(315b) Investigations Of The Coalescence Phenomena In Liquid-Liquid Taylor-FLOW

The general advantages of micro reaction engineering like a high surface to volume ratio and a resulting efficient high heat- and mass transport can also be observed by using Taylor-Flow a certain flow regime which occurs by combining two immiscible fluids in a capillary. Taylor-Flow can be used for chemical operations like (reactive) extraction. The type of Taylor-Flow differs with the used fluids; there is gas-liquid, liquid-liquid or even liquid-liquid-solid Taylor-Flow. The adjustment of different slug lengths is well investigated in gas-liquid Taylor-Flow both as numerical simulation [1] or even experimentally concerning capillary diameter [2], mixer geometry [3], liquid viscosity [4] and many other parameters. Numerical work has already been done in liquid-liquid Taylor-Flow but capillary lengths in these investigations were always shorter than a few meters. The undesired phenomenon of coalescence which describes an elongation of slugs along the capillary normally occurs at higher capillary lengths. To keep the extraction performance on a high level this coalescence must be avoided.

The aim of the ongoing research is to find parameters which are related to the formation of slugs in Taylor-Flow and their coalescence. This information could be very useful in following extraction experiments.

The used model system of water and 1-octanol is a common system in chemistry to determine the lipophilic and hydrophilic properties of a substance. For the investigation on the stability in liquid-liquid Taylor-Flow in consideration of different parameters a modularly set-up was built. This set-up made changes of the mixer and the capillary easy. To determine the resulting slug length distribution several sensors were tested. One was produced by ultrasonic hot embossing at KEµ RWTH Aachen University [5] the others were commercially available. The commercial sensors work with detection of light transmission and the newly produced sensors measures the electric conductivity directly inside the capillary.

From the commercially available sensors a laser light beam sensor was found to be the most suitable. It detects slugs with a length of at least 0.5 mm. For the detection with the laser light beam it was necessary to dye the aqueous phase with methylene blue which causes no significant change in surface tension or density of the phase. This was not necessary for the sensor from KEµ RWTH due to the measurements via electric conductivity.

To get a deeper understanding of the slug generation in Taylor-Flow many test parameters were defined. The variation of the total volume flow, the volume flow ratio, the material, diameter and length of the capillary and the resulting pressure drop were analyzed first in the research. Following the expectations, the capillary diameter and the volume flow parameters had the strongest impact on the resulting flow pattern.

In flows with a wide slug length distribution coalescence could be observed which could be caused by the different velocities of the slugs with differing lengths. To support this hypothesis a new set-up which can produce a heterogeneous slug length distribution was built. The research is still ongoing.

As already seen in gas-liquid Taylor Flow [2] the resulting Taylor-Flow is much more stable in smaller capillary diameters. This leads to higher adjustable volume flows until the flow regime changes from Taylor-Flow to a non-consistent flow. The influence of the capillary material was negligible which could be explained by the quite even properties of the both materials, fluorinated ethylene propylene (FEP) and perfluoroalkoxy (PFA). The strongest impact on stability of Taylor-Flow was observed by varying the volume flow and the water/1-octanol ratio. A decrease in stability could be observed at high octanol to water values wherein the increase of the water flow had no strong influence beyond a critical value.

For further research the focus will be on the mixer, especially its geometry and the orientation of the outlet. Also the geometry of the capillary will be further investigated.

[1] R. Gupta et al., J. Comput. Multiphase Flows 2010, 2, 1.

[2] J.W. Coleman, S. Garimella, Int. J. Heat Mass Transf. 1999, 42, 2869.

[3] A. Leclerc et al., Chem. Eng. J. 2010, 165, 290.

[4] S. Waelchli, P. Rudolf von Rohr, Int. J. Multiphase Flow 2006, 32, 791.

[5] W.K. Schomburg et al., Micromachines 2011, 2, 157.