Gravitational wave detector OGRAN as multi-messenger project of RAS-MSU
V. N. Rudenko, Yu. M. Gavrilyuk, A. V. Gusev, D. P. Krichevskiy, S. I. Oreshkin, S. M. Popov, I. S. Yudin
NNovember 12, 2019 1:56 WSPC/INSTRUCTION FILE OGRAN˙as˙multi-messenger˙project
International Journal of Modern Physics Ac (cid:13)
World Scientific Publishing Company
Gravitational wave detector OGRAN as multi-messenger project ofRAS-MSU
V. N. Rudenko , , Yu. M. Gavrilyuk , , A. V. Gusev , D. P. Krichevskiy , , S. I. Oreshkin ,S. M. Popov , I. S. Yudin , Sternberg Astronomical Institute, Lomonosov State University, Universitetskii prospect 13,Moscow, 119234, Russia Baksan Neutrino Observatory of INR RAS, Neutrino village, Elbrus district, 361609,Kabardino-Balkar Republic, Russia Bauman Moscow State Technical University, Physics Department, 2-nd Baumanskaya street 5,building 1, Moscow, 105005, Russia Moscow Institute of Physics and Technology, Institutskiy per. 9, Dolgoprudny, 141701, MoscowRegion, [email protected]
Received Day Month YearRevised Day Month YearModernized version of the combined opto-acoustical gravitational wave detector OGRANis presented. Located in the deep underground of the Baksan Neutrino Observatory thissetup is aimed to work on the program of collapsing stars searching for in multi-channelmanner with the neutrino telescope BUST. The both instruments have the sensitivityallowing a registration of collapses in our Galaxy as rare events with the estimated prob-ability 0 .
03 year -1 . The OGRAN narrow band sensitivity at the kilohertz frequency islimited by its acoustical mode thermal noise achieving 10 − in term of metric per-turbations. A possible algorithm of the joint data analysis for the both instruments isdeveloped and resulting formulae of the right detection probability are given. A futureincreasing of the OGRAN sensitivity associated with the moderate cooling (nitrogentemperature) of the acoustical mode is also discussed. Keywords : Gravitational detector; neutrino telescope; multi-messenger astronomy.PACS numbers: 04.08.Nn, 95.55.Vj
1. Introduction
The first registration (September 2015) of the gravitational wave (GW) signal fromthe merger of a binary formed by two black holes with masses of order 30 M (cid:12) , from adistance of 400 Mpc was a landmark event, confirming the reality of hopes for grav-itational wave astronomy as a new unique information channel in our knowledge ofthe universe. The signal recorded by LIGO detectors in September 2015 amountedto h (cid:39) − in terms of the dimensionless amplitude of the gravitational wave (orspatial metric variations in geometric language) at an average carrier frequency of150 Hz. In form, the signal fit into the well known portrait of so called chirp signal a r X i v : . [ phy s i c s . i n s - d e t ] N ov ovember 12, 2019 1:56 WSPC/INSTRUCTION FILE OGRAN˙as˙multi-messenger˙project V. N. Rudenko, Yu. M. Gavrilyuk, A. V. Gusev et al. i.e. looks as a radio-pulse with a changing carrier frequency. It has the three charac-teristic phases: inspiral, merging and relaxation (ring down). Until the end of 2016,a couple more similar signals were registered. In 2017, two more significant eventstook place. Firstly, the registration of signals from the merger of massive black holesperformed already by three detectors. The LIGO interferometer was joined by theEuropean VIRGO interferometer, which increased the accuracy of the localizationof the source of GW in the celestial sphere by a hundred times. Secondly, the samedetectors recorded a signal from the merging of the components of a binary neu-tron star GW170817, which was accompanied almost simultaneously by the gammaray burst GRB170817A recorded by the Fermi and Integral spacecrafts. The 10hours later, in zone of localization of source of gravitational radiation a supernovabelonging to the Galaxy NGC was discovered by optical telescopes (including theMASTER - global network of robotic telescopes ). This confirmed the hypothesisabout the general nature of astrophysical sources of gravitational and gamma-raybursts associated with relativistic catastrophes of super dense stars. In this case,it was a merging of neutron components of the relativistic binary at the end of itsevolution, which was also accompanied by a supernova explosion. Naturally, thisachievement gave rise to a new surge in the activity of involved scientific collabora-tions and individual groups with a number of proposals for new long-term projectsfor high-sensitivity detectors, such as Einstein Telescope, Voyager and Cosmic En-dower. LIGO and VIRGO instruments have achieved an extraordinary sensitivityto small perturbation of their baseline (spectral density of deformation noise10 − − − Hz -1/2 ) in the region from 10 Hz up to 2 kHz. From the technicalpoint of view these setups are extremely complex systems of automatic control ofoperational regime with a high level of seismic isolation, high vacuum in big vol-umes, modern laser physics technology etc . It is a hard problem to keep a longtime continuous operation without losing the working point. In contrast the abilityfor long time (through years) continuous operation is the important positive fea-ture of resonance bar detectors. Having such property, a resonance GW detector ofmoderate sensitivity 10 − − − Hz -1/2 could be used in the program of searchfor rare events - catastrophic processes with relativistic objects in our Galaxy andclose environment (100 kpc) accompanied by different types of radiation. In factat present a similar programs are carried out also with underground neutrino tele-scopes. It consists in a search for Galactic collapsing stars on the neutrino channel.In this paper, we describe the present status of opto-acoustical gravitational antennaOGRAN constructed for a multi-channel mode monitoring relativistic events andlocated the deep underground of Baksan Neutrino Observatory (BNO) close to theBaksan Underground Scintillator Telescope (BUST).ovember 12, 2019 1:56 WSPC/INSTRUCTION FILE OGRAN˙as˙multi-messenger˙project Gravitational wave detector OGRAN as multi-messenger project of RAS-MSU
2. GW Detector OGRAN Modernization2.1.
Technical status of initial setup
The idea of construction of a GW detector as an acoustical resonance bar coupledwith the optical FP cavity composed by mirrors attached to the bar ends wasconsidered in Refs. 12, 13. A new quality was emphasized of such a combination:(a) a more complex structure of signal response, containing separately acousticaland optical parts, and (b) a possibility to get sensitivity at the level of bar Browniannoise in a limited spectral frequency range due to the small back action of opticalread out. Implementation of this idea was reported as the OGRAN project. Atpresent the full scale setup is constructed, tested and installed into undergroundfacilities of BNO INR RAS. The principal opto-electronic scheme of the setup isgiven in Fig. 1. Generally, it belongs to the design type called a comparator ofoptical standards. It is composed by two feedback loops. The first one couplesthe Fabry-Perot cavity (FP) on the large bar (OGRAN acoustical detector) withthe laser of optical pump which frequency automatically tuned in resonance withbar FP cavity. Any change of the FP optical length is converted into the pumpbeam frequency variation. The second loop is the measuring one. It consists of theadditional FP cavity (the frequency discriminator) illuminated by the same pumplaser but tuned in resonance by an independent piezo-ceramic driver attached toone of discriminators mirror. Its output signal is proportional to the frequencydifference between of laser pump and of discriminator’s cavity. Any perturbation ofthe detector bar cavity (above of feed back cut off frequency ∼
100 Hz) is reflectedin the output discriminator signal. It worth to emphasize once more the OGRANphysical specificity: the gravitational wave interacts not only with acoustical degreeof freedom (the resonance bar) but with the EM field in the cavity as well. Itresults in a complex structure of signal response with optical and acoustical parts.A payment for this originality is the technical problem of constructing a large scalehigh finesse FP cavity rigidly coupled with the acoustical resonator without loosingits high mechanical resonance quality factor Q.The initial version of the OGRAN antenna was described in details in the Ref. ? . It had the following physical and technical characteristics: acoustical detectoreffective mass M = 10 kg, length L = 2 m, acoustical resonance frequency ω / π =1 . Q = 10 . Parameters of optical part were the following: laser wavelength 1064 nm, laser total power W ≈ . F = 3000 of detector cavity, F = 1500 of discriminator cavity, contrastof the interference C ≈ .
2. The experimental sensitivity curve in the form of noisespectral density was presented in Ref. ? . It resulted in the minimal detectable strain(∆ L/L ) f ≈ − Hz -1/2 in the bandwidth ∆ f ∼
30 Hz, and (∆
L/L ) f ≈ − Hz -1/2 in the bandwidth ∆ f ∼ . V. N. Rudenko, Yu. M. Gavrilyuk, A. V. Gusev et al.
Technical improvements of the installation
Seismic and acoustic isolation
Despite the underground location of the OGRAN antenna, the residual mechanicalvibrations of the environment negatively affected the stability of its functioning.The measuring part of the setup was most susceptible to their influence. First inthat number is the frequency discriminator having the form of cylindrical FP cavity(made from the ULE optical glass - sitall) located in a vacuum chamber on a massiveoptical table. Due to the fact that the discriminator is installed in a vacuum chamberin a vertical position, there are increased requirements for its seismic isolation.For these purposes, an anti-seismic suspension was developed, which includes amulti-link anti-seismic filter and acoustic sound insulation of a vacuum chamber.This antiseismic filter was composed by a system of several stainless steel ringsseparated by elastic tablets of isolating material (viton). The size of the tabletswas selected from the calculation of the total load of the upper rings and the bodyof the discriminator. The lower tablets are larger in size. The conical system ofrings was used as amore resistant to lateral vibrations of the inclined mode. In theupper link of the filter, a spring suspension with damping by soft mesh sponge madeof stainless steel is applied to lower its eigen frequency. By selecting the size andrigidity of the dampers, as well as reducing the size of the viton tablets betweenthe filter rings, we were able to reduce the eigen frequency of the entire filter to3 Hz with a quality factor of about 0 .
3. The transient response of the filter lookslike a step with a slight rise at the 3 Hz resonance and subsequent decline. Thesteepness of the decline is determined by the eigen frequencies of the steps andtheir number. To reduce the external acoustic impact, the discriminator’s vacuumchamber was externally coated with a two-layer shell of sound-absorbing materialsuch as Comfort Mat 10 mm thick.
Fig. 1. Principle scheme of the OGRAN antenna. ovember 12, 2019 1:56 WSPC/INSTRUCTION FILE OGRAN˙as˙multi-messenger˙project
Gravitational wave detector OGRAN as multi-messenger project of RAS-MSU Optical and electronic parts (i) The optical parts of the OGRAN setup were drastically updated. The mirrors ofboth FP cavities were replaced by the high quality mirrors (manufactured in theLaboratoire des Matriaux Avancs (LMA), Lion France). The new mirrors havethe same geometry, but very low dissipative losses 1 − − ∼ ∼ . which uses phase modulation of the light beam. Withpure phase modulation in the PDH technique, the excess noise does not occur.However, the available phase electro-optical modulators due to their not perfectand inaccurate settings bring also the residual (spurious) amplitude modula-tion of the pump radiation. Detection of such radiation on the photodetectorsgenerates excess noise in electronic circuits. The effect of residual amplitudemodulation (RAM) in electro-optical phase modulators is not fully understood.Measurements have shown that it depends on the transverse inhomogeneity ofthe laser beam, as well as on the properties of the propagation medium. Weused the empirical method of actively suppressing RAM at the output of thephase modulator. A ray passing through a modulator enters an optical fiber(that preserves the polarization of light) through a mirror (on a piezo driver).At the fiber output, a small fraction of the light is cut off to the photo detec-tor to generate an electronic error signal. The latter is produced only by lightcomponents of amplitude modulation. The error signal goes to the driver of themirror, which holds the position of the wavefront of the beam at the entranceof the optical fiber corresponded to the minimum (zero) error signal (lack ofRAM). Experimental results showed that after applying such an active feedback system, RAM noise is suppressed by more than 30 dB.(iii) The magnitude of the output signal OGRAN is proportional to the opticalpump power. However, the available photodetectors have limited power dissi-pation. Also, when operating at high frequencies, the parallel shunt capacitanceof the photodiode is important (for a modulation frequency of 10 . V. N. Rudenko, Yu. M. Gavrilyuk, A. V. Gusev et al. photodiodes, which also increases their capacitance. In the improved OGRANinstallation, silicon photodiodes in the old version are replaced by photodiodesfrom InGaAsP, which have the better quantum efficiency - 0 . . −
50 mW at frequencies of tens MHz. At higherpower, the signal at 10 . ≥ ? ). Atthe level of 10 − Hz -1/2 , the reception band expanded from ∼ ∼
3. OGRAN Potential and Factual Sensitivity
Lets come back to the calculation of sensitivity of the OGRAN antenna , consideringthe both natural type of noises: the brown noise of its acoustical resonator andphoton short noise of the optical one. For simplicity we will neglect of the complexstructure of the antenna reaction, containing of acoustical and optical parts in itsresponse to GW . To estimate sensitivity it is sufficient to follow the simple schemein which the GW signal excites of the fundamental mode of bar acoustical resonancebut the optical FP cavity serves as a read out system.At the output of such read out one has some bar length displacement x ( t )produced by equivalent GW force f g ( t ) together with the Langevin force of thermaloscillation f T ( t ). The oscillation x ( t ) appear also at the additive read out (optical)noise back ground x a ( t ). Thus the total output displacement x tot ( t ) in the complexamplitude formalism is presented as x tot ( t ) = K ( p ) (cid:2) f g ( t ) + f T ( t ) + K − ( p ) x a ( t ) (cid:3) , where the oscillation mode transfer function K ( p ) = [ M ( p + 2 δp + w )] − .ovember 12, 2019 1:56 WSPC/INSTRUCTION FILE OGRAN˙as˙multi-messenger˙project Gravitational wave detector OGRAN as multi-messenger project of RAS-MSU The theory of optimal filtering suggests to perform the signal extraction comingback to the antenna input i.e. applying the inverse filter. Then the spectral densityof the noise background after the inverse filter reads as N ( ω ) = N T ( ω ) + K ( jω ) N a ( ω ) . Here N T ( ω ), and N a ( ω ) are the spectral densities of thermal fluctuation and additiveread out noise. Below they are considered as constant N T , N a .Considering the processes in a narrowband region around of the fundamentalacoustical mode resonance frequency ω and supposing the smallest dissipationindex δ << ω one comes to the approximation N ( ω ) = N ( ω + ω ) ≈ N T + (2 mω )2 N a ω + o ( δ ) , | ω | << ω . For estimation of the potential and real OGRAN sensitivity it is convenient topresent this noise background introducing a new parameter - the noise factor F : N ( ω ) = N T [1 + N a (2 M ω ) ω /N T ] = N T F ( ω ) . It is convenient to rewrite the noise factor in a more practical form: F ( ω ) = 1 + ( ω/ ∆ ω ) , (∆ ω ) = N T / ( N a (2 M ω ) ) . Here ∆ ω is the effective bandwidth of potential sensitivity: in the region around theantenna resonance frequency ω ± ∆ ω where the sensitivity is limited only by thethermal fluctuation of the acoustical degree of freedom. Fig. 2. Noise spectral density of the OGRAN antenna (doted line - the experimental curve of oldversion of the setup F ∼ F ∼ ovember 12, 2019 1:56 WSPC/INSTRUCTION FILE OGRAN˙as˙multi-messenger˙project V. N. Rudenko, Yu. M. Gavrilyuk, A. V. Gusev et al.
Lets apply these general formulae directly to the OGRAN setup, using the pa-rameters of the old version presented in Ref. ? . The minimum metric perturbationregistered by the opto-acoustical detector using the optimal filtering procedure isread as h min = (∆ L/L ) ≈ (4 /L ) [( kT /M ω )(1 /Q/ω τ )] / √ F .
The potential sensitivity corresponds to the noise factor F = 1. Substitution of theOGRAN parameters: M = 10 kg, L = 2 m, ω = 8 ∗ rad/s, Q = 1 . ∗ and the duration of GW burst τ ≈ (1 / ∆ f ) = 2 ∗ − s (2 periods of resonancefrequency) results in the sensitivity estimate h min = 1 . ∗ − . The correspondentreceiver band width is 500 Hz. In the more short bandwidth the sensitivity forecastis at the level h min = 2 ∗ − . At last in the very narrow bandwidth ∼ h min = 5 ∗ − .It is known that the relativistic catastrophic events type of SN explosion canprovides h = 10 − from the distance 10 kpc if the part of energy contributed in GWwill be 1% of the solar mass. However the number of papers with calculation of thecollapse efficiency forecast GW pulse signal in the best case at the level h ≤ − .It means that the reachable sources would be located at more close distances ∼ F ≤ ∼ . F = 2000. ModernizedOGRAN version uses the pump power 2 W and has the finesse in the detector FPresonator F = 30000, i.e. the requirement F <
4. Neutrino and Gravitational Radiation of SN
During of supernova flare, an energy of E ≈ × erg is released. Neutrinoscarry out ∼ . E ; the characteristic energy of runaway neutrinos is ∼
10 MeV.In accordance with the standard theory of supernova explosion, it is the neutrinocollapse mechanism that causes the shock wave of the rebound and, therefore, theexpansion of the outer layers. There are two main processes in which neutrinos areproduced: (i) capture of electrons by protons of the iron nucleus: p + + e − → n + ν e ; neutrinoradiation occurs at initial stage with a huge luminosity L ∼ erg/s, duringof a very short time: τ (cid:46) . e + + e − → (cid:101) ν i + ν i , where i = µ, τ, e ; here the neutrinoemission occurs during of a time τ ∼
10 s with luminosity L ∼ erg/s. Justin this process the electronic antineutrinos are released that can be detected byBUST. In the program of searching for a neutrino burst from SN, the BUSTsensitivity radius is approximately 20 kps.ovember 12, 2019 1:56 WSPC/INSTRUCTION FILE OGRAN˙as˙multi-messenger˙project
Gravitational wave detector OGRAN as multi-messenger project of RAS-MSU GW radiation from SN occurs during a nuclear rebound due to the non-spherical collapse (there is a third derivative of the quadrupole moment). The ex-pected value of the dimensionless GW amplitude, which came from a distance of ∼
10 kpc, is h (cid:46) − . The characteristic frequency of the wave packet is f ∼ to 10 erg. As itwas shown in the Section 2 this forecast is a bit below of OGRAN sensitivity in itspresent version.BUST is located at BNO RAS an effective depth of 850 m. The setup consists of3184 scintillation counters, the total mass of the scintillator is 330 t. The scintillationcounter is an aluminum container 0 . × . × . in size, filled with a scintillatorbased on white spirit ( C n H n +2 , n ≈ ν e + p → n + e + . At the average antineutrino energyof 12 −
15 MeV, the range of length of the positron born in such reaction will beenclosed in the volume of one counter. Thus, the search for a neutrino burst consistsin registering a cluster (group) of single events during the time interval τ = 20 s- a typical duration of a neutrino burst from SN. For SN at a distance of 10 kpc,the total energy emitted into the neutrino is 3 × erg and the target mass is130 t (three lower BUST planes), the estimate of the expected number of eventsfrom a single collapse is N ν ≈
35. Of course, there is a random background ofsuch events created by the radioactivity of the environment, muons of cosmic rays,false alarms of counters, etc. The background is such that it creates a cluster of k = 8 single events at a rate of 0 .
138 year -1 . In 10 years, no more than 2 events canbe expected. The cluster formation rate from k = 9 background events is alreadyfalling significantly and is equal to 7 × − year -1 . Thus, clusters with k > ∼
21, 22 forecasting of multi-radiation flux during of a totalcollapse duration ∼
20 s. In our analysis below we take it into account. The numer-ical modeling of GW radiation from non-spherical collapse
21, 22 gives of the enoughcomplex pulse form (Fig. 4). But in this paper for simplicity we use an approxima-tion of more intensive part of the burst as a very short radio pulse with a smoothGaussian envelope and a resonant carrier (1 − V. N. Rudenko, Yu. M. Gavrilyuk, A. V. Gusev et al.
5. Strategy of Two Channel Search for Collapsars
The search for collapsars at the BUST neutrino telescope has been conducted forover 25 years. The result is an experimental estimate of the upper boundary of thefrequency of occurrence of such events in the Galaxy ∼ .
03 year -1 . By connectingthe OGRAN gravitational antenna to the observations, it is potentially possible toincrease the search efficiency. In this regard, it is important to develop a procedurefor joint data processing of both tools. Below there is a variant of such processing. Itsprinciple is to improve the detection characteristics of the gravitational antenna dueto additional information from the neutrino telescope about the current backgroundof neutrino events. In turn, the control record of the gravitational detector canfurther reduce the number of candidates (suspicious bursts) for signal neutrinosfrom collapses. Specifically, the analysis of the neutrino background allows us tonarrow the time interval for the search for GW - disturbances on a gravitationalantenna. Let’s explain it more in detail. Fig. 3. Neutrino luminosity of collapsing star M = 20 M (cid:12) , t b − time after bounce (R.Mayle,J.R.Wilson and D.N.Schramm, Astrophys.J. , 288 (1987)). ovember 12, 2019 1:56 WSPC/INSTRUCTION FILE OGRAN˙as˙multi-messenger˙project
Gravitational wave detector OGRAN as multi-messenger project of RAS-MSU Description of GW detector data processing
Let’s consider random process on the OGRAN output: x ( t ) = λs ( t ) + n ( t ) , (1)where λ = (0; 1) - detection parameter, s ( t ) - signal response, n ( t ) - additive Gaus-sian noise with known spectral density N ( ω ). For a short GW-burst the equivalentforce at the detector input formally can be presented as: F s ( t ) = Aδ ( t − τ ); where A - unknown amplitude, τ - unknown moment of arrival time with uncertainty in aprior interval ( τ min ; τ max ). The additive mixture x ( t ) is processed according to theoptimal filtering scheme x ( t ) → M F → AD → R ( t ) → max R ( t + t ) ≥ c α f or t ∈ ( τ min ; τ max ) , (2)where the blocks of processing marked as: M F - matched filter; AD - amplitudedetector getting an envelope of a narrow-band realization at the output of M F , t -time delay introduced by M F , c α - threshold lever of Neyman-Pearson’s receiver defined by false alarm error α (an exceeding of threshold means a presence of thegravitational signal or λ = 1). Let σ is the variance of Gaussian noise at the M F -output, ∆ f - the bandwidth; then in first approximation c α ≈ σ (cid:114) Bα , (3)
Fig. 4. Structure of GW burst from collapse of rotation star M = 20 M (cid:12) , t b - time after bounce(H.Dimmelmeier, C.D.Ott, A.Marek and H.T.Janka, Phys. Rev. D , 064056 (2008)). ovember 12, 2019 1:56 WSPC/INSTRUCTION FILE OGRAN˙as˙multi-messenger˙project V. N. Rudenko, Yu. M. Gavrilyuk, A. V. Gusev et al. where the new dimensionless parameter B is introduced so that B ≈ ( τ min − τ max )∆ f . The physical information of a close time position of neutrino and GWsignals (neutrino-gravitational correlation) allows us to reduce the prior intervalof searching ( τ min , τ max ), choosing | τ − t c | < s ; where t c the start moment ofcollapse. It decreases of the threshold c α and corresponding value of R ( t + t ). Principle of joint data processing
In our paradigm of neutrino gravitational correlations, the probability of missing ofgravitational signal on GW detector output β ∗ is composed by two terms β ∗ = β + p e (1 − β ) , (4)where the first is β - probability of missing of GW signal with unknown arrival time τ at the selected observational interval, and the second contains a p e - probabilityof error in choosing of this interval, which is defined as p e = α + p m (1 − α ) , (5)where α - false alarm rate in the neutrino channel, p m - probability of error inchoosing of the maximum neutrino signal which is calculated below. By taking intoaccount expected duration of gravitational collapse T ≈
20 s we use a neutrinorolling counter which counts neutrino events at following T = 20 s intervals ∆ n ( t ) = n ( t ) − n ( t − T ): if the number of neutrino events at some interval ( t − T , t ) exceedsthe threshold level ∆ n = ∆ n ( t ) > ∆ n α ; then one considers of possible appearingof GW signal on this interval. Supposing of Poissone low for the neutrino particlesone can use of general formula of probability to get no more the m -counts for thepulse flux with average velocity of counting (cid:104) m (cid:105) : P ( m ; (cid:104) m (cid:105) ) = 1 − Γ( m + 1 , (cid:104) m (cid:105) )Γ( m + 1) , (6)where Γ( a, z ) - incomplete Gamma-function, Γ( a ) Gamma function. The quite(standard) condition of neutrino detector corresponds to the following relation1 − α = P (∆ n α + 1; (cid:104) ∆ n (cid:105) ) , (7)where (cid:104) ∆ n (cid:105) - average number of noise neutrino events on observation interval. Nowone can consider fine structure of neutrino events on selected individual interval.Let’s divide the random process ∆ n ( t − T ≤ t ≤ t ) into sub-intervals with length,say: ∆ t ≈ m i events. Fromthe set of sub-intervals one selects of those which satisfy of the local maximumcondition: ∆ m i ≥ ∆ m α ; ∆ m i ± < ∆ m i , where ∆ m α - threshold level for sub-intervals. This procedure reflects the presentsof local maximums on the theoretical picture of neutrino luminosity (Fig. 3) corre-sponding to core collapse bounces. In that case we can define α as1 − α = P (∆ m α ; (cid:104) ∆ m (cid:105) ) , ovember 12, 2019 1:56 WSPC/INSTRUCTION FILE OGRAN˙as˙multi-messenger˙project Gravitational wave detector OGRAN as multi-messenger project of RAS-MSU where (cid:104) ∆ m (cid:105) - average number of noise neutrino events on sub-intervals. At lastthe probability p m required above is calculated according to the formula p m = [1 − P (∆ m i + 1; (cid:104) ∆ m i (cid:105) )] P (∆ m i +1 ; (cid:104) ∆ m (cid:105) ) P (∆ m i − ; (cid:104) ∆ m (cid:105) ) . (8)Formulas (3), (4), (7), (8) present the statistical estimate of registration of a collaps-ing object through the joint processing of two channel data provided by neutrinoand GW detectors.A more detailed comment is as follows. Formula (4) gives the probability of miss-ing of the gravitational signal at the threshold level c α (3) with optimal processing(2) of the gravitational antenna output obtained in the time interval ( τ min ; τ max )prompted by the BUST neutrino detector data (in vicinity of the suspicious neutrinoreference). Above the probability of false alarm (chance of occurrence) c α due to thethermal noise for OGRAN and Poisson background for BUST was introduced. It isreasonable to accept the same value in both cases without serious damage to thegenerality of analysis. In fact, both statistics affect the solution to the detection ofa multi-channel event: the thermal one - to set the registration threshold (3), Pois-son one - to optimize the choice of the search interval of gravitational bursts. Totest this technique, a computer simulation is planned with the injection of artificial(calibration) signals into both detectors.
6. Discussion of Results and Development Prospects
In the first part of this article the new qualities of the GW antenna OGRAN afterits modernization were described. The main thing is the expansion of the receptionband due to the increased finesse of FP resonators. But the ultimate sensitivitystill cannot exceed the level 10 − in metric perturbations (limited by the thermalBrownian noise of the acoustic detector). The radical way to overcome this barrieris associated with deep cooling (up to ∼
10 K) of the acoustic resonator. Thisexpensive enterprise requires a significant increase in funding from the experienceof cryogenic detectors Explorer and Nautilus. However, in Ref. 25 an intermediate order of magnitude less costly option wasproposed, associated with cooling the acoustic detector to a nitrogen temperatureof ∼
80 K. The vacuum chamber of the OGRAN detector can relatively easilybe transformed into a nitrogen cryostat as a result of adding an internal nitrogenbath and enveloping the detector with a thermally insulating screen shell. Thecorresponding design was developed and created for the cryo-OGRAN pilot model. The results of test experiments on this setup do not refute the possibilityof increasing the sensitivity of such a nitrogen version of the antenna to a level of ∼ × − Hz -1/2 . This already corresponds to the comparable radii of the locationzone of detected collapses (supernova explosions) along both observational channelsof neutrino and gravitational.It is the place to recall a well known case of detecting the neutrino and grav-itational signals from SN1987A.
27, 28
This was the only example of a two channelovember 12, 2019 1:56 WSPC/INSTRUCTION FILE OGRAN˙as˙multi-messenger˙project V. N. Rudenko, Yu. M. Gavrilyuk, A. V. Gusev et al. correlation between the signals detected by a neutrino telescope and by bar grav-itational wave detectors with piezo sensors operating at room temperature. Lateron, only neutrino events were recognized as the first detection of a neutrino fluxfrom a collapsing star. However, in the process of analysis, some algorithm of amulti-channel detection was proposed and applied. That algorithm was based onthe empirical estimates of the coincidence between the signals of different naturewith arbitrary relative time shifts of both data sets. The fact of the neutrino -gravitational correlation in the case of SN1987A was not confirmed,
30, 31 becausethe sensitivity of the Webers type resonance bar detector operating at room tem-perature was insufficient for detecting of signals of astrophysical origin. The opto-acoustical antenna OGRAN even without cooling has the sensitivity 2 −
27, 28 because it takes into account the fine structure of neutrino events inareas of records where neutrino anomalies are present.In conclusion, we note the importance of the very problem of detecting neutrino-gravitational signals from collapse. Here there is a much richer Physics than thatcontained in gravitational bursts (chirps) emitted during the merger of relativisticbinaries. In fact, the chirp structure in the inspiral stage is predicted quite wellalready in the framework of Newtonian theory and provides information on theparameters of the binary (mass, half-axis, frequency). The subtle relativistic detailsof the binary are not yet resolvable, as are nuclear processes in the merging stage.On the contrary, the temporal structure of neutrino and gravitational bursts from acollapsing star is just an indicator of the nuclear processes taking place in it.
17, 20, 22
In particular, bounces in the monotonic compression course indicate a change in theequation of state of nuclear matter with increasing density, temperature, etc. Thisargument is the main motive of the BNO INR RAS program for the two-channelsearch for collapsars in the Galaxy.
Acknowledgments
The authors express their gratitude to the organizers of the Friedman Conferencein June 2019: V. M. Mostepanenko, G. L. Klimchitskaya and Yu. V. Pavlov, forthe invitation to participate and submit their works to this high-level Internationalmeeting.
References
1. LIGO Scientific Collab. and Virgo Collab. (B. P. Abbott et al .),
Phys. Rev. Lett. ,061102 (2016).2. LIGO Scientific Collab. and Virgo Collab. (B. P. Abbott et al .),
Phys. Rev. Lett. ,221101 (2017). ovember 12, 2019 1:56 WSPC/INSTRUCTION FILE OGRAN˙as˙multi-messenger˙project
Gravitational wave detector OGRAN as multi-messenger project of RAS-MSU
3. LIGO Scientific Collab. and Virgo Collab. (B. P. Abbott et al .),
Phys. Rev. Lett. ,141101 (2017).4. LIGO Scientific Collab. and Virgo Collab. (B. P. Abbott et al .),
Phys. Rev. Lett. ,161101 (2017).5. A. von Kienlin, C. Meegan and A. Goldstein,
GRB Coordinates Network, CircularService (2017).6. V. Savchenko et al ., The Astrophys. J. , L15 (2017).7. V. M. Lipunov et al ., The Astrophys. J. , L1 (2017).8. M. Punturo et al ., Classical and Quantum Gravity , 084007 (2010).9. LIGO Scientific Collab. Document Control Center
T1500290-v3 (2016).10. R. X. Adhikari,
Rev. Mod. Phys. , 121 (2014).11. S. N. Bagaev et al ., Review of Scientific Instruments , 114 (2014).12. V. V. Kulagin, A. G. Polnarev and V. N. Rudenko, Sov. Phys. JETP , 319 (1977).13. I. Bichak and V. N. Rudenko, Gravitational Waves in GR and the Problem of TheirDetection (Moscow University, Moscow, 1987).14. R. W. P. Drever et al ., Applied Physics B , 97 (1983).15. W. Zhang et al ., Opt. Lett. , 1980 (2014).16. H. A. Bethe and J. R. Wilson, Astrophys. J. , 14 (1985).17. R. Mayle, J. R. Wilson and D. N. Schramm,
Astrophys. J. , 288 (1987).18. Yu. F. Novoseltsev et al ., JETP , 73 (2017).19. H. Dimmelmeier et al ., Phys. Rev. D , 064056 (2008).20. T. Melson et al ., Astrophys. J. , L42 (2015).21. V. S. Imshennik and O. G. Ryazhskaya,
Astronomy Letters , 14 (2004).22. G. S. Bisnovatyi-Kogan and S. G. Moiseenko, Physics-Uspekhi , 843 (2017).23. B. R. Levin, Theoretical Basis of Statistical Radio Engineering , 3rd edn. (Radio andSvyaz, Moscow, 1989).24. IGEC-2 Collab. (P. Astone et al .),
Phys. Rev. D , 102001 (2007).25. V. V. Kulagin et al ., Physics of Atomic Nuclei , 1552 (2016).26. N. N. Kvashnin et al ., Physics of Atomic Nuclei , 1606 (2017).27. M. Aglietta et al ., Europhysics Letters , 1321 (1987).28. E. N. Alexeyev et al ., Physics Letters B , 209 (1988).29. K. S. Hirata et al ., Phys. Rev. D , 448 (1988).30. C. A. Dickson and B. F. Schutz, Phys. Rev. D , 2644 (1995).31. V. N. Rudenko et al ., JETP91