3. Results and Discussion
Fig. 1 shows the ISO raster maps at 120 and 200 µm, those at 150 and 180 µm look very similar. At the spatial resolution of 15 most of the emission is seen by two pixels, the central one and its southern neighbour. The footprint (120 µm: = , = ; 200 µm: = , = ) is larger than the pixel size of and the pointing accuracy of the ISO satellite is better than . First we examine, how far the emission could originate from a point source located near the border between the two pixels. Then the point source flux should be roughly the sum of their flux around 2 Jy. Since the ISO serendipity survey with the C200 detector could undoubtedly detect 2 Jy sources from much faster slews across the sky (Bogun et al. 1996, Stickel et al. 1998), a point source would be bright enough to show up as a clear peak in the signal time series, when the telescope slewed between the two raster positions. But we could not find such a peak as is illustrated in Fig. 2. Therefore we are convinced that the flux seen by the two bright pixels does not originate from a point source between them, rather it is due to resolved emission.
Fig. 3 illustrates the location of the two bright ISO map pixels with respect to the HI contours. While the central pixel covers the galaxy nucleus and the HI peak ( cloud 11) north of it (both also measured at 1.1 mm by Fich and Hodge, 1991), the bright HI maximum south of the nucleus coincides with the southern pixel. Thus the dust emission might be correlated with the HI emission.
Fig. 4 shows the cuts through the ISO maps along north-south direction. They further reveal that, while at 120 and 150 µm the center constitutes the brightest peak, at 180 and 200 µm the southern region reaches similar brightness as the center. This indicates that the center of NGC 205 is slightly warmer than the southern region.
Photometry was derived for the total galaxy as well as the center and the southern region, taking the outer map parts (2 pixels = wide) for background subtraction. The total flux is estimated from the inner 7575 (55 pixels), and the center and southern flux from the central and the southern pixel alone applying a footprint correction and accounting for the reciprocal contributions. The fluxes are listed in Table 1. Remarkably about 40-50 of the total flux comes from the extended ring area outside the (footprint corrected) central and southern regions, but within 7575. Also, the intermediate part of 4545 yields fluxes between the total one and the sum of the central and southern regions giving further support for a substantial contribution by extended dust. Attempts to integrate the flux over larger areas than 75 yielded no increase, but became more noisy, because of improper background determination. Though there could be more far extended flux, our total flux from 7575 might well approximate the ultimate one.
Table 1. Isophot photometry of NGC 205 and resolved components (uncorrected gives the flux of the single pixel before footprint correction, extended ring = total - center - south ). The fluxes are in Jy. The errors are about 10 of the quoted fluxes.
Fig. 5 shows the spectral energy distribution for the total galaxy and the center supplemented by IRAS and mm data. The ISO measurements now clearly reveal the maximum of the SED around 120-150 µm and the Rayleigh-Jeans decline longward thereof. In order to characterise the dust emission, power-law modified blackbody curves are considered with emissivity exponent = 1, 1.5 and 2. With = 1.5 and 2 the ISO and IRAS data points are nicely fitted and the effect of rather shows up longwards of 200 µm. Fig. 6 shows the deviation of the fits from the IRAS and ISO data for the total galaxy. The shallow = 1.0 provides the worst fit, and = 2.0 appears somewhat closer to the data than = 1.5. Therefore we reject = 1.0 and restrict the further discussion to = 1.5 and 2. The value of plays a crucial role for interpretations and is still a matter of debate. Exact observational determination remains difficult, since most sources exhibit a mixture of various dust temperatures. While moderate approaches prefer the shallower value of = 1.5 (and sometimes even = 1.0), more and more observational examples are found favoring the steeper value of = 2 for the typical interstellar dust at temperatures between 20 and 50 K (e.g. Chini et al. 1995, Lehtinen et al. 1998). Since we can not definitely decide on for NGC 205, we discuss the data for both cases.
As zoomed in Fig. 5 (bottom), for the total galaxy both = 1.5 and 2 yield extrapolated 1.1 mm fluxes (61 and 31 mJy) which are clearly above the one measured for the nucleus (21 mJy). This provides compelling evidence for additional 1.1 mm flux from outside of the nucleus, the missing flux must be the higher the shallower is. Since the mm beam was smaller than the ISO beam, one should try to obtain a more realistic 1.1 mm flux estimate which compares better to the ISO center. Simply scaling up the measured mm flux proportional to the five times larger area would yield about 100 mJy which appears somewhat high and arbitrary. Therefore we consider the 5 mJy observations of cloud 11 (though not a 3 detection, Fich and Hodge, 1991), which coincides with the HI emission north of the nucleus and is also covered in the ISO cen er pixel (Fig. 2). Then nucleus and cloud 11 together give a 1.1 mm flux of 21 + 5 = 26 mJy which may still represent a lower limit for the center. = 1.5 as well as = 2 yield extrapolated fluxes for the center (15 and 7 mJy) clearly below the one for the nucleus and cloud 11. If = 2, then the discrepancy (26 - 7 = 19 mJy) is large enough to infer already the existence of an additional component of very cold dust. But also for the case = 1.5 the discrepancy ( mJy) since a lower limit suggests an additional component of very cold dust. A component with a temperature of 8 K would ideally fulfill this requirement. The one shown in Fig. 5 has a 1.1 mm flux of 14 mJy ( = 2). A component warmer than 10K would show up at 200 µm.
Table 2 lists the derived dust parameters for the blackbody curves in Fig. 5. The dust mass was estimated with the same formula (2) and basic dust grain assumptions as stated by Fich and Hodge (1991). For comparison they predicted a temperature between 20 and 26 K depending on the emissivity and give a lower limit for the dust mass of about 3000 for = 2. Despite the rough agreement of their and our temperature and mass estimates for the cold dust which lead to similar conclusions about the extinction and clumpiness of the dust allocations, the main advantage of the new ISO measurements consists of measuring the SED at its maximum and of spatially separating the center, allowing to infer for the first time the presence of verycold dust below 10 K in an elliptical galaxy. Since the very cold dust is derived in the center of NGC 205, it is likely to find a larger fraction outside as well. For comparison, in the late type Andromeda galaxy very cold molecular clouds have been detected via CO observations by Allen and Lequeux (1993). The nature of very cold dust is quite unknown. Probably it has larger grain sizes up to 1 µm, like found in the solar system with the ULYSSES satellite (Grün et al. 1994). Further submm and mm observations should be able to extend the conclusions beyond the finding of the existence of very cold dust in NGC 205 reported in this letter.
Table 2. Dust temperature, luminosity ( = ) and mass for NGC 205 and the resolved components, derived from the blackbody curves shown in Fig. 5. The second component for the center at 8 K refers to a 1.1 mm flux of 7 and 19 mJy at = 1.5 and 2, resp.
© European Southern Observatory (ESO) 1998
Online publication: August 6, 1998