Astron. Astrophys. 363, 1123-1133 (2000)

3. Results

In the following presentation of our results, we focus on a comparison with the Galactic counterparts, especially for the FIR line-to-continuum ratio.

3.1. [CII] line emission

The derived [CII] flux () is shown in Table 1 with the spatial offset x (positive in the northeast) relative to the nucleus of M31 along its major axis. The [CII] line was unresolved in each spectrum at the resolution. Thus, we discuss the line flux integrated over wavelength only. The listed uncertainty in represents statistical one () based on the residuals of the spectral fit at the baseline regions. The correction for an extended source is not adopted because the correction factor is not yet fixed (Swinyard et al. 1998). The uncertainty in this correction does not affect the ratio (Sect. 3.3) based on the LWS observations significantly, because of the similar distributions (Sect. 3.2) of the two emissions. At an offset of , where the line emission is weak, the average spectrum of two adjacent observed positions is analyzed to further reduce the noise.

Table 1. Observed FIR emission along the major axis of M31. For an offset of , the average of two adjacent observed positions is listed. The line-to-continuum ratio is derived from the LWS line and continuum at ; the LWS line and HiRes continuum at .
Notes:
a) The uncertainty represents statistical one (). The upper limit corresponds to a level.
b) The uncertainty is 0.9 Jy ().
c) The uncertainty includes that in the flux calibration of the LWS and HiRes observations at .
d) The line flux was not derived because of the contamination due to cosmic-ray hits.

We adopted a conservative upper limit of a level for a non-detection because a low-level signal is sometimes hard to separate from glitches due to cosmic-ray hits. Representative spectra are displayed in Fig. 1 for detection at and non-detection at ; the derived line flux for the latter is slightly below our criterion for detection. At , we failed to derive the line flux because of the significant contamination due to cosmic-ray hits.

Fig. 1a and b. Representative [CII] line spectra observed along the major axis of M31. Second-order baselines were subtracted. a  At , detection. b  Averaged at and , non-detection.

Fig. 2a shows the observed [CII] flux as a function of the offset x. The flux uncertainties and upper limits are equivalent to those in Table 1. The horizontal bars indicate the size of the spatial sampling profile: the FWHM (; Lloyd 1999) of the LWS beam at ; at , the sum of the beam FWHM and the separation () between the two adjacent averaged positions.

Fig. 2a and b. Distributions of FIR emission along the major axis of M31 (crosses; bars with arrows for upper limits), plotted as a function of the spatial offset x (positive in the northeast) relative to the nucleus of the galaxy. The M31 data are quoted from Table 1. The horizontal bars indicate the size (FWHM) of the spatial sampling profile. The dotted curves simulate our Galaxy (Nakagawa et al. 1998; the ISSA), located at the distance of M31 and observed with the LWS. The flux and flux density of our Galaxy were multiplied by . Extended-source correction is not applied for the LWS nor IRAS data. a  [CII] line flux (; this work). The upper dotted curve indicates the flux corrected for the possible offset of the Galactic observations. The lower dotted curve is for that without the correction. b  Continuum flux density at . The dashed curve indicates IRAS HiRes brightness in M31 multiplied by the solid angle of the LWS beam.

For comparison, we simulated LWS observations of our Galaxy located at the distance of M31, as also shown in Fig. 2. The [CII] map of Nakagawa et al. (1998) was regridded under the assumption , where and are the distances to the centers of M31 and our Galaxy, respectively. Then the map was convolved with a Gaussian profile, in order to produce the final spatial resolution equal to that of the LWS. Since the BICE observations have a possible offset of in the [CII] intensity (Nakagawa et al. 1998), the simulation is shown for the offset-corrected flux as well as that without the correction.

An extended [CII] component was detected over the central 1.6 kpc () of M31, while upper limits were obtained at most of the outer regions, as shown in Fig. 2a. The [CII] emission observed in M31 is different from that in our Galaxy as follows:

1. The [CII] emission detected in M31 is fainter by one to two orders of magnitude.

2. The [CII] emission in M31 lacks a spatially unresolved nucleus component, which is obvious in our Galaxy.

3. The central kiloparsec region of M31 is brighter than its disk region just outside, while the Galactic center (except the nucleus component) is less bright (Paper I) than the Galactic disk shown.

Points 1 and 2 above may be accounted for by a smaller amount of the ISM and less active recent star formation in the inner disk and around the nucleus of M31. The weak emission in M31 was also pointed out in other tracers of the ISM, such as the FIR continuum (Walterbos & Schwering 1987), the CO (-0) line (Dame et al. 1993), and the HI 21 cm line (Brinks & Shane 1984).

3.2. 100 µm continuum emission

The continuum flux density, , at observed with the LWS is listed in Table 1. The LWS continuum data can be compared with the brightness in the band of the IRAS High Resolution processing (HiRes; Rice 1993) data. The calibration correction for extended sources is not adopted for nor (Wheelock et al. 1994). The distribution of the LWS flux density is shown in Fig. 2b along with the HiRes flux density , where (Lloyd 1999) is the effective solid angle of the LWS beam at . The two datasets agree well at as shown in Fig. 2b, in spite of the slightly larger beam ( in FWHM) of the HiRes data. This is probably due to the relatively smooth distribution of the FIR emission in the observed regions and to the similarity in the extended-source correction factor between the two datasets.

We evaluated the uncertainty from the deviation of from (Sect. 2) at , where the difference in the beam sizes hardly affect the brightness because the distribution of the FIR emission is quite smooth as shown in Fig. 2b. The estimated uncertainty is 0.9 Jy () in . This value is comparable to at , where this uncertainty was estimated; the estimated uncertainty is practically equivalent to the fluctuation in the derived at the background regions without detectable emission. Representative spectra at a wavelength range of are displayed in Fig. 3 for detection at and non-detection at ; the emission at shorter wavelengths does not contribute to discussed in the present paper.

Fig. 3a and b. Representative spectra at a wavelength range of observed along the major axis of M31. a  At . b  At .

The emission is distributed similarly to the [CII] emission in M31 as shown in Fig. 2. The central component has a width of (), approximately equal to that of the [CII] emission. For comparison, the simulation of our Galaxy located at the distance of M31 is shown for the continuum (the IRAS Sky Survey Atlas; ISSA) in Fig. 2b, as for the [CII] line in Fig. 2a. Extended-source correction for the ISSA (Wheelock et al. 1994) is not applied to Fig. 2b. In contrast to M31, the line and continuum FIR distribution is quite different in our Galaxy: the emission lacks the depression in the central kiloparsec seen in the [CII] emission.

3.3. [CII]/100 µm ratio

The FIR line-to-continuum flux ratio of is derived from the LWS data at (Table 1). The uncertainty in the ratio at these positions is based on those in and listed in Table 1. On the other hand, at the LWS continuum data have insufficient signal-to-noise ratios. Since the FIR brightnesses at these positions are hardly affected by the difference in beam sizes because of the smooth distribution of the emission (Fig. 2b), we took instead the ratio to the HiRes data there: . In this case, the uncertainty in the ratio includes the flux calibration uncertainties of the two different instruments (15% for the [CII], Sect. 2; 10% for , Xu & Helou 1994), as well as the statistic uncertainty in .

The distribution of in M31 is plotted in Fig. 4. The ratio is for the central kiloparsec of M31. The upper limits for the outer regions are consistent with a constant ratio of through the observed regions in M31. This distribution is in contrast to the Galactic distribution also shown in Fig. 4. For derivation of the Galactic ratios, the extended-source correction was applied to the ISSA data: the flux was multiplied by 0.72 (Wheelock et al. 1994). Our Galaxy has a nearly constant ratio of in the disk, with 2-3 times lower in the central kiloparsec, as reported in Paper I. The ratios in the central kiloparsec of M31 are 2-3 times higher than those in the Galactic counterpart and closer to those in the Galactic disk outside. In the outer regions, no difference between the galaxies is found.

Fig. 4. Distribution of the FIR line-to-continuum ratio along the major axis of M31 (crosses; bars with arrows for upper limits), plotted as a function of the spatial offset x relative to the nucleus of the galaxy. The M31 ratios with the uncertainties are from Table 1. The horizontal bars indicate the size (FWHM) of the spatial sampling profile. The dotted curves simulate our Galaxy (Nakagawa et al. 1998; the ISSA), located at the distance of M31 and observed with the LWS beam. The two dotted curves correspond to the cases with and without the correction for the possible [CII] offset, as in Fig. 2a. For derivation of the Galactic ratios, the extended-source correction was applied to the emission.

The lower ratios observed in our Galactic center were ascribed to the radiation from late-type stars in the Galactic bulge (Paper I): the soft radiation illuminating the neutral ISM heats the dust grains, but does not heat the gas effectively (de Jong et al. 1980). However, the central kiloparsec of M31 does not show low ratios, in spite of its bright (e.g., Martinez Roger et al. 1986) bulge.

© European Southern Observatory (ESO) 2000

Online publication: December 5, 2000
helpdesk@link.springer.de