Astron. Astrophys. 338, 807-812 (1998)

5. Temperature distribution and heating mechanisms: the origin of the dust

Actually, the uncertainty on the temperature distribution is directly connected to the study of the heating mechanisms and, therefore, to the knowledge of the UV interstellar radiation field.

To this aim, and in order to investigate the origin of the dust in elliptical galaxies, I plot in Figs. 3a and b the dust mass M_d and the FIR luminosity L_FIR versus the blue luminosity L_B for the galaxy sample. Both M_d and L_FIR are correlated to L_B, in contrast with what has been found by GJ95. Therefore, while the absence of correlation between dust mass and blue luminosity in their sample was considered to support the external origin of the dust in ellipticals and the evaporation flow picture, the present result indicates that a relationship between the dust content and the present day population of stars should be not excluded. But, as already pointed out by GJ95, a lack or a presence of correlation between M_d and/or L_FIR and L_B cannot be a definitive proof of the external or the internal origin of the dust, due to the fact that the dust destruction mechanisms and timescales in elliptical galaxies are expected to be different depending on the evolutionary state and on the hot gas content of the individual objects.

Fig. 3. a Dust mass (with the temperature distribution model) and blue luminosity relation: the solid line is the least squares bisector fit (confidence level greater than 95) while the dashed line is the loci where dust is replenished by stellar mass loss and destroyed by sputtering with the maximum destruction timescale ( yr). Most of the galaxies present an observed dust mass larger than the amount predicted by the dust formation and destruction model. b FIR and blue luminosity relation: the solid line is the least squares bisector fit (confidence level greater than 99.5). For comparison the theoretical prediction (dashed line) by Tsai & Mathews (1996) is also shown.

For this reason, further analysis and, above all, high resolution FIR observations are required to understand the origin and the fate of dust in these systems. Waiting for a complete analysis of the ISO data, a reasonable approach is the comparison between the available observations and models of star formation and dust heating mechanisms. I plot in Fig. 3a the loci where dust is replenished by stellar mass loss at the rate given by Faber & Gallagher (1976) and destroyed by sputtering with the maximum destruction timescale ( yr) in ellipticals which contain hot gas (see GJ95 for details). In 85 of the galaxies the dust having an internal origin does not account for the computed dust masses, which are larger. Therefore, one can argue that in 85 of the ellipticals an alternative supplying mechanism for the dust is required to account for the FIR observations. Mergers with spirals or small, dust-rich irregular galaxies and , then, the evaporation flow scenario are, in fact, strongly supported by the previous evidence (Fig. 3a). The critical point of this picture is to identify the observed diffuse dust, located within 2 Kpc from the center, as the dust accreted during the galaxy interaction when high hydrodynamic instabilities are expected. In particular, the external dust had to be somehow protected from the interaction with the hot gas when moving toward the center.

The problem could be solved with the "extra" dust component (Tsai & Mathews 1996) located in very large disks out to the effective radius and, therefore, cold enough to emit at FIR and submillimeter wavelengths. This dust can have, in principle, both internal and external origin. In particular, due to the low temperature "dust may re-form and grow in these cold disks" (Tsai & Mathews 1996). Unfortunately, due to the large uncertainties, it is not possible to evaluate the exact amount of this dust component. Therefore, since the external origin of the dust is far from fully verified, I experimented with an alternative approach to interpret the relation between the FIR and the blue luminosity. By using the least squares bisector method, I find L_FIR (Fig. 3b), in agreement with Bregman et al. (1998). The luminosity correlation supports the scenarios in which a significative amount of dust has internal origin (Tsai & Mathews 1995, 1996), coming from stellar mass loss and being heated by stellar photons and the general interstellar radiation field.

The problem of the coexistence of dust and hot gas might be resolved with a dust grain distribution containing grains larger than the maximum size suggested by Mathis et al. (1977) and usually adopted in the current models. In fact, this assumption would increase the sputtering time and, then, the dust grain density. Tsai & Mathews (1996) found that the FIR luminosity is proportional to when the maximum grain size increase from 0.3 µm to 0.9 µm. This theoretical prediction (dashed line in Fig. 3b) and the observed relation (solid line in Fig. 3b) is consistent in the luminosity range covered by the present sample ¹, therefore the luminosity correlation seems to support the cooling flow scenario.

Finally, the two comparisons between observations and empirical-theoretical models show a clear conflict: while the canonical star formation rate cannot account for the dust amount in elliptical galaxies, thus supporting the evaporation flow scenario, the trend of the L_FIR-L_B can be explained with a proper dust model which increases the sputtering time and which suggests an ad hoc dust distribution. Concerning this last possibility it has to be stressed that, although the L_FIR-L_B relation can be affected by different hot gas contents and stellar populations in the individual objects, the correlation between these quantities is characterized by a trend predicted by theoretical models. Furthermore, both in Bregman et al. (1998) and in the present work, severe selection criteria are applied (see Sect. 3), which take into account the reliability of the observations and the different contributions to the FIR emission. I compare the present sample and the sample of Bregman et al. (1998) which rejected the galaxies whose FIR fluxes can be contaminated by AGN emission, position uncertainties, background objects and inhomogeneity. The L_FIR-L_B relation in Fig. 3b is not affected by the exclusion from the sample of the four AGNs (NGC 4374, NGC 4486, I 4296 and I1459) and of NGC 3557 that has a non-homogenous background.

Taking into account the results of the present analysis it seems difficult to choose between the evaporation flow or the cooling flow scenario and to affirm that the dust in ellipticals has mainly en external or an internal origin. A resonable conclusion is that the individual objects have experienced different mechanisms of dust accretion during their evolution and that only a detailed study of the individual systems will allow to describe for each object the different evolution phases and the contributions of the different mechanisms of dust accretion.

© European Southern Observatory (ESO) 1998

Online publication: September 17, 1998
helpdesk.link@springer.de