Self-assembling properties of a series of homologous ester-diamides-from ribbons to nanotubes

Article abstract of DOI:10.1039/c3sm51369f

We have studied the aggregation properties of a series of compounds BHPB-n bearing two amide groups and an ester alkyl chain. The number (n) of carbon atoms in this alkyl chain has been varied from 5 to 16. These compounds form gels at low concentration i

Full text of DOI:10.1039/c3sm51369f

Soft Matter
View Article Online
View Journal
PAPER
Self-assembling properties of a series of homologous
ester-diamides – from ribbons to nanotubes†
Cite this: DOI: 10.1039/c3sm51369f
a
a
a
a
´
François-Xavier Simon, Thi Thanh Tam Nguyen, Nancy Dıaz, Marc Schmutz,
b
c
a
ad
*
*
´
´ ˆ
´
Bruno Deme, Jacques Jestin, Jerome Combet and Philippe J. Mesini
We have studied the aggregation properties of a series of compounds BHPB-n bearing two amide groups
and an ester alkyl chain. The number (n) of carbon atoms in this alkyl chain has been varied from 5 to 16.
These compounds form gels at low concentration in alkanes. The gel concentrations and the heats of
formation of the gels, measured by DSC, are divided into three different regimes: n # 7, 8 # n # 12 and
n $ 13. Freeze fracture TEM shows that all the compounds self-assemble into high aspect ratio objects,
i.e. flat ribbons for n # 7 and irregular helical ribbons for n $ 13. For 8 # n # 12, the compounds form
tubes (except BHPB-11) and were studied by SANS. The outer and inner diameter distributions of the
tubes were determined from the scattered intensities with a model of paired tubes. The diameters are
controlled by the length of the ester chain (n) and vary from 21 for BHPB-8 to 32 nm for BHPB-12
(outer diameters). The SANS measurements also provide an estimate of the ratio of paired tubes and
the proportion of non-tubular objects that eventually co-exist with these tubes. Variable temperature
FTIR and DSC proved that H-bonds between the amide groups are involved in all types of self-
assemblies. In addition, FTIR spectra of helical tapes and nanotubes exhibit a double ester nCO band
whereas flat tapes display a single band, which confirms the involvement of the ester in the self-assemblies.
Received 15th May 2013
Accepted 8th July 2013
DOI: 10.1039/c3sm51369f
www.rsc.org/softmatter
how an enantiomeric excess of chiral counterions control the
diameters of the self-assembled tubules formed from gemini
Introduction
surfactants.18 However they cannot explain the few examples of
non-chiral molecules forming nanotubes.19,20 These models also
rely on a phenomenological description of the elastic energy but
do not take into account the local packing or interactions
between the molecules and thus cannot identify the molecular
parameters that govern the formation of the tubes. Resolution of
the structures unveiling interactions at a molecular level has
been achieved only in a few cases, where the objects have been
successfully ordered.21,22 But generally these size-limited objects
give diffuse signals on selected area diffraction and powder
patterns in solution. As a result, one lacks predictability to
control the shape and the diameters of those self-assemblies.
Many efforts have been made to tune the size of self-assembled
nanotubes. It has been achieved by addition of salts in the case of
tubes formed from gemini surfactants.23 For other systems, the
molecular parameters controlling the diameters have been
successfully identied, e.g. the relative size of the polar heads in a
series of bolaphiles,24 the nature of the polar group in glycer-
olipids25 or the amino-acid residues in peptidic nanotubes.26
However, these examples are very rare and modication of a
given self-assembling compound most oen precludes the
formation of nanotubes.
Self-assembly is an effective tool to produce nano-objects with
well-dened shapes and controlled sizes. This is illustrated by the
formation of nanotubes: a few compounds can form such
structures simply by spontaneous self-association.1 The diame-
ters of the resulting tubes depend on the nature of the molecule
and range from 10 to 500 nm. Due to their unique shape and
their high aspect ratio, these nanotubes nd applications for
instance as scaffolds for cell growth,2 as supports for 2-D crys-
tallization of proteins3–5 or as templates to form silica nano-
tubes6–8 and mesopores in organic matrices.9 They also have
triggered much theoretical research by physicists to understand
the mechanism of their self-assembly.10–17 These studies provide
models where the chirality of the constitutive molecules induces
an intrinsic bending of the self-assembled layers and yields both
helical and tubular shapes. These models explain for instance
aCentre National de la Recherche Scientique, Institut Charles Sadron, 23 rue du
Loess, BP 84047, F-67034, Strasbourg, France. E-mail: mesini@ics-cnrs.unistra.fr;
combet@ics-cnrs.unistra.fr; Fax: +33 3 88 41 40 99; Tel: +33 3 88 41 40 70
bInstitut laue Langevin, 6 rue Horowitz, BP 156, F-38042, Grenoble, France
c
´
ˆ
Laboratoire Leon Brillouin, Bat. 563 CEA Saclay, 91191 Gif sur Yvette, France
d
´
International Center for Frontier Research in Chemistry, 8 allee Gaspard Monge, F-
67083, Strasbourg, France
We have shown previously that 3,5-bis-(5-hexylcarbamoyl-
pentyloxy)-benzoic acid decyl ester (BHPB-10, Fig. 1, n ¼ 10) was
able to self-associate to form nanotubes in alkanes.27
† Electronic supplementary information (ESI) available: Characterization data of
all synthesized compounds; best ts for all the compounds in log–log, Kratky
and Porod representation. See DOI: 10.1039/c3sm51369f
This journal is ª The Royal Society of Chemistry 2013
Soft Matter

View Article Online
Soft Matter
Paper
organic phases were dried (Na2SO4), ltered and evaporated
under reduced pressure. The residue was chromatographed
(SiO2, MeOH/CH2Cl2: 5/95 eluent) to afford pure DHB-5 as a
1
colorless oil (5.74 g, yield 79%). H NMR (400 MHz, CDCl3): d
[ppm] 7.01 (d, 2H, J ¼ 1.8 Hz, C2–H, C6–H), 6.98 (s large, 2H,
O–H), 6.60 (t, 1H, J ¼ 2.0 Hz, C4–H), 4.30 (t, 2H, J ¼ 6.5 Hz,
COOCH2), 1.76 (m, 2H, COOCH2CH2), 1.36 (m, 4H,
COOCH2CH2(CH2)2), 0.90 (t, 3H, J ¼ 6.5 Hz, CH3). 13C NMR
(100 MHz, CDCl3): d [ppm] 167.5 (COO), 156.8 (C3, C5), 132.1
(C1), 108.9 (C2, C6), 107.3 (C4), 66.1 (COOCH2), 28.2
(COOCH2CH2), 22.1 (COO(CH2)2CH2), 22.6 (COO(CH2)3CH2),
14.0 (CH3); FTIR (ATR-diamond) nmax: 3351, 2958, 2933, 2873,
1763, 1749, 1688, 1599 cmꢂ1; HRMS (FAB+) m/z 225.1118 (MH+,
calcd for C12H16O4: 225.1121). Anal. found C, 63.76; H 7.27.
Calcd for C12H16O4: C, 64.27; H 7.19.
Fig. 1 Structure of BHPB-n.
Herein we have studied the properties of a series of
analogues where the length of the ester chain has been varied.
When the length remains within a certain range, these
analogues yield tubes with controlled diameters as shown by
SANS. When this length goes beyond this range, the compounds
self-assemble into either at or helical tapes. The morphology
of the self-assemblies in the whole series has been explored by
freeze-fracture TEM. The interactions that govern the self-
assemblies have been studied by FTIR and DSC.
3,5-Bis-(5-hexylcarbamoyl-pentoxy)-benzoic acid pentyl ester
(BHPB-5). A solution of pentyl-3,5-dihydroxybenzoate (204 mg,
0.91 mmol, 1 equiv.), 6-bromo-N-hexylhexanamide27 (509 mg,
1.83 mmol, 2.05 equiv.) and Bu4NBr (115 mg, 0.36 mmol, 0.4
equiv.) in DMF (50 mL) was treated with K2CO3 (625 mg,
4.52 mmol, 5 equiv.) and stirred at 50 ꢀC for 12 h. The mixture was
combined with water (200 mL) and acidied to pH 1 to 2 with
aqueous HCl (10%). The precipitate was ltered and recrystallized
in acetonitrile to afford pure BHPB-5 as a white solid (473 mg,
Material and methods
Compounds
The chemicals were purchased from Aldrich or Acros and
used as received. NMR spectra were recorded on a Bruker
Avance 400 operating at 400 MHz for 1H and 100 MHz for 13C.
The FTIR spectra of the neat compounds were recorded on a
Brucker Vertex 70 spectrometer equipped with an ATR dia-
mond reection unit (MVPStar). Mass spectra were recorded
with a Bruker Daltonique microTOF operating with an elec-
trospray source.
The compounds were synthesized according to Scheme 1.
As an example, the synthesis and characterization of DHB-5
and BHPB-5 are given below. The data on the characterization of
all other compounds are given in ESI.†
1
ꢀ
84%): M.p. 82.5 C. H NMR (400 MHz, CDCl3): d [ppm] 7.14 (d,
2H, J ¼ 2.2 Hz, C2–H, C6–H), 6.60 (t, 1H, J ¼ 2.2 Hz, C4–H), 5.44 (s
broad, 2H, NH), 4.29 (t, 2H, J ¼ 6.7 Hz, COOCH2), 3.97 (t, 4H, J ¼
6.5 Hz, ArOCH2), 3.24 (q, 4H, J ¼ 6.7 Hz, CH2NHCO), 2.19 (t, 4H,
J ¼ 7.6 Hz, CH2CONH), 1.76 (m, 10H, COOCH2CH2, ArOCH2CH2,
ArOCH2CH2CH2), 1.53–1.29 (m, 24H, CH2), 0.88 (t, 9H, J ¼ 6.8 Hz,
CH3). 13C NMR (100 MHz, CDCl3): d [ppm] 172.7 (CONH), 166.5
(COO), 160.0 (C3, C5), 132.3 (C1), 107.7 (C2, C6), 106.2 (C4), 68.0
(ArOCH2), 65.3 (COOCH2), 39.5 (CH2NH), 36.7 (CH2CONH), 31.4
(NH(CH2)3CH2), 29.6 (NHCH2CH2), 28.9 (ArOCH2CH2), 28.1
(COOCH2CH2), 26.6 (NH(CH2)2CH2), 25.8 (COO(CH2)2CH2), 25.7
(CH2CH2CONH), 25.5 (CH2(CH2)2CONH), 22.6 (COO(CH2)3CH2),
22.5 (NH(CH2)4CH2), 14.1 (COO(CH2)4CH3), 14.0 (NH(CH2)5CH3).
IR (ATR-diamond) nmax: 3292, 3093, 2952, 2926, 2857, 1718, 1638,
1611, 1552, 1443, 1318, 1230, 1159 cmꢂ1. HRMS (FAB+) m/z
619.4659 (MH+, calcd for C36H62N2O6: 619.4681). Anal. found: C,
70.23; H, 10.19; N, 4.39. Calcd for C36H62N2O6: C, 69.86; H, 10.10;
N, 4.53.
Pentyl-3,5-dihydroxybenzoate (DHB-5).
A
solution of
3,5-dihydroxybenzoic acid (5 g, 32.4 mmol), pentanol (2.85 g,
32.4 mmol, 1 equiv.) and para-toluenesulfonic acid (279 mg,
1.62 mmol, 0.05 equiv.) in toluene (150 mL) was reuxed in a
Dean–Stark apparatus. Aer elimination of water, the reaction
mixture was cooled to 25 ꢀC and the toluene was evaporated
under reduced pressure to afford a yellow oil that was mixed
with aqueous saturated NaHCO3 (500 mL). The resulting
mixture was extracted with CH2Cl2 (3 ꢁ 50 mL). The combined
Formation of the gels
BHPB-n (20 mg) and cyclohexane (980 mg) were weighed in a
vial (diameter 14 mm). The vial was closed tightly with a screw
cap equipped with a Teon gasket. The mixture was heated at
80 ꢀC until complete dissolution of BHPB-n and the solution let
ꢀ
stand at 25 C until the formation of a gel.
DSC
The thermograms were measured with a microcalorimeter DSC
III (Setaram) at a heating rate of 0.3 and 0.1 ꢀC minꢂ1 and gave
the same peak positions. Two heating–cooling cycles were
carried out. The peak integrations were done on the second
heating phase for all the samples.
Scheme
1
Synthesis of compounds BHPB-n. (a) CnH2n+1OH, para-toluene-
sulfonic acid, toluene, 140 ꢀC; (b) C6H13NHCOC5H10Br, Bu4NBr, K2CO3, DMF.
Soft Matter
This journal is ª The Royal Society of Chemistry 2013

View Article Online
Paper
Soft Matter
coherent
solute
FTIR
Data modeling. The corrected scattered intensity dS
/
dU(q) (cmꢂ1) (denoted dS/dU(q) hereaer) can be compared to
several models in order to determine the form and the size of
the aggregates. In the case of monodispersed non-interacting
particles, it can be written as
Spectra were recorded on a Bruker Vertex 70 spectrophotometer
equipped with a thermostatic cell holder and a temperature
controlling unit (Specac West 6100+). The gels were placed in a
NaCl cell with an optical path of 0.1 mm. The temperature was
increased by 1 ꢀC before each spectrum was obtained (rate 1 ꢀC
dS/dU(q) ¼ fVDr2VPP(q)
(1)
minꢂ1
) and the measurement was performed aer the
temperature was stabilized (about half a minute). The spectra
were recorded at different temperatures, were corrected for CO2
and water vapor and compensated for C6H12. They were pro-
cessed and tted using Igor (Wavemetrics, Inc.). The maxima of
the peaks were found with the built-in algorithm using second
derivatives.
where fV is the volume fraction of the particles, VP their volume
(cm3), P(q) their normalized form factor, and Dr the difference
between their scattering length density r (cmꢂ2) and that of the
solvent. In some gels, large heterogeneities were sometimes
observed. Thus, the real volume fraction under investigation
was not known with a great accuracy and we chose to normalize
the data to the Porod invariant as dened by
ð
N
ꢀ
1
2p2
Q ¼
q2dS dUðqÞdq
(2)
Electron microscopy
0
The freeze fracture was performed according to protocols
previously described.28 The grids were observed with a Philips
CM 12 operating at 120 kV or a FEI Tecnai G2 20 Twin operating
at 200 kV. The images were taken with a CCD camera. The
distances were measured with the soware Analysis (SIS-
For an ideal two-phase system this quantity is related to the
contrast factor and the volume fraction of the scattering parti-
cles through the relation
Q ¼ Dr2fV(1 ꢂ fV)
(3)
¨
Olympus Munster Germany). The cross-sectional sizes of the
objects were measured on statistically large numbers of objects
(c.a. 400). The statistics obtained from different samples or
from different freeze fractures prepared from the same
compounds showed superimposable histograms.
Since the volume fraction is very low (around 1%), Q z
Dr2fV, and the normalisation to the invariant can nally be
written as
ꢀ
dS dUðqÞ
¼ VPPðqÞ
(4)
Q
Small angle neutron scattering
The Porod invariant was calculaꢂte1d from the integration of
Experimental. Small Angle Neutron Scattering (SANS)
experiments were performed on the diffractometer D22 at the
Institut Laue Langevin (Grenoble, France). Data were collected
at four sample-to-detector distances: d ¼ 1.5, 3.6, 11.20 m
q2dS/dU(q) between qmin and 0.4 A
.
˚
Analysis procedure. The nite instrument resolution was
evaluated from the size of the incident beam on the detector
and the wavelength spread as described by Pedersen29 and
derived for D22 by Grillo.30 Normalized data were then
compared to the models (isolated innite tubes, clusters of
tubes, and combination of both, as explained in the Results and
discussion section) using the IGOR Pro package and the
extension developed by the NIST Center for Neutron Research.31
This approach allowed the determination of the average
external radius rext, the distribution width Drext, the thickness of
the wall of the hollow cylinders e and the fraction of tubes
forming clusters fCluster. A more general scaling factor (scale)
was introduced in order to overcome problems with the Porod
invariant normalization. Finally, a small background contribu-
tion (bkg) was also considered.
0
˚
(collimation length d ¼ 11.20 m, wavelength l ¼ 6 A, wave-
length distribution Dl/l ¼ 0.1) and d ¼ 17.60 m (d0 ¼ 17.60 m,
˚
l ¼ 16 A, Dl/l ¼ 0.1). The scattered intensity was recorded with
3
a 2-dimensional He gas detector allowing translational offset
in order to increase the highest scattering angle. These cong-
urations allowed measurements in the q vector range 0.0011–
0.64 Aꢂ1 (q is the magnitude of the scattering vector dened by
˚
q ¼ 4p/l sin(q), where l is the wavelength of the incident beam
and 2q the scattering angle). Measurements were performed in
1 mm thick quartz cells (Helma). Deuterated cyclohexane was
used as the solvent in order to increase contrast. The temper-
ꢀ
ature was kept constant at 22 C during all measurements.
Data were corrected with standard ILL programs for
isotropic scattering. They were converted to an absolute scale
using H2O as a standard to yield the total differential scattering
Results and discussion
Thermodynamic properties
total
cross section per unit volume dS
/dU(q) (cmꢂ1). This
sample
quantity is the sum of the coherent differential scattering cross-
section of the solute dSsolute
coherent/dU(q) (cmꢂ1) related to the The structure of BHPB-n is depicted in Fig. 1. n is the number of
structure of the aggregates and a background that comprises carbon atoms in the ester chain. BHPB-n form organogels in
the scattering of the solvent and the incoherent scattering of the alkanes for n $ 5. Compounds with smaller chain lengths form
BHPB-n molecules. The latter contributions were subtracted precipitates in these solvents, and have not been studied in this
and slightly adjusted in order to get a coherent signal very close paper. Gels are formed simply by heating the compound in the
to zero in the highest q region (around 0.5–0.6 Aꢂ1).
solvent at a given wt concentration until complete dissolution
˚
This journal is ª The Royal Society of Chemistry 2013
Soft Matter

View Article Online
Soft Matter
Paper
The lowest concentration gives a weak but still detectable
enthalpy value (Fig. 3A). The gel-to-sol transition temperature
increases with concentration and the heat Q is proportional to
gelator concentration (Fig. 3B), meaning that the transition
enthalpy per mole of BHPB-n, DH, is constant in the studied
C-range. For a better accuracy, DH values are derived from the
linear t of the Q vs. C plot. The plot of DH values against the
number n (Fig. 3C) clearly displays three regimes that were
slightly visible for the mgc. In the rst regime, when n # 7, DH
values are below 46 kJ molꢂ1, in the third one, when n $ 13, they
are above 71 kJ molꢂ1 and, in both regimes, they do not vary
monotonously. In the second regime, when 8 # n # 12, DH
varies from 53 to 58 kJ molꢂ1 with a linear progression. These
regimes are separated by sharp gaps in enthalpy: 10 kJ molꢂ1 for
the rst one and 19 kJ molꢂ1 for the second one. These differ-
ences are much larger than the values of the melting heat per
methylene in paraffins33–35 (3.85 to 4.10 kJ per mol of CH2) or
polyethylene34 (3.97 kJ per mol of CH2). Therefore they cannot
be explained by the sole van der Waals interactions of the
additional methylene unit. The division into different regimes
reects a different kind of packing. Within the linear regime,
the slope is 1.14 ꢃ 0.47 kJ per mole of CH2. This slight variation
suggests that compounds BHPB-8 through BHPB-12 self-
assemble with a very similar mechanism.
Fig. 2 Minimum gel concentrations of BHPB-n in cyclohexane.
and cooling the resulting solution at 25 ꢀC. This gelation is
thermoreversible, which shows that the molecules self-assem-
bled through non-covalent interactions. The minimum gelator
concentrations required to form a self-supporting gel (mgc)
ꢀ
were measured in cyclohexane at 25 C and are reported as a
function of the ester chain length n (Fig. 2).
The mgc values are divided into two sets: smaller than
0.1 wt% for 8 # n # 12 and equal to or greater than 0.2 wt% for
n # 7 or n $ 13. The lowest value, found for BHPB-9 and BHPB-
10, is 0.05 wt%, which shows that they are efficient organo-
gelators.32 The enthalpy of the gel-to-sol transition was
measured by DSC for each compound at 4 different concen-
trations above the gel concentration.
Infrared spectroscopy
In order to identify the interactions involved in the self-
assemblies, FTIR measurements were carried out on the gels.
The spectrum of the gels of BHPB-10 in C6H12 (2 wt%) as an
example is shown in Fig. 4.
The presence of a NH-stretching band at 3299 cmꢂ1 (Fig. 4
le), an amide I band at 1642 cmꢂ1 and an amide II band at
1551 cmꢂ1 (Fig. 4 right) provides strong evidence of H-bonds
between all the amide groups within the self-assemblies. By
comparison, in chloroform, BHPB-10 is completely dissociated
and the same bands are strongly shied (nNH: 3449 cmꢂ1, amide
I: 1662 cmꢂ1, amide II: 1519 cmꢂ1), which proves that the amide
groups are not H-bonded. The spectra were recorded for
all species at different temperatures. As an example, the
Fig. 3 (A) DSC thermograms of BHPB-12 in cyclohexane, for different concen-
trations (endothermic flow down); (B) measured heat per gram of gels as a
function of the concentration. The dashed line gives the best linear fit and
corresponds to DH ¼ 58.8 ꢃ 1.6 kJ molꢂ1; (C) enthalpies of gel melting vs. the
number of carbon in the ester chain.
Fig. 4 FTIR spectra of BHPB-10 in C6H12 (gel – 2 wt%) ( ); BHPB-6 in C6H12 (gel –
2 wt%) ( ); BHPB-10 in toluene (gel – 2 wt%) ( ) (for the sake of clarity the
uncompensated peak of toluene at 1600 has been erased); BHPB-10 in CHCl3
(solution – 2 wt%) ( ). Left: NH stretching region. Right: CO stretching and CN
bending region.
Soft Matter
This journal is ª The Royal Society of Chemistry 2013

View Article Online
Paper
Soft Matter
frequencies and intensity of the amide I band and nNH bands of be also deconvoluted into two peaks (1644/1638 cmꢂ1 and 3307/
BHPB-8 are displayed in Fig. 5.
3285 cmꢂ1 resp.). Deconvolutions with similar peak shis can
When the gel is heated above the transition temperature, the be done for all species except BHPB-5 and BHPB-6. When the
amide I band broadens and diminishes while a shouldꢂe1r temperature rises above the gel-to-sol transition, the double
appears at 1680–1690 cmꢂ1 (Fig. 5A). The NH band at 3299 cm
ester band also evolves toward a single peak (Fig. 5A). These
also decreases while a weak band appears at 3455 cmꢂ1 observations prove that not only the amide groups but also the
(Fig. 5C). These new shis are characteristic of non-bonded ester groups are involved in the formation of self-assemblies
amides and are indeed close to the ones in CHCl3. Both the and justify a posteriori the need to modify the ester group to
frequencies and the intensities of the bands show a steep vari- probe its role.
ation at a temperature that coincides with the transition
As a conclusion, both the DSC and FTIR studies show that
observed by DSC (Fig. 5B and C). It proves that H-bonds between the main interactions driving the self-assembly are the H-bonds
amides are involved in the self-assemblies. The energy of such between the amide groups. They account for half of the tran-
bonds can be assumed to be equal to the enthalpy of self- sition enthalpy. The splitting of some bands corroborates the
association of N-methylacetamide in aprotic solvents, 16 kJ involvement of the esters in the assemblies. It also suggests that
molꢂ1 36 multiplied by two since two amides are involved. This the structure of the BHPB-n self-assemblies with n $ 7
,
value represents a large part of the transition enthalpies comprises two different sets of molecules with asymmetrical
measured by DSC (more than 50% when n # 12) indicating the orientations.
predominance of the H-bonds in the self-assembly.
The ester CO stretching band is split into two bands
(1725 and 1716 cmꢂ1). This double frequency is observed for all
species except BHPB-5 and 6. The lower frequency cannot be
attributed to a H-bond from one of the amide groups since the
corresponding nNH band should be visible at 3350 cmꢂ1. It may
reect the presence of two sets of molecules with different
orientations or environments within the self-assemblies.
Different orientations of the carboxyl group with regard to the
aromatic ring can lead to different shis for the ester mainly
due to different degrees of conjugations, and may explain the
double frequency also observed for the aromatic C]C stretch-
ing band (1599 cmꢂ1 and 1612 cmꢂ1). The amide I and the
NH stretching bands, even if they appear like single bands, can
Electron microscopy
The microscopic structure of the gels was explored by freeze
fracture TEM. This technique originally developed for aqueous
media has been successfully implemented in organic solvents
especially to observe organogels.37–40 The self-assemblies were
examined by electron microscopy at different times aer the
formation of the gels. The same structures were observed in the
gels aer a few minutes or aer 6 months, which shows their
stability. The width or diameters of the observed objects were
measured for a statistically relevant number of objects (about
400) to calculate their average values and distribution (reported
in Table 1). The uncertainty is rather large (3 nm) because the
replicas are obtained aer coating the samples with metal
particles.
We present hereaer the morphologies and the sizes of the
structures in each regime as identied by DSC. When n # 7,
Table 1 Width or diameters of BHPB-n self-assemblies measured from TEM
Shape of the
self-assemblies
Avg. width or
Std. deviationa
[nm]
Compound
diametera [nm]
BHPB-5
BHPB-6
BHPB-7
Flat ribbons
Flat ribbons
Flat + helical
ribbons
70.1
41.5
34.2
23.3
18.1
9.1
BHPB-8
BHPB-9
BHPB-10
BHPB-11
Tubes
Tubes
Tubes
Helical + twisted
ribbons
24.8
25.8
28.5
32.7
3.6
4.4
3.8
7.2
BHPB-12
BHPB-13
Tubes and ribbons
Helical + twisted
ribbons
34.1
56.4
6.9
19.7
BHPB-14
BHPB-15
BHPB-16
Helical + twisted
ribbons
Helical + twisted
ribbons
Helical + twisted
ribbons
40.7
39.5
23.3
13.2
8.4
Fig. 5 VT-FTIR spectra of gels of BHPB-8 in C2H12 (2 wt%) from 25 ꢀC to 80 ꢀC in
the CO stretching and CN bending region (A) and in the NH stretching region (C).
Frequencies ( ) and intensities ( ) of the amide I band (B) and of the nNH band (D)
vs. temperature. Dashed line: transition temperature measured by DSC.
5.1
a
Uncertainty ꢃ 3 nm.
This journal is ª The Royal Society of Chemistry 2013
Soft Matter

View Article Online
Soft Matter
Paper
previous ones. Even if BHPB-8 to BHPB-10 form more regular
tubes, a small proportion of them bear helical grooves (Fig. 7B,
arrowheads), indicating that they stem from wound tapes. The
presence of helical tapes among tubes has been observed by
TEM as early as the 80s by Kunitake, who identied them as
precursors of the tubes.41 This hypothesis has been lately
conrmed by scattering techniques.42 It has also been explained
by many theoretical models.10–12,14,17 In toluene, BHPB-8 to
BHPB-12 self-assemble with a different morphology from that
in C6H12: they form only thick and straight brils (Fig. 7F). In
addition, the FTIR spectra in toluene show a band at 1680 cmꢂ1
,
proving that a large proportion of the amides are not H-bonded,
which is in contrast to the self-assemblies in C6H12 where all the
amides are H-bonded. Therefore toluene weakens the self-
assembly, probably by interacting with BHPB-n through p–p
interactions. In addition, the ester band shows a major band at
1710 cmꢂ1 (a shouldering band is present at 1724 cmꢂ1 but with
a much lower intensity) (Fig. 4). The structures of the self-
assemblies of compounds BHPB-n, when n $ 12, are shown in
Fig. 8. They all exhibit ribbons with average widths varying from
56 nm to 25 nm (Table 1). Most of these ribbons tend to wind
into tapes, with irregular helical turns along their main axis. No
cylindrical structure was ever observed for these compounds.
To summarize, TEM shows that BHPB-n with n $ 7 form
ribbons with limited width that tend to wind into helical or
tubular shapes. BHPB-5 and 6 are the only compounds in the
series forming truly at ribbons. They also have a single ester
nCO band in their IR spectra, whereas all other compounds have
two nCO bands. This suggests that the tapes wind only when the
esters in the constituting molecules have two different confor-
mations or orientations. This hypothesis is also supported by
the fact that in toluene, only at ribbons form and, in
conjunction, one of the ester nCO peaks becomes predominant.
Fig. 6 Freeze fracture TEM of BHPB-n in C6H12 (2 wt%). (A) BHPB-5; (B) BHPB-6;
(C) BHPB-7.
compounds BHPB-n form ribbons of several mm length (Fig. 6).
For BHPB-5, ribbons with polydisperse widths (between 30 nm
and 150 nm) stack to form thick multilamellar objects. BHPB-6
and BHPB-7 form individualized ribbons with polydisperse
widths between 13 and 160 nm (average 41 nm) and between 15
and 85 nm (average 34 nm) respectively.
The subgroup of compounds BHPB-5 to 7 has a trend to form
narrower and thinner objects when the ester size increases.
Although BHPB-7 belongs to the same DSC regime as BHPB-5
and BHPB-6, it is distinguished by a different morphology: it
forms twisted ribbons in large proportions (Fig. 6C). For n ¼ 8, 9
and 10, BHPB-n form regular cylinders with lengths in the range
of micrometers (Fig. 7). The shadowing pattern of the visible
ends of the cylinder shows that they are hollow (Fig. 7, arrows).
The diameters of the tubes signicantly increase when n the
length of the ester chain increases. BHPB-11 forms no tubes but
a mixture of irregular helical tapes with different diameters
(Fig. 7D). BHPB-12 self-assembles into mixtures of tubes and
helical tapes (Fig. 7E).
Small angle scattering
TEM clearly proved the tubular shape of some aggregates and
can provide an estimate of their outer diameter. But the inner
and outer diameters can be measured more precisely only
by small angle scattering techniques.22,42–47 Therefore the
The mean diameter of the objects formed by BHPB-12 is
34.1 ꢃ 6.9 nm, which is larger and more polydisperse than the
Fig. 7 Freeze fracture TEM. (A) BHPB-8; (B) BHPB-9; (C) BHPB-10; (D) BHPB-11; (E)
BHPB-12 in C6H12 (2 wt%). Arrows: cross section of tube ends. Arrowheads: helical
grooves. (F) BHPB-10 in toluene (4 wt%).
Fig. 8 Freeze fracture TEM of (A) BHPB-13; (B) BHPB-14; (C) BHPB-15; (D) BHPB-
16 in C6H12 (2 wt%).
Soft Matter
This journal is ª The Royal Society of Chemistry 2013

View Article Online
Paper
Soft Matter
self-assemblies formed by compounds BHPB-8 to BHPB-12 were
studied by SANS. They were prepared in deuterated solvent
(1 wt% conc.) and subjected to a neutron beam. The intensities
scattered by the self-assemblieꢂs1were collected in a wide range
of scattering vector q (0.0011 A < q < 0.64 Aꢂ1). The corrected
intensities were normalized to the Porod invariant Q (see
Material and methods). The normalized intensities depend only
on the geometrical parameters of the particles (eqn (4)).
General overview of the normalized scattering cross
sections. The normalized scattering cross sections dS/dU(q)/Q
for the series of gelators are presented in Fig. 9.
˚
˚
BHPB-8, BHPB-9 and BHPB-10 exhibit very similar features:
in the lowest q region (q < 0.01 Aꢂ1), the normalized intensity
˚
Fig. 10
: Experimental scattering cross section normalized to the Porod
varies as a power-law with an exponent between 1.5 and 2. No
plateau is observed. If the scattering intensity is only due to
uncorrelated particles, it indicates that the Guinier region is not
reached with the experimental setup but is shied to lower q
values. This is consistent with very long particles as observed by
electron microscopy experiments and validates the choice of an
invariant (BHPB-10, 2% in C6D12); dotted line: single infinite hollow cylinder
model (eqn (4) and (5)) with rext ¼ 13.5 nm, e ¼ 3.5 nm, scale factor ¼ 1; solid line:
same model with experimental resolution and radial distribution Dr ¼ 0.68 nm
(eqn (4), (5) and (7)); dashed lines: qꢂ1 and qꢂ4 variation.
envelope curve of the fast theoretical oscillations. This deviation
originates from the combination of nite experimental resolu-
tion and external radius polydispersity. The latter effect is more
pronounced for BHPB-9 than for BHPB-8 and -10.
The resolution was calculated and applied to the t as
described in the Material and methods section. Radial poly-
dispersity was taken into account by a symmetrical Gaussian
distribution of the external radii g(rext) (eqn (6)).
innite cylinder form factor. Above 0.01 Aꢂ1, we enter the
˚
asymptotic region (Porod region). The scattered intensity has an
overall decrease in qꢂ4, characteristic of sharp interfaces. It also
displays sharp oscillations, related to the cross-section of the
brillar objects. Their positions move with the nature of the
compounds and indicate a variation of the external radius for
BHPB-8, BHPB-9 and BHPB-10. At higher q-values, a large
oscillation is found at a common value (around 0.25 Aꢂ1 for the
˚
!
2
three compounds).
The form factor of long hollow cylinders (e.g. for which L [
10rext, with L the length of the nanotubes) can be written as
1
ðrext ꢂ rext
Þ
pffiffiffiffiffiffi
gðrextÞ ¼
exp ꢂ
(6)
2
2
Drext
Drext 2p
ꢁ
ꢂ
2
In this equation, rext is the average external radius and Drext
the distribution width (or standard deviation). The poly-
dispersity is dened as Drext=rext. The thickness e was kept
constant and the normalised scattering cross section can then
be written as
p rext J1ðqrextÞ ꢂ ðrext ꢂ eÞJ1ðqðrext ꢂ eÞÞ
Pðq; rext; eÞ ¼
(5)
qL
qeðrext ꢂ e=2Þ
with rext the external radius of the tube, e the thickness of the
wall and J1 the Bessel function of rst kind and rst order. This
single tube model reproduces the main experimental features of
the experimental curve (Fig. 10) on an absolute scale.
But compared with the theoretical curve, the experimental
oscillations are attenuated, and for q > 0.25, reproduce only the
ꢀ
ð
N
dS dUðqÞ
¼
VPðrextÞgðrextÞPðq; rextÞdrext
(7)
Q
0
Structural renements. Even when the resolution and radial
polydispersity are taken into account, the intensity in the low q
region (below 5 ꢁ 10ꢂ3
ꢂ1) is underestimated (Fig. 10). More-
˚
A
over the positions of the rst minima and maxima mismatch
those of the best ts. The scaling exponent for low q is larger
than unity, as it should be for uncorrelated cylinders. This
intensity can be explained by the association of the tubes into
clusters and can be modelled successfully by clusters of three
tubes47 and even simple pairs.28 We have implemented the pair
model in the present study. When two identical parallel tubes
associate in close contact with a centre-to-centre distance 2rext
,
the initial intensity (eqn (5)) has to be multiplied by the inter-
ference term:48
(1 + J0(2qrext))
(8)
where J0 is the Bessel function of rst kind and zero order.
Therefore, the calculation was approximated by considering
pairs of tubes with the same radius and by varying this radius
Fig. 9 Normalized scattering cross sections dS/dU(q)/Q for the BHPB-n series.
For the sake of clarity a vertical scaling offset (ꢁ10) is applied to each curve except
for BHPB-8.
This journal is ª The Royal Society of Chemistry 2013
Soft Matter

View Article Online
Soft Matter
Paper
according to the distribution function g(rext). In the absence of
correlation between pairs, eqn (4) is replaced by
ð
N
dS=dUðqÞ
¼
¼
VPðrextÞ gðrextÞPðq; rextÞð1 þ J0ð2qrextÞÞ drext
VPðrextÞ gðrextÞPðq; rextÞ drext
Q
0
ð
N
0
ð
N
þ
VPðrextÞ gðrextÞPðq; rextÞJ0ð2qrextÞ drext
(9)
0
To calculate the exact intensity, we should consider pairs of
cylinders with different diameters, derive their scattered
intensity and then sum over all the pairs. But the simple eqn (8)
holds only for identical cylinders and the exact intensity is more
complicated to derive. The second term in eqn (9) represents
the extra scattering contribution from the pairing. It exhibits a
series of positive and negative oscillations at low q values and
tends to 0 with increasing q values. These maxima/minima,
restricted at low q values, alter the shape of the rst oscillations
(rst term in eqn (9)). In addition, the association increases the
intensity at low q values (both terms of eqn (8) are identical for
sufficiently low q). The qꢂ1 behaviour is still present but is now
shied to smaller q values than for isolated nanotubes. Finally,
solutions where only a fraction fCluster of the tubes form clusters
can be modelled simply by multiplying the second term in eqn
(10) by fCluster. Fig. 11 presents the best t derived from eqn (9).
For a better visualization, two other representations are shown:
q2dS/dU(q)/Q (Kratky plot, Fig. 11B) and q4dS/dU(q)/Q (Porod
representation, Fig. 11C).
The model reproduces quantitatively the normalized scat-
tering cross section with a scale factor close to unity (scale ¼
0.95). The presence of clusters greatly renes the t below
0.03 Aꢂ1 where the experimental low q behavior and the shape
˚
of the rst oscillation are reproduced more exactly. The same
procedure was applied for BHPB-8 and -9 (Fig. 12).
Fig. 11 Structural refinement of BHPB-10 using the cluster model of eqn (9)
The intensity scattered by BHPB-11 shows no oscillations
and no simple geometric model can reproduce the data. These
observations are consistent with the irregular and helical tapes
with different diameters or widths observed by TEM for BHPB-
11. No attempt was made to t the data with a more sophisti-
cated model. The scattering pattern observed for BHPB-12 is
very close to that of BHPB-11, especially with a similar Bragg
(solid line). (A) Logarithmic; (B) Kratky; (C) Porod representations.
1/(1 ꢂ fBHPB-11) to renormalize correctly to the modied Porod
invariant. The resulting scattering cross section (Fig. 12, n ¼
12*) becomes very similar to that of the other samples and can
contribution around 0.2 Aꢂ1. But it also shows attenuated
ꢂ1
˚
˚
be tted with a single hollow cylinder model below 0.05 A
.
oscillations, too weak to be modelled by tubes even with a large
external radius polydispersity. This feature results from the
mixture of particles of different shapes (tapes and nanotubes)
and cannot be addressed by our model although well-dened
nanotubes are clearly present.
A more accurate inspection needs to separate the different
contributions. Since the scattering cross sections of BHPB-11
and BHPB-12 are very similar, we have assumed that the tapes
present in BHPB-12 gels are the same as those in BHPB-11 gels.
The scattering of BHPB-11 was treated as the contribution of the
tapes in BHPB-12. Thus, we removed a fraction fBHPB-11
of dSBHPB-11/dU/Q from dSBHPB-12/dU/Q to enhance the
contribution of the sole tubes. The result was multiplied by
For higher q values, the renement has a lower quality
than for the other compounds but the subtraction allows
the extraction of the rst oscillations and a good estimate of the
parameters of the tubes in the mixture.ꢂW1 e also note that the
˚
Bragg peak is still present around 0.2 A and partially masks
the large oscillation around 0.25 Aꢂ1. The intensity of BHPB-9
˚
was treated by the same procedure as BHPB-12, the subtraction
of dSBHPB-11/dU(q)/Q, but with a smaller fBHPB-11 fraction. This
treatment enhanced the rst minimum around 0.02 Aꢂ1 with a
˚
better denition. Fig. 12 presents the best ts for the series of
self-assemblies. Final parameters are reported in Table 2. All the
ts in different representations (log–log, Kratky and Porod) are
presented in the ESI.†
Soft Matter
This journal is ª The Royal Society of Chemistry 2013

View Article Online
Paper
Soft Matter
Fig. 13 (A) Average external diameter De of the tubes vs. n; (B) polydispersity of the
tube diameters vs. n. Solid line: measured by SANS. Dashed line: measured by TEM.
which suggests that a lower diameter polydispersity tends to
favor the alignment of the tubes.
Fig. 12 Best fits (solid lines) for the series of gelators BHPB-n. Samples for which
a fraction of dSBHPB-11/dU(q)/Q has been removed are indicated with an asterisk.
A vertical scaling offset (ꢁ10) is applied to each curve except for BHPB-8.
Conclusions
The thickness of the wall of the nanotubes is constant within
the uncertainty of the measurements when n varies. The found
distances 3.4 to 3.5 nm correspond to the lengths of the molecules
in extended conformations. A monolayer could t within this wall
thickness, but more complex models cannot be ruled out.
The evolution of the external diameter with the ester chain
length n is plotted in Fig. 13A. This curve shows a strong
correlation between the size of the ester and the radii of the
tubes.
The polydispersity of the tubes (Fig. 13B) reaches its smallest
for BHPB-10 (about 5%). BHPB-8 forms tubes with a higher
radial polydispersity, but no subtraction was necessary to
implement the model, which proves that only tubes are present
in the mixture. BHPB-9 forms tubes with about the same poly-
dispersity as BHPB-8; however, intensities scattered by BHPB-9
have attenuated oscillations that have been modelled by the
presence of non-tubular objects in the mixtures. The quality of
the self-assembled nanotubes follows the order BHPB-10 > 8 >
9 > 12 [ 11. The length of 10 C is thus the optimal value,
yielding tubes with the lowest dispersity. For this length, the
fraction of paired tubes is also higher than for BHPB-8 or 9,
We have studied a family of ester-diamide compounds that are
organogelators at low concentration. We have identied the
length of the ester chain (n) as the molecular parameter that
governs the shape of the self-assemblies. When this length is
below a given threshold (n # 7), the self-assemblies are at
ribbons, and when the length is above (n $ 8) the ribbons tend
to wind into irregular helical tapes. For lengths between 8 and
10 C, the self-assemblies are mainly nanotubes with a well-
dened shape, and for the length of 12 C, they self-assemble
into more polydisperse tubes along with a majority of helical
tapes. The scattering methods and the models we proposed
successfully provide the diameter distribution and the wall
thickness of the tubes. They also yield an estimate of the ratio of
paired tubes and the proportion of non-tubular objects that may
eventually co-exist with these tubes. An ester length of 10 C
produces tubes with the lowest diameter polydispersity. A slight
variation of this length yields nanotubes with varying diameters
and with a broader distribution. FTIR and DSC studies show
that the H-bonds between the amide groups are the main
interactions driving the self-assemblies. Additionally, FTIR also
proves the involvement of the ester groups. Whenever the
Table 2 Parameters from the best fits for the series BHPB-8 to BHPB-12
BHPB-9,a
BHPB-12,a
Compound
BHPB-8
BHPB-9
fBHPB-11 ¼ 0.19
BHPB-10
fBHPB-11 ¼ 0.64
Average external radius (nm)
Standard deviation rext (nm)
Drext=rext
10.58 ꢃ 0.1
1.00 ꢃ 0.05
0.094
12.50 ꢃ 0.1
1.35 ꢃ 0.05
0.108
12.50 ꢃ 0.1
1.25 ꢃ 0.05
0.100
13.54 ꢃ 0.1
0.68 ꢃ 0.05
0.050
16.03 ꢃ 0.2
1.63 ꢃ 0.1
0.101
Thickness of the wall e (nm)
Fraction of tubes in
clusters fCluster
3.4 ꢃ 0.1
3.4 ꢃ 0.1
3.35 ꢃ 0.1
3.55 ꢃ 0.1
3.5 ꢃ 0.2
0.4
0.3
0
0.8
0
Scale factor
Background bkg (A )
0.96
70
0.88
30
0.95
40
0.96
25
1.06
60
b
3
˚
a
Sample for which a fraction of dSBHPB-11/dU(q)/Q has been removed; fBHPB-11: corresponding removed fraction. b See Experimental.
This journal is ª The Royal Society of Chemistry 2013
Soft Matter

View Article Online
Soft Matter
Paper
aggregates are curved, i.e. if they are either helical tapes or 20 A. Singh, P. E. Schoen and J. M. Schnur, J. Chem. Soc., Chem.
tubes, their IR signature involves two ester bands suggesting Commun., 1988, 1222–1223.
two different sets of molecules in the self-assemblies. Efforts to 21 R. Oda, F. Artzner, M. Laguerre and I. Huc, J. Am. Chem. Soc.,
elucidate the inner structure will be pursued by studies on
oriented samples.
2008, 130, 14705–14712.
22 C. Valery, M. Paternostre, B. Robert, T. Gulik-Krzywicki,
T. Narayanan, J. C. Dedieu, G. Keller, M. L. Torres,
R. Cherif-Cheikh, P. Calvo and F. Artzner, Proc. Natl. Acad.
Sci. U. S. A., 2003, 100, 10258–10262.
Acknowledgements
´
This work was supported by two fellowships from the Region 23 A. Brizard, C. Aime, T. Labrot, I. Huc, D. Berthier, F. Artzner,
Alsace (T. T. T. N. and F.-X. S.), one fellowship from the Mexican
CONACyT (N. D.) and by a grant from the International Center
for Frontier Research in Chemistry, Strasbourg.
B. Desbat and R. Oda, J. Am. Chem. Soc., 2007, 129, 3754–
3762.
24 M. Masuda and T. Shimizu, Langmuir, 2004, 20, 5969–
5977.
25 B. N. Thomas, C. M. Lindemann, R. C. Corcoran,
C. L. Cotant, J. E. Kirsch and P. J. Persichini, J. Am. Chem.
Soc., 2002, 124, 1227–1233.
26 C. Tarabout, S. Roux, F. Gobeaux, N. Fay, E. Pouget,
C. Meriadec, M. Ligeti, D. Thomas, M. IJsselstijn,
F. Besselievre, D.-A. Buisson, J.-M. Verbavatz, M. Petitjean,
Notes and references
1 T. Shimizu, M. Masuda and H. Minamikawa, Chem. Rev.,
2005, 105, 1401–1443.
2 S. Vauthey, S. Santoso, H. Gong, N. Watson and S. Zhang,
Proc. Natl. Acad. Sci. U. S. A., 2002, 99, 5355–5360.
3 E. M. Wilson-Kubalek, R. E. Brown, H. Celia and
R. A. Milligan, Proc. Natl. Acad. Sci. U. S. A., 1998, 95, 8040–
8045.
´
C. Valery, L. Perrin, B. Rousseau, F. Artzner, M. Paternostre
and J.-C. Cintrat, Proc. Natl. Acad. Sci. U. S. A., 2011, 108,
7679–7684.
4 P. Ringler, W. Muller, H. Ringsdorf and A. Brisson, Chem.– 27 N. Diaz, F. X. Simon, M. Schmutz, M. Rawiso, G. Decher,
´
Eur. J., 1997, 3, 620–625.
J. Jestin and P. J. Mesini, Angew. Chem., Int. Ed., 2005, 44,
5 P. Huetz, S. van Neuren, P. Ringler, F. Kremer, J. F. L. van
3260–3264.
Breemen, A. Wagenaar, J. B. F. N. Engberts, 28 T.-T.-T. Nguyen, F.-X. Simon, J. Combet, M. Schmutz and
´
J. G. E. M. Fraaije and A. Brisson, Chem. Phys. Lipids, 1997,
89, 15–30.
P. J. Mesini, So Matter, 2011, 7, 1121–1128.
29 J. S. Pedersen, J. Phys. IV, 1993, 03, 491–498.
6 Y. Ono, K. Nakashima, M. Sano, Y. Kanekiyo, K. Inoue, 30 I. Grillo, in Small-angle neutron scattering and applications in
J. Hojo and S. Shinkai, Chem. Commun., 1998, 1477–1478.
7 J. H. Jung, Y. Ono, K. Hanabusa and S. Shinkai, J. Am. Chem.
Soc., 2000, 122, 5008–5009.
so and condensed matter, ed. R. Borsali and R. Pecora,
Springer-Verlag, Heidelberg, 2008, pp. 3–60.
31 S. Kline, J. Appl. Crystallogr., 2006, 39, 895–900.
8 J. H. Jung, S.-H. Lee, J. S. Yoo, K. Yoshida, T. Shimizu and 32 P. Terech and R. G. Weiss, Chem. Rev., 1997, 97, 3133–3159;
S. Shinkai, Chem.–Eur. J., 2003, 9, 5307–5313.
9 F. X. Simon, N. S. Khelfallah, M. Schmutz, N. Diaz and
P. J. Mesini, J. Am. Chem. Soc., 2007, 129, 3788–3789.
10 W. Helfrich, J. Chem. Phys., 1986, 85, 1085–1087.
11 U. Seifert, J. Shillcock and P. Nelson, Phys. Rev. Lett., 1996,
77, 5237.
R. G. Weiss and P. Terech, Molecular gels. Materials with self-
assembled Fibrillar Network, Springer, Dordrecht, The
Netherlands, 2006; Low Molecular Mass Gelator - Design,
Self-Assembly, Function, Top. Curr. Chem., ed. F. Fages,
2005, vol. 256; M.-O. M. Piepenbrock, G. O. Lloyd,
N. Clarke and J. W. Steed, Chem. Rev., 2009, 110, 1960–
2004; A. Dawn, T. Shiraki, S. Haraguchi, S.-I. Tamaru and
S. Shinkai, Chem. –Asian J., 2011, 6, 266–282.
12 J. V. Selinger and J. M. Schnur, Phys. Rev. Lett., 1993, 71,
4091.
13 J. V. Selinger, M. S. Spector and J. M. Schnur, J. Phys. Chem. B, 33 F. W. Billmeyer Jr, J. Appl. Phys., 2004, 28, 1114.
2001, 105, 7157–7169.
34 P. J. Flory and A. Vrij, J. Am. Chem. Soc., 1963, 85, 3548–3553.
14 J. V. Selinger, F. C. MacKintosh and J. M. Schnur, Phys. Rev. 35 B. Wunderlich and G. Czornyj, Macromolecules, 1977, 10,
E: Stat. Phys., Plasmas, Fluids, Relat. Interdiscip. Top., 1996,
53, 3804–3818.
906–913.
36 M. Davies and D. K. Thomas, J. Phys. Chem., 1956, 60, 767–
15 C. M. Chen, Phys. Rev. E: Stat. Phys., Plasmas, Fluids, Relat.
Interdiscip. Top., 1999, 59, 6192–6195.
770.
´
37 M. Schmutz and P. J. Mesini, in Handbook of Cryopreparation
16 D. S. Chung, G. B. Benedek, F. M. Konikoff and
J. M. Donovan, Proc. Natl. Acad. Sci. U. S. A., 1993, 90,
11341–11345.
Methods for Electron Microscopy, ed. A. Cavalier, D. Spehner
and B. M. Humbel, Francis and Taylor CRC Press, New
York, 2008, pp. 411–430.
17 W. Helfrich and J. Prost, Phys. Rev. A: At., Mol., Opt. Phys., 38 R. Schmidt, F. B. Adam, M. Michel, M. Schmutz, G. Decher
´
1988, 38, 3065–3068. and P. J. Mesini, Tetrahedron Lett., 2003, 44, 3171–3174.
18 R. Oda, I. Huc, M. Schmutz, S. J. Candau and 39 R. Schmidt, M. Schmutz, M. Michel, G. Decher and
´
P. J. Mesini, Langmuir, 2002, 18, 5668–5672.
F. C. MacKintosh, Nature, 1999, 399, 566–569.
19 S. B. Lee, R. Koepsel, D. B. Stolz, H. E. Warriner and 40 G. Palui, F.-X. Simon, M. Schmutz, P. J. Mesini and
A. J. Russell, J. Am. Chem. Soc., 2004, 126, 13400–13405. A. Banerjee, Tetrahedron, 2008, 64, 175–185.
´
Soft Matter
This journal is ª The Royal Society of Chemistry 2013

View Article Online
Paper
Soft Matter
41 N. Nakashima, S. Asakuma and T. Kunitake, J. Am. Chem. 45 T. Imae, K. Funayama, M. P. Kra, F. Giulieri, T. Tada and
Soc., 1985, 107, 509–510. T. Matsumoto, J. Colloid Interface Sci., 1999, 212, 330–337.
42 C. V. Teixeira, H. Amenitsch, T. Fukushima, J. P. Hill, W. Jin, 46 K. Sakurai, Y. Ono, J. H. Jung, S. Okamoto, S. Sakurai and
T. Aida, M. Hotokka and M. Linden, J. Appl. Crystallogr.,
2010, 43, 850–857.
43 P. Terech, A. De Geyer, B. Struth and Y. Talmon, Adv. Mater.,
2002, 14, 495–498.
S. Shinkai, J. Chem. Soc., Perkin Trans. 2, 2001, 108–112.
47 D. Dasgupta, Z. Kamar, C. Rochas, M. Dahmani, P. Mesini
and J. M. Guenet, So Matter, 2010, 6, 3573–3581.
48 G. Oster and D. P. Riley, Acta Crystallogr., 1952, 5, 272–
276.
44 P. Terech and Y. Talmon, Langmuir, 2002, 18, 7240–7244.
This journal is ª The Royal Society of Chemistry 2013
Soft Matter