1. The Role of Baryons in Strong Lensing properties of Simulated Clusters
Madhura Killedar
INAF-UniTS
arXiv: ongoing
Collaborators
 Stefano Borgani (UniTS-INAF)
 Massimo Meneghetti (INAF, Bologna)
 Dunja Fabjan (SPACE-SI, Ljubljana)
 Luca Tornatore (UniTS-INAF)
 Klaus Dolag (MPA)

2. Giant Arcs CosmoComp Meeting Trieste Italy Sept 2012 Madhura Killedar
11/01/2019

3. Strong Lensing efficiency3
Strong Lensing efficiency
How many giant arcs?
Add Einstein radius Eqn for axisymmetric lens
E = Einstein radius
How far apart?
CosmoComp Meeting Trieste Italy Sept Madhura Killedar
11/01/2019

4. … and a niggling problem with LCDM4
Testing cosmology
… and a niggling problem with LCDM
No cosmological constant preferred over LCDM?
Clusters not selected self- consistently?
Uncertainties in source properties?
Simulations not realistic enough?
Bartelmann et al 1998
CosmoComp Meeting Trieste Italy Sept Madhura Killedar
11/01/2019

5. Baryons in Simulations
Puchwein et al 2005
Mead et al 2010
Baryons in Simulations
The inclusion of baryonic processes have been shown to affect strong lensing predictions
Rozo et al 2008
L/W
CosmoComp Meeting Trieste Italy Sept Madhura Killedar
11/01/2019

6. New Simulation Set
Planelles et al (in prep)
SPH code GADGET-3 (Springel 2005)
29 Lagrangian regions around massive z=0 clusters re-simulated at improved resolution (Tormen et al 2005)
4 flavours of simulations with the same clusters: one of these includes AGN Feedback
Clusters: M > 3x1014 Msun
17 of 42 clusters relaxed at z=0.25
14 of 34 clusters relaxed at z=0.5
CosmoComp Meeting Trieste Italy Sept Madhura Killedar
11/01/2019

7. Glossary Simulation description DM Dark matter only NR
DM+gas: Non-radiative gas
CSF
DM+gas: UV heating, Metallicity-dependent gas cooling, star formation, chemical evolution, stellar feedback
AGN
As above + AGN feedback
(Springel et al 2005; Sijacki et al 2007) DM NR
CSF
AGN
There is a counterpart of each cluster in each simulation
CosmoComp Meeting Trieste Italy Sept Madhura Killedar
11/01/2019

8. Same cluster: before and after baryons8
Same cluster: before and after baryons
CosmoComp Meeting Trieste Italy Sept Madhura Killedar
11/01/2019

9. (don’t) assume a spherical cow9
(don’t) assume a spherical cow
Image caustics (a.k.a. critical curves) aren’t circular
MACS J
Zitrin et al 2011
CosmoComp Meeting Trieste Italy Sept Madhura Killedar
11/01/2019

10. (don’t) assume a spherical cow10
(don’t) assume a spherical cow
We see different caustics when we view the same cluster along different lines of sight
CosmoComp Meeting Trieste Italy Sept Madhura Killedar
11/01/2019

11. (don’t) assume a spherical cow11
(don’t) assume a spherical cow
Find a good histo for ER, try g334
We measure different Einstein radii when we view the same cluster along different lines of sight
We analyse 50 lines of sight through each cluster
CosmoComp Meeting Trieste Italy Sept Madhura Killedar
11/01/2019

12. Einstein radius pmin pmax DM - NR 0.11 1 CSF <10-4 AGN 0.93
KS-test vs DM
pmin
pmax DM - NR
0.11 1
CSF
<10-4
AGN
0.93
Einstein radius
AGN predict similar Einstein radii to DM counterpart
Einstein radii are good proxies for mass at large overdensities
CosmoComp Meeting Trieste Italy Sept Madhura Killedar
11/01/2019

13. The distribution of stellar mass
Stacked density profiles of relaxed clusters
Stellar mass dominates in the core
Reduced stellar (and total) mass
CosmoComp Meeting Trieste Italy Sept Madhura Killedar
11/01/2019

14. Mass profiles and typical Einstein radii
… but compare this to the scale of typical Einstein radii DM
NR CSF
AGN DM
NR CSF
AGN
stacked cumulative profiles reveal central mass affected by baryonic physics …
Agrees with Puchwein+ ‘05, Duffy+ ‘10, Mead+ ‘10, Martizzi+ ‘12
CosmoComp Meeting Trieste Italy Sept Madhura Killedar
11/01/2019

15. What about other sub-samples?
zL=0.25
Higher significance in the increased lensing efficiency of CSF vs DM
Only the lowest SL efficiency events are affected by AGN
zs=1
Smaller Einstein radii (as expected)
AGN clusters have only slightly larger Einstein radii than DM
Unrelaxed clusters
Produce complex caustics
Usually smaller Einstein radii (for fixed virial mass)
Boost in Einstein radii when substructure fortuitously placed
CosmoComp Meeting Trieste Italy Sept Madhura Killedar
11/01/2019

16. Einstein radius (effects of dynamical state)16
Einstein radius (effects of dynamical state)
CosmoComp Meeting Trieste Italy Sept Madhura Killedar
11/01/2019

17. The MACS z>0.5 sample
Ebeling et al (2007)
Strong lensing mass reconstruction is complete (Zitrin et al 2011); see Einstein radii in Meneghetti et al 2011
11 of 12 MACS clusters are at 0.5<z<0.6 (one at z=0.7) Comparison with our z=0.5 clusters is underway; z=0.6 clusters to follow
fX>10-12 erg s-1 cm-2
fX -> LX -> MΔ? (NO! Luminosity is being mis-interpreted)
So use LX,AGN to select clusters (from AGN-sim and counterparts)
CosmoComp Meeting Trieste Italy Sept Madhura Killedar
11/01/2019

18. Previous tests with MACS z>0.5
Zitrin et al 2011
Meneghetti et al 2011
Previous tests with MACS z>0.5
Horesh et al 2011
Previous works found that z>0.5 MACS are stronger lenses than simulated clusters
CosmoComp Meeting Trieste Italy Sept Madhura Killedar
11/01/2019

19. Mock Sample MACS KS-test E,mock E,MACS
CosmoComp Meeting Trieste Italy Sept Madhura Killedar

20. LX,AGN > LX,cut In all mock samples E,mock/E,MACS < 1 20
12% of mock samples p < 0.05
38% of mock samples p < 0.05
6% of mock samples p < 0.05
12% of mock samples p < 0.05
CosmoComp Meeting Trieste Italy Sept Madhura Killedar
11/01/2019

21. LX,AGN > LX,cut and relaxed
No mock samples have p<0.05 when compared to MACS
Almost all mock samples have p>0.5 21
0.81< E,mock/E,MACS<1.12
0.8< E,mock/E,MACS<1.1
0.2 < p < 1.0
0.2 < p < 1.0
0.85< E.mock/E,MACS<1.13
0.8< E.mock/E,MACS<1.1
0.3 < p < 1.0
0.2 < p < 1.0
CosmoComp Meeting Trieste Italy Sept Madhura Killedar
11/01/2019

22. Summary
Baryons play an important role in strong lensing predictions, *BUT* AGN feedback returns lensing predictions to levels similar to DM-only simulations (for z=0.5 clusters!)
The most realistic simulations have more mass within 3% Rvir but Einstein radii are typically larger (projection dependent)
Triaxiality and mergers/substructures complicate the significance of any test
MACS z>0.5 (and CLASH) provide ideal candidates for comparison but we must take care to select clusters appropriately (by luminosity, not mass). Preliminary results for z=0.5 suggest that mock samples do not differ significantly from MACS - let’s see how z=0.6 mock samples do!
CosmoComp Meeting Trieste Italy Sept Madhura Killedar
11/01/2019