I've always found it interesting when we see real life taking its cues from fictional arts, especially when it comes to Science Fiction technology. There are plenty of examples of classic books, TV shows, or movies having introduced what was at the time a fantastic futuristic tool or capability. Then, decades later would be eventually developed and used in everyday society.

Some examples are George Orwell's book 1984, published in 1949, which predicted the wide use of PCs and the Internet.

This question relates to a more recent a series of books, and now TV show, The Expanse. Taking place centuries in the future, many of the technologies used in The Expanse are well beyond our current technological capabilities, some being more plausable than others.

My question relates to one of these examples which I hope is plausible, but I just don't know, the Epstein Drive.

The Epstein Drive is a commonly used propulsion system in The Expanse universe, one which makes interplanetary travel for humans both feasible and efficient. However, it seems to defy what I understand as current limitations in material/fuel efficiency capabilities. Unfortunately, I am not knowledgeable enough in this discipline to understand if something like the Epstein Drive is actually feasible, why it would be, or how it could be accomplished.

However, my asking how this could be feasible, which requires one to suspend reality momentarily in order to answer, seems outside the scope of this board. Instead, I will ask the following:

Are the concepts of the Epstein Drive, a fictional propulsion drive technology of the book and TV show The Expanse, in any way based on actual theorized scientific research which could render the fictional capabilities included in the story technologically plausable at some point in the future?

  • 6
    $\begingroup$ Its based on a fusion drive. You might be interested in perusing this: projectrho.com/public_html/rocket/enginelist3.php#epstein $\endgroup$
    – Polygnome
    Jun 14, 2020 at 18:36
  • 2
    $\begingroup$ Also the Project Rho torchship pages. Basically, it's 1. Fusion, and 2. Somehow contains an insanely large amount of energy in the plasma. $\endgroup$
    – ikrase
    Jun 14, 2020 at 19:01
  • 3
    $\begingroup$ Yea, fictional, made up stuff, but we sit here discussing things that were inconceivable 200 years ago. This a good going-over of the Epstein. toughsf.blogspot.com/2019/10/the-expanses-epstein-drive.html $\endgroup$ Jun 14, 2020 at 21:00
  • 6
    $\begingroup$ First time I've heard someone claim Nineteen Eighty-Four predicted the Internet. $\endgroup$
    – OrangeDog
    Jun 15, 2020 at 10:22
  • 4
    $\begingroup$ The Epstein drive research project didn't kill itself! $\endgroup$
    – Dai
    Jun 16, 2020 at 6:02

2 Answers 2


From what I've been able to find, the "fluff" is that the Epstein Drive electromagnetically accelerates the exhaust, improving propellant consumption over a "standard" inertially-confined fusion drive. Electromagnetically accelerating plasma is certainly physically plausible...real world ion and plasma thruster designs do just that. The problem is that electromagnetically accelerating plasma takes power, with the power requirements scaling with the square of the exhaust velocity. These drives are limited to low amounts of thrust not because they can't be built to handle more, but because current power sources don't have enough power density to let them do better.

The whole idea of a fusion rocket is to get at most of that power by fusion. The fusion heats the propellant directly and you don't have to generate and handle the power output in electrical form, you just need enough to keep the fusion reaction going and run the magnetic nozzles/etc. Fusion would have extremely high energy density, and hopefully, not having to convert it all into electrical form would let you achieve higher power densities and more thrust...you just have to fuse your fuel fast enough and not burn yourself up in the process.

An "Epstein Drive" that electromagnetically accelerates fusion rocket exhaust would improve propellant consumption as described, but it would also add terawatts to the electrical power requirements and render the fusion portion largely superfluous. Even just doubling the velocity (and halving the propellant requirements) with perfect efficiency would mean that the "accelerator" would account for 3/4 of the power output (and a much bigger fraction of the power input). Thus, in reality not all that useful.

  • 2
    $\begingroup$ I think this is a little bit of a clumsy description - it's all correct, but realistically anything with performance approaching The Expanse's Epstein Drive is just a straight fusion thermal thruster. $\endgroup$
    – ikrase
    Jun 14, 2020 at 21:42
  • 1
    $\begingroup$ While just using a higher power fusion "torch" is the realistic way (given sufficiently advanced fusion technology) to achieve this performance, the Epstein Drive as described doesn't appear to just be a more powerful fusion drive. $\endgroup$ Jun 15, 2020 at 14:08
  • $\begingroup$ @BrianDrummond: The distinguishing features of the Epstein Drive are drastically increased thrust and specific impulse, not improved control, reduced surges, or thrust vectoring. To achieve this, it does have to add exhaust velocity, and far more than a few percent...several hundred percent, in fact. $\endgroup$ Jun 16, 2020 at 11:17
  • $\begingroup$ @ChristopherJamesHuff This is feasible though, it all depends on the use of the fusion reactor and how much you can tweak it. First, to increase ISP, you would need to harvest more energy than normal from the reactor to run the accelerator. Second, to increase TWR, you would need to run more propellant out the engine bell (essentially just increase regularity of fusion activation.) Neither of these can achieve the goal by themselves, but it's quite far from impossible. $\endgroup$ May 1, 2023 at 7:57
  • 1
    $\begingroup$ @VladimirSilver The Epstein drive wasn't depicted as being a few percent more efficient or even doubling the performance, it allowed continuous acceleration for weeks on end. This is equivalent to adding several zeros to the power output figures. If the difference is just that the Epstein drive is actually using all of the fusion energy to produce thrust and older drives were wasting most of it, the fluff is more how those older ships weren't outright vaporized. $\endgroup$ Jan 16 at 2:30

The properties of the fictional Epstein Drive are:

  • Very high ISP with an exhaust velocity at least several percent of the speed of light
  • Very high thrust while achieving that ISP or near to it, enough for 5+ g burns.
  • Fusion powered.
  • Based on the first two parameters, the thrust power is insanely high - this is not a fission sail, NTR, or low-power fusion thruster.
  • The ships are not dominated by heat radiators, which means that of all the immense power handled by the engine, not even a very small fraction needs to be dissipated as waste heat by radiators.

The first four properties are within the theoretical capabilities of nuclear fusion. The fifth is very, very difficult to achieve, though it is not fundamentally forbidden by theoretical physics. It is not clear how it can be pulled off.

The basic idea is that nuclear fusion fuel is compressed and heated by some method so that it fuses, and then the resulting ultrahot plasma is allowed to expand outward through a rocket-nozzle-shaped magnetic field. This turns heat and pressure into motion, just like in a chemical rocket.

The slight problem is that while we have achieved breakeven with fusion, nobody has even approached doing it in a useful way, let alone with a compact reactor that can fit in a spaceship.

Most realistic designs that have been considered for this (and which have less performance than the Epstein Drive) use inertial confinement fusion, where discrete pellets of fuel are flash-compressed with converging lasers, particle beams, or pulsed magnetic fields (Z-pinch), causing them to basically explode. Most realistic designs also have the exploding pellets and the magnetic field nozzle outside of the spaceship so that waste heat can be radiated away without heating up a reactor wall.

A good example is the D-T VISTA concept. This doesn't reach anything like the Epstein Drive performance, and it's basically shaped like a cone made of heat radiators. We could do a lot better if we had quasi-magical forcefields able to reflect radiation and channel plasma.

For a lot more information (appropriately for this question, aimed at fiction writers but very much based in factual science), check out Project Rho's page on torchships (a name often used in science fiction for spacecraft with this level of performance).

  • 1
    $\begingroup$ Nitpick: "nobody has yet achieved breakeven with fusion", I believe some of the laser based fusion experiments have done so, possibly even some of the tokamaks, but the former in particular lack 'sustainability'. $\endgroup$
    – MikeB
    Jun 15, 2020 at 11:18
  • 3
    $\begingroup$ @MikeBrockington The inertial fusion experiments haven't done so. Ed Moses, the project manager, claimed the NIF achieved "breakeven" when it produced more fusion energy than the target absorbed from the lasers, but that's not what breakeven means (and not even how Moses previously used the term). They produced 14 kJ from a 1.8 MJ pulse. They would have needed 129 times as much output to reach breakeven (and never got anywhere close to ignition, which the NIF was built to do). $\endgroup$ Jun 15, 2020 at 14:23
  • 2
    $\begingroup$ Tokamaks are doing much better. Back in 1998, the JT-60 managed to maintain plasma conditions that would have produced 1.25 times as much fusion power as the JT-60 consumed, if it'd been using a deuterium-tritium plasma. They don't actually use tritium because it's radioactive and the JT-60 is for doing research on the plasma control side of things. ITER will probably work as expected, it's just abominably expensive, horribly mismanaged, and proceeding at a glacial pace (startup in 2025, fusion experiments starting 10 years later). $\endgroup$ Jun 15, 2020 at 14:30
  • 2
    $\begingroup$ Season 4 Episode 6 ("Displacement") shows a pellet-based design for Rocinante's reactor. Whether this is the drive or a separate power reactor isn't clear (from what I recall). $\endgroup$ Dec 18, 2021 at 22:02
  • 1
    $\begingroup$ @ikrase From about 08:32 to 08:34: Pellet enters from top, ignition attempt using lasers, failed ignition, pellet hits floor, pellet has bounced off – two other pellets are also visible in the chamber. From the subsequent dialog it sounds like the Roci uses a single reactor for both power and propulsion. $\endgroup$ Dec 20, 2021 at 3:43

Not the answer you're looking for? Browse other questions tagged or ask your own question.