Institute for Computational Cosmology, Durham University, U.K.

High-angular resolution studies of the rest-frame optical morphologies and kinematics of faint high-redshift galaxies may provide a wealth of information about the formation and evolution of galaxies. However, they necessitate the use of an Adaptive Optics (AO) system to track and correct the effects of the atmosphere on the view of the distant galaxy and such systems require that the galaxies are typically within 25" of V < 15 stars. This has restricted the impact of AO for such studies and has forced researchers to undertake surveys for galaxies around bright stars. Larkin & Glassman (1999, PASP, 111, 1410) used 4 nights on Keck to identify roughly 15 galaxies within 25" of 5 bright stars, while Baker et al. (2003, A&A, 406, 593) took 6 nights on the NTT to identify 350 galaxies, brighter than K~19.5, within 25" of 42 stars. There are, however, opportunities for obtaining comparably deep data for free from archival calibration observations.

Undertaken as a 3 week summer-student project, our study (Christopher & Smail 2005, MN, submitted) exploited the UKIRT archive and the integrated ORAC-DR reduction pipeline to efficiently create a catalogue of faint galaxies around 16 bright stars. UKIRT is particularly well-suited for this purpose since the bulk of the photometric calibration relies on a small number of standard stars - the UKIRT Faint Standards (FS) from Hawarden et al. (2001, MN, 325, 563). There are approximately 140 hrs of observations of FS stars in the JHK filters with UFTI in the archive, so the cumulative integrations can reach the depth needed to study the morphologies of z > 1 galaxies, i.e. K~20. Moreover, UFTI calibrations typically jitter the star around one 46"-square quadrant of the HAWAII-1 detector - well matched to the isoplanatic patch.

 

Figure 1: 60" x 60" true-colour images of FS27 and FS29. Each panel is centred on the FS with North up and East left. The reddest galaxies in both fields, with (J-K)>2, are likely to be at z~2.

Our first task was to query the archive for all the UFTI observations of FS stars in the JHK-bands with more than 2.5 ksec integration in K, as of August 1st 2005. We removed any FS with V>15 and those with extinction of Ak>0.02. From the initial 83 faint standards this leaves us with 16 fields. These data were then retrieved and reduced using the ORAC-DR pipeline. Due to changes in the file headers, slightly different incantations were needed for data taken on different dates. Nevertheless, the basic steps taken by the pipeline were identical: subtracting suitable dark frames, creating (using the data themselves) and applying a flat field and aligning the frames, before co-adding them to create a final image. For the most part the pipeline reduces the observations with no intervention using the reduction recipes given in the frame headers. The co-added frames from each of the jitter sets were then combined using CCDPACK routines. In total, data were retrieved for the 16 FS stars from over 360 nights from 1999 to 2004, with around 10,000 individual frames (10Gb) retrieved and processed in about a week. Calibration of these fields is of course trivial due to the presence of the FS.

Figure 2: The (J-K) and (H-K) colours for galaxies and stars in our catalogue. The stars show a tight locus and are generally bluer than the galaxies, which have a broader colour distribution. We plot two model redshift tracks for galaxies with spectral energy distributions comparable to present-day ellipticals or Scd galaxies. These start off blue at z=0 and become redder at higher redshifts; we mark z=2 on both tracks. The distribution of galaxy colours broadly spans that expected for passive or star-forming galaxies at z~0-2. The dividing line at (J-K)=2 roughly corresponds to galaxies at z~2. We show the median errors for the colours of our objects in the lower-right- hand corner.

The mean exposure times for the 16 fields are 3.2ks, 1.7ks and 3.7ks in J-, H- and K-band respectively. This is sufficient to reach 5-sigma point-source limits of J~21.4, H~20.3 and K~20.3 - adequate to detect L* galaxies at z~1 and beyond (Fig. 1). The seeing is 0.70", 0.67" and 0.57" in J-, H- and K-bands. As these are measured from the co- added images from observations spread over 5 years they can be taken to represent firm upper limits on the typical seeing achieved by UKIRT.

Catalogues were constructed from the K-band frames using SExtractor within GAIA. In total we detect 24 stars (excluding the FS stars) and 87 galaxies objects within 25" radius of the FS and brighter than the 5-sigma limits, K=19.7-20.7, over a surveyed area of 8.7 sq. arcmin. In Fig. 2 we show the (J-K)-(H-K) diagram for the stars and galaxies. We plot model tracks for the colour of galaxies as a function of redshift which show that those with the reddest (J-K) colours are likely to lie at z~2. Focusing on the 12 (J-K)>2.0 galaxies, we find that they are all well- resolved with FWHM (corrected for seeing) of ~1.4" (~10kpc at z~2) and have a median magnitude of K=19.0. These galaxies are well-situated for AO-assisted studies with a median separation from the FS star of 13.0".

Acknowledgments: We thank Malcolm Currie, Brad Cavanagh, Peter Draper, Jim Lewis, Nigel Metcalfe and Mark Swinbank for help. NMC acknowledges support from a Summer Studentship at Durham University. IRS acknowledges support from the Royal Society.

UIST: Leading the Way with IFU Observations of Submillimetre Galaxies

Scott Chapman¹, Chris Lindner¹, Mark Swinbank², Ian Smail², Rob Ivison³ & Andrew Blain¹

¹Dept. Astronomy, California Institute of Technology, USA
²Institute for Computational Cosmology, Durham University, U.K.
³ATC/IfA, University of Edinburgh, U.K.

The study of submm-luminous galaxies (SMGs), first discovered using SCUBA on the JCMT (Smail, Ivison & Blain 1997), has progressed rapidly in the last two years with the combination of deep optical imaging from HST (Chapman et al. 2003a; Smail et al. 2004), deep X-ray imaging and spectra (Alexander et al. 2005), UV and near- IR spectroscopy from 10-m telescopes (Chapman et al. 2003b, 2005; Swinbank et al. 2004), and the subsequent follow-up in rotational CO emission lines to probe the state of the molecular gas (Neri et al. 2003; Greve et al. 2005). From these observations, we are beginning to understand the processes which trigger the immense bolometric luminosity output in these galaxies, allowing us to address their true contribution to the star-formation rate history of the Universe.

To probe the dynamical structures of these frequently complex systems requires a reliable separation of the spatial and spectral information. The UIST IFU is an ideal tool to accomplish this task. 2-D spectroscopic maps from UIST allow us to trace the dynamical and structural properties of SMGs on scales of a few kpc, pin-point the sites of active star formation and identify non-thermal emission from active galactic nuclei (AGN). These observations allow us to directly understand the rapid evolution of SMGs, testing the claims that far-infrared luminous galaxies comprise merging systems which are likely to be the progenitors of local massive ellipticals, or whether instead they are simply high-luminosity episodes in the history of less massive galaxies.

Figure 1: HST image with UIST IFU data overlaid as contours, illustrating the Hα velocity field across ELAIS N2.850.4 (Swinbank et al. 2005).	Figure 2: HST image with UIST IFU contours overlaid showing the distribution of Hα emission in ELAIS N2.850.7.

Seven SMGs were identified as bright submm/mm sources from various survey fields, with radio counterparts (Chapman et al. 2002; Ivison et al. 2002; Greve et al. 2004) providing precise positions which enabled spectroscopic redshifts to be procured from Keck LRIS and NIRSPEC observations (Chapman et al. 2005; Swinbank et al. 2004). In addition, three sources selected at 24 microns with the Spitzer Space Telescope were followed up as galaxies with potentially comparable bolometric luminosity to the SMGs, from the surveys of Yan et al. (2005) and Borys et al. (2005). All ten of these targets were observed with the UIST spectrograph using the HK grism and exposure times of 2-8 hours each. All targeted sources yielded detections in the Hα (and sometimes [OIII] 5007) emission line, at the targeted radio position and in most (7/10 cases) additional components were identified offset spatially from the radio/submm position (Swinbank et al. 2005, 2006). The observations represent a triumph of efficiency for the UIST spectrograph, given that all spectroscopic studies to date of these galaxies have been conducted with 10m-class telescopes.

Deconvolving the exact contributions from the star formation and AGN activity within these galaxies is difficult using just the UV slit spectroscopy from Chapman et al. (2005). Ideally we need to use the well-developed spectral indicators based on rest-frame optical emission lines (Veilleux & Osterbrock 1987) falling in the near-IR for these high-redshift galaxies. Coupling the spatial coverage from the UIST IFU with coverage around the the rest-frame optical emission lines allows us to locate and isolate the components hosting the AGN in these systems, determine dynamical masses, and search for extended halos and/or companions.

Figure 3: A comparison of the Hα emission from the UIST IFU observations with CFHT I-band imaging of SA22-96.

The UIST observations have enabled the remarkable discovery that many, potentially the majority, of the SMGs reveal additional kinematic components and extensions. These companions exhibit luminous Hα emission, indicative of star formation in excess of 100 M-solar/yr, in structures up to ~20 kpc from the submm burst. Figures 1-3 illustrate several examples of SMGs surveyed in our UKIRT/UIST program (from Swinbank et al. 2005, 2006), all exhibiting multiple and extended structures on the scale of the UIST field of view (~5 arcsec, ~40 kpc). With typical component velocity separations of ~300 km/s, our estimates for the dynamical masses of the systems (assuming the components are in a state of merging free- fall) are of order 5e11 M-solar.

These observations provide clear support for the proposal that SMGs represent merging/interacting systems with high instantaneous star formation rates and actively fueled AGN. The multi-component nature of these galaxies indicates that they may be analogous (but scaled up) versions of local Ultra-Luminous infrared galaxies. Furthermore, the velocity offsets and line widths from the resolved components in these galaxies are consistent with those seen in local luminous ellipticals, furthering the case that these sources represent the most active formation phase in massive galaxies.

Acknowledgments: We thank Andy Adamson, Thor Wold and Tim Carroll for their superb support with this program.

REFERENCES:
Alexander, D., et al., 2005, Nature, 434, 738
Borys, C., et al., 2005, ApJ, in press
Chapman, S., et al., 2002, ApJ, 570, 557
Chapman, S., et al., 2003a, ApJ, 599, 92
Chapman, S., et al., 2003b, Nature, 422, 695
Chapman, S., et al., 2005, ApJ, 622, 772
Greve, T., et al., 2004, MNRAS, 354, 779
Greve, T., et al., 2005, MNRAS, 359, 1165
Ivison, R., et al., 2002, MNRAS, 337, 1
Neri, R., et al., 2003, ApJL, 597, 113
Smail, I., Ivison, R., Blain, A., 1997, ApJ, 490, L5
Smail, I., et al., 2004, ApJ, 616, 71
Swinbank, M., et al., 2004, ApJ, 617, 64
Swinbank, M., et al., 2005, MNRAS, 359, 401
Swinbank, M., et al., 2006, MNRAS, submitted
Veilleux, S., Osterbrock, D., 1987, ApJS, 63, 295
Yan, L., et al., 2005, ApJ, in press