The Physics and Cosmology of TeV Blazars

1 in collaboration with Avery E. Broderick 2, Phil Chang 3, Ewald Puchwein 1, Volker Springel 1 1 Heidelberg Institute for Theoretical Studies, Germany 2 Perimeter Institute/University of Waterloo, Canada 3 University of Wisconsin-Milwaukee, USA Aug 30, 2012 / ICM theory/computation workshop, Michigan

Outline Physics of blazar heating 1 Physics of blazar heating Propagation of TeV photons Plasma instabilities in beams Implications 2 Blazar luminosity density Thermal history of the IGM Properties of blazar heating 3

The TeV gamma-ray sky Propagation of TeV photons Plasma instabilities in beams Implications There are several classes of TeV sources: Galactic - pulsars, BH binaries, supernova remnants Extragalactic - mostly blazars, two starburst galaxies

Propagation of TeV photons Propagation of TeV photons Plasma instabilities in beams Implications 1 TeV photons can pair produce with 1 ev EBL photons: γ TeV + γ ev e + + e mean free path for this depends on the density of 1 ev photons: λ γγ (35... 700) Mpc for z = 1... 0 pairs produced with energy of 0.5 TeV (γ = 10 6 ) these pairs inverse Compton scatter off the CMB photons: mean free path is λ IC λ γγ /1000 producing gamma-rays of 1 GeV E γ 2 E CMB 1 GeV each TeV point source should also be a GeV point source

Propagation of TeV photons Plasma instabilities in beams Implications What about the cascade emission? Every TeV source should be associated with a 1-100 GeV gamma-ray halo not seen! limits on extragalactic magnetic fields? TeV spectra expected cascade emission TeV detections Fermi constraints Neronov & Vovk (2010)

Missing plasma physics? Propagation of TeV photons Plasma instabilities in beams Implications How do beams of e + /e propagate through the IGM? plasma processes are important interpenetrating beams of charged particles are unstable

Propagation of TeV photons Plasma instabilities in beams Implications Two-stream instability: mechanism wave-like perturbation with k v beam, longitudinal charge oscillations in background plasma (Langmuir wave): initially homogeneous beam-e : attractive (repulsive) force by potential maxima (minima) e attain lowest velocity in potential minima bunching up e + attain lowest velocity in potential maxima bunching up Φ p p e e

Propagation of TeV photons Plasma instabilities in beams Implications Two-stream instability: mechanism wave-like perturbation with k v beam, longitudinal charge oscillations in background plasma (Langmuir wave): beam-e + /e couple in phase with the background perturbation: enhances background potential stronger forces on beam-e + /e positive feedback exponential wave-growth instability Φ e + e + e p e p e e

Propagation of TeV photons Plasma instabilities in beams Implications Two-stream instability: momentum transfer particles with v v phase : pair momentum plasma waves: growing modes/instability particles with v v phase : plasma wave momentum pairs: damping (Landau damping)

Oblique instability Physics of blazar heating Propagation of TeV photons Plasma instabilities in beams Implications k oblique to v beam : real word perturbations don t choose easy alignment = all orientations greater growth rate than two-stream: ultra-relativistic particles are easier to deflect than to change their parallel velocities (Nakar, Bret & Milosavljevic 2011) Bret (2009), Bret+ (2010)

Beam physics growth rates Propagation of TeV photons Plasma instabilities in beams Implications excluded for collective plasma phenomena consider a light beam penetrating into relatively dense plasma maximum growth rate oblique 0.4 γ n beam n IGM ω p Broderick, Chang, C.P. (2012) IC oblique instability beats IC by factor 10-100 assume that instability grows at linear rate up to saturation

Propagation of TeV photons Plasma instabilities in beams Implications TeV emission from blazars a new paradigm γ TeV + γ ev e + + e IC off CMB γ GeV plasma instabilities heating IGM absence of γ GeV s has significant implications for... intergalactic B-field estimates γ-ray emission from blazars: spectra, background additional IGM heating has significant implications for... thermal history of the IGM: Lyman-α forest late time structure formation: dwarfs, galaxy clusters

Blazar luminosity density Thermal history of the IGM Properties of blazar heating TeV blazar luminosity density: today collect luminosity of all 23 TeV blazars with good spectral measurements account for the selection effects (sky coverage, duty cycle, galactic occultation, TeV flux limit) TeV blazar luminosity density is a scaled version (η B 0.2%) of that of quasars! Broderick, Chang, C.P. (2012)

Unified TeV blazar-quasar model Blazar luminosity density Thermal history of the IGM Properties of blazar heating Quasars and TeV blazars are: regulated by the same mechanism contemporaneous elements of a single AGN population: TeV-blazar activity does not lag quasar activity assume that they trace each other for all redshifts! Broderick, Chang, C.P. (2012)

Evolution of the heating rates HI,HeI /HeII reionization Blazar luminosity density Thermal history of the IGM Properties of blazar heating Heating Rates [ev Gyr 1 ] 10 3 10 2 10 1 0.1 2 10 photoheating photoheating standard blazar standard blazar fit optimistic blazar optimistic blazar fit blazar heating 10x larger heating 10 5 2 1 1 +z Chang, Broderick, C.P. (2012)

Blazar heating vs. photoheating Blazar luminosity density Thermal history of the IGM Properties of blazar heating total power from AGN/stars vastly exceeds the TeV power of blazars T IGM 10 4 K (1 ev) at mean density (z 2) ε th = kt m pc 2 10 9 radiative energy ratio emitted by BHs in the Universe (Fukugita & Peebles 2004) ε rad = η Ω bh 0.1 10 4 10 5 fraction of the energy energetic enough to ionize H I is 0.1: ε UV 0.1ε rad 10 6 kt kev photoheating efficiency η ph 10 3 kt η ph ε UV m pc 2 ev (limited by the abundance of H I/He II due to the small recombination rate) blazar heating efficiency η bh 10 3 kt η bh ε rad m pc 2 10 ev (limited by the total power of TeV sources)

Thermal history of the IGM Blazar luminosity density Thermal history of the IGM Properties of blazar heating 10 5 only photoheating standard BLF optimistic BLF T [K] 10 4 10 3 20 10 5 2 1 1 +z Chang, Broderick, C.P. (2012)

Blazar luminosity density Thermal history of the IGM Properties of blazar heating Evolution of the temperature-density relation no blazar heating with blazar heating Chang, Broderick, C.P. (2012) blazars and extragalactic background light are uniform: blazar heating rate independent of density makes low density regions hot causes inverted temperature-density relation, T 1/δ

Blazars cause hot voids Blazar luminosity density Thermal history of the IGM Properties of blazar heating no blazar heating with blazar heating Chang, Broderick, C.P. (2012) blazars completely change the thermal history of the diffuse IGM and late-time structure formation

Entropy evolution Physics of blazar heating temperature evolution entropy evolution 10 5 only photoheating standard BLF 10 2 only photoheating standard BLF optimistic BLF optimistic BLF T [K] 10 4 K e [kev cm 2 ] 10 1 10 3 20 10 5 2 1 1 +z 0.1 20 10 5 2 1 1 +z C.P., Chang, Broderick (2012) evolution of entropy, K e = ktn 2/3 e, governs structure formation blazar heating: late-time, evolving, modest entropy floor

When do clusters form? 10 15 mass accretion history 5 mass accretion rates M 200c [ M O ] 10 14 10 13 z 0.5 = 0.42 z 0.5 = 0.55 z 0.5 = 0.70 z 0.5 = 0.86 z 0.5 = 0.95 S = d log M 200c / d log a (z) 4 3 2 1 10 12 1 + z 1 0 1 + z 1 C.P., Chang, Broderick (2012) most cluster gas accretes after z = 1, when blazar heating can have a large effect (for late forming objects)!

Entropy floor in clusters Cluster entropy profiles ICM entropy at 0.1R 200 : Optically selected X ray selected Cavagnolo+ (2009) Hicks+ (2008) Do optical and X-ray/Sunyaev-Zel dovich cluster observations probe the same population? (Hicks+ 2008, Planck Collaboration 2011)

Entropy floor in clusters Cluster entropy profiles Planck stacking of optical clusters Cavagnolo+ (2009) Planck Collaboration (2011) Do optical and X-ray/Sunyaev-Zel dovich cluster observations probe the same population? (Hicks+ 2008, Planck Collaboration 2011)

Entropy profiles: effect of blazar heating varying formation time varying cluster mass 1000 1000 z = 0 z = 0.5 z = 1 z = 2 M 200 = 3 x 10 13 M O M 200 = 1 x 10 14 M O, z = 0.5 M 200 = 3 x 10 13 M O, z = 0.5 M 200 = 1 x 10 13 M O, z = 0.5 K e [kev cm 2 ] 100 K e [kev cm 2 ] 100 10 optimistic blazar 10 optimistic blazar 0.01 0.10 1.00 r / R 200 0.01 0.10 1.00 r / R 200 C.P., Chang, Broderick (2012) assume big fraction of intra-cluster medium collapses from IGM: redshift-dependent entropy excess in cores greatest effect for late forming groups/small clusters

Gravitational reprocessing of entropy floors Borgani+ (2005) 2 K 0= 25 kev cm preheated no preheating + floor no preheating greater initial entropy K 0 more shock heating greater increase in K 0 over entropy floor net K 0 amplification of 3-5 expect: median K e,0 150 kev cm 2 max. K e,0 600 kev cm 2 Borgani+ (2005)

Cool-core versus non-cool core clusters cool hot Cavagnolo+ (2009)

Cool-core versus non-cool core clusters cool hot early forming, t > t merger cool late forming, t merger < t cool Cavagnolo+ (2009) time-dependent preheating + gravitational reprocessing CC-NCC bifurcation (two attractor solutions) need hydrodynamic simulations to confirm this scenario

How efficient is heating by? 10 4 C.P., Chang, Broderick (2011) E b, 2500 (kt X = 5.9 kev) E b, 2500 (kt X = 3.5 kev) E b, 2500 (kt X = 2.0 kev) Ecav= 4PVtot [10 58 erg] 10 2 10 0 E b, 2500 (kt X = 1.2 kev) E b, 2500 (kt X = 0.7 kev) 10-2 cool cores non-cool cores 1 10 100 K e,0 [kev cm 2 ]

How efficient is heating by? 10 4 C.P., Chang, Broderick (2011) E b, 2500 (kt X = 5.9 kev) E b, 2500 (kt X = 3.5 kev) E b, 2500 (kt X = 2.0 kev) Ecav= 4PVtot [10 58 erg] 10 2 10 0 E b, 2500 (kt X = 1.2 kev) E b, 2500 (kt X = 0.7 kev) max K 0 10-2 cool cores non-cool cores 1 10 100 K e,0 [kev cm 2 ] AGNs cannot transform CC to NCC clusters (on a buoyancy timescale)

Challenges to the Challenge Challenge #1 (unknown unknowns): inhomogeneous universe universe is inhomogeneous and hence density of electrons change as function of position could lead to loss of resonance over length scale spatial growth length scale (Miniati & Elyiv 2012) we have reasons to believe that the instability is not appreciably slowed in the presence of even dramatic inhomogeneities Challenge #2 (known unknowns): non-linear saturation we assume that the non-linear damping rate = linear growth rate effect of wave-particle and wave-wave interactions need to be resolved Miniati and Elyiv (2012) claim that the nonlinear damping rate is linear growth rate

Conclusions on blazar heating explains puzzles in high-energy astrophysics: lack of GeV bumps in blazar spectra without IGM B-fields unified TeV blazar-quasar model explains Fermi source counts and extragalactic gamma-ray background novel mechanism; dramatically alters thermal history of the IGM: uniform and z-dependent preheating rate independent of density inverted T ρ relation quantitative self-consistent picture of high-z Lyman-α forest significantly modifies late-time structure formation: group/cluster bimodality of core entropy values suppresses late dwarf formation (in accordance with SFHs): missing satellites, void phenomenon, H I-mass function

Simulations with blazar heating Puchwein, C.P., Springel, Broderick, Chang (2012): L = 15h 1 Mpc boxes with 2 384 3 particles one reference run without blazar heating three with blazar heating at different levels of efficiency (address uncertainty) used an up-to-date model of the UV background (Faucher-Giguère+ 2009)

-2-2 Physics of blazar heating Temperature-density relation 7.0 no blazar heating intermediate blazar heating 6.5 6.0 log 10 ( N HI N H ) = 8-7 -6-8 -7-6 log 10 (T /K) 5.5 5.0 4.5-5 -4-5 -4 4.0-3 -3 3.5 Viel et al. 2009, F=0.1-0.8 log 10 (Mpix/(h 1 M )) Viel et al. 2009, F=0-0.9 5 6 7 8 9 10 3.0 2 1 0 1 2 3 2 1 0 1 2 3 log 10 (ρ/ ρ ) Puchwein, C.P., Springel, Broderick, Chang (2012)

The Lyman-α forest Physics of blazar heating

The observed Lyman-α forest

The simulated Ly-α forest 1.0 0.8 transmitted flux fraction e τ 0.6 0.4 e τ 0.2 0.0 0.25 0.20 0.15 0.10 0.05 no blazar heating intermediate b. h. 0.00 0 1000 2000 3000 4000 velocity [km s 1 ] Puchwein+ (2012)

Optical depths and temperatures effective optical depth τeff 0.5 0.4 0.3 0.2 no blazar heating weak blazar heating intermediate blazar heating strong blazar heating Viel et al. 2004 Tytler et al. 2004 FG 08 temperature T ( )/(10 4 K) 3.5 3.0 2.5 no blazar heating weak blazar heating intermediate blazar heating strong blazar heating Becker et al. 2011 2.0 0.1 1.8 2.0 2.2 2.4 2.6 2.8 3.0 redshift z 2.0 2.2 2.4 2.6 2.8 3.0 redshift z Puchwein+ (2012) Redshift evolutions of effective optical depth and IGM temperature match data only with additional heating, e.g., provided by blazars!

Ly-α flux PDFs and power spectra PDF of transmitted flux fraction 10 1 10 0 10 1 10 1 tuned UV background no blazar heating weak blazar heating intermediate blazar heating strong blazar heating Kim et al. 2007 z = 2.52 10 0 z = 2.94 10 1 0.0 0.2 0.4 0.6 0.8 1.0 transmitted flux fraction Puchwein+ (2012)

Voigt profile decomposition decomposing Lyman-α forest into individual Voigt profiles allows studying the thermal broadening of absorption lines

Voigt profile decomposition line width distribution 0.08 0.07 0.06 N HI > 10 13 cm 2 2.75 < z < 3.05 no blazar heating weak blazar heating intermediate blazar h. strong blazar heating Kirkman & Tytler 97 PDF of b [skm 1 ] 0.05 0.04 0.03 0.02 0.01 0.00 0 10 20 30 40 50 60 b[kms 1 ] Puchwein+ (2012)

Lyman-α forest in a blazar heated Universe improvement in modelling the Lyman-α forest is a direct consequence of the peculiar properties of blazar heating: heating rate independent of IGM density naturally produces the inverted T ρ relation that Lyman-α forest data demand recent and continuous nature of the heating needed to match the redshift evolutions of all Lyman-α forest statistics magnitude of the heating rate required by Lyman-α forest data the total energy output of TeV blazars (or equivalently 0.2% of that of quasars)

Dwarf galaxy formation Jeans mass thermal pressure opposes gravitational collapse on small scales characteristic length/mass scale below which objects do not form hotter IGM higher IGM pressure higher Jeans mass: M J c3 s ρ 1/2 ( ) 1/2 T 3 IGM M J,blazar ρ M J,photo ( Tblazar T photo ) 3/2 30 depends on instantaneous value of c s filtering mass depends on full thermal history of the gas: accounts for delayed response of pressure in counteracting gravitational collapse in the expanding universe apply corrections for non-linear collapse

Dwarf galaxy formation Filtering mass M F [ h -1 M O ] 10 12 10 10 10 8 10 6 1 + δ = 1, z reion = 10 blazar heating only photoheating linear theory non linear theory non-linear theory optimistic blazar standard blazar only photoheating M ~ 10 M F M ~ 10 M F 11 10 o o M F, blazar / M F 10 1 10 1 1 + z C.P., Chang, Broderick (2012)

Peebles void phenomenon explained? mean density void, 1 + δ = 0.5 10 12 1 + δ = 1, z reion = 10 10 12 1 + δ = 0.5, z reion = 10 M F [ h -1 M O ] M F, blazar / M F 10 10 10 8 10 6 10 1 linear theory non-linear theory optimistic blazar standard blazar only photoheating 10 1 1 + z M F [ h -1 M O ] M F, blazar / M F 10 10 10 8 10 6 10 1 linear theory non-linear theory optimistic blazar standard blazar only photoheating 10 1 1 + z C.P., Chang, Broderick (2012) blazar heating efficiently suppresses the formation of void dwarfs within existing DM halos of masses < 3 10 11 M (z = 0) may reconcile the number of void dwarfs in simulations and the paucity of those in observations

Missing satellite problem in the Milky Way Springel+ (2008) Dolphin+ (2005) Substructures in cold DM simulations much more numerous than observed number of Milky Way satellites!

When do dwarfs form? 10 Myr M110 100 Myr 10 Gyr 1 Gyr Dolphin+ (2005) isochrone fitting for different metallicities star formation histories

When do dwarfs form? Dolphin+ (2005) red: τ form > 10 Gyr, z > 2

Milky Way satellites: formation history and abundance satellite formation time satellite luminosity function Maccio & Fontanot (2010) late forming satellites (< 10 Gyr) not observed! no blazar heating: linear theory non linear theory Maccio+ (2010) blazar heating suppresses late satellite formation, may reconcile low observed dwarf abundances with CDM simulations

Galactic H I-mass function Mo+ (2005) H I-mass function is too flat (i.e., gas version of missing dwarf problem!) photoheating and SN feedback too inefficient IGM entropy floor of K 15 kev cm 2 at z 2 3 successful!