Angry Astronaut and Nextbigfuture commenter are making the case that SpaceX and Elon Musk must switch to nuclear thermal rockets to colonize Mars. I will review that the nuclear thermal rocket program is taking and will take far longer. Also, triple the ISP does not cut the travel time by one third. There is no discussion by supporters of nuclear thermal like Angry Astronaut about the cost completed systems that match or get close to Starship thrust or even the cost of a production system. There is no complete analysis of scaling up the supply chain. Angry Astronaut and NASA missed the problem of no facilities to test even the DRACO demonstration rocket.
Angry Astronaut describes a hypothetical nuclear thermal rocket assembled in orbit. He assumes a nuclear thermal rocket will be working by the end of 2030. He says the SpaceX approach to building a lot of rockets and doing 11 launches for refueling is more complicated than getting a nuclear thermal rocket working. Also, his plan is only for one nuclear thermal rocket and not mass producing them by the thousands as planned by SpaceX. Angry Astronaut has made his own crude graphics of his proposal for nuclear thermal rockets. This is not coming from Lockheed Martin, NASA, DARPA, General Atomics or BWXT. None of the agencies or companies involved are pushing to rapidly scale nuclear thermal rocket technology.
Basically, Angry Astronaut is saying everything will work so much better and safer if we pull nuclear Excalibur from the stone. Pulling nuclear Excalibur is clearly easier than building and mass producing thousands of larger chemical rockets. He does not mention it but maybe we could also make nuclear passenger jets to take people anywhere on earth in ten minutes instead of using 30,000 commercial passenger jets. Nuclear has better ISP and thrust specs so we should use it everywhere.
Jim Shoemaker, DARPA’s second DRACO program manager, says once a DRACO demonstration proves successful, it could take another 10-15 years before the technology is used on an operational basis. So youtuber Angry Astronaut says we should depend upon a nuclear thermal rocket that is not yet into its full design phase and has no actual specs (ISP or thrust). However, one of the program managers says it will 10-15 years after a demo in 2028 or so before nuclear thermal rockets could first be operationally used. This means 2028-2033. Rocket programs that depend upon completely new technology and that technology has to be scaled in size (millions of newtons of thrust or clusters of 250,000 newton engines) and mass produced by the thousands are far more risky.
A nuclear thermal rocket could potentially achieve double or triple the Delta-V of a SpaceX Starship with just the full orbital refueling. But to move large payloads of 100-200 tons, then the thrust must be increased far beyond the potential demonstration. There will also need to be larger hydrogen tanks.
NASA, Defense Advanced Research Projects Agency (DARPA) and Lockheed Martin (prime contractor) are design, build, and testing of NASA and DARPA’s nuclear-powered rocket demonstration, in collaboration with other industry partners. The Demonstration Rocket for Agile Cislunar Operations (DRACO) program will test a nuclear-powered rocket in space as soon as 2027. However, the 2027 launch date for the Demonstration Rocket for Agile Cislunar Operations (DRACO) is on indefinite hold.
UPDATE: Angry Astronaut also realized that the nuclear thermal rocket project has been delayed with lack of test facilities.
The DRACO program has not yet entered the implementation phase as originally planned. According to NASA’s FY 2025 Budget Estimate document, the project aimed to begin the implementation phase in September 2024. However, this target has been missed, and the program is still in earlier stages of development. A cold flow ground test of the reactor is planned for 2025 by Lockheed Martin and BWXT. The program’s design phase started two years ago and the DARPA-NASA management team has encountered the challenges inherent in sending a nuclear reactor into space for the first time in more than 60 years.
The DRACO team is working to maximize ground-based component and subsystem testing that can be done with existing capabilities, a subset of which will take place at Marshall Space Flight Center.
Even after the demonstration, DRACO’s challenges will not be behind it. Long-term storage of cryogenic hydrogen for the follow-on propulsion system remains a key challenge for the scientific community.
The engine would consist of a 1-m-long (3.2-ft.), ultra-high-temperature, high-assay low-enriched uranium-fueled, flow-through nuclear reactor. The fission reaction is harnessed to generate heat to energize a helium gas propellant, which is exhausted to produce thrust. A follow-on operational engine would replace helium with more energetic liquid hydrogen fuel.
The DRACO team has not signed off on a design for BWXT’s reactor. The reactor is at a “PDR level of maturity,” DARPA says it is examining design refinements meant to improve ground processing safety and enhance on-orbit data collection.
BWXT hasn’t recently completed any new reactors, the company has a long history of manufacturing nuclear components. BWXT has been manufacturing naval nuclear components and reactors since the 1950s, including components for the USS Nautilus, the world’s first nuclear-powered submarine. The company continues to manufacture reactor components for U.S. Naval submarines and aircraft carriers.
NASA is committing up to $499 million toward the DRACO partnership. This includes up to $250 million in costs for the design and development agreement for the nuclear-powered engine, as well as technical oversight and expertise from agency personnel. The U.S. Space Force will provide the DRACO launch and launch site support.
We do not know the exact thrust value for the General Atomics DRACO engine in 2025. It will probably be in the range of 100,000 to 250,000 newtons. Most of it has not been built yet. It will likely have 800-900 ISP in the first version.
How Fast Can SpaceX Starship Reach Mars -One Way?
We can calculate the travel times for the SpaceX Starship to reach Mars. It is relatively easy to get 90 day trips each way with SpaceX Starship. This is faster than the usual 180-270 one-way travel times. This can be faster because we will have a lot more fuel to enable more direct routes to Mars. We could catch up Mars in 1/6th of an orbit instead of half of an orbit around the Sun.
There are ways to use extra expandable Starship tankers that fly with the main Starship and then transfer the extra fuel for deceleration from higher speed.
If there is more things built and working in orbit around the Earth, then this can be used to enable more ways to save fuel for faster or bigger missions. This can be done with reusable tugs to move a fully fueled Mars bound ship to higher orbits or even to escape velocity.
There is online calculator for the Lambert’s Targeting Problem (LTP) to produce launch and arrival v-infinity pork-chop plots between solar system targets selected by the user.
Scott Manley explains how these calculations work.
In terms of drastically reducing the time in days to get from Earth to Mars then it helps a lot to be able to go faster. The SpaceX Starship can enable a lot of extra fuel and more powerful chemical engines. Extra fuel can come from orbital refueling of the Starship. Other spacecraft and the Starship might only have 10% of the starting fuel because they used most of it getting to orbit.
The performance of the SpaceX Raptor engines is already very good but SpaceX is working on better engines. LEET 1337 will have even higher chamber pressure which will enable more thrust. The SpaceX LEET 1337 engines will be simpler, lighter and cheaper. SpaceX will likely be able to build them at ten times the production volume from the same sized factory that will now make 4000 Raptor engines each year.
SpaceX is also developing improvements to Starship that could enable larger fuel tanks.
The travel time and fuel calculations are from a low earth orbit refueling location. This assumption and situation could be changed IF there was a reusable tug that moved the SpaceX Mars ship to near Earth escape velocity. This would let the Mars bound Starship get to a higher speed without using on board fuel. This would save fuel for deceleration and landing.
A lot of refueling and creating a reusable tug are things that can be done when the cost of ships and fuel to orbit are very low. SpaceX and Elon want to get the cost of fully reusable Starship down to less than $10 million for the upper stage. The mass produced rocket engines are targeting $250,000 each while the older Shuttle era RS25 engines cost about $100 million each. The methane fuel can be made from abundant natural gas or can be made in large quantities using solar power and factories using materials in Earth and Mars atmosphere.
The main game changers are the making the ship and engines 100 to 1000 times cheaper and fully reusable. Things that would be wasteful or too costly become affordable.
There is more fuel that is needed to slowdown once a ship gets to Mars. There are ways to use the Martian atmosphere to aerobrake to land on Mars using no fuel. There is a maximum speed for aerobraking to work.
Background
The solution method is based on the works:
Approximate Analytical Solution of the Lambert’s Targeting Problem. Claudio Bombardelli, Juan Luis Gonzalo, Javier Roa. In Journal of Guidance, Control and Dynamics, Volume 41, Issue 3, 2018, pp. 792-801. https://doi.org/10.2514/1.G002887.
Approximate Analytical Solution of the Lambert’s Targeting Problem. Claudio Bombardelli, Javier Roa, Juan Luis Gonzalo. Paper AAS 16-212 in 26th AAS/AIAA Space Flight Mechanics Meeting, Napa, CA, USA, 14-18 February 2016.
Solar system bodies move on Keplerian orbits (all perturbations are neglected) and transfer arcs are Keplerian.
There are errors of up to 15-20% because it is computation of very inefficient transfer arcs (i.e. far from minimum delta-V conditions).
SpaceX Starship in 45 Days to Mars One Way
People believe that exotic new propulsion systems are needed to reduce the one way trip times from Earth to Mars from 180-270 days down to 45 days each way. The slower mission times are for chemical rockets where we barely get out of Earth orbit with a small rocket engine. SpaceX Starship can refuel after reaching orbit to enable faster orbits (straighter and less looping paths) to go to Mars. This makes 90 day times each way easy with chemical Starship and even more wasteful but still chemical rockets to Mars in 45 days each way.
This is calculated by Ozan Bellik.
In 2033 there are opportunities to do a high thrust ~45 day outbound transit with a ~10.5km/s TMI (trans Mars injection). If you refill in an elliptical orbit that’s at LEO+2.5-3km/s then the TMI burn requirement goes down to 7.5-8km/s. A SpaceX Starship with 1200 tons of fuel should be able to do with roughly 150 tons of burnout mass. This is enough for ship, residuals, and a crew cabin with enough consumables to last a moderately sized crew for the 45 day transit. The trouble is that once you get there, you are approaching Mars at ~15km/s.
Related Background
SpaceX Starship version 3 will be 150 meters tall. The current Super Heavy Starship is 120 meters tall (397 feet). It wil go from as tall as a 36 story building to 45 stories.
SpaceX is expanding the factory to make more Starships to make one every 72 hours.
SpaceX will scale to hundreds of Starship built per year by 2026.
Details on the transfer orbits between Earth and Mars.
Continuing the Details of Getting to Mars in 45 Days With Starships
One way to solve this would be to powerbrake to 8.5km/s and aerobrake from there. This needs an additional ~6.5km/s, and the most straightforward way to solve that is to send a fleet of expendable tankers alongside the crew ship and refill en route. This would call for 5 or more fully filled 2000 ton tankers in elliptical orbit. SpaceX could do it.
Alright, here we go.
In 2033 there are opportunities to do a high thrust ~45 day outbound transit with a ~10.5km/s TMI.
If you refill in an elliptical orbit that's at LEO+2.5-3km/s, your TMI burn requirement goes down to 7.5-8km/s, which a Starship w/ 1200t of prop should be… https://t.co/Rf2dDOOzwn— Ozan Bellik (@BellikOzan) November 28, 2023
How can you refill an empty Starship either in Mars orbit or on the surface. Make the methane out of the Mars atmosphere or bring the return propellant from Earth. A slower and more fuel efficient 2000 tanker tanker topped off in elliptical Earth orbit can bring about 1200 tons of that fully propulsively to an elliptical Mars orbit.
A 7.5km/s burn from elliptical Mars orbit will top out at around a 60 day return time in our pretty optimal 2035 return window. We again need extra fuel to brake. Starship again must slow down and then aerobrake (15-16km/s vs. 12.5km/s ITS baseline). The same tanker fleet trick as for the outbound could credibly solve the return speed problem, and technically 45-day outbound + 60-day inbound does satisfy the 45-day transit is achievable with Starship claim.
More vehicle staging from elliptical Earth orbit for the outbound, and a cascade of tankers to deliver the requisite prop to HEO and HMO. This would enable a fully chemical orbital transfer vehicle setup that can do 45-day transit in both directions in the ’33-’35 mission window with a total of 25kt of payload to LEO and less then a dozen expended Starships.
SpaceX Starship has many ways with more ordinary refueling to enable 8.5 km/s delta-v which is far more than the 3.8-4.5 km/s for the low fuel and slower Hohmann orbital transfers that take 180 days. Starship with faster more elliptical orbits can regularly do 90 days each way from Earth to Mars.
Brian Wang is a Futurist Thought Leader and a popular Science blogger with 1 million readers per month. His blog Nextbigfuture.com is ranked #1 Science News Blog. It covers many disruptive technology and trends including Space, Robotics, Artificial Intelligence, Medicine, Anti-aging Biotechnology, and Nanotechnology.
Known for identifying cutting edge technologies, he is currently a Co-Founder of a startup and fundraiser for high potential early-stage companies. He is the Head of Research for Allocations for deep technology investments and an Angel Investor at Space Angels.
A frequent speaker at corporations, he has been a TEDx speaker, a Singularity University speaker and guest at numerous interviews for radio and podcasts. He is open to public speaking and advising engagements.
Zubrin made an interesting point in his latest book: if you take that six-month trip to Mars, and you aren’t able to stop at Mars because something broke, then in two years you’ll coast back to Earth. As long as you have two years of supplies, you can be rescued pretty easily. With any faster trajectory you’ll die somewhere in deep space. You might cross Earth’s orbit again but the Earth won’t be there where you cross.
So he argued that we should stick with the slow trip, and if we get better rockets, use them to just carry more cargo instead of going faster.
For a while at least, seems like maybe a good idea.
That’s my own view, though chiefly because I don’t think the time gained is worth the dramatic reduction in payload. Zubrin does have a good point there, though.
My understanding is that nobody has the thrust to weight ratio of a nuclear thermal rocket high enough for take off from the earth’s surface, so NTRs would be for getting from low earth orbit to low Mars orbit. For that purpose it looks like a nuclear reactor powering some sort of electric drive could be as good or better.
See this proposal to combine the kilopower reactor with a VASIMIR drive.
https://phys.org/news/2024-12-strategic-alliance-high-energy-nuclear.amp
Why not just use water as the working fluid for a nuclear thermal propulsion system?
No cryogenic boiloff issues, and water is freely available just about everywhere in space.
Water will still give you an ISP twice as good as Lox/Hydrogen.
Would be nice if you could run a solid fuel element white hot over 2,200 deg-C (~4,000 F) and just completely dissociate the water, but even the most refractory materials (e.g. tungsten, UO2) are no stiffer than gummy worms at the needed temperature. It is as if the basic mechanical properties bestowed on solid matter by the creator are related to valence electron binding energy and we have a hard upper limit on solid material reality.
Nuclear Thermal Rockets should happen. DRACO underground testing can be done. The boring company can easily make an air-tight underground tunnel and line it with lead so we can test nuclear engines safely. It’s insane to hold up progress over a simple tunnel.
Why not push Starship from LEO to LMO and back with a space nuclear tug (remains in space for multiple missions) for crew flights? The Starship could be mounted to a nuclear tug just as it is to the super heavy booster. If SpaceX can refuel cryogenic CH4 in orbit they could refuel a nuclear tug with Cryo H2.
“The DRACO program has not yet entered the implementation phase as originally planned.” Yeah, you’ll get that on these big pork jobs without a valid mission /s.
“None of the agencies or companies involved are pushing to rapidly scale nuclear thermal rocket technology.” -Yeah, much to the surprise of spacenukebros all over the internet.
Did you forget the sarcasm flag /s when you wrote the following?
“maybe we could also make nuclear passenger jets to take people anywhere on earth in ten minutes instead of using 30,000 commercial passenger jets. Nuclear has better ISP and thrust specs so we should use it everywhere.” What?!?! Do you mean nuclear rockets or jets? I presume you’re inferring suborbital ICBM type travel, but with folks on giant LH2 nuclear rockets?
“BWXT hasn’t recently completed any new reactors, the company has a long history of manufacturing nuclear components.” -Well they are sole supplier for the fuel for every DOD/DOE reactor.
Testing
It works. Now go tell me how a 2:1 thrust-to-weight is good enough and how I don’t understand what could be done with ISP = 800 and I’m a luddite.
Trouble posting too?
Yup.
I can explain why NTP isn’t happening while all the SpaceFans blame a regulatory quagmire overly concerned with safety. But first, to make it OBVIOUS that the DRACO program is going nowhere fast, some quotes from “Nuclear Reactor Test Requirements Put DRACO Launch Plans On Hold Vivienne Machi January 17, 2025” excerpted below:
“
The 2027 launch date for the Demonstration Rocket for Agile Cislunar Operations (DRACO) is on indefinite hold….
We’re bringing two things together—space mission assurance and nuclear safety—and there’s a fair amount of complexity,” he [Sambora] says….
As such, “2027 is not a date that we’re shooting for at this point,” he explains, stressing that an eventual on-orbit demonstration remains a primary goal…
“We are considering ourselves still pre-completed PDR,” he [Sambora] adds, referring to the preliminary design review… he asserts that DARPA and its NASA partners will get there and that DRACO’s mission is not “undoable” but rather “difficult.”
The U.S. has not launched a reactor since the 1960s, an era euphemistically referred to as “the time before safety was invented…
Scientists in the Nuclear Engine for Rocket Vehicle Applications program conducted six ground tests of radioactive reactors in open air between 1964 and 1969, “which we could never get approved to do today,” Shoemaker notes.
“We’re not set into, at this point, any one way of getting this demonstration done,” he says.
“
What is going on?
Well, it is kinda like Climate Change in that space exploration gets a lot of lip service, but nobody really believes in it outside of a small minority that likes to talk about LaGrange points and delta-V. Whether he is a believer or not, Musk eggs these folks on because they are useful for his enterprise. Smart engineers that are so excited by rocket equations, that they play with them are very useful people to have around should any need or commercial opportunity arise. Enough money will be granted for advanced concepts to keep the designers at work; generations of grads excited to work at LockMart or NASA will become seasoned cogs without ever flying DRACO or Brilliant Pebbles, tho there will be many ‘separate effects and partially integrated or resistively heated tests’. Apollo was canceled by 1972 without illuminating any obvious commercial avenues that would support of planting more flags on more dead rocks; there is only a steady demand to place satellites for communications, spying, observation, etc.. Since NTP was already demonstrated in the 1960s, it’s not even theoretical – it’s known to be suboptimal. There are fans, like the MSR fans, that really believe the baggage and tradeoffs associated with NTP are worth it, and the rest of us with basic knowledge just look at these projects like we look at our kids when they ask to wear a smock and paint; we think to ourselves: “Do I really want to let you make a mess when I already know what your fingerpainting is going to look like? Can’t you find something else to do?” Lip service, then delay. I use the tactic with my children. We don’t want to quash the creative thinking – we want to give them hope and keep them working. It’s not about safety – we’re just dragging feet because: why bother? For what? Space rocks? Really. Uh, huh. Oh, live in a can, yes, yes. Mine He3 and extract water and terraform. Yup. We’ll do that.
I *think* I’ve got my comments posting now.
I honestly don’t think the sort of nuclear thermal we’re talking about here is enough of an advance over chemical propulsion to justify the trouble for missions where chemical is adequate. Methalox is good for maybe a hair over 300 seconds, the techniques proposed for DRACO and demonstrated with NERVA might get you to 800 seconds. Given the exponential nature of the rocket equation, that’s not chopped liver, but neither are the political and technical obstacles. If you can do the mission without it, you probably should. For one thing, that way you get to do it in the next decade, rather than long after I’m dead of old age…
Now, if we were talking a nuclear lightbulb rocket at 1,800 seconds, or a nuclear salt water rocket at 6,700 seconds, then you’d be talking. That would be performance worth fighting for, that could enable novel missions, let you launch manned missions to the gas giants. (No accident that the rocket in 2001 A Space Odyssey was nuclear.) Not for 800 seconds.
I don’t think nuclear is going to get much use in space until people living in space are calling the shots, frankly.
Dear Editor, I have a question re this part of your very interesting article; “There are ways to use extra expa(e)ndable Starship tankers that fly with the main Starship and then transfer the extra fuel for deceleration from higher speed.” Could a nuclear powered tug pushing a chemical starship work? The tug would not slow when approaching Mars, it would detach from the Starship and swing around Mars and back to Earth. As it returned to Earth orbit it would slow sufficiently to be able to collect another Starship and then accelerate towards Mars, or park in Earth orbit to await the next transfer window. This woul allow a fully fueled Starship to self-brake when approaching Mars and perhaps have enough remaining fuel for a landing. Or several Starships could do this and then transfer fuel to the lander. Comments?
He’s basically just proposing to increase the amount of fuel in one ship dramatically by having a number of ships set out, and then transfer their remaining fuel to just one of them, and then just be lost at the Mars end of the trip.
It would be more efficient to use a drop tank system if you were doing this; Since you don’t need high acceleration for the transfer orbit insertion, there’s no need for the extra tankage to have its own engines. Just have a bunch of simplified tanks mounted to the one Starship, and run them dry during the insertion, then drop them.
Yes, a nuclear powered tug could work, too, *if you had one*.
As I’ve said before, I don’t think the enormous increase in launch cost just to shave a few months off the trip time is really worth it, when the same effort for the longer trip time gets you a dozen times the payload. The transit time to Mars isn’t totally outrageous if you compare it to the length of voyages during the age of exploration. Now, if you were going to Jupiter, it might be worth it, since the trip time is several years each way.