Michael G. Burton, J.E. Howe, T.R. Geballe, P.W.J.L. Brand, PASA, 15 (2), 194
The html and gzipped postscript versions of this paper are in preprint form.
To access the final published version, download the pdf file.
Next Section: CONCLUSIONS Title/Abstract Page: Near-IR Fluorescent Molecular Hydrogen Previous Section: RESULTS | Contents Page: Volume 15, Number 2 |
DISCUSSION
The relatively strong 2-1 S(1) line flux, compared with that of the 1-0 S(1) line, and the rich spectrum of higher vibrational lines observed clearly indicate that the bulk of the H emission arises from fluorescent excitation. This phenomenon was predicted by Black & Dalgarno (1976) and first observed by Gatley et al. (1987). Our results extend this later work, and provide further detail on the nature of the mechanism.
The 1-0 S(1) line flux is relatively stronger in the southern position, (-11'',-78''). This is evident in Table 3, where the intensity of the H line with greatest S/N in each vibrational level from v=1 to v=6 is listed. These are relative to that of the 2-1 S(1) line flux at each position. It can be seen that for v=2-6 the relative intensities are, within the errors of observation, the same. However, for the v=1-0 S(1) line the ratio at (-11'',-78'') is , significantly greater than the value of measured at (+33'', +105''). The second value is consistent with models for pure fluorescent excitation (Black & Dalgarno, 1976). The first is expected from `collisional fluorescence' (Sternberg & Dalgarno 1989; Burton et al. 1990b), whereby collisional de-excitation of fluorescently excited levels increases the v=0 and 1 populations over the pure fluorescent value. For this to occur the gas density has to be above critical (n ) with some molecules existing in a hot, self-shielded layer close to the surface of the molecular cloud (A). Collisional de-excitation of high-v levels can then be significant, as well as thermalisation of the v=0 and 1 level populations.
This suggests that the molecular gas in NGC 2023 is clumpy, with some dense gas at the southern position having been heated in a self-shielded layer. We estimate that the far-UV radiation field from HD 37903, between 6-13.6eV, to be times the ambient interstellar value ( erg ; Habing 1968) for a projected distance 0.18pc from the star, as at (-11'', -78''). For (+33'', +105'') the minimum distance is 0.25pc, for which the radiation field would be G. The former radiation field is just sufficient for collisional fluorescence to occur in dense gas, the latter is not as the gas temperature would not be high enough (Burton et al. 1990b). Thus the data are consistent with a pure fluorescent emission spectrum from the northern position observed, and with some collisional fluorescence at the southern location.
We can examine this conclusion more quantitatively through comparison with predictions of theoretical models, such as that of Burton et al.\ (1990b). For pure fluorescent emission, H line intensities are approximately proportional to the density of gas and the surface filling factor, f, of the emission region in the beam. For the 1-0 S(1) line it is given to a factor by , where is the density in units of . The flux is only weakly dependent on the strength of the far-UV field, the gas being excited over a constant column depth A into the cloud, and an increased pumping rate from a higher far-UV field being balanced by an increased dissociation rate. Thus the observed intensity at (+33'', +105''), , is consistent with pure fluorescent emission from molecular gas at a density of a few , a filling factor of 1-3 and an extinction at 2m of 0-1 magnitudes. At (-11'',-78'') the radiation field is high enough to heat the surface of the cloud to temperatures of K. At the inferred densities self-shielding of H will occur, putting molecular gas close to the surface of the cloud. Over a typical column depth A there will thus exist hot molecular gas. If, for instance , then the intensity of the thermal contribution to the 1-0 S(1) line flux would be from this column, comparable to the measured at (-11'', -78''). However, this temperature is not sufficient to populate the v=2 level. The thermal contribution to the v=2-1 S(1) line would be over 1,000 times weaker, leaving that of the UV-pump to dominate its line intensity. As concluded above, for fluorescence remains the primary contributor to the line emission.
Also shown in Table 3 are the predictions for the relative H line intensities from Black & van Dishoeck's (1987) model 14, for a fluorescent cascade. We note that, while the observed high-v lines are relatively strong, they are still somewhat less than the model predictions for them. In addition, the intensities of higher rotational lines within a given vibrational level tends to be less than the lower-J lines, contrary to the expectations from Sternberg & Dalgarno's (1989) model (see also Howe, 1992). This indicates that further refinements to the models are still likely to be necessary.
Line 1-0 S(1) Relative intensity compared to the 2-1 S(1) line. These have not been corrected for differential extinction.
(-11'', -78'') (+33'', +105'') Model 1.60 2-1 S(1) 1.0 1.0 1.0 3-2 S(1) 0.47 4-2 O(3) 0.67 5-3 Q(1) 0.68 6-4 Q(1) 0.53
Prediction of Black & van Dishoeck (1987) model 14.
Next Section: CONCLUSIONS Title/Abstract Page: Near-IR Fluorescent Molecular Hydrogen Previous Section: RESULTS | Contents Page: Volume 15, Number 2 |
© Copyright Astronomical Society of Australia 1997