Resolution of High Redshift QSOs with AOB

J. B. Hutchings, S. L. Morris a,d. D. Crampton

Dominion Astrophysical Observatory
Herzberg Institute of Astrophysics, NRC
Electronic-mail: john.hutchings@hia.nrc.ca

Abstract:

We have used the AOB to obtain high resolution images of QSOs at redshift $\sim$1 and higher. While signal levels are low with small pixels, we have resolved at least half the host galaxies in early processing, and in many, we detect faint tails or arms that imply tidal interactions. In the higher redshift QSOs the host galaxies have compact or faint old stellar populations, and appear to be dominated by young stellar populations.

Observations

Data were obtained in 3 observing runs: Jan 1997 with Monica and STIS CCD; Jan 1998 with KIR; March 1998 with KIR. Observing conditions and data were significantly better in the 1998 runs.

QSOs observed had guide stars of 15 mag and brighter within 20 arcsec. In a few cases (1334 and 3C273 and OJ 287, not reported here) it was possible to guide on the QSO nucleus itself. Dither patterns of several arcsec were executed to enable data frames to be used to generate sky frames, and to eliminate the effects of bad pixels or cosmic rays. In most cases the guide star was observed for short exposures for PSF modelling. Photometric standards were observed and crowded fields were observed to allow modelling of the PSF with distance from the guide star, and the guide star brightness.

More detailed analysis will involve PSF modelling and off-axis degradation for subtraction and deconvolution, plus modelling of the observed colours. Processing to date includes sky removal and flat fielding, after shift and addition of the dithered images, and comparison with the PSF nearest in time, usually the QSO guide star.

The image quality generally obtained depended on the nightly conditions, the guide star magnitude and offset from the QSO, and the wavelength. We used broad-band filters I, J, H, K variously for the observations, with more than one colour for each object. Filters chosen sometimes include strong emission lines: for most they are selected to distinguish stellar populations at the QSO redshift. Images generally had FWHM in the range 0.1 to 0.15 arcsec, with some worse in bad conditions. High Strehl rations were obtained in the H and K data in the best conditions.

NIR Imaging Results

The table summarizes the results to date.

QSO Mag z Radio Comment

1334+246 15 0.11 RQ Disturbed disk galaxy with knots

0915-213 17.5 0.85 RL Resolved host, knot, tail, compact cluster

1055+019 20.5 1.06 RQ Resolved, faint jet or extended companion

$$\sim$$ 0804+499 17.5 1.43 RL

1337-013 18.7 1.61 RL Incomplete reduction: tail?

1540+180 18 1.66 RL Resolved, tail, compact cluster

$$\sim$$ 1236-003 19.1 2.18 RQ

0104+022 19.7 4.16 RQ Incomplete reduction, unresolved?

Our sample selection is strongly determined by the date of observation and the QSOs which have a suitable guide star. Thus, we have a small, but unbiased sample of possible targets, but have concentrated mainly on those at redshifts near and above 1.0, where little is known about QSO hosts.

Our results demonstrate the following:

1.

AOB imaging of QSOs is feasible and productive with proper procedures.

2.

With the small pixels in the IR detectors, signal levels are very low. It is necessary to observe for about 2 hours per filter to get significant results. The larger pixels in the CCD are more suited to this low-light-level work, but on-chip binning is not possible in the NIR. We frequently used binned images in the analysis, as shown.

3.

There are two kinds of positive detections: resolution of small bright details in the field, and detection of faint smooth flux near the nucleus with high suppression of scattered nuclear light. The latter is more common in the NIR, where the signal is dominated by old stellar populations with little obvious structure.

Results

We have the following preliminary scientific results.

1.

In the higher redshift QSOs we have studied, the host galaxies are generally compact, smooth, and faint in the NIR. Structure that is seen in faint tails which must arise in tidal encounters or mergers. These tails are present in both radio-loud and radio-quiet objects, but may be more obvious or frequently found in radio-loud ones. There appears to be an old population in QSO hosts at redshifts 2 or more.

2.

Structure becomes more evident at shorter wavelengths (J or I bands), where younger stars and dust are more dominant. This is further suggested by the detection of very luminous structure in blue/visible bands with HST and non-AO ground-based observations. Thus, there are probably both young and old populations in high z QSOs, indicating fresh star-formation that accompanies the QSO episode.

3.

Several of the high redshift QSOs in our program, and others observed with HST are in very dense groups of galaxies. This suggests that merging is a major trigger of QSO activity at all redshifts, although the evolutionary state of the merger components must evolve strongly with redshift.

4.

This work needs to be extended to visible wavelengths to study the young stellar populations associated with the QSO host galaxies and their companions. The new PUEO beamsplitters and ability to bin on-chip will make the visible band observations significantly more efficient.

5.

The program needs to be continued to build up statistics on different redshifts and the difference between radio-loud and quiet QSOs.

We have published our early results in PASP 110, 374, 1998. In that paper we also show some PSF models and fits. Further papers are in preparation. We would like to thank CFHT staff for their excellent support of the instrument and our observations.