Jets in Blazars and Radio
Galaxies: Conical Opening
Angles and Superdisks
Paul J. Wiita
Georgia State University, USA
Shanghai Astronomical Observatory, 12 May 2008
Outline:
• Basic Properties of Blazars
• TeV blazars: inverse Compton mechanism
boosting to the highest energies
• Conical jets vs. cyclindrical jets: modest
opening angles can explain many
peculiarities, including high Lorentz factors,
but slow radio knots
• Wide gaps between some lobes in radio
galaxies imply jets launched after mergers
Blazar Characteristics
•
•
•
•
•
•
Rapid variability at all wavelengths
Radio-loud AGN
Optical polarization “high”  synchrotron domination
BL Lacs show extremely weak emission lines
Double humped SEDs: RBL vs XBL?
Core dominated quasars (or FSRQs) clubbed w/ BL
Lacs to form the blazar class
• Population statistics indicate that BL Lacs are FR I
RGs viewed close to jet direction (Padovani & Urry 1992)
• The more powerful Flat Spectrum Radio Quasars are
FR II RGs viewed nearly along the jet (Padovani 2007)
Microvariability & Intraday Variability too
Romero, Cellone & Combi (2000); Quirrenbach et al (2000)
Blazar Spectral Energy Distributions
• Radio/IR/optical is
dominated by
synchrotron
emission, with
e ~ 103-105
• X-ray may be
synchrotron if
e > 107; or Inverse
Compton, where
e ~ 102 is OK
• Gamma-rays likely
to be IC and to get
TeV photons
7 might

~
10
e
BL Lac: Boettcher & Reimer 2004, ApJ, 609, 576
be needed
SED of TeV Blazar Mrk 421 in High &
Low States (Konopelko et al. 2003, ApJ, 597, 851)
Here x-rays at
peak of
synchrotron
(HBL) and
powerful
gamma-rays are
modeled by
Synchrotron selfcompton process
3C 130 & 3C 449: FR I’s
z=0.109; z=0.017
Canonical FR II: Cygnus A (z=0.056)
Quasar: 3C 175 (z=0.770)
Only 1 jet seen; core relatively more prominent than in
RG
VLBA of
3C279:
Apparent
Superluminal
Motion
with Vapp=3.5c:
really V=0.997c
at viewing angle
of 2 degrees
(z= 0.536)
RG Jets Start off With Relativistic
Bulk Motions
• Apparent superluminal motions seen in some
FR II RGs, especially flat spectrum quasars
seen in VLBI
• Gross asymmetries seen between jets and
counter-jets in FR II RGs: Doppler favoritism
• Correlated one-sided-ness almost always
seen between VLBI (pc-scale) and multi-kpc
jets
• Only plausible explanation for blazars
Jet of Quasar
3C 273
in IR,
radio + optical
& X-ray
(Uchiyama et al.
2006, ApJ, 648, 910)
Part I: Bulk Speeds of AGN Jets
• Big questions:
• What is the bulk Lorentz factor ?
• What is the true jet orientation angle ?
• Most of this part is based on three papers:
• Gopal-Krishna, Dhurde & Wiita, ApJ, 615, L81 (2004)
• Gopal-Krishna, Wiita & Dhurde, MNRAS, 369, 1287
(2006)
• Gopal-Krishna, Dhurde, Sircar & Wiita, MNRAS, 377,
446 (2007)
Estimating Bulk Doppler Factors ()
•
•
•
•
•
Boosted brightness temperature
Intraday radio flux variability
Models of SED of TeV blazars
Rapid variability of gamma-ray flux
The most direct measures come from
VLBI knot motions (but may arise from
shock, not bulk, velocities)
Doppler Factor from -ray Variability
Several blazars show obs < 1 hr for GeV -rays
If stationary source: size < c obs
For corresponding photon densities: +XSSCe++eHigh cross-section means -rays should not escape
If moving relativistically, then: size < c  obs
Thus photon opacity can be reduced sufficiently if
~100 (e.g., Krawczynski & Kirk 2002)
• Also, Gamma-Ray Bursts seem to require
~100-1000 (e.g. Sari et al. 1999; Meszaros et al. 2002)
• Is there an underlying similarity for AGN and GRBs?
•
•
•
•
•
•
Direct Estimates from VLBI
• For normal blazars
(Piner et al. 2006, ApJ, 640, 196)
• 0235+164: C1: app=25.6±7.0
C2: app= 8.9±1.3
C3: app= 7.9±4.7
• 0827+243: C2: app=25.6±4.4
C3: app=19.2±3.7
C4: app=12.3±7.4
C5: app=12.1±8.1
C6: app= 3.2±3.7
• 1406-076: C1: app=15.6±13.2
C2: app=28.2±6.6
C3: app=22.5±8.9
C4: app=15.8±2.0
Most are
quite
superluminal
VLBI Knot Speeds for TeV Blazars
(Piner & Edwards 2004, ApJ, 600, 115)
Mrk 421: C4: app=0.04±0.06
C5: app=0.20±0.05
C6: app=0.18±0.05
C7: app=0.12±0.06
Mrk 501: C1: app=0.05±0.18
C2: app=0.54±0.14
C3: app=0.26±0.11
C4: app=-0.02±0.06
1ES 1959+650: C1: app=-0.11±0.79
C2: app=-0.21±0.61
PKS2155-304: C1: app=4.37±2.88
1ES 2344+514: C1: app=1.15±0.46
C2: app=0.46±0.43
C3: app=-0.19±0.40
Most are
subluminal
or only
modestly
superluminal
Slow VLBI Knots in PKS 2155-304
•
Top row, natural weighting; bottom, uniform weighting with speeds: C1-1.15c, C2--0.46c, C3---0.19c (Piner & Edwards 2004)
How to have Small app in TeV Blazars?
1. Dramatic deceleration between sub-pc
(gamma-ray) and pc (radio) scales
(Georganopoulos & Kazanas 2003, ApJ, 594, L27)
Energetics are difficult; where does it go?
2. Very close alignment of the jet:

< 0.1o if =100 (statistically unlikely)
3. Fast spine ( > 30) and slow sheath (~3);
the spine would produce X- and -rays, while
the sheath would yield the radio synchrotron
photons (e.g. Ghisellini et al. 2005, A&A, 432, 401)
Distinctly possible, but not necessary
Jets Start Out Wide
•
Opening angle vs distance for M87 (Biretta et al. 2002) and Cen A
(Horiuchi et al. 2006)
So We Consider Conical Jets
• Assume a uniform radio emitting knot with a finite
opening angle, which may be comparable to the
viewing angle, and allow for large values of ,
which may be a function of transverse location.
Relevant Analytical Expressions
(Gopal Krishna et al. 2004)
Sobs=  n ().Sem()d  A()Sem
[where, n=3 for radio knots and A()=mean amplification factor]

 app

1
n


   dSem  d






S obs
Probabilit y of viewing angle :
p d    sin   Aq  d
q
where N S em dS em  S em
dS em
integral
source counts; q  1.5
(Fomalont et al. 1991)
High Gammas Yet Low Betas
• app vs  for jet
and prob of app >
 for opening
angles = 0, 1, 5,
10 degrees and 
= 50, 10
(continuum 2
boosting)
• Despite high  in
an effective spine
population
statistics are OK:
high probability of
low app
• Predict
transversely
resolved jets show
different app
Apparent Velocities for Conical Jets
• For  = 100: 40% sub-luminal (=5o)
70% sub-luminal (=10o)
• For  = 50: 15% sub-luminal (=5o)
30% sub-luminal (=10o)
<app> = 6 c (=5o)
• So high  and low app for TeV blazars can be
reconciled
• Small fraction of blazars must show app > 50
• Both dense VLBI monitoring and unbiased
interpretation of the data needed to check
Inferred Values for  for Conical Jets
Implications of Jet Angle Results
• If jets are moderately conical, the standard
analysis, which assumes =0, would lead to
serious underestimates of the jet orientation
angle,  (if  < 10o)
• Standard analysis would grossly overestimate
the deprojection factor, hence the true radio size
of the jet
• In-situ acceleration of TeV electrons in hot-spots
may not be needed-- they could be transported
• Parent population of blazars is not
overpredicted even if very high Lorentz factors
are assumed
Conical Spine-Sheath Jets
• We also consider jets where Lorentz factor varies
• (r) = 0exp(-2rq/)
• q=0 for constant ,
q=1 for mild transverse gradient;
q=2 for strong gradient
• The expectation values of the viewing angles decline
rapidly with 0 regardless of the values of  or q.
• But they level off at <> ~ /3 when the jets become
ultrarelativistic (0 > 30), particularly if >5o
Effective
Speeds (left)
and Doppler
Factors (right)
for p=3
& 0=20 (top),
0=50 (middle)
0=100(bottom)
Results for Spine-Sheath Conical Jets
• Decline of eff with  is faster for knots with higher .
• For well collimated jets ( < 0.5o) eff for uniform  is
typically 1.5-2 times more than for q=1 and 2-4 times
higher for q=2.
• Therefore the fastest spine component, close to the
jet axis, would be concealed in VLBI measurements.
• Again, for good collimation, uniform  jets would have
2-4 times larger eff compared to stratified jets,
implying Doppler boost factors ~10 times greater.
• Different VLBI speeds for different knots in the same
jet could only mean that surface brightness
distributions across similar speed knots are different.
Part II: Superdisks in Radio
Galaxies
• A small fraction of FR II RGs have lobes with large separations
(~25-30 kpc) and sharp parallel inner edges extending (~75 kpc
or more)
• These huge strip-like gaps imply the presence of a “superdisk”
made of denser material
(Gopal-Krishna & Wiita 2000, ApJ, 529,189)
Previous Interpretations of the Radio Gaps were Either:
•
•
Back-flowing synchrotron plasma in the radio lobes is blocked by the
ISM of the parent galaxy
(ISM arising from stellar winds and/or captured disk galaxies)
Buoyancy led outward squeezing of the lobe plasma by the ISM
• BUT, these wide gaps cannot be explained this way:
the ISM is too small
3C33
4C14.27
3C192
Ref: DRAGN Atlas (P. Leahy)
3C381
3C401
A Plausible Mechanism for the Radio Gaps at High Redshift
• Dynamical Interaction of radio lobes with a powerful thermal wind
outflowing from the AGN (Gopal-Krishna, PJW, Joshi, 2007, MN, 380,703)
Key Emerging Pieces of Evidence
• Non-relativistic winds (vw>103 km/s) and mass outflow ~1 M/yr are
generic to AGN
(e.g., Soker & Pizzolato 2005; Brighenti & Mathews 2006)
• Thus, relativistic jet pair and non-relativistic wind outflow seem to co-exist
(e.g., Binney 2004; Gregg et al. 2006)
•
Evidence: Absorption of AGN's continuum, seen in UV and X-ray bands
(review by Crenshaw et al. 2003)
• Wind outflow probably PRECEDES the jet ejection and can last for
w > ~ 108 yrs
(e.g., Rawlings 2003; Gregg et al. 2006)
• Wind outflow is quasi-spherical, while the jets are well collimated
(e.g., Levine & Gnedin 2005)
The Wind-Jet Model: Sequence of Events, 1
• Wind outflow from AGN blows an
expanding bubble of metal-rich, hot gas
into intergalactic medium
• Later, the AGN ejects a pair of collimated
jets of relativistic plasma
• The jets rapidly traverse the wind bubble
and often overtake the bubble’s boundary
• From then on, the high-pressure backflow
of relativistic plasma of the radio lobes
begins to impinge on the wind bubble,
from outside
• This sideways compression of expanding
wind bubble by the two radio lobes
transform the bubble into a fat pancake,
or superdisk
The Wind-Jet Model: Sequence of Events, 2
• The AGN's hot wind escapes through the
superdisk region, normal to jets
• The superdisk is "frozen" in the space. It
manifests itself as a strip-like central
emission gap in the radio bridge
• Meanwhile, the galaxy can continue to move
within the cosmic web
It can move ~ 100 kpc in ~ 300 Myr, with a
speed of ~ 300 km/s
• Thus, within about 108 years the parent
galaxy can even reach the edge of the radio
emission gap
(sometimes, even cross over into the radio
lobe: e..g., 3C16, 3C19)
• From then onwards, the two jets propagate
through very different types of ambient media
(wind material and radio lobe plasma)
Jets Overtake Many Bubbles
• Distance where (or
if) jets catch up to
bubbles is a
function of relative
powers (LJ/LW) and
delay between
wind and jet, tJ
• (a) - (d) go from
weak to strong
winds, all lasting
100 Myr
• Gray bands
correspond to
realistic lobe
energy densities
Gopal-Krishna, PJW & Joshi, 2007, MNRAS, 380, 703
Mergers Can Yield Superdisks at Low-z
• At z<1, the T~104K IGM assumed above isn’t around:
instead, RGs emerge into Intracluster Medium (ICM)
with T>107K
• We have just considered this situation in the context
of very asymmetric RGs with SDs
(Gopal-Krishna & Wiita 2008, New Astr.)
• Of 22 SD-RGs, 16 are substantially asymmetric, with
central galaxies well offset from center of SD,
sometimes even inside one lobe
Asymmetric SD RGs
(DRAGN atlas, P. Leahy)
(Saripalli et al. 2002)
Hot-Spot Asymmetries
• 13 of those 16 have hot-spots more symmetrically placed to
the SD midplane rather than the host galaxy
• Shown is Number of Sources against ratio of hotspot
distances to SD center (solid) and host galaxy (dashed)
Mergers of Ellipticals
• Can trigger jet launching
• If smaller galaxy is >0.1 mass of larger then
the gas attached to that galaxy is likely to
deposit its (orbital) angular momentum into
the host galaxy’s halo
• This can cause the halo to expand to SD
dimensions
• The host can get a kick from the merger
which, along with its random motion can
produce asymmetries over ~10-100Myr
Conclusions
• Part I: Modest opening angles (5º – 10º) of AGN jets can resolve the jet
Lorenz factor paradox of TeV blazars
– The frequently observed subluminal motion of VLBI knots can be
reconciled with the ultra-high bulk Lorenz factors (j >30 – 50) inferred from
rapid TeV and radio flux variability.
– Conical jets also produce larger central angles to line of sight and thus
smaller deprojected sizes
• Part II: Wide strip-like emission gaps are seen in some Radio Galaxies
and can’t be understood as arising from backflow onto normal ISM
_ Dynamical interaction between thermal (wind) and non-thermal (jet)
outflows resulting from the AGN activity, can produce fat pancake or
superdisk shaped regions at high redshifts.
– Mergers between elliptical galaxies can also produce superdisks; this is
more likely for low-z RGs.
– The observed asymmetries in lobe/core distances come out of these
scenarios
Finding Jet Parameters
• Determining bulk Lorentz factors, , and misalignment angles, , are
difficult for all jets
• Often just set  =1/ , the most probable value
• Flux variability and brightness temperature give estimates:
TB,obs 
S
( obs ) 2
1/(3 )
min
TB,obs 
 

 Tmax 
 2 app   2 min  1

2min
2 app
tan   2
 app   2 min 1
S = change in flux over
time obs
Tmax= 3x1010K
app from VLBI knot speed
 is spectral index
Conical Jets Also Imply
• Inferred Lorentz factors can be well below the actual
ones
• Inferred viewing angles can be substantially
underestimated, implying deprojected lengths are
overestimated
• Inferred opening angles of < 2o can also be
underestimated
• IC boosting of AD UV photons by ~10 jets would yield
more soft x-rays than seen (“Sikora bump”) but if >50
then this gives hard x-ray fluxes consistent with
observations
• So ultrarelativistic jets with >30 may well be common
Inferred Lorentz Factors
inf vs.  for =100,
50 and 10 for =5o
P() and < inf>
Inferred Projection Angles
• Inferred angles can be well below the actual viewing
angle if the velocity is high and the opening angle even a
few degrees
• This means that de-projected jet lengths are
overestimated
Modeling the Dynamics of the Bubble and the Jets
(Gopal Krishna, Wiita & Joshi 2006)
(Uses the analytical works of Levine & Gnedin 2005; Scannapieco &
Oh 2004; Kaiser & Alexander 1997)
Asymptotic (equilibrium) radius of the wind bubble:
 3

E
Req ( Mpc )  3 10 
 60 w 
 4Pext 10 erg 
5
PIGM  1   m   nb kTIGM ,
1
3
where nb  3 10 7 1  z  cm 3
3
2
 B 
 dyn.cm  2
Plobe  3.2 10 12 
 10 G 
Key Blazar Conclusions
• Blazars are dominated by emission from jets
• Variations within the jet are Doppler boosted and
greatly amplified
• TeV blazars almost certainly require very high
Lorentz factors but often show slow VLBI knots
• Allowing for conical jets means ultrarelativistic jet
speeds can produce slow apparent speeds, even for
fast spine--slow sheath structures
• They also produce larger central angles to line of
sight and thus smaller deprojected sizes