Langmuir 2009, 25, 4065-4069
4065
Nano- and Microstructure of Air/Oil/Water Interfaces† Duncan J. McGillivray,‡,§ Jitendra P. Mata,‡ John W. White,*,‡ and Johann Zank‡,| Research School of Chemistry, Australian National UniVersity, Canberra, ACT 0200, Australia, Department of Chemistry, The UniVersity of Auckland, PriVate Bag 92019, Auckland 1142, New Zealand, and Orica Mining SerVices, George Booth DriVe, Kurri Kurri, NSW 2327, Australia ReceiVed September 1, 2008. ReVised Manuscript ReceiVed October 21, 2008 We report the creation of air/oil/water interfaces with variable-thickness oil films using polyisobutylene-based (PIB) surfactants cospread with long-chain paraffinic alkanes on clean water surfaces. The resultant stable oil layers are readily measurable with simple surface techniques, exhibit physical densities the same as expected for bulk oils, and are up to ∼100 Å thick above the water surface as determined using X-ray reflectometry. This provides a ready system for studying the competition of surfactants at the oil/water interface. Results from the competition of a nonionic polyamide surfactant or an anionic sodium dodecyl sulfate with the PIB surfactant are reported. However, this smooth oil layer does not account for the total volume of spread oil nor is the increase in thickness proportional to the film compression. Brewster angle microscopy (BAM) reveals surfactant and oil structures on the scale of 1 to 10 µm at the interface. At low surface pressure (π < 24 mN m-1) large, ∼10 µm inhomogeneities are observed. Beyond a phase transition observed at π ≈ 24 mN m-1, a structure with a spongy appearance and a microscale texture develops. These structures have implications for understanding the microstructure at the oil/water interface in emulsions.
1. Introduction The oil/water interface, between aqueous and hydrophobic regions, is significant in physics, chemistry, and biology. Such interfaces are ubiquitous in industrial, food-related, and biological contexts and provide important challenges to standard theoretical models of interfaces. However, it is technically challenging to measure surface structure at oil/water interfaces reliably using standard surface techniques, and in particular, both neutron and X-ray reflectometry experiments have been hindered by the need to pass a measuring beam through a strongly attenuating liquid phase.1-4 Of particular interest in our work is the structure of the oil/ water interface in emulsion systems, which are used in industrial applications and can be related to natural emulsions such as the milk system. Polyisobutylene-based surfactants such as polyisobutylene succimide have been shown to be powerful stabilizing materials for the oil/water interface, and their use in the preparation of high internal phase (90+% internal phase) water-in-oil emulsions has been explored by us in previous papers.5-9 These surfactants have been studied at the air/water interface,9 but † Part of the Neutron Reflectivity special issue. * Corresponding author. E-mail:
[email protected]. Tel: +61 2 6125 3578. Fax: +61 2 6125 4903. ‡ Australian National University. § The University of Auckland. | Orica Mining Services.
(1) Lee, L. T.; Langevin, D.; Farnoux, B. Phys. ReV. Lett. 1991, 67, 2678– 2681. (2) Zarbakhsh, A.; Querol, A.; Bowers, J.; Yaseen, M.; Lu, J. R.; Webster, J. R. P. Langmuir 2005, 21, 11704–11709. (3) Luo, G. M.; Malkova, S.; Yoon, J.; Schultz, D. G.; Lin, B. H.; Meron, M.; Benjamin, I.; Vanysek, P.; Schlossman, M. L. Science 2006, 311, 216–218. (4) Zarbakhsh, A.; Querol, A.; Bowers, J.; Webster, J. R. P. Faraday Discuss. 2005, 129, 155–167. (5) Reynolds, P. A.; Gilbert, E. P.; White, J. W. J. Phys. Chem. B 2000, 104, 7012–7022. (6) Reynolds, P. A.; Gilbert, E. P.; White, J. W. J. Phys. Chem. B 2001, 105, 6925–6932. (7) Reynolds, P. A.; Henderson, M. J.; Holt, S. A.; White, J. W. Langmuir 2002, 18, 9153–9157. (8) Reynolds, P. A.; Henderson, M. J.; White, J. W. Colloids Surf., A 2004, 232, 55–65. (9) Reynolds, P. A.; McGillivray, D. J.; Gilbert, E. P.; Holt, S. A.; Henderson, M. J.; White, J. W. Langmuir 2003, 19, 752–761.
because they consist of a large polymeric hydrophobic tail region (ca. 1100 MW), we believe that their behavior at the oil/water interface may be significantly different (depending on the solvency of the oil for the hydrocarbon tail). Studies at the oil/water interface have previously been limited to small-angle scattering on formed emulsions;5-8 however, such systems give poor control over the composition of the interface, and systems need to have bulk emulsion stability to be measurable. We show here that these surfactants are also capable of stabilizing oil films (up to ∼100 Å thick) at the air/water interface to create an air/oil/water interface that can be readily probed using standard techniques. These air/oil/water interfaces can be formed on a Langmuir trough, opening up the system to detailed control of the surface pressure, oil, subphase, and surfactant properties, including the ability to study surfactant competition through the use of mixtures of surfactants. Using X-ray reflectometry the structure of the interface perpendicular to the surface can be studied with nanometer resolution, whereas Brewster angle microscopy gives information about lateral structures at the surface on the micrometer scale. The combination of the two methods gives a more complete picture of the interaction of the surfactant with the hydrocarbon at a planar aqueous interface and sheds light on the structures that may also form at emulsion interfaces. Attempts have been made previously to stabilize oil films at this challenging thickness using surfactants on a water surface (e.g., Kellay et al.10 using AOT as the surfactant). In that case the spreading of the oil layer was limited to short-chain alkanes, and the oil layers formed were not structurally characterized. We have also been made aware of recent work in a similar approach by Deutsch et al. producing mono- and bilayers of oil at the interface using CTAB as the stabilizing surfactant.11
2. Materials and Methods Two polyisobutylene (PIB)-based surfactants of mean molecular weight 1100 g mol-1 were used for these experiments. The first was (10) Kellay, H.; Meunier, J.; Binks, B. P. Phys. ReV. Lett. 1992, 69, 1220– 1223. (11) Sloutskin, E.; Sapir, Z.; Bain, C. D.; Lei, Q.; Wilkinson, K. M.; Tamam, L.; Deutsch, M.; Ocko, B. M. Phys. ReV. Lett. 2007, 99, 4.
10.1021/la802865z CCC: $40.75 2009 American Chemical Society Published on Web 12/02/2008
4066 Langmuir, Vol. 25, No. 7, 2009 a highly monodisperse polyisobutylenesuccinimide (PIBSI) synthesized by cationic polymerization. The second was a polydisperse polyisobutylene succinamide (PIBSA) that was derived from a commercial source. These surfactants were dissolved at a concentration of 1 mg mL-1 in redistilled toluene, which had been tested for purity by spreading on a Langmuir trough and monitoring for surface activity. To form supported oil layers, the surfactants were mixed in the spreading solution with a hydrocarbon oil (Propar 32, CAS 64742-65-0, a paraffinic mix of n-alkane chains with a boiling-point range from 400 to 600 °C). The solutions were spread onto a clean water surface on a Langmuir trough (NIMA, U.K.) in sufficient quantity to form a surfactant monolayer at the surface and were compressed to measure surface pressure versus surface area isotherms (π vs A). All water used was of ultrahigh quality, derived from a Millipore filtration system with resistivity >18 MΩ cm and no measurable surface activity. The X-ray reflectivity measurements were performed on an inhouse angle-dispersive X-ray reflectometer at the Australian National University.12 Cu KR radiation from a rotating anode source was selected using a flat graphite (002) monochromator. The alignment of the instrument was tested by measuring the reflectivity of pure water and comparing the model parameters with known values. Measurements were made at angles of incidence in the range of 0.07° < θ < 3.4°, corresponding to a QZ range of 0.1 Å < QZ < 0.48 Å-1 (where QZ ) (4π sin θ)/λ, the modulus of the perpendicular X-ray momentum transfer). The reflectivity measurements were modeled using MOTOFIT.13 Model parameters used in this program are τ, the thickness (Å); FX, the X-ray scattering length density (Å-2); and σ, the rms Gaussian interfacial roughness (Å) of a series of homogeneous layers. For these experiments, the subphase scattering length density and the interfacial roughness were fixed (at 9.43 × 10-6 Å-2 and 4 Å, respectively), leaving only FX and τ to be determined for each layer. Uncertainties in the fitted parameters have been derived from the matrix of covariance of the fit, following Marquardt minimization.13 Reflectivity data have been plotted as RQ4 against QZ to highlight deviations from Fresnel reflectivity. Brewster angle microscopy was performed on the Nanostar BAM at ISIS, Rutherford Appleton Laboratory, U.K.
3. Results 3.1. Surface Properties of a Pure Surfactants without Oil. The surface structure of pure, monodisperse PIB-based surfactants has been extensively determined previously9 and is repeated here to validate the procedures used in forming free-standing oil layers. For very large trough areas, the pure PIBSI surfactant appears to be well distributed across the interface at low surface coverage in what is classically termed the gas phase. In this phase, the surfactant molecules do not interact together or form isolated domains, and the measured surface pressure is 0.2 mN m-1. At this low density, the film cannot be resolved in the reflectivity measurement. At higher surface pressures (10 mN m-1 in Figure 1), the surfactant can be seen from modeling the X-ray reflectivity to form a hydrophilic-head/hydrophobic-tail structure at the air/water interface in which the tail stands (19.1 ( 0.5) Å above the water interface. As in our previous work,9 we found that the headgroup was too small to resolve its thickness and density-independently; consequently, we have adopted the approach of that paper and fixed the headgroup thickness in all modeling to a physically reasonable 5 Å while fitting its density. Models that do not include this thin headgroup are unable to account for the X-ray data satisfactorily. For comparison, the surface film formed from the polydisperse PIBSA surfactant is shown in Figure 1b, measured at two compressions. A similar structure of head and tail is observed (12) Brown, A. S.; Holt, S. A.; Saville, P. M.; White, J. W. Aust. J. Phys. 1997, 50, 391–405. (13) Nelson, A. J. Appl. Crystallogr. 2006, 39, 273–276.
McGilliVray et al.
Figure 1. (a) X-ray reflectivity (displayed as RQ4 vs QZ) from monodisperse PIBSI surfactant at the air/water interface at two surface pressures (O, π ) 0.2 mN m-1; b, π ) 10 mN m-1), with modeled reflectivity (-). (b) The reflectivity from the polydisperse polyisobutylene amide at two surface pressures (O, π ) 3.7 mN m-1; b, π ) 25 mN m-1), with modeled reflectivity (-). The inset of each graph shows the X-ray scattering length density profiles as a function of distance from the water interface corresponding to the fitted model lines in each main graph.
at the interface, the only difference being that the surfactant tail layer is slightly thicker for the higher-polydispersity polymer surfactant. On compression of the surface film of the PIBSA, the fringe structure seen in Figure 1b moves to higher Q as the surfactant tail region becomes thicker, and the surfactant tails are confined to a smaller area/chain resulting in more upright conformations. It is noteworthy that some of the polymer is displaced from the surface on compression; in this case, 23% of the surfactant is displaced in going from π ) 3.7 to 25 mN m-1. The surfactant is very insoluble in water, and we do not believe that any significant portion is lost to the aqueous subphase. In our previous work, this loss of material was attributed to the formation of large-scale structures that are not visible on the reflectometry length scale.9 3.2. Forming an Air/Oil/Water Interface. Figure 2 shows the reflectivity from a surface film derived from spreading a mixture of the polydisperse PIBSA polymer surfactant mixed with an equal mass of paraffinic oil (Propar 32) on the water surface. The very different reflectivity from those of Figure 1 is obvious, and model refinement shows that the fringe pattern is well described by a thick single layer ((50 ( 2) Å thick at the lowest compression and rising to (60.0 ( 0.3) Å above the headgroup, with a scattering length density that is less than that of water ((8.4 ( 0.1) × 10-6 Å-2), dropping to (8.0 ( 0.1) × 10-6 Å-2 on compression). Increasing the model complexity to allow for the possibility of either surfactant residing at the hydrophobic/hydrophobic oil/air interface or an inhomogeneous oil layer did not improve the quality of the modeling. Assuming that the layer consists solely of alkanes from the oil, this corresponds to a physical density of (0.85 ( 0.01) g cm-3 compared to an expected 0.87 g cm-3 for the bulk oil at 15 °C. This thick oil layer is observed even when the area per molecule of the surfactant is too low for the surfactant’s headgroup to be
Nano- and Microstructure of Air/Oil/Water Interfaces
Figure 2. X-ray reflectivity (displayed as RQ4 vs QZ) from a monolayer formed from the spreading polydisperse PIBSA surfactant with 1 mass equiv of paraffinic oil at two surface pressures (b, π ) 0 mN m-1; O, π ) 26 mN m-1). The lines represent the modeled X-ray reflectivities fit to each set of data. (Inset) X-ray scattering length density profiles as a function of the distance from the water interface derived from the modeled fit lines, with an oil layer of ∼55 Å shaded.
Figure 3. X-ray reflectivity (displayed as RQ4 vs QZ) from the measurement of a surface film formed from spreading a mixture of polydisperse PIBSA polymer surfactant with 2 mass equiv of paraffinic oil at a surface pressure of 12 mN m-1 (b) with the modeled reflectivity (-). (Inset) X-ray scattering length density profile as a function of the distance from the water interface derived from the modeled reflectivity line in the main graph.
resolved, highlighting the surfactant’s ability to stabilize the oil layer. From the fact that the oil density measured in this layer is close to the expected bulk density of the oil, we can estimate that there are few defects in this oil layer. It can also be determined from the measured density profiles that the reduction in the area of the surface film due to compression is not matched by a proportional increase in the thickness of the film; that is, the volume of oil measured at the interface is not constant. Quantitative measurements indicate that about 54% of the oil that was visible in the low-pressure measurement is no longer visible in the X-ray reflectivity at π ) 25 mN m-1. This is evidence of the formation of large-scale interfacial structures above the water surface that are too rough to be resolved by specular reflectivity measurements (similar to the case for the surfactant alone). On re-expansion of the surface film, the oil layer returns to the previous measured thickness, showing that none of the oil is permanently lost to the system (e.g., through irreversible aggregation). The data of Figure 3 represents the approximate maximum of this layer thickness that we have measured so far. This thickness of the oil layer ((91.7 ( 0.3) Å above the surfactant headgroup) was formed through a mixture of surfactant/oil (1:2 by mass) and has a physical density of (0.85 ( 0.01) g cm-3, which is still nearly equivalent to the expected density of the bulk oil. As above, the oil layer is well described as being homogeneous; modeling allowing the oil layer to consist of sublayers does not
Langmuir, Vol. 25, No. 7, 2009 4067
Figure 4. X-ray reflectivity (displayed as RQ4 vs QZ) from a surface film formed from spreading a mixture of monodisperse polymer surfactant/ hexadecane (1:6 by mass). The data are as spread (b, A ) 400 cm2) and after compression (O, 2 h after spreading, A ) 180 cm2). The lines represent the fitted model reflectivities. (Inset) Modeled real-space X-ray scattering length density profile as a function of the distance from the water interface derived from the fitted model lines in the main graph.
provide a significant improvement in the fit quality. It is not possible using X-ray reflectometry to distinguish the oil-solvated polymeric chains of the stabilizing surfactant from the surrounding hydrocarbon oil because of their low X-ray scattering contrast. Thus, the extent of the change in tail conformation cannot be determined. However, later experiments using deuterated oils and surfactants, measured using neutron reflectometry, will enable these two components to be individually resolved. A counterpoint to this oil-film stabilization can be found by using pure hexadecane to form the oil film because it does not produce a stable oil film. Using a 1:2 (mass of surfactant/mass of hexadecane) mixture of surfactant and hexadecane, no visible oil layer is detected (the surface film is does not differ from that of a pure surfactant film (not shown)). When the amount of oil in the mixture was increased so that the solution spread contains 1:6 mass of surfactant/mass of hexadecane, the reflectivity in Figure 4 was initially measured, indicating the presence of an oil layer (60 ( 2) Å thick. The hexadecane oil layer was less dense than the Propar 32 layer, but again very near the density expected for the bulk oil (the scattering length density is (7.5 ( 0.2) × 10-6 Å-2, which gives a physical density of (0.77 ( 0.02) g cm-3, compared to 0.78 g cm-3 expected for bulk hexadecane). The quality of this fit is lower than for other data sets, where over a period of approximately 2 h the signal from the oil layer disappeared, and the X-ray reflectivity is measured in a pointby-point fashion across increasing QZ. Following the loss of signal, even with later film compression only the surfactant monolayer was detectable. This temporal instability could be due to the slow formation of large-scale structures similar to those observed using the paraffinic oil; however, the slow, complete, irreversible nature of the loss of oil suggests that other mechanisms are at play here. Perhaps most significant is the low, but measurable, vapor pressure of hexadecane,14 causing the loss of oil to evaporation. This instability limited our investigations using hexadecane as the oil layer. 3.3. Competition between Surfactants at the Oil/Water Interface. The competition between different surfactants at the oil/water interface has been studied using an anionic surfactant (sodium dodecyl sulfate, SDS) and a nonionic polyacrylamidebased surfactant (PAM, a polyacrylamide heptamer hydrophilic headgroup with a C12-alkane hydrophobic tail, total MW ∼780 g mol-1) to compete with the polyisobutylene based surfactant films. Both the SDS and the PAM surfactant are water-soluble, (14) Parks, G. S.; Moore, G. E. J. Chem. Phys. 1949, 17, 1151–1153.
4068 Langmuir, Vol. 25, No. 7, 2009
McGilliVray et al.
Figure 5. X-ray reflectivity (displayed as RQ4 vs QZ) measured from a monolayer of a mixture of polydisperse PIB surfactant/paraffinic oil (1:1 by mass) spread on an aqueous solution containing a polyamidebased water-soluble surfactant introduced into the subphase. The lines represent the modeled reflectivity fits to the data. The measurements were made at varying compressions (-b-, A ) 510 cm2, π ) 37 mN m-1; · · · O · · · , A ) 210 cm2, π ) 41 mN m-1) and after rapid expansion of the monolayer (- · -4- · -, A ) 510 cm2, π ) 24 mN m-1). (Inset) Modeled X-ray scattering length density profile as a function of the distance from the water interface from which the model fit lines in the main graph are derived.
Figure 6. Surface pressure vs trough area isotherms for a mixture of polydisperse PIBSA and paraffinic oil (1:1 by mass) spread from toluene at 25 °C on water. (a) Representative Brewster angle microscopy (BAM) image of the surface at π ) 1 mN m-1 showing large (∼10 µm) features. (b) Representative BAM image at π ) 24 mN m-1 measured just after the transition detected at 23 mN m-1, showing the much finer (1 to 2 µm) spongelike features. Each scale bar is 100 µm long.
in contrast to the strictly oil-soluble PIB-based surfactant, and were therefore introduced to the oil/water interface by injection into the subphase. The polyacrylamide-based surfactant showed strong competition for the air/water interface with the polyisobutylene oil layer. After the addition of PAM to the aqueous subphase of the trough at the largest area per molecule of the PIB surfactant, the surface pressure immediately increased from ∼0 to 37 mN m-1, and at the same time, the thin oil layer was lost from the interface (Figure 5). The modeled real-space density profile derived from X-ray reflectometry shows the appearance of a monolayer of the PAM surfactant at the interfacesa large headgroup region in the aqueous subphase with a thickness of (18.9 ( 0.5) Å, accounting for the polyacrylamide portion of the surfactant, and the thin C12 hydrophobic tail above the interface of (11.5 ( 0.2) Å (similar to the model to reflectivity measured from a PAM monolayer by itself (not shown)). Neither the oil layer nor the PIB surfactant itself can be detected at the interface. Upon compression of the surface monolayer from an area of 510 to 210 cm2, little change is observed, with the thickness and density of the PAM monolayer essentially unchanged. This is typical behavior for a water-soluble surfactant that can be readily displaced from the interface into the bulk aqueous phase on reduction of the surface area available for each molecule. Interestingly, on rapid (minutes) expansion of the surface film back to 510 cm2 (∼ 2.5× larger area) the stabilized oil layer is again visible on the surface, at least over the period required for a reflectometry measurement (∼1 h), and the surface pressure drops to only 24 mN m-1. From the modeled density profile, we can also determine that the PAM surfactant is forming less than a complete monolayer, at approximately 25% monolayer coverage. On standing, the surface pressure slowly returns to 37 mN m-1, and a complete PAM monolayer is reformed with the consequent displacement of the oil layer. The implication is that whereas the PAM is more surface active than the PIB surfactant that stabilizes the oil layer the rate of its adsorption to the surface is significantly slower than the rate of respreading of the oil layer. Thus, on rapid formation of the fresh surface on the trough, through expansion of the film, the deaggregation of the PIB-oil structures forms a supported oil layer before being slowly redisplaced by the PAM. A similar experiment was performed using the anionic surfactant sodium dodecyl sulfate (SDS) as the cosurfactant
competing for the oil/water interface. In this case, the surface pressure rises only slightly on addition of the SDS (from 0.4 to 2.9 mN m-1), and a thick oil layer remains detectable and unchanged at the interface (not shown). The measured surface film has the same structural properties as that formed from the PIB surfactant alone, including reversible loss of material on monolayer compression. Because the headgroups of the SDS cannot be distinguished from that of the PIB surfactant (unlike the PAM case) and the tail groups are indistinguishable from the oil layer overlaying the surface, it is not possible to determine clearly whether the SDS is adsorbed to the surface or the PIB monolayer remains unaffected. Again, neutron reflectometry will reveal the components of this mixture. 3.4. Air/Oil Interface at Micrometer Resolution. The evidence from the reflectometry measurements is that material is being lost from the interface on surface compression of the oil-surfactant mixture. Similar behavior for surfactant alone has previously been attributed to the formation of large-scale rough structures.9 Brewster angle microscopy provides a method to visualize the formation of any large-scale structures that appear at micrometer resolution. Images were taken from a surface film formed from the spreading of a mixture in toluene of the polydisperse PIB surfactant with an equal mass of the paraffinic oil (an equivalent system to that represented in Figure 2) compressed to various pressures (the measured π-A isotherm is shown in Figure 6). This system was found from the X-ray reflectometry discussed above to consist of an ∼55 Å supported oil layer, of which ∼50% is lost on compression to structures not giving rise to specular reflectivity. The surface textures were photographed at various points on the surface pressure-area isotherm, and examples are shown in the insets of Figure 6. At all pressures, large-scale structures were found at the interface. At the lowest pressures, these consist of large (∼10 µm) pools of oil, initially scattered randomly across the surface but becoming more apparent as the surface pressure increases (Figure 6a). At 23 mN m-1, the isotherm shows a phase transition that corresponds to a dramatic change in the structures observed at the interface. As can be seen from the change between the images in Figure 6a and b, the surface texture becomes much finer (approximately 1 to 2 µm features) and spongelike. We interpret this change as representing the formation of a continuous surfactant/oil complex above the smooth oil layer measured by X-ray reflectometry. This structure should give rise to significant off-specular
Nano- and Microstructure of Air/Oil/Water Interfaces
scattering, which will be able to give further information on the nature of these structures.
4. Discussion The X-ray reflectometry measurements show that cospreading a mixture of polyisobutylene-derived surfactant with a nonvolatile oil (a paraffinic oil, Propar 32) at the air-water interface causes the formation of a thin (
