Influence of Silicate Bonding on GEO600 Thermal Noise

1
Influence of Silicate Bonding on GEO600 Thermal
Noise
Joshua Smith, the GEO600 team, Albert-Einstein-Ins
titute, Syracuse University, Glasgow University,
Stanford and MIT
2
GEO600 optical layout
3
GEO600 sensitivity
4
Thermal noise decomposed
Thermal Noise sets a fundamental limit on how
still the mirrors that we are measuring as test
masses can be.
Suspension Thermal Noise
Internal Thermal Noise

• Motion of mirror surface due to internal thermal
  fluctuations
• Resonances lie above GEO detection band
  (fundamental 11 kHz)
• Proportional to (mirror loss)1/2

• Motion of mirror due to thermal fluctuations in
  pendulum
• Most resonances at low frequency except fiber
  violin modes at n650 Hz (n 1, 2,)
• Proportional to (pendulum loss)1/2

5
Thermal noise and loss

• Fluctuation-Dissipation Theorem thermal
  fluctuations related to energy dissipation
• Equipartition theorem
• low loss low off-resonance TN

Hz
6
Quantifying loss

• Loss Factor
• ? 2 ? ET / Etot
• Quality Factor
• Q ? fn ? ?-1
• ? time for A0 to decrease to A0/e

7
Monolithic suspensions
8
Silicate bonding

• room temp. chemical bond developed at Stanford,
  practiced at Stanford, Glasgow
• High-vacuum compatible
• Bonds fused silica and sapphire
• strength bulk fused silica
• No evidence of strength degradation with time

• Will it add additional loss to the GEO Test
  Masses?

9
Ringdown measurements
10
Ringdown results
Qcontrol 1x107
11
Analysis estimating ?bond

• Energy ratio Ebond / Etotal calculated using FEA
  (MIT)
• Loss of bonded rod decomposed into loss in the
  bulk and loss in the bond (using ? bulk
  Qcontrol 1)
• ? bondedrod (Ebulk / Etotal) ? bulk (Ebond
  / Etotal) ?bond
• Using (Ebulk / Etotal) ? 1 and (? bondedrod - ?
  bulk ) ? ? bondedrod we solve for the loss
  factor of the bond material
• ? bond ? (Etotal / Ebond) ? bondedrod
• For these samples ? bond ? 1
• Does not take Youngs modulus (Glasgow) into
  account. Glasgow calculates lower ? bond for
  samples with different geometry

12
Applying ?bond to GEO600 test mass

• Loss of a GEO test mass can be expressed as a
  sum of losses
• As worst case, assume ?bond 1. With no other
  loss sources and assuming energy equally
  distributed in test mass volume, loss due to the
  bond sets the limit Qeff ? 8x107
• bulk Q of FS has been measured as high as 2x108
  (Syracuse), so in worst case loss due to bonds
  could be greater than loss due to bulk
• Q of internal modes of GEO test masses recently
  measured to be as high as 4x106. Well below lower
  limit set by Qeff. Limited by other loss source,
  possibly recoil loss up the pendulum chain.