1. Photon breeding mechanism in jets and its observational signatures
Stern & Poutanen 2006 MNRAS (description of the mechanism, analytical estimates problem setup, 1D simulations)‏
Stern & Poutanen 2008 MNRAS (more extensive 2D simulations,astrophysical implications)‏
Stern & Tikhomirova 2009 MNRAS accepted (possible observational signature: synchrotron cutoff in GeV range)

2. Converter mechanism: an alternative way of energy transfer from a relativistic fluid to particles
Fermi acceleration: a charged particle crosses a shock front
back and forth gaining factor 2 in energy at each round trip.
It is not efficient in the case of relativistic shocks.
Derishev et al. (2003) and Stern (2003) independently
suggested an alternative: converter mechanism.
conversion
Fermi
P --> n
e --> 

3. Charge conversion:
1. proton + photon--> n + p+ (photomeson production)‏
2. electron + photon --> electron + photon
(inverse Compton scattering)‏
3. photon + photon --> e+ e-
A background of low energy photons with column
density > 1025/cm2 is required
Fermi mechanism: small energy losses and a slow energy gain at a large 
Converter mechanism: large energy losses at the charge exchange and a fast energy gain at a large 

4. Photon breeding is the version of the converter mechanism with
 --> e+e -->  conversion in the supercritical regime (nuclear pile).
Breeding rather than acceleration because the spectrum of particles evolves slowly while their number rapidly increases.
Stern & Poutanen (2006, 2008)‏
Dissipation of the jet bulk energy into gamma-rays via the photon breeding

5. The jet C2 ~ 2 C1 Criticality: C1 x C2 x C3 x C4 x C5 >1
C3 (Comptonization)‏
C2 ~ 2 C1
C4 (escape from the jet +pair production)‏
C5 Comptonization
Criticality:
C1 x C2 x C3 x C4 x C5 >1
The external environment

6. What we need to launch the mechanism:
1. Transversal magnetic field in the jet Hj --> energy gain.
2. Transversal magnetic field He outside --> reflection.
3. Soft background radiation. --> charge conversion
(i) Radiation of the disk, Fd, (mainly along the jet) - insufficient
(ii) A scattered, nearly isotropic component Ft.
6. A seed high energy photon.

7. ? Photon breeding sites The vicinity of the disk ~ 100 Rg
Disk radiation is sufficient ?
R ~ 1015 cm
Model C
No Ft

8. Photon breeding sites Broad line region R ~ 1017 cm dn/d ~ -1
Ft: La, scattered disk
radiation, dust emission
dn/d ~ -1
dn/d ~ -1.4

9. The numerical model (Model A – broad line region)‏
Jet: the cylinder split into 64 X 20 cells which can independently decelerate, (two dimensional ballistic approximation). R = cm, Rj = 1016 cm
Soft photon field: disk radiation (multicolor spectrum) along the jet +
isotropic radiation, spectral number index = -1.4, xmax = Tmax = 5 eV
F iso= Fd/20
Magnetic field:
(Bext = 10-3 G).
Boundary conditions:  = const at the inlet The fixed cylindrical boundary. The external medium fixed at rest.
The problem: time evolution of the system from a non-radiating state.
Start: uniform fluid, a small amount of seed high energy photons
f(r)‏ 
20Rj

10. Particle simulation is precise and model independent:
simulation: Large particle nonlinear Monte-Carlo code (Stern, 1988, Stern et al. 1995). Compton scattering, pair production, synchrotron radiation, fine tracking in magnetic field. 219 – 221 large particles.
Output: emitted synchrotron and Comptonized photons of evolving spectrum and intensity, differential deceleration of the jet
Note:
Particle simulation is precise and model independent:
no artificial injection except a negligible amount at the start.

11. Results R=2 1017 cm =20, magnetically dominated LB=Ld, L L/Lj
erg/s,
=20, LB = 0.2 Ld
LB = 0.2 Ld, Ld = erg/s
 f Lg/Lj
Breeding works down to =8 at a low magnetization

12. Beam-on photon spectra “at infinity”.
An important detail is missing or poorly expressed: the “synchrotron peak”

13. Electron spectra.
They share up to 4% of the jet energy.
Slowly cooling pairs
!!!
If these pairs were slightly reheated, this would solve the problem of the synchrotron peak

14. Deceleration of the jet
2 parsec
1045
1045  
BLR, L =
z=20
Radius
z = 2
Deceleration of the jet

15. Ld, Lj, erg/s
R = 1017cm!
1046
A very robust nonlinear regime.
Is not sensitive to the magnetic field and external radiation.
Works at favourable conditions,
low B, soft radiation
1045
A stable linear regime, requires a soft isotropic radiation ~1/20 Ld
1044
1043 G 10 20 30

16. The ultimate synchrotron cutoff
(Guilbert, Fabian, Rees, 1983)‏
Es,max = mec2/a G
Low luminosity spectra have a feature observable by Fermi:
a dip at several GeV
Ft ~ 0.05 Fd
Ft ~ 0.01 Fd

17. Costamante, Aharonian, Khangulyan (2007) noticed something like this
in the spectrum of
PKS
EGRET + HESS data
The condition for MeV synchrotron hump: huge opacity of UV across the jet e = 10 (G/30)3 (B/10G)‏

18. Concluding remarks
- In the case of powerful jets of AGNs it is difficult to avoid the dissipation via the photon breeding.
- It is motivated by the requirement of a high efficiency of Gamma-ray emission.
- It, when taken alone, can not explain the entire spectrum of blazars. The photon breeding should be followed by a moderate pair reheating, which is quite natural. .

19. -- The mechanism is not unique, it probably does not work in the case of M87 and some BL Lacs
-- More likely it works together with Fermi acceleration and provides 90% of the energy release.
To be observed by Fermi (if the breeding really works):
1. A few per cent of blazars shoud demonstrate the ulthmate synchrotron cutoff in GeV range
2. There should be a kind of bimodality: the photon breeding “on” and “off”

20. EGRET
blazars
reduced c2 > 2
after a fit with power law

21. Exponential regime: Start of the breeding from extragalactic gamma-ray background
Breeding due to downstream-upstream feedback loop
3 years
Upstream – downstream photon avalanche