Post on 20-Jan-2016
description
1
Test Mass Suspensions for AIGO
Ben Lee
The University of Western Australia
Australia-Italy Workshop 3rd-6th Oct 2005
Ben Lee 2
Introduction
Thermal noise in interferometers.Reducing the thermal noise: what we know so far.
Suspensions for AIGO.Removable modular suspensions.Reducing violin mode Q factors.
Australia-Italy Workshop 3rd-6th Oct 2005
Ben Lee 3
Thermal Noise
10-22
10-20
10-24
10-26
100 101 102 103 104
Thermal Noise
1st Generation Sensitivities
Advanced Sensitivity
The thermal noise must be reduced below the quantum noise limit.
Frequency (Hz)
H(f
) / (
Hz)
1/2
Australia-Italy Workshop 3rd-6th Oct 2005
Ben Lee 4
Suspension Thermal Noise
101
102
103
104
10-28
10-26
10-24
10-22
10-20
10-18
10-16
Quantum NoiseTest Mass Thermal NoiseSusp. Thermal NoiseTotal Noise
H(f
) / (
Hz)
1/2
Frequency (Hz)
Low thermal nose achieved with very high Q modes.
Suspension modes may also affect control systems.
Australia-Italy Workshop 3rd-6th Oct 2005
Ben Lee 5
Low Loss Materials
4nn
222n
2nn
2n
B2 L
Tk4x
From the dissipation dilution theorm:
Thus, materials with low Φ are better. Fused Silica: Φ~3×10-8 [1]Silicon: Φ~2.8×10-8 [2]Sapphire: Φ~3.7×10-9 [3]
There is much more to consider….Thermoelastic loss
(300K)
1. A.M. Gretarsson, G.M. Harry. Rev. Sci. Instrum. 70 (1999) 40812. J. Ferreirinho in: D.G.Blair (Ed), The Detection of Gravitational
Waves, Cambridge Universtiy Press, Cambridge, 1991.3. S. Rowan, et. al. Phys. Lett. A. 5 (2000) 265
References:
Australia-Italy Workshop 3rd-6th Oct 2005
Ben Lee 6
0 1 2 3 4
-7
-6
-5
-4
0 1 2 3 4
-7
-6
-5
-4
Thermoelastic Loss
2th
1
v
2
c
TE
Thermoelastic loss presents a significant frequency dependance to the loss value, Φ.
K
tc2
2v
K32.4
dc 2v
,for fibres
,for ribbons
where
and
1 10 100 1000 10000Frequency (Hz)
10-4
10-5
10-6
10-7
Loss
ang
le,
Φ
Sapphire
Niobium
Fused Silica
Silicon (All 200µm thick)
Thinner ribbon
Australia-Italy Workshop 3rd-6th Oct 2005
Ben Lee 7
Fibers vs Ribbons.
2mgL
EI
2
1
Dissipation dilution factor:•The ratio of restoring force supplied by bending elasticity to the restoring force supplied by tension.
•This phenomena has a significant effect on pendulum mode and violin mode Q factors.
FT
Fb
FT
Fb
The effective loss factor
This value can be lower for ribbons compared to fibres with similar strength.
Australia-Italy Workshop 3rd-6th Oct 2005
Ben Lee 8
AIGO suspension
Sensitivity: 2nm/Hz1/2 (100Hz bandwidth)
Force: 20µN (100Hz bandwidth)
Control mass 30kgTest mass 4.2kg
Removable Modular suspensions.Peg in Hole interface.
Ribbon Suspension
Australia-Italy Workshop 3rd-6th Oct 2005
Ben Lee 9
Peg – Hole Interface
1500.164E+07
.328E+07.492E+07
.656E+07.820E+07
.984E+07.115E+08
.131E+08.150E+08
NODAL SOLUTION
STEP=1SUB =1TIME=1SEQV (AVG)DMX =.752E-05SMN =1472SMX =.350E+08
1500.164E+07
.328E+07.492E+07
.656E+07.820E+07
.984E+07.115E+08
.131E+08.150E+08
NODAL SOLUTION
STEP=1SUB =1TIME=1SEQV (AVG)DMX =.752E-05SMN =1472SMX =.350E+08Test mass
Ribbon
Peg
High pressure contact points
Result:Stress at contact points approaches yield strength of Nb
FEM model of pegs
Australia-Italy Workshop 3rd-6th Oct 2005
Ben Lee 10
Will localised losses at the hole surface significantly affect the thermal noise?
Holes in the test mass
Question…
Model and results courtesy of S. Gras (ACIGA)
FEM model of test mass.Homogeneous structural loss Φs=10-8
Define the extra lossy elements.Apply Φs=10-2 to 10-8
Thermal noise observed by a gaussian beam
Structural losses only
Include non-homogeneous losses Φh=10-2
Australia-Italy Workshop 3rd-6th Oct 2005
Ben Lee 11
Holes in the test mass
Model and results courtesy of S. Gras (ACIGA)
Contour values, Ψ:
Sh+s (spectral density of inhomogeneous model)
Ss (spectral density of homogeneous model)Ψ=
10% increase line
For AIGO interferometer with r0 = 20mm, - Φh cannot exceed 4×10-3
Australia-Italy Workshop 3rd-6th Oct 2005
Ben Lee 12
Reducing Violin Mode Qs
It has been reported by Goßler et. al. [1] the need to reduce the Q factor of the fundamental and first harmonic violin mode.
The purpose is to prevent interference with interferometer length control servo.
This is achieved by adding lossy coatings.
Reference:1. Class. Quantum Grav 21 (2004) S923-S933
Fused silica fibre
Teflon coatings
Australia-Italy Workshop 3rd-6th Oct 2005
Ben Lee 13
Reducing Violin Modes
The Orthogonal Ribbon can reduce violin modes and Q factors.
High Contact Peg
End Flexure
Orthogonal Section
End flexures provide similar pendulum mode Q factors.
Orthogonal ribbon section exhibits fewer, lower Q violin modes in the critical direction.
Pendulum mode freq. and dilution factors
Normal Ribbon
Orthogonal Ribbon
x
y
fpend (Hz) dilu
x
y
0.91
1.12
2.8×10-3
0.27
0.91
0.92
3.1×10-3
3.0×10-3Little difference between normal and orthogonal ribbon.
Australia-Italy Workshop 3rd-6th Oct 2005
Ben Lee 14
Reducing Violin Modes
2
2
4
4
2
2
t
t,zY
z
t,zYEI
z
t,zYT
The violin modes for the orthogonal ribbon can be calculated by solving the beam equation:
Direction of laser field
x direction modes y direction modes
Test Mass
Australia-Italy Workshop 3rd-6th Oct 2005
Ben Lee 15
Reducing Violin Modesm
/Hz1
/2
Expected Suspension Thermal Noise Comparison,
Normal Ribbon and Orthogonal Ribbon
100
101
102
103
10-22
10-20
10-18
10-16
Frequency (Hz)
10-14
10-12
10-10
10-8
Normal RibbonOrthogonal RibbonSimilar low
frequency thermal noise
Lower number, and lower Q factor x direction violin modes.
Small increase in thermal noise, due to lower violin mode Q.
Australia-Italy Workshop 3rd-6th Oct 2005
Ben Lee 16
Conclusion
Removable modular suspension can be achieved with only a slight increase in test mass thermal noise.
Lowering all the violin mode Q factors can be achieved with an orthogonal ribbon.
The orthogonal ribbon has little effect on pendulum mode thermal noise.
AIGO facility can be used to test the practicality of the suspensions presented.