Mark Beilby noticed that the strong excitation available from the PZT squeezing device, along with the high sensitivity of the birefringence readout, made his apparatus eminently suitable for measurement of modal quality factors in the standard ``ring-down'' method as well. This observation was followed up with an set of measurements by him of modal Q factors of a 7940 fused silica cylinder, and an extensive survey by graduate student William Startin of the Qs of a set of fused silica blocks of various compositions, which had been kindly loaned to us by Corning, Inc.
The results of these measurements were both interesting and encouraging. The
most encouraging aspect of them was that for each fused silica sample, there
was at least one mode with a Q in excess of 1 
, and usually
several modes with quality factors almost as large. The largest Q we
observed in this series was 1.86 
, found
in a sample of 7958 fused silica. This is within a factor of 2 of the
highest Q ever reported for fused silica, 3 .[8] The
fact that these high Q values appeared in every sample, in spite of the
substantial differences in formulation, strongly suggests that the result is
generic. A review of the recent gravitational-wave literature by Startin
confirmed that others who used fused silica almost always found quality
factors at least in the high millions; by contrast, measurements made on
fused quartz (from naturally-occurring silica feedstock) found Q values at
most of a few million. This recognition by Startin was also being made
simultaneously by others around the world, as discussions in August with
Sheila Rowan (Glasgow), Stan Whitcomb (LIGO), and Luca Gammaitoni (Perugia)
revealed.
The other interesting aspect of the results was the wide scatter in Q from mode to mode in a given sample. The variation was much too strong a function of frequency to be due to intrinsic frequency dependence of the dissipation in the material; that can at most be of order a single power of frequency. There was also a pattern of which modes would show high or low Q, consistent with the variation being due to dissipation in the fine wire suspension used to support the sample. Thus these multi-mode measurements give a diagnostic of suspension losses, and also set a challenge -- good thermal noise performance in a gravitational wave detector will depend on achieving loss losses in many modes.
A paper summarizing this will soon be published (Oct 1998) in the journal Review of Scientific Instruments.
We will follow up this work by attempting to suspend fused silica samples in ways that reduce losses in the suspensions. Attachment of fused silica fibers by ordinary glass welding techniques and by silicate bonding[12] are two promising techniqes. We will also check the time-domain dissipation measurements on sapphire (described in the previous section) with multi-mode Q measurements. Here we will coordinate with the measurement program of the Glasgow group on the same set of samples.