Online Material

Table 1:   Journal of spectroscopic observations.

Table 2:   Journal of NACO observations.

	Figure 6: Magnetic field measurements using the data obtained in November 2004. The longitudinal magnetic field is derived following the equation from the slope of the linear fit through the data points (see Bagnulo et al. 2002). Large open and small filled symbols are obtained from the high number and low number Balmer lines, respectively.
Open with DEXTER

Appendix A: Observations and data reduction

The observational material discussed in this section includes FORS 1 spectropolarimetric data (Appenzeller et al. 1998) obtained in the framework of our programs carried out in May 2008 and in December 2008. Additionally, we used ESO archive data from 2004 obtained in circular spectropolarimetric mode.These data were used by Wade et al. (2007) to search for the presence of longitudinal magnetic fields in a larger number of Herbig AeBe stars. These older observations (see Table 1) were carried out with a slit width of 0 8 and grism 600B ( ). The flux spectra in the ordinary and extraordinary ray of the Wollaston prism were extracted for all position angle settings of the retarder waveplate. Since there were no available flux calibrators observed for this setup in the respective night, we used flux calibrators taken in normal spectroscopic mode without the Wollaston prism. The new spectropolarimetric data from 2008 were obtained with the same instrument, but now equipped with a higher UV response detector system with a smaller pixel size. During the December run we have used the higher dispersion grism 1200B in addition to the data with the 600B grism. In all our observations the slit width was set to 0 4 to optimize the resolution for the polarimetric measurements ( and , respectively). The spectrophotometric calibrators retrieved from the archive were observed without polarization optics on different nights. One source of uncertainty in the flux calibration is therefore the unclear spectral response of the wave retarder plates, for which we presume a high throughput over the full spectral range. However, significant slit losses are expected with this setup due to the narrow slit widths and the difficult acquisition of the target on the slit considering the extreme brightness of the target for 8 m-class telescopes. To remedy this difficulty related to the absolute calibration, we re-scaled the flux at 5500 Å to the monthly average of the AAVSO visual observations of 9 4 in November 2004 and 8.4 in December 2008. In Fig. 2 we present spectrophotometric fluxes for the Z CMa system in 2004 and 2008, respectively. Further we corrected the result with the chromatic zero angle of the super-achromatic half wave retarder plate. for the proper interpretation of the measured polarization angle. To perform linear polarization measurements, a Wollaston prism and a half-wave retarder plate rotated in 22.5 steps between 0 and 67.5 were used. For circular polarization measurements the quarter-wave retarder plate was used at the positions +45 and -45. Our FORS 1 data were supplemented by three high-resolution FEROS spectra (R=48 000) obtained in February and March 2009 to study the behavior of hydrogen lines. The journal of spectroscopic observations is presented in Table 1.

To better understand the parameters of the resolved binary system we retrieved all available NACO adaptive optics near-infrared data (Lenzen et al. 2003) from the ESO archive. As for the FORS 1 data sets, the target is rather bright for imaging-mode observations with 8 m-class telescopes, and accordingly the various data sets were taken with a mixed choice of neutral density and narrow band filters. As a consequence we cannot provide absolute photometry for NACO data with the required level of confidence, but only the flux ratio of the two components (see Table 2) and the astrometric parameters of the binary. In case of normal jittered observation sequences we have used the Eclipse data reduction package provided by ESO to NACO users. In other cases we had only two acquisition exposures or a few jitter positions.