6 Discussion and conclusion

We have shown that the HD 358623 primary A and its companion candidate B indeed show the same proper motion and that the spectral type of the companion (M2) is consistent with the observed colors and magnitude differences, so that it is a truely bound companion.

We would like to point out again the high precision achieved in the relative astrometry: after just one year, we could measure the proper motion ($\sim$100 mas) of both HD 358623 A and B with sufficient precision to show that they form a common proper motion pair, using the 150 mas pixel scale of the SofI small field and several non-moving background stars.

Figure 5: Dynamic ranges achieved. We plot the log of the flux ratio between the $3 \sigma$ background noise level and the peak intensity of HD 358623 (and, on the right hand side, the magnitude difference) versus the separation to the primary's photocenter (and, on the top axis, the projected physical separation at 48 pc), for the two detectors used: Sofi at the La Silla NTT (H-band) and SHARP-I at the NTT (K-band). The dotted lines show the expected flux ratios for 42 and 13  $M_{\rm Jup}$ masses (according to Burrows et al. 1997) for 48 pc and 20 Myrs, next to HD 358623. Every stellar companion above the H burning mass limit (75  $M_{\rm Jup}$) would have been detected outside of $0.5^{\prime \prime }$ (24 AU at 48 pc). The dynamic range curve for SHARP-I is limited by the $6^{\prime \prime } \times 6^{\prime \prime }$ size of the quadrants (we placed the primaries only onto the two best (lower, western) quadrants).

The mean apparent angular separation between HD 358623 A and B ( $2.205 \pm 0.028^{\prime \prime}$) corresponds to a projected physical separation of $105.2 \pm 6.6$ AU (at the Hipparcos distance of HD 199143, which is presumably the same as for HD 358623, see vdA00). The projected physical separation between HD 199143 A and B is $48.8 \pm 3.9$ AU, namely $1.023 \pm 0.031^{\prime \prime}$ (our separation measured with SHARP-I) at $47.7 \pm 2.4$ pc.

Let us investigate the sensitivity limits determined for the dynamic range achieved in the images: the flux ratio is determined from our actual images of the two stars in all SofI and SHARP-I images as the $3 \sigma$ background noise level on $7 \times 7$ pixel boxes as approximate PSF areas and devided by the peak intensity. We compare the observed dynamic ranges with expected flux ratios for possible companions of different masses (calculated following Burrows et al. 1997) next to HD 358623 (Fig. 5). The MPE speckle camera SHARP-I clearly gives the best dynamic range. In the SHARP images, we should have detected all stellar companions above $\sim$ $0.1~M_{\odot}$ outside of $\sim$ $0.5^{\prime \prime }$. Brown dwarf companions with $\sim$ $25~M_{\rm Jup}$ would have been detectable at $\sim$ $3^{\prime \prime}$ separations, more massive ones at smaller separations (between $\sim$0.5 and $3^{\prime \prime}$).

   

Table 4: Physical properties.

Object	$T_{\rm eff}$	B.C.	$L_{\rm bol}$	mass	age
	[K]	[mag]	[$L_{\odot}$]	[$M_{\odot}$]	[Myr]
HD 199143 A	6200	0.16	$2.40 \pm 0.25$	$\sim$1.25	$\sim$20
HD 199143 B	3720	1.43	$\sim$0.1	$\sim$0.60	$\sim$20
HD 358623 A	3955	1.00	$0.24 \pm 0.5$	$\sim$0.90	$\sim$20
HD 358623 B	3580	1.64	$0.05 \pm 0.01$	$\sim$0.55	$\sim$20

Next, we can compute the luminosities of the four objects studied. We assume the Hipparcos distance towards HD 199143 A ( $47.7 \pm 2.4$ pc) for all four objects. From the known V-H color indices and/or known spectral types, we can estimate the effective temperatures and bolometric corrections B.C. (taken from Kenyon & Hartmann 1995). Temperatures, B.C., and luminosities are listed in Table 4.

We placed the stars into the H-R diagram and compared their locations with theoretical tracks and isochrones by Palla & Stahler (1999) and Baraffe et al. (1998) to estimate masses and ages. Rough values are given in Table 4. All four stars appear to be co-eval with an age of $\sim$20 Myrs.

In conclusion, we find that all our data are consistent with HD 199143 and HD 358623 each having an early-M type stellar companion. In the case of HD 358623 B, the spectral type is confirmed by JHK colors and a spectrum and companionship is also confirmed by common proper motion. For HD 199143 B, spectrum as well as proper motion and, hence, companionship, still have to be confirmed.

Acknowledgements

We would like to thank the NTT team with O. Hainaut, L. Vanci, and M. Billeres for support during the SofI observations. We are grateful to Klaus Bickert and Rainer Schödel for their help with the Sharp run. Also, we would like to thank Wolfgang Brandner, João Alves, Fernando Comerón, and Ray Jayawardhana for useful discussions throughout the project. RN wishes to acknowledge financial support from the Bundesministerium für Bildung und Forschung through the Deutsche Zentrum für Luft- und Raumfahrt e.V. (DLR) under grant number 50 OR 0003. We have made use of the Simbad database operated at the Observatoire Strassburg.