Astron. Astrophys. 346, 243-259 (1999)

Appendix A: narrow band calibration

The narrow-band filters for IRAC2b are relatively broad ( µm), thus the line and corresponding continuum filters share a large overlapping region. For calibration, two strategies were used. The counts inside the standard's PSF were determined by using an IDL adaption of the DAOPHOT package by W. Landsman. Then the actual flux of HD169588 at the desired wavelength was interpolated from the H and fluxes given by Carter & Meadows (1995). Each of the object frames was then multiplied by the ratio of flux-to-counts at the corresponding wavelength. For flux measurements, we also did a continuum subtraction. As the broadness of the filters leads to a serious contamination of the continuum with line emission, we had to take care of this. The flux per pixel in a line image is given by:

is the measured Br flux in the image. The symbol "" means "at the point of". Thus means the transmission of the line filter at the wavelength of the line. As the images are already flux calibrated, transmissions are measured relative to the transmission of the line filter at the line's wavelength (i.e. is 1 in this case). is the flux of line photons contributing to the total flux in the pixel, while denotes the continuum flux. The latter can of course also be present at different wavelengths while the line is assumed to be so narrow that it virtually only exists at . The flux in the continuum filter should be given by:

When assuming and taking , we can combine this to

These equations ignore the fact that only flux densities are measured in finite filter widths, but since these widths are identical for all the filters, the effect can be neglected. The factor gives the ratio of the continuum photon number at the two wavelengths considered. It was determined by simultaneously minimising the fluxes of 5 selected field stars. To determine the Br flux density, the procedure was performed for two continuum filters (NB4 and NB7) with variations in the resulting Br flux below 4%. Together with another 4% uncertainty from background noise in the resulting image and approximately 2% from the calibration using the standard, we end up with a total flux uncertainty of 6%. For the line, only NB4 was used for calibration, because residual Br flux in NB7 rendered the result useless.

© European Southern Observatory (ESO) 1999

Online publication: May 6, 1999
helpdesk.link@springer.de