Transcript G050522-00 - DCC

The Mesa Beam
Juri Agresti1,2, Erika D’Ambrosio1, Riccardo DeSalvo1,
Danièle Forest3, Patrick Ganau3, Bernard Lagrange3,
Jean-Marie Mackowski3, Christophe Michel3,
John Miller1,4, Jean-Luc Montorio3, Nazario
Morgado3, Laurent Pinard3, Alban Remillieux3, Barbara
Simoni1,2, Marco Tarallo1,2, Phil Willems1
LIGO-G050XXX-00-R
1. Caltech/ LIGO
2. Universita’ di Pisa
3. LMA Lyon/ EGO
4. University of Glasgow
Gingin’s Australia-Italia workshop on GW Detection
Why mesa beams

Detectors limited by
fundamental thermal noise

Spectral density scales as
1/wn
»

Diffraction prevent
dramatically increasing
beam size

Gaussian beams
sample only a few
percent of the mirror’s
surface
1
Sh  n
w
clip
LIGO-G050XXX-00-R
 m2 
 exp  2 2 
 w 
n = 1 for the dominant coating
losses
Gingin’s Australia-Italia workshop on GW Detection
2
Why mesa beams
Wider, flatter, and
steeper edges beams
Better average over
the mirror surface
depress thermal noise
without compromising
diffraction losses
LIGO-G050XXX-00-R
Gingin’s Australia-Italia workshop on GW Detection
3
Mesa Beam

Optimisation produces
the mesa beam
(same integrated beam power)
Higher peak power
Slow exponential fall
Steeper fall
LIGO-G050XXX-00-R
Aspheric profile
Spherical
profile
Gingin’s Australia-Italia workshop on GW Detection
4
Molecular beam deposited mirror
• Profiled Deposition:
• Coating the desired Mexican Hat
profile using a pre-shaped mask
• precision ~60nm Peak to Valley
•
Corrective coating:
1.
Compare achieved to
desired shape
Correct with molecular
pencil
precision <10 nm.
2.
•
LIGO-G050XXX-00-R
Gingin’s Australia-Italia workshop on GW Detection
5
The test Cavity

7.32 m folded cavity

Rigid structure

Suspended in custom vacuum tank
Flat input
mirror
Flat folding
mirror
MH mirror
INVAR rod
Vacuum pipe
LIGO-G050XXX-00-R
Gingin’s Australia-Italia workshop on GW Detection
6
Cavity Suspensions
V~ 0.6 Hz
H ~ 1 Hz
Suspension System:
GAS spring
wires
LIGO-G050XXX-00-R
Gingin’s Australia-Italia workshop on GW Detection
7
Cavity Vacuum & Thermal Shield
Suspension view
Suspension wires
Vacuum pipe
Thermal shield
Spacer plate
INVAR rod
LIGO-G050XXX-00-R
Gingin’s Australia-Italia workshop on GW Detection
8
“Mexican hat” mirrors
Numerical eigenmodes for a ideal
MH Fabry-Perot interferometer:
The fundamental mode is the socalled “Mesa Beam”, wider and
flatter than a gaussian power
distribution
Cylindrical symmetry yields TEMs
close to the Laguerre-Gauss
eigenmodes set for spherical
cavities
LIGO-G050XXX-00-R
Gingin’s Australia-Italia workshop on GW Detection
“Mexican hat” mirrors

LMA laboratories provided three mirror prototypes

All affected with several imperfection
» Due to the excessively small mirror size

Beam Tested one with a not negligible slope on the central bump

First simulated using paraxial approximation to evaluate how mirrors
with these imperfections would affect the resonant beam
LIGO-G050XXX-00-R
Gingin’s Australia-Italia workshop on GW Detection
10
FFT simulations

The slope on the central bump
can be corrected applying the
right mirror tilt
LIGO-G050XXX-00-R
Gingin’s Australia-Italia workshop on GW Detection
11
Tilts of Spherical
Mirrors

Tilts of spherical
mirrors only
translate optical
axis
LIGO-G050XXX-00-R
Gingin’s Australia-Italia workshop on GW Detection
12
MH Cavity Alignment
Tilt on MH mirrors destroys
cylindrical symmetry
-> resonant beam phase front
changes with the alignment
 Folded cavity: no obvious
preferential plane for mirrors
alignment
-> very difficult align within
required rad precision
=> TEM00 difficult to identify

LIGO-G050XXX-00-R
Gingin’s Australia-Italia workshop on GW Detection
13
Experimental Results


No stable Mesa beam profile was initially acquired
Higher order modes were found very easily
LIGO-G050XXX-00-R
Gingin’s Australia-Italia workshop on GW Detection
14
Results

These modes exhibit good agreement with theory
MH10
Good fit

LIGO-G050XXX-00-R
TEM10
LG10
Bad fit

Gingin’s Australia-Italia workshop on GW Detection
15
Results - other HOM
Diffraction around beam baffle eliminated
LIGO-G050XXX-00-R
Gingin’s Australia-Italia workshop on GW Detection
16
Chasing the TEM00

-
-
-
Apply FP spectrum
analysis:
TEMs identification
and coupling analysis
Non-symmetric
spacing: as expected
TEM00 is the first of
the sequence,
independently of its
profile appearance
LIGO-G050XXX-00-R
Gingin’s Australia-Italia workshop on GW Detection
17
Chasing the TEM00
2-dimensional nonlinear regression:
Definitively not Gaussian
LIGO-G050XXX-00-R
Gingin’s Australia-Italia workshop on GW Detection
18
Experimental Results

TEM00 tilt simulation 4rad tilt
LIGO-G050XXX-00-R
TEM00 data
Gingin’s Australia-Italia workshop on GW Detection
19
Systematic and next steps



Any attempt to “drive” the beam in a centered
configuration failed
cylindrical symmetry is definitely not achievable
FP spectrum analysis: peaks are separated enough
-> we are observing the actual TEM00cavity modes
LIGO-G050XXX-00-R
Gingin’s Australia-Italia workshop on GW Detection
20
Cause of cylindrical symmetry loss




Mechanical clamping
stress deform the folder
and input mirrors
~ 60 nm deformation ->
three times the height of
the MH central bump
Marked astigmatism is
induced
FFT simulation with actual
IM profile confirm problem
LIGO-G050XXX-00-R
Gingin’s Australia-Italia workshop on GW Detection
21
Solving the problem

Flat mirrors too thin (1 cm)
Temporary fix: Distributed stress
with aluminum rings

Thicker substrates ordered

LIGO-G050XXX-00-R
Gingin’s Australia-Italia workshop on GW Detection
22
Other improvements

Improved atmospheric
isolation

Better stability ‘in lock’
LIGO-G050XXX-00-R
Gingin’s Australia-Italia workshop on GW Detection
23
Passing from Side to
Dither lock

Bad spectrum



Improved spectrum
More power in the
fundamental mode
Now can lock on the
TEM00 mode
LIGO-G050XXX-00-R
Gingin’s Australia-Italia workshop on GW Detection
24
Improving Alignment

The reference during alignment was changed from the intensity profile
to the transverse mode spectrum
LIGO-G050XXX-00-R
Gingin’s Australia-Italia workshop on GW Detection
25
The First Mesa Beam
LIGO-G050XXX-00-R
Gingin’s Australia-Italia workshop on GW Detection
26
Non-Linear Fit X
LIGO-G050XXX-00-R
Gingin’s Australia-Italia workshop on GW Detection
27
Non-Linear Fit Y
wtheory  6.68mm
wexperiment  7.60  1.19mm
LIGO-G050XXX-00-R
Gingin’s Australia-Italia workshop on GW Detection
28
Alignment
LIGO-G050XXX-00-R
Gingin’s Australia-Italia workshop on GW Detection
29
Best Mesa Beam

Rsq = 0.996
LIGO-G050XXX-00-R

Rsq = 0.992
Gingin’s Australia-Italia workshop on GW Detection
30
Best Mesa Beam
Jagged top due to
imperfect mirrors
LIGO-G050XXX-00-R
Gingin’s Australia-Italia workshop on GW Detection
31
Tilt Sensitvity

Controllability of beam is
key

Decided to first
investigate tilt sensitivity

Tilt MH mirror about a
known axis
LIGO-G050XXX-00-R
QuickTime™ and a
YUV420 codec decompressor
are needed to see this picture.
Gingin’s Australia-Italia workshop on GW Detection
32
Profiles

Profiles along tilt
axis
QuickTime™ and a
YUV420 codec decompressor
are needed to see this picture.
LIGO-G050XXX-00-R
2 rad
simulated
3.85 rad
experiment
2.57 rad
experiment
Gingin’s Australia-Italia workshop on GW Detection
33
Excuses

Lack of temporal
stability
» vacuum?

Stiction

PZTs are bad
LIGO-G050XXX-00-R
Gingin’s Australia-Italia workshop on GW Detection
34
Summary

We are able to produce acceptable flat-topped
beams with imperfect optics

We have begun to make a quantitative analysis
of mesa beam
» Beam size appears correct
» Tilt sensitivity shows correct trends but less than expected
by a factor of two
LIGO-G050XXX-00-R
Gingin’s Australia-Italia workshop on GW Detection
35
Further Work With This
Set Up

Improve profile using new, stiffer flat mirrors

Repeatability/ stability – vacuum operations

Complete tilt sensitivity measurements

Test other two MH mirrors – mirror figure error
tolerances

Long term – design and build half of a nearly
concentric MH Cavity
LIGO-G050XXX-00-R
Gingin’s Australia-Italia workshop on GW Detection
36
Concentric cavity MH
mirror profile
LIGO-G050XXX-00-R
Gingin’s Australia-Italia workshop on GW Detection
37
QuickTime™ and a
YUV420 codec decompressor
are needed to see this picture.
LIGO-G050XXX-00-R
Gingin’s Australia-Italia workshop on GW Detection
38