Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.
Ultraluminous Infrared Garages B.T. SOIFER California Institute of Technology ABSTRACT The IRAS all-sk r survey has provided astronomers with a first deep New of the sky in the "thermal infrared", i.e., from 10 microns to 100 microns. One of the major discoveries of this survey has been a population of sources having the bolometric luminosities of quasars, but where more than 90% of the luminosity emerges in the infrared. These objects, more numerous than quasars, are found exclusively in interact~ng/merging galaxies that are extremely rich in interstellar gas. We have accumulated evidence that suggests that these systems are indeed quasars obscured by many tens of magnitudes of extinction. We have suggested that these Ultraluminous Infrared Galaxies are the formation stage of quasars, and that colliding galaxies, ultraluminous infrared galaxies, and quasars might all be linked through an evolutionary sequence where the infrared bright phase is one in which the quasar Is formed in the nucleus of a merger system, and is enshrouded in gas and dust, while the UV excess quasars are the end state of quasar evolution where most of the enveloping dust cloud has been dissipated, and the quasar is visible directly. INTRODUCTION Viewing We universe through a new portion of the electromagnetic spectrum has always lead astronomers to major discovenes. Quasars and neutron stars are just two examples of discoveries made as a result of the Radio and X-ray sky surveys of the 1950's, 1960's, and 1970's. In 1983 the 344
HIGH-ENER~ ~TROP~ICS 345 Infrared Astronomical Satellite (IRAS) performed the latest of these all sky surveys that has lead to major discoveries. Before IRAS, our view of the infrared universe was limited to sly surveys at 2.2 microns (Neugebauer and Leighton 1969) and 4-30 microns (Pnce and Walker 1976) as well as studying objects found by other means. The infrared sky surveys vastly increased our understanding of the Galaxy and its constituents, but lacked the sensitivity to go significantly beyond the Galaxy. The Caltech 2 micron sky survey contained one external galaxy, M31, while the AFGL sly survey contained a handful of extragalactic objects. The IRAN sky survey, with its 3 orders of magnitude improvement in sensitivity over previous surveys, and extension to 60 microns and 100 microns, drastically increased the numbers of galaxies detected purely by their infrared emission, finding ~ 0.5 galaxies/square degree, and altered our understanding of the extragalactic sly, permitting an unbiased view of the local universe in the infrared. TF~ WAS SKY SURVEY The IRAS all sly survey was performed at 12 microns, 25 microns, 60 microns, and 100 microns using a 57 cm telescope and focal plane entirely cooled to 2.7K with superfluid helium. The telescope was contained in the toroidal shaped, 700 liter capacity helium cryostat. The satellite was launched on January 25, 1983, and collected data for the 300 days that the superfluid helium lasted. The primary scientific goal of the IRAS mission was to perform the all sly survey, and approximately 60% of the satellite's time was devoted to this purpose. 1b produce a highly reliable census of the inerdally fixed sources in the presence of many "local" contaminants such as cosmic ray hits on the detectors, dust particles crossing the field of view of the telescope, asteroids and comets, the telescope scanned the sky 6 tunes over the 300 day mission. The multiple sightings of particular sources were used to filter and separate the inertially fixed sources from the "moving sources" such as asteroids and comets, and the transient events local to the telescope environment. These data were combined to produce the first TRAS Point Source Catalog which covered 96% of the sly. At high galactic latitude the IRAS Point Source Catalog was complete to 0.4, 0.5, 0.6, arid 1.5 Jy at 12 microns, 25 microns, 60 microns, and 100 microns respectively with reliability > 99.8%. Recently the same data have been reprocessed to produce a catalog at high galactic latitudes (the IRAS Faint Source Catalog) that is a factor of ~ 2.5 more sensitive at all wavelengths, at the cost of reduced reliability (> 98%). The IRAS survey has a fairly modest sensivity compared to modern optical surreys. In flux per octave, i.e., uS,~, the IRAS Point Source Catalog
346 AMERICAN AND SOVIET PERSPECTIVES reached 3 x 10-14 W/m2 at 60 microns, equal to the flux per octave of a 15.0 mag object at B. In terms of detectable luminosity, extragalactic sources seen by IRAS are comparatively local. Sources with the infrared luminosity of the Milly Way are detected to a redshift z ~ 0.03, while sources with the bolometric luminosity of a quasar, where this energy emerges in the infrared, are detected to redshifts of z ~ 0.3 (in this paper, we adopt Ho = 75 ~n/s/Mpc and QO = 0~. Thus the IRAS survey is indeed a survey of the local universe. The combination of the sensitivity of the IRAS survey, and the wavelength at which the infrared luminosity emerges from galaxies results in the vast majored of the ~ 20,000 extragalactic objects discovered in the IRAS Point Source Catalog being found at 60 microns. Since the IRAS survey was done at wavelengths substantially longward of the peak in stellar photospheres, those sources seen in the infrared radiate via mechanisms substantially different from those found to be bright in optical suIveys. The predominant mechanism responsible for the infrared flow from galaxies is thermal emission from dust; the dust absorbs and reradiates energy originally emitted at shorter wavelengths. Infrared bright galaxies are quite dusW and therefore have either been missed, or appeared innocuously in visible surveys. TlIE LOCAL UNIVERSE IN THE INFRARED With an unbiased view of the infrared sly available for the first time, several groups have investigated the local luminosity function of infrared bright galaxies using IRAS data (Rieke and Lebofsky 1986; Soifer et al. 1986, 1987; Lawrence et al. 1987; Smith et al. 1987~. All these analyses agree remarkably well, and have demonstrated that the space density of infrared bright galaxies does not vary significantly over a factor 3 in distance (from a median redshift of < cz > ~ 2,200 Km/s in Soifer, et al. 1987; to < cz > ~ 7,000 Km/s in Lawrence e' al. 1987~. 1b understand the importance of such infrared bright galaxies requires placing this class of objects in the context of other lmown classes of extragalactic objects. We have done this by comparing the bolometnc luminosity functions of the major classes of extragalactic objects (Soifer et al. 1987~. Since most luminosity functions for extragalactic objects are derived at B (4400 A), it was necessary to apply bolometric corrections to most such luminositr functions, to place them in the same units. This comparison is shown in Figure 1. The luminosity functions of these sources immediately shows that infrared bright gal~es are an important, but not dominant source of lumi- nosi~ in the local universe. For Loot < 2 x 10~° Go, the infrared bright galaxies have a density about 20% of that for galaxies found in the vis~- ble. For L`'o' > 2 x 10~° L:>, infrared bright galaxies become increasingly
HIGH-ENERGY ASTROP~SICS 10° 10 lo-2 _~ 10-3 _ 10-4 _ . . - o D ~ 10-5 ~ Cal to -,.\\ · IRAS GA LAX I ES - ALL ° IRAS GALAXIES - NON-VIRGO -NORMAL GALAXIES X STARBURST GALAXIES ~ SEYFERT GALAX I ES ° QUASARS $ x ~ Y~\ 10-7 lo-8 lo-l9o! 109 1010 loll LBOL [Le] t ~ 1 lol2 lol3 347 FIGURE 1 The bolometric luminosity functions of the major class" of extragalactic emitted (taken from Soifer et al. 1987). Separate bolometric corrections have been applied for each class of object. The infrared galaxies become the dominant extragalactic population at Lbol > 3 X 1011 Lain.
348 AMERICAN AND SOVIET PERSPECT~ES important in the local universe, since the space density of normal galaxies drops exponential with luminosity, while the infrared bright galaxies ap- pear to decrease as ~ L-2. At Lbol > 3 x 1011 Lo, the infrared bright galaxies become the dominant population in the local universe, exceeding in space density the quasars. Overall the infrared luminosity of galaxies represents ~ 25% of that emerging in stellar photospheres, and 60-80~o of this infrared luminosity is due to young, massive stars (Soifer et al. 1987~. ULTRALUMINOUS INFRARED GATES We have studied the ~ 300 brightest Intragalactic sources at 60 microns in 14,500 square degrees at high galactic latitude. This IRAS Bright Galaxy Sample has been most amenable to study at other wavelengths, since it represents the brightest of the infrared luminous galaxies. Of these IRAS Bnght Galaxies 10 have bolomeuic luminosities equivalent lo those of quasars. We have studied these 10 objects in great detail (Sanders et al. 1988a), and have found them to possess remarkable properties. The observed properties of these objects are summarized in Able 1. Four of these Ultraluminous Infrared Galaxies were previously cata- loged in other surveys. To of them, ~ 231 and Mk 273, were found in the Markarian survey of UV excess gal~es, and were known to be peculiar objects. Markan~an 231 has been lmown for nearly two decades as an extraordinarily luminous infrared galaxy (Rieke and Low 1972; Young et al. 1972~. Arp 220 (IC4553) has also been identified as a galaxy with very peculiar properties. As can be seen immediately from Table 1, all of the Ultraluminous Infrared Galaxies are rather faint optical galaxies. The ratio of infrared to visible luminosity is quite large for the ultraluminous galaxies, ranging from 50-150, as compared to 0.3 for "typical" spirals or 1-5 for "apical" infrared selected galaxies (Soifer et al. 1987). In the infrared these Ultraluminous Infrared Gal~es are unresolved at the IRAS angular resolution of ~ 1 '. At optical wavelengths, CCD images show ALL these systems to be galaxies undergoing strong interac- tions/mergers. Contour plots of CCD images of these galaxies (Sanders et al. l9~a) are shown in Figure ~ All of these objects show evidence for recent or ongoing mergers, i.e., multiple nuclei, tidal tails, strongly distorted disks, eta Optical spectroscopy shows that all of these galaxies have strong emis- sion lines. Three of the systems, MK 231, UG{: 5101, and IRAS 051~25, have Ha line widths > 2,000 Km/s (FWZI). Mk 231 is a well known Seyfen 1 system, while U5101 and IRAS 051~25 would be classified as interme- diate Seyfert nuclei based on their optical spectra. Of the remaining 7 Ultraluminous systems, 6 show line widths > 1,000 Km/s, and line ratios (tOIII]/H,B, iSII]/Hc~, [NII]/Hc~, [OIl/Hc') characteristic of AGN spectra:
349 ~~' :: ~O u2 rid V, At: A! U) 3 _( O [,I] Z - o V) Lo o _ Q _ o Ct .~ _] ~ , o L. r . I_ S ~ O Z ~ V ~ O ~ ~ 0 00 ~ ~ ~O Cat Cut ~ ~ ~ ~ ~ ~t- en) _4 O ~ ~ ~ 00 ~ ~ ~ ID - ~ ~ ~ ° cry ° ° ° ° ° ° ° ~ 4 ~ _4 ~4 ~ ~ ~ of 0 ~ ~ rat 0 oo an 00 ~ 00 0 _4 _ _ ~ ~ _ _ _ ~ _ ~ oo C~ ~ 0` ax ~ ~ 0 ~ · . . ~ . . ~ . . - ~ ~ ~ So ~ ~ ~o ~ ~ ~ - - - - t - - - - - 4 ~ oN O O ~ tn ~4 ~ 0 ~ _ ~ ~ ~ ~ ~ ~ ~ ~ ~ c~ _ ~ _ _ _ _I ~ t_ _ _ ~ o o oo o ~ ~ - oo o oo ~ ~ ~ oo ~ o ~ ~ ~ ~ ~ ~ ~ ~ ~ - ~ - - - ~ - - ~ - ~ c ~o o ~ ~ - ~ ~ ~; ~; ~ o o u) c~1 ~ ~ ~ + + + ~ + ~ l oo ~ ~ ~ ~ ~ ~ ~ ~ a~ _ ~ ~ ~4 u) ~ c~ c~ ~ ~ ~ ~ ~ ~ 0 0 0 ~ _ _ _ ~ _ ~
350 my, ~IERICAN AND SOVIET PERSPECTIVES 4c 2C -4C . ~ 20 , 6C TIC 2C o -2C Mrk 273 ~ 3 . . UGC OS101 JO . ,. ;: 0. ,4' " To JO ~0 -= to FIGURE 2 Contour maps of optical images of the 10 Ultraluminous Infrared Galaxies found in the IRAS Bnght Galaxy Sample (from Sanders et al. 198Sa). Note the irreg~ shapes, tidal mils and double nuclei; this is clear evidence of mergers in all these systems. Only IRAS 2249-18 is classified as an HII region spectrum based on the above line ratio. PhotometIy of the nuclear regions of these galaxies reveals that the near infrared colors of these galaxies show a very large spread, spanning the entire range of colors from slightly reddened spiral gal~es to highly reddened quasars. As will be discussed later there is enough dust (As > 100 may) obscuring the underlying energy source that it is unclear whether these observations are a meaningful probe of the nuclear regions. The overall energy distn~utions of these objects are shown in Figure 3. All of the objects plotted here have very large ratios of infrared to visible luminosity, immediately demonstrating why they are found in 60 micron selected samples. All of these systems have similar ratios of infrared to radio continuum emission, with log [fir/uS~ (1.49Ghz)] ~ 6.5, so these are indeed
HIGH-ENERGY ASTROP~ICS 351 "radio quiet" sources. The energy distributions are ordered by increasing S~60 microns)/S~100 microns) ratio. Into systematic trends emerge from this plot. The fraction of energy emerging at shorter infrared wavelengths increases with increasing S`~60 microns)/S`, (100 microns) ratio, while the separation between cold dust and near infrared components becomes less obvious in the 'warmer" objects. The eno~ous infrared luminosities, coupled with the fact that the emission is clearly thermal emission by dust, implies that these are gas-nch systems. The amount of dust required to produce the observed infrared luminosity is Mauri > lO8M,3, so we would expect > 10~°M~> of gas lo accompany this dust. Surpnsingly, the vast masons of this gas is in molecular for, with 9 of the Ultraluminous Galaxies baring been detected in the 2.6mm J = 1-0 line of CO (table 1~. Even more surprising has been the finding that a large fraction, as much as 70-80% of this gas is confined to the central few Kpc of the galaxy (Scoville et al. 1986; Scoville e' al. 1989~. COLLIDING GALAXIES FORMING QUASARS? Based on the observations of the Ultraluminous Infrared galaxies found in the IRAS Bright Galaxy sample described above, we have suggested (Sanders et al. 1988a) that these systems are heavily dust enshrouded quasars. The evidence for this comes from the luminosities, emission lines, near infrared colors, and luminosity to gas mass ratios. We believe that a quasar is powering the vast luminosity emerging from these objects, however the quasar is so heavily enshrouded in dust that we new the quasar primarily through its bolometric luminosity. An alternate view of these Ultraluminous Infrared Galaxies, as systems undergoing "mega starbursts", has been presented by several groups (e.g., Rieke e! al. 1985~. The fact that all of these systems are found ~ merging, gas-rich systems has lead us to suggest that the merger of two gas-rich galaxies is fundamental to the process of quasar formation. We believe that when two gas-rich galaxies interact in a near direct collision, the nuclei merge rapidly (lbomre and lbomre 1972), so disrupting the angular momentum and gravitational field of the system that the gas is funneled rapidly into the nuclear region of the merged system. This rapid accumulation of gas in the central environment of the galaxy will lead inevitably to an enormous "starburst." Enough gas accumulates in the central Kiloparsec of the system that the gas becomes self-gravitating, triggering further collapse of the gas. If a massive black hole exists In the center of the galaxy, as suggested by much current observational data (e.g., Filippenko and Sargent 1985; Dressier and Richstone 1988), this rapid funneling of interstellar gas into the central environment of the galaxy is exactly the source needed to trigger
352 AMERICAN AND SOVIET PERSPECTIVES Am 220 IR 1211~03 I R 2249 -18 - ~n V) a (z< 0.081) IR OS18-2S Ark =1 TR 08S7~39 ~Ultroluminous IRAS Goloxies M,` 273 1 ~ \ ·. · ~ \; '" ma; it_ \\ ..:---- .----'~ IR 1525~36 / ~ ~· . · \ \ \ ·. UGC Oslo' ~\\ - .. - \ ! ... is \ ' ... '''I \ .. .l me, N O'er , . . %~ 1 1 ~ t04 3000 t00 10 A (Jim) - .~ xlo 2 1 ~1 1 0.1 FIGIJPE 3 The energy distributions form 0.44 microns to 350 microns for the Ult~lumi- nous Infrared Galaxies (from Sanders et al. 1988a). I-ne energy distributions are ordered from lowest 60 micronsf100 microns color temperature (top) to highest 60 micronsllOO microns color temperature (bottom).
HIGH-ENERGY ASTROPHYSICS 353 formation of the quasar that is powered by gravitational energy released as the gas falls onto the black hole. We expect that the formation of a quasar embedded in 10~°M~ of gas and dust in the center of a galaxy is an enormously disruptive event, one that will ultimately result in a visible quasar. Qualitatively, we expect that as the energy release of the quasar begins to disrupt the surrounding gas and dust, the quasar will take on more and more of the properties of a "normal" quasar, be. revealing a strong nuclear opticallUV/near infrared continuum, very broad emission lines, etc. We have found evidence (Sanders et al. 1988b) for such an evolutionary sequence in the ``warm" Ultraluminous galaxies found in the IRAS database as extragalactic objects with luminosi- ties of quasars, but with infrared properties intermediate between those of the Ultraluminous Infrared Galaxies and the "normal" quasars. These systems, again selected by purely infrared flux density and color criteria, show equal numbers of Seyfert 1 and Seyfert 2 nuclei. Of these Warm Ultraluminous Infrared Galaxies, 9/12 are in strongly interacting/merging systems, and 3 are indeed classified as "normal" quasars by the criteria of Schmidt and Green (1983) and Veron-Cetty and Veron (1985~. As we propose this evolutionary scenario, ultimately the disruption of the surrounding medium is complete and the quasar emerges from its cocoon as a UV excess quasar. The residual dust and gas in the environment still produces a significant, but not dominant, fraction of infrared luminosity. Recent~, we have studied in detail the energy distributions of the brightest UV excess quasars (Sanders et al. 1989), and found that they indeed emit approximately 20% of their bolometric luminosity in the infrared, consistent with this picture, while the infrared selected quasars studied by Low et al. (1989) might represent a slightly earlier phase in this sequence. The ideas for the formation of quasars that we have suggested as a result of our study of the Ultraluminous Infrared Galaxies are not new, but rather bring new data to support rather old ideas. Alar and Juri lbomre (1972), in a seminal work, outlined most of these ideas for galaxy interaction triggering active nuclei while modern, high dynamic range imaging and spectroscopy of quasars and their environments have lead many investigators to the conclusion that interactions play a significant role in quasars (Stockton and McKenty 1983; Hutchings e! al. 1984~. What we believe the IRAS data has provided is evidence for the formative stages of the quasar, where the gas and dust that converts the luminosity of the quasar into a bright infrared source effectively acts as a strong neutral density alter to allow us to study in detail the environment surrounding the quasar without the blinding effect of the exceedingly bright central source. Recent theoretical work (Norman 1989) suggests that the scenario we have described here is indeed a plausible outcome of a direct collision between two gas-rich gal~es.
354 AMERICAN AND SOVIET PERSPECTIVES The model we envision for quasar formation has the exceedingly attractive feature that it ties quasars, the most luminous and presumably violent objects in the universe to the rare occurrence of a direct merger of two gas-nch spiral galaxies. This model very nicely provides a physical process to explain the well known strong evolution of quasars, both in their rapid increase with redshift, and apparent CUtOD at high redstart. The former is simply related to the higher rate of collisions of galaxies in the past, as a result of their closer proximity and their higher mean gas content, while the latter Is a result of the fact that the galaxies must form and interact before a quasar can be formed. Clearly, for this model to work the ultraluminous infrared galaxies must follow the evolution found for the optically selected quasars (see e.g., Schmidt and Green 1983~. This prediction is directly testable with the next generation of space infrared obsenatones, ISO and SIRTF. CONCLUSIONS The new view of the universe provided by the IRAS survey has lead to the discovery of a new class of objects, the Ultraluminous Infrared Galaxies. We believe that these objects are the first formative stages of quasars in the nuclei of merging gas rich spiral galaxies. Such an explanation naturally ties the formation of quasars to a violent, but rare event in the evolution of spiral galaxies. ACKNOWLEDGEMENTS Many colleagues have contoured in major ways to the development Of the ideas presented here Most notably these are Dave Sanders, Gerry Neugebauer, Jay Elias, and Nick Scoville. In addition, Keith Matthews, Dave Carnco, and James Graham have made major contributions to obser- vations that have helped formulate these ideas. Finally, none of this work would have been done without the tremendously successful IRAS mission, and it is a great pleasure to acknowledge the many dedicated engineers, scientists, technicians, and managers who helped make this project the success it has been. This research has been supported by NASA through the TRAS B- tended Mission, and by the NSF. RE":RENCES Dressler, A, and D. Richstone. l9SS. Ap. J. 324: 701. F~lippenko, An, and W.LW. Sargent. 1985. Ap. J. Suppl. 57:503. Hutchings, J.B., D. Crampton, and B. Campbell. 1984. Ap. J. 280: 41. Lawrence, An, D. Balker, M. Rowan-Robinson, KJ. Leech, and
HIGlI- ENERGY AS TROP~ICS 355 M.V. Penston. 1986. M.N.R^S. 21~. 687. Low, F.J., RM. Cutn, S.G. Kleinmann, and J.P. Huchra. 1989. Ap. J. Letters 340: L1. Neugebauer, G., and R.B. Leighton. 1969. ~o Micron Sly Surv~y, A Preliminary Catalog, NASA Pub SP-3047. Norman, C 1989. Proceedings of Conference W~ndows on Gala~nes, in preparation. Pnce, S.D., and R.G. UBalker. 1976. The AFGL Four Color Infrared Sly Surve~. Catalog of Observations at 4.2, 11.0, 19.8, and 27 microns, AFGL Pub AFGL,TR 0208. Rieke, G.H., R.M. Cutri, J.H. Black, W.F. Kailey, C.W. McAlay, MJ. Lebofsky, and R. Elston. 1985. Ap. J. 290: 116. Rieke, G.H., and M. Lebofsly. 1986. Ap. J. 304: 326. Rieke, G.H., and FJ. Low. 1972. Ap. J. Letters 176: L95. Sanders, D.B., B.T. Soifer, J.H. Elias, B.F. Madore, K Matthews, G. Neugebauer, and N.Z. Scov~lle. 1988a. Ap. J. 325: 74. Sanders, D.B., B.T Soifer, J.H. Elias, G. Neugebauer, and K. Matthews. 1988b. Ap. J. Letters 328: L35. Schmidt, M., and RF. Green. 1983. Ap. J. 269 352. Scoville, N.Z., D.B. Sanders, AI. Sargent, B.l: Soifer, S.L~ Scott, and KY. Lo. 1986. Ap. J. Lette~s 311: L47. Scoville, N.Z., D.B. Sanders, AI. Sargent, B.T. Soifer, and CG. Tinney. 1989. Ap. J. Letter~ submitted. Smith, B.J., S.G. Kleinmann, J.P. Huchra, and Low, FJ. 1987. Ap. J. Soifer, B.T, D.B. Sanders, G. Neugebauer, G.E. Danielson, CJ. Lonsdale, B.F. Madore, and S.E. Persson. 1986. Ap. J. Letten 303: IA1. Soifer, B.T et al. 1987. Ap. J. 320: 238. Stockton, ~, and J.W. MacKenty. 1983. Nature 305: 678. Toomre, ~ and J. Toomre. 197Z Ap. J. 178: 623. Veron-Cetti, M.-P., and P. Veron. 1985. Eso Sci. Rpt., No 4. Young, E.l:, R.~. Knacke, and R.R. Joyce. 1972. Nature 238: 263.
