Dust scattering halo of 4U 1630-47: High resolution X-ray and mm observations constrain source and molecular cloud distances

TL;DR

This work uses dust scattering halos to constrain the distance to the black hole X-ray binary 4U 1630-47 by combining high-resolution APEX CO maps with a 3D reconstruction of molecular clouds and ML-driven cloud segmentation to generate synthetic DSH images. By fitting both radial and azimuthal halo profiles to Chandra data and incorporating the source’s X-ray flux history, the authors constrain cloud ordering along the line of sight and favor a distance of $D=11.5\ \mathrm{kpc}$, while ruling out $D=4.85\ \mathrm{kpc}$ and $D=13.6\ \mathrm{kpc}$ based on morphology, extinction, and energy-dependence constraints. The method highlights systematic uncertainties from cloud-distance ambiguities, extinction, and dust cross-sections, yielding an overall distance error of about $\sim 1$ kpc. The approach demonstrates the potential of combining high-resolution mm data with X-ray halos to test Galactic structure models and to refine distances to embedded X-ray sources, with implications for wind absorption, polarization studies, and jet kinematics.

Abstract

We re-investigated the distance to the black hole X-ray binary 4U 1630-47 by analyzing its dust scattering halo (DSH) using high-resolution X-ray (Chandra) and millimeter (APEX) observations. Dust scattering halos form when X-rays from a compact source are scattered by interstellar dust, creating diffuse ring-like structures that can provide clues about the source's distance. Our previous work suggested two possible distances: 4.9 kpc and 11.5 kpc, but uncertainties remained due to low-resolution CO maps. We developed a new methodology to refine these estimates, starting with a machine learning approach to determine a 3D representation of molecular clouds from the APEX dataset. The 3D maps are combined with X-ray flux measurements to generate synthetic DSH images. By comparing synthetic images with the observed Chandra data through radial and azimuthal profile fitting, we not only measure the source distance but also distinguish whether the molecular clouds are at their near or far distances. The current analysis again supported a distance of 11.5 kpc over alternative estimates. While the method produced a lower reduced chi-squared for both the azimuthal and radial fits for a distance of 13.6 kpc, we ruled it out as it would have produced a bright ring beyond the APEX field of view, which is not seen in the Chandra image. The 4.85 kpc estimate was also excluded due to poor fit quality and cloud distance conflicts. The systematic error of 1 kpc, arising from uncertainties in determining molecular cloud distances, dominates the total error.

Figures (17)

• Figure 1: Figure depicting the formation of dust scattering halos using the 3D shapes of the molecular clouds. The source distance is given by $D$, and the distance to the dust cloud from the observer is given by $xD$. $\theta$ is the observed scattering angle, $\theta_{sc}$ is the physical scattering angle. The distances of 3D clouds in this figure are for better representation and do not coincide with actual distances in this work.
• Figure 2: The rebinned and background-subtracted Chandra image in the 2.25--3.15 keV band. All point sources (including 4U 1630$-$47) have been removed. A Gaussian smoothing with 3 pixel radius was applied. The white box shows the APEX field of view, and green boxes show Chandra ACIS-S chip boundaries with chip numbers shown at the lower left corners.
• Figure 3: MAXI 2--4 keV and constructed unabsorbed 2.25--3.15 keV (E2) light curves. (1) to (3) denote regions in the light curve separated by vertical dashed lines. The solid line is a smoothed absorbed light curve, and the thick dashed-line is the smoothed unabsorbed 2.25--3.15 keV (E2) light curve.
• Figure 4: Top: $^{13}$CO spectrum from APEX. Middle: $^{12}$CO spectrum from APEX. Bottom: $N_{HI}$ spectrum from the Southern Galactic Plane Survey (SGPS) towards 4U 1630$-$47. $T_{B,HI}$, $T_{B,12CO}$ and $T_{B,13CO}$ are the brightness temperatures of the emission lines. Gaussian decomposition is applied to all spectra.
• Figure 5: The foreground shows the velocity spectrum with fitted Gaussians, while the background displays the 3D shapes of molecular clouds based on x, y, and velocity coordinates. The colors of the Gaussians and 3D shapes are matched.