(News & views on the paper by Xu Han et al., RAA, 2012, vol.12, 1597-1602)
Understanding the timing properties of ULXs
(Joint Institute for VLBI in Europe, Dwingeloo, Netherlands)
Ultraluminous X-ray sources (ULXs) in nearby galaxies are an astrophysical puzzle. Their off-nuclear location excludes an origin from an active galactic nucleus, although in some cases ULX candidates were identified with background quasars. There are three competing interpretations to explain X-ray luminosities exceeding 3x1039 erg/s, well above the Eddington limit for a few solar-mass black hole. The high luminosity is naturally explained for accreting compact objects of a few thousand solar masses, so-called intermediate-mass black holes (IMBHs, Colbert & Mushotzky 1999). Alternatively, these are stellar systems like Galactic black hole X-ray binaries (BHXRBs) with either a mild (geometric) beaming of the X-ray emission (King et al. 2001), and/or a super-Eddington accretion ratio (Begelman 2002). To better understand the nature of ULXs and constrain their mass, there are two promising approaches. One may constrain the mass using the fundamental plane of black hole activity relations (FP; Merloni, Heinz & Di Matteo, 2003) with milliarcsecond angular resolution in very long baseline interferometry (VLBI) observations (Paragi, Garrett & Biggs 2006), or one may compare the X-ray spectral and timing properties of ULXs to BHXRBs (see e.g. Belloni 2011). Both methods have various issues, mostly related to the fact that we do not know how ULXs accrete or at what rate.
Galactic BHXRBs have X-ray spectral and timing properties, according to their accretion states (Fender, Belloni & Gallo 2004). In the high-intermediate state they show quasi-periodic oscillations (QPOs) and additional broad-band noise, known as type C QPO (Fender, Homan & Belloni 2009). Detection of similar QPOs in ULXs is very rare. The first one was discovered in M82 X-1 by Strohmayer & Mushotzky (2003). They argue that the QPO must originate in a thin disk, and therefore beaming of the X-rays is almost certainly ruled out: it is hard to reconcile the data with a black hole mass less than 300 Msun. In the radio regime, VLBI observations gave an upper limit of about 500 Msun (assuming that M82 X-1 obeys the FP-relations) - there is still no strong evidence for the presence of an IMBH (Paragi, Garrett & Biggs 2006). Mucciarelli et al. (2006) found QPOs at various frequencies in XMM and archival RXTE data, and showed that their properties are strongly reminiscent of the type C QPOs seen in Galactic BHXRBs, except for their 1:2:3 frequency ratios. Assuming the frequency scales inversely with the BH mass, these detections would allow a black hole of only 1 Msun (Belloni 2011).
Strohmayer & Mushotzky (2009) reported a QPO at a frequency of 20 mHz in NGC 5408, and claimed that the corresponding black hole mass is within 1000-9000 Msun, using the QPO frequency-spectral index relations. They also detected a secondary feature at 15 mHz. The noise characteristics of the QPO are again similar to the type C QPOs in BHXRBs, except for the secondary peak (Belloni 2011). Whether the QPO frequency scaling relations can give a good handle on measuring black hole masses in these systems is still a question of debate though. The X-ray spectral states seen in ULXs are somewhat different from the spectral states in BHXRBs (e.g. Soria, 2011). In the case of NGC 5408, Middleton et al. (2011) argues that the system may be in a super-Eddington state, naturally limiting the black hole mass to <= 100 Msun. Strohmayer (2009) found an additional clue. He reported a 115.5 day (10^-7 Hz) periodicity in the X-ray data, although this result was not fully consistent with the analysis of the same data (but having a shorter temporal baseline) by Kaaret & Feng (2009). Han et al. used state of the art data analysis techniques to show that the 115.5 day periodicity is real. If this is due to orbital motion, in the binary system (not to be confused with QPOs, that are related to accretion disk instabilities), then the system likely consists of a giant or supergiant star orbiting an IMBH (Strohmayer 2009). The 115.5 day periodicity however can be interpreted with superorbital variations as well (Foster et al. 2010).
Similarly, arguments based on the FP-relation between the X-ray luminosity, radio luminosity and BH mass should be treated with caution. It does however require a cosmic conspiracy if both stellar-mass and supermassive black holes can produce powerful compact jets, detectable on milliarcsecond scales, but IMBHs cannot. In the case of NGC 5408, the radio emission is resolved, and has a steep spectrum, consistent with a black hole powered nebula (Lang et al. 2007); at 5 GHz there is a strong upper limit for a flat spectrum compact component of 150 microJy. In fact, it is likely to be much fainter because the observed spectrum is quite steep. For an X-ray luminosity of ~10^40 erg/s and at a distance of 4.8 Mpc (Strohmayer & Mushotzky 2009), the predicted range of 1000-9000 Msun would correspond to 5 GHz radio flux densities between about 30 and 200 microJy. An accreting system in a radio-loud state would have orders of magnitude higher radio flux densities. The current radio data therefore do not seem to support the IMBH scenario (at least at the high end of the proposed mass range) for NGC 5408, unless the radio jet is quenched like in the high-soft state of BHXRB. While none of the above methods provide the final answer yet of whether an IMBH exists in NGC 4508 or not, further studies with the methods presented by Han et al. would be highly valuable, in order to better understand the periodicities seen in ULXs.