2. Observations and data reduction
NGC 2023 was observed with the Redeye wide field camera at the F/8 Cassegrain focus of the Canada-France-Hawaii telescope. The present observations extend to the north and south of the field of the observations reported in F94. In addition, the present observations involve both S(1) and S(2) lines of H2. Observations were performed during the nights of December 1st and 2nd 1993. The seeing was measured to be 0.8". The camera was equipped with a Rockwell NICMOS 3 Hg:Cd:Te array of 256 x 256 pixels at 40 µm pitch. The image scale is 0.5" per pixel, giving an observed field of 128" x 128", noting that NGC 2023 has an angular extent of 5' x 3.5' according to plate 1 of Malin et al. 1987.
Wavelength selection was performed through the use of fixed wavelength filters at 2.179, 2.121, 2.099, 2.033 and 1.996 µm. The 2.121 µm filter is nominally centred on the v=1-0 S(1) ortho-H2 line and the 2.033 µm filter on the S(2) para-H2 line. All filters are 1% bandpass. Filters at 2.179, 2.099 and 1.996 µm record the IR background continuum emission, where wavelengths have been chosen adjacent to the S(1) and S(2) lines such that there are no H2 emission lines within the bandpass of the filters (Black and van Dishoeck 1987). Exposure times ranged from 2 to 100 seconds, and total exposure times for each filter ranged from 370 to 387 seconds.
The method adopted in order to obtain images of H2 emission in S(1) and S(2) has already been described in detail in F94 and in L96. In brief, it is necessary to take account of temporal variability of the sky background and the H2 signal (especially in the event of variable weather), of spatial variations in the detector sensitivity, of variation of sky brightness with wavelength and of differing efficiencies of the filter- detector combination at different wavelengths. These objectives were achieved by manipulating the data to obtain normalized, flat-fielded images for each filter, both on- and off-source, with an off-source sky reference at a position -60" in RA (west) and +180" in Dec. (north) from that of HD 37903 (RA(2000) 05h:41m:38.32s; Dec(2000) :15':32.6"). For the southern field, emission from three faint stars was then used to bring images obtained through different filters to the same scale, on the good assumption that these stars emit as black-bodies in this wavelength range. This procedure cannot however be employed for the northern field, where there are no suitable stars. In this case, spatial overlap between the northern and central fields (see below) was used to place images for different filters on the same brightness scale. Line emission maps in S(1) and S(2) were obtained by subtracting maps at 2.179 and 1.996 µm from those at 2.121 and 2.033 µm respectively. Absolute flux calibration was provided by imaging the star HD2007, with a K magnitude of 4.432 0.004 (Bouchet et al. 1991).
The first night of observation was devoted to the same field as in F94, but now including S(2) emission. In the succeeding night, the two adjacent fields to the north and south were observed. The image to the north of HD 37903 involved an overlap with the central image of about 40" on a seahorse-shaped feature. This overlap, involving strong H2 emission features and continuum emission allows us to ensure that images obtained through different filters are on the same scale, as indicated above. Using the data reduction techniques outlined in the preceding paragraph and in F94, we obtained an image in S(1), shown in the upper portion of Fig. 1. The most prominent new feature is emission in the form of a triangle, reminiscent in form of a (celestial) Welsh harp (Jones 1784), 120"N 30"E of the central illuminating star. We refer to this feature as the infrared triangle. The southern portion of the image shown in Fig. 1 was obtained by positioning the so-called star C (De Poy et al 1990, Sellgren et al. 1992), located south-west of HD 37903, in the northwest quadrant 29"W 74"N of the centre of the image. The most striking feature of this portion of the image is that centred around 33"W 36"N of star C. We refer to this feature as the infrared cross. This southern part of the image also shows, to the southwest of star C, a series of small intense blobs of H2 emission. These may be readily identified with the chain of Herbig-Haro objects reported in Malin et al. 1987. Poor weather conditions during the first night reduced the quality of the S(1) and S(2) data for the central part of the field. In Fig. 1 we have therefore used the image shown in F94 in order to make the composite figure. We have data of good quality however for both S(1) and S(2) lines for the southern field. Due to lack of time, data were only recorded in S(1) for the northern image. Fig. 2 displays the continuum dust emission at 2.179 µm. The infrared triangle and jet-like structure associated with star C are only very faintly detectable in the continuum whilst the infrared cross is altogether absent. Sellgren (1984) showed that 2.2 µm continuum emission may arise from thermal spiking to 1000K of very small grains due to absorption of VUV photons. Apparently this mechanism is not operative in these zones and thus by implication there are no small dust grains present in these regions. Fig. 3 provides a finding chart for objects referred to in this paper. In the bandpass of the S(1) 2.121 µm and S(2) 2.033 µm filters, the sky reference images yield an average of respectively 900 and 1200 counts per pixel in 100 seconds of integration time. Excluding the jet-like structure to the east of star C and the Herbig-Haro objects, the two brightest spots on-source are located 22"E 27"N from HD 37903 and 32"W 36"N from star C, and in the 2.121 µm image provide respectively 200 and 150 counts per pixel in 100 seconds. Values of around 250 are measured in the "jet" east of star C, 570 in HH 1A and 280 in HH 4 (Malin et al. 1987). An overall sensitivity of 66 incident photons per count was found at 2.121 µm. All signals are well below the detector saturation level of 12000 counts.
The peak flux in the S(1) line was 4.3 1.0 x 10-18 Wm-2 in a beam of 1", measured on the brightest filament 22"E 27"N of the illuminating star, where the error quoted is not an error due to noise but arises very largely due to uncertainties in absolute flux calibration. Using a fluorescent lifetime of 1.20 x 106 s for the J=3, v=1 upper state of the S(1) line (Black and Dalgarno 1976), we find a peak column density of 0.8 0.2 x 1021 m-2 in this state. The S(2) data yield a column density of 0.4 0.1 x 1021 m-2 in the J=4, v=1 upper state, using a fluorescent lifetime of 1.18 x 106 s (Black and Dalgarno 1976). The corresponding emissivity in S(1) is 1.7 0.4 x 10-7 Wm-2 sr-1 and in S(2) is 0.8 0.2 x 10-7 Wm-2 sr-1, where again errors arise from calibration. The infrared triangle to the north is 2 times weaker whilst the infrared cross to the southwest is 1.3 times weaker than the peak values recorded above. Values of peak flux for S(1) and corresponding column density are about a factor of two less than observed in NGC 7023, where the emission is far more localized.
© European Southern Observatory (ESO) 1998
Online publication: April 15, 1998