Joint mass and anisotropy modeling of the. Fornax Dwarf Spheroidal

Transcription

1 Joint mass and anisotropy modeling of the Fornax Dwarf Spheroidal Fornax Dwarf Spheroidal with Chris GORDON (Oxford) Andrea BIVIANO (Trieste) DSS 1

2 Motivations Large velocity data sets in Dwarf Spheroidals Walker et al member velocities in Fornax! constraints on Dark Matter normalization constraints on Dark Matter inner slope constraints on velocity anisotropy of stars 2

3 Mass anisotropy degeneracy Spherical stationary Jeans equation anisotropic dynamical pressure tracer density d ( ) νσr 2 dr +2 β(r) r β(r) = 1 σ2 θ (r) σ 2 r(r) νσ 2 r = ν GM(r) r 2 = velocity anisotropy isotropic orbits: β = 0 radial orbits: β = 1 circular orbits: β Mass / Anisotropy Degeneracy MAD 3

4 Exploratory Mass-Anisotropy Modeling 1) Model-fitting of Line-of-Sight velocity dispersion profile Mtot(r) + β(r) σlos(r) 4

5 1) assume both M(r) & β(r) & fit the LOS velocity dispersions for β = 0 line-of-sight velocity dispersion Tremaine et al. 94; Prugniel & Simien 97 R Σ(R) σ 2 los (R) = 2G R r 2 R 2 r 2 ν(r)m(r)dr tracer surface density kernels for other simple β(r) Mamon & Łokas 05b 5

6 Exploratory Mass-Anisotropy Modeling 1) Model-fitting of Line-of-Sight velocity dispersion profile Mtot(r) + β(r) σlos(r) 2) Anisotropy inversion { Σ(R) + Mtot(r) β(r) σlos(r) Binney & Mamon 82 Tonry 83; Bicknell et al. 89 Solanes & Salvador-Solé 90 Dejonghe & Merritt 92 6

7 Exploratory Mass-Anisotropy Modeling 1) Model-fitting of Line-of-Sight velocity dispersion profile Mtot(r) + β(r) σlos(r) 2) Anisotropy inversion { Σ(R) + Mtot(r) β(r) σlos(r) 3) Mass inversion { Σ(R) σlos(r) + β(r) Mtot(r) Mamon & Boué 10 Wolf et al. 10 7

8 3) Mass inversion Mamon & Boué 10; Wolf et al. 10 Kinematic deprojection & mass inversion of spherical systems with known anisotropy anisotropic kinematic projection p = ρ σ r2 = dynamical pressure P(R) = 2 R 1 β R2 r 2 p r dr r 2 R 2 P = Σ σ los2 = projected pressure deprojection MB10: simple β(r) (1 β) p = = r r K[β(s)] s dp dr L β (R, r) dp dr dr R dr R2 r 2 insert dynamical pressure into Jeans equation mass profile simple β(r): single integral! vc 2 1 { R d 2 P (r) = π [1 β(r)] ρ(r) R2 r 2 dr 2 C β(r, r) dp } dr r 8 dr

9 Radius where mass is independent of anisotropy Wolf et al. 10 Sérsic (or similar): at r -3 r1/2 Mamon & Boué 10 NFW: at 7 rs or 3 rs 9

10 Exploratory Mass-Anisotropy Modeling 1) Model-fitting of Line-of-Sight velocity dispersion profile Mtot(r) + β(r) σlos(r) 2) Anisotropy inversion { Σ(R) + Mtot(r) β(r) σlos(r) 3) Mass inversion { Σ(R) σlos(r) + β(r) Mtot(r) 4) Fitting stars in Projected Phase Space MAMPOSSt Mamon, Biviano & Boué 11, to be subm ν(r) + M tot (r) + β(r) + {v}(r) g(r,vlos) 10

11 Other popular modeling methods gaussian {vlos(r)} lose info on β ν(r) + M tot (r) + β[r] + {vlos}(r) g(r,vlos) dispersion-kurtosis Łokas 02; see Łokas, Mamon & Prada 05 Mtot(r) + β σlos(r) + κlos(r) must assume β = cst distribution function modeling Dejonghe & Merritt 92; Merritt & Saha 93 M tot (r) + f(e,j) or {fi(e,j)} g(r,vlos) don t know f(e,j); not sure that {fi(e,j)} = basis set orbit modeling Schwarzschild 79; de Lorenzi et al. 09 M tot (r) + {orbits} g(r,vlos) too slow for MCMC investigation of parameter space 11

12 Mass modeling of the Fornax dwarf spheroidal 12

13 LV = 1.9 x 10 7 Lsun Irwin & Hatzimiditriou 95 LV = 0.9 x 10 7 Lsun Walcher et al. 03 Fornax data Fornax metallicity 2633 velocities 2278 Fornax members Milky Way metallicity 2 Re m = 0.7 Sersic distribution Walcher et al. 03; Battaglia et al. 06 ellipticity: Battaglia et al. 06 main starburst: age = 5.4 Gyr Saviane et al. 00 Mstars/LV = 4.8 Walcher et al. 03 (uncertain) center: Battaglia et al. 06 data & MW flags: Walker et al. 09 Magellan/MMFS rejected outlier 13

14 Fornax: velocity dispersion profile out to 2 Reff Łokas 09 isotropic, M/LV = 7.3 order 2 fit M/LV = 8.8 isotropic, M/LV = 4.8 isotropic, M/LV = 4.8, low LV order 3 fit Reff order 4 fit 1/3 fraction of dark matter in inner regions? OR LV underestimated by 40%? M/L increases outwards? OR tangential outer orbits? 14

15 Fornax: anisotropy inversion INPUT stars radial OUTPUT isotropic moderately tangential cst M/L radial (low M/L) or tangential (high M/L) orbits 15

16 Fornax: isotropic mass inversion stars dark matter, even in inner regions? luminosity and/or Mstars/L too low? fit order 4 σ los (R) falls too fast! stars stars radial outer orbits! 16

17 low LV log total mass (1.5kpc)( M ) std LV max stars 1σ 2σ 3σ 4σ low LV (P=0.95 MAMPOSSt with MCMC well-defined Mtot(2Re) gaussian 3D velocities cst β cst Mstars/L Kazantzidis DM: ρ ~ exp(-r/a) / r MCMC: 9 chains of log DM scale radius (kpc) a > 2 kpc (95% conf.) anisotropy: -0.5 log(1-β) Isotropic orbits! β = -0.03±

18 log total mass (1.5kpc)( M ) std LV low LV max stars 1σ 2σ 3σ 4σ std LV Free inner dark matter slope... gaussian 3D velocities cst β cst Mstars/L gen l Kazantzidis DM: ρ ~ r γ exp(-r/a) MCMC: 11 chains of log DM scale radius (kpc) anisotropy: -0.5 log(1-β) inner DM slope inner DM slope: not too steep! 18

19 log total mass (1.5kpc)( M ) 1σ 2σ 3σ 4σ Increasing anisotropy... gaussian 3D velocities β=r 2 /(r 2 +a 2 ) Osipkov-Merritt cst Mstars/L gen l Kazantzidis DM: ρ ~ r γ exp(-r/a) MCMC: 11 chains of log DM scale radius (kpc) anisotropy: -0.5 log(1-β) inner DM slope anisotropy radius: 1.6 < ranis < 25 kpc (1σ) 19

20 Radial variations Wolf+10 mass pinch stars P=95% DM total P=98% appreciable velocity anisotropy at 2 Re but β=0 is consistent w data 20

21 Comparison with previous work Łokas 09 this work Walker+09 Walker et al. 09 Łokas 09 anisotropy poorly constrained: < β < β < σ limits vs < β < 0.13 inner slope not constrained γ < -0.3 vs. γ >

22 Conclusions Mass modelling = exploratory data analysis: try methods: mass & β inversions, & MAMPOSSt Fornax MAMPOSSt stronger constraints Isotropic inner orbits, radial outer orbits Dark matter present: likely > 50% at all radii seen in simulations: Klimentowski+07; Lokas+10 Dark matter inner slope shallower than -1.1 see Pennarubia for tidal effects on GCs 22

23 Perspectives MAMPOSSt with non-gaussian 3D velocity distributions model rotation & metallicity effects Battaglia+06 in Fornax higher-parameter MCMC fits 23