At a meeting at ESAC, Madrid (do we have a hashtag yet?) to discuss ultraluminous X-ray pulsars - some of the most extreme accretors in the Universe!
Tim Roberts (@TYorkshirebloke) opens on the observational features of ULXs/ULXPs. We spent 15 years having ferocious arguments about the size of the black holes in these objects... and then we started seeing pulsations!
Some open questions: what are we calling these objects? (ULPs? PULXs? ULXPs?) What proportion of ULXs are neutron stars (NSs) vs black holes (BHs) and how do we distinguish them? How does super-Eddington accretion work with NSs?
Frank Haberl presents the recent discovery of a fourth ULXP in NGC 300, with a rapidly increasing pulsation period and an enormous pulsed fraction (nearly 80%!) The flux in 2010 was far lower than in 2016 - due to far higher absorption, or something else?
Matt Middleton (on behalf of @MurrayBrightman - we miss you!): ULX-8 in M51 shows an absorption feature, evidence of a cyclotron line. Shows us there's a NS (even without pulsations) and also lets us estimate the magnetic field strength.
If it's an electron line, it implies a field strength of ~10^11 G. However, electron lines tend to be broad, this one is narrow. Might instead be a proton line - implying a very high field strength of 10^14 G!
Gianluca Israel - the luminosity and period of ULXPs is not consistent with a strong dipole magnetic field (otherwise propeller regime sets in and stops accretion). Instead, could be a quadrupolar field on small scales.
New search for pulsating ULX yields its first result - M51 ULX-7 is a pulsar! (My first paper was on this source - I thought it was an intermediate-mass BH but suggested it might instead be a NS. And that's what it turned out to be!)
Matteo Bachetti: obtaining the pulsation period of PULXs is not straightforward - need to correct for binary orbit, spin-up, and also erratic second period derivatives. Pulsations are also not persistent from observation to observation.
There is some good software out there to handle these - presto (used for radio pulsation searches), Matteo's own HENDRICS software, and methods developed for SETI may also be useful!
Fabio Pintore describes a phenomenological pulsar spectrum model, which is a good fit for most of the ULX sample with good data. Doesn't tell us a lot, but shows that previously used black hole models don't necessarily have strong interpretive power.
PULXs tend to have hard spectra - possibly a good place to look for new candidates. He presented results for a new ULX in NGC 5097 and two ULXs in NGC 925. No pulsations found, but hard NGC 925 ULX-1 is one to keep an eye on.
Andrea Belfiore takes a closer look at NGC 5907 ULX-1. It entered a new low state in summer 2017. An extended source (1.1", point source is ruled out at 5 sigma) can be seen around it - an X-ray bubble from shocked interstellar medium? Or dust scattering?
Filippos Koliopanos presents a possible model of accretion onto highly magnetised NSs: Multicolor Accretion Envelope, where at high accretion rates, the NS becomes entirely engulfed by accreting material. Can describe the ULX population spectra well, more detailed study to come.
Marianne Heida picks up after lunch. First off - dear scientists, please try to use existing naming schemes rather than making up your own! And mention source coordinates in your paper if there is any ambiguity whatsoever. PLEASE!
She's covering the optical/infrared counterparts of ULXs/ULXPs. Spectra are really important for distinguishing between companion stars and irradiated accretion discs. They can also give us rotation curves - from which we can place limits on BH mass.
We can reverse this process for NSs (whose mass we know, pretty much), and place limits on the nature of the companion star instead. Other counterparts we've seen: nebulae in optical and near infrared, jets or dust in the mid-infrared.
Christian Motch talks optical counterparts of ULPs - focussing on the bright, blue counterpart of NGC 7793 P13. Optical spectrum shows stellar absorption features, allowing supergiant B star companion to be characterised.
In earlier observations, optical emission is modulated with the orbit, as the star is irradiated by the X-ray source. Later, when the X-ray source is brighter, this modulation goes away. Possibly a thickening accretion disc, higher beaming of X-ray emission?
Jakob van den Eijnden: Galactic NS Swift J0243.6+6124, ~10s period, reaches super-Eddington accretion in outburst. During this outburst, a faint radio jet is detected, which fades with the X-rays. Strong magnetic field doesn't inhibit jet formation => ULXPs may also launch jets?
Natalie Webb: transitional pulsars show X-ray pulsations when accreting, then radio pulsations when not accreting. Can we spot radio pulsations from ULXPs when in a low X-ray flux state? This would help us to track whether accretion is stopped or just diminished in these states.
Ciro Pinto talks about relativistic outflows in ULXs - indicated by absorption features in the high-resolution X-ray spectra, at ~0.2c! Launching such winds uses ~50% of the total energy budget of the ULX (very high!) Wind is thick and clumpy, potentially very complex structure.
XMM-Newton is doing very well after 18 years to let us do this science, but in the future Athena (and potentially Arcus) is going to be amazing for the high-resolution spectroscopy required to characterise outflowing winds.
Peter Kosec: continuing the outflow theme, NGC 300 ULX-1 is nearby and uncrowded, making it a good candidate for outflow searches. It was observed twice: in the second observation, a potential ~0.2c outflow is detected - first evidence for such an outflow for a ULXP.
Dom Walton talks about pulse-resolved spectroscopy of the same source. In the pulsed spectrum, we see a broad absorption feature consistent with being an electron cyclotron line (like that seen in M51 ULX-8, but broader).
This implies a field strength of 10^12 G, consistent with previous estimates of the magnetic field strength from the spin-up. Lower than estimates for the other ULXPs, and also lower luminosity. Are these connected?
Andy Fabian explains the propeller effect: some of the ULXPs disappear from time to time, which could be due to the propeller effect (when the magnetic field is rotating faster than the disc at the magnetospheric radius - matter can't accrete in this state).
A sort of hysteresis - the pulsar is spun up, the propeller regime kicks in, the pulsar spins down again, accretion restarts. Spin-down can be related just to the magnetic field strength: B ~ 10^12 G leads to very slow spin-down - so how come NGC 300 ULX-1 is so slow?
A NS entering a supercritical propeller regime can launch a powerful outflowing wind... like the ones we see in other ULXs, perhaps?
Our final speaker of the day, Ryan Lau, talks about using the mid-infrared to explore the donor star and circumstellar medium of ULX systems. ULXs with mid-IR counterparts: bluer ones consistent with being red supergiants, redder ones with being B[e] supergiant stars.
NGC 300 ULX-1 has a mysterious counterpart - it starts off looking like a B[e]sg, but after the fake supernova event, appears bluer and is now gradually brightening. X-ray brightness appears to be correlated with the mid-IR - how are they connected?
Selma de Mink gives us an overview of stellar evolution, which could tell us where ULXs come from and how they will eventually end up. LIGO has shown us that there are BHs out there with masses 30-45 times the solar mass. So, large-BH ULXs are definitely still a possibility.
Rapidly rotating stars have strange evolution. Mixing enabled by the fast rotation means that *entire star* can be converted from H to He - instead of expanding at the end of their life, they shrink! If ULX companions are spun up, we should consider different evolutionary tracks.
ULXs are systems that stay together after the initial supernova explosion, rather than the companion star being kicked away. However, the explosion can impart enough momentum to bump the system hundreds of parsecs outside of its initial star-forming region.
Stars that form NSs are lower mass than stars that form BHs, right? Wrong! It turns out it's incredibly complicated/chaotic, very dependent on initial conditions - metallicity, stellar mass, perhaps also magnetic fields and stochastic effects. Not at all solved yet.
McKinley Brumback: LMC X-4 could be a nearby ULP analogue - persistent 13.5s pulsations, 30d superorbital period due to warped disc precession. Its pulse strength increases dramatically just before and during super-Eddington flares.
The pulse profiles are 0.25 cycles out of phase between the pre-flare and flare intervals. What causes the phase shift? Change in accretion geometry? Change in beam shape or movement of the hotspot on the NS star could be responsible.
I'm up next! 😬
tl;dl: Observe ULX host galaxies lots of times to find sources entering and leaving the propeller regime. Next!
Elena Ambrosi: modeling a super-Eddington system shows that irradiation is important and can dominate the optical emission. NS ULX binaries with intermediate stars are very faint, but those with massive companions are hard to model since mass transfer is very unstable.
Tassos Fragos: on binary evolution towards NS ULXs. We know that the companion stars must be filling their Roche lobes to reach luminosities in excess of 10^39 erg/s, and that they must be H-rich main sequence stars (He stars are too compact to fill their Roche lobe).
Comparing simulations to the NS ULXs observed so far show that their companion stars are likely 3-8 times the mass of the Sun, with an orbital period around 1-3 days.
Thomas Tauris, with more on origins of ULXPs. High- and low-mass X-ray binaries are common, intermediate-mass XRBs not so much. Wind accretion is very weak, and Roche-lobe overflow is unstable and doesn't last long - however, these could be plausible ULXP systems.
Still, we know that P13's companion is a high-mass B star. HMXBs undergoing Roche lobe overflow are unstable in orbital dynamics, however stability can be achieved with a steep helium gradient within the star, which keeps it more compact and stops it completely engulfing the NS.
As for IMXBs, with a relatively massive NS (2 solar mass), stable systems can exist with companion stars of up to 6 solar mass. The NGC 5907 ULX-1 and NGC 300 ULX-1 systems could potentially be IMXBs. LMXBs can only be ULXPs for large orbital periods (>30 days).
Grzegorz Wiktorowicz: what's the compact object population of ULXs - more NSs or more BHs? Population synthesis simulations show that over time in a star-forming region, BH ULXs form first (within 4-40 Myr), followed by NS ULXs which go on forming much longer (6-4400 Myr).
We would therefore expect the ULXs we see in regions where star formation has ended to be NS ULXs. Therefore, lots of NS ULXs expected - however, not all of them will be pulsars.
In low-metallicity environments, the number of NS ULXs created is unaffected but more BH ULXs are generated, meaning that there are relatively more BH ULXs expected in these environments.
Nanda Rea is here to tell us about magnetars (10^13-15 G magnetic fields). They don't look like we think - rather than a nice neat dipole all the way to the surface, close to the NS the field is multipolar and very complex. Unstable magnetic field leads to lots of outbursts.
Magnetic outbursts in ULXs will be too faint to detect compared to the accretion luminosity. Also, ULXP magnetic fields are not going to be pure dipoles - magnetic field geometry is complicated and strong accretion can change it over time as well.
A 10^14 G dipole will not last long at all. However, low-field magnetars with high-B near the surface and a lower-B (~10^12 G) dipole can last a lot longer.
Kyle Parfrey is talking about magnetospheres and accretion discs - and showing us some very cool relativistic magnetohydrodynamic simulations of accretion. For low magnetic fields, the disc moves in and opens up the field lines, accreting matter is channeled onto the poles.
For high magnetic fields, matter is ejected along the open field lines or kept away entirely (propeller mechanism). Interestingly, in the accreting case, the changing density of the inner flow as well as the accretion column could potentially provide a pulsed emission component.
You can watch some of the simulations here: youtube.com/playlist?list=…
