Can Multiband Observations Constrain Explanations for Knotty Jets?
aa r X i v : . [ a s t r o - ph . H E ] F e b CAN MULTIBAND OBSERVATIONS CONSTRAINEXPLANATIONS FOR KNOTTY JETS?
D. E. Harris
MS-3, CfA, 60 Garden St.Cambridge, MA 02138, [email protected]
ABSTRACT
One can imagine a number of mechanisms that could be the cause ofbrighter/fainter segments of jets. In a sense, jets might be easier to understandif they were featureless. However we observe a wide variety of structures whichwe call “knots”. By considering the ramifications of the various scenarios for thecreation of knots, we determine which ones or which classes are favored by thecurrently available multiwavelength data.
Subject headings: relativistic jets; jet structure
1. Introduction
With the advent of the Chandra X-ray Observatory, sub arcsec resolution became avail-able in the 0.5 to 10 keV band. This technological advance resulted in an increase in thenumber of sources with detections of X-ray jets and hotspots from a few to close to onehundred. During the last ten years, it became apparent that the conventional wisdom (thatknots in jets were the results of internal shocks) was only part of the story. We now realizethat for X-ray synchrotron features, any emission region must also be an acceleration regionsince the timescales for radiative losses of the emitting electrons is much shorter than traveltimes from one part of the jet to downstream regions. Characteristic loss times for X-raysynchrotron-emitting electrons is ten times shorter than for those responsible for the opticalemission.
One of the main obstacles in comparing features in FRI jets with those in quasar andFRII jets is the large change in physical resolution for a fixed angular resolution. In Table 1 2 –we give the physical resolution corresponding to one arcsec for a sample of sources. Typicalangular resolutions for adaptive optics, the VLA, MERLIN, and HST will be 0.1 arcsec andVLBA can provide resolutions of 1 to 10 milli arcsec.When we make our calculations for physical parameters of jet knots, we are most likelymaking gross errors because of this limitation. In fig. 1 we can understand that measuringthe size, brightness, intensity, and spectrum of a knot in the M87 jet, or in the jet of Cen A(if it were at the distance of M87), will not provide the data necessary to derive the correctphysical parameters for the features we can actually discern in the full resolution X-ray map(bottom panel).
While no definitive answer to this question has been demonstrated, we are reasonablyconfident that X-ray emission from FRI jets is dominated by synchrotron emission. Theremaining contentious issue is the nature of the X-ray emission from quasar jets. If it issynchrotron, then the electrons responsible for the radiation will have energies of γ ≥ ( γ is the Lorentz factor of the electrons). However, since the energy density of the cosmicmicrowave background (CMB) in the jet frame will be augmented by Γ (Γ is the bulk Lorentzfactor of the jet) if Γ ≥
5, the resulting inverse Compton (IC) emission comes from electronswith γ ≈ E losses (i.e. synchrotron and IC) result in halflives of ordera year, whereas in the latter case, it would be ≥ years, and even longer for regions withmagnetic field strengths significantly less than 1 mG (Harris & Krawczynski 2002). Knotmorphology and intensity ratios between radio and X-ray might be quite different dependingon which process dominates the X-ray emission.Synchrotron self-Compton (SSC) emission (for which the target photons are the syn-chrotron photons) has been found to provide reasonable fits to the spectra of some FRIIhotspots, but has generally not been able to explain knot emission (Hardcastle et al. 2004).
2. Multiband Aspects of Knots
For almost all X-ray detections of knots and hotspots, there is a very good correspon-dence between the X-ray and radio morphologies, but the intensity ratio (X-ray flux dividedby the radio flux) varies considerably. 3 –Table 1: Physical size corresponding to one arcsecSource redshift Distance size Description(Mpc) (pc)Cen A 3.5 17 FRI jetM87 0.00427 16 77 FRI jetPicA 0.035 152 700 FRII jet and hotspotCygA 0.056 247 1070 FRII hotspots3C273 0.158 749 2700 CDQ jet3C109 0.3056 3338 4500 FRII hotspot3C263 0.656 3934 7000 LDQ hotspot3C280 1 6600 8000 FRII hotspots3C9 2 15850 8500 LDQ jet1508 4.3 39809 6900 CDQ knotFig. 1.— Chandra images of the M87 jet (top) and the Cen A jet (middle and bottom). Allimages have been rotated and contours increase by factors of two. The scale bars are 1 kpclong. The M87 image has been regridded to one tenth of a native ACIS pixel and smoothedwith a Gaussian of FWHM=0.25 ′′ . The center image is how the Cen A jet would appear ifit were at the distance of M87. The bottom image shows the event file which demonstratesthat the lower resolution image misses the fine scale structures. 4 – R xr From a study of over a hundred knots and hotspots with both radio and X-ray detections,it has been found that R xr lies in the range 1 to 100 for knots of both FRI radio galaxiesand quasars whereas most hotspots (both FRII and quasar) have ratios in the range 0.03to 3 (Harris, Massaro, & Cheung 2010). In that work we did not sample the much largernumber of radio knots with only upper limits of X-ray intensity, so it is quite likely thatthere are many knots and hotspots with smaller R xr values. A priori , if quasar knots wereall dominated by IC/CMB X-ray emission, we might have expected to see a clear differencein R xr between quasars and FRI knots; instead we find essentially the same range for bothclasses of sources. “Offsets” is a term we use to describe a common (but not universal) property of in-dividual X-ray knots. In many cases, when observed with similar angular resolutions, thebrightness of the X-ray image peaks upstream of the lower frequency emissions. This be-havior is seen also when comparing optical and radio morphologies, and to the best of ourknowledge, always occurs in the sense that the higher frequency peaks upstream of lowerfrequencies. This subject is discussed in section 3.2.1.1 of Harris & Krawczynski (2006), andseveral examples are shown.The term “progressions” is used to describe a systematic change in the overall spectralproperties of knots as a function of distance from the nucleus. This has also been coveredin Harris & Krawczynski (2006) (section 3.2.1.2). The systematic change is best seen in theratio of X-ray flux to radio flux and occurs in the sense that the ratio is larger closer to thenucleus. In the case of 3C 273, the ratio changes by two orders of magnitude, whereas for4C 19.44, the effect is essentially absent.It is not difficult to see that if a jet segment that behaved like 3C 273 were observedwith a single resolution element, it would show a marked offset between the peak of theX-ray compared to the peak of the radio distribution.Since this sort of effect has been observed over many physical scales, it is likely thatsynchrotron loss time compared to travel time down the jet is not the only cause of offsets.Rather some other mechanism is at work such as a progressively larger field strength movingdown stream. That would produce higher radio intensities as well as perhaps curtailing theproduction of the very high energy electrons required to produce X-ray synchrotron emission. 5 –
3. Mechanisms for Knot Production
Our basic assumption is that a knot is a region of enhanced emissivity which is producedby the jet. It is moving relativistically, but not necessarily at the same velocity as “the jet”(i.e. the velocity of the power flow) (Harris & Krawczynski 2007). One can imagine severalmechanisms for modulating the emissivity along a jet. We consider several possibilities, andsuggest a few diagnostics. We do not consider the underlying reasons for the existence of anyparticular knot at any particular location (i.e. instabilities, interactions with stars, molecularclouds, etc).
Perhaps the most intuitive explanation for knotty jets is the common notion of a series ofshocks. Each would create a new supply of relativistic electrons with a power law distributiondetermined by the local conditions. The eventual dimming as the shocked plasma is advecteddown stream can be caused by E losses or expansion (first power of energy). For E losses,we expect the lower frequency emissions to last longer, leading to offsets in peak brightnessas we move downstream. This is often seen in synchrotron jets; for IC/CMB models, theX-rays come from low energy electrons and should last longer. This behavior is almost neverseen, although the end of the jet in 4C19.44 could be an example. Although quite similar to the shock scenario, there is no shock acceleration as such.The only changes to the electron energy distribution comes from the change in volume of theemitting region. In both the shock case and the change in volume, compression augments themagnetic field and boosts the energy of electrons; expansion reduces synchrotron emissionboth from the lowering of electron energies, but also by the drop in field strength andmoreover, for a fixed observing band, a lower field means you are observing higher energyelectrons than previously so you are sampling a segment of the electron spectrum that hasmany fewer electrons. In the case of IC/CMB X-rays, the emissivity drops only becausethe normalization factor of the power law distribution of electron energy drops; the changein magnetic field has no effect. Therefore, if expansion were to be the dominent operatorfor separating adjacent knots, it would mean that the contrast from knot to inter-knotshould be greater in the radio/optical than in the X-rays for the IC/CMB model whereasif synchrotron emission dominates the X-ray emissivity, the contrast should be sensibly the 6 –same at all frequencies.
If kpc jets are similar to pc scale jets, an episodic supply of power to the jet by the SMBHcould produce a series of moving knots: knots represent high power intervals of activity, gapsare when the power is low or absent (c.f. the “flip-flop” model of jet formation). If sucha mechanism were to be the only formative one, there would have to be two timescales:one for pc scale jets and the other for kpc scale jet knots. Current evidence does not favorthis scenario, e.g. the upstream edge of HST-1 (a jet knot close to the nucleus of M87)was thought to have an apparent velocity close to c (downstream blobs were estimated at6c (Biretta, Sparks, & Macchetto 1999)) from HST data in the 1990’s. More recently, wemeasured comparable values at 1.7 GHz (Cheung, Harris, & Stawarz 2007). However, theupstream edge of HST-1 has not moved during the intervening 10 years, consistent with aninterpretation in terms of a stationery shock. We suspect that both estimates for the motionof the upstream edge were centroid shifts caused by the ejection of new components.
If the path of a jet changes direction compared to the line of sight, either by thrashingor by a regular (e.g. helical) path, apparent knots can be produced even though the jet itselfhas a steady power flow. Once the jet has Γ ≥ a few, moving in and out of the beamingcone can produce the required brightness fluctuations.The simple expectation is that the radio and X-ray emissions will be coincident: eachknot will have the same location and morphology for all wavelengths, i.e. no offsets areexpected in the brightness distributions. If the X-rays come from IC/CMB, for most jetsthe contrast should be higher for the X-rays because of the extra beaming factor of IC/CMB(Harris & Krawczynski 2002; Massaro et al. 2009). If it were possible to “store” jet energy in some other form than in the bulk Lorentzfactor, it might be conceivable to envisage a jet with an oscillating value of Γ. However, if thetotal energy flow is proportional to Γ, it would seem difficult to allow Γ to drop substantiallyand then to increase again. If we separate the emitting region from the underlying jet, then 7 –the knot’s emission might well decay from a drop in Γ, and the subsequent knot would haveto rely on one of the other possible explanations to generate a new emitting volume.
4. Summary
What we observe is not necessarily “the jet”. “The jet” is whatever it is that carries theenergy from the environs of the SMBH to distances of hundreds of kpc. There are a numberof possibilities: magnetic field/ Poynting flux, hot or cold protons, or cold pairs. What we seeare hot electrons/positrons, but these cannot be the agent that transports the energy sincethere are inescapable IC losses for electrons with γ ≥ E losses), the larger theIC losses since the effective energy density of the CMB increases as Γ . Thus we have theriver analogy: “the jet” is a river with smoothly flowing water; the emission we see is likewhite water produced by turbulence around rocks in the river or waterfalls. The white wateris a product of the river and may well be carried along by the river’s flow, but not necessarilywith the underlying velocity of the water. All our observations describe the product, notthe jet itself. When we see a knot, we see a location where energy is transferred from thejet to produce hot (radiating) electrons. For many knots, the energy transferred is a smallfraction of the total power of the jet, whereas for terminal hotspots in FRII radio galaxies,the transfer is complete.If IC/CMB dominates the X-ray emission, it seems that creating knot structure is anon-trivial problem. Curved trajectories and episodic ejection from the SMBH may beamongst the few viable options, since once a substantial population of low energy electronsis generated, it is difficult to reduce the emissivity and then increase it again.We suggest a diagnostic of comparing the brightness contrast between the peak intensityof a knot and the preceding and following minimum brightness (the gap between knots). Ifexpansion and contraction is the dominant mechanism for knot production, the synchrotronemission should have a much larger contrast than IC/CMB emission. If however a change indirection of the beaming cone is the principal operative, then the IC/CMB emission should(statistically) display the higher contrast. Acknowledgments
It is a pleasure to acknowledge collaborators C. C. Cheung, F. Massaro, and L. Stawarz.Partial support for this work was provided by NASA grants AR6-7013X and G09-0108X. 8 –
REFERENCES
Biretta, J. A., Sparks, W. B. & Macchetto, F. D. 1999 ApJ 520, 621Cheung, C. C., Harris, D. E. & Stawarz, L. 2007 ApJ 663, L65Hardcastle, M.J., Harris, D. E., Worrall, D. M., & Birkinshaw, M. 2004 ApJ 612, 729Harris, D. E. and Krawczynski, H. 2002 ApJ 565, 244Harris, D. E. & Krawczynski, H. 2006 ARAA 44, 463Harris, D. E. & Krawczynski, H. 2007 Revista Mexicana de Astronomia y Astrofisica, Seriede Conferencias 27 Contents of Supplementary CD, p.188.Harris, D. E., Massaro, F., & Cheung, C. C. 2010 to appear in
AIP Conference Series, A.Comastri, M. Cappi, L. Angelini, editors
Massaro, F., Harris, D. E., Chiaberge, M., Grandi, P. Macchetto, F. D., Baum, S. A., O’Dea,C. P. & Capetti, A. 2009 ApJ 696, 980