rXXXX American Chemical Society A dx.doi.org/10.1021/la2002404 | Langmuir XXXX, XXX, 000–000

ARTICLE

pubs.acs.org/Langmuir

Self-Assembly, Foaming, and Emulsifying Properties of SodiumAlkyl Carboxylate/Guanidine Hydrochloride Aqueous MixturesAnne-Laure Fameau,†,‡ B�er�enice Houinsou-Houssou,† Jorge Luis Ventureira,§ Laurence Navailles,||

Fr�ed�eric Nallet,|| Bruno Novales,† and Jean-Paul Douliez*,†

†UR1268, INRA Nantes, Biopolym�eres Interactions Assemblages, rue de la G�eraudi�ere, 44316 Nantes, France‡Laboratoire L�eon Brillouin, CEA-CNRS, CEA Saclay, 91191 Gif sur Yvette Cedex, France§Centro de Investigation y Desarollo en Criotechnologia de Alimentos (CIDCA) (CONICET-CCT-La Plata-Facultas de CienciasExactas, UNLP), 47 y 116 (s/n), La Plata (1900), Buenos Aires, Argentina

)Centre de Recherche Paul Pascal, CNRS, avenue A. Schweitzer, 33600 PESSAC, France

bS Supporting Information

’ INTRODUCTION

The use of agricultural resources for industrial purposes willundoubtedly be one of the major challenges of the 21st century,both from the energy point of view (by contributing to thereplacement of fossil fuels) and with respect to nonenergy uses,resulting from the availability of organic “biosynthons” to thechemicals industry. Our work on dispersions of saturated fattyacids and hydroxylated derivatives forms part of these efforts inthat it seeks to demonstrate the potential contribution thesebiological compounds of plant origin couldmake as a new class ofgreen surface active agents.

However, soaps of long chain saturated fatty acids (carboxylicacids) are restricted as surfactants because of their low solubilityin water at room temperature. Indeed, whereas they can formmicelles at high temperature, they crystallize below a giventemperature, which is known as the Krafft point.1,2 For instance,sodium myristate crystallizes below 45 �C. An alternative, whichis commonly used to date, is to chemically modify the fatty acidsto increase their amphiphilicity and make them more solublein water.3 However, this requires additional organic chemicalsteps, which may be in some cases detrimental for industrial

applications. Another possible way to disperse fatty acids in wateris to mix them with alkylamine derivatives yielding catanionicsystems, which may self-assemble into original supramolecularstructures4,5 and also vesicles.6 Besides, self-assembly of fatty acidvesicles is of particular interest in the concept of the “lipid world”.7

Indeed, fatty acid vesicles are believed to be the first sacs (in theprebiotic conditions on the early Earth) that allowed the encapsu-lation of chemicals yielding the first cells.8

The last way to disperse fatty acids in water is to use solubleorganic counterions. First, tetrabutyl ammonium hydroxide(TBAOH) has been successfully used as a counterion (insteadof sodium), allowing the formation of stable micelles even forvery long chain carboxylic acids.9,10 In the same way, choline,which is a biological natural chemical, has been employed fordispersing fatty acids.11,12 Other amines combined (via ion-pairing) with saturated fatty acids were also shown to yield alarge polymorphism, but the mixtures finally crystallized upon

Received: January 19, 2011Revised: February 28, 2011

ABSTRACT: Unsaturated fatty acids may be extracted fromvarious agricultural resources and are widely used as soaps in theindustry. However, there also exist a large variety of saturatedand hydroxy fatty acids in nature, but their metal salts crystallizeat room temperature in water, hampering their use in biologicaland chemical studies or for industrial applications. Addition ofguanidine hydrochloride (GuHCl) to sodium salt of myristicacid has been shown to prevent its crystallization in water,forming stable flat bilayers at room temperature. Herein, weextend this finding to two other saturated fatty acids (palmiticand stearic acids) and two hydroxyl fatty acids (juniperic and 12 hydroxy stearic acids) and study more deeply (by using small angleneutron scattering) the supramolecular assemblies formed in both saturated and hydroxyl fatty acid systems. In addition, we take theadvantage that crystallization no longer occurs at room temperature in the presence of GuHCl to study the foaming and emulsifyingproperties of those fatty acid dispersions. Briefly, our results show that all fatty acids, even juniperic acid, which is a bola lipid, arearranged in a bilayer structure that may be interdigitated. Depending on the nature of the fatty acid, the systems exhibit goodfoamability and foam stability (except for juniperic acid), and emulsion stability was good. Those findings should be of interest forusing saturated long chain (and hydroxyl) fatty acids as surfactants for detergency or even materials chemistry.

B dx.doi.org/10.1021/la2002404 |Langmuir XXXX, XXX, 000–000

Langmuir ARTICLE

resting.13,14 For instance, we obtained unstable flat membranesor vesicles with saturated fatty acids using ethanolamine or lysineas a counterion.14,15

However, fatty acids are generally obtained by saponificationof natural oil (triglycerides); that is, they are available under theform ofmetal (Naþ or Kþ) carboxylates (soaps). Dispersing fattyacids under this soap form then would be of more practical value.Indeed, ion-pairing with TBAOH, ethanolamine, or other anyamine as counterion would require that the soap be previouslyacidified to yield carboxylic fatty acids (COOH group). Inter-estingly, in a preliminary study, we have shown that addition ofguanidine hydrochloride (GuHCl) to the sodium salt of myristicacid (see Figure 1 for the chemical structure of the various comp-onents used here) prevents its crystallization.16 One observedpeculiar phase behavior because flat membranes formed at roomtemperature, which transited at a given temperature to originalanastomosis-like superstructures. Here, we extend this study onlyat room temperature using two saturated fatty acids (palmitic andstearic acids) and two hydroxyl fatty acids, juniperic and 12hydroxy stearic acids (see Figure 1). Moreover, those systems arestudied by using small angle neutron scattering (SANS) to getvaluable information on the fatty acid assemblies obtained atroom temperature in the presence of GuHCl. Together withsolid-state NMR results, it provides information on the fatty acidarrangement within the edifices.

It has been outlined in the literature that fatty acids arepotentially promising surface active agents but their low solubi-lity at room temperature requires performing studies at hightemperature or inmixtures with other surfactants.17We then takethe advantage that the dispersions in the presence of GuHCl arestable at room temperature (no longer crystallize) to study theirinterfacial, foaming, and emulsifying properties.

’EXPERIMENTAL SECTION

Sample Preparation. Fatty acid and hydroxylated fatty acid(Sigma Aldrich, 99% purity) were weighed exactly in a tube, and ultrapure water was added so that the concentration was 10 mg/mL (1%).Next, the desired volumes of 1 M stock solutions of sodium hydroxide(Fishers Scientific, 97%purity) and of guanidium hydrochloride (GuHCl;Sigma�Aldrich, 99% purity) prepared in ultra pure water were incorpo-rated to reach equivalence (fatty acid/counterion/GuHCl molar ratio R =1/1/1). The mixture was melted at 80 �C for 15 min until all componentswere dispersed and then vigorously vortexed. Prior to being used, eachsample was heated at 80 �C for 15 min and cooled at room temperature.Phase-Contrast Microscopy. Phase-contrast microscopy was

used to investigate the variation of the tube diameter as a function oftemperature (20�80 �C). Microscopy observations were carried out at20� magnification using an optical microscope in the phase-contrastmode (Nikon Eclipse E-400, Tokyo, Japan) equipped with a 3-CCDJVC camera allowing digital images (768� 512 pixels) to be collected. Adrop of the lipid dispersion (about 20 µL) was deposited onto the glass-slide surface (76 � 26 � 1.1 mm, RS France) and covered with acoverslide (20 � 20 mm, Menzel-Glaser, Germany). The glass-slideswere previously cleaned with ethanol and acetone.Solid-State NMR. Deuterium solid-state NMR experiments were

performed at room temperature on a 400 MHz Bruker spectrometeroperating at 61 MHz for deuterium using a static double channel probe.The sample coil of the probe was adapted to load a 7 mm rotor such asthose used for magic angle spinning probes equipped with a stretchedstator. Typically, lipid dispersions were previously heated to 60 �C and avolume of ca. 500 μL was transferred into the rotor, which was sealedand then end-capped. AHahn quadrupolar echo sequence was used with

an inter pulse delay of 40 μs. Height k points in 1k accumulations (every2 s) were done with a 90� pulse and spectral width of 5 μs and 125 kHz,respectively. Free induction decay signals were zero-filled to 16k pointsprior to Fourier transform after a broad line exponential multiplicationof 10 Hz.

For deuterium spectroscopy, the general theory for lipid systemscan be found in the literature.18 Briefly, the deuterium NMR signal iscomposed of doublets with a splitting, Δν, which depends on theorientation of the C�D bond with respect to the magnetic field. In ananisotropic but disoriented medium, all of the orientations are allowed,and these doublets are superimposed to form a powder spectrum havingtwo main peaks with an increased intensity corresponding to the 90�orientation, separated byΔν90. The edge of the spectrum corresponds tothe 0� orientation, with a splittingΔν0 equal to twiceΔν90. In the case ofperdeuterated systems, the spectrum is composed of the superimposi-tion of signals from each labeled position.Small-Angle Neutron Scattering (SANS). Small-angle neutron

scattering (SANS) experiments were performed at Laboratoire L�eon-Brillouin (laboratoire mixte CEA/CNRS, Saclay, France) on spectro-meter PAXY. The neutron beamwas collimated by appropriately chosenneutron guides and circular apertures, with a beam diameter at thesample position of 7.6 mm. The neutron wavelength was set to 3 or 16 Åwith a mechanical velocity selector (Δλ/λ≈ 0.1), the 2D detector (128� 128 pixels, pixel size 5 � 5 mm2) being positioned at 3.1 m. Thescattering wave vector, Q, then ranges from typically 0.005 to 0.3 Å�1,with a significant overlap between the two configurations. The samples,prepared with deuterated water, were held in flat quartz cells with a 2mmoptical path and were temperature-controlled by a circulating fluid towithin (0.2 �C. The azimuthally averaged spectra were corrected forsolvent, cell, and incoherent scattering, as well as for background noise.

The general theory for fitting the SANS data can be found in theliterature.19 The overall procedure employed for fitting the data in thisArticle has already been described for other fatty acid containing systems.15

Preparation of o/w Emulsions. Emulsions were prepared with25 vol % n-hexadecane (Sigma Chemicals) as the oil phase and 75 vol %of the aqueous phase (fatty acid dispersion at 44mM). In all cases, a fixedvolume (6 mL) of emulsion was prepared with an ultrasound homo-genizer (SONICSVibra Cell VCX750) at a power level of 40%,making 3pulses of 10 s each with a 3 mm diameter probe.Preparation of Foams. Foams were produced and characterized

with a FOAMSCAN developed by IT CONCEPT (Longessaigne,FRANCE). Foam is generated in a glass column (21 mm in diameter)by sparging N2 gas through a fixed volume (12mL) of the aqueous phase(fatty acid dispersion) 1 mg/mL (0.1%) via a porous glass filter (poresize 10�14 μm). With this instrument, the foam formation and stabilitycan be determined by conductivity and optical measurements. The flowrate was fixed at 35 mL/min. All foams were allowed to reach a finalvolume of 45mL, after which the gas-flowwas stopped and the evolutionof the foamwas analyzed. The foam volumewas determined with a CCDcamera (Sony Hexwave HAD). The liquid volume in the foam wasassessed by mean of conductivity measurements along the foam column.Light Scattering Measurements. The stability of the emulsions

was determined through the use of a vertical scan analyzer Quick Scan(Beckman-Coulter Inc., U.S.). The samples were loaded into a cylind-rical glass measurement cell, and the backscattering percentage profiles(%BS) were immediately monitored as a function of the sample height(total height, 73 mm approximately). Cells were then stored at roomtemperature, and %BS measurements were done as a function of time.Therefore, creaming was determined as a function of time as the mean %BS measured at 20 mm from the bottom of the cell.Surface Tension Measurements. An automated drop tensi-

ometer (IT-Concept, Longessaigne, France) was used to measure theinterfacial tension between hexadecane and water and between air andwater. All experiments were performed in the rising bubble configuration. A

C dx.doi.org/10.1021/la2002404 |Langmuir XXXX, XXX, 000–000

Langmuir ARTICLE

15μLhexadecane drop and a 8μL air bubble were formed upward at the tipof a U-shaped stainless steel needle immersed in a cuvette filled with thefatty acid/counterion solution at 2 � 10�5 M, under gentle stirring, withtemperature controlled at 25 �C in all phases. The dynamic tension wasfollowed in time (10 000 s in our case) using axisymmetric drop-shapeanalysis. Image acquisition and regression of the interfacial tension wereperformed with Windrop software by fitting the Laplace equation to thedrop shape.

’RESULTS AND DISCUSSION

a. Self-Assembly inWater. In the present study, all fatty acids(see Figure 1) are initially used under their carboxylic (COOH)form, and the dispersions are obtained by adding NaOH andGuHCl in an equimolar ratio (see Experimental Section).

Generally, the sodium salt of the fatty acid is first realized(COO�/Naþ), and then GuHCl is added, but the three com-ponents can be mixed together in water without changing thephase behavior. Except when mentioned, all systems furtherstudied are sodium salts of fatty acids in the presence of GuHCl.For instance, when we refer to the palmitic acid system, thismeans the sodium salt of the palmitic (sodium palmitate) in thepresence of GuHCl.We have previously shown for the sodium salt of myristic acid

(sodium myristate) that addition of GuHCl prevents its crystal-lization at room temperature.16 For sodium palmitate andstearate, in the presence of GuHCl, both fatty acids formedturbid solutions at room temperature, the viscosity of whichincreased with the alkyl chain length. The solutions did notprecipitate when kept at 4 �C for more than 6 months, showing

Figure 1. Chemical structures and nomenclature of the fatty acids and GuHCl used in the present study.

Figure 2. Deuterium solid-state NMR spectrum recorded at room temperature for the palmitic acid system. That obtained with the stearic acid systemwas almost superimposable (not shown). The scheme on the right illustrates the fatty acid long molecular axis, Δ, the orientation with respect to thebilayer normal, n, and the conformation of the fatty acid in the bilayers. R would represent the angle of tilt (see text for details).

D dx.doi.org/10.1021/la2002404 |Langmuir XXXX, XXX, 000–000

Langmuir ARTICLE

that the presence of GuHCl also prevented the crystallization ofthose fatty acids as in the case of myristic acid system,16 and thesolutions also appeared as flat membranes by phase contrastmicroscopy16 (see Figure 3C). They transited (as viewed byvisual inspection) to viscous isotropic solutions at higher tem-perature, and this was probed by DSC as a single peak (see FigureSI1, Supporting Information). The transition temperature oc-curred at 42 and 57 �C, for palmitic and stearic acid systems,respectively. This increase of the transition temperature iscoherent with the increase of the alkyl chain length; however,wepresently did not further investigate the “high” temperature phase.The solid-state NMR spectrum at room temperature of the

sample made with the palmitic or stearic acid system was rathersimilar to that obtained previously for the myristic acid system16

(see Figure 2). An identical spectrumwas recorded again at roomtemperature for samples that were rested at 4 �C for a prolongedperiod of time (not shown). This showed that the presence ofGuHCl indeed prevented the long chain fatty acid crystallization.The spectrum in the case of the palmitic acid system is shown inFigure 2 and exhibits two large patterns on which two quad-rupolar splittings can be measured. The first quadrupolar split-ting of about 60 kHz stands for the positions along the chain, andthe second, at 19 kHz, is associated with the methyl end group.Those values are very close to those obtained for themyristic acidsystem.16 It shows that the fatty acid alkyl chains are in a stiff (all-trans) conformation. However, this also reveals an additionalfeature that has not been previously commented on. For an alkylchain in an all-trans conformation, the plane spanned by the CD2

makes an angle of 90� with respect to the long molecular axis, Δ(see Figure 2). The two deuteriums along the alkyl chain are thenoriented at 90�. It is generally assumed that this long molecularaxis is the axis of rotation of the molecule within the bilayer.18 Asa consequence, if this axis would be parallel to the normal of thebilayer, n (no tilt, see Figure 2), the expected quadrupolarsplitting would be one-half ((3 cos2(90�) � 1)/2 = �1/2) thatof the quadrupolar constant18 (125 kHz), that is, 62.5 kHz, closeto our experimental value. This argument also holds in the case ofthe CD3 groups where the quadrupolar splitting is alreadydecreased by a factor about 3 because of the axial rotation alongthe last carbon�carbon bond.20 This then means that the alkylchains are in an all-trans conformation and, in addition, not tiltedwith respect to the bilayer normal. Indeed, any tilt of Δ (of anangle R with respect to the bilayer normal, see Figure 2) woulddecrease the quadrupolar splitting by an additional term of (3cos2(R) � 1)/2.20

We also used juniperic acid and 12-hydroxy stearic acid ashydroxyl fatty acids. Those fatty acids are commercially availableunder their carboxylic form (COOH), and we then synthesizedthe sodium salts upon addition of NaOH (molar ratio 1/1). Bothsalts (in the absence of GuHCl) yielded isotropic solutions inwater (at 10 mg/mL) at 80 �C (probably forming micelles) butcrystallized upon cooling at room temperature. This was con-firmed by visual inspection of the samples and by opticalmicroscopy (Figure 3A and B). In the presence of GuHCl (atan equimolar ratio fatty acid:GuHCl:NaOH of 1:1:1), thesodium salts of juniperic and 12-hydroxy stearic acids did notcrystallize any longer, even when stored at 4 �C for 6 months. Asobserved by phase contrast microscopy, the juniperic acid systemformed stable flat membranes similar to those observed with thepalmitic acid system (Figure 3C) and for the ethanolamine salt ofthe juniperic acid in a previous study.15 Interestingly, the micro-graph of the samplemade with the 12-hydroxy stearic acid system

shows elongated randomly oriented rods (Figure 3D). Onceagain, this is very similar to what was observed with the ethano-lamine salt of that fatty acid.21�23 This strongly suggests that, inthe present case, 12-hydroxy stearic acid also self-assembles intotubular structures. Although we have not confirmed this supposi-tion by higher resolution microscopy (TEM or other), this isclearly further evidenced by SANS. However, we cannot performsolid-state NMR using those fatty acids because their deuteratedform is not commercially available. DSC was not performedon those samples because the solutions remained turbid uponheating and also because we further investigated those systemsonly at room temperature.To get more structural information on the assemblies formed

in those conditions, we performed small angle neutron scatteringexperiments (SANS).We report here the data for the palmitic, 12hydroxystearic, and juniperic acids systems (Figure 4). For allsamples, the spectra are qualitatively very similar. They undergo astrong small-angle scattering signal proportional toQ�2 forQf0, and two to three Bragg peaks can be identified at low Q in aratio 1:2 or 1:2:3 (Q0, 2Q0, 3Q0). This indicates that periodicmultilayer structures are formed in those systems. Bragg peaksare better defined for the rod-forming system using 12 hydroxystearic acid system, but the spectrum was similar to thatpreviously obtained with the ethanolamine salt of that fattyacid.23 This confirms that this system self-assembles into multi-lamellar tubes and that the rods observed by phase contrastmicroscopy are very probably hollow structures.We observed faint Bragg peaks for the other two samples,

indicating that bilayers exhibit strong thermal fluctuations orthat the number of diffracting objects is low. It is possible todescribe the main features of the spectra in terms of a simplemodel of planar membranes of well-defined thickness δ, themembranes being stacked with a period d = 2π/Q0. The period is480 Å (respectively, 470 and 340 Å) for the palmitic acid system(respectively, juniperic acid and 12-hydroxystearic systems). Formultilamellar tubes obtained with other counterions, similarvalues have been previously reported.23 Such a high capacity of

Figure 3. Micrographs obtained by phase contrast microscopy. (A)Crystals of sodium salt of 12 hydroxy stearic acid. (B) Crystals of sodiumsalt of juniperic acid (similar images were obtained for the sodium salt ofpalmitic acid). (C) Membranes obtained with the juniperic acid system(similar images were obtained for the palmitic acid system). (D) Rodsobtained with the 12 hydroxystearic acid system. The scale bar repre-sents 10 μm.

E dx.doi.org/10.1021/la2002404 |Langmuir XXXX, XXX, 000–000

Langmuir ARTICLE

lamellar swelling may be of interest for various applications suchas the encapsulation of drugs.At large Q, in the Porod regime, one probes the surface

scattering of the objects, and the scattering intensity decays likeQ�4. In the so-called Porod representation, where Q4 times thescattering intensity is displayed as a function of Q (see Figure 4),a main oscillation appears, related to the form factor of thebilayer. A fit to this form factor main oscillation yielded a value ofthe bilayer thickness δ of 33, 34, and 40 Å for the palmitic,juniperic acid, and 12 hydroxystearic systems, respectively.For the structures made of 12-hydroxy stearic acid system,

such a value of 40 Å is consistent with a fatty acid bilayerembedded in a gel Lβ phase as was previously observed andcommented on for the ethanolamine salt of that fatty acid at thattemperature.23 We can conclude that the overall structure of therods/tubes follows that shown Figure 5A on which d and δ are

also depicted, that is, a multilamellar tubular structure. In the caseof the palmitic acid system, the value of 33 Å is significantly lowerthan twice that of the extended conformation of the fatty acid (19Å). Indeed, in a bilayer conformation with alkyl chains embeddedin a gel Lβ phase (see NMR results), a value of 2*19 = 38 Åwouldbe expected (see Figure 5B). Because the NMR data also showedthat the chains are not tilted, the single possible conformation isthat shown in Figure 5B with the alkyl chains interdigitated.Interdigitation of alkyl chains is already known to occur insurfactant systems embedded in a gel Lβ phase.

24,25 This featurearises mainly because of an increase of the lateral area betweenthe surfactant headgroups, creating a void that can be filled by thealkyl chains of the opposite monolayer.In the case of the juniperic acid system, the situation is peculiar

because this fatty acid is known as a bola lipid.15,26�30 Bola lipidsare bipolar amphiphiles, and, as a consequence, they generally

Figure 4. SANS intensity profile for the fatty acid systems in the presence of GuHCl. Top, palmitic acid; middle, juniperic acid; and bottom, 12hydroxystearic acid. Both the Porod region (right) and the scattering profile in a logarithmic scale are shown (left). The continuous line corresponds tothe model discussed in the Experimental Section and in the text.

F dx.doi.org/10.1021/la2002404 |Langmuir XXXX, XXX, 000–000

Langmuir ARTICLE

self-assemble into monolayers instead of bilayers. Besides, wehave shown that the ethanolamine salt of juniperic acid yields flatmembranes made of monolayers.15 However, the thickness (34Å) we have obtained by SANS would be difficult to account for asit is higher than the extended conformation of the fatty acid,about 20 Å (see Figure 5C). This shows that a bilayer is formed inthe present case, probably tilted or interdigitated as previouslysuggested for palmitic acid. Another possible conformation isalso a bilayer but with their alkyl chains in a fluid LR phase (notshown in Figure 5C). The formation of bilayers in unsymmetricalbola lipids has already been reported30 and may occur via a head-to-tail, tail-to-tail, or head-to-head conformation within thebilayer. In that present case, the head is the carboxylic group,and the tail is the omega hydroxyl moiety. Note that junipericacid may also adopt an antiparallel configuration within a givenmonolayer.As mentioned in the literature, long chain fatty acids are

potentially good surface active agents, but their properties at theinterfaces can only be studied at high temperature or in mixtureswith other surfactants.17 We then take the advantage that stabledispersions of sodium fatty acid salts in water can be produced atroom temperature in the presence of GuHCl to study theirsurface activity.b. Adsorption Kinetics at the Air/Water or Hexadecane/

Water Interface. The adsorption kinetics of the various systemsat the air/water and oil/water interfaces was first studied usingdrop tensiometry at a given concentration of 0.02 mM. Theresults of the dynamic surface tension measurements of sodiumsalt of fatty acid solutions in the presence of GuHCl at the air/water interface are shown in Figure 6A. For C14 (the myristicacid system), one observed the fastest decrease of the surfacetension with a value lower than 50 mN/m after 10 000 s. For C16(the palmitic acid system), the adsorption kinetic was slower, but

the surface tension reached a similar value than for C14 at the endof the experiment. ForC18 (the stearic acid system), we observed nochange of the surface tension during 10 000 s, meaning that no fattyacid monomers were adsorbed at the interface during this period atthat concentration. Our measurements then showed that the moststriking factor influencing the surface tension is the fatty acid chainlength. This feature is in good agreement with data obtained withfatty acids salts at higher temperature by surface tension measure-ments with the pulsating bubble method.31�34

Surprisingly, for the 12-hydroxy stearic acid system, theadsorption kinetic was different from C18. For that hydroxyfatty acid, the adsorption rate was much faster than for C18, andthe decrease of the surface tension was in three steps. As shown inthe literature from Langmuir isotherm studies, this moleculepresents a bicompetitive adsorption between the primary andsecondary hydrophilic groups (carboxylic and hydroxyl groups).35

We can then hypothesize that the phenomenon observed inFigure 6A for that fatty acid may arise from rearrangements ofthe molecule at the interface upon further adsorption of fatty acids.Moreover for the 12-hydroxy stearic acid system, the surface tensionwas about 53 mN/m at 10 000 s. Thus, the surface tension is muchlower with the 12-hydroxy stearic acid system than for C18

Figure 5. Schematic illustration of the fatty acid self-assembly of the 12hydroxy stearic acid (A), palmitic acid (B), and juniperic acid (C)systems. The repeat distance, d, and the bilayer thickness, δ, as measuredby SANS are reported. Themultilayer tubular structure in the case of the12 hydroxy stearic acid system is shown (A). In the case of the palmiticacid system, the sketch shows the untilted bilayer, the thickness of whichdoes not match the SANS measurement (B). The most probableconformation, that is, interdigitated, is shown. In the case of the junipericacid system, the “double monolayer” tilted or untilted in a tail-to-tailconformation is depicted (C). Other conformations are also possible(see text).

Figure 6. (A) Dynamic surface tension measurements of fatty acid/NaOH/GuHCl at the air/water interface. (B) Dynamic surface tensionmeasurements of fatty acid/NaOH/GuHCl at the hexadecane/waterinterface.

G dx.doi.org/10.1021/la2002404 |Langmuir XXXX, XXX, 000–000

Langmuir ARTICLE

(its analogue) (see Figure 6A). Those two fatty acids have the samealkyl chain length anddiffer only by the presence of a hydroxyl groupin the alkyl chain (yielding a more hydrophilic molecule). It isknown that the surface activity of fatty acid soaps increases with thealkyl chain length, but their solubility decreases with the alkyl chainlength.31,32,36 In the present case, the 12 hydroxy stearic acid exhibitsthe advantage of both a relatively good solubility (because of thepresence of the hydroxyl group) and a long alkyl chain length thatcontribute to increase the surface activity.For the juniperic acid system, the surface tension was about 65

mN/m at 10 000 s; in other words, its surface activity is low. Bycomparison with its C16 analogue, this goes against our previousargument based on the relation between the solubility, the alkylchain length, and the surface activity. We could expect thatjuniperic acid would adopt a U-shape for exposing both thecarboxylic and the hydroxyl groups to the water. Such a con-formation has already been evidenced from monolayer experi-ments in the case of bola lipids when they are compressed at theinterface.37 However, it represents a high cost of energy for such afatty acid to freely adopt a U-shape at an interface. In comparisonto the 12 hydroxystearic acid, this shows that the position of thehydroxyl group along the alkyl chain is determinant.The results of the dynamic surface tensionmeasurement of the

fatty acid solutions in the presence of GuHCl at the hexadecane/water interface are also shown in Figure 6B. C14 and C16 havesimilar adsorption kinetics, with a first fast decrease and then asurface tension constant and slightly higher than 40mN/m. As inthe case of the air/water interface, we observed only a very smalldecrease of the surface tension for C18. Moreover, at thehexadecane/water interface, we observed that the 12-hydroxystearic acid system exhibited the fastest decrease of the surfacetension, with a value lower than 33 mN/m after 10 000 s. Thathydroxy fatty acid has a higher surface activity than its homo-logue, that is, stearic acid as observed at the air/water interface.Surprisingly, by contrast to what happens at the air/water inter-face for the 12-hydroxy stearic acid system, the variation of thesurface tension occurred in only one step at the hexadecane/water interface. This suggests that the orientation of that fattyacid is different at that interface.Moreover, somehow unexpectedly, we observed that the

surface tension after a period of 10 000 s was lower for the juni-peric acid system than for its analogue C16. This contrast withthe results obtained at the air/water interface. We suggest thatthe hydroxyl group of juniperic acid may have some affinity withthe oil component, allowing the fatty acid to stabilize the inter-face. As mentioned above, it is still unlikely that the juniperic acidadopts a U-shape with both polar heads (carboxylic and hydroxylgroups) in the water and with the alkyl chain facing the oil. Thesurface activity then depends on the fatty acid chain length andthe presence of a hydroxyl group and the nature of the interface(oil versus air).c. Foaming Properties.Next, for all fatty acid dispersions, we

produced foams and emulsions, and their stability was assessed.In the foaming experiment, the foam is generated by bubblingnitrogen until a given foam volume (45mL) is reached, after whichbubbling is stopped and the evolution of the foam is recorded as afunction of time. The Foamscan apparatus enables to measure thefoam volume as well as the liquid volume incorporated into thefoam. The variation of the foam volume with time for all fatty acidsystems is shown Figure 7A. For all systems, the given volume of45 mL was reached at about the same time of 75 s. Next, foams canbe produced with such systems. However, the foam structure as

well as the stability strongly differedwith the nature of the fatty acid.For C14, the foam was made of small bubbles (see SupportingInformation, SI2), and the foam volume decreased to reach a finalvalue of 22 mL at the end of the experiment. At that time, a visualobservation revealed that a slight coalescence has occurred in theupper part of the foam, exhibiting a shiny aspect. The liquid volumeincorporated in the foammadewithC14 at the end of bubbling (75s) was of 8.5 mL, which corresponded to a liquid fraction of 19%(Figure 7B). A very fast drainage was observed because the liquidvolume (and then the foam volume) decreased during 400 s, afterwhich the foam drained more slowly.For C16, we observed larger bubbles than for C14 (see

Supporting Information, SI2), but the foam was more stablebecause the final foam volume observed 30 min after the end ofbubbling was still above 40 mL. For this foam, less than 4.5 mLwas incorporated into the foam, a value much lower than for C14,which corresponded to a drier foam containing 10% of liquidfraction.For C18, the foam volume did not decrease during the exper-

iment, but such foam was constituted of very large and inhomo-geneous bubbles (see Supporting Information, SI2). The liquidvolume in this foamwas much lower with only 1.3 mL of solution

Figure 7. (A) Evolution of the foam volume for foams made of fattyacid/NaOH/GuHCl at a concentration of 1 g/L. (B) Evolution of theliquid volume in the foam for foams made of fatty acid/NaOH/GuHClat 1 g/L.

H dx.doi.org/10.1021/la2002404 |Langmuir XXXX, XXX, 000–000

Langmuir ARTICLE

being incorporated during the foaming process, corresponding tovery dry foam (less than 2% of liquid fraction). Our experimentsshow that the stability and the volume of liquid incorporated intothe foam depend on the alkyl chain length. The higher is the alkylchain length, the drier is the foam.The 12-hydroxy stearic acid system gave very stable foam with

fine homogeneous bubbles (see Supporting Information, SI2),and more than 7.5 mL was incorporated into the foam at theend of bubbling (17% of liquid fraction). At the end of theexperiment, the foam contained more than 1.1 mL of liquid. The12-hydroxy stearic acid system has a high surface activity (seeFigure 6A), which provides extremely stable foams.Finally, for the juniperic acid system, the foam volume decre-

ased rapidly, showing that the foam (although the foamwasmadeof fine homogeneous bubbles, see Supporting Information, SI2)was unstable as it disappeared in a few seconds after the end ofbubbling. This was due to a fast coalescence of bubbles inside thecolumn. The liquid incorporated into the foammade of junipericacid at the end of bubbling was of 7mL (15.5% of liquid fraction).This result may be correlated with the low surface activity ofthat fatty acid as measured previously. A good foamabilityassociated with unstable foams may be of practical interest forvarious applications. For instance, this is a prerequisite in washing(cleaning) processes or even for nuclear decontamination.d. Emulsifying Properties. We also produced emulsions for

all systems investigated above by sonication (see ExperimentalSection). Whatever the fatty acids, emulsions encapsulated allthe oil volume during the emulsification process and were stableat room temperature. The observation by optical microscopy justafter emulsification of the emulsions showed individual oildroplets dispersed in the aqueous phase (see Supporting In-formation, SI3). For all systems, the initial particle size distribu-tions weremonomodal withmean droplet diameters (d4,3) of 1.3( 0.5 μm. Emulsions were stable as we did not observe anymacroscopic changes (no phase separation) during one day.However, after 6 days, all emulsions were creamed. This isevident by visual inspection because creaming results from themigration of the emulsion droplets toward the top of the cell andresults in a clearing of the lower part of the solution. Coalescence

also occurred during that lap of time as a thin layer of oil was alsoobserved at the surface of the sample tubes (see SupportingInformation SI3 and SI4). The mean droplet diameters (d4,3)increased to 2.8( 0.2, 2.7( 0.4, and 3.5( 0.3 μm for C14, C16,and C18, respectively. For the hydroxylated fatty acids, the meandroplet diameters (d4,3) increased to 2.5( 0.4 and 2.6( 0.2 μmfor the 12-hydroxystearic acid and the juniperic acid systems,respectively.It is well-known that creaming can occur very rapidly even if

it has no consequence on the macroscopic aspect of the emul-sions. Thus, to get more accurate information on the creamingphenomenon, the percentage of backscattering of light profiles(%BS) was measured as a function of time. To evaluate thecreaming kinetics of the emulsions, the %BS profiles at 25 �Cwere analyzed as a function of time during 4500 min. Creamingis the consequence of the migration of the emulsion dropletstoward the top of the cell and results in a clearing of the lowerpart of the solution, yielding a decrease of %BS as illustrated inFigure 8. As the %BS was measured at 20 mm from the bottom ofthe cell, a lag is observed for all systems during a given time, andthen a decrease is observed and the %BS reaches a final value at4500 min. We further define that the lower is the %BS, the morestable is the emulsion.The %BS was about 30% for emulsions made of C18, whereas

it was of 39% for C14 and 42% for C16. Emulsions made withC14 and C16 were then more stable than emulsions made withC18. Clearly, the creaming kinetics depended on the chainlength, in agreement with previous experiments.17,38 Emulsionsmade with the 12-hydroxystearic and juniperic acids systemswere the most stable. During the experiment, the %BS onlydecreased from 80% to 55% and 50%, respectively. It is interest-ing to notice that juniperic acid is efficient to stabilize a water/oilinterface (emulsions), while it is unable to stabilize an air/waterinterface (foams). This is in agreement with the surface activitymeasurement of juniperic acid at the corresponding interfaces(see Figure 6B). All of these results show that relatively stableemulsion can be produced with these systems. The creamingkinetics depends on the chain length and the presence of ahydroxyl group. Thus, the emulsion stability can bemodulated bythe nature of the fatty acids. This behavior is in agreement withthe behavior observed at the model hexadecane/water interface(see Figure 6B).

’CONCLUSION

In our experimental conditions, the present various fatty acidsystems self-assembled into bilayers, even in the case of the bolalipid, juniperic acid. Foams and emulsions were successfullyproduced at room temperature, the stability of which can bemodulated by the nature of the fatty acid. Our present resultsconfirm that the presence of GuHCl is determinant to preventthe crystallization of sodium salts of long chain fatty acids andtheir hydroxyl derivatives. However, we still have to determinehow GuHCl interacts with fatty acids. This finding is unprece-dented and should be of interest for further applications aimed atusing such fatty acids as surfactants. To date, this is only restrictedto unsaturated fatty acids and short chain saturated fatty acids. Itis obvious that we should further examine the effect of the natureof other counterions (instead of sodium) in such fatty acidcontaining GuHCl to vary the polymorphism. Such systemsshould find applications in materials chemistry as templates oreven in biology, for instance, for studying the interaction of fatty

Figure 8. Kinetics of creaming (clarification) of the emulsions madewith 25 vol % n-hexadecane as the oil phase and 75 vol % of fatty acid/NaOH/GuHCl at 44 mM. %BS was measured at 20 mm from thebottom of the cell at 25 �C.

I dx.doi.org/10.1021/la2002404 |Langmuir XXXX, XXX, 000–000

Langmuir ARTICLE

acid binding proteins with long chain fatty acids or also in thefield of detergency where such fatty acids may be used as greendetergents.

’ASSOCIATED CONTENT

bS Supporting Information. Additional figures. This ma-terial is available free of charge via the Internet at http://pubs.acs.org.

’AUTHOR INFORMATION

Corresponding Author*E-mail: jean-paul.douliez@nantes.inra.fr.

’ACKNOWLEDGMENT

Access to the NMR facilities of the BIBS platform (Biopoly-m�eres Interactions Biologie Structurale) of INRA Angers-Nanteswas greatly appreciated. We would like to thank Fabrice Cousin(LLB, Saclay, France) for his precious help and powerful discus-sions during the neutron run and also María Cristina A~n�on(CIDCA, Buenos Aires, Argentina) for her support during thetraining period in her lab. A.-L.F. would like to thank l’INRA andle CEA for the allocation of her Ph.D. grant.

’REFERENCES

(1) Cistola, D. P.; Hamilton, J. A.; Jackson, D.; Small, D. M.Biochemistry 1988, 27, 1881–1888.(2) Cistola, D. P.; Atkinson, D.; Hamilton, J. A.; Small, D. M.

Biochemistry 1986, 25, 2804–2812.(3) Rico-Lattes, I.; Blanzat, M.; Franceschi-Messant, S.; Perez, E.;

Lattes, A. C. R. Chim. 2005, 8, 807–814.(4) Zemb, T.; Dubois, M.; Dem�e, B.; Gulik-Krzywicki, T. Science

1999, 283, 816–819.(5) Dubois, M.; Deme, B.; Gulik-Krzywicki, T.; Dedieu, J.-C.;

Vautrin, C.; Desert, S.; Perez, E.; Zemb, T. Nature 2001, 411, 672–675.(6) Hao, J.; Liu, W.; Xu, G.; Zheng, L. Langmuir 2003,

19, 10635–10640.(7) Segr�e, D.; Ben-Eli, D.; Deamer, D. W.; Lancet, D. Origins Life

Evol. Biospheres 2001, 31, 119–145.(8) Morigaki, K.; Walde, P. Curr. Opin. Colloid Interface Sci. 2007,

12, 75–80.(9) Zana, R. Langmuir 2004, 20, 5666–5668.(10) Zana, R.; Schmidt, J.; Talmon, Y. Langmuir 2005,

21, 11628–11636.(11) Klein, R.; Touraud, D.; Kunz, W. Green Chem. 2008,

10, 433–435.(12) Klein, R.; Kellermeier, M.; Drechsler, M.; Touraud, D.; Kunz,

W. Colloids Surf., A 2009, 338, 129–134.(13) Douliez, J. P.; Navailles, L.; Nallet, F. Langmuir 2006,

22, 622–627.(14) Novales, B.; Riaublanc, A.; Navailles, L.; Houssou, B. H.;

Gaillard, C.; Nallet, F.; Douliez, J. P. Langmuir 2010, 26, 5329–5334.(15) Novales, B.; Navailles, L.; Axelos, M.; Nallet, F.; Douliez, J. P.

Langmuir 2008, 24, 62–68.(16) Douliez, J. P.; Houinsou-Houssou, B.; Fameau, A.-L.; Novales,

B.; Gaillard, C. J. Colloid Interface Sci. 2010, 341, 386–389.(17) Golemanov, K.; Denkov, N. D.; Tcholakova, S.; Vethamuthu,

M.; Lips, A. Langmuir 2008, 24, 9956–9961.(18) Davis, J. Biochim. Biophys. Acta 1983, 737, 117–171.(19) Glatter, O.; Kratky, O. Small Angle X-ray Scattering; Academic

Press: London, 1982; p 3.(20) Douliez, J. P. Langmuir 2004, 20, 1543–1550.

(21) Fameau, A. L.; Houinsou-Houssou, B.; Novales, B.; Navailles,L.; Nallet, F.; Douliez, J.-P. J. Colloid Interface Sci. 2010, 341, 38–47.

(22) Douliez, J. P.; Pontoire, B.; Gaillard, C. ChemPhysChem 2006,7, 2071–2073.

(23) Douliez, J. P.; Gaillard, C.; Navailles, L.; Nallet, F. Langmuir2006, 22, 2942–2945.

(24) Kranenburg, M.; Vlaar, M.; Smit, B. Biophys. J. 2004,87, 1596–1605.

(25) Slater, J. L.; Huang, C.-H. Prog. Lipid Res. 1988, 27, 325–359.(26) Douliez, J. P. J. Am. Chem. Soc. 2005, 127, 15694–15695.(27) Fuhrhop, A. H.; Wang, T. Y. Chem. Rev. 2004, 104, 2901–2937.(28) Meister, A.; Blume, A. Curr. Opin. Colloid Interface Sci. 2007,

12, 138–147.(29) Visscher, I.; Engberts, J. B. F. N. Langmuir 2000, 16, 52–58.(30) Masuda, M.; Shimizu, T. Langmuir 2004, 20, 5969–5977.(31) Chang, C. H.; Wang, N. H. L.; Franses, E. I. Colloids Surf. 1992,

62, 321–332.(32) Coltharp, K. A.; Franses, E. I. Colloids Surf., A 1996,

108, 225–242.(33) Wen, X.; Lauterbach, J.; Franses, E. I. Langmuir 2000,

16, 6987–6994.(34) Yim, K. S.; Rahaii, B.; Fuller, G. G. Langmuir 2002,

18, 6597–6601.(35) Wen, X.; McGinnis, K. C.; Franses, E. I. Colloids Surf., A 1998,

143, 371–380.(36) Chang, C.-H.; Coltharp, K. A.; Park, S. Y.; Franses, E. I. Colloids

Surf., A 1996, 114, 185–197.(37) Meglio, C. D.; Rananavare, S. B.; Svenson, S.; Thompson, D. H.

Langmuir 2000, 16, 128–133.(38) Golemanov, K.; Tcholakova, S.; Denkov, N. D.; Gurkov, T.

Langmuir 2006, 22, 3560–3569.