Astron. Astrophys. 322, 455-459 (1997)

2. Observations and reductions

All observations for this program were obtained at the 2.56 m Nordic Optical Telescope (NOT), La Palma. The observed area covers 180 in IJK and is centred at RA , Dec (J2000.0).

The I data was taken with BROCAM1 (TEK1K), operated in cassegrain focus with 0.176"/pixel and 3'x3' field. For future proper motion measurements, the astrometry errors due to nonuninform pixels and possible effects of rotator position were reduced by using four different position angles for each I-field (0, 90, 180 and 270 degrees) exposing 5 min at each position. Debiassing and flatfield corrections were done in a standard fashion within IRAF (Image Reduction and Analysis Facility) ¹. Median seeing was ", varying from 0.46" to 0.80".

The JK observations were done with the ARNICA NICMOS3 (256x256) array, which was made available at NOT via a collaboration with Arcetri Astrophysical Observatory, Florence. The last third of 6 nights in Aug-Sep95 was used. Unfortunately the pixel size (0.55") was not very well matched to the seeing () and all images were undersampled. Another problem was that focus changed across the field, which due to astigmatism in the NOT optics led to stars of slightly different elongation across the field. Apparently this was caused by the chip not being perpendicular to the optical axis which could be seen as a slight change in pixel scale across the field.

The J data could be treated the same way as I, but high and rapidly varying background in K necessitated a different approach. The sky for each K -image was defined as the median of the four images nearest in time and subtracted from the image, which then was flatfielded by a differential flat.

2.1. Classification of objects

A major problem in this kind of survey is to distinguish stars from distant galaxies at faint magnitudes. The whole area was visually inspected and all objects were classified as stars, galaxies, binaries or unclassifiable (too faint for a meaningful classification). 1411 objects out of 3800 were classified as stars. Since the seeing conditions were excellent it is estimated that the discrimination between stars and galaxies is reliable to .

2.2. Photometry

Since point spread functions (PSFs) were undersampled in JK it was not suitable to use ordinary PSF fitting. For single aperture photometry it is necessary to always keep the same aperture radius. However for the JK data centering errors and focus shifts are non-negligible, thus in order to maximize the signal to noise ratio it was desirable to change aperture radius from one object to another. A new, more robust, method (used also in I) was developed, brief outlines:

1. All objects were measured in a series of 80 apertures, with radii ranging from 0.05" to 4" in steps of 0.05".

2. From the "best" stars in each field a growth curve (Stetson 1990) was constructed, i.e. magnitude shift as function of aperture radius.

3. Each objects' curve was subtracted by that frames' growth curve. Ideally this curve is completely flat, but due to centering errors, errors in the sky level etc... that is generally not the case. However, there is normally an almost flat part of the curve at FWHM. The magnitude difference was taken as the mean level of that part. The advantage with this method is that the right mean level will be found even if objects are differently focussed or elongated as compared to the reference growth curve.

Transformation of JK instrument magnitudes to the CIT system was done by observations of IR standard stars and transformation equations provided by the ARNICA team (Hunt et al. subm.). I instrument magnitudes were transformed to the Kron-Cousins system via observations of Landolt (1992) standard stars. Typical internal errors are 0.1 mag ( ; ; ) and 0.02 mag ( ; ; ). The zeropoint errors are 0.05 mag in J and 0.08 mag in K. In I they are estimated to be less than 0.02 mag. Zeropoint errors are not included in Table 1 and Table 3.

Table 1. Photometry of potential Pleiades members. Magnitude errors are internal.

Table 2. The number of Pleiades candidates in the magnitude interval from five different surveys and as expected from four different IMF-indices. The rightmost column is scaled to 500 . Normalization of the IMF was deduced from proper motion members in (Hambley et al. 1993). Statistical errors are within the parenthesis.

Table 3. M-dwarf candidates, including the Pleiades candidate NOT1. Magnitude errors are internal. Distance errors are of the order of 10%.

All objects recognized as binaries or stars very close to a galaxy in I were, as a check, also measured by PSF-fitting. For a few objects that were resolved in I, but not in JK, the whole system was measured as one object in all bands, thus getting the binary system correct in the colour magnitude diagrams.

2.3. Completeness limits

In this kind of surveys it is important to know the magnitude limit to which the survey is complete. The number of detected stars as a function of magnitude increases exponentially and gives a straight line in a log() vs mag plot as long as the survey is complete. The point at which the curve turns away from the straight line should be regarded as the completeness limit. Thus , and are the completeness limits of this survey (magnitude error 0.08 mag). The 50 % limit is 0.7 mag fainter.

© European Southern Observatory (ESO) 1997

Online publication: June 5, 1998

helpdesk.link@springer.de