Date of Award

Winter 1999

Document Type


Degree Name

Doctor of Philosophy (PhD)



Committee Director

James Cox

Committee Director

Viet Nguyen-Tuong

Committee Member

Gary E. Copeland

Committee Member

Rocco Schiavilla

Committee Member

J. Mark Dorrepaal

Committee Member

Larry Phillips


This experimental study has revealed and investigated many of the physical issues that affect accurate measurement of the surface resistance for small samples consisting of superconducting niobium films on copper substrates. It is believed that this work provides the groundwork for future research directed towards solving this important problem. Accurate measurement of surface resistance for such samples is needed to allow the rapid evaluation and optimization of the deposition parameters necessary for manufacturing low-loss superconducting niobium films.

A superconducting niobium triaxial cavity was investigated to determine its suitability for measuring the residual surface resistance of copper samples that were sputter-coated with niobium. Apiezon and GE varnish were found to be inadequate to bond the sample to the endplate of the test chamber. A new method using a screw bonded to the sample and a threaded hole in the endplate was designed to solve these problems. However, electromagnetic simulations indicated that losses from the backside of such a sample might be significant for low-loss films. These effects were investigated using in-situ tuning of the cavity's frequency for samples coated alternately on one and two sides. The surface resistance was found to be frequency-dependent and higher than anticipated.

Bulk niobium samples were tested to eliminate the possibility of backside losses. These samples also gave high surface resistances. Experiments were performed to assess losses in the bonding agent between the screw and sample and to assess losses in the screw. The thermal impedance between the sample and endplate was also investigated. These possibilities did not explain the observed losses.

To eliminate extraneous losses from the installation method, bulk niobium samples were bonded to the endplate with GE varnish. The loss was found to be dependent on the frequency and higher than anticipated. Localized losses at the edge were discovered.

The endplate flange was redesigned to allow the entire endplate to be replaced to eliminate the possibility of edge losses. While lower surface resistances were achieved with this method, they were still much higher than anticipated. These results indicate that some anomalous loss mechanism is present even in the absence of edge losses.