There are several ways to test prevailing models of AGN, and in particular those related to the properties of the accretion disc, and the physical processes operating therein. These investigations come under two generic headings. Detailed monitoring of a few AGN, mostly associated with reverberation mapping projects eg. Mathur et al. (2017) and Edelson et al (2019). It is clear that for low-redshift AGN the amplitude of variability increases at shorter wavelengths, hence the importance of UV studies, see figure 1(Left) for NGC5548. The second category statistical studies. For example the regular monitoring of specific fields eg. Sloan Digital Sky Survey (SDSS) Stripe 82, see Annis et al. (2014), and the recent cross-matching of the photometry of SDSS quasars with the same objects identified in the Dark Energy Survey (Rumbaugh et al, 2018). These studies address different aspects of the problem. The detailed high time cadence studies reveal time lags between the optical/UV and X-ray energy bands, which pose problems for models of reprocessing by an illuminating source. Whereas the statistical programs look for correlations between the UV variability and other basic properties such at the black hole mass and Eddington ratio, and can identify candidates for the so-called Changing Look AGN. These objects have undergone extreme variations, often detected by their optical spectra (figure 2 Right), but which could easily be found in our proposed study. Indeed, the UV region is far better suited for their discovery than the optical because of the greater amplitude intrinsic variability, and it is much less contaminated by the presence of the constant starlight component from the host galaxy. Note: the Eddington ratio is a fundamental AGN parameter. It is the ratio of the observed bolometric luminosity to the maximum possible luminosity reached when radiation pressure outwards equals gravity acting inwards. Its value is determined by the accretion physics. High values may indicate mass outflows, and its range may define the “low state” and “high state” of an AGN (see figure 3).
Our proposed study fails under the second heading. It requires techniques related to “Big Data” mining. The first stage will provide us with the sample of UV variable AGN. Of course all AGN are variable, but some are more variable than others. The extreme variable cases, using criteria we will define in the course of the study (could be a factor 2 in flux), will be the subject of more detailed investigations such, as spectroscopic follow-up, and also spectra energy distribution (SED) modelling to derive their bolometric luminosities. In figure 1 (Right) we show the SED of a typical AGN. This illustrates the observational gap imposed by neutral atom absorption in our own Galaxy, separating the UV emission and the X-rays. Above this is a cartoon showing the regions responsible for this emission (Done et al. 2012). Models such as this will be employed to interpret cases of extreme UV variability.
Ultra-violet Data-bases, from space and ground-based facilities
The UV photometry for this project will be extracted from archive data-bases of the following space missions; The XMM-Newton OM instrument (an ESA Cornerstone mission), the Neils Gehrels Swift Observatory UVOT instrument (a joint Italian NASA MIDEX mission, with Italian and UK involvement) and GALEX (a NASA small explorer mission). We will also utilise blue-optical (u-band) photometry from the SDSS and the Pan-STARRs survey.
To emphasize the sheer amount of data waiting to be exploited , the latest (2018) XMM OM Serendipitous Ultraviolet Source Survey catalogue contains 5,503,765 sources, of which 1,035,453 have different observations, which we can use for variability studies. Furthermore, the Swift/UVOT Serendipitous Source catalogue has 6,200,016 sources of which 2,027,265 have more than one observation. GALEX has made an all-sky UV survey in two UV filters. The number of sources in the GALEX “clean” field of view, is 69,772,677.
Of course only a small fraction of these UV sources are AGN, and so we will cross correlate them with the SDSS data release 14 quasar catalogue, Paris et al. (2018), which contains 526,356 quasars. The cross-correlation will result in many 1000’s of hits. The first part of our project will be to data mine these huge data-bases, and to assemble a sample of all AGN with multiple UV observations.
Limitations
It is clear that our data will be only sparsely sampled in time, and so it will not be possible to study time lags between the UV and emission at other energies. However we will find example of extreme UV variability. We will pay special attention to “quality control” such cases, since they may be artefacts. To quote Carl Sagan “extraordinary claims require extraordinary evidence.” In such cases human intervention will be needed. For example, to examine the 2-D images from which the photometry is derived. There will be field stars near to the AGN, and the reality of extreme variability can be tested by simple inspection with the constant flux field stars.
A Pilot Study
The team leader (Ward) has commenced a pilot study of this proposed project, carried out with a 4th year undergraduate student. This involved only a small sub-sample of low redshift AGN having multiple SDSS optical spectra, and also multiple epoch UV photometry. We have identified several interesting objects, one of these is shown in figure 2 (Left). This AGN appears to have varied in the UV by more than a factor of 2. The timescales are important, and these will be statistically analysed in our project. Clearly this limited pilot study has only sampled the “tip of the iceberg” of all the UV data available in these catalogs, and so the time is ripe for a coordinated major investigation such as we propose here.
Phase 1: The kick-off meeting in Bern will aim to scope the project. Identify the most efficient ways to mine these huge data-bases, and decide who is responsible for different parts of the work. Note: some of the basic “data-mining” will be done by graduate students, with oversight by members of our team.
Phase 2: We will now have our sample of UV variable AGN. We must be concerned about quality control (are the variations real, and in particular for the extreme cases?). See the section on Limitations. We will assemble all available multi-wavelength data for a sub-set of the most interesting cases, and undertake SED modelling of their high and low states. This will used to plot their positions on the UV/X-ray, versus Eddington ratio plot, shown in figure 3. In parallel with this, we will begin a program of follow-up spectroscopy of the most promising variable objects.
Phase 3: We would expect to present our final results at international conferences, and to publish in the top astronomy journals eg. The Astrophysical Journal, MNRAS and Astronomy and Astrophysics. If we discover an exceptionally interesting case, we will consider submitting it to the journal Nature or Science. We will continue to monitor our best candidates spectroscopically, using medium aperture telescopes to which we have access.
Outputs
This will be the first large scale statistical investigation of AGN UV variability using several very large and mostly under-exploited satellite data-bases eg. XMM-Newton (OM) , Swift (UVOT) and GALEX. In addition we will enhance the multi-wavelength element by using other ground-based facility archives eg. SDSS, Pan-STARRS and the Catalina Survey.
Our investigation will be complementary to previous studies of the optical variability of quasars, for which the spectra and photometry cover their accretion disc emission by virtue of the wavelength redshift. Our sample will focus on medium and low redshift AGN (z < 1), where we need the space based UV data to cover the wavelengths of the dominant disc emission. All targets for which there is UV data from XMM and Swift, will also have near-simultaneous X-ray data, and so we can measure the UV/X-ray over a range of luminosities for the same object, for 1000’s of AGN.
For the first time we will determine the UV variability (near to the peak of the accretion disc emission) ranging from weeks to decades, for a large sample of medium and low redshift AGN. This, when combined with Eddington ratios, will be used to test new ideas to explain the cross-over region between Low/Hard and High/Soft states of AGN, see figure 3 from Ruan et al (2019). This will help us to understand what factors influence the emission properties of the accretion disc.
Finally, it is statistically highly probable that we will discover some AGN with extreme UV variability, which will be strong candidates for Changing Look objects, ideal for follow up optical spectroscopy. Our international team has access to medium aperture telescopes which are suitable for this task.
Legacy: Our project will provide a major resource for the extragalactic community. In addition to publishing our results we will also produce a catalogue of all AGN with multiple UV observations, giving the fluxes and dates of the observations.
References
Annis, J et al. ApJ., 794,120 (2014)
Done, C et al. MNRAS., 420, 1848 (2012)
Edelson, R et al. ApJ., 870,123 (2019)
Lawrence, A Nature, Astronomy., 2, 102 (2018)
Kubota A., Done C., 2018, MNRAS, 480, 1247
Mathur, S et al. ApJ., 846, 55 (2017)
Paris, I et al. A & A., 613, 51 (2018)
Ruan, J et al. preprint arXiv:1903.02553v1 (March 2019)
Rumbaugh, N et al ApJ., 854,160 (2018)
Shappee et al. ApJ. 788,48 (2014)