We are searching for companions with separations in the two decades from to to include the separation where the distribution of main sequence binaries has its maximum (, at the distance of Taurus). The lower limit is determined by the theoretical diffraction limit of the telescope on Calar Alto at a wavelength of . The upper limit is chosen so that the contamination by background stars has little effect (for a detailed discussion of this problem see Sect. 4.2). To make maximum use of the resolution of the telescope, the main observational technique for our survey is speckle interferometry, complemented by direct imaging to find companions that are outside the limited field of view of today's speckle cameras.
The speckle observations were carried out at the telescope on Calar Alto in September 1994, December 1994, and October 1995. We used MAGIC, a pixel NICMOS 3-camera (Herbst et al. 1993), at in its high-resolution configuration at the f/45 focus. This gives a pixel scale of /pixel and a field of view of . For our speckle observations, we usually use only one quarter of the array, unless we see a companion in the rest of the field. At each observing run, we measured a number of well-known binary stars to calibrate the pixel scale and position angle.
The modulus of the complex visibility (i.e. the Fourier transform of the object brightness distribution) is determined from power spectrum analysis, the phase is computed using the Knox-Thompson algorithm (Knox & Thompson 1974), and from the bispectrum (Lohmann 1983). If the object turns out to be a binary, we obtain the brightness ratio, separation and position angle of the components from a fit of binary models to the complex visibility. Fits to different subsets of the data give an estimate for the standard deviation of the binary parameters. This procedure sometimes yields unbelievable small errors, we estimate the minimal error of the separation to be about or , the minimal error of the position angle to be , and the minimal error of the brightness ratio to be 0.001. These might be slightly larger for the binaries with the smallest separations.
Otherwise, if the object appears to be unresolved, upper limits for the maximum brightness of an undetected companion are determined by computing the maximum brightness ratio of a companion that could be hidden in the noise of the data. For details of this procedure see Leinert et al. (1996).
The additional imaging has been done at different telescopes on Calar Alto, namely the telescope in January 1995, the telescope at the f/10 focus in January 1996, and the telescope in February 1996. Again, we used MAGIC at in its high-resolution configuration. The pixel scales were /pixel at the telescope, /pixel at the telescope, and /pixel at the 1.23 m telescope. Some images were taken at the telescope on La Silla in March 1996, using the IRAC2b camera (another pixel NICMOS-3 array) with its objective "B", which gives a pixel scale of /pixel. Some of the imaging observations could also be used to obtain infrared photometry of the stars.
© European Southern Observatory (ESO) 1998
Online publication: March 3, 1998