Article pubs.acs.org/Langmuir
Emulsion/Surface Interactions from Quiescent Quartz Crystal Microbalance Measurements with an Inverted Sensor Roozbeh Mafi and Robert H. Pelton* Department of Chemical Engineering, McMaster University, Hamilton, Ontario L8S 4L7, Canada S Supporting Information *
ABSTRACT: Interactions of three oil-in-water emulsion types with polystyrenecoated quartz crystal microbalance (QCM) sensor surfaces were probed with the QCM cell in both the conventional orientation (i.e., polystyrene surface on the bottom, “looking up”) and the inverted orientation (polystyrene on top interior surface of sensor chamber, “looking down”). With the conventionally oriented QCM sensors, the adsorption of soluble and/or dispersed species quickly gave steady-state frequency and dissipation outputs. By contrast, the inverted sensors gave changing responses at long times because of the gravity driven buildup of a viscous consolidation layer next to but not necessarily bound to the sensor surface. Three emulsion types (a simple hexadecane/phosphatidylcholine emulsion, 2% homogenized milk, and a diluted commercial ophthalmic emulsion) displayed a wide range of behaviors. We propose that quiescent QCM measurement made with an inverted sample chamber is a new approach to probing emulsion behaviors near solid surfaces.
■
INTRODUCTION With the advent of robust commercial instrumentation, quartz crystal microbalance (QCM) measurements have been applied to a wide range of systems leading to many publications and a few textbooks.1 Nevertheless, we have found only a few references to studies employing QCM measurements to characterize the interactions of emulsions with surfaces, whereas there have been many studies of vesicle deposition and bilayer formation on QCM sensor surfaces.2,3 The lack of emulsion studies is surprising since water-in-oil emulsions are important for food and formulated chemical products. Herein we report a new way of using QCM to probe the interactions of emulsions with surfaces, based on comparing results from conventional QCM methods to results obtained when the QCM sensor chamber is inverted. Whereas conventional QCM measurements, with the sensor surface facing up, probe adsorption of emulsion components onto QCM sensor surfaces, inverted (i.e., sensor facing down) measurements of stagnant samples also detect creaming and the gravity-driven formation of a consolidation layer next to the sensor surfaces. The few published QCM emulsion studies we have found include the following. Stalgren and co-workers4 compared the behaviors of liposomes and phospholipid stabilized emulsions. They reported a slow spreading of adsorbed emulsion droplets, causing the displacement of some of the initially adsorbed material. Yang et al. used QCM to show that lecithin-stabilized emulsion droplets adsorb onto chitosan-coated QCM sensors.5 Finally, Kallio et al. employed QCM to probe the adsorption potential of materials typically found in papermaking suspensions, including wood pitch emulsions.6 As part of an ongoing investigation of the physical chemistry of ophthalmic emulsions used to treat dry eye syndrome,7 we © 2015 American Chemical Society
were interested in characterizing the interactions of the ophthalmic emulsions with solid surfaces. For this we employed two emulsions: SYSTANE BALANCE, a commercial ophthalmic emulsion, and model hexadecane emulsions stabilized with L-α- phosphatidylcholine, typically with average diameters of about 200 nm. When the QCM-D showed promise as a technique for emulsions, we expanded our study to include 2% homogenized milk, an emulsion displaying different behaviors in our QCM-D studies.
■
EXPERIMENTAL SECTION
Materials. Hexadecane, NaN3, (4-(2-hydroxyethyl)-1-piperazineethanesulfonic acid) (HEPES), and NaCl were purchased from SigmaAldrich (Oakville, ON, Canada); l-α-phosphatidylcholine was purchased from Avanti Polar Lipid (Alabaster, AL). SYSTANE BALANCE (Alcon) and 2% homogenized milk were purchased at a grocery store. Aqueous solutions were made with 18.2 MΩ·cm Barnstead Nanopure Diamond system (Dubuque, IA) water. Measurements were performed with an E1 QCM-D (Q-Sense AB, Gothenburg, Sweden) and monitored using Qsoft401 software version 2.5.2. All measurements were repeated at least twice. Processed data from the fifth overtone using Qtools software version 3.0.7 are shown herein. Emulsions. In a typical experiment, 10 mg of a 25 mg/mL solution of L-α-phosphatidylcholine in chloroform was poured into a roundbottom flask. Removal of the solvent was done using a rotary evaporator at room temperature and exposure to a stream of nitrogen gas. This resulted in a dried film. 140 mg of hexadecane was then added to the flask, followed by a 10 mL solution, which contained 5 ppm of NaN3, 10 mM HEPES buffer, and 5 mM NaCl. The mixture Received: April 6, 2015 Revised: June 16, 2015 Published: June 17, 2015 7238
DOI: 10.1021/acs.langmuir.5b01904 Langmuir 2015, 31, 7238−7241
Article
Langmuir was ultrasonicated (90% amplitude and output power setting of 60) for 30 min using a Misonix sonicator (S-4000 Ultrasonic Processors, USA, 20 kHz) with the ultrasonic probe immersed directly into the emulsion for 15 min.
Sauerbrey analysis is not correct. We have included some viscoelastic modeling results in the Discussion section. In summary, we conclude that in the initial stages with flow emulsion droplets deposited on the polystyrene sensors in both orientations, and subsequent spreading of adsorbed droplets is possible. At times beyond 20 h the inverted cells gave a significant further drop in frequency and a corresponding increase in dissipation, not seen in the conventionally oriented sensor. The detailed shapes of the curves beyond 20 h were not reproducible in detail; however, there was always a second frequency decrease after a long delay time. We performed a series of similar experiments where the hexadecane volume fraction was varied from 0.25 to 2%. From these results we extracted Δf I, the frequency change from when the flow was stopped (indicated by ×’s in Figure 1), to the end of the experiments. Δf I values are plotted as functions of the hexadecane volume fractions in Figure 2. The frequency increased with emulsion volume fraction. These behaviors were only observed with the inverted sample chamber orientation.
■
RESULTS The interactions of dilute hexadecane emulsion (205 nm diameter, 0.15 PDI) with polystyrene surfaces were monitored by QCM-D. Figure 1 shows two sets of results corresponding
Figure 1. Hexadecane emulsion (2% v/v) interaction with polystyrene QCM-D sensor. For “inverted experiments”, the QCM-D sensor surface was at the top of the flow chamber, facing down, whereas in the conventional orientation, the sensor surface was on the bottom, facing up. The ×’s indicate the time at which flow was stopped.
Figure 2. Frequency changes for hexadecane emulsions as functions of the emulsion concentration. Δf I is the frequency change from when the flow was stopped (indicated by ×’s in Figure 1) to the end of the experiments.
to conventional and inverted sample QCM-D chamber orientations. The conventional curves were obtained with QCM-D sensor surface oriented in the conventional manner with the polystyrene surface at the bottom of the flow chamber. By contrast, the inverted experiments were conducted with the sensor holder flipped upside down, putting the polystyrene surface at the top of the flow chamber. In both experiments the emulsion was initially pumped through the QCM-D chamber; however, the flow was stopped at the times marked by ×’s in the figure. Considering first the frequency change results in Figure 1, both orientations gave an initial frequency decrease of about −60 Hz with −ΔD/Δf n ∼ 6 × 10−7 s. The corresponding dissipation versus frequency change plots were approximately linear and are shown in Figure S1 of the Supporting Information. If we assume that the corresponding Sauerbrey mass is correct, this frequency change corresponds to a surface coverage of about 10% by 205 nm diameter hexadecane spheres. The Sauerbrey model is valid for surface layers that do not deform during the measurement. Dissipation results are an indicator of deformation, and Q-Sense claimed that the Sauerbrey model is valid when −ΔD/Δf n values are less than 1 × 10−7 Hz−1.8 Also, the results in Figure S2 (Supporting Information) show the other overtone results for the frequency data corresponding to Figure 1. The fifth and seventh overtone results were close, whereas the third overtone results gave larger changes. These observations suggest that the application of the
We propose that the inverted curves at long times depict the gravity driven formation of a consolidation layer next to the sensor surface. Simple Stokes law estimates suggest that emulsion droplets should concentrate near the surface at time scales much less than 20 h. Our QCM-D sensors are sensitive to liquid properties up to about a micrometer from the crystal surface (page 11 in Johannsmann’s book1). Furthermore, in the QCM models liquid phase properties appear as the square root of the product of solution density and viscosity.9 The density of an emulsion is a linearly decreasing function of the oil volume fraction. For dilute emulsions, the viscosity is a linearly increasing function of volume fraction (the Einstein regime). Therefore, at early stages of creaming, the density × viscosity product is approximately constant. But viscosity will increase exponentially with oil volume fraction at longer times when the volume fraction extends beyond the Einstein regime.10 In summary, we propose that the long delay time required for the second transition in the inverted cell reflects the time required for the viscosity term to dominate the density × viscosity product. Figure 3 shows a similar set of experiments employing SYSTANE BALANCE lubricant eye drops,11 a complex formulated mineral oil product containing boric acid, dimyristoylphosphatidylglycerol, edetate disodium, hydroxy7239
DOI: 10.1021/acs.langmuir.5b01904 Langmuir 2015, 31, 7238−7241
Article
Langmuir
Figure 4. Polystyrene sensor was conditioned with milk serum and then exposed to 2% homogenized milk. Figure 3. Interaction of SYSTANE BALANCE emulsion (1 part diluted with 2 parts buffer) with polystyrene.
conventional orientation. As before, we propose that these differences reflect the properties of the consolidation layer next to the inverted cell surface.
■
propyl guar, mineral oil, polyoxyl 40 stearate, POLYQUAD (polyquaternium-1) 0.001% preservative, sorbitan tristearate, sorbitol, and purified water. According to our dynamic light scattering measurement for a sample diluted by half in 10 mM HEPES + 5 mM NaCl, the average particle size of the emulsion is 320 nm. Several of the product components, in addition to the emulsion droplets, have the potential to adsorb on polystyrene. For example, we have shown that hydroxypropyl guar adsorbs onto polystyrene.12 Similarly, polyoxyl 40 stearate has the potential to adsorb onto polystyrene in water. The frequency (Δf) results for SYSTANE BALANCE in Figure 3 show very little difference between the inverted and normal orientation, suggesting that the adsorbed surface layers may be similar in the two cases. By contrast, the inverted dissipation results in Figure 3 were substantially higher than for the conventionally oriented sensor, possibly indicating a contribution from a thin consolidating layer next to the sensor surface. For a final example we chose 2% homogenized milk, a particularly complex mixture of emulsified fat, casein micelles, and lipoprotein particles, as well as a mixture of soluble components. In these experiments we preconditioned the QCM-D sensor with milk serum prepared by centrifuging 2% milk at 35000 rpm (115000g) for 30 min. Under these conditions we would expect the sensor surfaces to be saturated with adsorbed casein before the emulsion was introduced.13 The results in Figure 4 show positive frequency changes for both cell orientations. Although phenomena such as sliding of the adsorbed layer on the sensor surface can give positive changes, the most likely explanation is the following. We speculate that the introduction of the milk caused the displacement of adsorbed components from the serum with a lighter but more strongly adsorbed material. With milk, like the other two emulsions, inverting the sensor gave substantially higher dissipation values compared to the
DISCUSSION Comparison of inverted versus conventional QCM-D measurements with emulsions reveals a rich range of behaviors. With oil-in-water emulsions, we propose that the conventional orientation results, with the sensor surface “looking up” into the solution, give information about adsorbing surfactant and/ or emulsion drops. When flow is stopped, gravity will drive the oil droplets away from the sensor surface, limiting subsequent adsorption. By contrast, with the inverted measurements, where the sensor surface is “looking down” into the solution, emulsion drops will rise toward the surface with the cessation of flow, either depositing or accumulating next to the sensor surface. The resulting frequency and dissipation changes could either be due to changes in the adsorbed layer or to changes in solution properties next to the sensor surface. Using Q-Sense software, we fitted a Voigt model to a set of hexadecane emulsion data, and the results are summarized in Figure 5. The third, fifth, and seventh overtones were selected for modeling because frequency-dependent behavior has been observed. Fixed parameters used for the model were the bulk density (1000 kg m−3) and bulk viscosity (0.001 Pa·s). The density of the adsorbed layer was assumed to be constant (900 kg m−3), and the ranges of the other layer properties were restricted to adsorbed layer viscosity (0.001−100 Pa·s), shear modulus (0.1−10000 Pa), and thickness (10−9−10−6 m). The experimental data were from a 2% (v/v) hexadecane emulsion. The model predicts a significant increase in the adsorbed layer shear modulus starting about 25 h after stopping flow. The model assumes that all changes are occurring in the adsorbed layer, whereas we believe the layer properties could be constant, with the changes occurring in the concentrated emulsion layer next to the surface. Johannsmann in his discussion of biological cells on QCM surfaces states “Viscoelastic modeling from first principles in the author’s 7240
DOI: 10.1021/acs.langmuir.5b01904 Langmuir 2015, 31, 7238−7241
Langmuir
Article
■
ACKNOWLEDGMENTS We thank the National Sciences and Engineering Research Council of Canada (NSERC) and Alcon Laboratories for funding this project. Some measurements were performed in the McMaster Biointerfaces Institute funded by the Canadian Foundation for Innovation. R.P. holds the Canada Research Chair in Interfacial Technologies
■
Figure 5. Shear modulus of the QCM-D sensor surface layer estimated with the Q-Sense Voigt-based extended viscoelastic modeling (Q-tools 3.0.12 software).
view is hopeless” (page 306 in ref 1)the same situation likely holds for modeling emulsion/QCM surface interactions. Our three emulsion types showed a broad range of behaviors. The very stable SYSTANE BALANCE (Figure 3) emulsion showed little orientation dependence in the frequency change, whereas the dissipation was somewhat higher for the inverted measurement. By contrast, the hexadecane showed orientationdependent changes of both frequency and dissipation at long times. For the milk experiments (Figure 4), the polystyrene sensor surface was first equilibrated with milk serum (ultracentrifuged milk) presumably coating the polystyrene with adsorbed casein. Upon exposure to milk, the dissipation increased and the frequency change was positive. The common feature for the three emulsion types was the higher dissipation values obtained with the inverted sensor configuration, which we believe reflects the gravity-induced formation of a consolidation layer near, but not necessarily adhering to the sensor surface.
■
CONCLUSIONS In conclusion, emulsion technologists have many excellent tools to analyze their products, including rheological measurements, accelerated stability testing, particle sizing, and automated creaming rate measurements. We propose that the combination of conventional and inverted quiescent QCM-D measurements is unique in terms of probing interactions on or near (up to ∼250 nm) solid surfaces.
■
REFERENCES
(1) Johannsmann, D. The Quartz Crystal Microbalance in Soft Matter Research; Springer International Publishing: Berlin, 2015. (2) Reimhult, E.; Höök, F.; Kasemo, B. Intact Vesicle Adsorption and Supported Biomembrane Formation from Vesicles in Solution: Influence of Surface Chemistry, Vesicle Size, Temperature, and Osmotic Pressure. Langmuir 2003, 19, 1681−1691. (3) Keller, C. A.; Glasmästar, K.; Zhdanov, V. P.; Kasemo, B. Formation of Supported Membranes from Vesicles. Phys. Rev. Lett. 2000, 84, 5443−5446. (4) Stalgren, J. J. R.; Claesson, P. M.; Warnheim, T. Adsorption of Liposomes and Emulsions Studied with a Quartz Crystal Microbalance. Adv. Colloid Interface Sci. 2001, 89, 383−394. (5) Yang, X.; Tian, H.; Ho, C.-T.; Huang, Q. Stability of Citral in Emulsions Coated with Cationic Biopolymer Layers. J. Agric. Food Chem. 2011, 60, 402−409. (6) Kallio, T.; Kekkonen, J.; Stenius, P. The Formation of Deposits on Polymer Surfaces in Paper Machine Wet End. J. Adhes. 2004, 80, 933−969. (7) Mafi, R.; Gray, C.; Pelton, R.; Ketelson, H.; Davis, J. On Formulating Ophthalmic Emulsions. Colloids Surf., B 2014, 122, 7−11. (8) Quevedo, I. R.; Olsson, A. L. J.; Tufenkji, N. Deposition Kinetics of Quantum Dots and Polystyrene Latex Nanoparticles onto Alumina: Role of Water Chemistry and Particle Coating. Environ. Sci. Technol. 2013, 47, 2212−2220. (9) Voinova, M. V.; Jonson, M.; Kasemo, B. ‘Missing Mass’ Effect in Biosensor’s Qcm Applications. Biosens. Bioelectron. 2002, 17, 835−841. (10) Krieger, I.; Dougherty, T. A Mechanism for Non-Newtonian Flow in Suspensions of Rigid Spheres. J. Rheol. 1959, 3, 137. (11) Alcon Systane® Products; systane.com (accessed April 5, 2015). (12) Zhang, L.; Pelton, R.; Ketelson, H.; Meadows, D. Charge Regulation Enables Anionic Hydroxypropyl Guar-Borate Adsorption onto Anionic and Cationic Polystyrene Latex. J. Colloid Interface Sci. 2011, 353, 557−561. (13) Reimhult, K.; Petersson, K.; Krozer, A. Qcm-D Analysis of the Performance of Blocking Agents on Gold and Polystyrene Surfaces. Langmuir 2008, 24, 8695−8700.
ASSOCIATED CONTENT
S Supporting Information *
ΔD/Δf n plots for the data in Figures 1 and 3 as well as a comparison of overtone results for the experiments in Figure 1. The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acs.langmuir.5b01904.
■
AUTHOR INFORMATION
Corresponding Author
*E-mail:
[email protected] (R.H.P.). Notes
The authors declare no competing financial interest. 7241
DOI: 10.1021/acs.langmuir.5b01904 Langmuir 2015, 31, 7238−7241
