Light and waves
April 2019
The first image of a black hole
Event Horizon Telescope Collaboration
How do you build a telescope as big as the Earth?
Radio dishes across the Earth observed together, acting like one telescope the size of the planet. They imaged the glowing ring of gas around the black hole at the centre of galaxy M87, about 55 million light-years away.
- Circles used
- 3 of 256
- Shape captured
- 97.7%
Draw any closed shape on the canvas. It is split into spinning circles (a Fourier series): each circle turns at a whole-number speed, and the biggest ones come first. Watch how few circles you need before the shape appears.
How they figured it out
The sharpness of a telescope depends on its size compared with the wavelength it sees. To see the shadow of the black hole in M87, as small in our sky as an orange on the Moon, you would need a dish the size of the Earth.
So astronomers linked radio dishes around the world, each recording signals stamped by atomic clocks onto hard drives: petabytes of data. The drives from the South Pole had to wait months for the end of the Antarctic winter before they could be flown out. The 2017 observations became the first image of a black hole, released on 10 April 2019.
The equation
Each pair of dishes measures one wave-pattern ingredient of the image; together they rebuild it.
See it in real life
Phones at a concert
Imagine reconstructing a concert stage from photos taken by fans in seats all over the stadium: more seats, sharper picture. Hospital MRI scanners build images from frequency measurements in the same way.
Where engineers use it
Radio astronomy and MRI
Rebuilding images from measured Fourier components is how MRI scanners work too.
Atomic clocks
Each dish recorded data stamped by atomic clocks so it could be combined later.
Where it came from
- The first black hole solution 1916
Led to
The dark centre in the image is the shadow of a black hole, about the size theory predicts.
- Everything is waves 1822
Same technique
Each pair of telescopes measures one Fourier ingredient of the image.
Where the comparison stops: Only a scattered set of ingredients is measured, so rebuilding the picture needs extra assumptions, tested carefully by several independent teams.