3 PAH 7.7 $\mu$m/850 $\mu$m measurements of high extinction

Figure 1 (y-axis) shows the PAH 7.7 $\mu$m/850 $\mu$m distribution of the ULIRGs and the reference sample.

Strikingly, all ULIRGs except Arp 220 and UGC 5101 lie in a confined range, which is also the same as for the normal galaxies. This suggests that for both samples the PAH and the sub-mm emission are related, and that no extraordinary excitation conditions are needed. The ISM in the galactic disk exhibits PAH emission with a low ISRF over a large range of different environmental parameters (Mattila et al. 1996). Léger & Puget (1989) predicted that PAH formation takes place in molecular clouds, and these are typically bright at 850 $\mu$m.

 

 

Table 1: Observed and dereddened luminosities of Arp 220

	deredd.	$L_{\rm MIR}$	deredd.	$L_{\rm FIR}$	$L_{\rm MIR}$/
	factor	10 $^{\rm 11}L_{\odot}$	factor	10 $^{\rm 11}L_{\odot}$	$L_{\rm FIR}$
observed	1.0	2.1	1.0	9.0	0.23
MRN: AV=330	10 (4...20)	$\approx$ 21.0	3.2	28.8	0.72
KS: AV=110	4.0	8.4	1.7 (1.3 ... 2.0)	15.3	0.54
KS: $A_V{\rm screen} = 90^{\rm *}$	20.0	42.0	"	"	2.75

$^{\rm *}$ Screen case on central MIR continuum, while the mixed case extinction probably still applies to the FIR emission.

UGC 5101 has a high mixed case extinction $A_V \approx 50$ derived from NIR-MIR spectroscopy (Genzel et al. 1998). Dereddening shifts it clearly into the range of the other ULIRGs (Fig. 1).

The PAH 7.7 $\mu$m/100 $\mu$m ratio is lower for the ULIRGs than for the reference sample by a factor of about three (x-axis in Fig. 1). ULIRGs have warmer dust ( $30~{\rm K}<{T}<50~{\rm K}$) than normal galaxies ( $20~{\rm K}<{T}<30~{\rm K}$), and therefore their 100 $\mu$m flux relative to that at 850 $\mu$m is higher. Again, along the PAH 7.7 $\mu$m/100 $\mu$m distribution Arp 220 lies below the other obviously more "typical ULIRGs'' which populate a confined range. We consider three possibilities to explain the exceptional position of Arp 220 in Fig. 1:

1.

A low PAH 7.7 $\mu$m flux due to destruction of the PAH carriers by a hard UV radiation field is unlikely, since this should be present in the other ULIRGs with exposed AGN like Mrk 231 and 0518. Yet there is no hint for such a field in Arp 220. If present, it would already point towards a hidden AGN.

2.

An excess of 850 $\mu$m flux due to very cold dust which also has to be deficient in PAH emisssion: for example, the extremely quiet galaxy M 31 has very cold dust at $T \le 16$ K (Haas et al. 1998) and it could be a low PAH emitter (Cesarsky et al. 1998). But in order to achieve the desired shift in Fig. 1, such very cold dust (even without any PAH emission) has to provide about half of the FIR emission and more than three quarters of the sub-mm emission of Arp 220. In this case the dust mass would approach (or even exceed) the molecular gas mass derived from the CO luminosities (Downes & Solomon 1998), resulting in a gas-to-dust ratio of about 5, unrealistically low compared with the standard value of about 150. Also, bearing in mind the ultraluminosity of Arp 220, a low excitation of the PAH carriers is unlikely.

3.

Extinction: since the first two mechanisms do not provide a satisfactory explanation, we conclude that the low PAH 7.7 $\mu$m/850 $\mu$m flux ratio of Arp 220 is mainly caused by extinction. Further evidence for a significant extinction comes from the deep silicate absoption feature (Dudley 1999). A comparison with UGC 5101 immediately suggests that the MIR extinction is very high.

To summarize, 13 out of 15 ULIRGs as well as 20 comparison galaxies populate the same confined range of the PAH 7.7 $\mu$m/850 $\mu$m flux ratio. Their MIR extinction may be moderate ( -10), so that NIR-MIR spectroscopy can yield proper results. Also, it is not likely that they contain a hidden powerful AGN which has not yet been identified as such. Two ULIRGs, however, lie significantly below this range and their offsets are mainly due to high MIR extinction. The new PAH 7.7 $\mu$m/850 $\mu$m diagnostics is a promising tool to reveal high MIR extinction in ULIRGs.