Could a nuclear bomb split an asteroid?

The full story

The episode's narration, word for word. Numbers link to the sources below; timestamps play that moment.

Drill, then bomb 0:00

A rock the size of Texas is heading for Earth ¹. In the film Armageddon, the plan is to land on it, drill an eight-hundred-foot hole, drop a nuclear bomb in, and split it in two, so the halves sail past either side of the planet ^1,2,3. It's a great piece of cinema. But could a nuclear bomb really split an asteroid? And if it could, would you want it to?

A piece the size of Dallas 0:26

Armageddon opened on the first of July 1998, directed by Michael Bay ³. Bruce Willis plays an oil driller, recruited to lead the drilling teams sent up to the asteroid ¹. In the film, NASA has eighteen days ^4,3. The critic Roger Ebert put the physics problem in one line: what if a piece the size of Dallas is left ¹? Hold on to that question, because it turns out to be the right one.

What holds a rock together 0:55

First, what would it take to blow a rock apart for good? Not just crack it, but carry every piece away, because every bit of rock is pulling on every other bit ⁵. Imagine taking the asteroid apart one thin layer at a time, and hauling each layer off against the pull of everything still underneath ⁵. Add up all that lifting and you get its gravitational binding energy ⁵. It grows fast with size: double the width of the same kind of rock, and you need thirty-two times the energy ⁵.

Weighing Texas 1:27

So let's weigh the film's rock. Take it as a ball a thousand kilometres across, made of iron-rich rock, as a team at the University of Leicester generously assumed ². That works out at nearly four billion billion tonnes of rock ². Holding it together takes about a one with twenty-seven zeros after it, in joules ⁵. That number means nothing yet, so let's put a bomb next to it.

The biggest bang 1:55

The most powerful explosion humans have ever made was a Soviet test on the thirtieth of October 1961, over an Arctic test range ⁶. In the West it was nicknamed the Tsar Bomba ⁶. Its yield was fifty megatons, the energy of fifty million tonnes of TNT, and it could in principle have been twice that ⁶. Taking the usual estimates of fifteen and twenty-one kilotons for the Hiroshima and Nagasaki bombs, that works out at about fourteen hundred times the two combined ^6,7.

Five billion bombs 2:30

Fifty megatons is about two times ten to the seventeen joules ^6,8. Divide the asteroid's binding energy by that, and you need about five billion fifty-megaton Tsar Bombas ^6,8,5. Even with the hundred-megaton design, it's more than two and a half billion ^6,8. There's no rounding that rescues the plan.

The Leicester sum 2:52

In 2011, four physics students at the University of Leicester ran the film's numbers in a short paper titled Could Bruce Willis Save the World ². They were generous: they ignored gravity altogether, as if nothing held the rock together ². They only asked how fast each half would need to fly sideways to miss Earth ². Their answer was about eight hundred trillion terajoules: about two billion times the bomb, even at its full hundred-megaton design ².

Sideways, in under three hours 3:24

We redid their sum ². Using their reading of the film, the rock is ten kilometres a second and a hundred thousand kilometres out, so it arrives in under three hours ². In that time each half must swing more than seven thousand kilometres sideways, which means about seven hundred metres a second ². For something that heavy, that's about four billion fifty-megaton Tsar Bombas, and that's before you've paid gravity a cent ^2,6.

No Texas out there 3:55

The astronomer Phil Plait pointed out that an asteroid the size of Texas would be the largest in the solar system, and we'd have seen it decades ahead ⁴. The real champion of the asteroid belt is Ceres, about nine hundred and forty kilometres across, with a quarter of the belt's mass ^9,10. Even Ceres would take about three hundred and sixty million fifty-megaton Tsar Bombas to unbind ^9,6. And of the thousand or so near-Earth objects bigger than a kilometre, NASA says about ninety-five per cent have been found, and none is a current threat ¹¹.

Real rocks are loosely held 4:32

Now flip it. The asteroids that could really hit us are hundreds of metres across, not hundreds of kilometres ¹¹. Bennu, about half a kilometre wide, is held together by gravity worth about a fifth of a tonne of TNT ¹². Dimorphos, a hundred and fifty metres wide, is a rubble pile that, simulations suggest, is glued by almost nothing: a cohesion of less than a few pascals ^13,14. So for a real asteroid, gravity's grip is tiny: a bomb could break it apart easily ^12,13. And that, it turns out, is exactly the trap.

One bullet, many pellets 5:10

Break a rock a day before impact, and the pieces keep almost the same path ⁴. Even drifting apart at walking pace, after a day they'd spread about a hundred and twenty kilometres, a sliver of Earth's width ¹⁵. Plait put it bluntly: instead of one explosion the size of a billion atomic bombs, you get a billion atomic bomb sized explosions ⁴. Only pieces small enough to burn up in the air are harmless ¹¹. Splitting late doesn't remove the danger; it spreads it out ^4,16.

The film that knew 5:43

About eight weeks before Armageddon, Deep Impact told the same story with astronomers in the room ^4,3,17. Its producers hired working astronomers as consultants, and took their advice ⁴. In the film, the crew plant nuclear bombs in the comet, and the blast splits it in two, with both pieces still headed for Earth ^17,18,16. The smaller one hits the Atlantic, and a giant wave swamps the east coast ^17,19. Plait still had quibbles, but he called it simply better ⁴.

Seven minutes 6:15

So what does work? Earth is moving around the Sun at about thirty kilometres a second, so it travels its own width in about seven minutes ¹⁵. That means you don't have to destroy the asteroid at all: roughly speaking, make it arrive a few minutes early or late, and Earth isn't there ¹⁵. Years ahead, that takes only a tiny push ¹⁴. The catch is time: a spacecraft that simply rams the asteroid needs several years of warning, and preferably decades ¹⁴.

DART 6:46

In 2022, NASA tried it ²⁰. Its DART spacecraft, about five hundred and eighty kilograms, launched in November 2021 ^14,20. On the twenty-sixth of September 2022, it flew at six kilometres a second into Dimorphos, the hundred-and-fifty-metre moon of a bigger asteroid, Didymos ^14,21. In eastern Australia, that was the morning of the twenty-seventh ¹⁴. A moon was chosen on purpose: telescopes on Earth could time its orbit before and after ¹⁴.

Thirty-three minutes 7:21

Before the crash, Dimorphos went around Didymos once every eleven hours and fifty-five minutes ²¹. Afterwards, radar and telescopes measured about eleven hours and twenty-two minutes: thirty-three minutes shorter ²¹. NASA's minimum for success had been seventy-three seconds ²². If the spacecraft's push were all that mattered, the change should have been about seven minutes ²¹. It was nearly five times that ²¹.

The kick from the debris 7:53

The extra came from the debris ²³. The crash blasted rock back out of the crater, and that spray shoved the moonlet the other way, like a rocket's exhaust ²³. The shove slowed Dimorphos in its orbit by two point seven millimetres a second, and a slower moon drops onto a smaller, quicker orbit ^21,23. The debris carried more push than the spacecraft did ²³. That's a crawl, but a crawl applied years ahead adds up ¹⁴.

Energy is not the point 8:22

Here's the twist. DART hit with the energy of about two point six tonnes of TNT, roughly a thousand times Dimorphos's binding energy ^14,23. Yet it didn't scatter the moonlet; it nudged it, and probably reshaped it ^13,21. That's because a crash spends its energy near where it hits: the shock heats, crushes and flings out the rock close by, rather than sharing it evenly with every boulder ²⁴. Binding energy is a floor, not a recipe ⁵. What saves you is a push in the right direction, early ^14,23.

Where the bomb comes back 8:59

Nuclear explosives aren't off the table ¹¹. NASA's 2023 strategy says one may be the only option for a very large asteroid, or one found too late ¹¹. In 2024, a team at Sandia National Laboratories zapped mock asteroid material with a huge burst of X-rays from a lab machine ²⁵. The surface boiled off and pushed the rest at about seventy metres a second ²⁵. Scaled up to the energy of a proposed interceptor, they estimate this kind of push could deflect asteroids up to about four kilometres across ²⁵.

Back to the scene 9:37

Europe's Hera spacecraft launched in October 2024 and is due at Didymos in November 2026, to survey the damage ²⁶. Simulations that match DART's results suggest it may find not a neat crater, but a reshaped moon ¹³. That will tell us how much push a rubble pile really gives back ^13,23.

So, could a bomb split it? 10:02

So, could a nuclear bomb split Armageddon's asteroid? Not by billions ^2,6. Could one break a real asteroid? Easily, but a day before impact, that doesn't help, and can make things worse ^4,16. The real fix is a nudge, years ahead, with a spacecraft, or for the biggest and latest threats, perhaps a nuclear burst ^11,14,25. That doesn't make a third act, so films reach for the drill. Next time, another nuclear stunt: could you survive a nuclear blast in a fridge?

Sources

Every factual claim in the episode is tied to one of these. Spotted an error? Tell us.

1. Roger Ebert, "Armageddon" (review), RogerEbert.com, 1 July 1998, archived 2 January 2021 — web.archive.org
2. A. Back, G. Brown, B. Hall and S. Turner, "Could Bruce Willis Save the World?",… — web.archive.org
3. Wikipedia, "Armageddon (1998 film)" (release date, director, plot summary and the 2012… — en.wikipedia.org
4. Phil Plait, "Hollywood Does the Universe Wrong", Space.com, 17 February 2000, archived… — web.archive.org
5. Wikipedia, "Gravitational binding energy" (formula for a uniform sphere and its… — en.wikipedia.org
6. Atomic Heritage Foundation / National Museum of Nuclear Science and History, "Tsar… — ahf.nuclearmuseum.org
7. U.S. Department of Energy, The Manhattan Project: An Interactive History, "The Atomic… — osti.gov
8. NIST, Guide for the Use of the International System of Units (SP 811), Appendix B.8,… — nist.gov
9. NASA JPL Small-Body Database, 1 Ceres: diameter 939.4 km, GM 62.6284 km³/s² (Park et… — ssd.jpl.nasa.gov
10. NASA Science, "Ceres Facts" — science.nasa.gov
11. NASA, "NASA Planetary Defense Strategy and Action Plan", April 2023 — nasa.gov
12. NASA JPL Small-Body Database, 101955 Bennu: diameter 0.48444 km, GM 4.8904 × 10⁻⁹… — ssd.jpl.nasa.gov
13. S. D. Raducan, M. Jutzi, A. F. Cheng et al., "Physical properties of asteroid… — nature.com
14. R. T. Daly, C. M. Ernst, O. S. Barnouin et al., "Successful kinetic impact into an… — nature.com
15. NASA Science, "Facts About Earth" (equatorial diameter 12,756 km; 150 million km from… — science.nasa.gov
16. Phil Plait, "Review: Deep Impact", Bad Astronomy, archived 18 December 2005 ("Blowing… — web.archive.org
17. Wikipedia, "Deep Impact (film)" (release date, director, plot summary; consultants,… — en.wikipedia.org
18. Roger Ebert, "Deep Impact" (review), RogerEbert.com, 8 May 1998, archived 4 January 2021 — web.archive.org
19. Rita Kempley, "'Deep Impact': C'mon Comet!", The Washington Post, 8 May 1998, archived… — web.archive.org
20. NASA Science, "Planetary Defense – DART" mission page — science.nasa.gov
21. C. A. Thomas, S. P. Naidu, P. Scheirich et al., "Orbital period change of Dimorphos… — nature.com
22. NASA, "NASA Confirms DART Mission Impact Changed Asteroid's Motion in Space", news… — nasa.gov
23. A. F. Cheng, H. F. Agrusa, B. W. Barbee et al., "Momentum transfer from the DART… — nature.com
24. Wikipedia, "Impact crater" (hypervelocity impact physics: the shock wave heats, melts… — en.wikipedia.org
25. N. W. Moore, M. Mesh, J. J. Sanchez, M.-A. Schaeuble et al., "Simulation of asteroid… — nature.com
26. European Space Agency, "Hera" mission page — esa.int

Researched and scripted with AI assistance, fact-checked claim by claim, with synthetic narration and diagrams drawn in code. How we make episodes.