed among the three series so that the effect of

both BO and E0 block lengths on structure formation could be studied.

Table 2.2. Copolymers prepared by the Mini-plant Process.

 

 

 

 

 

 

 

Expected Actual Number of

Batch Composition Composition“ phases
M1038 E0128012 E01280” 4
M1096 E012B012 E011B0“ 4
M1101 E0128012 EOHBOIO 3
M 1 190 series E010_243010 E08,22B09 ~ 2 ea
M2013 series £010.24B020 E010_26B01 8 ~ 2 ea
M2075 series E010_24BO30 E08_24BO30 ~ 2 ea

 

 

 

 

 

 

*as determined by proton NMR

Surfactants with short E0 blocks (< 20 units) form clear liquids at room
temperature, however increasing the E0 block length leads to increasing self-association
and, consequently, higher viscosity and increased opacity.7 Due to E0 crystallinity,
when the average E0 block length exceeds ~ 25, the copolymer is a wax-like solid at
room temperature. After initially screening numerous B0/EO compositions, this work
focused on liquid copolymers with average E0 block lengths less than 20. Butylene
oxide oligomer is a liquid at block lengths up to at least 30, so increases in B0 block

length had less effect on the copolymer appearance. The purified products were clear,

48

colorless liquids with moderate viscosity (~ 95 cP). These surfactants had molecular

weights ranging from ~ 500 to 2200 amu and were soluble in a range of organic solvents.

2.5 Copolym er Characterization

Each copolymer batch is a mixture of B0/E0 diblock copolymers having various
B0 and E0 block lengths. Because copolymer composition has a dramatic effect on
micellar structure formation,8 one must accurately determine the composition of each
copolymer batch to assess the impact of changes in block length on micellar structure
development. Because of E0-capping inefficiencies, the average B0/E0 compositions
calculated from reactant concentrations are insufficient to describe batch composition or
predict batch performance. A more extensive examination of the copolymer is required.

Copolymer characterization was performed using three analytical methods: NMR
spectroscopy, gel-permeation chromatography (GPC) and electro-spray ionization mass
spectrometry (ESI-MS). NMR spectroscopy provided a bulk composition,9 but was used
primarily to detect selected contaminants. GPC was useful in determining the
approximate mass range and polydispersity of the materials,lo but ESI-MS provided the
most accurate evaluation of the copolymer composition.3 ESI-MS was also useful in

determining the level of contaminants and byproducts in the batch.

2.6 Analysis by Nuclear Magnetic Resonance Spectroscopy

NMR spectroscopy has been used to study the composition of ethoxylated block
copolymers for many years, and the techniques used here have been partially described
before.3 Our NMR studies were conducted using FT-NMR on a Varian Mercury Vx 300

instrument (Varian, Palo Alto, CA). The copolymer was dissolved at about 10 wt % in

49

deuterated chloroform (CDC13) with 0.1 vol % TMS as a reference. A relaxation rate test
was performed, and complete copolymer relaxation was determined to require about 8
seconds, so samples were analyzed with a first delay time 2 10 seconds. The pulse width
was set at 6 or 7 ms with an acquisition time of 3.7 seconds.

The methyl protons on the initiator (3.) produced a unique peak at 1.24 ppm, while
the methyl protons of the B0 units (on) appeared at 0.92 ppm (Figure 2.2). The ratio of
the areas of the BO and initiator methyl peaks (or/2.) gave the average number of B0 units
per initiator and hence, per chain. A multiplet from the methylene peak (13) of the ethyl
substituent of the BO unit is found from 1.4 —1.7 ppm. To confirm the average BO block
length value, the BIA. area ratio was also calculated and multiplied by 3/2 to adjust for the
difference in the number of protons on the two groups. This value matched the 0th value
within i 5 %. NMR analysis of E0 blocks in BO/EO copolymers showed only
methylene signals that overlap similar peaks from the backbone methylene of the B0
oligomer. Therefore, calculation of E0 block length required determination of the E0
contribution to the total methylene peak area by subtraction of the methylene peak areas
generated by the initiator and B0 blocks from the total methylene peak area. The
contribution of each group to the total number of methylene protons per chain, MPT, is: 2
from each B0 block, 2 from the initiator and 4 from each E0 block. If n is the number of
BO blocks and m the number of E0 blocks, then the total number of methylene protons

per chain is:

50

38a? based mm Ema mm; “a flux: defied

8582:: cc as 302 .381 68 3:2 38:88 3.83% omen do 588m «22 .3 28E

 

 

 

o; m; 9N Wm

m6

 

 

> > d a w > > .«
mica:camovafioA«moamoioamovofommu

 

 

<02MDX as :5 o4

 

 

 

51

MPT=2n+2+4m eq.2.1
The number of B0 blocks per chain, n, is equal to all. The total number of methylene
protons is represented by the peak area yT. To convert the NMR peak areas to number of

protons, we use the conversion factor of 30., so yT(3/7t) = MPT. Substitution of these

values into equation 2.1 produces equation 2.2.
yT (30) = 2(0r/7t) + 2 + 4m eq. 2.2
Or, rearranging equation 2.2 to isolate the average number of E0 blocks per chain yields

equation 2.3:

M) _ l : m eq. 23
411 2

Compositions calculated from NMR data were considered to be within i 10 % of the
actual composition and were used primarily to confirm that the product was similar to
that expected from the formulation. The molecular weights derived from these

calculations represent the number-averaged molecular weight (Mn) of the compositions.

Synthesis of the B0 block prior to the E0 block allows for selection of both a
strong hydrophobe to initiate B0 polymerization and a hydrophilic hydroxyl group to
terminate the E0 block. However, because of the low E0 capping efficiency inherent in
this method, the expected and actual block lengths of the final product will probably
differ. Unfortunately, these differences are not detectable by NMR analysis. BO/EO
average composition is calculated by determining the molar ratios of total E0 units and
total BO units to total ethoxide (initiator) units. Since the total number of ethoxide units
is not changed by the inefficiency of the capping process, the average E0 length

determined by NMR will be the same regardless of the number of copolymers that

52

actually contain E0 blocks. Thus, several batches of copolymer could have the same
calculated average composition, but vary significantly in actual block composition
depending on the capping efficiency.

Oxidation of the monomer and copolymer can produce peroxides, aldehydes,
ketones, and other byproducts that are readily detectable using NMR. The two most
probable contaminants are toluene (residual solvent) and crotyl alcohol from the B0
rearrangement (Section 2.4). Toluene, which produces a distinct NMR peak at 2.32 ppm
and a pair of triplets between 7.1 and 7.3 ppm, is easily removed by evaporation. Crotyl
alcohol produces 3 NMR peaks: a singlet at 2.45, a triplet at 4.05 and a broad multiplet at
5.68 ppm. However neither crotyl alcohol, nor toluene, nor any of the other possible
byproducts, were detected in the copolymers used in this study.

For 13C NMR spectra, the integration was considered to be indicative, rather than
definitive, of actual copolymer composition, so this technique was used primarily to
identify the presence of contaminants. The vinyl group of the crotyl alcohol and the

carbonyl of an aldehyde (CH3CH2CH2C(=O)H) could be easily identified. Carbon-13

NMR showed that aldehyde formation was ~ 4 times greater in materials synthesized in
the mini-plant than in copolymers prepared using the Parr Reactor. However in neither
case were contaminants present at high enough concentrations (< 1 mol % of total B0) to

interfere with micellar structure formation.

2.7 Gel Permeation Chromatography

The copolymers in this work have a distribution of E0 and BO block lengths and

therefore a distribution of masses. Gel permeation chromatography (GPC) is frequently

53

used to separate the components of a polymer mixture according to molecular weight.
The GPC employed in this work consisted of a Waters 590 pump (Waters Corp., Milford,
MA), a Waters 712 Autosampler, a Waters CHM column heater, a Waters TCM
temperature control module and a Waters 410 refractive index detector. The eluent,
tetrahydrofuran (THF), was degassed and blanketed with helium throughout the
operation. GPC samples were prepared at 1.0 wt % copolymer in THF with 0.1 wt %
dodecane added as a flow marker. The system was fitted with two Waters columns, an
HRO.5 and HR2 (cat # WAT044231 and WAT044237), that were maintained at 35 °C.
The system was controlled using the Polymer Labs Cirrus software that provides both
data acquisition and processing capabilities, and calibration was performed using
polyethylene glycol (PEG) standards obtained from Polymer Labs (Polymer Laboratories
Inc., Amherst. MA).

This technique was used primarily to examine the polydispersity of the oligomers,
both prior to E0 addition and in the final product. However, GPC alone is not applicable
for determining the actual average copolymer composition because only ethylene oxide
standards are available for calibrating the instrument; no butylene oxide standards are
available. Surprisingly, however, if the value of the B0 MW taken from the NMR
analysis is substituted into the GPC MW calculations, the value of the E0 block
determined by the GPC data closely matches the theoretical value. Nonetheless, while an
absolute molecular weight could not be determined by the GPC data alone, an overall
polydispersity could be measured for the B0 oligomer and the final copolymer.

Figure 2.3 contains chromatographs of copolymers L1150-51A and L1150-51B,

whose synthesis was described in section 2.2. Since these copolymers are generated from

54

the same BO oligomer, they should vary only in E0 block length. Molecular weight

calculations based on the peak molecular weights (Mp) of E0 standards indicate that
batch L1150-51A had an Mp of 1303 amu, and batch L1150-51B had an Mp of 1501

amu. However, molecular weight assignments are dubious because the E0 standards are
being applied to B0. (As noted above, BO standards are not available.) The most

important point to note from the chromatograms in Figure 2.3 is that the peak shapes are
relatively narrow and unimodal. However, the small peaks between retention times of 19

and 23 minutes probably represent residual, uncapped B0 oligomer. The GPC data

 

I I I I I I I I I

 

—L1150-51A 7.
----L1150-51B -

q Dodecane

 

 

 

Differential Refractometry Response

 

 

 

 

 

14 16 18 20 22 24 26 28 30
Retention Time (min)

Figure 2.3. GPC chromatographs of the two copolymers whose synthesis was described

in section 2.2.

55

showed that all of the oligomers and copolymers used in this work had essentially
unimodal MW distributions. The B0 oligomers typically had a polydispersity index
(PDI) between 1.05 and 1.15, while the copolymer PDI was between 1.09 and 1.17. I
should note, however, that at low molecular weights, low polydispersities can still occur

with a wide range of oligomeric compositions.

2.8 Copolymer Analysis by Mass Spectrometry

Mass spectrometry provides the most detailed information about the copolymer
composition. Mass spectrometric analysis was performed by Bruce Bell of the Analytical
Laboratories at The Dow Chemical Company using a liquid chromatography/ electro-
spray ionization time of flight mass spectrometer. To prepare testing solutions, the
copolymers were dissolved in methanol at a concentration of 0.5 mg/ mL. The samples
were injected into a Waters Alliance 2690 ternary gradient liquid chromatograph coupled
to a Micromass QTOF2 quadrupole! time of flight MS/MS system via a Micromass Z-
spray electrospray ionization (ESI) interface operating in the positive ion (PI) mode. No
column was used in the liquid chromatograph. The mobile phase was 0.01 M ammonium
acetate in 50/50 water/ methanol flowing at a rate of 1.0 ml] min with a 2:1 split away
from the electrospray interface. The system used a diode array with a range of 200 to 400
nm as the UV detector. In the E81, the source block was held at 120 °C and desolvation
occurred at 250 °C. The capillary was held at 2.5 kV and the cone was set at —15 V. The
mass spectrometer was operated at a multi-channel plate (MCP) setting of — 2550V in
the +TOFMS mode as a continuum. The resolution, which is the ratio of the mass over

the delta mass, where delta mass is the peak width at half height, was set at 10000 (+).

56

(This is the same as saying that the instrument sensitivity was 1 part in 10000.) The scan
range was set from 50 to 3100 mlz (+) with a scan rate of 1.0 sec/ scan. The raw data
were collected and transformed using MaxEnt3 software to remove multiply charged
ions. The transformed data were imported into an Excel spreadsheet, and calculations

using the entire data set were made to determine MW, MN, Mp, Mw/MN, and PDI. These

values, and the raw data, were then forwarded to me for additional processing.

Using masses of 46.0419 amu for the ethoxide initiator, 44.0262 amu for the
ethylene oxide unit and 72.0575 amu for the butylene oxide unit, the molecular weight of
each possible combination of E0 and B0 block lengths, from 303 oligomer to

E030B025, was calculated and arranged in a spreadsheet. (The initiator mass includes

the proton of the terminal hydroxyl group.) The values are unique (differences between
neighboring values are > 4 amu) in the compositional range defined by an array bounded

by BO3EOO, BO3E017, B013E00 and B013E017. However, once the B0 length

exceeds 13, or the E0 length exceeds 17, duplication of masses begins. The BO
oligomer was analyzed first to determine the minimum and maximum values of the B0
block length, thus limiting the range of possible BO/EO compositions and making mass-
to-composition matching easier and more accurate.

To determine the copolymer composition, the transformed mass spectrometry data
were arranged by decreasing intensity to ensure that those components most prominent in
the copolymer were included in the calculations. The corresponding mass (actual mass

detected minus 18 for the NH4+ ion) was then matched to a composition. The difference

between the theoretical mass and that recorded by the instrument generally fell into one

of two groups: less than 0.09 amu, which was viewed as instrumental noise, and between

57

0.95 and l. 15 amu, which was attributed to slight variations in isotropic distribution.
More than 98 % of the recorded values were in these categories. If the mass could be
assigned to more than one composition, it was awarded to the composition with the
theoretical mass most closely matching the measured mass.

The intensity value from the actual mass measurement was then entered into a
second spreadsheet under the corresponding composition. Due to instrument sensitivity,
minute variations in mass (0.0001 amu) were recorded as separate moieties. Therefore,
multiple entries could be recorded for the same composition. When this occurred, each
intensity value was recorded for the composition and the intensities were summed to
determine the final total intensity per composition. This process was repeated until more
than 70 % of the total intensity produced by the analysis had been matched to an BO/EO
composition. The composition spreadsheet was then converted into a 3-dimensional
surface plot showing E0 length versus BO length versus signal intensity (See Figure 2.4
below). This generated a visual image of copolymer polydispersity, dominant
compositions and the range of compositions.

The intensity “cut-off point” of ~ 70 % of the total intensity was chosen for three
reasons. First, the values were arranged by decreasing intensity so values at the cut-off
were typically less than 5 % of the largest intensity values. Therefore, each subsequent
entry was a small fraction of the total by this point and had little impact on the population
distribution. Second, there was generally little change in the calculated average
composition when using 25 % of the intensity data versus 50 % versus 70 %, so further
sorting would most likely not prove to be illuminating with respect to compositional

distribution. Third, about this level, unassignable peaks began to appear, so to make

58

comparisons between batches more credible, I stopped including data below this level.
Overall, the 70 % cut-off usually involved the first ‘/3 to V2 of all of the peaks, when
arranged by decreasing intensity.

The average copolymer composition was calculated using a mean-intensity
procedure. The compositional spreadsheet was arranged with increasing rows for
increasing E0 block length and increasing columns for increasing BO length. The
intensity values for a given row in the two-dimensional array were summed, and then
multiplied by the E0 block length of that row to produce the total number of E0 units
represented by that row. This was repeated for each row, and the products were summed
to give the total number of E0 units represented in the spreadsheet. This sum was
divided by the total intensity recorded on the spreadsheet to produce an intensity-
averaged EO block length. This same procedure was repeated for the BO blocks and the
two average values were combined to give a composite, intensity-averaged copolymer
structure. This value should correspond to a number average molecular weight if
sensitivities for ions of differing masses are the same. A more detailed description of the

procedure and additional data are provided in Appendix B.

2.9 Mass Spectrometry Results

In using this procedure, I make four assumptions. The first is that the copolymer
polydispersity is low, and therefore, the range of compositions within a given batch is
manageable. This is generally true for polyethers prepared by anionic polymerization
and is confirmed by the GPC data. The second assumption is that the amount of

molecular fragmentation during mass spectrometry analysis is low, and thus the vast

59

majority of mass spectrometry peaks represent complete copolymers and not fragments
or rearrangement products. This is likely true given the gentle conditions of electro-spray
ionization. I also assumed that the instrument sensitivity is independent of copolymer
molecular weight, probably valid given the molecular weight range I examined. The
final assumption is that the monomers undergo little rearrangement or produce few
byproducts during synthesis so that the product is relatively pure. This is generally true
for the lab-scale synthesis, with the only common, possible rearrangement product being
crotyl alcohol. Replacing the ethoxide initiator with crotyl alcohol (CA) increases the

oligomer molecular weight by 26.0156 amu.6 For example, an EtO(B0)3H oligomer
weighs 622.502 amu while a CA(BO)8H weighs 648.518 amu. Such a mass difference

(+ 26.02 or — 46.04 amu) would be unmatchable to an ethoxide-initiated BO/EO
composition and is rarely found in the mass spectrometry data prior to the “70 %” cut-off.
(The vinyl group of the crotyl alcohol is also detectable by NMR when present at
sufficient levels to affect copolymer performance (sec 2.5), and none was found.)
Oxidation of the monomer and polymer can produce peroxides, aldehydes and ketones,
many of which would be detectable by mass spectrometry as unmatchable masses. In
fact, several batches produced by the mini-plant showed higher levels of materials that
could not be matched to an BO/EO composition than did the lab scale batches of similar
compositions. Nonetheless, even these levels are low (< 2 mol %) and considered to be
insignificant factors in structure formation.

Figure 2.4 is a 3-dimensional surface plot of the composition of copolymer
L1150-51A (section 2.2, first E0 addition). The aliquot of B0 oligomer removed prior

to E0 addition was lost, so BO block analysis could not be performed for these batches.

60

The plot shows that a substantial fraction of the B0 present (~ 4.0 mole %) is uncapped
oligomer. The molecular weight distribution of uncapped butylene oxide oligomer is

fairly symmetric around BO] 1, and no B03, B04 or B05 oligomers were detected. The
maximum peak for copolymer L1150-51A occurs at a composition of B09E08, however

the intensity-averaged composition is BOIOEOIO, close to the composition expected from

Copolymer Batch L1150-51A

Intensity

H

H H 0
EO Block

 

Length

Figure 2.4. Compositional profile of copolymer L1 150-51A (angled view) determined

from ESI Mass spectrometry.

the formulation, B010E01 1. This is also similar to the ratio-averaged structure
determined by NMR, B01 1E0] 1.

Figure 2.5 shows an overhead view of the same distribution. Here the dozen or so
highest peaks are represented by the dark, center region, which is approximately bounded
by the compositions B08EO7, B03E012, B011E07, and 801 1E012. This area

represents about 27 % of the total intensity plotted, and the copolymer distribution is

Copolymer Batch L1150-51A

 

8
7
6
5
4
3
2
1
O

90 N V $0 co 0 N V to (D O N
H H H H N N

24
26

E E0 Block Length

Figure 2.5. Compositional profile of copolymer L1 150-51A (overhead view) as

determined from ESI Mass spectrometry.

62

fairly symmetrical about this region. The average composition of this region, the “peak-
averaged composition”, has a formula of ~ BOIOEOIO, while the highest peak is found at
B09E08. The “kinks” in the grid represent smaller peaks.

Figure 2.6 shows the compositional distribution of the copolymer batch prepared
by the second addition of E0 monomer (section 2.2). Note that the average peak

intensity is about 50 % of that found in Figure 2.4. The overall range of E0 block

Copolymer Batch L1150-51B

Intensity

_,,,, j .
n a: (lithium :
lilllllvirrlrlrtsrll ..

l|l

 

E0 Block Length 2 30 32 0

Figure 2.6. Compositional profile of copolymer L1150-51B (angled view) as

detemtined by ESI-mass spectrometry.

63

lengths has increased by about 50 % as would be expected from the additional E0. The
level of B0 oligomer has dropped to ~ 2.6 mole %. The intensity-averaged composition
for this batch is BOIOEOIS. This formula is slightly lower in E0 block length than the
BOIOE019 composition determined by NMR.

Figure 2.7 shows the overhead View of the same distribution. The largest two dozen
peaks are represented by the dark, center region. This region is approximately the center
of a rectangle whose comers are the compositions B06E012, B06E021, B013E012, and

B013E021. The average structure defined by this region is B09_5E016.5, and the highest

peak is found at BOIOEOM.

Copolymer Batch L1150-51B

 

 

 

 

15
14
13
12
11
10
9 BO
8 Block
2 Length
5
4
3
2
1

4 6 10 12 14 16 18 0 2 4 6 28 O 32 0

E0 Block Length

 

Figure 2.7. Compositional profile of copolymer L1 150-51B (overhead view) determined

by mass spectrometry.

Based on these results, we find several designations for the composition of

copolymers as shown in Table 2.3. The number-averaged (Mn), weight-averaged (MW),
and peak molecular weights (Ml), as well as PDI, are based on the entire data set for a

given sample as opposed to the intensity-averaged composition, which is calculated using

only data with the highest intensity (70% of the total). Mn, MW, and M2 were converted

to compositions using the B0 block length found in the intensity-averaged composition.
Table 2.3 compares ESI-MS values for copolymers synthesized in both lab—scale and

mini-plant methods.
Table 2.3. Copolymer compositions derived from mass spectrometry data.

Intensity- Most
Batch averaged Abundant Mn based MW based Mz based
Name Composition Composition Formula Formula Formula PDI

 

L1150—51A 1301013010 BOgEOg 1301015011 130101308 1301015014 1.12

 

L1150-51B 1301013015 1301015017 1301015015 1301015011 1301013018 1.14

 

L120-15A 30101506 130101306 130101304 130101506 130101309 1.13

 

L150-66 BO **** **** 3010 13011 3012 1.10

 

M1038-BO **** **** 1301] 13012 13013 1.06

 

M1038A 130121309,5 130313017 1301213014 1301113012 301313015 1.11

 

M1038D 130le014 Bowl-:0ls 130125015 1301113013 130,3Eol7 1.13

 

 

M1096-BO **** **** 13010 13011 13012 1.06

 

 

 

 

 

 

 

The data in Table 2.3 can be compared to the molecular weights and probable
compositions obtained from NMR and GPC. Table 2.4 is a list of the copolymers
produced and their compositions calculated using different methods. As mentioned

previously, the GPC measured mass for BO oligomers is generally higher than the actual

65

 

mass because the calculation is based E0 standards. (B0 is more soluble in the mobile
phase and therefore is a fully elongated, “larger” molecule than E0. Thus it moves
through the column more rapidly than its actual molecular weight would suggest that it
should.) NMR results are quite consistent with mass spectrometry data for the B0
oligomer. The results from the various methods are especially comparable when the B0
block length obtained from NMR is substituted for the BO value determined by GPC in
the E0 block length calculation from the GPC data. GPC shows that the copolymer
polydispersities are generally low, though the polydispersity index (PDI) is included
primarily for comparison, because at low molecular weights such as these, it provides
little useful information. Other copolymer batches that were synthesized are not included

in this discussion for the sake of brevity.

Table 2.4. Comparison of compositions based on analytical method.

 

 

 

 

 

 

 

 

 

Batch NMR GPC GPC MS MS

Name Composition Composition PDI Compositionb PDI
L1150-51A B011E012 13011130“8 1.21 BOIOEOIO 1.12
L1150-51B B011E019 1301113015a 1.25 BOIOEOIS 1.14
L120-15A B06EO4 130613052‘ 1.34 BOIOEO6 1.13
L1150-66 BO B010 B014 1.13 B010 1.10
M 1038-B0 B011 B016 1.04 B011 1.06
M1038A BOllEOlz B016E010 1.04 B012E010 1.14
M1038D BOleOlg B016E012 1.04 BOIZEOM 1.13
M 1096-BO B010 B012 1.07 B010 1.06
M 1096B B09E01 1 B012E015 1.08 ----- ---

 

 

 

 

 

 

 

 

a No BO oligomer was available for measurement so NMR values for BO were substituted
b Intensity-averaged molecular weight.

66

2.10 Conclusions
Analysis of the copolymers is important for discerning the effects that changes in
block length have on structure formation. Analysis by NMR quickly indicates the

presence and level of contaminants in the material and provides an Mn value, but it

cannot distinguish provide information on copolymer homogeneity. GPC provides a

polydispersity and an MD value that can be converted into a composition, but due to

differences in solubility between the E0 and B0 blocks, GPC overestimates the length of
BO blocks. Mass spectrometry requires the most data manipulation, but also provides the
greatest detail about copolymer composition, polydispersity and the extent of E0
capping. While data provided by mass spectrometry are probably the most accurate in
describing average copolymer composition, the real advantage of this method is the
compositional profile, which provides the best illustration of the distribution of
copolymer compositions. However, given the relative speed and accuracy of NMR, it

was selected as the initial method for copolymer analysis.

2.11 References

(1) Inoue, S.; Aida, T. Living polymer systems, epoxides and lactones. In Encycl.
Polym. Sci. Eng, 1990; Vol. Supplemental, pp 412-420.

(2) Shinoda, K.; Carlsson, A.; Lindman, B. Advances in Colloid and Interface
Science 1996, 64, 253-271.

(3) Kalinoski, H. T. Chemical analysis of polyoxyalkylene block copolymers. In
Nonionic Surfactants: Polyoxyalkylene Block Copolymers; Nace, V. M., Ed.;
Marcel Dekker: New York, NY, 1996; Vol. 60, pp 31-66.

(4) Harris, J. K.; Rose, G. D.; Bruening, M. L. Langmuir 2002, I 8, 5337-5342.

67

(5)

(6)

(7)

(8)

(9)

(10)

Thier, R.; Bolt, H. M. Critical Reviews in Toxicology 2000, 30, 595-608.

Whitmarsh, R. H. Synthesis and chemical modification of polyoxyalkylene block
copolymers. In Nonionic Surfactants: Polyoxyalkylene Block Copolymers; Nace,
V. M., Ed.; Marcel Dekker: New York, NY, 1996; Vol. 60, pp 1-30.

Nace, V. M. Properties of polyoxyalkylene block copolymers. In Nonionic
Surfactants: Polyoxyalkylene Block Copolymers; Nace, V. M., Ed.; Marcel
Dekker: New York, NY, 1996; Vol. 60, pp 145-183.

Dan, N.; Safran, S. A. Europhysics Letters 1993, 21, 975-980.

Ribeiro, A. A.; Dennis, E. A. Structure and Dynamics by NMR and Other
Methods. In Nonionic Surfactants: Physical Chemistry; Schick, M. J ., Ed.; Marcel
Dekker: New York, NY, 1996; Vol. 23, pp 971-1009.

Provder, T., Ed. Chromatography of Polymers: Characterization by SEC and
FFF.; American Chemical Society, 1993; Vol. 521.

68

Chapter 3

Effect of Butylene Oxide and Ethylene Oxide
Block Length on Structure Formation

3.1 Introduction

BO/EO diblock copolymers offer special opportunities in the study of vesicle
formation because the surfactant head group properties change with increasing E0 block
length as well as solution conditions. The E0 block forms hydrogen bonds with water;
therefore larger E0 blocks increase amphiphile hydrophilicity and hence, the critical
micelle concentration (CMC), which is the surfactant concentration required to initiate
structure formation. However, increasing either E0 block length or surfactant
concentration can lead to pronounced interaction of aggregates and loss of discrete
structures.l Longer E0 blocks also change the head group diameter, altering the critical
packing factor and potentially the type of micellar structures that form.2

Solvation of E0 is enthalpically driven, so the hydrophile has decreasing
solubility with increasing temperature.3 Like many ethoxylated surfactants, BO/EO
copolymers can form liquid crystalline structures that are temperature sensitive and have
an obvious phase-transition point.4'5 Above this temperature, loss of E0 hydrophilicity
eventually leads to structure degradation. As with most surfactants, high concentrations
of selected salts or organic solvents also alter aggregate formation.6 Thus, variation of
block length, temperature, solvent, or ionic strength should allow for manipulation of
ethoxylated structures.

As discussed in section 1.6, Israelachvili et al. described micellar structure

formation in terms of a critical packing factor (CPF), (I), which is defined in equation 3.1.

69

In this equation, V is the hydrophobe volume in the micelle core, a0 is the cross-sectional
area of the head group, and IC is the hydrophobic chain length.2

(I) = V/(lc a0) eq. 3.1
Israelachvili’s model hypothesizes that changes in V, 16, or a0 affect micelle formation by

altering the amphiphile shape and hence, the geometry of structures it can form. These
changes include modifications of the surfactant composition or changes in the local
environment that lead to alternate amphiphile geometries. Here I focus on surfactant
composition changes and their effects on micelle formation particularly the impact of E0
block length.

Once formed, the micelles themselves can organize into structures called meso-
phases. The top row of Figure 3.1 shows the most common micellar structures as they
exist under dilute conditions. Rows 2 and 3 indicate the types of micellar aggregates that
can form as surfactant concentration increases. Several of these meso-phases, such as
hexagonal, tetragonal and lamellar, are optically anisotropic and produce a birefringence
pattern when viewed under plane-polarized light. Examination of such meso-phases
reveals the type of micelle involved and provides insight into the effects of E0 and B0
block length on micelle formation.

Thus two factors dominate meso-phase formation; the critical packing factor
because it determines micellar shape and the amphiphile concentration. However with
ethoxylated surfactants these two factors are not necessarily independent. Amphiphile

concentration changes may lead to changes in head group area, a0, which can alter the

micelle shape and therefore change the meso-phase.6

70

 

Spherical (L1) Cylindrical

  

 

Micellar cubic (11)

6'6“" “3“”)
«5,9111%»;
{5'9’33‘93‘33
..))§£ )))»)"? ))) Water in oil
Rectangular Hexagonal (H1) bicontinuous (V 1)

   
 
  

 
 

    

Figure 3.1. Micellar aggregate structures and the resulting meso—phases. (Meso—phase

designation in parentheses.)

This chapter will discuss the process for selecting BOIEO compositions that

demonstrate the potential for vesicle formation and evaluation of the structures that these

materials form. I begin by examining the meso-phases formed by numerous

compositions as a function of concentration in water, searching for the birefringent

patterns indicative of liquid-crystalline, meso-phase formation. Ithen identify vesicle-

71

forming copolymers and discuss the physical characteristics of vesicles produced by
these amphiphiles. Finally, I describe the copolymer properties, e.g., average E0 block
length and polydispersity, that are most critical to vesicle formation and how those
properties can be manipulated to alter the anrphiphile packing geometry and influence

structure formation.

3.2 Screening Tests for Liquid Crystalline Phases

It has been previously shown that planar lamellae can be converted to vesicles
using shearing techniques.7'lo Therefore my initial work focused on identifying BO/EO
copolymers with block lengths appropriate for larrrellae formation. If either the E0 or
B0 block is too short (less than 4 units) the copolymer will not demonstrate amphiphilic

properties. While BO block length has little impact on the critical packing factor (V/lc is

almost independent of B0 length), it does impact the free energy requirement to disperse
the molecules in water, a critical step in surfactant self-assembly. Thus the upper limit of
B0 block length is set by the dispersibility of the surfactant. Once a B0 block length is

selected, V and lc are essentially constant.

Conversely, if the E0 block is too long, the copolymer will show increased
solubility in water and structures will form only at high concentrations. Therefore a
copolymer which forms vesicles must have not only a B0 block length suitable for
forming micelles in water, it must have a corresponding E0 length sufficient to produce

the proper a0. Although previous studies showed that the head group area at the air!

water interface changes with E0 block length,11 no accurate means of predicting head

group size in solution is currently available. Because a priori prediction of structure

72

geometry is not currently possible, I used a screening method for determining the
structure-formation capabilities of BOIEO copolymers with various block lengths.

Clint describes a screening test that quickly indicates the types of meso-phases a
given surfactant will form with increasing levels of hydration.‘2 The surfactant is placed
on a glass microscope slide adjacent a dr0p of water. A slide cover is then placed over
both drops, forcing them into contact. The sample is allowed to equilibrate for ~ 20
minutes before examination during which time a concentration gradient arises from the
surfactant] water interface into the surfactant rich area. Structures develop along this
gradient as a function of hydration level. Meso-phase formation is monitored using
plane-polarized light microscopy (PLM). Meso-phase type is indicative of micelle type

and reveals how a0 is changing with hydration. The primary goal of this test was to

identify compositions that produced lamellae, although all meso-phase structure
formation was recorded. Planar lamellar structures can be precursors to vesicle
formation, so identification of such structures suggests that a particular composition may
form vesicles under selected conditions.

Crystalline structures are often optically anisotropic and are common among
inorganic materials occurring whenever molecules become organized in a periodic
fashion.l3 Liquid crystalline structures also display this property,12 but the molecules are
not held in as rigid an array as in crystalline solids, and thus the structures are frequently
temperature and shear sensitive”14 I examined BO/EO dispersions using plane-
polarized light to determine if liquid crystal formation had occurred.

Six crystalline structures were identifiable by birefringence patterns observed

using PLM (Olympus Vanox Transmission microscope, Olympus Optics, Tokyo):

73

tetragonal (C1), hexagonal (H1), reverse tetragonal (C2), reverse hexagonal (H2), planar
lamellar (La) and spherical lamellar phases. (1 follow the phase designations proposed

by Gray and Windsor”) Domain sizes of 500 nm or greater were necessary to produce a
recognizable pattern, but for the majority of samples, this was not an issue.

Figure 3.2 shows an example of the tetragonal phase. It has been proposed that
these patterns arise from the two-dimensional square tetragonal array of micellar rods, or
worms, as shown schematically in Figure 3.1.15 A “bundle” of rods associate to form a
thick filament which weaves throughout the structure, but can remain fairly linear.
Figure 3.2a shows the thicker, larger, more visible “fibers” produced by this meso-phase
whereas Figure 3.2b shows a finer version of the same structure. The resulting optical
pattern appears fuzzy and indistinct because of the weak impact each filament has on
light transmission. The extent of optical rotation, and hence regional brightness, is
largely a function of the tetragonal meso-phase thickness. This meso-phase is a highly

viscous gel due to the extensive entanglement of these structures.

74

 

Figure 3.2. Tetragonal meso-phase produced by B01 1E012 in water (70 wt %).

75

The hexagonal meso-phase also develops from the packing of micellar rods, in this
case forming a two-dimensional hexagonal array Giigure 3.1). The hexagonal packing of
the rods leads to less entanglement and produces a more defined shear-face when
shearing occurs. Thus, in contrast to the tetragonal phase, this meso-phase has prominent
facets, well-defined repeat units and sharp edges (Figure 3.3). The only difference
between the normal and reverse phases of the tetragonal and hexagonal phases is the
volume fractions of the “oil” and water; structural organization is essentially the same.
However, the reverse hexagonal phase is of interest because vesicles are frequently
formed by the controlled dilution of this phase in water often under slight shearing. '6

Three types of lamellar structures exist: flat planar as discs and sheets, tubular as
myelin, and spherical as vesicles. (Numerous micrographs of these structures can be
found in the literature and the optical patterns are sufficiently distinct as to allow facile

12"4'”‘22) Planar lamellae are described as having a “Streaky-oil” or

identification.
“railroad track” pattern. Myelin appear as fingers of copolymer which extent outward
into the water-rich phase and display this same pattern along the structure edges.
Figure 3.4 shows an example of planar lamellae sheets. Notice the broken
“railroad track” pattern prominent in Figure 3.4a and running the length of the lamellae
band. Moving from the left side of the image, between two strong lamellar bands is a
region of hexagonal structures. Multiple phase states commonly occur with these
materials because the high viscosity makes formation of a homogenous sample a slow
process. Note also how planar lamellar structures can exist as both straight strands or

produce tight bends and twists (Figure 3.4b). These shape differences are explained by

the CPF model (equation 3.1).

76

 

Figure 3.3. Hexagonal meso-phases produced by 1301113019. (Copolymer! water/

dodecane at 71/ 9.5/ 19.5 wt %.)

77

 

Figure 3.4. Planar lamellar formed by B01 IEOIZ in water (30 wt %).

78

A typical screening test result is shown in Figure 3.5. The pure water region is at
the top of the figure with the copolymer concentration increasing from top to bottom.
Three meso-phases can be identified when moving from low to high copolymer
concentration. The first is the lamellar phase which consists of both myelin extending
into the water region and of planar lamellar sheets. Next is the bicontiuous phase (V1)
which is not anisotropic and appears as a black band. The final phase, reverse hexagonal

(H2), is apparent by the sharp edges and numerous facets of the structure. The dark

region following the hexagonal phase is the copolymer.

‘
D

. g)“ ..
s £63,
n Flt?“ Cm 'kfi'fl) I'-

‘rfr

 

Figure 3.5. Example of screening test results of B01 1E0” copolymer.

79

It is important to note that the first structures encountered at the copolymer/ water
interface are myelin, which are curved lamellar structures. This indicates that under
dilute conditions the copolymer will probably form lamellar structures, possibly vesicles.
It should also be noted that the myelin have a fairly large diameters (greater than 25 pm)
which suggests a moderate value of a that would almost certainly increase when fully
hydrated. An increase in a leads to greater curvature in the lamellar structure resulting in
either vesicle or bicontinuous meso-phase formation.

I examined more than 30 BO/EO diblock copolymers using this screening
technique. Copolymers with B0 block lengths greater than 20 units did not readily
produce liquid crystalline structures and, in fact, did not absorb the water droplet at all.
Copolymers having BO lengths of ~ 10 units and E0 block lengths of greater than 30
units were solubilized by the water until meso-phase formation began. (Liquid crystal
formation occurred if the copolymer concentration was typically greater than about 40 wt
%.) Compositions with BO blocks of approximately 10 units and E0 blocks less than 25
units were most likely to form lamellae, and the tendency improved with decreasing E0

block lengths.

3.3 Spherical Lamellar Structures

As with the other lamellar structures, vesicles are optically anisotropic and
produce a visually unique birefringence pattern called the Maltese Extinction Cross. The
Maltese cross results from the uni-axial nature of the structures14 and is visible in
structures with diameters greater than ~ 500 nm. Light passing through the polarizing

filter interacts weakly with the liquid crystalline structures oriented parallel to the light

80

path and is therefore easily blocked by the analyzing filter. This produces a horizontal
black band (Figure 3.6). Light passing through liquid crystals which are not oriented
parallel to the filter is rotated slightly by the structures and passes through the analyzing
filter. Light perpendicular to the analyzer is blocked, however, and the result is a pattern
of two dark lines crossed at right angles to one another. Figure 3.7 shows a typical
sample containing vesicles in both a PLM image and a white—light image. PLM reveals
the Maltese cross patterns indicative of spherical lamellae, while the white-light image

confirms the spherical nature of the structures.

 

Crystal orientation Optical pattern

Figure 3.6. Maltese Cross formation occurs in plane-polarized light because of the uni-

axial liquid crystalline structure of spherical lamellae. Adapted from Collings.14

81

 

Figure 3.7. Vesicles of B01 1E012 (20 wt %) viewed through white light (above) and

plane-polarized light (below). Note that vesicles are intermixed with planar lamellae.

(Micrograph by Deborah Rothe, The Dow Chemical Company.)

82

In contrast to the BO/EO copolymer, dispersions of B012 oligomers do not form
vesicles. However, with proper agitation, these materials will form droplets that are
similar in size to BO/EO vesicles. Figure 3.8 shows that the spherical BO droplets are
easily distinguished from vesicles due to the lack of any birefringence properties under
plane-polarized light. (This micrograph is taken using partially polarized light because
when the filters are aligned 90° to one another, completely polarized, nothing is visible.)

Note the grayness of particles and absence of the Maltese cross pattern.

 

Figure 3.8. Butylene oxide oligomer, B012, (2 wt %) dispersion viewed under partially

plane-polarized light microscopy.

83

Figure 3.9 shows an interesting phenomenon observed during screening tests of
several B0/EO compositions. Here lamellae tubes, myelin, reach from the interface into
the aqueous phase. Myelin lengthen as water at the interface diffuses up the tube to the
end.23 At the top-center of the micrograph, a vesicle can be seen forming at the end of
the myelin as the copolymer becomes water saturated. The vesicle expands until it
detaches from the myelin, and floats away to be replaced by another growing vesicle.

This cycle can be repeated several times and these “budding” myelin are quite common.

 

Figure 3.9. Vesicle of B01 1E012 “budding” from myelin tip formed at copolymer!

water interface.

84

Of the copolymers tested, those having B0 lengths between 8 and 12 and having E0
lengths between 10 and 14 most commonly displayed this pattern. Based on these

observations, we refer to BO/EO vesicle formation as “spontaneous”.

3.4 Vesicle Structures

Having identified several BO/EO copolymers that spontaneously produce vesicles
on contact with water, I began to study the vesicular structures themselves. The
following micrographs provide some indication of the size and types of vesicles that
could be formed using only the surfactant and water. The dispersions shown in Figures

3.10 and 3.1 l were prepared by placing a drop of copolymer in a vial, adding about 0.5

 

Figure 3.10. Plane-polarized light micrograph of B0] 113012 ( 1 wt %) without slide

cover.

85

mL of deionized water and agitating the sample by hand. Dispersions were examined
within 2 hours of preparation. Figure 3.10 shows the range of vesicle diameters formed
when the vials experience minimal agitation during vesicle formation. In this figure, the
slide is not covered therefore the sample thickness is much greater than in Figure 3.11.
The dispersion shown in Figure 3.11 was vigorously agitated for about 1 minute
immediately upon addition of water. Note the improved uniformity in structure size as
compared to Figure 3.10. Given the size range of all of these spontaneously-formed

structures, they are probably multi-lamellar vesicles (MLV, Figure 1.7).

 

Figure 3.11. Plane-polarized light micrograph of vesicles from B01 1E012 dispersion (1

wt %).

86

3.5 Cryo-SEM

Because the Maltese cross patterns may result from either uni-lamellar or multi-
lamellar vesicles, we examined dispersions using cryo-SEM (JEOL JSM-6320FV) to
differentiate between these two possibilities. The dispersions were flash frozen in liquid
propane and stored in liquid nitrogen. They were then transferred to a cryo preparation
chamber at -90 °C, fractured with a cooled scalpel, held at -90 °C for 2 min in vacuum to
sublime the surface ice, and sputter coated with conductive chromium. Samples were

examined at -125 °C using a 3 keV electron beam. Figure 3.12 shows that the larger

 

Figure 3.12. Cryo-SEM micrograph of B011E012 vesicle (1.0 wt %) taken 10 weeks

after formation. (Micrograph by David Williams, The Dow Chemical Company.)

87

structures formed from BOHEOIZ are multi—lamellar vesicles (MLV). The concentric
layers are apparent from spots where the structure was punctured during freeze fracture.
Figure 3.13 also shows that the structures are composed of numerous layers, although in

this micrograph the layers appear less uniform and spherical.

 

Figure 3.13. Cryo-SEM of vesicle from B01 1E0” (1 wt %) showing numerous layers

of MLV structure. (Micrograph by David Williams, The Dow Chemical Company.)

88

3.6 Effect of E0 Length on Vesicle Formation

The critical packing factor (equation 3.1) can be re-written as (I) = (V/lc)(1/a0).
For a given hydrophobe, e. g. BO 11. V/lc is constant and independent of the head group.

Vesicle formation then requires finding a hydrophile with the appropriate value of a0 to

yield 0.5 < (I) < 1. For an ionic surfactant, where the hydrophile is generally short and
composed of a few atoms, the head group is envisioned as a sphere whose diameter
determines the head group area. For nonionic hydrophiles, which may be 20 times longer
and consist of dozens of atoms, the head group “shape” is not as obvious, and
determination of head group area is more complicated.

The relationship between E0 block length and head group area for micellar
structures is not first order as it is at the air-water interface.24 Instead, as the E0 block is
hydrated, it forms an expanding coil whose diameter is dependent on E0 length.25 The
most pronounced changes in coil size, and hence head group area, occur with increases to

shorter E0 blocks. Alexandridis et al. showed that the head group area for E02 to E010

varies with the square root of the number of E0 units.ll Thus, using an E0 block as the

hydrophile, one can generate various values of a0 simply by changing the block length
during synthesis. Note that do is inversely proportional to (I), so increasing the head

group area is expected to produce smaller structures.

Evaluation of numerous copolymers shows that surfactants with a B0 block
length of ~ 10 and an E0 block length of between 6 and 12 will spontaneously produce
vesicles even under dilute conditions. As the E0 length increases the concentration for
vesicle formation also increases. Table 3.1 shows that increasing E0 length produces

slightly smaller structures, consistent with an increase in ac. Continued increases in E0

89

block length should eventually lead to a copolymer with ac too large to produce vesicles.
Consistent with this pattern, B010E024 did not form vesicles at any concentration, and
dynamic light scattering showed the mean particle diameter in a 5 wt % B010E024

dispersion to be 14 nm, a value appropriate for micelles formed from a copolymer of this

size.26 Others have found that B010EO40 produces only micelles.27

Table 3.1. Composition, molecular weight, density and concentration range for PLM-

detectable vesicle formation for BO/EO diblock copolymers.

Density Concentrations for
Copolymera Mnb (g/mL) MLV Formation Ave. Diameter

 

 

 

 

 

B011E06 1154 1.00 0.05 — 20 wt % 5 pm c
130101-3014 1322 1.01 0.05 - 20 wt % 5 umc
301013014 1384 1.02 0.5 — 20 wt % 3 pm 6
1301113019 1661 1.03 2.0 — 20 wt % 1 11111 d
B010E024 1 824 --- 110116 ""

 

 

 

 

 

 

 

a) H1 NMR, b) mass spectrometry, c) 1 wt % dispersions, d) 5 wt % dispersions

To confirm that head group size is the primary reason for the structure-forming
differences between copolymers, we attempted to induce micron-sized vesicle formation

from a dilute B01 1E019 dispersion through chemical changes that affect head group

diameter. One method for doing this is the addition of selected salts that compete with
the hydrophile for free water.”28 Thus, a saline solution should “dehydrate” the E0
group, decrease the head group size and make large MLV formation more favorable.2 A

1.3 wt % sample of B01 1E019 was prepared using 12.5 wt % NaCl with gentle agitation.

90

After 24 h the sample contained an abundance of multi-lamellar structures with diameters
of ~ 1.0 pm. No MLVs were observed at the same concentration without salt.
Alternately, the head group size may be reduced using thermal dehydration of the

E0 block. A 1.0 wt % dispersion of BO, 1E019 was heated to 85 °C for 3 h, agitated

while cooling and then examined using PLM. Figure 3.14 shows the changes in the
dispersion due to heating. The large MLVs (> 3 mm dia.) were visible for up to 24 h after
heating. (This provides some information as to the time required for complete
rehydration of the E0 block after thermal dehydration.) Based on these results, we

conclude that the differences in structure formation between E012B011 and E019B011

are based primarily on head group size.
We also explored the effect of E0 length on structure formation using a series of

copolymers prepared by a sequential addition process. An BOIOEOIO copolymer was

prepared using a standard procedure. After copolymer formation, however, instead of
quenching the entire reaction, only an aliquot was removed and quenched. A second
reaction was performed to add another ~2 E0 units to the remaining copolymer. An
aliquot of this material was also removed and quenched, followed by another addition of

E02 to the remaining copolymer. This procedure was repeated until a series of

copolymers having the same B0 block length and E0 block lengths ranging from 10 to
24 was prepared.

Dispersions of 1 wt % surfactant were prepared from these copolymers. For
copolymers with E0 < 18, phase separation occurred after 5 days, forming a sediment of
large structures and leaving a dilute, opaque, upper phase of structures less that 1 um in

diameter. After allowing the two phases to equilibrate for 3 weeks at room temperature,

91

an aliquot of the upper phase was removed and ultra-sonicated. Ultra-sonication should
yield smaller structures which should have a head group area, a, closer to the optimal

area, a0 and reduce structure polydispersity. The structures were then evaluated using

dynamic light scattering (Brookhaven 90Plus, Brookhaven Instruments, USA).

 

Figure 3.14. Plane-polarized light micrographs of an B011E019 aqueous dispersion (1

wt %) before and after heating. No vesicles are visible prior to heating (left), but

numerous structures are visible after heating at 85° C for 3 hours (right).

Figure 3.15 shows structure diameter as a function of the E0 block length. These
data clearly show that sonicated structures produced from copolymers with shorter E0
blocks are much larger than those produced from copolymers with E0 lengths greater
than 16 units, which is in general agreement with the geometric model (eq. 3.1).
Structures produced by EOIOBOIO to E014B010 are too large to be spherical micelles

and are probably uni-lamellar vesicles. Conversely, the structures produced by E020 to

E024 have diameters appropriate for spherical micelles formed from surfactants of this

92

size. Particles produced from E016 may represent transitional aggregate structures, such

as rods or worms, and E013 structures are likely swollen micelles.

 

100 I I I W l I I I i I I I I rm I I I I I I I I I I I I I I

Diameter (nm)
:1 on
c c
1E!-
19

1151

N .5
G G
III'Ijj—[IIIIfrIlIlI

E

1111111111111111111

In 13

I1111111111411111111111111111

10 12 14 16 18 20 22 24
E0 block length

 

 

 

6

Figure 3.15. Mean structure diameters as a function of E0 block length for B01 lon

copolymers. (Error bars represent standard deviations of 5 consecutive measurements.)

3.7 Effect of Concentration on Vesicle Formation

The head group diameter for an ethoxylate group is concentration dependent, so I
also studied the effect of copolymer concentration on vesicle formation. In a typical
procedure, one drop of a particular copolymer was added to a tared 5-dram vial and the
amount of water required for the desired concentration was added. The vial was
moderately agitated by hand (~ I shake sec") for 30 seconds and allowed to stand for a
minimum of 24 h before examination. Aqueous dispersions of each copolymer were
prepared in concentrations ranging from 0.005 to 20 wt %, but the majority of work

focused on 0.5, 1.0, 2.0, 5.0 and 10.0 wt % samples. This concentration range was

93

selected to ensure that the E0 head groups were fully solvated, and to minimize the
effects of micelle-micelle interactions. (Samples greater than 40 wt % copolymer were
known to produce gels.) Samples were visually inspected for clarity, relative viscosity,
and examined using plane-polarized light microscopy (PLM) to determine the number

and type of any meso-phases present.

Copolymers with the shorter E0 blocks (< E012) make PLM-detectable MLVs at
all concentrations studied (Table 3.1). PLM images also showed that B011E012 formed

the highest density of large, stable MLVs. At 0.5 wt %, B01 1E012 formed abundant

structures with diameters ranging from 3 to 30 pm as determined by light scattering. As
the copolymer concentration increased, so did the apparent vesicle concentration as
viewed by PLM. At copolymer concentrations of 1.0 and 2.0 wt %, a sediment of MLVs
formed after a few days, and no gel fragments or other liquid crystal structures could be
observed in the sediment. In contrast, at 5.0 wt % the copolymer formed a stable
dispersion with no indications of phase separation or sediment formation for at least 2
months. MLVs were visible in samples with concentrations greater than 5.0 wt % for at

least 2 months when stored at room temperature. From 10 to 40 wt % B01 1E012, PLM

revealed planar lamellae with MLVs (Figure 3.16), with the level of planar structures

increasing until samples of 50 wt % B011E012 were exclusively planar lamellar.
For B011E019, vesicle formation depended on copolymer concentration, with

identifiable vesicles only visible at concentrations of 2.0, 5.0 and 10 wt %. At 0.1 wt %,
the sample formed a single, translucent phase with no crystalline structures observable by
PLM. Sample opacity increased without PLM-detectable vesicle formation until the

copolymer concentration reached 2.0 wt %. At concentrations less than 2 wt %,

94

 

Figure 3.16. B011E012 dispersion (15 wt %) showing vesicles with planar lamellae.

B011E019 may form vesicles that are too small to be detected with PLM. (Smaller
vesicles would be consistent with an increased an, and light scattering data (Brookhaven
90Plus, Brookhaven Instruments, USA) for a 0.1 wt % aqueous dispersion showed
particles with diameters of ~ 60 nm.) Within a few days of preparation, the 2.0 wt %
dispersions formed a sediment consisting of MLVs having a mean diameter of 1 pm. No
other liquid crystal phases were visible in either the sediment or the continuous phase,
although the continuous phase was not clear. The composition of the sediment did not
change after standing for 1 month at room temperature. Sediment formation, if it
occurred, was less apparent in 5.0 and 10.0 wt % dispersions due to increased opacity.
Planar lamellar structures first appear at a concentration of about 35 wt % and aggregates

were exclusively planar lamellar at 70 wt %.

95

These results are consistent with the current understanding of ethoxylate head
group changes as a function of block length and copolymer concentration. Increasing the
block length increases a, and shifts micelle formation from vesicles to spherical micelles.
Increases in copolymer concentration have two effects; reduced hydration of the E0
block and increases block interactions both leading to a decrease in the head group size.

This is evidenced by the formation of vesicles as the copolymer concentration increases.

3.8 Effects of Polydispersity on Vesicle Formation

As discussed in chapter 2, the BO/EO copolymers synthesized for this work are
not monodisperse, and copolymer polydispersity may impact structure formation.29 In a
preliminary investigation of the effects of polydispersity, we examined BL50-1500, a
commercial surfactant (The Dow Chemical Company, USA) described as having an

average composition of B010E017. Based on the results described above, B010E017

should spontaneously produce MLVs at elevated concentrations. However for 5 wt %
dispersions of this copolymer, dynamic light scattering (Brookhaven) showed only
micelles with diameters of ~ 18 nm. The bulk copolymer was a waxy solid, suggesting
that a significant portion of the surfactant had an E0 length greater than 25, so I decided
to fractionate the material. The copolymer was suspended in hexane and then centrifuged
at room temperature for up to 16 hours to produce a two-phase sample.30 The clear,
upper phase was decanted while the solid, lower phase, which contained the longer,
insoluble E0 chains, was removed after melting. The upper phase was re-centrifuged at -

5 °C. The newly formed clear, upper phase was also decanted while the white solid was

96

melted and removed. Finally, the hexane was removed from the upper phase to produce
a clear liquid.

This clear liquid, which had an average composition of B011E08 as shown by H1
NMR, spontaneously produced 1 — 5 um-diameter MLVs just as our BOllEOll had,

even at concentrations well below 1 wt %. This fraction accounted for ~ 30 wt % of the
BL50-1500 copolymer, while the remaining material had an average EO length greater
than 25. This result confirms that determination of the mean composition alone is not
sufficient to indicate whether a given copolymer will form vesicles.

A screening test (section 3.2) of the B011E03 copolymer also produced an

unexpected structure. Upon hydration the copolymer produced myelin, as frequently
occurs with the other BO/EO surfactants we studied. However, many of these structures
were twisted into helical shapes that persisted for hours (Figure 3.17). This may be an
indication that a substantial fraction of these molecules have an average head group size
at this level of hydration that is close to that required for planar lamellae structures,
however further investigation is necessary. It should be noted that no vesicles were

observed “budding” from the helical structures.

97

 

Figure 3.17. Plane-polarized light micrograph of B01 lEOg (BL50—1500A) shows

helical myelin at water/ surfactant interface upon hydration.

3.9 Giant Vesicles

Based on my observations of MLV formation and the effects of E0 block length
on structure size, I proposed that if the proper copolymer were allowed to hydrate slowly
and without agitation, it might produce much larger structures than those produced by
agitation after a short hydration period. Thus I chose copolymer M1038, which mass
spectrometry indicated has a mean composition of BOIZE09. 5, because the shorter E0
block length, even when fully hydrated, will produce a small head group and yield a large
radius of curvature, therefore larger, more easily detected structures.

As a “proof-of-concept” test I deposited 4 drops (~ 0.11 g) of copolymer dropwise

into a glass bottle containing 10 g of DI water. The droplets flattened and fell to the

98

bottle bottom. I examined the structures using plane-polarized light and found that the
surface of the drops had immediately become birefringent. Within 20 minutes
birefringent filaments of copolymer were visible rising from the coalesced droplets. The
filaments were laced with white spheres too small for me to identify as vesicles using a
15X magnifying glass. After 45 minutes the filaments were much longer and in several
cases moving straight up from the copolymer. I still could not identify any Maltese cross
patterns in the copolymer or filaments. I let the sample stand undisturbed at room
temperature for 3 days.

After about 65 hours, I examined the sample to find several changes. The
copolymer was now a swollen white gel about 4 mm thick above which was an opaque
liquid layer also 4 mm thick. Approximately 10 mm above the copolymer surface the
solution was completely clear. When viewed through the crossed polarizing filters, the
copolymer surface appeared covered with numerous bright spots that appeared to be
Maltese crosses. Several vesicular structures were obvious close to the bottle wall
however I could not confirm structure formation within the copolymer.

As the filaments passed through the opaque layer, they became chains of vesicles
(up to 10) that were about 0.5 mm in diameter. The structures were faint suggesting that
they were composed of few lamellar layers. Concentric layers (3 to 5) were visible inside
the structures. Above the opaque layer the filaments did not produce vesicles.

I re-examined the sample 138 hours after the copolymer addition. The copolymer
surface now displayed considerably more Maltese crosses than previously. In addition to
the large structures (~ 1 mm in dia.) there were hundreds of structures with diameters of

0.1 mm. The large structures did not appear any more intense than they had previously.

99

The opaque layer above the c0polymer was now filled with smaller, shorter strands of
copolymer creating a “fuzzy” surface appearance. The opaque layer volume had doubled
and the remaining liquid above it was no longer clear. The extent of vesicle formation by
the filaments had increased; they now contained twice as many vesicles as they had at 65
hours. The vesicles were also larger, averaging almost 1 mm in diameter although they
were no brighter suggesting that the number of lamellar layers that each contained was
unchanged.

The expansion of the “fuzz” and opaque layers continued until I ceased
observations after about 520 hours. At this point the “fuzz” layer was 12 mm above the
copolymer surface and the opaque layer reached the top of the bottle. The number of
vesicles had increased by 5X, with diameters ranging from 0.01 to 1 mm. The filaments
were covered with vesicles (they appeared to be “budding” from the filament surface)
however the structures were no brighter than when originally observed.

To repeat this experiment under conditions more amenable to recording my
observations, I prepared a rectangular, glass cell as the sample holder and capped it with
cork (Figure 3.18). To minimize vibration, I glued the cell to a wooden dowel that I
inserted into a clamp holder and mounted it to a lattice. I also attached a plastic ruler to
the cell face as a reference scale. I then positioned 150 W spotlight, a glass bottle (5” x5”
x8”) of water to act as an IR filter and polarizing filter mounted on a quartz plate
sequentially behind the sample and a camera (Minolta X700) equipped with a second
polarizing filter in front of a macro zoom lens at a 4:1 setting. I used 400 ASA TMax
black and white film (Kodak Co., Rochester, NY) for increased clarity. This way, I could

easily record the any large structure formation using plane-polarized light.

100

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

— l—_l
K J ,1 ‘ '\ Camera
Bottle Polarizing Quartz Sample w/ polarizing
Lamp of water filter P1316 7 holder filter

 

 

 

 

 

 

Figure 3.18. Experimental set-up for photographing giant vesicle formation.

I added about 4 drops (~ 0.12 g) of copolymer to about 15 g of DI water
depositing the drops on the cell bottom. Within nrinutes the copolymer surface was
birefringent and in less than an hour spherical structures were apparent. They followed
the same developmental sequence as structures in the first experiment. However
overnight the temperature in the lab rose to more than 32 °C and I realized that
temperature increases each evening were probably common place. Because of the
sensitivity of the E0 block to temperature changes, I transferred the entire apparatus to a
constant temperature lab (22 °C), switched the light source to a 120 W lamp (less IR) and
let the experiment continue.

After a total of four days numerous large vesicles were visible on the bottom of
the cell. Figure 3.19 shows the abundance and size of these structures. Most have
diameters in excess of 0.1 mm with many exceeding 0.5 mm. This is a smaller average

diameter than for structures in the original experiment, but clearly shows the proliferation

101

of these structures. Note that few of the vesicles are suspended in the aqueous phase.

Most remain on the copolymer surface or along the bottom of the cell.

 

Figure 3.19. Giant vesicles from copolymer B012E095 in glass cell.

102

There are also several larger structures which have a much less distinct Maltese cross
pattern. These may contain fewer lamellae or they may not be vesicles at all.

To evaluate the number and fragility of the structures I gently tilted the cell about
45° and held it in that manner for ~ 15 seconds. The vesicles were quickly and
thoroughly dispersed through out the aqueous phase and remained so for several hours.

Figure 3.20 shows the dispersion 22 minutes after tilting.

 

Figure 3.20. Giant vesicles dispersed after slight agitation.

103

I extracted several vesicles and aqueous phase using a pipette, placed them into a
vial and set them aside. After 3 days I transferred several structures to a deep-well slide
and obtained micrographs using PLM. In Figure 3.21 the multi-lamellar composition of
these structures is quite apparent. In spite of rough handling, these structures appeared to
be relatively unaffected. They easily rolled along the bottom of the vial for weeks. They

could be damaged, but it required considerably more effort than expected.

0.1 mm

 

Figure 3.21. Plane-polarized light micrograph of giant vesicle isolated from copolymer

of second experiment.

I repeated the experiment a third time using the same apparatus, but having the
entire experiment take place in the constant temperature lab. The drops were placed on
the cell bottom and the structural development sequence was similar to the first and
second attempts. However, the filaments did not produce structures in the same manner
as the first two experiments. Only the filament tips showed the spherical structure
formation and while they were as large as the previous vesicles they were not as bright
and appeared to be composed of few layers. Instead of development via filaments, it
appeared that the copolymer itself slowly expanded to produce the structures, eventually
becoming saturated with vesicles after about 3 weeks, much longer than required by the
first two runs. A fourth attempt under exactly the same conditions as the third produced
essentially the same results; no vesicle growth from the filaments and the structures
which did develop were much fainter.

I decided to see if post-heating would have an affect on the structures. I
transferred the entire cell and holding apparatus into an oven to stand at 30 °C for 90
hours. Upon re-examination I could see no change in the nature or quantity of structures.
It appears then that even slight heating of the copolymer during the first 48 of hydration
has an effect on both the type and number of vesicles formed.

The most probable cause of these differences concerns changes in the head group
as a function of temperature which directly affect the amphiphile packing and nature of
the resulting structure. If the copolymer hydrates at a temperature which yields an
equilibrium head group diameter consistent with planar lamellar structures, then the
copolymer simply swells with time eventually reaching an equilibrium state. However if

the temperature drops so that the head group diameter increases the new packing factor

105

indicates a curved structure will result. The strain of the larger head groups could cause
large sections of lamellae to delaminate from the copolymer and form curved structures,
the giant vesicles observed.

Conversely, if the head group diameter is also large enough to produce a strain in
the copolymer, then as soon as a few layers become hydrated they may delaminate from
the bulk copolymer and form structures. These would have a similar diameter to those
formed in the two-temperature process because the final head group diameter is the same.
However they would not contain as many lamellae because the strain initiated curvature
before the copolymer microns below the surface was hydrated enough to curve as well.

Further study on this process is warranted.

3.10 Conclusions

The use of B0 in amphiphilic BO/EO diblock copolymers results in large
hydrophobe volumes. This leads to surfactants with compositions that spontaneously
produce vesicular structures in water at room temperature. However for spontaneous
vesicle formation, the B0 block must be less than 20 units long, otherwise the copolymer
is not dispersible in water. Differences in structure formation between the various
BO/EO compositions follow the general trend put forth by Israelachvili’s geometric
model. For a given B0 length, shorter E0 chains favor the formation of vesicles by
decreasing the area of the head group. The tendency for vesicle formation can be
manipulated with concentration, thermal cycling or ionic strength.

Cryo-SEM micrographs indicate that the structures formed are multi-lamellar

spheres, and this is not unexpected given the particle diameters of > 1 um. MLVs

106

spontaneously form in concentration ranges from 0.05 to 20 wt % for the copolymers

with E0 < 15 and in concentrations of 2.0 to 20 wt % for B011E019. If the block length

is short (< 10), slow copolymer hydration will produce giant MLVs having diameters in

the 0.5 mm range. However, vesicles do not form for E0 > 23. Also the mean E0 block

length is not sufficient alone to predict vesicle formation. The polydispersity of the E0

block length must be narrow enough that the actual head group size is similar to that

predicted by the mean.

3.11 References

(1)

(2)

(3)

(4)

(5)

(6)

(7)

Derici, L.; Ledger, S.; Mai, S.-M.; Booth, C.; Hamley, I. W.; Pedersen, J. S. Phys.
Chem. Chem. Phys. 1999, 1, 2773-2785.

Israelachvilli, J. N.; Mitchell, D. J .; Ninham, B. W. J. Chem. Soc., Faraday Trans.
2 1976, 72, 1525-1568.

Bailey, F. E., Jr.; Koleske, J. V. Configuration and hydrodynamic properties of
the polyoxyethylene chain in solution. In Nonionic Surfactants: Physical
Chemistry; Schick, M. J ., Ed.; Marcel Dekker: New York, NY, 1987; Vol. 23, pp
927-969.

Hamley, I. W.; Pople, J. A.; Booth, C.; Yang, Y. W.; King, S. M. Langmuir 1998,
14, 3182-3186.

Bedells, A. D.; Arafeh, R. M.; Yang, 2.; Attwood, D.; Padget, J. C.; Price, 0;
Booth, C. J. Chem. Soc., Faraday Trans. 1993, 89, 1243-1247.

Rosen, M. J. Surfactants and Interfacial Phenomena, Second ed.; Wiley-
Interscience: New York, 1989; p 194.

Mueller, S.; Boerschig, C.; Gronski, W.; Schmidt, C.; Roux, D. Langmuir 1999,
15, 7558-7564.

107

(8)

(9)

(10)

(11)

(12)

(13)

(14)

(15)

(16)

(17)

(18)

(19)

Diat, 0.; Roux, D. Journal de Physique II 1993, 3, 9-14.

Escalante, J. I.; Hoffmann, H. J. Phys. : Condens. Matter 2000, 12, A483-A489.

Panizza, P.; Colin, A.; Coulon, C.; Roux, D. Eur0pean Physical Journal B:
Condensed Matter Physics 1998, 4, 65-74.

Alexandridis, P.; Athanassiou, V.; Fukuda, S.; Hatton, T. A. Langmuir 1994, 10,
2604-2612.

Clint, J. H. Surfactant Aggregation; Chapman & Hall: New York, NY, 1992; pp
158-171.

Gray, 0. W.; Windsor, P. A. Liquid Crystals & Plastic Crystals; Physico-
Chemical Properties and Methods of Investigation; John Wiley and Sons: New
York, NY, 1974; Vol. 1; p 223.

Collings, P. J. Liquid Crystals: Nature 's delicate phase of matter; Princeton
University Press: Princton, NJ, 1990; p 90.

Gray, G. W.; Windsor, P. A. Liquid Crystals & Plastic Crystals; Physico-
Chemical Properties and Methods of Investigation; John Wiley and Sons: New
York, NY, 1974; Vol. 1; p 280.

Marsh, D. Handbook of Lipid Bilayers; CRC Press: Boca Raton, FL, 1990; p 387.

Wanka, G.; Hoffmann, H.; Ulbricht, W. Macromolecules 1994, 27, 4145-4149.

Yu, G.-E.; Li, H.; Fairclough, J. P. A.; Ryan, A. J.; McKeown, N .; Ali-Adib, Z.;
Price, C.; Booth, C. Langmuir 1998, 14, 5782-5789.

Kunieda, H.; Nakamura, K.; Davis, H. T.; Evans, D. F. Langmuir 1991, 7, 1915-
1919.

108

(20)

(21)

(22)

(23)

(24)

(25)

(26)

(27)

(28)

(29)

(30)

Gradzielski, M.; Mueller, M.; Bergmeier, M.; Hoffmann, H.; Hoinkis, B. Journal
of Physical Chemistry B 1999, 103, 1416-1424.

Dorfler, H. D.; Knape, M. Tenside, Surfactants, Deterg. 1993, 30, 196-203.

Franses, E. 1.; Hart, T. J. J. Colloid Interface Sci. 1983, 94, 1-13.

Buchanan, M.; Egelhaaf, S. U.; Cates, M. E. Langmuir 2000, 16, 3718-3726.

Schick, M. J. J. Colloid Sci. 1962, 17, 801-813.

Schick, M. J. J. Am. Oil Chemists' Soc. 1963, 40, 680-687.

Harris, J. K.; Rose, G. D.; Bruening, M. L. Langmuir 2002, 18, 5337-5342.

Yu, G.-E.; Yang, Y.-W.; Yang, 2.; Attwood, D.; Booth, C.; Nace, V. M.
Langmuir 1996, 12, 3404-3412.

Lange, H.; Jeschke, P. Surface monolayers. In Nonionic Surfactants: Physical
Chemistry; Schick, M. J ., Ed.; Marcel Dekker: New York, NY, 1987; Vol. 23, pp
1-44.

Dan, N.; Safran, S. A. Europhysics Letters 1993, 21, 975-980.

Yang, Y.-W.; Brine, G.; Yu, G.-E.; Heatley, F.; Attwood, D.; Booth, C.;
Malmsten, M. Polymer 1997, 38, 1659-1668.

109

Chapter 4
Vesicle Characterization and Stability

4.1 Introduction to Vesicle Stability

This chapter examines several properties of BO/EO vesicles that may affect their
potential utility in encapsulation applications. Specifically, I discuss the fraction of
BO/EO that forms vesicle structures, the stability of the structures under a variety of
conditions, and the ability of these vesicles to encapsulate hydrophilic and hydrophobic
materials. Vesicles are attractive encapsulation materials because their bilayer structure
consists of both hydrophobic and hydrophilic portions that can house selected materials.
Encapsulated hydrophilic materials may have little association with the amphiphile head
group, and therefore the volume of solution encapsulated by the vesicle generally limits
the loading of these materials. Limits on loading of hydrophobic guests depend on how
significantly the incorporation of these molecules alters the packing geometry of the
amphiphile. MLVs are especially attractive because their large hydrophobic volume
should permit encapsulation of relatively large numbers of hydrophobic molecules.

Vesicles stability is an issue in encapsulation applications because there is a
constant exchange of surfactant between structures and the bulk aqueous phase.
Although the surfactant exchange rate for vesicles is similar to that of spherical micelles,l
the larger size of even small uni-lamellar vesicles (ULVs) yields lifetimes thousands of
times longer than those of micelles. Additionally, the exchange rate between the inner
and outer layer of the curved bilayer (the flip rate) is generally low.2 The combination of

these effects can result in MLV lifetimes of weeks, months or even years.

110

Below, I first discuss the fraction of surfactant that is incorporated into vesicles.
Subsequently, I examine structure stability, both thermal and mechanical, by monitoring
the particle size distribution in vesicle dispersions after various post-formation
treatments. Finally, I investigate the structural effects of introducing model hydrophobes

or hydrophiles into BO/EO vesicles and monitor the release rate of a model compound.

4.2 Yield of Structures

The first question addressed was the fraction of copolymer contained in the
vesicles at equilibrium. I had previously prepared numerous aqueous copolymers
samples at various concentrations. Based on these samples I decided to initially focus on
low concentration dispersions that were much less likely to produce gels or meso-phases.

Samples of B011E012 and B011E019 had formed structures that settled out of the bulk

phase. Vesicle sedimentation is consistent with MLV formation because a multi-layered
structure should have a density slightly higher than water. Given copolymer densities of
1.00 to 1.03 g/mL (increasing with increasing BO block length, Table 3.1) and assuming
that the copolymer occupies 50 % of the sphere volume, an MLV would have a density
about 0.5 % greater than water. This density difference, in addition to the large particle
size, explains why MLV sedimentation occurs and makes centrifugation a plausible
means of isolating large structures.3

To estimate the fraction of copolymers included in MLVs, I centrifuged 1.0 wt %

samples of BO] 1E012 at 3000 rpm for 24 h. Examination of the resulting sediment by

plane-polarized light microscopy (PLM) revealed only vesicular structures. I removed

the supernatant and dried the precipitate to determine the mass of the MLVs and found

111

that > 65 wt % of the copolymer was in these structures. In contrast, the yield of MLVs
from a 2.0 wt % B011E019 dispersion, which preferentially forms spherical micelles, or
small vesicles, was less than 10 %, again showing the greater propensity for vesicle

formation by B011E012.

4.3 Thermal Stability of Liquid Crystalline Structures.

BO/EO structures are liquid crystalline and as such undergo a phase transition
with increasing temperature. I examined structures from a 5.0 wt % dispersion of
B011E012 as a function of temperature using plane-polarized light microscopy (PLM)
with a microscope hot stage. The temperature was controlled using a Linkam THS600
hot stage and a Linkam PR600 Controller (Linkam Scientific Instruments Ltd, UK). The
sample holder consisted of a stainless steel washer glued to a glass slide with epoxy resin.
After the dispersion was deposited into the washer center, a glass slide cover was glued to
the washer to seal the holder. The samples were heated at 1 °C per min with a 10 min
pause every 5 °C to allow for equilibration.

Room temperature liquid crystalline structures were identified by the presence of
the Maltese cross pattern. However, upon heating, the E0 group undergoes dehydration
and becomes increasingly hydrophobic.4’5 Thus, these crystalline structures persisted
only until the sample reached 38 °C. At that point the Maltese crosses disappeared, but
non-polarized light microscopy showed that the structures maintained their spherical
shape. No liquid crystalline structures were evident between 38 and 65 °C, the maximum
temperature examined, and no coalescence or change in structure shape was observed.

Upon cooling from 45 to 35 °C the Maltese cross patterns reappeared within 5 minutes.

112

The same procedure was followed for structures formed from a 5.0 wt % B011E019

dispersion with similar results (phase transition ~ 42 °C). (This transition temperature is

greater than that of many phospholipids?) Results with both B011E012 and B011E019

show that there is no obvious redistribution of the block copolymers that make up the
vesicles during this thermal cycling and that the formation of liquid crystalline structures
in the vesicles is thermally reversible. When dispersions were heated overnight to
temperatures above 75 °C, complete phase separation of the block copolymers from
water is observed. Reformation of liquid crystalline structures is achieved by gentle
agitation of the two-phase system at room temperature, producing a similar particle size

distribution (PSD) to that formed upon initial agitation.

4.4 Thermal Stability of Vesicle Dispersions.

Light scattering measurements are the most suitable technique for the study of
vesicle stabilities. Particle size distributions were obtained using a Coulter LS230
Particle Size Analyzer (Beckman Coulter, Miami, FL) that can detect particles with
diameters from 40 nm to 2000 pm. The instrument was fitted with the Small Volume
Module and all analyses were carried out using the ‘pslref’ (polystyrene latex reference)
model. Instrument performance was confirmed using 0.357 and 1.0 um polystyrene latex
standards. Screening experiments showed that the dilution process used by this
instrument did not alter the particle size distribution of the sample over the time involved.
Particle sizes obtained in light-scattering experiments were consistent with MLV sizes
viewed in PLM of undiluted samples. For particles smaller than 50 nm, a Brookhaven

90Plus (Brookhaven Instruments Corp., Holtsville, NY) particle size analyzer was used.

113

Dispersions of B0] 1E0” were thermally cycled to evaluate the effects of heating

on PSD. Samples were heated to 85 °C for periods of 15 to 60 min, vigorously agitated
by hand and returned to the oven. This procedure was repeated until the samples had
been heated for a total of 3 h. After thermal cycling, the samples were gently agitated
and set aside to stand undisturbed for 24 h at room temperature. Light scattering data
showed no significant difference between the PSD of a B011E012 dispersion before and

after thermal cycling. In contrast, thermal cycling of B01 1E019 dispersions that did not

spontaneously produce MLVs at room temperature showed MLV formation (as
determined by PLM) after heating. Dispersions of 0.5 and 1.0 wt % B01 1E019 that were
thermally cycled showed vesicle formation with a PLM-determined size distribution
similar to that for B01 1E012 samples. A 10.0 wt % B011E019 dispersion appeared
unchanged by thermal cycling, showing the same PLM size distribution and apparent
abundance of vesicles before and after heating. Thus, although BO/EO vesicle size
distributions are stable to thermal cycling, MLVs may undergo phase or shape changes at

high temperatures before cooling and returning to MLV structures.

4.5 Structural Stability

Because B01 1E012 readily formed MLVs over a wide range of concentrations,

we conducted extended structural studies using this copolymer. To test the mechanical

stability of these structures, I first prepared 25 g of a 5 wt % dispersion of B01 1E0” by

the standard hand-shaking method. Light scattering data from this dispersion (Figure 4.1)
Showed that structure diameters ranged from 3.0 to 30 um with a mean of 5 .1 pm.

(Similar distributions occurred with 0.5, 1.0 and 2.0 wt % dispersions.) Interestingly, we

114

observed no particles with diameters between 0.04 and 2.0 pm (Based on copolymer
length, spherical micelles would be expected to have diameters less than 0.02 pm and
would be undetectable in these measurements using the Coulter LS230.) The dispersion
was then separated into 5 g aliquots and each was exposed to post-formation treatments.
One aliquot was exposed to moderate speed magnetic stirring overnight, and a
second for 2 minutes to high speed vortex mixing (Therrnolyne Vibrational Mixer, model

37615). A third was mildly sonicated in a sonication bath (Branson Sonic Bath, model

 

 

12 _ '- -— ‘
t )3 ........- Untreated . . q

1‘1 - Magnetrc strmng -

10 L I . "'"Vortex mixing _
- X Sonication -

 

 

 

on
I

5
' l

N
l' I I

O

I I I
)-
W __
+- I
_

 

 

Number Fraction of Particles (%)

 

 

‘ c- __--_-A
A A___________,

 

 

15 20 25 30 35 4o
Particle Diameter (um)

UI
H 1-
6

Figure 4.1. Particle size distribution of MLVs in a 5.0 wt % aqueous dispersion of

B011E012 before and after various post-formation treatments.

115

2200) for 2 min. The particle size distribution (PSD) was relatively unaffected by these
treatments. These techniques are routinely used to generate or alter the size of
phospholipid MLVs, so we would not expect them to destroy the vesicle structures.°
However, we were surprised that these treatments had so little effect on particle size.

The MLV structures are stable in the presence of sonication, moderate and high
shear mixing, as well as other post-formation treatments. We think that the stability of
these MLVs results from contributions by both the B0 and E0 blocks. Hydration of an
E0 block which is in a constrained, close-packed orientation leads to a liquid crystal
structure.7 This crystalline “shell” can be robust and presents an additional energy barrier
to vesicle disintegration. Collisions between vesicles are unlikely to penetrate this
“shell” or have any impact on structural stability.

The BO block contributes to the vesicle stability because of its length and
branching. The free energy of micellization for a CH2 group is about the same as for a
BO unit.8 Therefore, for an BO/EO copolymer to posses the same hydrophobic character
as a phospholipid, the B0 length would be twice that of the di-lipid tails. The additional

BO length (and ethyl branches) allow for greater hydrophobe entanglement, again

slowing the process of disintegration.

4.6 Extrusion of Vesicles
Extrusion is frequently used to reduce MLV sizes or convert MLVs to uni-
lamellar vesicles.4 To evaluate the effect of extrusion, a 5.0 wt % dispersion of

801 1E012 was forced through a 5 pm Nylon® filter repeatedly (7 times) using a syringe

pump. Figure 4.2 shows the change in structure size due to extrusion and indicates that

116

the mean particle diameter decreased from 5 pm to 70 nm. Greater than 95 wt % of the
copolymer was recoverable after extrusion. The same initial dispersion (5.0 wt %

301113012) was also extruded through a 0.45 pm nylon filter with essentially the same

results as with the 5.0 pm filter. This PSD after extrusion was stable for at least 2 months
with no reversion to larger structures. As tortuosity may have some effect on extrusion, a

5 .0 wt % dispersion of B011E012 was also extruded through a 0.4 pm track-etched

polycarbonate membrane. This resulted in the same particle sizes as found with extrusion
through the nylon membranes.

In contrast to the shearing results in section 4.4, extrusion reduced particle size from
an initial mean diameter of 5 pm to 70 nm (Figure 4.2). The differences between
stability to extrusion and stability to sonication or mixing can be explained by examining
the different methods of shear application. In the extrusion experiment, the MLV size or
shape must change to pass through the membrane pores, physically altering the liquid
crystal structure. Upon entering and exiting the pores, the MLVs experience elongational
forces and during the passage through the pores they experience a steady, laminar flow
field. This may allow for “delarnination” of the lamellae from the vesicles without
breaking the liquid crystal structure. These layers quickly close to form spherical
lamellae. Steady laminar flow was previously used to transform planar, lamellar liquid
crystal structures into MLVs.9'll These transformations depended on a critical
rheological property such as a critical shear rate, shear stress or strain. Consistent with
those results, the shear field presented by extrusion through 0.45 um pores is evidently
more effective at separating the lamellae than sonication or high shear mixing.

Furthermore, the particle sizes after extrusion through pores of this diameter suggests that

117

70 nm is the minimum stable size for vesicles formed from this copolymer. This may

represent the lower diameter limit of uni-lamellar vesicles.

 

 

 

 

 

 

 

 

 

 

 

10 _
A 8 _
Q
E
5 1
:3 6 r:
l
a
In *1 t
In 4 1 \
1
s , 1‘
5 ’ )1
z 2 - u
\\
*1,
.5}!
0 \t
LLI I l I llLill l 1 1411111 4 L4 1
01 1 10
Particle Diameter (m)

Figure 4.2. Particle size distribution before and after extruding a 5.0 wt % dispersion of

B011E012 through a 5.0 pm or 0.45 pm filter.

4.7 Ultra-sonication of Vesicles

Although a sonication bath had no discernible impact on vesicle size distribution,
I decided to evaluate the effects of ultra-sonication because it is a more focused and more
powerful application of ultra-sonic energy. I prepared a standard 1 wt % dispersion of

B011E012 using the hand-shaking rrrethod. The structures had a mean particle diameter

118

of about 3 um. I transferred about 5 mL of the dispersion into a vial that I placed into an
ice-water bath. Finally, I inserted the 3 mm diameter horn of the ultra-sonicator (Cole-
Parmer Ultrasonic Processor, model CP 130) into the dispersion and used a typical
sonication procedure of a “4 sec on! 1 see off” cycle at 10 watts, for 2 rrrinutes. This
procedure reduced the mean particle diameter to about 70 nm, similar to the size
produced by extrusion.

I then decided to test the concentration limits of this process. I prepared 20.0 wt

% BOl 1E012 in water and then agitated the mixture by hand for about 30 seconds and

placed it on the vibrational mixer for 30 seconds. The result was a thick, white fluid
saturated with bubbles. I placed the dispersion onto the mechanical shaker table for 2
minutes, but saw no change in the sample.

I ultra-sonicated the same dispersion using the standard “4 sec on/ 1 see off” cycle
at 10 watts. After 13 minutes the dispersion appeared considerably clearer and its
viscosity was much lower. However, there were several gel fragments visible, so I
returned it to the sonicator. After 35 minutes, the dispersion was much clearer with only
a few gels attached to the bottle wall. I removed the dispersion from the ice-water bath
and placed it onto the vibrational mixer for 2 minutes. I then refilled the water bath with
ice and returned the dispersion to the ultra-sonicator. After a total of 70 minutes of ultra-
sonication, the dispersion was very clear, only slightly opaque, and only slightly more
viscous than water.

I re-examined the dispersion after it stood at room temperature for 3 months. The
sample was still a slightly opaque liquid, but there was evidence of particulate settling. A

thin (0.1 m) white coating was visible on the bottom of the vial. By rotating the vial it

119

was apparent that this was a low viscosity fluid. The volume of material seemed far too
low to have been a significant portion of the copolymer. Otherwise the dispersion
seemed quite typical; water-like viscosity and uniform opacity. Iran PSD tests on the
dispersion using the Brookhaven 90Plus and found an average particle diameter of 99

run. I returned the sample to the hood, and after 7 months, it was unchanged. It remained
a clear liquid with a viscosity slightly above that of water and no indications of
precipitation. Based on these observations, I would speculate that the upper limit of
concentration for this process is greater than 25 wt % (250 mM) copolymer. This is
comparable with the “high loading level” of phospholipids which is typically about 30 wt

% or 200 mM.‘5

4.8 Examination of The Effect of Added H ydrophobe on Structure Formation

The screening test (section 3.7) showed that selected copolymers produced
lamellar structures, however it was important to determine the effect of a hydrophobe
(dodecane) on structure formation for two reasons. First, the mass spectrometry data
showed that uncapped B0 oligomer was present in many copolymers, and understanding
its impact on structure formation is critical for establishing a maximum acceptable level
of B0. Second, one possible application of BO/EO vesicles is the encapsulation and
delivery of hydrophobic materials. Understanding the impact of increasing hydrophobe
concentrations on structure formation as a function of BO/EO composition will determine
which copolymers might prove useful for these types of encapsulation applications.

After identifying compositions that produced lamellar structures, I began

developing phase diagrams for selected compositions. For such studies, the ratios of

120

copolymer, water and dodecane, the model hydrophobe selected for these studies, were
varied at a constant temperature. Samples were prepared by placing the copolymer into a
vial, adding dodecane and agitating the clear mixture on a vibrational mixer (Fisher
Scientific Instruments, Chicago, IL) until the liquid was homogeneous, typically ~ 15
seconds. Water was then added to give a total mass of ~ 2 g, and the vial was returned to
the mixer for about 2 minutes. The sealed dispersions were heated at 80° C, with
intermittent mixing, for up to 20 hours, or until the dispersions were uniform They were
then agitated vigorously after cooling (~ 2 minutes) and were stored undisturbed at room
temperature for a minimum of 40 days before examination. Samples were visually
inspected for clarity, relative viscosity, and number and type of any mesophases present.

Sample examination was conducted using three optical probes: random white
light, HeNe laser light, and plane-polarized light. The surfactant! dodecane mixtures
were transparent with neither white- nor laser-light scattering nor any visible features
under plane-polarized light. Addition of water, however, generally produced a dispersion
with several notable optical characteristics. White-light scattering was indicative of large
domain sizes (diameters >100 nm) and phase separation such as occurs in colloidal
dispersions, emulsions, and suspensions. (These materials differ primarily in size and
nature of the continuous and dispersed phases.) However, random white light is
ineffective in identifying small domains (diameter < 50 nm), particularly under dilute
conditions, so laser light was used to search for the Tyndall Effect and smaller domains.l2
For example, spherical micelles are too small to produce noticeable scattering from white
light, and micellar solutions of greater than 30 wt % surfactant can be optically

transparent. The same material illuminated with a HeNe laser will typically display a

121

faint shaft of red light indicating the presence of some type of structures. (The intensity
of the light shaft increases with particle size and concentration.) Conversely, vesicles are
generally large enough to brightly scatter laser light, even when present in such low
concentrations that they do not produce a detectable opacity. Therefore, examination of a
dilute, clear sample using both laser light and white light can discriminate between
homogeneous solutions and those having micellar structures. It should be noted,
however, that there is no obvious method of distinguishing between a small vesicle
(diameter < 40 nm) and a swollen micelle using these techniques. For identification of
those types of structures, dynamic light scattering must be employed.

Samples were also examined for the previously described liquid crystalline states
(sec. 3.2). Vials were initially placed between two sheets of polarizing material (90°
difference between polarizations) and visually examined. Those showing liquid
crystalline regions were set aside to be examined using plane-polarized light microscopy
(PLM). Three other states, spherical micelle, rod-shaped micelle and the bicontinuous
phase, were implied by a combination of properties such as duration of birefringence

(none, temporary or permanent), sample viscosity and phase separation.

4.9 BO/EO Copolymer Phase Diagrams

Once equilibrated, dispersions were characterized using the methods described
above to prepare phase diagrams. Only single-phase samples were evaluated, since the
composition of the different phases in a multi-phase sample was unknown. Thus, not

every possible mixture of components was examined. Detailed sample analysis by PLM

122

occurred whenever the dispersion showed a viscosity increase (compared to the neat
copolymer viscosity) or a birefringence pattern.

Figure 4.3 is the phase diagram for B06E04. At high dodecane concentrations (>

40 wt %) the mixture formed stable emulsions (a liquid in liquid dispersion), and the
majority of the samples were multi-phase. Only three liquid crystalline phases were

detected: H1, La, and C2. The La phase contained both planar lamellae and vesicle

structures, but I found no region containing only vesicles.

10
90
20
80
30
70
40
Water 50 6O Dodecane
(wt %) 50 (wt %)
60
40
70
30

80
20

90
© 1.

/ La/Vesicle \

/ f / 7 / / / / /
10 20 30 40 50 60 70 80 90

Copolymer (wt %)

Figure 4.3. Phase behavior of B06EO4 with dodecane and water at room temperature.

123

A comparison of the phase behavior of BO6EO4 (Figure 4.3) and B011E012

(Figure 4.4) shows some interesting differences. While both copolymers formed
emulsions at elevated levels of dodecane, emulsion formation was significantly reduced

for B011E012. The copolymer concentrations that produced H1 phases were similar in
both diagrams, but required much higher levels of dodecane for B01 1E0 12. B011E012
also showed a larger compositional range for the La phase and a region that produced

only vesicles. Additionally, the B011E012 showed an H2, C2, and V2 regions.

10
90
20
80
30
70
40
Water 60 Dodecane
50
(wt %) 50 (wt %)
60
4O
70
30
80

20
90

10
Vesicles CSICICS/La C2 m
/ / 7
0 80 90

/ / / / / /
10 20 30 40 50 60 7

Copolymer (wt %)

Figure 4.4. Phase behavior of B01 1E012 with dodecane and water at room temperature.

124

Comparison of the phase diagrams for B011E019 (Figure 4.5) and B01 1E012

shows the effects of increased E0 block length on vesicle formation. The longer E0

chains make formation of reverse micelles (Ll) easier and more likely. The expanded
region for reverse tetragonal formation (C2) is also likely due to the increased interaction
of the larger E0 blocks. The LOL phase region is reduced because the larger head group

size favors formation of spherical micelles rather than bilayers.

Compared to B011E019 and B011E012, we would expect 30125010 (MS

composition used for clarity) also to contain a large region of vesicles or lamellae
because of its shorter E0 block. However, Figure 4.6 shows that this is not the case.

Again, the mass spectrometry profile helps explain why. 301213010 contains about 2.5
times as much BO oligomer as 80, 1E019 or B011E012 and, as with BO6EO4, the

concentration of this hydrophobe has a strong influence on structure formation. (Table
4.1 shows the relative molar levels of B0 oligomer found, as determined by mass
spectrometry, in each of the copolymers listed.) The presence of B0 oligomer has the
same effect on copolymer structure formation as the presence of dodecane. Both
hydrophobes associate with the BO block of the copolymer and changes the average
hydrophobe volume. Thus the V in the critical packing factor increases, which in turn
makes the formation of planar lamellae more probable. However if the hydrophobe
concentration becomes too large, the copolymer forms a stable dispersion like that shown

in Figure 3.8.

125

Table 4.1. B0 oligomer concentrations in BOIEO copolymers.

 

 

 

 

 

 

 

 

 

 

 

Batch NMR MS BO Oligomer

Name Composition Composition Cone. (mol %)
L1150-51A B011E012 BOIOEOIO 4.0
L1150-51B B011E019 BOIOEOIS 2.6
L120-15A 306E04 B010E06 8.0
M1038A B011E012 BOleOIO 10.3
M1038D 30115018 BOIZEOIO 10.7

10
90
20
80
30
70
40
Water 60 Dodecane
WI % 50 Wt %
50
60
40
70
30

8O
20
9O

/ / / / / / / / /
10 20 30 40 50 60 70 80 90

Copolymer
wt %

 

Figure 4.5. Phase behavior of B0] 1E019 with dodecane and water at room temperature.

126

Additionally, although the intensity—averaged composition is B012E010, the

compositional profile shows that much of the copolymer has an E0 block length of
greater than 16, the length at which spherical micelle formation begins to predominate.
Mass spectrometry data also indicate that the copolymer has a broad compositional

distribution. This may account for the larger number of structures observed in this

 

 

diagram.
10
90
20
80
30
70
40
Water 60 Dodecane
wt % 50 wt %
50
60
40
70
30
80
20

9O

Vesicles L“ 10

/ / / / / / /

10 20 30 40 50 60 70 80 90

Copolymer
wt %

Figure 4.6. Phase behavior of BOIZEOIO (mass spectral composition) with dodecane

and water at room temperature.

127

In addition to showing which compositions of copolymer, dodecane and water
will produce vesicles, the phase diagrams indicate a maximum loading level of
hydrophobe for encapsulation. If we assume that the hydrophobe is restricted to the
region of the vesicle occupied by the BO block, that it is not associated with the head
group nor in the water phase, we can see that dodecane can be added to about a 1:1 mole
concentration with the copolymer and still produce vesicles. As expected however, the
B0 oligomer has a detrimental impact on hydrophobe loading by competing for those
sites. For those copolymers with high BO oligomer levels, less dodecane was required

for vesicle formation to be halted.

4.10 Hydrophile Loading - Experimental Procedure

Hydrophiles are generally water-soluble and therefore may be found in the
aqueous continuous phase, in the aqueous region between bilayers or in the hydrophilic
region of the bilayer, adjacent to the head groups. Any study of hydrophilic loading
capacity must first remove the hydrophile from the aqueous continuous phase. This is
generally accomplished using gel permeation separation. Then a means of accurately
determining the level of the hydrophile encapsulated in the vesicles must be identified.

I chose the sodium salt of fluorescein, which has been used in numerous

9'3"“ as a model hydrophile. Though it is a large hydrophile (376

encapsulation studies,
dal), it has good water solubility (> 2 g dye! g water) and is easily detected at
concentrations less than 0.05 ppm using UV or fluorescence spectroscopy (absorbance

max at ~ 493.5 nm).'5

128

To prepare the gel column I weighed out 1 g of the dry Sephadex G-50
polydextran resin (Sigma-Aldrich, Milwaukee, WI) and added 40 g of deionized (DI)
water. The mixture stood overnight so that the resin could swell with water. I selected a
glass column that was 22 mm in diameter, 200 mm long and fitted with a sintered—glass
frit at the bottom. I chose a large diameter column to accommodate between 8 and 10
mL of dispersion per pass while still providing a clean separation of vesicles from dye.
After hydration, I carefully poured the resin into the glass column, making sure that no
bubbles were entrained in the gel. Using a rubber tube, I connected the stopcock to a
syringe and gently drew water through the column. This was done to compact the resin
and minimize the column porosity. However, it is critical that the resin NEVER be
allowed to dry, therefore the headspace above the resin was always filled with DI water.
(To reduce damage to the vesicles, the aqueous continuous phase in the gel column
should have contained a solution with the same ionic strength as the dispersion, however
the salt concentration was low enough that I ignored this effect.) I then placed a plug of a
hydrophobic, open-cell foam (polypropylene, I believe) onto the top of the gel to act as a
baffle and minimize the disturbance of the resin while adding the dispersion. After
copolymer addition, a second glass column was attached above the first and filled with
deionized water to replace the permeate flow.

I prepared several test dispersions of B012E01 1 vesicles prepared using a 300

ppm fluorescein dye solution. The dispersions were ultra-sonicated to produce structures
with diameters less than 100 nm, which were probably uni-lamellar vesicles. I passed the
dispersions through the gel column to evaluate the column’s separation efficacy and

found that it could easily separate about 4 mL of a 15 wt % copolymer dispersion in a

129

single pass. Once I confirmed the utility of the gel column, I turned my efforts a method
of dye detection.

Use of ion exchange resin to isolate and pre-concentrate ionic analytes is well
known,“"'17 but the process has been generally used to collect ions from solution not
extract them from vesicles. To determine the efficacy of such a procedure, I passed a test
dispersion of vesicles loaded with 300 ppm fluorescein dye through the gel column and
collected the permeate. I then exposed half of the permeate dispersion to Dowex MSA
ion exchange resin (The Dow Chemical Company, Midland, M1) to extract the dye from
the structures. After allowing the resin! vesicle mixture to agitate for several hours, I
removed the resin, which was now green, by filtration and saved the white dispersion. I
added enough methanol to both dispersions to produce a clear solution. (An aqueous
solution of greater than 40 wt % MeOH is sufficient to dissolve the copolymer and
destroy the vesicle structures.) The sample that had contacted the ion exchange resin
produced a clear solution without a hint of color, indicating that there was no residual dye
in the vesicles. The second sample, which had not seen the ion exchange resin, produced
a green-yellow solution. To evaluate the sensitivity of the coloration to dye
concentration, I added a single drop of 1500 ppm fluorescein dye solution to about 300
mL of a 1:1 MeOszater solution. The resulting liquid had a faint, but distinct green hue
at a concentration of about 0.2 ppm. The dispersion which had contacted the resin had no
color and must have had a dye concentration lower than 0.2 ppm. Thus, contacting the
vesicles with the ion exchange resin alone is sufficient to extract the dye from the

Sil’UCIUI‘CS.

130

The reason behind the fluorescein extraction may be the high charge density of
the resin, given as 1.51 eq/L of resin. The dye dissolves to produce 3 ions (two Na+ and
divalent fluorescein) per molecule and, at a concentration of 300 ppm, yields a mean
charge concentration of 2.4x10‘3 eq/L or about 0.2 % that of the resin. The resin’s high
surface-charge density may produce an osmotic pressure gradient that bursts the vesicles,
causing them to release the dye. (The bursting of liposomes under osmotic pressure
gradients is a common problem for controlled delivery”) The structures quickly reform
once the pressure has equilibrated, typically with a much lower concentration of
encapsulated materials.

Fluorescence spectroscopy can be used to determine the dye concentration on the
resin, however, I did not have easy access to an instrument, so I decided to circumvent
this problem by using UVN is spectroscopy for dye analysis. To do so, I needed to

extract the dye from the ion exchange resin which I did using a 30.0 wt % NaNO3

solution. The resin has a high affinity for the N03" ion, which over time displaces the

dye. Once the dye was in solution, the ion exchange resin could be removed by filtration.
Several dye extractions were necessary to completely remove the dye from the resin.
(The resin visibly changes from green to golden as the dye is released.) When complete,
the extraction solutions were combined, and UV absorbance was used to determine the
dye concentration. From the concentration of fluorescein in the extraction solution, I
then calculated the dye level in the structures.

I evaluated the potential of this method by mixing the resin with solutions having
a known dye concentration and used the data generated to produce a calibration curve. I

prepared 5 solutions of fluorescein dye with concentrations of 500, 300, 150, 50 and 10

131

ppm, and mixed each solution with 1.5 g of ion exchange resin and let the mixtures
agitate overnight. The following day each solution was clear with no hint of color. I
isolated the resin by filtration and extracted the dye using 5 mL of a 30.0 wt % NaNO;
solution. To be consistent, each sample was extracted 5 times with N aN03 solutions
regardless of whether or not the extraction solution was colorless prior to the fifth rinse.
(Most were colorless after just three extractions.) These extractions were collected and
evaluated. Because of the large concentration range, I monitored 5 different wavelength;
450, 463, 473, 484 and 485 nm and produced a calibration curve for each wavelength.
Based on the color of the resin, I believe I recovered more than 95 % of the dye from

each sample.

4.11 Hydrophilic Loading and Release

My first attempted at studying the hydrophile release rate of the B012E010
copolymer involved a high hydrophile concentration and a high copolymer concentration.
I prepared a 15 wt % dispersion of copolymer using a 1500 ppm fluorescein dye solution.
The initial product was a thick gel, however after 45 minutes of ultra-sonication (using
the standard procedure), the resulting sample was a clear, orange-yellow liquid with
water-like viscosity. Structures in the dispersion were probably uni-lamellar vesicles as
suggested above.

After allowing the samples to stand overnight, I separated them into 4 aliquots of
about 4 g each. Aliquot A was a practice run. I could easily identify the location of the
dye by a color band that moved slowly through the gel column. I detected the presence

of vesicles by using the hand-held HeNe laser to look for a scattering pattern in the liquid

132

exiting the column. Vesicles, even small ones, produce a distinct, bright shaft of light if
present in the liquid. I observed that between the end of vesicle elution and the beginning
of dye elution, there was a gap of about 60 seconds; more than enough time to remove the
collection bottle without fear of dye contamination. The vesicle dispersion had a slight
greenish hue, but otherwise looked the same as an ordinary vesicle dispersion.

To accurately determine the dye level, I prepared a standard means of testing the
dispersions. After the initial gel separation, a release period, and a second gel separation,
I added 0.5 g of the resin to each of the vesicle dispersions and placed them onto the
shaker table overnight. The following morning the dispersions were white and the ion
exchange resin was colored bright green, yellow, orange or red depending on the dye
concentration on the resin surface. I isolated the resin by filtration and transferred the
beads into a 20 mL vial. I then added 5 mL of the 30 wt % NaN03 solution to the vials
and placed them on the shaker table overnight. The following day, I removed the saline
solution from the resin, replaced it with fresh saline and returned the sample to the shaker
table. I repeated this procedure five times, and then measured the dye concentration in
the saline solution. I calculated the amount of dye extracted by the resin and compared
this to the amount of dye in a dispersion that did not undergo a release period.

I ran aliquots B, C and D through the gel column and collected the permeate of
each. (The volume of the vesicle dispersion increased ~ 9-fold as it passed through the
column.) Upon collecting the last of aliquot B, I immediately added ion exchange resin
to the permeate. The dye concentration found in dispersion B became my reference value
for fluorescein loading. The other dispersions, C and D were allowed to stand at room

temperature for several hours before being passed through the gel column a second time

133

to separate the released dye from the vesicles. The dye remaining in the vesicles was
then extracted using ion-exchange resin, and the amount of dye on the resin was
compared to that in dispersion B.

Dispersion C, which had not produced as clean a separation in the gel column as
dispersion B (the reference dispersion), showed about 64 % retention of dye in the
vesicles over a 5 hour period. Dispersion D, which when passing through the column had
a greater separation between the aqueous dye and the vesicles, showed a retention level
of 81 % after 24 hours. Because of the high dye concentrations, I used the calibration
curve generated by absorbances at 485 nm wavelength.

I decided to repeat the experiment using a lower dye concentration and allowing
the structures to stand for a longer interval before measuring the release rate. The
reduced concentration was intended to minimize the osmotic pressure gradient across the
bilayer and limit rupture of the vesicles primarily to that which occurred in the presence
of ion-exchange resin. I also used a BO/EO copolymer with a longer B0 block that I
believed would increase the vesicle stability and reduce the release rate. I prepared a

13.25 wt % dispersion using a B020E012 copolymer and a 100 ppm fluorescein dye

solution. The dispersion was ultra-sonicated to produce an opaque, pale-yellow liquid
with water-like viscosity.

I passed the dispersions through the gel column in the manner described above
and combined aliquot A with ion-exchange resin immediately after elution. Thus, aliquot
A served as my reference material. I allowed the other aliquots to stand for a period of

time before a second gel separation and subsequent extraction of dye from the vesicles.

134

The amount of dye extracted by the ion exchange resin was measured and compared to
the dye level of dispersion A. The results are plotted below.

Figure 4.7 shows the fraction of fluorescein dye released as a function of time.
The dye concentration in batch A was the reference concentration and is shown as the
100% value at “0” hours. The graph shows that the release rate is quite low for these uni-
lamellar vesicles; they retain almost 60 % of the encapsulated dye for more than 3 weeks.
This is significantly better than typical phospholipid structures, which can lose from 15 %

of encapsulated fluorescein in 300 minutes to 80 % in 200 hours.°"8

 

' 1
80” I I n
t 4

Fluorescein Dye Retained (%)

 

 

 

Time (hr)

Figure 4.7. Fraction of fluorescein dye retained by B020E012 uni-lamellar vesicles as a

function of time.

135

(NOTE: Gel filtration results are somewhat dependent on technique, and the initial small
rise in retention may be due to improvements in my ability to isolate the vesicles from the
released dye.)

It should also be noted that the resin release period was as long as several weeks,
and no precautions were taken to retard photo-oxidation of the dye. To determine if this
was a significant concern, I re-tested the 50 ppm standard used to prepare the calibration
curve after 140 days, about 20 % longer than the samples had been extracted. I found an
~ 10 % decline in the sample absorbance, and would conclude that a 3 to 5 % decrease in
fluorescein dye concentration would be reasonable for the dye samples shown here.

During the second gel separation, I collected both phases: vesicles and released
dye. In this way I could do a fluorescein dye mass-balance calculation. Figure 4.9 plots
the interval between the first and second gel separations versus the total dye
concentration (dye in vesicles plus dye released) for each sample. The shaded area at the
top of the graph represents a i 10 % variation in the dye mass measured in batch A, the
standard. Ten percent seems to be a reasonable level of error given the technique.
Clearly the first three batches produce sums that are within the range of errors allowed
and thus account for all of the dye. However, the final run is outside this range,
indicating that some of dye is unaccounted for. A problem arose during this run that may
explain this deviation. As the dispersion passes through the column the vesicles separate
from the dye, which is slowed by the gel. Therefore the vesicles exit first. Following the
vesicle collection for the final run, I inadvertently allowed the column to dry out. I do

not believe that this affected the separation of vesicles from dye, but I may not have

136

collected all of the released dye. If so, then the retention value of ~ 60 % is valid, but the

sum is short of the actual value.

 

 

 

 

 

 

I

22— ' -
E ' -
g 20- -
3
g .
5 18— -
i "
E
E 16- -
O
[-1

14- -

0 100 200 300 400 500 600

Time (hr)

F igurc 4.8. Sum of retained and released fluorescein dye for each batch of vesicles.

4.12 Conclusions

The use of B0 in amphiphilic BO/EO di—block copolymers results in large
hydrophobe volumes that lead to surfactants with compositions that spontaneously
produce vesicular structures in water at room temperature. Mean particle diameter is
about 3 um for hand—shaken dispersions, and typical structure yields are approximately
65 % of the total amount of copolymer. The vesicular structures are mechanically stable
and resistant to a variety of post-formation treatments, but they can be extruded or ultra-

sonicated to form structures of ~ 70 nm diameter. Additionally, both B011E012 and

137

BO. 1E019 vesicles are thermally stable to ~ 40 °C, at which point they undergo a
transition to a non-crystalline phase.

Ultra-sonication allows for the formation of vesicles at copolymer levels similar
to or higher than those for typical phospholipids. These structures may be used to
encapsulate either hydrophobic or hydrophilic compounds. Hydrophobe loading levels of
up to a 1:1 mole ratio of copolymer to dodecane still produce vesicles structures. The
hydrophilic loading occurs up to about 1 wt % of salt in solution. The release rate for

hydrophiles is equal to or slower than that for conventional phospholipids.

4.13 References

(l) Aniansson, G. E. A. Progress in Colloid & Polymer Science 1985, 70, 2-5.

(2) Marsh, D. Handbook of Lipid Bilayers; CRC Press: Boca Raton, FL, 1990; pp
353-355.

(3) Giddings, J. C. Unified Separation Science; Wiley-Interscience: New York, NY,
1991; p 173.

(4) Bailey, F. E., Jr.; Koleske, J. V. Configuration and hydrodynamic properties of
the polyoxyethylene chain in solution. In Nonionic Surfactants: Physical
Chemistry; Schick, M. J., Ed.; Marcel Dekker: New York, NY, 1987; Vol. 23, pp
927-969.

(5) Rosen, M. J. Surfactants and Interfacial Phenomena, Second ed.; Wiley-
Interscience: New York, 1989; p 21.

(6) Lasic, D. D. Liposomes in Drug Delivery. In Vesicles, First ed.; Rosoff, M., Ed.;
Marcel Dekker: New York, NY, 1996; Vol. 62, pp 447-476.

138

(7)

(8)

(9)

(10)

(11)

(12)

(13)

(14)

(15)

(16)

(17)

Gray, G. W.; Windsor, P. A. Liquid Crystals & Plastic Crystals; Physico-
Chemical Properties and Methods of Investigation; John Wiley and Sons: New
York, NY, 1974; Vol. 1; p 4.

Bedells, A. D.; Arafeh, R. M.; Yang, 2.; Attwood, D.; Padget, J. C.; Price, 0;
Booth, C. J. Chem. Soc., Faraday Trans. 1993, 89, 1243-1247.

Bergmeier, M.; Gradzielski, M.; Hoffmann, H.; Mortensen, K. Journal of
Physical Chemistry B 1998, 102, 2837-2840.

Zipfel, J .; Lindner, P.; Tsianou, M.; Alexandridis, P.; Richtering, W. Langmuir
1999, 15, 2599-2602.

Mueller, 8.; Boerschig, C.; Gronski, W.; Schmidt, 0; Roux, D. Langmuir 1999,
15, 7558-7564.

Ingle, J. D. C., Stanley, R. Spectrochemical Analysis; Prentice-Hall: Englewood
Cliffs, NJ, 1988; pp 513-515.

Gregoriadis, G. Fate of injected liposomes: Observations on entrapped solute
retention, vesicle clearance and tissue distribution in vivo. In Liposomes as Drug
Carriers. Recent Trends and Progress; Gregoriadis, G., Ed.; Wiley & Sons: New
York, NY, 1988; pp 176-188.

Betageri, G. V.; Jenkins, S. A.; Parsons, D. L. Liposome Drug Delivery Systems;
Technomic Publishing Company Inc: Lancaster, PA, 1993; pp 17,18.

Budavari, S., Ed. The Merck Index, Twelfth ed.; Merck Research Laboratories:
Whitehouse Station, NJ, 1996.

Ellis, D. W. Flame Emiss. At. Absorption Spectrom. 1971, 2, 263-287.

Schwartz, M. K. Analytical Chemistry 1972, 44, 9R-33R.

139

(18) Gerasimov, O. V.; Rui, Y.; Thompson, D. H. Triggered Release from Liposomes
Mediated by Physically and Chemically Induced Phase Transistions. In Vesicles,
First ed.; Rosoff, M., Ed.; Marcel Dekker: New York, 1996; Vol. 62, pp 679-746.

140

Chapter 5

Spontaneous Vesicle Formation from Poly(l,2-
Butylene Oxide) Sulfate Oligomers

5.1 Introduction

Chapter 3 showed that poly (1,2-butylene oxide! ethylene oxide) (BO/E0) diblock
copolymers spontaneously form multi—lamellar vesicles (MLVs) upon addition of water if
the average E0 block length is between 5 and 18 units. If the E0 block is larger than 18
units, spherical micelles result; if smaller than 5 units, an emulsion forms. This behavior
is consistent with a geometric packing model (section 1.6) that suggests that micellar
structure is determined by surfactant geometric properties such as head group area and
hydrophobe length and volume.l

Interestingly, while numerous ethoxylated surfactants exist, few demonstrate
spontaneous vesicle formation. The propensity of certain BO/EO copolymers to form
MLVs is thus most likely due to the large volume-to-length ratio of the B0 hydrophobe.
If this is the case, BO surfactants with other head groups (anionic, cationic or nonionic)

should also form vesicles as long as the head group has the appropriate size. Based on
the similarity in head group areas calculated for an E011 block and measured for -OSO3'

, we thought that sulfonated BO surfactants would also form vesicles.“3 In this chapter
we show that sulfonated B0 is indeed capable of forming vesicles, thus expanding the
versatility of B0 surfactants.

An ionic surfactant that spontaneously produces vesicles offers several options as
a dispersant that ethoxylated materials do not. Due to charge repulsion, vesicles prepared

from ionic surfactants are less likely than ethoxylated structures to form gels at elevated

141

concentrations.4 Anionic vesicles should also be stable at higher temperatures than
structures formed from nonionic surfactants. Finally, strongly ionic head groups, such as

sulfate, are relatively unaffected by changes in pH, although hydrolysis of the —OSO3‘

group could be problematic.5

5.2 Oligomer Synthesis and Characterization
BO oligonrers (Scheme 5.1a) were prepared following the procedure we
previously used to make EO/BO diblock copolymers, except that the butylene oxide

polymerization was terminated with acetic acid instead of continuing with the addition of

the E0 block. The hydroxyl terminated B0 was then sulfonated using a DMF*S03
complex.6 In a typical procedure, I transferred 10.00 g (0.0287 mol) of the B04 oligomer
to a bottle and added about 45 mL of anhydrous methylene chloride (MeClz) to form a
clear solution. Next, about 100 mL of anhydrous MeC12 was placed into a reaction flask,
which was cooled in an ice-bath and stirred under dry N2. To this was added 4.8 g of the

sulfur trioxide! DMF complex (Sigma-Aldrich, cat # 37,368-0). The complex was not

completely soluble in the MeClz, so a white solid was visible. The solution containing

BO oligomer was then added dropwise (over a 1 h period) to the cooled solution using an
addition funnel. The ice completely melted about 90 minutes after BO addition began,
and the flask was allowed to warm to room temperature with constant stirring. The

mixture, now a clear solution, was allowed to stir overnight under N2 at room

temperature.

The following morning I added another 90 mL of MeC12 to dilute the solution and

decrease its viscosity. The sample was evaporated to produce a golden, viscous liquid.

142

Deionized water (~ 50 mL) was added to the flask to produce a yellow liquid, and then
4.6 mL of a 5 N NaOH was added dropwise to the sample over 2 minutes while stirring.
(This is about 80% of the calculated amount of NaOH required to neutralize the sample.)
The pH of the solution was still quite acidic. Another ~ 0.7 mL of the 5 N NaOH was
added dropwise over 10 minutes, and the pH of the solution was examined after every
four drops. After the pH reached 7, I stirred the sample for l h and found the pH
unchanged.

After evaporation of the water at 75 °C for 2 hours, the resulting product (Scheme
5.1b) was a grainy yellow solid at room temperature. (The grains may have been residual

Na2304; B012 and B017 produced clear, yellow oils after drying.) The 803 group

insertion was confirmed by FT-IR (Figure 5.1), where new peaks at 1270, 750 and ~ 600

cm‘l indicate the formation R-O-SOz-O"’Na.7 Average BO lengths were determined by

‘H NMR in CDCI3.

CH3 CH3
Ia | 1b l
CH2 CH2 0

I so I |
CH3—CH2—(O—CH2—CH)n—O—H Ki? CH3—CH2—(O—CHZ—CH)n—O—S—O‘+Na

l |
0

Scheme 5.1. Poly(l,2-butylene oxide) oligomer structure before (1a) and after (1b)

sulfonation.

143

After the material cooled, I weighed out 0.0395 g into a vial and added 3.243 g of
DI water to create a 1.20 wt % dispersion. I agitated the vial vigorously and set it aside.
Light scattering data suggested a particle diameter of between 170 and 190 nm.

Interestingly, after a few months, the EtBO4SO4Na and Et801 ISO4Na underwent

phase separation. The oligomers formed a clear, upper phase and a viscous, opaque,
lower phase that contained a white precipitate. Dispersions prepared from the upper
phase did not form vesicles, but phase separated from water after mixing. All of the

vesicle data presented here were obtained from samples prior to any phase separation.

 

Absorbance

 

o v ,..'..,. 0'

1600 "”1400 1200 1000 "s00 600
Wavenumbers (cm")

 

Figure 5.1. FT-IR spectra of the 3017 oligomer before and after sulfonation.

Vesicles were formed by placing 1 drop (~ 0.03 g) of the surfactant into a 5 dram

vial and adding 3 g of deionized (>17 MQcm) water. The vial was then agitated by hand

144

to produce structures. Dispersions were allowed to stand for 24 hours before

examination.

To compare the thermal stabilities of EtBOl 1E0“ and EtBOl ISO4Na, 1 wt %

dispersions of each material were prepared by hand agitation and subsequent ultra-
sonication (Ultrasonic Processor, Model 130, Cole-Farmer Instruments, Vernon Hills, IL;

3 mm probe). The EtBOl 1E0“ copolymer forms large MLV structures using the hand-

shaking methods, so ultra-sonication was used to reduce these structures to a diameter

similar to that spontaneously formed by the EtBOl ISO4Na oligomer. To be

experimentally consistent, both dispersions were ultra-sonicated using the same
conditions. The dispersions were exposed to ultra-sonication for 2 min at 10 watts using
a “4 sec on/ 1 see off” cycle, while the vials were cooled in an ice-water bath.

Prior to light-scattering measurements (Brookhaven 90Plus, Brookhaven
Corporation, Holtsville, New York, USA), the 1 wt % dispersions were allowed to
equilibrate in a water bath at the desired temperature for at least 16 h. The dispersions
were evaluated at the equilibrium temperature, and returned to the water bath to
equilibrate at a new temperature.

Light scattering data were analyzed using the CONT IN algorithm and intensity-
weighted measurements. The average diameter was calculated from the mean values of
at least 5 separate measurements. Structures were also examined using cryo-SEM (JEOL

JSM-6320FV) as described previously in section 3.5.

145

5.3 Vesicle Formation and Stability

Vesicle formation requires that the hydrophobe length exceed some minimal
value. Nakamura et al. showed that BO sulfonates with less than 3 BO units displayed no
interfacial tension or film formation characteristics on water, so we felt that the
combination of the ethoxide initiator and 4 BO units would be about the minimum BO
block length necessary to initiate micellization.8 We chose a maximum BO length of less
than 20 units because our experience with EtB020E0n copolymers showed that
spontaneous vesicle formation from materials containing such a large hydrophobe was
difficult. This is likely due to the high energy needed to disperse such a large

hydrophobe prior to self-assembly into an ordered structure. Therefore, the three

oligomers synthesized for this study (EtBO4SO4Na, EtBOl ISO4Na, and EtBOl7SO4Na)

were intended to represent probable minimum and maximum BO lengths for spontaneous
vesicle formation as well as a composition comparable to the most common BO/EO

material we previously studied (EtBOl 1E0”). While having low polydispersity, these

materials are not monodisperse, and the formulae listed represent the average oligomeric
compositions of each surfactant.

Gentle agitation was sufficient to produce structures from EtBO4SO4Na and
EtBOl ISO4Na; however, EtB017SO4Na required sonication for complete surfactant
dispersion. Light scattering showed that the EtBO4SO4Na oligomer spontaneously

produced vesicles with an intensity weighted mean particle diameter of 180 nm, with
more than 95% of the particles having diameters between 88 and 263 nm. The

EtBOl ISO4Na oligomer spontaneously formed vesicles with a mean diameter of 140 nm

and greater than 95% of these particles have diameters between 62 and 188 nm. A

146

second dispersion of EtBOl ISO4Na was ultra-sonicated for 2 min to yield vesicles with

diameters of 130 nm.

EtB017SO4Na produced vesicles with a mean diameter of 150 nm after 30 sec of
ultra—sonication. This size, which varied by less than 30% from measurement to
measurement, is on the order of that seen with BO/EO surfactants after ultra-sonication.
The vesicles showed less than a 10 % change in particle size after standing at room
temperature for 30 days.

Cryo-SEM images are consistent with light-scattering data taken from the same

dispersion. Images such as Figure 5.2 show that EtBOl ISO4Na spontaneously forms

 

Figure 5.2. Cryo-SEM image of spontaneously-formed vesicles in a 1 wt%

EtBOl ISO4Na dispersion.

147

spherical particles with 30— to 300-nm diameters, and the majority of structures are in the
50 to 150 nm range. This sample was aged for 3 weeks prior to examination.

As would be expected for an anionic surfactant, structures formed from sulfonated
copolymers are more stable at high temperatures than vesicles formed from ethoxylated

copolymers. Figure 5.3 shows the diameters of EtBOl ISO4Na and EtB011E012

structures as a function of temperature. The graph indicates that both ethoxylated and

 

 

 

 

 

 

350 u- I ' fi r ‘ l ' V ‘ V I V v v v I v v u v I . t . .4
E 300 5- .‘
5 C 3
E», 250 i
Q) ' _
E :
.53 .
G 4
3 200 ' .4
.9 I j
g t .
a. 150 ~ l

- 1

100 t / é/ ;

20 30 40 50 60 70

Equilibrium Temperature (C)

Figure 5.3. Diameters of vesicles formed from ultra-sonicated EtBOl 113012 (e) and
EtBOl 1804N a (I) (1 wt %) as a function of temperature. (Error bars represent standard

deviations of the mean of 5 measurements.)

148

sulfonated BO structures increase in diameter at about the same rate upon heating. For
both surfactants, the increase in particle size is likely due to a decrease in the head group
area resulting from thermally induced dehydration.3'9 Dehydration decreases the radius
of curvature for the bilayer and increases the vesicle diameter. However, at temperatures
greater than 65 °C, the ethoxylated surfactant loses its amphiphilic properties and phase

separates from the aqueous medium, while EtB01 1804N a does not. (The EtB011S04Na

structures are stable enough that reproducible light-scattering data can still be collected at
75 °C, though the ability of the instrument to maintain that temperature accurately is
questionable.)

To further evaluate the thermal stability of the anionic surfactant, a 1 wt %

dispersion of EtBOl ISO4Na was sealed in a glass bottle and placed in an oven at 100 °C

for 9 days. Examination of the hot sample showed that while the dispersion had become
more opaque, there was no indication of phase separation or an increase in sample
viscosity. The cooled sample returned to its original appearance.

Since hydrolysis of the sulfate group in an aqueous environment is a possibility,
the chemical stability of the vesicles was evaluated. A 1 wt % dispersion of

EtB017SO4Na was prepared with ultra-sonication and found to have a mean particle

diameter of 140 nm by light scattering. The sample then stood at room temperature for 5
months with less than a 10% decrease in average particle diameter. Thus, at pH 6.8 these

vesicles are extremely stable.

149

5.4 Conclusions

Spontaneous formation of vesicles from both sulfonated BO oligomers and
BO/EO copolymers indicates that the B0 hydrophobe, not the E0 head group, is
primarily responsible for the formation of BO/EO vesicles. Spontaneous structure
formation by sulfonated BO oligomers occurs for average hydrophobe lengths of 4 and
11 BO units, and vesicles also form from sulfonated 3017 oligomers upon ultra-
sonication. Structures of all three sulfonated oligomers were between 120 and 175 nm in
diameter, which is consistent with the size expected from a surfactant geometry model.2
Furthermore, the B0 sulfate structures are stable for at least 5 months at room
temperature and give no sign of phase separation after standing at 100 °C for more than a
week. These structures are much more temperature stable than vesicles formed from

EtBO] 1E0”, as the ethoxylated materials phase separate at 65 °C.

5.5 References
(1) Israelachvilli, l. N.; Mitchell, D. J.; Ninham, B. W. J. Chem. Soc., Faraday Trans. 2
1976, 72, 1525-1568.

(2) Crook, E. H.; Fordyce, D. B.; Trebbi, G. F. J. Phys. Chem. 1963, 67, 1987-1994.

(3) Rosen, M. J. Surfactants and Interfacial Phenomena, Second ed.; Wiley-
Interscience: New York, 1989, pp 70-80.

(4) Schmolka, I. R. J. Am. Oil Chemists Soc. 1991, 68, 206.
(5) Asinger, F.; Geiseler, G.; Muller, G. Chem. Ber. 1960, 93, 2491-2497.
(6) Kelly, K. K.; Matthews, J. S. J. Org. Chem. 1971, 36, 2159-2161.

150

(7) Colthup, N. B. Journal of the Optical Society of America 1950, 40, 397-400.

(8) Nakamura, M.; Sakashita, T.; Mori, T.; Yano, W.; Kimura, W. Yukagaku 1972, 21,
95-99.

(9) Corkill, J. M.; Goodman, J. T.; Tate, J. R. Trans. Faraday. Soc. 1966, 62, 979.

151

Chapter 6
Summary and Future Work

6.1 Background

This research was based primarily on the micellar model developed by
Israelachvili et al. that states that micellar shape formed during self-assembly is
dependent on surfactant geometry. Limitations on the surfactant packing configuration,
which determine the shape that the aggregate will form, are a function of hydrophobe
volume and length, and hydrophilic head group diameter. The twin tails of phospholipids
facilitate vesicle formation because the large hydrophobe volume makes packing the tail
into a micelle core unfavorable. We hypothesized that a single tail surfactant with a
similar volume to length ratio would also produce vesicles.

The BO oligomer has a much larger volume-to~length ratio than most
hydrophobes, making it a potential substitute for the two tails in phospholipids vesicles.
This hydrophobe also offers the advantage of being much more hydrophobic that the
propylene oxide monomer commonly used in nonionic surfactants. Increased
hydrophobicity enhances the driving force for self-assembly and increases the stability of
structures formed from BO containing surfactants.

Selection of an appropriate head group was not straightforward because the actual
size of surfactant head groups in solution is not well known. Thus, we chose to use a
nonionic head group because we could alter its size easily and to a greater extent than
with typical ionic head groups. By varying the E0 block length we hoped to find the

correct hydrophile to form vesicles from the B0 hydrophobe.

152

6.2 Summary

Copolymer synthesis is a two-step anionic polymerization using an alkoxide
initiator. The reaction is fast and efficient with few byproducts or side reactions.
However, the transition from polymerization of the B0 block to initiation of the E0
block does not have 100 % efficiency, and the reaction can produce hydroxy-terminated
BO oligomer instead of diblock copolymer. This has two effects: copolymer
contamination with a hydrophobic material and an increase in the average E0 block
length for the copolymers that do contain an E0 segment.

After evaluating numerous BO/EO diblock compositions I found that when the
E0 block length is less than 16 and greater than 6, and the B0 block length is less than
17 and greater than 4, the copolymer forms vesicles spontaneously upon mixing with
water. For similar E0 block lengths, and for BO blocks less between 17 and 25 units,
vesicles could be formed using ultra-sonication of the dispersion. However,
determination of composition using NMR analysis alone is not sufficient to predict
vesicle formation. I found that copolymer batches that had a mean composition of

BOIOEOn and a high polydispersity did not form vesicles. The physical properties of

the sample indicated an average that a substantial fraction of these copolymers had an E0
block length greater than 17 units. Fractionation of this material using hexane and low
temperature centrifugation isolated the copolymer with shorter E0 blocks to yield
copolymer that did spontaneously produce vesicles. Thus, while mean BO/EO
composition is important to vesicle formation, a narrow distribution of E0 block lengths

is also important. To more accurately determine the range of E0 block lengths in a

153

batch, I used mass spectrometry, which provided important information concerning BO
oligomer concentration and surfactant composition.
For relatively mono-disperse copolymers with a nominal composition of

BOIZEO 12, vesicles formed at copolymer concentrations between 0.05 and 10 wt % upon

simple hand—shaking. However, at concentrations greater than 25 wt %, vesicle
formation required ultra-sonication. BO/EO structures were stable for more than a 24
months at room temperature with little or no discernible change in particle size
distribution. Typically more than 80 % of the copolymer added was found in vesicular
structures. I also discovered that while vesicle formation is spontaneous, the size
distribution of the structures is a function of formation method. Simple agitation
produces micron-sized vesicles, but if the copolymer/ water mixture is ultra-sonicated,
the particles will have a diameter of about 70 nm. The same effect occurs if the mixture
is extruded through a polycarbonate filter. Solvent exchange processes also generate
structures that tend to be in the 70 — 100 nm range. If the copolymer is brought into
contact with water at ~1 wt % level and allowed to hydrate without agitation, giant
vesicles with diameters of 1 mm or greater can be produced. However, the process must
include thermal oscillations (> 10 °C) in order for multi-lamellar giant vesicle formation
to occur.

Once formed, these structures demonstrate good mechanical stability. Shear
applied by mixing, sonication or shaking had no effect on mean particle size distribution
when performed at room temperature. This stability is probably due in part to the liquid
crystalline nature of these structures. However, plane—polarized light microscopy shows

that the structures undergo a phase transition at about 40 °C that would probably reduce

154

this stability. Cooling the structures reformed the liquid crystalline morphology. In fact,
samples heated to 80 °C for three hours, then cooled and agitated showed the same mean
particle diameter as the original dispersion.

To confirm that the B0 oligomer was critical to vesicle formation, I replaced the

E0 block with an 803' group that is similar in size to the “predicted” diameter of an
E015 block. The sulfonated compositions also formed vesicles upon mixing with water
when the B0 block was 12 units long or less. The 803' oligomers showed improved

thermal stability versus the ethoxylated copolymers.

Because BO/EO vesicles consist of both hydrophilic and hydrophobic regions, I
examined the capacity of the structures to encapsulate both hydrophilic and hydrophobic
substrates. I used a fluorescein dye as the hydrophilic compound and loaded the vesicles
by mixing the copolymer with 100 ppm aqueous fluorescein. The vesicles were then
passed through a Sephadex column to separate the unencapsulated dye from the vesicles.
Less than 20 % of the fluorescein leaked out of the vesicles over a one month period. I
also found that I could add dodecane to the copolymer in a 1:1 mole ratio and produce
vesicular structures. There was no evidence of leakage from these vesicles over a two

month period.

6.3 Future Research

Because of the versatility of these materials, there are several areas that would be
of interest to explore. Phase diagrams showed that the presence of a hydrophobe has an
impact on structure formation. We also know from mass spectrometry data that there is

typically residual BO oligomer in each batch of copolymer. It would be interesting to

155

develop a technique for removing the B0 oligomer without extracting diblock copolymer
that has short EO blocks. One solvent which is promising for such work is dimethyl
sulfoxide (DMSO), which is quite effective at solubilizing EO but poor at dissolving B0.

A second area of study would be the effect of EO block polydispersity on vesicle
formation. Surfactant polydispersity and its impact on vesicle formation is a
controversial topic. Some have argued that polydispersity is necessary to allow those
molecules on the concave portion of the structure to assume a smaller head group
diameter than those on the convex side. We have shown that excessive polydispersity
can hinder vesicle formation, even if the mean composition is within the desired range.
One means of investigating the effect of polydispersity on structure formation would
involve the synthesis of mono-disperse BO/EO diblock copolymers.

Lithium-based initiators allow for the addition of single EO units per run. 1'2 Since
an E0 block of only 8 or 10 units is sufficient to produce structures, making a mono-

disperse batch of BOIOEOB would be a realistic target. This material could be used as a

model to study the effects of added B0 or other BO/EO copolymers on vesicle formation.
Further vesicles from these materials could be used to accurately determine the mean EO
head group diameter in solution as a function of EO block length. Preparation of a
second mono-disperse copolymer with a composition of BO 105025 would be time

consuming, but mixing this with BOIOE08 would show the impact of EO polydispersity

on vesicle formation.
Another area of interest is to use the static vesicle formation process employed

with B012E012 to form giant vesicles with other copolymers, preferably those with

larger EO blocks. Clearly identifying the conditions necessary for giant structure

156

formation could yield a better understanding of the mechanism of vesicle formation from
these materials. This would provide additional information on the thermodynamics
involved and the stability of vesicles formed from BO/EO copolymers.

The BO oligomer sulfonation also presents some research opportunities. We have
shown that the anionic oligomer forms vesicles spontaneously upon hydration, and while
there is no evidence to suggest that cationic materials would not do the same, it would be
worth demonstrating. Others have shown that the blending of anionic and cationic
single-tailed surfactants can lead to spontaneous vesicle formation, so the blending of
anionic and cationic BO oligomers would also be interesting. In fact, preparation of a
variety of B0 oligomers with various head groups could provide some insight into the
size of the head group in solution, an area not well understood at the moment.

Finally, ethoxylation of oligomers that have the same or similar V/lc ratios as the

B0 oligomer would provide an expanded source of potential vesicle-forming
hydrophobes. These new hydrophobe may offer advantages over B0 in terms of cost,
ease of preparation, or versatility. The resulting surfactants may provide increased

structural stability, improved bio-compatibility or even enhanced biodegradation.

6.4 References

(1) Guilbert, Y.; Brossas, J. Polymer Bulletin (Berlin, Germany) 1979, I, 293-298.

(2) Cabasso, I.; Zilkha, A. Journal of Macromolecular Science, Chemistry 1974, A8,
1313-1324.

157

APPENDICES

158

Appendix A
Geometric Limitations Imposed by Surfactant Packing Shape

Al The Critical Packing Factor Model

The critical packing factor model is based on the limitations imposed on a structure by its
geometric shape. As a first example consider a sphere. Previous discussions (section
1.5) have described the number of aggregates in a spherical micelle as N. The volume of
a sphere is shown in equation A1.1.

V:%flR3 eq.A1.l

The surface area of a sphere is given by equation A1.2.
A = 47rR2 eq. Al.2

Dividing A1.1 by A1.2 yields R/3 or the following relationship (equation Al.4) that is
fixed for any sphere.

ILK

e .A1.3
3 A q

Here R is the radius of the sphere which, in terms of the micelle, corresponds to the
hydrophobe length, lc. This equation can be rearranged to produce equation A1.4 which

shows the relationship between V, a0, and IC for a spherical micelle.

 

eq. A1.4

0C

This value of 1/3 sets an upper limit on the ratio that these values can have and still

produce a sphere. Therefore, in order to form a sphere, the ratio of the hydrophobe
volume to the product of the hydrophile head group area and the hydrophobe length must
lie between 0 and 1/3. These same geometric limitations can be shown for cylindrical

micelles and bilayer structures.
A2. Aggregation number versus Head Group Size

The volume of each surfactant molecule in the structure is that of a right circular cone
(equation A21).

159

v =—rzr2h eq.A2.1

Here r is the radius of the head group while h is the length of the hydrophobe, which is
the same as R for the micelle. We can divide equation A1 .1 by A2.] to determine the
number of surfactant molecules, N, in a given micelle.

-4153.
3 v

N eq. A2.2

By substitution of a for *rtr2 in equation A2.], equation A2.2 can be re-written as equation
A2.3.

_ 47rR2
a

N

 

eq. A2.3

This equation shows an inverse relationship between the aggregation number, N, and the
head group area, a which is consistent with Figure 1.3.

160

 

Appendix B

Processing the Mass Spectrometry Data

Detailed information such as residual BO oligomer concentration and the
distribution of copolymer block lengths can be obtained by analysis of mass spectrometry
data. Table B.1 presents one set of mass spectrometry (MS) raw data generated by the
Analytical Lab at The Dow Chemical Company and forwarded to me for additional
analysis. These data had been processed to remove contributions from multiply charged

ions. Values of MW, MN, Mp and Mw/MN, were calculated for the entire data set and

were provided as well. Also included were graphs of each data set shown in Figure B. 1.
The data were initially arranged in order of increasing mass as shown in Table
B. 1. In the second column, M W— 18, the mass of the ammonium ion (18.0344 amu) was
subtracted from the detected mass to determine the actual mass of the molecule. The
signal intensity for that mass was recorded in column 3, Intensity. I assumed that little
molecular fragmentation occurred during analysis, and that all data represented intact
copolymers. Thus, the ratio of the intensities of any two masses that represent specific
copolymer compositions should correspond to the molar ratio of those compositions
present in the sample.
I prepared a second table that lists the calculated masses of BO/EO compositions

ranging from BO3EOO to BOngOzg. (Table B.2 shows a portion of these data.) I used

masses of 72.0575 amu for each BO group, 44.0262 amu for each 130 group and 46.0419
amu for each chain initiator. These are the average masses of each unit assuming a

typical distribution of the isotopes. However the masses of each composition are unique

161

only within a limited range. Note that the value of BOI4EOO is quite similar to the value
of BO3E013, differing by less than 0.20 amu. This occurs because 11 BO units have
approximately the same mass as 18 E0 units, so copolymers BOXEOy (y > 18) have a
mass similar to copolymers BOX+11EOy_13. Thus, copolymer batches having narrow BO

polydispersity (< 11 units) minimize the confusion over assignment of composition to a
given mass by restricting the mass to a unique composition. When a specific m/z value
could be assigned to more than one composition, I initially assigned it to the composition
whose calculated mass was closest to the actual mass, then later made adjustments if
needed when all of the intensities had been assigned. For example, if an intensity value

was assigned to BOI4EO3, but at the end of the analysis there were no other

compositions having a B0 length of 14, I would probably transfer that intensity value to

BO3E021, particularly if there were other long EO compositions in the batch, or if there
were no intensity values for any compositions containing B013.

To begin the analysis, I sorted the raw data by decreasing intensity and inserted
the column labeled Range (Table B.3). Using the MW — 18 mass and the compositional
masses found in Table B.2, I matched the recorded mass to a BO/EO composition and
entered the intensity for that mass into a third Excel spreadsheet like the one shown in
Table B4. The differences between the recorded mass and the compositional mass
usually fell into one of 4 groups. When the actual masses differed by less than 1.15 amu

‘61”

from the calculated mass, I labeled the accuracy range value . For range “2”, the

actual mass was ~18 amu less than the calculated compositional mass. (These materials

may have lost an H20 molecule.) A third type of data (range “3”) had masses ~29 amu

less than a calculated composition (perhaps the loss of a CH3CH2 group), and m/z values

162

 

 

that did not match an BO/EO composition were given a range value of “u” (unknown).
The majority of the data used for compositional mapping (> 95 %) had range values of
“1”, and in general, the actual masses were between — 0.8 and + 0.5 amu of their
corresponding compositions.

As an example, the actual mass (M W— 18) of the first entry of Table B.3 is
1162.866 dal. The corresponding composition (Table B.2), with a mass of 1162.853, is

found at EOIOBOg. Thus, on Table B.4, under composition “BOloEOg”, an intensity

value of 4830 was entered. Finally on Table B.3, in the Range column, a 1 is recorded to
note that that the (M W- 18) value (1162.866) was within 1.5 amu of the calculated mass

for BOIOE09. If another (M W— 18) value were found that matched BOIOE09, for

example, 1162.861, then the intensity of that mass would also be entered under the

composition “BOlOEOg”. Summation of all of the intensities for a particular composition
yielded a total intensity. (The value shown for “BOIOE09” in Table B4 is 4880 because

the total of all of the intensities recorded for that composition is 4880.) This process was
continued until the first 70 % of the total intensity, when arranged by decreasing
intensity, was matched to a composition.

When all of the entries were made to Table B.4, the sum of the intensities for a
given composition was plotted as a function of B0 and E0 block lengths. Figure B.2 is
the graph for copolymer L120-15A, a material shown by NMR to have a composition of

BOGEO4. The MS data indicate an average composition of BOIOEO6, but they also show

high levels of B0 oligomer. The presence of uncapped oligomer is an important factor in

explaining the phase behavior of this material.

163

In cases when the B0 oligomer was available for analysis, MS data of this
material aided in determining the copolymer compositional profile. The BO oligomer
distribution for batch M1038 is shown in Figure B.3. I first plotted the B0 data to
determine the minimum and maximum BO oligomer lengths as well as relative amounts
of each composition. This was done by subtracting the ethoxide initiator from each block
(- 46.0419) then dividing the remaining mass by 72.0575. The result was the BO block
length, which was generally close to integer values. (Integer values are represented by
vertical lines in Figure B.3.) Isotopic compositions were typically within 1 amu of the
most intense peak and appear as lower intensity points on the same line. Contaminants or
byproducts appear as materials not found on an integer value and can be seen in the
figure as the low intensity deviations. Note that the range of BO oligomers for BO-

M1038 is about B06 to B022. I used these data limits to help resolve questions about

matching a mass to a composition when two calculated masses were similar to the actual

mass.

164

Table B.l. A portion of the mass spectrometry data for copolymer L1 150—5 1A

(30] lEOlz), arranged by increasing mass, after being processed to remove multiple-

charge signals.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Molecular
Weight MW - 18 Intensity
319.1746 301.1402 3.78E+01
327.2177 309.1833 7.99E+01
335.2996 317.2652 7.26E+01
341.2262 323.1918 2.22E+01
345.3176 327.2832 1.34E+02
349.2039 331.1695 3.88E+01
358.2583 340.2239 1.01E+02
363.2011 345.1667 2.77E+01
372.2481 354.2137 5.17E+01
380.3179 362.2835 4.13E+01
381.3273 363.2929 1.88E+01
388.2802 370.2458 2.69E+01
389.2590 371.2246 5.03E+01
393.2210 375.1866 3.89E+01
402.2861 384.2517 1.94E+02
407.2470 389.2126 8.79E+01
415.2693 397.2349 3.64E+01
416.2724 398.2380 8.01E+01
424.3780 406.3436 5.48E+01
430.3210 412.2866 5.56E+01
432.2984 414.2640 1.20E+02
433.2211 415.1867 3.79E+01
437.2598 419.2254 7.08E+01
444.3050 426.2706 6.22E+01
446.3056 428.2712 1 .94E+02
447.3542 429.3198 3.65E+01
451.2766 433.2422 9.00E+01
452.3735 434.3391 1.28E+02
456.9733 438.9389 4.1 1E+01
460.3132 442.2788 1.20E+02
465.2775 447.2431 3.59E+01
472.3450 454.3106 4.81E+01
474.3278 456.2934 9.84E+01
476.3259 458.2915 1.7ZE+02
477.3170 459.2826 9.49E+01
479.4138 461.3794 9.89E+01
481.1236 463.0892 7.91E+01
483.3810 465.3466 7.68E+01

 

165

 

Alt-30 Run #1101 poly(BO)

 

 

 

 

 

 

 

MAY-2001 10:10:46

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Harris Midland ASL um Spec
Flow Injection momma/11114111: Boll Q-TOF
md2001-002397bb5 5 (0.610021) Cm (3213) 1: TOF MS E84-
712. 5.9863
mm M1101 BO
1 640.5499\ ”56'7258
. /928.7844 Mw - 828
1000.6536
566.4904 / Mn - 781
I %_ 1072.9227
. Mz - 875
1073. 9274
496.4402 1146. 0023 Man , 1.06
””206 [1111. 1269.1193
c rune-EHLLHHA ...... ,...,....,
md2001-002397bb4 51(20'6104) Cm (3112) 1: TOF MS ES+
5.8563
100: 64054997: /856. 7136 M1096 80
6.7644
‘ (92 Mw - 830
4 1000.6406
%_‘ 5684904 1072.9094 M" ' 732
‘ 1073.9274 Mz - 879
496.4402 [(1145-9883 M /M
1289.1045 W " " 1-06
479.4117 1 . 113424-55, T .--,.---,.--
md2001-002397bb2 5 (0,351.6)201: (2:11) M1038 80 1: TOP MS ES+
100 1000. 8408 5 4163
1 712.6024\ / Mw _ 914
1 1072.9094
640.5499 /
\ 1/144.9645 Mn - 862
%~ ,11459663 Mz - 966
1 566.5002
1216. 0343
* Mw/Mn - 1.06
« 4964402 {[I136111508582676
of--,....3,h..l£‘d.'.¢1:.l isr : ,...., ,...-, -fim/z
500 1000 2000 2500

Figure B.1. Mass spectra of three BO oligomers along with calculated average

molecular weights and polydispersities.

166

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

22 2 a 24.28 22.32 42.22 22.22 52.2: 22.22 48.82 65.22 @222; .382 32.22
84.52 2.22 22.22 5.22: 822: 2.8: 2222 23.82 8322 «2.22 22.22 :32:
42.22 62.32 82.22 22.82 2 .2: 62.82 98.32 :32 $5.82 83.2 3.82 3:22
22.22 22.82 22.22 236: 2 .32 2 2 .22 29:2 23:; 232; 22.22 22.22 22.2:
22.22 22.82 62.5: 2 .2: :2 :2 48.22 68.82 25.2: :32: 42.22 62.22 224:
2.2. .2: 22.22 233: m: .22 2 2 .82 2o. 2 22 8°22 22.22 22.2 2 5.22 2.2:: N :22
22:2 :22: 3.2.8: 2: . 2 22 26.22 232; 48.2: 2.2.2.2 2222 532: 32.2: 26.22
22.82 222: w: .22 o2 .22 28.22 28.2.: 25.22 02.22 22.22 232: 22.22 2.622
62.2: 62.22 2 .22 48.2.2 23:; 68.22 23.22 42.22 62.2: 22.2 : 36.22 26626
22.2: 22 .22 2. :2 28.82 232: 22.22 22.22 22.22 22.2: 22.82 23.62. 86.26
222 2 .22 28.22 «3.22 42.22 32.2 2 @222 :22: 4232 62.22 2352 232
22.22 22.22 28.2: 23 :2 26.22 86.22 2.3.2: 222: 2822 233 26.68 22.2
2.22 2 .2 :2 :52: 26.82 23.62 5.22 :32. 62.22 «2.82 $6.42 222 222
22.22 2322 2.8.22 22.22 25.82 2.2.2: 832: 2522 26.8.2 232 83; 82.2;
2 .2: N22: 3322 2322 65.82 N22: 3262 838 2.6.2 a 832 42.4: :32
25. 2 2: 20.22 22.82 23.22 2222 2 2:52 2; _2 26.26 23.; 222 232 24.26
232; 83.22 2.362 42.8: 22.2 : 22.2.2 233 42.25 822 232 22.23 424.30
26.82 28.82 2222 22.2.: 82.22 22.82 232 232 02.2: 23.: 24.26 232

E22 2.6.2.2 223:. “2.8: E22 :32 62.62 86.2w 3.2: 22$ 62.22 22.62
26.22 23.82 22.2: 62.22 22.26 2%.: 2 26.2.2 23: 23% 84.6% 83.2 2222.
2222 4222 _ 22.22 2:22 2526 $6.22 8622 2.32 23.2 42.22 E2 2 2 2.24
68.2: 22.2 : 5.22 22.26 26.22 26.62 22.42 22.26 2426 M2422 2.2; 22.42

16%: _ 22.22 42.2... 5.68 25.32 26.2: 42.0: 84.26 2362 22.42. 42.24 $32
22.22 222.22 25.2 522 28.22 232 238 :32 24.22 6262. 232 3.82
3.22 822 233 232 :32 62.4% 82% 33.02 222:. 2262. 232 232

Eon 2% 2% Zen Sam «on new “on new 262 new «on

.vamEOm smacks: oOmmOm 2028258 28 2224238 283% Oman go 322: 322330 .Nfi 039—.

VNQN
MNQQ
NNQM
NNQW
cNQQ
Q~Qm
%NQW
N~Qm
b~©fl
WNQQ
VHO”
MNQW
N~QW
-°H
%NQMN

«am

new

new

462

wow
new

Sm
$2

167

Table B.3. A portion of the copolymer L1 150—5 1A mass spectrometry data, arranged by

increasing intensity, and including a “range” designation.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Molecular
Range Wg’Lht MW - 18 Intensity
1 1 180.9 1162.866 4.83E+03
1 1225 1206.966 4.74E+03
1 1 108.9 1090.866 4.68E+03
1 1 152.9 1134.866 4.63E+03
1 1136.9 1118.866 4.61E+03
1 1064.8 1046.766 4.45E+03
1 1253 1234.966 4.41E+03
1 1036.8 1018.766 4.40E+03
1 1269 1250.966 4.39E+03
1 1020.8 1002.766 4.39E+03
1 1208.9 1190.866 4.20E+03
1 1297 1278.966 4.16E+03
1 1080.8 1062.766 4.13E+03
1 1196.9 1178.866 4.09E+03
1 1092.9 1074.866 4.07E+O3
1 992.8 974.7656 3.97E+03
1 1124.9 1106.866 3.95E+03
1 1164.9 1146.866 3.81E+O3
1 1341.1 1323.066 3.81E+03
1 948.7 930.6656 3.80E+03
1 1325 1306.966 3.79E+03
1 1313 1294.966 3.76E+03
1 1048.8 1030.766 3.75E+03
1 1281 1262.966 3.70E+03
1 1120.9 1102.866 3.57E+03
1 976.8 958.7656 3.54E+03
1 1168.9 1150.866 3.53E+03
1 964.8 946.7656 3.52E+03
1 1008.8 990.7656 3.50E+03
1 1240.9 1222.866 3.48E+03
1 784.7 766.6656 3.45E+03
1 856.7 838.6656 3.42E+03
1 1052.8 1034.766 3.42E+03
1 1237 1218.966 3.34E+03
1 928.8 910.7656 3.31E+03
1 1385.1 1367.066 3.26E+03
1 712.6 694.5656 3.24E+03
1 1369.1 1351.066 3.21E+03

 

 

 

 

 

168

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

:2: KS cm: 22 38 c3
ow: ohm _ ca: cmom oma 83 82 on _ _
oflm 2.: 88 8mm Sow Sam 28 Sam 22 8:
ONE 83 cmmm 83 8% Oman coon emom omfi cm: onc—
mg 2.2 oovm owmm mmmv owmv anon Saw 2? oamm ommm 32 22
ONE 83 ovom cm; 8% Once coco ofim 8%. 80m 85 ONE Omn—
om: 32 cm?” wmon 8% RS 020 cm? omen 2.5m oven 8:
$2 82 ovmm can omwv 8mm 92% owom cwmv omen omfim o:—
oafl 82 2mm gen 89. Raw ovmv ommm owmm comm o9:
moa cam: 03m wwmm 2 S gem 943 83 82
33 cm? emom cmom on _ m ova an: m3
8: 2S ovfl 8: 89 2m— 2‘:
$2 _ 8: $3 Ommm ovmm ow? co: 8% out
:2— so: 2am :2. 2o: 29. :9. son mom. MWcm 5m can mom 5m mo:

.cotmmomEoU nose mo £225 33 mcsofi 805m 32% .832:me .acozmmanoQ Bax—once go oEmem in flash.

63282 3qu €385 HmonwE om C

169

Intensity

 

Figure B.2. Compositional profile of copolymer L120-15A (angled view). The mean
composition of the above peaks is BOIOEOG, while NMR indicated a composition of

BO6EO4. The expected composition based on synthesis formulation was EOIOBOIO

170

 

 

 

+ T4 . q 1 q . a q . .L 2
.2 :3......-.....-..-......--....-....:i.:t..‘1 m.
1. ii! i . i.......o.....ilu...-:3.3-2.-.32.383.582.5“I

 

 

 

 

 

 

iii .3 on :u a: 30.2.48:

13...! ..... A:..‘2.2:ti.‘t .....
l... 3.... 3. 3:6.-- d

I A A u- 2 3:93! ill
T.;... i. - .. 35..-... -

,tl‘ii u-- - 3 .48.. ll .3

133.43! in- - - At: 2...

 

 

Number of B0 Units

Figure B.3. The compositional distribution of the B0 oligomer from Batch M1038.

Note that the majority of the material (peaks with highest intensity) lie on integer values

of B0 length.

171

IIIIIIIIIIIIIIIIIIIIIIIIIIIIII

llHLHHllllljflllllllljjllljlflllltfllljlljfllil