Part I: Bubble dynamics
The investigations of bubble formation dynamics at a cross-junction in a straight milli-channel were reported. The bubble formation process could be divided into the filling and squeezing stages, and their frequencies were compared at various conditions. It was found that the filling and squeezing frequencies were controlled mainly by the gas and liquid superficial velocities, respectively. The bubble formation frequencies could be related to the gas-liquid superficial velocity ratios, the liquid superficial velocities, and the bubble length. The bubble formation process was then analyzed considering the length of the gas finger right after the bubble pinch-off. Two patterns were identified depending on whether the pinch-off occurred inside of the cross-junction or not. The transition between these two patterns was described by a critical liquid Capillary number. Furthermore, the squeezing and dripping patterns were distinguished as the gas finger could fully block the channel or not. The transitions between the bubble formation patterns were determined by the gas Weber number and liquid Capillary number. Finally, the bubble length, the liquid slug length, and the bubble length normalized by the unit cell length could all be predicted based on the gas-liquid superficial velocity ratios η0.
Part II: Gas-liquid mass transfer
Using the resazurin-based colorimetric technique and advanced image acquisition, the equivalent oxygen concentration fields inside the liquid slugs were measured during and after the bubble formation stage at a cross-junction in a straight millimetric channel. Firstly, two different mechanisms were identified for the development of oxygen concentration fields depending on the two-phase Reynolds numbers (ReTP). Under low ReTP, a jet-like central oxygen concentration ‘finger’ occurred between the newly formed bubble and the gas finger at the bubble pinch-off point. Right after the bubble pinch-off, the dissolved oxygen was transported first by the entering liquid from two side inlets, and later by the developing recirculation loop inside the liquid slug. Under higher ReTP, two highly concentrated oxygen concentration spots were formed near the bubble rear and the channel wall region, and a much more complex flow structure in the liquid slug appeared. Then, whatever the operating conditions, it was observed that the averaged oxygen concentrations inside the liquid slug followed a nearly linear relation as a function of the axial position in the channel, and that the related mass flux density decreased as far as the bubbles flowed along the channel’s length. The bubble formation process could be decomposed into three stages when considering the temporal evolution of the cumulated oxygen mass and depending on the development of the liquid slug. At last, the overall volumetric liquid side mass transfer coefficients (kL a) were deduced from the concentration fields, and found to linearly increase with the recirculation frequencies, leading to a scaling law.