All Figures

$\begin{figure} \par\includegraphics[angle=270,width=14.7cm,clip]{8493fig1.ps} \end{figure}$ Figure 1: PA of the companion relative to GQ Lupi A versus time. a) Absolute PA values with their absolute errors for all NACO observations (Table 2). b) Change in PA since May 2005 with relative errors, the error includes orbital motion of the HIP 73357 calibration binary since 2005.4 only (not since 1991.25 as in the left panel). We also show the data point at the first new epoch 2005.4, set to 0.0, with its error bar just from the Gaussian centering fit. For both a) and b), the straight dotted line is for constant PA, as expected if both objects are bound. To allow for orbital motion, the maximum change in PA (if pole-on orbital plane) is indicated by dashed lines. The data stay constant or vary slowly within those limits. There is no strong evidence for orbital motion in PA; the first data point from 2004 should have lower weight here, because it was obtained with different astrometric calibration binaries than in 2005 to 2007. The full wobbled lines with strong positive slope indicate the change in PA, if the companion was background ( $\pm$ expected errors from errors in proper motion and parallax in panel a); this hypothesis can be rejected with large significance, because the data points strongly deviate from this hypothesis.
Open with DEXTER

In the text

$\begin{figure} \par\includegraphics[angle=270,width=14.5cm,clip]{8493fig2.ps} \end{figure}$ Figure 2: Separation between GQ Lupi A and its companion plotted versus time. Data are from Janson et al. (2006) for 1994, N05 for 2002, and this paper for the rest. Constant separation is indicated as a straight dotted line (through 2005.4 value), separation changes due to maximum expected orbital motion are indicated by dashed lines. All values are well within the expectations for orbital motion. a) Absolute values from 1994 to 2007. Here, we also indicate as a wobbled line with strongly decreasing values the expected separation change for the background hypothesis, computed from the known GQ Lupi A proper motion and parallax and assuming that the faint object has negligible proper and parallactic motion ( $\pm$ errors as expected from proper motion and parallax errors as additional wobbled lines); the data are strongly deviant from the background hypothesis. The full line from 1999 to 2007 with smaller negative slope is the best fit for the orbital motion ( $\sim$ 2 to 3 mas/yr separation decrease from a linear fit to the last six data points), see text. b) Separation change since May 2005 with relative errors (including the data point from the first new epoch 2005.4, set to 0.0, with error bar just from Gaussian centering fits): the data points from Feb. and May 2006 are deviant from the dashed constant line (for constant separation) by 4.2 and 2.6 $\sigma$ , respectively, significance for orbital motion. The wobbled line again is for the background hypothesis. c) Separation in pixel with absolute errors (since June 2004). The full line with negative slope is our best fit for orbital motion as in a); this fit has a reduced $\chi ^{2}$ of 0.250 with 5 degrees of freedom and, hence, a probability of 0.94 (free fit parameters being starting separation and slope of decreasing separation). The wobbled line here is for assuming that GQ Lupi A and the faint object would have different parallaxes (different by 1.5 mas as best fit); this fit has a reduced $\chi ^{2}$ of 0.221 with 4 degrees of freedom and, hence, a probability of only 0.92 (free fit parameters being starting separation, slope of decreasing separation, and difference in parallaxes between GQ Lupi A and its co-moving companion). Hence, the full line (for identical parallaxes for GQ Lupi A and companion) is a better fit than the wobbled line. This gives evidence against the hypothesis that the deviations in separations observed, in particular in 2005.6 and 2006.4 (compared to 2005.4), are due to different parallaxes for GQ Lupi A and its companion.
Open with DEXTER

In the text

$\begin{figure} \par\includegraphics[width=7cm,clip]{8493fig3.eps} \end{figure}$ Figure 3: Proper and parallactic motion of GQ Lupi A in right ascension ( top) and declination ( bottom) relative to the background object GQ Lupi/cc2. Changes in separations in mas (since 2004.5) are plotted versus observing epoch as Julian date JD. We also show our best fit yielding a parallactic wobble of $7.4 \pm 1.9$ mas, from which we have to subtract the probable parallax of the background object cc 2 ( $1.0 \pm 0.3$ mas, see text), thus obtaining $6.4 \pm 1.9$ mas or $156 \pm 50$ pc for GQ Lupi A.
Open with DEXTER

In the text

$\begin{figure} \par\includegraphics[width=7cm,clip]{8493fig4.eps} \end{figure}$ Figure 4: Proper and parallactic motion of GQ Lupi b in right ascension ( top) and declination ( bottom) relative to the background object GQ Lupi/cc2. Changes in separations in mas (since 2004.5) are plotted versus observing epoch as Julian date JD. We also show our best fit yielding a parallactic wobble of $8.2 \pm 2.1$ mas, from which we have to subtract the probable parallax of the background object cc2 ( $1.0 \pm 0.3$ mas, see text), thus obtaining $7.2 \pm 2.1$ mas or $139 \pm 45$ pc for the GQ Lupi companion.
Open with DEXTER

In the text

$\begin{figure} \par\includegraphics[angle=270,width=13.7cm,clip]{8493fig5.ps} \end{figure}$ Figure 5: $K_{\rm s}$ band magnitude versus time for difference between HIP 73357 A and B ( bottom panel), GQ Lupi A ( 2nd from bottom), GQ Lupi companion ( 2nd from top), and the background object GQ Lupi/cc2 (top panel). We always show as a long-dashed line the mean value. In the case of GQ Lupi A, we also show the data points from Broeg et al. (2007), for which absolute photometry is available (C. Broeg, priv. comm.) as small dots in April 2005; we also show as dotted lines the range in variability ( $\pm$ 0.22 mag) found by Broeg et al. (2007) during their monitoring of GQ Lupi A (yielding the rotation period to be $\sim$ 8.4 days). We find variability in GQ Lupi A (amplitude $\pm$ 0.20 mag, 4 $\sigma$ ).
Open with DEXTER

In the text

$\begin{figure} \par\includegraphics[width=6.3cm,clip]{8493fig6.eps} \end{figure}$ Figure 6: Deepest image of GQ Lupi (102 min total exposure) after subtraction of PSF of GQ Lupi A. North is up, east to the left. The companion GQ Lupi b is seen well 0.7'' west of GQ Lupi A, whose PSF was subtracted. The background object GQ Lupi/cc2 is seen in the upper right (NW) corner.
Open with DEXTER

In the text

$\begin{figure} \par\includegraphics[angle=270,width=8cm,clip]{8493fig7.ps} \end{figure}$ Figure 7: Dynamic range from deepest image shown in Fig. 6. We plot (log of) noise flux ( ${\it S/N}=3$ ) devided by the peak flux of GQ Lupi A (before PSF subtraction) on the left axis (and the magnitude difference at the right axis) versus separation in arcsec on the bottom axis (versus separation in AU at 156 pc on the top axis). The lower full line with points is obtained after PSF subtraction, the upper line is without PSF subtraction. GQ Lupi and its companion are indicated as star symbols. The hatched area in the upper left is the region in this parameter space, where companions can be excluded by high-resolution spectroscopic monitoring of the radial velocity (rad vel) by Broeg et al. (2007), see text.
Open with DEXTER

In the text

$\begin{figure} \par\includegraphics[angle=270,width=8cm,clip]{8493figA1.ps} \end{figure}$ Figure A.1: NACO pixel scales as determined from binary HIP 73357 observed within one hour from GQ Lupi (but see footnote 1 for Feb. 2007). We show absolute values and errors in the upper panel (including errors due to (orbital) motion in HIP 73357 since Hipparcos at epoch 1991.25) and relative values for pixel scale changes with relative errors in the bottom panel (including errors due to (orbital) motion in HIP 73357 only since our first new epoch 2005.4), and also showing the data point at the first new epoch 2005.4, set to 0.0, with its error bar just from the Gaussian centering fit. Both panels show the same range in pixel scale on the y-axis, 0.17 mas in both panels.
Open with DEXTER

In the text

$\begin{figure} \par\includegraphics[angle=270,width=8cm,clip]{8493figA2.ps} \end{figure}$ Figure A.2: NACO detector orientations as determined from binary HIP 73357 observed within one hour from GQ Lupi (but see footnote 1 for Feb. 2007). Upper and lower panels are as in Fig. A.1. The dashed line in the upper panel shows the best fit to the data consistent with a detector orientation of $0^{\circ }$ at 1991.25, the Hipparcos epoch, indicating that the small increase in detector orientation with time can be interpreted as orbital motion in HIP 73357 A and B. Both panels show the same range in PA on the y-axis, $1.2^{\circ }$ in both panels.
Open with DEXTER

In the text

$\begin{figure} \par\includegraphics[width=6.3cm,clip]{8493fig6.eps} \end{figure}$	Figure 6: Deepest image of GQ Lupi (102 min total exposure) after subtraction of PSF of GQ Lupi A. North is up, east to the left. The companion GQ Lupi b is seen well 0.7'' west of GQ Lupi A, whose PSF was subtracted. The background object GQ Lupi/cc2 is seen in the upper right (NW) corner.
Open with DEXTER