Segregation & Tidal Disruption of Dark Matter Substructure: Fact or Fiction? Frank van den Bosch Yale University
Radial Segregation of Satellite Galaxies vdBosch+08 Satellite galaxies are segregated with respect to stellar mass, luminosity, color, SFR, etc. This is rarely accounted for in HOD/CLF modeling. How are subhalos segregated? Watson+15 Yale University Frank van den Bosch for
Subhalo Segregation present-day mass mass at accretion accretion redshift retained mass fraction vdBosch+16 Yale University Frank van den Bosch
Subhalo Segregation Bolshoi Chin250 Chin400 vdBosch+16 Virtually all subhalo properties show some level of radial segregation Segregation of M acc is partially imprinted at accretion, is magnified due to dynamical friction during first radial orbit, and is subsequently suppressed due to tidal disruption... Yale University Frank van den Bosch
Subhalo Disruption in Bolshoi Jiang & vdB, 2016 Tidal Stripping Pericentric Passage Disruption Fractional Disruption Rate ≈ 13 percent per Gyr Tidal Heating Subhalo-Subhalo Encounter Mechanisms Only ~35 percent of subhaloes accreted at z=1 survive to z=0 Numerical overmerging Frank van den Bosch Yale University
Does Stripping cause Disruption? As first pointed out by Hayashi+03, instantaneous stripping of outer layers of NFW halo can leave a remnant with positive binding energy. For an isotropic NFW halo, the core has positive binding energy if r cut < r bind = 0.77 r s. (corresponding core mass is ~0.08 M vir ) Spontaneous disintegration once r tid < r cut ? This assumption is made in several models or subhalo evolution (e.g., Zentner & Bullock 2003; Taylor & Babul 2004; Klypin et al. 2015) Frank van den Bosch Yale University
N=10 5 r cut =0.67r s tree-code E tot >0 r t
NO! Does Stripping cause Disruption? As first pointed out by Hayashi+03, instantaneous stripping of outer layers of NFW halo can leave a remnant with positive binding energy. For an isotropic NFW halo, the core has positive binding energy if r cut < r bind = 0.77 r s. (corresponding core mass is ~0.08 M vir ) Spontaneous disintegration once r tid < r cut ? E(r t )<0 E(r t )>0 However: particles have broad distribution of binding energies, and majority of particles remain bound. Simulations confirm that remnant rapidly re-virializes to a bound system with somewhat smaller, but non-zero mass. vdBosch+18, Frank van den Bosch Yale University
Numerical Simulations subhalo r t Simulate NFW halo orbiting on circular orbit inside static potential of host halo. r orb No impulsive (tidal) heating No dynamical friction host Naive Prediction: all matter outside of tidal radius will be stripped of over time... More `Sophisticated’ Prediction: all matter with an apocenter r apo > r t will be stripped of over time... Frank van den Bosch Yale University
N=10 5 r orb =0.1 r vir,h tree-code r t =0.11 r s r t
Numerical Simulations subhalo r t Simulate NFW halo orbiting on circular orbit inside static potential of host halo. r orb N=10 5 c h =5 c s =10 M h =10 3 m s m(r t )/m s m(r apo <r t )/m s host r orb shrinks r t shrinks dyn. friction vdBosch+18 mass loss Analytical predictions fail to predict amount of mass stripped r t shrinks virialization Mass loss continues for >50 Gyr modified ρ (r) Frank van den Bosch Yale University
Tidal Stripping on Circular Orbits N=10 5 c h =5 c s =10 M h =10 3 m s 0.2 0.1 0.15 0.05 Disruption for r orb < 0.15 r vir ......or numerical artefacts? r orb = 0.15 r vir vdBosch & Ogiya, 2018 Frank van den Bosch Yale University
Tuning the Softening Length ε opt ≃ 0.05 NFW halo N=10 5 vdBosch & Ogiya, 2018 ε too large ➢ force bias ➢ central cusp unresolved ε too small ➢ force noise ➢ artificial large-angle deflections ➢ isothermal core Frank van den Bosch Yale University
Force Softening ε =0.01 ε =0.03 ε =0.05 ε =0.07 ε =0.09 ε =0.11 vdBosch & Ogiya, 2018 ε opt ∝ r half N -1/3 (Dehnen+01; Power+03) Mass evolution and disruption extremely sensitive to softening length As subhalo looses mass, its optimal softening length decreases Frank van den Bosch Yale University
Towards Numerical Convergence r orb =0.1 c h =5 c s =10 M h =10 3 m s N=1,000,000 N=300,000 N=100,000 N=30,000 vdBosch & Ogiya, 2018 Frank van den Bosch Yale University
Numerical Criteria to Judge Reliability Runaway Disruption Instability vdBosch & Ogiya, 2018 Disruption if characteristic acceleration drops below central acceleration: a char /a 0 < 1 . 2 a char = G M ( r h ) G M ( r ) a 0 = lim r 2 ε r h r ↓ 0 (Power+03) Discreteness driven runaway instability kicks in when |dN/dt| > 100/ τ dyn 0.2 For average subhalo mass loss rate this implies N < 80 N acc Frank van den Bosch Yale University
Conclusions Abundance & demographics of dark matter substructure important for variety of astrophysical applications. Subhalo segregation mainly consequence of hierarchical formation. Impact of dynamical friction is modest Subhalo disruption is prevalent in numerical simulations What causes subhalo disruption? Dynamical friction (physical) Inadequate force resolution (numerical) Discreteness noise (numerical) Current generation of cosmological simulations still su ff ers from severe overmerging. serious road-block for small-scale cosmology program serious road-block for understanding galaxy formation Yale University Frank van den Bosch
Related Papers On the Segregation of Dark Matter Substructure van den Bosch F ., Jiang F ., Campbell D., Behroozi P ., 2016, MNRAS, 455, 158 Dissecting the evolution of dark matter subhaloes in the Bolshoi simulation van den Bosch F ., 2017, MNRAS, 468, 885 Disruption of Dark Matter Substructure: Fact of Fiction? van den Bosch F ., Ogiya G., Hahn O., Burkert A., 2018, MNRAS, 474, 3043 Dark Matter Substructure in Numerical Simulations: A Tale of Discreteness Noise, Runaway Instabilities and Artificial Disruption van den Bosch F ., Ogiya G., 2018, MNRAS, 475, 4066 Yale University Frank van den Bosch
Recommend
More recommend
