Milky Way Satellite Galaxy Kinematics and Scaling Relations for Dark Matter Searches. Andrew B. Pace. Texas A&M University. Louie Strigari (TAMU)

Transcription

1 Milky Way Satellite Galaxy Kinematics and Scaling Relations for Dark Matter Searches Andrew B. Pace Texas A&M University Louie Strigari (TAMU)

2 Identifying the Particle Nature of Dark Matter Production via Collider Indirect Direction SM SM Direct Detection

3 f the number of photons (above 3 GeV) produced by DM annihilation dicted flux in the Aquarius (Via Lactea II) setup. Dark Matter Gamma Ray Sky Pieri+ 009

4 Dark Matter Gamma Ray Sky: Targets MW Satellite galaxies (dsph) Drlica-Wagner+ 015

5 Dark Matter Flux f ( DW, E, E ) = min max 1 ásvñ Emax dn ò g d g 4p g m de E DM Emin particle physics ò ò ò rdm( r ( l)) dldw, Emin DW l.o.s. Jfactor Astrophysics: determine dark matter profile ( 1

6 mber vely. nent, The Astrophysical Journal Letters, 808:L36 (5pp), 015 August 1. Likelihood functions..1 Binned and unbinned analyses () on to nteronly (3) y disness) (4) d dimetric mpute s that o not spy), fit the on of (5) DM annihilation and dec Velocity anisotropy profile. We use th Schwarzschild and Made-To-Measure well Before fitting the actual dsph kinematic data, methods, we tested as both a as Jeans binned and an unbinned function on a Helmi set of mock data parametrization to describe the velocity analysis (see recent likelihood reviews by Battaglia, & Breddels 013; (mimicking ultrafaint and 013). classical dsphs, seewe Appendix A).the latter, Strigari 013; Walker In this work focus on β0 + β (r/ra )η Baes Both methods have been used as in ingredients the literature, but spherically to date, no symmetβani (r) =, using parametric functions of the 1 + (r/ra )η systematic comparison has been undertaken to test the merits and The Astrophysical Journal, 808:95 (14pp), July 0 Simon et al. ric Jeans equation. This015 technique has already been widely applied limits of each approach (binned analyses can be found in Strigari where the four free parameters are the to dsphs (Strigari et al. 007; Essig et al. 010; Charbonnier et al. et al. 007; Charbonnier et al. 011; unbinned in Strigari et al. anisotropy at large radii β, and the sh 011; Geringer-Sameth et al. 015c). Here, we apply the findings of 008; Martinez et al. 009; Geringer-Sameth et al. 015c). For at the scale radius ra. This parametrizat et al. (015a), where an optimized strategy are proposed the Bonnivard binned analysis, the velocity dispersion profiles σ obs (R)was some of the biases arising in the Jeans ana tofrom mitigate possible stellar biases velocities introduced the Jeans modelling. built the individual (seebysection 3), and the Ret IIwith fewer free ible anisotropy functions likelihood function we use is Osipkov Merrit see Bonnivard et al. + " #, Simon+ 015 N.1.1 binsspherical 1/ Jeans equation * 1 σobs (Ri ) σp (Ri ) (π) Light profile. We use a generalized exp, (8) Lbin = Figure 1. Projected stellar density profile of Ret II, derived from the +σi (Ri ) are considered +σas i (Rcollisionless i) dsph systemsphotometric described (Hernquist 1990; Zhao 1996) the stel i=1galaxies catalog of Koposov et al. (015a). Overplotted (red line)for is the Dark Matter Profiles in dsph Galaxies best-fitting model (we note that the fit is to the unbinned data), which is the by their phase-space distribution function, which obeys the colwhere νs (see Section sum of contributions from Ret II itself and a constant background Zhao (r)for=, & ' lisionless Boltzmann equation. Assuming steady-state, spherical..3). Dotted lines enclose ν 68% CIs the projection of n ( r ) )γ [1 + (r/r )α ](β γ )/α F. (r/r 1 s s J and the Jeans σp (Rirotational ++Ri ) σsupport,. (9) + σsymmetry i = + σobs (R i )+negligible p (Ri +R i ) second-order r Figure 1. (a) DES color magnitude diagram dots, and stars selected for of Reticulum II. Stars within of the center of Ret II are plotted as small black the five free parameters of which are the equation is obtained integrating moments of the phase-space Spherical Jeansby Equations distribution function (Binney & Tremaine 008): spectroscopy with MFS, GIRAFFE, and GMOS (as described in Section.1) are plotted as filled gray circles. Points surrounded by black outlines represent the isochrones used stars for which we obtained successful and velocity those we identify as Ret II members is themeasurements, error on the dispersion at the are filled in with red. The four PARSEC The quantity +σ obs (Ri )velocity s half-light radius to determine membership probabilities are displayed as black lines. (b) Spatial distribution of the observed stars. Symbols are as in panel (a). The standard of the radii distribution Ri, and i isisthe of Retradius II from Bechtol et al.+r (015) outlined as a blackdeviation ellipse. (c) Radial velocity distribution of observed stars, combining all three spectroscopic data sets. The 1 highlighted in red is the signature of Ret II. The hatched histogram indicates stars that are not members of Ret II; note clear narrow peak of stars at v ~ 60 km s in the ith bin. This likelihood ani r allows the uncertainties on both σ obs that there are two bins containing non-member stars near v = 70 km s-1 that are over-plotted on top of the red histogram. β (r)v GM(r) 1 d ) +, (1) (ν v = r andνrdr for each bin to be rtaken into account. r For the unbinned analysis, we assume that the distribution of line where ν(r),velocities v r (r), and βani (r) centred 1 v θon the stellar stellar number of-sight stellar is Gaussian, Compare With Stellar Velocities /vrthearemean density, velocity dispersion, velocity v. The likelihood function and readsvelocity (Strigari anisotropy, et al. 008) respectively. + " #,the stellar component, Neglecting the 1/ (< 1 per cent) contribution of N stars * (π) 1 v ) (v i the=enclosed mass at exp radius as, / r can be written (10) Lunbin σ (R )++! i p v r i i=1 σp (Ri )++vi ρdm (s)s ds, () M(r) = 4π where the dispersion 0 of velocities at radius Ri of the ith star comes radius r, the inner slope γ, the outer s slope α. Many studies have used less flex King, Plummer, or exponential profiles) bias the calculated astrophysical factors ( 0. Likelihood functions..1 Binned and unbinned analyses Before fitting Bonnivard+ the actual dsph kinematn 015 binned and an unbinned likelihood funcw

7 Dark Matter Profiles in dsph Galaxies Classical hundreds of stars The Astrophysical Journal, 801:74(18pp),015March10 Geringer-Sameth, Koushiappas, & Walker ellar velocity dispersion profiles observed for the Milky Way s eight classical dwarf spheroidal satellites, adopted from Walker et al.(009c). each projected radius, the median velocity dispersion of models sampled in the Markov-Chain Monte Carlo analysis. Dashed and dotted 68% and 95% of velocity dispersion values from the sampled models. The model profiles are fit to the unbinned kinematic data, but clearly h the binned data plotted here. Table 1 Properties of Milky Way Satellites a and Stellar-kinematic Samples R.A. (J000) Decl. (J000) Distance M V R half N sample r max (hh:mm:ss) (dd:mm:ss) (kpc) (mag) (pc) (pc) 06:41: :57: ± ± ± :0: :54:55 76 ± ± ± :39: :6: ± ± ± Ultra-Faint tens of stars Geringer-Sameth

8 J-Factor of Classical and Ultra-Faints Pace and Strigari 018

9 J-Factor of Classical and Ultra-Faints 1/d^ Pace and Strigari 018

10 J-Factors for Satellites without. Relationship between the distances and spectroscopically determined J-factors of known dsphs is derived with three different techniques: ( ve priors (Geringer-Sameth et al. 015a), (center) Dynamical Bayesian hierarchical modeling Modeling (Martinez 015), and (right) allowing for more flexible parametri distribution and orbital anisotropy profile (Bonnivard et al. 015a). We also include recently derived J-factor estimates for Reticulum II (Simon e d et al. 015b) and Tucana II (Walker et al. 015b) with J-factors for other dsphs that were calculated in a similar manner (see references for each p actor scaling relation (Equation ()) to the data in each panel, yielding log ( J GeVcm ) = { 18.1, 18.3, 18.4}, for the left, center, and rig ely; these relationships are plotted as solid, short dashed, and long dashed red lines. log 10 J J pred 0 =-log 10 D 100 kpc, Figure 3. Drilica-Wagner + 015, Albert + 017

11 J-Factors for Satellites without Dynamical Modeling Drilica-Wagner + 015, Albert Pace and Strigari 018

12 J-Factor Scaling with Dynamics J(0.5 ) GeV cm s los 4 5kms 1 d r1/ kpc 100pc Pace (9 and Strigari 018

13 J-Factor Scaling with Dynamics Residuals Pace and Strigari 018

14 Predicted Number LSST DES Massive in the past (V peak > Pre-reionization fossils (z > 8) Earliest Infall (z peak > 3) dsphs/j-factors in the LSST era 1 km/s) Survey Limiting Magnitude (r) Regular Dwarfs (L > 10 3 L ) Predicted Number Hargis Massive in the past (V peak > Pre-reionization fossils (z > 8) Earliest Infall (z peak > 3) LSST DES 1 km/s) Hyperfaint Dwarfs (L < 10 3 L ) Survey Limiting Magnitude (r) Figure 3. Predicted number of ultra-faint dwarfs for each of the three toy models as a function of survey r-band limiting magnitude for LSST and DES. The results for the brighter and fainter subsets of the ultra-faints are shown in the top and bottom panels, respectively. The error bars show the 10%/90% confidence intervals as described in Section 4. (A color version of this figure is available in the online journal.) L> Mass Pre-r Earli L< Mass Pre-r Earli L> Mass Pre-r Earli L< Mass Pre-r Earli stars tude and W dwa ( 5 som the aver each the 1 moc resu spat limi T num of r perc our

15 J-Factors for Satellites without Dynamical Modeling Can we do better than just distance? Replacing velocity dispersion with luminosity J(0.5 ) GeV cm 5 LV L 0.3 d r1/ kpc 100pc (1) Pace and Strigari 018

16 J-Factors for Satellites without Dynamical Modeling Can we do better than just distance? Pace and Strigari 018

17 dsphs/j-factors in the LSST era Giant Magellan Telscope/GMACS

18 Conclusions dsphs are excellent targets for the indirect detection of dark matter. Astrophysical J-Factors are required to compute the dark matter flux. There is a simple relationship to estimate the J-Factor with dynamical modeling: J(0.5 ) GeV cm s los 4 5kms 1 d r1/ kpc 100pc We will discover many satellites in the LSST era and their follow-up requires the next generation of telescopes. (9

19 Dark Matter Gamma Ray Sky: Limits Figure 10. Upper limits on the DM annihilation cross section ( bb channel) derived from the sub-sam nominal sample (right). Green curves show the limits obtained when these samples Albert are analyzed usin available) and fixed J-factor uncertainties of 0.4, 0.6, and 0.8 dex. The solid black line shows the obser