Surfactant-Induced Phases in Water-Supported Alkane

Surfactant-Induced Phases in Water-Supported Alkane Monolayers:II. StructureShai Yefet,†,§ Eli Sloutskin,†,§ Lilach Tamam,†,∥ Zvi Sapir,†,⊥ Moshe Deutsch,*,† and Benjamin M. Ocko*,‡

†Physics Department and Institute of Nanotechnology, Bar-Ilan University, Ramat-Gan 52900, Israel‡Condensed Matter Physics & Materials Sciences Department, Brookhaven National Laboratory, Upton, New York 11973, UnitedStates

ABSTRACT: The structure of the Langmuir−Gibbs films ofnormal alkanes Cn of length n = 12−21 formed at the surfaceof aqueous solutions of CmTAB surfactants, m = 14, 16, and18, was studied by surface-specific synchrotron X-ray methods.At high temperatures, a laterally disordered monolayer ofmixed alkane molecules and surface-adsorbed surfactant tails isfound, having thicknesses well below those of the alkanes’ andsurfactant tails’ extended length. The mixed monolayerundergoes a freezing transition at a temperature Ts(n,m),which forms, for n ≤ m + 1, a crystalline monolayer of mixedalkane molecules and surfactant tails. For n ≥ m + 2, a bilayer forms, consisting of an upper pure-alkane, crystalline monolayerand a lower liquidlike monolayer. The crystalline monolayer in both cases consists of hexagonally packed extended, surface-normal-aligned chains. The hexagonal lattice constant is found to decrease with increasing n. The films’ structure is discussed inconjunction with their thermodynamic properties presented in an accompanying paper.

■ INTRODUCTION

When placed on the surface of a dilute surfactant solution, adroplet of alkane wets the surface, forming a monolayer ofmixed surfactant tails and alkane molecules, called a Langmuir−Gibbs (LG) film.1−4 This LG film is in equilibrium with theexcess alkane molecules residing as macroscopic lenslikedroplets on the surface.1,3,4 In a previous paper,5 we exploredthe phase diagram and thermodynamics of these LG films as afunction of the carbon number of the alkane (n) and thesurfactant’s tail (m) for surfactants of the alkyl-trimethylammo-nium bromide family at a few concentrations. While that studyprovided a detailed picture of the thermodynamics of thesesystems, and demonstrated the important role played by theinterchange energy of the two species in determining the phasediagram, no molecular-resolution structural information wasprovided by the macroscopic surface tension measurementsemployed.We present here a detailed study of the surface-normal and

surface-parallel structure of the LG films, both above and belowthe surface freezing temperature, for the same n, m, andsurfactant concentration ranges over which the thermody-namics of the system was studied. This study confirms theconclusion, drawn in ref 5, that the transition occurring at Ts isa surface freezing (SF) transition from a liquidlike, laterallydisordered, monolayer of mixed surfactant tails and alkanes, toan ordered solid phase of closely packed, extended molecules,exhibiting long-range hexagonal lateral order. For alkanelengths, n, below the crossover point n = m + 2 of the SFline, Ts(n), by the bulk freezing line, Tb(n), the frozen layer is amonolayer of mixed extended alkane molecules and extended

surfactant tails. For alkane lengths above the crossover point,the frozen layer is a bilayer, where the upper layer is acrystalline monolayer of pure alkanes, and the lower layer is adisordered, liquidlike layer of surfactant tails. An intermediatestructure, where both the lower and upper layers are frozen,was also found.

■ EXPERIMENTAL SECTIONMaterials and Procedures. The further purification of the

highest-purity commercially purchased materials, normal-alkanes,CnH2n+2 (denoted hereafter as Cn) and alkyltrimethylammoniumbromides (denoted as CmTAB), was described in the previous paper.5

Millipore water (18.2 Ω m) was used throughout.The samples for the X-ray measurements were either 1 mL of

surfactant solution spread on a glass slab 3 cm wide by 6 cm long, or asolution-filled shallow KelF (poly(chlorotrifluoroethylene)) trough of6 cm inner diameter. The sample was placed inside a cell consisting ofa cylindrical beryllium shell of 8 cm diameter, 5 cm height, and 0.5 mmwall thickness, sealed at the top and the bottom by copper flanges. Thethin beryllium shell provided low X-ray absorption, along with auniform-temperature environment due to the beryllium’s high thermalconductivity. The temperature of the cell was controlled to ≤0.01 °Cby a Lakeshore 340 temperature controller, using flat Minco heaters atthe top and bottom of the cell, and 100 kΩ (at 25 °C) YSI thermistorsas temperature sensors. A few microliters of freshly sonicated solutionof alkane in chloroform (see more detail in ref 5) were deposited onthe solution’s surface. After a 5 min wait for the chloroform toevaporate, and moving any excess alkane lenses to the substrate edges

Received: April 24, 2014Revised: June 9, 2014Published: June 11, 2014

Article

pubs.acs.org/Langmuir

© 2014 American Chemical Society 8010 dx.doi.org/10.1021/la501589t | Langmuir 2014, 30, 8010−8019

pubs.acs.org/Langmuir

to avoid their illumination by, and consequent scattering of, theincident X-ray beam, the cell was sealed and ready for themeasurements. Identical results were obtained for alkanes vapor-deposited from above using a hot reservoir of alkanes contained in acopper wick.X-ray Measurements. The X-ray measurements were carried out

using the Brookhaven-Harvard liquid surface diffractometer atbeamline X22B, National Synchrotron Light Source, BrookhavenNational Laboratory. To reduce beam damage, X-ray exposure timeswere minimized by using an automatic shutter, which was opened onlyfor counting and kept closed during spectrometer movements, waittimes, and so forth. Three surface-specific methods were employed, X-ray reflectivity (XR), grazing incidence diffraction (GID), and Braggrod (BR) measurements, with the scattering geometry shown in Figure1. Since these methods are by now well documented in theliterature,6−10 only a short summary will be given here.

X-ray Reflectivity. X-ray reflectivity (XR) is the measured fraction,R(qz), of the incident X-ray beam’s intensity reflected from thesolution surface as a function of the angle of incidence of the incomingbeam, α. qz = (4π/λ)sin(α) is the surface-normal component of thescattering vector q = kout − kin, and λ = 1.5173 Å is the wavelength ofthe X-rays used. In the scattering geometry plot in Figure 1, XR ismeasured at 2θ = 0° and β = α. R(qz) provides information on thesurface-normal structure of the interface. Specifically, the surfacenormal electron density profile, ρ(z), is related to R(qz) through theBorn approximation of a weak reflection by the Master Formula:6,7,10

∫ρ ρ= | ⟨ ⟩ − |−R q R q z z iq z z( )/ ( ) (d ( ) /d )exp( )dz z zF b1 2

(1)

where ρb is the bulk electron density and ⟨ρ(z)⟩ is the electron densityat a depth z below the surface, averaged over a surface-parallel plane.RF(qz), the Fresnel reflectivity, is the reflectivity from an ideally flat andabrupt interface. The density profile ρ(z) is extracted from themeasured R(qz) by constructing a physically motivated model for ρand calculating the corresponding analytic R(qz) through eq 1. This isthen fitted to the measured XR values to yield the model-definingparameter values.9

The level of structural detail extractable with confidence from ameasured R(qz) strongly depends on the extent of modulationsappearing in the measured curves, into which the model fit could“lock”, on the qz-range of the curve and on the error bars of themeasured points (which increase with qz). Thus, in modeling thestructure it is sometimes necessary to adopt approximations whichallow keeping the number of fit parameters low enough to render theirrefinement independent and confident. Specifically, it was foundnecessary to approximate the complex Stern−Helmholtz−Gouy−Chapman11−13 ionic double layer of the surfactants’ headgroups by asingle uniform slab, as done in previous ellipsometric studies.3,14 Thus,for the monolayer phase (liquid and solid), we use a model consistingof two constant-density slabs to represent the LG film. The upper slab

represents the mixed monolayer of alkyl tails and alkanes, and thelower slab, the headgroups. For the bilayer phase, two alkyl slabs areused, in addition to the headgroup layer. A very thin low-density slabseparating the two alkyl ones was found to be necessary for obtaining agood fit. The underlying bulk is represented by an additional infinite-thickness layer of a constant 0.333 e/Å3 electron density. Finally, thisso-called “box model” is convoluted with a single Gaussian ofadjustable width to account for the surface roughness generated by thethermally-excited capillary waves, present at all liquid surfaces andinterfaces.10 This choice reduces the number of fit parameters, whilestill allowing excellent fits to be obtained, as shown below. Test fitswith a separate roughness at each interface did not improve the fitquality.

Grazing Incidence Diffraction. Grazing incidence diffraction (GID)measures the surface-parallel structure of the LG films. This is done byscanning the detector out of the reflection plane by an angle 2θ, thusvarying the surface-parallel wavevector transfer,15 q∥:

π λ α β α β θ

π λ θ

= + −

≈

q (2 / ) cos cos 2cos cos cos 2

(4 / )sin(2 /2)

2 2

(2)

Here, β is the grazing angle of exit of an X-ray diffracted to thedetector. To minimize the scattering from the underlying bulk whichcould overwhelm the weak signal from the thin LG film, the incidenceangle is kept at α < αc where αc is the critical angle for total externalreflection.7 At these α values, only an evanescent wave penetratesbelow the surface, its intensity decaying exponentially with depth witha decay length of Λ = 2/(qc

2 − qz2)1/2, where qc = (4π/λ)sin(αc).

7 Inour GID measurements, done at qz/qc = 0.8−0.9 with qc ≈ 0.022 Å−1,the penetration depth is only Λ ≲ 100 Å, thus reducing bulk scattering.The GID measurements are carried out using a Soller-slits-precededlinear position sensitive detector, aligned normal to the surface, andscanned over a 0.1 Å−1 ≤ q∥ ≤ 1.8 Å−1 range. At a given azimuthalangle 2θ, the intensity integrated over the length of the detector is theGID signal, while the intensity distribution along the detector providesthe Bragg rod at this q∥ position.

Bragg Rod Measurement. The Bragg rod (BR) is the surface-normal (i.e., in Figure 1, the β- or qz-) distribution of the intensitydiffracted by the LG film at a given q∥ position. For q∥ positions notexhibiting a GID peak, only scattering from the underlying bulk isobserved. At a position where the GID pattern exhibits a peak, theshape of the BR provides information on the angle of tilt from thenormal and its azimuthal direction, the length of the scatteringmolecules, and their shape, thus allowing their form factor to bedetermined.10,15

■ RESULTS AND DISCUSSIONSurface-Normal Structure. The surface-normal structure

has been studied by X-ray reflectivity, and the correspondinglaterally averaged surface-normal density profiles, ⟨ρ(z)⟩, wereextracted by modeling the XR curves by a box model, asdiscussed above. The main types of ⟨ρ(z)⟩ observed are shownschematically in Figure 2 and discussed in the followingsubsections.

The Bare Surface of the Solution. The measured, Fresnel-normalized, reflectivity curves (symbols) of the surfaces of thesurfactant solutions before deposition of the alkane are shownin Figure 3. For a simple liquid, for example, water16,17 orsimple organic liquids,10,18 a monotonically decreasing R/RF isobserved with increasing qz. The modulated R/RF observedhere, showing a dip separating two higher-intensity regions, isakin to the Kiessig fringes of conventional optics. Themodulations indicate, therefore, the presence of a surfacelayer, denoted hereafter as the ”bare” surface layer (shown inFigure 2a), of an electron density differing from that of theunderlying bulk, and of a thickness inversely proportional to thedip’s qz position. The dip’s position is observed to decrease with

Figure 1. X-ray scattering geometry. kin and kout are the wavevectors|kin| = |kout| = 2π/λ of an incident and scattered X-ray, and q, qz, and q∥are the total, surface-normal, and surface-parallel scattering vectors,respectively. The grazing angle of incidence, α, and exit, β, and thescattering angle out of the reflection plane, 2θ, are also marked.

Langmuir Article

dx.doi.org/10.1021/la501589t | Langmuir 2014, 30, 8010−80198011

the surfactant’s length, indicating that the thickness of thesurface layer increases with m. The two CTAB curves in Figure3a, measured for 60% (□) and 100% (▽) of the criticalmicelle concentration (cmc), show that an increase in thesurfactant’s bulk concentration also shifts the dip to a lower qzvalue, indicating a corresponding increase in the alkyl layerthickness. This is a consequence of the change in the surfacecoverage by the surfactants upon increasing the bulkconcentration.19,20

Following previous studies,4 we use a two-slab model toobtain the density profiles ρ(z) from the measured R/RF, asdiscussed in the Experimental Section, and shown in Figure 2aabove. The water-adjacent slab represents the layer of theheadgroups, and the air-adjacent slab represents the layer ofalkyl tails. Good agreement is obtained in the fit (lines) withthe measured curves (symbols) for all three surfactants inFigure 3a. The corresponding density profiles, shown in Figure3b, exhibit for the headgroup slab an average thickness of 6.6 ±0.5 Å, and a slightly increased electron density, ∼0.36 eÅ−3, ascompared to 0.33 eÅ−3 for water. The significantly lowerdensity obtained for the alkyl chains’ slab, ∼0.27 eÅ−3, agreeswell with 0.266 eÅ−3 calculated from the known 30 Å3

volume21 and the eight electrons of a CH2 group.The alkyl slab thickness values obtained from the fit (Table 1

and Figure 3c) are about 1−1.5 Å lower than those inferred22

from sum frequency measurements.23 However, as discussedabove, the thickness is concentration dependent. Our lower

concentrations, as compared with those of the sum frequencystudy, yield, therefore, lower thickness values. These values are

Figure 2. Laterally-averaged, surface-normal density profiles found inthis study. The cartoons show the arrangements and conformations ofthe headgroups, tails (Cm alkyl groups of the surfactant molecules),and alkanes (Cn) in each density profile. The bulk contains solvatedsurfactant molecules which coexist with the surface adsorbed ones.

Figure 3. (a) Measured (symbols) and model-fitted (lines) Fresnel-normalized X-ray reflectivity curves from the bare solution surfacebefore alkane deposition for STAB, CTAB, and TTAB at a bulkconcentration of ∼60% of the cmc. For CTAB, the reflectivity curve at100% of the cmc is also shown in blue inverted triangles. Curves areshifted from each other vertically by a decade for clarity. (b) Thesurface-normal electron density profiles obtained from the fits areshown in the same color and line type as the lines in (a). Curves areshifted from each other vertically by 0.08 e/Å3. (c) Fit-derivedthicknesses, listed in Table 1 (symbols), and their linear fits (lines) forthe bare (solid line) and alkane-wetted liquid (dashed line) surfacelayers, at a bulk concentration of ∼60% of the cmc. The red trianglerepresents the bare CTAB surface at ∼100% of the cmc.

Table 1. Alkyl Slab Thickness, d, and Densities of Alkyl, ρt,and Headgroup, ρh, Slabs, Obtained from Fits to theMeasured XR Curves for the Alkane-Free Surface (bare) andthe Alkane-Wet Surface in Its Liquid Phase (liquid)a

alkyl slab head slab

surfactant lt, Å c, mM cmc, mM d, Å ρt, e/Å3 ρh, e/Å

3

bareTTAB 19.2 2 3.5(3) 7.4(7) 0.26(1) 0.37(1)CTAB 21.7 0.6 0.93(5) 9.9(10) 0.28(1) 0.38(1)CTAB 0.93 15.9(9) 0.30(2) 0.41(3)STAB 24.7 0.16 0.32(4) 12.0(8) 0.28(1) 0.36(1)

liquidTTAB 2 9.3(6) 0.26(1) 0.37(1)CTAB 0.6 11.7(8) 0.27(1) 0.37(1)STAB 0.16 14.3(11) 0.28(2) 0.37(1)

aSurfactant concentration, c, critical micelle concentration cmc (at 20°C, obtained from the linear fits in Figure 8b in Ref 5.), and literaturevalues for the surfactants’ extended tail lengths, lt,

22 are also listed. Theliquid-phase quantities are the same for all alkanes measured for agiven surfactant, as discussed in the text and shown in Figure 7.

Langmuir Article


all much lower than the extended lengths of the tails of theCmTAB molecules22 even for a maximal surface coverage,achieved at the cmc (0.93 mM for CTAB). Our d valuessupport therefore previous neutron reflection,22 ellipsometry,and sum frequency23 measurements which conclude that thelayer of surfactant tails is liquidlike and consists of flexible tailscontaining gauche segments rather than extended ones.3 Thisconclusion is further supported by the absence of anydiffraction peaks in the GID scans we measured for this layer,indicating the absence of surface-parallel order.The alkyl layer thickness should depend in principle on the

surface coverage by, and thus on the bulk concentration of, thesurfactant. In our case, the maximum coverage is determined bythe cross section of the headgroup, which is significantly largerthan that of the alkyl tail. Thus, the identical headgroups of thesurfactants studied here should yield a roughly equal coveragefor all three surfactants at concentrations which are an equalfraction of their respective cmc’s. This is the case for three ofthe solutions in Table 1, where concentrations of ∼60% of therespective cmc’s were used. Dividing the alkyl slab thickness din Table 1 by the number of methyls + methylenes residing inthis layer, m, yields an average effective height of 0.6 Å (±15%)per carbon within the “bare” layer. This height and the knownvolume of a methylene group in the liquid phase,21 30 Å3/CH2,yield, in turn, a surface area per surfactant molecule of ∼50 Å2,which corresponds to a coverage of Γ ≈ 3.3 μmol/m2. Thisvalue agrees with previous surface tension20 and neutronreflection19 measurements for the 0.6 mM CTAB solution and,within ≲20%, with the values obtained in ref 5 for STAB andTTAB from the measured γ(c) curves.The Liquid Phase of the LG Film. Alkane droplets were

placed on the bare surfactant-decorated water surface, at atemperature above the corresponding surface freezing temper-ature, Ts, listed in Table 1 of ref 5. At this T > Ts, they form atthe surface a monolayer of mixed surfactant tails and alkanemolecules, coexisting with 3D alkane lenses.4,14 The R/RFcurves measured here for such films are shown in Figure 4.Fits to the same two-slab model discussed in the previoussection are shown in solid lines in this figure, and the refinedparameters are listed in Table 1 under “liquid”. The generalshape of the corresponding density profiles is shown in Figure2b. For each surfactant, the dip in the R/RF curve is observed toshift from its position for the bare surface to a lower qz,indicating that the mixed monolayer is thicker than themonolayer of pure surfactants’ tails. However, the alkane-independent dip positions for each surfactant also show that theincrease in thickness is independent of the alkane length, n.Moreover, the increase, ∼1.8 Å, is also independent of thesurfactant length m, as demonstrated in Figure 3c by the dashedblack line, denoted “liquid”, being almost parallel to the solidblue line, denoted ”bare”. The fitted electron densities, ρt, of theliquid slab in Table 1 coincide with those of the bare surfactantlayer, indicating that the mixed layer is also a liquid film.Indeed, no GID peaks were found for the mixed layer at T > Ts.Bare and Liquid Surface Layer Thicknesses. To understand

why the bare and mixed surface layers have differentthicknesses, it is important to note that while both layers areliquidlike, their thicknesses are determined by different effects.For the bare layer, the areal density of the surfactant moleculesis determined by the bulk surfactant concentration through theadsorption thermodynamics. As the fitted ρt values in Table 1show, for this adsorption-determined area per molecule, thelayer thickness is set by the system to a value such that the

resultant electron density of the film coincides (within themeasurement error) with that of the bulk liquid alkane.The mixed layer’s thickness is determined very differently.

Here the droplets constitute a macroscopic reservoir of alkane(oil) molecules. Spreading, or otherwise, of these droplets onthe surfactant-decorated surface depends on the initialspreading coefficient, Si = γa̅w − (γow + γao), where γaw, γow,and γao are the interfacial tensions of the air/water, oil/water,and air/oil, respectively, and the overbar indicates the absenceof oil.2,24,25 If Si > 0, the free energy of the alkane-coveredsurface is lower than that of the air/water interface andspreading occurs. If Si < 0, the opposite is true and the dropletremains condensed and does not spread. The later is thesituation for all alkanes of n ≥ 7 on pure water.26 Adding asurfactant to the water at the concentrations used here switchesthe sign of Si from negative to positive, and a spreading ofalkane across the interface is induced.2,3,25

The equilibrium layer thickness is determined by theminimum of the system’s free energy. In addition to therelevant surface tensions, this free energy includes also a termaccounting for the long-range dispersive forces acting across thefilm’s two interfaces.2,11 The free energy is well approximated inour D range by F(D) = (γow + γao)−AH/(12πD

2), where D isthe film’s total thickness (see Figure 2) and AH is the Hamakerconstant of the air/alkane/water interface, which is positive forall alkanes studied here.11,25 This approximation for F(D) must,however, break down as D → 0, since without an alkane filmF(D) → γa̅w. For our positive AH, the dispersive term causesF(D) to decrease with decreasing D, that is, pushes for athinner wetting layer. However, since Si > 0, γa̅w is larger than(γow + γao). Thus, to reach γa̅w at D = 0, F(D) must switch froma decreasing to an increasing function at some nonzero D, thusforming a minimum at some value Deq ≠ 0, which is theequilibrium thickness of the wetting film. This wetting scenario

Figure 4. Measured (symbols) and model-fitted (lines) Fresnel-normalized X-ray reflectivity curves from the surface of the STAB,CTAB, and TTAB surfactant solutions of concentrations c = 0.6cmc,after deposition of droplets of the indicated alkanes, at a temperatureabove their respective surface freezing ones. An n-independent but m-dependent dip position is clearly observed and discussed in the text.The reflectivities measured before the alkanes deposition (“bare”) arealso shown for comparison.

Langmuir Article


is called pseudopartial wetting.24 The microscopic wetting filmthus formed coexists at the interface with the macroscopicalkane droplets.25 This is the mixed layer detected in ourmeasurements shown in Figure 4.The following conclusions emerge from the discussion above.

First, for the bare monolayer, the roughly equal 0.6cmc bulkconcentrations yield an equal areal density of surfactantmolecules for the different surfactants studied here. Con-sequently, the equal (liquid) density requirement leads to abare layer thickness which increases linearly with m (Figure 3c).Second, the liquid layer thickness favored by F(D) is obviouslygreater than that of the bare one. Thus, the positive spreadingcoefficient drives sufficient molecules from the droplets into thebare layer to achieve the thermodynamically favored equili-brium thickness. Space filling considerations require that, for agiven bare layer, the same amount of CH2 groups is drawn fromthe droplets in forming the mixed layer, regardless of the lengthof the alkane molecules in the droplets. This is the reason forthe equal dip positions, and hence equal layer thicknesses,observed in Figure 4 for all liquid layers originating in the samesurfactant, regardless of the alkane length forming the droplets.This conclusion of an equal density of CH2 groups’incorporation into the mixed layer implies that the number ofalkane molecules incorporated scales inversely with themolecular length n, supporting a similar conclusion of earlier,mostly macroscopic, studies.2,25,27−29

In the following, we show that the mole fraction of thesurfactants’ alkyl tails, ϕ, a quantity used in the thermodynamiccalculations of ref 5, can be obtained from the ratio of the bareand alkane-wetted liquid surface layer thicknesses. Using thelengths presented in Table 1, we find that the ratio is equal to0.80 ± 0.1, 0.85 ± 0.10, and 0.84 ± 0.10 for TTAB, CTAB, andSTAB, respectively. In calculating ϕ from this ratio, the exactchain lengths and the density differences between the terminalCH3 and CH2 moieties should be taken into account, a smalloverall effect due to the low abundance of CH3 groups. Also,this calculation is based on an assumption that the surfaceadsorption of the surfactant does not change upon alkanedeposition. This assumption is consistent with an absence ofany significant variation in headgroup slab’s thickness anddensity between the bare surfactant monolayer and the alkane-comprising liquid monolayer, in agreement with earlier studies.4

In addition, we neglect here the difference between mole andvolume fractions, which is in any case very small for these chainlengths, well below the sensitivity of thermodynamiccalculations5 to the precise value of ϕ. Therefore, in thislimit, ϕ is equal to the slab thickness ratio. Based on thisanalysis, we have set ϕ = 0.8 for all n and m in thethermodynamic calculations in ref 5. Also, with the moleculararea of A = 19.9 Å2, derived below from GID for the solid phase,the analysis above implies that for the liquid phase A ≈ 19.9/0.8= 25 Å2, 19.9/0.85 = 23.4 Å2, and 19.9/0.84 = 23.7 Å2, forTTAB, CTAB, and STAB, respectively, which are used in ref 5.Solid Surface Layer. Plots of the measured XR for T < Ts for

a range of alkane lengths are shown in Figures 5a and 6a forCTAB and STAB, respectively, with the top curve in each figureshowing, for comparison, the n-independent XR curvemeasured for the mixed liquid surface phase above Ts for thatsurfactant. Even without detailed modeling, one can see that forthe frozen monolayer the first and second dips in R/RF shift toqz values lower than those of the liquid monolayer, indicatingthe formation of a thicker layer. The sharpness of the seconddip implies a uniform-thickness layer. These characteristics

prevail up to n ≤ m + 1, while for n ≥ m + 2, a more complexXR curve is obtained. Here, shorter-period modulations,showing two periodicities, are observed in R/RF. These indicatean overall thicker surface film, consisting of two different-thickness layers.To gain quantitative understanding, the T < Ts-measured

R/RF for n ≤ m + 1 were fitted by the same model used for theliquid phase at T > Ts, discussed above. The resultant best fitsare shown in solid lines in Figures 5a and 6a, with thecorresponding density profiles shown in Figures 5b and 6b. Forboth surfactants, for n ≤ m + 1, we obtain a monolayer of n-dependent thickness, corresponding roughly to the extendedlength of the alkane, and having a density ρ = 0.31 e/Å3, equalto that of solid alkane. These properties prove the T < Ts

Figure 5. (a) Measured (symbols) and model-fitted (lines) Fresnel-normalized XR curves for 0.6 mM CTAB solution for T < Ts. An XRcurve (for C16) of the n-independent liquid phase at T > Ts is alsoshown at the top, for comparison. (b) The surface-normal electrondensity profiles corresponding to the model fits in (a) are shown insame color lines.

Figure 6. Same as Figure 5, but for a 0.16 mM STAB solution and C13for the liquid-phase curve.

Langmuir Article


monolayer to be a solid, as shown in Figure 2c. For n ≥ m + 2, abilayer structure is observed. The upper layer’s thickness equalsthe extended alkane molecule’s length, and its densitythat ofa solid alkane. The lower layer is thinner and has a density of aliquid alkane. Thus, the bilayer consists of an upper solidmonolayer supported on a lower liquid one, as shownschematically in Figure 2d.As mentioned above, for all bilayers, a thin depletion slab had

to be used between the upper solid layer and the lowerliquidlike layer to obtain a good fit. Following work by Ocko etal.,30 we suggest that this depletion slab originates in the lower-density methyl groups terminating both upper and lower layers,and is similar to that found between two basal planes of a 3Dalkane crystal.31 X-ray reflectivity lacks the sensitivity toindependently refine both width, δw, and density, ρdep, of adeeply-buried thin slab. Only their product can be confidentlydetermined from model fits.9,30 We refine therefore Φ = δw δρ,where δρ = ρav − ρdep and ρav is the average of the densities onboth sides of the slab. Our bilayer R/RF could be fit well for any0.1 Å ≤ δw ≤ 2 Å, with the fitted R/RF’s and density profilespractically coinciding with those shown in Figures 5 and 6. Thefits yield of course different ρdep for different δw, but a narrowlydistributed Φ = (0.48 ± 0.05) e/Å2. This Φ agrees to within20% with Φ = 0.58 and 0.56 e/Å2 obtained for crystallinealkanes and lipid bilayers,30−32 respectively, and thus supportsthe identification suggested here for this slab.The XR curve of C20 on STAB in Figure 6a deserves

additional discussion. Based on present, and previous,4 results,the two carbon difference between n and m should haveresulted in this R/RF being that of a bilayer. In fact, however,while it does exhibit the shorter-period modulations of abilayer, it lacks some of the other characteristics of bilayer XRcurves shown in Figures 5a and 6a, for example, the sharp dip atqz ≈ 0.24 Å−1. The model fits reveals that a bilayer densityprofile where both the top and bottom layer are solid (seeFigure 6b) fits this XR curve well. This could imply either auniform solid bilayer or a coexistence of domains of a solidmonolayer phase with domains of a “normal” bilayer phase,where the bottom layer is liquid and top layer is solid. As XRaverages over a macroscopic surface area, these two possiblestructures can yield (for certain range of relative domainabundances) very similar results that cannot be distinguished.We note that two-phase coexistence regions involving a solidmonolayer, a solid bilayer, and a liquid phase were very recentlyobserved in macroscopic optical measurements on a systemvery similar to ours.33 This point deserves further detailedstudy, perhaps by methods like GISAXS (grazing incidencesmall-angle X-ray scattering) and GTSAXS (grazing trans-mission small-angle X-ray scattering), which could resolve thedomain structure of the surface film.The thickness values of the surface layers supported on the

surfactant solutions studied here are shown in Figure 7. Theseare the thicknesses denoted by d and D in Figure 2. For themonolayer phase, n ≤ m + 1, as well as for the upper layer ofthe bilayer phase, the layer thickness d derived from the fitsagrees with the calculated extended length of the alkanemolecules,9,21,34,35 Ln = 1.27(n − 1) + 1.5 Å, shown in solidlines in Figure 7. This and the electron density (0.31 ± 0.01)e/Å3, which is equal to that of bulk solid alkanes, show that themonolayer is a solid consisting of a mixture of surface-normal-oriented, closely packed, extended alkane molecules andsurfactant tails. This conclusion is supported by our GID

measurements discussed below, which clearly demonstrate thatthe monolayer phase has a lateral hexagonal crystalline order.The measured layer thickness d of the solid monolayer phase

for STAB solutions is shown in Figure 7c to deviatesystematically and increasingly above the alkane’s length Lnwith decreasing n. The deviation can be rationalized by takinginto account the fact that the monolayer is a mixture of alkanesand surfactant tails, and thus, if both are fully extended andsurface-normal, the expected mixed monolayer’s thickness issimply a concentration-weighted average, davg, of the lengths ofthe shorter alkanes and the longer surfactant tails. This can beexpressed in the following equation:

= + −d L x L L( )n n navg 18 (3)

Here L18 is the length of the STAB molecule’s tail, taken to beequal to the length of an octadecane molecule, Ln is the lengthof the extended Cn molecule, and xn is the concentration of thesurfactant tails in the solid monolayer phase. xn is obtainedfrom the fit of the theory, discussed in ref 5, to the measuredsurface transition temperatures, Ts, as shown in Figure 6c of ref5. For mixtures with a carbon number mismatch of Δn ≤ 2, thedifference between davg and Ln is less than the measurementerror; hence, the fitted monolayer thickness agrees well with Ln.davg, as calculated from eq 3, is shown in a dashed red line inFigure 7c, and observed to agree well with the experimentalvalues, thus supporting the interpretation above.

Figure 7. Thickness values of the alkane layers supported on (a)TTAB, (b) CTAB, and (c) STAB solutions of a concentration of∼0.6cmc in the liquid and surface-frozen phases. d and D are thesurface layer thickness values (see Figure 2) derived from model fits tothe measured XR (symbols). The calculated extended molecularlengths (solid lines) and monolayer/bilayer phase boundaries (verticaldash-dot lines) are also shown. The red dashed lines in (b,c) are xn-weighted averages of the calculated extended lengths of the surfactantsand the relevant alkanes. See text for discussion.

Langmuir Article


The effect discussed in the previous paragraph is moredifficult to observe for CTAB solutions. Note that for STAB thelargest n where d shows a visible, out-of-error-bar, deviationfrom Ln in Figure 7c is n = 14, that is, four carbons less than theSTAB’s alkyl tail length. Moreover, the deviation for that n isstill less than two error bars, 2σ. Correspondingly, for CTABsuch a (minimal) deviation is expected to show up only at n =12. Indeed, the dashed line in Figure 7b, calculated from eq 3,agrees within 1σ with the measured values for n > 12, and evenfor n = 12, the shortest alkane measured, the deviation is stillwell within 2σ.The bilayer phase has, for each n, a top layer of a density

equal to that of the corresponding monolayer phase, 0.31 e/Å3.However, here, for all surfactants, the top layer’s thicknessagrees closely with the calculated extended lengths of thealkanes studied. This supports the identification of this layer asa solid layer consisting of pure alkanes only.4 For the bottomlayer, the liquidlike densities, (0.28 ± 0.01) e/Å3, andthicknesses, obtained from the fits, are close to those of thecorresponding liquid monolayer phases (Table 1), leading tothe conclusion that this layer is liquidlike. The GID resultsdiscussed next strongly support these conclusions. Finally, nosystematic variations with m, n, or phase (bare, liquid, andsolid) were observed in the surface roughness width refinedfrom the reflectivity fits. The average value of all fits, (3.7 ± 0.3)Å, agrees well with the value expected from capillary wavetheory.10,18

Surface Parallel Structure. Only the surface-frozen phasebelow Ts exhibits GID peaks, indicating that above Ts thesurface has no long-range lateral order, while below Ts suchorder exists. For the surface-frozen phase a single GID peak wasfound for all surfactant/alkanes combinations studied here,implying a hexagonal packing, in line with the similar packingfound for the surface frozen phases of molten alkanes9,36 andalcohols,37 and for the highly-condensed phases of Langmuirfilms comprising surface-normal-aligned, end-functionalizedalkane molecules.15,38 The GID peak measured for thesurface-frozen phase of C20/CTAB combination is shown inFigure 8a. The measured peak positions for all alkanes/surfactants combinations have an average value of q∥ = (1.515± 0.01) Å−1, corresponding to a hexagonal lattice constant of a= 4π/(√3q∥) = (4.789 ± 0.035) Å and a molecular area of A =√3a2/2 = (19.9 ± 0.3) Å2/molecule, in good agreement withthe corresponding value for surface frozen layers in alkanemelts.9,36 The full width at half-maximum (fwhm) of a GIDpeak, Δq∥, and the resolution width of our experimental setup,Δres, yield through the Debye−Scherrer formula,39 ξ ≈ 0.9 ×2π/[(Δq∥)2 − Δres

2]1/2, an estimate for the average crystallinecoherence length ξ within the film. We obtain ξCTAB = (420 ±50) Å and ξSTAB = (330 ± 50) Å for the average of all filmssupported on CTAB and STAB solutions, respectively. Thesetwo values agree, within their combined error bars, with eachother. Similar few-hundred-angstrom crystalline coherencelengths were found to be typical of surface-frozen rotatorphases in previously studied Langmuir−Gibbs films4 and forsome Langmuir films on water15 and mercury.38,40

The Bragg rod corresponding to the typical GID peak shownin Figure 8 extends along the qz axis (Figure 8b) and peaks at qz≈ 0 Å−1 (Figure 8c). These properties demonstrate,respectively, that the GID peak originates in a quasi-2D surfacelayer, rather than a 3D crystal,7,9,10 and that the diffractingmolecules are surface-normal oriented. The length extractedfrom the GID model fit, (24.3 ± 0.8) Å, is smaller than, but

close to, the extended length of a single C20 molecule, 25.6 Å,calculated from Ln above. The slight difference in these valuesmay result from a small fraction of chain-end gaucheconformations within the layer. The Bragg rod characteristicsfound here for this particular surface-frozen layer hold for allsurfactant/alkane combinations studied. Specifically, Bragg rodsalways correspond to a single molecular layer, where themolecules are closely packed, extended, and oriented along thesurface normal. This result is not surprising for alkane/surfactant combination satisfying n ≤ m + 1, where the XRmeasurements discussed above find a single surface-frozenmonolayer. For n ≥ m + 2, where XR reveals a bilayer, the BRs,like that in Figure 8c, demonstrate that only one of the layers isordered laterally. Were both upper and lower layers laterallyordered and in registration, a BR corresponding to twice theextended molecular length would have resulted. Such acalculated BR, shown in a dash line in Figure 8c, clearlydisagrees with the measured BR. The single ordered layer is, forall n, the pure alkane upper monolayer whose XR-derivedthickness is identical with that derived from the BR fit. Thethinner, lower layer does not exhibit long-range lateral orderand is therefore liquidlike, consisting of flexible tails containinggauche conformations and/or tilted surfactant tails.Based on the packing arrangement discussed above, the

observed coherence lengths, and the Bragg rod results, weconclude that the surface-frozen monolayers, that is, the mixedalkane/surfactant layer in the monolayer phase and the upper,pure alkane monolayer in the bilayer phase, are highly likely tobe in a rotator phase, same as the surface-frozen monolayer onalkane melts.9

To get a measure of the elastic properties of these films, wemeasured the temperature dependence of the lattice constant aof the hexagonal packing in surface-frozen films of C16/CTABand C15/STAB. These are shown in Figure 9a. The linearthermal expansion coefficients α2D = (da/dT)/a obtained for

Figure 8. GID and BR for the surface-frozen phase of C20 on a 0.6 mMCTAB solution. (a) Measured, background-subtracted GID peak(points), fitted by a single Gaussian (line). (b) Equal-intensity contourplot of the Bragg rod corresponding to the GID peak in (a). (c)Measured (points) and model-fitted (solid line) Bragg rod. The Braggrod calculated for two solid layers in registration is also shown (dashline).

Langmuir Article


the two films coincide to within their combined error bars.They are very close to the α2D values obtained for other van derWaals dominated quasi-2D structures, like 6.5 × 10−4 and 5 ×10−4 K−1 for mercury-supported monolayers of standing-upalcohol40 and fatty acid41 molecules, respectively, and 3 × 10−4

K−1 measured normal to the alkyl chains for mercury-supportedmonolayers of lying-down diacid molecules.42 The surface-frozen monolayer of a C20 alkane melt, the only one for whichα2D was measured,43 exhibits, surprisingly, a higher expansioncoefficient of (9 ± 0.05) × 10−4 K−1. While the reason for this30% difference is not clear, we note that this value is still of thesame order of magnitude as those found here.The a values obtained for the frozen phase of all CTAB and

STAB solutions, interpolated to T = 25 °C using α2D in Figure9a, and then averaged for each n, are plotted in Figure 9b and c,respectively. A clear decreasing trend is observed in these values(black circles). A linear fit to these values yields the lateralexpansion coefficient due to chain length variation, δ =(da/dn)/a. The negative values indicate a small decrease inthe lattice constants of (1.6 ± 0.1) × 10−2 and (1.1 ± 0.2) ×10−2 Å as n is increased by unity for CTAB- and STAB-solution-supported surface-frozen films, respectively. Within thescatter of the data points, no difference could be discernedbetween monolayer and bilayer phases for any surfactant. Thecontraction of the lattice constant with increasing alkane lengthis likely due to the increase in the van der Waals attractionbetween the alkane molecules as their lengths increase. α2Dmeasurements for surface-frozen monolayers on alkane melts,highly challenging because of the small variation of α2D over thesmall ≲3 °C temperature range of existence of surface freezinghere, were carried out to date for C20 only.

43 Thus, it is not

clear whether the n-dependent contraction observed here existsin those monolayers as well, and whether its magnitude differsin the three different surface phases exhibited by thesemonolayers as n is varied from 15 to 50.9 Such measurementsare clearly desirable.A comparison of the contraction found here for LG films to

the bulk alkane behavior is complicated by the rich (n, T) phasediagram of alkanes in the T-region just below their bulk freezingTb. In particular, the bulk phase closest in structure to oursurface-frozen LG films is the RII rotator phase, which exhibitshexagonal packing of molecules with their long axis alignednormal to the basal planes of the alkane crystal, and theirbackbone planes are free to rotate around that axis. However,this phase appears in the bulk only for alkane lengths n = 22−26, outside the n-range studied for our LG films.44 Within thatn-range, however, no change exceeding the error bars wasfound in the hexagonal lattice constant, indicating that nolattice contraction occurs with increasing n in the bulk RIIrotator phase. Lower-T, fully-crystalline, rhombohedral phasesof alkanes also show no n-dependent contraction of thehexagonal lattice spacing.45−47 Thus, the contraction with nmay be a unique feature of our LG films.

■ SUMMARY AND CONCLUSIONSWe presented here a detailed study by X-ray methods of thesurface parallel and surface normal structure of the “bare”surfactant solution, and the Langmuir−Gibbs film formed atthe solution’s surface when placing on it a macroscopic dropletof liquid alkane, resulting in pseudopartial wetting.2,26 Afterdepositing the droplet, we find above the surface-freezingtemperature Ts a mixed monolayer of alkanes and surfactanttails, exhibiting no lateral long-range order. Surface freezing atTs yields two different structures: for n ≤ m + 1, the mixedmonolayer simply freezes, while for n ≥ m + 2, the mixedmonolayer phase separates vertically, forming a frozen uppermonolayer of pure alkanes and a lower liquidlike monolayer ofsurfactants alkyl tails. The frozen layers in both cases comprisefully extended, surface-normal-aligned alkane molecules, packedin a hexagonal crystalline structure of a lattice constant whichdecreases with increasing alkane length, a behavior observedhere, to our knowledge, for the first time for any type ofinterfacially frozen monolayers. No further structural transitionswere found in these layers beyond the freezing of the liquid LGfilm. The layer thicknesses in the bare and liquid surface phasesdepends on the surfactant’s tail length, but is independent ofthe alkane length, an effect rationalized by the requirements ofspace filling, and the n-independent cohesive energy of theliquid layer. The linear thermal expansion coefficients of thefrozen monolayer phase were also determined for both CTABand STAB and were found to coincide with each other. Thevalues obtained are similar to those of standing-up phases ofmercury-supported Langmuir films of alcohol and fatty acidmolecules, and close to those of surface-frozen monolayers ofalkane melts.The results derived here from the structural X-ray measure-

ments, and those derived in the accompanying paper5 fromsurface tension measurements, agree well with, and support,each other, forming a full picture of the thermodynamics, phasediagram, and structure of the Langmuir−Gibbs films studiedhere. Nevertheless, a few puzzles still remain. One is the natureof the transition observed in the slope of γ(T) at a temperaturedenoted as Tr, which yields an n-dependent line closely parallelto, but a few degrees below, the Tb(n) line. The surfactant

Figure 9. (a) Temperature dependence of the hexagonal latticeconstant a for the alkane/surfactant combinations specified. α2D is thethermal linear expansion coefficient calculated from the linear fit(lines) for each of the two samples. (b,c) Lattice constant a versusnumber of alkane carbons n for the two surfactants. Each point is anaverage of two to six individual measurements for a given n, eachadjusted to T = 25 °C using α2D in (a). δ is the lateral expansioncoefficient per carbon due to molecular length n variation, defined inthe text, as obtained from linear fits to the temperature-adjusted values(lines).

Langmuir Article


concentration dependence should also be explored well beyondthe two concentrations measured here for STAB and TTAB.Also, the important surfactant-length dependence of theinterchange energy ω was measured here only for CTAB andSTAB, which are too few and too closely spaced to yield insightinto, and the elucidation of, this dependence and its origin.Measurements for additional CmTAB surfactants are clearlyindicated.We hope that the detailed experimental results presented in

this and the accompanying paper5 will stimulate further work,in particular theoretical and simulational, on the molecular-levelorigin and nature of the surfactant-induced surface freezingeffect.

■ AUTHOR INFORMATIONCorresponding Authors*E-mail: [email protected].*E-mail: [email protected] Addresses∥L.T.: Nova Measuring Instruments Ltd., Rehovot, Israel.⊥Z.S.: Intel (Israel) Ltd., Kiryat Gat, Israel.Author Contributions§S.Y. and E.S. contributed equally to this work.NotesThe authors declare no competing financial interest.

■ ACKNOWLEDGMENTSSupport by the U.S.-Israel Binational Science Foundation,Jerusalem (MD) and beamtime at X22B, NSLS, BrookhavenNational Laboratory, are gratefully acknowledged. This researchwas supported by the U.S. Department of Energy, Office ofBasic Energy Sciences, Materials Sciences and EngineeringDivision, under Contract No. DE-AC02-98CH10886 (BMO).

■ REFERENCES(1) McKenna, C. E.; Knock, M. M.; Bain, C. D. First-order phasetransition in mixed monolayers of hexadecyltrimethylammoniumbromide and tetradecane at the air-water interface. Langmuir 2000,16, 5853−5855.(2) Wilkinson, K.; Bain, C.; Matsubara, H.; Aratono, M. Wetting ofsurfactant solutions by alkanes. ChemPhysChem 2005, 6, 547−555.(3) Wilkinson, K. M.; Lei, Q.; Bain, C. D. Freezing transitions inmixed surfactant/alkane monolayers at the air-solution interface. SoftMatter 2006, 2, 66.(4) Sloutskin, E.; Sapir, Z.; Bain, C. D.; Lei, Q.; Wilkinson, K. M.;Tamam, L.; Deutsch, M.; Ocko, B. M. Wetting, mixing, and phasetransitions in Langmuir−Gibbs films. Phys. Rev. Lett. 2007, 99, 136102.(5) Yefet, S.; Sloutskin, E.; Tamam, L.; Sapir, Z.; Cohen, A.; Deutsch,M.; Ocko, B. M. Surfactant-induced phases in water-supported alkanemonolayers: I. Thermodynamics. Langmuir 2014, DOI: 10.1021/la501567s.(6) Als-Nielsen, J.; Jacquemain, D.; Kjaer, K.; Leveiller, F.; Lahav, M.;Leiserowitz, L. Principles and applications of grazing-incidence X-rayand neutron-scattering from ordered molecular monolayers at the air-water-interface. Phys. Rep. 1994, 246, 252−313.(7) Als-Nielsen, J.; McMorrow, D. Elements of Modern X-ray Physics;Wiely: New York, 2001.(8) Daillant, J.; Gibaud, A. X-ray and Neutron Reflectivity: Principlesand Applications; Springer: Berlin, 2009.(9) Ocko, B. M.; Wu, X. Z.; Sirota, E. B.; Sinha, S. K.; Gang, O.;Deutsch, M. Surface Freezing in Chain Molecules: Normal Alkanes.Phys. Rev. E 1997, 55, 3164.(10) Pershan, P. S.; Schlossman, M. L. Liquid Surfaces and Interfaces:Synchrotron X-ray Methods; Cambridge University Press: Cambridge,2012.

(11) Israelachvili, J. Intermolecular and surface forces, 2nd ed.;Academic Press: London, 1992.(12) Schmickler, W. Interfacial electrochemistry; OUP: New York,1996.(13) Grahame, D. C. The electrical double layer and the theory ofelectrocapillarity. Chem. Rev. 1947, 41, 441−501.(14) Lei, Q.; Bain, C. D. Surfactant-induced surface freezing at thealkane-water interface. Phys. Rev. Lett. 2004, 92, 176103.(15) Kaganer, V.; Mohwald, H.; Dutta, P. Structure and phasetransitions in Langmuir monolayers. Rev. Mod. Phys. 1999, 71, 779−819.(16) Shpyrko, O.; Fukuto, M.; Pershan, P.; Ocko, B.; Kuzmenko, I.;Gog, T.; Deutsch, M. Surface layering of liquids: The role of surfacetension. Phys. Rev. B 2004, 69, 245423.(17) Vaknin, D.; Bu, W.; Sung, J.; Jeon, Y.; Kim, D. Thermally excitedcapillary waves at vapor/liquid interfaces of water-alcohol mixtures. J.Phys.: Condens. Matter 2009, 21, 115105.(18) Ocko, B. M.; Wu, X. Z.; Sirota, E. B.; Sinha, S. K.; Deutsch, M.X-ray reflectivity study of thermal capillary waves on liquid surfaces.Phys. Rev. Lett. 1994, 72, 242−245.(19) Manning-Benson, S.; Parker, S. R. W.; Bain, C. D.; Penfold, J.Measurement of the dynamic surface excess in an overflowing cylinderby neutron reflection. Langmuir 1998, 14, 990−996.(20) Gilanyi, T.; Varga, I.; Stubenrauch, C.; Meszaros, R. Adsorptionof alkyl trimethylammonium bromides at the air/water interface. J.Colloid Interface Sci. 2008, 317, 395−401.(21) Small, D. M. The Physical Chemistry of Lipids; Plenum: NewYork, 1986.(22) Lu, J. R.; Thomas, R. K.; Penfold, J. Surfactant layers at the air/water interface: structure and composition. Adv. Colloid Interface Sci.2000, 84, 143−304.(23) Bell, G. R.; Manning-Benson, S.; Bain, C. D. Effect of chainlength on the structure of monolayers of alkyltrimethylammoniumbromides (CnTABs) at the air−water interface. J. Phys. Chem. B 1998,102, 218−222.(24) Brochard-Wyart, F.; Di-Meglio, J. M.; Quere, D.; De-Gennes, P.G. Spreading of nonvolatile liquids in a continuum picture. Langmuir1991, 7, 335−338.(25) Ash, P. A.; Bain, C. D.; Matsubara, H. Wetting in oil/water/surfactant systems. Curr. Opin. Colloid Interface Sci. 2012, 17, 196−204.(26) Bonn, D.; Eggers, J.; Indekeu, J.; Meunier, J.; Rolley, E. Wettingand spreading. Rev. Mod. Phys. 2009, 81, 739−805.(27) Aveyard, R.; Binks, B. P.; Cooper, P.; Fletcher, P. D. I.Incorporation of hydrocarbon into surfactant monolayers. Adv. ColloidSurf. Sci. 1990, 33, 59−77.(28) Lu, J. R.; Thomas, R. K.; Aveyard, R.; Binks, B. P.; Cooper, P.;Fletcher, P. D. I.; Sokolowski, A.; Penfold, J. Structure andcomposition of dodecane layers spread on aqueous-solutions oftetradecyltrimethylammonium bromide - neutron reflection andsurface-tension measurements. J. Phys. Chem. 1992, 96, 10971−10978.(29) Lu, J. R.; Thomas, R. K.; Binks, B. P.; Fletcher, P. D. I.; Penfold,J. Structure and compsoition of dodecane layers spread on aqueoussolutions od doddecyl- and hexadecyltrimethylamonium bromidesstudied by neutron reflection. J. Phys. Chem. 1995, 99, 4113−4123.(30) Ocko, B. M.; Dhinojwala, A.; Daillant, J. Comment on HowWater Meets a Hydrophobic Surface. Phys. Rev. Lett. 2008, 101,039601.(31) Craievich, A. F.; Denicolo, I.; Doucet, J. Molecular motion andconformational defects in odd-numbered paraffins. Phys. Rev. B 1984,30, 4782−4787.(32) Wiener, M. C.; Suter, R. M.; Nagle, J. F. Structure of the fullyhydrated gel phase of dipalmitoylphosphatidylcholine. Biophys. J. 1989,55, 315−325.(33) Matsubara, H.; Takaichi, T.; Takiue, T.; Aratono, M.; Toyoda,A.; Iimura, K.; Ash, P. A.; Bain, C. D. Morphological transformationsin solid domains of alkanes on surfactant solutions. J. Phys. Chem. Lett.2013, 4, 844−848.(34) Kjaer, K.; Als-Nielsen, J.; Helm, C. A.; Tipman-Krayer, P.;Mohwald, H. Synchrotron X-ray diffraction and refelction studies of

Langmuir Article


mailto:[email protected]

mailto:[email protected]

arachidic acid monolayers at the air-water interface. J. Phys. Chem.1989, 93, 3200−3206.(35) Petrov, J. G.; Pfohl, T.; Mohwald, H. Ellipsometric chain lengthdependence of fatty acid Langmuir monolayers. A heads-and-tailsmodel. J. Phys. Chem. B 1999, 103, 3417−3424.(36) Wu, X. Z.; Sirota, E. B.; Sinha, S. K.; Ocko, B. M.; Deutsch, M.Surface crystallization of normal-alkanes. Phys. Rev. Lett. 1993, 70, 958.(37) Gang, O.; Ocko, B. M.; Wu, X. Z.; Sirota, E. B.; Deutsch, M.Surface freezing in hydrated alcohol melts. Phys. Rev. Lett. 1998, 80,1264−1267.(38) Ocko, B. M.; Kraack, H.; Pershan, P. S.; Sloutskin, L.; Tamam,L.; Deutsch, M. Crystalline phases of alkyl-thiol monolayers on liquidmercury. Phys. Rev. Lett. 2005, 94, 017802.(39) Guinier, A. X-Ray Diffraction in Crystals, Imperfect Crystals andAmorphous Bodies; Freeman: San Francisco, 1963.(40) Kraack, H.; Ocko, B. M.; Pershan, P. S.; Sloutskin, E.; Tamam,L.; Deutsch, M. The structure and phase diagram of langmuir films ofalcohols on mercury. Langmuir 2004, 20, 5386−5395.(41) Kraack, H.; Ocko, B. M.; Pershan, P. S.; Sloutskin, E.; Tamam,L.; Deutsch, M. Fatty acid langmuir films on liquid mercury: X-ray andsurface tension studies. Langmuir 2004, 20, 5375−5385.(42) Tamam, L.; Ocko, B. M.; Deutsch, M. Two dimensional orderin mercury-supported Langmuir films of fatty diacids. Langmuir 2012,28, 15586−15597.(43) Ocko, B. M.; Sirota, E. B.; Deutsch, M.; DiMasi, E.; Coburn, S.;Strzalka, J.; Zheng, S. Y.; Tronin, A.; Gog, T.; Venkataraman, C.Positional order and thermal expansion of surface crystalline N-alkanemonolayers. Phys. Rev. E 2001, 63, 032602.(44) Sirota, E. B.; Kink, H. E.; Singer, D. M.; Shao, H. H. Rotatorphases of the normal alaknes-an X-ray scattering study. J. Chem. Phys.1993, 98, 5809−5824.(45) Dirand, M.; Bouroukba, M.; Chevallier, V.; Petitjean, D.; Behar,E.; Ruffier-Meray, V. Normal alkanes, multialkane synthetic modelmixtures, and real petroleum waxes: Crystallographic structures,thermodynamic properties, and crystallization. J. Chem. Eng. Data2002, 47, 115−143.(46) Nyborg, S. C.; Potworow, J. A. Prediction of units cells andatomic coordinates for normal alkanes. Acta Crystallogr., Sect. B 1973,B 29, 347−352.(47) Nyborg, S. C.; Pickard, F. M.; Norman, N. X-ray powderdiagrams of certain normal-alkanes - corrigenda and extension. ActaCrystallogr., Sect. B 1976, 32, 2289−2293.

Langmuir Article


Documents

Surfactant-Induced Phases in Water-Supported Alkane