GALAXIES, INFRARED EMISSION BARUCH T. SOIFER Infrared emission from galaxies is comprised of three major components. The first is emission from stellar photospheres that peaks at 1-3 **, and dominates the energy output of most galaxies. Emission lines due to fine structure transitions in atomic and ionic constituents of interstellar gas contribute measurable emission at some infrared wavelengths as well, as do molecular rotational lines, and in some cases molecular vibration-rotation lines are seen. Such lines can contribute as much as a few percent of the total infrared luminosity of galaxies. By far the dominant component of infrared emission from galaxies at wavelengths longer than 3 ** is that due to thermal radiation by dust in the wide variety of environments found in galaxies. Mechanisms for producing infrared emission in active galaxies and quasistellar objects are discussed in another entry. Dust is a ubiquitous component of the matter surrounding the stars and constituting the interstellar medium. This dust absorbs the photons radiated at shorter wavelengths by stars (or other energy sources), and reradiates this energy in the infrared. The wavelengths at which this energy is radiated are dependent on the environment in which the dust is found. In a steady state the dust is at a temperature where it radiates an amount of energy equal to the amount of energy it absorbs at shorter wavelengths. This temperature ranges from **1000 K for dust in circumstellar environments, to <20 K for dust in interstellar space, with corresponding wavelengths of peak emission ranging from *3 ** for the hot dust to >150 ** for the coldest dust. The dust grains that radiate via this steady-state emission are comparatively large, with particle radii *0.1 **. There is also a significant fraction of interstellar dust composed of much smaller grains, with radii less than 0.001 **, that contribute significantly to the output of infrared radiation from galaxies. This smaller dust does not radiate in a steady state, but rather each individual photon that is absorbed (these particles can be thought of equally well as large molecules) heats the absorbing grain to a substantial temperature of many hundreds of degrees for a very short time. The dust at such a high temperature radiates energy very effectively, and cools rapidly. This non-steady-state emission process is important to the production of emission at the shorter infrared wavelengths (i.e., ********). Almost from the beginning of infrared astronomy (mid 1960s), galaxies have been known to produce copious amounts of infrared radiation, well in excess of that expected from the stellar photospheres in these galaxies. Two of the galaxies that are brightest in the infrared, the nearby irregular galaxy, M82, and the closest Seyfert 2 galaxy, NGC 1068, were discovered to be bright infrared sources, emitting far more luminosity at infrared wavelengths than in the visible, almost as soon as infrared observations began. Such galaxies were observed from ground-based and airborne telescopes over the entire infrared spectrum. Nearly all such galaxies that were found to be bright infrared sources were chosen based on peculiar properties at other wavelengths. It was not until the launch of the Infrared Astronomical Satellite (IRAS) in 1983, with its program of performing an unbiased all-sky survey at 12, 25, 60, and 100 **, that astronomers were able to assess the significance of infrared emission in the energy budget of galaxies. By conducting a survey of objects based on their infrared brightness, astronomers were able, for the first time, to make an unbiased evaluation of the fraction of galaxies that are bright at infrared wavelengths, and thereby to determine the fraction of the total luminosity of galaxies that is emitted in the infrared. Based on this survey it was found that infrared emission compromises *20% of the radiant energy in the local universe (i.e., the universe within a radius of *100 Mpc or 3x10* light years). At the highest luminosities, the infrared bright galaxies represent the most populous constituent of the local universe, having a higher space density than quasars at the same total power. It was found that the kinds of galaxies that are bright in the infrared are very dusty galaxies, typically spiral galaxies. Three local galaxies, the Milky Way, the great nebula in Andromeda (M31), and the nearest external spiral galaxy, M33, represent three different kinds of galaxies based on their infrared brightness. M31 is a comparatively dust-free galaxy and less than 10% of its total luminosity emerges in the infrared. M33 is a face-on spiral galaxy whose infrared luminosity is *25% of the total energy output. Finally, the Milky Way is a comparatively bright galaxy in the infrared, radiating roughly half its total luminosity in the infrared. Figure 1 illustrates the energy distributions of a variety of galaxies from the radio (log v=9) through to the visible (log v=15), including the infrared. The infrared luminosity ranges from a few percent of the total stellar luminosity in the central 4' of M31 to roughly 98% of the total energy output of the galaxy Arp 220. The environments that produce the infrared radiation are found to depend on the level of infrared radiation being produced in the galaxy. In quiescent galaxies like M31, the infrared emission is dominated by the interstellar dust absorbing starlight from the ambient radiation field and reemitting it in the infrared. In galaxies like M33 or the Milky Way the vast majority of the infrared emission is emitted by dust in the environments of giant molecular cloud complexes, where many massive, high-luminosity stars are currently forming or have recently been formed. These complexes are still very dusty, and the dust is very efficient at absorbing the ultraviolet radiation emitted by the hot, massive, young stars that dominate the luminosity output of these regions. This absorbed radiation is then reemitted in the infrared. In this way the amount of infrared luminosity from such galaxies is a measure of the total luminosity in recently formed stars, and by assuming a relation between the luminosity and mass of such stars, the rate at which stars have formed over the last -10* years can be calculated. The amount of infrared radiation emitted by galaxies depends to a great extent on the amount of dust that they contain. This in turn corresponds to the amount of interstellar gas, since typically there is 100-200 times more gas than dust in interstellar material. For a galaxy like M31, the amount of dust responsible for the far-infrared radiation is roughly 10* M* (1 M* is the mass of the Sun or 2x10** g), corresponding to roughly 10* M* of interstellar gas. In a galaxy like the Milky Way, approximately 10* M* of dust produce the infrared emission, with a corresponding amount of gas of 10* M*. In the most luminous infrared galaxies, there is more than an order of magnitude more gas and dust than is found in the Milky Way. Because infrared emission from galaxies is a process that requires the presence of both dust and higher-energy photons to heat the dust, the infrared emission extends only as far as the optical limits of the galaxies, with a decreasing brightness that is roughly equivalent to that seen in the starlight from the galaxies. In detail it appears that the infrared emission front galaxies decreases somewhat faster than the starlight in the galaxies. This is consistent with the fact that there are less "metals" (elements heavier than helium, which are needed to form dust grains) in the outer parts of galaxies than in the inner regions. In infrared-bright galaxies of high luminosity it appears that much of the emission is generated in the central regions of the galaxy, within 1-2 kpc from the center (as compared to a typical radius for a spiral galaxy of 10 kpc). A process that has been shown by IRAS observations to be important in causing production of infrared bright galaxies is the interaction of two gas-rich galaxies. Such interacting galaxies are identified by such visible phenomena as overlapping disks, multiple nuclei, gravitationally induced tails of stars streaming out from the body of the galaxy, etc. Interacting galaxies comprise an increasing fraction of the infrared bright galaxies as the luminosity of the galaxies increases. Indeed, among galaxies with infrared luminosities >10** L*, or 10 times that of the Milky Way (1 L* is the total luminosity of the Sun, or -4x10** erg s**), interacting systems are the majority. This shows that the process of galaxy interaction/merging is an important one for triggering very luminous galaxies in the infrared. It is expected that in such colliding systems the molecular gas is in some way compressed, either by global compression associated with the interaction of the interstellar media of the two galaxies, or perhaps via more direct cloud-cloud collisions. This compressed molecular gas forms stars quite rapidly and thereby produces the strong infrared emission seen in such galaxies. The most luminous galaxies found in the infrared emit 100-1000 times more luminosity in the infrared than the Milky Way, or 10**-10** L*. This is as much power as is emitted by the most luminous objects in the universe, the quasars. The objects that emit this amount of infrared radiation are all quite peculiar, being found in massive, strongly interacting/merging galaxy systems. Because all these galaxies are strongly interacting, it is believed that the interaction has triggered the process that produces the luminosity. Young, massive stars are believed to power a significant fraction of the luminosity in these galaxies. It is also possible that the bulk of the luminosity from these galaxies is not powered by stars, but rather by quasars enshrouded in so much dust that their optical signatures are hidden from view, and the major signature of the quasar, is the bolometric luminosity that emerges in the infrared. This picture, if correct, provides an explanation as to how quasars form, in collisions of two gas-rich normal galaxies. The subsequent evolution of the nucleus of the merged galaxy would disrupt the cloud in which it forms, and lead to the appearance of a visible quasar. Elliptical galaxies, previously thought to be virtually devoid of dust, have been shown to have a significant amount of interstellar dust, via the detection of infrared emission in excess of that expected from the stars in these systems. IRAS observations of these galaxies have shown that there is 10*-10** M* of dust in many of these galaxies. The source of the interstellar dust is believed to be material from old stars in the later stages of evolution. These stars are ejecting their outer envelopes in the form of gas; dust then condenses in the outflowing material. Typically in these galaxies it takes roughly 10* yr for the evolving stars to eject as much dust as is observed via their far-infrared emission. This in turn requires a mechanism for removing this interstellar material, since the time to accumulate it is only 1% of the age of the galaxies. One possible removal mechanism is the formation of new stars. Additional Reading Lonsdale, C.J., ed.(1987). Star Formation in Galaxies (NASA Conference Publication No. 2466). U.S. Government Printing Office, Washington, DC. Rieke, G.H. and Lebofsky, M.(1979). Infrared emission of extragalactic sources. Ann. Rev. Astron. Ap. 17 477. Soifer, B.T., Beichman, C.A., and Sanders, D.B.(1989). An infrared view of the universe. American Scientist 77 46. Soifer, B.T., Houck, J.R., and Neugebauer, G.(1987). The IRAS view of the extragalactic sky. Ann. Rev. Astron. Ap. 25 187. Telesco, C.M.(1989). Enhanced star formation and infrared emission in the centers of galaxies. Ann. Rev. Astron. Ap. 26 343.