Astron. Astrophys. 330, 990-998 (1998)

2. Observations

2.1. Optical and near infrared observations

Table 1 presents the details of the observations in the six wavebands. The observations in the V, R and I bands were carried out during June 1992 at the 3.6m CFHT (Canada France Hawaii Telescope) on Mauna Kea, Hawaii, using the FOCAM instrument with the RCA4 1024² CCD detector. Flat fields have been obtained in each photometric band on the sky during twilight. The CCD electronics offset was measured several times during the night. For each photometric band, various standard stars were repeatedly observed during the night.

Table 1. Log of observations in the V, R, I, J, H and K bands

At the F/8 focus of the telescope, the CCD pixel scale was 0.21". The observations in the J and H bands have been obtained in November 1992, also at the CFHT, using the visitor "MONICA" Nicmos 3 infrared camera (Nadeau et al. 1994) at the same F/8 focus. At these wavelengths, the resolution is 0.25" per pixel. The observations in the K band have been obtained on September 1992 at the 2.2m telescope of the ESO observatory of La Silla, Chile, using the common-user IRAC2a CCD camera with a resolution of 0.49" per pixel. In the optical bands, two fields slightly overlapped were observed, surveying an area of 16 square arcminutes. In the three NIR bands, sixteen fields were observed toward the Serpens cluster, approximately covering an area of 12 square arcminutes in J and H, and 19 square arcminutes in K. These fields were arranged in a 4x4 mosaic centered on the Serpens Reflection Nebula ( = 18h27m22s, = 1^o 12'30"). The fields were spatially overlapped by 30" in both right ascension and declination, allowing an accurate positioning of the mosaicked fields. Per image, the integration times used were of 10 minutes in the optical bands, five seconds in J and H, and two seconds in K, allowing the quoted sensitivity limits presented in Table 1.

2.2. Data analysis

The NIR data were reduced by first subtracting from each data frame a median filtered sky frame obtained from five nebulosity-free frames, observed immediately before and after the target observation. The J and H band images were then flat-field and distorsion corrected with a dedicated software. Finally, the images were mosaicked together. Nominal atmospheric extinctions for Mauna Kea are , per air mass, and nominal atmospheric extinction for ESO La Silla is per air mass. All infrared images were air-mass corrected.

Data analysis was done with standard Image Reduction and Analysis Facility (IRAF) and Interactive Data Language (IDL) routines. As a first step, several isolated stars of different intensities were chosen manually to determine the Full Width at Half Maximum (FWHM). Thus, for each image, source extraction and aperture photometry were performed using DAOPHOT (Stetson, 1987), and the routine DAOFIND was used to extract stellar-like sources whose fluxes were significantly above the background (that is, sky) noise in each image. The results were visually compared to the images at several contrast levels to ensure that spurious identifications were minimized. Such spurious detections were a problem in the area of the bright reflection nebula where probably non stellar emission knots could be interpreted as stars. We removed all these spurious sources from our data sample listed in Table 2. This led undoubtedly to the non-detection of faint sources in the area of the image where contamination from extended emission was present. In addition, sources that were not bright enough to be detected by the finding routine, but visually identified as stars, were appended to the coordinate list. Finally, the resulting images were mosaicked with an IDL routine. This routine eliminated bad pixels and adjusted the relative background level of overlapping frames to a common value. Using these procedures, 5, 12, 20, 44, 86 and 138 stars were found in the V, R, I, J, H and K mosaic images, respectively. Aperture photometry was performed for all the extracted stars in each image. Fluxes were determined for each star with the size of the software aperture used varying along with the brightness of the source (the brighter the source, the larger the software aperture).

Sky levels were determined around each star in a 5-pixel wide annulus. Sky levels were also obtained for annuli with smaller inner radii and larger outer radii with no significant change in the resulting stellar fluxes. Photometry was performed manually for stars which were confused with nearby nebulosity or other stars. For objects with associated NIR nebulosity, we considered that the true stellar flux is represented by the signal in the object aperture minus the contribution from the average sky plus nebulosity in the sky annulus. In the K band, a cross-correlation between our survey and the one of Eiroa & Casali (1992) on the brighter common sources has been used to assess and coordinates of all our sources. Positions for the cluster members observed in the other wavebands were then directly derived from the K mosaic image by applying the same procedure.

2.3. Magnitude uncertainties and sensitivity limits

For each detected source, we evaluated the magnitude uncertainty by , where is the stellar flux in the source aperture calculated with N pixels, and the standard deviation of the sky annulus. Thus, the final magnitude is given by:

where , , and

In order to determine the sensitivity limits, we considered a 3 detection limit, being the standard deviation of the sky calculated on nebulosity- and star-free regions. A star can be approximated by a gaussian whose parameters are the position centre, the Full Width at Half Maximum along the two axes ( and ), the stellar luminous intensity (I) and the sky level. Thus, the stellar flux is given by:

where .

In this case, and , leading to the values quoted in Table 1.

© European Southern Observatory (ESO) 1998

Online publication: January 27, 1998
helpdesk.link@springer.de