Cross-linked normal hexagonal and bicontinuous cubic assemblies via polymerizable gemini amphiphiles

Article abstract of DOI:10.1021/ja0208106

The synthesis and lyotropic liquid-crystalline (LLC) phase behavior of a homologous series of intrinsically cross-linkable gemini surfactants are described. These novel bis(alkyl-1,3-diene)-based phosphonium gemini amphiphiles exhibit "normal" hexagonal (H_I), Type I bicontinuous cubic (Q_I), and lamellar (L_α) phases in water, and can be photocross-linked with retention of phase architecture in each case. On the basis of their locations on the phase diagram, their powder X-ray diffraction profiles, and the physical properties of the cross-linked materials, the Q_I phases formed by these gemini monomers are consistent with four possible bicontinuous cubic architectures with P or I space group symmetry that have been identified previously for small molecule amphiphiles. The extent of polymerization (i.e., the degree of diene conversion) achieved in the LLC phases was determined to be in the 23% to 71% range using UV-vis spectrometry, which is more than sufficient to extensively stabilize the systems. The resulting cross-linked H_I, L_α, and Q_I phases are stable up to 300 °C in air. To our knowledge, these reactive amphiphiles constitute the first example of a polymerizable gemini surfactant, and the first example of a cross-linkable amphiphile system that can be polymerized in both the H_I and a Q_I mesophase with retention of phase microstructure.

Full text of DOI:10.1021/ja0208106

Published on Web 02/12/2003
Cross-Linked Normal Hexagonal and Bicontinuous Cubic
Assemblies via Polymerizable Gemini Amphiphiles
Brad A. Pindzola,†,‡ Jizhu Jin, and Douglas L. Gin*
§
,§
Contribution from the Department of Chemistry and Biochemistry, and Department of Chemical
Engineering, UniVersity of Colorado, Boulder, Colorado 80309, and Department of Chemistry,
UniVersity of California, Berkeley, California 94720
Received June 7, 2002 ; E-mail: gin@spot.colorado.edu
Abstract: The synthesis and lyotropic liquid-crystalline (LLC) phase behavior of a homologous series of
intrinsically cross-linkable gemini surfactants are described. These novel bis(alkyl-1,3-diene)-based
phosphonium gemini amphiphiles exhibit “normal” hexagonal (H
) phases in water, and can be photocross-linked with retention of phase architecture in each case. On
the basis of their locations on the phase diagram, their powder X-ray diffraction profiles, and the physical
properties of the cross-linked materials, the Q phases formed by these gemini monomers are consistent
I I
), Type I bicontinuous cubic (Q ), and lamellar
(L
R
I
with four possible bicontinuous cubic architectures with P or I space group symmetry that have been identified
previously for small molecule amphiphiles. The extent of polymerization (i.e., the degree of diene conversion)
achieved in the LLC phases was determined to be in the 23% to 71% range using UV-vis spectrometry,
which is more than sufficient to extensively stabilize the systems. The resulting cross-linked H , L , and Q
I R I
phases are stable up to 300 °C in air. To our knowledge, these reactive amphiphiles constitute the first
example of a polymerizable gemini surfactant, and the first example of a cross-linkable amphiphile system
that can be polymerized in both the H
I
and a Q
I
mesophase with retention of phase microstructure.
Introduction
hydrocarbon domains can adopt different relative locations with
respect to each other (Figure 1). LLC aggregate structures range
1
Lyotropic liquid crystals (LLCs) are amphiphilic molecules
that spontaneously organize in the presence of water into highly
ordered yet fluid, phase-separated assemblies with periodic
features on the 1-10 nm size regime. Polymerizable versions
of these molecules and their assemblies have recently generated
from simple micelle systems to condensed anisotropic as-
semblies composed of ordered arrays of cylinders (i.e., hex-
agonal (H) phases), and bilayers (i.e., lamellar (LR) phases).
Depending on whether the hydrophobic/hydrophilic interface
has a net positive mean curvature (i.e., curves toward the
hydrophobic domains) or a net negative mean curvature (i.e.,
1
a great deal of interest as a means of constructing functional
nanostructured materials. By polymerizing these ordered arrays
2
1
a
curves toward the aqueous domains), these phases can be
further classified into Type I (oil-in-water, or “normal”) or Type
II (water-in-oil, or “inverted”) phases. The LR phase is consid-
ered to have zero mean curvature and serves as a midpoint in
an idealized progression of LLC phases with increasing water
of molecules into heavily cross-linked polymer networks, robust
organic materials with unique nanoporous architectures can be
generated that are ideal for the nanoscale reaction, encapsulation,
and/or transport of molecules in either the aqueous or hydro-
carbon domains.2
A large part of the interest in using reactive LLC assemblies
for nanomaterials design stems from the fact that LLC phases
offer exceptional control over nanoarchitecture because of the
range of structures they form and the fact that the aqueous and
1
content and curvature. In addition to the LR and H phases, a
number of bicontinuous cubic (Q) phases have also been
identified, which can exist between any of the LLC phases
previously described and the simple micellar systems.1 These
Q phases are termed bicontinuous because they consist of two
,3
†
‡
or more unconnected but interpenetrating hydrophobic and/or
Department of Chemistry, University of California, Berkeley.
Current address: ABS, Inc., 701-4 Cornell Business Park, Wilmington,
1,3,4
aqueous networks with overall cubic symmetry.
Depending
Delaware 19801.
on where they appear on the phase diagram relative to the central
LR phase, these Q phases can also be classified as Type I or
Type II (i.e., they have the same cubic symmetries but inverted
locations of the aqueous and hydrocarbon domains).1a
A number of reactive amphiphile systems have been designed
that form LR and HII phases and retain their original LC phase
§
Department of Chemistry & Biochemistry, and Department of Chemical
Engineering, University of Colorado.
(
1) For general reviews on LLC phases and their classifications, see: (a) Tate,
M. W.; Eikenberry, E. F.; Turner, D. C.; Shyamsunder, E.; Gruner, S. M.
Chem. Phys. Lipids 1991, 57, 147. (b) Seddon, J. M. Biochim. Biophys.
Acta 1990, 1031, 1. (c) Tiddy, G. J. T. Phys. Rep. 1980, 57, 1.
2) For reviews of nanostructured materials synthesis using polymerizable
LLCs, see: (a) Gin, D. L.; Gu, W.; Pindzola, B. A.; Zhou, W.-J. Acc.
Chem. Res. 2001, 34, 973. (b) Miller, S. A.; Ding, J. H.; Gin, D. L. Curr.
Opin. Colloid Interfac. Sci. 1999, 4, 338. (c) O’Brien, D. F.; Armitage, B.;
Benedicto, A.; Bennett, D. E.; Lamparski, H. G.; Lee, Y.-S.; Srisiri, W.;
Sisson, T. M. Acc. Chem. Res. 1998, 31, 861. (d) Ringsdorf, H.; Schlarb,
B.; Venzmer, J. Angew. Chem., Int. Ed. Engl. 1988, 27, 113.
(
(3) For reviews on bicontinuous cubic LLC phases, see: (a) Fontell, K. Colloid
Polym. Sci. 1990, 268, 264. (b) Mariani, P.; Luzzati, V.; Delacroix, H. J.
Mol. Biol. 1988, 204, 165.
(4) Benedicto, A. D.; O’Brien, D. F. Macromolecules 1997, 30, 3395.
2940
9
J. AM. CHEM. SOC. 2003, 125, 2940-2949
10.1021/ja0208106 CCC: $25.00 © 2003 American Chemical Society

Assemblies via Polymerizable Gemini Amphiphiles
A R T I C L E S
Figure 1. Schematic representation of an ideal progression of LLC phases as a function of system composition and mean curvature. The various bicontinuous
cubic (Q) phases exist between the illustrated LLC phases and micellar structures.
Figure 2. Schematic representations of three QII phases, showing the relative locations of the hydrocarbon (shaded regions) and aqueous domains (open
white regions). The arrangement of the amphiphiles in these QII assemblies is illustrated in the inset. Partially reproduced from ref 4.
2
assemblies.12-15 These LLC monomers are all polymerizable
single-tail surfactants, which require added cross-linkers to form
robust networks, as opposed to soluble linear polysoaps. Even
more rare are LLC systems that afford cross-linked Q phases
with retention of phase structure. Early work by Anderson and
Str o¨ m showed that replicas of QII networks with Pn3m sym-
microstructure upon cross-linking. These nanostructured ma-
terials have been successfully used in applications ranging from
5-9
templated nanocomposite synthesis to heterogeneous catalysis
(
i.e., organic molecular sieve analogues).10,11 However, some
LLC phases have remained more elusive with respect to in-
phase polymerization and, thus, far less developed with respect
to potential applications. Only four amphiphilic monomer
systems have been reported that afford polymerized HI
2
24
metry (Type II Q (Pn3m)) (see Figure 2) can be formed via
polymerization of conventional monomers (e.g., methyl meth-
acrylate) in the organic regions of a QII phase formed by
(
5) Smith, R. C.; Fischer, W. M.; Gin, D. L. J. Am. Chem. Soc. 1997, 119,
16
nonpolymerizable surfactants. However, only one report of
4
092.
(
(
(
6) Gray, D. H.; Hu, S.; Juang, E.; Gin. D. L. AdV. Mater. 1997, 9, 731.
7) Gray, D. H.; Gin, D. L. Chem. Mater. 1998, 10, 1827.
8) Sellinger, A.; Weiss, P. M.; Nguyen, A.; Lu, Y.; Assink, R. A.; Gong, W.;
Brinker, C. J. Nature 1998, 394, 256
9) Ding, J. H.; Gin, D. L. Chem. Mater. 2000, 12, 22.
10) Miller, S. A.; Kim, E.; Gray, D. H.; Gin, D. L. Angew. Chem., Int. Ed.
the successful cross-linking of reactive amphiphiles in a Q phase
(12) Herz, J.; Reiss-Husson, F.; Rempp, P.; Luzzati, V. J. Polym. Sci., Part C
1963, 4, 1275.
(13) Shibasaki, Y.; Fukuda, K. Colloids Surf. 1992, 67, 195.
(14) McGrath, K. M. Colloid Polym. Sci. 1996, 274, 399.
(15) Pindzola, B. A.; Hoag, B. P.; Gin, D. L. J. Am. Chem. Soc. 2001, 123,
4617.
(
(
1
999, 38, 3021.
(
11) Gu, W.; Zhou, W.-J.; Gin, D. L. Chem. Mater. 2001, 13, 1949.
J. AM. CHEM. SOC.
9
VOL. 125, NO. 10, 2003 2941

A R T I C L E S
Pindzola et al.
has been published. O’Brien, Gruner, and co-workers showed
that a mixture of two polymerizable dienoyl derivatives of
naturally occurring phospholipids can be radically polymerized
at elevated temperatures (60 °C) to afford a cross-linked Type
and packing preferences of the constituent amphiphiles (i.e., a
surfactant packing parameter) to extrapolate the geometry of
2
3
the preferred LLC phase. Depending on the situation, one
approach is sometimes better for rationalizing the behavior of
certain LLC systems than the other. For example, changes in
LLC phases with temperature and composition can be better
rationalized by considering the interfacial energy and curvature
2
24
17
II Q (Pn3m) phase. Unfortunately, neither amphiphile by
itself affords this QII phase, and the hydrated amphiphile mixture
does not form the phase at ambient temperature.1 More recently,
O’Brien and co-workers designed a mixture of polymerizable
lipids that form a Type II Q2 (Ia3d) phase (Figure 2); however,
only soluble non-cross-linked polymers were formed by this
system.18
7
2
2
of the entire system, whereas the “shape” of the amphiphile
apparently plays the more crucial role in determining mesophase
30
23
geometry at fixed temperature and composition. Amphiphiles
that adopt the H phase have generally been rationalized by the
I
There is a great deal of interest in designing new LLC
monomer systems that efficiently form cross-linked HI and Q
phases because such stabilized assemblies offer new opportuni-
ties for applications as a result of their unique architectures.
For example, cross-linking of the HI phase would afford ordered
arrays of polymer nanofibrils that could be useful as one-
dimensional scaffolds for anisotropic nanocomposites, or as
shape-based approach, which is represented mathematically by
the “critical packing parameter,” q ) V/a l , where V is the
0 c
effective hydrocarbon chain volume a is the polar headgroup
0
2
3
area; and l . is alkyl chain length of the surfactant. According
c
to this model, single-tailed amphiphiles with a “truncated cone”
shape (q ≈ 0.5) tend to pack into cylindrical micelles to form
the H phase, whereas twin-tailed amphiphiles with an overall
I
novel mineralization platforms2 in much the same way lipid
a,15
cylindrical shape (q ≈ 1) tend to pack into L phases with no
microtubules have been employed.1
9-21
Cross-linked Q as-
23
curvature, etc. Thus, H monomers can be designed based on
I
semblies have been proposed for applications ranging from
membrane separations to bioencapsulation.2 Compared to HII
phases, cross-linked Q assemblies have even greater potential
as nanoporous organic catalysts and membrane materials
because they more closely mimic the interconnected channel
structure of zeolites.2a,c Unlike the uniaxial HII phase, Q
materials would not require macroscopic alignment to facilitate
substrate entry and transport because of their interconnected
channel structures.2 Herein, we report a new family of
intrinsically cross-linkable LLCs (1) based on gemini surfactants
that provide convenient access to both cross-linked HI and Q
phases. These monomers are readily synthesized from nonbio-
logical starting materials, and a number of homologues in this
series exhibit HI and QI phases under mild conditions in water
monomer shape and common structural elements. However, this
shape-based approach appears to break down when it comes to
understanding the Q phases which are considered curvature
a,c
1
a
“saddle points”. It turns out that a large number of single-
tailed amphiphiles and two-tailed phospholipids (q ≈ 1) with
diverse structures can form Q phases, usually at elevated
3
a
temperatures. Thus, the rational design of polymerizable
amphiphiles that form Q phases based on similar structural
motifs and packing principles is not very straightforward. The
ability of amphiphiles to form Q phases is better rationalized
in terms of balancing curvature and hydrocarbon chain stretching
energies at the hydrophilic/hydrophobic interface;1 however,
this approach makes molecular design difficult.
a,c
a,b
Recently, our research group designed a single-tailed LLC
monomer, tetradeca-11,13-dienyl-trimethylphosphonium bro-
mide (2), which can be photopolymerized in the HI phase with
(some even at ambient temperature). In addition, these mono-
mers can be photocross-linked in both the HI and QI phases
with retention of phase microstructure, without the need for
added cross-linkers or comonomers.
15
and without added divinylbenzene (DVB) as cross-linker.
Compound 2 is a polymerizable phosphonium analogue of
alkyltrimethylammonium bromide surfactants which have a
Results and Discussions
“
truncated cone” shape and are known to form the HI phase as
(a)LLC Monomer Design and Synthesis. For all practical
24-26
well as a QI phase with Ia3d symmetry.
Although this
purposes, the design of LLC monomers that adopt the HI phase
and ones that adopt the Q phases utilize two different sets of
design considerations. The ability of a hydrated amphiphile
system to preferentially form a particular LLC mesophase has
been explained in two different ways: The first is a global
approach which considers the interfacial energetics, tension, and
intrinsic curvature of the amphiphile/water mixture as a col-
phosphonium diene monomer also forms a sizable cubic phase
in water at temperatures above 35 °C; attempts to cross-link
this Q phase proved to be very difficult and were only partially
27
successful. The reason for this difficulty was that the addition
of DVB reduced the stable cubic phase regime to a very small
28
region in the ternary phase diagram at ambient temperature.
To overcome these problems, an intrinsically cross-linkable
derivative of tetradeca-11,13-dienyl-trimethylphosphonium bro-
mide was desired that would provide assess to HI and Q phases
without the need for added comonomers.
1
a,1b,22
lective ensemble.
The second approach examines the
system on the microscopic level in terms of the molecular shape
(
16) Anderson, D. M.; Str o¨ m, P. In Polymer Association Structures. Microel-
musions and Liquid Crystals; El-Nokaly, M. A., Ed.; ACS Symposium
Series 384; American Chemical Society: Washington, DC, 1989; Chapter
The synthesis of an intrinsically cross-linkable monomer from
a monofunctional monomer can simply be accomplished by the
1
3.
(17) Lee, Y.-S.; Yang, J.-Z.; Sisson, T. M.; Frankel, D. A.; Gleeson, J. T.; Aksay,
E.; Keller, S. L.; Gruner, S. M.; O’Brien, D. F. J. Am. Chem. Soc. 1995,
1
17, 5573.
(23) Israelachvili, J. N. Intermolecular and Surface Forces with Applications
to Colloidal and Biological Systems; Academic: London, 1985; pp 249-
257.
(24) Pindzola, B. A.; Gin, D. L. Langmuir 2000, 16, 6750.
(25) McGrath, K. M. Langmuir 1995, 11, 1835.
(26) Auvray, X.; Petipas, C.; Anthore, R.; Rico, I.; Lattes, A. J. Phys. Chem.
1989, 93, 7458.
(27) Pindzola, B. A., Ph.D. Thesis, University of California at Berkeley, 2001.
(28) See Supporting Information for ref 15.
(18) Srisiri, W.; Benedicto, A.; O′Brien D. F.; Trouard; T. P. Langmuir 1998,
1
4, 1921.
(19) Schnur, J. M.; Price, R.; Schoen, P.; Yager, P.; Calvert, J. M.; Georger, J.;
Singh, A. Thin Solid Films 1987, 152, 181.
(20) Chappell, J. S.; Yager, P. J. Mater. Sci. Lett. 1992, 11, 633.
(21) Archibald, D. D.; Mann, S. Nature, 1993, 364, 430.
(22) (a) Gruner, S. M. J. Chem. Phys. 1989, 93, 7562. (b) Tate, M. W.; Gruner,
S. M. Biochemistry 1989, 28, 4245.
2942 J. AM. CHEM. SOC.
9
VOL. 125, NO. 10, 2003

Assemblies via Polymerizable Gemini Amphiphiles
A R T I C L E S
Scheme 1
Figure 3. Schematic representations of three general approaches to making
an intrinsically cross-linkable LLC monomer from a monofunctional LLC
monomer.
addition of a second chain-addition polymerizable moiety to
the molecule. In our case, this is complicated by the need to
maintain an appropriate molecular geometry for the desired LLC
2
3
phase behavior. Consequently, a second polymerizable tail
cannot be simply attached to the amphiphile because this
modification would drastically change the overall molecular
shape. Figure 3 shows several alternative approaches for creating
a cross-linkable analogue of 2 that should retain the desired
LLC architecture on polymerization. The first approach (Figure
3
a) is to employ a polymerizable counterion such as a (meth)-
2
9
acrylate or 4-vinylbenzoate, following the work of Ringsdorf
and Kline.30 This method has the benefit of being a relatively
simple and minor modification to the existing amphiphile;
however, that the connection between the polymerizable groups
in each monomer would no longer be a covalent bond but rather
a weaker, ionic interaction. This approach is currently being
explored in our group.31 A second approach (Figure 3b) is to
append a polymerizable moiety to the headgroup. This route
would maintain the covalent connection between the polymer-
izable groups, but it would also be very synthetically challenging
and may detrimentally alter the overall molecular shape. The
final potential solution to this problem (Figure 3c) is to make
a polymerizable dimeric or “gemini” version of 2. The term
(
dimethylphosphino)alkanes with two equivalents of ω-bro-
15
moalkyl-1,3-diene in 2-propanol at elevated temperature gave
the desired products, which were recrystallized from ethyl
acetate. A series of nine homologues of 1 with three different
tail lengths (x ) 8, 10, and 14) and three different headgroup
spacer lengths (y ) 2, 6, and 8) were synthesized in order to
examine the effect of these two parameters on LLC behavior
and in-phase polymerization. It should be noted that these
polymerizable gemini surfactants are not stable in air at room
temperature for extended periods of time, so they were stored
at 0 °C under a nitrogen atmosphere.
“gemini surfactant” was originally coined by Menger to describe
a relatively new class of amphiphiles composed of two classical
single-tail amphiphiles connected covalently at the headgroup
(b)LLC Phase Behavior and Characterization. Once the
polymerizable gemini amphiphiles were synthesized, their two-
component phase diagrams with water were mapped out using
polarized light microscopy (PLM) as a function of temperature.
The identities of the observed LLC phases in each case were
assigned as follows, based on their PLM optical textures, low-
angle powder X-ray diffraction (XRD) patterns, and relative
positions on the phase diagram: The LR phase, which is
generally considered the midpoint of an ideal LLC phase
progression, is identified by a bright mosaic-type optical texture
and an XRD profile in which the observed d spacings proceed
in the ratio 1:1/2:1/3:1/4..., (i.e., the d100, d200, d300, d400,...
32
by a spacer. Gemini surfactants of the general structure CH3-
+
+
(
CH2)nN (CH3)2(CH2)mN (CH3)2(CH2)nCH3, with a flexible
spacer of moderate length (m ) 6-10), have demonstrated LLC
behavior similar to that of the monomeric amphiphiles of the
same tail length (e.g., spherical micelles, cylindrical micelles,
LR phases, and vesicles).33 By synthesizing a dimeric version
of our single-tail phosphonium monomer, it was hoped that an
intrinsically cross-linkable LLC exhibiting HI and Q phases
would be accessible. To our knowledge, polymerizable gemini
amphiphiles are unprecedented, and their LLC and polymeri-
zation behavior are completely unknown.
The general synthesis of gemini surfactants (1a-i) based on
polymerizable dienyltrimethylphosphonium bromide salts is
depicted in Scheme 1. 1,2-bis(Dimethylphosphino)ethane is a
commercially available reagent, but the other two bis(dimeth-
ylphosphino)alkanes had to be prepared from the reaction of
two equivalents of sodium dimethylphosphide (generated in situ
from sodium and tetramethyldiphosphine) with either 1,6-
dibromohexane or 1,8-dibromooctane in an adaptation of a
procedure developed by Kordosky.34 Reaction of the bis-
1
diffraction planes). The H phases exhibit a bright striated optical
texture, and their XRD profiles exhibit a d spacing pattern that
proceeds in the ratio: 1:1/x3:1/x4:1/x7... (i.e., d100, d110, d200,
1
d210,...). An H phase appearing on the water-rich side of the
LR phase on the phase diagram is generally assumed to be an
HI phase. Conversely, an H phase that appears on the water-
poor side of the LR phase is regarded as an HII phase.1 In the
absence of observable secondary or higher order XRD reflec-
tions, LR and H phases can be distinguished via their optical
textures and their location on the phase diagram with respect
a,b
1
(
29) (a) Ringsdorf, H.; Schlarb, B.; Tyminski, P. N.; O’Brien, D. F. Macro-
to other phases (see Figure 1). In contrast, Q phases are all
molecules 1988, 21, 671. (b) Ringsdorf, H.; Schlarb, B. Makromol. Chem.
optically isotropic by PLM but exhibit an XRD pattern with
symmetry-allowed d spacings that ideally proceed in the
ratio: 1:1/x2:1/x3:1/x4:1/x5:1/x6:1/x8:1/x9:1/x10... (cor-
1
988, 189, 299.
(
30) Kline, S. R. Langmuir 1999, 15, 2726.
(
31) Markevitch, D. M.; Pindzola, B. A.; Gin, D. L. Polym. Prepr., Am. Chem.
Soc. DiV. Polym. Chem. 2001, 42(1), 537.
(
32) Menger, F. M.; Littau, C. A. J. Am. Chem. Soc. 1991, 113, 1451.
33) Buhler, E.; Mendes, E.; Boltenhagen, P.; Munch, J. P.; Zana, R.; Candau,
S. J. Langmuir 1997, 13, 3096.
(
(34) Kordosky, G.; Cook, B. R.; Cloyd, J. J.; Meek, D. W. Inorg. Synth. 1973,
14, 14.
J. AM. CHEM. SOC.
9
VOL. 125, NO. 10, 2003 2943

A R T I C L E S
Pindzola et al.
Figure 4. Representative PLM optical textures of the LLC phases of the cross-linkable gemini surfactants: (a) an unidentified LLC phase (1e at 39 wt %
in water, 24 °C); (b) the HI phase (1e at 39 wt % in water, 50 °C); (c) the QI phase (1i at 85 wt % in water, 24 °C); and (d) the LR phase (1e at 90 wt %
in water, 95 °C). All images were taken at 12.6 X magnification.
Figure 5. Representative XRD patterns for the LLC phases observed in the cross-linkable gemini surfactants: (a) Unidentified LLC phase (1e, 64 wt % in
water, 24 °C); (b) HI phase (1e, 75 wt % in water, 24 °C); (c) QI phase (1e, 85 wt % in water, 35 °C); and (d) LR phase (1e, 90 wt % in water, 62 °C). The
broad peak just above 18° in 2θ in (c) and (d) results from the Mylar windows on the programmable XRD capillary oven.
responding to the d100, d110, d111, d200, d210, d211, d220, d221 (or
d300), d310,... diffraction planes).1 It should be noted that there
are a number of different Q phase architectures possible,
and that systematic XRD absences in the XRD peaks result as
the cubic unit cell becomes more complex. The three-
dimensional structures of many Q phases have been elucidated
for each observed LLC phase for the homologues of 1 are shown
in Figures 4 and 5, respectively.
Using these systematic phase identification guidelines, it was
found that the three gemini monomers with the shortest tails
(1a-c, x ) 8) exhibit relatively simple phase behavior. The
phase diagram for 1a (y ) 8, x ) 2) (Figure 6a) shows only a
,3
3
b,35
from detailed electron microscopy and two-dimensional XRD
relatively small H phase region that extends from 67 to 85 wt
I
studies.3
,35-40
For convenience, the presence of Q phases with
% amphiphile at temperatures of up to 55 °C. Similarly, the
phase diagram for 1b (x ) 8, y ) 6) (Figure 6b) only shows an
3
P or I symmetry in polydomain small molecule amphiphile
and phase-separated block copolymer systems35-40 has generally
been identified on the basis of a black optical texture and a
powder XRD profile in which the 1/x6 and 1/x8... d spacings
H regime that extends from 68 to 85 wt % amphiphile and up
I
to a maximum temperature of 58 °C. Compound 1c (x ) 8, y
) 8) in contrast does not display any LLC behavior at all: it
only exhibits a fluid isotropic phase (I) below 95 wt %
amphiphile.
The polymerizable gemini surfactants with the midrange tail
length (1d-f, x ) 10) have a more complex phase behavior in
general. The phase diagram for 1d (x ) 10, y ) 2) exhibits
two distinct LLC phases (Figure 6c). At lower percentages of
amphiphile (55-80 wt %) and slightly elevated temperature
1
7,35
(
i.e., the d211 and d220 reflections) are at least present.
To
unequivocally distinguish between the various 3-D cubic phase
architectures, however,3b a sufficient number of higher order
XRD reflections must be observable.3
b,35,36
As in the case of
the H phases, Q phases that appear on the water-rich side of
the LR phase of the phase diagram are considered QI phases,
and those that appear on the water-poor side are regarded as
1a,b
QII phases. Representative optical textures and XRD profiles
(30-75 °C), a rather broad Q phase is observed. Moving to
higher concentration of amphiphile (80-90 wt %) and higher
temperatures (55 to >100 °C), an LR phase is observed. Because
this Q phase appears on the water-rich side of the phase diagram
with respect to the LR phase, it should be classified as a QI
phase with net positive mean curvature. The phase diagram for
1e (x ) 10, y ) 6) (Figure 6d) exhibits four different LLC
phases. An HI phase extends from 48 to 81 wt % amphiphile at
ambient temperature to more than 100 °C, depending on
(
35) Hajduk, D. A.; Harper, P. E.; Gruner, S. M.; Honeker, C. C.; Kim, G.;
Thomas, E. L.; Fetters, L. J. Macromolecules 1994, 27, 4063.
36) Hajduk, D. A.; Harper, P. E.; Gruner, S. M.; Honeker, C. C.; Kim, G.;
Thomas, E. L.; Fetters, L. J. Macromolecules 1995, 28, 2570.
37) Sakurai, S. Trends Polym. Sci. 1997, 5, 210.
(
(
(
38) Lipic, P. M.; Bates, F. S.; Hillmyer, M. A. J. Am. Chem. Soc. 1998, 120,
8
963.
(39) Lee, M.; Cho, B.-K.; Kim, H.; Yoon, J.-Y.; Zin, W.-C. J. Am. Chem. Soc.
1
998, 120, 9168.
(40) Shefelbine, T. A.; Vigild, M. E.; Matsen, M. W.; Hajduk, D. A.; Hillmyer,
M. A.; Cussler, E. L.; Bates, F. S. J. Am. Chem. Soc. 1999, 121, 8457.
2944 J. AM. CHEM. SOC.
9
VOL. 125, NO. 10, 2003

Assemblies via Polymerizable Gemini Amphiphiles
A R T I C L E S
Figure 6. Phase diagrams for the eight polymerizable gemini surfactants that displayed LLC behavior in water: (a) compound 1a, (b) compound 1b, (c)
compound 1d, (d) compound 1e, (e) compound 1f, (g) compound 1h, and (h) compound 1i. The phase diagram for compound 1c is not included because
1
c did not display any LLC behavior. (LC ) unidentified LLC phase; I ) fluid isotropic phase; X ) crystalline phase. Vertical hashing indicates areas where
two or more phases coexist.) These phase diagrams were mapped out at U. C. Berkeley (near sea level, >50% relative humidity. They will be slightly
different at higher altitudes.
amphiphile concentration. A narrow QI phase region extends
from 82 to 88 wt % amphiphile at 25 °C to 87 °C. This QI
phase is followed by an LR phase that extends from 87 to 93
wt % amphiphile at temperatures of 75 °C and above. Finally,
an as yet unidentified LC phase exists from 45 to 73 wt %
amphiphile at ambient temperature up to a maximum of 60 °C.
From the PLM micrographs, it can be seen that the optical
texture for this unidentified LLC phase (Figure 4a) is similar
to that observed for the HI phase surrounding it on the higher
concentration amphiphile side. However, the phase in question
has distinct and reversible differences which do not correspond
to any of the generally accepted LLC phases. Unfortunately,
powder XRD was not useful in identifying this phase because
it only exhibits a single, strong reflection. One of two possibili-
domain and an LR domain. The HI phase of 1f spans 61 to 83
wt % amphiphile and exists up to 51 °C. The L region extends
R
from 95 to 98 wt % amphiphile and exists above 33 °C.
The three members in the homologous series with the longest
tails (1g-i, x ) 14) also display rich and diverse LLC phase
behavior. The phase diagram for compound 1g (x ) 14, y ) 2)
(Figure 6f) is the simplest of the three, displaying only an LR
phase from 49 to 93 wt % amphiphile and above ca. 55 °C.
Interestingly, this homologue is the only compound in the entire
nine compound series which has a Krafft point (i.e., an order-
disorder transition temperature for the alkyl chains) above room
temperature across the entire range of concentrations studied.
Presumably, the long alkyl tails and the close proximity of the
tethered headgroups (due to the short spacer length) favor more
ordered hydrocarbon tail packing than in the other molecules.
In contrast, compound 1h (x ) 14, y ) 6) (Figure 6g) displays
four different LLC phases: the same unidentified LLC phase
as in 1e, plus an HI, a QI, and an LR phase. The unidentified
mesophase extends from 52 to 74 wt % amphiphile and from
less than 25 °C to greater than 100 °C in temperature. From 74
to 78 wt % amphiphile and from 27 °C to greater than 100 °C,
the HI phase is observed. Following the HI phase and extending
to 83 wt % amphiphile from 46 °C to greater than 100 °C is a
QI phase. Finally, an LR phase exists from 83 to 88 wt %
amphiphile above 36 °C. Compound 1i (x ) 14, y ) 8) (Figure
ties for this mesophase seem most likely. It could be one of the
so-called “intermediate” phases,4
1,42
or simply a mixtures of
phases. “Intermediate” LLC phases generally appear between
41,42
the H and LR phases, frequently in place of Q phases.
These
optically anisotropic phases are proposed to have rectangular,
layered, rhombohedral, and tetragonal symmetries; and are
believed to have ribbon or mesh-like structures. The fact that
the unidentified LLC phase appears between the fluid isotropic
(I) phase and the HI phase, rather than between the HI and LR
phases, argues against it being one of the other better known
intermediate phases. It seems most likely that it is simply a
mixture of phases, but an uncommon intermediate phase cannot
be ruled out from the data available. Finally, compound 1f (x
6
h) displays the same four LLC phases seen in the previous
case. The unidentified LLC phase exists between 35 and 64 wt
amphiphile and from below ambient temperature to a
)
10, y ) 8) (Figure 6e) exhibits two LLC phases: an HI
%
(
41) Luzzati, V.; Mustacchi, H.; Skoulios, A.; Husson, F. Acta Crystallogr. 1960,
3, 660.
42) Hagslatt, H.; Soderman, O.; Jonsson, B. Liq. Cryst. 1994, 17, 157.
maximum temperature of 62 °C. The HI phase extends from 66
to 78 wt % amphiphile and from less than 25 °C to greater
1
(
J. AM. CHEM. SOC.
9
VOL. 125, NO. 10, 2003 2945

A R T I C L E S
Pindzola et al.
than 100 °C. The QI phase exists from 80 to 88 wt % amphiphile
and from about 30 °C to 90 °C. The LR phase exists at higher
temperatures and on the water-poor side of the QI phase,
extending from 80 to 98 wt % amphiphile and above 50 °C.
image the QI structures of 1e by transmission electron micros-
copy (TEM), and atomic force microscopy (AFM). Unfortu-
nately, the TEM images obtained with extensively stained and
embedded samples were of insufficient resolution to conclu-
43,44
sively show a cubic structure,
and the sample surfaces were
Certain trends in LLC phase behavior for the gemini
homologues can also be observed by considering the headgroup
spacer length (y) instead of the hydrocarbon tail length (x). It
is known that the LLC properties of gemini surfactants are
highly dependent on the length of the flexible spacer between
too rough to obtain high-resolution AFM images (see Supporting
Information). Consequently, it was not possible to unambigu-
ously identify which of the possible cubic architectures3b the
observed Type I Q phases might be. However, based on their
one-dimensional XRD profiles, their locations on the phase
diagrams, and the physical properties of the cross-linked
materials, it was possible to narrow down the possibilities.
According to Luzzati and co-workers, six geometrically
distinct Q phase architectures have been identified so far in small
molecule amphiphile systems, based on their detailed XRD data,
33
the polar headgroups. The reason for this strong dependence
33
is based on shape-based principles and packing parameter: The
headgroup area of a gemini surfactant (ao) has been found to
increase in a nonmonotonic manner with spacer length. It
increases rapidly for short spacers (<10 methylene units),
reaches a maximum for medium spacers (10-12 methylene
units), and then decreases for longer spacers (g14 methylene
230
symmetry, and common structural elements (i.e., Q (Ia3d),
2
24
229
212
227
223
3
3
Q
(
(Pn3m), Q (Im3m), Q (P4332), Q (Fd3m), and Q
units). A closer examination of the phase diagrams of 1a-i
presented in Figure 6 shows that the polymerizable gemini
surfactants exhibit the same sensitivity to headgroup spacer
length. On going across each row of phase diagrams, the number
and size of stable LLC domains exhibited by gemini monomers
with the same tail length (x) generally increases as y goes from
3
b
Pm3n); see Supporting Information). It turns out that cubic
mesophases consisting entirely of closed normal or reverse
micelles are quite rare, and their structures have not been
1
a,3a
substantiated.
Of the six Q phase architectures so far
3
b
identified in amphiphile systems, the combined data obtained
2
30
224
for the gemini QI phases are consistent with a Q (Ia3d), Q
2
to 6, but then decreases when y reaches 8. These gemini
2
29
212
(Pn3m), Q
(Im3m), or Q
(P4332) phase of Type I
monomers are in the “short” spacer regime (<10 methylenes),
so these changes reflect the fact that ao increases rapidly with
increasing spacer length. However, it is important to observe
the interplay between spacer length and tail length in determin-
ing the overall ability of these gemini systems to form certain
LLC phases. At short tail lengths (x ) 8), the gemini monomers
can only form the HI phase with varying spacer length,
indicating that ao is sufficiently large in the y ) 2-8 range
that the surfactants cannot adopt the general cylindrical shape
configuration. The fact that 1/x6 and 1/x8 peaks are exhibited
by all the QI phases of the gemini monomers indicates that they
2
27
227
cannot have the Q (Fd3m) architecture. The Q (Fd3m)
phase has F symmetry and does not have a symmetry-allowed
1
3
b,35
223
/x6 (i.e., d211) reflection.
The Q (Pm3n) phase can also
be ruled out with a high degree of confidence, based on our
polymerization results described in the proceeding section. A
2
23
Type I Q (Pm3n) phase is believed to be either a system
composed entirely of closed micelles with the hydrophobic tails
located in the micelle interiors, or at most a system composed
of independent micelles and a single channel network (see the
(q ≈ 1) needed to form bilayer type L (and Q) phases. In fact,
as y reaches 8 for this tail length, it appears that ao is so large
that even the truncated conical shape needed to form HI phases
is disrupted, and no LLC phases are observed. However, as the
tail length increases (x ) 10 and 14) (and presumably so does
the effective hydrocarbon tail volume), the resulting longer
gemini surfactants are better able to adopt an overall cylindrical
shape to form L and Q phases, even with increasing spacer
length. As can be seen in Figure 6, the gemini homologues that
are best able to form L and Q type phases are the ones with
3b
Supporting Information). If this system were to be polymerized
at the tail ends, there would be minimal inter-micelle connectiv-
ity and perhaps only a single polymerized channel system,
resulting in a polymer with very poor mechanical and thermal
integrity. As described in the polymerization results section, all
of the cross-linked QI phases of the gemini monomers exhibit
excellent mechanical integrity and thermal stability. Thus, the
QI phases of the gemini monomers can be inferred to be one of
the remaining four Q architectures with P or I symmetry which
have a more extensive interpenetrating channel system. A Type
“
matched” spacer and tail lengths (i.e., medium spacer with
medium length tails, or longer spacer with longer tails).
Qualitatively at least, it appears that for these gemini surfactants,
certain aspects of the shaped-based LLC model apply. That is,
increases in ao via spacer effects can be offset by increasing
tail length and hydrocarbon chain volume so that certain shapes
and values of the packing parameter q can be maintained.
2
30
I Q (Ia3d) phase appears to be the most likely candidate for
two reasons: First, Q phases with Ia3d symmetry are the most
common Q phases observed between the LR and H phases.3a
Second, the monomeric alkyltrimethylammonium salts from
which the gemini amphiphiles are derived exhibit a Q phase
Unfortunately, only two powder XRD peaks (1/x6, 1/x8)
were generally observed for the observed QI phases of the 1
homologues, indicating that they do not possess a great deal of
long range order. In an attempt to obtain more data on the
architecture of the QI phases, the samples were also analyzed
by two-dimensional X-ray scattering with a higher intensity
X-ray source. Unfortunately, the QI samples were all polydomain
in nature and could not be easily aligned resulting in diffraction
rings rather than spots in the 2-D XRD images (see Supporting
Information). In addition, higher order reflections were not
observed under these conditions. Attempts were also made to
24,25
with Ia3d symmetry on the water-rich side of the LR phase.
Regardless of which may be the actual case, the four possible
QI phases for the polymerizable gemini surfactants are all
bicontinuous, with aqueous and organic networks that permeate
the sample in three dimensions. They would still make intriguing
candidates for nanoporous organic transport and catalysis media
if they can be stabilized by cross-linking.
(43) Because of the poor quality of the TEM images, electron diffraction studies
I
were not performed on the gemini Q phases.
(
44) To our knowledge, successful STM and AFM imaging of cross-linked LLC
phases has not reported.
2946 J. AM. CHEM. SOC.
9
VOL. 125, NO. 10, 2003

Assemblies via Polymerizable Gemini Amphiphiles
c)In Situ Cross-Linking of the LLC Phases. Once the
A R T I C L E S
(
general LLC phase behavior of the polymerizable gemini
surfactants was characterized, their polymerization behavior in
the various LLC phases was subsequently examined. Three
issues with respect to the in-phase polymerization of these
monomers were investigated. The first issue was whether these
reactive gemini LLCs can be cross-linked without significant
disruption of the phase structures. This was determined by
comparing the PLM optical textures and XRD patterns of the
samples before and after radical photopolymerization. The
second issue was determining the degree of polymerization (or
at least the extent of 1,3-diene conversion) in these systems.
Typically the degree of polymerization in cross-linkable 1,3-
diene-based LC monomers can be estimated using FT-IR
spectroscopy by monitoring the decrease in intensity of the Cd
Figure 7. Representative XRD profiles and PLM optical textures of the
-
1 15,45
H phases of the gemini monomers before and after photopolymerization:
C stretching bands at 1602 and 1651 cm .
In the case of
I
(a) XRD pattern of 1e (75 wt % in water, 24 °C) before photopolymerization;
(b) XRD pattern of 1e (75 wt % in water, 24 °C) after photopolymerization;
(c) PLM micrograph of 1e (75 wt % in water, 24 °C) before photopolym-
LLC systems, D2O must be used in place of H2O to eliminate
-
1 15
the strong overlapping H-O-H band at 1600 cm . Unfor-
tunately, FT-IR spectroscopy could not be employed because
the baselines in the FT-IR spectra of the gemini LLC mixtures
did not remain level or consistent. The degree of 1,3-diene
polymerization was estimated instead using UV-vis spectros-
erization; and (d) PLM micrograph of 1e (75 wt % in water, 24 °C) after
photopolymerization.
Table 1. Summary of Data for Molecules Polymerized in the HI
Phasea
copy, because polymerization of a 1,3-diene moiety leads to
d100
before
irradiation
(Å)
d100
after
irradiation
(Å)
polymerized
phase
thermal
extent
of
conversion
(%)
loss of conjugation.46 The degree of conversion can be
determined by comparing the intensity of the 1,3-diene absorp-
tion band (λmax ) 226 nm) in the samples before and after
irradiation, similar to protocols used to study the polymerization
compd
stability (°C)
1a
1b
1e
31.3
29.8
34.0
35.9
42.5
41.7
31.0
29.4
33.6
33.1
41.4
42.3
> 300
> 300
> 300
> 300
> 300
> 300
60
71
40
44
N/A
23
4
6
of the related sorbate moiety in amphiphile/water mixtures.
With very thin films of the room-temperature LLC phases on
quartz or CaF2 plates, it was possible to obtain reproducible
UV-vis spectra of the monomers and polymers in those
phases.47 The third issue was the stability and robustness of the
cross-linked LLC phases. This was assessed by heating the LLC
networks in air and observing the temperature at which the LLC
order was disrupted as indicated by changes in the PLM optical
textures.
1f
1
1
h
i
a Extent of conversion via UV-vis is unavailable for compound 1h in
the HI phase because that phase does not exist at room temperature.
From the PLM micrographs in Figure 7, it is obvious that
the HI phase is being maintained upon polymerization since there
is little change in the optical textures before and after irradiation.
After polymerization, variable temperature PLM showed that
the HI phase is stable up to at least 300 °C, above which severe
degradation of the material itself begins. In contrast, the
unpolymerized phases either clear below, or begin to rapidly
lose water around 100 °C. As can be seen in Figure 7 and Table
It was found that the LLC mixtures of the gemini monomers
can be photopolymerized by 365 nm light either with 1-2 wt
%
2,2-dimethoxyphenylacetophenone as a radical photoinitiator,
or without any added photoinitiator at all. No appreciable
difference was observed in the extent of conversion achieved
between samples that included 2,2-dimethoxyphenylacetophe-
none and those that did not. Thus, all photopolymerization
studies were performed without added photoinitiator for sim-
plicity of sample preparation.
1
, XRD shows that there is slight disordering of the phases upon
cross-linking based on the loss in peak intensity observed in
all cases. However, the overall symmetry of the phase is
retained. Additionally, the dimensions of the HI phases slightly
decrease (at most by 3 Å) upon polymerization in almost every
As described in the previous section, six of the nine
homologues of 1 display an HI phase. In all six cases, radical
photopolymerization of the monomers in the HI phase proceeded
with retention of the phase architecture, as seen by PLM and
XRD. A representative view of polymerization of monomers
in the HI phase is shown in Figure 7. A summary of the positions
of the primary d100 XRD peak before and after irradiation, the
thermal stability of the phase after irradiation and the extent of
conversion for the six molecules polymerized in the HI phase
are presented in Table 1.
48
case, as would be expected upon cross-linking. Finally, a
general decrease is seen in the extent of conversion with
increasing tail length. Unfortunately, not enough data is available
to see any trends with regard to the length of the spacer between
the two headgroups. Extents of polymerization in the HI phases
vary from 23 to 71% diene conversion by UV-vis spectroscopy.
However, even with the lowest observed extent of conversion
observed in our HI assemblies (23%), an extensively stabilized
cross-linked structure is still achieved. The gel point for chain
addition polymerization, Fc, is given by the equation: Fc ) ([A]
(
(
(
45) Hoag, B. P.; Gin, D. L. Macromolecules 2000, 33, 8549.
+
[B])/([B]Xw), where [A] is the concentration of monofunc-
46) Lamparski, H.; O’Brien, D. F. Macromolecules 1995, 28, 1786.
47) Unfortunately, it was not possible to accurately determine the extent of
polymerization for LLC phases at elevated temperatures because our UV-
Vis instrument does not have variable temperature capabilities.
(48) Odian, G. Principles of Polymerization, 3rd ed.; John Wiley & Sons: New
York, 1991.
J. AM. CHEM. SOC.
9
VOL. 125, NO. 10, 2003 2947

A R T I C L E S
Pindzola et al.
Table 2. Summary of Data for Molecules Polymerized in the LR
Phasea
d
100
d
100
polymerized
phase
thermal
before
irradiation
(Å)
after
irradiation
(Å)
compd
stability (°C)
1
1
1
1
1
1
d
e
f
g
h
i
31.4
29.2
31.2
44.1
34.5
33.1
31.9
30.3
31.8
45.3
39.7
38.9
> 300
> 300
N/A
> 175
> 300
> 300
a
Extent of conversion data via UV-vis are unavailable for compounds
in the LR phase because they do not exist at room temperature.
Figure 9. Representative XRD profiles and PLM optical textures of the
QI phases of the gemini monomers before and after photopolymerization:
(a) XRD pattern of 1e (85 wt % in water, 35 °C) before photopolymerization;
(b) XRD pattern of 1e (85 wt % in water, 24 °C) after photopolymerization;
(c) PLM micrograph of 1e (85 wt % in water, 39 °C) before photopolym-
erization; and (d) PLM micrograph of 1e (85 wt % in water, 24 °C) after
photopolymerization.
supposition, since longer homologues would be expected to have
5
0
a greater tendency for tail interdigitation. A comparison of
Tables 1 and 2 shows that this effect appears to be more
pronounced in bilayer type assemblies (e.g., the LR and Q
phases) than in “monolayer” H phases in which there is more
interfacial curvature. With respect to the stability of the cross-
linked LR phases, the thermal behavior of the irradiated phases
is greatly improved over the nonirradiated phases. The optical
textures of the polymerized samples are generally stable to
greater than 300 °C. In the unpolymerized samples, water begins
to boil off around 100 °C. Accurate extent of conversion
information is unavailable for the LR phases because these
phases are only accessible at elevated temperatures.
Only four of the nine homologues (1d, 1e, 1f, and 1i) form
QI phases but each of these QI phases can be photocross-linked
with retention of phase architecture. Representative XRD
patterns and PLM micrographs taken before and after polym-
erization of these QI assemblies are presented in Figure 9. Table
Figure 8. Representative XRD profiles and optical textures of the LR phases
of the gemini monomers before and after photopolymerization: (a) XRD
pattern of 1e (90 wt % in water, 62 °C) before photopolymerization; (b)
XRD pattern of 1e (90 wt % in water, 24 °C) after photopolymerization;
(c) PLM micrograph of 1e (90 wt % in water, 50 °C) before photopolym-
erization; and (d) PLM micrograph of 1e (90 wt % in water, 24 °C) after
photopolymerization.
tional monomer double bonds, [B] is the concentration of cross-
linker double bonds, and Xw is the weight average degree of
4
8
polymerization. In our case, there is no monofunctional
monomer present so [A] ) 0, and the equation reduces to Fc )
1/Xw. Thus, to reach the gel point at 23% conversion, the weight
average degree of polymerization only needs to be greater than
five; and the higher the degree of polymerization, the less
conversion is needed to cross-link the entire system.
3
summarizes the data from the polymerization runs of the three
gemini monomers that form the more well-defined QI phases
that appear between the HI and LR phases. The PLM optical
textures remain essentially unchanged after polymerization.
However, in this case complete loss of phase structure cannot
be ruled out based on PLM data alone because the starting
optical textures are pseudo-isotropic. The XRD patterns, how-
ever, confirm that the QI phases are intact after photopolym-
erization in each case, albeit with very slightly decreased peak
intensities (Figure 9). Continuous long-term XRD analysis of
some of the cross-linked QI phases (i.e., over 3-4 d) shows
Six molecules in the series (1d-i) exhibit an LR phase, and
the data from their polymerization in this phase are summarized
in Table 2. Again, photopolymerization proceeded with retention
of the phase, and representative before and after XRD and PLM
profiles are shown in Figure 8. The PLM micrographs are
essentially unchanged after irradiation, and the intensities of
the observed XRD peaks are slightly reduced, similar to that in
the HI case. For the bilayer LR phases, it appears that a general
expansion of the unit cell occurs upon polymerization, which
is unusual because cross-linking generally results in a small
volume contraction.48 These results were reproduced several
times with different gemini homologues on a recalibrated XRD
(49) According to Bragg’s law, d spacing (in Å) is inversely proportional to
2
sin θ (in degrees). Consequently, small changes in 2θ result in progres-
sively larger changes in d spacing at lower and lower angles, so the error
in calculating d also gets progressively larger at lower 2θ angles. We are
able to correct for a great deal of this error in the 1.5-3 degree range by
instrument, so this phenomenon is real and not due to XRD
calibration error.49 This expansion can be rationalized by loss
using silver behenate (Kodak) as a low-angle diffraction standard (d
00
)
1
58 Å) and collecting the XRD data as a high-resolution data file, so that
the accuracy is within 1 Å in d spacing.
of some tail interdigitation within the lamellar bilayers if the
polymerization process causes the tail ends to shift up to join
together. The fact that this apparent increase in d spacing upon
polymerization is more pronounced in the gemini homologues
with the longest tails (1g-i, x ) 14) lends support to this
(50) Unfortunately, the degree of tail interdigitation cannot be determined
experimentally from the XRD data. Although d100 directly gives the bilayer
repeat distance, the bilayer spacing is the sum of an aqueous layer and the
hydrocarbon tail segment. The thickness of these two components cannot
be measured independently, so it is not possible to compare the actual
hydrocarbon layer distance against the calculated extended length of the
surfactants.
2948 J. AM. CHEM. SOC.
9
VOL. 125, NO. 10, 2003

Assemblies via Polymerizable Gemini Amphiphiles
A R T I C L E S
Table 3. Summary of Data for Three Gemini Monomers
Polymerized in the QI Phase
posites. With respect to the cross-linked QI systems, we are
exploring methods for fabricating thin films of these nanoporous
materials onto supports in order to examine their potential as
novel membrane media for gas and liquid nanofiltration. Future
work will focus on derivatizing the ionic headgroups in these
assemblies with reactive moieties for new heterogeneous
catalysis applications.
a
d
211
d
211
polymerized
phase
thermal
extent
of
conversion
(%)
before
irradiation
(Å)
after
irradiation
(Å)
compd
stability (°C)
1e
1h
1i
32.2
34.1
37.0
32.9
40.4
39.1
> 300
> 300
> 300
50
N/A
N/A
Conclusions
a
Extent of conversion data via UV-vis are unavailable for compounds
h and 1i in the QI phase because they do not exist at room temperature.
1
The synthesis and LLC phase behavior of a series of
intrinsically cross-linkable gemini LLCs has been described.
These novel polymerizable amphiphiles exhibit HI, QI, and LR
phases in water and can be intrinsically photocross-linked with
retention of phase architecture in each case. The QI phases
formed by the reactive amphiphiles in this series have P or I
space group symmetry from XRD analysis, and are most likely
2
30
224
229
212
either a Q (Ia3d), a Q (Pn3m), a Q (Im3m), or a Q
P4332) phase of Type I configuration, as inferred from their
(
locations on the phase diagram and the physical properties of
the cross-linked materials. The most likely candidate appears
2
30
to be a Type I Q (Ia3d) phase, based on the aforementioned
data and the phase behavior of monomeric amphiphiles of
similar structure. The extent of conversion achieved in the LLC
phases was determined by UV-vis spectrometry and found to
be more than sufficient to extensively cross-link and stabilize
the systems. The resulting cross-linked HI, LR, and QI assemblies
exhibit excellent thermal robustness in air.
Figure 10. XRD profile of a cross-linked QI phase of 1e showing higher
order diffraction peaks (collected over a continuous scanning period of 84
2
2
2
2
h). Reflection dhkl for a cubic phase has the relationship m ) h + k + l ,
such that dhkl peak position is assigned a value of (1/xm) in terms of d
4
0
17,35,36.
spacing ratio. The 1/x20 peak is often observed in Q phases.
The
smallest observed XRD peak at 11.1 Å matches with a cubic symmetry-
allowed 1/x43 peak, where h, k, l can be of the values 3, 5, and 3. (The
d-values for this polymerized QI sample are slightly different from those
presented in Figure 9 because this sample was prepared with slightly
different water content.)
Acknowledgment. This research was supported in part by
grants from the National Science Foundation (DMR-9642533),
and the Office of Naval Research (N00014-00-1-0271 and
N00014-02-1-0383). B.A.P. thanks Eli Lilly for a graduate
fellowship during the initial part of this work. D.L.G. thanks
the Alfred P. Sloan Foundation for a Research Fellowship. We
also thank Prof. J. Arnold for use of his UV-vis spectropho-
tometers; J. Mellott and Prof. D. Schwartz for preliminary AFM
studies; and Prof. E. Y.-X. Chen for his assistance in the TEM
imaging studies.
the presence of additional (albeit weak) higher-order d spacings
(Figure 10) that definitively confirm a Q phase with P or I space
group symmetry. As can be seen in Table 3, the QI phase unit
cells expand upon cross-linking for the mesogens with the
longest tail lengths (e.g., 1h, 1i), similar to that observed for
the LR phases. Again, this expansion upon cross-linking can be
rationalized by loss of some tail interdigitation within the Q
bilayers upon joining of the tail ends, and is more pronounced
for the longer tail surfactants. The extent of 1,3-diene polym-
erization in the room-temperature QI phase of 1e was found to
be ca. 50% by UV-vis spectroscopy, again well above the
expected gel point for these systems. All the resulting cross-
linked QI assemblies possess good mechanical integrity and can
be isolated as brittle solids or free-standing films that are
insoluble in water and conventional organic solvents. These
cross-linked LLC phases are also stable to at least 300 °C in
air, as determined by variable temperature PLM (and supported
by XRD).
Supporting Information Available: Detailed experimental
procedures for the synthesis, characterization, LLC phase
formation, and polymerization of the gemini monomers. A
complete listing of the compositions, temperatures, and XRD
data for the LLC phases identified in plotting the phase diagrams
of the polymerizable gemini surfactants. Preliminary 2-D XRD
and TEM images of the cross-linked QI phase of 1e. Schematic
representations of the six Q phase architectures observed so far
for small molecule amphiphiles (PDF). This material is available
free of charge via the Internet at http://pubs.acs.org.
We are currently exploring the use of these cross-linked HI
arrays are templates for forming anisotropic hybrid nanocom-
JA0208106
J. AM. CHEM. SOC.
9
VOL. 125, NO. 10, 2003 2949