Thermal Compensation SystemTCS
V. Fafone for the TCS Subsystem
AdV 1st Project Review, Cascina - 3.11.2008
2
Thermal Effects in Advanced Virgo
• Thermal compensation for Advanced Virgo, must take into account the much greater circulating laser power (~800kW)
• Optical power is predominantly absorbed by the HR coating and converted into heat, producing a gradient of temperature inside the substrate.
• Two different effects originate from the heating of the test mass:– nonuniform optical path length distortions (thermo-optic effect, or thermal
lensing) mainly due to the temperature dependency of the index of refraction.– change of the profile of the high reflective surface, due to thermal expansion
(thermo-elastic deformation).
AdV 1st Project Review, Cascina - 3.11.2008
3
3
Thermal Effects in Advanced Virgo• In AdV, thermal effects will create distortions both in the recycling and in the Fabry-Perot cavities:
– Wavefront distortion in the PRC• in presence of thermal lensing, which changes the cavity mode, the input laser no longer matches the PRC cavity,
the coupling coefficient between the laser TEM00 and the cavity TEM00 is less than one.– HR surface elastic deformation
• The cavity becomes less concentric, and the spot sizes at the mirrors will shrink (increase of thermal noise and reduction of arm cavity coupling to the input beam)
• Necessity to control the radii of curvature of all test masses, TCS will have to act on both ITM and ETM.
From S.Chelkowski presentation at 1.7.08 beweekly meeting
AdV 1st Project Review, Cascina - 3.11.2008
4
4
Simulation of thermal effects in Advanced Virgo Test Masses
Effect of the absorptions on ITM: 0.5ppm (coating) and 2ppm/cm (substrate).FP cavity power ~ 800 kWAbsorbed power ~ 0.5 W
Thermo-elastic deformation of the HR surface
Change in the ROC
Optical path length increase in the substrate.The focal length of the equivalent lens is 4.5 km
AdV 1st Project Review, Cascina - 3.11.2008
5
TM heated on both faces, with the same profile, but different powers: 2.5 W on the HR face, 7.5 W on the AR face.
Simulation of thermal effects in Advanced Virgo Test Masses
Effect on the ROC
Evaluated possibility to correct both thermal effects with one actuator:It is not possible if we shine only one faceIt is possible with CO2 heating on both faces•Front heating increases ROC (~3.1 m·W)•Back heating decreases ROC (~2 m·W)
AdV 1st Project Review, Cascina - 3.11.2008
6
Advanced Virgo 1st Review - 23.10.2008 6LIGO Laboratory 6
TCS Noise Coupling Mechanisms• Thermoelastic (TE)- fluctuations in locally
deposited heat cause fluctuations in local thermal expansion
• Thermorefractive (TR)- fluctuations in locally deposited heat cause fluctuations in local refractive index
• Flexure (F)- fluctuations in locally deposited heat cause fluctuations in global shape of optic
AdV 1st Project Review, Cascina - 3.11.2008
7
7
Ad Virgo TCS noise
In case the CO2 is applied on the HR face,to have TCS noise 10 times smaller than AdVirgo sensitivity @ 50 Hz RIN must be 5·10-9/Hz
Flat RIN over the entire frequency band
No relevant change with TM 30 cm thick
In case of heating on both sides of the TM, RIN requirements are released, due to cancellation mechanisms, to 10-8/Hz.
10 100 1 103 1 1041 10 26
1 10 25
1 10 24
1 10 23
1 10 22
1 10 21TCS noise totalFL noiseTR + TE noiseRad pressure noiseAdVirgo sensitivity
AdV 1st Project Review, Cascina - 3.11.2008
8Advanced LIGO and Advanced Virgo share the same thermal issues. A possible TC System has already been proposed in the LIGO Collaboration
Green dots: shielded heating ringBlue rectangles: CP
Compensation plates will correct thermal effects in the PRCShielded heating rings will compensate HR surface deformations
TCS scheme
In the case of Advanced Virgo, the positioning of CPs in the PRC is a delicate issue due to its interaction with the suspension systems
AdV 1st Project Review, Cascina - 3.11.2008
9
Compensation plate• Transmissive optic that lives entirely in the recycling cavity;• TCS noise coupling is much lower: RIN requirement is 10-6/Hz;• CPs interact with the IFO beam: it must satisfy requirements like those
of a core optics with respect to displacement noise, absorption and scattering, index homogeneity, antireflection coatings, and the like
– In order to minimize etalon effect in the CP, a wedge or a tilt is foreseen. – If the CP is wedged, there will be sensitivity to transverse motion. A 10 arcmin wedge
is considered.
safety factor of 10 wrt the AdV sensitivity
AdV 1st Project Review, Cascina - 3.11.2008
10
Compensation plate– The CP is foreseen to be attached to the ITM
reference mass, noise motion requirement is expected to be easily met.
– The most feasible solution is to fix the CP on the back of the reference mass.
– The clear aperture due to the coil-magnet actuators on the back of the RM is 280 mm, a maximum diameter of 280 mm for the CP is foreseen, while the thickness will be of about 7-10 cm.
– Design is in progress
AdV 1st Project Review, Cascina - 3.11.2008
11
CP effect on Longitudinal Sensing and Control: simulation
• Simulation of CP displacement noise coupling to longitudinal error signals and gravitational channel
• A mirror with 100ppm reflection added behind IMY, inside PRC
• No wedge on CP, worst case• Full 125 W input power• Transfer function from CP motion to all longitudinal signal
simulated with Optickle
Compensation plate
AdV 1st Project Review, Cascina - 3.11.2008
12
LSC: estimate noise projection• CP motion estimate projected linearly to h• Motion of CP reconstructed in a very conservative way:
– Assume in AdVirgo we actuate on input RM– Assume the AdVirgo longitudinal correction is the same as in Virgo– Assume RM and Mirror to have the same mass
• Therefore CP motion is assumed to be the same we would have in Virgo
AdVirgo correction should be much lower
Inserting wedges (or tilting) will largely reduce this coupling
Another reason to have a wedged or tilted CP is that if a plate with faces perpendicular to the beam is moving longitudinally in a cavity, it induces power fringes; with 100 ppm AR coating, the relative power variations would be 4%, on a time scale of 1 s. These variations may fool the DC locking of DARM.
AdV 1st Project Review, Cascina - 3.11.2008
13
• Simulation done with Finesse• A mirror with 100ppm reflection added
behind IMX• Longitudinal and Angular Lock active
Simulations:CP etalon effect on the Alignment optical gainsEffects of the CP misalignment with respect the IMX mirror (1 deg wedge on CP - worst case)Introduction of thermal effects
CP effect on Automatic Alignment: simulation
Compensation plate
AdV 1st Project Review, Cascina - 3.11.2008
14
Etalon effects on the Alignment optical gains
The CP etalon tuning produces an optical gain variation of the order of 1‰ on the AA error signals
CP misalignment effects on the AA control loop
The misalignment of the CP front face produces a big effect on the alignment error signals (1.5 nrad of correspective EMX misalignment for a CP misalignment of 10 urad)
Thermal effects included in the CP: lens with ROC of about 2700m (IFO at full power). No relevant changes from the CP plane configuration
Angular and longitudinal locks active
AdV 1st Project Review, Cascina - 3.11.2008
15
Conclusions
• Longitudinal system– Noise introduced in h by CP should be negligible– No significant effect on longitudinal control system
• Angular system– The CP etalon effect results to be negligible– The CP front face misalignment produces strong effects on the AA
error signals up to 0.1 mrad
The CP has to be misaligned more than 0.1 mrad with respect to the IMX front face
AdV 1st Project Review, Cascina - 3.11.2008
16
RH
TM
RMShielded ring heater should be embedded in the reference mass. It is necessary to avoid any heat transfer between the ring heater and the RM. The heating element should have the highest emissivity, while its shield should have the lowest (gold coating).
The heating element must be designed to avoid emitting any magnetic field that could couple with Advanced Virgo main beam or with local controls.
Geometry, shielding, materials must be optimized
Shielded ring heater design guidelines
AdV 1st Project Review, Cascina - 3.11.2008
17
Shielded ring heater design guidelines
RHTM
RH
HR Surface
Developing FEA to optimize the position/power of the heating ring.
Result for a TM heated by a simple ring - no shielding included yet, lower power is expected to give the same ROC correction
AdV 1st Project Review, Cascina - 3.11.2008
18
The full TC System to be implemented on Advanced Virgo will comprehend sensors which sample the thermal wavefront distortions within the interferometer and the ROC,
The TCS sensing scheme is to be defined
Need to investigate a dedicated TCS wavefront sensor. At present the Hartmann sensor seems to be a promising solution.The University of Adelaide’s Hartmann sensor has been demonstrated to have a shot-to-shot reproducibility of λ/580 at 635 nm, which can be improved to λ/16000 with averaging, and with an overall accuracy of λ/6800. Hartmann sensor compliant with AdVirgo spec should be developed within one year.
TCS sensing scheme
AdV 1st Project Review, Cascina - 3.11.2008
19
Status of the R&D• The definition of the TCS of AdV can benefit of the many years of R&D
already performed by LIGO• Last spring, an R&D program has been proposed to EGO Stac, it has been
evaluated as necessary for AdV.• The budget for the first year will be allocated within 2008.• The activity consists of development of FEA and measurements of thermal
effects induced in the optics and their compensation:– FEA is used to determine the amount of thermal effects both on the TM and in the CP and
the compensator parameters (such as dimensions of the CP, profile and power of the CO2 beam, dimensions and power of the heating ring). The FEA modeling also serves as consistency check for the measurements results. This modeling effort is in a well advanced state and has already supplied useful indications to define the TCS parameters.
– The efficiency of a scaled down prototype compensator (CP heated by CO2 laser and shielded HR on TM) in correcting at the same time both wavefront distortions and thermo-elastic deformations will be measured. The possibility to monitor both thermal effects will give information for the development of a closed loop control scheme. In the frame of this R&D we also plan to define a dedicated wavefront sensor to be implemented on AdV.
AdV 1st Project Review, Cascina - 3.11.2008
20
Directional Radiative Cooling (DRC) Principle• Establish thermal radiation heat exchange between a cold surface (masking
the environment) and the mirror– Heat radiated from the spot is absorbed by the cold target– The cold target returns less heat to the laser spot
• The cold target could be a surface cooled with LN2 or a cryocooler• Simulations have been performed with target at 77K:
– Highly efficient 99.6%, emits only 0.4% thermal radiation than a body at room temperature
AdV 1st Project Review, Cascina - 3.11.2008
21
21
DRC Issues
Matched10% larger10% smaller
• Radiative cooling is effective in correcting thermal lensing and ROC at the same time
• High sensitivity to profile mismatches• Scattered light: is the presence of large reflectors
placed in front of the TM at small angle (~20 deg) compatible with AdV specifications?
• Since most of the apparatus lives in-vacuum, how to make it more flexible, to be easily upgraded as new understanding of the IFO is realized?
• What if the absorptions are non uniform?• Methods to tune the cooling power must be
investigated• Need to investigate noiseless cooling systems• Evaluate interactions with other subsystems• Small R&D foreseen to evaluate the feasibility
AdV 1st Project Review, Cascina - 3.11.2008
22