Date of Award

Spring 2001

Document Type


Degree Name

Doctor of Philosophy (PhD)


Biomedical Engineering

First Advisor

Roy Schubert


An alternative way of blood oxygenation is proposed by introducing small bubbles directly into the blood stream. To be used for blood oxygenation, the bubble should be small enough to pass through the capillary beds. To explore the feasibility of producing small bubbles, bubble formation phenomena in a crosscurrent liquid flow in a silicon microchannel were studied. Small orifices with 4 μm, 4.2 μm, and 6.6 μm hydraulic diameter were machined in a direction perpendicular to a trapezoidal channel that has a hydraulic diameter of 42 micrometers. These orifices were connected to three different chambers of 2.5 × 10−7 cm3, 5 × 10 −7 cm3, and 10 × 10−7cm3 in volume. The glass plate was anodically bonded to the silicon. Water and oxygen were supplied to the channel and orifice through the chamber, respectively.

Between the liquid flow rate of 1 to 3 ml/hr and gas flow rate of 0.036 to 0.072 ml/hr very uniform bubbles, sizes of about 45 micrometer in diameter, were produced from the orifice with smallest chamber volume. In these ranges, the regression model shows that the bubble size has a weak relation to the gas and liquid flow rate. As the chamber volume gets bigger, the bubble sizes are increased and eventually become non-uniform.

From the experimental results, it is concluded that the chamber volume plays a significant role in determining bubble size. A mathematical model of appropriate phenomena was validated by comparison to experimental data. The data and the validated model suggest a simpler design and model. This model, with no chamber volume suggests that the bubble size can be decreased further if the orifice and crosscurrent liquid flow channel diameter can be decreased without causing pressure increase in the channel. This model predicts it is possible to produce bubbles of 16 micrometer in diameter at a rate of about 10,000 Hertz.