Solid-Liquid Phase Transition in the Octadecanoic Acid Film Adsorbed on the Toluene-Water Interface
aa r X i v : . [ c ond - m a t . s o f t ] O c t Solid-Liquid Phase Transition in the Octadecanoic Acid FilmAdsorbed on the Toluene-Water Interface
Aleksey M. Tikhonov a,b ∗ a Kapitza Institute for Physical Problems, Russian Academy of Sciences, Moscow, 119334 Russia b Institute of Solid State Physics, Russian Academy of Sciences, Chernogolovka, Moscow region, 142432 Russia
Abstract
The structure of the soluble protonated (pH=2) octadecanoicacid film adsorbed on the saturated hydrocarbon (n-hexane) -water and aromatic hydrocarbon (toluene)-water interfacesis studied by X-ray reflectometry using synchrotron radia-tion. The experimental data demonstrate that a solid phaseof a Gibbs monolayer ± ˚A thick, in which aliphatic tailsare perpendicular to the surface and the area per molecule is A = 18 ± ˚A , forms in the film at the n-hexane - water in-terface. The solid monolayer on the toluene - water interfacein the adsorbed film melts when temperature increases, andthis transition is caused by disordering the hydrocarbon tailsof the acid. During the solid - liquid transition, the Gibbsmonolayer thickness remains almost the same, ± ˚A. Inthe solid phase, we have A = 20 ± ˚A and the angle of de-viation of the molecular tails from the normal to the surfaceis about ≈ ◦ . The density of the liquid monolayer phasewith A = 24 ± ˚A corresponds to liquid n-octadecane. INTRODUCTION
Thermotropic phase transitions between surfacemesophases are observed in the soluble amphiphylicsubstance film adsorbed on the nonpolar organic solvent(oil)-water interface. These transitions can be both ex-tended in temperature and characterized by sharp changesin the state of surface. The works on studying such surfacephenomena can be conventionally divided into the followingtwo types. The works of the first type investigate thestructure of the internal interfaces in the material volumethat appear due to the microscopic separation of phaseswith the formation of micelle and liposome solutions ormicroemulsions [1]. The works of the second type dealwith the interfaces between macroscopically large oil andwater volumes [2-6]. The authors of [7, 8] were the firstto demonstrate the possibility of application of X-rayreflectometry using synchrotron radiation to determine themolecular ordering on the macroscopically flat saturatedhydrocarbon (n-hexane)water interface. Later [9-12],we used this technique to study the thermotropic phasetransitions at this interface in adsorbed fatty alcohol andacid layers. The purpose of this work is to investigate thesolid - liquid phase transition in the soluble octadecanoicacid film adsorbed on the aromatic hydrocarbon (toluene)- water interface by X-ray reflectometry (see Fig. 1). This ∗ [email protected] Figure 1.
Molecular structures of (a) toluene C H , (b) n-hexane C H , and (c) octadecanoic acid C H O . interface is considered as a model interface to study, e.g.,the adsorption of the high-molecular-weight oil components(asphaltens) that do not dissolve in saturated hydrocarbons[13]. EXPERIMENTAL
All chemical components for experiments were bought atSigma-Aldrich. Saturated hydrocarbon n-hexane (C H ,the density at 298 K is ≈ .
65 g/cm , the boiling temper-ature is T b ≈
342 K) and aromatic hydrocarbon toluene(C H , the density at 298 K is ≈ .
87 g/cm , the boilingtemperature is T b ≈
384 ) were preliminarily cleaned bymultiple filtration in a chromatographic column [14]. Oc-tadecanoic acid C H COOH (stearic acid, or C -acid) isa monocarboxylic acid of the aliphatic series, does not dis-solve in water, and is well dissolved in toluene and n-hexane.This acid was purified by recrystallization from a supersat-urated solution in n-hexane at room temperature [12, 15].The samples of the flat toluene - water (n-hexane - wa-ter) interface, which was oriented by the gravitational force,was studied in a stainless steel temperature controlled cellaccording to the technique from [16]. The surface tension of1 the interface γ ( T ) in Fig. 2 was measured by the Wilhelmyplate method [17]. A solution of sulfuric acid (pH = 2) in ∼
100 mL of deionized water (Barnstead, NanoPureUV) wasused as the lower phase. The upper phases consisted of a ∼
50 mL solution of octadecanoic acid in toluene (n-hexane)with a volume concentration ≈
46 mmol/kg ( ≈ . · − ).Before being placed in the cell, these fluids were subjectedto degassing in an ultrasonic bath. When reflection coef-ficient R was measured, a sample was annealed: the fluidtemperatures in the cell was increased to ∼
330 K and wasthen decreased to the chosen temperature, and the samplewas brought in equilibrium in several hours when the lowerphase was accurately mechanically stirred [18, 19].C H O acid molecules from the solution in the hydro-carbon solvent are adsorbed onto the toluene - water in-terface, which significantly decreases its energy. As followsfrom Fig. 2, a phase transition takes place in the mono-layer on the interface when temperature T increases (ata pressure p = 1 atm). The phase-transition temperature( T c ≈
319 K) is determined by the C -acid concentration c in the solvent volume, which serves as a reservoir for surfac-tant molecules. The change in the slope of γ ( T ) is relatedto the relatively small change in the surface enthalpy duringthe transition, ∆ H = − T c ∆( ∂γ/∂T ) p,c = 0 . ± .
01 J/m .Note that the octadecanoic acid film adsorbed on the n-heaxane-water interface exhibits no specific features in thebehavior of surface tension at p = 1 atm in wide concen-tration (10 −
100 mmol/kg) and temperature (290 −
330 K)ranges.The transverse structure of the toluene-water (n-heaxane- water) interface was studied by X-ray reflectometry onthe X19C station of the NSLS synchrotron [20]. In exper-iments, we used a focused monochromatic beam with anintensity of about ≈ photons/s and a photon energy( λ = 0 . ± .
002 ˚A). The design of the X19C stationspectrometer makes it possible to investigate the surfaces ofsolids, liquids, and liquid-liquid interfaces [21-26].Figure 3 shows the kinematics of surface scattering by theinterface. In the reflectometry experiment, we have α = β ,where α is the grazing angle and β is the angle between thesurface plane and the direction to the point detector in theplane of incidence yz . Here, X-rays pass through the oilphase and are specularly reflected by the structure formedby the surfactant on the interface. If k in and k sc are the wavevectors of the incident and reflected beam in the detectordirection, respectively, scattering vector q = k in - k sc inthis experiment is normal to the surface along axis z oppositeto the gravitational force. When reflection coefficient R ismeasured as a function of q z = (4 π/λ ) sin α , it is averagedover a large illumination surface area ( ∼ . ) becauseof the height of the incident beam ( > µ m) in the yz planeand the width ( ∼ α c ≈ λ p r e ∆ ρ/π (where r e = 2 . · − ˚A is the classic electron radius and ∆ ρ isthe difference between the volume electron concentrationsof the fluids), the incident beam undergoes total externalreflection ( R ≈ ρ w ≈ . e − / ˚A ( e − (e − is the electron charge)in water, ρ h ≈ . ρ w in n-hexane, and ρ t ≈ . ρ w in Figure 2.
Temperature dependence of the interfacial tension ofthe toluene-water interface at an octadecanoic acid concentration c ≈ mmol/kg in the hydrocarbon solvent. The lines are drawnby eye and the inflection point corresponds to T c ≈ K. toluene. Thus, we have α c ≈ − rad ( ≈ .
06 deg) for then-hexane - water interface and α c ≈ · − rad ( ≈ .
03 deg),which is significantly lower, for the toluene - water system.which is significantly lower, for the toluenewater system.Figure 4 shows the experimental dependences of reflectioncoefficient R on q z normalized by the Fresnel function R F ( q z ) = q z − p q z − q c q z + p q z − q c ! , (1)where q c = (4 π/λ ) sin α c . The triangles correspond to thevalues of R ( q z ) /R F ( q z ) for the n-hexane-water interface at T = 295 K, and the circles and the squares, to the valuesfor the toluenewater interface at T = 308 and T = 328 K(below and above T c ), respectively. THEORY
From the experimental data R ( q z ), we restored the elec-tron concentration distribution ( z ) along the normal to thesurface using the qualitative one-layer model based on theerror function [2729] (Fig. 5) ρ ( z ) = 12 ( ρ + ρ ) + 12 ( ρ − ρ )erf (cid:18) zσ √ (cid:19) + 12 ( ρ − ρ )erf (cid:18) z + z σ √ (cid:19) , erf( x ) = 2 √ π Z x exp( − y ) dy, (2)where ρ and ρ are the electron concentrations in water andtoluene (n-hexane), respectively; ρ is the electron concen-tration in the Gibbs monolayer; z = 0; z is the monolayer Figure 3.
Kinematics of X-ray surface scattering by the toluene- water interface. α = β in the reflectometry experiment. thickness; and is the root-mean-square deviation of the po-sitions of the interfaces from their nominal values z and z ,which was taken to be equal to the capillary width in thecalculations. This width is determined by the surface spatialfrequency range covered in the experiment [3033], σ ≈ k B T πγ ln (cid:18) Q max Q min (cid:19) , (3)where Q max = 2 π/a is the short-wavelength spectral limit( a ≈
10 ˚A is the intermolecular distance on the order of mag-nitude), Q min = q maxz ∆ β/ β ≈ · − rad is the angular resolution of the detectorin the experiment, and q maxz ≈ .
25 ˚A − . Under the ex-perimental conditions, estimation (3) gives σ = 4 . ± . γ ≈
35 mN/m) for the n-hexane - water interface and σ = 5 . ± . σ = 6 . ± . T = 308 and T = 308 K, respectively, for the toluene - water interface.Profile (2) corresponds to the reflection coefficient [34, 35] R ( q z ) R F ( q z ) ≈ (cid:12)(cid:12)(cid:12)(cid:12)(cid:12)(cid:12) ρ X j =0 ( ρ j +1 − ρ j ) e − iq z z j (cid:12)(cid:12)(cid:12)(cid:12)(cid:12)(cid:12) e − σ q z √ q z − q c . (4) RESULTS AND DISCUSSION
The solid lines in Fig. 4 demonstrate that the experimen-tal
R/R F dependences are well described by Eq. (4), whichhas two adjustable parameters, namely, monolayer thickness z and electron concentration in it ρ . The electron concen-tration profile of the adsorbed layer δρ ( z ) is obtained fromEq. (1) by the subtraction of the contributions of the bulkphases to ρ ( z ), δρ ( z ) = ρ ( z ) − ρ (cid:20) − erf (cid:18) zσ √ (cid:19)(cid:21) − ρ (cid:20) (cid:18) z + z σ √ (cid:19)(cid:21) . (5) Figure 4.
Normalized reflection coefficient
R/R F as a func-tion of q z for interfaces with an adsorbed octadecanoic acid film:(triangles) n-hexane - water interface at T = 295 K; (circles,squares) toluene - water interface at T = 308 and 328 K, respec-tively. Solid lines are calculations with qualitative model (2). Thenumerals at the curves indicate their shifts along the coordinateaxis for convenience of presentation. The δρ ( z ) profiles normalized by ρ w are shown in Fig. 6.The thermodynamic properties of the soluble adsorbedfilm, which is considered to be a monolayer in a first ap-proximation (Gibbs monolayer), are described by parame-ters ( p, T, c ) [2, 36-38]. In the interface plane, this film canbe both isotropic and anisotropic despite the isotropy of thebulk phases [39]. In this system, first-order phase transitionsare formally prohibited, since the formation of an equilib-rium spatially inhomogeneous structure, in which the do-mains of two homogeneous phases coexist, from adsorbedmolecules is thermodynamically favorable in a certain vicin-ity of T c [40]. Both phases tend toward intermixing, sincethe formation of one-dimensional interfaces leads to a sig-nificant decrease in the energy [41]. As follows from thelyotropic and thermotropic phase transitions between thebulk mesophases in aqueous fatty acid solutions, one of theparameters that determine the thermodynamic state of asystem is solution pH, which affects the degree of ionizationof the COOH group [36]. Since the hydroxyl groups in themonolayer are not ionized at pH=2 according to H protonNMR data, it is conventionally called protonated [42].According to the fitting of the experimental data withEq. (4), the electron concentration in the Gibbs monolayerat the n-hexanewater interface is ρ = 0 . ± . e − / ˚A andits thickness is z = 26 ± A = Γ / ( z ρ )= 18 ± area per molecule, where Γ = 160is the number of electrons in the C H O molecule. Thecalculated total length of the octadecanoic acid molecule is L ≈ . × .
27 ˚A(-) + 1.5 ˚A(- ) + 2.5 ˚A(-)). Thus, asolid phase of the C -acid monolayer with aliphatic -C H tails, which are fully ordered and extended along the normalto the surface, exists at the n-hexane - water interface.The electron concentration in the Gibbs monolayer atthe toluenewater interface at T = 308 K is ρ = 0 . ± . e − / ˚A and z = 22 ± A = Γ / ( z ρ )= 20 ± , which also corresponds tothe solid phase of the monolayer but with tilted aliphatictails (deviated by θ = arccos( z /L ) ≈ ◦ from the nor-mal). Finally, at T = 328 K, we have ρ = 0 . ± . e − / ˚A , z = 22 ± A = 24 ± . This phase can be conven-tionally called liquid, since its density corresponds to thatof liquid n-octadecane C H [36].The values of parameters A and θ = 0 of the soluble oc-tadecanoic acid Gibbs monolayer on the n-hexane - waterinterface are close to the characteristics of the untilted solidphases of the insoluble octadecanoic acid Langmuir mono-layer on the water surface [43, 44]. These phases can be rep-resented by either hexagonal phase LS or distorted hexag-onal phase S [45]. The parameters of the Gibbs monolayerat the interface with toluene correspond to the character-istics of the tilted hexatic L d and Ov solid phases of theLangmuir C R I , R II ) of a high-molecular-weight saturated hydrocarbonnear its melting temperature [47-49].Thus, octadecanoic acid molecules form an ordered solidmonolayer on both the n-hexane - water and toluene - wa-ter interfaces. The microscopic mechanism of forming thesolid monolayer is likely to be based on the formation ofa two-dimensional network of hydrogen bonds between thecarbonyl (C=O) and hydroxyl (OH) groups of neighboringmolecules [50]. Our experimental data also illustrate a solid -liquid phase transition in the Gibbs monolayer on the tolue-newater interface, which is accompanied by the disorderingof -C H hydrocarbon tails. As temperature increases ina certain vicinity of T c , some adsorbed C -acid moleculesleave the interface and dissolve in the toluene volume. Inthis case, A increases by 1020% and the detected change inthe monolayer thickness z is insignificant.The use of toluene as the upper phase for structural stud-ies has both advantages and disadvantages. On the onehand, the difference between the volume electron concen-trations at the toluene - water interface (∆ ρ = ρ w − ρ t ≈ . e − /˚A ) is noticeably lower than that at the hexane -water interface (∆ ρ = ρ w − ρ h ≈ . e − /˚A ); therefore,the structure factor oscillation amplitude for the former in-terface is higher than that for the latter by a factor of(∆ ρ / ∆ ρ ) ≈ R/R F ∝ ∆ ρ − ac-cording to Eq. (4). This fact explains the higher R/R F oscillation for the toluene - water interface as compared tothat for the n-hexane - water interface in Fig. 3.On the other hand, the interfacial tension of the toluene- water interface is lower than that of the n-hexane - waterinterface by a factor of 23. Therefore, the intensity of non-specular (diffuse) scattering by the capillary wave roughnessof the former interface is substantially higher than that ofthe latter. This fact and the smaller depth of penetration ofphotons with E ≈
15 keV into toluene as compared to that
Figure 5.
Model of the interface. into n-hexane (18 and 24 cm, respectively) substantially (byabout 30%) decrease the grazing angle range to be measuredand to determine q maxz in Eq. (3).Thermotropic phase transitions, which usually have a des-orption origin, were detected in the layers of high-molecularfatty acids adsorbed on, e.g., the n-hexane - water inter-face [5, 15]. These are solid - gas monolayer and liquid -gas monolayer transitions, where almost all adsorbed lipidmolecules leave the interface and dissolve in the oil volumewhen temperature increases [10, 51, 52]. Multilayers werefound to exist in the normal alkanol films adsorbed on theneutral n-hexane water and n-hexadecane - water interfaceswhen the surfactant hydrocarbon chain length exceeds thesolvent molecule chain length by approximately six carbonatoms ( ∼ T c is preceded by tran-sition to multilayer adsorption at T ∗ > T c when temperature T increases [19].Note that the noncapillary wavy width of the toluene -water interface ( σ ≈ σ is of the same order of magnitude), we attribute thisfinding to relatively low contrast in the interfacial structure[18].As follows from our experimental data, the joint appli-cation of reflectometry and diffuse scattering to the systemunder study in the vicinity of T c can give additional usefulinformation about both the possibility of transition to mul-tilayer adsorption of octadecanoic acid at the toluene - waterinterface and the integrated characteristic of the roughnessspectrum ( σ ).Thus, the phenomena that occur at the interfaces in wa-ter - oil emulsions in the presence of impurity surfactantsinfluence the efficiency of oil technological processes [58-60].This investigation of the aromatic hydrocarbon - water in- Figure 6.
Model electron concentration profiles for octadecanoicacid monolayer δρ ( z ) normalized by the electron concentrationin water under normal conditions ( ρ w = 0 . e − /˚A ): (a) n-hexane - water interface and (b) toluene - water interface. Modelof a solid monolayer phase at T ≈ K (solid line) and modelof a liquid monolayer phase at T ≈ K (dashed line). terface and our earlier studies of the phase transitions at thesaturated hydrocarbon - water interface demonstrate funda-mentally new experimental abilities for revealing the essenceof the processes that occur in oil dispersed systems by X-ray reflectometry and diffuse scattering using synchrotronradiation.
ACKNOWLEDGMENTS
This work was performed using the resources of the Na-tional Synchrotron Light Source, US Department of Energy(DOE) Office of Science User Facility, operated for the DOEOffice of Science by the Brookhaven National Laboratoryunder contract no. DE-AC02-98CH10886. The X19C beam-line was supported by the ChemMatCARS National Syn-chrotron Resource, University of Chicago, University of Illi-nois at Chicago, and Stony Brook University. The theoreti-cal part of the work was supported by the Russian ScienceFoundation (project no. 18-12-00108).