The EXPLORER gravitational wave antenna: recent improvements and performances

Transcription

1 INSTITUTE OF PHYSICSPUBLISHING Class. Quantum Grav. 19 (2002) CLASSICAL ANDQUANTUM GRAVITY PII: S (02) Te EXPLORER gravitational wave antenna: recent improvements and performances PAstone 1, M Bassan 2,3,PBonifazi 1,4, P Carelli 3,5, M G Castellano 3,6, G Cavallari 7, E Coccia 2,3, C Cosmelli 1,8, S D Antonio 9,VFafone 9, G Federici 1, Y Minenkov 3,GModestino 9, I Modena 2,3,AMoleti 2,3, G Pizzella 2,9, G V Pallottino 1,8, L Quintieri 9, A Rocci 2,FRonga 9, R Terenzi 3,4, G Torrioli 3,6 and M Visco 3,4 1 INFN, Sezione di Roma 1, Piazzale A Moro 2, Roma, Italy 2 Dip. Fisica, Università di Roma Tor Vergata, V. le della Ricerca Scientifica, Roma, Italy 3 INFN, Sezione di Roma 2, V. le della Ricerca Scientifica, Roma, Italy 4 CNR, Istituto Fisica Spazio Interplanetario, Via Fosso del Cavaliere, Roma, Italy 5 Dipartimento di Energetica, Università dell Aquila, Roio Poggio, Italy 6 CNR, Istituto di Fisica dello Stato Solido, Via Cineto Romano 42, Roma, Italy 7 CERN, Geneva, Switzerland 8 Dipartimento Fisica, Università di Roma La Sapienza, Piazzale A Moro 2, Roma, Italy 9 INFN, Laboratori Nazionali di Frascati, Via Enrico Fermi, Frascati, Italy massimo.visco@ifsi.rm.cnr.it Received 31 October 2001, in final form 13 December 2001 Publised 18 Marc 2002 Online at stacks.iop.org/cqg/19/1905 Abstract Since te beginning of 2000 te EXPLORER gravity wave (GW) detector as been operated continuously after a stop devoted to improve te apparatus. Te antenna as been equipped wit a new read-out. Te use of a new transducer, caracterized by a very small gap, and a dc-squid wit a ig coupling, led to a better sensitivity and a larger bandwidt. Te EXPLORER sensitivity in terms of spectral noise amplitude, at present (June 2001), is Hz 1/2 over a bandwidt of 35 Hz and Hz 1/2 wit a bandwidt of about 6 Hz, corresponding to a sensitivity to sort conventional GW bursts of = Te performance is stable and te apparatus is taking data wit a duty cycle in excess of 80%. PACS numbers: 0480N, 9555Y 1. Introduction EXPLORER [1], installed at CERN Laboratories in Geneva, is one of te two resonant gravitational antennas of te Rome group. It as been in operation since 1984 and it as performed long-term observations since 1990 (see figure 1). Te data acquired during its /02/ $ IOP Publising Ltd Printed in te UK 1905

2 1906 P Astone et al Figure 1. Operation during te past 10 years. During 1999, te experimental apparatus was upgraded. long activity were used to calculate different upper limits bot for pulse signals [1 5] and for stocastic background [6, 7]. EXPLORER is a part of te international network of resonant-mass detectors (IGEC Collaboration) [5] wic includes ALLEGRO at te Louisiana State University, AURIGA at te INFN Legnaro Laboratories, NAUTILUS at te INFN Frascati Laboratories and NIOBE at te University of Western Australia. During 1999, we modified te detection apparatus of EXPLORER to improve its sensitivity and its immunity from external seismic noise. Since te beginning of 2000 it as started to gater data again. Te performance is stable and te periods wit data of good quality are at te moment (June 2001) more tan 80% of te total time. Te remaining periods are mostly lost due to te periodic cryogenics operations. 2. Sensitivity of te detector In a gravitational wave (GW) detector, two unavoidable sources of noise limit te sensitivity: te termal noise associated wit dissipation in te antenna and te electronic noise of te amplifier. Te first source of noise can be reduced by cooling te antenna. Several efforts were made in recent years in tis direction and nowadays, as a result, some detectors [8, 9] cooled at termodynamic temperature close to 100 mk are in continuous operation. Te reduction of te amplifier noise influence was equally faced. Te use of dc-squid as permitted te contribution of te electronic noise to be strongly decreased, but tere are still opportunities to make significant progress in improving te electromecanical transducer and its coupling to te SQUID. Te sensitivity of a resonant bar can be conveniently expressed by means of te noise S referred to te input of te detector as if it were a GW spectral density. Te sape of S is strongly peaked around te frequencies of te two modes were te lowest values S min are reaced. S min does not depend, in a first approximation, on te electronic noise, but only on pysical parameters of te antenna (lengt L, massm, resonant frequency ω 0 ), on te termodynamic temperature T and on te mecanical quality factor Q: S min S min T QMLω0 3. (1) gives te maximum sensitivity for monocromatic or stocastic sources. If sort pulses of GWs are considered, te sensitivity can be calculated integrating S (ω) over te spectrum and te performances of a detector can be conveniently expressed by means of te minimum modification of te metric tensor detectable by te apparatus. Te sensitivity to burst depends not only on te minimum value of S, but also on te bandwidt.

3 Te EXPLORER gravitational wave antenna 1907 Te bandwidt ω 0 of te detector is strongly affected by te electronic noise of te apparatus: ω 0 = ω 0 Q Ɣ were Ɣ T n 2βQT. (2) T n is te noise temperature of te amplifier used, β is a factor representing te coupling between te mecanical and electrical parts of te system and depends on te transducer and its matcing to te amplifier used. Ɣ gives te ratio between te wide-band noise in te resonance bandwidt and te narrow-band noise. Te pysical parameters of te antenna (M, L, ω 0 ) and te termodynamic temperature T are fixed in a given detector, so te most significant improvements of a bar detector sensitivity can be acieved by decreasing te contribution of te electronic noise T n and increasing te coupling β of te transducer to te SQUID. Te groups involved in resonant detectors, during te last several years, ave devoted substantial efforts in tis direction and te first significant results ave appeared. 3. Experimental configuration Te antenna is made of ig Q alloy Al 5056, as a mass M = 2200 kg and a resonant frequency around 900 Hz. It can be cooled in a cryostat by superfluid liquid elium at a temperature around 2.5 K. Te antenna is equipped wit a capacitive resonant transducer and uses a superconductive interferometer dc-squid as amplifier (see figure 2). Figure 2. Read-out sceme. Te transducer, developed by our group [12], as an innovative design. It as been used for te first time on tis detector since te beginning of Te peculiar geometry of te resonator, rosette saped, allows a gap as small as 10 µm and consequently a capacitance C t = 12 nf tat is more tan tree times larger tan in te transducers we used in te past. Its mecanical Q is about and te overall Q expected for te system is around Te dc-squid is a commercial device produced by Quantum Design, it as an input flux noise n comparable to tat measured wit te SQUID previously used, but its input coil mutual inductance M s = 10 nh provides a coupling tree times larger tan in te past. Te use of tis new transducer and SQUID as increased te coupling between te mecanical and electrical parts of te circuit, decreasing te Ɣ by a factor larger tan 100. Wit tis configuration, we expected a spectral sensitivity S suc as te dased line in figure 3 [11] wit a bandwidt larger tan 10 Hz, corresponding to a minimum detectable for conventional standard pulses.

4 1908 P Astone et al S frequency (Hz) Figure 3. S over 10 of data. 4. Results and perspectives Te EXPLORER new run started at te beginning of te year 2000 but only starting from te end of November were we able to get a set-up tat assures reliable operations and very good and stable performances. Te data at te SQUID output are sampled at 5 khz and filtered wit an adaptive matced filter. Figure 3 sows a typical plot of te S (ω) relative to 10 of data, wereas te dotted line represents te sensitivity predicted by a numerical model of te detector [11]. Te frequencies of te two modes of oscillation of te antenna transducer system, wit a biasing field E = 7.3 MV m 1,areν = Hz and ν + = Hz. Te mecanical quality factors are respectively Q = and Q + = During tese measurements te dc-squid exibited a flux noise of 12 µ 0 Hz 1/2. Te EXPLORER sensitivity in terms of spectral noise amplitude, at present (June 2001), is Hz 1/2 over a bandwidt of 35 Hz and Hz 1/2 wit a bandwidt of about 6 Hz, corresponding to a sensitivity to conventional GW bursts of = Te values of te mecanical quality factor are lower and te noise of te SQUID is iger tan expected. Terefore, te sensitivity does not matc tat expected (dased line in figure 3) wit all te parts of te apparatus working at teir best. We are confident tat we will be able to obtain tese results in te near future as all te operations required to possibly obtain te goals can be performed witout warming up te apparatus. In figure 4 we report, for te period February June 2001, te value of averaged every alf an our. For most of te time its value is Te duty cycle in tis period is larger tan 80% of te total time. In figure 5 we sow a istogram of values over an entire day of good operation: te distribution is well fitted by a Gaussian wit standard deviation equal to , altoug a small tail of a few tens of samples is present. Te sensitivity expected wit te present experimental configuration is not te ultimate reacable by EXPLORER: te use of an amplifier made wit a double SQUID, a new transducer wit a iger capacitance and a superconductive transformer wit lower dissipation, will lead us to increase te sensitivity enoug to detect Te future detection apparatus is under test in our lab and could be used for te next improvement of EXPLORER antenna.

5 Te EXPLORER gravitational wave antenna doy Figure 4. Operation from January to June min averaged data Number of samples x Figure 5. Distribution of 24 of EXPLORER data day 161 of Next year, similar to NAUTILUS [10], te apparatus will be equipped wit a cosmic ray detector consisting of plastic scintillators, one layer of 6 m 2 under te cryostat and two layers of 13 m 2 above it. Tis improvement will make it possible to study signals induced in te antenna by cosmic rays crossing it. Conversely, te experience gained wit te new read-out of EXPLORER will be used for te next upgrade of te NAUTILUS antenna, foreseen in early Acknowledgments We tank F Campolungo, R Lenci, G Martinelli, E Serrani, R Simonetti and F Tabaccioni for teir valuable tecnical support.

6 1910 P Astone et al References [1] Astone P et al 1993 Pys. Rev. D [2] Amaldi E et al 1989 Astron. Astropys [3] Amaldi E et al 1990 Europys. Lett [4] Astone P et al 1999 Pys. Rev. D [5] Allen Z et al 2000 Pys.Rev.Lett [6] Astone P et al 1996 Pys. Lett. B [7] Astone P et al 1999 Astron. Astropys [8] Astone P et al 1997 Astropart. Pys [9] Prodi G A et al 1998 Initial operation of te gravitational wave detector AURIGA 2nd Edoardo Amaldi Conf. on Gravitational Wave Experiments (Geneva, Switzerland, 1997) ed E Coccia, G Veneziano and G Pizzella (Singapore: World Scientific) pp [10] Astone P et al 2000 Pys. Rev. Lett [11] Bassan M Te GASP numerical model for resonant antennas (unpublised) [12] Bassan M, Minenkov Y and Simonetti R 1997 Advances in linear transducers for resonant gravitational wave antennas Proc. Virgo Conf. on Gravitational Waves: Sources and Detectors (Cascina, Mar. 1996) ed I Ciufolini and F Fidecaro (Singapore: World Scientific) pp 225 8