¹University of Tokyo, Japan
²National Astronomical observatory, Japan
³The Graduate University for Advanced Studies, Japan

Pair (Triple) quasars separated by a few arcmin on the sky sometimes have common metal absorption lines at almost the same redshift ([1],[2],[3]). This implies the existence of intervening absorbers clustering at the redshift. In fact, several galaxies have been detected as counterparts of absorption lines around quasars at the redshifts identical to those of absorption lines ([4],[5],[6],[7]).

The quasar triplet (KP76/KP77/KP78) is located on the sky within a small FOV of 3 arcmin ([3]). All of them have C IV absorption line complexes at z=2.24 in their spectra ([8]). The radial velocity separations of these absorption lines are within 1600 km/s of each other. The linear angular distance on the sky between these systems at z=2.24 corresponds to ~1 Mpc (Omega₀=1, H₀ = 75 km/s/Mpc), which is too large for a single absorber to intercept all three lines of sight. Such common absorption lines are plausibly produced by gas clouds associated with galaxies that constitute a cluster (group) of galaxies at z=2.24

Figure 1 : K98 and 2.122(S1) images of the field around KP78. The quasar and the two candidates for star forming galaxies at z=2.24 are marked with circles. The size of the field is 90" x 90", which corresponds to 480kpc x 480kpc (Omega₀ = 1, H₀ = 75 km/s/Mpc).

Based on this assumption, we carried out a deep imaging observation of the field around one of the quasar triplet (KP78) with UKIRT + UFTI on May 24 - 25, 2002. We used the K98 filter as a broad-band, and the narrow-band 2.122(S1) filter which covers the H-alpha emission lines of the galaxies at z=2.24. With effective exposure times of 3.3h for K98, and 9.25h for 2.122(S1), we acquired the wonderful images in Figure 1 to depths of 20.9 mag for K98 and 20.0 mag for 2.122(S1), with S/N=5.

Figure 2 : 2.122(S1) magnitude versus K98 - 2.122(S1) colour diagram. The filled stars denote objects that are detected in the 2.122(S1) images, while the thin dots represent the artificial galaxies for comparison. The obj1 and obj2 are candidates for the star forming galaxies at z=2.24.

Figure 2 is the color-magnitude diagram for the comparison of detected objects with artificial ones. For the color excess analysis to isolate H-alpha emitting objects, accurate evaluation of the photometric errors is necessary. We have created about 25,000 artificial galaxies on the observed images. A pair of dashed, dash-dotted, and dotted lines in Figure 2 indicate the 1-sigma, 2-sigma and 3-sigma distributions of artificial galaxies, respectively. Two objects were found to be deviated beyond the 3-sigma level. We could not detect them in the K98 image even with S/N=2, but they are evident in the 2.122(S1) image, which suggests that they have very strong line fluxes compared with continuum fluxes.

If they are attributed to H-alpha emission lines on the band-pass of the 2.122(S1), we can estimate the lower limits of their star formation rates to be >125, and >54 Mo per year for obj1 and obj2, respectively. The colour excess, however, could be caused not by H-alpha but by Ly-alpha, [OII], [OIII], or other lines ([9],[10]). Spectroscopic follow-up observation is clearly required for further discussion.

If we confirm that our technique is useful for finding clusters (groups) of galaxies at high redshift, we will apply it to the fields around other quasar pair (triplet). UIST with its wide FOV (120" x 120") can be expected to play an important role for the achievement of our purpose.

References
[1] Shaver, Boksenberg, and Robertson, 1982,ApJ,261,7
[2] Jakobsen et al., 1986,ApJ,303,27
[3] Crotts & Fang, 1998,ApJ,502,16
[4] Bergeron & Boisse, 1991,A&A,243,344
[5] Lanzetta et al., 1998,astro-ph/9812272
[6] Chen, Lanzetta, and Webb, 2001,ApJ,556,158
[7] Chen et al., 2001,ApJ,559,654
[8] Crotts, Burles, and Tytler, 1997,ApJ,489,L7
[9] Teplitz, Malkan, and McLean, 1999,ApJ,514,33
[10] Teplitz et al., 2000,ApJ,542,18

Institute for Astronomy, University of Hawaii

While the number of known brown dwarfs is growing rapidly, the origin of these objects is an unanswered question. One insight into the formation mechanism(s) for brown dwarfs is whether young substellar objects possess circumstellar disks. There is abundant observational evidence and theoretical expectation for accretion disks around young solar-type stars. Thus, the presence of disks around young brown dwarfs would be naturally accommodated in ``star-like'' formation scenarios. On the other hand, scenarios involving dynamical interactions (e.g., collisions and/or ejections) are likely to be hostile to circumstellar disks.

In collaboration with Joan Najita of NOAO and Alan Tokunaga of the University of Hawaii, I have recently completed a large L'-band (3.8 micron) survey for disks around young brown dwarfs and very low mass stars using IRCAM/TUFTI on UKIRT (Liu, Najita, & Tokunaga 2002, ApJ, in press). Brown dwarf disks can be readily identified by excess L'-band emission, which arises from warm material within a few stellar radii (<0.1 AU). In addition, our survey is sensitive enough to detect young brown dwarf photospheres, and hence the absence of a disk can be discerned --- therefore we can determine the frequency of disks around brown dwarfs for the first time. Our sample comprises nearly all the known sources in the Taurus and IC 348 star-forming regions which have been spectroscopically classified to be very cool, with masses of ~15 to ~100 M_Jup based on current models. (At ages of a few Myr, young brown dwarfs have spectral types of mid to late-M. As they age and consequently cool, they will become the L and T-type objects found in the field.)

Figure 1: Dereddened (K_s - L') colors (2.0-4.1 micron) as a function of spectral type for our sample of young brown dwarfs and very low mass stars. Approximate mass estimates are listed at the top. The heavy line represents the colors of field M dwarfs. Most of our targets show intrinsic IR emission in excess of that expected from their bare photospheres.

Figure 1 shows that most objects in our sample have intrinsic colors which are redder than expected from their bare photospheres, i.e., they possess IR excesses. Also, the lower envelope of the observed color distribution agrees well with the heavy line, indicating that field M dwarfs provide a legitimate comparison. Like the case for T Tauri stars, such IR excesses are naturally explained by circumstellar disks; non-disk (e.g. spherical) configurations of dust which could produce the IR excesses would lead to very high line-of-sight extinctions, which is not observed.

The disk frequency in our sample appears to be independent of mass; however, some objects, including the very coolest (lowest mass) ones, lack IR excesses. Disks around young brown dwarfs and very low mass stars appear to be very common --- we find an excess frequency of ~75% for the sample as a whole. This disk fraction is comparable to the disk fraction observed for young stars in the same regions.

For T Tauri stars, both the optical line emission and the IR excesses are believed to originate from disk accretion. The IR excesses come from warm dust grains in the disk, while H-alpha emission arises from accretion of disk material onto the central star, e.g., via a boundary layer or a magnetically regulated accretion flow. Therefore, the H-alpha emission and IR excesses should be correlated, and indeed such a correlation is seen among T Tauri stars (e.g., Kenyon & Hartmann 1995, ApJS, 101, 117). Figure 2 shows a comparison for our sample of young brown dwarfs and very low mass stars, using H-alpha data from the literature. A 3-sigma correlation is observed between the intrinsic (K_s - L') excess and H-apha emission, based on the Spearman rank correlation coefficient. This level of correlation is comparable to that observed for T Tauri stars and provides strong circumstantial evidence for accretion disks around young brown dwarfs. Furthermore, the mere existence of accreting brown dwarfs at ages of a few Myr argues for mass-dependent accretion rates, since brown dwarfs with typical T Tauri star accretion rates would not remain substellar.

Figure 2: Comparison of the intrinsic (K_s - L') excesses and H-alpha equivalent width in Angstroms for most of our sample. The two quantities are well-correlated, supporting the idea that the optical and near-IR emission both originate from the same phenomenon, namely circum(sub)stellar accretion disks.

Altogether, we find that (1) disks around brown dwarfs are common, and (2) brown dwarf disks are contemporaneous with disks around T Tauri stars. The latter shows that brown dwarf disks are at least as long-lived as disks around stars, assuming that the stars and brown dwarfs are roughly coeval. These observations are naturally accommodated in a picture where brown dwarfs form in a similar manner as stars --- our results offer prima facie evidence for a common origin for objects from the stellar regime down to the substellar and planetary-mass regime.

Alternative formation scenarios, such as disk-disk collisions and/or premature ejection, involve dynamical interactions in creating brown dwarfs. While specific predictions are lacking due to the stochastic nature of these scenarios, brown dwarfs formed by these mechanisms are generally expected to have smaller and less massive disks, and consequently shorter disk lifetimes, compared to brown dwarfs formed in isolation. This expectation conflicts with our finding that brown dwarf disks are at least as long-lived as disks around young stars.

Further studies will be important for determining the properties of brown dwarf disks, and also for placing these objects in context with our physical understanding of the star formation process. Finally, the very high disk fraction of young brown dwarfs raises the possibility of forming planets around brown dwarfs. Such planetary systems would represent an fascinating alternative to the numerous planetary systems found around solar-type stars.

California Institute of Technology, Pasadena, CA 91125

The spectral shape of the far-infrared background (FIRB) detected by FIRAS at 140 and 240 micron (Puget et al. 1996; Fixsen et al. 1998) indicates a peak at ~200 micron with energy comparable to the optical/UV background. This peak arises from optical/UV radiation from star formation and AGN activity in obscured galaxies at z >> 0 which is absorbed by dust and reradiated in the far-infrared. This obscured population of galaxies could host approximately half of the massive star formation activity over the history of the Universe (e.g. Blain et al. 1999).

Far-infrared surveys at wavelengths close to the peak of the FIRB provide a powerful route for understanding the properties of the obscured activity in the distant Universe and its relevance to the formation and evolution of both galaxies and super-massive black holes. The FIRBACK (Far-InfraRed BACKground) survey obtained wide-field imaging at 170 micron with the PHOT instrument on-board ISO satellite in three separate regions of the sky chosen for low Galactic cirrus foreground. FIRBACK is the most reliable and deepest (\sigma(170 micron) ~ 40 mJy) infrared census at wavelengths near the peak in the FIRB. The FIRBACK sources down to 120 mJy account for about 10% of the FIRB seen by COBE at 140-240 micron (Puget et al. 1996). Evolutionary models (e.g. Dole et al. 2001) suggest that the sources identified by FIRBACK comprise both star-forming galaxies at low redshifts, z ~ 0.1, and a population of much more luminous galaxies at higher redshifts. The models predict that a quarter of the FIRBACK sources should have z > 0.5, with a tail reaching beyond z=1.5. The sources in this high-redshift tail provides the strongest constraints on the evolution of the population contributing to the peak of the FIRB at ~200 micron. For this reason the identification and study of these galaxies is of particular interest.

The coarse beam of ISO at 170 micron produces large uncertainties in the source positions, (100 arcsec diameter, 99% error circle). However, using the empirical radio-infrared correlation for star forming galaxies (Helou et al. 1985), we can exploit deep 1.4-GHz radio data from the VLA in C-array configuration (Ciliegi et al. 1999) to identify the FIRBACK sources with positional uncertainties of only ~1" (15" VLA beamsize). We can also identify counterparts using shorter wavelength observations which are expected to correlate with the 170 micron emission, such as the ISOCAM 15 micron maps over the same region of sky.

A large fraction of the proposed radio and mid-IR counterparts to the FIRBACK galaxies are not detected in shallow sky surveys in the optical (DPOSS2) or near-IR (2MASS). We have thus undertaken a K-band survey with the UFTI camera on UKIRT with the purpose of identifying the faint and red galaxy counterparts to the FIRBACK sources. We have also taken advantage of the superb seeing afforded by UKIRT to study the morphologies of some of the brighter and more nearby galaxy identifications to the FIRBACK population.

Figure 1: A montage of 30"x30" cutouts of our UKIRT K-band imagery of the FIRBACK population. The radio or ISOCAM identification is always centered in the frame. The lower S/N detections have been smoothed with the PSF. Sources are brightest at K to faintest (left to right) FN1-4, 24, 45, 44, 8, 68, 13, 42, 64, 78, 40, 48, 111, 117, 34, 32, and 59.

Here we provide a snapshot view (Figure 1) of our complete UKIRT observations, presented in a K-magnitude sequence from brightest to faintest, outlining the diversity in FIRBACK galaxy counterparts seen in the optical (all sources fall into a narrow range of 170 micron fluxes from 140-300 mJy). All panels show a 30" by 30" K-band image, where contours are used to highlight the structure of the brighter sources, while the faintest sources have been smoothed with the PSF for display purposes. An accompanying spectroscopic analysis based on Palomar 5-m observations of many of these galaxies will appear in a forthcoming paper.

Bright nearby spiral and elliptical galaxies form the brightest members of our sample. The best fit dust temperatures of ~20 K suggest that FIRBACK has selected galaxies similar in properties to the Milky Way, although often with significantly higher bolometric luminosities likely related to higher rates of ongoing star formation. Fine detail of the spiral and bar structures can be discerned with the superb ~0.4 arcsec UKIRT images.

Fainter galaxies appear resolved relative to a stellar PSF, but often do not reveal much more than a single near-IR component, detected at only ~5-10 \sigma and barely resolved, with red colors relative to the optical bands (where they are often undetected to R ~ 24). All targets attempted were detected with UKIRT, suggesting that the FIRBACK population with radio identifications are completely detected to K=21. Spectroscopic followup to these faint red galaxies is extremely difficult, even with 10m telescopes, and only three members of this class currently have robust redshift identifications. Two of these FIRBACK galaxies (FN1-40 and FN1-64) observed in our UKIRT program have already been presented in an ApJ paper (Chapman et al. 2002). The near-IR observations identified extremely red counterparts to both galaxies, and revealed complex merger morphologies, contrasting the more isolated examples in the greater sample. When available, the redshifts suggest that the dust temperatures are lower than expected for such luminous galaxies, by comparison to local ultra-luminous examples.

Ongoing study of this galaxy population will provide the first glimpse of the new classes of galaxies to be discovered with the SIRTF satellite, scheduled for launch January 2003. SIRTF will map large regions of the sky over similar wavelength bands to ISO, but at much greater sensitivities.

References:
Blain, A. W., Smail, I., Ivison, R. J., \& Kneib, J.-P. 1999, MNRAS, 302, 632
Chapman, S., Smail, I., Ivison, R., Helou, G., Dale, D., Lagache, G., 2002, ApJ, 573, 66
Ciliegi, P., et al., 1999, MNRAS, 302, 222
Dole, H. et al. 2001, ApJ, 591, 345
Fixsen, D. J., Dwek, E., Mather, J. C., Bennett, C. L., \& Shafer, R. A. 1998, ApJ, 508, 123
Helou, P., et al., 1985, ApJ 440, 35
Puget, J.-L., Abergel, A., Bernard, J.-P., Boulanger, F., Burton, W. B., Desert, F.-X., \& Hartmann, D. 1996, A\&A, 308, L5

Joint Astronomy Centre, Hilo.

The image below shows a combination of an UFTI K-band image with a HST/WFPC2 F814W I band image of the brightest cluster galaxy in Abell 193. The UFTI image revealed 3 nuclei in this galaxy for the first time (Seigar et al. 2002, in prep.), demonstrating the excellent spatial resolution available with UKIRT (FWHM~0.35"). This is only a factor of ~2 worse than the resolution capabilities of HST in the I band (as measured from the F814W image, although this is not the diffraction limit of HST due to undersampling of the psf with WFPC2).

Figure 1 - UFTI and HST imaging of the brightest cluster galaxy in Abell 193, showing the triple nucleus system revealed by UFTI.

Previously this galaxy was only believed to have a double nucleus (e.g. Hoessel et al. 1985, AJ, 90, 1648). Galaxies such as this one are thought to be constantly growing at the centres of clusters as they "cannibalize" their less massive neighbours (Hausman & Ostriker 1978, ApJ, 224, 300). As a result multiple nuclei are seen in approximately 50% of brightest cluster galaxies. In this galaxy (IC 1695) the resulting I-K colours of the nuclei (from north to south) are 1.74, 1.35, and 1.57 respectively, demonstrating that, if the cannibalism hypothesis is true, a merger has taken place in its recent history.