http://hdl.handle.net/10012/11940

Microfluidic technologies have started to show their potential in assisting with the probes into the complicated mechanical-chemical interactions of multiphase fluids at microscale geometries (e.g., regular channels, porous media micromodels). The benefits of appropriately implementing microfluidics in such research efforts may include, but are not limited to: (1) small-dimensions facilitated analogous mechanical behaviors, (2) precise and reliable controls over relevant operating parameters of the fluids, (3) approximated reproduction of the hydrostatic or hydrodynamic circumstances, and (4) implementations of advanced visualization technologies such as microscopic imaging in order to reveal the dynamic processes involved in those multi-fluid interactions. Following the early studies on two-phase flow such as oil and water in microscale devices driven by the understanding of oil recovery process and mechanisms, carbon dioxide (CO2) has drawn increasing attention because of their environmental impact such as greenhouse gas effects. Most studies target either enhancing the chemical reactions by using pressurized CO2 as a solvent or revealing physical properties as well as mass transfer performance of gaseous CO2 in common hydrodynamic scenarios. However, dense CO2 including liquid and supercritical states are rarely touched, which is mainly due to the technical difficulties in working with extreme pressures (tens to hundreds times of atmospheric pressure) and elevated temperatures (> 31°C). Driven by the literature voids, this thesis presents some preliminary studies of the hydrodynamic issues and mass transfer of dense CO2 in a form of flowing segments in microchannels.
Prior to the commencement of any experimental work concerning dense CO2, a system capable of working at extreme pressures reliably and safely needs to be build first. Chapter 3 details the building of an experimental system which is dedicated to two phase microfluidic studies, especially for those related to extreme pressure/temperature conditions. Based on two principles of being extreme conditions durable and leakage free., a few goals, namely, reliability, flexibility and coordinability of this system are achieved.
The first part of this thesis (Chapter 4) presents an experimental study of a fluid pair, namely, liquid CO2 and deionized (DI) water, in a micro T-junction, where liquid CO2 and DI water are injected from the side and the main channel of the T-junction, respectively. Drop flow and co-flow are identified as two main flow regimes subjected to the various flow rate ratios applied. By focusing on the drop flow, a full period of liquid CO2 drop generation is divided into three stages, and each stage is meticulously described in terms of the interfacial developments (e.g., interface profile, pressures across the interface, size variations). The mass transfer mechanisms including CO2 hydration, diffusion on a perpendicular dimension of the interface and the advection parallel to the interface are considered and discussed in terms of their effects in CO2 molecules transport. An overall theoretical analysis of such mechanisms verifies that the transported CO2 portion is a small quantity compared with the bulk CO2 stream. Based on this verification, the generated liquid CO2 drop size, speed, and the spacing development within one drop generation period are probed. A formulation of drop size with the flow rate ratio shows a magnified effect of the later factor, which is interpreted by the extended time scale of an ‘elongating-squeezing’ stage of the period. Drop speed results show that they can be approximated by dividing the total flow rates over the channel cross-sectional area. And the speed differences between the generated drop and the emerging one in the T-junction lead to a model which details the spacing development within one drop generation period. The model is well validated by experimental results.
The second (Chapter 5) and the third part (Chapter 6) of the thesis are devoted to the investigations of hydrodynamics and mass transfer of liquid CO2 and scCO2 drops traveling simultaneously with water in a long straight microchannel (~15mm long), respectively. The production of such CO2 drops is realized by using the aforementioned micro T-junction. Distinctly, these studies focus on the drop size and drop speed at three specified positions of the channel and the mass transfer caused shrinkage of the CO2 drop quantified by the decreasing drop length. In order to calculate the mass transfer coefficient of CO2 drops, the detailed geometries of a Taylor drop in the square microchannel with a presence of wall films that separate the drop from the channel wall are considered, and consequently, the surface area and the volume of the drop are formulated based on the drop dimensions, channel geometries, contact angle and estimated film thicknesses. Furthermore, a specific mathematical model is developed to calculate the mass transfer coefficients based on the drop length reductions and drop flowing time in the channel. Discussions on these results indicate that surface-volume ratio and drop flow time are the two main factors in controlling the hydrodynamic shrinkage of the liquid CO2 and scCO2 drops. In addition, pressure declines of segmented flows in microchannels are considered and their effects are evaluated based on a pressure decline model and the Peng-Robinson equation of state (Eos) as well as the estimated initial pressures at the T-junction. Calculations of the resulted volume changes from the pressure declines show that the influences are small, and the observed CO2 drop shrinkage is confidently attributed to the mass transfer across the interface between the CO2 phase and water phase.
The last part (Chapter 7) presents a numerical study of the hydrodynamics of one single liquid CO2 drop and one single scCO2 drop traveling in a straight microchannel simultaneously with water as the carrier fluid. A two-dimensional (2D) computational domain of the straight microchannel is configured based on the experimental observations. Three liquid CO2 cases and three scCO2 cases are studied. It is found that the computed drop is disk-like shape encapsulated by thin films that separate the drop from the channel walls. The predicated film thicknesses agree very well with the literature. Besides, the flow domain within CO2 drops could be mainly composed of a few vortex regions, and small vortex regions at the front and the back cap of the drop start to vanish with increased velocities. Analysis of the mechanisms causing the vortexes is provided. The interfacial CO2 distributions of the drop show that both diffusion and local relative convection at the meniscus regions contribute to the concentration profile. Although no significant drop shrinkage is observed for typical Taylor drops, the one for the case with the highest capillary number (Ca ~ 10-2) defined from its pure diffusional gradient profile showed a similar development tendency over time as its experimental counterpart in Chapter 6.

http://hdl.handle.net/10012/11940