First Sagittarius A* Event Horizon Telescope Results. I. The Shadow of the Supermassive Black Hole in the Center of the Milky Way

Decades of observations have provided a picture of our local SMBH that is unmatched in any other galaxy [38, for details about the full spectrum, see]. Sgr A∗ has been detected from long radio wavelengths (∼1 m) to the hard X-ray band, excepting approximately 1 μm to 1 nm due to extinction from dust in the galactic plane. Sgr A∗ is remarkable for its feeble emission, producing a bolometric luminosity of <∼10³⁶erg/s, only ∼100 times that of the Sun. Were it located in another galaxy, it would likely go undetected. Nevertheless, by observing its spectrum and variability, its environment, and its influence on surrounding bodies, a great deal has been learned about this source specifically, and about the astrophysical processes that operate around supermassive black holes. In this section we describe how we assembled our current knowledge of Sgr A∗, discuss important theoretical uncertainties about its accretion and outflows, and compare it with the other horizon-scale EHT target, M87∗.

II.1 Properties of Sgr A*

The proximity of Sgr A∗ permits precise measurements of its gravitating mass via the monitoring of resolved individual stellar orbits. High-resolution infrared (IR) observations, using increasingly sophisticated instrumentation and analyses, have traced out the three dimensional orbits of several stars within the innermost arcsecond around Sgr A∗. These orbits jointly determine the mass and distance to Sgr A∗ to high precision, particularly the ratio M/D that determines the angular size of the black hole on the sky. As discussed in and, the current values for the mass and distance suggest an angular shadow diameter close to 50 μas, comparable to that of M87∗. The closest orbital periapses confine the mass to within ∼1,000 Schwarzschild radii (RS=2GM/c²).

Radio observations of Sgr A∗ have provided a significant motivation for the development of the EHT experiment. Sgr A∗ shows a flat/inverted radio spectrum, which often arises from compact jet emission in other low-luminosity active galactic nuclei (LLAGN; Ho 62, Nagar et al. 86). Such a spectrum can, however, result from any stratified, self-absorbed synchrotron source, where successively higher frequencies are produced at increasingly smaller scales, even without a jet [91, e.g.,]. Sgr A∗ shows an excess of millimeter emission above the flat centimeter-wave spectrum, the so-called “submillimeter bump,” that was inferred to indicate the presence of a very compact emitting region at these wavelengths [123, 44, e.g.,].

To clarify the nature of this source, the structure of Sgr A∗ was investigated using very-long-baseline interferometry (VLBI) at progressively shorter wavelengths [see 38, and references therein]. For wavelengths longer than several cm, the observed source size is entirely determined by scatter-broadening in the ionized interstellar medium, scaling with a wavelength dependence of λ². At wavelengths of 7 mm and shorter, the imprint of the intrinsic structure of Sgr A∗ became discernible through the scattering [108, 75, 24, 12, 115, 13, e.g.,]. The source grows more compact at shorter wavelengths, as expected [94, e.g.,], though it does not present a clear jet structure. For wavelengths as short as ∼1 mm, Sgr A∗ is only slightly blurred by scattering, and Falcke et al. [46] predicted that submillimeter VLBI could directly image a brightness depression within this region related to the black hole shadow.

In parallel with these developments and motivated by the goal to study black holes, the capabilities of mm-VLBI improved rapidly. At 1.4 mm, Padin et al. [96] reported the first VLBI fringes and Krichbaum et al. [71] obtained the first VLBI detections and associated source size measurements for Sgr A∗, each using a 2-element interferometer. Doeleman et al. [25] observed Sgr A∗ at 1.3 mm with a 3-element VLBI array and reported the discovery of an intrinsic source size comparable to the expected angular diameter of the black hole shadow. These observations provided important constraints for theoretical models [15, 22, 85, e.g.,] and strongly suggested that 1.3 mm VLBI has a clear view into the innermost region around the Sgr A∗ black hole. Subsequent VLBI studies of Sgr A∗ at 1.3 mm with progressively enhanced arrays revealed the compact emission to be variable and significantly polarized, with measured “closure phases” that are indicative of persistent asymmetry in the image structure.

Observations of Sgr A∗ outside the radio band were important for completing the picture of this source as an LLAGN. When X-ray and gamma-ray instruments ROSAT and Sigma/GRANAT could not identify a bright central source, it became clear that Sgr A∗ must either be obscured or anomalously faint compared to other known LLAGN. The first identification of Sgr A∗ as a compact and variable source in the X-rays was achieved with the Chandra X-ray Observatory (Chandra). This detection, together with the ∼4×10⁶ M⊙ mass of Sgr A∗ determined from stellar orbits, sets a maximum luminosity scale for the source, the so-called Eddington luminosity, and the X-ray measurements confirmed that Sgr A∗ is approximately 9 orders magnitude less luminous than its LEdd — the lowest Eddington ratio (OPENL/LEdd) observed for any black hole.

This particularly faint high-energy emission stimulated a lively debate regarding the nature of Sgr A∗’s accretion/outflow properties, and motivated new theoretical developments. If the emission primarily originates in the accretion inflow, either it must be exceptionally radiatively inefficient, or the accretion rate onto the black hole must be very small compared to what is captured, or possibly some combination of the two. Alternately the emission could be dominated by an outflow, in which case a small accretion rate would also be favored, even during flares. Fortunately, the proximity of Sgr A∗ enables unparalleled study of the accretion flow. Observations with Chandra marginally resolve thermal bremsstrahlung emission from near the gas capture radius, leading to an estimate of the captured accretion rate of ∼10⁻⁶−10⁻⁵M⊙/yr at ∼10⁵RS. The source of this gas can be connected to the observable stellar winds of ∼30 individual massive stars found in the inner parsec. The near-horizon accretion rate has also been estimated to be 10⁻⁹−10⁻⁷ M⊙/yr through measurements of the Faraday rotation of polarized millimeter emission from Sgr A∗. This Faraday rotation is orders of magnitude smaller than would be expected if the accretion rate at the gas capture radius persisted to small radii. This conclusion is further supported by radially resolved X-ray spectroscopy suggesting that only ∼1% of the captured mass makes it to the SMBH. Taken together, the low luminosity, low radiative efficiency, and weak Faraday rotation are consistent with a weakly bound, magnetized accretion flow so diffuse that the electron and ion temperatures are unable to remain strongly coupled.

Finally, Sgr A∗ exhibits flaring emission at most wavelengths, including continuous variability at millimeter wavelengths. In particular, approximately daily flares are observed at X-ray and NIR wavelengths that are often (but not always) simultaneous [26, e.g.,] and that change on timescales as short as minutes [3, 51, e.g.]. These timescales suggest an origin from within ∼5RS of the SMBH, which is consistent with astrometry of NIR flares using the GRAVITY instrument on the Very Large Telescope Interferometer.

II.2 Comparison to M87∗

M87∗ and Sgr A∗ have the largest angular sizes of any known SMBHs, making them the primary EHT targets. However, they differ in several important ways. First, they have substantially different luminosities and accretion rates, both in absolute terms and when scaled by mass. M87∗, roughly 1500× more massive [30, M=(6.5±0.7)×109M=(6.5\pm 0.7)\times 10^{9} M⊙M_{\odot};], has an inferred accretion rate of M˙∼10⁻³M⊙/yr and a bolometric luminosity of Lbol≈10⁴²erg/s measured in 2017. Estimates of the total kinetic power are typically ∼10−100 times larger, thus conservatively L/LEdd≳10⁻⁵. Most likely this higher power indicates that M87∗ is being fed directly from a larger reservoir of gas rather than a trickle from stellar winds. This difference may underlie another substantial divergence between the sources: the prominent, powerful jet launched from M87∗, which can be traced at multiple wavelengths and across nearly eight orders of magnitude in size. The M87∗ jet provides firm constraints on the source orientation with respect to the line of sight; thus, an inclination of ∼20° was used for numerical simulations in, while for Sgr A∗ we do not have any such constraints.

The difference in the masses of Sgr A∗ and M87∗ implies a similar difference in their variability timescales. The period of the innermost stable circular orbit (ISCO), which depends on the mass and spin of the black hole, serves as an approximate dynamical timescale. For prograde orbits, this period ranges from 4πtg (maximal spin) to 126πtg (zero spin), where tg≡GM/c³. For M87∗, this range corresponds to 5 days to 1 month, so the source structure is expected to be effectively unchanged over the course of an observing night. Indeed, EHT images of M87∗ reconstructed on consecutive nights are almost identical. However, for Sgr A∗, the range is only 4–30 min, so the source structure can evolve within a single night.