Forum Springer Astron. Astrophys.
Forum Whats New Search Orders

Astron. Astrophys. 345, 49-58 (1999)

Previous Section Next Section Title Page Table of Contents

3. Results

Some typical spectra are plotted in Fig. 2 showing the range of intensity variations. Immediately one can recognize the strong variations in the continuum, in the Balmer lines and especially in the HeII[FORMULA]4686 line. The continuum variations are most pronounced in the blue section. The continuum gradient changes as a function of intensity. The emission line profiles of the Balmer lines in Mkn 110 are quite narrow (FWHM(H[FORMULA])=1800 km s-1) similar to those of the so called narrow-line Seyfert 1 galaxies. Week FeII emission is present in the spectra blending the red line wings of [OIII][FORMULA]5007 and H[FORMULA]. We measured the integrated intensity of FeII line blends between 5134 Å and 5215 Å. The main FeII components in this region are the 5169 Å and 5198 Å lines belonging to the multiplets 42 and 49. The FeII line flux (2.0 10- 14 erg s-1 cm-2) remained constant during our variability campaign within the error of 10%.

[FIGURE] Fig. 2. Normalized spectra of Mkn 110 taken at different epochs in Oct. 1988, Oct. 1989, March 1988, Feb. 1989, May 1989, Jan. 1992 (from bottom to top).

Difference spectra with respect to our minimum stage in October 1988 are plotted in Fig. 3. All narrow line components cancel out. The FeII lines disappear in the difference spectra as well. In the Balmer profiles (e.g. H[FORMULA]) very broad, slightly redshifted components stand out in the high intensity stages.

[FIGURE] Fig. 3. Difference spectra with respect to our minimum stage in Oct. 1988; otherwise same epochs as in Fig. 2.

3.1. Line and continuum variations

The results of our continuum intensity measurements at 3750 Å, 4265 Å, and 5100 Å as well as the integrated line intensities of H[FORMULA], H[FORMULA], HeII[FORMULA]4686, HeI[FORMULA]5876, and HeI[FORMULA]4471 are given in Table 2. The individual light curves are plotted in Fig. 4.


Table 2. Continuum and integrated line fluxes
Continuum fluxes (2)-(4) in 10-15 erg sec-1 cm- 2 Å-1.
Line fluxes (5)-(9) in 10-15 erg sec- 1 cm-2.

[FIGURE] Fig. 4. Light curves of continuum flux at 5100 Å and 4265 Å (in units of 10-15 erg cm-2 s- 1 Å-1) and integrated emission line flux of H[FORMULA], H[FORMULA], HeII[FORMULA]4686, HeI[FORMULA]5876 and HeI[FORMULA]4471 (in units of 10- 15 erg cm-2 s-1). The points are connected by a dotted line to aid the eye.

The continuum intensities are mean values of the wavelength ranges given in Table 3, Column (2). Line intensities were integrated in the listed limits after subtraction of a linear pseudo-continuum defined by the boundaries given in Column (3). All wavelengths are given in the rest frame.


Table 3. Boundaries of mean continuum values and line integration limits

We started our monitoring program in 1987. Therefore, our 5100 Å continuum light curve covers up a larger time interval than the light curve of Peterson et al. (1998a) beginning in 1992. The observing epochs are partly complementary in the common monitoring interval. But, spectra taken nearly simultaneously from both groups (within one week) correspond with each other to better than 5% in the continuum fluxes. In Fig. 5 we compare our continuum light curve with that of Peterson et al. (1988a) in the common interval of observations. Both light curves are in very good accordance regarding to the intensity variations. Fig. 5 shows that the data sets are not significantly undersampled. However, our H[FORMULA] intensities are systematically higher than those of Peterson et al. (1998a) as we integrated over a larger wavelength range and we carried out a slightly different continuum subtraction. This method led to a lower pseudo-continuum flux at 4790 Å. The H[FORMULA] fluxes are in perfect agreement if we multiply the values given by Peterson et al. (1998a) by a factor of 1.15.

[FIGURE] Fig. 5. Comparison of our continuum light curve at 5100 Å (filled circles) with that of Peterson et al. (1988a) between JD 2448600 and JD 2500000.

The pattern of the continuum light curves at 5100 Å and 4265 Å (Fig. 4) is identical apart from their different amplitudes. The HeII[FORMULA]4686 light curve follows closely these continuum light curves. The light curves of the Balmer lines H[FORMULA] and H[FORMULA] are similar among themselves and the light curves of both HeI lines as well.

Statistics of our measured continuum and emission line variations are presented in Table 4. We list minimum and maximum fluxes Fmin and Fmax, peak-to-peak amplitudes Rmax = Fmax/Fmin, the mean flux over the entire period of observations [FORMULA]F[FORMULA], the standard deviation [FORMULA], and the fractional variation


as defined by Rodríguez-Pascual et al. (1997). The extreme variability amplitudes of Mkn 110 attract attention compared to other galaxies, e.g. NGC 4593 (Dietrich, Kollatschny et al. 1994) and the Seyfert 1 galaxies from the sample of Peterson et al. (1998a).


Table 4. Variability statistics
Continuum flux in units of 10-15 erg sec- 1 cm-2 Å-1.
Line flux in units of 10-15 erg sec- 1 cm[FORMULA]

The variability amplitudes of the continuum increase towards the short wavelength region. These amplitudes as well as those of the emission line intensities are exceptionally high. The variability amplitude of the HeII[FORMULA]4686 line is unique compared to the other emission lines in Mkn 110 and compared to optical lines in other Seyfert galaxies. In Fig. 6 we plot the line intensity ratios HeII[FORMULA]4686/H[FORMULA] and HeI[FORMULA]5876/H[FORMULA] as a function of continuum intensity at 5100 Å. These line intensity ratios increase slightly for the HeI line but strongly for the highly ionized HeII line.

[FIGURE] Fig. 6. Line intensity ratios HeII[FORMULA]4686/H[FORMULA] (filled circles) and HeI[FORMULA]5876/H[FORMULA] (open circle) as a function of continuum flux at 5100 Å.

3.2. Balmer decrement

We calculated Balmer decrement H[FORMULA]/H[FORMULA] values in the range from 3.2 to 4.3. Simple photoionization calculations (Case B) result in a value of 2.8 for this line ratio (Osterbrock 1989). Deviations of the observed Balmer decrement from the theoretical value are often explained by wavelength dependent dust absorption and/or by collisional excitation effects. We will show later on that the observed difference cannot be explained by dust absorption alone in the broad-line region clouds of Mkn 110. There is a clear anti-correlation of the Balmer decrement with the continuum flux (Fig. 7). One has to keep in mind that the individual Balmer lines and the continuum of each spectrum originate in distinct regions of the BLR at different times. This will be confirmed by the cross-correlation analysis later on. A very tight correlation cannot be expected because of the short-term variations in this galaxy. The solid line in Fig. 7 is a linear fit to all our data points except for the lowest continuum intensity point. At this epoch (JD +5959) the H[FORMULA] intensity was extreme low (Fig. 4) in contrast to H[FORMULA] and the continuum. An anti-correlation of the Balmer decrement with the continuum flux has been first noted in NGC 4151 by Antonucci & Cohen (1983).

[FIGURE] Fig. 7. The Balmer decrement H[FORMULA]/H[FORMULA] as a function of continuum flux at 5100 Å. The minimum value of F 5100 was omitted for the linear fit.

3.3. UV spectra

Two UV spectra haven been taken with the IUE satellite with a time interval of one day only. These spectra have been taken nearly simultaneously (within 8 days) to our optical observations in March 1988. Therefore, these two spectra are suitable for a determination of optical/UV line intensity ratios. Fig. 8 shows an overplot of both short wavelength UV spectra taken with an time interval of 1 day. The spectra are identical in the continuum and in the emission lines within the error limits. The different flux values in the center of the CIV[FORMULA]1550 line are due to saturation effects in one of the spectra.

[FIGURE] Fig. 8. Short wavelength IUE UV spectra of Mkn 110 taken on Feb. 28 and Feb. 29, 1988

We determined an integrated Ly[FORMULA] flux of [FORMULA] erg s-1 cm- 2. Comparison with the optical spectra results in a Ly[FORMULA]/H[FORMULA] ratio of 11.0 at the observing epoch March, 1988.

The HeII[FORMULA]1640 flux amounts to [FORMULA] erg s-1 cm-2. The HeII[FORMULA]1640/HeII[FORMULA]4686 ratio of 2.45 is about a factor of two lower than that of typical photoionization models but consistent with other AGN observations (Dumount et al. 1998).

3.4. CCF analysis

An estimate of size and structure of the broad-line region can be obtained from the cross-correlation function (CCF) of a continuum light curve with emission line light curves.

We cross-correlated the 5100 Å continuum light curve with all our emission line light curves (Fig. 4) using an interpolation cross-correlation function method (ICCF) described by Gaskell & Peterson (1987). In Fig. 9 we plot the cross-correlation functions of the individual emission line light curves of HeII[FORMULA]4686, HeI[FORMULA]5876, H[FORMULA] and H[FORMULA] with the continuum light curve. The cross-correlation functions of HeI[FORMULA]4471 and HeI[FORMULA]5876 are identical within the errors; therefore, only one curve is shown in the plot.

[FIGURE] Fig. 9. Cross-correlation functions CCF([FORMULA]) of HeII[FORMULA]4686, HeI[FORMULA]5876, H[FORMULA] and H[FORMULA] light curves with the 5100 Å continuum light curve.

First of all we determined an error of the centroids of the ICCFs by averaging the centroids [FORMULA] that were calculated for fractions of the peak ranging from 35% to 90% of the maximum value of the cross-correlation functions. Then we estimated the influence of two principal sources of cross-correlation uncertainties namely flux uncertainties in individual measurements and uncertainties connected to the sampling of the light curves. We used a method similar to that described by Peterson et al. (1998b). We added random noise to our measured flux values and calculated the cross-correlation lags a large number of times. Due to the large variability amplitudes of Mkn 110 these uncertainties were of lower weight compared to those introduced by the sampling of the light curves. The sampling uncertainties were estimated by considering different subsets of our light curves and repeating the cross-correlation calculations. Typically we excluded 37% of our spectra from the data set (cf. Peterson et al. 1998b). In Table 5 we list our final cross-correlation results together with the total error.


Table 5. Cross-correlation lags

Considering the entire observing period we got a lag of the H[FORMULA] light curve of [FORMULA] days. Peterson et al. (1998a) obtained for a similar extended observing campaign a lag of [FORMULA] days. However, they claim that their best lag estimate derived from an observing period of 123 days yielding the smallest error is about [FORMULA] days.

3.5. Line profiles and their variations

Normalized mean and rms profiles of HeII[FORMULA]4686, HeI[FORMULA]5876, H[FORMULA] and H[FORMULA] lines are shown in Figs. 10 and 11. The rms profile is a measure of the variable part in the line profile. There is a very broad line component in the mean and rms profiles especially to be seen in the HeII line. Even apart from this very broad component the mean and rms profiles of the individual lines are different with respect to their shape and full width at half maximum (FWHM). In Table 6 we list the widths of the mean and rms profiles. The mean and rms H[FORMULA] profiles are more similar to the HeI[FORMULA]4471 profiles than to H[FORMULA]. The rms profile of H[FORMULA] e.g. is significantly narrower than the profile of H[FORMULA]. The profiles of the HeI[FORMULA]4471 line are more noisy than the other ones. They are identical to those of the HeI[FORMULA]5876 line within the errors.


Table 6. Mean and rms line widths (FWHM)

All mean and rms profiles show a red asymmetry. The asymmetry is mainly caused by a second line component at v=1200 km s-1. This second component does not vary with the same amplitude as the main component. Furthermore, this second component was stronger during the first half of our campaign from 1987 until January 1992 than during the second half of the campaign. The H[FORMULA] spectra taken at the intensity minima of February 1989 and August 1994 are plotted in Fig. 12. The additional component centered at v=1200 km s-1 is clearly to be seen. The mean spectra of the first half of our campaign are broader by 400-500 km s-1 (FWHM) than those of the second half because of this component.

There is an independent very broad component present in the mean and rms HeII profiles (Figs. 10 and 11). This very broad component exists in addition to the broad component. There is no transition component visible in the profile. The peak of this very broad profile component is redshifted by 400[FORMULA]100 km s-1 with respect to the narrow lines. This shift was measured in the difference spectra (cf. Fig. 13). This very broad component is the strongest contributor to the HeII variability as can be seen from the rms profile. The very broad component is visible in the Balmer line profiles also, especially at high continuum stages (see Fig. 3). The HeII and H[FORMULA] profiles taken in January 1992 are shown in more detail in Fig. 13. We subtracted the minimum profile taken in October 1988 to remove the narrow line component. The HeII line intensity has been divided by a factor of 1.3 for direct comparison with the very broad H[FORMULA] profile. Immediately one can see the striking similarity. The blue wing of the H[FORMULA] profile is stronger than that of the HeII profile because of the blending with the red wing of the HeII line. The very broad line component has a full width at zero intensity (FWZI) of 12 000 km s-1.

[FIGURE] Fig. 10. Mean profiles of HeII[FORMULA]4686, HeI[FORMULA]5876, H[FORMULA] and H[FORMULA].

[FIGURE] Fig. 11. The rms profiles of HeII[FORMULA]4686, HeI[FORMULA]5876, H[FORMULA] and H[FORMULA].

[FIGURE] Fig. 12. H[FORMULA] spectra taken at the first minimum state of February 1989 (solid line) as well as at the second minimum state of August 1994 (dotted line) and their difference spectrum (dashed line).

Previous Section Next Section Title Page Table of Contents

© European Southern Observatory (ESO) 1999

Online publication: April 12, 1999