Document Type

Restricted Campus Only

Publication Date

4-24-2012

Abstract

It was requested by the Department of Engineering Science at Trinity University, on behalf of Dr. Joshua Schwartz, to research and develop a sensor able to identify and characterize substances using controlled amounts of microwave energy. Primarily, the chamber must hold 3 microliters, interface with a Vector Network Analyzer, and be transparent.

It was determined that a buried microstrip would be the most ideal structure. It is simple to fabricate without sacrificing accuracy of measurement. The buried microstrip is a design that consists of the micros trip surrounded by a single dielectric. There is a baseplate below the dielectric which provides the grounding to which the generated electric field can travel to. The dielectric is the material through which the electric field permeates and that holds the test fluid. The microstrip is the trace guiding the microwave through the dielectric. The top layer is the portion of the dielectric to which the microstrip is attached.

Research was conducted into what materials would meet the design constraints. Aluminum was the final choice for the base and for a mold to form the dielectric. Aluminum is an inexpensive, lightweight, and easily-machined substance that can provide defined edges when used as a mold. Polydimethylsiloxane, or PDMS, was chosen as the dielectric. PDMS has a low loss tangent in the microwave range, is easily shaped, and is transparent. Copper was chosen as the microstrip material. It is readily available, easy to use, and can be cut in desired configurations.

In order to connect to the Vector Network Analyzer (VNA), an end launch connector was chosen. The connector provides a press fit and is optimized for signals at or below 40 GHz.

Simulations were performed with the software ANSYS High Frequency Structural Simulator (HFSS). Simulations were performed using the materials specified above, and the results show that the total loss is inside of acceptable range of 20 dB.

HFSS was also used to ensure proper functioning of the sensor as well as confirm that the mathematical calculations needed to obtain the permittivity from the measured s-parameters were correct. The results from these tests showed that the method used to obtain the dielectric permittivity did indeed work and can be used for the fluid identification tests. These tests also established the acceptable range of relative permittivities which could be tested using the mathematical model.

A final design was constructed with two varying lengths of the fluid chamber. The different lengths were needed for the calibration of the sensor so only the test fluid, and not the entire sensor, will be measured. The baseplate was made into a rectangular shape which enabled it to be used with both lengths, but different sections of microstrip and dielectrics were used for each of the lengths.

Using the final design with the fluid chamber, a second test was planned to guarantee that the sensor could accurately identify ethanol, but this was not performed due to time constraints.

Comments

Advisor: Dr. Joshua Schwartz

Download

Share

COinS