2. Observations and data reduction
The observations of 56 HAEBES long slit spectra were obtained during three observing runs in La Palma on the Isaac Newton Telescope (INT) with the Intermediate Dispersion Spectrograph (IDS) in September 1991, July 1993 and December 1994. Further observations were kindly taken by Alan Moorhouse on La Silla with the ESO/MPI 2.2m Telescope using EFOSC-II in December 1991.
The La Palma data were taken at three different resolutions using the H1800V holographic grating and the R1200Y and R632V ruled gratings with GEC6 (Sept '91), EEV5 (July '93) and TEK3 CCDs (Dec '94) as detectors. The pixel size for all CCDs was 22µm. The corresponding dispersions for the three gratings were 0.22 Å/pixel for the H1800V grating, 0.36 Å/pixel for the R1200Y grating and 0.7 Å/pixel for the R632V grating, resulting in a velocity resolution (2 pixels) at H of 21 kms , 34 kms and 70 kms respectively. The La Silla data was taken using a grism with a dispersion of 1.1 Å/pixel and a Thompson 1024 x 1024 CCD with a pixel size of 19µm and a velocity resolution of 105 kms . Data reduction of the spectroscopy was carried out using standard IRAF 1 routines. Bias subtraction and flat-fielding corrections were determined from zero second exposures and tungsten lamp exposures respectively. The adjacent sky spectrum was subtracted from the object spectra, and the dispersion solutions were determined from CuAr arc exposures.
The velocity calibrations with respect to the stellar rest velocity are very important. Unfortunately in the wavelength range of the exposures encompassing the [OI] 6300 lines, the coverage (e.g. 5850 - 6750Å with the R632V grating) does not include many photospheric lines that might serve to accurately determine the stellar rest velocity, V . Finkenzeller & Jankovics (1984) published radial velocities for 27 HAEBESs, measured from the shift of the narrow interstellar absorption lines of the Na D and CaII lines. They argued that both the interstellar lines of CaII and Na D have mean residual velocities close to those of the molecular clouds associated with the observed Herbig stars. Assuming that the majority of the line of sight absorption from Na D and CaII arises close to the HAEBESs in their associated clouds and that the young stars have velocities characteristic of their parent cloud velocity, the interstellar lines should have velocities that are approximately the same as those of the stars themselves. Finkenzeller & Jankovics (1984) found very good agreement between the measured velocities of Na D and CaII and the molecular cloud velocities measured by mm molecular line observations (see, for example, Edwards & Snell 1983) and quoted a mean residual velocity of -3.1 3 kms for the interstellar Na D absorption lines with respect to the associated molecular clouds. We have therefore used the interstellar absorption lines of Na D to determine the radial velocities of the stars in our sample.
In a number of cases it is possible to compare the radial velocity shift measured from the Na D with photospheric lines of iron, FeI 6495 & FeII 6456, and from SiII 6347/6371. The lines of silicon are at least partly chromospheric and as such may not truly represent the photospheric radial velocity. However, where the SiII lines are symmetric, the velocity shift agrees with that measured from the Na D lines and the FeI & FeII lines. The overall errors in the velocity measurements are between 5 kms and 20 kms , depending on the resolution of the grating. An average of 8 kms is adopted for the INT data and an average of 18 kms is adopted for the ESO data (as identified in Table 1).
Table 1. [OI] 6300 line parameters and characteristic H profiles for the Herbig Ae/Be sample. is the [OI] 6300 equivalent width, the [OI] 6300 centroid velocity relative to the estimated systemic velocity of the star, and the blue and red wing velocities respectively. Where there are two entries for the star shows doubled-peaked emission at [OI] 6300, and is the blue wing of the high velocity component and the red wing of the low velocity component respectively. The emission category is explained in the text. FWHM is corrected for the instrumental width. The profile of the H line is determined from our data, according to the criteria established by Finkenzeller & Mundt (1984). Those stars where the [OI] emission was unresolved in our spectra are listed as such. The average error in measuring is 8 kms except in those few stars observed at ESO which are marked with , where the error is 18 kms .
It must be noted that for the purposes of velocity calibration, we have only used the narrow symmetric absorption lines of Na D , to insure that we measure the cloud gas velocities. Broader redshifted absorption lines in Na D , as observed, for example, by Graham (1992) and Grinin et al. (1994) have been modeled by Sorelli et al. (1996) as due to either magnetospheric accretion or the evaporation of star-grazing planetesimals.
© European Southern Observatory (ESO) 1997
Online publication: June 30, 1998