Droplet core intermolecular interactions and block copolymer composition heavily influence oil-in-water nanoemulsion stability

Abstract

Colloidal oil-in-water nanoemulsions are gaining increasing interest as a nanoparticle delivery system due to their large oil droplet core that can carry a large payload. In order to formulate these particles with long-term stability, an appropriate oil media and block copolymer pair must be selected. The interaction between the nanoemulsion core and the polymer shell is critical to forming stable nanoparticles. Herein we probed how interactions between various polymers with hydrocarbon and perfluorocarbon oil media influenced nanoemulsion formation, stability, and size. Through a series of nanoemulsions with unique polymer/oil media combinations, we examined the effects of oil core hydrophobicity, fluorophilicity, surface charge, and volume as well as the effects of polymer tail composition. Surprisingly, we found that nanoemulsions formulated with pure perfluorocarbon oil cores versus perfluoro poly(ether) oil cores exhibited very different characteristics. We also found that both hydrocarbon and fluorocarbon polymer tails interacted favorably with perfluoro poly(ethers) as well as hydrocarbon oil cores forming stable nanoemulsions. We believe these results are focused on the unique properties of perfluorocarbons especially their rigidity, low polarizability, and near zero-surface charge. Interestingly, we saw that perfluoro poly(ethers) deviated from these expected properties resulting in an increased versatility when formulating nanoemulsions with perfluoro poly(ether) oil cores compared to pure perfluorocarbon oil cores. Nanoemulsion size, stability, growth rate, and life time were explored to probe these factors. Experimental and computational data will be presented as rationale.

Graphical Abstract

Introduction

Colloidal nanoemulsions are kinetically stable nanoparticles in the size range of 100-400 nm for biomedical applications.¹ These particles are formulated by the input of energy into a solution of two immiscible liquids and an appropriate surfactant. The focal point here will be oil-in-water nanoemulsions which result in oil droplet cores stabilized by a surfactant, surrounded by an aqueous environment. Nanoemulsions of this design are especially advantageous over other nanoparticle systems due to their large oil droplet core which is well suited for the solubilization of hydrophobic drugs or imaging agents.

Perfluorocarbons are a unique class of molecules where all carbon-hydrogen bonds have been replaced by carbon-fluorine bonds. This chemical change drastically alters the properties of the perfluorocarbons, as compared to their hydrocarbon analogs, due to the extremely strong carbon-fluorine bond (472 kJ/mol).² Compared to hydrogen, fluorine has a larger size, is much more electronegative, and yet has lower polarizability. This results in perfluorocarbons that exhibit minimal van der Waals interactions, rigid helical structures, high chemical and thermal stability, as well as biological inertness.^3,4 The low polarizability and weak van der Waals interactions lead to minimal cohesion forces between fluorous molecules resulting in dual hydrophobicity and lipophobicity.⁵ In solution, this phenomenon leads to perfluorocarbons forming a third, fluorous phase.⁶ The formation of the energetically favored fluorous phase is due to solvophobic effects rather than increased attraction between perfluorinated molecules.⁵ In the field of polymeric drug delivery, incorporating a fluorous moiety into a block copolymer greatly promotes polymer organization and assembly which leads to increased nanoparticle stability over time.⁴ Due to their biological inertness, semifluorinated polymers may also reduce recognition, binding, and clearance by blood serum proteins promoting in vivo particle stability⁷. For these reasons, semifluorinated polymers are a valuable niche of research to pursue in the field of drug delivery where nanoparticle assembly and stability are crucial factors.

Previous work in our laboratory has shown that our novel semifluorinated, triblock ABC copolymer forms extremely stable triphilic nanoemulsions and can deliver solid hydrophobic drugs and magnetic resonance imaging (MRI) contrast agents.⁸ This polymer, M2F8H18, is composed of a lipophilic tail, intermediate fluorophilic block, and hydrophilic corona. Interestingly, this ABC block copolymer formed stable nanoemulsions with a hydrocarbon, fluorocarbon, or combined hydrocarbon/fluorocarbon oil core. These results were surprising due to the expected interactions of the hydrocarbon tail with a purely hydrocarbon oil versus a fluorocarbon oil. This work probed the unusual polymer/oil core interactions observed in our previous studies at a deeper level as we compared a series of nanoemulsions formulated with varying polymers and oil media.

To perform this analysis, we selected four block copolymers (Figure 1) beginning with the aforementioned M2F8H18.⁸ As a comparison to M2F8H18, we also included its non-fluorous analog M2H18 that contains the exact same lipophilic and hydrophilic blocks but lacks the fluorophilic block. We also selected the triblock copolymer M2H10F13, which like M2F8H18 is triphilic but has a different orientation of blocks. For M2H10F13, the fluorophilic block is the tail - therefore interacting the most with the oil media core - followed by the intermediate lipophilic block and the same hydrophilic corona. We also included the non-hydrocarbon analog of M2H10F13, M2F13. These four polymers were expected to interact quite differently with a range of oil cores, based on their varying tails, and allowed us to assess the unique interactions observed in our previous work.⁸

Figure 1.

Chemical structures of synthesized block copolymer series. The nomenclature for the polymer series is based on the varying blocks: Mx represents methoxy capped poly(ethylene glycol) where x is the average molecular weight in the thousands, Fy represents the fluorocarbon block where y is the number of carbon atoms substituted with fluorine atoms, and Hz represents the hydrocarbon block where z is the number of carbon atoms substituted with hydrogens. Both diblock and triblock polymers follow this same nomenclature process.

For the oil cores, we selected four liquids with varying properties (Figure 2). First, we selected medium chain triglycerides and perfluoro-15-crown-5-ether, the hydrocarbon and perfluorocarbon media used in our previous work with M2F8H18.⁸ Medium chain triglycerides (MCT) are an FDA approved lipophilic medium derived from coconut oil. This hydrocarbon (Neobee M-5 grade) has a triglyceride backbone with three fully saturated fatty acids chains containing a range of eight to ten carbon units per chain.⁹ Perfluoro-15-crown-5-ether (PFCE) is a cyclic poly(ether) with a symmetrical chemical structure. This symmetry leads to all twenty fluorine atoms exhibiting a single resonance in ¹⁹F-NMR.¹⁰ Because the fluorine atoms are magnetically equivalent, this system is excellent for highly sensitive ¹⁹F-MRI applications and has been used extensively in research as an MRI contrast agent, cell labeling agent, oxygen carrier, and more.^11–14 Based on the surprising interactions between M2F8H18 and PFCE as well as the unprecedented loading of PFCE into these nanoemulsions, up to 35% v/v, we wanted to probe whether this phenomenon was generalizable for other perfluoro poly(ethers). As a result, we selected a linear perfluoro poly(ether) as a comparison: perfluorotriglyme (PFtriglyme). Short chain perfluoro poly(ethers) have gained increased interest in the industrial field in recent years. The addition of carbon-oxygen bonds between the fluorous groups leads to increased bond rotation in comparison to a typical, fully fluorinated molecule.^15,16 The increased rotation of perfluoro poly(ethers) is hypothesized to increase biodegradability and reduced persistence in the environment, making them a green alternative.¹⁷ This is especially key in large scale manufacturing facilities where product waste may enter the environment.^16,18 Finally, we selected a pure perfluorocarbon to contrast the behavior seen with the perfluoro poly(ethers): perfluorodecalin (PFdecalin). Similar to PFCE, perfluorodecalin is cyclic in structure but contains only carbon-fluorine bonds. Perfluorodecalin was used extensively in the past as an oxygen carrier in artificial blood applications. In fact, perfluorodecalin was the major component of the FDA-approved commercial blood substitute, Fluosol®. Perfluorodecalin was a common perfluorocarbon oil used in early stage fluorocarbon emulsions.^19–21

Figure 2.

Chemical structures of hydrocarbon and fluorocarbon oil media. Medium chain triglycerides (MCT), perfluoro-15-crown-5-ether (PFCE), perfluorotriglyme (PFtriglyme), and perfluorodecalin (PFdecalin).

Described in this article is the systematic formulation of a nanoemulsion series containing a range of polymer and oil core combinations (Figure 3). The resulting nanoemulsion properties will be discussed herein with an emphasis on the unique outcomes observed with perfluoro poly(ether) cores as compared with pure perfluorocarbon and hydrocarbon cores. We anticipate these findings will be useful in the design, prediction, and formulation of stable nanoemulsions for a variety of applications. Nanoemulsion size, stability, growth rate, and life time were explored to probe these factors. Experimental and computational data will be presented as rationale.

Figure 3:

Schematic representation of the series of oil-in-water nanoemulsion systems formulated here.

Experimental Section

Materials and Methods

Perfluoro-15-crown-5-ether was purchased from Oakwood Chemical (Estill, SC). Perfluorotriglyme was purchased from Exfluor Research (Round Rock, TX). Perfluorodecalin was purchased from Alfa Aesar (Ward Hill, MA). Medium chain triglycerides were purchased from Spectrum (Gardena, CA). Normal saline (AirLife sterile 0.9% sodium chloride for irrigation USP) was purchased from the University of Wisconsin Hospital Pharmacy. Solvents and all other reagents were purchased form Sigma Aldrich Co. (Milwaukee, WI) and Spectrum (Gardena, CA) and used as received, unless otherwise mentioned.

Polymer Synthesis and Characterization

Full details of polymer synthesis and characterization can be found in the Supporting Information (SI).

M2F8H18: This block copolymer was synthesized as previously reported.⁸

M2H18: The synthesis of M2H18 was performed, with slight modification (Scheme S1, Spectra S1-S2), as previously demonstrated by the Mecozzi group.²²

M2H10F18: The synthesis of M2H10F13 was performed, with some adaptation (Scheme S2, Spectra S3-S5), as previously demonstrated by the Mecozzi group.²³

M2F13: This block copolymer was synthesized with some modifications to the previous report (Scheme S3, Spectra S6-S8).²⁴

Microwave synthesis reactions were accomplished with an Anton Paar Monowave 300 Microwave Synthesis Reactor (Ashland, VA) equipped with an Anton Paar Autosampler MAS 24. Small molecule and polymer chromatography was accomplished with Silicycle 60 Å SiO₂ or by using a Teledyne CombiFlash® Rf 4x (Lincoln, NE) equipped with an ELSD for visualization and REDI-sep Rf high performance silica or C18 columns. ¹H- and ¹⁹F-NMR spectra were obtained on a Varian Unity-Inova 400 MHz. or Varian Unity-Inova 500 MHz. or Bruker Avance III HD 400 MHz. spectrometer using CDCl₃ as the solvent and TMS as the internal reference. Polymer purity was confirmed by MALDI-MS on a Bruker Ultraflex III MALDI TOF/TOF using α-cyano-4-hydroxycinnamic acid (CHCA) matrix at a ratio of 10:1 matrix: analyte.

Nanoemulsion Formulation

Aqueous polymer solutions were prepared freshly (10 mM) in sterile, normal saline and sonicated at room temperature until fully dissolved. Saline was composed of 0.9% (w/w) sodium chloride. Hydrocarbon or perfluorocarbon or a mixture of both were added to the polymer solution. The homogenizer and Microfluidizer® were first cleaned with 100% and 70% ethanol followed by 100% and 70% methanol and finally three rinses with Millipore Milli-Q water to remove all traces of any previous nanoemulsions. The prepared mixture was then homogenized with the high-speed homogenizer (Power Gen 500, Fisher Scientific, Hampton, NH) for 1 minute at 21,000 rpm at room temperature. The resulting crude emulsion was then further processed using the Microfluidizer® (model 110-S, Microfluidics Corp., Newton, MA) for 1 minute under 5,000 psi with the cooling bath kept at 0°C. The final emulsion was then filtered with a 0.8 μm cellulose acetate filter or 0.45μm nylon filter and stored in a sterile, polypropylene conical tube (Corning Inc., Corning, NY) at 4°C. It should be noted that the surfactant is used in excess and its concentration is always above the critical micelle concentration. Therefore, some of the surfactant is likely present as individual micelles rather than nanoemulsions particles. This is common in most emulsions. The presence of the micelles has no effect on the nanoemulsion stability and its potential applications.

Particle Size Determination via Dynamic Light Scattering (DLS)

Long-term nanoemulsion size and stability was monitored via dynamic light scattering (NICOMP 380ZLS, Particle Sizing Systems, Santa Barbara, CA). The nanoemulsions were diluted at the intensity factor of 500 by adding 2-50 μL of the nanoemulsion to 3 mL of Millipore Milli-Q water. Each particle size analysis was run for 5 minutes at room temperature in a disposable polystyrene cuvette and repeated three times. The data was analyzed using Gaussian analysis and reported as intensity weighted average diameters.

Nanoemulsion Imaging via Cryo-Electron Microscopy (Cryo-EM)

A M2F8H18 nanoemulsion sample was mailed to Case Western Reserve University where Dr. Sudheer Molugu performed sample preparation and imaging. A small aliquot of the nanoemulsion (3 μL) was applied to a Quantifoil holy carbon EM grid (R2/2, 300 mesh; EMS) and the excess sample was blotted away manually using Whatman filter. No dyes or stains were added to the sample. After blotting, the sample grid was rapidly plunged into liquid ethane (at approximately −168°C) to flash freeze the specimen. A JEOL 2200FS transmission electron microscope (200 kV, FEG, in-column energy filter) equipped with a Direct Electron DE20 detector was used for imaging. Due to the beam sensitivity of the nanoemulsion sample, just a few sub-second frames were averaged. Frames with noticeable beam damage were excluded from the final averaged micrographs. Micrographs were collected with a magnification of 25,000x and a defocus range of −1 to −4 μm.

Electrostatic Potential Surface Calculations

All calculations were carried out using the Spartan’10 software package from Wavefunction Inc. (Irvine, CA). All molecular geometries were optimized at the HF 6-31G* ab initio level. Electrostatic potential surfaces were HF 6-31G** single point calculations at the optimized geometries and color-coded. On all surfaces, electrostatic potential was mapped within the range of +30 kJ mol⁻¹ to −30 kJ mol⁻¹. Red color signifies a value greater than or equal to the maximum in negative potential and blue signifies a value greater than or equal to the maximum in positive potential.

Results and Discussion

Synthesis of Polymer Library

All four block copolymers were freshly synthesized for this study and thoroughly characterized via NMR and MALDI-MS. The synthesis of M2F8H18 was performed as previously reported.⁸ The synthesis of M2H18 was performed, with slight modification (Scheme S1), as previously demonstrated by the Mecozzi group and can be seen in the SI.²² The synthesis of M2H10F13 was performed, with some adaptation (Scheme S2), as previously demonstrated by the Mecozzi group and can be seen in the SI.²³ The synthesis of M2F13 was performed, with slight modification (Scheme S3), as previously reported and can be seen in the SI.²⁴

Formulation of Nanoemulsion Series

Formulation of the nanoemulsion series was very successful, resulting in opaque, milky colloidal solutions with particle sizes centered around 200 nm initially. High volumes of oil media were stabilized by the polymers. Nearly all of the nanoemulsions were highly stable , showing no signs of phase separation.

Nanoemulsion Size, Stability, and Organization

Long-term stability of nanoemulsions with oil volumes of 11-17% v/v

Initial nanoemulsions were formulated with a relatively small volume of oil media (11-17% v/v) to compare stability of a hydrocarbon core (MCT) contrasted against a perfluoro ether core (PFCE) with the full polymer series. We expected that the polymers containing a hydrocarbon tail (M2F8H18 and M2H18) would form the most stable particles with MCT while the polymers containing a fluorocarbon tail (M2H10F13 and M2F13) would form the most stable particles with PFCE. Interestingly, after formulation there was minimal to no difference in stability for the MCT core particles versus the PFCE core particles. All four polymers formed very stable, milky colloidal nanoemulsions with an MCT core, a PFCE core, and even a mixed 2:1 PFCE:MCT core (Figure 4). The series of nanoemulsions with a purely MCT core exhibited initial sizes in the narrow range of 216 – 232 nm with essentially no preference for particle formation with the polymers containing a hydrocarbon tail (Figure 4a, Table 1). The series of nanoemulsions with a purely PFCE core exhibited initial sizes in the range of 196 – 234 nm with slight preference for certain polymers over others (Figure 4b, Table 1). Surprisingly, the smallest nanoemulsion formulated here was the combination of M2H18 polymer with a PFCE core, in stark contrast to our initial hypothesis. The M2F13 polymer also formed especially small and stable nanoparticles with the PFCE core. Because the nanoparticle size differences observed in this four-polymer-series is so small, it may be due to the change in oil media or it may simply be due to the differences in size and packing of the diblock copolymers (M2H18 and M2F13) compared to the triblock copolymers (M2F8H18 and M2H10F13). Lastly, the series of nanoemulsions with a mixed 2:1 PFCE:MCT core exhibited initial sizes in the range of 214 – 236 nm with essentially no preference for any of the four polymers (Figure 4c, Table 1). All 12 nanoemulsions exhibited very minimal Ostwald ripening, one of the most common mechanisms leading to the growth and destabilization of emulsion particles. As a result, these nanoemulsions showed extremely high stability and consistent nanoparticle size over time; a lifetime of 250 days so far without any sign of phase separation (Tables S1–S3).

Figure 4.

Nanoemulsion particle size over time as measured by DLS: a) 11% v/v core media of medium chain triglycerides only, b) 11% v/v core media of perfluoro-15-crown-5-ether only, c) 17% v/v core media of a 2:1 mixture of perfluoro-15-crown-5-ether: medium chain triglycerides.

Note: Error bars are removed here for clarity. They are available in the Supporting Information, see Figure S1 and Table S1–S3.

Table 1.

Comprehensive nanoparticle data for nanoemulsions with oil volumes of 11-17% v/v.

Polymer Type	Oil Droplet Media Type and Volume	Particle Size Day 0 (nm)	Std. Dev. Day 0 (nm)	Number of Days Stablea
M2F8H18	MCT (11% v/v)	217.2	65.6	246+
M2H18		224.6	51.7	245+
M2H10F13		231.8	74.7	248+
M2F13		216.3	67.5	245+
M2F8H18	PFCE (11% v/v)	225.3	53.2	245+
M2H18		196.4	47.7	246+
M2H10F13		233.6	56.8	245+
M2F13		218.1	46.5	245+
M2F8H18	PFCE:MCT (2:1) (17% v/v)	235.5	56.3	246+
M2H18		214.2	68.1	246+
M2H10F13		234.8	62.0	245+
M2F13		233.3	34.8	245+

^aNanoparticle stability measurements are ongoing. No phase separation has occurred.

Long-term stability of nanoemulsions with oil volumes of 33% v/v

After observing the surprising stability of the polymer series with 11-17% v/v of MCT alone, PFCE alone, or a mixture of MCT with PFCE, we considered that this observed stability might be due to the relatively low volume of oil media in the core. We next formulated nanoemulsions with a very high volume of oil media (33% v/v). We also expanded our selection of oil media to include a traditional perfluorocarbon without any ether bonds (perfluorodecalin) as well as a second perfluoro poly(ether) that is linear (perfluorotriglyme) compared to the cyclic perfloro-15-crown-5-ether used initially. The M2F8H18 nanoemulsion series with varying perfluorocarbon oil cores exhibited initial sizes in the broad range of 223 – 270 nm (Figure 5a, Table 2). Both perfluoro poly(ether) cores formed very stable particles with the M2F8H18 and have been stable for 250 days so far with essentially no particle growth via Ostwald ripening. In stark contrast, the perfluorodecalin pure fluorocarbon core formed a very unstable nanoemulsion with M2F8H18 that exhibited fast Ostwald ripening and eventually phase separation after 14 days (Table S4). The M2H18 nanoemulsion series showed very similar results with initial sizes in the broad range of 195 – 255 nm (Figure 5b, Table 2). Once again, the perfluoro poly(ether) cores interacted favorable with the M2H18 polymer and formed extremely stable particles so far. Once again, the perfluorodecalin formed the largest and most unstable nanoemulsion with M2H18 that showed fast particle growth and phase separation after 14 days (Table S5). In order to rule out that these stability issues were due to the hydrocarbon tails of the M2F8H18 and M2H18 polymers, we also formulated a 33% v/v perfluorodecalin nanoemulsion with M2F13. In theory, M2F13 should form a very stable nanoparticle with a pure fluorocarbon oil core of PFdecalin due to favorable fluorophilic interactions when the perfluorocarbon polymer tails insert into the fluorous oil core. The initial formulation was successful with particles around 253 nm but once again nanoparticle growth was fast and phase separation occurred after 42 days (Table 2, Table S6). Though the presence of the perfluorocarbon tail in M2F13 did increase the PFdecalin nanoemulsion lifetime by 27 days, these particles were still quite unstable. It is interesting to notice how the similarity in hydrophobicity/fluorophilicity between the liquid forming the core droplet and the copolymer block in direct contact with the emulsion core, did not lead to a difference in the particle size and their standard deviation. Rather, it manifested itself through a dramatic change in the nanoemulsion stability as evidenced by an onset phase separation in the PFdecalin nanoemulsions (Table 2). These findings support the idea that perfluoro poly(ethers) have very different properties compared to pure perfluorocarbons. We pursued this idea further by looking at particle growth rates and surface charge through computational modeling.

Figure 5.

Nanoemulsion particle size over time as measured by DLS: a) M2F8H18 polymer formulated with 33% v/v core media, b) M2H18 polymer formulated with 33% v/v core media. Black X represents phase separation of the nanoemulsion.

Table 2.

Comprehensive nanoparticle data for nanoemulsions with oil volumes of 33% v/v.

Polymer Type	Oil Droplet Media Type (33% v/v)	Particle Size Day 0 (nm)	Std. Dev. Day 0 (nm)	Number of Days Stable
M2F8H18	PFCE	253.9	43.7	245+ a
M2H18		222.2	54.0	246+ a
M2F8H18	PFtriglyme	223.4	49.6	238+ a
M2H18		195.1	33.6	224+ a
M2F8H18	PFdecalin	269.9	62.4	14 b
M2H18		254.6	21.1	14 b
M2F13		253.3	17.0	42 b

^aNanoparticle stability measurements are ongoing. No phase separation has occurred.

^bNanoparticle stability measurements are finished. Phase separation occurred.

Cryo-EM images of multiphasic nanoemulsion

After observing the surprising stability of all four polymers with a 2:1 mixture of PFCE:MCT oil cores, we probed these unique nanoassemblies further by visualization through cryo-electron microscopy (cryo-EM). Cryo-EM has gained increasing traction in the past several years due to its improved resolution for biological samples, after three-dimensional image reconstruction, and to the expanding range of sample types that can be imaged, including nanoparticles²⁵. This technique is crucial for visualization of our nanoemulsion particles due to their dynamic, colloidal character that is not compatible with many microscopy techniques where samples need to be dried, stained, or manipulated in a way that distorts the natural environment of these nanoemulsions. By flash-freezing our particles, in the absence of any dyes or stains, we were able to obtain very clear images of the nanoemulsions in a near-natural state. We performed cryo-EM imaging on an M2F8H18 nanoemulsion with a 2:1 ratio of PFCE:MCT in the core. The cryo-EM images confirmed we have spherical, fairly uniform particles that are reasonably homogenous in size (Figure 6). These images also revealed unique interactions of the oil media in the droplet core leading to a biphasic core, or Janus droplet, which has been reported in the literature²⁶ but is a rare occurrence. The biphasic droplets that are visualized in Figure 6 provide excellent support for why we see both hydrocarbon and fluorocarbon polymer tails interacting well with this oil core. The discovery of a biphasic oil core also correlated well with a slow payload release from these nanoparticles, which confirmed previous work with M2F8H18.⁸

Figure 6.

Cryo-EM image of a M2F8H18 nanoemulsion containing: (red) 2 mL medium chain triglycerides (MCT), (green) 4 mL perfluoro-15-crown-5-ether (PFCE), 20 mM polymer, and 12 mL saline. The multiphasic oil core is visible here with MCT assigned to the light gray core (red arrow) and PFCE assigned to the black core (green arrow). The polymer shell cannot be appreciated here due to its very small size compared to the droplet core. This image indicates the particles have a spherical shape, are reasonably homogenous in composition, and have a very unique biphasic core composed of both MCT and PFCE.

Multiphasic Core Interactions Key to Nanoemulsion Stability

Ripening rate of nanoemulsions

Following the nanoemulsion DLS studies we observed no significant difference in nanoemulsion size and stability, regardless of polymer structure, with an oil core of MCT, perfluoro poly(ethers), or a mixture of the two media. In stark contrast, we observed high instability when the oil core was composed of pure fluorocarbon media, perfluorodecalin in this case. Polymers with hydrocarbon tails, M2F8H18 and M2H18, formed very unstable nanoemulsions with PFdecalin as did a polymer with a fluorocarbon tail, M2F13. We investigated this further by measuring the Ostwald ripening rate of these three nanoemulsions all containing 33% v/v PFdecalin.

$$\mathit{Ostwald\; ripening\; rate}=\frac{{\mathit{dr}}^3}{\mathit{dt}}$$ (1)

Where r is the average particle radius and t is time.^27,28 The nanoemulsion growth profiles follow our expected trend where the most unstable, fast growing nanoemulsion is the one formulated with M2H18 which contains a long hydrocarbon tail and no fluorous block (Figure 7). The interactions between perfluorodecalin and a hydrocarbon tail were expected to be disfavored as is confirmed by the growth rates and the phase separation that occurred after 14 days. The next most unstable, fast growing nanoemulsion is the one formulated with M2F8H18 which contains the identical long hydrocarbon tails but does contain an intermediate fluorous block (Figure 7). The same disfavored interactions between perfluorodecalin and the hydrocarbon tail were expected but it appeared the introduction of the intermediate fluorous block stabilizes the core slightly, as seen by the growth rates below. This is a subtle difference though and is not easily appreciated since phase separation for the M2F8H18 nanoemulsion also occurred after 14 days. The most stable nanoemulsion, with the slowest growth, was the one formulated with M2F13 which contains a long fluorocarbon block and no hydrocarbon block. The interactions between the fluorous tail and perfluorodecalin were expected to be favored as is confirmed by the slowest particle growth and longest living nanoemulsion that did not phase separate until 42 days. This trend was expected based on pure hydrocarbon/fluorocarbon interactions. This data adds further evidence to the unexpected nature of the favorable interactions observed with all of the hydrocarbon-tail-containing-polymers and the two perfluoro poly(ether) oil cores: PFCE and PFtriglyme.

Figure 7.

Ostwald ripening rates of nanoemulsion particles over time where the particle radius was measured by DLS. All nanoemulsions contained 33% v/v perfluorodecalin, 12 mL saline, and 10 mM polymer. As expected, the pure perfluorocarbon, perfluorodecalin, favored interactions with fluorous polymer tails over hydrocarbon polymer tails. This further emphasizes the drastic changes observed between pure perfluorocarbon oil cores and perfluoro poly(ether) oil cores.

Electrostatic potential surfaces of oil media

The surprising differences in the behavior of pure perfluorocarbons versus perfluoro poly(ethers) in our nanoemulsion cores can be further understood by observing the electrostatic potential on the surfaces of these molecules. The oil media were computationally modeled using Spartan’10 and their electrostatic potential surfaces were compared. Perfluorododecane (Figure 8A), our model pure perfluorocarbon, can be observed to have an evenly distributed surface with a near-zero charge across the entire molecule. This is the accepted expectation for surface charge of a pure perfluorocarbon and evidence of its extreme hydrophobicity.⁴ This can be contrasted with the hydrocarbon analog triglyme (Figure 8D) which was observed to have a very uneven distribution of charge due to polarization from the oxygen atoms and the absence of fluorine in the molecule which leads to patches of both positive and negative electrostatic potential on the surface. The surface charge differences between pure perfluorocarbons and hydrocarbons led to differences in hydrophobicity and polarity which have been quantified using pyrene fluorescence. The Py scale corroborates that pure perfluorocarbons, such as perfluorodecalin, are even more hydrophobic than pure hydrocarbons.²⁹ Perfluorotriglyme (Figure 8B) is observed to fall between the pure perfluorocarbon, perfluorododecane, and the pure hydrocarbon, triglyme. In comparison to the nearly neutral surface charge of perfluorododecane, perfluorotriglyme can be observed to have pockets of increased charge due to the polarizing ether oxygen atoms throughout the chain. The presence of the oxygen atoms led to increased polarization of the molecule and some amount of surface charge build up. Though the surface charge is more subtle compared to the triglyme hydrocarbon analog, it is reasonable to expect that it is enough to impact polymer/oil core interactions and nanoemulsion stability. Finally, perfluoro-15-crown-5-ether (Figure 8C) is observed to be even more similar to the pure hydrocarbon. Perfluoro-15-crown-5-ether has even larger pockets of increased charge due to the polarizing ether oxygen atoms and as a result, increased surface charge build up. These EPS support the observed data reported herein as well as our previous data⁸ that showed unprecedented loading of PFCE into M2F8H18 and M2H18 nanoemulsions containing long hydrocarbon polymer tails. It is hypothesized that these interactions are more favorable due to decreased hydrophobicity, decreased fluorophilicity, and decreased driving force for separation into the fluorous phase as well as increased molecular polarization, increased bond rotation, increased surface charge, and increased van der Waals forces on the perfluoro poly(ethers). These factors resulted in more favorable van der Waals interactions between the polymer tails and the perfluoro poly(ether) oil cores than predicted. These trends did not translate across all perfluorocarbons, based on the unfavorable interactions observed with perfluorodecalin due to the differences in intermolecular properties of pure perfluorocarbons, but have so far held true for two perfluoro poly(ether) derivatives.

Figure 8.

Spartan’10 ® (Wavefunction Inc., Irvine, CA) electrostatic potential surfaces of: a) a pure perfluorocarbon C₁₂F₂₆; b) a perfluoro poly(ether) C₈F₁₈O₄; c) a perfluoro crown ether C₁₀F₂₀O₅; and d) an oligo poly(ether) C₈H₁₈O₄. Red indicates an excess of negative potential while blue indicates an excess of positive potential. The ether oxygens in b) and c) contribute an increased molecular polarity and bond rotation as opposed to a pure perfluorocarbon a), which explains the decreased fluorophilicity and decreased hydrophobicity of the perfluoro poly(ethers). Electrostatic potential was mapped within the range +30 kJ mol⁻¹ (blue) to −30 kJ mol⁻¹ (red).

Summary and Conclusions

Herein we have prepared a series of nanoemulsions with varying polymer and oil core combinations to compare how the polymer/oil core intermolecular interactions and composition affect long-term nanoemulsion stability. At lower volumes (11-17% v/v) of MCT only, PFCE only, and a 2:1 mixture of PFCE:MCT, we found that there were minimal differences in nanoemulsion stability even when comparing long hydrocarbon or fluorocarbon polymer tail interactions with the oil droplet core. All 12 low-volume nanoemulsions exhibited excellent stability, so far up to 250 days, and nearly no Ostwald ripening. These results indicated that at low volumes, both hydrocarbon and perfluoro poly(ether) oil cores interacted favorably with hydrocarbon polymer tails (M2F8H18 and M2H18) and perfluorocarbon polymer tails (M2H10F13 and M2F13) which produced stable nanoemulsions. At higher volumes (33% v/v) of PFCE and a second perfluoro poly(ether) PFtriglyme, we observed the same stability with M2F8H18 and M2H18. In great contrast, the polymer tail/oil core interactions were disfavored when a pure perfluorocarbon, PFdecalin, was introduced as the oil media. This stark contrast in nanoemulsion stability when transitioning from a perfluoro poly(ether) oil core to a pure perfluorocarbon oil core can be rationalized by differences in surface charge, molecular polarity, flexibility, and van der Waals interactions, as well as overall hydrophobicity and fluorophilicity changes. Perfluoro poly(ethers) contain ether oxygen atoms through their structure which greatly impact their bond rotation, flexibility, polarity, and surface charge. In contrast to pure perfluorocarbons that show a near-zero surface charge and minimal molecular polarity, we showed that there is an increased surface charge and molecular polarity on both perfluoro-15-crown-5-ether as well as perfluorotriglyme through computational modeling. Perfluorodecalin exhibited the most hydrophobic and most fluorophilic profile, followed by perfluoro-15-crown-5-ether and perfluorotriglyme, both with decreased hydrophobicity and decreased fluorophilicity. This trend was corroborated by nanoemulsion particle growth comparisons that show the fastest Ostwald ripening occurring with a perfluorodecalin and M2H18 nanoemulsion combination, followed by perfluorodecalin and M2F8H18, and the most stable interaction with perfluorodecalin and M2F13. Overall, we showed that perfluoro poly(ethers) have very unique electrostatic and molecular properties when compared to pure perfluorocarbons. These perfluoro poly(ether) properties can be leveraged to form very stable nanoemulsion particles with block copolymers containing either hydrocarbon or perfluorocarbon tails. We also showed that polymer/oil core interactions are vital to predicting nanoemulsion formulation success, size, and long-term stability. To the best of our knowledge, this is the first time pure perfluorocarbons have been compared to perfluoro poly(ethers) in the context of nanoemulsion stability and polymer/oil core interactions. We expect these findings will be useful in the design, prediction, and formulation of stable nanoemulsions for a variety of applications. In addition, this work emphasizes the benefits of using perfluoro poly(ethers) in the drug delivery field and beyond.

ABBREVIATIONS

Cryo-EM

cryo-electron microscopy

DLS

dynamic light scattering

MALDI-MS

matrix assisted laser desorption ionization-mass spectrometry

MCT

medium chain triglycerides

PFCE

perfluoro-15-crown-5-ether

PFdecalin

perfluorodecalin

PFtriglyme

perfluorotriglyme