A changing accretion morphology exposed

X-ray binaries are most easily studied when they are devouring a lot of gas and therefore produce bright radiation at X-ray, UV, optical, infrared and radio wavelengths. During such outburst episodes, the gas that is being stripped off from the companion forms a disk that swirls around the black hole or neutron star. We think that this accretion disk extends very close to the cannibal star, maybe even touching it.

During quiescent episodes, X-ray binaries are consuming much less gas and are therefore orders of magnitude dimmer at all wavelengths. Current theories of accretion prescribe that during quiescence, the accretion disk cannot extend close to the black hole or the neutron star and must lie (tens of) thousands of kilometers away. In between the edge of this disk and the compact star, the gas flow might be very hot and vertically extended. However, because this gas is very tenuous and not producing strong radiation, it is very hard to test this idea with observations.

Accretion theory thus predicts that as an X-ray binary starts to fade from outburst to quiescence, the geometry of the accretion flow is strongly changing. However, it is highly challenging to measure this because 1) it’s not predictable exactly when X-ray binaries transition to quiescence, 2) once they do, the transition happens very rapidly, typically on a timescale of 1-2 weeks, and is therefore easy to miss, and 3) as X-ray binaries decay into quiescence they become increasingly dim and therefore long/sensitive observations are needed in order to study the properties of the accretion stream. Ready to take up this challenge, we recently designed an ambitious observing campaign aimed to reveal the changing accretion flow in an X-ray binary called 4U 1608-52.

One powerful way to measure the location of the inner edge of an accretion disk is X-ray reflection. This produces prominent emission features at certain X-ray energies that can be studied with sensitive X-ray telescope. NuSTAR is one of the telescopes that is optimally suited to study X-ray reflection. In a previous study, we observed the X-ray binary 4U 1608-52 during one of its accretion outbursts. Our NuSTAR observation revealed a beautiful X-ray reflection spectrum that allowed us to determine that the accretion disk was extending very close to the neutron star. Taking advantage of the fact that 1) this X-ray binary goes into outburst once every few years (opposed to some sources for which we have observed only 1 outburst over 5 decades), and 2) it produces a very strong reflection spectrum, we designed an observing campaign to capture the source again with NuSTAR, but then at a factor ~10 lower luminosity. The main aim was to use the reflection spectrum to determine if with this change in brightness, the accretion geometry changes a lot.

When our target 4U 1608-52 was seen to enter a new accretion outburst in 2018, we closely monitored how its brightness evolved by looking at the data obtained with the Japanese MAXI X-ray telescope that is installed on the International Space Station. Once MAXI showed that the X-ray brightness of 4U 1608-52 was decreasing, we performed observations with the much more sensitive Swift and NICER telescopes. This allowed us to continue watching our target once it became too faint to be detected with MAXI. We analyzed each new Swift observation immediately after it was performed and tried to predict how the brightness of our target would decay onward. With a few days lead time, we then triggered our NuSTAR observation, hoping that it would observe our target at exactly the right time.

After an intense 2 weeks of watching 4U 1608-52 closely every day, we succeeded to have NuSTAR point to our target at exactly the right time. Interestingly, our new NuSTAR observation showed that with a factor ~10 change in brightness, the reflection spectrum completely disappeared. We think that this is because the accretion flow is changing from a (flat) disk into a hot (spherical) structure. This constitutes one of the very few observations that supports standard accretion models, so we are very excited about these results.

van den Eijnden, Degenaar, Ludlam, Parikh, Miller, Wijnands, Gendreau, Arzoumanian, Chakrabarty, Bult, submitted to MNRAS: A strongly changing accretion morphology during the outburst decay of the neutron star X-ray binary 4U 1608-52

4U1608_combi

NICER and NuSTAR spectra of 4U1608-52. The left panel shows the data obtained during several instances along the 2018 outburst. Data obtained during the 2014 outburst is shown for comparison. A prominent feature can be seen between 5 and 10 keV that is referred to as the iron (Fe-K) line and gives information about the accretion geometry. The right panel shows that this emission feature disappears during the decay of the outburst, which shows that there is a dramatic change in the accretion flow.

 

 

 

 

Astrophysical pollution

The outflows, i.e. the jets and disk winds, that are produced by accreting black holes and neutron stars can potentially have a significant impact on the environment of X-ray binaries. The space in between stars and binary star systems is not empty: it’s filled with tenuous gas and dust that is referred to as the interstellar medium (ISM). Jets and disk winds can slam into this ISM, thereby stirring it and heating it. Moreover, extremely powerful thermonuclear X-ray bursts may eject material into the surroundings of the neutron star and create the same effect. These interactions can may have far-reaching consequences, perhaps influencing the formation of stars and thereby influence the evolution of the entire galaxy. However, it is not yet established if the majority of X-ray binaries truly impact their surroundings; this likely depends on the power of the outflows and the density of the ISM.

Whether or not an X-ray binary interacts with its environment may be determined by looking for shocks in the surrounding ISM. Such shocks produce ionized radiation that are characterized by strong emission lines, e.g. one produced by hydrogen gas at a wavelength of 650 nm (H-alpha emission). Some telescopes are equipped with filters that allow you to look at such a specific wavelength; by taking images with a H-alpha filter, shocked regions around X-ray binaries may be revealed. In addition, a camera with a very wide field of view is required, because the shocked regions may be lying quite far away from the X-ray binaries (and hence would be missed when looking with a camera that has a narrow field of view).

Determined to find out if X-ray binaries generally create shocks in their surroundings, we set up a very large campaign to take H-alpha images of many tens of X-ray binaries. For this purpose we are using the Wide Field Camera (WFC) mounted on the Isaac Newton Telescope that is located on La Palma, Spain. To be able to also access X-ray binaries that are located in the Southern hemisphere,  we are also using the Las Cumbres Observatory, which consists of network of telescopes located across the globe, and the Very Large Telescope located in Chile. To pull off this massive observing campaign, my group and I are joining forces with researchers from the University of Sounthampton in the United Kingdom, St Andrews University in Schotland, the Instituto de Astrofísica de Canarias in Spain, and New York University Abu Dhabi in the United Arab Emirates. Stay tuned for the results!

INT_bigger

The Isaac Newton Telescope on La Palma. Photo credit: see here

A mysterious source of heat

Understanding how neutron stars look like on the inside, i.e. what their structure is and what they are made of, is one of the main challenges of modern astrophysics. There are several different ways through which we try to unravel this. One avenue is to measure how the temperature of neutron stars evolve after they have been eating.

Humans gain energy by eating and so do neutron stars. The meal of a neutron star consists of gas that is supplied by a companion star. By consuming this gas, the outer layers of a neutron star get heated up. Once they stop eating, this gained energy is  radiated away and therefore the neutron star cools down again. How hot neutron stars become during their dinners, and how quickly they cool afterwards, depends strongly on their interior properties.

We use sensitive X-ray telescopes to measure how the temperature of neutron stars evolve after they have been consuming gas from their companion. Such studies have revealed that many neutron stars are much more strongly heated during their dinners than is predicted by theoretical models. This is referred to as the shallow heating problem, because the heat appears to be generated just below the surface of the neutron star. It has puzzled us for over a decade now where this shallow heat comes from.

In a previous study, we proposed that the well-known neutron star X-1 in the constellation Aquila, aka Aql X-1, could provide the key to unravel the mysterious source of shallow heat. This neutron star feasts off its companion about once a year and its heating and cooling can thus be studied and compared after multiple dinners.

Using the Chandra and Swift satellites, we studied Aql X-1 after three meals (in 2011, 2013 and 2016) that were very similar. Since the neutron star consumed approximately the same amount of gas during these three episodes, we expected that it would have been heated to similar extend and hence cool in the same way. However, we found that the temperature of the neutron star was strikingly different after its 2016 dinner than after the other two. Within our current understanding of heating and cooling of neutron stars, it is very difficult to understand why its temperature should be so different.

Instead of unraveling the mysterious source of shallow heat, our dedicated study of Aql X-1 seems to have further complicated the picture. We thus have to go back to the drawing board to think of a new experiment to find out what causes the shallow heat inside of neutron stars.

Degenaar, Ootes, Page et al. 2019, MNRAS in press: Crust cooling of the neutron star in Aql X-1: Different depth and magnitude of shallow heating during similar accretion outbursts

Paper link: ADS

aqlx1_cool_3outbursts

The observed temperature of the neutron star in Aql X-1 after three different outbursts. Although the outbursts were very similar and we thus expected the neutron star to heat up and cool down in the same way, the temperature evolution after the 2016 outburst (black stars) was very different from that seen after the 2011 (red circles) and 2013 (blue squares) outbursts.

Don’t overfeed the neutron star

The laws of physics dictate that there is a maximum amount of food that neutron stars and black holes can digest. Once you reach the so-called Eddington limit, the radiation that is produced by the consumption of gas becomes so strong that it blows away the in-falling material. Theoretically, it is therefore predicted that if you overfeed a neutron star or a black hole, that strong outflows are produced: in the regime of super-Eddington accretion, we expect both jets and winds to be created. Jets are usually detected at radio wavelengths, whereas winds mostly reveal themselves as narrow absorption lines in high-resolution X-ray or ultra-violet spectra.

There are a number of neutron stars and black holes identified that are likely accreting at very high rates. Most of these are located in other galaxies, and referred to as ultra-luminous X-ray sources (ULXs), because the high rate of food consumption makes them very bright X-ray emitters. For several of these ULXs, signatures of disk winds have been detected. A number of other ULXs have radio bubbles around them that suggest that these objects are producing strong jets. However, to date there is no ULX known that is known to produce both winds and jets at the same time. It therefore remains to be established if super-Eddington accretion indeed causes both types of outflows.

Swift J0243.6+6124 is an accreting neutron star that is located in our Milky Way galaxy and was discovered in late 2017 when it suddenly started to feed of its companion star. Following its discovery, the object kept brightening until after a few weeks it reached super-Eddington accretion rates. We previously detected a jet from this neutron star using the Very Large Array (VLA) radio telescope. Following this detection, and known that the source was in the super-Eddington regime, we also requested high-resolution X-ray observations with the Chandra telescope with the aim to search for the presence of a disk wind.

Detecting a disk wind in Swift J0243.6+6124 was not an easy task because it was so overwhelmingly bright that it was causing issues for all X-ray satellites: just as the NS cannot eat fast enough, our X-ray detectors couldn’t process the light received from the source fast enough. Luckily, Chandra could be operated in a very special setting that allowed us to look at the source anyway. Excitingly, the spectra that we obtained with Chandra contained a number of narrow absorption lines that can arise from a disk wind. The properties of these absorption lines suggest that the wind is blown away from the neutron star at a dazzling speed of 20% of the speed of light: a speed of about 200 million kilometers per hour! Similar wind speeds have been measured for ULXs in other galaxies.

Our Chandra and VLA observations thus revealed that indeed jets and winds are produced at the same time in the super-Eddington accretion regime, just like theory predicts.

van den Eijnden, Degenaar, Schulz et al. 2019, MNRAS 487, 4355: Chandra reveals a possible ultrafast outflow in the super-Eddington Be/X-ray binary Swift J0243.6+6124 

Paper link: ADS

Spacecraft labled new 2019

Schematic overview of the Chandra X-ray satellite, with which we performed this research. Image credit: NASA

A very radio-bright neutron star

Black holes and neutron stars are notorious for swallowing gas from their surroundings. However, these extreme objects also spit large amounts of matter and energy back into space via collimated streams of gas that are called jets. These jets emit radio emission that can be detected with large radio telescopes such as the Australian Telescope Compact Array (ATCA) that is located in Australia. Black holes seem better at producing jets, since their radio emission is on average a factor ~10 brighter than that emitted by neutron stars.

In 2018 August, the X-ray telescopes orbiting the Earth detected a new X-ray source in the Sky that was named IGR J17591-2342 (after its position in the Sky and after the telescope that discovered it, ESA’s INTEGRAL mission). Within a few days after this discovery, we observed this new X-ray source with ATCA, to investigate if it was producing a jet. We detected such bright radio emission that we suggested that this object was likely a black hole. However, a pulsed X-ray signal was detected from IGR J17591-2342; such a signal requires an object with a solid surface and therefore rules out that this source contains a black hole. The detection of X-ray pulsations instead showed that IGR J17591-2342 contains a neutron star, spinning at a dazzling rate of 527 rotations per second, that is swallowing gas from a nearby companion star.

The distance to the new X-ray source IGR J17591-2342 is unknown, but its X-ray emission is strongly absorbed by interstellar gas, which would suggest that the source is relatively distant. For distances larger than 3 kpc, the radio brightness of IGR J17591-2342 is very similar to that of black holes and much brighter than that of neutron stars. It is not yet understood why this neutron star is able to produce such a bright radio jet.

Russell, Degenaar, Wijnands, van den Eijnden, Gusinskaia, Hessels, Miller-Jones 2018, ApJ Letters 869, L16: The Radio-bright Accreting Millisecond X-Ray Pulsar IGR J17591-2342

Paper link: ADS

Lx_Lr_IGRJ1759_russel2018

Radio and X-ray luminosities of a large collection of black holes (black circles) and different classes of neutron stars (grey circles, pink squares and cyan triangles). The location of IGR J17591-2342 in this diagram depends on its unknown distance and is indicated by the different coloured symbols. Unless the source is very nearby (less than 3 kpc), it is unusually radio bright for a neutron star.

A new class of jet sources

Accretion is a fundamental physical process that plays an important role at all spatial scales encountered in the universe. Whenever accretion occurs, it appears to be inevitable that jets are produced; collimated beams of matter and energy that are spit into space by the astrophysical object that is accreting. For decades, strongly magnetized neutron stars stood out as the only objects that accreted and did not seem to produce jets. This led to the paradigm that their strong magnetic fields prevent the formation of jets. Earlier this year, we made a ground-breaking discovery that disproves this.

Despite decades of jet studies of X-ray binaries, strikingly, no radio emission was ever detected from accreting neutron stars that have strong magnetic fields. For decades, it was therefore assumed that these objects do not produce radio emission because they are incapable of producing jets. Originally set out to provide more stringent upper limits on the radio emission, we exploited the upgraded sensitivity of the Very Large Array (VLA) radio telescope to perform deep radio observations of two strongly magnetic neutron stars, Her X-1 and GX 1+4. Somewhat surprisingly, we detected radio from both objects for this first time. Though very exciting, we were not able to prove that these  detections pointed to the presence of a jet, since other emission processes could produce the observed radio emission. Nevertheless, this motivated us to dig deeper into the question if strongly magnetic neutron stars could produce jets after all.

In late 2017, we were fortunate to run into an ideal test case. A previously unknown X-ray binary suddenly exhibited an accretion outburst, making it shine very bright in X-rays. When it was discovered that the accreting object in this newly discovered source, dubbed Swift J0243.6+6124, was a strongly magnetic neutron star, we requested observations with the VLA to search for radio emission from a jet. And this is exactly what we found.

Our observations of Swift J0243.6+6124 unambiguously proved that we were watching an evolving radio jet. Firstly, we clearly observed a coupling between the radio emission and the X-ray emission, as is seen in black holes and weakly magnetic neutron stars. Secondly, by performing the radio emission in multiple frequency bands, we were able to measure the radio spectral index and evolution therein, which too followed exactly the same behavior as seen for other X-ray binaries. Our observations thus disproved the long-lasting paradigm that strongly magnetic neutron stars cannot produce jets, which has far-reaching consequences.

This discovery opens up a completely new regime to study astrophysical jets. In particular is can shed new light on the open question how these outflows are launched. This is because strongly magnetic neutrons stars have a completely different accretion geometry than black holes and weakly magnetic neutron stars, because their strong magnetic field pushes the accretion disk out to hundreds of kilometers. Any jet launching model must thus be able to explain that material is accelerated into a jet from such large distances. Moreover, several models prescribe that the power of a jet should scale with the rotation rate of the accreting object. This has been very difficult to test with black holes, because their spin rates cannot be unambiguously measured, or with weakly magnetic neutron stars, because these exhibit only a very narrow range in spin rates. Neutron stars with strong magnetic fields, however, are observed with a very wide range of accurately measured rotation rates, from sub-seconds to thousands of seconds. This finally allows to test the predicted correlation between that the radio brightness and the spin rate.

Because of the important scientific impact for jet studies, our results are published in the October issue of Nature (2018). Following up on our discovery, we have already started to perform a large,  systematic radio survey of accreting strongly magnetized neutron stars. The important next steps are to test if, and how, these jets are coupled to the properties of the accretion flow, and if we can detect any dependence of the jet properties on the spin of the neutron star. Stay tuned.

van den Eijnden, Degenaar, Russell, Wijnands, Miller-Jones, Sivakoff, Hernández Santisteban 2018, Nature 562, 233: An evolving jet from a strongly magnetized accreting X-ray pulsar

Paper link: ADS
Selection of press items: NOVA and NRAO
Explanatory movie (English): youtube

lc_swj0432_short

Discovery of a radio jet launched by the strongly magnetic neutron star in Swift J0243.6+6124. Shown is the X-ray light curve from Swift/BAT in black together with our radio observations from VLA in red. After an initial non-detection in the radio, we detected the jet emission during the peak of the outburst and watched it fade in tandem with the decrease in X-rays.

Facts and myths about neutron star jets

Jets are collimated outflows of matter and energy produced by accreting astrophyical objects. Such jets are found on many different scales in the universe, ranging from young stars to supermassive black holes in the centers of galaxies. Black holes and neutron stars that accrete gas from a companion star in an X-ray binary are prominent jet producers too. In these systems, the collimated jets are most prominently detected at radio wavelengths.

Starting in the 1970s, the radio jets of X-ray binaries have been studied in great detail. One key characteristic is that there is a very strong correlation between the radio brightness and the X-ray luminosity, which suggests a strong coupling between the inflow of matter (traced by the X-rays) and the outflow (traced by the radio emission). Early studies suggested, however, that the coupling between the X-rays and the radio emission, parametrized by the coupling index beta, is different for black holes and neutron stars. A plausible explanation for this difference could be that neutron stars have a solid surface; whereas gas that reaches a black hole can be carried across the event horizon without emitting any radiation, all energy contained by the gas will be converted into X-ray radiation when it hits the surface of a neutron star. This could translate into a different X-ray/radio correlation.

Collecting the largest sample of radio/X-ray points of X-ray binaries to date, we set out to perform a rigorous statistical analysis to investigate if the jets of neutrons stars are  fundamentally different from those of black holes. Our analysis contained a total of 35 individual black holes, and 41 neutron stars and let to several important conclusions. Several common conjectures about neutron star jets were disproved by our analysis, while others were strengthened, leading to the following facts and myths:

Facts:

Our rigorous analysis reinforces previous conjectures that the radio emission of neutron stars is fainter, by a factor ~20, than that of black holes accreting at similar X-ray luminosity. Correcting for different factors that might influence the comparison (e.g. their difference in mass, different bolometric correction factors and the extra X-ray emission of neutron stars coming from their surface) does not lift this difference. Therefore, we are left to conclude that, in general, neutron stars produce less bright radio emission than black holes accreting at similar rates.

Myths:

1) For decades, the number of neutron stars observed in the radio band was much more modest than that of black holes, partly driven by the fact (see above) that neutron stars were considerably fainter in the radio band, hence more difficult to observe. However, exploiting the current generation of upgraded radio facilities, much more neutron stars have been observed in the radio band. In fact, our study included 41 different neutron stars, compared to 35 different black holes. Neutron stars are thus no longer underrepresented in radio studies.

2) It is commonly said that neutron stars display a larger scatter in the radio/X-ray plane, i.e. display more chaotic behavior. However, in our study we found that the statistical scatter in the neutron star sample is similar to that in the black hole sample.

3) It is often assumed that neutron stars, in general, show a different (namely steeper) correlation between their radio and X-ray luminosity. However, this conjecture is largely based on a detailed study of one individual neutron star. Considering the sample as a whole, we obtained a coupling index for the neutron stars that was consistent with being the same as that of the black hole sample. It thus appears that neutron stars do not show a different radio/X-ray coupling than black holes.

Apart from comparing the neutron star and black hole samples, we also investigated if  sub-samples among the neutron stars may behave differently. Interestingly, we found that the sub-population of transitional millisecond radio pulsars, statistically behaves differently from the other neutron stars. This suggest that their jet properties are fundamentally different.

Gallo, Degenaar & van den Eijnden 2018, MNRAS Letters 478, L132: Hard state neutron star and black hole X-ray binaries in the radio:X-ray luminosity plane

Paper link: ADS

gallo2018_Lx_Lr

Radio versus X-ray luminosity of about 36 black holes (black filled circles) and 41 neutron stars (red open diamonds). The solid lines and shaded areas represent statistical fits to the correlation between the radio and X-ray luminosity. The resulting coupling index beta is quoted for both populations and is consistent with being the same within the errors.