If we get our way and really put humans on Mars in the coming decades, they’re going to need power. NASA has had concrete plans to send people to the Red Planet since 2010–with target dates in the 2030s, while Elon Musk thinks SpaceX can make it to Mars faster. But no matter who gets there first, the power problem remains. Astronomer Frank Shu has a great idea that could work–a type of nuclear reactor that’s cheaper, safer, and more efficient than the ones currently in wide use.

But Shu, a former president of the American Astronomical Society, actually doesn’t care so much about Mars. He thinks his reactor could do wonders here on Earth–so he’s flipping the space technology script. Usually, scientists say things like, “If we develop a rocket to Mars, who knows what cool earthly spinoffs we’ll develop along the way. Stuff like Tang! You love Tang! Right?” But Shu’s plan does the opposite. He wants to develop technology that helps Earth–and then get Martian spinoffs. Like power systems for the space settlers you’re never going to see again!

The path to plopping new nuclear reactors–or any nuclear power source–on this fragile planet is politically and philosophically fraught, so they require big buy-ins from governments and funding agencies. “Even if you can convince people to do it, it’s probably a long-term thing,” says Shu. To manifest his vision, he needs significant, long-term investment–which is why Shu is promoting his reactors as a solution to Martian power problems. In sum, he says: “I want to build this on Earth, but I want NASA to pay for it.”

Bonus: “Development by NASA could bypass much of the red tape that now ties up the timely testing of novel prototype reactors,” says Shu.

Salt of the Earth

Shu’s nuclear power device is called a “two-fluid molten salt reactor.” The full details are in the patent, but the basic idea is this: The first batch of molten salt is full of a thorium compound, which eventually decays into uranium as neutrons bombard the mixture. That uranium goes into the second batch of molten salt and circulates into the reactor’s graphite-filled core.

There, it encounters slowed-down neutrons, which kick off a fission chain reaction–that’s the energy-producing part. The first batch of salt then absorbs the heat from the reactions, cooling the system. (In typical nuclear reactors, water does the cooling.)

This system is self-regulating: If the reactions happen too fast and the reactor gets too hot, the salt naturally expands out of the core, which cools it off–think transferring hot coffee from a tiny cup to a cookie sheet (don’t ask questions). That makes the reactor pretty meltdown-proof.

That’s great for Mars, obviously–any energy supply on a newly-formed colony is going to have to be pretty foolproof. Not worth setting up a trillion-dollar settlement if your reactor gives everyone radiation poisoning (at least not before the sun does, anyway). Plus, the planet has plentiful thorium, and nuclear power doesn’t care if dust storms dot out the Sun for months at a time (solar panels, on the other hand, care very much). That’s why Shu hopes NASA will back the R&D for his reactors.

It’s also great for Earth. While Shu’s specific design is novel, the idea comes from the 1960s, when Oak Ridge National Laboratory made a molten salt reactor. The World Nuclear Association calls them “a promising technology today,” and China and India are sinking effort into their own designs. That hasn’t happened yet, though–which is where Shu’s plan comes in.

Any new technology must prove that it has the potential to scale safely, while outcompeting other established options. That takes time, and money–time and money that a Mars-tech development plan can provide.

Some technologies–like electricity-makers–make easy sense in Shu’s Earth-first philosophy. People and this planet want things like cleaner energy and clean water. Solar power and water purification, says MIT’s Olivier de Weck, are big markets with big companies behind them. “They’re not specifically working on space problems, but the technology is progressing,” he says. That technology can then transform into space-colony-ready systems.

And Mars may need Earth technologies just as much as Earth needs Mars technologies. Back in 2014, De Weck and his colleagues at MIT made a big crater among the go-to-Mars set when they called out a Dutch plan to send people to Mars, one-way. The so-called Mars One team claimed that “each stage of Mars One mission plan employs existing, validated and available technology.” But the technology they pegged as nearly go for launch simply wasn’t.

Impatient Earthlings

But when will we actually need systems that are space-colony-ready? Just how anticipatory is all this development, and how long is it likely to remain Earth-bound?

“The true answer is it’s always one generation away,” says de Weck.

“Like fusion,” I say.

“It’s kind of like fusion,” he says.

But the limits aren’t strictly technological. “If we really had to establish a Mars colony, we could do this within 10 to 15 years,” he says. It would be expensive, but not impossible.

In the meantime, the most prominent plans for Mars colonization are still sketchy. Elon Musk hopes his first ships will reach Mars around 2022, and last month, he revealed a nebulous version of his blueprints for his later colonization plans–run largely on solar, not nuclear, power–at a presentation.

“Elon’s plan is huge; it’s monolithic; it’s very ambitious,” says de Weck. But Musk basically waves his hands at the architectural and infrastructural details of the colony itself. “I’ve been asking–and others have been asking–to see the roadmap from the current capabilities to the far-out vision. I just don’t see the logical pathway,” he continues.

While Musk fills auditoriums, it may be people like Shu–with their terrestrial technologies and their economically supported stepping stones–who actually make Mars a livable planet instead of somewhere settlers stand around and stare at each other until they die.


  1. And you believe that why?

    I started looking for a couple key words from comments sections on Wikipedia. From there I started looking for other sources to verify the accuracy of Wikipedia. It turned out that there is a lot of other technical information available. Nextbigfuture.com normally has current news(warning, the articles aren’t always easy to read). Key US projects are Oak Ridge National Lab(1960s), Integral Fast Reactor(early 1990s).

    Current work is being done in Russia, China, India, Japan, and the UK. The most important key words are Molten Salt(sodium or fluoride) Reactor and Fast Breeder Reactor. The operational Russian design is the BN-800 using a uranium and plutonium fuel cycle. It’s predecessor was the BN-600 which was designed to use weapons grade uranium and plutonium as fuel. The next version, the BN-1200 is being designed for a thorium fuel cycle. The Chinese, Indian and Japanese efforts are still in the research phase with China hoping to have a commercial reactor in 2030. The UK design is for a commercial plant but it only exists on paper at this point.

    I would post specific links but nuclear fission is politically unacceptable. For example the following Scientific American article does a good job covering the basic reaction concepts. At the same time it make proliferation seem like a major concern. In reality if someone wants to build a nuclear weapon there are many far easily options for obtaining weapons grade materials. Getting the full story requires understanding the author’s bias so dozens of sources are needed. It is safe to assume that most environmental sites are biased against nuclear in any form.


  2. Really? Interesting as hell. Where can one find out more about this?

  3. William Donelson

    The moon has so many advantages for the next 50 years. Tunnels under surface in deep craters, close to earth so less time (much within Van Allen), close to earth in case of supplies and rescue, half the gravity well, etc etc

  4. … and the moon doesn’t have radiation problems? Radiation and micro-meteorites are even worse on the moon than on Mars, since the moon has no atmosphere and Mars at least has a bit of one. Neither have a magnetosphere so it’s really a moot point. And the closer proximity to the sun increases the likelihood of a direct hit from a CME on the moon than on Mars, although given enough time Mars colonists will get fried by CMEs too.

  5. You should let NASA know this.. I’m sure they haven’t even considered it in their designs…

  6. There are several nuke plants in orbit. Read up on it.

  7. I’m intrigued by the concept of MSR’s, but I wish that Wired had supplied some links to diagrams, more detailed descriptions, or technical papers. So many of the articles in the print version include really useful infographics or other supporting material. This article, while tantalizing, seems a little … thin… on the details. Could we at least have a link to Prof. Shu’s work?

  8. Billy, time for you to do some math. Nuclear provides 20% of our power which is also 60% of carbon free sources of energy.
    All technologies (solar and wind) that are dependent on RARE EARTHs have a significant radioactive footprint.
    The cost of the clean up of solar’s toxic legacies are not
    included in your presentation.

  9. Tom Billings

    Easier, and lighter, to use multi-ship squadrons, with cargo ships on the exterior of the squadron’s formation, transmitting plasma magnets around the formation, conducting electricity. With the passenger ships in the interior of the formation none of the primary cosmic rays, or the secondaries they would generate, have any chance to interact with the hulls of the passenger ships.

  10. Tom Billings

    William, I asked for numbers, …plural. You provided 1. What are the the sources and dosage rates, in what circumstances? What are the transit times? What shielding tech are you assuming? That would include the the many kilometer diameter magnetic fields available from plasma magnets for the 80-day squadron transits Musk is proposing. Numbers, William.

  11. gordonmcdowell

    To appreciate how under-communicated the topic of Molten Salt Reactors are, the only motion-film footage of Oak Ridge experiments on Molten Sat Reactors (which peaked with “The Molten Salt Reactor Experiment”) were edited into an extremely technically detailed documentary in 1969… which was then lost for 45 years.

    As someone trying to communicate this technology, the difference between having access to this public domain footage and having absolutely ZERO film footage is a very big deal. Imagine trying to talk about Apollo Missions without any motion-picture footage, only a handful of still photos?

    HERE is ORNL’s 1969 documentary on MSRE. Created in 1969. Finally found & released to the public in 2016. 45 years later.


    My own communication efforts focus (like this article does) on the dual use of MSR for both sustaining civilization on Mars, and to help deliver ample clean energy here on Earth. It is called “THORIUM REMIX 2016 – NASA”. Please check it out on YouTube.

  12. It might be more realistic with hevey radition shielding on the transport ship. Cosmic rays can be insulated against with a few inches of boron and plastic.

  13. Like the article. Serious typo though: “De Weck and his colleagues at made a big crater”. Preposition is usually followed by a noun or article a noun.

  14. Molten salt is the most effective energy storage medium since the higher the temperature the more effective the rate of energy transfer and hence the availability of the stored energy and therefore its entropic advantage.

  15. Dr. de Weck actually had his students get destroyed by industry experts to the point they had to withdraw major portions of their critque’s findings for a 10-year, one-way colonization mission (much like SpaceX’s) because of bombastic laguage and lack of validated simulation results; it made their first submission to AIAA essentially unpublishable. The report was also troubling to people like Dr. Harry W. Jones, who gave the MIT students some friendly advice before publishing a feasibility study in the manner Dr. de Weck allowed them to. A senior systems engineer at NASA Ames, Dr. Jones said, while critques like that MIT study can be important ways to refine ideas, the MIT report could have benefitted from “a basic literature review.” Chief engineer Barry Finger at Paragon called the MIT group results Dr. de Weck is so proud to cite…”nonsensical in their initial conditions and nonsensical in their results.”

  16. Rather than investing $10K in a rooftop solar panel system and write it off in 20 years (at best), why not pay your energy bill and invest the $10K in Tesla? – an introduction to the concept “present value of money”.

  17. Thorium Rocks are the stepping stones to STAR TREK TECH.

  18. Alden Wilner

    Let’s keep in mind that Mars has really, really crappy weather.

  19. Alden Wilner

    But there are huge, planetary-scale dust storms. And there are certainly mineral deposits, likely including the stuff needed to build nuke plants on-planet. What there isn’t? A breathable atmosphere. I don’t know why more people aren’t worried about that little issue.

  20. Alden Wilner

    Batteries needn’t be free. They only have to be cheap enough to be economical. At the same time, power plants that provide peak load don’t need to be cheap, but they do have to be *fast*. Coal and nuclear are base-load power, not peak power. Gas turbine and hydroelectric are peak power. Pumped hydroelectric is also an excellent battery. One can imagine a future in which battery technology never matures and power utilities do little more than provide hydroelectric energy storage. But I suspect batteries will continue to improve.

  21. The fact that Kirk Sorensen, Alvin Weinberg, EBR, EBRII, or the Aircraft Reactor Experiment were never mentioned in this article — just goes to show how educated Wired is.

    This whole article comes off as another – Elon Musk – verbal felatio.

  22. William Donelson

    I am all in favour of starting with the Moon. Travelling to Mars incurs radiation exposures equivalent to 6,000 chest x-rays with the fastest transit technology likely to be available in the next 50 years. Wake up.

  23. Tom Billings

    OK, William, HUGE is not a number. Give us some calculated numbers. Better yet, give us some numbers with data behind the calculations. Musk’s numbers of $10 billion to start, about what several major industrial projects have cost, and $4-500,000 per passenger, coming from the settlers, are quite within reasonable calculable bounds from present data. What are your numbers, and how did you get them? Radiation dosage counts only during exposure. Speed reduces exposure in transit, and using as initial settlement sites the 100-500m diameter lava tube caves already discovered completely ends dosage once on Mars or the Moon.

  24. Use ISIS as fuel.

  25. And send idiot Helen “Chicken Little” Caldicott in the baggage compartment

Leave a Comment

Your email address will not be published. Required fields are marked *