Nanostructures in water-in-CO2 microemulsions stabilized by double-chain fluorocarbon solubilizers

Article abstract of DOI:10.1021/la400376g

High-pressure small-angle neutron scattering (HP-SANS) studies were conducted to investigate nanostructures and interfacial properties of water-in-supercritical CO₂ (W/CO₂) microemulsions with double-fluorocarbon-tail anionic surfactants, having different fluorocarbon chain lengths and linking groups (glutarate or succinate). At constant pressure and temperature, the microemulsion aqueous cores were found to swell with an increase in water-to-surfactant ratio, W₀, until their solubilizing capacities were reached. Surfactants with fluorocarbon chain lengths of n = 4, 6, and 8 formed spherical reversed micelles in supercritical CO₂ even at W₀ over the solubilizing powers as determined by phase behavior studies, suggesting formation of Winsor-IV W/CO₂ microemulsions and then Winsor-II W/CO₂ microemulsions. On the other hand, a short C2 chain fluorocarbon surfactant analogue displayed a transition from Winsor-IV microemulsions to lamellar liquid crystals at W₀ = 25. Critical packing parameters and aggregation numbers were calculated by using area per headgroup, shell thickness, the core/shell radii determined from SANS data analysis: these parameters were used to help understand differences in aggregation behavior and solubilizing power in CO₂. Increasing the microemulsion water loading led the critical packing parameter to decrease to ～1.3 and the aggregation number to increase to >90. Although these parameters were comparable between glutarate and succinate surfactants with the same fluorocarbon chain, decreasing the fluorocarbon chain length n reduced the critical packing parameter. At the same time, reducing chain length to 2 reduced negative interfacial curvature, favoring planar structures, as demonstrated by generation of lamellar liquid crystal phases.

Full text of DOI:10.1021/la400376g

Article
pubs.acs.org/Langmuir
Nanostructures in Water-in-CO2 Microemulsions Stabilized by
Double-Chain Fluorocarbon Solubilizers
Masanobu Sagisaka,*,† Shuho Iwama,† Shinji Ono,† Atsushi Yoshizawa,† Azmi Mohamed,‡
Stephen Cummings,§ Ci Yan,§ Craig James,§ Sarah E. Rogers,∥ Richard K. Heenan,∥ and Julian Eastoe*,§
†Department of Frontier Materials Chemistry, Graduate School of Science and Technology, Hirosaki University, 3 Bunkyo-cho,
Hirosaki, Aomori 036-8561, Japan
‡Department of Chemistry, Faculty of Science and Mathematics, Universiti Pendidikan Sultan Idris, Tanjong Malim, Perak 35900,
Malaysia
§School of Chemistry, University of Bristol, Cantock’s Close, Bristol BS8 1TS, United Kingdom
∥ISIS-CCLRC, Rutherford Appleton Laboratory, Chilton, Oxon OX11 0QX, United Kingdom
S
* Supporting Information
ABSTRACT: High-pressure small-angle neutron scattering (HP-SANS) studies were conducted to investigate nanostructures
and interfacial properties of water-in-supercritical CO2 (W/CO2) microemulsions with double-fluorocarbon-tail anionic
surfactants, having different fluorocarbon chain lengths and linking groups (glutarate or succinate). At constant pressure and
temperature, the microemulsion aqueous cores were found to swell with an increase in water-to-surfactant ratio, W0, until their
solubilizing capacities were reached. Surfactants with fluorocarbon chain lengths of n = 4, 6, and 8 formed spherical reversed
micelles in supercritical CO2 even at W0 over the solubilizing powers as determined by phase behavior studies, suggesting
formation of Winsor-IV W/CO2 microemulsions and then Winsor-II W/CO2 microemulsions. On the other hand, a short C2
chain fluorocarbon surfactant analogue displayed a transition from Winsor-IV microemulsions to lamellar liquid crystals at W0 =
25. Critical packing parameters and aggregation numbers were calculated by using area per headgroup, shell thickness, the core/
shell radii determined from SANS data analysis: these parameters were used to help understand differences in aggregation
behavior and solubilizing power in CO2. Increasing the microemulsion water loading led the critical packing parameter to
decrease to ∼1.3 and the aggregation number to increase to >90. Although these parameters were comparable between glutarate
and succinate surfactants with the same fluorocarbon chain, decreasing the fluorocarbon chain length n reduced the critical
packing parameter. At the same time, reducing chain length to 2 reduced negative interfacial curvature, favoring planar structures,
as demonstrated by generation of lamellar liquid crystal phases.
ature-tunable solvency (or CO2 density). Unfortunately,
supercritical CO2 can dissolve only nonpolar and small
molecular mass materials, and common polar or high molecular
1. INTRODUCTION
Supercritical CO2 (scCO2) is seen as a promising green solvent
in various fields, including organic synthesis, dry cleaning,
polymerization, extraction, and nanomaterial processing.1 The
reasons why scCO2 has attracted much attention for those
applications are based on CO2 and supercritical fluid properties,
which are low cost, nonflammability, environmentally benign,
natural abundance, high mass transfer, and pressure/temper-
2
mass materials always separate from neat scCO2 . Improving
Received: January 29, 2013
Revised: May 11, 2013
© XXXX American Chemical Society
A
dx.doi.org/10.1021/la400376g | Langmuir XXXX, XXX, XXX−XXX

Langmuir
Article
the poor solubility of polar materials is important for
developing applications of scCO2. One of the most promising
approaches for enhancing solubility in scCO2 is to form
reversed micelles with high-polarity aqueous cores in the
continuous scCO2 phase, that is, water-in-scCO2 micro-
emulsions (W/CO2 μEs).2 Since such organized fluids have
the attractive characteristics of scCO2, as well as the solvation
properties of water, they have potential as volatile organic
compound (VOC)-free and energy-efficient solvents for
nanomaterial synthesis, enzymatic reactions, dry-cleaning,
dyeing, preparation of inorganic/organic hybrid materials, and
so on.2
To be a viable green technology, the amount of surfactant
used should be as small as possible, and this needs to be
balanced against the need for large interfacial areas in W/CO2
μEs and appropriate levels of dispersed water for enhanced
process efficiencies. One approach to meet these requirements
is to explore or develop a highly efficient solubilizers for W/
CO2 μEs. Since 1990, much effort has been directed toward the
development of surfactants for W/CO2 μEs.3−30 Three main
kinds of surfactant have been examined so far, being
hydrocarbon, silicone, and fluorocarbon (FC) surfactants.
Development of CO2-philic hydrocarbon surfactants has also
been conducted for economic and environmental reasons.3−9
However, most commercial and known hydrocarbon surfac-
tants are insoluble and inactive in scCO2 systems.3 In this
regard, it became apparent that conventional surfactant-design
theory cannot be applied to W/CO2 systems directly and that
CO2-philicity is not directly comparable to oleo-philicity.
Therefore, advancing molecular-design theory for CO2-philic
surfactants has required new directions and paradigms in the
field of surfactants.
Only a few hydrocarbon surfactants have so far been
reported to stabilize a W/CO2 μEs. One of these successful
surfactants is the nonionic TMN-6,4−6 which has highly
branched alkyl tails and ∼8 oxyethylene units: TMN-6 was
reported to solubilize water up to a water-to-surfactant molar
ratio, W0 of 30. Custom-made anionic surfactants with highly
branched double- or triple-tails (sodium bis(3,5,5-trimethyl-1-
hexyl) sulfosuccinate7 or (sodium 1,4-bis(neopentyloxy)-3-
(neopentyloxycarbonyl)-1,4-dioxobutane-2-sulfonate8) were
also found to be soluble in scCO2 and yield transparent single
phase W/CO2 μEs (Winsor-IV W/CO2 μE or IVμE) with W0
less than 15. Note that the commercial analogue of these
surfactants, Aerosol-OT (sodium bis-(2-ethyl-1-hexyl) sulfo-
succinate, AOT) is inactive and insoluble in scCO2 and hence is
ineffective at stabilizing W/CO2 μEs. It has been recognized
that hydrocarbon surfactants with highly branched tails,
especially methyl branches, can be considered sufficiently
CO2-philic. On the other hand, ester and ether groups have
been reported to increase solubility in scCO2 as well as methyl
branches, and highly CO2-philic, and W/CO2-interfacially
active copolymers with these groups have been developed
and used for emulsification and polymerization.9 Unfortunately,
an efficient and cost-effective hydrocarbon stabilizer for W/
CO2 μEs, like the AOT used commonly for W/O μEs,10 has
not yet been found.
concentration (∼ 0.1 wt %), and none were observed to form
microemulsions or generate micellar structures in scCO2.12
A
copolymer of PDMS24-b-EO22 was shown to lower the water−
scCO2 interfacial tension from 20 to 0.2 mN m−1; a PDMS
surfactant did however form W/CO2 emulsions,13 which
flocculated and coalesced, limiting the stability of these systems.
New trisiloxanes ((CH3 )3 SiO)2 Si(CH3 )2 (CH2 )3 -
(OCH2CH2)nR were reported to be effective emulsifiers,
forming both W/CO2 emulsions at n < 7 and CO2/W
emulsions at n ≥ 7.14 This change in curvature is attributed to
variations in hydrophilic-CO2-philic balance (HCB),15 which is
accompanied by a minimum in interfacial tension between
water and CO2.
In earlier studies, several fluorinated surfactants were found
to dissolve in CO2 and have a high activity at the W/CO2
interface, suggesting the feasibility of forming W/CO2 μEs.16,17
Among others, surfactants described below are noteworthy for
generating W/CO2 μEs. Early on a perfluoropolyether (PFPE)
surfactant16 was found to stabilize IVμE, but with only a low W0
= 21 (also expressed as a corrected water-to-surfactant molar
ratio by subtracting the low background water solubility in
c
CO2, W0 = 14). After that, numerous reports dealing with W/
CO2 μEs focused on PFPEs.
Another successful class of CO2-philic surfactants are the
hydrocarbon−fluorocarbon (HC-FC) hybrids, for example,
sodium 1-pentadecafluoroheptyl-1-octanesulfate (F7H7,
(C7H15)(C7F15)CHOSO3Na): these have both a HC and a
FC chain in the same molecule. As such F7H7 is able to
c
solubilize up to W0 = 35 (or W0 = 32) to form stable Winsor-
IV W/CO2 μEs.18 Further studies,19 with hybrid surfactants
related to F7H7 but with different FC and HC chain lengths,
resulted in the formation of Winsor-IV W/CO2 μEs for most of
the analogues, but unfortunately smaller attainable W0 values
than for F7H7.
Other investigations20−22 studied a class of fluorinated AOT
analogues, for example sodium bis(1H,1H,5H-octafluoropen-
tyl)-2-sulfosuccinate (di-HCF4), which yield Winsor IVμEs
c
with W0 = 30 (W0 ∼ 20). In addition, double-FC-tail
phosphate surfactants were also found to be efficient μE
stabilizers,23 the most favorable case stabilizing W0 up to 45.24
Lu and Berkowitz25 and Cummings et al.26 studied
nanostructures of W/CO2 μEs with FC and HC surfactants
by molecular dynamics simulations. Lu and Berkowitz25
observed a quick self-assembly of the PFPE surfactant spherical
reverse micelles over time periods of 5 ns, irrespective of initial
conditions. In most cases, the self-assembled PFPE reverse
micelles have a spherical shape and properties consistent with
SANS results.17,25 When the FC surfactant is replaced by a HC
analogue, the HC assembly contains a region of direct contact
between water and carbon dioxide, indicating that HC
surfactants are likely to be inappropriate for formation of W/
CO2 μEs. On the other hand, Cummings et al.26 reported that
the HC tails in scCO2 had an average of 74 4% trans bonds
(cf. 89% in vacuum), while the FC tails had 91 2% (cf. 81%
in vacuum). This implies that the HC tails assume more
contracted conformations in CO2, indicating CO2-phobic
interactions while the FC tails assume more extended
conformations, consistent with favorable CO2-philic interac-
tions.
In the case of silicone surfactants, polydimethylsiloxanes
(PDMS) have attracted attention as potential nonfluorinated
CO2-philes. While they are known to be miscible with CO2,
they still require greater pressures to dissolve than is generally
needed for fluorocarbon surfactants.11 The majority of siloxane
surfactants tested were only found to be soluble at low
Recent molecular simulation studies27,28 have elucidated the
reasons why FC surfactants can act as efficient W/CO2
solubilizers; as compared with HC chains, FC groups have
(1) a stronger interactions with CO2 via quadrupolar and
B
dx.doi.org/10.1021/la400376g | Langmuir XXXX, XXX, XXX−XXX

Langmuir
Article
Table 1. Properties of nFS(EO)2 (m = 1) and nFG(EO)2 (m = 2) in Water at 1 bar or scCO2 at 350 bar24,28,29
a
b
c
Critical micelle concentration in water. Surface tension at CMC. Area per surfactant headgroup at CMC obtained by using the Gibbs adsorption
equation. Values of Krafft point, CMC, γCMC, and Ah,st in aqueous surfactant solutions were obtained from literature.24,28,29 Solubilizing powers
expressed as W0 for Winsor-IV W/CO2 μEs at 350 bar, which were reported in earlier studies.24,28,29 The uncertainties are 3 in W0. The solubilizing
power of 2FG(EO)2 at 45 °C was measured by visual observation.
d
dispersion interactions, and (2) weaker FC-FC chain−chain
interactions which are down to a weak repulsion, electrostatic
in origin. These properties conspire together to give FC
surfactant reversed micelles better solvation by CO2, and this
causes lower surfactant interfacial packing densities, and weaker
attractive intermicellar interactions as compared with hydro-
carbon surfactant analogues.
To obtain more economic and environmentally friendly W/
CO2 μEs, the minimum fluorine content necessary to render a
surfactant CO2-philic has been identified by using double-
pentyl-tail surfactants with different fluorination levels. It was
found that at least two fluorinated carbons (CF3CF2-) are
required to stabilize μEs.29 An increase in fluorine level leads to
a lower aqueous surface tension at CMC and a lower
stabilization cloud pressure for the W/CO2 μEs.
Recently, with the aim of optimizing surfactant structure of
fluorinated AOT analogues for W/CO2 μE, double-FC-tail
anionic surfactants with various FC lengths (n = 4, 6, 8) and a
sulfo-glutarate (nFG(EO)2) or sulfo-succinate (nFS(EO)2)
headgroup were synthesized (see Table 1 for their chemical
structures).24,30,31 As the structural difference between nFS-
(EO)2 and nFG(EO)2 is only one methylene unit in the group
linking the polar ionic headgroup to the pendant FC tails, these
surfactants examined the effect of not only FC length but also
the additional methylene spacer. From surface tension
measurements on aqueous surfactant solutions, no significant
differences were found in CMC, surface tension at CMC, and
area per headgroup between nFG(EO)2 and nFS(EO)2 at the
same FC length n.31 However, visual observation and UV−vis
spectral measurements with a hydrophilic dye methyl orange
(MO) clarified that the solubilizing powers of the glutarates
nFG(EO)2 were higher than those of the succinate analogues
nFS(EO)2: the most efficient found in that study was
4FG(EO)2 at 75 °C (max. W0 = 80 in IVμE) even though it
has the shortest FC tails.31 Significantly, this is the highest W0
reported to date for W/CO2 μEs. Many earlier papers
mentioned that fluorocarbon is one of a few CO2-philic
groups, with longer FC chains promoting the higher the
solubilizing powers.32 This is an exciting result leading to new
approach to equip a fluorine-light surfactant with high
solubilizing power. As well as the high solubilizing power,
4FG(EO)2 was found to give the fastest dissolution into scCO2,
being capable of solubilizing water in just a few seconds, even at
high W0 and close to its maximum solubilizing capacity.31 Such
a fast solubilization rate is very rare in this field.
It is very interesting to consider how the beneficial properties
of 4FG(EO)2 are related to the molecular structure, especially
for advancing surfactant design theory. These excellent
solubilizing properties could be related to a low HCB,15
a
high critical packing parameter (CPP),33 and weak interactions
between tail−tail and tail−head groups,34,35 if the spherical
reversed micelles supported the formation of IVμEs.
To further characterize the surfactant and explore the origins
of the superefficiency, this new study has examined aggregation
behavior and nanostructures of the custom-made double-FC-
tail surfactants, nFG(EO)2 (FC lengths n = 2, 4, 8) and
nFS(EO)2 (n = 4, 8) in W/CO2 μEs, by high-pressure small-
angle neutron scattering (HP-SANS). The significance of this
study to the field of surfactant science is that optimized,
superefficient, low fluorine content surfactants are now available
for stabilization of W/CO2 μEs. In addition, although surfactant
structure−aggregation property correlations in W/CO2 systems
have been reported,19−22,29,31 as mentioned above, changes in
CPP (or curvature) and aggregation number as a function of
microemulsion composition through W0, and surfactant
chemical structure in terms of FC and methylene spacer
lengths have not been examined in such detail as here. The
extensive HP-SANS experiments described here as a function of
these variables show that very different interfacial curvatures
can be obtained, being negative curvature in the droplet
microemulsions and “planar” in the lamellar phase systems. As
such this paper describes a new approach for controlling
interfacial curvature in CO2 self-assembly systems by using a
simple composition parameter and also a more sophisticated
surfactant molecular structure design. These approaches hold
promise for tuning of and control over scCO2 physicochemcial
properties which will be needed for future applications.
2. EXPERIMENTAL SECTION
2.1. Materials. The surfactants used in this study were sodium 1,5-
bis[(1H,1H,2H,2H-perfluorobutyl)oxy]-1,5-dioxopentane-2-sulfonate
(2FG(EO)2), sodium 1,5-bis[(1H,1H,2H,2H-perfluorohexyl)oxy]-1,5-
dioxopentane-2-sulfonate (4FG(EO)2), sodium 1,5-bis-
[(1H,1H,2H,2H-perfluorodecyl)oxy]-1,5-dioxopentane-2-sulfonate
(8FG(EO)2), sodium 1,4-bis[(1H,1H,2H,2H-perfluorohexyl)oxy]-1,4-
dioxobutane-2-sulfonate (4FS(EO)2), and sodium 1,4-bis-
[(1H,1H,2H,2H-perfluorodecyl)oxy]-1,4-dioxobutane-2-sulfonate
(8FS(EO)2). These surfactants except 2FG(EO)2 were synthesized
and evaluated in terms of interfacial properties as described
previously.24,30,31 2FG(EO)2 was newly synthesized to examine the
FC length effect in this study, as shown in Section 2.2. 1H,1H,2H,2H-
Pentafluoro-1-butanol (SantaCruz) and dimethyl glutaconate (Al-
C
dx.doi.org/10.1021/la400376g | Langmuir XXXX, XXX, XXX−XXX

Langmuir
Article
drich) were used without further purification. Reagent grade acetone,
dichloromethane, hexane, 1,4-dioxane, toluene, p-toluene sulfonic acid
monohydrate, and sodium hydrogen sulfite were obtained from Wako
Pure Chemical Industries and employed as received. Surfactant
structures are shown in Table 1 with interfacial properties of aqueous
solutions obtained by standard measurements and solubilizing powers
in scCO2.24,30,31 As compared to nFS(EO)2, nFG(EO)2 has an extra
−CH2− spacer between the FC chain and the sulfonate group.
Ultrapure water with a resistivity of 18.2 MΩ cm, obtained from a
Millipore Milli-Q Plus system, was used in the experiments. CO2 of
99.99% purity (Ekika Carbon Dioxide Co., Ltd.) was used. The
structures of these steric models and the length of one surfactant
molecule in the absence of other molecules were obtained using MM2
(Molecular Mechanics program 2) calculations (Chem 3D;
CambridgeSoft Corp., Cambridge, MA).
Predetermined amounts of D2O and surfactant, where the molar
ratio of surfactant to CO2 was fixed at 8 × 10−4 (= 17 mM at the
experimental condition), were loaded into the Thar cell. Then, CO2
(11.3 g) was introduced into the cell by using a high pressure pump,
and the surfactant/D2O/CO2 mixture was pressurized up to 350 bar at
45 °C by decreasing inner volume of the Thar cell. With vigorous
stirring, visual observation was carried out to identify the mixture to be
a transparent single IVμE or the other turbid phases. Finally, HP-
SANS measurements were conducted for not only IVμEs but also
turbid phases. The densities of CO2 were calculated using the Span−
Wagner equation of state (EOS).36
Neutrons are scattered by short-range interactions with sample
nuclei, with the “scattering power” of different components being
defined by a scattering-length density (SLD), ρ (cm−2). For CO2, ρCO
2
∼ 2.50 × mass density × 1010 cm−2;37 at the experimental pressure of
350 bar and temperature of 45 °C, the CO2 density is 0.917 g cm−3 so
2.2. Synthesis. 2.2.1. Synthesis of Bis(1H,1H,2H,2H-pentafluor-
obutyl) Glutaconate. A mixture of 9.79 g of 1H,1H,2H,2H-
pentafluoro-1-butanol (61.0 mmol), 4.80 g of dimethyl glutaconate
(30.4 mmol), and 1.45 g of p-toluene sulfonic acid monohydrate (7.4
mmol) in 200 cm3 toluene was refluxed under stirring at 130 °C for 40
h. During the reaction, the methanol liberated was removed
azeotropically from the reaction system to shift the equilibrium of
the transesterification reaction. After the reaction was complete, the
mixture was purified by column chromatography with dichloro-
methane/n-hexane (3:1) as a developing solvent and silica gel. Finally,
transparent sticky liquid, bis(1H,1H,2H,2H-pentafluorobutyl) glutac-
onate was obtained (yield 3.69 g, 28.8%): 1H NMR (500 MHz,
CDCl3, TMS), (δH/ppm): 2.40−2.51 (a, m, 4H), 3.28 (c, dd, 2H, J =
1.5, 7.2 Hz), 4.41(b, t, 2H, J = 5.6 Hz), 4.44 (b′, t, 2H, J = 5.6 Hz),
5.96 (e, dt, 1H, J = 1.6, 15.7 Hz), 7.02 (d, dt, 1H, J = 7.2, 15.7 Hz) for
that ρCO ∼ 2.29 × 1010 cm−2. The scattering length densities of
2
surfactant (ρsurf) and D2O (ρD O) were obtained using
2
ρ =
b/V
i m
∑
(1)
i
bi are the nuclear scattering lengths as given in the literature,38 and Vm
is the molecular volume, which can be obtained from the mass density.
Mass densities of surfactants were assumed to be 1.0−1.7 g/cm3 as 1.7
g/cm3 for a typical fluorinated compound17 and 1.0 g/cm3 for a
hydrocarbon surfactant. The calculated scattering length densities for
nFS(EO)2 and nFG(EO)2 were 3.24 × 1010 cm−2 (n = 8), 2.80 × 1010
cm−2 (n = 4), and 2.33 × 1010 cm−2 (n = 2), respectively. The
scattering length density of D2O at 45 °C was calculated to be ∼6.32 ×
1010 cm−2. Samples in pure CO2 (11.3 g) were run at the constant
molar ratio of surfactant to CO2 of 8 × 10−4.
For model fitting data analysis the μE droplets were treated as
spherical core−shell particles with a Schultz distribution in the core
radius.39 Full accounts of the scattering laws are given elsewhere.39−43
For polydisperse spherical droplets at volume fraction ϕ, radius Ri,
volume Vi, and coherent scattering length density ρp dispersed in a
medium of ρm, the normalized SANS intensity I(Q) (cm−1) may be
written as
a
b
c
b
a
C2F5CH2 CH2 OCOCH2 CHdCHeCOOCH2 ′CH2 C2F5; IR (film)
νmax/cm−1: 2979, 1749, 1663, 1464, 1348, 1277, 1195, 1159, 1080,
987, 720, 694.
2.2.2. Synthesis of Sodium Bis(1H,1H,2H,2H-pentafluorobutyl)-2-
sulfoglutarate (2FG(EO)2). Bis(1H,1H,2H,2H-pentafluorobutyl) glu-
taconate (3.29 g, 7.79 mmol) was dissolved in 1,4-dioxane (140 cm3);
then, the mixture was heated to 50 °C. A solution of sodium
hydrogensulfite (3.60 g, 34.4 mmol) in water (60 cm3) was added to
this mixture. The reaction mixture was stirred under reflux for 24 h.
After, solvents were evaporated, leaving a white solid residue, which
was washed with 1,4-dioxane to remove the unreacted diester. The
product was Soxhlet extracted with dry acetone to remove excess
NaHSO3 and recrystallized from acetone. It afforded a white powder,
2FG(EO)2 (with no regio-isomers as C2F5CH2CH2OCOCH2CH-
(−SO3Na)CH2COOCH2CH2C2F5), after vacuum drying (yield 3.11
g, 76.0%); 1H NMR (500 MHz, CF3COOD, TMS), (δH/ppm): 2.73−
2.59 (a, m, 4H), 3.07 (d, dd, 2H, J = 6.0, 17.0 Hz), 3.36 (c, dd, 2H, J =
7.5, 17.0 Hz), 4.34−4.44 (e, m, 1H), 4.69 (b, t, 4H, J = 6.5 Hz) for
I(Q) = Θ(ρ − ρ )2[ VP(Q , R )X(R )] × S(Q , R , Θ )
∑
i
i
i
hs
hs
p
m
i
(2)
P(Q, Ri) is the single-particle form factor. The Schultz distribution
X(Ri) defines the polydispersity using an average radius, Rav, and a
root-mean-squared deviation, σ = Rav/(Z + 1)0.5, Z being a width
parameter. S(Q, Rhs, ϕhs) is the structure factor, and a hard-sphere
model modified for polydispersity was used:39,42 the constraints were
ϕhs = ϕd and Rhs = Radv together with the known ρ values for solvents.
(The subscripts “d” and “hs” denote the droplet and hard-sphere,
respectively). Using the approach of Ottewill et al., eq 2 can be
modified to allow for sharp-step shells built onto a spherical core.43
The least-squares FISH program was used to analyze the SANS
data.19−21,40,41 The fitted parameters are the volume fraction ϕ, the
core radius Racv, polydispersity index σ/Rcav, and shell thickness ts; these
were initially set at physically reasonable values, radii obtained by
preliminary Guinier analyses Rsph, the typical σ/Racv value (0.15)
reported for many common W/O μE systems,44 and the estimated
length of the chains (12.7 Å for n = 8, 7.5 Å for n = 4, and 4.9 Å for n =
2),32 respectively. The sequence of fitting parameters was a follows;
first, at constant σ/Racv and ts values, the other two parameters were
adjusted to the experimental data, and then suitable σ/Racv and ts were
obtained by floating for the best fit (to minimize sum of weighted
squared residuals).
a
b
c
d
b
a
C2F5CH2 CH2 OCOCH2 CH2 CHe(SO3Na)COOCH2 CH2 C2F5; IR
(KBr) νmax/cm−1: 2977, 1736, 1600, 1408, 1348, 1311, 1197, 1080,
800, 720. Elemental analysis for C13H13O7F10Na: found C, 29.1; H,
2.5; S, 7.0. Calcd C, 29.7; H, 2.5.; S, 6.1.
2.3. High-Pressure Small-Angle Neutron Scattering (HP-
SANS) Measurements and Data Analysis. Due to the range of
neutron wavelengths available, time-of-flight SANS is suitable for
studying the shapes and sizes of colloidal systems. HP-SANS is a
particularly important technique for determining surfactant aggrega-
tion structures in supercritical CO2. The HP-SANS measurements of
the D2O/surfactant/scCO2 systems were performed at 45 °C and 350
bar. The LOQ time-of-flight instrument, at the Rutherford Appleton
Laboratory at ISIS UK, was used in conjunction with a stirred high-
pressure cell (Thar). The path length was 10 mm, and the neutron
beam diameter was 10 mm. The measurements gave the absolute
scattering cross section I(Q) (cm−1) as a function of momentum
transfer Q (Å−1), which is defined as Q = (4π/λ)sin θ. The accessible
Q range was 0.007−0.22 Å−1, arising from an incident neutron
wavelength of 2.2−10 Å. The data were normalized for transmission,
empty cell, solvent background, and pressure induced changes in cell
volume as before.20
3. RESULTS AND DISCUSSION
3.1. Characterization of Structure in W/CO2 Micro-
emulsions. Previous studies reported solubilizing powers of
each surfactant listed in Table 1,24,30,31 which were measured by
D
dx.doi.org/10.1021/la400376g | Langmuir XXXX, XXX, XXX−XXX

Langmuir
Article
visual observation and spectroscopically with the water-soluble
dye methyl orange (MO) as an indicator for the μE water pool.
At W0 lower than the solubilizing power, the water/surfactant/
CO2 mixtures form transparent single-phases W/CO2 μE
(IVμEs), but at higher W0 values turbid W/CO2 macro-
emulsions or a precipitate (liquid crystals) are observed.24,30−32
To examine the shape and size of aggregates in surfactant/
D2O/CO2 mixtures at different W0, SANS I(Q) profiles were
measured as a function of W0 at 45 °C and 350 bar, as shown in
Figure 1, with the fitted I(Q) functions added. SANS profiles
phases. Using d = 2π/Q the Q max value (001) allows to
estimate a layer spacing (d) comprising one 2FG(EO)2 bilayer
and an aqueous layer. The calculated d was 62.5 Å at W0 = 30,
increasing to 110.2 Å at W0 = 60. The molecular length of
2FG(EO)2 is approximately 10 Å by MM2 simulations, so
thickness of the bilayers could be ∼20 Å if CO2 was not
incorporated into the interior, and then thickness of the
aqueous layer would be ∼43 Å at W0 = 30 and ∼90 Å at W0 =
60. It is noticed that doubling the W0 ratio yielded almost
double the aqueous layer thickness, according to a typical
swelling behavior45,46 for lamellar LCs. For lamellar LCs, the
relation between d and surfactant volume fraction ϕsurf (ϕsurf
=
1 − ϕsolv, ϕsolv: solvent volume fraction) could be expressed as
follows:46
δ
d =
ϕ
(3)
surf
where δ is a bilayer thickness. On the assumption of absence of
CO2 layer in the lamellar LC, d values calculated by eq 3 were
50 Å at W0 = 30 (ϕsurf = 0.40) and 80 Å at W0 = 60 (ϕsurf
=
0.25). The differences of d values from Bragg peak and eq 3 are
10−30 Å, implying the presence of CO2 incorporated into the
lamellar LC.
One method to approximate radii from SANS data for the
spherical microemulsions is via Guinier plots47 (log [I(Q) vs
Q2]) as shown in Supporting Information (Figures SI1−SI3).
In the all plots, linearity was obtained in the appropriate low Q2
region, and the gradients allowed estimation of a radius of
gyration, Rg (the slope = −Rg2/3). This Rg may also be related
47
to a principal sphere radius Rsph as Rg = (3/5)0.5Rsph
.
The
values of Rsph were calculated and then employed as the starting
points for model fit analyses using the full polydisperse Schultz
sphere model. The parameter outputs are the average radii for
the D2O cores (Rcav) and reversed micelles (Rasv = ts + Racv) and
the polydispersity width (σ/Racv). These fitted parameters with
Rsph are shown in Table 2. As W0 increases, the change in
intensity at low Q, and crossover at high Q, are characteristic of
an increase in droplet size, which is also clear from the fitted
radii. The polydispersities σ/Racv of IVμEs at W0 ≤ 40 ranged
from 0.24 to 0.39, which was 1.2−2.0 times larger than typical
values for reversed microemulsions.44 Such a high polydisper-
sity was reported for W/CO2 IVμEs with the anionic
fluorinated double-tail surfactant di-HCF4 and the cationic
perfluoropolyether surfactant PFPE−TMMA; the σ/Racv values
being 0.17−0.40 at W0 = 5−30 for di-HCF420 and for 0.22−
0.49 at W0 = 19.4−38.1 for PFPE−TMMA.48 As the
fluorocarbon-hydrocarbon hybrid surfactant F7H719 and AOT
analogue surfactants (AOK and AO-Vac)40 were found to give
typical σ/Racv values (<0.2) for W/CO2 IVμEs, the origin of
these high polydispersities is unlikely to be due to the use of
compressed CO2 fluid (i.e., high diffusivity49 increasing
frequency of μE-droplet aggregation/separation).20,48 For the
surfactant/W/CO2 mixtures be thermodynamically stable
microemulsions (energy-minimum equilibrium state), longer
equilibration might be necessary (>1 h) than employed in this
study (∼5 min owing to expense of SANS time),24 even though
clear single phases have already appeared. The limited
equilibration time in this experiment may be one reason for
the higher polydispersity.
Figure 1. SANS profiles for surfactant/D2O/CO2 mixtures with
various W0 at 45 °C and 350 bar (CO2 density = 0.92 g/cm3). Fitted
curves were based on a model incorporating a Schultz distribution of
polydisperse spheres with a core/shell structure. These SANS profiles
are for (a−b) 2FG(EO)2, (c) 4FG(EO)2, (d) 4FS(EO)2, (e)
8FG(EO)2, and (f) 8FS(EO)2. The molar ratio of the surfactant to
CO2 was fixed at 8 × 10−4.
can be useful in determining the shape of colloid particles. In
the low Q region (typically in the case of droplet micro-
emulsions <0.01 Å−1), the scattering may scale as I(Q) ∼ Q−D
,
where D is a characteristic “fractal dimension” for the colloids;
hence, the gradient of a log−log plot will be −D. In the case of
noninteracting spheres, D should be zero in this low Q region,
whereas D = 1 for rods and 2 for disks.30,40−43 For all SANS,
profiles except for 2FG(EO)2 at W0 ≤ 20, the gradients in the
low Q region of log−log plots were found to be ∼0, suggesting
the presence of globular (spherical) nanodomains. On the
other hand, 2FG(EO)2 shows negative slopes at Q < 0.01 and
W0 ≥ 25 and shoulders and peaks at Q of 0.02−0.2. These were
ascribed to Bragg peaks, suggesting formation of liquid crystal
(LC) phases. Earlier lamellar LCs of nFS(EO)2 in water and/or
scCO232 and peaks from Miller indices (001), (002), and (003)
observed in SANS profiles were identified as lamellar LC
3.2. Adsorption and Aggregation Properties of the
Double-Chain Fluorocarbon Solubilizers in W/CO2
Systems. Table 2 also shows that Racv and Rsav increased with
E
dx.doi.org/10.1021/la400376g | Langmuir XXXX, XXX, XXX−XXX

Langmuir
Table 2. Radii for Water-in-CO2 Microemulsions at 45 °C and 350 bar
Article
a
W0
Racv/Å
Rasv/Å
σRacv
Rsph/Å
Rpmax/Å
W0
Racv/Å
Rasv/Å
σRacv
Rsph/Å
Rpmax/Å
4FG(EO2)
10
20
30
40
11.9
17.9
24.6
25.1
27.2
31.4
19.5
25.4
32.1
32.6
34.7
38.9
0.32
0.17
0.24
0.27
0.32
0.42
21.1
24.4
34.9
36.2
42.8
60.8
15.1
17.1
33.2
39.5
44.5
8FG(EO2)
8FS(EO2)
2FG(EO2)
22
44
60
18.6
30.8
41.4
44.2
40.4
10.9
18.9
29.2
39.0
39.0
10.5
24.7
31.0
43.2
53.8
56.4
52.8
23.4
31.3
41.6
51.4
51.4
15.5
29.6
0.39
0.28
0.32
0.27
0.32
0.39
0.38
0.35
0.32
0.32
0.35
0.35
38.0
47.7
63.6
64.0
63.0
23.0
37.2
50.3
60.9
60.6
18.1
35.4
31.5
40.8
54.6
56.9
54.7
b
29.9
34.6
44.0
66
b
b
60
80
0
80
b
10
22
40
4FS(EO2)
31.3
40.0
52.2
52.6
10
20
40
60
10.2
18.1
28.6
29.2
17.7
25.6
36.1
36.7
0.25
0.38
0.25
0.32
b
28.5
32.5
37.0
60
b
80
b
10
20
a
Radii of aqueous core and shell of reversed micelles (Racv and Rsav) and polydispersity index σRcav were obtained by fitting of theoretical curve of
polydispersed core/shell-structured sphere to experimental values. Rsph and Rpmax were obtained from the slopes of Guinier plots and first maxima in
b
Porod plots, respectively. Turbid phase.
W0 and Rasv was slightly different from Rsph in IVμE phases,
namely, at W0 below the maximum solubilizing powers listed in
{I(Q) × Q4}Q→∞ = 2π(Δρ)2Σ
(4)
Table 1. To confirm the reliability of the radii obtained by
fitting and to estimate an effective area per headgroup Ah,p for
nFG(EO)2 in W/CO2 IVμE, Porod plots were generated.
Figure 2 shows Porod plots for 4FGEO2, and those of the other
where Σ is the total area per unit volume. Another requirement
here is that the cmc (concentration of nonadsorbed free
surfactant) in CO2 is negligible compared with the
experimental surfactant concentration; that is, all N surfactant
molecules are at the CO2−D2O interface; given those
conditions then the area per headgroup, Ah,p ≈ Σ/N. In the
Porod plots of Figure 2 at Q > 0.16 Å−1, the data are essentially
asymptotic, suggesting these approximations are reasonable.
The Δρ in eq 4 was assumed to be {(ρD O + ρsurf)/2 − ρCO
}
2
2
because the scattering will arise from not only the D2O core but
also the surfactant shell (ρsurf were 2.33−3.24 × 1010 cm−2
against ρCO ∼ 2.29 × 1010 cm−2) and volume fractions of total
2
surfactant molecular fragments in the W/CO2 μE interfacial
region were calculated as ∼0.5 (see Supporting Information,
Figure SI6 and Tables SI1−SI2). The broken lines at each W0
indicate asymptotes {I(Q)·Q4}Q→∞ giving Ah values as listed in
Table 3. Except at W0 = 60 gives rise to a macroemulsion, Ah,p
values of same surfactant are similar within an experimental
uncertainty 10 Å2.
For Porod plots of spherical particles, with these character-
istic polydispersities, the position of the first maximum and first
minimum typically occurs at ∼2.7/R and ∼4.5/R in Q,
respectively.50 The Porod plots display only the first maxima,
but not the first minima clearly. By using the first maxima the
μE droplet radii were estimated (Table 2); in the table, Rpmax
means the radius given from the first Porod maximum. As seen
in the table Rsav and Rpmax were similar at same W0, supporting
the validity of these analyses.
Figure 2. Porod plots to obtain μE droplet radii Rpmax and areas per
headgroup Ah,p in D2O/CO2 mixtures with 4FG(EO)2 for several W0
values at 45 °C and 350 bar. Each arrow shows the first maximum used
for calculating Rpmax. Broken lines display {I(Q)Q4}Q→∞ for the
calculation of Ah,p.
surfactants are displayed in Supporting Information (Figures
SI4 and SI5). For W/O μE droplets, Ah,p can be estimated from
high Q SANS data, by assuming a sharp interface and applying
the Porod equation:50
Table 3. Area (Ah,p and Ah,l), Volume (νh,l), and Radius (rh,l) of Head Group and Chain Length (lc) of nFG(EO)2 in D2O/CO2
IVμE at 45 °C and 350 bar
a
Ah,p /Å2
b
b
b
c
W0
10
20
30
40
60
Ah,1 /Å2
νh,1 /Å3
rh,1 /Å
lc /Å
f
f
4FG(EO2)
4FS(EO2)
4FG(EO2)
4FS(EO2)
131
140
136
146
129
116
111
151
123
167
129
128
129
125
117
231
213
215
199
3.8
3.7
3.7
3.6
7.7
7.6
d
d
e
114
108
12.4
12.5
f
a
b
Calculated by Porod plot on the assumption of Δρ = (ρD O + ρsurfactant)/2 − ρCO . The uncertainties are 10 Å2. Obtained from a linear function of
2
2
c
Racv vs W0 as shown in Figure 2. The uncertainties of Ah are 5 Å2. Yielded by subtracting intercept of a linear function of Rcav vs W0 from that of Rasv
vs W0, in other words, lc = lsurf − lh, as shown in Figure 2. At W0 = 22. At W0 = 44. Turbid phase.
d
e
f
F
dx.doi.org/10.1021/la400376g | Langmuir XXXX, XXX, XXX−XXX

Langmuir
Article
Figure 3 shows changes in the radii obtained by the
theoretical curve fitting for D2O cores and reversed micelle
addition, the Kraft temperature shown in Table 1 shows that
4FG(EO)2 seems to be more soluble in water, meaning the
actual area per headgroup of 4FG(EO)2 should be lower than
Ah,P and could be similar to the Ah,l.
In a previous work,21 fluorinated double-tail surfactants (di-
HCF4, di-HCF6, and di-CF4) having the same head groups as
nFG(EO)2 and nFS(EO)2 were investigated in a similar fashion
to as described above, giving 115 5 Å2 for Ah (15 °C and 500
bar). Interestingly, and importantly, that value is almost similar
as found here, not only for the sulfosuccinate nFS(EO)2 but
also the analogous glutarate surfactant nFG(EO2), on account
of difference of 30 °C in experimental temperature. For AOT in
a range of n-alkanes from propane to decane, a mean value of
72 Å2 for Ah was reported.41 The difference in Ah between W/
CO2 and W/O μEs suggests that packing density of AOT type
surfactants at water surface increases in CO2 as compared to
typical hydrocarbon oil. Earlier papers reported Ah of AOT to
Figure 3. Dependence of droplet radius for the 8FG(EO)2 D2O/CO2
μEs as a function of W0 at 45 °C and 350 bar.
52
be 155 Å2 at a W/scCO2 but 75 Å2 at the water/air53 and 72
Å2 at the water/alkane interface.41
shells as a function of W0. When comparing nFS(EO)2 and
nFG(EO)2, no significant difference in D2O core radii could be
seen at constant W0 for the IVμEs. Differences in FC length n
did not affect the radius in the IVμE region, but at W0 ≥ 60 an
increase of ∼10 Å was seen in the radius between n = 4 and 8
surfactants. This suggests the following: (1) the turbid two-
phase system at high W0 was identified as a Winsor-II W/
CO2μE, and (2) excess water molecules (= W0 − the
solubilizing power) formed W/CO2 “macro”-emulsion droplets
and separated from scCO2 in the absence of stirring while the
equilibrium reversed micelles still remain in scCO2. This
behavior was suggested by changes in UV−vis light absorption
of the surfactant/aqueous methyl orange solution/CO2
mixtures as a function of W0, as seen previously.51 In addition,
the linear dependences for radii vs W0 suggested that D2O
molecules are incorporated into spherical μE droplets. This
linear behavior for Rcav as a function of W0 can be used to
provide another estimate for the effective headgroup area
using10
Although the values of radii Racv in Figure 3 are approximately
half those seen for AOT W/alkane μEs,10 this could be
accounted for owing to the difference in interfacial packing and
the determined Ah values. In Figure 3, intercepts of linear
functions of Racv and Rsav vs W0 give mean radii of the headgroup
core and the dry reversed micelle, in other words, dimensions
of headgroup (lh) and overall surfactant molecule (lsurf),
respectively. The head groups −SO3Na of nFG(EO)2 and
nFS(EO)2 displayed the same radius in dry micelle cores (∼5.2
Å). The lsurf values obtained from Figure 3 were almost
consistent with those calculated by the MM2 calculation,
suggesting Racv and Rasv values in Figure 3 to be reasonable: for
example, lsurf of 8FG(EO)2 was 17.6 Å from the intercept
(radius of dry reversed micelles) of Figure 3 and 18 Å obtained
by MM2 calculation.
By using the intercept for lh and eq 5, the effective headgroup
volume was calculated as vh = 199−231 Å3, and assuming the
headgroup to be spherical, the radius (rh) would be 3.6−3.8 Å,
as summarized in Table 3. AOT analogues in W/O μEs were
found to have vh = 200−236 Å3 and rh = 3.8 Å,10 and these are
consistent with the values calculated here for surfactants with
identical polar head groups. This suggests the headgroup
volume is not affected by being in supercritical CO2, making it
unlikely the CO2 penetrates into the polar region of the
micellar structure.
3v
w w0 +
Ah
3vh
Ah
α(p)Rcav
=
(5)
where vw is the volume of a water molecule, p is the
polydispersity index (σ/Rcav), and α(p) depends on the Schultz
distribution α(p) = 1 + 2p2. Therefore, assuming the
polydispersity to be independent of W0, the broken line for
Racv in Figure 3 is expected to have a slope which depends on
the interfacial area Ah and an intercept which depends on the
average headgroup volume vh via the headgroup radius rh. The
headgroup areas calculated in this way span 117−129 Å2 with
an uncertainty of 10 Å2 (Table 3), which is consistent with
the previous results obtained by Porod analyses of high Q
SANS data in W/CO2 μEs. Areas per headgroup Ah,p of
4FG(EO)2 obtained by Porod analysis were slightly larger than
Ah,l obtained from the slope of Figure 3. The assumption in
these calculations for Ah,p is that all surfactant molecules are at
the W/CO2 interface, which may not be the case: the
computed area will be too large if the solubilities of surfactant
in CO2 and water are high. In a previous study,31 the
dissolution pressures of 0.08 mol % surfactant in scCO2 were
found to decrease at shorter FC chain lengths for nFG(EO)2
and nFS(EO)2, suggesting the solubility of 4FG(EO)2 in pure
CO2 (i.e., number of free 4FG(EO)2 molecules not at the W/
CO2 interface) is larger than those of longer FC surfactants. In
These parameters relating to the size of the chain and
headgroup allow estimates of the aggregation number numbers
Nagg and critical packing parameters (CPP)33 by using
4π(Rcav)2
N
=
agg
Ah
(6)
(7)
v
Ahlc
c
CPP =
where vc is the hydrophobic chain volume. If the hydrophobic
part is assumed to be a truncated core, the volume should be;
lc(Ah + Ac + AhAc )
v =
c
(8)
3
where Ac is area per hydrophobic chain terminus, obtained
using Ac = 4π(Rasv)2/Nagg. To calculate Nagg and CPP of
surfactants used in this study, Ah values obtained from eq 5
were employed here. The calculated Nagg and CPP are shown as
G
dx.doi.org/10.1021/la400376g | Langmuir XXXX, XXX, XXX−XXX

Langmuir
Article
a function of W0 in Figure 4 with previous data for AOT/W/n-
heptane μEs.10 As expected from typical behavior of W/O
example, HCB,15 Winsor R,34,35 and/or W/CO2 interfacial
tension5,52 mainly affect the solubilizing power of the double-
FC-tail surfactants in IVμEs.
On the other hand, CPPs of the sulfosuccinate (8FS(EO)2
and 4FS(EO)2) and the glutarate (8FG(EO)2 and 4FG(EO)2)
were found to be similar at same chain lengths n in the W/CO2
μEs, suggesting that the extra methylene in nFG(EO)2 did not
affect CPP: that is, CPP is not the main parameter explaining
differences in solubilizing power. Then, the question “Why can
the glutarate surfactants solubilize a higher W0 than the
succinate surfactants can?” still remains unsolved.
The subtle structural change in the headgroup region (i.e.,
succinate → glutarate) has also been investigated before for
hydrocarbon AOT-like surfactants.56 Interestingly, the addition
of just one extra −CH2− in the headgroup had little noticeable
effect on most interfacial properties at both air/water and oil/
water interfaces; for example, in aqueous systems, the
differences in CMC, surface tension, and area per headgroup
at CMC were 0.8 mM, <0.1 mN/m, and <7 Å2, respectively.56
However, a maximum attainable W0 ratio with the glutarate
analogue in W/n-heptane μEs was higher by >10 as compared
with the normal succinate (10−40 °C).56 Previously studies
have been made of succinate and glutarate fluoro-AOT-type
analogues in hydrocarbon W/CO2 μEs, suggesting that there is
a notable enhancement in stability for the glutarate21,22 (i.e.,
comparing bis(1H,1H,5H-octafuoropentyl)-2-sulfosuccinate or
di-HCF4, versus bis(1H,1H,5H-octafuoropentyl)-2-glutarate or
di-HCF4GLU). However, that work was only limited,
comparing the pressure−temperature (P−T) phase stability
of the W/CO2 μEs at one fixed W0 = 10: at 25 °C the phase
transition pressure Ptrans was ∼12 bar lower for the glutarate
analogue (181 vs 193 bar).21,22 The higher solubilizing powers
of glutarates have been often reported not only in scCO2 but
also in typical organic solvents. Without a hypothesis based on
CPP for the enhanced solubilizing power, there are another two
possibilities: the extra methylene spacer results in (1) an
increase in structural disorder by lowering molecular symmetry
leading to a weaker molecular packing, and (2) a decrease in
HCB by increasing hydrophobicity of linking group promoting
(possible) CO2-ester interactions57 instead of hydration. These
possibilities could be investigated by measuring FT-IR spectra
for the ester groups in the two water/scCO2 mixtures with
nFG(EO)2 and nFS(EO)2.58
The effect of FC length n of the double-FC-tail surfactants
on CPP can be seen in Figure 4: longer chains result in larger
CPP values (longer surfactants have larger shells in spite of the
core radius being similar, independent of FC length). As listed
in Table 1, the solubilizing power of nFG(EO)2 at 45 °C
increases with an increase in n, and this trend would be
accelerated by the higher CPP at larger n values. On the other
hand, the solubilizing power of 4FG(EO)2 at 75 °C was the
highest. An earlier paper30,31 discussed the reason why
4FG(EO)2 is the most efficient solubilizer at temperature >65
°C, and one of the proposed reasons was the formation of
bicontinuous microemulsions at high temperatures. Unfortu-
nately, the HP-SANS cell used is restricted and cannot be used
at high temperature, so that bicontinuous microemulsions
could not be directly confirmed. However, lamellar LCs were
observed in 2FG(EO)2/W/CO2 mixtures at W0 ≥ 25, raising
the possibility of formation of 4FG(EO)2 bicontinuous
microemulsions. In general, a negative curvature of W/O (or
W/scCO2) μEs approaches zero with decreasing CPP33 and/or
increasing HLB (or HCB),15 and the microemulsion often
Figure 4. Change in aggregation number (Nagg) and critical packing
parameter (CPP) of nFG(EO)2 and nFS(EO)2 in D2O/CO2μEs or
AOT in D2O/n-heptane μEs10 as a function of W0 at 45 °C and 350
bar.
reversed micelle systems,54,55 Nagg for these W/CO2 μEs
increased with W0. For example, Nagg was close to 10 at W0 =
10, but it reached up to ∼100 and ∼200 for 4FG(EO)2 at W0 =
80 and 8FG(EO)2 at W0 = 66, respectively. When the Nagg data
were compared at a constant W0, significant differences were
not observed between these FC surfactants. On the other hand,
Nagg values of the FC surfactants in scCO2 μEs were found to
be quite small, one-fifth to one-sixth of those of AOT/W/n-
heptane μEs. Even if compared with AOT in the other organic
solvents (e.g., Nagg = ∼90 at W0 = 10 and ∼290 at W0 = 20 in
isooctane and cyclohexane54,55), the difference in Nagg is quite
large, resulting from the small Rasv and large Ah for nFG(EO)2
and nFS(EO)2 in IVμEs.
The critical packing parameter (CPP)33 is an indicator of
“bulkiness” of a given hydrophobic group compared to the
headgroup, and dictates interfacial curvature of a surfactant
assembly. Negative interfacial curvature for a W/O or W/CO2
μEs is only stabilized for surfactants with CPP > 1, and if CPP
∼1 the interfacial curvature is planar: in other words, the
reversed curvature μE droplets would tend to more planar
macroemulsions. At W0 = 10 in Figure 4, the CPP values are
1.92, 1.76, and 2.57 for 4FG(EO)2, 4FS(EO)2, and 8FG(EO)2,
respectively. These values were found to decrease with
increasing W0 and then finally reached constant values of
∼1.3 at W0.
3.3. Effects of Fluorocarbon and Methylene Spacer
Lengths on Solubilization. Earlier SANS data for AOT/W/
n-heptane μEs10 were used to calculate CPPs, represented by
closed circles in Figure 4. The AOT-analogue glutarate in W/n-
heptane μEs also had CPP values similar to the regular AOT
(e.g., 1.18 at 25 °C and W0 = 40).56 These are always smaller
than those for the FC-surfactants in scCO2 at any W0 and
demonstrate that even at small CPP of ∼1.15 can stabilize
Winsor-IV W/O μEs, whereas such a small CPP is not possible
in the W/CO2 μEs. It implies that other parameters, for
H
dx.doi.org/10.1021/la400376g | Langmuir XXXX, XXX, XXX−XXX

Langmuir
Article
turns into a lamellar LC or a bicontinuous system bearing net
zero interfacial curvature. In the case of 4FG(EO)2, the CPP
and HCB values should be close to those of 2FG(EO)2, which
prefers zero interfacial curvature, but the ability to lower W/
CO2 interfacial tension could be higher than that of 2FG(EO)2
owing to the longer FC chains. In addition, affinities of
4FG(EO)2 tail-to-CO2 and head-to-water would be enhanced
at higher temperature, which would in turn enhance the ability
to lower interfacial tension. The discussion about CPP and
HCB suggests that the 4FG(EO)2 bicontinuous micro-
emulsions, which should require not only zero interfacial
curvature but also an extremely low W/CO2 interfacial tension
<1 mN/m, are highly probable.
be first example of single-phase bicontinuous microemulsions
for a ternary single-surfactant/W/CO2 mixture. However, in
earlier papers bicontinuous microemulsions were found but in
equilibrium with separated water and CO2 phases59 or by using
mixed fluorinated surfactants (Zonyl FSH/Zonyl FSN 100
mixtures60). High water content bicontinuous microemulsions
could represent new generation solvents with unique proper-
ties. Further studies on 4FG(EO)2 and its analogues should be
continued to advance surfactant design theory for W/CO2 μE
and bicontinuous systems.
ASSOCIATED CONTENT
* Supporting Information
■
S
Guinier and Porod plots for W/CO2 μEs with nFG(EO)2 and
nFS(EO)2 (Figures SI1−5). Calculation of the volume fraction
of total surfactant molecular fragments in the W/CO2 μE
interfacial region, Φsurf (Tables SI1−2). This material is
available free of charge via the Internet at http://pubs.acs.org.
4. CONCLUSIONS
A previous study found the glutarate headgroup compound
4FG(EO)2 to be an effective and efficient double-FC-tail
surfactant for water-in-scCO2 microemulsions: this compound
can solubilize water of W0 up to 80 at 75 °C, in spite of the
short C4 FC tails. At high temperature >65 °C, the solubilizing
power was the highest in for FG(EO)2 analogues with different
FC tail lengths (n = 2, 4, 6, and 8). Microemulsions in CO2 are
promising solvents for green chemistry and therefore should be
prepared with low levels of surfactant, being also inexpensive
and environmentally benign. Therefore, finding short FC-tail
surfactants, which generate high solubilizing power, is key to
designing useful CO2-philic surfactants. Further surfactant
structure-performance studies are needed to develop CO2 as
an “environmentally-benign” and “energy-saving” solvent, for
applications such as extraction, dyeing, dry cleaning, metal-
plating, and organic or nanomaterial synthesis.
This study characterized microemulsions with the double-
FC-tail solubilizers nFG(EO)2 and nFS(EO)2, and clarified the
effects of FC tail and methylene spacer lengths on the
nanostructure (radii of D2O cores and reversed micelle shells,
area per headgroup, aggregation number, and critical packing
parameter CPP) of the microemulsions. Through SANS
experiments and data analyses, relationships between nano-
structure and the surfactant chain length or composition W0
were revealed: (1) the core and shell radii and aggregation
number increased with W0, but these were almost independent
of addition of extra methylene spacer at same FC length and
W0; (2) values of area per headgroup for each surfactant were
almost the same based on the common −SO3Na headgroup
employed, and (3) the CPPs of FC surfactants decreased for
the shorter FC lengths and larger W0 values. These changes
account for the observed transition of microemulsions into
lamellar LCs as a function of surfactant structure and
microemulsion composition. It was also interesting to note
that the nanostructures (core radius) and surfactant properties
(Ah,l, Nagg, and CPP) in the CO2 IVμEs were significantly
different from those in typical AOT/W/O μEs at same W0, in
spite of similar surfactant structure and concentration. Those
differences could be down to significant effects of intermo-
lecular interactions between hydrocarbon chains and hydro-
carbon solvents, as compared with fluorocarbon chains and
CO2.
AUTHOR INFORMATION
Corresponding Author
■
*M.S.: E-mail: sagisaka@cc.hirosaki-u.ac.jp. Phone and fax:
+81-172-39-3569. J.E.: E-mail: julian.eastoe@bristol.ac.uk.
Phone: +44-117-928-9180. Fax: +44-117-925-1295.
Notes
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
■
M.S. thanks the Japan Society for the Promotion of Science
(JSPS) for a 1-year fellowship (Excellent Young Researcher
Overseas Visit Program) and Hirosaki University. S.C. thanks
the EPSRC for a PhD scholarship under the Next Generation
Facility User scheme (EP/F020686). A.M. thanks the Ministry
of Higher Education of Malaysia and Universiti Pendidikan
Sultan Idris for the provision of a Ph.D. studentship. We also
acknowledge STFC for the allocation of beam time, travel, and
consumables grants at ISIS. This project was supported by JSPS
[KAKENHI, Grant-in-Aid for Young Scientists (A), No.
23685034], JST [A-STEP, No. AS231Z02302D], and Leading
Research Organizations, namely, NSERC, ANR, DFG, RFBR,
RCUK, and NSF as Partner Organizations under the G8
Research Councils Initiative for Multilateral Research Funding,
the Noguchi Institute Foundation, and Nissan Chemical
Industries. The EPSRC is thanked for funding under grants
EP/C523105/1, EP/F020686, and EP/K020676/1.
REFERENCES
■
(1) Beckman, E. J. Supercritical and Near-Critical CO2 in Green
Chemical Synthesis and Processing. J. Supercrit. Fluids 2004, 28, 121−
191.
(2) Goetheer, E. L. V.; Vortaman, M. A. G.; Keurentjes, J. T. F.
Opportunities for Process Intensification Using Reverse Micelles in
Liquid and Supercritical Carbon Dioxide. Chem. Eng. Sci. 1999, 54,
1589−1596.
(3) Consani, K. A.; Smith, R. D. Observations on the Solubility of
Surfactants and Related Molecules in Carbon Dioxide at 50 °C. J.
Supercrit. Fluids 1990, 3, 51−65.
(4) Ryoo, W.; Webber, S. E.; Johnston, K. P. Water-in-Carbon
Dioxide Microemulsions with Methylated Branched Hydrocarbon
Surfactants. Ind. Eng. Chem. Res. 2003, 42, 6348−6358.
(5) Sagisaka, M.; Koike, D.; Mashimo, Y.; Yoda, S.; Takebayashi, Y.;
Furuya, T.; Yoshizawa, A.; Sakai, H.; Abe, M.; Otake, K. Water/
Supercritical CO2 Microemulsions with Mixed Surfactant Systems.
Langmuir 2008, 24, 10116−10122.
Although the reason why 4FG(EO)2 displays the highest
solubilizing power (maximum W0 = 80) at the high
temperatures is still unclear, one conceivable reason is
formation of W/CO2-type bicontinuous microemulsions, and
it seems highly probable based on fact (3) mentioned above. If
bicontinuous μEs form in the 4FG(EO)2 mixtures, this would
I
dx.doi.org/10.1021/la400376g | Langmuir XXXX, XXX, XXX−XXX

Langmuir
Article
(6) Sagisaka, M.; Fujii, T.; Koike, D.; Yoda, S.; Takebayashi, Y.;
Furuya, T.; Yoshizawa, A.; Sakai, H.; Abe, M.; Otake, K. Surfactant-
Mixing Effects on the Interfacial Tension and the Microemulsion
Formation in Water/Supercritical CO2 System. Langmuir 2007, 23,
2369−2375.
(7) Johnston, K. P.; Cho, D.; DaRocha, S. R. P.; Psathas, P. A.; Ryoo,
W.; Webber, S. E.; Eastoe, J.; Dupont, A.; Steytler, D. C. Water in
Carbon Dioxide Macroemulsions with a Hydrocarbon Surfactant.
Langmuir 2001, 17, 7191−7193.
(8) Hollamby, M. J.; Trickett, K.; Mohamed, A.; Cummings, S.;
Tabor, R. F.; Myakonkaya, O.; Gold, S.; Rogers, S.; Heenan, R. K.;
Eastoe, J. Tri-Chain Hydrocarbon Surfactants as Designed Micellar
Modifiers for Supercritical CO2. Angew. Chem., Int. Ed. 2009, 48,
4993−4995.
(9) Lee, H.; Pack, J W.; Wang, W.; Thurecht, K. J.; Howdle, S. M.
Synthesis and Phase Behavior of CO2-Soluble Hydrocarbon
Copolymer: Poly(Vinyl Acetate-alt-Dibutyl Maleate). Macromolecules
2010, 43, 2276−2282.
(10) Nave, S.; Eastoe, J.; Heenan, R. K.; Steytler, D.; Grillo, I. What is
So Special About Aerosol-OT? 2. Microemulsion Systems. Langmuir
2000, 16, 8741−8748.
(11) Hoefling, T. A.; Newman, D. D.; Enick, E. J.; Beckman, E. J.
Effect of Structure on the Cloud-Point Curves of Silicone-Based
Amphiphiles in Supercritical Carbon Dioxide. J. Supercrit. Fluids 1993,
6, 165−171.
(12) Psathas, P. A.; da Rocha, S. R. P.; Lee, C. T.; Johnston, K. P.;
Lim, K. T.; Webber, S. Water-in-Carbon Dioxide Emulsions with
Poly(dimethylsiloxane)-based Block Copolymer Ionomers. Ind. Eng.
Chem. Res. 2000, 39, 2655−2664.
(13) Johnston, K. P. Block Copolymers as Stabilizers in Supercritical
Fluids. Curr. Opin. Colloid Interface Sci. 2000, 5, 351−356.
(14) da Rocha, S. R. P.; Dickson, J.; Cho, D.; Rossky, J. P.; Johnston,
K. P. Stubby Surfactants for Stabilization of Water and CO2
Emulsions: Trisiloxanes. Langmuir 2003, 19, 3114−3120.
(15) Lee, C. T., Jr.; Psathas, P. A.; Johnston, K. P.; deGrazia, J.;
Randolph, T. W. Water-in-Carbon Dioxide Emulsions: Formation and
Stability. Langmuir 1999, 15, 6781−6791.
(16) Johnston, K. P.; Harrison, K. L.; Klarke, M. J.; Howdle, S. M.;
Heitz, M. P.; Bright, F. V.; Carlier, C.; Randolph, T. W. Water-in-
Carbon Dioxide Microemulsions: A New Environment for Hydro-
philes Including Proteins. Science 1996, 271, 624−626.
(17) Zielinski, R. G.; Kline, S. R.; Kaler, E. W.; Rosov, N. A Small-
Angle Neutron Scattering Study of Water in Carbon Dioxide
Microemulsions. Langmuir 1997, 13, 3934−3937.
(18) Niemeyer, E. D.; Bright, F. V. The pH within PFPE Reverse
Micelles Formed in Supercritical CO2. J. Phys. Chem. B 1998, 102,
1474−1478.
(19) Eastoe, J.; Bayazit, Z.; Martel, S.; Steytler, D. C.; Heenan, R. K.
Droplet Structure in Water-in-CO2 Microemulsion. Langmuir 1996,
12, 1423−1424.
Preparation of a W/scCO2 Microemulsion Using Fluorinated
Surfactants. Langmuir 2003, 19, 220−225.
(25) Lu, L.; Berkowitz, M. L. Molecular Dynamics Simulation of a
Reverse Micelle Self Assembly in Supercritical CO2. J. Am. Chem. Soc.
2004, 126, 10254−10255.
(26) Salaniwal, S.; Cui, S. T.; Cummings, P. T.; Cochran, H. D. Self-
Assembly of Reverse Micelles in Water/Surfactant/Carbon Dioxide
Systems by Molecular Simulation. Langmuir 1999, 15, 5188−5192.
(27) Dalvi, V. H.; Srinivasan, V.; Rossky, P. J. Understanding the
Effectiveness of Fluorocarbon Ligands in Dispersing Nanoparticles in
Supercritical Carbon Dioxide. J. Phys. Chem. C 2010, 114, 15553−
15561.
(28) Dalvi, V. H.; Srinivasan, V.; Rossky, P. J. Understanding the
Relative Effectiveness of Alkanethiol Ligands in Dispersing Nano-
particles in Supercritical Carbon Dioxide and Ethane. J. Phys. Chem. C
2010, 114, 15562−15573.
(29) Mohamed, A.; Sagisaka, M.; Guittard, F.; Cummings, S.; Paul,
A.; Rogers, S. E.; Heenan, R. K.; Dyer, R.; Eastoe, J. Low Fluorine
Content CO2-Philic Surfactants. Langmuir 2011, 27, 10562−10569.
(30) Sagisaka, M.; Iwama, S.; Hasegawa, S.; Yoshizawa, A.;
Mohamed, A.; Cummings, S.; Rogers, S. E.; Heenan, R. K.; Eastoe,
J. Super-Efficient Surfactant for Stabilizing Water-in-Carbon Dioxide
Microemulsions. Langmuir 2011, 27, 5772−5780.
(31) Sagisaka, M.; Iwama, S.; Yoshizawa, A.; Mohamed, A.;
Cummings, S.; Eastoe, J. An Effective and Efficient Surfactant for
CO2 Having Only Short Fluorocarbon Chains. Langmuir 2012, 28,
10988−10996.
(32) Sagisaka, M.; Koike, D.; Yoda, S.; Takebayashi, Y.; Furuya, T.;
Yoshizawa, A.; Sakai, H.; Abe, M.; Otake, K. Optimum Tail Length of
Fluorinated Double-Tail Anionic Surfactant for Water/Supercritical
CO2 Microemulsion Formation. Langmuir 2007, 23, 8784−8788.
(33) Israelachvili, J. N. Measurements of Hydration Forces Between
Macroscopic Surfaces. Chem. Scr. 1985, 25, 7−14.
(34) Winsor, P. A. Hydrotropy, Solubilisation and Related
Emulsification Processes. Trans. Faraday. Soc. 1948, 54, 376−398.
(35) Harrison, K. L.; Johnston, K. P.; Sanchez, L. C. Effect of
Surfactants on the Interfacial-Tension Between Supercritical Carbon-
Dioxide and Polyethylene-Glycol. Langmuir 1996, 12, 2637−2644.
(36) Span, R.; Wahner, W. A New Equation of State for Carbon
Dioxide Covering the Fluid Region from the Triple-Point Temper-
ature to 1100 K at Pressures up to 800 MPa. J. Phys. Chem. Ref. Data
1996, 25, 1509−1596.
(37) McClain, J. B.; Londono, D.; Combes, J. R.; Romack, T. J.;
Canelas, D. A.; Betts, D. E.; Wignall, G. D.; Samulski, E. T.;
DeSimone, J. M. Solution Properties of a CO2-Soluble Fluoropolymer
via Small Neutron Scattering. J. Am. Chem. Soc. 1996, 118, 917−918.
(38) Sears, V. F. Neutron Scattering Lengths and Cross Sections.
Neutron News 1992, 3, 26−37.
(39) Kotlarchyk, M.; Chen, S.-H.; Huang, J. S.; Kim, M. W. Structure
of Three-Component Microemulsions in the Critical Region
Determined by Small Angle Neutron Scattering Data. Phys. Rev. A
1984, 29, 2054−2069.
(20) Eastoe, J.; Cazelles, B. M. H.; Steytler, D. C.; Holmes, J. D.; Pitt,
A. R.; Wear, T. J.; Heenan, R. K. Water-in-CO2 Microemulsions
Studied by Small-Angle Neutron Scattering. Langmuir 1997, 13,
6980−6984.
(40) Eastoe, J.; Gold, S.; Rogers, S.; Wyatt, P.; Steytler, D. C.; Gurgel,
A.; Heenan, R. K.; Fan, X.; Beckman, E. J.; Enick, R. M. Designed
CO2-Philes Stabilize Water-in-Carbon Dioxide Microemulsions.
Angew. Chem. 2006, 118, 3757−3759.
(41) Eastoe, J.; Robinson, B. H.; Young, W. K.; Steytler, D. C.
Scattering Studies of Microemulsions in Low-Density Alkanes. J.
Chem. Soc., Faraday Trans. 1990, 86, 2883−2889.
(42) Hayter, J. B.; Penfold, J. Determination of Micelle Structure and
Charge by Neutron Small Angle Scattering. Colloid Polym. Sci. 1983,
261, 1022−1030.
(21) Eastoe, J.; Downer, A.; Paul, A.; Steytler, D. C.; Rumsey, E.;
Penfold, J.; Heenan, R. K. Fluoro-Surfactants at Air/Water and Water/
CO2 Interfaces. Phys. Chem. Chem. Phys. 2000, 2, 5235−5242.
(22) Eastoe, J.; Paul, A.; Downer, A.; Steytler, D. C.; Rumsey, E.
Effects of Fluorocarbon Surfactant Chain Structure on Stability of
Water-in-Carbon Dioxide Microemulsions. Links Between Aqueous
Surface Tension and Microemulsion. Langmuir 2002, 18, 3014−3017.
(23) Xu, B.; Lynn, G. W.; Guo, J.; Melnichenko, Y. B.; Wignall, G. D.;
McClain, J. B.; DeSimone, J. M.; Johnston, C. S., Jr. NMR and SANS
Studies of Aggregation and Microemulsion Formation by Phosphorus
Fluorosurfactants in Liquid and Supercritical Carbon Dioxide. J. Phys.
Chem. B 2005, 109, 10261−10269.
(43) Markovic, I.; Ottewill, R. H.; Cebula, D. J.; Field, I.; Marsh, J.
Small Angle Neutron Scattering Studies on Non-Aqueous Dispersions
of Calcium Carbonate. Part 1. The Guinier Approach. Colloid Polym.
Sci. 1984, 262, 648−656.
(44) Almgren, M.; Johannsson, R.; Eriksson, J. C. Polydispersity of
́
(24) Sagisaka, M.; Yoda, S.; Takebayashi, Y.; Otake, K.; Kitiyanan, B.;
Kondo, Y.; Yoshino, N.; Takebayashi, K.; Sakai, H.; Abe, M.
AOT Droplets Measured by Time-Resolved Fluorescence Quenching.
J. Phys. Chem. 1993, 97, 8590−8594.
J
dx.doi.org/10.1021/la400376g | Langmuir XXXX, XXX, XXX−XXX

Langmuir
Article
(45) Figueiredo Neto, A. M.; Salinas, S. R. A. The Physics of Lyotropic
Liquid Crystals, Phase Transition and Structual Properties; Oxford
University Press: New York, 2005.
(46) Roux, D.; Safinya, C. R. A Synchrotron X-Ray Study of
Competing Undulation and Electrostatic Interlayer Interactions in
Fluid Multimembrane Lyotropic Phases. J. Phys. (Paris) 1988, 49,
307−318.
(47) Guinuier, A.; Fournet, G. Small-Angle Scattering of X-Rays;
Wiley: New York, 1956.
(48) Lee, C. T.; Psathas, P. A.; Ziegler, K. J.; Johnston, K. P.; Dai, H.
J.; Cochran, H. D.; Melnichenko, Y. B.; Wignall, G. D. Formation of
Water-in-Carbon Dioxide Microemulsions with a Cationic Surfactant:
a Small-Angle Neutron Scattering Study. J. Phys. Chem. B 2000, 104,
11094−11102.
(49) Gopalan, A. S.; Wai, C. M.; Jacobs, H. K., Eds. Supercritical
Carbon Dioxide; Separations and Processes; American Chemical Society:
Washington, DC, 2003.
(50) Porod, G. Die Rontgenkleinwinkelstreuung von dichtgepackten
̈
kolloiden Systemen. Kolloid Zeit. 1951, 124, 83−114.
(51) Sagisaka, M.; Hino, M.; Oasa, J.; Yamamoto, M.; Yoda, S.;
Takebayashi, Y.; Furuya, T.; Yoshizawa, A.; Ochi, K.; Otake, K.
Characterization of Water/Supercritical CO2 Microemulsion by UV-
Visible Spectroscopy and Dynamic Light Scattering. J. Oleo Sci. 2009,
58, 75−83.
(52) Sagisaka, M.; Fujii, T.; Ozaki, Y.; Yoda, S.; Takebayashi, Y.;
Kondo, Y.; Yoshino, N.; Sakai, H.; Abe, M.; Otake, K. Interfacial
Properties of Branch-Tailed Fluorinated Surfactants Yielding a Water/
Supercritical CO2 microemulsion. Langmuir 2004, 20, 2560−2566.
(53) Antelmi, D. A.; Kekicheff, P. Measurement of the Compressi-
́
bility Modulus in a Lyotropic Lamellar Phase Stabilized by Undulation
Forces. J. Phys. Chem. B 1997, 101, 8169−8179.
́
(54) Izquierdo, C.; Moya, M. L.; Usero, J. L.; Casado, J. Estimation of
the Aggregation Number and Core Radius of Microemulsions.
Monatsh. Chem. 1992, 123, 383−389.
(55) Kawai, T.; Hamada, K.; Kon-No, K. Formation of AOT
Reversed Micelles and W/O Microemulsions. Bull. Chem. Soc. Jpn.
1992, 65, 2715−2719.
(56) Nave, S.; Eastoe, J.; Heenan, R. K.; Steytler, D.; Grillo, I. What
Is So Special About Aerosol-OT? Part III − Glutaconate versus
Sulfosuccinate Headgroups and Oil-Water Interfacial Tensions.
Langmuir 2002, 18, 1505−1510.
(57) Kazarian, S. G.; Vincent, M. F.; Bright, F. V.; Liotta, C. L.;
Eckert, C. A. Specific Intermolecular Interaction of Carbon Dioxide
with Polymers. J. Am. Chem. Soc. 1996, 118, 1729−1736.
(58) Zhou, G.-W.; Li, G.-Z.; Chen, W.-J. Fourier Transform Infrared
Investigation on Water States and the Conformations of Aerosol-OT
in Reverse Microemulsions. Langmuir 2002, 18, 4566−4571.
(59) Lee, C. T., Jr.; Bhargava, P.; Johnston, K. P. Percolation in
Concentrated Water-in-Carbon Dioxide Microemulsions. J. Phys.
Chem. B 2000, 104, 4448−4456.
(60) Holderer, O.; Klostermann, M.; Monkenbusch, M.; Schweins,
R.; Lindner, P.; Strey, R.; Richter, D.; Sottmann, T. Soft Fluctuating
Surfactant Membranes in Supercritical CO2-Microemulsions. Phys.
Chem. Chem. Phys. 2011, 13, 3022−3025.
K
dx.doi.org/10.1021/la400376g | Langmuir XXXX, XXX, XXX−XXX