Astron. Astrophys. 354, 411-422 (2000)

4. Data analysis results: individual objects

Below, we first provide a brief review of the multi-wavelength properties of the individual sources and then report the results from our analysis of the X-ray data.

4.1. QSO 0117-2837

QSO 0117-2837 (1E 0117.2-2837) was discovered as an X-ray source by Einsteinand is at a redshift of z=0.347 (Stocke et al. 1991). Grupe (1996) classified it as NLSy1. It is serendipitously located in one of the ROSAT PSPC pointings; the steep X-ray spectrum was briefly noted by Schwartz et al. (1993) and Ciliegi & Maccacaro (1996). We present here the first detailed analysis of the ROSAT observations of this AGN.

When the X-ray spectrum is fit by a single powerlaw continuum with Galactic cold absorption of cm^-2 (Dickey & Lockman 1990), we derive a photon index (-4.3, if is treated as free parameter). The overall quality of the fit is good (), but there are slight systematic residuals around the location of absorption edges.

A successful alternative description is a warm-absorbed flat powerlaw of canonical index. We find a very large column density in this case, and the contribution of emission and reflection is no longer negligible; there is also some contribution to Fe K. For the pure absorption model, the best-fit values for ionization parameter and warm column density are , ( is now consistent with the Galactic value), with = 0.74. Including the contribution of emission and reflection for 50% covering of the warm material as calculated with Cloudy gives ( = 0.65). We note that for these large column densities, the optical depth to electron scattering becomes significant. The main purpose of the present study was to check under which conditions a warm absorber model fits at all; more detailed modelling should await the availability of deeper observations and improved spectral resolution.

Several strong EUV emission lines are predicted to arise from the warm material. Some of these are: FeXXI2304/H = 10, HeII1640/H = 16, FeXXI1354/H = 37, FeXVIII975/H = 16, NeVIII774/H = 9, and FeXXII846/H = 113. No absorption from CIV and NV is expected to show up. Both elements are more highly ionized.

Alternatively, the spectrum can be fit with a flat powerlaw plus soft excess (Table 2). E.g., assuming a black body shape we derive keV for ().

Table 2. Comparison of different spectral fits to QSO 0117-2837, RX J0134-4258 and Mrk 1298: (i) single powerlaw (pl), (ii) accretion disk model after Shakura & Sunyaev (1973), and (iii) warm absorber. was fixed to -1.9 in (ii) and (iii), except for RX J0134-4258, where = -2.2 (see text). Instead of individual error bars we provide several models that successfully describe the data.
Notes:
(1) free, if
(2) fixed to
(3) in 10²¹cm^-2
(4) in 10⁵, fixed
(5) survey obs. pointed obs.

An analysis of the temporal variability reveals constant source flux within the 1 error during the observation.

4.2. RX J0134-4258

Discovered in the ROSAT survey (Greiner 1996), the object was optically identified as NLSy1 galaxy (Grupe 1996) with redshift z=0.237. The later pointed PSPC observation led to the detection of strong spectral variability (Greiner 1996, Grupe 1996, Mannheim et al. 1996, Komossa & Fink 1997d). Here, we present the first detailed analysis of the X-ray properties of this peculiar source. ³The kind of variability of RX J0134-4258 is rare, and provides important constraints on the intrinsic X-ray emission mechanisms and/or the properties of surrounding reprocessing material.

RASS. When fit by a single powerlaw, the spectrum of RX J0134-4258 turns out to be one of the steepest among NLSy1s with -4.4 (absorption was fixed to the Galactic value in the direction of RX J0134-4258, cm^-2). A warm-absorbed, intrinsically flat powerlaw provides a successful alternative fit to the RASS data. Due to the low number of available photons, a range of possible combinations of U and explains the data with comparable success. A large column density (of the order 10²³ cm^-2) is needed to account for the ultrasoft observed spectrum. When we fix = -2.2, the value observed during the later pointing, and use = , we obtain and . This model gives an excellent fit ().

A number of further models were compared with the observed spectrum. E.g., an accretion disk model was fit. Again, we fixed =-2.2. The black hole mass is not well constrained by the model and was fixed ( ). We find 0.1 and, again, a very good fit is obtained (Table 2). If instead the spectrum is fit by a single black body, one derives a temperature keV.

Pointed observation. The fit of a single powerlaw to the spectrum of RX J0134-4258 yields a photon index = -2.2 ( = 1.4), much flatter than during the RASS observation. The amount of cold absorption was fixed to the Galactic value (if treated as free parameter, the Galactic value is underpredicted). For this model fit, two kinds of residuals are visible: (i) the first data point (below 0.15 keV) indicates a higher countrate than predicted by the model. This data point significantly influences the value of , and if it is excluded from spectral fitting, we obtain = 1.0 and = -2.1. Formally, a very low-temperature soft excess could be present in the spectrum of RX J0134-4258. Indeed, such a model can be fit with 0.1 keV. Hints for a similar very soft excess have been found in the ROSAT spectra of TON S180 (Fink et al. 1997) and NGC 4051 (Komossa & Fink 1997d). However, since such a component is essentially only constrained by the first few data bins we do not discuss this possibility in further detail. Another possibility is uncertainties in the calibration at these low energy channels. The second deviation from the powerlaw is (ii) an underprediction of the countrate in the energy range 0.4-0.9 keV (Fig. 2) indicative of the presence of absorption edges, as observed in AGNs where warm absorbers are present. However, again, the deviations from the powerlaw are only defined by few bins, and we thus assume in the following that the spectrum during the pointed observation essentially represents the intrinsic, un-distorted continuum (a complete description might invoke both, a weak soft excess and weak warm absorption, but fitting such models would definitely be an overinterpretation of ROSAT data).

Fig. 2. X-ray spectrum (pointed obs.) of RX J0134-4258 and residuals. Left : The upper panel gives the observed X-ray spectrum (crosses) and powerlaw model fit (solid line), the lower panel the residuals. Right : The same for a powerlaw plus black body fit (the quality of the fit is improved, but some systematic residuals around 0.4-0.9 keV remain). The spectrum was binned to a signal/noise of 8 per bin. The amount of cold absorption was fixed to the Galactic value.

Temporal analysis: The countrate during the pointed observation turns out to be variable by about a factor 2. The lightcurve is displayed in Fig. 3.

Fig. 3. X-ray lightcurve of RX J0134-4258 (pointing) binned to time intervals of 400s. The time is measured in seconds from the start of the observation.

4.3. NGC 4051

NGC 4051 has been classified as Seyfert 1.8 (e.g., Rosenblatt et al. 1992) or NLSy1 (e.g., Malkan 1986) and is at a redshift of z = 0.0023. This galaxy has been been observed with all major X-ray satellites (e.g., Marshall et al. 1983, Lawrence et al. 1985, Matsuoka et al. 1990, Mihara et al. 1994, McHardy et al. 1995, Guainazzi et al. 1996, Komossa & Fink 1997a; for brief summaries of these papers see Sects. 1 and 5 of Komossa & Fink 1997a). Recently, first BeppoSAX results have been presented by Guainazzi et al. (1998a), who report the detection of a strong drop in source flux which lasted the whole observing interval of 2 d.

Here, we present an analysis of all ROSAT PSPC data of this source, including previously unpublished observations and a homogeneous re-analysis of published ones (McHardy et al. 1995). Since NGC 4051 is strongly variable in X-rays, the large set of ROSAT data is very valuable to create a long-term lightcurve of this source and to study variability mechanisms. It also provides an excellent data base to study long-term spectral changes due to the presence of the warm absorber and places tight constraints on the ionization state of the warm material.

To investigate the long-term trend in the variability of NGC 4051, in countrate as well as in ionization parameter U and column density of the warm absorber, we have fit our warm absorber model to the individual data sets. We find that in the long term all features are variable, except for the cold absorption which is always consistent with the Galactic value within the error bars. Ionization parameter U and column density change by about a factor of 2. The slope of the powerlaw remains rather steep (Table 3).

Table 3. Log of ROSATPSPC observations of NGC 4051 and warm absorber fit results. = effective exposure time, CR = mean count rate, = mean (0.1-2.4 keV) luminosity corrected for cold and warm absorption.
Notes:
*) results of model fits uncertain due to off-axis location of source

The long-term lightcurve reveals large-amplitude variability by a factor 30 in countrate within the total observing interval. The X-ray lightcurve is displayed in Fig. 5.

Fig. 5. Long-term X-ray lightcurve of NGC 4051, based on all pointed ROSAT PSPC observations of this source. NGC 4051 is variable by a factor 30 in countrate. The lightcurve of Nov. 16, 1991 was earlier shown in McHardy et al. (1995), the one of Nov. 1993 in Komossa & Fink (1997a). The time is measured in s from the beginnings of the individual observations; the insets in each panel give the starting times.

4.4. Mrk 1298

Mrk 1298 (PG 1126-041) is a luminous Seyfert 1 galaxy at redshift z = 0.06 (Osterbrock & Dahari 1983). Its optical spectrum (Rafanelli & Bonoli 1984, Miller et al. 1992) is characterized by strong FeII emission line complexes. Mrk 1298 was part of several studies of correlations between strength of FeII and other spectral properties (Boroson & Green 1992, Wang et al. 1996 (WBB96 hereafter)). A UV spectrum of Mrk 1298 was presented by Wang et al. (1999). The ROSAT PSPC X-ray spectrum was first analyzed by WBB96 who detected an absorption edge which they interpreted as arising from a warm absorber. We present here a more detailed analysis of the properties of the warm absorber (see also Komossa & Fink 1997d,e), including predictions of non-X-ray emission lines expected to arise from the warm material, and test for the presence of a dusty warm absorber. We also analyze the temporal behavior of the X-ray flux.

A single powerlaw does not provide a successful X-ray spectral fit. We find =3.3 and the amount of cold absorption underpredicts the Galactic value in the direction of Mrk 1298, cm^-2. If Galactic absorption is enforced, the quality of the fit becomes worse (=4.3). Therefore, a number of further spectral models was fit, including an intrinsically flat powerlaw plus black-body like soft excess. The latter model gives =3.4 (Table 4), still unacceptable.

On the other hand, a warm absorber fits the X-ray spectrum well. We fixed the photon index of the intrinsic powerlaw to =-1.9. In a first step, cold absorption was fixed to the Galactic value. We then obtain and and the quality of the fit is acceptable (=0.95). Slight systematic residuals remain at very low energies. Thus, in a second step, was treated as free parameter. In this case we find some excess absorption, the fit is further improved (=0.78), and the residuals disappear (Fig. 1). The warm absorber parameters change to and 22.5. The cold absorption amounts to cm^-2.

Fig. 1. X-ray spectra and residuals of the fit for QSO 0117-2837(left) and Mrk 1298(right). The upper panel gives the observed X-ray spectrum of each galaxy (crosses) and the model fit (solid line). The lower panel shows the residuals. QSO 0117-2837 (left): upper panels: single powerlaw, lowest panel: warm-absorbed flat powerlaw. Mrk 1298 (right): upper panels: single powerlaw, lowest panel: warm-absorbed flat powerlaw.

Finally we note that the model of a dusty warm absorber does not give a successful X-ray spectral fit provided the intrinsic powerlaw spectrum is close to =-1.9.

The X-ray temporal analysis (Fig. 6) reveals rapid variability of the source with repeated changes in countrate by a factor 2 within 800 s.

Fig. 6. X-ray lightcurve of Mrk 1298. Each point encloses a time interval of 800s. Rapid variability on short timescales is revealed.

4.5. 4C +74.26

4C +74.26 is a radio-loud quasar at z=0.104 (Riley et al. 1988). The RASS data of this source were analyzed by Schartel et al. (1996a) who derived . In a study of ROSAT and ASCAdata, Brinkmann et al. (1998) confirmed the unusually flat soft X-ray ROSAT spectrum ( ; as compared to typically seen in nearby radio-loud quasars), found a steeper ASCApowerlaw spectrum, and evidence for the presence of a warm absorber. We re-analyzed this source, since the description given in Brinkmann et al. was highly suggestive of the presence of a dusty warm absorber.

The fit of a single powerlaw model yields an acceptable fit () but an extremely flat spectrum with =-1.4. Applying the model of a dusty warm absorber to the ROSAT spectrum we get a successful spectral fit with a steeper intrinsic powerlaw. In particular, we fixed the photon index to =-2.2 since we wanted to test whether the data are consistent with the general expectation for radio quasars. In this case we obtain a column density of the dusty warm gas of = 21.6 and an ionization parameter of ( = 1.0).

© European Southern Observatory (ESO) 2000

Online publication: February 9, 2000
helpdesk.link@springer.de