A full spectral-timing model to map the accretion flow in black hole binaries: the low/hard state of MAXI J1820+070

Abstract

The nature and geometry of the accretion flow in the low/hard state of black hole binaries is currently controversial. While most properties are generally explained in the truncated disc/hot inner flow model, the detection of a broad residual around the iron line argues for strong relativistic effects from an untruncated disc. Since spectral fitting alone is somewhat degenerate, we combine it with the additional information in the fast X-ray variability and perform a full spectral-timing analysis for NICER and NuSTAR data on a bright low/hard state of MAXI J1820+070. We model the variability with propagating mass accretion rate fluctuations by combining two separate current insights: that the hot flow is spectrally inhomogeneous, and that there is a discontinuous jump in viscous time-scale between the hot flow and variable disc. Our model naturally gives the double-humped shape of the power spectra, and the increasing high-frequency variability with energy in the second hump. Including reflection and reprocessing from a disc truncated at a few tens of gravitational radii quantitatively reproduces the switch in the lag-frequency spectra, from hard lagging soft at low frequencies (propagation through the variable flow) to the soft lagging hard at the high frequencies (reverberation from the hard X-ray continuum illuminating the disc). The viscous time-scale of the hot flow is derived from the model, and we show how this can be used to observationally test ideas about the origin of the jet.

The accretion flow geometry of MAXI J1820+070 through broadband noise research with Insight-HXMT

“…It is worth noting that the characteristic frequency of 𝐿 4 above 90 keV is almost five times the frequency in 1-10 keV (from ∼ 2 Hz to ∼ 10 Hz). Similar results in LE band can be found in Kawamura et al (2022). They used NICER data to study the relationship between asymmetric Lorentzian function P1, P2 and energy which actually reflects the evolution of characteristic frequency with energy for 𝐿 2 , 𝐿 3 and 𝐿 4 components (P1 corresponds 𝐿 2 ; P2 corresponds 𝐿 3 and 𝐿 4 ) .…”

“…As Figure 13 shows, the characteristic frequency of 𝐿 2 component shows energy-dependence: the emitting region spans from ∼ 34𝑅 𝑔 to ∼ 27𝑅 𝑔 corresponding to 1-150 keV photon energy. Like Dziełak et al (2021) and Kawamura et al (2022) indicated, with frequency-resolved spectral analysis, the 𝐿 2 component is supposed to come from variable disc emission. However, making use of the ME and HE data from Insight-HXMT, we actually detect the emission from high energy emission > 100 keV from the 𝐿 2 component which cannot be attributed to a simple standard accretion disc.…”

“…They used NICER data to study the relationship between asymmetric Lorentzian function P1, P2 and energy which actually reflects the evolution of characteristic frequency with energy for 𝐿 2 , 𝐿 3 and 𝐿 4 components (P1 corresponds 𝐿 2 ; P2 corresponds 𝐿 3 and 𝐿 4 ) . In order to fit the PDS phenomenally, Kawamura et al (2022) used two asymmetric Lorentzian functions. Nevertheless, considering the Lorentzian function that fluctuation propagation predicts, we decide to use three standard Lorentzian functions to fit PDS.…”

“…The ObsID we choose and Kawamura et al (2022) choose are no more than one day apart, so we can easily match P1, P2 with 𝐿 2 , 𝐿 3 and 𝐿 4 . P1 corresponds to 𝐿 2 , P2 corresponds to 𝐿 3 and 𝐿 4 .…”

“…At least two Comptonization regions with different temperatures and optical depths are required to fit both the variability spectra and the time-averaged spectra. Kawamura et al (2022) proposed a model based on the fluctuations propagation by considering that the hot inner flow is spectrally inhomogeneous, and the viscous time-scale is discontinuous between the disk and the hot flow. This model reconstructs the shape of the broadband noise below 10 keV in MAXI J1820+070.…”