Description
The Gravitational Waves (GWs) interferometers are very big facilities and
the choice of the material to build their arms is crucial within the scope of the
project.
The second generation (2G) of GWs antennas have been built using austenitic
stainless steel like 304L (LIGO, Virgo, KAGRA) and 316L (GEO600), which
are no magnetic materials. However, the third-generation (3G) detectors, being
larger and more sensitive than their predecessors, may find cost-prohibitive the
use of these steels.
Einstein Telescope (ET) is a third-generation GWs antenna, which present de-
sign foresees six interferometers with 10 km arms (“xylophone” configuration)
but also a “2L-shape” configuration formed by interferometers with 15 km arms
is being discussed. For ET, the use of ferritic steel for the pipes is being explored
as a more cost-effective alternative than an austenitic solution.
However, the residual magnetization of ferritic steels should be considered as a
potential source of noise that could affect the ET sensitivity curve.
This study aims to present a model, named the Magnetic Dipole Model, which
predicts how a ferritic tube impacts the sensitivity curve of a given instru-
ment, using actual seismic noise data. The model primarily uses data from Sos
enattos, more precisely the 90th percentile from the north-south channel of a
seismometer (HHN instrument), and a key magnetic parameter like the coer-
cive force (Fc). The model has been supported by the characterisation of three
different ferritic steels samples (AISI 430, 444 and 441). The whole process was
carried-out in collaboration with CERN, using the split-coil permeameter for
the measurements of the coercive force.
Our model indicates that the primary contribution of magnetic noise from the
tested ferritic materials occurs between 1 and 2 Hz. It is critical to highlight
also that the model is very sensitive to the distance “d” between the tube and
the mirror, correlating to the length of the cryotrap.
According to the model, the magnetic noise contribution from the tube results to be five orders of magnitude lower than the ET sensitivity. Although this model
has numerous potential enhancements, two notable improvements include in-
corporating seismic data from different potential sites and updating the model
to account for the “2L-shape” configuration.
