A recent article on Slashdot got me thinking.

While the article is about using electric turbopumps for moving fuel from tanks to engines during launch phase and is just a small, incremental advancement on current design, it got me thinking about using electric pumps in space flight phase in zone where solar energy is abundant - zones where we normally use ion engines.

We can store extra energy in fuel by pressurizing it. The extra pressure converts to extra delta-V. Water-bottle rockets work that way trivially; ion propulsion uses highly pressurized xenon to give it initial boost as it enters the acceleration chamber. It's nothing new.

What puts a limit on pressure available is durability->weight of the container. In case of some fuels, like solid propellant, compression is pointless. In other cases like liquid hydrogen, the pressure isn't all that high because the volume and pressure if kept highly pressurized, would necessitate unduly heavy container.

But I can picture a tiny solar-powered pump capable of creating very high pressures with minuscule throughput - possibly multi-stage piston pump with very small cross-section pistons, although choice of the exact pump design is an engineering detail, a subject of discussion and brainstorming later on.

The pump can create enormous pressure in a tiny volume of a small "buffer tank". Keeping these things small means they don't weight all that much despite handling maybe gigapascals of pressure. Releasing that pressurized propellant through a nozzle would give it a very high initial velocity. It can then still be further accelerated by a ion drive or react with similarly pressurized oxidizer creating additional thrust.

Nevertheless, we're considerably increasing initial speed of the propellant, without changing the amount of propellant used. Unlike with ion engines we're not limited by available voltages and length of the acceleration chamber - by scaling the device down we can achieve staggering pressures in a small and lightweight package, and as result - exit energies, without increasing fuel usage. And still we can use "second-stage acceleration" be it chemical, ion, or whatever we can think of, on the pressure-accelerated propellant.

Would it be a viable avenue of research or am I out on a limb here?

  • $\begingroup$ The new Vulcan 2nd Stage is planning on using this idea. $\endgroup$
    – tl8
    Commented Apr 19, 2015 at 23:29
  • 3
    $\begingroup$ @tl8: no. The Vulcan 2nd stage will use the exhaust of an internal combustion engine for thrust. The propulsive force comes from combustion, not pumping at high pressure. $\endgroup$
    – Hobbes
    Commented Apr 20, 2015 at 13:10
  • 1
    $\begingroup$ as always, there is a relavent xkcd ;) $\endgroup$
    – Baldrickk
    Commented Jun 6, 2017 at 9:37
  • $\begingroup$ I think this could work on a magnetoplasmadynamic-thruster $\endgroup$
    – Muze
    Commented Dec 27, 2018 at 23:19
  • $\begingroup$ Could this be applied to a MPD thruster? $\endgroup$
    – Muze
    Commented Jan 6, 2019 at 19:38

3 Answers 3


Edit: second attempt, my initial post was completely incorrect.

My intuition tells me you'll run into a limit on the exhaust velocity with a system like this. My initial thought, "The exhaust speed of a rocket is limited by the speed of sound" is incorrect, so where could the limit be? So let's see where we end up if we use the highest pressure possible and a liquid propellant.

This list of typical fluid or gas pressures found in various systems shows that water jet cutters use some of the highest pressures available outside a lab environment. This example uses a pressure of 648 MPa (6480 bar) and its exhaust velocity is given as Mach 4 (or 1.3 km/s).

As a comparison, the SSME has an exhaust velocity of 4.43 km/s, and ion engines have an exhaust velocity of 20-50 km/s.

Now, the efficiency of a rocket depends directly on its exhaust velocity. The thrust of a rocket is Isp * M (i.e. the specific impulse, directly related to exhaust velocity times the mass flow). So you can swap exhaust velocity for mass: half the speed means double the mass.

At current launch prices, that's not a tradeoff people are willing to make. So a fluid pump doesn't seem a promising approach.

There are other ways to use electric power for propulsion: a mass driver uses an electromagnetic rail system to throw inert mass at high speed. The only mass drivers in existence today are rail gun prototypes, with an exhaust speed of 2.4 km/s (as of April 2015). Better than the water jet, but still not as good as a rocket engine.

As for SF's suggestion to igniting a high-pressure jet: if you do this, you'll want to use the expansion of the combustion gases as much as possible, since that gives most of the thrust. When you enclose the jet in a tube, you can force all gases to expand in one direction. But now you're back to building a rocket engine, where the usable pressure is limited by the temperature in the combustion chamber.

Edit: responding to SF's suggestion to add an ion drive to the high-pressure jet: there's little point in doing that. If the ion drive applies 20 km/s of delta-V to the exhaust stream, the pump only provides 5% of that. And the pump has to be very heavy to withstand 650 MPa of pressure. You can put that weight to better use improving the ion drive.

  • 1
    $\begingroup$ If I'm reading the Wikipedia page you link to correctly your point that "The exhaust speed of a rocket is limited by the speed of sound" is incorrect. The exhaust speed would only be limited by the speed of sound if "upstream air pressure at atmospheric pressure and vacuum conditions downstream of an orifice". I don't know of any chemical rocket that has such low pressure exhaust gasses. $\endgroup$ Commented Apr 20, 2015 at 8:12
  • 1
    $\begingroup$ @ForgeMonkey: good catch, I've changed my answer. $\endgroup$
    – Hobbes
    Commented Apr 20, 2015 at 13:07
  • $\begingroup$ Glad to be of service. +1 for the interesting way of looking at the problem. $\endgroup$ Commented Apr 20, 2015 at 13:15
  • $\begingroup$ What about the aspect of creating a high pressure jet, and then igniting it? SF suggests this could mean higher Isp without any extra propellant. $\endgroup$
    – kim holder
    Commented Apr 20, 2015 at 15:02
  • $\begingroup$ What about propelling the high-pressure jet through electric field - ion drive propulsion? It's not limited by temperature but attainable voltages; makes no physical contact with the propellant. Also, pressurizing the propellant would also increase its temperature (adiabatic compression) and as result increase the ejection velocity (although that would be also available through plain "dumb" resistive heaters likely with higher efficiency.) $\endgroup$
    – SF.
    Commented Apr 20, 2015 at 15:35

The problem is 'exact pump design is an engineering detail' is a really big detail. The material science to accelerate material to high mach numbers is hard. Yes a 'pump' type system works well with 'low speed', but scaling it up a bit more only really gets you so far. The order of magnitude improvement in velocity (2 orders of magnitude higher energies and pressures) needed to get this close to chemical rockets strains credulity in the short term. To get much higher still, and justify the tiny TWR, at least another order of magnitude jump in velocity would be required. This is 10,000 times the current limit of the necessary pressures. It's hard to see this happening anytime soon.

Note, while its hard to give an intuitive 'why not', it may be helpful to think that to reach the pressures needed to get a water jet to compete with chemical rockets, would require near diamond anvil pressures even if no energy was lost.


The goal is to eject propellant mass at high velocity; the higher the propellant velocity, the less mass per unit impulse. But, the higher the velocity, the more energy required per unit impulse (energy per unit mass goes up with the square of the velocity).

Ion drives seek to maximize propellant velocity, period. The goal is to maximize specific impulse - make the most of the propellant mass they carry. For the kind of exhaust velocities they achieve, the amount of energy per unit propellant mass is far beyond what could be stored as chemical energy in the propellant itself. So, these systems look to solar or nuclear power. Such systems have limits on the amount of power they can provide, which limits the amount of thrust, despite the high specific impulse. These systems are all about accelerating ions to extreme velocities, far beyond what could be achieved by any mechanical pump, and the challenge is about the ultimate velocities achievable. Giving such a system an initial "kick" from a mechanical pump adds complexity without any benefit.

Chemical rockets seek to get the most out of the chemical energy stored in their propellants. They tend to go in either of two directions. One is maximum thrust - get the biggest possible push. Widely known examples of this include the kerosene/liquid oxygen first stage of Saturn V and Falcon 9, or solid fuel strap-ons of Space Shuttle. The other direction is maximum impulse - get the most delta-V. Widely known examples include the liquid hydrogen/liquid oxygen second/third stages of Saturn V and Falcon 9 or main engines of Space Shuttle. Efficiency in such systems comes from minimizing the conversions between different forms of energy. It starts out as chemical and becomes thermal in the the combustion chamber. From there, the nozzle aligns the random kinetic energy of the gas molecules so that it becomes mostly linear along the engine's axis. It's a very efficient conversion from chemical energy in the propellant to kinetic energy in exhaust flow. Using a pump to propel mass for thrust adds complexity and additional steps to convert energy from one form to another, incurring losses at every step.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

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