The International Pulsar Timing Array Maura McLaughlin West Virginia University June 13 2011 Outline • Pulsar timing for gravitational wave detection • Pulsar timing arrays • EPTA, NANOGrav, PPTA • The International Pulsar Timing Array Science Goals • The International Pulsar Timing Array organization • The future of the International Pulsar Timing Array Gravitational Waves GWs are fluctuations in the fabric of spacetime traveling at the speed of light. plus-polarized cross-polarized h + hx We describe them by their strain (or amplitude) and their frequency. h = strain = ΔL/L = 0.5 Gravitational Wave Sources Masses with varying quadrupole moments will emit GWs. Which of the following are GW sources? - spinning spherical stars - pulsing stars - spinning “bumpy” spheres - stars in binary orbits - supernovae Proof of Gravitational Radiation 5/3 2/3 Pb − 1 2/3 ω˙ =3T 2 (mp + mc) , (1) ! !2π " 1 e 1/3 − 2/3 Pb mc(mp +2mc) γ = T e 4/3 , (2) ! !2π " (mp + mc) r = T mc, (3) ! 2/3 2/3 1/3 Pb − (mp + mc) s = sin i = T − x , (4) ! !2π " mc 5/3 ˙ 192π 5/3 Pb − mpmc Pb = T f(e) 1/3 , (5) Hulse & Taylor, 1975, ApJ, 195, L51 − 5 ! !2π " (mp + mc) Weisberg & Taylor, 1984, PRD, 52, 1348 How can we directly detect them? Search for light travel time changes between objects. LIGO LISA PTAs f ~ 1/ms (100 – 1000 Hz) f ~ 1/(mins-hrs) (10-2 – 10-3 Hz) f ~ 1/yrs (10-8 – 10-9 Hz) GW Sources Continuous Burst Stochastic Pulsar Timing Arrays PTAs depend on extreme stability of MPs Residual = measured – expected pulse arrival times -5 σRMS = root-mean-square residual < 200 ns (3 x 10 P). At midnight on April 6th 2001, the period was J0437-4715 P = 5.75 ms 5.757451924362137(2) ms!!!! Verbiest et al., 2008, ApJ, 400, 951 Pulsars for Direct GW Detection f ~ 1/weeks to 1/years (10-6 – 10-9 Hz) -15 h ~ σrms/Τ ~ 100 ns/5 years ~ 10 d ∆t ninj hij(x, t)dr (1) ∝ !0 Δt ~ h/f and will have pulsar and Earth terms λgw ~ 1-10 lyr; Dpsr ~ 1000 lyr First discussed by Detweiler (1979, ApJ, 234,1100) and Sazhin (1978,SA,22,36). Pulsars for Direct GW Detection P. Demorest Single Source: 3C66B Proposed z=0.02 SMBH binary with period 1.05 yr and mass of 5x1010 solar masses (Sudou et al. 2003). Pulsar term and Earth term Single Source: 3C66B Proposed z=0.02 SMBH binary with period 1.05 yr and mass of 5x1010 solar masses (Sudou et al. 2003). Analysis of the residuals from a single high-precision pulsar shows that this binary cannot exist (Jenet et al. 2004). Pulsar term and Earth term f α h (f)=A (1) Stochastic Background c 1 !yr− " P. Demorest Model A α References Supermassive black holes 10-15 – 10-14 -2/3 Jaffe & Backer, 2003, ApJ, 583, 616 Wyithe & Loeb, 2003, ApJ, 590, 691 Enoki et al., 2004, ApJ, 615, 19 Sesana et al., 2008, MNRAS, 290, 192 Relic GWs 10-20 – 10-15 -1 to -0.8 Grishchuk, 2005, PU, 48, 1235 Boyle & Buonanno, 2008,PRD, 78,043531 Cosmic Strings 10-16 – 10-14 -7/6 Maggiore, 2000, PR, 331, 283 Single Pulsar Limits -14 -1 hc < 2 x 10 (f=1 yr ) 1 1 2 13/3 P (f)= h (f) f − (1) 12π2 f 3 c ∝ 1 1 2 13/3 P (f)= 2 3 hc(f) f2− (1) 1 dρgw(f) 1 π12π2 f 2 2∝π 2 2 2/3 Ωgw(f)= = f hc(f) = 2 f hc(f) f (2) ρc dlnf ρc 4 3 H0 2 ∝ 1 dρgw(f) 1 π 2 2 2 π 2 2 2/3 Ωgw(f)= = f hc(f) = 2 f hc(f) f Kaspi, Taylor, & Ryba(2) , 1994, ApJ, 428, 713 ρc dlnf ∞ ρc 4 2 3 H0 ∝ P (f)df = σg (3) 0 ! ∞ P (f)df = σ2 (3) !0 Stochastic Background Limits from a Pulsar Timing Array Expected correlation of residuals for pairs of pulsars versus angular separation on sky. “Pulsar” terms uncorrelated. “Earth” terms correlated. N 1 1 − r(θ)= R(t , kˆ )R(t , kˆ )(1) N i 1 i 2 !i=0 r(θ) = σ2ζ(θ)(2) ! " g Jenet et al. 2005, ApJL, 625, 123 3 x 1 1 ζ(θ) = xlogx + + δ(x)(4) ! Clock" 2 errors− 4monopole.2 2 Ephemeris errors dipole. x = [1 cos(θ)]/2(5) GWs quadrupole− . δ(x)=1forx = 0 and is 0 otherwise. Hellings & Downs, 1983, ApJ, 265, L39 Sensitivity of a PTA T = total timespan of observations PATIENCE NTOAs = number of TOAs HIGHER CADENCES (MORE TELESCOPE TIME) NPSR = number of pulsars PULSAR SEARCHES (MORE TELESCOPE TIME) σrms = timing residual RMS INTRINSIC and EXTRINSIC factors Intrinsic: rotational and emission stability, ISM, pulse shape, brightness Extrinsic: observation length, collecting area, bandwidth, frequency, instrumentation, algorithms Also depends on distribution of sources! Sensitivity of a PTA A pulsar timing array is a detector (or telescope) just like LIGO is. We can calculate properties just like LIGO. Antenna response for the pulsar-Earth system. MSP Sky Distribution Galactic MSPs are local (distances ~ kly) and roughly isotropically distributed. There are roughly 100 MSPs (P < 20 ms) in our Galaxy. Timing precision w w Tsys 1 σTOA (1) ∼ SNR ∝ SPSR A √BT w – pulse width SNR - signal-to-noise ratio SPSR – pulsar flux Tsys – system temperature A – telescope area B – receiver bandwidth T – integration time Pulsar Timing Array A pulsar timing array is a detector (or telescope) just like LIGO is. We can calculate properties just like LIGO. Antenna response for the pulsar-Earth system. Collaboration and Pooling of Data Necessary! NANOGRav includes 22 members from North America nanograv.org Current Results J1713+0747 < 100 ns RMS! J1909-3744 P. Demorest 26 MSPs being timed with the Arecibo and Green Bank telescopes. 2-3 frequencies at each telescope. Bi-monthly to monthly observations. Residuals from 100 ns to 1.5 μs NANOGrav Limit after Four Years -15 -1 hc < 7 x 10 (f=1 yr ) or P. Demorest Sensitivity of a PTA T = total timespan of observations PATIENCE NTOAs = number of TOAs HIGHER CADENCES (MORE TELESCOPE TIME) NPSR = number of pulsars PULSAR SEARCHES (MORE TELESCOPE TIME) σrms = timing residual RMS INTRINSIC and EXTRINSIC factors Intrinsic: rotational and emission stability, ISM, pulse shape, brightness Extrinsic: observation length, collecting area, bandwidth, frequency, instrumentation, algorithms The International Pulsar Timing Array The International Pulsar Timing Array The International Pulsar Timing Array J. Verbiest 2012 Meeting • Gravitational waves are a key prediction of general relativity. • Their direct detection will allow us to probe new source classes. • Both ground-based interferometers and PTAs are based on searching for changes in light travel time between objects. • Source classes that will be studied by both are complementary. • Detection is challenging and improvements are need to both interferometers and current PTAs. The IPTA has been formed to facilitate a detection with pulsars. • Time to direct detection ~ 5-10 years. Summary • Gravitational waves are a key prediction of general relativity. • Their direct detection will allow us to probe new source classes. • Both ground-based interferometers and PTAs are based on searching for changes in light travel time between objects. • Source classes that will be studied by both are complementary. • Detection is challenging and improvements are need to both interferometers and current PTAs. The IPTA has been formed to facilitate a detection with pulsars. • Time to direct detection ~ 5-10 years. .