Probing dark matter (sub)structure with strong gravitational - PowerPoint PPT Presentation

Probing dark matter (sub)structure with strong gravitational lensing Simon Birrer, UCLA Collaborators: Tommaso Treu, Daniel Gilman, Anowar Shajib (UCLA) Adam Amara, Alexandre Refregier (ETHZ) Chuck Keeton, Anna Nierenberg IFT, Madrid, 29.6.2018

Koopmans 2005 Birrer+ 2017 Moustakas & Metcalf 2003 Nierenberg+ 2014, 2017 Bradac+ 2002 Hezaveh+ 2016 Dalal & Kochanek 2002 Vegetti+ 2010, 2012, 2018 Metcalf & Madau 2001 can be dark! unknown unknown data Source Lens Image Strong gravitational lensing

Resolved an un-resolved lensing effects: a simplified example un-resolved Observable degeneracies: clump mass - clump profile - clump position - source size - resolved credit: Daniel Gilman, UCLA

Method 1: Quasar flux ratio anomalies exclusion regions for a unresolved strong lensing from certain type of sub-clump quasar narrow line emission region ii) F140W G141 [OIII] D D A A G C C B B Image credit: Nierenberg+2017 small physical source size allows for sensitivity to very low masses Image credit: Nierenberg+2017 Dalal & Kochanek 2002 Moustakas & Metcalf 2003 Nierenberg+2014, 2017 Hsueh+2017, Gilman, Birrer+2018

Method 2: gravitational imaging resolved strong lensing from direct detection through galaxy surface brightness 2 × 10 8 M � lens modelling of sensitive to individual clumps Image credit: Vegetti+2012 near the Einstein ring sensitivity depends on spatial resolution and source structure Koopmans 2005 , Vegetti+2010, 2012 … Hezaveh+ 2016, Birrer+2017

Method 2: gravitational imaging example with perfect lens model Lensing: Birrer+ 2015, 2016 software available: $pip install lenstronomy Shapelets: Refregier 2003 https://github.com/sibirrer/lenstronomy Software: Birrer&Amara 2018

Method 2: gravitational imaging example with missing substructure Lensing: Birrer+ 2015, 2016 software available: $pip install lenstronomy Shapelets: Refregier 2003 https://github.com/sibirrer/lenstronomy Software: Birrer&Amara 2018

Method 2: linear source reconstruction computational cost of linear Requirement: Simultaneous inversion and number of non- reconstruction of source and linear parameters as limitations lens on all relevant scales High resolution reconstruction Lensing: Birrer+ 2015, 2016 of source with Shapelet basis set Shapelets: Refregier 2003 Software: Birrer&Amara 2018

software package publicly available installation: $pip install lenstronomy https://github.com/sibirrer/lenstronomy

sub-structure statistics: cold vs warm (sub-halos only) CDM Credit: Daniel Gilman (UCLA) software: lenstronomy

sub-structure statistics: cold vs warm (sub-halos only) CDM WDM Credit: Daniel Gilman (UCLA) software: lenstronomy

sub-structure statistics: main halo vs LOS (born approximation) main halo Credit: Daniel Gilman (UCLA) See e.g. Despali+2017 software: lenstronomy

sub-structure statistics: main halo vs LOS (born approximation) main halo main halo + LOS Credit: Daniel Gilman (UCLA) See e.g. Despali+2017 software: lenstronomy

sub-structure statistics: born approximation vs non-linear multi-plane born approximation Credit: Daniel Gilman (UCLA) software: lenstronomy

sub-structure statistics: born approximation vs non-linear multi-plane multi-plane born approximation (with main deflector) Credit: Daniel Gilman (UCLA) software: lenstronomy

sub-structure statistics: warm vs. cold in full LOS CDM Credit: Daniel Gilman (UCLA) software: lenstronomy

sub-structure statistics: warm vs. cold in full LOS CDM WDM Credit: Daniel Gilman (UCLA) software: lenstronomy

substructure quantification Data may show signatures of multiple substructure - Inherent degeneracies are present in the observables - propagating the complex observables into quantitative - statements about dark matter is difficult end-to-end forward modelling

Forward modelling of gravitational imaging Turn a physical model stochastically into simulated data look for the same features in your simulated data Accept/reject simulations Approximate Bayesian based on Computing (ABC) summary statistics Birrer+ 2017 no line-of-sight included

Dark Matter thermal relic mass constraints from lensing substructure Birrer+ 2017 no line-of-sight included

Dark Matter thermal relic mass constraints from lensing substructure excluded ( ) ≥ 2 σ Birrer+ 2017 no line-of-sight included

Dark Matter thermal relic mass constraints from lensing substructure Viel et al. 2014 (Lyman-alpha forest) Polisensky & Ricotti 2011 (MW satellites) excluded ( ) ≥ 2 σ Birrer+ 2017 no line-of-sight included

Forward modelling of quasar flux ratios Simpler observables - Forecast lots of degeneracies small physical source size allows for sensitivity to low masses Statistical statement of an ensemble of lenses Forward modelling allows for a hierarchical bayesian analysis with correlated priors Image credit: Gilman, Birrer+2018 no line-of-sight included Gilman, Birrer+ submitted (ABC application to flux ratios) Gilman, Birrer+ in prep (LOS contribution)

The way forward 1: combining flux ratios and imaging combining data sets and methods that probe different scales within the same framework… … if possible on the same lens discovered: Ostrovski+, Lemon+, Agnello+, Schechter+, … HST follow-up, PI: Treu modelling: Shajib, Birrer+ in prep software: lenstronomy

see SHARP for Keck AO The way forward 2: gravitational imaging with extreme AO (in the ELT era) or interferometry FWHM 0.02”

The challenges • forward modelling relies on realistic simulations: any limitation that may not be identical to the data may impact your statistic • luminous (sub) structure: globular clusters, stellar discs, .. e.g, Hsueh+ 2016, Gilman+ 2017, … • precise predictions of (sub- and LOS) halo properties: dynamical friction, tidal stripping, resolution limit, computational cost, baryonic physics, … e.g. Bullock & Boylan-Kolchin 2017, van den Bosch+2018, …

Summary • Strong lensing is an unique probe to test different dark matter scenarios in the cosmological context • Dark substructure has been directly detected down to 10^8-9 M_sol, statistical signal may be present down to 10^6-7 M_sol in quasar flux ratios (mass definition dependent) • Statistical constraints based on one single lens excludes a thermal relic mass < 2 keV to 2 sigma confidence level • Combined flux ratios and imaging applied may provide constraints on the mass function over a wide range in mass scale • lens modelling software package “lenstronomy” is publicly available

Thank you! Collaborators: Tommaso Treu, Daniel Gilman, Anowar Shajib (UCLA) Adam Amara, Alexandre Refregier (ETHZ) Chuck Keeton, Anna Nierenberg, … Simon Birrer IFT Madrid, 29.6.2018