Nanoemulsions obtained via bubble-bursting at a compound interface

Transcription

1 Nanoemulsions obtained via bubble-bursting at a compound interface Jie Feng 1, Matthieu Roché 1, #, Daniele Vigolo 1, $, Luben N. Arnaudov 2, Simeon D. Stoyanov 2,3, Theodor D. Gurkov 4, Gichka G. Tsutsumanova 5 and Howard A. Stone 1 1. Department of Mechanical and Aerospace Engineering, Princeton University, Princeton, New Jersey, 08544, USA 2. Unilever Research and Development, 3133AT Vlaardingen, The Netherlands 3. Laboratory of Physical Chemistry and Colloid Science, Wageningen University, 6703 HB Wageningen, The Netherlands; Department of Mechanical Engineering, University College London, Torrington Place, London WC1E 7JE, UK 4. Department of Chemical Engineering, Faculty of Chemistry & Pharmacy, University of Sofia, James Bourchier Avenue 1, Sofia 1164, Bulgaria 5. Department of Solid State Physics & Microelectronics, Faculty of Physics, University of Sofia, James Bourchier Avenue 5, Sofia 1164, Bulgaria Control experiments Control experiments allowed us to show that surfactants in the water phase, the presence of the oil layer, and bubble bursting are necessary to the dispersal of stable submicrometre objects in the water column. There are no detectable submicrometre droplets after bubbling without surfactant or a very low surfactant concentration (e.g. using a 1-mm-thick oil layer with [C 12 TAB] = 0.02 mm). The addition of more surfactants into water leads to the dispersal of the submicrometre droplets after bubbling as listed in Table S1. # Present address: Laboratoire de Physique des Solides, Université Paris Sud-CNRS, Orsay, France $ Present address: Department of Chemistry and Applied Biosciences, Institute for Chemical and Bioengineering, ETH Zurich, 8093 Zurich, Switzerland Corresponding author: hastone@princeton.edu NATURE PHYSICS 1 1

2 Relationship between the wetting state and dispersal Here, we describe a set of experiments involving a 1-mm-thick layer of hexadecane and a solution of the surfactant dodecyltrimethylammonium bromide (C 12 TAB) over a range of surfactant concentrations [C 12 TAB]. Hexadecane switches from a partial-wetting state to a pseudo-partial-wetting state on an aqueous solution of C 12 TAB when [C 12 TAB] mm 1-2. These experiments helped us to uncover the importance of molecular-scale oil-surfactant interactions in setting the properties of the dispersed oil droplets, thus establishing the connection between wetting and dispersal. We were able to distinguish three situations. In the first case, when [C 12 TAB] = 0.02 mm << mm, hexadecane is in a partial-wetting state on water, and no dispersal occurs. In the second case, an increase of [C 12 TAB] to 0.09 mm, for which, according to the literature value, the system is supposedly in a partial-wetting state at equilibrium, leads to the observation of dispersal. Finally, dispersal always occurred when [C 12 TAB] mm. The observation reported in the second above case indicates that not only the equilibrium wetting state is important, but also the wetting transition induced by surfactant dynamics at the alkane/water interface. Indeed, the collapse of the bubble in our experiments leads to a compression of the surfactant monolayer 3 sitting at the interface between hexadecane and water, thus changing the local surfactant interfacial concentration. We think that this dynamic change in interfacial concentration triggers a wetting transition from the equilibrium partial-wetting state to a pseudo-partial wetting state usually observed at higher equilibrium surfactant concentrations. 2 2 NATURE PHYSICS

3 SUPPLEMENTARY INFORMATION Ellipsometry measurements Ellipsometry experiments were carried out using hexadecane and an aqueous solution with [C 16 TAB] = 0.9 mm which corresponds to the pseudo-partial wetting state. The data show that the deposition of a droplet of pure hexadecane on the interface of the surfactant solution (not very far from the laser spot) leads to oil spreading, which forms an initially rather thick film. After the spreading, in a matter of ~1 minute, we observed that the film thins and breaks down. The process is illustrated in Figs. S1 and S2. When the initially spread film ruptures, the oil forms a very thin precursor layer whose thickness is O(1.5-2 nm), which is in coexistence with submicron lenses with a thickness of the order of ~110 nm (some macroscopic lenses are also embedded in this film). Figure S3 provides a visualization of the oil layer in terms of the ellipsometric angles, and. In the region of coexistence of precursor layer and liquid lenses, and represent averaged values, but the distribution of the experimental points is characterized by large deviations in and almost constant. A plausible explanation of this pattern is to have a predominant lens thickness, h, around 115 nm (Fig. S3), with some distribution in the values of h. Then, the changes of between 5.2 and 9 degrees (at const.) would correspond to different proportions of lenses and precursor film. At higher surfactant concentration ([C 16 TAB] = 1.8 mm), the processes of layer evolution described above happened significantly faster and we could not follow it with our experimental protocol for the ellipsometry Surface pressure measurements Figure S4 shows data for the surface pressure, s, jump upon addition of oil to the interface between the air and the C 16 TAB solution. The initial non-equilibrium spreading is manifested as 3 NATURE PHYSICS 3

4 63 64 a sudden increase of reaches saturation with a plateau value of s ; when the lenses/precursor layer coexistence is established, the surface s The estimation for the shape of the lens To estimate the shape of the lens, as shown in Fig. S5, we assume the oil lens at the interface has a shape of a spherical cap and the oil-water interface is flat. The thickness of the lens is h while the lateral size is denoted by λ. Using geometry we can show that the radius for the mean curvature of the cap is given by 70 R h R h R, (1) 2h 2 71 Hence, the mean curvature of the cap R 2, which is used in equation (1) in the manuscript, is 72 4h 2 h 2. For the volume of the lens V, we have 73 V lens R r sin d d dr R. (2) 3 74 Since 2 2 R cos 1 as shown in Fig. S5 and R 2 h R 2h 2, we finally obtain 75 V lens h 2 (3 h 6 2 ), (3) 76 which is used for the estimation of the volume of the submicron oil droplets Data to check the trend of the droplet size with different alkanes The data of surface tension oa, Hamaker constant A and the thickness of the nanometric film for different alkanes from the literature are listed in Table S NATURE PHYSICS

5 SUPPLEMENTARY INFORMATION Calculation of energy efficiency Classic high-shear-rate methods require significant energy input, and hence these processes are typically inefficient. Only around 0.1% of the energy input is used for emulsification 4. Our system has a higher energy efficiency, which is estimated as follows. The input energy E i is the work done by the pump to produce bubbles into the system, while the output energy E o is the energy stored at the newly created oil/water interface of the droplets in water, assuming the system is working in the atmosphere at room temperature. E i can be expressed as E i = (P w V b + γ aw S b ) n b /ε c (4) where P w is the hydrostatic pressure, V b is the volume of the bubble, γ aw is the interfacial tension between air and the surfactant solution, S b is the surface area of the bubble, n b is the number of the generated bubbles, and ε c is the pump efficiency. In a typical experiment, bubbles were produced for 2 days at a frequency on the order of 1 Hz. Thus, n b bubbles. The efficiency of air compressors is generally around ε c = The height of the water column is 25 mm so that P w = 245 Pa and the bubble diameter is O(1 mm). With aqueous solution of [C 16 TAB] = 0.09 mm, γ aw = 70 mn/m. Thus, we obtain E i 4 J. After bubble bursting, submicrometre oil droplets are formed in the water column. Hence, the energy output E o is E o = γ ow S b (5) where γ ow is the surface tension between the oil and surfactant solution and S b is the magnitude of the area created by dispersing the droplets. We estimate S b ~ N d r 2 ~ΦV tot r 2 /r 3 ~ΦV tot /r, where Φ is the volume fraction of the oil in water, V tot is the total volume of the system and r is the mean radius of the droplets. We estimate the oil volume fraction Φ of emulsified oil 5 NATURE PHYSICS 5

6 gravimetrically and we found < Φ < 0.01, thus giving 1 < S b < 10 m 2. Taking γ ow = 40 mn/m, we obtain E o J. The estimated energy efficiency of our emulsification method η = E o / E i is thus between 1 and 10%. This estimate is at least an order of magnitude larger than the energy efficiency of high pressure homogenizers and microfluidizers, which is on the order of 0.1%. In contrast to high-pressure emulsification devices that operate at system-wide pressures 6 on the order of 10 8 Pa, our method works under atmospheric conditions. The difference lies in the fact that the energy density in our system is strongly dependent on the distance to the bubble while pressure (or energy density) is homogeneous in high-pressure systems. In order to understand this, we can estimate the energy density of our system, in the vicinity of the bubble. 113 We assume that the totality of the surface energy of a bursting bubble, E γ aw R b 2 is dissipated in the volume of a thin film V h f R f 2, where R b is the radius of the bubble, R f is the radius of one of the thin films, while h f is the film thickness. For a single film, the film radius can be estimated to 116 be of the order of R f h f R m 7, where R m is the radius of the meniscus around it. We can approximate that with the bubble radius, R b. Subsequently, the volume of the thin film, where all the energy is dissipated will be ~ R b h 2 f giving an energy density of φ = E/V γ aw R b /h 2 f, which for typical parameters R b 1 mm, h f 100 nm and γ aw = 70 mn/m gives φ 10 9 Pa. This value exceeds the energy density (pressures) in high pressure homogenizers or microfluidizers by one order of magnitude. This order-of-magnitude estimate gives an idea of why bubble bursting assisted emulsification can be a more efficient process compared to classic methods. 6 6 NATURE PHYSICS

7 SUPPLEMENTARY INFORMATION Tests for scalability For industrial applications, the scaling up of the current system is one issue we have to consider. Experiments show that the scattering intensity signal of the sample was increased by using multiple outlets to generate more bubbles or adding co-surfactants into the oil phase, such as SPAN 80, during the same time period, which means the concentration of oil was increased and hence the production rate. It demonstrates that our system is readily scaled up through parallelization Report of PDI To compare the DLS results, the cumulant model was used in all DLS measurement. The cumulant model gives a mean value for the size, and a width parameter known as the polydispersity index (PDI), which is an indication of the variance of the droplet size. A larger PDI means a wider distribution. Results of PDI for the measurements in Fig. 1d and Fig. 3c are reported in Fig. S High-speed videos Supplementary Movie 1 This movie shows a top-view of the bubble bursting process at a compound interface made of air/hexadecane/water at [C 16 TAB] = 0.09 mm. The two films burst at different times. 7 NATURE PHYSICS 7

8 Supplementary Movie 2 This movie shows a side-view of the bubble-bursting process at a compound interface made of air/hexadecane/water at [C 16 TAB] = 0.09 mm. The formation of a spray from the boundary of the cavity is observed Supplementary Movie 3 This movie shows a close-up view of the spray formed by bubble bursting at a compound interface made of air/hexadecane/water at [C 16 TAB] = 0.09 mm during dispersal. Smaller droplets are injected into the water column while bigger ones rise quickly up to the interface Supplementary Movie 4 This movie shows how 10- m latex beads are dispersed into the water column by the bursting of a bubble at an interface made of air/water at [C 16 TAB] = 0.09 mm. Small beads are injected from the left wall of the cavity into the bulk NATURE PHYSICS

9 SUPPLEMENTARY INFORMATION First drop Hexadecane on 0.9 mm CTAB Second drop , deg , deg time, s Figure S1. Raw ellipsometric data after placing drops of hexadecane on the surface of C 16 TAB solution. (1) Gradual thinning of the initially spread thick oil layer. (2) Very thin precursor layer (1.5-2 nm). (3) Lenses in coexistence with the precursor layer NATURE PHYSICS 9

10 500 thickness of uniform layer, nm time, s Figure S2. The process of film thinning at the stage (1) in Fig. S1 for the case of hexadecane on C 16 TAB solution. Note the sudden breakdown of the film that happens at a critical thickness NATURE PHYSICS

11 SUPPLEMENTARY INFORMATION Drop of C16 on 0.9 mm solution of CTAB Lenses 1 Lenses 2 Film thinning Precursor 190, deg nm 115 nm thinning , deg Figure S3. Plot of ( ) in two time intervals: 590 s 1028 s and s, from the experiment in Fig. S1. Note the three distinct regimes of film behavior. The coexistence of lenses and precursor layer is characterized by deviations in but not. These results suggest a lens thickness of ~115 nm (on average). 181 NATURE PHYSICS 11

12 drops of C16 on 0.9 mm CTAB s, mn/m nd drop 3-rd drop st drop time, s Figure S4. Surface pressure, s (measured using Langmuir trough - Nima Technology Ltd., UK, model 302 LL/D1), when drops of pure hexadecane are placed on a solution of 0.9 mm C 16 TAB (with already formed adsorption layer). The initial sharp increase of s is attributed to non-equilibrium spreading; after some time the layer reaches saturation. The fact that surface pressure changes from ~38 mn/m, which is an equilibrium value for the air/water interface with presence of 0.9 mm C 16 TAB to around a new value of ~41mN/m indicates the presence of oil at the interface NATURE PHYSICS

13 SUPPLEMENTARY INFORMATION Figure S5. Geometry of the oil lens as a spherical cap at the interface. The radius of the cap is represented by R. r is the radial distance, while θ and φ are the azimuthal and polar angles, respectively, in the spherical coordinate NATURE PHYSICS 13

14 Figure S6. PDI of all measurements in Figs. 1d and 3. (a) PDI of measurements for the same sample over a week period in Fig. 1d. (b) PDI of all measurements in Fig. 3a. (c) PDI of all measurements in Fig. 3b. (d) PDI of all measurements in Fig. 3c. Note that in (d) The PDI for undecane (N c = 11) is defined as the standard deviation of size measurements based on microscope images NATURE PHYSICS

15 SUPPLEMENTARY INFORMATION Table S1. Relationship between detectable submicrometre-sized droplets, concentration of C 12 TAB and equilibrium wetting state between the oil and surfactant solution. Equilibrium wetting Detectable state between the oil Oil phase [C 12 TAB] (mm) submicrometre and surfactant droplets solution 2 Hexadecane 0 No Partial wetting Hexadecane 0.02 No Partial wetting Hexadecane 0.09 Yes Partial wetting Hexadecane 1.4 Yes Hexadecane 7.0 Yes Pseudo-partial wetting Pseudo-partial wetting NATURE PHYSICS 15

16 Table S2. Surface tension oa, Hamaker constant A and the thickness of the nanometric film with the predicted droplet size r for different alkanes (Hamaker constant A is for the airalkane-water system, i.e. different alkanes on the same pure water 10. The thickness of the nanometric film is for the system of alkanes on the same aqueous surfactant solution ([C 12 TAB] = 20 mm) 8-9.) Oil phase oa (mn/m) A (10-20 J) ζ nm Predicted r (nm) Decane Undecane Dodecane Tridecane Tetradecane Pentadecane Hexadecane NATURE PHYSICS

17 SUPPLEMENTARY INFORMATION References 1. Wilkinson, K. M., Bain, C. D., Matsubara, H. & Aratono, M. Wetting of surfactant solutions by alkanes. Chemphyschem 6, (2005). 2. Matsubara, H. et al. First-Order Wetting transition and line tension of hexadecane lens at air/water interface assisted by surfactant adsorption. B. Chem. Soc. Jpn. 83, (2010). 3. Boulton-Stone, J. M. The effect of surfactant on bursting gas-bubbles. J. Fluid Mech. 302, (1995). 4. Solans, C. & Sole, I. Nano-emulsions: formation by low-energy methods. Curr. Opin. Colloid In. 17, (2012). 5. Cornelissen, R. L. & Hirs, G. G. Exergy analysis of cryogenic air separation. Energ. Convers. Manage 39, (1998). 6. Mason, T. G., Wilking, J. N., Meleson, K., Chang, C. B. & Graves, S. M. Nanoemulsions: formation, structure, and physical properties. J. Phys-Condens Mat. 18, R635-R666 (2006). 7. Ivanov, I. B. & Dekker, M. Thin Liquid Films (CRC press, 1988). 8. Aveyard, R., Cooper, P. & Fletcher, P. D. I. Solubilization of hydrocarbons in surfactant monolayers. J. Chem. Soc. Faraday T. 86, (1990). 9. Aveyard, R., Binks, B. P., Cooper, P. & Fletcher, P. D. I. Incorporation of hydrocarbons into surfactant monolayers. Adv. Colloid Interfac. 33, (1990). 10. Hough, D. B. & White, L. R. The calculation of Hamaker constants from Lifshitz theory with applications to wetting phenomena. Adv. Colloid Interfac. 14, 3-41 (1980). 11. Israelachvili, J. N. Intermolecular and Surface Forces. 2nd edn, (Academic Press, 1991) NATURE PHYSICS 17