Article history:Received 15 February 2012Accepted 22 June 2012Available online 4 July 2012
Keywords:Supercritical CO2
MicroemulsionsSurfactants
Supercritical CO2 is considered as a promising alternative for volatile organic solvents currently used in cer-tain industrial processes and products, however, the poor solubilizing power of CO2 towards polar substancesremains a significant barrier to applications. Employing effective surfactants which generate stable disper-sions and water/CO2 microemulsions is accepted as one way to improve the physico-chemical propertiesof CO2. This article reviews recent studies on microemulsions in liquid CO2, as well as the development ofCO2-philic surfactants.
Volatile organic compounds (VOCs) are commonly used as solventsfor chemical processing, and disposal of these environmentally hazard-ous compounds is a necessary but costly task. Although carbon dioxideis an abundant gas in the atmosphere, under accessible conditions(Pc=72.8 bar and Tc=31.1 °C), gaseous CO2 can be turned into a su-percritical fluid (scCO2). As such scCO2 has numerous advantages as asolvent, such as non-flammability, being also cheap, non-hazardousand non-toxic. Most importantly, the fluid properties of scCO2 such asdensity, and hence solvent quality, are tunable under temperature andpressure control, which further enhances the ease of solvent removalthrough rapid evaporation. Furthermore, it could be considered as oneof a number of potential solutions to the greenhouse effect if industrialand sequestration applications are made a reality for CO2 as a solvent.The dense forms (liquid, supercritical fluid) of this greenhouse gascould make major contributions towards improvements in enhancedoil recovery, chemical synthesis, food processing, etc.
However, due to the low dielectric constant (~1.33 at 200 bar,350 K) [1,2], scCO2 is generally a very poor solvent, especially forpolar and high molecular weight solutes, such limitations obviouslyrule out many potential applications. Efforts have been made to im-prove the solvent quality of scCO2 at accessible pressures, and fornearly two decades formation of water-in-CO2 (w/c) microemulsionshas been widely accepted as a promising approach, ever since fluoro-carbon amphiphiles were first shown to be effective surfactants forstabilizing water droplets in scCO2 [1,2].
Microemulsions are clear, thermodynamically stable liquidmixturesof two immiscible fluids, stabilized by surfactants. With correct selec-tion of surfactants, micelles or reversed micelles are formed with thedispersed water encapsulated in the micellar cores to form dispersed
rights reserved.
nanodroplets [3]. This article is focused on the recent development ofsurfactants for w/c microemulsions, which may potentially enhancethe capability of scCO2 to dissolve/disperse polar molecules. Studieson other systems such as scCO2 solutions of polymers, w/c mini-emulsions or macroemulsions are beyond the scope of this review.
2. Fluorinated CO2-philic surfactants
Fluorinated surfactants are the earliest systems being reported as ac-tive in scCO2. The field was pioneered by Beckman et al. [1], who firstlysuggested that fluorocarbons are comparable to CO2 and therefore afluorinated analogue of AOT surfactant (Table 1a) should display CO2
solubility/activity. Based on these observations, Johnston and O'Rear'sgroup synthesized a hybrid surfactant [2], (C7F15)(C7H15)CHSO4
−Na+,or F7H7 (Table 1b), and stable w/c microemulsions were reported forthe first time at 30 °C and 260 bar, later confirmed by small-angle neu-tron scattering (SANS) experiments [4].
Fluorinated sulfosuccinate surfactants have also been investigatedintensively [5,6], it has been shown that the chain length and fluori-nation at the terminal carbon atoms are key factors for the surfactantsto effectively stabilize w/c microemulsions. Among the compoundsinvestigated so far, the di-fluorocarbon chain di-CF4 (Table 1c) ap-pears to have the optimum molecular structure. A systematic studyof different molecular structures has been performed recently byMohamed et al. [7], which showed that di-CF2 is the CO2 active sur-factant containing the minimum amount of fluorine.
In addition, phosphate surfactants [8,9] and surfactants with fluori-nated co-polymer tails [10–15]were also reported as effective stabilizersfor w/c microemulsions. Since Eastoe et al. [16,17] have extensivelyreviewed that work, as well as the early development of CO2 active sur-factants, the reader is referred elsewhere for details.
Sagisaka et al. synthesized a different analogue of fluorinated AOT,mFS(EO)n (Table 1e) [18], and it has been shown that, subtle alterationsof the oxyalkylene unit change the surfactant behavior dramatically: the
Table 1Surfactants and CO2 philic analogues. The hydrocarbon chain di-C5SS has the sametemplate structure as diCFn, but when n=0 and the terminal groups are CH3–.
1a AOT
1b FmHn
1c diCFn
1d di-HCFn
1e mFS(EO)n
1f mFG(EO)n
1g Hybrid CF2/AOT4
1h diH8
1i diF8
267J. Eastoe et al. / Current Opinion in Colloid & Interface Science 17 (2012) 266–273
cloud pressure (or phase transition pressure, Ptrans) has been examinedat different water contents for two of the mFS(EO)n surfactants (m=8 and n=2,4) which showed superior stabilization of the w/c micro-emulsions compared to other di-HCFn surfactants (n=4, 6, 8, Table 1d)which were examined for comparison [18,19]. The molecular structureof 8FS(EO)2 has been suggested as optimum, since it can effectively sta-bilize w/c microemulsions up to water contents Wo
c=32, (Woc is the
corrected water–surfactant molar ratio Wco ¼ water½ �tot− water½ �CO2
surfactant½ � ), where
[water]tot is the total addedwater and [water]CO2 is the residual amountpartitioning into the CO2 continuum. On further addition of water to thesystem Wo
c>32, however, a liquid crystal (LC)-like precipitate was
observed. Interestingly, 8FS(EO)2 itself is insoluble in pure scCO2 evenat the highest temperature and pressure used in the study [19], but inthe presence of a small amount ofwater, themixture formed a transpar-ent single phase above the cloud pressure Ptrans. On the other hand, thesame amount of 8FS(EO)4 dissolved easily in scCO2 without the aid ofwater. This difference clearly arises from the addition of oxyethyleneunits, which are known to promote the intermolecular self-associationof surfactant molecules to form small assemblies in scCO2 due to stron-ger interactionswith the oxygen lone pairs [2,20,21]. The group has alsoevaluated the effect of oxyethylene and fluorocarbon units withnFS(EO)m surfactants [22]. The phase behaviour of water–scCO2 mix-tures has been examined at various water contents, pressures and tem-peratureswith six different surfactants: 6FS(EO)2, 8FS(EO)2, 10FS(EO)2,4FS(EO)4, 6FS(EO)4 and 8FS(EO)4; a phase diagramwas then generatedfor each surfactant/water/scCO2 mixture for comparison. The com-pound 8FS(EO)2 again appears to have an optimum structure alongwith the analogue 6FS(EO)4 which showed a very similar efficiency(W0
c=30 when CO2 density=0.75 g ml−1).In later studies [23,24], positive synergism has been observed by
mixing 8FS(EO)2 with a hybrid F-carbon+H-carbon surfactantFC6-HCn. Using FT-IR the swelling of single phase Winsor-VIw/scCO2 microemulsions droplets was observed with increasingwater content: when W0 exceeded 24, the system transformed intoa Winsor-II w/scCO2 microemulsion (w/scCO2 microemulsion inequilibrium with separated excess water). Interfacial tension experi-ments showed a much larger interfacial area per molecule for themixed surfactant system, interpreted in terms of a lower molecularpacking after mixing, perhaps being important for stabilizing a micro-emulsion over the liquid crystal (LC) (see Fig. 1).
Recently, Sagisaka's made further structural modifications with8FS(EO)2 by introducing one extra methylene unit on the headgroupspacer region [25] (structure 1f by contrast with 1e in Table 1). As con-firmed by UV–vis spectroscopy with a probe dye (Fig. 2), the new sur-factant, 8FG(EO)2 dramatically enhanced the maximum watercontent, being now up to W0=60 which is the highest value reportedfor any w/c system so far. The microemulsion nanodroplets were alsoexamined by using High-pressure Small-angle Neutron Scattering(HP-SANS) which showed that nanodroplets are present in the singlephase systems, and the droplet radius increases linearly with watercontent for W0b60. It is surprising to find that with only this subtlemodification on 8FS(EO)2, the addition of one –CH2– group, the waterstabilization capacity of the resulting surfactant is nearly doubled. Al-though a robust explanation of this remarkable result is still lacking,molecular packing at the interface should again play an important roleto account for the high microemulsifying power of 8FG(EO)2 in scCO2.
3. The concept of fractional free volume (FFV)
Stone et al. have provided a detailed study on the w/c interface byusing molecular simulations on two types of surfactants [26]: DiH8and its fluorinated analogue DiF8 which is CO2 active (see compounds1 h and 1i in Table 1). The results suggested that in the simulatedsurfactant/water/scCO2 systems, both CO2 and water molecules pen-etrate the hydrocarbon surfactant layer to a greater extent than thefluorocarbon surfactant, which results in unfavorable interactions,thus destabilizing the microemulsion. This finding was explored inmore detail [27], in terms of the concept of fractional free volume(FFV equation 1) which was proposed as an index of the free spaceavailable in the interfacial film.
FFV ¼ 1− Vt
lAh
where Vt is the physical volume of the tail groups, l is the thickness ofthe interface, and Ah is the interfacial area per surfactant moleculewhich can be calculated from a swelling law [28,29], or tensiometric
Fig. 1. Schematic molecular structures for nFS(EO)2 and FC6-HC4 and packing at the water/scCO2 interface.(Reprinted with permission from Ref. [24]. Copyright American Chemical Society).
268 J. Eastoe et al. / Current Opinion in Colloid & Interface Science 17 (2012) 266–273
methods [5]. The lower the FFV the more bulky is the surfactant, andthe more efficiently the molecular fragments fill space in the inter-face. This goes hand in handwith a reduction in water and CO2 molec-ular contact, which is expected to promote stability of the w/cinterface necessary for microemulsion formation.
This FFV concept has been tested for a wide range of surfactants, andthe correlation between lower FFV with greater w/c microemulsion sta-bility is consistent with numerous experimental observations [24,27,30].In general,fluorocarbon chains are suggested as beingmuch bulkier than
Fig. 2. UV–visible absorption spectra of the probe dye methyl orange (MO at 0.1 wt.%) in 8Fatures and CO2 densities were: (a) 45 °C, 0.92 g cm−3 and (b) 75 °C, 0.81 g cm−3. The molaW0 value.(Reprinted with permission from Ref. [25] Copyright American Chemical Society)
the hydrocarbon counterparts (similar l and Ah butmuch larger Vt), lead-ing to lower FFV, and enhanced w/c microemulsion stability.
4. Low fluorine content CO2‐active surfactants
There is no doubt that fluorocarbon surfactants are very effective forstabilizing w/c microemulsions, but they are limited for applicationsbeing both expensive and environmentally hazardous [24,31–35]. Onthe other hand, as shownbyConsani and Smith [36] in their preliminary
G(EO)2/W/CO2 mixtures for different W0 values at 350 bar. The experimental temper-r ratio of 8FG(EO)2 to CO2 was fixed at 8×10−4. Values for each spectrum indicate the
269J. Eastoe et al. / Current Opinion in Colloid & Interface Science 17 (2012) 266–273
studies, most conventional commercially available surfactants are in-compatible with CO2, thus there is clearly a need to develop efficientlow-fluorine, or non-fluorous surfactants which would be more com-mercially viable, and environmentally responsible.
Early studies have compared the CO2 activities of numerous fluo-rocarbon surfactants with fluorination levels [5,6], and it was con-cluded that an effective surfactant for scCO2 should bear a highlyfluorinated structure, especially at the chain tips. However, the ques-tion ‘what is the minimum fluorine atom required for a fluorocarbonsurfactant to stabilize w/c microemulsions’ has not been resolved,until recently. Mohamed et al. [7] compared a series of di-CFn surfac-tants and revealed that di-CF2 has the minimum fluorine content,while also retaining sufficient CO2 activity, therefore, this structureand fluorination level could be considered as a baseline when design-ing low fluorine content surfactants.
In addition, the study proposed a predictive model for CO2-philicsurfactant design based on a direct link between the packing of sur-factants at air–water interface and the performance of the compoundfor stabilization of w/c microemulsions. This has an important advan-tage since normal laboratory-based surface tension measurements,merely on aqueous solutions, can be applied as a convenient and rea-sonably accurate method to test the CO2-philicity of newly designedcompounds before embarking on involved synthesis and high pres-sure studies.
The packing density at the interface is characterized by using rel-ative surfactant coverage index:
Φsurf ¼Vcal
Vmeas
where Vcal is the total physical volume of surfactant molecular frag-ments from literature, and Vmeas is that measured volume occupiedby a molecule at the reference air–water interface calculated usingthe limiting headgroup area at the aqueous phase cmc, Acmc, and
Fig. 3. Correlation between relative surfactant coverage, Φsurf, limiting aqueous surface tenCO2-philic surfactants with fully fluorinated (red), and partially fluorinated or hydrocarwhich are not soluble in CO2 (non-CO2-philic), but are included for comparison purposes.going hand in hand with both lower γcmc, and Ptrans. For example, hybrid CF2/AOT4 Φsurf=10 at 25 °C) and γcmc=23.5 mN m−1. The line is a guide to the eye. The error bars represe
interfacial thickness τ: Vmeas=Acmc×τ. By comparing a series of sur-factants with different fluorination levels, a linear correlation wasobtained between the surface coverage index, not only with the lim-iting surface tension at cmc (γcmc), but also the cloud point pressurePtrans for the associated w/c microemulsion (Fig. 3). In later studies[37], Mohamed et al. further examined this approach with F7H7(1b) and a CF2/AOT4 hybrid surfactant (structure 1g in Table 1),both surfactants comply with a well established trend found forother amphiphiles. In Fig. 3 surface tension values for hydrocarbonsurfactants di-C5SS (1c, see caption), AOT (1a) and its highlybranched analogue AOT4 (structure 2f, see later in Table 2) are alsoincluded for comparison, although they themselves do not stabilizew/c phases.
5. Fluorine-free and hydrocarbon-only CO2-active surfactants
Han et al. [38–40] reported studies on Dynol-604, a non-ionicacetylenic glycol-based surfactant, which appears to be soluble insc-CO2, and also stabilizes w/c microemulsions to as much as0.6 wt.% water content. Three other commercial surfactants, Ls-35,Ls 45 and Ls-54 (2a in Table 2) have also been studied by the samegroup, it appears that all three molecules are soluble in sc-CO2 (atca. 4 wt.% at 35–47 °C and 190–220 bar), and solubilization capacityfor water in CO2 is apparently enhanced, however, no SANS or SAXSexperiments have as yet been published to confirm these phaseobservations.
Work by da Rocha et al. [41] has shown promising results fortrisiloxanes (2b) being stabilizers for w/c or c/w emulsions, butmicroemulsions were not stabilized.
Based on an extensive systematic study of different surfactanttypes, Pitt et al. had pointed out that t-butyl chain tips promotethe lowest aqueous surface tensions for hydrocarbon surfactants[42]. Therefore, introducing highly branched chain tips could be
sion γcmc and cloud point pressure, Ptrans * data from Ref. [5]. Circled points representbon (black) surfactants. The un-circled points are for hydrocarbon-only surfactants,The correlation is a linear relationship linking Φsurf, γcmc, and Ptrans: with higher Φsurf
0.70 and this compound displays Ptrans=289 bar ([Surfactant]=0.05 mol dm−3 w=nt the uncertainty of Ptrans±40 bar.
270 J. Eastoe et al. / Current Opinion in Colloid & Interface Science 17 (2012) 266–273
considered as a reasonable approach to design effective hydrocarbonsurfactants for CO2, since low surface energies will be beneficial.
Following this hypothesis, two t-butyl tipped analogues of AOT(AOT3 and AOT4) were synthesized and studied by Eastoe et al.,and low w/c interfacial tensions were obtained for both surfactantscompared to normal AOT. Furthermore, formation of reversed mi-celles was later confirmed by HP-SANS [43], although addition ofwater immediately resulted in phase separation. This study is abreak-through in design of hydrocarbon surfactants for CO2.
Further enhancements in CO2 compatibility have been made byintroducing a third highly methylated chain [44], to make a familyof TCS tri-chain surfactants. Such molecular structures are expectedto lower surface energy and enhance compatibility with CO2, alsoconsistent with the fractional free volume [27] and cohesive densityarguments [45] used to account for CO2-philicity. A series of TCS triplechain surfactants was then designed and tested in CO2, and TC14 (2gTable 2) owing to its highly branched chain tips, exhibited superiorsolubility and also the best performance in stabilizing w/c hydratedreversed micelles (however, it cannot be claimed that these low Wsystems were true microemulsions). Recently, HP-SANS experimentshave been performed which confirmed hydrated reverse micelles areformed by TC14 with W0=5 at 360 bar and 25 °C [46].
Air/water surface tensions have also been determined for thesehighly branched TCS analogues [47], from which, knowledge of sur-factant molecular packing could be obtained. Results showed thatTC14 has a considerably lower FFV than AOT4 which may be one
plausible reason for why TC14 forms stable hydrated reversed mi-celles in CO2. It should be noted that the FFV value of TC14 is evenlower than that of the most efficient fluorinated surfactant DiCF4(FFV=0.42). This may indicated some limitations of considering FFVparameter as the only guideline for designing CO2-philic surfactants.
Johnston et al. have also demonstrated that a highly branchednon-ionic hydrocarbon surfactant, 5b-C12E8 (compound 2c) effectivelystabilizes w/c microemulsions at water contents up to 1.1 wt.%(W0~28) [48]. More recently, the behavior of another branched hydro-carbon nonionic surfactant 2EH–PO4.5–EO8 (compound 2d) has beenexamined in w/c microemulsions as well as in macroemulsions [49],results show it is CO2-philic owing to its highly branched structure. Nextsystematic investigations of the effects of branching using surface tensionat the air–water and high pressure interfacial tension CO2–waterinterface with a series of non-ionic hydrocarbon surfactants werepreformed [50]. It was revealed that the driving forces for adsorptionare different between air–water and CO2–water interfaces, in thelater case both tail solvation and hydrophobicity, which increasewith branching, become increasingly significant promoting surfac-tant adsorption.
6. Carbonyl CO2-active surfactants
It has been shown that CO2 exhibits strong Lewis acid–Lewis base in-teractions with carbonyl groups on solute molecules [47], and based onthat a series of acetylated sugars demonstrated good solubility in scCO2
[51–53]. Oxygenated hydrocarbon surfactants have been considered asanother promising approach since the CO2-philic acetylated sugars couldbe easilymodified andused as surfactants for eitherwater-in-CO2micro-emulsions or even oil-in-CO2 microemulsions [51].
Based on Wallen and Raveendran's experiments and simulations[51], Fan et al. [54] reported a series of AOT analogueswith different ox-ygenated hydrocarbons, and oligo(vinyl acetate)-functionalized surfac-tants showing promising results: water loading values of 50 wereattained with the twin-tailed sodium bis(vinyl acetate)8 sulfosuccinateAOT analogue (AO-VAc, compound 3a Table 3) at surfactant concentra-tions as high as 1 wt.%, while the solubility in CO2 is about 3 wt.%. Inter-estingly, it has been suggested that the presence of both hydrophilic andhydrophobic (methylene backbone) sections on the oligo(vinyl acetate)tails provided the correct balance of forces necessary to drive water outof the bulk phase into reversedmicelles, which further lead to stabiliza-tion of w/c microemulsion interfaces.
The stabilization of w/c microemulsions by oxygenated chains wasalso examined by Eastoe et al. [55], who synthesized a ketone ana-logue of AOT surfactant, AOK, bearing one carbonyl group in eachtail and t-butyl at the chain tips (Table 3). The w/c microemulsionsstabilized by both AOK and AO-VAc were examined by SANS andassigned to a spherical core-shell structure (Fig. 4). Best fit parame-ters to the structure such as core radii, shell thicknesses and polydis-persities for both CO2-philes were consistent with core-shellnanodroplet structures.
Fig. 5. Self-assembly of water in scCO2 inverse microemulsions stabilized byPFPECOO-NH4
+. (i) High pressure 129Xe NMR spectrum (32 scans) after 105 min ofself assembly at 30 MPa. (ii) Spectrum (256 scans) of Xe gas at 2 MPa. (iii) Contourplot of 129Xe NMR spectra as a function of time. (iv) A schematic showing theself-assembly process. In which large water droplets (a) reduce in size as smaller drop-lets (b) till equilibrium has been reached (c).(Reprinted from Ref. [62]).
Fig. 4. (a) SANS data fitted with spherical core-shell form factor (lines) forsurfactant-stabilized D2O-in-CO2 microemulsion droplets as a function of Wo
c. Surfac-tant concentration cAOK=50 mmol dm−3 (~2.4 wt.%) and cAO-Vac=6.2 mmol dm−3
(~1 wt.%). Experiment was performed under 500 bar and 45 °C for AOK but 25 °C forAO-VAc. Example error bars are shown for one AOK sample. (b) Schematic scatteringlength density (sld or ρ×1010 cm−2) profile fitted to SANS data fromsurfactant-stabilized D2O-in-CO2 microemulsion droplets. The values fixed in theform factor analyses for the external CO2 phase, AOK surfactant shell, and D2O corewere 2.4×1010, 0.3×1010, and 6.4×1010 cm−2 respectively.(Both figures are reprinted with permission from Ref. [55]. Copyright Wiley-VCH).
271J. Eastoe et al. / Current Opinion in Colloid & Interface Science 17 (2012) 266–273
Compared with other carbonyl surfactants which are inactive inCO2, the enhanced activity of AOK is largely owing to the nature ofthe highly branched chain tips, whereas for AO-VAc it is thought tobe due to the presence of more extensive CO2-philic groups. Theseobservations have pointed to new directions for designing stabilizersfor water–CO2 interfaces using unusual chain structures which areuncommon in normal water and water/oil systems.
7. Theoretical studies
Though experimental techniques such as IR and neutron scatter-ing have been used to characterize the stability of w/c micro-emulsions, there is still a lack of fundamental understanding aboutw/c interfaces due to the practical difficulties of experimental studiesat high pressures. However, deeper insight is essential for the devel-opment of more effective surfactants.
The theoretical studies on scCO2 using Molecular Dynamic (MD)simulation were pioneered by Rossky et al. [56], who initially investi-gated structural properties of the binary c/w fluid–fluid interface.Later, the behavior of amphiphiles at the interface was also examinedby MD with various surfactants [27,28,57]. Based on these results the
important concept of FFV was proposed (see previous section) whichhas been widely used in later studies as a surfactant design principle.In their recent studies, Rossky's group performed MD simulations onfluorocarbons as capping ligands used for stabilizing nanoparticle dis-persions in CO2 [58]. The results highlighted that the strong electrondipole of the C\F bond does not simply make the ligand CO2-philic,but also gives rise to a unique property of fluorocarbon chainswhich always present an adverse dipole to themselves so preventingparticle instability. In addition, the packing density of the perfluoro-carbon chains gives relatively weak fluorocarbon–ligand–passivatedsurface interactions, also enhancing the effectiveness as CO2-philic li-gands. These MD simulations have also been performed onn-dodecanethiol as ligand in supercritical ethane and CO2 [59]. Thebehavior of the ligands in both media has been compared suggestingthat the interaction between CO2 molecules is much more significantin ethane due to the strong quadrupole, which is not adequately com-pensated by the hydrocarbon ligands. As a result, the alkanthiol li-gand is less effective in supercritical CO2 compared with ahydrocarbon solvent under similar conditions.
Following Rossky's work, MD simulation has been extensively ap-plied to theoretical studies in behavior of amphiphiles in supercriticalCO2. A CO2 induced microstructural transition between a lamellar andmicroemulsion phase has been simulated by Zhuo et al. [60] for six dif-ferent amphiphiles, including three of the recently synthesized hydro-carbon surfactants AOT4, AOK, and AO-VAc. The self-assembly processof AOK surfactants in the w/c system has also been studied from a mo-lecular viewpoint by Wu et al. [61]. Aqueous cores have been observedwithin the reversed micelles, consistent with SANS experiments [55],and the estimated micellar size is very close to the reported value. Thegroup further explored the self‐assembly structure of AOK surfactantin CO2 without the presence of water; interestingly, irregular andnon-spherical reversed micelles are predicted with a loose inner coreregion allowing penetration of the solvent molecules.
Recently, Blakey et al. reported the first real time-monitoring tech-nique for surfactant self-assembly in CO2 by using high pressure 129Xe
272 J. Eastoe et al. / Current Opinion in Colloid & Interface Science 17 (2012) 266–273
NMR [62]. Owing to the fact that Xe can be easily introduced to CO2 bysimply pressuring the gas, and its chemical shift is extremely sensitiveto the local chemical environment, Xe can act as an environmentalprobe. An example experiment was set up with a PFPECOONH4
(which is an amphiphile containing a perfluoropolyether chain and anamide headgroup) solution in a pressure cell, Xe and CO2 were then in-troduced with pressurization up to 30 MPa. A single peak at 149.7 ppmwas immediately observed after pressurization with CO2 with respectto the signal from free xenon, two peaks corresponding to Xe atomsnext to large and small water droplet interface started to appear after45 and 120 min. The intensity increased andfinallymerged into a singlepeak at 97 ppm after 300 min,which indicates the process of spontane-ous formation of w/c reversed micelles from droplets of water (Fig. 5).
8. Applications
With compatible surfactants being developed, the applications ofCO2 as a green, tunable, and easy to process solvent have received in-tensive interest. Su et al. [63] reported that oxymatrine, a potentiallyimportant pharmaceutical, could be solubilized in w/c micro-emulsions with assistance of surfactant PFPENH4. A series of workhas been done by Jun et al. [64–67], which demonstrated applicationsof different analogues of PFPE surfactants in dispersing dyes in scCO2.
It has been recently discovered that by substituting the Na+ coun-terion with Co2+ or Ni2+ for surfactant di-HCF4 [68], rod-like micro-emulsions are formed instead of spherical droplets normally foundwith the unmodified Na+ surfactant. HP-SANS experiments andhigh pressure viscosity tests on ion exchanged di-HCF4 areself-consistent, showing that with 10 wt.% surfactant the viscosity ofCO2 can be increased by 90%. Although such increase is still toosmall to be suitable for direct applications in enhanced oil recovery,however, the work has successfully demonstrated that the viscosityof CO2 can be tuned using surfactant self-assembly.
Later, the hydrocarbon surfactant TC14 has also been modifiedthrough ion exchange [69], and tested as a solvent thickening agent inboth CO2 and CO2/C6D12 mixtures. Although such metallosurfactantshave been shown to be soluble in pure CO2 at high concentrations, norod-likemicelles were formed on addition of water. Hence, the challengeof developing a hydrocarbon-only viscositymodifier for CO2 still remains.
9. Synthesis of nanoparticles in w/c microemulsions
Water in oil (w/o) microemulsions have been widely applied fornanoparticle synthesis, as a good size control of nanoparticles canbe obtained with reversed micelles, which act as nanoreactors [70].As a continuous phase solvent scCO2 has various advantages over or-dinary organic solvents as introduced earlier, in addition, direct depo-sition of nanoparticle products could be achieved with scCO2 bysimply adjusting the temperature or pressure, owing to its supercrit-ical nature, which leads to advanced processing of nanoparticles notpossible with conventional solvents. Various inorganic nanoparticleshave been successfully synthesized in w/c microemulsions, includingmetals Cu [71], Pd [72] and silver [73–78]; metal oxides and TiO2 [79],silver halides [80–82], hydrous zirconia [83], metal sulfides of Ag2S[84], CdS and ZnS [85]. In addition, scCO2 has also been reported asan effective anti-solvent used to recover a variety of nanoparticlesfrom reversed micelles, all these studies have been well reviewedby Zhang et al. [86].
However, all studies mentioned above apply fluorocarbon surfac-tants to stabilize w/c microemulsions, which are both expensiveto source and environmentally persistent. Recently, Hollamby et al.[87] reported studies using the hydrocarbon surfactant TC14 as ananoparticle stabilizer in a CO2-rich solvent. The TC14-stabilized CeO2
nanoparticles were prepared under ambient conditions in water-in-heptane microemulsions, extracted and then re-dispersed into aheptane–CO2 mixed solvent. Solubility tests by UV–vis show that the
nanoparticles are well dispersed in CO2 and remain stabilized by thesurface-tethered TC14. This is the first report showing that hydrocarbonsurfactants could be considered as promising candidates for disper-sion of nanoparticles in CO2 solvents, and will hopefully open thedoor to more applications of this kind with hydrocarbon surfactants.
10. CO2 in water microemulsions
While w/c microemulsions have received intensive interest, littlehas been done on aqueous systems containing CO2 swollen micelles(i.e. c/w microemulsions). Johnson et al. [88] reported the first obser-vations of a water-rich system containing dense CO2. As shown byDynamic Light Scattering experiments stable CO2 in water (c/w)microemulsions are formed by swelling of cylindrical, rod-like mi-celles of PFPE-based surfactants. Later, Schwan et al. [89] reportedformation of stable c/w microemulsions with equal volumes ofwater and CO2 using a commercial non-ionic, hydrocarbon surfactantLutensol® XL 70. The study also showed that with increasing amountof CO2 the microemulsion system exhibits a phase inversion from aCO2-in-water to water-in-CO2, via a balanced scCO2 microemulsion.
O'Callaghan and co-workers [90] applied SANS to study the swell-ing of CTAB and Pluronic F127 micelles by CO2 at different pressuresand temperatures. Interesting results were obtained for the cationicsurfactant CTAB: by varying the pressure, not only the size but alsothe shape of micelles was seen to change, which is very differentfrom the swelling behavior of the non-ionic surfactant F127. Recent-ly, a systematic study on the microstructure of efficient water-rich(c/w)-microemulsions has been performed by Klostermann et al.[91]. Using HP-SANS-measurements it has been found that themean droplet radius increases almost linearly with increasing frac-tion of CO2 content and temperature, showing that the temperaturedependence of the mean curvature of the amphiphilic film found innormal oil containing microemulsions also holds for water richscSC microemulsions.
11. Ionic liquid in scCO2 microemulsions
Room temperature ionic liquids (ILs) are organic salts, composedof organic cations and organic or inorganic anions, and are liquids atroom temperature. Such compounds have been considered as prom-ising green solvents due to their unique properties [92], such aschemical stability, non-flammability and low volatility. CombiningILs with scCO2 could potentially introduce advantages of both sol-vents to the mixture, however, it had been reported that the solubilityof most ILs in scCO2 is in fact quite low [93].
Work by Han et al. [94] has shown, for the first time, the existenceof IL domains in a continuous CO2 phase from spectroscopic experi-ments on mixtures of the ionic liquid 1,1,3,3-tetramethylguanidiumacetate ([TMG][Ac]), supercritical CO2, and N-ethyl perfluorooctylsulfonamide (N-EtFOSA) as surfactant. By comparing the absorbanceof the tracer probe dye MO (methyl orange), it was concluded thatIL microemulsion droplets are entrapped within N-EtFOSA reversedmicelles. In a later study [95], such system was simulated by Senapatiet al. using MD, and nanometer sized, ellipsoidal ionic liquid dropletsstabilized in the polar cores of surfactant aggregates were predicted,consistent with SANS experiments on IL-in-oil microemulsion sys-tems [96]. Very recently [97], Zhang and co-workers have demon-strated that a CO2 in IL microemulsion could be formed with thesame mixture (TMGA, scCO2 and N-EtFOSA).
12. Conclusions
Studies on CO2-philic surfactants have been conducted to obtainstable w/c microemulsions. After significant experimental and theo-retical effort, it has been recognized that fluorocarbon chain surfac-tants might be the most effective compounds for stabilizing water
273J. Eastoe et al. / Current Opinion in Colloid & Interface Science 17 (2012) 266–273
nanodroplets and perhaps other polar species in scCO2. These studieshighlight the significance of three major factors for the design of effi-cient CO2-philic surfactants:
1. The fractional free volume FFV of fluorocarbon surfactants is smallenough to minimize penetration of CO2 and water into the interfa-cial film, thereby lowering the interfacial energy sufficiently topermit microemulsion formation.
2. The interaction between fluorocarbon chains and CO2 must be suf-ficiently strong.
3. Fluorocarbon chains always present an adverse dipole and thusprevent self-coagulation.
These rules can now be used as guidelines for designing cheaperand environmental responsible, effective fluorine-free surfactantsfor scientific, technological and industrial applications.
John Wiley & sons, Ltd; 2010.[4] Eastoe J, Bayazit Z, Martel S, Steytler DC, Heenan RK. Langmuir 1996;12:1423.[5] Eastoe J, Downer A, Paul A, Steytler DC, Rumsey E, Penfold J, et al. Phys Chem
Chem Phys 2000;2:5235.[6] Eastoe J, Paul A, Downer A, Steytler DC, Rumsey E. Langmuir 2002;18:3014.[7] Mohamed A, Sagisaka S, Guittard F, Cummings S, Paul A, Rogers, et al. Langmuir
2011;27:10562.[8] Steytler DC, Rumsey E, Thorpe M, Eastoe J, Paul A, Heenan RK. Langmuir 2001;17:
7948.[9] Keiper JS, Simhan RJ, DeSimone M, Wignall GD, Melinchenko YB, Frielinghaus H.
J Am Chem Soc 1834;2002:124.[10] Heitz MP, Carlier C, deGrazia J, Harrison KL, Johnston KP, Randolph TW. J Phys
Chem B 1997;101:6707.[11] Clarke MJ, Harrison KL, Johnston KP, Howdle SM. J Am Chem Soc 1997;119:6399.[12] Fremgen DE, Smotkin ES, Gerald III RE, Klingler RJ, Rathke JW. J Supercrit Fluids
2001;19:287.[13] Zielinski RG, Kline SR, Kaler EW, Rosov N. Langmuir 1997;13:3934.[14] Lee Jr CT, Bhargava P, Johnston KP. J Phys Chem B 2000;104:4448.[15] Lee Jr CT, Johnston KP, Dai HJ, Cochran HD, Melnichenko YB, Wignall GD. J Phys
Chem B 2001;105:3540.[16] Eastoe J, Dupont A, Steytler DC. Curr Opin Colloid Interface Sci 2003;8:267.[17] Eastoe J, Gold S, Steytler DC. Langmuir 2006;22:9832.[18] Sagisaka M, Yoda S, Takebayashi Y, Otake K, Kitiyanan B, Kondo Y, et al. Langmuir
2003;19:220.[19] Sagisaka M, Yoda S, Takebayashi Y, Otake K, Kondo Y, Yoshino N, et al. Langmuir
2003;19:8161.[20] Dong X, Erkey C, Dai HJ, Li HC, Cochran HD, Lin JS. Ind Eng Chem Res 2002;41:
et al. Langmuir 2002;18:7371.[22] Sagisaka M, Koike D, Yoda S, Takebayashi Y, Furuya T, Yoshizawa A, et al.
Langmuir 2007;23:8784.[23] Sagisaka M, Fujii T, Koike D, Yoda S, Takebayashi Y, Furuya T, et al. Langmuir
2007;23:2369.[24] Sagisaka M, Fujii T, Mashimo Y, Yoda S, Takebayashi, Furuya Y, et al. Langmuir
2008;24:10116.[25] Sagisaka M, Iwasa S, Hasegawa S, Yoshizawa A, Mohamed A, Cummings S, et al.
Langmuir 2011;27:5772.[26] Stone MT, da Rocha SRP, Rossky PJ, Johnston KP. J Phys Chem B 2003;107:10185.[27] Stone MT, Smith PG, da Rocha SRP, Rossky PJ, Johnston KP. J Phys Chem B
2004;108:1962.[28] Nave S, Eastoe J, Penfold J. Langmuir 2000;16:8733.[29] Nave S, Eastoe J, Heenan RK, Steytler DC, Grillo I. Langmuir 2000;16:8741.[30] Mohamed A, Trickett K, Chin SY, Cummings S, Sagisaka M, Hudson L, et al.
Langmuir 2010;26:13861.[31] Lau C, Butenhoff JL, Rogers JM. Toxicol Appl Pharmacol 2004;198:231.[32] Butenhoff J, Costa G, Elcombe C, Farrar D, Hansen K, Iwai H, et al. Toxicol Sci
2002;69:244.[33] Houde M, Martin JW, Letcher RJ, Solomon KR, Muir DCG. Environ Sci Technol
2006;40:3463.[34] Falandysz J, Taniyasu S, Gulkowska A, Yamashita N, Schulte-Oehlmann U. Environ
Sci Technol 2006;40:748.[35] Olsen GW, Church TR, Miller JP, Burris JM, Hansen KJ, Lundberg JK, et al. Environ
Heal Perspect 1892;2003:111.
[36] Consani KA, Smith RD. J Supercrit Fluids 1990;3:51.[37] Mohamed A, Sagisaka M, Hollamby M, Rogers SE, Heenan RK, Dyer D, et al. Lang-
muir 2012;28:6299.[38] Liu J, Han B, Zhang J, Li G, Zhang X, Wang J, et al. Chem Eur J 2002;8:1356.[39] Liu J, Han B, Zhang J, Li G, Zhang X, He J, et al. Langmuir 2001;17:8040.[40] Liu J, Han B, Zhang J, Li G, Zhang X, Wang J, et al. Langmuir 2002;18:3086.[41] da Rocha SRP, Dickson J, Cho D, Rossky JP, Johnston KP. Langmuir 2003;19:3114.[42] Pitt AR, Morley SD, Burbidge NJ, Quickenden EL. Colloids Surf A 1996;114:321.[43] Eastoe J, Paul A, Nave S, Steytler DC, Robinson BH, Rumsey E, et al. J Am Chem Soc
2010;364:455.[51] Raveendran P, Wallen SL. J Am Chem Soc 2002;124:7274.[52] Potluri VK, Xu JH, Enick RM, Beckman E, Hamilton AD. Org Lett 2002;4:2333.[53] Ma SL, Wu YT, Enick RM, Wallen SL, Grant CS. J Phys Chem B 2010;114:3809.[54] Fan X, Ptluri VK, McLeod MC, Wang Y, Liu J, Enick RM, et al. J Am Chem Soc
2005;127:11754.[55] Eastoe J, Gold S, Rogers S, Wyatt P, Steytler DC, Gurgel A, et al. Angew Chem
2006;45:3675.[56] Rocha SRP, Johnston KP, Westacott RE, Rossky PJ. J Phys Chem B 2001;105:12092.[57] Rocha SRP, Johnston KP, Rossky PJ. J Phys Chem B 2002;106:13250.[58] Dalvi VH, Srinivasan V, Rossky PJ. J Phys Chem C 2010;114:15553.[59] Dalvi VH, Srinivasan V, Rossky PJ. J Phys Chem C 2010;114:15562.[60] Zhuo S, Huang Y, Xu Z. J Phys Chem B 2010;114:6344.[61] Wu B, Yang X, Xu Z. Colloids Surf A Physicochem Eng Aspects 2010;367:148.[62] Blakey I, Thurecht KJ, Wittaker AK. Chem Commun 2010;46:2850.[63] Chen W, Su BG, Xing HB, Ren QL. J Chem Eng Data 2008;53:1705.[64] Sawada K, Jun JH, Ueda M. Color Technol 2003;119:336.[65] Sawada K, Jun JH, Ueda M. Dyes Pigm 2004;61:197.[66] Sawada K, Jun JH, Ueda M. Dyes Pigm 2004;60:17.[67] Sawada K, Jun JH, Ueda M, Sugimoto M, Urakawa H. Color Technol 2005;121:315.[68] Trickett K, Xing D, Enick R, Eastoe J, Hollamby MJ, Mutch KJ, et al. Langmuir
2010;26:83.[69] Mohamed A, Tricett K, Chin SY, Cummings S, Sagisaka M, Hudson L, et al. Lang-
muir 2010;26:4732.[70] Eastoe J, Hollamby MJ, Hudson LK. Adv Colloid Interface Sci 2006;5:128.[71] Ohde H, Hunt F, Wai CM. Chem Mater 2001;13:4130.[72] Ohde H, Wai CM, Kim H, Kim J, Ohde M. J Am Chem Soc 2002;124:4540.[73] Fan X, McLeod MC, Enick RM, Roberts CB. Ind Eng Chem Res 2006;45:3343.[74] Kometani N, Toyoda Y, Asami K, Yonezawa Y. Chem Lett 2000;6:682.[75] Fernandez CA, Wai CM. Small 2006;2:1266.[76] Meziani MJ, Pathak P, Beacham F, Allard LF, Sun YP. J Supercrit Fluids 2005;34:91.[77] Ji M, Chen XY, Wai CM, Fulton JL. J Am Chem Soc 1999;121:2631.[78] Sun YP, Torngitjawat PA, Meziani MJ. Langmuir 2001;17:5707.[79] Lim KT, Hwang HS, Lee MS, Lee GD, Hong S, Johnston KP. Chem Commun 2002:
1528.[80] Liu JC, Raveendran P, Shervani Z, Ikushima Y, Hakuta Y. Chem Eur J 1854;2005:11.[81] Ohde H, Rodriguez JM, Ye XR, Wai CM. Chem Commun 2000;23:2353.[82] Ohde H, Rodriguez JM, Ye XR, Wai CM. Phys Chem Chem Phys 2000;2:2353.[83] Lee MH, Lin HY, Thomas JL. J Am Ceram Soc 2006;89:3624.[84] Liu JC, Raveendran P, Shervani Z, Ikushima Y. Chem Commun 2004;22:2582.[85] Ohde H, Ohde M, Bailey F, Kim H, Wai CM. Nano Lett 2002;2:721.[86] Zhang J, Han B. J Supercrit Fluids 2009;47:531.[87] Hollamby MJ, Trickett K, Vesperina A, Rivett C, Steytler DC, Schnepp Z, et al. Chem
Commun 2008:5628.[88] Lee CT, Ryoo W, Smith PG, Arellano J, Mitchell DR, Lagow RJ, et al. J Am Chem Soc
2003;125:3181.[89] Schwan M, Kramer LGA, Sottmann T, Strey R. Phys Chem Chem Phys 2010;12:
et al. Langmuir 2010;26:7725.[91] Klostermann M, Strey R, Sottmann T, Schwein R, Linder P, Holderer O, et al. Phys
Chem Chem Phys 2011;13:20289.[92] Rogers RD, Seddon KR. Science 2003;302:792.[93] Blanchard LA, Hancu D, Bechman EJ, Brennecke JF. Nature 1999;399:28.[94] Liu J, Cheng S, Zhang J, Feng X, Fu X, Han B. Angew Chem Int Ed 2007;46:3313.[95] Chandran A, Prakash K, Senapati S. J Am Chem Soc 2010;132:12511.[96] Eastoe J, Gold S, Rogers SE, Paul A, Welton T, Heenan RK, et al. J Am Chem Soc
2005;127:7302.[97] Zhang J, Han B, Li J, Zhao Y, Yang G. Angew Chem 2011;123:10085.
