Powered by Indico
v3.2
The goal of this series of online meetings is to provide a forum for regular discussions between the teams that work on common design aspects of next-generation gravitational-wave detectors Einstein Telescope and Cosmic Explorer.
The plan is to have a meeting each 2-3 months and start with topics that are more urgent, i.e., that have a strong impact on the detector infrastructure including optical layout, stray-light noise, Newtonian noise, ...
Some material including the recordings is not shared publicly and can be accessed through the following links:
https://wiki.et-gw.eu/ISB/XGCD (accessible by ET collaboration members)
https://dcc.cosmicexplorer.org/cgi-bin/private/DocDB/DisplayMeeting?conferenceid=1056 (accessible by CE consortium members)
The URL shortcut to this indico page is
You can click on the number below to switch between pages of XGCD meetings:
Minimum aperture of the tube. Important to fix early in the design. Starting values 84cm ET-HF, 62cm ET-LF.
Baffles along the beam pipe prevent photons from directly hitting the tube wall.
Upconversion noise was modeled dependent on spectra of the two candidate sites.
The dependence of scattered light loss on beam offset was studied.
Used finite-element analysis to study baffle vibrations and transfer function from tube to baffle. First resonance at 130Hz, which should be high enough not to introduce observable peaks in the spectrum of the stray-light noise.
HL: Can you cut down the Q-value of the baffles?
MM: Can be done. Foreseen to be attached through a welded ring and screws. Could probably damp the mechanical system.
LB: I was expecting that for some reasonable offsets or misalignment, this is what you get, and then bring it back to requirements.
MA: We plan to do this, and it is a relatively easy step to do from our results.
MM: We have a plan for a second publication on non-ideal configurations.
ME: Might there be an important effect from beam-pipe resonances?
AE: There is evidence for resonances around 80Hz in the LIGO tube. The tube moves more than ground, which can also be connected to acoustics. Some issues seem to come from the LIGO baffle design (not the beam-pipe baffles, but the others).
PF: IS the whole cavity being offset or is the beam offset on the mirror?
MA: The beam is moved on the mirror surface.
PF: Interesting. The power drops strongly only for offsets larger than about 10cm.
Match forecasts of stray-light noise with observations. Use baffle locations, material BRDF, and LIGO mirror maps.
Recently there has been an indication that baffle noise can enter DARM at LLO. Shaking the beam tube at particular points, there has been some signal in the GW data between 80-90Hz. Need to compare with models.
Investigate a different way of placing baffles along the pipe. Is typically done with ray optics. What if you consider a realistic beam shape? The field gets cut by the baffles, then propagates and diffracts towards the beam tube. Simulating this effect (with SIS), we obtain an optimal placement of baffles, which is quite different from what you obtain with ray optics.
How much light on the tube is too much? First analyses have started.
How do the results depend on the mirror roughness? Roughness influences backscatter noise.
HL: Does this mean that parts of the tube are directly visible from the center of the mirror?
AK: Yes, part of the tubes are visible, which is reasonable if one considers only the scattering due to mirror roughness. If one takes into account defects which scatter at large angles, a good solution might be to use ray optics for the closer part of the pipe where also large-angle scattering matters, and the beam simulation for the more distant parts of the pipe.
AE: What material did you use?
AK: Black nickel coating. For LLO, I used blackened steel.
PF: There are three different curves in the plot of slide 2. Why are they so different?
AK: I used the assumption that the ground motion directly goes on the baffle. The blue curve uses the reference (ground) seismic data for LLO. The red one is based on data from accelerometers on the beam tube during quiet time, while the yellow one uses data from a louder time (AC is on).
MM: Once you give up on shielding the whole tube, then you need to consider vacuum pumps, which need to be shielded.
LB: Can you say a bit more about the overall logic. Now you can go beyond ray tracing. How do you plan to continue this work? You suggested a mix of the two.
AK: We are currently considering simple baffles (same size, no suspension). We analyze what is the best configuration in this case. After that baseline, we are going to work on different scenarios (passive vibration isolation, variable-size baffles).
Currently in LIGO, the opening angle towards the other end of the arm is 0.0076deg. This corresponds to about 0.8cm spatial wavelength of the mirror profile, which can scatter onto the beam tube. With 40km, we must consider spatial scales of the mirror profile up to 8cm.
The total loss towards the beam tube will be more complicated than today.
Even now, for LIGO and Virgo, the arm loss cannot be well explained by the measured surface map. Also, the measurement of the map with spatial wavelength ~ 1mm has uncertainty of order of magnitude.
Due to the uncertainty of the surface map on larger optics with different beam size, the noise estimation by the beam tube baffle can have large systematic errors, and we need to estimate using various different map cases.
We introduced the concept of instrumented baffle, which was triggered by Virgo discussions. Use photo sensors behind the plate (for example 76 sensors for TM baffles). Baffles coated as usual, but with holes of 4mm diameter for light to enter the sensors.
R&D activity for Si-based sensors with Hamamatsu. Response (linearity) of the sensors was characterized.
Readout currently foreseen at 1kHz sampling.
Vacuum compatibility was a major point of concern. Certification obtained.
A baffle was installed in Virgo at the IMC.
Large baffles being produced (35cm IMC -> 80cm TM). 120 sensors in 5 rings.
Installation plans: considering placing them at the entrance of the tower instead of hanging it at the mirror.
A plan at ETpathfinder now includes instrumented baffles at the entrance of towers. The interferometer is operated at 1.5microns. Needs to redesign the instrumented baffles.
Starting now detailed ET simulations of pre-alignment. Look forward to extending our work in ET to CE.
ME:When I first saw this, it came to my mind that a camera looking at the specular reflection from the baffle with many pixels gives you plenty of information with less effort.
MM: The camera only gives you information from a single angle. We had to substitute baffles, which made this redesign convenient. We did not look into the analysis of better camera arrangements as an alternative.
HL: Should we be worried about wireless?
MM: We tested wireless in vacuum near test mass, and we did not see a problem. We promised to do a full EM scan. This was under the plan to stay at 1cm from the TM. We now consider scenarios with the baffle a few meters away from the TM.
HL: From the photo, it seems that the suspension is done with metal springs, and higher Q should be expected.
MM: Best suspension method is still under investigation.
HL: You mentioned gold coating on circuit boards.
MM: Gold is used to spread heat quickly and then there is the coupling to the baffle.
RS: Digital electronics very close to the test masses might couple to actuators.
MM: We shield all electronics and we will shield on the back of the baffle.