DocDB Home ]  [ Search ] [ Recent Changes ]

Turn up the bass! Low-frequency performance improvement of seismic attenuation systems and vibration sensors for next generation gravitational wave detector

Document #:
Document type:
Submitted by:
Joris van Heijningen
Updated by:
Joris van Heijningen
Document Created:
05 Mar 2018, 18:44
Contents Revised:
05 Mar 2018, 18:48
DB Info Revised:
05 Mar 2018, 18:48
Viewable by:
Modifiable by:
Other Versions:
05 Mar 2018, 18:46
Since September 2015, the LIGO Virgo Collaboration is detecting gravitational waves. After the first detection of GW150914, many binary black hole mergers were recorded. GW170814 was detected by both the Advanced LIGO detectors and the Advanced Virgo detector, which joined LIGO on August 1, 2017. Three days later, GW170817 marked the first ever display of the power of multi-messenger astronomy; a binary neutron star merger and several electromagnetic counterparts were coincidentally detected.
The measurement of gravitational waves has been possible only because most (optical) elements of the detectors have been decoupled from the Earth’s ever-present vibrations. In the Advanced Virgo design, not only is this decoupling necessary for the core optics of the detector, but also for the auxiliary optics. Some of these auxiliary optics are housed on optical benches, which are isolated from the Earth’s vibration using so-called vibration isolation systems. Five such systems have been developed at Nikhef and installed at the Virgo site. The systems are called Multistage Seismic Attenuation System (MultiSAS).
MultiSAS has been designed, constructed and tested at Nikhef. Finite element (FEM) and state space modeling have taught Nikhef engineers much about the (internal) mechanical modes of the system, designing and installing dampers where necessary. The overall transfer function measurement did not show surprises. Five more systems were constructed with the lessons learned from the first characterization campaign and, so far, the systems are behaving according to expectation, meeting requirements set by the Advanced Virgo design. The prototype set-up now serves as an advanced sensors and controls test-bed at Nikhef. MEMS accelerometers and a monolithic accelerometer with interferometric readout are developed on the seismically isolated bench. The most quiet optical table in Europe is now available in Amsterdam for industry to test their sensors on.
Four systems are now suspending an optical table and are operated in vacuum. SIB2, an optical table of the injection system, is ready to be suspended. This was not critical for noise hunting of the 13 W input power, power recycled, Fabry-Perot Michelson interferometer operated at the end of observation run 2 with Advanced LIGO. The other four systems, SNEB, SWEB, SPRB and SDB2 suspend critical optical components for linear and angular alignment controls. SDB2 also houses the photodiodes which receive the signal containing the information about a passing gravitational wave. All systems have been installed and commissioned with a dummy mass. After this 2014 campaign, the dummy masses were removed and the MultiSASs were ready to be used. Later, the benches full of optics were attached to the lower suspension wire and in 2016 and 2017 optical tables were suspended. The MiniTowers, the vacuum vessels around the tables and suspensions, were subsequently evacuated.
Many tests have been performed to validate the expected performance of the isolation systems. The vacuum envelopes of the isolation systems at Virgo are below 1 mbar pressure and this has been tested to be enough at Nikhef not to have (measurable) acoustic coupling. The construction tolerances have been such that the so-called cradle effect has no effect on the inertial sensors on the top stage used for control. Thermal shielding has been designed and tested to protect the GAS filter blades from sagging too much. MultiSAS performance at the site shows to be compliant with set requirements and, apart from a GAS blade failure in April 2016, no significant problems have been encountered. The blade failure has been analyzed and modeling of Von Mises stress induced hydrogen migration, a possible culprit for the failure, is underway. Installation and commissioning of five MultiSASs has been reasonably smooth and already more systems are planned to suspend the squeezing optical table and filter cavity mirrors for Advanced Virgo. Further advancement in advanced control strategies will improve rms performance by better predicting system behavior and acting on it accordingly.
MultiSAS has provided suspended optical tables that are vibrationally so quiet that the best commercial sensors will only measure self-noise from 5 Hz onwards. Nikhef has set forth a combination of two proven concepts to create the world’s most sensitive vibration sensor in the 10 Hz to 100 Hz regime. This is pursued in order to monitor the suspended table motion down to MultiSAS specifications, which entails femtoradians and femtometer motions. The interferometric readout of the sensor features a tabletop interferometer and has a proven performance of 4 fm/rtHz from 5 Hz onwards. When this readout was used to determine the position of the proof mass of a monolithic accelerometer, an unprecedented 8 fm/ rtHz from 30 Hz onwards has been achieved. This is a factor ten better at 30 Hz than the best available sensor: the LIGO/ GeoTech GS-13 geophone.
The modeled shot noise limited sensitivity of the accelerometer with interferometric readout (3 fm/rtHz) has not been achieved yet. More advanced control can improve the proof mass rms motion, keeping the subtraction of common mode noise in the interferometer more optimal. Additionally, the thermal noise could be reduced significantly by using a different material for the mechanics. Already, titanium versions are being produced and these are expected to have 36 times less electrical susceptibility, which greatly reduces the viscous damping. These titanium monolithic accelerometers will most probably be dominated by structural damping. Looking into the future, quantum demolition techniques, such as squeezing, can possibly smash the fm/rtHz barrier.
A fiber-optic version of the same sensor design has been tested and achieved a sensitivity of 4 pm/rtHz. The readout is limited by thermal effects in the vicinity of the fibers, frequency noise and (non-isolated) seismic noise in the Nikhef lab. The next step is installing this sensor on the MultiSAS prototype optical bench in vacuum. Such a sensor can be installed in radiation or high magnetic field environments as the electrical components don’t have to be in the vicinity of the mechanics of the sensor. The interferometric readout in fiber could also be used as a high precision test mass motion sensor to replace e.g. the OSEM in the LIGO suspensions; this was the initial plan of the open air version of this readout.
The author has visited Japan as part of the ELiTES exchange program to help and learn while building the KAGRA gravitational wave detector. At NAOJ (Mitaka, Tokyo), control and sensor development for the room temperature payload for Type B(p) suspensions was performed. For the Type B(p) suspension payload, a simple inertial damping loop with appropriate roll-off filtering has shown to damp the translational, yaw and pitch modes of (dummy) test mass modes. The author also aided fabricating the cabling and set-up for (tests of) the intermediate mass and its recoil mass. All sensing and actuation is done using OSEMs (Optical Sensor and ElectroMagnetic actuation) and these have been calibrated and their designs improved.
Advanced sensor characterization and remote simulation on LVDT read out monolithic accelerometers for Type A or Type B pre-isolators has been performed. Low-frequency open loop operated sensors promise to be able to blend at lower frequency than the now used L4C geophones. This is important as no ground subtraction is foreseen in the KAGRA design. Simulation shows that, without tilt, 20 mHz blending gives better results. There is no proper tilt measurement available of the KAGRA site at this point, and this is almost surely impacting on the found simulation result.
Files in Document:
DocDB Home ]  [ Search ] [ Recent Changes ]

DocDB Version 8.7.10, contact Document Database Administrators