This is an image of the supermassive black hole at the center of our Milky Way galaxy, known as Sagittarius A*. The black hole itself doesn’t emit light, so what we’re seeing is the hot plasma swirling around it. This is only the second direct image of a black hole ever made. It was captured by the Event Horizon Telescope (EHT) collaboration, the same team that produced the famous first image of a supermassive black hole at the center of the galaxy M87.

Originally, the plan was to image Sagittarius A* first. Because it’s in our own galaxy, it’s about 2,000 times closer to us than M87*. However, it’s also more than 1,000 times smaller, so from Earth it appears only slightly larger than M87* in the sky. On top of that, there are several additional challenges.

First, there’s a huge amount of dust and gas between us and the galactic center, so you can’t see it with visible light at all. In a zoom-in video of the Milky Way’s core from the European Southern Observatory, as we get closer, we eventually have to switch to infrared light, which can penetrate the dust much better. That allows us to observe the center of the galaxy from Earth.

Over the past three decades, astronomers have peered into the heart of the Milky Way and seen something remarkable: a cluster of stars whipping around on highly eccentric orbits. They move incredibly fast. One star has been measured traveling at about 24 million meters per second—roughly 8% of the speed of light. All of these stars seem to be orbiting something extremely massive and extremely compact, but this object isn’t shining brightly like a star. It flickers occasionally, but otherwise it’s quite dim. This is what we believe to be a supermassive black hole.

From the motion of the stars around it, we infer that the black hole’s mass is about 4 million times that of the Sun, all crammed into an extremely tiny region—a singularity. Anything, including light, that comes within a distance known as the Schwarzschild radius cannot escape; it inevitably falls into the singularity.

So if we see any radiation associated with the black hole, it must come from outside this radius, typically from superheated plasma in the accretion flow as it falls inward. But for a black hole of its size, Sagittarius A* doesn’t actually eat very much matter. It’s unusually quiet and dark.

By contrast, the supermassive black hole at the center of M87 is much more active, feeding on material from its thick accretion disk. And because it’s more than 1,000 times larger than Sagittarius A*, orbital timescales near it are 1,000 times longer. That means its appearance changes much more slowly over time, making it easier to image. Sagittarius A*, on the other hand, can vary on timescales of minutes.

Some of the visualizations you may have seen come from Luciano Rezzolla and colleagues at Goethe University Frankfurt.

How Tiny Is a Black Hole on the Sky?

The biggest challenge in imaging either of these supermassive black holes is how incredibly small they appear in the sky.

To get a sense of scale, start with the whole sky and divide it into 180 degrees. The Andromeda galaxy spans about 3 degrees. Now take a single degree and divide it into 60 arcminutes. Each arcminute can be divided into 60 arcseconds. Then take one arcsecond and divide it by 100, then by 100 again, and then by 100 once more. The apparent size of these black holes on the sky is around that scale.

It’s equivalent to trying to take a picture of a donut sitting on the surface of the Moon.

No optical telescope on Earth has enough resolution to produce an image like this. So we want to answer two questions:

How did they actually do it?

What are we really looking at?

How the Image Was Made: Radio Astronomy and Resolution

First, these images were not made with visible light. They were made using radio waves with a wavelength of about 1.3 millimeters. That means all the observations were taken with radio telescopes—essentially giant satellite dishes.

When a distant source emits radio waves, they spread out in all directions. But because Earth is so far away, by the time the waves reach us, the wavefronts are almost perfectly flat and parallel. This is called a plane wave.

A radio telescope scans back and forth across the sky. When it is pointed directly at a radio source, it produces a strong signal. In that orientation, all the radio waves reflecting off the dish travel the same distance to the receiver and arrive at the same time. The peaks and troughs of the waves line up—this is called being “in phase”—so they interfere constructively, creating a bright signal.

As the telescope moves past the source, waves from different parts of the dish start traveling different distances. They arrive out of phase and interfere destructively, so the signal intensity drops toward zero.

To create a sharp image, you want this drop-off to be as steep as possible. Ideally, the telescope produces peak intensity only when aimed directly at the source, and the signal falls off quickly when the dish is moved even slightly.

There are two main ways to increase this angular resolution:

Use higher-frequency (shorter-wavelength) radio waves.
At shorter wavelengths, a small movement corresponds to a larger fraction of a wavelength, so destructive interference happens sooner.

Increase the diameter of the telescope.
A larger dish increases the difference in path length between waves hitting different parts of the telescope for a given angle, which also increases resolution.

A telescope’s ability to pinpoint where radio waves are coming from on the sky is called its angular resolution. You can think of it as the size of the “spot” on the sky that the telescope is sensitive to. It is proportional to the wavelength and inversely proportional to the diameter of the dish.

The challenge with imaging a black hole is that you’re trying to resolve structure within an extremely tiny patch of the sky. If you scanned a single radio telescope across the center of the black hole, you’d want to see a bright spot at the left edge of the ring, then a dark region in the middle, and then another bright spot at the right edge.

But for any single radio dish on Earth, the angular resolution is too low. As the telescope sweeps across the black hole, it will still be detecting radio waves from the left side at the same time it starts detecting waves from the right side. It can’t clearly distinguish a ring from a blob.

Observing at much shorter wavelengths isn’t practical, because those wavelengths are blocked either by our atmosphere or by material around the black hole. So if you want better resolution, the only option is to effectively increase the size of your telescope’s diameter.

If you do the math, you find that to resolve the ring structure around a black hole like Sagittarius A*, you’d need a telescope roughly the size of the Earth—which is obviously impossible to build as a single dish.

However, there’s a clever workaround: you don’t need a solid dish the size of Earth. You just need pieces of it—individual radio telescopes spread across the planet, separated by distances up to Earth’s diameter. As long as you can properly combine the signals from all those telescopes, you can get the same angular resolution as one Earth-sized dish. This method is called very long baseline interferometry (VLBI).

The Event Horizon Telescope is not a single instrument but a global network of radio observatories. All of these telescopes observe Sagittarius A* (or M87*) at the same time.

Unlike a single telescope, you can’t just bounce all the radio waves into one central receiver in real time. Instead, each observatory records the incoming signal at its location along with extremely precise timing information—down to femtoseconds (quadrillionths of a second). The result is petabytes of data.

That data then needs to be brought together. Surprisingly, the fastest way to do this was simply to fly hard drives to central processing locations as hand luggage.

Combining the Signals: Interference Patterns

So what do we have? Electrical signals and timestamps from multiple radio telescopes around the world. Each individual telescope still lacks the angular resolution to see the ring of the black hole. So how does combining the data reveal finer details than any single telescope can see?

The key is that there is information encoded in the relative positions of the telescopes and the slight time delays between when the same wavefront hits each one.

Imagine combining the signals from two widely separated telescopes. Suppose they both receive the same wave at the same time, and the peaks and troughs line up—they’re in phase. That tells us the source lies somewhere along a line in the sky where the distance to both telescopes is the same, so that the radio waves traveled equal distances to reach them.

But with just two telescopes, you don’t know where along that line the source is. And there’s another complication: the source might be one full wavelength closer to one telescope than the other, or two wavelengths closer, or three, and so on. Because waves repeat every wavelength, they could still arrive in perfect phase even with those path differences.

So from one pair of telescopes, the information you get about the source is a pattern of bright and dark fringes on the sky. Pairs of telescopes that are close together produce wide fringes; pairs that are far apart produce narrow fringes.

To reconstruct an image, you need many pairs of telescopes at different separations and orientations on Earth. Each pair provides a different fringe pattern. By combining all of these patterns mathematically, you can reconstruct an image of the radio source—in this case, the ring around a black hole.

What Are We Actually Seeing?

Now that we have this picture, what exactly are we looking at?

Here’s one way to think about it. Imagine a simple model black hole. A sphere around it represents the event horizon. Once inside this radius, there is no escape—not even for light. The radius of the event horizon is the Schwarzschild radius.

If we looked at an isolated black hole with nothing around it, we wouldn’t see anything. It would just absorb all the electromagnetic radiation that fell onto it. But the black hole observed by the EHT has matter around it in an accretion disk—gas and dust swirling in chaotic orbits, heated to millions of degrees and moving at a significant fraction of the speed of light.

This matter is what the black hole feeds on over time, gradually growing more massive. However, you’ll notice that the accretion disk doesn’t extend all the way to the event horizon. Why not?

Because there is an innermost stable circular orbit (ISCO). For matter orbiting a non-spinning (Schwarzschild) black hole, the ISCO lies at three Schwarzschild radii. Inside this radius, stable circular orbits are no longer possible. If matter moves inside the ISCO, it rapidly spirals into the black hole, and we never see it again.

In reality, the black hole at the center of our galaxy is almost certainly spinning, which complicates things. But for simplicity, let’s stick with the non-spinning case here.

There is one thing that can orbit closer than the ISCO: light. Because photons are massless, they can orbit at 1.5 Schwarzschild radii. This region is called the photon sphere. While we often draw it as a ring, it’s really a spherical shell of possible photon orbits in all directions.

If you could somehow stand there (which is impossible), you would be able to look straight ahead and see the back of your own head, because photons could loop around the black hole and meet your eyes from behind.

However, the photon sphere corresponds to an unstable orbit. Photons there eventually spiral inward into the black hole or outward to infinity.

Now, what does the dark “shadow” region in the EHT image correspond to in this picture? Is it the event horizon itself? The photon sphere? The ISCO?

The situation is complicated because the black hole strongly warps spacetime, bending the paths of light rays. Light travels in straight lines through spacetime, but spacetime itself is curved, so from our perspective, the paths look curved.

A helpful way to think about it is to imagine parallel light rays coming in from us toward the black hole. Any ray that crosses the event horizon is gone; we never see it. Those impact parameters define a totally dark region.

But light rays that pass just outside the event horizon also get bent so strongly that they still cross the horizon and fall in. Even rays passing at the same radial distance as the photon sphere get bent into the black hole.

For a light ray to come in and not fall into the black hole, it has to start out at a slightly greater distance. In fact, it must pass at about 2.6 Schwarzschild radii. Such a ray will just graze the photon sphere at its closest approach and then escape back out to infinity.

As a result, the apparent “shadow” of the black hole we see is about 2.6 times larger in radius than the event horizon itself.

At the very center of this shadow, in terms of mapping, lies the event horizon. But rays coming from above or below the black hole can also bend around and land on different parts of the horizon, including the back side relative to us. So the entire back surface of the event horizon gets mapped into a ring on this shadow.

From our single line of sight, we effectively get to see the entire event horizon projected into this dark region—even though, of course, the horizon itself emits no light.

It gets even stranger. Light can loop around the black hole multiple times before escaping. That means you can have photons that circle once and escape, forming another, fainter ring; or circle twice; and so on. In principle, there are infinitely many nested images of the event horizon, compressed closer and closer to the edge of the shadow.

The first light we see—the brightest ring—comes from rays that just skim the photon sphere and escape to our telescopes. They create the familiar ring whose diameter is about 2.6 times the event horizon’s diameter.

This is roughly what we’d see if we were looking straight down on the accretion disk, perpendicular to its plane. But in reality, we’re much more likely to view the disk at some random angle, possibly even nearly edge-on.

You might think that looking edge-on would obscure the black hole’s shadow. But because of the way the black hole warps spacetime and bends light, we can actually see the back side of the accretion disk as well. Light from the far side of the disk bends over the top of the black hole and reaches us; light from the underside can bend around and come up toward us from below.

This effect is what produces images like the black hole in the movie Interstellar—a bright ring that appears to wrap over and under the dark central region.

It gets even more dramatic. Light emitted from the top of the accretion disk can travel around the back of the black hole, skim the photon sphere, and emerge from below the shadow as a thin ring. Similarly, light from the bottom of the disk can curve under and emerge above. These additional lensed paths produce thin, secondary rings just inside or outside the main shadow.

If we were extremely close to the black hole, we’d see an incredibly complex and beautiful structure of warped rings and arcs of light.

Relativistic Beaming and the Bright Spot

There’s one more crucial effect: the material in the accretion disk is moving at speeds close to that of light. On the side of the disk that’s rotating toward us, the radiation is relativistically beamed in our direction, making it appear much brighter. On the side moving away from us, the light is beamed away, making it dimmer. This is called relativistic beaming or Doppler beaming.

As a result, one side of the ring around the black hole appears noticeably brighter than the other. That’s why in the EHT images, you see a sort of “hot spot” or brighter arc rather than a perfectly uniform ring.

Hopefully, this helps make sense of what we’re really seeing when we look at an image of a black hole.