Investigation of effects associated with electrical chargingof fused silica test mass

V. Mitrofanov, L. Prokhorov, K. Tokmakov

Moscow State University

P. Willems

LIGO Project, California Institute of Technology

LIGO-G040097-00-Z

2

Introduction

Fused silica mirrors — dielectric bodies suspended in high vacuum

They can trap and store electrical charges

These charges interact through electrostatic couplingwith the environment and electrostatic actuators in particular

Result of this interaction may be:

additional loss degradation of mechanical Q additional thermal noise

accidental variation of electrical charge  q variation of Coulomb force additional noise

3

Mechanical loss in fused silica oscillators due to electrical charges or electric field

A number of experiments which have demonstrated the effects of electrical charge on mechanical loss:

• Univ. of Glasgow (Class. Quantum Grav., 14 (1997) 1537

• Moscow State Univ. (Phys. Lett. A., 278 (2000) 25

• MIT (Rev. Sci. Instrum., 74 (2003) 4840

The mechanism of loss is not clear (only hypotheses were proposed)

Nevertheless it is likely that the additional thermal noise associated with charges is not dangerous for Adv. LIGO if the electrical charge on mirrors is not unduly large

4

Experimental setup Vacuum p < 10-7 Torr

All fused silica bifilar pendulum:

Mass M = 0.5 kg, Fibers: L = 25 cm, d =200 m,

Torsion mode f 1.14 Hz,

Quality factor Q 8107 ,

Relaxation time * 2.2107 sec,

Initial amplitude A 0.07 rad

Multistrip capacitive probe (two sets of gold strips sputter-depositedon fused silica plate) connected with high impedance amplifier.

Probe voltage U = kqA

q – electrical charge on the pendulum(distribution of charge is unknown)

5

Long-term measurements of probe signalin process of free decay of the cylinder oscillation

About 2 years of observation a lot of runs ( each run is about 30 days )

Averaged rate of charge variation: 104 e/cm2 day (assuming uniform distributionof charge on the end faceof the cylinder)

Charge variation corresponds to negative charging of the cylinder 

Resolution in measurement of charge is limited by seismic noise and parasitic effects

1.4

1.5

1.6

1.7

1.8

1.9

0 2 4 6 8 10 12 14 16 18 20 22 24 26 28 30 32 34

tim e, days

sign

al f

rom

pro

be,

arb

itra

ry u

nits

Tim e dependence of voltage from the probe

6

- - - - - - - - - -

Why we continue to study behaviorof electrical charges on fused silica test masses?

1) 2)

mass-q+q

d

Conductive plate

q=108e

q= 104e

D=3 mm

mass

High energy particle

Surfacelayer

Surfacecharge

•Generation of mobile charges within surface layer of the material

•Their separation in electric fieldof surface charges

gr11

20

el FN10d8π

δqqδF

7

Rare big jumps of signal from the probe synchronous with fast decrease of the cylinder amplitude

4 events in 12 runsof measurements

multistep structure of jump

Initial charge density: of order of 106 --107 e/cm2

After the jump:of order 108 e/cm2

(assuming uniform surface free charge density)

Change of amplitudein the process of jump corresponds to damping Q-1 of order of 10-4

Resolution of details is determined by averaging time – 70 sec

1.8

2

2.2

2.4

2.6

2.8

3

am

plit

ud

e,a

rbitr

ary

un

its

0 10 20 30 40

tim e, days

3 0

4 0

5 0

6 0

7 0

8 0

volta

ge

fro

m p

rob

e, a

rbitr

ary

uni

ts

0 10 20 30 40

tim e, days

Tim e dependence of amplitude of the cylinderand voltage from the probe

8

Behavior of the system after the first jump

After the first jump:

Repeated increase of torsion amplitude by means of electrostatic excitation results in a new cascade of jumps of amplitude and voltage from the probe (spoiled state of the system“fused silica pendulum nearby electrodes”)

With time elapsed from the first jumpfast changes of the probe’s voltage decrease. Spoiled state transfers to the original state within relaxation time of order of one month

4 0 8 0 1 2 0 1 6 0 2 0 0

tim e, hours

1.8

2

2.2

2.4

2.6

2.8

3

amp

litud

e, a

rbitr

ary

units

4 0 8 0 1 2 0 1 6 0 2 0 0

tim e, hours

1

2

3

4

5

6

pro

be s

igna

l, ar

bitr

ary

uni

ts

First jump

Excitation (increase)

of amplitude

Excitation (increase)

of amplitude

Amplitude of the cylinder

Voltage from the probe

9

Possibility that the pendulum modesof the cylinder causes it to touch the electrode

• A contact electrification can take place if the pendulum is swingingfar enough to touch the electrode plate due to seismic excitation

• Only torsional amplitude is well measured by the optical sensor

The electrometer signal on the pendulum frequencies give information about pendulum amplitudes with large uncertainty

So we can not control the touching or close approach of the pendulum to the electrode plate in this set up

10

Changes of torsional amplitude and probe signal in the case of provoked touching

For the provoked touching behavior of the system is much the same as it is in the case of our “jumps”: decrease of amplitude and increase of probe signal as well as transfer to the “spoiled” state

There are at least two distinctive property of provoked touching:

• In the case of provoked touching change of amplitude and probe signal is fast (within the time resolution) 2.72

2.76

2.8

2.84

2.88

2

2.4

2.8

3.2

3.6

4

0 20 40 60

tim e, m in

Change of torsional amplitude in the case of "jump" (black) and of touching (red)

11

Changes of torsional amplitude and probe signal in the case of provoked touching

• Admission of small portion of air into the chamber results in significant drop of probe signal (supposedly due to gas breakdown). After the “jump” the probe signal can be reduced only by electrical discharge in rough vacuum

These facts can not be the evidence for the absence of touching in our experiments. It is worth further investigation 0

0.1

0.2

0.3

0.4

0.5

sign

al f

rom

pro

be,

arb

itra

ry u

nits

0 20 40 60

tim e, m in

Change of probe signal after admission of air

12

Search for correlation between signals fromthe probe and cosmic ray detectors

Cosmic ray detector system

11 particle detectors (plastic scintillator paddles 180180.8 cm3 ) suppliedby photomultipliers were installed around the vacuum chamber with the pendulum

Selection of events with maximum summarized voltage from all paddles

Vacuumchamber

Scintillationdetectors

ADC

Threshold setting

Computer

Peak detectors

Wall

13

Result of measurements

We have not found statistically significant time coincidences between large signals from detectors and jumps of charge and amplitude in the “spoiled” state of the test mass

But we can not exclude correlation with:

low energy particles passing locally through the surface layer of the test mass

1 2 1 4 1 6 1 8

tim e, hours

3 2

3 6

4 0

4 4

4 8

5 2

prob

e volta

ge (q)

2

2.2

2.4

2.6

2.8

amp

litud

e (A

)

0

5

10

15

20

25

cosm

ic r

ay d

ete

ctor

sig

nal (

CR

)

A

q

CR

Fragment of record of am plitude and signalsfrom the probe and scintillation detectors

14

Conclusion

• We have a number of empirical facts but we can not

unambiguously interpret these facts yet

• We suppose that the electronic properties of the surface of

dielectric mirror can play the important role but they are poorly

studied till now

• We do not yet ready to say that electrical charges are not

dangerous for LIGO