3. ISO results
3.1. The light curves
The data and the corresponding light curves at 4.0 (SW5 filter), 14.3 (LW3), 60 (C1_60) and 90 µm (C1_90) are reported in Tables 4 and 5 and shown in Figs. 1 and 2. The discussion on the data analysis and error evaluation is given in Appendix A. At 170 µm (C2_160), the source is not detected: the three sigma upper limit at this wavelength is 1235 mJy (see Fig. 3).
When the purpose is to verify whether the flux is variable, the contribution of the pixel responsitivity to the absolute error can be neglected and a smaller uncertainty can be associated to the relative flux values of the light curves. However, this can be done only for the two light curves of the photometer (see Table 5), due to the way the photometric error was determined.
The relative errors on the flux are, in any case, quite large, about 10 - 12% for the camera observations and from 20 to more than 50% for the photometer (see Appendix A). Within these uncertainties the light curves show no evidence of variability. To quantify this statement, we fitted the light curves with a constant term and the reduced chi-square values were computed in order to test the goodness of the fits. We first fitted the values of the best sampled period, from 1996 May 13 to May 27. The results are mJy at 4.0 µm, mJy at 14.3 µm, mJy at 60 µm and mJy at 90 µm. To fit the data at 4.0 µm the lower limits were neglected. We then repeated the fits, taking the mean of the above-mentioned period and adding the other data, to look for possible longer-term variability. All the fits are acceptable within a confidence level of 95%. This means that PKS 2155-304 showed no evidence of variability at these wavelengths in the observed period.
However, the large uncertainty on the flux can hide smaller variations. We calculated the mean relative error and obtained 3 sigma limits for the lowest detectable variations of 32%, 36%, 76% and 132% at 4.0, 14.3, 60 and 90 µm, respectively.
3.2. The infrared spectrum
The infrared spectral shape of PKS 2155-304 was sampled, using 16 filters, from 2.8 to 170 µm. The photometer filter C2_200 was not considered reliable enough and its observation was discarded. The flux values are given in Table 6 and the spectrum is shown in Fig. 3, in a representation.
In Fig. 3 it is also shown the result of a power law fit, that gives an energy spectral index of . The lower and upper limits were not considered in the fit; the reduced chi-square is , with 9 d.o.f., that gives a confidence level of 77.4%.
From each simultaneous pairs of flux values of the SW5 and LW3 light curves, we obtained the spectral indices between 4.0 and 14.3 µm as
The mean value is , which is fully consistent with the index derived using 11 filters on a larger IR band.
The fit with a constant term of the spectral indices vs. time has a reduced chi-square of 0.26, with 13 d.o.f., which corresponds to a confidence level of less than 1%. This indicates that the source showed no spectral variability in the 4.0 - 14.3 µm range, during the observed period.
© European Southern Observatory (ESO) 2000
Online publication: March 28, 2000