Sometimes all it takes to spawn a new idea is a tiny smudge in a telescopic image. What counts, of course, is just what that smudge implies. In the case of the object called ‘Oumuamua, the implication was interstellar, for whatever it was, this smudge was clearly on a hyperbolic orbit, meaning it was just passing through our Solar System. Jim Bickford wanted to see the departing visitor up close, and that was part of the inspiration for a novel propulsion concept.
Now moving into a Phase II study funded by NASA’s Innovative Advanced Concepts office (NIAC), the idea is dubbed Thin-Film Nuclear Engine Rocket (TFINER). Not the world’s most pronounceable acronym, but if the idea works out, that will hardly matter. Working at the Charles Stark Draper Laboratory, a non-profit research and development company in Cambridge MA, Bickford is known to long-time Centauri Dreams readers for his work on naturally occurring antimatter capture in planetary magnetic fields. See Antimatter Acquisition: Harvesting in Space for more on this.
Image: Draper Laboratory’s Jim Bickford, taking a deep space propulsion concept to the next level. Credit: Charles Stark Draper Laboratory.
Harvesting naturally occurring antimatter in space offers some hope of easing one of the biggest problems of such propulsion strategies, namely the difficulty in producing enough antimatter to fuel an engine. With the Thin Film Nuclear Engine Rocket, Bickford again tries to change the game. The notion is to use energetic radioisotopes in thin layers, allowing their natural decay products to propel a spacecraft. The proper substrate, Bickford believes, can control the emission direction, and the sail-like system packs a punch: Velocity changes on the order of 100 kilometers per second using mere kilograms of fuel.
I began this piece talking about ‘Oumuamua, but that’s just for starters. Because if we can create a reliable propulsion system capable of such tantalizing speed. we can start thinking about mission targets as distant as the solar gravitational focus, where extreme magnifications become possible. Because the lensing effect for practical purposes begins at 550 AU and continues with a focal line to infinity, we are looking at a long journey. Bear in mind that Voyager 1, our most distant working spacecraft, has taken almost half a century to reach, as of now, 167 AU. To image more than one planet at the solar lens, we’ll also need a high degree of maneuverability to shift to multiple exoplanetary systems.
Image: This is Figure 3-1 from the Phase 1 report. Caption: Concept for the thin film thrust sheet engine. Alpha particles are selectively emitted in one direction at approximately 5% of the speed of light. Credit: NASA/James Bickford.
So we’re looking at a highly desirable technology if TFINER can be made to work, one that could offer imaging of exoplanets, outer planet probes, and encounters with future interstellar interlopers. Bickford’s Phase 1 work will be extended in the new study to refine the mission design, which will include thruster experiments as well as what the Phase II summary refers to as ‘isotope production paths’ while also considering opportunities for hybrid missions that could include the Oberth ‘solar dive’ maneuver. More on that last item soon, because Bickford will be presenting a paper on the prospect this summer.
Image: Artist concept highlighting the novel approach proposed by the 2025 NIAC awarded selection of the TFINER concept. This is the baseline TFINER configuration used in the system analysis. Credit: NASA/James Bickford.
But let’s drop back to the Phase I study. I’ve been perusing Bickford’s final report. Developing out of Wolfgang Moeckel’s work at NASA Lewis in the 1970s, the TFINER design uses thin sheets of radioactive elements. The solution leverages exhaust velocities for alpha particle decays that can exceed 5 percent of the speed of light. You’ll recall my comment on space sails in the previous post, where we looked at the advantage of inflatable components to make sails more scalable. TFINER is more scalable still, with changes to the amount of fuel and sheet area being key variables.
Let’s begin with a ~10-micron thick Thorium-228 radioisotope film, with each sheet containing three layers, integrating the active radioisotope fuel layer in the middle. Let me quote from the Phase I report on this:
It must be relatively thin to avoid substantial energy losses as the alpha particles exit the sheet. A thin retention film is placed over this to ensure that the residual atoms do not boil off from the structure. Finally, a substrate is added to selectively absorb alpha emission in the forward direction. Since decay processes have no directionality, the substrate absorber produces the differential mass flow and resulting thrust by restricting alpha particle trajectories to one direction.
The TFINER baseline uses 400 meter tethers to connect the payload module. The sheet’s power comes from Thorium-228 decay (alpha decay) — the half-life is 1.9 years. We get a ‘decay cascade chain’ that produces daughter products – four additional alpha emissions result with half-lives ranging from 300 ns to 3 days. The uni-directional thrust is produced thanks to the beryllium absorber (~35-microns thick) that coats one side of the thin film to capture emissions moving forward. Effective thrust is thus channeled out the back.
Note as well the tethers in the illustration, necessary to protect the sensor array and communications component to minimize the radiation dose. Manipulation of the tethers can control trajectory on single-stage missions to deep space targets. Reconfiguring the thrust sheet is what makes the design maneuverable, allowing thrust vectoring, even as longer half-life isotopes can be deployed in the ‘staged’ concept to achieve maximum velocities for extended missions.
Image: This is Figure 7-8 from the report. Caption: Example thrust sheet rotation using tether control. Credit: NASA/James Bickford.
From the Phase I report:
The payload module is connected to the thrust sheet by long tethers. A winch on the payload module can individually pull-in or let-out line to manipulate the sail angle relative to the payload. The thrust sheet angle will rotate the mean thrust vector and operate much like trimming the sail of a boat. Of course, in this case, the sail (sheet) pressure comes from the nuclear exhaust rather than the wind
It’s hard to imagine the degree of maneuverability here being replicated in any existing propulsion strategy, a big payoff if we can learn how to control it:
This approach allows the thrust vector to be rotated relative to the center of mass and velocity vector to steer the spacecraft’s main propulsion system. However, this is likely to require very complex controls especially if the payload orientation also needs to be modified. The maneuvers would all be slow given the long lines lengths and separations involved.
Spacecraft pointing and control is an area as critical as the propulsion system. The Phase I report goes into the above options as well as thrust vectoring through sheets configured as panels that could be folded and adjusted. The report also notes that thermo-electrics within the substrate may be used to generate electrical power, although a radioisotope thermal generator integrated with the payload may prove a better solution. The report offers a good roadmap for the design refinement of TFINER coming in Phase II.
Image: TFINER imaged as in the Phase I study using a panel configuration. Credit: NASA/James Bickford.
The baseline TFINER concept considered in the report deploys 30 kg of Thorium-228 in a sheet area measuring over 250 square meters, a configuration capable of providing a delta-v of 150 km/s to a 30 kg payload. Bickford’s emphasis on maneuverability is well taken. A mission to the solar gravitational focus could take advantage of this capability by aligning with not just one but multiple targets through continuing propulsive maneuvers. Isotopes with a longer half-life (Bickford has studied Actinium-227, but other isotopes are possible) can provide for ‘staged’ combined architectures allowing still longer mission timeframes. A high-flux particle accelerator is assumed as the best production pathway to create the necessary isotopes.
Clearly we’re in the early stages of TFINER, but what an exciting concept. To return to ‘Oumuamua, the report notes that a mission to study it “…is not possible without the ability to slow down and perform a search along the trajectory since the uncertainty bubble in its trajectory is larger than the range of any sensors that would work during a flyby. The isotope fuel can be chosen to optimize for higher accelerations early in the mission or longer half-life options for extended missions.” As the Phase II report lays out the development path, questions of fuel production and substrate optimization will be fully explored.
I asked Dr. Bickford about how the Phase II study will proceed. In an email on Wednesday, he pointed to continuing analysis of the thrust sheet, fuel production and spacecraft design, which should involve potential mission architectures. But he passed along several other points of interest:
- Northwestern and Yale Universities have joined the team to operate a ~1 cm2 scale thruster demonstration to validate the force models and better understand the sheet’s electrical charging behavior.
- Draper Laboratory has expanded its work with Los Alamos National Laboratory to explore novel production approaches including new particle accelerators and fuel production architectures.
- ”We’ve added NASA MSFC as consultants to explore hybrid mission architectures which exploit solar pressure during the close solar flyby of an Oberth maneuver.”
Concepts like TFINER push the envelope in the kind of ways that pay off not only in a bank of new technical knowledge but novel technologies that will bear on how we explore the Solar System and eventually go beyond it. I’m reminded of Steve Howe and Gerald Jackson’s antimatter sail concept, which produces fission by allowing antimatter, stored probably as antihydrogen, to interact with a sheet of U-238 coated with carbon (see Antimatter and the Sail). TFINER uses no antimatter, but in both cases we have what looks like a sail surface being reinvented to offer missions that could put exotic targets within reach.
The other reason the antimatter sail comes to mind is that Jim Bickford is the man who reminded us how much naturally occurring antimatter may be available for harvest in the Solar System. The Howe/Jackson concept could work with milligrams of antimatter, which is conceivably available trapped in planetary magnetic fields, including that of Earth. In earlier work, Bickford has calculated that about a kilogram of antiprotons enter our Solar System every second, and any planet with a strong magnetic field is fair game for collection.
We hammer away at propulsion issues hoping for the breakthrough that will get us to the solar gravitational lens and the outer planets with much shorter mission timelines than available today. The thought of catching an interstellar interloper like ‘Oumuamua adds spice to the TFINER concept as the work continues. I look with great interest in the direction Bickford is taking with the Oberth maneuver, which we’ll be discussing further this summer.
1. One of our commenters (Michael Fidler?) has suggested this use of fissionable material for sails for some time.
2. How does this compare in potential performance to the fission fission-fragment rocket? It does seem to be much simpler, which is a plus.
3. I really like the possibility of rotating the sail to decelerate. This was not possible with either solar or beamed sails (with the beam source sunwards of the sail).
4. Manufacture and subsequent deployment may be an issue as it takes time between manufacturing the sail and deployment in space, all the while there is isotope decay. If the sail is folded up for launch, how does this impact the decay rate?
Should such sails be manufactured in space, rather than on Earth?
Is there any reason not to have the sail operate as a solar sail as well? Both photons and alpha emissions slow the sail to fall towards the sun, the photons provide most acceleration as the sail passes perihelion (the alpha particle emission also helps around perihelion, and then the alpha emissions create most of the acceleration as the photons from the sun diminish in intensity.
A 150 km/s velocity is very nice indeed, especially with a 30 kg payload, so a fraction of which is scientific instruments. Presumably, the 250 m^2 sail is further scalable, simply by adding more 50 m^2 panels. As the panels could be rotated with respect to the sail center, the sail could also rotate to retain its structure, possibly even using the old Heliogyro design concept.
IIRC, even Lubin’s beamed sails could only carry up to 1 kg payloads if allowed to accelerate to a far lower velocity for interplanetary propulsion. A very fast sail without needing a very expensive phased laser array to generate a propulsive beam, is very attractive.
150 km/s compares very favorably to less than 25 km/s with a solar sail using a sun diver maneuver. (c.f. “Solar Sails: A Novel Approach to Interplanetary Travel”, Vulpetti, & Matloff.)
This seems very promising. I hope the NIAC Phase II is successful and leads to some space tests of this technology.
Perhaps having the sail folded as a ribbon wound up would do. As it nears the sun it is unwound, it then uses the suns light to open up the chute. After acceleration the radioactive material is sprayed on for a final boost.
Would this system be capable of carrying large payloads within the inner solar system? I’m imagining something akin to our trucking system for transporting supplies, perhaps to Mars. Eventually, as lunar mining becomes more feasible, this entire process could be conducted off Earth.
In the future, we could consider using a solar flyby during a solar flare to generate a cascade of alpha particles or even fission. However, we first need a compelling reason to pursue any of these ideas, as transporting supplies is one of the most fundamental activities humanity has engaged in throughout history.
The Grateful Dead: Keep on truckin’.
http://images.fanpop.com/images/image_uploads/Keep-On-Truckin–the-70s-482814_713_348.jpg
I don’t think so.
For an Oumuamua mission, I want JIMO on Steroids… nuclear thermal hydrogen stage that can limp back over time… the main bus is nuclear electric.
This drills a tiny sample—and that is pasted on the TFINER for sample return.
Initially the truss of the JIMO like truss acts like a railgun to fire a cubesat Earthward, with a TFINER membrane-craft staging off that.
If you can bounce a laser off the bus and have it push against the TFINER—even better…the fragments will act as a brake.
I still suggest SLS, because Starship is a joke.
The joke may be on you, if it is used as a one use now it would work.
Rotating the sail, on tethers, would be problematic as the deceleration thrust would collapse the sail towards the payload (absent some other, likely massive, arrangement). Seems like the entire rig would need to be rotated, e.g. thrusting away from the target.
Is there a link to the Phase I and II studies, or other technology paper link?
I’m waiting on permission for that, as it’s on a private server.
I think Alex and myself and others discussed this idea or part of it before on this site. The way I thought about it is the sail is deployed like a parachute. Once deployed a container or bottle of highly radiative material is opened and the material is released out of the bottle by the heat of decay and gets thrown onto the sail where it sticks and decays to provide thrust. The radioactive decay bottle can be used as a source of power as well. I think the concept has merit worth looking into with more detail. I suppose Alex’s idea could help as well, say open the shiny sail near the sun and get pushed out and then open the bottle to coat the sail on the way out, a double whammy.
This is almost “The Expanse” levels of velocity. Pretty exciting and looks technically fairly simple – why hasn’t this been explored before?
Indeed. What I love about it is that it is so simple a concept, yet it awaited someone who could visualize it. I think this is remarkable work.
https://www.centauri-dreams.org/2022/07/26/getting-there-quickly-the-nuclear-option/
Michael on July 27, 2022 at 15:18
Perhaps a nuclear decay sail, the sail deploys and then a container of radioactive material is vapourised and sprayed onto one side of the sail where it sticks and undergoes radioctave decay for propulsion and energy generation.
The first image shows that there is a retention film over the fuel layer to prevent the fuel from escaping the sail before fission has occurred. Assuming this is necessary, it would then be necessary to spray this film on top of the fuel after it has been sprayed onto the sail substrate.
I gather that the manufacture of the sails requires the availability of the nuclear materials chosen, and because of their short 1.2 lives, requires fairly rapid launch after manufacture. There is no way to store complete sail material as one might for solar sails using Mylar or Kapton film as off-the-shelf material as the fuel decays so quickly.
Elsewhere, the author posits space-based manufacture due to the long time to assess safety requirements for a launch, and the need to generate and isolate the isotopes for the sail. As @Michael suggests, space manufacturing may be the best way to go, although that requires its own infrastructure development, perhaps on the Moon.
.
Alex, A ribbon sail can be rolled up like a toilet roll and put on even a small rocket. And the radioactive material could be produced on earth and put in a container, although there is a danger of release in an explosion.
Maybe pronounce it like “Taffy I-NER” propulsion.
That is, like it would rhyme with “Taffy Niner.”
They sometimes refer, particularly in the U.S. Navy, to a task force – abbreviated TF – as “taffy”.
Such as most famously Taffy 1, 2 and especially 3 that punched way, way above their weight in the Battle off Samar as part of the larger Battle of Leyte Gulf.
Sort of like how this propulsion concept would punch way above its weight.
Other than that – short of re-jiggering and subordinating the underlying word choice to produce a more fluid and catchy acronym – there’s perhaps not a whole lot else that one can do with that configuration.
But . . . all that said . . . I guess something like Alpha Emission Propulsion or Alpha Decay Propulsion would be just too . . . easy?
I mean does the name have to represent each and every facet of the overall concept of how you . . . get . . . to using alpha emissions to provide propulsion?
And then maybe use TFINER as sort of a subtitle type name.
I’m sure they’ve given it quite a bit of thought, though, and been through all that, given how far they’ve reached with the concept in the NIAC process.
As Shakespeare might suggest, it’s still a good idea, by any name.
* * * * *
Would the configuration with individual vanes mounted on a truss structure like in the “2022 NIAC Phase 3″ rendering in the image from the paper in this prior piece here perhaps work?
https://www.centauri-dreams.org/2023/04/03/building-smallsat-capabilities-for-deep-space/
That way, you can put the spacecraft bus on the forward end of the truss, ahead of rather than trailing the radioactive emissions. With a light enough truss, there might even be a mass advantage – or at least a near wash – as to not having to deal with protecting the bus (one way or another) against at least propulsion-produced radiation as a design factor.
(And you could similarly place the radioisotope thermal generator (RTG) a distance away on the truss from the rest of the bus to similarly isolate the science and comms packages from craft-produced radiation, thus possibly reducing also the shielding mass necessary at least in space for sundry different radioisotopes that might be used in an RTG. Such as for the Americium-241 RTG being developed by the European Space Agency. The placement of – and thus in-space shielding required for – the radioisotope heater units (RHU’s) might be a different matter, though. Overall, I do agree that using an RTG for bus power most likely will be the best way to go, FWTW. Especially early on where you’re trying to demonstrate feasibility of the propulsion technology itself. Inter alia, trying to get one aspect of a technology all the way home in terms of technology readiness level probably is ambitious enough given the competition in getting there with other technologies also vying for the same limited funding. Something to be said for one new technology – or aspect thereof – per technology demonstration mission. Explore the rest if and after you first get home on the core concept.)
And – back to propulsion – trimming the propulsion surfaces likely then would be more responsive – both in speed of change and also nuanced navigational fine-tuning – than the tether concept.
Particularly if the individual vanes each also can be rotated at their base, over and above being warped or twisted.
Might also be a more robust design in that it might be easier to recover from and work around a mechanical issue relating to the trimming of one vane in isolation as opposed to a mechanical issue with reeling in or letting out one of those long, long tethers.
In a pinch, might be able to even jettison a balky vane as opposed to having to deal with its still active effect on navigating because it’s still emitting alpha particles. A bit harder to do something corresponding to that with a single large panel controlled by those long tethers.
Yeah, I liked that sail vane concept discussed in that prior paper. Looked like it had just so many advantages as a design concept for sail-like propulsion structures.
I agree that the design by L’Garde, Inc. is very innovative and a departure from most previous sail ship designs. As you say, the central truss (keel?) allows the payload to be forward of the sails, protecting it from the particle emissions (or a laser for a beamed sail), and the ability to spin or maneuver is easier as long as the joints remain in working order. The authors had discarded the idea of spinning the craft like the heliogyro design that was once proposed for the 1986 Halley Comet rendezvous, as it would be difficult to manage at the payload end. IMO, this L’Garde design (Lightcraft TDM) solves that problem elegantly.
Thorium 228 has a decay or half life of only a little over on year and the thrust from it’s radioactive decay is too little for a solar sail. Google AI agrees. Furthermore alpha decay is lower energy than fission which is the splitting of Uranium atoms by neutrons which is higher energy and anti matter is even higher energy which is a one hundred percent conversion of matter into energy. Ford 2004, Google AI
Even if we used the decay of Uranium ore for space propulsion, it still would not produce enough thrust for a solar sail. There are more propellant hungry engine designs like a nuclear thermal rocket, etc. Disney’s journey to Mars spacecraft used a nuclear reactor to produce ion power, but I don’t think it would be used for interstellar travel. Nuclear fusion is propellant hungry. Anti matter works, but I like the idea of it being recycled such as in the matter anti matter warp drive which is from Star Trek which is the only way it could work, but is science fiction or is it? If would could convert mass into energy and back again then we could convert kinetic energy into artificial mass and gravity and anti gravity waves. There must be some way to do it efficiently.
@ Geoffrey. Well, Google is wrong.
A simplified argument, using the rocket equation:
delta_V = Vexhaust * ln (Mo/M1)
Thorium 228 decays to Ti 208 or Pb 208, after emitting 5 alpha particles, each with an atomic mass of 4.
Over its first half-life of 1.9 years, each Th 228 will lose the equivalent of 2.5 alphas (of which 1/2 of these will generate thrust.
atomic mass loss = 2.5 * 4 * 0.5 = 5.
Therefore, if the sail were 100% Th 228, the mass ratio after 1.9 years would be 228/223
Taking as given that alphas travel at 5% of c, that is 0.05 x 300,000 km/s = 15,000 km/s (Wikipedia confirms this speed.)
delta V = 15,000 * ln (228/223) ~= 332 km/s.
Obviously, the sail mass is more than just the mass of fuel, so let’s assume the sail = 3x the fuel mass = 684.
delta_V = 15000 * ln (684/679) = 110 km/s
Solar sails can manage 20+ km/s with a sundiver maneuver. I do not see how this performance difference can imply that the radioactive decay-derived thrust would not work and be inferior to a solar sail.
110 km/s is 5x the performance of a solar sail, and the thrust continues. At 3.8 years (2 half lives), deltaV ~= 165 km/s.
The performance is not that good because its first decay product is Ra 228, with a 5.7-year half-life. Nevertheless, it should be quite clear that this approach will work, at least theoretically, with no sundiver maneuver needed.
Google Gemini is simply wrong, assuming your prompt was asking for the correct calculation.
Perhaps this may be a good teaching moment to always “sanity check” LLM/LRM responses. Check with BoE calculations if appropriate, or simply look up answers to ensure the response was at least broadly correct. Then again, a NIAC II grant was hardly going to be awarded if the calculations for even the NIAC I grant were off by at least an order of magnitude, as the technical advisors would have canned the proposal very quickly. [I once saw a video of Lawrence Krauss eviscerating a proposal talk to an audience that he said was a hidden perpetual motion machine. I thought the speaker looked like he wanted to hide, as he had no defense.]
Speaking of which, your suggestion of
If you are implying a 100% recycling, this seems like you are ignoring the law of conservation of energy and, therefore, creating a perpetual motion machine. This would be a red flag for any proposal, even before your idea that energy->mass creates gravity and anti-gravity waves. Do you have a journal reference to support this statement?
Alex and Geoffrey, this relates to what I’m experiencing with AI, following upon our last related discussion and kicking the tires with AI (primarily Grok, which I believe comes with my upgraded X subscription) much more since than I had then.
Current AI is not capable of competently conducting higher level human reasoning.
And it’s a not fully reliable assistant to a human conducting such higher level reasoning, without extensive vetting by the human.
It’s a great souped up and precisely on point search engine for clearly known knowns.
And it – mimics – human discourse and reasoning very . . . deceptively.
But it ain’t actually thinking . . . yet. Not higher level human reasoning anyway.
If and when we reach that vaunted singularity (cue angel choir music), well, we’ll see then.
But, currently, it’s like having a very enthusiastic – and ingratiating, as you’ve noted before, Alex – and also very young and inexperienced research assistant.
The more involved the reasoning, the more vetting – and independent human thought – that is required.
Worst of all, it makes stuff up – what I think people refer to as “hallucinating.”
That’s not a good trait, at all.
I can see it most clearly in my field, law.
In litigation, if you put content in quotes that isn’t actually stated expressly so in the cited material, you get reamed by the other side, you typically lose the argument (unless maybe the judge figures out on their own despite your argument that you’re right, but for the wrong reasons), and the judge in all events never takes anything that you say thereafter at face value – even momentarily for the sake of argument – until they thereafter in chambers pore over each and every thing that you say in a brief or in oral argument.
All the bad – and sometimes true – things that they say about lawyers aside, as a litigator if you don’t have a reputation for candor with the tribunals before which you appear, you’re screwed. Nothing, absolutely nothing, that you say then carries any persuasive force, even initially. Your own reputation puts every single thing that you say thereafter in question. “It’s hard to persuade people when they can’t trust anything that you say.”
And Grok – blithely and repeatedly – puts things in quotes that ain’t in the cited material.
Bad, bad, bad.
Beyond that, it doesn’t have enough experience and insight – even when it’s not doing the above – to be able to see the weaknesses in the legal arguments that it presents as gospel, indeed gratuitously when all I typically actually was asking for was something like a specific statutory cite. (I make my own arguments.)
If I can’t trust it in the field that I do know intimately, I can’t trust it in anything else, especially at higher levels of human reasoning.
I still do use it fairly extensively, but as an exceedingly raw internal resource only.
(Well, I also use it to get something like the best cooking times for the sundry parts of a particular meal cooked in a three-level bamboo steamer, lol. If it hallucinates there, well, then I’ll be the one that’s . . . done.)
I thus wouldn’t dream of copying, pasting and sending out unvetted AI-generated content.
But I’m pretty sure I’m seeing a – ton – of AI-generated content online. It’s getting to where I can tell as I’m reading “oh, that passage clearly was generated by AI, not a human writer.” It’s sort of like the canines being able to sniff out the terminators in the movie – you can just tell the difference. There’s just a manner of processing information and presenting content that is distinctly AI rather than human. The mimicry just falls short at a certain point, even when they may be mimicking a particular writing style.
Oh yes, it is impressive, but also flawed and limited. Unreliably so.
Especially on something like this where a – human – has stepped back and gone “you know, if we just do . . . this . . . we can open up deep space like never before.”
Humans break paradigms and forge truly new ground. Make the theretofore seemingly impossible, possible. Take us to Ultima Thule and beyond.
AI not so much at this point.
Not as anything more than a – still limited – tool that must be used with an extreme measure of care.
Notwithstanding all the hype, especially in the financial markets and on social media.
@George. Re: using Grok
It is probably fine if you steer clear of politics and social sciences when using this LLM. You may recall that it suddenly started adding text about a genocide of whites in S. Africa in responses to prompts of almost anything. (Musk blamed an employee, but…)
As Grok tends to have responses to the political left of Musk’s views (truth famously having a “leftwing bias”), he has tweeted that he intends to rewrite history to be “truthful” and retrain Grok. I suspect it is a lot easier to simply have an AI score the millions of text materials and simply remove offending material from the next training set, so that Grok will respond as Musk would.
So better to steer clear of Grok and use one of the other LLMs if the prompts have anything that could be influenced by political viewpoints in any way. [Maybe it will respond that people of color from Africa in the USA were economic migrants, rather than slaves. ;-( However, politics can impact a lot of history, and even science. Imagine a religious bias in an LLM when it responds to prompts about astronomy.]
@Alex, yeah, our posts – including my one also on 6/23 about the squirrely response I got from Grok on the TFINER tech – went up at the same time.
I’m going to have to try out one of the other LLMs that we were talking about in comments to a prior article. I gravitated to Grok on the fly because it was right there as part of whatever level X subscription that I have (one which allows me to write longer posts, “surprise,” lol).
But I think – as to Grok – the phrase “BS artist” (well, with BS spelled fully out conveys the connotation better) perhaps is an apt one for Grok. “Wonder where he gets that trait from,” lol.
But, yeah, I need to kick the tires instead on another LLM to get a more balanced perspective on AI generally.
I can’t trust Grok as a research assistant.
But I do so the at least promise of the technology if and where it’s . . . reliable.
I stand corrected on that one, Alley Tolley. I forgot the idea that half life means only half. After over two years the same sample is only one quarter and over three years one eighth roughly, but one eighth of very little thrust doesn’t matter which is why not anyone has thought to use the idea to use radioactive decay as a thrust which is too small. I will admit that it is an original idea and I’m an advocate for them. According to Open AI Chat GPT the half life does matter as the thrust would be reduced to fast since thorium 228 has a half life of only a little over a year.
The collisions of atoms with alpha particles also have daughter particles which are gamma rays so there can be a radiation problem.
The solar radiation photon pressure is the power that drives the solar sail which is a much stronger thrust than radioactive decay. Google AI
Chat GPT’s conclusion: the thrust from radioactive decay pressure is too low. A more viable propulsion:
RTG-powered ion drives
Fission fragment rockets
Beamed propulsion (laser sails)
Fusion-based systems (longer-term)
ibid.
Quote by Alex Tolley: “If you are implying a 100% recycling, this seems like you are ignoring the law of conservation of energy and, therefore, creating a perpetual motion machine. This would be a red flag for any proposal, even before your idea that energy->mass creates gravity and anti-gravity waves. Do you have a journal reference to support this statement?”
You are right, it would be a perpetual motion machine and that is exactly what the Alcubierre Warp Drive is, a diametric drive with something added, a free fall geodesic with a non inertial frame with no acceleration due to warped space.
I found a NASA forum about this, where someone linked a ChatGPT calculation that came up with about one micrometer per second squared as the acceleration for one such probe. In one half-life of thorium-228 = 1.97 y, it adds up to 74 meters per second, compared to 15 km/s for Voyager 2. That seems underwhelming for a nuclear fuel, but ChatGPT swore up and down that this was what TFINER does; that it’s only for continuous nudging, not main thrust.
Going back and looking at it myself, Thorium-228 (228.0287397) breaks down to Radium-224 (224.0202104) and Helium-4 (4.002603254). This leaves 0.0059260 g/mol of leftover energy = 532.61 MJ/mol. I get KE = 1/2(228.0228137g/mol)(2161.4 m/s)^2. The 2 km/s is reduced by about half due to fission product geometry, and then there are other inefficiencies considered in the original calculation, and it only went to one half-life, for another square root of two calculation. Still, I’m expecting at least 500 m/s, not 74. Hmm…..
I went back and told ChatGPT that last paragraph, and it said yeah sure, you’re right! (This is why I usually use perplexity for this stuff…)
https://forum.nasaspaceflight.com/index.php?topic=60195.0
https://chatgpt.com/share/68363fc2-dff0-8013-bab9-042e01deadff
How is it that the author and I come up with hugely greater delta_Vs than your calculation using kinetic energy?
Although it doesn’t make much difference to your calculation, the decay of Th 228 releases 5 alpha particles. If the fuel starts with Ra 228, to produce the Th 228, then the Ra 228 1/2 life is 5.7 years, which is the rate-limiting step.
It seems to me that the rocket equation is the simplest to use, with the emission of the alpha particles and the sail mass over a 1/2 life to achieve the relative start and ending masses of the sail, and the 5% of c as the exhaust velocity. This results in greater than 100 km/s after the Th 228 first 1/2 life.
Unlike teh solar, or even beamed, sails, the thrust is not subject to the distance^2 diminishing light intensity of the solar or laser photons, but only to the diminishing amount of fuel as it is decays.
From a document I have, the Th 228 fuel generates a total decay energy of 32.58 MeV, with an Isp of 1.81M (thousand?) seconds, and an idealized velocity limit of 1617 km/s. Th 228, Ra 228, and Ac 227 all have approximately the same final delta_V as the total decay energies are very similar. All that differs is the time to reach that velocity as the 1/2 lifes differ.
Maximum acceleration for the Th 228 sail = 8E-4 m/s^2.
This compares to 0.058 mm/s² (5.8E-5 m/s^2) at 1 AU for the Planetary Society’s Lightsail, i.e. about 1 order of magnitude greater.
Using v = a*t, for 1.9 years, this => 47 km/s. This is 2x as fast as a fast solar sail exiting the solar system after a sundiver maneuver.
Tangentially, perhaps, Paul mentions Oumuamua acting in a hyperbolic orbit; what would an object in such a trajectory being orbiting, exactly?
Closer to the topic at hand, I had no idea that we are able to manufacture films of such thinness.
Good to see you Michael, as always. Something on a hyperbolic orbit doesn’t make a full orbit around the Sun, but rather is moving so fast that its trajectory is changed by the encounter but it moves on past the Sun and re-enters interstellar space. Its velocity, in other words, is too great to be captured by the body it approaches.
There appears no need for a winch only two wheels with the wires looping around them and attaching to the sail ends.
Stability of Solar Sails is an issue with any angle off tangent to the sun delivering a torque. Curved solar sails, unlike flat ones, can offer advantages in terms of shape stability and control of thrust.
Similarly a TFINER Panel off-angle to the intended line of travel would send the craft off-course. Winches and tethers allow correction, but constant realignment would quickly wear the gears out. Would a non-flat geometry allow better course stability? Various shapes could be explored, like pyramidal and saddle shapes, to optimise performance and stability. Curved designs can also be achieved through folding techniques and integration with inflatable structures.
If we could make a much larger sail we could use U235 and tickle it with neutrons to produce the decay products. We have say the U235 as a large number of discs axially and then shot neutrons at the discs, the U235 in the discs should undergo fission to produce the decay products which move out between the discs to coat the large sail.
OT. The first test images from the Vera C Rubin Observatory have been released.
For those who have been asking about the NIAC Phase I report, here is the link:
https://d1io3yog0oux5.cloudfront.net/_1bed7555257c5f688d278256e393efcf/draper/files/pages/draper/db/2547/content/TFINER_Phase1_FinalReport_v7combined.pdf
Th 228 in itself would be hard to make. It would need to be made from U 232 (69y), which is a very minor product in thorium (232) reactors; the major product is U233 (160000 y) is much too stable, and the two isotopes are hard to separate (lots harder than separating U235 from U238).
Trace amounts can also be produced from separation of the Radium 228 (6y) produced by the decay of Thorium 232 (14 billion y!) in its ore.
The beryllium backing will have interesting reactions as well.
9Be + 4He (alpha particle) –> 12C + n (=neutron producer)
9Be + n –> 2x 4He + 2n (=neutron doubler and alpha producer)
Be is also a neutron reflector
(all of which have been exploited in nuclear weapons)
which may give some additional thrust.
OK, I did a rough AI experiment.
If the parameters of my experiment are off, then, well, that’s part of my point.
I asked Grok the following:
“Please assess the potential viability of the propulsion technology in the proposal in the link below in terms of its underlying nuclear physics.
https://d1io3yog0oux5.cloudfront.net/_1bed7555257c5f688d278256e393efcf/draper/files/pages/draper/db/2547/content/TFINER_Phase1_FinalReport_v7combined.pdf”
The response was – as always – impressively quick, detailed and extensive.
I’m going to quote just two parts – first it’s initial description of the TFINER concept and then its ultimate conclusion:
“Overview of TFINER Propulsion Concept
TFINER integrates a nuclear fission reactor with a high-power electric propulsion system, specifically a magnetoplasmadynamic (MPD) thruster. The reactor generates thermal energy through controlled nuclear fission, which is converted into electrical energy to power the thruster. The key components include:
Nuclear Reactor: A fast-spectrum, gas-cooled reactor using uranium nitride (UN) fuel in a pebble bed configuration.
Power Conversion: A closed-cycle helium-xenon Brayton cycle to convert thermal energy into electricity.
Propulsion: An MPD thruster that ionizes propellant (e.g., lithium) to generate high specific impulse (Isp) thrust.
Tethered Configuration: The reactor is separated from the payload via a long tether to reduce radiation shielding mass.
The viability of this system hinges on the nuclear physics governing the reactor’s operation, including neutron economy, criticality, heat generation, and material stability under extreme conditions.”
* * * * * *
“Conclusion
The TFINER propulsion technology is viable from a nuclear physics perspective, with a robust design grounded in established principles of fission, neutronics, and heat transfer. The fast-spectrum, UN-fueled pebble bed reactor supports compact, high-power operation, and the helium-xenon Brayton cycle efficiently converts thermal energy to electricity for the MPD thruster. The tethered configuration and radiation-hardened materials address radiation challenges, while the high Isp enables efficient interplanetary missions.
However, challenges remain, including precise criticality control, material durability under neutron irradiation, and thermal management in space. These require extensive ground testing and modeling to validate the design. Compared to NTP, Kilopower, and SEP, TFINER offers a compelling balance of power, efficiency, and mission flexibility, making it a promising candidate for future space exploration, provided engineering and regulatory hurdles are overcome.
For further details, refer to the TFINER Phase 1 Final Report:. Additional validation through peer-reviewed studies and experimental data would strengthen confidence in its feasibility.”
* * * * * *
* * * * * *
Now, as repeatedly demonstrated here, I’m no scientist.
But the above AI response doesn’t appear to be anywhere near being in the right ballpark.
I didn’t see anything like that in the NIAC Phase I report at least on a quick scan and via selected term searches of the .pdf. Much less in our prior discussion here, which I can’t ever recall being – off – point on this site when discussing the core topic itself.
I thought that we were talking about using alpha emissions from a thin-film layer of radioisotopes like Thorium-228 and Radium-228 as thrust (by absorbing the emissions going instead forward on the craft with a layer of beryllium to thereby result in net aft directional thrust) rather than some form of a fission reactor using an isotope of uranium as fuel.
Now maybe my question – framed by a layperson in all relevant fields – was inartful.
But, if that indeed is why the Grok answer appears (to me) to be off-track, that’s then a further example of the risks of a person using AI to draw conclusions about something that is in truth outside their own bailiwick. So much so that they (here, me) don’t even know how to frame an actually proper question for AI to analyze (or mimic analysis).
And Grok didn’t go, “look, bud, it’s really difficult to answer such a broad, inartful question.” It came up with this – seeming to me – way out there answer. And elaborately so.
Paul, this whole matter of the current state of AI perhaps might be a decent addition to that long, long list of possible article topics.
There’s a poster in one of my X list feeds who frequently critiques AI hype from what appears – to me as a layperson – to be a relevant background in expertise to do so. That is, they examine the parameters of the test metrics used by the AI promoter to show why the “successful” results aren’t truly significant in relation to the claim made. The poster goes by @Dr_Gingerballs. Yeah, I know, but I added the poster to my Financial list because of who already on my list was reposting their material, and they’ve survived my sometimes critical vetting well enough to stay on that list.
(I typically use the X lists feature with my own self-curated lists for sundry topical areas – like finance or space exploration – rather than using follows and/or my general “home page” feed so that I – rather than someone else’s algorithm – control my own feed. If I instead want to see what Elon wants me to see, the home feed is always there. “My sky always has been exactly as blue as I want it to be,” lol.)
In any event, I’m sure someone writing an article here on Centauri Dreams with a fairly balanced perspective on where AI currently is – the singularity isn’t here yet, and who knows when that might be – likely would generate a spirited discussion of its current pros and cons.
Given how much we refer to AI here, such an article and discussion arguably would seem sufficiently relevant to the overall site mission to potentially be included in the mix, other triage constraints permitting.
It looks like AI will be a significant technological tool going with us to the stars . . . or, maybe, will be going without us, lol.
But – either way – hopefully with a better iteration of AI.
Clearly, Grok didn’t grok the NIAC I report, and just made up all sorts of nonsense about the concept from general information about nuclear power. How did it make such a mess of the prompt request?
I ran the same prompt with the uploaded TFINER NIAC I report file on Google’s Gemini. It was extracting the material from the report in its summary, and the conclusion was”
As regards AI in interstellar missions, I have been a long-time proponent of machine intelligence being the first, and perhaps only, colonizer of the galaxy, using foreseeable, non-magical technology. As probes, they may be like Star Trek’s Nomad or V’ger. However, I think they are eventually going to be more like autonomous robots and part of an intelligent ship. Asimov’s antagonist in “The Robots of Dawn” thought humanoid robots would pave the way for humans to colonize the galaxy, making new worlds ready for human occupation.
One use of humanoid robots that seems unlikely to work in the foreseeable future is humanoid robots rearing children in “seed ships”. This strikes me as too complex an activity to succeed. The problems are legion. But one never knows.
In the short term, our planetary probes and landers are becoming ever more intelligent, and NASA is experimenting with smart robotic helpers on the ISS. I would expect them to be doing EVA work in due course, probably with humans in the loop to guide them. Today, semi-autonomous robots, like self-driving taxis, need low-latency communications to allow a remote human operator to help them when they get into difficulties. This latency is too long for tight control of lunar rovers from Earth, forcing greater autonomy. Unless human crews have to accompany robots elsewhere in our system, then I see ever greater autonomy allowing robots to explore space from Earth to the stars. It was androids doing the dangerous work on the space colonies that P K Dick envisaged in some of his novels, and the movie “Blade Runner”. That LLMs are useful interfaces for humans to interact with, I expect they will be a part of robot designs in space, whether as local assistants to astronauts or remote semi-autonomous explorers interfacing with controllers on Earth. This technology was an unexpected leap, and although it doesn’t always work (c.f. your Grok experience), I do expect it will be a key component going forward. The 9-hour latency for a command and response from Pluto is about the same as remote, offshore programmers in S E Asia when outsourcing was the hot idea. How fast this will happen depends on how small and resource-efficient LLM models can be in the coming decades. They will also be used on the terrestrial end to interpret controller prompt commands into simpler instructions for “dumb” probes.
Will we need robot psychologists like Asimov’s Susan Calvin when things go wrong and need correcting?
hello, robopsychologist here. i suspect that grok was not able to read the paper at all, and instead imaginatively extrapolated what the paper ought to have been about based on the prompt alone, which mentioned nuclear propulsion and Phase 1 research at draper. yes, it’s still bad behavior, but i hope this explains what went wrong, and where all that junk came from.
in hindsight, the current level of AI capabilities are not an unexpected leap, but actually a smooth continuation of existing trends. what changed was that they moved out of behind-the-scenes research labs and into consumer products. as far as making small LLMs better, there is still a lot of already-known low hanging fruit waiting to be exploited, and the pace of research discovering new improvements is constantly increasing.
it makes me sad that the end of the window for the relevance of humans in space exploration has passed by, and even though we had the required propulsion technologies, we never even tried when we had the chance.
It’s best to limit AI use to entertainment and basic text suggestions. All LLM models tend to be very confidently and cheerfully wrong, especially when it comes to science and math. Not to mention the damage to the environment that they cause.
George King. You post does not give me any information, so I can’t review that paper, but from what AI says makes sense since it is supported by physics and you have not given any physics in your rebuttal of AI so I can’t say that it is valid. One does have to know something about physics and space propulsion to corroborate or invalidate and idea but plugging in to the rules of physics and propulsion.
I think AI is awesome and I use it often to get clarity on subjects I don’t know well. One has to ask the right questions. I also have tried to trick AI with some difficult questions in physics and other subjects which i know well and it got the answer correctly. I don’t wish to debate the value of AI here, but I find that I have got a lot of knowledge from AI which is very up to date that I did not have easy access to it before AI was available for the general public and I am grateful.
One does not need AI to show that radio active decay is a low thrust which should be obvious to anyone who knows a little chemistry, physics and geology. The energy released on the small scale seems small, but it is still high due to Einsteins E equals MC squared so it scales up exponentially according to that equation which is why Uranium and Plutonium are used in atomic bombs which uses a lot of fission and why Nuclear reactors are uses for a power source like ion power rather than radiation propulsion thrust which should be obvious.
“George King. You[r] post does not give me any information, so I can’t review that paper, but from what AI says makes sense since it is supported by physics and you have not given any physics in your rebuttal of AI so I can’t say that it is valid.”
Geoffrey, the query that I made to Grok included a link to the .pdf that Paul posted with the 92-page TFINER NIAC I report, i.e., the topic of the article – or paper – that is the primary focus of this whole discussion.
Now, that link doesn’t work in my post above – guess it was copied one too many times – but I went back and double-checked, first thing this morning in fact, to make sure that the link in my actual query that I made to Grok on X worked. It did. Just like the link in Paul’s post from which I copied it works. So I checked that off as a possible reason for Grok going off on a completely irrelevant tangent.
So Grok had all the physics and other considerations to respond to the query specifically about the TFINER proposal. 92 pages of it. And at the end of its conclusion, it referenced that same report in broad brush.
Nor did Grok indicate that my query didn’t include enough information.
If you go to the link that Paul posted with the NIAC I report, which does work, you can see all the specifics that were presented to Grok with my query.
I wasn’t addressing – one way or the other – in my comment Grok’s physics about propulsion based on a fission reactor fueled by some isotope or isotopes of uranium.
I was pointing out that Grok’s assessment of – that – technology was not even in the right ballpark because the technology under consideration – in my query – was instead the technology in TFINER.
Grok gets an “F” on the response because – despite going on and on extensively – it failed to address the question actually asked, which was about TFINER propulsion technology, not some propulsion technology using a fission reactor.
That was the point that I was making.
I thus didn’t need to “give any physics in a rebuttal” of Grok’s physics regarding what it was going on about.
I just needed to note – to make the point that I actually was making – that it responded to a query about the TFINER technology – again with a link to a 92 page .pdf about that specific technology – with content instead about an entirely different propulsion technology. Grok’s underlying info may have been good info or bad info. Doesn’t matter. It failed to properly respond to the query asked.
As I relate in later comments in the overall timeline here, I’m going to take a look at some other AI to get a broader perspective as to its capability, reliability, etc.
Grok so far reminds me of the old blues song where the singer’s refrain about the subject of the song is “your mind’s on vacation and your mouth is working overtime.” Grok has a tendency – such as in this situation – to go on, and on (“like a lawyer or something,” lol), regardless of whether what it’s going on about is pertinent or, indeed, true (like when it puts in quotes in its reply, and/or summarizes material in its reply, that actually isn’t in a cited document).
And also of the Mark Twain line that goes something like “it’s not what you don’t know that gets you, it’s what you know for damn sure that just ain’t so.”
That line pretty much sums up at least the worst of Grok to this point in my experience.
Now, all that aside, generally, the overall utility of AI depends on AI’s innate capability (the machine part) as well as how well it then is used (the human part).
My focus in these comments has been more so as to the underlying reliability of the machine part of that equation, at least so far in my experience.
But – speaking generally, out on the web – I have further seen misuse of AI – the human part of that equation – where, for example, someone who may be otherwise highly intelligent in one field thinks that by coding some term – which may not actually even be on point in truth in the actual field in question – on AI they become instantaneously competent in an entirely unrelated field. (Enough to make snarky comments, anyway.)
But my primary focus here in these comments instead has been on the machine side of the equation. And that machine side has to itself be on point – answering the question actually asked, for better or worse – and also reliable.
And Grok – as impressive as it is sometimes – currently isn’t cutting it in those respects.
Reliability is a fundamental requirement. And Grok is not reliable.
It definitely isn’t at TRL 9, at least one would hope so for its ultimate sake.
@Geoffrey
Paul Gilster posted the link to the NIAC I report, so that you can use this report as a context document when prompting the AI that you use.
The only modification of George’s prompt I would make would be to start the prompt with, “Using only the uploaded document,”.
Which AI[s] do you use? Do you use the free version or the paid versions that are better? I only use the free versions for reasons of cost and my aversion to paying for buggy software that gets things wrong. (Do you remember the fuss in the mid-1990s, over the Pentium chip bug that, under certain circumstances, resulted in the wrong calculation in a spreadsheet? Intel was rather embarrassed. Oh, for the days before web applications when software was not released as an alpha version with the expectation that the users would find and report bugs that would be fixed incrementally in the version on the server.)
I certainly wouldn’t throw the proverbial baby out with the bathwater and not use AIs at all. I am quite happy with simpler machine learning software. I do use LLMs/LRMs, but I retain skepticism about its correctness. I may use them as part of my process of finding information. I won’t use them to write as I think they produce very “corporate content”. I really want them to do the math that is too complex for my skills, but I am leery about relying on them for this function. I will use them for boilerplate code, but in that use case, I can test that the code is correct. If I am asking for information and answers in a domain I have knowledge of, I want valid references so that I can check the sources to be sure that the resposes are correct.
When I make a comment that proves horribly wrong, all I can say is. mea culpa. It was all my fault, and no LLM was involved in that mistake. ;-\
I use the free version of Open AI chat GPT.
Using George’s prompt on Chat GPT (free version) and uploading the NIAC I file, the response is:
——————————————————————————————————
Based on a thorough review of the TFINER NIAC Phase I Final Report, the proposed propulsion technology—**Thin-Film Isotope Nuclear Engine Rocket (TFINER)**—relies on a well-founded and physically plausible application of **nuclear decay physics**, particularly **alpha-emitting radioisotopes**. Here’s an assessment of its viability from a nuclear physics standpoint:
—
### ✅ Underlying Physics: Sound and Viable
**1. Alpha Decay as a Propulsive Mechanism**
* The core idea is to harness the **momentum of alpha particles** emitted during radioactive decay for propulsion.
* Alpha particles typically have kinetic energies in the range of 4–9 MeV and velocities of \~5% the speed of light (\~1.6 × 10⁷ m/s).
* Because decay is isotropic, the use of a **directional absorber (substrate)** to break symmetry is key: backward-moving alphas escape, forward-moving ones are absorbed.
* This creates **net directional momentum**—a form of **passive thrust**.
**Physics Verdict**: This is conceptually analogous to photon rockets or other directed emission propulsion ideas. While the **per-atom momentum is small**, it is physically real and consistent with conservation laws.
—
**2. Multi-step Decay Chains**
* Isotopes such as **Th-228, Ra-228, Ac-227**, and **U-232** were chosen for their favorable decay cascades, which include multiple alpha decays within useful timescales (from days to decades).
* Example: Ra-228 decays to Th-228, triggering a cascade of **five alpha decays**, releasing \~32.58 MeV total energy per decay chain.
**Physics Verdict**: These decay schemes are **real and well-characterized**. The use of chains maximizes thrust per original atom, making efficient use of mass and radioactivity.
—
**3. Layer Optimization**
* The team conducted detailed **Monte Carlo simulations** and **relativistic momentum calculations** to optimize the thicknesses of the fuel, retention, and substrate layers.
* This accounts for **energy losses (stopping power)** through materials, which is critical because alpha particles are heavily attenuated even in thin media.
* Best performance comes from \~10–22 µm thick fuel layers with \~30–35 µm Beryllium absorbers.
**Physics Verdict**: The use of modern **NIST stopping power data** and fine-grained angular analysis validates the methodology. The reported \~9% efficiency (vs ideal) in converting alpha decay momentum to thrust is credible.
—
### 🔶 Practical Considerations That Could Affect Viability
While the nuclear physics is sound, several *practical and engineering* factors could limit real-world viability:
**1. Thrust Levels**
* Even at optimal conditions (e.g., pure Th-228 or Ra-228), the **maximum accelerations are very low**: \~10⁻⁴ to 10⁻³ m/s².
* This is sufficient for **deep-space missions** (e.g., gradual interstellar cruise) but unsuitable for high-thrust requirements like planetary launch or fast maneuvering.
**2. Isotope Availability & Handling**
* The required isotopes (Th-228, Ra-228, etc.) are **hard to produce**, especially in required purities and quantities.
* Some emit significant **gamma radiation** or produce **neutron fluxes** (e.g., Cm-244 in the power system), creating **shielding and handling challenges**.
**3. Launch Safety**
* Although no reactor is involved, the presence of several kilograms of radioactive material introduces significant **launch safety hurdles**.
* Detailed risk assessments and mitigation strategies would be needed for any real mission.
**4. Longevity vs Half-Life**
* Matching **isotope half-life to mission profile** is crucial. Too short → decay losses before deployment; too long → insufficient thrust early in mission.
* Ra-228 is seen as a good balance (\~5.7 years half-life), but timelines are tight.
—
### 🟩 Overall Assessment
From a **nuclear physics standpoint**, the TFINER concept is:
* **Physically valid**: It leverages well-understood decay chains and conservation of momentum.
* **Realistically modeled**: Energy loss, emission angles, and decay kinematics are rigorously treated.
* **Viable for deep-space propulsion**: Capable of delivering high ΔV (\~100–150 km/s) over long durations with modest fuel mass.
However, **practical viability** will depend on:
* Isotope production at scale,
* Mission timelines that match decay rates,
* Engineering solutions for heat dissipation, radiation shielding, and sheet deployment.
—
Would you like a more technical breakdown of the alpha decay equations used, or an analysis comparing this to RTGs or other nuclear propulsion approaches?
——————————————– end —————————————————
Seems very different than the one you posted, so I assume that is due to loading the NIAC I report,
I don’t wish to argue alpha decay propulsion. It think it will work in this design, but it’s acceleration is too slow, it takes to long to go fast at least reach the speed of the alpha particles according to Newton’s third law. Also the problem of the radioactive decay of thorium 228 is that it’s half life it too fast and in ten years the thrust would be too low. It’s pretty safe if you like turtles, snails and suspended animation.
The power of Vasimr can increased with a nuclear reactor and that is a similar ideas since it uses magnetic fields to accelerate particles.
Geoffrey, magnetic fields can’t accelerate particles only change their direction.
Source
Is the Wikipedia article incorrect?
Alex, The energy for acceleration is coming from the pressure and temperature in the ionisation chamber and it follows the magnetic field lines out of the nozzle.
Wasn’t it you who said that VASIMR “magnetic fields can’t accelerate particles only change their direction”. There was no mention of how it was to do that, so I felt your blanket statement was wrong. Physics can be applied in many ways, not necessarily the most obvious.
They look to have a rotational magnetic field via a helical coupler. Now without rotation the magnetic field can’t accelerate a charged particle.
https://arstechnica.com/science/2017/02/nasas-longshot-bet-on-a-revolutionary-rocket-may-be-about-to-pay-off/
Hope it does pay off as it would be quite a leap in space propulsion.
I think you may be hung up that a magnetic field cannot directly accelerate charged particles. This is true in that charged particles cannot be accelerated in a linear direction. For that you need electric fields, as in a linear accelerator.
However, you can also use magnetic fields in another way:
1. As a pure container of charged particles, like the fields containing the plasma in a Tokamak fusion reactor. If the gas is heated to a plasma, then it will expand. But the magnetic field acts as a containment field. As per Boyles Law, the pressure increases. Allow the plasma to escape via a venturi, as in a rocket, and the plasma will be accelerated to an “exhaust velocity”.
2. As per [1], but now move the magnetic field so that the containment volume decreases. This also increases the pressure. This would further increase the exhaust velocity.
You can argue this is not using a magnetic field to directly accelerate charged particles, and I would agree, but the magnetic field becomes a key component to of the acceleration mechanism. In the vacuum of space, there need not be any physical containment of the plasma, only the shaped magnetic chamber and exit, acting just like a chemical rocket. The efficiency is not going to be as high as with electric fields to accelerate the particle as in an ion engine, but the thrust is going to be much higher as the gas/plasma mass flow rate is higher.
The mechanism need not be continuous as in a rocket, but pulsed as in the V1 pulse jet. The propellant, e.g. water, is injected into the magnetic chamber, and microwaves convert the liquid to a plasma. The plasma is allowed to exit at high speed, and the next injection of water follows to repeat the cycle. The magnetic fields prevent corrosion of a physical engine. The main issue is the energy needed to maintain the magnetic field and its efficiency in in containing the plasma without leakage. If the magnetic field maintains a high field strength to keep the containment volume very small, the volume can be as narrow as the venturi throat before expanding out as the “bell” of the engine. It need not be on all the time, but cycled on/off with each pulse. [I expect superconducting coils would be used to reduce the energy losses of maintaining the magnetic field.]
3. A third method relies on the magnetic field to trap the charged particles as they spiral around the field lines. The field itself is then moved, dragging the trapped particles with it. At some point, the field is turned off, releasing the particles. I visualize this as a circular “engine” with the magnetic fields rotating around the circumference, releasing the particles in a pulse after the velocity is increased. This may be analogous to the rotating detonation engine where the detonation wave spreads around the circumference of the annular engine and accelerates until it exits at the exhaust, providing very efficient thrust.
IDK how VASIMR accelerates the plasma with magnetic fields, but I hope I have demonstrated that magnetic fields can be used to accelerate charged particles without violating physics, but using them in a different way .
Alex,
‘IDK how VASIMR accelerates the plasma with magnetic fields, but I hope I have demonstrated that magnetic fields can be used to accelerate charged particles without violating physics, but using them in a different way ‘
I think the magnetic field is only a containment method and does not accelerate the ions. I suspect the RF generators change polarity to push from the back (+) and pull (-) in a co-ordinated manner. The electrons then recombine as in an ion drive in the exhaust.
So, magnetic fields do not make electric motors. work? Or, maglev propulsion and rail guns cannot use magnetic fields to accelerate objects? Next, you will be saying that magnetic fields cannot induce electric currents.
All these cases require changing magnetic fields (or moving across fields).
Containing and manipulating a plasma is very involved in directing the movement of a plasma (or keeping it tightly compressed to initiate fusion in a Z-pinch arrangement).
It is somewhat analogous to saying that tire friction with the road surface is not important in vehicle motion. If the wheel doesn’t turn, then the friction is zero. The motor and axle rotation turns the wheel, but without friction, the vehicle won’t move, or conversely, cannot use the friction to brake when aquaplaning. If we think of the vehicle as a charged particle, without an engine it cannot move. But if we move the surface it is standing on (e.g. on a tow truck or delivery truck), the vehicle will move against the landscape. Moving or changing the magnetic field is like placing the vehicle on a moving surface.
Opps the RF coupler is helical and the magnetic field looks to be stationary.
Quote by Michael: “Opps the RF coupler is helical and the magnetic field looks to be stationary” The proton ions kinetic energy is still perpendicular to the magnetic field. The expansion of the field in the nozzle accelerates them in parallel motion. Electrons are accelerated by a magnetic field if they are perpendicular to the field like in solar flares helical, cork screw magnetic fields. Google AI
The energy for acceleration comes from the pressure and temperature via the RF generator in the chambers, the Magnetic field acts as guide.
Quote by Michael: “The energy for acceleration comes from the pressure and temperature via the RF generator in the chambers, the Magnetic field acts as guide.”
The energy comes from the RF generator not the temperature and pressure. Google AI. A magnetic field has to be changing like the expanding magnetic field to accelerate a charged particle which we see in the nozzle of VASIMR. The particles is not accelerated in a magnetic field unless it is changing, (expanding or collapsing). Ibid.
Quote by Michael, “Geoffrey, magnetic fields can’t accelerate particles only change their direction.” A magnetic field does accelerate electrons if they are perpendicular to the field which is what happens in VASIMR.
A magnetic field can only change the direction of a charged particle or you will be getting free energy, if the magnetic field is moving, say rotating, it can accelerate charged particles getting the energy from the rotation. A stationary charged particle for example would not move in a stationary magnetic field and therefore not be accelerated. I should have worded it better.
Sorry if it’s there and I missed it, but I don’t see that the scalar amount of thrust can be altered mid-journey; in particular it can’t be stopped, so putting the projectile into an orbit around anything seems challenging. Would it be chemically or mechanically feasible to fold it in half so the two halves smother each other or cancel each other out?