Langmuir 2007, 23, 3637-3645

3637

Hexosome and Hexagonal Phases Mediated by Hydration and

Polymeric Stabilizer
Idit Amar-Yuli, Ellen Wachtel, Einav Ben Shoshan, Dganit Danino,
Abraham Aserin, and Nissim Garti*,
Casali Institute of Applied Chemistry, The Hebrew UniVersity of Jerusalem, Jerusalem 91904, Israel,
Faculty of Chemistry, The Weizmann Institute of Science, RehoVot 76100, Israel, and Department of
Biotechnology and Food Engineering, Technion-Israel Institute of Technology, Haifa 32000, Israel
ReceiVed September 28, 2006. In Final Form: January 10, 2007
In this research, we studied the factors that control formation of GMO/tricaprylin/water hexosomes and affect their
inner structure. As a stabilizer of the soft particles dispersed in the aqueous phase, we used the hydrophilic nonionic
triblock polymer Pluronic 127. We demonstrate how properties of the hexosomes, such as size, structure, and stability,
can be tuned by their internal composition, polymer concentration, and processing conditions. The morphology and
inner structure of the hexosomes were characterized by small-angle X-ray scattering, cryo-transmission electron
microscope, and dynamic light scattering. The physical stability (to creaming, aggregation, and coalescence) of the
hexosomes was further examined by the LUMiFuge technique. Two competing processes are presumed to take place
during the formation of hexosomes: penetration of water from the continuous phase during dispersion, resulting in
enhanced hydration of the head groups, and incorporation of the polymer chains into the hexosome structure while
providing a stabilizing surface coating for the dispersed particles. Hydration is an essential stage in lyotropic liquid
crystal (LLC) formation. The polymer, on the other hand, dehydrates the lipid heads, thereby introducing disorder
into the LLC and reducing the domain size. Yet, a critical minimum polymer concentration is necessary in order to
form stable nanosized hexosomes. These competing effects require the attention of those preparing hexosomes. The
competition between these two processes can be controlled. At relatively high polymer concentrations (1-1.6 wt %
of the total formulation of the soft particles), the hydration process seems to occur more rapidly than polymer adsorption.
As a result, smaller and more stable soft particles with high symmetry were formed. On the other hand, when the
polymer concentration is fixed at lower levels (<1.0 wt %), the homogenization process encourages only partial
polymer adsorption during the dispersion process. This adsorption is insufficient; hence, maximum hydration of the
surfactant head group is reached prior to obtaining full adsorption, resulting in the formation of less ordered hexosomes
of larger size and lower stability.

Introduction

* To whom correspondence should be addressed. Tel: 972-2-658-6574/

5. Fax: 972-2-652-0262. E-mail; garti@vms.huji.ac.il.
The Hebrew University of Jerusalem.
The Weizmann Institute of Science.
Technion-Israel Institute of Technology.

characteristic is of great importance in applications where the

mesophase structure must remain intact in dilute aqueous media.
The large interfacial area, the three- and two-dimensional
symmetry of the cubic and hexagonal LLC phases with both
hydrophobic and hydrophilic domains, respectively, can provide
a complex diffusion pathway for controlled release of entrapped
molecules.9,10 Due to these exceptional physical and chemical
properties, study of the formation of liquid crystal dispersions
has been stimulated. Thus, one can take advantage of the complex
phases for controlled delivery, while simultaneously having the
benefit of the low viscosity of a system of soft nanoparticles
dispersed in a continuous aqueous phase. The preparation of
stable colloidal dispersions of the cubic and hexagonal liquid
crystalline phases, cubosomes and hexosomes, respectively, has
opened exciting new opportunities for applications of lyotropic
liquid crystals.11-13
The original dispersion agents for soft nanoparticles were bile
salts and caseins, which were assumed to form a lamellar envelope
on the surface of GMO-based cubosomes.1,14 A few years later
it was discovered that amphiphilic block copolymers, especially

(1) Larsson, K. J. Phys. Chem. 1989, 93, 7304-7314.

(2) Larsson, K. K.; Fontell, K.; Krog, N. Chem. Phys. Lipids 1980, 27, 321328.
(3) Qui, H.; Caffrey, M. Biomaterials 2000, 21, 223-234.
(4) Misquitta, Y.; Caffrey, M. Biophys. J. 2001, 81, 1047-1058.
(5) Caffrey, M. Biochemistry 1987, 26, 6349-6363.
(6) Frank, C.; Sottmann, T.; Stubenrauch, C.; Allgaier, J.; Strey, R. Langmuir
2005, 21, 9058-9067.
(7) Pena dos Santos, E.; Tokumoto, M. S.; Surendran, G.; Remita, H.; Bourgaux,
C.; Dieudonne, P.; Prouzet, E.; Ramos, L. Langmuir 2005, 21, 4362-4369.
(8) Mezzenga, R.; Meyer, C.; Servais, C.; Romoscanu, A. I.; Sagalowicz, L.;
Hayward, R. C. Langmuir 2005, 21, 3322-3333.

(9) Norling, T.; Lading, P.; Engstrom, S.; Larsson, K.; Krog, N.; Nissen, S.
S. J. Clin. Perodontol. 1992, 19, 687-692.
(10) Nielson, L. S.; Schubert, L.; Hausen, J. Eur. J. Pharm. Sci. 1998, 6,
231-239.
(11) Drummond, C. J.; Fong, C. Curr. Opin. Colloid Interface Sci. 2000, 4,
449-456.
(12) Gustafsson, J.; Ljusberg-Wahren, H.; Almgren, M.; Larsson, K. Langmuir
1997, 13, 6964-6971.
(13) Lopes, L. B.; Collett, J. H.; Bently, M. V. L. B. Eur. J. Pharm. Biopharm.
2005, 60, 25-30.
(14) Buchein, W.; Larsson, K. J. Colloid Interface Sci. 1987, 117, 582-583.

Polar lipids and certain surfactants are known to form

thermodynamically stable lyotropic liquid crystals (LLC) when
mixed with water. In particular, the phase diagram of glycerol
monooleate (GMO) and water as a function of temperature has
been extensively investigated. Several polymorphic mesophases
have been detected and studied, and some seem to have potential
application in industrial products.1-8 GMO can form at least
four distinct and well-characterized liquid crystalline mesophases: a lamellar phase, two reversed bicontinuous cubic phases,
Q230 and Q224, that form at room temperature, and a reversed
hexagonal phase that is obtained at relatively high temperature.1-5
A particularly attractive feature of GMO is the existence of
equilibrium of these phases with excess aqueous solution. This

10.1021/la062851b CCC: $37.00 2007 American Chemical Society

Published on Web 03/01/2007

3638 Langmuir, Vol. 23, No. 7, 2007

F127 (with 99 units of poly(ethylene oxide) (PEO) in each of

the two tails and 67 units of poly(propylene oxide) (PPO), PEO99PPO67PEO99), can provide very powerful steric stabilization.15-21
The formation of stable hexosomes was assumed to be related
to the specific and unique adsorption of F127 on the surface of
the particles. Furthermore, the observation of vesicular structures
at the periphery of the particles in the dispersions suggested that
further stabilization was provided by coexisting lamellar
structures. Simple magnetic stirrer agitation, followed by heat
treatment during the addition of the polymer, yielded a wide
range of particle sizes from 1 to 100 m.21 High shear
homogenization techniques that were later adopted formed soft
particles of 100-500 nm.12,16-20,22,23 Although several manufacturing methods have been developed for cubic and hexagonal
phase dispersions, there is still no validated process available for
controlling the properties and quality of hexagonal phase
dispersions and there is no clear understanding of the factors that
govern the formation and structure of the hexosomes.
Earlier studies of the GMO/F127/water phase diagram have
led to the selection of preferred lipid to polymer weight ratios
to produce cubosome particles.15 It was recognized that the
polymer caused phase transformations among cubic phases. Upon
addition of polymer, up to 80/20 weight ratio of GMO/F127,
diamond (Q224) and gyroid (Q230) cubic phases, which were
formed in its absence, transformed into the primitive cubic phase
Q229. With 70-90 wt % water, lamellar and sponge phases were
formed. This phase diagram indicates that, in order to produce
stable cubosome particles, the GMO/F127 weight ratio should
be kept in the range of about 94/6-80/20 where the primitive
cubic phase or a mixture of the primitive cubic, sponge, or lamellar
phases are formed in the presence of excess aqueous solution.15
The cryo-transmission electron microscope (cryo-TEM) images
of cubosomes confirm the coexistence of soft cubic particles
with a lamellar envelope and multilamellar vesicles.1,14
Although, as indicated above, the formation of cubosome
particles in the GMO/F127/water system has been studied,15-21
the effect of the polymer on the phase behavior of the hexagonal
liquid crystal phase and the resulting hexosome particles has not
yet been clarified. Earlier, we presented the phase diagrams of
GMO/triacylglycerol/water and demonstrated that tricaprylin was
the most effective compound in inducing the formation of reverse
hexagonal liquid crystals.24 A large region of reverse hexagonal
phase was detected even with a minimum tricaprylin concentration
of 3 wt % on the dilution line 96/4 at 20-25 wt % water and
up to a maximum concentration of 13.5 wt % tricaprylin on line
85/15, at 10-20 wt % water. At higher water contents (above
35-40 wt %), beyond full lipid head group hydration, a stable
reverse hexagonal phase was formed in equilibrium with excess
water.24 It should be noted that at relatively low water content
(<20 wt %), the hexagonal phase was relatively less ordered,
while at higher water content, the hexagonal phase was more
(15) Landh, T. J. Phys. Chem. 1994, 98, 8453-8467.
(16) Nakano, M.; Teshigawara, T.; Sugita, A.; Leesajakul, W.; Taniguchi, A.;
Kamo, T.; Matsuoka, H.; Handa, T. Langmuir 2002, 18, 9283-9288.
(17) Nakano, M.; Sugita, A.; Matsuoka, H.; Handa, T. Langmuir 2001, 17,
3917-3922.
(18) Monduzzy, M.; Ljusberg-Wahren, H.; Larsson, K. Langmuir 2000, 16,
7355-7358.
(19) Neto, C.; Aloisi, G.; Baglioni, P.; Larsson, K. J. Phys. Chem. B 1999,
103, 3896-3899.
(20) Larsson, K. J. Dispersion Sci. Technol. 1999, 20, 27-34.
(21) Kamo, T.; Nakano, M.; Leesajakul, W.; Sugita, A.; Matsuoka, H.; Handa,
T. Langmuir 2003, 19, 9191-9195.
(22) Spicer, P. T.; Hyden, K. L.; Lynch, M. L.; Boateng, A. O.; Burns, J. L.
Langmuir 2001, 17, 5748-5756.
(23) Barauskas, J.; Johnsson, M.; Johnsson, M.; Joabsson, F.; Tiberg, F.
Langmuir 2005, 21, 2569-2577.
(24) Amar-Yuli, I.; Garti, N. Colloids Surf. B 2005, 43, 72-82.

Amar-Yuli et al.

well ordered. The differences in lattice parameter and domain

size will be discussed below.
In this study, we demonstrate how hexosome properties such
as size, structure, and stability can be tuned by the polymer
concentration as well as by processing conditions that affect the
competition between hydration and adsorption of the polymer
on the hexosomes. The inner structure and morphology of the
well-organized hexosomes in mixtures of GMO/tricaprylin/water
in the presence of F127 were characterized by small-angle X-ray
scattering (SAXS) and cryo-TEM. Dynamic light scattering (DLS)
and the LUMiFuge technique were used to study particle size
distribution and stability. These hexosomes can be utilized for
topical applications and possibly for preparation of a patch applied
to the skin (will be published in the future).
Experimental Section
Materials. Monoolein, GMO, and distilled glycerol monooleate
that consists of 97.1 wt % monoglyceride and 2.9 wt % diglyceride
(acid value 1.2, iodine value 68.0, melting point 37.5 C, and free
glycerol 0.4%) were obtained from Riken (Tokyo, Japan). It should
be noted that when other manufacturers or batches of GMOs were
used similar results were obtained. Tricaprylin (97-98%) was
obtained from Sigma Chemical Co. (St. Louis, MO). Pluronic F127
(with 99 units of PEO on each of the two tails and 67 units of PPO,
PEO99PPO67PEO99) was a gift from BASF (Ludwigshafen, Germany).
Water was double distilled. All ingredients were used without further
purification.
Hexosome Preparation. The method used to prepare hexosomes
has two stages. First, the hexagonal liquid crystal was formed by
mixing all the components while heating to 70 C in sealed tubes
under nitrogen (to avoid oxidation of the GMO). Next, emulsification
was performed at room temperature by addition of the hexagonal
liquid crystalline mesophase (5 wt %) to the external aqueous phase
containing the polymer (F127). Mixing and shearing were done
with the Ultra Turrax homogenizer (Ultra Turrax, model T25, Janke
& Kunkel, IKA Labortechnik, Staufen, Germany) at 9500 rpm for
10-20 min at room temperature.
Light Microscopy. The samples were inserted between two glass
microscope slides and were observed with a Nikon light microscope
equipped with cross-polarizers and attached to a video camera and
monitor. The samples were analyzed at room temperature.
Small-Angle X-ray Scattering. Scattering experiments were
performed using Ni-filtered Cu KR radiation (0.154) from an Elliott
rotating anode X-ray generator that operated at a power rating of
1.2 kW. The X-radiation was further monochromated and collimated
by a single Franks mirror and a series of slits and height limiters
and measured by a linear position-sensitive detector. The samples
were held in 1.5-mm quartz X-ray capillaries inserted into a copper
block sample holder. The temperature was maintained at T (0.5 C
with a recirculating water bath. The camera constants were calibrated
using anhydrous cholesterol. The scattering patterns were desmeared
using the Lake procedure implemented in home written software.25
In order to estimate a lower bound for the sizes of ordered domains
(LH), the full width at half-height of the (10) diffraction peak was
measured and this value inserted into the Scherrer formula.15
Cryogenic Transmission Electron Microscopy. Samples were
equilibrated at 25 C either in the Controlled Environment Vitrification System (CEVS) or in the Vitrobot for 20 min, in the presence
of water and ethanol, to avoid evaporation of volatile components
during specimen preparation. Vitrified specimens were prepared on
400-mesh copper grids coated with a perforated Formvar film (Ted
Pella). A small drop (5-8 L) was applied to the grid and blotted
with filter paper to form a thin liquid film of solution. The blotted
sample was immediately plunged into liquid ethane at its freezing
point (-196 C). The procedure was performed manually in the
CEVS, and automatically in the Vitrobot. The vitrified specimens
were transferred into liquid nitrogen for storage. Some samples were
(25) Lake, J. A. Acta Crystallogr. 1967, 23, 191-194.

Hexosome and Hexagonal Phases

Langmuir, Vol. 23, No. 7, 2007 3639

Table 1. Structural Properties of the Less Ordered and More
Ordered (15 and 20 wt % Water Concentrations) Hexagonal
Structures upon Addition of an Amphiphilic Block Copolymer
(F127)
SAXS
lipid/polymer
wt ratio
(wt % polymer)

a (),
(0.5

100/0
(0 wt % polymer)
95/5
(4.25 wt % polymer)
90/10
(8.5 wt % polymer)
85/15
(12.75 wt % polymer)

Figure 1. X-ray scattering profiles of the GMO/tricaprylin 90/10

mixture with 10, 15, and 25 wt % water.
examined in a Philips CM120 transmission electron microscope,
operated at 120 kV, using an Oxford 3500 cryoholder maintained
below -178 C. Images were recorded on a Gatan 791 MultiScan
cooled charge-coupled device (CCD) camera. Other samples were
studied using a Philips Tecnai 12 G2 TEM, at 120 kV with a Gatan
cryoholder maintained below -173 C, and images were recorded
on an Ultrascan 1000 2k 2k CCD camera. In both microscopes,
images were recorded with the Digital Micrograph software package,
at low-dose conditions, to minimize electron beam radiation damage.
Brightness and contrast enhancement were done using Adobe
Photoshop 7.0 ME package.
Differential Scanning Calorimetry (DSC). A Mettler Toledo
DSC822 measuring model system was used. The heating rate of
calibration was 10 K min-1. The DSC measurements were carried
out as follows: 5-15 mg of hexagonal LLC samples were weighed,
using a Mettler M3 microbalance, in standard 40-L aluminum pans
and immediately sealed by a press. The samples were rapidly cooled
in liquid nitrogen from 30 to -20 C, at a rate of 10 C min-1. The
samples remained at this temperature for 20 min and then were
heated at 1 C min-1 to 12 C (for hexagonal LC formation), where
they remained for 10 min. Finally, the sample was further heated
to 50 C at the same rate. An empty pan was used as a reference.
The instrument determined the fusion temperatures of the solid
components and the total heat transferred in any of the observed
thermal processes. The enthalpy change associated with each thermal
transition was obtained by integrating the area of the relevant DSC
peak. DSC temperatures and enthalpies reported here were reproducible to (0.2 C and (2%, respectively.
LUMiFuge. The stability of the hexosomes was examined with
a LUMiFuge 114 (Gesellschaft fur Labor, Berlin, Germany), a
microprocessor-controlled analytical centrifuge that detects a large
variety of demixing phenomena (floating and clarification) of the
dispersed systems during centrifugation over the whole sample length
(25 mm). Centrifugation at 300-3000 rpm, which corresponds to
12-1200g (gravimetric factor), results in an accelerated migration
of the dispersed particles. Local variations in particle concentration
are detected due to changes in light transmission. A graphic
representation of transmission as a function of position constitutes
the transmission profile. Zones of well-mixed, concentrated dispersions will scatter and absorb the light; hence, the transmission will
be low. In contrast, any local clarification will raise the transmission.
In this work, centrifugation was at 3000 rpm and samples were
analyzed at intervals of 10 s for a total time of 3 h at room temperature.
Dynamic Light Scattering. The hexosome droplet size was
determined by Zeta-sizer model ZEN1600 (Malvern Instruments,
Worcestershire, England). One drop (0.02-0.03 mL) of the

100/0
(0 wt % polymer)
95/5
(4 wt % polymer)
90/10
(8 wt % polymer)
85/15
(12 wt % polymer)

DSC

LH (),
(50

15 wt % Water
52.3
772
49.5a

481a

water fusion
peaks (oC)
(0.2 C
0.20
-2.7

w/o micellar phase

none

w/o micellar phase

none

20 wt % Water
55.4
1014

0.2

0.3

48.2

647

-1.2

0.2

48.5a

554a

-2.7

0.3

w/o micellar phase

-6.0

a
Although only one peak is observed in SAXS measurements, we
assume these phases are HII, due to the typical hexagonal texture observed
in the polarizing light microscope.

hexosomes was dispersed in 2 mL of distilled water, and in each

case, two or three samples of the same composition were tested and
each measurement was repeated four times at 25 C. The size
distributions (by volume) were obtained from analysis of the data
with the CONTIN algorithm.

Results and Discussion

Reverse Hexagonal Liquid Crystals. In order to study the
role of the polymer in hexosome formations, understanding its
behavior in the bulk hexagonal LLC systems is necessary. Thus,
several mixtures of GMO/tricaprylin at weight ratio 90/10, F127
(GMO+tricaprylin/F127 ratios 100/0, 95/5, 90/10, and 85/15),
and water (15-25 wt %) were prepared. The samples that contain
the polymer were either transparent, LLC (at low polymer
concentrations), or liquidlike (clear for 1 week and then turning
into turbid emulsion at high polymer content). The polymer levels
that induce the LLC-liquid transformation were found to be
dependent on the initial level of LLC hydration. Polarized light
microscopy showed typical hexagonal-type birefringent textures
for all the LLC samples. All the samples were stable for at least
1 week, followed by polymer precipitation. Samples were studied
using SAXS and DSC techniques.
SAXS. SAXS measurements were carried out on mixtures of
GMO, tricaprylin, F127, and variable quantities of water along
dilution lines with GMO+tricaprylin/F127 ratios of 100/0, 95/5,
90/10, and 85/15. The degree of order (as determined by
diffraction peak broadening) of the hexagonal structures is a
function of both the internal water content and the quantity of
added polymer. In the absence of polymer, less ordered and
more well ordered structures were identified at relatively low
(<20 wt %) and higher (>20 wt %) water contents, respectively.
X-ray scattering profiles of the GMO/tricaprylin 90/10 mixture
with 10, 15, and 25 wt % water are shown in Figure 1. In each
case, the three Bragg peaks could be indexed as the 10, 11, and
20 reflections consistent with cylindrical micelles arranged with

3640 Langmuir, Vol. 23, No. 7, 2007

Amar-Yuli et al.

Figure 2. DSC thermograms of (a) GMO, (b) GMO/tricaprylin 90/10, (c) GMO/tricaprylin 75/25, (d) GMO/tricaprylin 25/75, (e) GMO/
tricaprylin 10/90, (f) tricaprylin, and (g) LLC of GMO/tricaprylin, 90/10 and 20 wt % water.

2D hexagonal symmetry.26 From the position of the peaks, we

calculated the corresponding mean lattice parameter (a). The
lattice parameters and Scherrer parameters (LH) for different
sample compositions (15 and 20 wt % water) are summarized
in Table 1 (and more extensively in the Supporting Information).
Upon addition of polymer (95/5 lipid/polymer weight ratio),
a decrease in the lattice parameter and in the domain size was
detected at all water contents (15, 20, and 25 wt %); however,
at lower water content the effect was more pronounced. The
incorporation of the polymer (95/5) at 15 wt % water content
resulted in only 10 diffraction peak being observed and reduced
the hexagonal domain size (LH) by 38%. At higher water content
(25 wt %), LH was reduced by 30%. With 4 wt % polymer
added, the lattice parameter of the sample with 20 wt % water
decreased from 55.4 ( 0.5 to 48.2 ( 0.5 , and the Scherrer
parameter (LH) decreased from 1014 in the absence of polymer
to 647 . It should be noted that the same effects (decrease in
the lattice and Scherrer parameters) were observed at higher
water content (25 wt %). Further increasing the polymer content
to 8 wt %, for the 15 wt % water sample, resulted in a HII
emulsion transformation. For the 20 wt % water sample with 8
wt % polymer, the Scherrer parameter (LH) decreased to 647 .
At high polymer contents8 (90/10 weight ratio) and 11.25 wt
% (85/15)sat 20 and 25 wt % water, respectively, only one peak
was detected by SAXS; however, both revealed a typical HII
texture in the polarized optical microscope. Further increase in
polymer content (>85/15) leads to the formation of an unstable
emulsion phase, even with 20 and 25 wt % water.
Competition for water between the polymer and the GMO can
be assumed to occur. The ability of the highly hydrophilic polymer
(HLB ) 17) to bind water is higher than that of GMO (HLB )
3.8). As a result, the LLC that is based on hydrogen bonding
between GMO and water is gradually destroyed by this
competition. GMO dehydration will decrease the effective size
of the GMO head groups, which will be reflected in shrinkage
of the lattice parameter. The results in Table 1 show that (1) the
polymer acts to reduce or terminate the liquid crystal domains,
a process which is essential for hexosome production; (2) up to
approximately 4 and 8 wt % polymer (at lower and higher water
contents, respectively), the hexagonal phase is compact; however,
above that polymer concentration, the LLC is completely

disrupted; and (3) the stability of the lower water content LLC
hexagonal structure is lower than that of the higher water content
LLC.
DSC. The salting-out effect27 (surfactant head group
dehydration) of the polymer hydrophilic chains on the head groups
of the lipids in the hexagonal mesophase can also be detected
in the thermal behavior of the mesophase as it is reflected in the
DSC measurements. A typical thermogram of the hexagonal
liquid crystal mesophase (GMO/tricaprylin 90/10 and 20 wt %
water) is shown in Figure 2 (thermogram g). The thermotropic
behavior during the heating scan from -20 to +40 C was
recorded and reveals two endothermic events (peaks) at 0.2 (
0.2 (peak A) and 6.1 ( 0.2 C (peak B). For identification of
peak A, we prepared a sample in which the water was replaced
by D2O. A shift of +1.4 C was detected for peak A, which
confirms that this peak is related to the fusion of water. A shift
was not detected for peak B, confirming that the endothermic
event of point B is not sensitive to the substitution of deuterium
for hydrogen. For identification of peak B, the thermal behavior
during heating of GMO and tricaprylin was examined
separately and each of them showed two endothermic peaks,
GMO at 24.7 ( 0.2 and 33.6 ( 0.2 C and tricaprylin at -8.1
and 10.1 ( 0.2 C (Figure 2, thermograms a and f, respectively).
Samples of GMO/tricaprylin at different weight ratios, 90/10,
75/25, 25/75, and 10/90, were examined as well (Figure 2 b-e,
respectively). At high ratios of GMO (90/10 and 75/25), two
endothermic peaks were observed. The higher melting temperature
peak appeared at the same temperature (33.4-32.7 ( 0.2 C);
however, the lower temperature peak moved to lower
temperatures with an increase in the tricaprylin fraction ((22.217.6) ( 0.2 C). Furthermore, decreasing the GMO fraction led
to a decrease in the enthalpy of both endothermic peaks from
34 (pure GMO) to 2 J g-1 (GMO/tricaprylin 75/25) and from 20
(pure GMO) to 6 J g-1 (GMO/tricaprylin 75/25), indicating that
increasing quantities of GMO are dissolved in the tricaprylin.
Further increasing the tricaprylin fraction (25/75 and 10/90 GMO/
tricaprylin) results in the appearance of only one endothermic
peak. This indicated that the GMO is completely dissolved in
the molten liquid tricaprylin, and we see only the endothermic
event at a significantly lower temperature (at 9-10 C, similar
to pure tricaprylin) and higher enthalpy values 70-112 J g-1,

(26) Hyde, S. T. In Handbook of Applied Surface and Colloid Chemistry, 1st

ed.; Holmberg, K., Ed.; Wiley: New York, 2001; Chapter 16.

(27) Takahashi, H.; Matsuo, A.; Hatta. I. Phys. Chem. Chem. Phys. 2002, 11,
2365-2370.

Hexosome and Hexagonal Phases

Langmuir, Vol. 23, No. 7, 2007 3641

Table 2. Effect of the F127 Polymer Concentration and

Homogenization Time on Hexosome Structural Parameters
lattice parameter a ( ( 0.5 )
polymer concn (wt % of aqueous phase,
95 wt % external water)
inner
water
(wt %)

hexagonal
phase
(bulk)

15

52.3

20

55.4

homogenization
time (min)

0.4

1.6

10
15
20
10
15
20

a
59.3
a
a
57.3
a

a
59.8
a
a
60.5
a

58.0
62.0
61.9
56.2
61.0
61.4

a
60.2
a
a
59.1
a

Not measured.
Table 3. Effect of the F127 Polymer Concentration on
Hexosome Size Distribution
particle size distribution
(by volume)

inner water
(wt %)

polymer
concn (wt %)

volume
(%)

diameter
(nm)

20

0.4

75-93
7-3
8-4
70
30
100

75-125
<20
>1000
78 5
300-330
130-150

1
1.6

yet lower than pure tricaprylin (141 J g-1). These results confirm
that peak B is due to a tricaprylin+GMO (eutectic) mixture, with
the position of the endotherm being determined by the relative
miscibility of one component in the other during the heating
process. It should be mentioned that peak B is affected by the
water-GMO interactions as well. When the GMO+water mixture
(73/20 weight ratio) was tested, peak B was observed at 9 C.
However, peak B is less sensitive to GMO hydration than GMOtricaprylin interactions.
The dehydrating effect of the polymer on the HII phase can
be observed by following the changes in peak A, which as shown
above, are attributed to water fusion (Figure 2). The melting
temperatures of the mixtures are summarized in Table 1. With
15 wt % water in the LLC and upon addition of polymer (95/5
lipid/polymer weight ratio), the amount of free water decreased;
i.e., the enthalpy associated with peak A decreased by 3 J g-1.
Moreover, the quantity of bound water increased, which is
expressed by the appearance of an additional endothermic peak
at -2.7 ( 0.2 C. Further increase in the quantity of polymer
(decreasing the lipid/polymer weight ratio to 90/10 and 85/15),
led to the disruption of the mesophase and disappearance of the
water fusion peak. We may conclude that the polymer binds all
the water that was previously free.
With 20 wt % water in the LLC, more stable behavior was
observed. The enthalpy of peak A decreased when polymer was
added up to 8 wt % polymer (90/10 lipid/polymer weight ratio).
Above this polymer concentration, the peak disappeared.
Simultaneously, the temperature of the bound water peak
decreased (to -6 C). Interestingly, following peak B at the
same lipid/polymer ratios (90/10) reveals that the midpoint of
this transition increased from 6.1 to 10.4 ( 0.2 C and from 7.1
to 10.0 ( 0.2 C for the lower and higher water contents in the
LLC, respectively.
Taken together, SAXS and DSC measurements confirm the
idea that head group dehydration, reduction in domain size and
lattice constant, and the LLC-emulsion transformation are all
polymer-dependent.

Figure 3. LUMIFuge stability analyses of the hexosome dispersions

that were stabilized with 0.4 wt % F127 of the external aqueous
phase with (a) 17.5 , (b) 20, and (c) 22.5 wt % inner water content,
following 15-min homogenization.

Characterization of Hexosome Particles. The effect of

polymer F127 on hexosome formation and stabilization has been
studied from the point of view of both microstructure and overall
morphology (particle size distribution and dispersion stability).
Hexosomes were prepared, as described in the Experimental
Section, from five samples taken along the dilution line 90/10
(GMO/tricaprylin weight ratio) with a range of internal water
contents (15-25 wt %) and with polymer concentrations 0, 0.4,
1, and 1.6 wt % of the external aqueous phase.
SAXS. The SAXS measurements presented here focus on
samples of 5 wt % LLC containing either 15 or 20 wt % internal
water dispersed in aqueous medium. The derived HII lattice
parameters for different quantities of added polymer and for
different times of homogenization are summarized in Table 2.
The hexosome preparation protocol, i.e., homogenization during
dispersion of the LLC aliquot in water without polymer, resulted
in rapid hydration during 15 min of homogenization. With 15
wt % internal water, where, as shown above, less-ordered LLC
structures were formed, the new hydration level is reflected in
an increase of the lattice parameter from 52.3 ( 0.5 for the
bulk hexagonal LLC (before dispersion) to 59.3 ( 0.5 for the
nonstable dispersion without polymer. However, at 20 wt %
water, the same conditions increased the lattice parameter, but

3642 Langmuir, Vol. 23, No. 7, 2007

Figure 4. LUMIFuge stability analyses of the hexosome dispersions

with 20 wt % inner water content that were stabilized with (a) 1 and
(b) 1.6 wt % F127, in the external aqueous phase following 15-min
homogenization.

Figure 5. LUMIFuge transmission profile at a fixed position (at

112 mm along the tube) as a function of measurement time for
hexosome dispersions prepared with initial water content of 17.5
(0), 20 (;), and 22.5 (+) wt % and with 1 wt % F127 in the external
aqueous phase.

to a smaller extent, and a more ordered hexagonal structure with

a lower lattice constant (57.3 ( 0.5 ) was formed. The addition
of polymer (0.4 wt % of the aqueous phase corresponding to 7.5
wt % of the dispersed LLC) increased the lattice parameter or
left it unchanged (at 20 and 15 wt % water, respectively). Further
increase in the polymer concentration, up to 1.6 wt % (corresponding to 30 wt % of the liquid crystal), did not affect the
lattice parameter of the hexosomes. However, the presence of
the polymer did reduce the domain size (Scherrer parameter) as
shown earlier for the bulk hexagonal systems. With 15 wt %
internal water, where less-ordered hexagonal structures were
formed, the inner structure was more sensitive to the presence
of polymer. The domain size was reduced from 772 ( 50 in
the absence of polymer and before dispersion to 493 ( 50
upon addition of 0.4-1.6 wt % F127 in the external water. It

Amar-Yuli et al.

Figure 6. LUMIFuge stability analyses of the hexosome dispersions

with 20 wt % inner water content stabilized with 1 wt % F127 in
the external aqueous phase, and homogenized for (a) 10 and (b) 20
min.

should be mentioned that increasing the polymer content from

0.4 wt % to 1 and 1.6 wt % did not further reduce the domain
size. With more water in the cylinder cores, the addition of
polymer (0.4-1.6 wt %) reduced the domain size to 730 ( 50
, compared to 1014 ( 50 in its absence.
On the other hand, we showed earlier (Table 1) that, for the
bulk hexagonal LLC, lower relative amounts of polymer (12 wt
%) destroyed the inner structure and caused a transformation of
the LLC into an unstable emulsion. At 1.6 wt % polymer (30
wt % polymer relative to the LLC), the highly hydrophilic F127
(HLB ) 17) would have led to the destruction of the HII structure
if no external water were present. However, in a hexosome
dispersion, the polymer can be hydrated by the external water
as well, leaving the HII structure relatively undamaged. A
competitive action of the external water and the polymer on the
head groups and free water of the cylinder cores can be
assumed. Such competitive action is evident from the fact that
an increase in polymer concentration from 0.4 to 1.6 wt % of
the aqueous phase did not further decrease the lattice parameter
or destroy the inner structure as shown for the bulk systems. We
can conclude that water penetration occurred in preference to
polymer adsorption. Furthermore, homogenization in water only
without the addition of polymer leads to a rapid swelling
manifested by a marked increase of the lattice parameter
(Table 2).
The relationship between homogenization time and hexosome
inner structure, size, and stability was also examined using SAXS.
Three homogenization periods were used (10, 15, and 20 min)
for a 1.0 wt % polymer concentration (Table 2). Hexosomes
were formed even after the minimum homogenization time of
10 min, but only partial hydration of the hexosomes occurred.
At 15 wt % water, the homogenization effect was more
pronounced: the new hydration state is reflected in the increase
of the lattice parameter from 52.3 ( 0.5 for the bulk hexagonal
LLC (before dispersion) to 58.0 ( 0.5 for the hexosome

Hexosome and Hexagonal Phases

Figure 7. Cryo-TEM images of hexosome dispersions with 20 wt

% inner water content stabilized with (a-b) 0, (c) 0.4, (d) 1, and
(e) 1.6 wt % F127 in the external aqueous phase. Images b and c
depict only the very large particles that were seen in the sample. The
small particles are similar to those seen in image d. The bar
corresponds to 50 nm and is appropriate to images a-e.

dispersion. However, when the more ordered structure was formed

with 20 wt % water, the same conditions of homogenization did
not change the lattice parameter. Increasing the homogenization
periods to 15 and 20 min increased the hydration and hence the
lattice parameter up to the maximum value (60-62 ),
regardless of the initial degree of order.
Hexosome Stability. The stability of the dispersed hexosomes
as a function of composition and homogenization time can be
evaluated by LUMiFuge tests; indirect determination can also
be made by DLS measurements of particle size distributions and
cryo-TEM imaging.
DLS. With 20 wt % water in the LLC phase (90/10 GMO/
tricaprylin), 15 min of homogenization, and at low polymer
concentration (0.4 wt % of the external aqueous phase), DLS
measurements reveal three major populations (Table 3); 75-93
vol % of the particles are in the size range of 75-125 nm and
the rest are either <20 nm or >1000 nm. We interpret this to
mean that there was insufficient polymer to fully cover all the
particles: some remained bare and some were reduced in size.
Increasing the polymer concentration to 1 wt % results in the
appearance of only two major populations; i.e., 70 vol % are 78
nm and 30 vol % are 300-330 nm in size. Further raising the
polymer concentration to 1.6 wt % narrowed the hexosome size
distribution to almost 100 vol % in the range of 130-150 nm.
These results stress the fine-tuning that is needed, under constant
shear conditions of the homogenization process, of the polymer/
LLC ratio.
We also fixed the polymer concentration (1 wt %) and varied
the length of the homogenization period. After 10 min of
homogenization, two major populations were obtained, 67-74
vol % are 68-83 nm and 25-33 vol % are 260-320 nm (the
narrow particles distribution are an intermediate particulation
of the dispersed systems); after 15 min, the distributions narrowed
somewhat. Surprisingly, after 20 min of homogenization, a
population of large particles (50-60 vol % are 1500-2600 nm)
was observed in addition to the small particles (40-50 vol %
76-140 nm). These results may suggest that since the penetration
of the polymer was sufficient even with short periods of
homogenization, as observed by the changes in the lattice
constants, the polymer effectively destroyed the partial LLC
structure and formed small particles even at these conditions.
LUMiFuge. Dispersion stability was further examined using
LUMiFuge measurements. With this instrument, lack of stability

Langmuir, Vol. 23, No. 7, 2007 3643

(stability to aggregation, particle precipitation, and separation)

is measured by the extent of centrifugation-induced changes in
light transmission as a function of both time and position in the
sample tube. A correlation between the initial degree of order
of the bulk HII and the stability of the colloidal dispersions is
illustrated (Figure 3). With 17.5 wt % water content in the LLC
phase (GMO/tricaprylin 90/10), 0.4 wt % polymer in the bulk
aqueous phase, and homogenization for 15 min at 9500 rpm, the
light transmission at the initial time (t0) was 15% along the whole
length of the sample centrifuge tube, indicating the presence of
a uniform milky dispersion (Figure 3a). With time, an increase
in transmission at the bottom of the tube appeared, meaning loss
of stability. Oil separation and movement to the top of the tube
is indicated by higher transmission at the lower part of the tube
relative to the upper part. At the final time (tf), transmission
increased over a larger length of the tube and a larger fraction
of the sample exhibited phase separation. At 20 wt % water
(Figure 3b), a similar but more moderate decomposition profile
is detected. However, with 22.5 wt % water in the LLC, more
stable dispersions were obtained (Figure 3c). The increase in
transmission during the course of the measurement was smaller
(20-45%), and only a small fraction of the sample exhibited
phase separation (oil floating).
Stability tests on hexosome dispersions with 20 wt % water
in the LLC and 1 or 1.6 wt % polymer in the bulk aqueous phase
are shown in Figure 4a,b, respectively. Higher polymer concentration exhibited a higher initial percentage of transmission
(25 and 35% for 1.0 and 1.6 wt % polymer, respectively),
suggesting the formation of smaller particles. Also the change
in transmission during the measurements decreased with an
increase in the polymer concentration, thereby proving higher
stability. Finally, oil separation was negligible at the highest
polymer concentration.
Figure 5 summarizes the transmission profile (at 112 mm
along the tube) as a function of measurement time and shows
almost equivalent stability profiles for hexosome dispersions
made with different initial water concentrations (17.5, 20, and
22.5 wt %), however, with high polymer concentration (1 wt %).
The narrow size distribution of these hexosomes as seen in DLS
measurements is consistent with the relatively high dispersion
stability.
The stability analyses of the hexosome dispersions by the
LUMIFuge technique confirm our previous findings concerning
the influence of the quantity of polymer and the length of the
homogenization periods. Results of the LUMiFuge measurements
are in agreement with those from DLS, which together show that
with 20 wt % water, either 10 or 15 min of homogenization leads
to similar size distribution and stability. After a short homogenization period (10 min), the change in transmission during the
measurement was small (25-35%) and only a minor fraction of
the sample exhibited phase separation (Figure 6a the bottom of
the tube).
Further homogenization, for 15 min, led to less phase separation
(Figure 4a). After 20 min of homogenization, a population of
large-sized (1500-2600 nm) particles in addition to the small
soft particles was observed by DLS. The presence of large particles
is also expressed in the low transmission (10%) at the beginning
of the LUMiFuge measurement (Figure 6b). Furthermore, during
the measurement, the increase in transmission was very noticeable
along the length of the capillary, indicating that a very large
fraction of the sample exhibited phase separation.
It should be mentioned that hexosomes which were prepared
with different homogenization periods (10-20 min) exhibited
similar behavior regardless of their location on the phase diagram

3644 Langmuir, Vol. 23, No. 7, 2007

Amar-Yuli et al.

Figure 8. Schematic presentation illustrating the polymer effect as domains and lattice reduction by dehydration.

(low or high water content), with only one exception. After 10

min of homogenization of the less ordered hexagonal structures,
dispersion with a larger size distribution and lower stability was
formed. These findings emphasize once more our assumption
that the less-ordered structures are much more sensitive to
modifications than the more-ordered ones. This property can be
utilized in different delivery systems with different demands.
Cryo-TEM. Cryo-TEM images provide additional evidence
for the existence of particles with hexagonal morphology and
internal hexagonal symmetry. At low polymer concentration (0
and 0.4 wt %) and 15 min of homogenization, a broad particle
size distribution can be seen (Figure 7A-C), while a narrow
hexosome size distribution is seen at higher polymer concentrations (1-1.6 wt %; Figure 7D,E). Images of the nonstable
dispersions in the absence of polymer (Figure 7A,B) show
hexosomes of <100 nm along with bulk hexagonal fractions
(>500 nm), which may be evidence for an aggregation process.
It should be mentioned that, visually, one can see a few large
liquid crystalline particles that were not dispersed or, alternatively,
rapidly aggregated. At 0.4 wt % of polymer (Figure 7C), very
large hexagonal particles (above 500 nm) are seen along with
a few vesicles. The lack of sufficient polymer is reflected in
insufficient cutting of the domains as well as in insufficient
interfacial covering, resulting in the existence of lamellar domains
that form vesicles. Large particles and the presence of vesicles
can promote aggregation, which eventually destabilizes the
hexosome dispersion. Dispersions with larger quantities of
polymer (1-1.6 wt %, Figure 7D,E) exhibit a much narrower
distribution of hexosome size (150-300 nm) along with a larger
number of vesicles. The polymer is efficiently incorporated into
the hexagonal structure but at the same time does not prevent

partial transformation of the hexagonal phases into lamellar

structures, resulting in the formation of vesicles.

Conclusions
Using the GMO/tricaprylin/water hexagonal LLC as well as
the hexosome particles, which were formed by dispersing the
LLC in water/polymer mixtures, the effects of the amphiphilic
polymer F127 and LLC hydration on hexosome structure were
studied. Competition between the hydration of the GMO and the
adsorption of the polymer plays a central role in hexosome
formation. Polymer adsorption presumably takes place via the
hydrophobic moiety of F127, which interacts with the outer
surface of the reverse cylindrical micelles, which is also
hydrophobic. Increasing the polymer concentration to an optimum
value was found to reduce the mean size of the hexosome particles
as well as the width of the size distribution. When the
concentration of polymer was too low, the result was nonstable
dispersions with a broad distribution of particle sizes, including
a population of very large particles.
SAXS and DSC measurements on the bulk LLC phases both
showed a dehydrating effect of the polymer on the hexagonal
structure, which leads first to the compaction of the LLC lattice
and eventually to its transformation into an unstable emulsion
(see schematic presentation in Figure 8). These polymer effects
were more pronounced for HII with lower water content (and
domain size) than with higher water content. It was found that
hexosome stability could be correlated to the initial order in the
hexagonal structure only when extreme conditions of polymer
concentration (<1 wt %) and length of shear periods (e10 and
g20 min) were used.
It can be seen that the role of the polymer in destabilizing the
hexagonal domains depends on the initial degree of order of the

Hexosome and Hexagonal Phases

hexagonal structure. We assume that the disordered structures

are less homogeneous, have more defects, and are more
susceptible to water penetration and hydration. Consequently, at
low polymer concentration, the polymer will cut the domains
but will not be homogeneously distributed and will not efficiently
cover the particle interface. Such a situation will result in unstable
particles with a very large size distribution. In a region of hydration
where structures are more homogeneous and better ordered, the
domains will be cut more gradually with better adsorption of the
polymer. As a result, a narrower size distribution and improved
stability of the polymer content hexosomes are observed. To the
contrary, at higher polymer concentrations (1-1.6 wt %), where
there is sufficient polymer to adsorb and stabilize the dispersion,
no differences between the initial structures are detected. It should
be stressed that hexosome stability depends not only on the
quantity of water in the LLC phase but also on the polymer
concentration and the shear rate. Hexagonal structures with both

Langmuir, Vol. 23, No. 7, 2007 3645

lower and higher water content can be stably dispersed if the

polymer concentration and shear rates are properly manipulated.
More importantly, one can manipulate the hexosome properties
by monitoring the polymer concentration, homogenization period,
and initial order structure.
Acknowledgment. This work was supported by a grant from
the U.S.sIsrael Binational Science Foundation number 2602003. D.D. also acknowledges the support of the Israel Science
Foundation, and the Pearl and Ernest Nathan Research Fund.
Supporting Information Available: An extensive table of
structural properties of the less ordered and more ordered (10 and 25
wt % water concentrations) hexagonal structures upon addition of F127.
This material is available free of charge via the Internet at http://pubs.acs.org.
LA062851B