First Sagittarius A* Event Horizon Telescope Results. VI. Testing the Black Hole Metric

It's two thousand seventeen, and telescopes spread across four continents are all pointing at the same patch of sky simultaneously — not to take a photograph in any ordinary sense, but to perform a single test. The question isn't whether there's a black hole at the center of our galaxy. We've known that for decades. The question is whether the black hole we see is the right kind of black hole. Whether the spacetime around it actually matches what Einstein's equations demand. That demand has a name: the Kerr metric. It's the unique solution to Einstein's equations for a rotating, uncharged black hole in vacuum — stationary, axisymmetric, asymptotically flat, with a horizon and no pathologies outside it. Crucially, it is the only such solution. This is the practical content of the no-hair theorem: once you know a black hole's mass and spin, its exterior spacetime is completely determined. No other parameters, no additional structure. Any persistent deviation from the Kerr prediction would require either extra fields or a breakdown of the vacuum Einstein equations themselves. Sagittarius A-star, the black hole at the center of our galaxy, is an unusually powerful place to run that test. It has a mass of roughly four million solar masses — sitting between the stellar-mass black holes detected by LIGO and the supermassive black hole in M87. More importantly, we have exquisite independent constraints on its mass and distance from decades of tracking orbiting stars, particularly the star S0-2. The Very Large Telescope Interferometer GRAVITY team reports a mass of four point two nine seven million solar masses and a distance of eight point two seven seven kiloparsecs. The Keck team's numbers are slightly different but consistent. These stellar-orbit measurements are entirely independent of the Event Horizon Telescope data, which makes them enormously useful: they let the Event Horizon Telescope form near-parameter-free predictions for what the shadow should look like before a single image is reconstructed. The phenomenological precession parameter for S0-2's orbit was measured at one point one plus or minus zero point one nine, consistent with general relativity at larger radii — setting the stage for a test much closer in. Now, the Event Horizon Telescope doesn't photograph the shadow directly. What it measures is a ring-like brightness distribution: a bright annulus surrounding a central depression. Translating that ring diameter into the true geometric shadow size requires careful calibration. The collaboration encodes this in a factor alpha-c, itself a product of two pieces. Alpha-one captures theoretical bias — how far the peak brightness sits from the actual shadow boundary in simulations. Alpha-two captures measurement bias — how faithfully the imaging pipelines recover the true ring diameter from sparse interferometric data. The deviation parameter delta is then defined as the inferred shadow diameter divided by the Schwarzschild shadow diameter, minus one. Delta equals zero means perfect agreement with Kerr. Pinning down those calibration factors required a massive simulation campaign. The team assembled roughly one hundred eighty thousand snapshots from time-dependent general relativistic magnetohydrodynamic — or GRMHD — simulations spanning black hole spins from minus zero point nine four to plus zero point nine four, multiple inclinations, both magnetically arrested disks and standard accretion states, and different electron heating prescriptions. A further two hundred thousand analytic models explored parameterized Kerr deviations and alternative spacetimes. To test the imaging pipelines themselves, one hundred forty-five synthetic Event Horizon Telescope datasets were generated using the SYMBA pipeline, which includes interstellar scattering and realistic atmospheric effects. Three independent imaging codes — EHT-imaging, SMILI, and DIFMAP — plus a visibility-domain ring-fitting method were run blind. The spread in reconstructed diameters was narrow for the real April seventh data, and datasets with large reconstruction ambiguity were excluded. The result: shadow diameters most likely between forty-six point nine and fifty point zero microarcseconds, with a sixty-eight percent credible envelope spanning forty-one point seven to fifty-five point six microarcseconds. The deviation parameter delta comes out at zero point zero four plus zero point one zero minus zero point zero nine using the Keck prior, and zero point zero eight plus or minus zero point zero nine using the Very Large Telescope Interferometer prior. Both are consistent with zero. The observed image size is within roughly ten percent of the Kerr predictions. The error budget has four intuitive contributions — the mass-to-distance prior, the formal ring-diameter measurement uncertainty, the theoretical spread in alpha-one, and the imaging uncertainty alpha-two — and they are well understood. But does Sagittarius A-star actually have an event horizon? The test here is elegantly simple: a surface should glow. If the black hole is replaced by some compact object with a physical surface, infalling gas deposits its kinetic energy there, the surface heats up, and it radiates. The collaboration considers two cases. For a thermally absorbing surface, even using the most conservative accretion rate — as low as ten to the negative ninth solar masses per year — the predicted steady surface luminosity is at least five times ten to the thirty-five ergs per second. The actual steady infrared flux from Sagittarius A-star sits nearly two orders of magnitude below even that lower limit. A thermal surface is ruled out, except in an extremely narrow fine-tuned window where the surface sits so close to the would-be horizon — within a fractional separation of between ten to the negative twenty-third and ten to the negative fourteenth — that steady state has never been reached. For reflective surfaces, the Event Horizon Telescope images themselves provide the constraint. The reconstructed central-to-ring brightness ratio is approximately zero point two, meaning the central region carries less than about thirty percent of the ring intensity. A perfectly reflecting surface at two point five gravitational radii produces synthetic images that fill much of the shadow with bright features — inconsistent with that depression. Partial reflectors with an albedo of zero point three are only marginally different from the black hole case, but any model that absorbs more than about ten percent of incident radiation would thermalize enough power to violate the infrared limits already discussed. Finally, horizonless objects with no surface at all — like certain boson star configurations — tend to produce either a bright interior like a radiating surface or source sizes inconsistent with the observed ring. The collaboration finds these explanations unlikely, though they acknowledge broader modeling is needed. When it comes to constraining specific departures from the Kerr metric, the shadow size measurement can be mapped onto two kinds of language: phenomenological deviation parameters and post-Newtonian coefficients. In post-Newtonian terms — an expansion of the metric in powers of the gravitational potential — the first-order deviation coefficients are constrained at roughly order unity, and the second-order ones at roughly order five. Explicit alternative spacetimes show mixed results. For charged and dilaton black holes from Einstein-Maxwell-dilaton-axion theories, charges approaching their theoretical maxima are ruled out; bounds are of order unity. A simple Morris-Thorne traversable wormhole predicts a shadow radius of E times the mass — about two point seventy-two times the mass — corresponding to a fractional deviation of roughly minus zero point four eight, immediately inconsistent with the data. Naked-singularity solutions tell a mixed story: the Reissner-Nordström naked singularity predicts shadows far too small and is ruled out, but other families — Janis-Newman-Winicour and the Joshi-Malafarina-Narayan class — survive the present test when they admit photon spheres. Throughout, the spin dependence is mild; the primary constraints trace the time-time component of the metric, not the spin. Where does this fit in the broader landscape of precision gravity tests? Solar system experiments probe weak fields — the Cassini Shapiro delay constrained the post-Newtonian parameter gamma minus one to about two times ten to the negative fifth. Pulsar timing probes strong-field matter couplings — the triple pulsar system places a Nordtvedt-type limit of eta less than three times ten to the negative fifth. The double pulsar validates quadrupolar gravitational-wave emission at the level of one point three times ten to the negative fourth at ninety-five percent confidence. Gravitational-wave observations from LIGO and Virgo probe the dynamics of stellar-mass black holes in the ten to one hundred solar mass range. Each test probes a different sector of the theory. The Sagittarius A-star result fills a specific gap. Taken together with M87 — at six point five billion solar masses — the two Event Horizon Telescope black holes span roughly three orders of magnitude in mass. Add LIGO's stellar-mass sources and the range stretches across eight orders of magnitude. All of them, so far, are consistent with Kerr. That span matters because it is precisely what general relativity predicts: the exterior spacetime of a black hole depends only on mass and spin, not on how the black hole formed or what it weighs. A hypothetical two hundred fifty-times deviation at second post-Newtonian order would have produced a shadow roughly four hundred twenty-five microarcseconds across — at least eight times larger than observed. Nothing like that appears. Pointing the largest Earth-sized telescope ever assembled at the center of our galaxy and finding agreement with a century-old theory is not an anticlimax. It is a precision statement about strong-field gravity, made by photons that threaded the deep gravitational well of a four million solar mass object. The silence — no deviation, no surface glow, no anomalous shadow — is the result. Future expanded arrays, space-based interferometers, and searches for time-delayed electromagnetic echoes could yet find a crack. For now, the data say: Kerr. This lecture was created by ennepō. Go to https://ennepo.ai to Discover, Create and Follow the latest research in your field. Read when you can. Listen when you want to.