Emulsion System

Emulsions are usually opaque and usually cream. Since we need to see into the emulsion sample, opacity is a problem. We also need control over the system's propensity to cream, both to monitor the impact of gravity on the dynamics, and to slow overall vertical drift sufficiently that confocal microscopy is possible. These two requirements mean that we need an emulsion for which the dispersed (oil) phase and the continuous (“water”) phase are refractive index matched, to remove opacity, and density matched, to eliminate creaming.

Matching either refractive index or density is not too difficult, but achieving both simultaneously and without polluting the system with additives such as salt or sucrose is much harder.

The emulsion system also needs to be stable against droplet coalescence and ripening, with the disperse phase insoluble in water. There also has to be a dopant present to enable NMR, and a fluorescent dye to enable confocal microscopy.

We have developed a new emulsion system, based on a mixed n-hexane and tetradecafluorohexane oil phase, that exhibits all the desired properties. The use of a perfluorinated oil is important, since these oils have refractive indices lower than that of water and densities greater than that of water. Tetradecafluorohexane is freely and cheaply available as the specialist refrigerant FC-72, and is well documented in the literature.

To create this emulsion we had to overcome a number of challenges. Firstly, though hexane and FC-72 are miscible, the consolute point (the upper temperature for immiscibility below which the oil mix separates into its component oils) is close to room temperature for useful mixtures. We added a diblock copolymer F6H6 to reduce the consolute point and prevent accidental separation. The image below shows one of the hazards of leaving one of our emulsion samples in an unheated microscopy suite over the weekend.

This “frogspawn” image shows an emulsion in which the dispersed phase is a hexane-FC-72 mix, but because the emulsion temperature dropped below the consolute point the oil phase separated, giving a double emulsion comprising droplets of one oil inside droplets of the other. The dark droplets are the FC-72. They sink within the hexane droplets since FC-72 is denser than hexane and in the image the direction of gravity is down.

A dispersed phase mixture of 39% weight-for-weight n-hexane, 58% FC-72 and 3% F6H6 gives an emulsion that is closely refractive index and buoyancy matched. The emulsion is stabilized using Brij-35. The continuous phase contains copper sulphate as NMR dopant and sodium azide as preservative, and the oil phase contains the fluorescent dye Nile Red. The droplets are flocculated via the depletion mechanism through the addition of polyethylene oxide (PEO). The resulting aggregated emulsion system is sufficiently transparent and languid that we are able to extract full 3D images using confocal microscopy.

We originally planned to use cross flow membrane emulsification (XME) technique on the grounds that this would yield monodisperse emulsions, the idea being that monodisperse emulsions would be easier to handle at the image processing stage. Below is the droplet size distribution for an XME emulsion as determined using a Coulter counter:

The droplet size distribution for an XME emulsion as determined using a Coulter counter

Blue denotes the normalized size distribution for an emulsion created using the XME (n-hexane FC-72 mix). For comparison, the size distribution for an emulsion created using a blender is shown in pink. In this case the oil was Bromohexadecane, though the oil type is not crucial.

There are two points to note about this plot. Firstly, the XME emulsion does not give a simple “spike” distribution of a truly monodisperse emulsion. Secondly, the emulsion droplets are quite large compared to those in an emulsion created by blending. We found that the XME created emulsions with a relatively narrow size distribution, though not really monodisperse, and with a relatively large mean droplet diameter (of 6.0 µm in this instance). The restricted size distribution is helpful, however, and does mean that the system is sufficiently polydisperse to qualify as an “everyday” emulsion free of artefacts resulting from monodispersity.

It transpires that the XME is indeed essential but for an unexpected reason. Both n-hexane and FC-72 have high vapour pressures (boiling points of 69 ºC and 55 ºC respectively) and the high-energy high-shear environment of a typical blender causes these oils to vaporize. It is not possible to create a hexane-FC-72 emulsion in a blender. However, no such problems are encountered in the low-shear environment of the XME, which is therefore well suited to creating emulsions based on volatile components or rendered “fragile” for some other reason, perhaps because they are stabilized by proteins.

Working with this emulsion system we discovered that it has a tendency to foam at the surface. We believe that this foaming is linked to gas bubbles that form in the bulk emulsion, evidenced by the appearance of bubbles in some of our confocal images. We presume that the bubbles are gaseous oil, and are a result of the volatile nature of the oil. These bubbles can be used as natural probes of the droplet network elasticity (see information on Network Ripples). Bubble and foaming activity can be reduced by the use of gas-tight sample containers.