Imaging a black hole is no easy task. Black holes are completely dark, meaning that taking a direct image of them is impossible. Instead, scientists must look at the glowing gas and dust that surround them. Even though supermassive black holes are intrinsically large, they’re so far away from us that their apparent size on the sky is extremely small, so small that no conventional telescope can discern their shape. To solve this problem, the EHT combines several radio telescopes located world-wide to form one huge, Earth-sized telescope, using a technique called Very Long Baseline interferometry.

Imaging Sgr A* proved challenging for scientists. In this blog, we spoke to three of the scientists in the EHT Collaboration to understand some of the similarities and differences between the images of Sgr A* and M87*, as well as getting to the bottom of why imaging our galaxy’s central black hole was so difficult.

These panels show the first two images ever taken of black holes. On the left is M87*, the supermassive black hole at the centre of the galaxy Messier 87 (M87), 55 million light-years away. On the right is Sagittarius A* (Sgr A*), the black hole at the centre of our Milky Way, 27 000 light-years away. The two images show the black holes as they would appear in the sky, with their bright rings appearing to be roughly the same size –– M87* is physically larger than Sgr A* but it’s much further away.

Credit: EHT Collaboration

Why do the images of Sagittarius A* and M87* look so similar, when we know the black holes are very different?

“There are two important things to note,” says Kazu Akiyama, an astrophysicist at MIT Haystack Observatory in the US. “Firstly, both images have a circular ring. Secondly, the rings are broadly speaking symmetric: they appear closed, without gaps.”

“The circularity is a beautiful result of the extremely strong gravity at the edge of a black hole. The size and shape of the black hole’s silhouette is governed simply by how its strong gravity bends the surrounding space-time. Einstein’s theory of relativity predicts that the shape of the silhouette is pretty circular no matter how fast the black hole is rotating or from where we are looking at it.”

You may think that the circular shadow that we see in the EHT images corresponds to the event horizon –– the surface beyond which not even light can escape. But that’s not the case. As the video below shows, the black hole bends the trajectory of light around it, casting a shadow that is about 2.5 times larger than the event horizon itself.

Hot matter orbits the black hole in a disc, and the orientation of this disc will determine whether the bright ring looks symmetric or not. “If the rotational axis of the matter around the black hole is tilted relative to our line of sight,” Kazu says, “the bright ring enclosing the shadow of the black hole will be asymmetric: the side where matter is approaching us will be brighter while the receding side will be dimmer. If the tilt is very pronounced we would see an unclosed ring where only a single side is visible, like a waxing crescent moon.”

“I was indeed surprised and didn’t even expect to see such a closed ring for Sgr A*,” says Kazu. “There was no guarantee that it should be. This second ring was basically a gift from nature.”

Black holes can also emit strong jets of material perpendicularly to the disc, which give astronomers a clue as to the orientation of the disc. Unfortunately, Sgr A* has been relatively quiet on this front.

“In contrast with M87*, which has a prominent jet indicating the orientation of the system, Sgr A* has never shown us a clearly visible jet,” says Katie Bouman, an assistant professor at the California Institute of Technology (Caltech). “For this reason, Sgr A* has been hiding its orientation from us for nearly 50 years until it was finally imaged with the EHT. As a symmetric ring appears in our image, we found that the theoretical models can explain EHT data only when the rotational axis is tilted within 30 degrees from our line of sight.”

So, what made imaging Sgr A* so difficult compared with M87*?

“There were two primary challenges in imaging Sagittarius A*, arising from its unique properties,” says José L. Gómez, a Research Scientist at the Instituto de Astrofísica de Andalucía in Spain. “Firstly, Sagittarius A* is about 1600 times lighter and smaller than M87*. Material takes days or weeks to travel around M87* but just minutes or hours to get around Sgr A*. The EHT fills up a gigantic, planet-sized virtual mirror using the Earth’s rotation. Each individual EHT telescope is like a segment of this mirror, gradually obtaining data over a night. M87* was easier to image, as it was steady when we were observing: the data obtained by all individual telescopes came from the same stable image.”

“However, Sgr A* is like a toddler who can’t stay standing still while we are taking its photo over the course of a night,” continues Katie. “The glowing gas around Sgr A* was dancing while we were taking the data. We needed to reconstruct how Sgr A* appears in the sky from a series of sparse information obtained while the gas was quickly moving around.”

José L. continues: “The second challenge is that radio signals from Sgr A* become somewhat distorted and blurred by turbulent gas located between Earth and the Galactic centre. The image of Sgr A* ripples like a distant mountain seen from a window of an aeroplane through its hot jet exhaust fumes. To see a true image of Sgr A*, we need to reconstruct the image as it was before such turbulent gas scattered the radio waves.”

“This was not the case for M87*,” adds Kazu, “which is located far away from the plane of the Milky Way, and therefore we did not have to image it through the dense material in our galaxy.”

How did you address these problems?

“For the scattering effects, we carefully examined the optics of propagation of radio waves from Sgr A* through the turbulent plasma that scatters them,” says Kazu. “Over the last several years, EHT scientists have built up a sophisticated model of the scattering for Sgr A* with an extensive series of observations at different wavelengths, and used it to assess how the distortion and blurring of Sgr A* appears in the EHT data. This allows us to reconstruct the image of Sgr A*.”

“As for the rapid motion of the gas around Sgr A*,” says José L., “we indeed saw in the EHT data that the glowing gas around Sgr A* is moving and dancing around during our observations. It is truly stunning that the EHT could see such complicated dynamics of the gas around a black hole, giving us a very important clue about how black holes interact with their surrounding environment.”

“We did a careful statistical analysis of the data to see how the appearance of Sgr A* changes from night to night,” explains Katie, “providing a statistical model of the dynamical structure in Sgr A*. This allows us to image the common structure of Sgr A* seen across the nights.”

As evidenced from the above answers, it is clear that a huge amount of analysis was undertaken to obtain an image of Sgr A*. And some very well developed algorithms were required for this analysis.

Could you describe how the different algorithms that you used work?

“The EHT creates the sharpest eye on the universe by computationally forming a planet-sized virtual telescope with radio telescopes spread across the Earth,” says Katie. “But we have only eight telescopes on six geographic sites, which provide a sparse coverage of telescopes that can fill up only a tiny fraction of this gigantic virtual mirror. It’s like having a fraction of the pieces of a jigsaw puzzle, from which we need to reconstruct what the entire picture of the puzzle looks like. Or like listening to a song from a piano where many of the keys are missing.”

“Because we have such limited information, you can imagine that there are an infinite number of ways to fill up the missing pieces of the puzzle,” continues Kazu, “though most of them wouldn’t make any sense, just like randomly adding a piece to a puzzle or a note to a song results in something awkward and very unlikely.”

“Imaging algorithms are mathematical detectives that we use to find the most reasonable images among the many possibilities that can explain the EHT’s measurements,” Katie explains. “Each algorithm has its own method for determining which image is most likely. It is a bit like hiring Sherlock Holmes, Hercule Poirot, Jane Marple, and Jules Maigret simultaneously to see what they commonly conclude and what they don’t.”

How accurately can the algorithms recover the true shape of Sgr A* or any other source?

“We vetted and evaluated each way of imaging EHT data according to whether it can distinguish and recover different shapes,” says Kazu. “including non-ring structures that may provide data similar to EHT data. Thousands of Sgr A* images were made from well-vetted methods that can distinguish different structures, and the vast majority show a ring shape.“

“The rapidly varying emission from Sagittarius A* and the limited amount of data introduced some uncertainty in where the brightest parts of the ring were,” continues José L. “The ongoing expansion of the EHT network and significant technological upgrades will allow us to better constrain the emission around the ring, and even obtain the first movies of black holes.”

The main image of Sgr A* (top) was produced by averaging together thousands of images created using different computational methods. These images can also be clustered into four groups based on similar features (bottom). The bar graphs show the relative number of images belonging to each cluster. As you can see, non-ring images (bottom right) are very rare.

Credit: EHT Collaboration

I can’t help but ask, how confident are you that the image is a ring? What could lead to a non-ring image?

“We have spent years developing sophisticated new tools to account for the challenges in imaging Sagittarius A*," José L. told us. “Tens of millions of images from test data (created to best resemble Sgr A*) have been produced in supercomputers around the world to refine our algorithms, much like searching for the best lens and filter in a camera to obtain the sharpest snapshot.”

“We found that our different algorithms can reliably distinguish between ring and non-ring images,” Katie adds. “Most importantly, as we found when imaging Sgr A*, only a very small fraction of non-ring images appear. Based on this detailed study we conclude that the non-ring images are caused by the limited amount of data and rapidly varying emission from Sagittarius A*, rather than being intrinsic to the black hole itself.”

“The image of Sagittarius A* now holds the record (previously set by M87*) of the most thoroughly vetted interferometric image that has ever been made,” José L. concludes.

Now that we have imaged Sgr A*, what’s next for the EHT?

“This is just the dawn of the exciting era opened by the EHT,” José L. tells us. “The EHT has been getting more powerful since 2017 when we obtained data providing these first images of black holes. Now, the EHT has a more sensitive and sharper eye, with new additional telescopes and upgraded instruments. We will work on new data from the upgraded EHT in the coming years, which will provide a sharper and more dynamic view of black holes.”

“Of course challenges lie ahead too,” Kazu explains. “As the EHT becomes more powerful, we will have much richer data. The sheer volume of EHT data, which already totaled 5 petabytes in 2017, has grown even more since then. How we handle this and hunt for treasures hiding in such enormous datasets will be our next exciting challenge—with accordingly enormous rewards.”

“The EHT is also looking at other supermassive black holes, investigating the hot plasma in their vicinity. This could one day lead to an eventual third image of a black hole — but only if we have underestimated the sizes of these black holes; if they are too small, we won’t be able to discern their shadows.”

Of course, just the story behind the image of Sgr A* is one for the history books, and we asked our three interviewees what advice they’d give to future astronomers.

“Dream big, stay curious, and don't let anyone tell you that you can't succeed,” says Katie. “If you set your mind to it, and work hard, oftentimes you can achieve what originally sounded impossible. Remember that you can be your own worst enemy –– once you start doubting yourself then the game is over. It's normal to feel unsure of yourself, but just put those blinders on and plow ahead!”

José L. agrees: “Follow your dreams! It is your imagination, and the fascination to understand how the Universe works. These are the only things you need to make every day of your life a journey of discovery.”

“Don’t be afraid of the risk of failure,” Kazu adds. “With the last ten years of my life at the EHT, I learned that whatever happens and however it ends up, the entire experience you will get will be a very fruitful element of your life.”

About the EHT scientists

Note that the images below are copyright of their respective owners, as mentioned in the credits, and are not released under our usual CC-BY license.

Credit: K. Bouman Katherine L. (Katie) Bouman is an assistant professor in the Computing and Mathematical Sciences, Electrical Engineering, and Astronomy Departments at the California Institute of Technology (Caltech). Before joining Caltech, she was a postdoctoral fellow in the Harvard-Smithsonian Center for Astrophysics. She received her Ph.D. in the Computer Science and Artificial Intelligence Laboratory (CSAIL) at MIT in EECS, and her bachelor's degree in Electrical Engineering from the University of Michigan. She is a Rosenberg Scholar, Heritage Medical Research Institute Investigator, recipient of the Royal Photographic Society Progress Medal and IST Electronic Imaging Scientist of the Year Award. As part of the Event Horizon Telescope Collaboration, she is co-organizer of the Imaging Working Group and acted as coordinator for papers concerning the first imaging of the M87* and Sagittarius A* black holes. She is also co-organizer of the Algorithms and Inference Working Group for the next-generation Event Horizon Telescope (ngEHT) effort.

Credit: K. Akiyama Kazu Akiyama is a research scientist at MIT Haystack Observatory. He joined the team of the Event Horizon Telescope (EHT) in 2010, and has worked on various aspects of EHT observations, from the development of new imaging algorithms to the calibration, imaging, and scientific interpretation of EHT data. He co-founded the Imaging Working Group of the EHT Collaboration in 2017, and has been a co-leader of the group since its establishment. As a lead of the EHT’s imaging team, he has coordinated the first imaging of two primary targets: M87* and Sgr A*. He also served as a coordinator of the EHT Sgr A* data calibration team. He developed one of the imaging software package, SMILI, used to create the first images of M87* and Sgr A*.

Credit: J. L. Gómez José L. Gómez is a Research Scientist, and head of the EHT group at the Instituto de Astrofísica de Andalucía (CSIC), in Granada, Spain. His research career has been focused on the study of relativistic jets often seen emanating as a byproduct of black hole accretion. He is a member of the EHT Science Council, and one of the coordinators of the EHT Imaging Working Group. He developed the scripts based on the classical CLEAN method for imaging interferometric data used for imaging M87* and Sgr A*, being one of the coordinators of the imaging papers for these two EHT primary targets. He also participates actively in other EHT scientific activities, in particular in the Polarisation, AGN, and Multi-wavelength Working Groups.