event horizon telescrope project

We’re scientists on the Event Horizon Telescope Project looking to capture an image of a black hole. Ask us anything about the telescope, astronomy, physics or black holes! AMA!

Hello, we are scientists that are a part of the Event Horizon Telescope project.

This telescope array is using Very Long Baseline Interferometry (VLBI) to create a composite image of the event horizon of the black hole, Sagittarius A*. Unlike a photograph – which is composed light hitting a single focal point on an optical lens which is captured by the camera – the EHT project is capturing data from 1.3mm radio wave detections from around the world to create a “virtual mirror” that will help create the first image of a black hole.

Proof or check out this PBS special

Please note that we will begin posting answers at 11am PDT/2pm EDT, as Avery, Dimitrios and I are in meetings/teaching this morning.

About the project:

  • Our group is currently using 9 telescope arrays with locations across the Earth in Mexico, Chile, Hawaii, and Spain
  • 3 other arrays will be incorporated as well, including a location at the South Pole
  • The project is capturing this data on 126 HGST Ultrastar helium-filled 6TB hard drives; currently 756TB of storage with plans to expand to 6PB
  • The hard drives are encased in a custom enclosure of eight drives each that process data at the speed of 64 Mb/sec
  • Each day an observation is run at a site, the site captures 350TB of data
  • 75 Scientists are currently contributing to the project
  • For context, EHT is processing ~10x the amount of data of the Large Hadron Collider in Switzerland

Today, you have three astrophysicists answering your questions:

  • Shep Doeleman, Assistant Director, MIT Haystack Observatory and Astronomer at Smithsonian Astrophysical Observatory
  • Dimitrios Psaltis, Professor of Astronomy and Physics, University of Arizona
  • Avery Broderick, Assistant Professor of Physics and Astronomy, University of Waterloo; Associate Professor, Perimeter Institute for Theoretical Physics

Ask us anything!

Thanks for attending – we’re wrapping things up here – we had a ton of fun! To learn a bit more, please see this month’s Scientific American:

http://www.scientificamerican.com/article/does-einstein-s-theory-of-gravity-hold-near-black-holes/

How much more can we learn about black holes before we hit the ceiling where we will need to get instruments close to black holes to make further discoveries?

[Shep Doeleman] A running joke among BH researchers is that we need to send one of us with a laser pointer to the event horizon and then point the laser back to earth. That would give us a lot of info, but also cost us an astronomer (good value?). The EHT will get us closer to the BH and Event Horizon than we have ever been, and in addition to a possible image, we can also trace orbits of material in time, so there are a number of cool experiments that will play out over the next 3-5 years. A lot to keep us occupied, so the ceiling is some ways away!

If a black hole swallows only electrons, will the electrical repulsion be eventually greater than the gravitational attraction?

[Shep Doeleman] This is a very interesting question, and it’s true that if a black hole ingested only particles of one charge it could accumulate charge and then repel like charges, but what may happen in reality is that protons or positrons would be attracted to the charged BH and cancel the charge. In other words, BH’s are always expected to be neutral.

You mentioned that this will be able to create the first image of a black hole. All we have now are drawings of we think it looks like. Do you think it will look similar, or completely different than our idea of a black hole?

[Shep Doeleman] The beauty of the EHT is that we really don’t know exactly what we’ll see. If we understand the physics and General Relativity, then we can predict what we’ll see: a bright ring near the last orbit photons can trace with a dim interior (caused by gravity pulling ferociously on escaping photons coming toward us). Einstein’s theory predicts what weshould see, but if we see something else (a warped shadow or a completely unexpected shape) then things get very interesting.

What are you most eager to see when the image is created? What questions do you hope this project will answer?

[Avery Broderick] Anything at all! In many ways this is a voyage of discover to the edge and back, and whatever we find will be transformative. More practically, the first key item is the shadow cast by the black hole event horizon cast against the bright surrounding plasma. The size and shape of this shadow will provide the first test of general relativity in the strong gravity regime, where the counter-intuitive aspects become most severe. Ultimately, we hope to move to high-precision tests of gravity in this heretofore unprobed regime, either supporting or disproving Einstein’s general relativity! Beyond that, we will learn a lot about how black holes interact with the universe around them, informing how they impact the formation and evolution of their host galaxies.

If we had a radio telescope on the surface of the Moon, do we currently have the technology and know-how to incorporate that into a super-VLBI? What other kinds of objects will you image with the EHT?

[Shep Doeleman] Wow! Someone is thinking BIG! We could certainly put a station on the Moon and the resulting magnification factor would be huge – 80,000 times better than Hubble (check me on that). The problem is that this might be too good in that we probably would not have the sensitivity (unless the antenna on the moon was very large) to detect a black hole on such a long baseline. The reason is that there has to be a lot of brightness on very small scales. Also, we’d have to wait a long time before the Moon data could get back to us! We hope to image other galaxies with the EHT – astronomers want to image the high speed jets launched by black holes and those show up in a number of sources. THe EHT will provide the sharpest view of these as well.

Was there any particular technologies or engineering challenges you and your team are proud of achieving to bring this project about?

[Shep Doeleman] Great question. In order to pull this off we had to create new systems that could record radio signal at ultra-high speeds in real time. The EHT works by then bringing these recordings together (later) the same way an optical mirror focuses light. Developing these systems has benefited enormously from Moore’s law, so our instruments have leaped ahead in capability, and the increased sensitivity that comes from recording more data was the key to making this work. So the international team is proud of these new purpose-built systems, and going to high altitudes (and being oxygen starved) to get them going.

Would you consider, in your opinion, a black hole to be a 4-dimensional hole in space, or just an infinitely dense 3-dimensional object?

[Dimitrios Psaltis] Even though there are many theories in which we live in a universe with more than 3 spatial dimensions, our working hypothesis for now is that a black hole is an object in 3 dimensions with a central singularity of practically infinite density. Of course, space and time are intimately coupled in Einstein’s theory, which is why we usually talk about the four-dimensional spacetime in this case.

6PB is a lot of storage! How do you guys manage that much data?

[Shep Doeleman] We have custom enclosures that house 16 disk drives, each at ~6TB, so a total of 96TB. We stripe the data across the drives and then beat any internet by shipping the filled disks via airplane back to a central location for processing.

What is your guys’ and gals’ opinion on Hawkings latest statement that a black hole is a portal to another universe?

[Dimitrios Psaltis] Stephen Hawking has been trying for years to find ways to resolve the information paradox by either tweaking the way that quantum field theory is combined with general relativity or by changing the way classical black holes look. If the correct answer is closer to the latter, then the Event Horizon Telescope will offer a unique look into it.

What did you think about interstellar depiction of a blackhole? And if you had to take a wild guess what other mysteries do you think blackholes hold (beyond the event horizon, in larger context of space etc)?

[Dimitrios Psaltis] The movie Interstellar used very sophisticated tools to calculate images of black holes. However, for cinematographic reasons, they made choices that look better for the movie but are not realistic for astrophysical black holes. For a more realistic view, have a look at this link and at a recent Scientific American article that discuss this.

I watch a lot of astronomy/physics documentaries, do you have any recommendation of an informative documentary?

[Shep Doeleman] I love the old COSMOS and the new one as well!

Based on information I have gathered, it is said that black holes actually grow in size when absorbing matter (or fusing with other black holes), is this all observed or mostly theory?

[Shep Doeleman] The event horizon of BHs are predicted to grow as they eat more matter. It’s hard to observe this for a single black hole (the time scales are too long), so astronomers do population studies across many black holes to see how big they get over time. It’s like an alien looking at the whole population on Earth (babies to older specimens) to study how people age – instead of watching one person grow old.

I’ve recently started to consider Astronomy as a potential field of study. What sort of advice would you give to aspiring astronomers?

[Avery Broderick] Keep looking up! For me astronomy and astrophysics is exciting in large part because I get to take flights of fancy with computers and telescopes to the far reaches of the universe. However, key to getting through the large, often exciting, sometime tedious, process of becoming a full-fledged astronomer is to maintain your original enthusiasm. Best of luck!

Why are you trying to image this particular black hole, Sagittarius A*? Were other black holes considered? Depending on the results will you attempt to image other black holes? Would it possible to image the black hole in the center of the Milky Way galaxy? What is the furthest black hole that could be imaged with this technique and equipment?

[Dimitrios Psaltis] Sagittarius A* is the black hole in the center of the Milky Way. Of all the known black holes in the Universe it is the one that is going to look the largest in the sky. There is a second black hole, the one in the center of the M87 galaxy, which is a good target. Even though it is 1000 times further away, it is also 1000 times more massive that Sagitarius A*, making it about the same size in the sky. As more massive black holes are found with other astronomical surveys, the number of targets will continue to increase.

This is a little off topic, but do you think there is a chance that some particles appearing and disappearing seemingly out of nowhere in particule physics can be in fact the observable part of a multidimensional object crossing our dimension, just like a 3d cube’s visible part while crossing a plane world would be only it’s perimeter intercrossing that plane?

[Avery Broderick] No worries! Vacuum fluctuations are a fundamental, and experimentally verified, prediction of quantum field theory. They happen everywhere all the time, so its hard to imagine that these are due to some passing higher-dimensional object. Moreover, they are virtual until some interaction pulls them out of the vacuum, so “particles appearing and disappearing” is a little to concrete. How such a passing higher-dimensional object would generate virtual particles is not obvious to me, at least.

Could you explain how a naked singularity works? I have never seen an intuitive explanation for how a singularity could exist without having an event horizon, and would very much appreciate it if you could provide one.

[Dimitrios Psaltis] A singularity is a region in spacetime where density is so high that modern physics breaks down. We think that there is a singularity behind the event horizon of all astrophysical black holes. Einstein’s theory also allow for the possibility that singularities exist without event horizons surrounding them. These will be “naked singularities”. Note, however, that as of now, we know of no way to make such naked singularities in astrophysics.

Beside needing an understandably absurd amount of storage, what are the other technical computing hurdles you face?

[Shep Doeleman] Excellent question. We also need to record the data quickly! The challenge is that we use relatively small radio dishes because we observe at high frequencies – and the higher the frequency, the smoother the dish has to be (and it’s hard to make large, smooth dishes). So to compensate for the small dishes we need to record huge bandwith (slices of the radio spectrum). So a big hurdle was to develop special systems that can digest data quickly.

What does your team hope to gain out of getting pictures of blackholes?

[Avery Broderick] There are two major pay-offs we are looking forward to:

  1. The first high-precision tests of general relativity in the strong gravity regime. To date GR has been well tested (and enormously successful) where it makes small corrections to Newtonian gravity. However, the EHT will extend these tests to where GR is 100% different.
  2. Black holes + stuff are the brightest things in the universe! As matter falls into the black hole enormous amounts of gravitational potential energy is released, powering quasars and relativistic jets that can impact the galaxies they live within. How this process works depends on how material moves around the horizon, which the EHT is uniquely capable of probing directly.

Presumably you are using custom software for collating, tracking and interpreting the data. What languages/tooling do you use?

[Shep Doeleman] We typically use Off the Shelf hardware at this point, and develop custom code. At the moment we use C for the number crunching in the cluster that compares data from around the globe, and a variety of languages for the imaging, calibration, modeling: python, C…

How much funding does this project receive? I’m interested in how a project of this undertaking to find out what a black hole looks like is funded.

[Shep Doeleman] Not enough! The cool thing about the EHT is that we put specialized equipment at existing telescopes to form an entirely new type of telescope (as big as the earth). So in that sense it is huge ‘bang for the buck’. We just got a very nice grant from the NSF for $6.5M that covers the build-out (on telescopes that cost over $1Billion) of our EHT equipment. What we need at this stage are resources to speed up data processing and analysis to answer to bigquestions: was Einstein right? How to BH’s feed, and how do they affect their host galaxies.

How much detail will the images provide?

[Avery Broderick] Nominally the EHT will put 5 pixels across the shadow. However, that underestimates the information by many orders of magnitude!

In practice, there are many pixels tiling the image, observed in many polarizations, observed at multiple frequencies, and over many time spans ranging from seconds to decades. Since the region around a black hole is typically very dynamic, we will have many different snapshots, which all together allow us to study gravity and black hole astrophysics in unprecedented detail!

Is Sagittarius A currently in an active or dormant phase? If dormant, do we know when it may begin absorbing material again?

[Dimitrios Psaltis] Sagittarius A* is active, in a sense that it is accreting matter from its surrounding, but it is also one of the least bright accreting black holes in the Universe because it is not accreting fast enough. We expect that this state will not change significantly during the next several years.

Do you expect to see interactions between the supermassive black hole and dark matter/dark energy? If so do you have have hypothesis what these interactions might be?

[Dimitrios Psaltis] We believe that dark matter exists because it pushes and pulls gravitationally other astrophysical objects that we can observe. In the same way, it is very likely that black holes in the centers of galaxies will also feel the presence of dark matter around them. Even though these interactions make long term changes to the positions of the black holes in the galaxies, they will be negligible for the short duration of the Event Horizon Telescope observations.

Is it possible to not feel like a badass saying “Event Horizon”?

[Shep Doeleman] No. This is an exciting project – we go to the top of extinct volcanoes with telescopes on top and we lug atomic clocks with us! All we need now are bullwhips and it’s Indiana Jones.

What methods are you using to analyze the data?

[Shep Doeleman] We use a technique called VLBI: capturing data at radio telescopes that are looking at the same black hole at the same time. We then compare the data from different sites in the same way an optical mirror combines data from a large mirror at the focal point. We play the data back in a large computing cluster (many servers linked together) to do this comparison. The trick is to make an image out of an sparse data set (we don’t have telescope all over the globe).

What are the furthest away and the closest black holes you’ve identified? Largest and smallest?

[Avery Broderick] The closest black hole candidate would be a stellar mass X-ray binary. However, these are a little different than the objects we are hoping to image — the objects the EHT is after are supermassive black holes that are millions to many billions of times more massive than a star. The closest of these is Sgr A*, the 4 millions solar mass black hole at the center of the Milky Way.

The farthest black hole candidate is the quasar ULAS J1120+0641, which has a redshift of 7.09, which corresponds to a distance of 8.85 Gpc for the standard cosmological parameters. That’s pretty close to the edge of the visible universe!

As a sophomore undergraduate philosophy major, I’ve been taking more and more space time and quantum based classes as of late. Is there any hope for a philosophy major to make it into an astrophysics grad school program, or is preference usually shown for physics majors?

[Shep Doeleman] Many roads lead to astronomy! When I left undergrad I spent an entire year in Antarctica doing a lot of cool(!) research, only some of which was astro-related. It was one of the highlights of my career. I then found ways to use that experience as I moved to astronomy. I also had a friend in graduate school who was a philosophy major and then focused on astronomy – she’s very very smart and now teaches at Columbia University – look her up: Janna Levin.

Is there a known minimum or maximum to the size a black hole can be? Or, if not known, a theoretical one?

[Dimitrios Psaltis] In principle, a black hole can have any size, from the subatomic to the size of the Universe. However, we only know of a few astrophysical ways to make black holes, either at the end of the life of a massive star, or during the formation and evolution of a galaxy. The first mechanism produces black holes that are more massive than about two times the mass of the Sun; the second mechanism produces black holes less massive than about 10 billion times the mass of the Sun.

Other than being interesting for the sake of documenting the event horizon in a composite image, what information are you hoping to glean from this project? Is there something in particular you hope an image can provide? Additionally, as often happens in astronomy related projects, say you learn something you weren’t expecting, any idea what you might find that you aren’t specifically looking for?

[Avery Broderick] Some of what we hope to find is described in

https://www.reddit.com/r/IAmA/comments/3ihgbr/were_scientists_on_the_event_horizon_telescope/cugisif.

To go a bit further, we will generate not only a single image but many images at many times, many wavelengths, in many polarizations, and for multiple objects. With that kind of information we can get at how gravity works near these objects and how they interact with the surrounding material. For example, general relativity makes strong predictions for the structure of the spacetime that we can directly test.

Of course, the unexpected things are the most exciting! As a voyage of discovery the EHT is avidly looking for these. An example would be to find that the shadow cast by the horizon is a factor of two too large, posing existential challenges to general relativity!

So I place an object on a counter or wall and it sits/hangs there for 30+ mins to multiple hours or multiple days. Then at one moment, without anything touching it or getting near it, it falls down (off the wall or counter). What caused this object to suddenly fall or slip, etc? Was it a variance in gravity or gravitational force? For instance, did gravity get very slightly stronger for a millisecond?

[Avery Broderick] Eventually gravity always wins. But gravity is in it for the long haul, and every other force can hold it at bay. The small fluctuations in the large number of other possible forces, e.g, a slight breeze, a catastrophic breaking of weak molecular bonds, minor tremors too small to be felt, are almost certainly responsible for your falling object.

Having said that, there are terrestrial experiments to probe gravity on millimeter scales that spend a tremendous amount of effort to eliminate non-uniform and time-varying gravitational perturbations caused by stuff in the vicinity of the lab!

What is the most mind blowing aspect of your job?

[Shep Doeleman] For me (your mileage may vary) it’s the adventure aspect: traveling by planes, trains, automobiles and schlepping equipment to the tops of mountains in the middle of nowhere. Then you successfully link up such sites around the world, and it all works, which still seems like magic. There’s also an incredible team element – everyone working together. While we are at one telescope, we know that colleagues ‘have our back’ at the other sites.

Are white holes possible? And can you explain why?

[Avery Broderick] Yes and no. There is no observation to date that excludes their existence somewhere in the universe at the moment. However, if general relativity is correct a white hole cannot form dynamically from other stuff (e.g., by the collapse of a star). So if they are not already there, they aren’t going to be there.

Could black holes occur in pairs, orbiting around a common center of gravity?

[Dimitrios Psaltis] Yes, we believe that black holes in the centers of colliding galaxies eventually come close to each other, orbit for some time, and then collide! When black holes collide, they generate intense gravitational radiation, which we hope to observe in the near future with gravitational wave detectors.

Since you mentioned it, I am curious about your decision to use HGST Helium drives. Any reason?

[Shep Doeleman] The HGST drives are helium filled and hermetically sealed. When we tried to use conventional drives the low air pressure at altitude (15,000ft) caused head crashes and most drives failed. That’s the main reason to use these new He drives, though the generally higher capacity is also very useful (fewer drives to ship back home).

What your team thought on Interstellar and their seemingly fictional science on black holes?

[Dimitrios Psaltis] The movie Interstellar used very sophisticated tools to calculate images of black holes. However, for cinematographic reasons, they made choices that look better for the movie but are not realistic for astrophysical black holes. For a more realistic view, have a look at this link and at a recent Scientific American article that discuss this.

Why are super massive black holes located at the center of galaxies?

[Shep Doeleman] Great question and one that many astronomers are working on. Currently we believe the galaxies and black holes evolve together with black holes affecting how galaxies grow, and then the black holes growing as galaxies merge.

How do you know what to look for in that HUGE data? What programming language/framework do you use for analyzing data? Are there plans to host a subset of data/project in GitHub so others can learn/contribute?

[Shep Doeleman] We record data at many sites, but the final data we use comes when we compare the data from these sites. So really we are looking only at the data that ‘correlates’ or is ‘in common’ between two widely separated sites. After the comparison (equivalent to light across a lens combining at the focus) the total data volume is greatly decreased (by more than a factor of a million). We typically use C and python. We are thinking of ways for more people to contribute – stay tuned.

Does research into black holes have any pragmatic function for society outside of expanding our understanding of our universe?

[Shep Doeleman] Great question. It’s instructive to ask what answer Einstein might have given to this same question asked about his theory of General Relativity (GR), developed 100 years ago. At that time, there was no space travel, and almost no way to detect any difference between Newtonian and Einsteinian gravity. I don’t think he would have been able to come up with anything ‘pragmatic’. But of course today we use GR effects everyday: GPS wouldn’t work without corrections derived from GR. So the answer is that the payoff from BH research may not be evident now, but if we reveal something fundamental about gravity, the consequences in the future could have high impact. Fundamental science is usually a solid investment.

What is the resolution of the images you are receiving and can it be easily livestreamed?

[Shep Doeleman] Great Question. The EHT technique requires us to wait until all the data recorded at different sites around the globe come together at a central facility (right now at MIT). So it’s the ultimate in delayed gratification – sometimes we have to wait weeks or months until we know things worked out, so no livestreaming. We are aiming to achieve ~20 micro arcseconds, or about the apparent size of a baseball on the moon.

What percentage of your replies are guesses?

[Dimitrios Psaltis] Before the Event Horizon Telescope takes the actual picture of a black hole, we are mostly making informed guesses, based on many decades of studying and analyzing non-imaging data. This is why we are building the Event Horizon Telescope: to go from informed guesses to hard data!

When processing data, what is your performance bottleneck – CPU or storage?

[Shep Doeleman] The storage issue is largely solved for the moment – disk drives are getting faster and higher capacity all the time. We are focused now on the computing cluster which compares data recorded at different places around the world. Currently we can process data at only a fraction of real time and we need to speed that up to efficiently churn through the data. So CPU’s on the ‘correlator’ is a bottleneck now that we hope to address with more resources (anyone have a spare cluster in their attic???).

For a project like this, how do you keep the data safe? I.e. do you use RAID and/or off-site or cloud backup?

[Shep Doeleman] At the moment we have only single copies, but striped across many drives, so if one drive dies, we can recover most of the data.

What is the advantage of using helium-filled hard drives instead of SSDs? Cost?

[Shep Doeleman] Helium drives are sealed, so they are robust at the high altitude sites we use (~15,000ft).

Where are your black hole candidates located? Are you looking at supermassive ones like the center of the milky way or “smaller” ones scattered throughout? Or both?

[Dimitrios Psaltis] We are primarily looking at two black holes: Sagittarius A, which is the black hole in the center of the Milky Way, and the black hole in the center of the M87 galaxy, which is 1000 times further away but is also 1000 times more massive than Sagitarius A, making it about the same size in the sky. The smaller black holes scattered in our Galaxy are too small to be resolved

What happens in or at the event horizon?

[Avery Broderick] Five years ago the answer was straightforward: absolutely nothing (except for the angry emails from everyone who isn’t getting your texts).

However, over the past few years it has become clear that unifying general relativity (which describes the large and energetic) and quantum mechanics (which describes the small and cold) may very well manifest quantum gravity effects at the horizon, not down near the singularity. All kinds of things have been posited, and while we hope that the EHT might have something to say about this very important problem in theoretical physics, but most proposals are designed to be very, very close to a general relativistic black hole outside the horizon.

It seems blackhole has a limit on its range of influence. Would it be possible to create a mini blackhole to study it?

[Dimitrios Psaltis] In principle, a black hole can have any size, from the subatomic to the size of the Universe. However, we only know of a few astrophysical ways to make black holes, either at the end of the life of a massive star, or during the formation and evolution of a galaxy. The first mechanism produces black holes that are more massive than about two times the mass of the Sun; the second mechanism produces black holes less massive than about 10 billion times the mass of the Sun.

Looking at your link to VLBI, local timing is kept with an atomic clock, but how are those clock synchronized and how accurate does that sync need to be? What is the effect of crappy timing (Loss in resolution? Signal? Contrast?). Is it obvious to you when some data are out of sync?

[Shep Doeleman] We use Hydrogen Maser clocks that are accurate to 1 sec in 100 million years. That stability is required because we compare radio signals from a black hole received in, say, Chile, with signals from the same black hole received in Hawaii (with both sites looking at the black hole at the same time). Without stable clocks to control the data capture, the two two signals would jitter back and forth too much and any similarity between the two signals would be washed out. We synchronize using GPS and get to within 1 microsecond, which is close enough to get us started. We do a search over different delays between the signals and lock onto the exact synchronization that way.