3. CCD observations of parallax candidates
The use of CCD frames for parallaxes determination have shown a growth during the 10 last years. Monet & Dahn (1983), Tinney (1993), Tinney et al. (1995) showed the advantages of CCD observations for parallaxes determinations especially for faint objects.
3.1. CCD observations
Observations were carried out at 4 epochs (Aug 1994, Mar 1995, Aug 1995, Feb 1996) at the maximum of the parallactic angle of our targets. The telescope used was the 1.23m Telescope of Calar Alto, equipped with the RC Cassegrain camera in direct imaging. The CCD detector was a Tektroniks chip (TEK#6), with 24 µm pixels giving an image scale of per pixel and a field of view of . All observations were made through a R filter due to the probable nature of faint nearby objects.
Due to restricted number of allocated nights we could observe only 20 of our 32 targets. These 20 candidates are located in the northern part of the Schmidt field. The candidates of the southern part remain to be observed. Each of the targets, including a nearby star (Gliese 1986, star G180-060) with known trigonometric parallax, were observed typically 12 times (see Table 5) at hour angles between -2 hours and hours from meridian. Exposure times were between 10min and 20min depending on the target luminosity.
We give in Table 4 the coordinates of these candidates (positions at the date in equinox B1950, obtained from the Schmidt master plate).
3.2. Astrometric reduction
Each CCD frame has been measured in the following way : images of each object of the field has been centroided using DAOPHOT (Stetson 1987). The ALLSTAR routine has been used for this. Point spread functions (PSF) were fit by hand using PSF routine. A Moffat function analytic approximation to the observed PSF was used which allows for elliptical images an arbitrary orientation. Finally for each frame we had the (x,y) position and the internal magnitude of each detected object (typically 30 per frame).
The frames were cross-identified to an arbitrary master frame and objects not present on a minimum number of frames were excluded. In what follows we will consider all these common stars (excluding the target star) as reference stars.
The astrometric reduction itself is identical to the one described for the treatment of Schmidt plates. Nevertheless due to the necessary accuracy that we wish to reach it was no more possible to ignore the singularity of the solution. The constraint that we added was the following one: after each iteration each object with a parallax non significant at a 3 level with respect to the field mean parallax is considered to be located at the infinite: in the system of equations, the parallax of this object will not be an unknown, it is a priori set to zero.
Many recent works (Monet 1983, Tinney 1993) showed that Differential Color Refraction (DCR) must be taken into account for the determination of accurate parallaxes. Unfortunately our observations of each source were not spread over a large enough number of hour angles to model these effects. Our goal here was mainly to confirm the proximity of the stars selected from Schmidt observations which can be achieved without these DCR corrections. The effects of DCR are sufficiently small not to change our conclusion but may modify the values of the parallaxes. However it is clear that we will have to take them into account to determine more accurate parallaxes for our stars.
3.3. Results and discussion
We have treated in that way the 20 observed fields. We derived for any detectable object of each field its projected proper motions and its parallax. We present in Table 5 the results for the 20 target stars. We give in column 1 the identification of the objects, in column 2 the parallax in mas, in column 3 and 4 the projected proper motions in right ascension and in declination in mas per year, in column 5, 6, 7 and 8 the number of frames at each of the 4 epochs of observations.
For comparison purpose we also included in our program a brighter star with a previous determination of its parallax (Gliese et al. 1986). Our result, given in Table 6, is within from Gliese's one for the parallax. We first notice the good agreement between the two values which give us good confidence that the other parallaxes derived are reliable. We must nevertheless keep in mind that we did not apply DCR corrections which may explain a fraction from this discrepancy.
For 6 of the 20 targets a detectable parallax has been measured (identification charts are given in Fig. 7). The rms reached are rather low and only represent the internal errors. The analysis of the covariance matrix of the stellar parameters shows that the correlations between the estimates of proper motion and parallax is week despite the short time base (three years). These correlation coefficients are roughly 0.2 for all the target stars. We can therefore conclude that the effects of proper motions and of parallaxes have been well separated.
It is interesting to note that 3 of the parallax stars (P5, P15, P18) belong to large proper motion surveys. Their parallax value are ones of the largest of the sample. It is also interesting to notice that P1 which has a large proper motion is absent to our knowledge from large proper motions catalogs. Its parallax is large with respect to the two last parallax objects (P4 and P9). The star P3 which has no detectable parallax appeared to be a galaxy on the CCD frames. This explains why the accuracy on both and µ are so catastrophic.
One third of our selected candidates appear to be indeed nearby stars. Due to the poor accuracy reached with the Schmidt plates and the poor number of CCD frames this result was not obvious and must be considered as a full success, validating the method used in searching for nearby stars.
Aside from this, in some of our fields some very faint objects appeared also to behave as nearby stars. It is possible that these objects are unresolved binaries or extragalactic objects, although the criterion of DAOPHOT for separating stars from extragalactic objects labels them as stars. These objects, excessively faint (), will motivate a new program of observations to determine their nature.
© European Southern Observatory (ESO) 1998
Online publication: April 28, 1998