Dear Readers,
Welcome to the October 19 edition of the MIPS/IRAC GTO newsletter. After a one-issue hiatus, George Rieke has returned with two new contributions to the newsletter. The first is a lengthy white paper outlining how we could spend all of our GTO time measuring the spectral energy distributions of quasars. George also submitted a brief MIPS perspective on Michael Pahre's program to study normal galaxies at 0 < Z < 1.
For those of you attending the SIRTF science working group meeting in mid-November, just a gentle reminder that the November 2 issue of the newsletter is your last chance to post your white papers prior to the meeting. I will be finishing radiation testing of the MIPS detectors at UC Davis that weekend, so the deadline for submissions is effectively Saturday, October 31st.
Doug Kelly, editor (dkelly@as.arizona.edu)
In his white paper on galaxy SEDs and evolution for 0 < z < 1, Michael asked what plans MIPS had for observations of `normal' galaxies in this range, possibly as a foundation for the program at higher redshifts.
MIPS has only thought about this issue a little bit, so far as I know. One specific program we will probably carry out is to determine the far infrared properties of low metallicity galaxies (less than 3% solar, say). Only a few such objects are known locally, but it seems important to us to get good data on them to establish some understanding of how the far infrared properties may be affected at large redshift by the possibly lower metallicity.
The broader issue of establishing a suitable database at intermediate redshifts depends for us on a couple of factors. We need to understand what ISO has accomplished in this area. Also, we need to look at how many intermediate redshift galaxies we would find in our deep fields, a total of about one square degree. We are a bit nervous about devoting a lot of time to surveying large areas of sky, though, because it is just the kind of program that would be difficult to protect from pre-emption by a Legacy team that surveyed a larger area somewhere else.
We should get a good overview of the ISO results at the conference October 20-23. We are just defining the regions we would survey deeply, and we will then be able to estimate the number of galaxies at intermediate redshift. We should keep communicating, to establish the best coordination given these boundary conditions.
Meanwhile, I encourage the IRAC team to think about observing low metallicity galaxies to help interpret the Early Universe data they hope to obtain.
To TopIn this white paper, I describe a number of investigations centered around measurement of quasar spectral energy distributions with MIPS. A number of these programs also involve data obtained with IRAC and/or IRS. More work is needed to refine the objectives. In addition, it is possible to save observing time by carefully coordinating the samples of quasars observed in a way that a single sample can be used to address a number of questions. Although the time estimates will vary with additional objectives and further coordination, for rough planning purposes I have generated lists of targets and very crude observation times, finding that about 200 hours could easily be devoted to these objectives.
The broad and bright spectral energy distributions of quasars, extending from the radio through the x-ray and gamma ray, is one of their unique features. Other astronomical sources tend to be prominent over only a portion of the electromagnetic spectrum. High sensitivities achieved in the radio, near infrared, optical, ultraviolet, and x-ray allow probing the quasar SED in complete samples containing large numbers of objects, leading to robust conclusions that are free of selection biases. However, the far infrared SEDs have only been probed extensively by IRAS and ISO to similar detection levels of 0.1 to 0.5 Jy, adequate to reach significant samples but not to detect all members of samples comparable with those that can be studied in the other spectral regions. The high sensitivity of MIPS in the far infrared offers an opportunity to advance our understanding of quasar SEDs to a level comparable with what has been achieved over the rest of their emission range.
There are two general classes of investigation possible for SIRTF. It has become popular during the past decade to try to relate differing types of AGN to each other. The apparent differences are ascribed to the effects of viewing angle or other non-fundamental parameters, working on a fundamentally identical type of underlying source. Such unification theories make specific predictions about the similarities of the different source types across the spectrum, which can be tested readily with complete and extensive observations of large samples of the different purportedly unified sources. Secondly, there is a specific paradigm for the emission of quasars, based on a central supermassive black hole with a surrounding accretion disk. The physical nature of the emission from this model can be probed in many ways with SIRTF, including tests of its overall validity.
The Broad Absorption Line (BAL) QSOs are a subset of radio-quiet QSOs which show strong, broad absorption lines just blueward of the corresponding emission line. All known BAL QSOs have been identified on the basis of optical spectra, with the degree of absorption near CIV being a useful measure of the strength of the BAL system. BAL QSOs have continua and emission-line properties which are very similar to those observed from other radio-quiet but non-BAL QSOs (Weymann et al. 1991). A prominent exception to this statement is the weak [OIII] lambda 5007 emission observed from low-redshift BAL QSOs which also show MgII (low- ionization) BAL systems. Four such QSOs have been observed (Turnshek et al. 1985, Boroson & Meyers 1992), and these data imply that the covering factor of the BAL material must be large, thereby inhibiting ionization of the narrow [OIII]-emitting gas. Questions have remained as to what range of covering factors are needed to explain what is observed (see Morris 1988 and Hamann et al. 1993 who argue for low covering factors and complete unification of BAL QSOs with radio-quiet QSOs). Another possibility is that perhaps the low-ionization BAL QSOs, somewhat different from other BAL QSOs in several respects (Weymann et al. 1991), have an [OIII] behavior that is atypical for all BAL QSOs. To investigate this latter question we have obtained 1.6-micron spectra of a sample of z=2-2.5 BAL and non-BAL QSOs. We found that there is a significant difference in the [OIII] between radio loud and radio quiet quasars in this sample, but that the BAL QSOs have [OIII] consistent with their radio properties. Thus, a large covering factor remains a possibility for the radio quiet BAL QSOs.
If the BAL material indeed has a large covering factor, then this material must be absorbing radiation. Any dust present would presumably radiate more strongly in the mid- and far-infrared than would be seen from a non- BAL QSO lacking this material. Sprayberry and Foltz (1992) demonstrate that at least in the case of the low-ionization BAL QSOs dust does appear to be present in the BAL region. We propose to measure the rest-frame 3-60 micron spectral energy distributions of the sample of BAL QSOs for which we have near-infrared spectroscopy. We will also observe a limited number of non-BAL QSOs in the same redshift range. If the infrared emission, assumed to be an isotropic property, is the same for all classes, then a unified model may still be viable, but the lack of [OIII] would require a more complicated model of the broad absorption line region with collisional de-excitation causing the reduced [OIII]. If the infrared emission differs, then the hope of unifying BAL QSOs and non-BAL QSOs may have to be abandoned. In either case we will also learn something about the nature of the broad absorption line region. Our sample of high redshift BAL QSOs includes three which show MgII BAL absorptions as well as higher ionization BAL systems so we may also be able to see whether the MgII systems differ from the others.
Little is known about the mid- and far-infrared emission from BAL QSOs with only a few low-redshift, MgII BALs having been detected by IRAS (Low et al. 1989). The rate of finding these rare MgII BAL QSOs in IRAS data suggests that the IR emission may be enhanced in these objects, but the statistics are very poor. In addition Cutri (private communication) has found that the rate of detecting higher redshift BALs in IRAS data does not appear to be larger than that for non-BALs. However, it must be emphasized that these results are based on statistics of less than 10 sources. One can speculate that BALs may have enhanced IR emission because they tend to have enhanced FeII emission which in turn is correlated with enhanced IR emission (Low et al. 1989).
The fluxes from a sample of BAL and non-BAL quasars will be observed at 8 or 10, 24, 70, and 160 microns. The shortest wavelength can be obtained either from existing ISOCAM images, from the ground (brightnesses are 1 - 3 mJy), or with IRAC. The remaining three wavelengths would require MIPS. The list of BAL QSOs was selected on the basis of having redshifts in the range where H-beta, [OIII] lambda-5007, and H-alpha can all be observed with groundbased spectrometers. We are proposing to observe 8 BALs. We are also proposing to observe 4 radio- quiet, non-BAL QSOs and 2 radio-loud QSOs for comparison. These 6 QSOs have been selected to have similar redshifts and brightnesses to the BAL sample.
Boroson, T. A. and Meyers, K. A. 1992, ApJ, 397, 442
Hamann, F., Korista, K. T., and Morris, S. L. 1993, ApJ, 415, 541
Low, F. J., Cutri, R. M., Kleinmann, S. G., and Huchra, J. P. 989, ApJL,
340, L1
Morris, S. L. 1988, ApJL, 330, L83
Rieke, G. H., Young, E. T., and Gautier, T. N. 1994, paper at Saclay
conference
Sanders, D. B., Phinney, E. S., Neugebauer, G., Soifer, B. T., and Matthews,
K. 1989, ApJ, 347, 29
Sprayberry, D. and Foltz, C. B. 1992, ApJ, 390, 39
Turnshek, D. A., Foltz, C. B., Weymann, R. J., Lupie, O., McMahon, R. G.,
and Peterson, B. M. 1985, ApJL, 294, L1
Weymann, R. J., Morris, S. L., Foltz, C. B., and Hewett, P. C. 1991, ApJ,
373, 23
Much of extragalactic astronomy for the past four decades has been devoted to the nature of quasars. Among the advances resulting are: 1.) powerful quasars generally lie at the nuclei of significant galaxies; 2.) there are three distinct types of quasar, radio quiet, radio loud, and violently variable/BL Lac. The proposed observations address whether these findings are complete. Specifically, we will find out if there is a fourth type of quasar that is radio loud but optically quiet and non-variable, and that may not need to be at the nucleus of a significant galaxy. The members of this fourth class have been classified as "empty fields" in optical campaigns to identify radio sources and have been omitted from consideration because of our optically-oriented classification techniques.
Most flat spectrum radio sources appear to have a nonthermal spectral energy distribution extending into the optical, and the ratio of radio (say at 5 GHz) to optical flux is very closely the same from one source to another. This proportionality, which continues with few exceptions into the X-ray, suggests some kind of commonality among all these objects in their underlying engines, even though optical spectroscopy and variability characteristics have led us to divide them into distinct categories such as quasi-stellar object, optically violently variable quasar, and BL Lac object, and current unification theories would place the QSOs and the BL Lac objects in different types of host galaxy with different underlying characteristics. We propose to investigate a small class of flat spectrum radio sources that does NOT share this proportionality, and which therefore is a candidate to operate by some fundamentally different mechanism.
Our examples of such sources are from the sample of some 518 high frequency radio sources originally collected by Kuhr et al. (1981) as the 1 Jy (at 5 GHz) survey. Extensive work has been conducted on this sample, including optical identifications and spectroscopy. After the low frequency 3C survey, the 1 Jy survey can lay claim to be the most thoroughly studied complete radio sample. Because the 1 Jy sample is roughly 50% flat spectrum sources, it is ideal for statistically unbiased studies of this type of object. The sample has been virtually completely identified, with general results consistent with the proportionality discussed above. If a different type of source lies within the sample, it must be among the very faintest (in the optical) members.
Over the past fourteen years, we have monitored 15 of the optically faintest identifications in the 1 Jy Catalog in the near infrared to determine their nature. Most of these objects have spectral energy distributions that fall very steeply from the near infrared to the optical, resembling the red QSOs which a number of workers, including ourselves, have found among the very faint optical identifications in other radio samples (Rieke, Lebofsky, and Kinman 1979; Smith and Spinrad 1980; Beichman et al. 1981; Bregman et al. 1981; Rieke, Lebofsky, and Wisniewski 1982; Webster et al. 1995). The very red infrared-to-optical continua appear to arise from two causes. In many cases the source is obscured, either by the host galaxy or by a foreground galaxy (e.g., Stocke et al. 1992); in some cases, the foreground galaxy lenses a background source as well as reddens it (e.g., 0218+35: Patnaik et al. 1993; Stickel and Kuhr 1993). However, in other cases (e.g., 0202+14) the continuum is a steep power law falling with the same slope from 2 microns through 0.4 microns; in these cases, reddening is not a likely explanation and the object appears to be a BL Lac or close relation to one, but with an intrinsically very steep continuum in the near infrared to optical (we find that most of these sources are violently variable). Our large scale study is described in Stickel, Rieke, Kuhr, and Rieke (1996).
If we include the near infrared, nearly all flat spectrum members have an optical or infrared identification with a brightness closely proportional to the radio flux density. There are, however, a few striking exceptions. For example, in the 1 Jy sample there are two sources which are far fainter than expected from this proportionality, well outside the distribution determined by the other 516 sample members. These objects appear to represent a genuinely optically quiet form of quasar, i.e., one that is not artificially made dim by obscuration in the host galaxy or by a galaxy along the line of sight. Furthermore, they must may lie in anomalously low luminosity host galaxies. They may therefore represent a new emission mechanism or geometry as well as a unique local environment for quasars.
One of these sources is 0742+10, with alpha {rad} = -0.1 (we define F_nu proportional to nu^alpha). In addition, there is a very similar object, 1413+349, that just missed classification as "flat spectrum" in the radio because its spectral index is alpha{rad} = -0.56 (flat spectrum requires alpha{rad} > -0.5). In repeated aperture photometry and imaging, spanning the past 14 years, the former source has not varied from a K magnitude of ~ 19.7, i.e., 9 microJy, and we have only recently detected it in the optical at R ~ 23.6; thus, the source is not particularly red in color (compared with other faint flat spectrum identifications), with a spectral index between R and K of about alpha{OIR} = -1.7. The latter object is at K ~ 19.5 and R > 24 (Stickel et al. 1996, Stanghellini et al. 1993). Not only are these sources by a substantial margin the faintest members of the sample in the optical and near infrared, but 0742 is among the brighter objects in the radio (3.7 Jy at 5 GHz). They therefore deviate from the typical ratio of radio to near infrared flux by a factor of ~ 100 -- 300. They are also considerably more extreme in this regard than any of the sources discussed by Webster et al. (1995). Unlike the situation for many of the other optically faint radio sources, the faintness at 2 microns and the R - K color of 0742 suggest that reddening does not play an important role with these objects. Thus, they are very strong candidates to be the "different" sources we hypothesized might be found in the survey. In support of this possibility, we note that the ratio of radio to IR flux is not just at the extreme of the distribution, but is removed from the extreme by factors of > 30 for 0742 and > 10 for 1413 (the slopes from 5 GHz to 2 microns are respectively alpha{RIR} = -1.3 and -1.1, whereas the range for other objects is generally from -0.5 to -0.9). We have achieved a weak detection of 1413 with ISO at 10 microns, consistent with interpolation between K and the high frequency radio points, further emphasizing that the faintness of these objects is unlikely to arise from reddening and that it is relatively broad in spectral range.
These two objects represent less than 1% of the 1 Jy sample. However, related sources may lie at the nuclei some of the flat spectrum radio galaxies, which constitute about 11% of the 1 Jy flat spectrum sample and 18% of the closely related S4 sample (Stickel et al. 1994; Stickel and Kuhr 1994). Thus, it is not clear whether we are dealing with a rare type of quasar or with a moderately widespread phenomenon, on a scale, say, with BL Lac objects which represent 13% and 10% respectively of the two surveys.
We propose to observe them both with IRAC (four bands) and with MIPS at 24, 70, and 160 microns to characterize their spectra between the near infrared and the radio in enough detail to understand their behavior relative to other flat spectrum quasars.
Beichman et al. 1981, Nature, 293, 711
Bregman et al. 1981 Nature, 293, 714
Kuhr et al. 1981, A&AS, 45, 367
Patnaik et al. 1993, MNRAS, 261, 435
Rieke, Lebofsky, and Kinman 1979, ApJL, 232, L151
Rieke, Lebofsky, and Wisniewski 1982, ApJ, 263, 73
Smith and Spinrad 1980,ApJ, 236, 419
Stanghellini et al. 1993, ApJS, 88, 1
Stickel et al. 1994, A&AS, 105, 211
Stickel and Kuhr 1993, A&AS, 101, 521
Stickel and Kuhr 1994, A&AS, 105, 67
Stickel, Rieke, Kuhr, and Rieke 1996, ApJ, 468, 556
Stocke et al. 1992, ApJL, 400, L17
Webster et al. 1995, Nature, 375, 469
It is widely believed that radio loud quasars are the same type of source as luminous radio galaxies of type FRII (Barthel 1989; Bregman, 1990; Urry & Padovani 1995). In this picture, the object appears as a quasar when we are close to the axis of its radio jet. It is a radio galaxy when the radio jet is close to the plane of the sky and a direct view of the nucleus is hidden by the circumnuclear accretion torus that is oriented with its plane toward us. Possibly, broad line radio galaxies appear when the torus is sufficiently face on that we see into the nuclear engine region, but we are sufficiently off-axis relative to the jet that it does not appear brightly (e.g., Barthel 1989). Yet another version is radio galaxies far off axis, steep spectrum radio quasars somewhat off axis, and flat spectrum radio quasars close to on axis (e.g., Urry & Padovani 1995).
The IRAS data were used by Hes et al. (1995) to probe whether the infrared properties of the two source types were identical, under the assumption that any accretion disks would be optically thin in the FIR and hence would radiate isotropically there. They studied complete samples of radio galaxies and quasars by selecting all objects in the 3CR catalog with 0.3 < z < 0.8. They found that the quasars in this sample appear to be significantly brighter than the radio galaxies, as can be seen even from the detection statistics in the appendix.
Although this result appears to be contradictory to what is otherwise a successful unification, a number of explanations have been offered. Hoekstra et al. (1997) compute the amount of beamed emission that might be expected in the far infrared, using as a boundary condition the ratio of compact nuclear to extended emission in the radio and assuming a similar ratio holds in the FIR. Their model shows that the results are consistent with the unification, if the far infrared includes a significant component that is generated nonthermally and is beamed in a manner similar to the radio emission. As discussed below, this model can probably be tested by observation of a sample of radio quiet quasars, since one expects much less powerful beamed emission from them and hence there is prediction that a suitably matched sample would be of lower far infrared luminosity than the radio loud objects. Alternately, the circumnuclear torus may be optically thick, even in the far infrared. In that case, it emits relatively little when seen edge on, as in the radio galaxies as opposed to the quasars (e.g., Pier & Krolik 1992: Granato & Danese 1994).
Finally, the detection statistics in the IRAS data are very poor - for example, 3 radio galaxies in the Hes et al. (1995) sample are detected, out of 49, and 7 quasars out of 19. By detecting a much higher portion, presumably all, of the sample members, SIRTF has the potential to provide a much more stringent test of the unification. However, to take advantage of the better data will require careful modeling along the lines of that of Hoekstra et al. (1997), along with appropriate tests such as through observation of radio quiet quasars matched in properties to the radio loud ones.
To illustrate the scope of a program in this area, the appendix lists the 3CR sources in the study of Hes et al. (1995). Other samples that might be considered are the 2 Jy sample (Wall & Peacock 1985) and the 1 Jy sample (Kuhr et al. 1981). These samples are selected respectively at 178MHz, 2.7 GHz, and 5 GHz, and hence have differing biases in their sample members. For example, the 3CR sample is heavily weighted toward steep spectrum sources, while the 1 Jy sample is about 50/50 steep and flat spectrum ones, as a result of the natural weighting given by the frequency of observation. All three samples have the advantage of being extensively studied in the optical, with nearly complete identifications.
Barthel, P. D. 1989, ApJ, 336, 606
Bregman, J. N. 1990, A&A rev., 2, 125
Granato, G. L., & Danese, L. 1994, MNRAS, 268, 235
Hes, R., Barthel, P. D., & Hoekstra, H. 1995, A&A, 303, 8
Hoekstra, H., Barthel, P. D., & Hes, R. 1997, A&A, 319, 757
Kuhr, H., et al. 1981, A&A, 45, 367
Pier, E. A., & Krolik, J. 1992, ApJ, 401, 99
Urry, C. M., & Padovani, P. 1995, PASP, 107, 803
Wall, J. V., & Peacock, J. A. 1985, MNRAS, 216, 173
Considerable effort has been invested in demonstrating a probable link between ultraluminous infrared galaxies and quasars. Although some skepticism persists, it seems likely that about 30% of the local ultraluminous galaxies are powered predominantly by dust embedded AGNs whose optical, UV, and soft x-ray outputs are absorbed and reradiated in the mid and far infrared (Sanders et al. ; Lonsdale, Smith, & Lonsdale 1995; Genzel et al. 1998; McLeod et al. 1998). For nearby ultraluminous galaxies, this possibility has been probed thoroughly and will be tested again with IRS spectroscopy.
A prediction of this hypothesis is that the ultraluminous galaxies should share in the strong cosmic evolution of quasars. Thus, a deep MIPS survey might be expected to find a large number of luminous dust-embedded quasars at high redshift. A reliable way to identify such objects is through a combination of infrared and hard x-ray data, the infrared to detect the signature of high reradiated luminosity and the x-rays to probe for the presence of a powerful AGN. Although AXAF is not a hard x-ray mission, at high redshift it actually probes the hard x-ray region in the source rest frame since at z = 1, 8keV rest is moved well into the AXAF sensitivity range. Therefore, if we survey the deep AXAF fields with MIPS, we should be able to identify a large number of quasars and active galaxies and should be able to classify them through a combination of their far infrared SEDs and optical and near infrared imaging and spectroscopy.
The relevant AXAF GTO programs for SIRTF mapping are:
HRC-I has a 31x31 arcmin FOV
ACIS-S has a 8 x 50 arcmin FOV
ACIS-I has a 16x16 arcmin FOV
Non-HDF south refers to the presence of a bright star near the HDF south
field that would blind the AXAF aspect sensors, so AXAF will do a deep
field but it will not coincide with the field selected for the southern HDF.
The sensivity limit for HRC-I is expected to be 5-sigma = 2.0E(-15)
erg/cm^2 s
The effective collecting areas for ACIS are 2-3 times larger, implying
greater sensitivity by a factor of 1.5-2.
In addition, there is a 100 ksec exposure with ACIS-S on a high latitude cluster with C. Canizares as PI. This is the only other GTO program of interest for deep imaging.
Genzel, R., et al. 1998, ApJ, 498, 579
Lonsdale, C. J., Smith, H. E., & Lonsdale, C. J., 1995, ApJ, 438, 632
McLeod, K. K., Rieke, G. H., Storrie-Lombardi, L., & Weymann, R. J.,
1998, preprint
Sanders, D. B., Soifer, B. T., Elias, J. H., Neugebauer, G., & Matthews, K.
1988, ApJL, 328, 35
There has been substantial interest in the possibility that the fall-off in quasar numbers at z > 2.5 may be due to reddening that suppresses the rest frame UV and hence makes most quasars difficult to detect through standard optical/UV search techniques. For example, Fall and Pei (1993) suggest that 10 to 70% of bright quasars are missing from optical samples by z = 3, and that 90% may be missing by z = 4. If this hypothesis is correct, it not only is important in understanding quasar evolution, but it also can be significant in understanding the far infrared background at high z.
The deep surveys in AXAF fields discussed in the preceding section would appear to be an optimum approach to looking for reddened quasars. If the surveys are coordinated with IRAC, we could also make use of mid infrared imaging that would potentially be very helpful both in constraining the SED of the quasars, important in determining if their colors are really due to reddening, but also in locating the identifications accurately enough for groundbased followup, for example to obtain redshifts. Further assessment is required of the expected detection rate in these limited area fields, but for now we will assume that they are large enough for this project.
Fall, S. M., & Pei, Y. C. 1993, ApJ, 402, 479
One of the advantages of unified models is, if the model can be taken to be strictly correct, it can be used to derive other aspects of the source. For example, assume the radio galaxy - steep spectrum radio quasar - flat spectrum radio quasar unification is correct. Then from the relative numbers of each class of object, we can deduce the opening angle of the obscuring torus (e.g., transition from galaxy to steep spectrum radio quasar) and the typical angle of source beaming (e.g., transition from steep spectrum to flat spectrum quasar).
An application of this principle would invert the arguments in Section 2.3. Comparing the luminosities of a sample of steep spectrum radio galaxies with those of the unified steep spectrum radio quasars would reveal the degree of anisotropy in the far infrared emission. Assuming the mid to far infrared emission arises predominantly from energy from the central engine, absorbed by a large-scale circumnuclear disk and thermally reradiated (e.g., Sanders et al. 1989), the SED in the infrared could constrain models for the geometry of this disk. It is implicit in the usual arguments based on number counts of different manifestations of a unified source type that one is exploring the covering factor of the purported circumnuclear disk. Combining such information with the information from the SEDs should further constrain the disk models.
Similarly, if beamed nonthermal emission makes a significant contribution to the far infrared emission of steep spectrum radio quasars (Hoekstra et al. 1997), then there should be a correlation of SED shape with the strength of the compact radio component. Also, the evidence suggests that radio quiet quasars truly have central engines very similar to those of radio loud ones. In that case, the absence of a strong compact central beamed source in the radio quiet quasars should result in a subtly different SED, which might be detected by study of carefully matched sets of quasars.
Sun and Malkan (1989) obtained broad SEDs for 60 quasars and fitted them with accretion disk models. A reasonably large selection of quasars for similar studies would be available from the samples already listed in this white paper. We add a small portion of time for observations of specific sources of interest to direct accretion disk model fitting.
Hoekstra, H., Barthel, P. D., & Hes, R. 1997, A&A, 319, 757
Sanders, D. B., Phinney, E. S., Neugebauer, G., Soifer, B. T., & Matthews,
K. 1989, ApJ, 347, 29
Sun, W.-H., & Malkan, M. A. 1989, ApJ, 346, 68
In the case of Seyfert galaxies, comparison of small beam groundbased mid infrared photometry with the measurements with IRAS Band 1 (12 microns) reveals that much of the far infrared emission arises from the host galaxies. In the case of Seyfert 2 galaxies, the hosts appear to have significantly higher luminosity than field galaxies (Maiolino et al. 1995), presumably powered by an elevated rate of star formation. A similar result has been achieved for luminous infrared galaxies (Carico et al. 1990), as is also supported by radio images of these objects (Condon et al. 1991).
To what extent are the far infrared fluxes of quasars generated in a similar manner? This question can be probed in a manner exactly analogous to the studies of Maiolino et al. (1995) and Carico et al. (1990). Groundbased imaging at 10 microns can now be obtained with a resolution up to ~ 0.3". The resulting flux densities can be compared with SIRTF measurements, where the diffraction limit is about 3" near this wavelength. Unfortunately, SIRTF does not have a band exactly coincident with the atmospheric window. The IRAC band at 8 microns is not only short of the groundbased N band, but even at zero redshift it is likely to be strongly affected by PAH feature emission (a strong feature lies at 7.7 microns), as shown by ISO measurements of nearby galaxies. A better choice is the IRS peakup band at 15 microns, or perhaps a combination of the 8 and 15 micron bands (although it would be inefficient to observe with two instruments).
If it turns out that quasar host galaxies are exceptionally luminous, as they would have to be to be detectable above the emission of the quasar itself, then very powerful star formation is implicated in these galaxies. Such an observation would be one of the strongest indications of a unification between ultraluminous infrared galaxies and quasars, since various lines of evidence suggest that most of the ULIRGs are currently dominated by star formation (Rieke 1988; Genzel et al. 1998). A suitable sample for such a study is the low redshift, low luminosity PG quasars listed in the appendix. These objects are nearly all resolved in the near infrared with a resolution of ~ 1" (McLeod & Rieke 1994), so extended emission from the host galaxies should be readily detected in this experiment.
Carico, D. P., Sanders, D. B., Soifer, B. T., Matthews, K., & Neugebauer, G.
1990, AJ, 100, 70
Condon, J. J., Huang, Z.-P., Yin, Q. F., & Thuan, T. S. 1991, ApJ, 378, 65
Genzel, R., et al. 1998, ApJ, 498, 579
Maiolino, R., Ruiz, M., Rieke, G. H., & Keller, L. D. 1995, ApJ, 446, 561
McLeod, K. K., & Rieke, G. H. 1994, ApJ, 420, 58
Rieke, G. H. 1988, ApJL, 331, L5
UM 275 [BAL]
0h 43m 39.5s 0d 48' 03.0" 1950
0059-2735 [BAL]
0h 59m 52.4s -27d 35' 57.0" 1950
CSO 203 [BAL]
8h 42m 30.4s 34d 31' 41.0" 1950
0226-1024 [BAL]
2h 26m 12.7s -10d 24' 32.0" 1950
1011+0906 [BAL]
10h 11m 03.3s 9d 06' 21.0" 1950
1246-0542 [BAL]
12h 46m 38.8s - 5d 42' 58.0" 1950
1309-0536 [BAL]
13h 09m 00.7s - 5d 36' 43.0" 1950
1331-0108 [BAL]
13h 31m 53.7s - 1d 08' 29.0" 1950
2212-1759 [BAL]
22h 12m 48.3s -17d 59' 07.0" 1950
UM 208 [RQ]
0h 07m 42.8s - 0d 04' 15.7" 1950
UM 288 [RQ]
0h 49m 59.5s 1d 24' 24.0" 1950
POX 42 [RQ]
11h 58m 11.3s -18d 43' 03.0" 1950
2310+3831 [RQ]
23h 10m 36.3s 38d 31' 23.0" 1950
PHL 1305 [RL]
2h 26m 22.1s - 3d 50' 58.0" 1950
1435+6349 [RL]
14h 35m 37.2s 63d 49' 36.0" 1950
Total time at 0.5 hours/source: 7 hours
2.2. Optically Quiet QSOs:QSO-0742+100 (optically quiet) 07 45 33.05 10 11 12.7 2000
QSO-1413+349 (optically quiet) 14 16 04.15 34 44 36.2 2000
Total time at 1 hour/source (to include IRAC and MIPS): 2 hours
2.3 RLQ - FRII radio galaxy unification:3CR galaxies with 0.3 < z < 0.8, not detected by IRAS:
3C16 3C200 3C306.1 3C19 3C220.1 3C313 3C34 3C220.3 3C320 3C41 3C225B 3C323 3C42 3C228 3C327.1 3C44 3C244.1 3C330 3C46 3C247 3C337 3C49 3C268.3 3C340 3C55 3C274.1 3C341 3C67 3C275 3C343.1 3C107 3C277 3C411 3C142.1 3C277.2 3C427.1 3C169.1 3C292 3C434 3C172 3C293.1 3C435 3C187 3C295 3C441 3C299
detected:
3C99
3C109
3C318
3CR quasars with 0.3 < z < 0.8 and not detected by IRAS:
3C93 3C254 3C138 3C263 3C147 3C275.1 3C154 3C277.1 3C175 3C380 3C216 3C455
detected:
3C47 3C334 3C48 3C345 3C207 3C351 3C249.1
Total time, assuming all sources are observed @ 0.5 hours/source: 35 hours
2.4 Assume we map all three deep AXAF fields, with 2000 seconds integration per point, and that the net efficiency is 50% (due, for example, to overscan to cover the fields). Then the total time for this program is 72 hours. However, the observations would also be usable by the deep survey program.
3.1 A possible sample, at least correct in scale, would be the nearby high luminosity PG quasars:
0026+129 1226+023 0052+251 1302-102 0157+001 1307+085 0923+201 1309+355 0947+396 1322+659 0953+414 1352+183 1004+130 1354+213 1013+008 1402+261 1048+342 1427+480 1116+215 1444+407 1121+422 1545+210 1151+117 1613+658 1202+281 1700+518
To estimate the time requirement, we assume these objects are observed photometrically and with the SED mode, with supporting groundbased data for the near and mid infrared. We assume that the 19 radio loud quasars in the previous investigation are also observed in SED mode for a comparison sample. Taking one hour per PG source, and 0.5 hours (extra) per 3CR source, the time is then 36 hours. We allow 10 more hours for objects specifically of interest for accretion disk model fitting (or for expanded SIRTF coverage of objects already listed). The total time required is then 46 hours.
3.2. Low redshift, low luminosity PG quasars
PG0050+124 PG1415+451 PG0804+761 PG1416-129 PG0838+770 PG1426+015 PG0844+349 PG1435-067 PG1001+054 PG1440+356 PG1114+445 PG1519+226 PG1115+407 PG1552+085 PG1211+143 PG1612+261 PG1229+204 PG1617+175 PG1351+640 PG1626+554 PG1404+226 PG2130+099 PG1411+442 PG2214+139
Total time, assuming observations by MIPS. IRAC. and IRS imaging at 1.5 hour/source: 36 hours