Astron. Astrophys. 325, L21-L24 (1997)

3. Results and Discussion

3.1. Spectral Energy Distributions

The fluxes are listed in Table 2 and the spectral energy distributions are shown in Figure 1. For comparison also the IRAS values are plotted which are in excellent agreement. The remarkable features of the SEDs are:

Table 2. ISOPHOT photometry.

Fig. 1a-c. Spectral energy distributions of Arp 220, NGC 6240 and Arp 244: ISOPHOT (errors generally adopted to 30 ), IRAS, SUBMM (from Rigopoulou et al. 1996), the solid and the dotted lines represent modified blackbody fits with emissivity , for NGC6240 and Arp220 the dashed-dotted lines connect the ISOPHOT data points in order to emphasize the spectral features shortwards of 20 µm.

• 1) For each galaxy the maximum of the SED can now be very accurately determined. It lies between 70 and 100 µm and the ISOPHOT long wavelength filters clearly outline the Rayleigh-Jeans branch. This branch follows excellently the shape of a modified blackbody with emissivity proportional to and temperature about 30-50 K. We call this the cold component. For Arp 220 the Rayleigh-Jeans tail is nicely continued by the submm data points measured by Rigopoulou et al. (1996).
• 2) In addition, shortward of 60 µm there is an excess over the cold component. In order to model the mid infrared SED we fitted a second modified blackbody function with temperature of about 120-160 K. We call it the warm component. Note that the functions are best eye fits and are not simply obtained by least square minimising.
• 3) For Arp 220 and NGC 6240 shortwards of 20 µm spectral line features appear in addition to the warm component. They include the PAH emission feature at 7.7 µm and the 10 µm silicate absorption, already reported by Smith et al. (1989). Moreover a small emission bump between 12 and 16 µm is present, possibly due to small grains enhanced by the [NeII]12.8 µm and [NeIII] 15.6 µm lines. As Acosta-Pulido et al. (1996) reported this for NGC 6090, it seems to be a common feature in the 12 to 16 µm SEDs of IR luminous galaxies. Note that due to the MIR spectral features the blackbody fit to the warm component is likely to be more uncertain than the fit to the cold component.
• 4) Between 10 and 60 µm the SED rises much steeper for Arp 220 than for NGC 6240 and Arp 244, or, in other words, the "knee" in the SED at 25 µm is less pronounced for Arp 220. This may be the consequence of increasing extinction from Arp 244 over NGC 6240 to Arp 220, as discussed by Lutz et al. (1996).

In particular for NGC 6240 and Arp 220 we have obtained the most detailed infrared spectral energy distributions which can better constrain galaxy dust models. Note that the fluxes quoted in Table 2 refer to a spectral shape of passing the filter and detector bands. As this is not the real spectral energy distribution for our galaxies, for a detailed modelling it might be necessary to apply color corrections. Therefore color correction factors for a modified blackbody spectrum proportional to B(,T), which is a good first order approximation, are given in Table 2. In the following we will confine to the simplest dust model and discuss the observed spectral energy distribution fitted with two modified blackbody functions. We did not apply any color correction, because the effect is small and it does not influence our basic results.

3.2. Luminosities, dust temperatures and masses

In Table 3 the first two columns PHT (= the IRAS range) and PHT (= the ISOPHOT range) list the IR flux contributions in synthetic non-overlapping bandpasses. Column 3 () gives the integrated flux between 1 and 1000 µm derived from the blackbodyfits. In column 4 we give for comparison the extrapolated mid and far infrared flux according to the definition in Table 1 of Sanders and Mirabel (1996). The main results of this comparison are:

Table 3. IR flux obtained from IRAS and ISOPHOT using various wavelength ranges. For more details of see Table 4. The fluxes are given in [10 Wm ].

• 1) Including the far infrared capabilities of ISOPHOT beyond 100 µm shows that the additional flux amounts to about 10 for Arp 220, 15 for NGC6240 and 18 for Arp 244.
• 2) Assuming that the gives the best estimate of the total flux, the IRAS range comprises already 80-90 and the ISOPHOT range 90-100 .
• 3) While for Arp 244 and NGC 6240 the PHT fluxes correspond well with the total flux from 1-1000 µm, a considerably smaller 3-220 µm flux is derived for Arp 220. This flux deficit can be explained by the coarse wavelength coverage between 25 and 60 µm which underestimates the flux due to the steep rise of the Arp 220 SED.
• 4) The IRAS extrapolation and the double blackbody fit results agree well within the adopted error range.

From the blackbody fits we derive the IR luminosity contributions between 1 and 1000 µm for the cold and warm component, respectively (H = 75 km/sec/Mpc). They are given in Table 4 together with their sum. Note that the total luminosities are dominated by the cold component.

Table 4. Infrared luminosities (1-1000 µm), dust temperatures and masses derived from two modified blackbodies with emissivity
(H = 75 km/sec/Mpc). For comparison also the total molecular gas masses M(H ) are listed.

The dust masses of the cold and warm components are compiled in Table 4, derived from the luminosities and dust temperatures using the formula given in Klaas and Elsässer (1993). E.g. for ARP 220 our estimate agrees within the errors with that obtained by Scoville et al.(1991) from IRAS and submm data ( = 5 10 [ ]). For comparison Table 4 also lists the gas masses from the literature. The dust-to-gas ratio lies about 1/300, a factor of three lower than the standard galactic dust-to-gas ratio.

The sequence Arp 244 - NGC 6240 - Arp 220 exhibits significant trends, the cold component showing an increase in luminosity, temperature and mass. The warm component, surprisingly, shows an apparent decrease in temperature, probably the consequence of increasing extinction mentioned above. In Arp 244 a relatively small dust mass amounts to a significant fraction of the IR luminosity. In Arp 220 nearly the whole luminosity is generated by the cold component, though the strong extinction could displace the ratio against the warm component.

3.3. Heating of the cold and warm component

Arp 244: The temperature of the cold dust component is close to that found for normal spirals (M101) or weakly interacting objects (M51), even including HII-regions (e.g. Hippelein et al. 1996). The bimodal appearance of the SED is similar to that of Seyfert galaxies, where the warm component is interpreted as being powered by the active nucleus (Rodriguez Espinosa et al. 1996). This mechanism, however, can be ruled out for Arp 244. Spatially resolved IR imaging between 12 and 18 µm by Vigroux et al. (1996) shows that the main activity is taking place in a few hot spots in the overlap region of the two galaxy disks and not in the nuclei.

NGC 6240 and Arp 220: In these tightly interacting and merging systems huge dust masses (3 to 5 10 ) are efficiently heated to temperatures 10 - 20 K higher than found in normal galaxies. Although for Arp 220 a large fraction of the molecular gas mass is concentrated within a very small region, the derived dust parameters argue for more spatially extended heating sources. It could be direct heating by a starburst as suggested by Lutz et al. (1996). Also the prominent PAH features argue against an AGN as the only heating source. The optical spectra of both objects are LINER types (Fried and Ulrich 1985, Sanders et al. 1988). NGC 6240 is the most luminous source of 1-0 S(1) H emission indicative for shocks (Goldader et al. 1997), and extended X-ray emission has been found for Arp 220 (Heckman et al. 1996). Therefore, collision of clouds similar to a model proposed by Harwit et al. (1987) or bipolar superwinds (Heckman et al. 1990) could be responsible for heating up a large mass of the ISM. For Arp 220 many typical features of an AGN are reported, too (Sanders and Mirabel 1996). An AGN could therefore power the warm component.

© European Southern Observatory (ESO) 1997

Online publication: April 28, 1998

helpdesk.link@springer.de