Our targets are so faint that we had to use quite long integration times. Nevertheless, the spectra are rather noisy. Therefore our measurements are not very accurate. One major problem was to define the continuum, which showed in some cases an unrealistic apparent wavelength dependence (see in particular Fig. 1 and Fig. 4). We did not subtract the off-source from the on-source spectra since the off-source spectra do not correspond to white noise but show spurious structures. This introduces an additional uncertainty in the measurement of the continuum. In particular the equivalent width of the 53 µ line may be affected by this.
It appears that our measurements were at the limit of what could be achieved with the ISO spectrometers. To our knowledge this is the first detection of OH lines in LWS spectra of OH megamaser galaxies.
3.1. IRAS 20100-4156
The lines at 34 µ and 53 µ are clearly detected in absorption (Figs. 1 and 2). There appears to be some absorption at the position of the 119 µ line (Fig. 3), but it is definitly much weaker than one would expect according to the strengths of the lines at 34 µ and 53 µ. From the latter one may derive an estimate for the OH column density. The spectral resolution (1400 and 200, respectively) is not sufficient to resolve the line profiles. (In the case of the 34 µ line, however, one can see that the line is wider than the instrumental profile, as expexted from the splitting.) The quantity to be measured is the equivalent width
For the redshift independent quantity we measured the values given in Table 2.
where is the column density of the molecules in the lower level and of those in the upper level. is the Einstein coefficient for absorption. Eq. (4) gives a (possibly rather crude) estimate of the coulumn density of the molecules in the lower level, which in our case is the rotational ground state. This estimate is close to the true value only if the line is optically thin and if If one wants to account for optical thickness effects one has to measure at least two lines and to calculate curves of growth. For this one has to know the broadening mechanism. In our particular case it is important to note that we measure the absorption by unresolved multiplets rather than by single lines. In calculating a curve of growth one therefore has to account for the individual components caused by doubling and hyperfine splitting.
We calculated curves of growth assuming that the individual components of the multiplets are broadened by a turbulent velocity field. We accounted for this in the microturbulent approximation. Furthermore we considered only the case of pure absorption , which is a reasonable assumption as long as the excitation temperature is well below the excitation energy of the upper level. Finally we assumed that the relative population of the individual levels of the rotational ground state corresponds to their statistical weight. This means, we calculated from (2) with
L is the number of components of a given multiplet (6 for the 34 µ line and 8 for the 53 µ line). The index i refers to the individual components. The designate the central frequencies while is an arbitrary reference frequency. The Doppler width is given by
where and () are the thermal and the turbulent velocities, respectively. The thermal velocity is negligible as compared to the turbulent velocity derived from the measured data. Since the number of components and the splitting is different for the two lines, we calculated separate curves of growth for the 34 µ and the 53 µ line, respectively, and determined the parameter pair and for which the computed values of of both lines coincide with the measured ones. In Table 3 we collect the most essential line data. The first column gives the (rest frame) wavelength, the second gives the corresponding excitation energy in units of K, the third one gives the Einstein value for absorption of the strongest hyperfine component, and the last column gives the doubling in units of km s-1.
The microturbulent velocity (b-parameter) is estimated to be km s-1. We note that this is the parameter used to calculate the curves of growth. The actual turbulent velocity may be larger if spatial correlations are important since in this case the curve of growth for a given value of starts to flatten at a lower column density (Levshakov & Kegel 1996).In addition an analysis based on a mesoturbulent model also may lead to a higher value of the column density. However, in view of the limited quality of our data we do not consider it appropriate to go into a more detailed analysis (which also would introduce one more free parameter). - The value (9) for the column density is in good agreement with the value one derives from using the crude estimate (4). This, as well as the curve of growth analysis, indicates that the 34 µ line is still close to the linear part of the curve of growth. This latter fact implies in retrospect that the value derived for the column density is quite independent of the details of the analysis.
It is obvious that the tentatively measured equivalent width of the 119 µ line is inconsistent with that of the other lines. Since the Einstein value for absorption of the 119 µ line is a factor of 57 larger than that of the 53 µ line, the equivalent width also should be larger. According to the model parameters determined from the 34 µ and the 53 µ line, the equivalent width of the 119 µ line should be a factor of 3 larger than the measured value. According to our observations (Fig. 3) so large a value can be ruled out. - One way to interpret the weakness of the 119 µ line would be to assume that the continuum source is more extended at 119 µ than at 34 µ and 53 µ and that at 119 µ the absorbing cloud does not fully cover the continuum source.(A covering factor of 0.3 would be consistent with the observation.) This assumption appears to be plausible if one considers the IR continuum to be emitted by dust heated by a central source. In this case the warmer dust, emitting at 34 µ and 53 µ, would be more locally concentrated than the cooler dust emitting at 119 µ. - Another explanation for the low absorption observed for the 119 µ line may be the possibility that the absorption in the light of the bright central source is partially compensated by line emission from other parts in the galaxy. The upper level of the 119 µ line may be excited by collisions in warmer regions (see Table 3) leading to an emission line if there is no bright IR background. The spatial resolution of ISO ( 200 kpc for ) is low compared to the size of the galaxy.
As mentioned in the introduction, the primary aim of this project was to search for an emission in the 115 µ line. Such an emission could not be found with any certainty (see Fig. 4). At the (redshifted) position of the 115 µ line there is with certainty no absorption. The apparent very weak emission is not reliable. One problem is the steep rise of the apparent continuum towards shorter wavelengths, which probably is caused by an instrumental effect since we see a similar rise in the off-source spectrum. - In view of the poor quality of the data the non-detection of the line does not allow any firm conclusion. Moreover, the argument given with respect to the line, that the continuum source may be sustantially more extended than at and , applies to the line as well.
3.2. 3 Zw 35
The spectra we obtained for 3 Zw 35 are considerably more noisy than those for IRAS 20100-4156. There seems to be some absorption at the positions of the 34 µ (Fig. 5) and the 53 µ line (Fig. 6). In the vicinity of the position of the 53 µ line there are other line-like structures which may be fringes. However a more detailed analysis could not identify a strict periodicity. The wavelength interval covered by our measurement is so short that this problem could not be resolved. (The quality of the LWS01 spectrum was also not sufficient.) If we take the measured equivalent widths of the 34 µ and the 53 µ line, respectively, at face value, we derive for the column density
and for the microturbulent velocity a value of km s-1. The spectra at the positions of the 119 µ and the 115 µ line are extremely noisy and no indication for a spectral line, neither in absorption nor in emission, could be detected
© European Southern Observatory (ESO) 1999
Online publication: November 3, 1999