LOW FREQUENCY NOISE SUPPRESSION FOR THE DEVELOPMENT OF GRAVITATIONAL ASTRONOMY

The existence of gravitational radiation, predicted by the General Relativity theory, was indirectly demonstrated by the observation of the orbital decay in the binary pulsar 1913+16, for which R.A. Hulse and J.H. Taylor were awarded with the Nobel Prize in 1993. From then on, the direct detection of gravitational waves became a main issue in the experimental physics, not only for the verification of the theory itself but, most important, because it can open a new "observation window" of the universe. In fact, many astronomical objects, such as neutron stars and black holes, can be directly studied only through their gravitational emission. Moreover, since its interaction with matter is intrinsically weak, the degradation of informations carried by gravitational waves is negligible, and their revelation will allow us to understand the internal structure of massive objects which emit them, and will also provide a complementary point of view to the traditional astronomy and cosmology. The direct detection must face the extreme weakness of gravitational radiation, hence very high sensitive detectors are required in order to reveal the quadrupolar effect produced by the passage of gravitational waves. The first attempts in this field were based on massive resonant bars, relying on the pioneering technique developed by J. Weber. In recent decades a more promising strategy based on interferometry was developed, providing the advantage of a wide-frequency detection-band (from few Hz to some kHz) jointly to an extreme sensitivity (the detectable strain is smaller than the size of a proton). The global network of first generation interferometric detectors, composed of Virgo, LIGO, GEO600 and TAMA300, demonstrated the feasibility of such a technique; in particular the kilometric-scale detectors Virgo and LIGO achieved a sensitivity high enough to determine the first upper limits for the gravitational emission of some known neutron stars, such as the Crab and Vela pulsars. In the next few years the upgraded version of these detectors, namely the second generation of detectors (such as Advanced Virgo and Advanced LIGO) will become operational and are expected to achieve the first direct detections of gravitational waves. However, the signal-to-noise ratio (SNR) of these first detections will be too low for precise astronomical studies of the gravitational wave sources and for complementing optical, radio and X-ray observations in the study of fundamental systems and processes in the Universe. For this reason the investigation on the design of a new, namely third, generation of detectors is already started, leading to the proposal of the European Einstein Telescope (ET). With a considerably improved sensitivity these new machines will open the era of routine gravitational wave astronomy, leading to the birth of a complete multimessenger astronomy. In particular, to enlarge the detector bandwidth in the range of 1 Hz, where interesting gravitational signals, such as those emitted by rotating neutron stars, can be detected, a further reduction of the so-called low-frequency noise, with respect to the second generation detectors, is required. In this low-frequency band the main limitation to the sensitivity of an interferometric detector arises from the thermal noise, and at lower frequencies, from the seismic and Newtonian noises. The suppression of the thermal noise will require the implementation of a cryogenic apparatus, in order to cool the test masses down to about 10 K, so that the development of position-control devices capable of cryogenic operations will be also necessary for the suspension and payload control. The seismic attenuation was already obtained in first generation detectors by means of long suspension chains of vertical and horizontal oscillators (e.g. the superattenuator of Virgo), so that a further reduction requires a smaller seismic noise at the input of the suspension system; moreover, mass density fluctuations produced by the seismic motion induce also a stochastic gravitational field (the so-called Newtonian or gravity-gradient noise) which shunts the suspension and couples directly to the mirrors of the interferometer. In order to suppress these two seismically-generated noises, third generation interferometers will be constructed in underground sites, where Rayleigh surface waves are attenuated, and the surrounding rock layers are more homogeneous and stable, reducing the density fluctuations. The feasibility of a cryogenic and underground interferometer was already tested by the Japanese prototype-detector CLIO, in the same site where is currently under construction KAGRA (formerly known as LGCT), the first full-scale interferometric detector based on these approaches. For these aspects, this second generation detector will be the forerunner of third generation interferometers such as ET, therefore a collaboration between the two scientific collaborations has been established. My experimental work is focused on the suppression of these low noise sources, so that this thesis is structured in two parallel fields of research: the seismic characterization of a potential site for the construction of the Einstein Telescope, and the development, calibration and test of a cryogenic vertical accelerometer, which can be used as a position control device, analogously to those used in the actual room-temperature superattenuator of Virgo, but also to check the vibrations introduced by the cryogenic apparatus, as I did with the measurements I performed on the cryostats of KAGRA, presented at the end of this thesis. The scheme of this thesis is subdivided in three main parts: in the first part I introduce the foundations of the gravitational astronomy, from the theory and the astrophysical sources to the experiments which will lead to the gravitational observations; in the second part I discuss the theory of low frequency noise sources and their suppression; in the third part I present the experimental work I performed in this context. Every part is composed of two chapters, structured as follows. In the first chapter I describe the derivation of gravitational waves from the Einstein's field equations, discussing their properties and the astrophysical and cosmological sources, especially those whose emission is expected at low frequencies. In the second chapter I describe the direct interferometric detection of gravitational waves and the main noise sources which limit the sensitivity, concluding with an overview of present and future detectors. In the third chapter I discuss the main features of the seismic and Newtonian noises, and the strategies necessary to suppress them, especially in third generation detectors. In the fourth chapter I discuss the theory of thermal noise, from the ideal case of the damped harmonic oscillator to the real dissipative mechanical systems and optical components of the interferometer. In the fifth chapter I present my experimental work on the long-period characterization of the Sos Enattos site in Sardinia (proposed for hosting the Einstein Telescope), from the construction and instrumentation of an underground array of sensors to the analysis of seismic and meteorological data collected in one year of observations. Finally, in the sixth chapter I present my experimental work on the development of a cryogenic vertical accelerometer, from the designing to the cryogenic calibration and tests at T=20 K. In this chapter I also present the results of the implementation of this device into the cryostats dedicated to the test masses of KAGRA, where I verified the operations of the accelerometer at T=8 K and I measured the vibrations of the inner radiation shield of the cryostats. These measurements led to a first experimental estimate of the additional vibrational noise which will be injected by the cryogenic refrigerators to the detector test masses