1. Anisotropic AGN Outflows Filling The Cosmological Volume
Paramita Barai
Collaborators: Joël Germain, Hugo Martel
Université Laval
Québec City, Canada
5th June, 2008
The Central Kpc (Ierapetra, Crete)
MAKE ~20 SLIDES --- SINCE ONLY 15 MINS FOR TALK …

2. Introduction
Outflows observed in a large fraction of AGNs
Goal: Calculate the volume fraction of the Universe filled by AGN outflows over the Hubble time
Energy density
Magnetic field
Outflows observed in a large fraction of AGNs, both at low and high z’s, which are argued to have impact on several galaxy and star formation and large scale structures.
In order to quantify the global impact which these AGN outflows might have we ask the ques …
To assess their importance in the cosmic evolution of the Universe, the key question we ask is how much volume of the Universe do these outflows fill.
31-Dec-18
P. Barai, U. Laval

3. Motivation Spherical Outflow Previous studies
Furlanetto & Loeb 2001, ApJ, 556, 619
Scannapieco & Oh 2004, ApJ, 608, 62
Levine & Gnedin 2005, ApJ, 632, 727
Spherical
Outflow
Anisotropically expanding outflow
Implement in cosmological simulations
Our Improvement
31-Dec-18

4. Anisotropic Outflow
Blue : density iso-surface, showing the pancake (a circular ripple results from the explosion), Red : temperature iso-surface, showing the outflow.
Shows explosion inside a DG that is forming at the intersection of 2 emerging filaments inside a cosmological pancake. Using Adaptive SPH algo. The outflow is anisotropic, bipolar, and goes perpendicular to the pancake plane.
As told by Hugo:
Snapshots of a hydrodynamic simulation of an explosion inside a cosmological pancake.
Shows intersection of 2 filaments with a cluster forming at the center / junction of filaments.
Whenever density gets higher than a threshold, eject some amount of hot gas outflow (RED).
Einstein De-Sitter model, so z and scale invariant.
Center can represent a DG. Then the box dimension is < 1 Mpc.
Cosmological outflows expand anisotropically in large scales
Away from high-density regions, into low-density regions, along the path of least resistance
(Martel & Shapiro 2001, RevMexAA, 10, 101)

5. Outflow Geometry (Pieri, Martel & Grenon 2007, ApJ, 658, 36)
Bipolar spherical cone
Spherical coordinates (r, , )
Radius = R
Opening angle = 
Direction of Outflow = ê (unit vector)

6. Direction of Least Resistance (DLR)
In large-scale filamentary structures, outflow direction is obtained from pressure of surrounding medium
Implementation
Find DLR around density peaks
Taylor expansion of density around a peak inside sphere of radius R
Rotate Cartesian coordinates to make cross-terms vanish
Largest of the coefficients A, B, C  DLR
Density peak means the density of a cell is greater than 6 neighboring cells.
AGN locations are in density peak.
31-Dec-18
P. Barai, U. Laval

7. Methodology N-body Cosmological Simulation Semi-analytical model of
AGN outflows
Cosmological
AGN population
from observed
luminosity function
Distribute AGN
in the simulation volume
(at density peaks)
How? (Procedures): N-body dynamics of dark matter. Our approach.
Evolve AGN and
their outflows within
Simulation Box
Compute the volume of
the cosmological box
filled by the outflows
31-Dec-18

8. Cosmological Simulation
N-body simulations of a cosmological volume
Box size (comoving)
= 128 h–1 Mpc
Triply periodic boundary conditions
Expanding with the Hubble flow
2563 dark matter particles
5123 grid
Evolve from z=25 up to z=0
P3M code
(particle-particle/particle-mesh)
Grav. softening length = 0.3  cell size
= 75 h–1 kpc
Particle mass
= 1.32 1010 M
CDM model :
In subroutine GetUnits :
SOME SIMULATION VALUES AND COMPUTATIONAL UNITS:
Mean MATTER density at z=0 (g/cm^3) = E-30
Mean BARYON density at z=0 (g/cm^3) = E-31
Total Mass inside Cosmological Box:
Mbox (g) = E Mbox/Msun = E+17
Total number of DM particles =
Mass of each DMP (Msun) = E+10
One side of cubic cosmol. box (Mpc) = E+02
One side of cubic cosmol. box (cm) = E+26
Grid size of P-M mesh (L_box) = E-03
For a virial mass (Msun) = E+12
Virial Radius, R_200 (L_box) = E-03
Virial Radius (cell size of PM mesh) = E-01
Computational Units :
Unit of mass : E+50 g
Unit of length : E+26 cm
Unit of time : E+18 s
Unit of luminosity : E+48 ergs/s
Time unit: (s) = E (Myr) = E+04
H0 : (cgs) = E (comp unit) = E+00
Mean IGM pressure at z=0 (cgs) = E-19
31-Dec-18
P. Barai, U. Laval

9. Ambient Medium for AGN Outflows
Assume: baryonic gas distribution follows dark matter in the simulation box
Ambient gas density :
Pressure :
Temperature (assuming a photoheated medium)
Mean molecular mass :
31-Dec-18
P. Barai, U. Laval

10. Redshift & Luminosity Distribution
Observed AGN bolometric luminosity function
(Hopkins, Richards & Hernquist 2007, ApJ, 654, 731)
Constant AGN lifetime, TAGN = 108 yr
Fraction of AGN hosting outflows = 0.2
Number of AGN
Locate AGN at local density peaks
Filter density above a minimum halo mass
Quoting from Hugo:
This whole issue is confusing. In our simulations, the mass per particle is 1.32e10. So we don't even resolve objects of mass 1.e10. But for locating quasars, we are not using groups of particles to identify halos. Instead, we are using the density on the grid, as in Levine & Gnedin. The mean mass in a cell is smaller by a factor of 8, so it is 1.65e09. But then when a density of 200 times the mean is reached, the mass inside a cell is 3.3e11. If we wanted to actually identify small-mass halos from particles, we would need a much larger resolution. But we are merely looking for places where to locate AGN's.
Gnedin & Levine could locate AGN's at any location, with a probability given by the density (their equation 2). As far as I can tell, there is no relation between the density at a chosen grid point, and the mass of the object they will actually locate there. In our case, we have to choose density peaks, and we filtered in order to eliminate local peaks caused by noise. But again, there should be no relation between the filter mass and the mass of the object.

11. 31-Dec-18
P. Barai, U. Laval

12. All Sources in Box from QLF. NAGN,total = 929805
31-Dec-18
P. Barai, U. Laval

13. New AGN locations in a slice of box at z = 0
New AGN locations in a slice of box at z = 0.5 Black - Particles (PM), blue - Peaks, red - AGNs
Slice at time step z = 0.5. Filtering mass = 109 M.
There are 872 peaks and 18 AGNs in this slice which spans x = [0,1], y = [0,1], z = [0.5, ].
Quoting Joel’s
The x and y ranges are (all the box). The z range is (This is the middle to 91 Mpc / 180 Mpc).
The peaks and the AGNs on the graph is only the "new ones" in this time step, in
this slice.
P. Barai, U. Laval

14. Active-AGN Life Jet Kinetic power Jet advance Energy Pressure
Relativistic outflow with  = 4/3
Overpressured : p >> px
Fraction to convert bolometric to kinetic luminosity, epsK = 0.1.
To Convert bolometric luminosity to kinetic luminosity in c.g.s: bLQ(i)=epsK*bLQ(i)*Lsun

15. Post-AGN Evolution Anisotropic Expansion when overpressured, p > px
Sedov-Taylor adiabatic expansion
Total AGN energy
Adiabatic loss
Bipolar conical outflow
When reach pressure equilibrium, p = px
Passive Hubble evolution, Rcomoving = Constant

16. Volume Filled
Count mesh cells in the simulation box occurring inside the volume of one/more outflow
Total number of filled cells, NAGN = total volume of box occupied by outflows
Express the total volume filled as a fraction of volumes of various overdensities in the box
31-Dec-18
P. Barai, U. Laval

17. Fractional Volumes of box of various overdensities filled by AGN outflows (N with C = 0, 1, 2, 3, 5, 7)
31-Dec-18

18. 31-Dec-18
P. Barai, U. Laval

19. Volume Averaged Energy Density
31-Dec-18
P. Barai, U. Laval

20. Volume Averaged Magnetic Field
31-Dec-18
P. Barai, U. Laval

21. Summary
Implemented a semi-analytical model of anisotropic AGN outflows in N-body simulations
AGN outflows pervade 13 – 24% of the volume of the Universe by the present
In some cases occupy 100% of the overdense regions by z > 0
Volume averaged quantities in the filled regions at z = 0
Energy Density = 5 10–18 erg cm–3
Magnetic field = 10–9 Gauss
These energy densities are consistent with / within limits of the energy density obtained from the cosmic X-ray background background.
The magnetic field is consistent with intergalactic large scale magnetic fields computed by Ryu et al (1998).
31-Dec-18
P. Barai, U. Laval

22. References Furlanetto, S.R. & Loeb, A. 2001, ApJ, 556, 619 (FL01)
Ganguly, R. & Brotherton, M.S. 2008, ApJ, 672, 102
Hopkins, P.F., Richards, G.T. & Hernquist, L. 2007, ApJ, 654, 731
Levine, R. & Gnedin, N.Y. 2005, ApJ, 632, 727
Pieri, M. M., Martel, H. & Grenon, C. 2007, ApJ, 658, 36 (PMG07)
Scannapieco, E. & Oh, S.P. 2004, ApJ, 608, 62
31-Dec-18
P. Barai, U. Laval

23. Backup Slides

24. Outline Motivation Introduction
N-body simulation + semi-analytical approach
Outflow expansion : Early and Late phases
Anisotropic (biconical) late expansion
Results: volume filling fractions
Conclusions
31-Dec-18
P. Barai, U. Laval

25. Anisotropic Outflow
31-Dec-18
P. Barai, U. Laval

26. Active-AGN Life Kinetic luminosity Jet advance Energy
Shock half-opening angle, s = 5
Energy
Spherical outflow
Pressure
Relativistic outflow with  = 4/3
31-Dec-18
P. Barai, U. Laval