Forum Springer Astron. Astrophys.
Forum Whats New Search Orders

Astron. Astrophys. 364, 517-531 (2000)

Previous Section Next Section Title Page Table of Contents

2. Sample selection and photometry

The Hubble Deep Field North has been observed in 4 broad bands (F300W, F450W, F606W, F814W) for a total of 150 HST orbits by Williams et al. (1996), and constitutes the deepest ever exposure on a small sky area. Accurate photometric data in the four bands have been published for thousands of faint galaxies by the authors.

Connolly et al. (1997) observed the HDF-North in the near-IR with the IRIM camera on the KPNO 4 m telescope. The camera employs a 256 [FORMULA] 256 NICMOS-3 array with 0".16 pixel-1, but the released images were geometrically transformed and rebinned into a 1024 [FORMULA] 1024 format. IRIM exposures have been secured in the J, H and K filters, for a total of 12, 11.5 and 23 hours, respectively. Formal 5 [FORMULA] limiting magnitudes for the HDF/IRIM images, computed from the measured sky noise within a 2" diameter circular aperture, are 23.45 mag at J, 22.29 mag at H, and 21.92 mag at K, whereas the image quality is [FORMULA] 1".0 FWHM.

Our sample of galaxies has been extracted from the HDF/IRIM K-band image through a preliminary selection based on the automatic photometry provided by SExtractor (Bertin & Arnouts 1996). It is flux limited in the K band and it excludes early-type galaxies, i.e. objects whose surface brightness distribution is dominated by a de Vaucouleurs profile.

To determine the limit of completeness in the K band (hereafter [FORMULA]) for inclusion in our sample, we followed the same empirical procedure described in FA98. Briefly, a large number of toy galaxies with exponential profiles (produced with the IRAF-MKOBJECTS tool) have been used to check the performances of Sextractor in estimating the K-band magnitudes ([FORMULA]) of late-type galaxies. This allowed us to determine the magnitude [FORMULA] below which the scatter [FORMULA] of the measured magnitudes turns out to be lower than some given value [FORMULA] (the vertical line in Fig. 1a corresponds to [FORMULA] and [FORMULA]). Moreover, using only galaxies with [FORMULA], we have derived the following empirical relation between the bias [FORMULA] and the effective surface brightness [FORMULA] (see Fig. 1b):


where [FORMULA]. This relation provides the true total K magnitudes from the SExtractor measured flux. Then simulations have shown that the K-band image has a surface brightness limit of [FORMULA]

[FIGURE] Fig. 1. Left panel: standard deviation of the SExtractor magnitude estimates as a function of the average magnitude for galaxies in the simulated images: the standard deviation is very small for [FORMULA]. Right panel: difference between true flux and the SExtractor flux as a function of the average surface brightness for the subsample of simulated galaxies with [FORMULA].

By analogy with FA98, we have first produced a catalog of morphologically selected late-type objects with [FORMULA]; then we have used effective radius estimates from high resolution (HST) optical imaging to derive the effective surface brightness of galaxies ([FORMULA]); finally, we have applied to the [FORMULA] magnitudes the statistical corrections given by Eq. (1) and we have included in the final sample only galaxies with corrected magnitudes less than or equal to [FORMULA] mag (we assume [FORMULA] from Fig. 1b).

A total of 176 objects with [FORMULA] mag were detected by SExtractor in the IRIM K-band image. After careful inspection of the high resolution HST images, we rejected all elliptical and S0 galaxies (including the 35 early-type objects identified in FA98) and the stars. A few objects were also rejected from the sample due to their position in the frame (at the edges of the image the noise is higher and the magnitude estimate is likely to be uncertain). We then produced a first preliminary, incomplete sample of late-type galaxies.

The effective radii [FORMULA] were estimated running SExtractor on the WFPC2 [FORMULA] frame with the parameter [FORMULA], providing the radius containing half of the total emitted flux. The surface brightness [FORMULA] was evaluated for each galaxy and the statistical corrections [FORMULA] were computed using the Eq. (1).

The final complete sample of late-type galaxies with [FORMULA] consists of 52 objects over the HDF area of 5.7 square arcmin. For 36 objects we have the spectroscopic redshift (Cohen et al. 1996; Fernandez-Soto et al. 1998), while for the remaining 16 we measured it from our photometric analysis as described below.

A procedure analogous to that outlined for the K band was used to derive, for each object of the selected sample, the corrected magnitudes in the J and H bands.

The optical magnitudes in the F300W, F450W, F606W and F814W bands (U, B, V and I in Table 1, respectively) have been computed again with SExtractor on the high resolution WFPC2 images (no corrections being applied in this case). Magnitudes are in the AB system, defined by the relation (Oke & Gunn 1983):


where [FORMULA] is the flux in [FORMULA], the constant beeing choosen so that [FORMULA] for an object with flat spectrum.


Table 1. Photometric data on the sample galaxies

Some data on the sample are listed in Table 1. Column 1: our identification; Columns 2-4: coordinates [FORMULA] and [FORMULA] (at J2000). To these must be added 12 hours 36 minutes (RA) and 62 degrees (Dec); Column 5: [FORMULA] is the effective radius, in arcsec, derived from HST images; Columns 6-9: optical U, B, V, I magnitudes in the AB system (see above); Columns 10-12: near-infrared J, H, K corrected magnitudes in the standard system; Column 13: redshift of each object. Values in brackets are photometric redshifts, while the other are all spectroscopic.

Previous Section Next Section Title Page Table of Contents

© European Southern Observatory (ESO) 2000

Online publication: January 29, 2001