A distant quasar’s changing soft X-rays have given astronomers a way to estimate where part of its ultrafast wind is located. In a preprint based on a March 2024 observation of PDS 456, the researchers argue that a low-ionization outflow crossed the compact region producing the quasar’s X-rays. Their calculations place that absorber at more than roughly 4,000 gravitational radii from the central black hole, a lower limit rather than an exact address.
The result comes from a striking split in the source’s behavior. The flare and quieter parts of the observation differed mainly around 1 keV in the quasar’s rest frame, while the hard-X-ray spectral shape stayed nearly unchanged. The authors interpret the localized soft-band change as variable partial covering by an outflow.
A moving patch in the line of sight
The team analyzed one object, PDS 456, using simultaneous observations from XRISM’s Xtend instrument and NuSTAR. They divided the XRISM data into seven time slices, choosing longer intervals when the count rate was lower so that each slice would contain enough photons for spectral fitting. The main time-sliced analysis linked the low-ionization outflow’s ionization parameter and outflow speed across all seven slices, while a separate test allowed both quantities to vary independently.
In plain terms, ionization describes how strongly the gas has been stripped of electrons by radiation. Broadband fitting characterized the changing soft-X-ray structure as a partially covering low-ionization ultra-fast outflow, or UFO, with an ionization parameter of about log ξ = 3.1 and a speed near 0.30 times the speed of light. Partial covering means the gas blocks only part of the X-ray source, so changes in the blocked fraction can alter the observed spectrum.
The covering fraction generally rose during the observation. A test of the idea that it stayed constant returned p = 8.5 × 10−8, while a straight-line fit gave a rise of (7.9 ± 1.7) × 10−5 per kilosecond across all slices. The increase was steeper, (1.3 ± 0.2) × 10−4 per kilosecond, when the final slice was left out, because that last slice did not follow the mainly upward trend.
The authors interpret that pattern as the absorber moving into the line of sight across the X-ray corona. That interpretation is central to the distance estimate, but it is not the only possible source of changing coverage: internal spreading, clumping or holes in the gas could also affect how much of the source is obscured.
Distance changes the picture of the wind
Using the observed rate of change to limit the absorber’s transverse, or side-to-side, speed, then combining that limit with angular-momentum conservation and an escape-radius condition, the researchers obtained a crossing speed below roughly 0.005 times the speed of light. The resulting absorber distance is greater than roughly 4,000 gravitational radii, where a gravitational radius is a length scale set by the black hole’s mass.
A separate estimate based on the size of the emitting region and the gas’s ionization gave a similar scale: more than roughly 5 × 10^17 centimetres, or about 7,000 gravitational radii. That check was treated as an order-of-magnitude comparison and assumes the wind is not significantly shielded from the ionizing radiation.
The inferred low-ionization phase therefore sits substantially farther out than the high-ionization phase, whose cited location was about 200 to 600 gravitational radii. Yet the two phases have comparable outflow speeds. The authors say that combination disfavors a self-similar magnetocentrifugal wind, a model in which wind speed follows a fixed relation with launch radius, and is compatible with either CAK-like radiation-pressure acceleration or a compact magnetic-reconnection event followed by nearly ballistic travel. More general magnetic-wind explanations remain possible.
The absorber is not the only changing component
The analysis also found a possible change in low-ionization emission. If that change is real, the fitted emission normalization rose from below 0.3 × 10−4 in the second slice to (2.1 ± 0.5) × 10−4 in the third, a nominal increase of more than about sevenfold within less than about 90 kiloseconds. The authors therefore favored an additional compact emitting component near the corona instead of assigning all of the emission to the more distant low-ionization outflow.
That possible emission variability did not erase the covering-fraction trend in a further test. Holding the low-ionization emission normalization fixed worsened the fit by ΔC = 37.1 for six additional degrees of freedom, corresponding to a formal significance of about 4.8 sigma. The covering fraction still rose, with a slope of (6.8 ± 1.1) × 10−5 per kilosecond. The significance is formal and depends on the spectral model and on the assumptions behind the statistical test.
A useful constraint, not a final map
The findings come from one quasar and one observing campaign, so they do not show that the same geometry applies to other systems. Nor do they prove that transverse motion alone caused the changing coverage, determine an exact distance or uniquely identify how the wind was launched. The low-ionization emission result also depends on treating that variability as genuine, and a single-power-law model for the intrinsic continuum leaves open the possibility that changes in a soft excess could affect the inferred component.
The next test will require longer monitoring and sharper tracking of individual wind structures. The authors point to future observations that could measure covering-fraction changes in high-ionization clumps, refine the distance ordering of the two phases and help distinguish continued radiation-pressure acceleration from compact magnetic acceleration. The manuscript is an arXiv version 1 preprint dated 28 August 2026, not a journal publication.
Paper data and sources
Original title: The Geometry of Ultra-Fast Outflows Probed by Soft X-ray Variability in PDS 456
Authors: Riki Sato, Kouichi Hagino, Toshiya Iwata et al.
Journal/Repository: arXiv
Status: Preprint, not yet peer-reviewed
First online: 2026-08-28
DOI: Not available
Original paper · Full text