Transcript Advanced Virgo: Wedges vs Etalon

Advanced Virgo:
Wedges vs Etalon
Stefan Hild, Andreas Freise
University of Birmingham
Roland Schilling, Jerome Degallaix
AEI Hannover
January 2008, Virgo week, Pisa
Motivation:
Input mirror without wedge
 Initial Virgo has no wedges in the input mirrors
 The etalon effect could be used for adjusting the cavity
finesse (compensating for differential losses)
 If etalon effect is not controlled it might cause problems
Stefan Hild
Virgo week January 2008
Slide 2
Motivation:
Input mirror featuring a wedge
 Used by initial LIGO
 Reflected beams from AR coating can be separated from
main beam => pick off beams provide additional ports for
generation of control signals.
 No etalon effect available.
Stefan Hild
Virgo week January 2008
Slide 3
What to use for Advanced VIRGO?
Etalon or Wedges ??
 For AdV possibility to adjust cavity finesse gets more
important (higher cavity finesse, DC-readout).
 For AdV possibility to create more and better control
signals seem desirable.
Is there a possibilty to have both for Advanced Virgo ??
Stefan Hild
Virgo week January 2008
Slide 4
Advanced Virgo:
waist inside the cavity
 Increase beam size at mirrors => reduce
thermal noise contribution of the test masses.
 Move beam waist away from input test mass
Is there still an etalon effect in the (flat/curved) input mirror ?
Stefan Hild
Virgo week January 2008
Slide 5
Etalon effect:
flat/flat vs curved/flat
 Flat/flat etalon:
 Perfect overlap of wavefronts
 Curved/flat etalon:
 Mismatch of wavefront curvature
 Fortunately mirror curvature of a
few km is not so far from “flat”.
 Simulations show: a reduced
etalon effect in curved/flat input
mirror is still present
Stefan Hild
Virgo week January 2008
Slide 6
Etalon effect:
flat/flat vs curved/flat
 Flat/flat etalon:
 Perfect overlap of wavefronts
 Curved/flat etalon:
 Mismatch of wavefront curvature
 Fortunately mirror curvature of a
few km are not so far “flat”.
 Simulations show: a reduced
etalon effect in curved/flat input
mirror is still present
Stefan Hild
Virgo week January 2008
Slide 7
IDEA: Wedges at input mirrors and
etalon effect at end mirrors
 Wedge at input mirrors:
 Allows for additional pick off beams
 (Concentrate on compensating thermal lensing in input mirror)
 Use etalon effect at end test mass
 Replace AR-coating by a coating of about 10% reflectivity.
 Ideally use a curved back surface (same curvature as front).
 End mirror behaves similar to flat/flat etalon.
Stefan Hild
Virgo week January 2008
Slide 8
Stefan Hild
Virgo week January 2008
Slide 9
Starting with a single
AdV arm cavity
 Using a single AdV arm cavity
(no IFO).
 Figure of merrit = intra cavity
power.
 Parameters used:
IM trans = 0.007
IM loss = 50 ppm
EM trans = 50 ppm
EM loss = 50 ppm
AR coatings = 0ppm
IM curvature = 1910m
EM curvature = 1910m
Input = 1W
Stefan Hild
Parameters taken from these 2 documents:
Virgo week January 2008
Slide 10
Influence of losses inside the cavity
 Imperfection of optics (surface + coatings) might cause
different losses in the arm cavities := differential losses.
 What are the
expected differential
losses of AdV ?
5ppm? 50ppm?
 A differential loss of
15ppm corresponds
to a change of 2W
intra cavity power in
this example.
 GOAL: Compensate
2W using etalon
effect in end mirror.
Stefan Hild
Virgo week January 2008
Slide 11
End mirror as curved etalon
(optimal solution)
 Simulation done with
finesse.
 Back surface of end
mirror curved (1910m).
 AR coating replaced
by coating of 10% or
20% reflectivity.
 R=0.1 allows
adjustment range
of 10W (65ppm)
 R=0.2 allows
adjustment range
of 16W (95ppm)
Stefan Hild
Virgo week January 2008
Slide 12
Etalon changes optical phase
 When changing the etalon tuning the optical-phase changes
as well. (noise!)
 The two etalon surfaces build a compound mirror, whose
position depends on the etalon tuning.
 A single FSR of the etalon corresponds to about 3pm.
Stefan Hild
Virgo week January 2008
Slide 13
Stefan Hild
Virgo week January 2008
Slide 14
Optical design: Check system
integrity for deviations from specs
 A deviation in the
refelctivity of the etalon
coating:
 Only changes tuning range (no
problem)
 A deviation in the curvature
of the etalon surface:
Analyzing “worst case”
scenario: curved/flat etalon
 Imperfect wave front overlap…
 Reduces tuning range …
 Beam shape distortions …
Stefan Hild
Virgo week January 2008
Slide 15
FFT-simulation of a
flat/curved etalon
 Using R. Schilling’s WaveProp,
(http://www.rzg.mpg.de/~ros/WaveProp/)
 Parameters:
 Field: 128x128
 Computing 3000 roundtrips
 End mirror front:
 50ppm transmission
 R_c = 1910m
 End mirror back:
 10, 20, 50% reflectivity
 R_c = flat
Stefan Hild
Virgo week January 2008
Slide 16
WaveProp simulation
 R=0.1 allows
adjustment
range of 3W
(20ppm)
 R=0.2 allows
adjustment
range of 5W
(35ppm)
 R=0.5 allows
adjustment
range of 9W
(60ppm)
Stefan Hild
Virgo week January 2008
Slide 17
Comparison of flat/flat and
curved/flat etalon
For a curved/flat etalon the tuning range (etalon effect) is
reduced by about a factor of 3.
Stefan Hild
Virgo week January 2008
Slide 18
Comparison of WaveProp and
finesse simulations
WaveProp
finesse
Waveprop and finesse are in excellent agreement.
Stefan Hild
Virgo week January 2008
Slide 19
Simulated beam inside arm cavity
Curved/curved etalon
Curved/flat etalon
 Simulated beam shape inside the arm cavity (using finesse)
 Simulation done with etalon on resonance (worst case scenario)
 By eye: no change …
Stefan Hild
Virgo week January 2008
Slide 20
Changes the beam shape inside
the arm cavity ??
 Subtracted beams
indicate the change in
beam shape.
 Residual light field is a
factor of 60000
weaker than the intra
cavity beam.
Stefan Hild
Virgo week January 2008
Slide 21
Summary
 Advanced Virgo CAN feature
wedges in the input mirrors AND
use the etalon effect at the end
mirrors.
 Proposed concept allows us to
build ‘arm cavities with adjustable
losses’.
 A curved/curved etalon would be
ideal (a curved/flat etalon should
work as well, but with reduced
tuning range).
Stefan Hild
Virgo week January 2008
Slide 22
Outlook
Potential issues to be investigated:
 How does misalignment of the etalon influence the alignment
control signals:
 For curved/curved etalon: probably fine (to 1st order)
 For curved/flat etalon: needs simulation
 Check that optical-phase noise from fluctuations is no
problem.
 Need a control system for etalon tuning (error signal +
actuator).
 Need a value for the expected differential losses in AdV in
order to choose the reflectivity of the etalon
Stefan Hild
Virgo week January 2008
Slide 23
END
Stefan Hild
Virgo week January 2008
Slide 24
It is important to compensate the
differential losses
AdV simulation (dummy parameters)
A differential loss of 70ppm causes already 100mW of waste
light at the dark port.
 increased shot noise !!
Stefan Hild
Virgo week January 2008
Slide 25