Optical/UV studies

Since these objects first came to light, astronomers have been searching for more informa- tion on their nature from multiple wavelengths. This task was difficult before the advent of Chandra, due to the large number of sources available within the positional error circle of ROSAT. However, in the decade since its launch, much progress has been made. Start- ing from the largest size scale and moving inwards, we find that ULXs have been found preferentially in actively star-forming galaxies (Kaaret 2005), along with starburst, inter- acting/merging and dwarf galaxies (Swartz et al. 2008) preferably with low metallicities (Mapelli, Colpi & Zampieri 2009). In the case of dwarf galaxies there does appear to be a preference for low metalicities (Pakull & Mirioni 2002).

An observed correlation between the number of ULXs and the global star-formation rate in spirals has also been noted (Swartz et al. 2004; Liu, Bregman, & Irwin 2006). The sheer number of sources observed in starburst galaxies together with their short lifetimes (implied by their location in regions of active star formation) mean that, if these systems were all to contain IMBHs, an unrealistically large underlying population must be present (see Figure 1.22; King 2004). Instead it is much more likely that the bulk of the population are the most extreme examples of High Mass X-ray Binaries (HMXBs) that somehow

Figure 1.22: Cartwheel galaxy: this image combines data from four different observatories: the Chandra X-ray Observatory (purple); the Galaxy Evolution Explorer satellite (ultra- violet/blue); the Hubble Space Telescope (visible/green); the Spitzer Space Telescope (in- frared/red). The insert highlights the large X-ray population contained within this galaxy (Image composite: NASA/JPL/Caltech/P.Appleton et al. X-ray: NASA/CXC/A.Wolter & G.Trinchieri et al.).

exceed the Eddington limit. These high mass companions provide a natural origin for the high mass transfer rates required to power the observed luminosities (Rappaport, Podsiadlowski & Pfahl 2005).

An association with star-forming regions (Fabbiano, Zezas & Murray 2001; Lira et al. 2002; Gao et al 2003) and the unbroken luminosity function connecting ULXs to the standard X-ray binary population (Grimm, Gilfanov & Sunyaev 2003) also indicates that we may be observing HMXB systems. A recent survey by Swartz, Tennant, & Soria (2009)

sought to investigate the nature of the regions surrounding ULXs (100 by 100 pc2_regions).

and ULXs. This result links in to the array of recent studies that have sought to identify potential counterparts of these sources and look at their immediate environments (see below).

Stellar counterparts

The search for optical counterparts has mainly focused on nearby systems (∼< 10 Mpc), with

individual counterparts observed over the range mV∼ 22-26 (Roberts et al. 2008). Many

of these are observed to be blue, suggesting an OB star and so the presence of HMXB, or that we are observing reprocessed emission from the accretion disc (e.g. Copperwheat et al 2005). Much of this work is discussed further in Chapter 2.

Of course one of the main reasons to identify these counterparts is to perform spec- troscopic studies of these systems. This will provide a means of answering the question of their nature conclusively by a direct mass measurement based on constraints placed by the binary orbit. To date this has not been achieved, but similar methods have been applied to numerous Galactic stellar mass black hole systems (van Paradijs & McClintock 1995; Charles & Coe 2006), and to an extra-galactic stellar mass black hole in IC 10 X-1, a Wolf-Rayet black hole binary. A radial velocity curve of this source was constructed from

repeated optical observations, which provided a mass estimate of 23 – 34 M⊙(Prestwich et

al. 2007; Silverman & Filippenko 2008). Although the line features used here are emitted from the star, it proves that the concept of radial velocity curves for extra-galactic sources, using a method that was instrumental in proving the first observational evidence of the existence of black holes (see Section 1.3.1), can work well for extra-galactic sources.

Only a small number of pilot optical spectra of ULX counterparts have been taken and published to date. Roberts et al. (2001) studied the counterpart of NGC 5204 X-1, finding a blue, featureless spectrum. Similar results were also found in the spectrum of NGC 1313 X-2, possibly the most explored counterpart to date (Zampieri et al. 2004;

Pakull, Gris´e & Motch 2006; Liu et al. 2007; Gris´e et al. 2009). A He ii 4686 ˚A line was

identified in its spectum, and an observed velocity shift in the line was reported in VLT

observations of this source. The shift was of ∼ 300 km s−1 _{over approximately a 3 week}

ULX nebulae

Pakull & Mirioni presented a study of a small number of ULX environments in 2002. They showed that each of the ULXs in the sample resided in emission nebulae that were up to a few hundred parsecs in diameter. They also reported that may of the nebulae showed evidence of both low and high ionisation emission lines within their spectra. This indicated that the nebula could be either shock ionised or photoionised. Such photoionisation could easily be caused by emission from the ULX. It can also be used to place limits on the X-ray emission powering the photoionisation of the nebula (if we assume this is the source of the energy). By comparing this to the observed X-ray luminosity of the source, Pakull & Mirioni (2002) were able to conclude that the optical emission was consistent with emission from an isotropic X-ray emitter. This appears to rule out beamed emission, suggesting that we are observing either IMBHs or some form of super-Eddington accretion state. Later results also showed evidence of the presence of cavities in the nebulae of these systems, possibly cleared by shocks (Roberts et al. 2002a). Further explorations of the spectra of ULX nebulae also showed evidence of both photoionisation and shocks (e.g. Pakull & Mirioni 2003), and in some cases the ULX must power strong winds and/or jets in order to power the nebulae (Abolmasov et al. 2007). From these results, it would suggest that we may be looking at a heterogeneous sample of sources.