The X-ray Properties of Million Solar Mass Black Holes

TL;DR

This study probes whether X-ray properties of low-mass black holes with $M_{BH}\sim10^6\,M_\odot$ depend on Eddington ratio by combining seven new Chandra observations ( spanning $-2.0<log(L_{bol}/L_{Edd})<-1.5$ ) with a literature sample of 73 mBHs. Using Chandra imaging/spectroscopy and stacking analyses, it measures X-ray luminosities, X-ray-to-UV ratios (aox), and spectral slopes, finding no systematic X-ray differences between low- and high-Eddington subsamples across a wide dynamic range in $L_{bol}/L_{Edd}$. A persistent X-ray weak tail is observed, which could reflect either a population of intrinsically X-ray weak AGN or the effects of a geometrically thick, radiatively inefficient slim disk, possibly connected to accretion physics at low Eddington ratios. The results have implications for SMBH growth in the early universe and for AGN-driven reionization, while highlighting substantial uncertainties in bolometric corrections for low-mass AGN and the need for broader multiwavelength, high-resolution data to calibrate these corrections.

Abstract

We present new Chandra X-ray observations of seven low-mass black holes (~1e6 Msun) accreting at low Eddington ratios between -2.0<log L/Ledd<-1.5. We compare the X-ray properties of these seven low-mass active galactic nuclei (AGN) to a total of 73 other low-mass AGN in the literature with published Chandra observations (with Eddington ratios extending from -2.0<log L/Ledd<-0.1). We do not find any statistical differences between low- and high-Eddington ratio low-mass AGN in the distributions of their X-ray to ultraviolet luminosity ratios (aox), or in their X-ray spectral shapes. Furthermore, the aox distribution of low-L/Ledd AGN displays an X-ray weak tail that is also observed within high-L/Ledd objects. Our results indicate that between -2<log L/Ledd<-0.1, there is no systematic change in the structure of the accretion flow for active galaxies hosting 1e6 Msun black holes. We examine the accuracy of current bolometric luminosity estimates for our low-L/Ledd objects with new Chandra observations, and it is plausible that their Eddington ratios could be underestimated by up to an order of magnitude. If so, then in analogy with weak emission line quasars, we suggest that accretion from a geometrically thick, radiatively inefficient `slim disk' could explain their diverse properties in aox. Alternatively, if current Eddington ratios are in fact correct (or overestimated), then the X-ray weak tail would imply that there is diversity in disk/corona couplings among individual low-mass objects. Finally, we conclude by noting that the aox distribution for low-mass black holes may have favorable consequences for the epoch of cosmic reionization being driven by AGN.

Figures (6)

• Figure 1: Chandra images of each galaxy. Each image is $1 \times 1 \arcmin$ on a side, smoothed using a Gaussian kernel with $\sigma=3$ pixels. The regions for measuring source and background counts are shown as white solid circles and dashed annuli, respectively. An X-ray source is detected at the $>$95% confidence level within galaxies 40, 53, 56, and 111.
• Figure 2: Spectral fits ( phabs*powerlaw) to the four Chandra observations with X-ray detections (with residuals displayed as $\Delta C / \left|\Delta C\right|^{0.5}$, where $C$ is the Cash statistic). The best-fit parameters are listed in Table \ref{['tab:xrayspec']}.
• Figure 3: X-ray properties of our seven lower-Eddington ratio Chandra targets (red histograms), with cross-hatched histograms indicating upper limits. The X-ray properties of the high-$L_{\rm bol}/L_{\rm Edd}$ sample are shown for reference (open histograms). For clarity, only high-$L_{\rm bol}/L_{\rm Edd}$ objects with X-ray detections are included. (a) Hard X-ray luminosity from 2-8 keV. (b) Soft X-ray luminosity from 0.5-2 keV. (c) Best-fit photon spectral index $\Gamma$ (only four Chandra targets with X-ray detections are shown). (d) X-ray to UV luminosity ratio parameterized by $\alpha_{\rm ox}$. More negative numbers are "X-ray weaker."
• Figure 4: Top panel: BPT diagram showing narrow emission line ratios for our seven Chandra X-ray targets (star symbols), color coded by X-ray brightness $\alpha_{\rm ox}$ (see color bar). Open star symbols denote X-ray non-detections. The solid curve shows the "maximum starburst line" from kewley06, derived from pure stellar photoionization models. Galaxies above the solid curve are "Seyfert-like." The dashed curve shows the empirical dividing line between star forming and active galaxies kauffmann03. AGN from the high-$L_{\rm bol}/L_{\rm Edd}$ sample are overplotted for comparison (circles), with open symbols denoting X-ray non-detections. The bottom panels show SDSS $gri$ color composite images of each of our Chandra targets. The green scale bar at the top left of each image represents 5$\arcsec$.
• Figure 5: $\alpha_{\rm ox}$ as a function of Eddington ratio for our Chandra targets (red filled star symbols), the four mBH AGN with $\log L_{\rm bol}/L_{\rm Edd} < -1$ from yuan14, the two mBH AGN with $\log L_{\rm bol}/L_{\rm Edd} < -1$ from gultekin14, and the high-$L_{\rm bol}/L_{\rm Edd}$ sample (circles). Arrows denote upper limits on $\alpha_{\rm ox}$. The open star symbols represent our two less confident AGN candidates (galaxies 7 and 82; see Section \ref{['sec:optical']}).