1. Michele Punturo INFN Perugia and EGO On behalf of the ET Design Study Team http://www.et-gw.eu/ 8th Amaldi Conference, New York, 2009 1

2.  Why 3 rd generation GW observatories?  Science potentialities  Technologies for the 3 rd generation GW detectors  The Einstein Telescope design study 8th Amaldi Conference, New York, 2009 2

3. 3 Scientific Run / LIGO - Virgo Commis -sioning/ Upgrade eLIGO, Virgo+ Commis- sioning Scientific run(s) 200720092008 Same Infrastructures, improvements of the current technologies, some prototyping of the 2 nd generation technologies

4. 8th Amaldi Conference, New York, 2009 4 Scientific Run / LIGO - Virgo Commis -sioning/ Upgrade eLIGO, Virgo+ Commis- sioning Scientific run(s) Upgra des & Runs advLIGO, advVirgo Com- mis- sioning Scientific run(s) 200720092011-122015200820172022 Same Infrastructures, improvements of the current technologies, some prototyping of the 2 nd generation technologies Same Infrastructures, engineering of new technologies developed by currently advanced R&D BNS Either the detection is reached or there is something fundamentally wrong

5. 8th Amaldi Conference, New York, 2009 5 Scientific Run / LIGO - Virgo Commis -sioning/ Upgrade eLIGO, Virgo+ Commis- sioning Scientific run(s) Upgra des & Runs advLIGO, advVirgo Com- mis- sioning Scientific run(s) 200720092011-122015200820172022 Upgrades (High frequency oriented?) and runs Same Infrastructures, improvements of the current technologies, some prototyping of the 2 nd generation technologies Same Infrastructures, engineering of new technologies developed by currently advanced R&D

6.  GW detection is expected to occur in the advanced detectors. The 3 rd generation should focus on observational aspects:  Astrophysics:  Measure in great detail the physical parameters of the stellar bodies composing the binary systems  NS-NS, NS-BH, BH-BH  Constrain the Equation of State of NS through the measurement  of the merging phase of BNS  of the NS stellar modes  of the gravitational continuous wave emitted by a pulsar NS  Contribute to solve the GRB enigma  Relativity  Compare the numerical relativity model describing the coalescence of intermediate mass black holes  Cosmology  Measure few cosmological parameters using the GW signal from BNS emitting also an e.m. signal (like GRB)  Probe the first instant of the universe and its evolution through the measurement of the GW stochastic background  Astro-particle:  Contribute to the measure the neutrino mass 6 8th Amaldi Conference, New York, 2009

7.  Advanced detectors will be able to determine BNS rates in the local Universe  “Routine” detections at low to medium SNR  But high precision fundamental physics, astrophysics and cosmology may be limited  would require good quality high-SNR events  Extremely rare in advanced detectors 8th Amaldi Conference, New York, 2009 7  Let suppose to gain about one order of magnitude in sensitivity in the full 1-10000Hz range  It will permit to access a larger amount of information embedded in the BS (BNS, BH-NS, BH-BH) chirp signal  Higher harmonics  Merging phase

8.  The GW emission of the coalescence of a NS binary system is often computed through the Post-Newtonian approximation  Amplitudes and phases of GWs are obtained as series expansions in v/c.  The current detection pipelines are based on the matched filter algorithm:  More sensitive to the phase than to the amplitude modulation  The current “template” construction is essentially based on the fundamental harmonic (f GW =2f Orb ), with an higher approximation order in phase than in amplitude  “Restricted PN approximation”  But, if the two star masses in the BS are unbalanced, higher multipole moments are associated to the source  Higher harmonics  Need of a full PN approximation 8th Amaldi Conference, New York, 2009 8

9.  Higher harmonics could play an important role depending on the masses, mass asymmetry and the inclination angle 8th Amaldi Conference, New York, 2009 9 Harmonics PN corrections McKechan et al (2008)

10.  The first consequence of the higher harmonics is a richer spectrum of the signal detected by the ITF 8th Amaldi Conference, New York, 2009 10 McKechan et al (2008) Plots normalized to LIGO I Dominant harmonic 5 harmonics

11.  Higher harmonics do not greatly increase overall power, but push power toward higher frequencies, which can make higher-mass systems detectable even if quadrupole signal is outside the observing band  BBH improved identification 8th Amaldi Conference, New York, 2009 11 ET Restricted ET Full Van Den Broeck and Sengupta (2007)

12.  Harmonics increasing the structure, greatly enhance parameter determination, by breaking degeneracy between parameters 8th Amaldi Conference, New York, 2009 12  Antenna response is a linear combination of the two polarizations:  H + and H × contain the “physics of the source” (masses and spins) and time and phase at coalescence  A + ( , , ,D L,i) and A × ( , , ,D L,i) contain the “geometry of the source- detector system”  Right ascension  Declination  Polarization angle  Luminosity distance  Orientation of the binary wrt the line of sight Credits: B.Sathyaprakash To fully reconstruct the wave one would need to make five measurements: ( , , ,D L,i)

13.  Restricted PN approximation can only measure the random phase of the signal at the coalescing time  To fully determine a source are needed  either 5 co-located detectors (“a la sphere”)  or 3 distant detectors (3 amplitudes, 2 time delays)  Detecting the harmonics one can measure the random phase of the signal with one harmonic, orientation of the binary with another and the ratio A + /A × with the third  Two detectors at the same site in principle allow the measurement of two amplitudes, the polarization, inclination angle and the ratio A + /A × – the source can be fully resolved  In practice, because of the limited accuracy, two ET observatories could fully resolve source:  4 amplitudes from two sites, one time delay 8th Amaldi Conference, New York, 2009 13 Credits: B.Sathyaprakash Networking is still a “must” !

14.  In principle, the right model of the merging of a black holes binary should fully use the General Relativity (GR)  Unable to analytically solve the Einstein Field Equation:  Use of the Numerical Relativity (NR)  There is no fundamental obstacle to long-term (i.e. covering ~10+ orbits) NR calculations of the three stages of the binary evolution: inspiral, merger and ringdown  But NR simulations are computationally expensive and building a template bank out of them is prohibitive  Far from the merging phase it is still possible to use post-Newtonian approximation  Hybrid templates could be realized and carefully tested (in case of high SNR detection) overlapping in the phenomenological template the PN and NR waveforms 8th Amaldi Conference, New York, 2009 14 Red – NR waveform Black – PN 3.5 waveform Green – phenomenological template Credits: Bruno Giacomazzo

15.  The late coalescence and the merging phase contain information about the GR models  Test it through ET will permit to verify the NR modeling  This is true also for the NS-NS coalescence where the merging phase contains tidal deformation modeling and could constrain, through numerical simulations, of the Equation Of State (EOS) 8th Amaldi Conference, New York, 2009 15

16.  EOS of the NS is still unknown  Why it pulses?  Is it really a NS or the core is made by strange matter?  Gravitational Wave detection from NS will help to understand the composition of the NS:  Trough asteroseismology, revealing the internal modes of the star  Trough its continuous wave emission 8th Amaldi Conference, New York, 2009 16

17. Measuring the frequency and the decay time of the stellar mode it is possible to reconstruct the Mass and the Radius of the NS Measuring the frequency and the decay time of the stellar mode it is possible to reconstruct the Mass and the Radius of the NS Knowing the Mass and the Radius it is possible to constrain the equation of state (EOS) Knowing the Mass and the Radius it is possible to constrain the equation of state (EOS) 8th Amaldi Conference, New York, 2009 17 Credits: B.Schutz

18.  A fraction of the NS emits e.m. waves:  Pulsars  These stars could emit also GW (at twice of the spinning rotation) if a quadrupolar moment is present in the star:  ellipticity  The amount of ellipticity that a NS could support is related to the EOS through the composition of the star:  i.e. high ellipticity  solid quark star?  Crust could sustain only  ~10 -6 -10 -7  Solid cores sustains  ~10 -3  Role of the magnetic field? 8th Amaldi Conference, New York, 2009 18 Imagine.gsfc.nasa.gov

19. 8th Amaldi Conference, New York, 2009 19 Upper limits placed on the ellipticity of known galactic pulsars on the basis of 1 year of AdVirgo observation time. Credit: C.Palomba Credits: B.Schutz

20.  Improving the sensitivity of the advanced detectors by an order of magnitude will permit to access, for the BNS observation, cosmological distances in the universe 8th Amaldi Conference, New York, 2009 20 Credits: G. Jones Virgo Coma Observed mass Intrinsic mass Z=1 Z=3

21.  BNS are considered “standard sirens” because, the amplitude depends only on the Chirp Mass and Effective distance  Effective distance depends on the effective Luminosity Distance and the antenna pattern  Through the detection of the BNS gravitational signal (using three detectors or through the higher harmonics) it is possible to reconstruct the luminosity distance D L :, but with an ambiguity due to the red-shift 8th Amaldi Conference, New York, 2009 21  In fact, the red-shift is entangled with the binary’s evolution:  D L  D L (1+z)     (1+z)  The observer wrongly reconstruct the total mass of the binary system Credits: B. Sathyaprakash, B. Schutz, C. van den Broeck M tot  M tot (1+z)

22.  Hence, GWs, measuring the chirp mass and the waveform amplitude are able to determine the luminosity distance D L of the red-shifted BNS, but need an e.m. counterpart to measure the red-shift z.  i.e. Short GRB are currently considered to be generated by coalescing BNS  Coupling GW and e.m. measurements it is possible to determine the origin of GRB  Knowing (1+z) and D L it is possible to test the cosmological model and parameters that relate these two quantities: 8th Amaldi Conference, New York, 2009 22  M : total mass density   : Dark energy density H 0 : Hubble parameter w: Dark energy equation of state parameter  B.Sathyprakash and co., show that, thanks to the huge number of BNS coalescences visible by a 3 rd generation GW detector, the cosmological parameters can be determined with only a few percent of error (B. Sathyaprakash, B. Schutz, C. van den Broeck in prep).

23. 8th Amaldi Conference, New York, 2009 23

24.  Supernova Explosions generates visible light, huge neutrino emission and GW emission  Large  t between light arrival and neutrino arrival times could occur, due to the interaction of the light in the outer layer of the collapsing star  3 hours according to the SNEWS (SuperNova Early Warning System) in the SN1987A  Small  t expected for GW detectors:  Where  t SN depends on the different models adopted to describe the neutrinos and GW emission (   t SN <1ms)   t det depends on the detectors reciprocal distance and on the source direction   t prop depends only on the mass of the neutrino and on the distance L of the source: N.Arnaud et al, Phys.Rev.D65,2002 24 8th Amaldi Conference, New York, 2009  A mass measurement accuracy lower than 1 eV could be reached  But, the event rate?

25.  Current SNe models estimates about 10 -8 M  emitted in GW; in the past larger fractions were expected  1 st generation were limited to our galaxy:  Event rate: 1evt every 20 years  To reach an acceptable event rate (1evt/year) a sight sphere of 3-5Mpc should be considered (Christian D Ott, 2008)  Events, optically detected, from 1/1/2002, to 31/8/2008 < 5MPc: 25 8th Amaldi Conference, New York, 2009

26.  Upper limit SNR estimated 26 advLIGO ET: ~ × 10 8th Amaldi Conference, New York, 2009

27.  The target for the design of a 3 rd generation GW detector is to gain roughly a factor 10 in sensitivity (broadband) with respect to the advanced detectors  Driven by the “precision astronomy” objective  Need of high SNR  What are the ideas under evaluation? 8th Amaldi Conference, New York, 2009 27

28. 8th Amaldi Conference, New York, 2009 28 10 -25 10 -16 h(f) [1/sqrt(Hz)] Frequency [Hz] 1 Hz10 kHz Seismic Thermal Quantum Seismic Newtonian Susp. Thermal Quantum Mirror thermal 3 rd generation ideal target

29. Actions:Drawbacks/Difficulties  Step 1: increase of the arm length:  h=  L/L 0 : 10 km arm, reduction factor 3.3 respect to Virgo  New infrastructure!!  Step 2: reduction and optimization of the quantum noise (HF):  Increase of the laser power from 125W to 500W  Optimization of the optical parameters (signal recycling, …)  Introduction of a 10dB squeezing  New site, new costs  Angular noises  …  Laser R&D difficulties  Thermal lensing effects  Optical losses crucial 8th Amaldi Conference, New York, 2009 29

30.  The power step requested for the 3 rd generation GW detectors is linear with respect to the increase required for the advanced detectors 8th Amaldi Conference, New York, 2009 30 Laser power [W] Year  Key task:  Dominate thermo- optical effects Virgo Virgo + adV/adL ET  Reduce deposited heat and thermal gradients  Compensate thermal aberrations  “New” possible technology under “initial” investigation  Rare earth doped fiber amplifier  High power amplification available  NUFERN 150W single frequency amplifier  50 kW multimode amplifier  Still noisy : learn how to stabilize  New (longer) wavelenght probably needed Credits: A.Brillet  LZH is aiming to reach 1kW within 5 years (solid state laser @ 1064nm)

31.  Squeezed light states injection in a GW detector does not seem anymore a 3 rd generation technology, but it “risks” to become a noise reduction tool in the advanced detectors  GEO is installing his squeezing bench now  LIGO wants to make a test after S6 8th Amaldi Conference, New York, 2009 31  Driven by huge recent progresses Vahlbruch et al, PRL 97, 2006: Demonstration of squeezing on a Michelson inteferometer down to audio bands Vahlbruch et al, PRL 100, 2008: 10dB squeezing @5MHz:

32. 2 nd generation 10km+High Frequency 8th Amaldi Conference, New York, 2009 32 Credits: S.Hild 10 -26 10 -19

33. 8th Amaldi Conference, New York, 2009 33 10 -25 10 -16 h(f) [1/sqrt(Hz)] Frequency [Hz] 1 Hz10 kHz Seismic Thermal Quantum Seismic Newtonian Susp. Thermal Quantum Mirror thermal 3 rd generation ideal target

34.  To improve the sensitivity in the medium-low frequency range the mirror and suspension thermal noise must be decreased.  Minimization of the mirror thermal noise  Reduction of the mechanical dissipations and of the thermo-elastic and thermo-refractive couplings  New materials  Cryogenics  New materials  Larger beams  Larger test masses sizes or new beam geometries 8th Amaldi Conference, New York, 2009 34

35. 8th Amaldi Conference, New York, 2009 35 SourceSymbolSpectral densityCrucial Parameters Coating brownian T, 1/w,  coat Bulk brownian T, 1/w,  sub Coating Thermoelastic T 2, 1/w 2,  coat Bulk Thermoelastic T 2, 1/w 3,  sub Coating Thermorefractive T 2, 1/w 2,  Bulk thermorefractive T 2, 1/w 2,  Credits: J.Franc et al.

36. 8th Amaldi Conference, New York, 2009 36 SiO2SiliconSapphire 300KCRYO300 KCRYO300 KCRYO  5 10 -9 10 -7 assumed 2 10 -8 ~3 10 -9 2 10 -9 4 10 -9  - W/(m.K) 1.380.13 @10K 130297 @4K 2330 @10K 40110 @4.2K 1500 @ 12.5K C - J/(K.Kg) 7461-7@10K7110.276 @10K7900.1 @ 10K  - 1/K 0.51 10 -6 -0.25 10 -6 @ 10K 2.54 10 -6 4.85 10 -10 @10K 5.1 10 -6 5.7 10 -10 @10K  - 1/K 8 10 -6 1.01 10 -6 @30K 2.5 10 -6 5.8 10 -6 @30K 1.3 10 -5 9 10 -8 @20K Credits: J.Franc et al. Material adopted in the LCGT design New possible candidate investigated in the ET Design Study

37. 8th Amaldi Conference, New York, 2009 37 Christian Schwartz @ ET-WP2 workshop P.Amico et al. CQG, 2004

38. 8th Amaldi Conference, New York, 2009 38

39.  Mirror thermal noise will be dominated by the coating dissipation in the advanced detectors  What about 3 rd generation GW detectors? 8th Amaldi Conference, New York, 2009 39  The thermo-mechanical properties of the possible coating materials aren’t so well known at low temperature (and in film format) to permit a reliable simulation.  What the experimental measurements say? Loss angle Martin et al, ET-WP2 meeting Yamamoto et al, CQG. 21 (2004) S1075–S1081

40.  The reduction of the coating dissipation could be obtained reducing the Tantala layer thickness using the waveguide coatings 8th Amaldi Conference, New York, 2009 40 SIO 2 / Ta 2 O 5 : 99.1% reflectivity Brückner et al., Opt.Expr., 17,2009 Si 500 nm R>99.8% Brückner via Schnabel Credits: H.Lueck @ GWADW ‘09

41.  It is possible to gain even more increasing also the beam size  Laguerre-Gauss higher order modes could help:  lower thermal noise & lensing 8th Amaldi Conference, New York, 2009 41 Coatings Standard : SiO 2 -Ta 2 O 5 (HL) 19 HLL : Transmission 4 ppm Substrate Silicon Temperature 10K Mirror dimension Diameter : 62 cm Thickness : 30 cm Beam dimensions Ratio insuring 1 ppm diffraction losses versus order of the LG mode LG33 : w = a/4.3 = 7.2 cm LG00 : w=a/2.6 = 11.9 cm T : 300K Arm length : 3 km Beam radius : 6 cm 10 -27 10 -22 J.Franc et al @ ET-WP3 meeting TTN LG00 TTN LG33

42.  Adopting some* of the solutions listed in the previous slides it is possible to approach the “ideal” 3 rd generation sensitivity  Still many unresolved technical questions 42 Credits: S.Hild 10 -26 10 -19 8th Amaldi Conference, New York, 2009 * Large LG00, Silicon test mass, cryogenics

43. Reduction of the seismic excitation: Reduction of the seismic excitation: Learning from the LISM lesson in Kamioka (Japan) New underground facility ~1 Hz seismic filtering system 8th Amaldi Conference, New York, 2009 43 TrexEnterprises Reduction of the gravity gradient noise (NN) Reduction of the gravity gradient noise (NN) New underground facility Suppression of the NN through correlation measurements? Clever design of the cave doesn’t help (G. Cella) Reduction of the radiation pressure noise Reduction of the radiation pressure noise Heavier mirrors 400mm maximum dimension of Si bulk by wafer producers Large SiC Mirror used in Astronomy But optical quality (in transmission) poorer

44.  Going underground we could gain 2-3 orders in seismic excitation, but we still need to filter 8th Amaldi Conference, New York, 2009 44 Moscow June 2008 Credit: K.Kuroda LISM  A 50 m tall Virgo-like Super- Attenuator? Optimized at 1Hz Courtesy G. Cella Horizontal  Violin modes to be considered Noise requirements could be matched with a 7m long stage

45. 8th Amaldi Conference, New York, 2009 45  Site requirements?  Geological properties?  Human activities in the surface?  The depth?  Seismic noise decreases exponentially going underground  But analytical models foresee a minor Newtonian Noise reduction  Possibility to subtract the residual noise measuring related auxiliary quantities  Network of sensors around the test masses  Modeling in progress Reduction factor Frequency [Hz] G.Cella, Fujihara seminar ‘09

46.  High power on the main cavities is a requirement conflicting with the design options devoted to reduce the low frequency noises:  Cryogenics  Long suspensions  A single broadband detector is probably too difficult to be realized  Possible solution:  Realize a GW observatory composed by two (or more) detectors  Xylophone design (R. De Salvo, 2003) 8th Amaldi Conference, New York, 2009 46

47.  Low Frequency optimized detector:  Underground, Long suspensions, Cryogenic, Low Laser power  High Frequency detector  High laser power, room temperature, “standard suspensions” 8th Amaldi Conference, New York, 2009 47

48. 8th Amaldi Conference, New York, 2009 48 Credits: S.Hild

49.  The infrastructure (tunnels, deep underground cavities, …) are by far the dominant component of a 3 rd generation GW observatory cost  It will be difficult to realize more than one observatory per continent  Is it possible to find a geometry that optimizes the costs and presents some additional benefit? 8th Amaldi Conference, New York, 2009 49

50. 8th Amaldi Conference, New York, 2009 50 L L 45° Fully resolve polarizations 5 end caverns 4×L long tunnels 45° stream generated by virtual interferometry Null stream Redundancy 7 end caverns 6×L long tunnels 60° L’=L/sin(60°)=1.15×L Fully resolve polarizations by virtual interferometry Null stream Redundancy 3 end caverns 3.45×L long tunnels L Equivalent to A Freise et al 2009 Class. Quantum Grav. 26

51. 8th Amaldi Conference, New York, 2009 51

52.  ET is a “design study” of a 3 rd generation GW detector supported (3M€ in 3 years: 2008-2011) by the European Commission under the Framework Programme 7  FP7 - Identification of the future European research infrastructures  The aim is to deliver a conceptual design study of a 3 rd generation gravitational wave (GW) observatory  Feasibility study  Identification of the needed infrastructures  Identification of the crucial technologies  Identification of the fundamental elements of the design  Science case development  Technical details to be defined in a future phase 8th Amaldi Conference, New York, 2009 52 http://www.et-gw.eu/ Next ET general meeting: 14-16 October 2009, Erice, Sicily

53.  ET design study team is composed by all the major groups leading the experimental Gravitational wave search in Europe: Virgo & GEO 8th Amaldi Conference, New York, 2009 53 Participant no.Participant organization nameCountry 1European Gravitational ObservatoryItaly-France 2Istituto Nazionale di Fisica NucleareItaly 3Max-Planck-Gesellschaft zur Förderung der Wissenschaften e.V., acting through Max- Planck-Institut für Gravitationsphysik Germany 4Centre National de la Recherche ScientifiqueFrance 5University of BirminghamUnited Kingdom 6University of GlasgowUnited Kingdom 7NIKHEFThe Netherlands 8Cardiff UniversityUnited Kingdom

54. 8th Amaldi Conference, New York, 2009 54 Scientific Run / LIGO - Virgo Commis -sioning/ Upgrade eLIGO, Virgo+ Commis- sioning Scientific run(s) Upgra des & Runs advLIGO, advVirgo Com- mis- sioning Scientific run(s) 200720092011-1220152008 ET Conceptual Design ET Preparatory Phase and Technical Design Preliminary site preparation 2017 ET Construction 2022 Upgrades (High frequency oriented?) and runs 2010 call for Preparatory phase INFRA-2010-1.1.30: Research Infrastructures for dark matter search, neutrinos, gravitational waves

55.  Third generation gravitational wave observatories will open, together with the future space detectors (LISA), the window of the precision gravitational wave astronomy  Many contributions in Astrophysics, cosmology and fundamental physics  The technologies to realize these infrastructures are under identification, but an intense R&D programme is necessary for their development  Even if the success of the advanced detectors is mandatory, we need to think now about these infrastructures, since the path is very long  A Worldwide collaboration is necessary to design, develop and realize a network of 3 rd generation GW detectors 8th Amaldi Conference, New York, 2009 55

56. 8th Amaldi Conference, New York, 2009 56

57.  LIGO limited the fraction of energy emitted by the Crab pulsar through GW to ~6% (  <1.8×10 -4 )  Virgo, at the start of the next science run could in few weeks set the upper limit for the Vela. 8th Amaldi Conference, New York, 2009 57 Spin down limits (1 year of integration ) Credit: B. Krishnan

58.  BNS are considered “standard sirens” because, the amplitude depends only on the Chirp Mass and Effective distance  Effective distance depends on the Luminosity Distance, Source Location (pointing!!) and polarization  The amplitude of a BNS signal is: 8th Amaldi Conference, New York, 2009 58 where the chirp mass is: Credits: D. E. Holz, S.A.Hughes  The Redshift is entangled with the binary’s evolution:  The coalescing evolution has timescales (Gm i /c 3 ) and these timescales redshift  Since also D L scales with (1+z), a coalescing binary with masses [(1+z)m 1, (1+z)m 2 ] at redshift z is indistinguishable from a local binary with masses [m 1, m 2 ] Credits: B.Schutz

59. 8th Amaldi Conference, New York, 2009 59 Signal strength and noise amplitude [1/sqrt(Hz)] Credits: B.Sathyaprakash If n=0, the Big-Bang- Nucleosynthesis limit is  0 <1.1×10 -5.  The SGWB is characterized by the dimensionless energy density  GW (f): n=0 for standard inflationary theory n>1 for other theories (strings,..)  In principle, co-located ITF could measure the correlation needed to detect the SGWB, but local (low frequency) noises could spoil the measurement