5.5. Multiwavelength Continuum Variability

As noted earlier, one of the important conclusions reached from the original AGN Watch monitoring program was that the UV and optical continua vary with no apparent time delay between them, at least to the accuracy of the experiment ( 1 day in the best cases). An important consequence of this is that AGN continuum variability is not due either to mechanical accretion-disk instabilities or to variations in the accretion rate, since such effects would be expected to propagate through the disk (and thus across the spectrum) on much-longer sound-crossing (Eq. (11)) or drift time scales (Eq. (12)), respectively. The suggestion was made ^{17, 15} that the UV/optical variations might in fact be driven by X-ray variations, possibly in a manner consistent with the X-ray reprocessing models that were being developed to account for the 10 keV reflection hump and the equivalent width of the 6.4 keV Fe K line ^{32, 46, 29}; both of these features suggest that something like half of the emitted hard X-rays in AGN interact with "cold" (not highly ionized) matter that covers much of the sky as seen from the source (and which might in fact be the accretion disk itself). If it is supposed that hard X-rays are produced above the disk plane near the axis of the accretion disk, then hard X-ray radiation striking the disk should be reprocessed into lower-energy continuum photons. Moreover, the UV/optical variations should follow those in the X-rays, with the shortest time delays for the higher-energy photons that are produced predominantly in the central regions of the accretion disk. The temperature structure of a classical thin accretion disk is given by

(41)

and assuming that AGNs are accreting close to the Eddington rate with about 10% efficiency, the radial temperature structure becomes

(42)

from which we expect a peak in the spectral energy distribution in the UV/soft X-ray region, which as noted earlier might in fact be the origin of the big blue bump ^{81, 48}.

If we suppose that a thin accretion disk is irradiated by an X-ray source on the disk axis, we should see the inner, hotter part of the accretion disk respond before the outer, cooler parts. From Wien's Law, the radiation at wavelength arises primarily at a particular temperature corresponding to a particular location in the disk, i.e., T^-1 r^3/4, so the difference in the response times at different wavelengths should be naively

(43)

It is straightforward to show that this relationship holds for an irradiated disk that is heated locally by hard X-rays.

The predicted wavelength-dependent continuum time delays have been detected reliably in the case of NGC 7469, which was monitored intensively in 1996 ^{91, 16, 42} Fig. 40 shows the result of measuring the UV and optical fluxes across the entire monitored UV and optical spectrum, and cross-correlating each of the resulting light curves with the shortest-wavelength UV continuum. Relative to the short-wavelength UV, continuum lags in other line-free continuum bands are detected at no less than 97% confidence throughout the UV/optical region.

Figure 40. Lag as a function of wavelength for NGC 7469. The average flux in each wavelength band is cross-correlated with the shortest-wavelength UV (1315 Å) continuum, yielding a "lag spectrum". The expected dependence for an irradiated thin accretion disk, 4/3, is also shown. The broad emission lines are prominent because they have longer response times than the adjacent continuum.

While the evidence shows that there are in fact wavelength-dependent continuum lags in NGC 7469, this is not an unambiguous detection of a classical thin accretion-disk structure. There are several complications:

1. Similar wavelength-dependent lags have not been reliably detected in other sources. This effect may be marginally present in the case of NGC 4151 ⁷⁰, but it is not found in NGC 3516 ²². However, the sampling rates obtained for NGC 7469 are far superior to those of most other experiments, and the upper limits on UV/optical continuum lags for other AGNs seem to be consistent with predictions based on (highly uncertain) scaling of the other sources to the NGC 7469 result, except possibly in the case of NGC 3516.

2. Korista & Goad ⁴¹ have suggested wavelength-dependent lags are consistent with the expectations for diffuse emission from the emission-line clouds (Balmer and Paschen continuum). Observations over a wider wavelength range are needed to distinguish between an accretion-disk and a diffuse-continuum origin (specifically, the NGC 7469 observations did not cover the crucial wavelength range around the Balmer edge ~ 3650 Å).

3. The hard X-ray flux variations in NGC 7469 are not correlated with those in the UV/optical ⁵⁷. The hard X-ray spectrum was monitored with RXTE at the same time as the UV/optical monitoring campaign. Comparison of the light curves shows that the minima occurred at about the same time, but the hard X-ray maxima occurred approximately 4 days later than the corresponding maxima in the UV continuum light curve. This is a prima facie argument against the reprocessing scenario. However, a more complete spectral analysis ⁵⁶ enabled by an improved background model for RXTE finds that the UV flux variations are correlated with variations in the X-ray spectral index, in the sense that the X-ray spectrum is softer when the UV is brighter. This suggests that the soft X-rays are strongly correlated with the UV; an increase in the number of UV seed photons cools the Comptonizing corona above the accretion disk, softening the spectrum.

By way of contrast, a recent study of high-energy variability in NGC 5548 shows clear relationships between the hard X-rays and the lower-energy photons ¹²: the extreme ultraviolet (EUV) variations lead those in the soft X-ray region by about 3.5 hrs and those in the hard X-ray by about 10 hrs. The temporal order and time scales are consistent with production of both the soft and hard X-rays by Compton upscattering of lower-energy photons.

If the X-rays are indeed Comptonized lower-energy photons, are the seed photons in the EUV ( ~ 100 Å) or UV (~ 1000 Å)? The only good comparison available is based on simultaneous UV and EUV monitoring of NGC 5548 in 1993 ⁵⁰. Cross-correlation of the UV and EUV light curves shows that the UV leads, with _cent = 0.1^+0.7_-0.2 days, which is consistent with zero lag and leaves the question unanswered.

A final recent observation that should be mentioned is based on a three-year combined X-ray (RXTE) and optical monitoring program on the narrow-line Seyfert 1 galaxy NGC 4051 ⁷⁵. During the first two years of this program, NGC 4051 behaved in a "normal" fashion for narrow-line Seyfert 1s, with rapid, violent X-ray variability and much lower-amplitude optical variations. During the third year, NGC 4051 went into a very low X-ray state. In Fig. 41, the top panel shows the X-ray light curve. The bottom two panels show rms optical spectra, which isolate the variable parts of the spectrum, obtained during the two periods indicated, one during a high X-ray state and one during the low X-ray state. They are remarkably different. The high-state rms spectrum (lower left) shows that the optical continuum and the H and He II 4686 emission lines were all varying strongly. However, the low-state rms spectrum (lower right) shows that the optical continuum and the H emission line are still present and variable, but the He II line has vanished from the rms spectrum, meaning that it is absent or constant. Whether or not it has completely vanished is difficult to determine on account of the blending of He II with strong Fe II emission in the mean spectrum, but an attempt to remove the Fe II emission by subtraction of a template indicates that most of the He II emission must have vanished during 1998. This indicates that not only has the X-ray flux dramatically decreased, but the EUV flux (which drives the He II variations) must have also decreased by a significant amount. This may suggest that the entire inner accretion disk has undergone a transition to a low-radiation state, such as an "advection-dominated accretion flow" ⁵⁷.

Figure 41. Comparison of hard X-ray and optical spectral variations in NGC 4051 in different X-ray states. Panel (a) shows the 2-10 keV flux measured with RXTE as a function time for the three-year period 1996-1998. Panel (b) shows the rms optical spectrum during an X-ray active period in 1996, Panel (c) shows the rms optical spectrum during an X-ray quiescent period in 1998. Note the absence of strong He II 4686 emission in panel (c). From Peterson et al. 75 © 2000 AAS.