Eur. Phys. J.

E (2011) 34: 4
DOI 10.1140/epje/i2011-11004-1
THE EUROPEAN
PHYSICAL JOURNAL E
Regular Article

Rheological behaviour of polyoxometalate-doped lyotropic

lamellar phases

J.P. de Silva1,a , A.S. Poulos1 , B. Pansu1 , P. Davidson1 , B. Kasmi1 , D. Petermann1 , S. Asnacios2 , F. Meneau3 , and
M. Impéror1
1
Laboratoire de Physique des Solides-UMR 8502-Université Paris-Sud, F-91405 Orsay, France
2
Laboratoire Matière et Systèmes Complexes-10 rue Alice Domon et Lonie Duquet, 75205 Paris, France
3
SOLEIL Synchrotron-L’Orme des Merisiers Saint-Aubin, BP 48 91192 Gif-sur-Yvette, France

Received 29 March 2010 and Received in final form 23 August 2010

Published online: 10 January 2011 – 
c EDP Sciences / Società Italiana di Fisica / Springer-Verlag 2011

Abstract. We study the influence of nanoparticle doping on the lyotropic liquid crystalline phase of the
industrial surfactant Brij30
R (C12 E4 ) and water, doped with spherical polyoxometalate nanoparticles
smaller than the characteristic dimensions of the host lamellar phase. We present viscometry and in situ
rheology coupled with small-angle X-ray scattering data that show that, with increasing doping concentra-
tion, the nanoparticles act to decrease the shear viscosity of the lamellar phase, and that a shear-induced
transition to a multilamellar vesicle “onion” phase is pushed to higher shear rates, and in some cases com-
pletely suppressed. X-ray data reveal that the nanoparticles remain encapsulated within the membranes
of the vesicles, thus indicating a viable method for the fabrication of nanoparticle incorporating organic
vesicles.

1 Introduction transition from lamellae to MLVs; a final shear thinning

regime at high shear rates corresponding to shearing of
The rheological properties of complex fluids is an active the MLV state. It is suggested that lamellar phases flow
field of research: areas of particular interest concern if and under shear through the transportation of defects (such
how such fluids flow, as well as any structural changes as dislocations) through the bulk phase [5,6].
or transitions associated with an imposed shear rate or
In this paper we discuss the rheological proper-
stress. Fine control over the flow of structured fluids is
ties of a lyotropic, liquid crystalline lamellar phase
highly desirable for a wide range of applications, while the
of Brij30/water
R doped with polyoxometalate (POM)
rheological properties of non-Newtonian, structured fluids
nanoparticles derived from phosphotungstic acid [7]. One
is also an important subject, as such fluids often exhibit
interesting property of the POM doped lamellar phases
strongly nonlinear behaviour under shear [1].
that is immediately visible is that the shear viscosity of
Here we consider how the rheological properties of
the phase is significantly reduced with increased doping
a complex fluid, in this case a lamellar phase of water
concentration, as illustrated in fig. 1, where the difference
and nonionic surfactant, are modified by the addition of
in flow under the influence of gravity between doped and
spherical nanoparticles (1.1 nanometers in diameter) that
undoped phases can be visualised simply by tilting the
are smaller than the characteristic dimensions of the host
samples. The extended phase diagram and properties of
lamellar phase. If one considers the rheological proper-
this ternary system have been previously characterised in
ties of dilute surfactant/water lyotropic lamellar phases,
detail [8,9]. Brij30
R surfactant is industrial grade C12 E4
observations of shear-induced alignment, thinning, and
that contains a small amount of homologue impurities, but
formation of multilamellar vesicles (MLVs) of monodis-
is significantly cheaper than pure C12 E4 and is therefore
perse dimensions initiated by buckling instabilities, have
more suitable for wide scale application. It is observed in
been reported by Diat and coworkers using a variety of
the ternary system that the POMs are preferentially lo-
microscopy and scattering techniques [2,3]. In such sys-
cated close to the EO groups of the surfactant membrane
tems one may observe three regimes [4]: an initial shear
rather than remaining dispersed within water between the
thinning regime corresponding to shearing of the lamel-
bilayers, while the interlamellar spacing remains unmodi-
lar phase; a shear thickening regime corresponding to the
fied by the presence of the particles [8]. Previous studies
a
Current address: Département de Physique-Université libre of C12 E4 /water [10] and C10 E3 /water [11] lamellar phases
de Bruxelles, U.L.B CP223, B-1050 Brussels, Belgium. under shear have characterised the shear viscosity and
e-mail: j.desilva@physics.org, imperor@lps.u-psud.fr lamellar-MLV transition in these pure systems. Recent
Page 2 of 9 The European Physical Journal E
Fig. 1. A simple visual illustration of the significant difference
in flow properties for 40/0 and 40/4 doped lamellar phases
under the influence of gravity: (a) initial state and (b) after
10 seconds inclined at approximately 70 degrees. Note also the
higher turbidity of the 40/0 sample.
studies of the effect of temperature on the MLV transition
in nonionic surfactant lamellar phases have been made us-
ing coupled rheology and scattering techniques [12]. Pre-
ceding studies of the rheological properties of doped lamel-
lar phases have considered dopants such as water soluble
polymers [13–15], colloidal particles [16–18] or clays [19]
that are of the same order as or of dimensions much larger
than the interlamellar spacing. Doping with nanoparti-
cles such as POMs, where the nanoparticles are smaller
than the characteristic dimensions of the lamellar phase,
has not been investigated in detail; more precisely, the
1.1 nm diameter POMs considered here are less than the
bilayer thickness (3.4 nm) and about one quarter of the
water channel thickness (4.6 nm) for a lamellar period of
8 nm [8].
2 Experimental
2.1 Sample preparation
The nonionic surfactant polyethylene glycol dodecyl ether,
C12 E4 , forms a lyotropic liquid crystalline phase when
mixed with water. Lamellar phases are prepared using sur-
factant, water and POMs at ratios that place the resul-
tant doped and undoped systems in the Lα lamellar re-
gion of the phase diagram, as previously deduced for the
same experimental system [8,20]. Here we use the com-
mercial surfactant Brij30
R sourced from Sigma-Aldrich
in lieu of pure C12 E4 ; Brij30
R contains an unspecified
small amount of impurities by way of organic homologues.
Nanoparticles are derived from Sigma-Aldrich sourced 12-
phosphotungstic acid, 3H+ [PW12 O40 ]3− ; the POM par-
ticles are in the form of the Keggin structure [7] with a
central phosphor homoatom and a diameter of 1.1 nm. The
surfactant is mixed with water or an aqueous POM solu-
tion of various concentrations in a test tube; samples are
centrifuged in normal and inverted configurations in or-
der to achieve good homogeneity and evacuate trapped
air bubbles that will adversely affect rheological measure-
ments. The resulting birefringent lamellar phases have a
bilayer spacing of between 7 and 11 nm depending on the
surfactant/water ratio, but importantly invariant with re-
spect to POM doping concentration [8]. The samples used
for this study comprise either 40 or 50% C12 E4 volume
fraction with up to 4% POMs by volume in water; these
surfactant/water/POM ratios are chosen to keep the sys-
tem well inside the Lα lamellar region of the phase dia-
gram, and the upper limit on POM doping is imposed by a
phase separation that is observed at POM concentrations
greater than this. Hereafter the samples will be labelled
using the nomenclature of “surfactant volume fraction of
total sample/POM volume fraction of water”; thus 40/1
refers to a sample composed of 40% surfactant by volume
and 1% POMs by volume in water.
2.2 Optical microscopy
Optical observations of the sample texture between
crossed polarisers is achieved using Olympus microscopes
with digital image capture via CCD camera. Optical
grade, flat glass capillaries (2 × 0.2 mm, Vitrocom U.S.A.)
are filled by slowly drawing the lamellar phase into the
capillary, while a temperature stage permits the samples
to be accurately thermostated.
2.3 Laboratory rheology and rheo-microscopy
Laboratory rheology measurements are performed using
Anton-Paar Physica MCR 301 and 501 series rheome-
ters in cone-plate geometry. One and two degree trun-
cated cones (52 μm truncation) of metal and 0.5 degree
cones of glass, both of radius 50 mm, were available. Rheo-
microscopy experiments are performed using a Rheomet-
rics DSR500 rheometer retrofitted with a cross-polarised
optical capture system, using an inverted microscope and
CCD or digital camera allowing in situ capture of the
sheared texture [5]. A truncated glass cone (40 mm in di-
ameter) and plate are roughened using precision sand-
blasting, while two small polished windows are retained
in the cone and plate in order to provide clear optical im-
age capture. This procedure is performed in order to re-
duce slip at the polished surfaces of this geometry. Rheol-
ogy measurements are performed with both imposed shear
rate and stress, and anti-evaporation traps are used where
available; measurements are performed at room tempera-
ture including all laboratory and subsequent synchrotron
RheoSAXS experiments.
2.4 RheoSALS
Rheological small-angle light scattering experiments were
performed at the laboratories of Anton-Paar France us-
ing an AP 501 rheometer fitted with a solid-state laser
(658 nm wavelength) illuminating the sample vertically
down through a glass plate-plate geometry (0.5 mm gap)
between crossed polarisers. Scattering patterns are cap-
tured using a CCD camera and screen. A value of n =
 J.P. de Silva et al.: Rheological behaviour of polyoxometalate-doped lyotropic lamellar phases Page 3 of 9

1.38 is taken for the refractive index of Brij30/water

R
samples [10] where the expression of the scattering vec-
tor is given by q = 4πn sin(θ/2)/λ and θ is the scattering
angle.

2.5 RheoSAXS

RheoSAXS experiments were performed at the SWING

beamline of the Soleil Synchrotron, using a custom-built
concentric cylinder Couette shear cell constructed of poly-
carbonate. A synchrotron light source provides sufficient
beam intensity for kinetic studies, where the time for a
single acquisition is minimised to at most a few seconds
in order to acquire a clear snapshot of a dynamic process
such as this. The polycarbonate cell is 20 mm in diameter
with a gap of 0.5 mm, height of 17 mm and wall thick-
ness of 0.3 mm. As the radius of the interior cylinder is
much greater than that of the gap, a linear approxima-
tion to the velocity gradient across the gap is valid: thus
we derive the shear viscosity η as a function of the mean
shear rate γ̇ = Ω(R22 + R12 )/(R22 − R12 ), and shear stress
σ = M (R22 + R12 )/(4πHR12 R22 ), where R1 and R2 are
the interior and exterior cylinder radii, respectively, H is
the cylinder height, M is the measured torque and Ω is
the angular velocity. The shear viscosity simply relates the
mean shear rate and stress: σ = ηγ̇.
The polycarbonate Couette shear cell is utilised by two Fig. 2. (a) Measured shear viscosity against shear rate for
different experimental set-ups: a custom shear rheometer 40/0, 40/1, 40/2 and 40/4 samples for a sequential series of
that provides a torque measurement is installed onto the constant shear rates in cone-plate geometry, after 2 h at each
beamline, where the shear rate is imposed by controlling shear rate step. In situ optical textures between crossed po-
the speed of rotation of the external cylinder; an identi- larisers are shown for 40/2 sample corresponding to viscosity
cal polycarbonate shear cell has also been fabricated by data before shear, and after the corresponding 10 and 20 s−1
Anton-Paar to permit the Physica 501 controlled stress shear steps. (Scale bar is 50 μm.) (b) Corresponding SALS pat-
rheometer to be installed onto the beamline, giving access terns for the 40/2 sample at the indicated shear rates.
to both imposed shear rate and stress experiments. Con-
trary to the custom shear rheometer, the interior cylinder
is rotated by the Anton-Paar rheometer and thus the sys- 3 Results and discussion
tem is susceptible to Taylor-Couette instabilites at higher
shear rates. Further details of these experimental appa- 3.1 Macroscopic observations
ratus are available [21]. During shear, scattering data at
both radial and tangential positions of the shear cell (with A quantitative global overview of the effect of POM dop-
respect to the flow direction these correspond to a per- ing is illustrated by fig. 2(a) for 40% surfactant lamellar
pendicular and parallel alignment of the incident beam) phases at 0, 1, 2, 3 and 4% POM doping concentrations.
is recorded by horizontally displacing the rheometer with This figure represents a comparison of the measured shear
respect to the incident X-ray beam (horizontal and verti- viscosity of each particular phase during a single imposed
cal spot size approximately 400 × 200 μm) via a motorised shear rate rheology experiment consisting of a continu-
table. The X-ray transmission at both radial and tangen- ous series of constant shear rate steps, each for a duration
tial positions of the cell is accurately modelled taking into of 2 h in the case of these data. The data presented are
account the width of the beam, and data are treated using the value of the shear viscosity at the end of each shear
software provided by the SWING beamline. The typical rate step. A pre-shear protocol at a constant shear rate of
accessible q-range is 4×10−4 –0.55 Å−1 at an X-ray energy 0.1 s−1 for 15 minutes is implemented in order to attain a
of 11 keV, which minimises absorption by the polycarbon- reproducible starting state for each repeat measurement
ate cell. RheoSAXS measurements are limited to a max- without large deformation [22,23]. From fig. 2 one can see
imum of approximately 2 h as the interlamellar spacing that the trend is for the lamellar phase shear viscosity
will be rather sensitive to drying of the sample. Labo- to systematically decrease with increasing POM doping
ratory X-ray experiments indicate that the interlamellar concentration, where the basic form of the flow curve is
spacing remains constant under static conditions within offset towards higher shear rate and lower viscosity. This
the bulk of the Couette cell for at least 10 hours without is in contrast to the influence of adding a small amount of
any form of anti-evaporation measures. ionic surfactant such as sodium dodecylsulfate (SDS) to
Page 4 of 9 The European Physical Journal E

a C12 E4 lamellar phase for example, which results in an in-

crease in shear viscosity [24]. Between 0 and 4% POM dop-
ing the flow curve is displaced by approximately an order
of magnitude in both viscosity and shear rate. The general
form of each individual flow curve is in good agreement
with that observed for pure C12 E4 by Müller et al. [10]:
an initial shear thinning, followed by a shear thickening
region corresponding to a lamellar-MLV transition, and a
final shear thinning regime of the MLV or “onion” phase.
As fig. 2(a) illustrates the lamellar-MLV transition region
is also systematically shifted to higher shear rates with
increasing POM doping concentration, and the origin of
this effect will be discussed in detail later.
Optical micrographs between crossed polarisers of the
lamellar and MLV phase textures are in good agreement
with those observed by Müller et al. [10]: the optical im-
ages presented in fig. 2(a) for the 40/2 sample before and
after shearing at γ̇ = 10 and 20 s−1 closely resemble those
of the C12 E4 lamellar phase and MLV phase, respectively,
where the texture of the MLV phase becomes finer with
increasing shear rate as the close-packed MLVs decrease in
diameter. This behaviour is also supported by the typical
four-lobe SALS patterns observed between crossed polaris-
ers (see fig. 2(b)) for the same 40/2 sample at shear rates
from 10 to 40 s−1 . These SALS patterns are in good agree- Fig. 3. (a) Normalised measured steady-state shear viscosity
ment with those observed for a pure C12 E4 system indicat- and fit as a function of POM area fraction Fpom for 40 and 50%
ing the presence of densely packed MLVs of tens of microns surfactant volume fraction, for imposed shear rate γ̇ = 0.1 s−1 .
in diameter [25]. Here we confirm for the studied system (b) Measured yield stress σy and linear fit as a function of Fpom .
that an MLV phase can be formed in both doped and un- (c) Corresponding instantaneous elastic modulus G0 and linear
fit as a function of Fpom .
doped samples; from these patterns, an average size of a
few tens of microns for the MLV diameter is derived, and
this size decreases with increasing shear rate from about and A are the volume and cross-section of the POM par-
30-40 microns to 20 microns for the 40/2 sample from 10 to ticle, taken as 0.685 nm3 and 0.93 nm2 , respectively [27].
40 s−1 , for example, again consistent with previous studies In fig. 3(a) we present the normalised steady-state shear
of undoped lamellar phases [26]. The important conclusion viscosity of the lamellar phase for 40 and 50% surfactant
to draw from these data is that the classical lamellar-MLV concentration as a function of Fpom , measured for an im-
transition is observed in both doped and undoped sam- posed shear rate γ̇ = 0.1 s−1 . Here we propose a suitable
ples even though it is shifted to higher shear rates with model, where Fpom is the relevant parameter used to de-
increased POM doping concentration. The optical texture scribe the decrease in shear viscosity with increasing dop-
of the doped lamellar phases prior to any shear exhibits ing concentration:
a much finer texture than the classical “mosaic” texture    n 
of a nonoriented lamellar phase [10]. In fact the undoped η Fpom
lamellar phases are consistently visually more turbid than (Fpom ) = 1 − A 1 − exp − , (1)
η0 B
the doped samples, as seen in fig. 1 for instance, which
indicates that POM doping has some effect on the basic where η0 is the zero-shear viscosity of the undoped phase,
macroscopic texture of the prepared lamellar phase prior (1 − A) is the plateau reduced viscosity at high doping
to shear. concentrations and B is the critical value of Fpom . Here
the data for the two different surfactant concentrations
collapse onto the same curve, where the data for both 40
and 50% samples are well fitted by n = 3, A = 0.85 and
3.2 Bulk properties of doped lamellar phases B = 0.092. This result strongly indicates that the effect of
the POMs in reducing the shear viscosity depends purely
This viscosity decrease with increasing POM doping can on the areal density of the POMs adsorbed onto the surfac-
be quantified by examining the effect of POM doping on tant membrane, regardless of the surfactant/water ratio.
the shear viscosity as a function of the POM area fraction It can be seen from fig. 3(a) that there is no further reduc-
of the surfactant membrane, Fpom , for different surfac- tion in viscosity after Fpom ≈ 0.15, which probably corre-
tant concentrations. The POM area fraction is given by sponds to a physical limit on the density of nanoparticles
Fpom = [Aδ(1 − φs )φpom ]/2V0 φs , where φs and φpom are that can be adsorbed onto the surfactant membrane [8].
the surfactant and POM volume fractions, respectively, δ Also presented in fig. 3(b) are yield stress (σy ) mea-
is the surfactant bilayer spacing, taken as 3.4 nm, and V0 surements for the same range of surfactant concentration
 J.P. de Silva et al.: Rheological behaviour of polyoxometalate-doped lyotropic lamellar phases Page 5 of 9

and POM areal fraction; yield stress data are determined

from creep experiments, where a series of increasing shear
stress steps are applied to the sample until a detectable
deformation is recorded. A linear fit to the data shows
that the trend is for the yield stress to decrease with in-
creasing POM areal fraction, falling from around 8 ± 2 Pa
for undoped phases to around 2 ± 1 Pa at Fpom = 0.15.
The values of σy measured here for the undoped phases
are also in good agreement with the values measured by
Paasch et al. [28] for similar undoped lamellar phases. (It
should be noted that the measured value of a yield stress
for a lamellar phase, or indeed even the existence of a yield
stress, has been shown to often depend on the measuring
technique itself, for example [29].)
Finally, the instantaneous elastic modulus G0 is pre-
sented as a function of Fpom in fig. 3(c) for the two surfac-
tant concentrations, determined from creep-recovery ex-
periments, where a constant stress σ0 < σy is applied Fig. 4. Measured shear viscosity against shear rate and fits
to the sample for a time T and then removed, while the for lamellar phases before the onset of an MLV transition for
various POM doping concentrations. Inset: fitted consistency
strain is measured during the stress application and sub-
K as a function of POM volume fraction.
sequent relaxation period. The creep data are well de-
scribed by the Kelvin-Voigt model derived by Panizza et
al. for viscoelastic phases [30], where typical values from
the fits for the 40/4 samples, for example, are of the or-
der of J0 , Jr , Jn ≈ 10−2 Pa−1 , τv ≈ 500 s, τrel ≈ 500 s
and ηm ≈ 105 Pa s. The introduction of POMs tends to
slightly reduce the elastic modulus, as shown in fig. 3(c),
where the instantaneous elastic modulus decreases from
about 300 to 100 Pa by Fpom ≈ 0.15.
The correlated decrease in viscosity, yield stress and
elastic modulus with increasing POM doping suggests that
the POMs act to fluidify the lamellar phase. If one consid-
ers that the yield stress and elastic response of a lamellar
phase are due to defects within the lamellar phase that cre-
ate internal stress [31], we can postulate that the POMs
act to reduce the defect density resulting in the experi-
mentally observed reduction in σy and G0 . Such an ef-
fect would also explain the reduction in turbidity observed
with increasing POM doping concentration (see fig. 1 for
example), corresponding to the measurements made by Fig. 5. Measured shear viscosity against shear rate and fit
Horn et al. relating turbidity to the density of defects for onion phases at various POM doping concentrations. Inset:
within the phase [31]. measured shear viscosity against time for a 40/0 sample at
various imposed shear rates for cone plate (c-p) and Couette
cylinder (cyl) geometries for ascending and descending shear
rate ramps. Dashed line is a guide to the eye.
3.3 Shear rate dependence of lamellar and MLV phase
viscosity

It is now pertinent to examine the shear rate dependence
of the viscosity for these doped phases. The shear thinning
behaviour of both a lamellar and MLV phase can be mod-
elled using the Sisko equation [32]: η(γ̇) = η∞ + K γ̇ −n ,
where η∞ is the infinite shear viscosity, K the consistency
and n the power law fit parameter. We present in fig. 4
shear viscosity against shear rate data for various POM
doping concentrations for the lamellar phase, i.e. strictly
prior to the onset of the lamellar-MLV transition. By con-
trast, in fig. 5 the equivalent data for the MLV phase is
presented for the various POM doping concentrations. The
inset in fig. 5 shows a typical complete rheological curve for
an undoped 40/0 sample, where the classical hysteresis of
a lamellar phase that has undergone a lamellar-MLV tran-
sition when subjected to an ascending and descending se-
ries of constant shear rate steps is observed, cf. Oliviero et
al. [11], for example. This figure also illustrates that data
derived from a polycarbonate concentric cylinder shear
cell (cyl) and metal cone-plate geometries (c-p) are in good
agreement. The data in fig. 4 and 5 are obtained by multi-
ple experiments; again with imposed shear rate steps over
narrow shear rate ranges. It should be noted that, because
the shear history of the sample is different, in contrast to
the data presented in fig. 2 that are from a single experi-
ment over a wide shear rate range, the exact point of the
lamellar-MLV transition may appear to be slightly shifted
Page 6 of 9 The European Physical Journal E

between the two data sets for samples of the same doping
concentration.
Firstly, upon examination of the shear rate dependence
of the viscosity for the lamellar phases, data presented in
fig. 4, it can be seen that the shear thinning behaviour
of the doped and undoped lamellar phases when fitted by
the Sisko equation are all well described by a η ∝ γ̇ −0.5
dependence that is typical of a lamellar phase that flows
by the displacement of defects through the bulk phase [5,
11]. However, it is clear that the viscosity is systematically
decreased by the addition of POMs without affecting this
fundamental behaviour, while the consistency K falls from
around 25 to 3 Pa s−n with 0 to 4% POM doping fraction,
as presented in fig. 4 inset. This behaviour is consistent
with our supposition that the POMs act to fluidify the
lamellar phase via a reduction of defect density.
In fig. 5 we present similar data and analysis for the Fig. 6. Mean critical shear stress against shear rate and fit for
viscosity of the MLV phase for various POM doping con- 40% surfactant samples for a range of doping concentrations.
centrations. Here we see a rather different behaviour from Inset: measured shear stress against shear rate data for a series
of shear rate steps for 40/0 and 40/4 samples.
that of the lamellar phase: all the data fall on the same
curve described by a power law index of n = 1 ± 0.1 from
a fit to the Sisko equation with η∞ ≈ 2 Pa s. This result
there is an invariant (with respect to POM doping concen-
indicates that POM doping has a negligible effect on the
tration) critical stress, σc = 130±20 Pa for 40% surfactant
viscosity of the corresponding MLV phase (note that the
phases, and it is this invariant critical stress that lies at the
viscosity of the MLV phase is typically higher than that of
root of the displacement of the lamellar-MLV transition to
the lamellar phases, K ≈ 200 Pa s−n compared to the 25-
higher shear rates with increased POM doping (see fig. 2):
3 Pa s−n range for the lamellar phases), a rather surprising
the POMs act to systematically reduce the viscosity of the
result considering the profound effect on the viscosity of
lamellar phase but σc remains constant, thus the doped
the lamellar phase seen previously. This η ∝ γ̇ −1 depen-
system must be sheared at a correspondingly higher shear
dence is of a comparable order to that of γ̇ −0.8 observed
rate in order to attain an average stress of σc across the
by Bergenholtz et al. [33] for an AOT/brine system and
sample and induce the lamellar-MLV transition.
γ̇ −0.89 observed by Nettesheim et al. for a Laponite clay
particle doped C12 E4 /water system [26], but greater than
the γ̇ −0.53 dependence observed for a pure C12 E4 /water
system by Müller et al. [10]. The calculation of such expo- 3.5 Microscopic observations
nents depends on the range of shear rates studied and the
uniformity of the sample at high shear rates, where voids SAXS experiments can be employed to shift the focus of
and solvent lubrication layers forming on the cell walls can the study towards the nanoscopic scale structure corre-
result in a measured viscosity lower than the true viscos- sponding to the various observed bulk states and transi-
ity. Here we may also see an effect of the small amounts tions. Presented in fig. 7 are rheological, SAXS and optical
of Brij30
R impurities versus pure C12 E4 , as the undoped data for a 50% surfactant system at two different doping
phase is also affected by the same discrepancy. concentrations for a range of sequentially increasing shear
rate steps, obtained through a combination of RheoSAXS
and rheo-optical experiments. In fig. 7 we present data for
3.4 Critical shear stress for MLV formation 50/2 and 50/4 samples sheared at 1, 4, 8, 10, and 15 s−1
for 10 minutes per step in the Couette cylinder geome-
Presented in fig. 6 is the mean shear stress, σm , against try of the RheoSAXS instruments. For the 50/4 sample
the shear rate during which the lamellar-MLV transition comparative data from glass cone-plate geometry is pre-
occurs for a range of POM doping concentrations. σm sented in order to give an indication of the consistency
is the average stress recorded during the lamellar-MLV between data from the two different measurement sys-
transition over the total time of the imposed shear rate tems [35]. Also presented are optical micrographs between
step, where the error bar gives an indication of the to- cross-polarisers of the sample texture and corresponding
tal stress range during the shear step. Typical measured X-ray scattering patterns at selected shear rates, in this
shear stress against shear rate data are shown in the in- case at the tangential position for the 50/2 sample and
set of fig. 6, where the curves for 40/0 and 40/4 samples the radial position for the 50/4 sample.
exhibit a similar form that is offset in shear rate by an The most striking feature of this data is that the 50/2
order of magnitude. The form of these data is consistent samples undergo a lamellar-MLV transition at approxi-
between doped and undoped samples and similar to pre- mately 8 s−1 while the 50/4 sample continues to exhibit
vious studies such as those of cationic lamellar phases for the shear thinning behaviour of a lamellar phase up to
example [34]. A zero-order fit to the data indicates that shear rates of 15 s−1 . If one estimates the critical stress
 J.P. de Silva et al.: Rheological behaviour of polyoxometalate-doped lyotropic lamellar phases
Fig. 7. Measured shear viscosity against time for 50/2 and
50/4 samples at shear rates of 1, 4, 8, 10 and 15 s−1 at 10 min-
utes per shear step (15 min pre-shear data not shown). Cor-
responding cross-polarised optical textures and comparative
rheological data derived from glass cone-plate geometries un-
der identical shear conditions. (Scale bar is 100 microns, arrow
indicates direction of shear.) Corresponding SAXS spectra are
acquired at the indicated cell position.
for the 50% surfactant systems to be about 160–200 Pa,
then at a shear rate of 20 s−1 only approximately 20 Pa
of stress is generated across the sample; thus a vastly in-
creased shear rate, probably well above that practically
possible before the sample is ejected from the shear cell,
would be required to reach the transition threshold. Such
suppression of MLV formation is similar to that observed
by Berghausen et al. upon the addition of 0.5% poly(N-
isopropyl acrylamide) (PNIPAm) to an SDS/water lamel-
lar phase [15].
Optical micrographs for the 50/4 sample at 1 s−1 are
representative of a disoriented lamellar phase, and the cor-
responding SAXS spectra exhibit typical isotropic, narrow
Bragg rings corresponding to the interlamellar repeat dis-
tance and indicating no preferential alignment in the sam-
ple; here the first two Bragg orders are clearly visible by
eye. The broad “halo” signal towards the extremities of the
pattern is due to strong scattering from the POM parti-
cles. As the shear rate increases to 10 s−1 the 50/2 sample
has already formed an MLV phase, where a characteris-
tic uniform texture is seen in the optical micrograph and
the corresponding SAXS pattern remains isotropic in both
tangential (pictured) and radial cell positions, as the MLV
phase is isotropic. It is interesting to note that the Bragg
rings observed for the 50/2 sample MLV phase are slightly
broadened and displaced from those of the corresponding
lamellar phase prior to any lamellar-MLV transition, and
the origin of this effect will be discussed in detail below.
Page 7 of 9
Fig. 8. Radially averaged SAXS data for 40/0 and 40/2 sam-
ples at shear rates pre- and post-transformation from a lamellar
to MLV phase (radial cell position).
Up to a shear rate of 15 s−1 the 50/4 sample contin-
ues to shear align, showing fine graining of the defects
in the lamellar phase into streaks parallel to the flow di-
rection, similar in structure to that observed in shear-
aligned C12 E5 /water lamellar phases [36] or AOT/brine
and SDS/oil/water lamellar phases [37]. This is also con-
firmed by SAXS data that shows an anisotropic enhance-
ment attributed to the scattering by the lamellae inside
defects oriented under the flow. This enhancement must
be attributed to the alignment of defects, as in the radial
cell position the lamellae oriented by the flow (with their
normal along the X-ray beam) do not contribute to the
scattering.
3.6 Vesicle formation in doped phases
One can more closely study the kinetics of the lamellar-
MLV structural transformation using in situ rheoSAXS
data. Radially averaged SAXS data for 40/0 and 40/2
samples at shear rates of 1 and 10 s−1 are presented in
fig. 8, where the data are representative of the lamellar
phase and MLV phases respectively; the peaks correspond
to the first and second Bragg orders and the strong scat-
tering profile from the POM particles is once again visible
in the 40/2 sample data. Future detailed line shape anal-
ysis of such data should provide some insight into the ef-
fect of adsorbed POMs on the bilayer bending moduli [38,
39]; here we focus only on the Bragg peaks in the region
0.06 < q < 0.2 Å−1 because the polycarbonate cell scat-
ters significantly at small q. The characteristic features of
X-ray data from doped and undoped phases have been
previously described in detail [8], where one finds that
the lamellar period is invariant with POM doping and
the POMs preferentially adsorb onto the surface of the
surfactant bilayers, as indicated by the change in relative
intensities of the first- and second-order Bragg peaks with
POM doping.
Page 8 of 9 The European Physical Journal E
With regards to the lamellar-MLV transition the X-
ray data reveal two important points: upon forming an
MLV state there is a broadening and shift to higher q of
the Bragg peaks from that of the pre-transition lamellar
phase, consistently around six percent for the first and
second Bragg orders. Pre-transition first and second or-
der Bragg peaks that are relatively sharp can be seen for
both the 40/0 and 40/2 samples, giving a well-defined in-
terlamellar spacing of around 8 nm for both the 40/0 and
40/2 samples at γ̇ = 1 s−1 . In the MLV state the peak cen-
tre shifts towards larger q, indicating a reduced lamellar
spacing, possibly a result of the compression or collapse
of the structure within the vesicle or the expulsion of sol-
vent [25,40,41]. It is important to note that this is not the
effect of drying in the sample over the duration of the ex-
periment, as SAXS data from lamellar phases sheared for
many hours at a constant shear rate below the lamellar-
MLV shear transition show no change in interlayer spacing
from that of the lamellar phase.
We derive a coherence length from the FWHM of the
broadened Bragg peak in the MLV state (correcting for the
experimental resolution) of around 30 lamellae, or around
250 nm, which shows that the X-rays do not probe the
total size of the MLV (20–40 microns or more than 3000
lamellae) but rather some smaller coherent length scale
with a reduced lamellar spacing, perhaps the result of im-
perfect stacking in the multilayers where several coherent
domains are separated by less ordered regions. The sec-
ond important point derived from data in fig. 8 is that the
nanoparticles remain encapsulated within the surfactant
layers of the MLVs, as the relative intensities of the peaks
are preserved pre- and post-transformation between the
doped and undoped samples, indicating that the nanopar-
ticles remain adsorbed to the surfactant membranes after
the lamellar-MLV phase transition. This result implies a
viable a method of encapsulating nanoparticles within in-
expensive and relatively easily formed surfactant vesicles,
which may have a wide variety of applications including
biological payload delivery.
In fig. 9 we present the interlamellar spacing, d, as a
function of shear rate for a range of POM doping frac-
tions, as derived from radially averaged rheoSAXS data.
At low shear rates in a lamellar state, d is well defined
for all POM doping concentrations at 8.05 nm. At the on-
set of the lamellar-MLV transition d begins to decrease in
all cases, reaching a steady 7.4 nm for the undoped sam-
ple but increasingly smaller values for the doped phases,
down to around 6.5 nm for the 4% POM doped sample.
The origin for such a change in compression with POM
doping is unknown; surfactant vesicles have already been
shown to deform into polyhedral shapes that fill available
space without excess space-filling solvent [42]. Such an ef-
fect may also be responsible for the reduced length scales
derived from X-ray experiments, as discussed previously.
Moreover, it was observed that after cessation of shear the
interlamellar spacing may increase again at rest, indicat-
ing the reabsorption of solvent and the swelling or rupture
of the vesicle phase, consistent with freeze-fracture elec-
tron microscopy measurements on AOT/brine phases [42].
Structural differences in doped and undoped vesicle phases
Fig. 9. Interlamellar spacing d derived from rheoSAXS data
as a function of shear rate and POM doping concentration.
could be revealed by electron microscopy experiments, for
example [42,34].
4 Conclusions
We have studied the influence of [PW12 O40 ]3− poly-
oxometalate nanoparticle doping on the rheological prop-
erties of a Brij30/water
R lyotropic lamellar phase. We
show that the shear viscosity, yield stress and instanta-
neous elastic modulus of the lamellar phase are all sys-
tematically decreased with increased POM doping frac-
tion. A shear-induced transition to a multilamellar vesicle
phase can be offset to higher shear rates, or even com-
pletely suppressed, with increasing POM doping fraction.
We measure a critical stress of around 130 Pa for the
lamellar-MLV transition in the 40% surfactant systems,
which is invariant with respect to POM doping fraction;
the constant nature of this critical stress causes a shift
of the lamellar-MLV transition to higher shear rates with
increasing POM doping. We show that the areal density
of POMs adsorbed to the surfactant bilayer is the key pa-
rameter required to quantitatively describe the viscosity
variation with POM doping. From the decrease in tur-
bidity, elastic modulus and yield stress, we postulate that
POMs act to suppress defects within the lamellar phase.
X-ray data shows that the nanoparticles remain encapsu-
lated within the MLV membranes, thus indicating a rather
inexpensive methodology for the encapsulation of metallic
nanoparticles within surfactant vesicles.
JPdS thanks “Triangle de la Physique” for a postdoctoral fel-
lowship (under project RheoSAXS), C. Baravian, I. Bihan-
nic, L. Michot, E. Paineau and D. Constantin for help with
RheoSAXS experiments and useful discussions. ASP benefited
from a Marie Curie fellowship under contract no. MEST-CT-
2004-514307. We thank D. Langevin for the use of an AP MCR
301 rheometer, Anton-Paar France for RheoSALS experiments
and the impromptu loan of an air-compressor, Laboratoire
 J.P. de Silva et al.: Rheological behaviour of polyoxometalate-doped lyotropic lamellar phases Page 9 of 9

FAST for the use of a glass rheometer measurement system seen in the C12 E4 phase diagram [20] and confirmed by ob-
and SOLEIL for beamtime allocation. We thank the referees servations of cross-polarised optical texture after a quench
for pertinent comments that improved the manuscript. to 0 ◦ C.
23. B. Medronho, M. Miguel, U. Olsson, Langmuir 23, 5270
(2007).
References 24. F. Nettesheim, J. Zipfel, P. Lindner, W. Richtering, Col-
loids Surf. A 183-185, 563 (2001).
25. O. Diat, D. Roux, F. Nallet, Phys. Rev. E 51, 3296 (1995).
1. R.G. Larson, The Structure and Rheology of Complex Flu-
26. F. Nettesheim, J. Zipfel, U. Olsson, F. Renth, P. Lindner,
ids (Oxford University Press, New York, 1999) pp. 3-4.
W. Richtering, Langmuir 19, 3603 (2003).
2. O. Diat, D. Roux, J. Phys. II 3, 9 (1993).
27. A.S. Poulos, PhD thesis (2009).
3. O. Diat, D. Roux, F. Nallet, J. Phys. II 3, 1427 (1993).
28. S. Paasch, F. Schambil, M.J. Schwuger, Langmuir 5, 1344
4. D. Roux, F. Nallet, O. Diat, Europhys. Lett. 1, 53 (1993).
(1989).
5. C. Meyer, S. Asnacios, M. Kléman, Eur. Phys. J. E 6, 245
29. J. Munoz, C. Gallegos, V. Flores, J. Disp. Sci. Technol. 7,
(2001).
453 (1986).
6. P. Oswald, M. Kléman, J. Phys. (Paris) Lett. 12, L411
30. P. Panizza, D. Roux, V. Vuillaume, C.-Y.D. Lu, M.E.
(1982).
Cates, Langmuir 12, 248 (1996).
7. J.F. Keggin, Proc. R. Soc. London, Ser. A 144, 75 (1934).
31. R.G. Horn, M. Kleman, Ann. Phys. (Paris) 3, 229 (1978).
8. A.S. Poulos, D. Constantin, P. Davidson, M. Impéror, B.
32. P. Versluis, J.C. van de Pas, J. Mellema, Langmuir 13,
Pansu, P. Panine, L. Nicole, C. Sanchez, Langmuir 24,
5732 (1997).
6285 (2008).
33. J. Bergenholtz, N.J. Wagner, Langmuir 12, 3122 (1996).
9. A.S. Poulos, D. Constantin, P. Davidson, M. Impéror, P.
34. P. Partal, A.J. Kowalski, D. Machin, N. Kiratzis, M.G.
Judeinstein, B. Pansu, J. Phys. Chem. B 114, 220 (2010).
Berni, C.J. Lawrence, Langmuir 17, 1331 (2001).
10. S. Müller, C. Börschig, W. Gronski, C. Schmidt, D. Roux,
35. Shear viscosities measured here for the pure undoped
Langmuir 15, 7558 (1999).
lamellar phase at 40% surfactant concentration are around
11. C. Oliviero, L. Coppola, R. Gianferri, I. Nicotera, U. Ols-
an order of magnitude higher than those given by Müller
son, Colloids Surf. A 228, 85 (2003).
et al. for a pure 40% C12 E4 lamellar phase at equivalent
12. Y. Kosaka, M. Ito, Y. Kawabata, T. Kato, Langmuir 26,
shear rates [10]. We can identify two reasons for this: previ-
3835 (2010).
ous experiments demonstrate that there is a clear effect of
13. H.E. Warriner, S.H. Idziak, N.L. Slack, P. Davidson, C.R.
the nature of the cell material and surface roughness [43];
Safinya, Science 271, 969 (1996).
the second is the possible influence of impurities, however
14. S.L. Keller, H.E. Warriner, C.R. Safinya, J.A. Zasadzinski,
this is less likely to explain the observed discrepancy. Our
Phys. Rev. Lett. 78, 4781 (1997).
measurements indicate that this discrepancy is due to the
15. J. Berghausen, J. Zipfel, P. Lindner, W. Richtering, J.
use of smooth glass shear cells rather than metal or poly-
Phys. Chem. B 105, 11081 (2001).
carbonate; a 40% Brij30 R sample sheared with a smooth
16. V. Ponsinet, P. Fabre, J. Phys. II 6, 955 (1996).
glass geometry very closely reproduces the viscosities mea-
17. J. Arrault, C. Grand, W.C.K. Poon, M.E. Cates, Euro-
sured by Müller et al.
phys. Lett. 38, 625 (1997).
36. C.-Y.D. Lu, P. Chen, Y. Ishii, S. Komura, T. Kato, Eur.
18. Y. Suganuma, M. Imai, K. Nakaya, J. Appl. Crystallogr.
Phys. J. E 25, 91 (2008).
40, s303 (2007).
37. L. Courbin, P. Panizza, Phys. Rev. E 69, 021504 (2004).
19. F. Nettesheim, I. Grillo, P. Lindner, W. Richtering, Lang-
38. J.T. Brooks, C.M. Marques, M.E. Cates, J. Phys. II 1, 673
muir 20, 3947 (2004).
(1991).
20. D.J. Mitchell, G.J.T. Tiddy, L. Waring, T. Bostock, M.P.
39. M.-F. Ficheux, A.-M. Bellocq, F. Nallet, Eur. Phys. J. E
McDonald, Faraday Trans. I 79, 975 (1983).
4, 315 (2001).
21. J.P. de Silva, D. Petermann, B. Kasmi, M. Imperor, P.
40. A. Lutti, P.T. Callaghan, Eur. Phys. J. E 24, 129 (2007).
Davidson, B. Pansu, F. Meneau, J. Perez, E. Paineau, I.
41. A. Leon, D. Bonn, J. Meunier, J. Phys.: Condens. Matter
Bihannic, L. Michot, C. Baravian, J. Phys. Conf. Ser. 247,
14, 4785 (2002).
012052 (2010).
42. T. Gulik-Krzywicki, J.C. Dedieu, D. Roux, C. Degert, R.
22. We use a pre-shear treatment rather than a quench
Laversanne, Langmuir 12, 4668 (1996).
into the low-temperature isotropic region of the phase
43. N. Jager-Lézer, J.F. Tranchant, V. Alard, J. Doucet, J.L.
diagram [10], as at 50% surfactant concentration there is
Groissiord, J. Rheol. 42, 417 (1998).
no low-temperature lamellar-isotropic phase transition, as