Astron. Astrophys. 354, L1-L5 (2000)

3. Results

Signal is detected in the central channels of both the CO(2-1) and CO(5-4) spectral data sets at 10 and 5 sigmas respectively. The averages of the channels with detected emission were used to produce the integrated maps shown on Figs. 1a and 1c. The CO(2-1) integrated flux is 1.4 Jy km s^-1. There is no indication for any extension of the emission at this resolution. The peak of the source is offset by about from the position of the optical source as measured by HST (Dey et al. 1999). This offset is well within the astrometric accuracy of HST but if real it may be indicative of a spatial extinction.

Fig. 1. a CO(2-1) integrated intensity map; the contours are at 0.25 Jy beam-1 km s-1. b CO(2-1) spectrum at the center of the source. c CO(5-4) integrated intensity map; the contours are at 0.75 Jy beam-1 km s-1. d CO(5-4) spectrum at the center of the source. The contribution from the continuum is not subtracted from the line flux.

The corresponding CO(2-1) profile (Fig. 1b) is roughly gaussian, with a FWHM of almost exactly 400 km s^-1, and a central frequency that corresponds to a redshift of . The appearance of a flat-top or even double peak profile around km s^-1 (see Fig1b) cannot be checked with the present data and only observations at higher signal-to-noise ratio (those exploiting the full spectral resolution) can settle its reality. The line-width observed is fairly large, although not atypical for this kind of sources (e.g. SMMJ02399, Frayer et al. 1998) and could be due either to an edge-on system or to various separate components. The redshift deduced from the CO line is apparently shifted from that deduced from the H line (), but corresponds to the redshift deduced from the [O II ] line at 3727 Å ().
This shift is still within the uncertainties in the optical redshift but if real it would be different from what is typically found for low redshift luminous galaxies (Sanders & Mirabel 1996), where systematic blue-ward offsets of optical lines from the CO redshift are attributed to outflows with dust obscuration (see e.g., Gonzalez-Delgado et al. 1998). The question remains, however, completely open since a recent analysis by McIntosh et al. (1999) of a sample of quasars shows how high-redshift objects present H lines with a systematic mean red-ward shift of 500 km/s with respect to the systemic redshift of the objects (that defined by the narrow line region). Even though the comparison with quasars may not be fair since the line emitting regions could be different, HR10 seems to show similar properties with the CO redshift corresponding to that of the narrow forbidden lines and coinciding with the centre of mass of the system, while the H line is shifted with respect to that.

The integrated map corresponding to the CO(5-4) map (Fig. 1c) shows a source at the same position as the CO(2-1) integrated map. The integrated intensity of that source (corrected for the contribution of the continuum) is 1.35 Jy km s^-1. The corresponding CO(5-4) profile (Fig. 1d) is roughly gaussian, with a FWHM of 380 km s^-1 similar to that of the CO(2-1) profile. The central frequency corresponds to a redshift of similar to that deduced from the CO(2-1) line.
Although the CO level is above the ground state the integrated flux (in Jy km s^-1) of CO(5-4) is equal to that of the CO(2-1) line, . CO luminosities, in solar units, are 1.5 and for the CO(2-1) and CO(5-4) line respectively, while the total line luminosities L are and K km s^-1 pc² , for the CO(2-1) and CO(5-4) line respectively. When expressed in these latter units, the ratio between the CO(2-1) and CO(5-4) luminosities of the same source is proportional to the line intrinsic brightness (Rayleigh-Jeans) temperature ratio integrated over the area of the source:

where is the source solid angle. If the spatial extent of the CO(5-4) emission region is similar to that of the CO(2-1) - a plausible hypothesis since both transitions have same line-width and profile - and corresponds to a value of the excitation temperature of K (see e.g. Maloney, & Black 1988). If gas and dust are in thermodynamic equilibrium the kinetic temperature would equal , but it is usually found that . In HR10 the dust temperature was estimated to be (Cimatti et al 1998 , Dey et al. 1999) and this value can be taken as the upper limit to the gas temperature. If we assume that K the gas density implied by this ratio is less than . As an example, at K and K the estimated excitation temperature is K for a CO density of . The present data cannot distinguish between a picture where the dominant component of the ISM in this system is a diffuse () gas or whether the medium is clumpy. The spatial shift between the CO and H lines would be more compatible with this latter picture.

If the ratio between L and the mass of molecular gas is similar in HR10 and Arp 220 (Scoville et al. 1997b), the molecular gas mass in HR10, using the CO(2-1) line, is , larger than what is usually found in local ULIRGs (Solomon et al. 1997 , Braine & Dumke 1998) but similar to that of other detected high-z sources (Frayer et al. 1998).

The 1.3mm continuum map shows only a marginal () detection at the position of the source, with an integrated flux of mJy beam^-1. The flux detected with the IRAM 30m telescope was mJy (Cimatti et al 1998) at 240 GHz. Scaling the flux as , the PdBI detection would correspond to an expected flux of mJy at the observed frequency of the 30m. We exclude that this discrepancy is due to an extended component; it is more likely due to the difference in the calibration of the two instruments since the two values are consistent within the error bars.

© European Southern Observatory (ESO) 2000

Online publication: January 31, 2000
helpdesk.link@springer.de