Radio Evidence for AGN Activity: Relativistic as Tracers of SMBHs
aa r X i v : . [ a s t r o - ph . H E ] D ec Star Clusters and Black Holes in Galaxies and Across Cosmic TimeProceedings IAU Symposium No. 312, 2015Fukun Liu ed. c (cid:13) Radio Evidence for AGN Activity:Relativistic Jets as Tracers of SMBHs
K. I. Kellermann
National Radio Astronomy Observatory,520 Edgemont Rd., Charlottesville, VA, USAemail: [email protected]
Abstract.
Although the radio emission from most quasars appears to be associated with star formingactivity in the host galaxy, about ten percent of optically selected quasars have very luminousrelativistic jets apparently powered by a SMBH which is located at the base of the jet. Whenthese jets are pointed close to the line of sight their apparent luminosity is enhanced by Dopplerboosting and appears highly variable. High resolution radio interferometry shows directly theoutflow of relativistic plasma jets from the SMBH. Apparent transverse velocities in these so-called “blazars” are typically about 7c but reach as much as 50c indicating true velocities withinone percent of the speed of light. The jets appear to be collimated and accelerated in regionsas much as a hundred parsecs downstream from the SMBH. Measurements made with Earth tospace interferometers indicate apparent brightness temperatures of ∼ K or more. This iswell in excess of the limits imposed by inverse Compton cooling. The modest Doppler factorsdeduced from the observed ejection speeds appear to be inadequate to explain the high observedbrightness temperatures in terms of relativistic boosting.
Keywords.
AGN, quasars, radio galaxies, jets, SMBHs
1. Why Radio?
Historically the first speculations about the existence of active galactic nuclei (AGN)and super massive black holes (SMBHs) came from the huge energy requirements impliedby the discovery of distant powerful radio galaxies and quasars. Today, radio observa-tions remain crucial to understanding the role of SMBHs in astrophysics. Only at radiowavelengths is it possible to image the region immediately surrounding the SMBH cen-tral engine and the relativistic jets which apparently originate with the SMBH. Typicalresolution obtained with Very Long Baseline Interferometer (VLBI) observations at cen-timeter wavelengths is of the order of 0.001 arccsecond (1 milliarcsec). Thus, for nearbysources such as those located in the Virgo cluster, a linear resolution of 1 milliarcseccorresponds to only about 0.1 parsec or about 100 Schwartzchild radii for the SMBHlocated in the nucleus of M87. For z ∼ ∼ ∼ ∼
10 GeV moving in weak mag-netic fields with B ∼ − to 10 − Gauss. The high energy electrons are thought to beaccelerated in one of two ways; either by a central engine associated with accretion onto1 K. I. Kellermanna SMBH in elliptical galaxies or in quasars, or by supernovae following massive star for-mation (starbursts) in the nucleus of early type galaxies. Unfortunately, both processesare often referred to as AGN, and this has led to considerable confusion in the literature.Above ∼ ± per ster, or 60 times greater than the density of thefaintest (mag 29) galaxies in the Hubble Ultra Deep Field (Condon et al. 2012). Thusif the excess background temperature observed by ARCADE is real and due to discretesources, these sources cannot be associated with any known galaxy population. It will beimportant to verify that the ARCADE 2 results were not contaminated by unrecognizedGalactic radio emission.Radio emission due to AGN can usually be distinguished from that due to star forma-tion in a variety of ways. • Morphology:
Star formation sources may have dimensions of the order of a fewtenths of an arcsecond or a few kiloparsecs at cosmological distances while SMBH drivenAGN sources are typically very small, of the order of 0.001 arcsec (10 pc) or less, andare coincident with the galaxy nucleus or QSO. Quasars and AGN powered by SMBHsare often variable on time scales as short as days with corresponding changes in theirmorphology indicating highly collimated outflows with apparent superluminal velocity.SMBH driven AGN may also contain extended lobes tens or hundreds of kiloparsecsdistant from the compact nucleus, and sometimes show optical and radio jets joining thenucleus and radio lobes. • Radio Spectra:
Star forming sources and the extended radio lobes of AGN gener-ally have steep radio spectra. Due to synchrotron self absorption, the compact sourcesgenerally have flat or even inverted spectra. • Brightness Temperature:
Star forming sources mostly have measured brightnesstemperature up to ∼ K while the compact flat spectrum sources associated withAGN have brightness temperature 10 − K or more. • Radio Luminosity:
Star forming regions typically have a radio luminosity close to10 − W/Hz and follow the well known correlation between radio and FIR luminosityCondon(1992). Radio galaxies and quasars driven by SMBHs may be 10 − times moreluminous so their radio luminosity greatly exceeds that expected from the radio/FIRrelation characteristic of star forming regions. • X and γ -ray emission: Star forming regions are only weak x-ray sources withtypical luminosity ∼ ergs/sec while SMBH driven AGN can be strong x-ray, γ -ray,and TeV sources. adio Galaxies, AGN, and SMBHs • Host Galaxies:
SMBH driven radio sources are located in the nuclei of ellipticalgalaxies or are associated with quasars which themselves are thought to be the brightnuclei of elliptical galaxies that greatly outshine their host galaxy. Low (optical) lumi-nosity AGN are typically found in early type spiral (often classified as Seyfert) galaxies.Radio emission from star forming regions is typically associated with spiral galaxies butmay also be found in the host galaxies of radio quiet quasars (see Section 4).
2. Early Evidence for AGN and SMBHs
Perhaps the first suggestions that the nuclei of galaxies may contain more than juststars came from Sir James Jeans in 1929 who remarked in his book on Astronomy andCosmogony (Jeans 1929),The centres of the nebulae are of the nature of singular points at which matter ispoured into our universe from some other and entirely spatial dimension so that to adenizen of our universe, they appear as points at which matter is being continuouslycreated.The modern understanding of the important role of galactic nuclei probably beganwith the famous paper by Karl Seyfert (1943) who reported on his study of broad strongemission lines in the nucleus of seven spiral nebulae. Interestingly, although Seyfert’sname ultimately became attached to the broad category of spiral galaxies with activenuclei, his 1943 paper received no citations until 1951, and apparently went unnoticeduntil Baade and Minkowski (1954) drew attention to the similarity of the Cygnus A radiosource spectrum with that of the galaxies studied by Seyfert.Not until the 1949 Nature paper by Bolton, Stanley, and Slee (1949) did astronomers fi-nally recognize the vast energy requirements of radio galaxies. Bolton et al. had identifiedthree of the strongest discrete radio sources with the Crab Nebula, M87, and NGC 5128,Until that time the discrete radio sources were widely thought to be associated withgalactic stars. This was understandable, as Karl Jansky and Grote Reber had observedradio emission from the Milky Way. The Milky Way is composed of stars, so it was natu-ral to assume that the discrete radio sources had a stellar origin. Bolton et al. understoodthe importance of their identification of the Taurus A radio source with the Crab Nebulawhich was widely recognized as the remnant of the 1054 supernova reported by Chineseobservers. BSS correctly identified two other strong sources with M87 and NGC 5128,but realizing that if they were extragalactic, their absolute radio luminosity would needto be a million times more luminous than that of the Crab Nebula, they argued that“NGC 5128 and NGC 4486 (M87) have not been resolved into stars, so there is littledirect evidence that they are true galaxies.” So they concluded that they are withinour own Galaxy. Indeed their paper carried the title “Positions of Three Discrete RadioSources of Galactic Radio Frequency Radiation.” John Bolton later argued that he reallydid understand that M87 and NGC 5128 were very luminous radio sources, but that hewas concerned that that in view of their apparent extraordinary radio luminosity, Naturemight not publish their paper.The following years saw the identification of more radio galaxies, and the changedparadigm which had previously considered all discrete radio sources to be stellar to onewith most high latitude sources were assumed to be extragalactic. The energy require-ments were exacerbated in 1951 with the identification of Cygnus A, the second strongestradio source in the sky with a magnitude 18 galaxy at what was then considered a highredshift of 0.056 and a corresponding radio luminosity about 10 times more luminous K. I. Kellermannthan M87 and NGC 5128 (Baade and Minkowski 1954). The total energy contained in rel-ativistic particles and magnetic fields in the radio lobes of Cygnus A and other powerfulradio galaxies was estimated to be at least 10 − ergs (Burbidge 1959).Hoyle, Fowler, Burbidge and Burbidge (1964) were apparently the first to call attentionto gravitational collapse as a possible energy source to power radio galaxies. By the middleof 1960, many radio sources had been identified with galaxies having red shifts up to 0.24(Bolton 1960). Typically the optical counterpart of strong radio sources was identifiedwith an elliptical galaxy that was the brightest member of a cluster. In 1960, RudolphMinkowski (1960) identified 3C 295 with a mag 20 galaxy at z=0.46. 3C 295 is about tentimes smaller than Cygnus A and ten times more distant consistent with the idea that thesmallest radio sources might be path finders to finding very distant galaxies. But a fewmonths later Caltech radio astronomers identified the first of several very small sourceswith what appeared to be galactic stars, thus raising questions about the extragalacticnature of other small diameter radio sources.
3. The First Quasars
While searching for ever more distant radio galaxies, Caltech radio astronomers JohnBolton and Tom Matthews identified 3C 48 with an apparent stellar object. At the 107thmeting of the American Astronomical Society held in New York in December 1960, AllanSandage (1960) reported the discovery of “The First True Radio Star.” Before he leftto return to Australia, John Bolton (1990) speculated that 3C 48 had a high redshift of0.37, but was apparently dissuaded by Jesse Greenstein and Ira Bowen on the groundsthat there was a 3 or 4 Angstrom discrepancy among the corresponding rest wavelengths.In a subsequent analysis of the complex emission line spectrum, Jesse Greenstein (1962)interpreted the 3C 48 spectrum in terms of emission lines from highly ionized states of rareearth elements. He briefly considered a possible redshift of 0.37, but quickly dismissed thepossibility that 3C 48 was extragalactic. Nearly two years would pass, and other compactradio sources would be identified as galactic stars before a series of lunar occultationswould lead to the identification of 3C 273 with a star like object at a redshift of 0.16 andthe immediate realization that 3C 48 was also extragalactic with a redshift of 0.37 leadingto the recognition of quasi stellar radio sources or “quasars” as the extremely bright nucleiof galaxies. The apparent high radio as well as optical luminosity of quasars, coupled withtheir very small dimensions presented a further challenge to understanding the source ofenergy and how this energy is converted to relativistic particles and magnetic fields.
4. Radio Loud and Radio Quiet Quasars
The following years led to the identification of more quasars at ever larger redshiftsand the suggestion that quasars are powered by accretion onto super massive black holes(SMBH) with masses up to 10 solar masses or more (Lynden-Bell 1969). Generally,the identified quasars had a significant UV excess compared with stars, so due to theredshift of their spectrum, they appeared blue on photographic plates facilitating theiridentification with radio sources with even modest position accuracy.In 1965, Sandage noted that the density of blue stellar objects on the sky was somethousand time greater than that of 3C radio sources. Sandage argued that what hecalled “quasi stellar galaxies” are related to quasars, except that they are not strongradio sources. But, his paper was widely attacked, perhaps in part because of the per-ceived irregular treatment by the Astrophsyical Journal. Sandage’s paper was receivedon May 15, 1965 at the Astrophysical Journal, but S. Chandrasekar, the ApJ editor was adio Galaxies, AGN, and SMBHs M i = -23 so were genuine quasars that presumably contained aSMBH. The observations were made with the Jansky Very Large Array at 6 GHz reachingan rms noise of 6 µ Jy. All but about 6 quasars were detected as radio sources with anobserved radio luminosity sharply peaked between 10 and 10 Watts/Hz characteristicof the radio luminosity typically observed from star forming galaxies. About ten percentof the SDSS sample are strong radio sources with radio luminosities ranging up to 10 W/Hz. Kimball et al. concluded that the radio emission from radio quiet quasars is dueto star formation in the host galaxy. Similar conclusions were reached by Padovani et al.(2011, 2014) based on the identification and classification of the microJy radio sourcesfound in a deep VLA survey of the Extended Chandra Deep Field South. Based on radio,optical, IR, and X-ray data, Padovani et al. concluded that the microJy radio emissionfrom AGN, like that of galaxies, is powered primarily by starbursts, and not the SMBHswhich powers the AGN. Condon et al. (2013) argue that these starbursts are fueled bythe same gas that flows into the SMBH that powers the quasar and thus accounts forthe co-evolution of star formation and SMBHs.
5. Jet Kinematics and Relativistic Beaming
Shortly after the recognition of quasars, radio source observations in both the SovietUnion (Sholomitskii 1965) and the U.S. (Dent 1965) demonstrated variability on timescales of months or less. This presented a problem. Causality arguments suggested lineardimensions, d c τ where c is the speed of light and τ the characteristic time scale ofthe observed variability. Knowing the quasar redshift and corresponding distance puts alimit to the angular size which for many variable sources was only ∼ − arcsecondsand the corresponding lower limit to the brightness temperature which appeared to besignificantly in excess of the inverse Compton limit of ∼ . KFor most variable sources, the apparent violation of the inverse Compton limit isnow understood in terms of relativistic beaming. Due to relativistic effects, we observeapparent jet speeds, luminosities, and brightness temperatures which are related to thecorresponding intrinsic quantities in the AGN rest frame through the Doppler factor, K. I. Kellermann δ , the Lorentz factor, γ , and the jet orientation, θ , with respect to the line of sight(Cohen et al. 2007.The apparent velocity transverse to the line of sight, β app , the apparent luminosity, L ,the apparent brightness temperature, T app and the Doppler factor, δ , can be calculatedfrom the Lorentz factor, γ , θ , and the intrinsic luminosity, L o . The apparent transversevelocity β app is given by β app = β sin θ − β cos θ , (5.1)where β = v/c For small values of θ , because the radiating source is almost catching up with its ownradiation, equation 5.1 shows that the apparent transverse can exceed the speed of light,which is commonly referred to as ”superluminal motion.” The apparent luminosity, L , isgiven by L = L o δ n , (5.2)where the Doppler factor, δ , is δ = γ − (1 − β cos θ ) − , (5.3)and where L o is the luminosity that would be measured by an observer in the AGNframe. The quantity n depends on the geometry and spectral index and is typically inthe range between 2 and 3.The Lorentz factor, γ , is given by γ = (1 − β ) − / . (5.4)Quasars or AGN with highly Doppler boosted relativistic jets pointed nearly along theline-of-sight are often referred to as “blazars.” Blazars are characterized by rapid fluxdensity variability, apparent superluminal motion, and strong x-ray and γ -ray emission.High resolution observations of blazars provide unique insight to the process by whichrelativistic jets are accelerated and collimated in the region close to the SMBH. We wantto understand: • How and where is the relativistic beam accelerated and collimated into narrow jets?Are there accelerations or decelerations? Do all parts of the jet move at the same speed? • What causes the curvature of jets? Does the flow follow a curved trajectory or is themotion ballistic and characteristic of a rotating nozzle? • Does the observed apparent velocity reflect the true bulk velocity of motion? Whatdetermines the jet velocity? Is the velocity related to other properties such as radio,optical, x, or γ -ray luminosity? • What is the maximum observed brightness temperature? Does it exceed the inverseCompton limit? • What is the energy production mechanism? • What can we learn from radio observations about the nature of the SMBH?Very Long Baseline observations made since 1971 have confirmed the apparent super-luminal motion expected from highly relativistic bulk motion. Since 1995, the NRAOVery Long Baseline Array (VLBA) has been used to study the motions of a large sam-ple of quasars and AGN at 7mm by a group from Boston University (Marscher 2012)and by the international MOJAVE group (Kellermann et al. 2004, Lister et al. 2009,Homan et al. 2009, Lister et al. 2013, Homan et al. 2014). More detailed informationmay be found on the respective web sites: adio Galaxies, AGN, and SMBHs
MOJAVE 2 cm program:
The results of these programs may be summarized as follows. • Radio loud quasars and AGN show highly relativistic bulk motion with a broaddistribution of apparent velocities. In general, the jets appear one sided, probably dueto differential Doppler boosting so that the approaching jet appears much brighter thanthe receding one. Each jet appears to have its own characteristic velocity but there is anappreciable spread in the apparent velocity of the different features within a given jet.The typical apparent velocity, β app ∼ β ∼ . β ∼
50 corresponds to an intrinsic value of β ∼ . • The jets with the fasted apparent velocities have the highest apparent luminosity,likely reflecting a correlation between intrinsic speed and intrinsic luminosity rather thansimply being the result of Doppler boosting (Cohen et al. 2007). • Apparent inward motions are uncommon and are likely the result of a feature movingoutward along a curved trajectory approaches the line-of-sight so that the apparentseparation from the jet base transverse to the line-of-sight appears to decrease withtime. • Individual jet features may show both apparent accelerations and decelerations. Boththe apaprent speed and direction of motion may change with time, but changes in speedare more common than changes in direction, indicating real changes in the Lorentz factoras features propagate down the jet. In general the apparent speed is greater further downthe jet, so that acceleration must take place at distances at least up to ∼
100 pc fromthe base of the jet (Homan et al. 2009, 2014). • Many jets show a curved structure and in a few cases there is evidence of an oscillatorybehavior. Sometimes the flow appears to follow pre-existing channels; other times the flowappears ballistic as from a rotating nozzle, perhaps due to precession possibly resultingfrom a binary black hole pair (Lister et al. 2013). • In some cases the direction of ejection appears to vary within a well defined coneforming what appears to be an edge brightened jet (Lister et al. 2013) such as shownby the jet in the nearby radio galaxy M87 where there is sufficient linear resolution toresolve the jet transverse to its structure (Kovalev et al. 2007). • There appears to be a relation between radio and gamma ray emission. There isstatistical evidence that radio outbursts follow a γ -ray event by ∼ γ -ray events (Pushkarev et al. 2010).
6. Brightness Temperature Issues
As described above, inverse Compton scattering limits the maximum observed bright-ness temperature, T < ∼ . . At the inverse Compton limit, the energy contained inrelativistic particles greatly exceeds that in the magnetic field which is perhaps not unrea-sonable in a very young source. If the particle and magnetic energies are in equilibrium,then the corresponding brightness temperature is only ∼ . .The observed brightness temperature may be calculated from (Kovalev et al. 2005) K. I. Kellermann T b = 2 ln kπ Sλ θ K = 1 . × S λ θ (1 + z ) K , (6.1)where k is the Boltzman constant, θ the angular size in milliarcsec, S is the flux densityin Janskys, and λ the wavelength in cm. The resolution of a radio interferometer, is givenby the ratio of the observing wavelength, to the interferometer baseline, D; or θ = λ /D.Putting this back into eqn. 6.1 gives T b = 80 SD (1 + z ) K , (6.2)so the maximum brightness temperature that may be measured depends only on the fluxdensity and baseline length, and is independent of wavelength. For ground based obser-vations with a maximum baseline of ∼ ,
000 km, the highest brightness temperatureswhich can be reached are ∼ K. Recent observations with the Russian RadioAstronspace VLBI satellite have suggested lower limits to brightness temperatures of 3C 273and other sources ∼ K (Kellermann et al. 2014, Kovalev 2014), or at least two tothree orders of magnitude greater than the limit set by inverse Compton cooling. Severalexplanations are possible (Kellermann et al. 2014).1. For a relativistically beamed source the apparent brightness temperature is boostedby a factor δ . To explain the high observed brightness temperatures in this way wouldrequire Doppler factors, δ ∼ to 10 . But typical observed values of δ ∼ γ ∼
10 withmaximum observed values ∼
50, and for 3C 273, γ ∼
15 (Lister et al. 2013). Possiblythe bulk flow which is related to the Doppler boosting might be much greater than thepattern flow observed by the VLBA, but Cohen et al. (2007) have shown that this isunlikely.2. The observed emission may be coherent such as observed in pulsars or the Sunincluding possible stimulated synchrotron emission.3) The radio emission might be the result of synchrotron emission from protons ratherthan electrons which would enhance the upper limit to the brightness temperature byabout the ratio of the proton to electron mass or more than a factor of 1000. However,proton synchrotron radiation would require a magnetic field strength more than 10 timesstronger than needed for electron synchrotron radiation of the same strength at the samewavelength.4) There may be a continuous acceleration of relativistic particles which balances theenergy losses due to inverse Compton cooling.
7. Summary and Issues
About ten percent of quasars and bright elliptical galaxies are strong radio sourceswith radio luminosity, P r > W/Hz and are thought to be driven by accretion ontoa SMBH. The weaker radio sources, P r < W/Hz are mostly due to star formationin the host galaxy. The observed properties of radio jets can be interpreted in terms ofa highly relativistic outflow from a central engine driven by a SMBH of up to 10 solarmasses. But, many questions remain. • Most quasars and AGN are not strong radio sources. Why are only ∼
10% of quasarsstrong radio sources, although all quasars presumably contain a SMBH to account fortheir extraordinary optical luminosity. • How do SMBHs generate relativistic jets? • How are the jets confined and shaped as they propagate away from the SMBH? adio Galaxies, AGN, and SMBHs • Why do only some jets produce γ -rays? How and where are the γ -rays produced?What is the relation between radio and γ -ray emission? • Is there evidence for binary black hole pairs? See the paper by Ekers (2015) in thisvolume. • If confirmed, observations of 3.3 GHz sky brightness combined with the density offaint sources detected in deep VLA observations suggest the possible existence of newpopulation of faint radio sources not due to star formation or to AGN and unrelated toany known galaxy population? • Are there other emission processes which play a role beside incoherent synchrotronradiation?
8. Acknowledgment
The National Radio Astronomy Observatory is operated by Associated Universities,Inc. under cooperative agreement with the National Science Foundation. I am indebtedto many colleagues, especially Ron Ekers and members of the MOJAVE team for numer-ous discussions that have contributed to this paper.
References
Baade, W., & Minkowski, R. 1954,
ApJ , 119, 206Bolton, J. G. 1960, paper presented to the URSI General Assembly, London, September 1960,
Observations of the Owens Valley Radio Observatory , 1960, No. 5Bolton, J. G. 1990,
Proceedings of the Astronomical Society of Australia , 8, 381Bolton, J. G., Stanley, G. J., & Slee, O. B. 1949,
Nature , 164, 101Burbidge, G. R. 1959,
ApJ , 129, 849Cohen, M. H., Lister, M. L., Homan, D. C., et al. 2007,
ApJ , 658, 232Condon, J. J. 1992,
ARAA , 30, pg. 575Condon, J. J., Cotton, W. D., Fomalont, E. B., et al. 2012,
ApJ , 758, 23Condon, J. J., Kellermann, K. I., Kimball, A. E., Ivezi´c, ˇZ., & Perley, R. A. 2013,
ApJ , 768, 37Dent, W. A. 1965,
Science , 148, 1458Ekers, R. D. 2015,
IAU Symposium 312, Star Clusters and Black Holes in Galaxies and AcrossCosmic Time
Fixsen, D. J., Kogut, A., Levin, S., et al. 2011,
ApJ , 734, 5Greenstein, J. L., 1962, manuscript submitted to the Astrophysical Journal, but subsequentlywithdrawnHoman, D. C., Kadler, M., Kellermann, K. I., et al. 2009,
ApJ , 706, 1253Homan, D. C., Lister, M. L., Kovalev, Y. Y., et al. 2014, arXiv:1410.8502Hoyle, F., Fowler, W. A., Burbidge, G. R., & Burbidge, E. M. 1964,
ApJ , 139, 909Jeans, J. H. 1961,
Astronomy and Cosmogony , New York: Dover, 1961, p. 360Kellermann, K. I., Lister, M. L., Homan, D. C., et al. 2004,
ApJ , 609, 539Kellermann, K. I., & RadioAstron AGN Early Science Team 2014,
American AstronomicalSociety Meeting Abstracts
ApJL , 739,L29Kinman, T. D. 1965,
ApJ , 142, 1241Kovalev, Y. Y., Kellermann, K. I., Lister, M. L., et al. 2005, AJ , 130, 2473Kovalev, Y. Y., Lister, M. L., Homan, D. C., & Kellermann, K. I. 2007, ApJL , 668, L27Kovalev, Y. 2014,
IAU Symposium 304 , pg. 78
Lister, M. L., Cohen, M. H., Homan, D. C., et al. 2009,
ApJ , 138, 1874Lister, M. L., Aller, M. F., Aller, H. D., et al. 2013, AJ , 146, 120Lynden-Bell, D. 1969, Nature , 223, 690Lynds, C. R., & Villere, G. 1965,
ApJ , 142, 1296Marscher, A. P. 2012,
International Journal of Modern Physics Conference Series , 8, 151Matthews, T. A., Bolton, J. G., Greenstein, J. G., Munch, G., and Sandage, A. H., 107th Meetingof the American Astronomical SocietyMinkowski, R. 1960,
ApJ , 132, 908Padovani, P., Miller, N., Kellermann, K. I., et al. 2011,
ApJ , 740, 20Padovani, P., Bonzini, M., Miller, N., et al. 2014,
IAU Symposium , 304, pg. 79Pushkarev, A. B., Kovalev, Y. Y., & Lister, M. L. 2010,
ApJL , 722, L7Seyfert, C. K. 1943,
ApJ , 97, 28Sholomitskii, G. B. 1965,
Soviet Astronomy , 9, 516Zwicky, F. 1965,
ApJ , 142, 1293Zwicky, F., & Zwicky, M. A. 1971, Guemligen: Zwicky,
Catalogue of Selected Compact Galaxiesand of Post-Eruptive Galaxies pg.pg.