Controlling Complex Nanoemulsion Morphology Using Asymmetric

Oct 31, 2017 - Complex nanoemulsions, comprising multiphase nanoscale droplets, hold considerable potential advantages as vehicles for encapsulation a...
1 downloads 11 Views 2MB Size
Subscriber access provided by READING UNIV

Article

Controlling complex nanoemulsion morphology using asymmetric co-surfactants for the preparation of polymer nanocapsules Mengwen Zhang, Patrick Corona, Nino Ruocco, David Alvarez, Paula Malo de Molina, Samir Mitragotri, and Matthew E. Helgeson Langmuir, Just Accepted Manuscript • DOI: 10.1021/acs.langmuir.7b02843 • Publication Date (Web): 31 Oct 2017 Downloaded from http://pubs.acs.org on November 2, 2017

Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a free service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are accessible to all readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.

Langmuir is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.

Page 1 of 46

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Langmuir

Controlling complex nanoemulsion morphology using asymmetric co-surfactants for the preparation of polymer nanocapsules Mengwen Zhang†, Patrick T. Corona†, Nino Ruocco†, David Alvarez†, Paula Malo de Molina†, Samir MitragotriϮ, Matthew E. Helgeson*† †

Department of Chemical Engineering, University of California, Santa Barbara, California

93106, United States

KEYWORDS: complex nanoemulsions; multiple emulsions; oil-in-water nanoemulsions; nanoparticles; nanocapsules

ACS Paragon Plus Environment

1

Langmuir

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 2 of 46

ABSTRACT Complex nanoemulsions, comprising multi-phase nanoscale droplets, hold considerable potential advantages as vehicles for encapsulation and delivery as well as templates for nanoparticle synthesis. Although methods exist to controllably produce complex emulsions at the micron-scale, very few methods exist to produce them at the nano-scale. Here, we examine a recently developed method involving a combination of high-energy emulsification with conventional co-surfactants to produce oil-water-oil (O/W/O) complex nanoemulsions. Specifically, we study in detail how the composition of the conventional ethoxylated cosurfactants Span80 and Tween20 influences the morphology and structure of the resulting complex nanoemulsions in the water-cyclohexane system. Using a combination of small-angle neutron scattering and cryo-electron microscopy, we find that the co-surfactant composition controls the generation of complex droplet morphologies including core-shell and multi-core shell O/W/O nanodroplets, resulting in an effective “state diagram” for the selection of nanoemulsion morphology. Additionally, the co-surfactant composition can be used to control the thickness of the water shell contained within the complex nanodroplets. We hypothesize that this degree of control, despite the highly non-equilibrium nature of the nanoemulsions, is ultimately determined by a competition between the opposing spontaneous curvature of the two co-surfactants, which strongly influences the interfacial curvature of the nanodroplets due to their ultra-low interfacial tension. This is supported by a correlation between co-surfactant compositions that produce complex nanoemulsions and those that produce homogeneous mixed micelles in equilibrium surfactant-cyclohexane solutions. Ultimately, we show that the formation of complex O/W/O nanoemulsion is weakly perturbed upon addition of hydrophilic polymer precursors, facilitating their use as templates for the formation of polymer nanocapsules.

ACS Paragon Plus Environment

2

Page 3 of 46

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Langmuir

INTRODUCTION There has been growing interest in using complex emulsions (i.e. multi-phase droplets consisting of one liquid phase dispersed in one or more other liquid phases1) to create structured nanomaterials for use in fields such as pharmaceuticals,2–4 foods,5–8 and chemical separations.9 The hierarchical structure of complex emulsions allows for the controlled encapsulation and release of small molecules,6,10 and could be used to template highly compartmentalized nanomaterials such as microcapsules and polymersomes.11–15 Recent technological advances in microfluidic devices have enabled the fabrication of highly sophisticated multiple emulsions on the order of 10-100 µm with precise control over their size and internal structure.16 Driven by a rapid need for the use of nanoscale structures in different applications (i.e. drug delivery17,18, imaging,19,20 and catalysis21), new approaches for producing nanoscale complex emulsions are sought after. However, such methods are still scarce, and are typically limited to a single morphology or internal configuration. For the case of microfluidic devices, the difficulty in producing nanoscale complex droplet structures lies in device limitations and the large energies required for their formation for the case of bulk emulsifications.22 This difficulty is compounded by a lack of systematic understanding in obtaining, controlling and stabilizing the complex morphologies. There have been several singular cases of successful production of water-in-oil-in-water (W/O/W) double nanoemulsions via either spontaneous emulsification or phase inversion methods.23–26 Compared to W/O/W double nanoemulsions, development of oil-in-water-in-oil (O/W/O) double nanoemulsions are even more lacking. Several groups have employed surfactants and high-energy methods to form reverse vesicle-type structures with an aqueous

ACS Paragon Plus Environment

3

Langmuir

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 4 of 46

shell encapsulating an oil core.27,28 Recently, Zhang et al. demonstrated the production of oil-inwater-in-oil (O/W/O) double nanoemulsions using sequential high-energy emulsification, and used them to produce oil-laden nanogels for drug delivery.29 Despite these scattered successes for forming complex nanoemulsions, there are currently no strategies that exist to produce a range of different complex nanodroplet morphologies (e.g. core-shell, multi-core and multi-shell) using the same emulsification method, nor fundamental guidelines by which such strategies could be used to select between these different morphologies. Recently, we successfully developed a method for producing O/W/O nanoemulsions with a range of morphologies involving high-energy emulsification in the presence of mixed surfactants, and showed how these nanodroplets could be used to template polymeric nanoparticles.30 The approach takes advantage of conventional co-surfactants, Tween20 (T20) and Span80 (S80), that possess opposite spontaneous curvature at the water-oil interface. While T20 prefers the formation of O/W emulsions,31 S80 prefers the formation of W/O emulsions.32 This specific combination of surfactant and co-surfactant is further known to impart ultra-low interfacial tension,33 and as such can influence the resulting interfacial curvature at the water-oil interface. We hypothesized that this frustration of spontaneous curvature of the surfactant mixture at the water-oil interface leads to a destabilization of the interface, and thus provides a driving force for the formation of complex nanoemulsions. We previously demonstrated that the preferred droplet microstructure created using this method – including microemulsions, core-shell O/W/O nanoemulsions, multi-core shell O/W/O nanoemulsions and uniform W/O nanoemulsions – could be selected by tuning the amount of water emulsified in a cyclohexane-T20/S80 mixture at a fixed surfactant concentration and surfactant ratio.30 However, it is currently unknown how the composition of the surfactant

ACS Paragon Plus Environment

4

Page 5 of 46

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Langmuir

mixture might be used to influence either the selection of the preferred morphology, or the internal features of the complex O/W/O nanoemulsions. In particular, it is unclear how the equilibrium self-assembly and interfacial behavior of the surfactant mixture influences the properties of highly non-equilibrium complex nanoemulsions. Experimentally testing this potential connection could ultimately allow for the formulation of near-equilibrium thermodynamic models to predict and control the morphology of complex nanoemulsions. The present work is therefore devoted to gaining a deeper understanding of how cosurfactant composition and equilibrium surfactant phase behavior influences the formation, morphology and structure of complex O/W/O nanoemulsions produced by T20 and S80 in the water-cyclohexane system. Specifically, we seek to determine how the selected droplet morphologies evolve with changes in both total surfactant concentration (γ) and co-surfactant mole fraction (δ). We hypothesize that tuning these compositional parameters varies the competition of spontaneous interfacial curvature at the water-oil interface, and will lead to changes in either the inner microstructures or selected morphology of the complex nanoemulsions. Additionally, we assess whether the presence of complex nanodroplet morphologies may be related to the underlying surfactant microstructure and phase behavior in solution. To address these points, a combination of small angle neutron scattering (SANS) and cryogenic-transmission electron microscopy (cryo-TEM) are used to determine the internal morphology and localization of components for O/W/O double nanoemulsions, and quantify their internal structure. Subsequently, we demonstrate the capability of using these complex morphologies as template for producing empty nanocapsules, a key limitation in previous studies of the method.30

ACS Paragon Plus Environment

5

Langmuir

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 6 of 46

THEORY To give context to the results to follow, we wish to provide a theoretical basis the hypothesis that complex nanoemulsions form due to frustrated spontaneous curvature of the two co-surfactants. To do so, we assume that that thermodynamics of a single nanodroplet are dominated by the interfacial free energy, for which we use the well-established Helfrich construction based on the local interfacial curvature.34 For a surfactant monolayer at the water-oil interface, the interfacial energy per unit area is expressed by an expansion of curvature to second order,35 E 2 = σ + 2 k ( H − cs ) + kK A

(1)

where σ is the surface tension, H is the mean curvature of the two principal curvatures [ =  

( +  )],  is the spontaneous curvature of the surfactant that minimizes the curvature

energy, K is the Gaussian curvature ( =  ), k is the bending modulus associated with the energy cost of bending the film away from its spontaneous curvature cs, and is the saddle-splay modulus accounting for energy costs associated with saddle-like deformations.36 For a spherical droplet with a single surfactant at the interface, Eqn. (1) becomes 2

1 1  E 1 = σ + 2k  −  + k 2 A R  R Rs 

(2)

where R is the droplet radius and Rs is the spontaneous radius of curvature. Nondimensionalizing Eqn. (2) by σ yields the following dimensionless equation, 2

E 2k  R k = 1+ 1−  + 2  σA σ R  Rs  σ R2

(3)

ACS Paragon Plus Environment

6

Page 7 of 46

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Langmuir

From Eqn. (3), we observe a dimensionless parameter ⁄  that determines the relative contributions of surface tension and surfactant spontaneous curvature to the overall interfacial energy. If ⁄  ≪ 1 , the surface tension term dominates and the impact of spontaneous curvature due to surfactant molecular geometry is negligible; if ⁄  ≥ 1, the surfactant contribution is significant and any difference between the droplet size R and spontaneous radius of curvature Rs will contribute significantly to the interfacial energy. Examining this energy scale, we observe that there exists a critical radius ∗ =  ⁄ at which the surfactant’s spontaneous curvature begins to dominate the interfacial free energy. Thus, if the droplet size can be reduced to R ~ R*, the molecular geometry of surfactants plays a considerable role in setting the interfacial energy, and thus the choice of surfactants plays a significant role in determining the preferred curvature of the resulting droplet interface. This concept is well-known and most commonly employed in so-called “spontaneous” nanoemulsification processes involving phase inversion, in which the values of k and σ precisely at the phase inversion point set the preferred radius of the resulting nanodroplets.37 This framework can be generalized to a mixed surfactant system with the surfactants of different spontaneous curvatures cs,1 and cs,2. If we assume ideal mixing of the surfactants at the interface, then 1⁄  =  ⁄ , + (1 − ⁄ , where δ is the mole fraction of surfactant 1 at the interface. Eqn. (3) then becomes, E 2k = 1+ σA σ R2

  δ 1−δ + 1 − R    Rs ,1 Rs ,2

2

 k   + 2   σ R

(4)

In this work, we are specifically interested in the case of surfactants possessing opposite spontaneous curvature, e.g. cs,1 = 1⁄ , is negative and cs,2 = 1⁄ , is positive. This case is interesting to the production of complex nanoemulsions for two reasons. First, the use of co-

ACS Paragon Plus Environment

7

Langmuir

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 8 of 46

surfactants with such asymmetric spontaneous curvature typically leads to situations of ultra-low surface tension, which facilitates the mechanical production of droplets whose size approaches R ~ R*. Second, previous theoretical studies established that interfaces with a mixture of surfactants of opposite curvature could experience thermodynamic instability and segregate into 2 interfaces, with asymmetric distribution of surfactants in the interfaces, one enriched in one surfactant and another in the other surfactant.38,39 In the context of complex nanoemulsions, this instability could provide a driving force for the energetically favorable formation of complex structures with internal interfaces (e.g. vesicles). Thus, the above analysis provides a route for producing nanodroplets with complex morphologies using surfactants with opposite spontaneous curvatures. In this work, this particular surfactant and co-surfactant pairs are known to produce ultralow surface tensions with σ ~ 0.01 mN/m at the oil-water interface.33 By assuming reasonable values of bending modulus (k ~ 10-20 to 10-19 J) for surfactants,40 we find that ∗ ~ 30-100 nm.

It is worth noting that in our previous work using a particular composition of the

abovementioned surfactant pairs, we were able to produce nanodroplets with complex morphologies on the order of R ~ 30-120nm,30 which is in this critical range. In this work, we thus hypothesize that varying the surfactant composition (i.e. surfactant ratio and concentration) may affect both k and σ to such an extent that would alter the dominant contributions of surface tension and surfactants’ spontaneous curvature to the interfacial curvature of the nanodroplet, and eventually influence the resulting nanodroplet morphology.

ACS Paragon Plus Environment

8

Page 9 of 46

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Langmuir

RESULTS AND DISCUSSION Effect of varying surfactant ratio We investigated the effect of the co-surfactant Tween20 mole fraction, δT20, on the selection of complex nanodroplet morphologies as the water volume fraction ɸw is increased. To do so, we fixed the total surfactant concentration (γ = 80mM) and studied six T20 mole fractions (δT20) spanning δT20 = 0 (pure Span80) to δT20 = 0.44. For this value of γ, we previously demonstrated the formation of complex nanoemulsions over a particular range of water volume fraction, ɸw.30 For each value of δT20, we prepared nanoemulsions with varying amounts of water (ɸw = 0 to 0.1). In our previous study, we found that the range of compositions for which complex nanoemulsions were formed correlated strongly with increases in the average hydrodynamic radius, RH, and dynamic viscosity, η, of the nanoemulsion compared to uniform (W/O) nanodroplets. We therefore used measurements of RH and η using DLS and viscometry, respectively, to screen for the probable range of δT20 and ɸw over which complex nanoemulsions form (Figure 1). Figure 1a compares the measured viscosity of the samples to that predicted for each ɸw and δT20 assuming that the emulsion is comprised of uniform (W/O) nanodroplets. The predicted viscosity is based on a correction to Einstein’s equation for hard spheres41 assuming that the dispersed phase is comprised only of the volume of water and surfactant,

η r − 1 = 2.5φ + 10.05φ 2

(5)

where ɸ is the total volume fraction of the dispersed phase, and includes both water (ɸw) and surfactants (ɸs), and ɳr = ɳ/ɳo where ɳ is the dynamic viscosity and ɳo the viscosity of

ACS Paragon Plus Environment

9

Langmuir

cyclohexane. If the measured viscosity equals that predicted by Eqn. (5), the system is likely comprised of uniform water droplets in cyclohexane, indicating the presence of either W/O microemulsions or nanoemulsions. However, if the measured viscosity is greater than that predicted by Eqn. (5), then the effective volume fraction ɸeff of the dispersed phase exceeds that of the water and surfactant alone. This suggests that the interior of the nanoemulsion contains some amount of cyclohexane enclosed within water droplets, indicating the presence of complex nanoemulsions. Thus, ɸeff would include contributions from water ɸw, surfactants ɸs, and the encapsulated cyclohexane ɸo.

η−1

a)

b)

10

I.

II.

III.

IV.

1

0.1 600

δΤ20=0 δΤ20=0.15

500

R (nm)

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 10 of 46

400

δΤ20=0.25

300

δΤ20=0.32 δΤ20=0.375

200

δΤ20=0.44

100 0 0.00

0.02

0.04

φw

0.06

0.08

0.10

Figure 1. (a) Reduced dynamic viscosity, ɳr and (b) Z-average hydrodynamic radius RH of nanoemulsions (γ = 80 mM) with varying water content (ɸw) for the T20 mole fractions, δT20, indicated. Solid lines in (a) indicate the viscosity predicted for uniform W/O nanoemulsions predicted from Eqn. (5). Error bars in (b) correspond to standard deviation (ε) of the droplet size distribution, where PDI = (ε/D)2.

Based on our previous study for δT20 = 0.375, we anticipate four regimes of composition with different preferred microstructures30 within the range of ɸw studied (Figure 1). For δT20 = 0.375, at ɸw < 0.015 where the predicted and measured viscosities agree quantitatively,

ACS Paragon Plus Environment

10

Page 11 of 46

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Langmuir

microemulsions are formed (region I). We distinguish microemulsions from nanoemulsions by examining the stability of the average droplet size over time (Figure S1, S2). Whereas nanoemulsions exhibit coarsening over time, the microemulsions exhibit no discernable change in RH during the course of experiments. For 0.015 < ɸw < 0.025 where measured viscosity begins to deviate significantly from that predicted by Eqn. 5, both SANS and cryo-TEM suggest the formation of core-shell type O/W/O nanoemulsions, in which a thin water film between a surfactant bilayer separates a cyclohexane droplet from the surrounding continuous phase (region II). For 0.025 < ɸw < 0.06 where we observe an even larger deviation from the predicted viscosity, SANS and cryo-TEM reveal the presence of multi-core shell nanoemulsions, in which multiple core-shell type O/W/O nanoemulsions are contained within a nanodroplet (region III). For ɸw > 0.06 where the measured viscosity is again equal to that predicted for uniform droplets, predominantly uniform W/O nanoemulsions are formed (region IV). Representative cryo-TEM images of these structures are shown in Figure 2 to verify the presence of the abovementioned microstructures associated with regions I-IV.

ACS Paragon Plus Environment

11

Langmuir

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 12 of 46

Figure 2. Evolution of microstructures (γ = 80mM, δT20 = 0.375) with increasing volume fraction of water, ɸw, added to the system. (left) Core-shell O/W/O nanoemulsions formed at ɸw = 0.02-0.025. (center) Multi-core O/W/O/W/O nanoemulsions formed at ɸw =0.03-0.05. (right) Predominately uniform W/O nanoemulsions formed at ɸw =0.10. The aqueous phase consists of 2wt% PTA as a positive stain. Red, blue, and black arrows are used to indicate core-shell, multi-core shell and uniform nanodroplets, respectively.

ACS Paragon Plus Environment

12

Page 13 of 46

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Langmuir

From these previous observations at δT20 = 0.375, we attempt to use viscosity and size measurements to decipher how changes in δT20 influence the selection of morphology (Figure 1). As shown in Figure 1, for δT20 = 0 (pure Span80), we expect only W/O microemulsions or nanoemulsions to form, which is confirmed by the agreement between the predicted and measured values of the suspension viscosity. The results are nearly identical for δT20 = 0.15 and δT20 = 0.25. As such, we surmise that a minimum amount of Tween20 in the co-surfactant mixture is required in order to form complex nanoemulsions. This suggests that at a critical surfactant ratio, the surface tension is dramatically reduced so that the spontaneous curvature of the surfactants at the oil-water interface dominates over the surface tension contribution to the interfacial free energy. At this point, the opposite spontaneous curvatures of the surfactants frustrate the interface, resulting in complex nanodroplets (i.e. with multiple interfaces within a droplet). For δT20 > 0.25, we observe significant viscosity deviations from that predicted by Eqn. (5) with increasing water volume fraction. For larger values of δT20 ≥ 0.32, we begin to observe behavior consistent with the previously observed transition from uniform W/O nanoemulsions to complex nanoemulsions, evident in deviations from the predicted viscosity values within the range of ɸw =0.015 to 0.05. As δT20 is increased further, the deviations of the observed viscosity become more pronounced. This indicates a larger relative volume of cyclohexane within the water droplets. This is consistent with the trends observed for the average radius of the emulsion droplets (Figure 1b). Specifically, there is a more pronounced increase in radius at higher δT20 in the same range of ɸw for which deviations in viscosity are observed. To confirm this, Eqn. (5) was used to estimate the amount of cyclohexane ɸo entrapped in the nanodroplets (region II) as

ACS Paragon Plus Environment

13

Langmuir

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 14 of 46

we increase δT20 (Table S1), assuming that the extra volume fraction of the droplets comes solely from cyclohexane.. We observe that the predicted amount of encapsulated cyclohexane in nanostructure increases from 5% to 27% as the mole fraction of Tween20 δT20 increases from 0.25 to 0.44. To characterize how the detailed features of the complex nanoemulsions are influenced by the surfactant ratio δT20, we focus on compositions exhibiting the core-shell type O/W/O nanoemulsion morphology, as they exhibit clearly distinguishable features (oil core and water shell thicknesses) and are representative of the other complex morphologies observed. SANS measurements on a series of nanoemulsions with fixed ɸw = 0.02, γ = 80 mM and varying δT20 (Figure 3a) was performed using H-surfactant/D2O/D12-cyclohexane in order to isolate scattering contrast from the surfactant film. SANS data indicate an apparent change in the qualitative shape of the scattering profile with δT20, suggesting changes in the features of the core-shell type O/W/O nanoemulsion with changes in surfactant composition. For δT20 = 0.32–0.44, we observe a power law slope of I ~ q-2 at moderate q-values and a shoulder at q ~ 0.9 nm-1, which indicates the distance between the surfactant layers stabilizing the water shell.30 In this range of δT20, the SANS data can be accurately fit to a 1-core three shell model consisting of a cyclohexane core, an inner surfactant layer (ti) consisting of the alkyl surfactant tails, an aqueous shell (taq) consisting of water and ethoxylated surfactant head groups, and an outer surfactant layer (to) consisting of the surfactant alkyl tails (Table 1). The model fitting was constrained in a similar manner as in previous studies, and further details of the fitting procedure are described in Supplementary Information (Section IV).30 Discrepancies in the oscillatory features of the SANS data and model fits at high q-values are indicative of polydispersity in the various thicknesses of the model. Attempts to incorporate such polydispersity in the model were insensitive to the

ACS Paragon Plus Environment

14

Page 15 of 46

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Langmuir

particular length scale for which polydispersity was included. Nevertheless, the model fits confirm that core-shell type O/W/O nanoemulsions are formed in this range of δT20. For δT20 < 0.25, both the shoulder at high q-values and the q-2 slope gradually disappear with decreasing δT20. For δT20 = 0.15–0.20, the SANS curves could be fitted to a bimodal distribution of large and small structures consisting of one population of core-shell structures and another population of triaxial ellipsoid structures. The core-shell model can be used to fit structures consisting of an aqueous core and a surfactant shell, which indicates the formation of uniform water-in-oil nanoemulsion droplets. The triaxial ellipsoid model can be used to fit the smaller structures, which corresponds to the formation of elongated surfactant micelles. For both δT20 = 0.15 and 0.20, we observe a mixture of predominantly W/O nanoemulsions (Rg ~ 10nm) and small amounts of ellipsoid micelles (Rg ~ 2nm). We therefore conclude that δT20 = 0.25 marks the transition from W/O nanoemulsions (δT20 < 0.25) to core-shell type O/W/O nanoemulsions (δT20 > 0.25), and this particular composition can be fitted by a mixture of 1-core 3-shell and core-shell models. These results confirm that, for fixed φw, a critical amount of T20 surfactant in the T20/S80 mixture is required for the formation of core-shell type O/W/O nanoemulsions.

ACS Paragon Plus Environment

15

Langmuir

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 16 of 46

Figure 3. SANS curve and fits for a) ɸw =0.02, and b) ɸw =0.01, at various δT20 and fixed γ = 80 mM

using H-surfactant/D2O/D12-cyclohexane samples; scattering intensities are vertically shifted to better differentiate SANS curves at each δT20. Detailed SANS fitting protocols can be found in SI Section IV.

Table 1. Features of microemulsions and core-shell type O/W/O nanoemulsions at γ = 80mM from SANS

δT20

Rc (nm)

ɸw = 0.01 (µE) σc tshell + ∆t (nm)

ɸeff

Rc (nm)

σc

ɸw = 0.02 (oil-in-water-in-oil nE) ti + ∆t taq+ ∆t to (nm) (nm) (nm)

ɸeff

0.25 4.53 0.32 4.45 0.375 5.34

0.33 0.32 0.29

2.181+0.015 2.052+0.015 1.348+0.010

0.024 0.034 0.039

--15.62 0.31 18.86 0.29

-1.678+0.004 0.923+0.007

-4.735+0.006 6.167+0.011

-2.052 1.348

-0.044 0.078

5.42

0.38

0.954+0.011

0.041

23.25 0.31

0.658+0.006

6.649+0.010

0.954

0.093

0.44

To gauge the importance of near-equilibrium interfacial mechanics in determining the O/W/O droplet morphology, it is useful to compare the core-shell nanoemulsion structures just described to the equilibrium microemulsion structures that form at lower φw. Therefore, SANS measurements were made on H-surfactant/D2O/D12-cyclohexane samples with identical surfactant composition, but with φw in the microemulsion region (Figure 3b). The microemulsion

ACS Paragon Plus Environment

16

Page 17 of 46

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Langmuir

SANS data were fit to a core-shell model consisting of a water core and surfactant shell, which provides an estimate of the surfactant shell thickness tshell. We observe that the surfactant shell decreases from 2.2 nm to 1.0 nm as δT20 increases from 0.25 to 0.44 (Table 1). This is expected because the fully extended alkyl chain length of S80 (2.30 nm) is considerably larger than that of T20 (1.55 nm), and indicates mixing of the two surfactants in the microemulsion shell. We now compare the surfactant and water shell thicknesses of core-shell type O/W/O nanoemulsions for different δT20 (Table 1). To accurately determine the aqueous shell thickness taq, we first analyzed the thickness of the inner (ti) and outer (to) surfactant layers. As discussed previously, theoretical predictions indicate that the difference in curvature between the inner and outer surfactant layers can facilitate a slight asymmetry in the distribution of surfactants on the two layers,39 which would lead to a difference in the inner and outer layer thicknesses ti and to. In order to account for this difference, the two surfactant layer thicknesses were allowed to be independently adjustable in the model fits. To fit the core-shell O/W/O structures, we performed the fitting in two steps. First, we fixed all the scattering length densities (i.e. the cyclohexane core, inner and outer surfactant shell, aqueous shell, and solvent) and set the thicknesses of both inner and outer surfactant shells (ti and to) to that of microemulsion shell thickness tshell in order to obtain reasonable fits for the core radius Rc, polydispersity σc, aqueous shell thickness taq, and effective droplet volume fraction ɸeff. We then tried to assess any asymmetry of the two surfactant films by letting ti and to and taq vary while fixing all the other parameters to the previous obtained values (See detailed fitting parameters in Table S9). To further validate the model fits, mass balances using viscosity and DLS data were performed to calculate taq (Table S1). The trend for taq with respect to δT20 was in approximate agreement with trend observed from SANS.

ACS Paragon Plus Environment

17

Langmuir

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 18 of 46

Interestingly, we find that the model fits well to the SANS curves when the outer layer thickness (to) equals that of the microemulsion surfactant shell thickness (tshell), whereas the inner layer thickness (ti) and the aqueous shell thickness (taq) vary with δT20 (Table 1). Furthermore, we observe that the fitted values of ti and to decrease with increasing δT20. This confirms that some mixing of the surfactants occurs in both layers, and reflects the shorter alkyl tail length of T20. However, the inner layer thickness ti is always significantly smaller than the outer layer thickness to, suggesting that there is more T20 residing in the inner layer compared to the outer layer. Because T20 has a spontaneous curvature that favors the formation of oil-in-water emulsion, it would prefer to reside in the inner surfactant layer of the O/W/O double nanoemulsion. Therefore, this difference in layer thickness suggests that there is indeed a compositional difference between the two surfactant layers. In principle, this could be further tested by allowing the scattering length densities (SLDs) of the two layers to vary in the experiments. However, this would require the use of surfactants with more significant SLD contrast than those used in the present study (ρS80 = 0.594∙10-6 Å-2 and ρT20 = 0.414∙10-6 Å-2). At the same time, the aqueous shell thickness taq increases with increasing δT20 (Table 1), indicating that more water can be imbibed between the oil regions of the O/W/O nanoemulsion with a higher amount of co-surfactant T20. This is consistent with the trends in ti and to just described. The increase in asymmetry of the layer thickness drives an increase in the asymmetry of the curvature of the layers, which accommodates an increase in the water film thickness. This could be aided by the larger ethoxylated head group of T20 in comparison with S80, which significantly reduces the water-oil interfacial tension33 and, in accordance with Eqn. (4), facilitates a larger degree of asymmetry between the surfactant layers.

ACS Paragon Plus Environment

18

Page 19 of 46

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Langmuir

Although the change in water shell thickness with δT20 measured here is relatively small, these results provide clear evidence that the water film thickness can be manipulated through the choice of co-surfactants, and in particular the degree of asymmetry in molecular geometry which may drive segregation of the two surfactants between inward- and outward-pointing interfaces within the complex nanoemulsions. Interestingly, we observe that the water shell thickness of the core-shell type O/W/O nanoemulsions correlates well with the radius of microemulsions at each value of δT20 studied. This suggests that the size of microemulsions formed by a given surfactant composition can be used as an approximate predictor of the water film thickness in complex nanoemulsions formed from the same surfactant composition.

Comparison to surfactant phase behavior in cyclohexane Given the similarity in various features of the O/W/O complex nanoemulsions and microemulsions that form under similar surfactant compositions, we also performed measurements to determine whether there is any possible relationship between the structure of the complex nanoemulsions and those formed by surfactant-cyclohexane solutions in the absence of water. Returning to the size measurements of surfactant mixtures by DLS (Figure 1), it is interesting to note that small stable microstructures are only formed for certain surfactant compositions in the absence of water. For example, for 0.15 < δT20< 0.25, large unstable structures on the order of hundreds of nanometers form and coarsen during the time of the measurements, and the solutions eventually phase separate macroscopically over time into an upper clear phase and a lower turbid phase. By contrast, for 0.32 < δT20 < 0.44, small stable structures with RH on the order of tens of nanometers are formed. Based on previous SANS

ACS Paragon Plus Environment

19

Langmuir

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 20 of 46

measurements at δT20 = 0.375, these small stable structures correspond to ellipsoidal mixed micelles of S80 and T20.30 For even higher values of δT20, the solubility limit of T20 is eventually exceeded, leading to macroscopic immiscibility of cyclohexane-rich and T20-rich phases. To further elucidate the surfactant nanostructures, SANS measurements were made for mixed S80/T20 H-surfactant solutions in D12-cyclohexane at values of δT20 corresponding to the range probed when water is added to the system (Figure 4a). For 0.32 < δT20 < 0.44, the SANS curves are fit well to a triaxial ellipsoid model, consistent with previous work,30 indicating the formation of elongated mixed micelles. However, for values of δT20 < 0.32, we observe both a gradual shift of the plateau in intensity to higher q values, as well as the sudden appearance of a significant excess scattering at low q-values. While the data at high q-values can still be fitted to triaxial ellipsoidal models, the upturn at the low q corresponds to large or aggregated structures.

ACS Paragon Plus Environment

20

Page 21 of 46

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Langmuir

Figure 4. a) SANS curve and fits for mixed micelles at various δT20 using H-surfactant/D12-cyclohexane samples; scattering intensities are vertically shifted to better differentiate SANS curves at each δ. b) CryoTEM images for δT20 = 0.25.

Table 2. Features of triaxial ellipsoid micelles (γ=80mM) at different δT20 δT20

Minor axis (nm)

Major Axis Polar Axis R Rgunier avg (nm) (nm)

0.44

2.65

33.76

8.67

17.64

16.92

0.375

2.64

24.40

8.61

13.22

14.34

0.32

2.26

17.66

7.36

9.84

9.82

0.25

1.74

13.30

5.55

7.42

7.67

0.20

1.44

7.87

3.80

4.58

4.62

0.15

1.34

5.76

2.95

3.45

2.45

This suggests that there is a change in surfactant aggregate microstructures at δT20 ~ 0.25. To further elucidate this transition, cryo-TEM was performed for δT20 = 0.25 and 0.375 (Figure 4b). For δT20 = 0.25, we observe large circular structures with an average size of 300-500 nm encapsulating smaller circular structures, indicative of multilamellar vesicles (MLVs). The

ACS Paragon Plus Environment

21

Langmuir

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 22 of 46

highly unstable nature of these vesicles suggests that they are non-equilibrium structures that will eventually undergo coarsening into a coexisting oil-rich phase and a surfactant-rich lamellar phase. For δT20 = 0.375, where we believe equilibrium mixed micelles to exist, we did not observe any large structures (Figure S3), which is consistent with the presence of nanometerscale mixed micelles. Thus, the combination of DLS, TEM, and SANS results are consistent with a transition with increasing δT20 from surfactant lamellar structures to small elongated micelles at δT20 ~ 0.32. Based on the theoretical framework we have proposed for the formation complex nanoemulsions involving near-equilibrium interfacial energy arguments, one might anticipate a correlation between the compositions for which complex nanoemulsions form upon the addition of water and those for which lamellar structures form from solutions of the co-surfactants in cyclohexane in the absence of water. However, surprisingly, we do not observe any specific correlation between the presence of these structures at particular surfactant compositions. For example, we find that for regions of γ and δT20 in which lamellar or vesicular structures are formed, we do not observe core-shell type nanoemulsions with the addition of water, but rather uniform W/O nanoemulsions. It thus appears that while surfactant compositions preferring lamellar structures result in simple O/W nanoemulsions upon water addition, surfactants mixtures that prefer to form mixed micelles results in the formation of core-shell type O/W/O nanoemulsions and complex nanoemulsions. We hypothesize that the reason for this counterintuitive behavior again relates to the limited yet non-negligible miscibility of the asymmetric co-surfactants.

ACS Paragon Plus Environment

22

Page 23 of 46

Effect of varying total surfactant concentration To determine the effect of total surfactant concentration (γ) on the formation of complex O/W/O nanoemulsions as we vary the water volume fraction ɸw, we first studied the reduced viscosity and size data of samples for various γ (Figure 5). We observe that as γ increases, the features of the viscosity curves indicating complex nanoemulsions shift to higher ɸw, and the range of ɸw where complex morphologies occur widens. A similar trend is observed in the effective hydrodynamic radius of nanodroplets (Figure 5b). This suggests that increasing the surfactant concentration merely delays the onset of complex structures; i.e., the value of ɸw that marks the onset of complex morphologies increases proportionally to the total surfactant concentration.

η-1

a)

b)

10

1

0.1 400 γ = 40mM γ = 80mM

300

R (nm)

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Langmuir

γ = 100mM γ = 130mM

200 100 0 0.00

0.02

0.04

0.06

φw

0.08

0.10

Figure 5. (a) Reduced dynamic viscosity, ɳr, and (b) radius, RH, of nanoemulsions with varying water content (ɸw) for different total surfactant concentrations, γ, and δT20 = 0.375. Error bars in (b) correspond to standard deviation (ε) of the droplet size distribution, where PDI = (ε/D)2.

ACS Paragon Plus Environment

23

Langmuir

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 24 of 46

To confirm this observation, SANS measurements were made for three distinct values of γ (80mM, 100mM and 130mM) at corresponding ɸw values (0.02, 0.025 and 0.033, respectively) that mark the onset of complex O/W/O nanoemulsions for a fixed co-surfactant composition δT20 = 0.375 (Figure 6). The samples were made using H-surfactant/D2O/D12-cyclohexane to isolate contrast from the surfactants. In all three cases, the scattering curves can be reasonably fitted to the 1-core 3-shell model characteristic of core-shell type O/W/O nanoemulsions described previously (Figure 6a). We note that normalization of the scattering intensity by ɸw collapses all three curves, with only a slight variation in the relative scattering intensity at low q-values (Figure 6b). This indicates that, unlike changing the surfactant ratio, changing total surfactant concentration does not significantly alter the internal features and length scales of the complex nanoemulsions (e.g. the water shell thickness or core radius), but rather merely delays the onset of complex structures, as shown by the viscosity measurements.

ACS Paragon Plus Environment

24

Page 25 of 46

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Langmuir

Figure 6. a) SANS curve and fits for oil-in-water-in-oil nanoemulsions at various γ using H-surfactant/ D2O/D12-cyclohexane; scattering intensities are vertically shifted to better differentiate SANS curves at each γ; b) normalization of scattering intensity with respect to ɸw for each γ.

Table 3. Features of core-shell type O/W/O nanoemulsions at δT20=0.375 from SANS oil-in-water-in-oil nE γ (mM) Rc (nm) 80 100 130

18.86 22.05 21.21

σc

ti + ∆t (nm)

taq + ∆t (nm) to (nm)

ɸeff

0.29 0.31 0.28

0.923+ 0.007 0.929 + 0.004 0.848+ 0.005

6.167 + 0.011 1.348 5.824 + 0.007 1.348 5.750 + 0.009 1.348

0.078 0.087 0.121

The aqueous shell thickness for each value of γ can be determined from fitting SANS data. Similar to before, the outer surfactant layer thickness (to) is the same as that of the microemulsion surfactant shell thickness (tshell) for δT20 ~ 0.375, which is approximately 1.35 nm. The inner surfactant layer thickness (ti), aqueous shell thickness (taq), core radius (Rc) and effective droplet volume fraction (ɸeff) were also determined (Table 3). Although there are slight variations in Rc, ti, and the aqueous shell thickness taq with varying γ, these could be explained by

ACS Paragon Plus Environment

25

Langmuir

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 26 of 46

small experimental error in selecting ɸw. The only significant change observed is for the effective volume fraction ɸeff as we increase γ. Simple mass balance calculations using DLS and viscosity data also indicate that while taq remains the same at different γ, ɸeff increases linearly with increase in γ (Table S2). This increase is consistent with an increase in γ providing a proportional increase in the amount of interfacial area available to stabilize O/W/O nanoemulsions, and therefore a proportional increase in the effective volume fraction of these structures. This confirms that increasing surfactant concentration does not alter the internal features of complex nanoemulsions, but merely alters the compositions for which they form due to limitations on the total available surface area of surfactant.

A state diagram for complex O/W/O nanoemulsions From the combination of viscosity, DLS, SANS, and cryo-TEM measurements over a wide range of compositions, we observe that the preferred morphologies of the microemulsions and complex nanoemulsions are uniquely determined by the composition of the fluid, namely the surfactant and water concentration parameters δT20, γ, and ɸw. With this in mind, we used the data to construct an effective “state diagram” of complex morphologies. The results are summarized in Figure 7 for varying δT20, γ, and ɸw, including regions for microemulsions, core-shell O/W/O nanoemulsions, multi-core shell O/W/O nanoemulsions, and W/O nanoemulsions. We find that varying the surfactant mole fraction δT20 produces a relatively rich state map, with complex changes to the boundaries between complex morphologies (Figure 7a). This is likely due to the significant changes in internal features of the complex nanoemulsions that occur upon changing δT20, and their influence on the interfacial energies required to produce them. We note that, in all

ACS Paragon Plus Environment

26

Page 27 of 46

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Langmuir

cases where complex nanoemulsions are observed, the nanoemulsions coarsen over time (Figures S1 and S2) toward an equilibrium state similar to that observed at large values of δT20 and ɸw (the “Two-phase” region in Figure 7a), involving coexistence of an upper oil-rich phase containing microemulsion droplets and a lower water-rich phase containing no measureable microstructure.

Figure 7. State diagrams for the morphology of nanostructures as a function of a) ɸw and δT20 with γ =

80 mM, and b) ɸw and γ with δT20 = 0.375.

ACS Paragon Plus Environment

27

Langmuir

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 28 of 46

By contrast, we find that varying the total surfactant concentration γ produces a systematic linear shift in the boundaries between various nanoemulsion morphologies (Figure 7b). Recalling that variations in γ do not significantly alter the internal features of the complex nanoemulsions, this suggests that there is a direct proportionality between the water volume fraction φw at which a particular nanoemulsion morphology becomes preferred and the total interfacial area for which there is surfactant available to occupy. In other words, the transitions between various morphologies are bounded in each case by an interfacial area balance constrained by the ratio φw/γ. In its simplest form, this area balance is given by ,  ,

= !"# $% 

(6)

where Aw,drop and Vw,drop are the total surface area and volume, respectively, of water contained in the various droplet morphologies, Nav is Avogadro’s number, and % is the average area per molecule that the co-surfactants occupy at the interface assuming ideal mixing of the two surfactants. Solving this expression for the ratio &'∗ ⁄$ ∗ given by the slope of the boundaries between morphologies yields ∗ 

(∗

=

∗ , )*+ ∗,

%

(7)

This equation can be interpreted in one of two ways. If one assumes ideal mixing of the cosurfactants at the water-oil interface, i.e. % = ,- %,,- + (1 − ,- %,./- , then one can ∗ ⁄5∗',1234 , which is related to the critical dimensions of estimate the critical shape factor 0',1234

the droplets at which one expects the preferred morphology to change (see Table 4). However, when we use previously reported values for %,,- and %,./- (0.81nm2 and 0.30nm2, respectively42,43), the resulting values of the droplet radii are more than an order of magnitude

ACS Paragon Plus Environment

28

Page 29 of 46

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Langmuir

smaller than those we observe in SANS measurements, and infeasibly small given the known dimensions of the surfactant molecules.

Table 4. Critical features of nanoemulsions estimated from Figure 5b and Equation 7 Morphology Dimensions

6∗7,89:;⁄∗ ? (nm2) (×10-4 mM-1) @

2 3

− R43 )

+ R42 )

4.46+0.89 0.26+0.05

9.69+0.94 0.56+0.05

If instead the dimensions of the nanodroplets are insensitive to the particular values of &'∗ and $ ∗ along the boundary between morphologies, which appears to be the case for the conditions we study (Figure 5b), then the equation above can be used to estimate the critical % required for a transition between the various morphologies. Using average area per molecule the values of radii for the different morphologies obtained from SANS to compute ∗ ⁄5∗',1234 (see Table 4), the corresponding estimates of 0',1234 % are shown in Table 4. Here, we

have assumed that the geometry of a multicore-shell nanodroplet can be approximated as an inner water shell encapsulated within an outer water shell with an intervening oil film, although we find that changes in the assumed number of inner cores negligibly affects the calculated value of . % We find that the estimated value of % increases significantly as the droplet morphologies become more complex, i.e. along a progression from microemulsion droplets to core-shell

ACS Paragon Plus Environment

29

Langmuir

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 30 of 46

nanoemulsions to multicore-shell nanoemulsions. One possible explanation for this is consistent with the hypothesis described earlier for the mechanism of formation of the complex nanoemulsions, which potentially involves demixing of the two surfactants between interfaces of different curvature, resulting in an effective increase of the average area occupied by surfactant molecules. Surprisingly, these results indicate that despite the highly non-equilibrium nature of these W/O and O/W/O nanoemulsions, it seems that their morphology is selected uniquely by the composition of the fluid from which they are prepared. This indicates the interesting possibility that near-equilibrium interfacial free energy models, such as those resulting in Eqs. (1-4), might be adequate for predicting a priori the preferred nanoemulsion morphology, akin to what was argued for in similarly highly non-equilibrium particle-stabilized emulsions.44 Such predictions would require careful measurement of the interfacial tension as well as the bending and saddlesplay moduli of the S80/T20 mixture at the cyclohexane-water interface. Unfortunately, such measurements are prohibitively challenging due to the extremely low surface tensions exhibited by this particular system (10-2 mN/m),33 and so we leave such efforts to future studies.

Synthesis of nanocapsules from core-shell type O/W/O complex nanoemulsion templates The demonstrated degree to which fluid composition can be used to select the morphology of complex O/W/O nanoemulsions facilitates their use as templates to form nanoparticles with complex morphologies. However, a key challenge in such templating in soft nanostructures is the ability of the template structure to be retained upon the addition of material precursors. Therefore, it is critical to establish whether the selection of complex nanoemulsion morphologies is

ACS Paragon Plus Environment

30

Page 31 of 46

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Langmuir

compromised in the presence of other components including monomers, polymers and other reactive species. In previous work,30 we demonstrated the formation of hydrogel nanoparticles using W/O nanoemulsion templates with a water volume fraction ɸw = 0.10 at δT20 = 0.375 and γ = 80 mM by incorporating 10 vol% of a crosslinkable hydrogel precursor, polyethylene glycol diacrylate PEGDA (Mn = 700) into the aqueous phase prior to emulsification. Because this composition lies in the W/O nanoemulsion region of the state diagram (Figure 7), this leads to the production of uniform solid nanogels (Figure 8, bottom row). However, in that work, we were unable to produce hydrogel nanocapsules from core-shell O/W/O nanoemulsions. We hypothesize that this was due to an inability to form percolated polymer network. This is because 10 vol% is at or near the percolation threshold of PEGDA,45 and heterogeneities in the PEGDA network structure on the order of the length scale of the water shell30,45 requires a significantly larger concentration in order to obtain a network that spans the entire water film. Experimentally, this was indicated by the reversion of the system to W/O microemulsion structures upon dilution with cyclohexane, even after UV exposure.

ACS Paragon Plus Environment

31

Langmuir

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 32 of 46

Figure 8. Templating nanocapsules using oil-in-water-in-oil nanoemulsion at ɸw =0.02 with 33 vol% PEGDA in aqueous phase, and templating of nanoparticles at ɸw =0.10 and using 10 vol% PEGDA. All samples are in cyclohexane with the exception of ɸw =0.10 after crosslinking, which is centrifuged down and redispersed in water. Red arrows correspond to core-shell type O/W/O nanoemulsions with and without PEGDA, and nanocapsules; blue arrows correspond to complex nanoemulsions and the resulting crosslinked structures; pink arrows correspond to W/O nanoemulsions and the resulting nanogels.

Here, we overcome this limitation to demonstrate the successful formation of nanocapsules using core-shell O/W/O nanoemulsions as templates at ɸw = 0.02, for δT20 = 0.375 and γ = 80mM. Specifically, for ɸw = 0.02, because of the ultra-thin aqueous shell (taq ~ 6 nm), a larger crosslinking density of PEGDA is required to form a self-sustaining network, which we achieve by increasing the concentration of PEGDA in the aqueous shell. This is confirmed by dilution tests after UV crosslinking of the templates, which show that while no permanent structures remain after crosslinking with 10 vol% PEGDA, permanent structures on the order of

ACS Paragon Plus Environment

32

Page 33 of 46

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Langmuir

100 nm were formed using 33 vol% PEGDA (Table S11). Thus, we used 33 vol% PEGDA in the aqueous phase to demonstrate the formation of nanocapsules. To confirm that the core-shell O/W/O nanoemulsion structure is retained as a template for forming nanocapsules in the presence of PEGDA, cryo-TEM images of the nanoemulsion before and after UV crosslinking were acquired (Figure 8, top row). As evident in the micrographs, core-shell O/W/O nanoemulsions form for ɸw = 0.02 with 33 vol% PEGDA, and resemble those formed without PEGDA. To further confirm the morphology, measurements of droplet growth rate and viscosity were made for the template with 33 vol% PEGDA, and compared to the template without PEGDA (Figure S7 and S8). The data suggest that adding 33 vol% PEGDA may slightly alter the boundaries for the transition from core-shell to multi-core shell O/W/O nanoemulsions, such that both morphologies are present at ɸw = 0.02. This observation is confirmed by cryo-TEM images of the nanocapsules formed after UV polymerization of PEGDA (Figure 8 and S5), in which we observe populations of nanocapsules with two distinct structures. The first has a single, diffuse polymeric shell resulting from the crosslinked PEGDA in core-shell O/W/O nanoemulsions (red arrows). The second consists of microstructures with internal features, which we attribute to crosslinking of PEGDA within multi-core shell O/W/O nanoemulsion structures (blue arrows). Nevertheless, we conclude that all of the various nanoemulsion morphologies observed in this work including uniform W/O nanodroplets, core-shell O/W/O nanodroplets, and multi-core shell O/W/O nanodroplets can be used to template the formation of hydrogel nanoparticles, giving rise to a range of different nanogel morphologies.

ACS Paragon Plus Environment

33

Langmuir

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 34 of 46

CONCLUSION We have investigated the preparation of core-shell and multi-core shell O/W/O complex nanoemulsions by a combination of high-energy emulsification with careful selection of cosurfactant pairs. The asymmetry in both molecular geometry and spontaneous curvature of the particular co-surfactants studied in this work (Tween20 and Span80) provide control over the interfacial curvature at the water-oil interface, and drive the formation of the various nanoemulsion morphologies. Under equivalent emulsification conditions, the preferred morphology depends only on the composition of the fluid including the co-surfactant ratio δT20, total surfactant concentration γ and water volume fraction ɸw. From this, we constructed a detailed state diagram by which complex nanodroplets with a particular morphology can be selected. Finally, the complex nanoemulsions can be used as templates to form complex nanoparticles and nanocapsules with potential uses for the encapsulation of molecules in drug delivery, foods, and chemical separations. In the course of these studies, we found that the co-surfactant composition, characterized here by the parameter δT20, is particularly important in determining the range of water concentration over which complex nanoemulsions form, as well as the particular internal features (e.g. the water film thickness) of the resulting nanodroplets. By contrast, we find that the total surfactant concentration merely sets the total interfacial area available for the formation of complex nanodroplet structures. We hypothesize that this is due to the strong influence of the spontaneous curvature of the surfactant mixture on the interfacial free energy which, due to the molecular asymmetry of the co-surfactants, may result in segregation of the two surfactants between interfaces of different curvature, and thereby facilitate the formation of complex nanoemulsions. This is evident in several experimentally observed trends, including changes in

ACS Paragon Plus Environment

34

Page 35 of 46

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Langmuir

the surfactant layer and water film thicknesses upon changes in δ. Moreover, it is evident in the connection we find between the ability of the surfactant mixture to form complex O/W/O nanodroplets upon addition of water and the ability of the same surfactant composition to form mixed micelles in solution in the oil phase. Based on these observations, this opens the possibility that the formation and morphology of complex O/W/O nanoemulsions might be predicted using near-equilibrium interfacial energy models, despite the highly non-equilibrium nature of the nanoemulsions themselves. Ongoing studies are aimed at testing this hypothesis, and determining the generality of these results to other co-surfactant systems.

EXPERIMENTAL METHODS Materials. Cyclohexane (ACS reagent grade > 99%), Tween20, Span80, Poly(ethylene glycol) diacrylate (Mn = 700g/mol), 2,2-diethoxyacetophenone, and phosphotungstic acid hydrate (PTA) were acquired from Sigma Aldrich and used as received. Deuterated cyclohexane (D12cyclohexane) and deuterated water (D2O) were acquired from Cambridge Isotope Laboratories. Preparation of Nanoemulsions. Nanoemulsions were prepared by emulsifying various amounts of deionized water into a mixture of surfactants and cyclohexane at a specific total surfactant concentration γ, where γ = (CT20 + CS80), and a Tween20 surfactant mole fraction δT20, defined as δT20 =CT20 /γ. δT20 was varied from 0 to 0.5, while γ was varied from 40mM to 130mM. These ranges of δT20 and γ were chosen in order to capture the regime where we observed complex morphologies and at the same time to avoid non-emulsifiable regimes (i.e. for δT20 values exceeding 0.5, macroscopic phase separation occurs). For SANS studies, deuterated solvents (D2O and D12-cyclohexane) were used to replace hydrogenated water and cyclohexane. For high-

ACS Paragon Plus Environment

35

Langmuir

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 36 of 46

energy emulsification, a 700 Watt Fischer sonicator dismembrator at 50% amplitude was used. Each sample was sonicated for 3 minutes with 5 s pulse on and 10 s pulse off in an ice bath. Synthesis of Nanogels from Nanoemulsions. To produce nanogels, a crosslinkable polymer, poly(ethylene glycol) diacrylate (PEGDA), was added to the aqueous phase before being emulsified in the cyclohexane-surfactant mixture. The morphology of the nanogels depends on the specific water volume fraction ɸw, which would dictate the minimum amount of PEGDA necessary to form a crosslinked network. For ɸw =0.02, 33 vol% of PEGDA was added to form nanocapsules; for ɸw =0.10, only 10% of PEGDA was needed to form solid nanoparticles. After emulsification, 2 vol% of photo initiator 2,2-diethoxyacetophenone was added and subsequently UV polymerized for 10 min using a 365 nm long-wave UV lamp. The nanoparticles were centrifuged down and stored in DI water for further use. Dynamic Viscosity Measurements. An Anton Paar densitometer with an inline falling ball viscometer (DMA 4100M) was used to measure the dynamic viscosity of nanoemulsions at 20 ° C. A glass Lovis capillary with a diameter of 1.59 mm and a 1.5 mm steel ball were used to perform viscosity measurements. 3 mL of sample was filled into the viscosity and measurements were repeated 3 times at a Lovis angle of 70°. Dynamic Light Scattering (DLS) Measurements. To measure the size of nanoemulsions, a BI200SM (Brookhaven Instruments) multi-angle detector along with a 632.8nm HeNe continuous wave laser was used. Autocorrelation functions were obtained at 20 °C and 90° scattering angle and fitted to a quadratic cumulant fit to obtain the average hydrodynamic size and polydispersity of the nanoemulsions.

ACS Paragon Plus Environment

36

Page 37 of 46

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Langmuir

Cryogenic Electron Microscopy (Cryo-TEM) A conventional vitrification robot system (Vitrobot Mark IV) was used to prepare cryo-TEM samples. For nanoemulsions prepared for cryo-TEM, 2 wt% PTA was added into the aqueous phase before emulsification in cyclohexane to enhance contrast between the liquid phases. To vitrify the samples, a fixed amount of sample (1.2 µL for aqueous samples and 2.0 µL for organic phase samples) was deposited onto copper grids (200 mesh) coated with lacey carbon films (Electron Microscopy Sciences). The grid was then blotted (blotting force: 2, blotting time: 1s, number of blots: 1) to remove excess liquid and quenched rapidly in the corresponding vitrifying liquid (liquid ethane for aqueous phase samples and liquid nitrogen for organic phase samples). The samples were then placed in a FEI Tecnai G2 Sphere TEM operated at an accelerating voltage of 200kV. Gatan Digital Micrograph software was used to record images. Small Angle Neutron Scattering (SANS) SANS measurements were performed at the NIST Center for Neutron Research on the NGB 30 m SANS instrument. Data were collected at four detector configurations to span the desired q-range (0.03 to 4nm-1) using a wavelength of λ = 8.4 and ∆λ / λ = 0.11 at a detector distance of 13.5m, and a wavelength of λ = 6 and ∆λ / λ = 0.11 at detector distances of 1m, 4m, and 13.5m. The nanoemulsions were loaded into 1 mm thick titanium scattering cells sandwiched between two quartz windows for SANS measurement and at a controlled temperature of 20 °C. SANS data were reduced using the NCNR IGOR software package46 and analyzed using the SasView47 software package.

ASSOCIATED CONTENT

ACS Paragon Plus Environment

37

Langmuir

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 38 of 46

Supporting Information. This document is available free of charge. Detailed protocol for SANS analysis in SI section I (Figure S1, Table S1-S8). Additional cryo-TEM images (Figure S2 to S4). Droplet growth data (Figure S5 and S6). Additional viscosity data and DLS data for nanocapsule templating (Figure S7, S8, Table S11). AUTHOR INFORMATION Corresponding Author * [email protected] Author Contributions MZ and MEH designed the experiments; MZ, DA, PC, NR and PMdM carried out the experiments. MZ, MEH, and SM prepared the manuscript. All authors have given approval to the final version of the manuscript. ACKNOWLEDGMENTS MZ was funded by the Hellman Foundation. PC and NR were funded by the Department of Energy, Award No. DE-SC0014127. PMdM was funded by the Defense Threat Reduction Agency under the Natick Soldier Research, Development and Engineering Center Agreement No. W911QY-13-2-0001. The MRL Shared Experimental Facilities are supported by the MRSEC Program of the NSF under Award No. DMR-1720256; a member of the NSF-funded Materials Research Facilities Network (www.mrfn.org). Access to the NGB30M SANS instrument was provided by the Center for High Resolution Neutron Scattering, a partnership between the National Institute of Standards and Technology and the National Science Foundation under Agreement No. DMR-1508249. We acknowledge the support of the National Institute of Standards and Technology, U.S. Department of Commerce, in providing the neutron

ACS Paragon Plus Environment

38

Page 39 of 46

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Langmuir

research facilities used in this work. We would also like to thank Dr. Yun Liu from NIST Center for Neutron Research for useful discussion on performing neutron scattering experiments. ABBREVIATIONS SANS - small-angle neutron scattering; PEGDA - poly(ethylene glycol) diacrylate; DLS dynamic light scattering; cryo-TEM - cryogenic transmission electron microscopy; W/O - waterin-oil-in-water; O/W/O - oil-in-water-in-oil; μE - microemulsion; MLV - multilamellar vesicle; nE - nanoemulsion. REFERENCES (1)

Aserin, A. Multiple Emulsions; Aserin, A., Ed.; John Wiley & Sons, Inc.: Hoboken, NJ, USA, 2007.

(2)

Okochi, H.; Nakano, M. Preparation and Evaluation of W/o/w Type Emulsions Containing Vancomycin. Adv. Drug Deliv. Rev. 2000, 45, 5–26.

(3)

Onuki, Y.; Morishita, M.; Takayama, K. Formulation Optimization of Water-in-Oil-Water Multiple Emulsion for Intestinal Insulin Delivery. J. Control. Release 2004, 97, 91–99.

(4)

Windbergs, M.; Zhao, Y.; Heyman, J.; Weitz, D. A. Biodegradable Core − Shell Carriers for Simultaneous Encapsulation of Synergistic Actives. J. Am. Chem. Soc. 2013, 135, 7933–7937.

(5)

Zeeb, B.; Saberi, A. H.; Weiss, J.; McClements, D. J. Retention and Release of Oil-inWater Emulsions from Filled Hydrogel Beads Composed of Calcium Alginate: Impact of Emulsifier Type and pH. Soft Matter 2015, 11, 2228–2236.

(6)

Muschiolik, G. Multiple Emulsions for Food Use. Curr. Opin. Colloid Interface Sci. 2007,

ACS Paragon Plus Environment

39

Langmuir

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 40 of 46

12, 213–220. (7)

Sapei, L.; Naqvi, M. A.; Rousseau, D. Stability and Release Properties of Double Emulsions for Food Applications. Food Hydrocoll. 2012, 27, 316–323.

(8)

Matalanis, A.; Jones, O. G.; McClements, D. J. Structured Biopolymer-Based Delivery Systems for Encapsulation, Protection, and Release of Lipophilic Compounds. Food Hydrocoll. 2011, 25, 1865–1880.

(9)

Kawasaki, J.; Kosuge, H.; Egashira, R.; Asawa, T. Mechanical Entrainment in W/O/W Emulsion Liquid Membrane. Sep. Sci. Technol. 2009, 44, 151–168.

(10)

McClements, D. J. Advances in Fabrication of Emulsions with Enhanced Functionality Using Structural Design Principles. Curr. Opin. Colloid Interface Sci. 2012, 17, 235–245.

(11)

Arriaga, L. R.; Amstad, E.; Weitz, D. A. Lab on a Chip. Lab Chip 2015, 15, 3335–3340.

(12)

Vilanova, N.; Kolen’ko, Y. V.; Solans, C.; Rodríguez-Abreu, C. Multiple Emulsions as Soft Templates for the Synthesis of Multifunctional Silicone Porous Particles. J. Colloid Interface Sci. 2015, 437, 235–243.

(13)

Abbaspourrad, A.; Carroll, N. J.; Kim, S.-H.; Weitz, D. a. Polymer Microcapsules with Programmable Active Release. J. Am. Chem. Soc. 2013, 135, 7744–7750.

(14)

Datta, S. S.; Abbaspourrad, A.; Amstad, E.; Fan, J.; Kim, S.-H.; Romanowsky, M.; Shum, H. C.; Sun, B.; Utada, A. S.; Windbergs, M.; et al. 25Th Anniversary Article: Double Emulsion Templated Solid Microcapsules: Mechanics and Controlled Release. Adv. Mater. 2014, 26, 2205–2218.

ACS Paragon Plus Environment

40

Page 41 of 46

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Langmuir

(15)

Kim, S.-H.; Nam, J.; Kim, J. W.; Kim, D.-H.; Han, S.-H.; Weitz, D. a. Formation of Polymersomes with Double Bilayers Templated by Quadruple Emulsions. Lab Chip 2013, 13, 1351–1356.

(16)

Abate,

a

R.;

Weitz,

D.

a.

High-Order

Multiple

Emulsions

Formed

in

Poly(dimethylsiloxane) Microfluidics. Small 2009, 5, 2030–2032. (17)

Hoare, T. R.; Kohane, D. S. Hydrogels in Drug Delivery: Progress and Challenges. Polymer (Guildf). 2008, 49, 1993–2007.

(18)

Hamidi, M.; Azadi, A.; Rafiei, P. Hydrogel Nanoparticles in Drug Delivery. Adv. Drug Deliv. Rev. 2008, 60, 1638–1649.

(19)

Matsushita, H.; Mizukami, S.; Sugihara, F.; Nakanishi, Y.; Yoshioka, Y.; Kikuchi, K. Multifunctional Core-Shell Silica Nanoparticles for Highly Sensitive 19 F Magnetic Resonance Imaging. Angew. Chemie 2014, 126, 1026–1029.

(20)

Schleich, N.; Sibret, P.; Danhier, P.; Ucakar, B.; Laurent, S.; Muller, R. N.; Jérôme, C.; Gallez, B.; Préat, V.; Danhier, F. Dual Anticancer Drug / Superparamagnetic Iron OxideLoaded PLGA-Based Nanoparticles for Cancer Therapy and Magnetic Resonance Imaging. 2013, 447, 94–101.

(21)

Gawande, M. B.; Goswami, A.; Asefa, T.; Guo, H.; Biradar, A. V; Peng, D.; Zboril, R.; Varma, R. S. Core–shell Nanoparticles: Synthesis and Applications in Catalysis and Electrocatalysis. Chem. Soc. Rev. 2015, 44, 7540–7590.

(22)

Helgeson, M. E. Colloidal Behavior of Nanoemulsions: Interactions, Structure, and

ACS Paragon Plus Environment

41

Langmuir

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 42 of 46

Rheology. Curr. Opin. Colloid Interface Sci. 2016, 25, 39–50. (23)

Lee, H. S.; Morrison, E. D.; Frethem, C. D.; Zasadzinski, J. A.; McCormick, A. V. Cryogenic Electron Microscopy Study of Nanoemulsion Formation from Microemulsions. Langmuir 2014, 30, 10826–10833.

(24)

Lee, H. S.; Morrison, E. D.; Zhang, Q.; Mccormick, A. V. Cryogenic Transmission Electron Microscopy Study: Preparation of Vesicular Dispersions by Quenching Microemulsions. J. Microsc. 2016, 263, 293–299.

(25)

Zhao, Y.; Zhang, J.; Wang, Q.; Li, J.; Han, B. Water-in-Oil-in-Water Double Nanoemulsion Induced by CO(2). Phys. Chem. Chem. Phys. 2011, 13, 684–689.

(26)

Shakeel, F.; Haq, N.; Al-Dhfyan, A.; Alanazi, F. K.; Alsarra, I. A. Double W/o/w Nanoemulsion of 5-Fluorouracil for Self-Nanoemulsifying Drug Delivery System. J. Mol. Liq. 2014, 200, 193–190.

(27)

Yang, D.; Zhao, J. Reverse Vesicles Formed by Polypentadeca- and PolyeicosaOxyethylene Mono N-Hexadecylethers in Cyclohexane. Colloid Polym. Sci. 2016, 294, 1037–1043.

(28)

Nakamura, K.; Uemoto, A.; Imae, T.; Solans, C.; Kunieda, H. Stability and Size Control of Reverse Vesicles. Journal of Colloid and Interface Science, 1995, 170, 367–373.

(29)

Zhang, M.; Nowak, M.; Malo De Molina, P.; Abramovitch, M.; Santizo, K.; Mitragotri, S.; Helgeson, M. E. Synthesis of Oil-Laden Poly(ethylene Glycol) Diacrylate Hydrogel Nanocapsules from Double Nanoemulsions. Langmuir 2017, 33, 6116–6126.

ACS Paragon Plus Environment

42

Page 43 of 46

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Langmuir

(30)

Malo De Molina, P.; Zhang, M.; Bayles, A. V.; Helgeson, M. E. Oil-in-Water-in-Oil Multinanoemulsions for Templating Complex Nanoparticles. Nano Lett. 2016, 16, 7325– 7332.

(31)

Porras, M.; Solans, C.; González, C.; Martínez, a.; Guinart, a.; Gutiérrez, J. M. Studies of Formation of W/O Nano-Emulsions. Colloids Surfaces A Physicochem. Eng. Asp. 2004, 249, 115–118.

(32)

Capdevila, M.; Maestro, A.; Porras, M.; Gutiérrez, J. M. Preparation of Span 80/oil/water Highly Concentrated Emulsions: Influence of Composition and Formation Variables and Scale-Up. J. Colloid Interface Sci. 2010, 345, 27–33.

(33)

Hashimoto, M.; Garstecki, P.; Stone, H. a.; Whitesides, G. M. Interfacial Instabilities in a Microfluidic Hele-Shaw Cell. Soft Matter 2008, 4, 1403.

(34)

Morse, D. C.; Milner, S. T. Statistical-Mechanics of Closed Fluid Membranes. Phys. Rev. E 1995, 52, 5918–5945.

(35)

Helfrich, W. Elastic Properties of Lipid Bilayers: Theory and Possible Experiments. Zeitschrift fur Naturforsch. - Sect. C J. Biosci. 1973, 28, 693–703.

(36)

Claessens, M. M. A. E.; Van Oort, B. F.; Leermakers, F. A. M.; Hoekstra, F. A.; Stuart, M. A. C. Bending Rigidity of Mixed Phospholipid Bilayers and the Equilibrium Radius of Corresponding Vesicles. Phys. Rev. E - Stat. Nonlinear, Soft Matter Phys. 2007, 76, 1–6.

(37)

Roger, K.; Olsson, U.; Schweins, R.; Cabane, B. Emulsion Ripening through Molecular Exchange at Droplet Contacts. Angew. Chemie - Int. Ed. 2015, 54, 1452–1455.

ACS Paragon Plus Environment

43

Langmuir

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

(38)

Page 44 of 46

Safran, S. A.; Pincus, P.; Andelman, D. Theory of Spontaneous Vesicle Formation in Surfactant Mixtures. Science (80-. ). 1990, 248, 354–356.

(39)

Safran, S. A.; Pincus, P. A.; Andelman, D.; MacKintosh, F. C. Stability and Phase Behavior of Mixed Surfactant Vesicles. Phys. Rev. A 1991, 43, 1071–8.

(40)

Delorme, N.; Bardeau, J. F.; Carrière, D.; Dubois, M.; Gourbil, A.; Mohwald, H.; Zemb, T.; Fery, A. Experimental Evidence of the Electrostatic Contribution to the Bending Rigidity of Charged Membranes. J. Phys. Chem. B 2007, 111, 2503–2505.

(41)

Thomas, D. G. Transport Characteristics of Suspension: VIII. A Note on the Viscosity of Newtonian Suspensions of Uniform Spherical Particles. J. Colloid Sci. 1965, 20, 267–277.

(42)

Gurkov, T. D.; Horozov, T. S.; Ivanov, I. B.; Borwankar, R. P. Composition of Mixed Adsorption Layers of Non-Ionic Surfactants an Oil/water Interfaces. Colloids Surfaces A Physicochem. Eng. Asp. 1994, 87, 81–92.

(43)

Politova, N.; Tcholakova, S.; Denkov, N. D. Factors Affecting the Stability of Water-OilWater Emulsion Films. Colloids Surfaces A Physicochem. Eng. Asp. 2017, 522, 608–620.

(44)

Park, B. J.; Lee, D. Particles at Fluid–fluid Interfaces: From Single-Particle Behavior to Hierarchical Assembly of Materials. MRS Bull. 2014, 39, 1089–1096.

(45)

De Molina, P. M.; Lad, S.; Helgeson, M. E. Heterogeneity and Its Influence on the Properties of Difunctional Poly(ethylene Glycol) Hydrogels: Structure and Mechanics. Macromolecules 2015, 48, 5402–5411.

(46)

Kline, S. R. Reduction and Analysis of SANS and USANS Data Using IGOR Pro. J. Appl.

ACS Paragon Plus Environment

44

Page 45 of 46

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Langmuir

Crystallogr. 2006, 39, 895–900. (47) SasView for Small Angle Scattering Analysis. http://sasview.org/index.html

ACS Paragon Plus Environment

45

O 1 2 3 4 5 6 7 8 9 10 11 12 13 14

W

Polymer nanocapsules Page 46 of 46

Langmuir Complex nanoemulsions

UV

O O W O ACS Paragon Plus Environment

W